Lysed-cell extract of a Pseudomonas sp. was shown to catalyse bioconversion of dimethylarsinoylacetate to arsenobetaine and dimethylarsinate. Provision of the universal methyl donor S-adenosylmethionine to bioconversion mixtures promoted both the rate and extent of arsenobetaine formation. These findings suggest that in the proposed biosynthesis of arsenobetaine from dimethylarsinoylethanol, oxidation (i.e. the formation of the carboxymethyl group of dimethylarsinoylacetate) would precede the reduction and methylation at the arsenic atom. The presence of enzyme(s) capable of methylating dimethylarsinoylacetate in a bacterial isolate from marine mussel (Mylitus edulis), highlights a possible direct involvement of prokaryotic organisms in the biosynthesis of organoarsenic compounds within marine animals.
It has been known for some years that marine animals naturally contain appreciable quantities of arsenic and that, in most cases, this arsenic is present almost entirely as trimethylarsonioacetate (arsenobetaine; Fig. 1, 5) [1,2]. Arsenobetaine is metabolically stable – it is rapidly excreted through the human kidney, and presents no toxic hazard. Until relatively recently it was thought that the natural occurrence of arsenobetaine was restricted to the marine environment but it has now been shown to occur in mushrooms  and in earthworms  living in arsenic polluted environments.
Although arsenobetaine was discovered as early as 1977  and is widespread in marine biota, information about its biosynthesis is lacking. It is speculated that arsenobetaine is formed via degradation of arsenosugars. The overwhelming bulk of algal arsenosugars are dimethylarsinoylribosides (Fig. 1, 1) and thus contain two methyl groups and a 5-deoxyriboside group attached to the arsenic atom. The biosynthesis of the dimethylarsinoylribosides has been previously considered at length . It is very likely that S-adenosylmethionine (SAM) is the source of both the methyl groups and the ribose moiety attached to the arsenic atom. Probably the two-carbon (carboxymethyl) side-chain in arsenobetaine is derived from breakdown of the ribose-containing portion of the dimethylarsinoylribosides. Anaerobic microbial activity derived from marine sand very quickly converts algal dimethylarsinoylribosides to dimethylarsinoylethanol (Fig. 1, 2) which is readily seen as a precursor of arsenobetaine. Curiously, attempts to bring about the further reduction and methylation of the arsenic atom of this compound by the sort of unspecified microbial activity that facilitated its formation have been unsuccessful . In addition, fish when supplied with either dimethyarsinoylethanol or dimethylarsinoyl acetic acid (Fig. 1, 4) in their food also failed to accumulate them as arsenocholine (Fig. 1, 3) or arsenobetaine .
There have been, however, several studies of the microbial degradation of arsenobetaine [10,11]; we have recently reported on two bacterial isolates from marine mussel (Mylitus edulis) capable of degrading arsenobetaine to dimethylarsinate via dimethylarsinoylacetate . In the present work we report for the first time the biosynthesis of arsenobetaine from dimethylarsinoylacetate by a lysed-cell extract of one of these isolates.
Materials and methods
Materials and microorganisms
Arsenobetaine was prepared using the procedure described by Lagarde et al. . 2-Dimethylarsinoyl acetic acid was prepared according to the method of Wigren [14,15]. Sodium dimethylarsinate (Me2As(O)O−Na+) was obtained from Sigma–Aldrich. Pseudomonas fluorescens A NCIMB 13944, previously described as degrading arsenobetaine to dimethylarsinate via dimethylarsinoylacetate in pure culture , is deposited with the National Culture of Industrial and Marine Bacteria, Aberdeen, UK.
Culture and preparation of lysed-cell extracts
Pseudomonas sp. NCIMB 13944 was grown in 6 × 1 l flasks containing 200 ml minimal salts medium supplemented with sodium succinate (2 g l−1), in the absence of arsenic compounds, at 25 °C on a rotary shaker (150 rpm) for 3 days. The minimal salts medium contained (g l−1): KH2PO4 (2); NH4Cl (3); MgSO4· 7H2O (0.4); trace element solution (2 ml) ; NaCl (30), final pH, 6.8 adjusted with KOH. The trace element solution comprised (g l−1): ethylenediaminetetraacetic acid (50); ZnSO4· 7H2O (22); MnCl2· 4H2O (5.54); FeSO4· 7H2O (4.99); (NH4)6Mo7O2· 4H2O (1.1); CuSO4· 5H2O (1.57); CoCl2· 6H2O (1.61), final pH, 7.2. Cells were harvested by centrifugation at 6000g for 30 min at 4 °C, washed twice in 50 mM potassium phosphate buffer (pH 7.2), and resuspended in 30 ml of the buffer. Cells were disrupted using a MSE Soniprep 150 ultrasonic disintegrator (MSE Scientific Instruments, Crawley, UK), at an amplitude of 10–12 μm for 20 × 30 s, with cooling on ice between cycles. Debris was removed by centrifugation at 100,000g for 60 min; the supernatant was decanted and termed the lysed-cell extract.
Protein concentration was determined by the Bradford method, using Biorad (Hercules) protein assay reagent and bovine serum albumin (Sigma–Aldrich) as standards (0–0.2 mg protein ml−1). The protein concentration of the lysed-cell extract was 2.0 mg ml−1.
In vitro assay for arsenobetaine biosynthesis
Lysed-cell extract (0.1 ml) was incubated at 30 °C for up to 300 min in 0.9 ml of 50 mM potassium phosphate (pH 7.2) buffer. All incubations contained a final concentration of 1 mg As l−1 as dimethylarsinoylacetate; SAM was added to some assay mixtures to a final concentration of 1 mg l−1. Aliquots (0.1 ml) of assay mixtures were taken at intervals, diluted 10-fold in water and frozen at −20 °C ready for analysis. The arsenic compounds were determined by high-performance liquid chromatography/inductively coupled plasma mass spectrometry (HPLC/ICPMS).
An integrated Hewlett Packard HPLC/ICPMS system, fitted with a Babbington nebuliser was used. Anion-exchange chromatography was on a Hamilton PRP-X100 column (250 mm × 4.6 mm) at 30 °C, using 20 mM NH4H2PO4 as mobile phase (adjusted to pH 6.0 with 25% NH3 (aq)). Cation-exchange chromatography was on a Zorbax 300-SCX column (150 mm × 4.6 mm) at 30 °C, using 20 mM aqueous pyridine as mobile phase (adjusted to pH 2.6 with formic acid). For both columns, flow rate was 1.5 ml min−1 and sample injection volume was 20 μl.
Bioconversion of dimethylarsinoylacetate by a lysed-cell extract of the P. fluorescens A strain, both in the presence and absence of added SAM (supplied as potential methyl donor), is shown in Fig. 2. In both types of incubations, arsenobetaine and dimethylarsinate were formed as sole detectable arsenic containing products of dimethylarsinoylacetate bioconversion; recovery of arsenic in these products was ≥98% of that supplied in the form of dimethylarsinoylacetate. Confirmation of the formation of arsenobetaine from dimethylarsinoylacetate by the lysed-cell extract was confirmed by HPLC/ICPMS analysis on a cation-exchange column (Fig. 3). Co-injection of assay incubation mixtures and (organo)arsenic standards (arsenobetaine, dimethylarsinoylacetate, dimethylarsinate, methylarsinate, trimethylarsine oxide, arsenocholine, dimethylarsinoylethanol, tetramethylarsonium ion, arsenate and arsenite), with subsequent chromatographic separations on both anion- and cation-exchange columns, was also used to confirmed the formation of arsenobetaine from dimethylarsinoylacetate by the lysed-cell extract.
Supply of SAM promoted both the rate and extent of arsenobetaine formation (Figs. 2and 4). In the presence of SAM, the maximum specific rate of arsenobetaine formation was 6.0 μg As min−1 (mg protein)−1 (67% higher than in absence of SAM), with 134 μg of arsenic in the form of arsenobetaine after 180 minutes incubation (almost 4-fold higher than in absence of SAM). SAM however did not influence the maximum specific rate of dimethylarsinate formation (3.7 μg As min−1 (mg protein)−1), and after 180 minutes incubation had increased the amount of arsenic in this bioconversion product by only 23% to 94 μg As. After 300 minutes incubation in the presence of SAM, 17 and 15% of supplied arsenic (as dimethylarsinoylacetate) was recovered as arsenobetaine and dimethylarsinate respectively (Fig. 2).
Chemical synthesis of arsenobetaine involves nucleophilic attack of the trivalent arsenic atom of trimethylarsine on the bromine-bearing carbon of ethyl bromoacetate . Methylation of arsenic (e.g. by the bread mould Scopulariopsis brevicaulis– as in the ‘Challenger’ pathway ) follows a process that parallels the chemical synthesis, i.e. nucleophilic attack by a reduced (electron-rich) arsenic species on a positively charged carbon (the methyl-bearing carbon of SAM). However, the natural transfer of a carboxymethyl group to arsenic (necessary for the biosynthesis of arsenobetaine) by an analogous route is mechanistically unlikely.
The presence of arsenosugars in the early stages of marine food chains and their occurrence with arsenobetaine in some animals suggested that arsenosugars might be precursors of arsenobetaine in a biogenerative process. Considerable breakdown of that part of the molecule containing the sugar residue would have to occur. Ideally such a degradative process should leave a two-carbon chain attached to arsenic together with the two methyl groups and oxygen already present. A reduction followed by an additional methylation, together with oxidation of the new two-carbon side chain to form a carboxymethyl group will yield arsenobetaine. Considerable experimental support for such a scheme was provided by a non-specific anaerobic microbial treatment of marine algal fragments; the arsenosugars they contained were quantitatively converted to dimethylarsinoylethanol, a compound that could very easily be seen as a precursor of arsenobetaine in the terms just outlined . In particular, the ribosides attached to arsenic had been degraded to the required simple two-carbon unit. Thus it seemed feasible that all groups attached to arsenic in arsenobetaine (the three methyl and the carboxymethyl) were derived from a single source (SAM) and were transferred to arsenic by a single process, namely, nucleophilic attack by reduced arsenic on a positively charged carbon.
The demonstration of the microbial breakdown of dimethylarsinoylribosides to yield the necessary 2-carbon chain only provides a plausible biosynthetic route to arsenobetaine if the final reduction and methylation of the arsenic atom is possible. Although mechanistically there would seem to be no impediment to such a reaction as it parallels any single stage in the ‘Challenger’ pathway (reduction of arsenic followed by oxidative methylation) demonstration of its occurrence has not been achieved until now.
Additional support, in particular for the degradative aspects of this biosynthetic scheme for arsenobetaine, has recently been provided by a chromatographic and mass spectrometric study of the wide range of arsenic compounds present in the kidney of the giant clam Tridacna derasa. In this case both arsenosugars and arsenobetaine were present, and several compounds could easily be seen as intermediates. It appeared that in the clam kidney the aglycone chains of the arsenosugars were progressively degraded by oxidation and decarboxylation, the ribose ring was then opened and the resulting sugar alcohol chain degraded in the same manner. The penultimate product was dimethylarsinoylacetate, which could be converted to arsenobetaine by reduction and methylation at the arsenic atom. Earlier work  left it unclear whether in the conversion of dimethylarsinoylethanol to arsenobetaine, reduction and methylation at the arsenic atom preceded or followed the oxidation of the terminal carbon in the two-carbon side chain. The work reported here has indicated that the oxidation (i.e. the formation of the carboxymethyl group) precedes the reduction and methylation at the arsenic atom.
We have recently reported on the capability of P. fluorescens A NCIMB 13944, a bacterium isolated from Mytilus edulis (marine mussel), to degrade arsenobetaine in whole cell bioconversions by initial cleavage of a methyl-arsenic bond to form dimethylarsinoylacetate, with subsequent cleavage of the carboxymethyl-arsenic bond to yield dimethylarsinate . During this study on bacterial degradation of organoarsenic compounds, we noted that growth of P. fluorescens A NCIMB 13944 in the presence of dimethylarsinoylacetate resulted in 94% removal of the organoarsenical from culture supernatants with appearance of some dimethylarsinate and arsenobetaine (each ca. 1–2% of the As supplied) (data not shown). This stimulated us to investigate the possibility of arsenobetaine formation by a prokaryotic organism, using lysed-cell extract of the P. fluorescens A strain.
In the present study, we have found that a lysed-cell extract of this organism converts dimethylarsinoylacetate not only to dimethylarsinate but also to arsenobetaine. We have also demonstrated that SAM markedly promotes the bioconversion of dimethylarsinoylacetate to arsenobetaine, which suggests that the bioconversion is catalysed by a methyltransferase enzyme using SAM as methyl donor. We have suggested previously  that bioconversion of arsenobetaine to dimethylarsinoylacetate may be catalysed through the reversible action of a methyltransferase, and the data reported here support that view. Although bioconversion of dimethylarsinoylacetate by the P. fluorescens A strain does not require induction of substrate, preincubation of cells with inorganic arsenic or organoarsenic compound(s) could greatly enhance the specific activity of the enzymes involved.
As discussed above, two routes of biological formation of arsenobetaine from dimethylarsinoylethanol have been proposed: one involving arsenocholine the other involving dimethylarsinoylacetate as intermediate . The significance of the work reported here is that it provides evidence for enzymatic methylation of dimethylarsinoylacetate, a key step in arsenobetaine biosynthesis via one of these routes. The presence of such enzymatic activity in a bacterial species, isolated from marine mussel, indicates a possible direct involvement of prokaryotic organisms in biosynthesis of organoarsenic compounds within marine animals.
The Engineering and Physical Sciences Research Council of the UK supported this work.