The M. xanthus strains used in this study are listed in Table 1. Unless otherwise specified, all M. xanthus strains were grown and maintained on Casitone-yeast extract (CYE) agar plates or in CYE liquid media (Campos et al., 1978) at 32°C on a rotary shaker at 300 rotations per minute (RPM). A stock solution of 1% (wt/vol) CV (ACROS Chemicals) was prepared in 20% (vol/vol) ethanol (Decon Laboratories). Glacial acetic acid (Fisher) was used to make a 30% (vol/vol) acetic acid solution. The MOPS buffer contains 10 mM morpholinepropanesulfonic acid (pH 7.6) and 2 mM MgSO₄.

The clear tissue culture (TC)-treated flat-bottom 96-well microplates (Falcon) were used for the development of M. xanthus SBFs per the Dahl protocol (Dahl et al., 2011) or according to the procedures as described later in the manuscript. For the Dahl protocol, M. xanthus cells in the logarithmic growth phase were harvested and resuspended in CYE media to an optical density at 600 nm (OD₆₀₀) of 0.8. 100 μl aliquots of the cell suspension in quadruplicate were added to the wells of a microplate. After incubation at 28°C for 12 h under static conditions for overnight seeding, the media was removed and the wells were washed with the MOPS buffer. For biofilm development, 100 μl of fresh CYE was added to each well and the microplate was incubated under static conditions for 24 h at 32°C. To develop our protocol, M. xanthus cells in logarithmic growth were harvested and resuspended in CYE media to various OD₆₀₀ as indicated in the text. Aliquots of 75, 100, or 125 μl of the cell suspensions were dispensed into the microwells in quadruplicate. Biofilms were allowed to develop at 32 or 27°C for 24 h under static conditions or on a rotary shaker at 230 RPM.

The SBFs developed above were quantified by the widely adopted CV-based method (Christensen et al., 1985; Kwasny and Opperman, 2010; Xi and Wu, 2010; Dahl et al., 2011; Redder and Linder, 2012; Naher et al., 2014; Bordeleau et al., 2018). Briefly, after the production of SBFs in microwells, the media and unattached cells were gently removed by a multichannel pipette. The wells were washed with equal volumes of MOPS buffer to the culture volume. Staining was conducted with 150 μl of 1% CV solution for 20 min before washing thrice with 175 μl of H₂O. After air drying, 200 μl of 30% acetic acid was added to each well and incubated for 20 min. 125 μl of the acetic acid solution was then transferred to the microwells of a clear polystyrene 96-well microplate (ExtraGene). The CV absorbance (A_cv) was measured at 600 nm using an Infinite F200 PRO plate reader and the A_cv values were used as the quantifier of SBF amounts per well. For some experiments, the biofilm amounts (A_cv values) were normalized to either the total area of the submerged surfaces or the final OD₆₀₀ of the samples. The submerged surface area for a given sample was calculated from the culture volume in a microwell based on the specifications of the microtiter plate by the manufacturer. The submerged surface areas for the 75 μl, the 100 μl, and the 125 μl samples were determined to be 0.79, 0.95, and 1.10 cm², respectively. To normalize SBF amount to cell density in a microwell, the OD₆₀₀ of the cell culture after SBF development was measured 16 times by the Multiple Reads function of the plate reader. The average of these measurements was given as the final OD₆₀₀. In some instances, Grubbs’ test identified one of the quadruple samples as an outlier which was expunged from the dataset. Statistical differences were determined using the Student’s t-test.

The linear range of the F200 PRO under our experimental conditions was determined by the measuring A_cv of serial dilutions of a CV solution. In total, 15 concentrations from 0 to 100 ppm were analyzed in quadruplicates in a 96-well microplate (Supplementary Figure 1). This analysis showed the linearity of A_cv vs. CV concentration extends up to 60 ppm of CV and A_cv values above 3.0. The coefficient of determination or R² values are 0.9977 and 0.9999 for the CV concentration range of 0–60 ppm and of 0–30 ppm, respectively. When CV concentrations were higher than 60 ppm and the A_cv values went above 3.5, the linear relationship no longer holds (data not shown).

We first examined if the conventional microplate-based biofilm assay (O’Toole and Kolter, 1998) without the overnight seeding step (Dahl et al., 2011) could be applied to M. xanthus SBFs under vegetative growth. For this, we compared the results from two sets of experiments. The first set was conducted according to the Dahl protocol such that a microwell was inoculated with 100 μl of M. xanthus cells at OD₆₀₀ of 0.8 and incubated overnight. The microwell was then washed and replenished with fresh media to allow SBF development for 24 h at 32°C. For the second set, each well of the microtiter plate was inoculated with 100 μl of a cell suspension at OD₆₀₀ of 0.1 and SBFs were allowed to develop directly for 24 h at 32°C. Three M. xanthus strains were used for the initial experiments: the wild type (WT) DK1622, the EPS^– strain YZ603 (ΔdifE) and the T4P^– strain YZ690 (ΔpilA). It should be noted that the T4P^– strain is also deficient in EPS production because T4P is required for wild-type levels of EPS production in M. xanthus (Black et al., 2006). As shown in Figure 1, these two sets of experiments yielded similar trends of SBF formation for these strains. In both protocols, the WT produced significantly more biofilms than the EPS^– and T4P^– strains as reflected by CV absorbance (A_cv). These trends were expected because both EPS and T4P have been demonstrated to be critical for biofilm formation in multiple bacteria (Bahar et al., 2009; Colvin et al., 2011; Maunders and Welch, 2017; Fiebig, 2019). Moreover, the amounts of SBFs as quantified by CV retention were comparable for the two protocols for all three strains. The A_cv values for the WT were 0.34 ± 0.03 and 0.31 ± 0.04 in these two protocols, respectively. Those for the ΔdifE strains were 0.08 ± 0.01 and 0.07 ± 0.01, and the ΔpilA strain, 0.13 ± 0.01 and 0.15 ± 0.01, respectively. These observations indicate that a protocol without the seeding step performed comparably with the Dahl protocol. Overnight seeding was therefore eliminated from experimental procedures for the remainder of this study.

We next examined the effects of culture volume and cell density on SBF formation in the microplate-based assay under static conditions as above (Figure 1). M. xanthus cells from an overnight culture were harvested and resuspended in fresh CYE at OD₆₀₀ from 0.1 to 0.8. Aliquots of 75, 100, or 125 μl of these cell suspensions were placed into the microwells for SBF development for 24 h at 32°C under static conditions, followed with analysis by CV retention. The results as shown in Figure 2A indicated a general trend of increasing SBF amounts with increasing cell density up to a certain point or threshold, beyond which this trend is lost. For the 75 μl samples, the OD₆₀₀ threshold was 0.5 as the amounts of biofilm dropped precipitously at OD₆₀₀ of 0.6 or higher (Figure 2A). For the samples with 100 μl and 125 μl culture, the reduction in biofilms occurred at OD₆₀₀ of 0.5 or higher (Figure 2A). Upon further examination, the decrease in SBFs at high cell density was found to coincide with the appearance of biofilms at the liquid-air interface under these experimental conditions (Supplementary Figure 2). These PBFs were removed with the culture media during the washing step before analysis by CV retention. Significant numbers of cells developed into PBFs rather than SBFs under these conditions. This explains the drastic decrease in SBF amount at higher cell densities (Figure 2A).

The formation of PBFs at the liquid-air interface has been observed and investigated for many bacteria including Escherichia coli and Bacillus subtilis (Yamamoto et al., 2011; Holscher et al., 2015; Kovacs and Dragos, 2019; Arnaouteli et al., 2021; Golub and Overton, 2021). PBFs are commonly observed on the surface of a liquid culture under static conditions in response to oxygen depletion in the liquid media (Armitano et al., 2014; Holscher et al., 2015; Kovacs and Dragos, 2019). Evidence suggests that cells may float to form aggregates at the liquid-air interface where the concentration of oxygen is the highest under these conditions (Yamamoto et al., 2011; Armitano et al., 2013, 2014; Holscher et al., 2015). These cells and their aggregates may further develop into mature PBFs by increasing the production of EPS and other biofilm matrix materials (Armitano et al., 2014; Holscher et al., 2015). As an obligate aerobe, it is perhaps not surprising that M. xanthus forms PBFs under static condition at high cell density. The consumption of dissolved oxygen in the media is expected to result in oxygen limitation and thus the formation of PBFs at the liquid-air interface where oxygen is more readily available. The formation of visible PBFs can therefore explain the observed reduction in M. xanthus SBFs at high cell densities (Figure 2A).

Indeed, the trend of SBF quantity with varying volumes of culture appeared consistent with an oxygen effect when normalized to the submerged surface areas for a sample. SBF amount (A_cv) per cm² of submerged surface area showed a seemingly decreasing trend with increasing culture volume (Figure 2B). At the same starting cell density, the higher the culture volume, the lower the A_cv/cm² value. For example, at the starting OD₆₀₀ of 0.1, the A_cv/cm² values are 0.33 ± 0.01, 0.31 ± 0.00, and 0.27 ± 0.02 for wells with the 75, 100, and 125 μl of samples, respectively. At OD₆₀₀ of 0.4, the values for these wells are 0.64 ± 0.06, 0.60 ± 0.02, and 0.52 ± 0.02, respectively. It can be assumed that when the depth of liquid in a microwell increases with increasing culture volumes, cells at or near the bottom of the well experience more severe oxygen limitations under static conditions. This explains the formation of PBFs at high cell density (Figures 2A,B and Supplementary Figure 2) and suggests that oxygen availability greatly influences the formation of SBFs by M. xanthus.

The impact of oxygen availability on the formation of SBFs by M. xanthus was investigated next. Here we conducted experiments wherein cultures in the microtiter plates were aerated on a rotary shaker. For these experiments, samples were prepared as in Figure 2, except that the microtiter plates were incubated with rotary shaking. Under these conditions, the amount of SBFs per well increased steadily with increasing cell density (Figure 3A and Supplementary Figure 3). This is in stark contrast to those under static conditions where the amount of SBFs showed drastic decreases when the starting OD₆₀₀ was equal to or greater than 0.6 (Figure 2A). Under this aerating condition, the amount of SBF continued a positive trend with increasing cell density up to OD₆₀₀ of 0.8, the highest in our experiments (Figure 3A and Supplementary Figure 3). In addition, no PBF was observed for any of the samples, suggesting that M. xanthus PBF formation is sensitive to oxygen levels in the culture (Figures 2, 3).

Recall that the amounts of SBFs normalized to submerged surface areas decreased with increasing culture volume under static conditions (Figure 2B). This trend no longer holds when cultures are aerated on a rotary shaker. For wells with the same starting cell density, the amounts of SBF/cm² showed no decrease as culture volume increased (Figure 3B). For example, at the starting OD₆₀₀ of 0.1, the 75, 100, and 125 μl samples had A_cv/cm² values of 0.33, 0.41, and 0.45, respectively. At OD₆₀₀ of 0.3, these samples gave values of 0.77, 0.77, and 0.91, respectively. In some cases, there are statistically significant increases at higher volumes (100 and 125 μl) over the 75 μl cultures. At OD₆₀₀ of 0.2 and 0.6, for instance, the 100 and 125 μl samples produced significantly more SBFs than the 75 μl cultures (Figure 3B). The amounts of SBFs for the two higher volumes are generally not statistically different after normalization to submerged area. Although the relationship between SBF formation and culture volume under aerating conditions has yet to be fully investigated, it is clear that the inverse relationship seen under static conditions (Figure 2B) disappears when cultures are aerated through rotary shaking (Figure 3B).

Most importantly, there are significant increases in M. xanthus SBF when cultures are aerated in comparison with the static condition (Table 2). It is not surprising that when their counterparts formed PBFs under static conditions (shaded wells in Table 2), the corresponding samples formed significantly greater amounts of SBFs under shaking conditions. The remaining samples may be divided into two categories by culture volume. The first category includes those with 75 μl of cultures; for these samples, there does not appear to be significant differences between shaking vs. static conditions. For those with higher culture volumes (100 and 125 μl), there are generally significant increases in SBF/cm² under shaking conditions (Table 2). For the 100 μl cultures, when the starting OD₆₀₀ increased from 0.1 up to 0.4, the increases under shaking conditions ranged from 85 to 97%. For the 125 μl cultures, the increases ranged from 67% to over 160%. These results indicate that aeration through rotary shaking significantly enhanced M. xanthus SBF formation and it was adopted for M. xanthus SBF development for the remainder of the study.

Myxococcus xanthus grows optimally at 32°C in aerated liquid culture (Janssen et al., 1977). Yet, we have observed that this bacterium produced higher levels of EPS on agar surfaces at 27°C or at room temperature (Black et al., 2017; Dye et al., 2021; unpublished data). Since EPS is a major component of the bacterial biofilm matrix, we compared the amount of SBF developed at 27 and 32°C. Here, two identical sets of experiments were initiated as in Figure 3, except one was incubated at 27°C while the other at 32°C. As shown in Figure 4A and Supplementary Figure 4A, differences in SBFs per microwell at these two temperatures are generally not statistically significant. However, because M. xanthus grows slower at 27°C than 32°C, the samples at 27°C were anticipated to have less growth and fewer cells. It therefore remained a possibility that a higher percentage of cells might be in SBFs at 27°C than 32°C relative to the planktonic population. We measured the optical density of the culture after biofilm development as described in section “Materials and Methods.” As expected, the OD₆₀₀ of the culture was significantly higher at 32°C than at 27°C (Figure 4B and Supplementary Figure 4B). When SBF was normalized to the OD₆₀₀ of the culture, it is clear that the proportion of cells in SBFs is significantly higher at 27°C than at 32°C (Figure 4C and Supplementary Figure 4C). For example, for the wells with starting OD₆₀₀ of 0.4, the biofilm amounts by this measure are between 40 to 60% more at 27°C than at 32°C for all samples at the three volumes examined (Figure 4C). These observations indicate that M. xanthus cells form SBFs more readily at 27°C. Based on these analyses and previous observations, 27°C was selected as the temperature for the development of M. xanthus vegetative SBFs in our assay moving forward.

To finalize the remaining parameters for our microplate-based assay, we chose to use 100 μl of culture per microwell with the starting OD₆₀₀ of 0.4. The 100 μl volume was chosen for three reasons. First, this is the most common volume in similar assays for other bacteria (Merritt et al., 2005; Kwasny and Opperman, 2010). Second, this is the volume used by Dahl et al for the analysis of M. xanthus SBFs previously (Dahl et al., 2011). Lastly, the difference in culture volumes per well generally did not translate into significant differences in SBF amounts under aerating condition when normalized to submerged surface area (Table 2). For the starting cell density, we took into consideration the linear range of the instrumentation (Supplementary Figure 1), aiming for an A_cv of ∼1.0 for DK1622 (WT) (Supplementary Figure 4). We anticipate that mutations may either enhance or diminish biofilm formation. An A_cv reading of ∼1.0 for the wild-type would leave room for analysis of mutants with either an increase or a decrease in biofilms formation. With 100 μl sample volume at 27°C, the starting OD₆₀₀ of 0.4 yielded the nearest A_cv reading to 1.0 (Figure 4A and Supplementary Figure 4A). The following is a summary of the experimental parameters for our finalized assay for vegetative SBF of M. xanthus. 100 μl of a cell suspension in CYE at OD₆₀₀ of 0.4 is inoculated into a microwell of a 96-well microplate in quadruplicates. The plate is incubated at 27°C for 24 h with rotary shaking for SBF development. The amounts of SBFs are then analyzed by CV retention using a plate reader (Supplementary Figure 1) as in similar assays for other bacteria (Christensen et al., 1985; Merritt et al., 2005; Xi and Wu, 2010).

It is known that bacterial T4P and EPS play critical roles in SBF development as adhesins and biofilm matrix materials. In M. xanthus, the levels of T4P and EPS are known to be intertwined in a mutual relationship. On one hand, piliation levels have been demonstrated to positively modulate EPS levels. pilA and pilB mutants, which are un-piliated, produces very low levels of EPS in both liquid culture and on agar plates (Black et al., 2006, 2009, 2017Yang et al., 2010). pilT mutants, which are hyperpiliated because they assemble non-retractable pili (Wu et al., 1997), produces higher amounts of EPS than the wild-type in liquid culture (Black et al., 2006). On the other hand, studies suggest that EPS levels in turn can influence piliation levels. Experimental evidence supports a model wherein the retraction of T4P is triggered by interactions with EPS in M. xanthus (Li et al., 2003; Nudleman and Kaiser, 2004; Zhou and Nan, 2017). In other words, M. xanthus EPS is the preferred anchor and trigger for T4P retractions. This explains the hyperpiliated phenotypes of certain EPS^– mutants (Bellenger et al., 2002; Li et al., 2003; Black and Yang, 2004) because the pilus does not retract without EPS as an anchor and trigger (Li et al., 2003). In addition, it is known that EPS levels in M. xanthus are regulated in part by the Dif chemotaxis-like pathway (Black and Yang, 2004; Black et al., 2006, 2017; Yang et al., 2014). DifE, which resembles the CheA kinase in bacterial chemotaxis pathways, is a positive regulator of EPS. The deletion of difE leads to the lack of detectable EPS, absence of S motility and increased piliation levels (Bellenger et al., 2002; Li et al., 2003; Black et al., 2006). There are also negative regulators of EPS in the Dif pathway, namely, DifD and DifG, which are homologs to the chemotaxis proteins CheY and CheC, respectively (Black and Yang, 2004; Black et al., 2006). Deletions of difD or difG lead to EPS overproduction and their mutations have additive effects such that a difD difG double mutant produces more EPS than their respective single mutants (Black and Yang, 2004; Black et al., 2006). It is known that the Dif pathway functions downstream of T4P in EPS regulation, in part because a ΔdifD ΔdifG double mutation suppressed the EPS defect resulting from a pilA deletion (Black et al., 2006, 2017).

We analyzed the amounts of SBFs of a few M. xanthus mutants with altered levels of EPS and T4P using our assay. The strains here included three that were used in in earlier experiments, namely the ΔdifE (YZ603), the ΔpilA (YZ690), and the WT (DK1622) strains (Figure 1). We selected six additional mutants with varying levels of EPS and T4P as established in multiple studies under different experimental conditions previously (Wu et al., 1997; Wall et al., 1998; Black et al., 2006, 2017; Wang et al., 2011; Perez-Burgos et al., 2020). These include an un-piliated ΔpilB mutant (DK10416) and a hyperpiliated ΔpilT mutant (DK10409). We also included a ΔdifG (YZ604), a ΔdifD (YZ613), and a ΔdifD ΔdifG double (YZ641) mutants. Finally, we included the ΔdifD ΔdifG ΔpilA triple mutant YZ646. This strain is un-piliated but produces similar amounts of EPS as the WT (Black et al., 2006). All of these strains were allowed to form SBFs as specified in the preceding section at 27°C, and the amounts of their SBFs were analyzed by CV retention (Figure 5).

The analysis of these results shows that M. xanthus SBF formation has no correlation with piliation levels under our experimental conditions. For example, both the ΔpilT and the ΔdifE mutants are hyperpiliated. Yet, the SBF of the former gave a A_cv of 1.51, more than 10-fold higher than 0.11 for the ΔdifE mutant (Figure 5). Similarly, the ΔpilA, the ΔpilB and the ΔdifD ΔdifG ΔpilA mutants are all un-piliated due to the deletion of either pilA or pilB. Yet the A_cv value for the triple mutant (1.21) is significantly higher than those for the ΔpilA (0.24) and the ΔpilB (0.18) mutants (Figure 5). Although the WT strain is piliated and the ΔdifD ΔdifG ΔpilA triple mutant is not, they produced comparable levels of SBFs in this assay. These observations clearly demonstrate that, under our experimental conditions, M. xanthus SBF formation has no direct correlation with piliation levels.

However, there is a clear correlation between EPS levels and SBF amounts by the different strains we examined (Figure 5). It is well established that ΔdifE, ΔpilA, and ΔpilB mutants produce undetectable or significantly lower levels of EPS in comparison with the wild-type strains (Black et al., 2006, 2017; Yang et al., 2010). Both ΔpilA and ΔpilB mutants produced more EPS than ΔdifE with the ΔpilA mutant producing slightly more ESP than a ΔpilB mutant (Black et al., 2006; Yang et al., 2010). It has also been demonstrated that a ΔdifD ΔdifG double deletion is able to suppress and restore EPS production to a ΔpilA mutant to about the wild-type level (Black et al., 2006). For mutants that overproduce EPS, the ascending order is ΔpilT, ΔdifG,ΔdifD and finally the ΔdifD ΔdifG double mutant (Black and Yang, 2004). To recap, previous studies indicate that the order of strains used here going from low to high EPS levels is YZ603 (ΔdifE)➔DK10416 (ΔpilB)➔YZ690 (ΔpilA)➔DK1622(WT)/YZ645(ΔdifD ΔdifG ΔpilA)➔DK10409 (ΔpilT)➔YZ604 (ΔdifG)➔YZ613 (ΔdifD)➔YZ641 (ΔdifD ΔdifG). As shown in Figure 5, the amounts of SBFs formed by these strains followed exactly the same order as their EPS levels. These results collectively demonstrate that the level of SBF formation in M. xanthus under our experimental conditions tightly correlate with the amount of EPS produced by M. xanthus under vegetative growth. We suggest that our newly developed SBF protocol here may be utilized to conveniently and reliably quantify the relative EPS levels in M. xanthus under vegetative conditions in a high throughput format. Most importantly, this assay will allow further studies of M. xanthus SBFs to probe the mechanisms of SBF formation by an obligate aerobe adapted to living and translocating on solid surfaces in its natural environment.