A non-hypervariable human minisatellite strongly stimulates in vitro intramolecular homologous recombination1

J. Mol. Biol. (1998) 278, 499±505

COMMUNICATION

A Non-hypervariable Human Minisatellite Strongly Stimulates in Vitro Intramolecular Homologous Recombination Francisco BoaÂn, Jose Manuel RodrõÂguez and Jaime GoÂmez-MaÂrquez* Departamento de BioquõÂmica y BiologõÂa Molecular, Facultad de BiologõÂa, Universidad de Santiago, 15706 Santiago de Compostela, La CorunÄa Galicia, Spain

Several features indicate that the low polymorphic human minisatellite MsH42 region could be involved in recombination. It contains different well-known recombination motifs, is able to generate single-stranded loops and is speci®cally recognized by nuclear proteins. These characteristics led us to investigate the possible recombinogenic activity of the MsH42 region in terms of intramolecular recombination. We constructed two plasmids, one of them carrying two copies of the minisatellite region and the other one containing sequences upstream of this repetitive region. We showed that MsH42 strongly stimulates intramolecular in vitro recombination, 22 times more than the control sequence, solely when the source of biological extract is mouse testes, suggesting that MsH42 could be a hotspot involved in meiotic recombination. Furthermore, there is a direct relationship between the frequency of equal crossovers and the enhancement of recombination. Interestingly, the third repeat of the minisatellite array is always involved in the resolution of unequal crossovers leading to minisatellite shortening. As far as we know, our results provide the ®rst evidence that a non-hypervariable minisatellite can enhance homologous recombination. # 1998 Academic Press Limited

*Corresponding author

Keywords: human repetitive DNA; non-hypervariable minisatellite; homologous recombination; equal crossover; in vitro recombination assay

Homologous recombination is a process of genetic information exchange between stretches of homologous DNA that generates new combinations of alleles favoring the plasticity and, therefore, the evolution of genomes. Recombination events are required for the segregation of chromosomes at the ®rst division of meiosis in most organisms (Hawley, 1989) as well as for repairing some types of DNA damage (Resnick, 1976; Szostak et al, 1983). The rate of genetic exchange between two loci during homologous recombination is roughly proportional to the physical distance that separates them. Nevertheless, some chromosomic regions are prone to undergo a much higher recombination rate than others. These sites, that originate a disparity between genetic and physical maps, are known as recombination hotspots. An important requirement of the recombination models (Meselson & Radding, 1975; Radding, 1982; Szostak et al., 1983, Smith & Stahl, 1985) is the ability of DNA sequences to form loop structures with 0022±2836/98/180499±07 $25.00/0/mb981714

single-stranded conformation that could invade homologous double-stranded DNA. In this sense, some minisatellite DNA sequences can misalign tandem repeats in vitro via strand-slippage, resulting in the formation of single-stranded DNA loops (Coggins et al., 1992). On the other hand, another main requisite of recombinational models is that the DNA sequence is able to interact with speci®c nuclear proteins. Related to this, in nuclear extracts from several species minisatellite-speci®c DNAbinding proteins have been detected (Collick & Jeffreys, 1990; Wahls et al., 1991; Yamazaki et al., 1992; Shinder et al., 1994), some of them with a single-stranded binding activity (Collick et al., 1991; Yamazaki et al., 1992). There are several lines of evidence indicating that some hypervariable minisatellite sequences can function as recombinogenic hotspots for the initiation of homologous recombination. Thus, there is a similarity between the consensus core sequence of some minisatellites (Jeffreys et al., 1985) and the well-known prokaryotic Chi recom# 1998 Academic Press Limited

500 bination hotspot (Smith et al, 1981; Smith & Stahl, 1985). Moreover, the presence has been reported of hypervariable minisatellite sequences at the meiotic recombinational hotspot within the mouse major histocompatibility complex locus (Steinmetz et al., 1986) as well as in meiotic chiasmata of human chromosomes (Chandley & Mitchell, 1988). Other indications that this type of repetitive sequence can promote recombination are the enhancement of homologous recombination shown by a synthetic hypervariable minisatellite sequence (Wahls et al., 1990), and the existence of hypervariable minisatellites whose polymorphism is generated by complex recombination events (Jeffreys et al., 1994; Buard & Vergnaud, 1994). These and other ®ndings strongly support a role for hypervariable minisatellites in recombinational events. Nonetheless, whether or not minisatellites that are not hypervariable could participate in the enhancement of recombination remains unknown. We have recently reported the structural characterization of a new human DNA sequence that contains a low polymorphic G ‡ C-rich minisatellite termed MsH42 (BoaÂn et al., 1997). This minisatellite has different recombination signals within and surrounding it and is very similar to immunoglobulin regions involved in class-switch recombination. MsH42 undergoes slipped-strand mispairing, indicating its ability to generate singlestranded loops. Furthermore, it is speci®cally recognized by nuclear proteins present in extracts prepared from human cells and mouse tissues (BoaÂn et al., 1997). Altogether these ®ndings point to a possible role of this repetitive sequence in recombination events. Recombinational activity of the MsH42 region To investigate the hypothetical capacity of MsH42 to act as a hotspot for homologous recombination, we constructed the plasmid pMsH42Lac carrying two copies in the same orientation of a fragment that includes the MsH42 region; between these copies we cloned the lacZ gene to facilitate the identi®cation of recombinants. The control plasmid p50 MsH42Lac was constructed in the same way but using two copies of a sequence located upstream of the minisatellite region. In Figure 1 the construction of both plasmids is explained. The intramolecular recombination of pMsH42Lac and p50 MsH42Lac, in the presence of nuclear extracts as a source of recombinogenic proteins, should produce the excision of the lacZ gene from the original constructs. This allows an easy identi®cation of recombinant products, since they generate lacZÿ/ampr colonies after transformation of lacZÿ Escherichia coli cells. In contrast, the lacZ‡ recombinant products excised from pMsH42Lac and p50 MsH42Lac cannot propagate in E. coli because they lack a replication origin (Figure 2(a)). On the other hand, since in vivo and in vitro experiments revealed differences in eukaryotic hotspot activity between meiotic and mitotic cells (Kell & Roeder,

MsH42 Region Enhances Intramolecular Recombination

1984; Edelmann et al., 1989), we compared the ability of MsH42 to stimulate in vitro recombination using either somatic (mouse liver and human NC37 cells) or germinal extracts (mouse testes). In the recombination experiments, equal amounts of pMsH24Lac and p50 MsH42Lac were incubated together with each of the nuclear extracts employed in this study. After incubation, DNAs were extracted and used to transform E. coli DH5a cells. The identi®cation of homologous recombination products was carried out by digestion with PstI. With this enzyme, the original pMsH42Lac generates six DNA restriction fragments (two of them identical), whereas the control plasmid yields four restriction fragments. However, the digestion of the recombination products derived from pMsH42Lac or p50 MsH42Lac produces three or two PstI restriction fragments, respectively (Figure 2(b)). The outcome of several in vitro intramolecular homologous recombination experiments, each of them carried out three times with a different nuclear extract (mouse testes, mouse liver, or human NC-37 cells), is summarized in Table 1A. Our data indicate that the different nuclear extracts ef®ciently catalyzed the homologous recombination reactions. Remarkably, when the recombination assay was carried out with the mouse testis nuclear extracts, the intensity of homologous recombination between the two copies of the minisatellite region was 22 times higher than that obtained with the control sequence. However, neither mouse liver nor human NC-37 nuclear extracts showed enhancement of the recombination (Table 1B). According to this, the MsH42 region could be a hotspot involved in meiotic homologous recombination. Nevertheless, whether or not this sequence stimulates meiotic versus mitotic recombination will require further experimentation. Crossover analysis The intramolecular recombination in the pMsH42Lac produced equal and unequal crossovers, whereas all recombinant products generated from the control plasmid were always as the original sequence (Figure 2(b)). The recombinant products from equal crossover showed a PstI pattern exhibiting a 832 bp DNA fragment that contained the entire minisatellite region. However, the unequal crossover produced variations in the size of that DNA fragment as a consequence of MsH42 shortening or growing. Figure 2(b) shows the PstI restriction pattern of the most representative recombinant variants that we have found. Furthermore, the size of the minisatellite region in the recombinant products, derived from equal and unequal crossovers, was con®rmed by PCR ampli®cation (Figure 2(c)). After restriction analysis of all pMsH42Lac recombinant products (794 lacZÿ colonies) obtained with the different nuclear extracts (Table 1), we observed that the amount of recombinants showing minisatellite growing or

MsH42 Region Enhances Intramolecular Recombination

shortening was approximately the same (data not shown). We found a rare recombinant product with a PstI pattern corresponding to a recombinational event leading to minisatellite growing (Figure 2(b), lane 4). Unexpectedly, although the PCR-ampli®ed product of this recombinant should be longer than 585 bp (the ampli®cation size of the original minisatellite region), this DNA fragment has a size

501 (480 bp; Figure 2(c), lane 5) that corresponds to the short allele of this locus (BoaÂn et al., 1997). Direct sequencing of this PCR product con®rmed its identity with the short allele. The divergency between the results obtained from the PstI digestion and the PCR ampli®cation suggests a complex recombination origin. One possible explanation is that this product originated from a deletion in the minisatellite (generating a sequence identical with the short

Figure 1. Generation of the plasmid constructs used in the recombination experiments. (a) Schematic representation of the H42 sequence showing the localization of the fragments employed in the recombination assays (recombination control fragment and the minisatellite region). (b) Construction of pMsH42Lac, a plasmid containing two copies of the minisatellite MsH42 region. We subcloned an 880 bp EcoRI-HindIII fragment derived from pRep42 into pBR322, a plasmid that includes MsH42 as well as its 50 and 30 proximal DNA ¯anking sequences (BoaÂn et al., 1997). With this cloning we obtained a new plasmid called pMsH42.1. The 880 bp fragment was cloned again in the same orientation downstream of the tetr gene generating plasmid pMsH42.2. Between both EcoRI-HindIII fragments, we cloned the lacZ gene obtaining the ®nal plasmid construct called pMsH42Lac. The steps for the synthesis of the control plasmid p50 MsH42Lac were similar to those described for pMsH42Lac. This control plasmid contains two cloned copies of an 850 bp EcoRI-BglII fragment located upstream from the MsH42 region. P1 and P2 represent the primers utilized in the PCR ampli®cations (BoaÂn et al., 1997). A, AvaI; B, BglII; D, DdeI; E, EcoRI; H, HindIII; P, PstI; X, XbaI. Plasmids are not shown to scale.

502

MsH42 Region Enhances Intramolecular Recombination

allele) and a duplication in the ¯anking region of MsH42. Interestingly, there is a direct relationship between the enhancement of recombination and the frequency of equal crossovers. In nuclear extracts that show strong stimulation of the recombinational activity (mouse testes), the frequency of equal crossovers was the highest (Table 1). It seems that when the biological system recognizes the MsH42 region as a recombinogenic point, an ef®cient pairing is established between the two homologous regions of the plasmid construct. If

this happened in vivo, it would provide an explanation for the low polymorphism detected in the MsH42 region (BoaÂn et al., 1997). In other words, the exact pairing would maintain the structure of this locus in the human population. The existence of unequal crossover products and the presence of different repeats in the MsH42 sequence opened up the possibility of knowing the region where the intramolecular exchanges took place during homologous recombination. Sequence analysis demonstrated the conservation of the minisatellite array in the recombinant products

Figure 2 (legend opposite)

503

MsH42 Region Enhances Intramolecular Recombination Table 1. In vitro intramolecular homologous recombination Extract A. Data collection Mouse testes Mouse liver NC-37 cells Extract

Experiments

lac Z‡ (non-recombinants)a pMsH42Lac p50 MsH42Lac Total

3 3 3

5275 4540 5381 pMsH42Lac (%)

B. Recombination frequencies Mouse testes Mouse liver NC-37 cells

11.27 2.15 0.44

5865 4130 4809 p50 MsH42Lac (%) 0.51 2.01 0.47

11,140 8670 10,190 Totalc 5.91 2.09 0.44

lac Zÿ (recombinants)b pMsH42Lac p50 MsH42Lac 670 100 24

30 85 23

Total 700 185 45

Enhancement ratiod

Equal crossover frequency (%)e

22.09(3.2) 1.07(0.5) 0.92(0.3)

85.10 43.75 44.80

a The pMsH42Lac and p50 MsH42Lac non-recombinants were distinguished by hydridization with a MsH42 probe as indicated in the legend to Figure 2(a). b The pMsH42Lac and p50 MsH42Lac recombinants were distinguished by restriction enzyme analysis as described in the legend to Figure 2(b). c The total recombination frequency represents the ratio between the total recombinants and non-recombinants. d Enhancement ratio represents the rate between pMsH42Lac and p50 Ms42Lac recombination frequencies. e The equal crossover frequency was calculated as the ratio of pMsH42Lac crossovers in total pMsH42Lac recombinants.

derived from equal crossovers (Figure 3(a)). In the molecules that originated from unequal recombinant events there were deletions or duplications that modi®ed the tandem repeat organization of MsH42 (Figure 3(b) and (c)). When the minisatellite is shortened, the deletion always takes place between the repeat number 3 and another one of the same type (Figure 3(b)). These results indicate that this repeat is a cardinal point for crossover resolution. It is worth noting that the third repeat of

the MsH42 array was also involved in the synthesis of several fragments generated by slippedstrand mispairing (BoaÂn et al., 1997). Likewise, the sequence analysis of recombinant products derived from minisatellite growing (Figure 3(c)) revealed that the exchange was again between repeats implicated in the slippage behaviour of MsH42. These coincidences suggest that the single-stranded loops generated by slippage mispairing, could be employed as invasive molecules between both

Figure 2. Rationale of the intramolecular homologous recombination and analysis of the recombination products. (a) The pairing between the homologous sequences of pMsH42Lac and the generation of intramolecular recombination products is shown. When an equal crossover occurs, the recombinant products have the entire minisatellite sequence, whereas if the crossover is unequal the minisatellite undergoes growing or shortening. The reasoning for the intramolecular recombination with the control fragment, depicted as an open box with the name 50 MsH42, was the same. Nuclear extracts were prepared from mouse tissues (testes and liver) and human NC-37 cells as described (Dignam et al., 1983). These extracts, with a protein concentration of 5 to 10 mg/ml measured by the Bradford (1976) method, were frozen under liquid N2 and stored at ÿ80 C. The in vitro recombination reactions were performed according to Edelmann et al. (1989). The reactions were carried out in a ®nal volume of 100 ml containing 20 mM Tris-HCl (pH 7.5), 10 mM MgSO4, 1 mM ATP, 0.1 mM of each dNTP, 500 ng of plasmid DNA (250 ng of each construct) and different amounts (5 to 20 mg) of nuclear proteins. After incubation for 30 minutes at 37 C, the DNA was extracted with saturated phenol and used to transform E. coli DH5a cells (lacZÿ, recAÿ). The bacteria were plated onto agar plates containing Xgal and isopropyl-b-D-thiogalactopyranoside as lacZ gene indicators. To differentiate between the pMsH42Lac and p50 MsH42Lac non-recombinants, the lacZ‡ colonies were blotted onto Hybond-N‡ nylon membranes (Amersham) and hybridized with the 32P-labeled pRep42 insert as described by GoÂmez-MaÂrquez et al. (1985). The white colonies (lacZÿ) were used for minipreparation of plasmid DNA following standard procedures (Sambrook et al., 1989). The lacZÿ colonies due to mutations in the lacZ gene were identi®ed by restriction analysis and tetracycline resistance and made up less than 1% of the total lacZÿ colonies. The number of the lacZÿ colonies in reactions without nuclear extract or with a heat-inactivated extract (15 minutes, 100 C) was about 2  10ÿ5. (b) PstI analysis of the recombinant products. Restriction analysis was utilized to distinguish the recombinants originating from either pMsH42Lac or p50 MsH42Lac. Aliquots of 300 ng were digested with PstI, separated on a 1.5% (w/v) agarose gel and stained with ethidium bromide. Lanes: 1, pMsH42Lac; 2, p50 MsH42Lac; 3 to 8, different variants of pMsH42Lac recombinant products; 9, p50 MsH42Lac recombinant product; M, size marker (BstEII fragments of l DNA). Lane 3 represents a product derived from equal crossover and lanes 5 to 8 represent products from unequal crossovers. Lane 4 represents a product originating from a complex recombination event. (c) PCR analysis of the recombination products. For PCR ampli®cation of the MsH42 sequence in the recombination products, we used the conditions described by BoaÂn et al. (1997). The DNA template (1 ng) was obtained from minipreps of plasmid DNA. The ampli®ed fragments were fractionated on 3% Metaphor (FMC) agarose gels and stained with ethidium bromide. Lanes: 1, positive control corresponding to genomic DNA from a heterozygous individual; 2, ampli®cation product of the plasmid containing the MsH42 region (pRep42); 3, p50 MsH42Lac recombinant product; 4 to 9, pMsH42Lac recombinant products arranged in the same manner as in lanes 3 to 8 (b). M, size marker (100 bp ladder). Arrowheads indicate the long and the short alleles.

504 duplexes. Such invasion of single-stranded DNA into a duplex is central in the recombination models (Szostak et al., 1983; Meselson & Radding, 1975).

MsH42 Region Enhances Intramolecular Recombination

Concluding remarks Here we have shown that a non-hypervariable human minisatellite is able to enhance in vitro

Figure 3. Sequence analysis of the recombination products. (a) Schematic representation of the sequence of a recombinant product derived from an equal crossover. The crossover resolution point cannot be assigned due to the exact pairing between the MsH42 sequences. (b) Representation of three unequal crossover events that produced minisatellite shortening. The sequence of the different recombinant products demonstrated that in all cases the third repeat is involved in the resolution of the recombination event. (c) Representation of an unequal crossover event which provokes minisatellite growing. For the sequence analysis, PCR products obtained after ampli®cation of recombinant plasmids were separated on a 1.5 % agarose gel, electroeluted and sequenced with the PCR product sequencing kit (Amersham). Five different clones of each recombinant type were sequenced.

MsH42 Region Enhances Intramolecular Recombination

intramolecular homologous recombination. Until now little attention has been paid to these kind of minisatellites. Our results provide an explanation for the low polymorphism observed in the MsH42 locus (BoaÂn et al., 1997), since the frequency of equal crossovers is highest when the recombination is truly enhanced. The low polymorphism in this and other similar loci might be re¯ecting the importance of their biological role, perhaps, in recombinational events.

References BoaÂn, F., GonzaÂlez, A. I., RodrõÂguez, J. & GoÂmezMaÂrquez, J. (1997). Molecular characterization of a new human minisatellite that is able to form single stranded loops and is recognized by nuclear proteins. FEBS Letters, 418, 251± 257. Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of proteing utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248± 254. Buard, J. & Vergnaud, G. (1994). Complex recombination events at the hypermutable minisatellite CEB1 (D2S90). EMBO J. 13, 3203± 3210. Chandley, A. C. & Mitchell, A. R. (1988). Hypervariable minisatellite regions are sites for crossing over in man. Cytogenet. Cell Genet. 48, 152± 155. Coggins, L. W., O'Prey, M. & Akhter, S. (1992). Intrahelical pseudoknots and interhelical associations mediated by mispaired human minisatellite DNA sequences in vitro. Gene, 121, 279± 285. Collick, A. & Jeffreys, A. J. (1990). Detection of a novel minisatellite-speci®c DNA-binding protein. Nucl. Acids Res. 18, 625± 629. Collick, A., Dunn, M. G. & Jeffreys, A. J. (1991). Minisatellite binding protein Msbp-1 is a sequencespeci®c DNA-binding protein. Nucl. Acids Res. 19, 6399± 6404. Dignam, J. D., Lebowitz, R. M. & Roeder, R. G. (1983). Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucl. Acids Res. 11, 1475±1489. Edelmann, W., KroÈger, B., Goller, M. & Horak, I. (1989). A recombination hotspot in the LTR of a mouse retrotransposon identi®ed in an in vitro system. Cell, 57, 937± 946. GoÂmez-MaÂrquez, J., Puga, A. & Notkins, A. L. (1985). Regions of the terminal repetitions of the Herpes simplex virus type 1 genome. J. Biol. Chem. 260, 3490± 3495. Hawley, R. S. (1989). Exchange and chromosomal segregation in eukaryotes. In Genetic Recombination

505 (Kucherlapati, R. & Smith, G. R., eds), pp. 497± 527, American Society for Microbiology, Washington DC. Jeffreys, A. J., Wilson, V. & Thein, S. L. (1985). Hypervariable minisatellite regions in human DNA. Nature, 314, 67 ±73. Jeffreys, A. J., Tamaki, K., MacLeod, A., Monckton, D. G., Neil, D. L. & Armour, J. A. (1994). Complex gene conversion events in germline mutation in human minisatellites. Nature Genet. 6, 136± 145. Kell, R. L. & Roeder, G. S. (1984). Cis-acting recombination-stimulating activity in a fragment of the ribosomal DNA of S. cerevisiae. Cell, 39, 377±386. Meselson, M. S. & Radding, C. M. (1975). A general model for genetic recombination. Proc. Natl Acad. Sci. USA, 72, 358± 361. Radding, C. M. (1982). Homologous pairing and strand exchange in genetic recombination. Annu. Rev. Genet. 16, 405± 437. Resnick, M. A. (1976). The repair of double-strand breaks in DNA: a model involving recombination. J. Theoret. Biol. 59, 97 ± 106. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Shinder, G. A., Manam, S., Ledwith, B. J. & Nichols, W. W. (1994). Minisatellite DNA-binding proteins in mouse brain, liver and kidney. Exp. Cell Res. 213, 107± 112. Smith, G. R. & Stahl, F. W. (1985). Homologous recombination promoted by Chi sites and RecBC enzyme of Escherichia coli. BioEssays, 2, 244±249. Smith, G. R., Kunes, S. M., Schultz, D. W., Taylor, A. & Triman, K. L. (1981). Structure of chi hotspots of generalized recombination. Cell, 24, 429± 436. Steinmetz, M., Stephan, D. & Lindahl, K. F. (1986). Gene organization and recombinational hotspots in the murine major histocompatibility complex. Cell, 44, 895± 904. Szostak, J. W., Orr-Weaver, T. L., Rothstein, R. J. & Stahl, F. W. (1983). The double-strand-break repair model for recombination. Cell, 33, 25 ± 35. Wahls, W. P., Wallace, L. J. & Moore, P. D. (1990). Hypervariable minisatellite DNA is a hotspot for homologous recombination in human cells. Cell, 60, 95 ±103. Wahls, W. P., Swenson, G. & Moore, P. D. (1991). Two hypervariable minisatellite DNA binding proteins. Nucl. Acids Res. 19, 3269 ±3274. Yamazaki, H., Nomoto, S., Mishima, Y. & Kominami, R. (1992). A 35-kDa protein binding to a cytosine-rich strand of hypervariable minisatellite DNA. J. Biol. Chem. 267, 12311± 12316.

Edited by M. Yaniv (Received 7 November 1997; received in revised form 9 February 1998; accepted 11 February 1998)