Most previous studies of biofilm formation have been performed under one or two conditions to present this phenotype. However, biofilm is a kind of "smart community" that seems able to cope with different environments. Therefore, a single condition may lead to misunderstanding regarding the elaborate function of a specific gene. To provide sufficient and rigorous evidence, we demonstrate that the LuxS/AI-2 system is involved in the regulation of biofilm formation under different conditions. In contrast to the most commonly used microtitre plate assay, flow cell is increasingly used for detecting biofilm growth, of which the dynamic three-dimensional image could be monitored by CLSM dynamically. This setting simulates the environment of flowing surfaces in vivo, such as some interfaces between body fluids and implants. The result under this condition may offer more accurate information about flow surroundings in vivo. In addition, we also investigated biofilm formation under anaerobic conditions, which the biofilm bacteria undergo. The oxygen partial pressure of air-equilibrated medium is high in vitro, whereas hypoxic environments may prevail in body implants and human tissues distant from arterial blood flow [58, 61]. Moreover, most locations in which the biofilm bacteria accumulate are usually niches of local low oxygen because of inflammatory cell aggregation [59, 62].
The mouse model was used here to compare biofilm formation of the WT and the ΔluxS strains and our data were consistent with the in vitro data. Nevertheless, there are few studies regarding AI-2 complementation in the mouse model to date, and the specific mechanism of these signal molecules in vivo remains elusive. In general, we offer consistent results under different conditions to emphasise that the conclusion drawn is consistent and worthy of reference in most cases.
LuxS and AI-2 affect biofilm development, whereas the results may be different in the same strain due to various factors. Previous work has shown that AI-2 was produced in rich medium under aerobic and anaerobic conditions peaking during the transition to stationary phase, but cultures retained considerable AI-2 activity after entry into the stationary phase under anaerobic conditions. In addition, the S. aureus RN6390BΔluxS strain showed reduction in biofilm formation in TSB containing 1% glucose and 3% sodium chloride under anaerobic conditions . However, in our study, analysis of biofilm growth in vitro and in vivo led to the conclusion that the ΔluxS strain exhibited increased biofilm formation compared to the WT strain. Importantly, the luxS mutation could be complemented by chemically synthesized DPD, indicating the effect of AI-2-mediated QS on biofilm formation in S. aureus. Hardie and Heurlier  summarised six main factors that influence the experimental results for doing research on the LuxS/AI-2 system: experimental design; genetic complementation; chemical complementation; conditioned supernatant complementation; and complementation with molecules linked to AI-2 production and that independent of luxS status. With detailed analysis, we found that the inconsistency of the results is in part owing to the different growth medium provided to the biofilm bacteria, especially the different concentrations of glucose and sodium chloride, which are both important factors enhancing biofilm formation .
In addition to the present evidence of AI-2-regulated biofilm formation in S. aureus, we found that the transcription of icaR is activated by AI-2, which is barely reported, although we have not yet identified the mechanism of the interaction between them. This is partly because the detailed mechanism of transport and action of AI-2 has only been described in several strains and different mechanisms are applied to different species, although AI-2 has been proven to act as a signalling molecule that could regulate series of gene expression. The first mechanism revealed was in Vibrio harveyi, which responds to AI-2 by initiating a phosphorylation cascade . In Salmonella typhimurium and E. coli[66, 67], AI-2 seems to be taken up by an ABC transporter. However, the mechanism of AI-2 transport and functional performing in Staphylococci was still unknown. Therefore, the detailed mechanism through which AI-2 functions in S. aureus should be highlighted here, and the interaction between AI-2 and IcaR requires further study.
In addition to PIA, we do not observe any obvious increase of extracellular protein (Additional file 2: Figure S2) or bacterial autolysis in the ΔluxS strain (Additional file 3: Figure S3). Our results showed no significant differences in the transcriptional levels of several main adhesion molecules. Moreover, previous work indicated that S. aureus strains 8325-4 and RN4220 formed PIA-dependent biofilms . We thus propose that AI-2 signalling represses the icaA expression, and subsequently leads to decreased biofilm formation in S. aureus.
More and more studies concerning multispecies biofilms gradually uncover the mechanisms of the interaction and communication of the different species inside the biofilms. One of the most popular approaches of the signalling regulation is directed towards the AI-2-controlled QS system for its extensive use of interspecies. The plaque biofilms on tooth surfaces consist of various oral bacteria including S. aureus and involve complex microbial interactions [69–71]. There is evidence that AI-2-mediated QS may play a significant role in oral biofilm formation . As reported by others, airway infections by Pseudomonas aeruginosa afflicting patients with cystic fibrosis (CF) are among the most enigmatic of biofilm diseases . S. aureus is also found to be a major pathogen associated with P. aeruginosa in CF lung infection . Previous work has shown that PIA is expressed in lungs infected with S. aureus, whereas CP8 is not expressed because of limited oxygen . Here, we provide evidence that AI-2 can regulate icaA expression under anaerobic conditions, suggesting a potential role of AI-2 in influencing S. aureus infection in lungs. However, few studies about biofilm formation cooperated by S. aureus and the other species are reported. Therefore, could S. aureus and the other species in their focus areas form multispecies biofilms? Could AI-2 play an important role in this process? It is interesting to discuss the actual complex-flora interaction in human and social behaviour of the bacteria. Therefore, revelation of the AI-2-regulated biofilm formation in S. aureus possesses instructive meaning for these related studies.