Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. 2006 Nov 21; 103(47): 17933–17938.
Published online 2006 Nov 13. doi:  10.1073/pnas.0606673103
PMCID: PMC1635973

Mapping the Sinorhizobium meliloti 1021 solute-binding protein-dependent transportome


The number of solute-binding protein-dependent transporters in rhizobia is dramatically increased compared with the majority of other bacteria so far sequenced. This increase may be due to the high affinity of solute-binding proteins for solutes, permitting the acquisition of a broad range of growth-limiting nutrients from soil and the rhizosphere. The transcriptional induction of these transporters was studied by creating a suite of plasmid and integrated fusions to nearly all ATP-binding cassette (ABC) and tripartite ATP-independent periplasmic (TRAP) transporters of Sinorhizobium meliloti. In total, specific inducers were identified for 76 transport systems, amounting to ≈47% of the ABC uptake systems and 53% of the TRAP transporters in S. meliloti. Of these transport systems, 64 are previously uncharacterized in Rhizobia and 24 were induced by solutes not known to be transported by ABC- or TRAP-uptake systems in any organism. This study provides a global expression map of one of the largest transporter families (transportome) and an invaluable tool to both understand their solute specificity and the relationships between members of large paralogous families.

Keywords: ATP-binding cassette, expression, high throughput, transport, tripartite ATP-independent periplasmic

The rapid increase in sequenced bacterial genomes has led biologists to try and predict the function of thousands of proteins in hundreds of different organisms. A significant proportion of these proteins are transporters, of which there are many large paralogous families with significant sequence identity, but unknown solute specificity and function. Perhaps nowhere is this more apparent than for the periplasmic solute-binding protein (SBP)-dependent transporters of the rhizobia, where a very large number of such systems are found (14). These systems consist of the ATP-binding cassette-transporter (ABC-T) and tripartite ATP-independent periplasmic transporter (TRAP-T) families. ABC-Ts possess a common minimum structure consisting of four domains: two hydrophobic integral membrane domains and two ATP-binding cassettes (ABC) (5). In addition, bacterial ABC uptake systems contain an SBP. TRAP-Ts possess one small and one large integral membrane protein and an SBP. Unlike ABC-Ts, which use ATP hydrolysis to energize uptake, TRAP-Ts use the proton motive force (6).

There are 200 ABC genes in Sinorhizobium meliloti, 216 in Mesorhizobium loti, 269 in Rhizobium leguminosarum, and 240 in Bradyrhizobium japonicum compared with 67 in Escherichia coli and 124 in Pseudomonas aeruginosa (14, 79). The ABC genes of S. meliloti are organized into 146 uptake systems and 18 export systems (2). Whereas the TRAP-Ts number 15 systems in S. meliloti, 6 in R. leguminosarum, 22 in B. japonicum, and curiously, only two in M. loti, these numbers compare with 1 in E. coli and 6 in P. aeruginosa (14, 79). The increase in the number of systems may be due to the high affinity of these transporters, enabling the acquisition of a broad range of nutrients from the oligotrophic and nutrient-limited soil and rhizosphere habitats of free-living rhizobia. Although this increased number of transporters presents a tremendous challenge to analyze, it also represents a unique opportunity among the available bacterial genomes to use high-throughput screening to determine their induction profiles and, by inference, their likely solute specificity. Array and proteomic analyses are extremely powerful approaches to examine induction profiles in bacteria and have been used to study rhizobia. However, they only allow screening against a few selected conditions (1013) and therefore are unlikely to allow the function to be assigned to a significant number of the ABC-T and TRAP-T uptake systems in S. meliloti. However, the use of reporter gene fusions in a high-throughput platform offers the possibility of screening hundreds of compounds against all of these uptake systems. The ABC-T and TRAP-T families are ideal candidates for such screening because they have an extremely high affinity for transported solutes and are usually tightly regulated at the transcriptional level, presumably because they are large protein complexes that are energetically expensive to synthesize (5, 6, 14).

In this study we cloned, upstream of reporter genes, the putative promoter regions for the ABC-T and TRAP-T operons of S. meliloti by using both a multicopy plasmid and an integrating vector. Thus, the high sensitivity of a multicopy-replicating plasmid was complemented by the advantage of the comparatively undisturbed regulation of an integrated fusion. In total, 300 plasmid fusions were prepared along with 198 integrated fusions, and these were tested with up to 174 inducing conditions. This work attempts a genome-level characterization of the solute induction of both the ABC-T superfamily and the TRAP-T family. In most cases, the induction profile corresponds to the specificity of the solute transported. Therefore, this analysis represents a powerful approach to understand the function of two of the most important classes of transport systems.

Results and Discussion

The 76 systems found to be specifically induced by one or more compounds are summarized in Table 1, and the results are shown in Table 2, which is published as supporting information on the PNAS web site. They represent 47% of the ABC-uptake systems and 53% of the TRAP-Ts in S. meliloti. Each fusion was tested with up to 174 biologically active compounds including sugars, amino acids, amino acid derivatives, purines, pyrimidines, polyols, organic acids, and limitation conditions, shown in Table 3, which is published as supporting information on the PNAS web site. The results are shown in full in Tables 4–6, which are published as supporting information on the PNAS web site. Of these transporters, 64 systems were not previously characterized in Rhizobium, and 29 do not have a characterized ortholog in any organism. In addition, 24 were induced in response to solutes not previously shown to be transported by ABC- or TRAP-uptake systems. The success of this approach is demonstrated by the same induction profile being obtained in this study for 13 of 14 previously characterized ABC systems in Rhizobium (Table 7, which is published as supporting information on the PNAS web site). The one exception (Frc) is likely to have two differentially regulated promoters (see below). In the detailed analysis that follows, the induction profiles have been analyzed according to the structures of the inducing compounds. Full details of all compounds tested are given in Table 3. Stereoisomers are only referred to in the text when more than one form was tested. For consistency, we refer to operons by the gene for the SBP (e.g., ABC-T SMc02324 and TRAP-T SMb20442) and routinely refer to the classification system of Saier (15). This system is based on phylogenetic comparison of the ABC subunits and known solute specificity of systems (Table 8, which is published as supporting information on the PNAS web site).

Table 1.
Compounds that induce (≥3-fold) ABC-T and TRAP-T in S. meliloti

Induction in Response to Sugars.

The 33 operons identified as being specifically induced by sugars or their derivatives include systems primarily induced by pentose and hexose monomers (eight and six, respectively), oligosaccharides (seven), polyols (six), sugar amines (three), and sugar phosphates (three).

As a model system, we consider transport of the methyl-pentose, rhamnose, in detail because it is well characterized in R. leguminosarum bv trifolii, where a single mRNA transcript encompasses the rhaRSTPQUK genes (16). In S. meliloti, the equivalent system shows the same genetic organization. Our analysis showed that a plasmid fusion to SMc02323 (rhaR) was induced by rhamnose and erythritol (both >6-fold) (Table 2), whereas a plasmid fusion to SMc02324 (rhaS) gave no expression. The integrated fusion in SMc02325 (rhaT) was induced 9-fold by rhamnose (Table 2). These data for S. meliloti are consistent with a single promoter upstream of SMc02323 (rhaR) driving transcription of the entire operon rhaRSTPQUK when an inducer is present. Thus, comparison of our findings, in both the plasmid and integrated systems, with this documented example, validates the approach that examination of the inducing compound may shed light on the transported solute. For many of the systems that follow, multiple plasmid and integrated fusions allow a similar analysis of operon structure, although for brevity, this level of detail has been omitted but can be determined from Tables 4–6.

There are four transporters whose genes were induced primarily by the methyl pentose, fucose (Table 1). ABC-T SMc02774 was induced by d-fucose (7-fold) and l-fucose (4-fold) as reported by an integrated fusion in ABC SMc02773. Integrated fusions do not indicate whether the promoter lies immediately in front of SMc02773 or further upstream. With fucose as the observed inducer, it is noteworthy that directly upstream of SMc02774, and possibly part of the same mRNA transcript as SMc02773, is a gene encoding l-fucose dehydrogenase (SMc02775) and a sugar hydrolase (SMc02776).

Fucose also induced expression of reporter gene fusions in both the plasmid and integration systems in ABC-T SMb21103. The integrated reporter fusion was induced 14-fold by d-fucose and 6-fold by l-fucose and the plasmid system ≈5-fold by both l-fucose and d-fucose, thus mapping the promoter directly in front of SMb21103.

ABC-T SMb21587 was induced by d-fucose (5-methyl-l-arabinose) (40-fold), l-arabinose (10–33-fold), and talose (11-fold).

The following TRAP-T systems are particularly notable because there are no described examples of TRAP-Ts transporting sugars. TRAP-T SMb20442, as measured from an integrated fusion in SMb20444, was induced 4-fold by both d-fucose and mannose and 3-fold by l-fucose. A plasmid reporter fusion revealed a constitutive promoter upstream of SMb20441 (GntR-type regulator), but no expression was obtained on cloning the region upstream of SMb20442 (SBP). The region upstream of SMb20441 may act as the promoter for a large transcript, including SMb20441, SMb20442 (SBP), SMb20443 (small permease), SMb20444 (large permease), SMb20445 (putative alcohol dehydrogenase), and SMb02446 (d-mannonate dehydratase), because the intergenic regions are very small. In the plasmid system, the regulation of expression of this operon may be disrupted. To examine this possibility, a plasmid including all of SMb20441 (including its upstream region), the intergenic region between SMb20441 and SMb20442, and the 5′ region of SMb20442 was fused to gfp and found to be induced 3-fold by mannose. The most likely explanation is that this is a negatively regulated inducible system (similar to the E. coli lac operon), with SMb20441 being a repressor and mannose an inducer with the operator lying upstream of SMb20441. A plasmid containing the region upstream of SMb20441 alone would bind the negative regulator, resulting in its removal and constitutive activation of the promoter.

For TRAP-T SMa0252, none of the fusions to genes encoding transport components showed significant induction, but a plasmid fusion to SMa0247 (fumarylacetoacetate hydrolase), which appears to be in the same operon, was induced by the aldose l-arabinose (5-fold). It also was induced by the disaccharide lactose (Galβ1-4Glc) and the trisaccharide raffinose (Galα1-4Gluα1-4Fru) (both 3-fold). The structures of the sugars appear to have little in common, other than that the di- and trisaccharide have the aldose hexose galactose at one end, which has the same stereochemistry on carbons 2–4 as l-arabinose. However, this fusion was induced only 2-fold by galactose.

The other pentose-induced systems include ABC-T SMb20854 induced by deoxyribose (7-fold) and ABC-T SMa0067, which is induced 4-fold by both ribose and mannose (a hexose).

Within the carbohydrate-induced systems, the second largest group responds to hexoses, which may reflect the importance of uptake of these carbon sources by S. meliloti. This observation concurs with phylogenetic classification that the largest groups of ABC transporters in S. meliloti are CUT2 and CUT1 (37 and 26 members, respectively) (Table 8). ABC-T SMa0203 (CUT2) was primarily induced by galactose (19-fold), [but also by l- and d-arabinose (pentose sugars) (4- to 7-fold), l- and d-fucose (4- to 8-fold) and glucose (4-fold)]. A second system was induced by galactose: ABC-T SMb21345, albeit to a lesser extent (6-fold) than ABC-T SMa0203. Tagatose induced ABC-T SMa2305 (CUT1) (5-fold) and ABC-T SMb21421 (CUT2) (3-fold). Directly upstream of SMb21421 and probably part of the same transcription unit is a gene involved in sugar metabolism; SMb21420 encodes a putative l-arabinose isomerase. ABC-T SMb20410 was induced by mannose and dulcitol (a polyol) (both 3-fold) and ABC-T SMb20904 was induced by mannose, l-lyxose (a pentose), glucose, and sorbose (3- to 8-fold).

The one discrepancy between the results here and in the literature is the Frc system. Lambert et al. (17) showed that frcC (SMc02170 integral membrane domain) was induced by fructose and mannose. However, in this study both plasmid and integrated fusions to frcB (SMc02171 SBP), which is upstream of frcC, were constitutive. These data suggest that the 352-bp and 285-bp intergenic regions before frcB and frcC, respectively, contain separate promoters, and only frcC is inducible by fructose and mannose. Indeed, consensus sequences for −35 and −10 sequences of S. meliloti (18) were found in these intergenic regions.

Disaccharide-induced systems identified include the trehalose transporter system ABC-T SMb20325 (thuE), which was induced 7-fold by trehalose, as shown by an integrated fusion to SMb20328 (Table 2). A plasmid fusion to SMb20325 was constitutive, but a fusion containing the upstream regulator (SMb20324) and the proposed promoter of SMb20325 was induced 6-fold by trehalose. This result agrees with Jensen et al. (19), who found that although the Thu system of S. meliloti transported the α-glucosides trehalose and maltose, it was only induced by trehalose. In contrast, the S. meliloti α-glucoside transporter, Agl (20), transports sucrose, maltose, and trehalose and was induced by sucrose and to a lesser extent trehalose (19). In our hands, a reporter gene fusion integrated into aglE, (SMc03061) encoding the SBP, was induced by sucrose and maltose but also by the previously undescribed inducers maltotriose (Glc1-4Glc1-4Glc) and turanose (Glc3-4Fruc) (Table 2), with the caveat that the trisaccharide may either contain, or be converted to, a disaccharide. Both of these compounds are also α-glucosides. The lack of observed induction of SMc03061 by trehalose may have been due to a lack of sensitivity.

ABC-T SMc04393 was induced nearly 10-fold by dextrin, a polyd-glucoside of indeterminate length formed during the hydrolytic breakdown of starch (Table 2). The characterized cyclodextrin transporter from Klebsiella oxytoca CymDEFG (21) shows limited identity (26–45%) with ABC-T SMc04393.

A single system was induced by β-glucosides. An integrated-fusion to ABC-T SMc04259 is induced by gentiobiose (Glcβ1-6Glc) (19-fold), salicin (Glcβ1-aromatic ring) (6-fold), and cellobiose (Glcβ1-4Glc) (6-fold). A plasmid clone containing the upstream gene encoding a putatitive LacI-type regulator (SMc04260), and the reporter gene fused to SMc04259 showed 8-fold induction in the presence of gentibiose. A clone lacking the whole of SMc04260 was expressed constitutively at a high level. The previously described bacterial β-glucoside transporters of Thermotoga maritima CbtABCDF and BglpABCDF (22) are in the PepT family, whereas ABC-T SMc04259 is classified as CUT1, and there is limited sequence identity (<30%) between these members of different classes.

As illustrated above, the nature of the chemical linkage between subunits is crucial in the specific recognition and transport of sugar oligomers. One characterized system, the S. meliloti α-galactoside (Agp) transporter (23, 24) (ABC-T SMb21647), was induced by α-galactosides: melibiose (67-fold) and raffinose (25-fold), and also by galactose (18-fold) in the plasmid system (Table 2). The promoter cloned lies in front of SMb21648 (agaL1 or melA), a melibiase, which is upstream of agpA (SMb21647) encoding the SBP. A plasmid reporter fusion to the region upstream of SMb21647 was not expressed under any of the conditions tested, so it is likely that expression of all of the transporter components is under the control of the promoter of melA. Bringhurst et al. (24) found that a reporter gene fused to melA was induced by galactose and α-galactosides melibiose and raffinose, and (to a lesser extent) by β-galactosides lactose and lactulose. We found that an integrated reporter gene in SMb21644, encoding an ABC subunit, and downstream of SMb21647 was induced, not only by raffinose and galactose, but by the tetra-saccharide (α-galactoside) stachyose (Galα1-6Galα1-6Gluα1-β2Fruf), and by the galactose derivatives, galactosamine and dulcitol (galactitol). This is consistent with the results of Bringhurst et al. (24) and Gage and Long (23).

We found a previously undescribed second system, ABC-T SMb20931, which was induced not only by α-galactosides raffinose and melibiose, but also by the β-galactosides lactose and lactulose. Only slight induction was seen with galactose (1- to 2-fold) (Table 2). ABC-T SMb20931 belongs to the CUT2 family and the SBP shows little sequence identity with that of ABC-T SMb21647 (classified as a PepT). The cluster of genes making up ABC-T SMb20931 is located close to a two-component sensor regulator and genes concerned with export and polysaccharide degradation.

ABC-T SMb21652 (homologous to Lac of Agrobacterium radiobacter; ref. 25) was induced by the β-galactosides lactose (Galβ1–4Glu) and the closely related lactulose (Galβ1-4Fruf).

Induction of transport systems by polyols is summarized in Table 1. ABC-T SMb20316 was induced by erythritol (3-fold) and is located in a 90-kb region that contains seven transport systems.

Two systems were induced by myo-inositol. The first is ABC-T SMb20712, the SBP of which has high identity (77%) to the rhizopine SBP of S. meliloti L5-30 (MocB) (26), which is part of a cluster of catabolic genes not present in strain 1021. However, the proteins encoded by the whole ABC-T SMb20712 operon have high identity (74–85%) to the myo-inositol transporter from R. leguminosarum (IntABC) (2, 27). The second SBP induced by myo-inositol (SMb20072) has high identity to SMb20712 (54%), MocB (52%), and IntA (60%) but is not part of an obvious ABC operon.

ABC-T SMc01628, as shown by the integrated fusion SMc01624, was induced by xylitol (C5, 15-fold), adonitol (C5, 12-fold), erythritol (C4, 6-fold) and sorbitol (C6, 6-fold) and is clustered with genes encoding proteins with high identity to those for metabolism of sugars and polyols, such as erythritol kinase (SMc01623).

Two systems were primarily induced by dulcitol. ABC-T SMb21377 (CUT2) was induced by dulcitol (C6, galactitol) but also by C6 monosaccharides; galactose (an aldose), tagatose and sorbose (both ketoses), and, to a lesser extent, the C5 aldose, l-lyxose. ABC-T SMc01496 (CUT1) was induced by the C6 polyols dulcitol, sorbitol, mannitol, and maltitol (Gluα1-4sorbitol) but not by sugars. It has high identity to the Smo system of Rhodobacter sphaeroides (28), with genes encoding a putative sorbitol dehydrogenase (smoS; SMc01500) and mannitol dehydrogenase (mtlK; SMc01501) downstream of SMc01496.

Three systems were induced by sugar amines, ABC-T SMb21135 and ABC-T SMb21221 by both galactosamine and glucosamine, whereas ABC-T SMb21151 was induced weakly by galactosamine. The three systems are tightly clustered on pSymB but are in two different families: ABC-T SMb21135 is a PAAT, whereas ABC-T SMb21221 and ABC-T SMb21151 belong to the CUT1 family. The binding of sugar amines may be shared by these sugar and amino acid ABC-T families.

Three systems were induced by sugar phosphates. Two of them, ABC-T SMa1427 and ABC-T SMb21273 (as reported by a plasmid fusion to SMb21272), were induced by both glucose-6-phosphate and glycerol-3-phosphate. A plasmid fusion to the region upstream of SMb21273 was not expressed, implying that the region upstream of SMb21272 contains the operon promoter. SMb21272 encodes a LacI-type regulator immediately upstream of SMb21273 (annotated as potD), which has significant identity to a putrescine and spermidine SBP from A. tumefaciens. The third system, ABC-T SMc02514, was induced by glycerol-3-phosphate and glycerol (as measured with a plasmid fusion to SMc02520 and an integrated fusion to SMc02519). SMc02520 (glpD) encodes a glycerol-3-phosphate dehydrogenase. ABC-T SMb20416 has high identity to the glycerol-3-phosphate (UgpBAEC) system in E. coli (29) but was constitutively expressed, as measured by a plasmid fusion.

Induction by Amino Acids and Their Derivatives.

Eighteen systems were induced in response to either one or more defined amino acids or by CAS amino acids (Table 1). ABC-T-SMb21526, which is homologous to the taurine transporter (TauABCD) of E. coli (30), was induced by taurine >100-fold by a plasmid fusion and 30-fold by an integrated fusion. The known S. meliloti ectoine transporter ABC-T SMb20428 (31) was induced at least 3-fold by ectoine, although it was highly expressed under noninducing conditions. The characterized S. meliloti histidine-transporter (Hut) ABC-T SMc00672 was induced 3-fold by histidine as measured by an integrated fusion to SMc00672 (Table 2). A plasmid fusion to SMc00672 was inactive; however, a plasmid fusion to the upstream gene (SMc00673), encoding a putative hydrolase, was induced 4-fold by histidine, suggesting that the promoter lies in front of SMc00673. The reported 5-fold induction of a LacZ integrated fusion to the SBP SMc00672 (hutX) (32) also is likely to be from a promoter upstream of SMc00673. Furthermore, when the upstream regulator (SMc00674) was cloned along with the promoter region of SMc00673 in a plasmid fusion, the induction by histidine was increased to 9-fold (Table 2).

ABC-T SMb21097 was induced by citrulline as shown by an integrated fusion (Table 2). It is unlikely that the promoter lies directly upstream of SMb21097, because a plasmid fusion to this region was not induced under any condition tested. Sequence analysis indicates that a promoter may lie upstream of SMb21094, encoding a putative l-arginosuccinate lyase (argH2). Citrulline, l-arginosuccinate, and arginine are three successive structurally related compounds in the urea cycle. No ABC transporter of citrulline has been documented.

ABC-T SMa0104 was strongly induced by the basic amino acids lysine, ectoine, and asparagine. ABC-T SMa0104 was reported in an array study of S. meliloti to be strongly induced on tryptone yeast (TY) medium, presumably due to the presence of such basic amino acids (10).

Three systems were induced by methionine: ABC-T SMb21196, ABC-T SMc00078 and ABC-T SMc03121. ABC-T SMb21196 has high sequence identity to an oligopeptide transport system in Bacillus subtilis (Opp) (33), whereas SMc00078 is an orphan SBP. SMc03121 is located close on the S. meliloti chromosome to ABC-T SMc03124 (induced by arginine) and ABC-T SMc03131 (induced by d-l-2-aminoadipic acid). None of the systems induced by methionine are in the MUT family of methionine uptake transporters of which there is probably a single member in S. meliloti, ABC-T SMc03157.

TRAP-T SMb20320 was induced 7-fold by hydroxyproline. There are no reports of a TRAP system transporting this amino acid. ABC-T SMc02219 was induced 2- to 3-fold by leucine, isoleucine, valine, hydroxyproline, and homoserine. ABC-T SMc02832 was induced by taurine, leucine, isoleucine, and valine (2- to 3-fold).

There are five systems that were induced by CAS amino acids but for which no clear single amino acid tested was an inducer, although ABC-T SMc00786, an ortholog of the Dpp dipeptide transport system of R. leguminosarum (34), also was induced by hydroxyproline (2-fold).

Induction by Purine and Pyrimidine Derivatives.

There are no known fully characterized ABC uptake systems for this group of compounds. ABC-T SMb20127 was induced by the purines xanthine and xanthosine (xanthine riboside) (Table 1). ABC-T SMc02415 was induced by the purine derivative, allantoin. This is an interesting gene cluster because it has two genes for SBPs (SMc02415 and SMc02417) interspersed with genes encoding two hydrolases and cytosine aminase. The SBPs show high identity (57%), suggesting that they have arisen from gene duplication. If these SBPs do interact with the same membrane components (SMc02418, SMc02419, SMc02423, and SMc02424), then it is plausible that they may transport related, but different, purine derivatives.

ABC-T SMc01827 was induced by the pyrimidine uracil (7-fold) and its derivative, uridine (uracil riboside) (3-fold). Upstream of the cluster encoding the components of the transporter are genes for a dihydropyrimidinase (SMc01821) and an N-carbamyl l-amino acid amidohydrolase (SMc01820), suggesting the whole region is involved in the metabolism of pyrimidine derivatives.

Induction by Quaternary Amines and Putrescine.

The quaternary amine choline not only supports growth of S. meliloti as a sole source of carbon, but is also a precursor of both phosphotidyl choline, an important lipid component in the organism, and the osmoprotectant glycine betaine. The identified choline transporter in S. meliloti, ChoXWV (ABC-T SMc02737) belongs to the QAT class, and its transport was reported to be induced 6-fold by choline (35). An integrated fusion to SMc02737 was induced 2-fold by choline (Table 2). Dupont et al. (35) suggested that there are other transport systems in S. meliloti for choline, because a choX mutant still has significant choline transport. ABC-T SMb20570 and ABC-T SMc02344 (both NitT systems) were induced by choline and may be these transport systems (Table 1).

Finally, a single transporter, ABC-T SMc01652, was induced by the polyamines putrescine (6-fold) and agmatine (3-fold), in accord with its classification as a member of the POPT family.

Induction by Organic Acids.

Of the eight systems induced by organic acids, five of them are TRAP-Ts. Of these, two were induced by solutes not previously associated by expression, binding, or transport studies with TRAP-T systems. TRAP-T SMa0157 was induced by the C3-dicarboxylic acid, malonic acid, and TRAP-T SMb20036, was induced by quinic acid (Table 1). Contiguous with the gene encoding the SBP in TRAP-T SMb20036 is a gene encoding a putative shikimate dehdrogenase (SMb20037), which is part of the quinic acid degradation pathway. This colocalization adds weight to the conjecture that this region might be involved in the transport and metabolism of quinic acid and/or its derivatives.

TRAP-T SMc00265 and TRAP-T SMb21438 were induced by 2-oxobutyric acid and 3-methyl-2-oxovaleric acid. In Rhodobacter capsulatus, the TRAP-SBP RRC01191 has been shown to bind a range of oxo-acids including pyruvate and 2-oxobutyric acid (36). TRAP-T SMb21353 was induced by pyruvic acid (11-fold) and methyl-pyruvic acid (3-fold). Rather confusingly, TRAP-T SMb21353 was named DctPQM after the first characterized TRAP-T system from R. capsulatus (37), despite the two having only modest sequence identity (13–34%).

The three ABC-T systems that are induced by organic acids may be particularly novel because there are no characterized ABC import systems for organic acids. ABC-T SMb20568 was induced by the two closely related aromatic carboxylic acids protocatechuic acid (7-fold) and p-hydroxybenzoic acid (3-fold), as well as by quinic acid (2-fold). ABC-T SMb20568 is classified as a HAAT, a family that normally transports hydrophobic amino acids, and binding of similar hydrophobic aromatic compound would not be inconsistent with this classification. ABC-T SMb20144 was induced by glycolic acid (10-fold) and is classified as a PepT. ABC-T SMc02021, classified as a CUT2, was induced by butyric acid (5-fold) and valeric acid (4- fold); however, the absolute activities were low, indicating this might be a nonspecific effect.

Induction by Nutrient Limitation.

Thirty-eight plasmid fusions were induced >3-fold by TY medium and not by any other condition tested (Tables 4 and 5). Although many of these fusions may be induced by specific compounds present in TY, such a rich medium will chelate micronutrients and metals, causing their limitation. Ten fusions were shown to be induced by the lack of specific trace elements when omitted from minimal medium. These systems include four with high amino acid sequence identity to SBPs of characterized systems from other organisms. ABC-T SMc02509 was induced by both manganese and iron limitation (≈110-fold) and is the identified manganese transporter (Sit) from S. meliloti (38). ABC-T SMc03869 was induced by thiamine limitation (11-fold) and has 57% identity with the characterized thiamine transporter (Thi) from S. typhimurium (39). ABC-T SMc03196, induced 12-fold by molybdate limitation, has 58% identity with the characterized molybdate transporter (Mod) from B. japonicum (40). Finally, ABC-T SMc04245 induced 37-fold by zinc limitation has 33% identity with the characterized zinc transporter (Znu) from E. coli (41).

In addition, ABC-T SMa1746 (FeCT class) was induced by iron limitation (4-fold), and ABC-T SMb21133 (SulT class) was induced by sulfur limitation (5-fold). Finally, three systems ABC-T SMa0585, ABC-T SMb20605, and ABC-T SMb21114 were induced by nitrogen-limitation and one characterized system ABC-T SMb21176 (PhoCDET) by phosphate limitation (42).


Overall the strategy used here has generated remarkably rich information concerning ABC-T and TRAP-T systems and should be applicable to orthologous systems in many other organisms. Specific inducers for 76 transport systems were identified, and in most cases, the inducers will reflect the transported solutes. Although a large range of compounds were tested, this selection represents a limited subset of biologically active compounds. Indeed, 28 plasmid fusions were induced only on TY with no specific inducer or limitation condition identified. In future, the fusion libraries can be tested against specific subsets of compounds of interest. Although detailed mutational and transport studies are required to verify and characterize the transport systems, the data in this report is a significant step toward determining the substrate specificity of ABC- and TRAP-uptake systems. It is also apparent that high-throughput screening of plasmid and integrated fusions for induction is a powerful tool for systems biology approaches to understanding interacting networks. Although extraordinarily powerful tools now exist for determining protein interactions (43, 44), we are less knowledgeable regarding which ligands induce and bind to these complexes. This problem is particularly acute for systems induced in response to complex environmental cues. The method described here provides a way to integrate these approaches with systems biology.

Materials and Methods

Bacterial Strains and Growth Conditions.

The bacterial strains and plasmids used in this study are listed in Tables 9 and 10, which are published as supporting information on the PNAS web site. S. meliloti strains were grown either in acid minimal salts (45) with the EDTA lowered to 1 μM or M9 minimal media with MgSO4 (1.0 mM), CaCl2 (0.25 mM), biotin (0.3 mg/ml), and cobalt chloride (10 ng/ml). Both media were supplemented with test compounds as described in Table 3. When required, glycerol (5 mM) or pyruvate (30 mM) was used as a background carbon source as generally neither causes significant catabolite repression. NH4Cl (10 mM) was used as a nitrogen source when required.

Generation of Reporter Gene Fusions.

The multicopy replicating vector pRU1097 (46) containing gfp mut 3.1 was used for cloning either in the BD In-Fusion system (Palo Alto, CA) or adapted by Invitrogen (Carlsbad, CA) to produce pRU1097/d-TOPO (by covalently attaching topoisomerase at the HindIII site of the vector and adding a GTGG extension for directional cloning). PCR products (0.5–1 kb) containing putative promoter regions of target genes were cloned into pRU1097 or pRU1097/d-TOPO, according to the manufacturers' instructions. PCR primers for d-TOPO cloning had a CACC at the 5′ end of the forward primer and a HindIII site at the 5′ end of the reverse primer (Table 11, which is published as supporting information on the PNAS web site). PCR primers for BD cloning had 20-bp overhangs on both primers complementary to pRU1097. All plasmids were conjugated into S. meliloti as described by Simon et al. (47).

The majority of the integrated transcriptional fusion strains were obtained from the S. meliloti random fusion library that carries gene fusions to lacZ/gfp and gusA/tdimer (see www.sinorhizobium.org) (48). To obtain a complete set of transport gene fusions, additional integrated fusions were made by cloning 250- to 500-bp fragments from the 3′ regions of putative ABC transport clusters into the suicide-plasmid pTH1360 generating gusA fusions (49). The plasmids were conjugated into S. meliloti and recombined into the genome.

Reporter Gene Assays.

S. meliloti strains were grown in 48-well (Greiner) or 96-well (Co-Star) plates, in triplicate, and shaken at 280 rpm at 26°C for 1–3 days until mid-log growth stage. Gfp fluorescence was measured by using a Tecan GENios fluorometer (excitation 485 nm, emission 510 nm) (46). β-Glucuronidase and β-galactosidase assays were performed according to Cowie et al. (48).

Supplementary Material

Supporting Tables:


We thank the Biotechnology and Biological Sciences Research Council U.K. and Genome Canada through the Ontario Genomics Institute and Ontario Research and Development Challenge Fund and Natural Sciences and Engineering Research Council for funding this work.


ABCATP-binding cassette
ABC-TABC transporter
SBPsolute-binding protein
TRAP-Ttripartite ATP-independent periplasmic transporter
TYtryptone yeast.


The authors declare no conflict of interest.

This article is a PNAS direct submission.


1. Galibert F, Finan TM, Long SR, Puhler A, Abola P, Ampe F, Barloy-Hubler F, Barnett MJ, Becker A, Boistard P, et al. Science. 2001;293:668–672. [PubMed]
2. Young JP, Crossman L, Johnston A, Thomson N, Ghazoui Z, Hull K, Wexler M, Curson A, Todd J, Poole P, et al. Genome Biol. 2006;7:R34. [PMC free article] [PubMed]
3. Kaneko T, Nakamura Y, Sato S, Asamizu E, Kato T, Sasamoto S, Watanabe A, Idesawa K, Ishikawa A, Kawashima K, et al. DNA Res. 2000;7:331–338. [PubMed]
4. Kaneko T, Nakamura Y, Sato S, Minamisawa K, Uchiumi T, Sasamoto S, Watanabe A, Idesawa K, Iriguchi M, Kawashima K, et al. DNA Res. 2002;9:189–197. [PubMed]
5. Davidson AL, Chen J. Annu Rev Biochem. 2004;73:241–268. [PubMed]
6. Kelly DJ, Thomas GH. FEMS Microbiol Rev. 2001;25:405–424. [PubMed]
7. Blattner FR, Plunkett G III, Bloch CA, Perna NT, Burland V, Riley M, Collado-Vides J, Glasner JD, Rode CK, Mayhew GF, et al. Science. 1997;277:1453–1462. [PubMed]
8. Stover CK, Pham XQ, Erwin AL, Mizoguchi SD, Warrener P, Hickey MJ, Brinkman FSL, Hufnagle WO, Kowalik DJ, Lagrou M, et al. Nature. 2001;406:959–964. [PubMed]
9. Ren Q, Paulsen IT. PLoS Comput Biol. 2005;1:e27. [PMC free article] [PubMed]
10. Becker A, Berges H, Krol E, Bruand C, Ruberg S, Capela D, Lauber E, Meilhoc E, Ampe F, de Bruijn FJ, et al. Mol Plant–Microbe Interact. 2004;17:292–303. [PubMed]
11. Barnett MJ, Tolman CJ, Fisher RF, Long SR. Proc Natl Acad Sci USA. 2004;101:16636–16641. [PMC free article] [PubMed]
12. Djordjevic MA, Chen HC, Natera S, Van Noorden G, Menzel C, Taylor S, Renard C, Geiger O, Weiller GF. Mol Plant–Microbe Interact. 2003;16:508–524. [PubMed]
13. Uchiumi T, Ohwada T, Itakura M, Mitsui H, Nukui N, Dawadi P, Kaneko T, Tabata S, Yokoyama T, Tejima K, et al. J Bacteriol. 2004;186:2439–2448. [PMC free article] [PubMed]
14. Hosie AHF, Allaway D, Jones MA, Walshaw DL, Johnston AWB, Poole PS. Mol Microbiol. 2001;40:1449–1459. [PubMed]
15. Saier MH. Microbiol Mol Biol Rev. 2000;64:354–411. [PMC free article] [PubMed]
16. Richardson JS, Hynes MF, Oresnik IJ. J Bacteriol. 2004;186:8433–8442. [PMC free article] [PubMed]
17. Lambert A, Osteras M, Mandon K, Poggi MC, Le Rudulier D. J Bacteriol. 2001;183:4709–4717. [PMC free article] [PubMed]
18. MacLellan SR, MacLean AM, Finan TM. Microbiology. 2006;152:1751–1763. [PubMed]
19. Jensen JB, Peters NK, Bhuvaneswari TV. J Bacteriol. 2002;184:2978–2986. [PMC free article] [PubMed]
20. Willis LB, Walker GC. J Bacteriol. 1999;181:4176–4184. [PMC free article] [PubMed]
21. Pajatsch M, Gerhart M, Peist R, Horlacher R, Boos W, Bock A. J Bacteriol. 1998;180:2630–2635. [PMC free article] [PubMed]
22. Nelson KE, Clayton RA, Gill SR, Gwinn ML, Dodson RJ, Haft DH, Hickey EK, Peterson LD, Nelson WC, Ketchum KA, et al. Nature. 1999;399:323–329. [PubMed]
23. Gage DJ, Long SR. J Bacteriol. 1998;180:5739–5748. [PMC free article] [PubMed]
24. Bringhurst RM, Cardon ZG, Gage DJ. Proc Natl Acad Sci USA. 2001;98:4540–4545. [PMC free article] [PubMed]
25. Williams SG, Greenwood JA, Jones CW. Mol Microbiol. 1992;13:1755–1768. [PubMed]
26. Rossbach S, Kulpa DA, Rossbach U, Debruijn FJ. Mol Gen Genet. 1994;245:11–24. [PubMed]
27. Fry J, Wood M, Poole PS. Mol Plant–Microbe Interact. 2001;14:1016–1025. [PubMed]
28. Stein MA, Schafer A, Giffhorn F. J Bacteriol. 1997;179:6335–6340. [PMC free article] [PubMed]
29. Overduin P, Boos W, Tommassen J. Mol Microbiol. 1988;2:767–775. [PubMed]
30. van der Ploeg JR, Weiss MA, Saller E, Nashimoto H, Saito N, Kertesz MA, Leisinger T. J Bacteriol. 1996;178:5438–5446. [PMC free article] [PubMed]
31. Jebbar M, Sohn-Bosser L, Bremer E, Bernard T, Blanco C. J Bacteriol. 2005;187:1293–1304. [PMC free article] [PubMed]
32. Boncompagni E, Dupont L, Mignot T, Osteras M, Lambert A, Poggi MC, Le Rudulier D. J Bacteriol. 2000;182:3717–3725. [PMC free article] [PubMed]
33. Perego M, Higgins CF, Pearce SR, Gallagher MP, Hoch JA. Mol Microbiol. 1991;5:173–185. [PubMed]
34. Carter RA, Yeoman KH, Klein A, Hosie AHF, Sawers G, Poole PS, Johnston AWB. Mol Plant–Microbe Interact. 2002;15:69–74. [PubMed]
35. Dupont L, Garcia I, Poggi MC, Alloing G, Mandon K, Le Rudulier D. J Bacteriol. 2004;186:5988–5996. [PMC free article] [PubMed]
36. Thomas GH, Southworth T, Leon-Kempis MR, Leech A, Kelly DJ. Microbiology. 2006;152:187–198. [PubMed]
37. Hamblin MJ, Shaw JG, Curson JP, Kelly DJ. Mol Microbiol. 1990;4:1567–1574. [PubMed]
38. Platero RA, Jaureguy M, Battistoni FJ, Fabiano ER. FEMS Microbiol Lett. 2003;218:65–70. [PubMed]
39. Webb E, Claas K, Downs D. J Biol Chem. 1998;273:8946–8950. [PubMed]
40. Delgado MJ, Tresierra-Ayala A, Talbi C, Bedmar EJ. Microbiology. 2006;152:199–207. [PubMed]
41. Patzer SI, Hantke K. Mol Microbiol. 1998;28:1199–1210. [PubMed]
42. Bardin S, Dan S, Osteras M, Finan TM. J Bacteriol. 1996;178:4540–4547. [PMC free article] [PubMed]
43. Butland G, Peregrin-Alvarez JM, Li J, Yang WH, Yang XC, Canadien V, Starostine A, Richards D, Beattie B, Krogan N, et al. Nature. 2005;433:531–537. [PubMed]
44. Krogan NJ, Cagney G, Yu HY, Zhong GQ, Guo XH, Ignatchenko A, Li J, Pu SY, Datta N, Tikuisis AP, et al. Nature. 2006;440:637–643. [PubMed]
45. Poole PS, Blyth A, Reid CJ, Walters K. Microbiology. 1994;140:2787–2795.
46. Karunakaran R, Mauchline TH, Hosie AHF, Poole PS. Microbiology. 2005;151:3249–3256. [PubMed]
47. Simon R, Priefer U, Pühler A. Biotechnology. 1983;1:784–791.
48. Cowie A, Cheng J, Sibley CD, Fing Y, Zaheer R, Patten CL, Morton RM, Golding B, Finan TM. Appl Environ Microbiol. 2006;72:7156–7167. [PMC free article] [PubMed]
49. Yuan ZC, Zaheer R, Morton R, Finan TM. Nucleic Acids Res. 2006;34:2686–2697. [PMC free article] [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...