Molecular Ecology (2011) 20, 4695–4706 doi: 10.1111/j.1365-294X.2011.05319.x

Drift and selection influence geographic variationat immune loci of prairie-chickens

JENNIFER L. BOLLMER,* ELIZABETH A. RUDER,* JEFF A. JOHNSON,† JOHN A. EIMES* and

PETER O. DUNN*

*Department of Biological Sciences, University of Wisconsin-Milwaukee, P.O. Box 413, Milwaukee, WI 53201, †Department of

Biological Sciences, Institute of Applied Sciences, University of North Texas, 1155 Union Circle, #310559, Denton, TX 76203, USA

Corresponde

E-mail: bollm

� 2011 Black

Abstract

Previous studies of immunity in wild populations have focused primarily on genes of the

major histocompatibility complex (MHC); however, studies of model species have

identified additional immune-related genes that also affect fitness. In this study, we

sequenced five non-MHC immune genes in six greater prairie-chicken (Tympanuchuscupido) populations that have experienced varying degrees of genetic drift as a

consequence of population bottlenecks and fragmentation. We compared patterns of

geographic variation at the immune genes with six neutral microsatellite markers to

investigate the relative effects of selection and genetic drift. Global FST outlier tests

identified positive selection on just one of five immune genes (IAP-1) in one population.

In contrast, at other immune genes, standardized G¢ST values were lower than those at

microsatellites for a majority of pairwise population comparisons, consistent with

balancing selection or with species-wide positive or purifying selection resulting in

similar haplotype frequencies across populations. The effects of genetic drift were also

evident as summary statistics (e.g., Tajima’s D) did not differ from neutrality for the

majority of cases, and immune gene diversity (number of haplotypes per gene) was

correlated positively with population size. In summary, we found that both genetic drift

and selection shaped variation at the five immune genes, and the strength and type of

selection varied among genes. Our results caution that neutral forces, such as drift, can

make it difficult to detect current selection on genes.

Keywords: genetic drift, grouse, immunity, selection, single-nucleotide polymorphism, Tym-

panuchus cupido

Received 11 May 2011; revision received 8 September 2011; accepted 15 September 2011

Introduction

Selection on hosts from parasites is expected to vary in

time and space leading to geographic variation in allele

frequencies at immune genes. Understanding the geo-

graphic patterns of variation at immune genes is

important for identifying local adaptation (Dionne

et al. 2007), as well as for making management deci-

sions regarding threatened populations (Allendorf et al.

2010). The vertebrate immune system can be classified

into innate and adaptive responses, with the innate

nce: J. Bollmer, Fax: 414-229-3926;

er@uwm.edu

well Publishing Ltd

response producing an immediate but generalized

defence against pathogens and the adaptive response

producing specialized cells that recognize and remem-

ber antigens from a particular parasite. Most studies of

selection on vertebrate immune genes have focused on

class I and II genes of the major histocompatibility

complex (MHC), which are involved in the adaptive

immune response (Piertney & Oliver 2006). In some

cases, MHC genes exhibited less differentiation across

populations than neutral markers, consistent with bal-

ancing selection acting to maintain variation across

populations (Sommer 2003; Fraser et al. 2010). In other

studies, MHC differentiation was greater than neutral,

presumably due to selection favouring different alleles

Fig. 1 Historic (dashed line) and contemporary distribution of

greater (GPC) and Attwater’s (APC) prairie-chicken. Our Texas

samples are from the APC population. KS, Kansas; MN, Min-

nesota; MO, Missouri; NE, Nebraska; WI, Wisconsin.

4696 J . L . BOLLMER ET AL.

between populations (Ekblom et al. 2007; Cammen

et al. 2011). Lastly, a number of studies have

also shown that MHC population structure did not dif-

fer from that of neutral genes, suggesting that the

MHC was influenced more by drift and migration than

selection (Landry & Bernatchez 2001; Miller et al.

2010).

Less is known about the roles of selection and neu-

tral processes on non-MHC immune genes, even

though they may be responsible for more of the

genetic variation in disease resistance than the MHC

(Jepson et al. 1997; Hill 1998). One of the better-stud-

ied groups of non-MHC genes is the toll-like receptors

(TLRs), which are part of the innate immune

response. TLR genes recognize conserved pathogen

structures and initiate a signalling cascade leading to

the production of cytokines that amplify and regulate

the inflammatory response. Recent studies of humans

(Barreiro et al. 2009), chickens (Gallus gallus; Downing

et al. 2010) and other birds (Alcaide & Edwards 2011)

have found evidence of purifying selection on some

TLR genes, but there is also evidence of positive

selection, which increases the frequencies of new,

advantageous mutations (Nielsen 2005). It has been

suggested that this variation in selection relates to the

role and location of particular TLR genes, with stron-

ger purifying selection acting against new mutations

in intracellular TLRs that target conserved viral

nucleic acids and weaker selection on cell-surface

TLRs that recognize a wider variety of pathogens

(Barreiro et al. 2009). At interleukin receptors, which

are cytokines involved in cell signalling and part of

both the innate and adaptive immune response, bal-

ancing selection appears to predominate in chickens

(Downing et al. 2010). Thus, there is evidence for

positive, purifying and balancing selection on immu-

nity genes, and there is no clear pattern yet, even in

model species. A few studies have examined non-

MHC immune genes in wild vertebrate populations

(e.g., Jensen et al. 2008; Tonteri et al. 2010; Tschirren

et al. 2011), and in many cases studies have focused

on selection across species using summary statistics

[e.g., Tajima’s D (Tajima 1989)], rather than geographic

variation in selection pressure within species, as in

many of the MHC studies cited above.

In this study, we took advantage of markers devel-

oped for domestic chicken to explore patterns of varia-

tion at non-MHC immune genes in populations of the

greater prairie-chicken (Tympanuchus cupido pinnatus),

including the endangered Attwater’s prairie-chicken

(T. c. attwateri), a subspecies in Texas. Prairie-chickens

were once abundant on grasslands throughout mid-

western North America, but loss of habitat has resulted

in fragmented and bottlenecked populations that differ

in the degree of recent population decline (Fig. 1;

Johnsgard 2002; Johnson et al. 2011). These populations

show a positive relationship between population size

and genetic variation at neutral markers (Johnson et al.

2003) and a strong pattern of differentiation between

some populations because of genetic drift following

population declines in the last century (Bellinger et al.

2003; Johnson et al. 2003, 2004). Furthermore, we

recently found that MHC class IIB variation dropped

after a bottleneck in Wisconsin, indicating that the

effect of genetic drift overwhelmed any selection acting

to maintain variability at these genes (Eimes et al.

2011).

In this study, we examined five non-MHC immune

genes to determine whether drift has also over-

whelmed selection at these functional genes. We chose

these immunity genes because they are associated

with health and fitness in domestic chickens; polymor-

phisms at these genes are associated with differences

in bacterial load (Malek & Lamont 2003; Liu & La-

mont 2003), antibody response (Zhou & Lamont 2003)

and general mortality and growth (Ye et al. 2006). The

five immune genes were chicken B-cell marker 6

(ChB6), inhibitor of apoptosis protein-1 (IAP-1), inter-

leukin-2 (IL-2), transforming growth factor b3 (TGF-

b3) and tumour necrosis factor-related apoptosis-

inducing ligand-like protein (TRAIL-like). As part of

� 2011 Blackwell Publishing Ltd

D RIFT AND SE L EC T ION ON IMM UNE L OC I 4697

the adaptive immune system, ChB6 codes for a recep-

tor that is expressed on B lymphocytes and, when

bound, initiates the apoptosis of self-reactive B cells,

thus preventing autoimmune disease (Tregaskes et al.

1996; Pifer et al. 2002). IAP-1 suppresses cell death

and may play an important role in innate and adap-

tive immunity to intracellular bacteria (Prakash et al.

2009). As part of the innate and adaptive immune

response, IL-2 stimulates the proliferation of T cells

and natural killer cells among others (Gaffen & Liu

2004). Also part of the innate and adaptive immune

response, TGF-b genes encode signalling molecules

with a number functions, including regulating the pro-

liferation and activation of leucocytes (Letterio & Rob-

erts 1998). The function of the TRAIL-like protein is

largely unstudied. The related TRAIL gene is

expressed on both innate and adaptive immune sys-

tem cells, and its functions include the clearance of

tumours and some viral infections (Falschlehner et al.

2009). Both are members of the tumour necrosis factor

superfamily of genes, which participate in cell-signal-

ling pathways concerning inflammation and apoptosis,

among others (Bodmer et al. 2002; Chang et al. 2006).

To our knowledge, these five immune genes have not

been previously studied in any wild population of

vertebrates.

We analysed sequence data from a subset of previ-

ously sampled birds in six prairie-chicken populations

that vary in size from 176 to >170 000 birds. Prairie-

chicken populations have varying demographic

histories and span a large geographic range, so they

have likely experienced different selective pressures. In

prairie-chickens, a pattern of stronger population differ-

entiation at immune genes, compared to neutral micro-

satellites, could be attributed to positive selection if

different alleles at immune genes are locally advanta-

geous. On the other hand, a pattern of weaker popula-

tion differentiation at immune genes compared to

microsatellites could be attributed to three types of

selection: balancing, positive or purifying. In MHC

studies, lower population differentiation is often attrib-

uted to a higher effective migration rate, because new

MHC alleles that enter a population are more likely to

be favoured than neutral alleles (Schierup et al. 2000;

Bamshad & Wooding 2003). However, low FST values

could also be caused by positive or purifying selection

for a common allele across populations. Lastly, if neu-

tral forces such as drift or migration overwhelm the

effects of selection, then we expect a positive relation-

ship between immune gene diversity and population

size, similar to neutral markers, and similar patterns of

population differentiation for immune genes and micro-

satellites (e.g., Miller et al. 2010; Radwan et al. 2010).

We also analysed our sequence data with summary sta-

� 2011 Blackwell Publishing Ltd

tistics, such as Tajima’s D, to test for evidence of long-

term selection.

Materials and methods

Laboratory methods

A total of 247 adult prairie-chickens from six popula-

tions were used in this study. The six populations con-

sisted of Kansas (N = 42), Minnesota (N = 37), Missouri

(N = 32), Nebraska (N = 44), Texas (Attwater’s subspe-

cies; N = 39) and Wisconsin (Buena Vista subpopula-

tion; N = 53; Fig. 1). We used DNA previously

extracted from blood and feather samples following

procedures described in Johnson et al. (2003) and Bel-

linger et al. (2003). Except for the samples from Texas,

these individuals were previously analysed using mi-

crosatellite and mtDNA markers by Johnson et al.

(2003). The Texas samples were collected during 1990–

1994 (N = 19; Johnson et al. 2007), when the population

dropped below 1000 (Morrow et al. 2004), and 2006

(N = 20) using samples from the captive population ini-

tiated in the early 1990s. The 2006 samples consisted of

a subset of individuals that were not first-order rela-

tives (i.e., parent ⁄ offspring and full-sibs) determined by

a pedigree analysis.

Sequences from the five immune genes were amplified

using previously published primer sets (Table 1). For

ChB6, the exon sequenced encodes the extracellular

region of the molecule, and for IAP-1 the exon sequenced

primarily encodes the BIR motif essential for gene func-

tion. We primarily sequenced promoter (controlling gene

expression) from IL-2, and the TGF-b3 exon sequenced

encodes a portion of the propeptide that is cleaved off of

the mature peptide. The function of the exon sequenced

from the TRAIL-like gene is not clear. All genes were

amplified in 20 lL polymerase chain reactions (PCRs):

0.5· buffer, 0.2 mM each dNTP, 1.5 mM MgCl2, 0.5 lM

each primer and 1 U GoTaq Flexi DNA polymerase (Pro-

mega, Madison, WI). Reaction conditions were as fol-

lows: 94 �C for 5 min; then 35 cycles of 94 �C for 50 s,

annealing temperature (Table 1) for 50 s, 72 �C for

1 min; and a final extension of 72 �C for 10 min. Using

the same primers, we sequenced the PCR products at the

University of Chicago Cancer Research Center DNA

Sequencing Facility or the University North Texas, in

both forward and reverse directions.

All the individuals were previously genotyped at six

microsatellite loci (Johnson et al. 2003): ADL44,

ADL146, ADL230, LLST1, LLSD4 and LLSD9. All of the

individuals (except Texas) were a subset of the individ-

uals used in Johnson et al. (2003), which describes the

methods used for screening the loci. Texas individuals

were genotyped using identical protocols.

Table 1 Amplification information and polymorphism of the five immune genes screened

ChB6 IAP-1 IL-2 TGF-b3 TRAIL-like

Annealing temperature

(primer source)

56 �C

(Zhou &

Lamont 2003)

61 �C

(Zhou &

Lamont 2003)

54 �C

(Zhou et al. 2001)

56 �C

(Ye et al. 2006)

54 �C

(Malek &

Lamont 2003)

Fragment region Exon 3 Exon 2 Promoter, exon 1 Intron 4, exon 5 Exon 1, intron 1

No. of SNPs ⁄ variable sites

Total 3 ⁄ 4 3 ⁄ 5 2 ⁄ 7 5 ⁄ 10 9 ⁄ 21

Exon (NS) 2 ⁄ 2 (114 bp) 1 ⁄ 2 (323 bp) 0 ⁄ 0 (71 bp) 1 ⁄ 1 (55 bp) 1 ⁄ 1 (103 bp)

Exon (S) 1 ⁄ 2 2 ⁄ 3 0 ⁄ 0 0 ⁄ 1 0 ⁄ 0Intron or promoter 2 ⁄ 7 (412 bp) 4 ⁄ 8 (685 bp) 8 ⁄ 20 (555 bp)

No. of SNP haplotypes 6 4 3 7 12

The number of polymorphic single-nucleotide polymorphisms (SNPs) (variable sites with minor allele frequencies >3% within a

population) out of the total number of variable sites for exon [nonsynonymous (NS) and synonymous (S)], intron and promoter

regions is given. The number of SNP haplotypes (haplotypes formed from the polymorphic SNPs) is also given. For each gene,

fragment lengths of exons, promoter (IL-2) and introns (TGF-b3 and TRAIL-like) are given in parentheses.

4698 J . L . BOLLMER ET AL.

Sequence analysis

Sequences from the five immune genes were assembled,

edited, and aligned in the program Geneious 5.4

(Drummond et al. 2011). We phased sequences from

each gene separately using the PHASE algorithm (Ste-

phens & Donelly 2003) in DnaSP 5.0 (Librado & Rozas

2009). A burn-in of 1000 iterations followed by 1000

more iterations with a thinning interval of 10 was used,

and haplotypes with probabilities <0.95 were discarded

(N = 8). Standard sequence diversity measures were cal-

culated in DnaSP, including number of mutations and

segregating sites. We calculated the number of nonsyn-

onymous substitutions per nonsynonymous site (KA)

and the number of synonymous substitutions per syn-

onymous site (KS) for coding regions in DnaSP and

used their ratio (KA ⁄ KS) as a measure of selection. For

this analysis, the exons from four of the genes were

used (IL-2 was excluded as the exon sequence was

monomorphic).

Summary statistics

Previous work on the domestic chicken (Downing et al.

2009a,b, 2010) has demonstrated selection at immune

genes (TLRs and cytokines) using summary statistics

such F* and D* (Fu & Li 1993) and D (Tajima 1989). We

used these statistics to test for signatures of selection on

the exon and promoter regions of the immune genes

(introns were excluded). Tajima’s D, which compares

the number of nucleotide differences between pairs of

sequences and the number of polymorphic sites, was

estimated in Arlequin 3.5 (Excoffier & Lischer 2010)

with P values based on 1000 simulations. F* assesses

the number of mutations appearing only once relative

to the average number of differences between

sequences, and D* assesses the number of mutations

appearing only once relative to the total number of

mutations; both of these statistics were estimated using

DnaSP. For all three statistics, a positive value indicates

balancing selection (or a population bottleneck) and a

negative value indicates purifying or positive selection

(or population expansion).

Global FST outlier tests

We also tested for selection on the immune genes by

performing FST outlier tests, which identify loci that

show higher or lower levels of population differentia-

tion than neutral loci (Beaumont & Balding 2004). For

these tests, we constructed single-nucleotide polymor-

phism (SNP) haplotypes to use as input data, treating

them as alleles at each gene. We categorized variable

sites as polymorphic SNPs if the minor allele occurred

in at least 3% of the individuals sampled. This

excluded rare haplotypes, the inclusion of which could

have led to a potentially inflated type I error when

comparing FST values. We then constructed SNP haplo-

types for each gene separately using the phased data.

SNP haplotypes were based on the entire region

sequenced, including introns. Hardy–Weinberg exact

tests were performed on microsatellite and immune loci

(SNP haplotypes) within each population in Genepop

4.0 (Raymond & Rousset 1995). Also, linkage disequilib-

rium for each pair of loci within each population was

assessed using the log likelihood ratio statistic in Gene-

pop.

For the global FST outlier tests, we used two different

programs: FDIST2 (Beaumont & Nichols 1996) and Arle-

quin 3.5. First, we employed FDIST2 as implemented in

� 2011 Blackwell Publishing Ltd

D RIFT AND SE L EC T ION ON IMM UNE L OC I 4699

Lositan v. 1 (Antao et al. 2008). Using coalescent simu-

lations, this method identifies FST values that are out-

side the 95% and 99% confidence intervals estimated

for the neutral distribution (based on microsatellites) for

a given level of heterozygosity. We analysed each

immune gene separately with the six microsatellite loci,

as well as a final run of all eleven loci together, running

50 000 simulations assuming the infinite allele mutation

model. Two coalescent approaches were implemented

in Arlequin: one that is similar to FDIST2 (finite island

model), and one that extends FDIST2 to account for hier-

archical population structure that may result in false

positives (Excoffier et al. 2009). For the hierarchical

analysis, we classified the six populations into four

groups based on the results from the genetic distance

analyses (see below): Wisconsin, Texas, Missouri and

the remaining three populations. We included all eleven

loci in the analyses and performed 5000 simulations.

Pairwise population differentiation

To assess the degree of pairwise population differentia-

tion at immune genes compared to neutral loci, we cal-

culated a standardized measure of differentiation

(Hedrick 2005) to control for differences in heterozygos-

ity between the marker types using the program Geno-

dive 2.0 (Meirmans 2006). Genodive calculates the

maximum FST values possible given the within-popula-

tion variation and then calculates a standardized differ-

entiation measure (G¢ST) by dividing the original FST

values by the maximum FST values. For the immune

genes, the SNP haplotypes were again treated as alleles.

An analysis of isolation by distance was performed

using Mantel tests in IBDWS 3.16 (Jensen et al. 2005).

We also calculated conventional pairwise FST values

(Weir & Cockerham 1984) in FSTAT 2.9.3 (Goudet 2001)

and performed Fisher’s exact tests for genotypic differ-

entiation in Genepop. To further investigate genetic

structuring, we constructed unrooted neighbour-joining

trees based on Cavalli-Sforza & Edwards (1967) dis-

tances using PHYLIP 3.69 (Felsenstein 2009) and visual-

ized them in TreeView 1.6.6 (Page 1996). All other

statistical analyses were performed in JMP 5.0.1 (SAS

Institute, Inc.).

Results

Each gene had between four and 21 variable sites, with

a total of 47 variable sites across the five immune genes

(Table 1). We classified a subset of those sites (N = 22)

as polymorphic SNPs, while the remaining variants had

minor alleles that were present in <3% of individuals

within a population. Each gene had between two and

nine SNPs (Table 1). For each of the five genes, we con-

� 2011 Blackwell Publishing Ltd

structed haplotypes from the phased SNPs, resulting in

three to twelve haplotypes per gene (Table 1). Twelve

SNPs were in introns, two were in a promoter (IL-2),

five were nonsynonymous and three were synonymous.

All of the SNPs and haplotypes were present in two or

more populations except for a single intron SNP (and

its corresponding haplotype) in TGF-b3, which was

present only in Wisconsin.

In the 247 prairie-chickens, the microsatellite loci

ADL44, ADL146, ADL230, LLST1, LLSD4 and LLSD9

had 10, 10, 11, 8, 26 and 12 alleles, respectively. All of

the microsatellite and immune (SNP haplotypes) loci

were in Hardy–Weinberg equilibrium within each pop-

ulation except for two comparisons with P-values below

the adjusted critical value of 0.0008: TGF-b3 in Minne-

sota and ChB6 in Texas (P < 0.0001, with homozygote

excess for both). Pairwise tests for linkage disequilib-

rium found no evidence for linkage between any of the

microsatellite or immune loci within any population (all

P > 0.002; adjusted P < 0.001).

Selection on sequences

Immune gene sequences showed evidence of selection,

as nonsynonymous substitutions (KA) were significantly

greater than synonymous substitutions (KS) when

paired by gene and population (paired sample t-test

t = 3.89, d.f. = 23, P < 0.001; Table 2). Individually, each

gene followed this pattern except for IAP-1, which had

a KA greater than KS in only one population (Wisconsin;

Table S1, Supporting information). We found no evi-

dence that either KA (Kruskal–Wallis test H = 0.713,

d.f. = 5, P = 0.982) or KS (H = 0.665, d.f. = 5, P = 0.985)

differed among populations. The summary statistics D,

D* and F* showed no evidence of departure from neu-

trality except at IAP-1 (Table S2, Supporting informa-

tion). All three tests were significant and negative at

IAP-1 in Nebraska, indicating purifying or positive

selection.

Genetic diversity and population size

Consistent with the effects of genetic drift, smaller

populations had lower diversity at the immune genes

and microsatellites (Fig. 2). At the immune genes,

smaller populations had fewer haplotypes per locus

(Fig. 2; r2 = 0.73, F1,4 = 11.0, P = 0.029) and fewer seg-

regating sites (r2 = 0.96, F1,4 = 88.9, P = 0.0007) in

regressions on log-transformed population size

weighted by sample size (data in Table 2). Similarly,

microsatellite variation as measured by mean number

of alleles per locus was lower in smaller populations

(Fig. 2; r2 = 0.78, F1,4 = 14.5, P = 0.019; Table 2).

Although genetic variation decreased with smaller

Fig. 2 Mean haplotypes per immune gene or alleles per micro-

satellite locus in relation to population size (log-transformed).

Data are from Table 2. Slope for immune genes: 0.45 ± 0.14

(b ± SE). Slope for microsatellites: 1.27 ± 0.33.

Table 2 Summary of within-population genetic diversity across the five immune genes and six microsatellite loci

Gene region(s) WI NE MN MO KS TX

N 53 44 37 32 42 39

Population size 794 131 484 1868 1000 178 000 176

All regions

S 19 30 23 22 34 17

Exon only (IL-2 excluded)

KA 0.0061 ± 0.0022 0.0062 ± 0.0028 0.0058 ± 0.0023 0.0054 ± 0.0022 0.0057 ± 0.0029 0.0044 ± 0.0027

KS 0.0016 ± 0.0015 0.0030 ± 0.0027 0.0011 ± 0.0007 0.0030 ± 0.0018 0.0037 ± 0.0029 0.0023 ± 0.0023

KA ⁄ KS 3.81 2.07 5.27 1.80 1.54 1.91

Nonsynonymous SNPs 5 5 5 5 5 4

Synonymous SNPs 2 3 2 3 3 1

Intron only (TGF-b3 and TRAIL-like only)

Intron SNPs 5 10 9 7 10 8

SNP haplotypes using all regions

Total no. of haplotypes 20 26 24 25 27 19

Mean haplotypes ⁄ gene 4.0 ± 0.4 5.2 ± 1.5 4.8 ± 1.1 5.0 ± 0.8 5.4 ± 1.4 3.8 ± 1.1

Haplotype diversity 0.536 ± 0.057 0.515 ± 0.103 0.590 ± 0.061 0.542 ± 0.068 0.527 ± 0.077 0.395 ± 0.131

Microsatellites

Alleles ⁄ locus 6.5 ± 1.4 10.2 ± 2.4 9.0 ± 2.1 8.8 ± 1.7 10.3 ± 2.2 6.3 ± 1.1

Ho 0.560 ± 0.130 0.692 ± 0.082 0.696 ± 0.060 0.724 ± 0.084 0.739 ± 0.072 0.645 ± 0.047

He 0.557 ± 0.123 0.733 ± 0.079 0.728 ± 0.061 0.711 ± 0.086 0.753 ± 0.064 0.719 ± 0.049

For the genetic statistics, data from all five immune genes are included except where noted. The following statistics are given: the

number of individuals genotyped (N), estimated population size (Svedarsky et al. 2000), total number of segregating sites (S) across

all immune genes, mean (±SE) nonsynonymous substitutions per nonsynonymous site (KA) and synonymous substitutions per

synonymous site (KS), measure of diversifying selection (KA ⁄ KS), number of polymorphic single-nucleotide polymorphisms (SNPs)

that are nonsynonymous or synonymous, total number of SNP haplotypes, mean number of SNP haplotypes per gene (±SE), mean

SNP haplotype diversity (±SE), mean number of microsatellite alleles per locus (±SE), mean observed microsatellite heterozygosity

(±SE) and mean unbiased expected microsatellite heterozygosity (±SE).

4700 J . L . BOLLMER ET AL.

population size for both markers (Fig. 2; F1,8 = 84.0,

P = 0.001), the immune genes tended to decrease in

variation at a slower rate than the microsatellites

(interaction between marker and population size:

F1,8 = 5.2, P = 0.051) in an ANCOVA (r2 = 0.93)

weighted by sample size (marker effect: F1,8 = 83.9,

P < 0.001).

Global FST outlier tests

Analysis of the SNP haplotypes with global FST outlier

tests indicated that IAP-1 was under strong positive

selection. Using the FDIST2 method implemented in

Lositan, each immune gene was analysed separately in

simulations with the microsatellites and then together

in a final simulation. IAP-1 was the only outlier, having

a FST greater than expected under neutrality (Fig. 3a),

indicating divergent positive selection among popula-

tions. The other four genes fell within the 95% confi-

dence intervals of the neutral space. IAP-1 also showed

higher differentiation than expected under neutrality

with both the Arlequin finite island model (FDIST2

method, P < 0.001; Fig. 3b) and the hierarchical model

(P = 0.0001). In contrast, there was only one gene (IL-2)

in one outlier test (Arlequin finite model) that showed

lower than neutral levels of FST (P = 0.048; Fig. 3b). It

is possible that the inclusion of synonymous and intron

� 2011 Blackwell Publishing Ltd

Fig. 3 Plots of results of FST outlier tests from (a) Lositan and

(b) Arlequin (finite island model). Black dots denote FST values

based on single-nucleotide polymorphism (SNP) haplotypes,

while black triangles denote microsatellite loci. The upper and

lower 95% and 99% confidence interval boundaries are indi-

cated by dashed and solid lines, respectively. (CH = ChB6,

IA = IAP-1, IL = IL-2, TG = TGF-b3, TR = TRAIL-like).

Fig. 4 Pairwise G¢ST values for microsatellites and each

immune gene separately. 95% confidence intervals (calculated

in Genodive) are shown for the microsatellites.

D RIFT AND SE L EC T ION ON IMM UNE L OC I 4701

SNPs in the SNP haplotypes masked selection acting

on nonsynonymous SNPs; however, an analysis in Los-

itan using SNPs as individual loci did not reveal any

significant outliers except a nonsynonymous IAP-1 site

under positive selection (Fig. S1, Supporting informa-

tion).

Pairwise G¢ST comparisons

Pairwise comparisons of population differentiation

using a standardized measure (G¢ST) revealed patterns

of selection not apparent at the global FST level. Pair-

wise comparisons showed that the positive selection

on IAP-1 identified above was driven by the Wiscon-

sin population. G¢ST values for IAP-1 in Wisconsin

comparisons were higher than those for the other ten

markers in Wisconsin and for IAP-1 in non-Wisconsin

comparisons (Fig. 4). In general, however, G¢ST values

were lower for the immune genes (mean ± SE:

� 2011 Blackwell Publishing Ltd

0.109 ± 0.025, N = 15 population pairs; Table S3, Sup-

porting information) than for the microsatellites

(0.162 ± 0.020; paired t14 = 2.4, P = 0.03). This differ-

ence between microsatellites and immune genes

became stronger when IAP-1, which had large values

of G¢ST (mean ± SE: 0.247 ± 0.069), was removed from

the analysis (four immune genes: 0.063 ± 0.010; paired

t14 = 7.0, P < 0.001). Overall, 55% (33 ⁄ 60 tests without

IAP) of immune gene G¢ST values fell below the mi-

crosatellite 95% confidence intervals, and most of

these (82%, 27 ⁄ 33) were for Wisconsin and Texas

comparisons, which had the largest G¢ST values for mi-

crosatellites (Fig. 4; Table S3, Supporting information).

FST (Table S3, Supporting information) and Cavalli-

Sforza–Edwards distances (Fig. S2, Supporting infor-

mation) showed a similar pattern of differentiation,

with Wisconsin, Texas and Missouri being most dis-

tinct.

Isolation by distance

The immune genes exhibited a weaker pattern of

isolation by distance than did the microsatellites

(Fig. 5). Standardized microsatellite G¢ST values were

positively correlated with geographic distance (IBDWS

Mantel test: Z = 7.28, r = 0.633, P = 0.034), while G¢ST

values from the four immune genes combined (IAP-1

excluded) were not correlated with geographic dis-

tance (Z = 2.80, r = 0.437, P = 0.126). The slope for the

four immune genes (b = 0.161, 95% CI: 0.074–0.247)

was significantly lower than that of the microsatellites

(b = 0.323, 95% CI: 0.173–0.473) when comparing 95%

CIs. G¢ST values for IAP-1 were not related to

geographic distance when all populations were

included (Z = 11.10, r = 0.311, P = 0.220), but they

Fig. 5 Isolation by distance assessed using G¢ST values plotted

against log of geographic distance. The IAP-1 comparisons

involving Wisconsin (all higher than the other comparisons)

are labelled with a ‘WI’.

4702 J . L . BOLLMER ET AL.

showed a stronger positive relationship when compar-

isons involving Wisconsin were removed (Z = 2.25,

r = 0.666, P = 0.072; b = 0.217, 95% CI: 0.085–0.348).

Discussion

Many different genes contribute to immunity, and an

increasing number of studies are investigating non-

MHC immune genes in nonmodel species (e.g., Alcaide

& Edwards 2011; Tschirren et al. 2011). We studied

five immune genes in bottlenecked prairie-chicken pop-

ulations, identifying variations in protein-coding

regions, as well as a promoter and introns. Polymor-

phisms at these genes are known to be associated with

fitness-related traits in domestic chickens, including

resistance to bacterial infections and antibody response

(Malek & Lamont 2003; Zhou & Lamont 2003). While

not all of our polymorphisms affect protein structure,

other studies of chickens (Hasenstein & Lamont 2007)

and humans (Chen et al. 2010) have identified synony-

mous SNPs and SNPs in untranslated genic regions

that exhibit disease associations and are probably

linked to functional polymorphisms. Global FST outlier

tests detected strong positive selection between popula-

tions at only one of the five immunity genes (IAP-1),

whereas at the other immune genes, pairwise popula-

tion differentiation was weaker than at microsatellites,

suggesting that either balancing selection or species-

wide positive or purifying selection was more impor-

tant. However, the mean number of haplotypes per

gene was correlated with population size, so genetic

drift has also played an important role in shaping

diversity at these immune genes, as it has done at

other neutral (Johnson et al. 2003, 2004, 2007) and

immune-related (Eimes et al. 2011) markers we have

studied in these populations.

Geographic patterns in immune variation

The five immune genes exhibited differing geographic

patterns of variation. We found consistent evidence for

positive selection for different alleles across populations

at IAP-1 using global FST outlier tests, and pairwise

comparisons (Fig. 4) revealed that this was driven by

strong differentiation between Wisconsin and the rest of

the populations. A single nonsynonymous site was

responsible. A thymine base was common in five popu-

lations, with a frequency ranging from 0.8 to 1, but in

Wisconsin a guanine base was more frequent (0.63),

resulting in the high measures of differentiation

between Wisconsin and the other populations. Other

studies have found selection for different alleles in dif-

ferent populations. In domestic chickens, an IAP-1 SNP

in the same region we sequenced tended to have a posi-

tive effect on body weight of chicks in one environ-

ment, but a negative effect in another (Ye et al. 2006).

Ye et al. (2006) also found that particular SNP alleles at

ChB6, IL-2, TRAIL-like and TGF-b3 varied in their effect

on fitness traits in different inbred lines of domestic

chickens. Similarly, at the MHC, there is evidence that

selection on particular alleles varies with geographic

differences in parasite strains (Bonneaud et al. 2006;

Loiseau et al. 2009) and pollution (Cohen 2002). A vari-

ety of other immunity genes have been identified as FST

outliers and associated with latitude and temperature in

Atlantic salmon (Salmo salar; Tonteri et al. 2010). Further

study is needed to determine the fitness effects of the

nonsynonymous SNP we identified in the prairie-

chicken.

At the other immune genes studied here (IL-2, ChB6,

TRAIL-like and TGF-b3), results from the global FST

tests, pairwise G¢ST comparisons with microsatellites

and isolation by distance were more consistent with

weaker population differentiation than expected under

neutrality. While only the finite model implemented in

Arlequin identified IL-2 as a significant global outlier,

the pairwise G¢ST values for IL-2 consistently fell at or

below the lower limits of the microsatellite 95% confi-

dence intervals for all 15 comparisons (Fig. 4). The

majority of TGF-b3 G¢ST values were also low, while

TRAIL-like and ChB6 showed more variation, with val-

ues falling below, within and above the 95% CI for mi-

crosatellites. The G¢ST values that involved Wisconsin

and Texas were especially low for the immune genes

relative to the microsatellites, as 75% (27 ⁄ 36 tests) of

the immune gene values (excluding IAP-1) were lower

than the 95% confidence intervals calculated for the

� 2011 Blackwell Publishing Ltd

D RIFT AND SE L EC T ION ON IMM UNE L OC I 4703

microsatellite loci. This difference occurred primarily

because the microsatellite G¢ST values at these two pop-

ulations were relatively high, allowing for more separa-

tion between them and the low immune gene values

(Fig. 4).

The low global FST and pairwise G¢ST values are likely

due to similar selection pressures across populations. A

number of MHC studies have found low FST values rel-

ative to neutral loci, which they attributed to balancing

selection or positive selection for the same advanta-

geous allele in multiple populations (e.g., Sommer 2003;

Fraser et al. 2010). Balancing selection has also been

found by analysis of summary statistics at several cyto-

kine genes in domestic chickens, particularly the inter-

leukins (Downing et al. 2009a,b, 2010). Cytokine genes

bind to many receptor types, and as a consequence, fre-

quency-dependent selection from a variety of pathogens

may favour a high number of diverse alleles (Downing

et al. 2010). Alternatively, the lower differentiation

between populations could be due to purifying selec-

tion for the same haplotypes across populations. Evi-

dence for purifying selection has been found at other

immune genes including defensins and some TLRs

(Barreiro et al. 2009; Mukherjee et al. 2009), likely to

conserve their function. In contrast to the peptide-bind-

ing genes of the MHC, none of the gene regions we

sequenced interact directly with pathogens and are

probably less likely to be selected for variability.

Outlier tests can result in false positives and nega-

tives. Certain demographic scenarios, including bottle-

necks, can cause false positives (Narum & Hess 2011).

Wisconsin prairie-chickens have gone through a bottle-

neck (Bellinger et al. 2003; Johnson et al. 2003, 2004),

and neutral markers show elevated differentiation

between this population and the others because of drift

(Fig 4). The IAP-1 outlier could be a false positive;

however, differentiation at IAP-1 was disproportion-

ately higher than for the remaining four immune genes

and the six microsatellites, suggesting that drift alone is

not responsible. Furthermore, the population that has

undergone the most extreme bottleneck (Texas) did not

show any evidence of positive selection with FST outlier

tests. False negatives may also have occurred. Despite

their low pairwise G¢ST values, three of the genes (TGF-

b3, ChB6 and TRAIL-like) were not identified as low

global FST outliers, and IL-2 was identified in only one

test. A number of factors could explain why the global

outlier tests generally failed to detect more outliers.

First, the tests may not be sensitive enough to detect

smaller changes in FST caused by weak selection (Na-

rum & Hess 2011). Second, low FST outliers are more

difficult to detect than high outliers, because in many

cases the candidate genes would need to have FST val-

ues very close to zero to be significant outliers, depend-

� 2011 Blackwell Publishing Ltd

ing on the level of differentiation present at neutral loci

(Beaumont & Balding 2004). Lastly, if selection pres-

sures vary geographically, then global outlier tests may

fail to detect selection that occurs in only a subset of

populations (Vitalis et al. 2001).

Sequence-level selection

We found an overall pattern of more frequent nonsyn-

onymous substitutions (KA) than synonymous substitu-

tions (KS) when paired by gene and population. This

indicates that positive or balancing selection has acted

on these genes historically, although not necessarily

recently (Garrigan & Hedrick 2003). In contrast, the

opposite pattern was found at 13 random protein-cod-

ing loci (nonimmune) in the closely related red grouse

(Lagopus lagopus; Quintela et al. 2010), suggesting puri-

fying selection. Recent studies in model bird species

have found similar patterns, with immune genes exhib-

iting higher levels of positive selection than other parts

of the genome (Berlin et al. 2008; Ekblom et al. 2010).

The summary statistics suggested that positive or puri-

fying selection was acting on IAP-1 in Nebraska, but

there was no significant deviation from neutrality in the

other populations. The nonsignificant results from the

summary statistics were somewhat surprising, because

other studies have detected selection using these statis-

tics on non-MHC immune genes, including other inter-

leukins in domestic chickens (Downing et al. 2009a,b,

2010). Our analyses may have been hindered by the rel-

atively low number of variable sites, a result of the

genetic bottlenecks these populations have experienced.

All of the gene regions we sequenced had polymor-

phisms associated with fitness in the domestic chicken,

but it is possible these regions were not under strong

selection. The TGF-b3 region we sequenced is part of the

molecule’s propeptide, which is excised from the final

mature peptide, and the ChB6 exon we sequenced forms

an extracellular region that is the most variable, presum-

ably due to relaxed selection compared to the more con-

served intracellular region (O’Laughlin 2010). TGF-b3

and TRAIL-like had a mix of SNPs occurring in both

introns and exons, although most were in the introns.

While the polymorphisms we sequenced within the

introns are unlikely to be directly under selection them-

selves, they may be linked to variation that is under

selection. For example, a TGF-b3 SNP associated with

antibody response in domestic chickens (Zhou & Lamont

2003) is located within the intron we sequenced.

Effects of drift on immune genes

In addition to selection, neutral forces appear to have

influenced patterns of variation at the immune genes.

4704 J . L . BOLLMER ET AL.

Like microsatellites, immune gene diversity was lower

in smaller prairie-chicken populations, consistent with

the effects of genetic drift. Similarly, within-population

MHC diversity often decreases as a result of population

bottlenecks (Radwan et al. 2010), and we have evidence

that drift reduced MHC class II variation in the Wiscon-

sin population (Eimes et al. 2011). It is not clear

whether populations are negatively affected by a loss of

MHC variability (Radwan et al. 2010), and there is

almost no information about the importance of variation

at non-MHC immune genes to fitness in wild bird pop-

ulations, but it is likely to depend on the type of selec-

tion acting on the gene and its role in immune function

(e.g., whether it acts in pathogen recognition or mediat-

ing the response to pathogens).

Conclusions

Despite a call for more studies of non-MHC immune

genes over 5 years ago (Acevedo-Whitehouse & Cunn-

ingham 2006), relatively few studies have assessed

geographic variation in selection pressures on these

genes in natural populations. In the greater prairie-

chicken, we expected to find local adaptation at the

non-MHC immune genes. Instead, the more common

pattern was one of weak population differentiation at

immune genes, suggesting similar selection pressures

on these genes across the species range. We found lit-

tle evidence for strong selection and more evidence

for weak selection and genetic drift. This is consistent

with evidence at MHC loci within the Wisconsin pop-

ulation, which has gone through a bottleneck and

been affected more strongly by drift than selection (Ei-

mes et al. 2011). These results underscore the impor-

tance of neutral forces in determining patterns of

genetic variation, even at functional loci that are pre-

sumably under selection. Future studies should

attempt to characterize the relationship between indi-

vidual variation at these immune genes and fitness to

determine the strength and direction of current selec-

tion on these genes.

Acknowledgements

Funding was provided by grants from the National Science

Foundation (DEB-0948695), United States Fish and Wildlife Ser-

vice, University of North Texas, and Research Growth Initia-

tive, University of Wisconsin-Milwaukee Graduate School.

References

Acevedo-Whitehouse K, Cunningham AA (2006) Is MHC

enough for understanding wildlife immunogenetics? Trends

in Ecology and Evolution, 21, 433–438.

Alcaide M, Edwards SV (2011) Molecular evolution of the toll-

like receptor multigene family in birds. Molecular Biology and

Evolution, 28, 1703–1715.

Allendorf FW, Hohenlohe PA, Luikart G (2010) Genomics and

the future of conservation genetics. Nature Reviews Genetics,

11, 697–709.

Antao T, Lopes A, Lopes RJ, Beja-Pereira A, Luikart G (2008)

LOSITAN: a workbench to detect molecular adaptation

based on a Fst-outlier method. BMC Bioinformatics, 9, 323.

Bamshad M, Wooding SP (2003) Signatures of natural selection

in the human genome. Nature Reviews Genetics, 4, 99–111.

Barreiro LB, Ben-Ali M, Quach H et al. (2009) Evolutionary

dynamics of human toll-like receptors and their different

contributions to host defense. PLoS Genetics, 5, e1000562.

Beaumont MA, Balding DJ (2004) Identifying adaptive genetic

divergence among populations from genome scans.

Molecular Ecology, 13, 969–980.

Beaumont MA, Nichols RA (1996) Evaluating loci for use in

the genetic analysis of population structure. Proceedings of the

Royal Society of London, Series B, 263, 1619–1626.

Bellinger MR, Johnson JA, Toepfer J, Dunn P (2003) Loss of

genetic variation in greater prairie chickens following a

population bottleneck in Wisconsin, U.S.A. Conservation

Biology, 17, 717–724.

Berlin S, Qu L, Li X, Yang N, Ellegren H (2008) Positive

diversifying selection in avian Mx genes. Immunogenetics, 60,

689–697.

Bodmer JL, Schneider P, Tschopp J (2002) The molecular

architecture of the TNF superfamily. Trends in Biochemical

Sciences, 27, 19–26.

Bonneaud C, Perez-Tris J, Federici P, Chastel O, Sorci G (2006)

Major histocompatibility alleles associated with local

resistance to malaria in a passerine. Evolution, 60, 383–389.

Cammen K, Hoffman JI, Knapp LA, Harwood J, Amos W

(2011) Geographic variation of the major histocompatibility

complex in Eastern Atlantic grey seals (Halichoerus grypus).

Molecular Ecology, 20, 740–752.

Cavalli-Sforza LL, Edwards AWF (1967) Phylogenetic analysis:

models and estimation procedures. American Journal of

Human Genetics, 19, 233–257.

Chang MX, Nie P, Xie HX, Wang GL, Gao Y (2006)

Characterization and expression analysis of TNF-related

apoptosis inducing ligand (TRAIL) in grass carp

Ctenopharyngodon idella. Veterinary Immunology and

Immunopathology, 100, 51–63.

Chen R, Davydov EV, Sirota M, Butte AJ (2010) Non-

synonymous and synonymous coding SNPs show similar

likelihood and effect size of human disease association. PLoS

ONE, 5, e13574.

Cohen S (2002) Strong positive selection and habitat-specific

amino acid substitution patterns in Mhc from an estuarine

fish under intense pollution stress. Molecular Biology and

Evolution, 19, 1870–1880.

Dionne M, Miller KM, Dodson JJ, Caron F, Bernatchez L (2007)

Clinal variation in MHC diversity with temperature:

evidence for the role of host-pathogen interaction on local

adaptation in Atlantic salmon. Evolution, 61, 2154–2164.

Downing T, Lynn DJ, Connell S et al. (2009a) Evidence of

balanced diversity at the chicken interleukin 4 receptor alpha

chain locus. BMC Evolutionary Biology, 9, 136.

� 2011 Blackwell Publishing Ltd

D RIFT AND SE L EC T ION ON IMM UNE L OC I 4705

Downing T, Lynn DJ, Connell S et al. (2009b) Contrasting

evolution of diversity at two disease-associated chicken

genes. Immunogenetics, 61, 303–314.

Downing T, Lloyd AT, O’Farrelly C, Bradley DG (2010) The

differential evolutionary dynamics of avian cytokine and

TLR gene classes. Journal of Immunology, 184, 6993–7000.

Drummond AJ, Ashton B, Buxton S et al. (2011) Geneious v5.4.

Available at: http://www.geneious.com/.

Eimes J, Bollmer J, Whittingham L, van Oosterhout C, Dunn P

(2011) Rapid loss of MHC class II variation in a bottlenecked

population is explained by drift and loss of copy number

variation. Journal of Evolutionary Biology, 24, 1847–1856.

Ekblom R, Sæther SA, Jacobsson P et al. (2007) Spatial pattern

of MHC class II variation in the great snipe (Gallinago media).

Molecular Ecology, 16, 1439–1451.

Ekblom R, French L, Slate J, Burke T (2010) Evolutionary

analysis and expression profiling of zebra finch immune

genes. Genome Biology and Evolution, 2, 781–790.

Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: a new

series of programs to perform population genetics analyses

under Linux and Windows. Molecular Ecology Resources, 10,

564–567.

Excoffier L, Hofer T, Foll M (2009) Detecting loci under

selection in a hierarchically structured population. Heredity,

103, 285–298.

Falschlehner C, Schaefer U, Walczak H (2009) Following

TRAIL’s path in the immune system. Immunology, 127, 145–

154.

Felsenstein J (2009) PHYLIP (Phylogeny Inference Package)

Version 3.69. Distributed by the author. Department of

Genome Sciences, University of Washington, Seattle, WA.

Fraser BA, Ramnarine IW, Neff BD (2010) Selection at the

MHC class IIB locus across guppy (Poecilia reticulata)

populations. Heredity, 104, 155–167.

Fu Y-X, Li W-H (1993) Statistical tests of neutrality of

mutations. Genetics, 133, 693–709.

Gaffen SL, Liu KD (2004) Overview of interleukin-2 function,

production and clinical applications. Cytokine, 28, 109–123.

Garrigan D, Hedrick PW (2003) Detecting adaptive molecular

polymorphism: lessons from the MHC. Evolution, 57, 1707–

1722.

Goudet J (2001) FSTAT, a program to estimate and test gene

diversities and fixation indices (version 2.9.3). Available at:

http://www2.unil.ch/popgen/softwares/fstat.htm. Updated

from Goudet (1995).

Hasenstein JR, Lamont SJ (2007) Chicken Gallinacin gene

cluster associated wtih Salmonella response in advanced

intercross line. Avian Diseases, 51, 561–567.

Hedrick PW (2005) A standardized genetic differentiation

measure. Evolution, 59, 1633–1638.

Hill AVS (1998) The immunogenetics of human infectious

diseases. Annual Review of Immunology, 16, 593–617.

Jensen JL, Bohonak AJ, Kelley ST (2005) Isolation by distance,

web service. BMC Genetics, 6, 13.

Jensen LF, Hansen MM, Mensberg K-LD, Loeschcke V (2008)

Spatially and temporally fluctuating selection at non-MHC

immune genes: evidence from TAP polymorphism in

populations of brown trout (Salmo trutta, L.). Heredity, 100,

79–91.

Jepson A, Banya W, Sisay-Joof F et al. (1997) Quantification of

the relative contribution of major histocompatibility complex

� 2011 Blackwell Publishing Ltd

(MHC) and non-MHC genes to human immune responses to

foreign antigens. Infection and Immunity, 65, 872–876.

Johnsgard PA (2002) Grassland Grouse and Their Conservation.

Smithsonian Institution Press, Washington, DC.

Johnson JA, Toepfer JE, Dunn PO (2003) Contrasting patterns

of mitochondrial and microsatellite population structure in

fragmented populations of greater prairie-chickens. Molecular

Ecology, 12, 3335–3347.

Johnson JA, Bellinger MR, Toepfer JE, Dunn PO (2004)

Temporal changes in allele frequencies and low effective

population size in greater prairie-chickens. Molecular Ecology,

13, 2617–2630.

Johnson JA, Dunn PO, Bouzat JL (2007) Effects of recent

population bottlenecks on reconstructing the

demographic history of prairie-chickens. Molecular Ecology,

16, 2203–2222.

Johnson JA, Schroeder MA, Robb LA (2011) Greater prairie-

chicken (Tympanuchus cupido). In: The Birds of North America

Online (ed. Poole A), doi:10.2173/bna.36. Cornell Lab of

Ornithology, Ithaca, NY.

Landry C, Bernatchez L (2001) Comparative analysis of

population structure across environments and geographical

scales at major histocompatibility complex and microsatellite

loci in Atlantic salmon (Salmo salar). Molecular Ecology, 10,

2525–2539.

Letterio JJ, Roberts AB (1998) TGF-b regulation of immune

responses. Annual Review of Immunology, 16, 137–161.

Librado P, Rozas J (2009) DnaSP v5: a software for

comprehensive analysis of DNA polymorphism data.

Bioinformatics, 25, 1451–1452.

Liu W, Lamont SJ (2003) Candidate gene approach: potential

association of caspase-1, inhibitor of apoptosis protein-1, and

prosaposin gene polymorphisms with response to Salmonella

enteritidis challenge or vaccination in young chicks. Animal

Biotechnology, 14, 61–76.

Loiseau C, Richard M, Garnier S et al. (2009) Diversifying

selection on MHC class I in the house sparrow (Passer

domesticus). Molecular Ecology, 18, 1331–1340.

Malek M, Lamont SJ (2003) Association of INOS, TRAIL, TGF-b2,

TGF-b3, and IgL genes with response to Salmonella enteritidis in

poultry. Genetics Selection Evolution, 35, S99–S111.

Meirmans PG (2006) Using the AMOVA framework to

estimate a standardized genetic differentiation measure.

Evolution, 60, 2399–2402.

Miller HC, Allendorf F, Daugherty CH (2010) Genetic

diversity and differentiation at MHC genes in island

populations of tuatara (Sphenodon spp.). Molecular Ecology,

19, 3894–3908.

Morrow ME, Rossignol TA, Silvy NJ (2004) Federal listing of

prairie-grouse: lessons from the Attwater’s Prairie-chicken.

Wildlife Society Bulletin, 32, 112–118.

Mukherjee S, Sarkar-Roy N, Wagener DK, Majumder PP (2009)

Signatures of natural selection are not uniform across genes

of innate immune system, but purifying selection is the

dominant signature. Proceedings of the National Academy of

Sciences USA, 106, 7073–7078.

Narum SR, Hess JE (2011) Comparison of FST outlier tests for

SNP loci under selection. Molecular Ecology Resources, 11,

184–194.

Nielsen R (2005) Molecular signatures of natural selection.

Annual Review of Genetics, 39, 197–218.

4706 J . L . BOLLMER ET AL.

O’Laughlin JP (2010) Genetic Diversity of the chB6 Alloantigen,

DePaul University, Chicago, IL.

Page RDM (1996) TREEVIEW: an application to display

phylogenetic trees on personal computers. Computer

Applications in the Biosciences, 12, 357–358.

Piertney SB, Oliver MK (2006) The evolutionary ecology of the

major histocompatibility complex. Heredity, 96, 7–21.

Pifer J, Robison D, Funk PE (2002) The avian Chb6 alloantigen

triggers apoptosis in a mammalian cell line. Journal of

Immunology, 169, 1372–1378.

Prakash H, Becker D, Bohme L et al. (2009) cIAP-1 controls

innate immunity to C. pneumoniae pulmonary infection. PLoS

ONE, 4, e6519.

Quintela M, Berlin S, Wang B, Hoglund J (2010) Genetic

diversity and differentiation among Lagopus lagopus

populations in Scandinavia and Scotland: evolutionary

significant untis confirmed by SNP markers. Molecular

Ecology, 19, 2380–2393.

Radwan J, Biedrzycka A, Babik W (2010) Does reduced MHC

diversity decrease viability of vertebrate populations?

Biological Conservation, 143, 537–544.

Raymond M, Rousset F (1995) GENEPOP (version 1.2):

population genetics software for exact tests and

ecumenicism. Journal of Heredity, 86, 248–249.

Schierup MH, Vekemans X, Charlesworth D (2000) The effect

of subdivision on variation at multi-allelic loci under

balancing selection. Genetical Research, 76, 51–62.

Sommer S (2003) Effects of habitat fragmentation and changes

of dispersal behaviour after a recent population decline on

the genetic variability of noncoding and coding DNA of a

monogamous Malagasy rodent. Molecular Ecology, 12, 2845–

2851.

Stephens M, Donelly P (2003) A comparison of bayesian

methods for haplotype reconstruction from population

genotype data. American Journal of Human Genetics, 73, 1162–

1169.

Svedarsky WD, Westemeier RL, Robel RJ, Gough S, Toepfer JE

(2000) Status and management of the greater prairie-chicken

Tympanuchus cupido pinnatus in North America. Wildlife

Biology, 6, 277–283.

Tajima F (1989) Statistical method for testing the neutral

mutation hypothesis by DNA polymorphism. Genetics, 123,

585–595.

Tonteri A, Vasemagi A, Lumme J, Primmer CR (2010) Beyond

MHC: signals of elevated pressure on Atlantic salmon (Salmo

salar) immune-relevant loci. Molecular Ecology, 19, 1273–1282.

Tregaskes CA, Bumstead N, Davison TF, Young JR (1996)

Chicken B-cell marker chB6 (BU-1) is a highly glycosylated

protein of unique structure. Immunogenetics, 44, 212–217.

Tschirren B, Raberg L, Westerdahl H (2011) Signatures of

selection acting on the innate immunity gene Toll-like

receptor 2 (TLR2) during the evolutionary history of rodents.

Journal of Evolutionary Biology, 24, 1232–1240.

Vitalis R, Dawson K, Boursot P (2001) Interpretation of

variation across marker loci as evidence of selection.

Genetics, 158, 1811–1823.

Weir BS, Cockerham CC (1984) Estimating F-statistics for the

analysis of population structure. Evolution, 38, 1358–1370.

Ye X, Avendano S, Dekkers JCM, Lamont SJ (2006) Association

of twelve immune-related genes with performance of three

broiler lines in two different hygiene environments. Poultry

Science, 85, 1555–1569.

Zhou H, Lamont SJ (2003) Associations of six candidate genes

with antibody response kinetics in hens. Poultry Science, 82,

1118–1126.

Zhou H, Buitenhuis AJ, Weigend S, Lamont SJ (2001)

Candidate gene promoter polymorphisms and antibody

response kinetics in chickens: interferon-c, interleukin-2, and

immunoglobulin light chain. Poultry Science, 80, 1679–1689.

The authors are interested in molecular evolution at immune

genes in natural bird populations. Current studies focus on

grouse and warblers, using both functional and neutral molec-

ular markers to assess population structure and mating sys-

tems.

Data accessibility

The GenBank accession numbers for the immune gene

sequences are JN573105–JN573162. The data were deposited in

the Dryad repository: doi:10.5061/dryad.j2q14vn5.

Supporting information

Additional supporting information may be found in the online

version of this article.

Fig. S1 Lositan analysis of the 22 individual SNPs and six mi-

crosatellite loci.

Fig. S2 Unrooted neighbor-joining phylograms of six prairie-

chicken populations.

Table S1 Nonsynonymous substitutions per nonsynonymous

site (KA) and synonymous substitutions per synonymous site

(KS) by gene and population.

Table S2 Tajima’s D and Fu and Li’s D* and F* values for each

population at each gene.

Table S3 Pairwise (A) FST and (B) G¢ST values for microsatel-

lites and immune genes.

Please note: Wiley-Blackwell are not responsible for the content

or functionality of any supporting information supplied by the

authors. Any queries (other than missing material) should be

directed to the corresponding author for the article.

� 2011 Blackwell Publishing Ltd