Michael A. Resnick - US grants

Summary of Work: Genetic defects as well potentially unstable at-risk DNA motifs (ARMs) can cause genome instability and the combination can lead to synergistic increases in instability and disease in humans. Yeast provides an in vivo test tube for functional analysis of human DNA metabolic genes and ARMs. We used ARMs to address genome stability and to detect subtle defects in DNA metabolic genes, reasoning that variants with a small effect might exhibit strong synergistic interactions between alleles, or with ARMS or with environmental factors. A) Alterations in the essential DNA polymerase e impact on genome stability. An allele was isolated that specifically increases +1 frameshift mutations in long homonucleotide runs by lowering DNA replication fidelity. In combination with a proofreading defect the double mutant is the strongest polymerase e mutator identified and is lethal in a mismatch repair background. B) The 5' DNA flap endonuclease hFEN1, which is important for human replication and repair, could fully complement a yeast null RAD27 mutant. The several genetic effects of a nuclease-deficient allele led to the isolation of novel genotoxic hFEN1 mutants. A mutant RAD27/FEN1 that lacks interaction with PCNA appears to have little effect on genome stability; however, it exhibited synergy with double-strand break (DSB) repair mutants. C) We discovered (through an IRA collaboration with the Kunkel lab) strong negative interactions between a subtle allele of RAD27/FEN1 and defects in the DNA polymerase d 3'->5' exonuclease (but not in other domains of Pol d) that caused hyper mutation and recombination, DSBs and even cell death. This demonstrates a novel role for this exonuclease in addition to proofreading. D) Many combinations of subtle mutator alleles of DNA polymerase d combined with strong recessive mutators can result in a strong mutator phenotype in diploid cells. E) collaborating with the Wilson lab we developed a yeast system to investigate human DNA polymerase b, which can play a major role in the repair of DNA damage.

Summary of Work: Repair of double-strand breaks (DSBs) in the chromosomal DNA of humans and other eukaryotic organisms involves a complex interplay between proteins associated with repair of broken DNA ends, maintenance of chromatin structure and DNA damage-sensitive checkpoints. Mutations within genes involved in the two major pathways of DSB repair, referred to as homologous recombination and nonhomologous end-joining (NHEJ), and in cell cycling responses to DNA strand breaks have been implicated in the etiology of several human cancers and in processes leading to cell senescence. Our research efforts have focused on DSB repair in the genetically tractable yeast Saccharomyces cerevisiae. Most genes comprising the two major pathways of DSB repair are structurally and functionally conserved in yeast and human cells. We have developed an in vivo system for enyzmatically generating defined DSBs and conducted a systematic analysis of genes involved in repair using multiple assay systems. We have a) provided the first demonstration that yeast genes involved in recombination-independent NHEJ play an essential role in the repair of EcoRI endonuclease-induced DSBs in vivo; b) shown that the functions of two repair complexes (involving Ku70:Ku80 and Rad50:Mre11:Xrs2) in NHEJ are genetically separable; c) developed the first functional assay for yeast DNA Ligase IV based on quantitation of endonuclease-induced checkpoint activation in dnl4 strains; d) confirmed that the mismatch repair exonuclease encoded by EXO1 and the excision repair endonuclease Rad1/Rad10 are involved in the DSB recombinational repair pathway; e) revealed that low-level expression of PvuII endonuclease (generating DSBs with blunt termini) is lethal in haploid cells, but not in diploid cells. This latter result establishes that the structures at the ends of DSBs are critical for determining their mechanism(s) of repair.

p53 MASTER REGULATORY NETWORK TRANSACTIVATION . The DNA binding activity of the master regulator p53 is critical to its tumor suppressor activity in response to cellular and environmental stresses. The role of binding to individual response elements (REs) in transactivation specificity was investigated using a highly regulatable p53 assay in yeast and human cells. Nearly all REs have at least 1 mismatch from the 20 base consensus sequence and spacers tend to consist of only a few bases at most. We are deconstructing the canonical p53 consensus sequence in order to understand the role of sequence, organization and level of p53 on transactivation in budding yeast and human cell systems. Contrary to early reports for in vitro binding, increases in spacer length of only a few bases greatly reduces p53 transactivation. While p53 lacked transactivation capacity from many full-sized RE canonical sequences, it functioned to different extents from several noncanonical sites that are frequent in the genome including 3/4 and 1/2 sites. Surprisingly, there can be substantial transactivation even at half-sites depending on sequence and p53 expression level. Efficient transactivation from canonical and noncanonical elements requires tetrameric p53. Thus, RE sequence and organization can have a large impact on p53-mediated transactivation. Furthermore, the functionality of noncanonical sequences greatly expands the p53 transcriptional network. In collaboration with the Dr. Douglas Bell group we have been involved in the development of a high-throughput genomics system using cellular extracts and a microsphere assay that detects protein-DNA binding (MAPD) for elucidating functional impacts of p53 protein and RE sequence variation and spacers in transcriptional networks.[unreadable] FUNCTIONALITY OF HUMAN p53 MUTANTS. p53 missense mutations in the DNA binding domain are often cancer-associated. As shown with yeast-based systems, p53 mutants can alter spectrum and intensity of transactivation from individual REs. We addressed in human cells the relationship between changes in the p53 master regulatory network and biological outcomes. Expression of integrated, tightly regulated DNA binding domain p53 mutants resulted in many patterns of apoptosis and survival following UV, ionizing radiation or spontaneously. These patterns reflected changes in the spectra and activity of p53 target genes and showed that p53 mutations in human cells can differentially influence target gene transactivation resulting in a variety of biological consequences which, in turn, might be expected to influence tumor development and therapeutic efficacy. We are also investigating how WT and mutant p53 can regulate genes through novel (noncanonical) sequence. [unreadable] EVOLUTION OF p53 TRANSCRIPTIONAL NETWORK. Networks can evolve through variation of master regulators and/or by changes in regulation of genes within networks. Using a combination of custom bioinformatics and multispecies alignment of promoter regions, we investigated the functional evolution of REs in terms of responsiveness to the sequence-specific transcription factor p53. We identified REs orthologous to known p53 targets in human and rodent cells or alternatively REs related to the established p53 consensus. The orthologous REs were assigned p53 transactivation capabilities based on rules determined from model systems, and a functional heat map was developed for p53 transactivation towards 38 genes in 14 species. This approach emphasizes not only conservation of functionality but also addresses conservation of level of responsiveness. Individual REs exhibited marked differences in potential transactivation as well as widespread turnover of functional binding sites during p53 network evolution. Functional differences were often not predicted from consensus sequence evaluations. Of the 43 established human p53 REs analyzed, 50% were nonfunctional in rodents. Surprisingly, there was almost no conservation of functional REs or compensatory REs for genes involved in DNA metabolism or repair suggesting important differences in p53 stress responses as well as cancer development between humans and rodents. Currently our efforts are directed to evaluating the conservation of noncanonical p53REs as well as spacer length.[unreadable] VARIATION WITHIN THE p53 NETWORK. We had developed a system to identify SNPs in potential p53 REs that are predicted to modify p53 control of target genes and establish their functional transactivation consequences. Recently we characterized a SNP in the promoter of the angiogenic factor FLT-1 receptor that might result in p53 transactivation. Interestingly the isolated p53 RE, which was actually a half-site, did not function alone and its transactivation required another upstream site that corresponded to an estrogen receptor half-site. The presence of both functional REs resulted in synergistic transcriptional activation of FLT1. Following these findings we had used the FLT1 promoter as a model, to understand in detail the synergistic mechanism of ERE with different p53 REs and in response to several environmental stresses.[unreadable] EXPANSION OF p53 TRANSCRIPTIONAL NETWORK. In a more global sense, these findings dramatically increase the size of a master regulatory network because half-sites contribute to regulation, and they demonstrate new opportunities for cross-talk between networks. Using the p53 transactivation rules in combination with bioinformatic approaches we are employing genome-wide approaches to determine other putative p53 target genes containing p53 half-site REs that could present similar regulation. In a preliminary search 15 novel putative p53 target genes were identified including some p53REs. We demonstrated that the RAP80 gene, involved in the DNA damaging signaling, is indeed a p53 target gene. In addition we identified a link between these two proteins (RAP80 and p53) and showed that they are part of a novel auto-regulatory loop consisting of RAP80, HDM2 and the p53. This loop includes transcriptional regulation of the RAP80 gene by p53 through a noncanonical p53 target sequence, interaction of RAP80 with p53 and HDM2, and negative modulation of p53 stability and activity by RAP80. The second gene analyzed is toll-like receptor 3 (TLR3) involved in the innate immune pathway and detection of invading pathogens particularly double-stranded RNA viruses. We demonstrated that different DNA damaging agents which activate p53 resulted in TLR3 gene activation. We also observed that p53 binds the RE identified in the promoter region of TLR3 under these stresses. In addition we have also identified and established that the promoters of several Toll receptors have canonical and noncanonical p53 target REs as well as SNPs in the target REs.[unreadable] CLINICAL IMPLICATIONS. The systems developed in yeast have provided opportunities to address the functional consequences of p53 mutations identified in tumors. As part of a screening of patients undergoing neoadjuvant treatment for breast cancer, p53 mutations are identified. The mutations are then characterized in terms of transcriptional function and related to disease characteristics. [unreadable] FRATAXIN. Friedreichs ataxia is a quantitative disease in that the level of reduction of frataxin appears to correlate with disease severity. Previously we had found with a yeast system that reduced levels of frataxin led to a) iron accumulation within mitochondria; b) mitochondrial DNA damage/loss and protein damage; and c) nuclear damage. To assess the biological consequences of reduced frataxin levels in human cells, we are utilizing RNA interference to lower frataxin levels 70 to 90%, similar to levels in FRDA. Preliminary results indicate that reduced levels of frataxin results in much high levels of mitochondrial lesions following treatment with ROS (hydrogen peroxide).

Site-directed mutagenesis systems in which specific DNA sequences are targeted for alteration in vitro have been instrumental in dissecting genetic pathways, gene regulation and in understanding structure-function relationships in proteins. Often there is a need for direct in vivo modification. However, the modification of genomic DNA within cells, such that no heterologous material is retained, is generally an elaborate and inefficient process, that includes various cloning steps for each mutant allele that is to be created. We developed an oligonucleotide targeting approach that eliminated many of the steps required to produce multiple changes in the genome. While developed in the yeast Saccharomyces cerevisiae where homologous recombination is highly efficient, the approach could be applied to other organisms. Briefly, the first step involves integration of a COunterselectable REporter (CORE) cassette that is targeted to a desired genomic locus. The gene modification step occurs by transformation with the appropriate transforming oligonucleotides such that the CORE cassette is lost and the appropriate change is generated. This versatile and accurate system for in vivo targeted mutagenesis has been referred to as delitto perfetto [Italian for perfect murder and idiomatic for prefect deletion] since there is complete elimination of the marker sequences that are used for selection. In this way, no heterologous sequence remains and multiple rounds of mutagenesis can be accomplished by what might be considered as a self-cloning process applicable to any target sequence. The system has been applied extensively in many studies in and beyond our laboratory to rapidly generate mutants of yeast genes or genes from other organisms cloned in yeast. A primary function of homologous recombination in the cells of all organisms is to repair double strand breaks (DSBs) that may occur during meiosis, programmed DNA rearrangements, faulty DNA metabolism and as a result of DNA damage. We proposed that an oligonucleotide might be capable of repairing a DSB and thereby provide a new tool to address DSB repair. A DSB-CORE cassette was created that contained a regulatable I-SceI endonuclease and an I-SceI double-strand cut site within the original CORE cassette described above. A site-specific DSB could, therefore, be generated by the I-SceI just prior to oligonucleotide transformation. We found that the DSB stimulated oligonucleotide targeting more than 1000-fold, with targeting efficiencies as high as 20% of all cells. This is over 2 orders of magnitude higher than any reported DNA integration frequency in yeast. We also found that a DSB can strongly stimulate recombination with single-strand DNA, without significant strand bias, suggesting new twists on present models of DSB repair. The extremely high transformation frequencies and versatility of the break-mediated delitto perfetto system, has resulted in new powerful tools for dissecting mechanisms of homologous recombination as well as rapid genome modification from point mutations to gross chromosome rearrangements such as chromosome circularization, chromosome fusion and reciprocal translocations. We demonstrated the mechanisms of DSB repair mediated by oligonucleotides by examining the effects of null mutations in all known DSB recombination/repair genes. We found that only the RAD52 gene, which is necessary in almost all homologous recombination events in yeast, is absolutely essential for the oligonucleotide targeting process. Rad52 can have two different roles in recombination: single-strand annealing (SSA) between complementary sequences in a Rad51 independent manner or facilitating Rad51-mediated strand invasion and recombination through recruitment of Rad51 protein. We demonstrated that oligonucleotide targeting to a DSB in haploid yeast is independent of Rad51. We also examined oligonucleotide targeting to a single DSB in diploid cells. Only 4% of repair events were due to recombination with the oligonucleotides, suggesting that most of the repair was via recombination with the unbroken homologous chromosome. Since DSB induced recombination between homologous chromosomes is prevented in a rad51 mutant, we examined oligonucleotide-targeted repair in rad51 homozygous cells. The efficiency of targeting increased nearly 14-fold and the vast majority of DSB repair events were due to oligonucleotide-mediated targeting. We propose that oligonucleotide targeting to a DSB occurs mainly via a Rad52-dependent SSA pathway. Rad51 constrains oligonucleotide targeting by directing the repair into recombination with a sister chromatid or a homologous chromosome and possibly by suppressing SSA. Based on these observations in yeast, we are currently pursuing the development oligonucleotide targeting in human cells.

The p53 tumor suppressor is central to human DNA repair, damage checkpoints and many aspects of human biology. Importantly, most cancers are altered for p53 function. There is considerable variation in p53 dependent expression across over 200 targeted genes leading to differences in p53-mediated biological consequences, due in part to variation in target response element (RE) sequence. We found that the target sequence motif differs considerably from the in vitro derived RE consensus target sequence previously described as 2 copies of RRRCA/TT/AGYYY separated by a spacer of up to 13 bases. We have focused on RE functionality, i.e., the ability of REs to support transactivation by p53. To directly assess functionality of human REs, i.e., transactivation responsiveness, we developed promoter systems in budding yeast for variable human p53 expression and have translated many of the findings to human cells in culture and ex vivo. Recently, we identified super-transactivating sequences that provide high transactivation at low p53 levels. The yeast system has been used to establish the functional evolution of REs across species and was summarized in our piano model that describes functional variability within a transcriptional network. We found that p53 can transactivate through noncanonical binding sequences including half-sites greatly expanding the p53 regulatory network. We extended these studies genome-wide using ChIP-Seq analysis to identify p53 binding sites and associated gene expression changes following p53 activation by different agents in cancer cells. Binding often is not associated with transactivation. A de novo motif search for the in vivo consensus binding sequence revealed the p53 canonical motif is, in fact, 2 tandem RRRCWWGYYY decamers without spacer. Unexpectedly, 10% of p53 bound targets were 1/2-sites. Importantly, we showed that p53 can engage transcription also through recognition of half-sites across the genome and went on to define the minimal binding unit for p53-mediated transcription as a 1/2-site. INTERACTION OF p53 AND ER NETWORKS. We had identified a 1/2-site estrogen receptor RE that greatly increased p53 transactivation at the FLT1 1/2-site p53 RE, establishing a new dimension to the p53 network. Recently, we addressed the generality of synergistic transactivation by p53 and ER. p53 transactivation was greatly enhanced by ligand-activated ER acting in cis. The increase extends to several cancer-associated p53 mutants, suggesting ER-dependent mutant p53 activity for at least some REs and possibilities for reactivation of cancer mutants. We propose a general synergistic relationship between p53 and ER master regulators in transactivation of p53 target canonical and noncanonical REs which might be poorly responsive to p53 on their own. CANCER-ASSOCIATED P53 MUTANTS. Nearly all cancers have mutant or reduced expression of the p53 tumor suppressor gene. We found that p53 mutations can lead to considerable diversity in the spectrum of responses from REs including 1) decrease/loss-of-function; 2) subtle changes; 3) altered specificity; and 4) super-transactivation. With inclusion of immune response-related TLR genes into the p53 network, we evaluated the effect of 25 tumor-associated p53 mutants on TLR gene family expression after transient transfection in p53-null cancer cell. Changes in TLR transactivation patterns, including change-of-spectrum, were observed, suggesting that p53 tumor status might be an important factor in adjuvant therapy employing TLR pathways to treat cancer. Furthermore, we demonstrated that tumor-associated p53 mutants that induced expression of TLR3, enhanced cytokine and chemokine responses mediated by this receptor after exposing cells to TLR3 ligand poly-I:C alone or in presence of Doxorubicin. We also found that functional rescue of loss-of-function p53 mutants by the p53 reactivating drug RITA, restored TLR gene expression in a mutant p53 cell line and also enhanced DNA damage induced-apoptosis via TLR3 signaling. Having established that WT and p53 mutants identified in somatic and germline-associated tumors can modulate TLR expression differentially, we propose that chemotherapeutic manipulation of normal or mutant p53 responses along with immune challenges that include TLRs could enhance inflammatory/immune type responses to environmental factors. p53 NETWORK EVOLUTION. We are investigating evolution of REs in terms of responsiveness to p53. Individual REs exhibited marked differences in potential transactivation as well as widespread turnover of functional REs during p53 network evolution. Only 1/3 of the REs found in humans are predicted to be functionally conserved in rodents, although nearly 2/3 of the REs were conserved at the sequence level. Importantly, we found functional conservation of weakly responding REs including 1/2 sites. Among validated p53 REs conserved between rodents and humans, one third were comprised of 1/2- or 3/4-sites, each with a perfect consensus-site suggesting a selective advantage in retaining weak p53 REs. Similar to those observations, we identified in our recent ChIPseq cancer cell study that 60% of the p53 REs were conserved at the sequence level in rodents. Using Geneontology for gene function classification to identify those potential p53 target genes associated with DNA metabolism and repair functions, we found that only a few of the p53 RE sequences were conserved in rodents in agreement with our earlier findings. These results may indicate evolutionary selection on p53 dependent chromosomal stress responses. REGULATION OF THE IMMUNE RESPONSE. The immune system can impact tumor development. We found that among the 10 human TLRs, nine had canonical and noncanonical p53 REs. Using primary human cells, we examined expression of the entire TLR gene family following exposure to anti-cancer p53 inducing agents. Expressions of all TLR genes in blood lymphocytes and alveolar macrophages from healthy volunteers were inducible by DNA metabolic stressors. However, there is considerable inter-individual variability. Similarly, p53 dependent TLR expression is detected in human cancer cell lines. For some TLRs the p53 control seems to enhance the inflammatory responses, mediated by activation of TLRs in the presence of natural ligands. In particular, we found a p53-dependent increase in response to the TLR5 ligand flagellin as measured by both mRNA and protein production of the downstream cytokines IL-6 and IL-8 accompanied by a specific increase in phosphorylation of p38 MAP kinase, which is one of the mediators of TLR signaling in breast cancer cells. Global gene expression analysis revealed a group of 200 genes that exhibited this p53/ligand synergistic response including genes related to immune/inflammation processes. We also found that in the absence of TLR ligands, p53 that is induced by anticancer drugs can cooperate with Nuclear Factor-kappa beta (NFkappa-beta) to induce pro-inflammatory cytokines in primary human macrophages. Transcriptome analysis identified a global cooperative p53/NF-kappaB effect on expression of immune response genes, including several chemokines such as CXCL1, CXCL3, and CCL5. Recently, we established that the genome guardianship role of p53 extends to the innate immune system beyond the TLR genes. In our p53 ChIPseq studies in cancer cells, we identified 3 members of the APOBEC3 family, A3B, -C and H, that were directly regulated by p53 in a stress dependent manner. There are 7 members of the APOBEC3 family and their products are nucleic acid deaminases that can mutagenize and inactivate infecting RNA viral genomes, serving as warriors in the innate immune response. We found from our p53 cistrome studies that several members of the APOBEC3 family are directly regulated by p53 in human cancer and primary cells.