This work presents the differences between two methods, PFGE and WGS, when used to determine bacterial variability. On one hand PFGE determines a partial bacterial variability (variability of the restriction sites), statistical conformation (multivariate cluster statistical analysis) of approximate groups with the PFG running patterns of the genomes of the bacteria studied. However, some of the constitutive groups do not discriminate the pathophysiological behavior of alleged bacteria [17]. On the other hand, the WGS determines the variability of the bacterial genome and relates or associates it with its pathophysiological behavior [18].
The PFGE method described in other studies [19,20,21] is used to determine genomic variability in bacteria. The bacterial genomes have sequences that are targeted by restriction enzymes, becoming a strategy used among bacteria to defend themselves against infections and eliminate foreign DNA. PFGE uses a restriction enzyme to digest bacterial genomic DNA using the respective restriction sequences. Depending on the restriction sequences, a certain number of restriction fragments will be generated [17, 10]. The use of the enzyme SmaI generated about 20 restriction fragments from the genome of different strains of S. mutans, serotype c. An adequate number of restriction fragments that allowed to run patterns by PFGE, enabled multivariate cluster statistical analysis [10]. However, there are certain disadvantages: 1. PFGE does not discriminate between all unrelated isolates, 2. bands of the same size may not come from the same part of the chromosome, and 3. It does not differentiate isolates to the same degree when compared to WGS. Nevertheless, this method is still used to determine genomic variability and discriminate groups according to their pathophysiological behavior [22,23.24]. The studies that used PFGE to determine variability and identify bacteria by groups, concluded that they did not obtain any discrimination among the groups of the studied bacteria despite finding variability, although partial [17, 10]. Moreover, PFGE is a basic tool for epidemiological screening between pathogenic and innocuous bacteria [22, 23,24]. In the work of Rincon-Rodríguez et al., genetic variability was demonstrated in strains of S. mutans, serotype c, and although eleven (11) different clusters were formed, the authors were unable to completely separate the isolates according to their pathophysiology [10]. In this study, the variability obtained was partial, as observed in Fig. 2, where it is visible that the position of the restriction sequences varied in each of the strains studied. The restriction fragments generated were of different sizes. However, the sequences of these fragments were not the same, neither was their order.
The use of WGS is biotechnologically complex. Nonetheless, the running costs are affordable and allows to accurately determine bacterial genomic variability and its pathophysiological relationships (Palmer et al, 2013). The use of bioinformatic tools help to identify the global and specific comparative variability of the total sequences, genomic contents and chromosomal arrangements. The global and specific variability obtained in this study (Fig. 2) was consistent with the variability obtained by Maruyama et al. when two strains of S. mutans, serotype c were compared for the first time, showing differences in lengths, genomic contents, and chromosomal arrangements between the UA159 and NN2025 strains [18].
Restriction sequences are believed to be inherited by bacteria from their parental strain, which are conserved as they are inherited vertically, generating a fixed number of fragments. However, HGT may alter the number of these restriction sequences, equally affecting the number and size of the restriction fragments, since HGT randomly inserts or removes DNA within the bacterial genome by altering the length and order of DNA within each restriction fragment [25]. When the bacterial genome is subjected to enzymatic digestion and PFGE is run, it may occur that each bacterial genome generates a unique and different running pattern according to the variability of its restriction sequences and restriction fragment size due to the HGT event. Subsequently, the software takes each of the PFGE patterns for analysis, grouping similar PFGE patterns into clusters, as to each cluster formed differs from another cluster. Figure 2 shows how the selected strains presents different PFGE patterns, which is a consequence of the variability of the location sites of the restriction and HGT sequences that generate variable restriction fragments that is reflected in the patterns of each strain of S. mutans, serotype c, studied.
Regarding the WGS results of the five strains, it was observed that each strain, despite their differences in their bp length and the number of orthologous and non-orthologous proteins, belonged to the same species of S. mutans, serotype c. This was confirmed by the CG percentage in each strain, corresponding to 36.7% and 36.8% CG approximately. The variation is probably the product of the HGT event, bacteriophages, factors related to the micro-niche of origin and other events related to the host. The factors rather influenced the genomic variability of each strain and exerted an indirect effect on the variability of the position of the sequences of restriction. Thus, it is presumed that PFGE shows variability, although partially. Alternatively, sequencing provides more complex and deeper information about variability.
Another limitation of PFGE presented in the study by Rincón-Rodríguez et al., was the low number of restriction fragments used for the PFGE analysis, Fig. 1 (approximately 10 to 12 restriction fragments) [10]. The PFGE simulation carried out in this study used 20 to 21 restriction fragments (Fig. 2). It is likely that taking a smaller number of restriction fragments for PFGE analysis influences a more specific conformation of patterns with better arrangement of the clusters, thus allowing more discrimination. Nevertheless, in PFGE studies, the low count of restriction fragments alone does not explain why mixed groups cannot be constituted, decreasing the probability of screening related to the physiopathological function of the bacteria studied.
When the complete genomes of bacteria of the same species are compared using bioinformatics tools, for example MAUVE, variability in the sequence, contents and arrangements of chromosomal DNA between the compared bacteria may be observed (Fig. 3). Here, we obtained results compatible with those obtained in other studies, either with S. mutans, serotype c, or with other species of bacteria analyzed. The results from this comparison suggest which bacterial genomic portions were probably inherited vertically or acquired by HGT, and, in parallel, shows the variability for each member of the bacterial species analyzed. The various arrangements present in the analyzed bacterial genomes may also be observed (see the description in Materials and Methods). Furthermore, the WGS results allowed us to identify both global and specific genomic variability, variability of genomic contents that may be associated with pathophysiological behaviors, as well as the variability in the ordering of the genome.
This study demonstrated that despite the differences in the results of the PFGE and WGS methods, both show bacterial genomic variability. First PFGE, considering the recommendations of the respective protocols, shows variability in the position of the restriction sites of the analyzed bacteria. However, it does not discriminate based on their pathophysiology. Finally, WGS provides variability in nucleotide sequences and in gene content. It also showed the chromosomal ordering set in the genome of the analyzed bacteria, which is important information to connect or associate the variability of bacteria with their pathophysiological behavior.