Strain selection and preparation
To study the potential of Pseudomonas putida as BCA against Salmonella, several field strains from broiler houses were used (Table 1). The P. putida strains were in a previous study (Maes et al., 2019) classified as weak (P1), moderate (P2) and strong (P3) biofilm formers in 96-well microtiter plates (MTPs). This classification was based on the absorbance measured at 590 nm after crystal violet staining of the biofilms, which was divided into groups according to Stepanović et al. (2000).
For the preparation of the bacterial suspension for inoculation in the in vitro biofilm model, strains were streaked on Plate Count Agar (PCA, Oxoid, CM0325, Basingstoke, Hampshire, England) from their glycerol stocks at -80°C and incubated for 24h at 37°C for Salmonella strains and 48h at 30°C for Pseudomonas strains. Subsequently, one colony from PCA was transferred to a test tube containing 10mL of Tryptone Soya Broth (TSB, Oxoid, CM0129). An overnight culture was obtained by incubating the broth for 18h at 30°C for Pseudomonas (8 log CFU/mL) or 18h at 37°C for Salmonella (9 log CFU/mL) strains. Quantification of the overnight culture was done by plating on PCA and incubation for 72h at 30 or 37 °C depending on the species. Finally, overnight cultures were diluted in sterile ¼ Ringer’s solution (Biokar, BR00108, Beauvais, France) to the desired density (3 or 6 log CFU/mL depending on the corresponding biofilm set-up) and the resulting suspension is called the inoculum suspension. Actual inoculum densities were calculated based on the CFU quantification of the overnight cultures and were taken into account in the study of the interactions between strains in dual-species biofilms.
Coupons, 35x10x2mm, were cut from new plastic drinking water lines commonly used in broiler houses (Swii’Flo, Roxell, Maldegem, Belgium). Before use, these coupons were sterilized in 70% ethanol for 10min and dried in a laminar flow cabinet. The sterilized coupons were vertically placed in the wells of a 6-well MTP (Novolab, SPL30006, Geraardsbergen, Belgium) using sterilized tweezers in a way that only the 2mm sides of the coupons touch the wells.
Mono- and dual-species biofilm formation
Attachment of the bacterial strains
For each independent test (n) of each tested strain (combination) and condition, a 6-well MTP was used for biofilm formation. In this plate, 11mL of inoculum suspension was added per well (technical replicates = r) to completely submerge the coupons. A second 6-well MTP was used as blank control. In this plate, 11mL of diluted TSB (equally diluted as the inoculum suspension) was applied in three wells. Only for the validation of the model, a third MTP was used for biofilm formation. Well plates were incubated in an incubator with shaker (Adolf Kuhner ag, LT-V 89799.89, Basel, Switzerland) for four hours at 25°C and 50 rpm making attachment of the bacteria to the coupons possible. After incubation, coupons were removed from the 6-well MTPs and transferred to 15mL falcon tubes (Sigma-Aldrich, Z720461-50EA, Overijse, Belgium) using sterilized tweezers. To remove non-attached bacteria, coupons were rinsed once by submerging them in 10mL sterile ¼ Ringer’s solution in the falcon tubes. Afterwards, the ¼ Ringer’s suspension was discarded and coupons with attached bacteria were placed vertically in new 6-well MTPs.
Biofilm formation by the attached bacterial strains
The new MTP’s with coupons were filled with 11mL of sterile 1/20 diluted TSB per well and subsequently incubated for 18h at 25°C and 50rpm to allow the attached bacteria to form biofilm. After incubation, coupons were removed from the 6-well MTPs and transferred to 15mL falcon tubes using sterilized tweezers. To remove non-attached bacteria, coupons were rinsed three times by consecutively adding 10mL of sterile ¼ Ringer’s solution in the falcon tubes and discarding the suspension. Finally, coupons were transferred to new sterile 15mL falcon tubes.
Quantification of biofilm formation based on bacterial counts
Coupons originating from the first MTP were used for quantification of biofilm formation by conventional microbial enumeration methods. Three blank control coupons from the second MTP were also counted to ensure no contamination did occur during analysis. First, 10mL of sterile ¼ Ringer’s solution was added to the falcon tubes containing the coupons. Then, three consecutive rounds of sonication for 30s at 42kHz in a ultrasonic water bath (Branson, 2510, Eemnes, The Netherlands) and vortexing for 30s were performed to harvest the biofilm.
The liquid suspension containing the detached biofilm cells was plated on Tryptone Soya Agar (TSA, Oxoid, CM0131) for enumerations of total aerobic count (TAC) and a second, more selective, medium. This selective medium was Xylose Lysine Desoxycholate Agar (XLD, Oxoid, CM0469) for Salmonella and Pseudomonas Agar Base (PAB; Oxoid, CM0559) with Pseudomonas CFC Selective Agar Supplement (Oxoid, SR0103) for Pseudomonas. Appropriate 10-fold dilutions were made in sterile 0,1% w/v Peptone Water with 0,85% w/v Salt (BioTrading, K110B009AA, Mijdrecht, The Netherlands) and pour plated. TSA plates were incubated for 72h at 30°C or 37°C for Pseudomonas or Salmonella biofilms, respectively. XLD plates were incubated for 72h at 37°C and PAB plates were incubated for 72h at 30°C. The limit of quantification (LOQ) for microbiological enumerations was 1,16 log CFU/cm².
Quantification of biofilm formation based on biomass
For the validation of the model, six coupons from the third MTP used for biofilm formation and three blank control coupons from the second MTP were used for quantification of biofilm formation based on biomass. 10mL of a 0.1% crystal violet solution (containing 0.1g/100mL crystal violet (Merck, 101418, Darmstadt, Germany) dissolved in one part of methanol (Biosolve, 13687802, CE Valkenswaard, The Netherlands), one part of isopropanol (Merck, 1.09634) and 18 parts of Phosphate Buffered Saline (Oxoid, BR0014G)) was added to each of the falcon tubes for 20min and shaken (Fisher Bioblock Scientific, KL2 6118 CU 00246, Merelbeke, Belgium) at 350rpm for the staining of the total biomass of the biofilm on the coupons. The excess stain was removed by placing the tubes under gently running tap water. Retained crystal violet was dissolved by adding 10mL of 33% acetic acid (Merck, 1.00063) for 15min at 350rmp. The absorbance was measured at 590nm using a spectrophotometer (Jasco, V-660, Pfungstadt, Germany). OD-measurements of the blank control coupons were subtracted from the OD-measurements of the biofilm coupons.
Study of interactions between bacterial strains in dual-species biofilms
In this study, the cooperation criterion and the biodiversity effect were calculated to determine social interactions between S. Java and P. putida and to consequently assess the potential of P. putida as BCA. The cooperation criterion requires that the inoculation density in co-culture equals the sum of inoculation densities of the monocultures whereas the biodiversity effect imposes that the inoculation density of each species in co-culture should be its inoculation density in monoculture divided by the number of species in co-culture (Parijs and Steenackers, 2018). A preliminary experiment was conducted growing mono-species biofilms of S. Java in both set-ups but no differences in final biofilm growth were observed. Therefore, both the cooperation criterion and the biodiversity effect were calculated based on the results of dual-species biofilms where the inoculation density equals the sum of the inoculation densities of the monocultures.
Concerning the cooperation criterion, counts for TAC of the dual culture were compared with the sum of the counts for TAC of the two monocultures of Salmonella and Pseudomonas. Also, counts for Salmonella spp. on XLD and Pseudomonas spp. on PAB were compared between mono and dual cultures.
The biodiversity effect can be calculated as follows (Loreau and Hector, 2001): see formula 1 in the supplementary files.
The selection effect comprises deviations from the expected productivity due to relative enrichment of strong biofilm or weak biofilm formers. A positive selection effect indicates enrichment of the strongest monoculture biofilm formers, whereas a negative selection indicates that the weaker biofilm producers are enriched. The complementarity effect measures to what extent deviations from the expected relative productivity are compensated by the other strains. It comprises all deviations from the expected productivity not explained by the selection effect. A positive complementary indicates some degree of niche separation between the different strains whereas a negative complementary points towards interference competition. Interpretation of these effects is further explained in the results and discussion section.
Statistical analyses on the obtained microbiological and biomass results were carried out using Statistical Analysis System software (SAS®, version 9.4, SAS Institute Inc., Cary, NC, USA). First, normal distribution of the OD measurements and of the log transformed enumerations per microbiological parameter per biofilm experiment were evaluated based on the histogram and QQ plot. For the evaluation of the reproducibility of the model system, a Kruskal Wallis test was used to compare results for OD measurement and enumerations between the three experiments per strain. For the comparison of mono-species biofilm formation of different bacterial strains, enumerations of TAC were evaluated per experiment using ANOVA. Post-hoc pairwise comparisons were made using Scheffe test. For the comparison of mono-species biofilm formation with different inoculum densities, enumerations of TAC were evaluated per experiment using a Kruskal Wallis test. Post-hoc pairwise comparisons were made using Dunn test. For the comparison of the quantification of different dual-species biofilms (with different strains, different inoculum densities or different application order) again a Kruskal Wallis test was performed on enumerations of TAC, Pseudomonas spp. and Salmonella spp. followed by a post-hoc pairwise comparison using Dunn test to indicate possible differences. P-values ≤ 0.05 were considered significant.