Biosci. Biotechnol. Biochem., 67 (4), 698–703, 2003

Ribonuclease Inhibitors in Malus x domestica (Common Apple): Isolation and Partial Characterization

Takao KOSUGE,1 Mamoru ISEMURA,2 Yoshiaki TAKAHASHI,3 Sumiko ODANI,4 and Shoji ODANI1,†

1Department of Biology, Faculty of Science, Niigata University, Ikarashi, Niigata 950-2181, Japan 2Department of Cellular Biochemistry, School of Food and Nutritional Sciences, University of Shizuoka, Tanida, Shizuoka 422-8526, Japan 3Department of Medical Technology, School of Health Science, Faculty of Medicine, Niigata University, Asahimachi, Niigata 951-8518, Japan 4Department of Home Economics, Faculty of Education and Human Science, Niigata University, Ikarashi, Niigata 950-2181, Japan

Received August 22, 2002; Accepted January 6, 2003

A inhibitory activity was detected in the sion, and carbohydrate- and ATP-binding.1) Self- fruits of common apple, Malus x domestica,cv.Fuji, incompatibility in the ‰owering plants is also and puriˆed by a‹nity chromatography on ribonuclease mediated by prevention of self-pollination by S- A-Sepharose. It inhibited hydrolysis of cyclic-2?:3?- ribonuclease, which is a homologue of ribonuclease CMP by bovine A with an T2.2) Therefore, a collective name of RISBASES apparent inhibition constant of about 5×10－8 M. ( with Special Biological Actions) was 1) Matrix-assisted laser desorptionWionization time-of- proposed for these RNase homologues. These ‰ight mass spectrometry of the puriˆed gave two diverse activities must be under precise control peaks corresponding to the mass numbers of 55,658 and responding to individual physiological conditions, 62,839, while three bands of 43-, 34-, and 21-kDa were but their mechanisms seem not fully understood. It detected by SDS-PAGE. These results suggested that may be possible that protein inhibitors of RNase are the inhibitor preparation was a mixture of two regulatory factors. RNase inhibitors, either synthetic comprised of 43- and 21-kDa subunits or of 34- and or natural, have been intensively sought after for 21-kDa subunits. Attempts to separate these two therapeutic purposes for , because angiogene- proteins were unsuccessful. Amino acid composition sis is frequently promoted by a pancreatic RNase A and N-terminal amino acid sequence of these subunits homologue, , and inhibition of angioge- were also identiˆed and N-terminal sequences showed nin is expected to result in suppression of growth and some similarity to that of cottonseed storage globulin. metastasis of solid tumors.3–5) RNase inhibitors are of The signiˆcance of the presence of ribonuclease inhibi- another practical use to protect RNA molecules from tors in apple fruits is not clear, but it might allow some unwanted degradation by RNases during prepara- speculation about their possible involvement in the tion, and a recombinant human placental RNase control of the self-incompatibility ribonuclease of inhibitor is commercially available. In contrast to a Rosaceae plants. vast number of proteinase- and -inhibitors, reports on protein inhibitors of RNase are quite Key words: Malus x domestica; ribonuclease inhibi- rare.6–8) We searched for inhibitors of RNase in some tor; amino acid sequence; seed storage plants and fungi and found a substantial inhibitory protein; protein puriˆcation activity in the fruits of apple (Malus x domestica). In this paper, we describe puriˆcation and partial Riboncleases (RNases), which have long been characterization of apple RNase inhibitors. regarded as mere digestive of animal and microorganisms, have a wide range of hitherto unexpected biological activities such as angiogenesis, neurotoxicity, aspermatogenicity, immuno-suppres-

† To whom correspondence should be addressed. Fax: ＋81-25-262-6174; E-mail: sodani＠bio.sc.niigata-u.ac.jp Abbreviations: MALDI, matrix-assisted laser desorptionWionization; PBS, phosphate buŠered saline; PVDF, poly(vinylidene di‰uoride); RNase, ribonuclease; TOF-MS, time-of-‰ight mass spectrometry Apple Fruit Ribonuclease Inhibitors 699 Materials and Methods was placed and incubated for 30 min at 379C. The solution was removed and the tube was washed four Materials. Fruits of M. x domestica cv. Fuji, other times with 200 ml of PBS. Reproducibility of ˆxation plant materials, and mushrooms were obtained from of RNase to the tubes was conˆrmed by the activity a local market. Bovine pancreatic RNase A (EC measurement. To this was added 100 mlofasample 3.1.27.5) and cytidine 2?:3?-cyclic monophosphate or PBS and the tube was incubated for 30 min at were obtained from Sigma-Aldrich Co. (St. Louis) . 49C, washed four times with PBS, and further incu- Sepharose 4B and cyanogen bromide-activated batedwith100mlofyeastRNA(0.43mgWml in PBS) Sepharose were products of Amersham-Pharmacia for 15 min at 379C. Then 400 mlof95z ethanol

BioTech (Uppsala). Ribonucleic acid (yeast) and sol- containing 10 mM MgCl2 was added and the mixture vents for HPLC were purchased from Wako Pure was centrifuged to precipitate large RNA molecules. Chemical Industries (Osaka). Reagents for protein The supernatant (400 ml) was mixed with 3.0 ml of sequencing were from Applied Biosystems Japan 66z ethanol and absorbance at 260 nm was meas- (Tokyo). ured. Suitability of the method for measuring RNase inhibitory activity was veriˆed by a commercial Isolation of RNase Inhibitor by A‹nity Chro- placental RNase inhibitor (Rnasin, Promega KK, matography. Whole mature apple fruits (500 g) were Tokyo). RNase activity was also measured by the homogenized in a blender with 500 ml of 50 mM Tris- absorbance change at 286 nm using 2?:3?-cyclic CMP HCl, pH 8.0, containing 0.1 M KCl, 10 mM EDTA, as a according to the procedure of Black- 5z glycerol, 4z soluble poly(vinylpyrrolidone), burn et al.6) A Hitachi 320 recording spectrophotom- 10z dimethylsulfoxide, 5 mM dithiothreitol, 10 mM eter was used. Concentrations of the and 6-amino-n-caproic acid, and 1 mM phenylmethylsul- substrate were usually 0.073 mM and 1.0 mM, fonyl ‰uoride. The extract was centrifuged at 105,000 respectively, in 0.1 M Tris-acetate (pH 6.5). RNase- ×g for 1 h. The centrifuged supernatant was ˆrst put inhibiting activity was assessed by the diŠerence in on a column of Sepharose (2.5×10 cm) to remove absorbance change in the presence and absence of an proteins having a‹nity for the agarose matrix, and inhibitor sample. Mode of the inhibition was exam- then to an RNase A-Sepharose column (1.5×20 cm), ined by a Lineweaver-Burk plot of the results, and which had been prepared by coupling 25 mg of the inhibition constant was roughly estimated by a RNase to 20 ml of CNBr-activated Sepharose accord- Dixon plot (i.e. plotting 1W[against [I ]atdiŠerent ing to the manufacturer's instruction. Both columns ˆxed substrate concentrations [S]).11) These plots were equilibrated with 50 mM Tris-HCl, pH 7.5, were analyzed by the linear regression method using 0.3 M KCl, and 1 mM EDTA. The RNase A- the PRISM software (Graph Pad Software, Inc., San Sepharose column was extensively washed with the Diego, U.S.A.). same buŠer until absorbance at 280 nm of the eluent became negligible. The bound protein was eluted AminoAcidAnalysis.Samples (1–3 nmol) blotted with 45 mM acetic acid, and then with 4 M urea. onto PVDF membrane were hydrolyzed in sealed The eluted fractions were dialyzed against 100 tubes in vacuo with 5.7 M HCl, 4z (vWv) thioglycolic volumes of 45 mM sodium phosphate buŠer, pH 6.5. acid for 22 h at 1109C. Amino acid compositions All operations were done at 49C. were analyzed on a Hitachi Model 835 amino acid analyzer. Polyacrylamide Gel Electrophoresis. Protein sam- ples were electrophoresed in the presence and absence Amino Acid Sequencing. Amino acid sequences of of the denaturant (sodium dodecyl sulfate, SDS).9,10) proteins (about 200 pmol) were analyzed on an A semilograrithmic plot of molecular mass versus Applied Biosystems 476A gas-phase sequencer after relative mobility was generated with marker proteins blotting onto the PVDF membrane. andusedtoestimatethemolecularmassesofsample proteins. Comparison of Amino Acid Sequence. Amino acid sequences were compared using the data bases at the Measurement of RNase Inhibitory Activity. DNA Information and Stock Center, National Screening of RNase inhibitor activity in the crude Institute of Agrobiological Resources (Tsukuba) by tissue extracts was done by measuring RNase activity the program FASTA.12) using a solid phase method, because the presence of large amounts of UV-absorbing substances in the Mass spectrometry. Protein mass analysis was extract interfered with the spectrophotometric assay. done by matrix-assisted laser-desorptionWionization To a polystyrene round-bottom tube (12×75 mm, time-of-‰ight mass spectrometry using a Shimadzu- Falcon 2008 RIA tube, Becton Dickinson & Co., Kratos instrument (KOMPACT MALDI II) at an Lincoln Park, U.S.A) 100 ml of bovine pancreatic accelerating voltage of 20 keV. Sinapic acid was used RNase (1 mgWml) in phosphate buŠered saline (PBS) as the matrix. 700 T. KOSUGE et al.

Fig. 1. A‹nity Chromatography of Apple Extract on RNase A- Sepharose. High-speed supernatant of the extract was put on a column (2.5×7 cm) of RNase A-Sepharose equilibrated with 50 mM Tris-HCl, pH 7.5, 0.3 M KCl, and 1 mM EDTA. The column was extensively washed with the same buŠer. The bound protein was eluted with 45 mM aceticacid(arrowA),andthenwith4M urea Fig. 2. SDS-Polyacrylamide Gel Electrophoresis of Apple RNase (arrow B). Two-ml fractions were collected. Proteins were de- Inhibitor Preparation. tected by their absorbance at 280 nm. Fractions indicated by (1) Fraction eluted with 45 mM acetic acid. (2) Fraction eluted bars were pooled separetely and dialyzed against 50 volumes of with 4 M urea. (3) Marker proteins. Molecular masses are indi- 45 mM sodium phosphate buŠer, pH 6.5. All operations were cated on the right margin. A semilograrithmic plot of molecular done at 49C. mass versus relative mobility was generated with marker pro- teins and used to estimate molecular mass of sample proteins.

Results and Discussion SDS, three major protein bands corresponding to 43, Screening for RNase Inhibitory Activity 34, and 21 kDa were obtained (Fig. 2). A minor band First, a number of vegetables and edible of 35 kDa was also seen. In the non-denaturing gel- mushrooms were examined for RNase inhibitory electrophoresis system, the protein migrated as a activity by the solid-phase method. Extracts from single band on the separating gel of pH 8.8 (ˆgure fruiting bodies of mushrooms were largely inhibitory not shown). Because this gel system can not identify active, but the substances bound to RNase A- basic proteins that would not enter into the separat- Sepharose appeared to be some low-molecular-mass ing gel, another gel system was used, where the pH of compounds rather than protein as judged from SDS- the separating gel is 4.3.9) However, no protein band PAGE. Unexpectedly, the extract of apple fruits was observed in this gel system (i.e., all the protein showed substantial inhibition of RNase A activity, components had migrated into the anode buŠer), which appeared to be due to a proteinaceous sub- indicating that the fraction eluted with acetic acid did stance, and we decided to further characterize the not contain any basic protein components. RNase A-inhibitory activity in the apple fruits. Figure 1 shows the elution proˆle of the 105,000× Inhibitory Activity g supernatant of apple fruit extract from the RNase- Increasing amounts of the inhibitor preparation Sepharose 4B column. After extensive washing with were added to a ˆxed amount of RNase A, and Tris-HCl buŠer, a protein peak was eluted with remaining RNase activity was measured at four 45 mM acetic acid, and an additional small peak was diŠerent concentrations of cyclic-2?:3?-CMP. The also eluted by 4 M urea solution. The urea-eluted data was analyzed by a double reciprocal plot using fraction showed the same band pattern on SDS- the PRISM software. From Fig. 3(a), it may be seen PAGE with that of the fraction eluted with acetic that the inhibition appears to be competitive. The acid but rather low RNase inhibitory activity, possi- result was further interpreted by a Dixon plot11) as bly due to partial denaturation. This fraction was shown in Fig. 3(b). The x-axis intercepts of the lines supposed to be a tailing region of the ˆrst peak being generated by linear regression fell between －6and forced to elute by urea and not analyzed further. The －4×10－8 M. Therefore an approximate inhibition acetic acid-eluted fraction was dialyzed against constant may be roughly estimated to be 5×10－8 M. 45 mM phosphate buŠer, and analyzed for protein The value is much larger than dissociation constants content. Upon gel electrophoresis in the presence of of the extremely tight complexes between Apple Fruit Ribonuclease Inhibitors 701

Fig. 3. Inhibitory Activity of Apple RNase Inhibitor Preparation. (a) A Lineweaver-Burk plot for hydrolysis of 2?:3?-cyclic CMP by RNase A in the presence of various amounts of RNase inhibitor 6) －3 preparation in 0.1 M Tris-acetate (pH 6.5) at 259C. Reaction rate ([) is expressed by the absorbance change at 286 nm per min (×10 ). Inhibitor concentrations were 0.070 mM (solid circles), 0.018 mM (open circles), 0.009 mM (open squares), and 0 mM (solid squares). (b) 11) Dixon plot of the results. Substrate concentrations were 0.1 mM (solid squares), 0.05 mM (open squares), 0.025 mM (solid circles), and 0.0125 mM (open circles). Lines were generated by linear-regression using the PRISM software (GraphPad Software, Inc.).

Table 1. Amino Acid Composition of Apple RNase Inhibitor mammalian RNase inhibitor and pancreatic RNase Subunits －14 13) (Kd＝4.4×10 M), and between microbial RNase Subunits (43k, 34k and 21k) were separated by SDS-PAGE and inhibitor (barstar) and RNase (barnase) from Bacil- blotted onto a PVDF membrane. Protein bands were excised from －14 14) lus amyloliquefaciens (Kd＝1.3×10 M). Rather membrane and hydrolyzed with 5.7 M HCl for 22 h at 1109C. weak inhibition of pancreatic RNase A by the apple Amino acid compositions are shown as residues per mole protein rounded to the nearest integers. inhibitor preparation suggests that the true target enzyme of the RNase inhibitor may be some other Amino acid 43k 34k 21k RNases. Hydrolysis of yeast RNA by pancreatic Aspartic acid 33 34 26 RNase A was also inhibited by the inhibitor prepara- Threonine867 tion (data not shown). Serine 17 17 16 Glutamic acid 122 87 27 Subunit Structure of Apple RNase Inhibitor Proline 12 13 7 Three major bands of 43-, 34-, and 21-kDa were Glycine474013 Alanine191516 detected by SDS-PAGE (Fig. 2) and they were blot- Half-cystine 4 3 1 ted onto a PVDF membrane for further analysis. The 16 14 13 amino acid compositions of these 3 bands are given 3 3 2 in Table 1. It is seen that 43- and 34-kDa bands have 13 10 12 similar compositions rich in glutamic acid and Leucine222117 Tyrosine556 glycine, while the 21-kDa band is somewhat diŠerent 13 15 7 from them in having a higher content of aspartic Lysine632 acid. The three bands were examined for their N- Histidine642 terminal amino acid sequences and the results are Arginine 34 24 15 summarized in Fig. 4. Although amino acids in some Total 380 314 189 positions were diŠerent between 43- and 34-kDa proteins, they were very similar. A minor 35-kDa band had the sequence identical with 34-kDa band up Sequences of 43- 34-, and 21-kDa proteins showed to the 20th residues from the N-terminus and its some similarity to the acidic chain (303 residues) or amino acid composition was very similar to that of the basic chain (185 residues) of a cottonseed protein 34-kDa protein (data not shown). It appears that (legumin A) which belongs to the 11S seed storage these two proteins are identical and diŠerence in .15) Theplantstorageproteins molecular mass may be due to glycosylation or often have - and amylase-inhibitory activi- proteolytic processing. Amino acid sequences ties, which suggests multiple functions for storage obtained were searched for similar proteins against proteins including defense mechanisms against pests the databases at the DNA Information and Stock and pathogens.16–19) Center, National Institute of Agrobiological To identify the subunit structure of apple RNase Resources (Tsukuba) using the program FASTA.12) inhibitor, the puriˆed protein was put through MAL- 702 T. KOSUGE et al.

Fig. 4. Comparison of the Primary Structures of Apple RNase Inhibitor Subunits and Cottonseed Globulin (legumin A precursor). N-terminal amino acid sequences of the 43-kDa, 34-kDa and 21-kDa bands of RNase inhibitor preparation were analyzed and a search was done for similar proteins in the data bases at the DNA Information and Stock Center, National Institute of Agrobiological Resources (Tsukuba). Numbering is that of cottonseed legumin A precursor. X denotes an unidentiˆed residue. Sequences were aligned by the program Clustal W.25) Identical residues are shown by asterisks and similar ones by periods. Gaps (－) are introduced for maximal .

however, not clear whether the two complexes are further associated in the native state, though no peaks were seen around 100 kDa in the mass spectro- gram. Therefore, we supposed that the RNase inhibi- tor preparation obtained here was a mixture of at least two closely similar proteins. Attempts to separate them by hydroxyapatite column chro- matography or gel ˆltration were unsuccessful. Signiˆcance of the presence of RNase inhibitors in apple fruit is not readily explainable. Some specula- tion might be allowed about a possible regulatory function of the RNase inhibitors for the self-incom- patibility RNase in the Rosaceae plant. Although recent studies suggested the presence of possible RNase S inhibitors in the pollen grains of Nicotiana alata and Petunia hybrida,22,23) a protein Fig. 5. Matrix-assisted Laser DesorptionWIonization Time-of- (SP11) for S-locus receptor , has been estab- ‰ight Mass Spectrometry of RNase Inhibitor Preparation. lished as a self-incompatibility pollen protein in A Shimadzu-Kratos instrument (KOMPACT MALDI II) was 24) operated at an accelerating voltage of 20 keV using sinapic acid Brassicaceae plants. It might be also possible that as matrix. There were no peaks in the mass range below 30,000. the inhibitor has a protective function against secret- ed RNases of pests and pathogens. Preliminary analysis of tissue distribution of RNase inhibitor DI-TOF-MS. Two peaks corresponding to 55,658 activity within a single mature fruit suggested its and 62,839 were observed (Fig. 5), and no peaks of concentration in the inner parts (ovary) of the fruit, molecularmasslessthan30kDawerepresent.Ithas but not in receptacle or seeds. Further investigation, been shown that multimeric proteins non-covalently especially an inhibition study on S-RNases and other associated are ionizable without dissociation by the RNases, is necessary for functional analysis of the MALDI method and can be detected by MS analy- apple RNase inhibitor. sis.20,21) Consequently, the absence in the mass spectrogram of 21-kDa protein may be attributed its Acknowledgment complex formation with the 43- and 34-kDa proteins. It seems that peaks of 55,658 and 62,839 Da respec- We thank Dr. Tokuji Ikenaka, Professor Emeritus tively correspond to the 43-kDa and 34-kDa proteins, at Osaka University, for constant encouragement, each of which is associated with the 21-kDa protein. and Prof. Saburo Hara (Kyoto Institute of Technolo- This subunit structure is also similar to that of gy) for mass-spectrometric measurement. cottonseed storage protein (Fig. 4), where two subunits are linked by disulˆde bonds.15) It is, Apple Fruit Ribonuclease Inhibitors 703 References barnase with its polypeptide inhibitor barstar studied by . Biochemistry, 32, 5145–5150 1) D'Alessio, G., Di Donato, A., Parente, A., and (1993). Piccoli, R., Seminal RNase: a unique member of the 15) Chlan,C.A.,Pyle,J.B.,Legocki,A.B.,andDure, ribonuclease superfamily. Trends Biochem. Sci., 16, L., Developmental biochemistry of cottonseed 104–106 (1991). embryogenesis and germination. XVIII. cDNA and 2) McClure, B. A., Haring, V., Ebert, P. R., Anderson, amino acid sequences of the members of the storage M.A.,Simpson,R.J.,Sakiyama,F.,andClarke,A. protein families. Plant Mol. Biol., 7, 475-489 (1986). E., Style self-incompatibility gene products of 16) Odani, S., Koide, T., Ono, T., and Ohnishi, K., Nicotiana alata are ribonucleases. Nature, 342, Structural relationship between barley (Hordeum 955–957 (1989). vulgare) inhibitor and castor-bean (Ricinus 3) Fett, J. W., Olson, K. A., and Rybak, S. M., A communis) storage protein. Biochem. J., 213, monoclonal antibody to human angiogenin. Inhibi- 543–545 (1983). tion of ribonucleolytic and angiogenic activities and 17) Shewry,P.R.,Miles,M.J.,andTatham,A.S.,The localization of the antigenic epitope. Biochemistry, prolamin storage proteins of wheat and related 33, 5421–5427 (1994). cereals. Prog.Biophys.Mol.Biol., 61, 37–59 (1994). 4) Olson, K. A., French, T. C., Vallee, B. L., and Fett, 18) Yeh,K.W.,Chen,J.C.,Lin,M.I.,Chen,Y.M., J. W., A monoclonal antibody to human angiogenin and Lin, C. Y., Functional activity of sporamin from suppresses tumor growth in athymic mice. Cancer sweet potato (Ipomoea batatas Lam.): a tuber storage Res., 54, 4576–4579 (1994). protein with trypsin inhibitory activity. Plant Mol. 5) Nobile,V.,Russo,N.,Hu,G.F.,andRiordan,J.F., Biol., 33, 565–570 (1997). Inhibition of human angiogenin by DNA aptamers: 19) Hou, W. C., Chen, Y. C., Chen, H. J., Lin, Y. H., nuclear colocalization of an angiogenin-inhibitor. Yang,L.L.,andLee,M.H.,Antioxidantactivities Biochemistry, 37, 6857–6863 (1998). of , a 33 KDa root storage protein of 6) Blackburn, P., Wilson, G., and Moore, S., sweet potato (Ipomoea batatas (L.) Lam cv. Tainong Ribonuclease inhibitor from human placenta. Puriˆ- 57). J. Agric. Food Chem., 49, 2978–2981 (2001). cation and properties. J. Biol. Chem., 252, 5904–5910 20) Borchers, C., and Tomer, K. B., Characterization of (1977). the noncovalent complex of human immunodeˆcien- 7) Hartley, R. W., and Smeaton, J. R., On the reaction cy virus glycoprotein 120 with its cellular receptor between the extracellular ribonuclease of Bacillus CD4 by matrix-assisted laser desorptionWionization amyloliquefaciens (barnase) and its intracellular inhi- mass spectrometry. Biochemistry, 38, 11734–11740 bitor (barstar). J. Biol. Chem., 248, 5624–5626 (1999). (1973). 21) Nedelkov, D., and Nelson, R. W., Analysis of native 8) Bisbal,C.,Martinand,C.,Silhol,M.,Lebleu,B., proteins from biological ‰uids by biomolecular and Salehzada, T., Cloning and characterization of a interaction analysis mass spectrometry (BIAWMS): RNase L inhibitor. A new component of the inter- exploring the limit of detection, identiˆcation of non- feron-regulated 2-5A pathway. J. Biol. Chem., 270, speciˆc binding and detection of multi-protein 13308–13317 (1995). complexes. Biosens. Bioelectron., 16, 1071–1078 8) Pasloske, B. L., Ribonuclease inhibitors. Methods (2001). Mol. Biol., 160, 105–111 (2001). 22) Golz,J.F.,Oh,H.Y.,Su,V.,Kusaba,M.,and 9) Bollag, D. M., and Edelstein, S. J., Gel electrophore- Newbigin, E., Genetic analysis of Nicotiana pollen- sis under nondenaturing conditions. In ``Protein part mutants is consistent with the presence of an S- Methods'', Wiley-Liss, Inc., New York, pp. 143–160 ribonuclease inhibitor at the S locus. Proc. Natl. (1991). Acad.Sci.U.S.A., 98, 15372–15376 (2001). 10) Bollag, D. M., and Edelstein, S. J., Gel electrophore- 23) Sims, T. L., and Ordanic, M., Identiˆcation of a S- sis under denaturing conditions. In ``Protein ribonuclease-binding protein in Petunia hybrida. Methods'', Wiley-Liss, Inc., New York, pp. 95–142 Plant Mol. Biol., 47, 771–783 (2001). (1991). 24) Takayama, S., Shimosato, H., Shiba, H., Funato,

11) Dixon, M., The graphical determination of Km and M.,Che,F.S.,Watanabe,M.,Iwano,M.,and Ki. Biochem. J., 129, 197–202 (1972). Isogai, A., Direct ligand-receptor complex interaction 12) Pearson, W. R., and Lipman, D. J., Improved tools controls Brassica self-incompatibility. Nature, 413, for biological sequence comparison. Proc. Natl. 534–538 (2001). Acad.Sci.U.S.A., 85, 2444–2448 (1988). 25) Thompson, J. B., Higgins, D. G., and Gibson, T. J., 13) Lee, F. S., Shapiro, R., and Vallee, B. L., Tight- CLUSTAL W: improving the sensitivity of progres- binding inhibition of angiogenin and ribonuclease A sive multiple through sequence by placental ribonuclease inhibitor. Biochemistry, 28, weighting, position-speciˆc gap penalties and weight 225–230 (1989). matrix choice. Nucleic Acids Res., 22, 4673–4680 14) Schreiber, G., and Fersht, A. R., Interaction of (1994).