MOLECULAR AND FUNCTIONAL CHARACTERIZATION OF THE lpdA1 GENE IN S. meliloti

any microorganisms, such as Escherichia coli, have a single gene encoding dihydrolipoamide dehydrogenase, which can function as the E3 component of different multi-enzyme dehydrogenase complexes involved in carbon metabolism and energy production. In contrast, the genome of the symbiotic N2fixing soil bacterium Sinorhizobium meliloti encodes three lpdA alleles. Based on their position relative to genes encoding the other complex subunits, each of the lpdA alleles are predicted to function in a different enzyme complex. The lpdA1 is presumed to encode the E3 component of pyruvate dehydrogenase (PDH); lpdA2, the E3 component of alpha-ketoglutarate dehydrogenase (KGD); and lpdA3, the E3 subunit of a branched-chain alphaketoacid dehydrogenase (BKD). To date, no functional characterization of these lpdA genes has been done in S. meliloti. Analysis of the LpdA amino acid sequences revealed conserved functional domains, suggesting that the three S. meliloti lpdA alleles encode functional proteins and that each may be specific to the complex encoded by the adjacent genes. To test this hypothesis, insertion mutation was isolated for the lpdA1 allele. Internal fragment of lpdA1 allele cloned into plasmid pVIK112 recombined into the S. meliloti genome by single crossover yielded lpdA1 mutant. Driscoll and Finan (2006)

any microorganisms, such as Escherichia coli, have a single gene encoding dihydrolipoamide dehydrogenase, which can function as the E3 component of different multi-enzyme dehydrogenase complexes involved in carbon metabolism and energy production.In contrast, the genome of the symbiotic N 2fixing soil bacterium Sinorhizobium meliloti encodes three lpdA alleles.Based on their position relative to genes encoding the other complex subunits, each of the lpdA alleles are predicted to function in a different enzyme complex.The lpdA1 is presumed to encode the E3 component of pyruvate dehydrogenase (PDH); lpdA2, the E3 component of alpha-ketoglutarate dehydrogenase (KGD); and lpdA3, the E3 subunit of a branched-chain alphaketoacid dehydrogenase (BKD).To date, no functional characterization of these lpdA genes has been done in S. meliloti.Analysis of the LpdA amino acid sequences revealed conserved functional domains, suggesting that the three S. meliloti lpdA alleles encode functional proteins and that each may be specific to the complex encoded by the adjacent genes.To test this hypothesis, insertion mutation was isolated for the lpdA1 allele.
Internal fragment of lpdA1 allele cloned into plasmid pVIK112 recombined into the S. meliloti genome by single crossover yielded lpdA1 mutant.Driscoll and Finan (2006)

Bioinformatics analysis
The sequence analysis was performed using several tools available at the GenBank (http://www.ncbi.nlm.nih.gov/genbank)notably the BLAST suite (blastn and blastp), using non-redundant nucleotide M and protein sequences (nr) database.The conserved domain investigation was done using the GenBank CD-search, as well the fingerprint Scan (EMBL-EBI).The multiple sequence alignment analysis was done using the CLUSTALW program (Thompson et al., 1994) using Gonnet series for weight matrix for both pairwise and multiple alignment parameters.The rooted phylogram was made using CLUSTALW alignment and PHYLIP's DRAWTREE program (Phylip, 2000).The protein sequences used for the phylogenetic analysis were selected by searching the GenBank database for dihydrolipoamide dehydrogenases identified as being the part of the PDH complexes in bacteria.The identity and similarity between the lpdA genes sequences was estimated using the EM-BOSS global pairwise alignment algorithm NEEDLE (Rice et al., 2000), and the possibility of the presence of the transmembrane regions within the protein sequences was tested using the TMHMM program (Krogh et al., 2001).

Enzyme assay
In preparation for the enzyme assays, the cultures were grown in LB mc supplemented with the appropriate antibiotics.After centrifugation and washing of obtained cell pellets with the sterile saline (0.85% NaCl), the cell suspension was used to inoculate M9 arabinose (15 mM) and succinate (15 mM) medium, supplemented with 1% of the LB broth.The cultures were grown in minimal media to minimize the unspecific enzyme activities observed under control conditions with no substrate.Cell growth and preparation of cell-free sonicated extracts were preformed essentially as described (Finan et al., 1988).Cells from late-log phase cultures were washed twice the with 20 mM Tris pH 7.8, and 1 mM MgCl 2 , resuspended in 4 mL/g cells of sonication buffer containing 20 mM Tris pH 7.8, 1 mM MgCl 2 , 10% glycerol and 10 mM βmercaptoethanol, then disrupted by sonication.Protein concentration was determined by the Bradford method (Bradford, 1976), using the BioRad protein assay dye with bovine serum albumin as standard.

Molecular biology techniques
Standard techniques were used for alkaline extraction of plasmid DNA, digestion of DNA with restriction endonucleases, DNA ligations, transformation of CaCl 2 -competent E. coli cells, and agarose gel electrophoresis (Sambrook et al., 1989).DNA fragments were isolated from agarose gels using the QIAEX II Gel Extraction Kit (Qiagen, Mississauga, Ontario, Canada).Bacterial genomic DNA was extracted as previously described (Meade et al., 1982).Restriction endonuclease digested genomic DNA was transferred by Southern blot to positively charged nylon membranes as described by (Sambrook et al., 1989).Hybridization was detected using the colorimetric method nitroblue tetrazolium (NBT), and 5-bromo-4-chloro-3-indolylphosphate (BCIP).

Plant growth conditions
Alfalfa (Medicago sativa var.Iroquois) nodulation experiments in Leonard jars containing 1:1 (v/v) sand and vermiculite were done as previously described (Yarosh et al., 1989).S. meliloti strains were inoculated into a minimum of three pots with 10 seedlings each.Controls were inoculated with sterile distilled H 2 O alone.Growth chamber (Conviron Inc., Winnipeg, Manitoba, Canada) conditions were 16 h of day (25C) and 8 h of night (20C).Shoot dry weights were determined from plant tops cut at the stemroot interface 28 days after inoculation.Shoots from each pot were pooled and dried for 3 weeks at 50C prior to weighing.Percent nitrogen and carbon content analyses were done by applying 2-mg samples of ground-dried shoots to an NC2500 elemental analyzer (CE Instruments, Milan, Italy).

Bioinformatics analysis
Initial investigation of the possible functionality of the lpdA genes and the corresponding LpdA proteins of S. meliloti were done by analyzing their gene and protein sequences that are available at the GenBank database.Comparison of the lpdA sequences using NEEDLE showed 38.7% identity and 56.0%similarity between lpdA1 and lpdA2, 31.8%identity and 52.9% similarity between lpdA1 and lpdA3, 42.6% identity, and 60.4% similarity between lpdA2 and lpdA3 protein sequences.
The search of the Genbank nucleotide database using the blastn algorithm revealed significant matches (hi score and low e-value) with the gene sequences from numerous organisms for the lpdA1 gene sequence and lpdA gene(s) sequences from R. leguminosarum as plant-microbe associations more than lpdA gene(s) sequences from A. vinelandii as free living symbiosis organism, and E. coli as standard bacterial model (Fig. 1).Phylogenetic analysis of the functionally related sequences using the multiple sequence alignment (CLUSTALW) suggests that the proteins belonging to the same functional group are evolutionary closer to each other than to the other groups (Fig. 2).Interestingly, the lpdA of E. coli that is shared by both pyruvate dehydrogenase and α-keto glutarate dehydrogenase complexes is evolutionary closer to the LpdA1 and LpdA2 proteins than to the LpdA3.The S. meliloti mercuric reductase, MerA1, although in this case used as an out-group, also shares the common ancestry with pyruvate, and α-ketoglutarate dehydrogenase group.

Induction of the lpdA1 mutant
To examine the hypothesis that lpdA1 gene encodes a functional protein as dihydrolipoamide dehydrogenase E3 and to incorporate a reporter gene expressed from the lpdA1 promoters it was necessary to make mutant in S. meliloti that produce non-functional LpdA1 proteins and have integrated lacZ reporter gene.To make the lpdA1 mutant, PCR products of the amplification of the corresponding lpdA1 gene fragment of 263 bp long were inserted directly into the pVIK112 suicide vector (Table 2).The vectors were constructed to contain the forward orientated gene fragment insertion, with respect to the target gene and vector's reporter gene (lacZ) and kmr as antibiotic resistant marker.The incorporation of the modified vector DNA into S. meliloti G212 genome was performed using bacterial conjugal (triparental) mating during which a single cross-over recombination of the introduced plasmid to the genome of the recipient producing the lpdA1 mutants.Both resistance marker and expression of the lacZ were used to verify the correct insertion of the modified vectors into the S. meliloti RmG212 genome lpdA1 mutnat was confirmed by southern plot using DIG label system as shown in Fig. (3).

Growth phenotypes
The growth phenotypes of the mutant strains (Table 3) were determined by incubating cultures on solid M9 medium supplemented with acetate, arabinose, glucose, glutamate, malate, pyruvate, succinate, and leucine.The choice of carbon sources was based on predicted inability of mutants to metabolize certain carbon sources, availability and suitability of other compounds to be used for the growth evaluation.The effect of the mutation in the S. meliloti lpdA1 gene was evident from the total inability of the organism to grow on pyruvate as a sole carbon source, which was expected due to the block in the PDH.Similarly to the B. subtilis PDH mutants that were unable to grow on glucose (Gao and Jiang, 2002), and the PDH deficient P. aeruginosa (Jeyaseelan and Guest, 1980) that failed to grow on malate, succinate, glucose, and glutamate, the S. meliloti lpdA1 mutant was unable to metabolize those compounds growing on solid media.The inability to grow on pyruvate is different from the ability of the previously reported putative S. meliloti arylesterase mutant (Soto et al., 2001) in which the mutation presumably had an effect on lpdA1 transcription due to the polar effects of the Tn5 insertion.
In these complexes the LPDs act as the E3 subunits, catalyzing oxidation of dihydrolipoamide to lipoamide while reducing FAD + cofactor, and ultimately NAD + to NADH.LPD also participates in the glycine cleavage multienzyme system (GCV), in which it functions as the L protein (Wilson et al., 1993).The results in  4).

Gene expression study
The pVIK112 plasmid used for the site-targeted mutagenesis carries the promoterless lacZ gene that, once recombined with the targeted gene, creates a transcriptional lpdA1:lacZ fusion which can be used as a reporter for the examination of the expression of the lpdA1 gene.To test the potential up-or down-regulation of the lpdA1 gene the cultures were grown overnight in the minimal media supplemented with different carbon sources, and their effects on the transcription of the lpdA1 gene were measured indirectly through the activity of the β-galactosidase.Test tubes containing the LB broth, and M9 minimal media supplemented with 15 mM arabinose, glutamate, glucose, leucine, malate, pyruvate and succinate were chosen to take into account the ability of the mutants to grow in the presence of the particular carbon source.Particular interest was examining the effect of the possible lpdA1 inducers such as pyruvate, arabinose, glutamate, and leucine, which could provide an insight into the regulation of the PDH, KDH, and BKD enzyme complexes.The results (Table 3) indicate that the expression of lpdA1 was somewhat higher under the influence of the complex medium, arabinose, malate, and succinate.It seems that both glucose and pyruvate do not have direct effect on the expression of the PDH component.The lack of substantial growth on malate and succinate can be explained by the interruption in the anaplerotic reaction in which the acetyl-CoA is replenished through the malic enzyme conversion of malate to pyruvate, and subsequently by the action of the PDH to obtain acetyl-CoA from pyruvate.Obviously, the lpdA1 mutant's inactivated PDH is not able to complete the final conversion to the acetyl-CoA.The inability of the LpdA1 to metabolize glutamate is also not unusual, since glutamate enters the TCA cycle by the action of the glutamate synthase that converts glutamate to αketoglutarate, so the block in the PDH again affects the anaplerotic pathway in which the malate is used to provide the acetyl-CoA for the continuation of the TCA cycle.However, the ability to grow on arabinose is quite surprising.In rhizobia, arabinose can enter the TCA cycle through pyruvate in slow growing rhizobia, or α-ketoglutarate in fast growers, such as R. leguminosarum and S. meliloti (Duncan, 1979;Duncan and Fraenkel, 1979;Sarma and Emerich, 2006).If it is true that the arabinose enters the cycle as it was published, it is not clear why the mutant would grow better on arabinose than on the malate, succinate, and glutamate, especially knowing that it enters into the cycle at the same place as glutamate.Although much has been done to examine the ways arabinose is metabolized, some of the genes, notably arabinose dehydrogenase, and arabinolactonase, have yet to be identified in the S. meliloti genome (Poysti and Oresnik, 2007).
It appears that in the lpdA1 mutant rate limiting step in functioning of TCA cycle is the formation of the acetyl-CoA, as seen inability of the mutant to grow on acetate, leucine and beta hydroxybuyrate (data not shown) all of them entering the cycle as acetyl-CoA.Unlike the decreased activity of PDH in putative arylesterase mutant (Soto et al., 2001), that might have been the result of the "leaky" Tn5 trancription terminator, the activity of PDH in lpdA1 mutant was completely absent, similarly to the activity of PDH in B. subtilis LPD mutant (Gao and Jiang, 2002).The absence of detectable PDH activity, inability of the lpdA1 mutant to grow on pyruvate, and inability to fix atmospheric nitrogen strongly indicate that it is highly unlikely that any of the other two LpdA enzymes can substitute defective LpdA1 in S. meliloti.Contrary to the expression of the pdhA gene (Cabanes et al., 2000) in S. meliloti grown in M9 minimal media with succinate or pyruvate as the sole carbon sources, the investigation of the expression of the lpdA1 indicated that the transcription was increased in presence of succinate and not affected by pyruvate, under the same conditions.This finding gives more credibility to the hypothesis that lpdA1 is expressed from the promoter separate from the one that regulates the pdhABC cluster.

Plant assay
To test the ability of S. meliloti strains to form a functional symbiotic relationship with the plant hosts, the plant growth evaluation was performed by infecting alfalfa seedlings (Medicago sativa var.Iroquois) with selected bacterial strains.The average mass of the plants inoculated with the wild type RmG212 and the lpdA mutants varied greatly (Table 5).The lowest mass per plant was measured for the plantsinoculated with lpdA1 mutant.The average mass per plant was just above the uninoculated control, with several plants appearing Fix + (Fig. 4), however, analysis of variance (ANOVA, F (4, 10) = 147.793,P<0.001) indicated that the difference between the average plant weights of the plants inoculated with the lpdA1 mutant, and the controls was not statistically significant at P=0.05.This strain produced fewer number of pink root nodules, compared to the wild-type RmG212.The phenotype of the lpdA1 mutant was clearly Fix -four weeks postinoculation.
This study clearly confirmed the theoretical probability that the S. meliloti lpdA1 gene encode functional protein, established its association with PDH enzyme complexes.However further studies for lpdA2 and lpdA3 genes would need to examine possible substitution of inactive lpdA2 with any of the other two lpdA, or yet unknown functionally similar enzymes.It would be also interesting to investigate possibility of existence of an alternative arabinose metabolism pathway.

SUMMARY
The alfalfa-Sinorhizobium symbiosis is one of the best studied for plantmicrobe associations.The pyruvate dehydrogenase complex catalyzes the key metabolic step connecting the glycolysis to the Tricarboxylic acid (TCA) cycle.The PDH also participates in anaplerotic synthesis of acetyl-CoA in bacteroids, or when grown in the presence of dicarboxylic acids as the sole carbon sources.The complex consists of multiple copies of the three subunits, pyruvate dehydrogenase (EC 1.2.4.1), that acts as the E1 component of the complex, dihydrolipoamide transacetylase (EC 2.3.1.12),the E2 component, and dihydrolipoamide dehydrogenase (EC 1.8.1.4),that participates as the E3 subunit.The lpdA1 gene was mutated through the site-directed, single cross-over recombination, using modified pVIK112 suicide plasmids.The lpdA1 mutant failed to grow on pyruvate, and was significantly delayed on other tested carbon sources.The activity of the pyruvate dehydrogenase complex was not detected in this mutant, and the regulation of the lpdA1 was not dependent on the presence of pyruvate.This mutant was Nod+, but appeared Fix -.Yarosh, O. K., T. C. Charles and T. M. Finan (1989).Analysis of C4dicarboxylate transport genes in Rhizobium meliloti.
Mol. Microbiol., 3: 813-823.The wild-type strain, RmG212, was used as the reference strain and it was able to grow in the presence of every carbon source that was tested.The numbers after the ± sign represent the standard error of the means (n = 3).The plants were grown in groups of 10-12 plants per pot.RmG212 is the wild-type strain, Rm30282 is lpdA1 mutant, * ANOVA: F (4, 10) = 147.793,p<0.001.The average plant weights with the same letter in the superscript are not significantly different.

Table ( 2
): PCR primers used for the amplification of lpdA genes and their fragments, and for the synthesis of the Southern blot probes.WhG: the primer amplifies the whole gene; ∆: indicates that the primers amplify only a fragment of the gene Table (3): Plate phenotypes of the lpdA1 mutant in different carbon sources.
a FW: forward oriented primer; RW: reverse oriented primer;

Table ( 5
): Average weight of the individual plants per strain.