Glycosylation of bacterial antigens changes epitope patterns

Abstract

Introduction: Unlike glycosylation of proteins expressed in mammalian systems, bacterial glycosylation is often neglected in the development of recombinant vaccines.

Methods: Here, we compared the effects of glycosylation of YghJ, an Escherichia coli protein important for mucus attachment of bacteria causing in urinary tract infections (UTIs). A novel method based on statistical evaluation of phage display for the identification and comparison of epitopes and mimotopes of anti-YghJ antibodies in the sera was used. This is the first time that the effect of glycosylation of a recombinant bacterial antigen has been studied at the peptide epitope level.

Results: The study identifies differences in the immune response for (non)-glycosylated antigens in rabbits and pigs and compares them to a large group of patients with UTI, which have been diagnosed as positive for various bacterial pathogens. We identified glycosylation-specific peptide epitopes, a large immunological similarity between different UTI pathogens, and a broad peptide epitope pattern in patients and animals, which could result in a variable response in patients upon vaccination.

Discussion: This epitope analysis indicates that the vaccination of rabbits and pigs raises antibodies that translate well into the human immune system. This study underlines the importance of glycosylation in bacterial vaccines and provides detailed immune diagnostic methods to understand individual immune responses to vaccines.

Keywords: ACFD; Escherichia coli; SSLE; YghJ; epitope; protein glycosylation; uropathogenic; vaccine development.

Figure 1

Sequences sharing motifs of the epitope motif 1178-PKFYKDGEWK enriched by binding to different patient sera’s antibodies. Patients (PS) 10, 13, 14, and 15 seem to form a group different from 11 and 12. Displayed are only those sequences found at least four times and sharing at least five amino acid identities in the alignment selected from a total of 2,815 sequences from 12 datasets with an average frequency of 2; the naïve library has an average of ca. 1.1 for all sequences in a comparable dataset.

Figure 2

Differences in the enrichment pattern of epitope motifs between gYghJ and ngYghJ comparing sera before (t0) and after (t1) immunization. (A) Prior to immunization, the sera does not contain YghJ-specific antibodies. (B) Rabbits K12, K13, and K14 (red lines) have been immunized with the glycosylated antigen and strongly enrich sequences containing proline-rich motifs. This enrichment is not found in the Rabbits K15 and K16 immunized with non-glycosylated antigen (blue line). Rabbit K17 was also immunized with non-glycosylated antigen (hatched line), but probably an additional contact to another (glycosylated) antigen led to the different enrichment pattern.

Figure 3

Results from screening sera from UTI patients with confirmed infections, pigs either vaccinated or infected directly in the bladder, and immunized rabbits that have been tested in an epitope peptide array with triplicates. The signal cutoffs are signal over background (empty spots). Color codes are signal over background; >100× (dark green), >20× (light green), >4× (yellow), >2× (orange), 2× and less (gray); no data due to too high error, printing errors in the array or not at least values from 2 spots. Blue boxes: Animal sera taken prior to immunization.

Figure 4

Absolute values measured in peptide microarrays with rabbit sera prior to immunization (D0) and after immunization with the glycosylated and non-glycosylated antigen. The color code is adjusted for each peptide individually from maximal (green) to lowest (red) value in direct comparison. The lower lines compare the signals for each peptide giving the average and the ratios between sera pre- and post-immunization with the different antigens. Only for the rabbits labeled “K” was a pre-immune serum taken; hence, the different ratios were calculated at the end of the table.

Figure 5

Summarizing the results of this study: The epitopes are shown on the AF-model for a related YghJ (E3PJ90) (–37). Epitopes are shown in red, with the exception of the epitopes positively (dark blue) or negatively (light blue) influenced by glycosylation. The N-terminal proline-rich loop (compare Figure 3 ) protrudes as an undefined orange tube structure towards the reader.