swgavi(croBa(wmbiuwpfcwDaDwcvstmsp

AI6h

Genomics 59, 143–149 (1999)Article ID geno.1999.5869, available online at http://www.idealibrary.com on

Fine-Mapping of a Region of Variation in Recombination Rateon BTA23 to the D23S22–D23S23 Interval Using Sperm Typing

and Meiotic Breakpoint Analysis

Chankyu Park, Matthew T. Frank, and Harris A. Lewin1

Laboratory of Immunogenetics, Department of Animal Sciences, University of Illinois at Urbana–Champaign, Urbana, Illinois 61801

Received July 16, 1998; accepted April 27, 1999

dl1ivRtropmRaO1rM(eta

t1a(maefi(saaoi(a

Meiotic recombination rate (u) within chromosomeegments of similar physical size is known to varyidely throughout the genome. This variation has aenetic component, occurring between the sexes andmong individuals of the same sex. We reported pre-iously the existence of variation in u between malesn the DYA–PRL interval on bovine chromosome 23BTA23). This region contains the bovine major histo-ompatibility complex and has been shown to containecombination hotspots in humans and mice. The aimf this study was to map more finely the interval(s) onTA23 where variation in u occurs using sperm typingnd meiotic breakpoint analysis. By adding a markerDRB3) between DYA and PRL, the DYA–PRL intervalas subdivided into two adjacent intervals, thus per-itting evaluation and comparison of u among five

ulls. Significant variation in u was found for bothntervals; uDYA–DRB3 ranged from 13.2 to 28.1%, andDRB3–PRL ranged from 2.4 to 13.0%. The variation in uas individual- and region-specific. A meiotic break-oint strategy employing PCR amplification productsrom recombinant sperm was then used to refine thehromosomal location associated with variation in uithin the DYA–DRB3 interval. The subinterval23S22–D23S23 exhibited the greatest degree of vari-tion among bulls having high and low u within theYA–DRB3 interval. To confirm this result, uD23S22–D23S23

as directly evaluated in three additional randomlyhosen bulls using sperm typing. The region showingariation in u was narrowed to the D23S22–D23S23ubinterval, ranging from 4.6 to 9.2%. Identification ofhe molecular basis for variation in u may be useful forap-dependent applications, such as marker-assisted

election and positional cloning of genes affectinghysiologically important traits. © 1999 Academic Press

1 To whom correspondence should be addressed at Department ofnimal Sciences, 206 Edward R. Madigan Laboratory, University of

llinois at Urbana–Champaign, 1201 W. Gregory Drive, Urbana, IL1801. Telephone: (217) 333-5998. Fax: (217) 244-5617. E-mail:-lewin@ux1.cso.uiuc.edu.

143

INTRODUCTION

Recombination rate (u) can vary dramatically amongifferent chromosomal segments of similar physicalength (Nachman and Churchill, 1996; Robinson,996). Although the molecular basis for variation in us not understood, selection experiments proved thatariation in u has a genetic component (Detlefsen andoberts, 1921; Chinnici, 1971). Heterogeneity in u be-

ween the same loci on different linkage maps mayepresent genetic variation in some aspect of the mei-tic recombination machinery among individuals com-rising the mapping pedigrees. This variation may beultifactorial, including differences in sex (Cook, 1965;eeves et al., 1990; Shiroishi et al., 1991; Ellegren etl., 1994), genetic background (Reeves et al., 1990;llivier, 1995), haplotype (Uematsu, 1986; Heine et al.,994), age (Elston et al., 1976; Zetka and Rose, 1990),ecombination-promoting sequences (Chandley anditchell, 1988; Wahls et al., 1990), chromosome size

Kaback et al., 1992), sequence homology (Sant’Angelot al., 1992; Yoshino et al., 1995), and sites for initia-ion of chromosome pairing (Goldway et al., 1993; Wund Lichten, 1994).Sperm typing is a powerful method for high-resolu-

ion linkage analysis (Cui et al., 1989; Park et al.,995), detection of germline mutation rate (Zhang etl., 1994), and localization of recombination hotspotsHubert et al., 1994). In addition, the large number of

eiotic products in a single semen sample permitsccurate comparison of u among different males (Parkt al., 1995; Yu et al., 1996; Simianer et al., 1997) andne-mapping of regions where variation in u occursHubert et al., 1994). Variation in u within and betweenexes may have important consequences for humannd animal genetics, including accuracy of marker-ssisted selection and genetic diagnosis, identificationf infertility and nondisjunction, and the understand-ng of linkage disequilibrium and haplotype evolutionRobinson, 1996). Variation in u may also dramaticallyffect the success of positional cloning or positional

0888-7543/99 $30.00Copyright © 1999 by Academic Press

All rights of reproduction in any form reserved.

ct

ttpmmcBuDswrD;

ohawD((mDobpsFihss

escR23oP

PP1wra51Mda9fmdAhfmrmmiW

pfcaa

tmLlbtaFirTwm(

Du

144 PARK, FRANK, AND LEWIN

andidate identification of disease genes and quantita-ive trait loci (QTL).

We previously demonstrated the existence of varia-ion in u in the DYA–PRL region, an interval on BTA23hat contains the bovine major histocompatibility com-lex (MHC; Park et al., 1995). The MHC of humans andice is known to contain recombination hotspots thatay contribute to maintenance of heterozygosity and

hromosome evolution (Uematsu, 1986; Yu et al., 1996;and et al., 1998). To investigate this variation further,s within the adjacent subintervals DYA–DRB3 andRB3–PRL were compared among five bulls using the

perm typing technique. Significant differences in uithin both subintervals were detected. Chromosome

egions responsible for variation in u within the DYA–RB3 interval were then fine-mapped to an interval of4.0 cM using meiotic breakpoint analysis.

MATERIALS AND METHODS

Sire selection and genotyping. Five bulls were selected for studyn the basis of DYA, DRB3, and PRL genotypes; four bulls wereeterozygous at all three loci, and one bull was heterozygous at DYAnd PRL. The exact age of the bulls at the time of semen collectionas not available, but did not differ by more than 3 years. GenomicNA was prepared from 20 ml of sperm as described previously

Heyen et al., 1997). Seven bulls, including two from previous studiesvan Eijk et al., 1993; Park et al., 1995), were genotyped for three

icrosatellite markers that mapped between DYA and PRL:23S22, D23S7, and D23S23 (Kappes et al., 1997). The two previ-usly studied bulls (17 and 198) were included in the present studyecause they were shown to differ in u in the DYA–PRL interval. Therimer sequences and PCR conditions used to amplify the micro-atellites were based on published information (Kappes et al., 1997).ollowing identification of the smallest interval exhibiting variation

n u by meiotic breakpoint analysis (described below), three bullseterozygous at D23S22 and D23S23 were selected for additionalperm typing experiments. Heterozygotes were identified by PCRcreening of genomic DNA as described above.

Sperm sorting, lysis, and whole genome amplification by primerxtension preamplification (PEP). Sperm were prepared for floworting and PEP as described by Li et al. (1991) with minor modifi-ations. Briefly, sperm were sorted into 96-well V-bottom plates (MJesearch Inc., Cambridge, MA), lysed in 10 ml of 100 mM KOH and5 mM dithiothreitol for 20 min at 65°C and neutralized with 5 ml of00 mM KCl, 900 mM Tris–HCl (pH 8.3), and 200 mM HCl. All 15 mlf lysed sperm solution was used for whole genome amplification byEP as described (Zhang et al., 1992; Park et al., 1995).

TAB

Oligonucleotides Used for Amplific

Locus Primer site Sequence

D23S22 59 GTATGTATTTTTCCC39 GAGTCAGACATGAC59 GAGGATCTTTAAGC

D23S7 59 AGAGTGTCTTATAA39 AACTCTTTCAGTTGG59 GAAATATACAAGGT

D23S23 59 GCTCATTCTCCTGG39 GCTCCTTTAGTTTTC39 GAGCAAGTGACTTA

Note. PCR1 and PCR2 denote primer concentrations in first and s

Amplification of PRL, DRB3, DYA, D23S22, D23S7, and D23S23.CR conditions, primers, methods, and strategy for discrimination ofRL, DRB3, and DYA alleles were described previously (van Eijk et al.,993; Park et al., 1995). Amplification of D23S22, D23S7, and D23S23as performed in two stages with primers listed in Table 1. A first-

ound PCR was performed separately for each locus using a 5-ml PEPliquot in a final volume of 20 ml containing 100 mM Tris–HCl (pH 8.3),0 mM KCl, 1.5 mM (D23S22 and D23S23) or 0.5 mM (D23S7) MgCl2,U of Taq DNA polymerase, and a 50 mM concentration of each dNTP.anual hotstart was performed by adding Taq DNA polymerase and

NTPs during a 5-min incubation at 85°C following a 4-min denatur-tion step at 94°C. Parameters for the first-round PCR (D23S22) were4°C for 4 min, 85°C for 5 min, 10 cycles of 94°C for 1 min, 59°C (62°Cor D23S7 and 60°C for D23S23) for 2.5 min, 72°C for 1 min, and 25

ore cycles as described above except that the annealing step wasecreased to 1 min. A final extension at 72°C for 5 min was performed.

second-round PCR was performed separately for each locus usingeminested primers. The composition of the PCR mix was as describedor the first round of PCR except that 1 mCi [a-32P]dCTP (3000 mCi/mol; DuPont NEN, Boston) was added. Parameters for the second-

ound PCR were 94°C for 4 min, 85°C for 5 min, 25 cycles of 94°C for 1in, 60°C (D23S22 and D23S23) or 58°C (D23S7) for 30 s, 72°C for 1in, and a final extension at 72°C for 5 min. All PCRs were performed

n a 96-well V-bottom plate thermal cycler (PTC-100, MJ Research Inc.,atertown, MA).

Analysis of PCR products. D23S22, D23S7, and D23S23 PCRroducts were diluted with an equal volume of loading buffer (95%ormamide, 10 mM EDTA, 0.1% bromophenol blue, 0.1% xyleneyanol), denatured at 100°C for 5 min, and electrophoresed on 6%crylamide sequencing gels for 3 h at 70 W. Gels were exposed toutoradiographic film at 270°C for 12–24 h.

Statistical analysis. Estimates of u, amplification efficiency, con-amination rate, and sorting efficiency using sperm typing data wereade using maximum-likelihood (ML) methodology as described byazzeroni et al. (1994). This method accounts for multiple typing of

oci using PEP samples, the incomplete data associated with meioticreakpoint analysis, and allows simultaneous likelihood formula-ions for several statistical models. For example, constraints onmplification efficiency and contamination rate may be considered.or all sperm donors, the general model (no locus or allele preference

n PCR amplification and no locus- or allele-specific contaminationate) was the most parsimonious model consistent with the data.herefore, the likelihood estimates used in the subsequent analysisere calculated for all data sets using the same model, which per-itted the analysis of variation in u simultaneously for all donors

Yu et al., 1996).Homogeneity of recombination rate for the DYA–PRL, DYA–RB3, DRB3–PRL, and D23S22–D23S23 intervals was evaluatedsing likelihood-ratio tests according to the formula

2@~L1 1 L2 1 . . . 1 LN! 2 ~L1121. . .1N!#,

1

ion of D23S22, D23S7, and D23S23

3 39) PCR1 (mM) PCR2 (mM)

CCTGC 0.2AGCCTG 0.1 0.2CAGAG 0.1GCCAGGAAAG 0.2CCTGT 0.1 0.2

GCCCAG 0.1TGCAGA 0.1 0.2GTGGGAG 0.2

TCCTC 0.1

nd rounds of PCR, respectively.

LE

at

(59

ACTGCATTA

TTGAGATT

AT

eco

wdtair

safewDDc

wnnrbtCau

V

wwPhbSeo

91ta(sl1

vTto(Doht

csAf1ivcuatss

F

D

145FINE-MAPPING OF INTERVAL ON BTA23

here L1 through LN equals the likelihood estimates for individualonors 1 through N, and L1121. . .1N equals the likelihood estimate ofhe combined data set for all donors. The test statistic follows anpproximate x2 distribution with N 2 1 degrees of freedom, where Ns the number of donors. The null hypothesis of homogeneity ofecombination rate was rejected at P , 0.05.Meiotic breakpoint analysis. Meiotic breakpoint analysis is a

imple method for localizing a recombination event. This can beccomplished easily with sperm by identifying recombinant spermrom the ML analysis followed by “walking in” on the recombinationvent with additional markers (Hubert et al., 1994). PEP products inells containing sperm that were classified as recombinant in theRB3–DYA interval were typed for the three internal markers:23S22, D23S7, and D23S23. Recombinants for each interval were

ounted and u was computed according to the formula

uAB 5 ~n1/n2! 3 ~N1/N2!

u BC 5 ~n 3/n 2! 3 ~N 1/N 2!,

here “A” is DRB3, “B” is D23S22, D23S7, or D23S23, and “C” is DYA;1 is the number of recombinants between loci A and B; n2 is theumber of informative sperm for all three loci among sperm with aecombination event between loci A and C; n3 is the number of recom-inants between loci B and C; N1 is the number of recombinants be-ween loci A and C; N2 is the number of informative sperm for loci A and. Recombination rates between two internal markers were estimateds follows: uD23S22–D23S23 5 uDYA–D23S22 2 uDYA–D23S23; uD23S7–D23S23 5 uDYA–D23S7 2

DYA–D23S23; and uD23S22–D23S7 5 uDYA–D23S22 2 uDYA–D23S7.

RESULTS

ariation in Recombination Rate within theDYA–DRB3 and DRB3–PRL Intervals

Sperm typing data on 1959 sperm from seven bullsere analyzed (Table 2). Two of the donors (17 and 198)ere used in previous studies (van Eijk et al., 1993;ark et al., 1995). One of the five new donors waseterozygous for PRL and DYA only, whereas fourulls were heterozygous at PRL, DRB3, and DYA.perm typing data were consistent with negligible lev-ls of contamination (range 0 to 1.6%), a high frequencyf wells containing only one sperm (range 86.9 to

TAB

Maximum-Likelihood Estimates of uDYA–PRL, uD

Bull Breed

DYA–PRLa

No.Sperm uDYA–PRL (SE)

Loglikelihood

NoSper

7 Red Angus 378 0.280 (0.029) 2326.346 2417d Red Angus 254 0.138 (0.025) 2201.695 1930 Simmental 405 0.301 (0.027) 2338.994 24

154 South Devon 343 0.247 (0.028) 2288.586 21166 Red Angus 201 0.303 (0.040) 2184.012 12191 South Devon 202 0.337 (0.039) 2168.432 NA198d South Devon 176 0.306 (0.040) 2138.217 NA

Overall NA 1959 0.270 (0.012) 21669.077 101

a Likelihood ratio test x2(6) 5 45.59, P , 0.001.b Likelihood ratio test x2(4) 5 46.90, P , 0.001.c Likelihood ratio test x2(4) 5 29.54, P , 0.001.d Data from van Eijk et al. (1993).e Not applicable because animal is a DRB3 homozygote.

6.6%), and a low frequency of wells (range 2.2 to5.1%) having no sperm or amplification failure of allhree loci (data not shown). Estimates of allele-specificmplification efficiency ranged from 70.5% for DRB3bull 166) to 95.7% for DRB3 (bull 154; data nothown). Mendelian segregation of alleles at all threeoci was not significantly different from the expected:1 ratio for each bull (data not shown).As expected, the ML analysis showed significant

ariation in u among bulls for the DYA–PRL interval.he ML analysis supports our previous conclusionshat were made on the basis of comparing estimatesbtained for donors 17 and 198 using a pairwise t testPark et al., 1995). The recombination rate within theYA–PRL interval ranged from 13.8 to 33.7%, with anverall estimate of u 5 27.0% for the seven doublyeterozygous donors. The high-recombination pheno-ype was most prevalent.

We then investigated whether the variability in uould be localized within the DYA–PRL interval byubdividing the interval with the DRB3 marker.mong the five DRB3 heterozygotes, uDYA–DRB3 ranged

rom 13.2 to 28.1%, and uDRB3–PRL ranged from 2.4 to3.0% (Table 2). Significant differences in u for bothntervals were detected; however, the magnitude ofariation in adjacent intervals among bulls was notorrelated (Table 2). For example, bull 30 showed highDRB3–DYA and low uPRL–DRB3. Variation in u was foundmong Red Angus donors for both the DYA–DRB3 andhe DRB3–PRL intervals (Table 2). These data demon-trate that variation in u detected among donors in thistudy is not due to breed effect.

ine-Mapping of the Region of Variation inRecombination Rate to the D23S22–D23S23Interval

Recombination rates within subintervals of DYA–RB3 were studied further by meiotic breakpoint anal-

2

RB3, and uDRB3–PRL Using Single Sperm Typing

DYA–DRB3b DRB3–PRLc

uDYA–DRB3 (SE)Log

likelihoodNo.

Sperm uDRB3–PRL (SE)Log

likelihood

0.200 (0.026) 2330.606 243 0.130 (0.022) 2325.6120.132 (0.024) 2179.057 193 0.024 (0.012) 2183.0980.281 (0.029) 2369.891 251 0.046 (0.014) 2327.0800.163 (0.024) 2301.325 255 0.058 (0.015) 2279.2440.257 (0.041) 2188.722 130 0.120 (0.030) 2173.858

NA NA NA NA NANA NA NA NA NA

0.203 (0.013) 21393.052 1072 0.072 (0.008) 21303.663

LE

YA–D

.m

17153e

e

7

y8tawlrdDb3bnth

tbvctfwDtadeids

C

rb2wwoDs3bD8Mnilrial

iThD

T1of

146 PARK, FRANK, AND LEWIN

sis using PEP products. A total of 180 recombinants of04 sperm coinformative for DYA and DRB3 were re-yped with three internal markers, D23S22, D23S7,nd D23S23. These three internal markers are frame-ork markers on the BTA23 consensus framework

inkage map (Beever et al., 1996). For each DYA–DRB3ecombinant, the locations of meiotic breakpoints wereetermined. The coamplification efficiency of DYA,RB3, and the three internal markers used for meioticreakpoint analysis was 72, 52, 72, and 69% for bulls 7,0, 154, and 166, respectively (data not shown); allulls were doubly heterozygous for two of three inter-al markers (see Fig. 1). Bull 17 was not included inhe meiotic breakpoint analysis because this donor wasomozygous for all internal markers except D23S22.Recombination rates within the DYA–DRB3 subin-

ervals were estimated for four bulls using meioticreakpoint analysis. Recombination rates for the inter-als DYA–D23S23 and D23S22–DRB3, intervals adja-ent to D23S22–D23S23, showed no significant varia-ion among bulls and were consistent with us estimatedrom BTA23 framework map data (Fig. 1). Increased uas localized to the interval between D23S22 and23S23 in bull 30 (uD23S22–D23S23 5 9.2%). By contrast,

he average male u for the intervals D23S22–D23S7nd D23S7–D23S23 was 2.0 and 2.0%, respectively, inata used to build the BTA23 framework map (Beevert al., 1996). Increased u within the D23S22–D23S7nterval in bull 166 compared with the framework mapata suggests that this interval might be the smallestubinterval exhibiting variation in u (Fig. 1).

FIG. 1. Localization of the region of variation in u to the D23S22he different markers are distinguished using different symbols (DY66, us in the subintervals between DRB3 and DYA were determinether intervals was determined using the ML methods describedrequencies estimated from BTA23 workshop data (Beever et al., 19

omparison of Recombination Rate within theD23S22–D23S23 Interval

Following identification of the relatively large rate ofecombination between D23S22 and D23S23 in bull 30y meiotic breakpoint analysis, additional donors (120,18, and 266) heterozygous for D23S22 and D23S23ere selected for a direct comparison of uD23S22–D23S23. Itas necessary to select additional donors because allther bulls were homozygous for either D23S22 or23S23. Sperm typing was performed on a total of 792

perm from the three bulls, and the data (including bull0) were analyzed using the ML methodology. Recom-ination rates for bulls 30, 120, 218, and 266 within the23S22–D23S23 interval were 9.2 6 3.0%, 4.6 6 1.7%,.5 6 1.9%, and 5.8 6 2.0%, respectively (Fig. 1). TheL analysis revealed significant variation in recombi-

ation rate among the 4 bulls in the D23S22–D23S23nterval (x2(3) 5 72.56, P , 0.001; data not shown). Theargest difference was between bulls 30 and 120. Theseesults demonstrate that the D23S22–D23S23 intervals the smallest DYA–DRB3 subinterval exhibiting vari-tion in u and that sequences within this interval areikely to be responsible for these differences.

DISCUSSION

We previously demonstrated individual differencesn u for the DYA–PRL interval (Park et al., 1995).hese results are supported by the analysis reportederein and show that high recombination within theYA–PRL interval is the predominant phenotype. For

3S23 interval using sperm typing and meiotic breakpoint analysis.; D23S23, F; D23S7, E; D23S22, Œ; DRB3, ■). For bulls 154, 7, andmeiotic breakpoint analysis. The recombination rate shown for allLazzeroni and co-workers (1994). Pairwise male recombination

are shown for comparison.

–D2A, h

d byby

96)

aD(DbwPviPvbreubciit

vfvsv2avrBtm

awmeQmavirmfcap

DlfcsslbD

Pqita12adlrvppstebhm

h1Ogrbotict(iabsrssr

DftDmfimoucDtrsdta

147FINE-MAPPING OF INTERVAL ON BTA23

more detailed study of variation in u within theYA–PRL interval, an internal marker was added

DRB3), and us in the two adjacent subintervals, DYA–RB3 and DRB3–PRL, were estimated for seven bullselonging to three breeds. Significant variation in uas observed for both the DYA–DRB3 and the DRB3–RL intervals among the donors. Recombination ratearied twofold between extremes for the DYA–DRB3nterval and fivefold between extremes for the DRB3–RL interval. Recently, Simianer et al. (1997) reportedariation in u for the CYP21–D23S21 interval, which isounded by DRB3–PRL, consistent with the findingseported herein. For the DYA–PRL interval, differ-nces in u were attributable to bull 17, for which bothDYA–DRB3 and uDRB3–PRL were lower (Table 2). Except forull 17, high u in the DYA–DRB3 interval was notorrelated with the magnitude of u in the adjacentnterval DRB3–PRL. This indicates that variation in us independently controlled within the smaller subin-ervals DYA–DRB3 and DRB3–PRL.

The BTA23 workshop data (Beever et al., 1996) pro-ide a basis to confirm the reliability of u estimatedrom sperm typing experiments and to compare indi-idual us to a sex-specific average. The average male-pecific us for the DYA–DRB3 and DRB3–PRL inter-als calculated from the BTA23 workshop data were0.0 and 6.0%, respectively. In the present study, theverage us for the DYA–DRB3 and DRB3–PRL inter-als estimated from five bulls were 20.5 and 7.2%,espectively, very close to estimates obtained by theTA23 workshop. Thus, we consider sperm typing data

o be reliable and reflective of population estimatesade from live offspring.Significant differences in u for the same interval

mong males suggest that caution should be appliedhen average map distances are used as a basis forarker-assisted selection or map-based cloning. For

xample, the 95% confidence interval for mapping ofTL in outbred species rarely exceeds 20 cM, andarker maps are not sufficiently dense to identify un-

mbiguously regions of identity by descent. Therefore,ariation in u could affect the correct choice of matingndividuals with desirable QTL alleles because doubleecombinants within an interval containing a QTLay go undetected. This may be extremely important

or BTA23, as this chromosome has been shown toontain genes that influence disease resistance (Xu etl., 1994), carcass traits (Beever et al., 1990), and milkroduction (Cowan et al., 1990).A possible cause of differences in u for the DYA–RB3 and DRB3–PRL intervals is the presence of al-

elic sequences controlling recombination events. A dif-erence in physical length due to a large deletion or ahromosomal rearrangement is less likely because weuccessfully amplified all the markers used in thistudy in all bulls. No association between specific al-eles and u was detected, possibly due to a small num-er of unrelated individuals being compared; however,RB3 homozygous bulls showed higher u for the DYA–

RL interval in general (Table 2). Putative DNA se-uences influencing meiotic recombination have beendentified; for example, a poly(dA z dT) tract was showno be a component of the recombination initiation sitet the ARG4 locus in yeast (Schultes and Szostak,991). Hubert and co-workers (1994) showed that a80-kb region on HSA 4p16.3 undergoes recombinationt a six- to ninefold greater rate per unit of physicalistance than the adjacent interval by identifying theocation of crossovers. Using a similar approach, weefined the region of variation in the DYA–DRB3 inter-al. We are unable to make a direct comparison ofhysical length and u because there are no data on thehysical size of the interval. Determining the physicalize of the interval would allow us to ascertain whetherhere is suppression of recombination in low u bulls ornhancement of u in high-recombination bulls. It wille possible to address this question in future studies asigh-quality YAC and BAC libraries for cattle becomeore generally available.Although DRB3 and PRL are syntetic in cattle and

umans, these loci are located on mouse chromosomes7 and 13, respectively (discussed by Park et al., 1995).ur results raise the interesting possibility that re-ions exhibiting variation in meiotic recombinationate are associated with evolutionary chromosomereakpoints or rearrangements. In support of this the-ry the D23S22–D23S23 interval, which was found inhe present study to be the smallest subinterval exhib-ting variation in u within the DYA–DRB3 interval, islose to the site of an ancient inversion that separatedhe two MHC class II subregions in the Bos lineageBand et al., 1998). Furthermore, both intervals exhib-ted distortion in map distances estimated using radi-tion hybrid analysis, with centiray distances alwayseing proportionally greater than the distances mea-ured in centimorgans. Comparisons of recombinationates across other intervals associated with chromo-ome rearrangements, and recombination rates withinegments that are evolutionarily stable, will help toesolve this issue.By comparing frequencies of recombination withinYA–DRB3 subintervals among different bulls, we

ound that uD23S22–D23S23 of one bull (30) was greater thanhat of other bulls. The lower amplification efficiency ofRB3 alleles in sperm from bull 30 did not bias esti-ates of u because there was no evidence that ampli-

cation was either locus- or allele-specific (the generalodel with amplification efficiency being independent

f locus or allele had the best fit to the data). Increasedwithin the D23S22–D23S23 interval of bull 30 ac-

ounted for most of the variation in u across the entireYA–DRB3 interval. Thus, we have reduced the size of

he region of interest from uDYA–DRB3 5 13.2% in a low-ecombination bull to uD23S22–D23S23 5 4.6%. Likewise, theame interval in a high-recombination bull was re-uced from u 5 28.1% to u 5 9.2%. These data suggesthat the DNA sequences responsible for variation in ure located within the D23S22–D23S23 interval. The

mDFwDf7ifDvcsDtottutamft

Ustaat

B

B

B

C

C

C

C

C

D

E

E

G

H

H

H

K

K

L

L

N

O

P

R

R

S

S

S

148 PARK, FRANK, AND LEWIN

arker D23S7 is located midway between D23S22 and23S23, but was not informative in the high u bull (30;ig. 1); however, in one of the low u bulls (154), D23S7as informative, having 1.2% recombination with23S22. In addition, bulls 7 and 166 were informative

or D23S7 and D23S23, with uD23S23–D23S7 5 4.3 and.2%, respectively. Increased u for the D23S23–D23S7nterval in bull 166 was consistent with the resultsrom bull 30, thus providing evidence that the D23S23–23S7 interval may be the smallest interval exhibitingariation in u. Additional experiments are necessary toonfirm this finding. Refinement of the interval to aegment that is 1–2 cM, about the size of the D23S7–23S22 interval, is a first step toward understanding

he molecular genetic basis for the observed variationf u. A large-insert contig of the D23S22–D23S23 in-erval will permit determination of the physical dis-ance between these markers. This information can besed to identify bulls that might have deletions, inser-ions, or rearrangements that may contribute to vari-tion in recombination rate in this region. Moreover, itay be possible to identify DNA sequences responsible

or variation in u at this site as well as at other loca-ions in the genome.

ACKNOWLEDGMENTS

The authors thank Gary Durack of the Flow Cytometry Facility,niversity of Illinois Biotechnology Center for expert assistance with

perm sorting. The authors also thank Dr. Ken Lange for providinghe ML analysis program and Dr. Sandra Rodriguez-Zas for helpfuldvice. This work was supported in part by Grants 94-37205-1165nd 97-35205-4738 from the United States Department of Agricul-ure, National Research Initiative Competitive Grants Program.

REFERENCES

and, M., Larsen, J. H., Womack, J. E., and Lewin, H. A. (1998). Aradiation hybrid map of BTA23: Identification of a chromosomalrearrangement leading to separation of the cattle MHC class IIsubregions. Genomics 53: 269–275.eever, J. E., George, P. D., Fernando, R. L., Stormont, C. J., andLewin, H. A. (1990). Associations between genetic markers andgrowth and carcass traits in a paternal half-sib family of Anguscattle. J. Anim. Sci. 68: 337–344.

eever, J. E., Lewin, H. A., Barendse, W., Andersson, L., Armitage,S. M., Beattie, C. W., Burns, B. M., Davis, S. K., Kappens, S. M.,Kirkpatrick, B. W., Ma, R. Z., McGraw, R. A., Stone, R. T., andTaylor, J. F. (1996). Report of the first workshop on the geneticmap of bovine chromosome 23. Anim. Genet. 27: 69–75.

handley, A. C., and Mitchell, A. R. (1988). Hypervariable minisat-ellite regions are sites for crossing-over at meiosis in man. Cyto-genet. Cell Genet. 48: 152–155.

hinnici, J. P. (1971). Modification of recombination frequency inDrosophila: Selection for increased and decreased crossing over.Genetics 69: 71–83.ook, P. J. (1965). The Lutheran–Secretor recombination fraction inman: A possible sex difference. Ann. Hum. Genet. 28: 393–397.owan, C. M., Dentine, M. R., Ax, R. L., and Schuler, L. A. (1990).Structural variation around prolactin gene linked to quantitativetraits in an elite Holstein sire family. Theor. Appl. Genet. 79:577–582.

ui, X., Li, H., Goradia, T. M., Lange, K., Kazazian, H. H. J., Galas,D., and Arnheim, N. (1989). Single-sperm typing: Determination ofgenetic distance between the Gg-globin and parathyroid hormoneloci by using the polymerase chain reaction and allele-specificoligomers. Proc. Natl. Acad. Sci. USA 86: 9389–9393.etlefsen, J. A., and Roberts, E. (1921). Studies on crossing over. I. Theeffect of selection on crossing over values. J. Exp. Zool. 32: 333–354.llegren, H., Chowdhary, B. P., Fredholm, M., Høyheim, B.,Johansson, M., Nielsen, P. B., Thomsen, P. D., and Andersson, L.(1994). A physically anchored linkage map of pig chromosome 1uncovers sex- and position-specific recombination rates. Genomics24: 342–350.lston, R. C., Lange, K., and Namboodiri, K. K. (1976). Age trends inhuman chiasma frequencies and recombination fractions. II.Method for analyzing recombination fractions and applications tothe ABO:Nail-Patella linkage. Am. J. Hum. Genet. 28: 69–76.oldway, M., Arbel, T., and Simchen, G. (1993). Meiotic nondisjunc-tion and recombination of chromosome III and homologous frag-ments in Saccharomyces cerevisiae. Genetics 133: 149–158.eine, D., Khambata, S., Wydner, K. S., and Passmore, H. C. (1994).Analysis of recombinational hot spots associated with the p hap-lotype of the mouse MHC. Genomics 23: 168–177.eyen, D. W., Beever, J. E., Da, Y., Evert, R. E., Green, G. C., Bates,S. R. E., Ziegle, J. S., and Lewin, H. A. (1997). Exclusion probabil-ities of 22 bovine microsatellite markers in fluorescent multiplexesfor semi-automated parentage testing. Anim. Genet. 28: 21–27.ubert, R., MacDonald, M., Gusella, J., and Arnheim, N. (1994).High resolution localization of recombination hot spots usingsperm typing. Nat. Genet. 7: 420–424.aback, D. B., Guacci, V., Barber, D., and Mahon, J. W. (1992).Chromosome size-dependent control of meiotic recombination. Sci-ence 256: 228–232.appes, S. M., Keele, J. W., Stone, R. T., McGraw, R. A., Sonstegard,T. S., Smith, T. P. L., Lopez-Corrales, N. L., and Beattie, C. W.(1997). A second-generation linkage map of the bovine genome.Genome Res. 7: 235–249.

azzeroni, L. C., Arnheim, N., Schmitt, K., and Lange, K. (1994).Multipoint mapping calculations for sperm-typing data. Am. J.Hum. Genet. 55: 431–436.

i, H., Cui, X., and Arnheim, N. (1991). Analysis of DNA sequencevariation in single cells. Methods Companion Methods Enzymol. 2:49–59.achman, M. W., and Churchill, G. A. (1996). Heterogeneity in ratesof recombination across the mouse genome. Genetics 142: 537–548.llivier, L. (1995). Genetic differences in recombination frequency inthe pig (Sus scrofa). Genome 38: 1048–1051.

ark, C., Russ, I., Da, Y., and Lewin, H. (1995). Genetic mapping ofF13A to BTA23 sperm typing: Difference in recombination ratebetween bulls in the DYA–PRL interval. Genomics 27: 113–118.

eeves, R. H., Crowley, M. R., O’Hara, B. F., and Gearhart, J. D.(1990). Sex, strain, and species differences affect recombinationacross an evolutionally conserved segment of mouse chromosome16. Genomics 8: 141–148.obinson, W. (1996). The extent, mechanism, and consequences ofgenetic variation for recombination rate. Am. J. Hum. Genet. 59:1175–1183.

ant’Angelo, D. B., Lafuse, W. P., and Passmore, H. C. (1992).Evidence that nucleotide sequence identity is a requirement formeiotic crossing over within the mouse Eb recombinational hotspot. Genomics 13: 1334–1336.

chultes, N. P., and Szostak, J. W. (1991). A poly(dA–dT) tract is acomponent of the recombination initiation site at the ARG4 locusin Saccharomyces cerevisiae. Mol. Cell. Biol. 11: 322–328.

hiroishi, T., Sagai, T., Hanzawa, N., Gotoh, H., and Moriwaki, K.(1991). Genetic control of sex-dependent meiotic recombination inthe major histocompatibility complex of the mouse. EMBO J. 10:681–686.

S

U

v

W

W

X

Y

Y

Z

Z

Z

149FINE-MAPPING OF INTERVAL ON BTA23

imianer, H., Szyda, J., Ramon, G., and Lien, S. (1997). Evidence forindividual and between-family variation of the recombination ratein cattle. Mamm. Genome 8: 830–835.ematsu, Y. (1986). Molecular characterization of a meiotic recom-binational hotspot enhancing homologous equal crossing-over.EMBO J. 5: 2123–2129.

an Eijk, M. J. T., Russ, I., and Lewin, H. (1993). Order of bovineDRB3, DYA, and PRL determined by sperm typing. Mamm. Ge-nome 4: 113–118.ahls, W. P., Wallace, L. J., and Moore, P. D. (1990). Hypervariableminisatellite DNA is a hotspot for homologous recombination inhuman cells. Cell 60: 95–103.u, T-C., and Lichten, M. (1994). Meiosis-induced double-strandbreak sites determined by yeast chromatin structure. Science 263:515–518.

u, A., van Eijk, M. J. T., Park, C., and Lewin, H. A. (1994). Poly-morphism of BoLA-DRB3 exon 2 correlates with resistance and

susceptibility to persistent lymphocytosis caused by bovine leuke-mia virus. J. Immunol. 151: 6977–6986.

oshino, M., Sagai, T., Lindahl, K. F., Toyoda, Y., Moriwaki, K., andShiroishi, T. (1995). Allele-dependent recombination frequency:Homology requirement in meiotic recombination at the hot spot inthe mouse major histocompatibility complex. Genomics 27: 298–305.

u, J., Lazzeroni, L., Qin, J., Huang, M-M., Navidi, W., Erlich, H.,and Arnheim, N. (1996). Individual variation in recombinationamong human males. Am. J. Hum. Genet. 59: 1186–1192.

etka, M. C., and Rose, A. M. (1990). Sex-related differences incrossing over in Caenorhabditis elegans. Genetics 126: 355–363.

hang, L., Cui, X., Schmitt, K., Hubert, R., and Navidi, W. (1992).Whole genome amplification from a single cell: Implications forgenetic analysis. Proc. Natl. Acad. Sci. USA 89: 5847–5851.

hang, L., Leeflang, E. P., Yu, J., and Arnheim, N. (1994). Studyinghuman mutations by sperm typing: Instability of CAG trinucleotiderepeats in the human androgen receptor gene. Nat. Genet. 7: 531–535.