Candida glabrata Metallothioneins

Southern blot analysis has identified several metallothionein gene sequences in a human pathogenic yeast Candida glabrata. Two of these genes encoding proteins designated MT-I and MT-I1 have been cloned and sequenced. No introns were found in either of the genes. The complete primary structure of MT-I1 was also determined by protein sequencing methods. As isolated, MT-I and MT-I1 consist of 62 and 51 amino acids, respectively. The only residues predicted from the nucleotide sequence but not present in the isolated protein are the amino-terminal methionines in each sequence. MT-I contains 18 cysteines, 14 of which are present as Cys-X-Cys motifs and two additional cysteines in a Cys-X-X-Cys sequence. The sequence of MTI1 contains 16 cysteinyl residues, 14 of which are in Cys-X-Cys sequences. Fluorescence spectroscopy indicates the presence of Cu(1)-thiolate bonds in both proteins. The binding stoichiometries are 11-12 for MTI and 10 for MT-11. Under certain nutritional conditions, a truncated form of MT-I1 was also produced. Northern analysis of the total cellular RNA from copper-treated cells showed that both MT-I and MT-I1 genes are regulated by this metal ion in a concentration-dependent fashion. The concentrations of MT-I1 mRNA appeared to be higher than that of MT-I mRNA at all concentrations of copper sulfate tested. Both genes are inducible by silver but not by cadmium salts. Cadmium ions, however, are effective in reducing the control levels of both MT-I and MT-I1 mRNAs.

Southern blot analysis has identified several metallothionein gene sequences in a human pathogenic yeast Candida glabrata. Two of these genes encoding proteins designated MT-I and MT-I1 have been cloned and sequenced. No introns were found in either of the genes. The complete primary structure of MT-I1 was also determined by protein sequencing methods. As isolated, MT-I and MT-I1 consist of 62 and 51 amino acids, respectively. The only residues predicted from the nucleotide sequence but not present in the isolated protein are the amino-terminal methionines in each sequence. MT-I contains 18 cysteines, 14 of which are present as Cys-X-Cys motifs and two additional cysteines in a Cys-X-X-Cys sequence. The sequence of MT-I1 contains 16 cysteinyl residues, 14 of which are in Cys-X-Cys sequences. Fluorescence spectroscopy indicates the presence of Cu(1)-thiolate bonds in both proteins. The binding stoichiometries are 11-12 for MT-I and 10 for MT-11. Under certain nutritional conditions, a truncated form of MT-I1 was also produced. Northern analysis of the total cellular RNA from copper-treated cells showed that both MT-I and MT-I1 genes are regulated by this metal ion in a concentration-dependent fashion. The concentrations of MT-I1 mRNA appeared to be higher than that of MT-I mRNA at all concentrations of copper sulfate tested. Both genes are inducible by silver but not by cadmium salts. Cadmium ions, however, are effective in reducing the control levels of both MT-I and MT-I1 mRNAs.
Cells have evolved a variety of mechanisms to deal with elevated concentrations of essential as well as nonessential metal ions. Sequestration and reduced uptake and/or facilitated efflux of potentially toxic metal ions are among the most commonly used mechanisms. Prokaryotes generally limit the intracellular concentrations of the offending ions by reduced uptake or facilitated efflux (1-3). Eukaryotes detoxify heavy metal ions by sequestering the ions within stable complexes (4-8). Two families of molecules, one represented by cysteine-rich polypeptides called metallothioneins (4, 5) and the other by enzymatically synthesized glutathione or the related y-EC peptides of general structure (7-EC),G (6-8), are involved in the cellular sequestration of toxic metal ions.
* This work is supported by Grants ES03817 and ES00147 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

to the GenBankTM/EMBL Data Bank with accession number(s) 505133
The nucleotide sequence(s) reported in this paper has been submitted and 505134.
To whom correspondence should be addressed.
Animals use metallothionein (MT)' as the main metal ion sequestering molecule while plants use y-EC peptides for the same purpose. Previous studies have suggested that fungi synthesize either MT-like polypeptides or y-EC peptides but not both (7-10). For example, Saccharomyces cereuisiae and Neurospora crassa respond to copper toxicity by synthesizing MT-like polypeptides (9, 10) whereas y-EC peptides are the major metal-binding species in Schizosaccharomyces pombe (8). We recently demonstrated that the yeast Candidaglabrata (also known as Torulopsis glabrata) has the ability to synthesize both MTs and y-EC peptides in a metal-specific manner (11). C. glabrata synthesizes two MTs in response to copper salts (11). MTs constitute a multigene family in many species (4, 5). All animal species studied to date, with the exception of chicken, have two or more MT genes, the products of which show a high degree of sequence homology. Drosophila constitutes the lone example in which members of the MT gene family share little sequence homology (12, 13). The fungi N. crassa and S. cereuisiae each have only one MT gene.
As part of our efforts to elucidate the pathways of metal ion detoxification in C. glabrata, we have cloned and sequenced two metallothionein genes from this yeast. The present report demonstrates that C. glabrata MTs constitute a multigene family consisting of two distinct classes of genes with multiple isoforms of one class. The two cloned C. glabrata genes are intronless like most genes (including MT) from S. cereuisiae. There is little sequence homology between the two C. glabrata MTs, a situation previously encountered only in the Drosophila MT genes (12, 13). The C. glabrata MT genes, like S. cereuisiae MT gene (14), are regulated transcriptionally by copper and silver but not by cadmium ions.

RESULTS
Purification of C. glabrata Cu-metallothioneins-Ion-exchange chromatography of the extracts prepared from the cells grown in a complete synthetic medium containing 0.5 mM CuS04 showed the presence of two main copper-binding components (11). Selective pools of the ion-exchange fractions were chromatographed on Sephadex G-50 to obtain pure samples of two distinct MTs designated MT-I and MT-11. In contrast, cells cultured in YTD medium supplemented with The abbreviations used are: MT, metallothionein; TFA, trifluoroacetic acid SDS, sodium dodecyl sulfate; kb, kilobase(s); HPLC, high performance liquid chromatography. * Portions of this paper (including "Materials and Methods" and Figs. 1, 2, 4, 5, and 7) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.
CuSO, contained at least three copper-binding components resolved by ion-exchange chromatography (Fig. 1). After further fractionation of these components on Sephadex G-50, amino acid analysis revealed that peaks I and I11 in the ionexchange profile corresponded to MT-I and MT-11, respectively. The amino acid composition of peak I1 was substantially different from both MTs, although the composition was closer to MT-11. To ascertain the nature of this copperbinding protein, further fractionation of the apoprotein prepared by KCN treatment was carried out using reverse phase HPLC. Elution of these proteins with an acetonitrile gradient showed three peaks (Fig. 2). HPLC pools I and I11 were identified as MT-I and MT-11, respectively, by amino acid analysis. The amino acid composition (data not shown) and amino-terminal sequence (see below) of peak I1 showed that it was MT-11-truncated at the amino terminus.
Amino Acid Sequences of C. glabrata MTs-The sequence of the amino-terminal 16 residues of MT-I and 19 residues of MT-I1 was determined previously by Edman degradation (11). An internal peptide (peptide D in Fig. 2) of MT-I1 obtained by chymotrypsin cleavage was also sequenced previously (11). Since peptide D accounted for two of the three lysines present in the protein, it was clear that tryptic digestion of MT-I1 would produce four peptides one of which should overlap peptide D. HPLC purification of the tryptic digest of carboxymethylated MT-I1 resulted in the identification, among others, of a peptide (peptide C) which provided good overlap with peptide D. To complete the sequence of MT-I1 between the amino-terminal 19 residues (peptide A) and peptide C, carboxymethylated protein was digested with Staphylococcus aureus Vs protease. The choice of this enzyme was based on the observations that the positions of all the expected glutamic acid residues were known and that one of the expected Vs peptides should begin with Asn'" and overlap peptide C. Indeed, such a peptide (peptide B) was identified following Vs digestions of carboxymethylated MT-11. Microheterogeneity was observed in the sequence of this peptide inasmuch as significant amounts of glutamine were also found besides asparagine at position 22.
Alignment of the sequences of all the above peptides (A-D) indicated the presence of 51 amino acids, although a previous amino acid analysis had predicted MT-I1 to consist of 53 amino acids (11). The carboxymethylated protein was digested with three different carboxypeptidases, B, Y, and P, to determine if there were only 51 residues in the protein making Cys5' the carboxyl-terminal residue. Carboxypeptidases B and Y did not release any significant amounts of amino acids. Cysteine followed by lysine was released upon incubation of MT-I1 with carboxypeptidase P suggesting that Cyssl was at the carboxyl terminus. This result was confirmed by the sequence of a genomic clone of MT-I1 (see below).
A partial sequence of C. glubrata MT-I1 published previously (11) had assigned position 13 to alanine. Reinvestigation of the sequence of MT-I1 and MT-11' showed that serine is present at position 13 of MT-11. The sequence of a genomic clone of MT-I1 confirmed the above result (see below).
The amino acid composition (data not shown) and aminoterminal sequence of MT-11', the minor species produced in YTD medium cultures, showed that it was a truncated version of MT-11. The amino terminus of MT-11' begins with Gin7 of unprocessed MT-I1 (Fig. 3).
Metal-binding Characteristics of C. glabrata Metallothioneins-A typical feature of vertebrate MTs is the presence of repeating Cys-X-Cys sequences that make these proteins effective metal-binding molecules (4, 5). C. glubrata MT-I1 contains seven Cys-X-Cys sequences (where X is any other MT 11' amino acid) and thus is expected to show metal-binding characteristics similar to those of vertebrate MTs. Native samples of MT-I1 contained 9.9 * 1.0 (mean f S.E., n = 5) mol eq of copper. The stoichiometry of metal binding was also determined by reconstitution of the apoprotein with Cu(1). Titration of the apoprotein with increasing mole equivalents of Cu(1) resulted in increased absorbance at 250 nm until 8 mol eq were added (Fig. 4a). No further increase in absorbance occurred when up to 10 mol eq of Cu(1) were added. The absorbance rose sharply thereafter. Examination of the UV spectra of the reconstituted protein showed that the absorption shoulder near 255 nrn characteristic of Cu(1) binding to thiolate sulfur (8, 17, 22, 23) appeared on the addition of 1 mol eq of Cu(1) and increased in intensity until 9 mol eq of Cu(1) were added (data not shown). This characteristic shoulder was lost upon the titration of the protein with additional quantities of Cu(1) although a general increase in absorbance did occur (Fig. 4a). The loss of the characteristic absorption shoulder was suggested to be indicative of disruption of Cu(1). thiolate clusters (8, 17, 22). Thus, the equivalence point of titration of the apoprotein with Cu(1) may be defined as the minimum number of mole equivalents of Cu(1) required to disrupt the metal-thiolate clusters as manifested in the disappearance of the characteristic absorption shoulder.

30
Previous studies have shown that Cu(1). thiolate clusters luminesce upon irradiation with ultraviolet light (8, 17, 22, 23). These luminescence properties can be used to determine Cu(1) binding stoichiometry of peptides and polypeptides (17, 22). Titration of apo-MT-11 with increasing mole equivalents of Cu(1) resulted in the appearance of luminescence. The intensity of emission increased until 9 mol of eq of Cu(1) had been added (Fig. 46). Addition of higher mole equivalents of Cu(1) led to progressive decline in luminescence. Moreover, the Amax of emission was also shifted (data not shown). The decline in luminescence indicated the disruption of Cu(1). thiolate clusters (17,22), analogous to the loss of the characteristic absorption shoulder. These reconstitution experiments and stoichiometries of the native protein suggested that C. glabrata MT-I1 binds nearly 10 mol eq of Cu(1).
The Cu(1). thiolate complex is extremely stable. Proton displacement studies have shown the pH at which 50% of bound Cu(1) ions dissociate from rat and probably other mammalian MTs is 2.7 (23). The pH of half-dissociation of Cu(I) binding to y E C peptides and S. cereuisiae MT is 1.3 and 0.3, respectively (8, 22). The C. glabrata MT-I1 loses 50% of Cu(I), as determined by loss of absorption at 250 nm, a t pH 0.8 (Fig. 5a).
As noted above, Cu(1)-containing y-EC peptides and metallothioneins luminesce upon irradiation with ultraviolet light. The luminescence of some y-EC peptides and MTs increases at acidic pH (17,22). Similar observations are made with C. glubrutu MT I1 in that the maximal luminescence was recorded at pH 5.0 (Fig. 5b) compared to maximal emission at pH 3.0 in S. cereuisiue CuMT. We have previously proposed that increased luminescence at acidic pH reflects conformational changes leading to greater shielding of Cu(1) -thiolate clusters from the solvent (17, 22).
Minimal experiments were carried out on MT-I due to limited availability of this protein. Analysis of the native samples of MT-I suggested that the protein bound -11-12 mol eq of copper. Reconstitution of a sample of apo-MT-I with Cu(1) showed that protein could bind up to 12 mol eq of the metal before the disruption of Cu(1). thiolate clusters as judged by the loss of luminescence and characteristic absorption shoulder (data not shown). Interestingly, the quantum yield of luminescence of the apo-MT-I reconstituted with Cu(1) was considerably higher than that of the native molecule (11).
Cloning and Sequencing of the MT-I Gene-Sequence analysis of MT-I by Edman degradation was complicated by minimal quantities of pure MT-I. To complete the sequence of this isoform, a degenerate oligonucleotide mixture was synthesized for cloning of the MT-I. gene. Southern analysis of EcoRI-digested genomic DNA using the MT-I-specific probe showed that the MT-I gene was present as a -3.7-kb fragment (Fig. 6, left). Screening of a partial genomic library in X g t l O identified a positive clone. DNA from this clone was digested with EcoRI to obtain the cloned fragment and restriction mapping showed that MT-I gene was located in a -0.7-kb EcoRI/ApuI fragment. This fragment was cloned into the appropriate sites of the Bluescript plasmid. The complete sequence of the cloned MT-I fragment was obtained by sequencing of the plasmid in both directions using m13 and T3 primers (Fig. 7u). Additionally, sequence of the coding region of the protein was confirmed by sequencing a 0.38-kb Suu3All Suu3Al fragment obtained from the 0.7-kb EcoRI/ApuI fragment. The complete sequence of the 0.7-kb fragment is shown in Fig. 8. A consensus polyadenylation signal AATAAA is found 169 nucleotides from the stop codon. A comparison with the amino-terminal sequence determined previously (11) shows that the amino-terminal methionine predicted from the nucleotide sequence is not found in the isolated protein. This finding is consistent with the observation that amino-terminal methionine, when followed by alanine, is cleaved by an aminopeptidase (24). Although C. glubrutu MT-I shows little sequence homology with mammalian MTs, it has seven Cys-X-Cys sequences, the motifs typical of MTs from vertebrates (4,5). MT-I also has one Cys-X-X-Cys sequence.
Cloning and Sequencing of the MT-11 Gene-Although the primary structure of MT-I1 had been determined by protein sequencing methods, cloning of the MT-I1 gene was performed with a view to understand the regulation of MT synthesis. Southern hybridization of EcoRI-digested genomic DNA with an MT-11-mixed oligonucleotide probe showed three bands (Fig. 6, right). Screening of a partial genomic library in X g t l O containing 1.0-kb piece with the above MT-I1 probe identified a large number of positive clones. Restriction mapping of the cloned DNA showed that the MT-I1 gene was located in a 0.7-kb EcoRI/SmuI fragment. The complete sequence of this DNA was determined after cloning into the appropriate sites of mp18 and mp19 (Figs. 7b and 9).
The amino acid sequence of the C. glabrutu MT I1 deduced from the nucleotide sequence ( Fig. 9) was identical with that determined by protein chemical methods (Fig. 3)

TCC AAT
GAA TGC TCC TGC CAA ACT TGC AAG TGT CAA ACA TGC AAG TGC the amino-terminal methionine predicted from the nucleotide sequence was not present in the isolated protein.

A T A C A T T A C G T A T A A A A A A A T A T A A A A A T A T A A A A A T A T A T C A C T T polyadenylation signal
Effects of Metals on M T mRNA Prodwtion-Logarithmically growing C. glabrata were exposed to various metal salts to determine the influence of these metal salts on MT mRNA production. Northern analysis showed that copper and silver salts had a positive influence and cadmium salt had a negative influence on the synthesis of both MT-I and MT-I1 mRNA (Fig. 10). These results may be used to compare the relative amounts of the two mRNAs as the probes used were similar in size, base composition, and specific activity (25). It is noteworthy that MT-I1 mRNA was induced far more exten-sively than MT-I mRNA upon exposure of cells to copper sulfate. This result is consistent with the finding that coppertreated cells produced very small amounts of MT-I protein (11). It is of significance that both MT-I and MT-I1 mRNA declined below control levels when the cells were treated with cadmium sulfate.

DISCUSSION
We previously showed that C. glabrata responds to copper toxicity by synthesizing two distinct metallothioneins (11). Cloning of the genes encoding these proteins has confirmed these predictions. Furthermore, Southern analyses indicate that there may be multiple related MT-I1 genes. This result is consistent with the observed microheterogeneity in the purified MT-I1 protein sequence. Southern analyses using either an oligonucleotide directed toward MT-I sequence, or a cloned DNA fragment having the coding sequence of MT-I (data not shown) showed the presence of a single hybridizable band. This suggests that there is probably only one MT-I gene. It is noteworthy that the overall organization of the C. glabrata MT gene family is very similar to MT gene families seen in higher vertebrates as each family consists of two principle genes and only one of these is present in multiple forms.
Vertebrate MTs constitute a family of highly conserved proteins and the positions of cysteinyl residues involved in metal binding are invariant (4, 5). Isometallothioneins from the invertebrate Scylla serrata show significant homology to each other as well as to vertebrate MTs (4, 5). In contrast, the two MT genes in the invertebrate Drosophila encode proteins which show little sequence homology to each other (12, 13). C. glabrata MTs likewise exhibit limited sequence homology with each other as indicated by alignment of the two sequences (Fig. 11). Neither of the C. glabrata MTs shares any appreciable sequence similarity with vertebrate, invertebrate, or known fungal MTs. Despite this lack of sequence similarities, both C. glabrata MTs exhibit the typical metallothionein sequence motif Cys-X-Cys. The role played by these sequence motifs in the formation of metal clusters in MTs is well recognized (4, 5). Thus C. glabrata MTs are structurally analogous, although not homologous, to other well characterized MTs.
None of the animal MTs studied to date contains tyrosine or other aromatic amino acids. Cyanobacterium MT (26) and C. glabrata MT-I1 are the only MTs known to contain tyrosine, the former contains two adjacent tyrosines and the latter contains a single tyrosine. Histidine is rarely present in MTs. Chicken (27), S. cereuisiae (17), and cyanobacterium (26) were the only MTs known to contain this amino acid. Both C. glabrata MT-I and MT-I1 contain histidine; MT-I has three, whereas MT-I1 has a single residue of this amino acid. A significant feature of the primary structures of C. glabrata MTs, not seen in any other MT, is the presence of internal sequence repeats. The pentapeptide Gln-Thr-Cys-Lys-Cys is repeated twice in MT-11. MT-I has two sequence repeats: the pentapeptide Cys-X-Cys-Pro-Asn and the octapeptide sequence Cys-Gly-Asp-Lys-Cys-Glu-Cys-Lys (Fig. 8). C. glubrata produces a processed MT-I1 designated MT-11' with Gln7 of the full-length protein as the amino terminus. The possibility was considered that MT-11' was a product of a unique gene coding for a sequence Met-Gln.. . . However, Southern analysis using oligonucleotide probes directed towards the peptide sequences Met-Pro-Glu-Gln-Val-Asn-Cys (amino-terminal sequence of  or Gln-Tyr-Asp-Cys-His-Cys-Ser (amino-terminal sequence of MT-11') showed the same bands with either of the probes suggesting that MT-I1 and MT-11' were product of the same gene. At present, it appears likely that MT-11' is produced by proteolytic cleavage of MT-I1 at the junction of Cys' and Gln7. Similar posttranslational processing has been observed in S. cereuisiae MT (16). The fact that the peptide cleaved from C. glabrata MT-I1 contains a cysteine (Cys6) may indicate lack of involvement of this particular residue in the formation of metal clusters.
The metal composition of vertebrate MTs depends on factors such as age, sex, nutritional conditions, and metal exposure (29). A variety of metals ions including cadmium, zinc, copper, and mercury have been found associated with animal MTs (4, 5). In contrast, fungal MTs have been found to contain only copper (10, 16, 30), and we also found copper only in C. glabrata MTs. Both native and reconstituted samples of C. glabrata MT-I1 show luminescence characteristic of Cu(1). thiolate clusters (8,17,22,23) indicating that the metal is bound to the protein in its monovalent state. This finding is consistent with previous studies on all copper-containing metallothioneins (4, 5).
The availability of only very limited quantities of MT-I precluded a detailed investigation of metal-binding stoichiometry of this protein. Preliminary experiments showed that the nature of copper binding in the native molecule is considerably different from that of the apoprotein reconstituted with Cu(1). The quantum yield of luminescence of reconstituted protein is much higher than that of the native MT. It has been noted previously that the quantum yield of luminescence of the native MT-I1 is considerably higher than that of the native MT-I (11). The differences in the absorption spectra of these two MTs were also recorded (11). The relatively low luminescence of the native MT-I may be due either to lack of shielding of Cu(1). thiolate cluster from the solvent (31) or because not all of the copper ions are present as Cu(1).
Mammalian MT genes are regulated by a variety of metals as well as other factors (5). The fungal MT genes respond to copper and in some cases to silver salts (4,5) but other metals do not have regulatory effects. The regulation of C. glubrata MT genes appears to be similar. The activation of the S. cereuisiue MT gene occurs via a transcriptional activation protein which binds regulatory DNA sequences only when monovalent copper is bound to it (14, 32). It has been suggested that the binding of Cu(1) to the transcription factor causes conformational changes in the molecule enabling it bind the upstream activating sequences. Similar mechanisms of MT gene regulation in C. glabrata cells are conceivable. The treatment of C. glabrata cells with cadmium salts decreased the levels of both MT-I and MT-I1 mRNA, an observation which could be explained by binding of cadmium to cysteinyl thiols in the putative transcription factor(s) which normally bind Cu(1). This could lead to conformational changes in the molecule making it unsuitable for interaction with upstream activating sequences. Current investigations in our laboratory are focused on the delineation of upstream activating sequences that may be present in the C. glabrata MT-I1 gene and the factor(s) that activate these sequences.    -1 (a) a d UT-I1 [ b ) clones. ~rlmert used for each sequence run are indicated.