The poly (ADP-ribose) polymerase (PARP) proteins have been characterized as enzymes that catalyse the attachment of the ADP-ribose subunits to itself and to multiple target proteins by using NAD⁺ as the substrate (Citarelli, Teotia, & Lamb, 2010; Ray Chaudhuri & Nussenzweig, 2017). This post-transcriptional modification is called Poly(ADP-ribosyl)ation (PARylation) (Citarelli et al., 2010). PARylation is a reversible modification: it is accomplished by the concerted actions of poly(ADP-ribose) polymerase (PARP) enzymes and poly(ADP-ribose) (PAR) hydrolysing enzymes such as PAR glycohydrolase (PARG) and ADP-ribosyl hydrolase 3 (ADPRHL2) (Virág, Robaszkiewicz, Rodriguez-Vargas, & Oliver, 2013). The removal of terminal ADP-ribose unit is achieved by the hydrolytic activity of macrodomain proteins (MACROD1, MACROD2, and OARD1) (Perina et al., 2014).
PARylation is a widely used process in eukaryotes. In eukaryotic species, the distribution of PARP proteins strictly follows the distribution of PARGs and at least one of the macrodomain proteins is also always present (Perina et al., 2014). On the other hand, PARP proteins are less common in bacteria and are thought to be acquired through horizontal gene transfer (Alemasova & Lavrik, 2019; Perina et al., 2014). In thermophilic archaeon Sulfolobus solfataricus, a protein with oligo(ADP-ribosyl) transferase activity was identified (Faraone-Mennella, Gambacorta, Nicolaus, & Farina, 1998) and in a number of dsDNA viruses have also been found to possess PARP homologues (Perina et al., 2014).
Based on the sequence homology, humans are assumed to express 17 defined PARPs (Vyas, Chesarone-Cataldo, Todorova, Huang, & Chang, 2013). PARP1, the first PARP purified and cloned from human, is a constitutive and the best studied member of the PARP family of proteins (Citarelli et al., 2010). The PARP1 gene is conserved in chimpanzee, Rhesus monkey, dog, cow, mouse, rat, chicken, zebrafish, fruit fly, mosquito, C.elegans, A.thaliana, rice, and frog (Table 1).
Table 1. Pairwise alignment of PARP1 gene (in distance from human) (HomoloGene:1222, NCBI).

Gene Species	Gene Symbol	Identity (%) DNA
vs. P.troglodytes	PARP1	99,2
vs. M.mulatta	PARP1	97,7
vs. C.lupus	PARP1	88,4
vs. B.taurus	PARP1	88,4
vs. M.musculus	Parp1	86,5
vs. R.norvegicus	Parp1	86,1
vs. G.gallus	PARP1	75,2
vs. X.tropicalis	parp1	72
vs. D.rerio	parp1	69,5
vs. D.melanogaster	Parp	49,5
vs. A.gambiae	AgaP_AGAP003230	53,1
vs. C.elegans	pme-1	48,2
vs. A.thaliana	PARP1	50,2
vs. O.sativa	Os07g0413700	51,7

The PARP1 gene is a protein-coding gene. It is located at 1q42.12 on the minus strand and consists of 23 exons spanning ∼43 kb (starts at 226360691 and ends at 226408093; GRCh38.p13 Assembly, NCBI).

This gene has 10 transcripts (splice variants) depending on Ensembl release 98 - September 2019 (Table 2 and Figure 2).
The human PARP1 promoter region does not contain typical regulatory elements, such as TATA or CAAT boxes. A near 40-base-pair region surrounding the transcription start site described as containing a near-consensus initiator element capable of initiating RNA polymerase II transcription (Abbotts & Wilson, 2017). Detailed analyses of the promoter regions of PARP1 genes in humans, rats, and mice showed that PARP1 promoter sequences have binding sites for transcription factors SP1, AP-2, YY1, ETS1, and NF1. In the distal promoter region of human PARP1 gene Candidate binding sites for several other factors including CDE, KLF4 (GKLF), BARB, RRM1 (MAZF), RREB1, HOX, GSX1 (GSH-1), CEBPB, NFIL3 (E4BP4), STAT6, cETSZ-1, PBX1, LEF1 (TCF), NF-kB, REL, ZNF148 (ZBP-89), KLF6 (CPBP), USF, CDF-1, EGR1, and IKZF1 (Ikaros 1) were also identified (Doetsch, Gluch, Poznanovic, Bode, & Vidakovic, 2012). Additionally, at the post-transcriptional level, miR-124, MIR223, let-7a, miR-7-5p, and MIR125B2 were shown to regulate cellular PARP1 expression (Dash et al., 2017; J. Lai et al., 2019; Wielgos et al., 2017).
Table 2. Transcripts of human PARP1 gene (Ensembl release 98 - September 2019)

PARP1-201	ENST00000366790.3	570	-	-	Protein coding
PARP1-202	ENST00000366792.3	553	-	-	Protein coding
PARP1-203	ENST00000366794.10	3978	CCDS1554	NM_001618.4	Protein coding
PARP1-204	ENST00000463968.5	830	-	-	lncRNA
PARP1-205	ENST00000468608.1	438	-	-	lncRNA
PARP1-206	ENST00000469663.1	542	-	-	lncRNA
PARP1-207	ENST00000490921.5	3165	-	-	lncRNA
PARP1-208	ENST00000491816.1	416	-	-	lncRNA
PARP1-209	ENST00000498787.1	628	-	-	lncRNA
PARP1-210	ENST00000629232.1	477	-	-	Protein coding

There are two known pseudogenes of PARP1: PARP1P1 and PARP1P2 located on chromosomes 13 and 14, respectively (Table 3). A germline, two-allele (A/B) polymorphism of PARP1P1 on chromosome 13q34-qter has been identified. In the B-allele, a 193 bp deletion was determined and this deletion has been shown to be associated with cancer predisposition to multiple myeloma, monoclonal gammopathies, prostate cancer, and lung cancers in African Americans. On the other hand, more recent studies do not support the earlier findings which suggest that the PARP1P1 genotype plays a critical role in cancer susceptibility (Lockett, Snowhite, & Hu, 2005). Analyses, comprising a larger cohort that is selected based on case/control differences rather than racial/ethnic differences, are needed to clarify if there is any significant role of this pseudogene in cancer predisposition.
Table 3. Pseudogenes of human PARP1 gene (GRCh38 Assembly, NCBI)

Name/Gene ID	Description	Location (bp)	Aliases
PARP1P2 (ID: 145)	poly(ADP-ribose) polymerase 1 pseudogene 2	Chr 14, NC_000014.9 (63123001..63123935)	ADPRTP2, PPOLP2
PARP1P1 (ID: 144)	poly(ADP-ribose) +B7:D7polymerase 1 pseudogene 1	Chr 13, NC_000013.11 (110936624..110940232)	ADPRTP1, PPOLP1

Encoded proteins by PARP1 gene in human are given in Table 4 and PARP1 protein similarity across species are given in Table 5.
Table 4. Protein products of human PARP1 gene (Ensembl release 98 - September 2019)

Name	Transcript ID	Protein	Charge	Isoelectric Point	Molecular Weight	CCDS	UniProt	RefSeq
PARP1-201	ENST00000366790.3	155aa	10,0	9,4862	17,324.99 g/mol	-	Q5VX85	-
PARP1-202	ENST00000366792.3	108aa	3,0	7,7272	12,234.02 g/mol	-	Q5VX84	-
PARP1-203	ENST00000366794.10	1014aa	31,5	9,3322	113,083.79 g/mol	CCDS1554	A0A024R3T8 P09874	NM_001618.4
PARP1-210	ENST00000629232.1	108aa	3,0	7,7272	12,234.02 g/mol	-	Q5VX84	-

Table 5. Pairwise alignment of PARP1 protein sequences (in distance from human) (HomoloGene:1222, NCBI)

Gene Species	Gene Symbol	Identity (%) PROTEIN
vs. P.troglodytes	PARP1	99
vs. M.mulatta	PARP1	98,2
vs. C.lupus	PARP1	94,1
vs. B.taurus	PARP1	90,4
vs. M.musculus	Parp1	92,2
vs. R.norvegicus	Parp1	91,6
vs. G.gallus	PARP1	79,5
vs. X.tropicalis	parp1	75,7
vs. D.rerio	parp1	72,1
vs. D.melanogaster	Parp	43,8
vs. A.gambiae	AgaP_AGAP003230	46,5
vs. C.elegans	pme-1	41,1
vs. A.thaliana	PARP1	42,3
vs. O.sativa	Os07g0413700	42,7

PARP1 encodes poly (ADP-ribosyl) transferase (EC 2.4.2.30) (NCBI Homo sapiens Annotation Release 109). Poly (ADP-ribose) polymerase1 (PARP1) is an isoform of the PARP enzyme family (Pacher & Szabó, 2007). Full length PARP1 protein comprises three major functional domains: an amino-terminal DNA-binding domain, a carboxy-terminal catalytic domain (CD; also called as CAT), and a central auto modification domain (called as AMD or AD) (Altmeyer, Messner, Hassa, Fey, & Hottiger, 2009; Gross, Kotova, Maluchenko, Pascal, & Studitsky, 2016). DBD contains two zinc finger domains (ZFI/ZF1 and ZFII/ZF2; also known as Zn1 and Zn2). Langelier et al. reported that the ZF2 domain exhibits high binding affinity to DNA compared to the ZF1 domain and Gradwohl et al. showed that disruption of the metal-binding ability of the ZF2 dramatically reduces the binding to target DNA (Gradwohl et al., 1990). In addition, in both in vitro and in vivo, the ZF1 domain was found to be necessary for DNA-dependent PARP1 activity whereas the ZF2 domain was not required strictly (Langelier, Planck, Roy, & Pascal, 2011). An additional zinc finger domain (ZFIII/ZF3; also known as Zn3) presents after DBD and it mediates inter-domain contacts, important for the PARP1 activation. (Langelier et al., 2011; Tao, Gao, Hoffman, & Liu, 2008). A bipartite nuclear localization signal (NLS) also lies in DBD and contains a caspase cleavage site DEVD214 (Castri et al., 2014). The AMD region is located in the central region of the enzyme and the region has acceptor amino acids for the covalent attachment of PAR. Moreover, a weak leucine-zipper motif has been described in the amino-terminal region of the AMD, which suggests that this motif may function in homo- and/or hetero-dimerization. The AMD of PARP1 also includes a breast cancer 1 protein (BRCA1) C-terminus (BRCT) domain as well as an unstructured loop that connects the AMD with the PARP homology domain (Altmeyer et al., 2009). The carboxy-terminal CD is the most conserved region across PARP family of proteins in different species. This domain has the PARP signature characterized by NAD acceptor sites and critical residues involved in the initiation (the attachment of the first ADP-ribose moiety onto an acceptor amino acid), elongation (the addition of further ADP-ribose units onto already existing ones) and branching (the generation of branching points) of PAR (Altmeyer et al., 2009; Simonin et al., 1990). Followed by the CD, there is a WGR region, named after the identification of conserved amino acid sequence in the motif: Tryptophan-W, Glycine-G, Arginine-R. WGR region functions in DNA binding and inter-domain contacts essential for DNA damage-dependent activation (Altmeyer et al., 2009; Dawicki-McKenna et al., 2015; Langelier, Planck, Roy, & Pascal, 2012). The domain organization of PARP1 is shown in Figure 3.
PARP1 is known to be activated by mono-ADP-ribosylation, acetylation, increased cellular calcium concentration, or by binding to tyrosyl tRNA synthase. On the other hand, self-PARylation and sumoylation were shown to inhibit PARP1 activity. PARP1 can also be phosphorylated in a reversible manner and the phosphorylation can activate (e.g., AMP-activated protein kinase [AMPK]) or inhibit (e.g., protein kinase C) PARP-1 activity (Bai, 2015). In addition, physical interactions with other proteins, including histones, HPF1, HMGN1, XPA, NEIL1, OGG1, DDB2, TP53, and MAPK1 (ERK2) found to regulate PARP1 activity (Alemasova & Lavrik, 2019). It was also reported that dimerization of PARP1 enhances its enzymatic activity while further multimerization or dissociation to single PARP1 molecules leads to decreased enzymatic activity (Alemasova & Lavrik, 2019).

PARP1 is an abundantly and ubiquitously expressed protein in most tissues (Schiewer & Knudsen, 2014). Compared to normal counterparts, enhanced PARP1 expression is found in various types of tumors, but the most striking differences in PARP1 expression have been found in breast, ovarian, endometrial, lung, skin cancers and non-Hodgkins lymphoma (Galia et al., 2012; Ossovskaya, Koo, Kaldjian, Alvares, & Sherman, 2010).

PARP1 is primarily localized to the nucleus, but a distinct fraction was also detected in the mitochondria. Unlike nuclear PARP1, mitochondrial PARP1 has been shown to affect mitochondrial DNA repair negatively (Szczesny, Brunyanszki, Olah, Mitra, & Szabo, 2014).
PARP1 is one of several known cellular substrates of CASP3 and CASP7 (caspase 3 and caspase 7) and cleavage of PARP1 by these caspases is considered to be a hallmark of apoptosis. Upon cleavage, two specific fragments of PARP1 are generated: an 89 kDa fragment containing AMD and the catalytic domain of the enzyme and a 24 kDa containing DBD. The 89 kDa fragment has a greatly reduced DNA binding capacity and is liberated from the nucleus into the cytosol. On the other hand, the 24 kDa cleaved fragment with 2 zinc-finger motifs does not leave the nucleus where it binds to nicked DNA irreversibly and therefore acts as a trans-dominant inhibitor of active PARP1 (Chaitanya, Alexander, & Babu, 2010).

ADP-ribosylation is a posttranslational modification. By using the oxidized form of NAD⁺ as a substrate, PARP enzymes bind and cleave NAD⁺ to nicotinamide (NAM) and ADP-ribose (ADPR) and catalyze the covalent binding of ADPR units onto glutamate, aspartate, tyrosine, lysine, and serine residues of target proteins (Rodrèguez-Vargas, Oliver-Pozo, & Dantzer, 2019). PARP1, PARP2, tankyrase 1 ( TNKS), and tankyrase 2 ( TNKS2) synthesize branched PAR polymers and the remaining PARP enzymes are either mono or oligo ADPR-transferases. No enzymatic activity has been identified for ZC3HAV1 (PARP13) (Bai, 2015). PARP1 is the prototype member of the PARP family of enzymes and more than 80% of stimulated and basal cellular PARP activity are exerted by PARP1 (Bai, 2015; Rajamohan et al., 2009).
Although PARP1 has been long defined as a DNA-damage response protein, recent investigations highlight multiple functions of PARP1 including transcription, replication, aging, viral protection, cell cycle regulation, modification of chromosome structure, differentiation, inflammation, metabolic regulation, proteasomal degradation, and RNA processing (Bai, 2015; Rodrèguez-Vargas et al., 2019) (Figure 4).
PARP1 functions in DNA repair
Genotoxic stress results in various types of DNA lesions, including DNA single-strand breaks (SSBs) and double-strand breaks (DSBs). If not repaired, accumulated damage can disrupt genomic integrity. Fortunately, cells have evolved different DNA-damage repair responses that repair these DNA lesions to insure genomic stability (C. Liu, Vyas, Kassab, Singh, & Yu, 2017).
PARP1 recognizes both SSBs and DSBs and transfers the ADP-ribose moiety of NAD⁺ to the side chains of asparagine, aspartic acid, glutamic acid, arginine, lysine, serine and cysteine residues on its target proteins. Through their PAR-binding domains, these PAR chains form a platform and recruit DNA repair proteins. Therefore, PARP1 is an important DNA damage sensor for both SSBs and DSBs (Ray Chaudhuri & Nussenzweig, 2017).
PARP1 modulates chromatin structure and transcription
PARP1 functions in chromatin compaction, decondensation and it modulates epigenetic marks via PARylating histones and chromatin remodeling enzymes (Quénet, El Ramy, Schreiber, & Dantzer, 2009). Being as a component of enhancer/promoter-binding complexes, besides its effects on chromatin structure, PARP1 can bind to most of the RNA polymerase II transcribed genes and mediate around 3.5% of all transcribed RNAs covering a broad range of functions from inflammation to metabolism (Ke, Zhang, Lv, Zeng, & Ba, 2019; Kraus, 2008). PARP1 can also enhance the accessibility of promoters via histone and nucleosome replacements and can enhance transcription by replacing negative transcriptional cofactors with positive ones (Kraus & Hottiger, 2013; Muthurajan et al., 2014).
Recent findings described new roles of PARP1 in the regulation of RNA binding proteins, rRNA synthesis, ribosome biogenesis, and mRNA regulation (Ke et al., 2019; Ryu, Kim, & Kraus, 2015). Accordingly, PARP1 can regulate gene expression at the post-transcriptional level.
Cell death and PARP1
As mentioned before, PARP1 is known to be cleaved and inactivated by active caspases 3 and 7 and this cleavage is accepted as a hallmark of apoptosis (Castri et al., 2014; Desroches & Denault, 2019). The cleavage causes the formation of 24 kDa and 89 kDa fragments. Depending on the intensity and type of stimuli resulting in the cleavage, two main consequences have been reported: (1) reduced PARylation during DNA repair processes; (2) the modification of PARP1 transcriptional activity (Castri et al., 2014).
Recent studies indicate that PARP1 hyperactivation, ie. excessive PARylation by PARP1, can lead parthanatos, a form of necrotic cell death which PAR induces the nuclear translocation of apoptosis-inducing factor ( AIFM1) from mitochondria to initiate chromatinolysis and cell death independently of caspase activation. For a long time, it has been thought that the cell death caused by excessive activation of PARP occurs via the catalytic consumption of NAD⁺ followed by ATP reduction and bioenergetic collapse. However, Andrabi et al. showed that not the decreased NAD⁺, but PAR-dependent inhibition of hexokinase activity leads to defects in glycolysis and therefore causes the bioenergetic collapse. On the other hand, PARP1 activity is kept at much lower levels during normal unstressed cellular conditions (Andrabi et al., 2014; Dawicki-McKenna et al., 2015; Gupte, Liu, & Kraus, 2017).

A list of PARP1 mutations in cancer can be found in: COSMIC, the Catalogue of Somatic Mutations in Cancer, https://cancer.sanger.ac.uk/cosmic/gene/analysis?ln=PARP1

Parp1^-/- mice are viable, fertile, normal in size and do not display any gross physical or behavioral abnormalities (Tha Jackson Laboratory; www.jax.org/strain/002779).

Lys933Asn and Lys945Asn mutations were found to be significantly correlated with colorectal cancer (CRC) in the Saudi population. Since these mutations were identified to be localized in PARP1 catalytic domain (CD), mutations of these lysine residues were suggested to affect the PARP1 catalytic activity (Alshammari, Shalaby, Alanazi, & Saeed, 2014). In addition, Val762Ala polymorphism in the CD has been reported as the most common variant of PARP1 associated with an increased risk of many tumors (Toss & Laura, 2013).
Using genome-wide and high-density CRISPR-Cas9 tag-mutate-enrich mutagenesis screens, Pettitt et al. identified PARP1 mutant alleles that cause in vitro and in vivo PARP inhibitor resistance. The results reveal that point mutations in the ZnF domains were sufficient for the inhibitor resistance (Pettitt et al., 2018).
Polymorphisms: In addition to mutations, through modulation of PARP1 expression level and enzyme activity, PARP1 gene polymorphisms can affect the outcome and response to therapy of cancer. For example, PARP1 SNP rs1805407, found in perfect linkage disequilibrium with two PARP1 promoter SNPs (rs2077197 and rs6665208), was shown to be associated with higher PARP1 expression (Abecassis et al., 2019). In another study, expression quantitative trait locus (eQTL) analysis in melanocytic cell types revealed that presence of the 1q42.1 melanoma risk allele (rs3219090[G]) is correlated with higher PARP1 levels. Furthermore, a proteomic screen identified that RECQL helicase binds to the insertion allele of PARP1 (indel SNP rs144361550) in melanoma cells and primary human melanocytes (J. Choi et al., 2017). In another study, using a new data integrative approach applied on multi-modal -omics, and clinical data, Abecassis et al. demonstrated that response to chemotherapy is directly linked to the gene expression, four methylation variables and PARP1 SNP rs1805407 in a cohort of metastatic melanoma patients (Abecassis et al., 2019). According to the results of another genotyping study, Val762Ala, Asp81Asp, and Lys352Lys polymorphisms and the haplotype-ACAAC in PARP1 are associated with reduced risk of non-Hodgkin lymphoma in Korean males (Jin et al., 2010). In a case-control study conducted in the Hexi area of China, PARP1 2819G allele was shown to be associated with an increased risk of gastric cancer (He, Liu, Shan, Zhu, & Li, 2012).
In addition, PAR metabolism is also involved in malignancies. For instance, PARylation of proteins in peripheral blood leukocytes was shown to be reduced by more than 50% in head, neck, breast and cervical cancers (Lakadong, Kataki, & Sharan, 2010).