Primary Structural Variation in Anaplasma marginale Msp2 Efficiently Generates Immune Escape Variants Telmo Graça,a,b Lydia Paradiso,a,c* Shira L. Broschat,a,b,c Susan M. Noh,a,b,d Guy H. Palmera,b,c The Paul G. Allen School for Global Animal Health,a Department of Veterinary Microbiology and Pathology,b and School of Electrical Engineering and Computer Science,c Washington State University, and Animal Disease Research Unit, Agricultural Research Service, U.S. Department of Agriculture,d Pullman, Washington, USA
Antigenic variation allows microbial pathogens to evade immune clearance and establish persistent infection. Anaplasma marginale utilizes gene conversion of a repertoire of silent msp2 alleles into a single active expression site to encode unique Msp2 variants. As the genomic complement of msp2 alleles alone is insufficient to generate the number of variants required for persistence, A. marginale uses segmental gene conversion, in which oligonucleotide segments from multiple alleles are recombined into the expression site to generate a novel msp2 mosaic not represented elsewhere in the genome. Whether these segmental changes are sufficient to evade a broad antibody response is unknown. We addressed this question by identifying Msp2 variants that differed in primary structure within the immunogenic hypervariable region microdomains and tested whether they represented true antigenic variants. The minimal primary structural difference between variants was a single amino acid resulting from a codon insertion, and overall, the amino acid identity among paired microdomains ranged from 18 to 92%. Collectively, 89% of the expressed structural variants were also antigenic variants across all biological replicates, independent of a specific host major histocompatibility complex haplotype. Biological relevance is supported by the following: (i) all structural variants were expressed during infection of a natural host, (ii) the structural variation observed in the microdomains corresponded to the mean length of variants generated by segmental gene conversion, and (iii) antigenic variants were identified using a broad antibody response that developed during infection of a natural host. The findings demonstrate that segmental gene conversion efficiently generates Msp2 antigenic variants.
M
icrobial pathogens utilize diverse mechanisms to persist in their mammalian hosts, evolutionarily driven to increase the likelihood of onward transmission and propagation of self. Antigenic variation is a convergent strategy used by pathogens ranging from RNA viruses to protozoa (1–4). Among bacteria and protozoa, recombinatorial mechanisms are commonly used to generate sequential antigenic variants from otherwise relatively stable genomes (2–5). For example, bacterial pathogens in the genera Anaplasma, Borrelia, Mycoplasma, Neisseria, and Treponema utilize gene conversion of a repertoire of silent alleles into expression sites to encode unique antigenic and structural variants (2–4). Gene conversion is unidirectional, leaving the donor allele unchanged but replacing the expression site sequence (6–8). Thus, the donor alleles are under long-term selective pressure to provide a template for expressing a variant that is sufficiently antigenically unique but also retains a viable structure (Fig. 1A). Pathogens in these genera also utilize segmental gene conversion, whereby an oligonucleotide segment, representing only a portion of the donor allele, is recombined into the expression site, resulting in a mosaic not present elsewhere in the genome (Fig. 1B and C) (6, 9). This process of segmental gene conversion has the potential to tremendously amplify the capacity to generate antigenic variants but can be limited by structural constraints, as these assembled mosaics, unlike the whole donor alleles, have not been under long-term selection for structural fitness. Using Anaplasma marginale infection in its natural bovine host, Futse et al. demonstrated that a unique Major Surface Protein 2 (Msp2) variant derived from a whole msp2 donor allele was sufficient to evade an existing broad antibody response generated against a repertoire of Msp2 variants (10). The unique donor allele encoded a hypervariable region (HVR) that shared 63% identity with HVRs to which the host had been previously exposed. This
4178
iai.asm.org
level of difference in encoded HVRs is reflected in the allelic repertoire of the multiple A. marginale strains examined, consistent with their deterministic role in antigenic variation (10, 11). In contrast, the degree of sequence difference that allows a variant generated by segmental gene conversion to escape immune recognition is unknown. If large differences are required to generate a variant capable of immune escape, the size of the potential repertoire would be reduced and segmental gene conversion would only marginally expand the repertoire compared to that generated by whole allelic conversion. To address this knowledge gap, we identified A. marginale St. Maries strain Msp2 variants, represented by their HVR microdomains that were recognized by IgG antibodies induced during infection of cattle, a natural host. A second set of expressed structural variants that had defined differences in the corresponding microdomains (12) but were not represented in the St. Maries strain and to which the animals had thus not been exposed were used to test whether these differences were
Received 30 June 2015 Returned for modification 14 July 2015 Accepted 4 August 2015 Accepted manuscript posted online 10 August 2015 Citation Graça T, Paradiso L, Broschat SL, Noh SM, Palmer GH. 2015. Primary structural variation in Anaplasma marginale Msp2 efficiently generates immune escape variants. Infect Immun 83:4178 –4184. doi:10.1128/IAI.00851-15. Editor: C. R. Roy Address correspondence to Telmo Graça, [email protected]. * Present address: Lydia Paradiso, School of Biology, University of Edinburgh, Edinburgh, United Kingdom. Supplemental material for this article may be found at http://dx.doi.org/10.1128 /IAI.00851-15. Copyright © 2015, American Society for Microbiology. All Rights Reserved.
Infection and Immunity
November 2015 Volume 83 Number 11
A. marginale Msp2 Segmental Variants
FIG 1 A. marginale Msp2 mechanism of antigenic variation. The mosaics generated through genomic recombination result in expression of new antigenically distinct Msp2 variants, which compose sequential bacteremic waves and allow persistence within the host. (A) Chromosomal msp2 loci, 7 silent alleles, and a single active expression site (ES) with a central hypervariable domain (HVR). P1, E6/F7, G11, 1, 2, 9H1, and 3H1 are the 7 conserved donor alleles for the Msp2 HVR, with 5 of them being unique (allele P1 is a duplication of allele G11, and allele 2 is a duplication of allele 3H1). (B) Segmental gene conversion from an oligonucleotide segment of a donor allele into the expression site, generating an expression site mosaic. (C) Bacteremic peaks characterized by expression of unique Msp2 variants, variants generated by recombination of a “whole” donor allele, predominate in acute bacteremia, followed by a persistent phase in which the mosaics generated by segmental gene conversion predominate. The modeled protein structure with the peptide sequences generated by a donor allele or segmental gene conversion is indicated above the bacteremic peak.
sufficient to evade immune recognition. This strategy incorporated three key elements: (i) the Msp2 variants were expressed during in vivo infection and thus represented at least minimal structural fitness; (ii) the microdomains tested represent the encoded variable regions; and (iii) the antibody response, induced by infection in the natural host, represented the relevant breadth of epitope specificity. Infection of a natural host is critically important to reflect the long-term evolutionary pressures associated
November 2015 Volume 83 Number 11
with mechanisms of pathogen persistence. Here, we present the results and discuss the findings in the context of the antigenic variant repertoire and the consequences for evasion of immunity at the levels of both the individual host and the host population. MATERIALS AND METHODS Variant selection. The immunogenic Msp2 HVR microdomains were matched between the complete repertoire of microdomains encoded by
Infection and Immunity
iai.asm.org
4179
Graça et al.
FIG 2 Graphical representation of the A. marginale St. Maries HVR microdomains and corresponding flanking tethers. T1, N-terminal flanking tether; T2, spacing tether between microdomain 1 and microdomain 2; T3, spacing tether between microdomain 2 and microdomain 3.
the St. Maries strain (GenBank accession no. CP000030) and those expressed by genetically distinct Nayarit strains (GenBank accession no. JQ933614 to JQ933796) (5, 12). Msp2 HVR microdomains 1 and 2 encompass the immunodominant B cell epitopes in the HVRs encoded by the allelic repertoire in the St. Maries strain (13, 14) and are defined by flanking conserved amino acids, previously designated “structural tethers” (Fig. 2). Conservation of the tethers was verified by analysis of ⬎1,600 expressed variants of the St. Maries strain and Nayarit strains (12). Microdomain 1 has a 6-amino-acid tether on the N terminus, a single conserved G internally spaced, and a CG tether on the C terminus. Microdomain 2 extends from the CG tether to a KI/L tether on its C terminus. To identify the corresponding microdomains in the Nayarit strains, pClust was used to cluster the different sets of microdomains (15, 16). The input parameters used for filtering and alignment were as follows: AlignOverLongerSeq ⫽ 7 (70%), MatchSimilarity ⫽ 8 (80%), OptimalScoreOverSelfScore ⫽ 5 (50%), SlideWindowSize ⫽ 3, and ExactMatchLen ⫽ 2. Following the clustering step, those clusters without representation from the
reference St. Maries strain were discarded. Multiple-sequence alignment was performed on the remaining clusters using Clustal X (17). Default parameters were used, except for the scoring matrix, which was changed from the default Gonnet series to the percent accepted mutation (PAM) series. Antibody selection. To identify Msp2 microdomain-specific antibodies that developed during infection in a natural host, five age-matched, competitive enzyme-linked immunosorbent assay (CELISA)-seronegative Holstein calves were infected with the St. Maries strain. The bovine lymphocyte antigen (BoLA) class II histocompatibility antigen, DR beta chain (DR3) haplotypes of the major histocompatibility complex (MHC) were determined for each animal as described below. All the animals were infected, as confirmed by microscopic detection of A. marginale bacteremia and CELISA seroconversion. Sera were collected weekly for 5 months and used to identify high-titer antibody specific to the St. Maries microdomains 1 and 2. Protocols were approved by the Washington State University Institutional Animal Care and Use Committee. Hightiter antibodies for these microdomains were identified in four calves.
FIG 3 Primary sequences of the Nayarit microdomain 1 variants compared to the homologous St. Maries strain. The microdomain 1 sequences encoded by the Nayarit strains are indicated below the corresponding sequence encoded by the St. Maries strain alleles 2, E6F7, and G11 (boldface). The flanking N-terminal (L/VSK/QKVC) and C-terminal (CG) tethers and conserved amino acids within the hypervariable microdomain are indicated in purple. Insertions relative to the St. Maries sequence are indicated in green, deletions in yellow, and substitutions in red. Changes in charge and hydrophobicity relative to the St. Maries sequence are indicated as follows: a, positive to neutral; b, negative to neutral; c, neutral to positive or insertion of a positive charge; d, neutral to negative or insertion of a negative charge.
4180
iai.asm.org
Infection and Immunity
November 2015 Volume 83 Number 11
A. marginale Msp2 Segmental Variants
FIG 4 Primary sequences of the Nayarit microdomain 2 variants compared to the homologous St. Maries strain. The microdomain 2 sequences encoded by the Nayarit strains are indicated below the corresponding sequence encoded by St. Maries strain alleles 9H1, 1, and E6F7 (boldface). The flanking N-terminal (CG) and C-terminal (KL/I) tethers and conserved amino acids within the hypervariable microdomain are indicated in purple. Insertions relative to the St. Maries sequence are indicated in green, deletions in yellow, and substitutions in red. Changes in charge and hydrophobicity relative to the St. Maries sequence are indicated as follows: a, positive to neutral; b, negative to neutral; c, neutral to positive or insertion of a positive charge; d, neutral to negative or insertion of a negative charge.
Peptides representing microdomains 1 and 2 of the complete St. Maries allelic repertoire were synthesized and purified by high-performance liquid chromatography (NeoBiolab) and used to coat Immulon-II 96-well plates (ThermoScientific). Replicate wells were coated with 50 l containing 20 g peptide dissolved in buffer (50 mM Na2CO3, pH 9.4) and incubated overnight at 4°C. The plates were washed 5 times with phosphatebuffered saline (PBS) containing 0.2% (vol/vol) Tween 20 and then blocked with 300 l/well of PBS containing 5% (wt/vol) milk and 0.2% Tween 20 for 1 h. To titrate the sera collected at each time point, 50 l of sera diluted at 1:300, 1:50, 1:10, and 1:5 in blocking buffer was added to triplicate wells and incubated for 1 h at room temperature. Following washing 5 times with PBS containing 0.2% Tween 20, IgG binding was detected by adding 50 l/well of recombinant protein G-horseradish peroxidase (Life Technologies) diluted 1:500, and the plates were incubated for 1 h at room temperature. After additional washes, 100 l/well of Sureblue microwell peroxidase substrate (KPL) was added for 15 min, the reaction was terminated with 100 l of 1% HCl, and binding was determined by measuring the optical density at 450 nm (OD450). Positive bind-
November 2015 Volume 83 Number 11
ing was defined statistically as exceeding the mean plus 2 standard deviations of the OD450 of preinfection serum from the same animal for the specific peptide. The dilution of each serum for each microdomain peptide that represented approximately 50% binding was determined. This midpoint allows sensitive detection of either increased or decreased binding to variant peptides. Antibody binding to variant peptides. Peptides representing the variant microdomains 1 and 2 of the Nayarit strains were synthesized in a fashion identical to that described above and tested for binding of antibodies specific to the homologous St. Maries microdomains 1 and 2 using the 50% breakpoint dilution, as described above. The tethers flanking the microdomains were synthesized to be identical, ensuring that only differences within the structural microdomains were analyzed for antibody binding (Fig. 3 and 4). Binding to variant peptides was determined statistically in comparison to binding to the homologous St. Maries strain microdomain peptide and to a negative-control peptide, P1, which is the Msp2 signal peptide that is absent from the mature protein and consequently does not elicit an immune response. Each serum was tested in
Infection and Immunity
iai.asm.org
4181
Graça et al.
TABLE 1 Microdomain variants resulting in complete ablation of antibody bindinga Peptide nameb Peptide sequencec SM.E6F7 Md-1 VP.E6F7-1.1 VP.E6F7-1.3 SM.E6F7 Md-2 VP.E6F7-2.5 VP.E6F7-2.7 VP.E6F7-2.8 VP.E6F7-2.10 SM.G11 Md-1 VP.G11-1.5 SM.StM1 Md-2 VP.StM1-2.2 VP.StM1-2.5 VP.StM1-2.6 VP.StM1-2.7 VP.StM1-2.15
GNGTGSSGSN GNGTDSSGSNH GNGTGS---N KNTTDSTNNNG -NTADGNATTQRKN-TSGTAAQTNG KN-TAGTAAQTNG KN-TS-TAGATT KGTGSTGSSGNK -GNG-TSG-GSTGT VNATSGSTNNG VN-T-GNTNNG VN-T-GNTNNNG VN-T-G-TNNNG VN-T-NTN-VNAASGTTT---
BoLA-DR3 haplotyping (*0601, *1001) (*0902, *1101) (*0103, *0101) (*0101, *2703) (*0902, *1101)
(*0601, *1001) (*0101, *2703) (*0902, *1101) (*0103, *0101) (*0601, *1001) (*0902, *1101)
a Variant peptides that resulted in complete ablation of antibody recognition, defined as binding that was statistically indistinguishable from that of preinfection serum in all biological replicates. b Boldface indicates the parent St. Maries strain homologous peptide; VP indicates the specific variant peptide, as defined in Fig. 2 and 3. c Italics indicate insertions relative to the parent St. Maries sequence, dashes indicate deletions, and underlining indicates substitutions.
triplicate for the variant peptide, the homologous St. Maries strain peptide, and the negative-control P1. Each peptide was tested with sera from at least two infected animals to provide a biological, as well as a technical, replicate. Binding was analyzed using one-way analysis of variance (ANOVA) with Tukey’s correction for multiple comparisons to maintain 95% joint confidence levels with a P value of ⱕ0.005. Determination of MHC DR3 haplotypes. DNA was extracted from whole blood of each animal using a Puregene Blood DNA kit (Gentra). The BoLA-DR3 loci were amplified using Polymerase PCR SuperMix High Fidelity (Invitrogen) with forward (5=-CGCTCCTGC[C/T]CAGAT CTATCC-3=) and reverse (5=-CACCCCCGCGCTCACC-3=) primers (18, 19). Briefly, amplification was performed using an initial step of 95°C for 5 min, followed by 30 cycles (94°C for 2.5 min, 57°C for 30 s, and 72°C for 1 min) and a final elongation step at 72°C for 5 min. The amplified PCR products were electrophoresed on 1% agarose gels, stained with SYBR Safe (Invitrogen), and visualized under UV light to confirm the presence of an amplicon of the expected size. Sequencing reactions were performed with a BigDye version 3.1 Cycle Sequencing kit from Applied Biosystems with an ABI 3730X genetic analyzer, and the haplotypes were assigned using ASSIGN 400ATF.
RESULTS
Microdomain structural variants. Application of pClust and ClustalX identified 38 unique primary structure variants in the Nayarit strains corresponding to microdomains 1 and 2 of the St. Maries strains for which high-titer antibodies were identified. All Nayarit microdomain 1 variant peptides (n ⫽ 12) differed from the microdomain 1 sequences encoded by the St. Maries strain alleles by amino acid substitutions, insertions, and deletions (Fig. 3). The minimum change from a St. Maries microdomain 1 sequence was a single amino acid insertion (VP.E6F7-1.2 compared to E6F7 microdomain 1), and the maximum change was 10 amino acids (VP.G11-1.5 compared to G11), incorporating substitutions, insertions, and deletions (Fig. 3). Identity between a given microdomain 1 variant and the corresponding St. Maries amino acid sequence, excluding the flanking tethers, ranged from 29% to 91% and included changes in charge without changes in hydropathy. As with microdomain 1, microdomain 2 variants (n ⫽ 26) differed from the sequences encoded by the St. Maries strain alleles by amino acid substitutions, insertions,
4182 iai.asm.org
TABLE 2 Microdomain variants resulting in significant loss of antibody bindinga Peptide nameb
Peptide sequencec
SM.StM2 Md-1 VP.StM2-1.1d VP.StM2-1.2d SM.G11 Md-1 VP.G11-1.1d VP.G11-1.2d VP.G11-1.3d VP.G11-1.4d SM.9H1 Md-2 VP.9H1-2.1d SM.E6/F7 Md-2 VP.E6F7-2.1 VP.E6F7-2.2d VP.E6F7-2.3d VP.E6F7-2.4d VP.E6F7-2.6d VP.E6F7-2.9d SM.StM1 Md-2 VP.StM1-2.1 VP.StM1-2.3d VP.StM1-2.4 VP.StM1-2.8d VP.StM1-2.9d VP.StM1-2.10 VP.StM1-2.11 VP.StM1-2.12d VP.StM1-2.14d
GKGEGSNGTKK -KGTGSTGT-K -KGTGTGSDGTKK KGTGSTGSSGNK -GKG-TGSTGTT -GNG-TGSTGTT KGTG-TDSSGNK -GNG-TSGTGGTQ KNSGDTNGSSTTQH KNSGDTSGS-TTQH KNTTDSTNNNG KNTTDGTT--KNTADGTTATT KNTADGTT--KNTADGNATTQRKNATSGT-AQTNG ---T-GTAGATT VNATSGSTNNG VNATSGSTNNNG VNATSG-TAT T VNATSGTAQTN-G VD-T-GNTNNNG VNATSGTAAQTN-G VNATSGTAQT--VNATSGTAQR--VNAASGTTAQT--VNATSGTTT----
BoLA-DR3 haplotyping (*0601, *1001)
(*0101, *2703)
(*0902, *1101)
(*0601, *1001)
(*0101, *2703)
(*0902, *1101)
(*0103, *0101)
(*0101, *2703)
(*0902, *1101)
(*0103, *0101)
(*0101, *2703)
(*0902, *1101)
(*0103, *0101)
(*0601, *1001)
(*0902, *1101)
a Variant peptides that resulted in a significant decrease in antibody recognition, defined as binding that was statistically significantly less than binding to the homologous St. Maries strain peptide in all biological replicates. b Boldface indicates the parent St. Maries strain homologous peptide; VP indicates the specific variant peptide as defined in Fig. 2 and 3. c Italics indicate insertions relative to the parent St. Maries sequence, dashes indicate deletions, and underlining indicates substitutions. d Variant peptide that resulted in complete ablation of antibody binding in at least one biological replicate.
and deletions (Fig. 4). The minimum change from a St. Maries microdomain 2 sequence was a single amino acid insertion (VP.StM1-2.1 compared to microdomain 2 encoded by St. Maries allele 1), and the maximum change was 10 amino acids (VP.E6F7-2.9 compared to allele E6F7), incorporating substitutions, insertions, and deletions (Fig. 4). Identity between a given microdomain 2 variant and the corresponding St. Maries amino acid sequence, excluding the flanking tethers, ranged from 18% to 92% and included changes in charge and hydropathy. Microdomain antigenic variants. Antigenic variants were defined by a statistically significant reduction (P ⬍ 0.005) in antibody binding compared to the corresponding homologous St. Maries microdomain. Complete ablation of antibody recognition, defined as binding that was statistically indistinguishable from that of preinfection serum, was observed across all biological replicates for 12/38 structural variants (Table 1). An additional 22 structural variants had statistically significantly reduced binding compared to the homologous microdomain across all biological replicates (Table 2). Of these 22, 17 demonstrated complete ablation of binding in at least one replicate, with reduced binding, defined as statistically significantly lower than that of the homologous microdomain but greater than that of preinfection serum, in the remaining replicates. Collectively, loss of antibody binding, including complete ablation, occurred using sera from all biological replicates for 34/38 structural variants. Loss of reactivity, either complete or partial, was not MHC specific, as similar binding patterns could be observed using sera from animals that did not share identical haplotypes (Tables 1 and 2). The remaining four
Infection and Immunity
November 2015 Volume 83 Number 11
A. marginale Msp2 Segmental Variants
TABLE 3 Microdomain variants resulting in inconsistent loss of antibody bindinga Peptide nameb
Peptide sequencec
SM.E6F7 Md-1 VP.E6F7-1.2 VP.E6F7-1.4 SM.StM2 Md-1 VP.StM2-1.3 SM.StM1 Md-2 VP.StM1-2.13
GNGTGSSGSN GNGTGSSGSNH GKGTGTSG-NH KGKEGSNGTKK KGKTTSGSDGNTKK VNATSGSTNNG VN--SGTTGST-NGN
BoLA-DR3 haplotyping (*0601, *1001)
(*0902, *1101)
(*0601, *1001)
(*0101, *2703)
(*0902, *1101)
(*0103, *0101)
(*0601, *1001)
(*0902, *1101)
a
Variant peptides that resulted in a statistically significant decrease in antibody recognition in a minimum of one but not in all biological replicates. b Boldface indicates the parent St. Maries strain homologous peptide; VP indicates the specific variant peptide as defined in Fig. 2 and 3. c Italics indicate insertions relative to the parent St. Maries sequence, dashes indicate deletions, and underlining indicates substitutions.
structural variants met the definition of an antigenic variant when tested using serum from at least one infected animal but did not show a significant loss of binding across all biological replicates (Table 3). There was no consistent pattern of reactivity among these structural variants associated with a specific MHC haplotype, but rather, they differed among the individual microdomains. A single amino acid difference was sufficient to generate an antigenic variant, as shown by the failure of antibody against microdomain 2, encoded by St. Maries allele 1, to equally bind variant VP.StM1-2.1. This variant differed from the homologous St. Maries microdomain by a single codon insertion, encoding a third asparagine (Fig. 4 and Table 2). Indels were present in the overwhelming majority of the antigenic variants with either complete or partial loss of binding (Fig. 3 and 4 and Tables 1 and 2), with or without associated amino acid substitutions. However, amino acid substitutions alone were sufficient to create an antigenic variant, as illustrated by variant VP.E6F7-2.2 (Fig. 4 and Table 2). Detailed enzyme-linked immunosorbent assay (ELISA) data are provided in Fig. S1 in the supplemental material. DISCUSSION
The Msp2 variant repertoire encoded only by recombination of the HVRs represented by the “whole” allele is defined by the number of alleles in the genome (5, 20). This allelic repertoire is limited among A. marginale strains and is ⬍10 per genome; the prototype St. Maries strain has only five unique msp2 alleles. Segmental gene conversion uses combinatorial recombination to generate hundreds to thousands of novel sequence variants not represented elsewhere in the genome other than transiently in the expression site (9). For this expansion to be functional in pathogen persistence requires that they evade the extant antibody response (12, 21, 22). The fraction of structural variants that represent antigenic variants has been unknown. The current study supports the notion that the HVR, composed of highly conserved tethers flanking variable microdomains, allows both structural fitness and antigenic variation. Of the 38 expressed primary structure variant microdomains examined, 12 were unrecognized by high-titer antibody against the homologous microdomain across all biological replicates: antibody binding to the variant was diminished to the level of preimmune serum. An additional 22 variant microdomains had a statistically significant reduction in binding across all replicates. Collectively, 89% of the expressed structural variants were also antigenic variants across all replicates, independent of a specific MHC haplotype. The biological relevance of these findings is supported by the following four lines of evidence: (i) all
November 2015 Volume 83 Number 11
structural variants were expressed during infection of a natural host, (ii) the structural variation observed in the microdomains corresponded to the mean length of variants generated by segmental gene conversion in the St. Maries strain (29 ⫾ 13 nucleotides, based on examination of 1,183 variants [9]), (iii) antigenic variants were identified using a broad antibody response that developed during infection of a natural host, and (iv) the ablation of antibody binding to a variant microdomain is in agreement with the results demonstrating lack of binding to a newly emergent full-length HVR (22, 23). The significant loss of antibody binding to a structural variant with only a single amino acid insertion demonstrates how minimal change can result in a new antigenic variant. While it is well established using monoclonal antibodies that a single amino acid change in an antigen is sufficient to eliminate binding, the significance in the present study is that the single insertion must retain minimal protein structure for bacterial viability and evade a polyclonal response. Interestingly, microdomain 2 variant VP.StM12.1 differed from the homologous microdomain by a single codon insertion that resulted in an additional asparagine. Whether this reflects a preferred genetic mechanism of codon duplication, a structural selection to retain function, or a combination of the two is unknown. Our lack of understanding of which changes allow retention of protein function while generating an antigenic variant represents a significant gap in knowledge and will require additional structure-function studies. We examined sera from animals with four unique MHC haplotypes. Fifteen structural variants, including 12 for which complete ablation of binding was observed, had consistent binding patterns across all replicates, regardless of the MHC haplotype. The remaining structural variants represented antigenic variants with complete ablation of binding with serum from at least one animal but either a significant reduction in binding, but not complete ablation (Table 2), or no loss of binding (Table 3) using sera from other animals, representing independent biological replicates. There was no consistent pattern of binding, or lack thereof, across MHC haplotypes. This very likely reflects sequential selection for variants, where multiple new variants are generated by segmental gene conversion, structurally fit variants are viable and initiate replication, and then the selective pressure of the immune response allows only those antigenic variants to continue replication and form a bacteremic peak. Which variants compose the bacteremic peak depends on the extant antibody response of the infected animal, which in turn is dependent on both the MHC and prior variant exposure. The fact that a very high proportion of structural variants represented true antigenic variants supports the efficiency of segmental gene conversion in generating Msp2 antigenic variants. Antigenic variation allows A. marginale persistence within the individual infected animal, allowing vector-borne transmission (22, 24, 25). Importantly, Msp2 antigenic variation has also been shown to drive strain superinfection at the host population level (10, 12, 26, 27). Strains with different msp2 allelic repertoires are capable of infecting hosts previously infected with a primary strain that had mounted an immune response against it. Even a single unique msp2 allele is sufficient to allow strain superinfection, with the requirement that the unique allele be expressed at the time of superinfection (10). Within regions of endemicity with high population immunity against a dominant strain, there is strong selective pressure for emergence of new strains, and consequently,
Infection and Immunity
iai.asm.org
4183
Graça et al.
strain superinfection is common (12, 21, 24, 27). While there is genomic evidence that a new strain may emerge by insertions or deletions within a previously duplicated msp2 allele, the magnitude of change in the Msp2 primary structure required to generate a truly antigenically unique strain capable of superinfection is unknown. The current study, which uses expressed variants from the Nayarit and St. Maries strains, supports the idea that very minor structural changes, especially insertions/deletions, are sufficient to create an escape variant. This is consistent with the results of two prior studies. The first is the comparative genomics approach that identified otherwise identical alleles in the St. Maries and Florida strains (alleles 1 and NAI1011, respectively) that differed by two indels and amino acid substitutions in microdomain 2 (11), variation within the range observed for antigenic variants in the current study. The second study mapped 1,183 expressed variants to the encoding alleles (9). While 99% of the variants mapped identically to the alleles, 1% of the variants contained unique sequences that differed by small insertions or deletions similar to those identified here as antigenic variants (9). Collectively, these studies suggest that relatively small genetic changes, predominantly insertions and deletions, can generate antigenic variants and, when maintained in the genome as a novel allele, would provide the basis for emergence of a new strain with selection for the capacity to superinfect.
11.
12.
13.
14.
15. 16.
17.
ACKNOWLEDGMENTS Research was supported by National Institutes of Health grant R37 AI44005. Telmo Graça was partially supported by a scholarship through Fundação para a Ciência e Tecnologia SFRH/BD/68377/2010. We appreciate the excellent technical support of Beverly Hunter.
18. 19.
REFERENCES 1. Holland J, Spindler K, Horodyski F, Grabau E, Nichol S, Vandepol S. 1982. Rapid evolution of RNA genomes. Science 215:1577–1585. http://dx .doi.org/10.1126/science.7041255. 2. Deitsch KW, Lukehart SA, Stringer JR. 2009. Common strategies for antigenic variation by bacterial, fungal and protozoan pathogens. Nat Rev Microbiol 7:493–503. http://dx.doi.org/10.1038/nrmicro2145. 3. Palmer GH, Bankhead T, Lukehart SA. 2009. ‘Nothing is permanent but change’—antigenic variation in persistent bacterial pathogens. Cell Microbiol 11:1697–1705. http://dx.doi.org/10.1111/j.1462-5822 .2009.01366.x. 4. Barbour AG, Restrepo BI. 2000. Antigenic variation in vector-borne pathogens. Emerg Infect Dis 6:449 – 457. http://dx.doi.org/10.3201/eid0605 .000502. 5. Brayton KA, Kappmeyer LS, Herndon DR, Dark MJ, Tibbals DL, Palmer GH, McGuire TC, Knowles DP, Jr. 2005. Complete genome sequencing of Anaplasma marginale reveals that the surface is skewed to two superfamilies of outer membrane proteins. Proc Natl Acad Sci U S A 102:844 – 849. http://dx.doi.org/10.1073/pnas.0406656102. 6. Palmer GH, Brayton KA. 2007. Gene conversion is a convergent strategy for pathogen antigenic variation. Trends Parasitol 23:408 – 413. http://dx .doi.org/10.1016/j.pt.2007.07.008. 7. Santoyo G, Romero D. 2005. Gene conversion and concerted evolution in bacterial genomes. FEMS Microbiol Rev 29:169 –183. http://dx.doi.org /10.1016/j.fmrre.2004.10.004. 8. Centurion-Lara A, LaFond RE, Hevner K, Godornes C, Molini BJ, Van Voorhis WC, Lukehart SA. 2004. Gene conversion: a mechanism for generation of heterogeneity in the tprK gene of Treponema pallidum during infection. Mol Microbiol 52:1579 –1596. http://dx.doi.org/10.1111/j .1365-2958.2004.04086.x. 9. Futse JE, Brayton KA, Knowles DP, Jr, Palmer GH. 2005. Structural basis for segmental gene conversion in generation of Anaplasma marginale outer membrane protein variants. Mol Microbiol 57:212–221. http://dx .doi.org/10.1111/j.1365-2958.2005.04670.x. 10. Futse JE, Brayton KA, Dark MJ, Knowles DP, Jr, Palmer GH. 2008.
4184
iai.asm.org
20.
21.
22. 23.
24. 25.
26.
27.
Superinfection as a driver of genomic diversification in antigenically variant pathogens. Proc Natl Acad Sci U S A 105:2123–2127. http://dx.doi.org /10.1073/pnas.0710333105. Dark MJ, Herndon DR, Kappmeyer LS, Gonzales MP, Nordeen E, Palmer GH, Knowles DP, Brayton KA. 2009. Conservation in the face of diversity: multistrain analysis of an intracellular bacterium. BMC Genomics 10:16. http://dx.doi.org/10.1186/1471-2164-10-16. Ueti MW, Tan Y, Broschat SL, Castañeda Ortiz EJ, Camacho-Nuez M, Mosqueda JJ, Scoles GA, Grimes M, Brayton KA, Palmer GH. 2012. Expansion of variant diversity associated with a high prevalence of pathogen strain superinfection under conditions of natural transmission. Infect Immun 80:2354 –2360. http://dx.doi.org/10.1128/IAI.00341-12. Zhuang Y, Futse JE, Brown WC, Brayton KA, Palmer GH. 2007. Maintenance of antibody to pathogen epitopes generated by segmental gene conversion is highly dynamic during long-term persistent infection. Infect Immun 75:5185–5190. http://dx.doi.org/10.1128/IAI.00913-07. Abbott JR, Palmer GH, Howard CJ, Hope JC, Brown WC. 2004. Anaplasma marginale major surface protein 2 CD4⫹-T-cell epitopes are evenly distributed in conserved and hypervariable regions (HVR), whereas linear B-cell epitopes are predominantly located in the HVR. Infect Immun 72: 7360 –7366. http://dx.doi.org/10.1128/IAI.72.12.7360-7366.2004. Wu CJ, Kalyanaraman A, Cannon WR. 2012. pGraph: efficient parallel construction of large-scale protein sequence homology graphs. IEEE Trans Parallel Distrib Syst 23:1923–1933. http://dx.doi.org/10.1109/TPDS.2012.19. Wu CJ, Kalyanaraman A. 2008. An efficient parallel approach for identifying protein families in large-scale metagenomic data sets, p 362–371. Proc ACM/IEEE Conf Supercomputing (SC’08), Austin, TX, November 15–21, 2008. IEEE Press, Piscataway, NJ. Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R, Thompson JD, Gibson TJ, Higgins DG. 2007. Clustal W and clustal X version 2.0. Bioinformatics 23:2947–2948. http://dx.doi.org/10.1093/bioinformatics/btm404. Miltiadou D, Law AS, Russell GC. 2003. Establishment of a sequencebased typing system for BoLA-DR3 exon 2. Tissue Antigens 62:55– 65. http://dx.doi.org/10.1034/j.1399-0039.2003.00080.x. Takeshima SN, Matsumoto Y, Miyasaka T, Arainga-Ramirez M, Saito H, Onuma M, Aida Y. 2011. A new method for typing bovine major histocompatibility complex class II DR3 alleles by combining two established PCR sequence-based techniques. Tissue Antigens 78:208 –213. http: //dx.doi.org/10.1111/j.1399-0039.2011.01708.x. Brayton KA, Knowles DP, McGuire TC, Palmer GH. 2001. Efficient use of a small genome to generate antigenic diversity in tick-borne ehrlichial pathogens. Proc Natl Acad Sci U S A 98:4130 – 4135. http://dx.doi.org/10 .1073/pnas.071056298. Vallejo Esquerra E, Herndon DR, Alpirez Mendoza F, Mosqueda J, Palmer GH. 2014. Anaplasma marginale superinfection attributable to pathogen strains with distinct genomic backgrounds. Infect Immun 82: 5286 –5292. http://dx.doi.org/10.1128/IAI.02537-14. French DM, Brown WC, Palmer GH. 1999. Emergence of Anaplasma marginale antigenic variants during persistent rickettsemia. Infect Immun 67:5834 –5840. Agnes JT, Brayton KA, LaFollett M, Norimine J, Brown WC, Palmer GH. 2011. Identification of Anaplasma marginale outer membrane protein antigens conserved between A. marginale sensu stricto strains and the live A. marginale subsp. centrale vaccine. Infect Immun 79:1311–1318. http://dx.doi.org/10.1128/IAI.01174-10. Palmer GH, Brayton KA. 2013. Antigenic variation and transmission fitness as drivers of bacterial strain structure. Cell Microbiol 15:1969 – 1975. http://dx.doi.org/10.1111/cmi.12182. Scoles GA, Broce AB, Lysyk TJ, Palmer GH. 2005. Relative efficiency of biological transmission of Anaplasma marginale (Rickettsiales: Anaplasmataceae) by Dermacentor andersoni (Acari: Ixodidae) compared with mechanical transmission by Stomoxys calcitrans (Diptera: Muscidae). J Med Entomol 42:668 – 675. Rodriguez JL, Palmer GH, Knowles DP, Brayton KA. 2005. Distinctly different msp2 pseudogene repertoires in Anaplasma marginale strains that are capable of superinfection. Gene 361:127–132. http://dx.doi.org /10.1016/j.gene.2005.06.038. Castaneda-Ortiz EJ, Ueti MW, Camacho-Nuez M, Mosqueda JJ, Mousel MR, Johnson WC, Palmer GH. 2015. Association of Anaplasma marginale strain superinfection with infection prevalence within tropical regions. PLoS One 10:e0120748. http://dx.doi.org/10.1371/journal.pone .0120748.
Infection and Immunity
November 2015 Volume 83 Number 11
Antigenic variation allows microbial pathogens to evade immune clearance and establish persistent infection. Anaplasma marginale utilizes gene conversion of a repertoire of silent msp2 alleles into a single active expression site to encode unique Msp2 variants. As the genomic complement of msp2 alleles alone is insufficient to generate the number of variants required for persistence, A. marginale uses segmental gene conversion, in which oligonucleotide segments from multiple alleles are recombined into the expression site to generate a novel msp2 mosaic not represented elsewhere in the genome. Whether these segmental changes are sufficient to evade a broad antibody response is unknown. We addressed this question by identifying Msp2 variants that differed in primary structure within the immunogenic hypervariable region microdomains and tested whether they represented true antigenic variants. The minimal primary structural difference between variants was a single amino acid resulting from a codon insertion, and overall, the amino acid identity among paired microdomains ranged from 18 to 92%. Collectively, 89% of the expressed structural variants were also antigenic variants across all biological replicates, independent of a specific host major histocompatibility complex haplotype. Biological relevance is supported by the following: (i) all structural variants were expressed during infection of a natural host, (ii) the structural variation observed in the microdomains corresponded to the mean length of variants generated by segmental gene conversion, and (iii) antigenic variants were identified using a broad antibody response that developed during infection of a natural host. The findings demonstrate that segmental gene conversion efficiently generates Msp2 antigenic variants.
M
icrobial pathogens utilize diverse mechanisms to persist in their mammalian hosts, evolutionarily driven to increase the likelihood of onward transmission and propagation of self. Antigenic variation is a convergent strategy used by pathogens ranging from RNA viruses to protozoa (1–4). Among bacteria and protozoa, recombinatorial mechanisms are commonly used to generate sequential antigenic variants from otherwise relatively stable genomes (2–5). For example, bacterial pathogens in the genera Anaplasma, Borrelia, Mycoplasma, Neisseria, and Treponema utilize gene conversion of a repertoire of silent alleles into expression sites to encode unique antigenic and structural variants (2–4). Gene conversion is unidirectional, leaving the donor allele unchanged but replacing the expression site sequence (6–8). Thus, the donor alleles are under long-term selective pressure to provide a template for expressing a variant that is sufficiently antigenically unique but also retains a viable structure (Fig. 1A). Pathogens in these genera also utilize segmental gene conversion, whereby an oligonucleotide segment, representing only a portion of the donor allele, is recombined into the expression site, resulting in a mosaic not present elsewhere in the genome (Fig. 1B and C) (6, 9). This process of segmental gene conversion has the potential to tremendously amplify the capacity to generate antigenic variants but can be limited by structural constraints, as these assembled mosaics, unlike the whole donor alleles, have not been under long-term selection for structural fitness. Using Anaplasma marginale infection in its natural bovine host, Futse et al. demonstrated that a unique Major Surface Protein 2 (Msp2) variant derived from a whole msp2 donor allele was sufficient to evade an existing broad antibody response generated against a repertoire of Msp2 variants (10). The unique donor allele encoded a hypervariable region (HVR) that shared 63% identity with HVRs to which the host had been previously exposed. This
4178
iai.asm.org
level of difference in encoded HVRs is reflected in the allelic repertoire of the multiple A. marginale strains examined, consistent with their deterministic role in antigenic variation (10, 11). In contrast, the degree of sequence difference that allows a variant generated by segmental gene conversion to escape immune recognition is unknown. If large differences are required to generate a variant capable of immune escape, the size of the potential repertoire would be reduced and segmental gene conversion would only marginally expand the repertoire compared to that generated by whole allelic conversion. To address this knowledge gap, we identified A. marginale St. Maries strain Msp2 variants, represented by their HVR microdomains that were recognized by IgG antibodies induced during infection of cattle, a natural host. A second set of expressed structural variants that had defined differences in the corresponding microdomains (12) but were not represented in the St. Maries strain and to which the animals had thus not been exposed were used to test whether these differences were
Received 30 June 2015 Returned for modification 14 July 2015 Accepted 4 August 2015 Accepted manuscript posted online 10 August 2015 Citation Graça T, Paradiso L, Broschat SL, Noh SM, Palmer GH. 2015. Primary structural variation in Anaplasma marginale Msp2 efficiently generates immune escape variants. Infect Immun 83:4178 –4184. doi:10.1128/IAI.00851-15. Editor: C. R. Roy Address correspondence to Telmo Graça, [email protected]. * Present address: Lydia Paradiso, School of Biology, University of Edinburgh, Edinburgh, United Kingdom. Supplemental material for this article may be found at http://dx.doi.org/10.1128 /IAI.00851-15. Copyright © 2015, American Society for Microbiology. All Rights Reserved.
Infection and Immunity
November 2015 Volume 83 Number 11
A. marginale Msp2 Segmental Variants
FIG 1 A. marginale Msp2 mechanism of antigenic variation. The mosaics generated through genomic recombination result in expression of new antigenically distinct Msp2 variants, which compose sequential bacteremic waves and allow persistence within the host. (A) Chromosomal msp2 loci, 7 silent alleles, and a single active expression site (ES) with a central hypervariable domain (HVR). P1, E6/F7, G11, 1, 2, 9H1, and 3H1 are the 7 conserved donor alleles for the Msp2 HVR, with 5 of them being unique (allele P1 is a duplication of allele G11, and allele 2 is a duplication of allele 3H1). (B) Segmental gene conversion from an oligonucleotide segment of a donor allele into the expression site, generating an expression site mosaic. (C) Bacteremic peaks characterized by expression of unique Msp2 variants, variants generated by recombination of a “whole” donor allele, predominate in acute bacteremia, followed by a persistent phase in which the mosaics generated by segmental gene conversion predominate. The modeled protein structure with the peptide sequences generated by a donor allele or segmental gene conversion is indicated above the bacteremic peak.
sufficient to evade immune recognition. This strategy incorporated three key elements: (i) the Msp2 variants were expressed during in vivo infection and thus represented at least minimal structural fitness; (ii) the microdomains tested represent the encoded variable regions; and (iii) the antibody response, induced by infection in the natural host, represented the relevant breadth of epitope specificity. Infection of a natural host is critically important to reflect the long-term evolutionary pressures associated
November 2015 Volume 83 Number 11
with mechanisms of pathogen persistence. Here, we present the results and discuss the findings in the context of the antigenic variant repertoire and the consequences for evasion of immunity at the levels of both the individual host and the host population. MATERIALS AND METHODS Variant selection. The immunogenic Msp2 HVR microdomains were matched between the complete repertoire of microdomains encoded by
Infection and Immunity
iai.asm.org
4179
Graça et al.
FIG 2 Graphical representation of the A. marginale St. Maries HVR microdomains and corresponding flanking tethers. T1, N-terminal flanking tether; T2, spacing tether between microdomain 1 and microdomain 2; T3, spacing tether between microdomain 2 and microdomain 3.
the St. Maries strain (GenBank accession no. CP000030) and those expressed by genetically distinct Nayarit strains (GenBank accession no. JQ933614 to JQ933796) (5, 12). Msp2 HVR microdomains 1 and 2 encompass the immunodominant B cell epitopes in the HVRs encoded by the allelic repertoire in the St. Maries strain (13, 14) and are defined by flanking conserved amino acids, previously designated “structural tethers” (Fig. 2). Conservation of the tethers was verified by analysis of ⬎1,600 expressed variants of the St. Maries strain and Nayarit strains (12). Microdomain 1 has a 6-amino-acid tether on the N terminus, a single conserved G internally spaced, and a CG tether on the C terminus. Microdomain 2 extends from the CG tether to a KI/L tether on its C terminus. To identify the corresponding microdomains in the Nayarit strains, pClust was used to cluster the different sets of microdomains (15, 16). The input parameters used for filtering and alignment were as follows: AlignOverLongerSeq ⫽ 7 (70%), MatchSimilarity ⫽ 8 (80%), OptimalScoreOverSelfScore ⫽ 5 (50%), SlideWindowSize ⫽ 3, and ExactMatchLen ⫽ 2. Following the clustering step, those clusters without representation from the
reference St. Maries strain were discarded. Multiple-sequence alignment was performed on the remaining clusters using Clustal X (17). Default parameters were used, except for the scoring matrix, which was changed from the default Gonnet series to the percent accepted mutation (PAM) series. Antibody selection. To identify Msp2 microdomain-specific antibodies that developed during infection in a natural host, five age-matched, competitive enzyme-linked immunosorbent assay (CELISA)-seronegative Holstein calves were infected with the St. Maries strain. The bovine lymphocyte antigen (BoLA) class II histocompatibility antigen, DR beta chain (DR3) haplotypes of the major histocompatibility complex (MHC) were determined for each animal as described below. All the animals were infected, as confirmed by microscopic detection of A. marginale bacteremia and CELISA seroconversion. Sera were collected weekly for 5 months and used to identify high-titer antibody specific to the St. Maries microdomains 1 and 2. Protocols were approved by the Washington State University Institutional Animal Care and Use Committee. Hightiter antibodies for these microdomains were identified in four calves.
FIG 3 Primary sequences of the Nayarit microdomain 1 variants compared to the homologous St. Maries strain. The microdomain 1 sequences encoded by the Nayarit strains are indicated below the corresponding sequence encoded by the St. Maries strain alleles 2, E6F7, and G11 (boldface). The flanking N-terminal (L/VSK/QKVC) and C-terminal (CG) tethers and conserved amino acids within the hypervariable microdomain are indicated in purple. Insertions relative to the St. Maries sequence are indicated in green, deletions in yellow, and substitutions in red. Changes in charge and hydrophobicity relative to the St. Maries sequence are indicated as follows: a, positive to neutral; b, negative to neutral; c, neutral to positive or insertion of a positive charge; d, neutral to negative or insertion of a negative charge.
4180
iai.asm.org
Infection and Immunity
November 2015 Volume 83 Number 11
A. marginale Msp2 Segmental Variants
FIG 4 Primary sequences of the Nayarit microdomain 2 variants compared to the homologous St. Maries strain. The microdomain 2 sequences encoded by the Nayarit strains are indicated below the corresponding sequence encoded by St. Maries strain alleles 9H1, 1, and E6F7 (boldface). The flanking N-terminal (CG) and C-terminal (KL/I) tethers and conserved amino acids within the hypervariable microdomain are indicated in purple. Insertions relative to the St. Maries sequence are indicated in green, deletions in yellow, and substitutions in red. Changes in charge and hydrophobicity relative to the St. Maries sequence are indicated as follows: a, positive to neutral; b, negative to neutral; c, neutral to positive or insertion of a positive charge; d, neutral to negative or insertion of a negative charge.
Peptides representing microdomains 1 and 2 of the complete St. Maries allelic repertoire were synthesized and purified by high-performance liquid chromatography (NeoBiolab) and used to coat Immulon-II 96-well plates (ThermoScientific). Replicate wells were coated with 50 l containing 20 g peptide dissolved in buffer (50 mM Na2CO3, pH 9.4) and incubated overnight at 4°C. The plates were washed 5 times with phosphatebuffered saline (PBS) containing 0.2% (vol/vol) Tween 20 and then blocked with 300 l/well of PBS containing 5% (wt/vol) milk and 0.2% Tween 20 for 1 h. To titrate the sera collected at each time point, 50 l of sera diluted at 1:300, 1:50, 1:10, and 1:5 in blocking buffer was added to triplicate wells and incubated for 1 h at room temperature. Following washing 5 times with PBS containing 0.2% Tween 20, IgG binding was detected by adding 50 l/well of recombinant protein G-horseradish peroxidase (Life Technologies) diluted 1:500, and the plates were incubated for 1 h at room temperature. After additional washes, 100 l/well of Sureblue microwell peroxidase substrate (KPL) was added for 15 min, the reaction was terminated with 100 l of 1% HCl, and binding was determined by measuring the optical density at 450 nm (OD450). Positive bind-
November 2015 Volume 83 Number 11
ing was defined statistically as exceeding the mean plus 2 standard deviations of the OD450 of preinfection serum from the same animal for the specific peptide. The dilution of each serum for each microdomain peptide that represented approximately 50% binding was determined. This midpoint allows sensitive detection of either increased or decreased binding to variant peptides. Antibody binding to variant peptides. Peptides representing the variant microdomains 1 and 2 of the Nayarit strains were synthesized in a fashion identical to that described above and tested for binding of antibodies specific to the homologous St. Maries microdomains 1 and 2 using the 50% breakpoint dilution, as described above. The tethers flanking the microdomains were synthesized to be identical, ensuring that only differences within the structural microdomains were analyzed for antibody binding (Fig. 3 and 4). Binding to variant peptides was determined statistically in comparison to binding to the homologous St. Maries strain microdomain peptide and to a negative-control peptide, P1, which is the Msp2 signal peptide that is absent from the mature protein and consequently does not elicit an immune response. Each serum was tested in
Infection and Immunity
iai.asm.org
4181
Graça et al.
TABLE 1 Microdomain variants resulting in complete ablation of antibody bindinga Peptide nameb Peptide sequencec SM.E6F7 Md-1 VP.E6F7-1.1 VP.E6F7-1.3 SM.E6F7 Md-2 VP.E6F7-2.5 VP.E6F7-2.7 VP.E6F7-2.8 VP.E6F7-2.10 SM.G11 Md-1 VP.G11-1.5 SM.StM1 Md-2 VP.StM1-2.2 VP.StM1-2.5 VP.StM1-2.6 VP.StM1-2.7 VP.StM1-2.15
GNGTGSSGSN GNGTDSSGSNH GNGTGS---N KNTTDSTNNNG -NTADGNATTQRKN-TSGTAAQTNG KN-TAGTAAQTNG KN-TS-TAGATT KGTGSTGSSGNK -GNG-TSG-GSTGT VNATSGSTNNG VN-T-GNTNNG VN-T-GNTNNNG VN-T-G-TNNNG VN-T-NTN-VNAASGTTT---
BoLA-DR3 haplotyping (*0601, *1001) (*0902, *1101) (*0103, *0101) (*0101, *2703) (*0902, *1101)
(*0601, *1001) (*0101, *2703) (*0902, *1101) (*0103, *0101) (*0601, *1001) (*0902, *1101)
a Variant peptides that resulted in complete ablation of antibody recognition, defined as binding that was statistically indistinguishable from that of preinfection serum in all biological replicates. b Boldface indicates the parent St. Maries strain homologous peptide; VP indicates the specific variant peptide, as defined in Fig. 2 and 3. c Italics indicate insertions relative to the parent St. Maries sequence, dashes indicate deletions, and underlining indicates substitutions.
triplicate for the variant peptide, the homologous St. Maries strain peptide, and the negative-control P1. Each peptide was tested with sera from at least two infected animals to provide a biological, as well as a technical, replicate. Binding was analyzed using one-way analysis of variance (ANOVA) with Tukey’s correction for multiple comparisons to maintain 95% joint confidence levels with a P value of ⱕ0.005. Determination of MHC DR3 haplotypes. DNA was extracted from whole blood of each animal using a Puregene Blood DNA kit (Gentra). The BoLA-DR3 loci were amplified using Polymerase PCR SuperMix High Fidelity (Invitrogen) with forward (5=-CGCTCCTGC[C/T]CAGAT CTATCC-3=) and reverse (5=-CACCCCCGCGCTCACC-3=) primers (18, 19). Briefly, amplification was performed using an initial step of 95°C for 5 min, followed by 30 cycles (94°C for 2.5 min, 57°C for 30 s, and 72°C for 1 min) and a final elongation step at 72°C for 5 min. The amplified PCR products were electrophoresed on 1% agarose gels, stained with SYBR Safe (Invitrogen), and visualized under UV light to confirm the presence of an amplicon of the expected size. Sequencing reactions were performed with a BigDye version 3.1 Cycle Sequencing kit from Applied Biosystems with an ABI 3730X genetic analyzer, and the haplotypes were assigned using ASSIGN 400ATF.
RESULTS
Microdomain structural variants. Application of pClust and ClustalX identified 38 unique primary structure variants in the Nayarit strains corresponding to microdomains 1 and 2 of the St. Maries strains for which high-titer antibodies were identified. All Nayarit microdomain 1 variant peptides (n ⫽ 12) differed from the microdomain 1 sequences encoded by the St. Maries strain alleles by amino acid substitutions, insertions, and deletions (Fig. 3). The minimum change from a St. Maries microdomain 1 sequence was a single amino acid insertion (VP.E6F7-1.2 compared to E6F7 microdomain 1), and the maximum change was 10 amino acids (VP.G11-1.5 compared to G11), incorporating substitutions, insertions, and deletions (Fig. 3). Identity between a given microdomain 1 variant and the corresponding St. Maries amino acid sequence, excluding the flanking tethers, ranged from 29% to 91% and included changes in charge without changes in hydropathy. As with microdomain 1, microdomain 2 variants (n ⫽ 26) differed from the sequences encoded by the St. Maries strain alleles by amino acid substitutions, insertions,
4182 iai.asm.org
TABLE 2 Microdomain variants resulting in significant loss of antibody bindinga Peptide nameb
Peptide sequencec
SM.StM2 Md-1 VP.StM2-1.1d VP.StM2-1.2d SM.G11 Md-1 VP.G11-1.1d VP.G11-1.2d VP.G11-1.3d VP.G11-1.4d SM.9H1 Md-2 VP.9H1-2.1d SM.E6/F7 Md-2 VP.E6F7-2.1 VP.E6F7-2.2d VP.E6F7-2.3d VP.E6F7-2.4d VP.E6F7-2.6d VP.E6F7-2.9d SM.StM1 Md-2 VP.StM1-2.1 VP.StM1-2.3d VP.StM1-2.4 VP.StM1-2.8d VP.StM1-2.9d VP.StM1-2.10 VP.StM1-2.11 VP.StM1-2.12d VP.StM1-2.14d
GKGEGSNGTKK -KGTGSTGT-K -KGTGTGSDGTKK KGTGSTGSSGNK -GKG-TGSTGTT -GNG-TGSTGTT KGTG-TDSSGNK -GNG-TSGTGGTQ KNSGDTNGSSTTQH KNSGDTSGS-TTQH KNTTDSTNNNG KNTTDGTT--KNTADGTTATT KNTADGTT--KNTADGNATTQRKNATSGT-AQTNG ---T-GTAGATT VNATSGSTNNG VNATSGSTNNNG VNATSG-TAT T VNATSGTAQTN-G VD-T-GNTNNNG VNATSGTAAQTN-G VNATSGTAQT--VNATSGTAQR--VNAASGTTAQT--VNATSGTTT----
BoLA-DR3 haplotyping (*0601, *1001)
(*0101, *2703)
(*0902, *1101)
(*0601, *1001)
(*0101, *2703)
(*0902, *1101)
(*0103, *0101)
(*0101, *2703)
(*0902, *1101)
(*0103, *0101)
(*0101, *2703)
(*0902, *1101)
(*0103, *0101)
(*0601, *1001)
(*0902, *1101)
a Variant peptides that resulted in a significant decrease in antibody recognition, defined as binding that was statistically significantly less than binding to the homologous St. Maries strain peptide in all biological replicates. b Boldface indicates the parent St. Maries strain homologous peptide; VP indicates the specific variant peptide as defined in Fig. 2 and 3. c Italics indicate insertions relative to the parent St. Maries sequence, dashes indicate deletions, and underlining indicates substitutions. d Variant peptide that resulted in complete ablation of antibody binding in at least one biological replicate.
and deletions (Fig. 4). The minimum change from a St. Maries microdomain 2 sequence was a single amino acid insertion (VP.StM1-2.1 compared to microdomain 2 encoded by St. Maries allele 1), and the maximum change was 10 amino acids (VP.E6F7-2.9 compared to allele E6F7), incorporating substitutions, insertions, and deletions (Fig. 4). Identity between a given microdomain 2 variant and the corresponding St. Maries amino acid sequence, excluding the flanking tethers, ranged from 18% to 92% and included changes in charge and hydropathy. Microdomain antigenic variants. Antigenic variants were defined by a statistically significant reduction (P ⬍ 0.005) in antibody binding compared to the corresponding homologous St. Maries microdomain. Complete ablation of antibody recognition, defined as binding that was statistically indistinguishable from that of preinfection serum, was observed across all biological replicates for 12/38 structural variants (Table 1). An additional 22 structural variants had statistically significantly reduced binding compared to the homologous microdomain across all biological replicates (Table 2). Of these 22, 17 demonstrated complete ablation of binding in at least one replicate, with reduced binding, defined as statistically significantly lower than that of the homologous microdomain but greater than that of preinfection serum, in the remaining replicates. Collectively, loss of antibody binding, including complete ablation, occurred using sera from all biological replicates for 34/38 structural variants. Loss of reactivity, either complete or partial, was not MHC specific, as similar binding patterns could be observed using sera from animals that did not share identical haplotypes (Tables 1 and 2). The remaining four
Infection and Immunity
November 2015 Volume 83 Number 11
A. marginale Msp2 Segmental Variants
TABLE 3 Microdomain variants resulting in inconsistent loss of antibody bindinga Peptide nameb
Peptide sequencec
SM.E6F7 Md-1 VP.E6F7-1.2 VP.E6F7-1.4 SM.StM2 Md-1 VP.StM2-1.3 SM.StM1 Md-2 VP.StM1-2.13
GNGTGSSGSN GNGTGSSGSNH GKGTGTSG-NH KGKEGSNGTKK KGKTTSGSDGNTKK VNATSGSTNNG VN--SGTTGST-NGN
BoLA-DR3 haplotyping (*0601, *1001)
(*0902, *1101)
(*0601, *1001)
(*0101, *2703)
(*0902, *1101)
(*0103, *0101)
(*0601, *1001)
(*0902, *1101)
a
Variant peptides that resulted in a statistically significant decrease in antibody recognition in a minimum of one but not in all biological replicates. b Boldface indicates the parent St. Maries strain homologous peptide; VP indicates the specific variant peptide as defined in Fig. 2 and 3. c Italics indicate insertions relative to the parent St. Maries sequence, dashes indicate deletions, and underlining indicates substitutions.
structural variants met the definition of an antigenic variant when tested using serum from at least one infected animal but did not show a significant loss of binding across all biological replicates (Table 3). There was no consistent pattern of reactivity among these structural variants associated with a specific MHC haplotype, but rather, they differed among the individual microdomains. A single amino acid difference was sufficient to generate an antigenic variant, as shown by the failure of antibody against microdomain 2, encoded by St. Maries allele 1, to equally bind variant VP.StM1-2.1. This variant differed from the homologous St. Maries microdomain by a single codon insertion, encoding a third asparagine (Fig. 4 and Table 2). Indels were present in the overwhelming majority of the antigenic variants with either complete or partial loss of binding (Fig. 3 and 4 and Tables 1 and 2), with or without associated amino acid substitutions. However, amino acid substitutions alone were sufficient to create an antigenic variant, as illustrated by variant VP.E6F7-2.2 (Fig. 4 and Table 2). Detailed enzyme-linked immunosorbent assay (ELISA) data are provided in Fig. S1 in the supplemental material. DISCUSSION
The Msp2 variant repertoire encoded only by recombination of the HVRs represented by the “whole” allele is defined by the number of alleles in the genome (5, 20). This allelic repertoire is limited among A. marginale strains and is ⬍10 per genome; the prototype St. Maries strain has only five unique msp2 alleles. Segmental gene conversion uses combinatorial recombination to generate hundreds to thousands of novel sequence variants not represented elsewhere in the genome other than transiently in the expression site (9). For this expansion to be functional in pathogen persistence requires that they evade the extant antibody response (12, 21, 22). The fraction of structural variants that represent antigenic variants has been unknown. The current study supports the notion that the HVR, composed of highly conserved tethers flanking variable microdomains, allows both structural fitness and antigenic variation. Of the 38 expressed primary structure variant microdomains examined, 12 were unrecognized by high-titer antibody against the homologous microdomain across all biological replicates: antibody binding to the variant was diminished to the level of preimmune serum. An additional 22 variant microdomains had a statistically significant reduction in binding across all replicates. Collectively, 89% of the expressed structural variants were also antigenic variants across all replicates, independent of a specific MHC haplotype. The biological relevance of these findings is supported by the following four lines of evidence: (i) all
November 2015 Volume 83 Number 11
structural variants were expressed during infection of a natural host, (ii) the structural variation observed in the microdomains corresponded to the mean length of variants generated by segmental gene conversion in the St. Maries strain (29 ⫾ 13 nucleotides, based on examination of 1,183 variants [9]), (iii) antigenic variants were identified using a broad antibody response that developed during infection of a natural host, and (iv) the ablation of antibody binding to a variant microdomain is in agreement with the results demonstrating lack of binding to a newly emergent full-length HVR (22, 23). The significant loss of antibody binding to a structural variant with only a single amino acid insertion demonstrates how minimal change can result in a new antigenic variant. While it is well established using monoclonal antibodies that a single amino acid change in an antigen is sufficient to eliminate binding, the significance in the present study is that the single insertion must retain minimal protein structure for bacterial viability and evade a polyclonal response. Interestingly, microdomain 2 variant VP.StM12.1 differed from the homologous microdomain by a single codon insertion that resulted in an additional asparagine. Whether this reflects a preferred genetic mechanism of codon duplication, a structural selection to retain function, or a combination of the two is unknown. Our lack of understanding of which changes allow retention of protein function while generating an antigenic variant represents a significant gap in knowledge and will require additional structure-function studies. We examined sera from animals with four unique MHC haplotypes. Fifteen structural variants, including 12 for which complete ablation of binding was observed, had consistent binding patterns across all replicates, regardless of the MHC haplotype. The remaining structural variants represented antigenic variants with complete ablation of binding with serum from at least one animal but either a significant reduction in binding, but not complete ablation (Table 2), or no loss of binding (Table 3) using sera from other animals, representing independent biological replicates. There was no consistent pattern of binding, or lack thereof, across MHC haplotypes. This very likely reflects sequential selection for variants, where multiple new variants are generated by segmental gene conversion, structurally fit variants are viable and initiate replication, and then the selective pressure of the immune response allows only those antigenic variants to continue replication and form a bacteremic peak. Which variants compose the bacteremic peak depends on the extant antibody response of the infected animal, which in turn is dependent on both the MHC and prior variant exposure. The fact that a very high proportion of structural variants represented true antigenic variants supports the efficiency of segmental gene conversion in generating Msp2 antigenic variants. Antigenic variation allows A. marginale persistence within the individual infected animal, allowing vector-borne transmission (22, 24, 25). Importantly, Msp2 antigenic variation has also been shown to drive strain superinfection at the host population level (10, 12, 26, 27). Strains with different msp2 allelic repertoires are capable of infecting hosts previously infected with a primary strain that had mounted an immune response against it. Even a single unique msp2 allele is sufficient to allow strain superinfection, with the requirement that the unique allele be expressed at the time of superinfection (10). Within regions of endemicity with high population immunity against a dominant strain, there is strong selective pressure for emergence of new strains, and consequently,
Infection and Immunity
iai.asm.org
4183
Graça et al.
strain superinfection is common (12, 21, 24, 27). While there is genomic evidence that a new strain may emerge by insertions or deletions within a previously duplicated msp2 allele, the magnitude of change in the Msp2 primary structure required to generate a truly antigenically unique strain capable of superinfection is unknown. The current study, which uses expressed variants from the Nayarit and St. Maries strains, supports the idea that very minor structural changes, especially insertions/deletions, are sufficient to create an escape variant. This is consistent with the results of two prior studies. The first is the comparative genomics approach that identified otherwise identical alleles in the St. Maries and Florida strains (alleles 1 and NAI1011, respectively) that differed by two indels and amino acid substitutions in microdomain 2 (11), variation within the range observed for antigenic variants in the current study. The second study mapped 1,183 expressed variants to the encoding alleles (9). While 99% of the variants mapped identically to the alleles, 1% of the variants contained unique sequences that differed by small insertions or deletions similar to those identified here as antigenic variants (9). Collectively, these studies suggest that relatively small genetic changes, predominantly insertions and deletions, can generate antigenic variants and, when maintained in the genome as a novel allele, would provide the basis for emergence of a new strain with selection for the capacity to superinfect.
11.
12.
13.
14.
15. 16.
17.
ACKNOWLEDGMENTS Research was supported by National Institutes of Health grant R37 AI44005. Telmo Graça was partially supported by a scholarship through Fundação para a Ciência e Tecnologia SFRH/BD/68377/2010. We appreciate the excellent technical support of Beverly Hunter.
18. 19.
REFERENCES 1. Holland J, Spindler K, Horodyski F, Grabau E, Nichol S, Vandepol S. 1982. Rapid evolution of RNA genomes. Science 215:1577–1585. http://dx .doi.org/10.1126/science.7041255. 2. Deitsch KW, Lukehart SA, Stringer JR. 2009. Common strategies for antigenic variation by bacterial, fungal and protozoan pathogens. Nat Rev Microbiol 7:493–503. http://dx.doi.org/10.1038/nrmicro2145. 3. Palmer GH, Bankhead T, Lukehart SA. 2009. ‘Nothing is permanent but change’—antigenic variation in persistent bacterial pathogens. Cell Microbiol 11:1697–1705. http://dx.doi.org/10.1111/j.1462-5822 .2009.01366.x. 4. Barbour AG, Restrepo BI. 2000. Antigenic variation in vector-borne pathogens. Emerg Infect Dis 6:449 – 457. http://dx.doi.org/10.3201/eid0605 .000502. 5. Brayton KA, Kappmeyer LS, Herndon DR, Dark MJ, Tibbals DL, Palmer GH, McGuire TC, Knowles DP, Jr. 2005. Complete genome sequencing of Anaplasma marginale reveals that the surface is skewed to two superfamilies of outer membrane proteins. Proc Natl Acad Sci U S A 102:844 – 849. http://dx.doi.org/10.1073/pnas.0406656102. 6. Palmer GH, Brayton KA. 2007. Gene conversion is a convergent strategy for pathogen antigenic variation. Trends Parasitol 23:408 – 413. http://dx .doi.org/10.1016/j.pt.2007.07.008. 7. Santoyo G, Romero D. 2005. Gene conversion and concerted evolution in bacterial genomes. FEMS Microbiol Rev 29:169 –183. http://dx.doi.org /10.1016/j.fmrre.2004.10.004. 8. Centurion-Lara A, LaFond RE, Hevner K, Godornes C, Molini BJ, Van Voorhis WC, Lukehart SA. 2004. Gene conversion: a mechanism for generation of heterogeneity in the tprK gene of Treponema pallidum during infection. Mol Microbiol 52:1579 –1596. http://dx.doi.org/10.1111/j .1365-2958.2004.04086.x. 9. Futse JE, Brayton KA, Knowles DP, Jr, Palmer GH. 2005. Structural basis for segmental gene conversion in generation of Anaplasma marginale outer membrane protein variants. Mol Microbiol 57:212–221. http://dx .doi.org/10.1111/j.1365-2958.2005.04670.x. 10. Futse JE, Brayton KA, Dark MJ, Knowles DP, Jr, Palmer GH. 2008.
4184
iai.asm.org
20.
21.
22. 23.
24. 25.
26.
27.
Superinfection as a driver of genomic diversification in antigenically variant pathogens. Proc Natl Acad Sci U S A 105:2123–2127. http://dx.doi.org /10.1073/pnas.0710333105. Dark MJ, Herndon DR, Kappmeyer LS, Gonzales MP, Nordeen E, Palmer GH, Knowles DP, Brayton KA. 2009. Conservation in the face of diversity: multistrain analysis of an intracellular bacterium. BMC Genomics 10:16. http://dx.doi.org/10.1186/1471-2164-10-16. Ueti MW, Tan Y, Broschat SL, Castañeda Ortiz EJ, Camacho-Nuez M, Mosqueda JJ, Scoles GA, Grimes M, Brayton KA, Palmer GH. 2012. Expansion of variant diversity associated with a high prevalence of pathogen strain superinfection under conditions of natural transmission. Infect Immun 80:2354 –2360. http://dx.doi.org/10.1128/IAI.00341-12. Zhuang Y, Futse JE, Brown WC, Brayton KA, Palmer GH. 2007. Maintenance of antibody to pathogen epitopes generated by segmental gene conversion is highly dynamic during long-term persistent infection. Infect Immun 75:5185–5190. http://dx.doi.org/10.1128/IAI.00913-07. Abbott JR, Palmer GH, Howard CJ, Hope JC, Brown WC. 2004. Anaplasma marginale major surface protein 2 CD4⫹-T-cell epitopes are evenly distributed in conserved and hypervariable regions (HVR), whereas linear B-cell epitopes are predominantly located in the HVR. Infect Immun 72: 7360 –7366. http://dx.doi.org/10.1128/IAI.72.12.7360-7366.2004. Wu CJ, Kalyanaraman A, Cannon WR. 2012. pGraph: efficient parallel construction of large-scale protein sequence homology graphs. IEEE Trans Parallel Distrib Syst 23:1923–1933. http://dx.doi.org/10.1109/TPDS.2012.19. Wu CJ, Kalyanaraman A. 2008. An efficient parallel approach for identifying protein families in large-scale metagenomic data sets, p 362–371. Proc ACM/IEEE Conf Supercomputing (SC’08), Austin, TX, November 15–21, 2008. IEEE Press, Piscataway, NJ. Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R, Thompson JD, Gibson TJ, Higgins DG. 2007. Clustal W and clustal X version 2.0. Bioinformatics 23:2947–2948. http://dx.doi.org/10.1093/bioinformatics/btm404. Miltiadou D, Law AS, Russell GC. 2003. Establishment of a sequencebased typing system for BoLA-DR3 exon 2. Tissue Antigens 62:55– 65. http://dx.doi.org/10.1034/j.1399-0039.2003.00080.x. Takeshima SN, Matsumoto Y, Miyasaka T, Arainga-Ramirez M, Saito H, Onuma M, Aida Y. 2011. A new method for typing bovine major histocompatibility complex class II DR3 alleles by combining two established PCR sequence-based techniques. Tissue Antigens 78:208 –213. http: //dx.doi.org/10.1111/j.1399-0039.2011.01708.x. Brayton KA, Knowles DP, McGuire TC, Palmer GH. 2001. Efficient use of a small genome to generate antigenic diversity in tick-borne ehrlichial pathogens. Proc Natl Acad Sci U S A 98:4130 – 4135. http://dx.doi.org/10 .1073/pnas.071056298. Vallejo Esquerra E, Herndon DR, Alpirez Mendoza F, Mosqueda J, Palmer GH. 2014. Anaplasma marginale superinfection attributable to pathogen strains with distinct genomic backgrounds. Infect Immun 82: 5286 –5292. http://dx.doi.org/10.1128/IAI.02537-14. French DM, Brown WC, Palmer GH. 1999. Emergence of Anaplasma marginale antigenic variants during persistent rickettsemia. Infect Immun 67:5834 –5840. Agnes JT, Brayton KA, LaFollett M, Norimine J, Brown WC, Palmer GH. 2011. Identification of Anaplasma marginale outer membrane protein antigens conserved between A. marginale sensu stricto strains and the live A. marginale subsp. centrale vaccine. Infect Immun 79:1311–1318. http://dx.doi.org/10.1128/IAI.01174-10. Palmer GH, Brayton KA. 2013. Antigenic variation and transmission fitness as drivers of bacterial strain structure. Cell Microbiol 15:1969 – 1975. http://dx.doi.org/10.1111/cmi.12182. Scoles GA, Broce AB, Lysyk TJ, Palmer GH. 2005. Relative efficiency of biological transmission of Anaplasma marginale (Rickettsiales: Anaplasmataceae) by Dermacentor andersoni (Acari: Ixodidae) compared with mechanical transmission by Stomoxys calcitrans (Diptera: Muscidae). J Med Entomol 42:668 – 675. Rodriguez JL, Palmer GH, Knowles DP, Brayton KA. 2005. Distinctly different msp2 pseudogene repertoires in Anaplasma marginale strains that are capable of superinfection. Gene 361:127–132. http://dx.doi.org /10.1016/j.gene.2005.06.038. Castaneda-Ortiz EJ, Ueti MW, Camacho-Nuez M, Mosqueda JJ, Mousel MR, Johnson WC, Palmer GH. 2015. Association of Anaplasma marginale strain superinfection with infection prevalence within tropical regions. PLoS One 10:e0120748. http://dx.doi.org/10.1371/journal.pone .0120748.
Infection and Immunity
November 2015 Volume 83 Number 11
