Bacteria modulate the CD8+ T cell epitope repertoire of host cytosol-exposed proteins to manipulate the host immune response

Abstract

The main adaptive immune response to bacteria is mediated by B cells and CD4+ T-cells. However, some bacterial proteins reach the cytosol of host cells and are exposed to the host CD8+ T-cells response. Both gram-negative and gram-positive bacteria can translocate proteins to the cytosol through type III and IV secretion and ESX-1 systems, respectively. The translocated proteins are often essential for the bacterium survival. Once injected, these proteins can be degraded and presented on MHC-I molecules to CD8+ T-cells. The CD8+ T-cells, in turn, can induce cell death and destroy the bacteria's habitat. In viruses, escape mutations arise to avoid this detection. The accumulation of escape mutations in bacteria has never been systematically studied. We show for the first time that such mutations are systematically present in most bacteria tested. We combine multiple bioinformatic algorithms to compute CD8+ T-cell epitope libraries of bacteria with secretion systems that translocate proteins to the host cytosol. In all bacteria tested, proteins not translocated to the cytosol show no escape mutations in their CD8+ T-cell epitopes. However, proteins translocated to the cytosol show clear escape mutations and have low epitope densities for most tested HLA alleles. The low epitope densities suggest that bacteria, like viruses, are evolutionarily selected to ensure their survival in the presence of CD8+ T-cells. In contrast with most other translocated proteins examined, Pseudomonas aeruginosa's ExoU, which ultimately induces host cell death, was found to have high epitope density. This finding suggests a novel mechanism for the manipulation of CD8+ T-cells by pathogens. The ExoU effector may have evolved to maintain high epitope density enabling it to efficiently induce CD8+ T-cell mediated cell death. These results were tested using multiple epitope prediction algorithms, and were found to be consistent for most proteins tested.

Figure 1. Algorithm for epitope prediction and SIR score computation.

Each protein is divided into all nine-mers and the appropriate flanking positions (a). For each eleven-mer (a nine-mer and the C and N flanking positions), a cleavage score is computed (b). We compute for all peptides with a positive cleavage score a TAP binding score and choose only supra-threshold peptides (c). The MHC binding score of all TAP binding and cleaved nine-mers is computed (d). Nine-mers passing all these stages are defined as epitopes. We then compute the number of epitopes per protein per HLA allele (e). The ratio between the number of predicted epitopes and the parallel number on a random sequence with a random amino acid distribution is defined as the SIR score.

Figure 2. Upper row and lower left drawings - Histograms of SIR score values for real and scrambled sequences of bacterial proteins.

The x axis is the SIR scores and the y axis is the frequency of sequence with such an SIR score. Each drawing represents a different bacterium. The dark lines are the real sequences and the gray lines are the values obtained for scrambled versions of the same sequences. The distributions of the real and scrambled sequences overlap showing that bacteria, unlike viruses, do not generally accumulate escape mutations in their CTL-epitopes (P = 0.4306,0.1574 and 0.1942 for Shigella, E.coli and Pseudomonas, respectively). The lower left drawing is the SIR score histogram in human viruses (dark line) and the parallel in non human viruses (gray line). The average human virus SIR score is lower than the one of non-human viruses, revealing the accumulation of escape mutations in human viral proteins (P value<1.e-7).

Figure 3. SIR score of T3SS effectors and non-effectors proteins in Shigella flexneri, Pseudomonas aeruginosa and Escherichia coli.

A) Comparison between the average SIR score in real sequences and in scrambled sequences in Pseudomonas aureginosa. The first column is the average over 400 randomly selected proteins and the second column is the average of ExoT, ExoS and ExoY (non-necrotic effectors). The third column is the average of the necrotic effector ExoU. The first column has a similar average for the real and scrambled sequences (P value>0.18). In the second column, the real T3SS effectors sequences have a lower averaged SIR score than randomly selected proteins and also than expected from their scrambled sequences (P values = 2.4e-6 and 0.011, respectively). The third column demonstrates that ExoU has a higher SIR score than randomly selected proteins and than expected from its scrambled sequences (P-value<1e-4). B) Comparison between the average SIR score in real sequences and in scrambled sequences in Shigella flexneri. The first column is the average over 400 randomly selected proteins. The second column is the average over all T3SS effectors. The third and fourth columns are the averages over all early secreted and late secreted effectors, respectively. The first column has a similar average for the real and scrambled sequences (P value>0.3). Again, as in P. aureginosa, effectors have a lower SIR score average than randomly selected proteins, and also than expected from their scrambled sequences (second, third and fourth columns). Moreover, this bias is much stronger in early secreted effectors (*P-value<8e-4.). The differences in the second and fourth columns (overall and late effectors) are not significant (P-value>0.2). C) Comparison between the average SIR score in real sequences and in scrambled sequences in Escherichia coli. The left column is the average over 400 randomly selected proteins and the right column is the average over all T3SS effectors. The left column has a similar average for the real and scrambled sequences (P value>0.15). In the right columns, the real T3SS effectors sequences have a lower SIR score average than randomly selected proteins (P-value = 2.4e-3). Although E.coli effectors have a lower averaged SIR score than expected from their scrambled sequences, the difference is not significant (P-value>0.3). Bordered bars represent results that are not consistent with the other MHC-I binding algorithms, MLVO and NetMHC (figures 6 and 7, respectively).

Figure 4. Average scores of epitope for all steps of epitope presentation (proteasomal cleavage, TAP binding and MHC-I binding) in Shigella flexneri, Pseudomonas aeruginosa and Escherichia coli.

The y axis values represent the score in each step. P. aeruginosa effector, ExoU, being a unique necrosis effector was examined separately. In all bacteria tested, the average scores of epitopes in effectors in all stages are lower than the scores of epitopes in randomly selected proteins (with one exception of proteasomal cleavage of E.coli). ExoU epitopes had higher proteasomal cleavage and MHC-I binding scores in comparison to these scores of randomly selected proteins. Note that TAP is the less limiting factor in epitope processing. All differences (except proteasomal cleavage of E.coli with P-value = 0.388) are significant with 1.e-10<P-value<0.06.

Figure 5. SIR score of cytosolic proteins and randomly selected proteins in Listeria monocytogenes and Mycobacterium tuberculosis.

A) Comparison between the average SIR score in real sequences and in scrambled sequences in Listeria monocytogenes. The cytosolic proteins Listeriolysin O and ActA (second and third columns, respectively) have lower SIR scores than randomly selected proteins and than their scrambled sequences (P-value<1.e-12 for both proteins). In randomly selected proteins (first column), the differences between real and scrambled sequences was insignificant (P-value = 0.8652). B) Comparison between the average SIR score in real sequences and in scrambled sequences in Mycobacterium tuberculosis. The first two columns are the average SIR scores of EsxA and EsxB. Columns 3-20 are the scores of ESAT-6 homologues hypothetical proteins. Column 21 is the averaged SIR score over all proteins in the Esx family, and the last column is the averaged scores of 400 randomly selected proteins. In EsxA (ESAT-6), as well as in 15 out of 18 ESAT-6 homologues, the average SIR score in the real sequence is lower than the SIR score in randomly selected proteins and their scrambled sequences. In randomly selected proteins, the differences between real and scrambled sequences was insignificant (P-value = 0.2212). These results argue that the hypothetical ESAT-6 homologues - like ESAT-6 itself – might be localized in the host cytosol. *NS-not significant. All other differences are significant with P-value<0.05.

Figure 6. Validation of the results with MLVO algorithm.

In Pseudomonas, Shigella, Micobacterium and Listeria, the cytosolic proteins have lower SIR scores than expected, consistent with our previous results. However, the E.coli effectors had higher SIR scores than expected from their sequence and ExoU have shown a lower SIR score than expected, in contrast with the results using BIMAS. (P-value<0.08 for late Shigella proteins and P-value<0.04 for other proteins).

Figure 7. Validation of the results with NetMHC algorithm.

In Pseudomonas, Shigella, Micobacterium and Listeria, the cytosolic proteins have lower SIR scores than expected, consistent with our previous results (P value<0.02). However, E.coli effectors showed no significant differences between the real and neutral SIR scores and ExoU have lower SIR score than expected (P value = 0.012), in contrast with the results using BIMAS.

Figure 8. Fraction of peptides derived from proteasomal cleavage in Escherichia coli, Pseudomonas aeruginosa and Shigella flexneri.

T3SS effectors have less possible cleavage- derived nine-mers than other proteins. Moreover, in S. flexneri, among effectors, early secreted effectors have less cleavage-derived nine-mers than the late secreted ones. *P-value<1e-4 for P.aeruginosa and late effectors of S. flexneri, and P-value<1e-44 for early effectors of S. flexneri. In E.coli, the differences in fraction of cleavage derived peptides between effectors and other proteins was not significant (P-value = 0.1).