LEE- encoded proteins considered critical for O157 adherence to FAE cells at the RAJ did not appear to have a role in O157 adherence to RSE cells at this same site. Both O157 strains grown in DMEM and pre-incubated with pooled, polyclonal antisera generated against the LEE (Tir, EspA, EspB, and Intimin) and flagellar H7 proteins, or the anti-Intimin antisera alone, at 1:5 and 1:10 dilution, continued to adhere to the RSE cells, irrespective of the presence/absence of D
Mannose. Data is shown for one of the O157 strains in the presence of D
Mannose (Additional file 1
, Figure , panel A, Figure ). These results were consistent between all trials, irrespective of toluidine blue or immunofluorescent staining, and did not show any differences in the adherence patterns compared to the controls. The same O157-RSE cell-adherence pattern was observed in the controls with normal rabbit sera added at 1:5 dilution (data not shown), and in the absence of any sera (Additional file 1
, Figure , panel B; Figure ) [5
], irrespective of the presence/absence of D
Mannose. The continued adherence of O157 to the RSE cells in the presence of antibodies to the LEE proteins may have been due to the masking of these antigens and the unmasking of other O157 adhesins targeting the receptors on the RSE cells. To that effect an increase in the total number of RSE cells with adherent bacteria and decrease in the total number of RSE cells with no adherent bacteria was observed, in the presence of pooled and anti-Intimin antisera (Figure ). We intentionally included antisera targeting the flagellar antigen H7 as flagella have been demonstrated to play a role in initial adherence to plant cells and the FAE [28
]. These results suggest that additional mechanisms of adherence, distinct from those attributable to LEE, Intimin and flagellar H7 proteins, are involved in O157 attachment to the RAJ squamous epithelial cells.
Figure 1 Adherence patterns of O157 strain EDL 933 on RSE cells, in the presence of D+Mannose and +/− antisera.Panel A, in the presence of “pooled antisera” against LEE, Intimin and flagellar H7 proteins, and the anti-Intimin (more ...)
Figure 2 Quantitative representation of the adherence patterns of O157 strains EDL 933, and 86–24 along with its mutant derivatives, on RSE and HEp-2 cells. Percent mean±standard error of mean of cells with adherent bacteria or (more ...)
On the other hand, the LEE-encoded proteins were critical to O157 adherence to HEp-2 cells as demonstrated previously [22
], with or without D
Mannose. As shown in Additional file 2
and Figure , panel A and Figure , preincubation with the pooled polyclonal antisera and the anti-Intimin antisera significantly interfered with and prevented O157 adherence to HEp-2 cells, at all dilutions tested. An increase in number of HEp-2 cells without any adhering bacteria was observed in the presence of either antiserum, accordingly (Figure ). However, pre-incubation with normal rabbit sera at 1:5 dilution (data not shown) showed the same diffuse, moderate adherence as in the absence of any antisera (Additional file 2
, Figure panel B and Figure ).
Figure 3 Adherence patterns of O157 strains on HEp-2 cells, in the presence of D+Mannose and +/− antisera.Panel A, O157 strain EDL933, in the presence of “pooled antisera” against LEE. Intimin and flagellar H7 proteins, (more ...)
The results observed with the adherence inhibition assays were further verified by the adherence patterns of O157 strain 86–24 (86–24) and its mutant derivatives on HEp-2 and RSE cells (Figure , panel C, Figures and ). The intimin-negative mutant 86-24eae
Δ10 did not adhere well to the HEp-2 cells compared to the intimin-positive, wild-type 86–24 or complemented mutant, 86-24eae
Δ10(pEB310) that demonstrated diffuse, moderate adherence (Figure , panel C, Figure , and Additional file 2
). Actin accumulation observed in the majority of HEp-2 cells with 100x magnification only in the presence of 86–24 and 86-24eae
Δ10(pEB310), along with an increase in the number of HEp-2 cells without adhering bacteria in the presence of 86-24eae
Δ10, further verified these observations (data not shown). This confirmed the role of intimin in O157 adherence to HEp-2 cells. On the otherhand, 86–24 and all its mutant derivatives demonstrated diffuse, strong adherence to RSE cells, irrespective of intimin expression (Figures and , and Additional file 1
). Infact with 86-24eae
Δ10, the number of RSE cells with adhering bacteria actually increased, which suggested that intimin did not have a role in the adherence of O157 to RSE cells.
Figure 4 Adherence patterns of O157 strain 86–24 (Intimin-positive) and its mutant derivatives, 86-24eaeΔ10 (Intimin-negative) and 86-24eaeΔ10 (pEB310) (Initmin-positive), on RSE cells, in the presence of D+Mannose. The (more ...)
In order to verify if proteins other than LEE proteins were being expressed by O157 upon growth in DMEM which could have a possible role in O157 adherence to RSE cells, we analyzed the O157 proteome as expressed in DMEM. While the proteome of O157 has been analyzed under various other growth conditions [30
] we decided to evaluate the same following growth in DMEM for several reasons, such as (i) this was the media used to culture both bacteria and the RSE cells, separately, prior to the adherence assays, (ii) the media closely mimicked the nutrient-limiting conditions seen in vivo, and (iii) this media closely matched that used to develop a commercially available cattle, O157 vaccine [15, 16; http://www.bioniche.com
. Our observations did not support a role for other host (RSE-cell)-derived factors in this adherence of O157 and hence, we did not evaluate RSE-cell adherence of O157 cultured in eukaryotic cell-conditioned media. This inference came from the fact that similar adherence results were obtained when DMEM was supplemented with norepinephrine (NE; DMEM-NE), a host neuroendocrine hormone that is encountered by O157 in vivo during the actual process of infection (data not shown). NE is reportedly a mimic of autoinduer 3 (AI-3), which regulates O157 virulence gene expression via quorum sensing [34
]. Further, Intimin, its receptor, Tir, as well as EspB were expressed in equivalent amounts in both DMEM and DMEM-NE, as observed using western blotting by others [34
], and by us, and also using top down proteomics by us (data not shown).
A total of 684 proteins were identified as being part of the O157 DMEM-proteome (13% of the O157 sequenced proteome), and these included several characterized and hypothetical/unknown proteins besides the TTSS proteins. While 171 of these proteins were uncharacterized with hypothetical functions assigned in the O157 genome [21; Figure , Additional files 3
and 5–12], the remaining 513 proteins localized to various bacterial cell compartments with functions including metabolic, cell division, regulatory, transport, environmental adaptation, and previously characterized O157 virulence factors [21
]; Figure , Additional files 4
. Proteins associated with O157 virulence or adherence in the DMEM-proteome included Tir, Intimin, EspB, LuxS, Iha, OmpA, KatP, ChuA, EspP, Stx1A, Stx1B, and Stx2B [20
]; Additional files 4
. Interestingly, 64 of the 684 (9.4%) proteins comprising the O157 DMEM-proteome were also part of the O157 immunoproteome in cattle, defined using the innovative proteome mining tool, Proteomics- based Expression Library Screening (PELS) [23
]; Additional files 3
. In addition, nine members of the DMEM-proteome were also part of the O157 immunome in humans [26
Bacterial cell localization of proteins comprising the O157 DMEM-proteome.
Given that a subset of pathogen proteins targeted by the host immune system help pathogens overcome hostile in vivo environments and rapidly adapt to host niches, counter host defenses, survive, propagate, and establish infection, it is very likely that proteins other than LEE proteins in the O157 DMEM-proteome play a role in O157 adherence to the RSE cells at the RAJ. The identification and inclusion of such proteins in anti-adhesion vaccine preparations, and their optimal administration to the host immune system, may enhance the efficacy of such vaccines in reducing or eliminating O157 not only from the FAE cells but also from the RSE cells of the bovine RAJ. Our strategy for selecting proteins constituting the O157 DMEM proteome with potential to function as adhesins for further evaluation was to employ the SPAAN algorithm [27
], which is part of the web-based vaccine target prediction and analysis system at http://www.violinet.org
. In particular, we shortlisted proteins that were shared by four sequenced O157 strains, namely, strain EDL933, strain EC4115, strain Sakai, and strain TW14359. Our analysis identified 36/684 components of the O157 DMEM-proteome to have adhesin potential, including the extensively reported primary O157 adhesin, intimin-γ, and its protein receptor, Tir (Table ). Three O157-specific proteins, namely ChuA, TerE, and a putative outermembrane porin protein on O-island #62 (Table ), have been prioritized and are under experimental evaluation for a role in O157 adherence to RSE cells. Also, towards studies of common STEC adherence mechanisms, we have determined homologs of several of these putative adhesins including those encoded on O-islands such as, Tir, Intimin, ChuA, TerE, EspP, bioinformatically (psiBLAST, National Center for Biotechnology Information; http://www.ncbi.nih.gov
), in sequenced non-O157 Shiga-toxin-producing Escherichia coli
(STECs) of clinical importance. Experimental evaluation of the contribution of these homologs to adherence of these non-O157 STEC, E. coli
O26:H11, O103:H2 and O111:NM, are underway.
Bioinformatically determined putative adhesins in the O157 DMEM-Proteome