Supplementary Material

Supplementary Material 2 4 6 Single-locus gene screening methods We screened for single nucleotide polymorphisms (SNPs) at 16 candidate immune genes (Table S1) by initially sequencing three captive-bred Attwater s prairie-chickens and one greater prairie-chicken from either the genetically diverse Kansas or Nebraska populations. If we found polymorphism among these sequences, we randomly selected an additional 16 Attwater s prairie-chickens to 8 genotype and identified exons by aligning each sequence with the domestic chicken (Gallus gallus) genome using the BLAST algorithm in Geneious version 7.0. If there were non-synonymous SNPs that resulted in different alleles (n = 20 individuals), we 10 genotyped the remaining Attwater s prairie-chickens used in this study (n = 125) In total, we genotyped Attwater s prairie-chickens at seven non-mhc genes, however, three of these (TLR4, TLR5 and AvBD11) had only two alleles and were included only in the 12 immune response and survival models where standardized heterozygosity and heterozygosity at individual loci were used as predictor variables (Tables S4 and S6). All immune gene sequences that were found in at least two individuals or two independent PCRs are 14 deposited in Genbank (accession nos. KX239674-KX239750). 16

18 20 22 Table S1. Primers and PCR conditions used to amplify the seven single-locus immune genes used in the study (A) and the nine other immune genes not used (B) because they did not contain non-synonymous SNPs. Each immune gene was amplified in PCRs (final volume 20ul) consisting of 1x Green GoTaq Flexi Buffer (Promega), 1.5 mm of MgCl 2, 10% DMSO, 0.5 µm of each primer, 0.4 mm of dntps, 1.0 U of GoTaq DNA polymerase and approximately 50 ng of genomic DNA. PCRs were performed with an initial denaturation step at 94 C for 2 min followed by 35 cycles of 20 s at 94 C, 20 s at locus-specific annealing temperature (T A ) and extension time (T E ) at 72 C. 24 (A) 26 28 T A T E Locus ( C) (sec) Primer names Primer sequences (5'-3') References Toll-like receptor 1B (TLR1B) 59 80 avtlr1lbf TCCAGGYTWCAAAATCTGACAC 1 Alcaide and Edwards 2010 avtlr1lbr CGGCACRTCCARGTAGATG Toll-like receptor 4 (TLR4) 50 80 avtlr4f GAGACCTTGATGCCCTGAG Alcaide and Edwards 2010 avtlr4r CCATCTTRAGCACTTGCAAAG Toll-like receptor 5 (TLR5) 55 80 avtlr5f GTAATCTTACCAGCTTCCAAGG Alcaide and Edwards 2010 avtlr5r GCTGGAGTTCATCTTCATC Toll-like receptor (TLR15) 62 45 TLR15F1 GCTGGGTGCTGTTTTGGAGT This study TLR15R1 GAGGTGCTGCAGAGAGATCG C-type lectin-like receptor 62 45 Blec1F AGCTCCCACGTTTCTCATCC This study (Blec1) Blec1R CCAAGGCAAGGATGGGAACT Interleukin 4 (IL4) 54 45 IL4ex1-2F ATGAGCTCCTCACTGCCCAC This study IL4ex1-2R CTGGCTTTCCTCTTACCTTA β-defensin 11 (AvBD11) 60 45 AvBD11F1mat GACTGATCCTGCAGCACAAC 2 Hellgren and Sheldon 2011 AvBD11R1 AGGGCTCCCACACGTACC 1 Alcaide M, Edwards SV. 2011. Molecular evolution of the toll-like receptor multigene family in birds. Molecular Biology and Evolution 28:1703-1715. 2 Hellgren O, Sheldon BC. 2011. Locus-specific protocol for nine different innate immune genes (antimicrobial peptides: β-defensins) across passerine birds species reveals within-species coding variation and a case of trans-species polymorphism. Molecular Ecology Resources 11:686-692.

30 (B) Locus T A ( C) T E (sec) Primer names Primer sequences (5'-3') References Toll-like receptor 3 (TLR3) 53 80 pctlr3falt ACCTCTCACTGAGCCATGTG This study pctlr3r TGTTGTTATTGCTGATGTC Toll-like receptor 7 (TLR7) 59 60 pctlr7f2 GAAGCTTATCCCCAGTCTTG This study pctlr7r3 AGACAGGTAGCAGAATTCGC Interleukin 1B (IL1B) 59 60 IL1Bex5-6F CTTCGACATCTTCGACATCAAC 1 Downing, et al. 2009 IL1Bex5-6R ATACGAGATGGAAACCAGCAAC Interleukin 6 (IL6.1) 63 45 IL6ex2-3F CGAGAACAGCATGGAGATGC This study IL6ex2-3R GTGGCCGCCAGGTGCTTTGT Interleukin 6 (IL6.2) 57 45 IL6ex4F GTGATAAATCCCGATGAAGT This study IL6ex4R TCAGGCACTGAAACTCCTGG Interleukin 10 (IL10) 62 45 IL-10_F1 CCACTGCTGGGGTTCAGATT This study IL-10_R1 CTCTCTCCCCACCCACTGAA β-defensin 02 (AvBD02) 55 45 AvBD2ex2F ATGAGGATTCTTTACCTGC This study AvBD2ex3R CATTTGCAGCAGGAACGGAA β-defensin 04 (AvBD04) 56 45 pcavbd4ex3f TGTTCAGGCTTTCCCCGTCC This study pcavbd4ralt TCAGTTTAGCCATAGTCAAG Lysozyme (LYZ) 60 45 LYSOex1_Falt GAGACAGGTGCAAGAGAGCC This study LYSOex1_Ralt GGGAAAGGAGCGTAAAGGGA 1 Downing T, Lynn DJ, Connell S, Lloyd AT, et al. 2009. Contrasting evolution of diversity at two disease-associated chicken genes. Immunogenetics 61:303-314.

32 34 Table S2. Description of the sequences that did not contain non-synonymous SNPs, obtained from the nine single-locus immune genes in the initial screening of prairie-chickens (n = 20). Included are the number of birds (n), particular exons sequences and the corresponding exon length in base pairs (intron length in parentheses if amplified), synonymous (Syn) and intron SNPs. Immune gene n Exon(s) Sequence length Syn SNPs Toll-like receptor 3 (TLR3) 20 4 773 1 - Toll-like receptor 7 (TLR7) 20 2 609 1 - Interleukin 1B (IL1B) 20 4,5,6 422 (185) 3 4 Interleukin 6 (IL6.1) 4 2,3 161 (121) 0 0 Interleukin 6 (IL6.2) 20 4 144 1 - Interleukin 10 (IL10) 20 3,4 219 (509) 1 2 β-defensin 02 (AvBD02) 20 2,3 180 (231) 1 3 β-defensin 04 (AvBD04) 4 2 102 0 - Lysozyme (LYZ) 20 1 136 (188) 2 3 Intron SNPs 36 38 40 42 44

46 48 50 Table S3. Models for post-release survival of Attwater s prairie-chickens using higher thresholds for the number of reads (compare with Table 2). Survival model A includes individuals (n = 93) with MHC class I genotypes based on 30 pyrosequencing reads, whereas survival model B includes individuals (n = 100) with class II genotypes based on 300 pyrosequencing. Predictor variables are specific alleles at immune genes, genome-wide heterozygosity, body mass (g), age class and sex. Family ID was also included in the models as a random factor. Presented are the parameter estimates (β), standard error (SE), hazard ratio (Exp(β)), P values from the Cox proportional mixed model and q values (the probability (P) values corrected for the false discovery rate). Predictors with P and q values 0.07 are in bold. Model A Model B β SE Exp(β) P q β SE Exp(β) P q Genome-wide het 0.123 1.274 1.130 0.92 0.63 0.070 0.989 1.073 0.94 0.75 MHC class I Tycu-IA*20 0.095 0.588 1.100 0.87 0.63-0.397 0.554 0.672 0.47 0.61 Tycu-IA*22-0.789 0.355 0.454 0.03 0.06-0.820 0.342 0.441 0.02 0.09 Tycu-IA*24-0.774 0.356 0.461 0.03 0.06-0.754 0.335 0.470 0.02 0.10 MHC class II Tycu-BLB*08-1.312 0.413 0.269 <0.01 0.01-1.278 0.430 0.279 <0.01 0.03 Blec1 Blec1*01 0.110 0.373 1.116 0.77 0.63-0.014 0.342 0.986 0.97 0.75 Blec1*02 0.250 0.331 1.284 0.45 0.48 0.118 0.378 1.126 0.75 0.73 Blec1*04-1.203 0.388 0.300 <0.01 0.01-1.168 0.367 0.311 <0.01 0.03 TLR1B TLR1B*01 0.471 0.392 1.602 0.23 0.29 0.442 0.362 1.556 0.22 0.37 TLR1B*02 0.295 0.456 1.343 0.52 0.49-0.021 0.445 0.980 0.96 0.75 TLR1B*03 0.829 0.351 2.291 0.02 0.06 0.587 0.326 1.798 0.07 0.20 TLR1B*04 0.582 0.390 1.790 0.14 0.19 0.123 0.366 1.131 0.74 0.73 TLR15 TLR15*01-0.063 0.745 0.939 0.93 0.63 0.296 0.672 1.344 0.66 0.73 TLR15*02 0.436 0.409 1.546 0.29 0.33 0.387 0.420 1.473 0.36 0.51 TLR15*03-0.113 0.479 0.894 0.81 0.63 0.076 0.447 1.079 0.87 0.75 TLR15*04 1.024 0.518 2.785 0.05 0.09 0.548 0.524 1.729 0.30 0.46

IL-4 IL4*01-0.029 0.333 0.971 0.93 0.63 0.084 0.287 1.088 0.77 0.73 IL4*02-0.973 0.630 0.378 0.12 0.18-0.910 0.546 0.403 0.10 0.20 IL4*03 0.347 0.491 1.415 0.48 0.48-0.175 0.494 0.839 0.72 0.73 52 Age class 1.180 0.533 3.255 0.03 0.06 0.831 0.494 2.295 0.09 0.20 Body mass -0.006 0.002 0.994 <0.01 0.01-0.004 0.002 0.996 0.03 0.11 Sex 0.601 0.319 1.824 0.06 0.10 0.474 0.300 1.606 0.11 0.21 54

56 58 60 Table S4. Post-release survival of Attwater s prairie-chickens (n = 116) in relation to standardized heterozygosity (A) and single-locus heterozygosity (B) at immune genes, genomewide heterozygosity, body mass (g), age class and sex. In both analyses, family group was used as a random factor. Presented are the parameter estimates (β), standard error (SE), hazard ratio (Exp(β)) and corresponding P and q (P values corrected for the false discovery rate) values from the Cox proportional mixed models. Predictors with P and q values < 0.05 are given in bold. (A) 62 Predictor β SE Exp(β) P q Genome-wide het 0.383 0.876 1.466 0.66 0.73 Immune gene het 0.115 0.338 1.121 0.73 0.73 Age class 0.979 0.375 2.661 0.01 0.02 Body mass -0.003 0.001 0.996 0.02 0.03 Sex 0.644 0.232 1.904 0.01 0.02 (B) 64 Predictor β SE Exp(β) P q Genome-wide het 0.446 0.857 1.562 0.60 0.90 MHC class I 0.016 0.257 1.016 0.95 0.95 Blec1-0.097 0.257 0.908 0.71 0.95 TLR1B 0.379 0.218 1.462 0.08 0.24 TLR4 0.330 0.469 1.391 0.48 0.90 TLR5-0.044 0.434 0.957 0.92 0.95 TLR15-0.142 0.244 0.868 0.56 0.90 IL4-0.028 0.217 0.972 0.90 0.95 AvBD11-0.583 0.544 0.558 0.28 0.67 Age class 1.085 0.391 2.959 <0.01 0.06 Body mass -0.003 0.002 0.997 0.03 0.10 Sex 0.649 0.252 1.913 0.01 0.06 66 68 70

72 74 76 78 Table S5. Immune response of Attwater s prairie-chickens in relation to specific alleles at immune genes, genome-wide heterozygosity, body mass (g), age class and sex. Family group was used as a random factor. Presented are the estimates (β), standard error (SE), P and q (P values corrected for the false discovery rate) values for each predictor variable in the models for Bacteria killing assay (n = 116) and Agglutination assay (n = 107). Note that no predictors had significant q values. Bacteria killing Agglutination Predictor β SE P q β SE P q Genome-wide het -0.433 0.358 0.23 0.94 2.257 1.131 0.05 0.55 MHC class I Tycu-IA*20-0.087 0.176 0.62 0.94 0.291 0.569 0.61 0.85 Tycu-IA*22 0.023 0.112 0.84 0.94 0.235 0.380 0.54 0.85 Tycu-IA*24-0.170 0.125 0.18 0.94 0.063 0.430 0.88 0.96 MHC class II Tycu-BLB*08-0.095 0.138 0.49 0.94 0.455 0.465 0.33 0.77 Blec1 Blec1*01 0.237 0.130 0.07 0.94 0.046 0.452 0.92 0.96 Blec1*02-0.039 0.121 0.75 0.94 0.170 0.402 0.67 0.87 Blec1*04-0.073 0.117 0.53 0.94 0.278 0.370 0.45 0.83 TLR1B TLR1B*01-0.043 0.131 0.74 0.94-0.239 0.428 0.58 0.85 TLR1B*02 0.181 0.148 0.23 0.94 0.777 0.485 0.11 0.62 TLR1B*03-0.014 0.109 0.90 0.94 0.045 0.361 0.90 0.96 TLR1B*04-0.030 0.128 0.82 0.94 0.011 0.435 0.98 0.98 TLR15 TLR15*01-0.080 0.228 0.73 0.94 0.195 0.783 0.80 0.96 TLR15*02 0.048 0.152 0.75 0.94-0.586 0.496 0.24 0.77 TLR15*03 0.049 0.147 0.74 0.94-0.682 0.471 0.15 0.67 TLR15*04 0.049 0.149 0.74 0.94 0.414 0.475 0.39 0.77 IL-4 IL4*01-0.035 0.098 0.72 0.94-0.531 0.327 0.11 0.62 IL4*02 0.038 0.208 0.85 0.94-0.351 0.701 0.62 0.85 IL4*03-0.101 0.190 0.60 0.94-0.569 0.600 0.35 0.77 Age class 0.081 0.078 0.30 0.94-0.634 0.256 0.02 0.34 Body mass 0.001 0.001 0.30 0.94-0.002 0.002 0.36 0.77 Sex -0.006 0.047 0.89 0.94 0.161 0.156 0.31 0.77

80 82 84 Table S6. Immune response of pre-released Attwater s prairie-chickens in relation to standardized heterozygosity (A) and single-locus heterozygosity (B) at eight immune genes, genome-wide heterozygosity, body mass (g), age class and sex. In these models, family group was included as a random factor. Sample sizes (n), parameter estimates (β), standard error (SE), P and q (P values corrected for the false discovery rate) values are given from the generalized linear mixed models. Single-locus heterozygosity models (B) also include false discovery rate q values for each variable. Significant q values are given in bold. A) Immunoassay Genome-wide het Immune genes het Age class Body mass Sex Lysozyme β 0.512-0.062 0.053 <0.001-0.113 n = 127 SE 0.205 0.089 0.054 <0.001 0.033 P 0.02 0.49 0.33 0.87 <0.01 q 0.04 0.61 0.54 0.87 0.01** Haemolysis β -1.421 0.268-0.276 0.001-0.197 n = 129 SE 0.935 0.336 0.189 0.001 0.124 P 0.13 0.43 0.15 0.66 0.11 q 0.26 0.56 0.26 0.69 0.26 Bacteria-killing β -0.566-0.056 0.026 < 0.001-0.015 n = 128 SE 0.304 0.110 0.062 < 0.001 0.040 P 0.07 0.61 0.67 0.54 0.71 q 0.34 0.74 0.74 0.74 0.74 Agglutination β 1.565-0.244-0.604-0.002 0.023 n = 118 SE 0.989 0.363 0.210 0.001 0.137 P 0.12 0.50 <0.01 0.16 0.87 q 0.27 0.63 0.02* 0.27 0.87 * indicates that second year after hatch birds had increased agglutination activity; ** indicates that males had stronger lysozyme activity 86

2 Table 6 (continued) 4 B) Immunoassay Genomewide het MHCI Blec1 TLR1B TLR4 TLR5 TLR15 IL4 AvBD11 Age class Body mass Sex Lysozyme β 0.451 0.059 0.020 0.026-0.045-0.066-0.021 0.035-0.122 0.079 <0.001-0.128 n = 114 SE 0.234 0.051 0.043 0.038 0.077 0.079 0.040 0.039 0.105 0.061 <0.001 0.039 P 0.06 0.25 0.63 0.49 0.55 0.41 0.60 0.37 0.25 0.20 0.98 <0.01 q 0.35 0.61 0.69 0.69 0.69 0.69 0.69 0.69 0.61 0.61 0.98 0.02* Haemolysis β -0.786 0.131-0.008-0.093-0.281-0.020 0.063-0.076-0.040-0.340 0.001-0.015 n = 113 SE 0.682 0.116 0.111 0.093 0.198 0.200 0.101 0.095 0.233 0.151 0.001 0.100 P 0.25 0.26 0.94 0.32 0.16 0.92 0.53 0.42 0.87 0.03 0.46 0.88 q 0.77 0.77 0.94 0.77 0.77 0.94 0.80 0.79 0.94 0.32 0.79 0.94 Bacteria-killing β -0.438 0.059-0.089-0.018-0.069 0.013 0.005 0.038-0.078 0.025 <0.001-0.018 n = 115 SE 0.325 0.050 0.051 0.042 0.091 0.089 0.046 0.043 0.104 0.069 0.001 0.045 P 0.18 0.25 0.08 0.67 0.45 0.89 0.91 0.38 0.45 0.72 0.40 0.69 q 0.78 0.78 0.78 0.86 0.78 0.91 0.91 0.78 0.78 0.86 0.78 0.86 Agglutination β 2.372-0.136-0.145-0.018-0.147 0.251 0.111 0.234-0.434-0.589-0.001 0.050 n = 105 SE 1.084 0.169 0.171 0.141 0.316 0.315 0.153 0.141 0.338 0.232 0.002 0.150 P 0.03 0.42 0.40 0.90 0.64 0.43 0.47 0.10 0.20 0.01 0.41 0.74 q 0.19 0.63 0.63 0.90 0.77 0.63 0.63 0.41 0.61 0.15 0.63 0.81 *indicates that males had stronger lysozyme activity 6 8

92 94 Figure S1. Plot of standardized heterozygosity at 20 990 SNPs (genome-wide) and eight single-locus immune genes in Attwater s prairie-chickens released into the wild in 2011 and 2012. 96 Genome-wide heterozygosity Immune gene heterozygosity