SEE COMMENTARY
The Drosophila cohesin subunit Rad21 is a trithorax group (trxG) protein Graham Hallson†, Monika Syrzycka†, Samantha A. Beck‡, James A. Kennison§, Dale Dorsett¶, Scott L. Page储, Sally M. Hunter储, Rebecca Keall储, William D. Warren储, Hugh W. Brock‡, Donald A. R. Sinclair†, and Barry M. Honda†,†† †Department
of Molecular Biology and Biochemistry, Simon Fraser University, Burnaby, BC, Canada V5A 1S6; ‡Department of Zoology, University of British Columbia, Vancouver, BC, Canada V6T 1Z4; §Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD 20892-2785; ¶Department of Biochemistry and Molecular Biology, School of Medicine, St. Louis University, St. Louis, MO 63104; and 储Comparative Genomics Centre, James Cook University, Townsville 4811, Queensland, Australia
The cohesin complex is a key player in regulating cell division. Cohesin proteins SMC1, SMC3, Rad21, and stromalin (SA), along with associated proteins Nipped-B, Pds5, and EcoI, maintain sister chromatid cohesion before segregation to daughter cells during anaphase. Recent chromatin immunoprecipitation (ChIP) data reveal extensive overlap of Nipped-B and cohesin components with RNA polymerase II binding at active genes in Drosophila. These and other data strongly suggest a role for cohesion in transcription; however, there is no clear evidence for any specific mechanisms by which cohesin and associated proteins regulate transcription. We report here a link between cohesin components and trithorax group (trxG) function, thus implicating these proteins in transcription activation and/or elongation. We show that the Drosophila Rad21 protein is encoded by verthandi (vtd), a member of the trxG gene family that is also involved in regulating the hedgehog (hh) gene. In addition, mutations in the associated protein Nipped-B show similar trxG activity i.e., like vtd, they act as dominant suppressors of Pc and hhMrt without impairing cell division. Our results provide a framework to further investigate how cohesin and associated components might regulate transcription. Hedgehog 兩 heterochromatin 兩 Nipped-B 兩 Polycomb 兩 cohesion
I
n eukaryotic mitosis, accurate chromosome segregation requires paired sister chromatids to attach to opposite spindle poles. Sister chromatids are held together by the cohesin protein complex, which consists of four core subunits, Rad21/SCC1, stromalin (SA) and structural maintenance of chromosome (SMC) proteins SMC1 and SMC3. A widely accepted model postulates that cohesin forms a ring-like structure via interaction of the N- and C-termini of Rad21 with a SMC1/SMC3 heterodimer. With the participation of SCC2/Nipped-B, SCC4, EcoI/Ctf7, and Pds5 proteins, sister-chromatid cohesion is maintained until the onset of mitosis. Cleavage of Rad21 and the resulting removal of cohesin then allow separation of sister chromatids in anaphase (see refs. 1–3 for reviews). Mutation of genes encoding these subunits leads to errors in chromosome segregation and aneuploidy, which are hallmarks of cancer and a leading cause of birth defects in humans (4). Given the highly conserved role for cohesin in sister chromatid cohesion, it was unexpected to discover that cohesin and associated proteins might also play a distinct, independent role in regulating gene expression. Reduction in Nipped-B expression in Drosophila affects expression of the cut and Ultrabithorax genes (5–7), and mutations in the human orthologue, NIPBL, result in Cornelia de Lange Syndrome (8). In zebrafish, mutations in rad21 or Smc3 affect embryonic runx gene transcription in heterozygous mutant animals without compromising cell division, suggesting that these proteins may have functions in transcription that are distinct from a mitotic role (9). Recently, Misulovin et al. observed extensive overlap of Nipped-B and cohesin components with RNA polymerase II binding at active genes and apparent exclusion from genes silenced by Polycomb group (PcG) genes (10). This intriguing www.pnas.org兾cgi兾doi兾10.1073兾pnas.0801698105
chromatin immunoprecipitation (ChIP) result strongly suggests a role in transcription for cohesin and Nipped-B, although the mechanisms are unknown. Trithorax group (trxG) genes encode proteins implicated in transcriptional regulation. These genes were initially characterized as regulators of homeotic genes in Drosophila (11, 12). The trxG genes are required to maintain activation of homeotic and other genes; many that have been molecularly characterized encode members of multimeric complexes with roles in transcriptional initiation and/or elongation. Typically, mutations in trxG genes suppress the phenotypes of mutations in PcG genes, whose function is to maintain the repressed state of homeotic genes (13–16) and other developmentally important genes like hedgehog (hh) (17), a gene required for cell signaling (18). As part of our work toward a functional annotation of heterochromatin of D. melanogaster (19), we characterized the verthandi (vtd) locus, a member of the trxG gene family with Suppressor of Polycomb [Su(Pc)] activity (11, 20). The vtd locus also affects hh expression, as vtd mutations are dominant suppressors of Moonrat (Mrt), a dominant gain of function allele of hh (20, 21). However, because of its location deep within the centric heterochromatin of the left arm of chromosome 3 (22, 23), vtd has resisted characterization at the molecular level. Here, we report that vtd mutations, isolated on the basis of their trxG phenotypes, map to the gene encoding the cohesin subunit Rad21 and exhibit corresponding defects in mitosis and sister chromatid cohesion. Mutations in Nipped-B also show trxG phenotypes, and as is the case for vtd, heterozygous mutant flies show trxG phenotypes without significantly affecting cell division. Our results provide a link between sister chromatid cohesion proteins and trxG functions, thus suggesting that cohesion factors may act by facilitating transcription activation and/or elongation. Results Mutations in the trxG Gene vtd Show Corresponding Lesions in the rad21 Gene. To establish the location of known coding regions
relative to the heterochromatic genetic map (19, 20), we performed single embryo PCR (SEPCR) (22) on embryos carrying homozygous 3L deletions. In this way, the gene encoding the Author contributions: G.H., M.S., S.A.B., J.A.K., D.D., S.L.P., W.D.W., H.W.B., D.A.R.S., and B.M.H. designed research; G.H., M.S., S.A.B., J.A.K., S.L.P., S.M.H., R.K., and D.A.R.S. performed research; D.D. and R.K. contributed new reagents/analytic tools; G.H., M.S., S.A.B., J.A.K., D.D., S.L.P., W.D.W., H.W.B., D.A.R.S., and B.M.H. analyzed data; and G.H., M.S., S.A.B., J.A.K., D.D., S.L.P., W.D.W., H.W.B., D.A.R.S., and B.M.H. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. See Commentary on page 12097. ††To
whom correspondence should be addressed. E-mail:
[email protected].
This article contains supporting information online at www.pnas.org/cgi/content/full/ 0801698105/DCSupplemental. © 2008 by The National Academy of Sciences of the USA
PNAS 兩 August 26, 2008 兩 vol. 105 兩 no. 34 兩 12405–12410
GENETICS
Edited by Kathryn V. Anderson, Sloan-Kettering Institute, New York, NY, and approved June 3, 2008 (received for review February 20, 2008)
Table 1. Progeny resulting from w1118 P{UAST-Drad21WT-2HA}/ Y; ⴙ/ⴙ; vtd␥26 – 6/TM3, Ser males crossed with w*; P{Act5C-GAL4}25FO1/CyO; Df(3L)K2, e/TM3, Ser females 3rd Chromosome Genotype vtd␥26–6/TM3,
Ser vtd␥26–6/TM3, Ser vtd␥26–6/TM3, Ser vtd␥26–6/TM3, Ser Df(3L)K2/TM3, Ser Df(3L)K2/TM3, Ser Df(3L)K2/TM3, Ser Df(3L)K2/TM3, Ser vtd␥26–6/Df(3L)K2 vtd␥26–6/Df(3L)K2 vtd␥26–6/Df(3L)K2 vtd␥26–6/Df(3L)K2 Total
Fig. 1. vtd mutants contain defects in the rad21 gene. (A) Schematic map of the rad21 gene. The black bars represent the eight exons, the small arrows below indicate the positions of PCR primers used in DNA sequencing, and the extent of rad21 DNA removed in vtd4 and vtd14 is indicated by the lines below (the exact break points within intron 2 have not been determined). The exons are drawn approximately to scale, as indicated. (B) Multiple sequence alignment of the first 100 aa of D. melanogaster Rad21 with Drad21 homologues present in other species. vtd1-166-38 contains a nonsense mutation in the first exon of Rad21, changing W18 (TGG) to stop (TAG). vtd35 and vtd36 contain missense mutations in the first exon of Rad21, changing V49 (GTA) to E49 (GAA). (C) The strong EMS allele vtd␥26-6 shows an altered splice donor site at the 3⬘ end of the 5th exon, resulting in a premature stop codon.
Drosophila rad21 cohesin subunit (Rad21), known to map to centric heterochromatin (23) was localized to the genetically defined region of 3L containing the verthandi (l3/vtd) locus [see supporting information (SI) Text]. To determine whether lethal mutations in the vtd locus are associated with lesions in the rad21 coding region, homozygous mutant vtd embryos were isolated and gene-specific primers used to sequence across portions of the rad21 genomic locus (Fig. 1A). Typical of most heterochromatic genes, rad21 is very large, with eight small exons encoding a transcript of 2.2 kb spread over ⬎20 kb in the genome (23). The results, summarized in Fig. 1, strongly suggest that vtd encodes the Rad21 protein. The ethyl methanesulfonate (EMS)-induced allele vtd1–166-38 exhibits homozygous early embryonic lethality and is genetically a null or strongly hypomorphic allele. In vtd1–166-38, we identified a nonsense mutation in exon 1 of rad21, with a conserved tryptophan (W18) changed to a stop codon (Fig. 1B). Another strong EMS allele, vtd␥26 – 6, was found to have a mutated splice donor site at the end of exon 5 (Fig. 1C). The altered mRNA encoded by this allele is predicted to yield a truncated Rad21 protein that would be unable to interact with SMC1, and therefore, would be unable to form a closed cohesin ring structure. Homozygous vtd35 and vtd36 mutants have a weaker phenotype: they pupate but fail to emerge as adults. Consistent with this phenotype, the vtd35 and vtd36 alleles contain identical missense mutations that convert a conserved valine (V49) to glutamic acid within the N-terminal winged-helix domain (Fig. 1B) that is required for binding to SMC3. Both vtd35 and vtd36 12406 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0801698105
Act5C-GAL4
UAST-Drad21
Sex
Number
⫺ ⫺ ⫹ ⫹ ⫺ ⫺ ⫹ ⫹ ⫺ ⫺ ⫹ ⫹
⫺ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫹
Male Female Male Female Male Female Male Female Male Female Male Female
119 112 77 104 119 115 102 135 0 0 0 123 1006
were isolated in the same screen in which males were EMStreated, so it is possible that both derive from the same premeiotic mutagenic event. Several other vtd mutations have been characterized as genetic deficiencies (20, 21); vtd4 and vtd14 are shown in Fig. 1A. Together, our results clearly show that mutations in the trxG gene vtd are associated with molecular lesions within the rad21 gene. vtd Mutations are Rescued by rad21 Transgenes. If rad21 and vtd are the same gene, then it should be possible to rescue vtd mutations with a rad21 transgene. Appropriate crosses were performed to combine an X chromosome carrying a pUAST-rad21-HA transgene with a second chromosome carrying a ubiquitously expressed Act5C-Gal4 driver, into a vtd␥26 – 6/Df(3L)K2 genetic background. The results in Table 1 show that the rad21 transgene rescues the lethality—because the transgene is X-linked, only females are rescued (bottom row in Table 1). We recovered 123 vtd␥26-6/Df(3L)K2 females carrying the Act5C-Gal4 and UASTrad21-HA transgenes compared with zero vtd␥26-6/Df(3L)K2 females that carried UAST-rad21-HA without the driver. The number of vtd␥26-6/Df(3L)K2 females recovered (123) was similar to the numbers of vtd␥26-6/TM3, Ser (104) and Df(3L)K2/TM3, Ser (135) heterozygote females carrying both transgenes that were produced in the cross, indicating a robust rescue by the rad21 transgene. The rescue of vtd lethality by the rad21 transgene establishes that vtd and rad21 are the same locus. rad21 RNAi Effects Are Enhanced in a vtd Background. We hypoth-
esized that vtd mutations might enhance the level of lethality produced by depletion of wild-type Rad21 protein by using RNAi. We tested this using transgenic lines that produce rad21specific siRNA under the control of the GAL4-UAS promoter. When driven by the Act5C promoter, expression of rad21specific siRNA causes a reduction in the number of both male and female adults, with males more strongly affected than females. Three different rad21-RNAi transgenic lines constructed by Rollins et al. (6) were tested, with the results for RNAi transgenic line 3F-1M shown in Fig. 2; similar results were observed for two other RNAi lines, 2F-1M and 8M-1M (data not shown). Heterozygosity for the vtd allele (1-166-38) dramatically enhanced the lethal effects of rad21 RNAi. This is consistent with our previous evidence that rad21 and vtd are the same gene. Henceforth we will refer to this gene as vtd, and the encoded protein as Rad21. vtd Mutations Affect the Embryonic Cell Cycle. If vtd encodes the
Rad21 cohesin subunit, then vtd mutants might be expected to Hallson et al.
show mitotic defects consistent with failure of sister chromatid cohesion. Cell cycle progression is readily monitored in Drosophila embryos because the first 13 nuclear divisions are rapid and synchronous and lack or have very short gap phases. These early cell cycles are entirely controlled by maternally deposited proteins or RNAs, and defects resulting from mutations in Pc or mitotic checkpoint genes can be detected in embryos derived from heterozygous mutant mothers, regardless of the embryo’s zygotic genotype (ref. 24, and references therein). Mitosis occurs as a wave beginning at each pole, so that more central nuclei are at an earlier mitotic stage than are nuclei at the poles. In embryos derived from wild-type mothers, most nuclei exhibit tight alignment of chromosomes at the metaphase plate in mitosis 13. Close to the poles, a small number of nuclei in the metaphase-anaphase transition can be recognized because of their looser alignment. Nuclei at the poles are in anaphase or telophase, as shown by the separated chromosomes. In embryos derived from mothers carrying the vtd␥26 – 6 allele (i.e., only one functional wild-type copy), there is an increase in the number of nuclei exhibiting the loose alignment of chromosomes seen at the metaphase to anaphase transition, and the chromosomes appear to be less condensed (Fig. 3). We never observed this phenotype
vtd Mutations Disrupt Sister Chromatid Cohesion in Larval Neuroblasts. In budding and fission yeast and in mammalian cells,
depletion of Rad21 results in loss of cohesion establishment and maintenance (29–33). To ensure that the embryonic mitotic defects we observed are not anomalies resulting from the unusually rapid and synchronous nuclear division observed in embryos, we examined sister chromatid cohesion in neuroblast chromosomes from third instar larvae that were transheterozygous or hemizygous for rad21 alleles. For these experiments, we used the pupal lethal vtd35 allele in trans to either the deficiency Df(3L)K2, the vtd␥26-6 allele, or the vtd4 allele. In each case, the mutant hemizygotes or transheterozygotes remained viable through the third larval instar stage, although we noted that imaginal discs were not evident in the larvae, which is a phenotype associated with mutants that cause mitotic defects (data not shown). Results for metaphase chromosomes from control vtd35/ TM6B, P{Ubi-GFP.S65T}PAD2, Tb1 larvae indicated intact sister chromatid cohesion in centromeric regions (Fig. 4A). Cohesion along the chromosome arms was disrupted by the hypotonic conditions used during preparation of the chromosome spreads. In the control, a diploid complement of eight chromosomes comprised of 16 sister chromatids was apparent (12 sister chromatids represent the large X, 2nd, and 3rd chromosomes, whereas the small size of the 4th chromosome precludes visualization of individual sister chromatids). In contrast, metaphase chromosome spreads from vtd␥26-6/vtd35, vtd4/vtd35, or vtd35/ Df(3L)K2 consisted of individual sister chromatids that were not conjoined at the centromere (Fig. 4 B–D). Thus, in the presence of mutations in the rad21 gene, sister chromatid cohesion was abolished. In addition, the chromosome complements were frequently aneuploid. For example, the numbers of large chromatids (chromosomes X, 2, and 3) in Fig. 4 B and C are 22 and 18, respectively, versus the normal 12, whereas in Fig. 4D, the number is ⬎50. A lack of sister chromatid cohesion would prevent the normal coorientation of sister centromeres toward the poles at mitotic metaphase, leading to the uneven segregation of sister chromatids and abnormal chromosome numbers in the daughter cells. Taken together, the results establish that vtd mutations have mitotic defects consistent with failure of initiation and maintenance of sister chromatid cohesion. Nipped-B Mutations also Show trxG Phenotypes. Our results support
Fig. 3. vtd mutations result in altered embryonic cell cycle progression. Embryos from wild-type Oregon R (A), or heterozygous mutant vtd␥26-6/TM3 (␥-26-6/TM3) (B) mothers were DAPI-stained and photographed as described in Materials and Methods. Different zones are labeled as follows: A/T, nuclei in anaphase or telophase, shown by the separated chromosomes; M/A, nuclei in the metaphaseanaphase transition recognizable by their looser alignment; and M, nuclei at metaphase, showing condensation and alignment.
Hallson et al.
a model that the role of cohesin proteins is similar to the role of trxG proteins in regulating transcription. If so, mutations in other cohesin subunits or accessory proteins might be expected to exhibit trxG phenotypes. The Nipped-B protein is required for cohesion, presumably by facilitating binding of cohesin to chromosomes (6). Because vtd mutations are dominant suppressors of Pc and the gain of function hhMrt mutation, we tested Nipped-B mutations in the same genetic assays. Indeed, heterozygous PNAS 兩 August 26, 2008 兩 vol. 105 兩 no. 34 兩 12407
GENETICS
Fig. 2. Rad21 RNAi lethality is enhanced in a vtd background. The rad21 RNAi construct 3F-1M was introduced by appropriate genetic crosses and expressed under the control of the Actin5C-driven GAL4 promoter. Resultant survival of females (F) and males (M) was determined for flies carrying either one copy of vtd1-166-38(⫹) (Right), as compared with wild-type (⫹/⫹) (Left), and flies carrying a TM3 balancer chromosome (Center).
SEE COMMENTARY
in wild-type embryos in this or in previous studies (24). No mitotic abnormalities were observed before mitosis 11, and these embryos subsequently achieve normal metaphase figures with proper condensation and alignment (data not shown). This relatively weak embryonic phenotype presumably results from a reduction in maternally deposited Rad21 relative to wild-type. In yeast, Xenopus, and human, Rad21 homologues are cleaved to relieve sister chromatid cohesion at anaphase (25–27). Therefore, our interpretation is that the embryonic phenotype results from premature entry into anaphase. Syncytial embryos injected with rad21 dsRNA show multiple severe defects mostly consistent with chromosome breakage (28). These more severe defects are likely due to cumulative mitotic failures resulting from depletion of Rad21 below the levels present in embryos derived from heterozygous vtd(⫹) mothers.
Fig. 4. Failure of sister chromatid cohesion in vtd mutant neuroblasts. (A) Metaphase chromosome spread from vtd35/TM6B, P{Ubi-GFP.S65T}PAD2, Tb1 demonstrating cohesion between sister chromatids (arrowheads) in the centromeric region (arrow) of the chromosomes. Chromosomes X, 2, 3, and the internally rearranged 3rd chromosome balancer TM6B are indicated. The two small 4th chromosomes are marked with a single numeral 4. (Inset) 2X enlarged image of chromosome 3. (B) Metaphase chromosomes from a vtd␥26-6/vtd35 cell showing a deficiency of cohesion between sister chromatids (arrowheads), including the centromeric region (arrow). (Inset) 2X enlarged image of a pair of sister chromatids. (C and D) Metaphase chromosomes from vtd4/vtd35 (C) and vtd35/Df(3L)K2 (D) in which sister chromatids (arrowheads) are separated from each other and do not display cohesion (centromeric region of one chromatid is indicated with an arrow) (Scale bar, 5 m).
Nipped-B mutations moderately suppressed the wing phenotypes of hhMrt (Fig. 5), and weakly suppressed Pc4 (data not shown). In addition, we also isolated the Nipped-B29 mutation in the same screens for EMS-induced dominant suppressors of hhMrt, from which many of the vtd mutations were isolated (20). None of the mutations in other genes encoding cohesion factors tested, including pds5e3, pds5e6, and Smc1exc46, had consistent effects on Mrt or Pc4 (data not shown). Discussion Our data demonstrate that the Drosophila trxG gene vtd encodes the Rad21 cohesin subunit, and that mutations in Nipped-B likewise show trxG phenotypes, thus implicating cohesin and associated components in transcriptional activation. Alleles of vtd have lesions in rad21, mutations or knockdowns of rad21 have vtd phenotypes, and vice versa, and a transgene containing rad21 rescues the lethality of vtd. It is also noteworthy that reductions in Rad21 or Nipped-B dosage alter gene expression without seriously affecting chromatid cohesion, suggesting that these may be separable functions for cohesin and associated proteins. As noted in the Introduction, evidence has accumulated that cohesin and associated proteins have important roles in gene regulation, but the functional basis for this has been unclear. The simplest model that explains the existing data is that Rad21, like most other trxG proteins, facilitates transcription. In Drosophila, many trxG proteins are subunits of complexes with diverse roles in transcriptional activation. Trx and Ash1
encode SET domain proteins that methylate lysine 4 of histone H3 (H3K4), and Ash2 is a member of a complex that also methylates H3K4 (34, 35). Other trxG proteins (e.g., Brahma, Osa, Moira, Kismet) are members of ATP-dependent nucleosome remodeling complexes (13–15). However, despite concerted efforts from many laboratories, the precise mechanisms by which trxG proteins regulate transcription remain unclear. In addition to chromatin modification, trxG proteins appear to be directly involved in recruiting factors required for transcription elongation, and noncoding RNAs may also play a role in regulating some of the affected genes (36). Our hypothesis that cohesin facilitates transcription is supported by the results of a recent genome-wide ChIP study, which shows preferential binding of Nipped-B and the cohesin subunits SMC1 and SA to transcribed regions, overlapping with RNA polymerase II (Pol II) binding sites (10). The colocalization of Nipped-B with cohesin on chromosomes, and physical association with SA and Rad21 in extracts (7, 10) further suggests that Nipped-B and cohesin act together. There are strong correlations between binding of cohesin components and active gene expression. The dosage sensitive suppression of the hhMrt gain of function allele by both vtd and Nipped-B mutations suggests that Nipped-B and cohesin both promote expression of hh. It is unknown, however, if this effect is direct. Cohesin or Nipped-B do not bind to the hh gene in any of the three cell lines examined, however, in at least two of these, PcG proteins actively silence hh (10, 37). Genome-wide, PcG
Fig. 5. vtd and Nipped-B mutations result in dominant suppression of Mrt. (A) Wild-type wing. (B) Mrt(⫹) wing with vein defects (interior loss, extra veins at margin) and anterior overgrowth. (C and D) Typical results for dominant suppression by Df(3L)vtd(⫹) and NippedB29(⫹), respectively. 12408 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0801698105
Hallson et al.
Hallson et al.
SEE COMMENTARY
this idea, Rollins et al. (6) did observe effects of rad21 dosage on cut expression if they used RNAi to deplete rad21 mRNA, presumably to levels lower than those available in our vtd(⫹) heterozygotes. They also reported that Nipped-B expression is not directly proportional to gene dosage. Our data here also show that reductions in Nipped-B and rad21 dosage act in the same direction i.e., suppress Mrt and Pc, suggesting that both genes may contribute to gene activation. The fact that both the rad21 and Nipped-B genes are resident within a late-replicating, heterochromatic environment may also explain some differences in outcomes of genetics tests of cohesin subunit function. Our results provide a link between cohesin binding and trxG gene function. It will be an interesting challenge for the future to determine how components involved in chromatid cohesion act at the molecular level to regulate transcription, particularly given other very recent evidence implicating cohesin in gene regulation (40 – 43). The discovery that vtd encodes the Rad21 cohesin subunit expands the known roles of cohesin and Nipped-B in Drosophila development to include regulation of hh, which like cut, Ubx, and EcR, has many developmental roles. Similar modulation of key developmental regulators in humans, each with multiple roles, could explain why Cornelia de Lange syndrome patients have multiple diverse developmental deficits. Materials and Methods Culture Conditions. Fly cultures were grown on a standard cornmeal, yeast, and molasses medium with Tegosept and propionic acid added as mold inhibitors. All stocks and crosses were reared and maintained at 25°C unless otherwise stated.
Drosophila Stocks, Transgene Constructs, and Genetic Crosses. The majority of strains are listed on the FlyBase website (www.flybase.org), and most vtd alleles have been genetically characterized (20). Deficiencies TTT, e*#54, e#406, e#301, and 7B-75 were recovered from X-ray screens (M.S., unpublished data) as described in ref. 19. Deficiency Df(3L)vtd14 was recovered as an EMS-induced dominant enhancer of brm2 (J. Kennison, unpublished). Assays for Su(Mrt) activity were performed as described in ref. 20. The pP{UAST-Drad21WT-2HA} cDNA construct used for phenotypic rescue was generated as follows. Two copies of a linker sequence encoding the HA epitope were ligated into the BstBI site located 16 bp before the stop codon in the rad21 cDNA clone pLD02527. The modified cDNA was digested with KpnI and EcoRI and cloned into KpnI/EcoRI digested pUAST (44) by using standard methods (45). Transgenic lines carrying these constructs were generated in a w1118 background using standard procedures (46), and an X-linked transgene was chosen for these analyses. The rad21-RNAi lines have been described (6). For testing phenotypic rescue, w1118 P{UAST-Drad21WT-2HA}/Y; (⫹/⫹); vtd␥26-6/ TM3, Ser males and w*; P{Act-GAL4}25FO1/CyO; Df(3L)K2, e/TM3, Ser females were generated using standard genetic methods and crossed together under standard conditions. Progeny genotypes were identified via visible phenotypic markers and counted until the 18th day following cross setup. Single Embryo PCR. Several vtd mutant alleles were balanced over a TM3 chromosome containing a GFP transgene. Genomic DNA was isolated from individual embryos of each strain and PCR was used (22) to identify samples originating from homozygous embryos containing vtd defects. Primers were designed to amplify the GFP transgene (330 bp fragment) and an X-linked gene used to evaluate DNA integrity (Grp84, 215 bp fragment). See SI Text and Fig. S1 for further details. DNA specimens for which PCR amplification detected the presence of Grp84 and the absence of the GFP transgene were used in subsequent analysis. Using DNA from homozygous embryos, PCR was performed using gene specific primers designed to anneal to regions flanking the eight exons of rad21. See SI for further details. The rad21 genomic sequence was obtained by aligning the known rad21 cDNA sequence to the corresponding FlyBase genomic scaffold (47) using the BLASTN algorithm (www.flybase.org). Sequence Analysis. DNA sequencing using rad21 gene specific primers was done by Macrogen Inc. (Seoul, Korea). The rad21 coding sequence of vtd mutants was compared with the annotated sequence present in the FlyBase June 2007 assembly (47) using BLASTN. Mutant sequences were translated using the WISE2 (www.ebi.ac.uk/Wise2/) program. These translated sePNAS 兩 August 26, 2008 兩 vol. 105 兩 no. 34 兩 12409
GENETICS
silencing and the resulting histone H3 lysine 27 trimethylation strongly anti-correlates with Nipped-B and cohesin binding. Thus, we would not expect cohesin to bind hh in these cell lines even if it directly regulates hh. For example, although Nipped-B regulates Ubx expression in vivo (5, 7), Nipped-B and cohesin are excluded from the silenced Ubx and Abd-A genes in Sg4 cells, but bind to the transcribed Abd-B gene (10). In cells in which Abd-B is silenced, cohesin does not bind to the Abd-B promoter region. Thus, it remains possible that Nipped-B and cohesin directly stimulate hh transcription in vivo. Identification of loss of function zebrafish rad21 alleles in a genetic screen for mutations that reduce expression of runx genes also suggests that cohesin promotes gene expression, but again, it is unknown if this effect is direct (9). Stronger evidence supporting the idea that cohesin directly stimulates transcription arises from a recent study on axon pruning in the Drosophila mushroom body (38). In this study, loss of function alleles of the Smc1 and SA genes were isolated in a screen for mutations that block pruning. The lack of pruning correlated with reduced expression of the ecdysone receptor (EcR) gene, and could be partially rescued by ectopic EcR expression. Nipped-B and cohesin bind to the transcribed portion of the EcR gene in all three cell lines examined, including the ML-DmBG3 line derived from third instar central nervous system, suggesting that they directly facilitate EcR expression (10). The question remains as to whether the same cohesin complexes required for cohesion of sister chromatids also function in transcription regulation, or whether, analogous to trxG proteins, different cohesin subunits have different functions in transcription, presumably because they are members of different complexes. One might conclude the latter based upon the observation that reductions in Rad21, SA, or SMC1 all increase cut expression, whereas decreases in Nipped-B reduce cut expression (6, 39). These effects are likely direct because cohesin and Nipped-B bind to a 180 kb region that encompasses the entire upstream regulatory and transcribed regions of cut in ML-DmBG3 cells (10). The expression of RNAi transgenes encoding for SA and Rad21 decreases the severity of the cutK allele (6, 39), whereas Nipped-B mutations enhance the cutK phenotype, also suggesting that they have opposite effects at cut. Finally, in contrast to results with Nipped-B mutations, we observed no consistent effects on cut expression for vtd mutant heterozygotes (data not shown); moreover, we also report here that mutations in vtd and Nipped-B both suppress the phenotypes of Pc4 and Mrt, but mutations in Smc1 or pds5 did not. Similarly, Dorsett et al. (39) have reported that null alleles of the cohesion factors sans and deco have no effect on the expression of cut when a functional chromosomal copy is present. Based on all of the above evidence, one might therefore conclude that different cohesin components may act differentially, possibly because, like trxG proteins, they are members of different regulatory complexes. However, it is also possible that the same cohesin complex involved in chromatid cohesion also regulates transcription, if binding at different loci results in different, gene specific consequences. Thus, in cut, which is activated by a remote wing margin enhancer located ⬎80 kb upstream of the promoter, it has been proposed that cohesin could inhibit long range activation, and that Nipped-B facilitates activation by maintaining a dynamic cohesin binding equilibrium (39). In other genes, such as EcR or hh, cohesin might help maintain open chromatin to facilitate transcription by encircling a 10-nm fiber and preventing refolding to a higher order structure. As for the differences observed in the genetics of cohesin components, there are likewise other plausible explanations: differences in genetic background of mutant lines tested, differences in maternal expression/loading of required gene products in different heterozygous flies, or the possibility that the cutK or Pc alleles are less sensitive to changes in rad21/vtd, SMC1, or pds5 gene dosage than they are to the gene dosage of Nipped-B. Consistent with
quences were aligned to the Flybase annotated rad21 coding sequence using CLUSTALW (www.ebi.ac.uk/clustalw). CLUSTALW was also used to create a multiple sequence alignment of the Rad21 protein sequence across multiple species. Shading of conserved residues was done using BOXSHADE (www. ch.embnet.org/software/BOX㛭form.html).
Analysis of Chromatid Cohesion in Neuroblasts. Third instar larval neuroblast chromosomes were prepared using the methods described by Pimpinelli et al. (48). Fluorescence images were collected using an Olympus BX51 microscope equipped with an Optronics CCD camera and MagnaFire 2.0 software.
Drosophila Embryonic Cell Cycles. Approximately 150 embryos derived either from parents heterozygous for the vtd␥26 – 6 allele or from wild-type sibling controls were collected, fixed, and rehydrated as described in ref. 24. Rehydrated embryos were stained with 2.5 g/ml DAPI in PBT [1⫻ phosphatebuffered saline (45), 1% bovine serum albumin, 0.05% Triton X-100] for twenty minutes followed by three washes of 10 minutes each with PBT. Embryos were mounted in Vectashield mounting medium (Vector Laboratories) and images were collected using a Zeiss LSM 510 META confocal microscope and analyzed in ImageJ.
ACKNOWLEDGMENTS. We thank Nazanin Ghavam and Neahlanna McLeod for technical assistance, Alan Baxter for the use of laboratory equipment, and Oren Schuldiner for sharing data before publication. This work was supported in part by National Institutes of Health Grants R01 GM055683 and P01 HD052860 (to D.D.); Intramural Research Program of the National Institutes of Health National Institute of Child Health and Human Development (J.A.K.); Basil O’Connor Starter Scholar Research Award Grant 5-FY0319 and the March of Dimes Birth Defects Foundation (W.D.W.); and the Natural Sciences and Engineering Research Council, Canada (H.W.B. and B.M.H., and a PGSD award to M.S.).
1. Nasmyth K, Haering CH (2005) The structure and function of SMC and kleisin complexes. Annu Rev Biochem 74:595– 648. 2. Hirano T (2006) At the heart of the chromosome: SMC proteins in action. Nat Rev Mol Cell Biol 7:311–322. 3. Losada A (2007) Cohesin regulation: Fashionable ways to wear a ring. Chromosoma 116:321–329. 4. Warren W, Gorringe K (2006) A molecular model for sporadic human aneuploidy. Trends Genet 22:218 –224. 5. Rollins RA, Morcillo P, Dorsett D (1999) Nipped-B, a Drosophila homologue of chromosomal adherins, participates in activation by remote enhancers in the cut and Ultrabithorax genes. Genetics 152:577–593. 6. Rollins RA, Korom M, Aulner N, Martens A, Dorsett D (2004) Drosophila Nipped-B protein supports sister chromatid cohesion and opposes the stromalin/Scc3 cohesion factor to facilitate long-range activation of the cut gene. Mol Cell Biol 24:3100 –3111. 7. Gause M, et al. (2007) Functional links between Drosphila Nipped-B and cohesin in somatic and meiotic cells. Chromosoma 117:51– 66. 8. Dorsett D (2007) Roles of the sister chromatid cohesion apparatus in gene expression, development, and human syndromes. Chromosoma 116:1–13. 9. Horsfield JA, et al. (2007) Cohesin-dependent regulation of Runx genes. Development 134:2639 –2649. 10. Misulovin Z, et al. (2007) Association of cohesin and Nipped-B with transcriptionally active regions of the Drosphila melanogaster genome. Chromosoma 117:89 –102. 11. Kennison JA, Tamkun JW (1988) Dosage-dependent modifiers of Polycomb and antennapedia mutations in Drosophila. Proc Natl Acad Sci USA 85:8136 – 8140. 12. Kennison JA (1995) The Polycomb and trithorax group proteins of Drosphila: Transregulators of homeotic gene function. Annu Rev Genet 29:289 –303. 13. Breiling A, Sessa L, Orlando V (2007) Biology of Polycomb and trithorax group proteins. Int Rev Cytol 258:83–136. 14. Ringrose L, Paro R (2007) Polycomb/Trithorax response elements and epigenetic memory of cell identity. Development 134:223–232. 15. Schuettengruber B, Chourrout D, Vervoort M, Leblanc B, Cavalli G (2007) Genome regulation by Polycomb and trithorax proteins. Cell 128:735–745. 16. Petruk S, Sedkov Y, Brock HW, Mazo A (2007) A model for initiation of mosaic HOX gene expression patterns by non-coding RNAs in early embryos. RNA Biol 4:1– 6. 17. Maschat F, Serrano N, Randsholt NB, Ge´raud G (1998) engrailed and polyhomeotic interactions are required to maintain the A/P boundary of the Drosophila developing wing. Development 125:2771–2780. 18. Wang Y, McMahon AP, Allen BL (2007) Shifting paradigms in Hedgehog signaling. Curr Opin Cell Biol 19:159 –165. 19. Fitzpatrick KA, Sinclair DA, Schulze SR, Syrzycka M, Honda BM (2005) A genetic and molecular profile of third chromosome centric heterochromatin in Drosophila melanogaster. Genome 48:571–584. 20. Schulze S, et al. (2001) Essential genes in proximal 3L heterochromatin of Drosophila melanogaster. Mol Gen Genet 264:782–789. 21. Felsenfeld AL, Kennison JA (1995) Positional signaling by hedgehog in Drosophila imaginal disc development. Development 121:1–10. 22. Hatton LS, O’Hare K (1999) Deficiency mapping of genes in Drosophila using single embryo PCR. Technical Tips Online (http://www.bio.ic.ac.uk/research/ohare/ t01816.htm) T01816. 23. Warren WD, Lin E, Nheu TV, Hime GR, McKay MJ (2000) Drad21, a Drosophila rad21 homologue expressed in S-phase cells. Gene 250:77– 84. 24. O’Dor E, Beck SA, Brock HW (2006) Polycomb group mutants exhibit mitotic defects in syncytial cell cycles of Drosophila embryos. Dev Biol 290:312–322. 25. Uhlmann F, Lottspeich F, Nasmyth K (1999) Sister-chromatid separation at anaphase onset is promoted by cleavage of the cohesin subunit Scc1. Nature 400:37– 42. 26. Hauf S, Waizenegger IC, Peters JM (2001) Cohesin cleavage by separase required for anaphase and cytokinesis in human cells. Science 293:1320 –1323.
27. Waizenegger I, Gime´nez-Abia´n JF, Wernic D, Peters JM (2002) Regulation of human separase by securin binding and autocleavage. Curr Biol 12:1368 –1378. 28. Vass S, et al. (2003) Depletion of Drad21/Scc1 in Drosophila cells leads to instability of the cohesin complex and disruption of mitotic progression. Curr Biol 13:208 –218. 29. Guacci V, Koshland D, Strunnikov A (1997) A direct link between sister chromatid cohesion and chromosome condensation revealed through the analysis of MCD1 in S. cerevisiae. Cell 91:47–57. 30. Tatebayashi K, Kato J, Ikeda H (1998) Isolation of a Schizosaccharomyces pombe rad21ts mutant that is aberrant in chromosome segregation, microtubule function, DNA repair and sensitive to hydroxyurea: Possible involvement of Rad21 in ubiquitinmediated proteolysis. Genetics 148:49 –57. 31. Tomonaga T, et al. (2000) Characterization of fission yeast cohesin: Essential anaphase proteolysis of Rad21 phosphorylated in the S phase. Genes Dev 14:2757–2770. 32. Toyoda Y, Furuya K, Goshima G, Nagao K, Takahashi K, Yanagida M (2002) Requirement of chromatid cohesion proteins rad21/scc1 and mis4/scc2 for normal spindlekinetochore interaction in fission yeast. Curr Biol 12:347–358. 33. Hoque MT, Ishikawa F (2002) Cohesin defects lead to premature sister chromatid separation, kinetochore dysfunction, and spindle-assembly checkpoint activation. J Biol Chem 277:42306 – 42314. 34. Wysocka J, Myers MP, Laherty CD, Eisenman RN, Herr W (2003) Human Sin3 deacetylase and trithorax-related Set1/Ash2 histone H3–K4 methyltransferase are tethered together selectively by the cell-proliferation factor HCF-1. Genes Dev 17:896 –911. 35. Beltran S, Angulo M, Pignatelli M, Serras F, Corominas M (2007) Functional dissection of the ash2 and ash1 transcriptomes provides insights into the transcriptional basis of wing phenotypes and reveals conserved protein interactions. Genome Biol 8:R67. 36. Mazo A, Hodgson JW, Petruk S, Sedkov Y, Brock HW (2007) Transcriptional interference: An unexpected layer of complexity in gene regulation. J Cell Sci 120:2755– 2761. 37. Schwartz YB, et al. (2006) Genome-wide analysis of Polycomb targets in Drosophila melanogaster. Nat Genet 38:700 –705. 38. Schuldiner O, et al. (2008) piggyBac-based mosaic screen identifies a postmitotic function for cohesin in regulating developmental axon pruning. Dev Cell 14:227– 238. 39. Dorsett D, Eissenberg JC, Misulovin Z, Martens A, Redding B, McKim K (2005) Effects of sister chromatid cohesion proteins on cut gene expression during wing development in Drosophila. Development 132:4743– 4753. 40. Stedman W, Kang H, Lin S, Kissil JL, Bartolomei MS, Lieberman PM (2008) Cohesins localize with CTCF at the KSHV latency control region and at cellular c-myc and H19/Igf2 insulators. EMBO J 27:654 – 666. 41. Wendt KS, et al. (2008) Cohesin mediates transcriptional insulation by CCCTC-binding factor. Nature 451:777–778. 42. Parelho V, et al. (2008) Cohesins Functionally Associate with CTCF on Mammalian Chromosome Arms. Cell 132:422– 433. 43. Pauli A, et al. (2008) Cell-type-specific TEV protease cleavage reveals cohesin functions in Drosophila neurons. Dev Cell 14:239 –251. 44. Rorth P (1998) GAL4 in the Drosophila female germline. Mech Dev 78:113–118. 45. Sambrook J, Russell DW (2001) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Press, New York). 46. Rubin J, Spradling A (1982) Genetic transformation of Drosophila with transposable element vectors. Science 218:348 –353. 47. Smith CD, Shu S, Mungall CJ, Karpen GH (2007) The Release 5.1 annotation of Drosophila melanogaster heterochromatin. Science 316:1586 –1591. 48. Pimpinelli, S, Bonaccorsi S, Fanti L, Gatti M (2000) in Drosophila Protocols, eds Sullivan W, Ashburner M, Hawley RS(Cold Spring Harbor Laboratory Press, New York), pp 3–23.
12410 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0801698105
Hallson et al.
