The crystal structure of Vibrio cholerae (6-4) photolyase reveals interactions with cofactors and a DNA-binding region

Photolyases (PLs) reverse UV-induced DNA damage using blue light as an energy source. Of these PLs, (6-4) PLs repair (6-4)-lesioned photoproducts. We recently identified a gene from Vibrio cholerae (Vc) encoding a (6-4) PL, but structural characterization is needed to elucidate specific interactions with the chromophore cofactors. Here, we determined the crystal structure of Vc (6-4) PL at 2.5 Å resolution. Our high-resolution structure revealed that the two well-known cofactors, flavin adenine dinucleotide and the photoantenna 6,7-dimethyl 8-ribityl-lumazin (DMRL), stably interact with an α-helical and an α/β domain, respectively. Additionally, the structure has a third cofactor with distinct electron clouds corresponding to a [4Fe-4S] cluster. Moreover, we identified that Asp106 makes a hydrogen bond with water and DMRL, which indicates further stabilization of the photoantenna DMRL within Vc (6-4) PL. Further analysis of the Vc (6-4) PL structure revealed a possible region responsible for DNA binding. The region located between residues 478 to 484 may bind the lesioned DNA, with Arg483 potentially forming a salt bridge with DNA to stabilize further the interaction of Vc (6-4) PL with its substrate. Our comparative analysis revealed that the DNA lesion could not bind to the Vc (6-4) PL in a similar fashion to the Drosophila melanogaster (Dm, (6-4)) PL without a significant conformational change of the protein. The 23rd helix of the bacterial (6-4) PLs seems to have remarkable plasticity, and conformational changes facilitate DNA binding. In conclusion, our structure provides further insight into DNA repair by a (6-4) PL containing three cofactors. Photolyases (PLs) reverse UV-induced DNA damage using blue light as an energy source. Of these PLs, (6-4) PLs repair (6-4)-lesioned photoproducts. We recently identified a gene from Vibrio cholerae (Vc) encoding a (6-4) PL, but structural characterization is needed to elucidate specific interactions with the chromophore cofactors. Here, we determined the crystal structure of Vc (6-4) PL at 2.5 Å resolution. Our high-resolution structure revealed that the two well-known cofactors, flavin adenine dinucleotide and the photoantenna 6,7-dimethyl 8-ribityl-lumazin (DMRL), stably interact with an α-helical and an α/β domain, respectively. Additionally, the structure has a third cofactor with distinct electron clouds corresponding to a [4Fe-4S] cluster. Moreover, we identified that Asp106 makes a hydrogen bond with water and DMRL, which indicates further stabilization of the photoantenna DMRL within Vc (6-4) PL. Further analysis of the Vc (6-4) PL structure revealed a possible region responsible for DNA binding. The region located between residues 478 to 484 may bind the lesioned DNA, with Arg483 potentially forming a salt bridge with DNA to stabilize further the interaction of Vc (6-4) PL with its substrate. Our comparative analysis revealed that the DNA lesion could not bind to the Vc (6-4) PL in a similar fashion to the Drosophila melanogaster (Dm, (6-4)) PL without a significant conformational change of the protein. The 23rd helix of the bacterial (6-4) PLs seems to have remarkable plasticity, and conformational changes facilitate DNA binding. In conclusion, our structure provides further insight into DNA repair by a (6-4) PL containing three cofactors. The cryptochrome (CRY)/photolyase (PL) family is a large family of flavoproteins that absorb near UV–visible light (300–500 nm range) and has diverse functions depending on the type of organism (1Kavakli I.H. Baris I. Tardu M. Gul S. Oner H. Cal S. et al.The photolyase/cryptochrome family of proteins as DNA repair enzymes and transcriptional repressors.Photochem. Photobiol. 2017; 93: 93-103Crossref PubMed Scopus (52) Google Scholar, 2Sancar A. Photolyase and cryptochrome blue-light photoreceptors.Adv. Protein Chem. 2004; 69: 73-100Crossref PubMed Scopus (67) Google Scholar). CRYs are the circadian photoreceptors in Drosophila and plants, while they act as transcriptional repressors of the circadian clock in mammals (2Sancar A. Photolyase and cryptochrome blue-light photoreceptors.Adv. Protein Chem. 2004; 69: 73-100Crossref PubMed Scopus (67) Google Scholar, 3Chaves I. Pokorny R. Byrdin M. Hoang N. Ritz T. Brettel K. et al.The cryptochromes: blue light photoreceptors in plants and animals.Annu. Rev. Plant Biol. 2011; 62: 335-364Crossref PubMed Scopus (602) Google Scholar, 4Kavakli I.H. Sancar A. Analysis of the role of intraprotein electron transfer in photoreactivation by DNA photolyase in vivo.Biochemistry. 2004; 43: 15103-15110Crossref PubMed Scopus (74) Google Scholar). PLs are enzymes that repair UV-induced DNA damage (5Kavakli I.H. Ozturk N. Gul S. DNA repair by photolyases.Adv. Protein Chem. Struct. Biol. 2019; 115: 1-19Crossref PubMed Scopus (21) Google Scholar, 6Sancar A. Mechanisms of DNA repair by photolyase and excision nuclease (Nobel lecture).Angew. Chem. Int. Ed. Engl. 2016; 55: 8502-8527Crossref PubMed Scopus (164) Google Scholar, 7Zhong D. Electron transfer mechanisms of DNA repair by photolyase.Annu. Rev. Phys. Chem. 2015; 66: 691-715Crossref PubMed Scopus (70) Google Scholar). Although phylogenetic analyses revealed different classes of PLs, there are three different types of PLs based on their repair activities: Cyclobutane Pyr dimer (CPD) PLs, (6–4) PLs, and Drosophila, Arabidopsis, Synechocystis, and Human (DASH)-type CRY (CRY-DASH) (8Asimgil H. Kavakli I.H. Purification and characterization of five members of photolyase/cryptochrome family from Cyanidioschyzon merolae.Plant Sci. 2012; 185-186: 190-198Crossref PubMed Scopus (31) Google Scholar, 9Ozturk N. Phylogenetic and functional classification of the photolyase/cryptochrome family.Photochem. Photobiol. 2017; 93: 104-111Crossref PubMed Scopus (66) Google Scholar, 10Ozturk N. Kao Y.T. Selby C.P. Kavakli I.H. Partch C.L. Zhong D. et al.Purification and characterization of a type III photolyase from Caulobacter crescentus.Biochemistry. 2008; 47: 10255-10261Crossref PubMed Scopus (42) Google Scholar). CPD PLs repair mainly CPDs (Pyr<>Pyr), while (6–4) PLs repair Pyr-Pyr photoproducts (Pyr [6–4] Pyr) using blue light (350–500 nm) as the energy source (1Kavakli I.H. Baris I. Tardu M. Gul S. Oner H. Cal S. et al.The photolyase/cryptochrome family of proteins as DNA repair enzymes and transcriptional repressors.Photochem. Photobiol. 2017; 93: 93-103Crossref PubMed Scopus (52) Google Scholar, 6Sancar A. Mechanisms of DNA repair by photolyase and excision nuclease (Nobel lecture).Angew. Chem. Int. Ed. Engl. 2016; 55: 8502-8527Crossref PubMed Scopus (164) Google Scholar). CRY-DASHs have quite diverse physiological functions including ssDNA PL activity (11Selby C.P. Sancar A. A cryptochrome/photolyase class of enzymes with single-stranded DNA-specific photolyase activity.Proc. Natl. Acad. Sci. U. S. A. 2006; 103: 17696-17700Crossref PubMed Scopus (249) Google Scholar). PLs are monomeric proteins with molecular masses of 50 to 61 kDa and possess two chromophores (12Sancar A. Structure and function of DNA photolyase and cryptochrome blue-light photoreceptors.Chem. Rev. 2003; 103: 2203-2237Crossref PubMed Scopus (1056) Google Scholar). Flavin adenine dinucleotide (FAD) is an essential cofactor for the catalytic activity of PLs and acts as a catalytic cofactor in all PLs (5Kavakli I.H. Ozturk N. Gul S. DNA repair by photolyases.Adv. Protein Chem. Struct. Biol. 2019; 115: 1-19Crossref PubMed Scopus (21) Google Scholar, 12Sancar A. Structure and function of DNA photolyase and cryptochrome blue-light photoreceptors.Chem. Rev. 2003; 103: 2203-2237Crossref PubMed Scopus (1056) Google Scholar). The second chromophore, which acts as a photoantenna, varies depending on the organism. The following photoantenna are being found: methenyltetrahydrofolate, 8-hydroxy-7,8-didemethyl-5-deazariboflavin, FAD, FMN, or 6,7-dimethyl 8-ribityl-lumazin (DMRL) depending on the type of PL (13Zhang F. Scheerer P. Oberpichler I. Lamparter T. Krauss N. Crystal structure of a prokaryotic (6-4) photolyase with an Fe-S cluster and a 6,7-dimethyl-8-ribityllumazine antenna chromophore.Proc. Natl. Acad. Sci. U. S. A. 2013; 110: 7217-7222Crossref PubMed Scopus (80) Google Scholar, 14Zhang M. Wang L. Zhong D. Photolyase: dynamics and electron-transfer mechanisms of DNA repair.Arch. Biochem. Biophys. 2017; 632: 158-174Crossref PubMed Scopus (63) Google Scholar). The (6–4) PLs had previously been thought to exist only in eukaryotes. However, studies with different bacteria revealed the presence of (6–4) PLs (15Dikbas U.M. Tardu M. Canturk A. Gul S. Ozcelik G. Baris I. et al.Identification and characterization of a new class of (6-4) photolyase from Vibrio cholerae.Biochemistry. 2019; 58: 4352-4360Crossref PubMed Scopus (12) Google Scholar, 16Geisselbrecht Y. Fruhwirth S. Schroeder C. Pierik A.J. Klug G. Essen L.O. CryB from Rhodobacter sphaeroides: a unique class of cryptochromes with new cofactors.EMBO Rep. 2012; 13: 223-229Crossref PubMed Scopus (72) Google Scholar, 17Oberpichler I. Pierik A.J. Wesslowski J. Pokorny R. Rosen R. Vugman M. et al.A photolyase-like protein from Agrobacterium tumefaciens with an iron-sulfur cluster.PLoS One. 2011; 6e26775Crossref PubMed Scopus (50) Google Scholar). These bacterial (6–4) PLs are distinct in that they contain [4 Fe-4 S] as well as DMRL and FAD cofactors (15Dikbas U.M. Tardu M. Canturk A. Gul S. Ozcelik G. Baris I. et al.Identification and characterization of a new class of (6-4) photolyase from Vibrio cholerae.Biochemistry. 2019; 58: 4352-4360Crossref PubMed Scopus (12) Google Scholar, 17Oberpichler I. Pierik A.J. Wesslowski J. Pokorny R. Rosen R. Vugman M. et al.A photolyase-like protein from Agrobacterium tumefaciens with an iron-sulfur cluster.PLoS One. 2011; 6e26775Crossref PubMed Scopus (50) Google Scholar). Crystal structures of various PLs have been determined by X-ray crystallography. The first available crystal structure is Escherichia coli CPD PL (18Park H.W. Kim S.T. Sancar A. Deisenhofer J. Crystal structure of DNA photolyase from Escherichia coli.Science. 1995; 268: 1866-1872Crossref PubMed Scopus (503) Google Scholar). Subsequent structures have been obtained from different organismal PLs including Anacystis nidulans (19Tamada T. Kitadokoro K. Higuchi Y. Inaka K. Yasui A. de Ruiter P.E. et al.Crystal structure of DNA photolyase from Anacystis nidulans.Nat. Struct. Biol. 1997; 4: 887-891Crossref PubMed Scopus (196) Google Scholar), Thermus thermophilus (20Komori H. Masui R. Kuramitsu S. Yokoyama S. Shibata T. Inoue Y. et al.Crystal structure of thermostable DNA photolyase: pyrimidine-dimer recognition mechanism.Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13560-13565Crossref PubMed Scopus (106) Google Scholar), (6–4) PLs of Arabidopsis thaliana (21Hitomi K. DiTacchio L. Arvai A.S. Yamamoto J. Kim S.T. Todo T. et al.Functional motifs in the (6-4) photolyase crystal structure make a comparative framework for DNA repair photolyases and clock cryptochromes.Proc. Natl. Acad. Sci. U. S. A. 2009; 106: 6962-6967Crossref PubMed Scopus (102) Google Scholar) and Agrobacterium tumefaciens (13Zhang F. Scheerer P. Oberpichler I. Lamparter T. Krauss N. Crystal structure of a prokaryotic (6-4) photolyase with an Fe-S cluster and a 6,7-dimethyl-8-ribityllumazine antenna chromophore.Proc. Natl. Acad. Sci. U. S. A. 2013; 110: 7217-7222Crossref PubMed Scopus (80) Google Scholar). The comparison of the crystal structures of various homologous PLs revealed valuable information about their structural features and enzyme mechanisms. PLs are made up of two well-defined domains which are α/β and α-helical domains. The α-helical domain is associated with FAD, while α/β domain is the binding site for the second chromophore (12Sancar A. Structure and function of DNA photolyase and cryptochrome blue-light photoreceptors.Chem. Rev. 2003; 103: 2203-2237Crossref PubMed Scopus (1056) Google Scholar). These two domains are connected by a long flexible loop. Vibrio cholerae O1 bivar Tor str. N16961 has a CPD PL that repairs T<>T dimers and two CRY-DASHs that repair T<>T dimers in ssDNA (11Selby C.P. Sancar A. A cryptochrome/photolyase class of enzymes with single-stranded DNA-specific photolyase activity.Proc. Natl. Acad. Sci. U. S. A. 2006; 103: 17696-17700Crossref PubMed Scopus (249) Google Scholar, 22Worthington E.N. Kavakli I.H. Berrocal-Tito G. Bondo B.E. Sancar A. Purification and characterization of three members of the photolyase/cryptochrome family blue-light photoreceptors from Vibrio cholerae.J. Biol. Chem. 2003; 278: 39143-39154Abstract Full Text Full Text PDF PubMed Scopus (92) Google Scholar). Our previous study with a gene (Vca0809) in V. cholerae highly expressed upon exposure of the organism to blue light (23Tardu M. Bulut S. Kavakli I.H. MerR and ChrR mediate blue light induced photo-oxidative stress response at the transcriptional level in Vibrio cholerae.Sci. Rep. 2017; 7: 40817Crossref PubMed Scopus (32) Google Scholar). Our further characterization of this gene revealed that it encodes (6-4) PL with FAD and DMRL and consists of the [4Fe-4S] domain (15Dikbas U.M. Tardu M. Canturk A. Gul S. Ozcelik G. Baris I. et al.Identification and characterization of a new class of (6-4) photolyase from Vibrio cholerae.Biochemistry. 2019; 58: 4352-4360Crossref PubMed Scopus (12) Google Scholar). Here, we determined the crystal structure of Vc (6-4) PL to 2.5 Å resolution. Our crystal structure comparison revealed a putative DNA-binding region of Vc (6-4) PL. We recently identified a (6-4) PL from V. cholerae O1 bivar Tor str. N16961(Vc) which possesses the catalytic cofactor FAD, an antenna chromophore DMRL, and an additional cofactor [4Fe-4S] cluster (15Dikbas U.M. Tardu M. Canturk A. Gul S. Ozcelik G. Baris I. et al.Identification and characterization of a new class of (6-4) photolyase from Vibrio cholerae.Biochemistry. 2019; 58: 4352-4360Crossref PubMed Scopus (12) Google Scholar). We determined the crystal structure of Vc (6-4) PL at 2.5 Å resolution at cryogenic temperature at the Turkish Light Source known as “Turkish DeLight” (24Atalay N. Akcan E.K. Gül M. Ayan E. Destan E. Ertem F.B. et al.Cryogenic X-ray crystallographic studies of biomacromolecules at Turkish light source “Turkish DeLight”.bioRxiv. 2022; ([preprint])https://doi.org/10.1101/2022.09.03.506456Crossref Scopus (0) Google Scholar, 25Gul M. Ayan E. Destan E. Johnson J.A. Shafiei A. Kepceoğlu A. et al.Rapid and high resolution ambient temperature structure determination at Turkish light source.bioRxiv. 2022; ([preprint])https://doi.org/10.1101/2022.10.12.511637Crossref Scopus (0) Google Scholar). The structure contains all the cofactors with well-defined electron densities and contains a total of 24 α helixes and five β sheets (Fig. 1A). It consists of two domains: an α-helical domain that is associated with FAD and α/β domain which interacts with DMRL cofactor. The sequence alignment of Vc (6-4) PL with A. tumefaciens (At) and Rhodobacter sphaeroides (Rs) (6-4) PLs showed that it has 38.4% and 43.1% sequence similarity, respectively (Fig. 1B). DMRL is a photoantenna chromophore that absorbs light more efficiently than FAD for increasing catalytic efficiency (13Zhang F. Scheerer P. Oberpichler I. Lamparter T. Krauss N. Crystal structure of a prokaryotic (6-4) photolyase with an Fe-S cluster and a 6,7-dimethyl-8-ribityllumazine antenna chromophore.Proc. Natl. Acad. Sci. U. S. A. 2013; 110: 7217-7222Crossref PubMed Scopus (80) Google Scholar). In our crystal structure, there is a very well-defined electron density of the DMRL and interacting amino acid residues (Fig. 2A). The aromatic ring of DMRL interacts with nitrogen atoms of Gln11, Leu35, Gln39, and oxygen atom of Ala33 (Fig. 2B). Additionally, the aromatic ring of DMRL interacts with Trp488, Ala33, and Ile9. The DMRL's ribityl group forms hydrogen bonds with the oxygen atoms of Glu38, Tyr41, Asp12, and Asp10's amino group. When we compare the interaction partners of DMRL in Vc (6-4) PL with other known bacterial (6-4) PLs, almost all the interactions are well conserved with a few exceptions. First, the Cys32 carboxyl group of At (6-4) PL interacts with the aromatic ring in DMRL (Protein Data Bank (PDB) ID: 4DJA), while in both Vc and Rs (6-4) PLs, DMRL interacts with the main chain carbonyl group of Ala33. We identified a critical water molecule (W22) that interacts with Asp12, Gln13, Asp106, and DMRL in Vc (6-4) PL (Fig. 2C). Analyses of other bacterial (6-4) PLs have shown similar conserved motifs (Fig. 2C). Such motifs increase the overall stability of the DMRL within the enzyme. However, the W22 molecule interacts with Asp106 of the Vc (6-4) PL with a distance of 2.8 Å, while the same interaction occurs with Gly105 of other bacterial (6-4) PLs with a distance of over 3.3 Å. These differences might increase the affinity of DMRL with Vc (6-4) PL. FAD has a U-shaped conformation in the α-helical domain (Fig. 3A). The N5 and O4 of FAD form hydrogen bonds with W7 and additional hydrogen bonds with the carbonyl group of Tyr398 and NH1 atom of Arg376 residue (Fig. 3B). As previously suggested for other PLs (17Oberpichler I. Pierik A.J. Wesslowski J. Pokorny R. Rosen R. Vugman M. et al.A photolyase-like protein from Agrobacterium tumefaciens with an iron-sulfur cluster.PLoS One. 2011; 6e26775Crossref PubMed Scopus (50) Google Scholar) these critical interactions may also serve a stabilization function Vc (6-4) PL. Unlike At (6-4) PL or Rs (6-4) PL, Vc (6-4) PL Glu410 amino acid faces away from FAD and interacts with Asp395 rather than a His365 residue (Fig. 3B). This results in a much-relaxed interaction between the residues due to hydrogen bonding formation rather than salt bridge formation in Vc (6-4) PL. Although the reaction mechanism of the CPD PLs has been elucidated, (6-4) PLs are not yet fully characterized. There are different mechanisms that are proposed for the (6–4) PL mechanism during DNA repair (7Zhong D. Electron transfer mechanisms of DNA repair by photolyase.Annu. Rev. Phys. Chem. 2015; 66: 691-715Crossref PubMed Scopus (70) Google Scholar, 14Zhang M. Wang L. Zhong D. Photolyase: dynamics and electron-transfer mechanisms of DNA repair.Arch. Biochem. Biophys. 2017; 632: 158-174Crossref PubMed Scopus (63) Google Scholar, 26Faraji S. Dreuw A. Insights into light-driven DNA repair by photolyases: challenges and opportunities for electronic structure theory.Photochem. Photobiol. 2017; 93: 37-50Crossref PubMed Scopus (29) Google Scholar, 27Matsubara T. Araida N. Hayashi D. Yamada H. Computational study on the mechanism of the electron-transfer-induced repair of the (6-4) T-T photoproduct of DNA by photolyase: possibility of a radical cation pathway.B Chem. Soc. Jpn. 2014; 87: 390-399Crossref Scopus (4) Google Scholar). In fact, mutagenesis studies with Drosophila melanogaster (Dm (6-4)) PL indicated that replacement of His365 with Asn365 results in the loss of its DNA repair activity, while mutagenesis of His369 into Met369 results in a highly reduced activity of Dm (6-4) PL (28Maul M.J. Barends T.R.M. Glas A.F. Cryle M.J. Domratcheva T. Schneider S. et al.Crystal structure and mechanism of a DNA (6-4) photolyase.Angew. Chem. Int. Ed. Engl. 2008; 47: 10076-10080Crossref PubMed Scopus (161) Google Scholar). However, mutagenesis studies with corresponding amino acid residues in Xenopus (6-4) PL showed a complete loss of activity (29Hitomi K. Nakamura H. Kim S.T. Mizukoshi T. Ishikawa T. Iwai S. et al.Role of two histidines in the (6-4) photolyase reaction.J. Biol. Chem. 2001; 276: 10103-10109Abstract Full Text Full Text PDF PubMed Scopus (110) Google Scholar). Femtosecond spectroscopy and site-directed mutagenesis suggest initial electron transfer from excited flavin induces transfer of a proton from a histidine on the active site of the enzyme to the (6–4) photoproduct (7Zhong D. Electron transfer mechanisms of DNA repair by photolyase.Annu. Rev. Phys. Chem. 2015; 66: 691-715Crossref PubMed Scopus (70) Google Scholar, 30Li J. Liu Z. Tan C. Guo X. Wang L. Sancar A. et al.Dynamics and mechanism of repair of ultraviolet-induced (6-4) photoproduct by photolyase.Nature. 2010; 466: 887-890Crossref PubMed Scopus (171) Google Scholar). Catalytically active His373 (corresponds to His365 of the Dm (6-4) PL) is conserved in Vc (6-4) PL (Fig. 3C). In addition, there are other conserved amino acid residues in the motif. For instance, His372 and Arg376 (corresponds to His364 and Arg368 of the Dm (6-4) PL) are conserved in Vc (6-4) PL (Fig. 3C). However, the conformation of the Vc (6-4) PL His372 is significantly different from the Dm (6-4) PL His364, which may suggest an alternate function of this residue. The high-resolution Vc (6-4) PL structure contains an iron–sulfur cluster with an unknown function and a well-defined electron density (Fig. 4A). Its interactions with the surrounding cysteine amino acids are summarized in Figure 4B. Cys357, Cys445, Cys448, and Cys461 sulfide atoms make bonds with Fe atoms in the cluster with each bond at 2.3 Å distance. The function of the [4Fe-4S] cluster is currently unknown; however, its distance to the FAD is 17 Å, which indicates it does not have a catalytic role in the DNA repair due to its large distance to the active site region. [4Fe-4S] clusters are well-known for their oxygen sensitivity, and under aerobic conditions, they quickly decompose into [3Fe-4S] clusters and can further dissociate into [2Fe-2S]. Mostly, the decomposition of the cluster deactivates the cluster-containing enzymes, and handling and purification of these enzymes require using a glovebox to prevent decomposition of the oxygen-sensitive [4Fe-4S] clusters. In the case of Vc (6-4) PL, the entire expression, purification, and crystallization procedures were performed in atmospheric conditions that took months. In our structure, there was no observable damage in the [4Fe-4S] cluster, which might indicate the cluster in the PL is unusually oxygen tolerant. The DNA-binding region of the Vc (6-4) PL has a defined electron density that covers the amino acids Asn178, Phe179, Asp180, Ala181, Asp182, Asn183, Arg184, and Asn185 (Fig. 5A). In contrast, this region has not a well-defined electron density in At (6-4) PL structures obtained in either cryogenic (PDB ID: 4DJA) or ambient temperature (PDB ID: 6DD6) conditions (Fig. 5, B and C). A similar analysis was carried out with Rs (6-4) PL, where there is a well-defined electron density in the same region (Fig. 5D). These results suggest that while this region is highly stable in Rs (6-4) PL and Vc (6-4) PL, it is flexible and disordered in At (6-4) PL structure and, therefore, this region has species-specific stability differences. Notably, we have not observed a distinct electron density of amino acid residues between 478 and 485 Vc (6-4) PL (Fig. 6A). However, this region has well-defined structures in other bacterial PLs (Fig. 6, B–D). Interestingly, this region is adjacent to the 23rd helix, which is one of the positively surface-charged helices around the catalytic region (Fig. 7A). When we superimposed Vc (6-4) PL with Dm (6-4) PL (PDB ID: 3CVU) with (6-4) T-T, this region and the 23rd helix were seen to pass through damaged DNA (Fig. 7B). This observation suggests that the region between 478 and 485 Vc (6-4) PL might be a DNA-binding region and further stabilizes after interacting with the DNA lesion. Upon investigation of the C-terminus of Vc (6-4) PL and Dm (6-4) PL, two additional helices were identified on Vc (6-4) PL and the 23rd helix seems to be blocking the DNA lesion in superimposed structure. Further analysis of superimposed structures indicated that the catalytic cofactor FADs are in similar conformation and their RMSD is 0.7 Å (Fig. 7C). To DNA lesions to bind the bacterial (6-4) PL, while still interacting with the FAD, there may be conformational changes on the 23rd helix to provide a gap for DNA lesion. B-factor analysis showed that 23rd helix has the highest B-factor values in the structure and therefore, it has the highest probability to undergo a conformational change (Fig. 7D). According to the superimposition of DNA lesion from Dm (6-4) PL with Vc (6-4) PL, which is compatible with a prototypical PL–DNA-binding mode, helix α23 of Vc (6-4) PL interacts with the DNA, while helix α24 looks away from the DNA lesion. The surface charge of helix α23 is positive, which is consistent with its role in DNA binding. The last two helixes (α23 and α24) are missing in Dm (6-4) PL or most of the other PLs, which indicates they may play a regulatory role rather than their involvement in repair activity.Figure 7DNA-binding comparison between Vc (6-4) PL and Dm (6-4) PL. A, generated surface density charge of Vc (6-4) PL, where blue is positive while red is negatively charged. Helices with positive surface charge are indicated in the structure. B, superimposed (RMSD: 3.944 Å) image of Vc (6-4) PL and Dm (6-4) PL with (6-4) DNA lesion (PDB ID: 3CVU). Deep teal is for Vc (6-4) PL, while tv_orange is for Dm (6-4) PL. C, FAD comparison with superimposed images of Vc (6-4) PL and Dm (6-4) PL. All atoms RMSD between the structures for FAD cofactor is 0.7 Å. D, B-factor analysis of the C-terminal of Vc (6-4) PL, where the highest B-factor has a yellowish color, while the lowest B-factor has a darker blue color. α23 has the highest B-factor in the structure. FAD, flavin adenine dinucleotide; PDB, Protein Data Bank; PL, photolyase.View Large Image Figure ViewerDownload Hi-res image Download (PPT) PLs belong to a large family of CRY/PL responsible for repairing the UV-induced DNA damages. Different types of PLs have been discovered, and they are shown to repair two different types of DNA damages: CPDs (Pyr<>Pyr) and Pyr-Pyr photoproducts (Pyr[6–4]Pyr). CPD photoproducts are mainly repaired by CPD PLs, while (6-4) photoproducts are being repaired by (6-4) PLs (12Sancar A. Structure and function of DNA photolyase and cryptochrome blue-light photoreceptors.Chem. Rev. 2003; 103: 2203-2237Crossref PubMed Scopus (1056) Google Scholar). The reaction mechanism of CPD PLs is elucidated at in vivo levels (1, 5, & 6). On the other hand, different reaction mechanisms are proposed for the (6-4) PL. The (6-4) PLs are initially thought to be specifically present in eukaryotes (13Zhang F. Scheerer P. Oberpichler I. Lamparter T. Krauss N. Crystal structure of a prokaryotic (6-4) photolyase with an Fe-S cluster and a 6,7-dimethyl-8-ribityllumazine antenna chromophore.Proc. Natl. Acad. Sci. U. S. A. 2013; 110: 7217-7222Crossref PubMed Scopus (80) Google Scholar). However, studies revealed that bacteria also possess (6-4) PLs (15Dikbas U.M. Tardu M. Canturk A. Gul S. Ozcelik G. Baris I. et al.Identification and characterization of a new class of (6-4) photolyase from Vibrio cholerae.Biochemistry. 2019; 58: 4352-4360Crossref PubMed Scopus (12) Google Scholar, 16Geisselbrecht Y. Fruhwirth S. Schroeder C. Pierik A.J. Klug G. Essen L.O. CryB from Rhodobacter sphaeroides: a unique class of cryptochromes with new cofactors.EMBO Rep. 2012; 13: 223-229Crossref PubMed Scopus (72) Google Scholar, 31Hendrischk A.K. Fruhwirth S.W. Moldt J. Pokorny R. Metz S. Kaiser G. et al.A cryptochrome-like protein is involved in the regulation of photosynthesis genes in Rhodobacter sphaeroides.Mol. Microbiol. 2009; 74: 990-1003Crossref PubMed Scopus (37) Google Scholar). Comparison of the crystal structure of the Vc (6-4) PL with other known bacterial (6-4) PLs indicated that cofactor-binding regions are well conserved with some differences: a water molecule interacts with the Asp106 in Vc (6-4) PL with a very short distance compared to the At and Rs (6-4) PLs, where the water molecule interacts with Gly105 (Fig. 2B). The FAD-binding region of the Vc (6-4) PL is highly conserved. However, one notable difference is that Glu410 makes a hydrogen bond with Asp395; whereas, in the case of Glu399 (corresponds to Glu410 of Vc (6-4) PL), it makes a salt bridge with His384 (corresponds to Asp395 of Vc (6-4) PL) At (6-4) PL and Rs (6-4) PL (13Zhang F. Scheerer P. Oberpichler I. Lamparter T. Krauss N. Crystal structure of a prokaryotic (6-4) photolyase with an Fe-S cluster and a 6,7-dimethyl-8-ribityllumazine antenna chromophore.Proc. Natl. Acad. Sci. U. S. A. 2013; 110: 7217-7222Crossref PubMed Scopus (80) Google Scholar, 16Geisselbrecht Y. Fruhwirth S. Schroeder C. Pierik A.J. Klug G. Essen L.O. CryB from Rhodobacter sphaeroides: a unique class of cryptochromes with new cofactors.EMBO Rep. 2012; 13: 223-229Crossref PubMed Scopus (72) Google Scholar). Whether such structural difference affects its activity or conformation of FAD needs to be further investigated. The His-His-X-X-Arg motif in (6-4) PLs are shown to be important for their catalytic activity (16Geisselbrecht Y. Fruhwirth S. Schroeder C. Pierik A.J. Klug G. Essen L.O. CryB from Rhodobacter sphaeroides: a unique class of cryptochromes with new cofactors.EMBO Rep. 2012; 13: 223-229Crossref PubMed Scopus (72) Google Scholar). This motif is also conserved in the Vc (6-4) PL. Further conformational comparison of the Vc (6-4) PL and Dm (6-4) PL of this motif indicates His373 and Arg376 of Dm (6-4) PL have identical configuration, while His372 (His364 of Dm (6-4) PL) acquired a different configuration, which may reveal similar/different functionalities between species. One of the major differences between Vc (6-4) PL and At (6-4) PL structures is their predicted DNA-binding site (Fig. 5). The predicted site (amino acid residues between 178 and 185) of Vc (6-4) PL has a distinct structure compared to that of At (6-4) PL, where the corresponding region has no well-defined structure (17Oberpichler I. Pierik A.J. Wesslowski J. Pokorny R. Rosen R. Vugman M. et al.A photolyase-like protein from Agrobacterium tumefaciens with an iron-sulfur cluster.PLoS One. 2011; 6e26775Crossref PubMed Scopus (50) Google Scholar). The same region of Rs (6-4) PL also has a well-ordered electron density map except for its Arg residue (16Geisselbrecht Y. Fruhwirth S. Schroeder C. Pierik A.J. Klug G. Essen L.O. CryB from Rhodobacter sphaeroides: a unique class of cryptochromes with new cofactors.EMBO Rep. 2012; 13: 223-229Crossref PubMed Scopus (72) Google Scholar), which indicates species-specific stability differences of DNA-binding domains. Comparison of Vc (6-4) PL and At (6-4) PL revealed that Vc (6-4) PL has an unstructured region with eight amino acid residues (Fig. 6A), which might be a DNA-binding region and is also conserved in other bacterial PLs. Arg483 might form a salt bridge with (6-4) photoproduct that stabilizes the disordered region. The general structure of Vc and Dm (6-4) PLs are highly similar except in C terminus, where Vc (6-4) PLs contain additional two helixes. Due to the position of the 23rd helix, DNA lesions cannot bind the Vc (6-4) PL similar to Dm (6-4) PL (Fig. 7B). However, the FADs are located in an extremely similar manner, which indicates a similar DNA lesion binding between the PLs (Fig. 7C). A conformational change upon DNA binding at the Vc (6-4) PL might open a space for DNA to bind. The 23rd helix is the main candidate for the conformational change, as it is the major blocker of the DNA and has the largest B-factor (Fig. 7D). Full-length (6-4) PL from V. cholerae O1 bivar Tor str. N16961 in a pET28a(+) vector between NdeI and BamHI cut sites was purchased from GenScript Biotech Corporation as codon optimized for E. coli protein expression. The plasmid was transformed into E. coli BL21 Rosetta-2 strain and grown in 4.5 L of LB-Miller liquid growth media supplemented with 50 μg/ml kanamycin and 35 μg/ml chloramphenicol antibiotics at 37 °C with 110 rpm shaking. When A600 reached to 0.8, induction of bacterial protein expression was performed by adding 0.4 mM of IPTG as the final concentration, and induction was performed at 18 °C for 24 h. The culture was centrifuged at 2850 g for 45 min at 4 °C, and the pellet was stored at −80 °C until further use. Lysis buffer containing 500 mM NaCl, 50 mM Tris–HCl pH 7.5, 10% glycerol (v/v), and 0.1% Triton X-100 (v/v) was added into pellets, and sonication (Branson W250 sonifier) was performed for 45 s at 60% power 3 times. Samples were ultracentrifuged at 35,000 rpm with a Ti-45 rotor (Beckman Coulter) for 1 h at 4 °C; subsequently, the supernatant was filtered with a 45 μm cellulose mixed ester filter (ISOLAB) and then applied to Ni-NTA agarose resin (QIAGEN). The column was equilibrated with His A buffer containing 200 mM NaCl, 20 mM Tris–HCl pH 7.5 (v/v), and 5% glycerol (v/v); then the sample was loaded with 2.5 ml/min flow rate. Washing of the column was performed with His A buffer, and protein was eluted with His B buffer, containing 200 mM NaCl, 20 mM Tris–HCl pH 7.5, 250 mM imidazole, and 5% glycerol (v/v) into 5 ml of 100% glycerol to prevent protein precipitation. Yellow-colored (6-4) PL was flash frozen with liquid nitrogen and stored at −80 °C with a final concentration of 25% glycerol (v/v). Initial crystallization screening experiments were setup with a 5 mg/ml final concentration of Vc (6-4) PL with the sitting drop, microbatch screening (under oil) method at 4 °C by mixing 0.83 μl of protein with an equal volume of commercial crystallization screening conditions in 72-well Terasaki plates. The well containing the protein:cocktail mixture was covered with 16.6 μl of parafilm oil to allow slow evaporation of solvents. Approximately, 3500 commercially available crystal screening conditions were tested, and the yellow-colored crystals were observed at Wizard Synergy #40 (Rigaku Corporation), which contains 2 M of ammonium citrate/citric acid pH 7.5 and 5% PEG 400 (v/v) after 7 weeks. The conditions were further optimized by mixing 0.5 μl Cryo-Pro #45 (Hampton Research), which contain 1 M sodium sulfate decahydrate, 1 μl of protein, and 1 μl of Wizard Synergy #40. The crystals were flash frozen by quickly plunging them into liquid nitrogen, and data collection was performed at 100 K. Two X-ray diffraction datasets were collected at 2.0 Å and 2.5 Å resolutions from two large crystals with a wavelength of 1.54 Å at the University of Health Sciences with Rigaku’s XtaLAB Synergy Flow X-ray diffractometer (Rigaku Corporation). The PhotonJet-R (Rigaku Corporation) X-ray generator was operated at 40 kV and 30 mA with a 23% beam intensity to mitigate the radiation damage during data collection. HyPix-Arc-150° detector (Rigaku Corporation) was used with a 60 mm detector distance. To further minimize the exposure time, we preferred not to collect fine-sliced oscillation data rather 1-degree oscillation scan width low X-ray dose chosen with 15 and 30 s exposure times. The total run time for two data collections was 1 h 11 m 30 s and 2 h 33 min 0 s, respectively. Profit merge is done by using the data from both crystals and data reduction is performed with CrysAlis Pro (32Diffraction, R. O.CrysAlisPro Software System, Version 1.171.42.51a. Rigaku Corporation, Wroclaw, Poland2021Google Scholar) software version 171.42.51a (https://www.rigaku.com/products/crystallography/crysalis). Two datasets were merged with 99.6 % completeness and 60.6 fold multiplicity. Unit cell dimensions were a = 200.8 Å, b = 200.8 Å, c = 77.0 Å, α = 90, β = 90, and γ = 120, in space group P64 22. Swiss model (33Bienert S. Waterhouse A. de Beer T.A. Tauriello G. Studer G. Bordoli L. et al.The SWISS-MODEL repository-new features and functionality.Nucleic Acids Res. 2017; 45: D313-D319Crossref PubMed Scopus (940) Google Scholar) was used for building a search model for structure determination that gave the best result with PDB ID: 4DJA (13Zhang F. Scheerer P. Oberpichler I. Lamparter T. Krauss N. Crystal structure of a prokaryotic (6-4) photolyase with an Fe-S cluster and a 6,7-dimethyl-8-ribityllumazine antenna chromophore.Proc. Natl. Acad. Sci. U. S. A. 2013; 110: 7217-7222Crossref PubMed Scopus (80) Google Scholar), which was used in the automated molecular replacement program PHASER (34Mccoy A.J. Grosse-Kunstleve R.W. Adams P.D. Winn M.D. Storoni L.C. Read R.J. Phaser crystallographic software.J. Appl. Crystallogr. 2007; 40: 658-674Crossref PubMed Scopus (14920) Google Scholar) in PHENIX (35Adams P.D. Afonine P.V. Bunkoczi G. Chen V.B. Davis I.W. Echols N. et al.Phenix: A comprehensive Python-based system for macromolecular structure solution.Acta Crystallogr. D Biol. Crystallogr. 2010; 66: 213-221Crossref PubMed Scopus (17044) Google Scholar) software version 1.20.1-4487 (https://phenix-online.org/). Further refinements were performed with PHENIX, and the addition of water and cofactors was performed by using COOT (36Emsley P. Cowtan K. Coot: model-building tools for molecular graphics.Acta Crystallogr. D Biol. Crystallogr. 2004; 60: 2126-2132Crossref PubMed Scopus (23864) Google Scholar) software version 0.9.6 (https://www2.mrc-lmb.cam.ac.uk/personal/pemsley/coot/). The final Rwork is 22.54% and Rfree is 28.30%, completeness is 99.60% with 24.250 to 2.500 Å refinement resolution (Table 1). The structure contains no Ramachandran outliers with 98.01% of the residues in the favored regions. The structure was deposited to the RCSB PDB website with the PDB ID: 7YKN. Figures were generated at PyMOL software (DeLano Scientific) version 2.4.1 (https://pymol.org/2/). Data collection and structure determination information are summarized in Table 1.Table 1Statistics for data collection, processing, and structure refinementBeamlineTurkish light source (Turkish DeLight)Resolution range24.25-2.5 (2.589–2.5)Space groupP64 2 2Unit cell200.8200.8 77.0 90 90,120Total reflections1,935,882 (189,833)Unique reflections32,043 (3130)Multiplicity60.4 (60.6)Completeness (%)99.60 (99.97)Mean I/σ5.61 (1.70)Wilson B-factor15.97R-merge1.382 (3.765)R-meas1.394 (3.796)R-pim0.1776 (0.4855)CC 1/20.955 (0.546)CCaPearson correlation coefficient (CC).0.988 (0.84)Reflections used in refinement31,969 (3130)Reflection used for R-free1068 (105)R-work0.2254 (0.2891)R-free0.2830 (0.3620)CC (work)0.910 (0.712)CC (free)0.929 (0.542)Number of non-hydrogen atoms4778Macromolecules4120Ligands101Solvent557Protein residues507RMS (bonds)0.002RMS (angles)0.48Ramachandran favored (%)98.01Ramachandran allowed (%)1.99Ramachandran outliers (%)0.00Rotamer outliers (%)1.84Clashscore5.92Average B-factor22.46macromolecules22.31Ligands17.61Solvent24.49Number of TLS groups8a Pearson correlation coefficient (CC). Open table in a new tab All data generated or analyzed during this study are included in this article, and the structure was deposited to the RCSB PDB website with the PDB ID: 7YKN. The authors declare that they have no conflicts of interest with the contents of this article. Authors would like to dedicate this manuscript to the memory of Dr Albert E. Dahlberg and Dr Nizar Turker. The authors gratefully acknowledge use of the services and Turkish Light Source (Turkish DeLight) X-ray facility at the Sağlık Bilimleri University Deneysel Tıp Araştırma ve Uygulama Merkezi. The authors also gratefully acknowledge use of the services and facilities at the Koç University Isbank Research Centre for Infectious Diseases. This publication has been produced benefiting from the 2232 International Fellowship for Outstanding Researchers Program of the TÜBİTAK (Project No. 118C270). However, the entire responsibility of the publication belongs to the authors of the publication. The financial support received from TÜBİTAK does not mean that the content of the publication is approved in a scientific sense by TÜBİTAK. I. H. K. and H. D. conceptualization; I. H. K. and H. D. methodology; H. D. supervision; H. D. resources; B. C. and H. D. investigation; B. C. and H. D. validation; B. C., I. H. K., and H. D. formal analysis; B. C. visualization; H. D. funding acquisition; I. H. K. and H. D. project administration; B. C., I. H. K., and H. D. writing – original draft; B. C. and H. D. writing – review and editing. H. D. acknowledges support from NSF Science and Technology Center grant NSF-1231306 (Biology with X-ray Lasers, BioXFEL). I. H. K. would like to thank the Istanbul Development Agency grant (ISTKA-TR/14/EVK/0039) for their financial support.