Structural Biochemistry/DNA Repair/Polynucleotide kinase/phosphatase in DNA strand break repair

Introduction
DNA strand breaks are often caused by internal and external factors. After the termini of these strands break, they require processing before missing nucleotides can be replaced by DNA polymerase and its strands rejoined by DNA ligases. The enzyme polynucleotide kinase/phosphatase plays an important role in repairing DNA strand breaks by catalyzing the restoration of DNA’s termini. In addition to this, PNKP also helps in other DNA repair pathways through interactions with other DNA repair proteins such as XRCC1 and XRCC4. PNKP is important in maintaining genomic stability of normal tissues, like developing neural cells, and enhancing resistance of cancer cell to genotoxic therapeutic agents.

Polynucleotide kinase/phosphatase
When damage is done to cellular DNA, this causes aging, cancer etiology and treatment, and neurological disorders. DNA damage comes in different forms like: base modification and base loss and strand breaks. These damages can be triggered by intracellular agents like primary reactive oxygen species (ROS) and exogenous agents. In order to protect themselves from this damage, cells have evolved a battery of repair pathways. These counter mutational and cytotoxic consequences that occur due to DNA damage. Various mechanisms that cause strand breaks include: cleavage by physical and chemicals means such as ionizing radiation (IR) and ROS, and enzymatic processes. Therefore, strand breaks comes in a wide variety of forms and different strand breaks can be further classified or subdivided based on the nature of their termini. The enzyme PNKP carries 5’-kinase and 3’phosphatase activities that are essential for processing of single and double strand breaks at termini. Research into PNKP has shown that small molecule inhibitors of these enzymes sensitize cells to IR or chemotherapeutic agents. Researchers have also identified that mutations that have lead to changes in PNKP, similar to mutations in other genes that encode other strand break repair proteins, have been connected to a severe autosomal recessive neurological disorder.

IR-and free radical-induced breaks
Ionizing radiation (IR) causes strand breaks with a variety of end groups at 3’-termini by generating hydroxyl radicals. By generating hydroxyl radicals, reactions at different carbon atoms occur within the deoxyribose group to produce two predominant end groups: phosphate and phosphoglycolate. Phosphoglycolate formation is dependent on the presence of oxygen while 3’-phosphate groups are produced under normoxia and anoxia. On the other end, at the 5’-termini, the major end group is phosphate. In addition to causing strand breaks, ionizing radiation also generates complex lesions. These areas contain two or more damaged bases or strand breaks in close quarters and singly damaged sites. Complex lesions include frank DSBs with a ratio of SSB:DSB determined to be ~25:1. Another factor that causes strand breaks is hydrogen peroxide. Similar to IR-medicated damage, hydrogen peroxide causes far fewer frank DSBs. Bleomycin, a chemotherapeutic agent, additionally produces DSBs at the 3’-phosphoglycolate termini.

Camptothecin-induced breaks
The enzyme topoisomerase 1 creates a DNA cut with a 5’-OH terminus and a covalent 3’-phosphate-enzyme intermediate in order to relieve torsional strain. Using topoisomerase, camptothecin prevents resultant strand rejoining, leaving a DNA-enzyme ‘dead-end’ complex. By hydrolyzing this complex with tyrosyl-DNA phosphodiesterase, more cuts with 3’-phosphate and 5’-OH termini are made.

Repair-endonuclease induced breaks
Using DNA glycosylases, damaged bases can be removed. The abasic sites are then cleaved by one of two classes of enzymes. One of the enzymes, AP endonuclease, hydrolyses the phosphodiester bond 3’ to the abasic site in order to give 3’-OH and 5’-deoxyribose phosphate termini. By using DNA polymerase β, 5’-deoxyribose phosphate termini can be converted to 5’-phosphate. AP lyase works by cleaving the phosphodiester bond 5’ to the abasic site by a β-elimination reaction to give a β-unsaturated aldehyde attached to 3’-phosphate at one terminus and a 5’-phosphate at the other. Since many DNA glycosylases have this enzyme activity, the pentenal moiety can then be eliminated by an AP endonuclease to give 3’-OH or by an AP lyase to give 3’-phosphate. Enzymes NEIL1 and NEIL2, mammalian DNA glycosylases with β,δ-lyase activity, remove an extensive amount of mutagenic and cytotoxic oxidative pyrimidien lesions and purine-derived formamidopyrimidines.

Molecular architecture of PNKP
PNKP is constituted as a multidomain enzyme. It consists of 2 domains: an N-terminal forkhead-associated (FHA) domain and a C-terminal catalytic domain that is composed of fused phosphatase and kinase subdomains. Using a flexible polypeptide segment, the two domains, FHA and catalytic domain are linked together. This flexible polypeptide segment acts to selectively bind acidic casein kinase 2 (CK2)-phosphorylated regions in XRCC1 and XRCC4. XRcc1 and XRCC4 are important scaffolding proteins that repair DNA SSBs and DSBs. Aprataxin and APLF are DNA repair factors that also include FHA domains that likewise bind CK2-phosphorylated XRCC1 and XRCC4. This function could result in coordinated regulation of these proteins leading to binding of the phosphorylated scaffolding factors. PNKP and T4 polynucleotide kinases are similar in their catalytic domain in that they both contain contiguous kinase and phosphatase domains but different in that T4 enzyme lacks a FHA domain and that the kinase subdomain lies N-terminal to the phosphatasesubdomian.

The two catalytic active sites are positioned on the same side of the protein of murine PNKP. Murine and T4 kinase subdomain share resembling structure of a bipartite active site cleft that ahs separate ATP and DNA binding sites. The structure of the ATP binding site includes Walker A (P-loop) and B motifs conserved in various kinases. In addition it also carries aspartic acid that activates the 5’-hydroxly for attack on the ATP γ-phosphate. DNA binding sites between mammalian and phage enzymes are different. While phage PNK DNA binding cleft forms a narrow channel that leads to the conserved catalystic aspartic acid residue that accommodates single-stranded substrates, mammalian enzymes phosphorylates 5’hydroxyl termini within cut, gapped or DSBs with single-stranded 3’ overhanging ends since single-stranded 5’ termini are phosphorylated less efficiently. A broad DNA recognition grove composed of two distinct positively charged surfaces, selectively recognizes larger, double-stranded DNA substrates. By using structural information from small angle X-ray scattering experiments coupled with the effect of amino acid substitutions on surfaces of kinase, researchres found that DNA substrates bind across these surfaces in a defined orientation.

A typical process employed by many phosphatases is the haloacid dehalogenase fold. Mechanisms employed by these enzymes are dependent on Mg2+ while proceeding by a catalytic aspartate and acyl-phosphate intermediate. Mammalian PNKP executes its processes on a multitude of 3’-phoshate ends like those within nick,s gaps, DSBs, and single-stranded termini. Two narrow channels that are surrounded by large positively charged loops make a pathway to the phosphatase active site but aren’t wide enough to take in double-stranded substrates. This shows that either a requirement for remodeling of the phosphatase substrate binding surface or an unwinding of the DNA is needed to accommodate double-stranded substrates.

DNA repairing scaffold proteins, XRCC1 and XRCC4, interacts with PNKP function, mediated by binding of the PNKP FHA domain to phosphorylated motifs on XRCC1 and XRCC4. FHA domains, phospho-peptide binding modules, have a β-sandwich fold where a series of loops jut out from one side of the β-sandwich and provide a peptide binding surface with a marked preference for targets that contain a phospho-threonine residue. Even though XRCC1 and XRCC4 are structurally unrelated, they share similar motifs that are phosphorylated by CK2 and act as the binding sites for the PNKP FHA domain. A significant reduction in the efficiency of SSB repair occurs when a cluster of CK2 phosphorylation sites between residues 515 and 526 in XRCC1 is needed for interaction with PNKP and amino acid substitutions within this certain region. Similarly, a primary CK2 site in XRCC4, THr233, is needed for PNKP binding and for efficient repair of DSBs in vivo. Significant conservation of sequence is show around these sites. Phosphorylation of a conserved serine occurs and structure of the complex with regard to the primary phospho-threonine reveals a dynamic interaction of this residue with ARG35 or ARg44 of PNKP FHA domain. Tyrosine residue is conserved at the -4 position and asparagines residue is conserved at the +3 position. Some reactions aren’t conserved in the complex with XRCC4 in the FHA domain. The +3 position residue is a glutamic acid. Due to the peptides acidic properties and long-range electrostatic interactions between residues, the largely positively charged peptide-binding surface contributes to binding specificity. Threonine phosphorylation in the +4 position also plays a role to binding selectivity through the recruitment of a second PNKP FHA domain.

PNKP and single-strand break repair (SSBR)
The multienzyme pathway, SSBR, uses different participants depending on the causative agent. An example would be with IR-induced strand breaks that involves losing at least one nucleotide. The process of damage recognition and correction of the strand that is broken at the termini is carried out by enzymes poly(ADP-ribose)polymerase (PARP), XRCC1, AP endonuclease 1 and PNKP, with other proteins acting as backups to this functionality. By using a short patch pathway that involves DNA polymerase β and DNA ligase III or a long patch pathway that uses DNA polymerase δ and/or ɛ, the FEN1 endonuclease and DNA ligase I, consecutive replacement of nucleotides and strand resealing may occur. When IR occurs, APE1 removes 3′-phosphoglycolates while PNKP hydrolyses 3′-phosphate groups. This occurs when 3′-phosphatase activity of APE1 is much weaker than that of PNKP. Enzyme PNKP also plays a role in confirming that 5′-OH termini are phosphorylated. Due to the fact that phosphatase activity of PNKP is much more active than the kinase activity, when strand breaks with both 3′-phosphate and 5′-OH termini occur, the activity of PNKP is prioritized. Phosphatase activity in PNKP was shown to be important in the rapid repair of hydrogen peroxide-induced SSBs in mammalian cells when a failure of overexpression of phosphatase-defective PNIKP to compensate for Xrcc1defieicney occurred. Correspondingly, another important factor of PNKP phosphate activity involves a small molecule inhibitor that dramatically retards SSBR in irradiated human cells. While it is shown here that important phosphatase activity exists in PNP, the physiological important of the 5′-kinase activity has yet to be determined.

A commonly accepted model for repair of radiation-induced SSBS is when SSB catalyzes the polymerization of chains of ADP-ribose onto acceptor chromatin proteins and itself. BY doing this, SSBR attracts the scaffold protein, XRCC1 and maybe also tightly bound DNA ligase III. The proteins then in turn recruits PNKP or APE1 in order to restore the essential terminal groups for DNA polymerase β so that it can add the missing base and allow DNA ligase III to rejoin the strand. By researching and analyzing protein-protein interactions, it was found that direct interactions between XRCC1 and PNKP exist, as well as with DNA polymerase β and DNA ligase III. This shows that these connected partnerships include tetrameric complex between the four proteins. This formation could form for various models. While there is evidence that shows interactions between XRCC1 and PNKP, evidence also exists that counters the concept that XRCC1 recruits either PNKP or APE1 to the strand break. By using the technique of cross linking proteins to DNA substrates, experiments were conducted to track the temporal association of SSBR proteins in HeLa cell. Through this process of incubation, it was discovered that for substrates with either 3′-phosphoglycolate termini or 3′-phosphate termini, APE1 and PNKP, were recruited to the strand breaks before XRCC1/DNA ligase III. In addition to this discovery, it was found that immunodepletion of APE1 or PNKP diminished the binding of XRCC1 to the following substrates. This indicated that APE1 and PNKP inducted XRCC1 to sites of oxidative damage rather than in reverse. Conversely, PNKP foci were found to be in the nuclei of hydrogen peroxide-treated cells expressing XRCC1, but did not exist in cells lacking XRCC1. This shows that although XRCC1 might not be required in the beginning stages of PNKP or APE1, it expedites the focal accumulation and provocation of these specific enzymes at sites of chromosomal damage

Even though DNA repair protein XRCC1 lacks inherent enzymatic activity, it has the ability to enhance both kinase and phosphatase activities of PNKP. By using florescence measurements to work out the binding mechanism between PNKP and substrates that mimic different strand breaks, the mechanism surrounding XRCC1-induced stimulation was discovered. Even though PNKP bounded tightly to a nicked substrate with a 5′-OH terminus with a Kd value of 0.25 μM, this was only 5- to 6-fold tighter than PNKP binding to the identical duplex bearing a 5′-phosphate. This showed that PNKP stayed bounded to the product of its kinase activity. Results showed that the presence of XRCC1 did not influence the binding of PNKP to the nonphosphorylated substrate. But further results also showed that PNKP interaction with the phosphorylated duplex was abolished thus indicating that XRCC1 did influence the binding and displaced PNKP from the reaction product. By following the evidence of kinetics of product accumulation under limiting enzyme concentration, the result of the addition of XRCC1 increasing PNKP enzymatic turnover was confirmed. Further data has shown that similar kinetic data was observed for PNPK phosphatase activity.

The relationship between PNKP and XRCC1 is further complicated by CK2-mediated phosphorylation of XRCC1. While promoting interaction with other proteins, XRCC1 phosphorylation also works to stabilize the XRCC1-DNA ligase III complex. Observations were found of multiple sites of CK2-mediated XRCC1 phosphorylation involved in vitro, clustered within specific locations. In order to recruit XRCC1 and PNKP to nuclear foci in hydrogen peroxide-treated or γ-irradiated cells, XRCC1 phosphorylation is needed. XRCC1 phosphorylation is also needed to promote more rapid repair of SSBs. If a cell lacked XRCC1 phosphorylation, this would not impact cell survival. But through further research and analysis, it was found that cells without function XRCC1 with triple mutant XRCC1 would fail to fully restore rapid SSBR, showing that there indeed existed an important interaction with PNKP. Repair of the cell could easily be completed by overexpression of PNKP. This shows that XRCC1 plays an important role in increasing PNKP enzyme turnover, especially when the cell contains a limiting concentration of PNKP.

Phosphorylation of XRCC1 by CK2, compared to nonphosphorylated XRCC1, prompts the kinase and phosphatase activities of PNKP that are measured in vitro. In contrast, Stimulation by nonphosphorylated XRCC1 is due to enhanced enzymatic turnover of PNKP. This situation brings up problems since it can be seen that phosphorylated and nonphosphorylated XRCC1 bind PNKP at different site and with different affinities, but both are able to stimulate PNKP by a similar mechanism. Research found that while phosphorylated XRCC1 binds the FHA domain with a Kdvalue of 4 nM, the nonphosphorylated protein binds the catalytic domain of PNKP with a 10-fold weaker affinity. This indicates that a certain possibility of phosphorylation-independent interaction between PNKP and XRCC1 in human cells exists. Researchers found that PNKP co-immunoprecipitated with XRCC1 triple mutant that was expressed in human 293T cells. While 85–90% of the cellular XRCC1 is phosphorylated, this does not indicate that the key cluster of amino acids involved in interaction with the FHA domain is fully phosphorylated. An increase in phosphorylation at the cluster and an approximately 3-fold increase in PNKP copurifying with XRCC1 was due to treatment of cells with hydrogen peroxide. This shows that cells might play a role in enhancing CK2-mediated phosphorylation of XRCC1 and its subsequent interaction with PNKP FHA domain. This enhancement happens directly in response to a confrontation by hydrogen peroxide or radiation to deal with rather high levels of DNA damage in an efficient manner. On the opposite end of the spectrum, unstressed cells are able to cope with comparatively low level of endogenous DNA damage by using a different method. By using nonphosphorylated XRCC1, or XRCC1 with a restricted degree of phosphorylation, it is able to activate PNKP through binding to the catalytic domain.

Cells are sensitive to camptothecin due to PNKP depletion in its cells and Pnk1 deletion in fission yeast. XRCC1 overlooks the repair of these strand breaks by forming a complex with TDP1, DNA ligase III and PNKP. Neurodegenerative disorder, spinocerebellar ataxia with axonal neuropathy-1, is caused by mutation of TDP1. Research shows that SCAN 1 cells have a reduced capacity to repair Camptothecin-induced SSBs and also display slow repair of hydrogen peroxide-induced SSBs. This evidence proffers that TDP1 is important and required to repair lesions generated by oxidative processes, lesions that possibly justify neurodegeneration observed in SCAN1. Evidence for this was shown by experiments for fission yeast in G0, Tdp1 and Pnk1 that act sequentially in order to process the 3′-termini of naturally occurring SSBs.

PNKP and base excision repair (BER)
Cellular mechanism BER, base excision repair, is accountable for the repair of most minor base modifications determined by IR, ROS and alkylating agents. First step in the mechanism is to remove the modified base by DNA glycosylases and then cleave the DNA at the newly formed apurinic/apyrimidinic (AO) site using APE1. Another way would be to use glycosylases hydrolyze the AP site with its AP lyase activity. With the discovery of the nei endonuclease endonuclease VIII-like-1 (NEIL1) and NEIL2 mammalian DNA glycosylases, it was indisputable that PNKP was involved in the BER pathway. Nei endonuclease VIII-like-1 (NEIL1) and NEIL2 mammalian DNA glycosylases possess β,δ-AP lyase activity that generates 3′-phosphate termini. Instead of binding directly to PNKP, these glycosylases instead are associated with larger complexes that contain other BER components that include PNKP. The function of these glycosylases are to undertake a variety of base lesions that include: thymine glycol, 5-hydroxyuracil and 8-oxoguanine. In addition to this function, glycosylases can also cleave intact abasic sites that are generated by glycosylases that do not possess AP lyase activity, and the pentenal moiety generated by the β-elimination AP lyases of other DNA glycosylases. Because of this NEIL glycosylases would compete with APE1 thus forming the basis of a different, APE1-independent, BER pathway. Although current research can not indicate to what extent NEIL1- or NEIL2-catalyzed cleavage of abasic sites arises in cells, the cleavage of these sites could possibly explain for the increased sensitivity of PNKP-depleted cells to the alkylating agent methyl methanesulfonate (MMS). This sensitivity to MMS came as a surprise in the experiments due to major lesions inflicted by this agent being N7-methylguanine and N3-methyladenine, with little if at all any direct strand scission. Downregulating aprataxin expression also causes cells to be sensitive to MMS. But since human DNA glycosylase that are responsible for removing these methylated bases do not possess AP lyase activity, the ability to act upon the abasic sites generated by MPG to produce strand breaks with 3′-phosphate termini must fall to NEIL1 or NEIL2.

PNKP and double-strand break repair (DSBR)
In the two major double-strand break repair pathways, there is proof for PNKP participating in nonhomologous end joining. But in contrast, due to its failure to influence IR-induced sister chromatic exchange by PNKP deletion, this suggests that PNKP may actually not be involved in homologous recombination. In addition to the other pathways, PNKP plays an additional role as a back-up, XRCC1-dependent, DSB repair pathway. Experiments showed evidence for PNKP participation through using human cell-free extracts. This evidence showed that PNKP kinase activity was required before binding of linearized plasmid substrates bearing 5′-OH termini could happen. XRCC4 and DNA-PK were important in determining how successful phosphorylation was. In parallel to the role of XRCC1 linking PNKP to DNA ligase III, XRCC4 links PNKP to DNA ligase IV. CK2-mediated phosphorylation of XRCC4 Thr233 plays a role in interacting with the PNKP FHA domain and smoothly stimulating XRCC4–DNA ligase IV mediated ligation of a 5′-dephosphorylated plasmid substrate in vitro. In an Xrcc4-deficient cell line, when expression of XRCC4 occurs instead of wild-type XRCC4, the rate of survival is reduced by approximately 30% following irradiation and thus slowing down the rate of DSB repair.

The function role of the XRCC4-PNKP interaction was able to be determined by coming biophysical and biochemical examination. While phosphorylation of XRCC4 advocates a tight affinity for PNKP, nonphosphorylated XRCC4 also have the ability to bind to PNKP. Though in this particular case, binding is to the catalytic domain of PNKP thus weakening the affinity. Similar to the ability of XRCC1 stimulation of PNKP turnover from SSBs, nonphosphorylated XRCC4 has the ability to stimulate pNKP enzymatic turnover from DSBs. Research found that the presence of phosphorylated XRCC4 failed to stimulate PNKP and thus did block PNKP-mediated DNA phosphorylation. But with the additional attendance DNA ligase IV, the complex it forms with phosphorylated XRCC4 has the ability to reverses the inhibition and stimulate PNKP turnover. A ratio of XRCC4:DNA ligase IV:PNKP of ∼7:1:3 was found in the proteins in HeLA cells, with almost half of the XRCC4 vitally phosphorylated at Thr233. This shows that in cells, only a fraction of XRCC4 can be complexed to DNA ligase IV thus indicating a possibility for FHA-independent interaction between XRCC4 and PNKP. Using XRCC4 co-immunoprecipitation with PNKP, the FHA independent interaction between XRCC4 and PNKP was confirmed for expression in cells depleted of endogenous PNKP. PNKP also has an important function of processing DSB 3′-phosphoglycolate termini, especially 3′-overhanging and blunt-ended termini. These termini are produced by IR, bleomycin and enediyne compounds like neocarzinostatin. Even though APE1 has the ability to remove phosphoglycolate groups at SSB termini and recessed DSB termini, with blunt-ended DSB termini it loses its effectiveness and with overhanging termini it is completely ineffective.

Physiological roles and clinical potential of PNKP
PNKP is involved in several DNA repair pathways that work to protect cells from endogenous and exogenous genotoxic agents. Neurological disorders with various symptoms occur when disruption of NHEU genes and SSBR/BER genes occur. An example would be microcephaly. Microcephaly occurs in people with mutations in LIG4 that encodes DNA ligase IV. Deletion of Xrcc1 in mice causes seizures. Research has found that PNKP mutations are the cause of a sever neurological autosomal recessive disease that is characterized by microcephaly. Symptoms include intractable seizures and developmental delay. Through analysis of families, mutations were found in both the kinase and phosphatase domains. Through the collection of all the symptoms shown by patients with MCSZ, it shows the involvement of PNKP in multiple DNA repair pathways.

PNKP has also shown to be linked to pathophysiological conditions. It has been observed that elevated expression of PNP in arthrofibrotic tissue shows a role for PNKP in mitigating the effects of ROS generated by macrophages. It has also been observed in another experiment that physiologically and environmentally relevant doses of cadmium and copper are known to elicit neurotoxic and carcinogenic effects, thus inhibiting PNKP.

The concept of DNA repair capacity of tumor cells shows an important point in clinical response to many antineoplastic agents. Thus investigations are underway of inhibitors of several DNA repair enzymes like PNKP. They hold on to the ability to sensitize cells to radiation and chemotherapeutic drugs thus showing an important concept for research. Through this research, a small molecule inhibitor of PNKP phosphatase activity was identified and exhibited to heighten the sensitivity of cells to IR and camptothecin. This is the parent compound of two clinically important topoisomerase I poisons, irinotecan and topotecan, that are frequently used to treat colon and ovarian cancers.

Conclusion
PNKP is an important enzyme that is used in cellular processing of strand break termini. PNKP is involved in many DNA repair pathways due to its helpful properties. More research is needed to identify how it is regulated, how it collaborates with other repair enzymes, and physiological role in neurons and other tissues. PNKP is seen as a therapeutic target in treatment of cancer since it is involved in a variety of repair pathways. Therefore, new inhibitory compounds will need to be identified, researched, and optimized for clinical use. Further research should be invested in identifying synthetic lethal partners of PNKP in order to view its potential use as single agents against tumors deficient in proteins.