A crystallization and preliminary X-ray diffraction study of the Arabidopsis thaliana proliferating cell nuclear antigen (PCNA2) alone and in a complex with a PIP-box peptide from Flap endonuclease 1*

DNA replication is an important event for all living organ- isms and the mechanism is essentially conserved from ar-chaea, bacteria to eukaryotes. Proliferating cell nuclear an- tigen (PCNA) acts as the universal platform for many DNA transacting proteins. Flap endonuclease 1 (FEN1) is one such enzyme whose activity is largely affected by the interaction with PCNA. To elucidate the key interactions be- tween plant PCNA and FEN1 and possible structural change of PCNA caused by binding of FEN1 at the atomic level, crystallization and preliminary studies of X-ray diffraction of crystals of Arabidopsis thaliana PCNA2 ( At PCNA2) alone and in a complex with a peptide derived from At FEN1, which contains a typical PCNA-interacting protein (PIP)-box mo- tif, were performed. Both peptide-free and peptide-bound At PCNA2s were crystallized using the same reservoir so- lution but in different crystal systems, indicating that the peptide affected the intermolecular interactions in the crys- tals. Crystals of At PCNA2 belonged to the hexagonal space group P 6 3 , while those of the peptide-bound At PCNA2 belonged to the rhombohedral space group H3 , both of which could contain the functional homo-trimers.


INTRODUCTION
Genomic DNA replication requires over 50 kinds of protein which work in a highly coordinated manner by forming a large assembly called a replisome (Masai et al., 2010). The mechanism of DNA replication is conserved from archaea, bacteria to eukaryotes. PCNA is one of the most essential proteins involved in DNA metabolic processes such as replication, repair and recombination. PCNA recruits and regulates the DNA metabolizing proteins by tethering them onto DNA (Moldovan et al., 2007). PCNA possesses a ring structure formed by the three identical subunits suitable for encircling DNA. PCNA is topologically linked to DNA by the ATP-dependent action of clamp-loader called RFC (rep-lication factor C) (Yao & O'Donnell, 2012). PCNA binds different proteins involved in DNA metabolism by recognizing the canonical motif called the PCNA-interacting protein (PIP)-box, which is usually located at the C-terminus of the partner protein (Warbrick, 1998). Although the conservation of the PIP-box motif sequence is moderate, the binding mode of PIP-box peptides to PCNA is similar, as shown by many crystallographic studies (Gulbis et al., 1996;Matsumiya et al., 2002;Bubeck et al., 2011). Recently, other PCNA-binding motifs have been identified and characterized by structural biology, which showed the unique binding mode of these non-canonical motifs to PCNAs (Hishiki et al., 2009;Sebesta et al., 2017).
Previously, we studied the functional and structural aspects of Arabidopsis thaliana PCNA proteins. They are designated as AtPCNA1 (molecular mass M r : 30 508 Da) and AtPCNA2 (M r : 30 335 Da), which are in 96% identical to each other, and show significant homology to human and archaeal PCNAs (Strzalka et al., 2009). We determined the atomic structure of the two Arabidopsis PCNAs complexed with the canonical PIP-box peptide derived from human cyclin-dependent kinase inhibitor, p21/WAF-1, using X-ray crystallography. The structure of the plant PCNAs showed the basic conservation in comparison to human and archaeal PCNA, and a possible functional binding mode of the PIP-box. Flap endonuclease 1 (FEN1) is a structure-specific nuclease which removes the 5' protruded flap nucleotide generated during DNA replication and repair. In particular, the removal of RNA primers from the nascent strands in DNA replication by the action of FEN1 is essential for the maturation of Okazaki fragments. Human FEN1 activity is stimulated by the interaction with PCNA via the PIP-box peptide (Chapados et al., 2004;Bruning & Shamoo, 2004). The crystal structures of human and archaeal FEN1s in various states such as FEN1 alone (Hwang et al., 1998;Hosfield et al., 1998), in complexes with DNA (Chapados et al., 2004;Tsutakawa et al., 2011) or with PCNA (Sakurai et al., 2005;Doré et al., 2006) were investigated. However, corresponding studies related to the plant FEN1 were not reported yet.
To explore the structural aspects of plant PCNA and FEN1 interactions, including possible structural change of PCNA associated with binding of FEN1, crystallographic studies of the Arabidopsis PCNA2-FEN1 PIP-box peptide complex as well as the isolated form of AtPCNA2 were performed. Here, we describe the purification, crystallization and study of the preliminary X-ray diffraction of AtPCNA2 alone and its complex with a PIP-box peptide derived from the Arabidopsis FEN1. In particular, we for the first time succeeded in the crystallization of the peptide-free AtPC- NA2, which allows revealing the possible structural change of AtPCNA2 caused by the binding of DNA metabolizing proteins such as AtFEN1.

MATERIALS AND METHODS
Overexpression and Purification. The Escherichia coli BL21 (DE3) Rosetta strain (Novagen) was used for AtPC-NA2 protein production. Cells transformed with pET15b vector coding for AtPCNA2 (Strzalka et al., 2009) fused with histidine tag were grown at 310 K using LB medium supplemented with 30 mg l -1 chloramphenicol and 100 mg l -1 ampicillin. Protein overexpression was induced using 1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) when OD 600 of the culture reached ~ 0.5. Bacteria were cultured by shaking vigorously at 310 K for 4 h then followed by centrifugation (10 min, 12 000 × g, 277 K). The pellet from 1 l of culture was resuspended in 40 ml of binding buffer A (50 mM sodium phosphate, 300 mM NaCl, 20 mM imi-dazole, pH 7.5) which contained an EDTA-free protease inhibitor cocktail (Roche). Bacteria were sonicated for 15 min (5 s pulse followed by a 15 s pause, QSonica Q700 sonicator, USA). The cell lysate containing the AtPCNA2 protein was supplemented with 200 units of DNase I (Takara Bio, Japan) and incubated for 20 min on ice. Then, the lysate was centrifuged (20 min, 21 000 × g, 277 K) and the supernatant was loaded onto HisTrap FF Ni Sepharose (5 ml, GE Healthcare, Sweden). The protein was purified according to the protocols supplied by the manufacturers. After being eluted from the nickel resin, the AtPCNA2 was dialyzed into buffer B (50 mM Tris-HCl pH 7.6) and loaded onto the anion exchange HiTrap Q column (5 ml, GE Healthcare, Sweden). The protein elution was performed using a linear NaCl gradient from 0 to 1 M (10 volumes of the column). The protein was eluted when NaCl concentration was around 0.5 M NaCl. The purified AtPCNA2 was dialyzed against buffer C (50 mM Tris-HCl, 200 mM NaCl, 0.2 mM EDTA, pH 8.0), supplemented with glycerol (final Values for the outer shell are given in parentheses. concentration 15% (v/v)) and stored at 193 K for further experiments.
Crystallization. The protein was crystallized using the hanging-drop vapour diffusion method at 293 K. According to the previous crystallization experiment of At-PCNAs complexed with human p21 peptide (Strzalka et al., 2009), a reservoir containing 100 mM citric acid pH 4.5 and 1.2 to 1.8 M ammonium sulfate was used for both AtPCNA2 alone and the complex with the peptide. AtPCNA2 used for crystallization was concentrated to 20 mg ml -1 in buffer C deprived of glycerol. In the case of the complex crystallization, the peptide was added to the protein at a 1.5 molar ratio (assuming AtPCNA2 as a monomer) and incubated at 289 K for 30 min. Crystallization drops were produced by mixing equal volumes (1 µl) of the protein and reservoir solutions and then placed against the reservoir solution (500 µl). Crystals of the peptide-free and peptide-bound AtPCNA2s grew to a maximum dimension of approximately 200 µm within a week.
X-ray diffraction data collection. X-ray diffraction data were collected using an in-house imaging plate detector R-AXIS VII (Rigaku) equipped with an X-ray generator either MicroMax-007 HF or FR-E (Rigaku). Crystals were cryo-harvested by soaking them into the mother liquor supplemented with either 30% (v/v) ethylene glycol for AtPCNA2 crystals or 20% (v/v) glycerol for the AtPCNA2-AtFEN1 PIP-box peptide complex for several seconds, and then flash-cooled in liquid nitrogen. The detailed X-ray diffraction experiment conditions and the data statistics are listed in Table 1. The data were processed by the XDS (Kabsch, 2010), Pointless (Evans, 2006), and Aimless (Evans & Murshudov, 2013) programs. The number of molecules in the asymmetric units was estimated using the Matthews coefficients (Matthews, 1968).

Crystallization
We previously succeeded in crystallizing the human p21-derived peptide-bound plant PCNAs which were purified using HisTrap as the only column chromatography method (Strzalka et al., 2009). In this study, using two-step column chromatography we purified AtPC-NA2 (Fig. 1). The genomic DNA from the E. coli cells was degraded into small fragments during sonication and DNase I treatment. Next, most of these fragments were removed at the first stage of the protein purification when the HisTrap column was used. The DNA remaining in the protein sample eluted from the HisTrap column was separated from AtPCNA2 using the HiTrap Q column (data not shown). Using a protein sample of higher purity than previously (Strzalka et al., 2009) we were able to crystallize AtPCNA2 free of the ligand as well as complexed with the PIP-box peptide derived from the AtFEN1. Interestingly, both the peptide-free and peptide-bound AtPCNA2 crystallized under the same reservoir conditions but in different crystal forms (Fig. 2), as revealed by the following crystallographic analysis. The peptide used for the present study affected the packing interactions of AtPCNA2 in the crystals.

X-ray crystallographic study
The AtPCNA2 crystals diffracted the X-rays from an in-house generator to a resolution of 2.65 Å, and the AtPCNA2-AtFEN1 PIP-box peptide complex to 2.85 Å ( Fig. 3 and Table 1). The AtPCNA2 crystals alone belong to the hexagonal space group P6 3 , with unit cell constants of a=b=93.95 Å, c=63.54 Å, and probably contain the one subunit in the asymmetric unit with a Matthews coefficient (V M ) of 2.68 Å 3 Da -1 and a solvent content of 54%. These crystallographic parameters are similar to those for homologous PCNAs (Matsumiya et al., 2001;Ladner et al., 2011) which suggests that in the AtPCNA2 crystals monomers form homotrimers with crystallographic symmetry-related molecules. On the other hand, the AtPCNA2-AtFEN1 PIP-box peptide complex crystals belong to the space group H3 with unit cell constants of a=b=224.17 Å, c=199.73 Å. The space group and unit cell constants of the crystal of this complex are similar to those of the previous AtPCNA2human p21 peptide complex crystal which contains eight complex molecules in the asymmetric unit with two monomers that form the symmetry-related homotrimers and the two independent homotrimers. If the crystal of the AtPCNA2/AtFEN1 PIP-box peptide complex has a molecular arrangement in the crystal very similar to that Lanes: M, molecular mass standards with sizes indicated on the left; 1, E. coli supernatant after sonication containing AtPCNA2; 2, AtPCNA2 after purification using HisTrap Ni-chelating affinity chromatography; 3, AtPCNA2 purified using HisTrap Ni-chelating affinity chromatography followed by HiTrap Q anion-exchange chromatography. of the AtPCNA2-human p21 peptide complex, it leads to a V M of 3.66 Å 3 Da -1 and a solvent content of 66%. Despite the fact that both the peptide-free and peptidebound AtPCNA2 proteins were crystallized using the same reservoir solution, two different forms of crystals were obtained, one type for AtPCNA2 and the other for peptide-bound AtPCNA2. Determining the structure of these complexes at the atomic level would give us an answer whether the peptide-bound AtPCNA2 complex displays unique and different inter-homotrimer interactions in the crystal, as compared to those in the ligand-free AtPCNA2 crystal. This information would be significant in the context of a detailed understanding of the role of PCNA in FEN1-dependent DNA replication in plants. Although we solved both crystal structures by the molecular replacement method, the quality of the electron density maps was too poor to build accurate atomic models. Therefore, the preparation of the better quality crystals is in progress to obtain high-quality X-ray diffraction data at synchrotron radiation facility beamlines.

Acknowledgement
The Faculty of Biochemistry, Biophysics and Biotechnology is a partner of the Leading National Research Center (KNOW) supported by the Ministry of Science and Higher Education. We thank Drs. Tsuyoshi Inoue and Yohta Fukuda for giving us an opportunity to use the X-ray diffractometer in their laboratory, and Drs. Genji Kurisu, Hideaki Tanaka and Tetsuko Nakaniwa for their kind support and encouragement during this research.