Bacterial resistances to toxic metal ions - a review

GENE AN INTERNATIONAL

GENES

ELSEVIER

dOURNAL

ON

AND GENOMES

Gene 179 (1996) 9-19

B a c t e r i a l r e s i s t a n c e s to t o x i c m e t a l i o n s - a r e v i e w 1 S i m o n Silver * Department of Microbiology and Immunology, University of Illinois at Chicago, M/C 790, Room 703, 835 S. Wolcott Avenue, Chicago, IL60612-7344, USA

Abstract

Bacterial plasmids encode resistance systems for toxic metal ions, including Ag +, AsO2, AsO 3-, Cd 2+, Co 2+, CrO 2-, Cu 2+, Hg 2+, Ni 2+, Pb 2+, Sb 3+, TeO32-, TI + and Zn 2+. The function of most resistance systems is based on the energy-dependent efflux of toxic ions. Some of the efflux systems are ATPases and others are cherniosmotic cation/proton antiporters. The Cd2+-resistance ATPase of Gram-positive bacteria (CadA) is membrane cation pump homologous with other bacterial, animal and plant P-type ATPases. CadA has been labeled with 31p from [~-3/p]ATP and drives ATP-dependent Cd 2+ (and Zn 2+) uptake by inside-out membrane vesicles (equivalent to efflux from whole cells). Recently, isolated genes defective in the human hereditary diseases of copper metabolism, namely Menkes syndrome and Wilson's disease, encode P-type ATPases that are more similar to bacterial CadA than to other ATPases from eukaryotes. The arsenic resistance etttux system transports arsenite [As(III)], alternatively using either a double-polypeptide (ArsA and ArsB) ATPase or a single-polypeptide (ArsB) functioning as a chemiosmotic transporter. The third gene in the arsenic resistance system, arsC, encodes an enzyme that converts intracellular arsenate [As(V)] to arsenite [As(IIl)], the substrate of the efItux system. The triple-polypeptide Czc (Cd 2+, Zn 2+ and Co 2+) chemiosmotic efflux pump consists of inner membrane (CzcA), outer membrane (CzcC) and membrane-spanning (CzcB) proteins that together transport cations from the cytoplasm across the periplasmic space to the outside of the cell.

Keywords: Arsenic; Bacterial plasmids; Cadmium; Mercury; Ettlux; Antiporter; Menkes syndrome

1. Introduction Not quite, but almost! Bacteria have specific genes for resistances to the toxic ions of heavy metal elements * Corresponding author. Tel.: + l 312 9969608; Fax: + 1 312 9966415; e-mail: [email protected] 1Presented at the Chulabhorn Research Institute International Conference on 'Biotechnology Research and Applications for Sustainable Development (BRASD)', Central Plaza Hotel, Bangkok, Thailand, 7-10 August 1995. Abbreviations: aa, amino acid(s); Ars, arsenic resistance; ars, gene(s) encoding Ars; ATPase, adenosine triphosphatase; bp, base pair(s); Cad, cadmium resistance; cad, genes encoding Cad; Chr, chromate resistance; chr, genes encoding Chr; Cnr, cobalt and nickel resistance; cnr, genes encoding Cnr; Cop, copper resistance in Pseudomonas and copper transport in Enterococcus hirae (both); Cut, chromosomal gene products involved in copper uptake, intraeellular movement and etttux; Czc, cadmium, zinc and cobalt resistance; E., Escherichia; FAD, ravin adenine dinucleotide; kb, kilobase(s), or 1000 bp; Kin, half-saturation concentration; NADPH, nicotinamide adenine dinucleotide phosphate, reduced; Met, mercury resistance; mer, genes encoding Mer; Pco, E. coli plasmid copper resistance, homolog of Pseudomonas Cop; S., Staphylococcus; Smt, Synechoeoccus metallothionein; smt, genes encoding Smt synthesis and its regulation. 0378-1119/96/$15.00 © 1996 Elsevier Science B.V. All rights reserved Pll S 0 3 7 8 - 1 1 1 9 ( 9 6 ) 0 0 3 2 3 - X

including Ag +, AsO~-, AsO 3-, Cd 2+, Co 2+, CrO42 , C u 2+, H g 2+, Ni 2+, Sb 3 +, YeO32- , T1 + and Zn z+. There are reports (but less than satisfactory basis) for resistances to B4072-, Pb 2+, and organotin compounds. These for the moment are not discussed further. Figure 1 presents a current summary of resistance systems and their biochemical mechanisms. Mostly, these resistance systems have been found on plasmids, but frequently related systems are subsequently found determined by chromosomal genes in other organisms (examples are mercury resistance in Bacillus, cadmium efflux mediated by a P-type ATPase, again in Bacillus, and arsenic efflux by chromosomal E. coli genes). This leaves out Group Ia (such as Na and K), IIA (such as Ca and Mg), and higher atomic number elements (the lanthanides and actinides, including uranium and transuranium elements) of the Period Table, as lacking genes for metal ion resistances. There also are no resistance genes for Group VIIa halides, although halides are abundant in the environment and toxic in higher concentrations. It remains a long-time hypothesis of this laboratory that toxic metal resistance systems arose shortly after prokaryote life started, in an already metal-polluted world. As with intermediate

10

S. Silver/Gene179 (1996) 9 19

l' ×lG

PLASMID HEAVY METAL RESISTANCE SYSTEMS AND MECHANISMS

1. Hg z'. mer.

Hg 2' and organomercurials are enzymatically

detoxified. 7'~

2. As043, AsOz. ars. Arsenate is enzymatically reduced to arsenite

by ArsC. Arsenite is " p u m p e d " out by the membrane protein ArsB that functions chemiosmotically alone or with the additional ArsA protein as an ATPase. (~'~

3. Cd 2 . cadA. Cd 2÷ (and Zn z*) are pumped from Gram* bacteria by

a P-type ATPase with a phospho-aspartate intermediate. ~""'~

4. Cd z', Zn 2', Co 2~, and Ni 2~. c z c ( a n d c n r a n d n c c ) .

CdZ+, Zn2~,

Co 2~, and Ni 2~ are pumped from G r a m bacteria by t h r e e p o l y p e p t i d e membrane complex that is NOT an ATPase but functions as a divalent cation/2 H * antiporter. The complex consists of an inner membrane protein (CzcA), an outer membrane protein (CzcC) and a protein associated with both membranes (CzcB). 5. Cu 2~ . cop. Cu 2~ resistance results from a four polypeptide complex, consisting of an inner membrane protein, an outer membrane protein, and t w o periplasmic copper-binding proteins. In Pseudomonas, Cop results in periplasmic sequestration of Cu 2.. In addition, chromosomally-encoded P-type ATPases may provide resistance by effluxing Cu 2÷ or Cu ÷ . (~"

6. Cr042 . chr.

Chromate resistance results from a single membrane

polypeptide that causes reduced net cellular uptake, but efflux has not been demonstrated. 1~'"~

7. TeO3 z .

tel (or the). Tellurite resistance results from any of

several genetically-unrelated plasmid systems. Although reduction to metallic Te ° frequently occurs, this does not seem to be the primary resistance mechanism. Additional resistance systems await understanding for ions of bismuth (Bi), boron (B), lead (Pb), silver (Ag), thallium (TI) and tin (Sn).

Fig. 1. Overviewof plasmid heavy-metal-resistancemechanisms. metabolites and carbohydrates, whatever was frequently found in the environment provided selection for genes (thus enzymes) required for transport and metabolism. As with regard to antibiotics and their resistances, the recent activities of humans create local environments of high selection, but there is nothing new about toxic heavy metal resistance determinants. Since Anne Summers found that mercury-resistant bacteria could volatilize mercury (as monoatomic elemental Hg°; Summers and Silver, 1972), the subject of bacterial plasmid-determined resistances to toxic inorganic cations and anions has burgeoned. Our first paper

on mercury resistance was initially rejected for publication, largely on the basis that mercury 'gas' as a resistance mechanism was a ludicrous idea. It was also the case that two groups in Japan had independently discovered the mechanism of mercury resistance - and published before we did. For purposes of this summary, one can only briefly cover the range of mechanisms and some newer findings. Many metal-specific mini-reviews have appeared recently, including in a 1992 issue of the journal Plasmid, in a 1995 issue of Journal of Industrial Microbiology, and in separately-published reviews that are listed here ion by ion.

5: Silver/Gene 179 (1996) 9-19

In addition to mercury-resistance determinants, we know of highly specific resistance systems for arsenic (and antimony), copper, cadmium (and zinc in Gram + bacteria, but a different system shared with cobalt, nickel and zinc in plasmids of G r a m - bacteria), tellurite, silver and other toxic metal ions (Fig. 1). Three generalizations may be made: (i) The specificities of plasmid-determined metal resistances are similar to those for antibiotic resistances, or sugar or aa metabolism (i.e., very specific). There is no general mechanism for resistance to all heavy metal ions. (ii) Metal-ion resistance systems have been found on plasmids of every eubacterial group tested, from Escherichia coli to Streptomyces. Frequently, the genes (and mechanisms) initially found with plasmids (because of ease of experimental analysis) are subsequently found on the chromosomes of other bacteria. For example, the newly-released 1830-kb total genome sequence of Haemophilus influenzae (Fleischmann et al., 1995) includes predicted genes for arsenate reductase (arsC) and mercury transport (merT and merP) closely homologous to those we had earlier sequenced from bacterial plasmids. We do not know as yet whether these genes are functional, as is the chromosomal ars operon of E. coli (Carlin et al., 1995; Diorio et al., 1995). No one has reported mechanisms such as described here with Archaebacteria. Whether this reflects a difference in biology or a lack of effort is hard to tell. The Archaea grow in environments with high levels of toxic heavy metal ions and therefore are expected to have developed the same or alternative mechanisms of resistances. (iii) The mechanisms of resistance are generally efflux 'pumping' (removing toxic ions that entered the cell by means of transport systems evolved for nutrient cations or anions) and enzymatic detoxification (generally redox chemistry) converting a more toxic to a less toxic or less available metal-ion species. Why it would not be easier to keep toxic ions out (by altering the specificity of membrane uptake transport systems), rather than to expend metabolic energy bringing in toxic ions and then more energy pumping them out, requires an explanation. It seems that the metabolic penalty for having more specific uptake pumps is greater than the genetic cost of having plasmid genes in the population that can spread and become induced (almost all of these systems are transcriptionally regulated) when needed. Efflux pumps are the major currently-known group of plasmid resistance systems. They can be either ATPases (as the Cd 2 + ATPase of Gram + and the arsenite ATPase of G r a m - bacteria) or chemiosmotic (as the divalent cation efflux system of soil Alcaligenes and the arsenite efflux system of the chromosome of G r a m - bacteria and of plasmids in Gram + bacteria). The mechanisms are not precisely the same in all bacterial types: while the mercury-resistance systems are highly homologous in all bacteria studied, and the arsenic-resistance systems are

11

homologous (but differ in energy coupling), cadmium resistance involves unrelated ATPases in Gram + bacteria (including Staphylococcus, Listeria and Bacillus) and chemiosmotic antiporters in G r a m - bacteria. These systems appear to be of independent evolutionary origin. There is even a well-described bacterial metallothionein, found so far only on the chromosome of some cyanobacteria, and conferring resistances to C d 2+ and Zn 2+ (Turner and Robinson, 1995).

2. Plasmid mercury resistance

Homologous plasmid systems for resistance to inorganic mercury have been found on plasmids of Gram and Gram + bacteria. In the collection of N. Datta of some 800 antibiotic-resistance plasmids that had been mobilized from various G r a m - bacteria into E. coli, 25% carried mercury resistance. In most cases, the order and approximate number and functions of the genes are the same. Figure2 presents one example from a G r a m - and one from a Gram + microbe. Both mer systems start with a regulatory gene, merR, whose product is a unique positively-acting activator protein that twists and bends the operator DNA region, allowing RNA polymerase to synthesize mRNA (Misra, 1992: Summers, 1992; O'Halloran, 1993). In the mer systems of G r a m - bacteria, merR is transcribed separately and in the opposite direction from the remaining met genes, allowing tighter control of the mer operon than possible with Gram ÷ bacteria, with which merR is the first gene on the multigene mer operon (Fig. 2). There then follows one to three genes whose products are involved in transport of toxic Hg 2 ÷ across the cell membrane to the intracellular detoxifying enzyme mercuric reductase, the product of the long merA gene. In both cases shown in Fig. 2, merA is followed immediately by merB, which encodes the enzyme organomercurial lyase, which breaks the carbon-mercury bond in toxic substrates such as phenylmercury acetate. In one recently studied example from Pseudomonas, the merB gene is found between resistance genes merR and merT, together with an extra operator/promoter site. In most mer operons from Gram sources, the final gene merD encodes a secondary regulatory protein that binds to the same DNA site as MerR (Mukhopadhyay et al., 1991). In Bacillus mer operons, there is a 1.8-kb gap (of unknown function) between merA and merB. Recent developments on the bacterial mercury-resistance system have been covered elsewhere (Silver and Walderhaug, 1995; Silver and Ji, 1994; Ji and Silver, 1995; Silver and Phung, 1996). One version of mercuric reductase has had its crystal structure solved by X-ray diffraction (Schiering et al., 1991). As anticipated, the structure is very similar to that of glutathione reductase from mam-

12

S. Silver/Gene 179 (1996) 9-19

A. Plasmid pDU1358 (Gram) v

II IIII merR

II IIIII1~11

P/O merT merP

merA

merB

merD

1116aa 91aa

561aa

212aa

120 aa

144aa /

Cytoplasm =.ili:.i~ii: ',= iiii iiiiiii:,

~ iiiiiii

i lil il

!iiii i ::i:=ii i ~ : .

Periplasmic space

/

/

i:.ii=.,,=.=.,.i:.'=:i,i~

ii

i==iiiiii iii?,i ':i':i::il

~ ( ~

A. Plasmid pl258 (Gram+) v

P/O

merR

ORF3

ORF4

135 aa

161 aa

226 aa

ORF5 merT?. 128 aa

merA

merB

547 aa

216 aa

Fig. 2. Genes and enzymes of representative mercury-resistance systems in E. coli (plasmid pDU1358) and S. aureus (plasmid pi258) (from Ji and Silver, 1995, with permission). Dots (.) represent proposed Hg 2+ -binding Cys thiols.

malian sources. It is a homo-dimer, with each subunit containing a highly conserved active site with two cysteines involved. An FAD is bound to each subunit, and an NADPH-binding site is found for electron transfer from N A D P H to FAD to the substrate Hg 2+. The active site includes the redox-active disulfide region conserved also with glutathione reductase (on one subunit) and the substrate-binding site at the C-terminal (including conserved vicinal Cys r e s i d u e s ) o f the other subunit (Distefano et al., 1990).

st,q,h.,.l,,.,,.c . . . . . . . wet,.,

p1258 •

l(;ram+)plasmid

~a,

,,,.~ ~ a,-.,~ .~4,,,,- ~4Z9aa -.-.131 a a \ \ I "~ \\ 13~v~ L ~~ 5s~ \\\\18c~ \ "-. \'-. -..\ ~ b ~ \\ x,\ .., F..~,'herichiac,diq(;,'an,-)plasmidR773 \\ \ dI l l 7 a a ~I 120aa 429 aa ""..\141 m~" 538aa "" ~oa, ,,-,r ~ [ ,,~A apsB / / ~ . / .9. .~ I

~

I

Regulalion,

ATPase

Membrane Arsenate subunit / / reductase

subunit

/1 [

3. Arsenic resistance

]

/

I 74~'~ I L ll7aal

/ / /

/

436aa

/

/

//

91'%

/.

////

94%/ / /

//141aa /

-

Understanding of the bacterial arsenic resistance systems is less complete and more recent than that of mercury resistance. Fundamentally the same genes (and biochemical mechanism) are encoded on plasmids in E. coli and in S. aureus (Ji and Silver, 1992a, 1995; Fig. 3). However, the number of genes can vary (Fig. 3) and the details of their functions can differ. Five ars operons have been sequenced, of which three are shown in Fig. 3. A second Gram + system was studied (Rosenstein et al., 1992, 1994) from a plasmid of Staphylococcus carnosus, a microbe used in the meat processing industry. Its sequence was 96% identical overall to that shown in Fig. 3 for plasmid pi258. A second G r a m - plasmid system (from plasmid R46) has recently been sequenced

E~cherichia

coil

I(;ram-)

chromosome

Fig. 3. Genes and products for arsenic resistance in E. coli and S. aureus, Alignment and functions (below) of arsenic resistance genes (boxes) with aa sizes of predicted products (above and below genes) and percent identities between aa products (modified from Ji and Silver, 1995, with permission).

(Bruhn et al., 1996) and it has the same gene order and the protein products 85-92% identical to those from plasmid R773, the first arsenic system studied in detail. This is important since the R773 system as shown in Fig. 3 contains two extra genes, arsA and arsD, that are

13

s. Silver/Gene179 (1996) 9-19

missing from the recently identified E. coli chromosomal arsenic-resistance operon (Fig. 3; Carlin et al., 1995; Diorio et al., 1995) and the staphylococcal plasmids. The ArsD protein is a secondary regulator of ars operon transcription, so its presence or absence might have little effect on resistance. The ArsA protein, however, is a membrane-associated ATPase (Kaur and Rosen, 1993) attached to the ArsB inner-membrane protein (Wu et al., 1992) and energizing the arsenite efflux pump by ATP hydrolysis (Silver et al., 1993a; Silver and Ji, 1994; Ji and Silver, 1995; Silver and Phung, 1996). That an essential energy-coupling protein might be missing came as an enormous surprise. The resolution of this puzzlement is that the ArsB protein can function alone as a chemiosmotic (membrane-potential driven) arsenite-efflux transporter (Br6er et al., 1993; Dey and Rosen, 1995) or together with ArsA as an ATPdriven primary membrane pump. Such alternative energy coupling is novel as of now for known bacterial uptake or efflux transport systems. All others are either chemiosmotic or primary (frequently ATP-driven) transporters. The final gene product of the ars operon has also presented a major surprise. ArsC is an enzyme that reduces less toxic arsenate [-As(V)] to more toxic arsenite [As(III)] (Ji and Silver, 1992b), the substrate for the ArsB transport protein. It seems bizarre to convert a less toxic compound to a more toxic form, but ArsC activity is closely coupled with efflux from the cells (Ji and Silver, 1992b) so that intracellular arsenite never accumulates. Arsenate reductases from plasmids pi258 (Ji and Silver, 1992b; Ji et al., 1994) and R773 (Gladysheva et al., 1994) have been purified and studied. Although both enzymes reduce arsenate and both confer arsenate resistance, their Km's differ by 1000-times and their energy coupling is different. Arsenate reductase of pi258 derives reducing power from thioredoxin (glutaredoxin will not work), presumedly by recycling oxidized cystine to two reduced cysteine residues (of the four found in the sequence; Ji et al., 1994). In contrast arsenate reductase of plasmid R773 uses glutaredoxin (but not thioredoxin) and probably only one of the two cysteines found in the protein sequence is required. The arsenic resistance results are new and novel so we expect more surprises and changes of models. However, it seems clear that basically the same ars operon conferring resistances to As(III), As(V) and Sb(III) occurs widely in G r a m - and Gram + bacteria (Silver et al., 1993a; Cervantes et al., 1994; Diorio et al., 1995).

4. Plasmid-encoded copper resistance in Gram- bacteria

Strong copper resistance has been described with plasmids in Pseudomonas (Cooksey, 1994), Xanthomonas (Lee et al., 1994) and E. coli (Brown et al., 1994, 1995).

The systems are highly homologous (Cooksey, 1993, 1994; Brown et al., 1994, 1995) and contain the same genes. For Pseudomonas, the two regulatory genes are called copR and copS and the four structural genes copABCD (Fig. 4). The comparable E. coli genes are called pcoRS and pcoABCD (Brown et al., 1994, 1995). The PcoR and PcoS proteins are the only current example among the metal-resistance systems of transcriptional regulation by a classical 'two component' regulatory system. The sensor protein PcoS is found in the membrane and probably can be labeled by an autokinase activity at a specific conserved histidine residue with 32p from [7-3zp]ATP. The DNA-binding responder protein PcoR is probably trans-phosphorylated on a specific aspartate residue by 32p-labeled PcoS (Brown et al., 1994; Cooksey, 1994). The four structural proteins determining copper resistance are the inner membrane protein CopD, the outer membrane protein CopB, and two periplasmic proteins CopA and CopC (Fig. 4; Cooksey, 1993, 1994). CopA and CopC are blue copper proteins containing 11 and 1 Cu 2+, respectively (Cha and Cooksey, 1991; Cooksey, 1994). It is thought that storage of excess copper in the periplasmic space protects the cell from toxic copper. How CopD and CopB are involved in movement of copper across the membranes is not understood. However, a mutant cop operon containing copD but lacking one or more of the other genes confers hyper-sensitivity and hyper-accumulation of cellular copper (Cha and Cooksey, 1993), indicating a role for CopD in copper uptake by the cell. One problem here is that colonies of the copper-resistant

cdtEA\cul+

\ Eyo,0s /

Fig. 4. Copper transport and resistance. HypothesizedCutU (uptake) and Cute (efflux)P-type ATPases modeled after work of Solioz and Odermatt (1995) on Enterococcus hirae and more limited work on other microbes. CutB, proposed E. coli intracellular copper-binding protein (Brownet al., 1994), and CopABCD,the products of the plasmid-encoded copper-resistance system of P. syringae (Cooksey, 1994) (modifiedfrom Silver and Phung, 1996, with permission).

14

S. Silver/Gene179 (1996) 9-19

Pseudomonas turn bright blue when grown in high copper-containing media, while those of Xanthomonas and E. coli turn brown, and show no sign of periplasmic copper storage (Brown et al., 1994). Furthermore, there is preliminary evidence for copper efflux (not uptake) associated with the E. coli copper resistance system (Brown et al., 1994, 1995). Resolution of these problems awaits further work. There are also E. coli (Brown et al., 1994) and Pseudomonas syringae (Cooksey, 1994) chromosomal genes called cut that affect copper transport and resistance. The tentative models for these gene functions have changed from year to year (e.g., Brown et al., 1994, and earlier papers quoted there). Figure 4 shows three hypothesized chromosomal gene products, called CutU (for uptake) and CutE (for efflux) for proposed membrane ATPase pumps and CutB (for intracellular binding). The protein names are different from those of Brown et al. (1994), since newer data challenges earlier assignments of functions (see for example, Gupta et al., 1995). However, all models require the control of cellular uptake and efflux of copper, as well as specific intracellular copper-binding proteins to protect the cytoplasm from copper-related redox damage.

and CopZ-Cu + binds CopY-Cu +, forming an inactive (that is non-DNA-binding) complex. This tentative model explains the simultaneous induction of synthesis of both ATPases by 20~tM Ag + or 2raM Cu 2÷ (Odermatt et al., 1994). Other than the CadA cadmium efflux ATPase (below), the CopB copper efflux ATPase of E. hirae is the only bacterial cation efflux ATPase for which subcellular transport data are available. Solioz and Odermatt (1995) isolated inside-out, subcellular, membrane vesicles from E. hirae cells and the vesicles required ATP in order to accumulate 64Cu+ and 11°mAg+.The in vitro substrate for CopB is thought to be Cu + (Fig. 4) rather than Cu 2÷. Whether copper is taken up initially as Cu 2+ (Fig. 4) and subsequently reduced to Cu ÷ or whether copper is reduced at the cell surface (before or concomitant with transport) is not established. However, this is similar to the situation with eukaryotic cells where monovalent Cu ÷ is thought to be the intracellular form (Solioz and Odermatt, 1995).

5. Chromosomal copper resistance in Enterococcus hirae

Bacterial metallothioneins (MT), functionally homologous to the small (approximately 60-aa long), thiolrich (perhaps 20 of those 60 aa are Cys) metal-binding proteins of animal systems, have not been widely found. MT of apparently separate evolutionary origin from those of animal cells have been well established for only one prokaryotic group, cyanobacterial strains of the genus Synechococcus (Gupta et al., 1992, 1993; Turner and Robinson, 1995). The 58-aa polypeptide product of the smtA gene contains only nine Cys residues, which are clustered in groups of 4 and 5, as are the Cys in animal MT. Cys are clustered in two MT domains that bind divalent cations independently. The synthesis of MT is regulated at three levels (Turner and Robinson, 1995). Firstly, the repressor protein SmtB is inactivated by binding of divalent cations. Secondly, there is gene amplification so that tandem multiple copies of the MT locus are produced in metal-stressed cells (Gupta et al., 1992); and thirdly, a specific 'programmed' deletion removes most of the smtB gene for the repressor protein (Gupta et al., 1993).

The best understood copper transport and resistance system today is that of the Gram + pathogen Enterococcus hirae (previously called Streptococcus faecalls) (Solioz et al., 1994). Remarkably, the two genes, copA and copB, that determine, respectively, uptake and efflux P-type ATPases, are found in a single operon (Odermatt et al., 1993). The system is regulated in response to both copper-starvation (when the CopA uptake ATPase is needed) and copper-excess (when the CopB efflux ATPase is needed) (Odermatt and Solioz, 1995). Enterococcus CopA and CopB may correspond to CutU and CutE in Fig. 4, but the names are different in Fig. 4, since cop names has been used for the Pseudomonas plasmid genes for copper resistance (Cooksey, 1994). E. hirae mutants lacking the CopA uptake ATPase become somewhat copper-resistant and require higher levels of medium copper for growth. Bacterial mutants lacking the CopB etttux ATPase become copper-hypersensitive (Solioz et al., 1994). Regulation of the cop operon is governed by the products of the first two genes in the operon, copY (which determines a 145-aa repressor protein) and copZ (which determines a smaller, 69-aa, 'antirepressor'). The CopY apo-repressor is thought to be inactive and fails to bind to the operator/promoter DNA in the absence of intracellular Cu +. A moderate level of intracellular Cu ÷ binds to CopY converting it to a DNA-binding repressor (Odermatt and Solioz, 1995). At higher intracellular Cu ÷ levels, the CopZ antirepressor binds Cu ÷,

6. Bacterial metallothionein in cyanobacteria: a new system

7. Cadmium resistance in Gram + bacteria results from a P-type ATPase

The 727-aa Cd 2+ efflux ATPase from staphylococcal plasmid pi258 (Fig. 5) was the first of a system now found widely in Gram + bacteria (Silver et al., 1993b; Silver and Phung, 1996). The Cd 2+ ATPase has been found as well in soil Bacilli (Ivey et al., 1992) and clinical

S. Silver/Gene 179 (1996) 9-19

15

Phosphatase

Cd 2+binding

Aspartyl kinase

d 6

i

'

1 2

l l r l r l

3

4

5

6

Fig. 5. The Cd 2+ -resistance ATPase of S. aureus. The predicted motifs (Cd 2+-binding, phosphatase, membrane channel, and aspartyl kinase) regions are shown with key deduced aa. The + and - signs indicate the predicted locations of charged aa, with most positively-charged (Lys and Arg) aa on the cytoplasmic surface and most negatively-charged (Glu and Asp) aa on the extracellular surface (modified from Silver and Walderhaug, 1992, with permission).

Listeria (Lebrun et al., 1994). The protein structure as

diagramed is typical of P-type ATPases: it starts with a metal-binding motif, including a vicinal Cys pair. This motif is similar to other cadmium-, copper- and mercurybinding regions on efflux ATPases and other proteins (Silver et al., 1993b; Silver and Walderhaug, 1995; Silver and Phung, 1996). There follows a membrane ATPase region closely homologous to other P-type ATPases of bacteria, animals and plants. This includes the six predicted membrane-spanning regions shown in Fig. 5, the fourth of which is thought to be involved in the cation translocation pathway. It includes a conserved proline residue (as shown) between Cys that are found in the Cd 2+ ATPases and related proteins. Two intracellular domains shown in the model and common to P-type ATPases are the 'aspartyl kinase domain' and the 'phosphatase domain'. During the transport cycle, the protein is phosphorylated by ATP at the invariant Asp 415 residue shown. The name for this class of ATPases, 'P-type', is used since they are the only transport ATPases that have a covalent phospho-protein intermediate. The CadA cadmium ATPase is one of the few bacterial examples for which direct experimental evidence for phosphorylation is available (Tsai and Linet, 1993). Tsai et al. (1992) showed directly ATP-dependent transport of cadmium by vesicles containing the CadA ATPase. We recently identified directly in vitro CadC, the product of the second gene of this operon, to be a DNAbinding transcriptional regulatory protein (Endo and Silver, 1995). CadC is a member of a new family (Bairoch, 1993) of metal-binding repressor proteins, with the arsenic system repressor ArsR (Wu and Rosen, 1991;

Ji and Silver, 1992a; Rosenstein et al., 1994) and the cyanobacterial metallothionein repressor SmtB (Turner and Robinson, 1995) as the other members.

8. Cadmium resistance in Gram- soil bacteria results from a multi-protein chemi-osmotic antiporter

The large plasmids of the soil chemi-lithotrophic autotroph Alcaligenes eutrophus have numerous toxic metal-resistance determinants (including in strain CH34, three for mercury, one for chromate resistance, and two for divalent cations called czc (for Cd 2+, Zn 2+ and Co 2+ resistances) and cnr (for Co 2+ and Ni 2+ resistances)). The DNA sequences of czc (Nies et al., 1989) and cnr (Liesegang et al., 1993) have diverged sufficiently that Southern blot hybridization fails. Nevertheless, the predicted amino acid sequences (Nies, 1992a; Nies and Silver, 1995; Diels et al., 1995) show closely related systems with basically the same three structural proteins (Fig. 6). Indeed, mutations of the Cnr system give additional Zn 2+ resistance, again showing that the two systems are fundamentally the same. Czc is an efflux pump (Nies et al., 1989) that functions as a chemiosmotic divalent cation/proton antiporter (Nies, 1995; Nies and Silver, 1995). The proteins involved have become the paradigm for a new family of three-component chemiosmotic exporters, including members that efflux toxic cations and also organic compounds (Diels et al., 1995). CzcA functions as the basic inner membrane-transport protein (Fig. 6). There is no evidence to distinguish

S. Silver/Gene 179 (1996) 9-19

16

~ C ~ 2"1" ~djO 2"1"

Z n 2, IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIUF-

ILIlilTIITli I°i l]n3Il Ul l l IH Illllllllllllllltli|

.

.

.

.

.

.

z°

Fig. 6. Czc model for Cd 2÷, Zn 2+ and Co 2÷ elttux system functioning as proton/cation antiporter consisting of inner membrane (CzcA), outer membrane (CzcC) and 'membrane fusion' (CzcB) proteins functioning as a dimer (adapted from model of Q. Dong and S.S. (modified from Diels et al. (1995), with permission)).

whether Czc functions as a dimer complex as shown in Fig. 6 or as a complex consisting of one CzcA, one CzcB and one CzcC polypeptide, as shown in the comparable figure in Diels et al., 1995. CzcC is thought to function as an outer membrane protein as shown in Fig. 6 (Diels et al., 1995), although there remains some uncertainty about its location (Nies and Silver, 1995); and CzcB appears to function as a 'membrane-fusion protein' that bridges the inner and outer cell membrane of G r a m bacteria (references in Ji and Silver (1995) and Diels et al. (1995)). The transcriptional regulation of the czc and cnr are currently under active study. Two or more regulatory genes are involved but there is no agreement about their functions (Diels et al., 1995; Nies, 1992b; Nies and Silver, 1995).

1990) chromate resistance systems share homologous chrA genes, which encode membrane proteins. A third chrA was recently found on a plasmid of a cyanobacterium (Nicholson and Laudenbach, 1995). Therefore, we expect more examples of chromate-resistance operons will be found. The A. eutrophus chromateresistance determinant contains an additional upstream gene, chrB, that was proposed to be responsible for the inducibility of the resistance (Nies et al., 1990). However, the P. aeruginosa determinant as cloned lacks both the chrB gene and inducibility.

10. Tellurite resistance in Gram- bacteria 9. Chromate resistance and chromate reduction in Gram- bacteria Both occur, but resistance to chromate governed by bacterial plasmids appears to have nothing to do with chromate reduction. Furthermore, it is not clear whether chromate reduction ability that has been found with several bacterial isolates confers resistance to CrO 2(Ohtake and Silver, 1994). Plasmid-determined chromate resistance results from reduced uptake of CrO 2- by the resistant cells (Ohtake et al., 1987), but we have been unable to determine whether there is chromate efflux (as in other resistance mechanisms) or a direct block on uptake. The DNA sequences of the P. aeruginosa (Cervantes et al., 1990) and A. eutrophus (Nies et al.,

There are well-studied and indeed several sequenced determinants of plasmid-governed tellurite resistance, but in each case we do not understand the mechanism of tellurite resistance (Lloyd-Jones et al., 1991; Walter and Taylor, 1989, 1992; Ji and Silver, 1995). Tellurite resistance does not appear to involve reduction to black metallic tellurium - which indeed occurs, especially if resistance allows cell growth (Walter and Taylor, 1992). The two surprising conclusions from cloning and sequencing of tellurite-resistance operons are firstly that each system seems different: related sequences have not been found; and secondly that the predicted aa sequences of the protein products have not led to a mechanistic model for resistance. Thus, we are waiting for clarification of tellurite-resistance mechanisms.

S. Silver/Gene 179 (1996) 9-19

150

100 (D

50

b, .1

I

I

I

1

10

100

1000

A g N O 3 ( tM) Fig. 7. Plasmid-determined resistance to silver (Ag +) in E. coll. Growth overnight at 37°C of sensitive E. coli cells without a plasmid (O), with plasmid pMG 101 (A), or AgCI precipitate without cells (D) (K. M atsui and S.S., unpublished).

11. Other toxic metal resistances I n a d d i t i o n to the specific resistances d i s c u s s e d a b o v e , several a d d i t i o n a l resistances are listed in the S u m m a r y , the I n t r o d u c t i o n a n d Fig. 1. T h e s e h a v e b e e n s t u d i e d still less a n d are m a t e r i a l for f u t u r e research. T h i s year we h a v e s t a r t e d a g a i n w o r k i n g o n p l a s m i d - d e t e r m i n e d silver r e s i s t a n c e (Fig. 7), so we h o p e t h a t the n e x t B a n g k o k C o n f e r e n c e will p r e s e n t a n o p p o r t u n i t y to d e s c r i b e the m o l e c u l a r g e n e t i c a n d b i o c h e m i c a l basis for h i g h l y specific A G + r e s i s t a n c e (Haefeli et al., 1984). P r e l i m i n a r y w o r k i n d i c a t e s t h a t n e i t h e r r e d u c t i o n of A g + to A g o (which o c c u r s with b o t h r e s i s t a n t a n d sensitive cells) n o r efflux of the toxic i o n is the basis for this resistance. A n e w p l a s m i d resistance to t r i b u t y l t i n (used as a n t i - f o u l i n g c o m p o u n d s for ship hulls) h a s r e c e n t l y b e e n r e p o r t e d ( M i l l e r et al., 1995).

Acknowledgement T h e r e s e a r c h in o u r l a b o r a t o r y o n these topics has i n c l u d e d p a r t i c i p a t i o n of c o l l e a g u e s a n d friends (freq u e n t l y listed as c o - a u t h o r s below).

References Bairoch, A. (1993) A possible mechanism for metal-ion induced DNAprotein dissociation in a family of prokaryotic transcriptional regulators. Nucleic Acids Res. 21, 2515.

17

BrOer, S., Ji, G., Br6er, A. and Silver, S. (1993) Arsenic efflux governed by the arsenic resistance determinant of Staphylococcus aureus plasmid PI258. J. Bacteriol. 175, 3480-3485. Brown, N.L., Lee, B.T.O. and Silver, S. (1994) Bacterial transport of and resistance to copper. In: Sigel, H. and Sigel, A. (Eds.), Metal Ions in Biological Systems, Vol. 30. Marcel Dekker, New York, pp. 405-434. Brown, N.L., Barrett, S.R., Camakaris, J., Lee, B.T.O. and Rouch, D.A. (1995) Molecular genetics and transport analysis of the copper-resistance determinant (pco) from Escherichia coli plasmid pRJ1004. Mol. Microbiol. 17, 1153 1166. Bruhn, D.F., Li, J., Silver, S., Roberto, F. and Rosen, B.P. (1996) Arsenic resistance operon of IncN plasmid R46. FEMS Microbiol. Lett. 139, 149-153. Carlin, A., Shi, W., Dey, S. and Rosen, B.P. (1995) The ars operon of Escherichia coli confers arsenical and antimonial resistance. J. Bacteriol. 177, 981 986. Cervantes, C., Ohtake, H., Chu, L., Misra, T.K. and Silver, S. (1990) Cloning, nucleotide sequence and expression of the chromate resistance determinant of Pseudornonas aeruginosa plasmid pUM505. J. Bacteriol. 172, 287-291. Cervantes, C., Ji, G., Ramirez, J.L. and Silver, S. (1994) Resistance to arsenic compounds in microorganisms. FEMS Microbiol. Rev. 15, 355-367. Cha, J.-S. and Cooksey, D.A. (1991) Copper resistance in Pseudomonas syringae mediated by periplasmic and outer membrane proteins. Proc. Natl. Acad. Sci. USA 88, 8915-8919. Cha, J.-S. and Cooksey, D.A. (1993) Copper hypersensitivity and uptake in Pseudomonas syringae containing cloned components of the copper resistance operon. Appl. Environ. Microbiol. 59, 1671 1674. Cooksey, D.A. (1993) Copper uptake and resistance in bacteria. Mol. Microbiol. 7, 1-5. Cooksey, D.A. (1994) Molecular mechanisms of copper resistance and accumulation in bacteria. FEMS Microbiol. Rev. 14, 381-386. Dey, S and Rosen, B.P. (1995) Dual mode of energy coupling by the oxyanion-translocating ArsB protein. J. Bacteriol. 177, 385-389. Diels, L., Dong, Q., van der Lelie, D., Baeyens, W. and Mergeay, M. (1995) The czc operon of Alcaligenes eutrophus CH34: from resistance mechanism to the removal of heavy metals. J. Ind, Microbiol. 14, 142-153. Diorio, C., Cai, J., Marmor, J., Shinder, R. and DuBow, M.S. (1995) An Escherichia coli chromosomal ars operon homolog is functional in arsenic detoxification and is conserved in Gram-negative bacteria. J. Bacteriol. 177, 2050-2056. Distefano, M.D., Moore, M.J. and Walsh, C.T. (1990) Active site of mercuric reductase and resides at the subunit interface and requires Cys 135 and Cys 14° from one subunit and Cysss8 and Cys 5s9 from the adjacent subunit: evidence from in vivo and in vitro heterodimer formation. Biochemistry 29, 2703-2713. Endo, G. and Silver, S.L (1995) CadC, the transcriptional regulatory protein of the cadmium resistance system of Staphylococcus aureus plasmid pi258. J. Bacteriol. 177, 4437-4441. Fleischmann, R.D., Adams, M.D. and 36 others, Smith, H.O. and Venter, J.C. (1995) Whole-genome random sequencing and assembly of Haemophilus influenzae Rd. Science 269, 496-512. Gladysheva, T.B., Oden, K.L. and Rosen, B.P. (1994) Properties of the arsenate reductase of plasmid R773. Biochemistry 33, 7288-7293. Gupta, A., Whinon, B.A., Morby, A.P., Huckle, J.W. and Robinson, N.J. (1992) Amplification and rearrangement of a prokaryotic metallothionein locus smt in Synechococcus PCC 6301 selected for tolerance to cadmium. Proc. R. Soc. London B 248, 273-281. Gupta, A., Morby, A.P., Turner, J.S., Whitton, B.A. and Robinson, N.J. (1993) Deletion within the metallothionein locus of cadmium-tolerant Synechococcus PCC 6301 involving a highly iterated palindrome (HIP1). Mol. Microbiol. 7, 189 195.

18

S. Silver/Gene 179 (1996) 9 19

Gupta, S.D., Lee, B.T.O., Camakaris, J. and Wu, H.C. (1995) Identification of cutC and cutF (nlpF) genes involved in copper tolerance in Escherichia coli. J. Bacteriol. 177, 4207-4215. Haefeli, C., Franklin, C. a n d Hardy, K. (1984) Plasmic-determined silver resistance in Pseudomonas stutzeri isolated from a silver mine. J. Bacteriol. 158, 389 392. Ivey, D.M., Guffanti, A.A., Shen, Z., Kudyan, N. and Krulwich, T.A. (1992) The CadC gene product of alkaliphilic Bacillus firmus OF4 partially restores Na ÷ resistance to an Escherichia coli strain lacking an Na+/H ÷ antiporter (NhaA). J. Bacteriol. 174, 4878-4884. Ji, G. and Silver, S. (1992a) Regulation and expression of the arsenic resistance operon from Staphylococcus aureus plasmid pi258. J. Bacteriol. 174, 3684-3694. Ji, G. and Silver, S. (1992b) Reduction of arsenate to arsenite by the ArsC protein of the arsenic resistance operon of Staphylococcus aureus plasmid pi258. Proc. Natl. Acad. Sci. USA 89, 7974-7978. Ji, G. and Silver, S. (1995) Plasmid resistance mechanisms for heavy metals of environmental concern. J. Ind. Microbiol. 14, 61-75. Ji, G., Garber, E.A., Armes, L.G., Chen, C.-M., Fuchs, J.A. and Silver, S. (1994) Arsenate reductase of Staphylococcus aureus plasmid pi258. Biochemistry 33, 7294-7299. Kaur, P. and Rosen, B.P. (1993) complementation between nucleotide binding domains in an anion-translocating ATPase. J. Bacteriol. 175, 351 357. Lebrun M., Audurier, A. and Cossart, P. (1994) Plasmid-borne cadmium resistance genes in Listeria monocytogenes are similar to cadA and cadC of Staphylococcus aureus and are induced by cadmium. J. Bacteriol. 176, 3040 3048. Lee, Y.-A., Hendson, M., Panopoulus, N.J. and Schroth, M.N. (1994) Molecular cloning, chromosomal mapping, and sequence analysis of copper resistance genes from Xanthomonas campestris pv. juglandis: homology with blue copper proteins and multicopper oxidase. J. Bacteriol. 176, 173-188. Liesegang, H., Lemke, K., Siddiqui, R.A. and Schlegel, H.-G. (1993) Characterization of the inducible nickel and cobalt resistance determinant cnr from pMOL28 of Alcaligenes eutrophus CH34. J. Bacteriol. 175, 767-778. Lloyd-Jones, R., Ritchie, D.A. and Strike, P. (1991) Biochemical and biophysical analysis of plasmid pMJ600-encoded tellurite Miller, C.E., Wuertz, S., Cooney, J.J. and Pfister, R.M. (1995) Plasmids in tributyltin-resistant bacteria from fresh and estuarine waters. J. Ind. Microbiol. 14, 337 342. Misra, T.K. (1992) Bacterial resistances to inorganic mercury salts and organomercurials. Plasmid 27, 4 16. Mukhopadhyay, D., Yu, H., Nucifora, G. and Misra, T.K. (1991) Purification and functional characterization of MerD. A coregulator of the mercury resistance operon in Gram negative bacteria. J. Biol. Chem. 266, 18538-18542. Nicholson, M.L. and Laudenbach, D.E. (1995) Genes encoded on a cyanobacterial plasmid are transcriptionally regulated by sulfur availability and CysR. J. Bacteriol. 177, 2143-2150. Nies, D.H. (1992a) Resistance to cadmium, cobalt, zinc and nickel in microbes. Plasmid 27, 17-28. Nies, D.H. (1992b) CzcR and CzcD, gene products affecting regulation of cobalt, zinc, and cadmium resistance (czc) in Alcaligenes eutrophus. J. Bacteriol. 174, 8102-8110. Nies, D.H.: The cobalt, zinc, and cadmium efflux system CzcABC from Alcaligenes eutrophus functions as a cation-proton antiporter in Escherichia coli. J. Bacteriol. 177 (1995) 2707 2712. Nies, D.H. and Silver, S. (1995) Ion efflux systems involved in bacterial metal resistances. J. Ind. Microbiol. 14, 186 199. Nies, D.H., Nies, A., Chu, L. and Silver, S. (1989) Expression and nucleotide sequence of a plasmid-determined divalent cation efflux system from Alcaligenes eutrophus. Proc. Natl. Acad. Sci. USA 86, 7351-7355.

Nies, A., Nies, D.H. and Silver, S. (1990) Nucleotide sequence and expression of a plasmid-encoded chromate resistance determinant from Alcaligenes eutrophus. J. Biol. Chem. 265, 5648-5653. Odermatt, A. and Solioz, M. (1995) Two trans-acting metallorgulatory proteins controlling expression of the copper-ATPases of Enterococcus hirae. J. Biol. Chem. 270, 434 4354. Odermatt, A., Suter, H., Krapf, R. and Solioz, M. (1993) Primary structure of two P-type ATPases involved in copper homeostasis in Enterococcus hirae. J. Biol. Chem. 268, 12775 12779. Odermatt, A., Suter, H., Krapf, R. and Solioz, M. (1994) Induction of the putative copper ATPases, CopA and CopB, of Enterococcus hirae by AG ÷ and Cu 2+ , and Ag ÷ extrusion by CopB. Biochem. Biophys. Res. Commun. 202, 44-48. O'Halloran, T.V. (1993) Transition metals in control of gene expression. Science 261,715 725. Ohtake, H. and Silver, S.: Bacterial reduction of toxic chromate. In: Chaudhry, G.R. (Ed.), Biological Degradation and Bioremediation of Toxic Chemicals. Chapman and Hall, London, 1994, pp. 403-415. Ohtake, H., Cervantes, C. and Silver, S. (1987) Decreased chromate uptake in Pseudomonas fluorescens carrying a chromate resistance plasmid. J. Bacteriol. 169, 3853-3856. Rosenstein, R., Peschel, A., Wieland, B. and G6tz, F. (1992) Expression and regulation of the antimonite, arsenite, and arsenate resistance operon of Staphylococcus xylosus plasmid pSX267. J. Bacteriol. 174, 3676-3683. Rosenstein, R., Nikoleit, K. and G6tz, F. (1994) Binding of ArsR, the repressor of the Staphylococcus xylosus (pSX267) arsenic resistance operon to a sequence with dyad symmetry within the ars promoter. Mol. Gen. Genet. 242, 566-572. Schiering, N., Kabsch, W., Moore, M.J., Distefano, M.D., Walsh, C.T. and Pal, E.F. (1991) Structure of the detoxification catalyst mercuric ion reductase from Bacillus sp. strain RC607. Nature (London) 352, 168-171. Silver, S. and Ji, G. (1994) Newer systems for bacterial resistances to toxic heavy metals. Environ. Health Perspect. 102 (Suppl. 3), 107 113. Silver, S. and Phung, L.T. (1996) Bacterial plasmid-mediated heavy metal resistances: new surprises. Annu. Rev. Microbiol. 50, 753-789. Silver, S. and Walderhaug, M. (1992) Gene regulation of plasmid- and chromosomal-determined inorganic ion transport in bacteria. Microbiol. Rev. 56, 195-228. Silver, S. and Walderhaug, M. (1995) Bacterial plasmid-mediated resistances to mercury, cadmium and copper. In: Goyer, R.A. and Cherian, M.G. (Eds.), Toxicology of Metals. Biochemical Aspects. SpringerVerlag, Berlin, pp. 435-458. Silver, S., Ji, G., Br6er, S., Dey, S., Dou, D. and Rosen, B.P. (1993a) Orphan enzyme or patriarch of a new tribe: the arsenic resistance ATPase of bacterial plasmids. Mol. Microbiol. 8, 637-642. Silver, S., Nucifora, G. and Phung, L.T. (1993b) Human Menkes X chromosome disease and the staphylococcal cadmium resistance ATPase: a remarkable similarity in protein sequences. Mol. Microbiol. 10, 7-12. Solioz, M. and Odermatt, A. (1995) Copper and silver transport by CopB-ATPase in membrane vesicles of Enterococcus hirae. J. Biol. Chem. 270, 9217-9221. Solioz, M., Odermatt, A. and Krapf, R. (1994) Copper pumping ATPases: common concepts in bacteria and man. FEBS Lett. 346, 44 47. Summers, A.O. (1992) Untwist and shout: a heavy metal-responsive transcriptional regulator. J. Bacteriol. 174, 3097-3101. Summers, A.O. and Silver, S. (1972) Mercury resistance in a plasmidbearing strain of Escherichia coli. J. Bacteriol. 112, 1228-1236. Tsai, K.-J. and Linet, A.L. (1993) Formation of a phosphorylated enzyme intermediate by the cadA Cd2+-ATPase. Arch. Biochem. Biophys. 305, 267-270. Tsai, K.-J., Yoon, K.P. and Lynn, A.R. (1992) ATP-dependent cadmium

S. Silver/Gene 179 (1996) 9-19 transport by the cadA cadmium resistance determinant in everted membrane vesicles of Bacillus subtilis. J. Bacteriol. 174, 116-121. Turner, J.S. and Robinson, N.J. (1995) Cyanobacterial metallothioneins: biochemistry and molecular genetics. J. Ind. Microbiol. 14, 119-125. Walter, E.G. and Taylor, D.E. (1989) Comparison of tellurite resistance determinants from the IncP~ plasmid RP4Te r and the IncHII plasmid pHH1508a. J. Bacteriol. 171, 2160-2165.

19

Walter, E.G. and Taylor, D.E. (1992) Plasmid-mediated resistance to tellurite: expressed and cryptic. Plasmid 27, 52-64. Wu, J. and Rosen, B.P. (1991) The ArsR protein is a trans-acting regulatory protein. Mol. Microhiol. 5, 1331-1336. Wu, J, Tisa, L.S. and Rosen, B.P. (1992) Membrane topology of the ArsB protein, the membrane subunit of an anion-translocating ATPase. J. Biol. Chem. 267, 12570-12576.