let-7-repressesed Shc translation delays replicative senescence

The p66Shc adaptor protein is an important regulator of lifespan in mammals, but the mechanisms responsible are still unclear. Here, we show that expression of p66Shc, p52Shc, and p46Shc is regulated at the post-transcriptional level by the microRNA let-7a. The levels of let-7a correlated inversely with the levels of Shc proteins without affecting Shc mRNA levels. We identified ‘seedless’ let-7a interaction elements in the coding region of Shc mRNA; mutation of the ‘seedless’ interaction sites abolished the regulation of Shc by let-7a. Our results further revealed that repression of Shc expression by let-7a delays senescence of human diploid fibroblasts (HDFs). In sum, our findings link let-7a abundance to the expression of p66Shc, which in turn controls the replicative lifespan of HDFs.


Introduction
Animal longevity is controlled by multiple molecular mechanisms, involving genes such as the silent information regulator SIRT1 (Bordone & Guarente, 2005), superoxide dismutases (SOD1 and SOD2; Fabrizio et al., 2003;Parker et al., 2004), p66Shc (Migliaccio et al., 1999;Pinton & Rizzuto, 2008), as well as insulin and insulin-like growth factor-1 (IGF-1; Yang et al., 2005). p66Shc is one of the members of the Shc family of proteins consisting of three isoforms (p66Shc, p52Shc, and p46Shc) that arise through alternative initiation of translation (Luzi et al., 2000). p52Shc and p46Shc were found in every cell type with invariant reciprocal relationship, while p66Shc expression varies in different cell type, suggesting that the function of p66Shc may be different from those of p52Shc/p46Shc (Migliaccio et al., 1997). Initially, all three isoforms of Shc were recognized as 'adaptor' proteins forming a complex with Grb2, an adaptor protein in the Ras signaling pathway. However, unlike p52Shc and p46Shc, p66Shc has little effect on the Ras signal pathway (Migliaccio et al., 1997). Instead, p66Shc has been described as a regulator for the longevity of mammals. p66Shc À/À mice exhibited longer lifespan that was probably due to decreased cardiac diseases and reduced reactive oxygen species (ROS) production (Migliaccio et al., 1999;Napoli et al., 2003). However, the role of p66Shc in regulating other age-related processes has not been elucidated.
In response to oxidative stress, part of the cytosolic p66Shc translocates to the mitochondria and acts as an oxidoreductase. Phosphorylation at serine 36 by PKC is essential for its translocation, because a serine-phosphorylation defective mutant of p66Shc cannot restore the normal stress response in p66Shc À/À cells (Migliaccio et al., 1999). On the other hand, elevation of p66Shc protein has been observed in primary human prostate tumors (Veeramani et al., 2005), in replicative senescence (Zhang et al. 2010), in middle-aged mice (Lebiedzinska et al., 2009), as well as in cells exposed to oxidative stress and UVC (Favetta et al., 2004). Regulation of p66Shc at transcriptional level by p53 and DNA methylation has been described (Trinei et al., 2002;Ventura et al., 2002). However, the regulation of p66Shc in replicative senescence remains largely unknown.
In this study, we show that microRNA let-7a regulates lifespan of human diploid fibroblasts by repressing the translation of p66Shc. We describe that let-7a interacts with 'seedless' sites located in the coding region (CR) of p66Shc mRNA, prevents the association p66Shc mRNA with the polysome, and enhances the recruitment of p66Shc mRNA into PBs, thereby repressing the translation of p66Shc. These studies reveal a novel microRNA-mediated mechanism linking p66Shc and cellular lifespan.
Results let-7 represses the translation of p66Shc, p52Shc, and p46Shc This study was initiated from our findings that intervention of the let-7a levels altered the levels of p66Shc, p52Shc, and p46Shc proteins. Western blot analysis revealed that overexpression of let-7a by transfecting a vector that expressed pre-let-7a in IDH4 cells reduced p66Shc, p52Shc, and p46Shc proteins by~50-70% (Fig. 1A), while knockdown of let-7a by transfecting a vector expressing let-7a antisense (AS-let-7a) increased Shc proteins by~2.3-to 5.6-fold (Fig. 1B). IDH4 cells are derived from senescent IMR-90 cells, but through constitutive, dexamethasone (dex)-driven SV40 large T-antigen, IDH4 cells can proliferate in culture; upon dex removal from the medium, cells rapidly return to senescence (Wright et al., 1989). In contrast, neither knockdown nor overexpression of miR-30 substantially altered the expression of p66Shc, p52Shc, and p46Shc. To further address the mechanism by which let-7a regulates the expression of Shc proteins, the levels of p66Shc mRNA in cells described in Fig. 1(A,B) were analyzed by reverse-transcription (RT) followed by real-time, quantitative (q)PCR analysis. p52Shc and p46Shc were translated from the same transcript as p66Shc (p66Shc mRNA), by alterative initiation of translation (Fig. S1). The levels of p66Shc mRNA, which could potentially be used for synthesis of all Shc proteins, were not substantially altered by modulating let-7a or miR-30 abundance (Fig. 1C), suggesting that let-7a does not affect Shc expression at the level of mRNA turnover and instead may affect Shc translation. To test this hypothesis, IDH4 cells transiently expressing antisense let-7a or antisense miR-30 were incubated in medium containing L-[ 35 S] methionine and L-[ 35 S] cysteine for 20 min, cell lysates were then prepared and subjected to immunoprecipitation to analyze the level of nascent Shc proteins. As shown in Fig. 1(D), nascent Shc protein synthesis in AS-let-7a-expressing cells was~3.2to 4.1-fold higher than what was observed in control cells, while Shc translation in AS-miR-30-expressing cells was comparable with that measured in control cells. In control reactions, knockdown of let-7a or miR-30 did not influence the levels of nascent GAPDH. These results support the idea that let-7a represses the translation of p66Shc, p52Shc, and p46Shc.
To further investigate the repression of p66Shc, p52Shc, and p46Shc by let-7a, we examined the involvement of the Ago2containing RNA/microRNA-induced silencing complex (RISC). HeLa cells were transfected with siRNA targeting Ago2 or let-7a, or cotransfected with both siRNAs, whereupon the levels of Shc proteins were assessed by Western blot analysis. As shown in Fig. 1(E), knockdown of Ago2 and let-7a increased the levels of Shc proteins by~2.2-3.8-and~2.5-3.9-fold, respectively, while simultaneous knockdown of Ago2 and let-7a did not show further effect of elevating Shc protein levels. These results support the view that let-7a represses the translation of p66Shc, p52Shc, and p46Shc in an Ago2/ RISC-dependent manner.
We also assessed the expression of Shc proteins in cells transfecting with the siRNA or inhibitor of let-7a. As shown in the Fig. S2(A), transfection of IDH4 cells with let-7a siRNA, but not miR-30 siRNA, elevated the levels of Shc proteins. Transfection of IDH4 cells with an inhibitor of let-7a increased the levels of Shc proteins; transfection of cells with let-7b inhibitor moderately induced the Shc protein levels (Fig.  S2B). In addition, transfection of cells with inhibitor of let-7c, let7d, miR-9, miR-22, or miR-30 did not alter the levels of Shc proteins (Fig. S2B,C). These results confirmed that let-7a specifically represses the translation of p66Shc, p52Shc, and p46Shc.
The presence of an individual mRNA in polysomes and processing (P) bodies are indicators of the efficiency of translation of that mRNA. We therefore tested the presence of p66Shc mRNA in polysomes and p-bodies of cells with silenced let-7a. As shown in Fig. S3, knockdown of let-7a increased the presence of p66Shc mRNA in the polysome fraction while it reduced the presence of MS2-CR chimeric transcripts in P-bodies. . Forty-eight hour after transfection, Western blotting analysis was performed to assess the protein levels of Ago2, p66Shc, and GAPDH. All Western blotting data are representative from three independent experiments. The RT-qPCR data represent mean AE SD from three independent experiments. let-7 regulates p66Shc, F. Xu et al.

Reduction of let-7a is accompanied by elevation of Shc proteins in replicative senescence
Next, we assessed the levels of Shc proteins in early-passage (Young, PDL 27), middle-passage (middle,~PDL 37), and late-passage (Senescent,~PDL 57) human diploid fibroblasts (2BS). As shown in Fig. 3(A), Shc proteins were almost undetectable in Young 2BS cells, but they increased in middle-passage cells and reached highest levels in senescent cells; similarly, Shc proteins were also higher in senescent IDH4 cells (Fig. 3B). The activity of p38MAPK (phospho-p38, p-p38) and the levels of p16 increased with cellular senescence (Fig. 3A,B). let-7 was found to suppress the expression of Ras, c-myc, E2F1, and CDC34, as well as to elevate the expression of p21. However, both Ras and c-myc were undetectable in either 2BS or IDH4 cells (data not shown); the levels of CDC34 decreased~80% in senescent 2BS and~70% in senescent IDH4 cells; the levels of p21 increased in middle-passage (~6.5-fold) and latepassage (~3.5-fold) 2BS cells as well as in senescent IDH4 cells (~3.3-fold; Fig. 3A,B). Although the levels of E2F1 remained unchanged in senescent 2BS cells, a remarkable reduction of E2F1 was observed in senescent IDH4 cells (~90%). In agreement with the findings that p66Shc is implicated in the production of intracellular ROS (Migliaccio et al., 1999;Napoli et al., 2003), ROS levels were significantly increased with the senescent process of both 2BS (P = 0.0175) and IDH4 cells (P = 0.0072; Fig. 3C).
Using the cells described in Fig. 3(A,B), the levels of let-7a, miR-30, U6, as well as p66Shc mRNA levels in 2BS and IDH4 cells progressing toward senescence were determined by Northern blot analysis (Fig. 3D) and by conventional RT-PCR analysis (Fig. 3E) respectively. In agreement with previous findings (Marasa et al., 2010), the levels of let-7a were reduced in middle-passage (~60%) and senescent (~80%) 2BS cells as well as in senescent IDH4 cells (~90%). In contrast, the levels of miR-30, U6, and p66Shc mRNA remained unchanged during senescence of 2BS ( Fig. 3D) and IDH4 cells (Fig. 3E). These results suggest that the reduction of let-7a may contribute to the elevation of p66Shc, but not to the alterations of CDC34, E2F1, or p21 during replicative senescence.

Discussion
The findings presented in this study support the view that p66Shc negatively regulates the lifespan of human diploid fibroblasts (Fig. 4), in keeping with the finding that p66Shc deletion extends longevity in mammals (Migliaccio et al., 1999). We have found that the elevation of Shc proteins in replicative senescence is regulated at post-transcriptional levels by let-7a (Figs 1-3, 5, 6 and S2). By repressing the translation of Shc genes, let-7a extends the lifespan of human diploid fibroblasts (Figs 1, 5 and 6). Although the specific mechanisms by which let-7a represses the translation of Shc proteins are not fully understood, our data suggest that the association of let-7a with the seedless interaction elements located in the CR of Shc mRNA (Fig. 2C) lowers the presence of Shc mRNA in polysomes and enhances the recruitment of Shc mRNA into P-bodies (Fig.  S3). The effect of let-7a was specific, as inhibition of let7c, let-7d, miR-9, miR-22, and miR-30 did not affect the levels of Shc proteins, while inhibition of let-7b moderately increased the levels of Shc proteins (Fig.  S2). The let-7a-Shc regulatory process contributes at least in part to the elevation of Shc proteins in senescent cells, because expression of the Shc CR fragment antagonizes the effect of let-7a in regulating either Shc expression or cellular lifespan (Fig. S7). Given that the expression of Shc proteins in senescent cells is robustly elevated (Fig. 3) when let-7a levels decline, the let-7a-Shc regulatory paradigm may be in part responsible for the elevation of Shc proteins in replicative senescence.
P66Shc, p52Shc, and p46Shc are adaptors for the Ras-MAPK signaling pathway. In 2BS cells, p38 is activated by overexpression of p66Shc, p52Shc, or p46Shc, or by knockdown of let-7a, while p38 is inhibited by p66Shc knockdown or by let-7a overexpression ( pronounced in the presence of active p66Shc, activation of p52Shc and p46Shc also modestly elevated ROS levels and elevated the senescenceassociated b-galactosidase activity (Figs S5B and S6B). Given that overexpression of p66Shc and that let-7a lowered the levels of all three Shc isoforms (Figs S4 and 1), it is possible that p52Shc and p46Shc also regulate replicative senescence. In addition, other let-7-regulated mRNAs may also encode proteins involved in preventing senescence. Further work is needed to identify the complete set of effectors through which of let-7a impacts upon the cellular replicative lifespan. MicroRNA-mediated gene regulation typically occurs through the interaction of the microRNA with the 3′UTR of the target mRNA (Bartel, 2009). However, analysis using general bioinformatics tools (Targetscan, miRanda, and microTv3.0), the p66Shc mRNA did not reveal any let-7 seed matches. Instead, our results indicate that seedless interaction elements in the CR of p66Shc mediate the association between let-7a and p66Shc mRNA and elicit the effects of let-7a (Fig. 2). MicroRNAs influence the translation or turnover of mRNAs as part of a larger molecular complex (the RISC) that includes Ago2 (Pratt & MacRae, 2009;Czech & Hannon, 2011); as anticipated, Ago2/RISC was involved in the let-7a-mediated regulation of Shc, because knockdown of Ago2 increased the levels of Shc proteins and diminished the effect of let-7 knockdown in elevating Shc expression (Fig. 1E).
p66Shc was described as a positive regulator of the proliferation of human prostate cancer cells (Veeramani et al., 2005). In human diploid fibroblasts, p66Shc may not promote cell growth, because elevated p66Shc levels have been observed in senescent human diploid fibroblasts (Fig. 3A) and in oxidative stress-induced senescence (Favetta et al., 2004). Indeed, evidence obtained in the present study supports the view that p66Shc acts as a negative regulator for cell growth (Figs 4C and S4C). On the other hand, let-7a inhibits the proliferation of human glioblastoma (Lee et al., 2011), human nonsmall cell lung tumor (Johnson et al., 2007;Kumar et al., 2008;He et al., 2009), Burkitt lymphoma (Sampson et al., 2007), breast cancer cells (Yu et al., 2007), and primary fibroblasts (Legesse-Miller et al., 2009), through the repression of Ras, c-myc, E2F1, and CDC34 as well as elevation of p21. However, these regulatory processes by let-7a may not impact on replicative senescence, as Ras and cmyc proteins were undetectable 2BS cells (not shown), and the reduction of let-7a in senescent cells was not accompanied by elevation of E2F1 and CDC34 or reduction of p21 (Fig. 3A,B).
The fact that let-7 represses proliferation of tumor cells but promotes the growth of HDFs may reflect the view that cell senescence contributes to tumorigenesis (Rodier & Campisi, 2011). In general, genes highly expressed in senescent cells tend to show low expression levels in cancer, and vice versa. Therefore, strategies to inhibit the growth of tumor cells may also shorten the lifespan of normal cells by inducing senescence, while strategies to extend cellular lifespan may increase the risk of tumorigenesis. Targeting let-7a-p66Shc in cancer may avoid this apparent dilemma, as extending cellular lifespan by elevating let-7a or reducing p66Shc does not increase the risk of tumorigenesis.

RNA isolation, Northern blot, and PCR analysis
Total cellular RNA was prepared using the RNeasy Mini Kit (Qiagen, Hilden, Germany). Northern blot analysis was performed as described by Jing et al. (2005). For reverse-transcription (RT) followed by real-time, quantitative (q)PCR or semiquantitative PCR analysis of p66Shc and GAPDH, primers were described in 'Supporting Information'. The primers for qPCR or semiquantitative PCR analysis of let-7a, miR-30, and U6 were from Ambion (Austin, TX, USA).

Constructs and reporter gene assays
All vectors were constructed as described in 'Supporting Information'. For reporter gene assays, each of the pGL3-derived vectors was cotransfected with pRL-CMV vector by Lipofectamine 2000 (Invitrogen). Forty-eight hour after transfection, cell lysates were collected and the firefly and renilla luciferase activities were measured with a double luciferase assay system (Promega, Madison, WI, USA) following the manufacturer's instructions. All firefly luciferase measurements were normalized to renilla luciferase measurements from the same sample.

Analysis of nascent protein and RNP IP assays
Nascent protein analysis was performed as described in 'Supporting Information'. For RNP IP assays, HeLa Cells were cotransfected with reporter pSL-MS2 or pSL-MS2-CR along with the pSL-flag-MS2-GFP; 48 h later, cell lysates then were prepared and subjected to immunoprecipitation assays by using monoclonal antiflag antibody (Sigma, St. Louis, MO, USA). The IP materials were washed twice with stringent buffer (100 mM Tris-HCI, pH 7.4, 500 mM LiCI, 0.1% Triton X-100, 1 mM DTT, 2 lg mL À1 leupeptin, 2 lg mL À1 aprotinin, 1 mM phenylmethylsulfonyl fluoride), and twice with the IP buffer. RNA was isolated from the IP materials and analyzed by RT-qPCR to assess the levels of microRNAs (Zhang et al., 2012).

Supporting Information
Additional Supporting Information may be found in the online version of this article at the publisher's web-site.       Data S1 Experimental procedures. let-7 regulates p66Shc, F. Xu et al.