A ChIP-cloning approach linking SIRT 1 to transcriptional modification of DNA targets

Limiting food intake increases life span (1). For instance, calorie restriction (CR) extends mean and maximum life span in a variety of organisms, including yeast (2), fruit flies (3), nematodes (4), and rodents (5). Food restriction also appears to promote longevity in nonhuman primates (6), suggesting therefore that CR may be an untapped potential procedure for modifying life span in humans. The mammalian SIR2 ortholog, SIRT1 (Sirtuin1), mediates in part the physiological effects produced by CR (7). SIRT1 is nuclear and functions as a protein deacetylase, capable of modulating the actions of other proteins in a (NAD+)-dependent manner. SIRT1 targets a wide range of proteins, including p53 (8), NFkappaB (9), MyoD (10), and histone proteins that form the nucleosome. Much of the work on SIRT1 has focused on linking Sirt1-induced protein modification with age-onset diseases. In contrast, the search for specific target genes directly affected by SIRT1 have been largely neglected, particularly those genes associated with epigenetic changes of the histone code. In this regard, once tethered to DNA, SIRT1 influences chromatin structure by deacetylating specific lysine residues of histones H1, H3, and H4 (11,12). Histone deacetylation in eukaryotic cells is generally associated with heterochromatin-formation and transcriptional repression.

Two major obstacles for the analysis of previously unidentified DNA targets are the lack of detailed knowledge of SIRT1/DNA linkage and the generally poor intergenic resolution of genome scanning techniques such as those found in microarrays procedures. To circumvent these two limitations, we decided on a technique that combines both chromatin immunoprecipitation (ChIP) with cloning of the generated DNA piece, followed then by sequence analysis (13). In this context, all ChIP techniques (14) take advantage of the fact that proteins that link with chromatin can be crosslinked to the site of their association in cell lineages. Subsequent fragmentation of the endogenous DNA, followed by immunoprecipitation using an antibody specific for the protein in question, allows for the selection of a specific protein/DNA complex. The purified DNA fraction is most often screened for explicit targets by PCR using sequence-specific DNA primers. To find previously unknown DNA targets, the precipitated DNA can be used in a genome-spanning DNA microarray (ChIP-chip), which is often not only prohibitively expensive but also limited by the degree of intergenic coverage. This latter issue is especially problematic when working with proteins binding to the large extent of regulatory versus coding DNA. We therefore chose to pursue a ChIP-cloning approach followed by DNA sequence identification of the precipitated DNA. The application of this combinatorial approach provides an innovative novelty to the field of SIRT1 biology, allowing the identification of endogenous DNA targets to be viewed in a true genome-wide search format.

For the following experiment, adherent human embryonic kidney cells (HEK-293) were grown to 80%–90% confluence on 10 cm cell culture dishes under standard culture conditions (Dulbecco's Modified Eagle Medium, 10% fetal bovine serum, at 37°C with 5% CO₂). Ten 10 cm dishes (approximately 1–2 × 10⁸ cells) went into one single experiment. ChIP and subsequent DNA purification was accomplished by using the EZ ChIP kit (Upstate, Lake Placid, NY, USA) according to the manufacturer's specifications. The crosslinked DNA was sheared into approximately 200 to 1000 bp fragments by sonication of the cell lysate with five sets of 10 s pulses on wet ice using a Cell Disruptor (Heat-Systems Ultrasonics, Farmingdale, NY, USA) with a 2 mm tip and set to 30% of maximum power. SIRT1/DNA complexes were immunoprecipitated by either using 10 µg of a monoclonal (Clone 2G1/F7, Upstate) or a polyclonal anti-SIRT1 antibody (Bethyl Laboratories, Montgomery, TX, USA). Following the manufacturer's specifications, 5 µl of the purified DNA solution was ligated to an EcoRI adaptor and excess adaptor subsequently removed using spin columns prepared with Sephacryl S-400 (Universal Riboclone cDNA Synthesis System, Promega, Madison, WI, USA). Purified DNA was amplified by PCR using primers matching the sequence of the EcoRI adaptor (forward and reverse primer: AATTCCGTTGCTGTCG). The amplified DNA was briefly separated on a 1.5% agarose gel and fragments larger than 200 bp were extracted (Qiagen, Valencia, MD, USA). Purified DNA was ligated into the pGEM-T Easy Vector System and transformed into the E. coli strain JM109 (Promega). Plasmid DNA from positive transformants (blue-white screen) was subjected to restriction with the endonuclease EcoRI, and analyzed on a 1.5% agarose gel (Figure 1). Inserts larger than 200 bp were DNA-sequenced and identified using the “Basic Local Alignment Search Tool” (BLAST; www.ncbi.nlm.nih.gov/blast). A selection of identified DNA targets was verified by PCR using sequence-specific DNA primers (Integrated DNA Technologies, Coralville, IA, USA) on unamplified ChIP DNA (Figure 2). One of the identified target sequences corresponds to an intronic sequence of the gene encoding the neurooncological ventral antigen2 (nova2). The type of association of SIRT1 with nova2 is not yet clear. Because SIRT1 has not yet been demonstrated to bind DNA by itself, the concomitant presence of other protein factors linking SIRT1 to the nova2 gene locus seems likely to occur. Regardless of this circumstance, it is conceivable that SIRT1 might facilitate local histone deacetylation followed by the formation of heterochromatin and transcriptional repression of nova2. Interestingly, NOVA2 is known to regulate the alternative splicing of scn1a, which encodes the α-subunit of a neuronal sodium channel associated with the pharmacological effects of antiepileptic drugs (e.g., carbamazepine and phenytorin; 15). These drugs are thought to control seizures in people by acting on certain neuronal channels of the neuromuscular system. Of further interest, antiseizure drugs markedly extend mean and maximum life span of the nematode Caenorhabditis elegans (16). Our results open up the possibility that there may be new SIRT1 targets not yet explored that affect life span and neuromuscular function.