Res. Microbiol. 152 (2001) 311–321 2001 Éditions scientifiques et médicales Elsevier SAS. All rights reserved S0923-2508(01)01202-5/FLA
Molybdate transport ✩ William T. Selfa , Amy M. Grundenb , Adnan Hasonac , Keelnatham T. Shanmugamc∗ a NHLBI, NIH, Bethesda, MD 20892, USA b Department of Microbiology, North Carolina State University, Raleigh, NC 27695, USA
c Department of Microbiology and Cell Science, Box 110700, University of Florida, Gainesville, Florida 32611, USA
Received 30 October 2000; accepted 9 January 2001
Abstract – In both bacteria and archaea, molybdate is transported by an ABC-type transporter comprising three proteins, ModA (periplasmic binding protein), ModB (membrane protein) and ModC, the ATPase. The modABC operon expression is controlled by ModE-Mo. In the absence of the high-affinity molybdate transporter, molybdate is also transported by another ABC transporter which transports sulfate/thiosulfate as well as by a nonspecific anion transporter. Comparative analysis of the molybdate transport proteins in various bacteria and archaea is the focus of this review. 2001 Éditions scientifiques et médicales Elsevier SAS molybdate transport / genetics / biochemistry / bacteria / archaea
1. Introduction Molybdoenzymes catalyze a variety of oxidation/ reduction reactions in bacteria as well as in plants and animals [17, 25]. Nitrogenase and nitrate reductase play critical roles in the global nitrogen cycle. Other molybdoenzymes, such as formate dehydrogenase, nitrate reductase, dimethyl sulfoxide (DMSO) reductase and trimethylamine N-oxide (TMAO) reductase catalyze redox reactions allowing bacteria to grow under anaerobic conditions using formate as an electron donor and nitrate, DMSO and TMAO as electron acceptors. Molybdoenzymes are also essential for normal human development . Nonnitrogenfixing organisms produce a limited number of molybdoenzymes under a given physiological condition and thus require only trace amounts of molybdate in the medium. For the synthesis of molybdoenzymes, the cell needs to transport molybdate, activate it to an appropriate form and incorporate it into the organic part of the molybdenum cofactor, molybdopterin [11, 13, 51]. In many bacterial molybdoenzymes this Momolybdopterin is conjugated with a nucleotide. In hy-
✩ Florida Agricultural Experiment Station Journal Series No. R08044. ∗ Correspondence and reprints. E-mail address: [email protected]
perthermophiles, the Mo is usually replaced by W . In nature, the predominant form of Mo is molybdate oxyanion which is transported by cells as molybdate by an ABC-type transport system. Most of the early studies on molybdate transport and metabolism have been reviewed in detail [11, 48] and will not be a part of this current endeavor. The results from recent studies as well as the comparative information from the analysis of genome sequences will be the main focus of this review. Since the Escherichia coli molybdate transport system and its components are well-defined, emphasis will be on the E. coli system, and the molybdate transport systems in other bacteria will be compared to that of E. coli. 2. The molybdate transport system in E. coli In E. coli, molybdate is transported by an ABCtype transport system comprising three proteins, ModA, the periplasmic molybdate-binding protein, ModB, the integral membrane channel protein, and ModC, the energizer protein (figure 1). The three genes constitute a single operon [34, 64]. Diverging from this operon is another operon encoding ModE and ModF . The molybdate-bound form of the ModE protein regulates the expression of the modABC operon by binding to the operator region of the modABC operon. In addition to the components of
W.T. Self et al. / Res. Microbiol. 152 (2001) 311–321
Figure 1. Molybdate transport and regulation by ModE in E. coli. The high-affinity molybdate transporter is represented by the ModABC proteins. An ABC transporter for sulfate (CysUWA and SBP, or sulfate binding protein) as well as an anion transporter also contribute to molybdate transport. ModE-Mo serves as a repressor of the modABC operon and enhances the expression of hyc and narXL operons. See text for details.
the ABC transporter and the control protein, an additional ORF (modF) is also found downstream of the modE gene. The role of the ModF protein is not known. Deletion of modF has no detectable effect on the Mo metabolism of E. coli. However, ModF has the ABC signature motif and twin Walker motifs found in ModC as well as in other ABC-ATPases. Genetic analysis and lacZ fusion studies (both transcription and translation fusions) show that modF is coded by a single transcript originating upstream of modE (Ray and Shanmugam, unpublished results). It is possible that the ModF protein functions in molybdate export. In addition to these genes, the modABC, modEF intergenic region could potentially encode a small, 49 amino acid protein (Genbank Accession # AE000178) but its production in E. coli is yet to be established. Deleting this region of DNA did not have any effect on the cell’s ability to produce molybdoenzymes or regulate molybdate transport (unpublished results). Since E. coli growing under anaerobic conditions produces a limited number of molybdoenzymes, the growth medium of a wild type strain normally needs no supplementation with molybdate. Molyb-
date present as a contaminant in various medium components is sufficient to support production of an optimum level of formate hydrogenlyase (FHL) activity which contains the molybdoenzyme formate dehydrogenase. Under these conditions, molybdate is transported by the high-affinity molybdate transporter, ModABC. The apparent K m for molybdate uptake by E. coli is about 25–50 nM at pH 7.0 [5, 20]. With a mutation in any one of the three mod genes (modABC), the amount of molybdate required for molybdoenzyme synthesis and activity depends on the medium composition, especially the presence of S compounds [31, 54]. In such a mutant, molybdate is transported by either the ABC-type sulfate transport system or by a nonspecific anion transporter (figure 1). With the sulfate transporter transporting molybdate, the amount of molybdate required for optimum expression of FHL activity is about 1 µM. A mod, cysA double mutant requires about 100 µM molybdate for an optimum level of active molybdoenzyme production. Selenite inhibits the accumulation of molybdate by the nonspecific anion transporter, although in E. coli, sulfate, as high as 30 mM, had no effect on this process . Thus, E. coli has the potential of utilizing one of the three transporters with varying affinity for molybdate. Although the sulfate transporter has been implicated in molybdate transport based on genetic and physiological studies, direct transport studies are yet to be performed. Also, the role of SBP or CysP protein, the two periplasmic sulfate and thiosulfate binding proteins, respectively, in molybdate transport by the CysUWA proteins is not known. Based on structural constraints (discussed later), the SBP may not bind molybdate effectively and thus may not function in molybdate transport. Various proteins which participate in molybdate transport apparently also transport tungstate, since E. coli is capable of producing active W-containing TMAO-reductase . However, the kinetics of tungstate transport by the different molybdate transporters in E. coli is yet to be investigated. 3. Molybdate transport in other microorganisms Based on genome sequence information and similarity to E. coli mod genes, mod genes can be identified in at least 20 bacteria and archaea. These include Azotobacter vinelandii, Rhodobacter capsulatus, Haemophilus influenzae, Vibrio cholerae,
W.T. Self et al. / Res. Microbiol. 152 (2001) 311–321
Bacillus subtilis, Synechocystis sp., Staphylococcus carnosus, Pseudomonas aeruginosa, Helicobacter pylori, Arthrobacter nicotinovorans, Deinococcus radiodurans, Campylobacter jejuni, Mycobacterium tuberculosis, Aquifex aeolicus, Aeropyrum pernix, Methanobacterium thermoautotrophicum, Methanopyrus kandleri, Streptomyces coelicolor, Haloferax volcanii, Pyrococcus abyssi, Pyrococcus furiosus and Pyrococcus horikoshii [4, 6, 7, 14, 15, 22 – 24, 27, 33, 39, 44, 45, 47, 56, 58, 60, 63, 65 – 68]. The presence of one or more of the mod genes in these prokaryotes suggests that the mod genes also exist in other related organisms. Among this group, only in a select few including Azotobacter vinelandii, Rhodobacter capsulatus, Staphylococcus carnosus and Haloferax volcanii have the mod genes been functionally defined and the appropriate DNA cloned and identified [33, 44, 65, 66]. In most of these bacteria and archaea, the three mod genes are organized as a single operon. In B. subtilis and D. radiodurans, the gene coding for the ABC-ATPase is not in the same operon and is apparently located elsewhere in the genome. The hyperthermophilic archaeal Pyrococcus sp. lack the ModA homolog. In the genome sequence of the hyperthermophilic bacterium Aquifex aeolicus and archaeon Aeropyrum pernix ModB and ModC homologs are not readily discernable but a ModA homolog is detectable. Although genes for molybdate transport are found in various bacteria and archaea, ModE homologs have only been identified in a few bacterial species (see later section). In several organisms, the mod genes are a part of a gene cluster coding for a specific molybdoenzyme. In R. capsulatus, the mod genes are located within the nif gene cluster coding for the molybdoenzyme nitrogenase and are apparently regulated along with the nif genes . The σ 54 component of RNA polymerase required for nif gene expression is also required for transcription of mod genes in this organism. In the methanogens Methanobacterium and Methanopyrus, the mod genes are clustered with the genes encoding the molybdoenzyme formyl methanofuran dehydrogenase [56, 63]. In both of these examples, the molybdoenzymes constitute a significant fraction of the soluble proteins in the cytoplasm and apoprotein synthesis is coordinated with molybdate transport by common regulators.
4. ModA, the periplasmic binding protein 4.1. ModA from E. coli
The E. coli modA gene codes for a 257 amino acid precursor protein from which a 24 amino acid signal peptide is cleaved off to produce a 233 amino acid mature protein. The ModA protein binds molybdate as well as its analog tungstate at equimolar quantities with a KD of 20 ± 8 nM [20, 29, 53]. The oxyanions sulfate and phosphate are not effective ligands for ModA. Imperial et al.  also determined a KD value of 67 nM for ModA in vivo using an E. coli mutant which accumulated molybdate in the periplasm but not in the cytoplasm. The isoelectric point of the mature protein decreased from 7.0 to 5.6 in the presence of molybdate suggesting a conformational change associated with molybdate-binding and alteration of solvent exposed charge. In agreement with the conformational change, molybdate or tungstate binding also altered the intrinsic fluorescence of ModA. This conformational change associated with ligand-binding is similar to the changes observed with other periplasmic binding proteins, including the other oxyanions sulfate- and phosphatebinding proteins . The crystal structures of several periplasmic binding proteins are currently available and the tertiary structures of these proteins are similar, although amino acid sequence identity among these binding proteins is minimal . The periplasmic binding proteins consist of two well-separated globular domains connected by a hinge region. Upon ligand binding, the two domains come together creating a deep cleft in which the ligand interfaces the two domains. In the absence of bound ligand, the cleft is open and is accessible to bulk solvent. For the sulfate-binding protein (SBP) from Salmonella typhimurium, the binding of oxyanion sulfate is accompanied by dehydration and water expulsion from the cleft . The crystal structures of E. coli ModA2− MoO2− 4 and ModA-WO4 as well as A. vinelandii 2− ModA2-MoO4 show that the tertiary structures of the ligand-bound form of ModA are very similar to that of the other periplasmic binding proteins including the SBP [19, 29]. Both molybdate and tungstate bind to the E. coli ModA protein as a tetrahedral complex which is held by seven hydrogen bonds formed between the oxygen of the bound anion and protein groups from
W.T. Self et al. / Res. Microbiol. 152 (2001) 311–321
the two domains. Four of the hydrogen bond donors are the main chain NH groups of residues 12 (Ser), 39 (Ser), 125 (Ala) and 152 (Val). Three hydroxyl groups (Ser-12, Ser-39 and Tyr-170) provide the remaining hydrogen bonds . Although bound molybdate is also supported by seven hydrogen bonds in A. vinelandii ModA2, the amino acid donors to the tetrahedral molybdate differ from that of the E. coli protein. In A. vinelandii ModA2, the peptide chain NH group of Asn-10, Ser-37, Tyr-118 and Val-147 and the side chain groups of Thr-9, Asn-10 and Ser37 contribute to the seven hydrogen bonds . The ModA proteins have a very low affinity for sulfate (dissociation constant of > 2 mM) and likewise the SBP does not readily bind molybdate (KD value > mM) or tungstate [19, 20, 29, 53]. This substrate specificity has been proposed to be due to the size of the binding site. The average metal-oxygen distances for molybdate and tungstate are 1.77 Å and 1.78 Å, respectively, compared to 1.47 Å for sulfate (see [19, 30]). Since molybdate and tungstate are larger than sulfate, these oxyanions may not be able to fit in the sulfate binding site in SBP. Also, in in vivo studies, the SBP or the CysP (thiosulfate/sulfate binding protein) failed to substitute for ModA in molybdate transport. This raises interesting questions about the mechanism which enables transport of molybdate via the sulfate transporter when the molybdate concentration of the medium is less than 1 µM. As yet, there are no answers to these questions. In a sulfate transport-defective mutant of E. coli, the molybdate transporter is capable of transporting sulfate to partially suppress the Cys– phenotype (unpublished results). However, this requires derepression of the modABC genes since these genes are normally expressed at low levels in a wild type. 4.2. ModA from other organisms
As expected, there is considerable amino acid sequence identity/similarity among the ModA proteins from various prokaryotes. However, significant differences do exist among these orthologs (figure 2). Based on sequence similarity, the ModA proteins from various organisms can be grouped into four families with family 3 further divided into two subgroups. The ModA protein from all the members of family 1 and family 4 have highly conserved amino acid sequences at and near the amino acids which donate hydrogen bond(s) to molybdate in the E. coli ModA pro-
Figure 2. An unrooted radial phylogenetic tree of molybdate binding proteins. The phylogeny was inferred using the programs Clustal W and TREEVIEW [46, 59]. The scale represents 10% divergence between sequences. SBP, sulfate binding protein; CysP thiosulfate binding protein. The specific organisms represented in this study are V. cholerae, H. influenzae strain Rd, M. kandleri, A. pernix, M. thermoautotrophicum, A. vinelandii, R. capsulatus, P. aeruginosa, A. aeolicus, H. pylori, C. jejuni, D. radiodurans R1, S. coelicolor, M. tuberculosis, A. nicotinovorans, Synechocystis sp. PCC 6803, B. subtilis and S. carnosus. See text for details.
tein. Similarly, the ModA sequences from the members of family 3 are also highly conserved except for the threonine at position 9 which donates a hydrogen bond from the side chain hydroxyl group found only in ModA2 from A. vinelandii . In its place, the other homologs carry alanine. It is highly likely that the hydrogen bond donors to molybdate within each family of ModA would be the same. It is expected that the tertiary structure of all the ModA homologs would be similar to each other and to the general class of periplasmic binding proteins of the ABC transporters. The modA sequence is present in the sequenced genomes of at least 20 prokaryotes, including E. coli; however, the three Pyrococcus spp. only have homologs of ModB and ModC. In these hyperthermophilic archaea, molybdate is functionally replaced by tungsten . Although E. coli and A. vinelandii ModA proteins bind both molybdate and tungstate equally well, it is possible that a Pyrococcus ModA protein may not share significant amino acid sequence similarity to E. coli ModA, thereby preventing detection of the pyrococcal ModA. The sequence corresponding to a periplasmic binding protein is not readily discernable near the modBC genes in either P. horikoshii or P. abyssi genomes but in
W.T. Self et al. / Res. Microbiol. 152 (2001) 311–321
Pyrococcus furiosus there is an ORF upstream of the putative modBC genes which may participate in tungstate transport. An alternative explanation is that the tungstate concentration within the thermal vents is high enough to support tungstate transport without the need for the periplasmic binding protein (0.37 µg W/Kg compared to < 0.002 µg Mo/Kg in thermal vent fluids ). It has been reported that growth of Pyrococcus furiosus in rich media requires tungstate supplementation , probably a consequence of ModA deficiency. This finding is in sharp contrast to studies of E. coli which demonstrate that molybdate requirements are fully satisfied by trace molybdate present in rich media. This then raises the question as to the ability of the ModBC proteins by themselves to transport tungstate in these hyperthermophiles, especially considering the need for a ligand-bound form of the periplasmic binding protein to energize the histidine and maltose transporters in E. coli [1, 42]. In mesophilic bacteria, such as E. coli, A. vinelandii and R. capsulatus, mutations in modA eliminate the high-affinity molybdate transport indicating its necessity. However, the presence of secondary transport systems for molybdate makes it difficult to evaluate the absolute requirement for this periplasmic binding protein in bacteria such as E. coli. Based on the need for periplasmic binding protein in ATP hydrolysis and transport in other ABC transporters [1, 16, 42], it is readily conceivable that the ModA protein is an essential component of the high-affinity molybdate transporter. However, it should be noted that suppressor mutations in the membrane component or the ATPase did support maltose or histidine transport in E. coli mutants lacking the corresponding periplasmic ligand-binding protein, although at a significantly lower affinity for the substrate [57, 61]. 5. ModB, the membrane channel protein The ModB protein of E. coli (229 amino acid) is a hydrophobic protein which has an inner membrane signature motif characteristic of the ABC transporters . Based on the requirement of this protein for molybdate transport and the presence of this motif, it is presumed that ModB, as a homodimer, provides the channel for molybdate transit across the cytoplasmic membrane. Five transmembrane regions can be predicted in this protein and the ABC permease signature sequence is located in the cytoplasmic loop be-
Figure 3. A schematic representation of the transmembrane helices of E. coli ModB and the position of unique sequences in the protein. See text for details.
tween the transmembrane regions 3 and 4 (consensus sequence 2; figure 3). Based on the amino acid sequence, 19 ModB homologs can be grouped into the same four families seen with the ModA homologs (figure 2). Comparative analysis of these sequences revealed two additional regions of sequence similarity among the ModB homologs. The first sequence is located between amino acids 53 and 66 in E. coli ModB and is within the second transmembrane domain (figure 3). The consensus sequence among the various ModB homologs is ‘LPLVLPP(V/T/S)VhG(F/Y)XL’ in which h represents a hydrophobic amino acid while X stands for any amino acid. This consensus sequence can also be identified in the sulfate permease proteins CysU and CysW (10/14 amino acids are similar) and the PstA and PstC proteins of the phosphate permease of E. coli. However, this sequence is not readily detectable in the permeases of other ABC transporters (histidine, maltose, etc.) of E. coli suggesting that these amino acids play a significant role in anion transport. The third consensus sequence, FAR(S/T)LGEFG (A/V)(T/V), comprises the amino acids between 157 and 167 in the E. coli ModB. This sequence also extends to the E. coli CysU and CysW proteins (sulfate permease) but not to the E. coli ABC permeases for phosphate, histidine or maltose. The amino acids F and A are within the predicted transmembrane helix 4 and the rest of the amino acids extend from the membrane into the periplasm (figure 3). In Pyrococcus abyssi and Pyrococcus horikoshii, the two organ-
W.T. Self et al. / Res. Microbiol. 152 (2001) 311–321
isms which lack ModA homologs, the amino acid sequence GEFG is replaced by SEVG suggesting that these changes play a role in transporting tungstate without the aid of the binding protein. It has been proposed that the ligand-bound form of the binding protein and the ATPase component interact with each other across the membrane which leads to hydrolysis of ATP . The transmembrane helix 2 which carries the unique consensus sequence 1 and the EQAA signature sequence in the cytoplasmic loop between transmembrane helices 3 and 4 may represent the regions of ATPase contact with the permease, both in the cytoplasm and in the membrane. Consensus sequence 3 is unique to the molybdate permease and to a limited extent also to the sulfate permeases. These oxyanions are bound by their respective binding proteins in a similar manner with the dehydrated anions hydrogen bonded through the metal oxygen. It is interesting to note that the amino acids in the third consensus sequence extend from the membrane into the periplasm where the binding protein is located. The ligand-bound form of the binding protein may interact with the permease at this particular location and thus facilitate ATP hydrolysis by ModC and molybdate transport. Biochemical and mutational analysis of the amino acids in these two unique regions along with suppressor mutations in ModA/ModC would provide evidence in support of these possibilities. 6. ModC, the ABC-ATPase The ModC protein of E. coli is a 352-amino acid protein coded by the third gene of the modABC operon. The ModC protein has the Walker A and B motifs as well as the ABC signature sequence associated with the ABC-ATPases . Based on the amino acid sequence and the demonstrated requirement for this protein in molybdate transport, ModC and its homologs are presumed to serve as the energizing protein for molybdate transport. Besides the Walker motifs and ABC signature sequence, no additional unique sequence can be readily visualized among the various ModC homologs. It is possible that the specificity for molybdate transport comes from the unique sequences found in the ModA and ModB proteins. However, CysA, the ATPase associated with sulfate transport, is not known to replace ModC. Can specific mutation(s) in CysA or other ATPases allow
these proteins to replace the ModC for molybdate transport? This is a question for the future. 7. Regulation of molybdate transport In wild-type E. coli molybdate transport genes are expressed at extremely low levels [52, 54]. Experiments using a mod mutant which carries a lacZ fusion in any of the three mod genes indicated that the mod operon in E. coli is tightly regulated and requires molybdate starvation for detectable level of expression of modABC. This regulation is achieved by a repressor protein, ModE, coded by a gene in an operon which diverges from modABC . A search of the genome sequences in the database revealed that only 5 other bacteria (A. vinelandii, H. influenzae, Herbaspirillum seropedicae, P. aeruginosa and R. capsulatus) have a gene corresponding to E. coli ModE [7, 33, 58, 62, 65]. This is in contrast to over 20 prokaryotes which contain the homologs of genes coding for molybdate transport. Of the five ModE homologs, the proteins from A. vinelandii and R. capsulatus have been defined by genetic analysis and the modE gene from H. influenzae functionally complements an E. coli modE mutation [11, 28, 40]. The ModE protein of E. coli (262 amino acid) has a helix-turn-helix region in the N-terminal segment and is a member of the LysR family of transcriptional regulators. The ModE-molybdate complex (ModE-Mo) is the active form of the protein and binds to target DNA as a dimer [2, 10]. The K d for this association between modA operator DNA and ModE-Mo is 0.3 nM which is in agreement with other repressor/operator DNA interactions . ModE also interacts with tungstate and the affinity of ModE-tungstate for modA operator DNA is about 6-times lower than that of ModE-Mo. The molybdate-free form of ModE also binds to the modA operator DNA but with a 25fold lower affinity (Kd of 8 nM) than ModE-Mo. Although ModE binds to modA operator DNA in vitro, it is not expected to be a repressor in vivo in the absence of molybdate. Sulfate and vanadate failed to substitute for molybdate both in vivo and in vitro in the ModE-DNA association. Upon binding molybdate, the conformation of ModE changes and this is manifested as a decrease in intrinsic fluorescence of the protein [2, 10]. Using this decrease in fluorescence as a measure of Mo-binding, a Kd of 0.8 µM was determined for this association. The stoichiometry of Mo
W.T. Self et al. / Res. Microbiol. 152 (2001) 311–321
binding to ModE is 1:1. Apparently, the high-affinity molybdate transport system effectively transports sufficient molybdate, even when the cells are growing in a medium not supplemented with molybdate (present in the medium constituents as a contaminant), to produce significant amounts of ModE-Mo complex to optimally regulate molybdate transport. The ModE protein dimer was crystallized and the structure of the dimer has been elucidated . Unfortunately, the ModE crystal lacked molybdate although molybdate was present in the crystallization solution. The crystal structure reveals that the ModE contains two major domains. The N-terminal domain of up to amino acid 121 (domain I) forms a winged helix-turn-helix motif and interacts with DNA. The C-terminal domain II is the putative molybdatebinding component and can be further separated into two subgroups each of which forms a β-barrel. There is significant sequence identity between the two subdomains. The N-terminal domain plays a major role in the dimerization of the protein while the Cterminal domain also contributes to the stability of the complex. Hall et al.  proposed that the amino acids Arg-128, Ser-166, Lys-183 and Glu-218 are involved in molybdate binding. Mutant forms of the ModE which are molybdateindependent for repression of modABC provide some information about the molybdate-binding site in the ModE [9, 35]. Two of the mutations (T125I, G133D)  altered the amino acids near a unique region, 125(T/S)SARNQXXG-133, and the amino acid sequence SARNQ is conserved in all ModE homologs as well as in Clostridium pasteurianum molybdopterinbinding proteins . Based on the highly conserved nature of these amino acids and the finding that mutations in this sequence render the protein molybdate-independent for DNA binding, it has been proposed that this region contributes to molybdate binding. In addition, removal of the C-terminal 10 amino acids (part of subdomain IIa) also produces a Mo-independent phenotype. The Mo-independent mutant forms of ModE have lower intrinsic fluorescence resembling the native ModE-Mo . Based on domain-swapping experiments, McNicholas et al.  proposed that ModE dimerizes upon molybdatebinding. Modeling the T125I mutation into the crystal structure of ModE suggested that the mutation increased the dimerization potential of the protein. All these studies indicate that the conformation of ModE changes upon binding molybdate. This conformation
Figure 4. The E. coli modABC operator sequence and the effect of mutations on the ability of ModE-Mo to bind to the DNA. Top: modABC upstream region. Bases enclosed within the hatched area are protected by ModE-Mo from DNase I hydrolysis. Letters in reverse print represent unique pentamer sequences. The two arrows indicate the dyad symmetry in the sequence. Bottom: various mutants within the E. coli modABC operator DNA. The values in parenthesis represent the relative affinity of the altered DNA to that of the wild type operator DNA which is taken as 1.0 (unpublished results).
change promotes the dimerization of ModE-Mo. The protein dimer is competent to bind DNA and repress the modABC operon. 8. The ModE binding region in the modABC operator DNA The molybdate-bound form of ModE binds to the modA operator DNA between bases −15 and +14 in E. coli (figure 4) [2, 10, 38]. A region of dyad symmetry which is highly conserved in the operator regions of various operons regulated by molybdate availability can be readily detected in E. coli as well as in other organisms . The common feature of the DNA protected by ModE from DNase I hydrolysis in E. coli is a pentamer sequence, TAYAT (Y = pyrimidine). In E. coli, the first and second pentamers are separated by 7 bases and the third pentamer is separated by five bases from the second pentamer. The third pentamer is not essential for ModE-binding both in vivo and in vitro and is also not detected in other ModE-controlled operons except for the modEABC of A. vinelandii (figure 5) .
W.T. Self et al. / Res. Microbiol. 152 (2001) 311–321
of excess molybdate in the cytoplasm and enhance molybdoenzyme synthesis. 9. Concluding remarks
Figure 5. Sequence similarities among operator/promoter DNA from operons of different organisms which are known to be under molybdate control and repressed or activated by ModEMo or its homologs. The ModE-Mo binding regions in the E. coli operons were determined experimentally [2, 10, 36, 38, 55]. In DNA sequences from other organisms, the ModE-Mo binding regions are inferred based on mutational analysis and similarities to the E. coli DNA and the upstream location of the sequence in relation to the translation start site of the listed gene (H. influenzae ; A. vinelandii ; R. capsulatus ; P. aeruginosa ). Regulation of the P. aeruginosa modABC operon by the ModE-Mo complex has not been demonstrated.
Deletion of the third pentamer sequence in the E. coli mod operon decreases the affinity for ModE-Mo by about 50% while the first and second pentamer sequences are essential for ModE-Mo binding. Based on mutational analysis of various bases both in and near the pentamer sequences (figure 4; unpublished results) combined with the mod operator sequence from other organisms, a consensus DNA sequence for repression by ModE-Mo can be derived (figure 5). Besides its role as repressor of molybdate transport genes, ModE-Mo also enhances the expression of genes coding for molybdoenzymes such as fdhF (formate dehydrogenase-H; FDH-H), hyc (hydrogenase 3, part of the formate hydrogenlyase complex along with FDH-H) and narXL (regulators for respiratory nitrate reductase) both directly and indirectly . Based on studies of direct ModE-Mo binding to DNA, it was found that the promoter DNA sequence to which ModE-Mo binds is different from that of the operator consensus sequence (figure 5). The first pentamer sequence is replaced by a tetramer followed by seven bases and the second pentamer sequence (figure 5). These studies show that the molybdate sensor protein, ModE, as a ModE-Mo complex, serves to both repress molybdate transport in the presence
Despite the identification of the genes and proteins involved in molybdate transport and regulation in E. coli, considerable amount of work is still needed to fully understand the pathways of molybdate transport. Although molybdate transport mutants are defined as mutations which can be suppressed by a high concentration of molybdate, there are other mutations (moeA)  which are also suppressible by molybdate but affect molybdate activation in the cytoplasm. Thus, it is imperative that the actual genotype of each of these mutants be established. Even though molybdate is an essential element in both plants and animals, including humans, a molybdate transporter has not yet been identified in eukaryotes. Is there a specific high affinity molybdate transporter in humans or is molybdate transported by a nonspecific anion transporter found in various tissues ? What is the nature of the nonspecific anion transporter which transports molybdate in E. coli and how does this compare with its human or eukaryotic counterparts? The sulfate binding protein has a Kd for molybdate close to 1 mM but the sulfate transporter still transports molybdate from media containing a molybdate concentration of < 1 µM. What is the role of CysP, the thiosulfate/sulfate binding protein, in this process? What are the sites of interaction between the three components of the transporter and how do the unique sequences found in the permease contribute to this interaction? How is molybdate transport regulated in the organisms which clearly have molybdate-specific transporters but no obvious ModE-homologs? These are some of the interesting questions for the future. Acknowledgements The work from the authors’ laboratory was supported by funds from the National Institutes of Health and from the State of Florida Agricultural experiment station. The authors thank Drs R. McKenna and A. McKenna for help in modeling ModE mutant structure.
W.T. Self et al. / Res. Microbiol. 152 (2001) 311–321
References  Ames G.F.L., Liu C.E., Joshi A.K., Nikaido K., Liganded and unliganded receptors interact with equal affinity with the membrane complex of periplasmic permeases, a subfamily of traffic ATPases, J. Biol. Chem. 271 (1996) 14264–14270.  Anderson L.A., Palmer T., Price N.C., Bornemann S., Boxer D.H., Pau R.N., Characterisation of the molybdenumresponsive ModE regulatory protein and its binding to the promoter region of the modABCD (molybdenum transport) operon of Escherichia coli , Eur. J. Biochem. 246 (1997) 119–126.  Buc J., Santini C.L., Giordani R., Czjzek M., Wu L.F., Giordano G., Enzymatic and physiological properties of the tungsten-substituted molybdenum TMAO reductase from Escherichia coli , Mol. Microbiol. 32 (1999) 159–168.  Cole S.T. et al., Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence, Nature 393 (1998) 537–544.  Corcuera G.L., Bastidas M.A., Dubourdieu M., Molybdenum uptake in Escherichia coli K12, J. Gen. Microbiol. 139 (1993) 1869–1875.  Deckert G. et al., The complete genome of the hyperthermophilic bacterium Aquifex aeolicus , Nature 392 (1998) 353–358.  Fleischmann R.D. et al., Whole-genome random sequencing and assembly of Haemophilus influenzae Rd, Science 269 (1995) 496–512.  Gimenez I., Garay R., Alda J.O., Molybdenum uptake through the anion exchanger in human erythrocytes, Pflugers Arch. 424 (1993) 245–249.  Grunden A.M., Ray R.M., Rosentel J.K., Healy F.G., Shanmugam K.T., Repression of the Escherichia coli modABCD (molybdate transport) operon by ModE, J. Bacteriol. 178 (1996) 735–744.  Grunden A.M., Self W.T., Villain M., Blalock J.E., Shanmugam K.T., An analysis of the binding of repressor protein ModE to modABCD (Molybdate transport) Operator/Promoter DNA of Escherichia coli , J. Biol. Chem. 274 (1999) 24308–24315.  Grunden A.M., Shanmugam K.T., Molybdate transport and regulation in bacteria, Arch. Microbiol. 168 (1997) 345–354.  Hall D.R., Gourley D.G., Leonard G.A., Duke E.M., Anderson L.A., Boxer D.H., Hunter W.N., The high-resolution crystal structure of the molybdate-dependent transcriptional regulator (ModE) from Escherichia coli : a novel combination of domain folds, EMBO J. 18 (1999) 1435–1446.  Hasona A., Ray R.M., Shanmugam K.T., Physiological and genetic analyses leading to identification of a biochemical role for the moeA (molybdate metabolism) gene product in Escherichia coli , J. Bacteriol. 180 (1998) 1466–1472.  Heidelberg J.F. et al., DNA sequence of both chromosomes of the cholera pathogen Vibrio cholerae, Nature 406 (2000) 477–483.  Heilig R., Pyrococcus abyssi genome sequence: insights into archaeal chromosome structure and evolution, Genoscope (www.genoscope.cns.fr) (1999).  Higgins C.F., ABC transporters: from microorganisms to man, Ann. Rev. Cell Biology 8 (1992) 67–113.
 Hille R., The mononuclear molybdenum enzymes, Chem. Rev. 96 (1996) 2757–2816.  Holland I.B., Blight M.A., ABC-ATPases, adaptable energy generators fuelling transmembrane movement of a variety of molecules in organisms from bacteria to humans, J. Mol. Biol. 293 (1999) 381–399.  Hu Y., Rech S., Gunsalus R.P., Rees D.C., Crystal structure of the molybdate binding protein ModA, Nature Struct. Biol. 4 (1997) 703–707.  Imperial J., Hadi M., Amy N.K., Molybdate binding by ModA, the periplasmic component of the Escherichia coli mod molybdate transport system, Biochim. Biophys. Acta 1370 (1998) 337–346.  Johnson J.L., Wadman S.K., Molybdenum cofactor deficiency, in: Scriver C.R., Beaudet A.L., Sly W.S., Valle D. (Eds.), The metabolic and molecular basis of inherited disease, McGraw-Hill, New York, 1995, pp. 2271–2283.  Kaneko T., Tabata S., Complete genome structure of the unicellular cyanobacterium Synechocystis sp. PCC6803, Plant Cell Physiol. 38 (1997) 1171–1176.  Kawarabayasi Y. et al., Complete sequence and gene organization of the genome of a hyper-thermophilic archaebacterium, Pyrococcus horikoshii OT3, DNA Res. 5 (1998) 147–155.  Kawarabayasi Y. et al., Complete genome sequence of an aerobic hyper-thermophilic crenarchaeon, Aeropyrum pernix K1, DNA Res. 6 (1999) 83–101.  Kisker C., Schindelin H., Rees D.C., Molybdenum-cofactorcontaining enzymes: structure and mechanism, Ann. Rev. Biochem. 66 (1997) 233–267.  Kletzin A., Adams M.W., Tungsten in biological systems, FEMS Microbiol. Rev. 18 (1996) 5–63.  Kunst F. et al., The complete genome sequence of the grampositive bacterium Bacillus subtilis , Nature 390 (1997) 249– 256.  Kutsche M., Leimkuhler S., Angermuller S., Klipp W., Promoters controlling expression of the alternative nitrogenase and the molybdenum uptake system in Rhodobacter capsulatus are activated by NtrC, independent of sigma54, and repressed by molybdenum, J. Bacteriol. 178 (1996) 2010– 2017.  Lawson D.M., Williams C.E., Mitchenall L.A., Pau R.N., Ligand size is a major determinant of specificity in periplasmic oxyanion-binding proteins: the 1.2 D resolution crystal structure of Azotobacter vinelandii ModA, Structure 6 (1998) 1529–1539.  Lawson D.M., Williams C.E., White D.J., Choay A.P., Mitchenall L.A., Pau R.N., Protein ligands for molybdate, specificity and charge stabilisation at the anion-binding sites of periplasmic and intracellular molybdate-binding proteins of Azotobacter vinelandii , J. Chem. Soc. Dalton Trans. (1997) 3981–3984.  Lee J.H., Wendt J.C., Shanmugam K.T., Identification of a new gene, molR, essential for utilization of molybdate by Escherichia coli , J. Bacteriol. 172 (1990) 2079–2087.  Leimkuhler S., Angermuller S., Schwarz G., Mendel R.R., Klipp W., Activity of the molybdopterin-containing xanthine dehydrogenase of Rhodobacter capsulatus can be restored by high molybdenum concentrations in a moeA mutant
W.T. Self et al. / Res. Microbiol. 152 (2001) 311–321 defective in molybdenum cofactor biosynthesis, J. Bacteriol. 181 (1999) 5930–5939. Luque F., Mitchenall L.A., Chapman M., Christine R., Pau R.N., Characterization of genes involved in molybdenum transport in Azotobacter vinelandii , Mol. Microbiol. 7 (1993) 447–459. Maupin-Furlow J.A., Rosentel J.K., Lee J.H., Deppenmeier U., Gunsalus R.P., Shanmugam K.T., Genetic analysis of the modABCD (molybdate transport) operon of Escherichia coli , J. Bacteriol. 177 (1995) 4851–4856. McNicholas P.M., Chiang R.C., Gunsalus R.P., The Escherichia coli modE gene: effect of modE mutations on molybdate dependent modA expression, FEMS Microbiol. Lett. 145 (1996) 117–123. McNicholas P.M., Chiang R.C., Gunsalus R.P., Anaerobic regulation of the Escherichia coli dmsABC operon requires the molybdate-responsive regulator ModE, Mol. Microbiol. 27 (1998) 197–208. McNicholas P.M., Mazzotta M.M., Rech S.A., Gunsalus R.P., Functional dissection of the molybdate-responsive transcription regulator, ModE, from Escherichia coli , J. Bacteriol. 180 (1998) 4638–4643. McNicholas P.M., Rech S.A., Gunsalus R.P., Characterization of the ModE DNA-binding sites in the control regions of modABCD and moaABCDE of Escherichia coli , Mol. Microbiol. 23 (1997) 515–524. Menendez C., Otto A., Igloi G., Nick P., Brandsch R., Schubach B., Bottcher B., Molybdate-uptake genes and molybdopterin-biosynthesis genes on a bacterial plasmid – characterization of MoeA as a filament-forming protein with adenosinetriphosphatase activity, Eur. J. Biochem. 250 (1997) 524–531. Mouncey N.J., Mitchenall L.A., Pau R.N., The modE gene product mediates molybdenum-dependent expression of genes for the high-affinity molybdate transporter and modG in Azotobacter vinelandii , Microbiology 142 (1996) 1997– 2004. Mourez M., Hofnung M., Dassa E., Subunit interactions in ABC transporters: a conserved sequence in hydrophobic membrane proteins of periplasmic permeases defines an important site of interaction with the ATPase subunits, EMBO J. 16 (1997) 3066–3077. Mourez M., Jehanno M., Schneider E., Dassa E., In vitro interaction between components of the inner membrane complex of the maltose ABC transporter of Escherichia coli : modulation by ATP, Mol. Microbiol. 30 (1998) 353–363. Mukund S., Adams M.W., Molybdenum and vanadium do not replace tungsten in the catalytically active forms of the three tungstoenzymes in the hyperthermophilic archaeon Pyrococcus furiosus , J. Bacteriol. 178 (1996) 163–167. Neubauer H., Pantel I., Lindgren P.E., Gotz F., Characterization of the molybdate transport system ModABC of Staphylococcus carnosus , Arch. Microbiol. 172 (1999) 109–115. Oliver K., Harris D., Bentley S.D., Parkhill J., Barrell B.G., Rajandream M.A., Streptomyces coelicolor sequencing project, Sanger Centre (http://www.sanger.ac.uk/Projects/S. coelicolor/). Page R.D.M., TREEVIEW: An application to display phylogenetic trees on personal computers, Computer Applications in the Biosciences 12 (1996) 357–358.
 Parkhill J. et al., The genome sequence of the foodborne pathogen Campylobacter jejuni reveals hypervariable sequences, Nature 403 (2000) 665–668.  Pau R.N., Klipp W., Leimkuhler S., Molybdenum transport, processing and gene regulation, in: Winkelmann G., Carrano C.J. (Eds.), Transition metals in microbial metabolism, Harwood Academic Publishers, Amsterdam, The Netherlands, 1997, pp. 217–234.  Pflugrath J.W., Quiocho F.A., The 2Å resolution structure of the sulfate-binding protein involved in active transport in Salmonella typhimurium, J. Mol. Biol. 200 (1988) 163–180.  Quiocho F.A., Atomic structures of periplasmic binding proteins and the high-affinity active transport systems in bacteria, Phil. Trans. R. Soc. London B 326 (1990) 341–351.  Rajagopalan K.V., Johnson J.L., The pterin molybdenum cofactors, J. Biol. Chem. 267 (1992) 10199–10202.  Rech S., Deppenmeier U., Gunsalus R.P., Regulation of the molybdate transport operon, modABCD, of Escherichia coli in response to molybdate availability, J. Bacteriol. 177 (1995) 1023–1029.  Rech S., Wolin C., Gunsalus R.P., Properties of the periplasmic ModA molybdate-binding protein of Escherichia coli , J. Biol. Chem. 271 (1996) 2557–2562.  Rosentel J.K., Healy F., Maupin-Furlow J.A., Lee J.H., Shanmugam K.T., Molybdate and regulation of mod (molybdate transport), fdhF, and hyc (formate hydrogenlyase) operons in Escherichia coli , J. Bacteriol. 177 (1995) 4857–4864.  Self W.T., Grunden A.M., Hasona A., Shanmugam K.T., Transcriptional regulation of molybdoenzyme synthesis in Escherichia coli in response to molybdenum: ModEmolybdate, a repressor of the modABCD (molybdate transport) operon is a secondary transcriptional activator for the hyc and nar operons, Microbiology 145 (1999) 41–55.  Smith D.R. et al., Complete genome sequence of Methanobacterium thermoautotrophicum deltaH: functional analysis and comparative genomics, J. Bacteriol. 179 (1997) 7135–7155.  Speiser D.M., Ames G.F., Salmonella typhimurium histidine periplasmic permease mutations that allow transport in the absence of histidine-binding proteins, J. Bacteriol. 173 (1991) 1444–1451.  Stover C.K. et al., Complete genome sequence of Pseudomonas aeruginosa PA01, an opportunistic pathogen, Nature 406 (2000) 959–964.  Thompson J.D., Higgins D.G., Gibson T.J., CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice, Nucleic Acids Research 24 (1994) 4673–4680.  Tomb J.F. et al., The complete genome sequence of the gastric pathogen Helicobacter pylori , Nature 388 (1997) 539–547.  Treptow N.A., Shuman H.A., Genetic evidence for substrate and periplasmic-binding-protein recognition by the MalF and MalG proteins, cytoplasmic membrane components of the Escherichia coli maltose transport system, J. Bacteriol. 163 (1985) 654–660.  Voigt E.L., Pedrosa F.O., Souza E.M., Chubatsu L.S., Rigo L.U., Steffens M.B.R., Sequence of the modE gene
W.T. Self et al. / Res. Microbiol. 152 (2001) 311–321 from Herbaspirillum seropedicae, Genbank Accession # AAG12168 (2000).  Vorholt J.A., Vaupel M., Thauer R.K., A selenium-dependent and a selenium-independent formylmethanofuran dehydrogenase and their transcriptional regulation in the hyperthermophilic Methanopyrus kandleri , Mol. Microbiol. 23 (1997) 1033–1042.  Walkenhorst H.M., Hemschemeier S.K., Eichenlaub R., Molecular analysis of the molybdate uptake operon, modABCD, of Escherichia coli and modR, a regulatory gene, Microbiol. Res. 150 (1995) 347–361.  Wang G., Angermuller S., Klipp W., Characterization of
Rhodobacter capsulatus genes encoding a molybdenum transport system and putative molybdenum-pterin-binding proteins, J. Bacteriol. 175 (1993) 3031–3042.  Wanner C., Soppa J., Genetic identification of three ABC transporters as essential elements for nitrate respiration in Haloferax volcanii , Genetics 152 (1999) 1417–1428.  Weiss R. et al., Pyrococcus furiosus genome sequence, Genemate database (http://comb5-156.umbi.umd. edu) (1999).  White O. et al., Genome sequence of the radioresistant bacterium Deinococcus radiodurans R1, Science 286 (1999) 1571–1577.