3.7. Compatibility of biochemical and gene analyses and PCA results of some CFEC-ESBL
These results from MALDI-TOF MS-based PCA analyses pointed to important clues that some of the 35 ESBL-CFEC isolates had significant differences. In this respect, the biochemical characteristics of the isolates with the highest affinity (CFEC-ESBL-68) and the lowest (CFEC-ESBL-38) and low affinity (CFEC-ESBL-90) to all isolates and findings for beta-lactam resistance genes were detected. The data is given in Figure 3F confirm all the PCA-based analysis results given above. Unlike the other 34 isolates, bla CTX-M-15is one of the beta-lactam resistance genes in only CFEC-ESBL-38 isolates. It is considered that the presence of 41% variance causes the scattering profile to be located farthest in the scattering profile and with a different line in the dendrogram. Besides, in both (CFEC-ESBL-38 and CFEC-ESBL-90) isolates, bla CTX-M,bla CTX-M-1, bla TEM,bla OXA-10 The absence of their genes may be a reason for the 87% affinity between them, as well as leading to their separation from the main cluster (n=33).
Discussion
In the present study, 122 isolates were identified as E. coli by two methods, traditional (biochemical tests) and the new-generation method (MALDI-TOF MS). The identification scores obtained from MALDI-TOF MS are between ≥2.000, 16% were between 1.700-1.999 and there was no definition below the cut-off value (≤1.699). Within the scope of phyloproteomic analysis, firstly, PCA analyses of all E. coli isolates were carried out and it was determined that there are isolates with different characteristics. Then, more detailed analyzes were carried out on a smaller group of E. coli with ESBL characteristics. With the support of all the analyzes performed, it was concluded that there were significant differences in the light of the data on variance, CCI index values, dendrogram, and scattering profile placements, even though hints on bacteria belonging to the same species are provided.
ESBL was detected in almost one of the 3 isolates. It is a very high rate especially according to the scales of industrial production facilities. Because E. coli is a very important strain in terms of food pathogens and antibiotic resistance profile. Unfortunately, similar or even higher rates of ESBL were detected in some studies. For example, Yang et al. [38] similarly reported that 22.9% of the bacteria which were ESBL- E. coli. In other studies, Badr et al. [39] determined the rate of ESBL producing E. coli isolated from chickens is 46.7%, while Gazal et al. [40] detected 66% and Fournier et al. [41] detected ESBL 84%. Cormier et al. [1] reported that this rate reached up to 90%. According to the findings of the present study, predominantly bla CTX-M-1resistance genes were detected. In previous studies, the most frequently reported resistance genes were CTX-Ms and their derivatives [42]. These genes, which are reported to be common on a global scale, may be an indication that bacteria are in contact with each other [2,39].
CFEC-ESBL-38 is a strain that came to the fore with its difference in the study as the only atypical ESBL strain that shows only indole (-) and lac (-) among 122 isolates.E. coli is lac (+) commonly and lactose permease enzyme (LacY protein) is a very important protein that enables the use of lactose inE.coli. However, in some bacteria, lac (-) variants occur due to deficiencies in the level of this enzyme encoded by this LacY gene [43,44]. Stępień-Pyśniak et al. [8] showed thatEnterecoccus faecalis and E. mundtii isolates were separated in the dendrogram with phyloproteomic analysis by using the spectral profiles of the isolates. In the same study, it was also noted that there was clustering of very similar strains in terms of phenotype and genotype according to galactosidase and mellobiose characteristics, and it was stated in the study that a single gelatinase negative isolate gave a different peak. Unlike the other 34 CFEC-ESBL isolates, it is thought that the presence of bla CTX-M-15 only in CFEC-ESBL-38 causes it to settle furthest in the scattering profile and with a different line in the dendogram with a variance of 41%. The expression of a peptide/protein that is directly or indirectly related to phenotypic resistance might be cause the difference of this strain [45-47].
In the I. Phyloproteomic study, the CFEC-ESBL-90 was completely excluded from the large cluster with 14% variance in the full dendrogram (n=122, for CFEC) profile while it was located on a separate line in the dendrogram connected to the large cluster with 15% variance in the II. Phyloproteomic study (n=35, for CFEC-ESBL). Additionally, it was determined that the closeness ratios (CCI) to large cluster members, except for a few, were at a low level. Biochemically, this strain had weak catalase ability. The presence of catalase enzyme is very characteristic of E. coli. In general, E. coli harbors two different catalase genes: katG encodes hydroperoxidase I (HPI) and katE encodes HPII. The activity of both catalases increases when both are present together and the expression level of genes increases [48]. In the dendrogram profile, the difference in common with the members of the cluster including CFEC-ESBL-90 (CFEC-19 and CFEC-ESBL-38) is that it does not have hemolysis activity. The presence of the hemolysis enzyme and the observed hemolysis ability are mostly variable within the species in E.coli [49,52]. On the other hand, another possible feature makes this strain different from others is that the indole test is negative. Because the indole test is an indicator of the tryptophanase and tryptophan permease enzymes of the tryptophanase operon (TNA Operon), and the indole test of E. coli (90-95%) is mostly positive [51]. Deficiency of these proteins was also determined as a possibile affecting all PCA analysis results. Torres-Corral and Santos [52] pointed out that Lactococcus garvieae isolates gave characteristic 3 peaks and as the reason for grouping of isolates, enzymes such as epimerase, methyltransferases, and acetylphophatases possessed by the isolates. There are some studies suggesting that changes in enzyme structures may be affect results because protein analyses of biological structures are performed with MALDI-TOF MS [53,54]. On the other hand, according to the present study, catalase, tryptophanase, and tryptophan permease enzymes ofE . coli may be the reasons for the differences, but it must be supported by further studies. Because bacteria produce many specific and non-specific proteins [55].
Besides, other genes (which were not tested in the study) are considered to responsible for ESBL in CFEC-ESBL-90. The majority of genes responsible for ESBL appear as CTX-M, SHV, and TEM variants. However, there are also other genes responsible for resistance [39, 56-58]. In the study of Laudy et al. [59], similar to the present study, although ESBL-producing Pseudomonas aeruginosa strains obtained from phenotypic test results, they could not detect all the genes it screened at the same rate. They also reported new 3 different ESBL-producing genes with their further studies. Because many genes responsible for beta-lactamases were reported and continue to be reported [2,39,59]. Further studies can be carried out to detect ESBL genes in CFEC-ESBL-90.
In the present study, as well as the detection of differences betweenE.coli and ESBL -E.coli, it was observed that similar ones were numerically higher. For example, the greatest affinity was detected in the CFEC-ESBL-68 isolate. Although the ratio of CFEC-ESBL-68 to 13 cluster members was 49-69%, the closeness ratio to the remaining 21 cluster members was 70-99%. This might be because CFEC-ESBL-68 has typical biochemical features for E.coli like most isolates in the study (strong catalase property, alpha hemolysis ability, indole and lactose positive, etc.). However, performing further tests (other simple sugar fermentation, gelatinase, nitrate, arginine, biofilm, multidrug resistance profile, etc.) are needed to understand.
The present study was conducted to determine the differences and similarities between E. coli isolates with all PCA analysis (Dendrogram, scatter plotting, variance, and CCI). It was also found that the ESBL group generally differed from susceptible strains and there were some heterogeneities and homogeneities in the isolates. Alharbi et al. [17] showed that MSSA and MRSA can be separated in dendrogram cluster analysis. Similarly, in another study, it was shown that MSSA and MRSA were distinguished by peaks of different masses, and it was emphasized that MALDI TOF MS saves time according to molecular studies [60]. All these data suggest that ESBL producing E. coli are phylogenetically separated and may differ greatly in natural ecosystems. Previous studies have shown the feasibility of MALDI-TOF-MS for the clonal identification of bacteria. Our study is also an example for that.
In conclusion, phyloproteomic analyses with MALDI-TOF MS may be useful for characterization of phenotypic behaviors. Unfortunately, the high cost of analysis of multiple samples with high-cost methods such as Whole Genome Analysis is a major challenge for analytical studies. Thus, the most important result was found in the present study is that performing advanced analyzes as well as identification with the less costly MALDI-TOF MS contributes significantly to the validation of traditional analysis results. Also, this study represents a first in terms of ESBL screening and characterization in broiler chickens for the region (Duzce, Türkiye). There is no previous study that was conducted or published in the region. In recent years, epidemiological studies have focused on the spread of resistant strains which are extremely important in terms of clinically and food safety. In this sense, it is anticipated that this study will contribute to the monitoring of the data in the region. Although important clues were obtained, further analyzes are planned to make sense of the effect of biochemical characteristics on variance values.