Development and optimization of a hybridization technique to type the classical class I and class II B genes of the chicken MHC

Abstract

The classical class I and class II molecules of the major histocompatibility complex (MHC) play crucial roles in immune responses to infectious pathogens and vaccines as well as being important for autoimmunity, allergy, cancer and reproduction. These classical MHC genes are the most polymorphic known, with roughly 10,000 alleles in humans. In chickens, the MHC (also known as the BF-BL region) determines decisive resistance and susceptibility to infectious pathogens, but relatively few MHC alleles and haplotypes have been described in any detail. We describe a typing protocol for classical chicken class I (BF) and class II B (BLB) genes based on a hybridization method called reference strand-mediated conformational analysis (RSCA). We optimize the various steps, validate the analysis using well-characterized chicken MHC haplotypes, apply the system to type some experimental lines and discover a new chicken class I allele. This work establishes a basis for typing the MHC genes of chickens worldwide and provides an opportunity to correlate with microsatellite and with single nucleotide polymorphism (SNP) typing for approaches involving imputation.

Keywords: Avian; BF1; BF2; BLB1; BLB2; Recombination.

Fig. 1

A diagram representing the four steps of RSCA (amplification to make an FLR, amplification from an experimental sample, hybridization to create DNA duplexes and analysis by capillary electrophoresis) with an idealized trace for a two gene system from a a homozygote (with all possible DNA duplexes shown, including those without an FLR, which would not be detected) and b a heterozygote (with only the DNA duplexes that include an FLR for detection shown). A homozygote sample produces one homoduplex and two heteroduplex peaks in the trace due to the labelled FLR; unlabelled strands form heteroduplexes but are not detected by the machine. A heterozygote sample produces one homoduplex and up to four heteroduplex peaks for a two-locus gene family (BLB or BF), but duplexes between unlabelled DNA strands are not detected (not shown for heterozygote). This figure is based conceptually on Argüello et al 1995

Fig. 2

The genomic basis for simple RSCA typing of chicken classical class I and class II B genes. Top portion shows the BF-BL region, which contains only two BLB and two BF genes (based on Kaufman et al. 1999); middle portion shows intron-exon structures with the location of the primers (identified with local primer names) to amplify the exons encoding peptide-binding domains (Jacob et al. ; Shaw et al. 2007); bottom portion shows the RSCA trace for a heterozygote sample, with homo- and heteroduplexes indicated

Fig. 3

Work flow for the RSCA experiments described in this report. Each step of the flowchart (as well as several others not shown, some of which are discussed in the text) needed optimization in order to apply the typing method to unknown populations. Red text and boxes indicate the results of optimization steps

Fig. 4

For both class I and class II B gene sequences, the four most divergent sequences were chosen to make FLRs. Nucleotide sequence alignments of exon 2 to exon 3 from class I (BF1 and BF2) genes (left panel) and of exon 2 from class II B (BLB1 and BLB2) genes (right panel) were used to construct neighbor-joining (NJ) trees, which were viewed using dendroscope. Four of the most divergent sequences are highlighted by green circles for class I, purple circles for class II B, and labelled for the four FLR combinations

Fig. 5

Several steps of the RSCA procedure needed optimization of which five are shown. All panels are RSCA traces from GenMapper software, with red size standard peaks (ROX2500 except for Fig. 5a) and blue fluorescent peaks from the hybridization mix. a Comparison after hybridization with a purified (top panel) and an unpurified (bottom panel) FLR. In this experiment, a BF2*0401 FLR preparation (before and after purification by silica spin column) was hybridized with an experimental B12 homozygote sample (blue peaks) and subjected to capillary electrophoresis along with ROX500 size standards (red peaks). This experiment established that the FLR needed to be purified before use and also that ROX2500 size standards were needed to cover the class I peaks. b Comparison after loading with hybridization mixtures of class II B mixture to class I mixture v/v 1:4 (top panel) and 1:1 (bottom panel). FLRs from BLB1*1401 and BF2*0401 were hybridized to DNA from a B4/B12 heterozygote sample. Increasing the volume of class I mix compared with class II B mix allowed the class I peaks to be detected because the concentrations were based on mass rather than molarity. c Comparison after storage of the hybridization mix and assembled plate under different conditions. An FLR produced from BF1*0201 was hybridized with genomic DNA from a B4/B12 heterozygote sample followed by storage under different conditions: hybridization not stored but assembled plate stored at − 20 °C for 5 days (top panel), hybridization stored at 4 °C for 10 days and assembled plate stored overnight at 4 °C (panel second from top), hybridization stored at 4 °C for 10 days and assembled plate stored at 4 °C for 5 days (panel third from top) and assembled plate stored at 4 °C overnight (bottom panel). This experiment shows that long-term storage of the assembled plate at 4 °C is not optimal. d Comparison of hybridization mixes assembled in the plate with different dilutions. FLRs from BLB1*0201 and BF1*0201 were hybridized to DNA from a B12 homozygote sample and assembled to give 12 μl in a well (top panel) and with a B19 homozygote sample and assembled with added water (to prevent too much evaporation during the run) giving 15 μl in a well (bottom panel). The comparable heights of the peaks show that the added water did not detract from the quality of the analysis. e Comparison of four identical samples overlaid to show reproducibility within a run. FLRs from BLB2*0401 and BF2*0401 were hybridized to DNA amplified from a B2 homozygote sample

Fig. 6

Examples of RSCA traces from hybridization of genomic DNA from homozygous samples of haplotype B2, B4, B12, B14, B15, B19 and B21 (from top to bottom) with a FLR combination A (BLB1*1501 and BF2*1501), b FLR combination B (BLB1*0201 and BF1*0201), c FLR combination C (BLB2*0401 and BF2*0401) and d FLR combination D (BLB2*0201 and BF1*0401). Peaks of size standard are red, while peaks of fluorescence from FLRs are blue, with migration scores of peaks determined with reference to the size standards by the GeneMapper software. Homoduplexes are highlighted, purple for class II B peaks and green for class I peaks

Fig. 7

Chart of migration scores for peaks from samples of haplotype B2, B4, B12, B14, B15, B19 and B21 (from bottom to top) with a FLR combination A (BLB1*1501 and BF2*1501), b FLR combination B (BLB1*0201 and BF1*0201), c FLR combination C (BLB2*0401 and BF2*0401) and d FLR combination D (BLB2*0201 and BF1*0401) based on data in Fig. 6a–d. The migration scores are indicated by lines that are purple for homoduplex peaks and pink for experimental peaks from class II FLRs and dark green for homoduplex peaks and light green for experimental peaks from class I FLRs

Fig. 8

Inferred location of the third BLB gene by comparison of B12 and B19 haplotypes for the BG, TRIM and BF-BL regions, showing the recombination between the B12 and B15 haplotypes to give the B19 haplotype. Black horizontal line indicates sequence identical (or nearly so) with the B12 haplotype, red horizontal line indicates sequence nearly identical to the B15 haplotype, vertical arrow indicates recombination event between BF1 and BF2 and dashed line indicates the likely location of the B12c gene given the data in this report

Fig. 9

Chart of migration scores for peaks from samples of French experimental lines corresponding to haplotypes Sr1, Sr2, Sr3, Sr4, Sr5, Sr6 and Sr7 (from bottom to top) with a FLR combination A (BLB1*1501 and BF2*1501), b FLR combination B (BLB1*0201 and BF1*0201), c FLR combination C (BLB2*0401 and BF2*0401) and d FLR combination D (BLB2*0201 and BF1*0401). The migration scores are indicated by lines that are purple for homoduplex peaks and pink for experimental peaks from class II FLRs, and dark green for homoduplex peaks and light green for experimental peaks from class I FLRs

Fig. 10

The BF alleles from the French experimental lines are identical with alleles of standard haplotypes (at least in the region amplified by the primers), except for one previously unreported BF2 allele from the Sr1 haplotype. a neighbor-joining tree of nucleotide sequences (as amplified by the primers, but without the primer sequences or the introns) from standard haplotypes [(named as per current nomenclature (Miller et al. 2004)], followed by an accession number for a representative sequence, and from the haplotypes from the French experimental lines [names have a number from the MAFFT alignment, followed by the Sr haplotype number, followed by the identification using an older nomenclature (Wallny et al. 2006)]. Sequences are identical when a vertical line links them. Red arrow indicates the new BF2 allele from the recombinant Sr1 haplotype, which is most closely related to BF2*1801 and far away from BF2*2401. b and c Nucleotide and amino acid sequence alignments of the new class I allele of the Sr1 haplotype with BF2*1801, BF2*0401 and BF2*2401 [all as amplified by the primers, but without the primer sequences or the introns, (Hosomichi et al. 2008)], with long stretches of identical sequence highlighted in grey and with the beginning of exon 3 highlighted in yellow. The BF2 sequence from Sr1 differs from BF2*2401 throughout the amplified sequence but is identical with the first 217 nucleotides (72 amino acids) of the amplified sequence of BF2*1801 from exon 2 and with the last 119 nucleotides (48 amino acids) of the amplified sequence of BF2*0401 from exon 3. The new BF2 sequence from the Sr1 haplotype was deposited in GenBank with accession number MN103189