Increased production of the extracellular polysaccharide Psl can give a growth advantage to Pseudomonas aeruginosa in low-iron conditions

In infections, biofilm formation is associated with a number of fitness advantages, such as resistance to antibiotics and to clearance by the immune system. Biofilm formation has also been linked to fitness advantages in environments other than in vivo infections; primarily, biofilms are thought to help constituent organisms evade predation and to promote intercellular signaling. The opportunistic human pathogen Pseudomonas aeruginosa forms biofilm infections in lungs, wounds, and on implants and medical devices. However, the tendency toward biofilm formation originated in this bacterium’s native environment, primarily plants and soil. Such environments are polymicrobial and often resource-limited. Other researchers have recently shown that the P. aeruginosa extracellular polysaccharide Psl can bind iron. For the lab strain PA01, Psl is also the dominant adhesive and cohesive “glue” holding together multicellular aggregates and biofilms. Here, we perform quantitative time-lapse confocal microscopy and image analysis of early biofilm growth by PA01. We find that aggregates of P. aeruginosa have a growth advantage over single cells of P. aeruginosa in the presence of Staphylococcus aureus in low-iron environments. Our results suggest the growth advantage of aggregates is linked to their high Psl content and to the production of an active factor by S. aureus. We posit that the ability of Psl to promote iron acquisition may have been linked to the evolutionary development of the strong biofilm-forming tendencies of P. aeruginosa.

In this work, we find that when iron is a limiting resource, multicellular aggregates 62 of P. aeruginosa have a growth advantage over P. aeruginosa single cells that is linked 63 to iron acquisition and to the presence of Staphylococcus aureus. S. aureus is a Gram-64 negative bacterium that is often found as a co-pathogen in P. aeruginosa biofilm 65 infections, and can also be thought of as a token representation of the diverse microbial 66 populations in the ecologies in which P. aeruginosa evolved. We also find that the 67 enhanced growth of aggregates is linked to aggregates having higher amounts of the 68 extracellular polysaccharide Psl than have single cells. Therefore, we infer that the need 69 to acquire and use iron, as a vital growth substrate, could result in a competitive 70 advantage to the production of Psl, which both binds iron [21] and also promotes 71 intercellular aggregation and biofilm formation. 72 In contrast to the present work, we have previously shown that, in monoculture   NaOH, adjusted to a pH of 9, was prepared and stored at 4 o C in plastic tubes (deferrated 124 EDDHA was a gift from Shelley Payne, University of Texas, Austin). EDDHA primarily 125 chelates the ferric form of iron, though it can also chelate small amounts of ferrous iron.

126
EDDHA was added to M9+glucose at a final concentration of 32 Molar and 320 Molar, 127 representing low and high concentration of iron chelator.

128
Flow cell experiments 129 We used a standard flow cell system as previously described with certain    aeruginosa single cells that were each ~500 m away from the corresponding aggregate 145 (Fig 1). Single-cell regions were displaced from their corresponding aggregate in a 146 direction perpendicular to the direction of media flow, so that the aggregate and single-147 cell region were at the same distance from the media inlet, and therefor were in areas 148 with approximately the same concentration of growth resources such as carbon source, 149 oxygen, and iron, and of metabolic byproducts. A displacement of ~500 m subtends 150 approximately 10% of the width of the ~5mm flow channels and is ~10× greater than ~10 151 m aggregate diameter. This was chosen to achieve spacing between observed regions 152 such that the single-cell region was far from, and unperturbed by, the aggregate, but the    We checked for statistically significant differences between two data sets using 194 the two-tailed Student t-test. For comparison of a data set to unity, we used the one-

215
In monoculture, we find that aggregates of P. aeruginosa grow better than single 216 cells of P. aeruginosa at high cell density; aggregates grow worse than single cells at low 217 cell density; growth of aggregates and single cells is comparable at medium cell density.

218
These finding agrees with our earlier publication showing that P. aeruginosa aggregates 219 can have a competition-dependent growth advantage over P. aeruginosa single cells [22]. 220 We also find that co-culture with S. aureus does not harm the growth of P. aeruginosa 221 (Fig 2). Indeed, co-culture with S. aureus slightly benefits P. aeruginosa growth, although 222 in most cases the increase in growth is not statistically significant (Fig 2).  S. aureus is present in co-culture, but these differences are not statistically significance.

232
Error bars are standard error of the mean.

233
However, we find that co-culture with S. aureus enhances the growth of P. 234 aeruginosa aggregates more than it enhances the growth of single cells. At high and 235 medium cell density, in co-culture with S. aureus, the growth of P. aeruginosa aggregates 236 is greater than the growth of P. aeruginosa single cells and these differences are 237 statistically significant; at low cell density, in co-culture with S. aureus, the growth of P. 238 aeruginosa aggregates is greater than the growth of P. aeruginosa single cells but this 239 difference is not quite statistically significant (Fig 3). The fitness advantage of aggregates 240 over single cells is not statistically different for the three cell densities studied, and thus 241 does not depend on competition (Fig 3). This is strikingly different from our earlier work,  Fig 1B). Thus, we conclude that it is unlikely that a 272 poorly-growing culture of S. aureus would be a better competitor for resources than 273 heartily-growing P. aeruginosa, and therefore that the dependence of aggregate relative 274 fitness on local S. aureus does not arise from competition.

275
Thus, we find that co-culture with S. aureus increases the relative fitness of 276 multicellular aggregates of P. aeruginosa and erases the dependence of relative fitness 277 on competition that characterizes the monoculture system.

278
The relative fitness of P. aeruginosa aggregates depends on 279 the local ratio of S. aureus to P. aeruginosa 280 To explore the role that S. aureus plays in the relative fitness of multicellular 281 aggregates of P. aeruginosa, we varied the initial sample-wide ratio of S. aureus to P. 282 aeruginosa when combining them prior to inoculation (P. aeruginosa < S. aureus, P. were used for these measurements. We then binned the resulting relative fitness of P. 295 aeruginosa aggregates according to the initial local ratio of S. aureus to P. aeruginosa, 296 with bins as follows: P. aeruginosa : S. aureus < 1; 1 < P. aeruginosa : S. aureus < 10; P. 297 aeruginosa : S. aureus > 10 (Fig 4). We find that when initial local concentrations of S. 298 aureus > P. aeruginosa, the average relative fitness of aggregates is 2.5 ± 0.4. When the 299 initial local concentrations are such that S. aureus ~ P. aeruginosa, the average relative 300 fitness is 1.8 ± 0.3. These values for relative advantage are both statistically significant.

301
When the initial concentrations of S. aureus < P. aeruginosa (i.e. when approaching a 302 monoculture of P. aeruginosa) there is no statistically-significant relative fitness 303 advantage for aggregates. Thus, we find that the relative fitness of P. aeruginosa 304 multicellular aggregates depends on the initial local ratio of the two species, increasing 305 as the local population contains a higher proportion of S. aureus than P. aeruginosa.  If S. aureus were producing a factor harmful to P. aeruginosa, from which 312 aggregates are better-protected than single cells, we would expect P. aeruginosa growth 313 in co-culture to be slower than growth in monoculture. On the contrary, we find that P. 314 aeruginosa in co-culture grows as fast or faster than P. aeruginosa in monoculture, for all 315 cases examined (Fig 2). 316 Thus, our results indicate that P. aeruginosa aggregates are able to better obtain 317 some benefit from S. aureus than are P. aeruginosa single cells, and that the benefit  pel should make more Psl than the wild-type (WT); this is supported by our earlier study 340 (Fig 6 in [28]). For pel aggregates compared with pel single cells, we found a relative 341 fitness of 0.6 ± 0.15 for medium initial density (OD~0.01). Thus, the relative fitness found 342 for WT aggregates vanishes and pel aggregates are less fit than pel single cells. This for both WT and pel. We find that is greater for pel than for WT for both 0 349 aggregates and single cells (Fig 5); thus, both aggregates and single cells have increased  (Fig 6). This suggests that when iron is a scarce resource, aggregates are better able to 386 acquire iron and grow than are single cells, but when iron is abundant this advantage is 387 no longer important. concentrations of chelator, we find that aggregates are fitter than single cells (Fig 7). At 408 low chelator concentration, the relative fitness of aggregates is comparable to that with 409 no chelator present (Fig 2). At high chelator concentration, aggregate advantage over 410 single cells increased by ~90% for high inoculation density, and by ~30% for medium 411 inoculation density (Fig 7).  Increasing iron chelator is equivalent to decreasing the available iron. Thus, these 420 findings are consistent with aggregates being better able to acquire and use iron for 421 growth when iron is a scarce resource, as we also inferred from the data shown in Fig 6   422 in the previous subsection. To probe this more deeply, we examine the absolute fitness Surprisingly, we found that S. aureus fitness is unchanged by proximity to an 453 aggregate of P. aeruginosa (Fig 8). This was true for all initial cell densities, and all initial 454 ratios of the two species, except at moderate inoculation density where the S. aureus 455 fitness is actually higher next to a P. aeruginosa aggregate. These results imply that  aeruginosa is reduced and is no longer statistically significant (Fig 9) This suggests that 469 active production and release of a factor by S. aureus is important for conferring a fitness  The finding that increased Psl production benefits single cells more than 507 aggregates (Fig 5)  that also led to greater tendencies to aggregation. Such aggregative tendencies could 515 also have promoted the development of additional group benefits associated with P. 516 aeruginosa aggregates and biofilms [25]. Because we find that the iron-associated growth 517 benefit for aggregates does not depend on the density of P. aeruginosa competitors for 518 growth resources, there is no reason to believe that spatial structure is a major driver for 519 this interaction. This is strikingly unlike our earlier finding for P. aeruginosa monoculture, 520 in which competition and the spatial structure of both aggregates and the environment 521 are essential to the growth advantage for aggregates [22]. Releasing the requirement for 522 spatial structure greatly broadens the range of natural scenarios in which the aggregate 523 growth advantage could provide a selective advantage on which evolution might act.

524
Implications for biofilm disease 525 In addition to being a more realistic representation of natural environments, acquisition would be a chemical and metabolic advantage of Psl production that 539 complements and is orthogonal to its effects in increasing mechanical robustness.