Genomic co-existence of three highly conserved genes coding for levansucrase is a feature unique to the plant pathogen P. syringae despite the fact that numerous other bacterial species harbor just a single copy of this gene in their genomes. Artificial expression of lscA from P. syringae under the control of the Plac had been shown previously
. The same study also showed that lscA could not be expressed under its own promoter. Major differences between lscA and the natively expressed genes lscB and lscC are not found in the coding sequences but in their upstream DNA regions. The upstream regions of lscB and lscC represent a possible PAPE
. We previously hypothesized that this PAPE might harbor regulatory sites required for expression of levansucrase and general sugar metabolism in P. syringae. Herein, the PAPE of lscB was fused to the coding sequence of lscA and thus proven for its transcriptional activity in P. syringae.
The nucleotide sequence of the predicted PAPE consists of two parts, the upstream region of lscB and the first 48-bp coding for the N-terminus of LscB. The importance of these 48-bp of the ORF for the expression was tested by generating fusion constructs of the upstream region and lscA with or without these coding nucleotides. Transformants carrying either of the two fusion constructs produced levan similar to the PG4180.M6 mutant complemented with lscB. Western blotting, zymographic detection, and qRT-PCR analyses confirmed these results but also allowed a more detailed view; native lscB and the lscB
A fusion had similar mRNA expression levels while that of the fusion lscB
A, which lacked the 48-bp of N-terminal LscB-coding region, had less. Consequently, one might speculate that although the -450 bp upstream DNA region of lscB, which includes the TSS as determined in this study, is sufficient for expression of lscA, the first 48-bp of the lscB ORF increase the level of its expression. Since our respective results of Western blotting and zymographic detection of Lsc activity were indistinguishable from each other, it could be concluded that the N-terminus of LscB might not be involved in altering of enzymatic activities.
A peculiar observation was the electrophoretic migration of the individual proteins or fusion proteins in polyacrylamide gels. The observed faster migration of LscBUpNA as compared to LscB under denaturing conditions could potentially be attributed to the apparent mass shift for two proteins with nearly identical molecular masses as described earlier
. Interestingly, the migration of LscBUpNA was significantly slower than that of LscB under native conditions. This finding might demonstrate that modest changes in the protein’s surface charge might result in significant alterations of electrophoretic mobility
[22, 27, 28].
Although the different migration rates of the proteins or fusion proteins under native or denaturing conditions suggested that the synthesized proteins were indeed different from each other, a MALDI-TOF analysis of each of the proteins was conducted using protein samples from zymograms. The produced levan surrounding the proteins did not seem to impact mass spectrometric analysis. The MASCOT score for each of the identified proteins was above the significance threshold of 100. The sample from the PG4180.M6(lscB) sample gave LscB from P. syringae pv. phaseolicola 1448A as the first significant match which was in line with the high homology of the respective genes in the close relatives pv. glycinea and pv. phaseolicola
. The sample from PG4180.M6(lscBUpA) which should synthesize only LscA gave the first significant match as LscA from P. syringae pv. glycinea race 4 strain. This proved that the lscB
A fusion actually synthesized an active LscA and confirmed earlier findings that artificial expression of LscA of PG4180 leads to levan formation
. Although the majority of obtained peptides for the sample representing LscBUpNA were LscA-borne as expected, the unique N-terminal 2,122-Da peptide NSPLASMSNINYAPTIWSR could be detected. This peptide is a consequence of the presence of the NheI restriction site coding for the amino acid residues alanine and serine. Oxidation of methionine, which was chosen as a variable modification parameter, added another 16 Da to the peptide mass which subsequently increased the mass of the NSPLASMSNINYAPTIWSR fragment to 2,138 Da. This mass was exactly the same as the mass of a recovered peptide which did not find a match during the NCBI search since the respective fusion peptide is not present in the database. Thus, the synthesis of the LscBUpNA fusion protein could also be proven.
The majority of previous LscA-related studies have been performed with P. syringae pv. glycinea PG4180
[9, 10, 23, 24]. However, thus far, there was no evidence for a lack of lscA expression in other pathovars of P. syringae. Since the genomes of P. syringae pv. phaseolicola 1448A, pv. syringae B728a and pv. tomato DC3000 are fully sequenced
[19–21], template-specific oligonucleotide primers for cDNA-based mRNA detection could be designed. Although mRNA samples were extracted during different growth stages, namely, early-logarithmic and late-logarithmic phase, no amplicons could be detected in any of the strains suggesting that lscA variants were not expressed. PCR amplification, using respective genomic DNA as template, proved that the primers were binding correctly. An independent gene, hexR, coding for a conserved hexose metabolism regulator protein HexR, was chosen to see if the total mRNA had been reverse transcribed correctly
. This PCR amplification gave correct sized amplicon of 880-bp for all the four strains demonstrating the accuracy of the used method. PCR amplification was also performed on the cDNA obtained from mRNA samples of PG4180.M6 containing lscA under the control of P
. This experiment gave the same-sized amplicon as for genomic DNA again proving the accuracy of the method.
In summary, we propose that lscA could be an ancestral Lsc variant in P. syringae as suggested by Srivastava et al.
. During evolution, the inactive promoter perhaps did not allow expression of lscA after this gene had potentially been introduced to an ancestral P. syringae. An evolutionary gene duplication of lscA followed by an insertion of a prophage-borne PAPE might have led to a new lsc variant, i.e. lscB which in turn got duplicated yielding lscC or vice-versa. As a result of this evolutionary process, two functional and expressed lsc genes emerged in the plant pathogen, for which utilization of sucrose, and perhaps levan formation, might be particularly important. The advantage of an additional in planta fitness-increasing and possibly virulence-promoting factor
 could have helped this organism to selectively establish itself as a potent plant pathogen. As a consequence of this hypothesis, one could speculate on a loss of the supposedly non-expressed lscA during further evolutionary steps, a phenomenon also previously hypothesized by Smits et al.