TY - JOUR
T1 - Discrimination of non-typhoid salmonella serogroups and serotypes by Fourier transform infrared spectroscopy
T2 - a comprehensive analysis
AU - Campos, Joana
AU - Sousa, Clara
AU - Mourão, Joana
AU - Lopes, João
AU - Antunes, Patrícia
AU - Peixe, Luísa
N1 - Publisher Copyright:
© 2018
PY - 2018/11/20
Y1 - 2018/11/20
N2 - Simpler, quick and low-cost methods for routine Salmonella enterica typing are required for epidemiologic surveillance of this important zoonotic pathogen. In this study, using a comprehensive isolate collection, we investigated the potential of Fourier transform infrared spectroscopy (FTIRS) to discriminate the most clinically-relevant serogroups and serotypes of non-typhoid Salmonella. Moreover, the role of O-units composition on the FTIRS Salmonella discrimination was also explored. S. enterica isolates (n = 325; 2002–2015; different sources and countries), of 57 serotypes and 15 serogroups [including the most frequent ones, B-n = 122; C-n = 108; D-n = 43 and E-n = 33)] were analysed by FTIRS. Infrared spectra were analysed by Partial Least Square Discriminant Analysis (PLSDA) and/or Principal Component Analysis (PCA). The polysaccharides region provided the spectral sharpest differences being used in the subsequent Salmonella typing. Serogroups (B, C, D and E) discrimination was achieved with high accuracy (99.6% of correct assignments; PLSDA model). Differences in the O-unit structures composition of those serogroups are likely justifying the discrimination achieved. Other serogroups (G, H, K, L, M, N, O, T, U, Y, Z) were correctly predicted as not belonging to serogroups B, C, D nor E, except for 3 isolates of serogroups H (S. Sundsvall, n = 1) and K (S. Cerro, n = 2). In fact, O-unit structure of serogroup H and K shows some similarity with sub-serogroup C1 with the remaining serogroups presenting marked differences in this cellular component. The sub-serogroups discrimination was successfully achieved for C1, C2 and C3 (using PCA), and for E1-E2-E3 and E4 (by PLSDA). Appropriate serotype discrimination was obtained for most of S. Rissen from the remaining C1 serotypes (91.5%-PLSDA), and S. Enteritidis (D1) from the remaining D1/D2 serotypes (93.4%-PLSDA). The lack of available O-unit composition for particular serotypes prevents the elucidation of the role of this cellular component on the discrimination at serotype level obtained. FTIRS was able to discriminate relevant serogroups (B, C, D and E), sub-serogroups (C1, C2 and C3; E1-E2-E3 and E4) and particular important serotypes (S. Enteritidis, S. Rissen and S. Senftenberg). Further studies on O-antigen composition would clarify the fundaments of discrimination obtained by FTIRS.
AB - Simpler, quick and low-cost methods for routine Salmonella enterica typing are required for epidemiologic surveillance of this important zoonotic pathogen. In this study, using a comprehensive isolate collection, we investigated the potential of Fourier transform infrared spectroscopy (FTIRS) to discriminate the most clinically-relevant serogroups and serotypes of non-typhoid Salmonella. Moreover, the role of O-units composition on the FTIRS Salmonella discrimination was also explored. S. enterica isolates (n = 325; 2002–2015; different sources and countries), of 57 serotypes and 15 serogroups [including the most frequent ones, B-n = 122; C-n = 108; D-n = 43 and E-n = 33)] were analysed by FTIRS. Infrared spectra were analysed by Partial Least Square Discriminant Analysis (PLSDA) and/or Principal Component Analysis (PCA). The polysaccharides region provided the spectral sharpest differences being used in the subsequent Salmonella typing. Serogroups (B, C, D and E) discrimination was achieved with high accuracy (99.6% of correct assignments; PLSDA model). Differences in the O-unit structures composition of those serogroups are likely justifying the discrimination achieved. Other serogroups (G, H, K, L, M, N, O, T, U, Y, Z) were correctly predicted as not belonging to serogroups B, C, D nor E, except for 3 isolates of serogroups H (S. Sundsvall, n = 1) and K (S. Cerro, n = 2). In fact, O-unit structure of serogroup H and K shows some similarity with sub-serogroup C1 with the remaining serogroups presenting marked differences in this cellular component. The sub-serogroups discrimination was successfully achieved for C1, C2 and C3 (using PCA), and for E1-E2-E3 and E4 (by PLSDA). Appropriate serotype discrimination was obtained for most of S. Rissen from the remaining C1 serotypes (91.5%-PLSDA), and S. Enteritidis (D1) from the remaining D1/D2 serotypes (93.4%-PLSDA). The lack of available O-unit composition for particular serotypes prevents the elucidation of the role of this cellular component on the discrimination at serotype level obtained. FTIRS was able to discriminate relevant serogroups (B, C, D and E), sub-serogroups (C1, C2 and C3; E1-E2-E3 and E4) and particular important serotypes (S. Enteritidis, S. Rissen and S. Senftenberg). Further studies on O-antigen composition would clarify the fundaments of discrimination obtained by FTIRS.
KW - Salmonella enterica
KW - Typing
KW - Fourier transform infrared spectroscopy-attenuated total reflectance
KW - Multivariate data analysis
KW - Bacterial identification methods
KW - O-antigen
UR - http://www.scopus.com/inward/record.url?scp=85049795668&partnerID=8YFLogxK
U2 - 10.1016/j.ijfoodmicro.2018.07.005
DO - 10.1016/j.ijfoodmicro.2018.07.005
M3 - Article
C2 - 30015261
AN - SCOPUS:85049795668
SN - 0168-1605
VL - 285
SP - 34
EP - 41
JO - International Journal of Food Microbiology
JF - International Journal of Food Microbiology
ER -