Effect of oxygen and growth medium on in vitro biofilm formation by Escherichia coli

Abstract

The effects of oxygen availability on in vitro biofilm formation by an Escherichia coli K-12 strain and 13 clinical E. coli strains were compared. All E. coli strains were capable of forming monospecies biofilm on polystyrene in aerobic media. The K-12 strain produced biofilm in both aerobic glucose minimal medium (ABTG), and aerobic trypticase soy broth (TSB) whereas the majority of the clinical strains produced significant biofilm only in aerobic TSB (9 of 13). In anaerobic media, E. coli K-12 and 9 of the 13 clinical strains were capable of forming biofilm in vitro . Only three clinical strains formed biofilm in anaerobic TSB whereas six clinical strains produced detectable biofilm in anaerobic ABTG. None of the strains tested were capable of forming biofilm in both anaerobic ABTG and anaerobic TSB. Strains that were good biofilm formers in aerobic ABTG also produced the highest amount of biofilm in anaerobic ABTG ( R 2 = 0.90). Image analysis revealed notable differences in architecture for biofilms grown in the presence and in the absence of oxygen. In aerobic ABTG, the biofilm was dominated by tall, mushroom-shaped microcolonies with pores and channels whereas biofilm in anaerobic ABTG was thinner and less heterogeneous, resulting in reduced maximum thickness and biovolume. Analysis of phospholipid fatty acid (PLFA) profiles from E. coli K-12 and three clinical strains did not reveal a specific pattern associated with the biofilm phenotypes. Interestingly, the clinical E. coli strains adjusted their PLFA composition much more than did E. coli K-12 in response to changes in growth regimens. Collectively, the results indicate that oxygen availability may affect E. coli biofilm formation in minimal and complex media. The results confirm that E. coli K-12 and some clinical E. coli strains are capable of forming in vitro biofilm under anaerobic conditions. However, the data also suggest that this attribute is highly strain dependent and may vary significantly among clinical isolates.