Proteome Changes Induced by Knock-Down of the Deubiquitylating Enzyme HAUSP/USP7 Benedikt M. Kessler,†,‡ Elisabetta Fortunati,§ Monique Melis,| Cornelieke E. G. M. Pals,§ Hans Clevers,| and Madelon M. Maurice*,|,⊥ Department of Pathology, Harvard Medical School, 77 Avenue Louis Pasteur, Boston, Massachusetts 02115, Hubrecht Laboratorium, Netherlands Institute for Developmental Biology, Uppsalalaan 8, 3584 CT Utrecht, The Netherlands, and Department of Pulmonary Diseases, University Medical Center Utrecht, Heidelberglaan 100 3584 CX Utrecht, The Netherlands
Received April 17, 2007
Modification of proteins by ubiquitin plays a major role in a broad array of biological processes. Reversal of this process through deubiquitylation likely represents an important regulatory step in the maintenance of cellular homeostasis. However, the biological functions of deubiquitylating enzymes still remain poorly characterized. To investigate the biological role of the herpes virus-associated ubiquitin-specific protease HAUSP/USP7, we have generated stably transfected cells carrying inducible shRNA expression plasmids. USP7 mRNA and protein were strongly down-regulated 48-72 h after shRNA induction. We used a selected clone to compare whole-cell proteomes by 2D-SDS-PAGE before and after knockdown of USP7. Alterations in 36 proteins were detected and their identities were revealed by mass spectrometry analysis. Components of the replication machinery, DNA/RNA binding proteins, enzymes involved in apoptosis and metabolism were found to be down-regulated upon USP7 removal, representing proteins that are either more rapidly turned over or synthesized less efficiently in the absence of USP7-mediated deubiquitylation. Alix/HP95, a protein implicated in endosomal organization and virus budding, was confirmed by immunoblotting to become down-regulated when USP7 levels were reduced. Our results extend the current list of USP7-dependent biological processes and suggest a role for this enzyme not only in transcriptional regulation but also in DNA replication, apoptosis, and possibly endosomal organization. Keywords: Ubiquitin • HAUSP/USP7 • proteomics • inducible shRNA • mass spectrometry • LC-MS/MS • MALDI • 2D gel electrophoresis • Alix/Hp95
Introduction Covalent attachment of ubiquitin (Ub) to proteins is a crucial regulatory step in a diverse array of cellular and biological processes, including embryonic development, cell cycle, growth control, and prevention of neurodegeneration. Conjugation of ubiquitin chains linked through lysine 48 primarily serves as a targeting signal for proteasomal degradation.1,2 In addition, modification of proteins with a single Ub molecule or with K63linked Ub chain regulates processes that range from endocytosis and membrane transport to transcriptional regulation.3-5 * To whom correspondence should be addressed. Dr. Madelon Maurice, Department of Cell Biology, University Medical Center Utrecht, Rm G02.525, Heidelberglaan 100, 3584CX Utrecht, The Netherlands. Phone: +31 30 250 6551. E-mail:
[email protected]. † Harvard Medical School. ‡ Present address: Henry Wellcome Building for Molecular Physiology, Department of Clinical Medicine, University of Oxford, Roosevelt Drive, Oxford OX3 7BN, UK. § University Medical Center Utrecht. | Netherlands Institute for Developmental Biology. ⊥ Present address: Department of Cell Biology, University Medical Center Utrecht, Heidelberglaan 100, 3584 CX Utrecht, The Netherlands. 10.1021/pr0702161 CCC: $37.00
2007 American Chemical Society
Ubiquitylation of proteins is a reversible process. The removal of Ub is carried out by deubiquitylating enzymes (DUBs), a large family of proteases that can release poly-Ub chains from proteins targeted for degradation, recycle monomeric Ub, liberate Ub from Ub-fusion precursors, reverse regulatory ubiquitylation and edit inappropriately ubiquitylated proteins.6-8 Five structurally different DUB enzyme families have been identified, comprising over 90 members all together in the human genome and novel putative ubiquitin proteases are still being discovered (reviewed in ref 9). The diversity of this enzyme class suggests discrete biological roles for each family with probably a limited set of specific substrates, in analogy to the family of ubiquitin conjugating enzymes (E3 ligases).10 One of the few DUBs that have been assigned a biological role is HAUSP/USP7. USP7 was identified to bind to and deubiquitylate p53.11 In subsequent studies, USP7-mediated regulation of Mdm2, an E3 Ub ligase for p53, and its structural homologue Mdmx were described.12,13 Based on its conservation between species, the primeval function of USP7 is likely to be more conserved than the p53 pathway. USP7 has also Journal of Proteome Research 2007, 6, 4163-4172
4163
Published on Web 10/10/2007
research articles
Figure 1. Inducible RNAi against endogenous USP7 in LS174 cells. LS174T colon carcinoma cells stably carrying a Tet repressor (TR)-construct and a pTER-USP7 shRNA construct (referred to as LS88) were treated with 1 µg/mL doxycycline for 24, 48, and 72 h and tested by Western blotting for down-regulation of endogenous USP7, using actin as a loading control.
been implicated in the regulation of Polycomb-mediated epigenetic silencing of homeotic genes in Drosophila Melanogaster.14 In these studies, mono-ubiquitylated Histone 2B was proposed to be the substrate for USP7 function. Consistent with a role for USP7 in transcriptional regulation, it was demonstrated that USP7 processes monoubiquitylated FOXO, a forkhead box O transcription factor whose function appears to be controlled by monoubiquitylation.15 Importantly, USP7 function was also shown to be exploited by herpes viruses through its interaction with the herpes simplex virus-derived immediate early protein ICP0 and EBV-derived EBNA1,16-19 which underscores its critical role in transcriptional regulation but probably also other biological processes. Thus far, appointed DUB substrates were mainly identified through common protein-protein interaction techniques. However, the general strength of interaction between a DUB enzyme and its substrate is an unknown parameter. Biochemical assays to demonstrate deubiquitylation of substrates may suffer from a lack of specificity, as a number of DUBs can deubiquitylate a similar substrate (Ub-bgal, di-/tetra-Ub) in vitro or upon overexpression.8,20 Knock-down approaches using RNAi-based model systems have proven to be helpful in determination of the function of DUBs in specific signaling pathways.21 Therefore, to investigate the biological role of USP7, we used an RNAi-based proteomics approach with which we compared proteomes of cells before and after knockdown of USP7. This strategy anticipates that affected pathways or individual substrates may disappear upon USP7 removal, due to a lack of deubiquitylation and therefore increased turnover by ubiquitin-dependent proteolysis. Our results extend the current knowledge of multiple cellular pathways that are affected by USP7, including replication, DNA/RNA processing, apoptosis, and possibly endosomal organization.
Results To study the biological function of USP7, we made use of the pTER plasmid that allows doxycycline-inducible knockdown of gene expression by short hairpin (sh)RNAs.39 We generated stable shRNA-inducible LS174T colon carcinoma cell clones in which endogenous protein levels of USP7 were knocked down significantly 48-72 h after induction with doxycycline (Figure 1). LS174T cells were chosen as they have an intact p53 DNA damage response pathway and a normal number of chromosomes. Elongated knockdown of USP7 for 5 days resulted in a G1 growth arrest accompanied by accumulation of p53, as reported earlier (refs 12, 13, and unpublished data). To explore the functional consequences of USP7 knockdown, we used a selected clone (LS88) to compare whole-cell pro4164
Journal of Proteome Research • Vol. 6, No. 11, 2007
Kessler et al.
teomes on 2D-SDS-PAGE before and 48 h after knockdown of USP7 (Figure 2). At the 48 h time point of shRNA induction, an onset of G1 arrest but no sign of apoptosis was observed as detected by the absence of a sub-G1 population and the absence of a cleavage product of the PARP protein (data not shown). Comparison of whole-cell proteomes yielded an alteration in 25 2D gel spots that were up- or down-regulated upon removal of USP7 (Figure 2A and B). These spots were excised from the gel (both A and B), digested with trypsin, and analyzed by mass spectrometry (LC-MS/MS and MALDI). With some spots containing a mixture of proteins, information was gathered from 36 individual proteins in total, from which 26 were down-regulated and 10 were up-regulated at reduced levels of USP7 (Table 1 and Figure 2C). Proteins with both cytoplasmic and nuclear functions were affected by USP7 knock-down, with altered proteins being involved in metabolismrelated processes (9 proteins), the translation machinery (7 proteins), protein folding (5 proteins), protein trafficking (2 proteins), DNA repair (2 proteins), transcription (2 proteins), protein turnover (2 proteins), and other functions (7 proteins). As reported previously,22,23 the induction of shRNA expression may lead to an IFN-mediated cellular stress response with downstream changes in protein expression as a consequence. To rule out IFN-mediated effects in our cells, we analyzed the expression of three IFN target genes (OAS1, ISGF3, ISG20) by Northern blotting before and after induction of shRNA (Figure 3 and data not shown). As a control, we included a LS174T clone (Myc9.2) that stably expresses an inducible shRNA against the Myc transcription factor. Twenty-four hours after shRNA induction, Myc9.2 cells showed strong expression of all three IFN target genes. In LS88 cells, none of the IFN target genes were induced, indicating that the observed changes in protein expression were not due to an IFN-mediated stress response (Figure 3). To validate our mass spectrometry results, we obtained antibodies against a subset of the identified proteins and analyzed their expression levels after knockdown of USP7 by Western blotting (Figure 4 A-C). We confirmed reduced expression of one of the two isoforms of glycogen phosphorylase B upon knockdown of USP7 (Figure 4A). The most striking difference in protein levels before and after USP7 knockdown was observed for Alix/HP95. This protein is involved in endosomal trafficking,24,25 virus budding,26 and apoptosis27 and showed reduced expression upon USP7 removal with delayed kinetics as compared to the downregulation of USP7 itself (Figure 4A). Reduced levels of the DNA repair proteins Ku70, Ku80 (Figure 4B), and TRIM28, PDI, and MCM6 (data not shown) could not be confirmed by Western blotting, indicating that the abundance of these proteins is not dependent on USP7 protein expression. However, further analysis of one of these candidates, Ku70, by 2D gel electrophoresis followed by Western blotting revealed a shift in two additional polypeptides present in the same molecular weight range, suggesting that posttranslational modifications may occur when USP7 is down-regulated (Figure 4C). Despite the fact that we could not detect an IFN response in LS88 cells, we sought to exclude any indirect consequences of G1 arrest, a possible weak IFN response or other side effects of the induction of shRNA oligo expression as a cause for changes in protein levels observed, in particular Alix/HP95 down-regulation. To this end, we used the Myc9.2 clone in which shRNA (specific for Myc) induction induces a strong G1 arrest as well (data not shown). Alix protein levels remained
HAUSP/USP7-Dependent Changes in the Proteome
research articles
Figure 2. Altered proteome when USP7 is down-regulated. LS88 cells were treated or not with 1 µg/m doxycycline for 48 h to downregulate USP7. Cell extracts were prepared and separated by 2D-SDS-PAGE, and proteins were visualized by silver staining. One representative out of two independent experiments is shown. Comparison of extracts from untreated cells (A) and cells where USP7 was down-regulated (B) revealed spots that were present at different intensities (labeled in circles). Twenty-five spots per condition (50 total, corresponding spots are labeled with A and B, respectively) were excised, subjected to in-gel trypsin digestion, and identified by mass spectrometry, revealing 36 different proteins. (C) Major fraction of identified proteins is involved in transcription/translation, as exemplified in this region of the 2D gel map between 70 and 200 kDa; see text for details. Journal of Proteome Research • Vol. 6, No. 11, 2007 4165
research articles
Kessler et al.
Table 1. Proteins with Altered Expression upon USP7 Removal Identified by 2D-SDS-PAGE Analysis and Mass Spectrometry USP7 RNAi/Ctrl 2D spot ratiob
protein no.
2D spot no.a
1
1B
>10
Filamin
P21333
280 250 5.7
2
2A
