Planta

, Volume 218, Issue 2, pp 300–308

Localization and functional characterization of metal-binding sites in phytochelatin synthases

Authors

  • Thomas Maier
    • Leibniz-Institut für Pflanzenbiochemie
    • Bruker Saxonia Analytik GmbH
  • Chao Yu
    • Max-Planck Forschungsstelle "Enzymologie der Proteinfaltung"
  • Gerhard Küllertz
    • Max-Planck Forschungsstelle "Enzymologie der Proteinfaltung"
    • Leibniz-Institut für Pflanzenbiochemie
Original Article

DOI: 10.1007/s00425-003-1091-7

Cite this article as:
Maier, T., Yu, C., Küllertz, G. et al. Planta (2003) 218: 300. doi:10.1007/s00425-003-1091-7

Abstract

Metal-binding domains consisting of short, contiguous stretches of amino acids are found in many proteins mediating the transport, buffering, trafficking or detoxification of metal ions. Phytochelatin synthases are metal-activated enzymes that function in the detoxification of Cd2+ and other toxic metal and metalloid ions. In order to localize Cd2+-binding sites, peptide libraries of two diverse phytochelatin synthases were synthesized and incubated with 109Cd2+. Distinct binding sites and binding motifs could be localized based on the patterns of Cd2+-binding. The number of binding sites was consistent with previous findings for recombinant protein. Positions of binding sites appeared to be conserved even among diverse phytochelatin synthases. Mutant peptide analysis was used to assess the contribution of exemplary amino acids to binding. Several binding motifs contain cysteines or glutamates. For cysteines a strong correlation was found between binding activity and degree of conservation among known phytochelatin synthases. These findings indicate the suitability of peptide scanning for the identification of metal-binding sites. The functional role of several cysteines was investigated by expression of hemagglutinin-tagged phytochelatin synthases in phytochelatin synthase-deficient, Cd2+-hypersensitive Schizosaccharomyces pombe cells. The data are consistent with a model suggesting functionally essential metal-binding activation sites in the N-terminal catalytic part of phytochelatin synthases and additional binding sites at the C-terminus not essential for activity.

Keywords

CadmiumMetal bindingMetal tolerancePhytochelatinSchizosaccharomyces

Abbreviations

EMM

Edinburgh's minimal medium

GSH

glutathione

HA

hemagglutinin

PC

phytochelatin

PCS

phytochelatin synthase

Introduction

Binding sites for metal ions are found in a myriad of proteins. About 3% of all the predicted proteins in Caenorhabditis elegans, for instance, have been estimated to contain zinc-binding domains (Clarke and Berg 1998). Among them are predominantly hydrolytic enzymes and DNA-binding proteins. Thus, Zn plays a catalytic as well as a structural role. The availability of Cu and Fe in different oxidation states under physiological conditions is utilized for a large number of redox reactions. Many essential oxidases, peroxidases and oxygenases, for instance, depend on either Cu or Fe ions. Besides the binding of metal ions as cofactors there is binding to proteins involved in metal homeostasis. The majority of proteins and peptides that function either in the uptake, distribution or storage of essential metal ions or the detoxification of essential as well as non-essential metal ions possess metal-binding sites. The Cys-Xaa-Xaa-Cys and Cys-Cys motifs of various metallothioneins are well known (Kojima et al. 1999; Cobbett and Goldsbrough 2002). Putative metal-binding sites have also been identified in several metal transporters. Members of the ZIP (ZRT-, IRT-like Proteins) and the CDF (Cation Diffusion Facilitator) family show His-rich cytoplasmic loops between transmembrane domains (Paulsen and Saier 1997; Guerinot and Eide 1999). Most metal-pumping CPX-type ATPases such as RAN1 (Hirayama et al. 1999) display varying numbers of N-terminal metal-binding domains consisting of Cys-Xaa-Xaa-Cys motifs (Solioz and Vulpe 1996). These are also found at the surface of metallochaperones such as Atx1 and its plant homolog CCH that deliver copper ions to copper-transporting ATPases (Himelblau et al. 1998; Huffman and O'Halloran 2001).

An example of a metal-activated enzymatic activity that is part of a metal-detoxification pathway is the synthesis of phytochelatins (PCs). Phytochelatins are small, glutathione (GSH)-derived metal-binding peptides that are important for the detoxification of mainly cadmium and arsenic in plants, fungi and certain animals (Cobbett and Goldsbrough 2002). Their synthesis is triggered by exposure to metal ions and catalyzed by the constitutively expressed enzyme phytochelatin synthase (PCS). PCS activity was found to be strictly dependent on the presence of metal ions in the assay buffer (Grill et al. 1989). Various metal ions and metalloids can activate PCS with Cd2+ ions being the strongest inducers. Other inducing ions include Pb2+, Zn2+, Cu2+, Sb3+, Ag+, Hg2+ and As5+ (Zenk 1996). Addition of chelating agents such as EDTA or apo-PCs to the enzyme assay immediately terminated the reaction. It was proposed that this dependence on metal ions and their chelation by the products of the reaction, the PCs, results in a self-regulation of the enzyme (Löffler et al. 1989). The recent cloning of PCS genes from various organisms (Ha et al. 1999; Vatamaniuk et al. 1999; Clemens et al. 1999) made detailed studies with purified recombinant PCS proteins possible. FLAG-tagged AtPCS1 was indeed shown to bind Cd2+ ions with high affinity (KD=0.54±0.2 μM) and a stoichiometry of 7.09±0.94 mol Cd2+/mol protein (Vatamaniuk et al. 2000). However, different S-alkylated GSH derivatives were found to activate AtPCS1 in a metal-ion-independent manner and to serve as substrates, leading to the hypothesis that bis(glutathionato)cadmium (Cd·GS2) and other blocked thiols are sufficient for the activation of the enzyme and that metal ions play only a minor role in enhancing activity. These results were later questioned. Oven et al. (2002) observed only very small activity with S-methyl-GSH in the absence of metal ions and instead, activation of AtPCS1 by various metal-thiolates not accepted as substrates. They concluded that the exact mechanism of enzyme activation by free metal ions or metal-thiolate complexes remains to be determined.

In order to gain insight into the structural basis of PCS activation by metal ions we attempted to identify Cd2+-binding sites in PCS proteins. Because many known metal-binding sites in proteins, e.g. the octapeptide repeat in the prion protein PrPC (Lehmann 2002) or the Cys-rich motifs in metallothioneins and CPX-type-ATPases, are encoded by short contiguous stretches of amino acids we hypothesized that the use of peptide libraries representing PCS proteins could be utilized for the analysis of Cd2+-binding. Here we report on the localization of putative Cd2+-binding sites in plant and fungal PCS proteins and their functional characterization in a PCS-deficient S. pombe strain. Our findings are consistent with a model suggesting functionally essential metal-binding activation sites in the N-terminal catalytic part of PCS proteins and additional binding sites at the C-terminus not essential for activity (Cobbett 2000). Furthermore, the data provide evidence for the suitability of peptide scanning to identify and analyze those metal-binding domains in proteins that are indeed formed by consecutive amino acids.

Materials and methods

Schizosaccharomyces pombe strains and media

The S. pombe strain employed in this study was the previously described phytochelatin synthase-deficient Δpcs strain Sp27 (Clemens et al. 1999). Cells were cultivated in Edinburgh's minimal medium (EMM; Moreno et al. 1991) supplemented appropriately. Transformation of S. pombe was performed according to the protocol of Bähler et al. (1998).

Synthesis of peptide scans and Cd2+-binding assay

Arrays of immobilized peptides (13mers) comprising sequences of the S. pombe and wheat (Triticum aestivum L.) phytochelatin synthases, SpPCS and TaPCS1, respectively, with shift of two amino acids were prepared by automated spot synthesis (Frank 1992; Kramer et al. 1994; Kramer and Schneider-Mergener 1998) on Whatman 50 filter paper (Whatman, Maidstone, UK). Peptides were C-terminally attached to cellulose via a (β-Ala)-2 spacer. Before screening, the dry membranes were washed in methanol for 10 min and equilibrated for 3×20 min in Tris buffer (10 mM Tris–HCl, pH 7.8). For binding assays, the membranes were incubated in Tris buffer containing metal ions and chelators together with 37 kBq 109Cd2+ for 30 min at 20 °C with gentle shaking. Then the membranes were washed three times with Tris buffer containing 100 μM CaCl2 and autoradiographed on Biomax MS film (Kodak, Rochester, USA) for 1–3 days. The membranes could be regenerated for re-use by stripping twice with buffer containing 100 μM EDTA and re-equilibration. Quantification of signal intensity was performed by densitometric analysis of the autoradiogram using a GS-700 Imaging Densitometer (Bio-Rad, Hercules, CA, USA).

Expression of epitope-tagged protein and site-directed mutagenesis

The S. pombe SpPCS and the Arabidopsis thaliana AtPCS1 coding sequences were amplified with Pfu polymerase using primers carrying an XhoI and a BglII site. The resulting fragments were ligated into pSGP72 (kindly provided by Dr. Susan Forsburg, Salk Institute, La Jolla, USA) to add a triple hemagglutinin (HA) tag to the C-terminus of the proteins. Site-directed mutagenesis was performed by PCR using Pfu polymerase. Sequences were confirmed by automatic sequencing.

Metal toxicity tests and PC assay

To test metal sensitivity, cells grown to log phase were diluted to an OD600 of 0.1 and incubated in the presence of different Cd2+ concentrations. After 24 h cell density was measured. Alternatively, a series of dilutions of precultures was spotted onto EMM plates containing different Cd2+ concentrations. For metal-sensitivity assays, 20 μM thiamine was added to suppress the strong constitutive nmt1 promoter in pSGP72 to ensure low expression levels. This was done to minimize the contribution of Cd2+-binding by PCS proteins to Cd2+ detoxification. Phytochelatins were analyzed essentially as described by Sneller et al. (2000). S. pombe cells were lyophyllized and extracted with 0.1% trifluoroacetic acid. Following centrifugation the supernatant was derivatized with monobromobimane at 45 °C in the dark. Extracts were analyzed by HPLC on a C18 column (5 μM, 150 mm) using an acetonitrile gradient. SH-containing compounds were detected fluorimetrically (excitation wavelength 380 nm, emission wavelength 480 nm).

Miscellaneous procedures

Standard methods for the manipulation of DNA and for PCR as well as for SDS–PAGE and Western analysis were used (Ausubel 1987). HA-tagged proteins were detected with HA-monoclonal antibody (Covance, Philadelphia, USA). Protein sequences were aligned using Clustal W (Thompson et al. 1994).

Results

Identification of Cd2+-binding sites in peptide scans of SpPCS and TaPCS1

Two diverse phytochelatin synthases were chosen for the analysis of possible Cd2+-binding sites by peptide scanning. TaPCS1 and SpPCS are only about 26% identical at the amino acid level and homology is largely confined to the N-terminal halves of the proteins. A total of 202 13mers, overlapping by 11 amino acids, were synthesized on a filter to represent SpPCS. For TaPCS1 a library of 245 peptides was synthesized accordingly. Incubation of the SpPCS and TaPCS1 peptide scans with 109Cd2+ revealed several contiguous stretches of peptides with binding activity (Fig. 1a, c). The SpPCS scan was synthesized twice on the same filter in order to assess reproducibility. Binding patterns were identical for both scans, showing that the signals were consistent throughout the analysis. Also, several rounds of re-incubation of scans following stripping and washing yielded identical results (data not shown). The length of the Cd2+-binding peptide stretches varied between three and seven peptides. Quantification of signal strength revealed distinct peaks indicating the presence of six major binding sites for SpPCS and seven major binding sites for TaPCS1 (Fig. 1b, d). The core peptides showing the highest binding activity within each stretch of peptides are highlighted in an alignment of known PCS proteins (Fig. 2). In spite of the diversity of the amino acid sequences the positions of Cd2+-binding peptides appear to be well conserved.
Fig. 1a–d.

Cd2+ binding assays with peptide scans representing Schizosaccharomyces pombe SpPCS and Triticum aestivum TaPCS1. SpPCS (a) and TaPCS1 (c) peptide scans were synthesized on filters as 13mers overlapping in 11 amino acids. Incubation with 109Cd2+ (37 kBq in 10 μM CdCl2) and subsequent autoradiography revealed several distinct stretches of peptides showing binding activity. b, d Quantification of signal strength for the SpPCS and TaPCS1 scans, respectively. The six different putative Cd2+-binding sites in SpPCS and the seven different putative Cd2+-binding sites in TaPCS1 are labelled with roman numerals. Several repetitions of the Cd2+-binding assays following stripping and re-equilibration yielded virtually identical patterns

Fig. 2.

Alignment of PCS amino acid sequences highlighting the position of putative Cd2+-binding sites. The amino acid sequences of TaPCS1 (from T. aestivum), AtPCS1, AtPCS2 (from Arabidopsis thaliana), BjPCS (from Brassica juncea, Acc. No. CAC37692), GmhPCS (Glycine max homo-PCS, Oven et al. 2002), AyPCS (from Athyrium yokoscense, Acc. No. BAB64932), CePCS (from Caenorhabditis elegans) and SpPCS (from S. pombe) were aligned using Clustal W. Boxes highlight the sequences of peptides showing the highest signal strength within a contiguous stretch of peptides with Cd2+-binding activity

Not only cysteines and not all cysteines bind Cd2+

The binding activity of consecutive peptides suggested that particular amino acid sequences shared by the respective sets of neighboring peptides accounted for most of the Cd2+-binding. Depending on the size of these sets between three and seven peptides, major binding motifs of lengths between eight and one amino acid, respectively, could be identified. Figure 3 shows the sequences of the core peptides, highlighting the amino acids which, based on the overlap between the neighboring binding peptides, are the most important ones for the binding activity of each Cd2+-binding site. Most of the core peptides contain cysteine residues. Cys-Cys pairs, for instance, are shared by all peptides of binding sites II and VI in SpPCS as well as binding sites II, VI and VII in TaPCS1. Single cysteines are found in the overlaps of SpPCS binding site I and TaPCS1 binding site I.
Fig. 3.

Core peptide sequences and major binding motifs of SpPCS and TaPCS1 putative Cd2+-binding sites. The sequences of peptides with highest Cd2+-binding activity within the binding sites of SpPCS and TaPCS1 are shown. Highlighted and underlined are the amino acids that represent the overlap between all the Cd2+-binding peptides within each putative binding site. Numbering of binding sites is shown on the left. The right column shows the position of the highlighted amino acids in the respective proteins

However, not all cysteines present in the two amino acid sequences showed substantial Cd2+-binding activity. Examples for non-binding cysteines in SpPCS include Cys-51, Cys-97, Cys-157 and Cys-186. None of the Cys-51-containing peptides 20–26 showed binding (Fig. 1a). The same applied to peptides 88–93 which contain Cys-186. Cys-97 is present in peptides 43–49. The binding activity of peptides 43–46, however, correlated with the presence of Cys-91. Peptides containing only Cys-97 were not 109Cd2+-labelled. Similarly, only peptides containing Cys-157 and Asp-164/Glu-165 displayed binding. Cys-157 alone (as in peptides 73–76) did not contribute. In TaPCS1, cysteine residues 231 (present in peptides 110–116), 243 (in peptides 116–122), 280 (in peptides 135–140), 416 (in peptides 203–208), 488 (in peptides 239–245) and 494 (in peptides 242–245) were only weakly labelled by 109Cd2+ (Fig. 1c). Cys-138, which is part of binding site V in TaPCS1, did not display binding independent of the presence of Cys-144. Peptides 64 through 66 showed only background activity. Strongly binding peptides 67–69 contain both cysteines.

Some of the binding activity appeared to be dependent on glutamate residues. In SpPCS, Glu-Glu is the common motif of binding site III, a single Glu is shared by the peptides of binding site IV and Glu-Val-Glu represents the overlap between the binding peptides of binding site V. In TaPCS1 the binding activity of binding site V correlates with the presence of the sequence Asp-Glu-Asp (peptides 165–167). In contrast, a histidine was found only once — in binding site III of TaPCS1 — among the sequence overlaps of Cd2+-binding peptides.

With few exceptions, those cysteine residues that are identified as being part of a binding site are also highly conserved among the PCS proteins from various organisms, whereas those not displaying binding activity are not conserved. Among the 18 cysteine residues in SpPCS, 5 are 100% conserved in all known PCS proteins: Cys-91, Cys-125, Cys-126, Cys-144, Cys-405. Ten cysteines are part of a binding site as indicated by the results of the peptide scanning. All five conserved cysteines belong to this group. The remaining five binding-site cysteines are either conserved in all but one PCS sequence (Cys-148, Cys-173) or do not contribute to Cd2+-binding based upon the analysis of all peptides containing the particular amino acid. The binding activity of Cys-97-peptides is dependent on the presence of Cys-91 in peptides 43–46 (see above). The activity of Cys-398 and Cys-399 peptides is largely dependent on Cys-405 in peptides 197–199. Analysis of the TaPCS1 scan results revealed a similar pattern. A total of 21 cysteines is present in this protein. All five 100% conserved cysteines (Cys-56, Cys-90, Cys-91, Cys-109, Cys-369) are part of a binding site according to the peptide scan results and within those sites part of the major binding motif (as highlighted in Fig. 3). Nine additional cysteines are part of one of the seven major binding sites. Seven of these are either conserved among all higher plant PCS sequences (Cys-138, Cys-144, Cys-352, Cys-370, Cys-374) or are conserved among all higher-plant PCS sequences with the exception of AtPCS2 (Cys-113, Cys-351). Two binding site cysteines are not conserved: Cys-331 and Cys-334. Six of the seven cysteines residues that are not part of a binding site are unique to TaPCS1. In conclusion, there is a clear correlation between the presence of a particular cysteine residue in a putative binding site and its degree of conservation. Conserved cysteines display binding activity, not conserved cysteines do not show binding activity. For other core residues of binding sites this correlation is less clear. The presence of charged amino acids is conserved rather than the presence of a particular amino acid: Asp-Glu-Asp of TaPCS1 binding site V aligns well with Glu-Val-Glu of SpPCS binding site V. Other PCS proteins have Tyr-Glu-Asp, Asp-Gly-Asp, Val-Asn-Glu or Tyr-Ser-Glu in these positions (Fig. 2).

Mutant peptide analysis and Cd2+ binding assays in the presence of chelators or competing metal ions

In order to evaluate the specificity of signals and the contribution of specific amino acids to binding, various mutant peptides were synthesized on a filter. Four reversed sequences of TaPCS1 peptides were tested and found to show substantially lower binding activities than their wild-type counterparts (Fig. 4a). As examples of binding sites containing one or two cysteines, series of mutant peptides for Cys-173 in SpPCS and Cys-369/Cys-370 in TaPCS1 were analyzed, respectively. SpPCS Cys-173 was present in peptides 81–87 of the original scan (Fig. 1a). Four of these peptides showed binding. Changing the single C in peptides containing Cys-173 to an A resulted in complete loss of binding activity (Fig. 4b). Exchange of either one of the cysteines in Cys-369/Cys-370 in peptides 180 and 181 for an alanine also caused loss of binding activity (Fig. 4c).
Fig. 4a–c.

Analysis of Cd2+-binding activity of mutant peptides. Mutant peptides were synthesized on a filter and incubated with 109Cd2+ (37 kBq in 10 μM CdCl2). The results for reversed sequences of four TaPCS1 peptides with different binding strength (a), of peptides containing SpPCS Cys-173 or an alanine in its place (b) and of peptides containing the TaPCS1 cysteine pair Cys-369/Cys-370 or an alanine in place of one of the cysteines (c) are shown

Under physiological pH conditions and in the presence of excess GSH, most of the Cd2+ is present as Cd·GS2 (Vatamaniuk et al. 2000). The Cd2+-binding affinity of AtPCS1 in the sub-micromolar range that was determined previously would be too low to result in any appreciable binding (Vatamaniuk et al. 2000). We tested Cd2+-binding activity in the presence of 1 mM GSH and found the binding activity to be lower overall but the binding pattern unaltered for both the TaPCS1 and the SpPCS scan (data not shown). Next we tested the influence of other divalent metal cations on the Cd2+ binding activity. The non-activating cation Co2+ did not affect binding even when present at a concentration of 100 μM. An excess of the activating cation Cu2+, however, selectively suppressed binding to some of the putative binding sites. Figure 5 shows the autoradiograph of the TaPCS1 scan incubated with 37 kBq 109Cd2+ in the presence of 50 μM Cu2+. Cd2+-binding to TaPCS1 binding sites I, III and VI was abolished. Cd2+-binding to binding sites IV and V was reduced.
Fig. 5.

Cd2+-binding in the presence of another activating cation. The TaPCS1 peptide scan was stripped, re-equilibrated, incubated with 37 kBq 109Cd2+ in the presence of 50 μM CuCl2, washed and autoradiographed. Numbering of peptides is shown on the left and on the right

Functional analysis of various cysteine and glutamate residues

In order to investigate the functional significance of various cysteines that according to the peptide scan results were either binding or non-binding, respective mutant versions of SpPCS were expressed with a C-terminal HA tag in PCS-deficient Cd2+-hypersensitive S. pombe cells (Clemens et al. 1999; Fig. 6a). Also, one glutamate residue belonging to binding site V was mutated. Phytochelatin formation and Cd2+ tolerance were assayed in the different strains in order to differentiate between active and non-active versions of PCS. SpPCS–HA restored Cd2+-activated synthesis of PC2 and PC3 in Δpcs cells (Fig. 6b) and growth in the presence of 5 μM Cd2+ (Fig. 6c). Cys-91 and the cysteine pair Cys-125/Cys-126 were identified as candidate binding motifs of binding sites I and II (Fig. 3). Changing them into alanines completely abolished PCS activity and Cd2+ tolerance. In contrast, the Cys pair Cys-398/Cys-399 of binding site VI was not essential for activity. The three cysteines Cys-144, Cys-148 and Cys-173 are part of binding site III but not identified as the major binding motif (Fig. 3). Cys-144 was found to be indispensable while the Cys148Ala and Cys173Ala mutants conferred PCS activity and partial rescue of the Cd2+ hypersensitivity of Δpcs cells. When non-binding cysteine Cys-186 was mutated, no difference from SpPCS–HA-expressing cells could be detected. A mutation directed at the major binding motif of binding site V, Glu-375-Val-376-Glu-377, did not cause loss of activity.
Fig. 6a–c.

Functional analysis of various cysteines and a glutamate in SpPCS. C-terminally HA-tagged SpPCS and different mutant versions of SpPCS–HA were expressed in a pcs-deficient S. pombe strain. a Protein expression was assayed by Western blotting and immunostaining with a monoclonal HA-antibody (Covance, Philadelphia) following protein extraction and separation of 10 μg total protein by SDS–PAGE. b Formation of PC2 (light grey) and PC3 (dark grey) in the pcs-deficient strain carrying either the empty vector or expressing a version of SpPCS–HA. Cells were grown in 10 μM Cd2+ for 24 h, harvested, freeze-dried and extracted with 0.1% trifluoroacetic acid. Thiols were labelled by monobromobimane, separated and quantified by HPLC. The result of a typical experiment is shown. The experiment was repeated twice, each time yielding similar relative differences. c Growth of the different S. pombe strains in the presence of 5 μM Cd2+ in EMM. OD600 was measured after 24 h and is shown here as percent of growth in medium without Cd2+. Error bars indicate SD, n=3

The analogous functional analysis of cysteine residues in a plant PCS protein had to be performed with AtPCS1 because HA-tagged TaPCS1 did not express well in S. pombe Δpcs. AtPCS1 and TaPCS1 are highly homologous throughout (Fig. 2) and most of the cysteines are conserved between these two proteins. Again, cysteines implicated in Cd2+-binding and cysteines not implicated were analyzed by expressing C-terminally HA-tagged AtPCS1 and the respective mutant versions in Δpcs cells (Fig. 7a). AtPCS1–HA conferred the ability to grow in the presence of an otherwise toxic Cd2+ concentration (Fig. 7b). Cells expressing the Cys56Ala and Cys90Ala variants, however, did not show any increase in Cd2+ tolerance over empty-vector controls. Cys-56 and Cys-90 in AtPCS1 correspond to Cys-56 and Cys-90 in TaPCS1, which were identified as belonging to the major binding motifs of binding sites I and II in TaPCS1 (Fig. 3). Loss of cysteines 109 or 113, which are part of binding site III in TaPCS1, abolished Cd2+ tolerance almost completely, while changing Cys-138, Cys-144 or Cys-358 to an alanine had no effect. The Cys366Ala and Cys404Ala mutants conferred a slightly lower degree of Cd2+ tolerance than AtPCS1–HA. Cys-138 and Cys-144 are part of binding site IV of TaPCS1. AtPCS1 Cys-358 corresponds to TaPCS1 Cys-369 in binding site VI and Cys-366 corresponds to TaPCS1 Cys-377, which lies outside binding site VI. AtPCS1 Cys-404 was included as an example of a non-conserved cysteine.
Fig. 7a, b.

Functional analysis of various cysteines in AtPCS1. C-terminally HA-tagged AtPCS1 and different mutant versions of AtPCS1–HA were expressed in a pcs-deficient S. pombe strain. a Protein expression was assayed by Western blotting and immunostaining with a monoclonal HA-antibody (Covance, Philadelphia, USA) following protein extraction and separation of 10 μg total protein by SDS–PAGE. b Growth of Δpcs S. pombe cells carrying either the empty vector or expressing a version of AtPCS1–HA was assayed by spotting serial dilutions of cells (OD600 is shown on the right) on EMM plates with or without Cd2+ (20 μM)

A comparison of mutant analysis data and PCS alignment showed, that the cysteines in the major binding motifs of binding sites I and II (Cys-91, Cys-125/Cys-126 in SpPCS; Cys-56, Cys-90/Cys-91 in TaPCS1) are 100% conserved among all known PCS proteins and essential for function. Similarly, mutations in the first cysteine of binding site III (Cys-144 in SpPCS; Cys-109 in TaPCS1), which is also found in all PCS sequences, resulted in an almost complete loss of activity. Non-binding cysteines such as Cys-186 in SpPCS are not conserved and are not required for PC synthesis.

Discussion

Localization and characterization of metal-binding sites can provide valuable knowledge about a large number of proteins because the binding of metal cations is an integral aspect of their function. Preferred ligands for soft and borderline ions such as Cu2+, Zn2+ and Cd2+ are thiolates and amines (Frausto da Silva and Williams 2001). Consequently, known metal-binding sites in most cases contain cysteine or histidine residues. Many of these sites consist of contiguous short stretches of amino acids. Examples are the above-mentioned motifs found in metal transporters or metallochaperones. The systematic analysis of peptide libraries might therefore represent a means of identifying and characterizing at least this particular type of metal-binding site in proteins. Several lines of evidence obtained in the study reported here on Cd2+-binding to PCS proteins suggest that peptide scanning could indeed be utilized to that end. 109Cd2+-binding assays reproducibly yielded patterns of mostly contiguous peptide stretches and pronounced peaks in signal strength (Fig. 1). This indicated distinct Cd2+-binding sites and the existence of binding motifs within these sites that were identifiable through the overlaps between neighboring Cd2+-binding peptides. Interestingly, the position of putative binding sites appeared to be conserved between diverse PCS enzymes. Also, it is noteworthy, that the number of seven putative binding sites found for TaPCS1 corresponds well to the 7.09 mol Cd2+/mol protein stoichiometry of Cd2+-binding determined for purified AtPCS1–FLAG under identical buffer conditions (Vatamaniuk et al. 2000).

Sequence overlaps between neighboring peptides with binding activity were used to tentatively determine major binding motifs. Not surprisingly, many of these contain cysteines. However, only a subfraction of cysteines present displayed binding activity. This finding argues against non-specifc binding of Cd2+ to all thiolates. Moreover, the striking correlation between degree of conservation and binding activity in the peptide scans, which applied to nearly all cysteines in SpPCS and TaPCS1, strongly suggests functional significance of the Cd2+-binding determined for certain cysteines. Given the inherent restriction of peptide scanning to analyzing binding motifs consisting of contiguous amino acids, a contribution of apparently non-binding cysteines to binding sites formed by amino acid residues coming into vicinity only in the folded protein, can obviously not be ruled out.

The Cys-Cys pairs are known from metallothioneins as metal-binding motifs and they were identified as binding motifs in several PCS binding sites. Similarly, the Cys-Xaa-Xaa-Cys motif found in metallothioneins and at the N-terminus of CPX-ATPases was identified as a binding motif in TaPCS1. Cys-331 and Cys-334 are part of binding site V. A third cysteine-containing motif was Cys-Xaa-Xaa-Xaa-Cys, found in binding site III of SpPCS and TaPCS1. Besides allowing candidate binding sites to be assessed, the analysis of peptide libraries might lead to the localization of unsuspected binding sites. The major binding motifs of SpPCS binding sites III–V and TaPCS1 binding sites II, IV and V contain glutamates and other charged residues. While their sequences are diverse their positions appear to be conserved according to the aligned amino acid sequences. Future analysis will show what the functional role of these putative Cd2+-binding sites is. The synthesis of mutant peptides represents one way of elucidating the contribution of particular amino acids to metal-ion binding. The examples analyzed here demonstrated in one case the contribution of a single cysteine to binding and in another case the dependence of binding by a Cys-Cys pair on both cysteines.

Availability of "free" metal ions is extremely limited inside a cell (Rae et al. 1999; Outten and O'Halloran 2001). The affinities of binding sites for different metal ions therefore have to be determined to assess their potential physiological significance. Affinity data obtained with peptide scans, just like those obtained with other in vitro techniques, might be artefactual. Still, the Cd2+-binding that was detectable even in the presence of millimolar concentrations of GSH or a vast excess of the non-activating divalent cation Co2+ suggests that the affinity of binding sites in PCS proteins localized by peptide scanning could be sufficiently high in vivo. Furthermore, competition experiments with other activating metal ions such as Cu2+ allow the relative affinity of different putative binding sites to be indirectly assessed. Our results indicate that Cu2+ binds mainly to TaPCS1 binding sites I, III and VI. Binding assays with purified wild-type and mutant proteins will be required to further elucidate the affinities of the different putative binding sites. Characteristics of binding by some of the identified metal-binding peptides could be further assessed using electrospray ionization mass spectrometry.

With regard to metal activation of PC synthesis, two models for the activation of the enzyme were envisioned, a direct binding of Cd to the protein or a binding of the substrate GSH–Cd (Ha et al. 1999; Cobbett 2000). The conserved N-terminal half of the protein was hypothesized to represent the catalytic domain, whereas the variable C-terminus might serve as a sensor domain that initially binds either the metal or the metal–GSH chelate but is not functionally essential. For both scenarios it was suggested that Cd2+ would first interact with C-terminal sensor sites and subsequently with the actual activation sites residing in the N-terminal half. A third possibility was later introduced (Vatamaniuk et al. 2000): Cd·GS2 and other blocked thiols activate PCS and serve as substrates with no need for the binding of activating metal ions. However, independent experiments with purified recombinant AtPCS1 showed activation also by metal-thiolates that are not substrates of PCS (Oven et al. 2002), suggesting that a transfer of activating metal ions from thiolate complexes to activation sites might be necessary. Moreover, the recently described second enzymatic activity of AtPCS1, cleavage of GSH conjugates, was also found to be strictly dependent on activation by either Cd2+, Zn2+ or Cu ions (Beck et al. 2003). In accordance with our findings, metal binding occurred even in the presence of a vast excess of thiols. Beck et al. suggested that the apparent low affinity of AtPCS1–FLAG for Cd2+ (Vatamaniuk et al. 2000) is due to oxidation of the enzyme during purification.

Our data are in agreement with the notion that Cd2+-binding occurs in both the essential catalytic N-terminal half of PCS enzymes as well as the C-terminal "sensor" half. Qualitative functional analysis was carried out with two different PCS enzymes, SpPCS–HA replacing the native protein and AtPCS1–HA as a representative of higher-plant PCS proteins expressed in a heterologous system. Detection of the C-terminal tags showed correct expression of the different mutant versions. The results were consistent for most of the corresponding residues tested and showed that the first four of the 100% conserved cysteines in the N-terminal half of the proteins (Cys-91, Cys-125, Cys-126, Cys-144 in SpPCS; Cys-56, Cys-90, Cys-91, Cys-109 in TaPCS1/ AtPCS1) are essential for function. The data obtained with peptide scans suggested that these cysteines, which show identical amino acid spacing, are also part of major Cd2+-binding motifs of putative binding sites I, II and III, and are therefore candidates to represent the activation sites. Mutations directed at C-terminal cysteines with apparent Cd2+-binding activity, on the other hand, resulted in only minor losses of PCS-dependent Cd2+ tolerance. This is in agreement with the findings by Ha et al. (1999), that the cad1-5 mutant showed residual Cd2+ tolerance although a stop codon deletes most of the C-terminal half of AtPCS1. A quantitative assessment of the tested mutations will require PCS activity assays.

In conclusion, our data provide evidence that peptide scans represent one suitable way of identifying and studying those frequently occurring metal-binding domains that are formed by contiguous stretches of amino acids. Only a subfraction of the S-, N-, or O-containing potential ligands present on the filters was found to show 109Cd2+ binding, indicating selectivity. Interestingly, most of the binding cysteines are conserved whereas the non-binding cysteines are not conserved among known PCS proteins. Also, the localization of putative binding sites appears to be similar. Furthermore, the number of putative binding sites determined here matches the stoichiometry found for recombinant PCS protein.

Acknowledgements

Initial cysteine mutagenesis and analyses in yeast were conducted in the laboratory of Julian Schroeder (University of California San Diego) and were supported by NIEHS Superfund grant 1 P42 ES10337 to J.S., whom we thank for his support and comments on the manuscript. We thank Marina Häußler for expert technical assistance and Regina Weiss for sequencing. This work was supported by grants from the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 363) and the EU (QLK3-CT-2000-00479).

Copyright information

© Springer-Verlag 2003