Rapid loss of MHC class II variation in a bottlenecked populationis explained by drift and loss of copy number variation

J. A. EIMES*, J. L. BOLLMER*, L. A. WHITTINGHAM*, J. A. JOHNSON� ,

C. VAN OOSTERHOUT�1 & P. O. DUNN*1

*University of Wisconsin-Milwaukee, Milwaukee, WI, USA

�University of North Texas, Denton, TX, USA

�University of East Anglia, Norwich, UK

Introduction

Changes in genetic variation following a population

bottleneck depend on the interaction of drift, selection,

recombination and gene flow (Crow & Kimura, 1970).

The outcome of these interactions differs fundamentally

between neutral genes and functional genes, which are

under selection and often thought to be less affected

by drift (Hedrick, 2001; Le Corre & Kremer, 2003). For

example, in small populations experiencing genetic drift,

variation at neutral loci may decline, whereas variation

at expressed genes may be maintained by balancing

selection (Hedrick, 2001). Several studies have tested this

idea by comparing levels of genetic variation at putatively

neutral markers (e.g. microsatellites) and those involved

in immune function that are thought to be under

relatively strong selection (Garrigan & Hedrick, 2003).

One of the most common functional gene families used

to examine balancing selection is the major histocom-

patibility complex (MHC), which plays a central role in

vertebrate immunity and may influence population

viability (Piertney & Oliver, 2006; Siddle et al., 2007;

Radwan et al., 2010).

MHC genes are among the most polymorphic coding

loci known in vertebrates (Klein, 1986), which is

important for their role in the recognition and presen-

tation of foreign antigens as part of the adaptive immune

response. MHC molecules differ in their ability to

recognize pathogens due to sequence variation at their

peptide-binding region (PBR) (Potts & Wakeland, 1990;

Apanius et al., 1997), and associations between particular

MHC alleles or haplotypes and parasite resistance have

been demonstrated across vertebrate taxa (Briles et al.,

1983; Langefors et al., 2001; Schad et al., 2005). Variation

in the MHC is generated at multiple levels. First, the

MHC is prone to gene duplication and deletion (Klein

et al., 1993; Nei et al., 1997), which results in copy

number variation (CNV) among individuals (Hosomichi

et al., 2006; Anmarkrud et al., 2010; Doxiadis et al., 2010)

as well as between species (Trowsdale, 1995; Kelley et al.,

2005). Several models of multigene evolution have been

proposed that may explain the evolution of CNV,

including the accordion model (Klein et al., 1993), the

Correspondence: John A. Eimes, Department of Biological Sciences,

University of Wisconsin-Milwaukee, PO Box 413, Milwaukee,

WI 53201, USA.

Tel.: 414 229 3026; fax: 414 229 3926; e-mail: jaeimes@uwm.edu1Joint last authors.

ª 2 0 1 1 T H E A U T H O R S . J . E V O L . B I O L .

J O U R N A L O F E V O L U T I O N A R Y B I O L O G Y ª 2 0 1 1 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y 1

Keywords:

bottleneck;

copy number variation;

genetic drift;

major histocompatibility complex;

selection;

Tympanuchus cupido.

Abstract

Population bottlenecks may reduce genetic variation and potentially increase

the risk of extinction. Here, we present the first study to use historic samples

to analyse loss of variation at the major histocompatibility complex (MHC),

which plays a central role in vertebrate disease resistance. Balancing selection

acts on the MHC and could moderate the loss of variation expected from drift;

however, in a Wisconsin population of greater prairie-chickens (Tympanuchus

cupido), the number of MHC class II B alleles per individual declined by

44% following a population bottleneck, compared to a loss of only 8% at

microsatellites. Simulations indicate that drift likely reduced MHC variation at

the population level, as well as within individuals by reducing the number of

gene copies per individual or by fixing the same alleles across multiple loci.

These multiple effects of genetic drift on MHC variation could have important

implications for immunity and fitness.

doi: 10.1111/j.1420-9101.2011.02311.x

birth-and-death model (Nei et al., 1997) and Associative

Balancing Complex evolution (van Oosterhout, 2009).

Second, allelic variation at individual loci is generated

by recombination and mutation, and this variation is

maintained by balancing selection acting on the PBR

(Sommer, 2005; Piertney & Oliver, 2006; Spurgin &

Richardson, 2010). Selection may also operate against

deleterious mutations that have accumulated in the MHC

(Shiina et al., 2006), and this can maintain a balanced

polymorphism at the MHC as well (van Oosterhout,

2009).

In multigene families, such as the MHC, CNV could be

lost due to drift in a similar fashion as allelic diversity is

lost at single loci (Cutler et al., 2007; Schrider & Hahn,

2010). Different mechanisms have been demonstrated

to generate CNV, including replication slippage, retro-

transposition and recombination between duplicated

sequences (Schrider & Hahn, 2010). Stochastic removal

of particular haplotypes with a relatively large number

of gene copies could eliminate several interacting (and

co-evolving) genes at once, reducing genetic variation

above and beyond that expected from the loss of

heterozygosity at a single locus alone. Furthermore,

duplicated alleles can occur at multiple loci (Chaves et al.,

2010), which means that drift could cause the fixation of

the same allele at different loci within the gene family.

Such drift-across-loci could further exacerbate the loss of

genetic variation during a population bottleneck in

multigene families.

Despite the mechanisms generating variation at the

MHC, many small populations exhibit low levels of

variability, and some are even fixed for single MHC

alleles (Hedrick et al., 2000; Miller & Lambert, 2004;

Babik et al., 2009; Radwan et al., 2010). This suggests that

drift can overwhelm balancing selection, although in

some cases, MHC diversity appears to be maintained

despite the loss of variation at neutral loci (Aguilar et al.,

2004; van Oosterhout et al., 2006a; Mona et al., 2008).

It has also been suggested that if genetic variation is

depleted through a bottleneck, divergent alleles are more

likely to be retained because they will confer resistance to

a wider suite of potential pathogens (Hedrick et al., 2002;

Hedrick, 2003). To date, no studies have directly com-

pared the genetic diversity of the MHC in a wild

population before and after a bottleneck using both

historic and contemporary DNA samples. Furthermore,

theory and computer simulation models have focused

primarily on the effects of bottlenecks on single locus

heterozygosity (Borghans et al., 2004; Ejsmond &

Radwan, 2009) and do not consider CNV, which may

be important in immune function and disease resistance

(Siddle et al., 2010).

Here, we examined genetic variation in a population of

greater prairie-chickens (Tympanuchus cupido pinnatus)

before and after a bottleneck to test the relative effects of

genetic drift and balancing selection on putatively neu-

tral (microsatellite) and functional (MHC class II B)

genes. Over the past century, prairie-chickens have

experienced one of the largest declines in range size of

any bird species in North America (Anderson & Toepfer,

1999; Johnson et al., 2011), primarily because of habitat

loss (Johnsgard, 2002; Johnson et al., 2011). For exam-

ple, in Wisconsin, the species occurred in nearly every

county in the state in the 1930s, but by the 1950s its

range was reduced to seven (of 72) counties in the centre

of the state. The population in central WI experienced

a further decline of 50% during the 1950s to approx-

imately 1200 individuals, and it has remained at that

population size for the past 50 years (Anderson &

Toepfer, 1999).

Our previous studies of greater prairie-chickens

showed that genetic variation at neutral markers, such

as microsatellites and the mtDNA control region,

declined as a consequence of the bottleneck in Wisconsin

(Bellinger et al., 2003; Johnson et al., 2003, 2004, 2007).

In the present study, we used a subset of the same

contemporary and historic samples used in the microsat-

ellite study to directly measure the effect of the bottle-

neck on allelic variation in MHC class II B genes. We

found a loss of MHC polymorphism (number of alleles in

the population) similar to microsatellites (Johnson et al.,

2004), a pattern that is consistent with the effect of drift.

However, the loss of MHC variation within individuals

(fewer alleles per individual) significantly exceeded the

rate of loss expected from drift. Here, we show that this

pattern of loss can be explained by drift acting on the

number of loci (CNV) or the fixation of the same

duplicated allele across multiple loci (drift-across-loci).

Materials and methods

Sampling and DNA extraction

To measure the change in MHC polymorphism through

the population bottleneck, we examined sequences from

40 adults (20 contemporary and 20 historic) from Buena

Vista Marsh, Portage County, WI, USA. Our sample

consisted of a subset of the same historic (1951–1952)

and contemporary (1996–1999) DNA samples that were

previously analysed using microsatellites (Bellinger

et al., 2003; Johnson et al., 2004) and mtDNA control

region sequences (Johnson et al., 2004). The historic

DNA was extracted from wings collected by F.N. and

F. Hamerstrom in 1951–1952 during the last hunting

season in Wisconsin (Bellinger et al., 2003). All historic

DNA was extracted in a separate laboratory free of any

other DNA or tissue samples from grouse (Johnson et al.,

2004).

PCR, cloning and sequencing

We generated a 239-bp amplicon (excluding primers) of

the PBR of MHC class II B (BLB) genes using the primers

Blex2F (Eimes et al., 2010) and RNA R1 (Strand et al.,

2 J. A. E IMES ET AL.

ª 2 0 1 1 T H E A U T H O R S . J . E V O L . B I O L . d o i : 1 0 . 1 1 1 1 / j . 1 4 2 0 - 9 1 0 1 . 2 0 1 1 . 0 2 3 1 1 . x

J O U R N A L O F E V O L U T I O N A R Y B I O L O G Y ª 2 0 1 1 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

2007). This fragment begins at position 14 and ends at

position 252 of the 270-bp exon 2.

PCRs contained approximately 100 ng of DNA, 0.2 lMM

of each primer (5¢ phosphorylated), 25 lL of Phusion

Flash High Fidelity PCR Master Mix (New England

Biolabs, Ipswich, MA, USA) and water to a total volume

of 50 lL. Cycling conditions included an initial denatur-

ation step at 94 �C for two min followed by 30 cycles of

94, 62 and 72 �C each for 30 s and a final extension of

72 �C for 10 min. Target PCR products were gel excised

before purification in spin columns (Qiagen-QIAquick,

Germantown, MD, USA). PCR amplicons were then

cloned using Clonesmart blunt end high count kanamy-

cin cloning kits with chemically competent cells (Luci-

gen, Middleton, WI, USA). Colonies were diluted in

20 lL of water, and gene products were amplified by

whole colony PCR using the vector primers SL-1 and

SR-2 in the Clonesmart kit. Whole colony PCR amplifi-

cations contained 3 lL of template, 0.625 lMM of each

primer, 0.5 mMM of each dNTP, 1.5 mMM MgCl2, 1· PCR

buffer and 1.0 U of GoTaq Flexi polymerase in a total

volume of 20 lL. The cycling conditions included an

initial denaturation at 94 �C for 2 min followed by 30

cycles of 94, 55 and 72 �C each for 1 min and a final

extension of 72 �C for 10 min. Sequencing was per-

formed on Applied Biosystems (Carlsbad, CA, USA)

3730XL and 3130 automated DNA sequencers (Univer-

sity of Chicago Cancer Research Center). Sequences were

labelled BLB3-BLB26 and were aligned manually using

BIOIOEDITDIT v. 7.0.9.0 (Hall, 1999) and GENEIOUSENEIOUS PRORO v.

5.0.4 (Biomatters, Ltd., Aukland, New Zealand).

To avoid the inclusion of erroneous sequences into our

data set that may be caused by heteroduplex formation,

Taq error or deamination (Hansen et al., 2001), all

sequences (alleles) were verified by at least two inde-

pendent PCR, cloning and sequencing rounds (Cum-

mings et al., 2010). Thus, identical sequences from these

two independent samples are likely to be genuine

autapomorphies (Paabo et al., 2004; Sefc et al., 2007).

Furthermore, none of the eight single nucleotide substi-

tutions that we discovered in historic samples were

distinguished by a cytosine to thymine replacement,

which is often caused by deamination (Paabo et al., 2004;

Sefc et al., 2007).

We sequenced at least 23 colonies per individual for

38 of the 40 individuals; the remaining two individuals

had 17 and 22 clones sequenced because of difficulties

amplifying historic DNA. We chose 23 colonies per

individual, because this number is expected to result in

finding all alleles of an individual with > 90% probabil-

ity. Assuming there are five loci (with ten equally

frequent alleles), the binomial probability of missing

one allele equals P = 0.089. Overall, there was no

difference (t38 = 0.1; P = 0.90) between the contempo-

rary (25.5 ± 3.4) and historic (25.3 ± 4.3) samples in the

number of colonies screened per individual. All seq-

uences were deposited in Genbank (accession numbers

FJ232512–FJ232514, FJ232516–FJ232520, GQ176848–

GQ176851 and HM011573–HM011586).

RFLP and Southern blot

To estimate the number of MHC class II loci and to verify

CNV, we performed a random fragment length poly-

morphism (RFLP) analysis using a probe that binds

specifically to MHC class II of eight contemporary

individuals. We digested seven micrograms of gDNA for

3 h with 20 units of PvuII-HF (New England BioLabs

Inc., Ipswich, MA, USA), an enzyme that has been used

successfully in other avian MHC studies (e.g. Miller &

Lambert, 2004; Balakrishnan et al., 2010). The digested

DNA was run on a 0.8% agarose gel for 19 h at 60 V in

0.5· Tris-Borate-EDTA buffer and then transferred by

Southern blotting to a Hybond (GE Healthcare Life

Sciences, Piscataway, NJ, USA) membrane following the

manufacturer’s instructions. The blot was prehybridized

at room temperature overnight in 50% formamide, 5·standard saline citrate (SSC), 10· Denhardt’s solution,

and 250 lg ⁄ mL yeast RNA. The blot was then hybridized

at 42 �C overnight with 32P-labelled 334 bp probe using

random primer labelling. The probe was designed from

an amplicon of a contemporary individual using the

primers Blex2F and E3R (Chaves et al., 2010) that

extended from the 108th bp of exon 2 to the 85th bp

of exon 3 using the first PCR conditions described

previously. The homology of this PCR product to

Galliform MHC Class II genes was verified by cloning

and sequencing as described previously and by perform-

ing a BLASTBLAST (NCBI, Bethesda, MD, USA). The hybrid-

ization buffer was the same as the prehybridization, with

the addition of dextran sulphate to a concentration of

10%. The blot was washed in 0.5· SSC, 0.1% sodium

dodecyl sulfate (SDS) by rinsing briefly at room temper-

ature, followed by two 30 min washes at 65 �C with a

buffer change between each wash. Autoradiography was

carried out for 24 h on Kodak (Rochester, NY, USA)

BioMax film at )80 �C using an intensifying screen.

Sequence analysis

Ratios of nonsynonymous (dN) and synonymous (dS)

substitution rates are often used as a test for (historic)

selection at MHC loci, with dN ⁄ dS ratios greater than one

suggesting balancing or positive selection (Garrigan &

Hedrick, 2003). We tested for positive selection at

peptide-binding codons (PBCs) and non-PBCs in MHC

class II B sequences using the modified Nei-Gojobori ⁄Jukes-Cantor method (Nei & Gojobori, 1986) in MEGAMEGA

v. 4 (Tamura et al., 2007). PBCs in our sequences were

identified using the position of PBCs in HLA II (Brown

et al., 1993).

Average pair-wise nucleotide diversity (d) and segre-

gating sites (S) were calculated with DNANASP 4.50.3

(Rozas et al., 2003). In addition, a phylogenetic network

Reduced MHC variation following a bottleneck 3

ª 2 0 1 1 T H E A U T H O R S . J . E V O L . B I O L . d o i : 1 0 . 1 1 1 1 / j . 1 4 2 0 - 9 1 0 1 . 2 0 1 1 . 0 2 3 1 1 . x

J O U R N A L O F E V O L U T I O N A R Y B I O L O G Y ª 2 0 1 1 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

was constructed using the program SplitsTree4 (Huson

& Bryant, 2006) and employed the Neighbour-Net

method (Bryant & Moulton, 2002) with Jukes-Cantor

distances.

Loss of genetic variation

Genetic variation was expressed as the number of distinct

alleles (or sequences) in the population sample (Ap) and

the number of distinct alleles (sequences) per individual

(Ai). These measures were computed by summing the

alleles across loci, which allowed us to directly compare

the temporal loss at the MHC with that at microsatellite

loci. Mann–Whitney U tests were used to analyse whether

the population bottleneck reduced Ai between temporal

samples. The relative loss of Ai was calculated as the

contemporary Ai divided by mean historic Ai. The relative

loss of Ai was compared between the MHC and microsat-

ellite loci using a randomization test with 10 000 itera-

tions to simulate the loss of MHC and microsatellite alleles.

MHC simulations

We simulated the loss of MHC diversity using an

individual-based model and several demographic scenar-

ios from molecular estimates of Ne for the Buena Vista

population (Johnson et al., 2004). These values ranged

from a maximum value of Ne = 1500 birds before the

bottleneck down to an estimated Ne = 26 in the con-

temporary population for t = 30 generations (Johnson

et al., 2004). Simulations focused on the loss of number

of unique MHC sequences at the population level (Ap),

as well as the loss of alleles per individual (Ai).

We first examined the hypothesis that drift within loci

can explain our results by testing whether the loss of

MHC variation was due to simulated population bottle-

necks with Ne = 26, 50, 100, 500 and 1500 over 30

generations. We assumed there were five loci (based on

the maximum of 10 alleles found in one individual; see

Results later) that were completely linked and present

in all individuals. The simulations took into account the

probability of not observing all alleles per individual due

to sequencing a finite number of clones (N = 23).

Drift within loci did not completely explain our results

(see later), so we tested a second hypothesis that genetic

drift accelerates the loss of genetic variation in multigene

families due to the loss of CNV. For this hypothesis, we

simulated the effects of genetic drift on: (i) the number of

alleles within a single locus, (ii) fixation of the same allele

across multiple loci and (iii) stochastic changes in CNV. If

the same allele or sequence occurs at different loci, drift

may not only reduce gene diversity (i.e. heterozygosity)

within loci, but it can also fix the same allele at different

loci. This can reduce the apparent copy number, and

we refer to this as the ‘drift-across-loci’ hypothesis. For

example, an individual may have five MHC loci, but if drift

fixes all these loci for the same allele, its multilocus

genotype is identical to an individual that is homozygous

for a single locus. Haplotypes were constructed by

randomly allocating 1 £ k £ 5 loci per individual, and

drawing k from a uniform distribution. Alleles were then

randomly allocated to these loci by sampling from the

allele frequency distribution observed in the prebottle-

neck population. Hence, some alleles could occur at

multiple loci, which introduced the possibility that the

simulated loss of alleles per individual (Ai) could be

affected by fixation or loss of alleles at more than one locus

(drift-across-loci). Next, genotypes were generated by

randomly combining the MHC haplotypes, and simula-

tions were run with the same demographic scenarios and

simulated sequencing effort (23 clones ⁄ individual), as

described previously. The mean (95% CI) simulated Ap

and Ai values were then compared with the values

observed in the historic and contemporary population.

The drift-across-loci hypothesis provided a good fit to

our results, but for completeness we also tested a more

complex third hypothesis that a combination of genetic

drift, balancing selection (s), gene conversion (c) and

migration (m) could explain our results. Ultimately, we

rejected this hypothesis because extremely high levels

of gene conversion were required to explain our results

(Fig. S1). Simulations were run 1000 times using a macro

written for MINITABINITAB 12.1 (available from the authors

upon request).

Results

Loss of population variation (Ap)

Genetic drift following the population bottleneck led to a

reduction in MHC class II variation. The contemporary

greater prairie-chicken population had only 14 distinct

MHC class II B alleles, which was significantly fewer than

the 22 alleles in the historic population (randomization

test: P < 0.001). In total, we recovered 24 distinct

sequences (alleles) from the combined populations: 10

were unique to the historic population, two were unique

to the contemporary sample and 12 were present in both

samples (Fig. 1). The loss of less common alleles in the

population was consistent with that expected from

genetic drift (Fig. 2).

The population-level loss of MHC variation, as mea-

sured by the loss of alleles across all class II B loci, was

33% (8 lost ⁄ 24 total). We found previously that the

number of haplotypes at the mtDNA control region

declined by 50% from 10 in the historic period (haplo-

type diversity: 0.889 ± 0.013, N = 19) to five in the

contemporary period [0.511 ± 0.128, N = 20; (Johnson

et al., 2004)]. There was also a 26% reduction in number

of alleles at microsatellite loci from the historic (mean

of 8.2 ± 1.3 alleles ⁄ locus, N = 42) to the contemporary

(mean of 6.1 ± 1.1 alleles ⁄ locus, N = 87) period when

accounting for differences in sample size (Johnson et al.,

2004). Thus, the loss of MHC genetic variation, at least

4 J. A. E IMES ET AL.

ª 2 0 1 1 T H E A U T H O R S . J . E V O L . B I O L . d o i : 1 0 . 1 1 1 1 / j . 1 4 2 0 - 9 1 0 1 . 2 0 1 1 . 0 2 3 1 1 . x

J O U R N A L O F E V O L U T I O N A R Y B I O L O G Y ª 2 0 1 1 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

at the population level, was comparable to that found

previously using putatively neutral markers in many of

the same individuals (Johnson et al., 2004).

The historic and contemporary populations did not

differ in average nucleotide divergence at either the PBCs

(mean p-distance for both populations = 0.02 ± 0.03) or

over the entire 239-bp fragment (mean p-distance for

both populations = 0.09 ± 0.01). The contemporary and

historic sequences intermixed in the network showing

similar divergence among alleles in both samples (Fig. 3).

The ratio of nonsynonymous and synonymous substitu-

tion rates (dN ⁄ dS) at the PBCs (2.02) was significantly

larger than one (codon-based Z test, Z = 2.01; P = 0.047)

for the contemporary population and approached signif-

icance for the historic sample (codon-based Z test,

Z = 1.74; P = 0.062; Table S1).

Loss of individual variation (Ai)

Overall, the number of distinct MHC alleles per individ-

ual (Ai) declined by 44%, from a mean (±SD) of 6.20

(±2.14) in the historic population to 3.45 (±1.88) in

the contemporary population (Mann–Whitney U test,

Z = 3.48, P < 0.001; Fig. S2). For microsatellites, how-

ever, the number of alleles per individual (summed

across six loci) declined by only 8% from 10.19 (±1.17)

in the historic to 9.34 (±1.15) in the contemporary

samples (Fig. S2).

RFLP analysis

Individuals had between two and four RFLP fragments,

indicating at least four loci in some individuals. The

banding pattern suggested MHC class II CNV as there

were at least three distinct haplotypes among the eight

individuals screened (Fig. S3).

Computer simulations

We first tested the hypothesis that genetic drift at five loci

in combination with failed amplification of some alleles

could explain the empirical results mentioned earlier

(within-locus drift hypothesis). The simulation showed

that even in the most extreme bottleneck scenario

(Ne = 26), the simulated number of alleles per individual

(Ai = 5.61) remained considerably higher than the

observed value (Ai = 3.45; Fig. 4). Therefore, we rejected

the hypothesis that drift within loci in combination with

the stochastic failure to amplify alleles could explain our

results.

Next, we examined the drift-across loci hypothesis

by simulating the effects of genetic drift on CNV and

allowing for variation in the number of loci (between 1

and 5 per individual drawn from a uniform distribution).

In this simulation, identical alleles can occur at multiple

loci and drift can act across these loci. These simulations

Fig. 1 Amino acid alignment of 24 greater

prairie-chicken major histocompatibility

complex (MHC) class II B exon 2 sequences.

Putative peptide-binding codons identified

by Brown et al. (1993) are indicated by plus

signs. Asterisks denote sites under positive

selection with a posterior probability of at

least 95% in the historic and contemporary

samples as identified by omegaMap. The two

sequences present only in the contemporary

population are in bold, whereas the

sequences present only in the historic

population are in italics.

MHC allele identifier

2 4 6 8 10 12 14 16 18 20 22 24 26

Per

cent

age

of in

divi

dual

s w

ith M

HC

alle

le

0

20

60

80

100

Contemporary sampleHistoric sample

40

100

80

60

40

20

Fig. 2 The frequency of greater prairie-chicken MHC class II B

sequences (alleles) expressed as the percentage present in individ-

uals in the contemporary population sample (black bars) and

the historic population sample (grey bars).

Reduced MHC variation following a bottleneck 5

ª 2 0 1 1 T H E A U T H O R S . J . E V O L . B I O L . d o i : 1 0 . 1 1 1 1 / j . 1 4 2 0 - 9 1 0 1 . 2 0 1 1 . 0 2 3 1 1 . x

J O U R N A L O F E V O L U T I O N A R Y B I O L O G Y ª 2 0 1 1 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

showed that the observed value of Ai (3.45) in the

contemporary post-bottleneck population fell within

the distribution of simulated values (mean = 4.69, 95%

CI = 1.7–7.9; Fig. 5). The observed variance in Ai

between individuals (3.53), which is a measure of CNV,

also fell within the 95% CI (1.63–4.62) of variances in

Ai of individuals from the simulations. Furthermore, the

distribution of simulated Ap values (mean = 11.5, 95%

CI = 5–16) included the observed Ap (14) of the con-

temporary population. Thus, our computer simulations

support the hypothesis that genetic drift reduced varia-

tion both within and between loci to produce the large

decline in number of MHC alleles per individual (Ai) and

the relatively moderate decline in the number of alleles

in the population (Ap).

Discussion

Greater prairie-chickens in Wisconsin lost a significant

number of MHC class II alleles following a population

bottleneck in the late 1950s. This finding is consistent

with our previous study of the same population using

microsatellites and mtDNA (Johnson et al., 2004), and

it is consistent with random genetic drift. However, we

also observed a dramatic loss of MHC variation at the

individual level (the number of unique sequences per

individual, Ai), and we showed that this loss was

significantly greater for the MHC (44%) than for six

(neutral) microsatellite loci (8%). We then used an

individual-based model to test whether this loss of Ai

could be explained by drift acting simply within loci,

but rejected this hypothesis. We tested a more complex

multilocus drift hypothesis, which proposes that drift

reduced gene diversity (heterozygosity) within loci, as

well as diversity between loci, by either fixing the same

allele at different loci or by eliminating haplotypes with

high gene copy number. Our simulations show this can

result in a loss of CNV, and our empirical data appear to

be best explained by this hypothesis.

The observed loss of MHC diversity at the population

level is consistent with the results from other fragmented

or bottlenecked populations (Radwan et al., 2010), as

well as with simulations of bottlenecked populations

(Ejsmond & Radwan, 2009). These studies show that

selection at the MHC does not necessarily maintain

variation over relatively short time scales. In contrast,

other studies of bottlenecked populations have found

evidence that selection may have maintained MHC

variation (Aguilar et al., 2004; van Oosterhout et al.,

2006a). In greater prairie-chickens, the dN ⁄ dS ratios

suggested selection at the MHC class II B region, but it

is probably the result of selection over an evolutionary

time scale (Eimes et al., 2010), rather than just the

50 years between our historic and contemporary

samples. It has been suggested that if allelic variation is

lost through a bottleneck, then alleles that are more

0.01

Tycu03

Tycu21

Tycu25Tycu10

Tycu06Tycu23Tycu17

Tycu24Tycu18

Tycu07

Tycu05Tycu19

Tycu26Tycu11

Tycu13Tycu14

Tycu04Tycu22

Tycu09

Tycu12

Tycu15

Coja-DCB1

Mega-IIB3

Mega-IIB2

Coja-DAB1

Coja-DAB1

Coja-DBB1

Coja-DBB1 Coja-DBB1 Coja-DBB1

Coja-DBB1 Mega-IIB1

Gaga-BLB1

Gaga-BLB2Gaga-BLB1

Tycu08Tycu16

Tycu20

Gaga-BLB2

Gaga-BLB2

Gaga-BLB2Gaga-BLB1

Gaga-BLB1Gaga-BLB1

Fig. 3 Phylogenetic network of major histo-

compatibility complex class II B sequences

(based on 239 bp) from four Galliform

species: the greater prairie-chicken (Tycu;

Tympanuchus cupido), red jungle fowl (Gaga;

Gallus gallus), Japanese quail (Coja; Coturnix

japonica) and domestic turkey (Mega; Melea-

gris gallopavo). Greater prairie-chicken

sequences found only in the contemporary

population are in boxes, sequences occurring

only in the historic population are in bold

and the remaining sequences are present

in both time periods. All of the Gallus

(AM489767–AM489775), Meleagris

(DQ993255) and Coturnix (AB078884,

AB265804 and AB282649–AB282651)

sequences were downloaded from Genbank

and originate from mapped loci.

6 J. A. E IMES ET AL.

ª 2 0 1 1 T H E A U T H O R S . J . E V O L . B I O L . d o i : 1 0 . 1 1 1 1 / j . 1 4 2 0 - 9 1 0 1 . 2 0 1 1 . 0 2 3 1 1 . x

J O U R N A L O F E V O L U T I O N A R Y B I O L O G Y ª 2 0 1 1 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

divergent are more likely to persist, assuming they

increase the chance of recognizing new pathogens

(Hedrick et al., 2002; Hedrick, 2003). However, the

contemporary alleles in our population were no more

divergent from each other than the historic alleles

(Fig. 3), similar to results from a previous simulation

study (Ejsmond & Radwan, 2009). This finding suggests

that the loss of MHC variation is governed by stochastic

processes rather than by selection.

The MHC appears to follow a birth-and-death model

of evolution, in which new genes arise via gene

duplication and genes may be lost by deletion (Nei et al.,

1997; Nei, 2005). This results in variation in MHC gene

number between haplotypes both within (Hosomichi

et al., 2006; Doxiadis et al., 2010) and across species

(Trowsdale, 1995; Kelley et al., 2005). In the Japanese

quail (Coturnix japonica), Hosomichi et al. (2006) identi-

fied between one and six transcribed class II B loci per

haplotype. Our RFLP analysis indicates that MHC class II

locus number is variable in greater prairie-chickens.

Therefore, drift could have caused an increase in the

relative frequency of haplotypes with fewer loci and a

decrease in individual level MHC variability (Fig. 5).

Additionally, identical alleles may have been present at

multiple loci, as has been found in turkeys (Chaves et al.,

2010) and guppies (McMullan, 2010), and drift could

have caused multiple loci to become homozygous for the

same alleles.

The roles of selection and drift in generating CNV

are just beginning to be studied, particularly in natural

populations (Nozawa et al., 2007; Schrider & Hahn,

2010). MHC copy number has been reduced and become

fixed among inbred strains of laboratory mice (Cutler

et al., 2007) and isolated populations of ornamental

guppies (Poecilia reticulata; van Oosterhout et al., 2006b).

In the Tasmanian devil (Sarcophilus harrisii), which is

endangered as a consequence of a contagious cancer,

MHC class I diversity is low and most of the variation is

attributable to CNV, rather than sequence polymorphism

(Siddle et al., 2010). Interestingly, Siddle et al. (2010)

speculated that devils with a lower number of gene

copies might be most resistant to the cancer, possibly as

a consequence of a selective sweep. Our study suggests

that a reduction in the number of gene copies may also

occur simply through drift in the absence of any

selection.

In conclusion, there was a significant loss of MHC class

II B diversity following a population bottleneck in this

Wisconsin population of greater prairie-chickens. The

bottleneck decreased both the number of alleles in the

population (Ap) and the average number of alleles

present in individuals (Ai). Computer simulations suggest

that these results can be explained by drift acting on the

within- and between-locus gene diversity, as well as CNV

between haplotypes. The loss of alleles could have been

caused by fixation of the same allele across multiple loci

or a reduction in the number of haplotypes with high

Number of alleles per population (Ap)0 10 15 20 25

Num

ber o

f alle

les

per i

ndiv

idua

l (Ai

)

0

2

3

4

5

6

7

Historic sampleContemporary sampleSimulated populations with 26 < N < 1500

Fig. 4 The mean (±95% CI) observed number of major histocom-

patibility complex (MHC) alleles per population (Ap) vs. the mean

(±95% CI) observed number of MHC alleles per individual (Ai) for

the historic sample (solid circle) and contemporary sample (open

circle). Also shown is the mean (±SE) numbers of Ap and Ai in

simulated populations with effective size Ne = 26, 50, 100, 500 and

1500. Genetic drift alone without loss in copy number variation

cannot explain the large decline in Ai.

Number of MHC alleles per individual (Ai)1 2 3 4 5 6 7 8 9 10

Num

ber o

f ind

ivid

uals

0

1

2

3

4

5

6

7

8

9

10

11

12Contemporary observedHistoric observedContemporary simulatedHistoric simulated

Fig. 5 Number of observed greater prairie-chicken major histo-

compatibility complex (MHC) alleles per individual (Ai) detected

in the contemporary sample (black bars) and historic sample (grey

bars). The number of alleles per individuals was significantly lower

in the contemporary sample than in the historic sample (Mann–

Whitney test, W = 282, P = 0.0005). Also shown are the distribu-

tions of the mean (±95% CI) simulated Ai values for the historic

(grey circles) and contemporary (black circles) populations. In

simulations, the individuals have a variable number of MHC loci

(range: 1–5), the historic (prebottleneck) population has a size of

Ne = 1500 and the contemporary (post-bottleneck) population has

been through a bottleneck with Ne = 26 for t = 30 generations

(see text for details).

Reduced MHC variation following a bottleneck 7

ª 2 0 1 1 T H E A U T H O R S . J . E V O L . B I O L . d o i : 1 0 . 1 1 1 1 / j . 1 4 2 0 - 9 1 0 1 . 2 0 1 1 . 0 2 3 1 1 . x

J O U R N A L O F E V O L U T I O N A R Y B I O L O G Y ª 2 0 1 1 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

gene copy numbers. Changes in CNV at the MHC are

related to disease in humans and laboratory animals

(Aldhous et al., 2010; Cahan & Graubert, 2010; Ye et al.,

2010), but they are relatively unstudied in wild popula-

tions and could be an important contributor to popula-

tion viability.

Acknowledgments

Funding was provided by grants from the Research

Growth Initiative, University of Wisconsin-Milwaukee

Graduate School, National Science Foundation (DEB-

0948695), American Ornithologists’ Union, The Ameri-

can Museum of Natural History, and a Ruth Walker

Research Award from the University of Wisconsin-

Milwaukee. We thank J. Toepfer and R. Bellinger for

tissue collection and processing, and C. Wimpee for

advice and assistance in the laboratory.

References

Aguilar, A., Roemer, G., Debenham, S., Binns, M., Garcelon, D.

& Wayne, R.K. 2004. High MHC diversity maintained by

balancing selection in an otherwise genetically monomorphic

mammal. Proc. Natl Acad. Sci. USA 101: 3490–3494.

Aldhous, M.C., Abu Bakar, S., Prescott, N.J. et al. 2010.

Measurement methods and accuracy in copy number

variation: failure to replicate associations of beta-defensin

copy number with Crohn’s disease. Hum. Mol. Genet. 19: 4930–

4938.

Anderson, R.K. & Toepfer, J.E. 1999. History, status and

management of the greater prairie chicken in Wisconsin. In:

The Greater Prairie Chicken, a National Look (W.D. Svedarsky,

R.H. Hier & N.J. Silvy, eds), pp. 39–57. Minnesota Agricultural

Experiment Station, University of Minnesota, St. Paul.

Anmarkrud, J.A., Johnsen, A., Bachmann, L. & Lifjeld, J.T.

2010. Ancestral polymorphism in exon 2 of bluethroat

(Luscinia svecica) MHC class II B genes. J. Evol. Biol. 23:

1206–1217.

Apanius, V., Penn, D., Slev, P.R., Ruff, L.R. & Potts, W.K. 1997.

The nature of selection on the major histocompatibility

complex. Crit. Rev. Immunol. 17: 179–224.

Babik, W., Pabijan, M., Arntzen, J.W. et al. 2009. Long-term

survival of a urodele amphibian despite depleted major

histocompatibility complex variation. Mol. Ecol. 18: 769–781.

Balakrishnan, C., Ekblom, R., Volker, M., Westerdahl, H.,

Godinez, R., Kotkiwicz, H. et al. 2010. Gene duplication and

fragmentation in the zebra finch major histocompatibility

complex. BMC Biol. 8: doi: 10.1186/1741-7007-8-29.

Bellinger, R., Johnson, J., Toepfer, J. & Dunn, P.O. 2003. Loss

of genetic variation in greater prairie chickens following a

population bottleneck in Wisconsin, USA. Conserv. Biol. 17:

717–724.

Borghans, J.A.M., Beltman, J.B. & De Boer, R.J. 2004. MHC

polymorphism under host-pathogen coevolution. Immunoge-

netics 55: 732–739.

Briles, W.E., Briles, R.W., Taffs, R.E. & Stone, H.A. 1983.

Resistance to a malignant lymphoma in chickens is mapped

to subregion of major histocompatibility (B) complex. Science

219: 977–979.

Brown, J.H., Jardetzky, T.S., Gorga, J.C. et al. 1993. 3-dimen-

sional structure of the human class-II histocompatibility

antigen HLA-Dr1. Nature 364: 33–39.

Bryant, D. & Moulton, V. 2002. NeighborNet: an agglomerative

method for the construction of planar phylogenetic networks.

doi: 10.1007/3-540-45784-4_28.

Cahan, P. & Graubert, T.A. 2010. Integrated genomics of

susceptibility to alkylator-induced leukemia in mice. BMC

Genomics 11: doi: 638 10.1186/1471-2164-11-638.

Chaves, L.D., Faile, G.M., Krueth, S.B., Hendrickson, J.A. &

Reed, K.M. 2010. Haplotype variation, recombination, and

gene conversion within the turkey MHC-B locus. Immunoge-

netics 62: 465–477.

Crow, F. & Kimura, A. 1970. An Introduction to Population Genetics

Theory. Harper and Row, New York.

Cummings, S., McMullan, M., Joyce, D. & van Oosterhout, C.

2010. Solutions for PCR, cloning and sequencing errors in

population genetic analysis. Conserv. Genet. 11: 1095–1097.

Cutler, G., Marshall, L.A., Chin, N., Baribault, H. & Kassner, P.D.

2007. Significant gene content variation characterizes the

genomes of inbred mouse strains. Genome Res. 17: 1743–1754.

Doxiadis, G.G.M., de Groot, N., de Groot, N.G. et al. 2010.

Extensive DRB region diversity in cynomolgus macaques:

recombination as a driving force. Immunogenetics 62: 137–147.

Eimes, J.A., Bollmer, J.L., Dunn, P.O., Whittingham, L.A. &

Wimpee, C. 2010. Mhc class II diversity and balancing

selection in greater prairie-chickens. Genetica 138: 265–271.

Ejsmond, M.J. & Radwan, J. 2011. MHC diverstity in bottlenecked

populations: a simulation model. Conserv. Genet. 12: 129–137.

Garrigan, D. & Hedrick, P.W. 2003. Perspective: detecting

adaptive molecular polymorphism: lessons from the MHC.

Evolution 57: 1707–1722.

Hall, T.A. 1999. Bioedit is a user friendly sequence alignment

editor and analysis program for Windows 95 ⁄ 98 ⁄ NT. Nucleic

Acids Symp. Ser. 41: 95–98.

Hansen, A., Willerslev, E., Wiuf, C., Mourier, T. & Arctander, P.

2001. Statistical evidence for miscoding lesions in ancient

DNA templates. Mol. Biol. Evol. 18: 262–275.

Hedrick, P.W. 2001. Conservation genetics: where are we now?

Trends Ecol. Evol. 16: 629–636.

Hedrick, P.W. 2003. The major histocompatibility complex

(MHC) in declining populations: an example of adaptive

variation. In: Reproduction Science and Integrated Conservation

(W. Holt, A. Pickard, J. Rodger & D. Wildt, eds), pp. 97–113.

Cambridge University Press, Cambridge.

Hedrick, P.W., Parker, K.M., Gutierrez-Espeleta, G.A., Rattink,

A. & Lievers, K. 2000. Major histocompatibility complex

variation in the Arabian oryx. Evolution 54: 2145–2151.

Hedrick, P.W., Lee, R.N. & Garrigan, D. 2002. Major histocom-

patibility complex variation in red wolves: evidence for

common ancestry with coyotes and balancing selection. Mol.

Ecol. 11: 1905–1913.

Hosomichi, K., Shiina, T., Suzuki, S. et al. 2006. The major

histocompatibility complex (Mhc) class IIB region has greater

genomic structural flexibility and diversity in the quail than

the chicken. BMC Genomics 7: doi: 10.1186/1471-2164-1187-

1322.

Huson, D.H. & Bryant, D. 2006. Application of phylogenetic

networks in evolutionary studies. Mol. Biol. Evol. 23: 254–

267.

Johnsgard, P.A. 2002. Grassland Grouse. Smithsonian Institution

Press, Washington, DC.

8 J. A. E IMES ET AL.

ª 2 0 1 1 T H E A U T H O R S . J . E V O L . B I O L . d o i : 1 0 . 1 1 1 1 / j . 1 4 2 0 - 9 1 0 1 . 2 0 1 1 . 0 2 3 1 1 . x

J O U R N A L O F E V O L U T I O N A R Y B I O L O G Y ª 2 0 1 1 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

Johnson, J.A., Toepfer, J. & Dunn, P.O. 2003. Contrasting

patterns of mitochondrial and microsatellite population struc-

ture in fragmented populations of greater prairie-chickens.

Mol. Ecol. 12: 3335–3347.

Johnson, J.A., Bellinger, R., Toepfer, J. & Dunn, P.O. 2004.

Temporal changes in allele frequencies and low effective

population size in greater prairie-chickens. Mol. Ecol. 13: 2617–

2630.

Johnson, J.A., Dunn, P.O. & Bouzat, J.L. 2007. Effects of recent

population bottlenecks on reconstructing the demographic

history of prairie-chickens. Mol. Ecol. 16: 2203–2222.

Johnson, J., Schroeder, M. & Robb, L. 2011. Greater prairie-

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

(Poole, E. ed.). Cornell Lab of Ornithology, Ithaca, New York.

Kelley, J., Walter, L. & Trowsdale, J. 2005. Comparative

genomics of major histocompatibility complexes. Immunoge-

netics 56: 683–695.

Klein, J. 1986. The Natural History of the Major Histocompatibility

Complex. John Wiley, New York.

Klein, J., Ono, H., Klein, D. & O’hUigin, C. 1993. The accordian

model of MHC evolution. Prog. Immunol. 8: 137–143.

Langefors, K., Lohm, J. & von Schantz, T. 2001. Allelic

polymorphism in MHC class IIB in four populations of Atlantic

salmon (Salmo salar). Immunogenetics 53: 329–336.

Le Corre, V. & Kremer, A. 2003. Genetic variability at neutral

markers, quantitative trait loci and trait in a subdivided

population under selection. Genetics 164: 1205–1219.

McMullan, M. 2010. Host-parasite Co-evolution and Genetic Varia-

tion at the Major Histocompatibility Complex (MHC) in the Guppy,

Poecilia reticulata. PhD Thesis, University of Hull, UK.

Miller, H.C. & Lambert, D.M. 2004. Genetic drift outweighs

balancing selection in shaping post-bottleneck major histo-

compatibility complex variation in New Zealand robins

(Petroicidae). Mol. Ecol. 13: 3709–3721.

Mona, S., Crestanello, B., Bankhead-Dronnet, S. et al. 2008.

Disentangling the effects of recombination, selection, and

demography on the genetic variation at a major histocompat-

ibility complex class II gene in the alpine chamois. Mol. Ecol.

17: 4053–4067.

Nei, M. 2005. Bottlenecks, genetic polymorphism and specia-

tion. Genetics 170: 1–4.

Nei, M. & Gojobori, T. 1986. Simple methods for estimating

the numbers of synonymous and nonsynonymous nucleotide

substitutions. Mol. Biol. Evol. 3: 418–426.

Nei, M., Gu, X. & Sitnikova, T. 1997. Evolution by the birth-and-

death process in multigene families of the vertebrate immune

system. Proc. Natl Acad. Sci. USA 94: 7799–7806.

Nozawa, M., Kawahara, Y. & Nei, M. 2007. Genomic drift and

copy number variation of sensory receptor genes in humans.

Proc. Natl Acad. Sci. USA 104: 20421–20426.

van Oosterhout, C. 2009. A new theory of MHC evolution:

beyond selection on the immune genes. Proc. Biol. Sci. 276:

657–665.

van Oosterhout, C., Joyce, D.A., Cummings, S.M. et al. 2006a.

Balancing selection, random genetic drift, and genetic varia-

tion at the major histocompatibility complex in two wild

populations of guppies (Poecilia reticulata). Evolution 60: 2562–

2574.

van Oosterhout, C., Joyce, D.A. & Cummings, S.M. 2006b.

Evolution of MHC class IIB in the genome of wild and

ornamental guppies, Poecilia reticulata. Heredity 97: 111–

118.

Paabo, S., Poinar, H., Serre, D. et al. 2004. Genetic analyses from

ancient DNA. Annu. Rev. Genet. 38: 645–679.

Piertney, S.B. & Oliver, M.K. 2006. The evolutionary ecology of

the major histocompatibility complex. Heredity 96: 7–21.

Potts, W.K. & Wakeland, E.K. 1990. Evolution of diversity at the

major histocompatibility complex. Trends Ecol. Evol. 5: 181–187.

Radwan, J., Biedrzycka, A. & Babik, W. 2010. Does reduced

MHC diversity decrease viability of vertebrate populations?

Biol. Conserv. 143: 537–544.

Rozas, J., Sanchez-DelBarrio, J.C., Messeguer, X. & Rozas, R.

2003. DnaSP, DNA polymorphism analyses by the coalescent

and other methods. Bioinformatics 19: 2496–2497.

Schad, J., Ganzhorn, J.U. & Sommer, S. 2005. Parasite burden

and constitution of major histocompatibility complex in the

Malagasy mouse lemur, Microcebus murinus. Evolution 59: 439–

450.

Schrider, D.R. & Hahn, M.W. 2010. Gene copy-number poly-

morphism in nature. Proc. Biol. Sci. 277: 3213–3221.

Sefc, K.M., Payne, R.B. & Sorenson, M.D. 2007. Single base

errors in PCR products from avian museum specimens and

their effect on estimates of historical genetic diversity. Conserv.

Genet. 8: 879–884.

Shiina, T., Ota, M., Shimizu, S. et al. 2006. Rapid evolution of

major histocompatibility complex class I genes in primates

generates new disease alleles in humans via hitchhiking

diversity. Genetics 173: 1555–1570.

Siddle, H.V., Kreiss, A., Eldridge, M.D.B. et al. 2007. Transmis-

sion of a fatal clonal tumor by biting occurs due to depleted

MHC diversity in a threatened carnivorous marsupial. Proc.

Natl Acad. Sci. USA 104: 16221–16226.

Siddle, H.V., Marzec, J., Cheng, Y., Jones, M. & Belov, K. 2010.

MHC gene copy number variation in Tasmanian devils:

implications for the spread of a contagious cancer. Proc. Biol.

Sci. 277: 2001–2006.

Sommer, S. 2005. The importance of immune gene variability

(MHC) in evolutionary ecology and conservation. Front. Zool.

2: 1–18.

Spurgin, L. & Richardson, D. 2010. How pathogens drive genetic

diversity: MHC, mechanisms and misunderstandings. Proc.

Biol. Sci. 277: 979–988.

Strand, T., Westerdahl, H., Hoeglund, J., Alatalo, R.V. & Siitari,

H. 2007. The Mhc class II of the black grouse (Tetrao tetrix)

consists of low numbers of B and Y genes with variable

diversity and expression. Immunogenetics 59: 725–734.

Tamura, K., Dudley, J., Nei, M. & Kumar, S. 2007. MEGA4:

molecular evolutionary genetics analysis (MEGA) software

version 4.0. Mol. Biol. Evol. 24: 1596–1599.

Trowsdale, J. 1995. Both man and bird and beast – comparative

organization of mhc genes. Immunogenetics 41: 1–17.

Ye, Z.Q., Niu, S., Yu, Y. et al. 2010. Analyses of copy number

variation of GK rat reveal new putative type 2 diabetes

susceptibility loci. PLoS ONE 5: doi: 1407710.1371/journal.

pone.0014077.

Supporting information

Additional Supporting Information may be found in the

online version of this article:

Figure S1 The mean (±95% CI) observed number of

MHC alleles per population vs. the mean (±95% CI)

observed number of MHC alleles per individual for the

Reduced MHC variation following a bottleneck 9

ª 2 0 1 1 T H E A U T H O R S . J . E V O L . B I O L . d o i : 1 0 . 1 1 1 1 / j . 1 4 2 0 - 9 1 0 1 . 2 0 1 1 . 0 2 3 1 1 . x

J O U R N A L O F E V O L U T I O N A R Y B I O L O G Y ª 2 0 1 1 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

historic sample (solid circle) and contemporary sample

(open circle).

Figure S2 The relative loss in number of alleles per

individual (Ai) for MHC (solid circles) and microsatellites

(open circles) expressed as Ai divided by mean historic Ai

of the corresponding gene.

Figure S3 RFLP ⁄ Southern blot of eight contemporary

greater prairie chickens.

Table S1 Variation in MHC class II B sequences of

greater prairie-chickens before (Historic) and after (Con-

temporary) the population bottleneck.

As a service to our authors and readers, this journal

provides supporting information supplied by the authors.

Such materials are peer-reviewed and may be re-

organized for online delivery, but are not copy-edited

or typeset. Technical support issues arising from support-

ing information (other than missing files) should be

addressed to the authors.

Received 23 February 2011; revised 10 April 2011; accepted 28 April

2011

10 J. A. EIMES ET AL.

ª 2 0 1 1 T H E A U T H O R S . J . E V O L . B I O L . d o i : 1 0 . 1 1 1 1 / j . 1 4 2 0 - 9 1 0 1 . 2 0 1 1 . 0 2 3 1 1 . x

J O U R N A L O F E V O L U T I O N A R Y B I O L O G Y ª 2 0 1 1 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y