Genetics, Vol. 157, 63-77, January 2001, Copyright © 2001
Crossing Over Is Rarely Associated With Mitotic Intragenic Recombination in Schizosaccharomyces pombe
Jeffrey B. Virgina,b,
Jeffrey P. Baileya,
Farnaz Hasteha,
James Nevillea,
Amy Colea, and
Gerard Trompb
a Department of Pathology, Wayne State University and the Barbara Ann Karmanos Cancer Institute, Detroit, Michigan
b Center for Molecular Medicine and Genetics, Wayne State University and the Barbara Ann Karmanos Cancer Institute, Detroit, Michigan
Corresponding author:
Jeffrey B. Virgin, Department of Pathology, University of Washington, VA Medical Ctr., 1660 S. Columbian Way, Seattle, WA 90108., jvirgin{at}u.washington.edu (E-mail)
Communicating editor: L. S. SYMINGTON
 | ABSTRACT |
|---|
Chromosomal rearrangements can result from crossing over during ectopic homologous recombination between dispersed repetitive DNA. We have previously shown that meiotic ectopic recombination between artificially dispersed ade6 heteroalleles in the fission yeast Schizosaccharomyces pombe frequently results in chromosomal rearrangements. The same recombination substrates have been studied in mitotic recombination. Ectopic recombination rates in haploids were 
1–4 x 10-6 recombinants per cell generation, similar to allelic recombination rates in diploids. In contrast, ectopic recombination rates in heterozygous diploids were 2.5–70 times lower than allelic recombination or ectopic recombination in haploids. These results suggest that diploid-specific factors inhibit ectopic recombination. Very few crossovers occurred in ade6 mitotic recombination, either allelic or ectopic. Allelic intragenic recombination was associated with 2% crossing over, and ectopic recombination between multiple different pairing partners showed 1–7% crossing over. These results contrast sharply with the 35–65% crossovers associated with meiotic ade6 recombination and suggest either differential control of resolution of recombination intermediates or alternative pathways of recombination in mitosis and meiosis.
RECOMBINATION between homologous DNA duplexes serves both common and different purposes in mitotic and meiotic cells. The most basic common function is the repair of broken chromosomes. In mitotic cells, homologous recombination provides a means for restoring chromosomes to their original state after accidental breaks during replication or segregation, or from exogenous insults. In meiosis, physiological chromosome breaks have been demonstrated in two distantly related species of yeast, Schizosaccharomyces pombe (CERVANTES et al. 2000 
) and Saccharomyces cerevisiae (reviewed in LICHTEN and GOLDMAN 1995 ). In S. cerevisiae, there is substantial evidence that these breaks and their repair are part of a developmental program in meiosis that combines pairing of homologous chromosomes and recombination to promote genetic diversity and ensure proper chromosome segregation. The evolutionary conservation of the SPO11 gene, which encodes a nuclease that appears to be responsible for the meiotic double-strand breaks in fungi and animals, suggests that this developmental program is very widespread in nature (KEENEY et al. 1997 ; CELERIN et al. 2000 , and references therein).
The consequences of homologous recombination are not always beneficial to the cell. There are examples of at least two mechanisms for adverse outcomes. First, reciprocal exchanges or crossing over between dispersed repeated sequences (ectopic recombination) can lead to chromosomal rearrangements or abnormal meiotic segregation. By sequence analysis, multiple examples of germline chromosomal rearrangements that lead to human diseases appear to result from ectopic recombination (reviewed in LABUDA et al. 1995 ). Also, somatic rearrangements with exchange points within a repetitive sequence have been identified in human leukemia cells (SCHICHMAN et al. 1994 ). In fungi, reciprocal ectopic exchanges in meiosis lead to abnormal chromosome segregation (JINKS-ROBERTSON et al. 1997 ). Second, in mitosis, crossing over between homologous chromosomes during the postreplication phase of the cell cycle can lead to homozygosis of a large portion of a chromosome arm. This may be a prevalent mechanism for inactivation of tumor suppressor genes in human cancers (reviewed in TISCHFIELD 1997 ).
Several factors could contribute to a balance favoring the advantageous functions of homologous recombination and limiting the adverse outcomes. Restoration of a broken chromosome coupled with limited risk to other portions of the genome is promoted most effectively by nonreciprocal recombination, whereas genetic diversity is promoted most effectively by reciprocal recombination. Therefore, the different purposes of mitotic and meiotic homologous recombination could be achieved by differentially controlling the balance between gene conversion and crossing over. Mechanisms may have evolved to limit reciprocal exchanges in mitotic homologous recombination (RICHARDSON et al. 1998 ). In meiosis, a balance must be achieved between the beneficial effects of crossing over between homologous chromosomes and the potentially deleterious effects of ectopic reciprocal exchanges. In fungi, studies of ectopic recombination between naturally and artificially repeated sequences have suggested that the balance may be achieved in part by limiting reciprocal exchanges between naturally repeated sequences, but this inhibition does not extend to artificially repeated sequences (VIRGIN and BAILEY 1998 , and references therein). Another mechanism that may limit ectopic recombination is limitation of chromosomal movement within the nucleus, so that dispersed repetitive sequences are unable to physically interact. Pairing of homologous chromosomes appears to limit meiotic ectopic recombination in S. cerevisiae (GOLDMAN and LICHTEN 1996 ) and may affect mitotic recombination as well (WEINER and KLECKNER 1994 ; BURGESS and KLECKNER 1999 ; BURGESS et al. 1999 ). In mammalian cells, chromosomes appear to be restricted to specific domains of the interphase nucleus (MANUELIDIS 1990 ; NAGELE et al. 1995 ).
To better understand the mechanisms that limit chromosomal rearrangements, we have established a system for studying ectopic recombination between artificially dispersed sequences in S. pombe. Different alleles of the ade6 gene have been integrated at different sites on all three S. pombe chromosomes (VIRGIN et al. 1995 ; VIRGIN and BAILEY 1998 ). In meiosis, 35–60% of the ectopic ade6 recombinants show an associated crossover, resulting in reciprocal translocations (VIRGIN and BAILEY 1998 ). Approximately the same proportion of allelic ade6 recombinants shows exchange between flanking markers in meiosis (GRIMM et al. 1994 ). In combination, these results suggest that there is overlap in the mechanisms that control and execute meiotic allelic and ectopic recombination. In this study we have analyzed the same ectopic recombination reactions in mitotic cells. In contrast to meiotic recombination, there are very few crossovers associated with mitotic intragenic recombination. These results suggest that crossing over is suppressed in mitotic recombination, thereby limiting loss of genetic information or chromosomal rearrangements in somatic cells.
| MATERIALS AND METHODS |
|---|
S. pombe strains and recombination rates:
Strain genotypes are listed in Table 1. Growth of S. pombe strains and heterothallic matings were performed as previously described (PONTICELLI et al. 1988 ; PONTICELLI and SMITH 1992 ). Construction and characterization of strains carrying single ectopic integrations of ade6 alleles have been previously reported (VIRGIN et al. 1995 ; VIRGIN and BAILEY 1998 ). By matching the DNA sequences of the plasmid-chromosomal junctions with the S. pombe genomic sequence database (http://www.sanger.ac.uk/Projects/S_pombe/), we have more precisely characterized the insertion sites. As reported previously (VIRGIN and BAILEY 1998 ), at each insertion site some chromosomal DNA was deleted, most likely in conjunction with the integration event. The z15 integration occurred in the region of chromosome I represented by the cosmid clone c2F7 and includes a deletion of 626 bp of chromosomal DNA between nucleotide positions 22,609 and 23,235. Sequence analysis of this region does not detect any genes (http://www.sanger.ac.uk/Projects/S_pombe/). The z7 integration occurred in the region of chromosome II represented by the bacteriophage P1 clone p8B7 and includes a deletion of 1995 bp of chromosomal DNA between nucleotide positions 66,214 and 68,209. This deletion includes the entire open reading frame of the puc1 gene, which encodes a G-type cyclin. No other genes are detected in this region. The only reported phenotype of puc1 mutants is an accelerated exit from mitosis and entry into meiosis after nitrogen starvation (FORSBURG and NURSE 1994 ). puc1 mutants show normal mitotic cell cycle kinetics. On the basis of several facts we contend that the puc1 mutation in our strains does not affect mitotic recombination: first, the conditions of our assay did not include nitrogen starvation and therefore the only known phenotype would not be expressed; second, most recombination reactions involving the z7 integration gave results similar to other strains in which the puc1 gene was intact.
Haploid strains carrying two different ade6 alleles at different loci were created by mating appropriate strains and isolating spore colonies on the basis of formation of Ade+ papillae after replica-plating to minimal media. After construction of the original strains, some of the recombination products from reactions between the ade6 insertions at z7 or z15 and ura4 showed deletion of the ade6 insertion from the ura4 locus. This is most likely due to formation of hybrid DNA at the two ends of the ura4 gene with the ade6 gene forming an extruded heterologous loop (see Fig 1A and Fig B). Therefore, in strain JV550 the z15 integration was revised so that the structure was identical to the ura4 insertion (Fig 1C).
Nonsporulating diploid strains were constructed using the h90mat2-B102 allele. Strains with the genotype h90mat2-B102 will conjugate with either h+ or h- strains, but h-/h90mat2-B102 diploids fail to sporulate and grow as vegetative diploids (EGEL 1984 ). Closely linked his4 and lys4 heterozygous markers were used to maintain selection for diploids. During strain construction it was discovered that diploids homozygous for the ade6-D1 deletion (PONTICELLI and SMITH 1992 ) were unstable. Therefore, a new smaller deletion (ade6-D19) was created by deleting a 1887-bp DraI fragment spanning the entire ade6 coding region and covering 184 bp of the 5'-flanking region and 48 bp of the 3'-flanking region. It was initially constructed in a plasmid and then inserted into the chromosome by homologous transplacement. The structure of the chromosomal deletion was confirmed by Southern blot. ade6-D1/ade6-D19 heterozygous diploids and ade6-D19 homozygous diploids are stable.
Mitotic recombination rates were determined by the method of the median (LEA and COULSON 1949 ). Exponential growth phase cultures were diluted to <10 cells per 2-ml culture in Edinburgh minimal medium 2 (EMM2; NURSE 1975 ) plus adenine, followed by growth to saturation density. Cells were plated on EMM2 agar with appropriate additives to determine the number of viable cells and the number of Ade+ recombinants. Diploids showed lower saturation densities than haploids; therefore, the entire culture was plated for diploids and one-half the culture was plated for haploids. The number of recombinants and total cells for haploids was corrected to reflect the entire culture. For each rate measurement, the median number of recombinants (r0) was determined from 10–24 independent cultures. Multiple groups of 3–5 independent cultures were analyzed at different times, and the median number of recombinants from the entire group of cultures was used to estimate m, the mean number of recombination events per culture. The recombination rate per cell division (a) is m/n, where n is the mean number of cells per culture.
Due to low recombinant frequencies in the two diploid strains JV490 and JV516, m was estimated by the proportion of cultures without recombinants (p0), where m = -ln p0 (LURIA and DELBRUCK 1943 ). The rate of prototroph formation per cell division was calculated as above, and the standard deviation was calculated according to LEA and COULSON 1949 . Likewise, reversion rates in monoallelic haploid strains and homozygous diploid strains were measured by the p0 method. Statistical analysis of differences between recombination rates included pairwise comparisons by a two-tailed z-test with Bonferroni correction for multiple comparisons (SNEDECOR and COCHRAN 1989 ).
Assays for chromosomal rearrangements:
The proportion of crossovers among Ade+ recombinants was determined by replica-plating nonselected haploids or diploids and isolating independent Ade+ papillae from separate colonies. Ectopic crossovers were assayed in haploids by test crosses and Southern blotting using the genius nonradioactive blotting procedure (Boehringer Mannheim, Indianapolis), as previously reported for meiotic recombinants (VIRGIN and BAILEY 1998 ), and were assayed in diploids by Southern blotting. A PCR assay for translocations was also developed (Fig 3). Genomic DNA was prepared as for Southern blotting (VIRGIN et al. 1995 ). Oligonucleotide primers complementary to unique DNA sequences flanking each insertion site were used in a PCR reaction with Taq DNA Polymerase (Fisher Scientific, Itasca, IL) with conditions recommended by the manufacturer. Cycle parameters were as follows: 95° for 30 sec, 57° for 30 sec, and 72° for 3 min for 40 cycles followed by a 10-min extension at 72°. The products were 3 kb in length. Thermal cycle DNA sequencing of recombinant alleles was performed with the BigDye ready reaction mix (Perkin-Elmer, Foster City, CA) and analyzed on an ABI PRISM 377 DNA Sequencer (ABI Advanced Biotechnologies, Columbia, MD).
View larger version (33K):
In this window
In a new window
Download PPT slide
|
Figure 3.
PCR assay for crossovers. (A) Recombinants from the ade6-M375 x zzz7-3::ade6-469 reaction (JV548) were analyzed for crossovers by PCR and agarose gel electrophoresis. Four different PCR reactions with different primer pairs were used to analyze each recombinant. Two examples are shown. Template DNA for the left four lanes was from a nonreciprocal recombinant and for the right four lanes was from a reciprocal recombinant (determined by test crosses and Southern blot). PCR was performed as described in MATERIALS AND METHODS with the primers indicated. The PCR products are 3 kb in length. (B) The structures of the loci being amplified. The primer annealing sites 1–4 are indicated by small arrows. Alleles A and B are parental for flanking primer sites and alleles C and D have exchanged flanking primer sites. Diagrams are as in Fig 1.
|
|
| RESULTS |
|---|
A system for studying ectopic recombination in S. pombe:
We have recently completed a study of meiotic ectopic recombination between artificially dispersed ade6 alleles in S. pombe (VIRGIN and BAILEY 1998 ). The ectopic recombination substrates included the endogenous ade6 gene and artificially dispersed ade6 alleles, integrated either at random on chromosomes I and II (z15 and z7, respectively) or by homologous transplacement at the ura4 locus on chromosome III. The integrations contain 3 kb of transplaced genomic DNA, including the ade6 transcriptional unit and flanking DNA and some associated bacterial plasmid DNA (Fig 1). At each site different single base pair ade6 mutations were inserted separately. In this study, we have created haploid and diploid strains carrying two different ade6 alleles at the same or different loci to study mitotic allelic and ectopic intragenic recombination, respectively.
Mitotic recombination rates were determined by the method of the median (LEA and COULSON 1949 ). To ensure that the Ade+ prototrophs arose by recombination rather than by simple reversion of the single base pair mutations, we measured the rates of prototroph formation in ancestral haploid strains carrying a single ade6 allele and in homozygous ade6-M375 and ade6-469 diploid strains. In most cases too few revertants were obtained to calculate the rates by the method of the median; therefore the proportion of cultures without mutants was used to calculate the reversion rates (LURIA and DELBRUCK 1943 ). For each strain, 5–19 cultures were assayed. The p0 values ranged from 0.74 to 1.0 for haploids and 0.5 to 0.8 for diploids. Excluding one haploid strain that produced no revertants, the calculated reversion rates ranged from 6.6 x 10-9 ± 6.6 x 10-9 to 4.0 x 10-8 ± 1.8 x 10-8 revertants per cell division. The precision of the reversion rates is low due to the few revertants obtained, and the comparison between the reversion and recombination rates suffers from the caveat that they were calculated by different methods. In spite of these caveats, the results demonstrate that, with two exceptions (see below), the monoallelic haploid and homozygous diploid reversion rates were substantially lower than heteroallelic recombination rates. Pairwise comparisons showed that most monoallelic reversion rates were reduced 10- to 100-fold relative to heteroallelic rates, and, in all but two comparisons, the differences were statistically significant (P values ranged from 2.4 x 10-13 to 1.5 x 10-2). The two exceptions were heteroallelic diploid strains JV490 and JV516, in which the low rates of prototroph formation were not statistically different from reversion rates in monoallelic haploid strains and homozygous diploid strains (see below). Thus, with these two exceptions, the rates of Ade+ prototroph formation measured in heteroallelic strains reflect ectopic recombination rates.
Chromosomal location and ploidy affected mitotic recombination:
The dispersal of different ade6 alleles to multiple chromosomal sites allowed the study of chromosomal position effects on both allelic and ectopic recombination reactions. The allelic recombination rates varied only slightly at different chromosomal positions. The same recombination reaction (ade6-M375 x ade6-469) was studied at four different loci, and the recombination rates varied over a 2.6-fold range, from 1.3 to 3.4 x 10-6 recombinants per cell division (Table 2C). This is similar to the 3.6-fold range for the same recombination reactions that we reported previously for meiotic allelic recombination (VIRGIN and BAILEY 1998 ). Thus, chromosomal position did not have a major effect on allelic recombination rates for these four loci in either mitotic or meiotic recombination.
More substantial chromosomal position effects were observed in ectopic recombination. At least two different aspects of chromosomal location affected the recombination rates (Table 2; Fig 2). First, allelic recombination rates were 2.5- to 70-fold higher than ectopic recombination rates in diploids (1.3–3.4 x 10-6 vs. 0.46–5.2 x 10-7 recombinants per cell division for allelic and ectopic recombination, respectively). For all pairwise comparisons, the differences between ectopic and allelic recombination were statistically significant (P values ranged from 8.5 x 10-8 to 4.6 x 10-2). This result suggests that interactions between allelic positions on homologous chromosomes are favored over interactions between nonhomologous chromosomes in S. pombe vegetative cells. Second, the z7 x ura4 reaction in diploid strain JV516 showed a 10-fold reduction relative to the z7 x z15 and z15 x ura4 reactions in vegetative diploids (Table 2; Fig 2). This reduction was similar to the 8- to 13-fold reduction for the same comparisons previously observed in meiosis (VIRGIN and BAILEY 1998 ). The low rate of prototroph formation in JV516 (4.6 x 10-8 ± 1.6 x 10-8 prototrophs per cell division; Table 2) was not significantly different from the rates in the two monoallelic haploid strains carrying either the zzz7-2::ade6-M375 allele (4.0 x 10-8 ± 1.8 x 10-8 prototrophs per cell division; P = 1.0) or the ura4-155::ade6-469 allele (1.5 x 10-8 ± 1.1 x 10-8 prototrophs per cell division; P = 0.63). This indicates that events other than recombination may contribute substantially to prototroph formation in JV516. The comparison of these results in diploids during vegetative growth with the results from meiotic recombination (VIRGIN and BAILEY 1998 ) indicates conservation of chromosomal position effects in S. pombe during mitosis and meiosis.
There were several differences observed between haploids and diploids in mitotic recombination rates (Table 2; Fig 2). First, in haploid cells the chromosomal position effect on the z7 x ura4 reaction disappeared; the recombination rates for the z7 x ura4 reaction were similar to the z7 x z15 and z15 x ura4 reactions. Second, the recombination rates for all haploid strains were higher than the corresponding diploid strains, with increases of 2.5- to 8-fold, not including the exceptionally higher z7 x ura4 reaction. The differences between haploid and diploid rates in pairwise comparisons were statistically significant (P values ranged from 3.8 x 10-10 to 4.5 x 10-2). Recombination rates were compared between opposite allele configurations of the same reaction in haploids, with only slight differences ranging from 1.1- to 2.7-fold (Table 2). In only one comparison did the different rates between opposite allele configurations achieve statistical significance (JV576 compared to JV548; P = 0.03). The comparison of ectopic recombination between haploids and diploids, for both mitotic and meiotic recombination, suggests that homologous chromosome interactions and/or some other feature of diploidy, such as mating-type heterozygosity, play an important role in limiting ectopic recombination in S. pombe.
Chromosomal rearrangements were rarely associated with mitotic ectopic recombination:
Two different ectopic recombination reactions (ade6 x z7 and ura4 x z15) generated viable reciprocal translocation products, as determined by three independent assays. Initially, test crosses were used to identify recombinants that showed "pseudolinkage" of the recombining loci as an indication of reciprocal recombination (VIRGIN and BAILEY 1998 ). A subset of recombinants, with and without pseudolinkage, were further analyzed by either a Southern blot (data not shown) or a PCR assay (Fig 3), or both. In the PCR assay, primers were designed that annealed to sites flanking the ade6 DNA at each locus. By using different combinations of primers, reactions specific for parental (primer sets A and B) and crossover (primer sets C and D) alleles were performed on template DNA from each of a subset of recombinants (Fig 3). With only a few exceptions, there was agreement between the test crosses, Southern blots, and PCR assays for the recombinants tested. The exceptions were six recombinants from JV548 that showed pseudolinkage, but two of these showed a pattern of nonreciprocal recombination on the Southern blot (two parental-size bands), and four showed one parental-size allele and one allele with an uncharacterized rearrangement. The nature of these recombinants has not been further characterized and they were excluded from the analysis. Some of them could have arisen by a mechanism such as break-induced replication (BOSCO and HABER 1998 ), which would result in one parental and one crossover allele. Because of the segmental aneuploidy that would result from this type of event with the recombination substrates used in this study, cell survival following such events would require additional rearrangements to preserve a complete genome.
The frequency of chromosomal rearrangements due to crossing over in ectopic recombination was much less in mitosis than previously reported for meiosis (Table 3; VIRGIN and BAILEY 1998 ). The comparisons of mitotic and meiotic crossovers were made without correction for loss of recombinants due to aneuploid segregation because the correction factor for mitosis is uncertain. Aneuploid segregation is possible when recombination occurs after DNA replication, as in meiosis. In contrast, mitotic recombination is likely to generate less aneuploidy because it can occur either before or after DNA replication (reviewed in PETES et al. 1991 ). Therefore, the correction factor is likely to be less for mitosis than for meiosis, and the data presented in Table 3 represent the minimal differences between mitotic and meiotic recombination. In mitosis, in two separate experiments there was an average of 4% crossovers in the ade6 x z7 recombination reaction, whereas in meiosis 24% of events from the same recombination reaction were associated with crossovers. Similarly, 3% of ura4 x z15 recombination events in haploids were associated with crossovers in mitosis, compared to 22% in meiosis. These differences were highly statistically significant (Table 3). In the diploid strain JV515, among 27 independent ura4 x z15 recombinants analyzed by Southern blotting, two crossover recombinants (7%) were identified. In this case, a comparison with the same recombination reaction in meiosis did not show a statistically significant difference; however, relatively few diploid mitotic recombinants were analyzed. These results indicate a suppression of crossing over in ectopic recombination in mitosis as compared to meiosis.
The high proportion of crossovers previously observed with meiotic intragenic recombination is similar in ectopic and allelic ade6 recombination (GRIMM et al. 1994 ; VIRGIN and BAILEY 1998 ). We wanted to determine if the low proportion of crossovers in mitotic recombination was specific to ectopic recombination, or if crossing over was also suppressed in allelic recombination. We used homozygosis of the recessive tps16-23 marker located 7 cM centromere-distal to the endogenous ade6 locus (GRIMM et al. 1994 ) as a measure of intergenic recombination or crossing over in the cenIII-tps16 interval. Approximately 2% of the mitotic intragenic ade6 recombinants from JV583 expressed the recessive tps16-23 mutation (Table 3). This low level of mitotic crossing over in the cenIII-tps interval represents a marked reduction compared to the 52–65% crossovers associated with meiotic intragenic ade6 recombination in the shorter ura4+-aim-tps16 interval (GRIMM et al. 1994 ).
Expression of the recessive tps16-23 mutation could occur as a result of chromosome loss rather than homozygosis. In S. cerevisiae, chromosome nondisjunction is associated with mitotic recombination (CHUA and JINKS-ROBERTSON 1991 , and references therein). Also, in S. pombe diploids the rate of spontaneous haploidization has been reported as 2 x 10-4 events per cell generation (BODI et al. 1991 ). However, we consider this an unlikely explanation for our results for several reasons. First, the possibility of loss of all three chromosomes through haploidization was excluded by flow cytometric measurement of DNA content in a subset of 17 Ade+ recombinants from JV583. All 17, including the three recombinants that expressed the tps16-23 mutation (Table 3), showed diploid DNA contents (data not shown). Second, diploids monosomic for chromosome III are unlikely to be stable since aneuploidy in general is unstable in S. pombe (NIWA et al. 1986 ). Also, in experiment 2 with JV583 (Table 3) no selection was applied for Lys+ or Ura+, which were heterozygous markers on chromosomes II and III, respectively; therefore if random chromosome loss occurred, it would be expected that some Ade+ recombinants would become Ura- or Lys-. However, such recombinants were not detected. These results support the conclusion that expression of tps16-23 was due to homozygosis and not chromosome loss.
It is possible that the few intergenic recombinants associated with mitotic ade6 intragenic recombination were incidental events that were not mechanistically related to the intragenic recombination event. If this were true, then the cenIII-tps16 recombinant frequency should be similar between unselected cells and Ade+ recombinants. Therefore, we made a direct comparison of the frequency of tps16-23 homozygosis among total viable cells and among Ade+ recombinants from the same culture of strain JV583. In two separate experiments, the cenIII-tps16-23 recombinants were enriched 25- and 100-fold among Ade+ prototrophs compared to unselected cells (Table 4). This result strongly supports the conclusion that the cenIII-tps16 intergenic recombination events reflect crossing over associated with ade6 intragenic recombination and that crossing over is suppressed in mitosis relative to meiosis for both ectopic and allelic recombination.
To further characterize the few crossovers observed in mitotic recombination, we determined if they were associated with, or independent of, conversion of the intragenic markers. Both products of the reciprocal recombination events could be recovered, because with the loci being studied balanced translocations are required for viability in haploid cells (VIRGIN and BAILEY 1998 ). This allowed a direct assessment of the association between gene conversion and crossing over in mitotic recombination. The crossover-specific PCR reactions with primer sets C and D (Fig 3) were used to isolate the reciprocal products from six independent ade6-M375 x zzz7::ade6-469 mitotic crossover recombinants, and, for comparison, eight ade6-M26 x zzz7::ade6-469 meiotic crossover recombinants from the previous study (VIRGIN and BAILEY 1998 ). The allele structures of the PCR products were determined by DNA sequencing. A simple crossover not associated with conversion would produce a wild-type ade6 allele and a double mutant allele, whereas in crossovers associated with conversions the double mutant allele would not occur. In both mitosis and meiosis, the majority of the crossovers examined were associated with conversions (Fig 4). Among the eight meiotic recombinants examined, one showed both a crossover and a conversion that were separated by a nonconverted allele. This could be due to independent events, or, if the conversion and crossover were associated, then either a lack of mismatch correction or restoration-type correction of the distal allele could result in a discontinuous conversion tract. These results indicate that both mitotic and meiotic crossovers were associated with conversions in ectopic recombination.
View larger version (12K):
In this window
In a new window
Download PPT slide
|
Figure 4.
Association of crossing over and conversion in mitotic recombination. Recombinants with reciprocal translocations from the ade6-M375 x zzz7-3::ade6-469 mitotic recombination reaction (JV548) or the ade6-M26 x zzz7-3::ade6-469 meiotic cross [GP1540 ade6-M26 leu1-32 ura4-D18 x GP1123 ade6-D1 leu1-32 ura4-D18 zzz7-3::(ade6-469 ura4+) (VIRGIN and BAILEY 1998 )] were identified and confirmed as described in MATERIALS AND METHODS. The two recombinant allele structures C and D (Fig 3) from each crossover recombinant were characterized using allele-specific PCR and DNA sequencing. The alleles present at the M375 or M26 and the 469 positions are indicated. Bars, conversion events; X, crossover events.
|
|
| DISCUSSION |
|---|
A quantitative and qualitative comparison between mitosis and meiosis of the genetic products of recombination can be exploited to better understand the mechanisms that control and execute recombination. In this study we have examined ectopic and allelic recombination in mitosis and compared the results to a previous study of the same recombination reactions in meiosis. In both mitosis and meiosis, the rate of ectopic recombination was reduced relative to allelic recombination, and a chromosomal position effect was conserved in mitotic and meiotic diploid cells. In contrast, there was a dramatic difference between mitosis and meiosis in the proportion of chromosomal rearrangements due to crossing over. Together, these comparisons suggest some overlap and some important differences between mitosis and meiosis in the mechanism and control of recombination.
Allelic and ectopic recombination were increased in meiosis relative to mitosis:
The 100- to 1000-fold increase in allelic ade6 recombination in meiosis relative to mitosis is a direct demonstration of the induction of recombination in meiosis, consistent with the role of recombination in generating genetic diversity and proper meiotic chromosome segregation. Ectopic recombination in diploids was increased 100- to 200-fold in meiosis relative to mitosis (Fig 2). This result suggests overlap in the mechanisms that control allelic and ectopic recombination.
In comparison with mitotic recombination in haploids, meiotic ectopic recombination was increased only 2- to 40-fold. This range of differences is lower than that found in diploids and is due to rates of ectopic recombination that are higher in vegetative haploids than in diploids. The higher ectopic recombination rates in haploids could be explained in several different ways: (1) greater flexibility of chromosome movement in haploids compared to diploids (see below); (2) competition for recombination partners between homologous chromosomes and ectopic sites on nonhomologous chromosomes in diploids; and (3) a concentration effect, since the nucleus and genome sizes are smaller in haploids than in diploids. These results contrast with a previous study in S. cerevisiae, in which mitotic interchromosomal ectopic recombination was slightly lower in haploids compared to diploids (LICHTEN and HABER 1989 ). There may be some differences between these two distantly related yeasts in mitotic chromosome dynamics and the control of ectopic recombination.
Ectopic recombination was reduced relative to allelic recombination:
In diploids, ectopic recombination was suppressed relative to allelic recombination in both mitosis and meiosis (Fig 2). One possible mechanism for limiting ectopic recombination is the pairing of homologous chromosomes. Whole chromosome painting studies have demonstrated that homologous chromosomes are associated with each other in vegetative diploids in S. pombe (SCHERTHAN et al. 1994 ). In S. cerevisiae, fluorescence in situ hybridization studies with single-locus probes suggest a certain level of homologous chromosome pairing in vegetative diploids (WEINER and KLECKNER 1994 ; BURGESS et al. 1999 ). Somatic pairing of homologous chromosomes has also been demonstrated in D. melanogaster (reviewed in HENIKOFF and COMAI 1998 ). The association of homologous chromosomes may limit the capacity for ectopic interactions among repetitive sequences on nonhomologous chromosomes and promote interactions between allelic sequences, as proposed for meiotic recombination in S. cerevisiae (GOLDMAN and LICHTEN 1996 ). For the recombination reactions reported here, the range of differences between allelic and ectopic recombination rates was 2.5- to 70-fold for mitosis and 13- to 300-fold for meiosis (Fig 2). This difference in ranges is consistent with more close alignment of homologous chromosomes in meiosis than mitosis. An alternative explanation for the higher rates of allelic vs. ectopic recombination is that diploid-specific gene products suppress ectopic recombination.
In addition to these general effects, the chromosomal position effect specific to the z7 x ura4 reaction appears to depend on the presence of a diploid genome. Although the reason for this position effect is not understood, one possibility is that some features of nuclear architecture that limit ectopic recombination are diploid specific. An analysis of nuclear diffusion of specific chromosomal loci in S. cerevisiae diploid vegetative cells suggests that nuclear architecture constrains diffusion to a subnuclear domain (MARSHALL et al. 1997 ). Since the same studies were not reported for haploid cells, it is not known if constrained diffusion is diploid specific. An alternative possibility is that pairing and unpairing of specific loci on homologous chromosomes occurs at different times (BURGESS et al. 1999 ), thereby preventing the opportunity for interactions between certain loci. Finally, it has been suggested that homolog interactions in mitotic cells may simply reflect self-organization according to chromosome size within the limited space of the nucleus, rather than specific controls (JIN et al. 1998 ).
Studies of mitotic recombination in S. cerevisiae have shown variable results in the comparison of ectopic and allelic recombination. In one study using ura3 heteroalleles, ectopic recombination was approximately fivefold lower than allelic recombination (JINKS-ROBERTSON and PETES 1986 ), whereas in another study of leu2 heteroalleles, ectopic recombination was nearly equivalent or slightly increased relative to allelic recombination (LICHTEN and HABER 1989 ). Using an artificially inserted bacteriophage site-specific recombination system in S. cerevisiae, BURGESS et al. 1999 showed that recombination rates ranged from nearly equal to sixfold higher for allelic compared to ectopic interactions. On average, allelic interactions were approximately twofold higher than ectopic interactions. Some of the differences in ectopic recombination efficiency may be explained by the relative distances of two ectopic loci from their respective centromeres, consistent with a "Rabl" arrangement of chromosomes in the interphase nucleus (BURGESS and KLECKNER 1999 ). In S. pombe, the low ectopic recombination rates between the centromere-linked ade6 locus and the more distal ectopic insertions are consistent with a Rabl effect (Table 2; Fig 2). However, in both S. pombe and S. cerevisiae, the chromosomal position effects on ectopic recombination that suggest a Rabl effect are small differences (two- to threefold), and only a few loci have been tested.
Chromosomal rearrangements were rarely associated with mitotic ectopic recombination:
We previously demonstrated that 35–60% (or 22–24% without correction for aneuploid segregation) of ectopic intragenic recombination events during meiosis had an associated crossover that resulted in a reciprocal translocation (VIRGIN and BAILEY 1998 ). In mitotic recombination, with the same recombination substrates as studied in meiotic recombination, we found that 1–7% of ectopic intragenic recombination events had an associated crossover (Table 3). This remarkable decrease in the proportion of crossovers associated with intragenic recombination was not restricted to ectopic recombination, since we found only 2% crossovers associated with ade6 allelic recombination during mitosis, compared to 52–65% previously reported for meiosis (GRIMM et al. 1994 ). These results indicate that, for ade6 recombination in S. pombe, the relationship between reciprocal and nonreciprocal recombination is different in mitosis and meiosis.
Studies of meiotic recombination in several species of fungi led to the development and later refinement of recombination models in which gene conversion and crossing over were mechanistically linked (HOLLIDAY 1964 ; MESELSON and RADDING 1975 ; RESNICK 1976 ; SZOSTAK et al. 1983 ; STAHL 1996 ). Central to these models is a crossed-strand intermediate (Holliday junction) that can isomerize to allow resolution, by nuclease cleavage, to produce either nonreciprocal or reciprocal recombinants. The action of mismatch repair enzymes on the heteroduplex intermediate, combined with resolution, is believed to produce gene conversion events with or without an associated crossover. Physical, biochemical, and genetic evidence has accumulated in support of the role of the Holliday junction in recombination. Structures consistent with Holliday junctions have been isolated from Escherichia coli (POTTER and DRESSLER 1976 ) and S. cerevisiae meiotic cells (BELL and BYERS 1979 ; SCHWACHA and KLECKNER 1994 , SCHWACHA and KLECKNER 1995 ). Molecular modeling and biochemical studies indicate that model four-strand junctions undergo the required conformational changes (SIGAL and ALBERTS 1972 ; LILLEY 1997 ). In E. coli, both biochemical and genetic studies suggest that the RuvC protein functions in resolving four-strand recombination intermediates (reviewed in WEST 1997 ). Enzymes with similar activities have been identified in S. cerevisiae (KLEFF et al. 1992 ) and S. pombe (WHITE and LILLEY 1997 ), although these gene products are targeted to mitochondria rather than to the nucleus. Although these results are consistent with models in which Holliday junctions are resolved by nuclease cleavage, they do not exclude alternative mechanisms of resolution. Based on studies of mitotic recombination in S. cerevisiae, models have been proposed in which resolution of a heteroduplex intermediate occurs by completing chromosomal replication, rather than by specific endonuclease action (discussed in ESPOSITO and WAGSTAFF 1981 ; PAQUES and HABER 1999 ). Different pathways of resolution may have different levels of importance, depending on the type of junction and the physiologic condition of the cell.
The classical genetic evidence in support of the association of gene conversion and crossing over is that, in most cases, the proportion of gene conversion events associated with outside marker exchange is higher than that expected if the two events were independent (reviewed in HOLLIDAY 1964 ; FOGEL et al. 1981 ; WHITEHOUSE 1982 ; PETES et al. 1991 ). This result supports a mechanistic coupling of the two events. However, in meiotic recombination the degree of association varies with different markers from 20 to 70%, and the interval of exchange is sometimes separated from the converted allele by a nonrecombinant marker. Furthermore, some mutants in S. cerevisiae, as well as in D. melanogaster, show differential effects on gene conversion and crossing over (CARPENTER 1982 ; ENGEBRECHT et al. 1990 ; ROSS-MACDONALD and ROEDER 1994 ; HOLLINGSWORTH et al. 1995 ). Although these results do not exclude the possibility of a single concerted event that gives rise to both gene conversion and crossing over, they raise the possibility of two independent events that are associated due to favorable regional chromosomal conditions (discussed in PETES et al. 1991 ).
In mitotic cells, reduction of heterozygous markers to homozygosity is believed to result most frequently from crossing over between the marker and the centromere during the postreplication phase of the cell cycle. Recessive homozygosis is more frequent among mitotic intragenic recombinants than among unselected cells (Table 4; also see KAKAR 1963 ; ESPOSITO 1978 ). The combination of this result and the fact that the majority of intragenic recombinants are nonreciprocal for heteroallelic markers (ROMAN 1956 ) suggests that flanking marker exchange is positively associated with gene conversion in mitotic recombination. Similar to meiotic recombination, a wide range of association between mitotic intragenic recombination and flanking marker exchange has been observed. On average the association is lower than the 20–70% reported for meiotic recombination (reviewed in WHITEHOUSE 1982 ). However, recessive homozygosis is a relatively insensitive method for detecting flanking marker exchange because alternative segregation possibilities at anaphase could theoretically result in detection of only one-half the G2 crossovers. Furthermore, if G1 crossovers occurred, they would not result in recessive homozygosis. CHUA and JINKS-ROBERTSON 1991 demonstrated that in S. cerevisiae recombinant chromatids segregated to the same or opposite poles of the cell in equal proportions, as expected for random segregation, resulting in expression of recessive homozygosis in only one-half the crossover recombinants. In contrast, several studies in D. melanogaster have shown preferential cosegregation of recombinant chromatids (PIMPINELLI and RIPOLL 1986 ; BEUMER et al. 1998 ), which results in a weaker estimate of crossovers based on recessive homozygosis.
A more sensitive measure of the association between intragenic recombination and flanking marker exchange has been achieved by the isolation of mitotic intragenic recombinants from diploids, followed by induction of sporulation to make haploid derivatives. This method is equivalent to a "half-tetrad" analysis, in which two chromosomes, at least one of which is recombinant, can be recovered in isolated form. The status of flanking markers on the recombinant chromosome can then be determined directly, without loss of information due to segregation patterns. Using this method, HURST and FOGEL 1964 made a direct comparison between mitotic and meiotic allelic recombination with the same heteroallelic markers in S. cerevisiae. About one-half the intragenic recombinants showed multiple exchange events among several flanking markers. If consideration of their results is restricted to events involving intragenic recombination with or without exchange between the two closest flanking markers, then the frequencies of flanking marker exchange among intragenic recombinants were 0.2% in unselected cells, 19% in mitotic recombinants, and 58% in meiotic recombinants. Thus, in mitotic allelic recombination, although gene conversion is associated with crossing over, the association is not as strong as it is in meiosis. Based on the current study, the association in S. pombe mitotic cells is also diminished relative to meiotic cells, and may be even weaker than that in S. cerevisiae.
In haploid yeast cells, ectopic heterochromosomal recombination can provide an accurate assessment of crossing over frequencies, since there is often selection for cosegregation of the reciprocal translocation products to maintain a balanced genome. Similar to the studies of allelic recombination discussed above, a wide range of crossover proportions is associated with mitotic intragenic recombination between heterochromosomal repeats, from 2 to 60% in S. cerevisiae (JINKS-ROBERTSON and PETES 1986 ; LICHTEN and HABER 1989 ; HARRIS et al. 1993 ; JINKS-ROBERTSON et al. 1993 ). The reasons for this wide range are unclear, but may be related to the nature of the recombining sequences (discussed in PAQUES and HABER 1999 ). Only a few studies have reported direct comparisons between mitotic and meiotic interchromosomal ectopic recombination using the same substrates, as in this study. In S. cerevisiae, the proportion of crossovers is decreased in mitosis relative to meiosis in ectopic recombination with ura3 insertions (6% for mitosis vs. 43% for meiosis; JINKS-ROBERTSON and PETES 1986 ) and leu2 insertions (10–20% for mitosis vs. 22–45% for meiosis; LICHTEN et al. 1987 ; LICHTEN and HABER 1989 ). Thus, in both S. pombe and S. cerevisiae, there is direct evidence that ectopic intragenic recombination is associated with crossing over less frequently in mitosis than in meiosis.
In multicellular organisms, there is evidence for nonreciprocal and reciprocal recombination in somatic cells. In Drosophila melanogaster, Mus musculus, and Homo sapiens, somatic mosaicism associated with twin spots, or a pattern of homozygosis of multiple markers along the terminal region of a chromosome arm, provides evidence of mitotic crossing over (STERN 1936 ; GUPTA et al. 1997 ; SHAO et al. 1999 ). In these studies, it could not be determined if the crossovers were associated with conversions. In D. melanogaster and M. musculus, most of the recombination products analyzed after induction of a site-specific chromosomal double-strand break were nonreciprocal, indicating a low association of crossing over with mitotic double-strand break repair (NASSIF et al. 1994 ; MOYNAHAN and JASIN 1997 ; RICHARDSON et al. 1998 ). These results are consistent with those obtained in fungi and suggest that the suppression of crossing over in mitotic recombination is conserved through evolution.
One possible explanation for the low frequency of crossovers associated with conversions in mitosis is that the two events can occur independently, and the crossover pathway is reduced in mitosis to a greater extent than is the conversion pathway. In S. pombe, multiple mutants deficient in mitotic intergenic recombination have no effect on intragenic recombination (GYSLER-JUNKER et al. 1991 ). These mutants could exert different effects on conversion and crossing over either by influencing the resolution of a common intermediate in a pathway that gives rise to both conversions and crossovers or by selectively inactivating one of multiple separate pathways. In this study, the majority of the translocation chromosomes examined in both mitotic and meiotic ectopic recombination were associated with conversions (Fig 4). Furthermore, the crossovers that were not associated with conversions could have resulted from hybrid DNA limited to the region between the M375 and 469 markers, or from mismatch correction that resulted in restoration rather than conversion. Similar results have been reported for S. cerevisiae; among meiotic crossover recombinants, 40–60% show an associated gene conversion event (BORTS and HABER 1987 ; SYMINGTON and PETES 1988 ). The coupling of crossovers and conversions on the same chromosome, within a small region of homologous sequence, is consistent with the hypothesis that crossovers are mechanistically linked with conversions.
There are several other possible explanations for the low frequency of crossovers associated with conversions in mitosis. First, using the classical models discussed above as a guide, all mitotic gene conversion and crossover events could occur through a common mechanism, with the different outcomes resulting from alternative resolution of a stable joint molecule, such as a double Holliday junction. In this case, the low proportion of crossovers in mitosis could be due to resolution favoring nonreciprocal recombination products (Fig 5A). A topoisomerase-mediated resolution mechanism has been proposed as a means of eliminating crossovers (HASTINGS 1988 ; MCGILL et al. 1989 ; WANG 1996 ). Second, there could be two separate pathways of mitotic recombination: a major pathway that does not include crossing over and a minor pathway that includes both conversion and crossing over. The synthesis-dependent strand annealing (SDSA) pathway has been proposed as a mechanism for mitotic gene conversion without crossing over in Drosophila (NASSIF et al. 1994 ). Mechanisms with similar features have been proposed for other mitotic recombination events, including recombination-induced DNA replication of bacteriophage T4 (FORMOSA and ALBERTS 1986 ), mating-type switching (STRATHERN 1988 ) and direct-repeat recombination (SUGAWARA and HABER 1992 ) in S. cerevisiae, and concatenation of DNA transfected into mammalian cells (LIN et al. 1984 ). The important common feature of these models that could explain the lack of crossing over is the absence of a stable Holliday junction intermediate, from which crossover products are believed to arise. Instead, the SDSA model proposes formation of a moving D-loop replication bubble that dissociates, followed by annealing of unpaired DNA strands. In this case, the lower proportion of crossovers in mitosis could be due to a major recombination pathway with an unstable joint molecule and without crossing over, such as SDSA, and a minor pathway with crossing over, similar to meiotic recombination (Fig 5B). It is also possible that a potentially reversible joint molecule could rarely progress to a stable joint molecule including a Holliday junction, providing the few crossovers observed (PAQUES et al. 1998 ; PAQUES and HABER 1999 ). Further experiments to determine the mechanism of suppression of crossing over in mitotic cells will be important, since loss of this suppression is a potentially important component of genomic instability.
View larger version (13K):
In this window
In a new window
Download PPT slide
|
Figure 5.
Possible pathways for suppression of crossing over in mitotic recombination. (A) In model 1 conversion and crossing over are alternative outcomes of resolution of a common intermediate, a stable biparental joint molecule (JM), of which the prototype model is the double Holliday junction. Crossing over is suppressed during resolution of the JM. (B) In model 2, crossing over is suppressed at an earlier step by the formation of alternative intermediates. The major intermediate is an unstable JM that dissociates and gives rise only to conversions. The prototype model for the unstable JM is a D-loop replication bubble, as proposed in the synthesis-dependent strand annealing model (see text). The few observed crossovers are derived from a less common stable JM. Another possible variation is the transformation of an unstable JM to a stable JM (PAQUES et al. 1998 ). Dashed arrows indicate potential steps for inhibition of crossing over.
|
|
| ACKNOWLEDGMENTS |
|---|
We thank Gerry Smith for strains and for helpful discussion. DNA sequencing was performed by the core sequencing facility of the Center for Molecular Medicine and Genetics at Wayne State University, under the direction of Dr. Mike Hagan. This work was supported by grants from the American Cancer Society and The Barbara Ann Karmanos Cancer Center.
Manuscript received February 8, 2000; Accepted for publication September 28, 2000.
| LITERATURE CITED |
|---|
BELL, L. and B. BYERS, 1979 Occurrence of crossed strand-exchange forms in yeast DNA during meiosis. Proc. Natl. Acad. Sci. USA 76:3445-3449[Abstract/Free Full Text].
BEUMER, K. J., S. PIMPINELLI, and K. G. GOLIC, 1998 Induced chromosomal exchange directs the segregation of recombinant chromatids in mitosis of Drosophila. Genetics 150:173-188[Free Full Text].
BODI, Z., A. GYSLER-JUNKER, and J. KOHLI, 1991 A quantitative assay to measure chromosome stability in Schizosaccharomyces pombe. Mol. Gen. Genet. 229:77-80[Medline].
BORTS, R. H. and J. E. HABER, 1987 Meiotic recombination in yeast: alteration by multiple heterozygosities. Science 237:1459-1465[Abstract/Free Full Text].
BOSCO, G. and J. E. HABER, 1998 Chromosome break-induced DNA replication leads to nonreciprocal translocations and telomere capture. Genetics 150:1037-1047[Abstract/Free Full Text].
BURGESS, S. M. and N. KLECKNER, 1999 Collisions between yeast chromosomal loci in vivo are governed by three layers of organization. Genes Dev. 13:1871-1883[Abstract/Free Full Text].
BURGESS, S. M., N. KLECKNER, and B. M. WEINER, 1999 Somatic pairing of homologs in budding yeast: existence and modulation. Genes Dev. 13:1627-1641[Abstract/Free Full Text].
CARPENTER, A. T., 1982 Mismatch repair, gene conversion, and crossing-over in two recombination-defective mutants of Drosophila melanogaster.. Proc. Natl. Acad. Sci. USA 79:5961-5965[Abstract/Free Full Text].
CELERIN, M., S. T. MERINO, J. E. STONE, A. M. MENZIE, and M. E. ZOLAN, 2000 Multiple roles of Spo11 in meiotic chromosome behavior. EMBO J. 19:2739-2750[Medline].
CERVANTES, M. D., J. A. FARAH, and G. R. SMITH, 2000 Meiotic DNA breaks associated with recombination in S. pombe. Mol. Cell 5:883-888[Medline].
CHUA, P. and S. JINKS-ROBERTSON, 1991 Segregation of recombinant chromatids following mitotic crossing over in yeast. Genetics 129:359-369[Abstract].
EGEL, R., 1984 Two tightly linked silent cassettes in the mating-type region of Schizosaccharomyces pombe.. Curr. Genet. 8:199-203.
ENGEBRECHT, J., J. HIRSCH, and G. S. ROEDER, 1990 Meiotic gene conversion and crossing over: their relationship to each other and to chromosome synapsis and segregation. Cell 62:927-937[Medline].
ESPOSITO, M., and J. WAGSTAFF, 1981 Mechanisms of mitotic recombination, pp. 341–370 in The Molecular Biology of the Yeast Saccharomyces, Life Cycle and Inheritance, edited by J. STRATHERN, E. JONES and J. BROACH. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
ESPOSITO, M. S., 1978 Evidence that spontaneous mitotic recombination occurs at the two-strand stage. Proc. Natl. Acad. Sci. USA 75:4436-4440[Abstract/Free Full Text].
FOGEL, S., R. MORTIMER and K. LUSNAK, 1981 Mechanisms of meiotic gene conversion, or "wanderings on a foreign strand," pp. 289–340 in The Molecular Biology of the Yeast Saccharomyces: Life Cycle and Inheritance, edited by J. STRATHERN, E. JONES and J. BROACH. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
FORMOSA, T. and B. M. ALBERTS, 1986 DNA synthesis dependent on genetic recombination: characterization of a reaction catalyzed by purified bacteriophage T4 proteins. Cell 47:793-806[Medline].
FORSBURG, S. L. and P. NURSE, 1994 Analysis of the Schizosaccharomyces pombe cyclin puc1: evidence for a role in cell cycle exit. J. Cell Sci. 107:601-613[Abstract].
GOLDMAN, A. S. and M. LICHTEN, 1996
TOP
ABSTRACT
MATERIALS AND METHODS
