No surfactant production was observed in a ∆gacS mutant, and syfR and syfA transcription are nearly abolished. Additionally, we determined whether SyfR operates independently of SalA, or if SalA is upstream from SyfR function. Surprisingly, the ∆salA deletion mutant also exhibited abolished surfactant production and very low levels of syfR and syfA transcription. This suggests that these genes mediate the baseline expression of SyfR. It also suggests that these pathways are not responsible for the surface-dependent induction of syfA, but rather have an epistatic role in syringafactin production. We were surprised to observe that both ∆salA and ∆gacS mutants exhibited abolished surfactant production. We thus questioned if the strains are still capable of producing and secreting syringafactin or whether pleiotropic effects on cell metabolism that made them incapable of this metabolic process was operative in these mutants. To test this, we constitutively expressed SyfR by introducing plasmid p519n-syfR into both the ∆gacS and ∆salA mutant strains. These strains produced surfactant detectable both by the atomized oil spray as well as water drop collapse in both broth and plate cultures of these strains.
These mutants, in which syfR is not normally transcribed, thus produce and secrete syringafactin when SyfR is supplied constitutively. This demonstrates that neither GacS nor SalA are necessary for syringafactin transcription,flood and drain table or for supply of necessary intermediates for syringafactin production, but rather exert their influence on syringafactin production solely through their control of the SyfR regulator. We initially suspected that secondary mutations might have been the cause of what looked to be a poorly growing or stressed strain. However, all independently isolated syfA– and syfB transposon mutants as well as the site-directed ∆syfA deletion mutant displayed this same phenotype, suggesting that it is a direct response to the absence of syringafactin. Nonetheless, in order to rule out the possibility of secondary mutations in our syringafactin-deficient strains, we created a syfR– mutation in a ∆syfA deletion strain. Surprisingly, this second mutation abolished the fried egg phenotype normally exhibited by the ∆syfA mutant, and the appearance of this ∆syfA/syfR– double mutant was indistinguishable from that of a syfR– mutant alone. This suggested that SyfR transcriptionally regulates more genes than just those enabling syringafactin production, specifically including genes encoding whatever trait triggers development of the fried egg phenotype.
We initially hypothesized that perhaps SyfR induced both syringafactin production and also a system involved in either its transport or its perception, and we further postulated that syringafactin served not only as a surfactant, but also as a signaling molecule. To additionally support the conjecture that SyfR controls more than just syringafactin production, we introduced the plasmid conferring the constitutive expression of SyfR into a ∆syfA strain. Curiously, this strain also exhibited a fried appearance, but one that appeared to be an exaggerated and earlier-onset version of the rough “white” from a ∆syfA strain. Thus, the strong visual phenotype of this strain provides further evidence that SyfR transcriptionally regulates more than just syringafactin production. In comparison, a wild-type strain capable of syringafactin production and that constitutively expresses SyfR develops a slightly matted appearance, but does not exhibit a fried appearance. Assuming that this rough fried egg appearance is indeed indicative of stress, we might hypothesize that syringafactin normally plays a protective role for the cell, and that its absence makes the cells somehow more susceptible to other factors induced by SyfR in P. syringae itself. In order to identify genes under the control of SyfR, we screened 2,000 transposon mutants in a ∆syfA mutant background for any that had lost the fried egg phenotype. Several such mutants were identified. Prominent among the mutants found were several insertional events in the syringomycin biosynthetic genes and an associated secretion gene , prompting further investigation.
In order to confirm the requirement for syringomycin to initiate the fried egg phenotype, we constructed a site-directed knockout of syrE in a ∆syfA mutant background.One of the most surprising aspects of this finding was the fact that syringomycin and syringopeptin are assumed to have overlapping roles as plant virulence factors, and are typically co-regulated by SyrF which is downstream of SalA. However, a site-directed knockout in the syringopeptin biosynthetic gene sypA did not lead to a loss of the fried egg phenotype. These results strongly suggest that syringomycin has a specific role in this phenoptype that is independent from syringopeptin. However, while syringomycin appeared necessary for the fried egg phenotype, this was not proof that it was a factor regulated by SyfR, which was posited to be required for this phenomenon. We postulated that the production of syringomycin in the absence of syringafactin is altering cell physiology in a way that leads to a production of the fried egg phenotype. In order to confirm that syringomycin expression is under the control of SyfR, we developed a plasmid-based transcriptional reporter of syrB expression. Indeed, the GFP fluorescence indicative of syrB expression was much lower in cells of a syfR– mutant harboring pPsyrB-gfp than in either a ∆syfA mutant or the wild type strain; expression in the ∆syfA mutant was similar to that in the wild type strain. In broth cultures,rolling bench expression of syrB was similarly low in all strains, as further proof that genes downstream of SyfR are not activated in broth conditions. This confirms that syringomycin is induced in cells cultured on plates and is under the regulatory control of SyfR.Because the fried egg phenotype is observed only in colonies older than 3 days, we investigated syringomycin expression in different mutant strains over the course of several days.
Surprisingly, although syringomycin was highly expressed in wild-type and ∆syfA strains after one day of growth, only very low levels of expression were detected at any subsequent time. To further explore this apparent temporal regulation, we measured the GFP fluorescence of cells of a wild-type strain harboring pPsyrB-gfp over the course of 48 hours of growth on plates. In agreement with our initial observations, syrB expression, and thus presumably syringomycin production, is limited to a short period during initial phases of colony development, peaking after about 24 hours and thereafter diminishing. This pattern of expression was seen in both the wild-type and a ∆syfA mutant strain, but not in the syfR- mutant, in which syfA expression was always low. This temporal regulation of syrB is contrary to that of syfA expression, which remains stably induced over this time period. Thus the role of SyfR in stimulating syringomycin expression is distinct from its effect on syringafactin expression. We hypothesized that the strong temporal, and hence cell density-related, regulation of syringomycin synthesis may be due to its suppression by quorum sensing in older cultures. To test this we measured syrB expression in a ∆ahlR mutant incapable of quorum sensing. The temporal expression of syrB in this strain was identical to that in a wild type strain, with peak expression at 24 hours. This suggests that quorum sensing does not mediate temporal regulation of syringomycin production. Also, these results also cast doubt on the model that syringomycin is directly responsible for inducing the fried egg phenotype, since the colony phenotype appears after about four days of growth, while syringomycin production apparently peaks after only 24 hours. Although the biosynthetic pathway for syringomycin and regulation of its expression has been extensively investigated, SyfR has never been implicated in its regulation. We thus questioned if SyfR was, in fact, an overlooked necessary regulatory element for syringomycin production. To test this we measured syrB expression in a wild-type and a syfR– mutant mutant strain on media specifically formulated to induce syringomycin and syringopeptin production. Although the levels of GFP fluorescence exhibited by a syfR– mutant harboring pPsyrBgfp were reduced compared to that in the wild type strain, we still see substantial expression of syrB in the medium conducive to syringomycin production. This suggests to us that SyfR is not absolutely required for the induction of syringomycin production in this medium that mimics the plant environment, but plays a more ancillary role in its production. This moderate effect on syringomycin production might explain why SyfR has not previously been identified as a regulator of syringomycin. By chance, it was observed that the fried egg phenotype in colonies of ∆syfA mutants appeared much earlier when they were grown near colonies of a ∆syfA mutant blocked in any of various steps in the AlgT regulatory pathway. When colonies of a ∆syfA mutant were grown on the same plate with those of a ∆syfA/algT– mutant , the timing of the onset of the fried egg phenotype was directly correlated with the distance from the ∆syfA/algT– mutant. The fried egg phenotype was induced in ∆syfA mutants after as few as 2 days of incubation when cultured near a ∆syfA/algT– mutant.
When cultured by an algT– mutant that was still capable of syringafactin production, this early-onset property was diminished. Premature induction of the fried egg phenotype occurred only in a ∆syfA mutant, while colonies of wild type and syfR– strains did not change their appearance in response to this signal. This observation suggested that the fried egg phenotype must be a response to an extracellular compound that is only sensed by a component of the SyfR regulon. We earlier hypothesized that syringomycin is the compound that induces this fried egg phenotype. If syringomycin directly stresses the cell or otherwise induces this colony phenotype, and if syringomycin is produced in large quantities in an algT– mutant, then we should have seen a restoration of the fried egg phenotype in a ∆syfA/syrE– double mutant strain upon exposure to syringomycin. However, when a ∆syfA/syrE– mutant strain is placed in close proximity to a ∆syfA/algT– mutant, there is no restoration of the fried egg phenotype. Furthermore, a ∆syfA mutant still exhibits a strong fried egg phenotype when placed near a ∆syfA/algT-/syrE– triple mutant. Therefore, it does not appear that syringomycin acts as the direct extracellular signal that invokes this response, but rather is necessary for enabling other factors to induce the response. This evidence, in addition to the finding that syringomycin is only produced during the initial 24 hours of surface growth, lead us to assume that syringomycin instead acts as a signal that primes the colony for the fried egg phenotype that we later observe. We recently observed that an algT– mutant of P. syringae produces high levels of a surfactant termed BRF , whose production requires an rhlA homolog. This surfactant also exhibits a strong temporal pattern of regulation, with production increasing over time. Therefore, we hypothesized that this surfactant could be responsible for inducing the fried egg phenotype in the ∆syfA mutant. Colonies of a ∆syfA/∆brfA double deletion mutant did not express the fried egg phenotype at any age. However, this double mutant regained the fried egg phenotype in the presence of a ∆syfA/algT– mutant strain, suggesting that BRF could be a signal that induces this phenotype. Furthermore, neither a ∆brfA/algT– double mutant, nor a ∆syfA/∆brfA/algT– triple mutant is capable of inducing an early fried egg phenotype in a ∆syfA mutant. However, application of a BRF extract near a colony of a ∆syfA mutant does not induce the early appearance of the fried egg phenotype. It is thus possible that BRF is modified to gain its activity, or that BRF might play a role in delivering an insoluble signal over the long distances that separate colonies. Similar to the other characterized members of LuxR-type regulators, SyfR appears to form multimers in order to initiate transcription. Furthermore, in keeping with observations of LuxR and SalA, SyfR appears to have an autoregulatory role in its own transcription. However, it is unclear how the activation of SyfR is mediated by external conditions. In the case of LuxR, binding of a quorum signal induces dimerization which then allows LuxR to function as a transcription factor; thus, cell density is conveyed to the cell by increased availability and binding of the autoinducer signal, which stimulates increased LuxR activity. However, SyfR belongs to a class of LuxR-type regulators that do not contain characterized small molecule binding domains, and thus there is no factor that has been determined to limit SyfR dimerization and activity other than its own transcription levels.