Identification of antigenic proteins of the nosocomial pathogen Klebsiella pneumoniae

Abstract

The continuous expansion of nosocomial infections around the globe has become a precarious situation. Key challenges include mounting dissemination of multiple resistances to antibiotics, the easy transmission and the growing mortality rates of hospital-acquired bacterial diseases. Thus, new ways to rapidly detect these infections are vital. Consequently, researchers around the globe pursue innovative approaches for point-of-care devices. In many cases the specific interaction of an antigen and a corresponding antibody is pivotal. However, the knowledge about suitable antigens is lacking. The aim of this study was to identify novel antigens as specific diagnostic markers. Additionally, these proteins might be aptly used for the generation of vaccines to improve current treatment options. Hence, a cDNA-based expression library was constructed and screened via microarrays to detect novel antigens of Klebsiella pneumoniae, a prominent agent of nosocomial infections well-known for its extensive antibiotics resistance, especially by extended-spectrum beta-lactamases (ESBL). After screening 1536 clones, 14 previously unknown immunogenic proteins were identified. Subsequently, each protein was expressed in full-length and its immunodominant character examined by ELISA and microarray analyses. Consequently, six proteins were selected for epitope mapping and three thereof possessed linear epitopes. After specificity analysis, homology survey and 3d structural modelling, one epitope sequence GAVVALSTTFA of KPN_00363, an ion channel protein, was identified harboring specificity for K. pneumoniae. The remaining epitopes showed ambiguous results regarding the specificity for K. pneumoniae. The approach adopted herein has been successfully utilized to discover novel antigens of Campylobacter jejuni and Salmonella enterica antigens before. Now, we have transferred this knowledge to the key nosocomial agent, K. pneumoniae. By identifying several novel antigens and their linear epitope sites, we have paved the way for crucial future research and applications including the design of point-of-care devices, vaccine development and serological screenings for a highly relevant nosocomial pathogen.

Figure 1. Epitope Mapping of KPN_00363.

The results of the epitope mapping summarized as a box-whisker plot (n = 10). The relative fluorescence intensities of each peptide from KPN_00363 are displayed. As controls, a positive (Rabbit IgG in red) and negative reference (MBP in blue) are added to the diagram. The boxes encompass 50% of the data, while the whiskers span 98% of the values. The median is portrayed as a horizontal line and the mean embodied by a small rectangle. Seven distinct regions display intensities exceeding 1000 A.U. These include peptides 2 and 3, 16–18, 29 and 30, 45–47, 50 and 51, 60–63, and 69–71. The highest mean value of these peptides is obtained by peptide 16 amounting to roughly 10000 A.U. In contrast, the negative control MBP reached less than 500 A.U., while the positive control peaked at a mean value of more than 30000 A.U.

Figure 2. Specificity binding analysis of epitope peptides of KPN_00363.

The bar chart represents the mean values [n = 6] of each potential linear epitope after incubation with polyclonal antibodies reactive to K. pneumoniae (green), C. jejuni (orange), S. aureus (purple), E. coli (grey), and S. enterica (blue). For peptides 2 and 3 the signal intensities after incubation with K. pneumoniae antibodies significantly surpassed those of the other antibody incubations. Contrastingly, all other peptides tested failed to show substantial discrimination between the different antibodies used. Mind the interruption of the ordinate for better scale.

Figure 3. Epitope Mapping of KPN_00459.

The box-whisker plot (n = 10) shows the relative fluorescence intensities of the different peptides from KPN_00459 and includes both a positive reference, Rabbit IgG in red, and a negative reference, MBP in blue. The boxes encompass 50% of the values, whereas the whiskers enclose 98% of the data. The median is represented by a horizontal line, while the mean is displayed as a small rectangle. Several regions may be identified with signal intensities significantly above the remaining peptides and the negative control. These include peptides 50, 51, 59, 60, 79, 80, and 81. The intensities of these peptides varied from 1200 A.U. for peptide 81 to roughly 6000 A.U. for peptide 59 which albeit significantly above 500 A.U. scored by the negative reference are far off from the 30000 A.U. achieved by the positive reference.

Figure 4. Specificity binding analysis of epitope peptides of KPN_00459.

After incubation with different antibodies reactive to to K. pneumoniae (green), C. jejuni (orange), and S. aureus (purple), peptides 59 (ALALGIAFGAVELFD) and 60 (GIAFGAVELFDVSFA) displayed mean signal intensities of 6000 and 3000 A.U. respectively for K. pneumoniae antibody incubation. Contrastingly, the mean signal intensities of these two peptides dropped to less than 500 A.U. after incubation with the other antibodies.

Figure 5. Epitope Mapping of KPN_00466.

Box-whisker plot (n = 10) displaying the resulting relative fluorescence intensities of the overlapping peptides. As controls, Rabbit IgG (red) and MBP (blue) are indicated. Each box represents 50% of the values, while the whiskers entail 98% of the data. The median is represented by a horizontal line and the mean displayed as a small rectangle. Peptides 11 and 12 are the only two adjacent peptides reaching intensities of 2000 A.U. within close proximity of the positive control at 4000 A.U indicative of a potential linear epitope. All the remaining peptides fell significantly short of these values except for peptide 16 which topped all other intensities by peaking at approximately 7000 A.U. However, as neither peptide 15 nor 17 showed any elevated intensities, the presence of a linear epitope at position 16 is rather unlikely.

Figure 6. Specificity binding analysis of epitope peptides of KPN_00466.

Bar chart representing the mean relative fluorescence intensities (n = 10) of each peptide potentially harboring a linear epitope site after incubation with polyclonal antibodies reactive to K. pneumoniae (green) and C. jejuni (orange). None of the peptides shows a peculiar specific interaction; rather signal intensities are in the same vicinity for each peptide independent of the antibody used. This indicates mainly non-specific binding to occur.

Figure 7. Partial 3d model of KPN_00363.

The model incorporates residues 31 to 294 and thus spans most of the protein except for the first 30 residues. Coils are displayed in green, helices in light blue and strands in purple. The N-terminal region of the model, i.e. starting with amino acid 31, is highlighted in orange. The model was based on the bacterial nucleoside transporter Tsx of E. coli. A major feature of the given model is the prominent beta barrel structure that originates from the abundance of beta strands. This is a typical feature of transport and channel proteins spanning the outer bacterial membrane. Contrastingly, the identified linear epitope GAVVALSTTFA is located at the very beginning of the protein and thus not included in the given model. However, it is likely an extension of the truncated N-terminal region marked in orange.

Figure 8. 3d model of predicted structure of KPN_00459.

The model was predicted using the automated mode of the SWISS MODEL application by Expasy (University Basel). As a template the crystal structure of a NA(+)/H(+) antiporter NhaA of E. coli was used. The resulting model spans residues 12 to 390 of the full-length protein and was subsequently dyed using the Chimera 1.7 software. Coils are depicted in light green, beta strands in purple and alpha helices in blue. The potential linear epitope GIAFGAVELFD is highlighted in orange. It comprises part of an alpha helix, a connective coil and the start of the next alpha helix.

Figure 9. Homology of linear epitope sequence GAVVALSTFFA of KPN_00363.

The sequence derived from K. pneumoniae MGH 78578 was used as a reference. Identical residues are marked by dots, gaps by a horizontal dash and differences by the single letter amino acid code. Seven of nine K. pneumoniae strains show identical epitope sequences, while two strains display changes in two residues. Threonine replaces alanine at position 11, a change observed not only in these two strains but in almost all other bacteria within the list. Additionally, threonine at position 9 is substituted by either serine or phenylalanine. Bacteria other than K. pneumoniae show an additional number of amino acid substitutions, most prominently leucine for valine at position 4 and serine for threonine at position 8. In S.enterica the changes become more pronounced. Glycine at position 1 is replaced by serine, valine at position 3 replaced by alanine and threonine inserted for serine at position 7. In some rare cases, other residues have also been substituted, e.g. valine replaces alanine at position 2 in Shigella dysenteriae.

Figure 10. Alanine scan of GAVVALSTTFA of KPN_00363.

Box-whisker plot (n = 12) of GAVVALSTTFA after alanine/glycine scanning. The box comprises 50% of the data, while the whiskers enclose 98%. The median is represented by a small horizontal line and the mean by a small rectangle. Rabbit IgG served as a positive reference, whereas MBP was used as a negative control. If alanine was present in the original present it was replaced by glycine, otherwise each amino acid was stepwise replaced by alanine. Additionally, GAVLALSSSFT and SAALALTSSFT were included as they resemble sequences present in E. coli and S. enterica. Switching glycine (position 1), alanine (positions 2 or 5), or threonine (positions 8 and 9) to alanine or glycine, results in a significant drop in signal intensities to levels below or at the negative control. In contrast, substituting valine (position 3) or leucine (position 6) by alanine, leads to an increase in signal intensities to 1700 A.U. and more than 5600 A.U., respectively. Note the different axis scales prior and after axis break at 2000 A.U.

Figure 11. Specificity assay of GAVVALSTTFA and derivatives.

The bar chart represents the mean signal intensities (n = 12) of GAVVALSTTFA and several modified peptides with single amino acid replacements incubated with antibodies reactive to K. pneumoniae (green), E. coli (orange) or S. enterica (purple). The sequences on the left represent the original epitope and modified versions displaying an increase in signal intensity for K. pneumoniae antibodies. In contrast, sequences on the right harbor modifications causing a significant drop in intensity for K. pneumoniae. Rabbit IgG is used as a positive reference and MBP serves as a negative control. None of the sequences tested displayed any significant signal intensity above the negative control when incubated with either E. coli or S. enterica antibodies.