Functional plasticity in the type IV secretion system of Helicobacter pylori

Abstract

Helicobacter pylori causes clinical disease primarily in those individuals infected with a strain that carries the cytotoxin associated gene pathogenicity island (cagPAI). The cagPAI encodes a type IV secretion system (T4SS) that injects the CagA oncoprotein into epithelial cells and is required for induction of the pro-inflammatory cytokine, interleukin-8 (IL-8). CagY is an essential component of the H. pylori T4SS that has an unusual sequence structure, in which an extraordinary number of direct DNA repeats is predicted to cause rearrangements that invariably yield in-frame insertions or deletions. Here we demonstrate in murine and non-human primate models that immune-driven host selection of rearrangements in CagY is sufficient to cause gain or loss of function in the H. pylori T4SS. We propose that CagY functions as a sort of molecular switch or perhaps a rheostat that alters the function of the T4SS and "tunes" the host inflammatory response so as to maximize persistent infection.

Figure 1. Loss of the capacity to induce IL-8 in H. pylori recovered from rhesus monkeys is associated with changes in the gene encoding CagY, an essential protein in the T4SS.

(A–E) H. pylori was isolated from five rhesus macaques up to 14 months after experimental infection with H. pylori WT J166. Individual colonies were co-cultured with AGS cells, and ELISA was used to measure IL-8 levels, which were normalized to the WT J166 positive control. Each data point represents the results from a single colony. The capacity to induce IL-8 decreased over time in colonies recovered from four monkeys (A–D), but was largely unchanged in one (E). PCR-RFLP analysis showed that H. pylori colonies that lost the capacity to induce IL-8 were associated with a change in cagY (open circles), while those that maintained IL-8 induction typically had cagY that was indistinguishable from WT J166 (filled circles). Animal designation is shown in the upper left corner of each panel. (F) Output strains from each monkey were analyzed by cagY PCR-RFLP and compared to WT H. pylori J166 (dark blue) and to one another. Each pie chart represents all colonies recovered from one of the five monkeys (12–24 colonies/monkey); different colors represent different cagY variants.

Figure 2. Recombination in cagY during infection of rhesus monkeys is sufficient to reduce the capacity of H. pylori to induce IL-8 and translocate CagA.

Deletion of cagY (▵Y) from WT H. pylori J166 significantly reduced its capacity to induce IL-8 (mean ± SEM of 3 replicates), which was recovered when the chromosomal WT cagY allele was restored (▵Y [J166]) by complementation (black bars). Immunoblot showed that only the WT or ▵Y [J166] expressed CagY protein (α-CagY) and translocated CagA that was tyrosine phosphorylated (α-PY99). Two rhesus output strains with unique cagY alleles (rOut1, rOut2) lost the capacity to induce IL-8 (gray bars) and translocate CagA, although they expressed CagY. Replacement of ▵cagY with cagY from rOut1 (▵Y [rOut1]) or rOut2 (▵Y [rOut2]) recapitulated their failure to induce IL-8 induction (white bars) and translocate phosphorylated CagA. Similarly, complementation with cagY from an output strain (rOut3) that expressed a unique cagY but maintained the capacity to induce IL-8 (gray bar) and translocate CagA, also phenocopied its IL-8 induction and translocation of CagA. All strains expressed CagA (α-CagA), though only those that induced IL-8 had the capacity to translocate CagA that was tyrosine phosphorylated. Multiple bands in the CagY immunoblot could represent different transcription or translation products, or even protein fragments, but they are CagY-specific since they are absent in the cagY deletion mutant. **P<0.01; ***P<0.001.

Figure 3. Loss of the capacity to induce IL-8 and change cagY during infection of mice is dependent on an intact host immune system.

H. pylori was isolated from C57BL/6 WT (A) or RAG2−/− (B) mice (N = 3–6/time point) up to 16 weeks after experimental infection with H. pylori WT J166. Individual colonies (3–6/mouse) were co-cultured with AGS cells, and ELISA was used to measure IL-8 levels, which were normalized to the WT J166 positive control (line = mean). Each data point represents the results from a single colony. Induction of IL-8 in colonies isolated from WT mice was significantly lower than in RAG2−/− mice at 12 and 16 weeks PI (P<0.01). Changes in cagY (open circles) were detected by PCR-RFLP in 28 of 70 colonies from WT mice but in 0 of 64 colonies from RAG2−/− mice (Fishers exact test, P<0.0001). Output strains from WT C57BL/6 mice were analyzed by cagY PCR-RFLP and compared to WT H. pylori J166 (dark blue) and to one another (C). Each pie chart represents the unique cagY RFLP patterns identified in a single mouse from 2 to 16 weeks PI, and is positioned according to the mean IL-8 induction by colonies recovered from that mouse.

Figure 4. Recombination in cagY during infection of mice is sufficient to reduce the capacity of H. pylori to induce IL-8 and translocate CagA.

Deletion of cagY (▵Y) from WT H. pylori J166 significantly reduced its capacity to induce IL-8 (mean ± SEM of 3 replicates), which was recovered when the chromosomal WT cagY allele was restored (▵Y [J166]) by complementation (black bars). Two output strains from C57BL/6 mice with unique cagY alleles (mOut1, mOut2) lost the capacity to induce IL-8 (gray bars) and translocate CagA, although they expressed CagY (α-CagY). Complementation of ▵cagY with cagY from mOut1 (▵Y [mOut1]) or mOut2 (▵Y [mOut2]) recapitulated their lack of IL-8 induction (white bars) and translocation of phosphorylated CagA (α-PY99). Similarly, replacement with cagY from two output strains (mOut3, mOut4) that expressed a unique cagY but maintained the capacity to induce IL-8 (gray bars) and translocate CagA, also phenocopied their IL-8 induction and translocation of CagA. All strains expressed CagA (α-CagA), though only those that induced IL-8 had the capacity to translocate CagA that was tyrosine phosphorylated. Multiple bands in the CagY immunoblot could represent different transcription or translation products, or even protein fragments, but they are CagY-specific since they are absent in the cagY deletion mutant. **P<0.01; ***P<0.001.

Figure 5. cagY variants that fail to induce IL-8 and translocate CagA do not induce expression of NF-κB.

(A) Co-culture of H. pylori with AGS cells stably transformed with a reporter plasmid demonstrated that activation of NF-κB was seen in WT J166 but not in a strain with a deletion of the cagPAI (▵PAI). Reintroduction of J166 cagY into a cagY deletion mutant restored NF-κB activation. Introduction of cagY from monkey (B) or mouse (C) output strains showed that increased NF-κB activation compared to ▵cagY (▵Y) or ▵PAI was seen only in strains bearing a cagY allele that was competent for induction of IL-8 and translocation of CagA (rOut3, mOut3, mOut4). ***P<0.001.

Figure 6. H. pylori colonization of rhesus monkeys and mice is associated with changes in the motif structure of the CagY middle repeat region.

As in strains J99 and 26695 , the predicted amino acid sequence of CagY in WT H. pylori J166 is organized into a 5′ repeat region (residues 9–398), a 3′ region orthologous to VirB10 (residues 1784–2028), and a middle repeat region (residues 715–1512) that is composed of a series of B motifs (yellow) that bracket one to four A motifs (orange). Passage of WT J166 in rhesus monkeys and in mice results in some strains that lose one or more A or B motifs, which is sometimes sufficient to reduce the capacity to induce IL-8 (rOut1 and rOut2; mOut1 and mOut2) and other times is not (rOut3; mOut3 and mOut4).

Figure 7. Changes in the motif structure of the CagY middle repeat region that alter the function of the cagPAI do not affect expression of T4SS pili on the bacterial surface.

H. pylori was co-cultured with AGS gastric cells at an MOI of 100∶1 and imaged by FEG-SEM. T4SS pilus structures were readily apparent in the WT H. pylori J166 but not in the cagPAI deletion mutant (J166▵cagPAI). T4SS pili were also observed in H. pylori J166 in which the WT cagY allele was replaced with that from output strains with a functional (rOut3, mOut3) or a non-functional (rOut2 mOut2) cagPAI. Pili were also seen in H. pylori strains J166 and 26695 with deletions in cagY. Magnification bars indicate 500 nm.

Figure 8. CagY decorates the H. pylori bacterial surface but is not associated with T4SS pili.

H. pylori was co-cultured with AGS gastric cells at an MOI of 100∶1, incubated with antibodies to the CagY MRR or CagA, and imaged by FEG-SEM in the environmental mode. CagY was detected on the bacterial surface of the WT strain but was not associated with pili. CagA was detected both on the bacterial surface and in close approximation to the tips of the pili of the WT strain. There was markedly reduced CagA labeling on the surface of ▵cagY mutant strain compared to the WT strain. No staining was seen when primary antibody was omitted. Pili are sometimes not as well visualized and more often appear broken in these images compared to Figure 7 due to the lack of metal coating and more frequent washes. Magnification bars indicate 500 nm.

Figure 9. Mouse adapted H. pylori strain SS1 expresses a CagY that is not functional for induction of IL-8 or translocation of CagA.

(A) H. pylori was isolated from C57BL/6 WT or RAG1−/− mice (N = 3–6/time point) 8 weeks after experimental infection with H. pylori PMSS1. Individual colonies (3–6/mouse) were co-cultured with AGS cells, and ELISA was used to measure IL-8 levels, which were normalized to the PMSS1 positive control (line = mean). Each data point represents the results from a single colony. Induction of IL-8 in colonies isolated from WT mice was significantly lower than in RAG1−/− mice, and was associated with changes in cagY PCR-RFLP (open circles). (B) cagY in H. pylori strain SS1 is larger than that in the progenitor strain PMSS1, and has a different fingerprint on PCR-RFLP. (C) Deletion of cagY from WT H. pylori PMSS1 reduced the induction of IL-8 and eliminated translocation of CagA, which were recovered when the WT PMSS1 cagY gene was restored (▵Y[PMSS1]. However, replacement of the PMSS1 cagY gene with that from H. pylori SS1 (▵Y [SS1]) showed reduced levels of IL-8 and no CagA translocation. (D) WT H. pylori SS1 showed little induction of IL-8 and no CagA translocation, and it was unaffected by deletion of cagY or restoration of the WT SS1 cagY allele. However, replacement of the WT SS1 cagY allele with that from PMSS1 markedly increased IL-8 induction and CagA translocation, though not to the level of PMSS1. All assays represent the mean ±SEM of 3 replicates. **P<0.01; ***P<0.001.

Figure 10. Recombination of cagY during infection of rhesus macaques and mice can also restore the capacity to induce IL-8.

Rhesus macaques and mice were inoculated with mOut2, which does not induce IL-8 or translocate CagA. Single colony isolates were recovered and tested for induction of IL-8 and compared to mOut2 by cagY PCR-RFLP. (A) Colonies from three monkeys (36001, 35951, 35930) showed significantly increased capacity to induce IL-8 at 8 weeks compared to 2 weeks PI, which was associated with changes in cagY RFLP. The fourth monkey (36018) was colonized with a mixture of cagY genotypes that induced low IL-8 similar to mOut2. (B) Colonies recovered from WT and RAG2−/− mice typically induced low IL-8 similar to input mOut2, with no change in cagY. *P<0.05; ***P<0.001.