Antimony biomethylation by Scopulariopsis brevicaulis: characterization of intermediates and the methyl donor

Chemosphere 41 (2000) 1717±1725

Antimony biomethylation by Scopulariopsis brevicaulis: characterization of intermediates and the methyl donor Paul Andrewes a, William R. Cullen a

a,*

, Elena Polishchuk

b

Environmental Chemistry Group, Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver BC, Canada V6T 1Z1 b Biological Services Facility, Department of Chemistry, University of British Columbia, Vancouver BC, Canada V6T 1Z1 Received 18 October 1999; accepted 19 February 2000

Abstract The ®lamentous fungus Scopulariopsis brevicaulis biomethylates inorganic antimony(III) compounds to trimethylstibine, that can be detected in culture headspace gases. Dimethylantimony and trimethylantimony species have been detected in the medium of these cultures, but the origin of these species was controversial. We now show that the dimethylantimony species is a true intermediate on the pathway to trimethylstibine (rather than arising from trimethylstibine oxidation or as an analytical artifact) because no dimethylantimony species are formed on trimethylstibine oxidation, as determined by using HG±GC±AAS. Furthermore, the dimethylantimony and trimethylantimony species can be separated, by using anion exchange chromatography, and so the dimethylantimony species is not an analytical artifact, formed during the hydride generation process. The antimony biomethylation mechanism was further probed by measuring incorporation of the methyl group, from 13 CD3 -L -methionine and CD3 -D -methionine, into methylantimony species and, for comparison, into methylarsenic species. The percentage incorporation of the labeled methyl group into methylarsenic and methylantimony species was not signi®cantly di€erent. The incorporation from 13 CD3 -L -methionine was 54% and 47% for antimony and arsenic, respectively. The incorporation from CD3 -D -methionine was 20% and 16% for antimony and arsenic, respectively. It appears that the biomethylation of arsenic and antimony occur by very similar, perhaps identical, mechanisms. Ó 2000 Elsevier Science Ltd. All rights reserved. Keywords: Scopulariopsis brevicaulis; Methionine; Trimethylstibine oxidation; Dimethylantimony; Trimethylantimony; Arsenic

1. Introduction Biomethylation of metalloids in the environment is an important phenomenon that has a large in¯uence on the toxicity and biogeochemical cycling of metalloids. For example, the biomethylation of arsenic in humans has recently received considerable attention (National Research Council, 1999) because methylarsenic species

*

Corresponding author. Tel.: +1-604-822-4435; fax: +1-604822-2847. E-mail address: [email protected] (W.R. Cullen).

have di€erent properties, such as toxicity, compared with inorganic arsenic species. Some of the ®rst research on metalloid biomethylation was performed using the ®lamentous fungus Scopulariopsis brevicaulis (originally classi®ed as Penicillium brevicaule) that Gosio (1901) isolated from wallpaper. Gosio demonstrated that S. brevicaulis produces a volatile arsenic compound that Challenger (1945) subsequently identi®ed as trimethylarsine. This work was performed because some deaths had been attributed to arsenic poisoning, resulting from the biological production of toxic volatile arsenic species from damp wallpaper. The wallpaper contained inorganic arsenic compounds, notably copper arsenite, as pigments. Challenger also showed that

0045-6535/00/$ - see front matter Ó 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 5 - 6 5 3 5 ( 0 0 ) 0 0 0 6 3 - 1

1718

P. Andrewes et al. / Chemosphere 41 (2000) 1717±1725

S. brevicaulis could biomethylate selenium and tellurium to dimethylselenide and dimethyltelluride, respectively, but was unable to demonstrate the biomethylation of antimony to trimethylstibine. Challenger (1945) proposed a mechanistic pathway (Fig. 1) for arsenic biomethylation and presented evidence for it. Subsequently, by using modern analytical instrumentation, the work of Challenger was con®rmed (Cullen and Reimer, 1989), and methylarsenic(V) intermediates were identi®ed in the medium of S. brevicaulis cultures (Cullen et al., 1994), which added further evidence to support ChallengerÕs proposed mechanism. The biomethylation of metalloids, especially antimony, by S. brevicaulis has received renewed attention recently, because of the hypothesis that S. brevicaulis, or similar microorganisms, growing on infantsÕ bedding materials, can biovolatilize metalloids in the materials.

Fig. 1. The mechanism proposed by Challenger (1945) for the biomethylation of arsenic, and the structure of S-adenosylmethionine (SAM); the circled methyl group in SAM is donated during methylation steps.

These toxic volatile species could then be responsible for chronic poisoning (Richardson, 1994). It was demonstrated that S. brevicaulis can biomethylate inorganic antimony(III) compounds to trimethylstibine (Andrewes et al., 1999a; Craig et al., 1999), but the evidence obtained to date does not support the hypothesis that microbiological trimethylstibine production, in bedding materials, causes sudden infant death (Limerick, 1998); this evidence includes the ®nding that S. brevicaulis only produces trace amounts of trimethylstibine (Andrewes et al., 1999a; Craig et al., 1999). In all experiments performed so far with cultures of S. brevicaulis, containing inorganic antimony compounds (Andrewes et al., 1998, 1999a,b; Jenkins et al., 1998a,b; Craig et al., 1999), the yields of biomethylation products (trimethylstibine and nonvolatile dimethylantimony and trimethylantimony species) are orders of magnitude lower than those obtained when inorganic arsenic compounds are used as a substrate (Challenger, 1945; Cullen and Reimer, 1989). This phenomenon may be speci®c to S. brevicaulis or it may be that antimony is not readily biomethylated by most microorganisms. Understanding the mechanism of antimony biomethylation might help to explain why antimony is less readily biomethylated. Because of their similar chemistries, the mechanism of antimony biomethylation is likely to be similar to the mechanism of arsenic biomethylation; it is likely that antimony biomethylation is a fortuitous process that occurs when the methyltransferase enzymes, which normally act upon arsenic, act upon antimony. Examining the mechanism of antimony biomethylation requires performing experiments analogous to those used to examine arsenic biomethylation. In this paper, we describe some experiments of this type. Some of the best evidence to support ChallengerÕs proposed mechanism for arsenic biomethylation was the detection of some of the expected intermediates (Cullen et al., 1994). Dimethylantimony and trimethylantimony species, which are putative intermediates in the antimony biomethylation pathway to trimethylstibine, were identi®ed in the medium of S. brevicaulis cultures (monomethylantimony species were not detected) (Andrewes et al., 1998, 1999b). Others have suggested that these are not true intermediates (Craig et al., 1999); i.e., they might be analytical artifacts or formed by trimethylstibine oxidation. In this paper, we describe experiments to show that the dimethylantimony species, detected in S. brevicaulis cultures, is a true intermediate and thus demonstrate that antimony biomethylation is similar to arsenic biomethylation. We describe experiments that show dimethylantimony species are not readily formed by trimethylstibine oxidation, however we have no means of di€erentiating between trimethylantimony species that are intermediates on the pathway to trimethylstibine and trimethyl-

P. Andrewes et al. / Chemosphere 41 (2000) 1717±1725

antimony species that arise from trimethylstibine oxidation. Indeed this is also the situation for arsenic, but, at low arsenic concentrations, trimethylarsine oxide is the end-product of arsenic biomethylation with very little formation of trimethylarsine (Cullen et al., 1994). This situation is likely to be true for antimony as well. In some cases, when hydride generation is performed using pure trimethylantimony or dimethylantimony compounds as standards, four peaks are seen in the chromatogram assignable to stibine, methylstibine, dimethylstibine and trimethylstibine (Dodd et al., 1992). We can now perform hydride generation experiments without seeing any rearrangements (Andrewes et al., 1998; Koch et al., 1998) nevertheless it is remotely possible that the dimethylantimony species we detected in the medium of S. brevicaulis may be an analytical artifact as is maintained by Craig et al. (1999). In this paper, our aim was to separate the dimethylantimony and trimethylantimony species in order to show that the dimethylantimony species is not an artifact. S-adenosyl methionine (SAM, Fig. 1) is established as the methyl donor for arsenic biomethylation (Cullen and Reimer, 1989) and in a previous paper we demonstrated that L -methionine, probably as the active compound SAM, is a methyl donor for antimony biomethylation (Andrewes et al., 1999b). Although this work provided some indication that the mechanisms of arsenic and antimony biomethylation by S. brevicaulis are similar, the calculated incorporation of methyl groups into antimony compounds could not be compared with literature values for arsenic because the arsenic experiments were performed under di€erent conditions. Thus, for corroboration, we now present the results from experiments designed to establish the abilities of 13 CD3 -L -methionine and CD3 -D -methionine to act as a methyl donors during arsenic and antimony biomethylation under identical conditions. Any di€erences in the abilities of D -methionine and L -methionine to act as methyl donors would arise if the biomethylation processes were stereospeci®c, which is most likely if an enzyme is involved. Mammalian methyltransferase enzymes, that perform arsenic biomethylation, have been isolated (Aposhian, 1997). The results presented in this paper indicate that the mechanisms of antimony and arsenic biomethylation are probably very similar although other environmental systems will need to be examined to see if this is the case in general. 2. Experimental 2.1. HG±GC±AAS In the experiments described below organoantimony and organoarsenic species were determined by using

1719

hydride-generation gas-chromatography atomic-absorption-spectroscopy (HG±GC±AAS) as described in previous literature (Cullen et al., 1994; Andrewes et al., 1998). Semi-continuous ¯ow hydride derivatization was performed using 2% sodium borohydride and an appropriate bu€er (50 mM citrate, pH 6.0 for antimony species; 4 M acetic acid for arsenic species). Where necessary quanti®cation was made by using trimethylantimony dichloride and trimethylarsine oxide to perform standard additions.

2.2. Determination of trimethylstibine oxidation products Trimethylstibine was generated in septa-capped vials (15 ml) as follows: An aqueous solution (50, 100, 200 or 500 ll) of trimethylantimony dichloride (100 mg Sb/l) was placed in a vial and water was added to give a total volume of 4 ml. Sodium borohydride (1 ml of 6%) was then injected through the septa. Immediately after vigorously shaking the vial, the headspace gas (1 ml), containing trimethylstibine, was sampled with a syringe and injected, via a 0.2 lm ®lter, into another septacapped vial (15 ml) containing 2 ml of sterile minimalsalts/glucose medium, the same as used in all of our cultures (Cox and Alexander, 1973), and a headspace of laboratory air. During injection the needle of the gas-syringe was placed below the level of the medium so that the trimethylstibine bubbled through the medium. The vials were then placed on the shaker used for the incubation of S. brevicaulis, and left for a week. The medium was then analyzed by using a previously described solid phase extraction procedure (Andrewes et al., 1998) followed by HG±GC±AAS as described above.

2.3. Separation of trimethylantimony and dimethylantimony species by using anion exchange chromatography Basic alumina (30 g, 80±200 mesh, Brockman activity I, Fisher Scienti®c) was placed into a 60 ml syringe; a small glass wool plug was used to hold the alumina in place. The alumina was rinsed with 50 ml of water. Then 50 ml of medium containing dimethylantimony and trimethylantimony species (from a culture of S. brevicaulis that had been incubated with potassium antimony tartrate for one month (Andrewes et al., 1998)) was passed through the alumina, and the eluate was collected (Eluate 1). The alumina was then further rinsed with 50 ml of distilled water. Finally, the alumina was rinsed with 50 ml of ammonium carbonate (50 mM, pH 12), and the eluate collected (Eluate 2). The two samples of eluate were analyzed by using HG±GC±AAS as described above.

1720

P. Andrewes et al. / Chemosphere 41 (2000) 1717±1725

2.4. CD3 -D -methionine and 13 CD3 -L -methionine as methyl donors for antimony and arsenic biomethylation

ment) and maintained at 26°C. After one month of incubation, the cultures were autoclaved (121°C, 131 kPa (19 psi), 20 min) before analysis. The culture medium was passed through a basic alumina solid phase extraction (SPE) column which removed the inorganic arsenic(III) and antimony(III) species (the details of this are described in Andrewes et al. (1998)) and it should be noted that, without modi®cation, the procedure removed >95% of the inorganic arsenic(III) and antimony(III) species, whereas >80% of the trimethylarsenic and trimethylantimony species were recovered in the eluate). The SPE eluate was analyzed to determine the approximate concentration of organoarsenic and organoantimony species by using HG±GC±AAS as described above. Then the SPE eluate was analyzed by using a previously described hydride-generation gaschromatography mass-spectrometry (HG±GC±MS) procedure to determine the incorporation of labeled methyl groups (Andrewes et al., 1999b).

Minimal-salts/glucose medium (Cox and Alexander, 1973) (400 ml) was seeded with 40 ml of S. brevicaulis (ATCC #7903) mycelial balls (20±30 balls,  1 mm diameter). Cultures of S. brevicaulis were prepared containing potassium antimony tartrate (100 mg Sb/l) or sodium arsenite (10 mg As/l), and CD3 -D -methionine (synthesized in our laboratory) or 13 CD3 -L -methionine (BOC Prochem) as detailed in Table 1. The methionine (0.1 g) was dissolved in 10 ml of water and added to the cultures via a syringe ®lter (0.2 lm). Appropriate volumes of potassium antimony tartrate (5000 ng Sb/ml) and sodium arsenite (1000 ng As/ml) stock solutions were added to the cultures via 0.2 lm syringe ®lters. The cultures were maintained in 1 l Erlenmeyer ¯asks capped with cotton stoppers. The Erlenmeyer ¯asks were shaken horizontally ( 135 rpm, 4.45 cm displace-

Table 1 Estimates of the biological incorporation of labeled methyl groups into trimethylantimony and trimethylarsenic speciesa Contents of culture 1. S. brevicaulis + 13 CD3 -L -methionine (0.1 g) + PATd (100 mg Sb/l) 2. S. brevicaulis + 13 CD3 -L -methionine (0.1 g) + PAT (100 mg Sb/l)

3. S. brevicaulis + CD3 -D -methionine (0.1 g) + PAT (100 mg Sb/l) 4. S. brevicaulis + CD3 -D -methionine (0.1 g) + PAT (100 mg Sb/l)

Percent  Meb in trimethylantimony

Ion ratios

121

123

Averagec

121/123

151/153

155/157

159/161

57

53

55

1.08

1.45

1.25

1.82

52

55

53

1.10

1.09

1.57

0.92

Percent  Me in trimethylantimony

Ion ratios

121

123

Average

121/123

151/153

154/156

157/159

14 21 22 18 20

22 23 21 17 23

18 22 21 18 22

1.24 1.38 1.20 1.11 1.06

2.35 1.54 1.70 1.87 1.47

1.46 1.36 1.49 2.66 1.38

0.7 1.46 2.16 1.39 1.07

Percent  Me in trimethylarsenic 5. S. brevicaulis + CD3 -D -methionine (0.1 g) + sodium arsenite (10 mg As/l) 6. S. brevicaulis + CD3 -D -methionine (0.1 g) + sodium arsenite (10 mg As/l) 7. S. brevicaulis + 13 CD3 -L -methionine (0.1 g) + sodium arsenite (10 mg As/l) 8. S. brevicaulis + 13 CD3 -L -methionine (0.1 g) + sodium arsenite (10 mg As/l)

± ±

± ±

11 15

± ± ± ±

± ± ± ±

21 20 17 36

± ±

± ±

54 53

a

See text for discussion of the assumptions made in these calculations. Me ˆ 13 CD3 for L -methionine experiments and CD3 for D -methionine experiments. c Average of results for the two antimony isotopes 121 and 123. Arsenic results are for the one isotope, d PAT ˆ potassium antimony tartrate. b

75

As.

P. Andrewes et al. / Chemosphere 41 (2000) 1717±1725

1721

3. Results 3.1. The product of trimethylstibine oxidation Trimethylstibine was injected into minimal-salts/glucose media in 15 ml septa-capped vials containing a headspace of laboratory air, and the oxidation-products determined, by using HG±GC±AAS. The only signi®cant product of trimethylstibine oxidation is a trimethylantimony species (Fig. 2). 3.2. Separation of trimethylantimony and dimethylantimony species by using anion exchange chromatography We had previously observed that when using basic alumina SPE for sample cleanup, failure to rinse the alumina with ammonium carbonate (50 mM, pH 12) resulted in retention of the dimethylantimony species. In the present study a sample of medium that was known to contain dimethylantimony and trimethylantimony species (Andrewes et al., 1998) was passed through alumina that had only been rinsed with water. The dimethylantimony species was retained but the trimethylantimony species eluted (Eluate 1, Fig. 3). The dimethylantimony species was recovered by rinsing the alumina with ammonium carbonate (Eluate 2, Fig. 3). Because these

Fig. 3. HG±GC±AAS chromatograms of dimethylantimony and trimethylantimony species obtained by using anion exchange chromatography.

solution-species can be separated by using chromatography, the dimethylantimony species observed by using HG±GC±AAS is not an artifact arising from the demethylation of trimethylstibine during hydride generation. Unfortunately inorganic antimony(V) species behave in a similar manner to the dimethylantimony species so this procedure is not yet suitable for obtaining pure dimethylantimony species for further studies. We did not detect any monomethylantimony species but this does not necessarily mean that these species were absent from the medium since we had no monomethylantimony standards to determine how this species behaves on alumina. When S. brevicaulis is incubated with sodium arsenite or sodium arsenate monomethylarsenic species are not detected (Cullen et al., 1994). 3.3. CD3 -D -methionine and 13 CD3 -L -methionine as methyl donors for antimony and arsenic biomethylation

Fig. 2. HG±GC±AAS chromatograms of media samples containing the products of trimethylstibine oxidation. The volumes given are that of Me3 SbCl2 (100 mg Sb/l) which was placed in a 15 ml vial and derivatized with sodium borohydride to form trimethylstibine found in the vial headspace. This headspace gas (1 ml) was subsequently injected into a second vial, containing medium, where the oxidation occurs (see text for full details).

We have previously reported that dimethylantimony and trimethylantimony species obtained from the incubation of S. brevicaulis with potassium antimony tartrate and 13 CD3 -L -methionine contained 30±50% incorporation of 13 CD3 (Andrewes et al., 1999b). In the present experiments cultures of S. brevicaulis were prepared containing 13 CD3 -L -methionine and potassium antimony tartrate (Cultures 1 and 2) equivalent to our previously reported experiments (Andrewes et al.,

1722

P. Andrewes et al. / Chemosphere 41 (2000) 1717±1725

1999b); 13 CD3 -L -methionine and sodium arsenite (Cultures 7 and 8); CD3 -D -methionine and potassium antimony tartrate (Cultures 3 and 4); and CD3 -D methionine and sodium arsenite (Cultures 5 and 6). The concentrations of organoarsenic and organoantimony species in the media were determined by using HG±GC±AAS. The concentrations of trimethylantimony species in the media of cultures containing L -methionine and potassium antimony tartrate were approximately 6 ng Sb/ml. The concentrations of trimethylantimony species in the media of cultures containing D -methionine and potassium antimony tartrate were approximately 2 ng Sb/ml. In the cultures containing L -methionine and sodium arsenite, the concentrations of trimethylarsenic species in the media were approximately 80 ng As/ml. For the cultures that contained D -methionine and sodium arsenite, the concentrations of trimethylarsenic species in the media were approximately 30 ng As/ml. Hydride derivatization of the medium from Cultures 1±4 produced dimethylstibine and trimethylstibine. The trimethylstibine was collected in septa-capped vials and then injected into the GC±MS (ion trap). A typical mass spectrum, for trimethylstibine that contains no labeledmethyl groups is shown in Fig. 4(a). When trimethylstibine, obtained by hydride derivatization of medium originating from a culture that contained 13 CD3 -L -methionine, is injected into the GC±MS a number of extra peaks are evident in the mass spectrum (Fig. 4(b)). The trimethylstibine mass spectra, for cultures that had been incubated with CD3 -D -methionine, also contained a number of extra peaks (Fig. 4(c)). In both cases these are consistent with the incorporation of the appropriate intact labeled-methyl group. The fraction of 13 CD3 (from 13 CD3 -L -methionine) incorporated into trimethylantimony species, can be calculated for the 121 Sb isotope (f121 ), by using the following formula: f121 ˆ

…1=2  A155 † ‡ A159 ; A151 ‡ A155 ‡ A159

where A151 , A155 and A159 are the peak areas, at the trimethylstibine retention time, of the ion chromatograms for m=z ˆ 151, 155, and 159, respectively. Analogous formulas were used for the 123 Sb isotope and for calculating the fraction of CD3 (from CD3 -D -methionine) incorporated into the trimethylantimony species. In using these procedures we assume (1) stibines containing labeled methyl groups fragment in the same way as unlabeled stibines; (2) the probability of ®nding a labeled methyl group in a fragment ion is the same for all fragments ions; and (3) there are no other interfering fragment ions. The results from the quanti®cation are given in Table 1. The isotope ratios, shown in Table 1, represent the ratios of peak areas for the single ion

Fig. 4. Mass spectra of trimethylstibine obtained by using GC± MS (ion trap): (a) trimethylstibine containing no labeled methyl groups; (b) trimethylstibine that was obtained by hydride derivatization of medium from a culture that contained 13 CD3  L -methionine, Me ˆ 13 CD3 ; (c) trimethylstibine that was obtained by hydride derivatization of medium from a culture that contained CD3 -D -methionine,  Me ˆ CD3 .

chromatograms. The natural isotope ratio for antimony is 1.34. The signi®cant variation from the natural isotope ratio arises from instrumental error and from the presence of interfering fragments in the mass spectrum although there was no evidence in the GC±TIC of other species co-eluting with trimethylstibine (Andrewes et al., 1999b). Thus the assumptions made above are not fully substantiated. Hydride derivatization of medium from Cultures 5±8 produced trimethylarsine. The trimethylarsine was collected in septa-capped vials and then injected into the GC±MS (ion trap). A typical mass spectrum, for trimethylarsine that contains no labeled-methyl groups is shown in Fig. 5(a). When trimethylarsine, obtained by hydride derivatization of medium originating from a culture that contained 13 CD3 -L -methionine, is injected into the GC±MS a number of extra peaks are evident in the mass spectrum (Fig. 5(b)). The trimethylarsine mass spectra, for cultures that had been incubated with CD3 -

P. Andrewes et al. / Chemosphere 41 (2000) 1717±1725

Fig. 5. Mass spectra of trimethylarsine obtained by using GC± MS (ion trap): (a) trimethylarsine containing no labeled methyl groups; (b) trimethylarsine that was obtained by hydride generation using medium from a culture that contained 13 CD3  L -methionine, Me ˆ 13 CD3 ; (c) trimethylarsine that was obtained by hydride generation using medium from a culture that contained CD3 -D -methionine,  Me ˆ CD3 . D -methionine, also contained a number of extra peaks (Fig. 5(c)). The presence of these peaks is consistent with the incorporation of the appropriate intact labeledmethyl groups. Results from the quanti®cation of the amounts of labeled-methyl groups are tabulated in Table 1. The same assumptions that were made in quantifying label incorporation into trimethylstibine were made for trimethylarsine; however, because arsenic has only one naturally occurring isotope, isotope ratios cannot be monitored as a check.

4. Discussion Parris and Brinckman (1976) studied the oxidation of trimethylstibine and trimethylarsine to better under-

1723

stand the fate of these compounds in the environment. They reported that in unagitated methanol solution, at high concentration, trimethylstibine reacts with air to a€ord mainly the oxide ÔMe3 SbOÕ accompanied by limited (<10%) antimony-carbon bond cleavage to a€ord ill-de®ned products; they found evidence for the formation of a dimethylantimony species by using 1 H-NMR. Extrapolating these results to environmental conditions is dicult but they seem to indicate that at low concentration oxidation of trimethylstibine probably does not result in signi®cant antimony-carbon bond cleavage (the reactions that result in antimony-carbon bond cleavage are less likely at low trimethylstibine concentration). Nevertheless, if the conclusions of Parris and Brinckman (1976) are to be believed, it is possible that the dimethylantimony species detected in the medium of S. brevicaulis cultures could arise from trimethylstibine oxidation (Craig et al., 1999). In order to refute this possibility we examined the oxidation products of trimethylstibine by using HG±GC±AAS. Only trimethylantimony species were detected. Thus, dimethylantimony species in the medium are unlikely to have arisen as a result of trimethylstibine oxidation. Furthermore, the dimethylantimony and trimethylantimony species detected in the medium could be separated, by using anion exchange chromatography, and so the dimethylantimony species is not an analytical artifact, formed during the hydride generation process. Challenger (1945) found that cultures of S. brevicaulis, grown in the presence of sodium arsenite and 14 C labeled D ,L -methionine, produced trimethylarsine containing 28.3% of 14 C labeled methyl group. Cullen et al. (1977) then demonstrated that the methyl group is transferred intact, by using CD3 labeled methionine. The trimethylarsine produced by S. brevicaulis incubated with CD3 -methionine and sodium arsenite was collected, and then injected into an MS. It was then demonstrated that cultures of S. brevicaulis, grown in the presence of sodium arsenite, also produced mainly a nonvolatile trimethylarsenic species found in the medium (Cullen et al., 1994). However, the methyl donor for this nonvolatile species was not determined although it was shown that cultures of Cryptococcus humicolus (Candida humicola), grown in the presence of sodium arsenite and CD3 -L -methionine, produced nonvolatile methylarsenic species containing 20% CD3 (Cullen et al., 1995). In a previous paper, we reported that when S. brevicaulis was incubated with 13 CD3 -L -methionine and potassium antimony tartrate the trimethylantimony species produced contained 47% labeled methyl group (Andrewes et al., 1999b). However the results for arsenic and antimony in these previous studies could not be directly compared. In the present study we found that cultures of S. brevicaulis, grown in the presence of potassium antimony tartrate and 13 CD3 -L -methionine, produced a nonvolatile trimethylantimony species that contained

1724

P. Andrewes et al. / Chemosphere 41 (2000) 1717±1725

54% labeled methyl group. When cultures of S. brevicaulis were grown in the presence of sodium arsenite and 13 CD3 -L -methionine, a nonvolatile trimethylarsenic species was produced. This trimethylarsenic species contained 47% labeled methyl group. Thus, within experimental error, antimony behaves in the same way as arsenic with respect to the methyl source for biomethylation. When D -methionine was used as a methyl source for both potassium antimony tartrate (20% CD3 in trimethylantimony) and sodium arsenite (16% CD3 in trimethylarsenic) less label was incorporated. This is expected if biomethylation is an enzymatic process, the stereochemistry of D -methionine would exclude it from the enzyme, but some incorporation may occur if there is racemization of the D -methionine. Again the similar behavior of arsenic and antimony in this respect is a good indication that they are both biomethylated in the same way, Table 2. As can be seen from Table 2, there is much in common between arsenic and antimony. The signi®cant di€erences are: (i) S. brevicaulis rapidly reduces arsenate to arsenite (Cullen et al., 1994), but antimony(V) is not reduced (Andrewes et al., 1998). (ii) Arsenate is biomethylated, most likely after reduction within the cell to

arsenite (Cullen and Reimer, 1989), but antimony(V) is not (Andrewes et al., 1998). (iii) The yields of methylantimony species (Andrewes et al., 1998, 1999a,b; Jenkins et al., 1998a,b; Craig et al., 1999) are much lower than the yields of methylarsenic species (Cullen and Reimer, 1989; Cullen et al., 1994). The reduction of arsenate to arsenite is a well-known cellular process (Cullen and Reimer, 1989; Silver, 1996); arsenate is structurally very similar to phosphate and is actively taken up by cells and reduction followed by elimination of arsenite is probably a means of detoxi®cation (Cullen and Reimer, 1989; Silver, 1996). The structures of antimony(V) species, such as the octahedral hexahydroxyantimonate, are di€erent from the tetrahedral phosphate and arsenate so it is unlikely that antimony(V) species are taken up by the same active systems as phosphate consequently there is no need for the cells to reduce antimony(V) species. Thus, points (i) and (ii) are probably explained by cell transport phenomena. The meager rate of antimony biomethylation, point (iii), can be explained in terms of either a low rate of antimony turnover at the enzyme site, or a low concentration of antimony at the site, as a result of low uptake. More research needs to be performed to elucidate this.

Table 2 Evidence supporting the mechanism (Fig. 1) proposed by Challenger (1945) for arsenic biomethylation and comparable results for antimonya Arsenic

Antimony

References

Reduction of E(V) to E(III)

XXX

X

Oxidation of E(III) to E(V) Biomethylation of E(V)

X XXX

X Xb

Biomethylation of E(III)

XXX

X

Biomethylation of RE(V), R2 E(V)

Not attempted

Detection of intermediate: MeE(V)

X X X

X

Detection of intermediates: Me2 E(V) and Me3 E(V)

XXX

X

Detection of volatile ®nal product Me3 E

XXX

X

L -methionine

47% 16%

54% 20%

Cullen et al. (1994) Andrewes et al. (1998) Andrewes et al. (1998) Cullen and Reimer (1989) Andrewes et al. (1998) Jenkins et al. (1998a) Cullen and Reimer (1989) Andrewes et al. (1998) Jenkins et al. (1998a) Craig et al. (1999) Challenger (1945) Cullen et al. (1994) Cullen et al. (1994) Andrewes et al. (1998) Cullen et al. (1994) Andrewes et al. (1998) Andrewes et al. (1999b) Andrewes et al. (1999a) Craig et al. (1999) Challenger (1945) Cullen and Reimer (1989) This work This work

D -methionine a

as a methyl donor as a methyl donor

``E'' refers to arsenic or antimony. Results are uncertain, we found no evidence for biomethylation of antimony(V) (Andrewes et al., 1998). However Jenkins et al. found evidence of antimony(V) biomethylation even though it was less than for antimony(III) (Jenkins et al., 1998a).

b

P. Andrewes et al. / Chemosphere 41 (2000) 1717±1725

It has been shown that the mechanisms of antimony and arsenic biomethylation in cultures of the model microorganism S. brevicaulis are probably the same. This may probably be the case for other organisms. Very little work on the biomethylation of antimony in mammals has been performed although one study claimed to show that antimony biomethylation in mammals does not occur (Bailly et al., 1991), but detection limits for methylantimony compounds were not given. It was shown that antimony signi®cantly inhibits arsenic biomethylation in rat liver cytosol (Buchet and Lauwerys, 1985). The transformation of antimony in the environment is an exciting ®eld where much work still needs to be done. Acknowledgements The authors wish to thank Bianca Kuipers for providing technical assistance with the GC±MS. We are also grateful to Changqing Wang for synthesizing the CD3 -D -methionine. NSERC Canada provided funding for this work. References Andrewes, P., Cullen, W.R., Feldmann, J., Koch, I., Polishchuk, E., Reimer, K.J., 1998. The production of methylated organoantimony compounds by Scopulariopsis brevicaulis. Appl. Organomet. Chem. 12, 827. Andrewes, P., Cullen, W.R., Polishchuk, E., 1999a. Con®rmation of the aerobic production of trimethylstibine by Scopulariopsis brevicaulis. Appl. Organomet. Chem. 13, 659. Andrewes, P., Cullen, W.R., Feldmann, J., Koch, I., Polishchuk, E., 1999b. Methylantimony compound formation in the medium of Scopulariopsis brevicaulis cultures: 13 CD3 -L methionine as a source of the methyl group. Appl. Organomet. Chem. 13, 681. Aposhian, H.V., 1997. Enzymatic methylation of arsenic species and other new approaches to arsenic toxicity. Ann. Rev. Pharmacol. Toxicol. 37, 347. Bailly, R., Lauwerys, R., Buchet, J.P., Mahieu, P., Konings, J., 1991. Experimental and human studies on antimony metabolism: their relevance for the biological monitoring of workers exposed to inorganic antimony. Br. J. Ind. Med. 48, 93. Buchet, J.P., Lauwerys, R., 1985. Study of inorganic arsenic methylation by rat liver in vitro: relevance for the interpretation of observations in man. Arch. Toxicol. 57, 125.

1725

Challenger, F., 1945. Biological methylation. Chem. Rev. 36, 315. Cox, D.P., Alexander, M., 1973. E€ect of phosphate and other anions on trimethylarsine formation by Candida Humicola. Appl. Microbiol. 25, 408. Craig, P.J., Jenkins, R.O., Dewick, R., Miller, D.P., 1999. Trimethylantimony production by Scopulariopsis brevicaulis during aerobic growth. Sci. Tot. Environ. 229, 83. Cullen, W.R., Froese, C.L., Lui, A., McBride, B.C., Patmore, D.J., Reimer, M., 1977. The aerobic methylation of arsenic by microorganisms in the presence of L -methionine-d3 . J. Organomet. Chem. 139, 61. Cullen, W.R., Li, H., Hewitt, G., Reimer, K.J., Zalunardo, N., 1994. Identi®cation of extracellular arsenical metabolites in the growth medium of the microorganisms Apiotrichum humicola and Scopulariopsis brevicaulis. Appl. Organomet. Chem. 8, 303. Cullen, W.R., Li, H., Pergantis, S.A., Eigendorf, G.K., Mosi, A.A., 1995. Arsenic biomethylation by the microorganism Apiotrichum Humicola in the presence of L -methioninemethyl-d3 . Appl. Organomet. Chem. 9, 507. Cullen, W.R., Reimer, K.J., 1989. Arsenic speciation in the environment. Chem. Rev. 89, 713. Dodd, M., Grundy, S.L., Reimer, K.J., Cullen, W.R., 1992. Methylated antimony(V) compounds: synthesis, hydride generation properties and implications for aquatic speciation. Appl. Organomet. Chem. 6, 207. Gosio, B., 1901. Arch. Ital. Biol. 35, 201. Jenkins, R.O., Craig, P.J., Goessler, W., Miller, D., Ostah, N., Irgolic, K.J., 1998a. Biomethylation of inorganic antimony compounds by an aerobic fungus ± Scopulariopsis brevicaulis. Environ. Sci. Technol. 32, 882. Jenkins, R.O., Craig, P.J., Goessler, W., Irgolic, K.J., 1998b. Biovolatilization of antimony and sudden infant death syndrome (SIDS). Hum. Exp. Toxicol. 17, 231. Koch, I., Feldmann, J., Lintschinger, J., Serves, S.V., Cullen, W.R., Reimer, K.J., 1998. Demethylation of trimethylantimony species in aqueous solution during analysis by hydride generation gas chromatography with AAS and ICP-MS detection. Appl. Organomet. Chem. 12, 129. Limerick, S., 1998. Expert group to investigate cot death theories: toxic gas hypothesis. National Research Council, 1999. Arsenic in drinking water. Parris, G.E., Brinckman, F.E., 1976. Reactions which relate to the environmental mobility of arsenic and antimony. II. Oxidation of trimethylarsine and trimethylstibine. Environ. Sci. Technol. 10, 1128. Richardson, B.A., 1994. Sudden infant death syndrome: a possible primary cause. J. Foren. Sci. Soc. 34, 199. Silver, S., 1996. Bacterial resistances to toxic metal ions ± a review. Gene 179, 9.