27, 17-28 ( 1992)


to Cadmium, Cobalt, Zinc, and Nickel in Microbes

DIETFUCH H: NIES Institut fur Pjlanzenphysiologie und Mikrobiologie, Freie UniversitUt Berlin, Kiinigin-Luise-Strasse 12-16, 1000 Berlin 33, Germany The divalent cations of cobalt, zinc, and nickel are essential nutrients for bacteria, required as trace elements at nanomolar concentrations. However, at micro- or millimolar concentrations, Coz+, Zn*+, and Ni*+ (and “bad ions” without nutritional roles such as Cd”) are toxic. These cations are transported into the cell by constitutively expressed divalent cation uptake systems of broad specificity, i.e., basically Mg*+ transport systems. Therefore, in case of a heavy metal stress,uptake of the toxic ions cannot be reduced by a simple down-regulation of the transport activity. As a response to the resulting metal toxicity, metal resistance determinants evolved which are mostly plasmid-encoded in bacteria. In contrast to that of the cation Hg2’, chemical reduction of Co’+$ Zn’+, Ni2+, and Cd” by the cell is not possible or sensible. Therefore, other than mutations limiting the ion range of the uptake system, only two basic mechanisms of resistance to these ions are possible (and were developed by evolution): intracellular complexation of the toxic metal ion is mainly used in eucaryotes; the cadmium-binding components are phytochelatins in plant and yeast cells and metallothioneins in animals, plants, and yeasts. In contrast, reduced accumulation based on an active efflux of the cation is the primary mechanism developed in procaryotes and perhaps in Saccharomyces cerevisiae. All bacterial cation efflux systemscharacterized to date are plasmid-encoded and inducible but differ in energy-coupling and in the number and types of proteins involved in metal transport and in regulation. In the gram-positive multiple-metal-resistant bacterium Staphylococcus aureus, Cd** (and probably Zt?) efhux is catalyzed by the membrane-bound CadA protein, a P-type ATPase. However, a second protein (CadC) is required for full resistanceand a third one (CadR) is hypothesized for regulation of the resistancedeterminant. The czc determinant from the gram-negative multiplemetal-resistant bacterium Alcaligenes eutrophus encodes proteins required for Co’+, Zn’+, and Cd’+ efflux (CzcA, CzcB, and CzcC) and regulation of the czc determinant (CzcD). In the current working model CzcA works as a cation-proton antiporter, CzcB as a cation-binding subunit, and CzcC as a modifier protein required to change the substrate specificity of the system from Zn*+ only to CO*‘, Zn*+, and Cd’+. o 1992 Academic press, inc.

The four heavy metals (metals with a density 25 g/cm3) zinc, cobalt, cadmium, and nickel are used in industrial countries for a variety of applications. Metallic zinc is used for alloys like brass,to galvanize iron, and for fabrication of batteries. It is obtained by roasting zinc blende (ZnS), by heating smithsonite (ZnCO,), or by electrolysis methods. Industrial plants producing metallic zinc by roasting or heating are always potential sources of environmental zinc contamination becausethe melting (4 19’C) and boiling (907°C) points of metallic zinc are low compared to those of other metals (example given in Diels and Mergeay, 1990). Cadmium is in many respects chemically similar to zinc. It occurs most often in small quantities associated with zinc ores; the cadmium content is frequently between 0.1 and

5%. The melting (321 “C) and boiling (766°C) points of cadmium are even lower than the respective values for zinc; plants producing zinc by roasting are therefore also potential sourcesfor environmental contamination with cadmium. Cadmium and solutions of its compounds are toxic. Cadmium is used for electroplating, for batteries, and for TV tubes. It is usually obtained as a by-product of zinc production. Cobalt and nickel are chemically related to iron and are both used for the production of steel and for electroplating. Cobalt and nickel salts have been used for centuries for the production of blue and green pigments, respectively. These examples illustrate the extensive use of cadmium, zinc, cobalt, and nickel by humans which consequently leads to contami17

0147-619X/92 $3.00 Copyright Q I992 by Academic Press, Inc. All rights of reproduction in any form reserved.



nation of soil and freshwater habitats with the divalent cations of the four heavy metals. In addition to the anthropogenic contamination, heavy metal ions may leak from naturally occurring minerals into soil or freshwater habitats. In both casesit is not generally one cation that is present in toxic concentrations, but usually a major cation plus some accompanying ions, e.g., Zn*+ plus Cd*+ and Ni*+ plus Co*+ and CrO$-. As a response to this challenge, multiple-metal ion-resistant bacteria evolved which contain a variety of plasmid-encoded metal resistance determinants, e.g., Staphylococcusaureus (Novick and Roth, 1968) and Alcaligenes eutrophus strain CH34 (Mergeay et al., 1985). In addition to the czc and the cnr determinant discussed in detail (see below), A. eutrophus strain CH34 harbors three rner determinants (resistance to Hg*+; Diels et al., 1985; Dressler et al., 199I), chr (resistance to chromate; Nies et al., 1989a; seethe article written by Cervantes and Silver, 1992 (this issue)), and cop (resistanceto Cu*+; seethe article written by Brown et al., 1992 (this issue)). When copper-resistant Alcaligenes strains were isolated from wastewater, these strains contained czc, cnr, chr, and mer resistance determinants although the corresponding metal ions were not used for selection (Dressler et al., 1991). Moreover, CH34-like organisms were easily isolated from metalcontaminated habitats in Germany, Belgium, and Zaire (Schmidt and Schlegel, 1989; Kaur et al., 1990; Diels and Mergeay, 1990). Thus, the presence of a variety of metal resistance determinants in one bacterium seemsto occur frequently in nature. This article mainly introduces the plasmidbound cadmium-zinc resistance determinant from S. aureus as well as the cobaltzinc-cadmium resistance determinant from

A. eutrophus. TRANSPORT OF Zn2+, Cd*+, Co2+, AND Ni2+ INTO THE CELL

Zn*+ Co*+, and Ni*+ are essential for microorganisms as trace nutrients. Zn*+ is a component of many DNA and RNA poly-

merasesand in recent years many eucaryotic (and some procaryotic) gene regulatory proteins were found to contain a motif called a “zinc finger” for nucleic acid recognition (Te-Chung and Christie, 1990; Berg, 1990; Klug and Rhodes, 1987; Evans and Hollenberg, 1988). Alkaline phosphatase contains an essential Zn*+ playing a structural rather than a catalytic role; the cation servesto coordinate the phosphate group in the noncovalent enzyme-phosphate complex (Coleman, 1987). Co*+ is an essential component of vitamin B,, . Ni*+ has been identified as a trace element for bacteria more recently; Ni*+ is a component of hydrogenases, ureases, and some key enzymes in the metabolism of strictly anaerobic bacteria (Bartha and Ordal, 1965; Hausinger, 1987; Thauer et al., 1980). Like Zn*+, Ni*+ may play a structural role in addition to the functional; the Ni*+-containing hydrogenase from Nocardia opaca requires additional Ni*+ cations for correct subunit association (Schneider et al., 1984a,b). In contrast, Cd*+ has not been identified as a trace nutrient and is (in fact) thought to have no beneficial roles in bacteria. Trace elements such as Zn*+, Co*+, and Ni2+ must be transported into cells against concentration gradients-i.e., trace concentrations outside and substantial amounts within the cells. High-affinity and highly specific Ni*+ uptake systems were found in A. eutrophus (Lohmeyer and Friedrich, 1987), Anabaena cylindrica (Campbell and Smith, 1986), Methanobacterium bryantii (Jarrell and Sprott, 1982), Clostridium thermoaceticum (Lundie et al., 1988), and probably Bradyrhizobium japonicum (Stults et al., 1987). The specific Ni*+ transporter in A. eutrophus was identified as part of the plasmid-encoded hydrogenase (a Ni-containing enzyme) gene cluster on the 450-kb megaplasmid pHG1 (Eberz et al., 1989) and is a membranebound protein with a molecular mass of 33 kDa (Eitinger and Friedrich, 1991); it shows no homology with the Salmonella typhimurium Mg*’ transport systems already sequenced (Hmiel et al., 1989; Snavely et al., 1989a; Thomas Eitinger, personal communi-


cation) or other transport proteins in available sequence libraries. Specific uptake systems for trace elements and for macronutrients like Mg2+, phosphate, and sulfate can be synthesized by the cell in times of starvation or high need. However, under normal “housekeeping” conditions a few constitutively expressedbroad ion range uptake systems with relatively high rates satisfy the needs of bacterial cells for both macronutrients and trace elements. This conclusion is illustrated by the following examples: in A. eutrophus Zn2+, Co2+, Ni2+, and Cd’+ are transported into the cell by the Mg2+ uptake system and CrO:- is accumulated via the SOi- uptake system (Nies and Silver, 1989b). In S. typhimurium three Mg2+ transport systems, designated CorA (42-kDa protein), MgtA (9 1-kDa protein), and MgtB (102-kDa protein), were cloned and characterized (Hmiel et al., 1989; Snavely et al., 1989a). Mg2+ transport by CorA and MgtB is inhibited by Co2+and Ni2+, and transport by MgtA is additionally inhibited by Zn2+ (Snavely et al., 1989b). Ni2+ is also accumulated by the Mg2+ uptake system(s)in Clostridium pasteurianum (Bryson and Drake, 1988), Rhodobacter capsulatus (Jasper and Silver, 1978; Takakuwa, 1987) A. eutrophus (Lohmeyer and Friedrich, 1987), and even Neurospora crassa(Maruthi Mohan and Sivarama Sastry, 1983; Maruthi Mohan et al., 1984). In Escherichia coli and other enterobacteria Co2+ and Ni2+ are also transported into the cell by a constitutive Mg2+ transport system (Nelson and Kennedy, 1971; Park et al., 1976; Webb, 1970a,b). However, although conclusive data concerning the uptake of Zn2+ and Cd2+into the cell are lacking (Bucheder and Broda, 1974; Laddaga et al., 1985) both cations are possibly transported in gram-negative bacteria by the Mg2+ uptake system(s) (Abelson and Aldous, 1950). In contrast to this situation, Cd2+ is transported into cells of grampositive bacteria via the Mn2+ uptake system, e.g.,in Bacillus subtilis (Laddaga and Silver, 1985) and S. aureus (Weiss et al., 1978; Tynecka et al., 198la; Perry and Silver, 1982).


The uptake of a broad range of metal ions by a few high-capacity and relatively unspecific uptake systems is an economical solution for most cells and allows the accumulation of trace elements inside the celI for future needs. This physiological situation is the foundation for considerations concerning heavy metal toxicity and resistance: although Zn’+, Co2+, and Ni2+ are essential ions and may be transported by highly specific uptake systems at nanomolar concentrations, the function of those carriers should be repressedand might be much lower than that of the unspecific transporter at high (micro- or millimolar) concentrations. However, the fast broad ion substrate range uptake system(s)are constitutively expressed and cannot be shut down in times of metal ion stress: the metal cations are transported into the cell, bind to proteins, and interfere with the functions of other metal cations. Consequently, metal resistance determinants had to evolve. They are mostly plasmid-encoded, which could be the reason for the high frequence of metal resistance determinants in nature. RESISTANCE TO CADMIUM AND ZINC, COBALT AND NICKEL: METAL CATION BINDING VERSUS METAL CATION EFFLUX

In contrast to that of Hg’+, chemical reduction of Cd2+,Znzf, Co*+, or Ni2+ to the metallic form is not possible as a mechanism of resistance: NAD(P)H-dependent reduction of the divalent metal cations is energetically not favored. Moreover, the metals would not evaporate out of the cell like Hg” but remain inside and could be reoxidized. Therefore, possible mechanisms of resistance include only complexation and active efflux for Cd2+, Zn2+, Co2+,and Ni2+. Methylation and other covalent modifications of these metals are also not candidates for resistance mechanisms because the resulting organometallic compounds are unstable, mutagenic, and more toxic than the divalent cations. What mechanisms of resistance to Cd2+, Zn2’, Co2+,or Ni2+ have been realized in na-



ture? In general, eucaryotes produce binding factors and procaryotes prevent accumulation of Cd’+ Zn2+ Co2+, or Ni2+ by active cation effluxl However, there are exceptions to this rule, especially in yeasts and cyanobacteria. Moreover, binding factors could play a role asa second line of defensein bacteria with an efflux system as the main resistance system. In eucaryotes, two different metal-ionbinding factors are known. Small cysteinerich proteins named metallothioneins are synthesized under heavy metal stress and have been found in a wide variety of animals and lower eucaryotes (Kojima and Yagi, 1978; Hamer, 1986) and more recently in plants (de Miranda et al., 1990). These proteins are synthesized on ribosomes. Plants in general produce a different cysteine-rich polypeptide phytochelatin from gluthathione (Rauser, 1990). All plants examined are able to produce phytochelatin (or the closely related homophytochelatin) (Grill, 1989), and older publications dealing with cadmium-binding “proteins” or “complexes” in plants probably were describing phytochelatins (Fujita and Kawanishi, 1986, 1987). Phytochelatin is chemically a poly(yglutamylcysteinyl)n-glycine with n = 2 to n = 11 (Grill et al., 1987; Jackson et al., 1987). This compound is produced from glutathione (or homoglutathione) by a specific y-glutamylcysteine dipeptidyltranspeptidase; it is not a protein and is not made on ribosome. The activity of this phytochelatin synthase is Cd2+-dependentand therefore autoregulated; the product (phytochelatin) complexes Cd” and the synthase is deactivated when all the Cd2+ is bound (Grill et al., 1989). Inhibition of phytochelatin biosynthesis results in rapid Cd2+-dependent cell death. However, Cdsensitive cells are capable of synthesizing amounts of phytochelatin equivalent to those of Cd-resistant cells. Thus, other mechanisms must contribute to the Cd2+ tolerance of plant cells (Jackson et al., 1989). In some yeasts (including S. cerevisiae) metallothioneine-like polypeptides were found (Mehra et al., 1989). Cd2+-binding

peptides (phytochelatins) were found in Schizosaccharomycespombe (Murasugi et al., 1981, 1984). Mutants unable to synthesize yeast phytochelatin lack activity of y-glutamylcysteine dipeptidyltranspeptidase or glutathione synthetase (Mutoh and Hayashi, 1988; Hayashi et al., 1988). By still another system, S. cerevisiae shows cadmium and zinc resistance associated with a 442-aminoacid putative membrane-bound protein designated ZRC- 1 for “zinc resistance conferring” (Kamizomo et al., 1989) which may be involved in cation efflux. While cation efflux is the exception as a basic mechanism of metal resistance for eucaryotic organisms and metal-binding components are rare among the faster growing bacteria, apparently good examples of metallothioneins in Pseudomonasand a cyanobacterium, Synechococcus,have been described. Three related cysteine-rich small proteins in a cadmium-resistant strain of Pseudomonas putida have been described (Higham et al., 1984). NMR analysis of these proteins showed sulfur-cadmium domains similar to mammalian methallothioneins. Unfortunately, the strain lost its ability to produce those cadmium-binding proteins. The inducible cyanobacterial metallothionein-like protein is acidic (~1 4.3) and has a molecular mass of 8.1 kDa (Olafson et al., 1979, 1980). Transcripts of the metallothionein gene smtA were maximally abundant after exposure to heavy metal cations (e.g., 5 PM Cd’+, 10 CLM Zn2+, or 10 j~A4Cu2+) but did not increase in abundance following a heat shock (Robinson et al., 1990). The metallothionein molecule complexes Cd’+, Zn’+, and Cu2+ and has some unusual properties: (i) a pair of histidine residues is located in repeating glycinehistidine-threonine-glycine sequences near the carboxy-terminus; (ii) six long-chain aliphatic and two aromatic amino acid residues make this protein the most hydrophobic metallothionein described; (iii) there is no association of hydroxylated residues with the cysteine residues; and (iv) the cysteine content of 19% is much lower than that of eucaryotic metallothioneins (which is about 32%). More-



Staphylococcus aureus S. aureus S. aureus Pseudomonasputida P. putida Thiobacillus thiooxidans Synechococcussp. Aicaligenes eutrophus A. eutrophus

Resistance to Cd’+ Zn” a2” Zn2f Cd’+: (not Zn2+) Cd’+ Zn2+ Cd2+ Cd2’ Zn2+ Cd2+’ Cd2+,Zn2+ Co2+ Co2+Ni2C’

Determinant &CA cadl?

czcCBAD cnr



~1258 p11147 Chromosome pGu100 ? ? ? pMOL30 pMOL28

Efflux Binding? Efflux Efflux? Binding Binding Binding Efflux Efflux

Reference t i 4 g h i

References.(a) Nucifora et al., 1989; Tynecka et al.; 198lb. (b) ElSolh and Ehrlich, 1982; Perry and Silver, 1982; Novick et al., 1979; Shalita et al., 1980; Smith and Novick, 1972. (c) Witte et al., 1986. (d) Horitsu et al., 1986. (e) Higham et al., 1984. (f) Sakamoto et af., 1989.(g) Olafson et al., 1979, 1980.(h) Nies et al., 1989a.(i) Siddiqui et al., 1989; Nies et al., 1989b.

over, no significant homology was found between the cyanobacterial and the eucaryotic metallothioneins. Therefore, cyanobacterial and eucaryotic metallothioneins are probably the product of a convergent evolution (Olafson et al., 1988). Table 1 shows a survey of systems encoding resistance to cadmium, zinc, cobalt, and/ or nickel ions in bacteria. In addition to the binding factors described above, at least six different determinants are known to encode cadmium resistance in bacteria (Silver and Misra, 1988); three of them are known in the gram-positive bacterium S. aureus: the cadA determinant catalyzes efflux of Cd” and Zn2+ (Tynecka et al., 1981b). Some plasmids (e.g., ~1258) in S. aureus contain only the cadA determinant; others (e.g., ~11147)contain a second resistance determinant designated cadB (Novick et al., 1979; Shalita et al., 1980; Smith and Novick, 1972). Cadmium resistance encoded by cadB is independent of cad2 (ElSolh and Ehrlich, 1982), weaker compared to resistance mediated by cad4, not additive to the action of cad4, and possibly based on the synthesis of a cadmiumbinding factor (Perry and Silver, 1982). The third type of determinant is also based on cadmium cation efflux; however, this chromosomal determinant does not encode resistance to Zn2+ (Witte et al., 1986).

Nothing is known about the reported plasmid-borne cadmium resistance in Rhodococcus (Dabbs and Sole, 1988). The czc determinant of the gram-negative bacterium A. eutrophus encodes resistance to Cd2+,Zn2+, and Co2+ by metal-dependent efflux (Nies and Silver, 1989a) and the plasmid-encoded cadmium resistance from P. putida also functions in reducing the accumulation of the toxic cation (Horitsu et al., 1986). Plasmid-determined resistance to Co2+ and Ni2+ is an old example of metal cation resistance (Smith, 1967), but was only recently studied in more detail in A. eutrophus. The cnr resistance system in this bacterium is encoded by plasmid pMOL28 and is primarily designed as a nickel resistance: it is induced effectively only by Ni2+ in growth experiments (Siddiqui and Schlegel, 1987; Siddiqui et al., 1988) and experiments with cnr::lacZ operon fusions (D. H. Nies, unpublished). Resistance encoded by the cnr determinant is based on energydependent specific-cation efflux (Sensfuss and Schlegel, 1988; Varma et al., 1990; Nies and Silver, 1989a). The cnr determinant was cloned (Nies et al., 1989b; Siddiqui et al., 1989). However, it has not yet been sequenced. Therefore, no information about the subunit composition of the cnr efflux pump is currently available.



the DNA sequence encoding a small (122 aa)’ and a large (727 aa) protein. The second protein was essential for cadmium resistance and was therefore referred to asthe cadA gene product. When the primary sequence of the 2+ Mn t putative CadA protein was compared to a liCd 2+ 2+ brary of protein sequences,a strong homolZn ogy was found with members of a protein family known as “E 1-EZcation transporting I! ATPases” (now renamed “P-type ATPases”), which alter during their catalytic cycle be\ & CadCA ,/ tween two conformations called El (cation site outside) and E2 (cation site inside). One R member of this family of enzymes is the Cd2+ Zn 2+ Ca2+-transporting ATPase from rabbit musFIG. 1. Resistance to cadmium and zinc in Staphylocles. Becauseof the strong homology between coccusaureus. Cd’+ is transported into the cell by a man- CadA and members of this protein family, ganeseuptake system and Zn2+ is transported by an unknown system (possibly a magnesium uptake system). In models developed to explain the action of the presenceofplasmid ~1258 (containing the cadA deter- those transportes could be transferred onto minant located on a 3.5kb BgfiI-XbaI fragment) the the CadA transporter (Nucifora et al., 1989; CadCA protein complex is induced and actively trans- Silver et al., 1989). ports both cations out of the cell. G, Bg/II site; A, XbaI All E 1-E2 transporters are composed of a site. membrane tunnel and four interacting cytoplasmic domains: the ATP-binding site, the protein kinase site, a transduction domain, Two resistance systemsbased on active efand the substrate binding site which is loflux of the cation, the cudA system from S. cated near the N-terminus of the protein. The aureusand the czc system from A. eutrophus, homology between CadA and other members have been characterized in detail and are reof the El-E2 family does not include the Nviewed in the following two sections. terminal 100 amino acid residues which show homology in fact to the N-terminal reCADMIUM-ZINC RESISTANCE IN gion of the mercuric ion reductase and the Staphylococcus aureus: THE CadA MerP periplasmic mercuric ion-binding proCATION EFFLUX SYSTEM tein. In all three proteins a conserved pair of The cud4 determinant of S. uureus is lo- cysteine residues may be involved in the cated on plasmid ~1258and related plasmids “soft” binding of Cd’+ and Hg’+, respec(Novick et al., 1979); Fig. 1 illustrates cad- tively. When the cosubstrate ATP is bound to mium resistance in this bacterium. Cd’+ is the ATP-binding domain, a phosphate group transported into the S. aureus cell by the could be transferred to the phosphate-bindMn2+ uptake system (Weiss et al., 1978, Tyn- ing domain and the protein changes into a ecka et al., 198la) and Zn2+ possibly by the conformation of high energy by hydrolyzing Mg2+ uptake system. In the presence of the this phosphate bond. This state of high encadA determinant a specific efflux of Cd2+ ergy is possibly relaxed again by transport of has been demonstrated by loading the cells the cation through the membrane (Silver et with lo9Cd2+and diluting them into Cd’+-free al., 1989). The CadA protein alone is not sufficient to medium (Tynecka et al., 198lb). The cadA determinant was cloned on a confer full resistance to cadmium and zinc; 3.7-kb BglII-XbaI fragment and sequenced. Two open reading frames were identified in ’ Abbreviation used: aa, amino acid. Cd 2+ Mn 2+



the second reading frame of the resistance determinant (named cadC) also must be present (Yoon and Silver, 1991). The cadC gene product is a small protein ( 122 aa) containing three possible metal-binding motifs (Nucifora et al., 1989): CEIFC, 1, CQDEELCVC,, and SHHLR,, . The function of CadC in Cd2+ efflux is not clear; it may function in gathering the cations from the cytoplasm and channeling them onto the CadA transporter (Yoon and Silver, 1991). The two genes cadC and cadA overlap for eight nucleotides (Nucifora et al., 1989) and are transcribed together (Yoon et al., 1991). Although the regulator of the cad4 determinant has not been identified, cadmium resistance in inducible and cadA:: bla fusions show a regulation of ,&lactamase activity by Cd2+.Bi3+ and Pb2+are also good inducers; in contrast, Zn2+ and Co2+ induce only weakly (Yoon and Silver, 199I). The cad4 determinant from plasmid ~I258 is therefore mainly a cadmium resistance system. However, since cadmium is often accompanied by zinc, the “by-product” zinc resistance should be beneficial for the cells carrying the cadA determinant. CADMIUM-ZINC-COBALT RESISTANCE IN Alcaligenes eutrophus: THE czc CATION EFFLUX SYSTEM

The A. eutrophus strain CH34 was isolated from a zinc decant&ion tank (Mergeay et al., 1978) and contains two large plasmids (Gerstenberg et al., 1982) designated pMOL28 (163 kb) and pMOL30 (238 kb). These plasmids harbor a variety of metal resistance determinants (Mergeay et al., 1985). Figure 2 illustrates the function of the plasmid pMOL30-encoded czc determinant: the toxic cations Cd2+ Zn2+, and Co2+appear to enter the cell via ;he Mg2+ uptake system. In the presence of the czc determinant, all three cations are actively extruded from the cell. When accumulation of toxic cations by resistant (czc-containing) and sensitive strains of A. eutrophus was compared, the resistant cells showed reduced accumulation of the cat-

2+ Mg 2+






CzcABC 2+ co

2+ zn


2+ Cd

FIG. 2. Resistance to Co’+, Zn”, and Cd’+ in Alcaligenes eutrophus. Co’+, Zn2+, and Cd2’ are transported into the cell by a magnesium uptake system. In the presence of plasmid pMOL30 (containing the czc determinant located on a 8.0-kb KpnI-BarnHI fragment) the CzcCBA protein complex is induced and actively transports both cations out of the cell. CzcD is involved in regulation of czc. K, KpnI site; H, Hind111site.

ions at 30°C. However, the processleading to reduced accumulation of the cations was inhibited at 4°C. This situation led to the demonstration of cation efflux: if the cells were incubated at 4°C sensitive and resistant cells accumulated equal amounts of metal cations. However, when they were shifted to 30°C the resistant cells showed an efflux of the cations while the cation content of the sensitive cells remained constant. This basic mechanism of resistance could be demonstrated for resistance to Cd2+,Zn2+, and Co’+ encoded by czc and resistance to Co2+ and Ni2+ encoded by cnr (Nies and Silver, 1989b). The czc determinant was initially cloned with a 9.1-kb EcoRI fragment (Nies et al., 1987), subcloned, and sequenced (Nies et al., 1989a). Four open reading frames were identified in the sequence and assigned to polypeptides (CzcA, CzcB, CzcC, and CzcD) expressedin E. coli from the determinant under control of the phage T7 promoter. Based on the phenotype of mutants carrying deletions in the czc determinant and computer-assisted modeling of the primary sequenceof the four


DIETRICH H. NIES CzcC: modification of substrate sperificity

CzcB, cation binding

CzcA, main pump protein

CzcD, regulation

FIG. 3. Model of action of the czc cation resiitance determinant. The hypothetical functions the gene products of the CZCCBADdeterminant are illustrated. For details, seetext.

proteins, a model which explains all the currently available data could be developed (Fig. 3; Nies et al., 1989a). The largest of the four proteins, CzcA (1064 aa), is essential for cation transport and the heart of the efflux protein complex. Only CzcA shows potential transmembrane a-helices and is thus capable of forming a membrane tunnel. CzcA alone is probably able to catalyze a slow Co’+ efflux as judged by the residual cobalt resistance conferred by strains carrying deletions of czcB or CZCC.In contrast to CadA of S. aureus, no putative ATPbinding site has been identified in the primary sequenceof CzcA. Therefore, a cationproton antiport is the hypothesized mechanism of action for the czc-encoded efflux protein complex (Nies et al., 1989a). Secondary structure predictions show four domains as constituents of the CzcA protein: two hydrophilic and two hydrophobic domains. In contrast to the hydrophobic region in the middle of the two predicted cytoplasmic domains, the N-terminal membrane region contains transmembrane cy-heliceswith positively and negatively charged amino acid residues in the middle of the membrane span. This N-terminal domain therefore might be a proton “tunnel” transporting the protons

with a charge-relay system (Kaback, 1988) across the cytoplasmic membrane. The second hydrophobic domain which does not contain charged residues in the middle of the membrane accordingly could act as a cation tunnel. The CzcA protein is very low in cysteine and histidine content and thus low in possible metal-binding sites. Metal cation binding may be the function of the second largest protein of czc, the CzcB protein (521 aa). CzcB contains eight histidine residues in the middle domain of three putative domains. These histidine residues are arranged in two possible metal-binding sites containing four histidine residues each. Both sites show homology to each other. The CzcA and the CzcB proteins function together as a Zn2+ efflux pump. In contrast to CadA (which is a Cd2’ pump in the first place) no cysteine residue is present in the proposed initial cation-binding site of CzcAB (Nies et al., 1989a). The third protein of the efflux complex, the CzcC protein (346 aa), is predicted to function as a modifier switching the substrate specificity of the transporter from Zn2+ only to Cd2’, Zn2+, and Co2+ (Fig. 3; Nies et al., 1989a). CzcC does not contain histidines or cysteines and is dependent on the CzcB pro-


tein for function. With respectto this, there is an interesting parallel between the proposed function of the CzcC protein and that of the ArsC protein, which switches specificity of the arsenate-arseniteresistancecomplex (see the paper by Rosen and Kaur). The CzcD protein (200 aa) is not needed for cation efflux. Together with components encoded by the 1.2-kbKpnI-EcoRI fragment located upstream of the fragment encoding czcCBAD, CzcD is involved in regulation of czc (D. H. Nies, unpublished). ACKNOWLEDGMENTS Thanks to Simon Silver for critically reading the manuscript. My recent work on metal resistance in A. eutrophus was supported by a grant from the Bundesminister fur Forschung und Technologie as a project of the Genzentrum Berlin.

REFERENCES ABELSON,P. H., AND ALDOUS, E. (1950). Ion antagonism in microorganisms: Interference of normal magnesium metabolism by nickel, cobalt, cadmium, zinc and manganese. J. Bacterial. 60,40 I-4 13. BARTHA, R., AND ORDAL, E. J. (1965). Nickeldependent chemolithotropic growth of two Hydrogenomonas strains. J Bacterial. 89, 1015- 1019. BERG,J. M. (1990). Zinc fingers and other metal-binding domains. J. Biol. Chem. 265,6513-6516. BRYSON,M. F., AND DRAKE, H. L. (1988). Energy-dependent transport of nickel by Clostridium pasteurianum. J. Bacterial. 170,234-238. BUCHEDER,F., AND BRODA, E. (1974). Enetgydependent zinc transport by Escherichia coli. Eur. J Biothem. 45,555-559. CAMPBELL,P. M., AND SMITH, G. D. (1986). Transport and accumulation of nickel ions in the cyanobacterium Anabaena cylindrica. Arch. Biochem. Biophys. 244,470-477. COLEMAN,J. E. (1987). Multinuclear nuclear magnetic resonanceapproaches to the structure and mechanism of alkaline phosphatase. In “Phosphate Metabolism and Cellular Regulation in Microorganisms” (A. Torriani-Gotini, F. G. Rothman, S. Silver, A. Wright, and E. Yagil, Eds.). ASM, Washington, DC. DABBS,E. R., AND SOLE,G. J. ( 1988). Plasmid-borne resistance to arsenate, araenite, cadmium, and chloramphenicol in a Rhodococcus species. Mol. Gen. Genet. 211, 148-154. DE MIRANDA, J. R., THOMAS,M. A., THURMAN, D. A., AND TOMSET~,A. B. (1990). Metallothionein genes from the flowering plant Mimulus guttatus. FEBS Lett. 260, 277-280.


DIELS, L., FAELEN,M., MERGEAY,M., AND Nms, D. (1985). Mercury transposons from plasmids governing multiple resistance to heavy metals in Alcaligenes eutrophus CH34. Arch. Int. Physiol. Biochim. 93, B27B28. DIELS, L., AND MERGEAY,M. (1990). DNA probe-mediated detection of resistant bacteria from soils highly polluted by heavy metals. Appl. Environ. Microbial. 56, 1485-1491. DRESSLER,C., Kuss, U., NIES, D. H., AND FRIEDRICH, B. ( 1991).Determinants encoding multiple metal resistance in newly isolated copper-resistant bacteria. Appl. Environ. Microbial. 57, (Nov.), in press. EBERZ,G., EITINGER,T., AND FRIEDRICH,B. ( 1989).Genetic determinants of a nickel-specific transport syatern are part of the plasmid-encoded hydrogenaaegene cluster in Alcaligenes eutrophus. J. Bacterial. 171, 1340-1345. EITINGER,T., AND FRIEDRICH,B. (199 1). Cloning, nucleotide sequence and heterologous expression of a high-affinity nickel transport gene from Afcaligenes eutrophus. J. Biol. Chem. 266,3222-3227. ELSOLH, N., AND EHRLICH, S. D. (1982). A small cadmium resistance plasmid isolated from Staphylococcus aureus. Plasmid 7,77-84. EVANS, R. M., AND HOLLENBERG,S. M. (1988). Zinc fingers: Gilt by association. Cell 52, 1-3. FUJITA, M., AND KAWANISHI, T. (1986). Purification and characterization of a Cd-binding complex from root tissue of water hyacinth cultivated in a Cd*+-containing medium. Plant Cell Physiol. 27, 1317- 1325. FLJJITA,M., AND KAWANISHI, T. (1987). Cd-binding complexes from root tissues of various higher plants cultivated in Cd*+-containing medium. Plant Cell Physiol. 28,379-382. FUREY,W. F., ROBBINS,A. H., CLANCY,L. L., WINCE, D. R., WANG,B. C., AND STOUT,C. D. ( 1986).Crystal structure of Cd, Zn metallothionein. Science 231, 704-710. GERSTENBERG,C., FRIEDRICH, B., AND SCHLEGEL, H. G. ( 1982). Physical evidence for plasmids in autotrophic, especially hydrogen-oxidizing bacteria. Arch. Microbial. 133, 90-96. GRILL, E. (1989). Phytochelatins in plants. In “Metal Ion Homeostasis: Molecular Biology and Chemistry,” pp. 280-300. A. R. Liss, New York. GRILL, E., WINNACKER,E.-L., AND ZENK, M. H. ( 1987). Phytochelatins, a class of heavy metal-binding pep tides from plants, are functionally related to metallothioneins. F’roc. Natl. Acad. Sci. USA 84,439-443. GRILL, E., L&FLER, S., WINNACKER,E.-L., AND ZENK, M. H. (1989). Phytochelatins, the heavy-metal-binding peptides of plants, are synthesized from glutathione by a specific [gamma]-glutamyl-cysteine dipeptidyl transpeptidase (phytochelatin synthaae). Proc. Natl. Acad. Sci. USA 86,6838-6842. HAMER, D. H. (1986). MetalIothionein. Annu. Rev. Biothem. 55,9 13-95 1.



HAUSINGER,R. P. (1987). Nickel utilization by microorganisms. Microbial. Rev. 51,22-42. HAYASHI, Y., NAKAGAWA,C. W., UYAKUL, D., IMAI, K., ISOBE,M., AND GOTO, T. (1988). The change of cadystin components in Cd-binding peptides from the fission yeast during their induction by cadmium. Biothem. Cell. Biol. 66, 288-295. HIGHAM, D. P., SADLER,P. J., AND SCAWEN,M. D. (1984). Cadmium-resistant Pseudomonasputidu synthesizes novel cadmium proteins. Science 225, 10431046. HMIEL, S. P., SNAVELY,M. D., FLORER,J. B., MAGUIRE, M. E., AND MILLER, C. G. (1989). Magnesium transport in Salmonella typhimurium:Geneticcharacterization and cloning of three magnesium transport loci. J. Bacterial. 171,4742-475 1. HORITSU,H., YAMAMOTO,K., WACHI, S., KAWAI, K., AND FUKUCHI, A. (1986). Plasmid-determined cadmium resistance in Pseudomonasputida CAM- 1 isolated from soil. J. Bacterial. 165, 334-335. JACKSON,P. J., UNKERF,C. J., DOLLEN,J. A., WAN, K., AND ROBINSON,N. J. (1987). PoIy(gamma-glutamylcysteinyl)glycine: Its role in cadmium resistance in plant cells. Proc. Natl. Acad. Sci. USA 84,66 19-6623. JACKSON,P. J., ROBINSON,N. J., AND DELLAIZE, E. (1989). Cd-tolerance in plant cells: A comparison of biochemical and molecular properties of tolerant and sensitive cells. In “Metal Ion Homeostasis: Molecular Biology and Chemistry,” pp. 337-346. A. R. Liss, New York. JARRELL,K. F., AND SPROTT,G. D. (1982). Nickel transport in Methanobacterium bryantii. J. Bacterial. 151, 1195-1203. JASPER,P., AND SILVER,S. (1978). Divalent cation transport systemsof Rhodopseudomonascapsulata. J. Bacteriol. 133, 1323-1328. KABACK, H. R. (1988). Site-directed mutagenesis and ion-gradient driven active transport: on the path of the proton. Annu. Rev. Physiol. 50, 243-256. KAMIZOMO,A., NISHIZAWA,M., TERANISHI,Y., MURATA, K., AND KIMURA, A. (1989). Identification of a gene conferring resistanceto zinc and cadmium in the yeast Saccharomycescerevisiae. Mol. Gen. Genet. 219, 161-167. KAUR, P., Ross, K., SIDDIQUI, R. A., AND SCHLEGEL, H. G. (1990). Nickel resistanceofAlcaligenes denitrijicans strain 4a-2 is chromosomally encoded. Arch. Microbiol. 154, 133-138. KLUG, A., AND RHODES,D. (1987). “Zinc fingers.” A novel protein motif for nucleic acid recognition. Trends Biochem. Sci. 12,464-469. KOJIMA, Y., AND KAGI, J. H. R. (1978). MetaIlothionein. Trends Biochem. Sci. 3,90-93. LADDAGA,R. A., AND SILVER,S. (1985). Cadmium uptake in Escherichia co/i K-12. J. Bacteriul. 162, 11051100. LADDAGA, R. A., BE%?,!+R., AND SILVER, S. ( 1985). Cadmium-resistant mutant of Bacillus subtilis 168

with reduced cadmium transport. J. Bucteriol. 162, 1106-1110. L~HMEYER,M., AND FRIEDRICH,C. G. (1987). Nickel transport in Alcaligenes eutrophus. Arch. Microbial. 149, 130-135. LUNDIE, L. L., YANG, H., HEINONEN,J. K., DEAN, S. J., AND DRAKE,H. L. (1988). Energy-dependent, high-affinity transport of nickel by the acetogen Clostridium thermoaceticum. J. Bacterial. 170, 5705-5708. MARUTHI MOHAN, P., AND SIVARAMA SASTRY, K. (1983). Interrelationships in trace-element metabolism in metal toxicities in nickel-resistant strains of Neurospora crassa. Biochem. J. 212,205-210. MARUTHI MOHAN, P., PRATAPRUDRA, M. P., AND SIVARAMA SASTRY, K. (1984). Nickel transport in nickel-resistant strains of Neuropsora crassa. Curr. Microbial. 10, 125- 130. MEHRA, R. K., GAREY, J. R., TAUSEEF,R. B., GRAY, W. R., AND WINGE, D. R. (1989). Candida glabrata metallothioneins: Cloning and sequence of the genes and characterization of the proteins. J. Biol. Chem. 264, 19,747-19,753. MERGEAY,M., HOUBA,C., ANDGERITS,J. (1978). Extrachromosomal inheritance controlling resistance to cadmium, cobalt, and zinc ions: Evidence from curing in a Pseudomonas. Arch. Int. Physiol. Biochem. 86, 440-44 1. MERGEAY,M., NIES, D., SCHLEGEL,H. G., GERITS,J., CHARLES,P., AND VANGIJSEGEM,F. (1985). Alcaligenes eutrophus CH34 is a facultative chemolithotroph with plasmid-bound resistance to heavy metals. J. Bacterial. 162, 328-334. MURASUGI, A., WADA, C., AND HAYASHI, Y. (1981). Cadmium-binding peptide induced in fission yeast, Schizosaccharomyces pombe. J. Biochem. 90, 15611564. MURASUGI,A., NAKAGAWA,C. W., AND HAYASHI, Y. ( 1984). Formation of cadmium-binding peptide allomorphes in fission yeasts.J. Biochem. 96, 1375- 1379. MUTOH, N., AND HAYASHI, Y. (1988). Isolation of mutants ofSchizosaccharomycespombe unable to synthesize cadystin, small cadmium-binding peptides. Biothem. Biophys. Res. Commun. 151, 32-39. NELSON, D. L., AND KENNEDY,E. P. ( 1971).Magnesium transport in Escherichia coli: Inhibition by cobaltous ion. J. Biol. Chem. 246, 3042-3049. NIES,D., MERGEAY,M., FRIEDRICH,B., AND SCHLEGEL, H. G. (1987). Cloning of the plasmid coded resistance to cobalt, zinc, and cadmium from Alcaligenes eutrophus CH34. J. Bacterial. 167, 4865-4868. NIES, D. H., NIES, A., CHU, L., AND SILVER,S. (1989a). Expression and nucleotide sequence of a plasmiddetermined divalent cation efflux system from Alcaligenes eutrophus. Proc. Natl. Acad. Sci. USA 86,735 l7355. NIES, A., NIES, D. H., AND SILVER,S. (1989b). Cloning and expression of plasmid genes encoding resistances

Cd, Co, Zn, AND Ni RESISTANCE IN MICROBES to chromate and cobalt in Alcaligenes eutrophus. J. Bacterial. 171, 5065-5070. NIES, D. H., AND SILVER, S. (1989a). Plasmiddetermined inducible efflux is responsible for resistance to cadmium, zinc, and cobalt in Alcaligenes eutrophus. J. Bacterial. 171, 896-900. NIES, D. H., AND SILVER,S. (1989b). Metal ion uptake by plasmid-free metal-sensitive Alcaligenes eutrophus strain. J. Bacterial. 171,4073-4075. NOVICK, R. P., AND ROTH, C. (1968). Plasmid-linked resistance to inorganic salts in Staphylococcus aureus. J. Bacterial. 95, 1335-1342. NOVICK, R. P., MURPHY, R. P., GRYCZAN,T. J., BARONE,E., AND EDELMAN,I. (1979). Penicillinase plasmids of Staphylococcus aureus: Restriction-deletion maps. Plasmid 2, 109- 129. NUCIFORA,G., CHU, L., MISRA, T. K., AND SILVER,S. (1989). Cadmium resistancefrom Staphylococcus aureus plasmid ~1258 cadA results from a cadmium-efflux ATPase. Proc. Natl. Acad. Sci. USA 86, 35443548. OLAESON,R. W., ABEL, R. S., AND SIM, R. S. (1979). Procaryoticmetallothionein: Preliminarycharacteriization of a blue-green algae heavy metal-binding protein. Biochem. Biophys. Res. Commun. 89,36-43. OLAFSON,R. W., LOYA, S.,AND SIM, R. G. ( 1980).Physiological parameters of procaryotic metallothionein induction. Biochem. Biophys. Res. Commun. 95,14951503. OLAFSON,R. W., MCCUBBIN, W. D., AND KAY, C. M. (1988). Primary- and secondary-structural analysis of a unique prokaryotic metallothionein from a Synechococcussp. cyanobacterium. Biochem. J. 251,69 l-699. PARK, M. H., WONG, B. B., AND LUSK, J. E. (1976). Mutants in three genes affecting transport of magnesium in Escherichia coli. J. Bacterial. 126,1096-l 103. PERRY, R. D., AND SILVER, S. (1982). Cadmium and manganese transport in Staphylococcus aureus membrane vesicles. J. Bacterial. 150, 973-976. RAUSER,W. E. (1990). Phytochelatins. Annu. Rev. Biothem. 59,6 l-86. ROBINSON, N. J., GUPTA, A., FORDHAM-SKELTON, A. P., CROY,R. R. D., WHITTON,B. A., ANDHUCKLE, J. W. ( 1990).Prokaryotic metallothionein genecharacterization and expression: chromosomal crawling by ligation-mediated PCR. Proc. Roy. Sot. London B 242,241-247. SAKAMOTO,K., YAGASAKI, M., KIRIMURA, K., AND USAMI, S. (1989). Resistance acquisition of Thiobacillus thiooxidans upon cadmium and zinc ion addition and formation of cadmium ion-binding and zinc ionbinding proteins exhibiting metallothionein-like prop erties. J. Ferment. Bioeng. 67, 266-273. SCHMIDT,T., AND SCHLEGEL,H. G. (1989). Nickel and cobalt resistance of various bacteria isolated from soil and highly polluted domestic and industrial wastes. FEMS Microbial. Ecol. 62, 3 15-328. SCHNEIDER,K., CAMMACK, R., AND SCHLEGEL,H. G.


(1984a). Content and localization of FMN, Fe-S clusters and nickel in the NAD-linked hydrogenase of Nocardia opaca lb. Eur. J. Biochem. 142,75-84. SCHNEIDER,K., SCHLEGEL,H. G., AND JOCHIM, K. ( 1984b). Effect of nickel on activity and subunit composition of purified hydrogenase from Nocardia opaca 1b. Eur. J. Biochem. 138,533-541. SENSFUSS, C., AND SCHLEGEL,H. G. (1988). Plasmid pMOL28-encoded resistance to nickel is due to specific efflux. FEMS Microbial. Lett. 55,295-298. SHALITA, Z., MURPHY, E., AND NOVICK, R. P. (1980). Penicillinase plasmids of Staphylococcusaureus: structural and evolutionary relationships. Plasmid 3,29 l311. SIDDIQUI,R. A., AND SCHLEGEL,H. G. (1987). Plasmid pMOL28 mediated inducible nickel resistance in Alcaligenes eutrophus CH34. FEMSMicrobiol. Lett. 43, 9-13. SIDDIQUI, R. A., SCHLEGEL,H. G., AND MEYER, M. (1988). Inducible and constitutive expression of pMOL28-encoded nickel resistance in Alcaligenes eutrophus N9A. J. Bacterial. 170,4 188-4 193. SIDDIQUI, R. A., BENTHIN, K., AND SCHLEGEL,H. G. ( 1989).Cloning of pMOL28-encoded nickel resistance genesand expression of the genesin Alcaligenes eutrophus and Pseudomonas spp. J. Bacterial. 171, 50715078. SILVER,S., AND MISRA, T. K. (1988). Plasmid-mediated heavy metal resistances. Annu. Rev. Microbial. 42, 717-743. SILVER,S., NUCIFORA,G., CHU, L., AND MISRA, T. K. (1989). Bacterial resistance ATPases: Primary pumps for exporting toxic cations and anions. Trends Biothem. Sci. 14,76-80. SMITH,D. H. (1967). R factors mediate resistanceto mercury, nickel and cobalt. Science 156, 1114-I 116. SMITH, K., AND NOVICK, R. P. (1972). Genetic studies on plasmid-linked resistance in Staphylococcus aureus. J. Bacterial. 112, 761-772. SNAVELY,M. D., FU)RER, J. B., MILLER, C. G.: AND MAGUIRE,M. E. (1989a). Magnesium transport in Salmonella typhimurium: Expression of cloned genes for three distinct Mg’+ transport systems. J Bacterial. 171,4752-4760. SNAVELY,M. D., FLORER,J. B., MILLER, C. G., AND MAGUIRE,M. E. ( 1989b).Magnesium transport in Salmonella typhimurium: 2sMg2+ transport by CorA, MgtA, and MgtB systems. J. Bacterial. 171, 47614766. STULTS,L. W., MALLICK, S., AND MAIER, R. J. (1987). Nickel uptake in Bradyrhizobium japonicum. J. Bacteriol. 169, 1398- 1402. TAKAKUWA, S. (1987). Nickel uptake in Rhodopseudomonas capsulata. Arch. Microbial. 149, 57-6 1. TE-CHUNG,L., AND CHRISTIE,G. E. (1990). Purification and properties of the bacteriophage P2 ogr gene product. A prokaryotic zinc-binding transcription activator. J. Biol. Chem. 265, 7472-7477. THAUER, R. K., DIEKERT, G., AND SCH~NHEIT, P.



( 1980). Biological role of nickel. Trends Biochem. Sci. 5,304-306.

TYNECKA,Z., Gas, Z., AND ZAJAC,J. (198 la). Reduced cadmium transport determined by a plasmid in Stuphylococcus aureus. J. Bacterial. 147, 305-3 12. TYNECKA,Z., Gos, Z., AND ZAJAC,J. (198 lb). Energydependent efflux of cadmium coded by a plasmid resistance determinant in Staphylococcus aureus. J. Bacteriol. 147, 3 13-3 19. VARMA, A. K., SENSFUSS, C., AND SCHLEGEL,H. G. ( 1990). Inhibitor effects on the accumulation and efflux of nickel ions in plasmid pMOL28-harboring strain of Alcaligenes eutrophus. Arch. Microbial. 154, 42-49.

WEBB, M. (1970a). The mechanism of acquired resistance to Co’+ and N?’ in Gram-positive and Gramnegative bacteria. Biochim. Biophys. Acta 222, 440446.

WEBB,M. (1970b). Interrelationship between the utilization of magnesium and the uptake of other bivalent cations by bacteria. Biochim. Biophys. Acta 222,428439.

WEISS,A. A., SILVER,S., AND KINSCHERF,T. G. (1978). Cation transport alteration associated with plasmiddetermined resistance to cadmium in Staphylococcus aureus. Ant. Agents Chemother. 14,856-865. WI?TE, W., GREEN, L., MISRA, T. K., AND SILVER,S. (1986). Resistance to mercury and to cadmium in chromosomally resistant Staphylococcus aureus, Antimicrob. Agents Chemother. 29,663-669. YOON,K. P., MISRA,T. K., AND SILVER,S.( 1991).Regulation of the cadA cadmium resistance determinant of Staphylococcus aureus. Submitted for publication. YOON, K. P., AND SILVER,S. (1991). A second gene in the Staphylococcus aureus cadA cadmium resistance determinant. Submitted for publication.

Resistance to cadmium, cobalt, zinc, and nickel in microbes.

The divalent cations of cobalt, zinc, and nickel are essential nutrients for bacteria, required as trace elements at nanomolar concentrations. However...
2MB Sizes 0 Downloads 0 Views