Sumoylation-deficient Prdx6 repairs aberrant Sumoylation-mediated Sp1 dysregulation-dependent Prdx6 repression and cell injury in aging and oxidative stress

Progressive deterioration of antioxidant response in aging is a major culprit in the initiation of age-related pathobiology induced by oxidative stress. We previously reported that oxidative stress leads to a marked reduction in transcription factor Sp1 and its mediated Prdx6 expression in lens epithelial cells (LECs) leading to cell death. Herein, we examined how Sp1 activity goes awry during oxidative stress/aging, and whether it is remediable. We found that Sp1 is hyper-Sumoylated at lysine (K) 16 residue in aging LECs. DNA binding and promoter assays revealed, in aging and oxidative stress, a significant reduction in Sp1 overall binding, and specifically to Prdx6 promoter. Expression/overexpression assay revealed that the observed reduction in Sp1-DNA binding activity was connected to its hyper-Sumoylation due to increased reactive oxygen species (ROS) and Sumo1 levels, and reduced levels of Senp1, Prdx6 and Sp1. Mutagenesis of Sp1 at K16R (arginine) residue restored steady-state, and improved Sp1-DNA binding activity and transactivation potential. Extrinsic expression of Sp1K16R increased cell survival and reduced ROS levels by upregulating Prdx6 expression in LECs under aging/oxidative stress, demonstrating that Sp1K16R escapes the aberrant Sumoylation processes. Intriguingly, the deleterious processes are reversible by the delivery of Sumoylation-deficient Prdx6, an antioxidant, which would be a candidate molecule to restrict aging pathobiology.


INTRODUCTION
All cells or organisms encounter many types of environmental stresses and apoptotic stimuli. In young cells, the defense responses are efficient, but with age the responses are weakened by dysregulation of antioxidant response, which leads to impaired cellular function and increased vulnerability [1,2]. Protein homeostasis is essential to maintain cellular systems during the progressive rise of oxidative stress in aging [3][4][5]. Many biologically relevant factors in the cellular and external environments, such as chemical, ultraviolet B (UVB) radiation, H 2 O 2 and growth factors, have been shown to initiate reactive oxygen species (ROS)-evoked adverse signaling in cells due to malfunction of antioxidant defense system. The antioxidant defense system is equipped with antioxidant proteins, like CAT, SODs, GPX1, Prdxs. The stress factors significantly negatively alter their activity that results in age-related pathobiology [6]. We and other investigators have reported that Prdx6 maintains cellular homeostasis by maintaining calcium homeostasis and cell membrane and DNA integrity by optimizing ROS levels against the stressors such as H 2 O 2 , UVB, paraquat, endoplasmic reticulum (ER) stress and overstimulated growth factors [7][8][9]. Our published report has also shown that expression of the antioxidant defense gene Prdx6 and its regulator Sp1 decline significantly in an age-dependent The protective protein Prdx6 is a member of the Prdx family, which consists of conserved single cysteine (Cys) residue (known as 1-Cysteine). This has both GSH-Peroxidase and acidic Ca 2+ -independent phospholipase A 2 (PLA 2 ) activities, and is widely expressed at significant levels in eye lens, lung, liver, testis and brain. Prdx6 is predominantly localized in cytoplasm, including endoplasmic reticulum, mitochondria, lysosomes and plasma membrane [5,7,[12][13][14][15]. This localization pattern of Prdx6 in ROS-producing organelles emphasizes its biologically important function in maintaining redox homeostasis. We have shown that Prdx6 expression is significantly reduced in aging lenses and LECs [8,16], and these LECs/lenses are highly susceptible to cell death and lens opacity induced by stressors [5,[17][18][19]. In the current study, we found that TATA-less human Prdx6 promoters bore four active binding sites for Sp1, which are evolutionarily well conserved (This study, Table 1). Transcription factor Sp1 belongs to the family of Sp/KLF (Kruppel-Like Factor) nuclear proteins [20,21]. Sp1 is responsible for basal transcription of genes, and can also modulate its target gene transcription in response to physiological and pathological stimuli [22]. It binds with high affinity to GC-rich motifs (such as 5'-GGGGCGGGG-3' or 5'-G/T-GGGCGG-G/A-C/T-3' or 5'-G/T-G/A-GGCG-G/T-G/A-G/A-C/T-3') and regulates the expression of TATA-containing or TATA-less genes via protein-protein, protein-DNA interaction or interplay with other transcription factors, and also through a large number of genes involved in a variety of cellular physiological processes such as growth, apoptosis, differentiation and senescence [23]. Nonetheless, posttranslational modifications of proteins are an important mechanism that regulates protein function, activity and localization [24]. Posttranslational modifications of Sp1, such as glycosylation, phosphorylation, methylation, acetylation, ubiquitination, and Sumoylation modulate the Sp1 and its target gene expression by affecting its protein level, transcription and DNA-binding activities [20,[25][26][27]. Oxidative stress-induced aberrant protein modification has been implicated in the etiology and progression of many human diseases [2,[28][29][30][31]. Sp1 expression has been shown to be significantly reduced in aging cells [11] and cells facing oxidative stress [5,9,11,32]. Cells overexpressing Sumo display repressed expression of Sp1 and its target gene Prdx6, whereas Sumo-protease (Senp1) overexpression enhances Prdx6 expression and transcription activity by increasing the levels of Sp1 [5]. In this scenario, we postulated that aberrant Sumoylation of Sp1 during aging and oxidative stress appears to be a plausible cause for Sp1 dysregulation and modulation of its target genes.
Sumos play important roles in the regulation of diverse cellular processes, including gene transcription, protein stabilization, and subcellular localization [12,[33][34][35]. Most, if not all, Sumo modification of transcription factors results in repression of their activity. Three major Sumo isoforms, Sumo1, Sumo2 and Sumo3 have been identified in mammals [24,26]. Sumo1 exhibits ~50% amino acid sequence identity with Sumo2 and Sumo3, whereas Sumo2 and Sumo3 share about 97% amino acid sequence identity and are referred to as Sumo2/3 in most cases [24,33,[36][37][38][39][40]. Sumo binds covalently to €-lysine embedded in the ɸKXE motif (ɸ represents hydrophobic amino acid, and X stands for any amino acids) within the target protein in an enzymatic cascade reaction leads to protein Sumoylation [24,40]. Sumoylation can be reversed by a family of sentrin/Sumo-specific proteases (Senps) [41,42]. Senp1 is largely responsible for the deconjugation of both Sumo1 and Sumo2/3 modifications, and regulates the activity of many transcription factors including Sp1 in vivo and in vitro [5,11,12,41,43]. This process can be aberrantly affected during oxidative stress and aging, leading to aberrant Sumoylation processes of proteins like Sp1, and thereby altering protein functions (dysregulation of Sp1 activity in the current study).
In the study reported here, we observed that during aging and oxidative stress, a progressive decline of Prdx6 expression was linked to an increase of Sp1 Sumoylation with decrease in Sp1 expression wherein Sp1-DNA binding activity to Prdx6 promoter was greatly reduced. We also noted that reduction in Sp1-DNA binding activity was connected to increased Sumo1 and ROS levels, and decreased Senp1 and Prdx6 as well as reduction in Sp1-DNA activity and expression in aging LECs and cells facing oxidative stress. We found that Sp1 was Sumoylated in vivo at K16 residue in LECs, a major site for the Sumoylation of Sp1. Additionally, data revealed that overexpression of Sumoylation-deficient Sp1K16 improved DNA-binding activity by escaping the erratic Sumoylation that occurs in aging or oxidative stress. An important observation was that delivery to cells of Prdx6 mutant at Sumo1 motif(s) linked to TAT-transduction domain provided cytoprotection by restoring Sp1 stability and DNAbinding activity and protecting against oxidative cell injury by halting ROS-driven aberrant Sumoylation processes. The findings offer a new perspective for developing antioxidant Prdx6-based therapy to rescue cells and organisms from ROS-evoked aberrant Sumoylation signaling.

Age-related increases of ROS levels in LECs were connected to progressive decline in Sp1 and Prdx6 expression and Sp1-DNA binding activity to its GC rich elements
During aging, gene expression levels change, a situation which may be associated with the accumulation of high levels of ROS [44]. To determine a connection between levels of ROS, Prdx6 and Sp1, and binding efficiency of Sp1 to its response elements (GC-box), we first monitored the intracellular redox-state of primary hLECs of different ages. Quantification by staining with H 2 DCFDA dye showed an age-dependent progressive increase in ROS levels (Fig. 1A), which reached significantly higher levels in aged hLECs (Fig. 1A, 56y onward). Next, we isolated RNA from the same groups of aging cells and quantified mRNA by real-time PCR. We observed that the levels of both Sp1 and Prdx6 mRNA in hLECs declined with aging, and this loss was more significant in aged cells (Fig. 1B, 56y onward). Collectively the results revealed a significant inverse correlation between expression of Sp1/Prdx6 and ROS levels during aging. Because we found a direct correlation between expression levels of Prdx6 mRNA and its regulator Sp1 mRNA and protein ( Fig. 1), we surmised that this could be related to a loss of Sp1 cellular abundance or reduction in its binding efficiency to Prdx6 promoter due to increased levels of ROS in aging cells. To explore that possibility, nuclear protein isolated from hLECs of different ages was used to quantify the presence of active Sp1 by using TransAM Sp1 transcription factor assay (Active Motif) as well as Sp1 protein level. Data revealed that, indeed, Sp1-DNA activity declined (Fig.1C), and that reduction in Sp1-DNA activity was connected to decline of Sp1 cellular levels with increase in age (Fig. 1E), suggesting that an increase in ROS-induced oxidative stress could jeopardize Sp1 activity and lead to repression of Prdx6 mRNA. Figure 1E reveals that Sp1 protein declined with advancing age as evidenced by Western analysis. However, due to the limited supply of primary hLECs, we were able to perform Sp1 protein expression analysis on only cells of certain age points (as Western analysis requires larger amounts of protein extracts). Next we asked whether dysregulation of Sp1 was due solely to reduced cellular abundance or if a reduction in Sp1 binding efficiency in nuclear extracts of aging cells might have made a contribution. We equalized Sp1 contents in nuclear extracts of hLECs isolated from different age groups using Sp1 specific sandwich-ELISA as described in Materials and Methods, and measured the Sp1/ DNA binding activity with TransAM Sp1 transcription factor assay (Active Motif). We found that both reduced abundance and a decrease of Sp1 binding efficiency were responsible for dysregulation of Sp1-DNA binding activity during aging (Fig. 1D).

AGING
In vivo DNA binding assay revealed reduction in Sp1-DNA binding to Prdx6 promoter during H 2 O 2 or UVB-induced oxidative stress in LECs Major causes of age-related pathobiology include an increased burden of oxidative stress and its associated damage to cells/tissues due to accumulation of ROS, a progressive decline of antioxidant expression, and a deterioration of transcription factor activity in aging [10,[45][46][47]. Our previous report [5] and an analysis of data presented in Figure 1 demonstrate that a progressive increase of ROS levels in aging cells was linked to progressive reduction in expression and activity of Sp1 and its target gene Prdx6 mRNA. We surmised that there might be a decrease in Sp1 interaction with its elements present in Prdx6 promoter due to increase of oxidative load in vivo. To examine this possibility, we performed ChIP assay using antibody specific to Sp1, as described in Materials and Methods, using hLECs facing different concentrations of oxidative stress. Reports by others using other cell types have shown that H 2 O 2 -induced oxidative stress lowers Sp1-DNA binding [48,49]. We initially per-formed bioinformatics analysis and spotted out Sp1 binding sites in mouse, rat and human gene promoters. As shown in Table 1, Sp1 binding sites present in mouse, rat and human Prdx6 promoter are well con-served. Next, chromatin samples were prepared from mouse and human LECs exposed to H 2 O 2 or UVB multiple times as indicated in Figs. 2 and 3. Protein-DNA complex was pulled down with anti-Sp1 antibody or with control IgG. Immunoprecipitated complex was  AGING processed for RT-PCR (mLECs) and qPCR (hLECs) using the primer that encompassed mouse and human Prdx6 promoters containing Sp1-responsive elements ( Figs. 2A and 3A). We observed a significant decrease in the enrichment of Sp1 with its target elements at Prdx6 promoters in samples exposed to both H 2 O 2 (Figs. 2B and 3B) and UVB (Figs. 2C and 3C). Interestingly, that decrease of Sp1 binding was directly related to decline of Sp1 mRNA and protein as revealed by expression analysis, suggesting that oxidative stressinduced reduced cellular abundance of Sp1 was one cause of decrease in Sp1-DNA activity (Fig. 2D). No effect was observed on internal control as shown in Fig.  2D and E, as observed in aging hLECs (Fig. 1E). Furthermore, in the same experiment, qPCR analysis of Sp1 targeted gene, Prdx6 transcription also showed a decline in Prdx6 mRNA, which was directly related to a decrease of Sp1-DNA binding (as observed in the ChIP experiment), corroborating that aging or oxidative stress did not affect gene regulation globally and transcriptional machinery was in active state, at least in hLECs facing oxidative stress in the model system observed. Sp1 is a transregulator of mouse Prdx6 transcription [5,9]. To examine whether Sp1 binding to human Prdx6 promoter activated the promoter through its responsive elements (GC-boxes), we performed transactivation assay by using mithramycin A (Mithra A), an inhibitor of Sp1 (Fig. 3D). SRA-hLECs were transiently transfected with pCAT-hPrdx6 (-918/+30) promoter or empty CAT vector, and treated with increasing amounts of Mithra-A as indicated in Fig. 3D. We observed that Mithra-A inhibited Prdx6 promoter activity in dosedependent fashion, suggesting that Sp1 bound and activated human Prdx6 promoter. Taken together, our results revealed that oxidative stress attenuated Prdx6 transcription through Sp1. Sandwich/Sumo1-ELISA assays to examine the total and Sumoylated forms of Sp1 protein. LECs (a model for aging) as well as in aging hLECs, and that increase was linked to a progressive increase of ROS-induced oxidative stress [5]. These LECs also showed reduced Sp1, Prdx6 and Senp1 expression (as well as the presence of an inactive dimeric form of Senp1) with increased Sumo1 levels, suggesting oxidative-induced aberrant Sumoylation signaling as a possible cause of reduction in Prdx6 and Sp1 levels. To uncover the role of Sumo1 in dysregulation of Sp1 during oxidative stress, we analyzed deSumoylation and Sumoylation status of Sp1 in Prdx6 +/+ and Prdx6 -/-LECs with or without oxidative stress by using Sp1 sandwich/Sumo1-ELISA as shown in Fig. 4A. We observed a significantly higher level of the Sumoylated form of Sp1 in Prdx6 -/compared to Prdx6 +/+ LECs, suggesting that Prdx6 deficiency (increased ROS levels) may cause an increase in Sumoylation status of Sp1, at least in LECs. In an earlier report [5], we showed that Prdx6 -/-LECs facing oxidative stress displayed increased Sumo1 conjugation and reduced Sp1 expression. Furthermore, Prdx6 -/-LECs exposed to different concentrations of H 2 O 2 for 48h ( Fig. 4B) or 0.3mM H 2 O 2 for different time intervals (Fig. 4C) displayed dramatic increases in Sp1 Sumoylation in dose-and time-dependent manner. Similarly, we exposed Prdx6 -/-LECs to UVB (to see indirect effect) as eyes are maximally exposed to UVB as indicated ( Fig. 4D and 4E). At 60 and 90J/m 2 UVB exposure, Sp1 Sumoylation significantly increased, but at 30J/m 2 no significant change was detected (Fig. 4D).
At the higher dose of 180J/m 2 UVB, we observed a further significant increase in Sp1 Sumoylation in timedependent fashion (Fig. 4E). Data revealed that as Prdx6-deficient mLECs came under redox stress, they showed increased Sumo1 and increased Sp1 Sumoylation. Oxidative stress induced by UVB or H 2 O 2 increased production of ROS in Prdx6 -/-LECs, which may have further increased Sumo1 conjugation and thereby reduced Sp1 expression by increasing the Sumoylation status of Sp1.

Increases in Sumo1 levels were directly connected to increased Sumoylation of Sp1, and inversely related to Senp1 levels in aging/aged lenses/LECs
We next examined whether Sumoylation status of Sp1 has any link to level of Sumo1 or Senp1 during aging of hLECs, similar to the relationship observed in Prdx6deficient mLECs which showed increased levels of ROS. Because of the increased ROS levels (Fig.1A), we AGING posited that aging/aged hLECs should have increased amounts of Sumoylated Sp1. (Fig. 4). To examine this, total RNA isolated from human lenses/LECs of different ages was processed for mRNA analysis using qPCR. Data showed a significant increase in Sumo1 and reduction in Senp1 mRNA with increases in age ( Fig.  5A, 56y onward). To determine whether modulation of Sumo1 or Senp1 expression in aging LECs had any influence on Sumoylation or deSumoylation status of endogenous Sp1, we quantified levels of Sumoylated and deSumoylated forms of Sp1 in the same group of aging/aged hLECs by utilizing a sensitive Sp1 sandwich/Sumo1-ELISA as described in Materials and Methods. Equal amounts of nuclear extracts isolated from hLECs were processed. Data revealed that the ratio of deSumoylated and Sumoylated forms of Sp1 was 4:1 in the youngest age group (16-18y). The ratio gradually changed with aging: Age group 52-58y, 1:1.5; Age group 62-66y, 1:3; and Age group 74-76y, 1.4. Taken together, data disclosed that Sumoylation of Sp1 was increased with aging/aged hLECs and suggested that oxidative load might be a prime cause for aberrant Sumoylation of Sp1.

Sumoylation negatively affected the DNA binding activity of Sp1 while the deSumoylating agent Senp1 enhanced Sp1-DNA binding
Studies showed that Sumo1 conjugation might alter the DNA binding of proteins [11,27,[50][51][52]. Our current study (Figs.1 to 5) and previous report [5] revealed that Sumo1 overexpression reduced Sp1 mRNA and protein expression, in contrast to Senp1. Next, to determine whether Sumoylation directly affects Sp1-DNA binding activity in Prdx6 promoter, we carried out a gel-shift mobility assay. Nuclear extracts with equal amount of protein from LECs transfected with pEGFP-Sumo1 or pEGFP-Vector were incubated with radiolabeled WT or mutant probes and processed for gel-shift assay (Fig.  6B). The binding of Sp1 in nuclear extract from Sumotransfected cells to the WT probe was reduced significantly (Fig. 6B, lane 3; 6C, right panel, densitometry of bands) compared to vector-transfected cells (Fig. 6B, lane 1). To determine if Sumoylation/deSumoylation affects the DNA-binding of Sp1 in vivo, we carried out a ChIP assay. mLECs were extrinsically transfected with different concentrations of pEGFP-Sumo1 or pFlag-Senp1. Chromatin samples were prepared as described in Materials and Methods. The Sp1 antibody or control rabbit IgG immunoprecipitated complex was processed and analyzed by qPCR using primers specific to promoter region as shown in Fig. 2A. Fig. 6D shows that Sumo1 overexpression significantly reduced the Sp1-DNA binding in dose-dependent manner with the endogenous Prdx6 gene, and this binding was increased in cells overexpressing Senp1 (Fig. 6E) in concen-tration-dependent fashion. There was no detectable Sp1-DNA binding with IgG control immunoprecipitated samples. Sumo1 over-expression significantly reduced the level of Sp1 ( Fig. 6B and 6D). In contrast, increased abundance of Sp1 protein was detected in cells overexpressed with Senp1 ( Fig. 6E). Taken together, data indicate that increased abundance of Sumo1 dysregulated Sp1 activity.

K16 residue of Sp1 was a target for Sumo1 conjugation in LECs, and an increased level of Sumo1 affected Sumoylation status of Sp1
Genes/proteins and their functions can differ in different cell types and cell backgrounds. Our initial aim was to confirm the findings of others that Sp1 is a substrate for Sumo1, and it is Sumoylated at K16 residue [23,32,34]. To determine whether Sp1 is indeed Sumoylated at K16 in hLECs in vivo as reported by others, we performed an immunoprecipitation assay using antibody specific to Sp1 with nuclear extract of hLECs overexpressing EGFP-Sumo1 plasmid. We observed that protein precipitated with anti-Sp1 monoclonal antibody, when immunoblotted with anti-Sp1 or anti-Sumo1 polyclonal antibodies as shown in Fig ). However, the samples loaded on SDS-Gel did not contain equal amounts of protein (Supplementary figures, (S1A and C), but, qualitatively, data showed that exogenous Sp1 was Sumoylated in LECs as reported previously by other investigators. To further assess the Sumoylated/deSumoylated status of Sp1, nuclear extract was used to perform Sp1 sandwich/ Sumo1-ELISA (with anti-Sp1 monoclonal and polyclonal antibodies). Sumoylated Sp1 was subtracted from the total Sp1 amount, and is presented as deSumoylated Sp1 (gray bars). This experiment revealed that the level of the Sumoylated form of Sp1 in pEGFP-Sumo1 transfected LECs was four times higher than in pEGFP-Vector transfected LECs (Fig. S1B).
Because identifying which specific lysine (K) residue of Sp1 is Sumoylated requires mutagenesis, it was necessary to determine whether exogenous Sp1 is Sumoylated in LECs. We co-transfected SRA-hLECs with either pEGFP-Vector plus pCl-neo-HA-Sp1 or pEGFP-Sumo1 plus pCl-neo-HA-Sp1. Cells were processed for immunoprecipitation (IP) using anti-HA monoclonal antibody. Probing immunoprecipitates with anti-HA or anti-Sumo1 polyclonal antibodies revealed a discrete slower migrating band in monoclonal HA-IPs from extracts co-expressing pEGFP-Sumo1 plus pClneo-HA-Sp1 (Fig. S1Ca,  (pCl-neo-HA-Sp1-K16R). These plasmids were used to transfect LECs along with pEGFP-Sumo1 as indicated in Fig. 7. In vivo Sumoylation was conducted as described in Materials and Methods. As shown in Fig. 7B, a Sumoylated band of pCl-neo-HA-Sp1 WT plus EGFP-Sumo1 (Fig. 7B, a and b, ~145kDa, lane 2) was observed with both anti-HA and anti-Sumo1 polyclonal antibodies, while no . 72h later ChIP assay was carried out with anti-Sp1 and control IgG antibodies. Pulled DNA fragments were subjected to PCR analysis for Sp1 binding cis-elements of Prdx6 promoter. The product was analyzed through agarose-gel. Data represent three experiments. (E) Senp1 overexpression dramatically enhanced Sp1-DNA binding in concentration-dependent -manner. mLECs were transfected with increasing concentrations of pFlag-Senp1 (0, 0.5, 1 and 2μg) for 72h. ChIP analysis was carried out using chromatin samples prepared from transfected LECs with a ChIP grade antibody, anti-Sp1 and control IgG. The DNA fragments were used as templates for PCR by using primers designed to amplify −208 to +27 region of the Prdx6 promoter bearing Sp1 binding sites as shown. PCR product was analyzed through agarose gel as shown.
AGING Sp1K16R mutated at Sumo1 site or Senp1 promoted Sp1-DNA binding, and overexpression of Sumo1 did not affect Sp1's DNA binding potential (Figs 1-6) suggest a possible role for oxidative stress and aberrant Sumoylation in dysfunction of Sp1-DNA binding activity. We posited that Sumo1-deficient Sp1K16R should have increased DNA binding activity, by escaping oxidative stress-induced aberrant Sumoylation. To confirm this, chromatin extracted from LECs co-expressing Sumo1 along with Sp1WT or its mutant (Sp1K16R) at Sumo1-motif (Fig. 8B) immunoprecipitated with IgG control or anti-HA monoclonal antibody was subjected to qPCR analysis using Prdx6 gene region specific primers denoted in Fig.3. Sumoylation-deficient Sp1 (Sp1K16R) showed significantly increased binding to its site present in Prdx6 gene in comparison to WT Sp1 (Fig. 8B). Sumo1 overexpression did not attenuate its DNA binding activity. (Fig. 8B, black bars). Similarly, we performed ChIP assay with deSumoylating agent Senp1 along with either Sp1WT or its mutant (Sp1K16R). As expected, LECs overexpressing Senp1 with WT Sp1 showed significantly higher Sp1-DNA binding activity, whereas Sp1K16R co-expression with Senp1 showed further increase in Sp1-DNA binding in comparison to Sp1 (Fig. 8C), indicating the presence of some minor Sumoylation sites in Sp1. However, data indicated that Sumoylation-deficient Sp1 or deSumoylation agent Senp1 enhanced the Sp1-DNA binding activity in Prdx6 gene promoter, and suggested the involvement of Sumoylation in the loss of Sp1 activity. The DNA fragments were used as templates for RT-qPCR by using primers designed to amplify −342 to +30 region of the Prdx6 gene promoter bearing Sp1 binding sites as shown. Histogram shows the amplified DNA through real-time qPCR analysis. 0 µg vs 2 µg and 4µg pEGFP-Sumo1, pCl-neo-HA-Sp1 WT vs pCl-neo-HA-Sp1K16R (*p < 0.001). (C) Senp1 overexpression showed increased Sp1 DNA binding of Sp1WT and comparable to Sumoylation deficient Sp1K16R. SRA-hLECs were cotransfected with pFlag-Senp1 with either pCl-neo-HA-Sp1 or mutant pCl-neo-HA-Sp1K16R as indicated. ChIP assay was conducted as described above using anti-HA antibody. Histograms represent the concentration dependence of Senp1-induced enrichment of Sp1 at its binding sites in Prdx6 gene promoter. 0 µg vs 0.5 µg and 1µg pFlag-Senp1, pCl-neo-HA-Sp1 WT vs pCl-neo-HA-Sp1K16R (**p<0.05; *p < 0.001).

Mutation of Sp1K16R at Sumoylation site enhanced its transactivation potential by stabilizing its cellular availability
Transactivation assay was performed to determine if the increased binding of Sumo1-deficient Sp1 (Fig. 8B) was functional and provided more transactional activity in activating transcription of its target gene, Prdx6. As shown in Fig. 9A, hLECs co-transfected with Sumo1deficient pClneo-HA-Sp1K16R (Fig. 9A, Black bars) along with Prdx6 promoter displayed significantly higher Prdx6 transcription compared to pClneo-HA-Sp1 WT (Fig. 9A, Dark gray bar). Next, we tested whether this enhanced transactivation potential was related to its AGING abundant availability to its GC-box response element(s) in Prdx6 promoter due to Sp1's increased stability. We analyzed the cellular stability of Sp1WT and its mutant, Sp1K16R at different time points as shown in Fig. 9B, by eliminating de novo protein synthesis with cycloheximide (CHX) treatment, a translational inhibitor. SRA-hLECs transiently transfected with pCl-neo-HA-Sp1WT or its mutant (pCl-neo-HA-Sp1K16R) were treated with CHX as indicated in Fig. 9B. We observed that Sumo1 deficient Sp1K16R was more stable than the Sp1WT; the remaining protein of Sp1WT and its mutant forms after treatment with CHX is also shown in the histogram (percentages) right side of the protein bands based on densitometry quantitation analysis. An analysis of data revealed that cellular abundance of mutant K16R proteins was significantly higher than that of Sp1WT protein with 10 μg/ml and 20 μg/ml treatment of CHX for 24h, 36h and 48h (Fig. 9Ba, Bb and Bc), suggesting that Sumoylation destabilized Sp1 by mediating its degradation [23,34,53]. We also observed that at a higher concentration of CHX (40μg/ml) for 8h, Sp1K16R mutated at Sumo1 site (Fig. 9Bd), showing greater stability compared to Sp1 WT. Similar results were obtained with mLECs (data not shown). We conclude that the decline in Sp1 abundance in cells might be due to changes in Sumoylated Sp1 stability (Fig. 9). , and then exposed to oxidative stress. ROS intensity (C) and cell viability (D) are presented as histograms. Values represent mean ± SD of three independent experiments. 0 vs 100µM H 2 O 2 and pEGFP-Sumo1 vs pEGFP-Sumo1 plus pCl-neo-HA-Sp1 WT vs pEGFP-Sumo1 plus pCl-neo-HA-Sp1K16R (**p<0.05; *p<0.001). Sumoylation-deficient Sp1K16R (black bars) showed significantly higher protection and reduced ROS production, indicating that mutant Sp1K16R was more effective in protecting cells from oxidative stress Sumoylation-mediated insults.

Sp1K16 R gained potential for protecting cells from oxidative and aberrant Sumoylation stresses
Next, we asked whether Sumoylation-deficient Sp1K16R had increased capacity to protect cells against oxidative or aberrant Sumoylation-mediated stresses. SRA-hLECs overexpressing pCl-neo-HA-Sp1WT or pCl-neo-HA-Sp1K16R were exposed to different concentrations of H 2 O 2 as indicated in Fig. 10A and 10B. Quantitation of ROS by CellROX Deep Red dye and cell viability by MTS assay revealed that pCl-neo-HA-Sp1K16R transfected cells showed significantly reduced ROS (Fig. 10A) and increased viability (Fig.  10B) compared to pCl-neo-HA-Sp1WT. We measured the synergistic effect of Sumo1 and oxidative stress on protective efficacy of Sp1K16R, as we assumed that this condition might occur in vivo in aging or aged cells.
The experiments were similar to those described above, but used cells overexpressing Sumo1. When assayed for ROS and cell viability, the transfectants bearing Sp1K16R were found to be highly efficient in reducing ROS (Fig. 10C), and engendered more resistance against oxidative and Sumo1-induced insults (Fig.  10D). Collectively, data suggest that Sp1K16R rescued the cells by blunting oxidative and increased Sumoylation mediated stresses.

Sumoylation-deficient Prdx6K122/142R fused to transduction domain (TAT) internalized in cells and lessened oxidative stress and its induced aberrant Sp1 Sumoylation
Results from the current study as well as mounting evidence from other reports including our own indicate AGING that oxidative stress is a primary culprit in malfunction or pathobiology of cells and tissues. Conceivably, a logical strategy for overcoming the adverse processes might be reducing intracellular ROS-driven oxidative stress and Sumoylation signaling by means of naturally occurring antioxidants such as Prdx6 K122/142R mutated at Sumo1 site [12]. We previously demonstrated that Prdx6 is Sumoylated at K122 and K142, and Sumoylation of Prdx6 reduces its protective potential, while mutant Prdx6 at K122/142R gains significant protective potential and escapes oxidative stress-induced aberrant Sumoylation signaling. To test whether Prdx6K122/142R protects cells against oxidative stress-induced deleterious signaling, we delivered Sumoylation-deficient Prdx6 linked with TAT trans-duction domain (TAT-HA-Prdx6K122/142R) [5,12] to Prdx6-deficient LECs subjected to oxidative stresses. As shown in Fig. 11, oxidative stress induced by H 2 O 2 (Fig.11A) or UVB (Fig.11B) significantly enhanced the Sp1 Sumoylation. Intriguingly, that increase in Sumoylation was significantly blunted in Prdx6 -/-LECs treated with mutant Prdx6 at K122/142R (Fig. 11A, open bars vs gray bars vs black bars) or UVB (Fig. 11B, open bars vs gray bars vs black bars) compared to TAT-HA-Prdx6WT as evidenced through Sp1/Sumo1-ELISA assay. These findings suggest that Sumoylation-deficient Prdx6 may efficiently protect cells by rescuing Sp1 from deleterious signaling evoked by oxidative stress and aberrant Sumoylation. Previous experiments in this study showed that expression of Prdx6 in LECs reduced the Sp1 Sumoylation caused by oxidative stress and protected those cells. We next asked whether delivery of Prdx6 linked to TAT transduction domain protected cells by augmenting the Sp1-DNA binding mechanism. Oxidative stress and its induced adverse signaling has been shown to dysregu-late transcriptional activity by damaging DNA binding activity and reducing expression of antioxidants, and antioxidant application is known to reverse the process [3,10,[54][55][56]. We performed ChIP assay using anti-Sp1 antibody. SRA-hLECs or aging LECs transduced with TAT-HA-Prdx6 WT and its mutant at K122/142R were subjected to oxidative stressors as described in Materials and Methods (Fig. 12). Chromatin samples prepared from cells were immunoprecipitated with anti-Sp1 antibody, and IP samples were analyzed with qPCR by using specific primers (Fig. 3A) designed for ChIP assay. As shown in Figure 12, cells pretreated with TAT-HA-Prdx6 WT (Fig. 12, gray bars) had significantly higher Sp1-DNA binding both with and without H 2 O 2 or UVB in comparison to untreated LECs (Open bars). On the other hand, we found significantly higher enrichment of Sp1 in LECs transduced with Sumoylation-deficient (Prdx6K122/142R, Black bars) in comparison to untreated (Open bars) and even WT Prxd6 (gray bars). Younger cells were more responsive than aged cells. As a whole, data demonstrated that, during aging and oxidative stress, application of Prdx6 in the form of Sumoylation-deficient TAT-Prdx6K122/142R is deliverable, internalizes in cells (Fig. 11C), and can protect cells by restoring DNA binding of Sp1.
Sp1 expression can be affected by various stimuli including oxidative stress and aging as well as via its interaction with other proteins [9,49,66]. Its activity can be highly modulated by various posttranslational modifications, such as Sumoylation, phosphorylation, glycosylation, acetylation and oxidation [27,34,68]. In the present study, we found that levels of Sp1 mRNA and protein as well as expression of its target gene, antioxidant protein Prdx6, declined with increases in ROS-driven oxidative load in aging LECs. The decline was linked to reduction of Sp1's DNA binding efficiency as well as reduced cellular abundance (Fig.1), indicating that repression of Prdx6 is at transcriptional levels. It is justifiable to mention that there was no disruption of transcriptional machinery during aging as there was no change in internal control (internal control gene or protein expression). We previously showed that aging hLECs, Prdx6 -/-(a model for aging) and aged trabecular meshwork (TM) cells bear significantly higher ROS expression levels than cells of younger age [10,16,69]. Ammendola reported that Sp1 binding is dysregulated in aging due to changes in redox environment with no altered expression of other molecules examined [54]. Moreover, this age-related adverse phenomenon is not limited to Sp1; several reports show that activity of transcription factors, such as Nrf2, declines with age and the decline is correctable [10,56]. Furthermore, aging and oxidative stress can directly modulate the induction of several genes encoding transcription factors, including Fos, Jun, Myc, Erg1, NF-кB, HSF1 [56,[70][71][72][73] . In the current study we also extended our previous finding(s) [5,12] to explore causes of the observed dysregulation of Sp1 and repression of Prdx6 gene during oxidative stress. We observed that Sp1-DNA binding was reduced to its response element(s) present in both mouse and human Prdx6 gene promoter in cells exposed to oxidative stress, and that the binding capacity was directly connected to intracellular oxidative load and reduction in Prdx6 expression (Fig. 2). ROS-evoked oxidative stress can modulate expression levels and DNA binding efficiency of many transcription factors that may lead to reduced or increased expression of their target genes [27,[74][75][76]. We observed that reduced Sp1 binding to Prdx6 promoter was connected to reduced expression of Prdx6 mRNA (without affecting internal control gene), implicating that transcriptional machinery is not influenced during oxidative stress. Nonetheless, we recognize that it is difficult to give a general statement about oxidative and aging stress-evoked deleterious signaling-mediated cell damage as they initiate multiple signaling. We believe that this requires further study to AGING discern the effects of aging or oxidative stress on epigenetic/genetic levels.
Oxidative-induced injurious signaling is known to be a prime cause of cellular insult during aging [77]. In humans, oxidative stress-induced increased ROS along with reduced antioxidant defense capacity is now believed to cause many age-related degenerative disorders, including cataractogenesis, neurodegenerative disorders, glaucoma, macular degeneration, cancer, diabetes and cardiovascular disease [2,5,12,31,69,78]. Intracellular ROS is tightly regulated by antioxidant defense proteins, to maintain correct physiological levels of ROS. Dysregulation of transactivators of antioxidant defense genes can lead to reduced expression of antioxidant protein, which may adversely affect physiological levels of ROS. Thus genetically allotted physiological control of the transcriptional machinery of defense genes, including Prdx6, is of utmost importance.
Oxidative stress has been reported to accelerate/decelerate Sumo1 conjugation to proteins via direct or indirect mechanisms, thereby modifying their functions [5,40,43,63,[79][80][81]. In the current study we found that Sp1 was endogenously Sumoylated and its Sumoylation status was dramatically increased with an increase of oxidative stress, as shown in Sumoylation experiments with Sp1 (Fig. 4). The aberrant Sp1 Sumoylation was linked to increased Sumo1 expression and reduced expression of Senp1. An excess of oxidative stress has been shown to aberrantly modulate Sumoylation process by increasing or decreasing proteins Sumoylation and their function(s) [5,12,[82][83][84]. This can happen due to inactivation of Sumo isopeptidases either by dimerization or overoxidation of their catalytic cysteine [5,23]. Our results revealed increased Sumoylation of Sp1 with increased oxidative stress and Sumo1 expression, which possibly occurred due to reduced expression and inactivation of Senp1, a deSumoylating enzyme for Sumo1 conjugation (Fig. 4), during aging, arguing that gene regulatory transcriptional state is active in aging and the condition can be remedied (as observed in Figs. 11 and 12). Other possibilities may exist: (i) increased expression of Sumo1 during oxidative stress that can result in increased abundance of free Sumo1 in cells ( [5] and current study, Fig. 5), and (ii) ROS directly or indirectly affecting Sumo E3 ligase activity leading to increased conjugation [79,85]. Alternatively, because oxidative stress modifies the phosphorylation status of proteins including Sp1, it may in turn modify Sumoylation of Sp1. Among several specific substrates for Sumos that have been found to be affected in response to oxidative stress are Sp1, NF-кB, LEDGF, HIPK2, TP53INP1 and Prdx6 [5,12,23,34,60,83,[86][87][88]. Moreover, in studying the molecular effect of oxidative-induced aberrant Sumoylation of Sp1 on Sp1-DNA binding activity, both in vitro and in vivo DNA-binding experiments with Sp1 to its GC-response element in Prdx6 promoter revealed significant dysregulation of Sp1-binding in cells overexpressing Sumo1. In contrast, cells overexpressed with Senp1 displayed increased Sp1-binding. This reduction in Sp1 binding indicates an involvement of Sumo1 that plays negatively in Sp1-DNA binding activity as shown in Fig. 6. The data argue that the repression of Prdx6 in cells may be a cause of reduced Sp1 DNA binding to its response element in Prdx6 promoter in response to oxidativeinduced increased Sumoylation signaling that influences Sp1 cellular abundance and thereby reduces enrichment to its binding sites. The reduction in Sp1-DNA interaction appears to have two possible causes: (i) Sp1 became a target for quick degradation due to aberrant Sumoylation [32,79,83] Fig. 7 and Supplementary Figs. S1 and S2), a finding consistent with reports by others [23,32,34]. We called this construct Sumo1-deficient Sp1K16R and tested its transactivation potential. Fig. 8 shows that DNA binding activity of Sumo1-deficient Sp1 significantly increased in Prdx6 promoter even in cells overexpressing Sumo1. In contrast, Sp1WT activity was dramatically attenuated, indicating the role of Sumoylation in reduction of Sp1 activity. The data were further supported when cells overexpressing Senp1 showed restoration of the DNA-binding activity of Sp1 (Fig. 8B). While testing the functionality of their DNAbinding activity in transactivation experiments, we found that Sumoylation-deficient Sp1 had greater transactivation activity than its counterpart, Sp1WT AGING (Fig. 9A). This elevated activation potential of Sumo1deficient Sp1 was connected with its enhanced cellular abundance due to increased stability as evidenced by cellular stability assay (Fig. 9B). One of the most salient observations of the present study was that LECs overexpressing Sumoylation-deficient Sp1 gained resistance against oxidative stress and aberrant Sumoylation-mediated injurious signaling, and those cells survived better (Fig. 10). The improved condition of the Sumoylation-deficient Sp1 was due to its escaping the aberrant Sumoylation processes and thereby activating Prdx6 ( [5,9] and Fig. 9). Posttranslational modification of Sp1 has been shown to influence its transcription activity and stability [12,25,26,34,35]. Glycosylation of Sp1 can stimulate or repress DNA binding and transcription [89]. In basal condition Sp1 is Sumoylated at N-terminal, which negatively regulates Sp1 transcription activity [32]. However, Sumo1 conjugation of Sp1 has been reported to enhance its activity [53]. This discrepancy may be related to experimental conditions. Our results and other reports clearly show that Sumoylation of Sp1 leads to its dysregulation [11,32,35,90]. Furthermore, H 2 O 2 has been shown to decrease NOS-3 promoter activity by reducing the binding capacity of Sp1 [91].
It was intriguing to observe that the level of Sp1 Sumoylation was significantly higher in redox active Prdx6 -/cells and in aging/aged cells, and that these were much more susceptible to oxidative stress than were Prdx6 +/+ cells (Fig. 4) and younger hLECs (data not shown). This suggests a pivotal role for Prdx6 in abating oxidative stress-mediated aberrant Sumoylation signaling, at least in eye lens. Previous studies [5,12,43], as well as the current research show that oxidative stress and oxidative stress-induced overstimulation of Sumoylation pathways are a major culprit in initiating etiopathobiology of cells or cell injuries. We think that damaging oxidative load in aging cells or cells in redoxactive stage can be remedied by enhancing the natural defense via providing an extrinsic supply of antioxidant, like Prdx6, capable of optimizing oxidative load [5,9,12]. Indeed, we found that delivery of Prdx6 reduced the UVB and H 2 O 2 -induced Sp1 Sumoylation in Prdx6 -/-LECs (Fig. 11), restored Sp1-DNA binding activity (Fig. 12), and reactivated Sp1-DNA binding to Prdx6 promoter in aging hLECs (Fig. 12C). These data indicate the potential for Sumoylation-deficient Prdx6 [12] to be used as a therapeutic molecule to block oxidative-induced aberrant Sumoylation-driven pathogenic signaling.
In conclusion, we have described a novel mechanismoxidative-evoked aberrant Sumoylation signalingwhich dysregulates Sp1 and its target survival genes such as Prdx6 in eye lens cells affected by aging or oxidative stress. Taken together, findings of this study show that oxidative stress-evoked aberrant Sumoylation signaling attenuates Sp1 transactivation ability by aberrant Sumoylation of Sp1; lessening cellular stability and availability of Sp1 to GC-response elements in the target genes. Additionally, we found that Sumoylationdeficient Sp1 resisted aberrant Sumoylation processes induced by oxidative stress, further demonstrating the prime role of oxidative stress and its associated aberrant Sumoylation-mediated Sp1 dysregulation. Because adverse signaling is driven by oxidative stress and reduction in antioxidants, application of Sumoylationdeficient Prdx6 [12] may be considered to block or delay the oxidative and aberrant Sumoylation-mediated injurious signaling by reactivating survival transcription factors like Sp1.

Cell culture
Human LECs used were of two types: (i) a cell line (SRA01/04) immortalized with SV40, and (ii) primary human LECs isolated from deceased persons of different ages. To avoid confusion, the remaining text will designate the immortalized LECs as SRA-hLECs, and the primary human (h) LECs as primary hLECs or hLECs.
The SRA-hLECs were derived from 12 infants who underwent surgery for retinopathy of prematurity [92] (a kind gift of Dr. Venkat N. Reddy, Eye Research Institute, Oakland University, Rochester, MI, U.S.A). These cells were maintained in DMEM with 15% FBS, 100µg/ml streptomycin, and 100µg/ml penicillin in 5% CO 2 environment at 37°C as described previously [5,93]. AGING dishes precoated with collagen IV containing a minimum amount of DMEM containing 15-20% fetal bovine serum (FBS), with a brief modification [16,[94][95][96]. Capsules were spread by forceps with cell layers upward on the surface of plastic petri dishes. Culture explants were trypsinized and re-cultured. Cell cultures attaining 90 to 100 percent confluence were trypsinized and used for experiments [11,93,97]. Western analysis was used to validate the presence of αA-crystallin, a specific marker for LEC identity (data not shown). For the experiments, SRA-hLECs and/or hLECs were cultured in 96, 24, 48, 6 well plates, or 60 and 100 mm petri dishes according to the specific requirements of each experiment.

Quantitation of intracellular ROS level by H2-DCF-DA and CellROX deep red reagent
Intracellular ROS level was measured by use of fluorescent dye dichlorofluorescin diacetate (H 2 DCFDA), a nonpolar compound that is converted into a polar derivative (dichlorofluorescein) by cellular esterase after incorporation into cells [16]. On the day of the experiment, the medium was replaced with Hank's solution containing 10 µM H 2 DCFDA dye and cells were incubated. Following 30 min later, intracellular fluorescence was detected with excitation at 485 nm and emission at 530 nm by a Spectra Max Gemini EM (Mol. Devices, Sunnyvale, CA, USA).
ROS level were measure according to the company's protocol (CellROX Deep Red Oxidative Stress Reagent, Catalog No. C10422, Thermo Scientific, Carlsbad, CA, USA). In brief, LECs (5 × 10 3 ) transfected with pEGFP-Vector or pEGFP-Sumo1 with pCl-neo-HA-Sp1 or pClneo-HA-Sp1K16R cultured in 96-well plate, 48 h later cells were exposed with different concentration of H 2 O 2 . After 8h, CellROX deep red reagent was added with final concentration of 5μM and cells were incubated at 37°C for 30 min. Media containing CellROX deep red reagent were removed and fixed with 3.7% formaldehyde. After 15 min, fluorescence signal were measured at Ex640 nm/ Em665 nm [12].

Transcription factor activation assay
Sp1 activation assay was performed according to manufacturer's protocol (TransAM Sp1 Transcription Factor Assay Kit, Cat No 41296, Active motif, Carlsland, California, USA). In brief, 10µg of nuclear extract (up to 10µl diluted with complete lysis buffer) prepared from aging hLECs added to the strips well, following the addition of 40µl complete binding buffer contains 20pmol of the wild-type and/or mutated consensus oligonucleotide to each sample well. For blank well, 10µl of complete lysis buffer were used. The plate was incubated for 1h at room temperature (RT) with mild agitation. 100µl primary antibody (1:1000 in 1X antibody binding buffer) added after 3 wash with 1X washing buffer and incubated at RT for 1h without agitation. 100µl of diluted anti-rabbit HRPconjugated antibody (1:1000 dilution in 1X antibody binding buffer) after three washing was added and incubated for 1h at RT. 100µl of developing solution was added to wells after four washing and incubated at RT in dark for 2 to 10 min. Finally, by addition of 100µl of stop solution, OD was recorded at 450nm.
As noted above, the same cell extracts from transfectants or controls were used and Sumoylated Sp1 or Prdx6 or HA-Sp1 was detected by ELISA using an EpiQuik in vivo universal protein Sumoylation assay kit (Epigentek, Farmingdale, NY, USA). Equal amounts of proteins from the total or nuclear extracts were added to the strip wells, which were precoated overnight with either anti-Sp1 or anti-Prdx6 or anti-HA serum. They were then incubated in blocking buffer for 45 min, washed three times and incubated with Sumo assay buffer for 1h at room temperature. After three washes, Sumo1 antibody was added and the proteins were incubated for 15 min at room temperature. Subsequent to color development by a Sumo detection system, OD 450 nm was measured an ELISA plate reader. To obtain deSumoylated Sp1 or Prdx6 or HA-Sp1, we calculated total and Sumoylated Sp1 or Prdx6 or HA-Sp1 protein and subtracted the Sumoylated Sp1 or Prdx6 or HASp1 protein from total Prdx6 protein.

Generation and validation of LECs isolated from lenses of Prdx6 −/− and Prdx6 +/+ mice
All animal experiments followed the recommendations set forth in the "Statement for the Use of Animals in Ophthalmic and Visual Research" by the Association for Research in Vision and Ophthalmology (ARVO). The University of Nebraska Medical Center (UNMC) approved animal studies. LECs isolated from Prdx6targeted mutants (Prdx6 −/− ) and wild type (Prdx6 +/+ ) mice were generated and maintained in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (FBS), as described earlier [33]. We used Prdx6 −/− mutant and Prdx6 +/+ C57/B6, mice of the same sex and age. All animals were maintained under specific pathogen-free conditions in an animal facility. LECs were isolated from mice of identical age, and Western analysis was carried out to confirm the presence of αA-crystallin, a specific marker of LECs.

Cotransfection and promoter activity assay
The CAT assay was performed using a CAT-ELISA kit (Roche Diagnostic Corporation, Indianapolis, IN, USA). Cells were transfected with reporter construct AGING (pCAT-Prdx6), and treated with different concentrations of Sp1 inhibitor, Mithramycin A (Cat. No. M6891, Sigma Aldrich, St. Louis, MO, USA). After 72 h of incubation, cells were harvested, extracts were prepared and protein was normalized. CAT-ELISA was performed to monitor CAT activity in accordance with the manufacturer's instructions. A 405 was measured using a microliter plate ELISA reader. Transactivation activities were adjusted for transfection efficiencies using GFP values cotransfected during transfection assays.

Induction of ultraviolet (UV) B induced stress
For UVB treatment, mLECs and SRA-hLECs were precultured for 16h in 100mm petri dishes with DMEM-10% and 15% FBS. The medium was replaced with phosphate buffered saline (PBS, pH 7.2) and the plates containing the monolayers were exposed to UVB using UV-lamp emitting 270-320 nm peaking at 302 nm wavelength (UVP, Upland, CA, USA). The energy actually incident onto the working area was measured by a UVX Radiometer (UVP Inc., Upland, CA) and expressed in J/m 2 . The UV dosage of J/m 2 was selected on the basis of results from our previous work [19]. After irradiation, PBS was withdrawn and fresh medium was added. At different time points protein was isolated and processed for Sandwich and Sumo1 specific ELISA assay to measure total Sp1 and Sumoylated Sp1 levels, respectively.
SRA-hLECs were cotransfected with pCl-neo-HA Sp1 WT or pCl-neo-HA Sp1 K16R along with either pEGFP-Vector or pEGFP-Sumo1. 48h later, the medium was replaced with phosphate buffered saline (PBS, pH 7.2) and the plates containing the monolayers were exposed to UVB as indicated. At intervals of 8 and 24 h later, ROS and MTS assays were performed to monitor the levels of ROS and cell viability, and the percentage of ROS and cell survival levels were then calculated for each group.
For another set of experiments, Prdx6 -/-mLECs and SRA-hLECs were pre-cultured for 16h in 100mm petri dishes with DMEM-10% and 15% FBS. Cells were washed with PBS and transduced with TAT-HA-Prdx6 WT or its mutant TAT-HA-Prdx6 K122/142R for 3h; the medium was replaced with phosphate buffered saline (PBS, pH 7.2), and the plates containing the monolayers were exposed to UVB as indicated. After irradiation, PBS was withdrawn and fresh medium was added. Similar treatment was repeated for 2-3 days as indicated in the figures.

Induction of hydrogen peroxide (H 2 O 2 ) induced stress
For H 2 O 2 treatment, mLECs and SRA-hLECs were precultured for 16h in 100mm petri dishes with DMEM-10% and 15% FBS. Cells were washed twice with PBS, and the medium (0.2% BSA+ DMEM) was replaced with predefined concentrations of H 2 O 2 . At different time intervals protein was isolated and processed for Sandwich and Sumo1 specific ELISA assay to measure total Sp1 and Sumoylated Sp1 levels, respectively. SRA-hLECs were cotransfected with pCl-neo-HA Sp1 WT or pCl-neo-HA Sp1 K16R along with either pEGFP-Vector or pEGFP-Sumo1. 48h later, the medium (0.2% BSA+ DMEM) was replaced with predefined concentrations of H 2 O 2 . At intervals of 8 and 24h later, ROS and MTS assays were performed to monitor the levels of ROS and cell viability, and the percentage of ROS and cell survival levels were calculated for each group.
For another set of experiments Prdx6 -/-mLECs and SRA-hLECs were pre-cultured for 16h in 100mm petri dishes with DMEM-10% and 15% FBS. Cells were washed with PBS and transduced with TAT-HA-Prdx6 WT or its mutant TAT-HA-Prdx6 K122/142R for 3h, and the medium (0.2% BSA+ DMEM) was replaced with predefined concentrations of H 2 O 2 . Similar treatment was repeated for 2-3 days as indicated in figures.

Extraction of nuclear and cytosolic fraction
Nuclear extract was prepared following the method of Sambrook et al. (18) with certain modifications. Briefly, LECs (1 X 10 6 ) were cultured in 100-mm plates. The cells were washed gently with chilled phosphatebuffered saline (pH 7.4). Cells were collected by centrifugation using a micro-centrifuge and resuspended in 5 pellet volumes of cytoplasmic extract buffer [(10 mM HEPES (adjusted pH at 7.9), 10mM KCl, 0.1 mM EDTA, 0.4% (v/v) Nonidet P-40, 0.5mM phenylmethylsulfonyl fluoride (PMSF), 1mM DTT and Protease inhibitor]. After a short incubation on ice and centrifugation (4ºC) at 10000 rpm for 10min, the cytoplasmic extract was transfer in fresh tube from the pellet. Following careful washing with cytoplasmic extract without detergent (Nonidet P-40), the fragile nuclei were resuspended in nuclear extract buffer [(20 mM HEPES (adjusted pH at 7.9), 0.4M NaCl, 1mM EDTA, 10% (v/v) glycerol, 1mM DTT, 0.5mM PMSF and Protease Inhibitor] and incubated 2h at 4ºC continuous vortexing. Finally, the extract was spin at 14,000 rpm for 15 min to pellet the nuclei. After centri-fugation, the nuclear extract was transferred and aliquoted in fresh tubes, and individual aliquots were stored at −70 °C to avoid repeated freezing and thawing of the preparation. Protein was estimated according to the Bradford protein assay and/or Pierce TM BCA Protein assay methods and extract was used for experiment as required.

Plasmids or constructs detail
Construction of pEGFP-Sumo1: For eukaryotic expression, the full length of Sumo1 cDNAwas subcloned into pEGFP-C1 vector. The coding region of Sumo1 was amplified by PCR from human lens cDNA library using forward (5'-CCGTCGACATGTCTGACC AGGAG-3') and reverse primer (5'-TCGGATCCGTT TTGAACACCACA-3') with restriction enzyme sites, SalI and BamHI. The PCR product was digested and ligated into pEGFP vector.

Site-directed mutagenesis (SDM)
PCR-based site-directed mutagenesis was carried out using the QuikChange TM lightning site-directed mutagenesis kit (Agilent Technologies; Catalog No. 210518), following the company's protocol. Briefly, amino acid exchanges K16R were generated by point mutations in the pCl-neo-HA-Sp1 constructs. The following complementary primers, forward primer: 5′-GCTGTGGTGAGGATTGAAAAAGGAGTTGGTGGC -3′ and reverse primer: 5'-GCCACCAACTCCTTTTTA AATCCTCACCACAGC-3' were used (changed nucleotides are in boldface type and underlined).
Epicurean Coli XL1-Blue super-competent cells (Stratagene) were transformed with resultant plasmid. The plasmid was amplified, and the mutation was confirmed by sequencing as described previously [101].

In vivo Sumoylation assay
SRA-hLECs was transfected as indicated in figures. After 48h, nuclear extracts (as mentioned above) or total cell lysates were prepared in IP lysis/wash buffer (0.025 m Tris, 0.15 m NaCl, 0.001 m EDTA, 1% NP-40, 5% glycerol, pH 7.4 plus 5 μm MG132 and 30 μm N-ethylmaleimide was added, as provided in the Pierce Classic IP Kit (catalog number 26146; Pierce, Rockford, IL, USA), in accordance with the manufacturer's instructions. Nuclear extracts diluted in lysis buffer or total cell lysates were incubated with 4μg of anti-Sp1 (sc-17824, Santa Cruz)/anti-HA (ab18181, Abcam) monoclonal serum/800 μg of protein in IP lysis buffer provided in the IP kit (Pierce) and were rotated at 4°C overnight. That was followed by the addition of 20 μL of Protein A/G plus Agarose beads and further rotation for 4h at 4 °C. The immunoprecipitates were collected by centrifugation and washed several times with wash buffer and 1X conditioning buffer before being boiled in SDS sample buffer. Next, 10% input and IP samples were resolved on 4-20% SDS/PAGE and analyzed by western blotting using anti-Sp1, anti-Sumo1 or anti-HA rabbit polyclonal antibodies.

Cycloheximide (CHX), a translational blocker treatment
To inhibit translation/ protein synthesis, transfected cells as indicated were treated with 0-40 μg/ml CHX for 0h-48h. CHX inhibitor (Catalog no. C4859) was purchased from Sigma-Aldrich. On the day of termination of experiment, total cell lysate prepared and immunoblotted with specific antibodies as indicated in figure and legends.

Protein expression analysis
Cell lysates of LECs were prepared in ice-cold radioimmune precipitation buffer and protein blot analysis was performed as described previously [46,103,104]. The membranes were probed with anti-HA
The PCR products were purified by preparative agarose gel electrophoresis. The purified products were ligated into a TA-cloning vector (Invitrogen) and then transformed into a competent cell and the plasmids of selected colonies were purified. The purified TA vector containing Prdx6 cDNA was digested with NcoI and EcoRI and then subcloned into a pTAT-HA expression vector (a kind gift of Dr S. F. Dowdy, Howard Hughes Medical Institute, University of California, San Diego, CA) that had been digested with the same restriction enzymes. Wild type (WT) TAT-HA-Prdx6 was then mutated at K (lysine) 122/142 (arginine) R by using SDM kit. Recombinant protein was purified using QIAexpress ® Ni-NTA Fast Start kit column (Qiagen Inc., Valencia, CA, USA). The host Escherichia coli BL21 (DE3) was transformed with pTAT-HA-Prdx6 and the transformants were selected on a LB plate with ampicillin. The selected colonies were cultured in 10 mL of LB medium containing ampicillin at 37 °C with shaking at 200 r.p.m. overnight. After incubation, 10 mL of the overnight cultures were combined with 250 mL of prewarmed media (with ampicillin) and then grown at 37 °C with vigorous shaking until D 600 of 0.6-0.8 was reached, and then isopropyl thio-β-d-galactoside was added to a concentration of 1 mm and the incubation was continued for 4-5 h. Cells were harvested by centrifugation at 4000 g for 20 min. Pellets were suspended in 10 mL of lysis buffer (50 mm NaH 2 PO 4 , 50 mm NaCl and 10 mm imidazole, pH 8.0) containing lysozyme and Benzonase ® Nuclease (Qiagen Inc., Valencia, CA, USA) and incubated for 30 min on ice. The suspension was then centrifuged at 14 000 g for 30 min. Supernatant was added to the Ni-NTA fast start column and allowed to drain before washing twice with 4 mL of wash buffer (50 mm NaH 2 PO 4 , 50 mm NaCl and 20 mm imidazole, pH 8.0), followed by elution with 1 mL of elution buffer (50 mm NaH 2 PO 4 , 50 mm NaCl and 250 mm imidazole, pH 8.0).

Statistical methods
For all quantitative data collected, statistical analysis was conducted by Student's t test and /or one-way ANOVA when appropriate, and was presented as mean ± S.D. of the indicated number of experiments. A significant difference between control and treatment group was defined as P value of < 0.05 and 0.001 for two or more independent experiments.

AUTHOR CONTRIBUTIONS
DP and BC conceived and designed the experiments. BC, PS and EK performed the experiments, and DP, BC and EK did the data analyses. BC and DP wrote the main manuscript text. DP supervised the study and provided the fund. All authors read and provided inputs on the manuscript, and approved the final version.

CONFLICTS OF INTEREST
All authors declare no competing financial interests.

FUNDING
This study was supported by the National Eye Institute, NIH (EY024589) (to DPS) grant and Research to Prevent Blindness (R.P.B.) are gratefully acknowledged.