Incotporathtg Mouse Genome

MammalianOriginal Contributions GenomeB1 insertions as easy markers for mouse population studies

Pavel Munclinger, 1 Pierre Boursot,2 Barbara Dod2

'Biodiversity Research Group, Department of Zoology, Faculty of Science, Charles University, Vinicna 7, CZ 128 44 Praha 2,Czech Republic2Laboratoire Genome, Populations, Interactions, Universite Montpellier II, F-34095 Montpellier, France

Received: 22 November 2002 / Accepted: 5 February 2003

Abstract

Few simple, easy-to-score PCR markers are availablefor studying genetic variation in wild mice popula-tions belonging to Mus musculus at the populationand subspecific levels. In this study, we show theabundant B1 family of short interspersed DNA ele-ments (SINEs) is a very promising source of suchmarkers. Thirteen B1 sequences from different re-gions of the genome were retrieved on the basis oftheir high degree of homology to a mouse consensussequence, and the presence of these elements wasscreened for in wild derived mice representing M.spretus, macedonicus and spicilegus and the differ-ent subspecies of M. musculus. At five of these loci,varying degrees of insertion polymorphism werefound in M. m. domesticus mice. These insertionswere almost totally absent in the mice representingthe other subspecies and species. Six other B1 ele-ments were fixed in all the Mus species tested. Atthese loci, polymorphism associated with three re-striction sites in the B1 consensus sequence wasfound in M. musculus. Most of these polymorphismsappear to be ancestral as they are shared by at leastone of the other Mus species tested. Both insertionand restriction polymorphism revealed differencesbetween five inbred laboratory strains considered tobe of mainly domesticus origin, and at the six re-striction loci a surprising number of these strainscarried restriction variants that were either notfound or very infrequent in domesticus. This sug-gests that in this particular group of loci, alleles offar Eastern origin are more frequent than expected.

Although the house mouse (Mus musculus) is one ofthe most important mammalian laboratory models,

Correspondence to: P. Munclinger ; E-mail: muncling@natur.cuni.cz

there are still surprisingly few easy-to-score markersthat can be used in population genetics studies or todifferentiate among its different subspecies. Becauseof their high degree of polymorphism and the ho-moplasy (alleles of the same size resulting from in-dependent mutation events) associated with theirmode of evolution, microsatellite loci are difficultmarkers to use in such studies and there is still rel-atively little comparative sequence data for the dif-ferent subspecies of M. musculus that can be used todesign SNP markers. Another potential source ofmarkers are the orthologous SINEs, which have avery low probability of being inserted more thanonce in the same site. At the species level, youngunfixed SINE loci have proved to be very useful forstudying population structure and intraspecific de-mography (Shedlock and Okada 2000).

The youngest members of the Alu subfamiliesthat are currently active in the human genome havebeen shown to be a good source of markers for hu-man population genetics (Batzer et al. 1996), andsystematic genomic database mining (Roy-Engelet al. 2001) has been used successfully to identify thecandidate Alu elements in the genomic sequencesprovided by the Human Genome Project. The B1repeat family is the mouse equivalent of the Alufamily, and here we assess the utility of using asimilar strategy to find insertion polymorphisms ofBl elements within M. musculus that can be used asintra or intersubspecific markers.

The B1 SINE family in the mouse consists ofmore than 100,000 copies that have accumulated byretroposition throughout the course of rodent evo-lution (Rogers 1985). Like the human Alu repeats,they are derived from 7SL RNA (Ullu and Tschudi1984) and are copies of a limited number of activemaster or source copies that have inserted into thegenome by retroposition. Quentin (1989) showedthat throughout the evolution of the rodent lineages,B1 elements have been retroposed from a succession

DOI: 10.1007/s00335-002-3065-7 • Volume 14, 359-366 (2003) • © Springer-Verlag New York, Inc. 2003 359

Co — Ci 1010 GN Co O N 10 \0 — N'M rommmCoCoCO nmrorocoCo

bU H U

HHH

U0U HC7 HHC7UHH HC7UU^

HH(^^E C7UUUUUHHHHUUU<HU U<4 0UUUU HUEH H L U

^H^C7HC7^C7UC7UU0UHU-eo ^UHUC7^L UC7UU^HUUC7^ C7^^HHUL

U-<UHUC7UUL HC7^

UH^ U QH UH

UU H H r U

UHUUUHUUHU^^

HUU

EHHC^U H

^ U C7UEU

"UU^U UUUoUUU0

UU^H^EH HUCU'

H

-^UU

^ UU^UH^UUHHHUUHU^UUU<HUUU

E UHUH ^ UE FU^EC7 E .HUHH UU^HUHHUUUUUUUHUU HUHUU

UHHUUEH-^U^<^UC7

UUUUUUUUUHU HHHU0U7 UHHU^

N ^ C\ Co N- ' t Co N Ln O o, O\ ONN Co Co 010 G\ ON Vr G1 Co 1010 N0C\I^'NLno'ONNC\Cm r- N N n r^ Co \C n

ter,

-^ co CoCo N CoN Co N N N Co

— •'C - N N'

N-m^^cmO"N \ N

o\ mmC\ v Co

^m^Co^0 10 0\ rC\O1ncLO' 0"10moo m o'tOZoCNN . CCN0LnM0,o ,or.,-,C N -rooc\ao

aaa^a,aac aaaa m "M Is 1a Ct

"Pa"P4aa11 Ic tt M M to M

C m 10 1n 0 -

C N- C N- N C\ - C — CoN -N N ' d' N - N- Lr)

A ti

^^°gti Zwh^

C

C

D

C

C

y ^

C

C

C

CC

CCil

00

.,y ^ w

o ^

U UC7^

C

U ^

UC^

O ^

4.^

^ c"s

04^ o^ o

'" o-d

0

o^

cl

o °a.^u ¢'O

r

I—H

360 P. MUNCLINCER ET AL: B1 INSERTIONS IN MOUSE POPULATIONS

of variant progenitor sequences that have left sub-families of related elements in the mouse genome.The subfamily generated by a given progenitor copycan be identified by the presence of shared variantnucleotide position(s) relative to the Bl consensussequence. Kass et al. (2000) provide evidence that theprogenitors of the four most recent subfamiliesfound in M. musculus, known as A, B, C, and D,have been active concomitantly in the lineage of thegenus Mus that gave rise to M. musculus, M. spretus,M. spicilegus, and M. macedonicus.

Our knowledge of the dynamics of B1 insertionin M. musculus is not as well documented as the Alufamily in humans. Kass et al. (2000) showed thatpolymorphic B1 insertions associated with recentintegrations exist in M. musculus, but there is noinformation about how widespread they are in themouse genome. In this study, we undertook a pop-ulation survey involving 250 wild mice belonging tothe different M. musculus subspecies, to investigatethe degree of insertion polymorphism associatedwith 13 B1 insertions of presumably recent originthat show a high degree of homology to the con-sensus sequence.

Materials and methods

Mice. The reference panel of mouse DNA, isolatedfrom wild derived strains or wild mice representing asubstantial part of the M. musculus species range,came from the mouse DNA collections of the au-thors' laboratories in Montpellier and Prague. Mostof the individuals represent the M. m. domesticusand musculus populations, but a few wild derivedstrains representing M. m. castaneus (Taiwan,Thailand), M. m. molossinus, M. spretus (France andSpain), M. macedonicus (Bulgaria, Turkey, and Syr-ia), and M. spicilegus (Moldavia, Austria, and Uk-raine) were also included. The detailed list ofindividuals is available on request.

B1 insertions. Three hundred and ninety sevensequences showing significant alignments with aconsensus B1 sequence (Quentin 1989) werefound in a Blast search (Altschul et al. 1997) of7606 gene-containing genomic sequences (a totalof 2.8 Mb) extracted from GenBank. Thirteenhighly conserved B1 elements present in or close tomapped genes or pseudogenes were chosen fromamong these sequences. Their accession numbersand position in the sequence are summarized inTable 1. For the sake of simplicity, these PCR lociare referred to by the name of the nearest gene orpseudogene.

P. MUNCLINGER ET AL: BI INSERTIONS IN MOUSE POPULATIONS

361

Primers and PCR conditions. Table 1 shows thePCR primers flanking these B1 elements sequences,which were defined by using the program PRIMERS(Rozen and Skaletsky 2000). PCR reactions (10 µL)were performed in the presence of 2 mivt MgC12 and0.25 units of Taq polymerise. After an initial incu-bation at 95°C for 3 min, the reactions were ampli-fied for 35 cycles at 95°C for 45 s, 60°C for 45 s, and72°C for 45 s. The last step was a final extension at72°C for 10 min. The PCR products were run on 2%agarose gels, and the presence or absence of the givenB1 element was deduced from the size of the am-plified DNA product.

RFLP analysis. Certain amplified products weredigested with the enzymes MspI and Tagl. The re-action was done in a final volume of 20 µL con-taining a 10-µL aliquot of the amplification productto be digested, 2 U of restriction enzyme, and thecorresponding digestion buffer. The products wereanalyzed in 1.5 or 2% agarose gels.

Results

Search for insertion polymorphism of Bl elementswithin M. musculus. The presence or absence of the13 conserved B1 sequences in a panel of DNA sam-ples representing the different M. musculus subspe-cies and M. spretus, spicilegus, and macedonicus isgiven in Table 1. Six of these insertions (loci Psmb5,Mtm1, Unp, Igf2r, Tpardl, and Renbp) were presentin all the samples tested, so their insertion musthave occurred before the divergence of this group ofspecies. The insertion at locus Pirb was not found inany of the wild mice tested. The B1 element in thegene Ercc2 was not found in any of individuals rep-resenting M. spretus, macedonicus, and spiciligus,although it is apparently fixed in all the subspecies ofM. musculus. However, there is considerable varia-tion at this locus, almost certainly because of the(GAAA)n microsatellite just after its poly(A) tail.

The five remaining loci show varying degrees ofpresence/absence of polymorphism within M. mus-culus (Table 2). The two elements in the loci Btk andTsx were found in the majority of the domesticusmice we tested. The distributions of the insertionsfound in the loci Pahahlbl-psi and Nkrt suggest thepresence of geographical frequency gradients, as theformer appears to be close to fixation in NorthernEurope and is much less frequent in the moreSouthern parts of the range, while the latter is verycommon in North and Western Europe but muchless frequent in SE Europe, the Middle East, and NAfrica. The element close to Brd2 is found at lowfrequency throughout the domesticus range.

Search for RFLPs in the fixed Bl elements. Ingeneral, the CpG dinucleotides present in the Blelements have a higher tendency to show changesrelative to the consensus sequence than the otherpositions (Quentin 1989), and this holds for theconserved sequences we analyzed in this study (Fig.1). Restriction enzymes with sites that include thesedinucleotides are more likely to reveal polymor-phisms than enzymes with sites in the more con-served regions of the B1 sequence. We therefore usedMspl, which has one site, and TaqI, which has twosites [TaqI( 1) and Tagl(2)] in the consensus sequence(see Fig. 1) to look for restriction polymorphismsassociated with the six B1 elements that were pres-ent in all the Mus species tested. The panel of miceused to investigate the restriction polymorphism inthe amplified fragments obtained at these loci wasessentially the same as the one used to search forinsertion polymorphisms. Haplotypes for all threesites of interest could be reconstructed in all casesbecause none of the individuals tested possessed twopolymorphic sites. The distribution of these haplo-types is given in Table 3, and they are indicated asabsent (a) or present (p) in the order of the potentialsites in the B1 sequence. Other polymorphic sites inthe regions flanking the B 1 repeats are indicated initalics (one TaqI site 5' of the Renbp insertion andone MspI site 3' of the Tpardl insertion).

All the haplotypes found in M. musculus, exceptfor the haplotype pap at the locus Unp, also occur inat least one of the other three species analyzed. Thismeans that most of the polymorphisms we found inM. musculus, spretus, macedonicus, and spiciligusmust have already been present in the commonlineage leading to these species. Although it is pos-sible that certain of these shared haplotypes are theresult of different mutational events, it is very un-likely that most of them are homoplasies.

Even though all the haplotypes found in M.musculus are shared by two or more of the subspe-cies, strikingly different haplotype frequencies be-tween domesticus and musculus are found at four ofthe loci. At the loci Psmb and Tpardl a differenthaplotype is almost completely fixed in musculusand domesticus, and at the loci Mtm and Igf2r thehaplotypes fixed in domesticus are found only in15% and 35% of the musculus samples respectively.The sampling of castaneus and molossinus is toosmall for a reliable estimation of their haplotypefrequencies to be made.

Although the same Renbp haplotype (pppa) ispresent in most of the mice belonging to these twosubspecies, an insertion of 10-20 bp, which is absentin all the other mice tested, was found in both thefrequent and infrequent domesticus haplotypes.

362

P. MUNCLINGER ET AL: B1 INSERTIONS IN MOUSE POPULATIONS

Table 2. The insertion frequency in various geographical location of the 5 B1 elements that were only found in wild micebelonging to M. musculus. The figures in brackets give the number of mice tested when it differs from the indicatednumber. NT is not tested

Geographical region Localities Mice Brd2 Nktr Pafaha-psi Tsx Btk

M. m. domesticusN Europe Jutland 8 60 0 0.49 0.92 0.85 1

Westphalia, Germany 4 8 NT 0.82 0.88 NT NTN Bavaria, Germany 9 13 0.35 0.8 1 1 1England 1 6 0 1 0 0.67 1

S & SW Europe Switzerland 1 3 0 0.83 0 1 1N Italy 3 7 0.07 0.86 0.29 1 1S France 8 19 0.3 0.76 0.32 0.68 1Spain 3 3 0.03 0.83 0 1 1

SE Europe & S Bulgaria 2 2 0 0.25 0.25 1 1Near East N Greece 1 5 0 0.30 0 1 1

E. Turkey 5 10 0.20 0.35 0 0.80 1Israel 3 4 0.13 0 0.38 1 0.60Egypt 1 1 0 0 0 1 1

N & W Africa Libya 5 16 0.13 0.34 0.09 0.24 1Tunisia 1 1 0 0.50 0 1 1Algeria 1 1 1 0.50 0 1 1Maroco 2 2 0 0 0 1 1Senegal 1 1 0 0 0.50 1 1

Pacific Oceania 1 1 0 1 0.50 1 1Australia 1 3 0 0.67 0.67 1 1New Zealand 1 4 0 0.75 1 1 1

M. m. musculusNW Europe Jutland 9 40 1(1) 0.06 0 0 0C Europe Czech Republic 9 17 0.03 0 0 0.03 0

Slovakia 3 4 0 0 0 0 0Austria 1 1 0 0 0 0 0Slovenia 1 1 0 0 0 0 0

SE Europe Serbia 1 1 0 0 0 0 0Bulgaria 4 4 0 0.25 0 0 0Rumania 1 1 0 0 0 0 0

M. m. castaneusTaiwan 1 1 0 0 0 0 0Thailand 2 2 0 0 0 0 0

M. m. molossinusJapan 1 1 0 0 0 0 0

Inbred strainsAKR 1 0 1 1 1BALB/c 0 1 0 1 1C57BL/6 1 1 1 1 1DDK 1 1 1 1 1SJL 0 1 0 1 1

This insertion, which occurs in the region thatflanks 3' end of the B1 element, could be related tothe (CAAA) 3 repeat that follows the poly(A) tail ofthe repeat in the GenBank sequence. The size dif-ference is hardly perceptible when the intact am-plified fragments are compared, but is clearlydetectable in 2% agarose gels after digestion withTaql.

Polymorphism between classical inbredstrains. Five inbred strains were screened for thesame B1 insertion and restriction polymorphisms asthe wild mice populations. It can be seen in Table 2

that the three loci with B1 insertion polymorphismsin domesticus also reveal differences between theinbred strains we tested. Neither BALB/c or SJLcarried the insertion found in the loci Brd2 and Pa-faha-ps1, and the element in the locus Nktr is absentin AKR. However, all five strains carry the BI ele-ments that are nearly fixed in the domesticus pop-ulations (Btk and Tsx). The B1 insertion in Pirb,which is present in the inbred strain 129 but absentin all the wild derived mice we tested, is also absentin these strains.

The restriction haplotypes found in these strainsare given in Table 3, and it can be seen that

P. MUNCLINGER ET AL: B1 INSERTIONS IN MOUSE POPULATIONS 363

Msp I

^0 20 30 q0Blast query GCCGGGCATGGTGGCGCACGCCTTTAATCCCAGCACTTGGG-AGGCAGAB1 subfamily A . . . . G . . . . . .... . . . . . . . . . . . . . . . . . C . . - . . . . . . .consensus B . . . . . . . G . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C . . . - . . . . . . .

sequences D .......G .............................C... .......aPirb A .T....G .................................-.......a Brd2 B .......G .............................C...-.......a Nktrl A . .T....G . . . .... ... . .. . . .. . . ........ . .C. .

a Pafaha psl A . . .. . . . G . . . . . . . . . . T . . . . . . . . . . . . . . . . . . G . . . - . . . . . . .

a Tsx B A......G .................................-.......a Btk B&D .......G .............................C...G.......aErcc2 A .......G ............................AC...-.......b Psmb5 B .. TA. .TG .................................-.......b Mtm1 A .......G ..........T ..................C...-.......bUnp D .................................A...C...-.......blg(2r B .......G ............... .... .......C... - .......b Tpardl A ..T....G...........A .....................-.......bRenbp D .......G .................G...............-.......

Tag 1(1)

50 60 10 g0 q0Blast query GGCAG9CGGATTTCTGAGTTCGAGGCCAGCCTGGTCTACAAAGTGAGTTB 1 subfamily A . . . . . . .. . . . . . . . . . . . . . . . . . .. . . . . . .. . . . . . . . . .consensus B . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . .sequences D . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . G . . . . . . . .

a Pirb A . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . .a Brd2 B . . . . . . . . . . .. . . . . . . . . .. . . . . .......................a Nktr] A ......T ................................T.........a Pafaha psi A ......................... .................... T .. .

a Tsx B .A ...............................................a Btk B&D ........................................G........a Ercc2 A ......T .........................................GbPsmb5 B . ..... T . . ..... . ....... ... ...... ...... ... ........ .

b Mtml A ................................................GbUnp D . . . . . . . . . . . . . . . . . . . . G. C. . . . . . . . . . . . . . . . . G . . . . . . . .b Ifg2r B ......................... ................................................b Tpardl A ........................ . ........................bRenbp D .....................A..................G........

Tag 1(2)

100 >>0 \20 X30 140

Blast query CCAGGACAGCCAGGGCTACACAGAGAAACCCTGTCTCOBl subfamily A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AAAAACCAAAconsensus B ..................T.............................sequences D ................................................

a Pirb A ................................................3a Brd2 B ..................T.............................1a Nktrl A ................................................3a Pafahapsl A ................................................3a Tsx B ............A.A.C.T .............................6a Btk B&D ..................T........................A.C..2a Ercc2 A ....................................T..4bPsmb5 B ..................T.............................5bMtml A ................................................3b Unp D .............A.................C................6blfg2r B ..................T.................T.......A.C.Ib Tpardl A ........................ ........... .A .........G.3bRenbp D ................................................3

Fig. 1. The sequences of the 13 B1 elements studied aligned under the Blast query sequence. The consensus sequencescorresponding to the A, B, and D subfamilies are indicated. The positions of the restriction sites for Mspl and TagI areindicated above the sequence, and the sites containing the CpG dinucleotides in the subfamily consensus sequences areunderlined. The number at the end of each sequence corresponds to the number of changes that occur outside the sitesdefining the subfamilies, and a and b in the first column indicate respectively the B1 elements that are found only inM. musculus and those fixed in M. musculus, spretus, macedonicus, and spicilegus.

364

P. MUNCLINGER ET AL: BI INSERTIONS IN MOUSE POPULATIONS

Table 3. Frequency of the restriction haplotypes associated with the six B1 elements that are fixed in all the Mus speciestested. They are defined as the absence (a) or presence (p) of the MspI, Tagl(1) and Tagl(2) sites that occur in the B1consensus sequence. The additional sites in flanking sequences are given in italics.The number of different localitiesscreened in each species or subspecies is indicated in brackets. The presence of the insertion in the 5' flanking region ofRenbp indicated by ❑

Locus:haplotype:

Psmb

app apa

Mtm

ppp pap

Unp

pap aap

Igf2r

ppa apa aapp

Tpardl

aaap paap appp

Renbp

aapp pppp

M. m. domesticus (34) 0.94 0.06 1 0 0.94 0.06 1 0 0.95 0.05 0 0.81 ° 0.19 ° 0M. m. musculus (20) 0.04 0.96 0.15 0.85 1 0 0.35 0.65 0 1 0 0.91 0.09 0M. m. castaneus (2) 0.5 0.5 1 0 0 1 1 0 0 1 0 0 1 0M. m. molossinus (1) 0 1 0 1 0 1 0 1 1 0 0 0 1 0M. macedonicus (1) 0 1 1 0 1 0 1 0 0 0 1 1 0 0M. spicilegus (3) 0 1 1 0 1 0 0 1 0 1 0 0 0 1M. spretus (1) 1 0 0 1 1 0 1 0 0 0 1 0 1 0AKR 1 0 0 1 1 0 1 0 0 1 00 1° 0BALB 1 0 0 1 1 0 1 0 0 1 0 0 1° 0C57/B6 1 0 0 1 1 0 1 0 0 1 0 0 1°DDK 0 1 0 1 0 1 1 0 0 1 00 1° 0SJL 1 0 1 0 0 1 1 0 0 1 0 0 1° 0

differences also occur at three of these loci. At thelocus Psmb, DDK has the most frequent musculushaplotype, which is also found in molossinus andone of the castaneus strains tested. At the locusMtm, only SJL carries the haplotype fixed in do-mesticus, and all the others share the haplotype thatis very frequent in musculus and present molossi-nus. Unlike AKR, BALB/c, and C57BL/6, neitherDDK nor SJL shares the Unp haplotype that occursmost frequently in the European mice belonging toboth the domesticus and musculus subspecies. Nodifferences were found among the five strains at theother three loci, but it is interesting to note thatneither the Tardl nor Renbp haplotypes correspondto the most frequent domesticus ones.

Sequence comparisons. Figure 1 shows the 13B1 elements aligned under the B1 consensus se-quence used as the Blast query. This consensus isderived from 51 mouse B1 elements that were in-serted in the genome over a wide period of evolu-tionary time (Quentin 1989). The sequencescorresponding to the presumed progenitor sequencesof the recent A, B, and D subfamilies of B1 elements(Zietkiewicz and Labuda 1996) are also indicated.The number of changes, including insertions anddeletions, that occur outside the sites defining thethree subfamilies (positions 89 and 116 respectivelyfor the D and B subfamilies) are given at the end ofeach sequence. The Bi sequences that are fixed in allthe species tested are presumably older than thosefound only in M. musculus, so one might expect tofind a correlation between the time a given elementhas been present in the genome and its degree ofhomology relative to the consensus sequence of the

subfamily to which it belongs. However, we foundno significant difference in the ranges of nucleotidevariation between the fixed and unfixed groups ofrepeats (one to six variant positions). This overlapsuggests that it will be difficult to predict the pres-ence of an insertion polymorphism that is close tofixation within domesticus simply on the basis ofhomology to a subfamily consensus sequences.

Another feature of all the sequences we studiedis the presence of an (A) 5CC motif at their 3' ends.This motif, which was found by Zietkiewicz andLabuda (1996) to be frequent in the recent B1 se-quences in different rodent species, is not present inthe majority of the B1 sequences in the mouse ge-nome. It therefore appears to be labile, and its pres-ence could be a useful criterion for choosingcandidate loci to screen for polymorphic B1 inser-tions.

Discussion

This study shows that Bi elements that differ fromtheir presumed subfamily consensus sequence atbetween one and six positions (see Fig. 1) provide agood potential source of markers that can be used tostudy genetic variation in mouse populations be-longing to M. musculus. The 13 B1 sequences wereretrieved from a very small portion of the genome (atotal 2.8 Mb), and since we started this study, theamount of genome sequence available through thepublic mouse genome sequencing effort has in-creased enormously. As most of the B1 elements wetested show some degree of either insertion or re-striction polymorphism within M. musculus, sys-tematic genomic data mining for conserved B1

P. MUNCLINGER ET AL: BI INSERTIONS IN MOUSE POPULATIONS

365

sequences, combined with the rapid screeningstrategy detailed above, should now make it possibleto cover a large part of the genome. It should also bepossible to increase the number of informative re-striction polymorphisms associated with the B1 el-ements present in a given region of the genome byextending the search to slightly more diverged B1sequences and including more restriction enzymesin the screening protocol. For example, AvaI andN1aIII both have either a restriction site or a siteminus one nucleotide that includes other labile CpGdinucleotides in the B1 consensus sequence.

Most of the present-day laboratory inbred strainshave mosaic genomes that are mainly of domesticusorigin, but they also have contributions from theJapanese and Chinese mice of M. m. musculus,castaneus, and molossinus origin (Bonhomme et al.1987), the most notable example being the Asianvariant of the Y Chr that is carried by many of thesestrains (Bishop et al. 1985 ; Nagamine et al. 1992 ;

Tucker et al. 1992). All but one of the Bl elementsshowing insertion polymorphism in M. musculuswere essentially specific to domesticus, which is tobe expected for B1 elements that were all originallyfound in an inbred strain (most of the time, derivedfrom the strain 129). It is possible, however, that theB1 element present in the Pirb locus of 129, whichwas not found in any of the mice we tested, origi-nated in mouse populations from eastern Asia thatwere not well represented in the sample used in thisstudy.

There is an increasing demand for markers thatcan contribute to the understanding of the genealogyof the inbred mouse strains (Beck et al. 2000). Ourpreliminary results on just a few inbred strains sug-gest that loci containing fixed but recent Bl ele-ments are also likely to be a very useful source ofsuch markers. Not only do half of the six loci wetested for restriction polymorphism reveal differ-ences between strains, but they also provide cluesabout the subspecies origin of the gene carrying theinsertion. None of the strains in the panel we ana-lyzed (AKR, BALB/c, and C57BL/6 derived fromCastle's Mice ; SJL from the Swiss mice ; and DDKfrom old Japanese stock) carry the most frequentdomesticus haplotype at more than three of the lociwe tested (Table 3). If the haplotypes present in themajority of the strains are considered to be of do-mesticus origin, half of them would have to havebeen derived from domesticus mice carrying infre-quent variants. It seems rather unlikely that allthese 'atypical' haplotypes are of domesticus origin,and certain of them were probably drawn from farEastern populations of musculus, castaneus, ormolossinus mice. It is interesting to note that in

DDK the most frequent domesticus haplotype wasfound only at one locus (Igf2r). Females of this strainare almost completely infertile when crossed withother laboratory strains (Wakasugi et al. 1967),but are fully fertile in crosses with two strains ofAsian origin, CAS and MOM (Zhao et al. 2002). Ourresults suggest that, although this strain is consid-ered to be of mainly domesticus origin, contribu-tions from the Far Eastern subspecies might be quiteimportant.

Unlike the study of Kass et al. (2000), our searchfor recent B1 elements did not target the recent B1subfamilies, but used a sequence consensus derivedfrom a much wider range of B1 elements that doesnot represent any of these subfamilies. Even so, itcan be seen from Fig. 1 that all the sequences weretrieved possessed a nucleotide variant that definesone of recent A, B, or D subfamilies (Quentin 1989).Like Kass et al. (2000), we found sequences that canbe classed in the A and B subfamilies in both thegroup of B1 elements present only in M. musculusand in the group of fixed B1 elements. Our resultstherefore provide further evidence that differentprogenitors were active concomitantly in the lin-eages leading to M. musculus, M. spretus, M. spi-cilegus, and M. macedonicus. However, certainfeatures of the sequences we analyzed suggest thatthe B1 subfamily structure is more complex. Forinstance, the B1 element in the locus Btk that isfound only in domesticus carries the nucleotide po-sitions that define both the supposedly recent B andthe older D subfamilies. As we found this nucleotidecombination in two other fixed B1 sequences (un-published observations), it probably defines anotherrelated subfamily. Similarly, the G at position 97that is shared by the B1 sequences in Ercc2 andMtm 1 could be the signature of yet another variantsource copy that was active during the same period.The presence among the recently inserted Alu ele-ments in humans of a variety of small subfamiliesrelated to the major Alu subfamilies that have beenactive over a longer period of evolutionary time hasled to the suggestion that certain insertions acquirethe ability to retropose for a limited period of time(Roy-Engel et al. 2001). These act as source copiesthat generate small subfamilies of daughter copiescarrying the same sequence differences. Our datasuggest that a similar situation may exist in themouse, and the presence of unidentified subfamiliescould be one of the reasons why the estimated degreeof homology to the presumed consensus sequencedoes not discriminate between the sequences thatare polymorphic in M. musculus and the older B1elements that were inserted before the divergence ofM. musculus, spretus, macedonicus, and spicilegus.

366

P. MUNCLINGER ET AL: B 1 INSERTIONS IN MOUSE POPULATIONS

As B1 elements tend accumulate preferentiallyin the GC-rich areas of the genome, suitable B1 el-ements are not likely to be retrieved frequently fromthe AT-rich regions (Smit 1999). The Ll long inter-spersed repeats, which are another very abundantretroposon family in the mouse (Loeb et al. 1986), arepreferentially associated with the AT-rich isochores.Although complete L1 elements are too long to beused as practical markers, most of them are variablytruncated sequences that are incomplete reversetranscripts of the full RNA transcript. Several L1lineages have been active recently in the mouse ge-nome (Hardies et al. 2000), so it should be possible toapply a similar strategy to the one we used in thisstudy to identify insertion or restriction polymor-phisms associated with short Ll repeats.

Acknowledgments

The participation of P. Munclinger was supported bythe Grant Agency of the Czech Republic (Grant No.206/01/0989), the Grant Agency of the Czech Acad-emy of Sciences (Grant No. A6045902), and by theBarrande (Czech-French Bilateral Co-operation inResearch and Development) Project No. 99043. Thisstudy was partly supported by the CNRS GDR 1928(Evolution des genomes dans les populations).

References

1. Altschul SF, Madden TL, Schaffer AA, Zhang J, ZhangZ et al. (1997) Gapped BLAST and PSI-BLAST: a newgeneration of protein database search programs.Nucleic Acids Res 25, 3389-3402

2. Batzer MA, Arcot SS, Phinney JW, Alegria-Hartman M,Kass DH et al. (1996) Genetic variation of recentAlu insertions in human populations. J Mol Evol 42,22-29

3. Beck JA, Lloyd S, Hafezparast M, Lennon-Pierce M,Eppig JT et al. (2000) Genealogies of mouse inbredstrains. Nat Genet 24, 23-25

4. Bishop CE, Boursot P, Bonhomme F, Hatat D (1985)Most classical Mus musculus domesticus laboratorymouse strains carry a Mus musculus musculus Ychromosome. Nature 315, 70-72

5. Bonhomme F, Guenet J-L, Dod B, Moriwaki K, BulfieldG (1987) The polyphyletic origin of laboratory inbred

mice and their rate of evolution. Biol J Linn Soc 30,51-58

6. Hardies SC, Wang L, Zhou L, Zhao Y, Casavant NC etal. (2000) LINE-1 (L1) lineages in the mouse. Mol BiolEvol 17, 616-628

7. Kass DH, Raynor ME, Williams TM (2000) Evolution-ary history of B1 retroposons in the genus Mus. J MolEvol 51, 256-264

8. Loeb DD, Padgett RW, Hardies SC, Shehee WR, ComerMB et al. (1986) The sequence of a large L1Md elementreveals a tandemly repeated 5' end and several featuresfound in retrotransposons. Mol Cell Biol 6, 168-182

9. Nagamine CM, Nishioka Y, Moriwaki K, Boursot P,Bonhomme F et al. (1992) The musculus-type Y chro-mosome of the laboratory mouse is of Asian origin.Mamm Genome 3, 84-91

10. Quentin Y (1989) Successive waves of fixation of B1variants in rodent lineage history. J Mol Evol 28,299-305

11. Rogers JH (1985) The origin and evolution of retropo-sons. Int Rev Cytol 93, 187-279

12. Roy-Engel AM, Carroll ML, Vogel E, Garber RK,Nguyen SV et al. (2001) Alu insertion polymorphismsfor the study of human genomic diversity. Genetics159, 279-290

13. Rozen S, Skaletsky HJ (2000) Primer3 on the WWW forgeneral users and for biologist programmers. In: Kra-wetz S, Misener S, (eds.) Bioinformatics Methods andProtocols: Methods in Molecular Biology. (Totowa, NJ:Humana Press), pp 365-386

14. Shedlock AM, Okada N (2000) SINE insertions: pow-erful tools for molecular systematics. BioEssays 22,148-160

15. Smit AF (1999) Interspersed repeats and other me-mentos of transposable elements in mammalian ge-nomes. Curr Opin Genet Dev 9, 657-663

16. Tucker PK, Lee BK, Lundrigan BL, Eicher EM (1992)Geographic origin of the Y chromosomes in "old" in-bred strains of mice. Mamm Genome 3, 254-261

17. Ullu E, Tschudi C (1984) Alu sequences are processed7SL RNA genes. Nature 312, 171-172

18. Wakasugi N, Tomita T, Kondo K (1967) Differences offertility in reciprocal crosses between inbred strains ofmice. J Reprod Fertil 13, 41-50

19. Zhao WD, Ishikawa A, Yamagata T, Bolor H, WakasugiN (2002) Female mice of DDK strain are fully fertile inthe intersubspecific crosses with Mus musculus mol-ossinus and M. m, castaneus. Mamm Genome 13,345-351

20. Zietkiewicz E, Labuda D (1996) Mosaic evolution ofrodent B1 elements. J Mol Evol 42, 66-72