Heavy metals are becoming an increasingly serious problem to the environment. Their toxicity and the fact that they tend to get diluted or dissolved and get distributed from out of waste deposits make them a major environmental hazard, whereas at the same time those same metals compose a precious commodity which would make it worthwhile to harvest these metals from the environment. Environmental contamination with metals occurs for example in waste, or in ground or process water stemming from mining, the metal industry, viscose industry, rubber industry, paper industry, potato industry, starch processing, yeast industry, and so on. In addition, mining for metals in itself is a process where metals are harvested from the environment, and all above methods entail processes in which metals are recovered.

Citrate is very abundant in nature and most bacteria have transport proteins in the cytoplasmic membrane that mediate the uptake of citrate. The carriers belong to the class of secondary transporters that use the free energy stored in transmembrane electrochemical gradients of ions to drive the uptake of the substrates (for a review, see 18). The citrate transporter CitH of Klebsiella pneumoniae is driven by the proton motive force (23) and the transporters CitS and CitC of K. pneumoniae and Salmonella serovars by both the proton motive force and sodium ion motive force (9,14,19,24).

Mechanistically these transporters catalyze coupled translocation of citrate and H+ and/or Na+ (symport). A special case are the citrate carriers of lactic acid bacteria that take up citrate by an electrogenic uniport mechanism or by exchange with lactate, a product of citrate metabolism (citrolactic fermentation) (16,17,20). These citrate

transporters are involved in secondary metabolic energy generation (12).

A number of structural genes coding for citrate trans- porters have been cloned and the primary sequences have been deduced from the base sequences. The proton dependent citrate carrier of K. pneumoniae CitH belongs to a large family of proteins in which also many sugar transporters are found (22). The Na+ dependent citrate carriers CitS of K. pneumoniae and CitC of S. serovars form together with the citrate carriers of lactic acid bacteria CitP's a distinct family termed the 2-hydroxy-carboxylate carriers (4,15,24).

The malate transporter of Lactococcus lactis MleP that is involved in malolactic fermentation is also a member of this family (1).

Citrate is a chelator that forms stable complexes with various metal ions or metalsalt ions such as but not limited to Zn, Mg, Be2+, Ba2+, Ca2+ Cu2+ C 2+ 2+ Pb2+, Cd2+, U022+, and Ni2+. The presence of metal ions results in inhibition of citrate transport activity by the transporters mentioned above (16,23,24) showing that the metal ion/citrate complex is not a substrate of these citrate transporters. On the other hand, other bacteria including Pseudomonas and Klebsiella spp and Bacillus subtilis are known to preferentially take up and degrade citrate in the metal ion bound complex (2,3,10,26).

These microorganisms have been implicated in the pre- vention of mobilization of toxic metal wastes by chelators like citrate. Degradation of the metal ion/citrate complex would render the metal ion in an insoluble, immobilized state (7). A complication is that the nature of the metal ion in the complex determines whether or not the complex is degraded. Studies with Pseudomonas fluorescens have shown that at least for a number of metal ions the lack of degra- dation was limited by the lack of transport of the complex

into the cells and not because of the toxicity of the metal ion. The transporter seemed to recognize only the bidentate metal ion/citrate complexes that leave the hydroxyl group of citrate free; and not tridentate complexes (10).

Bacterial or biological technology, using selected naturally occuring microorganisms that can help concentrate and accumulate heavy metals, is currently in development.

Sulphate reducing bacterial strains have been detected in nature which specifically convert sulphate into sulphite (which takes place under oxygen-free, anaerobic conditions).

The sulphite than reacts with dissolved metals, producing a metal precipitate which can be removed. Such processes can be carried out in reactors. Other biological processes for recovering metals make use of bacteria that reduce the heavy metal itself and create metal precipitates, however, the above described biological processes are basically performed in an oxygen-free environment under anaerobic conditions, conditions which are not easy to achieve when large masses of waste or material containing dilute concentrations of the sought-after metal need to be handled. The present invention provides bacterial or biological technology, using (chimeric) micro organisms or (chimeric) proteins provided by the invention, that are suited for recovering metals, from for example waste, and such, which can be performed under aerobic and/or anaerobic conditions, using a wide variety of sources of metal, and a wide variety of metals.

The present invention provides a nucleic acid and derivatives thereof encoding genes of a new family of se- condary transporter proteins. A first gene, termed CitM, was identified in Bacillus subtils. Functional expression in Escherichia coli showed this gene to encode a citrate transporter protein that preferentially transports a metal- citrate complex. As an example, the citrate carrier or citrate transporter protein of B. subtils transports citrate

in a complex with a wide variety of metal ions. Studies with membrane vesicles showed that the highest uptake rates were observed with Mn2+, intermediate rates with Zn2+, Mg2+, Be2+, Ba2+, Ca2+ and Cu2+ and the lowest rates with Co2+ and Ni2+. It can however be expected that other citrate carriers have other metal specificities. The invention now thus provides a gene that can be introduced in any bacterial or biological host cell to be used in bacterial or biological processing or recovery of metal. Such host cells can for example be selected from any of the bacterial strains from the genera Pseudomonas, Klebsiella, or Bacillus, but many other bacteria or other host cells can also be used. The invention provides such host cells or micro-organisms that are capable of taking up the complex of citrate and (divalent) metal ions in for example environmental technologies and mining.

One may for example select those strains that thrive well in the presence of metal or metalsalt ions of various nature.

A second gene, termed CitH, was identified in Bacillus subtilis, by searching of available databases for protein sequences resembling the CitM gene. The invention thus reveals a further citrate transporter.

CitM and CitH provided by the invention are two secondary transporters of Bacillus subtils that catalyse the uptake into the cells of the complex of citrate with a divalent metal ion (M2+/citrate) and free citrate, respectively. Proteins provided by the invention show a considerable homology and structure prediction. Analysis of the hydropathy profile of the amino acid sequences of CitM and CitH indicate that the proteins have very similar structures. The invention provides gene (s) encoding protein (s) of which their functional differences relating to varying specificity and selectivity towards metal-ions lay therefore within their structural differences. The structure comprises a bundle of 12 transmembrane-helices connected by hydrophilic loops that contact the water phase. The structure

comprises a repetitive helix-loop motif. Within this structure lay structural differences that enable the design of citrate transporters (by recombinant techniques) and micro-organisms comprising citrate transporters with varying specificity and selectivity for a wide variety of metal ions.

The invention provides genes that code for a new family of secondary transporters. Studying the topology of the proteins encoded by these genes reveals a protein topology consisting of 12 transmembrane segments interspaced by a total of 11 loops. Within these transmembrane fragments and loops topology requirements and active sites are present which, when modified via recombinant technology, will alter the metal specificity of the encoded citrate transporter protein.

Thus modified transporter proteins provided by the invention have either a narrower or broader metal ion specificity in the metal ion/citrate complex that is recognized by the carrier. The present invention encompasses the possibility of generating a nucleic acid sequence encoding a modified citrate transporter protein via the further application of recombinant DNA technology. Such modifications may be single or multiple mutations, substitutions, deletions or insertions or combinations thereof that can be achieved via any recombinant DNA technology methods known in the art.

The present invention makes it possible to modify citrate transporter proteins that can be used to specifically interact with a particular metal or metals or a salt thereof, and can be used in processes recovering metals. Such taylor made proteins can be used in vivo, as being present as the active component in bacteria used for the recovery of metals from industrial waste and the like, but can also be used in vitro as active component in for example artificial biomembranes that will be used for metal recovery.

The invention provides a recombinant nucleic acid comprising a nucleic acid sequence as shown in SEQ ID NO: 6

or functional fragments or functional derivatives thereof or a recombinant nucleic acid which is at least 65%, preferably at least 75% or 85%, homologous to SEQ ID NO: 6, said nucleic acid sequence can be derived from a micro organism capable of transporting citrate. Functional fragments or derivatives can for example comprise any fragment of the nucleic acid encoding one or more of the domains as described below. The invention also provides a recombinant nucleic acid consisting of a gene encoding a citrate transporter protein and having the nucleic acid sequence as shown in SEQ ID NO : 8 or functional fragments or functional derivatives thereof.

The invention provides a nucleic acid which encodes a citrate transporter protein which transports free citrate or a nucleic which encodes a citrate transporter protein which transports a metal-citrate complex. In addition, the invention provides a vector, such as a virus or a plasmid, comprising a nucleic acid according to the invention, and the invention provides a host cell comprising a vector or a nucleic acid according to the invention. The invention also provides a host cell or micro-organism expressing a protein encoded by a nucleic acid according to the invention.

Furthermore, the invention provides a (citrate transporter) protein obtainable by growing a host cell according to the invention. Such a host cell or protein provided by the invention find its use for example in a process provided by the invention for recovering metal.

A preferred embodiment of the invention is the use of a microorganism comprising a nucleotide sequence as shown in SEQ ID NO : 6 or SEQ ID NO : 8 or comprising a nucleotide se- quence which is at least 65% homologous to the sequence as shown in SEQ ID NO : 6 or SEQ ID NO : 8 for the industrial recovery of metal. Another preferred embodiment of the invention entails a chimeric protein or functional fragments or derivatives thereof, as provided by the invention. A

chimeric protein provided by the invention is a protein in which part of the amino acid sequence is replaced by the corresponding part of an other, related or homologous protein. By constructing chimeras of for example CitM and CitH it is possible to locate and design a specific substrate binding site in the primary sequence. In this way, transporters are constructed with potentially different substrate specificity with respect to the metal ion in the M2+/citrate complex. It is possible to construct chimeras with a broader substrate specificity as well as chimeras with affinity for just one or a limited number of divalent metal ions.

In the structure of the transporters provided by the invention 25 domains (13 loops and 12 transmembrane segments) are discerned. Starting at the N-terminus, domain 1 is the N- terminal loop, domain 2 is the first transmembrane segment, domain 3 the second loop, domain 4 is the second transmembrane segment, etc. Such a chimera provided by the invention is encoded by a nucleic acid according to the invention wherein a nucleic acid sequence corresponding to the sequence as shown in figure 1 from position 7 to 71, or from 72 to 77, or from 78 to 144, or from 145 to 182, or from 183 to 243, or from 244 to 272, or from 273 to 338, or from 339 to 398, or from 399 to 458, or from 459 to 536, or from 537 to 596, or from 597 to 704, or from 705 to 773, or from 774 to 848, or from 849 to 923, or from 924 to 965, or from 966 to 1037, or from 1038 to 1055, or from 1056 to 1109, or from 1110 to 1148, or from 1149 to 1202, or from 1203 to 1228, or from 1229 to 1316, has been modified.

Such a nucleic provided by the invention is encoding a chimeric citrate transporter protein capable of binding to a metal ion. Several prererred types of chimeras and fragments thereof (and types of recombinant or chimeric micro- organisms) are provided by the invention. A single-fusion- site chimera provided by the invention comprises a N-terminal

fragment of one transporter and the C-terminal complement of the other transporter. In a double-fusion-site chimera provided by the invention, an internal fragment of one transporter is replaced by a corresponding fragment of the other transporter. And a multiple-fusion-site chimera provided by the invention represents a mosaic transporter built of many different, at least two, fragments from each of the two original transporters.

The invention provides a nucleic acid according to the invention wherein said chimeric protein is selected from any of the group of single-fusion-site, double-fusion-site and multiple-fusion-site chimeric proteins. The invention also provides a chimeric protein, such as a single-fusion-site chimeric protein which is created consisting of a N-terminal continuous string of domains of CitM followed by complementary domains of CitH. For instance, a chimeric protein consisting of the N-terminal 4 domains of CitM is followed by any number of C-terminal domains of CitH. A similar set of chimeras is provided in which domains of CitH preceed domains of CitM. A double-fusion-site chimeric protein is created in which one domain of CitM is replaced with the corresponding domain of CitH. For example, a chimeric protein in which domain 5 of CitM is replaced by domain 5 of CitH. A similar set chimeras exists in which one or more domains of CitH are replaced by (a) corresponding domain (s) of CitM. Single and double fusion site chimera are used as the starting point to create a multiple-fusion-site chimeric protein provided by the invention.

The invention furthermore provides a site directed mutant of nucleic acid according to the invention wherein at least one amino acid residue of said citrate transporter at or around original position K20, D76, D94, K97, R136, N152, K250, N285, Q323, E358, H381, or H391, has been modified. The invention provides such a site-directed mutant of CitM and CitH (chimera) wherein the different substrate specificity's

(free citrate and the M2+/citrate complex) of the two proteins are detailed in the structure.

The invention provides the modulation of the substrate specificity at a subtle level, i. e. different affinities for Citrate complexes with different metal ions. The required changes in the structure are located at the single amino acid level, as identified at or around position K20, D76, D94, K97, R136, N152, K250, N285, Q323, E358, H381, or H391. The transmembrane-helices in the structure consist of hydrophobic amino acids to cope with the hydrophobic environment in the membrane. The presence of the above identified charged or polar residues in the transmembrane segments is energetically unfavourable and, therefore, if present, involved in catalysis. The invention provides chimeras wherein charged and polar residues in the hydrophobic segments of CitM are replaced by naturally occurring amino acids. In addition, the invention provides chimeras wherein non-charged residues in the immediate vicinity (the residues that immediately precede and follow the above residues) of these charged and polar residues are mutated to the other residues to change the micro environment of the structure in the substrate binding site. The mutants provided by the invention differ in their substrate specificity for different divalent metal ions.

The invention provides strains and cultures of (chimeric) micro-organisms that accumulate divalent metal ions to high levels, for example bacteria that need several divalent metal ions for growth and possess specific uptake systems for these metals. The invention provides a process for recovering metal comprising using a host cell or micro-organism or (chimeric) protein according to the invention. CitM is a transporter that not only allows the cell to take up the carbon and energy source citrate but at the same time allows it to take up metal ions. The fate of the citrate is degradation, while the (divalent) metal ion accumulates in the cell. The invention provides chimeric micro-organisms wherein specific transporters, which reverse the specific chimera action and

efflux the metal ions again from the cells, are deleted.

The invention further provides use of a (chimeric) microorganism comprising a nucleotide sequence as shown in SEQ ID NO : 6 or SEQ ID NO : 8 or comprising a nucleotide se- quence which is at least 65% homologous to sequence as shown in SEQ ID NO : 6 or SEQ ID NO : 8 for the industrial recovery of metal. Furthermore, the invention provides using micro- organisms capable of taking up the complex of citrate and divalent metal ions in environmental technologies and mining, wherein the accumulation level of the metal ions in the cells is as high as possible. The invention provides use of knock- out micro-organisms in which the specific divalent metal ion transporters are deleted and that rely completely on CitM, or chimeric CitM, for their requirements for divalent metal ions. Such cells accumulate divalent metal ions up to high levels.

The invention also provides methods to select microor- ganisms that comprise said genes of a new family of secondary transporter proteins. Such selection methods can for example be based on a wide array of nucleic acid amplification techniques that is now available to the art. Such microorganisms can than advantagously be used for the in- dustrial recovery of metal.

Experimental part MATERIALS AND METHODS Bacterial strains and growth conditions.

E. coli strain JM101 harboring plasmid pWSKcitM coding for the divalent cation dependent citrate carrier of B. subtilis and strain JM109 (DE3) harboring plasmid pWSKcitH coding for the proton dependent citrate carrier of B. subtilis were grown in 6 1 flasks containing 1 1 medium at 37 °C and under

vigorous shaking. JM101 was grown in LB medium, JM109 (DE3) in LB medium or minimal medium containing citrate as the sole carbon source. Carbenicillin was added at a concentration of 100 g/ml and isopropyl-ß-D-thiogalactopyranoside (IPTG) was added when appropriate. The cells were harvested in the late exponential growth phase and used immediately for uptake experiment or the preparation of membrane vesicles. B. subtils strain 6GM was grown at 37 °C with vigorous aeration in medium containing 0.8% trypton (Difco), 0. 5% NaCl, 25 mM KC1 and 10 mM Na-citrate.

Cloning and sequencing of CitM.

Chromosomal and plasmid DNA isolations and all other genetic techniques were done using the standard protocols described by Sambrook et al. (21) or the manufacturers protocols.

Chromosomal DNA isolated from B. subtilis 6GM was partially digested with the restriction enzyme HindIII and the fragments were ligated in the multiple cloning site of vector pINIIIA (8) restricted with the same enzyme. Two fragments of 0.9 kb and 1.8 kb were restricted from clone pM5 by HindIII and ligated in the multiple cloning site of plasmid pBluescript II SK (Stratagene) yielding pSK0. 9 and pSKl. 8.

Sets of nested deletions starting at both ends of the inserts were constructed of pSK0. 9 and pSKl. 8 using the Erase-a-base System (Promega). The plasmids were digested with KpnI or SacI to create protected 3'overhangs and SalI or BamHI to allow digestion into the fragments. The subclones were sequenced on a Vistra 725 automated sequencer using Texas Red labeled forward and reverse primers of the pBluescript vector (Fig. 1). The sequencing reactions were performed using the Thermo Sequenase labelled primer cycle sequencing kit (Vistra Systems) with 7-deaza-dGTP according to the manufacturers protocol.

Southern blotting.

DNA probes were prepared by amplifying the regions on genomic DNA of B. subtilis that code for CitM and open reading frame N15CR by PCR. The CitM probe constitutes approximately the first 487 nucleotides of citM and the N15CR probe the last 463 nucleotides of N15CR. The probes were labeled with digoxygenin by including DIG-dUTP in the PCR reaction. The reaction contained in a total volume of 30 1, 3 pl superTaq buffer, 10 ng template DNA, 6 pl of a mixture of dATP, dCTP, dGTP (0.65 mM each) and DIG-dUTP 0.35 mM, 2.5 U superTaq and 1 Rl gelatine (5 mg/ml stock). The oligonucleotide primers were used at a concentration of 0.03 pg/pl. The forward and backward primers for the CitM probe were 5'-TTAAGGGGCCATGGA TGTGTTAGC-3'and 5'-CTCCCCAAGGAATCGTGTTC-3'and for the N15CR probe 5'-GGTGGATGCAATGGCGCATTC-3'and 5'-ATGAATCTTCTAGACACTCATAGGATCCTATTGATC-3'. The PCR reactions yielded fragments of the expected size and control reactions in the absence of DIG-UTP yielded identical fragments.

Restriction analysis confirmed that the correct regions had been amplified. The DIG-dUTP labeled products were purified over an QiaQuick column (Qiagen) and approximately 1 pg of labeled probe was used per hybridization.

DNA was electrophoresed in agarose gels and blotted on Zeta- Probe Blotting Membrane (Biorad) using the Vacuum Blotting Unit (LKB Bromma). Sample preparation and transfer was essentially performed as described by Sambrook et al. (21).

The blots were incubated with the labeled probes overnight at 65 °C. Subsequently, the membranes were washed twice with 300 mM NaCl, 30 mM Na-citrate, 0.1 % SDS for 5 min and at room temperature and twice with 15 mM NaCl, 1.5 mM Na-citrate, 0. SDS for 15 min at 65 °C. The membranes were used immediately for detection of hybridization.

Construction of the expression vectors. pWSKcitM. The citM gene was amplified by PCR using the Vent polymerase (Biolabs) from chromosomal DNA isolated from B. subtils 6GM. The forward primer 5'-TTAAGGGGCCAT GGATGTGTTAGC-3'contained the putative ribosomal binding site (indicated in italics) and the valine start codon (bold) and two mutations that result in a NcoI restriction site (CCATGG) in front of the start codon. The NcoI site was engineered for future purposes. The backward primer 5'- GTCATTACGCCTGAATTCCTCATACG-3'contained two mutations that create an EcoRI site (italics) immediately behind the TGA stop codon (bold). The EcoRI site at the 3'end of the PCR product was cut while the 5'end was left blunt. The fragment was ligated into plasmid pWSK29 (25) digested with SmaI and EcoRI yielding plasmid pWSKcitM. In pWSKcitM, the open reading frame coding for CitM is downstream of the lac promoter on the vector and the B. subtilis ribosomal binding site. pWSKcitH. Open reading frame N15CR coding for CitH was amplified from the same chromosomal DNA. The forward primer 5'-AAAAAAGCTTTTGAATAGGGGAGGTCATA CCATGGTTGCCATAC-3'contained three mutations resulting in a HindIII site in front of the ribosomal binding site and a NcoI site around the start codon. The construction of the NcoI site results in the Leu2Val mutation in the primary sequence. The backward primer 5'-ATGAATCTTCTAGACACTCATAG- GATCCTATTGATC-3'was complementary to sequences downstream of the stop codon. Four base changes resulted in the intro- duction of BamHI and XbaI sites (italic) in the PCR product.

The PCR product was digested with HindIII and BamHI and ligated into plasmid pWSK29 (27) digested with the same two enzymes. In the resulting vector pWSKcitH the citH gene is downstream of the T7 promoter and the B. subtils ribosomal

binding site. The base sequences of the inserts in pWSKcitM and pWSKcitH were verified by sequencing the sense strand.

Transport assays.

Whole cells. Cells of E. coli harboring plasmids pWSKcitM and pWSKcitH were grown in LB broth as described above and washed twice in 50 potassium phosphate pH 7. Uptake of [1,5- 19C] citrate was measured essentially as described by Lolkema et al. (14). The cells were resuspended in 95 1 of the same buffer with the indicated additions and incubated for 10 min at 37 °C. At time zero 5 pl [1, citrate was added to give a final concentration of 4.5 pM. Uptake was stopped by adding 2 ml of an ice cold 100 mM LiCl solution followed by immediate filtering over 0.45 m nitrocellulose filters. The filters were washed twice with the LiCl solution and immediately submerged in scintillation fluid to stop any further metabolic activity. The radioactivity retained on the filter was quantified in a liquid scintillation counter.

Membrane vesicles. Right-side-out (RSO) membrane vesicles were prepared of the E. coli cells harboring plasmids pWSKcitM and pWSKcitH by the osmotic lysis procedure as described by Kaback (11). E. coli JM101/pWSKcitM was grown in LB medium and JM109/pWSKcitH in minimal medium supplemented with 20 mM Na-citrate. The membranes were resuspended in 50 mM potassium 1, (Pipes) pH 6.5 at a protein concentration of 15 mg/ml and stored in aliquots in liquid nitrogen. The membrane vesicles were energized by the K-ascorbate/phenazine methosulfate (PMS) electron donor system. The membranes were diluted in 50 mM K-Pipes pH 6.5, and 10 mM K-ascorbate in a total volume of 100 1 and incubated for 5 min at 30 °C under a constant flow of water saturated air. PMS was added at a concentration of 100 UM and the proton motive force was allowed to develop for 1 min,

after which [1, citrate was added to a final concentration of 4.5 pM. The uptake was stopped and the samples were processed as described above.

Materials. [1, citrate (111 mCi/mmol) was obtained from Amersham Radiochemical Center. Mono potassium phosphate and potassium hydroxide with low Na+ content were obtained from Merck. All other chemicals were reagent grade and obtained from commercial sources.

Data bank submission. The CitM base sequence has been sub- mitted to the NCBI gene bank and will be accessible under number U62003.

RESULTS Cloning and sequencing of CitM. The Mg2+-dependent citrate carrier of B. subtilis that we will term CitM was cloned using conventional techniques. Chromosomal DNA of B. subtilis 6GM was partially digested with HindIII and the fragments where cloned in the expression vector pINIIIA (8).

Escherichia coli is an ideal host for the cloning of citrate carriers since it is not capable of taking up citrate but metabolizes it readily in the citric acid cycle. The expression vector containing the chromosomal fragments were transformed to E. coli JM 101 and grown on citrate indicator plates (Simmons agar). A number of blue colonies, indicative of citrate uptake and metabolism, were assayed for their ability to take up citrate in the presence and absence of 10 mM MgCl2. A clone (plasmid pM5) was selected that showed no uptake activity in the absence of Mg2+ and a high uptake in the presence of Mg2+ (not shown). Restriction analysis of pM5 revealed the presence of a 5.6 kb insert containing 3

additional HindIII sites. A 2.8 kbase EcoRI fragment of pM5 was subcloned into the EcoRI site of pINIIIA. Only one of the two orientations of the insert (plasmid pM6) showed Mg2+- dependent citrate uptake activity revealing the presence of the gene on the fragment and showing that the gene is expressed from the tandem promoter on pINIIIA. Further subcloning of pM6 resulted in the loss of the citrate utilizing phenotype. pM6 was used to subclone the 0.9 kb EcoRI/HindIII fragment (pSK0. 9) and the 1.8 kb HindIII-EcoRI fragment (pSKl. 8) in pBluescript. The two subclones were used to make sets of nested deletions to determine the nucleotide sequence of the gene. Both subclones were sequenced in both directions. Reconstruction of the base sequence of the insert on pM6 from the sequences of the two subclones revealed an open reading frame downstream of the promoter region on the vector with a length of 1302 bp and starting with a GTG codon (Fig 1). The length of the ORF conforms to the expected length of a gene coding for a bacterial secondary transporter. A putative ribosomal binding site is located upstream of the GTG start codon that shows extensive similarity to the 3'end of B. subtilis 16S rRNA. Neither a clear promoter sequence nor a terminator sequence were found upstream and downstream of the ORF, respectively. The complete base sequence has been deposited in the NCBI gene bank and will be available under accession number U63002.

A non redundant search of the available gene banks revealed an ORF of 1278 bp with 60% of base identity with the cloned gene. The ORF indicated as N15CR was detected in the bglS- katB intergenic region on the genome of B. subtilis 168. The ORF (Fig 2) starts with an ATG codon and is preceded by a ribosomal binding site. No clear promoter region could be detected upstream. The alignment (Fig 3) of the CitM sequence with the N15CR ORF shows that the ATG codon that lies in

between the ribosomal binding site and the putative CitM GTG start codon is unlikely to function as the initiator of translation (Fig. 2).

The presence of the citM gene and the N15CR ORF on the genome of B. subtils 6GM was confirmed by PCR and Southern hybridization. DNA probes were prepared by amplifying the first 487 bp of citM and the last 463 bp of N15CR by PCR using chromosomal DNA of B. subtils 6GM as the template. The probes were selected such that they contained no HindII restriction sites. The PCR resulted in distinct DNA fragments of the expected length. The nucleotide analog DIG-dUTP was incorporated into the fragments for the use as probes in Southern blotting. The two probes detected unique, but different fragments of B. subtilis 6GM genomic DNA restricted with HindIII. Both the fragments were of the expected length.

Plasmids pWSKcitM and pWSKcitH that contain the citM gene and ORF N15CR (see below) hybridized exclusively with the citM and N13CR probes, respectively, in spite of the high sequence identity between the two genes. The lack of cross-reaction reflects the high stringency of the hybridization and washing conditions. Under these conditions the two probes did not detect similar genes on the chromosome of E. coli and the thermophilic Bacillus species B. stearothermophilus.

Primary sequence analysis. Translation of the base sequences of the cloned ORF and the corresponding ORF from the B. subtils gene bank results in two proteins with approximately 60 % amino acid sequence identity and an additional 18 % of similar residues. The amino acid composition of the two proteins is typical for integral membrane proteins with an average hydrophobicity of 0.51 and 0.47 on the normalized scale of Kyte (5), respectively. The hydropathy profiles of the two sequences are remarkably similar. Significant

differences show up only in the region from position 125 to 145 and to a lesser extent in the region from position 310 to 330. In both regions, CitM is the more hydrophobic sequence.

The high similarity of the two sequences suggests a similar folding in the membrane. Secondary structure prediction (6) results in 12 membrane spanning, presumably a-helical, regions both for the CitM protein and the protein coded by the N15CR ORF. Assuming similar folds for the two proteins, merging of the two predictions results in 12 transmembrane segments, interspaced by 11 loops. The respective nucleic acid sequences corresponding to the 12 transmembrane segements and the 11 loops relate more or less to the sequence as shown in figure 1 from position 7 to 71, or from 72 to 77, or from 78 to 144, or from 145 to 182, or from 183 to 243, or from 244 to 272 ; or from 273 to 338, or from 339 to 398, or from 399 to 458, or from 459 to 536, or from 537 to 596, or from 597 to 704, or from 705 to 773, or from 774 to 848, or from 849 to 923, or from 924 to 965, or from 966 to 1037, or from 1038 to 1055, or from 1056 to 1109, or from 1110 to 1148, or from 1149 to 1202, or from 1203 to 1228, or from 1229 to 1316, respectively. It can be expected that within these transmembrane fragments and loops lay topology requirements and active sites which, when modified via recombinant technology, will alter the metal specificity of the encoded citrate transporter protein.

Substrate specificity. The citM gene and the N15CR ORF (CitH gene) were amplified by PCR using B. subtils chromosomal DNA as template. The citM fragment was cloned downstream of the lac promoter region on plasmid pWSK29 a low copy pBluescript derivative (25) yielding plasmid pSKcitM. The N15CR ORF was cloned behind the T7 promoter on the same plasmid yielding pWSKcitH. The sequence of the cloned PCR fragments was verified in one direction and found to be identical to the

base sequence of the original ORFs except for the second codon of citH which now codes Val instead of Leu (see the Methods section for details).

Plasmid pWSKcitM was transformed to E. coli JM101 and plasmid pWSKcitH to E. coli JM109 (DE3) a strain that contains a chromosomal copy of T7 polymerase and the cells were plated on Simmons agar indicator plates. Surprisingly, both plasmids conferred the citrate utilizing phenotype. Apparently, the N15CR ORF codes for a citrate transporter as well. We will term this transporter CitH. Figure 4 shows the uptake of citrate in cells harboring plasmids pWSKcitM (A) and pWSKcitH (B) in the presence of different concentrations of Mg2+.

Citrate uptake activity in cells expressing CitM is completely absent in the absence of Mg2+-ions. The uptake activity increases with increasing Mg2+ concentrations con- sistent with the Mg2+-citrate complex being the substrate of the carrier (2). In marked contrast, cells harbouring plasmid pWSKCitH expressing the citrate carrier coded by ORF N15CR (citH) readily take up citrate in the absence of Mg2+.

Increasing concentrations of Mg2+ in the assay buffer result in decreasing uptake rates. Apparently, the substrate of CitH is free citrate as is the case for the Na+ and H+-dependent citrate carriers of K. pneumoniae (23,24) and the membrane potential generating citrate carrier of L. mesenteroides (16).

Co-ion specificity. The involvement of Na+ ions in the uptake of citrate by CitM and CitH was investigated by measuring the uptake of citrate in E. coli strains JM101/pWSKcitM and JM109 (DE3)/pWSKcitH in the presence and absence of 10 mM NaCl. Prior to the experiments the cells were washed three times in large volumes of potassium phosphate pH 7 containing especially low amounts of Na+. The residual Na+ ion concentration was at most a few pM. Furthermore, the uptake

experiments were performed in plastic tubes to prevent Na+ contamination from glassware. For both transporters the uptake of citrate was not significantly different in the presence or absence of NaCl indicating that Na+ is not a co- ion for CitM nor CitH (not shown).

Studies with membrane vesicles prepared from B. subtils cells have demonstrated that the Mg2+ dependent citrate transporter CitM is a secondary transporter that is driven by the proton motive force (2). The high similarity between CitM and CitH suggests that the same is true for CitH. The energy coupling mechanism of both cloned transporters was investigated by preparing right-side-out membrane vesicles of E. coli cells expressing CitM or CitH. The membranes were energized by the ascorbate/PMS electron donor system. In the presence of a proton motive force, both transporters accumulated citrate indicating cotransport by CitM of the Mg2+/citrate complex and protons and by CitH of citrate and protons. In the presence of the K+ ionophore valinomycin which results in the dissipation of the membrane potential component of the proton motive force the uptake activity was slightly less. In the presence of nigericin, a K+/H+ antiporter, which dissipates the pH gradient across the membrane and results in a proton motive force that is solely composed of the membrane potential, significant uptake above background is still observed. It is concluded that both CitM and CitH are electrogenic transporters that translocate net positive charge into the cells. The complex between citrate and Mg2+ is monovalent anionic (MgCit-) and, therefore, CitM cotransports at least 2 protons per Mg2+/citrate complex.

Electrogenic transport by CitH indicates cotransport of at least 3 or 4 protons depending on translocation of Hcit2-or cit3-, respectively.

DISCUSSION The cloning and sequencing of the Mg2+ dependent citrate carrier of B. subtilis CitM let to the surprising finding of a second citrate carrier in B. subtils CitH for which the gene was deposited in the databanks as open reading frame N15CR. The two transporters share common properties at different levels: (i) the coding genes are 61. homologous and in the primary sequence alignment about 60% residues are identical, (ii) the transporters function as electrogenic proton symporters and (iii) the genes coding for the transporters are present on the chromosome of B. subtilis and the genes were not found in E. coli and B. stearo- thermophilus. The most striking difference between the two transporters is that CitM transports the Mg2+/citrate complex while CitH transports free citrate in the uncomplexed state and is hampered by the presence of Mg2. In B. subtilis, CitM is induced by citrate in the medium and the absence of citrate uptake by membrane vesicle in the presence of EDTA indicates that CitM is the only transporter induced under these conditions (2). The experiments with the cloned transporter in E. coli emphasize that CitM transports only citrate in the Mg2+ complexed form. Therefore, CitH is not induced under the same conditions in B. subtils. Open reading frame N15CR that codes for CitH lies in between the genes bglS and katB on the B. subtilis genome. It is coded in the opposite direction relative to these two genes. Gene bglS codes for lichanase, an exported enzyme, that hydrolyses mixed linked P-glucans (27) and katB codes for a catalase involved in sporulation (13). The citH gene is not preceded by a known promoter sequence and does not seem to be part of a operon structure suggesting that the gene is silent. On the other hand, the gene is preceded by a ribosomal binding site

and results in a functional transporter when expressed from a heterologous promoter. Either the gene has become silent only very recently on a evolutionary time scale or, more likely, the gene is expressed under special, unknown conditions using an unknown promoter sequence.

The high sequence identity of CitM and CitH suggests that the binding sites for the Mg2+/citrate complex and citrate are not very different. The two transporters may be very suitable for the construction of chimeric proteins to localize the substrate binding site in the primary sequence. The successful construction of active chimeras can be tested on citrate indicator plates and the Mg2+-dependency provides an easy way to discriminate between the activity of the two transporters. We constructed one chimera by making use of a conserved StuI restriction site in the two genes around position 490. The hybrid protein consisted of the N-terminal CitM fragment and the C-terminal CitH fragment. We are now in the process of constructing a series of chimeras by introducing unique restriction sites at selected sites in the citM and citH genes. Said chimeras (chimeric proteins and/or chimeric micro organisms) can be used to remove metals (specifically and/or selectively) from various sources containing said metals. This can be achieved in for example two steps. In a first step the metal is being chelated with citrate. In a second step, the metal ion-citrate complex is taken up by a micro organism provided by the invention, which uses the citrate as growth-substrate, thereby reducing the concentration of the metal in the environment and accumulating it in the bio mass.

REFERENCES 1. Bandell, M., Ansanay, V., Rachidi, N., Dequin, S. and Lolkema, J. S. 1996. The membrane potential generating citrate and malate transporters of lactic acid bacteria are homologous protein. The 2-hydroxy-carboxylate family of transporters. J. Biol. Chem., submitted 2. Bergsma, J. and Konings, W. N. 1983 The properties of citrate transport in membrane vesicles from Bacillus subti- lis. Eur. J. Biochem. 134: 151-156.

3. Brynhildsen, L. and Rosswall, T. 1989. Effects of cupper, magnesium and zinc on the decomposition of citrate by Klebsiella sp. Appl. Environm. Microbiol. 55: 1375-1379.

4. David, S., M. E. van der Rest, A. J. M. Driessen, G.

Simons, and W. de Vos 1990 Nucleotide sequence and expression in Escherichia coli of the Lactococcus lactis citrate permease gene. J. Bacteriol. 172: 5789-5794.

5. Eisenberg, D. 1984. Three-dimensional structure of mem- brane and surface proteins. Ann. Rev. Biochem. 53: 595-623.

6. Eisenberg, D., Schwarz, E., Komarony, M. and Wall, R. 1984 Analysis of membrane and surface protein sequences with the hydrophobic moment plot. J. Mol. Biol. 179: 125-142.

7. Francis, A. J., Dodge, C. J. and Gillow, J. B. 1992. Biode- gradation of metal citrate complexes and implications for toxic-metal mobility. Nature 356: 140-142.

8. Inouye, H. 1983. Experimental manipulation of gene ex- pression. Academic Press Inc., New York.

9. Ishiguro, N., H. Izawa, M. Shinagawa, T. Shimamoto, and T.

Tsuchiya 1992. Cloning and nucleotide sequence of the gene (citC) encoding a citrate carrier from several Salmonella serovars. J. Biol. Chem. 267: 9559-9564.

10. Joshi-Tope, G. and Francis, A. J. 1995. Mechanisms of

biodegradation of metal-citrate complexes by Pseudomonas fluorescens. J. Bacteriol. 177: 1989-1993.

11. Kaback, H. R. 1971. Bacterial membranes. Methods Enzymol.

22: 99-120.

12. Konings, W. N., Lolkema, J. S. and Poolman, B. 1995. The generation of metabolic energy by solute transport. Arch.

Microbiol. 164: 235-242.

13. Loewen, P. C. 1989. Genetic mapping of katB, a locus that affects catalase 2 levels in Bacillus subtilis. Can J Microbiol 35: 807-810.

14. Lolkema, J. S., H. Enequist, and M. E. van der Rest 1994.

Transport of citrate catalyzed by the sodium-dependent citrate carrier of Klebsiella pneumoniae is obligatory coupled to the transport of two sodium ions. Eur. J. Biochem.

220: 469-475.

15. Lolkema, J. S., Poolman B., and Konings, W. N. 1996.

Secondary transporters and metabolic energy generation in bacteria. In: Handbook of Biological Physics. Ed. Konings, W. N., Kaback, H. R. and Lolkema, J. S. Elsevier Science, Amsterdam.

16. Marty-Teysset, C., J. S. Lolkema, P. Schmitt, C. Divies, and W. N. Konings 1995. Membrane potential generating transport of citrate and malate catalyzed by CitP of Leuco- nostoc mesenteroides. J. Biol. Chem. 270: 25370-25376.

17. Marty-Teysset, C., Posthuma, C., Lolkema, J. S., Schmitt, P., Divies, C. and Konings, W. N. 1996. Proton motive force generation by citrolactic fermentation in Leuconostoc mesenteroides. J. Bacteriol. 178: 2178-2185.

18. Poolman, B., Konings. W. N. 1993. Secondary solute transport in bacteria. Biochim. Biophys. Acta 1183: 5-39.

19. Pos, K. M. and P. Dimroth 1996. Functional properties of the purified Na+-dependent citrate carrier of Klebsiella

pneumoniae: Evidence for asymmetric orientation of the carrier protein in proteoliposomes. Biochemistry 35: 1018-1026.

20. Ramos, A., B. Poolman, H. Santos, J. S. Lolkema, and W.

N. Konings 1994. Uniport of anionic citrate and proton consumption in. Leuconostoc oenos. J. Bacteriol.

176: 4899-4905.

21. Sambrook, J., Fritz, E. F. and Maniatis, T. 1989. Mole- cular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.

22. Van der Rest, M. E., E. Schwarz, D. Oesterhelt, and W. N.

Konings 1990. DNA sequence of a citrate carrier of Klebsiella pneumoniae. Eur. J. Biochem. 189: 401-407.

23. Van der Rest, M. E., T. Abee, D. Molenaar, and W. N.

Konings 1991. Mechanism and energetics of a citrate transport system of Klebsiella pneumoniae. Eur. J. Biochem. 195: 71-77.

24. Van der Rest, M. E., R. M. Siewe, T. Abee, E. Schwarz, D.

Oesterhelt, and W. N. Konings 1992. Nucleotide sequence and functional properties of a sodium-dependent citrate transport system from Klebsiella pneumoniae. J. Biol. Chem.

267: 8971-8976.

25. Wang R. F. and Kushner S. R. 1991. Construction of versa- tile low-copy-number vectors for cloning, sequencing and gene expression in Escherichia coli. Gene 100: 195-199.

26. Willeke, K., Gries, E-M. and Oehr, P. 1973. Coupled transport of citrate and magnesium in Bacillus subtilis. J.

Biol. Chem. 248: 807-814.

27. Wolf, M., Geczi, A., Simon, O. and Borriss, R. 1995.

Genes encoding xylan and ß-glucan hydrolysing enzymes in B. subtilis : characterization, mapping and construction of strains deficient in lichanase, cellulase and xylanase.

Microbiology 141: 281-290.

Figure legends Figure 1 Nucleotide sequence and corresponding amino acide sequence of the CitM gene.

Figure 2 Nucleotide sequence and corresponding amino acide sequence of the CitH gene.

Figure 3 Alignment of amino acid sequences of the CitM and CitH genes.

Figure 4 Mg2+-ion dependence of the uptake activity of CitM (A) and CitH (B). [1, citrate uptake by E. coli JM101/pSKcitM (A) and E. coli JM109 (DE3)/pSKcitH (B) was measured in 50 mM potassium phosphate pH 7 supplemented with 0 (-), 0.5 (+), 1 (V), 5 (A) and 10 (E) mM MgCl2. The cell protein concentrations were 0.6 (A) and 1.2 (B) mg/ml.