Enterococcus faecalis inhibits Klebsiella pneumoniae growth in polymicrobial

19 Catheter-related urinary tract infections are one of the most common biofilm-associated diseases. 20 Inside biofilms, bacteria cooperate, compete, or have neutral interactions. The aim was to study 21 the interactions inside polymicrobial biofilms formed by Klebsiella pneumoniae and 22 Enterococcus faecalis, two of the most common uropathogens. 23 24 Although K. pneumoniae was the most adherent strain, it was unable to maintain dominance in 25 the polymicrobial biofilm due to the lactic acid produced by E. faecalis in a glucose-enriched 26 medium. This result was supported using the E. faecalis V583 ldh-1/ ldh-2 double mutant, which 27 not inhibited the growth of K. pneumoniae since this mutant does not produce lactic acid. 28 Lyophilized cell-free supernatants (L-CFS) obtained from E. faecalis biofilms also showed 29 antimicrobial/antibiofilm activity against K. pneumoniae. Conversely, there were no significant 30 differences in planktonic polymicrobial cultures. 31 32 In conclusion, E. faecalis modifies the pH by lactic acid production in polymicrobial biofilms, 33 compromising the growth of K. pneumoniae. 34

showing higher prevalence in urine samples were K. pneumoniae/E. coli, E. coli/E. faecalis, K. 56 pneumoniae/E. faecalis, and K. pneumoniae/P. mirabilis accounting for 26 %, 10 %, 8.5 %, 57 and 7 % of the cases, respectively (Galván et al. 2016). 58 Inside biofilms, social interactions of cooperation or competition between cells occur, and could 59 drive many changes or alterations in the community (Flemming et al. 2016). In fact, in contrast 60 to liquid cultures, these interactions are allowed by high cell concentration and diffusion 61 limitation (Rendueles & Ghigo 2012). The new technological developments have allowed the study of the diversity of highly complex microbial communities. Nevertheless, there is a lack 63 of knowledge about the implication of interspecies relationships that needs to be addressed 64 room temperature. Afterwards, CV was removed, rinsed once with 1x PBS and dried at 65 °C 134 for 60 min. Biofilm formation was quantified by eluting the CV fixed to the biofilm in 33 % 135 glacial acetic acid and absorbance of each well was measured at 580 nm (OD580nm) using a 136 microplate spectrophotometer (EPOCH 2 microplate reader; BioTek, VT). The experiment was 137 carried out in three technical and biological replicates. 138 Percentage of biofilm formation inhibition 139

Cultivable cells quantification 147
The number of cultivable cells from disrupted biofilms was obtained by colony counting. In 148 brief, after the aspiration of supernatants, the wells were rinsed once with 1x PBS to remove 149 non-attached cells. The plates were then sonicated at 40 kHz for 1 min, following the protocol 150

Competition in planktonic cultures 172
The same volume of ~ 1 x 10 7 CFU mL -1 of each strain was mixed and incubated at 37 °C with 173 shaking at 180 rpm. Aliquots were taken after specific time points (24, 48, and 72 h), serially diluted 10-fold when needed and plated for colony counting as previously described. The cell 175 count values were expressed as CFU mL -1 . Assays were performed in triplicate. After each cell 176 count, the culture was centrifuged, the supernatant was discarded and replaced by fresh medium 177 and the cells resuspended in the new one. 178

Supernatant analysis 179
Lyophilized cell-free supernatants collection (L-CFS) 180 The E. faecalis supernatants from biofilms were collected after 24 h of incubation. A portion 181 of each supernatant was adjusted at pH 6.5 with sodium hydroxide (NaOH) 1 M. Then, the total 182 volume collected was centrifuged for 15 min at 12,000 × g (at 4 °C) and passed through a 0. incubated at 37 °C for 24 h in static. Each well was rinsed once with sterile 1x PBS and the 207 remaining biofilms were quantified following the CV staining procedure described previously. 208 To evaluate the effects of L-CFS on pre-formed K. pneumoniae biofilms, the following method 209 was carried on. After 24 h of incubation at 37 ºC in static conditions, each well containing the 210 established biofilm was carefully rinsed once with sterile 1x PBS and treated with L-CFS at 211 different concentrations. The microtiter plates were then incubated at 37 °C for another 24 h in 212 static, and quantified using the CV staining procedure. 213 In both assays, a negative control (culture medium without inoculum) and positive control 214 (culture medium with bacterial inoculum) were included in each plate. 215 Both inhibitory and eradication capacities of DL-Lactic acid 85% towards K. pneumoniae 216 biofilms were also measured following the same protocol. The experiments were carried out in 217 three technical and biological replicates.

Determination of lactic acid 219
Quantitative detection of lactic acid in cell-free supernatants was performed using the L-lactic 220 acid Kit (BioSystems S.A.). The method is based on lactic acid oxidation. L-lactic acid in the 221 sample generates, using the reaction described below, NADH, which can be measured by 222 spectrophotometry. Measurements were made on the Analyser Y15 (BioSystems S.A.). 223

Biofilm development using lactic acid E. faecalis mutant strains 226
E. faecalis V583 wt and its mutant strains with deletions in either ldh-1, ldh-2, or both genes 227 were used to assess the inhibition caused by lactic acid production. Development and 228 quantification of mono-and polymicrobial biofilms using counting of bacterial CFUs were 229 performed as described above. The pH of the supernatants was also measured. 230

Interspecies interaction using pooled human urine 231
To evaluate the interactions between some of the strains, human urine was collected from six 232 healthy volunteers of both sexes who had no history of urinary tract infection. Urinalysis 233 showed normal parameters (glucose, ketones, nitrites, leukocyte esterase, bilirubin, 234 urobilinogen, blood, and proteins). The urine was pooled, filter sterilized and stored at 4 ºC. 235 Urine pH was 6.5 at the beginning of the analysis. Mono and polymicrobial biofilms were 236 developed using the pooled human urine with or without glucose 1 %, and quantified by 237 counting of bacterial CFUs as previously described. The competitive index was also calculated. Tukey's multiple comparisons test was used to analyse adhesion to abiotic surfaces. 248 Confirmation of inhibition by lactic acid production, using E. faecalis V583 wt and its mutants, 249 was analysed by confidence intervals on the difference between means. Tests with P values < 250 0.05 were considered significant. 251

Biofilm assays 253
Adhesion to abiotic surfaces 254 The time-dependent adhesion to polystyrene plates was measured by conventional plating 255 ( Figure 1 and Figure S1). For all these strains, the number of adherent bacteria increased during 256 the incubation period. K. pneumoniae AT was the most adherent strain, with an increase of 2.42 257 % after 60 min of incubation compared to the initial inoculum. Among the E. faecalis strains, 258 Ef 2 increased its adhesion in a 1.20 % after 60 min of incubation, and Ef 3, Ef 5, and Ef V583 259 reached a percentage of 1.50 %, 1.51 %, and 1.18 % respectively, after 120 min of incubation.

Percentage of biofilm formation inhibition 266
When assessing the interaction between the two pathogens within the biofilm, the reduction of 267 the total biomass of polymicrobial biofilms formed by K. pneumoniae and E. faecalis compared 268 to the sum of the total biomass of monomicrobial biofilms of each strain was statistically 269 significant (p < 0.001) in all comparisons, being the inhibition observed expressed in 270 percentages in Table 2. The same effect was observed using the other K. pneumoniae strains 271 (Table S1). These results suggest that the co-cultivation of K. pneumoniae and E. faecalis in a 272 polymicrobial biofilm significantly compromised their biomass formation compared to those 273 formed individually. 274 [ Table 2 near here] 275

Cultivable cells quantification 276
Results on the cultivable bacterial quantification after polymicrobial biofilm growth at different 277 time points are presented in Figure 2 and Figure  The antibacterial and antibiofilm effects of commercial lactic acid were also measured against 331 K. pneumoniae, being the MIC value = 1.25 mg mL -1 , MBIC value = 4 mg mL -1 , and MBEC 332 value was > 256 mg mL -1 .
To confirm that the decrease in pH was due to the production of organic acids, the lactic acid 335 concentration of supernatants collected from biofilms was measured. An important 336 concentration of lactic acid was detected in supernatants (Table 3 and Table S2), which may 337 confer the observed antibacterial and antibiofilm activities against K. pneumoniae. These 338 results are consistent with the MIC values obtained with commercial lactic acid, where a 339 concentration of 1.25 mg mL -1 inhibited K. pneumoniae growth. As well as the pH, lactic acid 340 of the E. faecalis supernatant was lower in monomicrobial than in polymicrobial, because, in 341 the second one, two kinds of species with different growth rates are competing for nutrients, 342 and Klebsiella pneumoniae, which has a higher growth rate than E. faecalis, also use up the 343 glucose and E. faecalis has not enough to produce the same lactic acid than produced when it 344 grows in monomicrobial biofilms. 345 [Table 3 near here] 346

K. pneumoniae biofilm development at different pH conditions 347
To define the influence of pH in the growth and the subsequent biofilm development of K. 348 pneumoniae, TSB medium was adjusted with NaOH 1M at pH ranging from 3.5 to 7.0, with 349 intervals of 0.5. Biofilms were established following the protocol of development and 350 quantification of biofilms and then incubated in static conditions at 37 °C for 24 h. After 351 incubation, biofilm production was quantified using CV staining. The results of OD 580 nm 352 showed that the lowest pH at which K. pneumoniae AT can form biofilm was 5.0, being 7.0 the 353 optimal pH value to develop a strong biofilm. This condition corresponds to the pH used in 354 conventional culture media (Figure 6 and Figure S7). 355 [ Figure 6 near here]

Biofilm development using lactic acid E. faecalis mutant strains 357
E. faecalis possesses two cytosolic L -(+) -lactate dehydrogenases encoded by the ldh-1 and 358 ldh-2 genes. Most of the activity is associated with LDH-1, and LDH-2 plays only a minor role 359 (Fatima Rana et al. 2013). Therefore, polymicrobial biofilms formed by E. faecalis V583 wt or 360 V583 Δldh-2 displayed the same inhibitory effect over K. pneumoniae observed previously with 361 the other E. faecalis clinical strains tested in this study. However, when E. faecalis V583 Δldh-362 1 or Δldh-1/Δldh-2 double mutant were analysed, the colony count of K. pneumoniae was not 363 statistically affected when compared to monocultures (Figure 7). Confidence intervals on the 364 difference between means showed statistically significant differences between means of K. 365 pneumoniae AT monomicrobial and K. pneumoniae AT co-cultured with E. faecalis V583 wt 366 or E. faecalis V583 Δldh-2. The difference between means of K. pneumoniae AT 367 monomicrobial and K.pneumoniae AT co-cultured with E. faecalis V583 Δldh-1 or E. faecalis 368 V583 Δldh-1/Δldh-2 was not statistically significant. The same effect was observed using the 369 other K. pneumoniae strains ( Figure S8). The competitive index showed an advantage of E. 370 faecalis over K.pneumoniae with all the strains used, but the difference of the obtained values 371 when E. faecalis V583 or E. faecalis V583 Δldh-2 were used (-3.63 and -2.52 respectively after 372 24 h of incubation), is higher than the obtained when E. faecalis V583 Δldh-1 or E. faecalis 373 V583 Δldh-1/Δldh-2 were in the polymicrobial biofilm (-0.74 and -0.59 respectively after 24 h 374 of incubation). In the same way, the pH decrease in the polymicrobial cultures using E. faecalis 375 V583 or E. faecalis V583 Δldh-2 was enough to inhibit the K. pneumoniae growth. Although 376 the decrease in the pH could be done by other organic acids produced, the loss of lactic acid 377 production in these E. faecalis mutant strains (V583 Δldh-1 or Δldh-1/Δldh-2 double mutant) 378 made these values not as lower as the wt V583 or the V583 Δldh-2 strain, causing less alteration 379 on K. pneumoniae growth (Table S3). 380 [ Figure 7 near here]

Interspecies interaction using pooled human urine 382
Using pooled human urine with and without glucose, the urine conditions of diabetic and non-383 diabetic patients were simulated. Neutral interactions between the strains were found when the 384 urine without glucose was used. However, the same inhibitory effect of E. faecalis over K. to a benefit for one of the species involved, based on nutrient competition or by inhibiting the 397 proper growth of their counterparts, a mechanism known as antagonism (Harrison 2007). Thus, 398 the co-culture of different bacteria in the biofilm state can lead to an increase or decrease in 399 their biomass. The third scenario is in which neither synergism nor antagonism is evident 400 among the species involved. Therefore, in this case, their interaction is classified as neutral. 401 Considering that K. pneumoniae and E. faecalis are common uropathogens, and biofilm 402 formation is an important trait in their pathogenesis, the study of their interspecies interactions 403 within biofilms seems mandatory. This approach could help identify possible targets or new 404 antimicrobial compounds, mainly produced by predominant strains, with therapeutic activity. can resist and adapt to different pH ranges growing in highly acid conditions (pH 2.9) (Rince 505 et al. 2000;Mubarak & Soraya 2018). Moreover, it is well known that growth inhibition of 506 different Gram-negative pathogens in urine occurs at pH 5.0 and below (Kaye 1968 inhibitor due to lactic acid production when is adhered to catheters, avoiding their adhesion by 516 competition or by reducing available nutrients, which should be explored further in future 517 research. All these results make us continue the study of potential lactic acid bacteria as 518 biocontrol agents to tackle the problematic emergence of antibiotic resistance and, in this case, 519 against biofilm formation on indwelling devices related to urinary tract infections. 520

Conclusions 521
K. pneumoniae and E. faecalis interact competitively when grown in biofilms in a rich glucose 522 environment. Both microorganisms produce more biomass in monomicrobial than in 523 polymicrobial biofilms. E. faecalis has shown to exhibit inhibitory activity against K. 524 pneumoniae, modifying the pH as a result of lactic acid production, which originates deleterious 525 effects over K. pneumoniae but without compromising their growth. However, the complex 526 network of interspecies interaction between this polymicrobial biofilm and others needs further 527 investigation.