Given this finding, we cultured the strains that had exhibited putative surface-dependent regulation of surfactant production for their ability to induce drop collapse when grown in viscous broth. While P. syringae B728a does not produce a surfactant capable of conferring drop collapse from normal broth cultures, it did so when grown in a viscous broth. A similar induction of surfactant production was induced by growth of other environmental strains of P. syringae, as well as Pantoea strain PB64 in viscous broth. Interestingly, P. fluorescens strain PB59, which produced surfactant only in broth media, still produced abundant biosurfactant detectable by drop collapse when grown in viscous broth , suggesting that its biosurfactant production is regulated by a different mechanism. Although it is tempting to speculate that the P. syringae and Pantoea strains are sensing a surface by directly measuring viscosity, growth in viscous broth could be indirectly stimulating biosurfactant production via alteration of growth patterns such as cell aggregation which was stimulated by the reduced turbulent drag of this culture medium. Vigorous shaking of P. syringae cultures reduced pellicle formation and resulted in a lower induction of syfA.
Since leaves are a waxy habitat, we hypothesized that the phyllosphere is enriched for biosurfactant-producing bacterial taxa due to the benefits this phenotype may confer. To test this hypothesis we examined the incidence of this trait in bacteria from different habitats including leaf surfaces using the atomized oil assay. Using this assay,garden pots for vegetables we screened over 5,000 bacteria recovered from leaf surfaces, soil, and freshwater samples in close proximity to each other in the early spring, when there were many ephemeral pools of water and streams. To determine the frequency of surfactant production in bacterial populations this trait was assessed in approximately 50 random strains per sample, and at least 30 samples were collected for each environment. The frequencies at which surfactant producers were found in a community from a given sample ranged from zero to close to 90%. Overall, a much lower frequency of surfactant producers was observed in freshwater samples than from leaf surfaces or soil. Student’s t-test with unequal variance comparing the frequencies of surfactant production revealed that leaves and soil harbored significantly higher frequencies of bacteria with this phenotype than water. Interestingly, while soil and leaf surfaces harbored a similar average frequency of surfactant producers, there was a much higher deviation in this frequency between samples of leaves than soil; nearly 30% of the leaf samples harbored no surfactant producers compared to 17% and 6% for water and soil samples respectively.
Conversely, many leaves also harbored very high proportions of surfactant producing bacteria. Several features of leaves were examined in an attempt to account for the substantial sample to sample differences in frequency of biosurfactant-producing bacteria. Given that the leaf surfaces of different plant species differ in hydrophobicity, we addressed whether plant species or the degree of water-repellency of leaves was predicative of the fraction of surfactant-producing bacterial strains recovered. There was no correlation between leaf hydrophobicity, measured as the total area covered by a 10 ul droplet of water applied to the leaf, and the frequency of surfactant producers. Likewise no association between plant species and the frequency of surfactant producers was evident , although more species would need to be examined to rigorously test this conclusion. Overall, our observations suggest that leaf properties are not the dominant factor that leads to the occurrence of surfactant-producing strains on a given plant. However, since our collections were made in early spring, the leaves examined were all at early stages of growth and thus the microbial communities were also in early stages of colonization. The apparent random patterns of occurrence of bacteria on the leaves therefore suggests that colonization can be described by a neutral theory of competition. As such, the abundance of a given bacterial strain on a leaf is reflective of its early time of arrival on that plant, and largely dependent on chance. In comparison to leaf surfaces, a much more uniform frequency of occurrence of surfactant production was observed in bacteria from soil and water.
There was no apparent effect of the source of water on the incidence of surfactant production in these samples, since about 5% of the bacteria in all samples from streams, ephemeral pools and a lake produced biosurfactant. Additionally, the frequency of surfactant-producing bacteria found in a given soil sample was not correlated with that from adjacent plant samples , suggesting that mixing of bacterial members of these two communities was not prominent. The application of the atomized oil assay to a wide variety of environmental bacterial strains and synthetic surfactants revealed it to be both more versatile and sensitive than the more commonly used drop collapse assay. The atomized oil assay confirmed surfactant production in every bacterial strain in which surfactants were detected using the drop collapse assay. More importantly, several bacterial strains were identified that produced either low amounts of surfactant or apparently hydrophobic surfactants that were not detectable using the drop collapse assay. The atomized oil assay readily confirmed biosurfactant production in taxa in which it had previously been described. The majority of the strains that produced surfactants detectable by both tests belonged to the genera Pseudomonas and Bacillus , both of which have been described in the literature to produce biosurfactants that lower the surface tension of water. Likewise, the Pantoea strain PB64 may produce rhamnolipids as do other members of this genera , although this was not verified. While surfactant production has not been previously documented in Staphylococcus, some species of this genus have been observed to be motile on swarming plates , suggesting their production of surfactants.
The identification of such previously recognized surfactant-producing taxa emphasizes that while the drop collapse assay is suitable for finding such biosurfactant producers,flower plastic pots the atomized oil assay may be more readily employed due to its high-throughput capability and higher sensitivity. The atomized oil assay was particularly useful in identifying biosurfactants in taxa in which this trait had not previously been shown. The surface-active compounds that are produced by the seven strains that were detectable only with the atomized oil assay would have escaped attention in most other studies; these compounds may well have unique biological functions and/or potential industrial applications. For example, our assay detected the hydrophobic pumilacidins produced by Bacillus pumilis which have been documented for their potent antibiotic and antiviral properties , although their surfactant activity has previously been ignored due to their low water solubility. Likewise, we detected surfactant production by a Rhizobium strain ; although we have not verified the compound, we suspect it could be similar to the long-chain AHLs produced by Rhizobium etli, which cannot be detected with a drop collapse assay but are documented as surfactants with a dual role in quorum sensing and swarming motility. Furthermore, a biosynthetic gene cluster proposed to synthesize a surface-active lipopeptide virulence factor was identified in the genome sequence of the plant pathogen Xanthomonas axonopodis ; although incapable of imparting drop collapse, both an authentic culture of X. axonopodis pv. glycines as well as a related environmental strain found in this study produced compounds detectable with the atomized oil spray. Biosurfactants detectable only with the atomized oil assay were also observed in a Cedecea strain, a taxon not previously known to produce surfactants; this feature may prove biologically important to its success as an opportunistic pathogen. Therefore it appears that application of the atomized oil assay in environmental surveys might greatly expand our knowledge of novel biosurfactants. While the atomized oil spray assay has many advantages over other assays there are some limitations that could bias the detection of surfactant producers. This assay best identifies bacterial strains that produce “bright” halos around colonies , although we have previously shown that some highly hydrophilic synthetic surfactants can modify oil droplets to appear “dark” due to their flattened nature.
“Dark” halos are less visibly obvious and no strains that unambiguously exhibited this appearance were found in our survey even though we approached the study with the expectation that we would find biosurfactants of this type. We were surprised that we did not find any biosurfactants that yielded a water drop collapse and such a “dark” halo. Bacteria that produce such compounds must thus be quite uncommon, or it may be that such surfactants are not easily distinguished or detected by either assay. Another limitation of the atomized oil assay, which is shared with any culture-based assay, is that the nutrient medium that we used may have precluded us from detecting production of surfactants by some strains which require specific conditions for surfactant production. Furthermore, our assay is restricted to surfactant production by culturable organisms, although there is evidence that at least on leaves the most common cultured taxa are also among the most prevalent taxa identified by culture-independent methods. Metagenomic investigation into the prevalence of biosurfactant production could be fruitful in expanding our understanding of their prevalence in bacterial communities, although advances will be limited until more genetic determinants for their production are described. An unexpected finding from this study was that the production of surfactants that conferred a reduction of surface tension was very conditional on whether the bacteria were grown on a surface or cultured planktonically. Although a number of studies have connected surface sensing with swarming motility , we are only aware of one report, of Serratia liquefaciens, which has noted increased biosurfactant production in cells grown on a surface. In the current work we have shown that a surprisingly large proportion of bacterial strains restrict biosurfactant production to growth on a surface. Although most of these surface-dependent surfactant producers were strains of P. syringae isolates, this phenomenon was also seen in a Pantoea strain, suggesting that it may be a common trait. Commonly-used methods of screening for biosurfactants by drop collapse employ broth cultures and would likely not identify such strains. On the other hand, two strains were identified that only conferred drop collapse from broth culture and not from cells grown onplates and subsequently suspended in water drops. However, surfactant production was still detectable in these strains as a small halo of de-wetted oil droplets with the atomized oil spray when cells were grown on plates. The small halo size of these two strains indicates that the amount of surfactant produced by cells grown on plates was probably too low in concentration to be detected by the drop collapse assay; therefore surfactant production was not fully blocked at a surface, but rather dramatically reduced. Although we have not yet encountered such strains, there is the potential for us to overlook biosurfactants which are produced only in broth culture. However, such strains must be uncommon based on our extensive survey, and the high sensitivity of the atomized oil assay should enable even very low production on solid surfaces to be detectable. Presumably the strong environmental-dependent regulation of surfactant production at surfaces is linked to its role in the habitat of some strains. For example, surfactants contributing to biofilm growth or movement on a surface would be pointless if produced in an aqueous environment. Thus, it makes sense that bacteria with multiple habitats should survey their growth environment before committing to production of a biosurfactant. The surface trigger for surfactant production and its conservation among bacterial taxa remains an active area of research.A few specific mechanisms for surface sensing have been investigated, such as two-component systems and flagellar inhibition. Once a surface is perceived, there is growing evidence that cyclic-di-GMP levels control genes involved in cell surface features that participate in processes such as biofilm growth. It is intriguing that increases in viscosity led to increases in surfactant production in this study , much as it has been shown to induce production of flagella in Vibrio parahaemolyticus. However, our results lead us to believe that it is not viscosity sensing per se that is inducing surfactant production, but rather perception of a growth pattern such as cell aggregation that perhaps restricts movement of cells which, in turn is induced by the reduced turbulent drag of a viscous medium.