Deletion of the Cul1 gene in mice causes arrest in early embryogenesis and accumulation of cyclin E

Brief Communication

1191

Deletion of the Cul1 gene in mice causes arrest in early embryogenesis and accumulation of cyclin E Yisong Wang*, Sonya Penfold*, Xiaojing Tang*, Naka Hattori†, Paul Riley†, J. Wade Harper‡, James C. Cross† and Mike Tyers*

Correspondence: Mike Tyers E-mail: [email protected] Received: 3 September 1999 Revised: 29 September 1999 Accepted: 29 September 1999 Published: 11 October 1999 Current Biology 1999, 9:1191–1194 0960-9822/99/$ – see front matter © 1999 Elsevier Science Ltd. All rights reserved.

Results and discussion A targeting vector was designed to replace exon 2 of the murine Cul1 gene (Figure 1a), and thereby delete the translation initiation codon and an essential Skp1-binding site. Cul1+/– embryonic stem (ES) cell clones did not express any detectable truncated Cul1 protein products, indicating that the targeted locus is probably a null allele (Figure 1b,c). Chimeric animals were generated from the targeted ES cells to introduce the Cul1 mutation into the mouse

Analysis of progeny derived from crosses of Cul1 heterozygotes revealed that homozygous mutants did not survive

E2 a b

E3

d

ES clones

BamHl Spel

EcoRl

neo

1 kb

A (c)

Mutant

L1

ES clones

+/ +/+ +/– hC– U

Cul1

B

r

Spel

Stul

c

Targeting vector

hsv-tk

neo

Wild type

Smal

LINE-1

(b)

Smal

ATG

I*

EcoRl

Stul

Smal Spel

(a)

BamHl Spel

Figure 1

kDa 97 66

3.2 kb Cul1

45 (d)

+/+

+/–

Cul1 (e) 5 kb 4 kb

31 E6.5 embryos

–/ – +/ +/– +

Addresses: *Programme in Molecular Biology and Cancer, and †Programme in Development and Fetal Health, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, 600 University Avenue, Toronto M5G 1X5, Canada. ‡Verna and Marrs McLean Department of Biochemistry, Baylor College of Medicine, Houston, Texas 77030, USA.

germline (Figure 1d). We have not observed any abnormal phenotype in Cul1+/– animals at up to 15 months of age.

+/ – +/ +/– +/+ M– ar ke

The stability of many proteins is controlled by the ubiquitin proteolytic system, which recognizes specific substrates through the action of E3 ubiquitin ligases [1]. The SCFs are a recently described class of ubiquitin ligase that target a number of cell cycle regulators and other proteins for degradation in both yeast and mammalian cells [2–6]. Each SCF complex is composed of the core protein subunits Skp1, Rbx1 and Cul1 (known as Cdc53 in yeast), and substrate-specific adaptor subunits called F-box proteins [2–4]. To understand the physiological role of SCF complexes in mammalian cells, we generated mice carrying a deletion in the Cul1 gene. Cul1–/– embryos arrested around embryonic day 6.5 (E6.5) before the onset of gastrulation. In all cells of the mutant embryos, cyclin E protein, but not mRNA, was highly elevated. Outgrowths of Cul1–/– blastocysts had limited proliferative capacity in vitro and accumulated cyclin E in all cells. Within Cul1–/– blastocyst cultures, trophoblast giant cells continued to endocycle despite the elevated cyclin E levels. These results suggest that cyclin E abundance is controlled by SCF activity, possibly through SCFdependent degradation of cyclin E.

Cul1 1.3 kb 0.9 kb

Current Biology

Characterization and targeted disruption of the Cul1 locus. (a) Structure of the mouse Cul1 locus, targeting vector and the disrupted mutant allele following homologous recombination. A 3′ flanking probe (A), a 5′ internal probe (B), and PCR primers (a–d) were used to distinguish mutant from wild-type alleles. The targeting vector removes 3 kb of genomic sequence that encodes a region including the initiation codon (ATG) and the Skp1-binding site (asterisk) in exon 2 (E2). (b) DNA from ES cell clones was amplified by PCR using primers c and d, which yielded a fragment of 3.2 kb corresponding to the mutant allele. (c) Absence of truncated Cul1 protein products in Cul1+/– ES cells. The indicated lysates were immunoblotted with an anti-CUL1 antibody directed against the carboxyl terminus of human CUL1. (d) Germline transmission of the disrupted Cul1 allele. Tail DNA from a representative litter of a Cul1+/– intercross was digested with SpeI and hybridized with the 3′ flanking probe (probe A). The 4 and 5 kb fragments are derived from the mutant and wild-type alleles, respectively. (e) Genotype analysis of E6.5 embryos from a representative Cul1+/– intercross. DNA was amplified with primers a, b and d, and yields a 0.9 kb wild-type Cul1 product and 1.3 kb mutant Cul1 product. The Cul1 mutant allele segregated with the expected Mendelian ratio (see the Supplementary material).

1192

Current Biology Vol 9 No 20

Figure 2 (a)

E6.5 +/+

(b) +/+ EXE

–/– EPC

EPC EXE

EE

EE

VE

VE E7.5 +/– EPC

EXE

EPC EXE

EE

–/–

EPC

EE –/–

(c) +/–

Comparison of wild-type and Cul1 mutant embryos. (a) Whole mounts of embryos of the indicated genotypes at E6.5 and E7.5. (b) Sagittal sections of wild-type and Cul1–/– embryos at E6.5 were stained with hematoxylin and eosin. (c) Loss of Cul1 does not initiate widespread apoptosis. Nuclei with detectable free DNA termini (arrowheads) were visualized in E6.5 embryos by terminaldeoxynucleotidyl-mediated dUTP nick-end labeling (TUNEL) staining. EPC, ectoplacental cone; EXE, extraembryonic ectoderm; EE, embryonic ectoderm; VE, visceral endoderm.

–/–

EXE

EE Current Biology

beyond E7.5 (Figure 1e; see the Supplementary material). Although Cul1–/– embryos were comparable in size to wild-type embryos at E6.5, the mutant embryos stopped developing at this point and were much smaller than wildtype or heterozygous littermates by E7.5 (Figure 2a). Histological examination showed that mutant embryos arrested before gastrulation with near normal morphology, except that the embryonic ectoderm was thin with many condensed nuclei (Figure 2b). The developmental arrest of Cul1–/– embryos was accompanied by only a modest increase in the number of apoptotic cells (Figure 2c). As Cul1 is essential for early embryonic development, it is not redundant with any of the nine other mammalian cullin homologs that can be identified in current sequence databases (A. Willems, personal communication). To identify targets of Cul1-based SCF complexes, we examined candidate substrates by immunohistochemical staining of embryos. In wild-type and heterozygous embryos, cyclin E was detected in a fraction of cells (Figure 3a), consistent with its G1–S-specific cell-cycle expression [7]. In marked contrast, intense cyclin E staining was detected in virtually every cell of Cul1–/– embryos (Figure 3a). Importantly, cyclin E mRNA expression was not altered in Cul1–/– embryos (Figure 3b), indicating that cyclin E did not accumulate as a result of deregulated transcription. We examined directly the role of Cul1 in cell proliferation by following the development in vitro of E2.5 or E3.5 embryos derived from Cul1+/– intercrosses. Regardless of genotype, most embryos hatched and formed blastocyst

outgrowths. After 4 days in culture, all outgrowths had an expanded inner cell mass (ICM), some endodermal outgrowth and trophoblast giant cells (Figure 4a–d). In contrast to wild-type (n = 12) and heterozygous cultures (n = 23), cells within Cul1–/– outgrowths (n = 10) failed to proliferate beyond 7 days in culture. Attempts to derive either Cul1–/– ES or trophoblast stem cell lines were unsuccessful, consistent with an essential role for Cul1 in proliferation of early embryonic cell types. Cyclin E expression in wild-type blastocyst outgrowths was heterogeneous, and easily detectable in approximately one third of trophoblast cells (Figure 4e), as expected for an asynchronous population of cells in which cyclin E normally oscillates during the endocycle [8]. In contrast, outgrowths of Cul1–/– blastocysts contained abnormally large trophoblast giant cells that had uniformly strong cyclin E immunoreactivity (Figure 4f). To determine whether Cul1–/– trophoblast giant cells undergo repeated endocycles despite the elevated cyclin E levels, blastocyst outgrowths were pulse-labeled with bromodeoxyuridine (BrdU). Both Cul1+/– and Cul1–/– outgrowths had a similar pattern of BrdU incorporation in which only a fraction of trophoblast giant cells were BrdU-positive, indicating that repeated rounds of DNA replication still occurred in Cul1–/– trophoblast giant cells (Figure 4g–j). Relative fluorescence intensity of nuclei after DNA staining with bisbenzimide was used to estimate ploidy levels. Compared to nuclei of parietal endoderm cells, trophoblast giant cells in wild-type and mutant outgrowths had DNA contents that were on average 10-fold and 18-fold higher with average nuclear diameters

Brief Communication

of 32 µm (n = 776) and 51 µm (n = 61), respectively. These data indicate that Cul1–/– trophoblast giant cells may actually undergo additional rounds of re-replication.

1193

Figure 3

(a)

+/–

–/–

Cul1–/–

The reason for the developmental arrest of embryos is unknown. As cells can tolerate high levels of cyclin E expression [7,9], it seems unlikely that accumulation of cyclin E should cause arrest at E6.5. In view of the emerging roles of SCF complexes in many cellular pathways, numerous substrates are likely to accumulate in Cul1–/– embryos, one or more of which may account for the developmental arrest. Known substrates of SCF complexes in mammals include the signaling protein IκBα and the transcription factor β-catenin, which are targeted by the F-box protein β-TrCP [5]; and the transcription factor E2F-1 and the cyclin-dependent kinase (Cdk) inhibitor p27Kip1, which are targeted by the F-box protein SKP2 [6]. Although E2F-1 contributes to cyclin E expression [6], E2F-1 is not expressed in early embryogenesis [10], perhaps explaining why we did not observe accumulation of cyclin E mRNA in Cul1–/– embryos. As in mice, cul-1 mutants in the nematode Caenorhabditis elegans fail to undergo embryogenesis, provided that maternal mRNA supplies are depleted [11]. Unlike the mouse, maternal contributions of cul-1 suffice for complete development of cul-1 mutants into sterile adults which exhibit the remarkable propensity to undergo extra rounds of cell division in all tissues [11]. The idea that the additional divisions in cul-1 mutant nematodes might be due to inappropriate G1 cyclin activity is consistent with the accumulation of cyclin E in Cul1–/– cells in mice. Although ectopic expression of cyclin E causes premature S phase in Drosophila, continuous overexpression of cyclin E in the fly salivary gland blocks the normal endoreduplication cycle, suggesting that periodic cyclin E activity is required to trigger the repeated rounds of DNA replication [12–14]. In mice, however, the accumulation of cyclin E in Cul1–/– trophoblast giant cells did not appear to interfere with the endocycle. Intriguingly, mice that lack Cul3, another cullin gene family member, also arrest early in development with elevated levels of cyclin E in some cell types, including trophoblast cells [15]. Nevertheless, in contrast to Cul1–/– mutants, Cul3–/– trophoblasts fail to endocycle [15]. Thus, additional mechanisms may allow periodic activation of cyclin E–Cdk2 in Cul1–/– cells but not in Cul3–/– cells. The most parsimonious explanation for the accumulation of cyclin E protein in Cul1–/– cells is that Cul1 mediates cyclin E degradation, although it is conceivable that the effects of Cul1 on cyclin E are indirect. Proof that a Cul1based complex is the ubiquitin ligase for cyclin E awaits direct demonstration that cyclin E is stabilized in Cul1–/– cells, identification of the putative F-box protein for cyclin E and reconstitution of ubiquitination activity

(b)

+/–

EPC

–/–

EPC EXE

EXE

EE EE

Current Biology

Cyclin E protein is elevated in Cul1–/– embryos in the absence of altered levels of cyclin E mRNA. (a) Sagittal sections of Cul1+/– and Cul1–/– embryos at E6.5 stained with anti-cyclin E antibody (see the Supplementary material). Arrows indicate several representative trophoblast giant cells. The original magnification was 100×. (b) In situ hybridization of sagittal sections of Cul1+/– and Cul1–/– embryos at E6.5 with an anti-sense cyclin E probe, visualized by dark field microscopy. The original magnification was 200×.

towards cyclin E in vitro. Elimination of the yeast G1 cyclin Cln2 and mammalian cyclin E by the ubiquitin system proceeds through Cdk-dependent and -independent pathways [16–18]. As Cul3 appears to target Cdk2free forms of cyclin E [15], Cul1-based complexes may account for the degradation of Cdk2-bound, and hence presumably phosphorylated, cyclin E. Consistent with this possibility, Cul1–cyclin E complexes can be detected only in the presence of active Cdk2 in heterologous expression systems (data not shown). As cyclin E is a crucial rate-limiting activator of DNA replication and centrosome duplication [7,19,20], both Cul1- and Cul3-dependent degradation pathways may be necessary to appropriately restrain its activity. Indeed, cyclin E is abnormally regulated in many cancer cells [21] and recent evidence suggests that excessive cyclin E activity leads to chromosome instability [9].

1194

Current Biology Vol 9 No 20

Figure 4 Growth and development of blastocysts in vitro. (a–d) Cul1–/– blastocysts cease growth but still undergo differentiation. PE, parietal endoderm; VE, visceral endoderm; TG, trophoblast giant cells. (e,f) Cyclin E is elevated in Cul1–/– blastocyst outgrowths. After 11 days in culture, approximately onethird of identifiable trophoblast cells in Cul1+/+ and Cul1+/– outgrowths were positive for cyclin E, whereas all TG cells in Cul1–/– outgrowths contained detectable cyclin E. Representative trophoblasts scored as positive (black arrowheads) or negative (white) for cyclin E are shown. (g–j) Cul1–/– TG cells remain in the endocycle. Outgrowths were labeled with BrdU after 11 days in culture, and stained by anti-BrdU antibody and bisbenzimide. Green arrowheads, BrdU+ cells; white, BrdU– cells.

Day 4 (a)

Day 11 (c)

VE

Cyclin E PE

BrdU

(g)

(i)

(h)

(j)

ICM

ICM

+/–

DNA

(e)

TG (b)

(d)

(f)

TG –/–

ICM ICM PE

VE Current Biology

Supplementary material A table showing genotypic and phenotypic analysis of Cul1+/– intercrosses, a figure showing Cul1 tissue distribution and cyclin E antibody specificity, and further methodological detail are available at http://current-biology.com/supmat/supmatin.htm.

Acknowledgements We thank R. Johnson for kindly communicating unpublished results; S. Kulkarni, A. Elia, R. Sakai, A. Willems, E. Patton, T. Kunath, V. Lai and J. Yankulov for superb technical assistance and advice; Y. Xiong, P. Hamel, S. Coats, J. Roberts, R. Deshaies, E. Kipreos and T. Stearns for reagents; and J. Lees, S. Reed, J. Roberts, S. Elledge, T. Pawson, L. Harrington and A. Breitkreutz for helpful discussions. Y.W. was supported by a Medical Research Council (MRC) of Canada Fellowship. J.W.H. was supported by NIH grants GM54137 and AG11085 and the Welch Foundation. J.C.C. and M.T. are supported by grants from the MRC and the National Cancer Institute of Canada.

References 1. Hershko A, Ciechanover A: The ubiquitin system. Annu Rev Biochem 1998, 67:425-479. 2. Bai C, Sen P, Hofmann K, Ma L, Goebl M, Harper JW, Elledge SJ: SKP1 connects cell cycle regulators to the ubiquitin proteolysis machinery through a novel motif, the F-box. Cell 1996, 86:263-274. 3. Patton EE, Willems AR, Tyers M: Combinatorial control in ubiquitindependent proteolysis: don’t Skp the F-box hypothesis. Trends Genet 1998, 14:236-243. 4. Koepp DM, Harper JW, Elledge SJ: How the cyclin became a cyclin: regulated proteolysis in the cell cycle. Cell 1999, 97:431-434. 5. Laney JD, Hochstrasser M: Substrate targeting in the ubiquitin system. Cell 1999, 97:427-430. 6. Amati B, Vlach J: Kip1 meets SKP2: new links in cell-cycle control. Nat Cell Biol 1999, 1:E91-E93. 7. Sherr CJ: G1 phase progression: cycling on cue. Cell 1994, 79:551-555. 8. MacAuley A, Cross JC, Werb Z: Reprogramming the cell cycle for endoreduplication in rodent trophoblast cells. Mol Biol Cell 1998, 9:795-807. 9. Spruck CH, Won K-I, Reed SI: Deregulated cyclin E induces chromosome instability. Nature 1999, 401:297-301. 10. Tevosian SG, Paulson KE, Bronson R, Yee AS: Expression of the E2F-1/DP-1 transcription factor in murine development. Cell Growth Differ 1996, 7:43-52. 11. Kipreos ET, Lander LE, Wing JP, He WW, Hedgecock EM: cul-1 is required for cell cycle exit in C. elegans and identifies a novel gene family. Cell 1996, 85:829-839.

12. Edgar BA, Lehner CF: Developmental control of cell cycle regulators: a fly’s perspective. Science 1996, 274:1646-1652. 13. Weiss A, Herzig A, Jacobs H, Lehner CF: Continuous cyclin E expression inhibits progression through endoreduplication cycles in Drosophila. Curr Biol 1998, 8:239-242. 14. Follette PJ, Duronio RJ, O’Farrell PH: Fluctuations in cyclin E levels are required for multiple rounds of endocycle S phase in Drosophila. Curr Biol 1998, 8:235-238. 15. Singer JD, Gurian-West M, Clurman B, Roberts JM: Cullin-3 targets cyclin E for ubiquitination and controls S phase in mammalian cells. Genes Dev 1999, 13:2375-2387. 16. Lanker S, Valdivieso MH, Wittenberg C: Rapid degradation of the G1 cyclin Cln2 induced by CDK-dependent phosphorylation. Science 1996, 271:1597-1601. 17. Won KA, Reed SI: Activation of cyclin E/CDK2 is coupled to sitespecific autophosphorylation and ubiquitin-dependent degradation of cyclin E. EMBO J 1996, 15:4182-4193. 18. Clurman BE, Sheaff RJ, Thress K, Groudine M, Roberts JM: Turnover of cyclin E by the ubiquitin-proteasome pathway is regulated by Cdk2 binding and cyclin phosphorylation. Genes Dev 1996, 10:1979-1990. 19. Leatherwood J: Emerging mechanisms of eukaryotic DNA replication initiation. Curr Opin Cell Biol 1998, 10:742-748. 20. Winey M: Cell cycle: Driving the centrosome cycle. Curr Biol 1999, 9:R449-R452. 21. Keyomarsi K, Herliczek TW: The role of cyclin E in cell proliferation, development and cancer. Prog Cell Cycle Res 1997, 3:171-191.

S1

Supplementary material Deletion of the Cul1 gene in mice causes arrest in early embryogenesis and accumulation of cyclin E Yisong Wang, Sonya Penfold, Xiaojing Tang, Naka Hattori, Paul Riley, J. Wade Harper, James C. Cross and Mike Tyers Current Biology 21 October 1999, 9:1191–1194 Figure S1

Supplementary materials and methods Targeted disruption of Cul1 in mice

Phenotypic analysis Deciduae were collected from uteri of plugged females at E5.5, E6.5, E7.5 and E9.5. Samples were fixed for 4–12 h in 4% paraformaldehyde at 4oC, dehydrated, and embedded in paraffin. Sections 7 µm thick were cut by microtome and processed for hematoxylin and eosin staining and for immunohistochemical analysis. Embryo sections and

(a)

In te Ki stin d Sp n e y e ESleen bv ce lls Th hCU ym L Ad u 1 s M reno LN g la O nd v H ary ea Ly rt m M ph us no Li cle de v Lu er ng

Three overlapping phage clones from a 129/Sv mouse genomic library were isolated using a 0.8 kb EcoRI fragment of the 5′ human CUL1 cDNA [S1]. A full-length mouse Cul1 cDNA is represented in the expressed sequence tag (EST) databases (accession numbers AA014775, AA549095, AI527547, AF083216) and is predicted to encode a protein that is 96% identical to human CUL1. The structure of approximately 30 kb of the mouse Cul1 gene was characterized by restriction-enzyme mapping and sequencing. The initiation codon was identified in exon 2, which was about 30 kb downstream of exon 1. A 3.1 kb StuI–SmaI genomic fragment upstream of exon 2 and a 1.5 kb EcoRI fragment downstream of exon 2 were cloned into pPNT to create a targeting vector, which was designed to avoid a LINE element that is approximately 4 kb proximal to exon 2 (Figure 1a). Mouse R1 ES cells were electroporated and selected using standard procedures [S2]. Homologous recombinants were identified by PCR using primers specific for sequences in the Cul1 gene and the neo cassette (upper primer c, 5′-GGCTGGAGGAGGAATTGAGATGAACACAC-3′; lower primer d, 5′-CAACGCTATGTCCTGATAGCGGTCC-3′) and confirmed by Southern blotting using a 3′ flanking probe (probe A) and an internal probe (probe B; Figure 1a). Chimeric mice were generated by aggregation of three correct ES cell clones with CD-1 morulae [S3], one of which yielded germline transmission. Heterozygotes that gave germline transmission of the targeted allele were bred into both 129/Sv and 129/Sv × CD1 backgrounds. In every experiment involving embryos, blastocysts and tissue sections, the status of the Cul1 locus was determined by PCR using a three-primer combination (Figure 1a): upper primer a, 5′-CAGAGGCCATGTCAGGAAAGTAA-3′; lower primer b, 5′-ACTTTGCCATGCTTTGTCT-3′; lower primer d (see above). PCR products were visualized by hybridization with internal probes specific for wild-type and mutant alleles.

CUL1 α-Tubulin (b)

1

2

kDa 97

3

66 Cyclin E 45

31

Current Biology

(a) Ubiquitous expression of Cul1 protein in different tissues. Total cell lysates from adult mouse tissues, and insect-cell-expressed (bv) human CUL1, were immunoblotted with a human anti-CUL1 antibody [S8] and, as a loading control, with anti-α-tubulin antibody. (b) Specificity of the anti-cyclin E antibody. Total cell lysates were separated on a 10% gel by SDS–PAGE and blotted with anti-cyclin E antibody (M20, Santa Cruz). Lane 1, Cul1+/– ES cells; lane 2, Cul1+/+ ES cells; lane 3, human K562 cells. Note that the antibody does not cross-react with human cyclin E.

blastocyst outgrowths were stained by standard methods with an anticyclin E antibody (M20, Santa Cruz; [S4–S6]) at 1:100 dilution and

Table S1 Genotypic and phenotypic analysis of Cul1+/– intercross progeny. Total

Cul1+/+

Cul1+/–

Cul1–/–

Abnormal†

Unknown‡

E6.5

14(30)*

4(6)

5(16)

3(8)

0(8)

2

E7.5

32

8

13

10

10

1

E9.5

12

4

5

0

0

3

Newborn

54

20

34

0

0

–

Neonates and embryos were genotyped by PCR and confirmed by Southern analysis as described in the Supplementary materials and methods. *Numbers in parentheses were derived from genotypes of E6.5 embryonic sections. †Abnormal phenotypes in whole-mount assay

were characterized by reduced size and poor embryonic organization, and, in the case of embryonic sections, abnormal phenotypes were defined by the abundance of cyclin E staining and the presence of an increased number of apoptotic cells. ‡Empty deciduae.

S2

Supplementary material

detected with a DAB substrate kit (Vector Laboratories). For in situ hybridization, embryos were isolated in diethylpyrocarbonate (DEPC)treated PBS at E6.5 and processed as for immunohistochemical analysis. Hybridization was performed using [α-33P]UTP-labeled cyclin E anti-sense and sense probes. Apoptotic nuclei were detected with a TUNEL assay kit (Boehringer Mannheim). For blastocyst outgrowth experiments, superovulated females were mated and morulae collected at E2.5 or E3.5. After cultivation in KSOM medium for 48 h, blastocysts were individually cultured in 10 µl drops of ES cell media without leukemia inhibitory factor overlayed with paraffin oil in 5% CO2 at 37°C. For BrdU staining, outgrowths were labeled with 10 µM BrdU for 1 h after 11 days in culture. After fixation with 4% paraformaldehyde, outgrowths were treated with 2 M HCl, followed by staining with anti-BrdU antibody (Zymed). At the end of most outgrowth experiments, blastocysts were genotyped by PCR in combination with Southern blot analysis.

Immunoblot analysis Cell lysates were prepared in 50 mM Tris at pH 7.5, 150 mM NaCl, 5 mM NaF, 5 mM EDTA, 0.1% NP-40, 10% glycerol, 15 mM β-glycerolphosphate, 1 mM PMSF, 1 µg/ml leupeptin, 1 µg/ml pepstatin, 10 µg/ml TPCK and 10 µg/ml soybean trypsin inhibitor. Lysates were centrifuged at 10,000 × g for 15 min twice at 4°C. Immunoblots were performed as described previously [S7]. Anti-Cyclin E (M20, Santa Cruz) and anti α-tubulin antibodies (MAS 078, Harlan SERA-LAB) were used at 1:3000 and 1:1000, respectively. An antibody directed against a conserved carboxy-terminal peptide in the human CUL1 sequences was used as described previously [S8]. Specificity of CUL1 immunoreactivity was confirmed with two other independent anti-CUL1 antisera (data not shown).

Supplementary references S1. Kipreos ET, Lander LE, Wing JP, He WW, Hedgecock EM: cul-1 is required for cell cycle exit in C. elegans and identifies a novel gene family. Cell 1996, 85:829-839. S2. Wurst W, Joyner A: Production of targeted embryonic stem cell clones. In Gene Targeting: A Practical Approach. Edited by Joyner A. Oxford: Oxford University Press; 1993: 33-61. S3. Wood SA, Allen ND Jr, Nagy A: Non-injection methods for the production of embryonic stem cell-embryo chimaeras. Nature 1993, 365:87-90. S4. ElShamy WM, Fridvall LK, Ernfors P: Growth arrest failure, G1 restriction point override, and S phase death of sensory precursor cells in the absence of neurotrophin-3. Neuron 1998, 21:1003-1015. S5. We GL, Krasinski K, Kearney M, Isner JM, Walsh K, Andres V: Temporally and spatially coordinated expression of cell cycle regulatory factors after angioplasty. Circ Res 1997, 80:418-426. S6. Wang QS, Sabourin CLK, Wang H, Stoner GD: Overexpression of cyclin D1 and cyclin E in N-nitrosomethylbenzylamine-induced rat esophageal tumorigenesis. Carcinogenesis 1996, 17:1583-1588. S7. Willems AR, Lanker S, Patton EE, Craig KL, Nason TF, Mathias N, et al.: Cdc53 targets phosphorylated G1 cyclins for degradation by the ubiquitin proteolytic pathway. Cell 1996, 86:453-463. S8. Michel JJ, Xiong Y: Human CUL-1, but not other cullin family members, selectively interacts with SKP1 to form a complex with SKP2 and cyclin A. Cell Growth Differ 1998, 9:435-449.