[Cell Cycle 3:12, 1638-1644; December 2004]; ©2004 Landes Bioscience
Ufd2, a Novel Autoantigen in Scleroderma, Regulates Sister Chromatid Separation Report
ABSTRACT
UT E
SC
mitosis, ubiquitylation, securin, autoantigen, Ufd2
BIO
ABBREVIATIONS
05
LA
ND
ES
ubiquitin fusion degradation-2 ubiquitin-activating enzyme ubiquitin-conjugating enzyme ubiquitin-protein ligase in vitro transcription/translation protein kinase A human salivary gland N-ethylmaleimide propidium iodide anaphase promoting complex fluorescein-5-isothiocyanate
20
ACKNOWLEDGEMENTS
©
See page 1643.
NOTE
Supplementary material accompanies this manuscript. Figure 1S can be found at: http://www.landesbioscience.com/journals/cc/ spinetteCC3-12-sup.pdf.
1638
Human autoantibodies have proven to be powerful tools for revealing the molecular mechanisms that regulate various important cellular processes.1-3 For example, several scleroderma autoantibodies have provided tools to define structural, functional and regulatory components in mitosis. Little was known about the composition and function of the mitotic centromere until 1980, when it was discovered that scleroderma patient sera recognized centromere proteins CENP-A, CENP-B, and CENP-C.4 Human autoantibodies recognizing these centromere proteins enabled studies to be performed that provided information about centromere structure and function in mitosis.5-8 Scleroderma patient autoantibodies also facilitated the discovery and functional characterization of the centrosome protein, pericentrin, which is required for the establishment of organized microtubule arrays that are essential to mitosis and meiosis.9,10 Interestingly, several other scleroderma autoantigens described to date are associated with and/or regulate critical mitotic machinery. One of the most common autoantigens, topoisomerase I, is localized along the chromatids throughout mitosis and is essential for chromosome condensation.11-13 Since knowledge about the components that regulate mitosis in human cells is still fragmented, characterization of novel scleroderma autoantigens might identify additional pathways that play important roles in mitotic regulation. In this manuscript, we have identified Ufd2 as a scleroderma autoantigen, and examine the cell biology and mitotic function of this protein. Ufd2p was originally identified in a Saccharomyces cerevisiae screen for mutants in the UFD (ubiquitin fusion protein degradation) pathway. In the original assay, N-terminally fused ubiquitin-βgal was rapidly targeted for degradation in wild type yeast but was stabilized in ufd2∆ mutants.14 In subsequent studies, Ufd2p was found to have an “E4” function, catalyzing the formation of long multiubiquitylated chains on preformed ubiquitin conjugates generated by E1, E2 and E3 enzymes. These multiubiquitylated chains were essential for degradation in vitro.15 A human orthologue of Ufd2 shares significant homology (40–57% identity) with the yeast protein, primarily within a 150–200 amino acid C-terminal region, designated the “U-box’’. The U-box is a variation of the RING domain found in many E3 ligases,16 and is believed to be required for ubiquitylation and protein binding.15 While the human Ufd2 protein can ubiquitylate itself and bacterial proteins in vitro,17 its substrates and function in vivo still remain unknown. We
IEN
KEY WORDS
INTRODUCTION
CE
Previously published online as a Cell Cycle E-publication: http://www.landesbioscience.com/journals/cc/abstract.php?id=1345
IST
RIB
Received 11/02/04; Accepted 11/04/04
OT D
*Correspondence to: Antony Rosen; Department of Medicine/Rheumatology; Johns Hopkins University School of Medicine; Mason F. Lord Bldg.; 5200 Eastern Avenue, Suite 4100; Baltimore, Maryland 21224 USA; Tel.: 410.550.1900; Fax: 410.550.2072; Email
[email protected]
ON
6Molecular Biology Program; Memorial Sloan-Kettering Cancer Center; New York, New York USA
.D
Departments of 1Medicine (Division of Rheumatology); 2Dermatology; 3Cell Biology and Anatomy; 4Pathology; 5The Sidney Kimmel Comprehensive Cancer Center; Johns Hopkins University; Baltimore, Maryland USA
Ufd2 E1 E2 E3 IVTT PKA HSG NEM PI APC FITC
The characterization of molecules recognized by patients with autoimmune diseases has provided significant insights into important biological pathways. In these studies, we define Ufd2 (a novel E4 polyubiquitylating enzyme) as a new autoantigen in scleroderma, and show that it regulates chromosome condensation and separation during mitosis in human cells. Ufd2 is regulated by phosphorylation in mitosis. Inhibition of Ufd2 expression results in mitotic arrest at the metaphase-anaphase transition, where cells manifest abnormal chromosome morphology, missegregated chromosomes, irregular spindles, and premature separation of sister chromatids. This is accompanied by premature separase activation, and accumulation of securin in a novel modified form. We further demonstrate that Ufd2 directly and efficiently ubiquitylates securin in vitro and is required for securin polyubiquitylation in vivo. This is the first description of a physiologic substrate for Ufd2, establishing this E4 enzyme as an important regulator of chromosome condensation and separation during mitosis in human cells. Its targeting in scleroderma, together with many other components of the mitotic machinery, reinforces the concept that mitotic cells may be an important focus of the autoimmune response in this disease.
.
Sarah Spinette1 Christoph Lengauer5 James A. Mahoney1 Prasad V. Jallepalli6 Zhenghe Wang5 Livia Casciola-Rosen1,2 Antony Rosen1,3,4,*
Cell Cycle
2004; Vol. 3 Issue 12
UFD2 REGULATES SISTER CHROMATID SEPARATION
demonstrate here that Ufd2 is a scleroderma autoantigen which is dramatically regulated by phosphorylation during mitosis on a unique N-terminal domain present in the mammalian forms of the molecule, but not the yeast orthologue. Knockdown of Ufd2 generates a striking mitotic phenotype characterized by abnormal chromosome morphology, missegregated chromosomes, irregular spindles, and premature sister chromatid separation. We also show that securin is polyubiquitylated by Ufd2 in vitro and in vivo, thus delineating a novel Ufd2-dependent ubiquitylation pathway which regulates chromosome condensation and separation in human cells.
MATERIALS AND METHODS
Cell Lines and Nocodazole Treatment. HeLa and K562 cell lines were obtained from American Type Culture Collection (Rockville, MD). Human salivary gland (HSG) epithelial cells were received from Dr. Bruce Baum (NIDCR) from an original stock provided by Dr. Mitsunabo Sato (Tokushima University, Shikoku, Japan). HeLa and HSG cells were cultured in DMEM supplemented with 2 mM glutamine and 10% heat inactivated FBS. K562 cells were cultured in RPMI supplemented with 2 mM glutamine and 10% FBS. To arrest cells in prometaphase, 0.2 µg/ml nocodazole (Sigma Aldrich) was added to the medium of sub-confluent HeLa and HSG cells (plated 10–24 hrs previously) and incubated for 12 hours. Mitotic HSG and HeLa cells were obtained by mitotic wash off into PBS and centrifugation. K562 cells in log-phase growth were treated with nocodazole at 0.4 µg/ml for 15 hours. Patient Sera. Sera from 55 randomly selected patients with scleroderma (including patients with limited and diffuse disease) were obtained from a serum bank derived from patients evaluated at the Johns Hopkins Scleroderma Center, and were kindly provided to us by Drs. F. Wigley and L. Hummers. Sera from 30 healthy controls, with no history, symptoms or signs of scleroderma or other autoimmune diseases, were collected. All sera were collected in compliance with Institutional Review Board and HIPPA regulations. Western Blotting. Cell lysates were made in lysis buffer (10 mM Tris pH 7.6, 2 mM EDTA pH 7.6, 150 mM NaCl, 1% NP-40) with 1 µg/ml each of pepstatin, leupeptin, antipain, chymotrypsin, and 1mM phenylmethylsulfonylfluoride. Equal protein amounts of the lysates were electrophoresed on SDS-PAGE gels, transferred to nitrocellulose or polyvinylidene difluoride membrane and immunoblotted as described.18 Human patient sera were used to detect poly ADP (ribose polymerase) (PARP) and full length and C-terminal Ufd2, and polyclonal rabbit antibodies were used to probe the N-terminus of Ufd2 (gift of Dr. A. Koff, Sloan Kettering Institute), securin (Zymed), and active caspase 3 (New England Biolabs). A monoclonal antibody recognizing the N-terminus of separase was used to probe separase and its cleavage fragment. Visualization of the blotted bands was performed using horseradish peroxidase-labeled secondary antibodies (Jackson Immunoresearch) and enhanced chemiluminescence. Granzyme B and λ Phosphatase Treatment. Unsynchronized or nocadazole-treated K562 cells were lysed on ice in lysis buffer containing all the protease inhibitors above except chymotrypsin in the absence or presence of 5 mM N-ethylmaleimide (Sigma Aldrich). Purified human granzyme B (generous gift of N. Thornberry, Merck, Rahway, NJ) was added at 42 nM as previously described.19 λ phosphatase (New England Biolabs) was added at 200 U/75 µg total protein. Samples were incubated at 30˚C for 30 minutes before terminating the reaction by adding gel application buffer and boiling. Production and Purification of Recombinant Proteins. Full-length Ufd2 sequences containing C-terminal histidine tags were generated by PCR and inserted into pBacPAK8 (Clontech). Recombinant viruses were generated and used to infect Sf9 cells. Hi-5 cells were infected with Sf9 media supernatants. To produce full length Ufd2, the infected cells were grown for two days at 27˚C, and then moved to 4˚C for 24 hours. Infected cells were lysed, and recombinant proteins were purified by Ni2+-nitrilotriacetate chromatography. Eluted recombinant proteins were dialyzed against
www.landesbioscience.com
20 mM Tris/HCl, pH 7.5, containing 10% (v/v) glycerol, 150 mM NaCl, and 1 mM dithiothreitol and stored at -80˚C in small aliquots. cDNAs encoding His-tagged UbcH5c and UbcH4 (kindly provided by Keiichi Nakayama, M.D., Ph.D., Kyushu University, Japan), were produced in E. coli and purified via Ni2+-nitrilotriacetate chromatography as for Ufd2. Ubiquitylation Assays. Ubiquitylation reaction mixtures (20 µl total volume) consisted of 0.25 µg/mL ubiquitin (Sigma), 10 mM Tris-HCl pH 7.5, 25 mM NaCl, 5 mM MgCl2, 0.5 mM DTT, 2 mM ATP, 1 mM creatine phosphate, and 0.5 U creatine phosphokinase. The following reagents were added as noted in the figure legends: (i) purified Rabbit E1 (33 or 100 nM, Boston Biochem) (ii) GST-UbcH5c (2 µg) (iii) [35S] methioninelabeled securin or cyclin B1 made by in vitro transcription/translation (IVTT) (Stratagene). Purified Ufd2 (0.5–1 µg) was added for 7, 15 or 60 minutes at 30˚C. Reactions were terminated by adding SDS gel application buffer and samples were electrophoresed on SDS-PAGE gels. Intact and ubiquitylated proteins were visualized by fluorography. To observe securin ubiquitylation in vivo, HeLa cells were transfected with siRNA oligos as described above. After 48 hours, the proteasome inhibitors MG132 (50 µM) and lactacystin (20 µM, both from Sigma) were added for 12 hours. Cell lysates were generated as described above and Protein A beads (Pierce) precomplexed with the polyclonal antibody to securin were added for one hour at 4˚C. The immunoprecipitations were washed four times with lysis buffer and boiled in gel application buffer. One-tenth of the immunoprecipitation volume was removed for immunoblotting with the anti-securin antibody. The remaining volume was immunoblotted using a mouse monoclonal antibody recognizing ubiquitin (BD Biosciences). siRNA Studies. 21 nucleotide RNAs were chemically synthesized using ribonucleoside phosphoramidites (Dharmacon). The siUfd2 sequence used was 5’ AAUACUUCUCAGGGCCUGCCAdTCdT 3’. The mismatch RNA sequence used was 5’AAUACGUCUCAUGGCGUGCCAdTdT 3’ and contained three point mutations but retained GC content. The oligos were transfected into HeLa or HSG cells using Oligofectamine (Invitrogen). The separase siRNA sequence used was 5’ AAGCUUGUGAUGCCAUCCUGAdTdT 3’. In addition, a Golgin 160 siRNA oligo (gift of Dr. Carolyn Machamer, Johns Hopkins University) was used as a control. FACS and Immunofluorescence. To assay cell cycle progression by FACS, cells were fixed at various time points using 75% ethanol, stained with propidium iodide (PI, Molecular Probes) and analyzed by flow cytometry. For immunofluorescence assays,20 cells were grown on #1 coverslips, fixed with 4% paraformaldehyde in PBS, permeabilized with cold acetone and stained with DAPI, anti-cyclin B1 polyclonal antibodies (Pharmingen), anti-Eg5 monoclonal antibody (Pharmingen), or affinity purified rabbit polyclonal anti-Ufd2, followed by fluorescein-5-isothiocyanate (FITC) and Texas Red labeled secondary antibodies (Jackson Immunoresearch Laboratories). To visualize mitotic spindles, cells were rinsed in microtubule stabilizing buffer (MTSB) (0.1M Pipes pH6.9, 1mM EGTA, 1mM MgCl2, 4% PEG-6000, and 250 µM GTP), then permeabilized in MTSB containing 0.1% Triton X-100 for 5 min at 37˚C. After washing with MTSB and fixing for 20 min. in MTSB containing 4% paraformaldehyde, cells were incubated with mouse monoclonal anti-β-tubulin antibody (Sigma) and appropriate secondary reagents. Cells were mounted with Pro-long antifade reagent (Molecular Probes) and analyzed on a Zeiss Axioscope microscope. Metaphase Spreads. HSG cells were transfected as above with Ufd2 siRNA or mismatch siRNA and incubated for 64 hours. Cells were then treated with 0.1 µg/ml colcemid (KaryoMax, Life Technologies) for the indicated times and processed by standard methods.21 Enzyme-Linked Immunosorbent Assays (ELISAs). 96-well high-binding plates were coated overnight with 5 µg/ml purified His-Ufd2 in carbonate/ bicarbonate binding buffer (pH 9.5) at 4˚C. Plates were subsequently washed, then blocked for 2 hours at room temperature with buffer containing PBS, 5% BSA, 0.05% Tween 20 and 0.02% sodium azide. Patient sera were tested at 1:1,000 dilution in binding buffer (PBS, 1% BSA and 0.05% Tween 20) followed by horseradish peroxidase-conjugated goat anti-human IgG (Jackson ImmunoResearch, Avondale, PA) diluted 1:5,000 in binding
Cell Cycle
1639
UFD2 REGULATES SISTER CHROMATID SEPARATION
A
mean values, in at least 2 independent experiments, greater than the mean plus 2 standard deviations of 25 normal controls. Statistical Analysis. All immunoblots and photomicrographs are representative of ≥ 3 separate experiments. The NASS/PASS statistical analysis program was used to perform two-tailed T-tests and create box plots. Where no Y-axis bars are seen, standard deviations fall within the shaded box. Quantification of immunofluorescence data was performed by counting ≥ 1000 cells per coverslip for ≥ 2 separate experiments.
B
RESULTS
Figure 1. Ufd2 is Phosphorylated During Mitosis. (A) Lysates were prepared from unsynchronized HSG, HeLa and K562 cells (- noc), or cells arrested in prometaphase with nocodazole (+ noc). (B) K562 cells in interphase or mitosis were lysed in the absence or presence of NEM and incubated for 10 min on ice. Purified λ phosphatase was then added followed by incubation at 30˚C for 30 min.
buffer (both incubations for 2 hours at room temperature). Detection was performed with ABTS peroxidase substrate and hydrogen peroxide (H2O2) per manufacturer’s directions (Kirkegaard and Perry, Gaithersburg, MD). All sera were tested in duplicate. Positive sera were defined as those with
Antibodies to Ufd2 are Found in Scleroderma Patient Sera. Ufd2 was originally identified in an expression screen of a patient serum recognizing a 130 kD protein cleaved during apoptosis.17 We subsequently identified antibodies recognizing Ufd2 in 2 patients with scleroderma. To define the prevalence of Ufd2 autoantibodies in scleroderma, sera from 30 healthy controls and from 55 patients with scleroderma were analyzed by ELISA using recombinant Ufd2. Ufd2 antibodies were observed in 5/55 (9%) scleroderma patients screened, but in 0/30 (0%) controls (the difference between the cohorts was statistically significant; p = 0.002, Wilcoxon Rank Sum test). Ufd2 is Phosphorylated during Mitosis. Since many scleroderma autoantigens have roles in mitosis, we initially examined Ufd2 expression during the cell cycle. While Ufd2 migrated as a single band at 140kD in interphase cells, it migrated at 165kD in cells arrested in prometaphase by nocadazole (Fig. 1A) or other microtubule disrupting agents (data not shown). Elutriation of untreated cells confirmed that Ufd2 underwent a dramatic change in migration on SDS-PAGE gels only in mitotic cell populations (data not shown). When mitotic cell lysates were incubated at 30˚C, the 165kD species was extremely labile (Fig. 1). Addition of N-ethylmaleimide ((NEM), which alkylates free cysteine residues) stabilized the
A
C
B
D
Figure 2. Inhibition of Ufd2 expression causes cells to accumulate in mitosis and decreases progression into anaphase. (A) HeLa cells were transfected with mismatch control (C) or Ufd2 (U) siRNA, and incubated at 37˚C for the indicated times before lysing. Immunoblots were performed on equal protein amounts (vinculin was used as a loading control). (B and C) HeLa cells were transfected with mismatch or Ufd2 siRNA. At various time points, cells were fixed and stained with PI. Representative histograms at 68 hours after transfection are depicted in (B) and quantification of the percentage of cells in G2M (by 2N DNA content) at each time point is shown in (C). (D) HeLa and HSG cells were transfected with mismatch (C) and Ufd2 (U) siRNA for 62 hours, were stained with anti-cyclin B1 antibodies and DAPI to differentiate metaphase (cyclin B-positive) and anaphase (cyclin B-negative) cells. Box plots depict the percentage of mitotic cells in anaphase. Means ± SD are: HeLa-Mismatch 28.0 ± 3.6 vs. Ufd2 20.4 ± 8.0 *p = 0.003; HSG-Mismatch 35.9 ± 1.3 vs. Ufd2 7.1 ± 0.85; **p
