Surface display of Salmonella epitopes in Escherichia coli and Staphylococcus carnosus
© Nhan et al; licensee BioMed Central Ltd. 2011
Received: 20 December 2010
Accepted: 11 April 2011
Published: 11 April 2011
Skip to main content
© Nhan et al; licensee BioMed Central Ltd. 2011
Received: 20 December 2010
Accepted: 11 April 2011
Published: 11 April 2011
Salmonella enterica serotype Enteritidis (SE) is considered to be one of the most potent pathogenic Salmonella serotypes causing food-borne disease in humans. Since a live bacterial vaccine based on surface display of antigens has many advantages over traditional vaccines, we have studied the surface display of the SE antigenic proteins, H:gm and SefA in Escherichia coli by the β-autotransporter system, AIDA. This procedure was compared to protein translocation in Staphylococcus carnosus, using a staphylococci hybrid vector earlier developed for surface display of other vaccine epitopes.
Both SefA and H:gm were translocated to the outer membrane in Escherichia coli. SefA was expressed to full length but H:gm was shorter than expected, probably due to a proteolytic cleavage of the N-terminal during passage either through the periplasm or over the membrane. FACS analysis confirmed that SefA was facing the extracellular environment, but this could not be conclusively established for H:gm since the N-terminal detection tag (His6) was cleaved off. Polyclonal salmonella antibodies confirmed the sustained antibody-antigen binding towards both proteins. The surface expression data from Staphylococcus carnosus suggested that the H:gm and SefA proteins were transported to the cell wall since the detection marker was displayed by FACS analysis.
Apart from the accumulated knowledge and the existence of a wealth of equipment and techniques, the results indicate the selection of E. coli for further studies for surface expression of salmonella antigens. Surface expression of the full length protein facing the cell environment was positively proven by standard analysis, and the FACS signal comparison to expression in Staphylococcus carnosus shows that the distribution of the surface protein on each cell was comparatively very narrow in E. coli, the E. coli outer membrane molecules can serve as an adjuvant for the surface antigenic proteins and multimeric forms of the SefA protein were detected which would probably be positive for the realisation of a strong antigenic property. The detection of specific and similar proteolytic cleavage patterns for both the proteins provides a further starting point for the investigation and development of the Escherichia coli AIDA autotransporter efficiency.
Bacterial display of proteins on the cell surface using recombinant DNA technology has been used in microbiology, biotechnology and vaccine technology and has been an area of intense research since reports of this novel technology were published [1–5]. One of the most studied areas of application has been live vaccine delivery through the surface display of antigenic proteins [2, 4]. This technique gives potential advantages over traditional vaccines since they are less costly to produce (there is no need for extensive isolation and purification) , they are better recognised by the host immune system and therefore create better immune responses , they elicit strong long-lasting immunity , components of the host, e.g. outer membrane lipopolysaccharides (specifically in E. coli), may contribute to a very strong immune response acting as an adjuvant to the recombinant antigenic proteins , and finally the surface expression may also be safer than attenuated or inactivated vaccines because bacteria strains used for surface expression must be non-pathogenic.
Several surface display systems for both gram negative and gram positive bacteria have been described in the literature but the detailed knowledge of these systems varies considerably . Many naturally occurring proteins have been developed as carriers for a target protein, such as outer membrane proteins, lipoproteins, secreted proteins and subunits of surface appendages. These systems show some disadvantages, for example with respect to the size of foreign proteins or high sensitivity to the structure of inserts [6, 7]. E. coli has attracted a lot of interest due to its easy handling and the wealth of literature and knowledge accumulated. The successful transplantation of the pathogenic E. coli β-autotransporter (type V secretors of Enteropathogenic bacteria) to laboratory strains has provided a new technique for vaccine production through exposure of antigens on the surface [6, 8, 9]. This transport vehicle is the most abundantly expressed protein transporter for surface display in these cells and the "AIDA" autotransporter (Adhesin Involved in Diffuse Adherence)  was chosen for the present work. The original vector contains all the parts necessary for translocation to the environment and consists of its own unique elements: an AIDA-specific signal sequence, a passenger protein, a linker region known to influence translocation, and a translocation unit (AIDAC) intended for insertion as a system-specific pore in the outer membrane (C-terminal). A suggested transport mechanism for the AIDA autotransporter was described and the fate of the passenger is to stay anchored or to be cleaved-off . It is generally believed that the passenger protein must be kept in an unfolded state to achieve proper translocation across the outer membrane .
Staphylococcus carnosus is a Gram-positive bacterium frequently used for protein display and it is considered to be a food-grade bacterium due to its low DNA homology with the pathogenic strain, Staphylococcus aureus. The S. carnosus surface display vector uses the M and X domains of protein A from Staphylococcus aureus (SpA)  where the latter contains a charged repetitive domain responsible for binding to cell-wall peptidoglycan.
In this work we wished to explore the possibilities of expression of salmonella proteins using the E. coli AIDA autotransporter, in order to understand its potential for surface exposure and to compare this to surface expression in the Gram positive bacterium Staphylococcus carnosus, earlier reported to be used in live-vaccine applications . We believed this to be particularly interesting since E. coli is structurally similar to salmonella and the production strain is a K12 strain, which lacks the O-antigen. We selected the H:gm flagellar protein, recognized and targeted by both the innate and adaptive immune systems [12, 13], and the SefA fimbrial protein, that is restricted to SE and other closely related group D Salmonella, due to their central role in salmonella infections . We show that both proteins can be translocated by both E. coli and S. carnosus but with differing efficiencies and characteristics.
Primers used for amplification of gm and sefA
AAA CAC GTG GGG CAC AAG TCA TTA ATA CAA ACA G
AAA TCT AGA GCA CGC AGT AAA GAG AGG ACG
AAA CAC GTG GGG CTG GCT TTG TTG GTA ACA AAG
AGG TCT AGA GCG TTT TGA TAC TGC TGA ACG TAG
GGG CTC GAG GCA CAA GTC ATT AAT ACA AAC AG
AAA GTC GAC ACG CAG TAA AGA GAG GAC G
ATA CTC GAG GCT GGC TTT GTT GGT AAC AAA G
ATA GTC GAC GTT TTG ATA CTG CTG AAC GTA G
A 2.5 μl volume of E. coli cells was taken from the -80°C storage and inoculated into 25 ml LB medium (Scharlau Chemie S. A.) with 100 μg/ml ampicillin. The cells were cultivated overnight in 250 ml baffled shake flasks at 37°C and stirred at 180 rpm. Recombinant cells of S. carnosus with either of the two proteins were taken from glycerol stocks at -80°C and used to inoculate 25 ml of LB medium with chloramphenicol to 10 μg/ml. The culture was grown overnight at 37°C and 150 rpm in a 250 ml baffled shake flask. OD600 was used to monitor the growth (Novaspec II).
Outer membrane proteins from E. coli strains were extracted according to the literature . The outer membrane proteins were stored at -20°C before analysis.
Samples mixed with reducing buffer (0.0625 M Tris-HCl, pH 6.8, 20 g/l sodium dodecyl sulfate (SDS), 43% glycerol, 10% (v/v) 2-mercaptoethanol, 10 mg/ml bromophenol blue) were heated at 95°C for 10 minutes and analyzed on NuPAGE® 10% Bis Tris Gels (Invitrogen) in MOPS buffer . Gels obtained after SDS-PAGE were stained with Coomassie Brilliant Blue solution (7% acetic acid, 50% methanol and 2.5 g/l Coomasie Brilliant Blue R-250). SeeBlue® Plus2 Prestained Standard (Invitrogen) was used to assess the molecular weights of the protein.
Western blot detection was based on antibodies to the AIDAC domain fused to either of the target proteins. The SDS-PAGE gel, the nitrocellulose membrane and filter papers were first equilibrated in transblotting buffer (2.93 g l-1 glycine, 5.81 g l-1 Tris-HCl, 0.37 g l-1 SDS, 20% methanol), blotted at 15 V for 45 min (Trans-blot® SD Semidry transfer cell machine, Bio-Rad) and blocked in phosphate buffered saline, PBS (9 g l-1 NaCl, 0.21 g l-1 KH2PO4, 0.726 g l-1 Na2HPO4, pH 7.4) and 5% low fat milk for about one hour. The membrane was washed three times with PBS buffer (10 min each) before being incubated at room temperature for one hour with rabbit serum IgG specific for the AIDAC, diluted 1:50000 in PBS with 0.1% bovine serum albumin (BSA). After being rinsed three times with PBS to remove non-specific binding (10 min each), the membrane was incubated for one hour with alkaline phosphatase-linked goat anti-rabbit IgG secondary antibody (Sigma) diluted 1:10 000 in PBS. The membrane was then again washed three times with PBS. The substrate (FAST™ BCIP/NBT, 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium, Sigma) was dissolved in 10 ml of deionized H2O and poured onto the membrane to detect the antigen-antibody complex through alkaline phosphatase cleavage of the substrate, resulting in a dark brown staining of the membrane.
The proteins derived from E. coli were fused to an N-terminal His6 tag which can be used for detection by commercial antibodies. The procedure was the same as that described for the detection of AIDAC but with the use of antibodies and detection by the HisProbe-HRP and SuperSignal system (Pierce). Blots were scanned by a Fujifilm Image Reader 1000 V1.2 (Fujifilm Life Science) using the software: Image gauge 4.0.
The procedure for flow cytometry analysis has been described before . The proteins were visualised by fluorescence markers either to the E. coli His6 tag or to the albumin-binding domain (ABD) in S. carnosus. A sample of 10 μl of an overnight culture was incubated with biotinylated 6× His tag-specific rabbit polyclonal antibody (Abcam) and a streptavidin-AlexaFluor488 conjugate (Invitrogen). The labelled cells were resuspended in ice-cold PBS and protected from light until FACS analysis was performed (FACS Vantage SE, BD Biosciences). The procedure was repeated in a similar manner for S. carnosus except that the human serum albumin binding to the protein was evaluated with HSA-Alexa647 (HSA conjugate solution, Invitrogen).
Although OM expression was confirmed, the H:gm and SefA proteins might be subjected to two possible steric orientations: directed either inwards towards the periplasm or outwards towards the extracellular environment. The latter is the preferred, although the generation of efficient antibodies to evoke an immunological response cannot be conclusively established without further specifically designed experiments.
To compare the E. coli expression levels with a more common expression system used for vaccine production, S. carnosus surface expression was used for the same two salmonella proteins, H:gm and SefA. S. carnosus has earlier attracted considerable interest due to the possibility of using a one-step simple translocation, and the solidity of the cell wall allows robust production and handling. The thick cell wall of Gram-positive bacteria has however been a potential drawback leading to low frequency of transformation. Recently, however, electroporation-mediated transformation in S. carnosus has been optimised, and this potentially circumvents this problem. The anchoring to the cell wall is here provided by the X and M domains from Staphylococcus aureus protein A (SpA), where X is a charged repetitive domain binding to the cell wall and M consists of an LPXTG motif followed by hydrophobic amino acids and a short charged tail.
Two genes representing proteins of one salmonella fimbrium and one flagellum were inserted to full length and within the correct open reading frame within both E. coli and S. carnosus vectors. Proteins were expressed and readily transferred to the outer membrane and cell wall, respectively. One of the proteins, SefA, was conclusively shown to face the external environment to its full length in E. coli, but in the case of the other protein, H:gm, a degradation form was produced and the localisation could not be fully ascertained. Polyclonal antibodies to Salmonella enteritides bound to both protein antigens and, despite the degradation and unproven localisation of H:gm, the experiment showed that they were recognised by the antibody. The possibility that the E. coli expressed proteins to generate antibodies from a host immune response has not however been shown for either system, and it will have to be studied in further experiments. The further binding of the salmonella antibodies to several proteins in E. coli could imply that this could serve as an adjuvant for recombinant antigenic proteins and this would thus promote the use of this host cell. The expression in E. coli also seems to be the most promising due to the very narrow distribution of protein on each cell (SefA). In addition, the probable existence of multimeric forms of the SefA protein in E. coli could promote the development of a strong and potent antigenic property. Further, we believe that to overcome the partial proteolytic degradation in E. coli is a more straightforward research and development route than the formulation of a hypothesis as to why S. carnosus expression is so very broad. The similarity between the proteolytic patterns shown from both SefA and H:gm protein expression in E. coli will form an interesting basis for further research and optimisation of the system.
We would like to acknowledge Prof's Inga Benz and Alexander Schmidt for the kind supply of their plasmids for AIDA expression and the AIDAC antibodies. Polyclonal antibodies to Salmonella enteritides were kindly provided by Dr. Do Thi Huyen (IBT, Hanoi, Vietnam). Prof Stefan Ståhl is gratefully acknowledged for the use of the S. carnosus strain and vector. Dr John Löfblom is greatly acknowledged for his support with the FACS analysis. The Swedish International Development Agency (SIDA) and the Swedish Agency for Innovation Systems (Vinnova) are acknowledged for their financial support.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.