2014 年 37 巻 6 号 p. 968-978
RNase Po1 is a guanylic acid-specific ribonuclease member of the RNase T1 family from Pleurotus ostreatus. We previously reported that RNase Po1 inhibits the proliferation of human tumor cells, yet RNase T1 and other T1 family RNases are non-toxic. We determined the three-dimensional X-ray structure of RNase Po1 and compared it with that of RNase T1. The catalytic sites are conserved. However, there are three disulfide bonds, one more than in RNase T1. One of the additional disulfide bond is in the catalytic and binding site of RNase Po1, and makes RNase Po1 more stable than RNase T1. A comparison of the electrostatic potential of the molecular surfaces of these two proteins shows that RNase T1 is anionic whereas RNase Po1 is cationic, so RNase Po1 might bind to the plasma membrane electrostatically. We suggest that the structural stability and cationic character of RNase Po1 are critical to the anti-cancer properties of the protein.
RNase Po1 hydrolyzes single-stranded RNA via a 2′,3′-cyclic phosphate intermediate at the 3′-terminus of oligonucleotides, and is a guanylic acid-specific ribonuclease (RNase) (a member of the RNase T1 family of RNases). RNase Po1 has a molecular mass of approximately 11 kDa and exhibits optimal activity at pH 7.5, similar to RNase T1 from Aspergillus oryzae, the best-known member of this family.1,2) We previously isolated and purified RNase Po1 from Pleurotus ostreatus of the basidiomycota and reported its amino acid sequence.1) There is high sequence identity (40%) between RNase Po1 and RNase T1. The X-ray crystallographic structures of RNase T13) and RNase Ms4) have been reported, and the catalytic site and base recognition region for RNase activity have been elucidated. The catalytic site of RNase T1 consists of His40 and/or Glu58, Arg77, and His92.5) Moreover, Steyaert et al.6) and Nonaka et al.7) reported that Glu58, rather than His40, must be the general base catalyst of the RNase T1 family. The base recognition site of RNase T1 consists of Tyr42, Asn43, Asn44, Tyr45, Glu46, and Asn98.5) These residues are completely conserved in RNase Po1 except for Tyr45 (RNase T1) being changed to Phe (RNase Po1). A comparison of the primary structures of RNase Po1 and RNase T1 and other RNase T1 family RNases shows that RNase T1 contains four cysteine residues that form two disulfide bonds, whereas RNase Po1 has six cysteine residues1,8–23) (Fig. 1). The C9–C99 disulfide bond of RNase Po1 is superimposable on the analogous disulfide bond in RNase T1 (C6–C103). This disulfide bond is conserved in all known RNase T1 family enzymes, except for those of bacterial origin. The C48–C82 bond of RNase Po1 is superimposable on RNase U1 and RNase U2 from Ustilago sphaerogena. Therefore, these two disulfide bonds may exist in RNase Po1. The other cysteine residues in RNase Po1 (Cys7, Cys84) are not found in RNase T1 and other RNase T1 family RNases. Therefore, we have not considered whether or not RNase Po1 has three disulfide bonds, and which cysteine residues may form those bonds. Furthermore, RNase Po1 is an alkaline protein (isoelectric point (pI) 9.0), whereas RNase T1 (pI 2.9) and most members of the RNase T1 family of RNases are acidic proteins (pI 4.0–4.5).24,25) Thus, RNase Po1 is a unique member of the RNase T1 family. Recently, we reported that RNase Po1 exhibits anti-tumor activity towards several types of human tumor cells,26) in contrast to RNase T1 and other RNase T1 family RNases that are non-toxic to tumor cells, except for α-sarcin from Aspergillus giganteus.27) α-Sarcin is a single polypeptide chain protein composed of 150 amino acids, bigger than RNase Po1 (101 amino acids) and with low sequence identity.28) α-Sarcin exhibits anti-tumor activity by degrading the larger ribosomal RNA of tumor cells.29) The three-dimensional structure of α-sarcin has been solved,30) and it is necessary to compare its structure with that of RNase Po1 to understand the basis of their anti-tumor activities.
Po1: Pleurotus ostreatous RNase,1) U1, U2: Ustilago sphaerogena RNase,8,9) F1: Fusarium moniliforme RNase,10) FL1: Fusarium lateritium RNase,11) Th1: Trichoderma harzuanum RNase,12) Ms: Aspergillus saitoi RNase,13) T1: Aspergillus oryzae RNase,14) C2: Aspergillus clavatus RNase,15) Ap1: Aspergillus pallidus RNase,16) N1: Neurospora crassa RNase,17) Pch1: Penicillium crysogenum RNase,18) Pb1: Penicillium brevicompactum RNase,19) Sa: Streptomyces aureofaciens RNase,20) St: Streptomyces rythreus RNase,21) Bi: Bacillus intermedius RNase,22) Ba: Bacillis amyloliquefacience RNase.23) * Catalytic site. # Base recognition site. Disulfide bonds of RNases are shown as connected by bold lines.
There is another RNase family whose members have molecular masses of 13–14 kDa. This family is pyrimidine base-specific and is referred to as the RNase A family. Almost all RNase A family members are non-toxic, but some RNases have been reported to exhibit anti-tumor activity.31–35) RNases from Rana pipiens are the most extensively studied, and ranpirnase (Onconase) is currently in clinical trials as an anti-tumor drug.36,37) The X-ray crystallographic structures of RNase A and Onconase have been reported.38,39) Comparisons of their structures suggest a relationship between stability and anti-tumor activity.40,41) In the present work, we determined the X-ray crystallographic structure of RNase Po1 and investigated the relationship between structure and anti-tumor activity by comparing the X-ray structures of RNase Po1 and RNase T1, which have high sequence identity.
RNase Po1 was expressed in Escherichia coli. The cDNA was ligated to expression vector pET-pel-Po1, constructed following the procedure of Huang et al.42) from pET22b (Novagene, Darmstadt, Germany), then transferred to E. coli BL21(DE3) pLysS (Novagene). The cells were cultured in Terrific Broth at 25°C for 7 d, with the addition of 100 µg/mL of ampicillin and a final concentration of 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) (Wako Pure Chemical Industries, Ltd., Osaka, Japan). The supernatant of the culture was used for subsequent purification steps. The supernatant was fractionated with 90% saturated ammonium sulfate, and the precipitate was collected by centrifugation at 10000 rpm for 30 min. The precipitate was suspended in 10 mM acetate buffer (pH 6.0) and dialyzed overnight against de-ionized water. The dialysate was heated at 60°C for 10 min and then rapidly cooled in ice-cold water for 10 min. The precipitate was recovered by centrifugation at 10000 rpm for 30 min. Subsequent purification steps were carried out using previously described protocols.26)
RNase activity was measured as described previously using yeast RNA (Marine Biochemicals, Tokyo, Japan) as the substrate at pH 7.5 at 37°C.43)
The protein concentration of the final enzyme preparation was determined spectrophotometrically, assuming an absorbance of 0.54 for a 0.1% solution at 280 nm. This value was estimated from the amino acid composition of RNase Po1 (data not shown).
Tricine-SDS-PAGE was performed using a 15% polyacrylamide gel by Schagger’s method.44) Proteins on the gel were stained by silver staining. Activity staining of the RNases was conducted using the method of Blank et al.45)
RNase Po1 was crystallized at 293 K by the hanging drop vapor diffusion method. An initial crystal was obtained using condition #31 (3.5 M sodium formate, 0.1 M 1,3-bis[tris(hydroxymethyl)methylamino]propane-1,3-diol (Bis-Tris propane) (pH 7.0) in SaltRx1 (Hampton Research, Aliso Viejo, CA, U.S.A.). Drops were placed over a well containing 500 µL of reservoir solution (4 M sodium formate, 0.1 M Bis-Tris propane, pH 7.0, 10% polyethylene glycol (PEG) 400). Crystals suitable for data collection were obtained in drops containing 0.5 µL protein solution (10 mg/mL RNase Po1 in 20 mM Tris–HCl, pH 7.5) mixed with 0.5 µL reservoir solution (4 M sodium formate, 0.1 M Bis-Tris propane, pH 7.0, 10% PEG 400) and 0.3 µL 2 M cesium chloride after 3 d.
Diffraction data were measured at 100 K and a wavelength of 0.9800 Å using an ADSC Quantum 270 detector on beamline BL-17A at the Photon Factory (Tsukuba, Japan). Data from a single crystal were integrated and scaled using XDS46) and SCALA.47) The crystal belonged to trigonal space group P31. The structure was determined by molecular replacement with BALBES48) and MOLREP.49) The presence of the three subunits of RNase Po1 in an asymmetric unit gave a crystal volume per protein mass (VM) of 1.78 Å3 Da−1 and a corresponding solvent content of 31.0%.50) Structure refinement was performed using PHENIX.51) The molecular model was manually corrected in the electron density map using COOT.52) The final Rwork and Rfree values were 0.164 and 0.176, respectively. The Ramachandran statistics for the structure calculated using MolProbity53) were as follows: favored regions, 98.7%; allowed regions, 1.3%; outlier regions, 0% (Table 1).
Values in parentheses are for the highest-resolution shell. a) Rmerge=∑hkl ∑i |I(hkl; i)−〈I(hkl)〉|/∑hkl ∑i I(hkl; i), where I(hkl; i) is the intensity of an individual measurement, and 〈I(hkl)〉 is the average intensity from multiple observations. b) Rwork=S||Fobs|−|Fcalc||/|Fobs|, where |Fobs| and |Fcalc| are the observed and calculated structure factor amplitudes, respectively. c) Rfree is the same as Rwork, but for a 10.0% subset of all reflections for RNase Po1.
RNase Po1 and RNase T1 at 0.16 mg/mL were treated at a ratio of 1 : 17.5 (w/w) chymotrypsin : RNase in 10 mM Tris–HCl at pH 7.5 and 37°C. RNase Po1 and RNase T1 at 0.14 mg/mL were treated at 37°C with thermolysin at a ratio of 1 : 20 (w/w) thermolysin : RNase in 10 mM Tris–HCl at pH 7.5 containing 1 mM CaCl2. RNase activity in the reaction mixture was measured as described above.
Transfections of RNase Po1 and RNase T1 were performed according to the manufacturer’s instructions with cell-penetrating peptide (CPP) (TaKaRa Bio Inc., Shiga, Japan). The human leukemia cell line HL-60 was purchased from the Health Science Research Resources Bank (Osaka, Japan). HL-60 cells were cultured in RPMI 1640 supplemented with 10% fetal calf serum. The cells were collected by centrifugation and then suspended in in RPMI 1640 without fetal calf serum and diluted to 1.5×106 cells/mL. One hundred and ten microliters of cell suspension was added to each well of a 96-well plate, and then 20 µL of 0.1 or 1.0 µM final concentration of RNase Po1 or RNase T1 previously filtered through a Millipore filter (Millex GV, Billerica, MA, U.S.A.) was treated with Xfect Protein Transfection reagent for 30 min and added to the cells. Viable cells were counted by MTT yellow tetrazole assay, after 18 h of incubation at 37°C under 5% CO2. After the addition of 10 µL of 0.5% MTT solution (Dojindo Laboratories, Kumamoto, Japan) to each well, incubation was continued for an additional 2–3 h at 37°C under 5% CO2. The absorbance at 630 nm was then measured. Inhibition of cell proliferation was calculated as the percent decrease in final cell numbers as compared to the absence of RNase Po1 or RNase T1.
The amounts of RNase Po1 or RNase T1 in HL-60 cells were estimated using RNase activity as a marker. Two milliliters of HL-60 cell suspension containing 3×105 cells/mL were incubated with 200 µL of 6 µM final concentration RNase Po1 or 6 µM RNase T1 or phosphate buffered saline (PBS) as a control in RPMI 1640 medium (Invitrogen, Carlsbad, CA, U.S.A.) with 10% fetal calf serum (Bio West, Strasbourg, France) for 30 min at 37°C under 5% CO2. The cells were collected by centrifugation at 1500 rpm and washed with PBS twice. The cell pellet was suspended in 0.1 mL of 0.5% Triton X-100 in 0.1 M Tris–HCl buffer (pH 7.5) and frozen and thawed twice, then centrifuged again. The RNase activity of the supernatant was measured as described above and the RNase activity of the control was deducted (the endogenous RNase activity of HL-60 cells). The relative rate of internalization was then calculated.
Expression of RNase Po1 in E. coli provided about 35000 units (approximately 17 mg) of RNase Po1 in 2 L of culture supernatant. RNase Po1 was homogeneous according to SDS-PAGE (Fig. 2). Purified RNase Po1 was crystallized with space group P31 and unit cell dimensions a=b=75.56 (Å), c=34.80 (Å). The structure of RNase Po1 was solved using the Molecular Replacement method. The structure was refined at 1.85 Å resolution, with final Rwork and Rfree factors of 0.164 and 0.176, respectively.
Purified RNase Po1 was homogeneous according to Tricine-SDS-PAGE. Tricine-SDS-PAGE was performed using a 15% polyacrylamide gel. (a) Silver staining of molecular marker proteins. (b) Silver staining of RNase Po1. (c) Activity staining of RNase Po1 using RNA as the substrate.
The overall structure of RNase Po1 is an (α+β)-type structure consisting of an α-helix (residues 16–27) and seven β-strands (residues 7–9, 12–14, 36–38, 52–56, 71–76, 82–86, 97–98), the same as RNase T1 (Figs. 3, 4). The refined model of RNase Po1 has an α-helix with 3.5 turns, which is fewer than the helix in RNase T1 (4.5 turns). In RNase Po1, the helix runs like a “backbone” down the molecule, and a four-stranded anti-parallel β-sheet (36–38, 52–56, 71–76, 82–86) cross the α-helix. The catalytically active amino acid residues of RNase T1 (His40, Glu58, Arg77) are located in the β3–6 strands (His40, Glu58, Arg77) and next to the β6 strands (His 92). It has also been reported for RNase T1 that in the transesterification step of phosphodiester hydrolysis, His40 and/or Glu58 act as a general base toward the ribose 2′-hydroxyl group and that His92, as a general acid, donates a proton to the leaving 5′-hydroxyl group.5) The catalytically active amino acid residues of RNase Po1 corresponding to RNase T1 are His36, Glu58, Arg72, and His87. These amino acid residues are also located in the β3–6 strands (His36, Glu58, Arg72) and next to the β6 strands (His 87). The base recognition site of RNase T1 consists of Tyr42, Asn43, Asn44, Glu46, Tyr45, and Asn98 and is located in the loop between β3–4 strands (Asn43, Asn44, Tyr45, Glu46) and in the loop between β6–7 strands (Asn98). The aromatic rings of Tyr42 (in the β3 strand) and Tyr45 stack with the guanine base.5) In case of the base recognition site of RNase Po1, the amino acid residues Tyr38, Asn39, Asn40, Phe41, Glu42, and Asn94 correspond to those of RNase T1. These amino acid residues are also located in the loop between β3–4 strands (Asn39, Asn40, Phe41, Glu42) and in the loop between β6–7 strands (Asn94). The aromatic rings of Tyr38 (in the β3 strand) and Phe41 in RNase Po1 may stack with the guanine base. Two important roles for the β-sheet in RNase T1 have been reported.2) One is that the β-sheet forms an internal hydrophobic core, and the second is that the β-sheet structure forms the catalytic pocket. The three hydrophilic side chains of Glu58, Arg77, and His92 in RNase T1 form a cluster on the β-sheet surface opposite the α-helix and are involved in enzymatic activity. The Glu54 (in the β4-strand), Arg72 (in the β5-strand), and His 87 (next to the β6-strand) residues of RNase Po1 are conserved with them completely. The β4–6 sheet of RNase Po1 is inside the molecule and consists of 16 amino acid residues, of which eight are hydrophobic, as in RNase T1 (9 hydrophobic residues, 17 total). The hydrophobic amino acid residue clusters of the β4–6 sheet of RNase Po1 are at opposite to the α-helix and may interact with those of other strands and the α-helix, which then forms an internal hydrophobic core similar to RNase T1 (Fig. 5).
The figure was drawn with PyMOL (http://pymol.sourceforge.net). (a), (b) The N- and C-termini are labeled N and C, respectively. The α-helices and β-strands are marked α1 and β1–7, respectively. The circles enclose disulfide bonds. (a) RNase T1 is colored pink (PDB ID: 2BU4, Proteins 1999, 36, 117–134), 2′GMP is shown as sticks colored blue and red. (b) RNase Po1 is colored blue (PDB ID: 3WHO). (c) Structural overlay of RNase Po1 with that of the RNase T1/2′GMP complex. Active site residues of RNase Po1 and RNase T1 are colored blue and pink, respectively. The α-helices and β-strands are marked α1 and β1–7, respectively. The disulfide bonds of RNase Po1 are shown as sticks colored yellow. (d) The active site of RNase Po1 superimposed with that of the RNase T1/2′GMP complex from (c). 2′GMP is gray. In (c) and (d), the amino acid numbers of RNase Po1 are shown first and those of RNase T1 follow in parentheses.
Po1, RNase Po1; T1, RNase T1. Sequences in common are enclosed in boxes. Numbers above and below the matrix show RNase Po1 and RNase T1 numbering, respectively. The cysteine residues are shaded, and the disulfide bonds of RNase Po1 are shown as connected by bold lines. The catalytic site residues are indicated by arrows. Secondary structures are denoted as follows: α1, α-helix; βn, strand of β-sheet structure.
The figure was drawn with PyMOL (www.pymolwiki.org/index.php/Color_h). The molecules are shown as ribbons and surface models. The figure (b) is rotated horizontally by 180° by the center. βn, strand of β-sheet structure. The hydrophobic regions are colored gray and the deepness of the dark color shows hydrophobic strength. The circles enclose the hydrophobic core.
RNase Po1 has six cysteine residues. We determined the disulfide bond combinations of these residues to be Cys9–Cys99, Cys7–Cys84, and Cys48–Cys82 (Fig. 6). One disulfide bond in RNase Po1 (Cys9–Cys99) is superimposable on the analogous bond in RNase T1 (Cys6–Cys103). This disulfide bond is conserved in all known RNase T1 family enzymes, except for those of bacterial origin. The locations of the other two disulfide bonds in RNase Pol (Cys7–Cys84 and Cys48–Cys82) do not correspond to any disulfide bond in RNase T1 (Fig. 4). The Cys7–Cys84 bond is located parallel to the Cys9–Cys99 bond, thus making the C-terminus and N-terminus of RNase Po1 more rigid than those in RNase T1. The Cys48–Cys82 bond of RNase Po1 is located near the active site on the opposite side of the Cys7–Cys84 bond. Moreover, the Cys48–Cys82 bond connects the β6-strand to the loop between the β3-strand and β4-strand. In this loop, there are some amino acid residues (Asn39, Asn40, Phe41, Glu42) thought to constitute the base recognition site. Next to the β6-strand, there is one catalytic residue (His87). Therefore, the Cys48–Cys82 bond may help maintain the conformational stability of the base recognition region and catalytic site (Fig. 3).
The cysteine residues and disulfide bonds are shown as sticks colored blue and yellow, respectively. The contour level is 1.2σ. βn, strand of β-sheet structure. (Color images were converted into gray scale.)
We digested RNase Po1 and RNase T1 with chymotrypsin and thermolysin to study their conformational stabilities. These proteases have broad specificity, so the proteolytic cleavage sites are primarily determined by accessibility to the substrate rather than on the primary structure of the substrate. RNase Po1 was more resistant to proteolysis than RNase T1, and RNase Po1 retained 50–60% RNase activity after incubation with these proteases for 20 h, whereas RNase T1 retained only 15% RNase activity (Fig. 7).
Digested RNase Po1 and RNase T1 by chymotrypsin (a) and thermolysin (b) and measured RNase activity with reaction times. RNase activity just after starting the reaction was normalized to 100%. Circles show RNase Po1 and triangles show RNase T1.
We transfected RNase Po1 and RNase T1 with CPP and incubated them for 1 h. Viable cells were the same as for non-CPP cells. We then incubated them for 18 h, after which RNase Po1-CPP had decreased viable cells dependent on the RNase Po1 concentration, 20% (0.1 µM RNase Po1) or 60% (1.0 µM RNase Po1) lower than that of non-CPP cells. In contrast, RNase T1-CPP had a maximum of 10% decreased viable cells (1.0 µM RNase T1) (Fig. 8). We thus concluded that RNase Po1 caused greater inhibition of proliferation in HL-60 cells compared to RNase T1 after transfection.
HL-60 cells were treated with a given concentration of RNase Po1 or RNase T1 with or without cell-penetrating peptide (CPP). Viable cells were counted by MTT assay after 18 h of incubation at 37°C under 5% CO2. Cell proliferation without RNase was normalized to 100%. The graphs of −CPP show cell proliferation without CPP and are colored black, and those of +CPP show cell proliferation with CPP and are gray. The data represent the means and standard errors of three independent experiments, each performed in triplicate. (a), RNase Po1. (b), RNase T1.
We compared the amounts of RNase Po1 and RNase T1 in HL-60 cells internalized into HL-60 cells using RNase activity as a marker. The results are shown in Table 2. The RNase activity of RNase Po1 and RNase T1 deducted the endogenous RNase activity of HL-60 cells (7.0×10−3 units) by 17.78×10−3 units and 2.48×10−3 units, respectively. We calculated that the relative rate in HL-60 cells of RNase Po1 activity was four times higher than that of RNase T1. This suggested that RNase Po1 was internalized into HL-60 cells at a greater rate than RNase T1 and/or that RNase Po1 was considerably more stable in the cells.
There is high sequence identity (40%) between RNase Po1 and RNase T1, and complete conservation of the catalytic sequence.1) Both proteins are guanine-specific RNases with an optimum pH of 7.5 and with comparable specific activity towards RNA substrates. However, RNase Po1 exhibits anti-tumor activity towards several types of human tumor cells, whereas RNase T1 is non-toxic towards them.26) To investigate this difference, we determined the X-ray crystallographic structure of RNase Po1 and compared it with that of RNase T1. RNase Po1 has an α-helix with 3.5 turns, which is shorter than the helix in RNase T1 (4.5 turns); thus, RNase Po1 has a more spherical shape (Fig. 3). The catalytic sites of RNase Po1 and RNase T1 are conserved, and each are composed of four β-strands (β3–6). A comparison of the molecular surfaces of RNase Po1 and RNase T1 clearly shows that although the catalytic residues are conserved, the areas around the catalytic residues are poorly conserved (Fig. 9). A comparison of the hydropathy profiles54) of the amino acid sequences of RNase Po1 and RNase T1 shows that RNase Po1 is more hydrophobic between residues 50–60 compared to RNase T1, because two tyrosine residues (Y56, Y57) and one tryptophan residue (W59) of RNase T1 are changed to three phenylalanine residues (F52, F53, F55) in RNsase Po1. In that region, there is Glu 54 (RNase Po1) or Glu58 (RNase T1), part of the catalytic site in the β4-strand (Figs. 4, 10). The molecular surface hydrophobicities of RNase Po1 and RNase T1 confirmed that the hydrophobic region of RNase Po1 is larger than that of RNase T1 in the catalytic region, whereas the hydrophobicities of the other regions are similar (Fig. 11). Hence, the catalytic region of RNase Po1 should be more stable than that of RNase T1 by forming a wider hydrophobic region. We also compared the electrostatic potentials of the molecular surfaces of RNase Po1 and RNase T1.55) This showed that the surface of RNase Po1 is positively charged, but that the surface of RNase T1 is negative, especially behind the catalytic site (Fig. 9). RNase Po1 has eight arginine residues, much more than RNase T1 (one residue). Five of the eight arginine residues of RNase Po1 are located at the molecular surface and make it positive. Since cytotoxic RNases attack intracellular RNA, these RNases must be internalized.41) Johnson et al.56) have suggested that a family of cationic onconases, the RNase A family from Rana pipiens, which exhibit strong anti-tumor activity, bind to the plasma membrane electrostatically, and that enzyme binding is therefore critical to their anti-cancer properties. We compared RNase Po1 to Onconase in detail. The molecular weight of Onconase is 12 kDa and the optimum pH toward RNA is pH 7.5, the same as RNase Po1. Therefore, the base specificity of Onconase is pyrimidine base specific, whereas RNase Po1 is guanine specific. The RNase activity of Onconase toward RNA is much lower than that of RNase Po1. The catalytic residues of Onconase consist of two histidines and one lysine residue, and the sequence identity between Onconase and RNase Po1 is very low.32) Onconase is a cationic protein that contains a total of 15 positively charged residues (three arginines and twelve lysines), much more than RNase Po1 (eight arginines). The surface of Onconase has three high-density positively charged regions (patches) aside from the active site. These patches are very important in the liquid bilayer translocation step that is required for the cytotoxicity of Onconase.57) The three dimensional structure of Onconase39) is completely different from that of RNase Po1; however, the electrostatic potentials of their molecular surfaces are similar in having positively charged regions. This suggests that the positively charged regions of RNase Po1 might be important for binding to the plasma membrane of tumor cells electrostatically.
The surface potential was calculated and displayed using the PyMOL ABPS tool.55) The central figures are viewed from the same direction as Fig. 3. The lower figures are rotated horizontally by 180° by the center. The circles enclose the active site. (a), (b) Electrostatic potentials of the RNase Po1 and RNase T1 molecular surfaces, respectively. The molecules are shown as ribbons and surface models. Negatively charged regions are shown in red, and positively charged regions are shown in blue. RNase T1 (a) is pink and RNase Po1 (b) is blue. (c) Conserved regions between the RNase Po1 and RNase T1 molecular surfaces. RNase Po1 is shown as ribbon and surface models. The conserved regions are colored orange and the others colored white.
The hydropathy profiles were calculated by Expasy Tools (http://www.expasy.ch/tools/protscale html). Glu 54 (RNase Po1) and Glu58 (RNase T1) residues are denoted by arrows.
The figure was drawn with PyMOL (www.pymolwiki.org/index.php/Color_h). RNase Po1 and RNase T1 are shown as surface models. The figures are viewed from the same direction as Fig. 5. The hydrophobic regions are colored gray and the deepness of the dark color shows hydrophobic strength. 2′GMP is shown as sticks.
α-Sarcin from Aspergillus giganteus has been reported to be a ribotoxin and contains the same active site of the RNase T1 family.28) α-Sarcin is composed of a central antiparallel β-sheet packed against an α-helix (smaller than RNase Po1) and has a conserved active site located on the other side, similar to RNase Po1.30) However, the sequence identity between α-sarcin and RNase Po1 is low aside from the active site, and α-sarcin has some insertion sequences with positive charges. One of them is a β-hairpin structure at the N-terminus (Leu7–Arg22), which suggests its involvement in interactions with the cell membrane, while the others are unstructured loops that interact with the ribosomes58) (Fig. 12). RNase Po1 does not have such structures. Thus, RNase Po1 likely binds to the plasma membranes of tumor cells by means different from that of α-sarcin.
Po1: Pleurotus ostreatous RNase,1) T1: Aspergillus oryzae RNase,14) α-Sarcin: Aspergillus giganteus RNase.27) * Catalytic site. Disulfide bonds of RNases are shown connected by bold lines. The numbers at the top of the matrix are those of RNase Po1 and the bottom aero those of α-sarcin. The β-hairpin structure of α-sarcin is in italics.
We earlier reported that RNase Po1 has higher thermal stability than RNase T1.1,26) The optimum temperature for catalysis was measured using RNA as a substrate at 20–80°C for 20 min. The optimum temperature for catalysis by RNase Po1 (70°C) was higher than that for RNase T1 (50°C). The circular dichroism (CD) spectrum of RNase Po1 at [θ]210 nm, which reflects the peptide backbone conformation of the protein, was maintained up to 60°C, then decreased very rapidly with increasing temperature. In contrast, the [θ]210 nm value of RNase T1 decreased in a biphasic manner, with breaks at 40°C and about 60°C. The first sharp decrease around 40°C corresponds to a decrease in the enzymatic activity of RNase T1. We transfected RNase Po1 and RNase T1 into HL-60 cells. We found that RNase Po1 caused greater inhibition of cell proliferation than RNase T1 (Fig. 8). The optimal pH of both enzymes was 7.5, and the specific activity of RNase Po1 for RNA was slightly higher than that of RNase T1; nevertheless, RNase T1 might have inferior stability in cells. Proteolysis experiments showed that RNase Po1 is more resistant to chymotrypsin and thermolysin than RNase T1 (Fig. 7). This suggests that RNase Po1 may be more stable, even in human tumor cells. RNase Po1 contains six cysteine residues, two more than RNase T1. We investigated the disulfide bond combinations of these residues. One disulfide bond of RNase Po1 (Cys9–Cys99) is conserved in all known RNase T1 family enzymes from fungi. The other two disulfide bonds of RNase Po1 (Cys7–Cys84 and Cys48–Cys82) do not have corresponding bonds in RNase T1. The Cys7–Cys84 bond in RNase Pol connects the β1 to the β6 strand, and is located parallel to the Cys9–Cys99 bond, thus making the C-terminus and N-terminus of RNase Po1 more rigid. The Cys48–Cys82 bond connects the β6 strand to the loop containing the recognition site for the guanine base (Asn39, Asn40, Phe41, and Glu42).4) Next to the β6-strand is the His87 residue which is part of the catalytic site. Therefore, RNase Po1 is a more stable enzyme because the catalytic and binding sites (which recognize the guanine base) are stabilized by the Cys48–Cys82 bond in spite of the loop regions. Whereas, the disulfide bond Cys2–Cys10 which connects the N-terminus to the β2-strand in RNase T1 is absent in RNase Po1. The Cys2–Cys10 bond is located far from the catalytic site, and hence it might have little influence on protein stability. Therefore, RNase T1 is not stable as RNase Po1. A non-toxic (to tumor cells) RNase A family member and a toxic family member (Onconase) have been reported. Onconase has four pairs of disulfide bonds, one more than RNase A. Onconase is extremely stable due to the extra disulfide bond at its C-terminus (containing the active site), which contributes to both its stability and toxicity.40) This general hypothesis is supported by the fact that α-sarcin contains two disulfide bonds, one of which is common in all known RNase T1 family enzymes, and the other of which is superimposable with the Cys48–Cys82 bond of RNase Po1 (Fig. 11). We found that RNase Po1 is likely internalized into cells at a much greater rate than RNase T1 by comparing the RNase activities in tumor cells, and that RNase Po1 is probably more stable in the cells than RNase T1 given the results of our proteolysis and cell transfection experiments. These properties might be responsible for the inhibition of cell proliferation of RNase Po1.
In conclusion, we investigated the X-ray crystallographic structure of RNase Po1, which has anti-tumor activity, and found that it is more stable than RNase T1 because of an additional disulfide bond. Furthermore, RNase Po1 can bind to the plasma membrane of tumor cells because its surface is positive, in contrast with RNase T1, which is negatively charged. These differences might contribute to the anti-tumor activity of RNase Po1 towards human cancer cells. To construct RNase Po1 variants with stronger anti-tumor activity, we will make RNase Po1 recombinants utilizing the information gathered from this investigation. Further investigations into the relationship between the structure and anti-tumor activity of RNase Po1 may lead to the development of new anti-cancer drugs.
We thank Professor Hiroshi Iijima for helpful discussions. We also thank Dr. Ryuichi Kato and Ms. Taeko Sasaki for the crystallization of RNase Po1. This work was supported by the Platform for Drug Discovery, Informatics, and Structural Life Science from the Ministry of Education, Culture, Sports, Science and Technology of Japan.