Conserved cysteine residues determine substrate specificity in a novel As(III) S-adenosylmethionine methyltransferase from Aspergillus fumigatus

Abstract

Methylation of inorganic arsenic is a central process in the organoarsenical biogeochemical cycle. Members of every kingdom have ArsM As(III) S-adenosylmethionine (SAM) methyltransferases that methylates inorganic As(III) into mono- (MAs(III)), di- (DMAs(III)) and tri- (TMAs(III)) methylarsenicals. Every characterized ArsM to date has four conserved cysteine residues. All four cysteines are required for methylation of As(III) to MAs(III), but methylation of MAs(III) to DMAs(III) requires only the two cysteines closest to the C-terminus. Fungi produce volatile and toxic arsines, but the physiological roles of arsenic methylation and the biochemical basis is unknown. Here we demonstrate that most fungal species have ArsM orthologs with only three conserved cysteine residues. The genome of Aspergillus fumigatus has four arsM genes encoding ArsMs with only the second, third and fourth conserved cysteine residues. AfArsM1 methylates MAs(III) but not As(III). Heterologous expression of AfarsM1 in an Escherichia coli conferred resistance to MAs(III) but not As(III). The existence of ArsMs with only three conserved cysteine residues suggest that the ability to methylate MAs(III) may be an evolutionary step toward enzymes capable of methylating As(III), the result of a loss of function mutation in organisms with infrequent exposure to inorganic As(III) or as a resistance mechanism for MAs(III).

Graphical abstract

Abbreviated summary. ArsM As(III) S-adenosylmethionine (SAM) methyltransferases (ArsM) transform inorganic As(III) into methylarsenicals. Most bacterial and animal enzymes have four conserved cysteine residues that are required for methylation of As(III) to MAs(III). ArsM from Aspergillus fumigatus lacks one of the conserved cysteines and can only methylate and confer resistance to MAs(III), not As(III).

Introduction

Arsenic is a ubiquitous toxic and carcinogenic metalloid that is introduced into the environment primarily from geological sources and, to a less extent, through anthropogenic activities (Zhu et al., 2014). Arsenic speciation plays a significant role in its behavior and fate, and the various inorganic and organic arsenic species differ greatly in their toxicity to cells (Akter et al., 2005). In general, pentavalent arsenicals, whether inorganic or organic, are relatively nontoxic compared with the corresponding trivalent species. Trivalent inorganic arsenite, As(III), is more soluble, mobile and toxic than pentavalent species. It can be methylated enzymatically in three consecutive steps to considerably more toxic MAs(III), DMAs(III) and, to a lesser extent, nontoxic and volatile TMAs(III) (Marapakala et al., 2015, Dheeman et al., 2014, Le et al., 2000, Drobna et al., 2012). In microbes, arsenic methylation has been considered to be a detoxification process because the trivalent methylarsenicals are abiotically oxidized in air to the pentavalent species (Qin et al., 2009, Qin et al., 2006). Biomethylation of arsenic is widespread in nature and has been observed in bacteria (Qin et al., 2006, Yin et al., 2011), archaea (Wang et al., 2004 Wang et al., 2014), fungi (Bentley & Chasteen, 2002), algae (Qin et al., 2009), fish (Hamdi et al., 2012), and mammals (Lin et al., 2002), including humans (Waters et al., 2004). Arsenic biomethylation is catalyzed by the enzyme As(III) S-adenosylmethionine methyltransferase, which has been termed ArsM in microbes (Qin et al., 2006) and AS3MT in animals (Waters et al., 2004). For simplicity ArsM will be used as a general term for this enzyme family in this article. ArsM catalyzes three consecutive transfers of the methyl group from SAM to As(III). Bacterial, algal and mammalian ArsMs studied to date all have four conserved cysteine residues. Functionally, the two closest to the N-terminus form one pair, and the other two form a second pair (Marapakala et al., 2015, Marapakala et al., 2012, Dheeman et al., 2014). For clarity, they are labeled CysA, CysB, CysC and CysD according to their distance from the N-terminus. All four cysteines are required for the first methylation step (As(III)→MAs(III)), but the second methylation step (MAs(III)→DMAs(III)) requires only the second pair, CysC and CysD. CysC and CysD form the binding site for trivalent arsenicals, explaining why the enzyme is inactive in their absence. Mechanistically CysA and CysB are thought to be involved reduction of an enzyme-bound MAs(V) intermediate, producing a disulfide bond between CysA and CysB and an enzyme-bound MAs(III) complex (Dheeman et al., 2014, Marapakala et al., 2015). The disulfide bond is reduced by thioredoxin, regenerating the active form of the enzyme. Thus the two conserved cysteine pairs are considered essential for ArsM function.

Aspergillus fumigatus Af293, a common soil fungus often found in decaying organic matter such as compost heaps has four AfarsM genes, each with only CysB, CysC and CysD. Cells of the arsenic hypersensitive E. coli strain AW3110 (∆ars) expressing AfarsM1 are resistant to MAs(III) but not As(III). Purified AfArsM1 rapidly methylated MAs(III), but As(III) methylation was nearly absent. A C55S derivative methylated MAs(III), but neither C143S nor C195S mutants showed activity. These results demonstrate that the AfArsM1 of A. fumigatus Af293 is a functional MAs(III) SAM methyltranserase with little ability to methylate inorganic As(III). We postulate that soil fungi such as Aspergillus evolved three-cysteine ArsMs to cope with environmental MAs(III) produced by other soil microbes.

Results

Phylogeny of three-cysteine ArsMs

A phylogenetic analysis of ArsM sequences in 55 fungal species was conducted by aligning the sequence of AfArsM1 with orthologous sequences identified by BLAST searches in complete genomes (Fig. 1S). The majority of arsM gene products appear to have only three of the four conserved cysteine residues. In the 10 most closely related sequences, these correspond to CysB, CysC and CysD (Fig. 2S). These fungal ArsMs have low similarity to most bacterial ArsMs and animal AS3MTs. A representative fungus is Aspergillus fumigatus, a common soil organism. In the eight chromosomes of A. fumigatus Af293 there are six clusters of ars genes that are similar to their bacterial counterparts (Fig. 3S). Included in the ars clusters on chromosomes 1, 3, 5 and 8 are four AfarsM genes. AfArsM1 exhibits 93% sequence identity with AfArsM2, 79% identity with AfArsM3 and 56% identity with AfArsM4. Two closely-related orthologs with four cysteine residues are from the fungus Glarea lozoyensis ATCC 20868 (45%) and the bacterium Rhodopseudomonas palustris (43%) (Fig. 1). Comparing AfArsM1 with animal and plant orthologs, it shares only 26% identity with human AS3MT and 29% with algal CmArsM (Cyanidioschyzon sp). This low similarity indicates a more distant evolutionary relationship of fungal ArsMs with other eukaryotic As(III) SAM methyltransferases. Only four fungal species have ArsM sequences with four conserved cysteines: Glarea lozoyensis ATCC 20868, Trichosporon oleaginous, Rhizophagus irregularis DAOM 181602 and Spizellomyces punctatus DAOM BR117 (indicated by closed triangles in Fig. 1S). A. fumigatus Af293 AfArsM1, 2 and 3 cluster in one group, while AfArsM4 is in a more distant group. Similarly, two other fungal species, A. lentulus and Neosartorya fischeri NRRL 181, have multiple ArsMs that fall into the same two separate groups as the AfArsMs. This distribution in two groups in multiple organisms suggests that the ancestral genes may be been acquired by separate horizontal gene transfers from bacterial sources. The ancestor of AfArsM1 may have given rise to AfArsM1, 2 and 3 by gene duplications, leading to multiple copies in A. fumigatus Af293. To trace the bacterial origin of AfArsM, a phylogenetic analysis of ArsM sequences of 55 representative prokaryotes and eukaryotes organisms was conducted (Fig. 2). On the whole, the fungal sequences were most closely related to bacterial sources from Planctomycetes and Acidobacteria, which also have three conserved cysteines. These two bacteria sequences are quite distant from other prokaryotic ArsMs with four conserved cysteines. The fungal ArsM of Glarea lozoyensis ATCC 20868 (XP_008084726.1) has four conserved cysteines but clustered with other fungi. On the other hand, the fungal ArsMs from T. oleaginous, R. irregularis and S. punctatus, which each have four conserved cysteines, did not cluster with other fungi, suggesting that those three ArsMs were acquired from bacterial sources more recently than the initial horizontal gene transfer to fungi. This analysis suggests that the ArsM of most fungi originated from Planctomycetes and Acidobacteria.

Fig. 1. Multiple alignment of AfArsM orthologs.

ArsM sequences were compared from A. fumigatus Af293 AfArsM1 (XP_753155.1); AfArsM2 (XP_752977.1); AfArsM3 (XP_747465.1); Glarea lozoyensis ATCC 20868 (XP_008084726.1) and Rhodopseudomonas palustris (WP_011159102.1). Identities are highlighted in black, and conservative replacements in grey.

Fig. 2. A neighbor-joining phylogenetic tree showing the evolutionary relationships of fungi ArsM protein with arsenic methyltransferase proteins from members of other kingdoms.

ArsM sequences with three conserved cysteines are highlighted by closed squares. Fungal ArsM sequences with four conserved cysteines are indicated by black triangles.

AfArsM1 methylates and confers resistance to MAs(III) but not As(III)

To examine the physiological role of AfArsM1 in arsenic resistance, the DNA sequence of the AfarsM1 gene located in chromosome 5 was chemically synthesized, cloned into vector pET32a (pET32a-AfarsM) and expressed in the As(III)-hypersensitive E. coli strain AW3110(Δars). In M9 minimal medium, E. coli strain AW3110 is sensitive to 50 μM As(III) and 10 μM MAs(III) (Fig. 3). Expression of the synthetic AfarsM1 gene conferred resistance to MAs(III) but did not complement As(III) sensitivity. AfArsM1 was then expressed in E. coli BL21(DE3), and the products of methylation analyzed. After 30 min of incubation, arsenic in the supernatant solution was speciated by HPLC ICP-MS. When cells were incubated with 4 μM MAs(III), 90% was methylated to the trimethylated species in 30 min (Fig. 4A). In these experiments the reactions were terminated H₂O₂, which oxidizes and solubilizes the products. In contrast, when the cells were incubated with inorganic As(III), all of the arsenic was recovered as As(V) after H₂O₂ treatment (Fig. 4A). These results demonstrate that cells expressing AfArsM1 can carry out the second and third methylation steps, transforming MAs(III) to TMAs(III) but does not transform As(III) into MAs(III). AfArsM1 was purified from E. coli and assayed for methylation activity. The enzyme methylated MAs(III) in a time-dependent manner (Fig. 4A). but As(III) methylation was nearly absent (Fig. 4B). These results conclusively demonstrate that the preferred substrate of AfArsM1 is MAs(III), not inorganic As(III).

Fig. 3. AfArsM1 confer MAs(III) resistance, not As(III).

E. coli strain AW3110 bearing either pET32a-AfArsM1 (open symbols) or vector plasmid pET32a (closed symbols) were grown in M9 minimal medium with 0.3 mM IPTG for the indicated times at 37 °C with shaking, as described in Experimental Procedures. Additions: none (circles), 50 μM As(III) (triangles) or 10 μM MAs(III) (squares). Growth was estimated from A_600nm. Data are the mean ± SE (n = 3).

Fig. 4. Methylation of MAs(III) or As(III) by AfArsM1 in vivo and in vitro.

Methylation reactions in vivo and in vitro were assayed as described in Experimental Procedures. All reactions were terminated by addition of 6% (v/v) H₂O₂, final concentration. Arsenicals were speciated by HPLC using a C18 reverse phase column, and the amount of arsenic was estimated by ICP-MS. A) AfArsM1 methylates MAs(III) but not As(III) in vivo. Methylation As(III) and MAs(III) were assayed by cells of E. coli BL21(DE3) expressing AfarsM1 grown in M9 minimal medium with 0.3 mM IPTG for 30 min times at 37 °C, with shaking, as described in Experimental Procedures. Samples were treated with 6 % (v/v) H₂O₂, final concentration. B) AfArsM1 methylates MAs(III) but not As(III) in vitro. AfArsM1 was purified and assayed for methylation of MAs(III) or As(III), as described in Experimental Procedures. The reaction mixture (1 ml) containing 2 μM purified AfArsM1, 1 mM SAM, 8 mM GSH and either 10 μM MAs(III) or As(III) was incubated at 37 °C for the indicated times.

Role of conserved cysteine residues in MAs(III) methylation

Most fungal As(III) SAM methyltransferases have only three conserved cysteine residues (Fig. 1S), which are Cys55 (CysB), Cys143 (CysC) and Cys195 (CysD) in AfArsM1 (Fig. 1). To examine their role in MAs(III) methylation, each of the three cysteines was altered individually to serine residues. When the three derivatives were expressed in E. coli AW3110, cells with the C55S derivative methylated MAs(III), but no methylation was observed in cells expressing either the C143S or C195S mutants, consistent with the role of CysC and CysD in binding trivalent arsenicals (Table 1 and Fig. 4SA). These results clearly demonstrate that cells expressing AfArsM1 can methylate and detoxify MAs(III) but not As(III).

Table 1. Methylation of MAs(III) by AfArsM1 mutants in vivo and in vitro.

MAs(III) methylation by AfArsM1 mutants was assayed in vivo and in vitro.

AfArsM1 variants	MAs(III) concentration (μM)	Products found in culture medium (μM)a
		TMAsO(V)	MAs(V)	DMAs(V)
WT	4b	1.61±0.04	0.13±0.02	0.58±0.03
		(40.3±1.0%)d	(3.3±0.5%)	(14.5±0.8%)
C55S		1.52±0.07	0.14±0.03	0.62±0.05
		(38.0±1.8%)	(3.5±0.8%)	(15.5±1.3%)
C143S		NDe	1.73±0.08	0.01±0.00
			(43.3±2.0%)	(0.03±0.0%)
C195S		ND	1.79±0.09	0.01±0.00
			(44.7±2.3%)	(0.03±0.0%)
Control	10c	ND	3.91±0.05	ND
			(97.8±1.3%)
WT		0.35±0.03	1.21±0.06	2.19±0.13
		(8.8±0.8%)	(30.3±1.5%)	(54.8±3.3%)
C55S		0.31±0.03	1.25±0.07	2.21±0.09
		(7.9±0.8%)	(31.3±1.8%)	(55.3±2.3%)
C143S		ND	3.86±0.17	ND
			(96.5±4.3%)
C195S		ND	3.87±0.15	ND
			(96.8±3.8%)
AfArsM258		0.11±0.02	1.24±0.07	2.45±0.15
		(2.8±0.5%)	(31.0±1.5%)	(61.3±3.8%)
C240/241S		0.18±0.03	1.23±0.04	2.39±0.17
		(4.5±0.8%)	(30.8±1.0%)	(59.8±4.3%)

^aMethylation activity of E. coli cells expressing AfarsM1 or variants were assayed in vivo and in vitro, as described in Materials and Methods. Assays were incubated with MAs(III) at 4 μM in vivo or 10 μM in vitro, final concentration. All the samples were treated with 6% (v/v) H₂O₂, final concentration, and separated by HPLC using a C18 reverse phase column, and the amount of arsenic was estimated by ICP-MS. Data are the mean ± SE (n = 3).

^bMAs(III) methylation in vivo was assayed for 20 min in cells of E. coli expressing wild type and the C55S, C143S and C195S mutant AfarsM1 genes.

^cMAs(III) methylation was assayed for 40 min with purified wild type AfArsM1 and the C55S, C143S, C195S, C240/241S double mutant and AfArsM1₂₅₈.

^dNumbers in parentheses are the percentage of added arsenic.

^eND: non-detectable.

Each of the mutant proteins was purified, and methylation activity assayed (Table 1). Purified derivative C55S methylated MAs(III), but neither the C143S nor C195S enzymes exhibited methylation activity (Table 1 and Fig. 4SB). AfArsM1 has three additional nonconserved cysteine residues at positions 240, 241 and 280. To examine their contribution to methyltransferases activity, Cys240 and Cys241 were simultaneously changed to serine residues, creating a double AfArsM1_C240/241S derivative. In addition, a deletion mutant containing only the first 258 residues, AfArsM1₂₅₈ (which lacks Cys280), was constructed. This truncation was based on the CmArsM truncation that retains methylation activity (Marapakala et al., 2012). Each of the three purified proteins derivatives methylated MAs(III), although at a somewhat reduced levels (Table 1 and Fig. 4SC). These results clearly demonstrate that only Cys143 and Cys195, but neither conserved residue Cys55 nor the non-conserved cysteine residues, are required for MAs(III) methylation.

AfArsM1 binds but does not methylate inorganic As(III)

Conserved residues Cys143 and Cys195 are in the CysC and CysD positions, which correspond to the arsenic binding site observed in the crystal structure of CmArsM (Ajees et al., 2012). Although AfArsM1 does not methylate As(III), it is possible that it still binds inorganic As(III). To examine this possibility, 5 μM AfArsM1 was incubated with either 20 μM As(III) or MAs(III) for 10 min at 37°C to obtain the substrate-bound form. Unbound arsenic was removed by gel filtration using a Bio-Gel P-6 spin column, and the protein with bound As(III) or MAs(III) was denatured with guanidine HCl to release protein-bound arsenic (Dheeman et al., 2014). Denatured protein was removed by filtration, and the filtrate was speciated by HPLC-ICP-MS (Fig. 5A). The results show that either As(III) or MAs(III) bind to AfArsM1. In addition, As(III) inhibited MAs(III) methylation (Fig. 5B). MAs(III) methylation activity decreased as the concentration of inorganic As(III) increased. These results indicate that As(III) competes with MAs(III) for binding to Cys143-Cys195 even though it is not itself methylated. These results are consistent with the observation that both CysA and CysB residues are required for methylation of inorganic As(III).

Fig. 5. Competition of As(III) with MAs(III) for binding to AfArsM1.

A) As(III) and MAs(III) both bind to AfArsM1. Purified AfArsM1 (5 μM) was incubated at 37° C with 20 μM As(III) or MAs(III) containing 8 mM GSH. After 10 min, samples were separated through a Bio-Gel P-6 spin column. Portions (25 μl) were diluted with 6 M guanidine HCl to denature the protein and release arsenicals bound to the enzyme. The samples were analyzed by HPLC-ICP-MS with a C18 reverse phase column. B), As(III) inhibit AfArsM1 methylation activity with MAs(III) substrate. The reaction mixture (1 ml) containing 2 μM purified AfArsM, 1 mM SAM, 8 mM GSH 10 μM MAs(III) and As(III) as indicated. Reaction was assayed at 37°C for 40 min. samples were treated with 6% (v/v) H₂O₂, final concentration, and separated by HPLC using a C18 reverse phase column, and the amount of arsenic was estimated by ICP-MS.

MAs(V) is reduced to MAs(III) prior to methylation

What is the physiological role of a three-cysteine ArsM? We speculate that soil microbes reduce MAs(V) to generate MAs(III) as an antibiotic that kills off their competitors (Li et al., 2016). In response, the target organisms have evolved resistance mechanisms to escape MAs(III) toxicity. While application of the commercial herbicide MSMA (monosodium methylarsenate) introduces MAs(V) into soil, how is MAs(V) produced biologically? Soil microbes with four-cysteine ArsMs synthesize MAs(III), which is excreted and oxidized in soil and water to MAs(V). In microbial soil communities some bacteria such as the environmental isolate Burkholderia sp. MR1 reduce pentavalent arsenicals, and other soli organisms methylate or demethylate MAs(III) (Yoshinaga et al., 2011). To test this idea, we created an artificial microbial community by co-culturing E. coli Bl21(DE3) harboring the AfarsM1 gene with Burkholderia sp. MR1. When the mixed culture was incubated with MAs(V), both DMAs(V) and TMAs(V)O were produced (Fig 5S). In the absence of Burkholderia sp. MR1, no methylation was observed. These results are consistent with our hypothesis that the physiological role of three-cysteine ArsMs is detoxification of MAs(III) generated in microbial soil communities.

Discussion

Genes for ArsM As(III) SAM methyltransferases are widespread and found in members of every kingdom. A relatively small number of ArsM orthologs have been characterized. The majority of those have four conserved cysteine residues in the A, B, C and D positions from the N-terminus. The role of those cysteine residues has been examined in the algal CmArsM from Cyanidioschyzon sp (Marapakala et al., 2015, Marapakala et al., 2012) and human hAS3MT ortholog (Dheeman et al., 2014). In both enzymes CysA and CysB are both required for methylation of As(III) but not MAs(III). From the crystal structure of CmArsM, two cysteines are located in the SAM binding domain. CysC and CysD are in the arsenic binding domain and are required for methylation of As(III) and MAs(III). In this study we identify a class of enzyme with only three of the conserved cysteine residues, CysB, CysC and CysD. For the most part these are fungal enzymes, although there are a small number of bacterial ones that cluster with them, suggesting horizontal transfer of the gene for a three-cysteine ArsM from bacteria to fungi. Some fungi have four-cysteine arsM genes, but those appear to be divergent from other fungal ArsMs, perhaps the result of a second horizontal gene transfer event.

In this report we show that the three-cysteine AfArsM1 does not methylate inorganic As(III) at a measurable rate even though it binds As(III). In contrast, AfArsM1 methylates MAs(III) at rates comparable with four-cysteine orthologs. Even though binding of As(III) inhibits MAs(III) methylation, A. fumigatus has three arsB genes that encoded As(III) efflux permeases, which would reduce the intracellular concentration of As(III) to levels that would not inhibit MAs(III) methylation. From our previous studies with three-cysteine derivates of four-cysteine CmArsM and hAS3MT, it appears that the three-cysteine enzymes methylate As(III) very slowly, perhaps a few percent of the activity of the wild type, while the rate of MAs(III) methylation is nearly wild-type (Dheeman et al., 2014, Marapakala et al., 2015). Without either CysA or CysB, the enzymes are unable to form a disulfide bond that reduces a putative pentavalent enzyme-bound intermediate. In the absence of the disulfide bond, the rate at which the enzyme progresses to the second methylation is extremely slow, and we predict that AfArsM1 and other three-cysteine enzymes can methylate As(III) but at only a fraction of the rate of MAs(III). In this regard, an ArsM ortholog from a soil fungus, Westerdykella aurantiaca, was recently shown to methylate As(III) even though it only has three conserved cysteine residues (Verma et al., 2016). It is not clear if the rate of methylation of inorganic As(III) methylation is comparable with the rate of MAs(III) with this fungal enzyme.

What is the pathway of evolution of three- and four-cysteine ArsMs? Since AS3MT with only CysC and CysD is capable of methylating MAs(III) rapidly and As(III) slowly (Dheeman et al., 2014), it is reasonable to consider that a two-cysteine ArsM with only CysC and CysD was the ancestor of three- and four-cysteine enzymes. A mutation generating a third cysteine (either CysA or CysB) may have been a neutral event, but a subsequent mutation introducing the fourth cysteine was a beneficial event, and that arsM gene would have rapidly replaced the genes for two- and three-cysteine enzymes. Are the three-cysteine ArsMs found in most fungi and some bacteria the predecessor of the four-cysteine enzymes or the result of a loss-of-function mutation? While there are no data to distinguish between these two possibilities, since fungal arsB and acr3 genes could confer substantial As(III) resistance, we infer that the primary function of fungal ArsMs is to confer MAs(III) resistance. We have shown that a number of soil microbes produce MAs(III), either by methylation of As(III) (Qin et al., 2009, Qin et al., 2006) or by reduction of MAs(V) (Yoshinaga et al., 2011). We proposed that these organisms produce MAs(III) as a primitive antibiotic (Li et al., 2016). In 1942 Selman Waksman defined “antibiotic” as any substance produced by one microbe that inhibits growth of other microbes. By this definition, MAs(III) is an antibiotic, and a three-cysteine ArsM that transforms it to TMAs(III), which is nontoxic (Cullen, 2005), is an antibiotic resistance. Thus we envision that the role of three-cysteine ArsMs is to protect against the antibiotic generated by MAs(III) producers.

Finally, why does Aspergillus fumigatus have so many clusters of ars genes, including four arsM genes? It is not unusual to find multiple ars clusters in a single organism (Andres & Bertin, 2016), and many fungi have multiple copies of the arsM gene. On the one hand, several ars clusters may confer additive tolerance to arsenic (Ordóñez et al., 2005). On the other hand, multiple clusters may make it possible to express a cellular activity in a wider range of environmental conditions. For example, Pseudomonas putida KT2440 bears two virtually identical arsRBCH operons, but they are expressed at different temperatures (Paez-Espino et al., 2014). Thus, the number and expression of arsenic-related genes, and their acquisition, may depend on the selective pressure of varying environmental stresses.

Experimental Procedures

Chemicals

Unless otherwise indicated, chemicals were purchased from Sigma-Aldrich. MAs(V) was obtained from Chem Service, Inc., West Chester, PA, respectively. MAs(V) was reduced as described (Reay & Asher, 1977).

Strains, media and growth conditions

E. coli Stellar™ (Clontech Laboratories, Inc., Mountain View, CA) (F⁻, endA1, supE44, thi-1, recA1, relA1, gyrA96, phoA, Φ80d lacZΔ M15, Δ(lacZYA-argF)U169, Δ(mrr-hsdRMS-mcrBC), ΔmcrA, λ–) was used for plasmid DNA construction and replication. E. coli AW3110 (Δars∷cam F⁻ IN(rrn-rrnE) (Carlin et al., 1995), which is hypersensitive to As(III), was used for complementation studies. E. coli strains were grown aerobically at 37 °C in either Luria-Bertani (LB) medium or M9 medium (Sambrook et al., 1989), as noted, supplemented with 125 μg/ml ampicillin or 34 μg/ml chloramphenicol, as required. Bacterial growth was monitored by measuring the optical density at 600 nm (A_600nm).

Synthesis of AfarsM1 gene and construction of expression

An AfarsM1 gene corresponding to the mRNA sequence in NCBI (XM_748062.1) was chemically synthesized with 5′ NcoI and 3′ HindIII sites and subcloned into the EcoRV site of pUC57-Kan (GenScript, NJ, USA). The synthetic AfarsM1 gene was cloned as an NcoI/HindIII double digested fragment from pUC57-Kan-AfarsM into expression vector pET32a(+) that produces a fusion with a 109-residue thioredoxin and six-histidine tags at the 5′ end. All sequences were confirmed by commercial DNA sequencing (Sequetech, Mountain View, CA).

Phylogenetic analysis

Multiple alignment of ArsM homolog sequence was performed using Clustal Omega (http://www.ebi.ac.uk/Tools/msa/clustalo/). ArsM sequences with conserved cysteines were selected for phylogenetic analysis. Acquisition of sequences was performed by searching a list of reference organisms or from the National Center for Biotechnology (NCBI) protein database by BLASTP search (Johnson et al., 2008). Phylogenetic analysis was performed to infer the evolutionary relationship among the As(III) SAM methyltransferases of various organisms. The phylogenetic tree was constructed using the Neighbor-Joining method using MEGA 6.0.1 (Tamura et al., 2013). The statistical significance of the branch pattern was estimated by conducting a 1000 bootstrap (Saitou & Nei, 1987).

Mutagenesis of the AfarsM1 gene

AfArsM1 mutations were generated by site-directed mutagenesis using a Quick Change mutagenesis kit (Stratagene, La Jolla, CA). The mutagenic oligonucleotides used for both strands and the respective changes introduced (underlined) are listed in Table 1S. The codons for conserved residues Cys55, Cys143 and Cys195 were individually changed to serine codons, generating three single AfArsM1 mutants. The nonconserved Cys240 and Cys241 residues were simultaneously mutated to serine codons, producing a double mutant (AfArsM1_C240/241S). From multiple sequence alignment of putative ArsM orthologs, it appears that there is little sequence conservation in the C-terminal domain (Fig. 1S). For that reason, an AfArsM1 derivative encoding only the N-terminal 258 residues, designated AfArsM1₂₅₈, was constructed. Each AfArsM1 derivative was confirmed by commercial DNA sequencing (Sequetech, Mountain View, CA).

Arsenic resistance assays

For metalloid resistance assays, competent cells of AW3110 (DE3) were transformed with constructs with or without an AfarsM1 gene. Cells were grown overnight with shaking at 37 °C in LB medium with 100 μg ml⁻¹ ampicillin. Overnight cultures were diluted 100-fold in M9 medium containing various concentrations of As(III) or MAs(III) plus 0.3 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) and incubated at 37 °C with shaking for an additional 24 h. Growth was estimated from the absorbance at 600 nm.

Expression and purification of AfArsM1

E. coli BL21(DE3) bearing AfArsM1 in vector plasmid pET32a were grown in LB medium containing 100 μg ml⁻¹ ampicillin with shaking at 37 °C. At an A_600nm of 0.6, 0.3 mM IPTG was added as an inducer, and the culture was grown for an additional 4 h at 37 °C. The cells were harvested and suspended in 5 ml per gram of wet cells in buffer A (50 mM 4-morpholinepropanesulfonic acid, 20 mM imidazole, 0.5 M NaCl, 10 mM 2-mercaptoethanol and 20% glycerol (vol/vol), pH 7.5). The cells were broken by a single passage through a French pressure cell at 20,000 psi, and immediately treated with diisopropyl fluorophosphate (2.5 μl per gram wet cell). Membranes and unbroken cells were removed by centrifugation at 150,000 × g for 1 h, and the supernatant solution was loaded onto a Ni²⁺-nitrilotriacetic acid column (Qiagen, Valencia, CA) at a flow rate of 0.5 ml min⁻¹. The column was washed with more than 25 column volumes of buffer A. AfArsM1 was eluted with buffer A containing 0.2 M imidazole, and the purity was analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS PAGE). Protein concentrations were estimated from A_280nm (ε = 39080 M⁻¹ cm⁻¹). AfArsM1-containing fractions were divided into small portions, rapidly frozen, and stored at −80 °C until use.

Assay of arsenic methylation activity

Methylation of As(III) and MAs(III) was assayed both in cells of E. coli expressing AfarsM1 genes and by purified ArsM as described previously (Marapakala et al., 2012). For in vivo methylation activity assays, E. coli cells expressing AfarsM1 or variants were grown to A_600nm = 2 with 0.3 mM IPTG as an inducer at 37°C with aeration in LB medium. The cells were harvested and washed once with M9 medium, then suspended in M9 medium supplemented with 0.2% D-glucose and cell density was adjusted to A600 = 3.0. As(III) or MAs(III) were individually added at 4 μM, final concentration. Methylation activity was analyzed at the indicated time… Methylation activity of purified AfArsM1 was assayed at 37 °C in buffer consisting of 50 mM MOPS, pH 7.5, containing 0.3 M NaCl, 8 mM GSH and 1mM SAM. The reactions were terminated by adding 6% (v/v) H₂O₂ and heated at 80 °C for 5 min to oxidize all arsenic species. Denatured protein was removed by centrifugation using a 3 kDa cutoff Amicon ultrafilter. The filtrate was speciated by HPLC-ICP-MS. Where noted, H₂O₂ was not added to allow for determination of trivalent arsenicals. Without oxidation, the majority of the arsenic at early times was found to be bound to the enzyme. To determine the nature of AfArsM1-bound arsenicals, purified AfArsM1 (5 μM) was incubated at 37 °C with 20 μM As(III) or MAs(III) in 50 mM MOPS, containing 0.3 M NaCl and 8 mM GSH for 10 min. The samples were passed through a Bio-Gel P-6 column pre-equilibrated with reaction buffer, and portions (25 μl) were immediately diluted with 6 M guanidine HCl to denature the protein and release bound arsenicals. Arsenic was speciated by high pressure liquid chromatography (HPLC) (PerkinElmer Series 2000) using a C18 reversed-phase column eluted with a mobile phase consisting of 3 mM malonic acid, 5 mM tetrabutylammonium hydroxide, and 5% (v/v) methanol (pH 5.6) with a flow rate of 1 ml/min, and arsenic content was determined by inductively coupled plasma mass spectrometry (ICP-MS) using an ELAN DRC-e spectrometer (PerkinElmer, Waltham, MA) (Qin et al., 2006).

Abbreviations

MAs(III)

Methylarsenite

MAs(V)

methylarsenate

DMAs(V)

dimethylarsenate

TMAs(III)

trimethylarsenite

TMAs(V)O

trimethylarsenate

MSMA

monosodium methanearsenate

SAM

S-adenosylmethionine

HPLC

high pressure liquid chromatography

ICP-MS

inductively coupled plasma mass spectroscopy