Current Biology, Vol. 12, 1473–1483, September 3, 2002, 2002 Elsevier Science Ltd. All rights reserved. PII S0960-9822(02)01090-4

The Dynamics of Homologous Chromosome Pairingduring Male Drosophila Meiosis

ences, with mitosis [1, 2]. During mitosis, sister chroma-tids resulting from chromosome replication are held to-gether at the centromeres and along the arms by a

Julio Vazquez,1 Andrew S. Belmont,2

and John W. Sedat1,3

1Department of Biochemistry and BiophysicsUniversity of California, San Francisco conserved cohesin complex. Sister kinetochores attach

to microtubules emanating from opposite poles of theSan Francisco, California 941432 Department of Cell and Structural Biology spindle. At anaphase, cohesion between sisters is re-

leased and chromosomes segregate to opposite poles.University of Illinois, Urbana-ChampaignUrbana, Illinois 61801 In contrast, during meiosis, homologous chromosomes

must pair to ensure their disjunction at anaphase I. Co-hesion between sister chromatids is released in twostages: cohesion along the arms is released at anaphaseSummaryI, while cohesion at sister centromeres is released at themetaphase II-anaphase II transition. Finally, in contrastBackground: Meiotic pairing is essential for the properwith mitosis, sister kinetochores attach to microtubulesorientation of chromosomes at the metaphase plate andemanating from the same pole during meiosis I, whiletheir subsequent disjunction during anaphase I. In malethey assume a mitotic bipolar configuration in meiosis II.Drosophila melanogaster, meiosis occurs in the ab-

Different mechanisms have evolved to segregatesence of recombination or a recognizable synaptonemalchromosomes during meiosis. The most common mech-complex (SC). Due to limitations in available cytologicalanism involves pairing of homologs, recombination, andtechniques, the early stages of homologous chromo-the formation of a synaptonemal complex (SC). Whilesome pairing in male Drosophila have not been ob-SC formation is temporary, the sites of reciprocal ex-served, and the mechanisms involved are poorly under-change, or chiasmata, persist after SC disassembly instood.late prophase and are resolved during anaphase. TheResults: Chromosome tagging with GFP-Lac repressorphysical link between homologous chromosomes dur-protein allowed us to track, for the first time, the behavioring this critical phase is ensured by cohesion betweenof meiotic chromosomes at high resolution, live, at allexchange and nonexchange sister chromatids [3–6].stages of male Drosophila meiosis. Homologous chro-Such cohesion involves a meiosis-specific cohesin com-mosomes pair throughout the euchromatic regions inplex [1]. In Drosophila, MEI-S332 and ORD are also impli-spermatogonia and during the early phases of sperma-cated in the maintenance of sister cohesion in both maletocyte development. Extensive separation of homologsand female meiosis [7]. The exact relationship betweenand sister chromatids along the chromosome arms oc-pairing, recombination, SC formation, and chromosomecurs in mid-G2, several hours before the first meioticsegregation may vary [8]. For instance, in the yeast Sac-division, and before the G2/M transition. Centromeres,charomyces cerevisiae, recombination is required foron the other hand, show complex association patterns,both SC formation and chromosome segregation, whilewith specific homolog pairing taking place in mid-G2.mutations that abolish SC formation do not prevent re-These changes in chromosome pairing parallel changescombination. In contrast, both Drosophila and Caeno-in large-scale chromosome organization.rhabditis elegans can form an SC and segregate chro-Conclusions: Our results suggest that widespread in-mosomes in the absence of recombination.teractions along the euchromatin are required for the

In some organisms, meiosis can take place in theinitiation, but not the maintenance, of meiotic pairing ofabsence of recombination, synaptonemal complex for-autosomes in male Drosophila. We propose that hetero-mation, or both [9]. The mechanisms of chromosomechromatic associations, or chromatid entanglement,pairing and segregation in such organisms are not wellmay be responsible for the maintenance of homologunderstood. In Drosophila females, the small fourthassociation during late G2. Our data also suggest thatchromosome never recombines, while the X chromo-the formation of chromosome territories in the sperma-some recombines only 90%–95% of the time. Genetictocyte nucleus may play an active role in ensuring thestudies using either natural chromosomes carrying du-specificity of meiotic pairing in late prophase by dis-plications [10] or a minichromosome derivative [11]rupting interactions between nonhomologous chromo-indicated a role for heterochromatic homology in thesomes.segregation of homologous achiasmate chromosomes.Three-dimensional cytological studies further provided

Introduction direct physical evidence for heterochromatic interac-tions between nonrecombinant homologous chromo-

During meiosis, two rounds of cell division follow one somes [12]. Drosophila females also have a limited abil-round of chromosome replication, resulting in the halv- ity to segregate chromosomes that do not undergoing of the chromosome number from a diploid to a hap- recombination and lack homology (“distributive” segre-loid complement. Meiosis is a complex process that gation). Such chromosomes do not show detectableshares some similarities, but also several notable differ- attachments [12], suggesting that physical pairing may

not be an absolute requirement for chromosome segre-gation. It has been suggested that segregation mecha-3 Correspondence: sedat@msg.ucsf.edu

Current Biology1474

nisms related to those present in Drosophila achiasmate results suggest a three-step model for meiotic pairingin male Drosophila.meiosis might be used in humans as a backup in case

of recombination failure [13].In Drosophila males, meiosis occurs in the absence Results

of recombination [14] or formation of a detectable syn-aptonemal complex [15, 16]. Pairing is critical for correct Tracking Chromosomes Live in the Malesegregation during male meiosis, as sex chromosomes Drosophila Germlinedeleted for the rDNA locus both fail to pair and show a We used the GFP-Lac repressor/lac operator systemhigh degree of nondisjunction [17–19]. Genetic analysis to label specific chromosomal loci in live, developingof the sequences required for proper chromosome seg- spermatocyte nuclei [40]. A nuclear green fluorescentregation has revealed the presence of at least two differ- protein-Lac repressor fusion protein (GFP-Lac I) wasent segregation systems in male Drosophila. Pairing of expressed in the male germline and was used to labelsex chromosomes requires specific sequences from the an array of lac operator (lacO) sequences integrated atrDNA gene repeats present in the heterochromatin of the specific chromosomal locations (Figure 1A). TransgenicX and Y chromosomes [20–22]. Pairing of autosomes, on Drosophila lines heterozygous for a single autosomalthe other hand, appears to require multiple regions of insert on chromosome 2 showed one spot of fluores-homology in the euchromatin [23]. The existence of dif- cence in mitotically dividing spermatogonia (cysts withferent segregation systems in males is also supported 2–8 nuclei), and young spermatocytes (cysts with 16by the isolation of mutants that alter the segregation nuclei). In lines carrying multiple inserts, either on theof specific chromosomes [24–26]. With a few notable same or different chromosomes, a corresponding num-exceptions, mutations that affect chromosome segrega- ber of GFP spots was present, indicating the lack oftion in females do not affect segregation in males, and ectopic interaction between lacO inserts (Figures 1B,vice versa [27]. This finding suggests that distinct mech- 1J, and 1K). The spots were visible in all identifiableanisms are present in the two sexes. germline nuclei but were absent in somatic cyst nuclei,

Classic cytological studies of meiosis in male Dro- due to the lack of expression of GFP-Lac I in these cells.sophila have shown that homologous chromosomes are Spots, however, could also be detected in other tissuespaired in the prometaphase to anaphase stages of meio- and cell types by using appropriate GFP-Lac I-express-sis I [28–33], and possibly also in mitotically dividing ing lines (data not shown). In a minority of nuclei (�5%),spermatogonia [28]. Autosomes, for instance, appear two spots could be detected in close proximity to eachaligned along their entire length in metaphase squashes, other (�0.5 �m). Such unpaired loci were present inwith the four chromatids assuming a parallel orientation. spermatogonia and young spermatocytes. However, weThe X-Y chromosome pair, on the other hand, is con- never found cysts with unpaired sister chromatids innected at discrete sites (“collochores”) located at well- more than a few nuclei. These observations suggestdefined locations in the pericentric heterochromatin occasional transient local separation of sister chroma-[31]. Surprisingly, these sites do not seem to correspond tids in G2 or nuclei in the process of chromosome repli-to the rDNA pairing sites identified in genetic studies. cation.This finding suggests that different mechanisms may beresponsible for the establishment and maintenance of

Homologous Chromosome Pairing inmeiotic chromosome pairing. The study of meiosis inSpermatogonia and Young SpermatocytesDrosophila females has greatly benefited from the use ofIn lines homozygous for a single lacO insert, a singleelectron microscopy techniques [34] and, more recently,fluorescent spot was present in approximately half of thefrom the development of in situ hybridization methodsnuclei in spermatogonia and very young spermatocytesthat preserve the three-dimensional organization of the(Figures 1D–1F; note that these two stages cannot al-oocyte nucleus [12]. The early stages of male meioticways be reliably distinguished in our assay). Further-pairing, on the other hand, have been notoriously resis-more, we rarely observed nuclei with more than twotant to cytological analysis [33]. Both conventional lightspots, consistent with tight association of sister chroma-microscopy and electron microscopy approaches, fortids. The results indicate that homologous chromo-instance, are unable to resolve individual chromosomessomes are paired in a substantial fraction of interphaseduring early prophase [35, 36], and in situ hybridizationspermatogonia and in the very early stages of spermato-methods similar to those developed for female meiosiscyte development.have not been described.

Young spermatocytes can be recognized by theirHere, we report the use of a live cytological approachlarger nuclear size compared to spermatogonial nucleito address the dynamics of homologous chromosomeand by various cytological features readily observablepairing during male meiosis in Drosophila. The GFP-Lacby differential interference contrast microscopy. Inrepressor/lac operator system [37, 38] was used to markyoung spermatocytes, both sister chromatids and ho-a variety of euchromatic sites, while centromeres weremologous chromosomes were paired (Figures 1H, 1J,detected with a CID-GFP fusion protein [39]. This ap-and 1K). Occasionally, sister chromatids could be re-proach allowed for the direct observation of chromo-solved (Figure 1K). The frequency of paired homologssomes with a high degree of spatial resolution, at allat this stage was greater than 95%, considerably moredevelopmental stages, in live, unperturbed spermato-than in spermatogonia. The results suggest that homolo-cyte nuclei. We found complex patterns of homologousgous chromosomes pair very early during spermatocytechromosome pairing that are coordinated with changes

of nuclear organization during meiotic prophase. The development, possibly during, or shortly after, the last

Dynamics of Meiotic Chromosome Pairing1475

Figure 1. Tracking Homologous Chromosome Pairing Live in the Male Drosophila Germline

(A) Chromosomal location of the lacO arrays used in this study (only mapped inserts are shown; see the Experimental Procedures).(B–L) Pairing patterns in the male germline. The number and location of lacO sites is shown above the micrographs. (B and C) In spermatogonia,nuclei from a line heterozygous for three inserts have three distinct spots (B). A higher magnification view is shown in (C). Note that thisparticular insert on the X is less bright than the autosomal inserts, probably due to internal deletions in the lacO array, and can therefore berecognized. (D–F) The “somatic” pairing of homologous chromosomes in spermatogonia homozygous for a lacO insert on chromosome 2.Note the presence of nuclei with both paired and unpaired chromosomes. (F) shows an enlarged view of a portion of (D). (G–I) Homologouschromosome pairing in spermatocyte nuclei homozygous for a single insert. Fluorescence images show a field of nuclei at (G) different stagesof development or representative nuclei at stages (H) S2 or (I) S5. Note the pairing of homologs in early G2 and their separation in late G2.(J–L) The separation of sister chromatids in a line heterozygous for a chromosome with two inserts. Panels show a field of nuclei at (J) differentstages or a group of representative nuclei at stage (K) S2 or (L) S5. In early G2, sisters are generally paired, and there is no interaction betweendifferent inserts. The arrow in (K) shows two resolved sister loci. In late G2, sisters are separated at all loci. The developmental stages havebeen described [36]. Each panel is a projection of several optical sections. The location of second chromosome lacO inserts is (D–F and G–I)57A or (A and B and J–L) 60A and 60F. The location of the X-chromosome insert shown in (A) and (B) was not determined. Distances betweenGFP spots are consistent with a 30 nm fiber configuration for spermatocyte chromosomes, organized into a random coil ([40] and data notshown). Heterochromatin is represented as a filled box, centromeres are represented as open circles, and lacO inserts are represented asfilled circles. In all examples, chromosomes have replicated and consist of two sister chromatids (not shown in diagram). The scale barrepresents 5 �m.

gonial mitotic division, as was originally suggested by reached 5–6 �m (average value �2.5 �m; [40]). Assum-ing a level of chromosome compaction consistent withMetz [28].a 30-nm fiber organization in these nuclei [40], the dis-tances observed are consistent with a separation ofHomolog and Sister Separation

in Mature Spermatocytes chromatids over at least 500 kb on both sides of individ-ual lacO sites. Similar distances were observed at differ-In older spermatocytes, however, extensive separation

of homologs was observed (Figure 1I). Unexpectedly, ent lacO inserts on chromosomes X, 2, and 3 (14 sitesanalyzed), including 8 sites on the distal half of chromo-not two, but four GFP spots were present in lines homo-

zygous for a single lacO insert. This result indicates some arm 2R. These results suggest complete or nearlycomplete separation of chromatids along the euchro-separation of homologous, as well as sister, chromatids.

Separation of sister chromatids was also evident in lines matic regions of Drosophila chromosomes in maturespermatocytes. The majority of cysts consisted of nucleiheterozygous for either a single (data not shown) or a

double insert (Figure 1L). The separation of chromatids that either were all paired (young spermatocytes) or wereall unpaired (mature spermatocytes). Only a small frac-was extensive, as the distances between tagged loci

Current Biology1476

Figure 2. Simultaneous Separation of Homo-logs and Sister Chromatids Occurs before theG2/M Transition

(A) A cyst with chromosomes in the processof separation. The line was homozygous for asingle insert at position 57A, as diagrammedabove the micrograph. Arrows indicate nucleiin which a single chromatid has separated.(B–E) Nuclei heterozygous for a single lacOinsert at position 60F were examined in analy2/aly2 background. Chromatids are pairedin early (B and C) G2 and are separated in (Dand E) late G2. The scale bar represents 5 �m.

tion of cysts (�5%) contained nuclei with either paired ment (Figure 3). These stages were defined by using aset of morphological criteria, as previously describedor unpaired chromosomes. This result suggests that

chromatid separation must occur during a relatively nar- [36]. Sister lacO sites were generally paired in earlystages (S1–S2a, Figures 3B and 3C), except in rare in-row developmental window lasting less than 4 hr.stances (last panel in 3C). Sister loci were always sepa-rated in stages S3 and older (Figures 3D–3F). In a smallSimultaneous Separation of Homologsproportion of cysts, corresponding approximately to theand Sister ChromatidsS2b/S3 transition, a variable number of dots was ob-To address whether unpairing of homologous chromo-served among sister nuclei (see Figure 2A), indicatingsomes occurs before sister separation, we examineddifferent degrees of chromosome separation.those rare cysts with chromosomes in the process of

In order to correlate sister separation with the changesseparation. If separation occurs in two stages (i.e., ho-in nuclear organization that occur at the S2b/S3 transi-mologs first, and sisters second), we should expect totion, we counter-stained live cysts with the chromatinfind nuclei with either one spot (homologs paired), twodye Hoechst 33342 (Figure 4). In early stages, character-identical spots (homologs unpaired), one large and twoized by a single chromatin mass, sisters were paired.smaller spots (homologs unpaired, sister chromatidsBoth paired and unpaired chromosomes were observedseparated in one homolog), or four small spots (homo-in nuclei at the S2b/S3 transition, when chromosomelogs unpaired, all sister chromatids separated). All ofterritories first become apparent (Figures 4E–4G). Inthese combinations were found. An additional categoryolder nuclei, with well-defined chromosome territories,of nuclei, however, was also frequently observed (Figuresisters were generally separated (Figures 4H–4J). As2A). These nuclei were characterized by one large spotillustrated in Figure 4, the separation of sisters is exten-and one faint spot. Since allelic lacO sites always gavesive, with homologous loci spread over the entire chro-signals of similar intensity, the latter category must rep-mosome territory. Unpairing of homologs and sisterresent nuclei in which a single chromatid has separatedchromatids therefore appears to take place just after thefrom the bivalent. These results indicate that prior un-formation of chromosome territories in early stage S3.pairing of homologs is not a prerequisite for sister chro-

matid separation. Rather, individual chromatids appearto separate independently, suggesting a common Dynamic Pairing Patterns of Centromeresmechanism for homolog and sister chromatid cohesion The previous results indicate that extensive euchromaticand/or separation. associations are unlikely to provide the required physi-

cal linkage between homologous chromosomes duringlate meiotic prophase in male Drosophila. By analogySeparation Occurs before the G2/M Transition

Progression of spermatocytes through their normal de- to mitosis [42], it is possible that cohesion at the centro-meres might be responsible for such linkage. To addressvelopmental pathway is under genetic control. In partic-

ular, commitment to meiosis is dependent on a set of the pairing of centromeres in live spermatocytes, weused a transgenic line expressing a CID-GFP proteingenes, including always early (aly), cannonball (can),

spermatocyte arrest (sa), and meiosis I arrest (mia) [41]. ([39]; J.S. Platero and S. Henikoff, personal communica-tion). CID is the Drosophila homolog of human CENP-A,In these mutants, spermatocytes arrest in G2, and mei-

otic and postmeiotic stages are absent. The analysis of a histone H3 variant that is exclusively deposited atcentromeres, and plays a role in kinetochore functiontagged chromosome lines in a homozygous aly back-

ground revealed that sister chromatid separation oc- [39, 43]. The analysis of CID-GFP transgenic lines re-vealed unexpected and dynamic patterns of centromerecurred normally and therefore precedes the G2/M transi-

tion (Figures 2B–2E). behavior during spermatocyte development.In Drosophila nuclei, four CID-GFP spots are expectedTo determine the precise time of separation, we exam-

ined spermatocyte nuclei at different stages of develop- if homologous centromeres are paired, and eight spots

Dynamics of Meiotic Chromosome Pairing1477

Figure 3. Chromatid Separation Occurs at the S2b/S3 Transition

(A–F) (A) A transgenic line heterozygous for a single lacO array at position 57A was analyzed by (B and D) DIC or (C–F) fluorescence microscopy.Stages correspond approximately to (B and C) S2, (D and E) S3, and (F) S5. The scale bar represents 5 �m.

are expected if they are unpaired. In young spermato- ochromatin. Sister centromeres, on the other hand,remained associated throughout spermatocyte devel-cytes (stages S1–S2a), we observed a variable number

of spots, ranging from 1 to 8, with a majority of nuclei opment.showing 1–3 spots (Figure 5A). The preponderance ofnuclei with fewer than four spots indicates a strong Chromosome Dynamics during the First

Meiotic Divisiontendency of centromeres to cluster through nonhomolo-gous interactions, a situation reminiscent of chromocen- Metz suggested that the pairing observed in young

spermatocytes could be related to the somatic pairingter formation in Drosophila somatic tissues. On the otherhand, the existence of nuclei with more than four spots observed in Drosophila somatic cells [28]. The patterns

of pairing observed during early meiotic prophase couldalso indicates the presence of unpaired homologouscentromeres. This is in contrast to the virtual absence therefore be unrelated to the actual meiotic process.

We sought to determine whether de novo pairing ofof unpaired euchromatic loci at these same stages.The variable number of spots in early stages (Figures homologs or sister chromatids occurs in the spermato-

cyte nucleus in prometaphase. In male Drosophila, mei-5A and 5B) gave rise to four spots in early stage S3, withone spot present in each newly formed chromosome osis is relatively short compared to the length of the

complete spermatocyte developmental cycle (about 3–4territory (Figure 5C). As spermatocyte developmentcontinued, each CID-GFP spot appeared to duplicate, hr versus 80 hr), and cysts in the process of meiosis are

relatively rare. To capture these rare events, cysts weregiving rise to nuclei with eight spots, two per chromo-some territory (Figures 5D–5G). The results indicate that cultured in vitro under the microscope, and time-lapse

images of mature spermatocytes were collected as theythe exclusive pairing of homologous centromeres isachieved only in early stage S3, soon after the formation underwent meiosis (Figure 6). Meiotic prophase in male

Drosophila is not well defined. However, spermatocytesof chromosome territories, and coincides with, or shortlyfollows, the extensive unpairing of euchromatic regions. undergo visible morphological changes as they prepare

for meiosis [36, 44]. For instance, mature spermatocytesThe results suggest that formation of chromosome terri-tories may disrupt the nonhomologous heterochromatic are characterized by large, irregular nuclei of �14–16

�m diameter that become round (�12–14 �m diameter)associations observed in young spermatocytes. In lateG2, homologous centromeres were always separated, at prometaphase. In our lines, the predominantly nuclear

GFP-Lac I protein disperses within the entire cell as thesuggesting local unpairing of homologous chromo-somes over a sizeable portion of the pericentric heter- nuclear envelope breaks down when nuclei approach

Current Biology1478

Figure 5. Pairing Patterns of Centromeric RegionsFigure 4. Chromatid Separation Occurs Soon after the Formation

(A–G) Centromeres were labeled with (A–G) CID-GFP and simultane-of Chromosome Territories

ously labeled with (B–G) Hoechst 33342. (A) Nonhomologous associ-Live spermatocytes homozygous for a lacO insert at 57A were ations of centromeres in early G2. Examples of nuclei at stage S2,stained with Hoechst 33342 to simultaneously detect chromatin showing different numbers and sizes of dots and indicating theand bound GFP-Lac repressor protein (chromatin: red, GFP: green, clustering of nonhomologous centromeres (1–3 dots), as well as theoverlap: yellow). separation of homologous centromeres (�4 dots). (B and C) The(A–C) A field of nuclei showing either (A) GFP labeling only or (B variable numbers of CID-GFP spots present in early stages getand C) GFP and Hoechst. resolved into four spots, as chromosome territories become distinct(D–J) Representative nuclei at different stages of development. In in early S3. Soon afterwards, individual spots duplicate (arrow in [D]).early stages, chromosomes were generally paired, with �95% of (E–G) In mature spermatocyte nuclei (late S3–S6), eight centromericnuclei showing a single dot (stage S2a, [D]). (E–G) At the S2b/S3 spots can generally be observed, with two spots per chromosometransition (the earliest stage with distinct chromosome territories), territory. The inset in (F) is a different view of the area indicated byboth paired and unpaired chromosomes can be observed. In the the arrow to show the presence of three GFP spots. CID-GFP islater stages (S4 and later, [H–J]) with clearly defined nuclear territor- shown in green, chromatin is shown in red, and the overlap is shownies, individual chromatids can be observed. (H) shows a field with in yellow. The scale bar represents 5 �m.two nuclei. (I)–(J) show a magnified view of the two chromosome 2territories from the nuclei in (H). Note the extensive separation ofchromatids in late G2 within the entire chromosome territory. The

of physical attachments both between sister chromatidsscale bar represents 5 �m.and between homologs and are in agreement with earliercytological studies that showed lengthwise alignment ofautosomal homologs at metaphase of meiosis I [30–33].meiosis, and it reaccumulates in newly formed daughter

nuclei during anaphase/telophase. These changes pro- During anaphase and telophase, chromosomes rapidlydecondensed, while GFP-Lac I accumulated in thevide a convenient developmental marker (Figure 6A).

Time-lapse images revealed that chromosomes re- newly formed daughter nuclei (Figures 6G and 6H). Atthis stage, each nucleus of the secondary spermatocytemained unpaired and decondensed during the early

stages of nuclear envelope breakdown (Supplementary inherited one homolog and showed fluorescent spotsrepresenting both sister chromatids, as expected afterMaterial to [40], and data not shown). In prometaphase,

chromosomes moved away from the nuclear periphery reductional division. The distance between spots in-creased again during telophase, as sister loci dispersedand migrated toward the nuclear center while rapidly

condensing. During this process, the distance between within the nucleus of secondary spermatocytes.sister GFP spots rapidly decreased from an averagedistance of �2.5–3 �m to �0.5–1.0 �m (Figures 6E and Discussion6F). However, even at maximum chromosome conden-sation, individual chromatids could still be resolved and Cytological studies of meiotic chromosome pairing in

male Drosophila have focused mostly on the prometa-often appeared to be slightly out of register, suggestinga less intimate association than that observed in early phase-anaphase stages of meiosis I, due to the inability

of conventional cytological methods to resolve individ-G2. Nevertheless, sister chromatids remained in closeproximity to each other, including near telomeric re- ual chromosomes in earlier stages of spermatocyte de-

velopment. Consequently, the early stages of meioticgions, during and after condensation. A similar behaviorof tagged loci was observed in homozygous lacO lines pairing in male Drosophila have not been observed.

Here, we report the use of a live cytological approach(data not shown). These results suggest the presence

Dynamics of Meiotic Chromosome Pairing1479

Figure 6. Chromosome Dynamics during the First Meiotic Division

Nuclei from a line heterozygous for a double insert on chromosome 2were tracked by time-lapse microscopy from late prophase throughtelophase of meiosis I.(A) A diagram of the location of lacO sites (cytological location: 54Aand 60F). Figure 7. A Model for Meiotic Pairing in Male Drosophila(B–D) Time-lapse images of a cluster of spermatocyte nuclei at (A) A summary of spermatocyte development. Only nuclei are out-different stages of prometaphase. Images were collected at 10- lined. Shading indicates chromatin. The formation of chromosomemin intervals. Note the breakdown of the nuclear envelope and the territories marks the beginning of stage S3 [36].dispersion of nuclear GFP-Lac I in the cytoplasm. (B) A model for meiotic chromosome pairing in male Drosophila.(E and F) (E) Three nuclei at or near metaphase. A magnified view Pairing patterns of one telocentric and one metacentric chromo-of the indicated chromosomes (arrow) is shown in (F). Spots of some are represented, with their euchromatic regions shown in redsimilar intensities most probably belong to the same chromatid. and blue, respectively. Heterochromatin is shown in gray. Centro-(G and H) (G) The same cells at telophase. Note the reappearance meres are represented as filled circles. We propose that homologousof nuclear GFP-Lac repressor in the six newly formed secondary interactions in the euchromatin are responsible for the initiationspermatocyte nuclei. A magnified view of the indicated nucleus is of meiotic pairing early during spermatocyte development, whileshown in (H). The arrowhead points to the sister nucleus carrying heterochromatic regions associate nonspecifically. Separationthe unlabeled homolog. (sorting) of chromosomes into distinct territories at the beginning

of stage S3 in mid-G2 disrupts nonspecific heterochromatic interac-tions between nonhomologous chromosomes while preserving in-teractions between heterochromatic regions on homologous chro-based on GFP-Lac repressor/lac operator recognitionmosomes. At this stage, there is extensive separation of homologsto track chromosome behavior at all stages of develop-and sister chromatids along the euchromatic chromosome arms.

ment in the male Drosophila germline. Our results sug- Homologous centromeres also separate, possibly due to microtu-gest a three-step model for meiotic chromosome pairing bule forces (arrows), while sister centromeres remain tightly pairedin male Drosophila (Figure 7). until the second meiotic division. We propose that, in late G2, homol-

ogous chromosomes are held together by a non-sequence-specificmechanism (e.g., heterochromatic interactions or chromatid entan-

Initiation glement). These interactions are terminated at anaphase I.In Drosophila, homologous chromosomes are generallypaired in somatic cells. Such somatic pairing has beenwell documented in embryos and larval tissues [45, 46]. stages were not investigated further, the evidence sug-

gests that homologous chromosomes are paired duringSomatic pairing of homologs has also been observedduring prophase and anaphase of the last spermatogo- G1 and remain paired through S phase. Our results are

therefore consistent with a somatic-type pairing of chro-nial mitotic division [28, 29]. This led to the hypothesisthat meiotic pairing in male Drosophila could be related mosomes in young spermatocyte nuclei.

The labeling of centromeres with CID-GFP also re-to somatic pairing and that homologs might enter thespermatocyte life cycle in a paired state ([28]; reviewed vealed interactions in the pericentric heterochromatin

of young spermatocytes. Many of these interactions,in [33]).Our results confirm that pairing of homologs takes however, appeared to take place between nonhomolo-

gous chromosomes, as indicated by the large percent-place in spermatogonial interphase nuclei. Furthermore,homologs were paired in the earliest stages of sperma- age of nuclei with less than four CID-GFP spots. This

situation is reminiscent of chromocenter formation intocyte development. The virtual absence of unpairedchromosomes suggests that pairing is very efficient, somatic cells. These results altogether suggest that, in

early spermatocyte development, homologous interac-stable, and occurs very early in spermatocyte nuclei,possibly as early as anaphase of the last mitotic gonial tions are restricted to euchromatic regions of the au-

tosomes, while heterochromatic regions interact non-division, as suggested by Metz [28]. Although these early

Current Biology1480

specifically. The findings are consistent with a role of are unknown. The location of these territories just under-neath the nuclear envelope suggests that chromosome-euchromatin in the initiation of autosomal meiotic pair-

ing in male Drosophila and are in agreement with earlier nuclear envelope attachments may play a role in territoryformation. This hypothesis is further supported by ourgenetic studies [23, 47].observation that, in late G2, GFP-tagged loci often ap-pear in close proximity to the nuclear surface (data not

Sorting shown). Our studies on chromosome motion in sperma-In mid-G2, spermatocyte chromatin separates into three tocytes also suggested tethering of the chromatin fiberdistinct clumps. In prometaphase, each clump appears either to the nuclear envelope or to other internal struc-to condense into a set of paired homologous chromo- tures [40]. Therefore, the separation of bivalents intosomes, suggesting that each clump corresponds to a territories could be the result of the stable attachmentchromosome territory containing one set of paired ho- of the chromosomes to sites on the nuclear envelopemologs. While a single mechanism has generally been (or some other internal nuclear structure), followed bythought to account for continuous homolog pairing dur- nuclear expansion. How such a mechanism would en-ing spermatocyte development, our results clearly show sure that paired homologs are specifically recognizedthat euchromatic regions abruptly separate soon after and segregated remains to be determined.the formation of chromosome territories in mid-G2. Sur-prisingly, the separation affected not only homologous

Maintenancechromosomes, but also sister chromatids. Furthermore,While the associations along the euchromatic autoso-sister chromatids appeared to separate individually frommal arms and at the centromeres are disrupted afterthe bivalent. This observation suggests either a commonthe formation of chromosome territories, homologousmechanism for homolog and sister cohesion or a singlechromosomes remain in close physical proximity withinmechanism for cohesion release between both sisterseach territory during late spermatocyte development.and homologs.Similarly, classic cytological studies have shown thatThe formation of chromosome territories led to thehomologs (and sister chromatids) appear to be aligneddisruption of nonhomologous heterochromatic interac-and juxtaposed during prometaphase and metaphasetions. In early stage S3, each territory contained a single[30–33], suggesting the existence of physical associa-CID-GFP spot, indicating that homologous centromerestions between chromatids. Our analysis of GFP-taggedwere paired. This association was short lived, however,chromosomes confirms that homologous (or sister) lociand homologous centromeres separated shortly aftercome back into close proximity during chromosomethe separation of euchromatic regions. One notable dif-condensation, without, however, showing the intimateference, however, was the continuous pairing of sisterassociation seen in G1 and early G2 (Figure 6). Thesecentromeres, which persisted through the first meioticresults altogether indicate that homologous chromo-division. Some evidence for the separation of homolo-somes remain physically linked until the metaphase-gous centromeres in prometaphase spermatocyte nu-anaphase transition of meiosis I. These late associa-clei has been previously reported [48]. Our results indi-tions, however, cannot be mediated by the intimate,cate that the separation is more extensive, oftensequence-specific euchromatic interactions observedreaching several microns, and occurs much earlier dur-in young spermatocytes. Similarly, it appears that inti-ing spermatocyte development. The functional signifi-mate connections between homologs do not persist atcance of this separation is unclear. A separation of cen-centromeres.tromeres during mitotic prophase has been observed in

humans [49], and yeast [50], and could be part of thetension-sensing mechanism that takes place during pro- What Holds Homologs Together

during Meiotic Prophase?phase [51, 52]. There is evidence for a meiotic spindlecheckpoint in male Drosophila [53, 54]. The separation of What is the nature of the attachments during late pro-

phase and metaphase? Our results rule out extensivehomologous centromeres in spermatocyte nuclei couldtherefore represent the initial interaction of kinetochores pairing along the chromosome arms or pairing at centro-

meres. One possible mechanism could involve the en-with microtubules and their separation under tensionbefore the alignment of chromosomes at the metaphase tanglement of homologous chromosomes. Topoisomer-

ase II is required for mitotic sister chromatid separationplate. If this hypothesis is correct, the binding of kineto-chores to microtubules, and possibly the spindle check- [55]. In Drosophila, the barren product is also required

for sister separation and possibly acts by regulatingpoint, would occur relatively early during spermatocytedevelopment. Separation, however, could also simply topoisomerase II function [56]. It has been proposed

that chromosome replication produces highly entangledreflect a lack of cohesion in the adjacent heterochro-matin. sister chromatids and that resolution of such tangles by

topoisomerase II might require chromosome condensa-We propose that the formation of chromosome territo-ries in stage S3 may provide a sorting mechanism that tion or the tension generated by the spindle at anaphase

[57]. In spermatocyte nuclei, entanglement could also besegregates individual bivalents into physically distinctdomains. Such a sorting mechanism may ensure that the the result of sustained diffusive motion of chromosomes

during the early stages of spermatocyte developmentassociation between homologs, established in youngspermatocytes through euchromatic interactions, per- [40] or topoisomerase II activity. Paired chromatids,

however, undergo rapid and extensive separation insists in late spermatocytes after the unpairing of theeuchromatic chromosome arms. The mechanisms re- mid-G2, suggesting that entanglement does not offer a

substantial obstacle to the separation of decondensedsponsible for the formation of chromosome territories

Dynamics of Meiotic Chromosome Pairing1481

of the bacterial Lac repressor protein lacking sequences necessarychromatids. We cannot exclude, however, that somefor tetramer formation was used to prevent adventitious interactionstangles remain in late G2. Even a few tangles couldbetween individual inserts. The GFP-Lac I fusion also contains aprovide a relatively strong structural linkage betweenC-terminal SV40 NLS for nuclear targeting. Expression of GFP-Lac

chromosomes [58] and could be responsible for bringing I in the male germline was achieved from the hsp83 promoter. GFP-all the chromatids into close proximity during condensa- Lac I was also cloned in the pUASP vector [65], and transgenic lines

were screened for expression in the male germline. The differenttion. In this model, these tangles would persist untilvectors used have been described [40]. Transgenic lines expressinganaphase, at which their removal, possibly through aGFP-Lac I in the male germline, or carrying a 256-mer lacO array,regulated topoisomerase II-dependent reaction, wouldwere obtained by P element-mediated transformation. Chromo-allow the separation of homologs. Such a mechanismsomal insertion sites of the lacO array were determined by in situ

would be consistent with the progressive displacement hybridization on polytene chromosomes. The location of sites usedof chromosomal associations toward the distal ends of in this study is as follows: 1F, 28C–E, 28E, 50F, 52–54, 53–54, 54A,

57A, 57C, 60A–B, 60F, 67B–D, 79B–D, 100D–F. Several unmappedthe chromosome arms at anaphase, as was previouslyinserts were also used.observed [30–33].

A CID-GFP fusion protein under control of the cid promoter [36]Another possibility involves cohesion in the hetero-was used to generate transgenic flies (J.S. Platero and S. Henikoff;chromatin. In Drosophila, chromosomes with large het-personal communication). The aly alleles have been described [41].

erochromatic deletions are somatically unstable, sug-gesting a direct role for heterochromatin in sister

In Vitro Tissue Culturechromatid cohesion in mitotic cells [59]. Heterochroma- Spermatocyte cysts were cultured in a sealed microscope chambertin was also shown to be required for cohesion at the as described [40]. Under our culture conditions, tissues remained

healthy for up to 10–12 hr, as indicated by the absence of visiblecentromeres during yeast mitosis [60, 61], possiblymorphological changes and by the fact that spermatocytes werethrough the assembly of a distinct cohesin complex. Aable to complete both meiotic divisions. When time-lapse imagingrole for heterochromatin in achiasmate meiotic pairingwas not required, cysts were mounted in a drop of testis bufferin female Drosophila has also been established throughinside a ring of dried nail polish. Under these conditions, the three-

genetic and cytological studies [10–12]. Distinct mecha- dimensional structure of nuclei was partially preserved, and evapo-nisms for removal of cohesion along chromosome arms ration was delayed. Nuclei imaged under these conditions for short

periods of time (15–30 min) were undistinguishable from nucleiand near centromeres have been identified in Drosophilagrown in culture and were optically superior. When indicated,[42] and in vertebrates [62]. It is therefore possible thatHoechst 33342 (2 �g/ml) was added to counterstain chromosomesheterochromatin cohesion could act as a very generalin live nuclei. Developmental stages were identified with morpholog-system for maintaining sister chromatid or homolog as-ical criteria [36, 44].

sociation during the late stages of both mitotic and mei-otic prophase. Three-Dimensional Fluorescence Microscopy

A role for heterochromatin in the persistence of mei- Three-dimensional images were acquired with a scientific gradeotic pairing in male Drosophila is consistent with classic cooled CCD camera (Photometrics) mounted on an Olympus IMT-2

inverted fluorescence microscope, as described [40]. For three-cytological observations on the sex chromosome pair.dimensional data sets, optical sections were collected at 0.25- toWhile genetic studies have demonstrated that X-Y pair-0.4-�m intervals along the vertical axis. Out-of-focus informationing requires specific sequences in the rDNA genewas removed by using a constrained iterative deconvolution algo-repeats [20–22], studies by Cooper suggest that X-Yrithm [66].

pairing sites (“collochores”) map to specific heterochro-matic sequences in the X and Y chromosomes that are

Acknowledgmentsdistinct from the rDNA gene repeats [31]. In this model,initial sequence-specific interactions at the rDNA gene The authors are greatly indebted to S. Henikoff and J.S. Platero for

their generous gift of the CID-GFP transgenic line and to M. Fullerrepeats would be responsible for the initial pairing offor kindly providing aly and other meiotic mutant strains. We wouldthe X-Y chromosome pair, while less specific hetero-like to thank Bruce McKee, John Tomkiel, S. Henikoff, David Agard,chromatic interactions would be responsible for theand members of the Sedat and Cande labs for their critical com-maintenance of X-Y association in late spermatocytes.ments and suggestions. This research was supported by National

Since euchromatic homology is necessary for the initial Institutes of Health grants GM58460 (to J.W.S.) and GM58460 (torecognition and pairing of homologs, this model is con- A.S.B.). J.V. was supported in part by a senior postdoctoral fellow-sistent with the apparent lack of a role for heterochroma- ship of the American Cancer Society, California Division (#1-17-00).tin in meiotic pairing deduced from genetic studies [23,63, 64]. Received: May 7, 2002

Revised: July 1, 2002In both cases, i.e., heterochromatic interactions orAccepted: July 4, 2002chromatid entanglement, the mechanism responsiblePublished: September 3, 2002for the maintenance of homologous chromosome pair-

ing probably involves non-sequence-specific mecha-Referencesnisms. It is likely that the formation of chromosome

territories provides a mechanism to ensure that such 1. Nasmyth, K. (2001). Disseminating the genome: joining, resolv-interactions are restricted only to homologous chromo- ing, and separating sister chromatids during mitosis and meio-somes. To our knowledge, this would represent the first sis. Annu. Rev. Genet. 35, 673–745.

2. Uhlmann, F. (2001). Chromosome cohesion and segregation ininstance of a specific role for chromosome territories.mitosis and meiosis. Curr. Opin. Cell Biol. 13, 754–761.

3. Rockmill, B., and Roeder, G.S. (1994). The yeast med1 mutantExperimental Proceduresundergoes both meiotic homolog non-disjunction and preco-cious separation of sister chromatids. Genetics 136, 65–74.Fly Strains

4. Buonomo, S.B.C., Clyne, R.K., Fuchs, J., Loidl, J., Uhlmann, F.,The GFP-Lac repressor fusion protein (GFP-Lac I) and the lac opera-tor (lacO) array have been described [37, 38]. A truncated version and Nasmyth, K. (2000). Disjunction of homologous chromo-

Current Biology1482

somes in meiosis I depends on proteolytic cleavage of the mei- 27. Orr-Weaver, T. (1995). Meiosis in Drosophila: seeing is believing.Proc. Natl. Acad. Sci. USA 92, 10443–10449.otic cohesin Rec8 by separin. Cell 103, 387–398.

5. Bickel, S.E., Orr-Weaver, T.L., and Balicky, E.M. (2002). The 28. Metz, C.W. (1926). Observations on spermatogenesis in Dro-sophila. Z. Zellforsch. und mikroskop. Anat. 4, 1–28.sister-chromatid cohesion protein ORD is required for chiasma

maintenance in Drosophila oocytes. Curr. Biol. 12, 925–929. 29. Stevens, N.M. (1908). A study of the germ cells of certain Diptera,with reference to the heterochromosomes and the phenomena6. Miyazaki, W.Y., and Orr-Weaver, T.L. (1994). Sister-chromatid

cohesion in mitosis and meiosis. Annu. Rev. Genet. 28, 167–187. of synapsis. J. Exp. Zool. 5, 359–374.30. Darlington, C.D. (1934). Anomalous chromosome pairing in the7. Lee, J.Y., and Orr-Weaver, T.L. (2001). The molecular basis of

sister-chromatid cohesion. Annu. Rev. Cell Dev. Biol. 17, male Drosophila pseudo-obscura. Genetics 19, 95–118.31. Cooper, K.W. (1964). Meiotic conjunctive elements not involving753–777.

8. Walker, M.Y., and Hawley, R.S. (2000). Hanging on to your homo- chiasmata. Proc. Natl. Acad. Sci. USA 52, 1248–1255.32. Cooper, K.W. (1949). The cytogenetics of meiosis in Drosophila.log: the roles of pairing, synapsis and recombination in the

maintenance of homolog adhesion. Chromosoma 109, 3–9. Mitotic and meiotic autosomal chiasmata without crossing-overin the male. J. Morph. 84, 81–122.9. Wolf, K.W. (1994). How meiotic cells deal with non-exchange

chromosomes. Bioessays 16, 107–114. 33. Cooper, K.W. (1950). Normal spermatogenesis in Drosophila. InBiology of Drosophila, M. Demerec, ed. (New York: John Wiley10. Hawley, R.S., Irick, H., Zitron, A.E., Haddox, D.A., Lohe, A., New,

C., Whitley, M.D., Arbel, T., Jang, J., McKim, K., et al. (1993). and Sons), pp. 1–61.34. Carpenter, A.T.C. (1975). Electron microscopy of meiosis in Dro-There are two mechanisms of achiasmate meiosis in Drosophila

females, one which requires heterochromatic homology. Dev. sophila melanogaster females. I. Structure, arrangement andtemporal change of the synaptonemal complex in wild-type.Genet. 13, 440–467.

11. Karpen, G.H., Le, M.-H., and Le, H. (1996). Centric heterochro- Chromosoma 51, 157–182.35. Lin, H.-P.P., Ault, J.G., and Church, K. (1981). Meiosis in Dro-matin and the efficiency of achiasmate disjunction in Drosophila

female meiosis. Science 273, 118–122. sophila melanogaster. I. Chromosome identification and kineto-chore microtubule numbers during the first and second meiotic12. Dernburg, A.F., Sedat, J.W., and Hawley, R.S. (1996). Direct

evidence of a role for heterochromatin in meiotic chromosome divisions in males. Chromosoma 83, 507–521.36. Cenci, G., Bonnacorsi, S., Pisano, C., Verni, F., and Gatti, M.segregation. Cell 86, 135–146.

13. Koehler, K.E., and Hassold, T.J. (1998). Human aneuploidy: les- (1994). Chromatin and microtubule organization during premei-otic, meiotic and early post meiotic stages of Drosophila mela-sons from achiasmate segregation in Drosophila melanogaster.

Ann. Hum. Genet. 62, 467–479. nogaster spermatogenesis. J. Cell Sci. 107, 3521–3534.37. Robinett, C.C., Straight, A.F., Li, G., Willhelm, C., Sudlow, G.,14. Morgan, T.H. (1912). Complete linkage in the second chromo-

some of the male Drosophila. Science 36, 719–720. Murray, A.W., and Belmont, A.S. (1996). In vivo localization ofDNA sequences and visualization of large-scale chromatin or-15. Meyer, G.F. (1960). The fine structure of spermatocyte nuclei

of Drosophila melanogaster. In Proceedings of the European ganization using lac operator/repressor recognition. J. Cell Biol.135, 1685–1700.Regional Conference on Electron Microscopy, A.L. Houwink and

B.J. Spit, eds. (Delft, The Netherlands: Nederlandse Vereiging 38. Straight, A.F., Belmont, A.S., Robinett, C.C., and Murray, A.W.(1996). GFP tagging of budding yeast chromosomes revealsvoor Electroninmicroscopie), pp. 951–954.

16. Rasmussen, S.W. (1973). Ultrastructural studies of spermato- that protein-protein interactions can mediate sister chromatidcohesion. Curr. Biol. 6, 1599–1608.genesis in Drosophila melanogaster. Meigen. Z. Zellforsch. Mi-

krosk. Anat. 140, 125–144. 39. Henikoff, S., Ahmad, K., Platero, J.S., and van Steensel, B.(2000). Heterochromatic deposition of centromeric histone H3-17. Gershenson, S. (1933). Studies on the genetically inert region

of the X chromosome of Drosophila. I. Behavior of an X chromo- like proteins. Proc. Natl. Acad. Sci. USA 97, 716–721.40. Vazquez, J., Belmont, A.S., and Sedat, J.W. (2001). Multiplesome deficient for a part of the inert region. J. Genet. 28,

297–312. regimes of constrained chromosome motion are regulated inthe interphase Drosophila nucleus. Curr. Biol. 11, 1227–1239.18. Sandler, L., and Braver, G. (1954). The meiotic loss of unpaired

chromosomes in Drosophila melanogaster. Genetics 39, 41. Lin, T.-Y., Viswanathan, S., Wood, C., Wilson, P.G., Wolf, N.,and Fuller, M.T. (1996). Coordinate developmental control of365–377.

19. Peacock, W.J. (1965). Nonrandom segregation of chromosomes the meiotic cell cycle and spermatid differentiation in Drosophilamales. Development 122, 1331–1341.in Drosophila males. Genetics 51, 573–583.

20. McKee, B.D., and Karpen, G.H. (1990). Drosophila ribosomal 42. Warren, W.D., Steffensen, S., Lin, E., Coelho, P., Loupart, M.,Cobbe, N., Lee, J.Y., McKay, M.J., Orr-Weaver, T.L., Heck,genes function as an X-Y pairing site during male meiosis. Cell

61, 61–72. M.M.S., et al. (2000). The Drosophila RAD21 cohesin persistsat the centromere region in mitosis. Curr. Biol. 10, 1463–1466.21. McKee, B.D., Habera, L., and Vrana, J.A. (1992). Evidence that

intergenic spacer repeats of Drosophila melanogaster function 43. Blower, M.D., and Karpen, G.H. (2001). The role of DrosophilaCID in kinetochore formation, cell cycle progression and hetero-as X-Y pairing sites in male meiosis, and a general model for

achiasmate pairing. Genetics 132, 529–544. chromatin interactions. Nat. Cell Biol. 3, 730–739.44. Fuller, M.T. (1993). Spermatogenesis. In The Development of22. Merrill, C.J., Chakravarti, D. Habera, L., Das, S., Eisenhour, L.,

and McKee, B.D. (1992). Promoter-containing ribosomal DNA Drosophila, M. Bate and A. Martinez-Arias, eds. (Cold SpringHarbor, New York: Cold Spring Harbor Press), pp. 71–147.fragments function as X-Y meiotic pairing sites in D. melanogas-

ter males. Dev. Genet. 13, 468–484. 45. Hiraoka, Y., Dernburg, A.F., Parmelee, S.J., Rykowski, M.C.,Agard, D.A., and Sedat, J.W. (1993). The onset of homologous23. McKee, B.D., Lumsden, S.E., and Das, S. (1993). The distribution

of male meiotic pairing sites on chromosome 2 of Drosophila chromosome pairing during Drosophila melanogaster em-bryogenesis. J. Cell Biol. 120, 591–600.melanogaster: meiotic pairing and segregation of 2-Y transposi-

tions. Chromosoma 102, 180–194. 46. Fung, J.C., Marshall, W.F., Dernburg, A., Agard, D.A., and Sedat,J.W. (1998). Homologous chromosome pairing in Drosophila24. Baker, B.S., and Carpenter, A.T. (1972). Genetic analysis of sex

chromosomal meiotic mutants in Drosophila melanogaster. Ge- melanogaster proceeds through multiple independent initia-tions. J. Cell Biol. 141, 5–20.netics 71, 255–286.

25. Baker, B.S., Carpenter, A.T.C., Esposito, M.S., Esposito, R.E., 47. McKee, B.D. (1996). The license to pair: identification of meioticpairing sites in Drosophila. Chromosoma 105, 135–141.and Sandler, L. (1976). The genetic control of meiosis. Annu.

Rev. Genet. 10, 53–134. 48. Tang, T.T.-L., Bickel, S.E., Young, L.M., and Orr-Weaver, T.L.(1998). Maintenance of sister-chromatid cohesion at the centro-26. Tomkiel, J.E., Wakimoto, B.T., and Briscoe, A., Jr. (2001). The

teflon gene is required for maintenance of autosomal homolog mere by the Drosophila MEI-S332 protein. Genes Dev. 12, 3843–3856.pairing at meiosis I in male Drosophila melanogaster. Genetics

157, 273–281. 49. Shelby, R.D., Hahn, K.M., and Sullivan, K.F. (1996). Dynamic

Dynamics of Meiotic Chromosome Pairing1483

elastic behavior of alpha-satellite DNA domains visualized insitu in living human cells. J. Cell Biol. 135, 545–557.

50. He, X., Asthana, S., and Sorger, P.K. (2000). Transient sisterseparation and elastic deformation of chromosomes during mi-tosis in budding yeast. Cell 101, 763–775.

51. Rieder, C.L., Schultz, A., Cole, R., and Sluder, G. (1994). Ana-phase onset in vertebrate somatic cells is controlled by a check-point that monitors sister kinetochore attachment to the spindle.J. Cell Biol. 127, 1301–1310.

52. Nicklas, R.B., Ward, S.C., and Gorbsky, G.J. (1995). Kinetochorechemistry is sensitive to tension and may link mitotic forces toa cell cycle checkpoint. J. Cell Biol. 130, 929–939.

53. Basu, J., Logarinho, E., Li, Z., Williams, B.C., Lopes, C., Sunkel,C.E., and Goldberg, M.L. (1999). Mutations in the essential spin-dle checkpoint gene bub1 cause chromosome missegregationand fail to block apoptosis in Drosophila. J. Cell Biol. 146, 13–28.

54. Rebollo, E., and Gonzalez, C. (2000). Visualizing the spindlecheckpoint in Drosophila spermatocytes. EMBO Rep. 11, 65–70.

55. Holm, C., Goto, T., Wang, J.C., and Botstein, D. (1985). DNAtopoisomerase II is required at the time of mitosis in yeast. Cell41, 553–563.

56. Bhat, M.A., Philp, A.V., Glover, D.M., and Bellen, H.J. (1996).Chromatid segregation at anaphase requires the barren prod-uct, a novel chromosome-associated protein that interacts withtopoisomerase II. Cell 87, 1103–1114.

57. Holm, C. (1994). Coming undone: how to untangle a chromo-some. Cell 77, 955–957.

58. Duplantier, B., Jannink, G., and Sikorav, J.-L. (1995). Anaphasechromatid motion: involvement of type II topoisomerases. Bio-phys. J. 69, 1596–1605.

59. Wines, D., and Henikoff, S. (1992). Somatic instability of a Dro-sophila chromosome. Genetics 131, 683–691.

60. Bernard, P., Maure, J.-F., Partridge, J.F., Genier, S., Javerzat,J.-P., and Allshire, R.C. (2001). Requirement of heterochromatinfor cohesion at centromeres. Science 294, 2539–2542.

61. Nonaka, N., Kitajima, T., Yokobayashi, S., Xiao, G., Yamamoto,M., Grewal, S.I.S., and Watanabe, Y. (2002). Recruitment ofcohesin to heterochromatic regions by Swi6/HP1 in fissionyeast. Nat. Cell Biol. 4, 89–93.

62. Waizenegger, I.C., Hauf, S., Meinke, A., and Peters, J.M. (2000).Two distinct pathways remove mammalian cohesin from chro-mosome arms in prophase and from centromeres in anaphase.Cell 103, 399–410.

63. Yamamoto, M. (1972). Cytological studies of heterochromatinfunction in Drosophila melanogaster male: autosomal meioticpairing. Chromosoma 72, 293–328.

64. Hilliker, A.J., Holm, D.G., and Appels, R. (1982). The relationshipbetween heterochromatic homology and meiotic segregationof compound second autosomes during spermatogenesis inDrosophila melanogaster. Genet. Res. 39, 157–168.

65. Rørth, P. (1998). GAL4 in the Drosophila female germ line. Mech.Dev. 70, 113–118.

66. Agard, D.A., Hiraoka, Y., Shaw, P., and Sedat, J.W. (1989). Fluo-rescence microscopy in three dimensions. Methods Cell Biol.30, 353–377.