|
|
|
Original Article |
|
||
|
Acta
Biochim Biophys
Sin 2010, 42: 303 –310
|
|||
|
doi:
10.1093/abbs/gmq022.
|
Tetra-glutamic acid residues adjacent to Lys248 in HMG-CoA reductase are critical
for the ubiquitination mediated by gp78 and UBE2G2
Honghua Miao1, Wei Jiang1,2, Liang Ge1, Boliang Li1, and Baoliang Song1*
1The State Key Laboratory of Molecular Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences,
Chinese Academy of Sciences, Shanghai 200031, China
2Institute of Cancer Biology and Drug Screening, School of Life Sciences, Lanzhou University, Lanzhou 730000, China
*Correspondence address. Tel: þ86-21-54921629; Fax: þ86-21-54921629; E-mail: blsong@sibs.ac.cn
Sterol-regulated degradation of 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGCR) is a rapid feedback regulatory mechanism by which cells employ to control the cholesterol biosynthesis. This process is initiated by the sterol-induced interaction between HMGCR and Insig-1/gp78, a membrane-bound ubiquitin ligase complex. There are two Lys residues (Lys89 and Lys248) facing cytosol inthe membrane domain of HMGCR, and Lys248 is the major ubiquitination site. In this study, we investigated the mechanism of ubiquitination site selection in HMGCR. We find that the distance of Lys248 to membrane is dispensable for its ubiquitination. However, the conserved tetra-glutamic acid residues adjacent to Lys248 in HMGCR are essential. Replacement of these negatively charged residues with tetraarginine causes the resistance of HMGCR to sterol-induced ubiquitination and degradation, albeit this mutant HMGCR can still binds to Insig-1. We further find that the tetra-glutamic acid residues are necessary but not sufficient for the
modification on their adjacent Lys, since they are not functional on Lys89 of HMGCR or in SCAP. UBE2G2, a previously known E2 of gp78, is demonstrated to be involved in the sterol-regulated ubiquitination and degradation of HMGCR. In summary, these results identify the tetraglutamic acid residues as a critical motif in HMGCR for the ubiquitination reaction mediated by gp78 and UBE2G2.
Keywords HMG-CoA reductase; endoplasmic reticulum associated protein degradation (ERAD); tetraglutamic acid residue; gp78; UBE2G2
Received: November 18, 2009 Accepted: March 7, 2010
Introduction
The enzyme 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase (HMGCR) catalyzes the reduction in
role in the sterol-regulated degradation of HMGCR [3]. When the concentration of cellular sterol is high, the membrane domain of HMGCR can sense the excessive sterols and the whole protein is rapidly degraded, therefore the cholesterol biosynthesis is shut down [3]. This sterolregulated degradation of HMGCR is a very important negatively feedback regulatory mechanism that cells use to control their cholesterol level.
The degradation of HMGCR is through the ubiquitin–proteasome pathway and is synergistically regulated by sterol and non-sterol isoprenoid [1,4,5]. Two types of sterols are known to stimulate the ubiquitination and degradation of HMGCR: the oxysterol 25-hydroxycholesterol (25-HC) and the cholesterol precursor lanosterol [6,7]. Sterol promotes the degradation of HMGCR by triggering the binding of the HMGCR membrane domain to Insig-1 or -2, which constitutively interacts with gp78, a membrane-bound ubiquitin ligase (E3) [5,8]. Gp78 then catalyzes the ubiquitination of HMGCR at two specific lysine residues: Lys89 and Lys248, among which Lys248 is the major ubiquitination site [5]. The ubiquitinated HMGCR is finally degraded by proteasome. Through direct interaction with gp78, other proteins including VCP and Ufd1 are also involved in the sterol-regulated degradation of HMGCR [9].
In this study, we investigated the mechanism of ubiquitination site selection in HMGCR, and found that the tetra-glutamic acid residues immediately adjacent to Lys248 were necessary but not sufficient for the ubiquitination on their adjacent Lys residue.
Materials and Methods
Materials
MG-132 and digitonin were obtained from Calbiochem (San Diego, USA); sterols were from Steraloids, Inc.(Newport, USA); the solution of sodium mevalonate was described as in Kita et al. [10]; horseradish peroxidaseconjugated donkey anti-mouse and anti-rabbit immunoglobulin G (IgG) was from Jackson ImmunoResearch Laboratories (West Grove, USA); ubiquitins were from Boston Biochem (Cambridge, USA); lipoprotein-deficient serum (LPDS, d . 1.215 g/ml) was prepared from newborn calf serum by ultracentrifugation [11]; other regents were described previously [9].
The following plasmids were described previously as indicated in the sources: pCMV-gp78-Myc [8]; pCMV-HMGRed-T7 and pCMV-Insig-1-Myc [12]; pEF1a-HA-ubiquitin [5]; pCMV-UBE2G2, encoding human UBE2G2; pCMVUBE2G2 (C89S), encoding dominant-negative form of human UBE2G2, in which C89 was mutated to serine; pCMV-HMG-Red-T7 (KEEEEN), pCMV-HMG-Red-T7 (EKEEEN), pCMV-HMG-Red-T7 (EEKEEN), pCMV-HMGRed-T7 (EEEKEN), pCMV-HMG-Red-T7 (EEEEKN), pCMV-HMG-Red-T7 (E243EEE to DDDD), pCMV-HMGRed-
T7 (E243EEE to RRRR), pCMV-HMG-Red-T7 (E243R), pCMV-HMG-Red-T7 (E244R), pCMV-HMGRed-T7 (E245R), pCMV-HMG-Red-T7 (E246R), and HMG-1/2 were constructed based on pCMV-HMG-Red-T7 using a site-directed mutagenesis kit (Stratagene, La Jolla, USA); SCAP-1/2/3/4 were generated by the standard PCR methods based on the plasmid pCMV-SCAP, which encoded hamster SCAP [13].
Primary antibodies used in this paper for immunoblotting were as follows: mouse monoclonal anti-T7 (Novagen, Madison, USA); mouse monoclonal anti-Myc IgG-9E10 (Roche, Indianapolis, USA); mouse monoclonal anti-HA (Sigma, Milwaukee, USA); mouse monoclonal antiubiquitin IgG-P4D1 (Santa Cruz, Santa Cruz, USA); mouse monoclonal antibody against the catalytic domain of hamster HMGCR (amino acid 450–887) IgG-A9 [14]; polyclonal antibody against human HMGCR cytosolic domain was generated by immunizing rabbits followed by affinity purification with antigens [9]. IgG-9D5, mouse monoclonal antibody against hamster SCAP, was prepared from cell line CRL-2347 from ATCC (Manassas, USA) according to a previous reference [13].
Cell culture
Chinese hamster ovary-K1 (CHO-K1) cells were maintained in a monolayer culture at 378C in a 5%CO2 incubator. These cells were maintained in medium A [1:1 mixture of Ham’s F-12 medium and Dulbecco’s modified Eagle’s medium (DMEM) containing 100 units/ml penicillin and 100 mg/ml streptomycin sulfate] supplemented with 5% (v/v) fetal calf serum (FCS). SRD-13A cells were SCAP null mutant cells derived from CHO cells [15], which were maintained in medium B [medium A supplemented with 5% (v/v) FCS, 5 mg/ml cholesterol, 1 mM sodium mevalonate, and 20 mM sodium oleate] at 37oC in 5%CO2. SV589 cells were immortalized human fibroblasts by expressing the SV-40 large T-antigen and maintained in monolayer in DMEM containing
100 units/ml penicillin and 100 mg/ml streptomycin sulfate supplemented with 10% (v/v) FCS at 37oC in 5%CO2.
Transient transfection
Cells were transiently transfected with FuGENE 6 (Roche Applied Science, Indianapolis, USA). For each transfection, the FuGENE 6 DNA transfection reagent was added to 0.2 ml medium A at a ratio of 3 ml of FuGENE 6 per 1 mg of DNA, and the total amount of DNA was adjusted to 3 mg/60-mm dish with pcDNA3 (Invitrogen, Carlsbad, USA). After treatment, triplicate dishes of cells for each condition were harvested and pooled for analysis.
Immunoprecipitation
The harvested cells were lysed in a detergent-containing buffer, and immunoprecipitations were carried out with monoclonal anti-T7 IgG-coupled agarose (Novagen) against transfected HMGCR or polyclonal antibody against human endogenous HMGCR as described previously [9].
Ubiquitination of HMGCR
The harvested cells were lysed in a detergent-containing buffer, and immunoprecipitations were carried out with either polyclonal antibody against the human endogenous HMGCR or monoclonal anti-T7 IgG-coupled agarose (Novagen) against transfected HMGCR as described previously [9]. Aliquots of the immunoprecipitates were subject to 6% SDS–PAGE, transferred to nylon membranes, and then subject to immunoblot analysis.
RNA interference
Duplexes of small interfering RNA (siRNA) targeting UBE2G2 were synthesized by Genepharma (Shanghai, China). The sequences of siRNA oligos are as follows: 5'-GUGUGGAGAAGAUCCUGCU-3' (sense); 5'-AGCAGGAUCUUCUCCACAC-3' (antisense). RNA interference experiments were carried out as described previously [9].
Results
The distance of Lys248 from membrane was not important for the sterol-regulated ubiquitination of HMGCR
Our previous study showed that mutation of Lys248 caused the resistance to sterol-induced ubiquitination and
Previous study on yeast HMGCR isozyme Hmg2p showed that the distance of ubiquitination site Lys6 from the ER membrane was critical for its ubiquitination and degradation [16]. We then designed experiments to test whether the distance of Lys248 from the ER membrane was required for the sterol-regulated degradation of HMGCR. CHO-K1 cells were transfected with plasmids encoding different versions of HMGCR and Insig-1. After transfection, the cells were depleted from cholesterol and treated for additional 5 h with 25-HC plus 10 mM mevalonate, which allowed for theproduction of non-sterol isoprenoids that enhanced. HMGCR degradation. As expected, wild-type HMGCR was rapidly degraded in the presence of sterol and mevalonate [Fig. 2(A), compare lane 2 with lane 1]. When Lys248 was moved to different positions, these mutants of HMGCR were similarly degraded as wild-type protein did [Fig. 2(A)], even in the context of K89R (SupplementaryFig. S1). These data showed that the distance of the ubiquitination site of HMGCR from the membrane was irrelevant to the sterol-regulated degradation.
Acidic residues adjacent to Lys248 are critical for the degradation of HMGCR
From the sequence alignment results, we noticed that the tetra-glutamic acid residues adjacent to Lys248 were highlyconserved [Fig. 1(B)]. To test whether these residues are essential for the function of Lys248, we substituted them by acidic, basic, or neutral residues, and tested their effects on sterol-regulated degradation. The mutant of HMGCR with EEEE to DDDD was subjected to the sterol-regulated degradation as wild-type HMGCR did [Fig. 2(B), lanes 1–8]. However, substitution of all four glutamic acid residues of tetra-arginine completely prevented the protein from sterol-induced degradation [Fig. 2(B), compare lanes 13–16 with lanes 9–12]. Nevertheless, replacement of any single E with R had no significant effect on the sterolstimulated degradation of HMGCR [Fig. 2(C), lanes 5–12]. When the four glutamates were changed into four alanines, the mutant HMGCR did not undergo sterolregulated
degradation [Supplementary Fig. S2(A)], although its level was decreased when the Insig-1 was increased [Supplementary Fig. S2(A), lanes 7–12]. It is possible that the increased Insig-1 affects the expression of mutated HMGCR. Ubiquitination analysis showed that the ubiquitination of mutant HMGCR (tetra-E to tetra-A) was not affected by sterol or Insig-1 [Supplementary Fig. S2(B), lanes 5–8], suggesting that the substitution of tetra-A for tetra-E ablates the sterol-regulated ubiquitination and degradation of HMGCR. Together, the above data showed that the negative charges adjacent to Lys248 were
required for the sterol-regulated ubiquitination and degradation of HMGCR.
Mutation of the tetra-glutamic acids impairs the ubiquitination of HMGCR but not its interaction with Insig-1/gp78
We next tested whether the substitution of tetra-glutamic acid of four arginines impaired the ubiquitination and interaction with Insig-1. After transfection and sterol-depletion,cells were treated without or with 25-HC in the presence of 20 mM MG-132 (a proteasome inhibitor to block protein degradation) for 1 h. HMGCR was then immunoprecipitated and its ubiquitination was analyzed. In the presence of Insig-1, 25-HC promoted the ubiquitination of wild-type
HMGCR [Fig. 3(A), lanes 1–4]. However, 25-HC failed to induce the ubiquitination of mutant HMGCR in which EEEE was changed to RRRR in the presence of Insig-1 and 25-HC [Fig. 3(A), lane 8]. Together, the above results showed that the EEEE motif was required for sterolstimulated ubiquitination and subsequent degradation of HMGCR [Fig. 2(B), lanes 13–16].
It has been known that the sterol-induced binding of HMGCR to Insig-1/gp78 was prerequisite for the ubiquitination and degradation of HMGCR, and Insig-1 or -2 constitutively interacts with gp78 [8]. We then tested whether the mutation of EEEE to RRRR reduced the affinity of HMGCR to Insig-1 in the absence and the presence of sterol. After transfection and sterol-depletion, cells were treated with 25-HC in the presence of MG-132 for 20 min.
The cell lysate was prepared and subjected to immunoprecipitation with anti-T7-coupled agarose to pull down HMGCR-T7. As shown in Fig. 3(B), Insig-1 wasco-precipitated in the pellet fraction, no matter wild-type or mutant HMGCR [Fig. 3(B)]. These results suggested that the replacement of EEEE to RRRR did not affect the sterol-stimulated binding of HMGCR to Insig-1/gp78 complex, but impaired the ubiquitination reaction on Lys248.
The tetra-glutamic acid motif was necessary but not sufficient for sterol-regulated degradation of HMGCR
Previous study has shown that in HMGCR, Lys89 was a minor ubiquitination site [5]. We wondered whether the EEEE motif flanking Lys248 was sufficient to increase the ubiquitination efficiency of a Lys residue. To test this possibility, we inserted the tetra-glutamic acid motif into the site between Ser88 and Lys89 of HMGCR, meanwhile the Lys248 was mutated to Arg so that the ubiquitination could not happen on Lys248 [Fig. 4(A)]. We then transiently transfected the plasmids encoding the mutated and wild-type HMGCR together with Insig-1 into CHO-K1 cells. After transfection, the cells were depleted of sterol and treated with 25-HC plus high level of mevalonate (10 mM). The HMGCR with Lys248 to Arg mutation was resistant to sterol-induced degradation [Fig. 4(B), comparelane 8 with lane 7]. Insertion of the tetra-E motif did not render Lys89 to be a dominant ubiquitination site since the mutation with tetra-E insertion could not undergo sterolregulated degradation or ubiquitination [Fig. 4(B), compare lane 12 with lane 11, Supplementary Fig. S3].
SCAP is another Insig-binding protein [17,18]. However, unlike HMGCR that is rapidly degraded mediated by Insigs, SCAP is very stable even after it binding to Insigs. The mechanism for the different behaviors of HMGCR and SCAP is unclear. It is possible that SCAP lacks the ubiquitination site(s); therefore, it cannot be modified by gp78. We then introduced the EEEENK motif before the first, third, fifth, or seventh transmembrane helix of SCAP to see whether we could created a ubiquitination site, which would render SCAP undergo the sterol-regulated degradation. Figure 4(C) shows different forms of SCAP in which the framed amino acids were replaced by EEEENK motif, respectively. The SCAP-deficient SRD-13A cells were transfected with different combinations of plasmids. The wild-type HMGCR comprising EEEENK motif was degraded when exposed to 25-HC. However, the differentmutants of SCAP with EEEENK motif could not be degraded even in the presence of 25-HC and mevalonate [Fig. 4(D), compare lanes 5–12 with lanes 1–4].
Together, these results indicated that the tetra-glutamic acid residues are necessary but not sufficient for the ubiquitination
on the adjacent Lys residue.
UBE2G2 was required for the ubiquitination and degradation of HMGCR
Previous studies have shown that an E2 named UBE2G2 (Ubc7) binds gp78 and is involved in the ubiquitination and degradation of some proteins such as CD3-d [19]. However, whether UBE2G2 is required for the ubiquitination of endogenous HMGCR has not been documented. In order to test the role of UBE2G2 in the sterol-regulated ubiquitination of HMGCR, we transfected SV589 cells with siRNA to reduce the expression of UBE2G2 and measured the ubiquitination of endogenous HMGCR. In cells receiving control siRNA, 25-HC strongly induced the ubiquitination of HMGCR [Fig. 5(A), lane 2]. When
siRNA targeting UBE2G2 was introduced into cells, the25-HC-regulated ubiquitination of HMGCR was dramatically decreased [Fig. 4(A), compare lane 4 with lane 2]. In the same experiment, we used quantitative real-time PCR to show that the UBE2G2 mRNA was reduced to less than 10% by RNAi [Fig. 5(B)].
We next performed a transfection experiment to determine the effect of UBE2G2 on the sterol-regulated degradation of HMGCR. Wild-type UBE2G2 accelerated the degradation of HMGCR in a concentration-dependent manner [Fig. 5(C), lanes 3–6], whereas the dominantnegative UBE2G2 in which the catalytic Cys89 was changed to Ser blocked the degradation of HMGCR [Fig. 5(C), lane 8]. Together, these results showed that the UBE2G2 is required for the sterol-induced degradation of HMGCR.
Discussion
This study explores the mechanism of ubiquitination site selection in HMGCR. We find that the tetra-glutamic acids adjacent to Lys248 are necessary but not sufficient for the ubiquitination on Lys248 in HMGCR. When these tetraglutamic acids are substituted for four aspartic acids, or mutation of any single E to R [Fig. 2(C)], HMGCR can still undergo sterol-regulated degradation. However, when the tetra-E is replaced by tetra-R, the sterol-promoted
degradation of HMGCR was completely blocked [Fig. 2(B)]. These results suggest that the negativelycharged environment formed by multiple acidic residues is essential for the ubiquitination on Lys248. The wild-type and mutant HMGCR with tetra-E to tetra-R can both bind Insig-1 (Fig. 3), suggesting the tetra-E motif does not impair the interaction between HMGCR and Insig-1/gp78. Interestingly, it has been found that the positive charge of Lys91 in yeast E2 Ubc4 is important for the ubiquitination reaction [20]. It is possible that the carboxyl groups of tetra-E in HMGCR interact with the positively charged group in UBE2G2 through ionic interactions, which facilitates the ubiquitination on Lys248.
It was found that the lysine was a preference ubiquitination site in KEEE motif [21]. Furthermore, Catic et al. [20] found that Asp and/or Glu had a dominant enrichment at position ‘-2’and ‘-1’ (Lys attaching ubiquitin chains as position ‘0’) in transmembrane proteins and membrane-associated proteins
compared with either non-membrane-associated proteins or proteome of yeast. Our data shown in this paper are in accordance with these published observations. Nevertheless, if the tetra-negatively charged residues are inserted into the position adjacent to another ubiquitination site Lys89 of HMGCR [Fig. 4(A)], or six amino acids near transmembrane helix-1/3/5/7 of SCAP are replaced by ‘EEEENK’ motif, respectively [Fig. 4(C)], we do not find the modified proteins are subjected to sterol-induced degradation [Fig. 4(B,D)].
These findings suggest that other motifs may be important for the ubiquitination site selection, which need further investigation.
Supplementary Data
Supplementary data are available at ABBS online.
Acknowledgements
We thank Yu-Xiu Qu and Su-Zhe Pan for technical assistance.
Funding
This work was supported by grants from the Ministry of Science and Technology of China (2006CB910603 and 2009CB919000), National Natural Science Foundation of China (90713025), and Shanghai Science and Technology Committee (08JC1421300 and 08431900500).
References
1 Goldstein JL and Brown MS. Regulation of the mevalonate pathway. Nature 1990, 343: 425–430.
2 Roitelman J, Olender EH, Bar-Nun S, Dunn WA, Jr and Simoni RD. Immunological evidence for eight spans in the membrane domain of 3-
hydroxy-3-methylglutaryl coenzyme A reductase: implications for enzyme degradation in the endoplasmic reticulum. J Cell Biol 1992, 117:
959–973.
3 Gil G, Faust JR, Chin DJ, Goldstein JL and Brown MS. Membrane-bound domain of HMG CoA reductase is required for sterol-enhanced degradation of the enzyme. Cell 1985, 41: 249–258.
4 Ravid T, Doolman R, Avner R, Harats D and Roitelman J. The ubiquitin–proteasome pathway mediates the regulated degradation of mammalian
3-hydroxy-3-methylglutaryl-coenzyme A reductase. J Biol Chem 2000, 275: 35840–35847.
5 Sever N, Song BL, Yabe D, Goldstein JL, Brown MS and DeBose-Boyd RA. Insig-dependent ubiquitination and degradation of mammalian
3-hydroxy-3-methylglutaryl-CoA reductase stimulated by sterols and geranylgeraniol. J Biol Chem 2003, 278: 52479–52490.
6 Song BL, Javitt NB and DeBose-Boyd RA. Insig-mediated degradation of HMG CoA reductase stimulated by lanosterol, an intermediate in the synthesis
of cholesterol. Cell Metab 2005, 1: 179–189.
7 Song BL and DeBose-Boyd RA. Ubiquitination of 3-hydroxy-3-methylglutaryl-CoA reductase in permeabilized cells mediated by cytosolic
E1 and a putative membrane-bound ubiquitin ligase. J Biol Chem 2004, 279: 28798–28806.
8 Song BL, Sever N and DeBose-Boyd RA. Gp78, a membrane-anchored ubiquitin ligase, associates with Insig-1 and couples sterol-regulated ubiquitination to degradation of HMG CoA reductase. Mol Cell 2005, 19: 829–840.
9 Cao J, Wang J, Qi W, Miao HH, Wang J, Ge L and DeBose-Boyd RA, et al. Ufd1 is a cofactor of gp78 and plays a key role in cholesterol metabolism by
regulating the stability of HMG-CoA reductase. Cell metab 2007, 6: 115–128.
10 Kita T, Brown MS and Goldstein JL. Feedback regulation of 3-hydroxy-3-methylglutaryl coenzyme A reductase in livers of mice
treated with mevinolin, a competitive inhibitor of the reductase. J Clin Invest 1980, 66: 1094–1100.
11 Goldstein JL, Kita T and Brown MS. Defective lipoprotein receptors and atherosclerosis. Lessons from an animal counterpart of familial hypercholesterolemia. N Engl J Med 1983, 309: 288–296.
12 Sever N, Yang T, Brown MS, Goldstein JL and DeBose-Boyd RA. Accelerated degradation of HMG CoA reductase mediated by binding of
insig-1 to its sterol-sensing domain. Mol cell 2003, 11: 25–33.
13 Sakai J, Nohturfft A, Cheng D, Ho YK, Brown MS and Goldstein JL.Identification of complexes between the COOH-terminal domains of sterol
regulatory element-binding proteins (SREBPs) and SREBP cleavage-activating protein. Biol Chem 1997, 272: 20213–20221.
14 Liscum L, Luskey KL, Chin DJ, Ho YK, Goldstein JL and Brown MS. Regulation of 3-hydroxy-3-methylglutaryl coenzyme A reductase and its
mRNA in rat liver as studied with a monoclonal antibody and a cDNA probe. J Biol Chem 1983, 258: 8450–8455.
15 Rawson RB, DeBose-Boyd R, Goldstein JL and Brown MS. Failure to cleave sterol regulatory element-binding proteins (SREBPs) causes cholesterol
auxotrophy in Chinese hamster ovary cells with genetic absence of SREBP cleavage-activating protein. J Biol Chem 1999, 274: 28549–28556.
16 Gardner RG and Hampton RY. A ’distributed degron’ allows regulated entry into the ER degradation pathway. EMBO J. 1999, 18: 5994–6004.
17 Yang T, Espenshade PJ, Wright ME, Yabe D, Gong Y, Aebersold R and Goldstein JL, et al. Crucial step in cholesterol homeostasis: sterols
promote binding of SCAP to INSIG-1, a membrane protein that facilitates retention of SREBPs in ER. Cell 2002, 110: 489–500.
18 Yabe D, Brown MS and Goldstein JL. Insig-2, a second endoplasmic reticulum protein that binds SCAP and blocks export of sterol regulatory
element-binding proteins. Proc Natl Acad Sci USA 2002, 99: 12753–12758.
19 Fang S, Ferrone M, Yang C, Jensen JP, Tiwari S and Weissman AM. The tumor autocrine motility factor receptor, gp78, is a ubiquitin protein ligase
implicated in degradation from the endoplasmic reticulum. Proc Natl Acad
Sci USA 2001, 98: 14422–14427.
20 Catic A, Collins C, Church GM and Ploegh HL. Preferred in vivo ubiquitination sites. Bioinformatics 2004, 20: 3302–3307.
21 Jonassen I, Collins JF and Higgins DG. Finding flexible patterns in unaligned protein sequences. Protein Sci 1995, 4: 1587–1595.