Organofluorines (Handbook of Environmental Chemistry)

CHAPTER 1

Principles and Topical Applications of 19F NMR Spectrometry Paul D. Stanley Syngenta, Jealott’s Hill International Research Centre, Bracknell, RG42 6ET, UK E-mail: [email protected]

Fluorine occurs naturally in only a few organic compounds. In the chemical industry fluorine is a substituent of choice used to modify synthetic substrates in expectation of conferring beneficial physical properties. When incorporated, the spectroscopic properties of fluorine make it a useful tool to aid the structural elucidation of the derived substances. Fluorine is generally resistant to degradation and offers interesting possibilities as a probe for the determination of chemical residues and the investigation of metabolic processes. This chapter offers an overview of the ways in which the many recent advances in NMR technology can be exploited to derive useful qualitative and quantitative chemical information from fluorinated substances. 19F

NMR Experimental procedures, Chemical shifts and coupling constants, Structure elucidation, Metabolites, Quantitative analysis, Illustrative applications of 19F NMR, Fluorine tags Keywords.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3

1

Introduction

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8

General Principles of NMR Spectroscopy . . . . . . . . . . . . Application of NMR Spectroscopy to Chemical Structure Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . NMR Experiments for Chemical Structure Confirmation . . . NMR Experiments for Chemical Structure Elucidation . . . . Computational Methods . . . . . . . . . . . . . . . . . . . . . . 19F NMR Spectra – Chemical Shifts and Coupling Constants . Acquisition of 19F NMR Spectra – Instrumental Considerations Mass Spectrometry and NMR as Complementary Procedures .

2

Topical Applications of 19F NMR Spectrometry . . . . . . . . . . . 20

2.1 2.2 2.3 2.4 2.4.1 2.4.2

Structure Elucidation – Tefluthrin, a Case Study LC/NMR and Metabolite Identification . . . . Quantitation of Metabolites . . . . . . . . . . . Solid State NMR Applications . . . . . . . . . . Gels . . . . . . . . . . . . . . . . . . . . . . . . Semi-Solids . . . . . . . . . . . . . . . . . . . .

. . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . . . . .

. . . . . .

3

. 5 . 6 . 7 . 9 . 11 . 16 . 18

. . . . . .

20 36 45 49 49 50

The Handbook of Environmental Chemistry Vol. 3, Part N Organofluorines (ed. by A.H. Neilson) © Springer-Verlag Berlin Heidelberg 2002

2 2.4.3 2.5 2.5.1 2.5.2 2.5.3 2.6

P.D. Stanley

Solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biochemical Studies . . . . . . . . . . . . . . . . . . . . . . Labels and Tags . . . . . . . . . . . . . . . . . . . . . . . . Magnetic Resonance Imaging (MRI) . . . . . . . . . . . . . Drug Design . . . . . . . . . . . . . . . . . . . . . . . . . . Determination Optical Purity – Fluorinated Derivatisation Reagents . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

50 51 51 52 52

. . . . 53

3

Conclusions and a Prospective View . . . . . . . . . . . . . . . . . 55

4

Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

5

References

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

List of Abbreviations dimensionless unit (chemical shift scale) spin-spin coupling constant (Hz) parts per million (chemical shift scale) parts per billion (concentration) Attached Proton Test COrrelation spectroscopy via LOng range Coupling COrrelation SpectroscopY Distortionless Enhancement by Polarisation Transfer long range heteronuclear correlation experiment using Bird pulses GARP Globally optimised Alternating-phase Rectanglar Pulses HETCOR HETeronuclear CORrelation HETGOESY HETeronuclear Gradient Overhauser Enhancement SpectroscopY HMBC Heteronuclear Multiple Bond Correlation HOESY Heteronuclear Overhauser Enhancement SpectroscopY HSQC Heteronuclear Single Quantum Coherence INADEQUATE Incredible Natural Abundance DoublE QUAntum Transfer MRI Magnetic Resonance Imaging NOESY Nuclear Overhauser Enhancement SpectroscopY ROESY Rotating frame Overhauser Enhancement SpectroscopY TOCSY TOtal Correlation SpectroscopY WALZ a broadband decoupling pulse sequence WURTZ a broadband decoupling pulse sequence

d J ppm ppb APT COLOC COSY DEPT FLOCK

Principles and Topical Applications of 19F NMR Spectrometry

3

1 Introduction This chapter provides an account of the overall principles of NMR spectroscopy with particular reference to the acquisition and interpretation of 19F NMR data. It introduces the concepts of the correct selection of instrumental parameters, NMR pulse methodologies (such as COSY), instrumentation, and the application of LC-NMR interfaces. Rather than provide a summary of numerous applications, the structural determination of a specific compound is discussed in detail so as to illustrate the importance of 19F as a probe and reporter of chemical structure. In many biochemical applications, a complete knowledge of the structure of metabolites is of primary importance and thus is discussed in terms of qualitative identification, quantitation, and determination of optical configuration. Attention is drawn to applications in analytical chemistry and microbiology, both discussed elsewhere in this volume, where no theoretical background is presented, and to the fact that many other industrial applications involving perfluorinated compounds are not discussed. In this section we take a representative selection from the many experiments and tools that can be used to realise the potential of NMR as a generic tool to solve chemical structures. This account is biased by the substances handled within our laboratory and thus does not consider the large amount of work on perfluorinated materials that are important elsewhere in the chemical industry. The presence of 19F atoms, and fluorinated groups, in chemical structures are powerful probes of subtle structural information that can be accessed using many standard NMR methods. For complex, or unknown, substances where a result is needed against a tight deadline, computational methods may aid the process of structure elucidation. 1.1 General Principles of NMR Spectroscopy

Within the context of this chapter, detailed discussions of the physics of the NMR phenomenon and the ever increasing number of NMR experiments that may be applied to elicit unique pieces of structural chemical information are not appropriate. This information will be found readily in the many excellent general and specialist NMR texts available. Those currently popular amongst the chemists in our laboratory include the works by Hore [49], Sanders and Hunter [89] and the late Andy Derome [31]. The recent text by Claridge [27] promises to become the preferred general work for NMR users and practitioners alike. In reviewing this text Bladon [12] points out that the author caters for three classes of reader: (1) NMR users who have no interaction with spectrometers, (2) those who have also been trained to acquire their own data and (3) the non-specialist responsible for instrument maintenance. The identification of this last classification of user, together with hints on how to negotiate with manufacturers surely makes the book unique. This book gives a clear, readable and comprehensive introduction to NMR and is thoroughly recommended by the reviewer. It should be noted that none of these texts are specific to 19F NMR.

4

P.D. Stanley

Though out of print, the books by Dungan and Van Wazer [34] and Mooney [65] are worth searching out since they deal specifically with the topic and are still widely referenced today. The feature that is common to all NMR spectra, and that accounts for its widespread application to chemical structure determination, is that NMR spectra report substances as sets of optionally connected chemical substructures. The signals from each substructure give rise to a characteristic NMR fingerprint. These fingerprints are reproducible and generally predictable by nature. A retrospective assembly of all the identified substructures may reveal several structural alternatives. Most often, these can be discriminated by further NMR experimentation. The following pieces of information are readily extracted from NMR spectra and are the building blocks that lead to the successful elucidation of chemical structures. – Chemical shift (d) – NMR spectroscopy differentiates between atomic nuclei in different chemical environments. Nuclei in different chemical environment have signals with different positions along the x-axis of the spectrum (chemical shift). Nuclei in similar chemical environments have similar chemical shifts. Chemical shifts are normally measured relative to a small amount of a known compound (internal standard) added to the test sample. – Signal intensity – essentially, the NMR experiment is quantitative and with due care the areas of NMR signals can be measured (by electronic integration) to determine the relative number of nuclei giving rise to each signal. In mixtures, the ratio of peak integrals can be used to estimate molar composition; in pure materials, the same ratio can be used to propose an empirical formula. – Spin-spin coupling (J) – NMR signals may appear as sharp peaks or have characteristic splitting patterns that result from the magnetic interaction of one nucleus with another. Spin-spin coupling may be observed between nuclei of the same kind (e.g. 19F–19F) and between nuclei of different kinds (e.g. 19F–1H or 19F–13C). Chemically isolated nuclei have no such interactions and appear as single peaks. – Relaxation times (T1 –T2) – when radiofrequency energy is absorbed to generate an NMR signal it must be allowed to dissipate before further experimental data can be acquired. There are two time-dependant mechanisms by which this energy is dissipated: either to the environment (T1) or by interaction with local non-fluctuating magnetic fields (T2). Knowledge of the approximate relaxation times of different nuclei in the same sample is a key factor in achieving accurate quantitation of NMR spectra. Most fluorinated compounds of interest to the chemist will contain 19F, 1H and 13C atoms. The nuclear properties of these nuclides are shown in (Table 1). At first glance, there is little to differentiate between the nuclear properties of 19F and 1H in terms of sensitivity and natural abundance, additionally both nuclei also have T1 relaxation times that are short enough to allow the meaningful quantitation of the signals. With modern medium-field NMR instruments, it is possible to obtain good quality 1H and 19F spectra on sub-milligram amounts of

5

Principles and Topical Applications of 19F NMR Spectrometry Table 1. Nuclear properties of 1H, 13C and 19F

Nuclide

Nuclear spin, I

Relative sensitivity

Natural abundance (%)

n0 (MHz) at 9.4 T

g (107 rad T—1 s—1)

1H

1/2 1/2 1/2

100 1.59 83.3

99.98 1.11 100.00

400.0 100.4 376.4

26.752 6.728 25.181

13C 19F

substances (MW<500 Da) within a few minutes. In contrast, 13C nuclei have both low sensitivity and low natural abundance. 13C nuclei often have long T1 relaxation times that make the routine use of 13C spectra for quantitation difficult when small amounts of sample are available. Despite these characteristics, 13C spectra are key components of chemical structure elucidation. If properly equipped with a low volume NMR probe, a modern medium-field NMR instrument will deliver good quality 13C spectra on sub milligram amounts of substances with a molecular mass <500 Da in less than 1 h. NMR spectrometers are tuned to acquire spectra from only one kind of nucleus at a time, thus when tuned for 19F observation only signals from 19F containing substances are observed. 1.2 Application of NMR Spectroscopy to Chemical Structure Determination

The high sensitivity and enormous chemical shift range of 19F NMR nuclei make 19F NMR an attractive proposition. Although 19F NMR spectra are exceptionally selective in terms of establishing and quantifying the number of fluorinated species present, the chemical shift ranges for the different fluorinated functionalities overlap extensively. Alone, therefore, 19F NMR data often do not add a great deal to the process of chemical structure determination. The inclusion of fluorine substituents in substances, however, provides a powerful handle that enhances the process of structure determination through consideration of the resulting spin-spin coupling pathways established from the fluorine entities to nearby 1H and 13C atoms. There are four levels of refinement for the determination of chemical structure: – Primary chemical structure (elemental composition) is normally determined using elemental analysis or high resolution mass spectrometry (HRMS). – Secondary chemical structure (functional groups and connectivities within the molecule) is normally determined using NMR, vibrational spectroscopy and, to a lesser extent, MS. – Tertiary chemical structure (spatial arrangement of sub-structural motifs) is normally determined using NMR or X-ray analysis. – Quaternary structure (existence of functional domains) is normally determined using NMR or X-ray analysis.

6

P.D. Stanley

Although X-ray crystallography is undeniably the single method of choice for the determination of chemical structures, it is often difficult to obtain crystals suitable for X-ray analysis. The application of a combination of alternative instrumental techniques then becomes necessary. The text by Crews et al. [29] presents a logically ordered approach to the solution of the chemical structures of pure natural products by the application of multiple instrumental techniques and is to be recommended as a reference strategy document. NMR spectroscopy is probably the most vital step in the process. This is not always because it is the best, or most cost effective, technique but because of a combination of the certainty of the results, the relative ease of spectral interpretation and a familiarity with the technique on the part of most chemists. Familiarity with the technique is a powerful stimulus but it should be part of every chemists creed to take regular “reality checks” to ensure that the analyses being performed are the most appropriate and that they really are answering the questions being posed. Although chemical intuition is an invaluable tool routinely called upon it may, occasionally, be misleading. Knowledge of the chemical route producing the substance being investigated or an educated guess as to the metabolic process leading to expression of the substance being analysed is almost always useful, but should not be used as primary information. The real structural information is there to be had in the NMR data sets on the desk in front of you, and this should form the primary source for any structural conclusions. 1.3 NMR Experiments for Chemical Structure Confirmation

Modern NMR technology gives access to an enormous range of experimental methodologies that probe chemical structure and make the process of structure elucidation by this technique a truly stimulating occupation. The volume by Braun et al. [15] contains details of how to set up and measure 163 separate NMR experiments. At a practical level the volume is most useful to owners of Brüker NMR systems, but it serves as an invaluable reference text for all NMR instrument users. Details of how to set up the experiments described later may be found in this volume; users of instruments from other manufacturers (e.g. JEOL and Varian) will find similar guidance in their instrument manuals. Selecting the correct post acquisition processing needed to display the NMR spectra and massaging them to give the “best” results is an art in itself. The volume by Bigler [11] is a structured introduction to the post-acquisition processing of NMR spectra that is, again, aimed at owners of Brüker NMR systems. This is an interesting book to work through since it contains carefully selected examples of good NMR data and an academic copy of the Brüker WinNMR program so that the learning can take place away from the spectrometer. Transferring the learning experience to the instrument software of other vendors is not without problems but these are relatively minor compared to the benefits. There are two common situations to which the tactics of organic structure determination are applied. The simplest case involves proving that the sub-

Principles and Topical Applications of 19F NMR Spectrometry

7

stance at hand is identical to one previously reported. The other involves establishing the structure of a completely new substance. In both cases a command of the methods for translating spectroscopic data into structures is essential, a point we will come back to later. In our laboratory the first stage is usually to acquire a simple one-dimensional (1D) proton NMR spectrum. At first sight this spectrum will give a good idea of the complexity of the substance being investigated, the purity of the sample and give some high level pointers to the type of substance being examined (e.g. aliphatic, aromatic etc.). A particular structural isomer can often be distinguished by consideration of the patterns arising from spin-spin coupling; the presence of selected other nuclei (e.g. 19F) can be inferred by the presence of spin-spin couplings not explainable by the consideration of interactions between protons alone. If the presence of 19F substituents is suspected, or expected, the 1D 19F spectrum yields crude structural information about the chemical types of fluorine present and, more importantly, a sensitive doublecheck on the purity of the sample with respect to other fluorinated materials. For the purposes of proving that the substance being investigated is identical to one previously reported the chase usually stops here. If, however, the structure of the analyte is unknown, or if confirmation of a previously reported structure is required, this is the starting point for a more detailed study. Acquisition of 1D 13C and DEPT [9] or APT [78] will determine the number of carbon atoms present and label each in terms of the number of protons attached. In conjunction with the 1H and 19F spectra a list of proposed substructures can be constructed. On the basis of an empirical formula deduced from the molecular weight of the substance from MS, a series of two-dimensional structures can be written. 1.4 NMR Experiments for Chemical Structure Elucidation

Depending on the apparent complexity of the substance, a range of 1D and twodimensional (2D) NMR experiments can be planned that will experimentally verify or deny the existence of the connectivities required by the proposed structures. For less complex molecules 1D NMR experiments such as homonuclear decoupling (probing 1H–1H or 19F–19F interaction) and heteronuclear decoupling (probing 1H–19F, 1H–13C or 19F–13C interaction) experiments are often sufficient to establish the chemical substructures actually present. In the case of more complex substances 2D NMR methods such as COSY [5] (probing geminal and vicinal 1H–1H interactions) or TOCSY [16] (correlating all protons in a particular substructure) are usually applied. It should be noted that complexity is not always a function of molecular size; the spectra of small molecules are often sufficiently overlapped to preclude the use of spin-decoupling experiments. In principle, these 2D methods display every spin-spin coupling present in the molecule, rather than establishing them one at a time using spin-decoupling experiments – the habit of not using 2D methods for small molecules stems from the long time (typically several hours) required to set up, collect and process the 2D data sets. The introduction of gradient selected versions of COSY [51] and

8

P.D. Stanley

TOCSY [51] experiments, together with the increased stability of modern NMR instruments, removes this barrier since the 2D data can now be collected in tens of minutes. Correlations between different types of nuclei (e.g. 1H–19F or 1H–13C) are readily established using HETCOR [42] (one bond correlations) or COLOC [58]/FLOCK [20] (long-range correlations) experiments. These spectra can take many hours to acquire, especially if limited amounts of material are available. The introduction of inverse geometry NMR probes has increased the sensitivity of the complementary proton detected experiments HSQC [13] (one bond correlations) and HMBC [17] (long range correlations) to such an extent that the direct observation experiments are falling from popular usage. For the analysis of samples, where a reasonable amount of material is available, the gradient selected versions of HSQC [57] and HMBC [7] can often cut the acquisition time of the spectra by a factor of four. The edited gradient selected HSQC [77] experiment is the equivalent of heteronuclear correlation with signal of the low frequency nuclei being edited in DEPT fashion with “even” and “odd” multiplicity carbons being separated by the phase of the signal in the 2D display. Since this experiment also correlates carbon and proton chemical shifts, inspection of the proton spectrum usually removes the uncertainty as to whether the “even” multiplicity signals are methyl or methine carbons. In practice, unless there are unusual combinations of chemical shifts present, heteronuclear correlation experiments often do not offer useful information with reference to chemical structure determination. These spectra are, however, an indispensable part of the inevitable process of the complete retrospective unambiguous assignment of the spectra. To establish the basic stereochemistry of molecules (e.g. the E–Z configurations of alkenes) the nuclear Overhauser (NOE) effect can be profitably applied. NOE depends on the dipolar relaxation of one nucleus by another. The effect is proportional to the inverse sixth power of the distance between the participating nuclei and is thus sensitive to conformational changes. There are NMR experiments designed for either homonuclear or heteronuclear applications. The basic 1D NOE difference experiment [72] collects a spectrum with external radio frequency irradiation at the peak of interest followed by a spectrum without irradiation. When these spectra are subtracted the difference signals can be correlated with proximity. 2D versions of this experiment, NOESY [73] and its gradient selected version [107], are often used to study larger molecules; it is an essential method for determining the peptide conformation (tertiary structure) of proteins. NOESY cross peaks may “vanish” for molecules with molar masses in the range 1000–3000 since the sign of the NOE effect changes sign depending upon the molecular correlation time. The NOE is always positive under the spin-lock conditions that are used in the ROESY [14] experiment. Without special spin-lock conditions [52, 53] ROESY experiments may also show TOCSY correlations that may lead to confusion. Whereas most of the experiments mentioned above appear to focus on 1H and 13C NMR it should be remembered that, when 19F is incorporated into the chemical structure, its interaction with other nuclei through spin-spin couplings is usually far more useful as structural handles than simply observing

Principles and Topical Applications of 19F NMR Spectrometry

9

the 19F chemical shift. The clean baselines typical of 19F spectra often make it the nucleus of choice for the determination of the proximity relationships which lead to the successful determination of stereochemistry and 3D structures. So it seems that the collection of a carefully selected range of NMR spectra from a sample will yield information that should make it possible to assemble a list of candidate chemical structures consistent with the observed data that can be refined to a single structural entity by further experimentation. Whilst this is undeniably true, the process is far from easy. When results are required against a short deadline it may be appropriate to seek assistance from computational aids. 1.5 Computational Methods

The application to the treatment of spectroscopic data using computational methods is presently not well developed. The demands placed by the new areas of chemistry, such as solid phase synthesis and combinatorial chemistry, that have the potential to produce many thousands of samples with apparent ease have resulted in a resurgence in interest in the topic of 1H NMR spectrum prediction and appropriate display software [1]. The prediction of 1H NMR spectra is fraught with difficulties due to the unpredictable effects caused by throughspace effects, changes in NMR solvent, etc. These effects are not so important with respect to 19F and 13C spectra, due to the wide chemical shift ranges involved, and it is possible to predict the spectra of these nuclei with a good degree of precision. The computer software necessary to do these calculations is available commercially [3, 24]. The products from ACD and Chemical Concepts (SpecInfo) are based on enhanced applications of the sub-structural coding routine devised by Bremser [18]. Using this scheme, each atom (node) in the molecule is assigned a code based upon its chemical environment, described in terms of the number of bonded atoms together with their bonding scheme; it is extended to consider atoms up to four chemical bonds away. The extent of each code was determined by the computer word-length available at the time. A peculiar limitation of this approach is exemplified with reference to aromatic materials, whereby the nature of a substituent para to any particular node is not recognized. Each node is then assigned a chemical shift during spectrum assignment, the resulting correspondences being contained in an inverted database. When used to predict the spectrum of a compound not exemplified in the database, the program disassembles the novel structure into sub-structural units (the nodes) and seeks matches in the database. Exact node for node matches are therefore reported with high confidence and values for non-exact matches with a lower confidence. Whilst originating in the prediction of 13C NMR data, both ACD and SpecInfo offer 19F predictions based upon large authenticated databases. ACD and SpecInfo quote access to 15,000 and 23,500 records respectively. As well as prediction of the spectra of novel substances, both programs offer traditional chemical shift line searches to identify substances or sub-structures with similar chemical shifts and full sub-structural database searching to access the complementary data stored along with the

10

P.D. Stanley

chemical shift values. These typically include coupling constant information, experimental conditions and literature references. Both these programs are remarkable in as much as they accurately predict the spectra of substances that have been suggested by the analyst, who then uses experience to judge the goodness of fit between the experimental and the predicted data. A more attractive option, however, is to allow the spectral data to speak for itself. This proposal is not so wild as it may seem. There are several prototype programs available commercially that aim to apply logic to the assembly of spectroscopically identifiable substructures. With the present level of development, each of the software toolkits described below produces creditable results under idealized conditions. In our laboratory the quality of data we are able to generate routinely is often not of sufficient quality to guarantee success. Once in possession of a working empirical formula the Molgen [10] program will generate all chemical structures consistent with the empirical formula. The enormous numbers of proposed structures are constrained using a combination of 13C multiplicity and a good list – bad list principle. When the list of candidates has been reduced to a manageable number (say less than 200) the structures can be imported directly into SpecInfo; the SpecInfo program automatically predicts the 13C spectrum of each structure and presents a list of goodness of fit with the experimental data. Understanding the principle of good – bad lists is key to succeeding with this suite of programs. Pretsch’s paper [91] offers an authoritative summary of the strategy of good list – bad list formulation. There are commercial programs that go some way towards offering a fully automatic interpretation of spectroscopic data, and the products from ACD [2], Spectrum Research [99] and ScienceSoft [92] are worthy of note. ACDStructure Elucidator allows automatic or manual import of spectral data from other ACD software products. At present only 1D NMR data (plus IR and MS data) can be used; the 2D NMR module is presently in beta testing. The experimental data is subjected to a sub-spectrum search in the main ACD prediction database followed by a structure generation stage based on overlapping fragments by their common atoms, a process which does not require knowledge of the molecular formula. Alternatively, a unique “spectrum filtration” procedure based on spectral features can be used to constrain the list of possible structures. The generated structures are then presented in order of goodness of fit with the data. This program works well with structures that have partial or close precedents in the main ACD prediction database. NMR-SAMS (formerly CISOC-SES) from Spectrum Research was developed by Yuan et al. [80, 81] and utilizes NMR spectra prepared for analysis by the corresponding SpecMan program. Data from a wide range of 1D and 2D NMR experiments together with molecular formula information and structure fragment information from UV and IR methods generate structures compatible with the spectral data. The program automatically generates fixed bonds and “building blocks” using COSY, 13C and DEPT – APT information. The program proposes unique structures when given substantial data and partial structures when given sparse data. Since this program does not consider existing information, it has the potential to solve unprecedented chemical structures. In our hands the program performed well so long as the data was of high quality, but

Principles and Topical Applications of 19F NMR Spectrometry

11

was unforgiving when expected correlations were absent due to poor sensitivity etc. ScienceSoft’s NMRAnalyst is also available as FRED from Varian Inc. The package automates the interpretation of 2D NMR data. At inception the program dealt only with INADEQUATE (13C–13C COSY) data. This NMR experiment is very insensitive and the first stage is to use novel mathematical procedures (Full Reduction of Entire Data-sets) to determine the true 13C–13C COSY correlations present in the spectrum, thus establishing the carbon skeleton. AssembleIt is the NMRanalyst extension module for correlating the information extracted from different spectrum types. Currently, AssembleIt supports the challenging combination of short-range and long-range heteronuclear, DEPT, and 2D INADEQUATE information to derive molecular carbon skeletons. It is planned to extend AssembleIt for the complete structure elucidation of unknown compounds and for the 3D-structure determination of proteins and other bio-molecules. In this section we have mentioned some of the experiments and tools that can be used to realise the potential of NMR as a generic tool to solve chemical structures. The presence of 19F atoms, and fluorinated groups, in chemical structures are powerful probes of subtle structural information that can be accessed using many standard NMR methods. For complex, or unknown, substances where a result is needed against a tight deadline, computational methods may aid the process of structure elucidation. 1.6 19F NMR Spectra – Chemical Shifts and Coupling Constants

As indicated above, 19F nuclei have favourable NMR properties and since fluorine-containing substances are important industrially, 19F NMR has been actively studied since the discovery of the NMR experiment. Consequently, there are many thousands of literature references detailing the applications of 19F NMR to structural chemistry problems that span more than 40 years of industrial and academic research. Vast compilations of experimental data have been reported in the series Annual Reports on NMR Spectroscopy [66, 55, 56, 39, 21, 111, 112] and Progress in NMR Spectroscopy [35, 36] but many of the substances reported in these references are highly fluorinated and, as such, are not relevant to this chapter. Since these represent an enormous amount of information spread across several volumes these references are not convenient to use but, undeniably, include some data not exemplified in the databases supporting the 19F spectrum prediction approach described above. An additional complication with this literature is that it contains 19F chemical shifts reported using several different NMR chemical shift conventions, the relevance of which is discussed later. The only convenient way to access more recent data is to use conventional literature searching methods in the primary journals. The many journals dedicated to NMR techniques generally focus upon details of new applications for 19F NMR whereas the chemical literature is relied upon to report spectroscopic properties of new fluorinated substances. Everett [38] has produced the most recent authoritative review covering the NMR spectrometry of

12

P.D. Stanley

Table 2. Chemical shifts of alternative reference compounds based on CCl3F=d 0.00

Substance

Chemical shift (d)

Benzenesulfonyl fluoride – C6H5SO2F Trichlorofluoromethane – CCl3F Dichlorodifluoromethane – CCl2F2 a,a,a-Trifluorotoluene – C6H5CF3 1,2-Difluorotetrachloroethane – CCl2FCCl2F Methyl trifluoroacetate – CF3CO2CH3 Trifluoroacetic acid – CF3CO2H 1,1,1-Trichlorotrifluoroethane – CCl3CF3 Hexafluoroacetone – CF3COCF3 a Fluorobenzene – C6H5F 1,4-Difluorobenzene – FC6H5F Hexafluorobenzene – C6F6

+65.50 0.00 – 6.90 –63.90 –67.30 –74.21 –78.50 –82.20 –84.60 –113.15 –120.00 –162.90

Chemical shifts measured in non-polar solvent. a Note that, in aqueous solution, hexafluoroacetone exists as the hydrate (CF C(OH) CF with 3 2 3 a chemical shift of d –92.80.

organofluorine compounds. In this review, the author discusses and references sample spectroscopic data from a carefully selected range of fluorinated substances that are likely to be of interest to organic chemists. The references are backed up with reproductions of many high quality 19F spectra that illustrate the, sometimes unusual, appearance of 19F–19F and 19F–1H spin-spin coupled systems. The internationally agreed internal reference substance for 19F NMR is trichlorofluoromethane (CCl3F) and a representative selection of some of the alternative substances used in the past are shown in Table 2, in which their chemical shifts are shown referenced to (CCl3F). External referencing is occasionally reported. This is achieved by either insertion of a sealed sample of the referencing substance into the analyte prior to spectroscopic analysis or by simply assigning the reference position to a data point that represents the chemical shift of the standard as determined experimentally. This mode of referencing leads to significant errors in chemical shift measurement, particularly if the external reference substance is dissolved in a different solvent to the analyte. Solvent change induced shifts of >5 ppm are not unusual. 19F NMR data should be acquired and reported using the guidelines laid down by the International Union for Pure and Applied Chemistry (IUPAC) as shown in Appendix 1. The IUPAC guidelines ensure that signals for all commonly occurring organofluorine substances will be properly observed. However, these conditions demand a very wide spectrometer sweep width that may result in a problem with processing the data and that will necessitate the re-running of the spectrum over selected narrower areas if spin-spin couplings are to be measured with any precision. When CCl3F is used as an internal reference, the majority of commonly occurring organofluorine residues (with the exception of RCOF and RSO2F) have

13

Principles and Topical Applications of 19F NMR Spectrometry

Fig. 1. Chemical shift ranges of commonly occurring fluorine containing organic groups

negative chemical shifts (i.e. they appear to the right of the CCl3F signal). In contrast to 1H NMR spectra, in which the signals from protons bonded to sp3 and sp2 carbons are distinctly separated, the chemical shift ranges for fluorine atoms bonded to sp3 and sp2 carbons overlap extensively. Figure 1 illustrates the likely shift ranges for commonly occurring fluorine containing organic groups. The theoretical background that is necessary to explain the origin of 19F chemical shifts is currently incomplete. The wide chemical shift range for commonly occurring 19F functionalities arises from large paramagnetic contributions in the shielding constant. These contributions arise from the low-lying orbitals in 19F atoms, which are subjected to electronic excitations by the external magnetic field B0 of the NMR spectrometer, and result in variable down-field shifts for the NMR signal from 19F nuclei. In the case of 19F nuclei, diamagnetic contributions to the screening constant are very small (1%) and the effects of neighbouring groups are essentially negligible. For example, when fluorine is bonded to a sp3 hybridised carbon atom, a definite trend in chemical shift exists within the series R–CF3 , R2CF2 , R3CF. The signals from R3CF groups appear at highest field with respect to the others. This is accounted for by the fact that the increasing substitution of the sp3 carbon by fluorine reduces the ionic character of the resulting C–F bonds. Owing to reduced symmetry, the paramagnetic contribution to the shielding constant becomes larger, resulting in increasing downfield shifts for the signals from R2CF2 and R–CF3 groups respectively. The chemical shifts of 19F resonances may also be affected by steric (van der Waals) interactions. This can be illustrated with reference to the 19F chemical shifts of the series of ortho substituted fluorobenzenes shown in Table 3, in which the 19F resonances are progressively deshielded by their interaction with groups of increasing bulk. In addition to large chemical shift ranges, 19F NMR spectra are also characterized by relatively large values for spin-spin coupling interactions between Table 3. Chemical shifts of ortho substituted fluorobenzenes

X F Cl Br I

Chemical shift (d) –132 –109 –100 –87

14

P.D. Stanley

Table 4. Geminal, vicinal and longer-range 19F/19F and 19F/1H coupling constants (Hz)

CH3CH2F

CF3CF2CF2CO2H

2J FF

157

2J FH

49

2J FF

244

2J FH

49

2J FF

297

2J FH

85

2J FF

16

2J FH

73

2J FF

26

2J FH

55

3J FF

1.4

3J FH

27

3J FF

4

3J FH

20

3J FF

111 (trans)

3J FH

22 (trans)

3J FF

35 (cis)

3J FH

85 (cis)

3J FF

19

3J FF

125

3J FH

20

3J FH

4

3J FF

20

3J FH

6–10

CH3CH2F

15

Principles and Topical Applications of 19F NMR Spectrometry Table 4 (continued)

CF3CF2CF2CO2H

4J FF

10

4J FF 4J FF 3J FF

(cis) 22 (trans) 8 13

4J FH

3

4J FF

(meta)+/–20

4J FH

(meta) 6–8

5J FF

5–18

5J FH

2

“through space” interactions 170

8

both 19F–19F and 19F–1H nuclei. This often leads to the production of very complex signals at each chemical shift. Fortunately the relative magnitude of 19F–19F and 19F–1H spin-spin coupling constants compared with the potentially large chemical shift differences between coupling 19F nuclei leads to the situation whereby many 19F spectra can be interpreted using first-order rules. Table 4 gives a brief survey of geminal (2-bond), vicinal (3-bond) and longer-range spin-spin coupling constants. In this table, the signs of the spin-spin coupling constants are not given; the absence of this information should not be important except in the case of 1,3-difluorobenzenes where the experimental values of the 4-bond 19F–19F coupling constant lie in the range ±20 Hz, and they can be, and often are, 0 Hz. The rules used for proton spin-spin coupling cannot always be used to interpret the corresponding fluorine interactions. Spin-spin couplings are generally regarded as being transmitted through chemical bonds. In the case of fluorine nuclei, there is evidence that spin-spin couplings from fluorine nuclei may also be transmitted by a direct mechanism through space. The through space effect is envisaged as originating from scalar spin-spin coupling through non-bonding orbital overlap rather than from dipolar interactions. These long-range couplings can be detected through more than five bonds in fluorine, proton and carbon NMR spectra and provide insights into conformational preferences [106].

16

P.D. Stanley

1.7 Acquisition 19F NMR Spectra – Instrumental Considerations

NMR spectra can be acquired using either continuous wave (CW) or pulsed Fourier transform (PFT) NMR spectrometers. Although basically quite similar, they differ in the power and time dependence of the spectrum exciting RF wave and the mode of signal acquisition. CW instruments record a classical absorption signal whereas the PFT instruments use a digital process to acquire an interferogram, known as a free induction decay (FID), by means of an analog-todigital converter and a computer. Subsequent Fourier transformation of this time-domain data yields the frequency domain spectrum as acquired under CW conditions. The immediate advantages of PFT instruments are that they allow the fast acquisition of many spectra which are co-added, before processing, to increase the signal-noise ratio of the measured spectrum and the application of the multiple pulse experiments that are the basis of modern structure elucidation methods. Since their introduction, PFT spectrometers have become the instrument of choice for the observation of all NMR spectra. The early work reports spectrum characteristics obtained from continuous wave instruments at low field and is limited to the measurement of chemical shifts and spin-spin coupling constants. Due to the high sensitivity and wide spectral range of 19F nuclei, there were few technical problems in amassing the enormous libraries of data referred to in Sect. 1.5. Modern pulsed Fourier transform NMR spectrometers have multiple radiofrequency channels to allow observation, decoupling and access to multi-pulse NMR experiments. There are no NMR experiments commonly used that require more than four radio frequency channels. The major manufacturers often offer a choice between “routine” and “research” grade PFT instruments; often the differentiation between the “routine” and “research” grade instruments lies in the system’s potential for expansion. That is, most “routine” instruments have a fixed design that means it is not possible to add extra capability if the user’s needs change after purchase. In practice, “routine” instruments allow access to all the NMR experiments needed for the structural investigation of organic materials. The choice of magnet is fundamental in establishing the final cost of the NMR spectrometer.Whilst “routine” systems are normally offered with magnets delivering proton frequencies of 200–500 MHz,“research” systems are available with magnets delivering proton frequencies 200–900 MHz. The benefits of aspiring to the highest field magnet affordable are due to the accompanying higher sensitivity and spectral dispersion. Except in special cases, where the need for sensitivity is paramount (e.g. LC-NMR) or when extra spectrum dispersion is required to address particular classes of substances (e.g. carbohydrates), the need to acquire a system with a magnet delivering proton frequencies of above 500 MHz is not well developed for most problems in organic chemistry. In the case of 19F NMR, 400-MHz systems are normally adequate in terms of sensitivity and avoid operational problems that may occur from the extremely wide sweep widths encountered in 19F NMR spectra. In organic chemistry laboratories, spectrometers are normally configured to handle samples presented to the instrument in the “solution state”. In this con-

Principles and Topical Applications of 19F NMR Spectrometry

17

figuration there are a bewildering set of alternatives of NMR probes to select from. First, there is the choice between probes designed for conventional NMR tubes and flow-probes. Driven by experiments using LC-NMR and the need for the high throughput analysis of combinatorial chemistry samples, flow-probes are becoming a more popular choice for routine use due to their mass efficiency, measured in terms of ability to detect low masses of analyte. LC-NMR probes are optimised for proton observation, but some early designs of flowprobes allow the proton channel to be retuned for 19F observation. For conventional probes, no single design is universal and each configuration will find a place in a laboratory working in a specialist area. In most cases, more than one probe will be specified with each instrument to cover efficiently the complete range of experiments needed to effect structure elucidation by NMR. Except for specialist laboratories, the purchase of a dedicated 19F probe is often regarded as a luxury, but many chemists still require access to 19F NMR data. The usual compromise is to select a “tuneable” probe that allows the observation of 19F NMR spectra. For tuneable probes, the choice will be between a manually tuned probe with a fixed proton channel and a tuneable channel that covers a range of other lower frequency nuclei (typically 31P–15N) or a probe with three or four fixed frequencies (e.g. 1H, 19F, 31P and 13C). The latter is a popular choice since it covers the nuclei of most interest to organic chemists and removes the need to re-tune the probe between samples. In practice, these probes have limitations that need to be considered. In general, they will be calibrated using test-samples dissolved in a particular solvent (say CDCl3). Changing from one solvent to another (e.g. from CDCl3 to D2O) will introduce a change in performance of the probe, both in terms of overall sensitivity and a change in pulse duration that will, particularly, affect the appearance of DEPT spectra. One “work-around” is to generate calibration sets for each solvent likely to be used and implement them as part of the set-up for each new experiment, but this represents effort arguably best expended elsewhere. These effects become more dramatic as the field strength increases; at 300 MHz Varian offer an ABT (Always Beautifully Tuned) probe which is not sensitive to solvent changes. Above 400 MHz the effects are real and alternatives to get around this problem are offered in the form of the JEOL auto-tuneable probe and the PulseTune accessory from the Nalorac Corporation. Under computer control, each of these devices will tune and match the probe for each sample. As with flow probes, 19F is detected by retuning the proton channel. Tuneable probes are also likely to deliver a less than perfect spectrum baseline due to signals detected from the fluorinated components and glues used in the construction of the probe. These materials are used to ensure an acceptable spectrum baseline in the proton NMR spectra. Although the 19F spectrum baseline artefact signal is not normally intrusive when normal concentrations of analyte are used, it does assume importance when the spectrum is composed of a collection of very small signals. Even for normal strength samples, it is desirable to remove the baseline artefact for other than cosmetic purposes; subsequent spectrum processing (e.g. phase correction, peak listing etc.) is far more controllable when the spectrum baseline is flat. The baseline artefact can be removed effectively using post-acquisition linear prediction routines [84] nor-

18

P.D. Stanley

mally found in current instrument software; if these are not available, advanced baseline correction software (e.g. as found in ACD-NMR Processor) is worth considering. So far, we have a PFT spectrometer that acquires 19F NMR data comparable to that obtainable on a CW instrument, i.e. it allows recording of 19F NMR chemical shifts and the measurement of spin-spin coupling constants. Without a spectrometer deliberately configured to observe 19F spectra, access to more complex NMR experiments is not always facile. Whereas spin-decoupling in proton NMR and the broadband decoupling of protons during the acquisition of 13C spectra ({1H}13C) are commonplace, the decoupling of 19F nuclei is problematic due to the very wide spectral width of 19F spectra. Traditional methods of broadband decoupling (e.g. WALZ, GARP) do not work well because their limited bandwidths mean that high power levels need to be employed to cover wide spectral regions. Recent advances in adiabatic decoupling schemes (e.g. WURST) make it possible to generate very wide bandwidths with minimum power requirements. A recent article [23] describes the experimental considerations necessary to perform either proton decoupled fluorine ({1H}19F) or fluorine decoupled proton ({19F}1H) experiments on Varian Inova Spectrometers configured for specialist (e.g. dedicated H–F probe with a third broadband channel) or general (e.g. AutoSwitchable probe with a single high-band amplifier) use. When the instrument is configured to perform these experiments the acquisition of 2D 13C–19F correlation experiments is possible. In a subsequent article [105] the implementation of a variety of 19F–1H double resonance experiments is discussed. Due to the wide variation in protonfluorine spin-spin coupling constants, 1H–19F ge-COSY is an experiment that is ideal for proton-fluorine correlation since no fixed delays, incorporated to optimise for a particular value of spin-spin coupling, are required. 1D 19F filtered HMQC, HMQC-TOCSY and HMQC-NOESY are also described. 1.8 Mass Spectrometry and NMR as Complementary Procedures

In laboratories that have access to both mass spectrometers and NMR spectrometers there is normally little, or no, conflict regarding the application of each kind of instrument. Whereas MS gives primitive structural information that is ideal for confirmatory purposes, NMR gives precise structural information that is normally reserved for chemical structure determination. In special areas (such as protein sequencing) MS may be used to derive structural chemical information, but this should not be confused with the total structure determination of the functional structure of proteins as determined by X-ray or NMR. MS is a sublimely sensitive technique often requiring only picomoles of material; at a conservative estimate, MS is two orders of magnitude more sensitive than NMR, which is compromised when a sample of much less than a microgram of substance is to be analysed. In special applications (e.g. 19F NMR analysis of fluorinated metabolites) NMR can almost compete on sensitivity

Principles and Topical Applications of 19F NMR Spectrometry

19

grounds with MS and, because of the ease of sample preparation and the added selectivity conferred by wide spectral ranges, often becomes the analytical method of choice. To obtain MS data from a substance it must first be made to ionise and a portfolio of different methods are required to cater for the diversity of substances commonly analysed. Because of the small mass requirements for MS analysis it is often used in conjunction with a separation procedure such as gas chromatography (GC), liquid chromatography (LC) or capillary electrophoresis (CE). The text by Chapman [22] gives an up-to-date introduction to the topic of MS ionisation methods and interfaces. Presently, the major applications of MS focus on the qualitative confirmation of chemical structure with reference to a known (or proposed) material and the quantitative determinations of specific analytes using a combination of chromatography followed by MS. Presently, LC-MS is arguably the most widely used mass spectrometric technique in the areas of metabolite identification and quantitation. The volume by Willoughby et al. [109] gives an overview of the factors that determine the establishment and use of LC-MS technology in the modern laboratory. Although NMR is not a sensitive technique, it is non-destructive and does not consume or alter the sample. It is not regarded as a selective technique in conventional terms since it reports the presence of each substance equally, without the bias inherent to MS techniques that highlight the different ionisation characteristics of each substance present. When a sample is presented for NMR analysis and the instrument is tuned to observe (say) 19F nuclei, every fluorinated species present will give a signal (or set of signals) characteristic of its chemical structure. Since 19F is a relatively rare substituent in synthetic organic molecules, the huge chemical shift range of 19F signals (typically 100 kHz) compared with the relatively modest line width of most signals (<3 Hz) gives the potential to separate signals from more than 30,000 separate entities. Even greater selectivity can be conferred by recent developments that have successfully interfaced LC chromatographs with NMR instruments. LC-NMR technology is still under development and all the major NMR instrument manufacturers offer LC-NMR probes as standard accessories. First generation of LC-NMR probes were often re-tuneable to 19F and, as we will see later, there are several examples where 19F-detected LC-NMR has been used to identify chromatography peaks of interest prior to 1H NMR analysis. Sadly, the drive to improve NMR sensitivity for 1H observation has resulted in a generation of probes that cannot observe 19F and will necessitate the special manufacture of 19F observe LC-NMR probes. NMR is a concentration detector; this infers that signals from small amounts of material will be obtained if the volume of the analyte solution is small enough. In the context of 19F analysis, using an LC-NMR probe with a detect volume of (say) 60 µl it is possible to detect fluorinated substances, in real time, at levels of a few micrograms as the sample peaks pass from the LC though the NMR probe. Although presently limited to 1H NMR, NMR flow probes with ultra-small detection volumes (<2 µl) are at the point of production [61]; they are capable of generating good quality spectra of glucose at nanomolar concentrations (0.24 µg in cell) within 10 min. This cell size is compatible with the

20

P.D. Stanley

peak volumes in capillary LC and CE. The publication by Olson et al. [75] compares the efficiency of this microprobe with other NMR micro-detection devices. Except when extreme detection limits are required, the combined use of MS and NMR is a viable and indeed an essential part of chemical structure determination. For the amounts of analyte typically available from organic synthesis, or following enrichment, MS and NMR can be used as a triply hyphenated technique along with LC (i.e. LC-MS-NMR) as described in the paper by Bailey [6].

2 Topical Applications of 19F NMR Spectrometry Many examples could be given, but this would involve a massive undertaking that is not the primary objective of this volume. Instead, the structural determination of a specific compound is discussed in detail as an illustration of the possibilities opened up by 19F NMR. It is appropriate to note that applications in microbiology, where no theoretical background is presented, are discussed elsewhere in this volume: likewise the many applications involving perfluorinated compounds are not considered, since they are discussed elsewhere in this volume. 2.1 Structure Elucidation – Tefluthrin, a Case Study

Tefluthrin (Table 5) is an insecticide used to control a wide range of soil insect pests, particularly those of the orders Coleoptera, Lepidoptera and Diptera, in maize, sugar beet, wheat and other crops. The analysis of residues is a central issue that determines the acceptability of agrochemicals, and for tefluthrin no residues (at a limit of detection 0.01 mg/kg) have been found in major crops treated at recommended rates. The NMR spectra discussed below show how a range of experiments might be used in combination to successively refine the interpretation and assignment of the spectra of this commercially significant compound. These procedures may be regarded as hierarchical. All spectra were run on Varian Inova 400 MHz spectrometers. Figures 2–5 were acquired in our own laboratories (two-chan-

Table 5. Tefluthrin

2,3,5,6-Tetrafluoro-4-methylbenzyl (Z)-(1RS,3RS)-3-(2-chloro-3,3,3-trifluoroprop-1-en-yl)2,2-dimethylcyclopropanecarboxylate

21

Fig. 2. 1H spectrum of tefluthrin

Principles and Topical Applications of 19F NMR Spectrometry

22

P.D. Stanley

Table 6. Assignment of proton NMR spectrum of Tefluthrin

Chemical shift (d)

Integral (protons)

Description a

Coupling constant (Hz)

Assignment

6.89 5.24 5.18 2.29 2.18 1.97 1.30 1.29

1 1 1 3 1 1 3 3

bd bd bd t bt d s s

3J HH 2J HH 2J HH 4J FH 3J HH 3J HH

Propene 1-H Benzyl CH2 (one) Benzyl CH2 (one) Benzyl 4-CH3 Cyclopropane 3-H Cyclopropane 1-H Cyclopropane 2-CH3 (one) Cyclopropane 2-CH3 (one)

a

9.4 12.1 12.1 2.1 9.4 9.4

b=broad, s=singlet, d=doublet, t=triplet.

nel instrument) under automation, the remainder by Dr Péter Sándor at the Varian application laboratory in Darmstadt (three-channel instrument). In both, a 5-mm AutoSwitchable probe was used (outer coil doubly tuned to 1H and 19F, inner coil doubly tuned to 13C and 31P). No additional hardware, over and above the standard Inova configuration, was required, and the pulse sequences are from the standard Varian pulse library. The sample was prepared as a 20 mg/ml solution in chloroform-d to ensure reasonably fast acquisition times. The proton NMR spectrum (Fig. 2) shows all the features expected from an NMR spectrum, i.e. chemical shifts indicating the presence of a range of chemical environments, integration traces showing the relative number of protons in each chemical environment and spin-spin coupling constants giving information on nearby 1H and 19F nuclei. Under normal circumstances, and only with the benefit of hindsight, this spectrum is sufficient to confirm that the sample is tefluthrin. The analysis of the proton spectrum is shown in (Table 6). An important feature of this spectrum is that the 19F/1H spin-spin couplings observed in this spectrum are all transmitted over four chemical bonds (4J) and, in the main, are observed only as a broadening of the relevant proton signals and are therefore not diagnostic. The stereochemistry of the trans-cyclopropane ring is “confirmed” by the magnitude of the 3JHH coupling constant; in the case of cis-cyclopropanes in this class of substance, 3JHH is typically 5–6 Hz. The (Z) stereochemistry of the propene substituent is substantiated by the chemical shift of the alkene proton; in the case of (E) propenes in this class of compounds, the alkene proton is shifted to higher field. Although this type of analysis is readily carried out for a known class of compound, this example serves to illustrate that no real investigative analysis has actually been performed. The proton decoupled carbon spectrum and DEPT analysis (Figs. 3 and 4) add to the certainty of the structural assignments. The ACD spectrum prediction of the structure (Fig. 5) shows some small deviations from the observed spectra, attributable to the use of incomplete models in the prediction database.

Fig. 3. {1H}13C spectrum of tefluthrin; the inset shows an expansion of the aromatic region with the CF3 peak picked

Principles and Topical Applications of 19F NMR Spectrometry

23

carbons

Fig. 4. DEPT edited 13C spectra of tefluthrin showing (from bottom to top): all protonated carbons, only CH carbons, only CH2 carbons and only CH3

24 P.D. Stanley

25

Fig. 5. ACD prediction of 13C spectrum of tefluthrin

Principles and Topical Applications of 19F NMR Spectrometry

26

P.D. Stanley

Table 7. Assignment of carbon NMR spectrum of Tefluthrin

Chemical shift (d)

DEPT a

169.7 145.0 129.7 122.3 123.0 117.3 110.8 53.7 32.5 31.0 28.9 28.2 14.8 7.7

S S D S S S S T D D S Q Q Q

a

Coupling constant JFC (Hz) Complex 3J 4.5 CCCF3 2J CCF3 37.6 1J CCF3 271.0 2J 18.4 CCF 2J 17.7 CCF

Assignment

ACD prediction (d)

DACD–exp

COO Benzyl 2,3,5&6-C Propene 1-C Propene 2-C CF3 Benzyl 4-C Benzyl 1-C CH2COO Cyclopropane 1-C Cyclopropane 3-C Cyclopropane 2-C Cyclopropane 2-CH3 Cyclopropane 2-CH3 Benzyl 4-CH3

171.8 147.3 129.4 120.8 120.3 113.7 108.5 53.5 33.5 20.3* 29.1 22.8 19.9 4.6

2.1 –0.1 –0.3 –1.5 –2.7 –3.6 –2.3 –0.2 1.0 –10.7* 0.2 –5.4 5.1 –3.1

B=broad, S=singlet, D=doublet, T=triplet, Q=quartet denote the signal multiplicity reported by DEPT analysis (i.e. S is a quaternary carbon, D is a CH etc.). Coupling of the 13C signals to 19F is reported as text.

The large error in the prediction of the cyclopropane 3-C is an example of a program reporting error. The 32 models used to predict this chemical shift report an average value d 32.1 for this carbon atom, which represents an error of DACD–exp –1.1 ppm. The vendors of this software are keen to improve the program code and therefore welcome these observations so that the program can be improved. In this case, spectrum prediction is a useful tool that will be used to design NMR experiments to clarify deficiencies in the incomplete models. The analysis of the carbon spectrum is shown in Table 7. It is worth noting that the spin-spin couplings from the CF3 and aromatic fluorine nuclei are visible in the proton decoupled carbon spectrum. The magnitudes of these spin-spin couplings are useful markers that aid the assignment of individual signals and, with careful study, can be used to give conformational information. The signals from the benzyl 2,3,5 and 6-C appear as a single complex multiplet that is not interpretable. The 19F spectrum of tefluthrin (Fig. 6) shows signals consistent with CF3 (d69) and aromatic fluorine nuclei (d-144 and d-145) that are shown expanded in the inset. Under the IUPAC conditions for 19F NMR acquisition, integration of these signals is not meaningful. In this case, the substance was available in relatively large quantities and the previous data can be acquired reasonably quickly; all these data sets were acquired and processed within 1 h under total automation. When only smaller amounts of a sample are available, the 19F–13C couplings that we used to infer structural information are an inconvenience since the spin-spin coupling reduces the intensity of the signal (e.g. the intensity of a triplet is only 50% that

Fig. 6.

19F

spectrum of tefluthrin; the insert shows expansion of aromatic fluorine signals

Principles and Topical Applications of 19F NMR Spectrometry

27

Fig. 7. {1H/19F}13C spectrum of tefluthrin

28 P.D. Stanley

plets and resolution of signals at d 145

Fig. 8. Expansion of aromatic regions of {1H}13C (bottom) and {1H/19F}13C (top) spectra of tefluthrin showing simplification of fluorine coupled multi-

Principles and Topical Applications of 19F NMR Spectrometry

29

aromatic 2 and 6-F and broadband decoupling

Fig. 9. {19F}1H spectra of tefluthrin showing (from bottom to top): no irradiation, irradiation of CF3, irradiation of aromatic 3 and 5-F, irradiation of

30 P.D. Stanley

Principles and Topical Applications of 19F NMR Spectrometry

31

Fig. 10. Fluorine detected 19F-1H g-HMBC spectrum of tefluthrin

of the uncoupled signal). In such cases, where there is a suspicion that fluorine may be present, it is preferable to acquire preliminary data with both 19F and 1H decoupling as shown in Fig. 7. Figure 8 is an expansion of the aromatic carbon region of the {1H}13C and 1 { H–19F}13C spectra. In addition to an increase in the signal to noise ratio, the {1H/19F}13C spectrum shows that the complex multiplet at d145 can be resolved into two separate signals representing the benzyl 2 and 6, and benzyl 3 and 5 carbons. In the original proton spectrum we inferred the presence of the long-range 1H-19F spin-spin couplings that were expected in this substance through broadening of the coupled proton signals. When run under conditions of higher digital resolution these couplings are visible as shown in Fig. 9. This figure also illustrates a series of selective decoupling experiments that confirm the coupling pathways in the original assignment of the proton spectrum. This technique is selective enough to enable separate irradiation of the two signals from the aromatic fluorine nuclei, and confirms that the signal at d-144 in the fluorine spectrum represents the benzyl 3 and 5 fluorines. The same information is conveyed more graphically in the gradient selected 1H-19F HMBC spectrum shown in Fig. 10. In this spectrum, spin-spin couplings are displayed as correlation “islands” in the 2D spectrum. Thus the fluorine signal at d-144 (F2 dimension) shows a correlation with the proton signal at d1.97 (F1 dimension) that represents the benzyl 4-CH3. The CF3 signal shows correlations with both the propene 1-H and the cyclopropane 3-H that represents spin-spin coupling through both three and four chemical bonds.

Fig. 11. Fluorine detected g-HMQC spectrum of tefluthrin

32 P.D. Stanley

33

Fig. 12. Fluorine detected g-HMBC spectrum of tefluthrin

Principles and Topical Applications of 19F NMR Spectrometry

Fig. 13. Fluorine detected HOESY spectrum of tefluthrin

34 P.D. Stanley

Fig. 14. Fluorine detected HETGOESY spectrum of tefluthrin, with selection (from bottom to top) of propene 1-H, benzyl C-4 CH3 and benzyl C-1 CH2

Principles and Topical Applications of 19F NMR Spectrometry

35

36

P.D. Stanley

Figure 11 shows a gradient selected HMQC experiment optimised to determine one bond 19F–13C couplings (270 Hz) and confirms the assignment of the CF3 carbon. Interestingly, this experiment has sufficient resolution to discriminate between the two very close signals representing the fluorinated carbons in the benzyl group. Since they are so close, the result may seem academic but this is an interesting illustration of the capability of the technique. The longer-range spin-spin couplings between 19F–13C are probed using the gradient selected HMBC spectrum shown in Fig. 12. For example, correlations representing couplings from the fluorine signal at d-144 can be seen to the carbon signals representing the benzyl 4-C, benzyl 2 and 6-C and the benzyl 4-C CH3 . HOESY [85] is a 2D NMR experiment similar to NOESY that yields information on the spatial relationships between spins in the heteronuclear case. The primary use of this experiment is to determine distances between quaternary carbon atoms and protons where spin-spin coupling information is unhelpful. In this case, Fig. 13 shows the 1H–19F HOESY spectrum of tefluthrin, and proximity information between the fluorine signal at d-144 and the proton signal from the benzyl methyl group can be clearly seen. Often 1D NOE methods are to be preferred, and the HETGOESY experiment [102] originally developed to detect 13C–1H NOE enhancements, detected directly on the 1H channel, can be adapted to 19F measurements. This experiment uses pulsed field gradients for coherence selection, thereby completely avoiding the problems associated with difference spectroscopy. In this adaptation, selective excitation is used in the better-resolved decoupled 19F spectrum rather than in the 1H spectrum (Fig. 14). Although the information content of these spectra can be inferred from previous experiments, the sensitivity of this experiment is higher than with normal NOE methods, and the relatively small enhancements observed illustrate the problems associated with NOE measurements of small molecules. Using 19F detection minimises the number of artefacts in the spectra, allowing the easier identification of true correlations. 2.2 LC/NMR and Metabolite Identification

The analysis of complex mixtures (e.g. biological fluids) using high resolution NMR spectroscopy has been successfully applied to the study of altered levels of endogenous substances in pathological conditions [74] and to the identification of xenobiotics and their metabolites [43]. Naturally, complex mixtures generate complex NMR spectra that require spectral editing techniques and 2D NMR methods to assign resonances [41, 103]. If the mixtures can be separated, or simplified, the task of structure elucidation becomes easier. When metabolites with a range of polarities are present, the partial separation of the components using solid phase extraction prior to NMR analysis has been applied successfully [110]. Ideally each component of the mixture should be separated, for example by chromatography, prior to NMR analysis. Attempts to develop coupled LC-NMR

Principles and Topical Applications of 19F NMR Spectrometry

37

started in the 1980s [8, 32] but were not widely adopted due the lack of sensitivity of the NMR detector and the expense incurred by the necessity of using deuterated LC solvents. Recent advances such as higher magnetic field strength, developments in NMR probe design, and solvent suppression techniques have resulted in major improvements in LC-NMR technology. As a result, it has become a viable tool for obtaining NMR spectra directly from the small (microgram) amounts of separated compounds delivered from an on-line chromatography system. With this amount of material, however, only proton, fluorine and proton detected heteronuclear correlation experiments (e.g. HMQC etc.) are possible. The world’s major pharmaceutical businesses have been stimulated by these developments to undertake a major re-examination of the technique as a tool to solve a range of problems in biochemistry and drug metabolism. Despite aggressive investment in this area by industry, the major technical innovations in the area come from academia, for example, Nicholson’s group at Imperial College (University of London). Before we consider applications, it is important to recognize the operational choices and limitations of the LC-NMR technique. In its simplest form, the LC-NMR experiment involves the acquisition of NMR data from the effluent from the LC chromatograph that is used for separation of the components of the sample. The LC-NMR probe is designed to have a small volume (typically 60–100 µl) that is slightly smaller than the actual volume of solvent associated with a typical chromatography peak. This ensures that there are no concentration gradients across the NMR detection region in the probe leading to poor NMR line shape. With careful calibration of the instrument, NMR data from both isocratic and gradient elution regimes can be acquired successfully. LCNMR experiments are almost exclusively limited to chromatography systems containing either water/acetonitrile or water/methanol mobile phases. The common practice is to use deuterated water (D2O) and protonated acetonitrile or methanol since D2O is relatively cheap. The chromatography conditions required for successful LC-NMR are, however, often unlike those developed by a specialist chromatographer for purely analytical purposes. It is essential that the chromatography conditions are adjusted so that the peak of interest is well separated from closely eluting components. In addition, the chromatographic system should be tolerant of overloading since this condition is almost always required to ensure that sufficient material is present in the peak being measured by the NMR spectrometer. The major consideration in limiting the solvent system to a simple binary mixture is that the signals from the solvent, that appear as major peaks in the proton NMR spectrum, must be suppressed using a NMR pulse sequence. The resulting spectrum has areas that have been “burnt” out by the suppression process; more than two of these degrade the information content of the NMR spectrum to such an extent that the spectrum is useless. A recent review [98] describes recent developments in solvent suppression techniques and their application to LC-NMR. The possible modes of NMR operation are on-flow analysis, stopped flow analysis, and the retrospective analysis of components recovered from sample storage loops. When small amounts of the analyte are involved, the acquisition of useful proton NMR data “on-flow” is not viable due to the low sensitivity of

38

P.D. Stanley

even the highest field NMR spectrometer. “Stopped flow” analysis, where the chromatographic separation is paused during acquisition of the NMR spectrum (often for several hours), is far more widely used. To avoid possible degradation and diffusion of materials held on the chromatography column during these long periods of NMR spectrum acquisition, the chromatography is often carried out as normal except that the peaks of interest are transferred into sample storage loops. The contents of these loops can then be transferred sequentially to the NMR instrument for measurement. This method enables the NMR instrument to be used for other purposes whilst the chromatography is being completed. Sensitivity is the keyword for LC-NMR experiments. LC-NMR probes are optimised for sensitivity by designing detection cells with good filling factors and good NMR line shape. Based on detection of mass, these probes are the most sensitive conventional probes available. The other factor in the equation is, of course, the NMR magnet field strength. The higher the magnetic field, the more sensitive the instrument. It is unlikely that successful LC-NMR experiments on low abundance substances will be carried out effectively with magnetic fields of less than 500 MHz, with 600 MHz being the industry standard. Fields in excess of this are desirable, but in the main often unaffordable. Nicholson’s paper [100] was the first to report the use of 19F NMR in conjunction with LC. Whilst 19F NMR spectra contain little structural information, due to the absence of background signals, they are an ideal indicator of which peaks in a complex chromatogram represent fluorinated materials. This paper demonstrates the identification of the two major urinary metabolites of the anti-inflammatory drug Flurbiprofen: the glucuronide and 4´-hydroxyflurbiprofen, by a combination of on-flow 19F detected LC-NMR to identify the fluorine-containing chromatography peaks, followed by the acquisition of stopped flow 1H detected LC-NMR to enable structural characterization of the metabolites. This principle may be illustrated with reference to plant metabolites [6]. 5-Trifluoromethyl-pyridone (5TFMP) serves as a simple substrate to examine the possible occurrence of N- vs O-glycosylation in plants. Following spiking with 5TFMP, the 500-MHz 1H NMR spectrum (Varian Inova 500) of a typical aqueous extract of maize (Zea mays L.) shoot material is shown in (Fig. 15). Apart from signals due to carbohydrates (d 5.45–3.20) the spectrum contains few interpretable peaks. The corresponding 19F NMR spectrum (JEOL GSX270) is shown in (Fig. 16). In this spectrum, the chemical shift scale was referenced using 3-trifluoromethyl-pyridone (3TMFP), and clearly shows the presence of four other fluorinated materials, identified as 5TFMP together with three putative metabolites. The on-flow 19F NMR detected HPLC (Brüker DRX 500) chromatogram is shown in (Fig. 17), where the data are shown in “pseudo 2D” format representing several 1D 19F NMR spectra viewed from above, and acquired separately during a period of approximately 30–45 min during the elution of the chromatogram. It should be noted that the sensitivity of the experiment precludes the detection of a signal for the substance labelled “Metabolite III” in Fig. 16. From this data we conclude that the 19F NMR signal for “Metabolite I” is con-

Fig. 15. 500 MHz Proton NMR spectrum of maize extract; residual peak from water signal suppression indicated at d 4.8

Principles and Topical Applications of 19F NMR Spectrometry

39

19F NMR spectrum of maize extract 14 days after dosing with 5-trifluoromethyl-2-pyridone (5TFMP). Chemical shift referencing is to 3-trifluoromethyl-2-pyridone (3TFMP)

Fig. 16. 254.05 MHz

40 P.D. Stanley

mission from J Agric Food Chem (2000) 48(1):42–46, Copyright (2000) American Chemical Society

Fig. 17. On-flow 470.5 MHz 19F NMR detected HPLC chromatogram of maize extract after dosing with 5TMFP. Reproduced with per-

Principles and Topical Applications of 19F NMR Spectrometry

41

Fig. 18. UV detected HPLC chromatogram of maize extract after dosing with 5TMP. Metabolite peaks are indicated at 33.9 and 43.3 min (two co-eluting peaks). Reproduced with permission from J Agric Food Chem (2000) 48(1):42–46, Copyright (2000) American Chemical Society

42 P.D. Stanley

Principles and Topical Applications of 19F NMR Spectrometry

43

Fig. 19. Stopped-flow 500 MHz 1H detected HPLC chromatogram of (from bottom to top):

metabolite I, metabolite II contaminated with 5TFMP. Reproduced with permission from J Agric Food Chem (2000) 48(1):42–46, Copyright (2000) American Chemical Society

44

P.D. Stanley

Fig. 20. Stopped-flow 500 MHz 1H detected HPLC chromatogram of (from bottom to top):

metabolite III, metabolite II, metabolite I and 5TFMP

Principles and Topical Applications of 19F NMR Spectrometry

45

tained in an LC peak with a retention time of 33 min, and that “Metabolite II” and 5TFMP co-elute in an LC peak eluting at 43 min. These findings can be transposed onto the complex UV chromatogram of the maize extract shown in (Fig. 18). Using this information, the stopped flow 1D 1H NMR spectra (Brüker DRX 500) from the chromatography peaks identified above are shown in (Fig. 19). From these spectra it is possible to assign the chemical structure 5TFMP-N-glucoside to Metabolite I, and 5TFMP-O-6-malonylglucoside to Metabolite II. Since Metabolite II co-elutes with 5TFMP, signals from the latter are clearly observed in the 1H NMR spectrum, but the difference in molar concentration between the two components does not lead to confusion during the data interpretation step. Coupled with a modification to the chromatography conditions, stoppedflow 19F NMR detected HPLC (Brüker DRX 500) enabled the detection of a signal for the low abundance Metabolite III and a partial separation of the other metabolite containing peak now eluting at 39 min. Stopped flow 1D 1H NMR spectra (Brüker DRX 500) of the identified chromatography peaks are shown in Fig. 20. From the spectrum of the newly identified 19F-containing chromatography peak it is possible to assign the chemical structure 5TFMP-N-6-malonylglucoside to Metabolite III. “Time-slicing”, i.e. advancing the chromatograph flow slowly though a peak and acquiring NMR spectra at each stopping point, allows the homogeneity of chromatography peaks to be investigated. “Timeslicing” through the chromatography peaks with retention times of approximately 39 min enables the acquisition of a spectrum of pure Metabolite II and of a slightly contaminated 5TFMP. In all these spectra, the approximate mass of material detected in the NMR probe was in the order of 5–10 µg. A second novel application of 19F NMR detected LC-NMR has been reported by Nicholson et al. [97] in which the acyl migration reaction of drug 1-O-acyl glucuronides is monitored. This reaction is of significance because of the possible role of acyl glucuronides in covalent binding to serum proteins and consequent allergic reactions. 2.3 Quantitation of Metabolites

The registration of a chemical substance for use as a pesticide is preceded by many detailed studies that ensure the substance does not accumulate in crop species and that the substance and its metabolites do not have toxic, or adverse ecological, properties. The first step in this process is to identify the metabolites. Such studies are normally carried out using 14C labelled materials, whereby the putative metabolites are located using thin layer chromatography (autoradiogram) or liquid chromatography and radiochemical detection. The isolation procedure for the metabolites, prior to spectroscopic structure elucidation, is time-consuming and often entails complex extraction procedures. When the substances are fluorinated, 19F NMR can often be used to advantage in addressing the problem; often, little or no clean up of the crude plant extracts is required. Normally this approach is only viable when the final concentration of

46

P.D. Stanley

analytes is projected to be at levels greater than 1–5 µmol/l. It should be emphasised that this concentration will normally only be achieved following reduction of a much larger extraction volume. Such quantities are compatible with the application rates and legislative detection limits for the registration of the current generation of herbicidally active compounds. The detection limit for each analyte is derived from a combination of the measured sensitivity of the available NMR spectrometer coupled with a sensible total acquisition time for each analysis and will be variable dependant upon the nature of the fluorine reporter group. For variously configured NMR instruments in the 270–500 MHz range, an accumulation time of 4 h will detect the 19F NMR signal from around 0.5 µg of a CF3 substituted analyte (MW 300 Da) dissolved in 0.5 ml of NMR solvent. Using these guidelines, the concentration of the extracts from 25 g of crop material or 1L of groundwater to 0.5 ml will deliver, respectively, detection limits of 20 ppb following a 4 h accumulation time and 1 ppb following a 1-h accumulation time. In our laboratory, we have used such principles to establish validated experimental protocols to detect residues and metabolites of a range of fluorine containing pesticides. In particular they have been used to support the registration of the herbicidally active substances fomesafen and fluazifop-butyl, the chemical structures of which are shown in Table 8. The comparative detection of fluorinated xenobiotics and their metabolites by both 19F NMR and 14C labelling techniques is discussed in the paper by Serre et al. [93]. Apart from an identification strategy for unknown fluorinated metabolites, this reference discusses the procedure required to acquire quantitative NMR data that deliver an equivalent sensitivity to the 14C method. To obtain reliable quantitative results (±10%) the T1 relaxation rates of the analytes must be determined with some precision [19], since they will determine the cycle time for each separate NMR acquisition as they are summed, over several hours, to give the required detection limit. It is tempting to set up the NMR acquisition so as to use conditions that deliver maximum sensitivity per pulse; i.e. a p/2 (90°) pulse duration with a relaxation delay five times the T1 of the slowest relaxing analyte. However, for longterm data acquisition, using conditions that deliver the maximum sensitivity per unit time is preferable. The NMR pulse duration and relaxation time are related by the Ernst angle condition [37]. Under this condition, reducing the pulse Table 8. Fomesafen and fluazifop-butyl

fomesafen: 5-[2-chloro-4-(trifluoromethyl) phenoxy-N-(methylsulphonyl)-2-nitrobenzamide

fluazifop-butyl: 2-[4-[[5-(trifluoromethyl)2-pyridinyl]oxy]phenoxyl]propanoic acid, butyl ester

Principles and Topical Applications of 19F NMR Spectrometry

47

duration (e.g. to 30°) allows the relaxation delay to be shortened whilst retaining the quantitative nature of the derived NMR data. Accordingly the 4 s cycle time of the maximum sensitivity per pulse condition can be reduced to 0.3 s by adopting the maximum sensitivity per unit time conditions specified by the Ernst condition. The summation of many more of these less sensitive pulses is a more effective strategy, on a per time basis, than the summation of fewer more sensitive pulses. In this case, by using the Ernst angle conditions and a total acquisition time of 4 h, the signal to noise ratio of the spectra were increased by a factor of 1.26 over the maximum sensitivity per pulse conditions. In such cases the 19F NMR signals are very small and the acquisition conditions for the NMR spectrometer need to be set up with care. As discussed earlier, some NMR probes capable of 19F detection have a broad background 19F signal as a consequence of their manufacturing processes. During the collection of the time-domain NMR spectrum, the signal representing this broad baseline artefact will dominate the first few (n) points of the FID. As a consequence, the acquisition conditions for the sample will be imperfectly set, resulting in a less than optimum detection of these small signals. NMR acquisition software normally allows adjustment of the data acquisition conditions so that the first n points of the FID can be omitted. Under these conditions, the instrument will adjust the acquisition parameters to collect data based on the signals present rather than those from the artefact peak, the missing data points being re-inserted using post-acquisition backward linear prediction. Following acquisition of NMR data under quantitative conditions, the areas of the detected NMR peaks must be measured (integrated) to determine the concentration, and thus the mass, of each species present. Although it is possible to compare the relative amounts of material present in a series on NMR spectra from inspection of the instrumental parameters alone, this comparison is most often achieved by the addition of a “spike” of known substance to each analytical sample. The spike will normally have a 19F chemical shift close to that of the other analytes and can be used both as a chemical shift reference and a quantification standard. The manual, or automatic, integration of the data typically produced from these analyses is not generally easy to perform with any accuracy. Deconvolution of the spectra using the curve fitting routines that form part of the basic functionality of most NMR spectrometer software often produces accurate and reproducible results where manual integration fails. In their paper Table 9. Model fungicide

N-Ethyl-N-methyl-4-(trifluoromethyl)-2-(3,4-dimethoxyphenyl)benzamide

perimental spectrum of model and five of its metabolites at different concentrations. The lower trace shows the calculated deconvolution of the experimental spectrum. Reproduced with permission from J Agric Food Chem (1997) 45(1):242–248, Copyright (1997) American Chemical Society

Fig. 21. The 19F NMR spectrum of fungicide model and related compounds, mixed at different concentrations in CHCl3 . The upper trace shows the ex-

48 P.D. Stanley

Principles and Topical Applications of 19F NMR Spectrometry

49

Serre et al. [93] discuss the identification and quantification of the metabolites of a model fungicide N-ethyl-N-methyl-4-(trifluoromethyl)-2-(3,4-dimethoxyphenyl)benzamide, shown in (Table 9). The efficiency of the deconvolution method is shown in (Fig. 21), in which the upper trace shows the experimental spectrum of the model fungicide (labelled AI) and five of its metabolites that were added at different concentrations. The lower trace shows the calculated deconvolution of the experimental spectrum. The concentrations of the various components were calculated from the deconvolution data with CFCl3 (1 mmol/l) as a reference. From left to right the results are as follows: AI 2.57 mmol/l calculated concentration/2.50 mmol/l actual concentration; 3-mono-deOCH3 , 1.25/1.25 mmol/l; N-deCH3 , 1.84/1.88 mmol/l; N-deC2H5 , 1.31/1.25; N-didealkyl, 0.68/0.62 mmol/l; free acid, 2.33/2.50 mmol/l. Other applications in this area include a comparison of 19F NMR and GC-ECD for analysing trifluralin residues in field grown crops [69], the 19F NMR analysis of trifluralin in a range of crops and derivative products [70] and the generic detection of CF3-containing residues at or near the legislative tolerance levels in foodstuffs [63]. Applications to animal systems include an investigation of the toxic effects of 3-trifluoromethylaniline on earthworms [108] and the quantification of the metabolites of 2,3,5,6-tetrafluoro-4-trifluoromethylaniline by using 19F NMR as a part of a strategy incorporating a range of hyphenated techniques [90]. 2.4 Solid State NMR Applications 2.4.1 Gels

A detailed discussion of the theory of the NMR of solid samples is beyond the scope of this chapter. With reference to NMR spectroscopy, the difference between the liquid and the solid state is the timescale and geometry of molecular motion. In liquid samples the intramolecular effects – dipolar coupling, quadrupolar interactions and chemical shift anisotropy – are averaged to zero by rapid molecular motion. In solid samples intramolecular effects are pronounced and the NMR spectrum of a typical solid sample consists of broad overlapping signals whose widths and shapes provide information about molecular motion within the sample as well as giving the distribution of particular molecular orientations within an ordered sample. In semi-solid samples there is sometimes sufficient molecular motion to give useable NMR spectra. The NMR spectra of “gel-phase” samples such as swollen solid phase beads inserted into a conventional NMR tube are a case in point. For such samples, proton NMR is usually uninformative but Shapiro and Wareing [95] point out that 19F NMR is particularly useful because of its large chemical shift range and because structural modifications quite remote from the fluorine can give rise to useful chemical shift changes [104]. High quality 19F NMR spectra can be obtained from substances bonded to 20–500 mg of TentaGel resin in a very short time, in which the spectral line-widths are almost the same as in

50

P.D. Stanley

solution. Accordingly chemical shift differences of 0.5 ppm can be readily distinguished. A large variety of fluorine-containing building blocks and reagents are commercially available, allowing the use of fluorine as a convenient reporter for gel-phase chemistry without the need to resort to the synthesis of special materials. Gel-phase 19F NMR has been used successfully to monitor nucleophilic aromatic substitution reactions [94]. 2.4.2 Semi-Solids

For true solid samples, even 19F spectra are often too broad to be useful. Several techniques have been developed to overcome the problem of the spectrum broadening encountered in solid samples. In principle, the broadening can be removed by imitating the motions of molecules in a liquid. To a good approximation, all the intermolecular broadening effects are proportional to (3 cos2 q–1), where q is the angle between the coupling vector of two spins and the magnetic field. This term reduces to zero when q=54.7°, the “magic angle”. If the entire sample is rotated at this angle, within the magnetic field of the NMR spectrometer, with a rotational frequency greater than the frequency range of the interaction the effects of the interactions are removed from the spectrum. This technique is known as magic angle spinning (MAS). In semi-solid samples, such as tissue samples and those polystyrene beads with shorter linkers than TentaGel some components of a “solid” sample will still have reasonable molecular motion. This motion may well be slower or more hindered than in solution and may be anisotropic, resulting in moderate line broadening due to residual dipolar interactions and sample heterogeneity. Rapid spinning of the sample, typically 2–15 kHz, at the magic angle is often sufficient to significantly reduce or eliminate the broadening experienced by the more mobile components in the sample. The improvement in sensitivity and resolution obtained can be quite dramatic. It should be noted that this does not involve the use of high-power decoupling or any of the special techniques commonly used to obtain NMR spectra from true solids. To distinguish it from true solid state NMR, the technique has been dubbed “high-resolution magic angle spinning” (HR-MAS). In our laboratory we have successfully used a Varian Nano·NMR probe to observe 19F spectra from a range of semi-solid samples. The same type of equipment was used to generate the NMR data used to quantitate reaction products of solid phase synthesis [33]. The best quality spectra were obtained from samples consisting of 2–4 mg of resin swollen with DMF; spinning speed was in the range 1.5–1.7 kHz. 2.4.3 Solids

True solids require the higher spin rates that are only obtainable from purpose build MAS probes, which typically spin samples at up to 20 kHz. 19F solid state NMR was used to investigate the sorption of hexafluorobenzene to soil organic matter [59]. The authors demonstrated that the sorptive uptake of hexafluo-

Principles and Topical Applications of 19F NMR Spectrometry

51

robenzene gives direct spectroscopic evidence for the existence of dual-mode sorption to soil organic matter. The sorption process was shown to be rapid, with all the applied hexafluorobenzene being adsorbed within a few hours. Extractable lipids competed for high-energy sorption sites in the organic matter, and their removal increased the amount of rigidly sorbed immobile species present. Soil lipids enhance the sorption capacity of the solid-state dissolution domain of the organic matter and this dissolution domain was responsible for partitioning in the dual-mode phenomenon. Removing the lipids decreased the partitioning capacity of the soil organic matter. A second paper [28] describes the use of 19F NMR to probe the mechanism of sorption in a range of sediments and polymers. 2.5 Biochemical Studies 2.5.1 Labels and Tags

Since fluorine is infrequent as a substituent in naturally occurring substances (covered elsewhere in this volume) and is generally resistant to degradation, once incorporated into a biological substrate it functions as an indelible label that, with minimal background interference, can be used to monitor biochemical structure and dynamics in solution using spectroscopic methods. The intrinsic low sensitivity of the NMR experiment persistently limits the generic applicability of NMR techniques to biochemical systems: whereas NMR measurements are most often made at millimolar concentrations, many relevant biochemicals are only present at sub-micromolar levels. Although advances in technology are gradually improving the sensitivity of NMR instruments, this threefold order of magnitude deficit in abundance will remain a real challenge for the foreseeable future. Within limits, it is possible to increase the detectability of fluorinated substances by increasing the fluorine content of the label. For example the signal intensity of a trifluoroacetyl substituent, a commonly used reporter substituent for amino acids, carbohydrates and steroids has three times the sensitivity of a monofluoro derivative. The review by Everett [38], cited in Sect. 1.6, contains a comprehensive list of references that exemplify the ways in which 19F NMR has been used as a biochemical probe. Some more recent novel applications include the determination of the antioxidant capacity of bio-molecules using high resolution 19F NMR [4]. This method is based on the use of trifluoroacetylated reporters such as trifluoroacetanilide. Upon hydroxyl radical attack such fluorinated detectors yield trifluoroacetamide and trifluoroacetic acid that can be quantitatively determined by 19F NMR. Both 1D and 2D 19F NMR methods have been used to determine the conformational heterogeneity of modified DNA fragments [113]. The real-time refolding of 6-fluoro-tryptophan labelled proteins have been monitored using stopped-flow 19F NMR [48].

52

P.D. Stanley

2.5.2 Magnetic Resonance Imaging (MRI)

MRI is a non-invasive diagnostic technique commonly used in both medicine and materials science. A readily understandable guide the principle of MRI is to be found at Hornak’s website [50]. Using MRI techniques it is possible to measure a number of chemical and physical parameters, temporally and spatially resolved, in intact biological systems. Since MRI measurements do not perturb the metabolic and transport processes within organisms, mediation of these processes brought about by a change in environment and in response to a chemical challenge can be observed particularly well. Fluorine MRI is a relatively new technique and has been used clinically to investigate the oxygenation of tumours [64] and many dynamic processes such as cerebral oxygen consumption and blood flow [79]. MRI techniques have also been used extensively in materials science. In addition to normal imaging methodologies, the less well-known stray field imaging (STRAFI) technique enables NMR images to be generated using the large magnetic field gradients that occur outside a normal NMR magnet. Dental cements [60] and solid paramagnetic species [83] have both been studied using 19F STRAFI methods. 2.5.3 Drug Design

In drug or other “effect chemicals” design, ligand binding by a target protein induces the ultimate effect of interest, such as cell growth or cell death. With knowledge of the structure of the protein targeted by a particular disease state and its relevant ligand binding sites it may be possible to design inhibitors or activators to fit the architecture and chemical nature of that site and to elicit the desired response. To observe ligand binding in solution, the NMR spectrum of ligand-free protein is compared with that of the ligand-bound protein. When a ligand binds to a protein, the NMR signals from amino acid residues in, or close to, the binding site may move, broaden or disappear. To ensure that binding is occurring at the site of interest, it is essential to have access to the spectroscopic assignment of signals from the amino acids that form that binding site. The total assignment of the NMR spectra of proteins is a task not to be undertaken lightly since both data acquisition and analysis involve complex procedures. To speed up the acquisition of the spectroscopic data it is preferable to use labelled proteins whenever they are available. With proteins having molecular weights greater than around 7 kDa, 15N incorporation is generally sufficient; proteins with molecular weights greater than 15 kDa generally require both 15N and 13C labelled material; when the molecular weight is greater than 25 kDa it is preferable to use 15N, 13C, 2H triply labelled substrates. Measuring changes in chemical shifts and hydrogen exchange rates serves to elucidate the dynamics of ligand binding and determines the critical residues for ligand-protein interactions [101]. The landmark paper by Fesik et al. [96] describes how NMR may play a directive role in the process of discovering which ligands bind effectively to a spe-

Principles and Topical Applications of 19F NMR Spectrometry

53

cific protein thereby making a significant contribution to the process of discovering new active substances. This elegant method has become known as “SAR by NMR” (structure/activity relationships by NMR) and neatly spans the paradigm of rational design and high throughput screening strategies by applying mixtures of carefully chosen small molecules (ligands) to the target protein. The resulting [15N] HSQC spectra are capable of detecting ligand-protein bindings that are below the level of detection of conventional in vitro biochemical assays. After finding small molecules that bind, albeit weakly, at different yet proximal sites in the protein the two molecules can be linked together in the proper orientation to produce a tightly binding, and thus more potent, substance. At a recent conference it was revealed that this methodology and subsequent developments produce almost 20% of all lead substances within the author’s business. This immensely powerful method is relatively input intensive since it requires a backbone assignment of the target protein and a supply of 15N labelled material. This method has also been subjected to patent protection, which has no doubt played some part in the spawning of several related proton NMR and mass spectrometry [47, 88] based methods. These alternative NMR methods are based on either a single measurement or combination of NOE [67, 86], molecular translational motion [25] (diffusion) and NMR line-shape measurements [68]. All these methods share the common advantage that they measure changes in the ligand signal rather than the protein; therefore neither the complete spectrum assignment of the protein nor labelled material is required. The disadvantage of these methods is that there is no guarantee that the measurements are related to the binding site of interest. In principle all these processes can be observed using 19F NMR provided the ligands contain fluorine as either an essential constituent or as a tag. The 1991 review article by Jenkins [54] describes the study of ligand-macromolecular interactions monitored by 19F NMR. The revue covers the determination of protein/ligand equilibrium constants and stoichiometries, the co-binding and competition of ligands, allosteric and conformational effects together with reaction kinetics. Chemical models include the interaction of 5-fluorotryptophan, 5-fluorosalicylic acid, flurbiprofen and sulindac sulfide with human serum albumin. 2.6 Determination of Optical Purity – Fluorinated Derivatisation Reagents

For any chemical structure, the determination of the absolute stereochemistry is vital for a complete understanding of its chemical and biochemical reactions. Many chemists apply a mixture of degradative chemistry, total synthesis and Xray analysis as the arbiter between similar working structures. In addition to Xray analysis, optical rotatory dispersion (ORD) and circular dichroism (CD) are effective techniques providing that a chromophore is present. Many empirical methods exist to determine absolute stereochemistry, amongst them NMR. The NMR spectra of enantiomers of a chemical structure are identical, whereas the NMR spectra of diastereomers are potentially different.

54

P.D. Stanley

There are three ways of using this methodology to investigate the optical purity of substances, and hence make inferences about their absolute stereochemistry. All methods rely on inducing diastereoisomerism in an optically pure material by chemical reaction or complexation. The high sensitivity and wide chemical shift range of 19F nuclei make it the ideal tool to monitor these processes, either when 19F is present in the substance or when it can be introduced as part of a derivatisation process. The separate 19F signals from these diastereoisomeric derivatives can be identified and quantified, thus determining the optical properties of the original compound. It should be born in mind, however, that the 19F chemical shift differences between diastereoisomers are often small (less than 1 ppm). Although this limits the technique, especially when the 19F signal is extensively coupled, the resulting spectra are almost always easier to work with than the complex spectra derived from proton NMR. Analysis based on 19F NMR does not of course preclude the use of the proton spectrum when 19F chemical shifts are not diagnostic. Although quite old, the volume by Martin et al. [62] still serves as a useful source of practical information on the use of NMR methods to assess optical purity. A more recent review [17] discusses the preparation and properties of chiral organofluorine substances with respect to the reaction of chiral reagents with prochiral fluorinated substrates. Chemical derivatisation is often not the method of choice since it involves an additional step prior to analysis, and it can introduce uncertainties arising from the mechanism of the reaction. Many references, however, report the successful application of this approach and a wide range of suitably pure derivatising reagents are available commercially: these include a-methoxy-a-trifluoromethylphenylacetic acid (MTPA or Mosher’s acid) [30], 2,2,2-trifluoro-1-(9anthryl)ethanol (Pirkel’s reagent) [82] and 1-amino-2-fluoro-2-phenylethane [46]. Classes of compounds that may be analysed using this methodology include alcohols, amines and carboxylic acids. Reliance on amide formation to induce diastereoisomerism can be equivocal since rotation of amide bonds often results in the existence of two forms that are in dynamic equilibrium, and that are resolvable under the conditions of the NMR experiment. If it is desired to avoid chemical derivatisation, the use of either chiral lanthanide shift reagents or “chiral solvents” should be considered. Chiral lanthanide shift reagents are appropriate when the substrate contains basic functional groups capable of co-ordinating with a metal. Increased effectiveness is to be expected when the co-ordinating site in the substrate is close to the chiral centre. As for any lanthanide shift reagent, scrupulous laboratory techniques are required to guarantee success and reproducibility; these techniques are described in the volume by Martin et al. referred to above. Chiral lanthanide shift reagents may generically be represented as (L)3M(III), where L represents a chiral ligand and M a paramagnetic metal from the lanthanide series. A representative selection of the reagents that are available commercially uses either 3-trifluoromethylhydroxymethylene-d-camphor (facam), 3-heptafluoropropylhydroxymethylene-d-camphor (hfbc) or d,d-dicampholylmethane (dcm) as the ligand, and either europium or praseodymium as the paramagnetic metal. When the resultant diastereoisomeric complexes have the

Principles and Topical Applications of 19F NMR Spectrometry

55

same geometry, europium and praseodymium induce chemical shift changes in the substrate that are in opposite directions. Chiral shift reagents do not, however, work well in aqueous conditions, and for compounds with sufficient water solubility, the measurement of NMR spectra following complexation with cyclodextrins should be considered [44].

3 Conclusions and a Prospective View The aim of this chapter has been to give an overview of the relevance of fluorine as a sensitive probe of chemical structure and to illustrate how the rarity of fluorine, as a naturally occurring substituent, in organic substances can be used to advantage in locating and quantifying significant products of chemical reactions and metabolic processes. The high intrinsic NMR sensitivity of 19F coupled with its wide spectral range means that informative 19F NMR data can be obtained using NMR spectrometers with the most modest performance, often with no modification of the basic hardware. It is even possible to identify various fluorinated residues without actually performing 19F NMR [26]. This interesting procedure relies on the pictorial analysis of the contour patterns formed within the correlation islands of 2D 13C/1H correlation spectra. Using this method, the identification of the position of fluorination in a substance can be inferred relatively simply and, additionally, allows other significant NMR parameters (e.g. 19F/13C and 19F/1H spinspin coupling constants) can be obtained directly from the plots. As pointed out several times throughout this chapter, the sensitivity of the NMR experiment is rather low compared to other spectroscopic techniques and the optimisation of NMR sensitivity has almost become a research topic in its own right. There are several possibilities for increasing the signal-to-noise ration (S/N) of NMR spectra [45]: 1. Increase the static magnetic field (B0 ) 2. Increase the sample concentration within the detector coil 3. Increase the sample and coil volume when the solubility of the sample is limited 4. Reduce the sample coil volume when the amount of substance is limited 5. Increase the number of transients acquired 6. Choose the most sensitive nuclear species for excitation and detection 7. Use pulse sequences optimised for the molecular weight of interest [76] A different, purely electronic, approach for increasing NMR S/N is to reduce thermal noise within the signal pathway of the NMR spectrometer by cooling them to extremely low temperatures [71]. The first commercial cryogenic probes configured for either dual frequency or inverse triple resonance experiments were introduced by Brüker in 1999. In these devices both the receiver coil and the signal preamplifier are cooled to low temperatures with helium gas. The spatial requirements for insulating and cooling the coil result in a reduction of the probe filling factor and the absolute signal amplitude generated. Even so, the reduction in noise with these cryogenic NMR probes dominates and there is a

56

P.D. Stanley

significant increase in S/N per transient compared with the results obtained from an equivalent conventional probe. The comparison of the performance of a cryogenic and conventional probe is described in the paper by Martin [87]. Using a 3-mm cryogenic probe from another manufacturer (Nalorac), the acquisition of the HSQC spectrum from a small strychnine standard afforded a 12- to 16-fold reduction in data acquisition time for a comparable S/N ratio spectrum. Although the present major area of development for the manufacturers of these probes is to optimise inverse triple resonance performance for use in protein studies, 19F detection probes have been constructed that offer the 3- to 4fold sensitivity gains expected from theory. Using such devices will push the level of detection of suitable fluorinated materials into the low nanogram range using acceptable acquisition times. The commercial significance of fluorinated materials, coupled with the favourable NMR properties of 19F nuclei, resulted in an enormous amount of historical literature focused on 19F NMR principles and practice. At one time relegated to a confirmational tool, access to modern NMR instrumentation and pulse methods have revitalised the technique as a tool for chemical structure elucidation and quantitative analysis. The many new applications appearing in the current literature indicate that 19F NMR will continue to be used appropriately as a powerful member of the armoury of modern analytical techniques.

4 Appendix IUPAC guidelines for the acquisition and reporting of 19F NMR data are: 1. Use fluorotrichloromethane (CCl3F) as the standard reference. 2. Use the standard sign convention of (+) for signals downfield from (left of) CCl3F and (–) for signals upfield from (right of) CCl3F. The vast majority of C–F signals are negative. 3. Give a clear indication of solvent, concentration and temperature. These parameters have a much greater effect on chemical shifts and coupling constants for fluorine than for protons. 4. Give an exact description of instrumentation – magnetic field strength, pulse sequence, decoupling mode etc. 5. Use standardised plotting and spectral presentations, comparable to d 0–10 used for proton spectra. The range from d +50 to –250 covers most organofluorine signals but is too wide to show small spin-spin couplings. To measure spin-spin couplings accurately it is usual to re-acquire the particular areas of interest with higher digital resolution.

Principles and Topical Applications of 19F NMR Spectrometry

57

5 References 1. Advanced Chemistry Development Inc, Toronto, Canada ACD/CombiNMR, http //www acdlabs com/ 2. Advanced Chemistry Development Inc, Toronto, Canada ACD/Structure Elucidator, http //www acdlabs com/ 3. Advanced Chemistry Development Inc, Toronto, Canada http//www acdlabs com/ 4. Aime S, Calzoni S, Digilio G, Giraudo S, Fasano M, Maffeo D (1999) A novel 19F-NMR method for the investigation of the antioxidant capacity of biomolecules and biofluids. Free Radical Biol Med 27(3/4):356–363 5. Aue WP, Bartholdi E, Ernst RR (1975) Two-dimensional spectroscopy; application to nuclear magnetic resonance. J Chem Phys 64(5):2229–2246 6. Bailey NJC, Cooper P, Hadfield ST, Lenz EM, Lindon JC, Nicholson JK, Stanley PD, Wilson ID, Wright B, Taylor SD (2000) Application of directly coupled HPLC-NMR-MS/MS to the identification of metabolites of 5-trifluoromethylpyridone (2-hydroxy-5-trifluoromethylpyridine) in hydroponically grown plants. J Agric Food Chem 48(1):42–46 7. Bax A, Summers MF (1986) Proton and carbon-13 assignments from sensitivity-enhanced detection of heteronuclear multiple-bond connectivity by 2D multiple quantum NMR. J Am Chem Soc 108(8):2093–2094 8. Bayer E, Albert K, Nieder M, Grom E, Wolff G, Rindlisbacher M (1982) On-line coupling of liquid chromatography and high-field nuclear magnetic resonance spectrometry. Anal Chem 54(11):1747–1750 9. Bendall MR, Doddrell DM, Pegg DT (1981) Editing of carbon-13 NMR spectra. 1. A pulse sequence for the generation of subspectra. J Am Chem Soc 103(15):4603–4605 10. Benecke C, Grund R, Hohberger R, Kerber A, Laue R, Wieland T (1995) Molgen structure elucidation program from Chemical Concepts 11. Bigler P (1997) NMR spectroscopy processing strategies. VCH 12. Bladon P (2000) Personal communication to The NMR Newsletter 13. Bodenhausen G, Ruben DJ (1980) Natural abundance nitrogen-15 NMR by enhanced heteronuclear spectroscopy. Chem Phys Lett 69(1):185–188 14. Bothner-By AA, Stephens RL, Lee, J-M, Warren CD, Jeanloz RW (1984) Structure determination of a tetrasaccharide: transient nuclear Overhauser effects in the rotating frame J Am Chem Soc 106(3):811–813 15. Braun S, Kalinowski H-O, Berger S (1999) 150 and more basic NMR experiments. VCH 16. Braunshweiler L, Ernst RR (1983) Coherence transfer by isotropic mixing: application to proton correlation spectroscopy. J Magn Reson 53(3):521–528 17. Bravo P, Resnati G (1990) Preparation and properties of chiral fluoroorganic compounds. Tetrahedron Asym 1 (10):661–692 18. Bremser W (1978) HOSE – a novel substructure code. Anal Chim Acta 103(4):355–365 19. Canet D, Levy GC, Peat IR (1975) Time saving in 13C spin-lattice relaxation measurements by inversion recovery. J Mag Reson 18:199–204 20. Carpenter KA, Reynolds WF, Yang JP, Enriquez RG (1992) Further improvements in the FLOCK sequence. Mag Reson Chem 30:S35–S41 21. Cavalli L (1976) Annu Rep NMR Spectrosc 6B:43 22. Chapman JR (1995) Practical Organic mass spectrometry – a guide for chemical and biochemical analysis. Wiley 23. Cheatham S, Adams B, Kupce E (1999) 19F decoupling. Mag Moments 10(1):1–8 24. Chemical Concepts SpecInfo http//www chemicalconcepts com/ 25. Chen A, Shapiro MJ (1999) Affinity NMR. Anal Chem News Feat Oct 1:669–675 26. Cholli AL (1991) Pattern recognition in 2D NMR contour maps for studying molecules containing fluorines. Part I: Study of organofluorine ethers. Appl Spectrosc 45(5):839–848 27. Claridge TDW (1999) High resolution NMR techniques in organic chemistry. Pergamon

58

P.D. Stanley

28. Cornelissen G, Van Noort PCM, Nachtegaal G, Kentgens APM (2000) A solid-state fluorine-NMR study on hexafluorobenzene sorbed by sediments, polymers and active carbon. Environ Sci Technol 34(4):645–649 29. Crews P, Rodríguez J, Jaspars M (1998) Organic structure analysis. Oxford 30. Dale JA, Dull DL, Mosher HS (1969) a-Methoxy-a-trifluoromethylphenylacetic acid, a versatile reagent for the determination of enantiomeric composition of alchols and amines J Org Chem 34(9):2543–2549 31. Derome AE (1987) Modern NMR techniques for chemistry research. Pergamon 32. Dorn HC (1984) Proton NMR: a new detector. Anal Chem 56(6):747A–758A 33. Drew M, Orton E, Krolikowski P, Salvino JM, Kumar NV (2000) A method for quantitation of solid-phase synthesis using 19F NMR spectroscopy. J Comb Chem 2(1):8–9 34. Dungan CH,Van Wazer JR (1970) Compilation of reported 19F NMR chemical shifts 1951 to mid-1967. Wiley-Interscience 35. Emsley JW, Phillips L (1971) Prog Nucl Magn Reson Spectrosc 7:1 36. Emsley JW, Phillips L, Wray V (1976) Prog Nucl Magn Reson Spectrosc 10:83 37. Ernst RR, Anderson WA (1966) Application of Fourier transform spectroscopy to magnetic resonance. Rev Sci Instrum 37:93–102 38. Everett TS (1995) Nuclear magnetic resonance spectroscopy of organofluorine compounds. In: Hudlicky M (ed) Chemistry of organic fluorine compounds 11. ACS Monograph, vol 187, pp 1037–1086 39. Fields R (1977) Annu Rep NMR Spectrosc 7:1 40. Fields R (1972) Annu Rep NMR Spectrosc 5A:99 41. Foxall PJD, Parkinson JA, Sadler IH, Lindon JC, Nicholson JK (1993) Analysis of biological fluids using 600 MHz proton NMR spectroscopy: application of homonuclear twodimensional J-resolved spectroscopy to urine and blood plasma for spectral simplification and assignment. J Pharm Biomed Anal 11(1):21–31 42. Freeman R, Morris GA (1978) Experimental chemical shift correlation maps in nuclear magnetic resonance spectroscopy. J Chem Soc Chem Commun 684–686 43. Ghuari FYK, Wilson ID, Nicholson JK (1990) Fluorine-19 and proton-NMR studies of the metabolism of 4-trifluomethylbenzoic acid in the rat. Methodol Surv Biochem Anal 20:321–324 44. Greatbanks D, Pickford R (1987) Cyclodextrins as chiral complexing agents in water, and their application to optical purity measurements. Magn Reson Chem 25:208–215 45. Griesinger C, Schwalbe H, Schleucher J, Sattler M (1994) Two-dimensional NMR spectroscopy, 2nd edn. VCH, pp 457–580 46. Hamman SJ (1989) J Fluorine Chem 45:377 47. Heck AJR (1999) Ligand fishing by mass spectrometry. Spectrosc Eur 11:12–17 48. Hoeltzli SD, Frieden C (1996) Real-time refolding studies of 6– 19F-tryptophan labeled Escherichia coli dihydrofolate reductase using stopped flow nmr spectoscopy. Biochemistry 35(51):16,843–16,851 49. Hore PJ (1995) Oxford chemistry primers 32. Oxford University Press 50. Hornak JP (1999) The basics of MRI http://wwwcisritedu/htbooks/mri/insidehtm 51. Hurd RE (1991) Gradient enhanced spectroscopy. J Magn Reson 87(2):422–428 52. Hwang T-L, Kadkhodaei M, Mohebbi, A Shaka AJ (1992) Coherent and incoherent magnetisation transfer in the rotating frame. Magn Reson Chem 30:S24–S34 53. Hwang T-L, Shaka AJ (1992) Cross relaxation without TOCSY: transverse rotating-frame Overhauser effect spectroscopy J Am Chem Soc 114(8):3157–3159 54. Jenkins BG (1991) Detection of site-specific binding and co-binding of ligands to macromolecules using 19F NMR. Life Sci 48:1227–1240 55. Jones K, Mooney EF (1970) Annu Rep NMR Spectrosc 3:261 56. Jones K, Mooney EF (1971) Annu Rep NMR Spectrosc 4:391 57. Kay LE, Keifer P, Saarinen T (1992) Pure absorption gradient enhanced heteronuclear single quantum correlation spectroscopy with improved sensitivity. J Am Chem Soc 114(26):10,663–10,665

Principles and Topical Applications of 19F NMR Spectrometry

59

58. Kessler H, Griesinger C, Zarbock J, Loosli HR (1984) Assignment of carbonyl carbons and sequence analysis in peptides by heteronuclear shift correlation via small coupling constants with broadband decoupling in t1 (COLOC). J Magn Reson 57(2):331–336 59. Kohl SD, Toscano PJ, Hou WH, Rice JA (2000) Solid-state 19F NMR investigation of hexafluorobenzene sorption to soil organic matter. Environ Sci Technol 34(1):204–210 60. LLoyd CH, Scrimgeour SN, Hunter G, Chudek JA, Lane DM, McDonald PJ (1999) Solid state spatially resolved 1H and 19F nuclear magnetic resonance spectroscopy of dental materials by stray field imaging. J Mater Sci Mater Med 10(6):369–373 61. Magnetic resonance microsensors. Savoy, Illinois USA 62. Martin ML, Delpeuch J-J, Martin GJ (1980) Practical NMR Spectroscopy. Heyden, pp 191–193 63. Mazzola EP, Borsetti AP, Page SW, Bristol DW (1984) Determination of pesticide residues in foods by flourine-19 Fourier transform nuclear magnetic resonance spectroscopy. J Agric Food Chem 32(5):1102–1103 64. McIntyre DJO, McCoy CL, Griffiths JR (1999) Tumour oxygenation measurements by 19F magnetic resonance imaging of perfluorocarbons. Curr Sci 76(6):753–761 65. Mooney EF (1970) Introduction to 19F NMR Spectroscopy. Heyden 66. Mooney EF, Winson PH (1968) Annu Rep NMR Spectrosc 1:243 67. Moore JM (1999) NMR screening in drug discovery. Curr Opin Biotechnol 10:54–58 68. Moore JM (1999) NMR techniques for characterisation of ligand binding: utility for lead generation and optimization in drug discovery. Peptides Sci 151:221–243 69. Mortimer RD, Black DB, Dawson BA (1994) Pesticide residue analysis in foods by NMR. 3. Comparison of 19F NMR and GC-ECD for analysing trifluralin residues in field grown carrots. J Agric Food Chem 42(8):1713–1716 70. Mortimer RD, Dawson BA (1991) Using fluorine-19 NMR for trace analysis of fluorinated pesticides in food products. J Agric Food Chem 39(10):1781–1785 71. Moskau D, Richter C, Kovacs H, Salzmann M, Baselgia L, Haeberli M, Marek D, Schett O (2001) Highest sensitivity for cutting-edge NMR applications: 600 MHz CryoProbes. Bruker Report 149, pp 19–21 72. Neuhaus D, Williamson M (1989) The nuclear Overhauser effect in structural and conformational analysis. VCH, pp 217–240 73. Neuhaus D, Williamson M (1989) The nuclear Overhauser effect in structural and conformational analysis. VCH, pp 253–305 74. Nicholson JK, Wilson ID (1989) High resolution proton magnetic resonance spectroscopy of biological fluids. Prog Nucl Magn Reson Spectrosc 21(4/5):449–501 75. Olson DL, Lacey ME, Sweedler JV (1998) High-resolution microcoil NMR for analysis of mass-limited nanoliter samples. Anal Chem 70(3):645–650 76. Palmer AG III, Cavanagh J, Wright PE, Rance M (1991) Sensitivity improvement in proton-detected two-dimensional heteronuclear correlation NMR spectroscopy. J Magn Reson 93(1):151–170 77. Parella T, Sanchez-Ferrando F, Virgili A (1997) Improved sensitivity in gradient-based 1D and 2D multiplicity edited HSQC experiments. J Magn Reson 126(2):274–277 78. Patt S, Shoolery JN (1982) Attached proton test for carbon-13 NMR. J Magn Reson 46(3):535–539 79. Pekar J, Sinnwell T, Ligeti L, Chesnick AS, Frank JA, McLaughlin AC (1995) Simultaneous measurement of cerebral oxygen consumption and blood flow using 17O and 19F magnetic resonance imaging. Cereb Blood Flow Metab 15(2):312–20 80. Peng C, Yuan S, Zheng J (1994) Efficient application of 2D NMR correlation information in computer-assisted structure elucidation of complex natural products. J Chem Inf Comput Sci 34(4):805–813 81. Peng C, Yuan S, Zheng J (1995) Practical computer-assisted structure elucidation for complex natural products: efficient use of ambiguous 2D NMR correlation information. J Chem Inf Comput Sci 35(3):539–546

60

P.D. Stanley

82. Pirkle WH, Beare SD (1967) Nonequivalence of the nuclear magnetic resonance spectra of enantiomers in optically active solvents. IV. Assignment of absolute configuration. J Am Chem Soc 89(21):5485–5487 83. Randall EW (1997) 1H and 19F magnetic resonance imaging of solid paramagnetic compounds using large magnetic field gradients and Hahn echoes. Solid State Nucl Magn Reson 8(3):173–178 84. Reynolds WF,Yu M, Enriquez RG, Leon I (1997) Investigation of the advantages and limitations of forward linear prediction for processing 2D data sets. Magn Reson Chem 35(8):505–519 85. Rinaldi PL (1983) Heteronuclear 2D-NOE spectroscopy. J Am Chem Soc 105(15): 5167–5168 86. Roberts GCK (2000) Applications of NMR in drug discovery. Drug Discovery Today 5:230–240 87. Russell DJ, Hadden CE, Martin GE, Gibson AA, Zens AP, Carolan JL (2000) J Nat Prod 63(8):1047–1049 88. Rüdiger A-H, Rüdiger M, Carl UD, Chakraborty T, Roepstorff P, Wehland J (1999) Affinity mass spectrometry approaches for the analysis of protein-protein interaction and complex mixtures of peptides. Anal Biochem 275:162–170 89. Sanders JKM, Hunter BK (1993) Modern NMR spectroscopy: a guide for chemists. Oxford University Press 90. Scarfe GB, Clayton E, Wilson ID, Nicholson JK (2000) Identification and quantification of metabolites of 2,3,5,6-tetrafluoro-4-trifluoromethylaniline in rat urine using fluorine-19 nuclear magnetic resonance spectroscopy, high-performance liquid chromatography-nuclear magnetic resonance spectroscopy and, high-performance liquid chromatography-mass spectrometry. J Chromatogr B Biomed Appl 748(1):311–319 91. Schriber H, Pretsch E (1997) General characterisation of good-list and bad-list entries for structure generators from spectra. J Chem Inf Comput Sci 37:879–883 92. ScienceSoft, LLC, Salt Lake City, Utah USA NMRanalyst/FRED http //www sciencesoft net/index html 93. Serre AM, Roby C, Roscher A, Nurit F, Euvrard M, Tissut M (1997) Comparative detection of fluorinated xenobitotics and their metabolites through 19F NMR or 14C label in plant cells. J Agric Food Chem 45:242–248 94. Shapiro MJ, Kumaravel G, Petter RC, Beveridge R (1996) 19F NMR monitoring of an SNAr reaction on solid support. Tetrahedron Lett 37(27):4671–4674 95. Shapiro MJ, Wareing JR (1998) NMR methods in combinatorial chemistry. Curr Opinion Chem Biol 2:372–375 96. Shuker SB, Hajduk PJ, Meadows RP, Fesik SW (1996) Discovering high-affinity ligands for proteins: SAR by NMR. Science 274:1531–1534 97. Sidelmann UG, Nicholls AW, Meadows PE, Gilbert JW, Lindon JC, Wilson ID, Nicholson JK (1996) High-performance liquid chromatography directly coupled to 19F and 1H NMR for the analysis of mixtures of isomeric ester glucuronide conjugates of trifluoromethylbenzoic acids. J Chromatogr A 728:377–385 98. Smallcombe SH, Patt SL, Keifer PA (1995) WET solvent suppression and its application to LC NMR and high-resolution spectroscopy. J Magn Reson Ser A 117(2):295–303 99. Spectrum Research, LLC, Madison, Wisconsin USA http //www specres com/nmrsams asp 100. Spraul M, Hofmann M, Wilson ID, Lenz E, Nicholson JK (1993) Coupling of HPLC with 19F and 1H-NMR spectroscopy to investigate the human urinary excretion of flurbiprofen metabolites. J Pharm Biomed Anal 11(10):1009–1015 101. Stockman BJ (1998) NMR spectroscopy as a tool for structure based design. Prog Nucl Magn Reson Spectrosc 33(2):109–151 102. Stott K, Keeler J (1996) Gradient-enhanced one-dimensional heteronuclear NOE experiments with 1H detection. Magn Reson Chem 34(7):554–558 103. Sweatman BC, Farrant RD, Holmes E, Ghuari FYK, Nicholson JK, Lindon JC (1993) 600 MHz proton-NMR spectroscopy of human cerebrospinal fluid: effects of sample manipulation and assignment of resonances. J Pharm Biomed Anal 11(8):651–664

Principles and Topical Applications of 19F NMR Spectrometry

61

104. Svensson A, Fex T, Kihlberg J (1996) Use of 19F NMR spectroscopy to evaluate reactions in solid phase organic synthesis. Tetrahedron Lett 37 (42):7649–7652 105. Uhrin D, Beatty EJ, Sandór P, McSparron S, Wilken J, Starkmann B, Young DW, Barlow PN, Ramage R (1999) 19F-1H Correlation at no extra cost. Magn Moments 10(1):9–15 106. Vikic-Topic D, Meic Z (1986) Carbon-fluorine long range coupling for estimating molecular conformations. J Mol Struct: 142–148 107. Wagner R, Berger S (1996) Gradient-selected NOESY – a fourfold reduction of the measurement time for NOESY experiments. J Magn Reson Ser A 123(1):119–121 108. Warne MA, Lenz EM, Osborn D, Weeks JM, Nicholson JK (2000) An NMR-based metabonomic investigation of the toxic effects of 3-trifluoromethylaniline on the earthworm Eisenia venata. Biomarkers 5(1):56–72 109. Willoughby R, Sheehan E, Mitrovich S (1998) A global view of LC/MS – how to solve your most challenging analytical problems. Global View Publishing 110. Wilson ID, Nicholson JK (1987) Solid-phase extraction chromatography and nuclear magnetic resonance spectrometry for the identification and isolation of drug metabolites in urine. Anal Chem 59(23):2830–2832 111. Wray V (1980) Annu Rep NMR Spectrosc 10B:1 112. Wray V (1983) Annu Rep NMR Spectrosc 14:1 113. Zhou L, Rajabzadeh M, Traficante D, Cho BP (1997) Conformational heterogeneity of arylamine-modified DNA:19F NMR evidence. J Am Chem Soc 119(23):5384–5389

CHAPTER 2

Partitioning of Organofluorine Compounds in the Environment David A. Ellis 1, Thomas M. Cahill 3, Scott A. Mabury 2, Ian T. Cousins 4, Donald Mackay 5 1 2 3 4 5

Chemistry Department, University of Toronto, Toronto, Ontario M5S 3H6, Canada. E-mail: [email protected] Chemistry Department, University of Toronto, Toronto, Ontario M5S 3H6, Canada. E-mail: [email protected] Canadian Environmental Modelling Centre, Trent University, Peterborough, Ontario K9J 7B8, Canada. E-mail: [email protected] Canadian Environmental Modelling Centre, Trent University, Peterborough, Ontario K9J 7B8, Canada. E-mail: [email protected] Canadian Environmental Modelling Centre, Trent University, Peterborough, Ontario K9J 7B8, Canada. E-mail: [email protected]

This chapter describes the partitioning properties of organofluorine compounds in the environment. Partitioning in the air-water-octanol system is discussed using these pure substances as surrogates for real environmental systems consisting of air, water, natural organic matter and lipids. It is shown that the substitution of fluorine for hydrogen in alkanes, aromatics and carboxylic acids causes significant changes in properties such as vapor pressure, aqueous solubility, octanol-water partition coefficient and acid dissociation constant. These changes are quite different in magnitude from the corresponding changes caused by other better-studied halogens, namely chlorine and bromine. Perfluorinated substances have unique properties imparted by their minimal intermolecular interactions. The environmental implications of these properties are illustrated using simple multimedia partitioning models showing that organofluorine compounds behave quite differently than organochlorine analogs and more closely resemble the corresponding non-halogenated compounds. Keywords. Fluorine, Organofluorine, Haloacetic acids, Fluorocarbon, Halogen, Hydrofluoro-

carbon, Perfluorinated compounds, Halocarbon

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

1

Introduction

2

Elemental Fluorine . . . . . . . . . . . . . . . . . . . . . . . . . . 65

3

Physico-Chemical Properties

3.1 3.2 3.3

Short-Chain Fluorocarbons . . . . . . . . . . . . . . . . . . . . . 66 Fluorinated Benzenes . . . . . . . . . . . . . . . . . . . . . . . . 69 Fluorinated Acetic Acids . . . . . . . . . . . . . . . . . . . . . . . 71

4

Implication for Environmental Fate – Evaluative Modeling of the Fluoro- and Chlorobenzenes . . . . . . . . . . . . . . . . . . . . 75

5

Conclusions

6

References

. . . . . . . . . . . . . . . . . . . . 65

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 The Handbook of Environmental Chemistry Vol. 3, Part N Organofluorines (ed. by A.H. Neilson) © Springer-Verlag Berlin Heidelberg 2002

64

D.A. Ellis et al.

List of Abbreviations BCF CFC DDT DFA EPIWIN EQC HCFC HFC Ka KOW MFA PCB TFA VP

bioconcentration factor chlorofluorocarbon para-dichlorodiphenyltrichloroethane difluoroacetic acid Estimations Programs Interface for Windows Equilibrium Criterion Model hydrochlorofluorocarbon hydrofluorocarbons equilibrium acid dissociation constant octanol-water partition coefficient monofluoroacetic acid polychlorinated biphenyl trifluoroacetic acid vapor pressure

1 Introduction Since the publication of Rachel Carson’s “Silent Spring” in 1962, there has been a continuing focus on the environmental fate and effects of organochlorine compounds such as PCBs, “dioxins” and pesticides such as DDT and Mirex. The physico-chemical effects of substituting chlorine for hydrogen in organic molecules are now well known and highly predictable. In general, this substitution causes a marked increase in molar mass, which results in a reduced vapor pressure. The larger molecule has a larger surface area which usually induces a stronger hydrophobic effect. This is manifested as a decrease in water solubility and a corresponding increase in the octanol-water partition coefficient, KOW . The substitution of a hydrogen atom with chlorine also imparts stability to degradation by both abiotic reactions, as is the case with OH radical attack, and biotic reactions such as microbial degradation. The combination of increased hydrophobicity and stability results in molecular characteristics that heighten environmental concerns. Much less attention has been paid to the other halogens, although there is increasing concern about organobromine substances such as brominated fire retardants. To a first approximation, bromine appears to behave similarly to chlorine, but there are quantitative differences attributable to the size and mass of the bromine atom, i.e. bromine is approximately twice the mass of chlorine and has a slightly larger atomic radius. The carbon-chlorine bond is also considerably stronger than the carbon-bromine bond, namely 397±29 compared with 280±21 kJ/mol respectively. Organofluorine compounds have received relatively little attention when compared with organochlorine compounds.An obvious exception was the study of the volatile chlorofluorocarbons (CFCs) and related compounds that were used as refrigerants and were implicated as agents responsible for stratospheric ozone depletion. Although there have been several reviews of the commercial [2] and biological aspects of organofluorine com-

Partitioning of Organofluorine Compounds in the Environment

65

pounds, the first comprehensive critical review of their environmental sources, fate and effects was by Key et al. [26]. The goals of this review are to: 1. compile selected available data on the environmentally relevant physico-chemical properties of organofluorine compounds; 2. discuss the nature of the physico-chemical properties induced in the molecule through fluorination and, highlight the differences between fluorinated molecules and their corresponding chlorine and bromine analogs; 3. apply an evaluative environmental model based upon the physico-chemical properties of a selected class of fluorinated organics in order to investigate their likely environmental distribution relative to their non-halogenated and chlorinated counterparts. It transpires that, as is often the case, the first member of a “homologous series”, in this case fluorine, displays a number of unique characteristics that impart unusual partitioning properties in both model systems and the environment.

2 Elemental Fluorine Although this review focuses on the organofluorine compounds, a brief discussion of the properties of elemental fluorine is useful to assist in understanding some of the properties of organofluorine compounds. Table 1 shows some of the physical properties of fluorine relative to chlorine and bromine. Fluorine gas, F2 , is the strongest known oxidizing agent and it can directly react with most elements. The extreme reactivity of fluorine is exemplified by the fact that even water “burns” in the presence of fluorine gas to form hydrofluoric acid and oxygen. Hydrocarbons react directly with F2 to form fluorinated hydrocarbons. For example: CH4 +4 F2 Æ 4 HF+CF4 Fluorine can even react directly with the heavier noble gases, such as xenon, to form fluorinated compounds (XeF2 , XeF4 , XeF6) [11]. Fluorine has an electronegativity of 4.0 on the Pauling scale, which is the highest of all the elements. This extreme electronegativity results in fluorine drawing electrons from adjacent atoms, thus creating polar bonds. For example, the electronegativity difference between fluorine and carbon is 1.5 units, which indicates that the electrons are shared unequally between the two atoms and that there is a partial negative charge on the fluorine and a partial positive charge on the carbon.

3 Physico-Chemical Properties In this section we review the physico-chemical properties of organofluorine compounds focusing on the highly fluorinated compounds where fluorine is the dominant functional group. The properties of simple fluorocarbons, fluorobenzenes and fluorinated acetic acids are described and discussed, with particular

66

D.A. Ellis et al.

attention paid to the influence imparted to the molecule by fluorine. Some compounds, such as the widely used herbicide trifluralin, contain a trifluoro group, but the physico-chemical properties of the molecule are dominated by other functional groups, so they are not considered here. 3.1 Short-Chain Fluorocarbons

Fluorine is commonly used in the synthesis of volatile short-chain halocarbons that are used as refrigerants, aerosol propellants and foam blowing agents. Initially, the chlorofluorocarbons (CFCs) were used for this purpose, but they were implicated in stratospheric ozone depletion [39] and have subsequently been replaced by the hydrofluorocarbons (HFCs) [51]. The CFCs were typically fully halogenated methanes while the HFCs are generally partially fluorinated ethanes such as HFC 134a (1,1,1,2-tetrafluoroethane), which is the most commonly used HFC. The effect of fluorination on the vapor pressure of short chain hydrocarbons illustrates its unusual behavior. For alkanes, the substitution of one fluorine for hydrogen results in a decrease of the vapor pressure of the compound (Fig. 1). In this respect, fluorine behaves in a consistent fashion with respect to the other halogens. In each case, the vapor pressure of the monofluorocarbon is less than the vapor pressure of the parent compound by a factor of 2.5 to 5.0. The addition of multiple chlorine and bromine atoms to the hydrocarbon causes a similar trend in vapor pressure reduction; while the addition of multiple fluorines to a hydrocarbon results in anomalous behavior. Figure 2 shows the effect of adding multiple fluorines to the volatility (as expressed by boiling

Fig. 1. Influence of monohalogenation on the vapor pressure of the short chain hydrocarbons.

Hydrocarbon data from Lide et al. [29], fluorocarbon data from [6, 20, 35, 41], and the other halocarbon data from Mackay et al. [31]

Partitioning of Organofluorine Compounds in the Environment

67

Fig. 2. The boiling points of the fluorinated, chlorinated and brominated methanes. (Data from

[29])

point) of methane. The addition of one or two fluorines to methane causes the expected increase in the boiling point of the compound, which indicates that the partially fluorinated methanes are less volatile than methane. However, the addition of the third and fourth fluorines cause a decrease in boiling point, thus indicating that they are more volatile than the partially fluorinated methanes. The increase in volatility occurs despite the increase in molar mass. In contrast to fluorine, the addition of each chlorine or bromine to methane causes a consistent increase in the boiling point or a decrease in vapor pressure, which is expected since the molecule has a higher molar mass. A similar trend is observed with the fluorinated ethanes where the partially fluorinated ethanes have higher boiling points (–47.3 to 30.7 °C) than either ethane (–88.6 °C) or perfluoroethane (–79 °C) [29]. Fluorine is therefore unique in that perfluorination results in an increase in vapor pressure over the partially fluorinated hydrocarbons.

Fig. 3. Anticipated dipole-dipole interaction between two partially-fluorinated methanes

68

D.A. Ellis et al.

Table 1. Selected properties of fluorine, chlorine and bromine atoms where “X” is halogen [29]

Atomic mass Ionic radius (nm) Bond length in CX4 (nm) C–X Bond energy (kJ/mol) Electronegativity

Fluorine

Chlorine

Bromine

18.998 0.133 0.1323 552 4.0

35.453 0.181 0.1767 397±29 3.0

79.904 0.196 0.1935 280±21 2.8

A suggested mechanism for this anomalous behavior is that the high electronegativity of the fluorine creates a polar bond with the carbon. The unequal sharing of electrons creates a dipole moment in the partially fluorinated methanes. The dipole-dipole interaction between molecules (Fig. 3) results in a reduction of the vapor pressure and increases the boiling point. Tetrafluoromethane, however, has no overall net dipole moment since polarization of the C–F bonds occurs equally in all directions. The observed change in the boiling point trend observed for the tri- and tetrafluoromethanes, when compared with their other halogen counterparts, can be explained through changes in the atomic size and bond length (Table 1). As previously described, each halogen causes a polarization of the bond between it and the central carbon atom. The shorter bond length between carbon and fluorine results in the exterior negative shell being held more tightly to the carbon atom. These two inherent properties renderthe carbon atom less susceptible to intermolecular interactions causing these molecules to show enhanced volatility.

Fig. 4. Comparison of boiling points of the perfluoroalkanes to the normal alkanes. Fluorocarbon data from [1, 4, 8, 22, 29, 36–38, 43, 44]. Hydrocarbon data from [7, 12, 18, 21, 24, 29, 40, 42, 47, 48, 50]

69

Partitioning of Organofluorine Compounds in the Environment

Table 2. Comparison of the boiling points of noble gases, perfluorocarbons and hydrocarbons

of similar molecular weights. (Data from [29]) Molecular weight analogs (molecular wt. in parenthesis)

Boiling point (°C)

Difference (°C) between noble gas and organic b.p.

Kr (83.8) CF4 (88.0) Hexane (86.2)

–152 –129 69

– 23 221

Xe (131.3) C2F6 (138.0) Nonane (128.3)

–107 –79 151

– 28 258

–62 4 287

– 66 349

Rn (222) C4F10 (238.0) Hexadecane (226.5)

A related and unusual property of the perfluorinated hydrocarbons is that they become more volatile than the parent hydrocarbon when the chain length exceeds four carbons. Figure 4 shows the boiling point for hydrocarbons and perfluorocarbons of a variety of carbon numbers. Perfluorinated linear chain aliphatic compounds exhibit unique properties when compared to those of nonfluorinated analogs. The origins of these unusual properties lie in the interaction between neighboring –CF2– units. Strong electronic repulsion of adjacent fluorine atoms results in the backbone of the chain being held rigid. It is this rigidity, coupled with the large partial negative charge associated with each fluorine atom, results in the vapor pressure of these molecules being substantially higher than would be predicted based purely upon the mass of the molecules. It is interesting to compare the noble gases, perfluorocarbons and normal alkanes of similar molecular weights as in Table 2. In each case, the perfluorocarbon behaves more like a noble gas than a hydrocarbon of corresponding molecular weight. This suggests that there are minimal intermolecular interactions in perfluorocarbons and that their volatility may be related more to their molar mass than the hydrocarbons. In conclusion, the addition of fluorine to a hydrocarbon generally results in a drop in vapor pressure, as would be expected in a molecule of higher molar mass and increased dipole moment. The perfluorocarbons behave differently and have higher vapor pressures indicating that dipole induced intermolecular interactions in the liquid phase are reduced. 3.2 Fluorinated Benzenes

The effects of adding a single fluorine to benzene is similar to that of the alkanes. Figure 5 shows a decrease in the vapor pressure resulting from monofluorination, which is similar to that observed for the alkanes. The decrease in vapor pressure is consistent with the vapor pressure decrease caused by the addi-

70

D.A. Ellis et al.

Fig. 5. The octanol-water partition coefficient and vapor pressure of the monohalogenated ben-

zenes at 25 °C. (Data from [31])

tion of other halogens to benzene. In addition, the octanol-water partition coefficient (KOW) also increases by the addition of a single fluorine. The increase in log KOW is probably related to the increase in molar volume or surface area. Whereas multiple chlorination of benzene causes a consistent change in the physico-chemical properties of the molecule, the influence of multiple fluorination is less predictable. Figure 6 shows the boiling points of the fluorinated and chlorinated benzenes. An increase in fluorination of benzene has little effect on

Fig. 6. The boiling points for the fluorinated, chlorinated and brominated benzenes. The “boiling point” of hexachlorobenzene actually represents a sublimation point. Data for pentabromobenzene and hexabromobenzene where not available. All data from [29]

Partitioning of Organofluorine Compounds in the Environment

71

Fig. 7. The octanol-water partition coefficient (KOW) for the fluorinated and chlorinated ben-

zenes. (Chlorobenzene data from [31]; fluorobenzene data from [3, 17, 19, 23, 52])

the boiling point of the molecule. The partially fluorinated benzenes are slightly less volatile than either benzene or hexafluorobenzene. In contrast, the addition of chlorine or bromine to benzene increases the boiling point of the molecule by consistent increments as a function of the number of halogens. The effect of multiple fluorination on the octanol-water partition coefficient also contrasts with the behavior of chlorine. Figure 7 shows the log KOW of the fluorobenzenes and the chlorobenzenes. While the addition of each chlorine causes a consistent increase in log KOW , the addition of fluorine initially causes an increase in log KOW , but then the addition of the last two fluorines causes a decrease. This behavior results in difficulties in establishing quantitative structural property relationships based on constant atom or group contributions. 3.3 Fluorinated Acetic Acids

Fluorinated acetic acids are frequent degradation products of some organic fluorine compounds. Trifluoroacetic acid has received considerable attention since it is a known breakdown product of certain pesticides [13] and the HFC refrigerants and aerosol propellants [51]. Chlorodifluoroacetic acid may be formed via the degradation of CFC-113 in the stratosphere or through the direct tropospheric degradation of hydrochlorofluorocarbons (HCFCs) such as HCFC-142b [33]. Fluorinated acetic acids are now ubiquitous in the environment [5, 9, 14, 25, 45]. Table 3 gives the melting and boiling points for acetic acid and the fluorinated acetic acids, the latter indicating their relative volatilities. The replacement of one hydrogen with a fluorine increases the boiling point of the acid, which indicates

72

D.A. Ellis et al.

Table 3. The melting point and boiling point of the fluorinated acetic acids

Chemical

Melting point (°C)

Boiling point (°C)

CH3COOH CH2FCOOH CHF2COOH CF3COOH

16.6 35.2 –0.3 –15.2

118.9 165 134.2 72.4

Data from Lide [29].

that monofluoroacetic acid (MFA) is less volatile than acetic acid. The decreased volatility of the mono-fluorinated species follows the same general trend as the fluorinated alkanes where the first fluorine added to the molecule decreases the vapor pressure. The boiling point of difluoroacetic acid (DFA) is lower than MFA, which indicates that DFA is more volatile than MFA. Lastly, trifluoroacetic acid (TFA) has the lowest boiling point, and thus the greatest volatility. This trend is similar in nature to the vapor pressure of the alkanes where the species with perfluorinated carbons have higher vapor pressures than partially fluorinated species. In contrast to the alkanes, in which methane was more volatile than tetrafluoromethane, TFA is more volatile than the parent compound of acetic acid. The addition of halogens to acetic acid increases the acid-dissociation constant (Ka) of the acids. Ka is the equilibrium constant between the unionized and the ionized forms of the acid and is numerically defined as [H3O+][A–]/[HA]. Larger values of Ka imply stronger acids that are more likely to lose protons and be found in the ionized form. Figure 8 shows the effect of halogen substitution on the Ka of the monohalogenated acetic acid. Fluorine causes the greatest increase in Ka of the monohalogenated acetic acids, which indicates that MFA is more easily ionized than the other monohaloacetic acids.

Fig. 8. The influence of halogen electronegativity on the acid-dissociation constants of the

haloacetic acids. (Data from [46])

Partitioning of Organofluorine Compounds in the Environment

73

Fig. 9. The influence of multiple fluorination of the acid-dissociation constant of the fluorinated acetic acids. An increase in the number of fluorines causes an increase in Ka that indicates that the highly fluorinated acetic acids are more prone to ionization. (Data from [46])

The Ka is observed to increase with increasing fluorination (Fig. 9). The Ka values indicate that the perfluorinated acids will be completely dissociated under normal aqueous and physiological conditions. Upon deprotonation of the acid, the resultant negative charge is stabilized through the bond inductive effect of the fluorine atoms (Fig. 10). Thus, as the number and magnitude of electronegative elements that are contained within the acetic acid increases, so too does the Ka . Perfluorination of carboxylic and sulfonic acids also imparts a high degree of chemical stability or resistance to degradation. The stability of perfluorinated compounds is demonstrated by the lack of microbial degradation of perfluorinated carboxylates and sulfonates compared to their partially fluorinated counterparts. Recent studies by Ellis et al. [15], Emptage et al. [16] and Cahill et al. [10] have shown that TFA is stable under anaerobic and field microcosm conditions. Earlier research indicated that reductive dehalogenation of TFA could occur [49], although these results have not been replicated by other re-

Fig. 10. Inductive stabilization of negative charge on trifluoroacetate

74

D.A. Ellis et al.

Fig. 11. Examples of fluorinated compounds that are microbially degradable and compounds

that not been shown to be degraded by microbial mechanisms. The primary difference between the compounds is the presence of one or more hydrogens that represent a reactive site. (sulfonate data from [27], fluorinated acetate data from [16], trichloroacetate data from [15])

searchers [16]. A recent study by Kim et al. [28] demonstrated that TFA could be degraded in an anaerobic reactor at elevated temperatures in the presence of ethanol, so TFA degradation may occur in the field, but probably at rates too slow to be readily observed. Partially fluorinated acetates, especially fluoroacetate that is highly toxic to biota, have been shown to be degradable by microbes (references in Chap. 7 and [16]). Key et al. [27] demonstrated that only fluorinated sulfonates that contained hydrogen could be degraded by bacteria. In all cases, the compounds that lacked hydrogen were recalcitrant to degradation by aerobic bacteria [27]. Figure 11 shows the substances that were degraded compared to the non-metabolized compounds. In contrast to trifluoroacetate, trichloroacetate is degradable in field microcosm conditions [15].

Partitioning of Organofluorine Compounds in the Environment

75

4 Implication for Environmental Fate – Evaluative Modeling of the Fluoro- and Chlorobenzenes A full evaluation of the probable environmental fate of the fluorinated organic compounds discussed earlier is beyond the scope of this chapter. It is, however, important to appreciate the cardinal significance of these physico-chemical properties in determining their environmental fate. This is illustrated by a discussion of fluorobenzenes and a comparison with the analogous chlorobenzenes. As noted earlier, there may be a perception that a compound with fluorine substituents will behave somewhat similarly to its chlorinated analog. This proves not to be the case for partitioning, since fluorinated compounds clearly display anomalous behavior in the bromine-chlorine-fluorine series. The calculated behavior of a chemical in a model environment provides a basis for evaluating its environmental fate.An environmental impact assessment of a compound relies on a number of factors: its chemical properties, which largely determine the phase into which the chemical will partition, its tendency to bioaccumulate, the possibility of long-range transport, and the primary mechanisms for its loss that determine its environmental persistence. In addition, numerical results from modeling provide “benchmark” environmental data for a chemical that can be compared with those for chemicals with established properties in a comparable environment. The Equilibrium Criterion or EQC model [30] is a widely used evaluative model that treats an area of 105 km2 with about 10% of the area being covered by water. The temperature in the EQC environment is set at 25 °C, which is a common temperature at which physico-chemical properties are measured. Evaluative modeling assessments usually progress through three stages of complexity; Levels I, II and III. Each subsequent level requires more detailed information or includes additional processes providing a step-wise increase in understanding of the chemical behavior in the environment. Mackay [32] and Mackay et al. [30] provide a complete description of these calculations with examples at each level of complexity. The calculations are conducted using the concept of fugacity which simplifies the equations and facilitates interpretation of the results. Here we undertake evaluative fate modeling of benzene and selected fluoroand chlorobenzenes, specifically; monofluorobenzene, 1,2,4-trifluorobenzene, hexafluorobenzene (or perfluorobenzene), monochlorobenzene, 1,2,4-trichlorobenzene and hexachlorobenzene (or perchlorobenzene). The aim is to determine how substitution of fluorine and chlorine on the benzene ring affects environmental partitioning and fate. The physico-chemical data required for the EQC model are vapor pressures, water solubilities, octanol-water partition coefficients, melting points and reaction rate constants in air, water, soils and sediments. Physico-chemical property data for benzene and the chlorobenzenes were taken from a handbook of physico-chemical properties [31]. Melting points for the fluorobenzenes were taken from the CRC Handbook of Chemistry and Physics [29] and vapor pressures, water solubilities and octanol-water partition coefficients were obtained from Ellis and Mabury of the University of Toronto (personal communication).

76

D.A. Ellis et al.

Table 4. Physico-chemical property data for fluoro- and chlorobenzenes used as inputs to the EQC model

Compound

Benzene Monofluorobenzene 1,2,4-Trifluorobenzene Hexafluorobenzene Monochlorobenzene 1,2,4-Trichlorobenzene Hexachlorobenzene a b c

d

Molar mass

Water solubility (g/mol) (g/m3) a

Vapor pressure (Pa) a

Log Melting Assumed 1st order reaction KOW a point half lives (hours) c, d (°C) b

78.1

1780

12700

2.13

96.1

1610

12300

2.82

132.1

961

7670

186.1

120

112.6 181.5 284.8

Water Soil Sediment

209 1480

2960 13300

–42.2

372 1480

2960 13300

3.09

– 5.5

249 4320

8640 38900

2010

1.33

5.3

1500 4320

8640 38900

484

1580

2.8

–45.6

40

61

4.1

16

467 2520

5040 22700

0.0023 5.5

231

951 4320

8640 38900

0.005

5.53

Air

333

444

888

4000

Benzene and chlorobenzene data taken from Mackay et al. [31], fluorobenzene data obtained from Ellis and Mabury, University of Toronto by personal communication. Obtained from CRC handbook of chemistry and physics [29]. Estimated using the Syracuse Research Corporation EPIWIN software [34]. The following EPIWIN setting were used: BIOWIN (Ultimate)/AOP programs were used for estimating air, water, soil and sediment half-lives, the more conservative Alternate BIOWIN half-life values were used, and BIOWIN half-life factors water :soil:sediment were set to 1:2:9. The EPIWIN settings for calculation of half-life values are described in Meylan [34] and are not repeated here. It is noted that the reaction half lives estimated for hexafluorobenzene are conservative as it is expected that the compound will have oxidative and hydrolytic reaction rates markedly greater than hexachlorobenzene in all compartments of the environment.

First-order reaction rate constants were estimated using the Estimations Programs Interface for Windows (EPIWIN) software [34]. EPIWIN is a Windowsbased estimation program commercially available from the Syracuse Research Corporation that can estimate physico-chemical properties and reaction rates from structure alone. Although reaction rate constants for benzene and the chlorobenzenes have previously been estimated by Mackay et al. [31], it was decided, for consistency, to use EPIWIN for estimating the reaction rate constants of half-lives for all the compounds modeled. It must be appreciated that the halflives are subject to considerable error and variability, especially when they exceed a month in duration. For example, the assigned half-lives of 8640 h in soil for HCB and HFB (nearly one year) are highly speculative, although it is likely that they are correct in magnitude. Physico-chemical property and reaction rate data used as input to the EQC model are summarized in Table 4. A summary of results from the EQC Level I and Level II modeling is presented in Table 5. Level I EQC modeling indicates that under equilibrium, and steady state conditions, the monochloro- and trichlorobenzene partition mostly to air,

77

Partitioning of Organofluorine Compounds in the Environment Table 5. Summary of results of Level I and II simulations using the EQC model

Compound

Benzene Monofluorobenzene 1,2,4-Trifluorobenzene Hexafluorobenzene Monochlorobenzene 1,2,4-Trichlorobenzene Hexachlorobenzene a

Level I distribution

Level II

% in Air

% in Water

% in Soil

% in Sediment

99.0

0.88

0.11

0.002

98.9

0.67

0.39

0.009

99.0

0.47

0.51

0.011

99.8

0.16

0.003

0.00007

98.0

1.24

0.70

0.015

81.8

1.47

16.3

8.42

0.32

89.2

BCF a

6.7

Overall % Loss by residence reaction time (days) 3.2

24.9

33

3.5

15.7

62

3.3

21.8

4.0

4.4

32

3.5

17.4

0.36

630

4.4

13.1

1.98

16000

1.1

43

13.7

BCF is the bioconcentration factor deduced as 0.05 KOW , i.e. equilibrium partitioning into fish of 5% lipid.

whereas the hexachlorobenzene partitions mostly to the organic carbon component of soils. Thus each addition of chlorine tends to decrease the volatility and increase the hydrophobicity of the chlorobenzenes. The tendency for the more highly chlorinated, hydrophobic chlorobenzenes to be transported into biota is demonstrated by the bioconcentration factor (BCF) from water to fish for hexachlorobenzene being 2000 times that of benzene. In contrast, increasing the level of fluorination does not markedly increase hydrophobicity; indeed the hexafluorinated compound actually is less hydrophobic than benzene. The fluorobenzenes partition almost exclusively to air, i.e. their partitioning behavior is similar to benzene and monochlorobenzene. Hexafluorobenzene has unusual partitioning properties due to its low hydrophobicity and relatively high vapor pressure, so hexafluorobenzene, unlike its chlorinated analog, does not appreciably partition to soils or bioaccumulate. EQC Level II modeling illustrates the increasing persistence of the chlorobenzenes with increasing levels of chlorination. Increasing half-lives and increasing tendency to partition to soil combine to increase the overall residence time markedly as the degree of chlorination increases. Although the half-lives of the fluorobenzenes similarly increase with increasing degree of fluorination, the overall residence time (or persistence) does not markedly increase because increasing fluorination does not increase the amount partitioning to the soil and sediment compartments in which there is slower chemical degradation. Losses of the fluorobenzenes are mainly controlled by advection in air flowing out of the model world. Although hexafluorobenzene has a longer reaction half-life in air, its overall residence time in the evaluative environment is not much longer than benzene.

78

D.A. Ellis et al.

Table 6. EQC Level III results: chemical amounts in each medium based on single and multiple emissions Compound

Benzene

Monofluorobenzene

Emission medium

Amount at steady state (kg) (percent in brackets) Air

Water

Soil

Sediment

Overall residence time (days)

Air Water Soil “Total”

75,100 (99) 46,500 (15) 73,500 (49) 195,000 (37)

419 (0.55) 259,000 (84) 3,216 (2.2) 263,000 (49)

99.9 (0.13) 61.9 (0.02) 72,600 (49) 72,700 (14)

1.96 (0.003) 1212 (0.40) 15.0 (0.01) 1230 (0.23)

3.2 13 6.2 7.4

Air Water Soil “Total”

84,200 (99) 52,400 (17) 79,700 (27) 216,000 (31)

358 (0.42) 258,000 (82) 2,390 (0.80) 260,000 (37)

336 (0.40) 209 (0.07) 219,000 (73) 219,000 (31)

5.95 (0.007) 4,280 (1.4) 39.8 (0.01) 4,330 (0.62)

3.5 13 13 9.7

78,200 (99) 53,000 (16) 76,300 (21) 207,000 (26)

252 (0.32) 278,000 (82) 1,840 (0.51) 280,000 (36)

406 (0.52) 275 (0.08) 286,000 (79) 287,000 (37)

8.96 (0.01) 9,870 (2.9) 65.5 (0.02) 9,940 (1.3)

3.3 14 15 11

95,600 (100) 65,108 (19) 95,500 (93) 256,000 (47.6)

105 (0.11) 275,000 (81) 651 (0.64) 275,000 (51)

10.4 (0.01) 7.11 (0.002) 6,080 (5.9) 6,090 (1.1)

0.25 (0.0003) 661 (0.19) 1.57 (0.002) 662 (0.12)

4.0 14 4.3 7.5

82,600 (99) 39,500 (16) 62,100 (17) 184,000 (26)

506 (0.61) 204,000 (83) 2,800 (0.75) 207,000 (29)

475 (0.57) 227 (0.09) 310,000 (83) 311,000 (44)

6.3 (0.008) 2,540 (1.0) 34.8 (0.009) 2,580 (0.37)

3.5 10 16 9.8

86,900 (91) 55,300 (12) 34,600 (0.79) 177,000 (3.6)

1,010 (1.0) 280,000 (59) 2,980 (0.07) 284,000 (5.7)

7,140 (7.5) 4,550 (0.96) 4,360,000 (99) 4,370,000 (88)

482 (0.51) 134,000 (28) 1,426 (0.03) 136,000 (2.7)

4.0 20 180 69

92,600 (57) 47,000 (1.1) 2,030 (0.02) 142,000 (0.85)

2,200 (1.4) 294,000 (6.7) 691 (0.006) 297,000 (1.8)

37,200 (23) 18,900 (0.43) 12,200,000(100) 12,200,000 (73)

30,300 (19) 4,050,000 (92) 9,520 (0.08) 4,090,000 (24)

6.8 180 510 233

1,2,4-Trifluorobenzene Air Water Soil “Total” Hexafluorobenzene Air Water Soil “Total” Monochlorobenzene Air Water Soil “Total” 1,2,4-Trichlorobenzene Air Water Soil “Total” Hexachlorobenzene Air Water Soil “Total”

EQC Level III calculations allow non-equilibrium conditions to exist between connected media at steady state. They are useful in determining how media of release affects environmental fate and can identify important transformation and interphase transport processes. Table 6 shows the amount of chemical present in each medium of the EQC model environment, and the chemical persistence at steady state for individual emissions to air, water and soil, as well as a “Total” for simultaneous emissions to each compartment. The “Total” mass balance is the sum of the three individual, single compartment emission mass balances because the system of equations is linear.

Partitioning of Organofluorine Compounds in the Environment

79

The Level III EQC results for monochlorobenzene and 1,2,4-trichlorobenzene indicate that emissions tend to remain in the media of release, and are removed from the system by advection or chemical degradation before substantial partitioning to other media takes place. Hexachlorobenzene, which has a lower vapor pressure and higher hydrophobicity, is shown to partition rapidly out of air and water. Emissions to air partition significantly to soil, whereas water emissions partition to sediment. If emitted to soil, the chlorobenzenes are predicted to remain in that medium with only the relatively volatile benzene and monochlorobenzene partitioning significantly to air. The medium of release is also shown to affect greatly the environmental persistence of the chlorobenzenes. Direct emissions to soil lead to the longest environmental residence times because of lack of advection and slow degradation within this medium. Air emissions on the other hand are relatively rapidly removed from the evaluative environment by a combination of advection and more rapid degradation. It should be noted that rapid removal by advection is advantageous from the viewpoint of the receiving environment, but it merely transfers the contaminant to another “downwind” region by long range transport. The substance is thus not persistent locally, but can be persistent globally. This is the case with hexachlorobenzene and probably with hexafluorobenzene. The Level III EQC results for the fluorobenzenes are similar to those for benzene and monochlorobenzene in that a large fraction of emissions tends to remain in the medium of release, although emissions to soil tend to repartition rapidly to air. For example, hexafluorobenzene emissions to soil do not remain

Fig. 12. Diagram showing EQC Level III output for hexachlorobenzene

80

D.A. Ellis et al.

Fig. 13. Diagram showing EQC Level III output for hexafluorobenzene

in the soil because of the compound’s relatively high vapor pressure and very low hydrophobicity. Thus, unlike its chlorinated analog, hexafluorobenzene has a short local environmental residence time irrespective of the medium of discharge. Figures 12 and 13 compare Level III EQC output for hexachloro- and hexafluorobenzene. In summary, the Level I, II and III EQC results indicate that the environmental partitioning and fate of the fluorobenzenes are similar to those of benzene. Substituting fluorines does not cause an increase in hydrophobicity and a reduction in vapor pressure to the same extent as does chlorine substitution. Thus, in terms of environmental partitioning, fluorine substitution on the benzene ring is approximately equivalent to hydrogen substitution. Stability to degrading reactions does, however, increase with addition of fluorines because of the relative strength of the carbon-fluorine bond. The perfluorinated benzene appears to be a special case in that its hydrophobicity is very low, which increases its tendency to partition to air. A similar analysis could be undertaken for the halogenated alkanes and alkenes, resulting in generally similar conclusions that fluorine substitution does not cause a marked increase in hydrophobicity or a marked decrease in volatility. For organic acids such as the carboxylic acids, the substitution of fluorine tends to reduce pKa and increase the degree of ionization, and thus these substances have a greater affinity for the aqueous phase. They are also more stable than their chlorinated analogs as is exemplified by the extreme persistence of trifluoroacetic acid [15].

Partitioning of Organofluorine Compounds in the Environment

81

5 Conclusions This brief review of the partitioning properties of organofluorine compounds between the media of air, water and octanol as surrogates for environmental media has shown that fluorine as a substituent for hydrogen confers quite different properties from that of the more widely studied chlorine and bromine. The low atomic mass and a lower volume of fluorine do not cause the consistent increase in hydrophobicity and decrease in volatility induced by chlorine. The substituent effects of fluorine are smaller and generally less predictable, at least using existing structure-properties approaches. In particular, polyfluorinated compounds exhibit unique behavior probably because of significant intramolecular interactions (Sect. 3.1) and minimal intermolecular interactions. The strong electronegativity of fluorine causes appreciable dissociation of carboxylic acids and the strong C–F bond imparts considerable stability to degrading reactions. Evaluative mass balance modeling shows that the partitioning behavior of fluorinated benzenes is remarkably similar to that of benzene by exhibiting appreciable volatility and minimal hydrophobicity. Persistence is, however, greatly increased. Clearly there is a need to improve our knowledge about the partitioning and reactivity properties of this important class of organic compounds in order that their environmental fate and effects can be more fully and accurately evaluated.

6 References 1. Aten AHW Jr (1976) The boiling points of perfluoroparaffins. J Fluorine Chem 8:93–94 2. Banks RE, Smart BE, Tatlow JC (eds) (1994) Organofluorine chemistry principles and commercial applications. Plenum Press, New York 3. Bechalany A, Roethlisberger T, El Tayar N, Testa B (1989) Comparison of various non-polar stationary phases used for assessing lipophilicity. J Chromatogr 473:115–124 4. Benning AF, Park JD (1946) Fluorinated organic compounds. US Pat 2,490,764 5. Berg M, Müller SR, Mühlemann J, Wiedmer A, Schwarzenbach RP (2000) Concentrations and mass fluxes of chloroacetic acids and trifluoroacetic acid in rain and natural waters in Switzerland. Environ Sci Technol 34:2675–2683 6. Booth HS, Swinehart CF (1935) The critical constants and vapor pressures at high pressure of some gaseous fluorides of Group IV. J Am Chem Soc 57:1337–1342 7. Brekke T, Aksnes DW, Sletten E, Stocker M (1988) Solvent shifts and component identification of hydrocarbon mixtures by carbon-13 nuclear magnetic resonance spectrometry. Anal Chem 60:591–596 8. Brown JA (1963) Physical properties of perfluoropropane. J Chem Eng Data 8:106–108 9. Cahill TM, Seiber JN (2000) Regional distribution of trifluoroacetate in surface waters downwind of urban areas in northern California, U.S.A. Environ Sci Technol 34:2909–2912 10. Cahill TM, Thomas CM, Schwarzbach SE, Seiber JN (2001) Accumulation of trifluoroacetate in seasonal wetlands in California. Environ Sci Technol 35:820–825 11. Cotton AF, Wilkinson G (1998) Advanced inorganic chemistry, 5th edn. Wiley, New York, USA 12. Daniewski AR, Dabrowska H, Piasek Z, Urbanski T (1962) Infrared absorption spectra of some urea inclusion compounds. J Chem Soc 2340–2343

82

D.A. Ellis et al.

13. Ellis DA, Mabury SA (2000) The aqueous photolysis of TFM and related trifluoromethylphenols. An alternate source of trifluoroacetic acid in the environment. Environ Sci Technol 34:623–637 14. Ellis DA, Martin JW, Muir DCG, Mabury SA (2000) Development of an F-19 NMR method for the analysis of fluorinated acids in environmental water samples. Anal Chem 72:726– 731 15. Ellis DA, Hanson ML, Sibley PK, Shahid T, Fineberg NA, Solomon KR, Muir DCG, Mabury SA (2001) The fate and persistence of trifluoroacetic and chloroacetic acids in pond waters. Chemosphere 42:309–318 16. Emptage M, Tabinowski J, Odom JM (1997) Effect of fluoroacetates on methanogenesis in samples from selected methanogenic environments. Environ Sci Technol 31:732–734 17. Gago FA,Alvarez-Builla J, Elguero J (1987) Hydrophobicity measurements by HPLC: a new approach to p constants. J Liq Chromatogr 10:1031–47 18. Galska-Krajewska A (1961) Quaternary positive-negative system of acetic acid-pyridinenonane-p-xylene. Bull Acad Pol Sci Ser Sci Chim 9:455–459 19. Garst JE (1984) Accurate, wide-range, automated, high-performance liquid chromatographic method for the estimation of octanol/water partition coefficients. II. Equilibrium in partition coefficient measurements, additivity of substituent constants, and correlation of biological data. J Pharm Sci 73:1623–1629 20. Grosse AV, Wackher RC, Linn CB (1940) Physical properties of the alkyl fluorides and a comparison of the alkyl fluorides with the other alkyl halides and with the alkyls of the elements of period II. J Phys Chem 44:275–296 21. Haberditzl W, Koeppel H (1967) Faraday spectroscopy. I. Magnetooptic rotatory dispersion of organic liquids. Inorg Chem Org Chem Biochem Biophys Biol 22:691–698 22. Haszeldine RN, Smith F (1951) Organic fluorides. VI. The chemical and physical properties of certain fluorocarbons. J Chem Soc 603–608 23. Inel Y, Iseri R (1997) The octanol-water partition coefficient of benzene derivatives based on three dimensional structure directed molecular properties. Chemosphere 35:993–1002 24. Jones LB, Foster JP (1967) The nucleophilicity of chloride ion toward carbonyl carbon. J Org Chem 32:2900–2901 25. Jordan A, Frank H (1999) Trifluoroacetate in the environment. evidence for sources other than HFC/HCFCs. Environ Sci Technol 33:522–527 26. Key BD, Howell RD, Criddle CS (1997) Fluorinated organics in the biosphere. Environ Sci Technol 31:2445–2454 27. Key BD, Howell RD, Criddle CS (1998) Defluorination of organofluorine sulfur compounds by Pseudomonas sp. strain D2. Environ Sci Technol 32:2283–2287 28. Kim BR, Suidan MT, Wallington TJ, Du X (2000) Biodegradability of trifluoroacetic acid. Environ Engin Sci 17:337–342 29. Lide DR (ed) (2000) CRC handbook of chemistry and physics, 81st edn. CRC Press, Boca Raton, FL, US 30. Mackay D, Di Guardo A, Paterson S, Cowan C (1996) Evaluating the environmental fate of a variety of types of chemicals using the EQC model. Environ Toxicol Chem 15:1627–1637 31. Mackay D, Shiu WY, Ma KC (2000) Physical-chemical properties and environmental fate and degradation handbook. CRCnetBASE 2000, Chapman and Hall CRCnetBASE, CRC Press LLC., Boca Raton, FL. (CD-ROM) 32. Mackay D (2001) Multimedia environmental models: the fugacity approach, 2nd edn. Lewis Publishers, Boca Raton, Florida 33. Martin JW, Franklin J, Hanson ML, Solomon KR, Mabury SA, Ellis DA, Scott BF, Muir DCG (2000) Detection of chlorodifluoroacetic acid in precipitation: a possible product of fluorocarbon degradation. Environ Sci Technol 34:274–281 34. Meylan W. (1999) EPIWIN v.3.04 [computer program: USA EPA Version for Windows], Syracuse Research Corporation, Syracuse, NY, US 35. Michels A, Wassenaar T (1948) Vapor pressure of methyl fluoride. Physica, 14:104–110 36. Miller WT Jr, Bergman E, Fainberg AH (1957) Perfluoroalkylzinc compounds. I. The preparation and properties of perfluoroalkylzinc halides. J Am Chem Soc 79:4159–4164

Partitioning of Organofluorine Compounds in the Environment

83

37. Mohler FL, Bloom EG, Lengel JH,Wise CE (1949) Mass spectra of some cyclic and straightchain fluorocarbons. J Am Chem Soc 71:337–339 38. Moldavsky DD, Furin GG (1998) The purification of perfluorinated compounds for commercial use. J Fluorine Chem 87:111–121 39. Molina MJ, Rowland RS (1974) Stratospheric sink for chlorofluoromethanes: chlorine atom catalyzed destruction of ozone. Nature 249:810–812 40. Nose A, Kudo T (1990) Studies of reduction of various organic compounds with the nickel(II) chloride-zinc system. Chem Pharm Bull 38:2097–2101 41. Parish WR, Sitton DM (1982) Vapor-liquid equilibrium data for the propane, n-butane, isobutane, and propylene/isopropyl fluoride systems at 30 and 50 °C. J Chem Eng Data 27:303–306 42. Paul S, Chaudhury RT, Bhattacharyya B, Majumdar DK (1972) Suitability of Henderson’s partition function in the computation of heat capacity and compressibility of liquids. Indian J Chem 10:505–508 43. Pedler AE, Smith RC, Tatlow JC (1972) Synthesis and dehydrofluorination of some polyfluoroalkanes. J Fluorine Chem 1:337–345 44. Scott RL (1948) The solubility of fluorocarbons. J Am Chem Soc 70:4090–4093 45. Scott BF, Mactavish D, Spencer C, Strachan WM, Muir DCG (2000) Haloacetic acids in Canadian lake waters and precipitation. Environ Sci Technol 34:4266–4272 46. Serjeant EP, Dempsey B (1979) Ionisation constants of organic acids in aqueous solution. IUPAC Chemical Data Series, No 23. Pergamon Press, Oxford, England 47. Streng AG (1971) Miscibility and compatibility of some liquefied and solidified gases at low temperatures. J Chem Eng Data 16:357–359 48. Takahashi J, Kawabata Y, Yamada K (1965) The utilization of hydrocarbons by microorganisms.V. Screening of yeast for cell production from hydrocarbons and their RNA contents. Agr Biol Chem (Tokyo) 29:292–299 49. Visscher PT, Culbertson CW, Oremland RS (1994) Degradation of trifluoroacetate in oxic and anoxic sediments. Nature 369:729–731 50. Wadsworth WS Jr, Emmons WD (1962) Bicyclic phosphites. J Am Chem Soc 84:610–617 51. Wallington TJ, Schneider WF,Worsnop DR, Nielsen OJ, Sehested J, Debruyn WJ, Shorter JA (1994) The environmental impacts of CFC replacements – HFCs and HCFCs. Environ Sci Technol 28:320A-325A 52. Yalkowsky SH,Valvani SC (1980) Solubility and partitioning. I. Solubility of nonelectrolytes in water. J Pharm Sci 69:912–922

CHAPTER 3

Atmospheric Chemistry and Environmental Impact of Hydrofluorocarbons (HFCs) and Hydrofluoroethers (HFEs) Timothy J. Wallington 1 · Ole J. Nielsen 2 1 2

Ford Research Laboratory, MD 3083/SRL, Dearborn, Michigan 4812, USA E-mail: [email protected] University of Copenhagen, Department of Chemistry, Universitetsparken 5, 2100 København Ø, Denmark E-mail: [email protected]

We review the available data concerning the atmospheric chemistry and environmental impact of hydrofluorocarbons (HFCs) and hydrofluoroethers (HFEs). HFCs and HFEs have no impact on stratospheric ozone. The direct global warming potentials of HFCs and HFEs are approximately an order of magnitude less than the CFCs they replace. At the concentrations expected from their atmospheric degradation, none of the oxidation products of HFCs or HFEs, are noxious or toxic. Keywords. Atmospheric chemistry, Hydrofluorocarbons, Hydrofluoroethers, Environmental impact, Global warming, Stratospheric ozone

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

1

Introduction

2

Atmospheric Photochemistry . . . . . . . . . . . . . . . . . . . . . 87

2.1 2.2

Troposphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Stratosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

3

Atmospheric Lifetimes . . . . . . . . . . . . . . . . . . . . . . . . . 90

4

Gas-Phase Chemistry

5

Aqueous-Phase Chemistry

6

Contribution of HFCs and HFEs to Global Climate Change . . . . 97

6.1 6.2

Radiative Forcing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 Global Warming Potential . . . . . . . . . . . . . . . . . . . . . . . 99

7

Formation of Toxic/Noxious Degradation Products

8

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

9

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

. . . . . . . . . . . . . . . . . . . . . . . . . 91 . . . . . . . . . . . . . . . . . . . . . . 96

. . . . . . . . 100

The Handbook of Environmental Chemistry Vol. 3, Part N Organofluorines (ed. by A.H. Neilson) © Springer-Verlag Berlin Heidelberg 2002

86

T.J. Wallington · O.J. Nielsen

1 Introduction Recognition of the adverse impact of chlorofluorocarbons (CFCs) on stratospheric ozone [9] has prompted an international effort to replace CFCs with environmentally acceptable alternatives. Hydrofluorocarbons (HFCs) and hydrofluoroethers (HFEs) are two classes of CFC replacements. HFCs are used as refrigerant agents, foam blowing agents, fire extinguishing agents and aerosol propellants. From a commercial point of view, the most important HFC is CF3CFH2 (HFC-134a) which has found widespread use as a replacement for CFC-12 (CF2Cl2) in domestic refrigeration and automobile air-conditioning units. The annual production of CF3CFH2 has increased rapidly from essentially zero in 1990 to 112,000 tons in 1998 and is projected to increase still further as demand increases. HFEs are used in much smaller quantities than HFCs and find applications as cleaning agents for electronic equipment, heat transfer agents in refrigeration systems, and carrier fluids for lubricant deposition. HFCs and HFEs are volatile and insoluble in water. Following release into the environment these compounds reside in the atmosphere. Figure 1 shows the concentrations of HFC-134a and HFC-23 in the atmosphere observed at a remote monitoring site at Cape Crim, Tasmania [15]. Substantial levels of HFC-152a (CH2FCH3) are also present in the atmosphere. There are no natural sources of HFCs and HFEs, these compounds are entirely man-made in origin and are emitted into the troposphere as the result of industrial activities. In light of the atmospheric buildup of HFCs evident in Fig. 1 it is obviously important to study the atmospheric chemistry and quantify the environmental impact of such compounds. When HFCs were first proposed as CFC replacements in the 1980s little was known about their atmospheric chemistry. Initially there was speculation that fluorine-containing free radical species generated during the at-

Fig. 1. HFC-134a and HFC-23 concentrations in ambient air observed at Cape Grim, Tasmania

[15]

Atmospheric Chemistry and Environmental Impact of Hydrofluorocarbons and Hydrofluoroethers

87

mospheric degradation of HFCs and HFEs (CF3Ox , FCOx , and FOx ) might adversely impact stratospheric ozone [6] and that the atmospheric degradation of HFCs might form toxic products which could accumulate in the environment [24]. To assess the environmental acceptability of HFCs the chemical industry sponsored a number of experimental and theoretical studies of their atmospheric chemistry. As a result we now have a good understanding of the atmospheric degradation mechanisms of such compounds. In fact, it can be argued that of all the different classes of chemical compounds emitted into the atmosphere we understand the chemistry of HFCs and HFEs the best. It is now clear that initial concerns regarding the environmental impact of such compounds were unfounded. The recent work establishing the atmospheric degradation mechanism of HFCs provides a notable example of the successful application of fundamental research to solve a practical environmental problem. In this chapter we will examine the atmospheric degradation mechanisms of HFCs and HFEs. Our intent is not to give an exhaustive account of the photochemical oxidation of every HFC and HFE but rather to present a representative example of the degradation mechanism of one member of each class (HFC134a, CF3CFH2 and HFE-7100, C4F3OCH3) and make some general comments on the atmospheric chemistry of HFCs and HFEs. First, we need to consider the general features of atmospheric photochemistry.

2 Atmospheric Photochemistry The atmosphere is a giant inhomogeneous photochemical reactor in which temperature, pressure, radiation flux, and composition vary widely. All of the energy for atmospheric chemistry comes from solar radiation which is absorbed by various components of the atmosphere. Figure 2 shows the direct solar flux over the wavelength range 120–360 nm at the top of the atmosphere (labeled solar), and at 40, 20, and 0 km altitude. Molecular oxygen is responsible for the absorption of UV radiation of wavelengths less than 200 nm at altitudes above 40 km. Absorption by ozone in the stratosphere shields the earth’s surface from UV radiation of wavelengths less than 300 nm. Gas-phase and solution-phase reactions play important roles in the atmospheric degradation mechanism of HFCs and HFEs. In discussions of atmospheric processes it is useful to divide the atmosphere into different regions. Temperature profiles provide the most convenient basis for this division. Figure 3 shows the altitude profile of atmospheric temperature and pressure up to 50 km [25].

88

T.J. Wallington · O.J. Nielsen

Fig. 2. Direct solar flux over the wavelength range 120–360 nm at the top of the atmosphere

(labeled solar) and at 40, 20, and 0 km

Fig. 3. Temperature and pressure profiles of the atmosphere up to 50 km taken from the US Standard Atmosphere

Atmospheric Chemistry and Environmental Impact of Hydrofluorocarbons and Hydrofluoroethers

89

2.1 Troposphere

The first 10–15 km of the atmosphere is characterized by a temperature profile in which colder air overlays warmer air. This situation is caused by the fact that the predominant heat source for this region of the atmosphere is the warm surface of the earth. The temperature profile of the troposphere is inherently unstable and results in strong vertical convective mixing from which the region derives its name (tropos is derived from Greek, turning). Intense tropical thunderstorm systems can transport molecules from close to the earth’s surface to the top of the troposphere within a few minutes. However, more typical mixing times are of the order of days to weeks. More than 90% of the mass of the atmosphere is located in the troposphere and it is here that the vast majority of HFCs and HFEs are degraded. 2.2 Stratosphere

In contrast to the troposphere, the stratosphere which lies at approximately 15–50 km altitude is heated principally from above by the absorption of solar UV radiation. In the stratosphere, warm air lies on top of cooler air which is an inherently stable situation resulting in a layered structure which gives the region its name (strato is derived from Latin, stratum, layer). Vertical mixing in the stratosphere proceeds slowly, typically on a time scale of several years. The region which marks the boundary between the two different regions is called the tropopause. The driving force for most of the chemistry that occurs in the atmosphere is the formation of hydroxyl (OH) radicals via photolysis of ozone to form O(1D) atoms which react with water vapor, Eq. (1). O3 +hn(l<320 nm) Æ O(1D)+O2 (1Dg)

(1)

O(1D)+H2O Æ 2 OH

(2)

UV light, O3 , and H2O vapor combine to produce a source of OH radicals. OH radicals react with HFCs and HFEs. As with many other compounds, the atmospheric lifetimes of HFCs and HFEs are determined by their reactivity towards OH radicals. While the OH radical concentration in the atmosphere varies with location, time of day, season, and meteorological conditions, a reasonable global 24 hour global average is 1.0¥106 cm–3 [17]. The choice of HFCs and HFEs as replacements for CFCs is motivated by a number of factors, not least of which is that in contrast to CFCs, HFCs and HFEs contain one or more C–H bonds. Hence, unlike CFCs, HFCs and HFEs are susceptible to attack by OH radicals in the lower atmosphere (troposphere). HFCs and HFEs do not contain chlorine and so have no ozone depletion potential associated with the well established chlorine-based catalytic cycles [9]. To define the environmental impact of HFCs and HFEs, their ability to destroy stratospheric ozone, contribute to potential global warming, and pro-

90

T.J. Wallington · O.J. Nielsen

duce noxious degradation products must be assessed. To address these issues a detailed knowledge of their atmospheric chemistry is required. Atmospheric chemistry includes reactions in gaseous and aqueous phases together with any relevant heterogeneous processes. We present here an overview of the atmospheric chemistry and environmental impact of HFCs and HFEs.

3 Atmospheric Lifetimes The concept of atmospheric lifetime is useful in discussions of the environmental impact of HFCs and HFEs [17]. It can be defined in several ways. Most simply put, it can be expressed as the turnover time which is the atmospheric burden of a species divided by its rate of emission, assuming a constant emission rate and steady state condition. Alternatively, it can be given as the reciprocal of the pseudo-first order rate constant (k¢) for its removal: 1 Atmospheric Lifetime (t)= 3 k¢ HFCs and HFEs contain one or more C–H bonds and so are susceptible to attack by OH radicals in the troposphere. Reaction with OH radicals is the dominant loss process for all saturated HFCs and HFEs. In the stratosphere photolysis and reaction with Cl and O(1D) atoms make minor contributions to the overall loss. The starting point for lifetime estimations for HFCs and HFEs are laboratory generated kinetic data for their reaction with OH radicals. The bimolecular rate constants measured in laboratory kinetic experiments need to be converted into a pseudo-first order rate constant for loss of the compound, k¢. In principal this conversion is simple, i.e., the bimolecular rate constant merely has to be multiplied by the OH concentration [OH]. In practice there are difficulties associated with the choice of an appropriate value of [OH]. At present we cannot measure the global OH concentration field directly. The OH radical concentration varies widely with location, season, and meteorological conditions. To account for such variations requires use of sophisticated 3D computer models of the atmosphere. In such models the OH concentration field is computed using measured or estimated concentration fields of the precursor molecules and photon flux data (the term “field” refers to distribution as a function of latitude, longitude, and altitude). The resulting OH field is then tuned such that it correctly predicts the lifetime of methylchloroform (CH3CCl3) with respect to OH radical attack. From measurements of the atmospheric turnover time of CH3CCl3 (4.6 years), its lifetime with respect to loss in the stratosphere (43 years), and its lifetime with respect to loss in the oceans (80 years), the tropospheric lifetime of CH3CCl3 with respect to OH radical attack can be inferred to be 5.5 years. Methylchloroform is the calibration molecule of choice because it has a long history of precise atmospheric measurements, it has no natural sources, its industrial production is well documented, and the kinetics of reaction (3) are well established, k3 =1.8¥10–12 exp(–1550/T) cm3 molecule–1 s–1 [5].

 

OH+CH3CCl3 Æ products

(3)

Atmospheric Chemistry and Environmental Impact of Hydrofluorocarbons and Hydrofluoroethers

91

The use of 3D computer models to calculate atmospheric lifetimes is a rather cumbersome approach and access to such models is limited. A simpler technique to estimate the tropospheric lifetime of a given HFC with respect to OH attack is to scale the tropospheric lifetime of CCl3CH3 by the rate constant ratio k(OH+CCl3CH3)/k(OH+HFC). OH+HFC Æ products

(4)

If reactions (3) and (4) have the same temperature dependence then the rate constant ratio can be evaluated at any given temperature. If not, then the temperature chosen for the comparison is important. It has been shown that 272 K is the optimal temperature for comparison. Hence, a reasonable estimate of the atmospheric lifetime of a compound with respect to OH radical attack in the troposphere can be made using:





k(OH+CCl3CH3) t(compond)= 88818 ¥ 5.5 year k(OH+compound) Measurements of rate constants of OH radical reactions with fluorinated organic compounds are available [5] and can be used in estimations of atmospheric lifetimes. Table 1 lists atmospheric lifetimes for selected HFCs and HFEs along with those of CFC-11, CFC-12, and CO2 for comparison. Differences in the atmospheric lifetimes of the compounds in Table 1 can be ascribed to the number and strength of the C–H bonds in the molecules. C2H5F has 5 relatively weak C–H bonds and so has a lifetime of just a few months, in contrast CF3H with just one strong C–H bond has a lifetime of 243 years.

4 Gas-Phase Chemistry HFCs and HFEs have similar molecular structures and atmospheric oxidation mechanisms. To understand the general atmospheric oxidation mechanism of HFCs and HFEs it is useful to concentrate on the detailed mechanism of a representative compound for each class. We will consider the atmospheric oxidation of HFC-134a and HFE-7100 in detail. As mentioned above, HFC-134a is the most widely used HFC, it has attracted detailed study by a large number of research groups and its atmospheric oxidation mechanism is now well established. The atmospheric oxidation mechanism of HFC-134a is given in Fig. 4. Atmospheric oxidation is initiated by reaction with OH radicals which takes place with a rate constant k=1.5¥10–12 exp(–1750/T) cm–3 molecule–1 s–1 [5]. The lifetime of HFC-134a with respect to reaction with OH is 13.6 years. The fluorinated alkyl radical CF3CFH adds O2 rapidly (within 1 µs) to give the corresponding peroxy radical CF3CFHO2 . As with other peroxy radicals the fate of the CF3CFHO2 radical is reaction with either NO, NO2 , or HO2 radicals. Peroxy radicals react rapidly with NO2 to give alkyl peroxynitrates (RO2NO2). Using the laboratory kinetic data for the reaction of CF3CFH2O2 radicals with NO2 the lifetime of CF3CFH2O2 radicals with respect to reaction with NO2 can be estimated to be 10 minutes. Alkyl peroxynitrates are thermally unstable and decompose rapidly to regenerate RO2 radicals and NO2 . At room temperature in

92

T.J. Wallington · O.J. Nielsen

Fig. 4. Atmospheric degradation mechanism of HFC-134a.Values in parentheses are order-of-

magnitude lifetime estimates, species in boxes are observed products

one atmosphere of air the peroxynitrate derived from HFC-134a has a lifetime of <90 seconds. The lifetime of CF3CFHO2 radicals with respect to reaction with HO2 has been estimated to be approximately 5 minutes [20]. The reaction of peroxy radicals with HO2 radicals gives hydroperoxides and, in some cases, carbonyl products. Product data are available for two HFC-derived peroxy radicals: CH2FO2 and CF3CFHO2 . Reaction of CH2FO2 radicals with HO2 gives a 30% yield of the hydroperoxide, CH2FOOH, and 70% yield of the carbonyl product, HCOF. In the reaction of CF3CFHO2 with HO2 radicals less than 5% of the products appear as the carbonyl CF3COF. The factors which determine the relative importance of the hydroperoxide and carbonyl forming channels are unclear at present. More work is needed in this area. The hydroperoxide CF3CFHOOH is expected to be converted back into CF3CFHO2 radicals via photolysis and reaction with OH. Alkylperoxy radicals react rapidly with NO to give alkoxy radicals and NO2 as major products and alkyl nitrates as minor products. The experimental data suggest that nitrate formation from reactions of fluorinated alkylperoxy radicals with NO is of little, or no, significance. The reaction of CF3CFHO2 radicals gives CF3CFHO radicals and NO2 . It has been shown that this reaction produces

Atmospheric Chemistry and Environmental Impact of Hydrofluorocarbons and Hydrofluoroethers

93

a substantial fraction (approximately 60%) of chemically activated CF3CFHO radicals (denoted as CF3CFHO* in Fig. 4) [27]. The excited alkoxy radicals possess sufficient internal energy to overcome the approximately 8 kcal mol–1 barrier to C-C bond scission and decompose on a timescale of the order of 10–10 seconds [19] to give CF3 radicals and HCOF. The CF3CFHO radicals which become thermalized undergo both reaction with O2 and thermal decomposition via C–C bond scission. CF3CFHO+O2 Æ CF3COF +HO2

(5)

CF3CFHO+M Æ CF3 +HCOF+M

(6)

In one atmosphere (one atmosphere=101 mbar=760 Torr) of air dilutent the competition between reactions (5) and (6) is described by the rate constant ratio k5/k6 =(2.4+1.6 –1.0)¥10–25 exp(3590±150/T) cm3 molecule–1. The relative importance of the two channels is given by k5 [O2]/k6 . In one atmosphere of air at 288 K (temperature of US Standard Atmosphere at sea level, see Fig. 3) k5 [O2]/k6 =6.22¥10–20 ¥5.34¥1018 =0.33, hence 25% of CF3CFHO radicals undergo reaction with O2 while the remaining 75% decompose via reaction (6). With increasing altitude both temperature and pressure fall substantially (see Fig. 3). Decreasing temperature slows down the rates of both reactions (5) and (6). Reaction (6) is a unimolecular decomposition process and has an activation energy which is substantially larger than that of the bimolecular reaction (5). Hence reaction (6) slows down much more rapidly with increasing altitude (even after accounting for decreased [O2] at higher altitudes). Ignoring for the moment the pressure dependence of reaction (6) we can calculate that at an altitude of 10 km (T=223 K, total pressure=265 mbar) that k5 [O2]/k6 =2.35¥10–18 ¥1.81¥1018 =4.25, hence 81% of CF3CFHO radicals undergo reaction with O2 while the remaining 19% decompose via reaction (6). Accounting for the distribution of OH radicals in the troposphere, the temperature and pressure profiles shown in Fig. 3, and the chemical activation effect it has been estimated that 7–20% of the atmospheric degradation of HFC-134a gives CF3COF [27]. CF3 radicals formed in reaction (6) add O2 to give CF3O2 radicals which react with NO to give CF3O radicals. The usual modes of alkoxy radical loss are not possible for the CF3O radical. Reaction with O2 and decomposition via F atom elimination are both thermodynamically impossible under atmospheric conditions. Instead, CF3O radicals react with NO and hydrocarbons. CF3O+NO Æ COF2 +FNO

(7)

CF3O+CH4 Æ CF3OH+CH3O

(8)

Reaction with NO yields COF2 ◊ COF2 does not react with any gas phase trace atmospheric species and its photolysis is slow. COF2 is removed from the atmosphere by incorporation into water droplets and hydrolysis to give CO2 and HF and by photolysis in the upper stratosphere to give FCO radicals and F atoms. FNO photolyzes to give NO and an F atom. F atoms reversibly form FO2 radicals by combining with O2 , and also react with CH4 and H2O to give HF which will be rained out of the atmosphere. The reaction of CF3O radicals with hydrocar-

94

T.J. Wallington · O.J. Nielsen

bons such as CH4 produces CF3OH. The CF3O–H bond is unusually strong at 120 kcal mol–1 (1 kcal=4.18 kJ). CF3OH is not attacked by any trace atmospheric radical and is not photolyzed. CF3OH undergoes heterogeneous decomposition to give COF2 and HF and reaction with atmospheric water droplets to give CO2 and HF. Several years ago there was speculation that CF3O radicals could participate in catalytic ozone destruction cycles [1]. Experimental studies have shown that this is not the case [18]. As shown in Fig. 4 the gas-phase atmospheric degradation of HFC-134a gives four products: CF3COF, HCOF, COF2 , and CF3OH. Experimental studies of the atmospheric oxidation mechanisms of commercially important HFCs have been performed and their gas-phase atmospheric degradation products are, in general, well understood [26]. Table 1 lists the observed gas phase oxidation products of important HFCs [26]. We will now turn our attention to HFEs. The atmospheric oxidation mechanism for a typical HFE (HFE-7100, C4F9OCH3) is given in Fig. 5 and is very similar to that of HFC-134a. Atmospheric degradation is initiated by reaction with OH radicals which takes place with a rate constant of 1.2¥10–14 cm3 molecule–1 s–1 at 295 K leading to an estimate of 5 years for the atmospheric lifetime of HFE-7100 [28]. The fluorinated radical product, C4F9OCH2 , adds O2 rapidly (within 1 µs) to give the corresponding peroxy radical C4F9OCH2O2 , k(C4F9OCH2 + O2) = (2.5 ± 0.5) ¥ 10–12 cm3 molecule–1 s–1 [28]. The peroxy radical reacts with three trace species in the atmosphere: NO, NO2 , and HO2 radicals. The rate constants for reactions of C4F9OCH2O2 radicals with NO and NO2 have been determined to be (8.5 ± 2.5) ¥ 10–12 and (8.8 ± 1.8) ¥10–12 cm3 molecule–1 s–1, respectively [28]. In polluted urban air masses with [NOx]=1–50 ppb the lifetime of C4F9OCH2O2 radicals with respect to reaction with NOx is 0.1–5.0 seconds. In remote areas with [NOx] = 1–10 ppt C4F9OCH2O2 radicals have a lifetime of 8–80 minutes. As in the case of HFC-

Fig. 5. Atmospheric degradation mechanism of HFE-7100.Values in parentheses are order-ofmagnitude lifetime estimates, species in boxes are observed products

Atmospheric Chemistry and Environmental Impact of Hydrofluorocarbons and Hydrofluoroethers

95

Table 1. Atmospheric lifetimes, global warming potentials, and gas-phase atmospheric degradation products for selected HFCs and HFEs

Compound

Lifetime a (years)

GWP a

Degradation Products

HFC-23 (CF3H) HFC-32 (CH2F2) HFC-41 (CH3F) HFC-125 (CF3CF2H) HFC-134a (CF3CFH2) HFC-143a (CF3CH3) HFC-152a (CF2HCH3) HFC-161 (CH2FCH3) HFC-227ca (CF3CF2CHF2) HFC-227ea (CF3CHFCF3) HFC-236cb (CF3CF2CH2F) HFC-236fa (CF3CH2CF3) HCFC-245fa (CF3CH2CHF2) HFE-125 (CF3OCF2H) HFE-143a (CF3OCH3) HFE-7100 (C4F9OCH3) HFE-7200 (C4F9OC2H5) H-Galden 1040X (CHF2OCF2OC2F4OCHF2) CFC-11 (CFCl3) CFC-12 (CF2Cl2) CO2

243 5.6 3.7 32.6 13.6 53.5 1.5 0.25 32.6 36.5 14.6 226 7.6 165 5.7 5 0.77 48

13000 710 160 3600 1300 5000 140 10 2900 3200 1300 8400 860 14000 620 440 60 8700

COF2 , CF3OH COF2 HCOF COF2 , CF3OH HCOF, CF3OH, COF2 , CF3COF CF3COH, CF3OH, COF2 , CO2 COF2 HCOF, CH3COF COF2 , CF3OH COF2 , CF3OH, CF3COF HCOF, CF3OH, COF2 , C2F5COF CF3COCF3 COF2 , CF3COH, CF3OH COF2 , CF3OH CF3OCOH C4F9OCOH C4F9OCOH, HCHO COF2

45 100

4500 10600 1b

a b

Sihra et al. [21]. By definition.

134a, the reaction of C4F9OCH2O2 with NO2 gives a thermally unstable peroxynitrate, C4F9OCH2O2NO2 , whose sole fate is decomposition to reform C4F9OCH2O2 and NO2 . Reaction of C4F9OCH2O2 with NO gives the alkoxy radical C4F9OCH2O(•). The sole atmospheric fate of C4F9OCH2O(•) radicals is reaction with O2 to give the formate C4F9OCOH. Whereas the reaction of C4F9OCH2O2 radicals with HO2 has not been investigated, it has been shown that the analogous peroxy CF3OCH2O2 reacts with HO2 to give an (80±11)% yield CF3OCOH and a 20% yield of CF3OCH2OOH [4]. By analogy to the behavior of CF3OCH2O2 radicals, it seems reasonable to assume that reaction of C4F9OCH2O2 with HO2 will give the formate, C4F9OCOH, as the major product. Interestingly, irrespective of whether C4F9OCH2O2 radicals react with HO2 radicals or with NO the same product is formed; C4F9OCOH. This formate is relative unreactive to further gas-phase chemistry. The fate of C4F9OCOH is believed to be incorporation into cloudrain-sea water followed by hydrolysis (see following section). Results of studies of the atmospheric chemistry of CF3OCH3 [4, 7, 30], C2F5OCH3 [12, 22], C3F7OCH3 [11, 12, 22], C4F9OCH3 [13, 28], C4H9OC2H5 [3], C5F12OCH3 [12], CHF2OCHF2 [7, 16], CHF2OCF3 [7, 16], CF3CH2OCH3 [13, 14, 30], CF3CH2OCHF2 [14, 30], CHF2CF2OCH3 [13, 14, 23], CF3CHFCF2OCH3 [13,

96

T.J. Wallington · O.J. Nielsen

14], CF3CH2OCH2CF3 [14, 16, 29], and CHF2OCF2OC2F4OCHF2 [2] are available in the literature. The gas-phase atmospheric oxidation products for selected HFEs are given in Table 1. Thus far the oxidation of the halocarbons into halogenated carbonyl products has been discussed. While the gas-phase oxidation mechanisms are complex, the carbonyl products are well established and are also given in Table 1. The carbonyl products represent a convenient break point in our discussion. The sequence of gas-phase reactions that follow from the initial attack of OH radicals on the parent halocarbon are sufficiently rapid that heterogeneous and aqueous processes play no role. In contrast, the lifetimes of the carbonyl products (e.g., HCOF, CF3COF, and C4F9OCOH) are relatively long. As discussed in the following section, incorporation into water droplets followed by hydrolysis plays the dominant role in the removal of halogenated carbonyl compounds.

5 Aqueous-Phase Chemistry The final step in removal of any species from the atmosphere involves heterogeneous deposition to the earth’s surface. Removal processes include wet deposition via rainout (following uptake into tropospheric clouds) and dry deposition to the earth’s surface, principally to the oceans. The rates of these processes are largely determined by the species’ chemistries in aqueous solution. For the HFC/HFE the question is whether the rate of removal of any degradation product is slow compared to the OH reaction limited lifetime of the parent compounds listed in Table 1. Heterogeneous lifetimes of the parent compounds themselves are on the order of hundreds of years because of their low aqueous solubility and reactivity. As discussed in the preceding sections, the species listed in Table 2 are degradation products that have removal rates in the gas phase (via reaction or photolysis) that are slow enough (days or longer) that heterogeneous processing might be significant. All the halogen-containing species are thought to undergo aqueous interactions that are “fast enough” for efficient wet and dry deposition. For example, although CF3COF is relatively insoluble in water, it hydrolyzes to produce HF and CF3C(O)OH (see Table 2). Since HF and CF3C(O)OH are very water soluble, hydrolysis removes the halides from the gas phase irreversibly. Nohara et al. [12] have presented evidence suggesting that the formate nC5F11OCHO is readily hydrolyzed to n-C5F11OH and formic acid. n-C5F11OH Table 2. Aqueous-phase atmospheric degradation products [26]

Compound

COF2 CF3COF HCOF

1/2

H* khyd (M atm–1 s–1/2) 6 4

Lifetime

Degradation products

Clouds (days)

Ocean (years)

5–10 5–15 150–1500

0.3–1.5 0.3–3.0 80

HF, CO2 CF3C(O)OH, HF HF, HC(O)OH

Atmospheric Chemistry and Environmental Impact of Hydrofluorocarbons and Hydrofluoroethers

97

then decomposes via elimination of HF to give CF3(CF2)3COF which undergoes hydrolysis to give the perfluoroacid CF3(CF2)3C(O)OH. It is difficult to measure the relevant aqueous kinetics of such species. Atmospheric removal rates depend on both solubility (expressed in terms of the Henry’s law constant, H, M atm–1) and hydrolysis rate (khyd , s–1). Because these species do not form stable aqueous solutions, neither parameter is simply measurable in bulk solution. As a result, laboratory determinations involve heterogeneous processes that typically measure combinations of H and khyd . Estimation of atmospheric lifetimes requires deconvolution of these parameters. For the halocarbonyls, COF2 , CF3COF, and CF3COCl, in particular, there has been much effort to measure the relevant aqueous kinetics. The results of these experiments are summarized in Table 2 in terms of the product H¥k1/2 hyd , which is the parameter typically measured in gas/liquid mass transfer experiments. For the two related chlorinated species, COCl2 and CCl3COCl, a combination of studies by several groups have determined khyd 100–150 s–1, with about a factor of two uncertainty [26]. Assuming that khyd 100 s–1 is representative of all the halocarbonyl species, one obtains H 0.1–0.6 M atm–1. With H and khyd values, tropospheric lifetimes for heterogeneous uptake into clouds and into the ocean can be estimated, as listed in Table 2 [26]. The range of lifetimes reflects a conservative range of khyd =1–1000 s–1. For tropospheric cloud processing, the lower limit of 5 days is indicative of atmospheric transport limitations, i.e., the time taken to transport the species into the clouds. Two conclusions can be drawn from Table 2. (1) Most importantly, despite the considerable uncertainty in the laboratory kinetic results, heterogeneous removal of halocarbonyl species is fast enough to have no effect on the overall halogen lifetime compared to the lifetime of the parent halocarbon. (2) Tropospheric cloud rainout will predominate over deposition to the ocean. More precise estimates of halocarbonyl lifetimes require further study of the aqueous kinetics. For example the lifetime of HC(O)F listed in Table 2 is based upon a very conservative estimate of khyd =0.01 s–1. Experimental studies of the aqueous kinetics of HC(O)F are required to better define its atmospheric lifetime. Further studies of the aqueous chemistry of fluorinated formates such as C4F9OCOH are needed to complete our understanding of the atmospheric degradation of HFEs.

6 Contribution of HFCs and HFEs to Global Climate Change As shown in Table 1 the atmospheric lifetimes of certain HFCs and HFEs can be quite long. For example CF3H has a lifetime of 243 years while CF3OCF2H has a lifetime of 165 years. Such long lifetimes bring up the question “What will be the contribution of these compounds to global climate change?”. The Earth is warmed by absorption of short wavelength radiation from the Sun, and cooled by the emission of long wavelength radiation into space. The effectiveness of a

98

T.J. Wallington · O.J. Nielsen

Fig. 6. IR spectra of HFC-134a (top) and HFE-7100 (bottom)

Fig. 7. Irradiance at the tropopause

greenhouse gas in the atmosphere in delaying the emission of long wavelength radiation into space and hence warming the Earth is determined by two factors: (1) the strength and position of its IR absorption features and (2) by its atmospheric lifetime. Figure 6 shows the IR absorption spectra of HFC-134a (CF3CFH2) and HFE7100 (C4F9OCH3) over the range 700–1500 cm–1. Figure 7 shows the IR radiation passing through the tropopause (approximately 10 km altitude) on its

Atmospheric Chemistry and Environmental Impact of Hydrofluorocarbons and Hydrofluoroethers

99

way into space. The Earth’s surface acts as a black body and emits long wavelength radiation. The decreases in radiance centered at 667 and 1040 cm–1 in Fig. 7 are caused by absorption attributed to CO2 and ozone in the lower atmosphere. Comparison of Figs. 6 and 7 reveals that HFC-134a and HFE-7100 have strong absorption features at frequencies in the “atmospheric window” region between 700 and 1500 cm–1 where the main atmospheric constituents absorb weakly. To compare the relative impacts of different greenhouse gas control strategies it is useful to place climatic impacts of various greenhouse gases on a common scale. Two concepts are useful in this regard; radiative forcing and global warming potential. 6.1 Radiative Forcing

This is defined as the net change in irradiance at the troposphere resulting from the presence of the greenhouse gas. Radiative forcing is usually calculated using radiative transfer models of the atmosphere and is expressed in units of Wm–2 ppb–1. HFC-134a has a radiative forcing of 0.158 Wm–2 ppb–1 while HFE7100 has a forcing of 0.362 Wm–2 ppb–1 [21]. 6.2 Global Warming Potential

The global warming potential of a compound is a measure of its radiative forcing per mass relative to a given reference compound integrated over a given time horizon. Global warming potentials differ from radiative forcing values in three important respects. (1) Radiative forcings are calculated on a volume basis while global warming potentials are calculated on a mass basis. (2) Global warming potentials are relative, not absolute, measures of the potential impact on climate. Typically, but not always, the reference compound is CO2 , the global warming potential then provides a measure of the likely climatic impact of a given compound relative to the same mass of CO2 . (3) Global warming potentials are integrated over a specified time horizon. All other factors aside, compounds with shorter atmospheric lifetimes will have lower global warming potentials. The relationship between radiative forcing and global warming potential is given below where RFHFC-134a , RFCO2 , MHFC-134a , MCO2 , tHFC-134a , and tCO2 are the radiative forcings, molecular weights, and atmospheric lifetimes of HFC-134a and CO2 and t is the time horizon over which the forcing is integrated.



 30 t M 0 3009 1–exp(–t/t ) 

RFHFC-134a GWPHFC-134a = 3931 RFCO2

tHFC-134a MCO2 CO2

HFC-134a

1–exp(–t/tHFC-134a ) CO2

100

T.J. Wallington · O.J. Nielsen

As a result of the growing industrial significance and increasing atmospheric concentration of HFCs a number of studies have been performed over the past 5 years to establish the contribution of HFCs to climate forcing. Table 1 lists global warming potentials for selected compounds taken from Sihra et al. [21]. As seen in Table 1, the global warming potentials scale approximately linearly with atmospheric lifetime. This is not surprising as CFCs, HFCs, and HFEs all have broadly comparable molecular structures and chemical bonds. Hence, the strengths and positions of their infrared absorptions are similar. The global warming potentials of HFCs and HFEs are less than those of the CFCs they replace but are substantially greater than CO2. For example, the GWP of HFC134a is 8 times less than that of CFC-12 but 1300 times greater than that of CO2 . To place the potential climatic impacts of HFCs and HFEs in perspective, we need to consider their emission rates relative to CO2 . Previous CFC production provides a convenient likely upper limit for HFC and HFE emissions. In 1986 global CFC production was of the order of 106 tonnes [8], the current anthropogenic emission of CO2 is of the order of 1010 tonnes. While HFCs and HFEs are on a per mass basis much more powerful greenhouse gases than CO2 , because of their relatively small emission rates and short lifetimes the climatic impact of HFCs and HFEs is modest (of the order of a % of that of CO2).

7 Formation of Toxic/Noxious Degradation Products The atmospheric degradation of HFCs and HFEs gives rise to a wide variety of products (Tables 1 and 2). The atmospheric concentration of these products will be extremely small (ppb). There are no known adverse environmental impacts associated with these compounds at such low concentrations. The ultimate removal mechanism for all products is incorporation into rain-seacloud water where hydrolysis will take place. With the possible exception of CF3C(O)OH, the hydrolysis products are ubiquitous, naturally-occurring species that have no adverse environmental impact. Trifluoroacetic acid, CF3C(O)OH, has been detected in surface waters (oceans, rivers, and lakes) and in fog, snow, and rainwater samples around the globe and appears to be a ubiquitous component of the present hydrosphere. However, trifluoroacetic acid appears to have been absent from the ancient freshwater hydrosphere [10]. While it is clearly established that there is a substantial environmental burden of CF3C(O)OH, the sources of this compound are unclear. No significant natural sources of trifluoroacetic acid have been identified. Several man-made compounds, e.g., the anaesthetics isoflurane (CF3CHClOCHF2) and halothane (CF3CHClBr), and the CFC replacements HFC-134a (CF3CFH2) and HCFC-123 (CF3CHCl2), are emitted into the environment and produce CF3C(O)OH. However, the magnitude of these industrial sources is several orders of magnitude too small to account for the levels of CF3C(O)OH observed in the world’s oceans. The ongoing research effort to identify the sources and sinks of CF3C(O)OH in the environment is discussed in Chapter 5.

Atmospheric Chemistry and Environmental Impact of Hydrofluorocarbons and Hydrofluoroethers

101

8 Conclusions A substantial body of experimental and theoretical data concerning the atmospheric degradation of HFCs and HFEs is available. In general, we have a good understanding of the atmospheric chemistry of the commercially important HFCs and HFEs. The following summarizes the conclusions of this review. (1) HFCs and HFEs have no impact on stratospheric ozone. (2) The direct global warming potentials of HFCs and HFEs are modest (approximately an order of magnitude less than that of the CFCs they replace). (3) At the concentrations expected from the atmospheric degradation of HFCs and HFEs, none of the oxidation products are noxious or toxic.

9 References 1. Biggs P, Canosa-Mas CE, Shallcross DE, Wayne RP, Kelly C, Sidebottom HW (1993) The possible atmospheric importance of the reaction of CF3O radicals with ozone. Proceedings of the STEP-HALOCSIDE/AFEAS Workshop (pp 177–181), University College Dublin, Ireland, March 2. Cavalli F, Glasius M, Hjorth J, Rindone B, Jensen NR (1998) Atmospheric lifetimes, infrared spectra and degradation products of a series of hydro-fluoroethers. Atmos Environ 32:3767–3774 3. Christensen LK, Sehested J, Nielsen OJ, Bilde M, Wallington TJ, Guschin A, Molina LT, Molina MJ (1998) Atmospheric Chemistry of HFE-7200 (C4F9OC2H5): Reaction with OH Radicals, and Fate of C4F9OCH2CH2O(•) and C4F9OCHO(•)CH3 Radicals. J Phys Chem 102:4839–4845 4. Christensen LK, Wallington TJ, Gushin A, Hurley MD (1999) Atmospheric Degradation Mechanism of CF3OCH3 . J Phys Chem 103:4202–4208 5. DeMore WB, Sander SP, Golden DM, Hampson RF, Kurylo MJ, Howard CJ, Ravishankara AR, Kolb CE, Molina MJ (1997) Chemical kinetics and photochemical data for use in stratospheric modeling. JPL Publication 97–4 6. Francisco JS, Goldstein AN, Li Z, Zhao Y (1990) Theoretical investigation of chlorofluorocarbon degradation processes: structures and energetics of XC(O)Ox intermediates (X=F, Cl). J Phys Chem 94:4791–4795 7. Good DA, Kamboures M, Santiano R, Francisco JS (1999) Atmospheric oxidation of fluorinated ethers, E143a (CF3OCH3), E134 (CHF2OCHF2), and E125 (CHF2OCF3). J Phys Chem A 103:9230–9240 8. McFarland M, Kaye J (1992) Chlorofluorocarbons and ozone. Photochem Photobiol 55:911–929 9. Molina M, Rowland FS (1974) Stratospheric sink for chlorofluoromethanes: chlorine atom catalysed destruction of ozone. Nature 249:810–812 10. Nielsen OJ, Scott BF, Spencer C, Wallington TJ, Ball JC (2001) Trifluoroacetic acid in ancient freshwater. Atmos Environ 35:2799–2801 11. Ninomiya Y, Kawasaki M, Guschin A, Molina LT, Molina MJ, Wallington TJ (2000) Atmospheric chemistry of n-C3F7OCH3: Reaction with OH radicals and Cl atoms and atmospheric fate of n-C3F7OCH2O. radicals. Environ Sci Technol 34:2973 12. Nohara K, Toma M, Kutsuna S, Takeuchi K, Ibusuki T (2001) Cl atom-initiated oxidation of three homologous methyl perfluoroalkyl ethers. Environ Sci Technol 35:114–120 13. Nolan S, O’Sullivan N, Wenger J, Sidebottom H, Tracy J (1998) Kinetics and mechanisms of the OH radical initiated degradation of a series of hydrofluoroethers. In: Borrell PM, Borrell P (eds), Proceedings of EUROTRAC Symposium 1998, pp 120–123

102

T.J. Wallington · O.J. Nielsen

14. O’Sullivan N, Wenger J, Sidebottom H, Tracy J (1996) Kinetics and mechanisms of the OH radical initiated degradation of fluorinated ethers. In: Larsen B, Versino B (eds), Proceedings of the 7th European Commission symposium on physicochemical behavior of atmospheric pollutants, Venice, Italy, October, pp 77–79 15. Oram DE, Reeves CE, Sturges WT, Penkett SA, Fraser PJ, Langenfelds RL (1996) Recent tropospheric growth rate and distribution of HFC-134a (CF3CH2F). Geophys Res Lett 23:1949–1952; 25:35–38 16. Orkin VL, Villenave E, Huie RE, Kurylo MJ (1999) Atmospheric lifetimes and global warming potentials of hydrofluoroethers: Reactivity toward OH, UV spectra, and IR absorption cross sections. J Phys Chem A 103:9770–9779 17. Ravishankara AR, Lovejoy ER (1994) Atmospheric lifetime, its application and its determination: CFC-substitutes as a case study. J Chem Soc Faraday Trans 90:2159–2170 18. Ravishankara AR, Turnipseed AA, Jensen NR, Barone S, Mills M, Howard CJ, Solomon S (1994) Do hydrofluorocarbons destroy stratospheric ozone? Science 263:71–75 19. Schneider WF, Wallington TJ, Barker JR, Stahlberg EA (1998) CF3CFHO. Radical-decomposition vs. reaction with O2. Ber Bunsen Ges 102:1850–1856 20. Sehested J, Møgelberg T, Fagerstrom K, Mahmoud G, Wallington TJ (1997) Absolute rate constants for the self reactions of HO2 , CF3CFHO2 , and CF3O2 radicals and the cross reactions of HO2 with FO2 , HO2 with CF3CFHO2 , and HO2 with CF3O2 at 295 K. Int J Chem Kinet 29:673–682 21. Sihra K, Hurley MD, Shine KP, Wallington TJ (2001) Updated radiative forcing estimates of sixty-five halocarbons and non-methane hydrocarbons. J Geophys Res 22. Tokuhashi K, Takahashi A, Kaise M, Sekiya A,Yamashita S, Ito H (1999) Rate Constants for the Reactions of OH Radicals with CH3OCF2CH3 , CH3OCF2CF2CF3 and CH2OCF(CF3)2 . Int J Chem Kinet 31:847–853 23. Tokuhashi K, Takahashi A, Kaise M, Shigeo K, Sekiya A, Yamashita S, Ito H (2000) Rate Constants for the Reactions of OH Radicals with CH3OCF2CHF2 , CHF2OCH2CF2CHF2 , CHF2OCH2CF2CF3 , and CF3CH2OCF2CHF2 over the Temperature Range 250–430 K. J Phys Chem A 104:1165–1170 24. Tromp TK, Ko MKW, Rodriguez JM, Sze ND (1995) Potential accumulation of a CFCreplacement degradation product in seasonal wetlands. Nature 376:327-329 25. US Standard Atmosphere, (1976) NOAA, NASA, USAF, Washington DC 26. Wallington TJ, Schneider WF, Worsnop DR, Nielsen OJ, Sehested J, DeBruyn WJ, Shorter JA (1994) Atmospheric chemistry and environmental impact of CFC replacements: HFCs and HCFCs. Environ Sci Tech 28:320A–326A 27. Wallington TJ, Fracheboud J-M, Orlando JJ, Tyndall GS, Sehested J, Møgelberg TE, Nielsen OJ (1996) Role of excited alkoxy radicals in trifluoroacetic acid formation during atmospheric degradation of HFC-134a. J Phys Chem 100:18116–18122 28. Wallington TJ, Schneider WF, Sehested J, Bilde M, Platz J, Nielsen OJ, Christensen LK, Molina MJ, Molina LT, Wooldridge PV (1997) Atmospheric chemistry of HFE-7100 (C4F9OCH3): Kinetics of its reaction with OH radicals, UV spectra and kinetic data for the C4F9OCH2 and C4F9OCH2O2 radicals, and the atmospheric fate of C4F9OCH3O radicals. J Phys Chem 101:8264–8274 29. Wallington TJ, Gushin A, Stein TNN, Platz J, Sehested J, Christensen LK, Nielsen OJ (1998) Atmospheric Chemistry of CF3CH2OCH2CF3 : UV spectra and kinetic data for CF3CH(.)OCH2CF3 and CF3CH(OO.)OCH2CF3 radicals, and atmospheric fate of CF3CH(O.)OCH2CF3 radicals. J Phys Chem 102:1152–1161 30. Zang Z, Saini RD, Kurylo MJ, Huie RE (1992) Rate constants for the reactions of the hydroxyl radical with several partially fluorinated ethers. J Phys Chem 96:9301–9304

CHAPTER 4

Trifluoroacetic Acid and Longer Chain Perfluoro Acids – Sources and Analysis David A. Ellis · Cheryl A. Moody · Scott A. Mabury Department of Chemistry, University of Toronto, 80 St. George St., Toronto, Ontario M5S 3H6, Canada E-mail: [email protected]

Perfluorinated organic acids are ubiquitous in the environment at relatively low concentrations and appear to be highly persistent. Analytical techniques have relied primarily on GCbased methods, and for the lowest member TFA involve derivatization. The longer chain perfluoro acids have only recently been routinely analyzed at environmental concentrations using electrospray-MS/MS systems. Their analysis at low concentrations in environmental matrices requires extensive concentration. For the longer chain acids, this generally relies on liquid-liquid or SPE based methods and, for TFA anion-exchange or evaporation of water. Industrial and consumer applications are the primary source of emissions of the higher perfluoro acids while TFA results primarily from both the atmospheric degradation of CFC replacements and from the degradation of fluoropolymers in high temperature applications. Except for the question of a natural source, the current knowledge of the environmental chemistry of TFA is relatively mature, whereas that for the higher perfluoro acids is only beginning to emerge. Keywords. Trifluoroacetic acid, Analysis, Perfluorinated acids, Sources, Environmental fate, Perfluorooctanesulfonate

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

1

Introduction

2 2.1 2.2 2.3 2.4 2.5

Environmentally Significant Sources of TFA HFCs and HCFCs . . . . . . . . . . . . . . . TFA . . . . . . . . . . . . . . . . . . . . . . Trifluoromethyl Aromatic Compounds . . . Alternative Anthropogenic Sources . . . . . Natural Sources of TFA . . . . . . . . . . . .

3 3.1 3.2

Environmental Measurement of TFA . . . . . . . . . . . . . . . . 111 GC Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Alternative Procedures . . . . . . . . . . . . . . . . . . . . . . . . 114

4

Analysis of Longer Chain Perfluoro Acids

5

Biological and Environmental Measurements of Longer Chain Perfluoro Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

6

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

7

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

105 105 106 106 107 109

. . . . . . . . . . . . . 114

The Handbook of Environmental Chemistry Vol. 3, Part N Organofluorines (ed. by A.H. Neilson) © Springer-Verlag Berlin Heidelberg 2002

104

D.A. Ellis et al.

Abbreviations CFC CPFP 2,4-DFAn DMS ECD GC GC/ECD HCFC HFC HFP HF HPLC HS IC IEC LC/MS LC/MS/MS LSC m MS NMR PCTFE PFPD PFOA PFOS PFHxS PFOSA SAX SPE TFA TFM

Chlorofluorocarbon Chloropentafluoropropene 2,4-Difluoroaniline Dimethyl Sulfate Electron Capture Detection Gas Chromatography Gas Chromatography/Electron Capture Detection Hydrochlorofluorocarbon Hydrofluorocarbon Hexafluoropropene Hydrofluoric Acid High Performance Liquid Chromatography Head Space Ion Chromatography Ion Exclusion Chromatography Liquid Chromatography/Mass Spectrometry Liquid Chromatography/Tandem Mass Spectrometry Liquid Scintillation Counting Micro Mass Spectrometry Nuclear Magnetic Resonance Spectroscopy Polychlorotrifluoroethylene 1-(Pentafluorophenyl)diazoethane Perfluorooctanoic Acid Perfluorooctanesulfonate Anion Perfluorohexanesulfonate Anion Perfluorooctanesulfonylamide Strong Anionic Exchange Solid Phase Extraction Trifluoroacetic Acid 3-Trifluoromethyl-4-nitrophenol

1 Introduction The aim of this chapter is to assess current understanding of perfluorinated acids in the environment with specific attention directed to their analysis and a discussion of major sources. The extreme stability of the perfluorinated acids which have no established environmentally relevant degradation pathway, redefines persistence that is well exemplified by chlorinated pesticides and industrial compounds. Major interest has arisen from the demonstration that TFA is the major and persistent degradation product of the CFC replacement gases. It is biodegradable under laboratory anaerobic conditions using ethanol as a source of elec-

Trifluoroacetic Acid and Longer Chain Perfluoro Acids – Sources and Analysis

105

trons [32], and its toxicity to algae, terrestrial plants, and mammals, and its mutagenicity using the Salmonella typhimurium assay including the S-9 microsomal activation system have been described [8]. Due to the fact that more TFA occurs in rainwater than can be attributed to these replacements, great effort has been focused on determining alternate sources of TFA. It is intriguing to question whether there are other major industrial sources of TFA and whether a significant natural source exists. Current interest is focused on the longer chain perfluorinated acids (both carboxylic and sulfonic) that appear to be equally persistent and widespread in the environment. Unlike TFA, some of these longer perfluoro acids readily accumulate, so that residues have been detected in a variety of biota collected round the world. Currently almost nothing is known about the environmental chemistry of these compounds partly due to problems in handling them and to the fact that the required analytical procedures are not routinely available in most laboratories.An important area of significance is their influence in proliferating peroxisomes in rodents and this is covered in detail by DePierre in his chapter in this volume. This chapter highlights the current methods for analysis of both TFA and longer chain perfluoro acids and contains an evaluation of the strengths and weaknesses of the various approaches.We conclude the chapter with a short discussion on current knowledge of longer chain perfluoro acids and of their wide dissemination in the environment.

2 Environmentally Significant Sources of TFA 2.1 HFCs and HCFCs

In 1988 it was reported that three of the CFC replacement gases HCFC-123 (CCl2HCF3), HCFC-124 (CClHFCF3), and HFC-134a (CF3CFH2) were potential sources of TFA in the environment through atmospheric degradation processes [51], and the atmospheric chemistry of these and related compounds is discussed in detail by Wallington and Nielsen in this volume. Although HFC-134a yields only 37% TFA, compared with a 100% yield for the other two gases, it is the most significant source due to the relatively high rates of relative production and consumption [estimated global emissions for the year 2010 are HCFC-123 (23.3 kt), HCFC-124 (49.2 kt), and HFC-134a (221 kt)] [57]. The tropospheric mechanism for the production of TFA is similar for all three gases, and the general mechanism that is outlined in Fig. 1 for HFC-134a is thought to occur via initial hydrogen abstraction by a hydroxyl radical to form an alkyl radical. This alkyl radical then reacts with oxygen to form a peroxy radical that undergoes reaction with oxides of nitrogen to form an alkoxy radical. The alkoxy radical then rapidly decays to yield the trifluoroacetyl halide. The acetyl halide is then hydrolyzed to TFA.

106

D.A. Ellis et al.

Fig. 1. Production of TFA through the atmospheric degradation of HFC 134a

2.2 TFA

Modeling calculations were conducted in 1995 to estimate the potential for accumulation of TFA in seasonal wetlands by the year 2010 [57]. These calculations were based upon the known chemistry of the CFC replacement gases in the troposphere. The results of the modeling suggested that if these gases were the single source contributing to the atmospheric burden of TFA, the global rain concentration would be, on average, 0.16 mg/L by 2010. Subsequent environmental measurements [11, 19, 28, 64] of TFA, however, have shown that current levels cannot originate only from these sources. For example, in 1999 it was shown that in certain regions, concentrations of TFA in rainwater averaged 0.12 mg/L [28]. These observations have led other researchers to investigate possible alternatives including natural sources of TFA. 2.3 Trifluoromethyl Aromatic Compounds

Ellis and Mabury have shown that the incorporation of trifluoromethyl substituents into the aromatic ring of pesticides such as the Lampricide TFM, could provided a source of TFA in the aquatic environment [15]. A detailed outline of the proposed mechanism for the production of TFA from TFM is given in Fig. 2 and highlights the key step in the formation of TFA as the generation of a triflu-

Trifluoroacetic Acid and Longer Chain Perfluoro Acids – Sources and Analysis

107

Fig. 2. The production of TFA from the photolysis of TFM (Reproduced with permission from

Environ Sci Tech (2000) 34:623–637. Copyright 2000 American Chemical Society)

oromethylquinone through a photonucleophilic displacement of the original aromatic nitro group by water. This results in the production of a hydroquinone that is subsequently oxidized, and it was proposed that the potential for formation of TFA from the trifluoromethyl group on an aromatic ring exists if oxidation to such an intermediate can occur; the yield of TFA would clearly be dependent on the degree to which quinone is produced. Although the production of TFA from TFM is expected to be a significant source of TFA in surface waters where lampricide treatment has been applied, it is not expect to be a generally important source of TFA in rainwater. This is due to the high water solubility of TFM compared with its vapor pressure. It is hypothesized, however, that if the trifluoromethyl pesticide had a significantly high vapor pressure, oxidation might occur in the atmosphere in the gas phase in water droplets, or on particulate matter. If these processes were to occur they might contribute to the levels of TFA that have been observed in rainwater. 2.4 Alternative Anthropogenic Sources

It has been suggested that the most probable alternate anthropogenic sources of TFA are the combustion of fluoropolymers, electrolytic aluminum production, or the atmospheric degradation of fluorinated inhalation anesthetics [28]. Electrolytic aluminum production is unlikely to be a major ubiquitous source due to its regional use. Similarly, the degradation of anesthetics is unlikely to contribute significantly to the atmospheric burden of TFA, due to the low quantities which are being used or emitted (estimated at 6 kt/year [28]); these would yield an

108

D.A. Ellis et al.

Fig. 3. Proposed mechanism for the production of fluoro acids through the thermolysis of flu-

oropolymers

average rainfall concentration of less than 5 ng/L. It has even been suggested that the concentrations of atmospheric halogenated inhalation anesthetics are significantly less than the figures used in making these estimates [24]. It would therefore seem most likely that the major atmospheric source of TFA is from the thermal degradation of fluoropolymers such as polytetrafluoroethylene (PTFE). Ellis et al. have recently shown that, for all classes of fluoropolymer that were tested, TFA was produced at the onset of thermal breakdown of the compounds [18]. The mechanism for the production of TFA and other longer chain fluoro acids was believed to occur through initial scission of the polymer chain to produce a biradical species (Fig. 3). It was postulated that this biradical then abstracts a fluorine atom to yield a single alkyl radical. Reaction of the fluoroalkyl radical with oxygen would then produce an end-chain peroxy radical which degrades to an alkoxyl radical. This alkoxyl radical then abstracts a proton, presumably from atmospheric water, to form the corresponding alcohol. The degradation of this alcohol then proceeds by elimination of HF to yield an acyl fluoride that is then rapidly hydrolyzed to the corresponding acid. A second, or additional mechanism for the production of TFA is through the thermal oxidation of HFP that is also a thermolysis product of fluoropolymers. The oxidation of HFP results in the production of the transient trifluoroacetyl fluoride and again TFA is produced through hydrolysis (Fig. 4) [10]. Using environmental modeling calculations, it was suggested that this source may play an important role in explaining the atmospheric burden of TFA. Ellis et al. also showed that in addition to the production of TFA, many other longer chain fluoro acids were produced, and that in the case of fluoropolymers that contained chlorine atoms, such as PCTFE, chlorofluoro acids were produced along with fluoro acids. In addition to the direct production of TFA, other species such as HFP and CPFP were also observed during thermolysis. These compounds are believed to degrade in the troposphere to produce haloacetic acids [40]. These observation plausibly explain the long-range transport of haloacetates such as TFA, and a detailed scheme for the reaction of HFP with hydroxyl radicals that ultimately produces TFA is given in Fig. 5. The quantitative production of TFA from

Fig. 4. Oxidation of HFP produced via the thermolysis of fluoropolymers leading to TFA

Trifluoroacetic Acid and Longer Chain Perfluoro Acids – Sources and Analysis

109

Fig. 5. Proposed atmospheric source of TFA through degradation of HFP with hydroxyl radi-

cals

HFP was also shown to occur by chlorine radical initiation in place of the hydroxyl radical. If the atmospheric degradation of HFP were a significant source of TFA, the estimated 9-day lifetime of the parent molecule would allow for long-range transport of TFA [18]. A summary of the environmental implications of the thermal degradation of fluoro- and chlorofluoropolymers is given in Fig. 6. Environmental modeling for the direct production of TFA showed that fluoropolymers indeed have the potential to contribute a significant amount of TFA to the atmosphere from which they would be deposited primarily as rainfall (approx. 25 ng/L) [18]. 2.5 Natural Sources of TFA

Although nature herself is not known to produce trifluoromethylated compounds biologically [22], it has been suggested that geological events such as volcanic activity might yield such compounds and that these could in turn result in either the direct or indirect production of TFA [26]. Reference should be made to chapter by Gribble for a full discussion of naturally occurring fluorinated compounds. Although volcanoes have been shown to produce significant quantities of fluorinated organic compounds along with gaseous hydrogen fluoride [16], no potential precursors to TFA have been observed. The hypothesis of natural production is, however, supported by two key pieces of evidence. The first is the observation of TFA in glacier ice that formed during the 19th century [59], from

110

D.A. Ellis et al.

Fig. 6. Proposed environmental reaction pathways for the thermal degradation of fluoro- and chlorofluoro-polymers. Square boxes represent environmentally transient species (t1/2 <1 year) and curved boxes represent compounds with significant lifetimes (t1/2 >10 years) and important environmental impacts. A significant concentration of perhalo acids (x=0–2, y=1–3, z=1–2, n=1–13, m=1–7) are produced, primarily TFA and CDFA (1–10% w/w). Tropospheric oxidation of the perhalopropenes also predominantly leads to, or is expected to lead to, the formation of TFA and CDFA. (Reproduced with permission from Nature 2001, 412:321–324. Copyright 2001 Nature.)

which it is concluded that there existed a pre-industrial source of TFA which might be volcanic. It is possible, however, that this is the result of pre-industrial anthropogenic activities such as the burning of fossil fuels. Further support for this hypothesis could be obtained from a correlation between TFA levels in dated ice samples and the amount of fossil fuels combusted at those locations together with their geographical location. It is possible that application of stable carbon isotope ratios (d13C) might shed light on this issue [27].The second piece of evidence comes from the observation of TFA in deep ocean water samples which are extremely old [14]. If this observation were corroborated and it could be shown that TFA is found at comparable concentrations throughout the oceans of the world, effort would be justified in establishing a significant natural source. Geogenic processes have been linked to the production of TFA which might account for these observations. Although, for example, gaseous fluorinated compounds have been found as inclusion gases in igneous and metamorphic rocks, to date only methane-based fluorinated organics have been observed and these would presumably not lead to TFA. In addition, SF6 , that is known to be highly persistent in the environment, was observed in these rocks [58], although SF6 has

Trifluoroacetic Acid and Longer Chain Perfluoro Acids – Sources and Analysis

111

apparently not been observed at the levels remotely similar to those of TFA. This calls into doubt the significance of this potential source. Furthermore, freshwater samples collected from Greenland and Denmark which ought to have been exposed to such sources for >2000 years appear not to contain any significant amounts of TFA (<2 ng/L) [44]. These results suggest that local geogenic activity in these areas is not an important source of TFA. Published results on these issues are not sufficiently complete to determine whether multiply fluorinated TFA has a natural source. This issue remains open to further research and discussion.

3 Environmental Measurement of TFA Several procedures exist for the measurement of TFA in a wide range of matrices. The types of matrix, the extraction/derivatization procedures, the analytical methods, and the environmental concentrations are given in Table 1 along with the appropriate references. Several analytical techniques have been developed to monitor TFA including gas chromatography (GC) coupled to electron capture detection (ECD), mass spectrometry (MS) or atomic emission detection (AED). Other techniques include 19F-NMR spectroscopy, ion chromatography (IC), ion exclusion chromatography (IEC), and liquid scintillation counting (LSC) of 14Clabeled TFA. The choice of the appropriate analytical technique is predetermined by several factors, such as the concentration of TFA in the original matrix and the necessity for derivatization. Examples of pre-concentration methods include extraction of the TFA from an aqueous matrix using strong anionic extraction cartridges (SAX) or through evaporation of the water. The latter has the disadvantage that it is relatively time consuming and is not convenient when many analysis have to be carried out simultaneously. It does have the advantage, however, that when other anions that may interfere with the recovery of TFA from a SAX column are present in the sample, this is avoided by evaporation. 3.1 GC Methods

When GC analysis is to be preformed, a derivatization step must be employed due to the highly polar nature of TFA. Researchers have used a variety of reagents. Overall, it would appear that the method of Frank et al. [3, 19, 20, 28, 49, 59], shows the lowest limits of detection due to the high response of the derivatized samples to detection by electron capture detection (ECD). Frank’s method has proven to be generally applicable to a wide variety of matrices (see Table 1). The method has, however, the disadvantage that the derivatizing agent [1-(pentafluorophenyl)diazoethane] is not generally commercially available and therefore must be synthesized. In addition, the reagent is both toxic and highly reactive, and therefore requires care in its preparation and use. The method reported by Scott et al. [52, 53], has been shown to yield similarly low detection limits and has also been applied to a wide variety of matrices. This method has the added ad-

PFPD

Extraction with (CH3CH2)2O

c

b

a

IEC LSC

Direct Measurement N/A Direct Measurement N/A

20 µg/L N/G

N/G

0.2 ng/L

6.5 ng/L – water 34 ng g–1 – soil 34 ng g–1 – plant 0.18 ng/L

100 ng/L 25 ng/L – water 1 ng m3 – air 36 ng/L

LOQ – Limit of quantification. N/A – Not applicable. N/G – Not given. Values are given for data obtained in the reference containing the original method. References include publications that have employed the original method.

IC

GC-MS (or) GC-AED

GC-MS (NCI)

Direct Measurement N/A

2,4-DFAn

H2SO4/CH3OH

Water Evaporation (NH4)2CO3 added

Water Evaporation

HS-GC-ECD

H2SO4/CH3OH

SAX

Lake water Rain water Snow Ground water Drinking water Soil Vegetation Microbial Microbial

IC HS-GC-MS (or) HS-GC-ECD HS-GC-ECD

N/A H2SO4/DMS

SAX Water evaporation

16 ng/L

F-NMR

19

N/A

SAX

Method LOQ a

Analysis Method

Rain water River water Pond water Surface waters Air Surface waters Rain water Fog water Surface waters Soil Plants Rain-water-airsurface waters and biomass

Derivatization Method

Extraction/ Concentration Method

Environmental matrix

Table 1. Methods of analysis for TFA in environmental matrices

N/A N/A

<3–120 pg g–1 – rain <3– 3230 pg g–1 – air <2– 1030 pg g–1 – surface waters and biomass < 0.2 – 240 ng/L – lake water <0.2–60 ng/L – rain water <0.2–52 ng/L – snow < 0.2 ng/L – ground water <0.2–450 ng/L – drinking water N/A

74–850 ng/L – rain N/A – river N/A ≤ 280 ng/L – surface ≤ 3.3 ng m3 – air 81–275 ng/L – surface 149 ng/L – rain 2154 ng/L – fog N/A

Environmental Concentrations Observed b

[6, 7] [54, 61]

[4, 5, 38, 48]

[52, 53]

[3, 19, 20, 28, 49, 59]

[11, 12, 13]

[62, 63]

[17] [65]

[15, 16]

References c

112 D.A. Ellis et al.

Trifluoroacetic Acid and Longer Chain Perfluoro Acids – Sources and Analysis

113

Fig. 7. Methods of derivatization for TFA

vantage that the derivatizing agent is commercially available and is user friendly. The disadvantage of this method lies in the requirement for pre-concentrating the samples. It necessitates the evaporation under vacuum of large volumes of water which is time consuming and laborious, and limits the number of samples which can be processed at one time. The method developed by Wujcik et al. [63], while less sensitive than the two methods previously outlined, owes its advantages to its ease of use. Samples can be efficiently and rapidly pre-concentrated and derivatized with relative ease. The method has also proven to be generally applicable to a variety of matrices. These three methods of derivatization that incorporate a pentafluorinated aromatic ester (Frank), a difluorinated aromatic amide (Scott), and methylation (Wujcik) are shown in Fig. 7. Each of these methods fulfills the prerequisites of an efficient derivatization in that they are quantitative for the TFA. The limits of quantification using these derivatives follow the general trend: pentafluorinated aromatic ester >di-fluorinated aromatic amide >methylation. It would therefore appear that the choice of GC method is defined by the concentration of the TFA. Generally, if TFA concentrations are high, Wujcik’s method would be method of choice, and if concentrations are lower than those which can be analyzed using this method, then either Frank’s or Scott’s method should be employed. Wujcik’s method unlike the others does not, however, lend itself to the analysis of other haloacetic acids.

114

D.A. Ellis et al.

3.2 Alternative Procedures

Direct analysis of TFA in aqueous media has been achieved using several techniques including 19F-NMR, IC and IEC. Use of these methods has the advantage that no derivatization step is required which reduces the time required for a set of analyses and makes the overall process more efficient. These techniques do, however, have the drawback that they are typically an order or two less sensitive than their GC counterparts (cf. mg/mL for 19F-NMR and ng/mL for GC-ECD). They therefore invariably require a substantial pre-concentration step to achieve comparable results for the analysis of typical environmental concentrations. In summary, each of these methods has been successfully employed in the analysis of TFA in environmentally diverse matrices and at various concentrations (Table 1). For example, TFA has been measured in samples of rainwater (<0.2–850 ng/L), surface waters (<0.2–280 ng/L), fog condensate (2154 ng/L), drinking water (< 0.2 – 450 ng/L), air (≤ 3.3–3230 ng/L), and in plant cultures (20 mg/L). Each of the methods has been successfully employed in subsequent analysis by the same or independent researchers (see Table 1 for references).

4 Analysis of Longer Chain Perfluoro Acids Longer chain perfluoro acids, including PFOS and PFOA (Fig. 8), have attracted attention due to their detection in human and animal samples [9, 47, 60]. Although relatively little is known of their environmental behavior, an important aspect of their biochemical toxicology has been addressed by DePierre in his chapter.Additional information is required to evaluate these perfluorinated compounds, with emphasis on the need for analytical methods to determine longer chain perfluoro acids in a variety of matrices. Historically, a wide range of methods has been employed for the determination of perfluorinated acids and these are reviewed elsewhere [35–37]. This sec-

Fig. 8. Chemical structures of PFOS and PFOA

Trifluoroacetic Acid and Longer Chain Perfluoro Acids – Sources and Analysis

115

tion highlights several analytical methods, including structure-specific procedures for the determination of longer chain perfluoro acids. As early as the 1960s, organic and inorganic fluorine were observed in human blood [55, 56]. Analytical methods, however, primarily non-structure-specific techniques, limited the appreciation of the occurrence, role, and recalcitrance of longer chain perfluoro acids such as PFOS and PFOA. A commonly employed method for the analysis of perfluoro acids converted the organofluorine compounds to fluoride, which was then subsequently analyzed by fluoride-specific electrodes. Typically, an oxyhydrogen flame combusted the sample and the HF was trapped in an aqueous solution [1, 33]. One application of the oxyhydrogen torch method was the evaluation organofluorine contaminants in air, where total organofluorine concentrations (mg/L) were measured [34]. 19 F-NMR has been widely used to investigate the chemistry of longer chain perfluoro acids in solution. Guy et al. [23] fractionated and concentrated a large volume of pooled human blood to isolate the primary fluoroorganic compounds present; subsequent analysis by 19F-NMR yielded a spectrum similar to that of PFOA. More recently, 19F-NMR was employed for the determination of longer chain perfluoro acids, including PFOS and PFOA [43], where quantitation was based on the peak area of the terminal CF3 group. The chemical shift for longer chain perfluoro acids was approximately –79 ppm. 19F-NMR methodologies similar to those used for TFA [16] and fluorinated pesticides [39] may have additional application to longer chain perfluoro acids. A detailed review of the principles of 19F-NMR and exemplification of their application is given by Stanley in this volume. Several chromatography-based methods have been developed for longer chain perfluoro acids but have not had wide implementation. As an example, biological samples were liquid-liquid extracted, and the extracts which contained the fluorinated acids of interest (i.e., PFOA) were methylated with diazomethane and analyzed by GC/ECD [2]. Alternatively, Ohya et al. [45] determined perfluorocarboxylic acid (7–10 carbons) concentrations in liver tissue samples using HPLC with fluorescence detection. Mass spectrometric techniques provide structural information that aids in the identification of longer chain perfluoro acids in biological and environmental matrices. Moody and Field [41] developed a solid phase extraction (SAX) method with in-vial derivatization that employed gas chromatography and mass spectrometry to characterize and quantify the methyl esters of PFOA and lower chained homologues (6–7 carbons) in groundwater. Unfortunately, GC/MS has limited utility with longer chain perfluoro acids, as several are non-volatile and derivatization is required [42]. For the analysis of perfluoro acids, liquid chromatography/mass spectrometry has proven to be a useful technique for both biological and environmental application. Early LC/MS-based methodologies employed a thermospray interface to determine perfluoro acids in matrices such as serum [46] and wastewater [50]. More recently, Hansen et al. [25] provided details of an ion pair extraction method with detection by LC/MS/MS (negative ion electrospray) for low level (µg/L) concentrations of perfluoro acids (PFHxS, PFOS, PFOA, and PFOSA) in human sera and liver tissue. Limits of detection for PFOA, PFOS, PFOSA, and

116

D.A. Ellis et al.

PFHxS in unconcentrated sera (0.5 mL sample volume) were 1.0 mg/L, 1.7 mg/L, 1.5 mg/L and 2.0 mg/L, respectively [52]; detection limits for organic fluorochemicals in liver homogenate (1 g of liver tissue to 5 mL deionized water) for the same suite of perfluorinated compounds ranged from 2.0 mg/L to 8.5 mg/L. In addition to LC/MS/MS-based methods for biological samples, Moody et al. [43] employed LC/MS/MS for the determination of perfluoroalkanesulfonates (6 and 8 carbons) and perfluorocarboxylates in aqueous samples. Surface water samples (0.2 to 200 mL) were extracted by solid phase extraction (C18), the compounds of interest eluted with methanol, and extracts analyzed by LC/MS/MS. For a 100-mL surface water sample, the limits of quantitation for PFOS and PFOA were 17 and 9 ng/L, respectively. Mass spectrometry methods, especially LC/MS/MS methods for biological and environmental matrices, are robust, routine, and permit the detection of longer chain perfluoro acids at environmentally relevant concentrations (ng/L to mg/L).

5 Biological and Environmental Measurements of Longer Chain Perfluoro Acids Analytical methods that combine liquid chromatographic separation with MS spectrometry facilitate the determination of perfluoro acids in complex matrices. For biological systems, it was reported in the literature that PFOS concenTable 2. Concentrations of Perfluorinated Acids in Biological and Environmental Matrices by LC/MS/MS

Sample Matrix

n

PFOS

PFHxS

PFOA

Reference

Human plasma (occupationally exposed to fluorochemical production) Human plasma (nonoccupationally exposed) Ringed seal plasma (Baltic sea) Bald eagle plasma (Midwestern USA) River otter liver (West Coast, USA) Great egret liver (Florida, USA) Surface water (Ontario, Canada)

149

100–9930 µg/L

not reported

not reported

[46]

65

6.7–81.5 µg/L 16–230 ng/mL 1–2570 ng/mL 33.6–994 ng/g 81–1030 ng/g LOD-2210 µg/L

< LOD-21.4 µg/L



[21]

not reported not reported LOD-49.6 µg/L

not reported not reported 0.011–11.3 µg/L

[30]

18 26 5 3 9

[31] [43]

Method limit of detection for PFHxS in non-concentrated sera by LC/MS/MS is 2 µg/L [25]. Method limit of quantitation for PFOA in non-concentrated sera by LC/MS/MS is 1 µg/L [25]. Method limit of quantitation for a 100-mL sample by LC/MS/MS is 17 ng/L for perfluoroalkanesulfonates [43].

Trifluoroacetic Acid and Longer Chain Perfluoro Acids – Sources and Analysis

117

trates primarily in the liver and plasma [46, 47]. In 1995, Olsen et al [46] measured mg/L concentrations of PFOS in the serum of occupationally exposed fluorochemical production workers by LC/MS. Two years later, employing improved methodologies, PFOS concentrations (mg/L) were again measured in serum (Table 2). Employing LC/MS/MS methods [25], PFOS and PFOA concentrations in sera of 65 non-occupationally exposed individuals ranged from 6.7 to 81.5 mg/L and
6 Conclusions TFA and the longer chain perfluorinated acids are of considerable interest on account of their persistence, wide dissemination in the environment, and novel properties that differ from those of other ‘persistent organic pollutants’ such as neutral chlorinated hydrocarbons. The fact that they are ionized completely under all environmental conditions necessitates their derivatization for GC-based techniques (TFA and to a lesser extent some long chain perfluoroalkanoic acids). For the longer chain acids, their anionic character has made them ideal for electrospray-based techniques in the negative-ion mode making it possible to attain the sublimely low detection limits required for investigating these compounds in the environment.Although TFA analysis would presumably also benefit from this technique, LC/MS/MS instruments may also be used for the biochemical analysis of metabolites in buffers containing TFA, and this would inevitably lead to difficulties in obtaining clean blanks. Although the reasons behind the widespread distribution of the longer chain perfluoro acids are unresolved, it seems unlikely that there are major sources other than their industrial and consumer use as surface active agents. The primary sources of TFA in rainwater are emissions from fluoropolymer applications under conditions of high thermal stress, with lesser amounts from hydroxyl radical-mediated degradation of some CFC replacements. The question of a natural

118

D.A. Ellis et al.

source for TFA remains unanswered although no hitherto published data have shown either a demonstrable mechanism for its production or evidence for its source. The suggestion that there may be a natural reservoir arises from the observation of ng/L levels of TFA in samples of ocean water. If these levels are proven to be spatially consistent, extrapolation would result in a very high total amount of TFA in the environment. Current published data do not, however, permit any firm conclusion.

7 References 1. Belisle J, Hagen DF (1978) Method for the determination of the total fluorine content of whole blood, serum/plasma, and other biological samples. Anal Biochem 87:545–555 2. Belisle J, Hagen DF (1980) A method for the determination of perfluorooctanic acid in blood and other biological samples. Anal Biochem 101:369–376 3. Berg M, Muller SR, Muhlemann J, Wiedmer A, Schwarzenbach RP (2000) Concentrations and mass fluxes of chloroacetic acids and trifluoroacetic acid in rain and natural waters in Switzerland. Environ Sci Technol 34:2675–2683 4. Berger TW, Likens GE (1999) Effects of acid anion additions (trifluoroacetate and bromide) on soil solution chemistry of a northern hardwood forest soil. Water Air Soil Pollut 116:479–499 5. Berger TW, Tartowski SL, Likens GE (1997) Trifluoroacetate retention in a northern hardwood forest soil. Environ Sci Technol 31:1916–1921 6. Bott TL, Standley LJ (1999) Incorporation of trifluoroacetate, a hydrofluorocarbon decomposition by product, by freshwater benthic microbial communities. Water Res 33:1538–1544 7. Bott TL, Standley LJ (1998) Effects of trifluoroacetate, an atmospheric breakdown product of hydrofluorocarbon refrigerants, on acetate metabolism by freshwater benthic microbial communities. Bull Environ Contam Toxicol 60:472–479 8. Boutonnet JC, Bingham P, Calamari D, De Rooij D, Franklin J, Kawano T, Libre JM, McCulloch A, Malinverno G, Odom JM, Rusch GM, Smythe K, Sobolev I, Thompson R, Tiedje JM (1999) Environmental risk assessment of trifluoroacetic acid. Human Ecol Risk Assess 5:59–124 9. Brown D, Mayer, CE (2000) In: Washington Post: Washington, DC, 2000, pp A01. 10. Buravtsev NN, Kolbanovsky YA (1999) Intermediates of thermal transformations of perfluoro-organic compounds. New spectral data and reactions. J Fluor Chem 96:35–42 11. Cahill TM, Seiber JN (2000) Regional distribution of trifluoroacetate in surface waters downwind of urban areas in Northern California. USA. Environ Sci Technol 34 : 2909–2912 12. Cahill TM, Thomas CDM, Schwarzbach SE, Seiber JN (2001) Accumulation of trifluoroacetate in seasonal wetlands in California. Environ Sci Technol 35:820–825 13. Cahill TM, Benesch JA, Gustin MS, Zimmerman EJ, Seiber JN (1999) Simplified method for trace analysis of trifluoroacetic acid in plant, soil, and water samples using headspace gas chromatography. Anal Chem 71:4465–4471 14. Christoph EH, Frank H (2000) Oceanic distribution of trifluoroacetate. International Symposium, Haloacetic acids and short chain halocarbons: Sources and fate in the environment. University of Toronto, Toronto, ON Canada. 15. Ellis DA, Mabury SA (2000) The aqueous photolysis of TFM and related trifluoromethylphenols. An alternate source of trifluoroacetic acid in the environment. Environ Sci Technol 34:623–637 16. Ellis DA, Martin JW, Muir DCG, Mabury SA (2000) Development of an F-19 NMR method for the analysis of fluorinated acids in environmental water samples. Anal Chem 72:726–731

Trifluoroacetic Acid and Longer Chain Perfluoro Acids – Sources and Analysis

119

17. Ellis DA, Hanson ML, Sibley PK, Shahid T, Fineberg NA, Solomon KR, Muir DCG, Mabury SA (2001) The fate and persistence of trifluoroacetic and chloroacetic acids in pond waters. Chemosphere 42:309–318 18. Ellis DA, Mabury SA, Martin JW, Muir DCG (2001) Thermolysis of fluoropolymers as a potential source of halogenated organic acids in the environment. Nature 412:321–324 19. Frank H, Klein A, Renschen D (1996) Environmental trifluoroacetate. Nature 382: 34–34 20. Frank H, Renschen D, Klein A, Scholl H (1995) Trace analysis of airborne haloacetates. J High Resol Chromatogr 18:83–88 21. Giesy JP, Kannan K (2001) Global distribution of perfluorooctanesulfonate in wildlife. Environ Sci Technol 35:1339–1342 22. Gribble GW (1992) Naturally occurring organohalogen compounds – A survey. J Nat Prod 55:1353–1395 23. Guy WS, Taves DR, Brey WS (1976) Amer Chem Soc Symp Ser 28:117 24. Halpern DF (1994) In: Banks RE, Smart E., Tatlow JC, (eds) Organofluorine Chemistry: Principles and Commercial Applications. Plenum Press: New York, p 549 25. Hansen KJ, Clemen LA, Ellefson ME, Johnson HO (2001) Compound-specific, quantitative characterization of organic fluorochemicals in biological matrices. Environ Sci Technol 35:766–770 26. Harnisch J, Frische M, Borchers R, Eisenhauer A, Jordan A (2000) Natural fluorinated organics in fluorite and rocks. Geophys Res Lett 27:1883–1886 27. Harper DB, Calin RM, Hamilton JTG, Lamb C (2001) Carbon isotope ratios for chloromethane of biological origin: potential tool in determining biological emissions. Environ Sci Technol 35:3616–3619 28. Jordan A, Frank H (1999) Trifluoroacetate in the environment. Evidence for sources other than HFC/HCFCs. Environ Sci Tech 33:522–527 29. Jordan A, Harnisch J, Borchers R, Le Guern FN, Shinohara H (2000) Volcanogenic halocarbons. Environ Sci Technol 34:1122–1124 30. Kannan K, Koistinen J, Beckmen K, Evans T, Gorzelany JF, Hansen KJ, Jones PD, Helle E, Nyman M, Giesy JP (2001) Accumulation of perfluorooctanesulfonate in marine mammals. Environ Sci Technol 35:1593–1598 31. Kannan K, Franson JC, Bowerman WW, Hansen KJ, Jones PD, Giesy JP (2001) Perfluorooctanesulfonate in fish-eating water birds including bald eagles and albatrosses. Environ Sci Technol 35:3065–3070 32. Kim BR, Suidan MT, Wallington TJ, Du X (2000) Biodegradability of trifluoroacetic acid. Environ Eng Sci 17:337 –342 33. Kissa E (1983) Determination of fluoride at low concentration with the ion-selective electrode. Anal Chem 55:1445–1448 34. Kissa E (1986) Determination of organofluorine in air. Environ Sci Technol 20:1254–1257 35. Kissa E (2001) Fluorinated Surfactants and Repellents, 2nd ed, Marcel Dekker, Inc.: New York 36. Kissa E (1994) Fluorinated Surfactants: Synthesis, Properties, and Applications. Marcel Dekker: New York 37. Kissa E (1998) In: Cross J (ed) Anionic Surfactants: Analytical Chemistry. Marcel Dekker: New York, Vol 73 38. Likens GE, Tartowski SL, Berger TW, Richey DG, Driscoll CT, Frank HG, Klein A (1997) Transport and fate of trifluoroacetate in upland forest and wetland ecosystems. Proc Natl Acad Sci USA 94:4499–4503 39. Mabury SA, Crosby DG (1995) F-19 NMR as an analytical tool for fluorinated agrochemical research. J Agric Food Chem 43:1845–1848 40. Mashino M, Ninomiya Y, Kawasaki M, Wallington TJ, Hurley MD (2000) Atmospheric chemistry of CF3CF=CF2 : kinetics and mechanism of its reaction with OH radicals, chlorine atoms, and ozone. J Phys Chem A 104:7255–7260 41. Moody CA, Field JA (1999) Determination of perfluorocarboxylates in groundwater impacted by fire-fighting activity. Environ Sci Technol 33:2800–2806

120

D.A. Ellis et al.

42. Moody CA, Field JA (2000) Perfluorinated surfactants and the environmental implications of their use in fire-fighting foams. Environ Sci Technol 34:3864–3870 43. Moody CA, Kwan WC, Martin JW, Muir DCG, Mabury SA (2001) Determination of perfluorinated surfactants in surface water samples by two independent analytical techniques: liquid chromatography/tandem mass spectrometry and 19F NMR. Anal Chem 73:2200– 2206 44. Nielsen OJ, Scott BF, Spencer C,Wallington TJ, Ball JC (2001) Trifluoroacetic acid in ancient freshwater. Atmospheric Environment 35:2799–2801 45. Ohya T, Kudo N, Suzuki E, Kawashima Y (1998) Determination of perfluorinated carboxylic acids in biological samples by high-performance liquid chromatography. J Chromatogr B 720:1–7 46. Olsen GW, Burris JM, Mandel JH, Zobel LR (1999) Serum perfluorooctanesulfonate and hepatic and lipid clinical chemistry tests in fluorochemical production employees. J Occup Environ Med 41:799–806 47. Renner R (2001) Growing concern over perfluorinated chemicals. Environ Sci Technol pp 154A–160A 48. Richey DG, Driscoll CT, Likens GE (1997) Soil retention of trifluoroacetate. Environ Sci Technol 31:1723–1727 49. Rompp A, Klemm O, Fricke W, Frank H (2001) Haloacetates in fog and rain. Environ Sci Technol 35:1294–1298 50. Schröder HF (1991) Fluorine-containing surfactants – a further challenge for the environment? Part 1: anionic and cationic surfactants. Vom Wasser 77:277–290 51. Scientific Assessment of Stratospheric Ozone: 1989 Global Ozone Research and Monitoring Project, Vol II (Rep. No. 20, World Meteorological Organization, Geneva, 1989) 52. Scott BF,Alaee M (1998) Determination of haloacetic acids from aqueous samples collected from the Canadian environment using an in situ derivatization technique. Water Qual Res J Can 33:279–293 53. Scott BF, Mactavish D, Spencer C, Strachan WMJ, Muir DCG (2000) Haloacetic acids in Canadian lake waters and precipitation. Environ Sci Technol 34:4266–4272 54. Standley LJ, Bott TL (1998) Trifluoroacetate, an atmospheric breakdown product of hydrofluorocarbon refrigerants: Biomolecular fate in aquatic organisms. Environ Sci Technol 32:469–475 55. Taves DR (1966) Normal human serum fluoride concentrations. Nature 211:192–193 56. Taves DR (1968) Evidence that there are two forms of fluoride in human serum. Nature 217:1050–1051 57. Tromp TK, Ko MKW, Rodriguez JM, Sze ND (1995) Potential accumulation of a CFCreplacement degradation product in seasonal wetlands. Nature 376:327–330 58. Victor DG, MacDonald GJ (1999) A model for estimating future emissions of sulfur hexafluoride and perfluorocarbons. Climatic Change 42:633–662 59. Von Sydow LM, Grimvall AB, Boren HB, Laniewski K, Nielsen AT (2000) Natural background levels of trifluoroacetate in rain and snow. Environ Sci Technol 34:3115–3118 60. Weber J (2000) 3M to Pare Scotchgard Products. Business Week, pp 96–98 61. Wiegand C, Pflugmacher S, Giese M, Frank H, Steinberg C (2000) Uptake, toxicity, and effects on detoxification enzymes of atrazine and trifluoroacetate in embryos of zebrafish. Ecotox Environ Saf 45:122 –131 62. Wujcik CE, Zehavi D, Seiber JN (1998) Trifluoroacetic acid levels in 1994–1996 fog, rain, snow and surface waters from California and Nevada. Chemosphere 36:1233–1245 63. Wujcik CE, Cahill TM, Seiber JN (1998) Extraction and analysis of trifluoroacetic acid in environmental waters. Anal Chem 70:4074–4080 64. Wujcik CE, Cahill TM, Seiber JN (1999) Determination of trifluoroacetic acid in 1996–1997 precipitation and surface waters in California and Nevada. Environ Sci Technol 33:1747–1751 65. Zehavi D, Seiber JN (1996) An analytical method for trifluoroacetic acid in water and air samples using headspace gas chromatographic determination of the methyl ester. Anal Chem 68:3450–3459

CHAPTER 5

Naturally Occurring Organofluorines Gordon W. Gribble Department of Chemistry, Dartmouth College, Hanover, NH 03755, USA E-mail: [email protected] † This chapter is dedicated to the memory of Professor Richard E. Stoiber, 1911–2001, a Dartmouth College colleague and volcanologist who first discovered organic fluorine compounds in volcanic gases.

Although 3700 naturally occurring organohalogens are now known to exist, only relatively few contain fluorine. The presence of several fluoroalkanes in volcanic and other geothermal emissions is well documented, although exactly how these compounds are produced remains a mystery. Also unknown is the impact that these natural fluoroalkanes have on the global atmospheric budget compared to their anthropogenic counterparts, since the concentrations of the natural compounds vary widely depending on the source. The remarkable ability of a few plants to sequester and convert fluoride into the highly toxic fluoroacetate and other fluorocarboxylic acids is well recognized, and the mechanisms for their formation are becoming understood. Keywords. Organofluorine fluoroacetate, Fluoroacetic acid gifblaar fluorocitrate, Fluorinated fatty acids, Nucleocidin, 4-Fluorothreonine, Fluoroacetaldehyde, Fluorite, Hydrogen fluoride, Fluoroalkanes, Trifluoroacetic acid, Tetrafluoroethylene, Tetrafluoromethane

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

1

Introduction

2

Abiotic Production of Organic Fluorine Compounds

3

Biogenically Produced Organic Fluorine Compounds . . . . . . . 126

3.1 3.2

Range of Structures . . . . . . . . . . . . . . . . . . . . . . . . . 126 Biogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

4

Unresolved Issues and Future Developments . . . . . . . . . . . . 132

5

References

. . . . . . . 122

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

1 Introduction After several decades of skepticism from scientists and outright ridicule from environmental activists, the fact that nature produces organohalogen compounds is now well established and can no longer be ignored. Several reviews of naturally occurring organohalogens have appeared [6, 9, 14–22, 34, 59, 82, 90, 97].While only about a dozen organohalogen compounds were recognized The Handbook of Environmental Chemistry Vol. 3, Part N Organofluorines (ed. by A.H. Neilson) © Springer-Verlag Berlin Heidelberg 2002

122

G.W. Gribble

in the mid-1950s, the number today has grown to about 3650 as of December 2001 [23]. As summarized below most of these organohalogens contain chlorine or bromine (or both halogens). A far lesser number contain iodine or fluorine: – – – –

Organochlorines: 2150 Organobromines: 1850 Organoiodines: 95 Organofluorines: 30

Two quite distinct sources of natural organohalogen compounds exist: biogenic and abiogenic. Biogenic organohalogens are produced by bacteria, fungi, lichen, marine plants and animals, terrestrial plants, several insects, and some higher animals including humans. Abiogenic organohalogens are formed in, or released during, geothermal processes such as volcanoes, biomass fires, and other geological processes. Fluorine with a mean abundance of 950 ppm in the earth’s crust is the 13th most abundant element and occurs mainly as fluorite (calcium fluoride) (45%), with the remainder in the minerals fluoroapatite, cryolite, and to a minor extent in topaz [30]. These inorganic fluorides are not readily converted into organic fluorine except in the circumstances discussed in Sect. 2. By comparison, chlorine with a mean abundance of 130 ppm in the earth’s crust is the 26th most abundant element. For excellent previous reviews of naturally occurring organofluorine compounds, the reader is referred to reviews by Harper, O’Hagan and colleagues [32, 54, 65]. These articles are outstanding and treat natural organofluorine compounds in greater depth than in the present review.

2 Abiotic Production of Organic Fluorine Compounds Despite the fact that fluorine, as fluoride salts and hydrogen fluoride, was well known to be abundant on earth, the existence of natural fluoroalkanes remained unknown until relatively recently. Harnisch and Eisenhauer estimate that 5¥1021 g of calcium fluoride (fluorite) is present on earth [30]. For comparison the mass of the earth’s crust is 2.4¥1025 g. Interestingly, samples of the dark purple fluorite from Bavaria, called “Stinkspat” by local miners, when crushed exude the unmistakable smell of molecular fluorine! Chemists since the early 1800’s such as Wöhler, Morissan, Becquerel, and Schönbein have made this observation [30]. This molecular fluorine may originate from calcium fluoride via decay of radioactive uranium and thorium which are also present in this mineral. Kranz was the first to propose this theory to account for the presence of fluoroalkanes in fluorites [46] (see below). Volcanoes are a major source of hydrogen fluoride along with a myriad of other gases. Both large, but infrequent, volcanic eruptions and passively degassing volcanoes are major sources of stratospheric and tropospheric hydrogen fluoride, respectively. An estimate of 0.06–6¥1012 g/year of volcanic hydrogen fluoride has been given [87]. For example, the volcanoes Hekla in 1970 [67], El Chichón in 1982 [98], Guatemala

Naturally Occurring Organofluorines

123

in 1978 [7], and Kilauea [66] all have produced and continue to produce large quantities of hydrogen fluoride in addition to hydrogen chloride. It is estimated that Kilauea on Hawaii, which has been continuously erupting since 1983, produces 180 tons of hydrogen fluoride daily.A new technique,“Passive Infrared Spectroscopy”, has been developed for detecting gases in erupting volcanic gas plumes at a 17-km distance [47]. This study of the Popocatépetl volcano in Mexico over the period 1994–1997 has revealed that hydrogen fluoride is released to the extent of hundreds of tons per day. Hydrogen chloride is released on the order of thousands of tons per day. In contrast, sulfur tetrafluoride is only released to the extent of a few tons per day. It might be noted that Popocatépetl as of December 2000 was undergoing explosive eruptions. A related study of Mount Etna in 1997 revealed volcanic emission rates of 2.2 kg/s of hydrogen fluoride and 8.6 kg/s of hydrogen chloride, making Mount Etna the largest known point source of these two gases [10]. This latter study, which employed remote solar occultation spectroscopy, also summarized other volcanic eruptions and estimated the global emission rates of hydrogen fluoride from volcanoes and anthropogenic sources as 1.9 –190 and 15.8 kg/s, respectively. Interestingly, hydrogen fluoride has recently been detected in interstellar space [60]. This exciting observation parallels the earlier discovery of hydrogen chloride, the only two halogen-containing molecules to be identified in interstellar space thus far. Given the abundance of both hydrogen fluoride in volcanic gases and fluoridecontaining minerals in the earth’s crust, it perhaps should not be surprising that abiogenic fluoroalkanes are present in the environment. Stoiber et al. were the first to report in 1971 the presence of organofluorine compounds, including some CFCs, in the volcanic gases from the fumaroles of the Santiaguito volcano in Guatemala [83]. In addition to detecting methane, hydrogen fluoride, hydrogen chloride, and about 30 other organic compounds, these workers identified tetrafluoroethylene (1), hexafluoropropene (2), chlorodifluoromethane (3), chlorotrifluoroethylene (4), dichlorofluoromethane (5), trichlorofluoromethane (6), and 1,1,2-trichloro-1,2,2-trifluoroethane (7). The authors suggest that these compounds “probably resulted from reactions of hydrocarbons with hydrogen fluoride and hydrogen chloride, inorganic halides, or halide-containing minerals under high temperature conditions which prevailed at the source. The most likely sources of organic compounds are probably sediment or fossil soil layers beneath the volcano” [83]. Unfortunately, quantitative measurements were not carried out, but the organic compounds were separated and identified by gas chromatography-mass spectrometry.

124

G.W. Gribble

In extensive studies of the Kamchatka volcanoes on the Siberian peninsula, Isidorov and colleagues have also identified several organofluorines in these volcanic solfataric gases [36, 37, 39]. These workers reported 5, 6, dichlorodifluoromethane (8), and dimethyl difluorosilane (9). From one solfataric vent the concentration of these organofluorines exceeded background levels by 400 times. Isidorov estimates that 75% of the world’s 2000 active volcanoes have the requisite mineral composition and geological configuration to produce organofluorine compounds such as 1–9. He has proposed the set of reactions in Scheme 1 to account for the formation of fluoroalkanes [36, 37]. Methane, chloroform, and carbon tetrachloride are also present in these solfataric gases.

Scheme 1

These organofluorines have also been discovered in the gases from hydrothermal vents and thermal springs in the Kamchatka,Ashkhabad, and Tskhaltubo regions of the former Soviet Union [38, 40]. These organofluorines and numerous organochlorines, organobromines, and other organic compounds are released when certain rocks, shales, and minerals are crushed and processed in mining operations [6, 37, 40, 41]. In addition to 6 and 8, the new naturally occurring trifluoromethane (10) [40] and polyfluorinated propane 11 [6] were identified in these investigations. In his study of carbonaceous black shales from Central Asia, Buslaeva concludes that these haloalkanes could arise from the interaction of alkanes “with halogens in the rocks at the temperatures and pressures characteristic for geological processes” [6]. Isidorov suggests that the world’s mining industry, which processes hundreds of millions of tons of ore for the production of potassium salts alone, is a significant source of atmospheric halocarbons including 10–15 kilotons of chloroform and 0.1–0.15 kilotons of trichlorofluoromethane (6) and carbon tetrachloride [4].

Naturally Occurring Organofluorines

125

It has been suggested that trifluoroacetic acid (12) has an as yet unidentified natural source since the concentration of this compound in rainwater is higher than the levels predicted from the degradation of anthropogenic compounds [43]. Trifluoroacetate is discussed in Chap. 5 by Scott Mabury. Three recent studies of the composition of volcanic emissions have verified the earlier work of Stoiber and Isidorov. Thus, Wahrenberger et al. have identified both isomers of tetrachlorodifluoroethane, 1,2-difluoro-1,1,2,2-tetrachloroethane (13) and 1,1-difluoro-1,2,2,2-tetrachloroethane (14) from the Italy volcano Vulcano [94]. Another study by this same group has identified nine halocarbons from Vulcano and Kudriavy (Kuriles, Russia) including 1-chloro-1,1-difluoroethane (15), which is “HCFC-142b” [95]. The flux of 15 from these volcanoes is at least three orders of magnitude above tropospheric background concentrations.

Jordan et al. have examined the fumarole and lava gas samples from four volcanoes (Kuju, Satsuma Iwojima, Mount Etna, and Vulcano) [44]. They identified more than 300 organic compounds, including 135 that contain halogen. Of these, five are fluorinated: trichlorofluoromethane (6), a trifluoropropene (16), fluorobenzene (17), a tetrafluorobenzene (18), and a chlorofluorobenzene (19). In the cases of 16, 18, and 19 the exact structures were not determined. The authors conclude that volcanic CFC 6 contributes a negligible amount compared with the anthropogenic sources of this compound.

Tetrafluoroethylene (1), which was previously identified in the Santiaguito volcanic gases and which is the anthropogenic precursor to the commercial polymer Teflon, has been discovered in natural fluorites along with tetrafluoromethane (20) [29, 30]. Sulfur hexafluoride is also present in these minerals. Tetrafluoromethane has a large anthropogenic source from the production of aluminum (15,000 tons/year) [84], but the authors estimate that approximately half of the current atmospheric burden of tetrafluoromethane has accumulated from natural sources since the lifetime of this compound is thought to be 50,000 years! Natural gas also contains significant traces of tetrafluoromethane that probably result from weak radiogenic sources in the lithosphere. This observed cold degassing from the Earth’s crust is estimated to be 0.1–10 tons/year, which is negligible compared with current anthropogenic emissions [29]. Earlier Kranz sug-

126

G.W. Gribble

gested that these fluorite-containing organofluorines arise from reactions of molecular fluorine with encased organic material [46]. The formation of fluorine might occur as a result of the alpha radiation from the decay of uranium and thorium present in the fluorites [29, 46]. Harnisch and Eisenhauer have examined the gases from several volcanoes (Mauna Loa, Iilewa, Mount Erebus, Mount Etna, Vulcano, Mount Kuju, and Satsuma Iwojima) and find that they are not significant sources of either tetrafluoromethane or sulfur hexafluoride [30]. Although studies of biomass burning in North America and Canada have revealed the presence of high amounts of dichlorodifluoromethane (8) and other CFCs, the authors conclude that these compounds are solely anthropogenic [5, 33], since “dichlorodifluoromethane cannot be produced by fires” [33]. Hoekstra has recently disputed this conclusion on two grounds [35]. To quote Hoekstra, “The calculated global emission rate [from biomass burning] equals about half of the estimated total annual emission of dichlorodifluoromethane and the deposition of dichlorodifluoromethane will be evenly distributed around the globe because of its very long lifetime, while the area which is annually burned is relatively small” [35]. In other words, significantly more dichlorodifluoromethane is present in these forest fires than could be accounted for from the revolatization of preexisting dichlorodifluoromethane in the fuel. Both bromotrifluoromethane and bromochlorodifluoromethane have been identified in the atmosphere [68], but no natural source has been identified for these two compounds and they are presumed to be of anthropogenic origin. Although the concentration of fluoride in seawater is only 1.4 ppm, the sponge Halichondria moorei contains potassium fluorosilicate (K2SiF6) to the extent of up to 11.5% (dry weight) [12]. Whereas this animal is able to sequester and concentrate fluoride, another sponge species from the same taxonomic order and in a nearby location contains no fluoride or fluorosilicate. Interestingly, potassium fluorosilicate is a potent inflammatory agent [12].

3 Biogenically Produced Organic Fluorine Compounds 3.1 Range of Structures

Nearly all of the known biogenic organofluorine compounds are fluorinated carboxylic acids. Fluoroacetic acid (21) (hereafter fluoroacetate, which is the form at physiological pH values) was the first such compound to be isolated and characterized nearly 60 years ago (see below), although methyl fluoroacetate (22) was synthesized in 1896 [85, 86]. Fluoroacetate itself was also first prepared in 1896 [79, 85].

Naturally Occurring Organofluorines

127

Fluoroacetate was first isolated by Marais from the South African plant “gifblaar”, Dichapetalum cymosum, in 1943 [48, 49]. This plant was long recognized as being highly toxic to cattle and sheep in South Africa [13, 32, 54, 65]. For example, less than an ounce of the leaves of this plant will kill a sheep and one half a leaf was fatal to an ox [80]. Several other Dichapetalum species contain fluoroacetate, including D. stuhlmanii [25, 64], D. toxicarium [92], D. heudelotii [93], D. michelsonii [93], D. guineense [93], D. venenatum [93], D. braunii [64, 93], D. macrocarpum [93], D. ruhlandii [93], D. barteri [62], and D. edule [64]. Although plants in general contain only 0.1–10 ppm fluoride, D. toxicarium can sequester fluoride up to levels of 450 ppm in young leaves from an initial soil concentration of only 1–10 ppm [92]. The Tanzanian D. braunii can attain fluoroacetate levels of 7200 ppm in young leaves and 8000 ppm in the seeds [64]. Several analytical methods have been developed to assay plants and other biological materials for fluoroacetate. For example, 19F Nuclear Magnetic Resonance Spectroscopy (NMR) can detect fluoroacetate at levels of 4 ppm in plants [3], and High Pressure Liquid Chromatography (HPLC) can detect this compound at levels down to 0.1 ppm [55]. Several other tropical and semitropical countries have plants containing fluoroacetate. The Australian Acacia georginae, widespread in Northwest Queensland and the Northern Territory covering 28,000 square miles, has caused serious sheep and cattle losses due to the presence of natural fluoroacetate [2, 58, 63]. The highly toxic Brazilian Palicourea marcgravii, known as “rat weed”, contains fluoroacetate [8], as does the Australian Gastrolobium grandiflorum, which killed 2000 sheep in a single episode [50]. A few other fluoroacetate-containing plants are Oxylobium parviforum [24, 25], Spondianthus preussii [45], and Cyamopsis tetragonolobus, known as “Guar Gum” [91]. The latter plant is used medicinally in Finland following extraction of the fluoroacetate that is present in the gum at levels of 0.07–1.4 ppm [91]. Interestingly, none of these plants grows in high fluoride soil and there does not appear to be a strong correlation between soil fluoride and plant fluoroacetate [24, 25]. For a more extensive discussion of fluoroacetate in plants, the reader is directed to the Harper and O’Hagan review [32]. As we will see in the section on biosynthesis, the actual toxic metabolite may be, at least in part, fluorocitrate (23). Fluorocitrate is also present in commercial tea leaves (<30 ppm) and oatmeal (<62 ppm) [4, 75]. The latter also contains some fluoroacetate. These levels of fluorocitrate are apparently below toxic levels. The natural stereoisomer of fluorocitrate, and the only one to be toxic of the four stereoisomers, has the 2R,3R configuration as shown.

128

G.W. Gribble

In their studies with Acacia georginae, Peters and Shorthouse observed fluoroacetone (24) in the volatiles from this plant [73, 74, 76]. However, O’Hagan and Harper have questioned this result and have suggested that this volatile organofluorine may in fact be fluoroacetaldehyde [65]. The alleged fluoroacetone was isolated as the 2,4-dinitrophenylhydrazone derivative. The melting points of these two 2,4-dinitrophenylhydrazones are sufficiently different to make a distinction possible. Interestingly, several of the other possible monohalo acetones and acetaldehydes have been isolated from red algae [18]. Seeds of the Sierra Leone shrub Dichapetalum toxicarium, also known as ratsbane, contain several fluorinated fatty acids. These have been characterized as 18fluorooleic acid (25) [69–72], 16-fluoropalmitic acid (26) [96], 10-fluorocapric acid (27) [96], 14-fluoromyristic acid (28) [96], and threo-18-fluoro-9,10-dihydroxystearic acid (29) [31]. As we will see later, these even-numbered carbon chain fluorinated fatty acids are equally toxic as fluoroacetate, since they are metabolized by b-oxidation to fluoroacetate in accord with the theory of fatty acid oxidation.

A more recent examination of the seed oil from D. toxicarium by Hamilton and Harper has uncovered six additional novel 16-fluorinated fatty acids [27]. These are 16-fluoropalmitoleic acid (30), 18-fluorostearic acid (31), 18-fluorolinoleic acid (32), 20-fluoroarachidic acid (33), 20-fluoroeicosenoic acid (34), and 18-fluoro-9,10-epoxystearic acid (35). The latter compound is strongly suspected of being present but not yet confirmed. The major fluoro fatty acids in the seed oil of D. toxicarium are 18-fluorooleic acid (25) (75%) and 16-fluoropalmitic acid (26) (15%). This study did not find 10-fluorodecanoic acid (27) or 14-fluoromyristic acid (28), in contrast to the earlier work by Ward and co-workers [96].

Naturally Occurring Organofluorines

129

Although terrestrial plants provide most of the known natural organofluorines (see above), Streptomyces bacteria have furnished a few such compounds. The antibiotic nucleocidin (36) was isolated in 1957 from cultures of Streptomyces calvus [89], and its correct structure was finally determined in 1968 [56, 81]. Total synthesis confirmed the structure beyond all doubt [42]. This novel fluorinated nucleoside appears to be naturally derived as fluoride salts were not added to the culture medium. Unfortunately, recent attempts to reisolate nucleocidin have been unsuccessful [51]. An excellent discussion of the discovery and structure elucidation of nucleocidin is given by Harper and O’Hagan D [32].

Not only does Streptomyces cattleya produce fluoroacetate but also the novel amino acid 4-fluorothreonine (37) [78]. This bacterium produces the antibiotic thienamycin. The structure of 37 as the 2S,3S isomer has been confirmed by total synthesis [1]. As will be discussed in the section on biosynthesis, Harper and co-workers have presented strong evidence that fluoroacetaldehyde (38) is the intermediate biosynthetic precursor of both fluoroacetate and 4-fluorothreonine in S. cattleya [57].

130

G.W. Gribble

3.2 Biogenesis

Several biosyntheses have been proposed for the formation of fluoroacetate and the other natural fluorocarboxylic acids. These are discussed in detail by Harper and O’Hagan [32]. The first of these involves the well-known pyridoxal phosphate mechanism as proposed by Mead and Segal [52]. Although attempts thus far to support this mechanism have been unsuccessful [53], Harper and O’Hagan point out that these negative results do not necessarily invalidate this pathway (Eq. 1) [32].

(eqn. 1)

A second possible pathway involves fluoride displacing phosphate from phosphoglycolate in an SN2 reaction (Eq. 2). Such nucleophilic displacement reactions involving a-substituted carbonyl substrates are much faster than those involving simple alkyl substrates [84]. For example, a-chloroacetophenone is a powerful lachrymator (“MACE”) as are the haloacetones, except fluoroacetone. Thus far, evidence to support this mechanism has been negative [11].

(eqn. 2)

Other pathways that have been proposed are the generation of fluorophosphate as a fluoride carrier and fluorodecarboxylation of malonic acid by a fluoroperoxidase enzyme. The reader is referred to the review by Harper and O’Hagan for further discussion of these proposed pathways [32]. The research groups of Harper et al. [26, 28, 57, 61, 77] and Tamura et al. [88] have independently studied the biosynthesis of fluoroacetate and 4-fluorothreonine by Streptomyces cattleya. This research follows earlier biosynthetic work by Sanada et al. on the same organism [78]. The two most recent papers by Harper et al. summarize the current state of knowledge regarding the biosynthesis of these fluorinated metabolites [28, 57]. Thus, glucose, glycerol, serine, b-hydroxypyruvate, and glycine are all incorporated into fluoroacetate and 4-fluorothreonine to var-

Naturally Occurring Organofluorines

131

ious degrees. The most recent evidence indicates that fluoroacetaldehyde (38) is converted to both fluoroacetate and 4-fluorothreonine in Streptomyces cattleya [57].Additional evidence with deuterated glycerols indicates that a common fluorinating enzyme for both of these compounds is present in this organism [28, 61]. Based on these results, a plausible biosynthetic pathway is illustrated in Scheme 2 [65], although it may not be the exclusive one. The longer chain carboxylic acids 25–35 are presumably biosynthesized from fluoroacetic acid via fluoroacetyl CoA and malonyl CoA by the usual fatty acid biosynthesis pathway [32]. Interestingly, these long-chain fluorinated fatty acids

Scheme 2

132

G.W. Gribble

are equally toxic as fluoroacetic acid since they appear to be degraded to the latter by the fatty acid degradation pathway [13]. The toxicity of fluoroacetic acid is discussed in detail by others [13, 32, 54, 65].

4 Unresolved Issues and Future Developments While there is no disputing the presence of organic fluorine compounds in volcanic and hydrothermal emissions, the factors that determine the concentrations of these chemicals and the mechanisms of their formation remain largely not understood. Are these compounds formed in reactions between a carbon source and hydrogen fluoride or a fluoride mineral such as fluorite? What is the source of carbon? What is the source of chlorine in the chlorofluorocarbons? Lightning induced forest and brush fires rage regularly across our planet. Is biomass burning a significant source of organic fluorine compounds? What are the natural sources and quantities of tetrafluoromethane? How significant are the reactions shown in Scheme 1 for the formation of the chlorofluoroalkanes? What are the natural sources of trifluoroacetate? These and other questions regarding naturally occurring abiotic organic fluorine compounds remain unanswered. The presence of fluoroacetate, the most toxic low molecular weight natural compound known in numerous plants, is extraordinary.Although much is known about the biogenesis of fluoroacetate and the other natural fluorinated carboxylic acids, the full story has yet to be told. Why are some animals relatively impervious to the toxicity of fluoroacetate? What is the mechanism of toxicity? Can the fluorinecontaining antibiotic nucleocidin be reisolated? Will other organic fluorine natural products be discovered along with the growing number of other naturally occurring organohalogen compounds?

5 References 1. Amin MR, Harper DB, Moloney JM, Murphy CD, Howard JAK, O’Hagan D (1997) A short highly stereoselective synthesis of the fluorinated natural product (2S,3S)-4-fluorothreonine. Chem Commun 1471–1472 2. Aplin TEH (1967) Poison plants of western Australia. The toxic species of the genera Gastrolobium and Oxylobium. J Agric West Aust 8:42–52 3. Baron ML, Bothroyd CM, Rogers GI, Staffa A, Rae ID (1987) Detection and measurement of fluoroacetate in plant extracts by 19F NMR. Phytochem 26:2293–2295 4. Bennett LW, Miller GW, Yu MH, Lynn RI (1983) Production of fluoroacetate by callus tissue from leaves of Acacia georginae. Fluoride 16:111–117 5. Blake DR, Smith TW Jr, Chen TY,Whipple WJ, Rowland FS (1994) Effects of biomass burning on summertime nonmethane hydrocarbon concentrations in the Canadian wetlands. J Geophys Res 99:1699–1719 6. Buslaeva EY (1994) Halogen-substituted hydrocarbons in carbonaceous black shales. Geokhimiya 1130–1131 7. Cadle RD, Lazrus AL, Huebert BJ, Heidt LE, Rose WI Jr, Woods DC, Chuan RL, Stoiber RE, Smith DB, Zielinski RA (1979) Atmospheric implications of studies of Central American volcanic eruption clouds. J Geophys Res 84:6961–6968 8. de Oliveira MM (1963) Chromatographic isolation of monofluoroacetic acid from Palicourea marcgravii St Hil. Experientia 19:586–587

Naturally Occurring Organofluorines

133

9. Engvild KC (1986) Chlorine-containing natural products in higher plants. Phytochem 25:781–791 10. Francis P, Burton MR, Oppenheimer C (1998) Remote gas measurements of volcanic gas compositions by solar occultation spectroscopy. Nature 396:567–570 11. Goldman P, Milne GWA (1966) Carbon-fluorine bond cleavage II. Studies on the mechanism of the defluorination of fluoroacetate. J Biol Chem 241:5557–5559 12. Gregson RP, Baldo BA, Thomas PG, Quinn RJ, Bergquist PR, Stephens JF, Horne AR (1979) Fluorine is a major constituent of the marine sponge Halichondria moorei. Science 206:1108–1109 13. Gribble GW (1973) Fluoroacetate toxicity. J Chem Ed 50:460–462 14. Gribble GW (1992) Naturally occurring organohalogen compounds: a survey. J Nat Prod 55:1353–1395 15. Gribble GW (1994) Natural organohalogens – many more than you think! J Chem Ed 71:907–911 16. Gribble GW (1994) The natural production of chlorinated compounds. Environ Sci Technol 28:310A–319A 17. Gribble GW (1996) The diversity of natural organochlorines in living systems. Pure Appl Chem 68:1699–1712 18. Gribble GW (1996) Naturally occurring organohalogen compounds: a comprehensive survey. Prog Chem Org Nat Prod 68:1–498 19. Gribble GW (1998) Naturally occurring organohalogen compounds. Acc Chem Res 31:141–152 20. Gribble GW (1998) Chlorinated compounds in the biosphere, natural formation. In: Meyers RA (ed) Encyclopedia of environmental analysis and remediation. Wiley, New York, pp 972– 1035 21. Gribble GW (1999) The diversity of naturally occurring organobromine compounds. Chem Soc Rev 28:335–346 22. Gribble GW (2000) The natural production of organobromine compounds. Environ Sci Pollut Res 7:37–49 23. Gribble GW (2002) Unpublished compilation 24. Hall RJ (1972) The distribution of organic fluorine in some toxic tropical plants. New Phytol 71:855–871 25. Hall RJ, Cain RB (1972) Organic fluorine in tropical soils. New Phytol 71:839–853 26. Hamilton JTG, Amin MR, Harper DB, O’Hagan D (1997) Biosynthesis of fluoroacetate and 4-fluorothreonine by Streptomyces cattleya. Glycine and pyruvate as precursors. Chem Commun 797–798 27. Hamilton JTG, Harper DB (1997) Fluoro fatty acids in seed oil of Dichapetalum toxicarium. Phytochem 44:1129–1132 28. Hamilton JTG, Murphy CD,Amin MR, O’Hagan D, Harper DB (1998) Exploring the biosynthetic origin of fluoroacetate and 4-fluorothreonine in Streptomyces cattleya. J Chem Soc Perkin Trans 1 759–767 29. Harnisch J, Borchers R, Fabian P, Gäggeler HW, Schotterer U (1996) Effect of natural tetrafluoromethane. Nature 384:32 30. Harnisch J, Eisenhauer A (1998) Natural CF4 and SF6, on earth. Geophys Res Lett 25:2401–2404 31. Harper DB, Hamilton JTG, O’Hagan D (1990) Identification of threo-18-fluoro-9,10-dihydrostearic acid: a novel w-fluorinated fatty acid from Dichapetalum toxicarium seeds. Tetrahedron Lett 31:7661–7662 32. Harper DB, O’Hagan D (1994) The fluorinated natural products. Nat Prod Rep 11:123–133 33. Hegg DA, Radke LF, Hobbs PV, Rasmussen RA, Riggan PJ (1990) Emissions of some trace gases from biomass fires. J Geophys Res 95:5669–5675 34. Hoekstra EJ, De Leer EWB (1995) Organohalogens: the natural alternatives. Chem Brit 127–131 35. Hoekstra EJ (1999) On the natural formation of chlorinated organic compounds in soil. Vrije University, Amsterdam

134

G.W. Gribble

36. Isidorov VA, Zenkevich IG, Ioffe BV (1990) Volatile organic compounds in solfataric gases. J Atmos Chem 10:329–340 37. Isidorov VA (1990) Organic chemistry of the earth’s atmosphere. Springer, Berlin Heidelberg New York 38. Isidorov VA, Prilepskii EB, Fedorov YN (1991) Organic components of thermal source gases in seismically active regions. Dokl Akad Nauk SSSR (Geophysical) 319:1106–1109 39. Isidorov VA, Zenkevich IG, Karpov GA (1991) Volatile organic compounds in steam-gas discharges from some volcanoes and hydrothermal systems of Kamchatka.Vulkanol Seismol 19–25 40. Isidorov VA, Povarov VG, Prilepsky EB (1993) Geological sources of volatile organic components in regions of seismic and volcanic activity. J Ecol Chem 19–25 41. Isidorov VA, Prilepsky EB, Povarov VG (1993) Photochemically and optically active components of minerals and gas emissions of mining plants. J Ecol Chem 201–207 42. Jenkins ID,Verheyden JPH, Moffatt JG (1976) 4¢-Substituted nucleosides. 2. Synthesis of the nucleoside antibiotic nucleocidin. J Am Chem Soc 98:3346–3357 43. Jordan A, Frank H (1999) Trifluoroacetate in the environment. Evidence for sources other than HFC/HCFCs. Environ Sci Technol 33:522–527 44. Jordan A, Harnisch J, Borchers R, LeGuerm F, Shinohara H (2000) Volcanogenic halocarbons. Environ Sci Technol 34:1122–1124 45. Kamgue RT, Sylla O, Pousset JL, Laurens A, Brunet JC, Sere A (1979) Isolation and characterization of toxic principles from Spondianthus preussii var glaber Engler. Plant Med Phytother 13:252–259 46. Kranz R (1966) Organische Fluor-Verbindungen in den Gas-einschlüssen der Wölsendorfer Flusspäte. Naturwissenschaften 53:593–600 47. Love SP, Goff F, Counce D, Siebe C, Delgado H (1998) Passive infrared spectroscopy of the eruption plume at Popocatépetl volcano, Mexico. Nature 396:563–567 48. Marais JSC (1943) The isolation of the toxic principle “K cymonate” from “Gifblaar”, Dichapetalum cymosum (Hook) Engl. Onderstepoort J Vet Sci Animal Ind 18:203–206 49. Marais JSC (1944) Monofluoroacetic acid, the toxic principle of “gifblaar” Dichapetalum cymosum (Hook) Engl. Onderstepoort J Vet Sci Animal Ind 20:67–73 50. McEwan T (1964) Isolation and identification of the toxic principle of Gastrolobium grandiflorum. Nature 201:827 51. McGuire AR, Meng W, Roberts SM, Willetts AJ (1993) Synthetic approaches toward nucleocidin and selected analogues; anti-HIV activity in 4¢-fluorinated nucleoside derivatives. J Chem Soc Perkin Trans 1:1795–1808 52. Mead RJ, W Segal (1972) Fluoroacetic acid biosynthesis proposed mechanism. Aust J Biol Sci 25:327–333 53. Mead RJ, Segal W (1973) Formation of b-cyanoalanine and pyruvate by Acacia georginae. Phytochem 12:1977–1981 54. Meyer M, O’Hagan D (1992) Rare fluorinated natural products. Chem Brit 785–788 55. Meyer JJM, Grobbelaar N (1990) The determination, uptake and transport of fluoroacetate in Dichapetalum cymosum. J Plant Physiol 135:546–549 56. Morton GO, Lancaster JE, Van Lear GE, Fulmor W, Meyer WE (1969) The structure of nucleocidin. III (a new structure). J Am Chem Soc 91:1535–1537 57. Moss SJ, Murphy CD, Hamilton JTG, McRoberts WC, O’Hagan D, Schaffrath C, Harper DB (2000) Fluoroacetaldehyde: a precursor of both fluoroacetate and 4-fluorothreonine in Streptomyces cattleya. Chem Commun 2281–2282 58. Murray LR, McConnell JD, Whittem JH (1961) Suspected presence of fluoroacetate in Acacia georginae. Aust J Sci 24:41–42 59. Naumann K (1993) Chlorchemie der Natur. Chem Zeit 27:33–41 60. Neufield DA, Zmuidzinas J, Schilke P, Phillips TG (1997) Discovery of interstellar hydrogen fluoride. Astrophys J 488:L141–L144 61. Nieschalk J, Hamilton JTG, CD Murphy, Harper DB, O’Hagan D (1997) Biosynthesis of fluoroacetate and 4-fluorothreonine by Streptomyces cattleya. The stereochemical processing of glycerol. Chem Commun 799–800

Naturally Occurring Organofluorines

135

62. Nwude N, Parsons LE, Adaudi AO (1977) Acute toxicity of the leaves and extracts of Dichapetalum barteri (Engl) in mice, rabbits and goats. Toxicol 7:23–29 63. Oelrichs PB, McEwan T (1961) Isolation of the toxic principle in Acacia georginae. Nature 190:808–809 64. O’Hagan D, Perry R, Lock JM, Meyer JJM, Dasaradhi L, Hamilton JTG, Harper DB (1993) High levels of monofluoroacetate in Dichapetalum braunii. Phytochem 33:1043–1045 65. O’Hagan D, Harper DB (1999) Fluorine-containing natural products. J Fluorine Chem 100 :127–133 66. Olmez I, Finnegan DL, Zoller WH (1986) Iridium emissions from Kilauea volcano. J Geophys Res 91:653–663 67. Oskarsson N (1980) The interaction between volcanic gases and tephra: fluorine adhering to tephra of the 1970 Hekla eruption. J Volcanol Geother Res 8:251–256 68. Penkett SA, Jones BMR, Rycroft MJ, Simmons DA (1985) An interhemispheric comparison of the concentrations of bromine compounds in the atmosphere. Nature 318:550–553 69. Peters RA, Hall RJ (1959) Further observations upon the toxic principle of Dichapetalum toxicarium. Biochem Pharmacol 2:25–36 70. Peters RA, Wakelin RW, Martin AJP, Webb J, Birks FT (1959) Observations upon the toxic principle in the seeds of Dichapetalum toxicarium. Biochem J 71:245–248 71. Peters RA, Hall RJ, Ward PFV, Sheppard N (1960) The chemical nature of the toxic compounds containing fluorine in the seeds of Dichapetalum toxicarium. Biochem J 77: 17–23 72. Peters RA, Hall RJ (1960) Fluorine compounds in nature; the distribution of carbon-fluorine compounds in some species of Dichapetalum. Nature 187:573–575 73. Peters RA, Shorthouse M (1967) Observations on the metabolism of fluoride in Acacia georginae and some other plants. Nature 216:80–81 74. Peters RA, Shorthouse M (1971) Identification of a volatile constituent formed by homogenates of Acacia georginae exposed to fluoride. Nature 231:123–124 75. Peters RA, Shorthouse M (1972) Fluorocitrate in plants and foodstuffs. Phytochem 11:1337–1338 76. Peters RA (1973) Keynote and historical perspective of organic fluorides in plants. Fluoride 6:189–194 77. Reid KA, Hamilton JTG, Bowden RD, O’Hagan D, Dasaradhi L,Amin MR, Harper DB (1995) Biosynthesis of fluorinated secondary metabolites by Streptomyces cattleya. Microbiology 141:1385–1393 78. Sanada M, Miyano T, Iwadare S, Williamson JM, Arison BH, Smith JL, Douglas AW, Liesch JM, Inamine E (1986) Biosynthesis of fluorothreonine and fluoroacetic acid by the thienamycin producer, Streptomyces cattleya. J Antibiot 39:259–265 79. Saunders BC, Stacey GJ (1948) Toxic fluorine compounds containing the C-F link. Part 1. Methyl fluoroacetate and related compounds. J Chem Soc 1773–1779 80. Saunders BC (1957) Some aspects of the chemistry and toxic action of organic compounds containing phosphorus and fluorine. University Press, Cambridge, pp 114–170 81. Schuman DA, Robins RK, Robins MJ (1969) The synthesis of adenine 5¢-O-sulfamoyl nucleosides related to nucleocidin. J Am Chem Soc 91:3391–3392 82. Siuda JF, DeBernardis JF (1973) Naturally occurring halogenated organic compounds. Lloydia 36:107–143 83. Stoiber RE, Leggett DC, Jenkins TF, Murrmann RP, Rose WI Jr (1971) Organic compounds in volcanic gas from Santiaguito volcano, Guatemala. Geol Soc Amer Bull 82:2299–2302 84. Streitwieser A Jr (1962) Solvolytic displacement reactions. McGraw-Hill, New York, pp 28–29 85. Swans F (1896) Sur l’acide fluoracétique. Bull Acad Roy Belg 31:675–688 86. Swarts F (1896) Sur l’acide fluoracétique. Bull Soc Chim 15:1134 87. Symonds RB, Rose WI, Reed MH (1988) Contribution of Cl- and F-bearing gases to the atmosphere by volcanoes. Nature 334:415–418 88. Tamura T, Wada M, Esaki N, Soda K (1995) Synthesis of fluoroacetate from fluoride, glycerol, and b-hydroxypyruvate by Streptomyces cattleya. J Bacteriol 177:2265–2269

136

G.W. Gribble

89. Thomas SO, Singleton VL, Lowery JA, Sharpe RW, Pruess LM, Porter JN, Mowat JH, Bohonos N (1957) Nucleocidin, a new antibiotic with activity against trypanosomes.Antibiotics Ann 1956/1957:716–721 90. Thomson RH (1978) Halogenated metabolites from marine animals and plants. J Indian Chem Soc 55:1209–1215 91. Vartiainen T, Gynther J (1984) Fluoroacetic acid in Guar Gum. Food Chem Toxicol 22:307–308 92. Vickery B, Vickery ML (1972) Fluoride metabolism in Dichapetalum toxicarium. Phytochem 11:1905–1909 93. Vickery B,Vickery ML,Ashu JT (1973) Analysis of plants for fluoroacetic acids. Phytochem 12:145–147 94. Wahrenberger C, Dietrich V, Seward TM (1996) CFC’s: Natural background emission by volcanoes? EOS Transactions 77:F804 95. Wahrenberger C, Seward TM, Dietrich V (1998) Halocarbon compounds in high temperature volcanic gases. International Volcanological Congress, Cape Town, South Africa 96. Ward PFV, Hall RJ, Peters RA (1964) Fluoro-fatty acids in the seeds of Dichapetalum toxicarium. Nature 201:611–612 97. Winterton N (2000) Chlorine: the only green element – towards a wider acceptance of its role in natural cycles. Green Chem 2:173 – 225 98. Woods DC, Chuan RL, Rose WI (1985) Halite particles injected into the stratosphere by the 1982 El Chichón eruption. Science 230 :170 –172

CHAPTER 6

Degradation and Transformation of Organic Fluorine Compounds Alasdair H. Neilson 1 · Ann-Sofie Allard 2 1 2

Swedish Environmental Research Institute Limited, Stockholm, Sweden E-mail: [email protected] Swedish Environmental Research Institute Limited, Stockholm, Sweden E-mail: [email protected]

An overview is given of the degradation and transformation of organic fluorine compounds by aerobic and anaerobic bacteria, yeasts, and fungi. Representatives of the major structural types of aliphatic, aromatic carbocyclic and heterocyclic compounds are included and a mechanistic view of the enzymology has been given where possible. Potential applications of organic fluorine compounds in biotechnology and the metabolism of organic fluorine compounds in higher organisms are briefly discussed. The use of fluorinated substrates to elucidate the mechanism of irreversible enzyme inactivation is illustrated for both microbial and a few mammalian enzymes.A number of abiotic systems including chemical and photochemical reactions are noted and compared with biochemical pathways. Throughout, a comparison is made with chlorinated analogues, and this reveals broadly comparable pathways for biodegradation and biotransformation. Keywords. Biodegradation and biotransformation, Bacteria, Yeasts, Fungi, Higher organisms, Photometabolism, Aliphatic compounds, Aromatic compounds, Enzyme inhibition

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

1

Introduction

2

Aliphatic Compounds

2.1 2.1.1 2.1.2 2.2 2.2.1 2.2.2 2.2.3 2.3

Alkanes and Alkenes . . . . . . . . . . . . . . . . . Fluorohydrocarbons . . . . . . . . . . . . . . . . . Chlorofluorocarbons and Hydrochlorofluorocarbons Carboxylic Acids and Related Compounds . . . . . Fluoroacetate . . . . . . . . . . . . . . . . . . . . . Fluorofumarates . . . . . . . . . . . . . . . . . . . Di- and Trifluoroacetate . . . . . . . . . . . . . . . Perfluoroalkyl Sulfonates . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

142 142 143 146 146 147 147 148

3 3.1 3.1.1 3.1.2 3.2 3.2.1 3.2.2

Aromatic Compounds . . . . . . . . Introduction . . . . . . . . . . . . . Aerobic Conditions . . . . . . . . . . Anaerobic Conditions . . . . . . . . Carbocyclic Aromatic Hydrocarbons Aerobic Organisms . . . . . . . . . . Anaerobic Reactions . . . . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

148 148 149 150 152 152 154

. . . . . . . . . . . . . . . . . . . . . . . . 141

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

The Handbook of Environmental Chemistry Vol. 3, Part N Organofluorines (ed. by A.H. Neilson) © Springer-Verlag Berlin Heidelberg 2002

138

A.H. Neilson · A.-S. Allard

3.3 3.4 3.4.1 3.4.2 3.5 3.5.1 3.5.2 3.6 3.7

Heterocyclic Aromatic Compounds . . . . . . . . . . . . . Benzoates . . . . . . . . . . . . . . . . . . . . . . . . . . . Aerobic Degradation . . . . . . . . . . . . . . . . . . . . . Degradation Under Denitrifying and Anaerobic Conditions Fluorinated Phenolic Compounds . . . . . . . . . . . . . . Aerobic Conditions . . . . . . . . . . . . . . . . . . . . . . Anaerobic Conditions . . . . . . . . . . . . . . . . . . . . Fluorinated Anilines . . . . . . . . . . . . . . . . . . . . . Aromatic Trifluoromethyl Compounds . . . . . . . . . . .

4

General Metabolism and Application of Fluorinated Compounds

4.1 4.2 4.2.1 4.2.2 4.2.3 4.3 4.4

Miscellaneous Reactions . . . . . . . Inhibitors and Mutagenizing Agents . Inhibitors . . . . . . . . . . . . . . . Mutagenizing Agents . . . . . . . . . Fluorine as Label in Biosynthesis . . Biotechnology . . . . . . . . . . . . . Non-Microbial Biochemical Reactions

5

Enzyme Inhibition by Fluorinated Analogue Substrates . . . . . . 178

5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9

Introduction . . . . . . . . . . . . . . Alanine Racemase . . . . . . . . . . . Ribonucleoside Diphosphate Reductase Ornithine Decarboxylase . . . . . . . . DNA-Cytosine Methyltransferase . . . Thymidylate Synthase . . . . . . . . . Enolpyruvyl Transferase . . . . . . . . Adenosyl-L-Homocysteine Hydrolase . Estrogen Synthetase . . . . . . . . . .

6

Abiotic Transformations . . . . . . . . . . . . . . . . . . . . . . . 185

6.1 6.2

Chemical Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . 186 Photochemical Reactions . . . . . . . . . . . . . . . . . . . . . . . 186

7

Concluding Comments . . . . . . . . . . . . . . . . . . . . . . . . 189

8

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189

. . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . .

. . . . . . . . .

. . . . . . . . .

155 157 157 164 165 165 171 172 172 174

. . . . . . .

. . . . . . . . .

174 175 175 176 176 176 177

178 180 181 181 182 182 183 184 185

1 Introduction This chapter discusses the biochemical interaction of organic fluorine compounds with microorganisms. In conventional terms, these result in the biodegradation or biotransformation of the substrate.A more recent aspect is the determinative role of microbial metabolites – some of which are intermediates. In addition, fluorinated analogue substrates may react irreversibly with critical

Degradation and Transformation of Organic Fluorine Compounds

139

biosynthetic enzymes and thereby inhibit further cell growth. The biodegradation of phosphorofluoridates and their mode of action are discussed by DeFrank and White in Chap. 10. In a wider context, attention is directed to chapters on toxicological aspects: Chap. 7 by DePierre on peroxisome synthesis, and Chap. 9 by Cavalieri on the application of fluorinated aromatic substrates to probe the mechanism of the induction of cancer. All these aspects converge in determining the environmental impact of organic fluorine compounds. The chemical stability of organic fluorine compounds is well established and it is sufficient to provide two illustrations: (a) the destruction of the aromatic ring and retention of the fluorine substituents during the reaction of 3-aminobenzotrifluoride with chromic acid, and (b) hydrolytic removal of fluoride from trifluoroacetate that necessitates refluxing in 30% NaOH. The increased application of perfluorinated compounds has stimulated examination both of their toxicology and methods for their destruction ; some of these abiotic reactions are also discussed in Sect. 6. Whereas biodegradation of organic fluorine with loss of fluorine compounds has been established for a range of compounds, biotransformation reactions in which fluorine is retained are frequently observed. To avoid ambiguity the distinctions between these terms are summarized. The same principles apply to abiotic degradation and transformation. Biodegradation results in the mineralization of an organic compound to carbon dioxide and water and – if the compound contains nitrogen, sulfur, phosphorus, or halogen – with the release of ammonium (or nitrite), sulfate, phosphate, or halide. In the present review emphasis is placed on loss of fluoride as a measure of degradation. Studies in which only loss of substrate is used as a measure of degradation or transformation but which lack identification of metabolites or further details have been given less prominance. Such studies have frequently been carried out on agrochemicals where the emphasis on mechanistic detail is not obligatory. Biotransformation, on the other hand, is used for reactions in which only a restricted number of reactions take place, and the basic framework of the molecule remains essentially intact: fluorine may not necessarily be lost. In this context both abiotic (photochemical) and biotic (bacteria, fungi, and yeasts) are discussed. Although it is generally assumed that the fluoride resulting from biodegradation is the terminal product, its incorporation into organic fluorine compounds is noted in Sect. 2.2 and the synthesis of naturally occurring organic fluorine compounds is discussed by Gribble in Chap. 5. During the preparation of this review it became apparent that the mechanisms for biodegradation or biotransformation of organic fluorine compounds frequently paralleled those of the corresponding organochlorine compounds. In addition, a discussion is given on the metabolism of unsubstituted – though otherwise analogous – aromatic compounds. The opportunity has therefore been taken to adopt a reductionist approach and draw these together so that a wider topic has been addressed. In some compounds, however, the fluorine substituents are distant from the sites at which reaction occurs, and in these fluorination plays a minor part. The

140

A.H. Neilson · A.-S. Allard

biochemistry of the metabolism of organic fluorine compounds has attracted somewhat limited attention so that more extensive discussion is justified when such investigations have been carried out. The range of fluorinated compounds of industrial importance is wide and includes, for example, liquid crystals, reactive dyes, surfactants, plastics and coatings, chemotherapeutics, and agrochemicals. All of these have been covered in a book [10] to which reference should be made for details of their synthesis. Whereas, for example, agrochemicals enter the environment directly, manufacturing processes may give rise to by-products that may be recalcitrant to normal treatment practice. Illustrative examples are provided by the following: 1. The occurrence of isomeric 2-trifluoro-3-nitro-7-(8)chlorodibenzo[1,4]dioxins as impurities in the lampricide 3-trifluoromethyl-4-nitrophenol [92] 2. The presence in contaminated sediments of difluorodiphenylmethane and several biphenyls from the production of 4-chloro-(trifluoromethyl)benzene [105] 3. In a single incident, dust samples from an aluminum refining plant that used dichlorodifluoromethane (Freon-12), dibenzofurans, biphenyls, biphenylenes and naphthalenes containing both chlorine and fluorine substituents were identified [244] One the reason for the attractiveness of the trifluoromethyl group is the combination of its stability with its electron-attracting property.Although care must be exercised in making generalizations, although the trifluoromethyl group is appreciably recalcitrant to biological transformation it may be photolabile. It is not possible to make comparison with the trichloromethyl group since, except in alkanes, this is an unusual substituent. The environmental origin of trifluoroacetate has been discussed by Wallington and Nielsen in Chaps. 3 and by Ellis et al. Chap. 4, and its degradability is noted in Sect. 2.2. Biochemical reactions involve interaction of the substrate with an enzyme that functions as a catalyst. The enzyme is then released from the transformed substrate albeit at the increase in entropy of the system. In the context of biodegradation, an alternative has emerged in suicide metabolism (lethal synthesis): that is discussed in a wider context in Sect. 5.2. The enzyme metabolizes the substrate to products that react with the enzyme to form essentially irreversible complexes from which the active enzyme is not released for recycling: enzymatic activity is thereby destroyed. Extensive studies have been directed to the metabolism of organic fluorine compounds in higher organisms in the contexts of toxicity, drug action, and enzymology [170, 239, 245]. A discussion of only a few of these in a limited context is given in Sect. 5, but reference should be made to the extensive discussions in other chapters in this book. A few brief comments are made on the design of experiments and the extent to which the results of these can be extrapolated to natural ecosystems. The significance of concentration and the presence of several potential substrates and a wide microflora may profoundly effect the recalcitrance of a xenobiotic and has been reviewed [1, 162]. One further aspect is noted here: the association of sub-

Degradation and Transformation of Organic Fluorine Compounds

141

Table 1. Designations of selected taxa that have undergone alteration

Current designation

Former designation

Aminobacter aminovorans Brevundimonas diminuta Burkolderia cepacia Comamonas testosteroni Comamonas acidovorans Ralstonia pickettii Sphingomonas paucimobilis Stenotrophus maltophilia Ralstonia eutropha

Pseudomonas aminovorans Pseudomonas diminuta Pseudomonas cepacia Pseudomonas testosteroni Pseudomonas acidovorans Pseudomonas pickettii Pseudomonas paucimobilis Pseudomonas maltophilia Alcaligenes eutrophus

strates with organic matter that reduces their bioavailability is frequently neglected in studies on biodegradation and biotransformation.Although phase partition has been discussed by Ellis et al. in Chap. 2, the association of compounds as different as trifluoroacetate with soil [193], hexafluorobenzene with sediments [41], fluorinated quinoline carboxylic acid derivatives with clay minerals [169], and fluorinated compounds with humic and fulvic acids (references in [51]) is worth noting. An issue of cardinal importance is the determination of the structure of metabolites. In addition to the widely used 1H and 13C NMR, the additional value of 19F NMR deserves special mention. A comprehensive discussion of the principles and practice of 19F NMR is given by Stanley in Chap. 1, and a number of illustrations are provided in this chapter. The names of organisms are those given by the authors. Particularly for organisms designated Pseudomonas sp., increasingly detailed studies and genetic relationships have resulted in substantial changes in nomenclature, but no attempt has been made systematically to assign these organisms to the currently accepted taxa. Some of the reassignments are given in Table 1 and others in the text.

2 Aliphatic Compounds The first step in the metabolism of halogen-substituted alkanes and alkanoates is loss of halogen mediated by halohydrolases, oxygenases, reductases, or glutathione S-transferases [63]. For some C-1 compounds, the initial step is monoxygenation, whereas for dichloromethane aerobic dehalogenation is mediated by glutathione S-transferase. In both reactions, formaldehyde is produced and is assimilated by the ribulose monophosphate cycle [7]. For C-2 compounds, the elimination reaction is much less frequent, and metabolites produced by alternative dehalogenation pathways enter either the TCA or the glyoxylate cycles of central metabolism. Anaerobic transformation may involve partial or complete reductive loss of halogen.

142

A.H. Neilson, A.-S. Allard

Table 2. Acronyms for chlorofluorocarbons (CFCs), hydrofluorochlorocarbons (HCFCs), and hydrofluorocarbons (HFCs) (redrawn from Midgley and McCulloch 1999) [150]

Acronym Structure

Acronym

Structure

Acronym

Structure

CFC-11 CFC-12 CFC-113 CFC-114 CFC-115

HCFC-123 HCFC-22 HCFC-141b HCFC-142b HCFC-124 HCFC-225ca HCFC-225cb

CHCl2 · CF3 CHClF2 CH3 ⋅ CCl2F CH3 ⋅ CClF2 CHClF ⋅ CF3 CHCl2 ⋅ CF2 ⋅ CF3 CHFCl ⋅ CF2 ⋅ CF2Cl

HFC-134a HFC-227ea HFC-152a HFC-143a HFC-125 HFC-32 HFC-23 HFC-245ca

CH2F ⋅ CF3 CF3 ⋅ CHF ⋅ CF3 CH3 ⋅ CHF2 CH3 ⋅ CF3 CHF2 ⋅ CF3 CH2F2 CHF3 CHF2 ⋅ CF2 ⋅ CH2F

CCl3F CCl2F2 CCl2F ⋅ CClF2 CClF2 ⋅ CClF2 CClF2 ⋅ CF3

2.1 Alkanes and Alkenes

Chlorinated ethanes and ethenes have been used extensively as solvents and concern over their adverse environmental impact has led to extensive studies on their biodegradation both aerobically by methanotrophs and anaerobically by both a number of genera of sulfate-reducing bacteria and by mixed cultures of methanogens (references in [233]). A general discussion of anaerobic dechlorination by dehalorespiration is given in the introduction to Sect. 3. Brief attention is drawn here to important groups of dechlorinating anaerobic bacteria: 1. Dehalobacterium formicoaceticum that is able to degrade dichloromethane fermentatively by conversion to formate and acetate [157]. 2. Dehalorespiration of halogenated ethenes by Dehalospirillum multivorans, Dehalobacter restrictus, Desulfuromonas chloroethenica, and strains of Desulfitobacterium sp. (references in [97]), and the range of chloroethenes that support dehalorespiration of Dehalococcoides ethenegenes and its capability of reductively dehalogenating tetrachloroethene to ethene [136]. The analogous perfluorinated compounds including difluoromethane (HFC32), trifluoroethane (HFC-143a), tetrafluoroethane (HFC-134a), hexafluoropropane (HFC-236ea), and heptafluoropropane (HFC-227ea) are of potential interest as environmentally acceptable [85]. A full list of acronyms is given in Table 2 [150]. The biodegradation of hydrochlorofluorocarbons and hydrofluorocarbons has attracted considerable attention on account of their presumptive adverse effect on ozone depletion (references in [62]), and their biodegradation follows broadly that of chlorinated ethanes and ethenes.Valuable background on chlorofluorocarbons has been given in [53] and on alternatives to them in [183]. 2.1.1 Fluorohydrocarbons

Fluoromethane is the simplest organic fluorine compound and was prepared in 1835 by reacting methyl sulfate with potassium fluoride. It has been used as a selective inhibitor of ammonium oxidation and nitrification-linked synthesis of

143

Degradation and Transformation of Organic Fluorine Compounds

Table 3. Release of fluoride during complete degradation of dichlorofluoromethane (HCFC-21)

and 1-chloro-1,1-difluoroethane (HCHC-142b) by Methylosinus trichosporium strain OB3b and Mycobacterium vaccae strain JOB5. % Relative to values for complete biodegradation Substrate

Methylosinus trichosporium

Mycobacterium vaccae

% fluoride released Dichlorofluoromethane 1-Chloro-1,1-difluoroethane

109 NR

91 105

NR not reported.

N2O in Nitrosomonas europaea [152], while difluoromethane has been proposed as a reversible inhibitor of methanotrophs [151]. Consistent with the established substrate range of monooxygenases, ammonia monooxygenase in Nitrosomonas europaea is able to oxidize fluoromethane to formaldehyde [102]. The methanotroph Methylosinus trichosporium strain OB3b that produces the soluble methane monooxygenase system consisting of a 40-kDa NADH oxidoreductase, a 245-kDa hydroxylase, and a 16-kDa protein termed component B has a low substrate specificity [224] and has been shown to metabolize trifluoroethene to glyoxylate, difluoroacetate, and the rearranged product trifluoroacetaldehyde [66] by a reaction analogous to the formation of trichloroacetaldehyde from trichloroethene from the same strain [171]. This strain and Mycobacterium vaccae strain JOB 5 that produces propane monooxygenase have been used to examine the degradation of a number of hydrofluorocarbons and hydrofluorocarbons [221]. It was shown that during complete degradation by M. trichosporium and M. vaccae degradation was accompanied by release of fluoride (Table 3). It is worth noting, however, that the methylotrophs IMB-1 [201] and CC495 [42] are able to oxidize chloromethane, bromomethane, and iodomethane though not fluoromethane. On the basis of their 16S rRNA sequences, these organisms are related to organisms classified as Pseudaminobacter sp. but more distant from nitrogen-fixing rhizobia [42]. 2.1.2 Chlorofluorocarbons and Hydrochlorofluorocarbons

In general, halogen is lost from organic substrates in the order I >Br>Cl>F, and the following examples illustrate this for substrates containing both chlorine and fluorine substituents. As for the substrates in Sect. 2.1.1, substantial effort has been directed to degradation and transformation by methylotrophic bacteria. The hydrochlorofluorocarbons are more readily degraded that the corresponding compounds lacking hydrogen. 1. The soluble methane monooxygenase from Methylosinus trichosporium OB3b that has a wide substrate spectrum is able to oxidize some hydrochlorofluoroethanes (Table 4) though neither trichlorofluoromethane CFC-11 nor any hydrochlorofluoroethane with three fluorine substituents on the same carbon

144

A.H. Neilson · A.-S. Allard

Table 4. Incubation of chlorofluoroethanes and fluoroethanes with soluble methane oxygenase

from Methylosinus trichosporium Degraded

Undegraded

1,1-Dichlo-1-fluoro (HCFC-141b) 1,1,2-Trichloro-2-fluoro (HCFC-131) 1,1,2-Trifluoro (HFC-143)

2,2-Dichloro-1,1,1-trifluoro (HCFC-123) 1-Chloro-1,1-difluoro (HCFC-142b) 1,1,1,2-Tetrafluoro (HFC-134a) 1,1,2,2-Tetrafluoro (HFC-134)

atom were oxidized [45]. Although quantitative loss of fluoride and chloride was shown for dichlorofluoromethane, none of the organic products of oxidation were identified for the others. 2. A study using a mixed culture containing a methanotrophs and a plethora of other bacteria showed that substrate loss of chlorodifluoromethane (HCFC22) and 1-chloro-1,1-difluoroethane (HCFC-142b) took place although only ca. 50% of the substrate was degraded even after prolonged incubation [35]. Neither 1,1,1,2-tetrafluoroethane (HFC-134a) nor 1,1-dichloro-2,2-difluoroethane (HCFC-123) were appreciably degraded. Whereas the degradation of polychlorinated ethenes that are stable and have been used as solvents is mediated by monooxygenation, the corresponding polyfluorinated ethenes are not themselves used directly since they are susceptible to polymerization and are used for further synthesis. It is well established that dehalogenation of a range of chloroalkanes and chloroalkenes, and their brominated homologues may take place under anaerobic conditions. Several sulfate-reducing bacteria have been isolated, and in methanogens, cofactors including coenzyme F430, and corrinoids [95, 96, 124, 125] plausibly mediate the reaction [68]. Illustrative examples of fluorinated analogues include reactions carried out by both methanogens and sulfate-reducing bacteria. 1. The methanogen Methanosarcina barkeri is able to transform trichlorofluoromethane (CFC-11) by successive loss of chlorine to produce chlorofluoromethane (Fig. 1a) [123], and a similar transformation has been demonstrated in the presence of sulfate and butyrate with a mixed culture containing putatively Desulfovibrio baarsii and Desulfobacter postgatei [220]. 2. In anaerobic microcosms 1,1,2-trichloro-1,2,2-trifluoroethane (CFC-113) was transformed by successive reductive dechlorination to 1,2-dichloro-1,2,2trifluoroethane (HCFC-123a), and under methanogenic conditions to 1chloro-1,2,2-trifluoroethane (HCFC-133) and 1-chloro-1,1,2-trifluoroethane (HCFC-133b) without evidence for reductive replacement of fluorine (Fig. 1b) [126]. 3. 1,1,1-Trifluoro-2,2-dichloroethane (HCFC-123) was recalcitrant in aerobic soils but underwent reductive dechlorination anaerobically to produce 1,1,1trifluoro-2-chloroethane [173].

Degradation and Transformation of Organic Fluorine Compounds

145

Fig. 1. a Metabolism of trichlorofluoromethane by Methanosarcina barkeri. b Transformation

of 1,1,2-trichloro-1,2,2-trifluoroethane (CFC-113) under anaerobic conditions. c Dehalogenation by cytochrome P450CAM of trichlorofluoromethane and 1,1,1-trichloro-2,2,2-trifluoroethane

It is worth noting that considerable effort has been devoted to the study of model systems for dehalogenation by Castro [31], and that a strain of Shewanella alga strain BrY that is able to reduce Fe (III) and Co (III) using lactate or H2 as reductant is able to dechlorinate carbon tetrachloride to carbon monoxide in the presence of vitamin B12 and reductant [252]. Cytochrome P450CAM is a monooxygenase that is induced in Pseudomonas putida G786 by growth on camphor, and is responsible for the introduction of a hydroxyl group at C-5. Although this is the first step in the degradation of camphor, this enzyme is also able to carry out non-physiological reductive dehalogenation [32]: Monooxygenation Reductive dehalogenation

R · H+O2 +2 H+ +2 e → R · OH+H2O R · Hal+2 H+ +2 e → R · H+H·Hal

and it has been shown [130] that the electrons for reduction may be supplied by putidaredoxin. The production of carbon monoxide from trichlorofluoromethane catalyzed by cytochrome P450CAM proceeded through intermediate formation of the dichlorofluorocarbene [130] (Fig. 1c). Other reactions included b-elimination from 1,1,1-trichloro-2,2,2-trifluorethane (Fig. 1c). Pseudomonas putida strain G786 (pGH-2) was constructed to contain both the cytochrome P450CAM genes on the CAM plasmid and the tod C1, tod C2, tod b, and tod A genes of toluene dioxygenase. Toluene dioxygenase was constitutively expressed and cytochrome P450CAM after induction by camphor [238]. Under anaerobic conditions 1,1,1,2-tetrachloro-2,2-difluoroethane was dehalogenated to 1,1chloro-2,2-difluoroethene that could be oxidized by the dioxygenase under aerobic conditions to oxalate (Fig. 2) [100].

146

A.H. Neilson · A.-S. Allard

Fig. 2. Combined dehalogenation of 1,1,1,2-tetrachloro-2,2-difluoroethane by cytochrome

P450CA M and oxidation with toluene 2,3-dioxygenase

2.2 Carboxylic Acids and Related Compounds

Various aspects of the biochemistry of fluorinated aliphatic carboxylic acids have motivated their investigation: 1. Their occurrence in higher plants (Gribble, Chap. 5). 2. The mechanism of toxicity of fluoroacetate and the use of fluorinated substrates to probe details of the TCA cycle Walsh ([239]). 3. Environmental sources, distribution, and the biodegradability of trifluoroacetate (Wallington and Nielsen, Chap. 3 and Ellis et al., Chap. 4). 2.2.1 Fluoroacetate

This occurs in natural environments and has been found in a restricted number of higher plants (Gribble, Chap. 5), including species of Dichapetalum. The mechanism of fluoroacetate toxicity in mammals has been extensively examined and was originally thought to involve simply initial synthesis of fluorocitrate that inhibits aconitase and thereby the functioning the TCA cycle.Walsh has extensively reinvestigated the problem and revealed the complexity of the mechanism of inhibition and the stereospecificity of the formation of fluorocitrate from fluoroacetate [239]. As might be expected, bacteria have been isolated from the plants that produce fluoroacetate, and these include an unidentified Pseudomonas sp. [75], a strain of Pseudomonas cepacia from Dichapetalum cymosum [149], and a strain of Moraxella sp. [113]. Fluoroacetate is also an unusual product of microbial metabolism: 1. It is a terminal metabolite formed during the metabolism of 2-fluoro-4-nitrobenzoate by Nocardia erythropolis (Fig. 10) [28]. 2. It is formed together with 4-fluorothreonine during the late-stage growth of Streptomyces cattleya on a defined medium in the presence of fluoride [184]. Although it has been shown [81] that glycine is an effective precursor of both fluoroacetate and 4-fluorothreonine, and that glycine is metabolized via

Degradation and Transformation of Organic Fluorine Compounds

147

N5,N10-methylenetetrahydrofolate to serine and thence to pyruvate the mechanism of incorporation of fluoroacetate is unresolved (see also alanine racemase, Sect. 5). The metabolism of fluoroacetate proceeds by formation of fluoride and glycollate and use of H218O showed that the oxygen atom was introduced from water [76]. The enzyme has been purified from an unidentified pseudomonad [75] and is specific for fluoroacetate: it is inactive towards 2- and 3-fluoropropionate, di- and trifluoroacetates and fluorobenzoates. It is rapidly inhibited by 4chloromercuriphenylsulfonate and slowly by N-methylmaleimide; this suggests the involvement of an active thiol group in the enzyme. A strain of Moraxella sp. has also been shown to assimilate fluoroacetate using plasmid-determined dehalogenase activities [113]: one was active towards both fluoroacetate and chloroacetate H-1 whereas the other H-2 was active only towards chloroacetate. Further investigation [8] of the haloacetate hydrolase H-1 from Pseudomonas sp. strain A used 1H NMR of the (–)-a-methoxyl-a-(trifluoromethyl) phenylacetic acid ester of the phenacyl ester of the glycolate produced from (S)-[22 H1]fluoroacetate. This established that the major metabolite was the (R)-enantiomer of 2-[2H1]glycolate so that the reaction proceeded with inversion of the configuration at C-2. These observations may be viewed in the perspective of discussions on the various halohydrolases including the two groups of haloacetate halohydrolases and haloalkane halohydrolases [63]. 2.2.2 Fluorofumarates

Fumarase – a ubiquitous cellular enzyme – catalyzes the hydration of fumarate to L-malate (see [94] for a general review of classical studies). It reacts with monofluorofumarate by reaction from the si-si face of fumarate of HO at the fluorine-bearing carbon to produce a fluorohydrin from which fluorine is lost to produce oxalacetate. For unresolved reasons, the opposite is true for chlorofumarate. A more complex situation arises with difluorofumarate in which hydration produces a fluorohydrin that loses fluorine to produce 3S-fluorooxalacetate whose configuration was proved by formation of only L-threo-2R,3S-3-fluoromalate after reduction with malate dehydrogenase [140]. 2.2.3 Di- and Trifluoroacetate

The soluble methane monooxygenase from Methylsosinus trichosporium OB3 produces difluoroacetate as one of the main products from the oxidation of trifluoroethene, and a low yield of trifluoroacetaldehyde by rearrangement [66]. Chlorodifluoroacetic acid has been identified in rain and snow samples [143] and may plausibly be an atmospheric degradation product of 1,1,2-trichloro1,2,2-trifluoroethane. Trifluoroacetate has been found in a wide range of environmental samples [115] (Ellis et al., Chap. 4) and attention has been directed to

148

A.H. Neilson · A.-S. Allard

three issues: its origin, its toxicity, and its recalcitrance. Possible sources are discussed in Chap. 4, and, compared with chloroacetates, it is only mildly phytotoxic to algae [15] and higher plants [19]. It is important to draw attention to its formation during the photochemical degradation of 3-trifluoromethyl-4-nitrophenol [54] (see Sect. 6.2 and Fig. 39). Concern has been expressed over its biodegradability, and conflicting results over its recalcitrance under anaerobic conditions have been reported. This has been resolved by the results of a chemostat study using a mixed ethanol-degrading culture [117]. Clear evidence for anaerobic degradation was produced on the basis of fluoride release and formation of acetic acid and methane. 2.3 Perfluoroalkyl Sulfonates

There has been increasing interest in perfluoro sulfonates and carboxylates [115]. The sulfonates are valuable surfactants under extreme conditions and are components of fire fighting foams [155]. Perfluorooctane sulfonate accumulates in marine mammals [109] and has a global distribution in a range of biota [73]. Low levels of perfluoro compounds have been carefully quantified in samples of human sera by Hansen et al. [82], and determination of perfluorinated surfactants in surface water samples has been described: two independent analytical techniques – liquid chromatography/tandem mass spectrometry and 19F NMR – were used, and a summary of analytical methods was included [156]. References to toxicological studies in rats and epidemiological studies in man have been given [82] and the biochemical toxicology of the related perfluorooctanoic acid is discussed by DePierre in Chap. 7. A strain of Pseudomonas sp. strain D2 was used to evaluate the degradation of a range of fluorinated sulfones [116], and defluorination was observed with difluoromethane sulfonate, 2,2,2-trifluoroethane sulfonate, and 1H,1H,2H,2H-perfluorooctane sulfonate. For growth and defluorination of difluoromethane sulfonate, acetate or glucose were supplied as the source of carbon, ammonium as the source of nitrogen, and the substrate as the source of sulfur. It should be noted that perfluoro analogues that lacked hydrogen substituents did not support growth of this strain.

3 Aromatic Compounds 3.1 Introduction

A somewhat lengthy introduction is provided in which it is attempted to provide a perspective against which the metabolism of fluorinated aromatic compounds may be viewed. It covers the initial stages in the aerobic and anaerobic metabolism of substrates but excludes discussion of the lower pathways.

Degradation and Transformation of Organic Fluorine Compounds

149

3.1.1 Aerobic Conditions

The pathways and mechanisms for the biodegradation of hydrocarbons, phenols, benzoates, and anilines under aerobic conditions are well established. Details of the mechanisms and the genetic determinants in bacteria have been elucidated and a brief summary of salient issues that are relevant to substituted aromatic compounds is given. Further details are available from reviews [163, 217, 234]. Bacterial oxygenases (excluding cytochrome P-450 systems) 1. Source of oxygen O2-oxygenases 1.1 Monooxygenases 1.2 Dioxygenases 2. Source of oxygen H2O-hydroxylases 2.1 Oxidoreductases 2.2 Hydroxylating system: chemical reaction with cation radicals produced by peroxidases 1. Oxygenases 1.1 Monooxygenases

•

Flavoproteins introducing oxygen at site adjacent to existing hydroxyl group 4-hydroxybenzoate hydroxylase Anthranilate hydroxylase in yeasts Salicylate hydroxylase

hydroxylases introducing oxygen into non-oxygenated rings • Hydrocarbon Toluene, phenanthrene, pyrene monooxygenases epoxidases • Hydrocarbon Fungal and yeast biotransformations of PAHs: production of phenols by NIH shift 1.2 Dioxygenases

•

Dihydroxylating enzymes Producing cis dihydrodiols: soluble multicomponent enzymes requiring NAD(P)H as cofactor) Benzene 1,2-dioxygenase Toluene 2,3-dioxygenase Naphthalene 1,2-dioxygenase Biphenyl 2,3-dioxygenase

enzymes • Ring-cleavage (Carrying out ring fission and requlring no cofactor) Catechol 2,3-dioxygenase (extradiol cleavage to carboxyaldehydes) Catechol 1,2-dioxygenase (intradiol cleavage to dicarboxylates) 2. Hydroxylases 2.1 Oxidoreductases Mo multicomponent enzymes carrying out addition of H2O followed by dehydrogenation e.g. anaerobic hydratases coupled to dehydrogenases 2.2 Hydroxylation systems Secondary reaction with cation radicals formed by peroxidases Fig. 3. Outline of oxygenation reactions (redrawn from Neilson and Allard 1998 [163])

150

A.H. Neilson · A.-S. Allard

Pathways of Oxygenation

Aromatic hydrocarbons are generally degraded by two mechanisms under aerobic conditions: (i) dioxygenation by bacteria with the introduction of both atoms of oxygen and formation of cis-dihydrodiols or (ii) monooxygenation by fungi and yeasts to produce epoxides that are converted into trans-dihydrodiols by epoxyhydrolases. The cis-dihydrodiols are dehydrogenated to catechols, whereas phenols may be produced from the trans-dihydrodiols by rearrangement. For benzoates, dioxygenation to catechols generally involves concomitant decarboxylation, and for ortho-halogenated benzoates loss of both CO2 and halogen may occur: this is discussed in Sect. 3.4.1. Further degradation of the catechols is accomplished by dioxygenation with fission of the ring either between the substituents (intradiol or 1,2-fission), or between one hydroxyl group and the adjacent carbon atom (extradiol or 2,3-fission). Although extradiol fission is generally preferred, both may occur and whether extradiol or intradiol fission is followed depends on the substrate. This has important consequences as a result of the synthesis of metabolites that inhibit further degradation. Discussions of the various oxygenases are given in [83, 145, 159]. A tabular summary of bacterial oxygenases is given in Fig. 3.

Dehalogenation

Enormous effort has been expended on the biodegradation of halogenated aromatic compounds, and it is sufficient to summarize the essentially different pathways: (i) loss of halogen before fission of the aromatic ring and (ii) loss of halogen from muconates after ring fission. Loss of halogen from the ring by halohydrolase activity may occur but is less common, and for halogenated phenols depends on the number of substituents. For phenols with less than three halogen substituents, hydroxylation is used to initiate degradation and halide is lost after ring fission, whereas for those with three or more chlorine substituents, halogen may be lost before ring fission by the activity of halohydrolases and reductases. This is discussed in Sect. 3.5. 3.1.2 Anaerobic Conditions

Anaerobic degradation is important for compounds that are partitioned into the sediment phase from the aquatic environment. In sediments to which mixed effluents are discharged the system rapidly becomes anaerobic due to the consumption of oxygen by readily degraded low-molecular weight components in the effluent. The fate of other, often recalcitrant, contaminants is therefore largely determined by the outcome of these anaerobic reactions.

Degradation and Transformation of Organic Fluorine Compounds

151

Background

As noted previously, many agrochemicals contain fluorine or trifluoromethyl substituents and, although there is evidence for their partial degradation under anaerobic conditions in soils, the C–F bonds are generally recalcitrant even in highly reducing environments. Two examples are used as illustration. 1. Trifluralin is degraded to products of unestablished structure by microbial reactions under oxygen limitation by nitrate, or Fe (III) reduction [228]. 2. Under sulfate-reducing or methanogenic conditions flumetsulam is transformed with reduction of the triazolo[1,5-a]pyrimidine ring [251].

Organisms and Electron Acceptors

Anaerobic transformations may be carried out by fermentation, or by bacteria using sulfate or carbonate as electron acceptors (sulfate-reducing and methanogenic bacteria) or nitrate: the first are important in marine or brackishwater systems that contain high concentrations of sulfate. Although the pathways and mechanisms for anaerobic degradation of both benzoate and phenol have been explored extensively in denitrifying organisms, they may be different from those used by sulfate-reducing organisms. For example, although it has been shown [182] that toluene is degraded by the same pathway for both a sulfate-reducing and a denitrifying bacterium, the extent to which this is generally true has not been established. Considerable effort has been expended in elucidating the degradation pathways of benzoates and phenols by denitrifying bacteria. Except in poorly managed agricultural conditions or in municipal treatment systems, however, denitrification is uncommon in natural environments where nitrate is unlikely to be available, and where alternative electron acceptors including sulfate are available. In addition, pathways used by denitrifying organisms may not be representative of those used by other anaerobic bacteria, and may be different from those used even by the same organism when oxygen is used as electron acceptor [225]. Organisms that use the halogenated substrates as electron acceptors are discussed in the next section

Dehalogenation

Dechlorination of all structural groups of aromatic compounds has been examined in considerable detail. Only some general issues are summarized here by way of orientation to plausible reactions with fluorinated aromatic compounds. In general, it may be stated that dechlorination of highly halogenated substrates may often be partial rather than total. Emphasis is placed here on chlorinated benzoates and phenols for which details are most extensive.Attention is directed to a review on dehalorespiration in which reductive dehalogenation is coupled to energy generation in the cell [97].

152

A.H. Neilson · A.-S. Allard

1. Dehalogenation may be carried out using the chlorinated substrates as electron acceptor with electron donors including methanol, lactate, pyruvate, or H2. 2. Strains may be flexible enough to encompass dechlororespiration of ethenes and chlorophenols. a) Co-induction of tetrachloroethene and 3-chlorobenzoate dehalorespiration may occur in Desulfomonile tiedjei strain DCB-1 [39]. b) A strain of Desulfitobacterium sp. strain PCE1 can grow by reductive dechlorination of tetrachloroethene or ortho-chlorinated phenols [70]. 3. Desulfomonile tiedjei strain DCB-1 was the first pure culture able to dechlorinate 3-chlorobenzoate [49], and the membrane-bound dehalogenase has been purified [166]. 4. Dechlorination of phenols has been observed in Desulfomonile tiedjei strain DCB-1 [154], Desulfitobacterium dehalogenans [232], Desulfobacterium hafniense [38], and Desulfitobacterium frappieri [24, 47]. 5. Although the organisms belong to taxonomically diverse groups, the dehalogenases are generally membrane-bound (references in [36]). 3.2 Carbocyclic Aromatic Hydrocarbons 3.2.1 Aerobic Organisms

The basic mechanisms for the aerobic degradation of carbocyclic aromatic compounds are well established and initial dioxygenation is now recognized as one of the fundamental mechanisms for activating aromatic rings to aerobic de-

Fig. 4. Transformation of fluorinated substrates by toluene 2,3-dioxygenase and yields (%)

Degradation and Transformation of Organic Fluorine Compounds

153

Fig. 5. Metabolism of 4,4′-difluorobiphenyl production by Sphingomonas strain SS33

gradation. For toluene a number of distinct pathways have been elucidated, and an important one involves dioxygenation at the 2,3-positions. Gibson et al. [71] used a strain of Pseudomonas putida that was grown with toluene and was able to oxidize 4-chlorotoluene to (+)-cis-4-chloro-2, 3-dihydrodihydroxytoluene and this strain was also able to transform fluorobenzene to 3-fluorocatechol. Whereas dioxygenation without loss of chlorine takes place for chlorobenzenes in which there are vicinal unsubstituted positions on the ring, dioxygenation with concomitant loss of chloride occurs with 1,2,4,5-tetrachlorobenzene [199]. A strain of Pseudomonas strain T-2 in which toluene 2,3-dioxygenase is synthesized was used to study the transformation of a series of 3-substituted fluorobenzenes [189], and consistent with the mechanisms proposed above, two reactions were observed. Dioxygenation took place in both, and dioxygenation with concomitant loss of fluorine was generally dominant. The relative contributions of these reactions depended on the substrate (Fig. 4). Similarly, the biphenyl dioxygenase BphA in Burkholderia sp. strain LB400 was able to accept a number of 2,2′-substituted biphenyls including 2,2′-difluorodiphenyl from which 2′-hydroxy-2,3-dihydroxybiphenyl was produced by loss of fluoride from the initially formed fluorohydrodiol [214]. A strain of Sphingomonas strain SS33 was adapted to the growth of 4,4′-difluorodiphenyl ether and the transient intermediates 4-fluorophenol and 4-fluorocatechol were formed before loss of fluoride (Fig. 5) [206] presumably through the activities of phenolhydroxylase and catechol 1,2-dioxygenase that were confirmed in the parent strain [207]. It has already been noted that the metabolism of aromatic substrates by yeasts and fungi is distinct from that used by bacteria. Biotransformation is initiated by monooxygenation to the epoxide and subsequent hydrolysis to the trans-dihydrodiols, while phenols may be formed either by direct hydroxylation or by nonenzymatic rearrangement of the epoxide. Illustrative examples include the following: 1. The transformation of 4-fluorobiphenyl by the ectomycorrhizal fungus Tylospora fibrilosa was studied using 19F NMR [80], and 4-fluorobiphen-3′- and 4′-ol were the principal products (Fig. 5). 2. 1-Fluoronaphthalene was metabolized by Cunninghamella elegans to a number of metabolites whose synthesis was clearly initiated by formation of epoxides and trans-dihydrodiols (Fig. 6) [33]. This example illustrates the apparent indifference of the monooxygenase to the presence of the fluorine atoms.

154

A.H. Neilson · A.-S. Allard

Fig. 6. Transformation of 1-fluoronaphthalene by Cunninghamella elegans

3.2.2 Anaerobic Reactions

The bacterial metabolism of toluene has been achieved under different anaerobic conditions – denitrifying, sulfate-reducing, Fe(III)-reducing [134], syntrophic [148], and methanogenic conditions [13], and has been the subject of intense investigation. A number of different mechanisms and taxonomically different organisms have emerged.

Denitrifying Organisms

Some strains originally designated as Pseudomonas sp. belong to other genera such as Thauera or Azoarcus [5]. Broadly two pathways have been found: 1. In Pseudomonas sp. strain K172, the pathway involves successive dehydrogenation to benzyl alcohol and benzoate [4] followed by the reductive benzoate pathway noted in Sect. 3.4.2. 2. An alternative in Thauera aromatica strain K172 involves condensation of the benzyl radical with fumarate [128] to give benzylsuccinate. Benzylsuccinate synthase lies on the bss operon [128], and it has been shown that the enzymes required for further degradation of benzylsuccinate by b-oxidation lie on another operon bbs [127]. Degradation of toluene by Pseudomonas sp. strain T under anaerobic denitrifying conditions produced consecutively benzaldehyde and benzoate, and low yields of benzylsuccinate and benzylfumarate presumably by the alternative condensation pathway [216]: these pathways are not therefore mutually exclusive. Both pathways involve successive production of benzoate either by direct dehydrogenation or by b-oxidation of benzylsuccinate and formation after ring fission of alkanoic acids (acetate and in some cases propionate).

Degradation and Transformation of Organic Fluorine Compounds

155

Consistent with this, fluorotoluenes have yielded the corresponding fluorobenzoates: these are, however, merely biotransformations since the fluorobenzoates are terminal metabolites. Some examples illustrate this. 1. Pseudomonas sp. strain T oxidized 2-, 3-, and 4-fluorotoluenes to the fluorobenzoates without loss of fluorine, although the yields from 2- and 3-fluorotoluene were low [126]. 2. The toluene-degrading organism Azoarcus tolulyticus strain Tol-4 transforms toluene to (E)-phenylitaconate (benzylidene succinate) and to benzoyl-CoA [36] by a pathway that is consistent with the mechanism proposed above. In the presence of 2-fluorotoluene, 3-fluorotoluene, or 4-fluorotoluene, the corresponding fluorobenzoates were produced together with non-fluorinated (E)phenylitaconate. Presumably the fluorinated analogues were too transient for unambiguous detection. 3. Thauera sp. strain K172 was capable of oxidizing a range of substituted toluene to the corresponding benzoates: these included 3- and 4-fluorotoluene, and less effectively 2-fluorotoluene [16]. Sulfate-Reducing Bacteria

A number of strictly anaerobic sulfate-reducing bacteria have been implicated in the anaerobic degradation of toluene and alkyl benzenes, for example strains PRTOL1 [14], Tol2 [181], XylS1 [84], and TRM1 in coculture with Wolinella succinogenes [148]. A major fraction of the substrate is degraded to CO2 , and it has been shown [182] that the benzylsuccinate pathway is used by both the sulfatereducing bacterium Desulfobacula toluolica and a strain of denitrifying bacterium, and benzylsuccinate synthase activity has been found in a methanogenic consortium and a phototrophic bacterium (References in [13]); it is therefore plausible to conclude that this is a widely used pathway. It is important to underscore that the distinction made for aerobic organisms between the occurrence of dehalogenation before or after ring fission may be critical for anaerobic defluorination. Whereas the abiotic cycloisomerization of fluorinated muconates produced aerobically from benzoates has been examined, the anaerobic metabolism of the plausible more highly reduced fluorinated cyclohexene carboxylates produced after ring reduction is unresolved. Initial defluorination would then be preferable. 3.3 Heterocyclic Aromatic Compounds

The degradation or transformation of only a few fluorinated heterocyclic compounds has been examined under aerobic conditions. (1) Although nicotinic acid is degraded by a number of bacteria, the metabolism of its 5-fluorinated analogue has seldom been studied. Pseudomonas fluorescens strain N-9 can oxidize 5-fluoronicotinic acid though it is unable to grow at its expense. The initial reaction – analogous to that for nicotinic

156

A.H. Neilson · A.-S. Allard

Fig. 7. Degradation of 5-fluorouracil by Nectria haematococca

acid – is the formation of 5-fluoro-6-hydroxynicotinic acid [12]. Although this was oxidized further, the products were not identifiable by the methods then available. Pyruvate was formed from nicotinic acid and it was suggested by analogy that fluoropyruvate was produced from 5-fluoronicotinic acid and thence fluoroacetate that would inhibit effective functioning of the TCA cycle as a source of carbon and energy. On the other hand, the degradation of nicotinic acid forms maleic acid [11], and 5-fluorinicotinic acid would be expected to yield fluorofumarate whose degradation by fumarase is known to yield oxalacetate and fluoride [140]. (2) 5-fluoro-2′-deoxyuridine has been extensively used in studies of the mechanism of action of thymidylate synthase (Sect. 5.2), and 5-fluorouracil is an anticancer drug that has provided a lead to the development of others. The metabolism of 5-fluorouracil by the ascomycete fungus Nectria haematococca has been studied using 19F NMR [175]. a-Fluoro-b-alanine (2-fluoro3-aminopropionate) was found (Fig. 7) together with 5-fluorouridine-5′monophosphate, diphosphate and triphosphate in acid extracts of the mycelia, and the 2′- and 3′-monophosphates were recovered from RNA. (3) Fluoroquinolones are important antibiotics that have seen extensive use in veterinary medicine since they are effective against both Gram-negative organisms including Salmonella sp. and Helicobacter (Campylobacter) sp. and Gram-positive organisms. The degradation of the quinolone antibacterial drugs enrofloxacin [246] and ciprofloxacin [247] has been studied in the brown rot fungus Gloeophyllum striatum. A wide range of metabolites was isolated, among them a series in which the 6-fluorine substituent was replaced by hydroxyl (Fig. 8). The major metabolite produced from enrofloxacin by Mucor ramannianus was the N-oxide of the piperazine ring distal from the aromatic ring [176]. The important issue of the stability of aqueous solutions of these antibiotics is discussed in Sect. 6.

Fig. 8. Transformation of enrofloxacin by Gleophyllum striatum

157

Degradation and Transformation of Organic Fluorine Compounds

3.4 Benzoates 3.4.1 Aerobic Degradation Introduction

The metabolism of fluorobenzoates has been examined over many years. Studies with Nocardia erythropolis [28] and Pseudomonas fluorescens [98] showed that although rates of oxidation by whole cells were less than that for benzoate, they were comparable – and greater than for the chlorinated analogues. Details of the mechanisms, however, were dependent on the results of studies with nonfluorinated aromatic substrates, and on the availability of modern instrumentation for structural determination. This activity flourished during the 1970s and early 1980s when details of the bacterial dioxygenation of arenes became available through the work of Gibson and his collaborators (references in [72]), and these were quickly applied to the study of the degradation of halogenated aromatic compounds including fluorobenzoates by a number of workers. These investigations included both chlorinated and fluorinated aromatic compounds and both similarities and differences between them have emerged. For example, the metabolism of ortho fluorinated and ortho chlorinated benzoates differed mechanistically from that for both the meta and the para isomers. Although it is attractive to compare the rates of degradation and transformation of the isomers of halogenated benzoates, it is emphasized that there are marked differences between strains, and particularly in the details of the degradative pathways that are discussed below: – The rates of dioxygenation of the fluorobenzoates and chlorobenzoates by Pseudomonas sp. B13 grown with 3-chlorobenzoate were greater for the 3-halogenated compounds than those for the other isomers – though less than that for benzoate (Table 5) [186]. Reference to comparable supporting evidence is given above. – Whereas 3-fluoro- and 3-chloro-cis,cis-muconic acids prepared from Pseudomonas sp. B13 by catechol oxidation are cycloisomerized abiotically at pH values between 4 and 5, the corresponding 2-halogenated muconic acids are stable at pH 6 [205]. – Loss of halogen may take place before or subsequent to ring fission. Table 5. Rates of dioxygenation (%) of halogenated benzoates by whole cells of Pseudomonas

sp. strain B13 relative to benzoate (Reineke and Knackmuss 1978) [186] Substrate

Relative rate

Substrate

Relative rate

2-Fluorobenzoate 3-Fluorobenzoate 4-Fluorobenzoate

15 67 23

2-Chlorobenzoate 3-Chlorobenzoate 4-Chlorobenzoate

<1 31 <1

158

A.H. Neilson · A.-S. Allard

– For the ortho-halogenated compounds, dioxygenation may occur with simultaneous decarboxylation and loss of halide. – For the 4-halogenated compounds, loss of halogen may take place hydrolytically before ring fission. – The pathways for the degradation of 3-halogenated compounds are determined by the inhibitory effect of 3-halogenated catechols that is noted below. A few words are included on lethal metabolism (suicide) and this issue is taken up in another context in Sect. 5. Lethal metabolism occurs when a substrate is metabolized by an enzyme to a product that combines irreversibly with the enzyme and sabotages its activity. Catechols are produced by dioxygenation of the ring followed by dehydrogenation, and ring fission mediated by other dioxygenases occurs either between the hydroxyl groups (intradiol fission) or between one of them and the adjacent carbon atom (extradiol fission). For 3-halobenzoates, the metabolites are 3-halogenated catechols in which extradiol ring fission would produce acyl halides that would combine with and deactivate the catechol dioxygenase. Generally, therefore, although extradiol fission of catechols is preferred, it is avoided for 3-halogenated catechols, for example by the alternative 1,6 fission.

2-Fluorobenzoate

Although investigations on the degradation of 2-fluorobenzoate preceded the elucidation of the enzymatic details, the broad details were delineated on the basis of metabolites. Details of the pathways were later confirmed by enzymological studies, and it is significant that these investigations were carried out during the years in which details of the degradation of chlorinated analogues were being studied and from which many of the details emerged. The metabolism of 2-fluorobenzoate was described by Goldman et al. [77] in a strain of Pseudomonas sp. that was able to use 2-fluorobenzoate as sole source of carbon. They identified 3-fluorocatechol and 2-fluoromuconic acid as metabolites, while in studies with 18O2 [153] it was established that both atoms of catechol originated from oxygen. It was shown [187] that benzoate is transformed by a mutant strain of Alcaligenes eutrophus to 1,2-dihydrodihydroxybenzoate-1carboxylate, and a later study [185] showed that this strain produced comparable metabolites from fluorobenzoates. Alcaligenes eutrophus B9 and Pseudomonas sp. strain B13 can be adapted to growth with 2-fluorobenzoate and this involves dioxygenation to catechol with simultaneous loss of fluoride and decarboxylation followed by intradiol ring fission [55]. Further details on this reaction and a wider perspective were provided by a study of the metabolism of Pseudomonas cepacia strain 2CBS that is able to metabolize 2-fluorobenzoate, 2-chlorobenzoate, 2-bromobenzoate, and 2-iodobenzoate to catechol by concomitant decarboxylation and loss of halide (Fig. 9) [64]. In this strain, the inducible 2-halobenzoate 1,2-dioxygenase consisted of two components, an oxygenase A that is an iron-sulfur protein and consists of nonidentical subunits, a (Mr 52,000) and b (Mr 20,000) in an a3b3 structure. The re-

Degradation and Transformation of Organic Fluorine Compounds

159

Fig. 9. Transformation of 2-halogenated benzoates to catechol by concomitant decarboxylation

Fig. 10. Degradation of 2-fluoro-4-nitrobenzoate by Nocardia erythropolis

ductase B is an iron-sulfur flavoprotein (Mr 37,500) containing FAD. The enzyme is also active towards 2,4-difluorobenzoate and less so towards 2,6- and 3,4-difluorobenzoate. It is broadly similar to benzoate dioxygenase though different from the FMN-containing dioxygenases. This 2-halobenzoate-1,2-dioxygenase is also different from the three-component enzyme in Pseudomonas aeruginosa strain 142 [196] that contains a [2Fe-2S] ferredoxin and is active towards 2,5dichlorobenzoate and less to 2,6-, 3,5-, and 3,4-dichlorobenzoate. This is similar to other three-component dioxygenases containing ferredoxin such as benzene dioxygenase, toluene-2,3-dioxygenase, biphenyl dioxygenase, and naphthalene dioxygenases. A contrasting pathway was found for the metabolism of 2-fluoro-4-nitrobenzoate by Nocardia erythropolis that produced 2-fluoro-3,4-dihydroxybenzoate and 2-fluoroacetate which was identified as the hydroxamate: fluoride was not identified [28]. The pathway that was derived (Fig. 10) does not involve loss of fluorine concomitant with decarboxylation, and therefore differs from what has subsequently emerged as the principal pathway for metabolism of non-hydroxylated 2-halogenated benzoates.

3-Fluorobenzoate

It was shown [187] that the metabolism of 3-fluorobenzoate by a mutant of Alcaligenes eutrophus that was unable to grow with benzoate produced 3fluorobenzene-1,2-dihydrodihydroxy-1-carboxylate – tentatively the cis isomer.

160

A.H. Neilson · A.-S. Allard

Fig. 11. Degradation of 3-fluorobenzoate by regioselective dioxygenation

Fig. 12. Degradation of 3-chlorocatechol by Pseudomonas putida GJ31: a hydrolysis of the intermediate acyl chloride; b unproductive irreversible reaction with the enzyme EnzN

This provided the basis for further studies of the degradation of 3-fluorobenzoate. 3-Fluorobenzoate that is used for growth by a bacterial strain FLB300 that does not belong to the genus Pseudomonas [56, 209] minimizes the production of toxic 3-fluorocatechol as an intermediate by regioselective 1,6-dioxygenation with formation of 4-fluorocatechol (Fig. 11). This strategy minimizes the production of the inhibitory 3-fluorocatechol. It is worth inserting a few comments on the formally comparable situation for 3-chlorobenzoate and the corresponding cis,cismuconic acids that are the ring fission products. 1. The pathway for 3-fluorobenzoate is similar to that for the degradation of 3chlorobenzoate [203] though different from that for 3-chlorobenzoate in Alcaligenes strain BR60 in which a 3,4-dioxygenase (dechlorinating) produced 3,4-dihydroxybenzoate directly [158]. 2. The extradiol dioxygenase, chlorocatechol 2,3-dioxygenase from Pseudomonas putida GJ31 [141] is able to avoid inhibition of the intermediate acyl chloride by its hydrolysis to 2-hydroxymuconate [111] (Fig. 12). 3. For maleylacetates that contain a halogen atom (fluorine, chlorine, or bromine) at the 2 position, loss of halogen consumes 1 mol of NADH per mol of substrate, and the further reduction to 3-ketoadipate a further mol [110] (Fig. 13).

Degradation and Transformation of Organic Fluorine Compounds

161

Fig. 13. Degradation of halogenated maleylacetate to 3-ketoadipate in Pseudomonas sp. strain

B13

4-Fluorobenzoate

A mutant strain of Alcaligenes eutrophus was used [187] to demonstrate that the initial step in the metabolism of 4-fluorobenzoate involves 1,2-dioxygenation to 4-fluoro-1,2-dihydrodihydroxybenzene-1-carboxylate. 4-Fluorobenzoate may be used for growth by several bacteria, and metabolites produced by a strain of Pseudomonas sp. that was also able to degrade 4-fluorophenylacetic acid were identified and used to construct the pathway. Degradation took place by dioxygenation (decarboxylating), intradiol ring fission of the resulting catechol, and loss of fluoride from the muconolactone [88]. Part of this pathway (Fig. 14) was confirmed in later studies in which the important role of maleylacetate reductase was noted [203, 204, 209]. Strains of Alcaligenes RH021 and RH022 were able to utilize all the monofluorobenzoates for growth and two degradative pathways were observed: 1. In strain RH025, a hydrolytic mechanism brought about loss of fluoride from 4-fluorobenzoate to produce 4-hydroxybenzoate that was further hydroxylated to 3,4-dihydroxybenzoate before intradiol ring cleavage (Fig. 15a).

Fig. 14. Degradation of 4-fluorobenzoate

162

A.H. Neilson · A.-S. Allard

Fig.15. Alternative paths of degradation of 4-fluorobenzoate by Alcaligenes sp.: a strain RH 025; b strain RH 021

2. In strain RH021, a benzoate dioxygenase produced 4-fluorocatechol that was degraded by intradiol fission (Fig. 15b) [172]. This pathway is the same as that proposed for the degradation of 4-chlorobenzoate by a strain of Arthrobacter sp. [139], and for 2,4-dichlorobenzoate by Alcaligenes denitrificans strain NTB1 after initial reductive loss of the ortho chlorine substituent [236], and in a coryneform bacterium through an intermediate 2,4-dichlorobenzoyl coenzyme A [197]. Difluorobenzoates

Polyfluorinated benzoates have found several areas of application: 1. Fluorobenzoates have been used as tracers for the movement of water in soil and groundwater, so that their degradability as well as their potential toxicity to plants is important.A study using 11 crop species and pentafluorobenzoate, 2,6-difluoro- and 3,4-difluorobenzoate showed their generally low toxicity [20]; clearly their application depends on their recalcitrance. 2. A range of synthetic compounds that inhibit insect growth incorporate the – NH.CO.NH.CO.[2,6-difluorobenzene] – structure that is synthesized from 2,6-difluorobenzoate. The metabolism of 2,5-difluorobenzoate and 3,5-difluorobenzoates by Pseudomonas putida strain JT103. has been examined by 19F NMR [30]. 1,2Dioxygenation took place with both substrates (Fig. 16a,b), and for the 2,5-difluorobenzoate this dominated the alternative 1,6-dioxygenation and resulted in the production of 4-fluorocatechol that was then degraded further with loss of the additional fluorine substituent (Fig. 16 b). This pathway is used by Pseudomonas aeruginosa JB2 for the degradation of the analogous 2,5-dichlorobenzoate that also degraded 2-, and 3-chlorobenzoates [93].

Degradation and Transformation of Organic Fluorine Compounds

163

Fig. 16. a Transformation of 3,5-difluorobenzoate. b Degradation of 2,5-difluorobenzoate by

Pseudomonas putida JT 102

4-Fluorophenylacetate

A strain of Pseudomonas sp. was capable of growth with 4-fluorophenylacetate and fluoride was released into the medium [86].A number of metabolites were found in the medium including D-(+)-fluorosuccinate and trans-3fluoro-hex-3-enedicaboxylate. It is not clear, however, what metabolic pathway can be reconciled with these structures although some can be excluded including the dioxygenase (dechlorinating) pathway that is used for 4-chlorophenylacetate by Pseudomonas sp. strain CBS3 (references in [213]). Although a pathway was presented (Fig. 17) [87], there remain rather serious objections to this. 3-Fluorophthalate

The various pathways for the degradation of phthalate have been reviewed [191] and all of them involve decarboxylation with production of 3,4-dihydroxybenzoate. The metabolism of 3-fluorophthalate was examined in strains of Pseudomonas testosteroni that could not use this for growth [144]: both strains metabolized the substrate without loss of fluorine by the pathway used for the non-fluorinated substrate, whereas the mutant that lacked ring fission enzymes did not carry out decarboxylation (Fig. 18).

Fig. 17. Degradation of 4-fluorophenylacetate by Pseudomonas sp

164

A.H. Neilson · A.-S. Allard

Fig. 18. Transformation of 3-fluorophthalate by Pseudomonas testosteroni

3.4.2 Degradation Under Denitrifying and Anaerobic Conditions

The degradation of benzoate under anaerobic denitrifying conditions has been studied in detail and involves formation of the CoA thioester by a ligase followed by reduction of the ring, hydroxylation, and dehydrogenation before fission. This pathway for both denitrifying and anaerobic phototrophic bacteria has been reviewed [90] while the care required in extrapolating these mechanisms to other groups of organisms has already been noted. Degradation of 2-fluorobenzoate under denitrifying conditions has been reported several times over twenty years [202, 218, 227]. It was shown that, by analogy with benzoate itself, benzoyl-coenzyme A synthetase was induced by 2-fluorobenzoate in Pseudomonas sp. strain KB740 under anaerobic conditions and also by 3- and 4-fluorobenzoate that were not however degraded [202]. A number of bacterial strains isolated by enrichment with halobenzoates were examined for growth under aerobic conditions and under denitrifying conditions [218] and a rather complex pattern of assimilation was observed: 3fluorobenzoate was not used under anaerobic conditions, and 2- and 4-fluorobenzoate often under both aerobic and anaerobic conditions – though sometimes only anaerobically (Table 6). This difference was confirmed in a study with a strain designated Thauera aromatica genomovar chlorobenzoica [219]. The fate of the fluorine was not resolved in these studies, and different mechanisms apparently operate under aerobic or denitrifying conditions. Importantly, all the fluorobenzoate isomers were recalcitrant under sulfate-reducing, ironreducing, and methanogenic conditions [237]. This is in contrast to the anaerobic degradation of 3-chlorobenzoate that has been extensively investigated in the sulfate-reducing Desulfomonile tiedjei that synthesizes ATP by dehalogenation [133], and which produces benzoate by reductive dehalogenation of 3-chlorobenzoate, 3-bromobenzoate, and 3-iodobenzoate – but not 3-fluorobenzoate [50].

165

Degradation and Transformation of Organic Fluorine Compounds

Table 6. Substrate utilization of fluorobenzoates by various strains under aerobic and denitri-

fying conditions Strain

2 FB 11 4 FB 11 2 FB 6 4 FB 10

Substrate 2-Fluorobenzoate

3-Fluorobenzoate

4-Fluorobenzoate

Aerobic

Denitrifying

Aerobic

Denitrifying

Aerobic

Denitrifying

+ + + –

+ + + +

+ – – –

– – – –

+ + – –

– + – +

+ growth; – no growth.

3.5 Fluorinated Phenolic Compounds 3.5.1 Aerobic Conditions Introduction

Phenols are degraded aerobically by initial oxidation to catechols followed by ring fission. The oxidation (hydroxylation) is carried out by monooxygenases that are generally flavoproteins [83]. This pathway is used for chlorophenols with less than three substituents, and degradation involves ring fission to chloromuconates from which chlorine is lost: for example the degradation of 2,4-dichlorophenol proceeds with initial formation of 3,5-dichlorocatechol and ring fission to 2,4-dichloromuconate from which chlorine is successively lost to produce 3-ketoadipate. From pentachlorophenol, however, loss of chlorine takes place both by ring hydroxylation and by reductive dechlorination to produce 2,6dichlorohydroquinone before ring fission. These reactions have been briefly summarized [103, 161]. In the following, it is convenient to distinguish fluorophenols with or without carboxyl groups since details of the structure of the enzyme involved are available. Fluorinated Phenols Lacking a Carboxyl Group

An important use of fluorophenols is the synthesis of precursors used in liquid crystals that incorporate a range of fluorinated components including those derived from 3,4-difluorophenol and 2,3-difluoro-4-heptoxy-phenol. Bacterial metabolism. A number of studies have used various strains of Rhodococcus sp. and Mycobacterium sp. to examine the metabolism of fluorophenols and have illustrated important alternatives:

166

A.H. Neilson · A.-S. Allard

Fig. 19a,b. Transformation of fluorophenols by Rhodococcus sp. strain 1 cp

1. The metabolism of pentafluoro-, pentachloro-, and pentabromophenol by Mycobacterium fortuitum strain CG-2 is initiated by a monooxygenase that carries out hydroxylation at the para position [230], and cell extracts of Rhodococcus chlorophenolicus (Mycobacterium chlorophenolicum) strain PCP-1) in the presence of reductant transformed tetrafluoro-, tetrachloro-, and tetrabromohydroquinone to 1,2,4-trihydroxybenzene by reactions that clearly involve both hydrolytic and reductive loss of fluorine [231]. This contrasts with the metabolism of pentachlorophenol by a Flavobacterium sp. that involves glutathione-specific reductive dechlorinations [254]. 2. Whole cells of Rhodococcus opacus strain 1cp were used to study the metabolism of the fluorophenol isomers [65], and it was shown that both fluorocatechols and fluoropyrogallols were produced (Fig. 19a). It was tentatively suggested that further metabolism of 5-fluoropyrogallol produced 2-pyrone4-fluoro-6-carboxylate that is analogous to that whereby 5-fluoro-3,4-dihydroxybenzoate is converted to 2-pyrone-4,6-dicarboxylate by 3,4-dihydroxybenzoate 4,5-dioxygenase in pseudomonads (Fig. 19b) [114]. Table 7. Metabolites of fluorophenols produced by Rhodococcus corallinus strain 135. Relative contributions (%) are given in parentheses

Substrate

Metabolites (%)

2,3-Difluorophenol

3,4-Difluorocatechol (23)

2,5-Difluorphenol

4-Fluorocatechol (9) 3,6-Difluorocatechol (12) 3,5-Difluorocatechol (78) 3,4,6-Trifluorocatechol (2)

2,3,5-Trifluorophenol

2-Fluoromuconate (24) 2,3-Difluoromuconate (1) 2,5-Difluoromuconate (74) 2,4-Difluoromuconate (1) 2,3,5-Trifluoromuconate (5)

Degradation and Transformation of Organic Fluorine Compounds

167

Fig. 20. Transformation of 2,3-difluorophenol by Rhodococcus opacus strain 135

3. A 19F NMR study using a number of species of Rhodococcus and a range of mono-, di-, and trifluorophenols illustrated the occurrence of both hydroxylation and hydrolytic defluorination (Table 7) (Fig. 20a,b) [18]. 4. The transformations carried out by Rh. corallinus strain 135 illustrate important features including ring fission of the 3-fluorocatechol formed by hydrolytic defluorination of 2,3-difluorophenol to the major fission product 2-fluoromuconate, and of 2,3,4-trifluoromuconate from 2,3,4-trifluorophenol by Rh. erythropolis strain 1CP. It is clear that the metabolism for fluorophenols follows essentially hydrolytic dechlorination that is well established in chlorinated phenols as the first stage in the degradation of pentachlorophenol by Flavobacterium sp. [253] of 2,4,6-trichlorophenol by Azotobacter sp. strain GP1 [248] and of 2,4,5trichlorophenol to 1,2,4-trihydroxybenzene by Burkholderia cepacia strain AC1100 [44]. As an alternative to degradation with loss of fluoride, extensive studies on the bacterial O-methylation of halogen-substituted phenols showed that pentafluorophenol [2] and pentafluorothiophenol [160] formed the corresponding Omethyl ethers during incubation with cell suspensions or cell extracts of strains of Rhodococcus sp. and Pseudomonas sp. Metabolism in yeasts and fungi. The metabolism of fluorophenols has been examined quite extensively in yeasts, and some characteristic features of hydroxylase activity have emerged: 1. The reaction of phenol hydroxylase from the yeast Trichosporon cutaneum with a number of substituted phenols has been examined [165], and it was shown that when the ratio of NADPH to phenol does not exceed 1.0, the consumption of O2 is diverted to the production of H2O2 (Table 8). The production of H2O2 in the absence of an acceptable substrate is typical of a number of flavoprotein hydroxylases [164] and occurs when hydroxylation of the substrate is insufficient to decompose the oxygenated enzyme-flavin intermediate. This has also been observed with other flavoproteins including salicylate hydroxylase, 4-hydroxyphenylacetate hydroxylase, 2,4-dichlorophenol hydroxylase, and 2,4,6-trichlorophenol monooxygenase.

168

A.H. Neilson · A.-S. Allard

Table 8. Diversion of O2 consumption to H2O2 production by phenol hydroxylase from a range

of phenols Substrate

Ratio NADPH/phenol

% of O2 consumed in H2O2

Phenol Fluorophenols 4-Methylphenol 2- and 3-Methylphenol 2-Chlorophenol

0.25–0.40 0.20–0.40 0.80 0.50–0.60 0.25–1.00

10 8–23 20 47–75 100

2. The metabolism of a range of fluorophenols containing up to five fluorine substituents was examined using phenol hydroxylase from T. cutaneum [178]. Illustrative examples are given in Table 9. A 19NMR study [177] of the hydroxylation of fluorophenols by phenol hydroxylase from the same organism showed ortho hydroxylation of 2- and 4-fluorophenol to the corresponding, catechols and that for 3-fluorophenol ortho hydroxylation produced both 3- and 4-fluorocatechol and that the ratio of these was higher with increasing pH. 19

F NMR was used to examine the metabolites produced by the fungus Exophilia jeanselmei. Hydroxylation was observed for 4-fluorophenol, 2,4-difluorophenol, and 2,3,4-trifluorophenol, and the muconates produced from purified phenol hydroxylase and catechol 1,2-dioxygenase are given in Table 10 [17].

Table 9. Fluorinated catechols synthesized from fluorophenols by phenol hydroxylase from

Trichosporum cutaneum Catechol products

Phenolic substrates

4-Fluorocatechol 3,4-Difluorocatechol 3,4,5-Trifluorocatechol 3,4,5,6-Tetrafluorocatechol

3- and 4-Fluoro; 2,4- and 2,5-difluorophenol 2,3- and 3,4-Difluoro; 2,3,6-trifluorophenol 2,3,4- and 3,4,5-Trifluorophenol Pentafluorophenol

Table 10. Fluorinated muconates formed from fluorophenols by phenol hydroxylase and cate-

chol 1,2-dioxygenase from Exophilia jeanselmei Fluoromuconate metabolite

Phenolic substrate(s)

2-Fluoromuconate 2,4-Difluoromuconate 2,3,4-Trifluoromuconate Tetrafluoromuconate

2- and 3-Fluoro; 2,3- and 2,6-difluorophenol 2,4- and 3,5-Difluoro; 2,3,5-trifluorophenol 2,3,4- and 3,4,5-Trifluorophenol Pentafluorophenol

169

Degradation and Transformation of Organic Fluorine Compounds

Fluorinated 4-Hydroxybenzoates

It is convenient to discuss fluorinated 4-hydroxybenzoates separately since considerable effort has been devoted to the role of 4-hydroxybenzoate hydroxylase in their metabolism. The enzyme is a flavoprotein required for the metabolism of 4-hydroxybenzoate to 3,4-dihydroxybenzoate, and the structure of 4-hydroxybenzoate hydroxylase from Pseudomonas fluorescens has been determined by X-ray crystallography [210–212]. This makes it possible to confirm a mechanism for the mechanism put forward earlier (Fig. 21a,b) [48] for the reactions and for their stereoselectivity. Enzymatic studies have also confirmed the significance of the greater consumption of NADPH when fluorinated substrates rather than their natural substrates are hydroxylated (Fig. 21c). (1) 4-hydroxybenzoate hydroxylase from Pseudomonas fluorescens has been used to demonstrate hydroxylation or hydrolytic defluorination of 3,5-difluoro-4-hydroxybenzoate and 2,3,5,6-tetrafluoro-4-hydroxybenzoate: the stoichiometries of the reactions are given in Table 11 [101].A rationalization of the mechanism is given in Fig. 21a,b based on the established structure of the enzyme that is noted above. Substrate activation of the enzyme has been studied in mutants Tyr201 → Phe and Tyr385 → Phe in P. fluorescens and presents a plausible scheme for proton transfer [61], while 19F NMR studies have shown more extensive hydroxylative defluorination of 2,3,5,6-tetrafluoro-4-hydroxybenzoate (Fig. 22) by the active-site mutant Y385F of P. fluorescens [235]. Table 11. Stoichiometries of the action of 4-hydroxybenzoate hydroxylase from Pseudomonas fluorescens on fluorinated 4-hydroxybenzoates

Substrate

4-Hydroxybenzoate 3-Fluoro-4-hydroxybenzoate 3,5-Difluoro-4-hydroxybenzoate 2,3,5,6-Tetrafluoro-4-hydroxybenzoate

Ratios NADPH/ substrate

NADPH/O2

1.0 1.06 2.0 2.0

1.0 1.04 2.0 2.0

Fluoride/ substrate

0.9 0.95

Table 12. Hydroxylative decarboxylation by 4-hydroxybenzoate 1-hydroxylase from Candida

parapsilosis Substrate

Product

2-Fluoro-4-hydroxybenzoate 3-Fluoro-4-hydroxybenzoate 2,5-Difluoro-4-hydroxybenzoate 2,6-Difluoro-4-hydroxybenzoate 3,5-Difluoro-4-hydroxybenzoate 2,3,5,6-Tetrafluoro-4-hydroxybenzoate

2-Fluoro-1,4-dihydroxybenzene 2-Fluoro-1,4-dihydroxybenzene 2,5-Difluoro-1,4-dihydroxybenzene 2,6-Difluoro-1,4-dihydroxybenzene 2,6-Difluoro-1,4-dihydroxybenzene 2,3,5,6-Tetrafluoro-1,4-dihydroxybenzene

170

A.H. Neilson · A.-S. Allard

(a)

Fig. 21. a NADPH-dependent-transformation of 3-fluoro-4-hydroxybenzoate to 4-carboxybenzo[1,2]quinone by 4-hydroxybenzoate hydroxylase with elimination of fluoride. b Reduction of the quinone to the hydroquinone by NADPH. c Stoichiometry of the transformation of 3-fluoro-4-hydroxybenzoate to 3,4-dihydroxybenzoate and comparison with the reaction with the non-fluorinated natural substrate 4-hydroxybenzoate (redrawn from Detmer and Massey 1985 [48])

Degradation and Transformation of Organic Fluorine Compounds

171

Fig. 22. Hydroxylation of 2,3,5,6-tetrafluoro-4-hydroxybenzoate by Pseudomonas fluorescens strain Y 385F

(2) For the sake of completeness an alternative transformation that is catalyzed by 4-hydroxybenzoate 1-hydroxylase (decarboxylating) should be noted. This enzyme that is a flavoprotein has been isolated and purified from the yeast Candida parapsilosis strain CBS604 [60].A number of fluorinated substrates were metabolized and their structures demonstrated by 19F NMR (Table 12) and the mechanism proposed resembled that for the activity of 4-hydroxybenzoate hydroxylase involving FADH peroxide. It is worth noting that this organism is unusual among yeasts in being able to assimilate both 4-hydroxybenzoate and 3,4-dihydroxybenzoate [40]. 3.5.2 Anaerobic Conditions Anaerobic Metabolism of Phenols

The metabolism of phenols under anaerobic conditions is completely different from that under aerobic conditions and has been examined under denitrifying, sulfate-reducing, Fe(III)-reducing, and anaerobic non-methanogenic conditions. Collectively, it is plausible to suggest as a common pathway the one that has been elucidated for denitrifying bacteria. This involves several stages: (i) activation of phenol by formation of phenylphosphate, (ii) carboxylation at a position para to the hydroxyl group [23], and (iii) dehydroxylation of the 4-hydroxybenzoyl CoA to benzoate [22]. This is then further degraded by the reactions that have been noted in Sect. 3.4.2 [90]. It is worth noting that aniline is metabolized by Desulfobacterium anilinii by an apparently analogous reaction involving anaerobic deamination [208]. Consistent with this pathway, successive carboxylation and dehydroxylation reactions have been demonstrated with 2-fluorophenol and 3-fluorophenol, and produce 2-fluorobenzoate and 3-fluorobenzoate [69]. Under methanogenic conditions, 6-fluoro-3-methylphenol was degraded to 3-fluorobenzoate via 5-fluoro4-hydroxy-2-methylbenzoate and 3-fluoro-4-hydroxybenzoate (Fig. 23) [131]. In contrast to chlorophenols for which reductive dechlorination is described in Desulfitobacterium frappieri [47], reductive loss of fluorine has not apparently been observed. The use of fluorophenols to study the anaerobic metabolism of phenols is noted in Sect. 4.2.

172

A.H. Neilson · A.-S. Allard

Fig. 23. Anaerobic transformation of 6-fluoro-3-methylphenol

3.6 Fluorinated Anilines

A strain of Moraxella sp. strain G is able to grow with a number of halogenated anilines using them as a source of both carbon and nitrogen [259]. Aniline oxygenase was induced, and the pathway for the degradation of 4-chloroaniline was deduced using mutants to establish the formation of chlorocatechol and the synthesis of catechol 1,2-dioxygenase and the cis,cis muconate lactonizing enzyme. Since 4-fluoroaniline was used for growth it may plausibly be presumed that the same pathway occurs (Fig. 24). The deamination to catechol is apparently carried out in strain of Nocardia by a dioxygenase [9] although details of the enzyme are not fully resolved [67]. 3.7 Aromatic Trifluoromethyl Compounds

Compounds bearing trifluoromethyl substituents have a wide range of application and, consistent with their chemical stability, the strongly electron-attracting trifluoromethyl group is recalcitrant to defluorination. During the metabolism of aromatic compounds with trifluoromethyl substituents this group therefore generally remains intact even after ring fission. Although loss of fluoride does not occur with aromatic trifluoromethyl arenes, the trifluoromethyl group apparently does not inhibit either ring dioxygenation or ring fission. It is the recalcitrance of the ring-fission product that is responsible for the only partial degradation of these substrates. The following illustrative examples are given. 1. The cis-dihydrodiols may be produced from a number of trifluoromethyl benzoates: a) 4-Trifluoromethylbenzoate by Pseudomonas putida strain JTIO7 (Fig. 25a) [46]. b) 2-Trifluoromethylbenzoate by P. aeruginosa strain 142 (Fig. 25b) [215]. c) 4-Methyl- and 4-iodotrifluorobenzene by P. putida strain UV-4 [21].

Fig. 24. Degradation of 4-fluoroaniline by Moraxella sp. strain G

Degradation and Transformation of Organic Fluorine Compounds

173

Fig.25a,b. Dioxygenation of: a 4-trifluoromethylbenzoate by Pseudomonas putida strain JT 107; b 2-trifluoromethylbenzoate by Pseudomonas aeruginosa strain 142

2. The terminal product 2-hydroxy-6-keto-7,7,7-trifluorohepta-2,4-dienoate that is the ring-fission product retains the trifluoromethyl group of the original substrates and is produced both from trifluoromethyl benzoates and trifluoromethyl phenols: a) 4-Trifluoromethyl benzoate was co-metabolized by 4-isopropylbenzoate grown cells of P. putida strain JT (Fig. 26a) [57]. b) 3-Trifluoromethylbenzoate was co-metabolized by 3-methylbenzoate grown cells of P. putida strain mt-2 (Fig. 26b) [59].

Fig. 26a,b. Metabolism of production of: a 4-trifluoromethylbenzoate by Pseudomonas putida strain JT grown with 4-isopropylbenzoate; b 3-trifluoromethylbenzoate by Pseudomonas putida strain mt-2 grown with 3-methylbenzoate

Fig. 27. Transformation of 2-trifluoromethylphenol by Bacillus thermoleovorans

174

A.H. Neilson · A.-S. Allard

c) The hydroxylation of 2-trifluoromethylphenol to 3-trifluoromethylcatechol by Bacillus thermoleovorans strain A2 was followed by extradiol ring fission (Fig. 27) [188] and the transformation of 3-trifluoromethyl catechol [58] produced the stable terminal metabolite.

4 General Metabolism and Application of Fluorinated Compounds 4.1 Miscellaneous Reactions

A few further examples are given of fluorinated substrates in which, although the fluorine atom is not removed during the reaction, the presence of fluorine substituents does not inhibit microbial transformation: 1. Reduction of perfluoroalkylated ketones and of the double bond in perfluoroalkenoic acid esters has been accomplished with baker′s yeast (Saccharomyces cerevisiae), and the products had a high optical purity ranging from 67% to 96% [120]. 2. Whereas it is hardly to be expected that fluorine substituents will transmit stereochemical or electronic effects to sites remote from the fluorine atoms, they may alter the products of a reaction by blocking the normal sites. For example, the quaternary 9-methyl group of 5,5-difluorocamphor is hydroxylated to the 9-hydroxymethyl compound [52] (Fig. 28 a). 3. On the other hand, although both enantiomers of bicyclo[2.2.1]heptan-2-one are oxidized to the lactones by Acinetobacter strain NCIB 9871, the presence of both bromine and fluorine substituents makes it possible to obtain a single enantiomer of the ketone and of the lactone that was subsequently converted by chemical means to fluorinated nucleoside analogues (Fig. 28b) [129].

Fig. 28a,b. Transformation of: a 5,5-difluorocamphor; b the substituted bicyclo[2.2.1]heptan2-one by Acinetobacter strain NCIB 9871

Degradation and Transformation of Organic Fluorine Compounds

175

4.2 Inhibitors and Mutagenizing Agents

Fluorinated substrates have been used to obtain mutants, as inhibitors to study bacterial metabolism, and in biosynthetic studies that use the fluorine as a label. 4.2.1 Inhibitors

The use of fluoromethane as an inhibitor of ammonium oxidation and of difluoromethane as an inhibitor of methanotrophs has been noted in Sect. 2.1.1.As an example of their application, inhibition of the oxidation of bromomethane and dibromomethane by methyl fluoride was observed only in bacteria from freshwater samples and not those from saline habitats [78]. It was therefore concluded that different nonmethanotrophic organisms were responsible for the oxidations. Fluorinated compounds – frequently related to growth substrates – may be toxic to growth, for example, 4-, 5-, or 6-fluorotryptophan to E. coli [25], and 3fluorotyrosine to E. coli, Streptococcus faecalis, and Lactobacillus plantarum. These analogues may inhibit specifically the activity of enzymes normally used for metabolism of the non-fluorinated substrates; for example 4-fluorohistidine is a competitive inhibitor of histidine – ammonia lyase that produces urocanate from histidine by elimination of ammonia. It has already been noted in Sect. 3.4.1 that 3-halogenated catechols inhibit the activity of catechol 2,3-dioxygenases by different mechanisms. These enzymes carry out the preferred extradiol pathway for the degradation of non-halogenated catechols, and 3-fluorocatechol is generally more effective as an inhibitor than 3-chlorocatechol. It has been used to establish metabolic pathways by inhibiting further reactions of the catechols, for example the metabolism of both chlorobenzene and toluene – that are normally incompatible – by Pseudomonas putida strain GJ31 [142], and of styrene by Rhodococcus rhodochrous strain NCIMB 13259 that involves the less-common ring-dioxygenation pathway [242]. Fluorophenols have served as inhibitors and been used to delineate the pathway for the anaerobic metabolism of 3-methylphenol by a methanogenic consortium [132] (Fig. 23): – 6-Fluoro-3-methylphenol brought about the accumulation of 4-hydroxy-2methylbenzoate. – 3-Fluorobenzoate resulted in the transient accumulation of 4-hydroxybenzoate. – Addition of bromoethanesulfonate inhibited the synthesis of methane and caused the accumulation of benzoate. The corresponding fluorinated metabolites were identified, and this enabled construction of the complete pathway for the degradation that included the unusual reductive loss of a methyl group from the aromatic ring that has been noted in Sect. 3.5.2.

176

A.H. Neilson · A.-S. Allard

4.2.2 Mutagenizing Agents

Fluorinated substrates have been used to obtain mutants of aerobic organisms degrading aromatic compounds. For example, mutants of Pseudomonas putida strain U that were defective in the utilization of phenol or cresols were obtained using 3-fluorophenol and 4-fluorophenol, while a greater range of fluorinated substrates was used for other pairs of substrates (Table 13) [249]. Orotic acid is a direct precursor of uridine phosphate and its biosynthesis from cytosine provides a salvaging route. 5-Fluoroorotic acid has been used to obtain pyrE and pyrf mutants of Sulfolobus acidocaldarius that encode orotate phosphoribosyltransferase and orotidylate decarboxylase respectively. These have been used to study rates of forward mutation of these mutants and thereby estimate the genetic stability of the DNA in this extreme thermophile [104]. 4.2.3 Fluorine as Label in Biosynthesis

The application of 19F NMR has been discussed by Stanley in Chap. 1, and examples of its application may be found in this review. Synthetic fluorinated substrates that can be accepted by the natural enzyme have proved valuable in studies of biosynthesis. The classic example of the application of fluorinated substrates to elucidate details of biochemical reactions was made during investigations into the terpene biosynthetic pathway in which rates of solvolysis of methanesulfonates containing fluorine or trifluoromethyl substituents in the substrates [179, 180] were examined. Developments in procedures for the synthesis of a wide range of potential precursors [104] make this an increasingly attractive procedure. 4.3 Biotechnology

The degradation of aromatic compounds via initial formation of cis-dihydrodiols has been extensively studied in mutants of the parent strains, and it is then Table 13. Fluorinated analogues used to obtain mutants of their non-fluorinated analogues in

strains of Pseudomonas putida Fluorinated analogue

Substrate

5-Fluorosalicylate 3-Fluoro-4hydroxybenzoate 3-Fluoro-4methoxybenzoate

Salicylate 4-Hydroxybenzoate 4-Methoxybenzoate

Concentration (mmol/l) 5 2.5 10

No. of defective mutants relative to no. of survivors

Strain of Pseudomonas putida

94/94 33/152

P. putida AC-36 P. putida U

91/94

P. putida JT 101

Degradation and Transformation of Organic Fluorine Compounds

177

possible to accomplish dioxygenation without interference from the products of further metabolism. This reaction has been used to produce a number of valuable starting materials in biotechnology [192]. In practice, a combination of microbiological and chemical reactions has most often been used.A few illustrative examples include the following: 1. Transformation of fluorobenzene to 3-fluorocatechol by chemical or microbiological oxidation of the intermediate dihydrodiol produced by a mutant of Pseudomonas putida or directly with P. putida strain ML2 [135]. 2. Pseudomonas putida strain UV4 converts 4-iodofluorobenzene to cis-dihydrodiols that can be reduced chemically to the corresponding fluoro compounds [3]. 3. Fluorobenzene has been converted by Pseudomonas strain T-12 to the catechol and methylated chemically to 3-fluoroveratrole [106]. This compares with the four-step synthesis of the corresponding guaiacol from 2-fluorobenzene. 4. A mutant of P. putida strain JT 103 transformed 3,4-difluorobenzoate and 3,5difluorobenzoate to the 3,4-, 4,5-, and 3,5-difluorocyclohexadiene-cis-1,2-diol1-carboxylates [198]. 4.4 Non-Microbial Biochemical Reactions

In most of the examples that have been discussed the focus has been directed to bacterial degradation. In higher organisms, including fungi and yeasts, biotransformation is the dominant reaction and this is often the first step in detoxification by introducing hydroxyl groups into aromatic compounds followed by conjugation to form water-soluble sulfate esters or glucuronides that can be excreted. Fluorinated aromatic compounds can be metabolized (transformed) by higher organisms and this is also frequently accomplished by introduction of hydroxyl groups into the substrates. A few examples are given as illustration: 1. Enzymes from higher organisms have been used to examine the metabolism of fluoroanilines. a) The peroxidase-catalyzed oxidation of 4-fluoroaniline resulted in the elimination of fluoride and formation of the bis(4-fluorophenyl)-2-amino-5-(4fluoroanilino)-benzo[1,4]quinonimine [99]. b) Hepatic aryl hydroxylase produced 4-aminophenol from both 3- and 4-fluoroaniline [190], and the formation of this isomer from 3-fluoroaniline may plausibly be attributed to an NIH shift [43]. c) The principal reaction from cytochrome P450 in microsomes prepared from rat liver was hydroxylation at the position para to the fluorine substituents, and this increased relatively with the degree of fluorine substitution (Table 14) [194]. The ortho and para aminophenols were autoxidized in air. 2. The transformation of 4-fluorophenylalanine to tyrosine by liver enzymes is accomplished not by a hydroxylase but by a flavin monooxygenase, and this is supported by the NADP+/tyrosine ratio of 2.0 compared with the ratio of 1.0

178

A.H. Neilson · A.-S. Allard

Table 14. Hydroxylative defluorination of fluorinated anilines by liver microsomes (nmol/10 min/nmol cytochrome P450

Substrate

Fluoride anion

4-hydroxyaniline metabolite

4-Fluoroaniline 2,4-Difluoroaniline 2,4,6-Trifluoroaniline Pentafluoroaniline

8 15 30 62

2.0 2.3 1.1 1.6

when phenylalanine is used as the substrate [112]. This phenomenon has already been noted in Sect. 3.5.1. 3. trans-4-Fluoroproline is incorporated into collagen and is enzymatically hydroxylated in the bound state to 4-hydroxyproline that was recovered after hydrolysis by collagenase [79].

5 Enzyme Inhibition by Fluorinated Analogue Substrates 5.1 Introduction

Biochemical mechanisms of toxicity have been supplemented by the results of studies using substrates that function as inhibitors through irreversible inactivation of enzyme activity. Although these surrogate substrates are sufficiently similar to natural substrates to be accepted by the enzymes, they undergo reactions that result in covalent association and loss of enzyme activity. Fluorinated substrates have been particularly attractive on account of the small size of the fluorine atoms and the reactivity of the substrate-enzyme complexes. In the context of biodegradation, suicide metabolism resulted from the products of enzymatic reaction (Sect. 3.4.1) but may generally occur with compounds that mimic the natural substrate for the enzyme. Rational drug design seeks to produce compounds that deactivate specific enzymes associated with dysfunction and disease, and has been explored in various books (e.g., [179, 245]). Intense attention has naturally been focused on mammalian enzymes, but the same principles have been applied to microbial enzymes. Indeed the paradigm for ribonucleotide reductase is the enzyme from Escherichia coli. In the following discussion, attention is directed to microbial enzymes though some mammalian enzymes are included. The classic example of fluoroacetate [239] has been noted in Sect. 2.2.1. In another context, DeFrank and White discuss in Chap. 10 the development of irreversible inhibitors of acetylcholinesterase. Substantial effort has been directed to establishing plausible mechanisms for inactivation that generally involves loss of fluoride concomitant with irreversible binding of the residue to the enzyme. This is an enormous field in view of its importance in drug design. This lies beyond the expertise of the authors and only a few examples that illustrate the principles are addressed.

Fig. 29. Mechanism of inhibition of alanine racemase by fluorinated alanine (redrawn from Wang and Walsh 1981 [241])

Degradation and Transformation of Organic Fluorine Compounds

179

180

A.H. Neilson · A.-S. Allard

5.2 Alanine Racemase D-Alanine

is required for incorporation into bacterial cell walls and is formed from biosynthetically produced L-alanine by alanine racemase. The enzyme from Escherichia coli strain B is a dimer containing one molecule of pyridoxal phosphate per subunit [240]. In enzymes that contain pyridoxal entities, the inactivation proceeds by reaction of the fluorine-containing amine with formation of an intermediate enamine (Fig. 29). This is illustrated by the reaction of b-fluoroalanine and b,b,b-trifluoroalanine with alanine racemase. The observed pathways for both D- and L-fluoroalanine in which racemization does not take place are accommodated by a mechanism in which the enzyme may produce either pyruvate (reaction 1) or the alkylated enzyme (reaction 2) that results in enzyme deactivation via a common enamino complex (Fig. 29). Further investigations [241] used b,b-difluoroalanine and b,b,b-trifluoroalanine to add further supporting details in which 2-fluoropyruvate and 2,2′-difluoropyruvate were the products of the elimination reactions. In these cases, however, only b,b,b-trifluoroalanine led to inactivation whereas b,b-difluoroalanine produces labile intermediates that regenerate the enzyme. A comparable mechanism applies to other enzymes including decarboxylases [121] (Fig. 30) and, for example, b-fluoromethylornithine is an inhibitor of ornithine decarboxylase in E. coli [108]. It is worth pointing out that the reverse reaction has been proposed [147] as a possible sequence for the biosynthesis of fluoroacetate without enzymological evidence although this is contradictory to the established ease of loss of fluoride and the limited anionoid character of fluoride. The enzymatic reaction between 2,4-dinitrophenyl-mannoside and fluoride catalyzed by the b-mannoside from by a mutant of Cellulomonas fimi illustrates the possibility of such reactions [258].

Fig.30. Mechanism of inhibition of decarboxylases (redrawn from Kollonitsch et al. 1978 [121])

Degradation and Transformation of Organic Fluorine Compounds

181

Fig. 31. Mechanism of inhibition of ribonucleoside diphosphate reductase by 2′deoxy-2′(fluoromethylene) cytidine diphosphate (redrawn from McCarthy et al. 1996 [146])

5.3 Ribonucleoside Diphosphate Reductase

The enzyme plays a key role in DNA synthesis and catalyzes the rate-determining reduction of ribonucleotides to deoxyribonucleotides and has been the subject of review [223]. The enzyme has been divided into classes depending on the mechanism whereby the thiyl radical is generated, and in Escherichia coli that is the paradigm of class I reductases, the enzyme consists of two subunits R1 and R2. R1 contains the catalytically-required cysteine groups [137], while R2 is an iron-sulfur protein containing the tyrosyl radical µ-oxo bridged peroxy binuclear iron center [168] that may generate the thiyl radical in R1 through proton and electron transfers [222]. The inhibition by 2′-deoxy-2′-(fluoromethylene) cytidine diphosphate has been rationalized on the basis of the reactions in Fig. 31 [146]. 5.4 Ornithine Decarboxylase

Putrescine is required for the synthesis of intracellular particles and is produced by decarboxylation of ornithine. The effect of two irreversible inhibitors of ornithine decarboxylase has been examined in representatives of different groups of bacteria. DL-Difluoromethylornithine inhibited the activity of ornithine decarboxylase from Pseudomonas aeruginosa, though not the enzyme from Escherichia coli or Klebsiella pneumoniae [107]. Conversely, a-monofluoromethylputrescine inhibited the enzyme from E. coli but not that from P. aeruginosa [108]. It is plausible to attribute these observations to the mechanism pos-

182

A.H. Neilson · A.-S. Allard

Table 15. Fluorinated irreversible inhibitors of selected enzymes

Inhibitor

Enzyme

b-Fluoroalanine a-Fluoromethyl histidinea a-Fluoromethyl ornithine a-Fluoromethyl glutamate a-Fluoromethyl-3,4-dihydroxyphenylalanine

Alanine racemase Histidine decarboxylase Ornithine decarboxylase Glutamate decarboxylase 3,4-Dihydroxyphenylalanine decarboxylase

a

Only mammalian enzyme.

tulated for pyridoxal-dependent decarboxylases including the decarboxylated amines of mammalian histidine carboxylase (histamine) and 3,4-dihydroxyphenylalanine (dopa) [121] (Fig. 30). A list of inhibitors and the corresponding enzymes are given in Table 15 [121, 239]. 5.5 DNA-Cytosine Methyltransferase

In eukaryotes and some prokaryotes, the activated methyl group of S-adenosylmethionine is transferred to the C-5 position of cytosine, and in bacteria thereby prevents the digestion of the host DNA by the host restriction endonucleases. Studies with Haemophilus aegypytus DNA (cytosine-5)-methyltransferase that is overproduced in Escherichia coli have provided a mechanism for the inhibition by a 5-fluoro-2′-deoxycytidine-containing oligonucleotide that brings about methylation but that is inactivated by its inability to eliminate the cysteine residue of the enzyme at C-4 by loss of fluoride in place of a proton from the covalent intermediate that was characterized in detail by determination of the structure of the two peptides obtained from it by proteolysis, reduction and alkylation (Fig. 32) [37]. 5.6 Thymidylate Synthase

In the context of drug discovery, Heidelberger and Santi have extensively investigated the possibility of finding inhibitors of this enzyme. The enzyme reacts with 5,10-methylenetetrahydrofolate and 2′-deoxyuridine monophosphate to

Fig. 32. Mechanism of inhibition of DNA-cytosine methyltransferase by 5-fluoro-2′-deoxycy-

tosine (redrawn from Chen et al. 1993 [37])

Degradation and Transformation of Organic Fluorine Compounds

183

Fig. 33. Inhibition of thymidylate synthase by 5-trifluorovinyl-2′-deoxyuridine diphosphate (redrawn from Walsh 1982 [239])

form a ternary complex that carries out a function comparable to that of DNAcytosine methyltransferase. The irreversible inhibition of the enzyme by 5-substituted-2′-deoxyuridines has been examined, and once again a cysteine residue in the protein reacts at C-4 of the pyrimidine ring. The mechanisms of inhibition by 5-trifluoromethyl-2′-deoxyuridine monophosphate [200] and the vinyl analogue 5-trifluorovinyl-2′-deoxyuridine diphosphate [243] are different from that for the 5-fluoro compound and involve acylation of the enzyme. This may be rationalized by a mechanism involving the formation of an irreversible covalent bond with the enzyme (Fig. 33) [239]. 5.7 Enolpyruvyl Transferase

Muramic acid (3-O-a-carboxyethyl-N-acetylglucosamine) is a component of the peptidoglycan layer of bacterial cell walls. Its synthesis is accomplished by reaction of phosphoenolpyruvate with UDP-N-acetylglucosamine mediated by the enzyme MurZ, followed by reduction of the enolpyruvate by the reductase MurB to N-acetylmuramic acid. The activity of MurZ is inhibited by 3-fluorophosphoenolpyruvate and results in the formation of both an enzyme complex and a fluorinated tetrahedral intermediate (Fig. 34) from which a proton from the fluoromethyl group is not lost [118], and whose stereochemistry has shown that the intermediate from (E)-fluorophosphoenolpyruvate has the (R) configuration and that from the (Z) isomer the (S) configuration [119].

184

A.H. Neilson · A.-S. Allard

Fig. 34. Mechanism of reaction of 3-fluorophosphoenolpyruvate with UDP-N-acetylglucos-

amine (redrawn from Kim et al. 1995 [119])

5.8 Adenosyl-L-Homocysteine Hydrolase

Methionine, S-adenosylmethionine, S-adenosylhomocysteine, and homocysteine are key components in the cycling of activated methyl groups. S-Adenosyl-Lhomocysteine is formed from S-adenosylmethionine during methyl transfer, and is hydrolyzed by adenosyl-L-homocysteine hydrolase to adenosine and Lhomocysteine from which methionine is regenerated from N5-methyltetrahydrofolate. Inactivation of this enzyme may take place by two mechanisms designated type I or type II that have been discussed in the context of inactivation by (E)-5′,6′-didehydro-6′-deoxy-6′-halohomoadenosines [257]. In the first, inactivation of the enzyme is accomplished by irreversible substrate reduction of the enzyme-NAD+ complex to the enzyme-NADH complex. In the second, a chemically active intermediate is formed on the enzyme and then a covalent adduct with the enzyme and a protein. The basic mechanism was established by Palmer and Abeles [174] who showed that the oxidation of the 3′-hydroxyl group by enzyme-bound NAD+ to form the 3′-keto compound was a key reaction: homocysteine was then eliminated with the recycling of enzyme-bound NAD+ and adenosine. The synthetic substrate (Z)-4′,5′-didehydro-5′-deoxy-5′-fluoroadenosine that inactivates the hydrolase has been used to establish the details of the mechanism of the reaction (Fig. 35) [256]. The 5′-carboxaldehyde is formed initially with loss of fluoride by a hydrolase-catalyzed elimination, and is then oxidized by NAD+ to the 3′-keto-5′-carboxyaldehyde by an independent-catalyzed activity in which the enzyme forms strong – though not covalent – bonds with the aldehyde. This is a type I inhibition and the potential acyl fluorides that could deactivate the enzyme by covalent formation are not the source of inhibition as in Type II mechanisms. A comparable mechanism of inactivation

Fig.35. Mechanism of inhibition of adenosyl-L-homocysteine hydrolase by (Z)-4′,5′-didehydro6′-deoxy-6′-fluoroadenosine (redrawn from Yuan et al. 1993 [256])

Degradation and Transformation of Organic Fluorine Compounds

185

Fig. 36. Inhibition by 19,19-difluoroandrost-4-ene-3,17-dione of the aromatizing enzyme required for synthesis of estrone from androgens

has also been observed with (E)-5′,6′-didehydro-6′-deoxy-6′-halogenohomoadenosines (halogen = Cl and Br) [257] in which once again the oxidation to the 3′-keto compound by NAD+ is a cardinal reaction. Further references may be found in [195]. 5.9 Estrogen Synthetase

In mammalian systems, the synthesis of estrogens from androgens is a cardinal reaction in the placenta. The two critical stages in the synthesis of estrone from androst-4-ene-3,17-dione are microsomal hydroxylation at the C-19 position followed by aromatization of ring A with loss of formate. The irreversible inhibition of the aromatizing enzyme from placental microsomes by 19,19-difluoroandrost-4-ene-3,17-dione can be rationalized on the basis of hydroxylation at C-19 followed by elimination of one mol of hydrogen fluoride to form the acyl fluoride that reacts irreversibly with the enzyme (Fig. 36). In contrast, the 19-fluoroandrost-4-ene-3,17-dione reacts normally since the C-19 aldehyde is formed instead of the reactive acyl fluoride and this can then be aromatized to estrone [138]. In conclusion, it is worth drawing attention to two features of these enzyme inactivations: 1. The formation of acylfluorides that bring about deactivation of enzyme activity by forming stable covalent bonds with nucleophilic groups in the enzymes. 2. The significance of other mechanisms that impinge on central metabolism: an illustration is provided by the deflection of NAD+ and NADP from functional oxidation-reduction.

6 Abiotic Transformations Previous sections have been devoted to biochemical mechanisms for fission of the C–F bond, and a number of examples for direct replacement of fluoride have been illustrated. On the other hand, chemical reactions of alkyl fluorides with loss of fluoride are much less facile than with their chlorinated – and especially

186

A.H. Neilson · A.-S. Allard

brominated and iodinated – analogues. Purely chemical reactions involving displacement of fluoride are therefore unusual. Examples have been given in the introduction of the stability of the C–F bond. 6.1 Chemical Reactions

Although two reactions are noted here, it is only in the second group that fission of the C–F bond is accomplished: 1. a) Although nickel (I) octaethylisobacteriochlorin is a strong nucleophile and is effective in the reduction of a wide range of halogenated substrates, fluorotrichloromethane stands out as an example in which the C–F bond remains intact [91]. This recalcitrance has been confirmed in other studies. b) Biotic dehalogenation in organisms containing corrinoids and coenzyme F430 has been discussed in Sect. 2.1.2, and purely chemical models based on these reactions using corrinoids in the presence of a chemical reductant containing Ti(III) have been studied. For example, whereas reductive dechlorination of chloroethenes could be demonstrated, reductive loss of fluorine did not occur [74]. 2. Gas-phase plasma destruction of perfluorinated compounds has been examined: a) Surface wave plasmas have been used to examine the destruction of C2F6 [89], and CF4 and CHF3 [250]. High levels of transformation were observed for both, and the degradation products included CO2 , CO, COF2 , and HF. Although the last two of these could be removed by scrubbing, the ultimate fate of HF was unresolved: optimally it would be recovered. b) Using a different protocol, gas-phase oxidation has been examined. Experiments were been carried out on the destruction of C2F6 using dielectric barrier discharge with atmospheric pressure plasma processing in the presence of O2 [34]. The initial reaction is the production of CF3 radicals that then reacted with oxygen radicals to produce CO2 , CO, and COF2. Combination of the system with a metal oxide catalyst consisting of CuO/ZnO/Al2O3 packed within the discharge region increased the conversion of C2F6 by reactions involving the metal-oxygen bond. 6.2 Photochemical Reactions

In view of the preceding comments, it may be concluded that a determining factor in the susceptibility of organic fluorine compounds to photochemical degradation and transformation is access to the energy that is available by absorption of radiation. For the atmospheric environment this has been discussed in detail for aliphatic compounds by Wallington and Nielsen in Chap. 3 whereas, in aquatic and terrestrial environments, attention has been primarily directed to aromatic compounds.

Degradation and Transformation of Organic Fluorine Compounds

187

In view of concern over possible adverse effects of perfluoro compounds (Ellis et al., Chap. 4), increasing activity has been devoted to physico-chemical methods for their destruction. Although this may seem somewhat out of place in this chapter, a few summary comments are offered. Although the photochemical transformations of organic fluorine compounds in the atmosphere have been presented by Wallington and Nielsen in Chap. 3, some examples of reactions in the gas phase, in non-aqueous and aqueous solution and on surfaces may be given here: 1. The reaction of perfluoroalkyl (hydrofluoroalkyl) methyl ethers (HFEs) with hydroxyl radicals has been discussed by Wallington and Nielsen in Chap. 3, and these reactions may be contrasted with the reactions of these compounds with photochemically-produced chlorine atoms [167]: C2F5OCH3 +Cl →

C2F5OCH2 +HCl

C2F5OCH2 +O2 →

C2F5OCH2O2

2 C2F5OCH2O2 →

2 C2F5OCH2O+O2

C2F5OCH2O+O2 →

C2F5OCHO + HO2

C2F5OCHO + Cl →

C2F5OCO+HCl

C2F5OCO+O2 →

C2F5OCO.O2

2 C2F5OCO.O2 →

C2F5O+CO2

C2F5O →

CF3 +COF2

For the C2 to C5 congeners, degradation with loss of fluorine occurs by the elimination of COF2 , and for the C3 , C4 , and C5 congeners by a reaction involving elimination of HF: CnF2n+1O+HO2 → CnF2n+1OH →

CnF2n+1OH + O2 Cn-1F2n-1COF+HF

2. Nanocrystalline ZnS and CdS in dimethylformamide were used to examine the photoreductive dehalogenation of both chlorinated and fluorinated benzenes in the presence of triethylamine as electron donor [255]. The following defluorinations were observed using the ZnS catalyst irradiated at l<300 nm: a) Hexafluorobenzene → pentafluorobenzene. b) 1,2,4,5- and 1,2,3,4-tetrafluorobenzene → trifluorobenzenes. c) 1,2,4-Trifluorobenzene → difluorobenzene. 3. In the wider context of destruction of pollutants, photochemically-induced dechlorination involving OH radicals in “advanced oxidation processes” may be briefly noted, for example by slurries of TiO2 in which photochemical production of a free electron in the conduction band (e–cb) and a corresponding hole (h+vb) in the valence band produce H2O2 and thence OH radicals [122], and by irradiation of polyoxometalates (heteropolyacids) that generates OH radicals [6].

188

A.H. Neilson · A.-S. Allard

Fig. 37. Phototransformation of enrofloxacin

Fig. 38. Photolysis of 3-trifluoromethyl-4-nitrophenol (Carey and Fox) [29]

4. The following examples illustrate primarily the important issue of the fission of aromatic rings bearing fluorine and trifluoromethyl substituents in aqueous media: a) The photolysis in aqueous solution of fluoroquinolone carboxylic acids including the antibiotics enrofloxacin and ciprofloxacin has been examined and has been evaluated in the specific context of the degradation of drug preparations during storage [229]. The structures of the products have been established [26] and used to propose a pathway for the photodegradation (Fig. 37) [27]. b) The photolysis of 3-trifluoromethyl-4-nitrophenol resulted in the production of 2,5-dihydroxybenzoate from hydrolytic loss of the nitro group and oxidation of the trifluoromethyl group, together with a compound identified as a condensation product of the original compound and the dihydroxybenzoate (Fig. 38) [29]. The degradation of 3-trifluomethylbenzoate was shown to proceed by sequential microbial and photochemical reactions [226]. A number of 3-trifluoromethylphenols were examined

Fig. 39. Photolysis of 3-trifluoromethyl-4-nitrophenol (Ellis and Mabury) [54]

Degradation and Transformation of Organic Fluorine Compounds

189

in buffered aqueous media [54]: 3-trifluoromethylphenol produced fluoride, and 3-hydroxybenzoyl fluoride that was hydrolyzed to 3-hydroxybenzoate, and 3-trifluoro-4-nitrophenol was degraded at pH 7 to give trifluoroacetate in 8% yield by a pathway that is given in Fig. 39. The reactions may be contrasted with biochemical transformations in which ring fission does not occur.

7 Concluding Comments Throughout this overview, comparisons have been made between the microbial reactions of fluorinated substrates and their chlorinated counterparts. In spite of the substantial differences between the atoms in size, electronegativity and electronic configuration – and though in general chlorine is more readily removed than fluorine – it is concluded that their biochemical reactions are broadly comparable. This should be contrasted with the widely different physico-chemical properties of chlorobenzenes and fluorobenzenes (Ellis et al., Chap. 2).

8 References 1. Allard A-S, Hynning P-Å, Remberger M, Neilson AH (2000) Environmental hazard assessment: a laboratory approach to reality? Toxicol Environ Chem 78:127–197 2. Allard A-S, Remberger M, Neilson AH (1987) Bacterial O-methylation of halogen-substituted phenols. Appl Environ Microbiol 53:839–845 3. Allen CR, Boyd DR, Dalton H, Sharma ND, Brannigan I, Kerley NA, Sheldrake GN, Taylor SC (1995) Enantioselective bacterial transformation routes to cis-diol metabolites of monosubstituted benzenes, naphthalene and benzocycloalkenes of either absolute configuration. J Chem Soc Chem Commun 117–118 4. Altenschmidt U, Fuchs G (1992) Anaerobic toluene oxidation to benzyl alcohol and benzaldehyde in a denitrifying Pseudomonas strain. J Bacteriol 174:4860–4862 5. Anders H-J, Kaetzke A, Kämpfer P, Ludwig W, Fuchs G (1995) Taxonomic position of aromatic-degrading denitrifying pseudomonad strains K 172 and KB 740 and their description as new members of the genera Thauera, as Thauera aromatica sp nov, Azoarcus, as Azoarcus evansii sp nov, respectively, members of the beta subclass of the Protobacteria. Int J Syst Bacteriol 45:327–333 6. Androulaki E, Hiskia A, Dimotikali D, Minero C, Calza P, Pelizzetti E, Papaconstantinou E (2000) Light induced elimination of mono- and polychlorinated phenols from aqueous 3– . The case of 2,4,6-trichlorophenol. Environ Sci Technol 34:2024–2028 solutions of PW12O40 7. Anthony C (1982) The biochemistry of methylotrophs. Academic Press, London 8. Au KG, Walsh CT (1984) Stereochemical studies on a plasmid-coded fluoroacetate halohydrolase. Bioorg Chem 12:197–295 9. Bachofer R, Lingens F (1975) Conversion of aniline into pyrocatechol by a Nocardia sp: incorporation of oxygen-18. FEBS Lett 50:288–290 10. Banks RE, Smart BE, Tatlow JC (eds) (1994) Organofluorine chemistry principles and commercial applications. Plenum Press, New York 11. Behrman EJ, Stanier RY (1957) The bacterial oxidation of nicotinic acid. J Biol Chem 228:923–945 12. Behrman EJ, Stanier RY (1957) Observations on the oxidation of halogenated nicotinic acids. J Biol Chem 228:947–953

190

A.H. Neilson · A.-S. Allard

13. Beller HR, Edwards EA (2000) Anaerobic toluene activation by benzylsuccinate synthase in a highly enriched methanogenic culture. Appl Environ Microbiol 66:5503–5505 14. Beller HR, Spormann AM, Sharma PK, Cole JR, Reinhard M (1996) Isolation and characterization of a novel toluene-degrading sulfate-reducing bacterium. Appl Environ Microbiol 62:1188–1196 15. Berends AG, Bouttonet JC, de Rooij CG, Thompson RS (1999) Toxicity of trifluoroacetate to aquatic organisms. Environ Toxicol Chem 18:1053–1059 16. Biegert T, Fuchs G (1995) Anaerobic oxidation of toluene (analogues) to benzoate (analogues) by whole cells and by cell extracts of a denitrifying Thauera sp. Arch Microbiol 163:407–417 17. Boersma MG, Dinarieva TY, Middelhoven WJ, van Berkel WJH, Doran J, Vervoort J, Rietjens IMCM (1998) 19F nuclear magnetic resonance as a tool to investigate microbial degradation of fluorophenols to fluorocatechols. Appl Environ Microbiol 64:1256–1263 18. Bondar VS, Boersma MG, Golovlev EL, Vervoort J, van Berkel WJH, Finkelstein ZI, Solyanikova IP, Golovleva LA, Rietjens IMCM (1998) 19F NMR study on the biodegradation of fluorophenols by various Rhodococcus species. Biodegradation 9:475–486 19. Boutonnet JC et al. (1999) Environmental risk assessment of trifluoroacetic acid. Hum Ecol Risk Assess 5:59–124 20. Bowman RS, Schroeder J, Buluau R, Remmenga M, Heightman R (1997) Plant toxicity and plant uptake of fluorobenzoate and bromide water tracers. J Environ Qual 26: 1292–1299 21. Boyd DR, Sharma ND, Hand MV, Groocock MR, Kerley NA, Dalton H, Chima J, Sheldrake GN (1993) Stereodirecting substituent effects during enzyme-catalysed synthesis of cisdihydrodiol metabolites of 1,4-disubstituted benzene substrates. J Chem Soc Chem Commun 974–976 22. Breese K, Fuchs G (1998) 4-Hydroxybenzoyl-CoA reductase (dehydroxylating) from the denitrifying bacterium Thauera aromatica – prosthetic groups, electron donor, and genes of the molybdenum-flavin-iron-sulfur proteins. Eur J Biochem 251:916–923 23. Breinig S, Schiltz E, Fuchs G (2000) Genes involved in anaerobic metabolism of phenol in the bacterium Thauera aromatica. J Bacteriol 182:5849–5863 24. Breitenstein A, Saano A, Salkinoja-Salonen M, Andreesen JR, Lechner U (2001) Analysis of a 2,4,6-trichlorophenol-dehalogenating enrichment culture and isolation of the dehalogenating member Desulfitobacterium frappieri strain TCP-A. Arch Microbiol 175:133–142 25. Browne DT, Kenyon GL, Hegeman GD (1970) Incorporation of monofluorotryptophans into protein during the growth of Escherichia coli. Biochem Biophys Res Commun 39:13–19 26. Burhenne J, Ludwig M, Spiteller M (1997) Photolytic degradation of fluoroquinolone carboxylic acids in aqueous solution. Primary photoproducts and half-lives. Environ Sci Pollut Res 4:10–15 27. Burhenne J, Ludwig M, Spiteller M (1997) Photolytic degradation of fluoroquinolone carboxylic acids in aqueous solution. Isolation and structural elucidation of polar photometabolites. Environ Sci Pollut Res 4:61–67 28. Cain RB, Tranter EK, Darrah JA (1968) The utilization of some halogenated aromatic acids by Nocardia. Oxidation and metabolism. Biochem J 106:211–227 29. Carey JH, Fox ME (1981) Photodegradation of the lampricide 3-trifluoromethyl-4-nitrophenol (TFM). 1. Pathway of the direct photolysis in solution. J Great Lakes Res 7:234–241 30. Cass AEG, Ribbons DW, Rossiter JT, Williams SR (1987) Biotransformation of aromatic compounds Monitoring fluorinated analogues by NMR. FEBS Lett 220:353–357 31. Castro CE (1998) Environmental dehalogenation: chemistry and mechanism. Rev Environ Contam Toxicol 155:1–67 32. astro CE, Wade RS, Belser NO (1985) Biodehalogenation reactions of cytochrome P-450 with polyhalomethanes. Biochemistry 24:204–210 33. Cerniglia CE, Miller DW,Yang SK, Freeman JP (1984) Effects of a fluoro substituent on the fungal metabolism of 1-fluoronaphthalene. Appl Environ Microbiol 48:294–300

Degradation and Transformation of Organic Fluorine Compounds

191

34. Chang MB, Yu SJ (2001) An atmospheric-pressure plasjma process for C2F6 removal. Environ Sci Technol 35:1587–1592 35. Chang W, Criddle CS (1995) Biotransformation of HCFC-22, HCFC-142b, HCFC-123, HFC-134a by methantrophic mixed culture MM1. Biodegradation 6:1–9 36. Chee-Sanford JC, Frost JW, Fries MR, Zhou Z, Tiedje JM (1996) Evidence for acetyl coenzyme A and cinnamoyl coenzyme A in the anaerobic toluene mineralization pathway in Azoarcus tolulyticus Tol-4. Appl Environ Microbiol 62:964–97360 37. Chen L, MacMillan AW, Verdine GL (1993) Mutational separation of DNA binding from catalysis in a DNA cytosine methyltransferase. J Am Chem Soc 115:5318–5319 38. Christiansen N, Ahring BK, Wohlfarth G, Diekert G (1998) Purification and characterization of the 3-chloro-4-hydroxyphenylacetate reductive dehalogenase of Desulfitobacterium hafniense. FEBS Lett 436:159–162 39. Cole JR, Fathepure BZ, Tiedje JM (1995) Tetrachloroethene and 3-chlorobenzoate dechlorination avtivities are co-induced in Desulfomonile tiedjei DCB-1. Biodegradation 6:167–172 40. Cooper BH, Land APM (1979) Assimilation of protocatechuate acid and p-hydroxybenzoic acid as an aid to laboratory identification of Candida parapsilosis and other medically important yeasts. J Clin Microbiol 10:343–345 41. Cornellison G, van Noort PCM, Nachtegaal G, Kentgens APM (2000) A solid-state fluorineNMR study on hexafluorobenzene sorbed by sediments, polymers, and active carbon. Environ Sci Technol 34:645–649 42. Coulter C, Hamilton JTG, McRoberts WC, Kulakov L, Larkin MJ, Harper DB (1999) Halomethane: bisulfide/halide ion methyltransferase, an unusual corrinoid enzyme of environmental significance isolated from an aerobic methylotroph using chloromethane as the sole carbon source. Appl Environ Microbiol 65:4301–4312 43. Daly JW, Jerina DM,Witkop B (1972) Arene oxides and the NIH shift: the metabolism, toxicity and carcinogenicity of aromatic compounds. Experientia 28:1129–1149 44. Daubaras DL, Saido K, Chakrabarty AM (1996) Purifiation of hydroxyquinol 1,2-dioxygenase and maleylacetate reductase: the lower pathway of 2,4,5-trichlorophenoxyacetic acid metabolism by Burkholderia cepacia AC1100. Appl Environ Microbiol 62: 4276–4279 45. DeFlaun ME, Ensley BD, Steffan RJ (1992) Biological oxidation of hydrochlorofluorocarbons (HCFCs) by a methanotrophic bacterium. Bio/Technology 10 :1576 –1578 46. DeFrank JJ, Ribbons DW (1976) The p-cymene pathway in Pseudomonas putida PL: isolation of a dihydrodiol accumulated by a mutant. Biochem Biophys Res Commun 70:1129–1136 47. Dennie D, Gladu DDI, Lépine F,Villemur R, Bisaillon R, Beaudet R (1998) Spectrum of the reductive dehalogenation activity of Desulfitobacterium frappieri PCP-1.Appl Environ Microbiol 64:4603–4606 48. Detmer K, Massey V (1985) Effect of substrate and pH on the oxidative half-reaction of phenol hydroxylase. J Biol Chem 260:5998–6005 49. DeWeerd KA, Suflita JM (1990) Anaerobic aryl reductive dehalogenation of halobenzoates by cell extracts of “Desulfomonile tiedjei”. Appl Environ Microbiol 56:2999 – 3005 50. DeWeerd KA, Mandelco L, Tanner RS, Woese CR, Suflita JM (1990) Desulfomonile tiedjei gen nov and sp. nov., a novel anaerobic, dehalogenating, sulfate-reducing bacterium.Arch Microbiol 154 : 23 – 30 51. Dixon AM, Mai MA, Larvine CY (1999) NMR investigation of the interactions between 4′fluoro-1′-acetonaphthone and the Suwanee River fulvic acid. Environ Sci Technol 33:958–964 52. Eble KS, Dawson JH (1984) Novel reactivity of cytochrome P-450-CAM Methyl hydroxylation of 5,5-difluorocamphor. J Biol Chem 259:14,389–14,393 53. Elliott AJ (1994) Chlorofluorocarbons. In: Banks RE, Smart BE, Tatlow JC (eds) Organofluorine chemistry principles and commercial applications. Plenum Press, New York, pp 145–157

192

A.H. Neilson · A.-S. Allard

54. Ellis DA, Mabury SA (2000) The aqueous photolysis of TFM and related trifluoromethylphenols.An alternate source of trifluoroacetic acid in the environment. Environ Sci Technol 34:632–637 55. Engesser KH, Schmidt E, Knackmuss H-J (1980) Adaptation of Alcaligenes eutrophus B9 and Pseudomonas sp B13 to 2-fluorobenzoate as growth substrate. Appl Environ Microbiol 39:68–73 56. Engesser KH, Auling G, Busse J, Knackmuss H-J (1990) 3-Fluorobenzoate enriched bacterial strain FLB 300 degrades benzoate and all three isomeric monofluorobenzoates.Arch Microbiol 153:193–199 57. Engesser KH, Rubio MA, Ribbons DW (1988) Bacterial metabolism of side chain fluorinated aromatics: cometabolism of 4-trifluoromethyl (TFM)-benzoate by 4-isopropylbenzoate grown Pseudomonas putida JT strains. Arch Microbiol 149:198–206 58. Engesser KH, Rubio MA, Knackmuss H-J (1990) Bacterial metabolism of side-chain-fluorinated aromatics: unproductive meta cleavage of 3-trifluoromethylcatechol. Appl Microbiol Biotechnol 32:600–608 59. Engesser KH, Cain RB, Knackmuss HJ (1988b) Bacterial metabolism of side chain fluorinated aromatics: cometabolism of 3-trifluoromethyl (TFM)-benzoate by Pseudomonas putida (arvilla) mt-2 and Rhodococcus rubropertinctus N657. Arch Microbiol 149:188–197 60. Eppink MHM, Boeren SA,Vervoort J, van Berkel WJH (1997) Purification and properties of 4-hydroxybenzoate 1-hydroxylase (decarboxylating), a novel flavin adenine dinucleotide-dependent monooxygenase from Candida parapsilosis CBS604. J Bacteriol 179:6680–6687 61. Eschrich K, van der Bolt FJY, de Kok A, van Berkel WJH (1993) Role of Tyr201 and Tyr385 in substrate activation by p-hydroxybenzoate hydroxylase from Pseudomonas fluorescens. Eur J Biochem 216:137–146 62. Fabian P, Singh ON (eds) (1999) Reactive halogen compounds in the atmosphere. Springer, Berlin Heidelberg New York 63. Fetzner S, Lingens F (1994) Bacterial dehalogenases: biochemistry, genetics and biotechnical applications. Microbiol Revs 58:641–685 64. Fetzner S, Müller R, Lingens F (1992) Purification and some properties of 2-halobenzoate 1,2-dioxygenase, a two-component enzyme system from Pseudomonas cepacia 2CBS. J Bacteriol 174:279–290 65. Finkelstein ZI, Baskunov BP, Boersma MG, Vervoort J, Golovlev EL, van Berkel WJH, Gololeva LAA, Rietjens IMCM (2000) Identification of fluoropyrogallols as new intermedates in biotransformation of monofluorophenols in Rhodococcus opacus 1 cp. Appl Environ Microbiol 66:2148–2153 66. Fox BG, Borneman JG, Wackett LP, Lipscomb JD (1990) Halolkene oxidation by the soluble methane monooxygenase from Methylosinus trichosporium OB3b: mechanistic and environmental implications. Biochemistry 29:6419–6427 67. Fukumori F, Saint CP (1997) Nucleotide sequences and regulational analysis of genes involved in conversion of aniline to catechol in Pseudomonas putida UCC22 (pTDN1). J Bacteriol 179:399–408 68. Ganzer CJ, Wackett LP (1991)Reductive dechlorination catalyzed by bacterial transitionmetal coenzymes. Environ Sci Technol 25:715–721 69. Genther BRS, Townsend GT, Chapman PJ (1990) Effect of fluorinated analogues of phenol and hydroxybenzoates on the anaerobic transformation of phenol to benzoate. Biodegradation 1:65–74 70. Gerritse J, Renard V, Gomes TMP, Lawson PA, Collins MD, Gottachal JC (1996) Desulfitobacterium sp. strain PCE1, an anaerobic bacterium that can grow by reductive dechlorination of tetrachloroethene or ortho-chlorinated phenols. Arch Microbiol 165:132–140 71. Gibson DT, Koch JR, Schuld CL, Kallio RE (1968) Oxidative degradation of aromatic hydrocarbons by microorganisms II. Metabolism of halogenated aromatic hydrocarbons. Biochemistry 7:3795–3802

Degradation and Transformation of Organic Fluorine Compounds

193

72. Gibson DT, Resnick SM, Lee K, Brand JM, Torok DS, Wackett LP, Schocken MJ, Haigler BE (1995) Desaturation, dioxygenation, and monooxygenation reactions catalyzed by naphthalene dioxygenase from Pseudomonas sp strain 9816–4. J Bacteriol 177:2615–2621 73. Giesy JP, Kannan K (2001) Global distribution of perfluorooctane sulfonate in wildlife. Environ Sci Technol 35:1339–1342 74. Glod G,Angst W, Holliger C, Schwartzenbach RP (1997) Corrinoid-mediated reduction of tetrachloroethene, trichloroethene,and trichlorofluoroethene in homogeneous solution: reaction kinetics and reaction mechanisms. Environ Sci Technol 31:253–260 75. Goldman P (1965) The enzymatic cleavage of the carbon-fluorine bond in fluoroacetate. J Biol Chem 240:3434–3438 76. Goldman P (1966) Carbon-fluorine bond cleavage II. Studies on the mechanism of the defluorination of fluoroacetate. J Biol Chem 241:5557–5559 77. Goldman P, Milne GWA, Pignataro MT (1967) Fluorine containing metabolites formed from 2-fluorobenzoic acid by Pseudomonas species. Arch Biochem Biophys 118 : 178–184 78. Goodwin KD, Schaefer JK, Oremland RS (1998) Bacterial oxidation of dibromomethane and methyl bromide in natural waters and enrichment cultures. Appl Environ Microbiol 64:4629–4636 79. Gottlieb AA, Fujita Y, Udenfriend S,Witkop B (1965) Incorporation of cis- and trans-4-fluoro-L-proline into proteins and hydroxylation of the trans isomer during collagen biosynthesis. Biochemistry 4:2507–2513 80. Green NA, Meharg AA, Till C, Troke J, Nicholson JK (1999) Degradation of 4-fluorobiphenyl by mycorrizal fungi as determined by 19F nuclear magnetic resonance spectroscopy and 14C radiolabelling analysis. Appl Environ Microbiol 65:4021–4027 81. Hamilton JTG, Murphy CD, Amin MR, O’Hagan D, Harper DB (1998) Exploring the biosynthetic origin of fluoroacetate and 4-fluorothreonine in Streptomyces cattleya. J Chem Soc Perkin Trans 1:759–767 82. Hansen KJ, Clemen LA, Ellefson ME, Johnson HO (2001) Compound-specific quantitative characterization of organic fluorochemicals in biological matrices. Environ Sci Technol 35:766–770 83. Harayama S, Kok M, Neidle EL (1992) Functional and evolutionary relationships among diverse oxygenases. Annu Rev Microbiol 46:565–601 84. Harms G, Zengle K, Rabus R, Aeckersberg F, Minz D, Rosselló-Mora R, Widdel F (1999) Anaerobic oxidation of o-xylene, m-xylene, and homologous alkylbenzenes by new types of sulfate-reducing bacteria. Appl Environ Microbiol 65:999–1004 85. Harnisch J (1999) Reactive fluorine compounds. In: Fabian P, Singh ON (eds) Reactive halogen compounds in the atmosphere. Springer, Berlin Heidelberg New York, pp 81–111 86. Harper DB, Blakley ER (1971) The metabolism of p-fluorophenylacetic acid by a Pseudomonas sp I. Isolation and identification of intermediates in degradation. Can J Microbiol 17:635–644 87. Harper DB, Blakley ER (1971) The metabolism of p-fluorophenylacetic acid by a Pseudomonas sp II. The degradative pathway. Can J Microbiol 17:645–650 88. Harper DB, Blakley ER (1971) The metabolism of p-fluorobenzoic acid by a Pseudomonas sp. Can J Microbiol 17:1015–1023 89. Hartz CL, Bevan JW, Jackson MC, Wofford BA (1998) Inovative surface wave plasma reactor technique for PFC abatement. Environ Sci Technol 32:682–687 90. Harwood CS, Burchardt G, Herrmann H, Fuchs G (1999) Anaerobic metabolism of aromatic compounds via the benzoyl-CoA pathway. FEMS Microbiol Revs 22:439–458 91. Helvenstron MC, Castro CE (1992) Nickel(I) octaethylisobacteriochlorin anion. An exceptional nucleophile. Reduction and coupling of alkyl halides by anionic and radical processes. A model for factor F-430. J Am Chem Soc 114:8490–8496 92. Hewitt LM, Carey JH, Munkittrick KR, Parrott JL, Solomon KR, Servos MR (1998) Identification of chloro-nitro-trifluoromethyl-substituted dibenzo-p-dioxins in lampricide formulations of 3-trifluoromethyl-4-nitrophenol: assessment to induce mixed function oxidase activity. Anal Chim Acta 17:941–950

194

A.H. Neilson · A.-S. Allard

93. Hickey WJ, Focht DD (1990) Degradation of mono-, di-, and trihalogenated benzoic acids by Pseudomonas aeruginosa JB2. Appl Environ Microbiol 56:3842–3850 94. Hill RL, Teipel JW (1971) Fumarase and crotonase. Enzymes 3rd edn 5:539–571 95. Holliger C, Schraa G, Stuperich E, Stams AJM, Zehnder AJB (1992) Evidence for the involvement of corrinoids and factor F430 in the reductive dechlorination of 1,2dichloroethane by Methanosarcina barkeri. J Bacteriol 174:4427–4434 96. Holliger C, Kengen SWM, Schraa G, Stams AJM, Zehnder AJB (1992) Methyl-coenzyme M reductase of Methanobacterium thermoautrotrophicum delta H catalyzes the reductive dechlorination of 1,2-dichloroethane to ethylene and chloroethane. J Bacteriol 174:4435–4443 97. Holliger C, Wohlfarth G, Diekert G (1999) Reductive dechlorination in the energy metabolism of anaerobic bacteria. FEMS Microbiol Revs 22:383–398 98. Hughes DE (1965) The metabolism of halogen-substituted benzoic acids by Pseudomonas fluorescens. Biochem J 96:181–188 99. Hughes GMK, Saunders BC (1954) Studies in peroxidase action. Part IX. Reactions involving the rupture of the C–F, C–Br and C–I links in aromatic amines. J Chem Soc 4630–4634 100. Hur H-G, Sadowsky M, Wackett LP (1994) Metabolism of chlorofluorocarbons and polybrominated compounds by Pseudomonas putida G786pHG-2 via an engineered metabolic pathway. Appl Environ Microbiol 60:4148–4154 101. Husain M, Entsch B, Ballou DP, Massey V, Chapman PJ (1980) Fluoride elimination from substrates in hydroxylation reactions catalyzed by p-hydroxybenzoate hydroxylase. J Biol Chem 255:4189–4197 102. Hyman MR, Page CL, Arp DJ (1994) Oxidation of methyl fluoride and dimethyl ether by ammonia monooxygenase in Nitrosomonas europaea. Appl Environ Microbiol 60:3033–3035 103. Häggblom MM (1990) Mechanisms of bacterial degradation and transformation of chlorinated monoaromatic compounds. J Basic Microbiol 30:115–141 104. Jacobs KL, Grogan DW (1997) Rates of spontaneous mutation in an archaeon from geothermal environments. J Bacteriol 179:3298–3303 105. Jaffe R, Hites RA (1985) Identification of new, fluorinated biphenyls in the Niagara River Lake Ontario area. Environ Sci Technol 19:736–740 106. Johnston JB, Winicov H, Cunningham VL (1987) Microbial catabolites as chemical synthesis intermediates: a cost-benefit analysis. Biotechnol Prog 3:127–130 107. Kallio A, McCann PP (1981) Difluoromethylornithine irreversibly inactivates ornithine decarboxylase of Pseudomonas aeruginosa, but does not inhibit the enzymes of Escherichia coli. Biochem J 200:69–75 108. Kallio A, McCann PP, Bey P (1982) DL-–Monofluoromethylputrescine is a potent irreversible inhibitor of Escherichi coli ornithine decarboxylase. Biochem J 204:771–775 109. Kannan K, Koistinen J, Beckman K, Evans T, Gorzelany JF, Hansen KL, Jones PD, Helle E, Nyman M, Geisy JP (2001) Accumulation of perfluorooctane sulfonate in marine mammals. Environ Sci Technol 35:1593–1598 110. Kaschabek SR, Reineke W (1995) Maleylacetate reductase of Pseudomonas sp. strain B13: specficity of substrate conversion and halide elimination. J Bacteriol 177:320–325 111. Kaschabek SR, Kasberg T, Müller B, Mars AE, Janssen DB, Reineke W (1998) Degradation of chloroaromatics: purification and characterization of a novel type of chlorocatechol 2,3-dioxygenase from Pseudomonas putida GJ31. J Bacteriol 180:296–302 112. Kaufman S (1961) The enzymatic conversion of 4-fluorophenylalanine to tyrosine. Biochim Biophys Acta 51:619–621 113. Kawasaki H, Tone N, Tonomura K (1981) Plasmid-determined dehalogenation of haloacetates in Moraxella species. Agric Biol Chem 45:29–34 114. Kersten PJ, Chapman PJ, Dagley S (1985) Enzymatic release of halogens or methanol from some substituted protocatechuic acids. J Bacteriol 162:693–697 115. Key BD, Howell RD, Criddle CS (1997) Fluorinated organics in the biosphere. Environ Sci Technol 31:2445–2454

Degradation and Transformation of Organic Fluorine Compounds

195

116. Key BD, Howell RD, Criddle CS (1998) Defluorination of organofluorine sulfur compounds by Pseudomonas sp strain D2. Environ Sci Technol 32:2283–2287 117. Kim BR, Suidan MT, Wallington TJ, Du X (2000) Biodegradability of trifluoroacetic acid. Env Eng Sci 17:337–342 118. Kim DH, Lees WJ, Haley TM,Walsh CT (1995) Kinetic characterization of the inactivation of UDP-GlcNAc enolpyruvate transferase by (Z)-3-fluorophosphoenolpyruvate: evidence for two oxocarbenium ion intemediates in enolpyruvyl transfer catalysis. J Am Chem Soc 117:1494–1502 119. Kim DH, Lees WJ, Haley TM, Walsh CT (1995) Stereochemical analysis of the tetrahedral adduct formed at the active site of UDP-GlcNAc enolpyruvate transferase from the pseudosubstrates (E) and (Z)-3-fluorophosphoenolpyruvate, in D2O. J Am Chem Soc 117:6380–6381 120. Kitazume T, Ishikawa N (1983) Asymmetrical reduction of perfluoroalkylated ketones, ketoesters and vinyl compounds with baker’s yeast. Chem Lett 237–238 121. Kollonitsch J, Patchett AA, Marburg S, Maycock AL, Perkins LM, Doldouras GA, Duggan DE, Aster SD (1978) Selective inhibitors of biosynthesis of aminergic neurotransmitters. Nature (London) 274:906–908 122. Konstantinou IK, Sakellarides TM, Sakkas VA, Albanis TA (2001) Photocatalytic degradation of selected s-triazine herbicides and organophosphorus insecticides over aqueous TiO2 suspensions. Environ Sci Technol 35:398–405 123. Krone UE, Thauer RK (1992) Dehalogenation of trichlorofluoromethane (CFC-11) by Methanosarcina barkeri. FEMS Microbiol Lett 90:201–204 124. Krone UE, Laufer K, Thauer RH, Hogenkamp HPC (1989) Coenzyme F430 as a possible catalyst for the reductive dehalogenation of chlorinated C-1 hydrocarbons in methanogenic bacteria. Biochemistry 28:10,061–10,065 125. Krone UE, Thauer RH, Hogenkamp HPC (1989) Reductive dehalogenation of chlorinated C-1 hydrocarbons mediated by corrinoids. Biochemistry 28:4908–4914 126. Lesage S, Brown S, Hosler KH (1992) Degradation of chlorofluorocarbon-113 under anaerobic conditions. Chemosphere 24 :1225–1243 127. Leuthner B, Heider J (2000) Anaerobic toluene catabolism of Thauera aromatica: the bbs operon codes for enzynes of the b oxidation of the intermediate benzylsuccinate. J Bacteriol 182:272–277 128. Leuthner B, Leutwein C, Schulz H, Hörth P, Haehnel W, Schiltz E, Schägger H, Heider J (1998) Biochemical and genetic characterisation of benzylsuccinate synthase from Thauera aromatica: a new glycyl-radical catalysing the first step in anaerobic toluene degradation. Mol Microbiol 28:615–628 129. Levitt MS, Newton RF, Roberts SM, Willetts AJ (1990) Preparation of optically active 6′fluorocarbocyclic nucleosides utilising an enentiospecific enzyme-catalysed Baeyer-Villiger type oxidation. J Chem Soc Chem Commun 619 – 620 130. Li S, Wackett LP (1993) Reductive dehalogenation by cytochrome P450CAM : substrate binding and catalysis. Biochemistry 32:9355–9361 131. Londry KL, Fedorak PM (1993) Fluorophenols and 3-fluorobenzoate in phenol-degrading methanogenic cultures. Arch Microbiol 160:137–143 132. Londry KL, Fedorak PM (1993) Use of fluorinated compounds to detect aromatic metabolites from m-cresol in a methanogenic consortium: evidence for a demethylation reaction. Appl Environ Microbiol 59 :2229–2238 133. Louie TM, Mohn WW (1999) Evidence for a chemiosmotic model of dehalorespiration in Desulfomonile tiedjei DCB-1. J Bacteriol 181:40–46 134. Lovley DR, Lonergan DJ (1990) Anaerobic oxidation of toluene, phenol, and p-cresol by the dissimilatory iron-reducing organism, GS-15.Appl Environ Microbiol 56:1858–1864 135. Lynch RM, Woodley JM, Lilly MD (1997) Process design for the oxidation of fluorobenzene to fluorocatechol by Pseudomonas putida. J Biotechnol 58:167–175 136. Magnuson JK, Romine MF, Burris DR, Kingsley MT (2000) Trichloroethene reductive dehalogenase from Dehalococcoides ethenogenes: sequence of tceA and substrate characterization. Appl Environ Microbiol 66:5141–5147

196

A.H. Neilson · A.-S. Allard

137. Mao SS, Holler TP, Yu GX, Bollinger JM, Booker S, Johnston MI, Stubbe J (1992) A model for the role of multiple cysteine residues involved in ribonucleotide reduction: amazing and still confusing. Biochemistry 31:9733–9743 138. Marcotte PA, Robinson CH (1982) Inhibition and inactivation of estrogen synthase (aromatase) by fluorinated substrate analogues. Biochemistry 21:2773–2778 139. Marks TS, Smith ARW, Quirk AV (1984) Degradation of 4-chlorobenzoic acid by Arthrobacter sp. Appl Environ Microbiol 48:1020–1025 140. Marletta MA, Chenng K-F, Walsh C (1982) Stereochemical studies on the hydration of monofluorofumarate and 2,3-difluorofumarate by fumarase. Biochemistry 21:2637–2644 141. Mars AE, Kingma J, Kaschabek SR, Reinke W, Janssen DB (1999) Conversion of 3-chlorocatechol by various catechol 2,3-dioxygenases and sequence analysis of the chlorocatechol dioxygense region of Pseudomonas putida GJ31. J Bacteriol 181:1309–1318 142. Mars AE, Kasberg T, Kaschabek SR, van Agteren MH, Janssen DB, Reineke W (1997) Microbial degradation of chloroaromatics: use of the meta-cleavage pathway for mineralization of chlorobenzene. J Bacteriol 179:4530–4537 143. Martin JW, Franklin J, Hanson ML, Solomon KR, Mabury SA, Ellis DA, Scott BF, Muir DCG (2000) Detection of chlorodifluoroacetic acid in precipitation: a possible product of fluorocarbon degradation. Environ Sci Technol 34:274–281 144. Martin RE, Baker PB, Ribbons DW (1987) Biotransformations of fluoroaromatic compounds: accumulation of hydroxylated products from 3-fluorophthalic acid using mutant strains of Pseudomonas testosteroni. Biocatalysis 1:37–46 145. Mason JR, Cammack R (1992) The electron-transport proteins of hydroxylation bacterial dioxygenases. Annu Rev Microbiol 46:277–305 146. McCarthy JR, Sunkara PS, Matthews DP, Bitonti AJ, Jarvi ET, Sabol JS, Resvick RJ, Huber EW, van der Donk WA,Yu G, Stubbe J (1996) Design of a fluoro-olefin nucleoside as a bioprecursor of a mechanism-based inhibitor of ribonucleotide reductase. In: Ojima I, McCarthy JR, Welch JT (eds) Biomedical frontiers of fluorine chemistry. American Chemical Society Symposium Series 639, American Chemical Society, Washington DC, pp 246–264 147. Mead RJ, Segal W (1972) Fluoracetic acid biosynthesis: a proposed mechanism.Aust J Biol Sci 25:327–333 148. Meckenstock RU (1999) Fermentative toluene degradation in anaerobic defined syntrophic cocultures. FEMS Microbiol Lett 177:67–73 149. Meyer JJM, Grobbelaar N, Steyn PL (1990) Fluoroacetate-metabolizing pseudomonad isolated from Dichapetalum cymosum. Appl Environ Microbiol 56:2152–2155 150. Midgley PM, McCulloch A (1999) Properties and applications of industrial halocarbons. In: Fabian P, Singh ON (eds) Reactive halogen compounds in the atmosphere. Springer, Berlin Heidelberg New York 151. Miller LG, Sasson C, Oremland RS (1998) Difluoromethane, a new and improved inhibitor of methanotrophy. Appl Environ Microbiol 64:4357–4362 152. Miller LG, Coutlakis MD, Oremland RS, Ward BB (1993) Selective inhibition of ammonium oxidation and nitrification-linked N2O formation by methyl fluoride and dimethyl ether. Appl Environ Microbiol 59:2457–2464 153. Milne GWA, Goldman P, Holzman JL (1968) The metabolism of 2-fluorobenzoic acid. J Biol Chem 243:5374–5376 154. Mohn WW, Kennedy KJ (1992) Reductive dehalogenation of chlorophenols by Desulfomonile tiedjei DCB-1. Appl Environ Microbiol 58:1367–1370 155. Moody CA, Field JA (2000) Perfluorinated surfactants and environmental implications of their use in fire-fighting foams. Environ Sci Technol 34:3864–3870 156. Moody CA, Kwan WC, Martin JW, Muir DCG, Mabury SA (2001) Determination of perfluorinated surfactants in surface water samples by two independent analytical techniques: liquid chromatography/tandem mass spectrometry and 19F NMR. Anal Chem 73:2200–2206 157. Mägli A, Messmer M, Leisinger T (1998) Metabolism of dichloromethane by the strict anaerobe Dehalobacterium formicoaceticum. Appl Environ Microbiol 64:646–650

Degradation and Transformation of Organic Fluorine Compounds

197

158. Nakatsu CH,Wyndham RC (1993) Cloning and expression of the transposable chlorobenzoate-3,4-dioxygenase genes of Alcaligenes sp strain BR60. Appl Environ Microbiol 59:3625–3633 159. Neidle EL, Hartnett C, Ornmston LN, Bairoch A, Rekin M, Harayama S (1991) Nucleotide sequences of the Acinetobacter calcoaceticus benABC genes for benzoate 1,2-dioxygenase reveal evolutionary relationship among multicomponent oxygenases. J Bacteriol 173:5385–5395 160. Neilson AH, Lindgren C, Hynning P-Å, Remberger M (1988) Methylation of halogenated phenols and thiophenols by cell extracts of gram-positive and gram-negative bacteria. Appl Environ Microbiol 54:524–530 161. Neilson AH (1990) The biodegradation of halogenated organic compounds. J Appl Microbiol 69:445–470 162. Neilson AH (2000) Organic chemicals: an environmental perspective. CRC Press, Boca Raton, Florida 163. Neilson AH, Allard A-S (1998) Microbial metabolism of PAHs and heteroarenes. In: Neilson AH (ed) PAHs and related compounds. In: Hutzinger O (ed) Handbook of environmental chemistry, vol 3, part J. Springer, Berlin Heidelberg New York, pp 1–80 164. Neujahr HY (1991) Phenol hydroxylase. In: Muller F (ed) Chemistry and biochemistry of flavoenzymes, vol II. CRC Press, Boca Raton, Florida, pp 65–85 165. Neujahr HY, Kjellén KG (1978) Phenol hydroxylase from yeast. J Biol Chem 253:8835–8841 166. Ni S, Fredrickson JK, Xun L (1995) Purification and characterization of a novel 3chlorobenzoate-reductive dehalogenase from the cytoplasmic membrane of Desulfomonile tiedjei DCB-1. J Bacteriol 177:5135–5139 167. Nohara K, Toma M, Kutsuna S, Takeuchi K, Ibusuki T (2001) Cl atom-initiated oxidation of three homologous methylperfluoroalkyl ethers. Environ Sci Technol 35: 114–120 168. Nordlund P, Sjöberg B-M, Eklund H (1990) Three-dimensional structure of the free radical protein of ribonucleotide reductase. Nature (London) 345:593–598 169. Nowara A, Burhenne J, Spiteller M (1997) Binding of fluroquinolone carboxylic acid derivatives to clay minerals. J Agric Food Chem 45:1459–1463 170. Ojima I, McCarthy JR, Welch JT (eds) (1996) Biomedical frontiers of fluorine chemistry. American Chemical Society Symposium Series 639. American Chemical Society, Washington DC 171. Oldenhuis R, Vink RLJM, Janssen DB, Witholt B (1989) Degradation of chlorinated aliphatic hydrocarbons by Methylosinus trichosporium OB3b expressing soluble methane monooxygenase. Appl Environ Microbiol 55:2817–2826 172. Oltmanns RH, Müller R, Otto MK, Lingens F (1989) Evidence for a new pathway in the bacterial degradation of 4-fluorobenzoate. Appl Environ Microbiol 55 : 2499–2504 173. Oremland RS, Lonrergan DJ, Culbertson CW, Lovley DR (1996) Microbial degradation of hydrochlorofluorocarbons (CHCl2F and CHCl2CF3) in soils and sediments.Appl Environ Microbiol 62:1818–1821 174. Palmer JL, Abeles RA (1979) The mechanism of action of S-adenosylhomocysteinase. J Biol Chem 254:1217–1226 175. Parisot D, Malet-Martino MC, Martino R, Crasnier P (1991) 19F nuclear magnetic resonance analysis of 5-fluorouracil metabolism in four differently pigmented strains of Nectria haematococca. Appl Environ Microbiol 57:3605–3612 176. Parshikov IA, Freeman JP, Lay JO, Beger RD, Williams AJ, Sutherland JB (2000) Microbiological transformation of enrofloxacin by the fungus Mucor ramannianus. Appl Environ Microbiol 66:2664–2667 177. Peelen S, Rietjens IMC, van Berkel WJH, van Workum WAT, Vervoort J (1993) 19F-NMR study on the pH-dependent regioselectivity and rate of the ortho-hydroxylation of 3fluorophenol by phenol hydroxylase from Trichosporon cutaneum. Eur J Biochem 218:345–353

198

A.H. Neilson · A.-S. Allard

178. Peelen S, Rietjens IMC, Boersma MG,Vervoort J (1995) Conversion of phenol derivatives to hydroxylated products by phenol hydroxylase from Trichosporon cutaneum. Eur J Biochem 227 : 284 – 291 179. Poulter CD, Rilling HC (1978) The prenyl transfer reaction. Enzymatic and mechanistic studies of the 1′-4 coupling reaction in the terpene biosynthetic reaction. Acc Chem Res 11:307–313 180. Poulter CD (1996) Mechanistic studies of the prenyl transfer reaction with fluorinated substrate analogues. In: Ojima I, McCarthy JR,Welch JT (eds) Biomedical frontiers of fluorine chemistry.American Chemical Society Symposium Series 639.American Chemical Society, Washington DC, pp 158–168 181. Rabus R, Nordhaus R, Ludwig W, Widdel F (1993) Complete oxidation of toluene under strictly anoxic conditions by a new sulfate-reducing bacterium. Appl Environ Microbiol 59:1444–1451 182. Rabus R, Heider J (1998) Initial reactions of anaerobic metabolism of alkylbenzenes in denitrifying and sulfate-reducing bacteria. Arch Microbiol 170:377–384 183. Rao VNM (1994) Alternatives to chlorofluorocarbons (CFCs). In: Banks RE, Smart BE, Tatlow JC (eds)Organofluorine chemistry principles and commercial applications. Plenum Press, New York, pp 159–175 184. Reid KA, Hamilton JTG, Bowden RD, O’Hagan D, Dasaradhi L, Amin MR, Harper DB (1995) Biosynthesis of fluorinated secondary metabolites by Streptomyces cattleya. Microbiology (UK) 141:1385–1393 185. Reineke W, Otting W, Knackmuss H-J (1978) cis-Dihydrodiols microbially produced from halo- and methylbenzoic acids. Tetrahedron 34:1797–1814 186. Reineke W, Knackmuss H-J (1978) Chemical structure and biodegradability of halogenated aromatic compounds. Substituent effects on 1,2-dioxygenation of benzoic acid. Biochim Biophys Acta 542:412–423 187. Reiner AM, Hegeman GD (1971) Metabolim of benzoic acid by bacteria.Accumulation of (–)-3,5-cyclohexadien-1,2-diol,1-carboxylic acid by a mutant strain of Alcaligenes eutrophus. Biochemistry 10:2530–2536 188. Reinscheid UM, Zuilhog H, Müller R, Vervoort J (1998) Biological, thermal and photochemical transformation of 2-trifluoromethylphenol. Biodegradation 9 : 487–499 189. Renganathan V (1989) Possible involvement of toluene-2,3-dioxygenase in defluorination of 3-fluoro-substituted benzenes by toluene-degrading Pseudomonas sp strain T-12.Appl Environ Microbiol 55:330–334 190. Renson J, Bourdon V (1968) Hydroxylation et défluorination des dérivés fluores de l ániline. Arch Int Pharmacodyn 171:240–242 191. Ribbons DW, Keyser P, Kunz DA, Taylor BF, Eaton RW, Anderson BN (1984) Microbial degradation of phthalates. In: Gibson DT (ed) Microbial degradation of organic compounds. Marcel Dekker, New York, pp 371–395 192. Ribbons DW, Taylor SJC, Evans CT, Thomas SD, Rossiter JT, Widdowson DA, Williams DJ (1990) Biodegradations yield novel intermediates for chemical synthesis. In: Kamely D, Chakrabarty A, Omenn GS (eds) Biotechnology and biodegradation.Advances in Applied Biotechnolgy, vol 4. Gulf Publishing, Houston, Texas, pp 213–244 193. Richey DG, Driscoll CT, Likens GE (1997) Soil retention of trifluoroacetate. Environ Sci Technol 31:1723–1727 194. Rietjens IMCM, Tyrakowska B, Veeger C, Vervoort J (1990) Reaction pathways for biodehalogenation of fluorinated anilines. Eur J Biochem 194 : 945 – 954 195. Robins MJ, Neschadimenko V, Ro B-O, Yuan C-S, Borchardt RT, Wnuk SF (1998) Nucleic acid related compounds.101. S-adenosyl-L-homocysteine hydrolase does not hydrate (5′fluoro)vinyl or (6′-halo)homovinyl analogues deririved from 3′-deoxyadenosine or 3′(chloro or fluoro)-3′-deoxyadenosine. J Org Chem 63 :1205 –1211 196. Romanov V, Hausinger RP (1994) Pseudomonas aeruginosa 142 uses a three-component ortho-halobenzoate 1,2-dioxygenase for metabolism of 2,4-dichloro- and 2-chlorobenzoate. J Bacteriol 176:3368–3374

Degradation and Transformation of Organic Fluorine Compounds

199

197. Romanov V, Hausinger RP (1996) NADPH-dependent reductive ortho dehalogenation of 2,4-dichorobenzoic acid in Corynebacterium sepedonicum KZ-4 and cornefrom bacterium strain NTB-1 via 2,4-dichlorobenzoyl coenyze A. J Bacteriol 178:2656–2661 198. Rossiter JT, Williams SR, Cass AEG, Ribbons DW (1987) Aromatic biotransformations 2: production of novel chiral fluorinated 3,4-cyclohexadiene-cis-1,2-diol-1-carboxylates. Tetrahedron Lett 28:5173–5174 199. Sander P, Wittich R-M, Fortnagel P, Wilkes H, Francke W (1991) Degradation of 1,2,4trichloro- and 1,2,4,5-tetrachlorobenzene by Pseudomonas strains. Appl Environ Microbiol 57:1430–1440 200. Santi DV, Sakai TT (1971) Thymidylate synthetase. Model studies of inhibition by 5-trifluoromethy-2′-deoxyuridylic acid. Biochemistry 10:3598–3607 201. Schaefer JK, Oremland RS (1999) Oxidation of methyl halides by the facultative methylotroph strain IMB-1. Appl Environ Microbiol 65:5035–5041 202. Schennen U, Braun K, Knackmuss H-J (1985) Anaerobic degradation of 2-fluorobenzoate by benzoate-degrading, denitrifying bacteria. J Bacteriol 161:321–325 203. Schlömann M, Schmidt E, Knackmuss H-J (1990) Different types of dienelactone hydrolase in 4-fluorobenzoate-utilizing bacteria. J Bacteriol 172:5112–5118 204. Schlömann M, Fischer P, Schmidt E, Knackmuss H-J (1990) Enzymatic formation, stability, and spontaneous reactions of 4-fluoromuconolactone, a metabolite of the bacterial degradation of 4-fluorobenzoate. J Bacteriol 172:5119–5129 205. Schmidt E, Remberg G, Knackmuss H-J (1980) Chemical structure and biodegradability of halogenated aromatic compounds. Halogenated muconic acids as intermediates. Biochem J 192 : 331 – 337 206. Schmidt S, Fortnagel P, Wittich R-M (1993) Biodegradation and transformation of 4,4′and 2,4-dihalodiphenyl ethers by Sphingomonas sp strain SS33. Appl Environ Microbiol 58:3931–3933 207. Schmidt S, Wittich R-M, Erdmann D, Wilkes H, Francke W, Fortnagel P (1992) Biodegradation of diphenyl ether and its monohalogenated derivatives by Sphingomonas sp strain SS3. Appl Environ Microbiol 58:2744–2750 208. Schnell S, Schink B (1991) Anaerobic aniline degradation via reductive deamination of 4aminobenzoyl-CoA in Desulfobacterium anilini. Arch Microbiol 155:183–190 209. Schreiber A, Hellwig M, Dorn E, Reineke W, Kmackmuss H-J (1980) Critical reactions in fluorobenzoic acid degradation by Pseudomonas sp B13. Appl Environ Microbiol 39:58–67 210. Schreuder HA, Prick PAJ, Wierenga RK, Vriend G, Wilson KS, Hol WGJ, Drenth J (1989) Crystal structure of the p-hydroxybenzoate hydroxylasesubstrate complex refined at 19 Å resolution. J Mol Biol 208:679–696 211. Schreuder HA, Hol WGJ, Drenth J (1990) Analysis of the active site of the flavoprotein phydroxybenzoate hydroxylase and some ideas with respect to its reaction mechanism. Biochemistry 29:3101–3108 212. Schreuder HA, Mattevi A, Obmolova G, Kalk KH, Hol WGJ, van der Bolt FJT, van Berkel WJH (1994) Crystal structure of wild-type p-hydroxybenzoate hydroxylase complexed with 4-aminobenzoate, 2,4-dihydroxybenzoate, and 2-hydroxy-4-aminobenzoate, and of the Tyr222Ala mutant complexed with 2-hydroxy-4-aminobenzoate. Evidence for a proton channel and a new binding mode of the flavin ring. Biochemistry 33 10,161–10,170 213. Schweizer D, Markus A, Seez M, Ruf HH, Lingens F (1987) Purification and some properties of component B of the 4-chlorophenylacetate 3,4-diooxygenase from Pseudomonas species strain CBs 3. J Biol Chem 262:9340–9346 214. Seeger M, Cámara B, Hofer B (2001) Dehalogenation, denitrification, dehydroxylation, and angular attack on substituted biphenyls and related compounds by a biphenyl dioxygenase. J Bacteriol 183:3548–3555 215. Selifonov SA, Gurst JE, Wackett LP (1995) Regioselective dioxygenation of ortho-trifluoromethylbenzoate by Pseudomonas aeruginosa 142: evidence for 1,2-dioxygenation as a mechanism in ortho-halobenzoate dehalogenation. Biochem Biophys Res Comm 213:759–767

200

A.H. Neilson · A.-S. Allard

216. Seyfried B, Glod G, Schocher R, Tschech A, Zeyer J (1994) Initial reactions in the anaerobic oxidation of toluene and m-xylene by denitrifying bacteria. Appl Environ Microbiol 60:4047–4052 217. Smith MR (1994) The physiology of aromatic hydrocarbon degrading bacteria. In: Ratledge C (ed) Biochemistry of microbial degradation. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp 347–378 218. Song B, Palleroni NJ, Häggblom MM (2000) Isolation and characterization of diverse halobenzoate-degrading denitrifying bacteria from soils and sediments.Appl Environ Microbiol 66:3446–3453 219. Song B, Palleroni NJ, Häggblom MM (2000) Description of strain 3CB-1, a genomovar of Thauera aromatica, capable of degrading 3-chlorobenzoate coupled to nitrate reduction. Int J Syst Evol Microbiol 50:551–558 220. Sonier DN, Duran NL, Smith GB (1994) Dechlorination of trichlorofluoromethane (CFC11) by sulfate-reducing bacteria from an aquifer contaminated with halogenated aliphatic compounds. Appl Environ Microbiol 60:4567–4572 221. Streger SH, Condee CW, Togna AP, Deflaun MF (1999) Degradation of halohydrocarbons and brominated compounds by methane- and propane-oxidizing bacteria. Environ Sci Technol 33:4477–4482 222. Stubbe J (1989) Ribonucleotide reductases. Adv Enzymol 63:349–419 223. Stubbe J, van der Donk WA (1998) Protein radicals in enzyme catalysis. Chem Revs 98:705–762 224. Sullivan JP, Dickinson D, Chase HA (1998) Methanotrophs, Methylosinus trichosporíum OB3b, sMMO, and their application to bioremediation. Crit Revs Microbiol 24:335–373 225. Taylor BF (1983) Aerobic and anaerobic catabolism of vanillic acid and some other methoxy-aromatic compounds by Pseudomonas sp strain PN-1. Appl Environ Microbiol 46:1286–1292 226. Taylor BF, Amador JA, Levinson HS (1993) Degradation of meta-trifluoromethylbenzoate by sequential microbial and photochemical reactions. FEMS Microbiol Lett 110:213–216 227. Taylor BF, Hearn WL, Pincus S (1979) Metabolism of monofluoro- and monochlorobenzoates by a denitrifyijng bacterium. Arch Microbiol 122:301–306 228. Tor JM, Xu C, Stucki JM, Wander MM, Sims GK (2000) Trifluralin degradation under microbiologically induced nitrate and Fe(III) reducing conditions. Environ Sci Technol 34:3148–3152 229. Torniainen K (1997) Photodegradation of ciprofloxacinin aqueous solution. Dissertationes Biocentri Viikki Universitatis Helsingiensis, Vol 9 230. Uotila JS, Kitunen VH, Saastamoinen T, Coote T, Hägblom MM, Salkinoja-Salonen M (1992) Characterization of aromatic dehalogenases of Mycobacterium fortuitum CG-2. J Bacteriol 174:5669–5675 231. Uotila JS, Kitunen VH, Coote T, Saastamoinen T, Salkinoja-Salonen M, Apajalahti JHA (1995) Metabolism of halohydroquinones in Rhodococcus chlorophenolicus PCP-1 Biodegradation 6:119–126 232. Utkin I, Dalton DD,Wiegel J (1995) Specificity of reductive dehalogenation of substituted orthochlorophenols by Desulfitobacterium dehalogenans JW/IU-DC1. Appl Environ Microbiol 61:346–351 233. van Agteren MH, Keuning S, Janssen DB (1998) Chloroethenes. In: Handbook on biodegradation and biological treatment of hazardous organic compounds. Kluwer Academic Publishers, Dortrecht, chap 3.2, pp 93–109; van Agteren MH, Keuning S, Janssen DB (1998) Chloroethenes and related compounds. In: Handbook on biodegradation and biological treatment of hazardous organic compounds. Kluwer Academic Publishers, Dortrecht, chap 3.3, pp 111–125 234. van Agteren MH, Keuning S, Janssen DB (1998) Aromatic compounds. In: Handbook on biodegradation and biological treatment of hazardous organic compounds. Kluwer Academic Publshers, Dortrecht, chap 4, pp 189–286; van Agteren MH, Keuning S, Janssen DB (1998) Polyaromatic hydrocarbons (PAHs). Handbook on biodegradation and biological

Degradation and Transformation of Organic Fluorine Compounds

235.

236.

237. 238. 239. 240. 241. 242. 243. 244.

245. 246. 247. 248. 249. 250. 251.

252. 253.

201

treatment of hazardous organic compounds. Kluwer Academic Publshers, Dortrecht, chap 5, pp 287–249 van der Bolt FJT, van der Heuvel RHH,Vervoort J, van Berkel WJH (1997) 19F NMR study on the regiospecificity of hydroxylation of tetrafluoro-4-hydroxybenzoate by wild-type and Y385F p-hydroxybenzoate hydroxylase: evidence for a consecutive oxygenolytic dehalogenation mechanism. Biochemistry 36:14,192–14,201 van der Tweel WJJ, Kok JB, de Bont JAM (1987) Reductive dechlorination of 2,4dichlorobenzoate to 4-chlorobenzoate and hydrolytic dehalogenation of 4-chloro-, 4bromo-, and 4-iodobenzoate by Acaligenes denitrificans NTB-1. Appl Environ Microbiol 53:810–815 Vargas C, Song B, Camps M, Häggblom MM (2000) Anaerobic degradation of fluorinated aromatic compounds. Appl Microbiol Biotechnol 53:342–347 Wackett LP, Sadowsky MJ, Newman LM, Hur H-G, Li S (1994) Metabolism of polyhalogenated compounds by a genetically engineered bacterium. Nature (London) 368: 627–629 Walsh C (1982) Fluorinated substrate analogs: routes of metabolism and selective toxicity. Adv Enzymol 55 :187– 288 Wang E, Walsh C (1978) Suicide substrates for the alanine racemase of Escherichia coli B. Biochemistry 17:1313–1324 Wang EA, Walsh C (1981) Characteristics of b,b-difluoroalanine and b,b,b-trifluoroalanine as suicide substrates for Escherichia coli B racemase. Biochemistry 20: 1739–1746 Warhurst AM, Clarke KF, Hill RA, Holt RA, Fewson CA (1994) Metabolism of styrene by Rhodococcus rhodochrous NCIMB 13259. Appl Environ Microbiol 60:1137–1145 Wataya Y, Matsuda A, Santi DV, Bergstrom DE, Ruth JL (1979) trans-5-(3,3,3-Trifluoro-1propenyl)-2′-deoxyuridylate: a mechanism based inhibitor of thymidylate synthetase. J Med Chem 22:339–340 Weber R, Hagenmeier H (1997) Synthesis and analysis of mixed chlorinated-fluorinated dibenzo-p-dioxins and dibenzofurans and assessment of formation and occurrence of the fluorinated and chlorinated-fluorinated dibenzo-p-dioxins and dibenzofurans. Chemosphere 34:13–28 Welch JT, Eswarakrishnan S (1991) Fluorine in bioorganic chemistry. Wiley, New York Wetzstein H-G, Schmeer N, Karl W (1997) Degradation of the fluoroquinolone enrofloxacin by the brown-rot fungus Gloeophyllum striatum: identification of metabolites. Appl Environ Microbiol 63 :4272– 4281 Wetzstein H-G, Stadler M, Tichy H-V, Dalhoff A, Karl W (1999) Degradation of ciprofloxacin by basidiomycetes and identification of metabolites generated by the brown rot fungus Gloeophyllum striatum. Appl Environ Microbiol 65:1556 –1563 Wieser M,Wagner B, Eberspächer J, Lingens F (1997) Purification and characterization of 2,4,6-trichlorophenol-4-monooxygenase, a dehalogenating enzyme from Azotobacter sp strain GP1. J Bacteriol 179:202–208 Wigmore GJ, Ribbons DW (1981) Selective enrichment of Pseudomonas spp defective in catabolism after exposure to halogenated substrates. J Bacteriol 146:920–927 Wofford BA, Jackson MC, Hartz C, Bevan JW (1999) Surface wave plasma abatement of CHF3 and CF4 containing semiconductor process emissions. Environ Sci Technol 33:1892–1897 Wolt JD, Schwake JD, Batzer FR, Brown SM, McKendry LH, Miller JR, Roth GA, Stanga MA, Portwood D, Holbrook DL (1992) Anaerobic aquatic degradation of flumetsulam [N-(2,6difluorophenyl)-5-methyl[1,2,4]triazolo[1,5a]pyrimidine-2-sulfonamide]. J Agric Food Chem 40:2302–2308 Workman S, Woods L, Gorby YA, Fredrickson JK, Truex MJ (1997) Microbial reduction of vitamin B12 by Shewanella alga strain BrY with subsequent transformation of carbon tetrachloride. Environ Sci Technol 31:2292–2297 Xun L, Orser CS (1991) Purification and properties of pentachlorophenol hydroxylase, a flavoprotein from Flavobacterium sp strain ATCC 39723. J Bacteriol 173:4447–4453

202

A.H. Neilson · A.-S. Allard

254. Xun L, Topp E, Orser CS (1992) Purification and characterization of a tetrachloro-p-hydroquinone reductive dehalogenase from a Flavobacterium sp. J Bacteriol 174:8003–8007 255. Yin H, Wada Y, Kitamura T, Yanagida S (2001) Photoreductive dehalogenation of halogenated benzene derivatives using ZnS or CdS nanocrystallites as photocatalysts. Environ Sci Technol 35 : 227 – 231 256. Yuan C-S, Yeh J, Liu S, Borchardt RT (1993) Mechanism of inactivation of S-adenosylhomocysteine hydrolase by (E)-5′,6′-didehydro-6′-deoxy-6′-halohomoadenosines. J Biol Chem 268 :17,010 –17,137 257. Yuan C-S, Liu S, Wnuk SF, Robins MJ, Borchardt RT (1994) Mechanism of inactivation of S-adenosylhomocysteine hydrolase by (E)-5′,6′-didehydro-6′-deoxy-6′-halohomoadenosines. Biochemistry 33:3758–3765 258. Zechel DL, Reid SP, Nashiru O, Mayer C, Stoll D, Jakeman DL, Warren RAJ, Withers SG (2001) Enzymatic synthesis of carbon-fluorine bonds. J Am Chem Soc 123:43450–43451 259. Zeyer J, Wasserfallen A, Timmis KN (1985) Microbial mineralization of ring-substituted anilines though an ortho-cleavage pathway. Appl Environ Microbiol 50:447–453

CHAPTER 7

Effects on Rodents of Perfluorofatty Acids Joseph W. DePierre Unit for Biochemical Toxicology, Department of Biochemistry and Biophysics, Wallenberg Laboratory, Stockholm University, 109 61 Stockholm, Sweden E-mail: [email protected]

Perfluorofatty acids are used in increasing amounts as corrosion inhibitors, anti-wetting agents, surfactants, and in fire extinguishers. The perfluorofatty acids whose biological effects have been studied most extensively are perfluorooctanoic and perfluorodecanoic acids. The most dramatic effect of these xenobiotics in rats and mice is hepatic peroxisome proliferation, i.e., a considerable increase in the size and number of hepatic peroxisomes, which is almost invariably accompanied by potent up-regulation of peroxisomal fatty acid b-oxidation. However, these compounds also elicit numerous other responses in these rodents, including decreased body weight, liver hypertrophy, a decrease in the size of hepatic mitochondria, decreased circulating levels of thyroid hormones, altered expression of a number of other enzymes, and the appearance of tumors in the liver and testis. Perfluorooctanoic and perfluorodecanoic acids will continue to be important tools for investigating basic cellular processes and may even turn out to be of clinical use. The risk to human health posed by exposure to these compounds in the occupational and general environments remains to be elucidated. Keywords. Perfluorofatty acids, Perfluorooctanoic acid, Perfluorodecanoice acid, Peroxisome proliferation, Fatty acid b-oxidation, Catalase, Hepatic, Lipids, Hepatic lipid metabolism, Hypolipidemia, CYP4A1, Oxidative stress, Hepatocarcinogenecity, Mice, Rats

1

Introductory Remarks . . . . . . . . . . . . . . . . . . . . . . . . 205

2

Use and Occurrence of Perfluorofatty Acids

3

Experimental Systems for Studying the Effects of Perfluorofatty Acids on Mammalian Cells . . . . . . . . . . . . . . . . . . . . . . 208

4

Pharmacokinetics and Metabolism of PFOA and PFDA in Rats . . 210

5

Effects of Perfluorofatty Acids on the Number, Size, and Functions of Peroxisomes in Rodent Liver . . . . . . . . . . . . . . . . . . . 212

5.1 5.2 5.2.1 5.2.2 5.3

Morphological Studies . . . . . . . . . . . . . Effects on Peroxisomal Proteins and Functions Fatty Acid b-Oxidation . . . . . . . . . . . . . Catalase . . . . . . . . . . . . . . . . . . . . . Sex and Species Differences . . . . . . . . . .

. . . . . . . . . . . . 207

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

212 212 212 214 216

The Handbook of Environmental Chemistry Vol. 3, Part N Organofluorines (ed. by A.H. Neilson) © Springer-Verlag Berlin Heidelberg 2002

204 5.4

J.W. DePierre

5.5 5.6 5.7

Dependence on Chain Length, the Carboxylic Acid Moiety, and Fluorination . . . . . . . . . . . . . . . . . . . . . . Dependence on Dose and Time . . . . . . . . . . . . . . Reversibility/Persistence . . . . . . . . . . . . . . . . . . Tissue Specificity . . . . . . . . . . . . . . . . . . . . . .

6

Additional Effects of Perfluorofatty Acids in Rodents . . . . . . . 224

6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9

Additional Effects on Hepatic Lipid-Metabolizing Enzymes, Lipid-Binding Proteins, and Lipid Composition . . . . . . Hypolipidemia . . . . . . . . . . . . . . . . . . . . . . . . “Wasting Syndrome”: Loss of Body Weight and Body Fat . Hepatomegaly . . . . . . . . . . . . . . . . . . . . . . . . . Decrease in Mitochondrial Size . . . . . . . . . . . . . . . Decreases in the Levels of Thyroid Hormones . . . . . . . Up-Regulation of CYP4A1 . . . . . . . . . . . . . . . . . . Up-Regulation of UDP-Glucuronyltransferase . . . . . . . Oxidative Stress . . . . . . . . . . . . . . . . . . . . . . . .

7

Mechanism(s) Underlying These Effects of Perfluorofatty Acids . . 230

7.1 7.2

Formation of Perfluorofatty Acyl-CoA and/or of Dicarboxylic Fatty Acids and/or Disruption of Fatty Acid Homeostasis in Other Ways 230 Peroxisome Proliferator-Activated Receptor-Alpha . . . . . . . . . 231

8

Toxicity/Genotoxicity of Perfluorofatty Acids

8.1 8.2 8.3 8.4 8.5 8.6 8.6.1 8.6.2 8.6.3 8.6.4 8.6.5 8.6.6

Acute Toxicity: LD50 Values, the “Wasting Syndrome”, and Acute Tissue Damage . . . . . . . . . . . . . . . Developmental Toxicity . . . . . . . . . . . . . . . . Degeneration of Seminiferous Tubules . . . . . . . . Immunotoxicity? . . . . . . . . . . . . . . . . . . . . Genotoxicity . . . . . . . . . . . . . . . . . . . . . . Possible Genotoxic Mechanism(s) . . . . . . . . . . . Lack of Direct Genotoxicity . . . . . . . . . . . . . . Increased Oxidative Stress . . . . . . . . . . . . . . . Altered Xenobiotic Metabolism . . . . . . . . . . . . Stimulation of Hepatocyte Proliferation . . . . . . . . Inhibition of Hepatocyte Apoptosis . . . . . . . . . . Immunotoxicity? . . . . . . . . . . . . . . . . . . . .

9

Studies on Humans . . . . . . . . . . . . . . . . . . . . . . . . . . 237

10

Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . 237

10.1 10.2 10.3

Valuable Experimental Tools . . . . . . . . . . . . . . . . . . . . . 237 Possible Clinical Applications . . . . . . . . . . . . . . . . . . . . 238 Hazard to Human Health? . . . . . . . . . . . . . . . . . . . . . . 238

11

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239

. . . .

. . . .

. . . . . . . . .

. . . .

. . . . . . . . .

. . . .

. . . . . . . . .

. . . .

. . . . . . . . .

217 218 222 222

224 226 226 226 227 228 228 229 230

. . . . . . . . . . . 233 . . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

233 233 233 233 234 234 234 235 235 235 236 236

Effects on Rodents of Perfluorofatty Acids

205

1 Introductory Remarks Biochemical toxicology is a two-way avenue. In the one direction, biochemical approaches are employed in attempts to elucidate the mechanisms underlying toxicological/genotoxicological phenomena. For instance, how does the widely used pain-killer and anti-fever agent paracetamol kill hepatocytes? How does aflatoxin, a fungal metabolite, cause liver cancer in widespread areas of Africa? How does cigarette smoke cause chronic bronchitis and lung cancer? In the other direction, the responses of, in particular, mammalian cells to xenobiotics have found extensive use as valuable tools for studying basic cellular processes. For instance, how is the expression of certain genes regulated? How are particular cellular compartments expanded when circumstances so require and how does the enlarged compartment return to normal after removal of the provocative and often stressful stimulus? As will be seen, studies with peroxisome proliferators, including the perfluorofatty acids, are a perfect example of this two-way avenue. The peroxisome is a membrane-bound subcellular compartment, i.e., a cell organelle [102, 125 and references therein]. In electron micrographs peroxisomes appear to be circular with a wide variety of diameters (approximately 0.5–1.0 µm in hepatocytes and smaller microperoxisomes in many extrahepatic tissues), are surrounded by a single phospholipid bilayer, and exhibit a relatively amorphous matrix, with the exception of occasional crystalline structures (crystalline urate oxidase and, perhaps, other proteins as well) in hepatic peroxisomes (see Fig. 1). Even though peroxisomes account for no more than a few percent of cellular protein and volume under normal physiological conditions, a number of important functions are localized in this organelle. Many of these functions are related to global lipid homeostasis, e.g., b-oxidation of fatty acids (see also below); synthesis of bile acids, plasmologens (ether phospholipids), and isoprenoid compounds such as cholesterol, ubiquinone, and dolichol, and fatty acid elongation and rearrangement. Other peroxisomal functions include inactivation of reactive forms of oxygen (e.g., by catalase and epoxide hydrolase), oxidation of aliphatic alcohols (also via catalase); conversion of D-amino acids to the corresponding Lamino acids, and uric acid catabolism. Much characterization of peroxisomes remains to be performed and, undoubtedly, additional important cellular processes are also carried out by this organelle. As indicated by the term peroxisome proliferator, the definitive characteristic of such a compound is its ability to evoke an increase in the number of peroxisomes present in hepatocytes and/or other cell types (Fig. 1). However, a number of other physiological and biochemical changes also typically occur in association with this proliferation (see below). For instance, up-regulation of the three enzymes involved in peroxisomal fatty acid b-oxidation is virtually always observed in association with peroxisome proliferation. Thus, such up-regulation is routinely used as an indicator of peroxisome proliferation, since morphometric study of electron micrographs (i.e., actually counting the number of peroxisomal profiles present per hepatocyte) is a laborious and time-consuming procedure. However, it should always be kept in mind that under some circumstances pro-

206

J.W. DePierre

Fig. 1. Peroxisome proliferation in mouse hepatocytes in response to dietary exposure to PFOA. The upper electron micrograph depicts the liver of untreated mice, while the lower micrograph shows the liver of treated animals. Px=peroxisomes. The arrows point to crystalline cores within these organelles. A length of 1 micron is indicated in the lower electron micrograph. Electron micrographs courtesy of Professor Anders Bergstrand. Unpublished studies in our laboratory (1995)

Effects on Rodents of Perfluorofatty Acids

207

Fig. 2. Structures of commonly studied peroxisome proliferators related (left) and unrelated (right) to clofibrate

liferation of peroxisomes can occur without up-regulation of the fatty acid catabolism localized in this organelle (see, for example, Sect. 7.2) and the opposite situation is certainly also conceivable. Long-term exposure of rodents to peroxisome proliferators promotes the formation of liver tumors and is also associated with an elevated incidence of testis cancer. At present, more than 1000 different xenobiotics have been found to belong to the class of peroxisome proliferators (for reviews, see [7, 23, 28, 47, 48, 73, 89, 92]). These compounds include clinical drugs (e.g., hypolipidemic drugs of the fibrate family, acetylsalicylic acid, and other non-steroidal anti-inflammatory drugs), industrial chemicals (e.g., phthalate plasticizers, perfluorofatty acids, di(2-ethylhexyl)phosphate, trichloroethylene, chlorinated paraffins), and agricultural chemicals (e.g., phenoxyacetic acids). The structures of some of the most commonly studied peroxisome proliferators are depicted in Fig. 2.

2 Use and Occurrence of Perfluorofatty Acids The wide variety of chemicals which cause peroxisome proliferation in rodent liver makes it difficult to propose a unifying hypothesis concerning the molecular mechanism underlying this phenomenon (see Sect. 7). The present review focuses on the biological effects of perfluorofatty acids and the analogous sulfonic acid for a number of reasons. Because of their hydrophobicity and relative chemical and thermal stabilities, in addition to the fact that they can be produced at relatively low cost, these compounds are finding increasing use as, among other

208

J.W. DePierre

things, corrosion inhibitors, anti-wetting agents, fire extinguishers, and surfactants. Since perfluorofatty and perfluorosulfonic acids are also metabolized poorly, if at all, at least in rodents (see Sect. 4), these substances would be expected to accumulate in the general environment. However, to date, the carboxylic acid does not appear to have leaked into the environment. The levels of perfluorooctanoic acid in the serum of members of the general population are 10–100 parts per billion, although these levels are, as expected, considerably higher in occupationally exposed workers [36]. A study in 1974 reported that the average level of organic fluorine in human plasma was approximately 26 ng/ml and among the fluorine-containing compounds present, perfluorooctanoic acid was tentatively identified [54]. A recent study by Hansen and coworkers [54] confirmed the presence of perfluorooctanoic acid, perfluorooctane sulfonic acid, and perfluorohexane sulfonic acid in human serum at average levels of 6.4, 28.4, and 6.6 ng/ml, respectively. Perfluooctanoic acid is not among the various fluorinated long-chain carboxylic acids detected in plants, nor does there appear to be any other natural source of this compound [49]. The possible adverse biological effects of perfluorooctane sulfonic acid are of growing concern [127], since this compound has been found to occur ubiquitously in marine mammals inhabiting widely spread geographical biospheres [70]. Although much less is presently known about the responses of living organisms to this compound than to the corresponding perfluorooctanoic acid, it is clear that the sulfonic acid elicits the same degree of peroxisome proliferation and related effects in rodent liver [142] (Table 2).

3 Experimental Systems for Studying the Effects of Perfluorofatty Acids on Mammalian Cells Virtually all studies on the effects of exposure to perfluorofatty acids involve perfluorooctanoic acid (PFOA) and perfluorodecanoic acid (PFDA) (Fig. 3). The reasons for this are that the responses of mammalian cells to shorter perfluorofatty acids are considerably less pronounced (see below) and that longer perfluorofatty acids are not readily commercially available. As will become evident below, al-

Fig. 3. Structures of the most commonly studied perfluorofatty acids

Effects on Rodents of Perfluorofatty Acids

209

though the effects of PFOA and PFDA on rats and mice are in many respects similar, there are also important differences. Indeed, it is remarkable how large an effect the presence of two extra CF2 groups can have on certain responses. Obviously, PFOA and PFDA are analogues of naturally occurring non-fluorinated fatty acids, which is almost certainly of central relevance to the mechanism(s) by which they exert their effects. As has been examined in detail [160], extraction and isolation of PFOA and PFDA requires some care. These compounds can be readily quantitated as their benzyl [180] or methyl [84] esters employing gas chromatography or by highperformance liquid chromatography after derivatization with 3-bromoacetyl-7methoxycoumarin [111]. More convenient for many studies is the use of [14C]PFOA or -PFDA, allowing radiometric quantitation [e. g., 155, 157, 158]. Definitive quantitation requires confirmation of the structures and this has been carried out [54] using HPLC interfaced to an atmospheric pressure tandem mass spectrometer operating in the electrospray negative mode. All reports in the literature concerning the effects of perfluorofatty acids on mammals involve the use of rats and mice as the experimental animals. Sometimes isolated rat hepatocytes constitute the experimental system. The simple reason for this is that, as is the case for other peroxisome proliferators as well, rats and mice are more responsive to PFOA and PFDA than are other rodents and mammals. Other species (e.g., hamsters and guinea pigs) are only examined for purposes of comparison to the rat or mouse. No studies on non-mammalian organisms, such as fish and plants, have yet been reported. In most cases male animals are studied, which is of considerable importance in the case of rats, since the female of this species does not respond to PFOA, at least not at the doses commonly employed. The strains of rats most commonly utilized in these investigations are Sprague-Dawley,Wistar, and Fischer-144 rats. The mice most commonly employed are of the C57Bl/6 strain. Virtually all studies concerning the biological effects of PFOA and PFDA focus on the liver, although it is apparent that these substances also influence other tissues, including adipose tissue, the thymus and spleen, the testis, and the heart (see below). PFOA and PFDA are almost always administered to rats by intraperitoneal injection, although dietary exposure has also been employed in some studies. Usually, a single such injection is performed and the animals monitored thereafter for 1–4 weeks. However, other conditions – e.g., a total of four intraperitoneal injections at two-week intervals – are sometimes used. The dose injected varies between 0.3 mg (= 0.73 mmoles PFOA or 0.58 mmoles PFDA) and 80 mg (=190 µmoles PFOA or 160 mmoles PFDA)/kg body weight. In contrast, PFOA is always administered to mice in their diet. This exposure normally lasts for 1–2 weeks and the doses employed vary between 0.005 and 0.05wt%. Since a 20-g mouse consumes approximately 3 g of chow per day and the corresponding value for a 200-g rat is 10 g, these dietary levels result in ingestion of 7.5 mg (=18 mmoles of PFOA or 15 mmoles of PFDA) to 75 mg (=180 mmoles PFOA or 150 mmoles PFDA) per day. Although far more than 90% of published reports concerning the biological effects of PFOA and PFDA on rodents involve administration of these substances

210

J.W. DePierre

in the diet or by intraperitoneal injection, inhalation studies and administration by gavage have also been reported.Although it is the acid itself, rather than a salt, which is commonly administered (dissolved in an oil or mixed as a powder with the animal’s food), the carboxylic acid moiety of these substances must certainly be deprotonated to at least some extent in biological fluids. Thus, as is often the case in connection with studies in the area of biochemical toxicology, a wide variety of different routes, frequencies, and periods of administration, as well as different doses of PFOA and PFDA, are utilized by different investigators. However, with the exception of marked differences between responses to PFOA and PFDA and between the responses of male and female rats described below, I have not been able to discern any consistent differences resulting from the use of such different experimental conditions. Of course, the responses differ in their extent, i.e., quantitatively, but they appear to be qualitatively similar. Therefore, I have chosen to compare different investigations rather freely in the present review.

4 Pharmacokinetics and Metabolism of PFOA and PFDA in Rats Obviously the pharmacokinetics and metabolism of PFOA and PFDA in rats and mice are of fundamental significance with regards to their biological effects. Xenobiotics commonly exert primary effects on the organs in which they accumulate and if they do not accumulate at all, because of rapid metabolism to an inactive substance(s) and/or elimination, they are unlikely to have any major influence on the organism. In addition, the metabolites of a xenobiotic may actually elicit more or less pronounced and/or different responses compared to the parent compound itself. The pronounced differences in the pharmacokinetics of PFOA in male and female rats has received considerable attention [41, 43, 53, 71, 72, 81, 85, 154, 157, 158, 178, 179, 181]. Female rats eliminated 91% of a single intraperitoneal dose of PFOA in their urine within 24 h, whereas the corresponding value for male rats was only 6% [158]. Consequently, the half-time for elimination of this substance was 15 days for the male animals, but <1 day for the females [53, 158]. Four days after administration, the level of PFOA in the serum of male rats was 17–40 times higher than in female animals [179], and 28 days after administration the concentrations in various tissues were 6–9 times higher in males than in females [81]. The highest concentrations were observed in the liver and plasma of male rats and in the liver, plasma, and kidney of female rats [158]. Several studies involving castration, ovarectomy, and administration of sex hormones have demonstrated that estradiol promotes urinary excretion of PFOA, whereas testosterone decreases this excretion [71, 72, 84, 179]. The effect of testosterone appears to dominate over the effect of estradiol in determining the pharmacokinetics of PFOA in rats. The molecular mechanisms by which these sex steroids influence the excretion of PFOA remain to be elucidated. Interestingly, comparison of the pharmacokinetics of PFDA to those of PFOA in male and female rats reveals quite different patterns. Twenty-eight days after

Effects on Rodents of Perfluorofatty Acids

211

a single intraperitoneal administration of PFDA, 51% and 24% of the dose given had been eliminated in the feces of male and female rats, respectively [157, 178]. During this same period, <5% of the total dose was recovered in the urine of animals of either sex. The half-time for elimination of PFDA from male rats was calculated to be 23 days, whereas the corresponding value for the female animals was 45 days. The highest levels of PFDA were observed in the order liver>plasma>kidneys in both sexes, while much lower amounts were present in the heart, fat pads, testis, muscle, and ovaries [157]. Thus, PFDA is eliminated from the rat more slowly than PFOA (which probably explains the more pronounced toxicity of the former; see below), as well as via a different route (i.e., via the bile and urine, respectively). Furthermore, elimination of PFDA does not exhibit the same sex difference. This slower elimination of PFDA may reflect the fact that 99% of this compound in the serum is bound to protein [178]; however, the corresponding value for PFOA has not been reported. Curiously, one study reported covalent binding of small amounts (0.1–0.5% of the total dose) of both PFOA and PFDA to proteins in the liver, plasma, and testis of male rats, although no metabolism was observed (see also below) [159]. Sulfhydryl groups on the proteins were apparently involved in this binding. In another study the effects of perfluorofatty acids of different lengths on hepatic peroxisomal fatty acid b-oxidation in male and female rats were compared [85]. In the male animals perfluorohexanoic acid elicited no response, whereas C8 , C9 , and C10 perfluorofatty acids potently up-regulated this activity. In female rats the effects of these different compounds were the same as in the male, with the exception of perfluorooctanoic acid, which elicited much less pronounced responses in females, as expected from the pharmacokinetic differences described above. There was a significant correlation between the hepatic concentration of each compound and its potency. Peroxisomal fatty acid b-oxidation in hepatocytes isolated from male and female rats was equally responsive to perfluorooctanoic and perfluorononanoic acids [85], further demonstrating that the sex differences observed in vivo reflect pharmacokinetic differences. It has been suggested that the relatively higher water solubility of perfluoroheptanoic and perfluorooctanoic acid favors their excretion in the urine, while the greater hydrophobicity of C9–11 perfluorofatty acids promotes their elimination in the bile, with subsequent reuptake in the intestine, i.e., enterohepatic recirculation [43]. Investigations designed to detect metabolites of PFOA or PFDA have all failed [41, 155, 157, 158, 179]. These studies have attempted to identify polar metabolites of these compounds in urine or bile, lipids containing these fatty acid analogues, and/or loss of fluorine (which could potentially be catalyzed by the cytochrome P-450 system, which can dehalogenate certain organic substances [124]), all unsuccessfully. Nor were acyl-CoA species containing PFOA or PFDA detected in rat liver or in isolated rat hepatocytes [88]. Thus, PFOA and PFDA appear to be excreted from rats without prior metabolism. A finding which is not consistent with this conclusion is the report that chromosomal aberrations are caused by PFDA only after activation by a post-mitochondrial fraction from rat liver (see Sect. 8.6.1).

212

J.W. DePierre

Table 1. Initial time-course of the proliferation of peroxisomes in mouse liver upon dietary exposure to 0.02 wt% perfluorooctanoic acid

Period of treatment (days)

Average number of peroxisomes per unit area (% of control)

Average size of peroxisomes (% of control)

None (control) 1 2 3 4

100 125 170 210 300

100 130 200 220 220

Unpublished data from our laboratory (1998).

5 Effects of Perfluorofatty Acids on the Number, Size, and Functions of Peroxisomes in Rodent Liver 5.1 Morphological Studies

Definitive demonstration that a xenobiotic is a peroxisome proliferator requires electron microscopic studies, preferably quantitative (i.e., morphometry). Several studies have established that the number of peroxisomes in rat and mouse liver is increased upon exposure to PFOA or PFDA [56, 57, 68, 140, 143]. One of these studies revealed more extensive peroxisome proliferation in centrilobular than in periportal rat hepatocytes [68]. We have conducted a detailed morphometric study of peroxisome proliferation in the liver of mice exposed to 0.02 wt% PFOA in their diet. It can be seen from Table 1 (unpublished results from our laboratory, 1998) that after four days of such treatment, the number of hepatic peroxisomes had increased 3-fold, while their average size had increased 2.2-fold. (As can be seen from Table 5, these effects were almost maximal.) 5.2 Effects on Peroxisomal Proteins and Functions 5.2.1 Fatty Acid b -Oxidation

Peroxisomal fatty acid b-oxidation can be quantitated both as the activity of the total process, e.g., oxidation of palmitoyl-CoA, or as the activity of the initial enzyme in this pathway, i.e., acyl-CoA oxidase, which is thought to be rate-limiting for the entire pathway [67, 126, 131]. In contrast to the corresponding mitochondrial catabolism, which conserves the energy released by the first step in the form of FADH2 , the first step in peroxisomal fatty acid b-oxidation produces hydrogen peroxide (Fig. 4). Thus, quantitation of this hydrogen peroxide provides

Effects on Rodents of Perfluorofatty Acids

213

Fig. 4. Mitochondrial (on the left) and peroxisomal (on the right) fatty acid b-oxidation

both an assay for the peroxisomal pathway and a means of distinguishing this pathway from mitochondrial catabolism as well as raising the as-yet-unanswered question as to why the energy released by the first step of peroxisomal fatty acid b-oxidation is not conserved. Indeed, peroxisomes contain a number of other oxidases which also produce hydrogen peroxide, thereby potentially causing oxidative stress in the cell (see below). Numerous studies have demonstrated that exposure of rats and mice to PFOA or PFDA results in potent increases in the hepatic activities and levels of the peroxisomal enzymes catalyzing fatty acid b-oxidation [12, 32, 43, 66, 68, 71, 82, 83, 85, 139–143, 173].

214

J.W. DePierre

Table 2. Increases in the peroxisomal activity of acyl-CoA oxidase and in the peroxisomal and

cytosolic activities of catalase in the livers of mice upon dietary administration of peroxisome proliferators (unpublished findings from our laboratory, 1997) Treatment

Enzyme activities in mmol product formed/min-g liver (%) Peroxisomal acyl-CoA oxidase

Peroxisomal catalase

Cytosolic catalase

None (control)

202±8 ¥10–6

Perfluorooctanoic acid

4.10±0.40 ¥10–3 *** (2030)

24.4±2.3 ** (200)

21.3±4.7 *** (970)

Perfluorooctane sulfonic acid

3.20±0.70 ¥10–3 ** (1580)

37.3±1.7 *** (308)

13.8±2.2 **

Clofibrate

2.05±0.50 ¥10–3 ** (1010)

18.0±1.2 ** (149)

22.0±0.5 *** (1100)

Nafenopin

2.86±0.67 ¥10–3 ** (1420)

13.4±1.9

20.1±1.4 *** (1000)

(100)

12.1±1.5

(100)

(110)

2.0±0.3

(100)

(690)

n=4. **P<0.01, ***P<0.001 compared to the control value using Student’s t-test.

5.2.2 Catalase

Peroxisome proliferators can increase the total hepatocyte capacity to produce hydrogen peroxide in connection with the reaction catalyzed by acyl-CoA oxidase by 20-fold or more (see Table 2), and this hydrogen peroxide has the potential to damage the cell by oxidizing many of its components. Therefore, it might be expected that the cellular capacity to detoxify hydrogen peroxide would also be enhanced in association with peroxisome proliferation. Catalase, which is among the proteins expressed at the very highest levels in aerobic cells, is the predominant enzyme in this connection, catalyzing the conversion of hydrogen peroxide to water (Fig. 5) [20, 33, 137]. As also shown here, in addition to its catalytic activity, catalase can use hydrogen peroxide to oxidize other substances, a so-called peroxidative activity, which may actually be favored at the concentrations of hydrogen peroxide thought to be present in vivo [20, 33, 137]. Indeed, this peroxidative activity with some as-yet-unknown endogenous substrate may provide an explanation as to why acyl-CoA oxidase and other peroxisomal oxidases produce hydrogen peroxide. PFOA and PFDA behave as typical peroxisome proliferators by causing up-regulation of hepatic peroxisomal catalase in rats and mice [66, 139–143]. Also typ-

Fig. 5. The catalytic and peroxidative reactions catalyzed by catalase

Effects on Rodents of Perfluorofatty Acids

215

ical is the fact that catalase is only up-regulated approximately 100% by these compounds (Table 2). The fact that up-regulation of acyl-CoA oxidase by peroxisome proliferators is many times greater than the associated up-regulation of catalase has led numerous investigators to postulate that peroxisome proliferators increase the level of hydrogen peroxide and, consequently, of oxidative stress in hepatocytes. However, consideration of the increases in absolute activities, rather than the percentage or -fold increases, leads to a totally different conclusion (see Table 2, based on unpublished data from our laboratory, 1997). This table illustrates that the absolute activity of hepatic peroxisomal catalase, which is 60,000 times higher than the absolute activity of hepatic acyl-CoA oxidase in untreated mice, is still in 6000-fold excess after treatment with PFOA. Of course, these values were obtained employing in vitro assay procedures, which certainly have their limitations, but even if the actual activity of catalase in vivo is only 1% of that shown in Table 2, this activity would nonetheless be 60 times greater than the maximal acyl-CoA oxidase activity. Indeed, actual measurement of indicators of oxidative stress in the livers of rodents exposed to perfluorofatty acids and other peroxisome proliferators reveal only a slight increase in such parameters (see Sect. 6.9). Another interesting observation presented in Table 2 is that peroxisome proliferators cause a much more dramatic increase in the catalase activity recovered in the high-speed supernatant (the cytosolic fraction) than in peroxisomal catalase activity. This has been observed by a number of other investigators as well [60, 62, 98]. Apparently, there are pronounced species differences with respect to the intracellular compartmentalization of catalase in hepatocytes; sheep, monkey, pig and guinea pig demonstrate extensive cytosolic activity, whereas in the untreated rat and mouse most of the hepatic catalase appears to be localized within the peroxisomes [60, 62]. However, it is difficult to demonstrate definitively that this non-sedimentable catalase is actually cytosolic and not an experimental artifact due to leakage of peroxisomes during the subfractionation procedure. In an earlier study we have attempted to approach this problem from another direction [105]. In this series of experiments we used the fact that the plasma membrane of mammalian cells can be made permeable to large molecules selectively, i.e., without permeabilizing intracellular membranes, by appropriate treatment of intact cells with digitonin [93, 174]. Upon binding of this detergent to cholesterol molecules in the plasma membrane, this membrane becomes permeable to the components of the cytoplasm, which then leak from the cell. Thus, by subsequently collecting the cells by centrifugation and washing them, the cytoplasm can be separated from the plasma membrane and intracellular structures. Using this experimental approach, we found that approximately 5% of the total cellular catalase activity in hepatocytes isolated from untreated mice could be released by incubation with digitonin, whereas the corresponding value for hepatocytes from clofibrate-treated animals was 40–50% [105]. These values agree well with those obtained utilizing subcellular fractionation (Table 2). Therefore, we conclude that perfluorofatty acids actually do cause a dramatic increase in cytosolic catalase activity.

216

J.W. DePierre

The mechanism underlying and functional consequences of this phenomenon remain obscure. Upon isolation and comparison, the peroxisomal and cytosolic forms of catalase from the liver of mice treated with peroxisome proliferators were found to be identical with respect to physicochemical properties, peptide mapping, and internal sequences (unpublished data from our laboratory, 1996), so these two enzymes are apparently the same protein coded for by a single gene. We have presented evidence that the cytosolic catalase in peroxisome-proliferated mouse liver is on its way to the peroxisomes, i.e., its import has been delayed, perhaps by competition with the enzymes catalyzing peroxisomal fatty acid b-oxidation for the protein importing machinery [104]. It is possible that the cytosolic catalase acts as a “back-up” line of defense, detoxifying hydrogen peroxide which manages to diffuse out of the peroxisomes. 5.3 Sex and Species Differences

As expected, the pronounced differences in the pharmacokinetics of PFOA in male and female rats is reflected in the responses of these animals to exposure to this compound. The peroxisome proliferation, hepatomegaly (see below), and the marked increases in the activities of peroxisomal fatty acid b-oxidation, microsomal 1-acylphosphocholine acyltransferase, cytosolic long-chain acylCoA hydrolase [71, 139] and microsomal stearoyl-CoA desaturase (which is catalyzed by cytochrome b5 and NADH-cytochrome b5 reductase) [72] observed in male rats exposed to PFOA do not occur in female rats exposed in the same manner. In marked contrast to these observations with male and female rats are our findings with male and female C57/Bl6 mice who received 0.02–0.05 wt% PFOA in their diet for five to ten days [139]; see also [151]. In this study the hepatic activities of peroxisomal fatty acid b-oxidation, peroxisomal and cytosolic catalase, microsomal w- and w-1-hydroxylation of lauric acid and cytosolic epoxide hydrolase, glutathione transferase and DT-diaphorase, as well as the protein content of the mitochondrial fraction were increased to the same extent in male and female mice upon exposure, whereas male and female Wistar rats exposed in the same manner demonstrated the usual sex differences in response, i.e., the female animals were non-responsive. The pharmacokinetics of PFOA in mice have not yet been investigated, but it is highly probable that the marked sex difference in this process observed in rats is not exhibited by mice. It remains to be seen whether other mammals more closely resemble rats or mice with respect to the responses of males and females to PFOA. Apparently, only two studies examining the effects of perfluorofatty acids in rodent species other than the rat and mouse has been conducted. In one of these investigations administration of PFOA or PFDA to guinea pigs was found to result in hepatic peroxisome proliferation which was much less pronounced than in the case of the rat [129], in agreement with observations using other peroxisome proliferators [7, 89]. The second study involved a more detailed comparison of the effects of PFDA on the rat, mouse, hamster, and guinea pig [163].Again, peroxisome proliferation

Effects on Rodents of Perfluorofatty Acids

217

was most pronounced in the rat and mouse, almost absent from the guinea pig, and intermediate in the hamster. Two associated alterations (see below), i.e., liver hypertrophy and degeneration of seminiferous tubules, demonstrated a similar pattern among these species. In contrast, thymus atrophy was seen in the mouse, hamster, and guinea pig, but not in the rat. Furthermore, accumulation of lipid droplets in the liver was more pronounced in the hamster and guinea pig than in the mouse and rat. These findings suggest that the various responses elicited by exposure to PFDA may exhibit different patterns of species dependency. More extensive examination of this question is warranted in connection with attempts to extrapolate from animals studies to man. Some information concerning peroxisome proliferation in human cells in response to exposure to PFOA is also available (see Sect. 9). 5.4 Dependence on Chain Length, the Carboxylic Acid Moiety, and Fluorination

Although there is some discrepancy between the results of different investigations on the potency of perfluorofatty acids with carbon chains of different lengths, the general consensus is that a chain length of 8–12 carbon atoms is required for eliciting maximal responses [43, 66, 68, 83, 85, 118, 156]. In most reports perfluoroacetic, perfluorobutyric, and perfluorohexanoic acids have been found to cause little change in the number of hepatic peroxisomes, liver weight, peroxisomal fatty acid b-oxidation, and catalase. However, in two cases exposure to such short perfluorofatty acids has been observed to result in significant peroxisome proliferation [68, 83]. The consensus is that in order for a xenobiotic to cause peroxisome proliferation, it must contain a relatively large hydrophobic moiety and a carboxylic acid group (either originally or after metabolic processing) [7, 89, 99, 100]. With very few exceptions, reported observations concerning the peroxisome-proliferating effects of perfluorofatty acids are in agreement with this hypothesis, indicating that the carboxylic moiety is required in this case as well. Thus, perfluorooctanol (which is assumed to be oxidized to the corresponding carboxylic acid in vivo) is an effective peroxisome proliferator, whereas perfluorooctane and perfluorodecane are not [7, 66, 83, 142]. One striking exception is the finding in our laboratory that in mice, perfluorooctane sulfonic acid (see structure above) elicits peroxisome proliferation and related changes of the same magnitude as those caused by the corresponding carboxylic acid [142], which has also been shown to be the case in rats [7, 30, 58]. It is difficult to imagine how the sulfonic acid moiety could be converted into a carboxylic acid in mammalian cells and this strange finding must be taken into consideration when formulating hypotheses concerning the mechanism(s) underlying peroxisome proliferation (see below). Administration of the non-fluorinated octanoic or decanoic acid does not elicit peroxisome proliferation and related effects in rodent liver [e.g., 142]. Presumably, this reflects the fact that these naturally occurring non-fluorinated compounds are rapidly metabolized in mammalian cells.

218

J.W. DePierre

a

b Fig. 6 a – d. Hepatic levels of various parameters related to peroxisome proliferation after

dietary exposure of C57/Bl6 mice to 0.02 wt % PFOA for different periods of time *P<0.1, **P<0.01, ***P< 0.001 = significantly different from 0-time exposure as determined by Student’s t-test. Unpublished data from our laboratory (1997): a peroxisomal b-oxidation of palmitoyl-CoA and of lauroyl-CoA assayed in the mitochondrial subfraction; b peroxisomal acyl-CoA oxidase activity with palmitoyl-CoA and lauroyl-CoA as substrates assayed in the mitochondrial subfraction; c catalase activity in peroxisomes (mitochondrial fraction) and cytosol : d microsomal omega hydroxylation of lauric acid (catalyzed by cytochrome P45 4A)

5.5 Dependence on Dose and Time

Two investigations have reported the effects of different doses of PFOA and PFDA on hepatic peroxisome proliferation and related parameters in rats. In one of these, female rats received four intraperitoneal injections of PFDA – 0.3, 1.0, 3.0, 10, or 30 mg/kg each time – at two-week intervals [14]. Peroxisomal fatty acid b-oxidation was up-regulated only by total administration of at least 12 mg PFDA/kg body weight.

Effects on Rodents of Perfluorofatty Acids

219

c

d Fig. 6 c, d (continued)

In the other study male rats received 0.0025–0.04% PFOA or 0.00125–0.01% PFDA in their diet for one week [69]. These regimens correspond to total doses of approximately 9–140 mg (22–338 mmol) and 4.5–35 mg (9–68 mmol) per kg body weight for these two substances, respectively. The dose-response curves for the various parameters investigated – including induction of peroxisomal fatty acid b-oxidation and increases in microsomal 1-acylglycerophosphocholine acyltransferase and cytosolic long-chain acyl-CoA hydrolase activities – were very similar. We have conducted a detailed investigation of the effects of different doses of PFOA administered to male mice in their diet for ten days (unpublished data, 1997). The parameters monitored were body and liver weights, mitochondrial protein, peroxisomal fatty acid b-oxidation and acyl-CoA oxidase activity, peroxisomal and cytosolic catalase activities, and microsomal w- and w-1-hydrox-

220

J.W. DePierre

a

b Fig. 7 a – d. Hepatic levels of various parameters related to peroxisome proliferation after di-

etary exposure of C57/Bl6 mice to various doses (w/w) of PFOA for ten days. *P<0.1, **P<0.01, ***P<0.001=significantly different from no exposure (control=C) as determined by Student’s t-test. Unpublished data from our laboratory (1997): a peroxisomal β -oxidation of palmitoylCoA and of lauroyl-CoA assayed in the mitochondrial subfraction; b peroxisomal acyl-CoA oxidase activity with palmitoyl-CoA and lauroyl-CoA as substrates assayed in the mitochondrial subfraction; c catalase activity in peroxisomes (mitochondrial fraction) and cytosol; d microsomal omega hydroxylation of lauric acid (catalyzed by cytochrome P450 4A)

ylation of lauric acid (i.e., cytochrome P450 4A) (see below for explanations as to why these parameters were chosen).As seen in Fig. 6a–d, significant increases in almost all of these parameters were observed at the lowest dose tested (0.001% =approximately 1.2 mg (3 mmol)/kg-day). Indeed, with the exception of fatty acid b-oxidation and catalase activity, half- or almost half-maximal effects were obtained with this lowest dose. The observation that at the lowest doses employed, acyl-CoA oxidase was up-regulated to a greater extent than overall peroxisomal fatty acid b-oxidation suggests that this enzyme may not always be rate-limiting

Effects on Rodents of Perfluorofatty Acids

221

c

d Fig. 7 c, d (continued)

for the peroxisomal catabolism of fatty acids. The much greater increase in cytosolic than in peroxisomal catalase activity, and at lower doses, is also noteworthy. In terms of the dose required to obtain maximal effects, PFOA is the most potent of the many different peroxisome proliferators which we have administered to mice. With respect to the time-course of peroxisome proliferation in these same mice, it can be seen from Table 1 that an increase in both the number and size of peroxisomes was observed after only a single day of treatment. However, the number of peroxisomes continued to increase for at least four days, whereas these organelles appear to have expanded to their maximal size after only two days. The time-courses observed for the increases in peroxisomal fatty acid b-oxidation and acyl-CoA oxidase activities, peroxisomal and cytosolic catalase, and

222

J.W. DePierre

w- and w-1-hydroxylation of lauric acid upon dietary exposure to 0.02% (=approximately 25 mg (60 mmol)/kg-day) are documented in Fig. 7a–d (unpublished data from our laboratory, 1997). Again, in most cases significant changes are observed after a single day of treatment and the maximal increases are achieved within five days of exposure. Up-regulation of w- and w-1-hydroxylation of lauric acid occurred somewhat earlier than up-regulation of acyl-CoA oxidase, in agreement with the hypothesis that dicarboxylic fatty acids are the actual peroxisome proliferators (see Sect. 7.1). 5.6 Reversibility/Persistence

Although only two reports expressly designed to examine the persistence of the responses of rodents to perfluorofatty acids have appeared, a commonly employed experimental regimen is to treat rats with a single intraperitoneal injection of PFOA or PFDA and sacrifice these animals for examination of various parameters as long as four weeks later. This assumes, of course, that the responses elicited are persistent. In one study the up-regulation of UDP-glucuronyltransferase activity in male rat liver after a single intraperitoneal dose of PFDA was found to be stable for three weeks [4]. Presumably, such persistence reflects the relatively long period of time required for elimination of PFDA by the rat (see above). The only really detailed investigation of the reversibility/persistence of the responses to perfluorofatty acids was performed in our laboratory utilizing C57/Bl6 mice exposed to 0.05 wt% PFOA for one week in their diet, followed by a 20-day recovery period [141]. Two patterns of reversibility/persistence were observed. In one of these, two of the activities that were found to be elevated, i.e., cytosolic DT-diaphorase and glutathione transferase, returned to control levels within 20 and 2 days of recovery, respectively. In the other, most of the parameters that were examined remained elevated – i.e., palmitoyl-CoA oxidation (238% of the control value after 20 days of recovery), lauroyl-CoA oxidase activity (625%), peroxisomal catalase activity (190%), cytosolic catalase activity (575%), and mitochondrial protein (a measure of the decrease in mitochondrial size; see below (230%). One possible explanation for this difference is that the increases in cytosolic DT-diaphorase and glutathione transferase activities require the presence of higher concentrations of PFOA than does elevation of the other parameters examined. 5.7 Tissue Specificity

Upon treatment of mice with PFOA, there is extensive peroxisome proliferation in the liver and a small degree of peroxisome proliferation in the kidney (unpublished observations from our laboratory, 1992). Other tissues are affected very little, if at all. This pattern is typical of that found with other peroxisome proliferators as well [7, 89].

223

Effects on Rodents of Perfluorofatty Acids Table 3. Various effects of perfluorofatty acids on rats and mice

Effects

Known to be common for peroxisome proliferators?

References

Increase in the number and size of hepatic peroxisomes Increase in hepatic peroxisomal fatty acid β -oxidation

Yes

56, 57, 68, 140, 143

Yes

Increase in hepatic catalase activity Changes in the activities/levels of various hepatic lipid-metabolizing enzymes and binding proteins

Yes Yes

Changes in the lipid composition of the liver

Yes

Hypolipidemia (i.e., decreased levels of blood fat) Hepatomegaly (i.e., liver hypertrophy)

Yes Yes

Disruptive changes in hepatic morphology “Wasting syndrome” (i.e., decrease in body weight and/or amount of body fat), usually accompanied by decreased intake of food Decrease in hepatic mitochondrial size Decrease in the circulating level of thyroid hormone Up-regulation of hepatic cytochrome p-450 4 A

No No

Up-regulation of hepatic UDP-glucuronyl-transferase Down-regulation of hepatic sulfotransferase Mild hepatic oxidative stress

Yes No Yes

Up-regulation of hepatic cytosolic epoxide hydrolase Increase or decrease in hepatic cytosolic glutathione transferase activity and levels Decrease in the hepatic level of reduced glutathione Down-regulation of hepatic selenium-dependent glutathione peroxidase Increase in hepatic cytosolic DT-diaphorase activity Up-regulation of hepatic isozymes of microsomal esterases Down-regulation of hepatic xanthine dehydrogenase Changes in hepatic levels of stress proteins (primarily heat-shock proteins) Decrease in hepatic retinal palmitate hydrolase activity

Yes No

12, 32, 43, 66, 68, 71, 82, 83, 85, 139–143, 173 66, 139–143 11, 29, 68, 86, 87, 130, 138, 145, 156, 159, 175 1, 29, 40, 86, 87, 128, 129, 165, 175 11, 58, 68, 86, 153 16, 22, 24, 56, 71, 76, 112, 147, 163 35, 162, 169 16, 35, 41, 50, 56, 90, 112, 161, 163, 164 139–143 50, 51, 74, 75, 91, 162 24, 31, 32, 82, 101, 139–143, 161, 175 4, 71, 101 168 39, 52, 65, 77, 117, 146 139–143 135, 141

No No

21 21

No No

141, 142 30, 63

No No

134 169, 172

No

123

Yes Yes Yes

224

J.W. DePierre

Table 3 (continued)

Effects

Known to be common for peroxisome proliferators?

References

Decrease in serum retinol level Increase in hepatic level of mRNA encoding PPARalpha Inhibition of hepatic glucose transport Modification of the immunoglobulin heavy chain- binding protein in hepatic microsomes Inhibition of gap junctional communication between hepatocytes Alterations in hepatic levels of eicosanoids (prostaglandins, thromboxane, leukotrienes) Decrease in the number of beta-receptor binding sites in the heart and in the heart rate Liver necrosis Increased apoptosis in cultured hepatoma cells Thymus atrophy Degeneration of seminiferous tubules Promotion of liver cancer Cancer of the testis

No No No No

123 144 42, 44 170, 171

No

150

No

38, 166, 167

No

120, 121

No No No No Yes No

80, 82 116, 136 16, 56, 163, 176 163 13, 40, 109, 125, 153 9, 27

6 Additional Effects of Perfluorofatty Acids in Rodents It seems safe to say that under no conditions will treatment of any animal with a given xenobiotic elicit only a single response. Even if, although highly unlikely, there is only a single initial response, many secondary, tertiary, etc. responses will occur, since all processes in the cell are ultimately interdependent. Indeed, as illustrated in Table 3, perfluorofatty acids cause a wide variety of different effects in rats and mice. Certain of these effects will be discussed in more detail below. 6.1 Additional Effects on Hepatic Lipid-Metabolizing Enzymes, Lipid-Binding Proteins, and Lipid Composition

Although somewhat overlapping in substrate specificity, the peroxisomal and mitochondrial catabolism of non-fluorinated fatty acids complement each other in a number of ways: the peroxisomal system prefers long-chain (C14–C20) and verylong-chain (>C20 ) saturated and unsaturated fatty acids, long-chain dicarboxylic fatty acids, and branched-chain fatty acids. In contrast, the mitochondrial path-

Effects on Rodents of Perfluorofatty Acids

225

way prefers short-, medium- and long-chain monocarboxylic acids and is unable to catabolize uncommon fatty acids to any great extent [3, 64, 152]. Thus, it might be expected that peroxisome proliferators would up-regulate both peroxisomal and mitochondrial fatty acid b-oxidation coordinately. Such coordinate regulation has been demonstrated for a number of peroxisome proliferators, which up-regulate carnitine acyltransferase (the protein which transports fatty acyl groups into the mitochondrial matrix and which is thought to be rate-limiting for mitochondrial fatty acid b-oxidation) many-fold and the enzymes of the mitochondrial catabolic pathway moderately [8, 59, 148, 182]. We have found that dietary treatment of mice with 0.02% PFOA for one week results in a 16-fold induction of hepatic carnitine acetyltransferase activity. However, this process remains to be examined in more detail, as does possible up-regulation of the enzymes involved in mitochondrial fatty acid b-oxidation upon exposure to PFOA or PFDA. At the same time, alterations in the hepatic activities of a large number of other enzymes involved in lipid metabolism, as well as in the hepatic levels of lipid-binding proteins upon exposure to PFOA or PFDA, have been reported [11, 29, 68, 86, 87, 112, 130, 145, 156, 159, 175]. Under these conditions, increases have been observed in the activities of glycerol-3-phosphate acyltransferase, diacylglycerol kinase, phosphatidylserine decarboxylase, and 1-acylglycerophosphocholine acyltransferase, whereas the activities of CTP-phosphoethanolamine cytidyltransferase and phosphatidylethanolamine N-acyltransferase are lowered [87]. The rates of cholesterol and fatty acid synthesis are apparently decreased by such treatment [29], while the rate of triglyceride formation from glycerol is increased [86]. Administration of PFDA to rats increases the hepatic levels of both acyl-CoA binding protein and fatty acid binding protein [145, 159] and exposure to PFOA increases hepatic levels of short-chain acyl-CoA species almost tenfold [68]. It is not yet clear whether all or some of these changes simply reflect the hepatic hypertrophy and increased hepatic triglyceride content elicited by PFOA and PFDA, whether the lipid content of the liver is being remodeled in some subtle way, and/or whether these changes are involved in or reflect the mechanism of action and/or toxicity exerted by these compounds (see below). An intriguing and consistent observation in the liver of rats and mice exposed to peroxisome proliferators is a pronounced up-regulation of the cytosolic acylCoA hydrolase [7, 89]. Similar up-regulation is caused by PFOA in mice (Table 5). To date, this phenomenon remains totally unexplained. A number of reports demonstrate that the total hepatic content of phospholipids is increased upon treatment of rats with PFOA or PFDA [1, 128, 129, 165]. However, this is not surprising in light of the pronounced liver hypertrophy caused by such exposure (see Sect. 6.4). Actually, the levels of phospholipid and cholesterol per hepatocyte are unaltered [128]. On the other hand, the contents of triglycerides [29, 40, 86, 87, 128] and of cholesteryl esters [29, 128] per hepatocyte have been found to be increased after exposure of rats to perfluorofatty acids. Analyses of the levels of individual phospholipid species [1, 87] and even of individual species of fatty acids [87, 175] in the liver of rats treated with PFOA or PFDA have been performed, but have not as yet revealed any consistent pattern of change.

226

J.W. DePierre

6.2 Hypolipidemia

Clofibrate, the first substance shown to be a peroxisome proliferator [7], is a hypolipidemic drug (i.e., a drug which lowers the fat content of the blood). Since then, peroxisome proliferators have generally been found to elicit hypolipidemia in rats and mice. PFOA and PFDA are no exceptions [11, 58, 68, 86, 153]. PFOA and perfluorooctane sulfonic acid have both been observed to lower the levels of triglycerides and cholesterol in the serum of Wistar rats [58]. In this same study hepatic levels of triglcyeride and free cholesterol were increased by both compounds, whereas the level of esterified cholesterol was decreased by 50%. It was suggested that impaired production of lipoprotein particles in hepatocytes, as a consequence of decreased synthesis and esterification of cholesterol and increased fatty acid b-oxidation, was the direct cause of the hypolipidemia observed. 6.3 “Wasting Syndrome”: Loss of Body Weight and Body Fat

In general, exposure to peroxisome proliferators results in a decrease in body weight or, at least, in the rate of body weight gain in rats and mice [7, 89]. PFOA and, in particular, PFDA are extreme in this respect, giving rise to such severe weight loss that the phenomenon is referred to as “wasting syndrome” [16, 35, 41, 50, 56, 90, 112, 161, 163, 164]. In one study, rats exposed to PFDA lost 40% of their body weight within four days [35]; while in another investigation, PFOA caused a 50% decrease in body weight [112]. This weight loss is associated with decreased food intake, which is particularly severe in animals administered PFDA. In the case of mice exposed to PFOA, we consistently observe a 20–25% decrease in total body weight, without any significant decrease in total food or water intake during the one-week period of dietary administration (unpublished observations, 1995). At the same time, these animals lose virtually all of their body fat. Such loss of body fat is a common observation in rodents exposed to peroxisome proliferators [7, 89], but this phenomenon can only account for a small part of the total weight loss evoked by PFDA and PFOA. The molecular mechanisms underlying these losses of both body weight and fat are at present a total mystery. 6.4 Hepatomegaly

An effect invariably associated with peroxisome proliferation is hepatomegaly (i.e., liver hypertrophy) [7, 89]. PFOA and PFDA have been found in a large number of studies to cause enlargement of the liver in both rats and mice [16, 22, 24, 56, 71, 76, 112, 147, 163]. Indeed, in our laboratory we observed that dietary treatment of a 20-g C57/Bl6 mouse with PFOA causes an increase in its liver weight from approximately 1 g to more than 2 g (at the same time as the total body

227

Effects on Rodents of Perfluorofatty Acids

weight is decreasing; see above). The liver hypertrophy caused in rats by PFDA apparently reflects enhanced cell proliferation [22], which also seems to be the case for the hepatomegaly induced by other peroxisome proliferators [7, 89]. This system thus appears to offer a useful tool for investigating control of cell proliferation. 6.5 Decrease in Mitochondrial Size

In studies on peroxisome proliferators involving subcellular fractionation, a large increase in the amount of protein recovered in the mitochondrial fraction is invariably observed [8, 59, 97, 148, 182].At first, this increase was thought to reflect either the increased number of peroxisomes recovered in this fraction and/or simultaneous proliferation of mitochondria. However, in a study involving dietary treatment of mice with three peroxisome proliferators (i.e., clofibate, nafenopin, and WY-14.643), and using differential centrifugation and electron microscopy, we demonstrated that this increase in mitochondrial protein reflects a decrease in the size of individual mitochondria [97]. The average diameter of mitochondria in peroxisome proliferator-treated mice is one-third less than in untreated animals, as a result of which the mitochondria which are normally recovered in the nuclear fraction are pelleted instead at the higher centrifugal force required to obtain the mitochondrial subfraction (see Table 4). At the same time, the number of mitochondrial profiles, at least in mouse liver, doubles, so that the total mitochondrial volume remains unchanged. These effects are completely reversible within several days after termination of the exposure. As is the case for other peroxisome proliferators, treatment of mice with PFOA causes a four- to fivefold increase in the amount of protein recovered in the mitochondrial fraction [139–143].Although this effect has not yet been definitively shown to reflect a decrease in mitochondrial size, electron microscopic studies in our laboratory (unpublished, 1999) indicate that this is the case. It remains a total mystery why mitochondria should decrease in size in response to peroxisome proliferators. One possibility we have considered is that the

Table 4. Mitochondrial parameters in the liver of mice after dietary exposure to the peroxisome

proliferators clofibrate, nafenopin orWY-14.463 [97] Parameter

Untreated mice

Treated mice

Protein in the mitochondrial fraction (mg/g liver) Relative enzyme a levels in the nuclear fraction Relative enzyme a levels in the mitochondrial fraction

5 8 4

20–25 1 4–5

a

Outer mitochondrial membrane=monoamine oxidase, microsomal glutathione transferase. Intermembrane space=adenylate kinase. Inner mitochondrial membrane=cytochrome oxidase, cytochromes c, c1 , and a. Mitochondrial matrix=glutamate and malate dehydrogenases.

228

J.W. DePierre

smaller mitochondria represent a stage in mitochondrial biogenesis. However, even after continued dietary treatment of mice with 0.02 wt% PFOA for one month, these small mitochondria do not become larger. A second possible explanation is related to the extensive up-regulation of carnitine acyltransferase which occurs upon exposure to peroxisome proliferators (see above). Proteins are closely packed in the phospholipid bilayers of the mitochondrial membranes and perhaps additional bilayer is required to accommodate newly synthesized carnitine acyltransferase. Indeed, one major effect of packaging the same mitochondrial volume into twice as many units is that the total membrane surface area is approximately doubled. This increase in surface area also provides for more extensive contact between the mitochondria and cytoplasm, the origin of the fatty acyl-CoA which is b-oxidized in this organelle. 6.6 Decreases in the Levels of Thyroid Hormones

A number of the effects of peroxisome proliferators are thyromimetic (i.e., mimic those exerted by thyroid hormone). Consequently, a link of some kind between the mechanism of action of peroxisome proliferators and that of thyroid hormone has long been postulated [6, 18, 61, 106]. One such connection has now been established by the demonstration that PPARalpha can form a heterodimer with the thyroid hormone receptor (see below). Another possible interrelationship might involve the displacement of thyroid hormones from binding sites on serum proteins by peroxisome proliferators, so that larger proportions of these hormones can diffuse into tissues and exert their effects. A number of investigations have demonstrated that exposure of rats to PFDA results in decreased levels of both thyroxine (T3) and triiodothyronine (T4) in the serum [50, 51, 74, 75, 91, 162]. However, it is not clear whether this effect is due to displacement from binding sites on serum proteins, reduced responsiveness of the thyroid and/or pituitary glands to hormonal stimulation, and/or some other phenomenon. PFDA does cause increases in a number of enzymes known to be up-regulated by thyroid hormone, i.e., malic enzyme [74, 75], glycerol-3-phosphate dehydrogenase [51, 75], and glucose-6-phosphate dehydrogenase [75]. Thus, some of the effects of PFDA and, presumably, PFOA may be secondary to the displacement of thyroid hormones from serum proteins. Also of interest in this connection is the observation that simultaneous treatment with thyroid hormone does not prevent the severe weight loss caused by PFDA [50]. 6.7 Up-Regulation of CYP4A1

Typically, peroxisome proliferators up-regulate the isozyme of cytochrome P-450 designated CYP4A1, which catalyzes w- and w-1-hydroxylation of fatty acids (Fig. 8) [7, 89]. Similar reactions with derivatives of fatty acids, i.e., eicosanoids such as prostaglandins, leukotrienes, and thromboxanes, are also catalyzed. Thus

Effects on Rodents of Perfluorofatty Acids

229

Fig. 8. Reactions catalyzed by cytochrome P-450 4A

this member of the rather large family of cytochrome P-450 proteins is one of those which appear to be specialized for the metabolism of endogenous substrates, even if it can catalyze analogous reactions with certain xenobiotics, e.g., di(2-ethylhexyl))phthalate. There is ample evidence that both PFOA and PFDA also cause up-regulation of CYP4A1 in rat and mouse liver, both of the levels of the protein and of the mRNA which encodes it [24, 31, 32, 82, 101, 139–143, 161, 175]. The fact that CYP4A1 is invariably up-regulated in association with peroxisome proliferation has led, across the years, to many investigations based on the hypothesis that it is w- and w-1- hydroxylated fatty acids, or rather dicarboxylic acids produced in vivo by enzymatic oxidation of these hydroxylated products which actually cause many of the effects attributed to peroxisome proliferators [95, 96]. This up-regulation is certainly likely to contribute to the changes in eicosanoid levels observed in connection with peroxisome proliferation by PFOA and PFDA (see Table 3). 6.8 Up-Regulation of UDP-Glucuronyltransferase

As observed with other peroxisome proliferators, treatment of male rats with PFOA or PFDA results in potent up-regulation of one of the isozymes of UDPglucuronyltransferase, both at the protein and mRNA levels [4, 78, 101]. Since this enzyme is capable of conjugating carboxylic acid groups with glucuronic acid (Fig. 9) and does indeed conjugate certain other peroxisome proliferators in this way [101], it was suspected initially that this up-regulation by perfluorofatty acids reflected substrate induction, i.e., up-regulation of their own metabolism. However, it has not been possible to detect glucuronidation of either PFOA or PFDA in rats (see Sect. 4 above).

Fig. 9. Reactions catalyzed by UDP-glucuronyltransferase

230

J.W. DePierre

6.9 Oxidative Stress

As discussed in Sect. 5.2, the potent up-regulation of acyl-CoA oxidase and much more moderate increase in catalase activity observed upon exposure of rats and mice to peroxisome proliferators have led many investigators to believe that the level of oxidative stress in hepatocytes is enhanced in connection with peroxisome proliferation. Indeed, increases in certain indicators of oxidative stress have been observed in some but not all cases, but the increases reported have been relatively small [5, 17, 26, 45, 52, 107]. In agreement with such findings with other peroxisome proliferators, Handler and coworkers [52] have reported that in the liver of intact rats exposed to PFOA, the production of hydrogen peroxide is not increased, although minor increases have been observed in other laboratories [39, 117]. PFOA and PFDA do increase the level of 8-hydroxyguanine (another indicator of oxidative stress) in the DNA of rodent liver, but, again, only to a modest extent [65, 77, 146]. Of relevance here are unpublished findings from our laboratory (1996) which also demonstrate that the level of 8-hydroxyguanine in the hepatic nuclear DNA of mice exposed to PFOA is increased by approximately 25%, whereas the level of this indicator of oxidative stress in hepatic mitochondrial DNA is actually decreased by about 70%. Thus the level of oxidative stress in different subcellular compartments may vary, depending in part on the levels of antioxidant defenses present. Reduced ubiquinone is a major endogenously produced antioxidant [110, 122] and, since it is also an essential component of the electron transport chain involved in oxidative phosphorylation in the inner mitochondrial membrane, relatively large quantities of ubiquinone are present in the neighborhood of the mitochondrial genome. In summary, available evidence suggests that, as is the case with other peroxisome proliferators, PFOA and PFDA increase the level of oxidative stress in the hepatocytes of exposed rats and mice to a limited extent only.

7 Mechanism(s) Underlying These Effects of Perfluorofatty Acids 7.1 Formation of Perfluorofatty Acyl-CoA and/or of Dicarboxylic Fatty Acids and/or Disruption of Fatty Acid Homeostasis in Other Ways

Since the vast majority of peroxisome proliferators contain a large hydrophobic moiety and a carboxylic acid group (originally or after metabolic conversion) and are in this sense analogues of naturally occurring fatty acids (see Sect. 5.4), one early hypothesis concerning the mechanism underlying peroxisome proliferation involved the formation of acyl-CoA derivatives containing the xenobiotic peroxisome proliferator [7, 89]. Indeed, certain peroxisome proliferators (e.g., clofibrate) have been shown to form acyl-CoA metabolites [80]. However, it has not been possible to demonstrate formation of acyl-CoA species containing PFOA or PFDA, either in rat liver or isolated rat hepatocytes [88]. In light of the

Effects on Rodents of Perfluorofatty Acids

231

pronounced potency of these compounds as peroxisome proliferations (see Sect. 5.5), lack of such formation must be concluded to eliminate this hypothesis. As described in Sect. 6.7, peroxisome proliferators should lead to the formation of dicarboxylic fatty acids via up-regulation of cytochrome P-450 4A. However, such formation has not been demonstrated experimentally. Using a somewhat different approach, we treated mice with different non-fluorinated dicarboxylic fatty acids and subsequently determined their hepatic acyl-CoA oxidase activity as an indicator of peroxisome proliferation [96]. We found that straight-chain, saturated dicarboxylic fatty acids containing 12, 16, or 22 carbon atoms are moderately effective up-regulators of this activity in mouse liver. Thus, dicarboxylic fatty acids may be one of the signals involved in peroxisome proliferation [7, 89, 95]. As described in Sect. 6.1, peroxisome proliferators, including PFOA and PFDA, alter hepatic lipid metabolism in exposed rats and mice in a variety of ways. Although the consequences of these changes remain to be elucidated, it is possible that higher levels of long-chain, non-fluorinated unsaturated fatty acids known to be ligands for the nuclear receptor involved in peroxisome proliferation (see below) may be produced. 7.2 Peroxisome Proliferator-Activated Receptor-Alpha

In investigations designed to elucidate the molecular mechanism(s) underlying peroxisome proliferation, an orphan nuclear receptor was found to be involved in this process and this protein was named the Peroxisome Proliferator-Activated Receptor, or PPAR. Later, other forms of this receptor, expressed primarily in tissues other than the liver, were discovered and the original form was renamed PPAR-alpha. More detailed studies have led to the model depicted in Fig. 10. (For recent reviews, see [27, 28, 47, 48, 73, 92]).

Fig. 10. Model for the activation of gene expression by PPARalpha

232

J.W. DePierre

In this model PPAR-alpha is shown to form a heterodimer with RXR, the retinoic acid receptor. Other nuclear receptors, including the thyroid hormone receptor, have also been found to form heterodimers with PPAR-alpha [6, 106]. Formation of such heterodimers allows cross-regulation of different metabolic activities, so-called cross-talk. As also shown, the best ligand for binding to RXR is 9-cis-retinoic acid. The best ligand(s) for PPAR-alpha has not yet been definitively identified, but is probably a long-chain, polyunsaturated fatty acid or a derivative thereof, e.g., a prostaglandin. The PPAR-alpha:RXR heterodimer binds to a responsive element in the 5¢ upstream region of the genes to be up-regulated and functions there as an activating transcription factor. It is possible to test this model in vivo, since 9-cis-retinoic acid is a metabolite of vitamin A, which mammals do not synthesize themselves but take up from their diet. Thus, we made C57/Bl6 mice vitamin A-deficient and then exposed these animals to PFOA [140, 143]. As illustrated in Table 5, the activity of acylCoA oxidase was not up-regulated in vitamin A-deficient mice, as predicted from the model shown above. However, all of the other parameters examined – liver hypertrophy (i.e., hepatocyte proliferation), peroxisomal number and size, mitochondrial protein (reflecting a decrease in mitochondrial size), peroxisomal catalase and cytosolic palmitoyl-CoA hydrolase activities – are all increased in the same manner in vitamin A-deficient as in vitamin A-adequate animals. It may be that PPAR-alpha is nonetheless involved in mediating at least certain of these effects, but perhaps as a heterodimer with some nuclear receptor other than RXR or, together with RXR, but with a ligand other than a metabolite of vitamin A. Another approach for testing this model in vivo is the use of mice in which the gene coding for PPAR-alpha has been mutated to produce an inactive protein, so-

Table 5. Effects of perfluorooctanoic acid on various hepatic parameters in vitamin A-deficient

mice [140, 143] Parameter

Liver weight Number of hepatic peroxisomes per unit area Average size of hepatic peroxisomes Acyl-CoA oxidase activity (per gram liver) Peroxisomal catalase activity (per gram liver) Mitochondrial protein (per g liver) Cytosolic palmitoyl-CoA hydrolase activity (per g liver)

Vitamin A-adequate mice administered PFOA (fold-increase over the control value)

Vitamin A-deficient mice administered PFOA (fold-increase over the control value)

2.2 3.5

2.3 3.3

2.5 26.1

2.9 1.1

2.3

1.8

5.1 10.9

4.9 10.7

Effects on Rodents of Perfluorofatty Acids

233

called PPAR-alpha null mice [46].We have recently (2001) treated such mice with PFOA and found that, as expected, up-regulation of acyl-CoA oxidase does not occur. However, the liver hypertrophy (Sect. 6.4) and decrease in body weight/loss of body fat (Sect. 6.3) observed in normal mice are also seen in the PPAR-alpha null animals. Clearly, then, the molecular mechanisms underlying all of the different effects of peroxisome proliferators, including the perfluorofatty acids, are not identical.

8 Toxicity/Genotoxicity of Perfluorofatty Acids 8.1 Acute Toxicity: LD50 Values, the “Wasting Syndrome”, and Acute Tissue Damage

The LD50 values for PFOA and PFDA in rats have been reported to be 189 and 41 mg/kg body weight, respectively [35, 56, 112]. Even much lower doses cause extreme loss of body weight and signs of acute liver damage, including hepatocyte swelling and disruption of intracellular organelles [2, 56, 79, 163]. 8.2 Developmental Toxicity

In the only reported study on the developmental (or reproductive) toxicity of perfluorofatty acids, it was found that when female mice were treated with PFDA, fetal weight gain and viability were lowered [55]. However, since these alterations were observed only with doses which resulted in decreased maternal weight, it was concluded that these effects on the fetus were secondary rather than primary. 8.3 Degeneration of Seminiferous Tubules

One study has demonstrated degeneration of seminiferous tubules in rats and mice and, to a lesser extent, hamsters and guinea pigs exposed to PFDA [163]. This alarming observation merits examination in more detail. 8.4 Immunotoxicity?

As documented in Table 3, in at least four different studies, including one from our own laboratory, thymus atrophy was observed in animals treated with perfluorofatty acids [16, 56, 163, 176]. We also observed atrophy of the spleen to approximately half its normal size in mice exposed to PFOA [176]. These two organs play central roles in the immunological defenses of mammals and a highly interesting question is whether thymic and splenic atrophy result in and/or reflect an impairment of these defenses, i.e., immunotoxicity. Apparently, to date only a single study has addressed this question specifically [108]. In this investigation rats were administered PFDA and then injected with

234

J.W. DePierre

an antigen (keyhole limpet hemocyanin). Eight days later the production of specific antibodies by the treated animals was significantly lower than in untreated mice, although this difference had disappeared 30 days after immunization. The possible immunotoxicity of perfluorofatty acids is of considerable toxicological interest (see Sect. 8.6) and is now under intensive investigation in our laboratory. 8.5 Genotoxicity

Neither PFOA nor PFDA by itself elicits the formation of liver tumors in rats and mice [13, 40, 153]. However, both of these compounds do enhance the frequency of liver tumors caused by a given dose of another substance (e.g., diethylnitrosamine), i.e., both PFOA and PFDA are promoters of hepatocarcinogenesis. Indeed, peroxisome proliferators are in general promoters of liver tumors, although with varying degrees of potency [7, 89, 109, 125]. Both PFOA and PFDA are known to cause testis cancer – more precisely, Leydig cell adenomas – in rats and the mechanism underlying this form of genotoxicity has been elucidated in some detail [9, 10, 15, 27]. In rats exposed to PFOA or PFDA, the weight of the testis is decreased, the serum level of testosterone increased, and serum estradiol lowered. Perfluorofatty acids appear to inhibit the synthesis and/or release of testosterone by the Leydig cells. The relationship of this inhibition to the induction of testis cancer remains to be elucidated. 8.6 Possible Genotoxic Mechanism(s)

The mechanisms by which PFOA and PFDA exert their genotoxicity – in particular, how these compounds promote hepatocarcinogenesis – is of considerable interest in connection with attempts to determine whether these compounds can also increase the frequency of liver cancer in human beings. At present there are a variety of hypotheses concerning these mechanisms, no two of which are, of course, mutually exclusive. It is probably safe to say that the larger the number of hypotheses being considered, the less is known about the process in question. 8.6.1 Lack of Direct Genotoxicity

As is the case for most peroxisome proliferators [7, 89, 109, 125], present indications are that perfluorofatty acids are not directly genotoxic [13, 27, 40, 77, 153]. Thus, PFDA is non-mutagenic in the Ames test without activation [27, 40, 77], nor does it cause mutations in the hypoxanthine:guanine phosphoribosyltransferase (hprt) gene of the Chinese hamster ovary cell line or elicit sister chromatid exchange or unscheduled DNA synthesis [40]. However, in this latter study it was found that in the presence of an S9 fraction from rat liver (a post-

Effects on Rodents of Perfluorofatty Acids

235

mitochondrial fraction employed routinely to achieve xenobiotic metabolism in connection with the Ames test), PFDA does cause chromosomal aberrations. This is a rather peculiar finding, since rats appear not to metabolize PFDA (see Sect. 3). 8.6.2 Increased Oxidative Stress

Even though the increase in hepatic oxidative stress upon exposure to PFOA and PFDA is relatively minor (see Sect. 6.9), it is repeatedly suggested that this increase may be sufficient to promote hepatocarcinogenesis initiated by other xenobiotics [103, 177]. Massive accumulation of lipofuscin in the hepatocytes of rats exposed to peroxisome proliferators has been proposed to demonstrate than even a relatively modest increase in the level of oxidative stress over prolonged periods can have dramatic effects [177]. However, to date, no detailed mechanism by which oxidative stress could act as a promoter has been examined. Furthermore, mice which lack active acyl-CoA oxidase, the major source of increased hydrogen peroxide production in association with peroxisome proliferation, nonetheless develop liver tumors as a consequence of exposure to these substances [177]. 8.6.3 Altered Xenobiotic Metabolism

Since peroxisome proliferators, including PFOA and PFDA, up-regulate cytochrome P450 4A (see Sect. 6.7), it has been suggested that these compounds alter the metabolism of other xenobiotics via the cytochrome P450 system, catalyzed either by this isozyme or some other which is also up-regulated. Increased metabolism of xenobiotics to electrophilic reactive intermediates – including epoxides, free radicals, and carbonium ions – would, indeed, be expected to promote hepatocarcinogenesis. To date, no such enhanced formation of reactive metabolites as a consequence of exposure to PFOA or PFDA has, however, been demonstrated. 8.6.4 Stimulation of Hepatocyte Proliferation

It is well known that rapidly dividing cells are more susceptible to the mutagenic/genotoxic/carcinogenic effects of xenobiotics than are cells which divide more slowly [34, 133]. Thus, the dramatic increase in hepatocyte proliferation which results from exposure of rats and mice to PFOA or PFDA (see Sect. 6.4) may explain, at least in part, the ability of these compounds to promote the carcinogenicity of other xenobiotics. Of relevance in this connection is the observation by Takagi and coworkers [147] of a close relationship between the carcinogenicities of seven different peroxisome proliferators (not including perfluorofatty acids) in rats and the relative increases in liver weights after a oneweek exposure to carcinogenic doses.

236

J.W. DePierre

8.6.5 Inhibition of Hepatocyte Apoptosis

Apoptosis is a natural and active form of cell death or, more descriptively, cell suicide [19, 34, 102, 119, 132, 133, 149]. This process is involved in removing cells which have served their purpose (e.g., during morphogenesis of the fetus) or cells which are potentially dangerous to the organism (e.g., lymphocytes which produce antigens towards the organism’s own components, and mutated cells, which may progress to become cancerous and lead to the formation of life-threatening tumors). Thus, it has been postulated that peroxisome proliferators may promote hepatocarcinogenesis by preventing mutated hepatocytes from undergoing apoptosis, as they normally would. Indeed, certain peroxisome proliferators have been shown to inhibit apoptosis in rat and mouse liver, as well as in cell lines derived from these tissues [132]. However, under other conditions the opposite effect, i.e., enhancement of apoptosis by representatives of this class of compounds, has been reported [19, 119]. In attempt to test this hypothesis, we exposed the relatively well-differentiated human hepatoma HepG2 cell line to PFOA and looked for changes in apoptosis [116, 136].We found that at moderate concentrations PFOA induces apoptosis in HepG2 cells and at higher concentrations this compound leads to necrosis. The apoptosis induced in this manner seemed to involve typical pathways, i.e., an increase in oxidative stress (even though acyl-CoA oxidase was not up-regulated) and disruption of the mitochondrial membrane. Similar observations were later made with the rat hepatoma Fao cell line (unpublished observations from our laboratory, 2000). Thus, presently available information lends no definitive support to the hypothesis that perfluorofatty acids promote hepatocarcinogenesis by preventing apoptosis in mutated hepatocytes. 8.6.6 Immunotoxicity?

The role of immunological defenses in protecting against tumor cells has been debated intensely for decades now [176 and references therein]. On the one hand, tumor-specific transplantation antigens (or tumor rejection antigens) are well characterized. Furthermore, T-cell immunodeficient mice exposed to 3-methylcholanthrene develop tumors more rapidly and with a higher frequency that do normal mice exposed to this carcinogen. On the other hand, the incidence of common tumors in mice that lack lymphocytes is not much different than in animals with an intact immune system, which is also the case for humans deficient in T-cells. In addition, tumors are known to have various endogenous systems (e.g., low immunogenicity, internalization of antigens, tumor-induced immunosuppression, etc.) for avoiding immune surveillance. Most likely, the final resolution of this debate will prove to be complex, i.e., immune surveillance plays an important role in preventing the development of certain tumors, but not of others. Thus, if, as seems highly probable, exposure of rats and mice to perfluorofatty acids results in immunosuppression, this could

Effects on Rodents of Perfluorofatty Acids

237

be the mechanism by which these compounds promote the development of liver tumors. In addition, immunotoxic effects might increase the susceptibility of the exposed organism to infection by bacteria, viruses, fungi, and parasites.

9 Studies on Humans A number of reviews have claimed that since human cells are relatively resistant to the peroxisome-proliferating effects of perfluorofatty acids and other peroxisome proliferators, the risk posed by these compounds to human beings can be considered to be negligible [7, 89, 182]. However, such a conclusion assumes that the mechanism(s) underlying the toxic and genotoxic effects of perfluorofatty acids in rodents involves peroxisome proliferation. In light of the wide variety of other effects exerted by these substances and described above, such an assumption is, at present, not well-founded. To date, only a few studies concerning the effects of perfluorofatty acids on human beings have been conducted. In 1993 an epidemiological study of employees at a PFOA production plant concluded that the mortality rate from all causes, as well as from cardiovascular disease, were normal, but that the frequency of prostate cancer after 10 years of exposure was 3.3 times higher than the expected incidence [36]. However, only four cases of prostate cancer occurred among exposed workers, so this finding must certainly be interpreted with caution. Other epidemiological studies have failed to detect any effects on male workers exposed to PFOA – either with respect to hepatic toxicity (as reflected in plasma levels of cholecystokinin, hepatic enzymes, cholesterol, and lipoproteins) [37, 113], or circulating levels of reproductive hormones (estradiol or testosterone) [115]. Nor did occupational exposure to perfluorooctane sulfonate affect plasma levels of hepatic enzymes, cholesterol, or lipoproteins [114]. Two investigations dealing with the effect of PFOA and PFDA on human cell lines have appeared in the literature [25, 94]. In one of these, the expression and excretion of IgM antibodies by cell lines derived from human and murine B-lymphocytes were unaffected by PFDA, although significant increases in both the number of peroxisomes present and in acyl-CoA oxidase activity were observed [94]. Furthermore, these effects were enhanced by simultaneous addition of 9-cisretinoic acid to the culture medium. These cell lines were shown to express both PPARalpha and PPARgamma, and PFDA was shown to up-regulate the expression of the gamma form of this nuclear receptor.

10 Concluding Remarks 10.1 Valuable Experimental Tools

Clearly, PFOA and PFDA will certainly provide even more insight into a range of biochemical problems and will continue to be valuable experimental tools, e.g., in the mechanisms underlying peroxisome biogenesis and degradation, control

238

J.W. DePierre

of hepatocyte proliferation, global lipid homeostasis, modulation of the immune system, and promotion of liver and testicular tumors. It is, of course, of considerable interest to identify endogenous and naturally occurring dietary peroxisome proliferators, which are presumably involved in the physiological regulation of these processes. 10.2 Possible Clinical Applications

Possibly, these perfluorofatty acids may also be valuable in a number of clinical applications. For instance, their immunotoxicity might be turned to use in treating certain conditions, e.g., rejection of transplanted organs or autoimmune diseases, i.e., PFOA and PFDA may be potential useful as immunosuppressive drugs. In addition, the loss of body fat in response to exposure to, in particular, PFOA (and other peroxisome proliferators) may make this a useful anti-obesity drug. However, it is my strong belief that, rather than being useful in themselves as drugs, the basic biological information obtained through the use of these compounds as experimental tools (see above) will aid in the design of clinically useful immunosuppressive and anti-obesity agents and/or dietary strategies. 10.3 Hazard to Human Health?

An important question which remains to be explored is whether PFOA, PFDA, and other perfluorofatty acids in the occupational and general environments represent a hazard to human health. The relative chemical and metabolic stabilities of these compounds may result in their presence in increasing quantities in the general environment. At present, peroxisome proliferators are generally considered not to increase the risk for liver tumors in humans, since the peroxisomes in our liver respond not at all or only slightly to these compounds. However, this conclusion assumes that peroxisome proliferation is an integral part of the genotoxic effect of peroxisome proliferators and this assumption is, I feel, rather loose in light of the numerous other effects exerted by peroxisome proliferators. Furthermore, even if, as is to be hoped, peroxisome proliferators are not genotoxic in humans, many of their other effects could potentially have a negative influence on our health. Clearly, much research remains to be done before this question can receive its definitive answer. Acknowledgements. The experiments performed in our laboratory and described here were supported by grants from the Swedish Natural Science Research Council, the Environmental Fund of the Swedish Association of Civil Engineers, and the Knut and Alice Wallenberg Foundation (Stockholm).

Effects on Rodents of Perfluorofatty Acids

239

11 References 1. Adinehzadeh M, Reo NV (1998) Effects of peroxisome proliferators on rat liver phospholipids: sphingomyelin degradation may be involved in hepatotoxic mechanism of perfluorodecanoic acid. Chem Res Toxicol 11:428–440 2. Adinehzadeh M, Reo NV, Jarnot BM, Taylor CA, Mattie DR (1999) Dose-response hapatotoxicity of the peroxisome proliferator, perfluorodecanoic acid and the relationship to phospholipid metabolism in rats. Toxicology 134:179–195 3. Alexson SHE, Cannon B (1984) A direct comparison between perixosomal and mitochondrial preferences for fatty acyl b-oxidation predicts channeling of medium-chain and very-long-chain unsaturated fatty acids to peroxisomes. Biochim Biophys Acta 796:1–10 4. Arand M, Coughtrie MW, Burchell B, Oesch F, Robertson LW (1991) Selective induction of bilirubin UDP-glucuronosyl-transferase by perfluorodecanoic acid. Chem Biol Interact 77:97–105 5. Arnaiz SL, Travacio M, Llesuy S, Boveris A (1995) Hydrogen peroxide metabolism during peroxisome proliferation by fenofibrate. Biochim Biophys Acta 1272:175–180 6. Bannasch P, Klimek F, Mayer D (1997) Early bioenergetic changes in hepatocarcinogenesis: preneoplastic phenotypes mimic responses to insulin and thyroid hormone. J Bioenerg Biomembr 29:303–313 7. Bentley P, Calder I, Elcombe C, Grasso P, Stringer D, Wiegand H-J (1993) Hepatic peroxisome proliferation in rodents and its significance for humans. Food Chem Toxic 31:857–907 8. Berge RK, Madsen L, Vaagenes H (1999) Hypolipidemic 3-thia fatty acids. Fatty acid oxidation and ketogenesis in rat liver under proliferation of mitochondria and peroxisomes. Adv Exp Biol Med 466:125–132 9. Biegel LB, Liu RC, Hurtt ME, Cook JC (1995) Effects of ammonium perfluorooctanoate on Leydig cell function: in vitro, in vivo, and ex vivo studies. Toxicol Appl Pharmacol 134:18–25 10. Bookstaff RC, Moore RW, Ingall GB, Peterson RE (1990) Androgenic deficiency in male rats treated with perfluorodecanoic acid. Toxicol Appl Pharmacol 104:322–333 11. Borges T, Glauert HP, Chen LC, Chow CK, Robertson JW (1990) Effect of the peroxisome proliferator perfluorodecanoic acid on growth and lipid metabolism in Sprague Dawley rats fed three dietary levels of selenium. Arch Toxicol 64:26–30 12. Borges T, Glauert HP, Robertson LW (1993) Perfluorodecanoic acid noncompetitively inhibits the peroxisomal enzymes enoyl-CoA hydratase and 3-hydroxyacyl-CoA dehydrogenase. Toxicol Appl Pharmacol 118:8–15 13. Borges T, Peterson RE, Pitot HC, Robertson LW, Glauert HP (1999) Effect of the peroxisome proliferator perfluorodecanoic acid on the promotion of two-stage hepatocarcinogenesis in rats. Cancer Lett 72:111–120 14. Borges T, Robertson LW, Peterson RE, Glauert HP (1992) Dose-related effects of perfluorodecanoic acid on growth, feed intake and hepatic peroxisomal beta-oxidation.Arch Toxicol 66:18–22 15. Boujrad N, Vidic B, Gazouli M, Culty M, Papadopoulos V (2000) The peroxisome proliferator perfluorodecanoic acid inhibits the peripheral-type benzodiazepine receptor (PBR) expression and hormone-stimulated mitochondrial cholesterol transport and steroid formation in Leydig cells. Endocrinology 141:3137–3148 16. Brewster DW, Birnbaum LS (1989) The biochemical toxicity of perfluorodecanoic acid in the mouse is different from that of 2,3,7,8-tetrachlorodibenzo-p-dioxin. Toxicol Appl Pharmacol 99:544–554 17. Cai Y, Appelkvist EL, DePierre JW (1995) Hepatic oxidative stress and related defenses during treatment of mice with acetylsalicylic acid and other peroxisome proliferators. J Biochem Toxicol 10:87–94

240

J.W. DePierre

18. Cai Y, Nelson BD, Li R, Luciakova K, DePierre JW (1996) Thyromimetic action of the peroxisome proliferators clofibrate, perfluorooctanoic acid and acetylsalicylic acid includes changes in mRNA levels for certain genes involved in mitochondrial biogenesis. Arch Biochem Biophys 325:107–112 19. Canuto R, Muzio G, Bonelli G, Maggiora M,Autelli R, Barbiero G, Costelli P, Brossa O, Baccino FM (1998) Peroxisome proliferators induce apoptosis in hepatoma cells. Cancer Detect Prevent 22:357–366 20. Cattani L, Ferri A (1994) The function of NADPH bound to catalase. J Biol Chem 70:75–82 21. Chen LC, Borges T, Glauert HP, Knight SA, Sunde RA, Schramm H, Oesch F, Chow CK, Robertson LW (1990) Modulation of selenium-dependent glutathione peroxidase by perfluorodecanoic acid in rats: effect of dietary selenium. J Nutr 120:298–304 22. Chen H, Huang CY, Wilson MW, Lay LT, Robertson LW, Chow CK, Glauert HP (1994) Effect of the peroxisome proliferators ciprofibrate and perfluorodecanoic acid on hepatic cell proliferation and toxicity in Sprague-Dawley rats. Carcinogenesis 15:2847–2850 23. Chinetti G, Fruchart JC, Staels R (2000) Peroxisome proliferator-activated receptors (PPARs): nuclear receptors at the crossroads between lipid metabolism and inflammation. Inflamm Res 49:497–505 24. Chinje E, Kentish P, Jarnot B, George M, Gibson G (1994) Induction of the CYP4A subfamily by perfluorodecanoic acid: the rat and the guinea pig as susceptible and non-susceptible species. Toxicol Lett 71:69–75 25. Cimini A, Cristiano L, Bernardo A, Farioli-Vecchioli S, Stefanini S, Ceru MP (2000) Presence and inducibility of peroxisomes in a human glioblastoma cell line. Biochim Biophys Acta 1474:397–409 26. Conway JG, Popp JA (1995) Effect of the hepatocarcinogenic peroxisome proliferator Wy14.643 in vivo. No increase in ethane exhalation or hepatic conjugated dienes. Toxicol Appl Pharmacol 135:229–236 27. Cook JC, Murray SM, Frame SR, Hurtt ME (1992) Induction of Leydig cell adenomas by ammonium perfluorooctanoate: a possible endocrine-related mechanism. Toxicol Appl Pharmacol 113:209–217 28. Corton JC, Anderson SP, Stauber A (2000) Central role of peroxisome proliferator-activated receptors in the actions of peroxisome proliferators. Annu Rev Pharmacol Toxicol 40:491–518 29. Davis JW, Vanden Heuvel JP, Peterson RE (1991) Effects of perfluorodecanoic acid on de novo fatty acid and cholesterol synthesis in the rat. Lipids 26:857–859 30. Derbel M, Hosokawa M, Satoh T (1996) Differences in the induction of carboxylesterase RL4 in rat liver microsomes by various perfluorinated fatty acids, metabolically inert derivatives of fatty acids. Biol Pharm Bull 19:765–767 31. Diaz MJ, Chinje E, Kentish P, Jarnot B, George M, Gibson G (1994) Induction of cytochrome P4504A by the peroxisome proliferator perfluoro-n-octanoic acid. Toxicology 86:109–122 32. Dirven HA, van den Broek PH, Peters JG, Noordhoek J, Jongeneelen FJ (1992) Microsomal lauric acid hydroxylase activities after treatment of rats with three classical cytochrome P450 inducers and peroxisome proliferating compounds. Biochem Pharmacol 43:2621–2629 33. Fita I, Rossmann MG (1985) The active center of catalase. J Molec Biol 185:21–37 34. Foster JR (2000) Cell death and cell proliferation in the control of normal and neoplastic tissue growth. Toxicol Pathol 28:441–446 35. George ME,Andersen ME (1986) Toxic effects of nonadecafluoro-n-decanoic acid in rats. Toxicol Appl Pharmacol 85:169–180 36. Gilliland FD, Mandel JS (1993) Mortality among employees of a perfluorooctanoic acid production plant. J Occup Med 35:950–954 37. Gilliland FD, Mandel JS (1996) Serum perfluorooctanoic acid and hepatic enzymes, lipoproteins, and cholesterol: a study of occupationally exposed men. Am J Ind Med 29:560–568

Effects on Rodents of Perfluorofatty Acids

241

38. Glauert HP, Hong JT, Leung LK,Wilson MW (1996) Role of eicosanoid metabolism in carcinogenesis by peroxisome proliferators. Ann N Y Acad Sci 804:719–721 39. Glauert HP, Srinivasan S, Tatum VL, Chen LC, Saxon DM, Lay LT, Borges T, Baker M, Chen LH, Robertson LW (1992) Effects of the peroxisome proliferators ciprofibrate and perfluorodecanoic acid on hepatic cellular antioxidants and lipid peroxidation in rats. Biochem Pharmacol 43:1353–1359 40. Godin CS, Myhr BC, Lawlor TE, Young RR, Murli H, Cifone MA (1992) Assessment of the potential genotoxicity of perfluorodecanoic acid and chlorotrifluoroethylene trimer and tetramer acids. Fundam Appl Toxicol 18:557–569 41. Goecke CM, Jarnot BM, Reo NV (1992) A comparative toxicological investigation of perfluorocarboxylic acids in rats by 19F NMR spectroscopy. Chem Res Toxicol 5: 512–519 42. Goecke CM, Jarnot BM, Reo NV (1994) Effects of the peroxisome proliferator perfluoron-decanoic acid on hepatic gluconeogenesis and glycogenesis: a 13C NMR investigation. Chem Res Toxicol 7:15–22 43. Goecke-Flora CM, Reo NV (1996) Influence of carbon chain length on the hepatic effects of perfluorinated fatty acids. A 19F- and 31P-NMR investigation. Chem Res Toxicol 9:689–695 44. Goecke-Flora CM, Wyman JF, Jarnot RM, Reo NV (1995) Effect of the peroxisome proliferator perfluoro-n-decanoic acid on glucose transport in the isolated perfused rat liver. Chem Res Toxicol 8:77–81 45. Goel SK, Lalwani ND, Reddy JK (1986) Peroxisome proliferation and lipid peroxidation in rat liver. Cancer Res 46:1324–1330 46. Gonzalez FJ (1997) Recent update on the PPAR alpha-null mouse. Biochimie 79:139–144 47. Green S (1992) Peroxisome proliferators: a model for receptor-mediated carcinogenesis. Cancer Surveys 14:221–232 48. Green S,Wahli W (1994) Peroxisome proliferator-activated receptors: finding the orphan a home. Mol Cell Endocrinol 100:149–153 49. Gribble GW (2001) Naturally occurring organofluorines. Present volume, chap 5 50. Gutshall DM, Pilcher GD, Langley AE (1988) Effect of thyroxine supplementation on the response to perfluoro-n-decanoic acid (PFDA) in rats. J Toxicol Environ Health 24:491–498 51. Gutshall DM, Pilcher GD, Langley AE (1989) Mechanism of the serum thyroid hormone lowering effect of perfluoro-n-decanoic acid (PFDA) in rats. J Toxicol Environ Health 28:53–65 52. Handler JA, Seed CB, Bradford BU, Thurman RG (1992) Induction of peroxisomes by treatment with perfluorooctanoate does not increase rates of H2O2 production in intact liver. Toxicol Lett 60:61–68 53. Hanhijarvi H, Ylinen M, Kojo A, Kosma VM (1987) Elimination and toxicity of perfluorooctanoic acid during subchronic administration in the Wistar rat. Pharmacol Toxicol 61:66–68 54. Hansen KJ, Clemen LA, Ellefson ME, Johnson HO (2001) Compound-specific quantitative characterization of organic fluorochemicals in biological matrices. Environ Sci Technol 35:766–770 55. Harris MW, Birnbaum LS (1989) Developmental toxicity of perfluorodecanoic acid in C57BL/6N mice. Fundam Appl Toxicol 12:442–448 56. Harris MW, Uraih LC, Birnbaum LS (1989) Acute toxicity of perfluorodecanoic acid in C57BL/6 mice differs from 2,3,7,8-tetrachlorodibenzo-p-dioxin. Fundam Appl Toxicol 13:723–736 57. Harrison EH, Lane JS, Luking S,Van Rafelghem MJ,Andersen ME (1988) Perfluoro-n-decanoic acid: induction of peroxisomal beta-oxidation by a fatty acid with dioxin-like toxicity. Lipids 23:115–119 58. Haughom B, Spydevold O (1992) The mechanism underlying the hypolipemic effect of perfluorooctanoic acid (PFOA), perfluorooctane sulphonic acid (PFOSA) and clofibric acid. Biochim Biophys Acta 1128:65–72

242

J.W. DePierre

59. Hawkins JM, Jones WE, Bonner FW, Gibson GG (1987) The effect of peroxisome proliferators on microsomal, peroxisomal, and mitochondrial activities in the liver and kidney. Drug Metab Rev 18:441–515 60. Hemsley A, Pegg M, Crane D, Masters C (1988) On the compartmentalization of catalase, fatty acyl-CoA oxidase and urate oxidase in mammalian livers, and the influence of clofibrate on this microlocalization. Mol Cell Biochem 83:187–194 61. Hertz R, Kalderon B, Bar-Tana J (1993) Thyromimetic effect of peroxisome proliferators. Biochimie 75:257–261 62. Holmes RS, Masters CJ (1972) Species specific features of the distribution and multiplicity of mammalian liver catalase. Arch Biochem Biophys 148:217–223 63. Hosokawa M, Satoh T (1993) Differences in the induction of carboxylesterase isozymes in rat liver microsomes by perfluorinated fatty acids. Xenobiotica 23 : 1125–1133 64. Hryb DJ, Hogg JF (1979) Chain length specificities of peroxisomal and mitochondrial b-oxidation in rat liver. Biochem Biophys Res Commun 87:1200–1206 65. Huang CY, Wilson MW, Lay LT, Chow CK, Robertson LW, Glauert HP (1994) Increased 8hydroxydeoxyguanosine in hepatic DNA of rats treated with the peroxisome proliferators ciprofibrate and perfluorodecanoic acid. Cancer Lett 87:223–228 66. Ikeda T, Aiba K, Fukuda K, Tanaka M (1985) The induction of peroxisome proliferation in rat liver by perfluorinated fatty acids, metabolically inert derivatives of fatty acids. J Biochem (Tokyo) 98:475–482 67. Inestrosa NC, Bronfman M, Leighton F (1979) Detection of peroxisomal fatty acyl-CoA oxidase activity. Biochem J 182:779–788 68. Just WW, Gorgas K, Hartl FU, Heinemann P, Salzer M, Schimassek H (1989) Biochemical effects and zonal heterogeneity of peroxisome proliferation induced by perfluorocarboxylic acids in rat liver. Hepatology 9:570–581 69. Kawashima Y, Kobayashi H, Miura H, Kozuka H (1995) Characterization of hepatic responses of rat to administration of perfluorooctanoic and perfluorodecanoic acids at low levels. Toxicology 99:169–178 70. Kannan K, Koistinen J, Beckman K, Evans T, Gorzelany JF, Hansen KL, Jones PD, Helle E, Nyman M, Geisy JP (2001) Accumulation of perfluorooctane sulfonate in marine mammals. Environ Sci Technol 35:1593–1598 71. Kawashima Y, Uy-Yu N, Kozuka H (1989) Sex-related difference in the inductions by perfluoro-octanoic acid of peroxisomal beta-oxidation, microsomal 1-acylglycerophosphocholine acyltransferase and cytosolic long-chain acyl-CoA hydrolase in rat liver. Biochem J 261:595–600 72. Kawashima Y, N Uy-Yu N, Kozuka H (1989) Sex-related differences in the enhancing effects of perfluoro-octanoic acid on stearoyl-CoA desaturase and its influence on the acyl composition of phospholipid in rat liver. Comparison with clofibric acid and tiadenol. Biochem J 263:897–904 73. Keller JM, Collet P, Bianchi A, Huin C, Bouillaud-Kremarik P, Becuwe P, Schohn H, Domenjoud L, Dauca M (2000) Implications of peroxisome proliferator-activated receptors (PPARs) in development, cell life status and disease. Int J Dev Biol 44:429–442 74. Kelling CK,Van Rafelghem MJ, Drake RL, Menahan LA, Peterson RE (1986) Regulation of hepatic malic enzyme by perfluorodecanoic acid. J Biochem Toxicol 1:23–37 75. Kelling CK, Van Rafelghem MJ, Menahan LA, Peterson RE (1987) Effects of perfluorodecanoic acid on hepatic indices of thyroid status in the rat. Biochem Pharmacol 36:1337–1344 76. Kennedy GL (1987) Increase in mouse liver weight following feeding of ammonium perfluorooctanoate and related fluorochemicals. Toxicol Lett 39:295–300 77. Kim SC, Hong JT, Jang SJ, Kang WS, Yoo HS, Yun YP (1998) Formation of 8-oxodeoxyguanosine in liver DNA and hepatic injury by peroxisome proliferator clofibrate and perfluorodecanoic acid in rats. J Toxicol Sci 23:113–119 78. Kimura T, Owens IS (1987) Mouse UDP glucuronosyltransferase. cDNA and complete amino acid sequence and regulation. Eur J Biochem 168:515–521

Effects on Rodents of Perfluorofatty Acids

243

79. Kinney LA, Chromey NC, Kennedy GL (1989) Acute inhalation toxicity of ammonium perfluorononanoate. Food Chem Toxicol 27:465–468 80. Knights KM (1998) Role of fatty acid:coenzyme A ligases in the metabolism of xenobiotic carboxylic acids. Clin Exp Pharmacol Physiol 25:776–782 81. Kojo A, Hanhijarvi H, Ylinen M, Kosma VM (1986) Toxicity and kinetics of perfluorooctanoic acid in the Wistar rat. Arch Toxicol Suppl 9:465–468 82. Kozuka H, Watanabe T, Horie S, Yamada J, Suga T, Ikeda T (1991) Characteristics of peroxisome proliferation: co-induction of peroxisomal fatty acid oxidation-related enzymes with microsomal laurate hydroxylase. Chem Pharm Bull (Tokyo) 39 :1267–1271 83. Kozuka H, Yamada J, Horie S, Watanabe T, Suga T, Ikeda T (1991) Characteristics of induction of peroxisomal fatty acid oxidation-related enzymes in rat liver by drugs. Relationships between structure and inducing activity. Biochem Pharmacol 41 : 617–623 84. Kudo N, Bandai N, Kawashima Y (1998) Determination of perfluorocarboxylic acids by gas-liquid chromatography in rat tissues. Toxicol Lett 99:183–190 85. Kudo N, Bandai N, Suzuki E, Katakura M, Kawashima Y (2000) Induction by perfluorinated fatty acids with different carbon chain length of peroxisomal beta-oxidation in the liver of rats. Chem Biol Interact 124:119–132 86. Kudo N, Kawashima Y (1997) Fish-oil feeding prevents perfluorooctanoic acid-induced fatty liver in mice. Toxicol Appl Pharmacol 145:285–293 87. Kudo N, Mizuguchi H, Yamamoto A, Kawashima Y (1999) Alterations by perfluorooctanoic acid of glycerolipid metabolism in rat liver. Chem Biol Interact 118:69–83 88. Kuslikis BI,Vanden Heuvel JP, Peterson RE (1992) Lack of evidence for perfluorodecanoylor perfluorooctanoyl-coenzyme A formation in male and female rats. J Biochem Toxicol 7:25–29 89. Lake BG (1995) Peroxisome proliferation: current mechanisms relating to non-genotoxic carcinogenesis. Toxicol Lett 82/83:673–681 90. Langley AE (1990) Effects of perfluoro-n-decanoic acid on the respiratory activity of isolated rat liver mitochondria. J Toxicol Environ Health 29:329–336 91. Langley AE, Pilcher GD (1985) Thyroid, bradycardic and hypothermic effects of perfluoro-n-decanoic acid in rats. J Toxicol Environ Health 15:485–491 92. Latruffe N, Malki MC, Nicolas-Frances V, Clemencet MC, Jannin B, Berlot JP (2000) Regulation of the peroxisomal beta-oxidation-dependent pathway by peroxisome proliferator-activated receptor alpha and kinases. Biochem Pharmacol 60 :1027–1032 93. Lepers A, Cacan R, Verbert A (1990) Permeabilized cells as a way of gaining access to intracellular organelles: an approach to glycosylation. Biochimie 72:1–5 94. Levitt D, Liss A (1986) Toxicity of perfluorinated fatty acids for human and murine B cell lines. Toxicol Appl Pharmacol 86:1–11 95. Lock EA, Mitchell AM, Elcombe CR (1989) Biochemical mechanisms of induction of hepatic peroxisome proliferation. Annu Rev Pharmacol Toxicol 29:145–163 96. Lundgren B, Andersson K, DePierre JW (1992) Effects of dietary treatment with 11 dicarboxylic acids, diethylcarboxylic esters and fatty acids on peroxisomal fatty acid b-oxidation, epoxide hydrolases and lauric acid – ω -hydroxylation in mouse liver. Biochem Pharmacol 43:785–792 97. Lundgren B, Bergstrand A, Karlsson K, DePierre JW (1990) Effects of dietary treatment with clofibrate, nafenopin or WY-14.643 on mitochondria and DNA in mouse liver. Biochim Biophys Acta 1035:132–138 98. Lundgren B, Karlsson K, Messing-Eriksson A, DePierre JW (1990) Characterization of the cellular compartmentalizations of catalase, soluble epoxide hydrolase and carnitine acetyltransferase by differential centrifugation of the livers of mice exposed to clofibrate. Life Sci Adv 9:339–345 99. Lundgren B, Meijer J, Birberg W, Pilotti Å, DePierre JW (1988) Induction of cytosolic and microsomal epoxide hydrolases in mouse liver by peroxisome proliferators, with special emphasis on structural analogues of 2-ethylhexanoic acid. Chem-Biol Interact 68:219–240

244

J.W. DePierre

100. Lundgren B, Meijer J, DePierre JW (1987) Examination of the structural requirements for proliferation of peroxisomes and mitochondria in mouse liver by hypolipidemic agents, with special emphasis on structural analogues of 2-ethylhexanoic acid. Eur J Biochem 163:423–431 101. Magdalou J, Fournel-Gigleux S, Pritchard M, Siest G (1993) Peroxisome proliferators as inducers and substrates of UDP-glucuronosyltransferases. Biol Cell 77:13–16 102. Majno G, Joris I (1995) Apoptosis, oncosis, and necrosis. An overview of cell death. Am J Pathol 146:3–15 103. Masters C, Crane D (1998) On the role of the peroxisome in cell differentiation and carcinogenesis. Mol Cell Biochem 187:85–97 104. Messing-Eriksson A, Lundgren B, Andersson K, DePierre JW (1992) Is the cytosolic catalase induced by peroxisome proliferators in mouse liver on its way to the peroxisomes? FEBS Lett 308:211–214 105. Messing-Eriksson A, Zetterqvist M-A, Lundgren B, Andersson K, Beije B, DePierre JW (1991) Studies on the intracellular distributions of soluble epoxide hydrolase and of catalase by digitonin-permeabilization of hepatocytes isolated from control and clofibratetreated mice. Eur J Biochem 198:471–476 106. Miller RT, Scappino LA, Long SM, Corton JC (2001) Role of thyroid hormones in hepatic effects of peroxisome proliferators. Toxicol Pathol 29:149–155 107. Nakagawa M, Ishihara N, Shimokawa T, Kojima S (1987) Effects of clofibrate on lipid peroxidation in rats treated with aspirin and 4-pentenoic acid. J Biochem (Tokyo) 101: 81–88 108. Nelson DL, Frazier DE, Ericson JE, Tarr MJ, Mathes LE (1992) The effects of perfluorodecanoic acid (PFDA) on humoral, cellular, and innate immunity in Fischer 344 rats. Immunopharmacol Immunotoxicol 14:925–938 109. Nilsson R, Beije B, Préat V, Erixon K, Ramel C (1991) On the mechanism of the hepatocarcinogenecity of peroxisome proliferators. Chem Biol Interact 78:235–250 110. Nohl H, Gille L, Staniek K (1998) The biochemical, pathophysiological and medical aspects of ubiquinone function. Ann N Y Acad Sci 20:394–409 111. Ohya T, Kudo N, Suzuki E, Kawashima Y (1998) Determination of perfluorinated carboxylic acids in biological samples by high-performance liquid chromatography. J Chromatogr B Biomed Sci Appl 720:1–7 112. Olson CT, Andersen ME (1983) The acute toxicity of perfluorooctanoic and perfluorodecanoic acids in male rats and effects on tissue fatty acids. Toxicol Appl Pharmacol 70:362–372 113. Olsen GW, Burris JM, Burlew MM, Mandel JH (2000) Plasma cholecystokinin and hepatic enzymes, cholesterol and lipoproteins in ammonium perfluorooctanoate production workers. Drug Chem Toxicol 23:603–620 114. Olsen GW, Burris JM, Mandel JH, Zobel LR (1999) Serum perfluorooctane sulfonate and hepatic and lipid clinical chemistry tests in fluorochemical production employees. J Occup Environ Med 41:799–806 115. Olsen GW, Gilliland FD, Burlew MM, Burris JM, Mandel JS, Mandel JH (1998) An epidemiologic investigation of reproductive hormones in men with occupational exposure to perfluorooctanoic acid. J Occup Environ Med 40:614–622 116. Panaretakis T, Shabalina IG, Grandér D, Shoshan MC, DePierre JW (2001) Reactive oxygen species and mitochondria mediate the induction of apoptosis in human hepatoma HepG2 cells by the rodent peroxisome proliferator and hepatocarcinogen perfluorooctanoic acid. Toxicol Appl Pharmacol (in press) 117. Permadi H, Lundgren B,Andersson K, DePierre JW (1992) Effects of perfluoro fatty acids on xenobiotic-metabolizing enzymes, enzymes which detoxify reactive forms of oxygen and lipid peroxidation in mouse liver. Biochem Pharmacol 44:1183–1191 118. Permadi H, Lundgren B, Andersson K, Sundberg C, DePierre JW (1993) Effects of perfluoro fatty acids on peroxisome proliferation and mitochondrial size in mouse liver: dose and time factors and effect of chain length. Xenobiotica 23:761–770

Effects on Rodents of Perfluorofatty Acids

245

119. Perrone CE, Shao L,Williams GM (1998) Effect of rodent hepatocarcinogenic peroxisome proliferators on fatty acyl-CoA oxidase, DNA synthesis, and apoptosis in cultured human and rat hepatocytes. Toxicol Appl Pharmacol 150:277–286 120. Pilcher D, Gutshall DM, Langley AE (1987) The effects of perfluoro-n-decanoic acid on rat heart beta-receptors, adenylate cyclase, and fatty acid composition. Toxicol Appl Pharmacol 90:198–205 121. Pilcher GD, Langley AE (1986) The effects of perfluoro-n-decanoic acid in the rat heart. Toxicol Appl Pharmacol 85:389–397 122. Pobezhimova TP, Voinikov VK (2000) Biochemical and physiological aspects of ubiquinone function. Membr Cell Biol 13:595–602 123. Powers RH, Aust SD (1986) The effects of nonadecafluoro-n-decanoic acid on serum retinol and hepatic retinyl palmitate hydrolase activity in male Sprague-Dawley rats. J Biochem Toxicol 1:27–42 124. Raucy JL, Kramer JC, Lasker JM (1993) Bioactivation of halogenated hydrocarbons by cytochrome P4502E1. Crit Rev Toxicol 23:1–20 125. Reddy JK (1990) Carcinogenicity of peroxisome proliferators: evaluation and mechanisms. Biochem Soc Trans 18:92–94 126. Reddy JK, Mannaerts GP (1994) Peroxisomal lipid metabolism. Annu Rev Nutr 14:343–370 127. Renner R (2001) Growing concern over perfluorinated chemicals. Feature Article. Environ Sci Technol 35:154A–160A 128. Reo NV,Adinehzadeh M (2000) NMR spectroscopic analyses of liver phosphatidylcholine and phosphatidylethanolamine biosynthesis in rats exposed to peroxisome proliferatorsA class of nongenotoxic hepatocarcinogens. Toxicol Appl Pharmacol 164:113–126 129. Reo NV, Goecke CM, Narayanan L, Jarnot BM (1994) Effects of perfluoro-n-octanoic acid, perfluoro-n-decanoic acid, and clofibrate on hepatic phosphorus metabolism in rats and guinea pigs in vivo. Toxicol Appl Pharmacol 124:165–173 130. Reo NV, Narayanan L, Kling KB, Adinehzadeh M (1996) Perfluorodecanoic acid, a peroxisome proliferator, activates phospholipase C, inhibits CTP:phosphocholine cytidylyltransferase, and elevates diacylglycerol in rat liver. Toxicol Lett 86:1–11 131. Reubsaet FAG,Veerkamp JH, Bukkens SGF, Trijbels JMF, Monnens LAH (1988) Acyl-CoA oxidase activity and peroxisomal fatty acid oxidation in rat tissues. Biochim Biophys Acta 958:434–442 132. Roberts RA (1996) Non-genotoxic hepatocarcinogenesis: suppression of apoptosis by peroxisome proliferators. Ann NY Acad Sci 804:588–611 133. Roberts RA, James NH, Hasmall SC, Holden PR, Lambe K, Macdonald N, West D, Woodyatt NJ,Whitcombe D (2000) Apoptosis and proliferation in nongenotoxic carcinogenesis: species differences and role of PPARalpha. Toxicol Lett 112–113:49–57 134. Satoh A, Sasago S, Takahashi S, Kato N (1993) Regulation of xanthine dehydrogenase in rat liver in response to peroxisome proliferators. Biochem Biophys Res Commun 195:751–757 135. Schramm H, Friedberg T, Robertson LW, Oesch F, Kissel W (1989) Perfluorodecanoic acid decreases the enzyme activity and the amount of glutathione S-transferases proteins and mRNAs in vivo. Chem Biol Interact 70:127–143 136. Shabalina IG, Panaretakis T, Bergstrand A, DePierre JW (1999) Effects of the rodent peroxisome proliferator and hepatocarcinogen perfluorooctanoic acid on apoptosis in human hepatoma HepG2 cells. Carcinogenesis 20:2237–2246 137. Sichak SP, Dounce AL (1986) Analysis of the peroxidatic mode of action of catalase. Arch Biochem Biophys 249:286–295 138. Singer SS,Andersen ME, George ME (1990) Perfluoro-n-decanoic acid effects on enzymes of fatty acid metabolism. Toxicol Lett 54:39–46 139. Sohlenius A-K, Andersson K, DePierre JW (1992) The effects of perfluoro-octanoic acid on hepatic peroxisome proliferation and related parameters show no sex-related differences in mice. Biochem J 285:779–783

246

J.W. DePierre

140. Sohlenius A-K, Andersson K, Olsson J, DePierre JW (1995) Peroxisome proliferation and associated effects caused by perfluorooctanoic acid in vitamin A-deficient mice. ChemBiol Interact 98:45–50 141. Sohlenius A-K, Lundgren B, DePierre JW (1992) Perfluorooctanoic acid has persistent effects on peroxisome proliferation and related parameters in mouse liver. J Biochem Toxicol 7:205–212 142. Sohlenius A-K, Messing Eriksson A, Högström C, Kimland M, DePierre JW (1993) Perfluorooctane sulfonic acid is a potent inducer of peroxisomal fatty acid b-oxidation and other activities known to be affected by peroxisome proliferators in mouse liver. Pharmacol Toxicol 72:90–93 143. Sohlenius A-K, Reinfeldt M, Bäckström K, Bergstrand A, DePierre JW (1996) Hepatic peroxisome proliferation in vitamin A-deficient mice without a simultaneous increase in peroxisomal acyl-CoA oxidase activity. Biochem Pharmacol 51:821–827 144. Sterchele PF, Sun H, Peterson RE, Vanden Heuvel JP (1996) Regulation of peroxisome proliferator-activated receptor-alpha mRNA in rat liver. Arch Biochem Biophys 326: 281–289 145. Sterchele PF, Vanden Heuvel JP, Davis JW, Shrago E, Knudsen J, Peterson RE (1994) Induction of hepatic acyl-CoA-binding protein and liver fatty acid-binding protein by perfluorodecanoic acid in rats. Lack of correlation with hepatic long-chain acyl-CoA levels. Biochem Pharmacol 48:955–966 146. Takagi A, Sai K, Umemura T, Hasegawa R, Kurokawa Y (1991) Short-term exposure to the peroxisome proliferators, perfluorooctanoic acid and perfluorodecanoic acid, causes significant increase of 8-hydroxydeoxyguanosine in liver DNA of rats. Cancer Lett 57:55–60 147. Takagi A, Sai K, Umemura T, Hasegawa R, Kurokawa Y (1992) Hepatomegaly is an early biomarker for hepatocarcinogenesis induced by peroxisome proliferators. J Environ Pathol Toxicol Oncol 11:145–149 148. Totland GK, Madsen L, Klementsen B,Vaagenes H, Kryvi H, Froyland L, Hexeberg S, Berge RK (2000) Proliferation of mitochondria and gene expression of carnitine palmitoyltransferase and fatty acyl-CoA oxidase in rat skeletal muscle, heart and liver by hypolipidemic fatty acids. Biol Cell 92:317–329 149. Trump BF, Berezesky IK, Chang SH, Phelps PC (1997) The pathways of cell death: oncosis, apoptosis, and necrosis. Toxicol Pathol 25:82–88 150. Upham BL, Deocampo ND, Wurl B, Trosko JE (1998) Inhibition of gap junctional intercellular communication by perfluorinated fatty acids is dependent on the chain length of the fluorinated tail. Int J Cancer 78:491–495 151. Uy-Yu N, Kawashima Y, Kozuka H (1990) Comparative studies on sex-related difference in biochemical responses of livers to perfluorooctanoic acid between rats and mice. Biochem Pharmacol 39:1492–1496 152. Vamecq J, Draye J-P (1987) Interactions between the w- and b-oxidation of fatty acids. J Biochem (Tokyo) 102:225–234 153. Vanden Heuvel JP (1996) Perfluorodecanoic acid as a useful pharmacologic tool for the study of peroxisome proliferation. Gen Pharmacol 27:1123–1129 154. Vanden Heuvel JP, Davis JW, Sommers R, Peterson RE (1992) Renal excretion of perfluorooctanoic acid in male rats: inhibitory effect of testosterone. J Biochem Toxicol 7:31–36 155. Vanden Heuvel JP, Kuslikis BI, Peterson RE (1992) Covalent binding of perfluorinated fatty acids to proteins in the plasma, liver and testes of rats. Chem Biol Interact 82:317–328 156. Vanden Heuvel JP, Kuslikis BI, Shrago E, Peterson RE (1991) Inhibition of long-chain acylCoA synthetase by the peroxisome proliferator perfluorodecanoic acid in rat hepatocytes. Biochem Pharmacol 42:295–302 157. Vanden Heuvel JP, Kuslikis BI,Van Rafelghem MJ, Peterson RE (1991) Disposition of perfluorodecanoic acid in male and female rats. Toxicol Appl Pharmacol 107:450–459 158. Vanden Heuvel JP, Kuslikis BI,Van Rafelghem MJ, Peterson RE (1991) Tissue distribution, metabolism, and elimination of perfluorooctanoic acid in male and female rats. J Biochem Toxicol 6:83–92

Effects on Rodents of Perfluorofatty Acids

247

159. Vanden Heuvel JP, Sterchele PF, Nesbit DJ, Peterson RE (1993) Coordinate induction of acyl-CoA binding protein, fatty acid binding protein and peroxisomal beta-oxidation by peroxisome proliferators. Biochim Biophys Acta 1177:183–190 160. Vanden Heuvel JP,Van Rafelghem MJ, Menahan LA, Peterson RE (1989) Isolation and purification of perfluorodecanoic and perfluorooctanoic acids from rat tissues. Lipids 24:526–531 161. Van Rafelghem MJ,Andersen ME (1988) The effects of perfluorodecanoic acid on hepatic stearoyl-coenzyme A desaturase and mixed function oxidase activities in rats. Fundam Appl Toxicol 11:503–510 162. Van Rafelghem MJ, Inhorn SL, Peterson RE (1987) Effects of perfluorodecanoic acid on thyroid status in rats. Toxicol Appl Pharmacol 87:430–439 163. Van Rafelghem MJ, Mattie DR, Bruner RH,Andersen ME (1987) Pathological and hepatic ultrastructural effects of a single dose of perfluoro-n-decanoic acid in the rat, hamster, mouse, and guinea pig. Fundam Appl Toxicol 9:522–540 164. Van Rafelghem MJ, Noren CW, Menahan LA, Peterson RE (1988) Interrelationships between energy and fat metabolism and hypophagia in rats treated with perfluorodecanoic acid. Toxicol Lett 40:57–69 165. Van Rafelghem MJ,Vanden Heuvel JP, Menahan LA, Peterson PE (1988) Perfluorodecanoic acid and lipid metabolism in the rat. Lipids 23:671–678 166. Wilson MW, Lay LT, Chow CK, Tai HH, Robertson LW, Glauert HP (1995) Altered hepatic eicosanoid concentrations in rats treated with the peroxisome proliferators ciprofibrate and perfluorodecanoic acid. Arch Toxicol 69:491–497 167. Wilson MW, Leung LK, Hong JT, Glauert HP (1997) Effect of the peroxisome proliferators ciprofibrate and perfluorodecanoic acid on eicosanoid concentrations in rat liver.Adv Exp Med Biol 400A:439–445 168. Witzmann F, Coughtrie M, Fultz C, Lipscomb J (1996) Effect of structurally diverse peroxisome proliferators on rat hepatic sulfotransferase. Chem Biol Interact 99:73–84 169. Witzmann FA, Fultz CD, Lipscomb JC (1996) Toxicant-induced alterations in two-dimensional electrophoretic patterns of hepatic and renal stress proteins. Electrophoresis 17:198–202 170. Witzmann FA, Jarnot BM, Clack JW (1994) Charge modification in rodent hepatic Grp78/BiP following exposure to structurally diverse peroxisome proliferators. Appl Theor Electrophor 4:81–88 171. Witzmann FA, Jarnot BM, Parker DN, Clack JW (1994) Modification of hepatic immunoglobulin heavy chain binding protein (BiP/Grp78) following exposure to structurally diverse peroxisome proliferators. Fundam Appl Toxicol 23:1–8 172. Witzmann FA, Parker DN (1991) Hepatic protein pattern alterations following perfluorodecanoic acid exposure in rats. Toxicol Lett 57:29–36 173. Witzmann FA, Parker DN, Jarnot BM (1994) Induction of enoyl-CoA hydratase by LD50 exposure to perfluorocarboxylic acids detected by two-dimensional electrophoresis. Toxicol Lett 71:271–277 174. Wolvetang EJ, Tager JM, Wanders RJ (1992) Latency of peroxisomal palmitoyl-CoA oxidation in digitonin permeabilized fibroblasts: the effect of ATP on peroxisomal permeability. Prog Clin Biol Res 375:223–229 175. Yamamoto A, Kawashima Y (1997) Perfluorodecanoic acid enhances the formation of oleic acid in rat liver. Biochem J 325:429–434 176. Yang, Q, Xie Y, DePierre JW (2000) Effects of peroxisome proliferators on the thymus and spleen of mice. Clin Exper Immunol 122:219–226 177. Yeldandi AV, Rao MS, Reddy JK (2000) Hydrogen peroxide generation in peroxisome proliferator-induced oncogenesis. Mutat Res 448:159–177 178. Ylinen M,Auriola S (1990) Tissue distribution and elimination of perfluorodecanoic acid in the rat after single intraperitoneal administration. Pharmacol Toxicol 66:45–48 179. Ylinen M, Hanhijarvi H, Jaakonaho J, Peura P (1989) Stimulation by oestradiol of the urinary excretion of perfluorooctanoic acid in the male rat. Pharmacol Toxicol 65:274–277

248

J.W. DePierre

180. Ylinen M, Hanhijarvi H, Peura P, Ramo O (1985) Quantitative gas chromatographic determination of perfluorooctanoic acid as the benzyl ester in plasma and urine. Arch Environ Contam Toxicol 14:713–717 181. Ylinen M, Kojo A, Hanhijarvi H, Peura P (1990) Disposition of perfluorooctanoic acid in the rat after single and subchronic administration. Bull Environ Contam Toxicol 44:46–53 182. Youssef J, Badr M (1998) Extraperoxisomal targets of peroxisome proliferators: mitochondrial, microsomal, and cytosolic effects. Implications for health and disease. Crit Rev Toxicol 28:1–33

CHAPTER 8

Monofluorinated Polycyclic Aromatic Hydrocarbons: Standards in Environmental Chemistry and Biochemical Applications Gregor M. Luthe 1 · Freek Ariese · Udo A.T. Brinkman Department of Analytical Chemistry and Applied Spectroscopy, Vrije Universiteit, De Boelelaan 1083, 1081 HV Amsterdam, the Netherlands 1 E-mail: [email protected], Tel.: +49 1702438926, Fax: +31 20 4447543

We have recently synthesized a series of monofluorinated polycyclic aromatic hydrocarbons (FPAHs) and found that they provide a promising set of internal standards for environmental analysis of trace-level PAHs. Fluorine substitution has two antagonistic effects: the creation of a permanent dipole moment and the reduction of the London forces. As a result, F-PAHs are largely similar to the corresponding non-fluorinated PAHs in terms of physico-chemical properties. PAHs are routinely determined in many types of samples because of their carcinogenic properties, and in this chapter we address the suitability of F-PAHs as internal standards in PAH analysis during sample preparation, for liquid and gas chromatography, and for Shpol’skii spectroscopy. Additional work indicates that F-PAHs are also useful as marker compounds in mechanistic studies. Because of the strength of the fluorine-carbon bond, fluorine substitution can be used to block certain positions of the hydrocarbons, and the bond is highly resistant to scrambling.As illustrations, fluoropyrene metabolism in vivo and flash vacuum thermolysis are discussed. Keywords. Monofluorinated PAHs, F-PAHs, Internal standard, Solid-phase extraction, Shpol’skii

spectroscopy, Flash vacuum thermolysis, PAH metabolism, LC-UV, GC-MS

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250

1

Introduction

2

F-PAHs as Internal Standards for Extraction and Clean-Up Procedures . . . . . . . . . . . . . . . . . . . . . . 252

2.1 2.2 2.3 2.4

Introduction to Solid-Phase Extraction (SPE) . . . . . . . . . . . . 252 SPE Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 Comparison of Desorption Profiles of PAHs and F-PAHs . . . . . 253 F-PAHs as Internal Standards to Monitor PAH Recovery from Soil 255

3

F-PAHs in Chromatographic Separation and Detection . . . . . . 256

3.1 3.2 3.3 3.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chromatographic Procedures . . . . . . . . . . . . . . . . . . . . Liquid Chromatography with UV Absorption Detection (LC-UV) Gas Chromatography with Mass Spectrometric Detection (GC-MS)

4

F-PAHs in Shpol’skii Spectroscopy

256 256 257 260

. . . . . . . . . . . . . . . . . 262

The Handbook of Environmental Chemistry Vol. 3, Part N Organofluorines (ed. by A.H. Neilson) © Springer-Verlag Berlin Heidelberg 2002

250

G.M. Luthe et al.

4.1 4.2 4.3 4.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 Shpol’skii Spectroscopy Set-Up . . . . . . . . . . . . . . . . . . . 263 Comparison of Shpol’skii Spectra of F-PAHs and Parent PAHs . . 264 Quantitative Studies . . . . . . . . . . . . . . . . . . . . . . . . . 265

5

F-PAHs as Model Compounds to Study Reaction Mechanisms

5.1 5.2

In Vivo Studies of Pyrene Metabolism . . . . . . . . . . . . . . . . 267 Flash Vacuum Thermolysis of F-PAHs . . . . . . . . . . . . . . . . 270

6

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272

7

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273

. . 267

List of Abbreviations FF-PAH F-PAH FVT GC HOMO IS LB-ratio LC LOD LUMO LVI MS NMR PAH PLE QSAR RP SDB SPE

polyfluorinated polycyclic aromatic hydrocarbon monofluorinated polycyclic aromatic hydrocarbon flash vacuum thermolysis gas chromatography highest occupied molecular orbital internal standard length-to-breadth ratio liquid chromatography limit of detection lowest unoccupied molecular orbital large-volume injection mass spectrometry nuclear magnetic resonance polycyclic aromatic hydrocarbon pressurized liquid extraction quantitative structure-activity relationship reversed phase styrene-divinylbenzene solid-phase extraction

1 Introduction Monofluorinated polycyclic aromatic hydrocarbons (F-PAHs) show interesting properties that can be qualitatively understood if we consider the influence of the fluorine atom on the aromatic system. (1) Fluorine induces a dipole moment, though not as strong as one might have expected. The small fluorine atom is – using the terminology of the hard-and-soft-acids-and-bases (HSAB) model – too hard to retain a significant part of the electron charge. In addition, hyperconju-

Monofluorinated Polycyclic Aromatic Hydrocarbons

251

gation, caused by overlapping of the aromatic p-orbitals and the free fluorine porbitals, results in a partial back-transfer of the charge to the aromatic system. (2) As a result of reduced polarizability, the London forces are weaker for F-PAHs in comparison with the corresponding non-substituted PAHs, and this effect is antagonistic to the effect of the increased dipole moment. Consequently, the overall effect of fluorine substitution on the physical properties of PAHs is very small, and as a result F-PAHs show a behavior remarkably similar to that of the corresponding non-fluorinated parent compounds. A general discussion is presented in Chap. 1. This is also borne out by the physico-chemical properties of fluorinated aromatic hydrocarbons that are discussed by Ellis et al. in their chapter in this volume. Furthermore, F-PAHs have no known natural occurrence in the environment [17]; see also Gribble, this volume. For all these reasons, F-PAHs would be ideal compounds as internal standards in PAH analysis [32] and would also be valuable model compounds in mechanistic studies on PAHs. The first investigations in this area were carried out by the group of Andersson [2], who attempted to use polyfluorinated PAHs (FF-PAHs) as internal standards in gas chromatography using flame ionization detection (GC-FID). At the same time, we studied the behavior of poly- and perfluorinated PAHs in liquid chromatography (LC). These compounds were obtained from V. Platonov of Nowosibirsk, Russia, and we were able to describe the retention behavior of FFPAHs on a monomeric reversed-phase column in terms of the degree of fluorination, electronic factors such as dipole moments, and London forces. Unfortunately the physical properties of polyfluorinated PAHs and perfluorinated PAHs were found to be very different from those of non-substituted PAHs, which made them less suitable for use as recovery standards. On the other hand, our investigations with monofluorinated PAHs showed a surprising similarity to parent PAHs in physico-chemical properties and retention times in both GC and LC analysis. This prompted us to study in greater detail their usefulness as internal standards in analytical chemistry. PAHs are regularly determined in a variety of matrices (e.g., food products, water, soil and sediment, biota, air, and cigarette smoke) because of the carcinogenic properties of many PAH congeners. Internal standards are routinely added to the samples at various stages of the analytical process to monitor losses during sample treatment, extraction, or cleanup, or to correct for variability induced by such factors as solvent evaporation, changes in detector sensitivity, injection volumes, etc. Although a series of F-PAHs is now commercially available as certified calibration solutions or mixtures from Chiron, Trondheim, Norway, these were not available when we began our studies.We therefore developed new synthetic pathways involving intra-molecular rearrangements in flash vacuum thermolysis at high temperature [9, 20], photocyclization, oxidation, and Schiemann reactions at low temperatures. In this chapter we will first demonstrate that F-PAHs are useful as internal standards for a broad range of analytical applications. In Sect. 2 we will show their value in extraction and clean-up procedures for PAHs in wastewater using solid-phase extraction (SPE), and for soil samples using pressurized liquid extraction (PLE). The chromatographic behavior (LC and GC) of FF-PAHs and

252

G.M. Luthe et al.

F-PAHs will be discussed in Sect. 3. The use of F-PAHs to correct for specific sources of variability in low-temperature Shpol’skii spectroscopy will be the topic of Sect. 4. Finally, in Sect. 5 we will show some preliminary results on the use of F-PAHs as model compounds for the elucidation of chemical reaction mechanisms, rearrangements, and metabolic pathways.

2 F-PAHs as Internal Standards for Extraction and Clean-Up Procedures 2.1 Introduction to Solid-Phase Extraction (SPE)

The determination of PAHs in matrices such as soil, food, air, and water typically involves extraction, clean-up, and pre-concentration in order to isolate the analytes from the matrix, remove potential interference, and improve the sensitivity of the method [12]. First we focus on water samples. Many laboratories involved in the environmental control and monitoring of water systems use SPE routinely [27]. Solvent evaporation, breakthrough of analytes on SPE cartridges, incomplete desorption, matrix interactions, and/or other problems in sample pretreatment may cause low or inconsistent recoveries for PAHs. This is the case not only for low molecular mass/volatile PAHs such as naphthalene and acenaphthylene since problems have also been encountered with the larger PAHs. By using one or more well-chosen internal standards, it is possible to determine the magnitude of these variations and, if necessary, correct for them. In this section we will demonstrate the potential suitability of F-PAHs as internal standards in SPE. SPE characteristics of PAHs and F-PAHs were determined on two types of SPE sorbents: octyl-bonded silica (Bakerbond 60 Å SPE octyl disposable cartridges; 3 ml, 200 mg, 40 mm average particle diameter) and styrene-divinyl-

Fig. 1. Structures of selected monofluorinated polycyclic aromatic hydrocarbons discussed in this chapter

Monofluorinated Polycyclic Aromatic Hydrocarbons

253

benzene copolymer (Bakerbond 2,7 Å SPE SDB-1 disposable cartridges; 3 ml, 200 mg, 43–125 mm average particle diameter).We used eight PAHs that are listed as priority pollutants by the US-EPA, and eight corresponding F-PAHs as internal standards (ISs). In order to make the study representative for a broad group of PAHs, four compounds with three, four, or five fused rings (acenaphthylene, fluorene, fluoranthene, benzo[k]fluoranthene), one linear-condensed compound (naphthalene), one peri-condensed system (pyrene), and two angular-condensed (phenanthrene, chrysene) polyarenes were used (Fig. 1). SPE was carried out with spiked distilled water and wastewater. Full details can be found in [35]. 2.2 SPE Procedures

Distilled water and wastewater samples (after azide sterilization and filtration) were spiked with mixtures of PAHs and F-PAHs. Both types of SPE cartridges were conditioned with 15 ml of acetonitrile, followed by twice 3 ml distilled water/acetonitrile (85:15, v/v). The sample solutions were loaded onto the octyl columns (100 ml) and on the SDB-1 columns (1000 ml) under reduced pressure. After loading, the columns were washed with 3 ml of distilled water/acetonitrile (85:15, v/v) and dried. The analytes were then eluted from the columns with acetonitrile in 100-µl steps for the octyl, and in 1-ml steps for the SDB-1 cartridges into glass tubes, using centrifugation at 3000 rpm instead of vacuum sucking. To avoid evaporation the cartridges were fitted tightly in the centrifuge tubes with cleaned rubber rings and closed at the top with a stopper.With naphthalene, 1-fluoronaphthalene, acenaphthylene, and 1-fluoroacenaphthylene the breakthrough volumes were found to be about 100 ml of distilled water for the octyl cartridges and about 1000 ml for the SDB-1 cartridges. 2.3 Comparison of Desorption Profiles of PAHs and F-PAHs

To make sure that we would focus only on the adsorption and desorption behavior during SPE, we did not include any solvent evaporation steps. Two commonly used SPE sorbents were selected. Octyl-bonded silica material separates the analytes on the basis of weak, non-polar attractive interactions with the organic chains; SDB-1 is a cross-linked styrene-divinylbenzene copolymer which separates analytes on the basis of strong p–p and weak, non-polar interactions. As far as possible, the same SPE procedure was used for both sorbents. However, because of the much higher sorption strength of the copolymer, we loaded 100-ml water samples onto the octyl columns and 1000-ml samples onto the SDB-1 columns. For the same reason, the elution from the octyl cartridges was studied in 100-ml steps and from the SDB-1 cartridges in 1-ml steps. As an example, Fig. 2 shows the results for fluoranthene and 3-fluorofluoranthene as the corresponding F-PAH, spiked in a 1:1 mass ratio to distilled water. The experiments were carried out in triplicate, i.e., on three separate cartridges, and are shown here as separate curves.As is to be expected, there are distinct dif-

254

G.M. Luthe et al.

ferences between the elution volumes for the octyl sorbent (top frames) and the copolymer sorbent (center frames). However, it is much more important – and relevant for our purposes – that, whereas there are noticeable differences between the elution profiles recorded for a given compound on three separate cartridges (especially in the case of octyl material), the profiles found on each individual cartridge are very similar for fluoranthene and 3-fluorofluoranthene. Essentially the same results were obtained for all other PAH/F-PAH pairs, and for 10:1, 1:1 as well as 1:10 mass ratios, i.e., over two orders of magnitude [35]. As an example of a practical application, samples of municipal wastewater were analyzed.A typical result is included in Fig. 2 (bottom). The elution profiles

Fig. 2. Desorption profiles of fluoranthene (black symbols) and 3-fluorofluoranthene (open

symbols), eluted with acetonitrile from (top) octyl SPE, loaded with 100 ml of distilled water; (center) SDB-1, loaded with 1000 ml of distilled water; (bottom) octyl SPE, loaded with 100 ml of wastewater. Spiking level, 100 µg/l of each of the test analytes

Monofluorinated Polycyclic Aromatic Hydrocarbons

255

of fluoranthene and 3-fluorofluoranthene are virtually identical. Furthermore, for the higher molecular mass PAHs, we observed a distinct shift to lower elution volumes compared with distilled water samples. Probably, this was caused by overloading the active sites of the sorbent with non-polar compounds from the sample matrix. This effect can cause strong differences in the recoveries if the same elution volume is used for calibration solutions and for real samples with varying levels of matrix interferences.Very similar elution profiles were also observed for the other PAH/F-PAH pairs tested after spiking to wastewater [35]. Obviously, F-PAHs can indeed serve a most useful purpose as surrogate standards in SPE procedures. 2.4 F-PAHs as Internal Standards to Monitor PAH Recovery from Soil

Promising results were also obtained with liquid-liquid extraction and pressurized liquid extraction (PLE). In this study, the suitability of F-PAHs as ISs for soil analysis was evaluated. The selected compounds in a 1:1 methanol solution were spiked to an organic soil, yielding realistic levels of 75 ng/g soil. The slurry was homogenized by vigorous shaking and the methanol allowed to evaporate to dryness in a fume hood. This was taken as the zero time of the study. The dried spiked soil samples were then stored for 25 h in capped glass containers at room temperature, protected from light. PLE (10 min static plus dynamic extraction of a 50-mg sample with toluene at 200 °C and 10 MPa) was used to extract the PAHs and F-PAHs from the soil at specific times after spiking.Analysis was carried out by GC-MS (see Sect. 3.4). The concentrations of the PAHs found in the spiked samples were compared with and without IS-based correction.

Fig. 3. Recovery (%) of 1-fluoronaphthalene, naphthalene, 1-fluoropyrene and pyrene, spiked

to an organic soil. The samples were extracted by PLE at different times after spiking; analysis was carried out by GC-MS

256

G.M. Luthe et al.

As an example, Fig. 3 illustrates the recoveries of 1-fluoronaphthalene, naphthalene, 1-fluoropyrene, and pyrene as a function of time. Because of established evaporation losses during the initial drying, the recoveries at the onset of the experiment were rather poor for naphthalene and 1-fluoronaphthalene, but were greatly improved for pyrene/1-fluoropyrene. As is often observed, the recoveries decreased slowly during longer storage times, which probably reflects analytematrix interactions or aging effects. However, for all the PAH/F-PAH pairs that were studied, the recoveries were closely similar. The results demonstrate the suitability of the F-PAHs to mimic the behavior of their parent compounds, which makes them suitable as internal standards.

3 F-PAHs in Chromatographic Separation and Detection In most analytical procedures sample preparation is followed by a separation step such as column liquid chromatography (LC) [50, 52] or capillary gas chromatography (GC). The behavior of fluorinated PAHs in LC and GC is discussed below. 3.1 Introduction

First, we will discuss the retention behavior of higher fluorinated PAHs (FFPAHs) in reversed-phase liquid chromatography (RP-LC). Difluorinated naphthalene, di-, tri-, and tetrafluorinated phenanthrenes, difluorinated chrysenes, and difluorinated benzo[g]chrysenes with varying substitution patterns were used as model compounds. The influence of the degree of polymerization of the C18 phase, geometric factors like the maximum length-to-breadth ratio (LB ratio) of the FF-PAHs, and electronic factors like dipole moment and London forces were studied. The role of polymerization in LC analysis is discussed in detail by Poster et al. [42]. As a result of these studies, attention was subsequently focused on monofluorinated PAHs (F-PAHs). In GC the use of an internal standard has become routine. For the determination of PAHs with GC-MS, perdeuterated PAHs are commonly used as internal standards [40]. The use of methylated PAHs has also been reported [50], but these compounds often occur in contaminated environmental samples. Here, we will discuss the suitability of F-PAHs as internal standards.With the exception of polyfluorinated phenanthrenes and chrysenes [2], (poly)fluorinated-PAHs have not been studied by GC before. 3.2 Chromatographic Procedures

In addition to the F-PAHs shown in Fig. 1, the following FF-PAHs were used: 1,5difluoronaphthalene, 3,6-, 2,3-, 4,5-, 2,7-, 2,6-, 2,4-, 1,4-, 1,8-, 2,5-, and 3,5-difluorophenanthrene, 2,3,6-, 3,4,6-, and 2,4,6-trifluorophenanthrene, 1,3,6,8tetrafluorophenanthrene, 1,4-, 2,4-, and 1,2-difluorochrysene, 12,14- and 2,4-difluorobenzo[g]chrysene.

257

Monofluorinated Polycyclic Aromatic Hydrocarbons

The LC stationary phase used was a 9 µm monomeric Nukleosil RP-18 bonded silica with an aTBN/BaP ratio of 1.67. To determine the retention factors of the individual FF-PAHs, water/acetonitrile (40:60, v/v) was used for the substituted naphthalenes and phenanthrenes, and water/acetonitrile (25:75, v/v) for the chrysenes and benzo[g]chrysenes. GC-MS was carried out on an HP 6890 instrument, using a Restek XTI-5 capillary column, 50-µl large-volume injection, and detection in the total-ion scan mode. 3.3 Liquid Chromatography with UV Absorption Detection (LC-UV)

When the parent PAHs and polyfluorinated PAHs (FF-PAHs) were analyzed on the monomeric C-18 phase, the retention factors of the FF-PAHs were found to be much larger. Furthermore, there were rather large differences in retention between substitutional isomers. Table 1 gives an overview of the relevant data and also lists the dipole moments of the FF-PAHs, calculated with the Ph4 program assuming vacuum conditions. As mentioned before, fluorine substitution of PAHs causes two antagonistic effects that have an impact on the interaction with the solvent. (1) The creation of a permanent dipole would cause a greater hydrophilicity and consequently

Table 1. Retention factors of FF-PAHs as measured on a monomeric RP stationary phase and their calculated dipole moments. The compound numbers refer to Fig. 4

No.

FF-PAH

Retention factor

Dipole moment/ 10–30 Cm

PAH 1 2 PAH 3 4 5 6 7 8 9 10 11 12 13 14 15 PAH 16 17 18

Naphthalene 1-Fluoronaphthalene 1,5-Difluoronaphthalene Phenanthrene 2,7-Difluorophenanthrene 1,8-Difluorophenanthrene 2,5-Difluorophenanthrene 3,6-Difluorophenanthrene 3,5-Difluorophenanthrene 4,5-Difluorophenanthrene 1,4-Difluorophenanthrene 2,4-Difluorophenanthrene 2,3-Difluorophenanthrene 2,4,6-Trifluorophenanthrene 2,3,6-Trifluorophenanthrene 3,4,6-Trifluorophenanthrene 1,3,6,8-Tetrafluorophenanthrene Chrysene 1,4-Difluorochrysene 2,4-Difluorochrysene 1,2-Difluorochrysene

3.90 4.75 4.95 7.58 11.0 12.6 8.60 8.58 8.76 8.36 12.5 11.9 11.2 12.9 12.2 9.66 13.4 5.94 10.5 9.08 8.42

– 1.70 0.00 – 0.11 2.95 2.79 2.89 3.40 3.22 0.23 1.85 2.92 2.93 3.27 4.47 0.06 – 0.27 1.89 2.94

258

G.M. Luthe et al.

shorter retention times in RP-LC. Of course, the substitution pattern will also influence the magnitude of the overall dipole moment. When substitution occurs in a electron-poor region, the increase of the dipole will be relatively weak. (2) Fluorination also causes a weakening of the London forces, that will lead to increased retention times in RP-LC. Here also, the substitution pattern will be an important determinant. For instance, in the case of 1,8-difluorophenanthrene fluorination in electron-rich regions causes a relatively strong reduction of the London forces. Multiple fluorination in one aromatic ring has a larger impact on the London forces than substitution on different rings, as can be seen by comparing the set of 1,4-, 2,4-, and 2,3-difluorophenanthrene with that of the 2,5, 2,7-, 3,5-, 3,6-, and 4,5-difluorophenanthrene isomers: the former group shows higher retention factors. When comparing the retention factors of mono-, di-, tri-, and tetrafluorinated phenanthrenes, a positive correlation with the degree of fluorination is observed. These observations can be explained by a partitioning model, i.e., the distribution of the analyte between the non-miscible liquid phases. In other words, the monomeric C18 bonded silica phase can be regarded as a liquid lipophilic stationary phase, for which highly fluorinated PAHs with a small dipole moment have the highest affinity. In Fig. 4, retention factors are plotted vs the permanent dipole moment. FF-PAH isomers with the same degree of substitution on the same rings were grouped together. The three difluorinated chrysenes show essentially the same

Fig. 4. Correlation of the retention factor of FF-PAHs on a monomeric RP column and dipole moment. Isomers were grouped according to degree of substitution and substitution ring

Monofluorinated Polycyclic Aromatic Hydrocarbons

259

behavior as the corresponding difluorophenanthrenes. In addition we carried out retention measurements on a polymeric phase. Here, the behavior of FF-PAHs can also be understood as a combination of the slot model of Wise [51] and a partitioning model. Molecules that are able to penetrate deeper into the slots of the stationary phase, for instance molecules with a large LB ratio, will show increased retention [33]. In summary, all FF-PAHs investigated showed much longer retention times on RP-LC material than the corresponding parent PAHs. One advantage of using FF-PAHs as internal standards is the large choice of possible isomers. Because of their widely varying retention factors it is relatively easy to select a suitable set of polyfluorinated compounds that do not co-elute with any major PAH in the sample. For example, it is possible to synthesize up to 25 different difluorinated phenanthrenes, in contrast to just one perdeuterated phenanthrene. On the other hand, the fact that FF-PAHs have retention times and other properties that are rather different from those of the parent PAHs makes them less suitable as internal standards. In other words, polyfluorination may well be less than ideal. Further research was then focused on monofluorinated PAHs, which were found to show a behavior in RP-LC much more similar to that of the parent PAHs. Figure 5 illustrates the LC separation of eight PAH/F-PAH pairs on a 5-µm poly-

Fig. 5. LC-UV of eight PAHs and their monofluorinated analogues. Column: 5-µm polymer RP-18 Bakerbond PAH 16 Plus, Eluent: water/acetonitrile (50:50) (4.0 min hold) then to pure acetonitrile (4–15 min) and hold to 30 min, UV detection at 254 nm

260

G.M. Luthe et al.

mer RP-18 Bakerbond PAH 16 Plus column with a water-acetonitrile gradient as the eluent. The resolution of the pairs was found to vary from 1.25 (chrysene/3fluorochrysene) to 3.5 (benzo[k]fluoranthene/9-fluorobenzo[k]fluoranthene). The slightly stronger retention of the F-PAHs reflects their more lipophilic character caused by the reduction of the London forces. As regards UV absorption detection, fluorine substitution causes no changes in the general shape of the spectra, and only minor spectral shifts (0.5–5 nm) to longer wavelengths. These bathochromic shifts can be explained on the basis of hyperconjugation, which causes an extension of the conjugated system. However, the shifts are minor by comparison with typical detector bandwidths, so in most cases no adjustments in detector wavelength settings will be necessary. 3.4 Gas Chromatography with Mass Spectrometric Detection (GC-MS)

Eight F-PAHs were subjected to GC-MS, using 50-ml large-volume injections (LVI) of standard solutions containing 2 ng/ml of the individual F-PAHs. The results fully agreed with those obtained by 1-µl injections in the hot splitless mode of the same mass of the individual F-PAHs, showing that no degradation occurred in the liner. Using 70-eV electron impact ionization, the most abundant ions for the F-PAHs were [M+1]+, [M]+·, [M–1]+, [M–2]+·, [M–3]+, [M–20]+· and [M]++ [36]. In all cases the molecular ion was found to be the most abundant ion, and was consequently selected for quantification purposes throughout the study. Although the ion [M–20]+·, corresponding to the loss of HF, had a relative

Fig. 6. GC–MS of eight PAH/F-PAH pairs on a Restek XTI-5 column. Detection was over the

m/z 10–300 range

Fig. 7. RSD (n =7) of PAH peak areas in GC-MS without IS, with one or more F-PAHs as ISs, or using its ‘own’ (i.e., corresponding) F-PAH

Monofluorinated Polycyclic Aromatic Hydrocarbons

261

262

G.M. Luthe et al.

abundance of only 2–5%, it is characteristic for F-PAHs and was therefore selected for identification. Figure 6 shows that for six out of the eight PAH/F-PAH pairs, the fluorinated compound elutes slightly in front of, but not totally baseline-resolved from, its parent PAHs. For two pairs, however, 1-fluoronaphthalene/naphthalene and 2-fluorofluorene/fluorene, there is no resolution. The minor differences in retention times are another indication of the close similarity of the properties of the PAH/F-PAH pairs. Actually, preliminary studies in our laboratory have shown that even tandem GC (GC ¥GC) with a non-polar and a polar phase cannot further separate these compounds. Of course, in practice, such coelution should not create problems when using mass-selective detection. As regards analytical performance, responses for 1-fluoronaphthalene and 2fluorofluorene were linear over the tested range of 0.5–100 ng/ml with regression coefficients better than 0.97 (n=7). For the other F-PAHs, regression coefficients from 0.996 to 0.9997 (n =10) were found in the range 0.5–500 ng/ml. The experimentally determined limits of detection (LOD) were 0.4 and 0.2 ng/ml for 1-fluoronaphthalene and 2-fluorofluorene, respectively, and somewhat better, i.e., 0.09–0.1 ng/ml, for the other F-PAHs. The higher LODs, reported for the two fast eluting F-PAHs, are probably due to evaporation losses during the solvent elimination step of the LVI procedure. The present results were in the range of those previously found for native PAHs. In order to test the suitability of F-PAHs as ISs, standard mixtures of eight PAHs were analyzed by GC-MS after the addition of a mixture of the corresponding F-PAHs. The relative standard deviations (RSD; n=5) obtained for the parent compounds were calculated without IS and after ratioing with one or more ISs. The results of Fig. 7 clearly show that using at least one IS dramatically improved the RSD for all compounds. For most PAHs some further improvement was obtained when several ISs were used. Especially in the case of the rather volatile naphthalene the best result was obtained when its ‘own’ fluorinated IS, i.e., 1-fluoronaphthalene, was used.

4 F-PAHs in Shpol’skii Spectroscopy 4.1 Introduction

Many analytical approaches for PAHs make use of the fact that most PAHs and PAH derivatives show strong native fluorescence. Although straightforward fluorescence measurements can be used for very simple mixtures or as a fast screening method, fluorescence spectra are not usually sufficiently selective to analyze complex mixtures or identify individual compounds. The main reason why conventional, room-temperature fluorescence spectra are fairly broad and featureless is related to inhomogeneous broadening of the signal. In an amorphous environment each analyte molecule is subjected to a different local electric field, and this results in a Gaussian distribution of electronic transitions [44].

Monofluorinated Polycyclic Aromatic Hydrocarbons

263

In the Shpol’skii approach, the inhomogeneity of the matrix is strongly reduced by cooling the sample to cryogenic temperatures of 77 K or lower in a solvent such as an n-alkane that forms a crystalline structure at low temperatures. When the analyte molecules (guests) are compatible with the matrix lattice (host) in terms of polarity, size, and shape, the guest molecules will occupy specific sites in the crystal, replacing one or more solvent molecules [39, 41, 44]. In the ideal case each individual analyte molecule will experience the same interaction with its surroundings, and narrow-banded, vibrationally resolved fluorescence spectra will be obtained. The requirement for a good molecular fit imposes some limitations on the choice of the n-alkane solvent. Often the best spectral resolution is obtained if there is a close match between the longest dimension of the guest molecule and the length of the alkane molecule – the ‘lock-and-key’ principle. If the analyte is able to occupy several different, energetically non-equivalent sites, a multiplet structure is observed. The exact wavelength of the 0–0 transition(s) and the vibrational fine structure are highly specific for the analyte/matrix combination that is used. Shpol’skii spectra are therefore very suitable for fingerprint identification, and can be used for the quantitation of individual compounds (even minor constituents) in complex matrices without prior separation. The Shpol’skii technique has been applied extensively to PAH analysis in a variety of samples [3, 16]. Shpol’skii spectroscopy can be used in a straightforward manner for identification, but quantitative application is more difficult. Changes in optical alignment of the detector set-up or lamp/laser output will directly affect the signal intensities, and it is usually not possible to freeze all samples and calibration solutions in a single run. More importantly, differences in solvent purity or in cooling rate may influence the rate of matrix solidification and thus, the efficiency with which analyte molecules are ‘trapped’ during the crystallization process [22, 41]. The use of an internal standard is recommended to correct for these effects. In the past, perdeuterated PAHs have been used for this purpose [53]. As far as we know, F-PAHs have not yet been systematically studied in Shpol’skii spectroscopy. Although one might expect that such compounds will not readily fit into an alkane matrix in view of the high electronegativity of fluorine, it will nevertheless be shown below that good-quality spectra can be obtained. For structures and substituent numbering of the F-PAHs; see Fig. 1. 4.2 Shpol’skii Spectroscopy Set-Up

Low-temperature fluorescence spectra were recorded at 11 K using a closedcycle cryocooler. For most measurements this temperature was reached in approximately 60 min using only cryocooling, that is without fast precooling in liquid nitrogen. The sample holder was made from gold-plated copper and could hold four 10-µl samples at a time. Each sample was covered by a sapphire window. As a result of good thermal conductivity and the small sample volume relative to the heat capacity of the sample holder, temperature gradients within the alkane solution were minimal, and the matrix solidified almost instantaneously, i.e., in less than 1 s. The samples were excited with a 450-W xenon lamp and an

264

G.M. Luthe et al.

excitation monochromator with a 5-nm bandpass. The luminescence was collected with a large F/1 quartz lens and focused on the entrance slit of a high-resolution monochromator equipped with an intensified linear diode array detector; the spectral resolution was 0.1 nm. 4.3 Comparison of Shpol’skii Spectra of F-PAHs and Parent PAHs

Shpol’skii spectra of the F-PAHs were recorded in the conventional mode with lamp excitation. The initial experiments focused on the question whether F-PAHs would give well-resolved Shpol’skii fluorescence emission spectra that are sufficiently different from those of the parent PAHs. For all compounds fluorescence spectra were recorded in n-octane, n-heptane, n-hexane, and n-pentane matrices; concentrations were typically 10–6 mol/l. In Fig. 8, the fluorescence spectra are given for the analyte/matrix combination that gave the best Shpol’skii spectrum in terms of intensity and simplicity (no complex multiplet structure). The wavelength repeatability of the emission lines was better than 0.1 nm for six separate measurements. The spectra of the corresponding parent PAHs in the same matrix are included for comparison. For all analyte/alkane combinations phosphorescence measurements were also carried out. Four compounds (1-fluoronapthalene, 1-fluorophenanthrene, 3-fluorochrysene, and 2-fluorofluorene) yielded high-resolution Shpol’skii phosphorescence spectra in n-pentane, n-hexane, n-octane, and n-hexane, respectively (spectra not shown, see [38]). In all cases the phosphorescence spectra could be easily distinguished from those of the parent PAHs. As illustrated in Fig. 8, F-substitution brings about changes in the fluorescence emission spectra. The 0–0 transition region is significantly shifted; consequently, the vibronic transitions of the F-PAHs and their parent PAHs show no overlap, and simultaneous identification of F-PAHs and PAHs in a mixture will be possible. This is illustrated by the spectra for the F-PAH/PAH mixtures included in the figure. The observed spectral shifts in the fluorescence spectra were between 0.5 nm (3-fluorochrysene/chrysene) and 5.8 nm (2-fluorofluorene/fluorene). Upon fluorine substitution, the fluorescence (and most phosphorescence) spectra of the PAHs are shifted to longer wavelengths. The bathochromic shifts can be attributed to hyperconjugation, which leads to an extension of the dimensions of the conjugated p-electronic system and a narrowing of the HOMO-LUMO gap. In many cases, the intensities of the 0–0–transitions (relative to the vibronic transitions) of the F-PAHs are somewhat higher than observed for the parent PAHs, presumably because fluorine substitution causes a reduction of molecular symmetry [41]. As mentioned above, the positive outcome of these first experiments was somewhat unexpected in view of the electronegativity of fluorine. One should realize, however, that, in addition to purely electronic factors, the molecular shape of the F-PAHs strongly influences the fit in the matrix. Fluorine and hydrogen have rather similar atomic radii, especially in aromatic systems, caused by attractive p–p interactions between carbon and fluorine. Larger, bulkier halogen substituents would reduce the fit in the layered alkane matrix, but apparently such an effect is absent or insignificant in the case of fluorine.

Monofluorinated Polycyclic Aromatic Hydrocarbons

265

Fig. 8. Shpol’skii spectra of seven PAH/F-PAH pairs, both as pure solutions and in a 1:1 mixture. Solvent: n-hexane, with the exception of 1-fluoronaphthalene/naphthalene (n-pentane) and 3-fluorochrysene/chrysene (n-octane) [38]

4.4 Quantitative Studies

The intensity of Shpol’skii bands sometimes depends critically on experimental factors such as optical alignment, sample thickness, solvent evaporation, rate of freezing/solidification, inner filter effects, and the purity of the solvent matrix [4]. If not carefully controlled, these factors will diminish the value of Shpol’skii spec-

266

G.M. Luthe et al.

troscopy as a quantitative technique. Having demonstrated the suitability of Shpol’skii spectroscopy for qualitative purposes, we now studied its usefulness for quantification, using 10–6 mol/l solutions of 1-fluoropyrene and pyrene in n-octane and n-hexane as model systems. The repeatability of the procedure (n=4) was determined by doing each measurement using a different sample holder position and thereby deliberately changing the optical alignment. Sample thickness, solvent evaporation and optical alignment will affect all analyte signals to the same extent, and in such cases correction with an IS should be straightforward. The relative standard deviations (RSD) for pyrene without correction were acceptable (5.9% in n-hexane; 3.2% in n-octane), but could be considerably improved by ratioing with the 1-fluoropyrene 0–0 intensities in the same sample (RSD, 2.0% and 0.6%). Parameters such as the cooling regime may not have the same influence on the spectral properties of the parent and the fluorinated PAH. This was studied by comparing the standard cryocooling procedure and one involving fast precooling with liquid nitrogen. The main difference is that the temperature gradients within the samples, and thus the solidification rates, will not be the same with the two methods and this will affect the way in which the analytes are trapped in the crystalline environment [22]. This proved to be particularly critical in the case of n-hexane, with fast cooling resulting in approximately twofold higher fluorescence intensities. The pyrene and 1-fluoropyrene spectra were, however, both affected in the same way: the intensity ratios 1-fluoropyrene/pyrene were 2.06±0.04 for standard cooling, and 2.06±0.02 for fast precooling. These preliminary results illustrate the similarity in geometry and physico-chemical properties of pyrene and 1-fluoropyrene, and suggest that the introduction of F-PAHs may open a new way to use the Shpol’skii technique for quantitative analysis of PAHs [37]. As an application, a river sediment was extracted with n-pentane using PLE. Here, 1 ml of the crude extract in n-pentane was diluted with an equal volume of 1-fluoropyrene in n-octane. Nitrogen was used to evaporate selectively n-pentane and the volume was subsequently made up to 1 ml with n-octane. For external calibration, separate mixtures of pyrene, chrysene, benzo[a]pyrene and benzo[k]fluoranthene were prepared in the 10–5 –10–7 mol/l range with 1-fluoropyrene as IS. Each compound was determined separately at its optimum exci-

Table 2. PAH concentrations (mg/kg) in sediment internal reference material, determined by Shpol’skii spectroscopy using 1-fluoropyrene as IS. The reference data were obtained from regular quality control measurements with LC-fluorescence

Compound

Pyrene Chrysene Benzo[k]fluoranthene Benzo[a]pyrene

PAH concentrations (mg/kg) via Shpol’skii (n=3)

LC-Flu (n>30)

400±10 420±10 230±10 320±10

470±50 430±40 220±20 340±40

Monofluorinated Polycyclic Aromatic Hydrocarbons

267

tation wavelength. The relative intensities of the 0–0 transition(s) of the PAHs and IS were used to construct the calibration curves. For the analysis of the sediment extract the 0–0 intensities of the PAHs analytes and IS were measured, with an exposure time of 10 s/scan. The four target PAH analytes were readily detected.With natural samples, matrix interference present in the crude extract may well affect the spectral shape and line intensities of the analytes [4]. Indeed, the spectral resolution of the 0–0 lines was somewhat poorer than for the calibration solutions in n-octane. In such cases, the use of an IS is particularly important. Table 2 shows the PAH concentrations determined with Shpol’skii spectroscopy using 1-fluoropyrene as IS. The data agree very well with those found with LC and fluorescence detection. These are encouraging results that justify further work in this area [37].

5 F-PAHs as Model Compounds to Study Reaction Mechanisms 5.1 In Vivo Studies of Pyrene Metabolism

Toxicity of PAHs towards biota is of two kinds. In the first, the compounds themselves display toxicity [49], whereas in the second metabolites formed by the action of oxidative enzymes are the active agents responsible for the induction of cancer [10, 21] as a result of the formation of adducts with DNA [25, 43]. These metabolites have also been used as biomarkers [7. 30]. For example, the quantification of pyrene metabolites to estimate bioavailability has proven successful in both occupational [29] and environmental [15] health assessments. Enzymes responsible for the biotransformation of PAHs belong to the cytochrome P450 superfamily. Within one species a large number of genes can be found, each expressing a different cytochrome P450 with specificity for a given substrate [11, 18, 46]. The net sum of the accumulated enzyme activities will determine the formation of reactive intermediates and, hence, influence the probability of genotoxic effects. Knowledge of similarities and dissimilarities of biotransformation pathways is necessary for the development of biomarkers and the modeling of chemical fate in ecosystems. We have used 1-fluoropyrene as a model compound to study the in vivo metabolism of pyrene in the terrestrial isopod Porcellio scaber (Crustacea), and in the flatfish flounder (Platichthys flesus) [14, 34]. The formation of conjugated pyrene metabolites by Porcellio scaber [47] and the presence of pyrene metabolites in the bile of flatfish had been demonstrated previously [5, 28, 31].When exposed to non-fluorinated pyrene, 1-hydroxypyrene is formed in both species as virtually the only Phase-I metabolite. Nevertheless, the responsible enzymes may be different [1, 13]. 1-Hydroxypyrene is subsequently conjugated (Phase-II metabolism) to form polar metabolites that are subsequently excreted via urine or feces. Several types of substituted PAHs have been used to investigate electronic and steric effects on the metabolism of PAHs and the functioning of the various cytochrome P450 iso-enzymes. In this respect fluorinated analogues have distinct advantages. The similarity of F-PAHs to parent PAHs in terms of polarity and in-

268

G.M. Luthe et al.

teraction with the solvent have been described in previous sections. An additional advantage for enzyme-related studies are the comparable Van der Waals radii of fluorine and hydrogen, especially if C-F secondary binding (hyperconjugation) is taken into account [6]. Therefore, the overall dimensions and the steric fit in enzyme receptor sites will not be greatly affected by fluorine substitution. In this study, flatfish and isopods were exposed to 1-fluoropyrene via injection or via the food. The resulting metabolite mixtures were extracted from fish bile or from proteolyzed isopod hepatopancreas, and analyzed by LC with both fluorescence detection (lex/em =346/384 nm) and UV absorption detection. Since fluorinated metabolite standards are not yet available, the diode array UV spectra and a modified slot model [51] were used to help identify the different isomers formed. In addition, the electronic charge distribution measured by 13CNMR and nitration patterns of 1-fluoropyrene were used to rationalize the product ratios on the basis of the model of electrophilic aromatic substitution. Figure 9a shows the relevant portions of the LC-UV chromatograms (detection at 273 nm) of samples of flounder bile after exposure to 1-fluoropyrene (upper trace) or pyrene (lower trace). The bile from the pyrene-exposed flounder showed two main pyrene metabolites, which were identified on the basis of authentic standards as pyrene-1-glucuronide (at 20.0 min) and pyrene-1-sulfate (at 22.4 min). Both compounds are conjugates of 1-hydroxypyrene, the major hydroxylated species formed in the Phase-I oxidation of pyrene in flounder. After exposure to 1-fluoropyrene, four peaks were observed (Fig. 9a upper trace – marked by arrows). The peaks appear in two clusters, each less than 1 min later than the nearest pyrene conjugate. The UV absorption spectra of each peak of the cluster were very similar to that of the nearest-eluting pyrene conjugate. Furthermore, the ratio of the combined peak areas of both clusters is similar to the peak area ratio found for pyrene-1-glucuronide and pyrene-1-sulfate. Finally, pyrene-1-sulfate is the only conjugate present in both flounder and isopod and the two peaks at 23.1 and 23.4 min are also observed in the chromatograms of 1fluoropyrene-exposed isopods (see below). This suggests that the two clusters are glucuronide and sulfate conjugates, each consisting of two different fluorohydroxypyrene isomers. Figure 9b shows the LC-UV (273 nm) chromatograms of hepatopancreas samples of pyrene- and 1-fluoropyrene-exposed isopods. The peaks at 22.3 and 23.9 min from the pyrene-exposed isopod (lower trace) are pyrene-1-sulfate and pyrene-1-glucoside. The peak at 21.7 min is an unknown pyrene-1-conjugate that has been shown to yield 1-hydroxypyrene after enzymatic hydrolysis [26, 47]. The LC-UV chromatogram resulting from the 1-fluoropyrene exposure (Fig. 9b, upper trace) shows nine peaks which all display pyrene-like UV absorption spectra (data not shown). As was also observed for the flounder bile, the conjugates elute in clusters – although for the isopod, three clusters of three peaks each – which were tentatively assigned to three isomeric 1-fluorohydroxypyrene phase-I metabolites bound to three different conjugating groups. The electronic distribution in 1-fluoropyrene was used to rationalize the regioselective formation of different hydroxylated isomers. The electron-rich FPAHs used in our studies, including 1-fluoropyrene, react by electrophilic aro-

Monofluorinated Polycyclic Aromatic Hydrocarbons

269

Fig. 9 a, b. LC-UV (273 nm) chromatograms: a bile from flounder; b hepatopancreas from iso-

pod exposed to pyrene (lower traces) or 1-fluoropyrene (upper traces). For explanation, see text

matic reactions only at unsubstituted positions. For the hydroxylation and nitration of 1-fluoropyrene we can postulate a late transition state in terms of chemical kinetics according to the Hammond postulate. This means that the substitution pattern can be predicted on the basis of thermodynamic parameters, in this case the electron density distribution. Charge densities in 1-fluoropyrene were derived from chemical shifts in 13C-NMR [19]. Careful interpretation of the data showed that the C-6, C-8, and C-3 positions can be expected to be the preferred sites for electrophilic aromatic substitution, in the order C-6>C-8>C-3. To obtain a quantitative assessment of the selectivity of this type of reaction, nitration of 1-fluoropyrene was carried out at room temperature with nitric acid.

270

G.M. Luthe et al.

Characterization and quantification of the products was done by 19F-NMR, 1HNMR, and GC-MS. The reaction yielded three nitrated 1-fluoropyrenes: 1-fluoro6-nitropyrene, 1-fluoro-8-nitropyrene, and 1-fluoro-3-nitropyrene in the ratio 55:36:9. These results are in full agreement with the charge distribution in 1-fluoropyrene. In the highly symmetrical pyrene molecule there are four equivalent positions (1, 3, 6, and 8), and introduction of a hydroxyl group at any of these positions yields the same product: 1-hydroxypyrene. In 1-fluoropyrene, however, the 1-position is effectively blocked. The three positions that remain available for substitution are no longer equivalent. The isopod yielded all three possible Phase-I metabolites upon exposure to 1-fluoropyrene, in ratios similar to those formed by nitration (e.g., 48:42:10 for sulfate). This indicates that the Phase-I enzyme cytochrome P450 in the isopod is not very regioselective and that the metabolite profile depends mainly on electron density. There were no significant differences in the isomer ratios of the unknown fluoropyrene conjugates, the fluoropyrene sulfates and fluoropyrene glucuronides. Apparently, the Phase-II enzymes in the isopod, i.e., sulfo transferase and glucoside transferase, are not regioselective with regard to the position of the fluorine substituent relative to the hydroxy group. In flounder the product ratio was rather different (75:25:0 for sulfate). No 1,3isomers were observed, and the 1,6-isomers were formed in higher yields than the 1,8-isomers; which is consistent with the electron density. Again, there were no differences in the isomer ratios between the different Phase-II conjugates. As was found for the isopod, sulfo transferase and glucuronide transferase in the flounder are not regioselective. The present results are important for understanding the different metabolic pathways of diverse animal species. They are also relevant for environmental risk assessment when the same biomarker (pyrene metabolites) is determined in different species (for instance vertebrates vs invertebrates). The comparable physical properties of PAHs and F-PAHs, the small size of the fluorine substituent, and the strength of the C–F bond make F-PAHs excellent tools for investigating PAH metabolism. This aspect is discussed in greater detail by Cavalieri and Rogan elsewhere in this volume. The presence of fluorine offers the opportunity of taking advantage of 19F NMR in identifying metabolites (see Stanley, this volume). 5.2 Flash Vacuum Thermolysis of F-PAHs

PAHs that are found in the atmosphere derive mainly from anthropogenic combustion of fossil fuels, domestic heating, refuse incineration, coke production, and traffic exhaust. Natural sources such as forest fires and volcanic activity constitute about 15% of the total emission [20]. Mechanistic studies of thermal synthesis and interconversion during combustion are important in understanding the types of PAHs that may be formed [45]. The use of flash vacuum thermolysis (FVT) [9, 48], that is, gas-phase pyrolysis with a short contact time in the hot zone, makes it possible to study these mechanisms. Depending on the precursors and the temperature, different PAH product mixtures with variable complexity can be obtained. We demonstrated

Monofluorinated Polycyclic Aromatic Hydrocarbons

271

that monofluorinated PAHs (F-PAHs) provide an interesting alternative to 13Clabeled or deuterated precursors to study reaction pathways. The strength of the C–F bond and the comparable size of hydrogen and fluorine make F-PAHs interesting model compounds for FVT studies. Furthermore, 19F-NMR can provide additional information on the structure of the products obtained by pyrolysis. PAHs with five or more fused rings can be formed during high-temperature combustion and are considered hazardous as inducers of cancer in mammals [23, 24]. Following the experience with the parent acenaphthylene and fluoranthene series in FVT [8], we selected 1-fluoro-5-ethynylnaphthalene (7) as precursor to study the temperature profile of formation of fused cyclopentane rings, and the possibility of F–scrambling at temperatures of 800–1200 °C. Although the conversion of 1-substituted-5-ethynylnaphthalene to a 1-substituted acenaphthylene is a textbook example, the conversion pathway is still unresolved. Figure 10 shows the reaction mechanism of the conversion of 1-fluoro-5-(1chloroethene)-naphthalene (1) to 5-fluoroacenaphthylene (2) via 1-fluoro-5ethynylnaphthalene at T>800 °C and the subsequent conversion into 4-fluoro-

Fig. 10. Suggested pathway for the formation of 5-fluoroacenaphthylene (2) and its rearrange-

ment by 1,2-H shifts and 1,2-C shifts to different isomers under high-temperature conditions

272

G.M. Luthe et al.

Fig. 11. Relative product yield during flash-vacuum thermolysis of 1-ethynyl-4-fluoronaphthalene to 5-fluoroacenaphthylene and its rearrangement products as a function of temperature: 5-fluoroacenaphthylene (2), 3-fluoroacenaphthylene (4), 4-fluoroacenaphthylene (5), acenaphthylene (6) and 1-ethynyl-4-fluoronaphthalene (7)

(5) and 3-fluoroacenaphthylene (4) at T>850 °C up to 1200 °C. Identification of the pyrolysates was done using 19F-NMR and quantification with GC-MS. After initial elimination of hydrogen chloride from the precursor (1), the resulting 1fluoro-5-ethynylnaphthalene is quantitatively converted at 850 °C by cyclization and a hydrogen shift to 5-fluoroacenaphthylene (2).At even higher temperatures, mixtures of isomers and non-substituted acenaphthylene (6) were obtained. The relative product yields as a function of temperature are shown in Fig. 11. Apart from coal-forming combustion processes at T>850 °C, the substitution of fluorine by hydrogen is the predominant process. Free radicals of hydrogen are available from the increasing coal formation process. The formation of isomers by the addition of fluorine radicals can be excluded, since otherwise higher fluorinated isomers would also have been obtained. We explain the isomerization by unimolecular rearrangements with a combination of a 1,2 H-shift followed by a 1,2 C-shift, as shown in Fig. 10. Conclusive proof for the absence of a pathway involving scrambling could be obtained using 13C-labeled fluorinated analogues. Although fluorine-substituted precursors provide a powerful tool for investigating unknown pathways in the formation of PAHs during pyrolytic processes, the somewhat different electronic density of F-PAHs and PAHs should be fully appreciated. Future investigations carried out with F-PAH/PAH pairs in FVT should provide more information on these important issues.

6 Conclusions In this chapter we have demonstrated that F-PAHs are useful as internal standards for a wide range of applications.An additional important advantage is that there are no known natural sources of these compounds. Due to the comparable physico-chemical properties of F-PAHs and their parent PAHs, F-PAHs have the potential for use as surrogate standards to monitor recoveries during extraction, clean-up and other sample-handling steps in PAH analysis. The chromatographic

Monofluorinated Polycyclic Aromatic Hydrocarbons

273

behavior of monofluorinated PAHs in LC and GC and that of polyfluorinated PAHs in LC can be explained on the basis of dipole moments and London forces. The suitability of F-PAHs as internal standards in GC-MS and in Shpol’skii spectroscopy was demonstrated. Fluoro-substituted PAHs are also useful for the elucidation of reaction pathways, both in the laboratory and in biota; the fluorinated metabolites may be used as biomarkers in a variety of applications. We are currently working on the development of new classes of F-containing compounds, such as fluorinated PAH-DNA adducts for quality control in the monitoring of DNA damage. Acknowledgements. Many people have contributed to these studies with technical and analyt-

ical support, by carrying out long series of measurements, and with creative and stimulating ideas. We acknowledge Jan Scharp, Jan T. Andersson, Viktor Platonov, Steven J. Kok, Hadil EsSbai, Jolanda Broeders, Lourdes Ramos, Gerard Ph. Hoornweg, Jens Dallüge, Gerard J. Stroomberg, Johan Jol, Reyer J. Dijkstra, Henk Lingeman, Cees Gooijer, Dik van Iperen,Wilhelm Hesselinck, Ma Lan, Wim de Wolf, Frans J.J. de Kanter, Leo Jenneskens, Ulfert E. Wiersum, Jan Beens, René Vreuls, Mohamed Adahchour, and E. Maria Kristenson.

7 References 1. Anders MW (1982) Metabolic basics of detoxification. Academic Press, New York 2. Andersson JT, Weis U (1994) Gas chromatographic determination of polycyclic aromatic compounds with fluorinated analogs as internal standards. J Chromatogr A 659:151–161 3. Ariese F, Kok SJ, Hoornweg GP, Gooijer C, Velthorst NH, Hofstraat JW, Verkaik M (1993) Chemical derivatization and Sphol’skii spectrofluorometric determination of benzo[a]pyrene metabolites in fish bile. Anal Chem 65:1100–1106 4. Ariese F, Gooijer C, Velthorst NH, Hofstraat JW (1990) Shpol’skii spectrofluorometric determination of polycyclic aromatic hydrocarbons in biota. Anal Chim Acta 232:245–251 5. Ariese F (1993) Shpol’skii spectroscopy and synchronous fluorescence spectroscopy; (bio)monitoring of polycyclic aromatic hydrocarbons and their metabolites. Ph.D. thesis, Free University, Amsterdam, The Netherlands 6. Banks RE (2000) Fluorine chemistry at the millennium. Fascinated by fluorine. Elsevier, Amsterdam, The Netherlands 7. Beller HR, Ding W-S, Reinhard M (1995) Byproducts of anaerobic alkylbenzene metabolism useful as indicators of in situ bioremediation. Environ Sci Technol 29:2864–2870 8. Brown RFC (1988) Thermal rearrangements of alkynes under FVP conditions – the acetylene methylenecarbene rearrangement. Recl Trav Chim Pays-Bas 107:655–661 9. Brown RFC (1980) Pyrolytic methods in organic chemistry, applications of flow and flash vacuum pyrolytic techniques. Academic Press, New York 10. Cavalieri E, Rogan E (1998) Mechanisms of tumor initiation by polycyclic aromatic hydrocarbons in mammals. In: Neilson AH, Hutzinger O (eds) PAHs and related compounds. Handbook of environmental chemistry, vol 3J. Springer, Berlin Heidelberg New York, pp 82–117 11. Cnubben NHP,Vervoort J (1995) The effect of varying halogen substituent patterns on the cytochrome P450 catalyzed dehalogenation of 4-halogenated anilines to 4-aminophenol metabolites. Biochem Pharmacol 49:1235–1248 12. Colmsjö A (1998) Concentration and extraction of PAHs from environmental samples. In: Neilson AH, Hutzinger (eds) PAHs and related compounds. Handbook of environmental chemistry, vol 3I. Springer, Berlin Heidelberg New York, pp. 56–76 13. De Knecht JA, Stroomberg GJ, Tump C, Helms M, Verweij RA, Commandeur J, van Gestel CAM, van Straalen NM (2001) Characterisation of enzymes involved in biotransformation of polycyclic aromatic hydrocarbons in terrestrial isopods. Environ Toxicol Chem (in press)

274

G.M. Luthe et al.

14. De Maagt PG-J, Vethaak AD (1998) Biotransformation of PAHs and their carcinogenic effects in fish. In: Neilson AH, Hutzinger O (eds) PAHs and related compounds. Handbook of environmental chemistry, vol 3J. Springer, Berlin Heidelberg New York, pp. 265–309 15. Escartin E, Porte C (1999) Hydroxylated PAHs in bile of deep-sea fish. Relationship with xenobiotic metabolizing enzymes. Environ Sci Technol 33:2710–2714 16. Garrigues P, Ewald M (1985) Application of high resolution Shpol’skii luminescence spectroscopy to the analysis of polycyclic aromatic hydrocarbons in environmental samples. Int J Environ Anal Chem 21:185–197 17. Gribble GW (1998) Naturally occurring organohalogen compounds. Acc Chem Res 31:141–152 18. Guengrich FP, Shimada T (1991) Oxidation of toxic and carcinogenic chemicals by human cytochrome P450 enzymes. Chem Res Toxicol 4:391–407 19. Hansen PE (1979) 13C NMR of polycyclic aromatic compounds.A review. Org Magn Reson 12:109–142 20. Harvey RG (1997) Polycyclic aromatic hydrocarbons. Wiley-VCH, New York 21. Harvey RG (1985) Polycyclic hydrocarbons and carcinogenesis. American Chemical Society, Washington DC 22. Hofstraat JW, Freriks IL, de Vreeze MEJ, Gooijer C, Velthorst NH (1989) Thermal history and concentration effects on Shpol’skii spectra, study of acenaphthene in normal hexane. J Phys Chem 93:184–190 23. Howard JB, Longwell JP, Marr JA, Pope CJ, Busby WF, Lafleur AL, Taghizadeh K (1995) Effects of PAH isomerizations on mutagenicity of combustion products. Combust Flame 101:262–270 24. Jacob J (1996) The significance of polycyclic aromatic hydrocarbons as environmental carcinogens. Pure Appl Chem 68:301–308 25. Jankowiak R, Small GJ (1998) Analysis of PAH-DNA adducts – fluorescence line-narrowing spectroscopy. In: AH Neilson, Hutzinger (eds) PAHs and related compounds. Handbook of environmental chemistry, vol 3J. Springer, Berlin Heidelberg New York, pp 120–145 26. Koerts J, Soffers AEMF, Vervoort J, De Jager A, Rietjens IMCM (1998) Occurrence of the NIH shift upon the cytochrome P450-catalyzed in vivo and in vitro aromatic ring hydroxylation of fluorobenzenes. Chem Res Toxicol 11:503–512 27. Kootstra PR, Straub MHC, Stil GH, Vandervelde EG, Hesselink W, Land CCJ (1995) Solid phase extraction of polycyclic aromatic hydrocarbons from soil samples. J Chromatogr A 697:123–129 28. Krahn MM, Burrows DG, MacLeod WD Jr, Malins DC (1987) Determination of individual metabolites of aromatic compounds in hydrolyzed bile of English sole (Parophrys vetulus) from polluted sited in Puget sound,Washington.Arch Environ Contam Toxicol 16:511–522 29. Levin JO (1995) First international workshop on hydroxypyrene as a biomarker for PAH exposure in man – summary and conclusions. Sci Total Environ 163:165–168 30. Li X-F, Le X-C, Simpson CD, Cullen WR, Reimer KJ (1996) Bacterial transformation of pyrene in a marine environment. Environ Sci Technol 30:1115–1119 31. Livingstone DR (1998) The fate of organic xenobiotics in aquatic ecosystems: quantitative and qualitative differences in biotransformation by invertebrates and fish. Comp Biochem Physiol Part A 120:43–49 32. Luthe G, Brinkman UAT (2000) Monofluorinated polycyclic aromatic hydrocarbons: characteristics and intended use in environmental analysis. Analyst 125:1699–1702 33. Luthe G, Brinkman UAT (2002) Investigations into the retention behaviour of fluorinated analogues of polycyclic aromatic hydrocarbons in reverse phase liquid chromatography. In preparation 34. Luthe G, Stroomberg GJ, Ariese F, Brinkman UAT, van Straalen NM (2002) In vivo metabolism of 1-Fluoro-pyrene compared with that of pyrene in the marine flatfish Platichthys flesus and the terrestrial isopod Porcellio scaber (submitted) 35. Luthe G, Broeders J, Gooijer C, Brinkman UAT (2001) Monofluorinated polycyclic aromatic hydrocarbons as internal standards to monitor trace enrichment and desorption of their parent compounds during solid-phase extraction. Chromatogr A 933:27–32

Monofluorinated Polycyclic Aromatic Hydrocarbons

275

36. Luthe G, Ramos L, Dallüge J, Ariese F, Vreuls JJ, Brinkman UAT (2001) Mono-fluorinated analogues of polycyclic aromatic hydrocarbons as internal standards in GC-MS analysis. In preparation 37. Luthe G, Es-bai H, Ariese F, Brinkman UAT, Gooijer C (2002) Monofluorinated polycyclic aromatic hydrocarbons as internal standards in Shpol’skii spectroscopy. Anal Chim Acta, in press 38. Luthe G, Scharp J, Brinkman UAT, Gooijer C (2001) Monofluorinated polycyclic aromatic hydrocarbons in Shpol’skii spectroscopy Anal Chim Acta 429:49–54 39. Mastenbroek JWG, Ariese F, Gooijer C, Velthorst NH, Hofstraat JW, van Zeijl JWM (1990) Shpol’skii fluorometry as an independent identification method to upgrade routine HPLC analysis of polycyclic aromatic hydrocarbons. Chemosphere 21:377–386 40. May WE,Wise SA (1984) Liquid chromatographic determination of polycyclic aromatic hydrocarbons in air particulate extracts. Anal Chem 56:225–232 41. Nakhimovsky L, Lamotte M, Joussot-Dubien J (1989) Handbook of low temperature electronic spectra of polycyclic aromatic hydrocarbons. Elsevier, Amsterdam 42. Poster DL, Sander LC, Wise SA (1998) Chromatographic methods of analysis for the determination of PAHs in environmental samples. In: Neilson AH, Hutzinger O (eds) PAHs and related compounds. Handbook of environmental chemistry, vol 3I. Springer, Berlin Heidelberg New York, pp. 78–135 43. Ramanathan R, Gross ML (1998) Mass spectrometry techniques: DNA adducts of PAHs and related carcinogens. In: Neilson AH, Hutzinger O (eds) Handbook of environmental chemistry, vol 3J. Springer, Berlin Heidelberg New York, pp. 148–202 44. Renge I,Wild UP (2000) Principles of matrix-induced high-resolution optical spectroscopy and electron-phonon coupling in doped organic glasses. In: Gooijer C, Ariese F, Hofstraat JW (eds) Shpol’skii spectroscopy and other site selection methods. Wiley, New York, NY, USA, pp 19–71 45. Sarobe M, Jenneskens LW, Wesseling J, Snoeijer JD, Zwikker JW, Wiersum UE (1997) Thermal interconversions of the C16H10 cyclopentafused polycyclic aromatic hydrocarbons fluoranthene, acephenanthrylene and aceanthrylene. Liebigs Ann/Recueil 1207–1213 46. Snyder MJ (2000) Cytochrome P450 enzymes in aquatic invertebrates, recent advances and future directions. Aquat Toxicol 48:529–547 47. Stroomberg GJ, De Knecht JA, Ariese F, van Gestel CAM, Velthorst NH (1999) Pyrene metabolites in the hepatopancreas and gut of the isopod Porcellio scaber, a new biomarker for polycyclic aromatic hydrocarbon exposure in terrestrial ecosystems. Environ Toxicol Chem 18:2217–2224 48. Vallée Y (2000) Gas phase reactions in organic synthesis. Gordon and Breach, Sydney,Australia 49. Van Brummelen TC, van Hattum B, Crommentuijn T, Kalf DF (1988) Bioavailability and ecotoxicity of PAHs. In: Neilson AH Hutzinger O (eds) PAHs and related compounds. Handbook of environmental chemistry, vol 3J. Springer, Berlin Heidelberg New York, pp 203–263 50. Wise SA, Sander LC, May WE (1993) Determination of polycyclic aromatic hydrocarbons by liquid chromatography. J Chromatogr 642:329–349 51. Wise SA, Bonnett WJ, Franklin R, Guenther R, May E (1985) A relationship between reversed phase C18 liquid chromatographic retention and the shape of polycyclic aromatic hydrocarbons. J Chromatogr Sci 19:457–465 52. Wise SA, Schantz MM, Benner BA, Hays MJ, Schiller SB (1995) Certification of polycyclic aromatic hydrocarbons in a marine sediment standard reference material. Anal Chem 67:1171–1178 53. Yang Y, D’Silva AP, Fassel VA (1981) Deuterated analogues as internal reference compounds for the direct determination of benzo[a]pyrene and perylene in liquid fuels by laser-excited Shpol’skii spectrometry. Anal Chem 53:2107–2109

CHAPTER 9

Fluoro Substitution of Carcinogenic Aromatic Hydrocarbons: Models for Understanding Mechanisms of Metabolic Activation and of Oxygen Transfer Catalyzed by Cytochrome P450 Ercole L. Cavalieri · Eleanor G. Rogan Eppley Institute for Research in Cancer and Allied Diseases, University of Nebraska Medical Center, 986805 Nebraska Medical Center, Omaha, NE 68198–6805

Fluoro substitution of polycyclic aromatic hydrocarbons (PAHs), which only slightly affects their steric and electronic properties, has served two main purposes. One has been to gain evidence about the mechanism of metabolic activation leading to tumor initiation by PAHs. Studies with fluoro-substituted 5-methylchrysenes were useful in identifying the 1,2-dihydrodiol-3,4-epoxide as the ultimate carcinogenic form of this compound. In contrast, studies with fluoro-substituted 7,12-dimethylbenz[a]anthracenes or benzo[a]pyrenes (BPs) have not contributed much to understanding the mechanisms of activation for these two carcinogens. Specific fluoro-substituted BPs have, however, served as models to understand the mechanism by which the activated oxygen of cytochrome P450 is transferred to form BP metabolites. Using this knowledge, a unifying mechanism is proposed by which cytochrome P450 transfers its activated oxygen species to various substrates. For substrates containing nonbonded or p electrons, the electrophilic perferryl oxygen of cytochrome P450 is converted to a nucleophilic oxygen by abstraction of one electron from the substrate. On the other hand, the cytochrome P450 nucleophilic oxygen species for substrates having only s electrons is presumably an oxygen radical species capable of abstracting a hydrogen atom from the substrate to form a carbon radical that leads to carbon hydroxylation. Thus, the common mechanism in cytochrome P450-catalyzed oxygen transfer involves a nucleophilic oxygen species for all substrates. Keywords. Cytochrome P450, Fluorinated aromatic hydrocarbons, Mechanism of metabolic ac-

tivation, Mechanism of oxygen transfer

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

1

Introduction

2

Fluoro Substitution of PAHs as a Guideline to Determine the Mechanism of Metabolic Activation . . . . . . . . . . . . . . . 279

2.1 2.2 2.3

5-Methylchrysene . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 7,12-Dimethylbenz[a]anthracene . . . . . . . . . . . . . . . . . . 280 Benzo[a]pyrene . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282

3

Fluorobenzo[a]pyrenes as Models to Determine the Mechanism of Oxygen Transfer Catalyzed by Cytochrome P450 . . . . . . . . 283

4

A Unifying Mechanism of Oxygen Transfer from Cytochrome P450 to Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . 289

5

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290 The Handbook of Environmental Chemistry Vol. 3, Part N Organofluorines (ed. by A.H. Neilson) © Springer-Verlag Berlin Heidelberg 2002

278

E.L. Cavalieri · E.G. Rogan

List of Symbols and Abbreviations Ade BP DMBA Gua H 5-MeC OHBP PAH

adenine benzo[a]pyrene 7,12-dimethylbenz[a]anthracene guanine Harvey 5-methylchrysene hydroxybenzo[a]pyrene polycyclic aromatic hydrocarbon.

1 Introduction Polycyclic aromatic hydrocarbons (PAHs) are activated by two main pathways: one-electron oxidation to produce radical cations and monooxygenation to afford bay-region diol epoxides [14 ,16]. These two types of reactive electrophilic intermediates initiate the process leading to tumorigenesis by covalently binding to DNA to form two types of DNA adducts: stable ones that remain in DNA unless removed by repair and depurinating ones that are released from DNA by destabilization of the glycosyl bond [14, 16]. Stable adducts are formed when carcinogens react with the exocyclic N6 amino group of adenine (Ade) or N2 amino group of guanine (Gua), whereas depurinating adducts are obtained when carcinogens covalently bind at the N-3 or N-7 of Ade or the N-7 or sometimes C-8 of Gua. Loss of Ade or Gua by depurination leads to formation of apurinic sites. Identification and quantification of PAH-DNA adducts led us to discover that there is a correlation between depurinating adducts and oncogenic mutions, suggesting that these adducts are the primary culprits in tumor initiation by these compounds [16–18]. This discovery was made by determining the DNA adducts formed in mouse skin by the three potent carcinogens dibenzo[a,l]pyrene, 7,12-dimethylbenz[a]anthracene (DMBA) and benzo[a]pyrene (BP), and, at the same time, determining the mutations in the Harvey (H)-ras oncogene in mouse skin papillomas initiated by these three PAHs [17]. Fluoro substitution of a PAH, which only slightly affects its steric and electronic properties, has served two main purposes. One has been to gain evidence about the mechanism of metabolic activation leading to tumor initiation. In the second objective, fluoro-substituted BPs have been used as models to understand the mechanism by which the activated oxygen of cytochrome P450 forms the various BP metabolites. On the basis of this knowledge, a unifying mechanism by which cytochrome P450 transfers the active oxygen species to various substrates is proposed.

Fluoro Substitution of Carcinogenic Aromatic Hydrocarbons

279

2 Fluoro Substitution of PAHs as a Guideline to Determine the Mechanism of Metabolic Activation 2.1 5-Methylchrysene

Substitution of chrysene with a methyl group produces five methylated isomers with weak carcinogenic activity similar to that of the parent compound, whereas substitution of the methyl group at C-5 gives rise to the potent carcinogen 5methylchrysene (5-MeC, Fig. 1) [32, 33].All experimental data concerning tumor initiation by 5-MeC are consistent with activation by the diol epoxide pathway [34–36, 54, 57]. This PAH has too high an ionization potential and insufficient charge localization in its radical cation to be activated by one-electron oxidation [14, 16]. To study the activation of 5-MeC by the diol epoxide pathway, a series of fluorinated 5-MeC derivatives was synthesized and tested for mutagenicity, metabolism and tumorigenicity in mouse skin [34–36]. Studies of tumor-initiating activity and complete carcinogenicity in mouse skin showed that the 1-F- and 3-Fderivatives had borderline activity, and the 12-F-5-MeC displayed lesser tumorinitiating activity and carcinogenicity than the parent compound, 5-MeC. The 6F-, 7-F-, 9-F- and 11-F-5-MeC had activity similar to that of 5-MeC. These results suggest that the important metabolic pathway leading to tumor initiation involves formation of a diol epoxide in the C-1–C-4 ring, adjacent to the 5-methyl group. The interpretation of these results is supported by metabolic studies of these fluorinated compounds conducted with induced rat liver microsomes [35, 36]. 5MeC, 6-F-5-MeC and 7-F-5-MeC produced among other metabolites the 1,2-dihydrodiol that constitutes the first step in the formation of the critical bay-region diol epoxide. Virtually no 1,2-dihydrodiol was obtained in the metabolism of 1-

Fig. 1. Structures of three carcinogenic PAHs

280

E.L. Cavalieri · E.G. Rogan

F-, 3-F- and 12-F-5-MeC. This metabolite is not formed from the 1-F- and 3-F- derivatives because fluorination deactivates epoxidation in the C-1–C-4 ring. The limited amount of 1,2-dihydrodiol obtained by metabolism of 12-F-5-MeC presumably derives from restrictions imposed by the binding site of the enzyme. Thus, the metabolic pathway of activation of 5-MeC that involves epoxidation at C-1–C-2, followed by epoxide hydratase-catalyzed formation of the 1,2-dihydrodiol and further epoxidation at C-3–C-4, with formation of the ultimate carcinogenic diol epoxide, 5-MeC-1,2-dihydrodiol-3,4-epoxide, is consistent with the tumorigenicity and metabolism data reported. The major identified stable DNA adduct of 5-MeC contains the diol epoxide bound at the 2-amino group of Gua [54, 57], implying that formation of this adduct initiates tumors by 5-MeC. Both the syn- and anti-diol epoxides form depurinating DNA adducts of Ade and Gua [26]. To establish the adducts that are relevant in the process of tumor initiation in mouse skin, it will be necessary to identify and quantify the stable and depurinating adducts and to attempt to correlate the profile of these adducts with the H-ras mutations found in mouse skin papillomas induced by 5-MeC [37]. 2.2 7,12-Dimethylbenz[a]anthracene

Numerous studies conducted on the potent carcinogen DMBA (Fig. 1) support activation by one-electron oxidation, with formation of the specifically reactive carbenium ion at the 12-methyl group, and by formation of the bay-region diol epoxide (Fig. 2) [14, 16]. Fluoro substitution was also used as an approach to identify a specific pathway of activation for DMBA [13, 27, 28, 31, 39]. In mouse skin and rat mammary gland, the tumor-initiating activity and carcinogenicity of 9-F-, 10-F- and 11-FDMBA were similar to those of the parent compound, DMBA [13, 28, 39]. The 1-

Fig. 2. Metabolic activation of DMBA by the one-electron oxidation and diol epoxide pathways

Fluoro Substitution of Carcinogenic Aromatic Hydrocarbons

281

F-, 2-F-, 4-F- and 5-F- DMBA were much weaker tumor initiators and weaker as complete carcinogens [13, 27, 31, 39]. These results could suggest that the bay-region diol epoxide pathway has the most impact in the metabolic activation of DMBA. This is also supported by the greater tumor-initiating activity of DMBA3,4-dihydrodiol compared to DMBA [60]. Confirmation of this interpretation, however, requires the results from metabolism of the fluorinated DMBA derivatives, which has never been conducted due to the complexity in separating and identifying the numerous metabolites of DMBA [20]. There are a series of biological experiments with fluorinated DMBAs that conflict with the above interpretation. One is that the 1,2,3,4-tetrahydroDMBA is a strong carcinogen both in mouse skin and rat mammary gland [13, 27], despite being fully saturated in the angular ring and, thus, unable to be activated by the diol epoxide pathway. When reacted with dG and dA, the radical cation of this compound produces numerous adducts at the 12-methyl and 7-methyl groups [49], which are similar to those formed by DMBA itself [55]. Both in vitro and in vivo, DMBA binds specifically at the 12-methyl group to the DNA bases Ade and Gua to form predominantly depurinating adducts [25, 56].When the two methyl groups of the potent carcinogen DMBA were substituted with two ethyl groups, the resulting 7,12-diethylbenz[a]anthracene was not carcinogenic [53]. The inactivity of the ethyl-substituted compound is consistent with the lack of nucleophilic substitution at the benzylic methylene group of the radical cation of an ethyl PAH [9, 63]. Furthermore, 7-methyl-12-ethylbenz[a]anthracene is a much weaker carcinogen than DMBA, whereas 7-ethyl-12-methylbenz[a]anthracene shows a carcinogenicity similar to that of DMBA [53]. The involvement of the 12methyl group clearly indicates that one-electron oxidation plays a major role in the metabolic activation of DMBA. DMBA-3,4-dihydrodiol, precursor to the bay-region diol epoxide, has been found to be among the numerous metabolites of DMBA [20] and is a more potent tumor initiator in mouse skin than the parent DMBA [60]. The higher tumorigenicity of DMBA 3,4-dihydrodiol can be attributed not only to formation of the bay-region diol epoxide, but also to formation of DMBA 3,4-dihydrodiol radical cation because the chemistry of this intermediate is very similar to that of the well-investigated 1,2,3,4-tetrahydroDMBA [49]. Therefore, to determine the extent that the two pathways are involved in the metabolic activation of DMBA, it will be necessary to determine not only the depurinating adducts obtained from the radical cation of DMBA [25, 56], but also the depurinating adducts obtained from the DMBA 3,4-dihydrodiol radical cation and the DMBA diol epoxides. Identification and quantification of the stable and depurinating adducts of DMBA in the target organ mouse skin will provide the most reliable assessment of the extent that the two pathways contribute to tumor initiation by DMBA. In summary, studies with monofluorinated DMBAs have contributed very little to understanding the mechanism of metabolic activation of DMBA in tumor initiation because this PAH is activated by two distinct pathways.

282

E.L. Cavalieri · E.G. Rogan

2.3 Benzo[a]pyrene

Fluorination at the 6-, 7-, 8- and 10- position of BP (Fig. 1) leads to reduced tumor-initiating activity in mouse skin and carcinogenicity in rat mammary gland compared to the activity of BP [12]. Only 9-FBP is inactive in both models. Therefore, the lack of activity by 9-FBP can be explained by blockage of the formation of the bay-region diol epoxide. In fact, metabolism of 9-FBP shows a metabolic profile similar to that of the parent compound BP with the exception of the formation of the 9,10-dihydrodiol [21]. The weak-to-moderate tumorigenicity of 7-, 8- and 10-FBP, which show metabolic profiles similar to that of BP with the exception of not forming the dihydrodiol at the double bond in which the fluoro group is substituted [4, 21], cannot be explained by the bay-region diol epoxide pathway of activation. Analogously to DMBA, BP has been hypothesized to initiate tumors via formation of its radical cation, with specific reaction at C-6, as well as formation of the bay region diol epoxide (Fig. 3) [14, 16]. Monofluorination has not provided important clues for delineating the relative importance of the two mechanisms of activation. Instead, comprehensive analysis of the depurinating adducts obtained from the radical cation and bay-region diol epoxide of BP [19], along with determination of the H-ras mutations in tumors induced by BP [17], led us to discover that the mutations are related to the DNA base that depurinates as a BP adduct. For example, 71% of the adducts formed by BP in mouse skin are depurinating adducts, with 66% formed by one-electron oxidation [19]. Of the 29% of stable adducts, 23% are BP diol epoxide bound at C-10 to the 2-amino group of Gua.Among the depurinating adducts, 46% contain Gua and 25% have Ade. Fifty percent of the mouse skin papillomas induced by BP have G to T mutations at codon 13 and 20% have A to T mutations at codon 61 of the H-ras gene [16, 17]. This ratio mimics the distribution of depurinating adducts among Gua and Ade

Fig. 3. Metabolic activation of BP by the one-electron oxidation and diol epoxide pathways

Fluoro Substitution of Carcinogenic Aromatic Hydrocarbons

283

residues, confirming the relationship between depurinating adducts and oncogenic H-ras mutations. In conclusion, monofluorination has not played an important role in determining the mechanism of tumor initiation by DMBA or BP. Monofluorinated derivatives of 5-MeC have, however, provided the clue that the critical diol epoxide initiating the tumor process occurs in the ring adjacent to the 5-methyl group. These findings have put research on the right track to substantiate the role of the 1,2-dihydrodiol-3,4-epoxide in the activation of this compound.

3 Fluorobenzo[a]pyrenes as Models to Determine the Mechanism of Oxygen Transfer Catalyzed by Cytochrome P450 The condensation of five benzene rings to form BP produces twelve carbon atoms with different electronic densities and double bonds displaying some ethylenic character. The position of BP having the greatest reactivity with electrophiles is C-6, followed by C-1 and C-3 [6, 7]. In BP+ ∑ it is again C-6 that displays by far the major reactivity with nucleophiles [3, 5, 8, 22, 40, 41, 48, 58, 61, 64, 65], and C1 and C-3 follow in decreasing order [22]. Furthermore, experimental data indicate that the spin density in the high occupied molecular orbitals of BP+ ∑, which follows the same pattern as the charge density, is greater at C-6, followed by C-1 and C-3 [62]. These experiments are corroborated by molecular orbital calculations in the neutral molecule BP [46, 59] and BP+ ∑ [15, 50]. The overall data suggest that electrophilic and nucleophilic substitution in BP and BP+ ∑, respectively, should occur most easily at positions 6, 1 and 3 in decreasing order. Calculations of the bond orders of BP, 6-FBP, 1-FBP and 3-FBP show the highest electron density in the 4,5-double bond, followed by the 11,12, the 9,10 and the 7,8 [15, 50]. Two classes of primary metabolites are obtained when BP is converted by cytochrome P450, BP phenols and BP dihydrodiols (Fig. 4) [23, 38, 66]. The major phenol obtained is 6-hydroxyBP (6-OHBP) [42, 45, 51], not isolated because it is autoxidized to BP 1,6-, 3,6- and 6,12-quinone. The other major BP phenol is 3OHBP, whereas the minor phenols formed are 1-OHBP, 7-OHBP and 9-OHBP. The very small yield of 1-OHBP obtained is presumably due to restrictions imposed by the binding site of the enzyme [1]. The second class of metabolites is constituted by three dihydrodiols, the BP 4,5-, BP 7,8- and BP 9,10-dihydrodiol. The dihydrodiols are obtained via two enzymatic steps, namely, formation of BP epoxides, followed by their hydrolysis catalyzed by epoxide hydratase to form the dihydrodiols. The metabolic formation of dihydrodiols follows the same order as the calculated bond orders (see above) with the exception of the 11,12-dihydrodiol, which has never been found in the metabolism of BP, presumably impeded by the binding site of the enzyme. Fluorination of BP at C-6, C-1 and C-3 has generated excellent probes for determining the mechanism by which the active oxygen of cytochrome P450 is transferred from the enzyme to a substrate [11, 50]. Combined studies of the chemistry of BP and 6-FBP, as well as the metabolism and molecular orbital calculations of BP, 6-FBP, 1-FBP and 3-FBP, have enabled us to determine the mech-

284

E.L. Cavalieri · E.G. Rogan

Fig. 4. Cytochrome P450-catalyzed metabolism of BP

anism of oxygen transfer from cytochrome P450 to BP in the formation of 6OHBP, 3-OHBP and 1-OHBP [11, 50]. Metabolism of 6-FBP by cytochrome P450 [4, 11], horseradish peroxidase [10] or prostaglandin H synthase [10] affords the same quinone metabolites obtained from BP [4, 10, 11]. Displacement of the fluoro substituent from 6-FBP+ ∑ can occur only after attack by the nucleophilic oxygen of the enzyme. In fact, nucleophilic acetoxy substitution of 6-FBP+ ∑ obtained by oxidation of 6-FBP with manganic acetate occurs exclusively at C-6, with displacement of the fluoro ion [22]. In addition, reaction of the isolated 6-FBP+ ∑ perchlorate with H2O yields the BP 1,6-, 3,6- and 6,12- diones, as does BP+ ∑ under the same conditions [11]. In contrast, reaction of 6-FBP with the electrophile CF3COOD or pyridinium bromide perbromide affords 1,3-dideuteriated 6-FBP or a mixture of 1-Br-6-FBP and 3-Br-6-FBP, respectively, with retention of fluorine at C-6 [22]. Calculations of the electron and charge densities in 6-FBP and 6-FBP+ ∑ resemble those of BP and BP+ ∑, respectively, with the exception of C-6, which is positively charged in 6-FBP and even more so in 6-FBP+ ∑, due to the electron-withdrawing effect of the fluoro substituent [50]. These values are in agreement with the electrophilic chemistry of the neutral 6-FBP and the nucleophilic chemistry of the 6-FBP+ ∑. As previously suggested [30, 47], the remarkable reactivity of the activated oxygen of cytochrome P450 is attributed to an electrophilic oxo-Fe4+ porphyrin radical cation. The first step in the metabolism of 6-FBP to produce the 1,6-, 3,6-

Fluoro Substitution of Carcinogenic Aromatic Hydrocarbons

285

and 6,12-dione is the transfer of one electron to [Fe4+=O]+ ∑, with formation of 6-FBP+ ∑ and Fe4+=O (Fig. 5). The nucleophilic oxygen of the Fe4+=O reacts at C-6 of the 6-FBP+ ∑, with subsequent removal of the fluoro ion (Fig. 5, left). Then, the oxo-BP radical is oxidized by Fe4+ with formation of an oxo-BP cation. Alternatively, if the radical species of 6-FBP+ ∑ is localized at C-6 (Fig. 5, right), it reacts with Fe4+=O, with subsequent removal of the fluoro ion to form the oxo-BP cation. Nucleophilic attack of H2O on this reactive species affords 1,6-, 3,6- and 6,12-dihydroxyBP, which yield the corresponding quinones by autoxidation. Thus, removal of the fluoro substituent from C-6 can occur only via formation of

Fig. 5. Mechanism of oxygen transfer from cytochrome P450 to 6-FBP to form BP 3,6-dione.

The similar formation of BP 1,6- and 6,12-dione is not shown

Fig. 6. Proposed mechanism for the formation of BP 7,8-oxide

286 E.L. Cavalieri · E.G. Rogan

Fluoro Substitution of Carcinogenic Aromatic Hydrocarbons

287

an intermediate radical cation. This demonstrates that the metabolite 6-OHBP (and the three quinones formed by its autoxidation) derives from BP+ ∑, obtained by abstraction of a p electron from BP by the electrophilic [Fe4+=O]+ ∑ of cytochrome P450. Calculations of the electron densities in the neutral molecule 1-FBP show that the various carbon atoms have an electronic distribution similar to that of BP, with the exception of C-1, which is much more positively charged due to substitution of the electronegative fluoro group [50]. In its radical cation, C-1 becomes the center with the highest charge density, followed by C-6 and C-3.Analogously, in the 3-FBP neutral molecule, the most positively charged center is at C-3; in 3-FBP+ ∑, the highest charge density is at C-3, followed by C-6 and C-1 [50]. Formation of BP 1,6-dione from the metabolism of 1-FBP is obtained by removal of the fluoro ion from C-1 [50]. BP 3,6-dione is formed from 3-FBP by metabolic removal of the fluoro ion from C-3 [50]. This can occur only by an initial one-electron oxidation of the substrate, 1-FBP or 3-FBP, by the activated perferryl oxygen of cytochrome P450, with formation of the substrate radical cation. Analogously to 6-FBP+ ∑ (Fig. 5), attack of the nucleophilic oxygen of cytochrome P450 in 1-FBP+ ∑ and 3-FBP+ ∑ yields BP 1,6-dione and BP 3,6-dione, respectively. This proves that the abundant 3-OHBP metabolite of BP and the minor metabolite 1-OHBP derive from oxygenation of the BP+ ∑. Calculations of the bond orders of BP, 6-FBP, 3-FBP and 1-FBP show that the greatest electron density is in the 4,5-double bond, followed by the 11,12, the 7,8 and the 9,10 in decreasing order. The metabolic formation of dihydrodiols occurs at the same double bonds, with the exception of the 11,12, in which restriction is imposed by the enzyme binding site. These results lead us to hypothesize that the three dihydrodiols are obtained from abstraction of a p electron from one of these double bonds by the [Fe4+=O]+ ∑, as illustrated for the formation of BP-7,8dihydrodiol (Fig. 6). This is followed by a non-concerted oxygen rebound to the radical or cation of the double bond and subsequent closure, with formation of the epoxide. The hydrolysis of the epoxide, catalyzed by epoxide hydratase, leads to formation of BP 7,8-dihydrodiol. A hydride shift before the closure of the 7,8and 9,10-epoxide could presumably account for the formation of the small amount of 7-OHBP and 9-OHBP in the metabolism of BP. Therefore, all the major metabolites of BP are obtained by abstraction of an initial p electron from BP by the [Fe4+=O]+ ∑ of cytochrome P450, with formation of BP+ ∑ and Fe4+=O. This is followed by transfer of the nucleophilic oxygen of Fe4+=O to C-6, C-1 and C-3, the positions of greatest charge localization in BP+∑,with formation of 6-OHBP (and then the three quinones formed by its autoxidation), 3-OHBP and 1-OHBP respectively.Alternatively, if the BP+ ∑ is located on one of the double bonds, the BP 4,5-, 7,8- and 9,10-epoxides are formed. Addition of H2O catalyzed by epoxide hydratase produces the respective dihydrodiols.

288

E.L. Cavalieri · E.G. Rogan

Fig. 7. Catalytic cycle of cytochrome P450 illustrating the postulsted structures of the intermediates involved in activation of molecular oxygen and in oxygenation of substrates

Fluoro Substitution of Carcinogenic Aromatic Hydrocarbons

289

4 A Unifying Mechanism of Oxygen Transfer from Cytochrome P450 to Substrates Characterization of the critical intermediates in the mechanism of oxygen transfer from cytochrome P450 to specifically fluorinated BPs has provided the impetus to understand the oxygenation of a variety of substrates in a unified way. The overall reaction of cytochrome P450 monooxygenation is represented as follows: 2 e, 2 H+

RH +O2 æææÆ ROH +H2O where one oxygen atom of O2 oxidizes the substrate RH and the other is reduced to H2O. The stoichiometry of this reaction uses two protons and two electrons that are involved in the activation and cleavage of the O2 to yield the oxygenated substrate and H2O. The catalytic cycle in Figure 7 illustrates the mechanism of O2 activation and the mechanism of oxygen transfer from the hemoprotein cytochrome P450 to the substrate. The cycle begins with the low-spin six-coordinate Fe3+ in the resting phase (1), which has a cysteine thiolate and H2O as axial ligands. The binding of the substrate (RH) gives rise to the high-spin five-coordinate derivative (2), with a higher reduction potential than (1), to facilitate the electron transfer from cytochrome P450 reductase and yield the higher-spin five-coordinate Fe2+ (3). This intermediate binds O2 to form the ternary Fe2+/RH/O2 complex (not shown). Binding of O2 presumably produces a low-spin six-coordinate Fe3+ intermediate with a peroxy radical as the sixth ligand (4). One-electron reduction by cytochrome P450 reductase or cytochrome b5 yields the peroxy anion (5).Addition of two protons leads to the cleavage of the O–O bond to form H2O and the perferryl oxygen intermediate (6). This reactive species, formally an Fe5+, more likely corresponds to an Fe4+ porphyrin radical cation [24, 30, 47].An alternative pathway to produce directly the intermediate (6) from the high-spin Fe3+ (2), called the peroxide shunt, can occur in the presence of a cofactor peroxide that can donate an oxygen atom. The first step for substrates containing nonbonded or p electrons consists of the donation of one electron by the substrate to the perferryl oxygen radical cation (6) with formation of RH+ ∑. and a Fe4+=O (7). This is followed by attack of the nucleophilic oxygen atom of 7 on RH+ ∑ to yield 8, with subsequent formation of the oxygenated substrate and the Fe3+ porphyrin (9). When the substrate contains only s electrons such as in aliphatic hydroxylation, a mesomeric perferryl oxygen (10) is proposed, in which one electron of the Fe4+=O double bond neutralizes the radical cation and the other electron yields an oxygen radical. This oxygen intermediate, which is capable of abstracting a hydrogen atom from the substrate [29], can be considered nucleophilic [43]. Oxygen rebound from the Fe4+–OH intermediate (11) occurs with the carbon radical (R ∑) within a caged pair [2, 52] to yield the hydroxylated substrate and the Fe3+ porphyrin (9). Therefore, an electrophilic oxygen species (oxenoid mechanism) does not appear to have a role in cytochrome P450-mediated oxygenation [44].

290

E.L. Cavalieri · E.G. Rogan

In summary, for substrates containing nonbonded or p electrons the electrophilic perferryl oxygen species (6) is converted to a nucleophilic oxygen species (7) by abstraction of one electron from the substrate. On the other hand, the cytochrome P450 nucleophilic oxygen species for substrates having only s electrons would be a mesomeric form of 6, namely an oxygen radical species (10) capable of abstracting a hydrogen atom to form a carbon radical, leading to carbon hydroxylation. Thus, the common mechanism in cytochrome P450-catalyzed oxygen transfer involves a nucleophilic oxygen species for all substrates. Acknowledgements. We deeply thank Dr. Muhammad Saeed for valuable comments. Preparation of this article was supported by U.S. Public Health Service grants P01 CA49210 and R01 CA49917 from the National Cancer Institute. Core support to the Eppley Institute is provided by grant P30 CA36727 from the National Cancer Institute.

5 References 1. Alpert AJ, and Cavalieri EL (1980) Metabolism of 6-substituted benzo[a]pyrene derivatives: O-Dealkylation and regiospecificity in aromatic hydroxylations. J Med Chem 23: 919–927 2. Atkinson, JK, Hollenberg PF, Ingold KU, Johnson CC, LeTadic MH, Newcomb M, Putt DA (1994) Cytochrome P450-catalyzed hydroxylation of hydrocarbons: Kinetic deuterium isotope effects for the hydroxylation of an ultrafast radical clock. Biochemistry 33 : 10630–10637 3. Blackburn GM, Taussing PE,Will JP (1974) Binding of benzo[a]pyrene to DNA investigated by tritium displacement. J Chem Soc Chem Commun 907–908 4. Buhler DR, Unlü F, Thakker DR, Slaga TJ, Conney AH, Wood AW, Chang RL, Levin W, Jerina DM (1983) Effect of a 6-fluoro substituent on the metabolism and biological activity of benzo[a]pyrene. Cancer Res 43:1541–1549 5. Caspary W, Cohen B, Lesko S, Ts’o POP(1973) Electron paramagnetic resonance study of iodine-induced radicals of benzo[a]pyrene and other polycyclic hydrocarbons. Biochemistry 12:2649–2656 6. Cavalieri E, Calvin M (1971) Molecular characteristics of some carcinogenic hydrocarbons. Proc Natl Acad Sci USA 68:1251–1253 7. Cavalieri E, Calvin M (1972) 220 MHz nuclear magnetic resonance analysis and the selective protonation of benzo[a]pyrene and 6-methylbenzo[a]pyrene. J Chem Soc, Perkin Trans I:1253–1256 8. Cavalieri E, Auerbach R (1974) Reactions between activated benzo[a]pyrene and nucleophilic compounds, with possible implications on the mechanism of tumor initiation. J Natl Cancer Inst 53:393–397 9. Cavalieri E, Roth R (1976) Reaction of methylbenzanthracenes and pyridine by one-electron oxidation: A model for metabolic activation and binding of carcinogenic aromatic hydrocarbons. J Org Chem 41:2679–2684 10. Cavalieri, EL, Devanesan PD, Rogan EG (1988) Radical cations in the horseradish peroxidase and prostaglandin H synthase mediated metabolism and binding of benzo[a]pyrene to deoxyribonucleic acid. Biochem Pharmacol 37:2183–2187 11. Cavalieri E, Rogan E, Cremonesi P, Devanesan P (1988) Radical cations as precursors in the metabolic formation of quinones from benzo[a]pyrene and 6-fluorobenzo[a]pyrene: Fluoro substitution as a probe for one-electron oxidation in aromatic substrates. Biochem Pharmacol 37:2173–2182 12. Cavalieri E, Rogan E, Higginbotham S, Cremonesi P, Salmasi S (1988) Tumor-initiating activity in mouse skin and carcinogenicity in rat mammary gland of fluorinated derivatives of benzo[a]pyrene and 3-methylcholanthrene. J Cancer Res Clin Oncol 114:16–22

Fluoro Substitution of Carcinogenic Aromatic Hydrocarbons

291

13. Cavalieri EL, Rogan EG, Higginbotham S, Cremonesi P, Salmasi S (1990) Tumorigenicity of 7,12-dimethylbenz[a]anthracene, some of its fluorinated derivatives and 1,2,3,4-tetrahydro-7,12-dimethylbenz[a]anthracene in mouse skin and rat mammary gland. Polycyclic Aromatic Compounds 1:59–70 14. Cavalieri EL, Rogan EG (1992) The approach to understanding aromatic hydrocarbon carcinogenesis. The central role of radical cations in metabolic activation. Pharmacol Ther 55:183–199 15. Cavalieri EL, Rogan EG, Murray WJ, RamaKrishna NVS (1993) Mechanistic aspects of benzo[a]pyrene metabolism. In: Garrigues P, Lamotte M (eds) Polycyclic aromatic compounds. Gordon and Breach, Langhorne, PA, p 1047–1054 16. Cavalieri EL, Rogan EG (1998) Mechanisms of tumor initiation by polycyclic aromatic hydrocarbons in mammals. In: Neilson AH (ed) The handbook of environmental chemistry: PAHs and related compounds, vol. 3J Springer, Berlin Heidelberg New York, Germany, p 81–117 17. Chakravarti D, Pelling JC, Cavalieri EL, Rogan EG (1995) Relating aromatic hydrocarboninduced DNA adducts and c-Harvey-ras mutations in mouse skin papillomas: The role of apurinic sites. Proc Natl Acad Sci USA 92:10422–10426 18. Chakravarti D, Mailander P, Cavalieri EL, Rogan EG (2000) Evidence that error-prone DNA repair converts dibenzo[a.l]pyrene-induced depurinating lesions into mutations: Formation, clonal proliferation and regression of initiated cells carrying H-ras oncogene mutations in early preneoplasia. Mutation Res 456:17–32 19. Chen L, Devanesan PD, Higginbotham S, Ariese F, Jankowiak R, Small GJ, Rogan EG, Cavalieri EL (1996) Expanded analysis of benzo[a]pyrene-DNA adducts formed in vitro and in mouse skin: Their significance in tumor initiation. Chem Res Toxicol 9: 897–903 20. Chou MW, Yang SK, Sydor W, Yang CS (1981) Metabolism of 7,12-dimethylbenz(a)anthracene and 7-hydroxymethyl-12-methylbenz(a)anthracene by rat liver and microsomes. Cancer Res 41:1559–1564 21. Chou MW, Fu PP (1984) Stereoselective metabolism of 8- and 9- fluorobenzo[a]pyrene by rat liver microsomes: Absolute configuration of trans-dihydrodiol metabolites. J Toxicol Environ Health 14:211–223 22. Cremonesi P, Cavalieri EL, Rogan EG (1989) One-electron oxidation of 6-substituted benzo[a]pyrenes by manganic acetate. J Org Chem 54:3561–3570 23. Croy RG, Selkirk JK, Harvey RG, Engel JF, Gelboin HV (1976) Separation of ten benzo(a)pyrene phenols by recycle high pressure liquid chromatography and identification of four phenols as metabolites. Biochem Pharmacol 25:227–230 24. Dawson, JH (1988) Probing structure-function relations in heme-containing oxygenases and peroxidases. Science 240:433–439 25. Devanesan PD, RamaKrishna NVS, Padmavathi NS, Higginbotham S, Rogan EG, Cavalieri EL, Marsch GA, Jankowiak R, Small GJ (1993) Identification and quantitation of 7,12-dimethylbenz[a]anthracene-DNA adducts formed in mouse skin. Chem Res Toxicol 6: 364–371 26. Devanesan PD, Li K-M, Higginbotham S, Harvey RG, Rogan EG, Cavalieri EL (1998) Identification of DNA depurinating adducts of the potent carcinogen, 5-methylchrysene (5MeC). Proc Amer Assoc Cancer Res 39:637 27. DiGiovanni J, Diamond L, Singer JM, Daniel FB, Witiak DT, Slaga TJ (1982) Tumor-initiating activity of 4-fluoro-7,12-dimethylbenz[a]anthracene and 1,2,3,4-tetrahydro-7,12-dimethylbenz[a]anthracene in female SENCAR mice. Carcinogenesis 3:651–655 28. DiGiovanni J, Decina PC, Diamond L (1983) Tumor-initiating activity of 9- and 10- fluoro7,12-dimethylbenz[a]anthracene (DMBA) and the effect of 2,3,7,8-tetrachlorodibenzo-pdioxin on tumor initiation by monofluoro derivatives of DMBA in SENCAR mice. Carcinogenesis 4:1045–1049 29. Groves, JT, McClusky GA,White RE, Coon MJ (1978) Aliphatic hydroxylation by highly purified liver microsomal cytochrome P-450: Evidence for a carbon radical intermediate. Biochem Biophys Res Commun 81:154–160

292

E.L. Cavalieri · E.G. Rogan

30. Groves, JT, Haushalter RC, Nakamora M, Nemo T, Evans BJ (1981) High-valent iron-porphyrin complexes related to peroxidase and cytochrome P-450. J Am Chem Soc 103:2884–2886 31. Harvey RG, Dunne FB (1978) Multiple regions of metabolic activation of carcinogenic hydrocarbons. Nature (London) 273:566–568 32. Hecht SS, Bondinell WB, Hoffmann D (1974)Chrysene and methylchrysenes: Presence in tobacco smoke and carcinogenicity. J Natl Cancer Inst 53:1121–1133 33. Hecht SS, Loy M, Hoffmann D (1976) On the structure and carcinogenicity of the methylchrysenes. In: Freudenthal RI and Jones PW (eds) Carcinogenesis, vol. 1. Raven Press, New York, p 325–340 34. Hecht SS, Hirota H, Loy M, Hoffmann D (1978) Tumor-initiating activity of fluorinated 5methylchrysenes. Cancer Res 38:1694–1698 35. Hecht SS, LaVoie E, Mazzarese R, Hirota N, Ohmori T, Hoffmann (1979) Comparative mutagenicity, tumor-initiating activity, carcinogenicity and in vitro metabolism of fluorinated 5-methylchrysenes. J Natl Cancer Inst 63:855–861 36. Hecht SS, Mazzarese R, Amin S, LaVoie E, Hoffmann D (1979) On the metabolic activation of 5-methylchrysene. In: Jones, PW, Leber P (eds) Polynuclear Aromatic Hydrocarbons, Ann Arbor Science Pub., Ann Arbor, MI, p 733–752 37. Hecht SS, Ronai ZA, Dolan L, Desai D, Amin S (1998) Comparative mouse skin tumorigenicity and induction of Ha-ras mutations by bay region diol epoxides of 5-methylchrysene and 5,6-dimethylchrysene. Carcinogenesis 19:157–160 38. Holder G, Yagi H, Dansette P, Jerina DM, Levin W, Liu AYH, Conney AH (1974) Effects of inducers and epoxide hydrase on the metabolism of benzo[a]pyrene by liver microsomes and a reconstituted system: analysis by high pressure liquid chromatography. Proc Natl Acad Sci USA 71:4356–4360 39. Huberman E, Slaga TJ (1979) Mutagenicity and tumor-initiating activity of fluorinated derivatives of 7,12-dimethylbenz(a)anthracene. Cancer Res 39:411–414 40. Jeftic L, Adams RN (1970) Electrochemical oxidation pathways of benzo[a]pyrene. J Am Chem Soc 92:1332–1337 41. Johnson MD, Calvin M (1973) Induced nucleophilic substitution in benzo[a]pyrene. Nature (London) 241:271–272 42. Lesko S, Caspary W, Lorentzen R, Ts’o POP(1975) Enzymic formation of 6oxobenzo[a]pyrene radical in rat liver homogenates for carcinogenic benzo[a]pyrene. Biochemistry 14:3978–3984 43. Lewis, DFV, Ioannides C, Park DV (1989) Molecular orbital studies of oxygen activation and mechanism of cytochromes P-450-mediated oxidative metabolism of xenobiotics. ChemBiol Interact 70:263–280 44. Lewis DFV (1996) The P450 catalytic cycle and oxygenation mechanism. In: Cytochromes P450. Taylor & Francis, Bristol, PA p. 79–115 45. Lorentzen RJ, Caspary WJ, Lesko SA, Ts’o POP (1975) The autoxidation of 6-hydroxybenzo[a]pyrene and 6-oxobenzo[a]pyrene radical, reactive metabolites of benzo[a]pyrene. Biochemistry 14:3970–3977 46. Loew GH,Wong J, Phillips J, Hjelmeland L, Pack G (1978) Quantum chemical studies of the metabolism of benzo(a)pyrene. Cancer Biochem Biophys 2:123–130 47. Marnett LJ, Weller P. Battista JR (1986) Comparison of the peroxidase activity of hemeproteins and cytochrome P-450. In: Ortiz de Montellano PR (ed) Cytochrome P-450. Plenum Press, New York, p. 29–74 48. Menger EM, Spokane RB, Sullivan PD (1976) Free radicals derived from benzo[a]pyrene. Biochem Biophys Res Commun 71:610–616 49. Mulder PPJ, Chen L, Sekhar BC, George M, Gross ML, Rogan EG, Cavalieri EL (1996) Synthesis and structure determination of adducts formed by electrochemical oxidation of 1,2,3,4-tetrahydro-7,12-dimethylbenz[a]anthracene in the presence of deoxyribonucleosides or adenine. Chem Res Toxicol 9:1264–1277 50. Mulder PPJ, Devanesan P, van Alem K, Lodder G, Rogan EG, Cavalieri EL. Fluorobenzo[a]pyrenes as probes to determine the mechanism of cytochrome P450-catalyzed oxygen transfer in aromatic hydroxylations. (submitted)

Fluoro Substitution of Carcinogenic Aromatic Hydrocarbons

293

51. Nagata C, Y. Tagashira, Kodama M (1974) In: Ts’o POP, DiPaolo JA (eds) Chemical Carcinogenesis, Part A. Marcel Dekker, p 87–111 52. Ortiz de Montellano, PR, Stearns RA (1987) Timing of the radical recombination step in cytochrome P-450 catalysis with ring-stained probes. J Am Chem Soc 109:3415–3420 53. Pataki J, Balick R (1972) Relative carcinogenicity of some diethylbenz[a]anthracenes. J Med Chem 15:905–909 54. Peltonen K, Hilton BD, Pataki J, Lee H, Harvey RG, Dipple A (1991) Spectroscopic characterization of syn-5-methylchrysene 1,2-dihydrodiol 3,4-epoxide-deoxyguanosine adducts. Chem Res Toxicol 4:305–310 55. RamaKrishna NVS, Cavalieri EL, Rogan EG, Dolnikowski GG, Cerny RL, Gross ML, Jeong H, Jankowiak R, Small GJ (1992) Synthesis and structure determination of the adducts of the potent carcinogen 7,12-dimethylbenz[a]anthracene and deoxyribonucleosides formed by electrochemical oxidation: Models for metabolic activation by one-electron oxidation. J Am Chem Soc 114:1863–1874 56. RamaKrishna NVS, Devanesan PD, Rogan EG, Cavalieri EL, Jeong H, Jankowiak R, Small GJ (1992) Mechanism of metabolic activation of the potent carcinogen 7,12-dimethylbenz[a]anthracene. Chem Res Toxicol 5:220–226 57. Reardon DB, Prakash AS, Hilton BD, Roman JM, Pataki J, Harvey RG, Dipple A (1987) Characterization of 5-methylchrysene-1,2-dihydrodiol-3,4-epoxide-DNA adducts. Carcinogenesis 8:1317–1327 58. Rochlitz J (1967) Neue Reaktionen der carcinogenen Kohlenwasserstoffe-II. Tetrahedron 23:3043–3048 59. Shipman LL (1978) Ab initio quantum mechanical characterization of the ground electronic state of benzo[a]pyrene. Implications for the mechanism of polynuclear aromatic hydrocarbon oxidation to epoxides by cytochrome P-450. In: Jones PW, Freudenthal RI (eds) Carcinogenesis, Vol. 3, Polynuclear Aromatic Hydrocarbons. Raven Press, p 139–144 60. Slaga TJ, Gleason GL, DiGiovanni J, Sukumaran KB, Harvey RG (1979) Potent tumor-initiating activity of the 3,4-dihydrodiol of 7,12-dimethylbenz(a)anthracene in mouse skin. Cancer Res 39:1934–1936 61. Stack DE, Cremonesi P, Hanson A, Rogan EG, Cavalieri EL (1995) Radical cations of benzo[a]pyrene and 6-substituted derivatives: reaction with nucleophiles and DNA. Xenobiotica 25:755–760 62. Sullivan PD, Bannoura F, Daub G (1985) 13C and 1H EPR analysis of the benzo[a]pyrene cation radical. J Am Chem Soc 107:32–35 63. Tolbert LM, Khanna RK, Popp AE, Gelbaum L, Bottomley LA (1990) Stereoelectronic effects in the deprotonation of arylalkylradical cations: meso-ethylanthracenes. J Am Chem Soc 112:2373–2378 64. Wilk M, Bez W, Rochlitz J (1966) Neue Reaktionen der carcinogenen Kohlenwasserstoffe 3,4-Benzpyren, 9,10-Dimethyl-1,2-Benzanthracen und 20-Methylcholanthren. Tetrahedron 22:2599–2608 65. Wilk M, Girke W (1972) Reactions between benzo[a]pyrene and nucleobases by one-electron oxidation. J Natl Cancer Inst 49:1585–1597 66. Yang SK, Roller PP, Gelboin HV (1977) Enzymatic mechanism of benzo[a]pyrene conversion to phenols and diols and an improved high-pressure liquid chromatographic separation of benzo[a]pyrene derivatives. Biochemistry 16:3680–3687

CHAPTER 10

Phosphofluoridates: Biological Activity and Biodegradation Joseph J. DeFrank 1 · William E. White 2 U.S.Army Edgewood Chemical Biological Center,Aberdeen Proving Ground, MD 21010–5424, USA 1 E-mail: [email protected] 2 E-mail: [email protected]

The combination of phosphorus and fluorine has resulted in a variety of compounds that are unique in their physical and chemical properties as well as being some of the most toxic materials produced by man. A review of the chemistry and enzymatic basis for the toxicology of these compounds is presented. Catalytic enzymes that use these compounds as substrates and their potential applications are also reviewed. Keywords. Phosphorofluoridates, Organophosphates, Pesticides, Acetylcholinesterase, Nerve agents, Decontamination, Chemical warfare

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297

1

Introduction

2

Bonding in Phosphorus Compounds . . . . . . . . . . . . . . . . 298

2.1

Geometries of Phosphorus Compounds . . . . . . . . . . . . . . . 299

3

Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300

4

Inorganic Phosphorus Fluorides

4.1 4.2 4.3

Phosphorus Trifluoride . . . . . . . . . . . . . . . . . . . . . . . . 302 Phosphorus Pentafluoride . . . . . . . . . . . . . . . . . . . . . . 303 Ligand Exchange and Pseudorotation . . . . . . . . . . . . . . . . 303

5

Acid and Base Properties . . . . . . . . . . . . . . . . . . . . . . . 305

6

Nucleophilic Substitution

6.1 6.2 6.3

Unimolecular Mechanism . . . . . . . . . . . . . . . . . . . . . . 306 Bimolecular Mechanism . . . . . . . . . . . . . . . . . . . . . . . 307 Theoretical Studies on Mechanisms of Hydrolysis . . . . . . . . . 310

7

Neurotransmitters . . . . . . . . . . . . . . . . . . . . . . . . . . 310

7.1 7.2

Acetylcholine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310 Acetylcholinesterase (AChE) . . . . . . . . . . . . . . . . . . . . . 311

. . . . . . . . . . . . . . . . . . 302

. . . . . . . . . . . . . . . . . . . . . . 306

The Handbook of Environmental Chemistry Vol. 3, Part N Organofluorines (ed. by A.H. Neilson) © Springer-Verlag Berlin Heidelberg 2002

296

J.J. DeFrank · W.E. White

8

Development of Chemicals Designed to Inhibit Acetylcholinesterase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

9

Chemical Warfare Agents

9.1 9.2 9.3 9.4 9.5

Toxicities . . . . . . . . . . . . . . . . . . . . . . . . Quantitative Relation Between Reactivity and Toxicity Effects of Stereochemistry on Toxicity . . . . . . . . . Restoration of AChE Activity . . . . . . . . . . . . . . Signs and Symptoms of Organophosphate Poisoning

10

Enzymatic Hydrolysis of Phosphorus Fluoridates . . . . . . . . . 321

11

Organophosphorus Pesticide Biodegradation

12

Nerve Agent Enzymes

13

Enzyme Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . 325

14

Organophosphorus Hydrolase . . . . . . . . . . . . . . . . . . . . 325

15

Halophilic Bacterial OPAAs . . . . . . . . . . . . . . . . . . . . . 327

16

Enzyme Production and Applications . . . . . . . . . . . . . . . . 332

16.1 16.2 16.3

Field Decontamination . . . . . . . . . . . . . . . . . . . . . . . . 332 Prophylaxis and Therapy . . . . . . . . . . . . . . . . . . . . . . . 333 Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335

17

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336

. . . . . . . . . . . . . . . . . . . . . . 314 . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

315 316 318 320 320

. . . . . . . . . . . 323

. . . . . . . . . . . . . . . . . . . . . . . . 324

List of Abbreviations AChE DFP GA GB GD GF OPAA OPH

. . . . .

acetylcholinesterase diisopropylfluorophosphate tabun (N,N-dimethylethyl phosphoroamidocyanidate) sarin (O-isopropyl methylphosphonofluoridate) soman (3,3-dimethyl-2-butyl methylphosphonofluoridate) O-cyclohexyl methylphosphonofluoridate organophosphorus acid anhydrolase organophosphorus hydrolase (also parathion hydrolase and phosphotriesterase) RTECS Registry of Toxic Effects of Chemical Substances R-VX Russian VX (S-2(diethylamino)ethyl O-isobutyl methylphosphonothioate) VX S-2(diisopropylamino)ethyl O-ethyl methylphosphonothioate

Phosphofluoridates: Biological Activity and Biodegradation

297

1 Introduction Phosphorus was discovered in 1699 by Brand when he isolated it from urine [2]. Since then, many beneficial and deleterious applications have been identified or discovered. It is a critical component of biopolymers, bones, teeth, energetic metabolites, and other critical biomolecules [167]. In addition, various phosphorus compounds have been used in detergents, glasses, cements, pesticides, and metal treating processes. In a less humanitarian role during military conflicts, white phosphorus was a critical component of incendiary munitions used to destroy entire towns. Because fluorine is the most electronegative element, introduction of a P–F bond alters the chemistry and therefore the biology of phosphorus in different ways. First, electron density around phosphorus is reduced, thereby increasing or decreasing (depending on the substituents) the strength of the bonds to the other atoms. Second, the fluorine may serve as a leaving group in a nucleophilic substitution reaction. Because fluorine is the weakest leaving group among the halogens, phosphorus fluorides are frequently much more specific in their reactivity than other phosphorus compounds. By altering the electronic structure around the phosphorus atom, fluorine may increase or decrease the activation energy and thereby change the kinetics of reactions. The chemistry of the P–F bond may be compared with the analogous C–F bond that is discussed by Dixon in Chap. 1. One beneficial application of fluorine is the prevention of dental caries by replacing some of the OH groups in hydroxyapatite (a crystalline calcium phosphate material). Fluorine substitution hardens the enamel and makes the tooth less susceptible to decay [46]. Although calcium or stannous fluorides have been used extensively, the best material is the sodium salt of monofluorophosphoric acid because some of the problems associated with precipitation of the fluoride salts are obviated. During the early part of the twentieth century, the interest in organophosphorus fluorides was minimal until the discovery by Lange in the late 1920s and early 1930s that some of these compounds were toxic. His germinal discovery led to the establishment of major research programs in Europe and culminated in the development of a new class of chemical warfare agents (nerve agents) by the German military. The biochemistry and toxicity of organophosphorus fluorides depend on the chemistry of these compounds, which results from the electronic structure of phosphorus, and the vacant and filled electronic orbitals. This review describes the chemistry of organophosphorus fluoride compounds, the corresponding reactions that lead to enzyme inhibition, and the development of other enzymes that catalyze the degradation of these compounds without themselves becoming irreversibly inhibited. There are review articles on several aspects of phosphorus fluoride chemistry. The extensive review by Schmutzler describes the general principles of phosphorus-fluorine chemistry and covers the literature through the mid-1960s [150]. The chemistry, environmental fate, and toxicology of organophosphorus pesti-

298

J.J. DeFrank · W.E. White

cides are covered in a book edited by Chambers and Levi [31]. The recent review of the chemistry of the organophosphorus nerve agents by Black and Harrison provided a good synopsis of the current methods for synthesis, analysis, and detection [26]. Decontamination methods for chemical warfare agents were discussed by Yang et al. [173]. Books on chemical warfare agents by Compton [45], Somani, and Franke discuss nerve agents as well as other classes of compounds that have been used as chemical warfare agents. Kosolapoff provides insight into the focus of unclassified organophosphorus chemistry immediately after World War II. His discussion of halogen analogs is limited principally to chlorides and no mention is made of phosphonofluoridates [95]. Finally, there are several books on phosphorus chemistry that cover a wide range of topics and provide good insight into the differences in chemistry between fluorophosphorus chemistry and other phosphorus compounds [46, 65]. The recent extensive treatise on organophosphorus chemistry edited by Hartley consists of a fourvolume set published in alternate years starting in 1990 [75–78]. Because this series covers the entire spectrum of OP chemistry, the percentage devoted to phosphorus-fluorine chemistry is relatively small. In contrast, the single volume by Quin is a very readable text that provides excellent insight into current OP chemistry issues [135]. The cited references can provide information that is more extensive.

2 Bonding in Phosphorus Compounds Phosphorus is located in group V (the pnictides) of the periodic table. In many ways this second row element resembles the third row member, arsenic, more closely than the first row nitrogen. Nitrogen is found in nature in the negative three oxidation state (i.e., ammonia), the plus five state (i.e., nitrates) as well as others. In contrast, phosphorus is found in the environment almost exclusively in the plus five state. Trivalent nitrogen compounds like ammonia and amines are quite basic because the nitrogen atom is readily capable of donating its lone pair of electrons. In contrast, the phosphorus lone pair is less available to Lewis acids so most phosphines are not particularly basic. Trivalent phosphorus is much more prone to oxidation by sulfur and oxygen than corresponding nitrogen compounds. The five valence electrons and low-lying d orbitals offer numerous opportunities for bonding in a wide variety of organic and inorganic compounds. The s, p, and d orbitals and their hybrids afford a variety of bonding geometries not only for stable compounds but also for structures that occur transiently during the course of reactions. The need for d orbitals in penta- and hexacoordinate phosphorus is obvious; however, their role in lower valent structures is less apparent. Vacant d orbitals have been implicated in various degrees of back bonding particularly with nitrogen; however, the actuality of this phenomenon/conjecture is still uncertain. Some evidence supports d orbital back bonding [108], whereas other reports indicate little evidence [68–70]. This issue has been a point of discussion and controversy for many years and probably won’t be resolved soon [8].

299

Phosphofluoridates: Biological Activity and Biodegradation Table 1. Bond angles [48]

Compound

Bond angles

Compound

Bond angle

NH3 PH3 AsH3 SbH3

107.3 93.3 91.8 91.3

NF3 PF3 AsF3 SbF3

102.1 97.8 96.2 88.0

2.1 Geometries of Phosphorus Compounds

Monovalent phosphorus compounds normally contain a single triple bond. There were suggestions about H–C∫P as early as 1950; however, experimental evidence for its synthesis did not appear until 1961 [93]. This area of chemistry was almost ignored for many years; however, interest is increasing, particularly the involvement of these compounds in pericyclic reactions [104]. Binary phosphorus compounds obviously have only one bond and therefore a linear geometry. The bond length for the phosphazyne P∫N has been reported to be 1.49 Å. [72]. Many of the phosphazynes are prepared by loss of nitrogen from the corresponding azide, and all polymerize very rapidly if they are not stabilized in the form of a complex. Above 800 °C, elemental phosphorus exists as a dimer, though at lower temperatures white phosphorus exists as a tetramer with tetrahedral geometry. For compounds in which the phosphorus atom has two substituents, the arrangement usually consists of one single and one double bond. Aspects of this chemistry are the subject of reviews by Cowley [49] and by Cowley and Kemp [50]. The most common phosphorus compounds contain three or four substituents. Pentavalent phosphorus compounds are less plentiful but are significant. Because there is increasing research interest in hexavalent phosphorus, this class will probably increase in importance and visibility in the future. Most of the tertiary compounds like PH3 and PF3 are pyramidal with equivalent bonds formed by some degree of p3 and sp3 hybridization. The lone pair sits at the apex of the pyramid. Unlike the situation in ammonia, inversion of the pyramid is slow in phosphines due to the high-energy barrier. The bond angles for the pnictide hydrides and fluorides are indicated in Table 1 [48]. The angles decrease as the size of the central atom increases, thereby indicating an increased amount of p character. Because the s orbitals are closer to the nucleus than the p, an increasing amount of p character in the lone pair would normally increase the basicity; however, other factors cause the basicity to decrease. Tetracoordinated phosphorus exists in a tetrahedral or distorted tetrahedral geometry. Planar or square pyramid geometries are extremely rare if they exist at all. Tetra substituted compounds may be ionic like PH+4 or PR+4 or a pseudo pentavalent structure like the following where R represents almost any non-metal.

300

J.J. DeFrank · W.E. White

Although organophosphorus esters like triethyl phosphate are frequently written with a P=O double bond for simplicity, this bond is extremely polar so that a more accurate representation has a partial positive charge on the phosphorus and a partial negative charge on the oxygen.

Pentacoordinated phosphorus exists almost exclusively as a trigonal bipyramid with sp3d hybridization. The halogenated compounds such as PF5 and PCl5 typify this class. Various reactive intermediates, transition structures, and other transients also have this geometry. A square pyramidal configuration may exist as a transient when axial and equatorial ligands are exchanging position via pseudorotation. Hexacoordinated complexes exist in octahedral geometries with d 2sp3 hybridizations. Although the phosphorates were discovered in 1954, serious interest in these compounds did not begin until the 1980s [40]. Various ions such as PF6– fall into this category. For neutral hexacoordinate compounds both electrons must come from the ligand. Nitrogen, oxygen, and sulfur have been shown to form donor complexes with phosphorus [82]. In general, the donor bonds are longer than traditional covalent bonds. Phosphites as well as phosphates are capable of forming these complexes. In general, the bond order is greater for phosphites than for phosphates – probably because the back-bonding with the phosphoryl oxygen in the phosphates decreases the electrophilicity on the phosphorus. There is increasing evidence that some enzyme-catalyzed reactions may proceed through a hexacoordinate transition with the sixth bond supplied by interaction with an amino acid containing a basic side chain [83, 84].

3 Nomenclature The nomenclature is relatively straightforward for inorganic fluorophosphorus compounds like phosphorus pentafluoride (PF5) and phosphorus trifluoride (PF3). In contrast, the nomenclature for organophosphorus compounds is more complex and sometimes confusing because of the prevalence of common names and different nomenclature rules used in both the old and the current literature. In addition, the preferences of organic and inorganic chemists occasionally differ from those of biochemists and toxicologists. Phosphines are trivalent compounds having three alkyl, aryl, or hydrogen substituents. Oxidation of the phosphorus to a pentavalent species produces a phosphine oxide.

Phosphofluoridates: Biological Activity and Biodegradation

301

Phosphites such as trimethyl phosphite can be considered esters of phosphorus acid. For phosphorus acids the equilibrium between the trivalent and pentavalent tautomers is so far toward the pentavalent that phosphorus acid is dibasic rather than tribasic and should be considered phosphonic acid. Trimethyl phosphite is usually prepared by methanolysis of phosphorus trichloride, which in turn is prepared by the chlorination of red phosphorus.

Trimethylphosphite is a valuable building block for larger organophosphorus compounds because it undergoes the Arbusov reaction (Scheme 1) – one of the more general reactions for forming a phosphorus-carbon bond. The reaction of a phosphite with an alkyl halide produces a phosphonate and another alkyl halide.

Scheme 1

The esters of phosphinic, phosphonic, and phosphoric acids are phosphinates, phosphonates, and phosphates, respectively. Nomenclature for phosphorus esters is analogous to carboxylic esters with separate words for the alkoxy groups. The term phosphylate is sometimes used generically to refer to any of the phosphorus esters.

Phosphorus fluoridates are esters containing a phosphorus-fluorine bond. For example, replacement of one the methoxy groups in dimethyl methylphosphonate or trimethyl phosphate with fluorine produces the corresponding fluoridate.

302

J.J. DeFrank · W.E. White

For common organophosphorus compounds, old names frequently persist. The diisopropyl ester of fluorophosphoric acid is more frequently indicated by its common name, diisopropyl fluorophosphate (DFP), rather than by its appropriate scientific name, diisopropyl phosphorofluoridate. Substituents attached to oxygen, sulfur, and nitrogen are usually designated with O, S, and N respectively when there is the possibility of confusion. It is never incorrect to include them.

In general, the leaving group takes precedence in nomenclature. Therefore, the compounds indicated above are a phosphorothiolate and a phosphorofluoridate and not phosphoroamidates.

4 Inorganic Phosphorus Fluorides 4.1 Phosphorus Trifluoride

PF3 is a reactive compound (though less reactive than PCl3 , which hydrolyzes immediately in water) that is normally prepared by fluorination of PCl3 ; however, it can be prepared by reaction of red phosphorus with HF. PCl3 is the ultimate starting material for almost all organophosphorus compounds. The most common industrial method of production of PCl3 is by direct chlorination of red phosphorus; however, white phosphorus is sometimes used when a purer product is needed. Examination of the geometry of the smallest of the phosphorus fluorine compounds, phosphorus trifluoride (PF3), provides insight into the nature of the P–F bond. PF3 exists as a trigonal pyramid like ammonia and phosphine. The P–F bonds are 1.570 Å and the F–P–F angle is 98°. The bond angle increases with the larger halogens: 100° for PCl3 , 101° for PBr3 , and 102° for PI3 . The small bond angle indicates considerable p character for the P–F bond. Bonds formed with p orbitals only would have 90° bond angles reflecting the orthogonality of the three p orbitals. A pure sp3 hybrid would have bond angles of 109°. The result of the higher p character in the P–F bond is an increased s character in the lone pair on the phosphorus and reduced basicity. Because the s electron shell is closer to the nucleus, the lone pair electrons are held more tightly in the free PF3 molecule,

Phosphofluoridates: Biological Activity and Biodegradation

303

thereby reducing the basicity of the molecule. Even though phosphines are poor bases, hundreds of coordination complexes have been formed with PF3 ligands. The previous statements seem to be inconsistent (i.e., weaker Lewis base but stronger complexes). The usual explanation involves the formation of dative bonds or back-bonds between the metal atom and the vacant d orbitals of phosphorus [48]. The fluorine moiety reduces the electron density around the phosphorus thereby making it a better receptor of the electrons from the metal. The stronger back-bonds more than offset the weaker coordination bond between the phosphorus lone pair and the vacant s and p orbitals on the metal. 4.2 Phosphorus Pentafluoride

Phosphorus pentafluoride is much less stable than phosphorus trifluoride. PF5 hydrolyzes immediately in water to form POF3 and HF. It may be prepared by fluorination of PCl5 , by oxidation of PF3 with chlorine in the presence of CaF2 or by the dissociation of the PF3 · As5 complex. The bonding in PF5 is interesting, and this symmetrical molecule can serve as a model for the pentavalent transition structures for the nucleophilic substitution of tetravalent phosphorus esters that will be discussed in detail later. Electron diffraction studies indicate that the PF5 molecule exists as a trigonal bipyramid rather than in a square planer configuration. The two axial bonds are longer (1.534±0.004 Å) than the three equatorial bonds. (1.577±0.005 Å) [80]. The equatorial bonds are most likely the result of sp2 hybridization, which would create three equivalent bonds in a plane around the phosphorus. The axial bonds probably result from pd hybridization, which would generate a linear bonding arrangement perpendicular to the plane of the equatorial substituents. Because the d orbitals are larger than the s and p, the axial bonds should be longer and weaker than the equatorial. This assertion is supported by the electron diffraction data and the position of nucleophilic substitution. For pentavalent phosphorus compounds with different substituents, the smallest and most electronegative moieties occupy the axial positions. Substituting a methyl group for one of the equatorial fluorines lengthens the axial P–F bonds to 1.612±0.004 Å and the remaining equatorial bonds to 1.543±0.006 Å. Addition of a second methyl group lengthens the bonds further to 1.643±0.003 Å for the axial and 1.553±0.006 Å for the equatorial [80]. 4.3 Ligand Exchange and Pseudorotation

Because the energy barriers for the exchange of the axial and equatorial substituents are relatively small, both conformations of the PF5 molecule are interchangeable at room temperature. Measurements by electron diffraction are carried out in a very short period of time and, as indicated in the previous section, are able to differentiate the axial and equatorial fluorine atoms in the molecule. In contrast, nuclear magnetic resonance with its relatively slower time of measurement shows only a single average signal for the five fluorine atoms.

304

J.J. DeFrank · W.E. White

There are two mechanisms that have been proposed for exchange leading to a change in configuration. Turnstile rotation involves the exchange of three ligands. An initial distortion creates a structure in which one axial and two equatorial ligands are equivalent with respect to the phosphorus. Then a simultaneous rotation moves the axial ligand to the first equatorial position, the first equatorial ligand moves to the second equatorial position, and the ligand in the second position moves to the axial. A reversal of the distortion reforms the trigonal bipyramid with the ligands in the new positions. In Berry pseudorotation, the two axial ligands move to equatorial positions, and two of the three equatorial ligands become axial. To achieve this exchange, the four bonds are distorted about 15° to form a square pyramid in which the four ligands occupy essentially equivalent positions with respect to the phosphorus. The ligand that is not involved is positioned at the apex of the pyramid. Then the pyramid dissociates in opposite manner to form a new trigonal bipyramid with the ligands in the new configuration. Theoretical studies indicate low activation energies for the Berry type pseudorotation [165]. For PH5 , the values are 2.68 and 1.53 kcal M–1 for calculations performed at the HF and the MP4 levels of theory with large triple zeta basis sets. For PF5 , the corresponding values are 5.07 and 4.24 kcal M–1 respectively with the smaller 6–31G* basis. In general, the pseudorotational barriers for compounds having dissimilar ligands like PHF4 and PH2F3 were higher. The preference of methyl substituents for the equatorial position is so strong that only one stable geometry could be identified for P(CH3)F4 . During pseudorotation of this compound, the methyl group occupies the pivot position exclusively and therefore remains in the equatorial position. When two methyl groups are present in the molecule, one must move to an axial position during pseudorotation. Because of the strong preference of the methyl group for the equatorial position, no minimum with one of the methyl groups in an axial position could be located for the P(CH3)2F3 . Calculations on the unstable structure having one methyl group in an axial position and the other equatorial indicated an energy level of 15.04 (HF) and 14.01 (MP4) kcal M–1 higher than that for the stable isomer with both methyls in equatorial positions. The propensity of a substituent to occupy an apical position is termed apicophilicity. Theoretical studies at the Hartree-Fock and Møller-Plesset level established the following order of substituents [171]: FOH > HCH3 NH2 Based on this work, it is very reasonable that a methyl group would reside exclusively in the equatorial position. Earlier work in another laboratory using a tailored basis set (larger than double zeta quality) at the second order Møller-Plesset level on Hartree-Fock geometries produced an energy of activation value for PF5 Berry pseudorotation of 16±2 kJ (i.e., 3.8±0.5 kcal M–1) after making some small corrections to errors inherent in the basis set enlargement [110].

305

Phosphofluoridates: Biological Activity and Biodegradation

5 Acid and Base Properties Organophosphorus compounds have both acidic and basic properties. Basic properties result from the donation of the lone pair of electrons on phosphorus and are characteristic of the trivalent compounds in the group 5 elements.Acidic phosphorus compounds are mainly pentavalent species with OH substituents. The acidic properties result from donation of a proton similar to carboxylic and sulfuric acids. Table 2 [71, 72] compares the pKa’s for the conjugate acids of a series phosphines and amines. The phosphorus compounds are much less basic than the corresponding nitrogen analogs. The usual explanation for the reduced availability of the electrons involves the incorporation of the phosphorus lone pair into the 3d orbitals. The promotion of the electrons is possible in phosphorus because the energy levels of the 3d orbitals are not much higher than those of the 3s and 3p orbitals. In nitrogen, very little promotion occurs because the empty 3d orbitals are considerably higher than the occupied 2 s and 2p. The effect of the reduced basicity on intermolecular hydrogen bonding can be observed in the boiling points of amines and phosphines. Even though phosphine (PH3) has a higher molecular weight than ammonia (NH3), its normal boiling point is considerably lower (i.e., –87 °C vs –33 °C). The effect of hydrogen bonding on boiling point can be deduced by comparing the boiling point of methane (i.e., –161 °C), which essentially has no intermolecular hydrogen bonding, to that for ammonia and water (100 °C). Table 3 [46] indicates the effect of hydroxyl groups on the pKa’s of phosphorus acids. Phosphinic acid, with only one acidic hydrogen, is the strongest acid with a pKa of 1.1. Phosphonic acid is a dibasic acid with pKa’s of 1.3 and 6.7. Phosphoric acid is the most common by several orders of magnitude and is one of the major industrial chemicals. Even though it has three replaceable hydrogens, its pKa’s are higher than the corresponding values of the other two. Substitution at phosphorus affects the pKa’s of the various phosphorus acids. Fluorine, the most electronegative element in the periodic table, exerts an electron-withdrawing effect on phosphorus. The effect can be observed by compar-

Table 2. Basicities of phosphines and amines [71, 72]

Phosphine

pKa for BH+

Amine

pKa for BH+

PH3 CH3PH2 (CH3)2PH (CH3)3P

–14.0 –3.2 3.91 8.65

NH3 CH3NH2 (CH3)2NH CH3)3N

9.24 10.65 10.79 9.80

0.27 4.55

306

J.J. DeFrank · W.E. White

Table 3. Acidity of phosphorus acids [46]

pKa 1 pKa 2 pKa 3

Phosphinic acid

Phosphonic acid

Phosphoric acid

Fluorophosphoric acid

1.1

1.3 6.7

2.1 7.2 12.7

0.55 4.8

ing the pKa’s of fluorophosphoric acid with phosphoric acid. As a result, phosphoric acid is less acidic than its fluoride analog [46]. The pKa value is a thermodynamic property that depends only on the free energy differences between the free acid and the conjugate anion (assuming the free energy of the proton is constant in aqueous solution). In the anion of fluorophosphoric acid, the fluorine moiety withdraws electron density from the phosphoryl group, partially reduces the effect of the negative charge, and thereby stabilizes the entire system. Even though fluorine exerts a similar effect on the corresponding free acid, the magnitude of the effect is less because the system is uncharged.

6 Nucleophilic Substitution Nucleophilic substitution of tetravalent organophosphorus compounds occurs via two principal mechanisms. The bimolecular SN 2 reaction is the most general and occurs in solution as well as at the active center of enzymes such as ribonuclease, nucleic acid polymerase, and acetylcholinesterase. In contrast, the unimolecular SN 1 reaction is relatively rare and occurs only when there is an active hydrogen adjacent to the phosphorus. 6.1 Unimolecular Mechanism

For organophosphorus compounds having an active hydrogen adjacent to the phosphorus, nucleophilic substitution can occur via a first-order reaction in addition to the traditional second-order pathway [65]. A base removes the proton with immediate elimination of the fluorine substituent to form a planar

Scheme 2

307

Phosphofluoridates: Biological Activity and Biodegradation Table 4. Rate constants and activation energies for fluorophosphorodiamidates [79]

R1

R2

K(OH–) min–1 / 25 °C

E* (Kcal M–1)

MeNH EtNH PrNH Me2N Et2N Et2N

Me2N Me2N Me2N Me2N Me2N Et2N

17.6 12.3 8.43 4.70 ·10–3 2.87 ·10–4 2.50 ·10–5

11.2 11.4 11.9 14.7 16.6 17.1

metastable ortho phosphate intermediate. The nucleophile then adds to either side of the phosphate to form the racemic product (Scheme 2). Table 4 [79] compares the rate constants for the hydrolysis of the P-F bond in a series of phosphorodiamidates having three or four alkyl substituents [79]. Compounds with hydrogen attached to the nitrogen have rate constants 104 to 106 times greater than those with only alkyl groups on the nitrogen. Energies of activation for the SN 1 reaction were calculated to be about 11 kcal/mole in contrast to 15–17 kcal M–1 for the SN 2 reaction under the same conditions. There is some evidence that the biochemical hydrolysis of adenosine triphosphate, which is coupled to a variety of unfavorable reactions to achieve an overall favorable DG, proceeds via a similar mechanism [65]. 6.2 Bimolecular Mechanism

Although there is some evidence for equatorial attack that leads to a retention of configuration [47], the SN 2 reaction in phosphinates, phosphonates, and phosphates is usually thought to occur through a trigonal bipyramid with the nucleophile in one axial position and the leaving group on the other [39, 81]. The principles associated with the orientation of the reaction were developed by Frank Westheimer in a series of studies on hydrolysis of cyclic phosphates and described in a landmark synopsis. They have since become known as Westheimer’s rules or guidelines [168]. Nucleophilic attack on tetravalent phosphorus proceeds through a pentavalent trigonal bipyramid. Nucleophiles enter at one axial position; leaving groups depart at the other. Ligands may occupy either axial or equatorial positions depending on their electronegativity, size, etc. If the trigonal bipyramid is stable for some brief period, pseudorotation may occur to move a leaving group from an equatorial to an axial position. The asymmetrical nature of the reactant is maintained in the transition structure so the reaction proceeds with an inversion of configuration – just as it does in the SN 2 reaction of tetravalent carbon compounds (Scheme 3).

308

J.J. DeFrank · W.E. White

Scheme 3

The stability of the leaving group is one of the factors that determine the rate of nucleophilic substitution. In general, the reaction rate parallels the stability of the leaving group, which is reflected in their respective pKa’s (i.e., I– >Br–>Cl– >F–). Iodine and bromine are quite reactive. Chlorine is moderately reactive and is the most frequently used leaving group for reactions performed in nonaqueous solution. Fluorine analogs are the least reactive of the halogenated phosphorus compounds and have limited (though significant) applications as substrates in the preparation of other compounds. Because the rate of hydrolysis of the P–Cl bond is quite rapid, many of the kinetic studies on phosphorus chloridates were conducted in acetone or ethanol containing only 5% water [123–125]. In contrast, the hydrolysis of the fluoride analogs was sufficiently slow to be performed in water. The quantitative effect of the leaving group on rate of hydrolysis is indicated in Table 5 [66]. In triethyl phosphate, ethoxide is a very poor leaving group because ethanol is a weak acid with a pKa around 19. The hydrolysis rate is about six orders of magnitude less than that of the fluoride analog, which has a pKa of 4.73. Phenol is a stronger acid than ethanol because the negative charge can be partially distributed around the aromatic ring. Many commercial pesticides employ p-nitrophenoxide leaving groups (pKa =7.15) because most are sufficiently reactive against insects and other pests to be useful but are not so toxic that they pose a serious health hazard to humans. Because of regulatory action as well as market pressure, older pesticides are continually replaced by newer compounds having wider safety margins and less adverse environmental impact. Most theoretical calculations indicate the presence of one or more metastable intermediates separated by transition structures [164]; however, the results from a few studies are more consistent with a single transition structure separating the reactants from the products [60, 61]. The empirical evidence for the existence of the metastable intermediates is less compelling because they have never been isolated, trapped, or analyzed spectroscopically. The existence of these intermediTable 5. Relative rates of hydrolysis of diethyl phosphates [66]

X

pKa

Relative rate of hydrolysis

F SPh OPh OEt

4.73 8.3 9.95 19

1 2.3¥10–2 1.7¥10–4 1.7¥10–6

309

Phosphofluoridates: Biological Activity and Biodegradation Table 6. Effect of alkoxy groups on anion stability

Dimethyl phosphinic acid

O-Methyl methyl phosphonic acid

O, O-Dimethyl phosphoric acid

3.1 [63]

2.67 [43]

1.29 [63]

pKa

ates, even if only fleeting, is the basis of considerable controversy regarding the mechanisms of organophosphorus reactions. Replacing the hydrogens on phosphinic acid with methyl groups produces dimethylphosphinic acid. The electron-releasing alkyl groups increase the electron density in the O=P–O region. Because the impact of the effect is greater in the anion than in the free acid, the anion is effectively destabilized and the pKa increases to 3.1. In contrast, as shown in Table 6 [43, 63], the alkoxy groups on the phosphonic and phosphoric acids are electron-withdrawing and thereby partially stabilize the anion so that the pKa decreases. Table 7 [73] indicates the relation between structure of the fluorophosphorus compounds and rate of hydrolysis in aqueous solution at neutral pH. The phosphinic fluoride is the most reactive. The only example indicated, diethylphosphinic fluoride, is three times as reactive as the most reactive phosphonate. The size of the substituents on the phosphonates affects their reactivity, but in general the fluorophosphonates are about one tenth as reactive as the corresponding phosphinic compounds. The fluorophosphates have the smallest rate constant and therefore are the most stable – about ten times less reactive than the phosphonates. Increasing the size of the substituents decreases the rate of hydrolysis for both phosphonates and phosphates. Table 7. Rates of hydrolysis of fluorophosphorus compounds [73]

Class

R1

R2

Kaq (¥105)

Phosphinic Phosphonic

Et Me Me Me Me Et Et Et OMe OEt OPr OBu

Et OMe OEt OPr OBu OMe OEt OPr OMe OEt OPr OBu

94 32 16 15 10 10 0.50 0.49 1.2 0.29 0.26 0.12

Phosphoric

310

J.J. DeFrank · W.E. White

6.3 Theoretical Studies on Mechanisms of Hydrolysis

The reaction of methanol with dimethylphosphinic fluoride provides a good model for nucleophilic substitution of the phosphorus fluoride bond. Theoretical studies were performed on the isolated system (i.e., equivalent to a gas phase reaction) at the Hartree-Fock level and with density functional theory using the B3LYP hybrid functional [172]. As the methanol group approaches the phosphorus, a loose association forms between the two molecules and the energy decreases to a minimum when the methanol moiety is about 3 Å from the phosphorus. As the P–O distance shortens further and the P–O bond begins to form, the energy rises to a maximum (transition1) and the hydrogen begins to shift to the phosphinyl oxygen. The energy then drops to an intermediate (local minimum) with the P–OH pointed toward the methanol. The reaction then proceeds through a small transition as the hydroxyl rotates 180° from the methoxy side to the fluoride side. Next, as the P–F bond begins to lengthen, the energy rises to a maximum (third transition) and the hydrogen begins to shift from the phosphinyl oxygen to the fluoride. The P–F bond continues to lengthen and finally breaks with the formation of O-methyl dimethylphosphinate and HF.

7 Neurotransmitters Neurotransmitters are small molecules that transmit neural signals from the axon of one neuron across a small space called a synapse to a receptor. Neurotransmitters such as gamma aminobutyric acid (GABA) or the catecholamines stimulate or depress various physiological phenomena for significant periods. 7.1 Acetylcholine

Acetylcholine is a neurotransmitter that conducts impulses across neuro-neuro and neuro-muscular junctions and interacts with cholinergic receptors [163]. The signal travels down the axon by a sequential influx of sodium ions.When the impulse reaches the presynaptic terminus, acetylcholine is released. It diffuses across the synapse and binds reversibly to the postsynaptic receptor, where it initiates a new impulse (neuron) or contraction (muscle).Acetylcholine dissociates from the receptor and is quickly hydrolyzed into acetate and choline in a reaction catalyzed by the enzyme acetylcholinesterase (Scheme 4). Because neither com-

Scheme 4

Phosphofluoridates: Biological Activity and Biodegradation

311

pound is capable of stimulating the receptor, initiation of new signals is terminated. Without the enzymatic hydrolysis, acetylcholine would accumulate uncontrollably and continue to stimulate the receptor. 7.2 Acetylcholinesterase (AChE)

Mammalian acetylcholinesterase located on the cholinergic receptor consists of a tetramer of catalytic and structural subunits. X-ray crystallographic analysis of mammalian acetylcholinesterase indicates that the enzyme exists as a tetramer composed of two dimers. The recombinant enzyme contains 548 amino acid residues; however, posttranscriptional processing of mRNA leads to alternate forms of the enzyme in different tissues. Another form of AChE is found in the red blood cells. The biological function for this enzyme is unknown but probably has some role in hydrolyzing various carboxylic esters. This enzyme does have a prophylactic role against OP poisoning because much of the nerve agent traveling through the blood will be consumed by the irreversible inhibition of AChE. A similar enzyme butyrylcholinesterase (BuChE) is present in the plasma [64]. The active site of this enzyme differs structurally from AChE by having a larger pocket for the acyl group. It hydrolyzes the butyryl (four carbons) analog much better than the acyl (two carbons). In contrast, AChE is almost totally inactive toward butyrylcholine. The catalytic triad of AChE, consisting of ser203, glu334, and his447, sits in the active site at the bottom of a deep gorge through which reactants enter and products depart. The analogous enzyme from the electric eel, Torpedo californica, has a similar structure containing 537 amino acids. The catalytic triad consists of ser200, glu327, and his440. Serine proteases, which catalyze the hydrolysis of carbonyl-nitrogen bonds rather than carbonyl-oxygen bonds have an aspartate amino acid instead of a glutamate [51]. If neurotransmitters are to be effective in start/stop or off/on mechanisms, there must be an immediate maximal release of chemical into the synapse and a very rapid degradation after the receptor has been stimulated. For the cholinergic system, the high synaptic concentration is effected by storing acetylcholine in vesicles at the axon terminal and releasing the contents all at once. As indicated earlier, degradation of acetylcholine is achieved by hydrolysis of the ester. Because carboxylic esters are relatively labile (certainly in comparison to amides), the energy of activation for hydrolysis is relatively low. Therefore, enzymes with very rapid turnover have developed rather than enzymes that bind tightly to the transition structure of the substrate and significantly lower its activation energy. AChE is one of the fastest enzymes known with kcat –1.6¥104 s–1. The proper method for comparing the rate of an enzyme-catalyzed reaction with one occurring in solution is not, however, obvious. One method is to use kcat/Km for the enzyme rate. For the hydrolysis of acetylcholine, the enzyme rate is about 2¥108 M–1 s–1, whereas the rate for OH– attack is 2.2 M–1 s–1. Thus, the enzyme accelerates the rate by about 108, which is essentially the diffusion rate of the substrate through the gorge.

312

J.J. DeFrank · W.E. White

Scheme 5

AChE functions via a two-step mechanism [136]. In the first step, a transesterification reaction, the serine in the active center attacks the acetate moiety and frees the choline. The reaction proceeds through a trigonal bipyramid structure similar to those described for bimolecular nucleophilic substitution in solution. In the second step, water hydrolyzes the acetylated serine via the same mechanism to produce acetic acid and regenerate the enzyme. Acetylcholinesterase is capable of catalyzing this reaction because it generates an internal nucleophile that is much stronger than that available in physiological solutions. An internal proton transfer converts the weakly nucleophilic serine into a strongly nucleophilic serine anion. In AChE from Torpedo californica [160], the catalytic triad effectively transfers the proton on serine200 to an adjacent acidic glutamate residue. The transfer is facilitated (low energy and efficient coupling) by histidine440 in which the proton is accepted by a nitrogen on one side of the imidazole ring and lost by the nitrogen on the other. At the atomic level, it is more likely that the proton resides between histidine440 and glutamate327 rather than being bound exclusively to the glutamate. However, the concept is the same (Scheme 5). The activation energy for the OH– in a solvent cage was estimated to be about 36 kcal M–1.A more representative comparison would use histidine as a base. The value for this reaction is about 27 kcal M–1. Based on molecular simulations, the enzyme reduces the activation energy by 10–15 kcal M–1. This corresponds to an acceleration of 107–1011, which corresponds well with the empirical value of 108. It has been suggested that much of the acceleration is due to stripping the solvent away from the substrate in the active site. Computation studies do not support this hypothesis and indicate that the hydrophobic effect may increase the energy of activation. The catalysis comes principally from a stabilization of the transition structure by providing a polar environment and by reducing the reorganization energy. Successful functioning of the entire cholinergic system depends on the effective operation of the two-step mechanism. If the first step is much more effective than the second (i.e., serine200 is acetylated but not deacetylated), the active site

Phosphofluoridates: Biological Activity and Biodegradation

313

of AChE becomes blocked and the enzyme is no longer capable of further catalysis. Inhibition of acetylcholinesterase leads to the accumulation of acetylcholine in the synapse and continued stimulation of the receptor until the neuron is depolarized or the muscle becomes inactivated. Death usually results from respiratory collapse.

8 Development of Chemicals Designed to Inhibit Acetylcholinesterase The requirement for rapid turnover has led to rather loose binding in the active site and considerable flexibility with respect to substrate structure. Similarly, there is considerable flexibility in designing inhibitors. Irreversible cholinesterase inhibitors have two types of structural limitations – steric and reactive. Suicide substrates are a class of irreversible inhibitors that form a covalent bond through a reaction catalyzed by the enzyme itself and have been discussed by Neilson and Allard in Chap. 7. Compounds that are too reactive may react nonspecifically with other macromolecules or be hydrolyzed before reaching their designed target. Unreactive compounds may never react at all if the activation energy remains too high. In general, the approach is to design a molecule with a minimally reactive leaving group and rely on the enzyme to catalyze the reaction sufficiently for inactivation. For phosphorus compounds designed to inhibit acetylcholinesterase in vivo, the larger halogens such as iodide, bromide, and chloride are normally too reactive. Leaving groups like fluoride (F–), cyanide (CN–), azide (N3–), and thiol (RS–) are optimally active and provide the essential component for chemical warfare nerve agents. Leaving groups based on alcohols, phenols, as well as some thiols are less than optimally reactive and have been used in numerous pesticides because they are considerably less toxic and thereby provide higher safety margins. Although water is sufficiently nucleophilic to hydrolyze carboxylated acetylcholinesterase, it hydrolyzes carbamated and phosphorylated enzymes poorly. The period of inactivation may last dozens of minutes for carbamates and days for organophosphorus compounds. For phosphorylated AChE, restoration of normal neurofunction requires synthesis of new AChE. The chemical industry has exploited the inhibition of AChE by developing several carbamates and phosphates for use as insecticides [66]. Most of the carbamates are solids that are delivered as dusts or wettable powders. Organophosphorus insecticides may be liquids or solids. The most widely used carbamate is carbaryl that is marketed under the tradename Sevin by Union Carbide.

314

J.J. DeFrank · W.E. White

Most of the organophosphorus compounds used as pesticides are phosphates having alkoxy, phenoxy, or thiol leaving groups. Examples are shown below.

Relatively few of the organophosphorus pesticides have fluorine-leaving groups. Mipafox was developed by Fisons Pest Control Ltd and introduced in 1950. It was subsequently removed from the market because of delayed neurotoxicity.

Dimefox appeared the previous year. It is interesting that the pesticides with fluoride leaving groups that reached the commercial market were phosphamides (having P–N bonds) rather than phosphates (having P–O) bonds.

9 Chemical Warfare Agents The most serious application of phosphorus fluoridates that inhibit AChE is for chemical warfare. These compounds have the same mechanism of action as the insecticides; however, they are orders of magnitude more toxic to humans. In January 1937, Gerhard Schrader and associates in the pesticide development program at the German chemical conglomerate, I.G. Farben, discovered the high mammalian toxicity of a compound called tabun [45]. This compound was an organophosphate having a cyano leaving group and ethoxy and dimethylamino groups as modifiers.As prescribed by law, the Wehrmacht chemical laboratory at Spandau in Berlin was informed of this discovery. Further studies confirmed the potential of this new chemical warfare agent and led to the establishment of a production facility at Elberfeld in the Ruhr. Continued experimentation under the direction of the military led to the discovery that compounds having a P–C bond (i.e., phosphonates) were more toxic than the phosphorofluoridates. The first of the phosphonofluoridates having enormous military potential was the isopropyl ester, which Schrader named sarin after the investigators who were involved in its discovery (i.e., Schrader, Ambrose, Rudricker, and van der Linde). Further work led to a less volatile analog, soman, which was never produced in large quantities in Germany.

Phosphofluoridates: Biological Activity and Biodegradation

315

Although the German military produced large quantities of tabun and lesser amounts of sarin, they were never used against Allied forces. The reason for their non-use has perplexed Western leaders, scientists, and journalists for over 50 years. The best answer seems to be the fear of reprisal. There was the strong feeling among the German hierarchy that allied nations knew about their nerve agent research in the 1930s and had probably developed even more toxic agents [27]. Tabun had been described in the literature in 1902. Furthermore, a German patent was issued to I.G. Farben for sarin in 1938 before the start of World War II and the resulting cessation of scientific and technical exchange. If the full military potential of these compounds had been realized in the 1930s, it is unlikely the compounds would have been patented and the information disclosed outside Germany. After the revelation of the toxicity of phosphonofluoridates at the end of World War II, other countries began research programs to discover new agents that would be more effective than those synthesized in Germany. One of those was the cyclohexyl ester that is sometimes designated GF or cyclosarin. Like soman, this compound has a secondary alkoxy moiety and six carbon atoms; however, they are arranged in a closed ring instead of an open chain. GF is somewhat easier to produce than soman because cyclohexanol is more readily available than the soman component, pinacolyl alcohol, whose sale is restricted by the Australia Group agreement and whose production is limited by schedule 2B(14) of the Chemical Weapons Convention. Following the end of the Gulf War, Iraq declared that it had produced and weaponized significant quantities of GF.

9.1 Toxicities

Table 8 [144–147] includes the toxicities of sarin, soman, GF, and DFP for mice and rats by different routes of administration. These values were obtained from the RTEC database. Because the experiments were performed in different laboratories with different protocols using animals with different genetic histories, comparison of the data is difficult and fraught with error. Although there is significant variation in the data, some trends emerge. Soman is the most toxic, followed by GB and GF. DFP, which is frequently used as a simulant for chemical reactivity of sarin and soman, is about 10% as toxic.As would be expected, the i.v. route of administration is the most toxic because the com-

316

J.J. DeFrank · W.E. White

Table 8. Toxicity of chemical agents (LD50 mg kg–1)

Species Rat IP SC IV IM Mouse Skin IP SC IV IM

GB [144]

GD [145]

218 103 39 108

98 71 44 62

1080 283 60 109 164

7800 393 40 35 89

GF [146]

DFP [147]

225

1280 1440

80

1800

400

72000 2450 3000

Table 9. Estimates of toxicity for humans [149]

Agent

Inhalation LCt50 (mg min–1 m3)

Intravenous (mg kg–1)

Percutaneous (mg kg–1)

GA GB GD VX

135 70 70 30

0.018 0.014 – 0.008

14 24.3 5.0 0.143

pound is introduced directly into the circulatory system. Compounds are less toxic when administered percutaneously because the skin provides a moderately effective protective barrier, and some of the agent may evaporate prior to penetration. Considerable effort was expended in analyzing the toxicity data and developing a consistent set of human estimates that are listed in Table 9 [149]. GA and VX are included for comparison even though they don’t contain phosphorus fluorine bonds. 9.2 Quantitative Relation Between Reactivity and Toxicity

If a compound is to be toxic by inhibiting AChE, it must penetrate one or more cellular layers, travel through the circulatory system, penetrate the central nervous system, reach the postsynaptic neuron, and react with the enzyme. Compounds that are too reactive hydrolyze spontaneously before reaching their target and become inactivated. For example, the subcutaneous toxicity (LD50) of the chloride analog of soman is 153 and 185 mg/kg in rats and mice [143], respectively, in contrast to 0.040 mg/kg for soman in mice [105]. The 1000-fold lower toxicity results principally from hydrolysis and other deactivating reactions prior to reaching the critical target. In contrast, compounds that are too stable reach

Phosphofluoridates: Biological Activity and Biodegradation

317

the enzyme intact, but because their activation energy is too high, they fail to react. For many years, there has been considerable qualitative speculation about the relation between reactivity and toxicity. Recently, theoretical approaches were used to generate a quantitative relation [169]. Computational chemistry, at the semiempirical level of theory, was used to calculate the relative reactivity for a series of fluorinated phosphinates, phosphonates, and phosphates. These values were plotted against the i.v. toxicities of the compounds in rats. Toxicities of the compounds examined extended over four orders of magnitude. The plot resembled an inverted parabola with the phosphonates in the center having maximal toxicity and intermediate reactivity. The phosphinates were too reactive and had lower toxicity than the phosphonates. As would be expected, electron-withdrawing groups on the alkyl substituents reduced the reactivity and increased toxicity of the phosphinates. At the other end of the graph, the phosphates displayed the least reactivity and were less toxic than the phosphonates. Electronwithdrawing groups on the alkyl moieties also decreased reactivity but this time reduced toxicity. These results suggest that there is a maximal toxicity that can be achieved by inhibiting cholinesterase with suicide substrates because there is an optimal level of reactivity. Extremely reactive compounds never reach their target in their active form; stable compounds may reach their target intact but can’t do anything about it. It may be possible to increase toxicity slightly by reducing the amount of substrate that is degraded by butyryl cholinesterase, carboxyesterase, or other enzymes; however, it seems unlikely that reactive compounds having much greater toxicities await discovery. It is possible that some reversible inhibitors of cholinesterase may have greater toxicity, but none has appeared thus far. Huperzine and analogs are being studied extensively by empirical and theoretical methods as possible therapeutic agents for Alzheimer’s disease. These compounds are reasonably effective inhibitors of AChE and, because their toxicity is relatively low, provide a good safety index.

Although AChE inhibition can occur throughout the body at any neuro-neuro and neuro-muscular cholinergic receptor, severe toxicity including lethality is thought to result for inhibition within the central nervous system. Therefore, only those agents that are capable of reaching the central nervous system are expected to possess high toxicity.

318

J.J. DeFrank · W.E. White

9.3 Effects of Stereochemistry on Toxicity

Because proteins are formed by the polymerization of L-amino acids, enzymes possess a certain degree of chirality. As a result, one stereoisomer may be a better substrate than the other may. Soman has two chiral centers – one on the phosphorus and one on the alpha carbon atom. Therefore, unresolved soman consists of a mixture of four stereoisomers. Because the absolute configuration at each chiral center was unknown when the isomers were first resolved, they were designated as (+) or (–) to indicate the direction of optical rotation. The liquid nature of soman precluded X-ray crystallography, and the susceptibility of the phosphorus to nucleophilic substitution prevented the unambiguous conversion to solid analogs. The absolute configurations were determined indirectly by matching the observed biochemical properties with the orientation of the OP compounds inside the active site using computational methods [20, 128]. The absolute configurations of the four isomers are indicated below.

As would be expected, the rate or extent of inhibition of cholinesterase differs among the four isomers [24, 25]. The in vitro inhibition of acetylcholinesterase isolated from two species is indicated in Table 10 [24]. The two P (–) isomers inhibit both enzymes rapidly. In contrast, the P (+) isomers are essentially unreactive (five orders of magnitude less). The configuration on the alkoxy chain is less critical. The C (+) configuration is one to five times as active. Table 10. Inhibition of acetylcholinesterase by individual soman isomers [23]

Isomer

Eel AChE

Bovine erythrocyte AChE

C (+) P (–) C (–) P (–) C (+) P (+) C (–) P (+)

(2.8±0.02) ¥108 (1.8±0.1)¥108 <5 ¥103 <5 ¥103

(1.75±0.11) ¥108 (0.27±0.1)¥108 <1¥104 <1¥104

Phosphofluoridates: Biological Activity and Biodegradation

319

Table 11. Toxicities of individual soman isomers [24]

Compound

LD50 (mouse, SC) mg kg–1

Soman (no resolution) C (+) P (–) C (–) P (–) C (+) P (+) C (–) P (+)

156 38 99 ~7000 ~1000

Because the enzymatic reaction proceeds via the trigonal bipyramid, the serine and fluorine moieties must be in the axial positions.As mentioned earlier, the anion pocket where the alkoxy group binds is relatively large and can accommodate a wide variety of structures. In contrast, the esteratic site is so small that groups larger than isopropyl fit very poorly. With the P (–) isomer, the methyl group fits appropriately in the esteratic pocket and the anion site can accommodate the larger pinacolyl group. In the P (+) isomers, the methyl group can fit into the anion site; however, the pinacolyl group is much too large to fit into the esteratic pocket. The molecule could turn around and fit into the active site; however, the fluorine would be on the wrong side – adjacent to the serine rather than opposite. Such binding would inhibit the enzyme reversibly. Table 11 [24] depicts the toxicities (s.c. mouse) for the four isomers. As one would expect from the in vitro data, the C (+) P (–) isomer is the most toxic [23, 24]. The P (+) isomers are considerably less toxic. It is also possible that the toxicity for these two isomers is less than reported because traces of the P (–) isomers may be present because of partial racemization (most likely by nucleophilic substitution with F– to give the other isomer. The toxicity of the unresolved mixture of isomers is less than that for the P (–) isomers because the toxic forms are diluted by the minimally toxic P (+) isomers. A series of compounds developed by Tammelin were very effective at inhibiting AChE in vitro [162]. The phosphonofluoridate moiety provided optimal reactivity for inhibition, and the quaternary amine resembling choline optimized the binding. In vivo toxicity for the choline analogs (i.e., top compounds) when administered directly into the animal (subcutaneously) was higher than corresponding values for sarin.

320

J.J. DeFrank · W.E. White

The tertiary amine (i.e., bottom compounds) degraded too rapidly for accurate toxicology results. The degradation times of the choline and b-methylcholine analogs was 9 and 23 min respectively at pH 8. Under similar conditions, the halflife for sarin hydrolysis is greater than 5 h. The instability, physical properties, and difficulty penetrating membranes limited the choline analogs to laboratory applications. Similar compounds with thiocholine leaving groups (i.e., V agent analogs) are also effective inhibitors of AChE [161]. 9.4 Restoration of AChE Activity

The inhibition of the enzyme occurs because the serine anion is a much stronger nucleophile than water. If less reactive esters or other derivatives of various acids reach the enzyme, the serine may be sufficiently powerful, especially if the acid has a very good leaving group. After the serine is attached, the original leaving group is gone so the regeneration of the enzyme depends on the reactivity of the serine ester bond. Stronger nucleophiles are sometimes capable of reacting with the blocking group and restoring enzyme activity. Oximes have been used for many years to restore activity [170]. The active species is the oximate anion rather than the free amine because only oximes having pKas lower than physiological pHs are effective in restoring activity. As a sidelight, some of the phosphorylated adducts of the oximes used therapeutically spontaneously decompose into nitriles and phosphonates, thereby eliminating the possibility of subsequent enzyme inhibition (See Scheme 6). 9.5 Signs and Symptoms of Organophosphate Poisoning [1, 109]

The most frequent routes of non-laboratory exposure to chemical warfare nerve agents are by inhalation of vapor and aerosols and through the skin (percutaneous) by direct contact of liquid. The properties of the individual agents, particularly vapor pressure, determine the effectiveness of each route of exposure. The vapor hazard from volatile compounds like sarin is significant. In contrast, the vapor pressure of VX is so low that there is minimal hazard associated with the vapor. A low vapor pressure does not however exclude an inhalation hazard because the agent can be disseminated as a toxic aerosol. In contrast, the percutaneous hazard for sarin is somewhat less than for less volatile agents because much of the compound evaporates before it can penetrate the skin.

Scheme 6

Phosphofluoridates: Biological Activity and Biodegradation

321

Exposure to agents can lead to both local effects at the site of exposure and systemic effects resulting from the transport of the agent to other tissues. In general, the nerve agents are fast acting. With high exposure, adverse effects may occur in less than a minute. Frequently the first sign of vapor exposure (especially at low to moderate doses) is miosis or reduction in the size of the pupils in the eye. Normally, miosis does not occur from skin exposure unless the dose is large. Other local effects are rhinorrhea (runny nose) and constriction of the bronchial airways. Exposure to the skin usually leads to signs and symptoms in other parts of the body. Nerve agents increase the secretions in the walls of the GI track and may lead to nausea, vomiting, and to diarrhea if the dose is large. In the skeletal muscle, initial fasciculations and twitching may lead to fatigue and weakness and finally to flaccidity at high doses. A large dose of agent in the central nervous system leads to a loss of consciousness, seizures, and apnea. The effects of low doses are somewhat nonspecific but may include things like forgetfulness, insomnia, or depression. Nerve agents seldom cause confusion or hallucinations. Sometimes the observed signs or symptoms may be the opposite of what is expected. For example, agents normally decrease the heart rate because of stimulation of the vagus nerve; however, other effects (i.e., fright, psychological, hypoxia, adrenergic, stimulation) may increase heart rate.

10 Enzymatic Hydrolysis of Phosphorus Fluoridates Almost as soon as highly toxic phosphorus fluoridates were developed, there was interest in the ability of enzymes to hydrolyze and detoxify them. The earliest work dealing with enzymes capable of catalytically hydrolyzing organophosphorus esters was reported shortly after World War II by Mazur [111]. Enzymes isolated from human and rabbit tissues were shown to enhance the hydrolysis of DFP and thus its detoxification. Using partially purified enzyme preparations from rabbit kidney, he was able to determine that the activity was not related to phosphatase, cholinesterase, or other known esterases. During the 1950s three investigators, Aldridge [6], Augustinsson and Heimburger [13–19], and Mounter et al. [113–118] carried out much of the work in this field. Aldridge reported on what he designated an A-serum esterase from mammalian sources that could hydrolyze paraoxon. Paraoxon is a product of mammalian oxidation of the pesticide parathion, but is considerably more toxic.

This enzyme, more recently referred to as a phosphotriesterase or paraoxonase, was shown to differ from the phosphatases in that phosphatases only hydrolyze monoesters of orthophosphoric acid. In addition, he demonstrated that

322

J.J. DeFrank · W.E. White

his A-serum esterases could be stereospecific, hydrolyzing (+)-sarin, but not the more toxic (–)-sarin. Augustinsson confirmed the findings of Aldridge and extended the work to include enzymes that would hydrolyze organophosphorus compounds such as tabun. He determined that this phosphorylphosphatase or tabunase cleaved the P–CN bond of tabun to release hydrocyanic acid. Using a combination of electrophoretic separation, substrate specificities, and sensitivity to inhibition, he concluded that there were three types of esterase activity in plasma: arylesterase (aromatic esterase,A-esterase), aliesterase (carboxylesterase, B-esterase,“lipase”), and cholinesterase. He noted that there were species variations with respect to the properties of individual enzymes. He also confirmed the observation made by Aldridge that the hog kidney phosphorylphosphatases showed stereoselectivity against tabun. During the same period, Mounter was attempting to purify further and characterize the enzyme from rabbit kidney originally reported by Mazur. He referred to this enzyme as dialkylfluorophosphatase (DFPase, fluorophosphatase). After partial purification by ethanol fractionation, the enzyme was determined to be activated by Co2+ and Mn2+ ions. Reagents that reacted with metal ions, sulfhydryl, or carbonyl groups were found to inhibit DFPase activity. In addition to his work with mammalian enzymes, Mounter et al. were the first to report on the DFPases in microorganisms [114, 118]. Of the bacteria tested, the highest activity was observed with the Gram-negative Proteus vulgaris and Pseudomonas aeruginosa, which were also stimulated by Mn2+. Based upon enzyme activity with different metal ions and inhibitors, it was also demonstrated that there were a number of different DFPases. Studies conducted with preparations from Escherichia coli, Pseudomonas fluorescens, Streptococcus faecalis, and Propionibacterium pentosaceum showed that, while differences were observed in the relative hydrolysis rates of a variety of organophosphorus compounds, they were comparable to those observed with the hog kidney enzyme. During the late 1950s and early 1960s, a number of additional groups became involved in the investigation of the enzymatic hydrolysis of DFP, paraoxon, sarin, and tabun [3–5, 44, 85, 90, 106, 107].While considerable advances were made during this time in comprehending the diversity of enzymes, one of the most significant events was the beginning of the research efforts of Hoskin in this field. Of particular importance is his work, beginning in 1966, on the purification and characterization of a DFPase from squid [86, 87, 89]. The squid enzyme has a molecular weight of approximately 30,000, is found only in cephalopods, requires Ca2+ for activity, and hydrolyzes DFP five times faster than soman. The significance of the squid enzyme lies in the fact that its chemical and biological properties are completely different from all other types of DFPases. To this day, it still appears to belong to a unique class of enzyme. The gene for a squid enzyme has recently been cloned, sequenced, and expressed in Escherichia coli and in the yeast Pichia pastoris by Prof. Heinz Rüterjans, University of Frankfurt [148]. The crystal structure of the enzyme has been determined and is to be reported in the near future.

Phosphofluoridates: Biological Activity and Biodegradation

323

11 Organophosphorus Pesticide Biodegradation The interest in microbial enzymes for the degradation of organophosphorus compounds received a boost in the early 1970s by the isolation of bacteria capable of growing on a variety of pesticides. The initial report was by Sethunathan and Yoshida who isolated a diazinon-degrading Flavobacterium species (ATCC 27551) from rice paddy soil [154]. Cell-free extracts of this organism could also hydrolyze the insecticides chlorpyrifos (diethyl (3,5,6-trichloropyridyl) phosphorothionate) and parathion, the aromatic or heterocyclic products of which were not further metabolized. In 1979, Rosenberg and Alexander [141] described two Pseudomonas isolates capable of hydrolyzing a variety of organophosphorus compounds and using the products as sole phosphorus source. In 1973, a pseudomonad capable of hydrolyzing parathion and utilizing the p-nitrophenol product as a source of carbon and nitrogen was isolated [157]. Somewhat later, Daughton and Hsieh isolated a Pseudomonas stutzeri capable of hydrolyzing parathion from a chemostat culture [52]. Strains of Bacillus and Arthrobacter that can hydrolyze parathion were also described [126] as well as a Pseudomonas capable of utilizing isofenphos as sole carbon and energy source [137]. With the exception of the Flavobacterium, little if any additional information has been reported about any of these organisms or their enzymes.

In addition to the Flavobacterium mentioned above, the other major parathion-degrading bacterium described in the literature is Pseudomonas diminuta MG, which was isolated in 1976 by Munnecke [121, 122]. By far, these two enzymes have been the most widely studied of any capable of hydrolyzing organophosphorus compounds. However, it is worthwhile to note that while the two organisms containing the enzymes were isolated on opposite sides of the world and the plasmids that contain their genes are highly dissimilar, the genes (and enzymes) themselves are virtually identical [74, 112, 119, 151–153]. These phosphotriesterases, parathion hydrolases, or organophosphorus hydrolases are also the only well-characterized enzymes known to hydrolyze catalytically the PS bond of V-agents [94, 138]. This enzyme will be discussed in much more detail in a later section. Beginning in the 1980s, the research efforts in this field have been divided into two major areas: the isolation and characterization of microorganisms (and their enzymes) capable of growth on a variety of organophosphorus pesticides; and

324

J.J. DeFrank · W.E. White

the somewhat more random search for organisms that possess enzymes capable of hydrolyzing DFP and the related nerve agents. The relatively high spontaneous hydrolysis rates of the nerve agents make their use as substrates in microbial enrichment cultures rather problematic. Generally, microorganisms that grow on them actually utilize the hydrolysis products of the agents and do not possess the enzymes to deal with the agents themselves. Examples of the former are the isolation of Pseudomonas alcaligenes C1 that can hydrolyze and grow on fensulfothion [155], the isolation of additional Pseudomonas sp. and other unidentified bacteria that hydrolyze and grow on parathion and/or methyl parathion [32, 119], and the isolation of three distinct bacteria capable of metabolizing coumaphos [156].

12 Nerve Agent Enzymes In the search for nerve agent degrading enzymes, investigations that are more recent have gone in a number of directions. Landis and co-workers examined the ciliate protozoan Tetrahymena thermophila [97–99] and the clam Rangia cuneata [9]. Partial purification of extracts from Tetrahymena revealed that this organism had at least five enzymes with DFPase activity and molecular weights ranging from 67,000 to 96,000. The rate ratios for soman and DFP hydrolysis as well as the effect of Mn2+ on activity varied considerably from one enzyme to another. Preliminary investigations on the clam also resulted in the detection of several DFPases with differing substrate specificity and metal stimulation. Of particular interest was the presence of an enzyme in the clam digestive gland that appeared to have significant activity on the DFP analog mipafox. Most enzymes described to date are either indifferent to mipafox or subject to strong competitive inhibition. Little et al. [103] characterized an enzyme from rat liver that has a substrate preference of sarin>soman>tabun>DFP, but without activity on paraoxon. The enzyme had a molecular weight of 40,000 and was stimulated by Mg2+. Unlike an enzyme from Escherichia coli enzyme, all the stereoisomers of soman appeared to be hydrolyzed at equal rates [28]. A screen of 18 Gram-negative bacterial isolates by Attaway et al. [12] resulted in the finding that, while all showed at least some activity on DFP, only cultures with parathion hydrolase activity showed significant “DFPase” levels. In the mid 1980s, one of the authors (JJD) began an investigation of thermophilic bacteria as sources of DFPases. A Gram-positive, aerobic, spore-forming, rod-shaped, obligate thermophile was isolated from the soil of Aberdeen Proving Ground and found to possess activity against DFP and soman [57]. This isolate, designated as JD-100, was tentatively identified as a strain of Bacillus

Phosphofluoridates: Biological Activity and Biodegradation

325

stearothermophilus. Its temperature range for growth was 40–70 °C, with an optimum of ~55 °C. While the crude cell extracts of JD-100 had low levels of DFP activity [57], the purified enzyme showed no detectable DFP activity but retained soman hydrolyzing activity [41, 88]. The enzyme had a molecular weight of 82,000–84,000 and showed considerable stimulation by Mn2+ (~80-fold). It was unaffected by mipafox and degraded all the stereoisomers of soman.

13 Enzyme Nomenclature As illustrated in the discussion above, the nomenclature of these enzymes has been un-systematic and confusing. In general, the names utilized have been representative of the particular substrate used by an individual investigator. Hence, the literature is filled with references to enzymes such as phosphorylphosphatase, fluorophosphatase, DFPase, paraoxonase, parathion hydrolase, phosphotriesterase, phosphofluorase, somanase, sarinase, and tabunase. In 1992, the Nomenclature Committee of the International Union of Biochemistry and Molecular Biology developed a new nomenclature that could be used until the natural substrates and functions of these enzymes could be determined. Under the general category of Phosphoric Triester Hydrolases (EC 3.1.8), the enzymes are divided into two subgroups. The first, EC 3.1.8.1, Organophosphate Hydrolase (also paraoxonase and phosphotriesterase) is for those enzymes with paraoxon and other P-esters (P–O bonds) as preferred substrates. The second group, EC 3.1.8.2, Diisopropyl-Fluorophosphatase (also Organophosphorus Acid Anhydrolase or OPAA) is for enzymes with a preference for OP compounds with P–F or P–CN bonds.

14 Organophosphorus Hydrolase The Organophosphorus Hydrolase (OPH) (parathion hydrolase, or phosphotriesterase) from Pseudomonas diminuta or Flavobacterium has been one of the most studied enzymes in regards to its activity on nerve agents and pesticides. While many researchers have studied OPH, the primary information on the structure and function of the enzyme has come from the laboratories of Raushel and Wild, both at Texas A&M University. Most early work with the enzyme used the constitutively expressed form which is membrane associated [30]. When the gene for the enzyme was cloned into other hosts, the membrane association remained, making purification difficult. It was discovered that the enzyme was synthesized as a 365 amino acid precursor from which 29 amino acids were removed to generate the mature protein [153]. When this 29 amino acid leader sequence was removed from the clone, the recombinant enzyme was found as a soluble mature enzyme that maintained activity [153]. The enzyme has been expressed in a variety of hosts including insects [134], insect cells [54], fungi [53], and Streptomyces [142, 159]. The mature enzyme is a ~36,000 metalloprotein with two Zn2+ ions present in the native enzyme. However, a variety of other divalent metal ions (Mn2+, Cd2+,

326

J.J. DeFrank · W.E. White

Table 12. Comparison of hydrolytic constants for OPH [94]

Substrate

Bond type

kcat (s–1)

Km (mmol/l)

kcat/Km (M–1 s–1)

Paraoxon Parathion DFP Sarin Soman Demeton-S VX Acephate

P–O P–O P–F P–F P–F P–S P–S P–S

3170 630 465 56 5 1.25 0.3 2.8

0.058 0.24 0.048 0.7 0.5 0.78 0.44 160

5.5 ¥107 2.6¥106 9.7¥106 8.0 ¥104 1.0¥104 1.6 ¥103 45 18

Assays conducted at pH 7.2 and 37 °C.

Co2+, or Ni2+) can be substituted [127]. The Co2+ enzyme has the greatest activity on substrates with P–F and P–S bonds [94]. Originally thought to function as a monomeric enzyme, more recent information based on the crystal structure of the protein indicates that the active form is actually a homodimer [21]. OPH has a broad pH profile with optimum activity between pH 8 and 10 and a temperature optimum of ~50 °C [30]. OPH has the ability to hydrolyze a wide variety organophosphorus pesticides [42] as well as other compounds having P–O, P–F, and P–S bonds. Several selected substrates and the kinetic constants for OPH (Co2+ form) are shown in Table 12 [94]. The hydrolytic mechanism of OPH was originally believed to involve an SN 2 mechanism where an active site base of the enzyme abstracts a proton from a water molecule, thus making it a nucleophile that directly attacks the phosphorus in the substrate [102]. Based on the crystal structure of the enzyme [21, 22] it is now thought that the active site base is not required. The binuclear metal center of the enzyme is located at the C-terminus of a b-barrel with the metals separated by 3.8 Å. There are two bridging ligands to the metals: a water molecule (or possibly a hydroxide ion) and a carbamylated lysine residue. The bridging water molecule probably acts as the nucleophile in the reaction with the substrate by single in-line displacement attack [22]. The natural or original function of OPH is still unknown. So far, this type of enzyme has been restricted to prokaryotic organisms. However, with fluorescent in situ hybridization (FISH), cDNAs encoding proteins with a high degree of homology to OPH have been identified in mice [92], rats [55], and humans [7]. Site-directed mutagenesis of OPH has resulted in modification of its activity on a variety of substrates [96, 166]. Both the catalytic activity (kcat) and substrate specificity (kcat/Km) of OPH has been changed as well as its stereospecificity against chiral substrates. Increases in activity of 20–40-fold have been achieved with soman and to a lesser extent with DFP, VX, and acephate.

Phosphofluoridates: Biological Activity and Biodegradation

327

15 Halophilic Bacterial OPAAs Work began in the late 1980s on identifying OPAA enzymes in halophilic bacteria. The rationale behind this was based on the intent to develop enzyme-based decontamination systems that would use any available water, to include seawater and other saline sources. Isolates were obtained from water and soil samples of salt springs in the state of Utah. In particular, one isolate, designated JD6.5, was obtained from Grantsville Warm Springs, which has a relatively constant temperature of 24–32 °C and 25,000 ppm dissolved solids (96% NaCl). The isolate is a Gram-negative, aerobic short rod, and an obligate, moderate halophile. It required at least 2% NaCl for growth, with an optimum between 5 and 10% NaCl. Fatty acid analysis of the isolate identified it as a strain of Alteromonas, a common genus of marine bacteria. An intracellular OPAA enzyme from strain JD6.5 was purified and characterized [59]. It is a single polypeptide with molecular weight of 58,500. It has a pH optimum of 8.5 and temperature optimum of 50 °C. Maximum activity is seen with either Mn2+ or Co2+. The enzyme is inhibited by iodoacetic acid, p-chloromercuribenzoate, and N-ethylmaleimide, indicating that a sulfhydryl group is either essential for activity or in close proximity to the active site. It is subject to competitive, reversible inhibition by mipafox and is significantly stimulated by NH4+ ions (three- to fivefold). Its catalytic activity towards nerve agents and related compounds is shown in Table 13 [37]. The activity toward soman corresponds to a 109-fold increase in the rate of reaction compared to its spontaneous hydrolysis rate. It was also one to two orders of magnitude greater than for any other known enzyme. Having identified isolate JD6.5 as a strain of Alteromonas, a number of other Alteromonas strains were obtained from the laboratory of Dr. Rita Colwell, University of Maryland, and the American Type Culture Collection. These strains were evaluated for enzyme activity against DFP and the nerve agents [58]. Several showed high levels of activity and two of the enzymes were purified and characterized [36, 37]. Table 14 shows how these enzymes compare to the JD6.5

Table 13. OP substrate specificity for Alteromonas sp. JD6.5 OPAA [37]

Substratea

kcat (s–1)

GD (soman) DFP GF GB (sarin) Paraoxon GA (tabun) VX

3145 1820 1654 611 124 85 0

a

The activity on the substrates with fluoride leaving groups was measured with a fluoride ionselective electrode method [89]. Activity on paraoxon was determined by measurement of the increase in absorbance at 405 nm representing the release of the p-nitrophenol group. For GA (tabun), cleavage of the P-CN bond was determined by P31-NMR.

328

J.J. DeFrank · W.E. White

Table 14. Comparison of Alteromonas OPAAs

OPAA Property

Alteromonas sp. JD6.5

Alteromonas haloplanktis

Alteromonas undina

Molecular weight pH Optimum Temperature opt. (°C) Metal requirement Substrate specificity (kcat) DFP GA GB GD GF

60,000 8.5 50 Mn=Co

50,000 7.5 40 Mn

53,000 8.0 55 Mn

1820 85 611 3145 1654

691 255 308 1667 323

1403 368 426 2826 1775

strain OPAA in physical and biochemical properties.As can be seen, there are significant similarities as well as differences with these enzymes. In order to produce the enzymes in larger quantities, the genes for Alteromonas sp. JD6.5 and Alt. haloplanktis were cloned into Escherichia coli and expressed [35, 37]. In addition, the gene sequences were determined and translated into an amino acid sequence. The 10,000 Dalton molecular weight difference between the Alteromonas sp. JD6.5 and Alt. haloplanktis OPAAs was found to be due to the presence of an extended C-terminal region in the JD6.5 enzyme. The two enzymes were found to have a 77% amino acid homology. If the extended C-terminus of strain JD6.5 was excluded, the homology increased to ~90%. Previously, it had been assumed that the natural function of the OPAAs would have something to do with phosphorus metabolism (phospholipase, phosphodiesterase, etc.). Therefore, it came as a considerable surprise when the results of screening the amino acid sequence of Alteromonas sp. JD6.5 against the NCBI protein data base revealed a high degree of homology (48%) to the Escherichia coli X-Pro dipeptidase. Two other matches were for E. coli aminopeptidase P Table 15. Alteromonas OPAA dipeptidase profile [37]

Substrate

Leu-Pro Ala-Pro Leu-Ala Gly-Glu Met-Asn Ala-Ala Pro-Leu Pro-Gly Gly-Pro-Ala Ala-Pro-Phe

Specific activity (µmoles min–1 mg protein–1) Altermonas sp. JD6.5

Altermonas haloplanktis

Altermonas undina

636 510 82 <1 <1 <1 <1 <1 <1 <1

988 725 63 <1 <1 <1 <1 <1 <1 <1

810 658 220 1391 410 105 <1 <1 <1 <1

Phosphofluoridates: Biological Activity and Biodegradation

329

(31% homology) and Lactobacillus sake dipeptidase (19% homology). There was no homology observed between the OPAA and the Flavobacterium and Pseudomonas diminuta phosphotriesterase (OPH). X-Pro dipeptidases, also known as Prolidases (EC 3.4.13.9), are a ubiquitous class of enzymes that hydrolyze dipeptides with a prolyl residue at the carboxylterminal position. They are usually activated by Mn2+, are possibly thiol dependent, and usually do not act on tri- or tetrapeptides or dipeptides with proline at the N-terminus. They generally have a molecular weight of 40,000–50,000, a temperature optimum between 40 and 55 °C, and a pH optimum between 6.5 and 8.0. All these properties are very similar to those of the Alteromonas OPAAs. The three OPAAs listed in Table 14 were also tested against a variety of di- and tripeptides by measuring the release of amino acids by a modified Cd-ninhydrin method [62]. The results of these assays are shown in Table 15 [37]. The results clearly indicate that the Alteromonas OPAAs are prolidases rather than aminopeptidases. Another property of the prolidases is that they contain a conserved region with three histidine residues (Table 16, Alteromonas sp. JD6.5 amino acid residues 331–348) as well as two smaller, highly conserved regions nearer the C-terminus, the larger of which is shown in Table 17 (Alteromonas sp. JD6.5 amino acid residues 371–388). The residues that are identical in all the genes are boxed. Highly conserved or conservative substitutions are in uppercase bold text. Residues in lowercase are mismatched. Hoskin and Walker [91] examined several nerve agent/DFP-hydrolyzing enzymes to determine whether any of these might also be prolidases. Rather than measure dipeptide hydrolysis directly, they reasoned that if a DFP-hydrolyzing enzyme also hydrolyzes Leu-Pro, then Leu-Pro should inhibit the hydrolysis of DFP. Their results are shown in Table 18 [91]. This result demonstrated that the hog kidney and the Escherichia coli OPAAs most likely are prolidases and that the squid-type OPAA and the Pseudomonas diminuta OPH are not. This does not eliminate the possibility that the squid-type enzyme may be a peptidase with different substrate specificity. However, in the case of OPH, the fact that it has no sequence homology suggests that it has a very different natural function than the prolidases. Until the sequence of the squidtype enzyme has been published, a similar statement cannot be made. Since the Alteromonas prolidases have high levels of activity against the G-type nerve agents, the question arises whether other prolidases have OPAA activity as well. In the case of mammalian enzymes, partially purified human prolidase and porcine liver prolidase were obtained from Dr. Lin Liu, ChemGen Corporation, and Sigma Chemical Co. respectively. Both had low levels of activity against DFP and G-agents, in the range of 1/200–1/500th of that seen with the OPAAs [37]. Their activity on X-Pro dipeptides was comparable to that observed with the OPAAs. At the other extreme, a preparation of recombinant prolidase from the hyperthermophile Pyrococcus furiosus was obtained from Dr. Michael Adams, University of Georgia [67]. It was determined that it had measurable, but low levels of DFP activity at 80 °C. The enzyme activity was measured well below the optimum of this enzyme (100 °C), but at the maximum temperature that the fluoride

P P P P m i l p

H H H H H H H H

Alteromonas sp. JD6.5 OPAA (Prolidase) Alteromonas haloplanktis OPAA (Prolidase) Halophile JD30.3 OPAA Escherichia coli PepQ (Prolidase) Escherichia coli Aminopeptidase P Methanococcus jannaschii (Prolidase) Streptomyces lividans Aminopeptidase P-1 Human Prolidase

k l d i v l t h I I V L L L L L

Table 17. Conserved Region no. 2 in Prolidases/Peptidases

Alteromonas sp. JD6.5 OPAA (Prolidase) Alteromonas haloplanktis OPAA (Prolidase) Halophile JD30.3 OPAA Escherichia coli PepQ (Prolidase) Escherichia coli Aminopeptidase P Methanococcus jannaschii (Prolidase) Streptomyces lividans Aminopeptidase P-1 Human Prolidase

Table 16. Conserved region no. 1 in prolidases/peptidases

E E E q E k E q

G G G G G s G G

A k A p p e p p

L L L I L L t L

N N g g g g g g

G G G G s G G G

Q Q Q m m m m m

H H H H H H H H

V V V V V V V V

H H f p w g m f

F F F l l v l l

I L I I L V L L

T T T T T T T T

G G G G G G G G

I I I I V I V V

L A L L L L m I

E E E E E E E E

Q G G G d e d d

P P P P P P P P

V V V V V V V V

G G G G G G G G

H H H H H H H H

L L L I L L L I

D D D D D E D D

Y Y Y Y Y Y Y Y

V V V V V e c V

F F v F I l F F

G G G A G p A G

I I V I A k q I

G G G G V r A G

D D D E p D a D

l a Y

F F F F

330 J.J. DeFrank · W.E. White

331

Phosphofluoridates: Biological Activity and Biodegradation

Table 18. Effect of Leu-Pro on DFP hydrolysis by purified nerve gas/DFP hydrolyzing enzymes in relation to other properties [91]

Enzyme source

% Inhibition by Leu-Pro

Soman/DFP hydrolysis ratio

Mn2+ stimulation

Loligo pealei (squid) Pseudomonas diminuta OPH Hog kidney Escherichia coli OPAA Alteromonas

–3, –5 a 0, 2, 3 93, 93 72, 76 55, 68

0.2–0.25 0.125 ~5 ~50 ~2

0 0 ~5X ~5X Yes

a

Negative values = stimulation.

electrode could tolerate. For safety reasons, the high temperature tests were not repeated with the nerve agents. This enzyme is a homodimer with 39,400 MW subunits and a pH optimum of 7.0. In regards to metal requirements, Co2+ is preferred 4:3 over Mn2+. The preferred substrates are Met-Pro and Leu-Pro. Comparison of the amino acid sequence of this enzyme with other prolidases found the greatest degree of similarity with that from another archeon, Methanococcus jannaschii (69%). Considerable similarity was also found with other prolidases for which the sequence is known: Lactobacillus delbrueckii (61%), Haemophilus influenzae (58%), Escherichia coli (56%), human (53%), and Alteromonas sp. JD6.5 (51%). Based on the crystal structure of the E. coli methionine aminopeptidase, five amino acids (Asp97, Asp108, His171, Glu204, and Glu235) were shown to coordinate the binding of two Co2+ ions per active site [140].All these prolidases conserved the same five residues, even though some of the enzymes are monomers instead of dimers. In the case of strain JD6.5, the conserved residues are Asp244, Asp255, His336, Glu381, and Glu420. The question naturally comes up as to why prolidases are such efficient catalysts for the hydrolysis of organophosphorus compounds, in particular, for the nerve agents. Molecular modeling studies comparing the structures of soman and Leu-Pro have been carried out [11]. It was determined that the three-dimensional structure and the electrostatic density maps of the materials look nearly identical. The organophosphorus compounds such as soman appear to fit into the active site of the enzyme in an orientation that allows the hydrolysis of the target P–F, P–CN, or P–O bond. Using the crystal structure of the Escherichia coli methionine aminopeptidase, a postulated model of the Alteromonas sp. JD6.5 active site has been developed. It indicates two hydrophobic pockets, a large one where the side chain of the leucine (of Leu-Pro) can fit and a smaller one where the proline ring fits. The amide bond is positioned for a backside attack just above the two metal ions that have a bridging oxygen or hydroxyl group. When soman is substituted for Leu-Pro, the pinacolyl group fits into the large hydrophobic pocket and the methyl group into the small pocket. The phosphorus atom is located in the same position as the amide bond of Leu-Pro and the fluorine leaving-group extends out of the active site. The model suggests that the activity of the dipeptidases on soman and related compounds is primarily a matter of serendipity. They mimic the structure of the natural substrates for the

332

J.J. DeFrank · W.E. White

enzyme so well that the enzymes are able to catalyze their hydrolysis efficiently. Confirmation of this proposition will come when the crystal structure of the Alteromonas sp. JD6.5 enzyme has been determined.

16 Enzyme Production and Applications As mentioned earlier, recombinant OPH has been produced in a variety of host organisms. However, in most cases the level of production has been relatively low at 10–25 mg/l. In some instances, it is believed that the enzyme forms insoluble and inactive inclusion bodies, but this has not yet been confirmed. A variety of techniques such as varying growth temperature, varying the carbon sources in microbial fermentations, fed-batch fermentations, and others are being pursued. The use of a host that will secrete the active enzyme may also solve this problem, although production by Streptomyces was still quite low [142, 159]. The Alteromonas sp. JD6.5 prolidase has proven to be quite amenable to production by recombinant DNA technology. In initial studies, the opaA gene encoding the prolidase was cloned into pBluescript SK+ (pTC6513) and expressed in Escherichia coli. The expressed enzyme constituted about 5% of the total cell protein. To enhance production further, the gene was cloned downstream of a strong trc promoter in a high-level, regulated expression vector, pSE420 (Invitrogen, San Diego, California). After induction with IPTG, the enzyme was produced at levels up of to 50–60% of total cell protein for a yield in shake-flask cultures of 150–200 mg/l [38]. Fermentation studies with fed-batch systems have pushed production levels to nearly 1 g/l [33]. 16.1 Field Decontamination

In order to be useful in the decontamination of nerve agents during military operations or after terrorist attacks, the enzyme-based formulation needs to be stable for long periods and easy to use. The Alteromonas sp. JD6.5 prolidase has been lyophilized in the presence of trehalose (a-D-glucopyranosyl-a-D-glucopyranoside) and stored for extended periods at room temperature with no apparent loss of activity [34]. In the absence of trehalose, lyophilized enzyme lost >90% of its activity. The enzyme-based decontamination formulation is planned to be reconstituted in whatever water-based system the user has available. The systems being considered include fire-fighting foams and sprays, aqueous degreasers, laundry detergent, aircraft deicing solutions, etc. Table 19 [38] shows the effect of a variety of these types of materials on the activity of the prolidase. The materials were evaluated at the shown concentrations. OPH has been the workhorse for a number of groups in evaluating activity in unusual environments. Russell has reported on the development of water-in-oil microemulsions that maintain significant enzymatic activity. Both the prolidase and OPH have been shown to be active in generated foams [33, 101]. The use of foams offers several advantages. The foams are generally

333

Phosphofluoridates: Biological Activity and Biodegradation Table 19. DFP hydrolysis by Alteromonas sp. JD6.5 prolidase in different matrices [38]

Matrix (source)

Normal function

Conc. Spec. Act. (%) (U mg–1) a

Control (buffer only) AFC-380 (Sandia National Lab.; NM) AFFF (3 M; St. Paul, MN) BioSolve (Westford; Westford, MA) ColdFire (FireFreeze; Rockaway, NJ) Silv-EX (Ansul; Marinette, WI) Blue Base (Neutron; Torrance, CA) BV406LF (FireFreeze; Rockaway, NJ) Green Thunder (Jackson, MI) SC-1000 (Gemtek; Phoenix, AZ) Star Clean. Miracle (Hudsonville, MI) Supersolve (Gemtek; Phoenix, AZ) Odor Seal (FireFreeze; Rockaway, NJ)

– Modified fire-fighting foam Fire-fighting foam Fire-fighting wetting agent Fire-suppressing agent Fire-fighting foam Degreaser Degreaser/cleaner Degreaser/cleaner Wetting agent/ degreaser Wetting agent/oil removing Wetting agent/ degreaser Wetting agent/odor removing

– 6 6 6 10 6 8 10 10 5 30 10 10

a

1950 1050 460 1030 2340 320 140 1430 250 650 670 440 1980

The reaction medium contained 50 mmol/l (NH4)2CO3 (pH 8.7), 0.1 mmol/l MnCl2, 3 mmol/l DFP, and 0.3–0.4 units of enzyme in a total volume of 2.5 ml. One unit (U) of enzyme activity is defined as the release of 1.0 mmole of F – min–1.

made up of surface-active agents that may help in the solubilization of the substrates, and the foam will stick to vertical surfaces for sufficient time to allow the enzyme action. As can be seen in Table 19, there is considerable variation in activity, but even in the systems where inhibition occurs, the residual activity may be sufficient to carry out the necessary decontamination. It should be noted that the enzyme was not optimized for use in any of these materials and that considerable enhancement may be possible. In addition to these liquid matrices, the prolidase and OPH have been immobilized in polyurethane foams where they retained significant activity [33, 100]. The immobilized enzymes were considerably more stable than the free enzyme. This offers the potential use of the enzymes in sponges or wipes for the decontamination of personnel (including casualties) and small sensitive equipment. The potential for enzymes to function on a large scale against nerve agents has been demonstrated under the auspices of NATO Project Group 31 that deals with the development of “Non-Corrosive, Biotechnology-Based Decontaminants for Chemical and Biological Agents.” Several successful trials with soman and VX have been conducted in France, Germany, and the United Kingdom [56]. These trials have used enzymes in a variety of matrices (foams, sprays, and microemulsions) against agents on both porous and hardened painted surfaces. Excellent removal and hydrolysis of the agents was observed. 16.2 Prophylaxis and Therapy

In addition to their potential for decontaminating nerve agents and pesticides on external surfaces or in liquid streams, there has been considerable interest in enzymes medical applications. A group in Israel has shown that an OPH-type en-

334

J.J. DeFrank · W.E. White

Table 20. Protective effect of erythrocyte encapsulated OPH against Paraoxon in mice [130]

OPH/CRBC (i.v.)

Atropine (i.p., mg kg–1)

2-PAM (i.p., mg kg–1)

LD50 (i.v., mg kg–1)

No enzyme (control) No enzyme No enzyme No enzyme CRBC CRBC CRBC CRBC

– 10 – 10 – 10 – 10

– – 90 90 – – 90 90

0.95 2.28 5.67 53.67 119.9 594.3 597.7 991.2

zyme obtained from a strain of Pseudomonas was able to protect mice (7–26 µg injected i.v.) from multiple median lethal doses of paraoxon (3.8–7.3 LD50s) and diethyl phosphorofluoridate 2.9 LD50s) [10]. These studies were then expanded to demonstrate protection of up to 6.65 LD50s of tabun by this enzyme [139]. A similar result was observed at the U.S. Army Medical Research Institute for Chemical Defense with the Pseudomonas diminuta OPH. With subcutaneous injection of this enzyme (at 0.1 mg of enzyme per gram body weight), mice were protected against 2 LD50s of soman [29]. This is even more impressive when it is recognized that OPH does not have very significant activity against soman. A difficulty that will be faced when using any foreign protein/enzyme as a prophylactic or therapeutic is the potential for undesirable physiological responses and immunological reactions. In addition, rapid clearing of the enzyme from the blood stream may negate its potential benefits. One approach to deal with these difficulties is to encapsulate the enzyme in naturally occurring vehicles. OPH was encapsulated in murine erythrocytes (CRBC) by means of hypotonic dialysis and resealing. As shown in Table 20 [130], significant protection (100-fold) against paraoxon was observed [129, 130] when only the encapsulated enzyme was used. When combined with the classical antidotes for organophosphorus poisoning, atropine and pralidoxime (2-PAM), a synergistic effect was obtained and the protection increased to 1000-fold. Presumably, the atropine and 2-PAM are acting to protect/reactivate the cholinesterases while the OPH was scavenging the paraoxon in the bloodstream. While the use of erythrocytes as carriers for enzymes was quite successful, they do have problems in regards to production and long-term storage. An alternative approach is to encapsulate or entrap the enzyme in a synthetic matrix that will protect it from clearance by the body and render it invisible to the immune system. Sterically stabilized liposomes have been used with both OPH [132] and OPAA [131, 133]. In the case of OPAA, the studies were conducted against DFP. As in the example above dealing with the OPH in erythrocytes, the liposomes with OPAA provided significant protection (9.6 LD50s) when used alone. This is illustrated in Table 21 [131]. In combination with atropine and 2-PAM, a synergistic effect was observed with the protection increased to

335

Phosphofluoridates: Biological Activity and Biodegradation

Table 21. Protective effect of liposome (SL) encapsulated OPAA against DFP in mice [131]

OPH/CRBC (i.v.)

Atropine (i.p., mg kg–1)

2-PAM (i.p., mg kg–1)

LD50 (i.v., mg kg–1)

No enzyme (control) No enzyme No enzyme No enzyme SL SL SL SL

– 10 – 10 – 10 – 10

– – 90 90 – – 90 90

4.2 5.7 7.7 29.3 9.6 18.9 21.1 98.6

23.2 LD50s. By using several enzymes with differing specificities and activities, protection against a wide range of organophosphorus compounds can be achieved. 16.3 Detection

Numerous detection systems utilizing cholinesterase inhibition (AChE and BuChE) have been developed and in many cases commercialized. The advantage of such systems is that they can be very sensitive to anything that will inhibit the enzyme. However, since they will react with neurotoxins such as carbamate pesticides and other materials, these sensors are not very selective. In addition, they are primarily single-use systems. Even inclusion of 2-PAM to reactivate the enzymes will only be partially successful. The use of catalytic enzymes such as OPAA, OPH, squid DFPase, and others could result in sensors that not only give an alarm when an agent or pesticide is present, but also can generate an all-clear signal when the threat has passed. Being catalytic, the enzymes are not destroyed in the process. Additionally, if used in parallel or on an array, the differences in specificity will allow discrimination between substrates, thus giving identification as well as detection. The primary difficulty that has been faced in developing such sensors has been in the conversion of the enzyme reaction to a signal that can be picked up optically or electronically. OPH has been immobilized on a nylon membrane and attached to a fiber-optic probe [120].With substrates such as parathion, paraoxon, and coumaphos, the products of the OPH reaction (p-nitrophenol and chlorferon) have significant extinction coefficients and are readily detected at levels as low as 2 mmol/l in 2 min. The system can work with any chromogenic substrate for OPH or other enzymes that may be used. An alternative method taken with both OPH and OPAA is the detection of the proton released in the course of the P–F bond cleavage [158]. In one approach, OPAA was covalently immobilized on silica gel and applied to the surface of a flat pH electrode for use in a batch-mode system. In the second approach, OPAA was immobilized directly to one gate of a pH-sensitive field effect transistor (FET)

336

J.J. DeFrank · W.E. White

and used in a flow cell (measurements taken in a stop-flow mode). The enzyme was shown to function well in both systems and DFP concentrations of 25 and 20 µmol/l were detected respectively. Because of its different substrate specificity, it could be readily distinguished from OPH. New and modified enzymes with increased activities, broader specificities, changes in pH optimum, and resistance to certain metals and inhibitors are being pursued.As they are discovered or developed, these enzymes will have a profound effect on the applications described here. They may lead to entirely new technologies involving organophosphorus/fluorine chemistry.

17 References 1. U.S. Army Medical Research Institute of Chemical Defense (1995) Medical management of chemical casualties handbook. U.S. Army Medical Research Institute of Chemical Defense, Aberdeen Proving Ground, MD 2. Weast RC (ed) (1996) CRC handbook of chemistry and physics. CRC Press, Boca Raton 3. Adie PA (1956) Purification of sarinase from bovine plasma. Can J Biochem Phys 34:1091–1094 4. Adie PA, Hoskin FCG, Trick GS (1956) Kinetics of the enzymatic hydrolysis of sarin. Can J Biochem Phys 34:80–82 5. Adie PA, Tuba J (1958) The intracellular localization of liver and kidney sarinase. Can J Biochem Phys 36:21–24 6. Aldridge WN (1953) Serum esterase 1. Two types of esterase (A and B) hydrolyzing p-nitrophenyl acetate, propionate and butyrate and a method for their determination. Biochem J 52:110–124 7. Alimova-Kost MV, Imreh S, Buchman VL, Ninkina NN (1998) Assignment of phosphotriesterase-related gene (pter) to human chromosome band 10p12 by in situ hybridization. Cytogen Cell Gen 82:16–17 8. Allcock HR (1972) Theories of bonding in phosphazenes. Phosphorus-nitrogen compounds. Academic Press, New York, pp 16–32 9. Anderson RS, Durst HD, Landis WG (1988) Initial characterization of an OPA anhydrase in the clam, Rangia cuneata. Comp Biochem Phys 91C:575–578 10. Ashani Y, Rothschild N, Segall Y, Levanon D, Raveh L (1991) Prophylaxis against organophosphate poisoning by an enzyme hydrolysing organophosphorus compounds in mice. Life Sci 49:367–374 11. Ashman WP (personal communication) 12. Attaway H, Nelson JO, Baya AM,Voll MJ, White WE, Grimes DJ, Colwell RR (1987) Bacterial detoxification of diisopropyl fluorophosphate. Appl Environ Microbiol 53(7): 1685–1689 13. Augustinsson K-B (1954) The enzymatic hydrolysis of organophosphorus compounds. Biochem Biophys Acta 12:303–304 14. Augustinsson K-B, Heimburger G (1954) Enzymatic hydrolysis of organophosphorus compounds I. Occurrence of enzymes hydrolyzing dimethylamido-ethoxy-phosphoryl cyanide (tabun). Acta Chem Scand 8(5):753–761 15. Augustinsson K-B, Heimburger G (1954) Enzymatic hydrolysis of organophosphorus compounds II. Analysis of reaction products in experiments with tabun and some properties of blood plasma tabunase. Acta Chem Scand 8(5):762–767 16. Augustinsson K-B, Heimburger G (1954) Enzymatic hydrolysis of organophosphorus compounds IV. Specificity studies. Acta Chem Scand 8(9):1533–1541 17. Augustinsson K-B, Heimburger G (1955) Enzymatic hydrolysis of organophosphorus compounds V. Effect of phosphorylphosphatase on inactivation of cholinesterases by organophosphorus compounds in vitro. Acta Chem Scand 9(2):310–318

Phosphofluoridates: Biological Activity and Biodegradation

337

18. Augustinsson K-B, Heimburger G (1955) Enzymatic hydrolysis of organophosphorus compounds VI. Effect of metallic ions on the phosphorylphosphatases of human and swine kidney. Acta Chem Scand 9(3):383–392 19. Augustinsson K-B, Heimburger G (1957) Enzymatic hydrolysis of organophosphorus compounds VII. The stereospecificity of phosphorylphosphatases. Acta Chem Scand 11(8):1371–1377 20. Bencsura A, Enyedy IY, Kovach IM (1996) Probing the active site of acetylcholinesterase by molecular dynamics of it’s phosphonate ester adducts. J Am Chem Soc 118(36):8531–8541 21. Benning MM, Kuo JM, Raushel FM, Holden HM (1994) Three-dimensional structure of phosphotriesterase: an enzyme capable of detoxifying organophosphate nerve agents. Biochemistry 32:15,001–15,007 22. Benning MM, Kuo JM, Raushel FM, Holden HM (1995) Three-dimensional structure of the binuclear metal center of phosphotriesterase. Biochemistry 34:7973–7978 23. Benschop HP, Bijleveld EC, Jong LAPD, Wiel HJVD, Helden HPMV (1987) Toxicokinetics of the four stereoisomers of the nerve agent soman in atropinized rats – influence of a soman stimulator. Toxicol Appl Pharm 90:490–500 24. Benschop HP, Konings CAG, Genderen JV, Jong LPAD (1984) Isolation, anticholinesterase properties, and acute toxicity in mice of the four stereoisomers of the nerve agent soman. Toxicol Appl Pharm 72:61–74 25. Benschop HP, Konings CAG, Jong LPAD (1981) Gas chromatographic separation and identification of the four stereoisomers of 1,2,2-trimethylpropyl methylphosphonofluoridate (soman). Stereospecificity of in vitro “detoxification” reactions. J Amer Chem Soc 103(14):4260–4262 26. Black RM, Harrison JM (1996) The chemistry of organophosphorus chemical warfare agents. In: Hartley FR (ed) The chemistry of organophosphorus compounds, vol 4.Wiley, chap 10, pp 781–840 27. Borkin J (1978) The crime and punishment of I.G. Farben. Barnes and Noble, New York 28. Broomfield CA (personal communication) 29. Broomfield CA (1992) A purified recombinant organophosphorus acid anhydrase protects mice against soman. Pharm Toxicol 70:65–66 30. Brown KA (1980) Phosphotriesterases of Flavobacterium sp. Soil Biol Biochem 12:105–112 31. Chambers JE, Levi PE (eds) (1992) Organophosphates: chemistry, fate, and effects. Academic Press, San Diego 32. Chaudhry GR, Ali AN, Wheeler WB (1988) Isolation of a methyl parathion-degrading Pseudomonas sp. that possesses DNA homologous to the opd gene from a Flavobacterium sp. Appl Environ Microbiol 54(2):288–293 33. Cheng T-C (personal communication) 34. Cheng T-C, Calomiris JJ (1996) A cloned bacterial enzyme for nerve agent decontamination. Enz Microb Technol 18:597–601 35. Cheng T-C, Harvey SP, Chen GL (1996) Cloning and expression of a gene encoding a bacterial enzyme for decontamination of organophosphorus nerve agents and nucleotide sequence of the enzyme. Appl Environ Microbiol 62(5):1636–1641 36. Cheng T-C, Harvey SP, Stroup AN (1993) Purification and properties of a highly active organophosphorus acid anhydrolase from Alteromonas undina. Appl Environ Microbiol y 59(9):3138–3140 37. Cheng T-C, Liu L, Wang B, Wu J, DeFrank JJ, Anderson DM, Rastogi VK, Hamilton AB (1997) Nucleotide sequence of a gene encoding an organophosphorus nerve agent degrading enzyme from Alteromonas haloplanktis. J Ind Microbiol Biotechnol 18:49–55 38. Cheng T-C, Rastogi VK, DeFrank JJ, Sawiris GP (1998) G-type nerve agent decontamination by Alteromonas prolidase. Annals NY Acad Sci 864:253–258 39. Cherkasov RA, Ovchinnikov VV, Pudovik MA, Pudovik AN (1982) The reactivity of 1,3,2diheteropholans and 1,3,2-diheterophosphorinanes with a tetracoordinate phosphorus atom. Russian Chem Rev 51(8):746–764

338

J.J. DeFrank · W.E. White

40. Cherkasov RA, Polezhaeva NA (1987) 1,3,2-diheterophosphacyclanes containing a hexacoordinate phosphorus atom. Russian Chem Rev 56(2):163–181 41. Chettur G, DeFrank JJ, Gallo BJ, Hoskin FCG, Mainer S, Robbins FM, Steinmann KE, Walker JE (1988) Soman hydrolyzing and detoxifying properties of an enzyme from a thermophilic bacterium. Fund Appl Toxicol 11:373–380 42. Chiang T, Dean MC, McDaniel CS (1985) A fruit fly bioassay with phosphotriesterase for detection of certain organophosphorus insecticide residues. Bull Environ Contam Toxicol 34:809–814 43. Christol H, Levy M, Marty C (1968) Hydrolyse basique de phosphonates 1. Etude qualitative. J Organomet Chem 12:459–470 44. Cohen JA, Warringa MGPJ (1957) Purification and properties of dialkylfluorophosphatase. Biochem Biophys Acta 26:29–39 45. Compton JAF (1987) Military chemicals and biological agents. The Telford Press, Caldwell, NJ 46. Corbridge DEC (1995) Phosphorus: an outline of its chemistry, biochemistry, and uses. Elsevier, Amsterdam, p 247 47. Corriu RJP, Lanneau GF, Leclercq D (1983) Silicon-phosphorus analogies. Rate controlling factors in sn2(p) reactions. Influence of the nucleophile for a given stereochemistry. Tetrahedron Lett 24(40):4323–4326 48. Cotton FA, Wilkinson G (1988) Advanced inorganic chemistry. Interscience Publishers, New York, pp 58–68 49. Cowley AH (1984) Double bonding between the heavier main-group elements: from reactive intermediates to isolable molecules. Polyhedron 3(4):389–432 50. Cowley AH, Kemp RA (1985) Synthesis and reaction chemistry of stable two-coordinate phosphorus cations (phosphenium ions). Chem Rev 85:367–382 51. Daggett V, Schroder S, Kollman P (1991) Catalytic pathway of serine proteases: classical and quantum mechanical calculations. J Amer Chem Soc 113(23):8926–8935 52. Daughton CG, Hsieh DPH (1977) Parathion utilization by bacterial symbionts in a chemostat. App Environ Microbiol 34:175–184 53. Dave KI, Lauriano C, Xu B, Wild JR, Kenerley CM (1994) Expression of organophosphate hydrolase in the filamentous fungus Gliocladium virensi. Appl Microbiol Biotechnol 41:352–358 54. Dave KI, Phillips L, Luckow VA, Wild JR (1994) Expression and post-translational processing of a broad-spectrum organophosphorus-neurotoxin-degrading enzyme in insect tissue culture. Biotechnol Appl Biochem 19:271–284 55. Davies JA, Buchman VL, Krylova O, Ninkina NN (1997) Molecular cloning and expression pattern of rpr-1, a resiniferatoxin-binding, phosphotriesterase-related protein, expressed in rat kidney tubules. FEBS Letters 410:378–382 56. DeFrank JJ (personal communication) 57. DeFrank JJ (1986) Unpublished data 58. DeFrank JJ, Beaudry WT, Cheng T-C, Harvey SP, Stroup AN, Szafraniec LL (1993) Screening of halophilic bacteria and Alteromonas species for organophosphorus hydrolyzing enzyme activity. Chem-Biol Interactions 87:141–148 59. DeFrank JJ, Cheng T-C (1991) Purification and properties of an organophosphorus acid anhydrase from a halophilic bacterial isolate. J Bacteriol 173(6):1938–1943 60. Dejaegere A, Liang X, Karplus M (1994) Phosphate ester hydrolysis: calculation of gasphase reaction paths and solvation effects. J Chem Soc: Faraday Transactions 90(12):1763–1770 61. Dejaegere A, Lim C, Karplus M (1991) Dianionic pentacoordinate species in the basecatalyzed hydrolysis of ethylene and dimethyl phosphate. J Amer Chem Soc 113(11):4353–4355 62. Doi E, Shibata D, Matoba T (1981) Modified colorimetric ninhydrin methods for peptidase assay. Anal Biochem 118:173–184 63. Edmundson RS (1998) Dictionary of organophosphorus compounds. Chapman and Hall, London

Phosphofluoridates: Biological Activity and Biodegradation

339

64. Ekholm M, Konschin H (1999) Comparative model building of human butyrylcholinesterase. J Mol Struct (Theochem) 467:161–172 65. Emsley J, Hall D (1976) The chemistry of phosphorus. Harper and Row, London, p 315 66. Eto M (1979) Organophosphorus pesticides: organic and biological chemistry. CRC Press, Boca Raton, FL, p 62 67. Ghosh M, Grunden AM, Dunn DM,Weiss R,Adams MWW (1998) Characterization of native and recombinant forms of an unusual cobalt-dependent proline dipeptidase (prolidase) from the hyperthermophilic archaeon Pyrococcus furiosus. J Bacteriol 180(18):4781–4789 68. Gilheany, DG (1990) Structure and bonding in organophosphorus (III) compounds. In: Hartley FR (ed) Primary, secondary, and tertiary phosphines, polyphosphines and heterocyclic organophosphorus (III) compounds, vol 1. Wiley, New York, pp 9–49 69. Gilheany DG (1992) Structure and bonding in tertiary chalcogenides. In: Hartley FR (ed) Phosphine oxides, sulfides, selenides, and tellurides, vol 2. Wiley, New York, pp 1–44 70. Gilheany DG (1994) No d orbitals but Walsh diagrams and maybe banana bonds: chemical bonding in phosphines, phosphine oxides, and phosphonium ylides. Chem Rev 94:1339–1374 71. Goldwhite H (1981) Introduction to phosphorus chemistry. Cambridge University Press, Cambridge, p 42 72. Goldwhite H (1981) Introduction to phosphorus chemistry. Cambridge University Press, Cambridge, p 93 73. Gubaidullin MG (1982) Some relations holding with respect to the effects of substituents on the reactivity of phosphorus compounds. J Gen Chem USSR 52(11):2182–2184 74. Harper LL, McDaniel CS, Miller CE, Wild JR (1988) Dissimilar plasmids isolated from Pseudomonas diminuta MG and a Flavobacterium sp. (ATCC 27551) contain identical opd genes. Appl Environ Microbiol 54(10):2586–2589 75. Hartley FR (ed) (1990) Primary, secondary, and tertiary phosphines, polyphosphines and heterocyclic organophosphorus (III) compounds. The chemistry of organophosphorus compounds, vol 1. Wiley, New York 76. Hartley FR (ed) (1992) Phosphine oxides, sulfides, selenides, and tellurides. The chemistry of organophosphorus compounds, vol 2. Wiley, New York 77. Hartley FR (ed) (1994) Phosphonium salts, ylides, and phosphoranes. The chemistry of organophosphorus compounds, vol 3. Wiley, New York 78. Hartley FR (ed) (1996) Ter- and quinque-valent phosphorus acids and their derivatives. The chemistry of organophosphorus compounds, vol 4. Wiley, New York 79. Heath DF (1956) The effects of substituents on the rates of hydrolysis of some organophosphorus compounds. I. Rates in alkaline solution. J Chem Soc 3796–3804 80. Holmes RR (1980) Pentacoordinated phosphorus, vol 1. Structure and spectroscopy. American Chemical Society, Washington, DC 81. Holmes RR (1980) Pentacoordinated phosphorus, vol 2. Reaction mechanisms.American Chemical Society, Washington, DC 82. Holmes RR (1996) Comparison of phosphorus and silicon: hypervalency, stereochemistry, and reactivity. Chem Rev 96(3):927–950 83. Holmes RR (1998) Hexacoordinate phosphorus via donor interaction. Implications regarding enzymatic reaction intermediates. Acc Chem Res 31(9):535–542 84. Holmes RR (1999) Donor interactions at phosphorus. Implications regarding enzymatic reaction intermediates. Phosphorus Sulfur Silicon 144/146:1–4 85. Hoskin FCG (1956) The enzymatic hydrolysis products of sarin. Can J Biochem Phys 34:75–79 86. Hoskin FCG (1969) Possible significance of “DFPase” in squid nerve. Biol Bull 137: 389–390 87. Hoskin FCG (1971) Diisopropylphosphorofluoridate and tabun: enzymatic hydrolysis and nerve function. Science 172:1243–1245 88. Hoskin FCG, Chettur G, Mainer S, Steinmann KE, DeFrank JJ, Gallo BJ, Robbins FM, Walker JE (1989) Soman hydrolysis and detoxication by a thermophilic bacterial enzyme.

340

89. 90. 91. 92.

93. 94. 95. 96.

97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111.

J.J. DeFrank · W.E. White In: Reiner E, Hoskin FCG, Aldridge NW (eds) Enzymes hydrolyzing organophosphorus compounds. Ellis Horwood, Chichester, England, pp 53–64 Hoskin FCG, Roush AH (1982) Hydrolysis of nerve gas by squid-type diisopropyl phosphorofluoridate hydrolyzing enzyme on agarose resin. Science 215:1255–1257 Hoskin FCG, Trick GS (1955) Stereospecificity in the enzymatic hydrolysis of tabun and acetyl-b-methylcholine chloride. Can J Biochem Phys 32:963–969 Hoskin FCG,Walker JE (1998) A closer look at the natural substrate for a nerve-agent hydrolyzing enzyme in squid nerve. Biol Bull 195:197–198 Hou X, Maser RL, Magenheimer BS, Calvet JP (1996) A mouse kidney- and liver-expressed cDNA having homology with a prokaryotic parathion hydrolase (phosphotriesterase)encoding gene: abnormal expression in injured and polycystic kidneys. Gene 168: 157–163 Ishmaeva EA, Patsanovskii II (1985) Phospha-alkynes – compounds of monocoordinate phosphorus. Russian Chem Rev 54(3):243–248 Kolakowski JE, DeFrank JJ, Harvey SP, Szafraniec LL, Beaudry WT, Lai K, Wild JR (1997) Enzymatic hydrolysis of the chemical warfare agent VX and its neurotoxic analogues by organophosphorus hydrolase. Biocat Biotrans 15:297–312 Kosolapoff GM (1950) Organophosphorus compounds. Wiley, New York Lai K, Grimsley JK, Kuhlmann BD, Scapozza L, Harvey SP, DeFrank JJ, Kolakowski JE,Wild JR (1996) Rational enzyme design: computer modeling and site-directed mutagenesis for the modification of catalytic specificity in organophosphorus hydrolase. Chimia 50:430–431 Landis WG, Haley DM, Haley MV, Johnson DW, Durst HD, Savage REJ (1987) Discovery of multiple organofluorophosphate hydrolyzing activities in the protozoan Tetrahymena thermophila. J Appl Toxicol 7:35–41 Landis WG, Haley MV, Johnson DW (1986) Kinetics of the DFPase activity in Tetrahymena thermophila. J Protozool 33(2):216–218 Landis WG, Savage REJ, Hoskin FCG (1985) An organofluorophosphate-hydrolyzing activity in Tetrahymena thermophila. J Protozool 32(3):517–519 LeJeune KE, Mesiano AJ, Bower SB, Grimsley JK, Wild JR, Russell AJ (1997) Dramatically stabilized phosphotriesterase-polymers for nerve agent degradation. Biotechnol Bioeng 54:105–114 LeJeune KE, Russell AJ (1999) Biocatalytic nerve agent detoxification in fire fighting foams. Biotechnol Bioeng 62:659–665 Lewis VE, Donarski WJ, Wild JR, Raushel FM (1988) Mechanism and stereochemical course at phosphorus of the reaction catalyzed by a bacterial phosphotriesterase. Biochemistry 27(5):1591–1597 Little JS, Broomfield CA, Boucher LJ, Fox-Talbot MK (1986) Partial characterization of a rat liver enzyme that hydrolyzes sarin, soman, tabun and DFP. Fed Proceed 45(4):791 Lober O, Regitz M (1999) Pericyclic reactions in the chemistry of low-coordinated phosphorus. Main Group Chem News 7(2):3–10 Lotts KV (1960) Zur toxikologie und pharmakologie organischer phosphosaurester. Dtsch Gesundheitswes 15:2133–2179 Main AR (1960) The differentiation of the A-type esterases in sheep serum. Biochem J 75:188–195 Main AR (1960) The purification of the enzyme hydrolyzing diethyl p-nitrophenyl phosphate (paraoxon) in sheep serum. Biochem J 74:10–20 Marmann H, Groenweghe LCD, Schaad LJ, Wazer JRV (1970) An LCAO MO SCF study of the “p-d” bonding to phosphorus. The H3PO molecule. J Am Chem Soc 92(21):6107–6122 Marrs TC, Maynard RL, Sidel FR (1966) Chemical warfare agents: toxicology and treatment. Wiley, New York Marsden CJ (1984) Pseudorotation pathway and quadratic field for pf5 by ab initio calculations. J Chem Soc: Chem Commun (7):401–402 Mazur A (1946) An enzyme in animal tissues capable of hydrolyzing the phosphorus-fluorine bond of alkyl fluorophosphates. J Biol Chem 164:271–289

Phosphofluoridates: Biological Activity and Biodegradation

341

112. McDaniel CS, Harper LL,Wild JR (1988) Cloning and sequencing of a plasmid-borne gene (opd) encoding a phosphotriesterase. J Bacteriol 170(5):2306–2311 113. Mounter LA (1955) The complex nature of dialkylfluorophosphatases of hog and rat liver and kidney. J Biol Chem 215:705–709 114. Mounter LA, Baxter RF, Chanutin A (1955) Dialkylfluorophosphatases of microorganisms. J Biol Chem 215:699–704 115. Mounter LA, Dien LTH (1956) Dialkylfluorophosphatase of kidney V. The hydrolysis of organophosphorus compounds. J Biol Chem 219:685–690 116. Mounter LA, Dien LTH, Chanutin A (1955) The distribution of dialkylfluorophosphatases in the tissues of various species. J Biol Chem 215:691–697 117. Mounter LA, Floyd CS, Chanutin A (1953) Dialkylfluorophosphatase of kidney I. Purification and properties. J Biol Chem 204:221–232 118. Mounter LA, Tuck KD (1956) Dialkylfluorophosphatases of microorganisms II. Substrate specificity studies. J Biol Chem 221:537–541 119. Mulbry WW, Karns JS, Kearney PC, Nelson JO, McDaniel CS (1986) Identification of a plasmid-borne parathion hydrolase gene from Flavobacterium sp. by southern hybridization with opd from Pseudomonas diminuta. Appl Environ Microbiol 51(5):926–930 120. Mulchandani A, Pan S, Chen W (1999) Fiber-optic enzyme biosensor for direct determination of organophosphate nerve agents. Biotechnol Prog 15:130–134 121. Munnecke DM (1976) Enzymatic hydrolysis of organophosphate insecticides, a possible pesticide disposal method. Appl Environ Microbiol 32:7–13 122. Munnecke DM (1980) Enzymatic detoxification of waste organophosphate pesticides. Agric Food Chem 28:105–111 123. Neimyscheva AA, Ermolaeva MV, Knunyants IL (1970) Nucleophilic substitution in phosphorus acid derivatives IV: Kinetics of the hydrolysis of phosphonochloridic esters. J Gen Chem USSR 40(4):775–779 124. Neimyscheva AA, Knunyants IL (1966) Nucleophilic substitution in phosphorus I: Kinetics of the hydrolysis of dialkylphosphinic chlorides. J Gen Chem USSR 36(6): 1105–1111 125. Neimyscheva AA, Knunyants IL (1968) Nucleophilic exchange in derivatives of phosphorus II: Kinetics of the hydrolysis of dihalides of acids of phosphorus. J Gen Chem USSR 38(3):575–579 126. Nelson LM (1982) Biologically-induced hydrolysis of parathion in soil: Isolation of hydrolyzing bacteria. Soil Biol Biochem 14:219–222 127. Omburo GA, Kuo JM, Mullins LS, Raushel FM (1992) Characterization of the zinc binding site of bacterial phosphotriesterase. J Biol Chem 267(19):13,278–13,283 128. Ordentlich A, Barak D, Kronman C, Benschop HP, Jong LPAD, Ariel N, Barak R, Segall Y, Velan B, Shafferman A (1999) Exploring the active center of acetylcholinesterase with stereoisomers of an organophosphorus inhibitor with two chiral centers. Biochemistry 38(10):3055–3066 129. Pei L, Omburo G, McGuinn WD, Petrikovics I, Dave K, Raushel FM, Wild JR, DeLoach JR, Way JL (1994) Encapsulation of phosphotriesterase within murine erythrocytes. Toxicol Appl Pharm 124:296–301 130. Pei L, Petrikovics I, Way JL (1995) Antagonism of the lethal effects of paraoxon by carrier erythrocytes containing phosphotriesterase. Fund Appl Toxicol 28:209–214 131. Petrikovics I, Cheng T-C, Papahadjopoulos D, Hong K,Yin R, DeFrank JJ, Jiang J, Song ZH, McGuinn WD, Sylvester D, Pei L, Madec J, Tamulinas C, Jaszberenyi JC, Barcza T, Way JL (2000) Long circulating liposomes encapsulating organophosphorus acid anhydrolase in diisopropylfluorophosphate antagonism. Toxicol Sci 57:16–21 132. Petrikovics I, Hong K, Omburo G, Hu QZ, Pei L, Sylvester D, Tamulinas C, Papahadjopoulos D, Jaszberenyi JC (1999) Antagonism of paraoxon intoxication by recombinant phosphotriesterase encapsulated within sterically stabilized liposomes. Fund Appl Toxicol 22:70–75 133. Petrikovics I, McGuinn WD, Sylvester D,Yuzapavik P, Jiang J,Way JL, Papahadjopoulos D, Hong K, Yin R, Cheng T-C, DeFrank JJ (2000) In vitro studies on sterically stabilized li-

342

134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157.

J.J. DeFrank · W.E. White posomes (sl) as enzyme carriers in organophosphorus (op) antagonism. Drug Delivery 7:83–89 Phillips L, Xin J, Kirby K, Milne CP, Krell P, Wild JR (1990) Transfer and expression of an organophosphate insecticide-degrading gene from Pseudomonas in Drosophila melanogaster. Proc Nat Acad Sci USA 87:8155–8159 Quin LD (2000) A guide to organophosphorus chemistry. Wiley Interscience, New York Quinn DM (1987) Acetylcholinesterase: Enzyme structure, reaction dynamics, and virtual transition states. Chem Rev 87:955–979 Racke KD, Coats JR (1987) Enhanced degradation of isofenphos by soil microorganisms. J Agric Food Chem 35:94–99 Rastogi VK, DeFrank JJ, Cheng T-C, Wild JR (1997) Enzymatic hydrolysis of RussianVX by organophosphorus hydrolase. Biochem Biophys Res Commun 241(2) : 294–296 Raveh L, Segall Y, Leader H, Rothschild N, Levanon D, Henis Y,Ashani Y (1992) Protection against tabun toxicity in mice by prophylaxis with an enzyme hydrolyzing organophosphate esters. Biochem Pharm 44(2):397–400 Roderick SL, Matthews BW (1993) Structure of the cobalt-dependent methionine aminopeptidase from Escherichia coli – a new type of proteolytic enzyme. Biochemistry 32:3907–3912 Rosenberg A, Alexander M (1979) Microbial cleavage of organophosphorus insecticides. Appl Environ Microbiol 37:886–891 Rowland SS, Speedie MK, Pogell BM (1991) Purification and characterization of a secreted recombinant phosphotriesterase (parathion hydrolase) from Streptomyces lividans.Appl Environ Microbiol 57(2):440–444 Registry of Toxic Effects of Chemical Substabces (1996) O-Pinocolyl methylphosphonochloridate. Dec 1996, No TA3,650,000 Registry of Toxic Effects of Chemical Substabces (1999) Cyclohexyl methylphosphonofluoridate. Dec 1999, No TA8,225,000 Registry of Toxic Effects of Chemical Substabces (1997) O-Isopropyl methylphosphonofluoridate. Jan 1997, No TA8,400,000 Registry of Toxic Effects of Chemical Substabces (1998) O-Pinocolyl methylphosphonofluoridate. Jan 1998, No TA8,750,000 Registry of Toxic Effects of Chemical Substabces (1998) Diisopropyl phosphorofluoridate. Oct 1997, No TE5,075,000 Rüterjans H (2000) Personal communication Salem H, Sidel FR (1988) Nerve gases. In: Wexler P (ed) Encyclopedia of toxicology, vol 2. Academic Press, San Diego, CA, pp 380–385 Schmutzler R (1965) Fluorides of phosphorus. In: Stacey M, Tatlow JC, Sharpe AG (eds) Advances in phosphorus chemistry, vol 5. Butterworth, Washington Serdar CM, Gibson DT (1985) Enzymatic hydrolysis of organophosphates: Cloning and expression of a parathion hydrolase gene from Pseudomonas diminuta. Bio/Technol 2:567–571 Serdar CM, Gibson DT, Munnecke DM, Lancaster JH (1982) Plasmid involvement in parathion hydrolysis by Pseudomonas diminuta. Appl Environ Microbiol 44:246–249 Serdar CM, Murdock DC, Rohde MF (1989) Parathion hydrolase gene from Pseudomonas diminuta MG: Subcloning, complete nucleotide sequence, and expression of the mature portion of the enzyme in Escherichia coli. Bio/Technol 7:1151–1155 Sethanathan N, Yoshida T (1973) A Flavobacterium that degrades diazinon and parathion. Can J Microbiol 19:873–875 Sheela S, Pai SB (1983) Metabolism of fensulfothion by a soil bacterium, Pseudomonas alcaligenes c1. Appl Environ Microbiol 46(2):475–479 Shelton DR, Somich CJ (1988) Isolation and characterization of coumaphos-metabolizing bacteria from cattle dip. Appl Environ Microbiol 54(10):2566–2571 Siddaramappa R, Rajaram KP, Sethanathan N (1973) Degradation of parathion by bacteria isolated from flooded soil. Appl Environ Microbiol 26:846–849

Phosphofluoridates: Biological Activity and Biodegradation

343

158. Simonian AL, Grimsley JK, Flounders AW, Schoeniger JS, Cheng T-C, DeFrank JJ, Wild JR (2001) An enzyme-based biosensor for the direct detection of G-type chemical warfare agents. Anal Chim Acta (in press) 159. Steiert JG, Pogell BM, Speedie MK, Laredo J (1989) A gene coding for a membrane-bound hydrolase is expressed as a secreted soluble enzyme in Streptomyces lividans. Bio/Technol 7:65–68 160. Sussman JL, Harel M, Frolow F, Oefner C, Goldman A, Tolker L, Silman I (1991) Atomic structure of acetylcholinesterase from Torpedo californica: A prototypic acetylcholinebinding protein. Science 252:872–879 161. Tammelin LE (1957) Dialkoxy-phosphorylthiocholines, alkoxy-methylphosphorylthiocholines and analogous choline esters. Acta Chem Scand 11:1340–1349 162. Tammelin LE (1957) Methyl-fluoro-phosphorylcholines. Acta Chem Scand 11 : 859–865 163. Taylor P, Brown JH (1994) Acetylcholine. In: Siegel GJ, Agranoff BW, Albers RW, Molinoff PB (eds) Basic neurochemistry: Molecular, cellular, and medical aspects. Raven Press, New York, pp 231–260 164. Uchimaru T, Tanabe K, Nishikawa S, Taira K (1991) Ab initio studies of a marginally stable intermediate in the base-catalyzed methanolysis of dimethyl phosphate and nonexistence of the stereoelectronically unfavorable transition state. J Amer Chem Soc 113(11):4351–4353 165. Wassada H, Hirao K (1992) Theoretical study of the reactions of pentacoordinated trigonal-bipyramidal compounds: PH5 , PF5 , PF4H, PF3H2 , PF4CH3 , PF3(CH3)2 , P(O2C2H4)H3 , and PO5H4–. J Am Chem Soc 114(1):16–27 166. Watkins LM, Mahoney HJ, McCulloch JK, Raushel FM (1997) Augmented hydrolysis of diisopropyl fluorophosphate in engineered mutants of phosphotriesterase. J Biol Chem 272(41):25,596–25,601 167. Westheimer FH (1987) Why nature chose phosphates. Science 235:1173–1178 168. Westheimer FW (1968) Pseudo-rotation in the hydrolysis of phosphate esters. Acc Chem Res 1(3):70–78 169. White WE (1999) Effects of chemical reactivity on the toxicity of phosphorus fluoridates. SAR and QSAR in Environ Res 10:207–213 170. Wilson BW, Hooper MJ, Hansen ME, Nieberg PS (1992) Reactivation of organophosphorus inhibited AChE with oximes. In: Chambers JE, Levi PE (eds) Organophosphates: Chemistry, fate, and effects. Academic Press, San Diego, CA, pp 1108–1137 171. Wladkowski BD, Krauss M, Stevens WJ (1995) Apicophilicities of substituted oxyphosphoranes [P(OH)4X, PO–(OH)3X]. J Phys Chem 99:4490–4500 172. Wright JB,White WE (1998) A neutral gas phase mechanism for the reaction of methanol with dimethylphosphinic fluoride. J Mol Struct (Theochem) 454:259–265 173. Yang Y-C, Baker JA, Ward JR (1992) Decontamination of chemical warfare agents. Chem Rev 92(8):1729–1743