Identification of the Acidification Mechanism of the Optimal pH for RNase He1

Abstract

Ribonuclease (RNase) He1 is a small ribonuclease belonging to the RNase T1 family. Most of the RNase T1 family members are active at neutral pH, except for RNase Ms, U2, and He1, which function at an acidic pH. We crystallized and analyzed the structure of RNase He1 and elucidated how the acidic amino residues of the α1β3- (He1:26–33) and β67-loops (He1:87–95) affect their optimal pH. In He1, Ms, and U2, the hydrogen bonding network formed by the acidic amino acids in the β67-loop suggested that the differences in the acidification mechanism of the optimum pH specified the function of these RNases. We found that the amino acid sequence of the β67-loop was not conserved and contributed to acidification of the optimum pH in different ways. Mutations in the acidic residues in He1 promoted anti-tumor growth activity, which clarified the role of these acidic amino residues in the binding pocket. These findings will enable the identification of additional targets for modifying pH-mediated enzymatic activities.

INTRODUCTION

The edible mushroom Hericium erinaceus (Japanese name: yamabushitake) secretes He1, which is a ribonuclease (RNase) in the T1 family.¹⁾ The RNase T1 family includes guanylate specific ribonucleases that hydrolyze single-stranded RNA via a 2′, 3′-cyclic phosphate intermediate at the 3′-end of the oligonucleotides. The RNase T1 family includes two enzyme types: one with an acidic optimum pH and the other with a neutral optimum pH. Most of the RNase T1 family members belong to the neutral optimum pH group, with RNase He1, Ms, and U2 being the only three enzymes in the acidic optimum pH group, with an optimal pH of approximately pH 4.5.^2,3) The U2 is a member of the T1 family, but also cleaves adenine bases. The typical examples of neutral RNases include RNase T1 and Po1. RNase T1 is the most representative of the RNase T1 family, while RNase Po1 has the highest sequence identity with He1 among its family members with known structures.^4–9) The secondary structure of the RNase T1 family is highly conserved, with all members having the (α + β) motif,⁶⁾ while the amino acid sequence of the loops shows less conservation among these RNases. In particular, RNases that contain a more acidic amino acid in the α1β3- (He1:26–33, Ms: 31–38, U2:30–40) and β67-loops (He1:87–95, Ms:91–99, U2:102–110) are more specific to those active at an acidic pH (i.e., RNase He, Ms, and U2) than to those that function at a neutral pH. Of the acidic RNases, He1 is shorter than Ms by about one α-helix turn and shorter than U2 by three amino acids in α1β3-loop (Fig. 1). Po1 is a homolog of He1 with known anti-tumor progression activity; therefore, mutations in He1 that shift its optimal pH toward neutral and gain anti-tumor activity have clinical application potential.^6,10,11) Candidate residues in He1 that shift the optimal pH toward acidic have been identified; however, their mechanisms remain largely unknown. Recently, the structure of the He1/Zn complex was reported to be in the loop near the active center. This complex was disordered, and the base binding pocket was unable to bind the zinc ion, indicating a relationship between the optimal pH and the residues near the active center, although this relationship remains uncharacterized.¹²⁾

Fig. 1. Sequence Alignment of the T1, Ms, Po1, U2, and He1 RNases

Orange arrows indicate the β-sheet, and the green rectangles represent α-helices. The second structure is Po1. Pale blue shading indicates the catalytic residues. Red letters, yellow shading, and the black frame border indicate the acidic residues that are the focus of this study (e.g., Asp 31, Glu 92, and Asp 93) in He1. The red letter indicates the residues that were used for structural comparisons with the acidic residues of He1. The numbers at the top of the matrix are those of RNase He1. Asterisks indicate conserved residues.

In this study, we determined the structures of the He1/guanosine complex and He1 apo form, investigated the acidity mechanism of the optimal pH, and compared it with the structures of other RNase T1 family members.

MATERIALS AND METHODS

Enzymes

RNase He1 was expressed in Escherichia coli. The cDNA was ligated into the expression vector pET-pel-He1, which was constructed following the procedure for pET22b (Novagene, Darmstadt, Germany), and then transformed into E. coli BL21(DE3) pLysS (Novagene).¹³⁾ The cells were cultured in Terrific Broth with 100 µg/mL ampicillin and a final concentration of 0.5 mM isopropyl β-D-1-thiogalactopyranoside at 25 °C for 7 d (Wako Pure Chemical Corporation, Osaka, Japan). The supernatant of the culture was used for subsequent purification steps, which were conducted as previously described.¹²⁾

Enzyme Assay

RNase activity was measured as previously described using yeast RNA (Marine Biochemicals, Tokyo, Japan) as the substrate at 37 °C.¹⁴⁾

Protein Concentration

The protein concentration of the final enzyme preparation was determined spectrophotometrically at 280 nm with an absorbance of 0.905 in a 0.1% solution. This value was estimated from the amino acid composition of RNase He1 (data not shown).

Tricine-Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis

Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Tricine-SDS-PAGE) was performed using a 15% polyacrylamide gel following Schagger’s method.¹⁵⁾ Proteins in the gel were visualized with silver staining.

Crystallization of RNase He1 Complexed with 3′GMP

RNase He1 was mixed at a molar ratio of 1 : 5 with 3′GMP (Sigma Chemical. Co., St. Louis, MO, U.S.A.) on ice. The RNase He1 complexed with 3′GMP was crystallized at 293 K using the hanging-drop vapor diffusion method. Each drop was equilibrated against a 50 µL reservoir using the sitting-drop vapor diffusion method. Initially, we obtained a crystal in polyethylene glycol (PEG)/Ion HT (20% (w/v) PEG3350, 0.2 M magnesium sulfate heptahydrate). Further optimization was conducted around this condition using a 24-well crystallization plate with 0.5 µL protein solution mixed with 0.5 µL reservoir and equilibrated against the 500 µL reservoir using the hanging-drop vapor diffusion method. The crystal was obtained in 0.4 M magnesium sulfate and 20% PEG 3350 (pH 7.9) after two years of incubation.

Data Collection and Protein Structure Determination

RNase He1 Complexed with 3′GMP

Diffraction data were measured at 100 K and a wavelength of 1.100 Å using an Eiger X-4M detector on beamline BL-1A at the Photon Factory in Tsukuba, Japan. Data from a single crystal were integrated using the XDS package and scaled using Aimless from the CCP4 site.^16,17) The crystal belonged to the tetragonal space group I422. The structure was determined using molecular replacement with MOLREP using the structure of RNase He1 (PDB ID: 5gy6) as a search model.¹⁸⁾ There were two RNase He1 molecules in an asymmetric unit, one complexed with a phosphate ion and the other with guanosine. Structure refinement was performed with REFMAC5 and PHENIX.^19,20) The molecular model was manually corrected in the electron density map using Coot.²¹⁾ The final R_work and R_free values were 0.149 and 0.190, respectively. The Ramachandran statistics for the structure were as follows: favored regions, 99.5%; allowed regions, 0.5% (Table 1).

Table 1. Data Collection and Refinement Statistics Parameters

Crystal name	RNase He1 apo form and RNase He1/guanosine complex
Data collection
Beamline	Photon Factory BL-1A
Wavelength	1.100
Temperature (°K)	100
Rotation range per image (°)	1
Total rotation range (°)	360
Space group	I422
Resolution (Å)	48.88–1.58 (1.61–1.58)
a, b, c (Å)	97.76, 97.76, 68.91
α, β, γ (°)	90.00, 90.00, 90.00
Total number of reflections	611891 (30806)
Number of unique reflections	23162 (1131)
Completeness	100.0 (100.0)
Redundancy	26.4 (27.2)
⟨I/σ(I)⟩	23.2 (7.7)
Rmerge	0.108 (0.516)
CC1/2	0.998 (0.974)
Refinement statistics
Resolution	8.79–1.58
Number of reflections	22993 (2268)
Rwork	0.149 (0.168)
Rfree	0.190 (0.230)
No. of non-H atoms
Protein	1517
Ligand	32
Water	207
RMS (bonds)	0.015
RMS (angles)	1.57
Average B factors (Å2)
Protein	19.92
Ligand	20.73
Water	33.00
Ramachandran plot
Most favored	98.98
Additionally allowed	1.02
Rotamer outlier	0.63

Values in parentheses are for the highest resolution shell. R_merge = Σ_hkl Σ_i|I_i(hkl) − <I(hkl) > |/Σ_hkl Σ_iI_i(hkl), where I_i(hkl) is the intensity measured for the i-th reflection, and < I(hkl)> is the average intensity of all reflections with indices hkl. CC_1/2 is the correlation coefficient of the mean intensities between two random half-sets of the data. To calculate R_free, a subset of reflections (5%) was chosen as a test set. Ramachandran plot statistics is based on Rampage software available in the CCP4 software package.

RESULTS AND DISCUSSION

Purification and Crystal Structure Analysis of He1/Guanosine and He1 Apo Forms

The recombinant RNase He1 was expressed in E. coli and shown to be homogeneous using SDS-PAGE (Fig. 2). Crystals of the He1/guanosine and He1 apo forms were identified from space group I422 with unit lattice parameters a, b, and c (Å = 97.76, 97.76, 68.91, respectively). The structure was determined at 1.58 Å resolution using the molecular replacement method, which showed that the asymmetric unit comprised two molecules, the He1/guanosine complex and He1 apo form. The final refinement statistics were R_work = 0.149 and R_free = 0.190 (Table 1).

Fig. 2. Tricine-SDS-PAGE of RNase He1

Silver staining of the Tricine-SDS-PAGE of RNase He1, where (A) RNase He1 and (B) molecular protein markers are indicated. The molecular weight of He1 is 10.7 kDa.

The overall structure of the RNase He1/guanosine and He1 apo form had a 3.5-turn α-helix (residues 13–25) running like a spine down the two molecules. Both structures had seven β-sheets (residues 5–7, 10–12, 34–36, 51–55, 70–75, 80–85, and 96–97) and three intramolecular disulfide bonds (cysteine (Cys) 5–Cys 83, Cys 7–Cys 98, and Cys 47–Cys 81) (Figs. 3A, C). The base recognition sites of RNase He1 were tyrosine (Tyr) 36, histidine (His) 37, aspartic acid (Asp) 38, Tyr 39, glutamic acid (Glu) 40, and Asp 93, whereas the aromatic rings of Tyr 36 and Tyr 39 were stacked with a guanine base (Fig. 3B). The side chain of Asp 93 had a different conformation in the guanosine-bound complex and apo form, while the His 34, Glu 53, arginine (Arg) 71, and His 87 catalytic residues were in the same position in both complexes (Fig. 3D). Crystallization was performed using 3′GMP; however, the structure obtained included the phosphate group cleaved from 3′GMP. Therefore, it is possible that the time required for crystallization led to decomposition or the phosphate groups were unstable because the crystallization conditions were different from the protein’s optimal pH. The other molecule in the asymmetric unit showed a tetrahedral electron density in the active center. Sulfate ions in the crystallization reagent and the phosphate groups generated by the degradation of 3′GMP are possible candidates for this electron density; however, the structure was determined using sulfate ions in this model based on the ion concentrations. In reality, the sulfate and phosphate ions are possibly mixed together, with the sulfate ion being within the hydrogen bonding distance of His 34, Glu 53, Arg 71, and His 87 (Fig. 3D).

Fig. 3. Overall Structure of the RNase He1/Guanosine Complex and He1/Apo Form

(A) The structure of the RNase He1/guanosine complex. (B) The interaction between He1 (pale green) and the base of guanosine. (C) The structure of the RNase He1 apo form. (D) Catalytic residues of He1, where the pale green is the He1/guanosine, and pale cyan is the He1 apo form.

Acidification Mechanism of the Optimal pH for He1

In the RNase T1 family, there are acidic RNases Ms and He1 and neutral RNases Po1 and T1 enzymes, even though the sequence of the RNA binding pocket and the residues responsible for catalytic activity are conserved. Therefore, we focused on the amino acids in the α1β3- and β67-loops that were near the active center as domains that could contribute to the shift in the optimal pH.

Role of Asp 31

In acidic RNase T1 family members, an acidic amino residue (He1: Asp 31, Ms: Asp 36, and U2: Asp 37) was present three or four residues before the catalytic His residue in the β3 sheet (He1: His 34, Ms: His 39, and U2: His 41). In the He1/guanosine complex, the oxygen atom of the carboxylate group of the Asp 31 side chain formed a hydrogen bond, which was also observed in the He1/Zn complex (Fig. 4A). In the He1 apo form, the tetrahedral ion was located in the active center and formed a hydrogen bond with His 34 (Fig. 4B). Therefore, Asp 31 and His 34 did not form hydrogen bonds, even though the structure of Asp 31 and His 34 overlapped. Moreover, when the side chains of Asp 31 and His 34 were inverted, these residues formed hydrogen bonds.

Fig. 4. The Hydrogen Bond Network of Asp 31 and His 34 in the He1 Apo Form and He1/Guanosine Complex

The (A) RNase He1/guanosine complex and (B) RNase He1/apo form. The yellow lines indicate hydrogen bonds. In the He1/guanosine complex, His 34 forms hydrogen bonds with guanosine and Asp 31. In the He1 apo form, His 34 only forms a hydrogen bond the sulfate ion.

In RNase T1, the protonated His 40 and the hydrogen interaction between histidine and Glu 58 were hypothesized to be important for RNase activity.^22,23) In a neutral environment, the carboxylate group of the aspartic acid side chain becomes a carboxylic acid ion, and histidine forms a type III conformation with the nitrogen in the hydrogen π or τ position. The carboxylate ion of the Asp 31 side chain formed a hydrogen bond with the His 34 side chain, which prevented the hydrogen bond between His 34 and Glu 53, making the reaction less likely to occur in a neutral environment.

Roles of Glu 92 and Asp 93

The Glu 92 and Asp 93 residues were located in the β67-loop, which was constructed in the active center of the He1/guanosine complex (Fig. 5A). In the He1 apo form, the β67-loop was structurally stabilized by hydrogen bonds among His 87, threonine (Thr) 91, and glycine (Gly) 94. In contrast, Glu 92 and Asp 93 did not form hydrogen bonds with other amino acids, indicating that these residues had high flexibility in the β67-loop (Fig. 5B). In the He1 apo form, the carbon at the γ position and carboxylate group δ positions of Glu 92 and the carboxylate group at the γ position of Asp 93 were absent on the electron density map (Fig. 5C). This result suggested that the structures of these carboxylate groups were highly flexible. In the He1 apo form, the linear distance of the stable atoms in the β position of Glu 92 and Asp 93 was 5.1 Å, while that of the disordered atoms from the β to δ positions in the carboxylate group of the oxygen atom in Glu 92 was 3.3 Å, and the distance of the disordered atoms from the β to γ position in the carboxylate group of the oxygen atom of Asp 93 was 2.4 Å (Fig. 5D). Moreover, the carbonyl bases in the side chains of Glu 92 and Asp 93 were in the carboxylate ion form. In terms of distance and ionization state, the disorder of the side chains in Glu 92 and Asp 93 clashed and repulsed each other in the He1 apo form. One predicted conformation of the Asp 93 side chain included it facing toward the binding pocket, where it clashed with the ribose portion of the 3′GMP (Fig. 5E). Furthermore, Glu 92 may interact with the base of the downstream 3′ terminal of RNA, which was indicated by the ligand transfer from the structure of RNase Ms/2′-fluoroguanylyl-(3′-5′)-phosphocytidine (GpC) (PDB ID: 1RDS) to RNase He1 by fitting the main chain atoms (Fig. 5F). The ligand position of GpC suggested that the flexibility of Glu 92 was unfavorable for RNA binding with respect to its molecular contact. Conversely, in the He1/guanosine complex, the electron density map covered all atoms in Asp 93 but not the oxygen atom of the carboxylate base at the δ positions in Glu 92. Thus, Glu 92 and Asp 93 were more stable in the He1/guanosine complex than in the He1 apo form complex. The stability of the Asp 93 side chain was derived from the hydrogen bond between Thr 91 and Asp 93, which decreased the conformation of Asp 93, which lost its contact with Glu 92 and reduced the repulsion of their side chains. Hence, in a neutral environment, the carboxylate ions of the Glu 92 and Asp 93 side chains were unfavorable for 3′GMP binding and contributed to acidification of the optimum pH.

Fig. 5. Acidic Residues in the β67-Loop of He1

(A) Locations of Thr 91, Glu 92, and Asp 93 in the He1/guanosine complex. The black line indicates the hydrogen bond between Asp 93 and Thr 91 and Asp 93 and guanosine. (B) Location of Thr 91, Glu 92, and Asp 93 in the He1 apo form. (C) The upper panel is an electron density map of Glu 92 and Asp 93 in the He1 apo form (map level 1.0σ), and the lower panel is an electron density map of Glu 92 and Asp 93 in He1/guanosine apo form (map level 1.0 σ). (D) The location and distance of Glu 92 and Asp 93 in the He1 apo form. (E) Superimposed structures of He1/guanosine complex (green) and He1 apo form (cyan). (F) Superimposed structures of the He1/guanosine complex and GpC from the RNase Ms/GpC complex.

In a previous study, we mutated He1 to investigate the difference between Po1 and He1.¹¹⁾ The culture supernatant of two He1 point mutation variants (Asp 31 to asparagine (Asn) 31 and Glu 93 to glutamine (Gln) 93) indicated that the optimum pH shifted to a neutral environment (pH 5.5), while four other He1 variants (Asp 31 to Asn 31, Asp 38 to Asn 38, Asp 92 to Asn 92, and Glu 93 to Gln 93) shifted the optimum pH to an even more neutral environment (pH 6.5). The acidification mechanism of Asp 31, Glu 92, and Asp 93 was explained by the structures of He1 and the He1/guanosine complex.

Comparison of the He1/Guanosine, He1 Apo, and He1/Zn Complex Structures

Comparison of the He1/Zn complex¹²⁾ and He1/guanosine structures showed that the side chains of the catalytic residues His 34 and Glu 53 formed coordination bonds with zinc in the He1/Zn complex and that the positions of these residues were conserved (Figs. 6A, B). This conformational change was caused by the side chain of inverted His 37, which coordinated the bond to the zinc ion. With the conformational change in His 37, Asp 38 stabilized the β34-loop by forming hydrogen bonds with Gly 41, and phenylalanine (Phe) 42 formed a hydrogen bond with His 37, causing a conformational change from inside the β34-loop toward the β67-loop side (Fig. 6A). The β67-loop was disordered even though it did not directly interact with the zinc ion. In the He1/Zn complex, the hydrogen bond between His 87 and Gly 94 was lost, and His 87 formed a new hydrogen bond with the oxygen atom of the carboxylate group on the Asp 93 side chain, while the Thr 89 side chain formed a hydrogen bond with the main chain of Gly 94. In the He1 apo form, Asp 93 had a disordered carboxylate group in its side chain, which was flexible and moved freely. On superimposing the structures of He1/Zn complex and He1 apo form, we found that the shortest possible distance between the oxygen atom of the carboxylate base in the Asp 38 side chain of the He1/Zn complex and the oxygen atom of the carboxylate base in the He1 apo form Asp 93 side chain was 2.5 Å. This analysis indicated that these residues could repulse each carboxylate base (Fig. 6C); therefore, the inverted side chains of His 37 and Asp 38 and the disorder of the β34-loop due to zinc binding could have caused the repulsion between the Asp 38 and Asp 93 side chains and the loss of the interaction between the β34- and β67-loops. With this structural sequence change, the β67-loop may be important for destabilization, and the new hydrogen bond between His 87 and Asp 93 may cause conformational changes in the β67-loop such that it cannot directly interact with zinc ion (Fig. 6C).

Fig. 6. The Active Center of He1/Zn and He1 Apo Form

(A) The He1/Zn complex with the hydrogen bond (black line) and zinc ion (gray ball) indicated. (B) The He1 apo form with the hydrogen bond (black line) indicated. (C) The proposed mechanism of the disordered β67-loop in the He1/Zn complex. The zinc ion binds to His 37 and the β34-loop shifts toward the β67-loop, while Asp 38 and Asp 93 repulse each other to create a conformational change in the β67-loop, contributing to its disorder.

Comparison of the Acidification Mechanisms of RNases Ms and He1

In the RNase T1 family, He1 and Ms are RNases having optimal activity at an acidic pH. He1 and Ms have the same acidic amino acid residue in the α1β3-loop (He1: Asp 31; Ms: Asp 36) and two contiguous acidic residues in the β67-loop (He1: Glu 92 and Asp 93; Ms: Asp 97 and Asp 98). The structure comparison was conducted using the structures for Ms/3′GMP (PDB ID: 1RMS) and Ms/GpC (PDB ID: 1RDS).⁷⁾ Protein structure alignment also indicated that Asp 36 in Ms overlapped with Asp 31 in He1. Although there was no hydrogen bonding, the side chain of Asp 36 was oriented toward the side chain of His 39, and a hydrogen bond could form between Asp 36 and His 39 with a shift in the His 39 side chain. Thus, Asp 36 in Ms was predicted to have the same role as Asp 31 in He1 (Fig. 7A). The superposition of the main chain indicated that the location of the two contiguous acidic amino residues was different between the two acidic RNases (Ms: Asp 97 and Asp 98; He1: Glu 92 and Asp 93), while those in He1 were closer to the center of the β67-loop (Fig. 7B). In the Ms/3′GMP complex, Asp 97 recognized the hydroxy group at position 5 in the ribose with carboxylate groups on the main and side chains. In the Ms/3′GMP and GpC complexes, Asp 98 formed a hydrogen bond with the hydroxy group of the side chain of Tyr 44. In a neutral pH environment, aspartic acid was in a carboxylate ion form, and the hydrogen bond between the side chain of Asp 97 and the hydroxy group at position 5 in ribose of 3′GMP may have prevented 3′GMP from divergence and reduced the reaction rate. Moreover, the side chain of Asp 98 formed a hydrogen bond with that of Tyr 44, which fixed the position of Tyr 44 and lost the interaction between Tyr 44 and the base of 3′GMP. The aromatic ring of Tyr 39 in He1 corresponded to Tyr 44 in Ms and formed a CH–π interaction that contributed to the stability of the base in guanosine (Figs. 7C–E).

Fig. 7. Comparison of the Active Centers in RNases He1 and Ms

(A) Superposed structures of the He1/guanosine, Ms/3′GMP, and Ms/GpC complexes illustrating the overlap of Asp 31 and His 34 in He1 with Asp 36 and His 39 in Ms. (B) Superposed He1/guanosine, Ms/3′GMP, and Ms/GpC complexes illustrating the Tyr 39 in He1 and Tyr 44 in Ms in different conformations and the different locations of acidic amino acid in the β67-loop between He1 and Ms. The surfaces of the (C) He1/guanosine, (D) Ms/3′GMP, and (E) Ms/3′GpC complexes. (A, B) Green: He1/guanosine complex, orange: Ms/3′GMP, pink: Ms/GpC complex. (C–E) Green: the main chain carbon in He1/guanosine complex, orange: the main chain carbon in Ms/3′GMP, pink: Ms/GpC complex, red: oxygen atom, blue: nitrogen atom.

Therefore, in a neutral environment, the Asp 98 of Ms could reduce the base stability by interacting with the tyrosine that forms the base binding pocket on the β34-loops.

The side chains of Glu 92 and Asp 93 in He1 might not be the site of the direct interactions between He1 and 3′GMP or the important residues that recognize the base, but rather, repulsively act to support the interaction between each acidic residue and cause a steric hindrance between 3′GMP and the active pockets to decrease RNase activity in neutral environments. These results suggested that both He1 and Ms have acidic residues on their α1β3- and β67-loops, but the acidification of their optimal pH occurs through different mechanisms.

Comparison of the Acidification Mechanisms of RNases U2 and He1

RNase U2 is another acidic member of the RNase T1 family. The amino acid sequence alignment of Ms and He1 indicated that the Asp 31 in He1, which contributed to the acidification of optimum pH, corresponded to Asn 38 in U2. However, the structure alignment of RNase U2/adenosine 3′-monophosphate (3AM) (PDB ID: 3AGO) and He1 showed that Asp 31 in He1 closely aligned with Asp 37 in U2. In U2, to form the hydrogen bond between Asp 37 and His 41, the side chain of Asp 37 had to be flipped toward His 41 (Fig. 8A). The α1β3-loop of U2 was longer than He1 and varied in its conformation, including where Asp 37 binds to calcium and is not oriented towards His 41.^8,9) Therefore, Asp 37 does not contribute to acidification of optimum pH in all situations; however, Asp 37 in U2 could partially perform the same role as Asp 31 in He1. In the β67-loop of U2, Asp 108 overlapped with Asp 93, and Tyr 107 corresponded to Glu 92 on the superposed main chain. In the U2 apo form (PDB ID: 1RTU), serine (Ser) 106 formed a hydrogen bond with Asp 108, and the side chain of Tyr 107 was disordered. In the β67-loop of He1, Glu 92 and Asp 93 repulsed each other and had high flexibility. Conversely, in the β67-loop of U2, Asp 108 formed a hydrogen bond with Ser 106, and the side chain of Tyr 107 did not interact with other residues. Therefore, the side chain of Tyr 107 might have high flexibility and repulse 3′GMP and 3′AMP (Fig. 8B). Thus, U2 had only one acidic amino acid in the β67-loop, suggesting that its mechanism of acidification of the optimum pH was different from that of He1. The acidification of optimum pH for U2 could be via the α2 helix and the α1β3- and β67-loops. The α2 helix was not present in He1 and Ms and involved acidic amino acids in U2. In the α2 helix of U2, the carboxylate group of the Glu 49 side chain that recognized the guanosine base was located at 3.4 Å from the carboxylate group of the Glu 46 side chain (Fig. 8C). Thus, Glu 46 and Glu 49 might repulse each other and prevent base recognition in neutral environments. Given these findings, the acidification of optimum pH for U2 activity was caused by an acidic amino acid not only in the α1β3- and β67-loops but also in the α2 helix; this further suggested that the acidification of the U2 optimal pH occurs through different mechanisms than that of He1.

Fig. 8. Comparison of Active Centers of the RNases He1 and U2

(A) Superposed structures of He1/guanosine and U2/AMP complexes illustrating the overlap of His 34 in He1 and His 41 in U2. The side chain of Asn 38 is oriented in the opposite direction to His 41 in U2, while the side chain of Asp 37 is close to His 41 in U2. (B) The superposed He1 apo and Ms apo forms show that the Asp 93 in He1 overlaps with Asp 108 in U2. In U2, Ser 106 forms a hydrogen bond with Asp 108, and Tyr 107 has high flexibility. (C) The α2 helix of U2 illustrating that the distance between the carboxylate goup of Glu 46 and Glu 49 is 3.5 Å. (A, C) Green: the main chain carbon in the He1/guanosine complex, light brown: the main chain carbon in the U2/AMP complex. (B) Cyan: the main chain carbon in He1 apo form, yellow: the main chain carbon in U2 apo form. (A–C) Red: oxygen atom, blue: nitrogen atom, orange: phosphate atom.

CONCLUSION

The structure of the He1 apo form and He1/guanosine complex were determined using X-ray crystallography. In neutral pH environments, Asp 31 in the α1β3-loop interacted with His 34 and influenced the hydrogen network in the active center. Moreover, the Glu 92 and Asp 93 residues in the β67-loop repulsed each other and inhibited 3′GMP binding. This structure supported the optimal pH shift of the mutated He1 (Asp 31/Asn 31 and Asp 93/Asn 93) from pH 4.5 to 5.5 in a previous study.¹¹⁾ Asp 36 in Ms and Asp 37 in U2 corresponded to Asp 31 in He1 and played a similar role in the acidification of the optimum pH. While Asp 97 and Asp 98 in Ms formed a different hydrogen bond network than that of Glu 92 and Asp 93 in He1, Asp 108 in U2 had a similar role to that of Asp 93 in He1. In U2, no residue corresponded to Glu 92 in He1, and the α2 helix of U2 contributed to the acidification of the optimum pH.

Given these results, we hypothesize that the Asp residue in the α1β3-loop is a common feature of the acidic optimum RNase T1 family members. However, although the amino acid sequence of the β67-loop is not conserved, it contributes to the acidification of the optimum pH differently in each acidic RNase. In the development of new anti-tumor drugs, one important requirement is that the optimum pH is within physiological conditions. Focusing on the structural characteristics of acidic amino acids near the active center and shifting the optimum pH toward neutral using point mutations is a potentially important method for creating novel anti-tumor drug candidates.

Acknowledgments

We thank Dr. Ryuichi Kato for the crystallization of RNase He1. This research was partially supported by the Platform Project for Supporting Drug Discovery and Life Science Research from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) and the Japan Agency for Medical Research and Development (AMED) (Proposal Nos. 2012G001, 2013R-11). This work was performed in part under the Collaborative Research Program of the Institute for Protein Research, Osaka University, CR-15-05, CR-16-05, CR-17-05, CR-18-05, CR-19-56, CR-20-54, CR-21-45, CR-22-33, CR -23-02. The atomic coordinates and structure factors (PBD ID code 6LS1) have been deposited in the Worldwide Protein Data Bank (http:www.pdb.org/).

Conflict of Interest

The authors declare no conflict of interest.

REFERENCES

• 1) Yoshida H. The ribonuclease T1 family. Methods Enzymol., 341, 28–41 (2001).
• 2) Watanabe H, Ohgi K, Irie M. Primary structure of a minor ribonuclease from Aspergillus saitoi. J. Biochem., 91, 1495–1509 (1982).
• 3) Arima T, Uchida T, Egami F. Studies on extracellular ribonucleases of Ustilago sphaerogena. Purification and properties. Biochem. J., 106, 601–607 (1968).
• 4) Nomura H, Inokuchi N, Kobayashi H, Koyama T, Iwama M, Ohgi K, Irie M. Purification and primary structure of a new guanylic acid specific ribonuclease from Pleurotus ostreatus. J. Biochem., 116, 26–33 (1994).
• 5) Arni R, Heinemann U, Tokuoka R, Saenger W. Three-dimensional structure of the ribonuclease T1 2′-GMP complex at 1.9-A resolution. J. Biol. Chem., 263, 15358–15368 (1988).
• 6) Kobayashi H, Katsutani T, Hara Y, Motoyoshi N, Itagaki T, Akita F, Higashiura A, Yamada Y, Inokuchi N, Suzuki M. X-Ray crystallographic structure of RNase Po1 that exhibits anti-tumor activity. Biol. Pharm. Bull., 37, 968–978 (2014).
• 7) Nonaka T, Nakamura KT, Uesugi S, Ikehara M, Irie M, Mitsui Y. Crystal structure of ribonuclease Ms (as a ribonuclease T1 homologue) complexed with a guanylyl-3′,5′-cytidine analogue. Biochemistry, 32, 11825–11837 (1993).
• 8) Noguchi S, Satow Y, Uchida T, Sasaki C, Matsuzaki T. Crystal structure of Ustilago sphaerogena ribonuclease U2 at 1.8 A resolution. Biochemistry, 34, 15583–15591 (1995).
• 9) Noguchi S. Isomerization mechanism of aspartate to isoaspartate implied by structures of Ustilago sphaerogena ribonuclease U2 complexed with adenosine 3′-monophosphate. Acta Crystallogr. D Biol. Crystallogr., 66, 843–849 (2010).
• 10) Kobayashi H, Motoyoshi N, Itagaki T, Tabata K, Suzuki T, Inokuchi N. The inhibition of human tumor cell proliferation by RNase Pol, a member of the RNase T1 family, from Pleurotus ostreatus. Biosci. Biotechnol. Biochem., 77, 1486–1491 (2013).
• 11) Kobayashi H, Motoyoshi N, Itagaki T, Suzuki M, Inokuchi N. Effect of the replacement of aspartic acid/glutamic acid residues with asparagine/glutamine residues in RNase He1 from Hericium erinaceus on inhibition of human leukemia cell line proliferation. Biosci. Biotechnol. Biochem., 79, 211–217 (2015).
• 12) Kobayashi H, Sangawa T, Takebe K, Motoyoshi N, Itagaki T, Suzuki M. X-ray crystallographic structure of Hericium erinaceus ribonuclease, RNase He1 in complex with zinc. Biol. Pharm. Bull., 42, 2054–2061 (2019).
• 13) Huang HC, Wang SC, Leu YJ, Lu SC, Liao YD. The Rana catesbeiana rcr gene encoding a cytotoxic ribonuclease. Tissue distribution, cloning, purification, cytotoxicity, and active residues for Rnase activity. J. Biol. Chem., 273, 6395–6401 (1998).
• 14) Irie M. Isolation and properties of a ribonuclease from Aspergillus saitoi. J. Biochem., 62, 509–518 (1967).
• 15) Schägger H, von Jagow G. Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal. Biochem., 166, 368–379 (1987).
• 16) Kabsch W. XDS. Acta Crystallogr. D Biol. Crystallogr., 66, 125–132 (2010).
• 17) Evans PR, Murshudov GN. How good are my data and what is the resolution? Acta Crystallogr. D Biol. Crystallogr., 69, 1204–1214 (2013).
• 18) Vagin A, Teplyakov A. Molecular replacement with MOLREP. Acta Crystallogr. D Biol. Crystallogr., 66, 22–25 (2010).
• 19) Murshudov GN, Skubák P, Lebedev AA, Pannu NS, Steiner RA, Nicholls RA, Winn MD, Long F, Vagin AA. REFMAC5 for the refinement of macromolecular crystal structures. Acta Crystallogr. D Biol. Crystallogr., 67, 355–367 (2011).
• 20) Adams PD, Afonine PV, Bunkóczi G, Chen VB, Davis IW, Echols N, Headd JJ, Hung LW, Kapral GJ, Grosse-Kunstleve RW, McCoy AJ, Moriarty NW, Oeffner R, Read RJ, Richardson DC, Richardson JS, Terwilliger TC, Zwart PH. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr. D Biol. Crystallogr., 66, 213–221 (2010).
• 21) Emsley P, Lohkamp B, Scott WG, Cowtan K. Features and development of Coot. Acta Crystallogr. D Biol. Crystallogr., 66, 486–501 (2010).
• 22) Steyaert J, Hallenga K, Wyns L, Stanssens P. Histidine-40 of ribonuclease T1 acts as base catalyst when the true catalytic base, glutamic acid-58, is replaced by alanine. Biochemistry, 29, 9064–9072 (1990).
• 23) Zegers I, Verhelst P, Choe HW, Steyaert J, Heinemann U, Saenger W, Wyns L. Role of histidine-40 in ribonuclease T1 catalysis: three-dimensional structures of the partially active His40Lys mutant. Biochemistry, 31, 11317–11325 (1992).