Reverse mutants of the catalytic 19 kDa mutant protein (nanoKAZ/nanoLuc) from Oplophorus luciferase with coelenterazine as preferred substrate

Native Oplophorus luciferase (OpLase) and its catalytic 19 kDa protein (wild KAZ) show highest luminescence activity with coelenterazine (CTZ) among CTZ analogs. Mutated wild KAZ with 16 amino acid substitutions (nanoKAZ/nanoLuc) utilizes bis-coelenterazine (bis-CTZ) as the preferred substrate and exhibits over 10-fold higher maximum intensity than CTZ. To understand the substrate selectivity of nanoKAZ between CTZ and bis-CTZ, we prepared the reverse mutants of nanoKAZ by amino acid replacements with the original amino acid residue of wild KAZ. The reverse mutant with L18Q and V27L substitutions (QL-nanoKAZ) exhibited 2.6-fold higher maximum intensity with CTZ than that of nanoKAZ with bis-CTZ. The catalytic properties of QL-nanoKAZ including substrate specificity, luminescence spectrum, luminescence kinetics, luminescence products of CTZ, and luminescence inhibition by deaza-CTZ analogs were characterized and were compared with other CTZ-utilizing luciferases such as Gaussia and Renilla luciferases. Thus, QL-nanoKAZ with CTZ could be used as a potential reporter protein for various luminescence assay systems. Furthermore, the crystal structure of QL-nanoKAZ was determined at 1.70 Å resolution. The reverse mutation at the L18Q and V27L positions of α2-helix in nanoKAZ led to changes in the local structures of the α4-helix and the β6- and β7-sheets, and might enhance its binding affinity and oxidation efficiency with CTZ to emit light.


Introduction
Oplophorus luciferase (OpLase) is a secretory protein isolated from the deep-sea shrimp, Oplophorus gracilirostris, and catalyzes the oxidation of coelenterazine (CTZ, a luciferin) to emit

Preparation of mutated genes for nanoKAZ
The chimeric genes for wild KAZ (wK), nanoKAZ (nK), and their reverse mutants were prepared as follows: i) The chimeric genes for wK/nK chimera and nK/wK chimera were prepared by replacing the EcoRI-SalI fragment (amino acid residues 1-82) or SalI-XbaI fragment (amino acid residues 83-169) of wild KAZ with the corresponding fragments of nanoKAZ, respectively (Fig 2A).
ii) The reverse mutant genes of wK/nK chimera (Fig 2B) and nanoKAZ (Fig 2C-2E) were prepared by site-directed mutagenesis using the overlap extension polymerase chain reaction (PCR) procedures [22] with the specific primers (S1 Table). The nucleotide sequences of mutated genes were confirmed by DNA sequencing (Eurofins Genomics).

Expression of mutated proteins in bacterial cells
The cold-inducible expression vector of pCold-ZZ-P-X under the control of the cold shock protein A promoter and the lac operator in E. coli cells [23,24] was used to express as a fused protein of the ZZ domain, which is the synthetic IgG-binding domain of staphylococcal protein A and serves as a soluble partner of target proteins [23,24]. The vector consists of a histidine tag sequence for Ni-chelate affinity chromatography, the cleavage sequence of human rhinovirus 3C protease between the ZZ domain and a target protein, followed by the multiple cloning sites [24]. The EcoRI-XbaI fragment of the reverse mutated gene was inserted into the EcoRI-XbaI site of a pCold-ZZ-P-X vector to give the corresponding expression vector.
To determine luminescence activity of reverse mutants, the seed culture of E. coli strain BL21 possessing each expression vector was grown in 5 mL of Luria-Bertani broth (LB broth) containing ampicillin (50 μg/mL) at 37˚C for 18 h using a TAITEC model BR-3000LF shaker (Tokyo, Japan) with reciprocal shaking (130 rpm). After transferring 0.1 mL of the seed culture to 10 mL of LB broth, the bacterial cells were cultured at 37˚C for 3 h and then cooled on an ice bath over 30 min. To induce protein expression, 1 M IPTG was added to the culture medium at a final concentration of 1 mM and the bacterial cells were cultured at 15˚C for 20 h. After collecting 1 mL of cultured cells using a TOMY model MCX-150 micro-centrifuge (Tokyo, Japan) at 8,000 rpm for 30 s, cells were suspended in 0.5 mL of 50 mM Tris-HCl (pH 7.6) and disrupted by sonication using a Branson model 250 sonifier (Danbury, CT) for 5 s in an ice-water bath. The soluble fraction of cell extracts obtained by centrifugation at 12,000 × g for 3 min was used for determining the luminescence activity and 5 μL was subjected to SDS-PAGE analysis (Fig 3).

Expression and purification of QL-nanoKAZ and SNH-nanoKAZ
For purification of QL-nanoKAZ, the EcoRI-XbaI fragment of 18Q27L-nanoKAZ gene was replaced with nanoKAZ gene in pCold-dnKAZ [10] to give an expression vector, pCold-QL- nanoKAZ. The expressed QL-nanoKAZ consisted of 191 amino acid residues having the amino-terminal sequence of MNHKVHHHHHHMELGTLEGSEF (histidine-tag sequence underlined). The purification was carried out using Ni-chelate column chromatography [5,10]. Briefly, the seed culture (10 mL) in LB broth was transferred to 400 mL of LB broth containing 100 μL of antifoam (Disform CE475, NOF Co., Tokyo, Japan) in a 2-L Sakaguchi flask and cultured with reciprocal shaking (130 rpm) at 37˚C for 3 h. The absorbance at 660 nm measured by TAITEC model mini photo 518R was 0.97 and reached 1.22 after 18 h at 15˚C by the addition of IPTG to a final concentration of 0.2 mM. The harvested cells from 800 mL of culture medium (wet weight, 5.1 g) by using a Hitachi model CR20GIII high-speed refrigerated centrifuge (Tokyo, Japan) at 5,000 rpm for 5 min were suspended in 60 mL of 50 mM Tris-HCl (pH 7.6) and disrupted by sonication three times for 3 min in an ice-water bath. The soluble fractions obtained by centrifugation at 10,000 rpm for 20 min were applied on a Ni-chelate column (ø2.5 × 5 cm) at room temperature. After washing with 200 mL of 50 mM Tris-HCl (pH 7.6), the fractions of QL-nanoKAZ containing 154.2 mg proteins were obtained by eluting with 0.1 M imidazole (Wako Pure Chemicals) in 50 mM Tris-HCl (pH 7.6). For further purification, a portion of the eluted QL-nanoKAZ fractions (98.8 mg protein) from the 1st Ni-chelate column was diluted to 200 mL with 50 mM Tris-HCl (pH 7.6) containing 2 M NaCl and was applied on the 2nd Ni-chelate column (ø1.5 × 6 cm) and the proteins were eluted with 0.1 M imidazole. The highly purified QL-nanoKAZ (91.8 mg protein) was obtained (Fig 3F and S2 Table) and used as a protein source for crystal structure analysis.
For purification of SNH-nanoKAZ, the EcoRI-XbaI fragment of for SNH-nanoKAZ gene was chemically synthesized (Eurofine Genomics) and was replaced with the EcoRI-XbaI fragment of nanoKAZ gene in pCold-dnKAZ [10] to give an expression vector, pCold-SNH-nano-KAZ. The purification procedures were essentially the same as that of QL-nanoKAZ, and the yield was 135.6 mg protein from 800 mL of cultured cells (Fig 3F).

Expression of luciferase genes in mammalian cells
For the secretory expression of wild KAZ mutants in mammalian cells, a pcDNA3-GLsp vector that is a derivative of pcDNA3 (Invitrogen, Carlsbad, CA) having the signal peptide sequence of GLase for secretion was used, as previously described [5]. The expression vectors were constructed as follows. The fragment of the mutated KAZ gene obtained by PCR procedures was inserted into the HindIII-XbaI sites of pcDNA3 or the EcoRI/XbaI site of pcDNA3-GLsp [5] to give pcDNA3-dnKAZ, pcDNA3-AQQL-nK (this work), and pcDNA3-QL-nK (this work) for expression in the cytoplasm, or pcDNA3-GLsp-dnKAZ [6], pcDNA3-GLsp-QL-nK (this work), pcDNA3-GLsp-SNH-nK (this work), and pcDNA3-GLsp-EpGLuc [25,26] for secretory expression. These vectors had the identical nucleotide sequence between the promoter and the initial methionine codon in the same vector [25,26]. For transient expression in Chinese hamster ovary-K1 (CHO-K1) cells, CHO-K1 cells (1 × 10 5 cells) cultured in 2 mL of Ham's F-12 medium (Wako Pure Chemicals) containing 10% (v/v) heat-inactivated fetal calf serum (Biowest, France) without antibiotics were seeded in a 6-well plate (n = 4) at 37˚C in a humidified atmosphere of 5% CO 2 for 24 h. For transfection, the mixture of the expression vector (1 μg) and FuGENE HD (3 μL, Promega) in 100 μL of serum-free Ham's F-12 medium was added to the cells. After further incubation for 24 h, the culture medium was recovered and the cells were washed three times with 2 mL of PBS (D-PBS (-), Wako Pure Chemicals), and then lysed in 400 μL of lysis buffer (2 mM EDTA, 1% Triton-X 100 (Sigma) and 10% glycerol (Wako Pure Chemicals) in PBS), for 15 min at room temperature (24-26˚C). The culture medium and cell extracts (2 μL) were used for luminescent assays (n = 2) [25,26].

Assay for luminescence activity
The luminescence activity was determined using an ATTO (Tokyo, Japan) AB2200 luminometer (Ver.2.07, rev4.21) in the presence of a 0.23% neutral density filter (determined by using a laser at 544 nm) or an ATTO AB2270 luminometer with an F2-cut filter (HOYA, R62).
Under these measurement conditions, the luminescence intensity of AB2270 showed 1.8-fold higher than that of AB2200 using aequorin as a light standard [27].
The reaction mixture (100 μL) contained CTZ or CTZ analog (1 μg/μL dissolved in ethanol) in 30 mM Tris-HCl (pH 7.6)-10 mM EDTA, and the luminescence reaction was initiated by the addition of 1-5 μL of luciferase solutions. The luminescence intensity was recorded in 0.1 s-intervals for 10-60 s using an AB2200 or AB 2270 luminometer. The maximum intensity of luminescence (I max ) and the integrated luminescence intensity (Int.) were shown as relative light units (rlu).

Bioluminescence spectral analysis
Bioluminescence spectra were measured with a fluorescence spectrophotometer FP-6500 (Jasco, Tokyo, Japan) at 24˚C with the excitation light source turned off (emission bandwidth, 20 nm; sensitivity, medium; response, 0.5 s; scan speed, 1000 nm/min). The corrected bioluminescence spectra were obtained by the manufacturer's protocol. The reaction mixture (1 mL) in a quartz cuvette contained 5 μg of CTZ or CTZ analog (dissolved in 5 μL of ethanol) in 30 mM Tris-HCl (pH 7.6)-10 mM EDTA, and the luminescence reaction was initiated by adding 5 μL of the purified luciferase (1.0 μg protein) to the reaction mixture.

Protein analysis
SDS-PAGE analysis was carried out under reducing conditions using a 12% separation gel (TEFCO, Tokyo, Japan) and the gel was stained with a colloidal CBB staining kit (TEFCO). The protein concentrations of nanoKAZ, QL-nanoKAZ, SNH-nanoKAZ, GLase, and aequorin were determined by amino acid composition analysis, as previously described [27], and other luciferases were determined by the dye-binding method using a commercially available kit (Bio-Rad, Richmond, CA, USA) and bovine serum albumin as a standard (Pierce, Rockford, IL, USA).

Crystallization, data collection, and structure determination
The eluted fractions of QL-nanoKAZ (50 mg protein) from the Ni-chelate column were diluted twice with the buffer (20 mM Tris-HCl (pH 8.5) and 2 mM DTT) and then loaded on a HiTarp Q HP column (Cytiva, MA, USA). The fractions containing QL-nanoKAZ were eluted at 200 mM NaCl by a linear gradient from 50 mM to 400 mM of NaCl and concentrated by an Amicon Ultra centrifugal filter unit (MWCO 10,000, Merck-Millipore, Billerica, MA, USA), and separated by a HiLoad 16/60 Superdex 75 column (Cytiva) equilibrated with 20 mM Tris-HCl (pH 8.0) containing 150 mM NaCl and 2 mM DTT. The fractions of QL-nanoKAZ were concentrated to a concentration of 20 mg/mL by a centrifugal filter unit and stored at -80˚C.
The crystallization was performed by the method of sitting drop vapor diffusion. The crystals were grown in a mixture of 0.2 μL of QL-nanoKAZ and 0.2 μL of the precipitant solution (100 mM MES (pH 6.5) and 1.9 M MgSO 4 ) at 20˚C. After a few days of incubation for equilibration against the precipitant solution, the crystals with dimensions of 250 μm × 250 μm × 20 μm were obtained (S1 Fig). The crystals were cryoprotected in the reservoir solution supplemented with 25% (v/v) glycerol before flash-cooling in liquid nitrogen. An X-ray diffraction dataset was collected to 1.70 Å at a wavelength of 1.0 Å on beamline BL26B2 at SPring-8 [29]. The diffraction data were processed using the XDS programs [30] and the structure was solved by molecular replacement using the Phaser program [31] from the PHENIX programs [32], with the nanoKAZ coordinates (PDB ID: 5B0U) as the search model. The structural model was built into the electron density map using COOT [33] and refined using the PHENIX program. All structure figures were prepared using PyMOL (http:// pymol.sourceforge.net/).

Soluble expression of nanoKAZ mutants in bacterial cells
Previous studies have reported that wild KAZ and its mutated proteins were mainly expressed as inclusion bodies in E. coli cells [3,4]. To express soluble KAZ mutants in E. coli cells [5][6][7], the mutated proteins were expressed as the fusion protein with the ZZ domain of staphylococcal protein A using the cold-induced expression vector, pCold-ZZ-P-X [24]. The chimeric genes between wild KAZ and nanoKAZ (Fig 2A) and the reverse mutated genes of nanoKAZ were prepared using PCR procedures (Fig 2B-2E), and the mutated proteins fused to the ZZ domain were all successfully expressed as soluble form in E. coli cells.
The soluble fraction obtained by centrifugation was analyzed by SDS-PAGE (Fig 3A-3E), suggesting that the expression levels of each protein were not significantly different among them under the same culture conditions. Thus, the soluble fractions were used as luciferase sources for the luminescence assay without further purification. The luminescence activities of I max and Int. 60 s were used as tentative indicators for substrate specificity.
The reverse mutant of nK-18Q27L (hereafter referred to as "QL-nanoKAZ") showed the highest luminescence activity with CTZ. The reverse mutations at L18Q and V27L synergistically stimulated luminescence with CTZ. In addition, the reverse mutants of nK-27L, nK-11Q27L, and nK-18Q27L displayed luminescence with 6h-CTZ, similar to nK-AQQL and wK/nK- 124Q (Table 2E), though the reason for the high luminescence activity with 6h-CTZ was unclear. A schematic representation of the relationships between the reverse mutants and their luminescence activities with CTZ and bis-CTZ are summarized in Fig 2, based on the results shown in Tables 1 and 2.

Expression and purification of nanoKAZ, QL-nanoKAZ, and SNH-nanoKAZ from bacterial cells
To compare the luminescence properties of QL-nanoKAZ with those of nanoKAZ, SNH-nanoKAZ, and native OpLase, we expressed nanoKAZ, QL-nanoKAZ, and SNH-nanoKAZ as soluble proteins without the ZZ domain and highly purified them using Ni-chelate column chromatography (Fig 3F). It has been reported that SNH-nanoKAZ (teLuc) is a mutant of nanoKAZ with three amino acid substitutions (D19S, D85N, and C169H) and showed approximately 2-fold higher I max value than FMZ using diphenylterazine (DTZ) (Fig 1B) with a redshifted emission peak at 502 nm [15]. The protein concentrations of the purified luciferases were determined by amino acid composition analysis [27] and used for normalization of the luminescence activities and estimation of the luciferase concentration expressed in the mammalian cells.
The luminescence kinetics and emission spectrum of QL-nanoKAZ were determined using CTZ and its four analogs. QL-nanoKAZ with CTZ showed the highest luminescence activity with slow decay kinetics (Fig 4A). The emission peaks of QL-nanoKAZ with CTZ and its analogs were around 458 nm, similar to native OpLase [4] and nanoKAZ [5] (Fig 4B). The linearity of luminescence activity for QL-nanoKAZ under various concentrations of CTZ was determined (Fig 4C) and the relative values of I max and Int. 30 s were normalized using recombinant aequorin as a light standard ( Table 3). Among the CTZ-utilizing luciferases, GLase is known to have the highest I max value with fast decay kinetics [21,27]. The I max value of QL-nanoKAZ with CTZ showed a similar order to that of GLase with more than 80-fold higher luminescence activity of nanoKAZ, and the luminescence kinetics was the slow decay pattern with a 2.6-fold higher value of Int. 30 s than that of GLase (Table 3). Thus, QL-nanoKAZ is a potential candidate for the reporter protein in the glow luminescence assay system.

Expression of QL-nanoKAZ in mammalian cells
OpLase is a secretory protein, but the catalytic component of wild KAZ could not be secreted from mammalian cells using its own signal peptide sequence [4,8] or other signal peptide sequences such that from GLase [8]. However, the signal peptide sequence for GLase (GLsp) has been successfully used for the secretion of nanoKAZ in CHO-K1 cells [5]. To evaluate QL-nanoKAZ as a reporter protein in mammalian cells, it was transiently expressed in CHO-K1 cells. The amount of QL-nanoKAZ secreted into the culture medium was similar to nanoKAZ and GLase ( Table 4).
As previously reported, nanoKAZ could be secreted into the culture medium even without a signal peptide sequence by the unknown mechanism [8]. Approximately 10% of nanoKAZ and QL-nanoKAZ without the signal peptide sequence was secreted into the culture medium [8], contrary to the report for nanoLuc by Hall et al. [9] (Table 5). Thus, QL-nanoKAZ with CTZ can be used as a reporter protein in the cytoplasm, similar to nanoKAZ/nanoLuc with bis-CTZ and h-CTZ [5]. However, based on the slow decay luminescence pattern of QL-nano-KAZ (Fig 4A), it was not suitable for real-time imaging of protein secretion from mammalian cells like GLase (unpublished results) [34,35].

Luminescent products of CTZ catalyzed by CTZ-utilizing luciferases
In the oxidation process of CTZ with O 2 by luciferase, 2-peroxycoelenterazine (CTZ-OOH) might be an intermediate, from which CTMD is produced with light emission (Fig 1A). To A. Luminescence kinetics of QL-nanoKAZ with CTZ and its analogs as substrates. B. Normalized luminescence spectra of QL-nanoKAZ with CTZ and its analogs, based on the luminescence intensity of QL-nanoKAZ with CTZ. C. Linearity of luminescence intensity (I max ) of QL-nanoKAZ with CTZ, in comparison with nanoKAZ, SNH-nanoKAZ, GLase, and aequorin at the protein concentrations of 0.3 pg to 3 ng (n = 6). Solid and dashed lines represent blank + 3 SD for aequorin and the CTZ-utilizing luciferases, respectively. https://doi.org/10.1371/journal.pone.0272992.g004 identify the luminescent products of CTZ with the CTZ-utilizing luciferases, CTZ was incubated with each luciferase for 2 h and the products were analyzed by HPLC (Fig 5). The major and minor products from CTZ catalyzed by luciferases were CTMD and CTM, respectively. This was similar to the result obtained from CTZ-OOH in the Ca 2+ -triggered luminescence reaction of aequorin [36]. Recently, we investigated the degradation products from (S)-2-peroxycoelenterazine ((S)-CTZ-OOH) in aequorin under protein denaturing conditions and the major product was identified as CTM, but not CTMD, without light emission. In addition, we confirmed that CTMD was not hydrolyzed to CTM in aqueous solutions [36]. Thus, CTZ-OOH might be an intermediate in the luminescence reaction of CTZ catalyzed by these CTZ-utilizing luciferases. The efficiency of CTMD formation might be correlated with the luminescence activity of luciferases. Interestingly, even though CTZ is not a suitable substrate in nanoKAZ [5,9], CTZ was completely consumed and converted to CTMD and CTM with 63% and 37% in 2 h, respectively ( Table 6). As we recently reported, the products after incubation of CTZ (473 pmol) in 50 mM Tris-HCl (pH 7.6) at 25˚C for 2 h in the absence of luciferase were CTZ (149 pmol), CTMD (68 pmol), CTM (65 pmol), dCTZ (90 pmol) and an unknown product ( Table 6) [19]. Because the formation of dCTZ was not detected in the reaction mixtures of CTZ with nanoKAZ and other luciferases, the CTM and CTMD formation from CTZ by

PLOS ONE
nanoKAZ could be an enzymatic oxidation process. The low luminescence activity of nano-KAZ might be explained by that CTZ-OOH could not properly stabilized in the nanoKAZ molecule during the luminescence reaction and partially decomposes into CTM and CTMD without light emission [36].

Substrate specificities of nanoKAZ, QL-nanoKAZ, SNH-nanoKAZ, and native OpLase for CTZ analogs
As previously reported, native OpLase showed broad substrate specificity [11], and the catalytic component of 19 kDa protein (wild KAZ) in OpLase also showed similar substrate specificity [4,13]. As summarized in Table 7, nanoKAZ and SNH-nanoKAZ showed lower luminescence activity with CTZ, but other profiles of substrate specificity were similar to those of OpLase and wild KAZ. By contrast, QL-nanoKAZ could use CTZ and h-CTZ as the preferred substrates, similar to OpLase [11] and wild KAZ [13]. The I max value of QL-nanoKAZ with CTZ was 2.6-and 5.3-folds higher than that of nanoKAZ with bis-CTZ and FMZ, respectively (Table 7). Furthermore, QL-nanoKAZ with all C2-modified CTZ analogs possessing the hydroxy moiety (-OH) at the C6 benzyl group retained higher luminescence activity than those of nanoKAZ and SNH-nanoKAZ. Among these mutants, similar luminescence activities were observed for n-CTZ and hcp-CTZ, suggesting that the C2-and C8-groups of CTZ might not be essential for binding to these luciferases ( Table 7). In QL-nanoKAZ, stabilization of the hydroxy moiety at the C6-benzyl group of CTZ with specific amino acid residue(s) in the protein might be crucial to stimulate the luminescence activity with CTZ. In addition, the amino acid residues that interact with the C3-carbonyl group of the imidazopyrazinone structure also might be required to stabilize CTZ analogs lacking the hydroxy moiety at the C6-benzyl group of CTZ (i.e., bis-CTZ, 6h-CTZ, and FMZ).

Effect of deaza-analogs for CTZ and CTZ-OOH on luciferase activity
The chiral deaza-analogs of (S)-and (R)-deaza-CTZ ((S/R)-daCTZ) for CTZ and (S)-2-and (R)-2-hydroxymethyl-deaza-CTZ ((S/R)-HM-daCTZ) for CTZ-OOH (Fig 1C) have been used as inhibitors to investigate the oxidation process of CTZ with O 2 in RLase [12] and the regeneration process of the calcium-binding photoprotein, aequorin, from CTZ and O 2 [19]. The inhibitory effect on luminescence activity of various CTZ-utilizing luciferases with CTZ was examined in the presence of each deaza-analog ( Table 8). As previously reported, RLase was inhibited by (S/R)-daCTZ and by (S)-HM-daCTZ [12]. We proposed that the binding of CTZ to RLase might be a non-stereospecific process, while the following oxidation of CTZ with O 2 to CTZ-OOH might be a stereospecific process. Interestingly, the luminescence activities of GLase and RLase-547 were clearly inhibited by (S)-HM-daCTZ and (R)-HM-daCTZ, respectively, suggesting that the addition of O 2 to CTZ might be a stereospecific process. Notably, the inhibition stereospecificity of RLase-547 by (R) form of HM-daCTZ was different from that of RLase by (S) form of HM-daCTZ. This result suggested that the addition of O 2 to CTZ might be proceeded in the opposite direction between RLase and RLase-547 (Table 8). Table 7. Substrate specificities and luminescence properties for purified QL-nanoKAZ, nanoKAZ, SNH-nanoKAZ, and native Oplophorus luciferase (OpLase).
On the other hand, the luminescence activity of OpLase, nanoKAZ, QL-nanoKAZ, and SNH-nanoKAZ were not selectively inhibited by (S)-or (R)-HM-daCTZ, suggesting that the oxidation step of CTZ with O 2 by these luciferases might be a non-stereospecific process. However, it is possible that (S/R)-HM-daCTZ binds less efficiency to the catalytic pocket of luciferase due to steric hindrance by the hydroxymethyl group of daCTZ.

Crystal structure of QL-nanoKAZ and substrate specificity for CTZ
Based on the results showing inhibitory effect of deaza-CTZ analogs against QL-nanoKAZ ( Table 8), we attempted to determine the complex structure of daCTZ or HM-daCTZ with QL-nanoKAZ, but were unsuccessful. However, in the absence of deaza-CTZ analogs, the structure of QL-nanoKAZ was determined at 1.70 Å resolution (PDB ID: 7VSX). The statistical values of data collection and structure refinement are summarized in Table 9.
The overall structure of QL-nanoKAZ substituted at L18Q and V27L was almost the same as that of nanoKAZ (PDB ID: 5B0U) (Fig 6A). But some structural differences were observed at the α3-helix (29-37 aa), β6-sheet (105-109 aa), and β7-sheet (112-119 aa) (Fig 7). Interestingly, the amino acid residue at Tyr 109 in the β6-sheet showed significant change in position, which might be hydrogen bond formation between the hydroxy group of Tyr 109 and the C6-hydroxy group of CTZ, increasing the affinity with CTZ (Fig 6B).
Recently, we determined the structures of a complex of apoAequorin (apoprotein of aequorin) with (S)-daCTZ (PDB ID: 7EG2) or (S)-HM-daCTZ (PDB ID: 7EG3). The stabilization mechanisms of daCTZ and HM-daCTZ in apoAequorin were identical to that of the complex of apoAequorin and (S)-CTZ-OOH (aequorin: PDB ID: 1EJ3) [19]. Thus, the C6-hydroxy and the C3-carbonyl groups in both CTZ and CTZ-OOH were stabilized by the hydrogenbonding interactions via three amino acid residues (His 16, Tyr 82, and Trp 86) and His 169, respectively. On the other hand, the Ca 2+ -triggered Renilla luciferin-binding protein (RLBP) is   a complex of apoRLBP (apoprotein of RLBP) and CTZ [37,38], and apoRLBP can also bind h-CTZ and bis-CTZ [38]. The crystal structure of RLBP from R. muelleri containing CTZ has been determined (PDB ID: 2HPS) [39] and CTZ was stabilized with hydrogen-bonding interactions at the C2-hydroxy group with both Tyr 36 and Arg 19, the C6-hydroxy group with Asp 183 and Lys 139 though two H 2 O molecules, the C3-carbonyl group with Arg 22, and the N(7) nitrogen with Phe 190. As apoRLBP could stably bind to both h-CTZ and bis-CTZ to form semi-synthetic RLBP [38], the C3-carbonyl group and the N(7) nitrogen in CTZ might be responsible for the stabilization with apoRLBP without interacting the C2-and C6-hydroxy groups of CTZ. Thus, there are variations in the stabilization pattern of CTZ by the amino acid residues in each protein molecule. In QL-nanoKAZ, CTZ was the preferred substrate over bis-CTZ and FMZ, which lack the C2-and C6-hydroxy groups of CTZ. When the C2-CTZ analogs were used as a substrate, the decrease in luminescence activity of QL-nanoKAZ was not significant compared to CTZ ( Table 6). Thus, the C2-hydroxy group of CTZ might not be essential for binding QL-nano-KAZ, similar to nanoKAZ and SNH-nanoKAZ. Furthermore, hcp-CTZ with a C8 substitution in CTZ could be efficiently used as a substrate by QL-nanoKAZ, nanoKAZ, and SNH-nano-KAZ, indicating that the C8-phenyl group of CTZ was also not required for molecular recognition of CTZ by QL-nanoKAZ. From these results, the high luminescence activity of QL- Table 9. Statistics of data collection and structure refinement. nanoKAZ with CTZ might be explained by the high affinity between the C6-hydroxy group of CTZ and QL-nanoKAZ. The reverse mutations at V18Q and V27L in nanoKAZ caused local changes in its structure, resulting in stabilization of CTZ for efficient oxidation with O 2 .

Data collection and processing
Although the binding cavity in nanoKAZ for CTZ was not determined, Tyr 109 seems to be the potential amino acid residue that stabilizes the C6-hydroxy group of CTZ with hydrogenbonding interactions, resulting in efficient oxidation of CTZ.

Conclusions
To investigate the differences in substrate specificities between wild KAZ and nanoKAZ toward CTZ and bis-CTZ, the reverse nanoKAZ mutants substituted with the identical amino acid residue of wild KAZ at the same position were prepared. Among these mutants, a reverse mutant substituted with L18Q and V27L (QL-nanoKAZ) showed the highest luminescence activity with CTZ, and the luminescence properties of QL-nanoKAZ were compared with those of the CTZ-utilizing luciferases including nanoKAZ, Renilla luciferase, and Gaussia luciferase. The results showed that QL-nanoKAZ is an ideal candidate for the reporter protein in various luminescence assay systems. Furthermore, the crystal structure of QL-nanoKAZ