D. P. Singh, Department of Ophthalmology and Visual Sciences, University of Nebraska Medical Center, Omaha, NE 68198-5840, USA Fax: +1 402 559 8808 Tel: +1 402 559 8805 E-mail: email@example.com
Lens epithelium-derived growth factor (LEDGF), a ubiquitously expressed nuclear protein, acts by interacting with DNA and protein and is involved in widely varying cellular functions. Despite its importance, the mechanism(s) that regulate naturally occurring LEDGF activity are unidentified. In the present study, we report that LEDGF is constitutively Sumoylated, and that the dynamical regulatory mechanism(s) (i.e. Sumoylation and deSumoylation) act as a molecular switch in modulating the DNA-binding and transcriptional activity of LEDGF with the functional consequences. Using bioinformatics analysis coupled with in vitro and in vivo Sumoylation assays, we found that lysine (K) 364 of LEDGF was Sumoylated, repressing its transcriptional activity. Conversely, mutation of K364 to arginine (R) or deSumoylation by small ubiquitin-like modifier (Sumo)-specific protease-1, a nuclear deSumoylase, enhanced the transactivation capacity of LEDGF and its cellular abundance. The enhancements were directly correlated with an increase in the DNA-binding activity and small heat shock protein transcription of LEDGF, whereas the process was reversed in cells overexpressing Sumo1. Interestingly, cells expressing Sumoylation-deficient pEGFP-K364R protein showed increased cellular survival compared to wild-type LEDGF protein. The findings provide insights into the regulation and regulatory functions of LEDGF in Sumoylation-dependent transcriptional control that may be essential for modifying the physiology of cells to maintain cellular homeostasis. These studies also provide new evidence of the important role of post-translational modification in controlling LEDGF function.
Lens epithelium derived growth factor (LEDGF) is an ubiquitous and stress-inducible transcriptional nuclear protein that has important functions in regulating cell proliferation, differentiation and apoptosis [1–3]. In addition, the involvement of aberrant expression of LEDGF with cancer or progression of cancer has been documented [4,5]. The critical importance of LEDGF in cell biology was shown when genetic ablation studies revealed that a deficiency of LEDGF may lead to embryonic lethality or perinatal death, whereas some survivors have many abnormalities . LEDGF is identical to p75 and was originally identified as a transcriptional coactivator, as well as a transactivator of stress response genes, thereby providing cytoprotection against various stressors [3,7]. LEDGF belongs to the hepatoma-derived growth factor family of proteins . Its gene yields two proteins by alternate splicing: the coactivator p52 (exons 1–9 plus part of intron 9) and LEDGF (exons 1–15) [2,9]. Both are karyophilic proteins, although they differ in their nuclear localization patterns. Structurally, LEDGF contains several functional domains, which are involved in protein–protein and protein–DNA interactions and are responsible for various activities of LEDGF [2,10,11]. Studies have shown that LEDGF is tightly bound to chromatin during all phases of the cell cycle, and this binding is mediated by the functional interaction of the PWWP domain, the nuclear localization signal and two AT hook motifs located in the N-terminal region. A stretch of 58 amino acids, comprising the PWWP domain bound to stress response elements (STREs; A/TGGGGA/T) and C-terminal LEDGF with two helix-turn-helix-like domains, has been found to bind to a heat shock element (HSE; nGAAn) of stress-related genes . A transactivation assay using HSP27 promoter revealed that both helix-turn-helix domains contribute in a cooperative manner to the transactivation potential of LEDGF. The N-terminal domain is involved in stabilizing the LEDGF–DNA-binding complex. These studies highlight the importance of LEDGF with respect to controlling many cellular functions that may depend upon the cellular microenvironment and thus modifying the physiology of cells to maintain cellular homeostasis.
Recently, the role of LEDGF in modulating diverse cellular functions has been documented. It is involved in HIV-1 integration and the prevention of proteosomal degradation of HIV-1 integrase . LEDGF/p75 (but not p52) was identified as the prominent interaction partner of HIV-1 integrase [14,15]. This interaction of HIV-1 integrase with the C-terminal integrase-binding domain (IBD) of LEDGF/p75 is crucial for HIV-1 replication . The protein is also involved in autoimmune disorders and cancer through the formation of chimeric protein with NUP98 [5,16]. A high expression level has been reported in prostate tumours . We have shown that the expression of LEDGF is developmentally regulated and also that the level of expression influences cellular fate [17,18]. LEDGF plays a role in lens epithelial to fibre cell terminal differentiation and, more recently, the presence of this molecule in discrete regions and cell types within the fetal and adult brain suggests that it is involved in neuro-epithelium stem cell differentiation and neurogenesis . All these functions collectively highlight the wide range of roles played by this molecule, ranging from cellular protection to an association with cellular abnormalities. The mechanism by which LEDGF is involved in various cellular events and the specific ways in which it plays its diverse roles need to be investigated. We predict that the functions of LEDGF may be associated with its modifications, such as phosphorylation and/or Sumoylation, and that these modifications may be responsible for its diverse cellular functions. Recently, it was demonstrated that LEDGF is a target of the small ubiquitin-like modifier (Sumo), and that Sumoylation of LEDGF leads to the repression of small heat shock protein (HSP) transcription . Even so, it remains unclear how naturally occurring LEDGF functions within the transcriptional regulatory programmes that govern and regulate the transactivation of stress response genes. Several possibilities exist: LEDGF may repress transcription by interacting with corepressors after Sumoylation, or LEDGF Sumoylation may decrease its DNA-binding activity and definably contribute to the dynamic regulation of transcription that occurs in response to signals of cellular status. However, the mechanisms involved and the molecular components that mediate such a dynamic function of LEDGF or, at the same time, the transcriptional status of LEDGF have not been identified, nor have the functional consequences of LEDGF Sumoylation and deSumoylation on cell fate.
Sumo has been shown to regulate cellular processes by controlling the localization, function and expression, as well as stability, of large numbers of cellular proteins . The complexity of LEDGF gene regulation and function was explained recently by ectopically expressing LEDGF post-translational modification (i.e. Sumoylation). However, the fate of naturally occurring LEDGF Sumoylation is still not known, nor how Sumoylated LEDGF affects its own function(s) under normal physiological conditions, as well as in cells facing stress or during ageing . Sumoylation is reversible, and the removal of Sumo (i.e. deSumoylation) is catalyzed by Sumo-specific proteases (Senps). Lysines (K) are major sites of protein modification . Sumo modification of transcriptional protein is an important mechanism for achieving the dynamic regulation of gene expression. However, most Sumoylated proteins have been shown to repress gene transcription [24–28]. In a reversible post-transcriptional modification, Sumo(s) are covalently linked to lysine residues of the target proteins . Sumo is an 11.5-kDa ubiquitin-related protein and a close relative of ubiquitin [30,31]. Although Sumoylation is enzymatically similar to ubiquitinization, the two require different sets of enzymes. The Sumo-activating enzymes, SAE1 and SAE2, activate Sumo in an ATP-dependent manner. Activated Sumo is then transferred to the Sumo-conjugating enzyme Ubc9, which mediates the conjugation of Sumo to an exposed lysine residue in the target protein. Sumoylation is enhanced by Sumo ligases, comprising a diverse group of proteins that stabilize and direct the interaction between Ubc9 and its Sumoylation targets. Upon conjugation, Sumo can be efficiently removed from its targets by Sumo proteases, resulting in very low steady-state levels of the Sumo-modified forms for most Sumo targets [32,33]. A long list of transcriptional factors includes HSF1 and HSF2, which are modified by Sumo1. However, in contrast to HSF2, the HSF1 protein is not constitutively modified by Sumo1 and, instead, is only modified after cells are exposed to stress conditions [34,35].
Sumoylation can be readily reversed by a family of Senps. Senp-1 is a nuclear protease that appears to deconjugate a large number of Sumoylated proteins . Senp-2 is a nuclear-envelope-associated protease that has activity similar to Senp-1 [37,38]. Two additional Sumo-specific proteases, Senp-3/SMT3IP1 and Senp-6/SUSP1, have also been reported [39,40]. Even though the ability of Senps to reverse Sumoylation is well established, the specificity of each Senp and the difference in each regulatory pathway mediated by these Senps remain to be defined. We found that LEDGF transcription is markedly enhanced by Senp-1. Sumoylation has been linked to transcription repression in an increasing number of Sumoylated transcription factors or cofactors . Several lines of evidence indicate the ability of LEDGF to bind to many proteins [10,11,42], including Sumo1 . It appears that the Sumo motif (K364) present within the IBD (with the highest probability score) in the C-terminal of LEDGF protein is crucial for its constitutive Sumoylation and function(s), as identified by computational prediction software: sumoplot (Abgent, San Diego, CA, USA). At present, however, the influence of endogenous LEDGF Sumoylation/deSumoylation upon LEDGF activity is in infancy. It is also unclear whether Senp-1 or Sumo alters DNA-binding activity and function in cells under normal physiological conditions or in cells facing stress. Also unknown is whether Sumoylaion or deSumoylation influences the expression levels of LEDGF because these factors may be involved in downstream signalling by modulation of transcription.
Using several biochemical approaches, in the present study, we demonstrate that LEDGF is Sumoylated in vitro as well as in vivo, and that K364 present in the C-terminal domain (within IBD) is critical for the constitutive Sumoylation of LEDGF and its regulatory functions. We also show that the transcriptional activity of LEDGF is altered as a consequence of Sumo1 conjugation, and conversely, deSumoylation by Senp-1 or disruption of the Sumo1 Motif, lysine (K) residue to arginine (R), impairs the Sumoylation process and promotes the transactivation capacity of LEDGF (pEGFP-K364R), which is accompanied by increased DNA binding and activation of stress response HSPs and cell growth. Our data reveal a regulatory mechanism for naturally occurring LEDGF mediated by the transcriptional modulation of its own and target genes via a dynamic process of Sumoylation and deSumoylation, which may have implications in cellular survival when faced with stress and human disorders.
LEDGF is present in Sumoylated form and colocalizes with Sumo1 in human lens epithelial cells (hLECs)
LEDGF is both a coactivator and transactivator, and most of the transcriptional proteins have been shown to be Sumoylated . Considering the broad range of substrates in which the Sumoylation pathway is involved, it would not be surprising if LEDGF was a target of Sumo modification. Recently, it was shown that exogenous LEDGF is polySumoylated . Our purpose was to observe whether endogenous LEDGF is Sumoylated. Western blotting of LEDGF expression reveals two forms (duplet) that are detected only by polyclonal antibodies (Lifespan Biosciences, Seattle, WA, USA, or Santa Cruz Biotechnology, Santa Cruz, CA, USA). To test whether either of these forms is yielded by Sumo1 conjugation of LEDGF, we examined the possible presence of Sumo1-modified LEDGF forms by using a membrane that had been previously probed with polyclonal anti-LEDGF serum, and restriping and reprobing it with Sumo1 antibody (Santa Cruz Biotechnology). Of the two LEDGF-antibody-reacting bands (Fig. 1A, lane 1, upper and lower bands), only the upper band of ∼ 87 kDa was recognized by Sumo1 antibody (Fig. 1A, lane 2, upper band, a). Additionally, we examined mouse (m)LECs (Fig. 1B) to determine whether the LEDGF is in Sumoylated form as in hLECs. Interestingly, similar results (two bands) were obtained with these cells, suggesting that LEDGF of 75 kDa was modified with Sumo1 [29,44], giving rise to a band of ∼ 87 kDa (Fig. 1B). Moreover, the mobility of the shifted band was dependent upon the percentage of SDS gel (7.5–12%) used to resolve the protein.
Furthermore, LEDGF is predominately localized in the nucleus [9,12,45–47], and Sumo1 is known to be localized and exerts its genetically defined activity in the nucleus . Because our western blot data indicated that a fraction of naturally occurring LEDGF was constitutively Sumoylated, we next examined whether LEDGF and Sumo1 are colocalized. We performed immunocytochemistry using Sumo1 or LEDGF antibody, respectively. Immunofluorescence analysis determined that both molecules colocalized in the nucleus. Merged images of Sumo1 (Green) and LEDGF (Red) yielding a yellow colour or granules revealed that a portion of total LEDGF protein interacted with Sumo1 (Fig. 1C), suggesting that only a certain amount of LEDGF is Sumoylated.
Aiming to determine whether endogenous LEDGF is indeed Sumoylated, we performed an immunoprecipitation assay using antibody specific to LEDGF or IgG antibody (control) or Sumo1 antibody. Accordingly, hLECs were transiently transfected with enhanced green fluorescent protein (EGFP)-Sumo1. Forty-eight hours later, nuclear extracts were processed for immunoprecipitation and immunoblotted as described in the Materials and methods. Cells transfected with EGFP-empty vector were used as a control (Fig. 1D lanes 1, 3 and 5). The results obtained revealed the presence of three bands with polyclonal anti-LEDGF serum (Fig. 1D, a, lane 2: unSumoylated endogenous LEDGF, ∼ 75 kDa; endogenous Sumoylated LEDGF, ∼ 87 kDa; and endogenous LEDGF Sumoylated by exogenous EGFP-Sumo1, ∼ 115 kDa). The same membrane was probed with anti-Sumo1 that gave rise two bands (Fig. 1D, a, lane 4: endogenously Sumoylated LEDGF, ∼ 87 kDa and endogenous LEDGF Sumoylated with exogenous EGFP-Sumo1, ∼ 115 kDa), whereas only one band could be detected when the same membrane was probed with anti-GFP serum (Fig. 1D, a, lane 6: endogenous LEDGF Sumoylated with EGFP-Sumo1). To further validate the Sumoylation of endogenous LEDGF, nuclear extracts were immunoprecipitated with anti-Sumo1 serum and immunoblotted with anti-LEDGF serum (Fig. 1E, lane 1) or anti-Sumo1 serum. Cells transfected with EGFP-empty vector were taken as a control (Fig. 1E lanes 1 and 3). The data revealed the presence of two bands: Sumoylated LEDGF (∼ 87 kDa) and endogenous LEDGF Sumoylated with exogenous EGFP-Sumo1 (∼ 115 kDa). By contrast, nuclear extract immnoprecipitated with IgG antibody (control) did not produce any detectable specific band with either of the specific antibodies, validating specificity of experiments (Fig. 1D, b). Collectively, these results indicate that a fraction of LEDGF is present in mono-Sumoylated form. Taken together, endongenous LEDGF was Sumoylated with endogenous Sumo1, as well as exogenously expressed EGFP-Sumo1, suggesting that LEDGF is a substrate for Sumo1.
LEDGF is Sumoylated both in vitro and in vivo, and conserved K364 is subject to Sumoylation
Ubc9 is an E2 conjugation enzyme for protein Sumoylation, and interaction with UBC9 is a feature of many proteins subsequently modified by Sumo1. To determine whether this occurs in the LEDGF–Sumo1 conjugation, we performed an in vitro Sumo1 conjugation assay. We prepared constructs expressing glutathione S-transferase (GST)-LEDGF and GST-Sumo1 as described in the Materials and methods, and verified the expression of these two via western blotting (data not shown). Purified GST-LEDGF protein was incubated with purified Sumo1, activating enzyme E1 (SAE1/2) and conjugating enzyme E2 (Ubc9) in a buffer containing an ATP-generating system. As shown in Fig. 2A, in the presence of enzymes, a higher molecular mass was recognized by antibodies specific to GST (Fig. 2A, left, lane 1), Sumo1 (Fig. 2A, middle, lane 1) or LEDGF (Fig. 2A, right, lane 1) in the western blot, confirming the identity of a slowly migrating Sumoylated LEDGF band. The data indicate that LEDGF is substrate for Sumo1, and that Ubc9 is essential as an interacting partner for the Sumoylation process. By contrast, purified GST protein did not show any interactions, and so we could not detect such a band (data not shown). HSF1 is known to be Sumoylated. In a parallel experiment, we also used HSF1 Sumoylation experiment as a control to validate the experiments (Fig. 2B).
Next, we determined the critical region or Sumo1 conjugation site(s) of LEDGF protein using sumoplot prediction, a Web-based tool for predicting consensus Sumoylation (Abgent). Sequence analysis predicted six Sumo1 conjugation sites (Fig. 3A); however, the highest score was predicted with K364 (LKID) present at the C-terminal of LEDGF. To accurately identify the Sumoylation site(s) of LEDGF, we generated various deletion constructs of GST-LEDGF expression plasmids containing predicted sequences and we also utilized previously prepared LEDGF constructs for expressing prokaryotic vectors. The in vitro Sumoylation assay showed that C-terminal LEDGF ranging from 170 to 530 amino acids is indeed Sumoylated (Fig. 3B, lane 3), whereas N-terminal LEDGF (amino acids 1–250) was not, indicating that the C-terminal LEDGF bore a Sumo1 conjugation motif.
Furthermore, closer analysis of the predicted sequence revealed that the C-terminal of LEDGF contained only one Sumo1 conjugation motif with the highest score (sumoplot, score < 90), whereas the other two had the lowest scores. Additionally, the in vitro Sumoylation assay and sumoplot prediction supported the fact that K364 (LKID) can be an actual site for LEDGF Sumoylation, which is also evolutionarily well conserved (Fig. 4). Based on these findings, and to determine whether lysine within the IBD  is required for the Sumoylation of LEDGF, we mutated lysine (K) to arginine (R), generating LEDGF mutant (K364R) and performed an in vivo Sumoylation experiment as described in the Materials and methods. The coexpression of EGFP-Sumo1 and lysine mutated forms of EGFP-LEDGF (K364R) showed that EGFP-LEDGF (K364R) was not Sumoylated, and thus could not be detected by anti-Sumo1 serum (Fig. 5A, right, lane 3). Conversely, wild-type EGFP-LEDGF or endogenous LEDGF was Sumoylated, and therefore migrated more slowly on gel, yielding bands: EGFP-LEDGF plus EGFP-Sumo1 (Fig. 5A, ∼ 143 kDa, lane 2); endogenous LEDGF plus EGFP-Sumo1 (Fig. 5A, ∼ 115 kDa, lanes 1–3); and endogenous LEDGF plus endogenous Sumo1 (∼ 87 kDa, lanes 1–3). Next, the same blotted membrane was probed with anti-LEDGF (Fig. 5B) or anti-GFP (Fig. 5C) sera to determine whether the Sumo1 recognized band is endogenous/exogenous LEDGF. The results revealed that anti-LEDGF (Fig. 5B) or anti-GFP (Fig. 5C) sera detected the same protein bands as recognized by anti-Sumo1 serum (Fig. 5A). Thus, our data reveal that the disruption of K364R attenuated LEDGF Sumoylation (Fig. 5A, lane 3), and show that K364 was an active and effective Sumoylation site in LEDGF, although residual Sumoylation may have occurred at the lysine residues outside the IBD domain of LEDGF, although we were unable to detect them in our system .
To test whether Sumoylation/deSumoylation alters LEDGF localization, a nuclear protein, we transfected cells with pEGFP-K364R (mutant LEDGF) and wild-type pEGFP-LEDGF. Disruption of the Sumo1 motif did not affect the LEDGF localization pattern, which was predominantly localized in the nucleus as wild-type LEDGF (Fig. 5D). Next, we tested the identity and expression levels of LEDGF and mutant LEDGF EGFP-K364R by western blotting (Fig. 5E). We found that protein bands of mutant LEDGF EGFP-K364R and wild-type EGFP-LEDGF (Fig. 5E, lane 2 versus 3, upper band) were of the correct molecular mass (∼ 105 kDa), suggesting that protein integrity is maintained. In addition, this was consistent with the findings of a recent study showing that protein Sumoylation may lead to protein stabilization or vice versa . Our results also revealed that EGFP-LEDGF (K364R) is more stable than wild-type EGFP-LEDGF. However, our results are consistent with earlier studies showing that Sumoylation destabilizes LEDGF proteins and decreases their half-life .
Sumoylation/deSumoylation, a dynamical process, affects the transactivation capacity of LEDGF and regulates transcription
Given the roles of Sumoylation and deSumoylation of proteins in the modulation of transcriptional activity, and in vitro and in vivo Sumoylation of LEDGF (Figs 1–5), as well as the findings of previous studies , we first examined whether Sumo1/Senp-1 alters the endogenous expression of the LEDGF and its target gene, HSP27. Cellular extracts isolated from Sumo1- or Senp-1-transfected cells were analyzed by western blotting using antibodies specific to Sumo1, Senp-1 or LEDGF (monoclonal antibody), or HSP27. We found that Sumo1 down-regulated the expression of LEDGF (Fig. 6A, b) and HSP27 (Fig. 6A, c), and that the decrease of HSP27 was dependent upon the concentration of Sumo1 (Fig. 6A, a), whereas Senp-1 increased the expression of LEDGF (Fig. 6B, b) and HSP27 (Fig. 6B, c) proteins in a dose-dependent fashion. Our data indicated that the decreased expression of HSP27 in cells was associated with the increased Sumoylation of LEDGF.
Next, we performed a series of cotransfection transactivation assays using mutants of LEDGF pEGFP-K364R at the Sumo1 conjugation site (Fig. 6C), pEGFP-Sumo1 or pFlag-Senp-1 (Fig. 6D). The specificity of LEDGF-dependent activity was evaluated by using small interfering RNA (siRNA) specific to LEDGF (Fig. 6E). First, we compared the transactivation ability of wild-type LEDGF with that of Sumoylation-deficient mutant (LEDGF) pEGFP-K364R to activate the small HSP genes: αB-crystallin and HSP27. LEDGF binds to HSE (nGAAn) in HSP27 and αB-crystallin promoters to up-regulate their transcription [2,3,47]. LECs were cotransfected either with pEGFP-vector, pEGFP-LEDGF (wild-type) or pEGFP-K364R (mutant) plasmid along with reporter plasmid HSP27 or αB-crystallin promoters linked to chloramphenicol acetyltransferase (CAT) vector, and cell extracts were analyzed with CAT-ELISA. Both forms of LEDGF activated expression of the gene for CAT above the basal level (Fig. 6C, light grey bar). However, the transactivation potential of pEGFP-K364R in HSP27 or αB-crystallin promoter activity was significantly greater than that of pEGFP-LEDGF (Fig. 6C, black bar). Because single mutation disrupts the Sumoylation of LEDGF, our data indicated that the covalent modification of LEDGF by Sumo1 transrepressed the transactivation capacity of LEDGF. Collectively, the results revealed an increase in the transactivation potential of LEDGF when lysine (K) 364 was changed to arginine (R).
Sumoylation and deSumoylation are reversible processes , and the removal of Sumo1 (i.e. deSumoylation) is catalyzed by Sumo-specific proteases. Senp-1 is most abundantly engaged in the deSumoylation of proteins . However, it was not clear whether the cellular expression levels of Senp-1 or Sumo1 influenced LEDGF transcriptional activity. Thus, to identify the critical link among Sumoylation (EGFP-Sumo1), deSumoylation (Senp-1) and LEDGF (by applying siRNA of LEDGF) in vivo, we utilized transactivation assays by cotransfecting cells with the plasmid(s) of the above-mentioned reagents and reporter plasmids: pCAT-HSP27. Promoter activity was dramatically increased (Fig. 6D, left, black bar), and the activity was dependent specifically upon LEDGF, as indicated by LEDGF-specific interference (Fig. 6D, right, black bar). By contrast, reduced activity was observed in cells cotransfected with pEGFP-Sumo1 (Fig. 6D, light grey bars versus grey bars). In addition, western blotting revealed that the LEDGF siRNA used for the assay effectively reduced the expression of LEDGF protein (Fig. 6E, lane 2) and that reduced promoter activity was correlated with the reduced expression of LEDGF protein.
These data imply that Sumoylation and deSumoylation of LEDGF affects its transcriptional efficacy. Moreover, we also recognized that Sumo1 modification of LEDGF may change its conformation, making LEDGF less accessible to the nucleus. This possibility was ruled out when the transfection assay revealed the nuclear localization of mutant LEDGF- K364R, and that the localization pattern was indistinguishable from that of LEDGF (Fig. 5D). Collectively, the results provide evidence that Sumoylation/deSumoylation effectively influenced the regulatory activity of LEDGF, in turn affecting the expression of its target genes.
Sumoylation diminished the DNA-binding activity of LEDGF, suggesting that reduced transactivation activity was dependent upon LEDGF–DNA interaction
The transactivation experiments did not reflect whether modulation in the transactivation activity of LEDGF was associated with its DNA binding. Recent evidence revealed that Sumo1 conjugation may alter the DNA-binding activity of proteins [51,52]. To determine whether Sumoylation affected the DNA-binding activity of LEDGF, we carried out a gel-shift mobility assay. Nuclear extracts from cells transfected with pFlag-Senp-1, pEGFP-Sumo1 or empty vector were incubated with radiolabelled probes containing HSE or STRE and processed for a gel-shift assay to examine the influence of the Sumoylation of LEDGF on its DNA-binding activity (Fig. 7A). The binding of LEDGF in nuclear extract from Sumo1-transfected cells to the probe was reduced significantly (Fig. 7A, lane 2) compared to vector-transfected cells (Fig. 7A, lane 1). Conversely, LEDGF in nuclear extracts of Senp-1-transfected cells bound with a greater affinity to the same probe and formed a complex, Cm1 (Fig. 7A, lane 3). These data demonstrate that Senp-1 increased the LEDGF interaction with its binding sites. Previously, we had shown that LEDGF binds to HSE (nGAAn) and STRE and activates the transcription of stress-associated genes [3,17,47].
Next, we examined whether the increased binding of LEDGF was associated with a greater/lesser abundance in cells or with its modulated affinity to DNA binding as a result of Sumoylation. Initially, levels of LEDGF in cells transfected with Senp-1, Sumo1 or empty vector were subjected to western blotting using monoclonal anti-LEDGF serum. Cell extracts from pFlag-Senp-1-transfected cells contained increased levels of LEDGF compared to cells transfected with Sumo1 or vector (Fig. 6A,B). This indicated that Senp-1-mediated abundance of deSumoylated LEDGF may be at least one cause of increased DNA binding (Fig. 6B, b). Sumoylation and deSumoylation have been shown to influence protein stability and integrity [53,54], and deSumoylated LEDGF protein is known to have greater stability because it is less susceptible to degradation than Sumoylated LEDGF . Moreover, LEDGF is an inducible and constitutively transcriptionally active gene, and Sumo acts as a controller for constitutive transcription and also during the activation of inducible genes . The presence of Sumo1 at transcriptionally active gene LEDGF suggests that Sumo1 plays a role in regulating LEDGF transcription [41,56–58].
Interestingly, when evaluating the expression level of LEDGF in cells overexpressing Senp-1, we observed that the expression of LEDGF increased with an increase in Senp-1 concentration (Fig. 6). We speculated three possible explanations: (a) DeSumoylation of LEDGF increased its half-life by slowing its degradation processing by ubiquitization pathway(s) opposed to Sumo1 because Sumo1 conjugation of LEDGF decreased LEDGF half-life or stability ; (b) Senp-1 enhanced LEDGF transcription by enhancing the transactivator of LEDGF, thereby increasing the expression of LEDGF in cells; (c) and Senp-1 enhanced the abundance of LEDGF expression via both mechanisms; however, whether Sumoylation increased LEDGF degradation pathways is known . Therefore, we proceeded to explore whether mRNA of LEDGF was increased in cells harboring higher level of Senp-1.
Sumo-specific protease, Senp-1 enhanced expression of LEDGF mRNA by promoting LEDGF gene transcription in LECs
Endopetidase Senp-1 has been shown to remove Sumo from conjugated protein substrate including transcriptional proteins and to modulate their transcriptional activity, which involves Sumo conjugation in the repression of gene transcription . However, in the present study (Fig. 6), we observed that the level of LEDGF protein was elevated in cells after increased expression of Senp-1, and that the increased expression of LEDGF was one factor responsible for increased HSP expression. Assuming that the expression levels of Senp-1 may increase LEDGF expression by promoting LEDGF gene transcription, we determined first whether Senp-1 expression could regulate the endogeneous expression of LEDGF mRNA in LECs. By real-time PCR, we found that cells overexpressing Senp-1 displayed a greater abundance of LEDGF mRNA (Fig. 8A, black bar) than cells with vector or Sumo1 plasmids. Sumo1-transfected cells displayed a reduced expression of LEDGF mRNA, indicating that Sumo1 repressed the transcription of LEDGF as opposed to Senp-1 (Fig. 8A, open bar versus grey bar versus black bar). Because the modulation of LEDGF expression occurred at the mRNA level, we postulated that Sumoylation/deSumoylation may occur in the transcriptional protein responsible for transcriptional activation of LEDGF. To determine whether this was the case, we generated truncated constructs of LEDGF promoter linked to CAT vector, as described previously . Cells were cotransfected with a series of 5′-deletion mutant constructs of the LEDGF gene promoter linked to CAT containing a common 3′ end or their wild-type and pFlag-Senp-1, as shown in Fig. 8B and reported previously . Cellular extracts of transfectant(s) were analyzed for promoter activity by monitoring CAT reporter protein as described in the Materials and methods. Data analysis showed that Senp-1 dramatically increased LEDGF gene promoter activity (black bars versus grey bar), and the stimulation in promoter activity was Senp-1 dependent (Fig. 8B, black bar). Significantly increased promoter activity was observed with reporter plasmid containing −170/+35 and −127/+35 bp promoter regions, whereas the lowest promoter activity was shown by deletion mutant plasmid with −28/+35, suggesting that transcriptional protein-binding response element(s), which are plausibly involved and activated by Senp-1 in transactivating LEDGF, can be present between nucleotides −170 and −28 (Fig. 8B).
LEDGF gene promoter region spanning from −170 to −28 contained critical responsive element(s) for transcription factor sensitive to Senp-1 and/or Sumo1 modification
To determine whether the transcription factor-binding region of 5′-promoter between nucleotides −170/−28 is indeed responsive to Senp-1 and/or Sumo1-dependent regulation, we cotransfected cells with reporter plasmid containing LEDGF promoter spanning −170 to +35 linked to CAT and pFlag-Senp-1 (0.01, 0.05, 0.15, 0.5, 2 μg) or Sumo1, or control vector(s). Titration of Senp-1 showed a dose-dependent enhanced transcription of LEDGF (Fig. 8C). The transcriptional activity of LEDGF was at its peak (∼ 40-fold) at a concentration of 2 μg of Senp-1 (Fig. 8C, grey bar). Even at a very low concentration (150 ng), Senp-1 induced the transcription of LEDGF significantly compared to the control (black bars). This indicates the transcription of LEDGF by Senp-1, and showed that the transcription factors (activator/enhancer) were a target of Sumo1 modification.
Sumo proteins are known to influence the activities of other proteins by direct conjugation. We found that Senp-1 up-regulated LEDGF transcription, and so it is possible that Sumo conjugation to some factor(s) comprises the mechanism underlying the action of Sumo1 on LEDGF transcription. To test whether Sumo1 repressed the transcription of LEDGF, we transfected cells with pEGFP-Sumo1, pFlag-Senp-1 and empty vector, along with reporter plasmid, pCAT-LEDGF (−170/+35), and performed transactivation assays. Enforced expression of Sumo1 led to a decrease in the promoter activity of LEDGF (Fig. 8D, black bar versus grey bar). On the other hand, Senp-1 increased the promoter activity significantly (Fig. 8D, light grey bar versus dark grey or black bar). Collectively, our data demonstrated that LEDGF expression was regulated by dynamical processes of Sumoylation and deSumoylation. We propose that the transcriptional regulator of LEDGF is a target of such a process, and that the process controls LEDGF transcription in accordance with cellular needs and background.
Senp-1 potentiated the Sp1 activation of LEDGF transcription
To further investigate the potential cis-regulatory elements of LEDGF promoter involved in the transcriptional control of the expression of LEDGF, we analyzed 5′-LEDGF promoter sequences (within −170 to −28)  using Web-based computer software for predicting putative transcription factor-binding sites (matinspector, Genomatix, Munich Germany). Our recent finding [59a], when coupled with the present analysis of the LEDGF promoter, convinced us that three putative Sp1 sites in the LEDGF promoter positioned at −50/−43 (Sp1-3), −109/−102 (Sp1-2) and −146/−139 (Sp1-1) were responsible for transcription. Most importantly, it was shown recently that Sp1 is Sumoylated by Sumo1, and that Sumoylation attenuates Sp1-dependent transcription . Conversely, Senp-1 was found to stabilize Sp1 activity. Under this scenario, we predicted that modulation in the transcription of LEDGF by Senp-1 and Sumo1 should be associated with the regulation of Sp1 by these molecules. To determine whether the regulation of LEDGF is indeed under the control of the regulatory activity of Sp1, we performed cotransfection transactivation experiments using wild-type LEDGF promoter (−170/+35) or its mutant at all Sp1 sites linked to the reporter gene for CAT along with the pFlag-Senp-1 plasmid at variable concentrations (Fig. 9). Interestingly, cellular extract from Senp-1-transfected cells showed higher LEDGF promoter activity, and promoter activity increased with an increase of Senp-1 concentration (Fig. 9, black bar). However, the mutant promoter showed some activity to Senp-1, although this was significantly lower than that of wild-type promoter (Fig. 9, right, light grey bar versus black bar). We consider that the promoter activity in response to Senp-1 should be associated with other cofactors/factor that may be a target for the Sumoylation/deSumoylation process. Taken together, our results revealed that the expression and function of LEDGF is regulated at both post-translational and transcriptional levels.
Sumoylation and deSumoylation of LEDGF affected growth and survival of cells at 37 °C, as well as during heat stress
Sumoylation and deSumoylation are dynamic processes, and either the activation or inhibition of those processes would have a profound effect on subsequent cellular events. However, LEDGF expression and stability in response to stressors is imperative for maintaining its ability to coordinate the stress response and protect cells, and it acts by up-regulating stress response genes [3,22,47,49,59,61,62]. Under conditions of stress, the present study showed that the increased expression and stability of LEDGF appeared to be attributable, in part, to its transcriptional and post-transcriptional control by Sumoylation/deSumoylation. We next examined whether the protective activity of LEDGF against environmental stressors was altered. Cells transfected with wild-type LEDGF, mutant LEDGF pEGFP-K364R or pEGFP-vector were submitted to heat stress [47,63]. Interestingly, cells overexpressing mutant LEDGF pEGFP-K364R at the Sumo1 conjugation site showed better growth than wild-type LEDGF under normal physiological conditions (Fig. 10, 37 °C; grey bars). By contrast, cells with mutant LEDGF pEGFP-K364R did not have resistance against heat stress compared to wild-type LEDGF-transfected cells (Fig. 10, 43 °C, black bars). However, we could not explain how the protective mechanism of LEDGF and mutant LEDGF-K364R became different in cells facing stress versus cells under normal physiological conditions. Next, our interest was to examine the effect of Senp-1 or Sumo1 expression on cell survival. Cells were overexpressed with either pEGFP-Sumo1 or pFlag-Senp-1, and were subjected to heat stress. Increased cell growth was seen in cells overexpressing Senp-1 compared to cells overexpressing Sumo1 under normal physiological conditions (Fig. 10B, grey bars). However, the phenomenon was reversed when cells were subjected to heat stress (Fig. 10B, 43 °C, black bars), and the survival of cells overexpressing wild-type LEDGF was increased. The results of both experiments (Fig. 10A) were similar, and we consider that the results shown in Fig. 10A define the specific and selective functions of Sumo1 or Senp-1 with respect to LEDGF, whereas the results in Fig. 10B indicate the broader effects of Sumoylation and deSumoylation, and may also include the regulatory activity of Sp1 and Sp1-mediated regulation of LEDGF transcription. Taken together, these data suggest that Sumoylation and deSumoylation play a critical role in cellular integrity and the expression of LEDGF, one that defines the function of LEDGF depending upon the cellular microenvironment and cellular needs.
LEDGF belongs to the HDGF (hepatoma-derived growth factor) family of proteins that consists of a well-conserved N-terminal amino acid sequence called the HATH (homologous to amino terminus of HDGF) region [9,64,65]. It is stress-inducible and acts as a transcriptional survival factor and coactivator [3,22,59]. LEDGF plays a pivotal role in cellular survival, and is known to be a preventive factor in eye disorders [7,22,61,66]. Recently, elevated expression of LEDGF has been reported in cancer cells  and it has been implicated in carcinogenesis and cancer progression. The involvement of LEDGF has been reported in HIV-1 integration , autoimmune disorders [5,67–70] and cancer. It acts by forming a chimeric protein with NUP98 . In addition, LEDGF plays a role in lens epithelial-to-fibre cell terminal differentiation . These studies highlight the wide range of roles played by LEDGF, ranging from cellular protection to the promotion of cellular abnormalities. However, the mechanisms by which it is involved in various cellular events and how it specifically plays its various roles remain the subject of active investigation. In the present study, we attempted to identify certain regulatory roles, and to show that at least some functions of LEDGF are associated with its post-translational modification (i.e. Sumoylation/deSumoylation), which regulates LEDGF transactivation capacity (Fig. 6). The same reversible modifier, Sumoylation, regulates LEDGF expression by regulating its transactivator, specificity protein Sp1 (Fig. 9). Furthermore, the biological functions of molecules are primarily dependent upon their cellular localization, expression levels and post-translational modification. Post-translational modifications such as phosphorylation, acetylation, ubiquitization and/or Sumoylation can alter the activity of proteins by modifying their DNA-binding affinity and transactivation capacity [23,72–74]. Recently, Sumoylation has been implicated in diverse regulatory functions, including subcellular compartmentalization, protein stability, chromatin structure regulation, transcription factor activity, DNA binding and protein complex assembly [41,43]. A growing list of transcription factors and coregulators has been shown to be modified by Sumo, indicating that Sumo modification is important in the regulation of gene transcription [75,76]. The process involved in the Sumoylation pathway is analogous to ubiquitization and requires a specific E1 activating enzyme (SAE1/SAE2), a Sumo-specific E2-conjugating enzyme (Ubc9) and an E3 ligating enzyme [76,77]. The Sumoylation target is a lysine that occurs in the consensus motif: ¥ KXE, where ¥ is a hydrophobic amino acid and X is any residue. The present study, with deletion mutants and a point mutation, showed that LKID positioned at amino acid 364 of the LEDGF was a Sumo1 modification motif. We identified K364 as the primary site and found that it lies inside the IBD domain [78–80]. Furthermore, in the present study, we found that an amount of naturally occurring LEDGF is constitutively Sumoylated by Sumo1 to lysine (K) 364 within the IBD domain, which is an evolutionarily well conserved region (Fig. 4) . Using sumoplot (web-based software for predicting Sumo1 conjugation in proteins) coupled with in vitro and in vivo Sumoylation assays, we found that LEDGF is Sumoylated by Sumo1 and Senp-1 as the Sumo proteases for LEDGF deSumoylation. Interestingly, protein expression analysis showed that sufficient endogenous LEDGF exists in Sumoylated form in the nuclear extracts of hLECs and mLECs (Fig. 1). It was also shown that LEDGF is polySumoylated ; however, unexpectedly, we found that endogenous LEDGF is monoSumoylated (Figs 1–3). This discrepancy may be associated with cell types, cell background or cellular microenvironment, or be a result of the ectopic/forced expression of molecules.
The most important finding of the present study is that the expression and transcriptional activity of LEDGF is regulated by the dynamical process of Sumoylation and deSumoylation. Sumo1 conjugation with LEDGF was further demonstrated by immunoprecipitation experiments in which proteins were first immunoprecipitated with anti-Sumo1 serum and analyzed by immunoblotting with LEDGF antibody (Fig. 5). These data indicate that endogenous LEDGF with a molecular mass of ∼ 87 kDa was conjugated with Sumo1. However, the mobility of Sumoylated LEDGF was retarded, reflecting an ∼ 12–15-kDa band shift, which demonstrates that naturally occurring cellular LEDGF modification may involve only a single Sumo1 molecule in LECs under normal physiological conditions (Fig. 1). Moreover, we could not rule out the possibility of polySumoylation of LEDGF in other cell types or when cells are ectopically overexpressed with interacting molecules (Sumo1 and LEDGF). The Sumoylation of endogenous LEDGF may require conformational changes that are not readily available to EGFP-linked Sumo1 in the presence of endogenous Sumo1. Possibly, low levels of other enzymes involved in Sumoylation processing did not cooperate with exogenously expressed Sumo1 in the recruitment of proteins that are Sumoylated before they associate with a gene . Moreover, post-translational modification by Sumoylation has been reported for a variety of proteins, including several transcriptional regulators, such as p53, c-Jun, c-Myb, AP-2, androgen receptor, promyelocytic leukaemia protein and IκBα [83–85]. Covalent attachment of the Sumo1 protein to the negative regulatory domain of the c-Myb transcription factor modifies its stability and transactivation capacity . Sumo1 modification has been found to have diverse substrate-specific functions. It has been shown to act antagonistically to ubiquitization by enhancing protein stability, as exemplified by Sumoylation of the nuclear factor-κB inhibitor IκBα . Sumo1 modification of IκBα inhibits nuclear factor-κB activation. Our findings indicate that Sumoylation, in general, has a suppressive effect on LEDGF transcription (at least in LECs), a mutant LEDGF (K364R) cannot be Sumoylated, and overexpression of this mutant protein increases the transcriptional activity of LEDGF. Also, the overexpression of Senp-1 enhances the transcriptional activity of LEDGF. Senp-1 is a nuclear protease that appears to deconjugate a large number of Sumoylated proteins . It has been identified in organisms ranging from yeast to mammals. Although several Senps have been reported in mammals, Senp-1 is the most often studied and it is known to regulate the activity of many transcriptional factors . We found that Senp-1 profoundly enhanced LEDGF-dependent transcription, which is very important in light of our earlier research showing LEDGF to be implicated in tumour progression and to be highly expressed in cancer cells [85a]. Senp-1 is also highly expressed in prostate cancer cells, and the silencing of the gene for Senp-1 attenuates the progression and growth of such cells . We expect that the silencing of Senp-1 will also reduce LEDGF expression. Elevated LEDGF/p75 expression has also been found in human breast and bladder carcinomas, and its ectopic overexpression increases the tumourigenic potential of human cancer cells in murine models .
Sumo1 modification may alter other properties of its targets and consequently their activities. It affects the transactivation capacity of p53, as well as its intracellular localization [87,88]. A family of cystein proteases (Senps) specifically hydrolyzes Sumo isopeptide bonds. For example, mammals have at least nine Senps localized in different subcellular compartments, such as at the promyelocytic leukaemia protein nuclear bodies (Senp-1), the cytoplasm (Senp-6), the nucleolus (Senp-3) and the nuclear pore (Senp-2). In the present study, the enzymatic activity of Senp-1 was used to independently confirm the effect of Sumoylation on the transactivating capacity of LEDGF. We found that Sumoylation/deSumoylation is at least one mechanism of controlling LEDGF activity, and it may be associated with cellular background. In particular, the Sumoylation of LEDGF led to transrepression of stress response genes such as small HSP genes (Fig. 6). Our experiments showed that cells transfected with mutant EGFP-LEDGF (K364R) had a higher activation of small HSP genes. In addition, we found that Senp-1 up-regulated HSP transcription and endogenous expression levels more than did Sumo1 (Fig. 6). An interesting finding shown in Fig. 6A,B indicates that an increase in Sumo1 concentration was associated with a decrease in LEDGF protein level (Fig. 6A, a, b) and with a decline in the expression of HSP27 protein (Fig. 6A, c). By contrast, an increase in the expression of Senp-1 was associated with an increase of LEDGF and HSP27 protein levels (Fig. 6B).
Furthermore, our data demonstrate that LEDGF and mutant LEDGF pEGFP-K364R localization patterns were indistinguishable from one another, indicating that Sumo1 conjugation or deSumoylation of LEDGF did not alter the subcellular localization of LEDGF, making LEDGF available for interaction with DNA binding. Furthermore, using western blotting, we examined whether mutant LEDGF pEGFP-K364R is more prevalent than LEDGF in cells (Fig. 5), proposing that deSumoylated LEDGF is stable and does not comprise the target of a quick ubiquitization pathway, as is the case for Sumoylated LEDGF. Although these results were derived from exogenously expressed LEDGF tagged to EGFP, similar results were reported previously . Moreover, stabilization and regulation of the LEDGF response in favour of cellular survival during environmental stress is essential. In general, the regulation of biological processes can be achieved through controlling the transcription and translational level. In the present study, we found that LEDGF activity was under the control of both levels. Moreover, LEDGF activated stress response genes by binding STRE and HSE. However, Sumo1 modification of HSF2 has been reported to have greater DNA-binding activity. Goodson et al.  proposed that this increase of DNA-binding activity is associated with conformational changes. Furthermore, our experiments also revealed that Senp-1 enhanced the DNA-binding activity of LEDGF in nuclear extract of cells compared to controls, whereas it was decreased in cells treated with Sumo1, emphasizing that a change in DNA-binding activity of LEDGF can be related to either conformational changes or a lower abundance of LEDGF protein as a result of the repression of LEDGF expression or stability as a result of Sumoylation (Fig. 7A). Sumo1 conjugation of LEDGF has been reported to decrease the half-life of LEDGF . Our data, however, reveal an increased level of LEDGF protein in Senp-1 treated cells (Figs 5 and 6B), raising the possibility that this increase of LEDGF protein may be associated with the deSumoylation of LEDGF. Sumoylation was shown to both stabilize and destabilize the proteins and determine their fate and integrity . It is also possible that the Sumoylation of LEDGF may alter its localization patterns. In this case, Sumoylated LEDGF is not available to bind the DNA. The present study ruled out this possibility because the localization patterns of LEDGF and mutant LEDGF pEGFP-K364R were indistinguishable and both were predominantly localized in the nucleus (Fig. 5D). However, the level of LEDGF Sumoylation and its potential for repression can be regulated by the opposing activities of Sumoylases and deSumo proteases. Thus, a balance in the expression and activity of the two is essential for determining the cellular signalling that determines the fate of cells.
Furthermore, the present study showed that the protein level of LEDGF was increased with an increase of Senp-1, whereas levels were decreased in cells overexpressing Sumo1. Two possibilities appeared to exist: (a) LEDGF accumulated in the presence of Senp-1 (increase of half-life) as a result of the inhibition of degradation by deSumoylation of LEDGF [20,50] or (b) Senp-1-dependent up-regulation of LEDGF transactivation counteracted transrepression by Sumo1. Real-time PCR data revealed that the latter was the case; the expression of LEDGF mRNA was significantly increased in cells overexpressing Senp-1 (Fig. 8A) compared to cells overexpressing Sumo1 or control vector. Senp-1-induced the increased expression of LEDGF mRNA, indicating that LEDGF may be transcriptionally regulated. Our cotransfection experiments involving the promoter activity assay revealed that Senp-1 up-regulated LEDGF transcription, and promoter activity was dramatically increased in a dose-dependent fashion (Fig. 8B). The evidence of the role of Senp-1 in controlling other transcription factors is indeed compelling. It functions in two ways: by up- or down-regulating transcriptional activity and by modulating DNA-binding activity, which is exemplified by GATA-1, HSF1 or nuclear factor-kB, amongst others [34,89,90]. However, LEDGF activity was not fully eliminated by Sumoylation or deSumoylation. We assumed that LEDGF in cells is not absolutely Sumoylated or deSumoylated because both are dynamic phenomena. We surmise that the balance is established to maintain optimum cellular signalling beneficial to cells and cell background. Certainly, further investigations are required to determine how this balance of Sumoylation/deSumoylation occurs in cells. However, Sumoylation is a very dynamic process involving, on the one hand, the conjugation components and, on the other hand, the deconjugation machinery. Thus, short period of cycles of conjugation and deconjugation may support the existence of a usually very low steady-state of target gene modification within the cellular microenvironment . However, this may be true for transcription factors such as c-Jun and c-fos that are deSumoylated within minutes. By contrast, RanGAP-1 is particularly stably and constitutively Sumoylated and is almost (possibly) fully deconjugated, suggesting that each protein has differential sensitivity and selectivity. Our results also showed that a portion of LEDGF is constitutively Sumoylated and may have differential affinity to Sumo1 conjugation, and that this affinity may be changed according to the cellular background.
The present study is the first demonstration of the functional significance of LEDGF deSumoylation. Although Senp-1 enhanced the binding of LEDGF to its binding element, this may be attributable to a higher abundance of unSumoylated protein because of the higher transcriptional expression of LEDGF and its deSumoylation by Senp-1 (Figs 6 and 8). However, no effect of LEDGF Sumoylation on its chromatin-binding activity has been shown. The difference may lie in the conformational changes in exogenously expressed LEDGF and Sumo1 or the different cell types used in a study, or the cellular microenvironment. Several studies have emphasized that Sumo or Senp activities are dependent upon cell type and cell backround, or cellular microenvironment . Furthermore, we found that LEDGF promoter ranging from −170/+35 bore Sp-binding sites and predicted that the transrepression of LEDGF or its transactivation can be associated with Sp1. Our data revealed that Sp1 indeed regulated LEDGF transcription. Senp-1 up-regulation of Sp1 may be associated with its higher abundance in cells. Sumo1 conjugation is known to decrease Sp1 transcriptional activity . Our results clearly demonstrate that Sumo1 down-regulated LEDGF transcription, and that this repression of the LEDGF gene may be a result of destabilization and degradation of Sp1 by Sumo1 [60,93]. A previous study reported that constitutively Sumo-modified Sp1 was a poorer transactivator than Sp1 . Modification of Sp1 by Sumo1 was found to affect the stability of Sp1, thereby inhibiting Sp1-dependent transcription. Another study found that Sumoylation inhibits cleavage of the negative N-terminal regulatory domain of Sp1 and p21WAF1/CIP1 transcriptional activity [60,93]. Only a small fraction of total Sp1 is Sumo1-modified, and this difference in activity suggests that Sumoylation plays an important role in Sp1-dependent transcription. When a LEDGF transactivation assay was performed after overexpressing hLECs with Sumo1 protein, we noted reduced CAT values. In a similar assay after overexpressing Sumo hydrolase, Senp-1, the CAT value was much higher. Our finding demonstrated that mutant LEDGF pEGFP-K364R (where Sumo1 conjugation sites were disrupted) enhanced cell growth and cell survival. The results obtained with cells that received Senp-1 were similar, suggesting that deSumoylation is functionally important for cells under normal physiological conditions. Although Sumoylation under normal conditions may maintain cellular homeostasis by restricting cell growth, with such cells gaining resistance against stressors, Sumo- transfected cells were more resistant to heat stress and survived better than Senp-1-transfected cells. However, further work is needed to determine the underlying mechanism by which Sumo1 conjugation of LEDGF provides cytoprotection against stress. We are exploring whether Sumoylated LEDGF interacts with other proteins that stablize the protective function of LEDGF during stress.
Finally, we found that LEDGF activity was controlled by dual mechanisms that include the transcription and post-translational modification of LEDGF, and these processes are under the control of the reversible Sumoylation process. Based on our results, we propose that LEDGF Sumoylation and deSumoylation act as a molecular switch that determines cellular integrity/fate by regulating LEDGF-dependent cellular signalling within the cellular microenvironment. If this process goes awry, the aberrant expression of LEDGF leads to cellular abnormalities. In conclusion, our data show that naturally occurring LEDGF is constitutively Sumoylated in a process involving Ubc9 (E2), that Senp-1 is one peptidase involved in LEDGF deSumoylation, and that this dynamic process controls LEDGF activity by regulating its expression and protein modification.
Materials and methods
hLECs (a gift from Dr V. N. Reddy, Eye Research Institute, Oakland University, Rochester, MI, USA)  and mLECs were maintained routinely in our laboratory as described previously . Briefly, cells were cultured in a 75-mm tissue culture flask in DMEM supplemented with 15% or 10% heat-inactivated fetal bovine serum, 100 μg·mL−1 streptomycin and penicillin in a 5% CO2 environment at 37 °C in accordance with standard methods. Cells were harvested and cultured in 96-, 24-, 48- or six-well plates and 100-mm petri dishes in accordance with the requirement of the experiment.
Generation of prokaryotic expression vectors
Subcloning techniques described by Sambrook et al.  were used throughout the present study. A fusion protein between LEDGF and partial LEDGF with GST, generated by inserting the entire coding sequence of the LEDGF cDNA into the BamHI and EcoRI sites of a pGEX-2T vector (Pharmacia Biotech, Piscataway, NJ, USA), was used to transform Escherichia coli (BL21) . GST-LEDGF, N-terminal (amino acids 1–250) and C-terminal (amino acids 170–530) deletion constructs were also prepared. The expression of the GST-LEDGF fusion protein was induced with isopropyl thio-β-d-galactoside. Proteins were purified with glutathione-Sepharose 4B beads (Pharmacia Biotech) in accordance with the manufacturer’s instructions. Similarly, a construct containing GFP and LEDGF cDNA was also generated with the ‘Living Color System’ (Clontech, Palo Alto, CA, USA) using the plasmid vector pEGFP-C1 (Clontech) for eukaryotic expression . Also, the full length of Sumo1 was subcloned into pGEX-5X-3 (Pharmacia Biotech) vector for prokaryotic expression by creating BamH1 and Xho1 sites in the forward and reverse primers, respectively (forward: 5′-CAGGATCCTCATGTCTGACCAGGAGGCA-3′; reverse: 5′-CCGTCTCGAGCACCACATTACAAAAGAAC-3′). Additionally, the GST-HSF1 construct was made by subcloning the full length of the HSF1coding region into pGEX-5X-3 vector using specific primers and sequenced. The sequence was confirmed by automated DNA sequencing, the purified plasmid was transformed into E. coli BL21, and the protein was purified.
Construction of pEGFP-Sumo1
For eukaryotic expression, the full length of Sumo1 cDNA was subcloned into pEGFP-C1 vector. The coding region of Sumo1 was amplified by PCR from human lens cDNA library using forward (5′-CCGTCGACATGTCTGACCAGGAG-3′) and reverse primer (5′-TCGGATCCGTTTTGAACACCACA-3′) with restriction enzyme sites, SalI and BamHI. The PCR product was digested and ligated into pEGFP vector. pFlag-Senp-1 was a generous gift from Dr E. Yeh (University of Texas MD Anderson Cancer Center, Houston, TX, USA). All the Transfection experiments were carried out either with Superfactamine Reagent (Invitrogen, Carlsbad, CA, USA) or using the Neon Transfection system (Invitrogen).
Construction of LEDGF gene promoter linked to CAT
The 5′-flanking region of the human LEDGF gene was isolated and sequenced as reported previously . A construct of −5139 bp was prepared by ligating it to basic pCAT vector (Promega, Madison, MI, USA). Similarly, constructs of different sizes were prepared with were made prepared with appropriate sense primers bearing SacI or MluI and antisense with NheI and ligated into pCAT-basic vector as described previously . The plasmid was amplified and used for CAT assay.
Construction of HSP27 and αB-crystallin-CAT
pCAT-HSP27 and pCAT-αB-crystallin constructs were engineered as described previously . Briefly, the 5′-flanking region of the human HSP27 gene was isolated with a genomic PCR kit (Clontech) using specific primers. A forward primer containing a SacI site (5′-GCGTCGAGCTCTCGAATTCATTTGCTT-3′) and reverse primer with a XhoI site (5′-GCTCTCGAGGTCTGCTCAGAAAAGTGC-3′) were used to generate the fragment, which was cloned between the EcoRI sites of the TA vector (Invitrogen). Similarly, a fragment comprising the 5′-flanking region of the human αB-crystallin promoter (a gift from Dr J. Piatigorsky, NEI, NIH, Metheda, MD, USA) was prepared using specific primers. A forward primer with a SacI site (5′-CTCTCTTCCAAGAGCTCACAAAG-3′) and reverse primer containing a XhoI site (5′-ATGGTGGCTACTCGAGAGTGA-3′) were used to generate the above fragment, which was cloned between the EcoRI sites of the TA vector. The HSP27 and αB-crystallin/TA constructs were digested with SacI and XhoI and promoter fragments were ligated to pCAT-Basic vector (Promega), using the appropriate restriction enzymes.
Preparation of siRNAs
The LEDGF-specific siRNA expression plasmid was designed as described previously . The sequence was selected from location 1340–1360 (5′-AAAGACAGCATGAGGAAGCGA-3′). The sense and antisense oligonucleotides with the internal loop were synthesized by Invitrogen. These were annealed and ligated into the BamHI and HindIII sites of pSilencer 4.1-CMV hygro (Ambion, Austin, TX, USA). pSilencer 4.1- pCMVhygro expressing a scrambled siRNA (Ambion) was used as a control.
GST-K364R and pEGFP-K364R constructs carrying a substitution at lysine 364 for arginine was prepared using the Quickchange™ Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA, USA) in accordance with the manufacturer’s instructions. Briefly, amino acid exchanges were generated by point mutations in the pEGFP-LEDGF using the complimentary primers (K364R):
and reverse: 5′-TCAAGATTATCAATTCTGAGTGAATTCTTAATCTCAGCATGT-3′.
After site-directed mutagenesis, XL1-Blue super-competent cells (Stratagene) were subsequently transformed with mutated cDNA, and clones were grown on LB/kanamycin plates. The plasmid was amplified, and the mutation was confirmed by sequencing. The mutated plasmid was transfected and intracellular translocation was determined using fluorescent microscopy.
Cotransfection and promoter activity assay
The CAT assay was performed using a CAT-ELISA kit (Roche Diagnostic Corporation, Indianapolis, IN, USA). Cells were transfected/cotransfected with various reporter constructs (pCAT-LEDGF, pCAT-HSP27 and pCAT-αB crystallin) and/or pEGFP-Sumo1, pFlag-Senp-1, pEGFP-LEDGF or pEGFP-K364R expression vectors. After 48 or 72 h of incubation, cells were harvested, and extracts were prepared and the protein was normalized. CAT-ELISA was performed to monitor CAT activity in accordance with the manufacturer’s instructions. A405 was measured using a microtitre plate ELISA reader. Trans-activation activities were adjusted for transfection efficiencies using GFP values.
Electrophoretic mobility shift assay
An electrophoretic mobility shift assay was performed as described previously . Oligos containing HSE or STRE elements were commercially synthesized, annealed and end-labelled with [γ-32P] ATP using T4 polynucleotide kinase (New England Biolabs, Beverly, MA, USA). The binding reaction was performed in 20 μL of binding buffer containing 20 mm Tris-HCl (pH 8.0), 75 mm KCl, 5% glycerol, 50 μg·mL−1 BSA, 0.025% Nonidet P-40, 1 mm EDTA, 5 mm dithiothreitol and 1 μg of poly (dI/dC). Five ficomols (1000 c.p.m.) of the end-labelled probe were incubated on ice with GST-LEDGF or GST-K364R fusion protein. Samples were then loaded on 5% polyacrylamide gel in 0.5 × TBE buffer for 2 h at 10 V·cm−1. The gel was dried and autoradiographed.
sumoplot or Sumo site analysis and in vitro Sumoylation assay
To predict and identify the Simulation site(s) in LEDGF protein, computational prediction software was used: sumoplot (Abgent) and the pic-Based Sumo Site Prediction Server. In vitro Sumo modification reaction was performed in accordance with the manufacturer’s instructions (#K007 kit; LAE Biotech, Rockville, MD, USA). Bacterial expressed GST, GST-LEDGF, GST-K364R and GST-HSF1 were purified using glutathione-Sepharose 4B (Pharmacia) in accordance with the manufacturer’s instructions and then used as substrates in the in vitro reaction. Five hundred nanograms of GST-LEDGF, GST-K364R or GST-HSF1, 150 ng of SAE I/SAE II, 1 μg of Ubc9 and 1 μg of Sumo1 were incubated in the reaction buffer containing 20 mm Hepes (pH 7.5), 5 mm MgCl2 and 2 mm ATP in a total volume of 20 μL for 1 h at 37 °C. Reactions were stopped by the addition of 2 × SDS/PAGE sample buffer and subjected to SDS/PAGE followed by western blotting. Blotted membrane was immunostained with primary antibodies (anti-GST, anti-Sumo1 or anti-LEDGF). Membranes were incubated with horseradish peroxidase-conjugated secondary antibodies. Specific protein bands were visualized by incubating the membrane with luminal reagent (Santa Cruz Biotechnology) and recorded with a Fuji Film-LAS-4000 luminescent image analyzer (Fuji Film Medical Systems, Tokyo, Japan).
In vivo Sumoylation assay
hLECs were cotransfected with either pEGFP-LEDGF or pEGFP-K364R and pEGFP-Sumo1. After 48 h, cytoplasmic and nuclear extracts were prepared as described previously [12,47]. Nuclear extract was incubated with 3–4 μg of monoclonal anti-LEDGF serum (catalogue number 611714; BD Biosciences, Franklin Lakes, NJ, USA) in binding buffer provided in the immunoprecipitation kit (Pierce, Rockford, IL, USA), and maintained at 4 °C for 2 h followed by the addition of 40 μL of of Protein A–Sepharose pre-absorbed with BSA and rotated overnight at 4 °C. The immunoprecipitates were collected by centrifugation and washed several times before boiling in SDS-sample buffer. Precipitates were resolved on 4–20% SDS/PAGE and analyzed by western blotting using monoclonal anti-LEDGF, anti-Sumo1 or anti-GFP sera.
Whole cell extracts or nuclear extracts were prepared as described previously [12,47] and western blotting was performed. Equal amounts of protein samples were loaded on 10% or 7.5% or 4–20% SDS/PAGE, blotted onto a poly(vinylidene difluoride) membrane, and then immunostained with primary antibodies at the appropriate dilutions. LEDGF monoclonal (catalogue number 611714; BD Biosciences) and polyclonal (catalogue number LS-C31241; Lifespan Biosciences; catalogue number sc-33371, SantaCruz Biotechnology, CA, USA) antibodies were used. GFP (catalogue number sc-8334), GST (catalogue number sc-53909), Sumo1 (catalogue number sc-9060), Senp-1 (catalogue number sc-46634) and HSP27 (catalogue number sc-1048) antibodies were purchased from SantaCruz Biotechnology. The membranes were then incubated with horseradish peroxidase conjugated with secondary antibodies. Specific protein bands were visualized by incubating the membrane with luminal reagent (Santa Cruz Biotechnology, CA, USA) and recorded with an LAS-4000 luminescent image analyzer (Fuji Film Medical Systems). To confirm comparative expression and equal loading of the protein samples, the membrane stained earlier was stripped and re-probed with β-actin antibody (Abcam, Cambridge, MA, USA).
Real-time PCR or quantitative PCR
Total RNA was isolated using the single-step guanidine thiocyanate/phenol/chloroform extraction method (Trizol Reagent; Invitrogen) and converted to cDNA using Superscript II RNAase H-Reverse Transcriptase. Quantitative real-time PCR was performed with SYBR Green Master Mix (Roche Diagnostic Corporation) in a Roche® LC480 Sequence detector system (Roche Diagnostic Corporation). PCR conditions consisted of 10-min hot start at 95 °C,followed by 45 cycles of 10 s at 95 °C, 30 s at 60 °C and 10 s at 72 °C. The primer sequence was: LEDGF, 5′-CAGCAACAGCATCTGTTAATCTAAA-3′ and Reverse primer: 5′-GGGCTGTTTTACCATTTTGG-3′; β-actin, 5′-CCAACCGCGAGAAGATGA-3′ and 5′-CCAGAGGCGTACAGGGATAG-3′. The expression levels of target genes were normalized to the levels of β-actin as an endogenous control in each group.
Cell survival assay
A colorimetric MTS assay (Promega) was performed as described previously [12,47]. This assay of cellular proliferation uses 3-(4, 5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2 to 4-sulfophenyl)-2H-tetrazolium salt (Promega). Upon addition to medium containing viable cells, the salt is reduced to a water-soluble formazan salt. A490 was measured after 4 h with an ELISA reader.
Data are presented as the mean ± SD of the indicated number of experiments. Data were analyzed by Student’s t-test as appropriate. P < 0.05 and P < 0.001 were considered statistically significant.
Grants provided by the National Eye Institute, NIH (EY-13394 and EY017613) (to D.P.S.) and Research for Preventing Blindness are gratefully acknowledged. Grant support by American Health Assistance Foundation (to N.F.) is gratefully acknowledged.