Anti-(Raf-1) RNA aptamers that inhibit Ras-induced Raf-1 activation Michiko Kimoto 1,2 , Mikako Shirouzu 2,3 , Shin Mizutani 4 , Hiroshi Koide 4 , Yoshito Kaziro 4 , Ichiro Hirao 5 and Shigeyuki Yokoyama 1,2,3,5 1 Department of Biophysics and Biochemistry, Graduate School of Science, The University of Tokyo, Bunkyo-ku, Tokyo, Japan; 2 Cellular Signaling Laboratory, RIKEN Harima Institute, Mikazuki-cho, Sayo, Hyogo, Japan; 3 RIKEN Genomic Sciences Center, Tsurumi, Yokohama, Japan; 4 Faculty of Bioscience and Biotechnology, Tokyo Institute of Technology, Yokohama, Japan; 5 Yokoyama CytoLogic Project, ERATO, JST, Wako-shi, Saitama, Japan RNA aptamers with anity for the Ras-binding domain (RBD) of Raf-1 were isolated from a degenerate pool by in vitro selection. These aptamers eciently inhibited the Ras interaction with the Raf-1 RBD, and also inhibited Ras- induced Raf-1 activation in a cell-free system. The RNA aptamer with the most potent inhibitory eect speci®cally inhibited the RasáRaf-1 i nteraction and had no anity for the RBD of the RGL protein, a homolog of the Ral GDP dissociation stimulator. Although t he aptamer w as capable of binding to the B-Raf RBD, the RNA did not inhibit the interaction between Ras and the B-Raf RBD. Enzymatic and chemical probing experiments indicated that the apt- amer was folded i nto a pseudoknot structure, and some loop regions of th e p seudoknot were located at the binding interfacefortheRaf-1RBD. Keywords: Raf-1; Ras; RNA aptamer; in vitro selection. The Raf-1 protein is a cytoplasmic serine/threonine kinase that transmits cellular proliferative and developmental signals from the p lasma membrane to the cytosol and the nucleus. Raf-1 phosphorylates and activates the MAPKK/ MEKs, which in turn pho sphorylate the MAPK/ERKs [1]. Activated MAPK phosphorylates a number o f proteins, including protein kinases, transcription factors, and cyto- skeletal proteins, and these proteins are thought to b e critical for proliferation or differentiation. Accordingly, inappropriate activation of these pathways causes uncon- trolled proliferation o r differentiation of cells, and Raf-1, which integrates the extracellular signals through the pathways, is the cellular homologue of a viral oncogene [2]. The regulation mechanisms involving R af-1 are com- posed of complex signaling networks, where protein± protein interactions play a major role. The activity of Raf-1 is regulated b y t he interaction with the Ras protein on the plasma membrane and/or by Ras-independent mech- anisms [1]. Ras belongs t o a superfamily of proteins, termed the GTPases, which cycle between the GTP- and GDP- bound forms [3], a nd the GTP-bound Ras physically associates with the region comprising amino-acid residues 51±131 of Raf-1 (the Ras-binding domain, RBD) with high af®nity (K d  18 n M ) [4±6]. The RBD a lone is suf®cient for the Ras-mediated translocation of Raf-1 from the cytosol to the plasma membrane, while additional membrane-local- ized events are required for the Raf-1 activation [1]. In addition, th e GTP-bound Ras can bind to some other mammalian proteins, such as the Raf-1 isozymes (A-Raf and B-Raf) [7,8], the guanine nucleotide dissociation stimulators for the Ral protein (RalGDS and RGL) [9,10], and phosphatidylinositol 3-kinase [11]. Recently, it became clear that cross-talk occurs among the individual Ras-signaling p athways [ 12]. For example, Ras regulates the Raf-1 kinase activity indirectly, through the Ras/phosphat- idylinositol 3-kinase signaling pathway [12]. The molecular mechanisms of the cellular signaling involving Raf-1, however, are still not fully understood. The development of s peci®c modulators for the i ndividual interactions involving t he two p roto-oncogene products, Raf-1 and Ras, would a llow us t o understand and to regulate these i nteractions in the intricate cellular signaling pathways. Accordingly, many efforts have been made to generate speci®c inhibitors for the Ras/Raf signaling path- way. The s election o f phage-displayed Fab antibodies directed toward th e GTP-bound Ras yielded a h igh af®nity, conformation-speci®c antibody, and the antibody prevented the binding of Raf-1 kinase to Ras [13]. However, it would prevent a variety of Ras-mediated signaling pathways in addition to the Ras/Raf signaling p athway. Monoclonal antibodies that bind to the Raf-1 kinase domain are also potential inhibitors, and can block the kinase activity [14]. However, they inhibit both the Ras-dependent and -independent Raf-activation mechanisms. On t he other hand, the chemical i nhibitors of Raf, ZM3363 72 and SB203580, inhibit the Raf-1 kinase activity in vitro, but they paradoxically induce a remarkable activation in vivo because they block a negative feedback loop initiated by Raf-1 itself [15,16]. Thus, it is worthwhile to develop a speci®c modu- lator for only the RasáRaf interaction. Correspondence to S. Yokoyama, Department of Biophysics and Biochemistry, Graduate School of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan. Fax: + 81 35841 8057, Tel.: + 81 35841 4413, E-mail: yokoyama@biochem.s.u-tokyo.ac.jp Abbreviations: MAPKK, MAPK kinase; ERK, extracellular signal- regulated kinase; GST, glutathione S-transferase; MAPK, mitogen- activated protein kinase; RBD, Ras-binding domain; MEK, MAPK kinase/ERK kinase; GTPcS, guanosine 5 ¢-O-(3-thiotriphosphate); KN-MAPK, kinase-negative MAPK; DMS, dimethyl sulfate; CMCT, 1-cyclohexyl-3-(2-morpholinoethyl)carbodiimide metho-p-toluene- sulfonate. (Received 4 September 2001, revised 26 November 2001, accepted 27 November 2001) Eur. J. Biochem. 269, 697±704 (2002) Ó FEBS 2002 To develop such a speci®c inhibitor, we used an in vitro selection technique, which allows the rapid generation of RNA or DNA molecules (aptamers) with high af®nity and unique selectivity against target proteins [17,18]. So f ar, aptamers have been shown to inhibit a variety of targets and their functions in vitro and in vivo [19±24]. We previously showed that the Raf-1 RBD could be a target for RNA aptamers that inhibit the RasáRaf interaction [25]. However, the inhibition by the aptamers was not ef®cient, and l arge amounts of the aptamers to Ras were required for the inhibition. Here, we report the isolation and characteriza- tion of newly develop ed RNA aptamers that inhibit t he RasáRaf interaction more ef®ciently than the previous ones, and demonstrate the abilities of these RNA aptamers to inhibit Ras-induced Raf-1 activation in a reconstituted cell- free system. MATERIALS AND METHODS Materials Oligodeoxyribonucleotides were synthesized by standard phosphoramidite chemistry on a DNA synthesizer (model 392, PE Applied Biosystems). A nucleotide labeled with [c- 32 P]ATP was purchased from Amersham Pharmacia Biotech and NEN Life Science Products. RNase V 1 was purchased from Amersham Pharmacia Biotech, and mung bean nuclease, T4 polynucleotide kinase, and reverse transcriptase (RAV-2) were purchased from Takara (Tokyo, Japan). D imethyl sulfate (DMS) and 1-cyclo- hexyl-3-(2-morpholinoethyl)carbodiimide metho-p-toluene- sulfonate (CMCT) were purchased from Nacalai Tesque (Kyoto, Japan). The anti-(FLAG e pitope) Ig (M2) was purchased from Eastman Kodak Co. Protein preparation The Ras-binding domains (amino-acid residues 51±131 of human Raf-1, 149±226 of human B-Raf, and 632±734 of rat RGL) as GST fusion forms (designated as the Raf-1 GST± RBD, the B-Raf GST±RBD, and the RGL GST±RBD, respectively) and the wild-type human Ha-Ras protein were expressed in Escherichia coli and were puri®ed as described previously [25±27]. A recombinant Xenopus kinase-negative MAPK in the GST fusion form (GST±KN-MAPK) and a histidine-tagged Xenopus MAPKK were puri®ed from E. coli as described p reviously [28,29]. In vitro selection of RNA aptamers and sequencing In v itr o selection was carried out using an RNA pool that contained 45 randomized nucleotide positions. The RNA pool (Fig. 1A) was transcribed from the DNA template using an Ampliscribe T7 in vitro transcription kit (Epicenter Technologies, Madison, WI, USA). The RNA was dis- solved in binding buffer (NaCl/P i containing 5 m M MgCl 2 , buffer A), heated at 75 °C for 3 m in, and cooled to room temperature. To exclude the ®lter-binding RNA s pecies from the pool, the RNA was passed through a 0.45-lm HAWP ®lter ( Millipore, New Bedford, MA, USA) two or three times before the incubation with the target protein. The RNA was m ixed with the Raf-1 GST±RBD and was incubated for 1 h at 37 °C. Table 1 summarizes the conditions used in each round of the selection. After the incubation, the solution was gently vacuum-®ltered over the HAWP ®lter and was washed three times with 300±500 lL of buffer A. The RNA was eluted and was precipitated with isopropyl alcohol, as described previously [30]. In r ounds 6±9, an additional ®ltration step was carried out on the eluted R NA pool to further e xclude ®lter-binding species. In a ddition, we carried out the GST-counter selection to eliminate the RNAs that bind GST. Prior to the i ncubation with the target protein (in rounds 6±9) and after the elution (in rounds 8 and 9), the RNA pool was incubated with 1 00 pmol of GST and a matrix (glutathi- one±Sepharose 4B) for 30 min at 37 °C, and t he superna- tant of the solution was ®ltered. The c ollected RNA was reverse transcribed and was ampli®ed by PCR u sing primer 39.45 (5¢-GGTAATACGACTCACTATAGGGAGTGG AGGAATTCATCG) a nd primer 24.45 (5¢-GCAGAAGC TTGCTGTCGCTAAGGC). The RNA pool for the next round of selection was prepared from the ampli®ed cDNA. The PCR products of the ninth round pool were clon ed in the TA cloning vector using the TOPO TA Cloning Kit Dual (Invitrogen). Plasmid DNAs were isolated and were sequenced using the dRhodamine Terminator Cycle Seq- uencing Ready Reaction Kit (PE Applied Biosystems) on a DNA sequencer (model 377, Applied Biosystems). Fig. 1. The sequence of the RNA pool used for in vitro sele ction (A) and the sequences of the selected RNA aptamers from the ninth round pool (B). (A) The randomized region (N 45 ) is ¯anked by the 5¢ and 3¢ constant regions. The PCR prime r annealing sites within the constant regions are underlined. (B) The names of the r epresentative aptamers are shown on the left. As compared to RNA 9A or 9B, the mutated residues in other clones are shown in boldface. The number in parentheses indica tes the number of the same clone. 698 M. Kimoto et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Filter-binding assays RNA pools and individual RNAs were labeled with T4 polynucleotide k inase a nd [c- 32 P]ATP after 5 ¢ dephosphory- lation of the T7 transcripts. The 5 ¢-labeled RNA p ool (5 n M ®nal concentration) was mixed with the Raf-1 GST±RBD or GST (0.3 l M ®nal concentration) in 100 lL of buffer A for 1 h at 37 °C. The solution was then ®ltered and was washed with 200 lL of buffer A. F or competitive binding, the labeled RNAs were prepared b y transcription with [a- 32 P]UTP (Amersham Pharmacia Biotech). For t he determination of t he dissociation constant (K d ), t he 5¢-labeled RNA (2 n M ®nal concentration) was incubated with various concentrations (12.3±1230 n M )ofproteinsin 60 lL of buffer A for 30 min at 37 °C. The ®lter was exposed to a PhosphorImager plate, and the amount of the retained radioactivity was measured with a B io-imaging analyzer (Fuji BAS 2500). The fraction of bound RNA was calculated by comparing t he radioactivity o n the ®lter with the total radioactivity of the input RNA. The K d values were determined with the program KALEIDAGRAPH (Albelbeck Software, Reading, PA, USA) by using the general curve ®t for y  M 0 ´ M 1 /(M 0 + M 2 ), where y is the fraction of the R NA binding, M 0 is the protein concentration, M 1 is the binding capacity of the aptamer, and M 2 is K d . Competition experiments TheRaf-1,B-Raf,orRGLGST±RBD(20pmol)in 150 lL of buffer A containing 0.05% Triton X-100, was mixed with 1 0 lL of the matrix (glutathione±Sepharose 4B) suspended in NaCl/P i (50% slurry). The mixture was incubated at 4 °C for 30 min. After brief centrifugation, the s upernatant w as discarded. Then, the matrix was mixedwithRas(20pmol),whichhadbeencomplexed with GTPcS as d escribe d previously [31], and was incubated with the renatured RNAs (0±200 pmol) in buffer A (150 lL) at 4 °C for 30 min. After the incuba- tion, the matrix was washed with 500 lL of buffer B (20 m M Tris/HCl buffer (pH 7.5) containing 150 m M NaCl and 5 m M MgCl 2 ) th ree times. The proteins w ere eluted from the matrix b y denaturation with Laemmli's buffer and wer e fractionated by 15% SDS/PAGE. The immunoblots were probed with t he anti-Ras Ig (RAS004) [32] and were visualized by using the ECL immunodetec- tion system (Amersham Pharmacia Biotech). Inhibition experiments The Raf fraction from HEK293 cells expressing RafFH (human Raf-1 with the FLAG epitope and six histidine residues at its C-terminus) and the membrane fraction from Sf9 cells expressing H-Ras [G12V] were prepared as described previously [33,34]. The Raf fraction was incubated with the membrane fraction in the presence of RNAs (2 or 10 l M ®nal concentration) at 16 °C for 30 min. The reaction was terminated by the addition of Triton X-100 to a ®nal concentration of 0.5%. Solubilized fractions were subjected to immunoprecipitation with an anti-FLAG Ig, a rabbit anti-(mouse IgG) Ig, and protein A±Sepharose (Amersham Pharmacia Biotech). The RafFH kinase activity was determined a s described previously [26,33]. The immu- noprecipitates were incubated with the recombinant hist- idine-tagged MAPKK and GST±KN-MAPK in the presence of [c- 32 P] ATP a t 30 °C for 20 min. After t he incubation, the reaction was stopped b y adding Laemmli's buffer and boiling. The samples were fractionated by S DS/ PAGE, and the phosphorylation of GST±KN-MAPK was measured with a Bio-imaging analyzer (Fuji BAS 2500). Enzymatic structure probing The procedures for the enzymatic digestion experiments were modi®cations of that previously reported [30]. The 5¢-labeled RNA 9A (4 pmol) was renatured in 10 lLof buffer A containing 10 m M MgCl 2 , followed by mung bean nuclease digestion (8 U) for 5 min at 37 °CandRNaseV 1 digestion (0.007 U) f or 1 min at 37 °C. The d igested products were analyzed on a 15% polyacrylamide gel with 7 M urea at 50 °C. Chemical structure probing and footprinting experiments DMS, which modi®es the base-pairing face of C (at N-3) and A (at N -1), and CMCT, which m odi®es the base- pairing face of U (at N-3), were used for chemical Table 1. Summary of in vitro selection experiments. The ®lter-binding assay w as carried out at 37 °C for 1 h (Raf-1 RBD/RNA p oo l  300: 5 n M ) as described in the Materials and methods. Round No. Pool RNA (l M ) Raf-1 RBD (l M ) Total volume (lL) Filtration (pre-post) a GST counter (pre-post) b Filter-binding assay (%) Raf-1 RBD GST Filter 0 < 0.1 < 0.1 0.5 1 1.0 3.0 600 3-0 0-0 2 1.0 1.0 300 3-0 0-0 3 1.0 1.0 300 3-0 0-0 4 0.5 0.5 200 3-0 0-0 5 0.3 0.3 200 3-0 0-0 34 1.2 1.3 6 0.3 0.3 200 3-2 1-0 7 0.15 0.15 200 3-2 1-0 54 0.6 0.7 8 0.15 0.15 200 3-2 1-1 9 0.05 0.15 200 3-2 1-1 80 0.3 0.1 a Number of pre- and post-®ltrations. b Number of pre- and post-GST-counter selections. Ó FEBS 2002 Anti-(Raf-1) RNA aptamers (Eur. J. Biochem. 269) 699 modi®cation. The renatured RNA 9A ( 12 pmol) in 30 lL of buffer A, was mixed with 3 lL o f the Raf-1 GST±RBD (180 pmol) and was incubated for 20 min at 37 °C. The RNAáRBD complex solution was mixed with 30 lLof CMCT (42 mg ámL )1 inbufferA)orwith1lLofDMS (diluted 10-fold in ethanol) and was incubated for 30 min (for CMCT) or for 15 min (for DMS) at 37 °C. The reaction was stopped an d th e c omplex was precipitated w ith ethanol. The modi®ed RNAs were extracted with phenol and were precipitated with ethanol. The collected RNAs were dissolved in 24 lL of distilled water, and were reverse transcribed with the 5¢-labeled primer 24.45 a nd reverse transcriptase (RAV-2). The reverse transcription products were analyzed on a 10% polyacrylamide gel with 7 M urea. RESULTS AND DISCUSSION Isolation of RNA aptamers to the Raf-1 RBD For selective inhibition of the Raf-1 interaction with Ras, we utilized only the Ras-binding domain (GST±RBD) of Raf-1 instead of t he whole protein as a selection target. W e used an RNA pool with approximately 10 14 different molecules, and the i nitial (round 0) pool included about three copies of each RNA molecule, with a length of 100 nucleotides containing a r egion of 45 randomized nucleotides ( N 45 ) ¯ anked by constant regions (Fig. 1A). The sequences within the 5¢ and 3¢ constant regions next to the randomized region were designed to form a stem structu re, to prevent undesirable interactions between the constant and the randomized regions. The complementary sequences were not used for the annealing sites of the PCR primers, so they would not decrease the PCR ef®ciency. We cho se a shorter randomized region than that (N 60 )usedinourpreviousselection[25],to increase the population of RNA species in the pool that could be easily handled for more s tringent selection. In addition, in the previous selection, a nitrocellulose ®lter was better than the glutathione±Sepharose 4B matrix for ef®cient immobilization and separation of the protein± RNA complexes [25], and thus we used the nitrocellulose ®lter throughout the selection. Althou gh some ®lter- or GST-binding species were accumulated during the middle rounds of the selection, post®ltration of the selected RNAs and GST-counter selection (see Materials and methods for details) successfully removed these species. After nine rounds of selection, the binding ef®ciency of the pool to the GST±RBD had increased from < 0.1% to 80% (Table 1). We isolated and s equenced 37 clones from round 9. Alignment of the sequences revealed that the pool had been winnowed to two RNA molecules, 9A and 9B (Fig. 1B), and these sequences were completely different from those isolated by the previous selection. By using the ®lter- binding assay, the v alues of t he dissociation constant (K d ) of RNAs 9A and 9B for GST±RBD binding were determined to be 152  23 a nd 361  46 n M , re spectively (Fig. 2). On the other hand, for RNA 21.01, which was previously isolated by in vitro selection [25], it was much more dif®cult to determine its exact K d value because the binding capacity (M 1 )ofRNA21.01wasaslowas 50%. Then, we directly compared the competitive binding abilities of these RNAs to GST±RBD. Equimolar amounts of the 32 P-labeled RNA 9A, 9B or 21.01 and the protein (0.2 l M ®nal concentrations, respectively) were incubated in the presence of an equal amount or twofold or threefold molar excess of the nonlabeled RNA 9A, 9B or 21.01, and the binding of the labeled RNA was examined by using the ®lter-binding assay. The addition of the nonlabeled RNA 9A reduced the binding of the labeled RNAs 9A itself and 9B, but the addition of the nonlabeled RNA 21.01 hardly affected their binding (data not shown). Furthermore, the binding of the labeled RNA 21.01 was reduced more signi®cantly in the presence of the nonlabeled RNA 9A or 9B than with 21.01 itself (data not shown). These results demonstrated that the binding abilities of RNAs 9A and 9B were higher than that of RNA 21.01. Thus, the present selection provided more ef®cient aptamers than the previ- ous selection. Competition of RNA aptamers with Ras for binding to the Raf-1 RBD We next examined the inhibitory effects of the RNA aptamers on the RasáRBD interaction by the competition assay. The GTPcS-bound Ras (133 n M ) and an equal amount or ®vefold or 10-fold molar excess of the individual RNAs were incubated with GST±RBD immobilized on the Sepharose matrix. The inhibition of Ras-binding to GST± RBD in the presence or absence of the RNA was tested by immunoblotting with the anti-Ras Ig (Fig. 3). The addition of RNA 9A or 9B signi®cantly reduced Ras-binding to the Raf-1 RBD (Fig. 3A, lanes 3±5 and 7±9, and Fig. 3B), indicating that both RNA aptamers ef®ciently inhibited the interaction between Ras and the Raf-1 RBD. In contrast, the addition of RNA 21.01 showed little inhibition ( Fig. 3A, Fig. 2. Binding curves for RNA aptamers with the Raf-1 RBD, the B-Raf R BD, and the RGL RBD. Data points were obtained by the nitrocellulose ®lter-binding assay a s described in the Materials and methods (9A with the Raf-1 GST±RBD, the B-Raf G ST±RBD, and the RGL GST± RBD: ®lled circles, open circles, and cro sses; 9B with the Raf-1 GST±RBD: squares; 21.01 with the Raf-1 GST±RBD: tri- angles). The concentration of labeled aptamer was 2 n M . 700 M. Kimoto et al. (Eur. J. Biochem. 269) Ó FEBS 2002 lane 1), although the addition of 100-fold and 1000-fold molar excess amounts of RNA 21.01 to Ras reduced the Ras-binding to the Raf-1 RBD, as shown in the previous report [25]. Thus, the inhibitory effects of RNAs 9A and 9B were superior as compared to that of RNA 21.01. The binding of the RNA aptamer 9A to the B-Raf RBD and the RGL RBD and the inhibitory effect on the RasáRBD interactions We next examined the binding speci®city of RNA 9A to other downstream effector molecules of Ras, such as RGL and B-Raf, and its inhibitory effect on the RasáRBD interactions. Both the RGL and Raf-1 RBDs have the same ubiquitin-like folds and interact similarly with Ras, although the sequen ce i dentity b etween them is as low as 10% [35,36]. Nonetheless, RNA 9A had no af®nity for the RGL RBD (the GST fusion form) (K d >> 1 l M ) (Fig. 2) and showed no inhibition of the RasáRGL interaction (data not shown). On the other hand, the B-Raf RBD has 52% sequence identity to the Raf-1 RBD and shares almost the same binding-speci®city for Ras with the Raf-1 RBD [37,38], and RNA 9A was capable of binding to the B-Raf RBD (the GST fusion form), with a K d value of 285  63 n M (Fig. 2 ). However, clear inhibition of the interaction between Ras and B-Raf was not observed (Fig. 3C). These r esults showed that the RNA aptamer 9 A was likely to recognize different structural features of the Raf RBD from those of the RasáRBD recognition. First of all, the RNA sharply discriminated between the Raf and RGL RBDs. Second, the RNA selectively inhibited the RasáRaf-1 interaction, although it bound to both Raf-1 and B-Raf. The af®nity of the aptamer for the B-Raf RBD (K d  285  63 n M ) was lower than that for the Raf-1 RBD (K d  152  23 n M ), which m ight cause the selec- tive inhibition. However, the af®nity of Ras to the B-Raf RBD (K d  60 n M ) is also lower than that to the Raf-1 RBD (K d  18 n M ) [6,37]. In a ddition, note t hat RNA 9B, which showed lower af®nity for the Raf-1 RBD (K d  361  46 n M ), ef®ciently inhibited the RasáRaf- 1 interaction. Thus, the speci®c inhibition by the RNA aptamer 9A would be caused mainly by t he recognition differences between the Ras protein and the RNA aptamer with each Raf protein. The Raf-1 amino-acid residues involved in the interaction with Ras have been identi®ed from mutational analyses [39,40] a nd the crystal structure of the complex between the Ras-like mutant Rap and the Raf-1 RBD [41]. Certain residues, such as Gln66, Lys84, and Arg89, are particularly conserved in homologous B-Raf, suggesting that the Ras binding site on the B-Raf RBD would be similar to that on th e Raf-1 RBD. Although the binding site of the aptamer is not exactly known, the RNA probably recognized the structural features that are common between the t wo Raf RBDs. Nevertheless, the binding site on the Raf-1 RBD enabled the aptamer to inhibit the RasáRaf interaction ef®ciently, but that on the B-Raf RBD did not. Inhibition of Ras-induced Raf-1 activation by RNA aptamers To evaluate the potential of the anti-(Raf-1) R NA aptamers 9A and 9B as reagents for probing the function of full- length Raf-1 in vitro or in vivo, we examined the inhibition of the Ras-induced Raf-1 activation by the aptamers in a cell- free system. Stokoe et al. have shown that a membrane fraction from cells expressing oncogenic Ha-Ras [G12V] activates Raf-1 in a cytoplasmic fraction [42]. Similarly, in our system, RafFH (Raf-1 with a FLAG epitope tag and six histidine residues in its C-terminus) in the cytoplasmic fraction (denoted the Raf fraction) prepared from HEK293 cells is activated by an incubation with t he membrane fraction from Sf 9 cells, which were previously infected with the Ha-Ras [G12V] baculovirus (denoted by the Ras membrane), as shown in Fig. 4, lanes 1 and 2 [33,34]. The addition of RNA 9A to the Raf fraction ef®ciently prevented the Ras-induced RafFH activation (Fig. 4, lanes 4 and 6). I n contrast, RNA 0C, w hich was r andomly isolated from the initial RNA pool and had negligible af®nity for the Raf-1 RBD (K d >> 1 l M ), inhibited neither the RafFH activation (Fig. 4, lanes 8 and 10) nor the RasáRaf i nteraction (data not sh own). The RafFH activation was also reduced to about 30% by the addition of RNA 9B to a ®nal concentration of 10 l M (data not Fig. 3. RNA aptamers 9A and 9B inhibit the interaction between Ras and the Raf-1 RBD. The amount of Ras binding to the Raf-1 GST± RBD (A and B) or the B-Raf GST±RBD (C) was measured by immunoblotting with the anti-Ras Ig, RAS004. Ras (20 pmol), in either the GTPcS-bound (T) or the GDP-bound (D) form, and the GST±RBD (20 pmol) were incubated in the absence and presence of various amounts of RNAs. (A) Lanes 2, 6, and 10, no RNA; lane 1, 200 pmol of RNA 21.01; lanes 3, 4, and 5 , 20, 100, and 200 pmol of RNA 9A; lanes 7, 8, and 9: 20, 100, and 200 pmol of RNA 9B. (B)Lanes1and2,noRNA;lanes3,4,5,6,7,and8,10,20,40,60,80, and 100 pmol of RNA 9 A (C) Lanes 1 and 2, no RNA; l anes 3, 4, 5, 6, 7, and 8, 100, 200, 400, 600, 800, and 1000 pmol of RNA 9A. Ó FEBS 2002 Anti-(Raf-1) RNA aptamers (Eur. J. Biochem. 269) 701 shown). The addition of RNA 9A or 9B after the termination of the RafFH-activation reaction (during the immunoprecipitation or in the kinase assay), however, did not inhibit the Raf k inase activity (data not shown). Thus, RNA 9A selectively inhibits the Ras-induced Raf-1 activa- tion process, not the Raf-1 kinase activity itself. These inhibitory effects correlate with the abilities of the aptamers to prevent the Ras interaction with the Raf-1 RBD, indicating that the RNAs can also inhibit the Ras interac- tion with full-length Ra f-1. Characterization of the sequence and the structural motif of RNA 9A required for binding to the Raf-1 RBD We determined a possible secondary structure of RNA 9A by enzymatic and chemical probing experiments. These experiments showed that the aptamer was likely to be highly structured, and folded into a pseudoknot with an additional hairpin-loop, as shown in Fig. 5 . In the enzymatic probing experiments, the labeled RNA 9A was treated w ith mung bean nuclease and RNase V 1 . Mung bean n uclease specif- ically digests single-stranded regions of RNA, and RNase V 1 speci®cally digests double-stranded regions of RNA. In the chemical probing experiments, the nonlabeled RNA 9A was treated with CMCT (for probing U at N-3) and DMS (for probing A at N-1 and C at N-3), and the chemically modi®ed nucleotides were detected by reverse transcription using the labeled primer. Some nucleotides in stem 3 w ere digested by RNase V 1 . Although DMS reacted with the adenosines at positions 12 and 3 3 in s tem 3, t hese adenosines might be in t he syn conformation an d be base-paired with the g uanosines at positions 32 and 11, respectively. The nucleotides in the single-stranded regions (lo ops 1, 3.1, and 3.2) were more sensitive to the relatively small molecules, CMCT and DMS, compared with mung bean nuclease. Although the secondary structure of RNA 9A, as deter- mined by the MULFOLD program [43], was composed of a stem±loop structure including stem 1 a nd stem 2, all o f the data obtained by the enzymatic digestion and c hemical probing experiments support the formation o f a pseudokn ot additionally containing stem 3. We prepared several shortened variants of RNA 9 A a nd examined their b inding af®nities to the Raf-1 RBD and inhibitory effects on the Ras áRaf-1 interaction, as described above. An 80-mer fragment of RNA 9A (positions 1±80) retained its binding af®nity (K d  173  31 n M )andthe inhibitory effect (data not shown). However, further truncation from either the 5¢ end (positions 21±80) or the 3¢-end (positions 1±70) of the 80-mer f ragment caused a precipitous loss in af®nity (K d >> 1 l M ), and these fragments did not inhibit the RasáRaf-1 interaction (data not shown). This essential region w as also predicted by the structural features determined by the enzymatic and chem- ical probing experiments; the region from positions 1±80 formed the pseudoknot structure and the region from 81 to 100 was not involved in the folding. To ascertain which regions of RNA 9A interact with t he Raf-1 RBD, we performed footprinting experiments o n the Fig. 5. Predicted secondary structure of RNA 9A. The secondary structure of RN A 9A wa s predicted b y enzymatic and chem ical modi®cation mapping ( see Fig. 6). The 5¢ and 3¢ de®ned sequences are shown in lower case letters, and the rand omized region is shown in capital letters. Enzymatic digestion and chemical modi®cation patterns were mapped onto the secondary structure. The sequences protected from chemical m odi®cation b y DMS an d CMCT in t he pre sence of the Raf-1 GST±RBD are enclosed (see Fig. 6). Fig. 4. Inhibition of Ras-induced Raf-1 activation by RNA 9A in the cell-free system. The Raf fraction was i ncubated alone (open bars) or with the Ras me mbrane (solid bars) in the pre senc e and absence of RNA f or 30 min at 1 6 °C. RafFH was then immunoprecipitated with the anti-FLAG Ig, and the kinase activity of RafF H in the immu no- precipitates was measured a s described in the Materials and methods. The activity in the reaction containing only the Raf fraction and the Ras membrane (lane 2) was normalized to 100. RNA 0C was u sed as a negative control RNA, which was randomly isolated from the initial RNA pool and had negligible anity for the Raf-1 RBD (K d >> 1 l M ). The data represent the average of at le ast two experiments. 702 M. Kimoto et al. (Eur. J. Biochem. 269) Ó FEBS 2002 RNA c omplexed with the Raf-1 RBD. As shown in Fig. 6, the nucleotides at positions 30, 31, 33, 45±48, and 68±70 in loops 3.1 and 3.2 were protected from chemical modi®ca- tion by DMS a nd CMCT in the presence of the protein. In contrast, the nucleotides in loops 1 and 2 were not protected from chemical modi®cation in the presence of the protein. These results indicate th at loops 3.1 a nd 3.2 are located o n the surface that binds to Raf-1. CONCLUSIONS In the present study, we isolated and characterized RNA aptamers to the Raf-1 RBD, and inhibited the Ras-induced Raf-1 activation by these aptamers. The aptamers, RNAs 9A and 9 B, prevented t he RasáRaf-1 interaction much more ef®ciently than the RNA aptamer that we previously isolated. In addition , RNA 9A cou ld speci®cally inhibit the Ras interaction with Raf-1, but affected neither B-Raf nor RGL. The m ost potent aptamer, R NA 9A, would b e useful as a tool for interfering with and regulating t he Ras- and Raf-1-mediated signaling networks in cells. Recently, a system that allows the expression of aptamers in the cytoplasm has been developed to modulate cytoplas- mic protein targets in the context of a living cell [ 21,22]. For example, in vivo expression of the aptamers to the cytoplasmic domain of the human b2 integrin subunit can block the pathways mediated through the domain [22]. Thus,theanti-(Raf-1)aptamermayalsobeusedtodissect the speci®c RasáRaf-1 interaction in the intricate cellular signaling pathways. ACKNOWLEDGEMENTS We thank E. Nishida for providing the histidine-tagged Xenopus MAPKK and the GST±KN-MAPK expression system, and A. Kikuchi for providing the RGL GST±RBD expression system. We also thank K. Sakam oto for helpful discussions. REFERENCES 1. Kolch, W. (2000) Meaningful relationships: the regulation of the Ras/Raf/MEK/ERK pathway by protein interactions. Biochem. J. 351, 289±305. 2. Rapp, U.R., Heidecker, G., Huleihel, M., Cleverland, J.L., Choi, W.C., Pawson, T., Ihle, J.N. & Anderson, W.B. (1 988) raf family serine/threonine p rotein kinases in mitogen signal transduction. Cold Spring Harb. Symp Quant. Biol. 53 , 173±184. 3. Campbell, S .L., Khosravi-Far, R ., Rossman, K.L., Clark, G.J. & Der, C.J. (1998) Increasing c omplexity of Ras signaling. Oncogene 17, 1395±1413. 4. Vojtek, A.B., Hollenberg, S.M. & Cooper, J.A. (1993) Mamma- lian Ras interacts directly with the serine/threonine kinase Raf. Cell 74, 205±214. 5. Chuang,E.,Barnard,D.,Hettich,L.,Zhang,X F.,Avruch,J.& Marshall, M.S. (1994) Critical binding and regulatory interactions between Ras and R af occur through a small, stable N-terminal domain of Raf and speci®c Ras eector residues. Mol. C ell. Biol. 14, 5318±5325. 6. Herrmann, C., Martin, G.A. & Wittinghofer, A. (1995) Quanti- tative Analysis of the complex between p21 ras and the Ras-binding domain of the human Raf-1 protein kinase. J. Bi ol. Chem. 270, 2901±2905. 7. Dau m, G ., Eisenm ann-Tappe , I ., Fries, H W., Troppmair, J. & Rapp, U.R. (1994) The ins and outs of Raf kinases. Trends Biochem. Sc i. 19, 474±480. 8. Marais,R.,Light,Y.,Paterson,H.F.,Mason,C.S.&Marshall, C.J. (1997) Dierent regulation of Raf-1, A-Raf, and B -Raf by oncogenic Ras and tyrosine kinases. J. Biol. Chem. 272, 4378±4383. 9. Spaargaren, M. & Bischo, J.R. (1994) Identi®cation of the guanine nucleotide dissociation stimulator for Ral as a put ative eector of R-ras, H-ras, K-ras, and Rap. Proc. Natl Acad. Sci. USA 91, 12609±12613. 10. Kikuchi, A., Demo, S.D., Ye, Z H., Chen, Y W. & Williams, L.T. (1994) ralG DS fa mily membe rs interact w ith th e eector l oop of ras p21. Mol. Cell. Biol. 14, 7483±7491. 11. Rod riguez-Viciana, P., Warne , P.H., Dhand , R., Van haesebroeck, B., Gout, I., Fry, M.J., Water®eld, M.D. & D ownward, J. (1994) Phosphatidylinositol-3-OH kin ase as a direct target of Ras. Nature 370, 527±532. 12. Sun, H., King, A.J., Diaz, H.B. & Marshall, M.S. (2000) Regulation of the protein kinase Raf-1 by oncogenic R as through phosphatidylinositol 3-kinase, Cdc42/Rac and Pak. Curr. Biol. 10, 281±284. Fig. 6. Footprints of the aptameráRBD complex. Th e autoradiogram shows the primer extension stops for reverse transcriptase on RNA 9A extracted from the aptameráRBD complex that was m odi®ed with CMCT and DMS. A, G, C, and U a re dideoxy sequencing lanes using the a ptamer as the template. The sequences protec ted from chemical modi®cation by C MCT (gray arrows) and DMS (black arrows) are shownontheright. Ó FEBS 2002 Anti-(Raf-1) RNA aptamers (Eur. J. Biochem. 269) 703 13. Horn, I.R., Wittinghofer, A., de Bru È ine, A.P. & Hoo genb oom, H.R. (1999) Selection of phage-displayed Fab antibodies on the active conformation of Ras yields a high anity conformation- speci®c antibod y preven ting the binding of c-Raf kinase to Ras. FEBS Lett. 463, 115±120. 14. Kolch, W., Philipp, A., Mischack, H., Dutil, E.M., Mullen, T M., Feramisco,J.R.,Meinkoth,J.L.&Rose,D.W.(1996)Inhibition of Raf-1 signaling by a monoclonal antibody, which interferes with Raf-1 activation and with Mek substrate binding. Oncogene 13, 1305±1314. 15. Hall-Jackson, C.A., Eyers, P.A., Cohen, P., Goedert, M., Boyle, F.T., Hewitt, N., Plant, H. & Hedge, P. (1999) Paradoxical acti- vation of Raf by a novel Raf inhibitor. Chem. Biol. 6, 559±568. 16. Hall-Jackson, C.A., Goedert, M., Hedge, P. & Cohen, P. (1999) Eect of SB 203580 on the activity of c-Raf in vitro and in vivo. Oncogene 18, 2047±2054. 17. Ellington, A.D. & S zostak, J.W. (1990) In vitro selection of RNA molecules that bind speci®c ligands. Nature 346, 818±822. 18. Tuerk, C. & G old, L. (1990) systematic evolution o f ligands by exponential enric hment: RNA ligands to bacteriophage T4 D NA polymerase. Science 249 , 505±510. 19. Werstuck, G. & Green, M.R. (1998) Controlling g ene expression in living cells through small molecule±RNA interactions. Science 282, 296±298. 20. Famulok, M., M ayer, G. & Blind, M. (2000) Nucleic a cid aptamers-from selection in vitro to applications in vivo. Acc. Chem. Res. 33, 591±599. 21. Mayer, G., Blind, M., Nagel, W., Bo È hm, T., Knorr, T., Jackson, C.L., Kolanus, W. & Famulok, M. (2001) Controlling small guanine-nucleo tide-exchan ge factor fun ction through cy toplasmic RNA intramers. Proc. Natl Acad. Sci. USA 98, 4961±4965. 22. Blind, M., Kolanus, W. & Famulok, M. (1999) Cytoplasmic RNA molecules of an inside-out signal-transduction cascade. Proc. N atl Acad. Sci. USA 96, 3606±3610. 23. Shi, H., Homan, B.E. & Lis, J.T. (1999) RNA aptamers as eective protein antagonists in a multicellular organism. Proc. Natl Acad. Sci. USA 96, 10033±10038. 24. Thomas, M., Che  din,S.,Carles,C.,Riva,M.,Famulok,M.& Sentenac, A. (1997) Selective Targeting and Inhibition of Yeast RNA Polymerase II by RNA aptamers. J. Biol. Chem. 27 2, 27980±27986. 25. Kimoto, M., Saka moto, K ., Shirouzu, M ., Hirao, I. & Yokoyama, S. (1998) RNA aptamers t hat s peci® cally bind to the R as-binding domain of Raf-1. FEBS Lett. 441, 322±326. 26. Shirouzu, M., Morinaka, K., Koyama, S., Hu, C D., Hori- Tamura, N., Okada, T., Kariya, K., Kataoka, T., Kikuchi, A. & Yokoyoma, S. (1998) Interactions of the amino acid residue at position 31 of the c-Ha-Ras Protein with Raf-1 and RalGDS. J. Biol. Chem. 273, 7737±7742. 27. Koyama, S., Chen, Y W., Ikeda, M., Muslin, A.J., Williams, L.T. & Kikuchi, A. (1996) Ras-interacting domain of RGL blocks Ras-dependent signal transduction in Xenopus oocytes. FEBS Lett. 380, 113±117. 28. Kosako, H., N ishida, E. & Gotoh, Y. (1993) cDNA cloning of MAP k inase kinase reveals kinase cascade pathways in yeasts to vertebra tes. EMBO J. 12, 787±794. 29. Gotoh, Y., Matsuda, E., Takenaka, K., Hattori, S., Iwamatsu, A., Ishikawa, M., Kosako, H. & Nishida, E. (1994) Characterization of rec ombinant Xenopus MAP kinase k inase m utated at potential phosphorylation sites. Oncogene 9, 1891±1898. 30. Hirao,I.,Madin,K.,Endo,Y.,Yokoyama,S.&Ellington,A.D. (2000) RNA aptamers that bind to and inhibit the ribosome- inactivating protein, pepocin. J. Biol. Chem. 275, 4943± 4948. 31. Koide, H., Satoh, T., Nakafuku, M. & Kaziro, Y. (1993) GTP± dependent association of Raf-1 with Ha-Ras: identi®cation o f Raf as a target downstream of R as in mammalian cells. Proc. Natl Acad. Sci. USA 90, 8683±8686. 32. Kanai, T., Hirohashi, S., Noguchi, M., Shimoyama, Y., Shimosato,Y.,Noguchi,S.,Nishimura,S.&Abe,O.(1987) Monoclonal antibody highly sensitive for the detection of ras p21 i n immunoblotting analysis. Jpn. J. Cancer. Res. 78, 1314± 1318. 33. Mizutani, S., Koide, H. & Kaziro, Y. (1998) Isolation of a new p rotein factor requ ired for activatio n of Raf-1 by Ha-Ras: partial puri®cation from rat brain cytosols. Oncogene 16, 2781± 2786. 34. Inouye, K., Mizutani, S., Koide, H. & Kaziro, Y. (2000) Forma- tion of the Ras dimer is essential for Raf-1 activation. J. Biol. Chem. 275, 3737±3740. 35. Kigawa, T., Endo, M., Ito, Y., Shirouzu, M., Kikuchi, A. & Yokoyama, S. (1998) Solution structure of the Ras-binding domain of RGL. FEBS Lett. 441, 413±418. 36. Shirouzu , M., H ashimoto, K., Kikuchi, A. & Yokoyama, S. (1999) Double-mutant a nalysis of the interaction of Ras with the Ras-binding domain of RGL. Biochemistry 38, 5103±5110. 37. Weber, C.K., Slupsky, J.R., Herrmann, C., Schuler, M., Rapp, U.R. & Block, C. (2000) Mitogenic signaling of Ras is regulated by dierential interaction with Raf i sozym es. Oncogene 19, 169±176. 38. Shinkai, M., Masuda, T ., Kariya, K., Tamada, T., Shirouzu, M ., Yokoyama, S. & Kataoka, T. (1996) Dierence in the m echanism of interaction o f Raf-1 and B-Raf with H-Ras. Biochem. Biophys. Res. Commun. 223, 729±734. 39. Block, C., Jankn echt, R., Herrmann, C., Nassar, N. & Witting- hofer, A. (1996) Quantitative structure -activity analysis correlating Ras/Raf interaction in vitro to Raf a ctivation in vivo. Nat. Struct. Biol. 3, 244±251. 40. Shirouzu, M., Koide, H., Fujita-Yoshigaki, J., Oshio, H., Toyama, Y., Yamasaki, K., Fuhrman, S.A., Villafranca, E., Kaziro,Y.&Yokoyama,S.(1994)Mutationsthatabolish the ability of Ha-Ras to associate with Raf-1. Oncogene 9, 2153±2157. 41. Nassar,N.,Horn,G.,Herrmann,C.,Block,C.,Janknecht,R.& Wittinghofer, A. (1996) Ras/Rap e ect or sp eci®city determ ined by charge reversal. Nat. Struct. Biol. 3, 723±729. 42. Stokoe, D. & McCormick, F. (1997) Activation of c-Raf-1 by Ras and Src through dierent mechanisms: activation in vivo and in vitro. EMBO J. 16, 2384±2396. 43. Ja eger, J.A., Tu rn er, D.H. & Zuker, M. (1990) Predicting op tima l and sub optimal sec ondary s tructure for RNA. Methods Enzymol. 183, 281±306. 704 M. Kimoto et al. (Eur. J. Biochem. 269) Ó FEBS 2002 . Anti- (Raf-1) RNA aptamers that inhibit Ras-induced Raf-1 activation Michiko Kimoto 1,2 , Mikako Shirouzu 2,3 ,. surface that binds to Raf-1. CONCLUSIONS In the present study, we isolated and characterized RNA aptamers to the Raf-1 RBD, and inhibited the Ras-induced Raf-1