Isolation, cloning and characterization of two major satellite DNA families of rabbit (Oryctolagus cuniculus)

Gene 343 (2004) 271 – 279 www.elsevier.com/locate/gene

Isolation, cloning and characterization of two major satellite DNA families of rabbit (Oryctolagus cuniculus)B Csaba E´kes, Erika Csonka, Gyula Hadlaczky, Imre Cserpa´n* Institute of Genetics, Biological Research Center of the Hungarian Academy of Sciences, H-6701 Szeged, Temesva´ri krt. 62., P.O. Box 521, Hungary Received 23 March 2004; received in revised form 13 July 2004; accepted 23 September 2004 Available online 5 November 2004 Received by G. Pesole

Abstract We report here the isolation, cloning and characterization of two abundant centromeric satellite sequences (Rsat I and Rsat II) what are not related to each other, and that of a divergent subfamily (Rsat IIE) of rabbit (Oryctolagus cuniculus). The Rsat I monomers had a 375 base pair (bp) average length, while repeat units Rsat II and Rsat IIE were ~585 bp long. Variable amounts of Rsat I were detected by FISH at the centromeric region of 11 chromosome pairs of the complement. Rsat II hybridized to the centromere of 12 different chromosomes, and two of these were labeled also with the Rsat IIE probe. Two-color in situ hybridizations with the satellite probes and rDNA revealed that the NOR chromosomes carried different satellites. Rsat I was abundant on chromosome 20 and 21, but it was undetectable on chromosomes 13 and 16. Large Rsat II arrays were found on chromosomes 16, 20 and 21, but reduced amount was detected on chromosome 13. The variant Rsat IIE was prominent on chromosome 16, but was absent from the other rDNA-bearing chromosomes. The rDNA signal on chromosome 21 was localized to the 21q(ter) region, what can be a useful cytological marker in comparative cytological studies. These data show that rabbit chromosomes form at least four distinct groups based on the satellite composition of their centromeres. The differences in the chromosomal distribution of satellite families will help easy FISH identification of individual chromosomes, as well as to unveil the evolutionary history of the Leporidae karyotype. D 2004 Elsevier B.V. All rights reserved. Keywords: Chromosome; Heterochromatin; Tandem repeat; Centromere; Evolution

1. Introduction The genomes of higher eukaryotes harbor huge amounts of various types of repeated sequences. According to their organization, two major classes, i.e., interspersed and tandem repeats have been distinguished. Satellite DNAs, what can represent as much as 40% of the genome, consist of long tandem arrays of more or less well-defined repeat units. Except de novo formed neocentromeres (KeresI et al., 1996; Barry et al., 1999), all higher eukaryotic centromeres

Abbreviations: bp, base pair(s); p, plasmid; Rsat I, II, IIE, rabbit satellite DNA families; rDNA, DNA coding for rRNA. B AF498005-AF498016; AF527543-AF527548; AY114136AY114137. * Corresponding author. Tel.: +36 62 432 232; fax: +36 62 433 397. E-mail address: [email protected] (I. Cserpa´n). 0378-1119/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.gene.2004.09.029

examined so far have been found to consist of families of these highly repeated sequences (Tyler-Smith and Willard, 1993). Tandem-repeat centromeric DNA sequences have been characterized in a variety of nonprimate and primate mammals. Satellites from several Muridae species have been described; e.g., the mouse major and minor satellites (Wong and Rattner, 1988; Vissel and Choo, 1989; Gubatz and Cutler, 1990; Broccoli et al., 1991); the Acomys dimidiatus satellite (Kunze et al., 1999); that of Microtus chrotorrhinus (Modi, 1992, 1993), and a chromosome 2-specific centromeric repeat DNA from Chinese hamster (Fa´tyol et al., 1994). Both the Microtus and Chinese hamster sequence families have long (N2.5 kb) repeat units and they show characteristics of recently amplified arrays (Modi, 1993; Fa´tyol et al., 1994). In Ruminantia, sequences related to the 1.715 satellite of bovine (Gaillard et al., 1981; Plucienniczak et al., 1982) have

272

C. E´kes et al. / Gene 343 (2004) 271–279

been found in sheep (D’Aiuto et al., 1997), and are common in Artiodactyla (Modi et al., 1996). Studies of satellites from Bovini species with interrelated evolutionary histories led to the feedback model for the evolution of these DNA repeats (Nijman and Lenstra, 2001). However, both structurally and functionally, by far the best-characterized satellite sequences are those of primates, especially that of man (reviewed in Lee et al., 1997). The ~171 base pair (bp) a-satellite was found on all chromosomes of most primate species studied so far (Alexandrov et al., 2001). While different centromeric satellites have been described from various animal species, molecular analysis of key structural components of chromosomes of an important mammalian order (Lagomorpha) has been lagging behind. In spite of the fact that rabbit is one of the mammalian species frequently used in the laboratory and biotech industry, its genome, as well as that of the related species, has remained essentially untouched. Although high-resolution GTG-banding of rabbit chromosomes has been described more than a decade ago (Yerle et al., 1987), and recently, chromosome painting studies revealed basic evolutionary relationships within the order (Robinson et al., 2002), apart from a single effort (Ding, 1991), nothing has been published so far on the characterization of the DNA sequence forming the pericentromeric heterochromatin of domestic rabbit (Oryctolagus cuniculus). The present study was undertaken as the initial step of identifying the key repetitive DNA components of rabbit chromosomes. We report here cloning and characterization of two major centromeric satellite DNA families (Rsat I and Rsat II), and that of an Rsat II-related variant (Rsat IIE) from rabbit. Although these satellites do not provide a complete coverage of the complement, they are molecular markers for the majority of the rabbit chromosomes.

2. Materials and methods 2.1. Rabbit chromosomal DNA, construction of subgenomic libraries High molecular weight chromosomal DNA was isolated either from commercially available rabbit cell lines [SIRC, VX2 CCL-60 (cornea) or CRL-6504 (carcinoma) fibroblast cell line], or from the spleen, liver and peripheral blood lymphocytes of a sacrificed male domestic rabbit using standard procedures (Sambrook et al., 1989). Subgenomic libraries were constructed by cloning HindIII-, EcoRI-, or BglII-digested rabbit DNA in pUC19 cleaved with the appropriate enzyme. To achieve size fractionation prior to cloning, 10 Ag of the above digests was separated on 0.8% agarose gels and the ~0.4 and ~0.6 kb bands were cut out. DNA was electroeluted, extracted

with phenol and chloroform, then precipitated and dissolved in water. DNA was cloned in pUC19 cut with HindIII-, EcoRI-, or BamHI. Reaction products were then transfected into E. coli DH5a competent cells, and plated onto selective media. When size fractionation was omitted, 0.2 Ag aliquots of digested genomic DNA were used for cloning. White colonies were propagated, and plasmid DNAs were purified as described (Matthes et al., 1987) and insert lengths were determined. 2.2. DNA sequencing and data analysis All selected plasmid clones were sequenced on both strands with the standard M13 sequencing primers using Dye Terminator Sequencing Kit on an ABI model 365 automated DNA sequencer (Applied Biosystems). Assembly and proofreading of raw sequence data was performed with the DNAStar PC software. Occasionally, manual override was applied to remove conflicts between forward and reverse strand readings. Sequences of each clone were compared pairwise to all the rest with the Align two sequences (bl2seq) BLAST option of the NCBI web server, and those showing N75% identities along the whole insert were further analyzed. BLAST searches of the GenBank database with the rabbit queries were carried using the NCBI server. Multiple sequence alignments were performed at the INRA Multialign Interface homepage (Toulouse, France; Corpet, 1988). 2.3. Southern blot hybridization For genomic Southern hybridization, 10 Ag rabbit DNA was digested with restriction enzymes HindIII, BglII, Sau3AI, NsiI, HinfI, EcoRI and AseI, then separated on 0.8% agarose gels in duplicate panels. Gels were stained with ethidium bromide and photographed. After denaturation with alkali, gels were blotted onto Hybond N+ membrane (Amersham), and DNA was crosslinked to the filters by exposure to UV light, using standard procedures (Sambrook et al., 1989). Hybridization of 32P-labeled plasmid DNA of selected recombinants (Rsat I: pA06; Rsat II: pA43; Rsat IIE: pE2–3, respectively) to the blots was performed using standard techniques (Sambrook et al., 1989). High-stringency posthybridization washes were: 25 min in 2SSC and 0.1% SDS at room temperature, and 215 min in 0.1SSC and 0.1% SDS at 65 8C. 2.4. Fluorescent in situ hybridization Male rabbit peripheral blood lymphocyte metaphase spreads were obtained by the standard squash method, and fluorescent in situ hybridization was performed according to published procedures (Csonka et al., 2000). Satellite DNA clones pA06 (Rsat I), pA43 (Rsat II) and pE2/3 were labeled with DIG or biotin according to the instructions supplied

C. E´kes et al. / Gene 343 (2004) 271–279

273

dimers proved that bona fide satellite clones were being dealt with. Preliminary fluorescent in situ hybridization of the dimeric Rsat I and Rsat II probes to SIRC and VX cell line metaphases resulted in strong signals at the centromeric region of several chromosome pairs (not shown). The first group, Rsat I, had repeat unit lengths 373–380 bp. This family was found in seven monomeric clones derived from the HindIII library: pHind2, pA05, pA27, pA34, pA36, pA40 and pA42, respectively (GenBank accession nos. AF498006-AF498012). Clone pA06 carried a dimer (746 bp) with a mutated HindIII site between the two monomers (GenBank Acc. No. AF498005). In the BglII library, seven more monomeric Rsat I clones [pB09, pB18, pB19, pB59, pB65, pB70 and pB73 (GenBank accession nos. AF527543-AF527549)] were identified. Repeat unit length differences were mainly due to single nucleotide deletions/insertions without any particular pattern. Clones pA36 and pB18 had a tetranucleotide deletion, but at different locations. Multiple alignment of Rsat I sequences revealed that monomers A34, A06A, A06B and A27 were very similar, while clones A36 and B19 were the most divergent ones (not shown). Pairwise identities were calculated and are shown in Table 1. The second family, Rsat II, had repeat units between 585 and 590 bp. Sequence length differences were due to insertions/deletions of 1–3 nucleotides. No pattern could be observed in the distribution of these differences, probably because of the small number of sequences. Three monomers (pA02, pA29, pA37) and a dimer (pA43) were obtained from the HindIII subgenomic library (GenBank accession Nos. AF498013-AF498016). The two clones isolated from the EcoRI library, pE2/1 and pE2/3 (GenBank accession Nos. AY114136-AY114137), respectively, proved to be diverged variants of Rsat II. Furthermore, sequence alignments to HindIII monomers showed that the reiteration phase of the EcoRI repeats was shifted by 7 nucleotides upstream (Fig. 1). Therefore, these two clones were classified as a subtype of the

with the labeling kit. The rDNA probe (biotin- or DIGlabeled) was prepared from a cloned EcoRI fragment of the mouse pre-rRNA gene (I. Cserpan, unpublished; the clone spanned the region between nt. 9887 and 16498 of the GenBank entry with accession no. X82564). High-stringency washes were done by incubating the plates in 50% formamide/2SSC at 45 8C, five changes of washing solution for 1 min. [Low-stringency washes (50% formamide/4SSC at 35 8C, 51 min) gave essentially identical results to those obtained under high stringency and therefore, these conditions were not used apart from the first set of experiments.] Images were acquired through a CCD system mounted on an OLYMPUS Vanox microscope.

3. Results 3.1. Construction and screening of rabbit subgenomic libraries, identification of satellite DNA clones and their sequence analyses Subgenomic libraries were constructed from rabbit genomic DNA, digested with EcoRI, HindIII and BglII, with or without size fractionation. Due to the complete lack of sequence information on rabbit satellites, it was not possible to generate probes for hybridization screening of these libraries. Therefore, we had to resort to a less efficient procedure, i.e., plasmid DNAs were prepared from a total of 110 clones, and they were sequenced. Sequence readings were compared pairwise, and those showing N70% identities were further analyzed. Nucleotide sequence analyses yielded 21 satellite candidates (i.e., approximately 20% of the clones analyzed) after discarding interspersed repeats. These clones could be sorted into two groups, what were called Rsat I and Rsat II. One dimeric insert was also identified for both types, in which the monomer units were in head-to-tail arrangement. These Table 1 The calculated pairwise sequence identities of individual Rsat I clones is shown Rsat I

A05

A06A

A06B

A27

A34

A36

A40

A42

H2

B09

B18

B19

B59

B65

B70

B73

A05 A06A A06B A27 A34 A36 A40 A42 H2 B09 B18 B19 B59 B65 B70 B73

100 – – – – – – – – – – – – – – –

89.2 100 – – – – – – – – – – – – – –

87.8 93.7 100 – – – – – – – – – – – – –

88.4 92.1 91.5 100 – – – – – – – – – – – –

90.5 93.9 93.1 91.5 100 – – – – – – – – – – –

84.7 84.4 82.8 84.1 84.1 100 – – – – – – – – – –

92.4 91.8 89.4 89.4 90.8 83.9 100 – – – – – – – – –

91.0 91.8 91.3 91.3 92.1 85.4 91.6 100 – – – – – – – –

93.4 91.5 90.2 90.2 92.6 84.9 93.4 92.6 100 – – – – – – –

92.6 91.2 89.4 88.4 92.6 83.0 91.0 90.0 93.6 100 – – – – – –

89.9 92.6 91.3 90.5 94.7 82.8 88.9 90.8 90.8 91.8 100 – – – – –

80.1 80.9 80.9 81.5 81.2 77.1 80.0 81.9 80.6 81.0 80.2 100 – – – –

88.6 94.7 94.5 92.4 94.5 83.0 88.9 91.5 91.3 91.3 92.9 81.6 100 – – –

88.3 94.5 94.5 92.6 93.9 84.1 88.6 91.8 90.8 88.9 91.6 79.8 94.8 100 – –

86.8 90.0 88.7 88.2 91.8 80.2 87.0 87.6 89.2 92.4 91.0 78.6 90.8 88.7 100 –

88.5 89.8 89.0 85.6 92.7 82.1 88.7 87.7 88.5 92.1 90.6 78.2 90.4 88.5 88.4 100

Gaps were counted as mismatches.

C. E´kes et al. / Gene 343 (2004) 271–279

274

Fig. 1. Nucleotide sequence alignment of an Rsat II (AF498015) and one of the Rsat IIE (AY114136) monomers is shown. In order to obtain the best alignment, the Rsat IIE repeat unit was shifted 7 nucleotides upstream to that of Rsat II. This way, the two sequences were 80% identical.

Rsat II family, called Rsat IIE. When the Rsat II and Rsat IIE monomers were compared to each other, pairwise identities ranged between 77.8% and 92.0% (Table 2). Sequences of pE2/1 and pE2/3 showed the highest identity (92%), while the lowest level (77.8%) was observed between the 5V-end and 3V-end monomeric units of clone pA43. BLAST searches of the GenBank database with Rsat I, Rsat II and Rsat IIE sequences showed that no related sequences had been identified previously. Noteworthy is that

no CENP-B-like motifs (Muro et al., 1992) could be located in the rabbit satellite sequences. However, parts of the CENPB box sequence [CT(T/A)(C/T)G(T/G)TGGAAA(C/A)GG(G/A)A], or the less strictly defined 15 bp core element [TTCGNNNNANNCGGG] could be recognized: for example, the AAA(C/A)GGG element was found in approximately half of the repeat units.

Table 2 Pairwise comparison of sequence homologies of Rsat II clones

Southern blots prepared from rabbit genomic DNA digested with seven restriction enzymes were hybridized with Rsat I (clone pA06), Rsat II (clone pA43) and Rsat IIE (pE2/3) probes, respectively. The Rsat I probe detected clear ladders starting at the ~375 bp monomer unit, as well as its integer multiples, a pattern characteristic of tandem repeats, with HindIII, BglII and Sau3AI (Fig. 2a, lanes 1–3; note that the Sau3AI site is part of the BglII recognition sequence). With NsiI and HinfI, the strongest signal was that of the monomeric unit,

Rsat II

A02

A29

A37

A43A

A43B

E2-1

E2-3

A02 A29 A37 A43A A43B E2-1 E2-3

100 – – – – – –

89.8 100 – – – – –

91.5 90.6 100 – – – –

80.7 81.4 81.7 100 – – –

88.6 86.3 88.4 82.6 100 – –

80.4 78.8 81.1 78.9 78.5 100 –

80.3 79.5 81.8 79.5 78.4 92.0 100

Gaps were counted as mismatches, like in Table 1.

3.2. Genomic organization of Rsat I, Rsat II and Rsat IIE

C. E´kes et al. / Gene 343 (2004) 271–279

275

Fig. 2. Southern blot hybridization of 32P-labeled Rsat I, Rsat II and Rsat IIE is shown. Filters were probed with dimeric Rsat I clone pA06 (a), dimeric Rsat II clone pA43 (b) and Rsat IIE clone pE2/3 (c), respectively.

though half of the clones (pA05, pA27, pA34 and pA42, pB18, pB19, pB70 and pB73) had a second recognition site. Most of the signal remained in the unresolved region with EcoRI and AseI. The hybridization patterns shown in Fig. 2a did not indicate a significant proportion of higher-order repeats longer than dimers. Southern blot hybridization with Rsat II revealed a more complex pattern. Strong monomer-size (~0.58 kb) signals were detected in the HindIII, Sau3AI, EcoRI and AseI digests (Fig. 2b, lanes 1, 3, 6 and 7). However, bands with comparable intensities at dimer length were seen with HindIII, EcoRI and AseI (lanes 1, 6 and 7), indicating that part of Rsat II formed dimeric blocks. In contrast to Rsat I, NsiI and BglII cleaved only a small fraction of these arrays, but HinfI produced smaller-than-monomer fragments. Noteworthy is, that bands were detected out of register with respect to the ladder of the integer multiples of monomeric length with HindIII and AseI already below 2 kb, and with the exception of HinfI, in all other lanes N2 kb. The hybridization pattern with Rsat IIE resembled to that of Rsat II, but it was more regular (Fig. 2c). In the Sau3AI, EcoRI and AseI lanes signals corresponding to monomers were the strongest in them monomer-spaced ladders. In this case, HindIII appeared to cut less frequently, although similarly to the EcoRI and AseI digests, the band set out of register with that of multiples of monomer length was also seen it this lane. Digestion with HinfI yielded a major smaller-than-monomer band, together with two submonomeric fragments. Similarly to Rsat II, these arrays were relatively resistant to BglII and NsiI. The Rsat IIE probe also revealed the out-of-register band sets. 3.3. Chromosomal localization of the rabbit satellite DNA families and rDNA genes Fluorescent in situ hybridizations were carried out with the cloned satellite sequences, and their chromosomal

localizations were compared to that of ribosomal RNA gene clusters, as well as to each other. The result of the rDNA alone FISH is shown in Fig. 3a. In addition to the three NOR chromosomes (#13, #16 and #20) described previously (Yerle et al., 1987), a clear signal was seen at the tip of the long arm of chromosome 21 (21qter). When the Rsat I (red) probe was hybridized together with rDNA (green), strong satellite signals were observed at the centromeric regions of six out of the eleven chromosome pairs labeled (Fig. 3b). Both probes gave readily detectable signals on the two smaller rDNA-bearing chromosomes, #20 and #21, respectively. However, the two large NOR chromosomes (#13 and #16) remained unlabeled with Rsat I even under low-stringency conditions (not shown), and neither were chromosome X and Y. The other major family, Rsat II, hybridized to the centromeres of 12 chromosomes pairs (Fig. 3c, green) with variable intensities, including all chromosomes with rDNA. While on chromosome 16 (Fig. 3c, red rDNA probe), the Rsat II signal was clearly visible, it was barely detectable on chromosome 13. Two-colour FISH with the Rsat IIE probe pE2/3 (green) and rDNA (red) resulted in strong green signals at the centromeric regions of chromosomes #9 and #16, and the latter was also intensively labelled with the rDNA probe. Thus, the only NOR chromosome with Rsat IIE satellite subfamily was chromosome 16 (Fig. 3d). When Rsat I (red) and Rsat II (green) were hybridized together, 10 chromosome pairs showed dual labelling (Fig. 3e–h), including two of the rDNA-bearing chromosomes (#20 and #21). Both Rsat II and Rsat IIE probes hybridized to the centromeres of chromosomes #9 and #16 (Fig. 3i–j). Noteworthy is that no Rsat I signal was seen on these chromosomes. Comparative FISH analyses with Rsat I, Rsat II and Rsat IIE probes on lymphocytes of female and male lymphocytes revealed that none of these satellites were present on the sex chromosomes and seven somatic chromosome pairs (Table 3). In addition, hybridization

276

C. E´kes et al. / Gene 343 (2004) 271–279

Fig. 3. Fluorescent in situ hybridization of rDNA and satellite DNA probes to rabbit metaphase spreads. (a) rDNA only: chromosomes 13, 16, 20 and 21 were labeled; (b) rDNA (green)+Rsat I (red): both probes hybridized to chromosomes 20 and 21; (c) rDNA (red)+Rsat II (green): a relatively weak green signal was detected at the centromere of chromosome 16; (d) rDNA (red)+Rsat IIE (green): a strong Rsat IIE hybridization was seen to the chromosome 9 and 16 centromeres; (e–h) Rsat I (red)+Rsat II (green): cohybridization of the two probes was observed on 10 chromosome pairs; (f–g) Rsat II (red)+Rsat IIE (green): at the centromeres of chromosome 9 and 16, the Rsat II signal was seen to overlap with a substantially stronger Rsat IIE labelling.

C. E´kes et al. / Gene 343 (2004) 271–279 Table 3 Chromosomal distribution of rabbit satellites Rsat I, Rsat II, Rsat IIE and that of rDNA are shown FISH probe

Rsat I

Rsat II

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 X Y

++ +++ +

+ +

+++ ++

+++ +

++

Rsat IIE

rDNA chromosome

++

+

p ter

+

+++ +++ + +++ +++

+++ +++ + + ++ +++

+++

p ter

p ter q ter

The rDNA-bearing chromosomes differ in their satellite contents, and rDNA is localized at 21q(ter). +++ abundant, strong signal; ++ fainter, but usually detectable without signal amplification; + traces, signal amplification (digital or chemical) is necessary to detect.

signal intensities observed with a particular probe varied in a wide range on different chromosomes, indicating substantial array size polymorphism.

4. Discussion The pericentromeric heterochromatin of mammalian chromosomes is composed of tandemly repeated satellite DNAs, frequently forming arrays of several million nucleotides. Satellite compositions of centromere regions are variable, indicating that it is epigenetic modification, rather than the particular sequence of nucleotides, what supports centromere function (Henikoff et al., 2001). Therefore, overall sequence composition of centromeres can be very different even among closely related species. Conversely, even in a single species, e.g., in humans, variable amounts of several classes of unrelated satellites could contribute to forming pericentromeric DNA of individual chromosomes (Lee et al., 1997). Satellites have variable repeat unit lengths: in primates, the range is form the 5 bp monomers of satellite III to the 171 bp of a-satellite, what is known to form chromosomespecific higher-order repeat units (Lee et al., 1997), while in Chinese hamster, the CH2sat found on chromosome 2 was reported to have 2.8 kb monomeric blocks, organized into 375–440 kb reiterated domains (Fa´tyol et al., 1994).

277

The rabbit satellites described in this paper are arrays of medium-sized monomer units with family-specific features. They have not only different lengths (~375 and ~585 bp), but completely different sequences, as well. Individual Rsat I sequences showed differences in the 5.3–21.8% range (78.2–94.7% identities), and that of Rsat II/IIE was 8.0– 22.2% (77.8–92%; cf. Tables 1 and 2). These values are lower than the 10–40% divergence range found between monomers forming human a-satellite higher-order repeat units (Durfy and Willard, 1989). The rabbit satellites appear to be more homogenous than their human counterparts, owing to the less extensive higher-order repeat organization. However, elucidation of the large-scale organization of the individual satellite arrays will require further studies, including long-range mapping. FISH experiments with Rsat I, Rsat II and Rsat IIE showed, that the satellite sequences identified so far did not cover the complete chromosome complement of rabbit, although signals were detected with Rsat I on 11, with Rsat II to 12, and with Rsat IIE to 2 chromosome pairs (Table 3). It is remarkable that both Rsat I and Rsat II were found on nine chromosomes, but Rsat I was undetectable on the two chromosome pairs (#9 and #16) with Rsat IIE. Apparently, the two main satellite families have always been, or have become so different that they could coexist without much influence on each other. Probably, these two sequences are equally compatible with centromere function. The fact that the Rsat IIE-bearing chromosomes still carry Rsat II supports the view that probably the former arose from the local amplification of the latter. It has been shown recently that substantial shifts may occur in the relative proportions of different satellites as a consequence of evolutionary events (Mestrovic et al., 1998; Ugarkovic and Plohl, 2002). Processes in accordance with the blibraryQ hypothesis of satellite DNA sequence evolution (Mestrovic et al., 1998) appear to be at work in rabbit in two forms. First, both of the two main families, Rsat I and Rsat II, were detected together in varying ratios on a number of chromosomes (cf. Table 3), forming thus major components of the library. Second, the emergence of Rsat IIE on chromosomes #9 and #16 is the result of a relatively recent branching from Rsat II, thus creating new bvolumesQ. However, the catalogue of the rabbit satellite blibraryQ is still incomplete, since there are still nine chromosomes (seven somatic pairs and the sex chromosomes) for which no satellite has been characterized so far. The possibility that these chromosomes had some variant of either Rsat I or Rsat II seems unlikely, since no FISH signal was detected with any of these probes under low-stringency conditions. Therefore, some more work will be needed to identify, what kind of as yet sequences the heterochromatin of chromosomes devoid of the known satellites are composed of ? From evolutionary point of view, the precise relation of Rsat II and Rsat IIE deserves some consideration. These sequences are related to each other (Fig. 1); however, in addition to the differences in restriction site periodicities,

278

C. E´kes et al. / Gene 343 (2004) 271–279

they appear to have diverged sufficiently (approximately 20% difference) to be called as bsuprachromosomal familyQ along the same criteria applied for human a-satellite DNA (Alexandrov et al., 2001, Willard and Waye, 1996). Although both the Rsat II and Rsat IIE arrays are based on similar, but diverged repeat units, only Rsat II showed signs of higher-order periodicity (Fig. 2b). Therefore, it is plausible that the Rsat IIE arrays of these chromosomes formed via relatively recent large-scale amplification events of diverged repeat units, and consequently, they are expected to be more homogenous. Since the Southern hybridization pattern of Rsat IIE is more regular than that of Rsat II (cf. Fig. 2b and c), such putative saltatory amplification events could indeed have occurred. Taken together, it is plausible that Rsat IIE was derived from divergent Rsat II units, and not vice versa. Chromosome painting probes have been used recently to establish evolutionary relations within the order of Lagomorpha (Robinson et al., 2002). Comparative karyotype studies could also benefit from availability to cloned satellite DNA probes. Although in this paper, two dramatically different rabbit satellite families have been described, what cover two-thirds of the chromosome set, we are fully aware that there are at least six autosome pairs and the sex chromosomes, for which centromeric DNA sequences remain to be characterized. At present, the only reasonable statement regarding the nature of the satellites of the remaining chromosomes is that they must be quite different from those examined so far. The use of an rDNA hybridization probe showed that in addition to the three known NOR chromosomes (Yerle et al., 1987), there were signals at 21q(ter). This locus is as a useful cytological marker at least in some species of the Leporidae. While attempts are made to identify the missing satellites, those present on the NOR chromosomes could become components of rabbit satellite-based artificial chromosomes (Csonka et al., 2000). Acknowledgement We thank Dr. A. Udvardy, Dr. K. Fodor and Dr. I. Rasko´ for helpful comments and discussions on the manuscript. Publication of this paper was supported by the Ba´styaiHolzer Fund.

References Alexandrov, I., Kazakov, A., Tumeneva, I., Shepelev, V., Yurov, Y., 2001. Alpha-satellite DNA of primates: old and new families. Chromosoma 110, 253 – 266. Barry, A.E., Howman, E.V., Cancilla, M.R., Saffrey, R., Choo, K.H.A., 1999. Sequence analysis of an 80 kb human neocentromere. Hum. Mol. Genet. 8, 217 – 227. Broccoli, D., Trevor, K.T., Miller, O.J., Miller, D.A., 1991. Isolation of a variant family of mouse minor satellite DNA that hybridizes preferentially to chromosome 4. Genomics 10, 68 – 74.

Corpet, F., 1988. Multiple sequence alignment with hierarchical clustering. Nucleic Acids Res. 16, 10881 – 10890. Csonka, E., Cserpa´n, I., Fodor, K., Hollo´, Gy., Katona, R., J., KeresI, Praznovszky, T., Szaka´l, B., Telenius, A., deJong, G., Udvardy, A., Hadlaczky, Gy., 2000. Novel generation of human satellite DNAbased artificial chromosomes in mammalian cells. J. Cell. Sci. 113, 3207 – 3216. D’Aiuto, L., Barsanti, P., Mauro, S., Cserpan, I., Lanave, C., Ciccarese, S., 1997. Physical relationship between satellite I and II DNA in centromeric regions of sheep chromosomes. Chromosome Res. 5, 375 – 381. Ding, J., 1991. Cloning and sequencing of a highly repeated DNA fragment of rabbit. Zhongguo Yixue Kexueyuan Xuebao 13, 241 – 245. Durfy, S.J., Willard, H.F., 1989. Patterns of intra- and interarray sequence variation in alpha satellite from the human X chromosome: evidence for short-range homogenization of tandemly repeated DNA sequences. Genomics 5, 810 – 821. Fa´tyol, K., Cserpa´n, I., Praznovszky, T., KeresI, J., Hadlaczky, Gy., 1994. Cloning and molecular characterization of a novel chromosome specific centromere sequence of Chinese hamster. Nucleic Acids Res. 22, 3728 – 3736. Gaillard, C., Doly, J., Bernardi, G., 1981. The primary structure of bovine satellite 1.715. Nucleic Acids Res. 10, 6069 – 6082. Gubatz, J.W., Cutler, R.G., 1990. Mouse satellite DNA is transcribed in senescent cardiac muscle. J. Biol. Chem. 265, 17753 – 17758. Henikoff, S., Ahmad, K., Malik, H.S., 2001. The centromere paradox: stable inheritance with rapidly evolving DNA. Science 293, 1098 – 1102. KeresI, J., Praznovszky, T., Cserpa´n, I., Fodor, K., Katona, R., Csonka, E., Fa´tyol, K., Hollo´, G., Szeles, A., Ross, A.R., Sumner, A.T., Szalay, A.A., Hadlaczky, G., 1996. De novo chromosome formations by largescale amplification of the centromeric region of mouse chromosomes. Chromosome Res. 4, 226 – 239. Kunze, B., Traut, W., Garagna, S., Weichenhan, D., Redi, C.A., Winking, H., 1999. Pericentric satellite DNA and molecular phylogeny in Acomys (Rodentia). Chromosome Res. 7, 131 – 141. Lee, C., Wevrick, R., Fisher, R.B., Ferguson-Smith, M.A., Linc, C.C., 1997. Human centromeric DNAs. Hum. Genet. 100, 291 – 304. Matthes, H.W., Staub, A., Chambon, P., 1987. The segmented paper method: DNA synthesis and mutagenesis by rapid microscale bshotgun gene synthesisQ. Methods Enzymol. 154, 275 – 287. Mestrovic, N., Plohl, M., Mravinac, B., Ugarkovic, D., 1998. Evolution of satellite DNAs from the genus Palorus-experimental evidence for the blibraryQ hypothesis. Mol. Biol. Evol. 15, 1062 – 1068. Modi, W.S., 1992. Nucleotide sequence and genomic organization of a tandem satellite array from the rock vole Microtus chrotorrhinus (Rodentia). Mamm. Genome 3, 226 – 232. Modi, W.S., 1993. Rapid, localized amplification of a unique satellite DNA family in the rodent Microtus chrotorrhinus. Chromosoma 102, 484 – 490. Modi, W.S., Gallagher, D.S., Womack, J.E., 1996. Evolutionary histories of highly repeated DNA families among the Artiodactyla (Mammalia). J. Mol. Evol. 42, 337 – 349. Muro, Y., Masumoto, H., Yoda, K., Nozaki, N., Ohasi, M., Okazaki, T., 1992. Centromere protein B assembles human centromeric alphasatellite DNA at the 17-bp sequence, CENP-B box. J. Cell Biol. 116, 585 – 596. Nijman, I.J., Lenstra, J.A., 2001. Mutation and recombination in cattle satellite DNA: a feedback model for the evolution of satellite DNA repeats. J. Mol. Evol. 52, 361 – 371. Plucienniczak, A., Skowronski, J., Jaworski, J., 1982. Nucleotide sequence of bovine 1.715 satellite DNA and its relation to other bovine satellite sequences. J. Mol. Biol. 158, 293 – 304. Robinson, T.J., Yang, F., Harrison, W.R., 2002. Chromosome painting refines the history of genome evolution in hares and rabbits (order Lagomorpha). Cytogenet. Genome Res. 96, 223 – 227. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning. CHS Laboratory Press.

C. E´kes et al. / Gene 343 (2004) 271–279 Tyler-Smith, C., Willard, H.F., 1993. Mammalian chromosome structure. Curr. Biol. 3, 390 – 397. Ugarkovic, D., Plohl, M., 2002. Variation in satellite DNA profiles—causes and effects. EMBO J. 21, 5955 – 5959. Vissel, B., Choo, K.H., 1989. Mouse major (gamma) satellite DNA is highly conserved and organized into extremely long tandem arrays: implications for recombination between nonhomologous chromosomes. Genomics 5, 407 – 414. Willard, H.F., Waye, J.S., 1996. Chromosome-specific subsets of human alpha satellite DNA: analysis of sequence divergence within and

279

between chromosomal subsets and evidence for an ancestral pentameric repeat. J. Mol. Evol. 25, 207 – 214. Wong, A.K., Rattner, J.B., 1988. Sequence organization and cytological localization of the minor satellite of mouse. Nucleic Acids Res. 16, 11645 – 11661. Yerle, M., Echard, G., Gilois, M., 1987. The high-resolution GTG-banding pattern of rabbit chromosomes. Cytogenet. Cell Genet. 45, 5 – 9.