Biosurfactant production was evident as a bright zone of de-wetted or raised oil droplets

The drop collapse assay was performed as according to Bodour and Miller-Maier. 2 µl 10W-40 Pennzoil® was applied to delimited wells on the lid of a 96-well plate and allowed to equilibrate at room temperature. Next, 5 µl of either diluted surfactant samples or supernatant from bacterial cultures or re-suspended bacterial colonies were pipetted onto the oil surface. Drops which retained a spherical shape were scored as negative for surfactant content, while drops which had a visibly-decreased contact angle with the oil and spread were scored as positive for surfactant content. The atomized oil assay was conducted as follows: Bacteria were evenly spotted onto KB agar plates using sterile toothpicks and grown overnight. Alternatively, if visualizing surfactant from broth culture, 1mL of 2-day-old broth culture was centrifuged at 10,000xg for 2 min, and 5 µl of supernatant was pipetted onto the plate and allowed to equilibrate for 30 minutes before assaying.An airbrush was used to apply a fine mist of mineral oil onto the plate with an air pressure between 15 and 20 psi.A collection of 377 bacterial strains isolated from a variety of terrestrial and aquatic sources were grown on agar plates and tested for biosurfactant production using the atomized oil assay in which an airbrushed mist of oil droplets was applied to culture plates.Additionally, cells of each strain suspended from plates into water as well as drops of broth culture supernatants were tested for drop collapse on an oil surface. A total of 41 of these strains exhibited biosurfactant production in at least one assay. The identities of these strains were determined from partial 16S RNA sequences,draining plant pots and all isolates were assigned to described taxa based on 98% BLAST sequence identity. Pseudomonas and Bacillus species were the most common genera identified, in line with previous reports of limited surveys. 

All biosurfactant producers were members of the Gammaproteobacteria or Firmicutes except for a single Rhizobium species. After eliminating duplicate taxa from the same sampling location, a total of 23 unique environmental strains that produced surfactant detectable in at least one assay were identified and further characterized. All 23 isolates produced surfactant detectable by the atomized oil assay, although only 16 isolates conferred drop collapse of either cells suspended from plates or of broth culture supernatants. Furthermore, cells of only 9 of these 16 isolates conferred drop collapse from both culture conditions. Most of the other 7 strains that conferred drop collapse only under one culture condition did so for suspended plate-grown cells. P. syringae strains were typical of this group; cells of four representative isolates conferred drop collapse when suspended in water from plate cultures but not the supernatant of planktonic cultures. While 16 strains of P. syringae, P. fluorescens, or B. subtilis produced biosurfactant that could be detected by both assays, the 7 strains that exhibited biosurfactant activity that was detectable only by the atomized oil assay mostly consisted of a diversity of other taxa. Although not appreciated in most biological studies, surfactants differ greatly in their chemical properties in ways that could influence their ability to be detected by various assays. For instance, a fundamental property of a surfactant is its relative solubility in water and oil, which can be broadly described by its hydrophilic-lipophilic balance value. Some important synthetic surfactants with low hydrophilicity are not readily dispersible in water, and thus have unique functions such as forming inverse emulsions of water into oil. If a bacterial strain produced a biosurfactant with such low water solubility this could account for its inability to reduce the surface tension of water sufficiently to collapse a water drop.

In order for drop collapse to occur on an oil surface, a minimum surface tension reduction at the water/air interface from 72 dyn/cm to around 43 dyn/cm is required. Although a surfactant may be present in a sample of interest, it might not bedetected by the drop collapse assay if it is produced in low quantities or has a property preventing it from lowering the surface tension of water. Because the atomized oil assay can detect 10- to 100-fold lower concentrations of surfactant than that of the drop collapse assay , it is reasonable to hypothesize that the atomized oil assay can detect surfactant production in weakly producing strains. Therefore, it was possible that the 7 strains that did not confer drop collapse may simply produce too little surfactant to be detected with this method. Indeed many of these strains exhibited small halos in the atomized oil assay , suggestive of low surfactant concentrations. However, a few strains such as Bacillus pumilis that did not cause drop collapse produced biosurfactants that conferred halos of de wetted oil droplets around colonies that were at least as large as many strains whose biosurfactants did confer drop collapse. This observation led us to suspect that the surfactant had properties which hindered its ability to be detected by the drop collapse assay. To address the features of biosurfactants that could be detected by the atomized oil assay but not the drop collapse assay, we distinguished the extent to which the hydrophobicity of the surfactants might limit their detection with the later method or whether the higher sensitivity of the atomized oil assay was responsible for their detection. As a test of the relative hydrophobicity of the surfactant produced by B. pumilis we suspended colonies of it as well as P. syringae strain B728a in water to identical concentrations, removed the cells by centrifugation, and then tested the supernatant for surfactant activity using the atomized oil assay. The water soluble material washed from cells of P. syringae B728a, which contains syringafactin and readily causes drop collapse , contained sufficient surfactant to produce a large halo of de-wetted oil droplets when placed on an agar surface. 

However, very little biosurfactant was apparently washed from cells of B. pumilis, since no zone of de-wetted oil droplets was observed. Similarly, the surfactants produced by Pantoea ananatis and Pseudomonas fluorescens strains which were detected only by the atomized oil assay also appeared to have low water solubility when assayed after washing of cells. However, the washings of four other strains that exhibited the ability to de-wet atomized oil droplets but not to collapse water drops,drainage gutter retained the ability to de-wet oil droplets. This suggests that these strains produced only small amounts of a water-soluble surfactant that could be detected by the drop collapse assay if present in higher concentrations. In support of this conjecture was the observation that these later strains exhibited only relatively small halos in the atomized oil assay. The low production of water soluble surfactants in these strains was verified for P. syringae strain PB54 using mass spectroscopy. This strain was observed to produce the same syringafactins as P. syringae B728a, albeit in much lower quantities, confirming that the detection of surfactants in strain PB54 by the drop collapse assay was compromised by its low level of production. In order to confirm our conjecture that the lack of detection of biosurfactant production in our B. pumilis strain in the drop collapse assay was due to its low water solubility, we characterized it using MALDI mass spectroscopy. The mass spectrogram of the material extracted from the cell free region surrounding colonies on the surface of plates revealed a series of prominent peaks in the range of 1050-1130. Several B. pumilis strains have previously been shown to produce a family of pumilacidins in this mass range. The mass spectrogram of our strain shares the same masses of a sample containing a mixture of pumilacidin A, B, C, and D. The masses observed in Fig. 2 are a combination of [M+Na]+ and the [M+K]+ adducts commonly seen in MALDI mass spectroscopy. Therefore, we conclude that our strain is producing a mixture of low water solubility pumilacidins that are capable of readily diffusing away from cells on the surface of an agar plate, but which are not sufficiently water soluble to impart drop collapse. In order to demonstrate pumilacidin’s surfactant capabilities, the surface tension of a broth culture of B. pumilis was measured using a highly sensitive pendant drop analysis. The surface tension of the broth culture supernatant was lowered by production of a surface active compound to 50 dyn/cm; this surface tension is just above the minimum threshold necessary to impart a drop collapse. Since the highly hydrophobic pumilacidins were detectable using the atomized oil assay, we further determined the efficiency with which other characterized synthetic surfactants differing in chemical properties could be detected by this method.

The assay was performed on synthetic surfactants that possessed a broad range of hydrophobicities. As seen previously, the atomized oil assay readily detected surfactants having more balanced hydrophilic and lipophilic groups, which were also detected by the drop collapse assay. On the other hand, the hydrophobic surfactants Span® 85 and Span® 80 each yielded large bright halos in the atomized oil assay, but given their low water solubility, could not be detected in the aqueous phase by the drop collapse assay. This is in agreement with our observation that hydrophobic pumilacidins were also only detectable by the atomized oil assay and not by the drop collapse assay. Curiously, the synthetic surfactants not only caused bright halos of de-wetted atomized oil droplets, but those with balanced hydrophilic and lipophilic groups also caused the oil droplets to migrate away from the source of surfactant, traveling at a speed of up to 0.1 mm/minute. Such expanding halos may result from a strong surfactant gradient, such as explored by Angelini et al. , although it is unclear why this should not be also conferred by the hydrophobic surfactants. This property was commonly observed around biosurfactant producing bacterial colonies and might be used to infer the water solubility properties of the biosurfactants. In addition to the surfactants that were only revealed by the atomized oil assay, we also found that many surfactants were detectable in the drop collapse assay only when cells had experienced a particular growth condition. Most prominent among strains exhibiting such growth condition-dependent production of surfactants were strains of P. syringae; cultures of this species never conferred water drop collapse when grown planktonically. The factors determining surfactant production in P. syringae pv. syringae B728a, typical of this species, was thus investigated. While culture supernatants of this strain did not cause water drop collapse on an oil surface, plate-grown cells suspended to the same concentration as the planktonic culture conferred water drop collapse. Suspension of a syfA– mutant blocked in production of syringafactin did not cause water drop collapse, confirming that the drop collapse is due to syringafactin. We thus postulated that enhanced expression of syringafactin production in cells grown on a surface was responsible. In order to link syringafactin production to surface-mediated increases in surfactant production, we examined the transcriptional regulation of syfA using a GFP-based bioreporter. Greater than a 10-fold increased expression of syfA was observed when cells were grown on an agar surface compared to planktonic growth in broth culture. As a control, a strain constitutively expressing GFP exhibited similar levels of fluorescence in both cultures. Since there have been reports that production of some surfactants are influenced by growth stage , we examined syfA expression at a variety of times for up to 3 days during the growth of both liquid and solid cultures of P. syringae. GFP expression was higher in cells recovered from agar plates than broth cultures at all times, indicating that this is not growth-stage dependent phenomenon. Additionally, some reports have documented that surfactant production is activated in more dense cultures by quorum sensing. However, the GFP fluorescence of P. syringae harboring the pPsyfA-gfp fusion in the wild-type and a quorum sensing deficient strain was similar both in liquid and solid cultures, indicating that syringafactin production is not dependent on quorum sensing. Although not previously connected to surfactant production, one of the ways by which bacteria sense surfaces is apparently through monitoring the viscosity of their environment. When PVP-360, a viscosifying agent, was added to broth medium, the expression of syfA was increased to levels similar to that of cells on agar plates.