Identification of a region of Bacillus subtilis Ffh, a homologue of mammalian SRP54 protein, that is essential for binding to small cytoplasmic RNA.

Bacillus subtilis Ffh and scRNA are homologues of mammalian SRP54 and SRP RNA, respectively, which are components of the eukaryotic signal recognition particle (SRP). Ffh (446 amino acids) interacts with scRNA to form a stable complex in vivo. Here, we identified an RNA-binding domain of Ffh. The results obtained using a series of deletion mutants show that amino acid positions 364 to 432 in the C-terminal region of Ffh correlates with its ability to bind RNA. The amino acid sequence of this region is well conserved among members of the SRP54 protein family. This sequence contains two hydrophobic regions (h2, 364 to 391, and h3, 416 to 435), separated by the positively charged amino acid motif, 398RRKRIAKGSG407. Among the basic amino acid residues in this region, Arg-401 was essential for binding to scRNA, but Arg-399 and Lys-400 were not. The co-existence of Arg-398 and Lys-404 was necessary for the same affinity as wild type Ffh. The two glycine residues of the 405GSG407 were also essential. MH23 peptide (91 amino acids) encompassing from 356 to 446, consisting of h2-RRKRIAKGSG-h3, bound scRNA with the same affinity as wild type Ffh, whereas a 24-amino acid synthetic peptide 392DIINASRRKRIAKGSGTSVQEVNR415 did not. The region containing two hydrophobic segments separated by the positively charged motif is the minimal requirement of Ffh for RNA binding.

In mammals, the signal recognition particle (SRP) 1 plays an important role in targeting secretory proteins to the membrane of the endoplasmic reticulum. Mammalian SRP is a ribonucleoprotein particle composed of one RNA molecule (7 S RNA here referred to SRP RNA) and six polypeptides of 9, 14, 19, 54, 68, and 72 kDa (1,2). SRP interacts with a signal sequence of a nascent polypeptide as it emerges from the ribosome, then the elongation of the nascent chain is inhibited (3,4). The complex consisting of SRP, nascent polypeptide chain, and ribosome is targeted and binds to the heterodimeric SRP receptor in the endoplasmic reticulum membrane (3)(4)(5).
The SRP54 subunit of SRP has a central function in recog-nizing and binding signal sequences (6 -8). This protein comprises the structurally distinct, N-terminal G-and C-terminal M-domains (9,10). The G-domain contains three GTP-binding motifs, and it may play an essential role in mediating the interaction between SRP and the SRP receptor (11). The Mdomain is characterized by a high content of predominantly positively charged residues and an abundance of methionine residues. Moreover, there are three putative amphipathic helices in the M-domain (9). The M-domain is a binding site for the signal sequence and SRP RNA (12)(13)(14)(15). The three-dimensional structure of SRP54 protein has been determined by scanning transmission electron microscopy (STEM) (16). The scanning transmission electron microscopy image consisting of one larger and one smaller domain (probably representing the Gand M-domain, respectively) joined by slender linker, was in agreement with genetic and biochemical data. A protein homologous to SRP54 has been identified in Eukaryotes and Eubacteria (3,(17)(18)(19)(20)(21)(22)(23)(24). The features of the two structural domains (G-and M-domains) of SRP54 are highly conserved in these proteins. Furthermore, these proteins form an SRP-like complex with RNA molecules that are structurally related to mammalian SRP RNA (25)(26)(27)(28). The binding of Ffh to 4.5 S RNA, which are homologues of SRP54 and SRP RNA in Escherichia coli, respectively, is essential for the signal recognition function (29). Larsen and Zweib (30) proposed that SRP RNA consists of 8 helices (numbered 1 to 8). The nucleotide sequence in the region corresponding to the helix 8 is highly conserved among SRP RNA homologues, and this loop region mediates the binding to SRP54 and its homologues (17,31,32). The SRP RNA homologue of Bacillus subtilis scRNA (small cytoplasmic RNA) lacks helices 6 and 7, whereas those of E. coli 4.5 S RNA lack helices 1, 2, 3, 4, 6, and 7 as well as part of 5 (30). Furthermore, mammalian SRP54 can bind to E. coli 4.5 S RNA in vitro (33). These findings indicated that the RNAbinding form of SRP and related complexes contain common features. The RNA-binding domain of the homologue of SRP54 in Mycoplasma mycoides is located in the M-domain (34), and it has been proposed that the hydrophilic faces of the predicted amphipathic ␣-helices in the M-domain contact SRP RNA (11,14,15). However, the RNA-binding region of Ffh has not yet been analyzed in detail.
B. subtilis can produce large amounts of extracellular enzymes. We demonstrated that it contains SRP-like particles and that they are closely involved in enzyme secretion (26). B. subtilis Ffh protein has features very similar to these of SRP54 (23). The Ffh also has G-domain and M-domain. A phylogenetic comparison of SRP54 protein has revealed three highly hydrophobic regions (hydrophobic regions 1, 2, and 3 are referred to as h1, h2, and h3) in the M-domain (23). The h1, h2, and h3 regions correspond to amino acid positions 307 to 331, 364 to 391, and 416 to 435, respectively, in B. subtilis Ffh. To deter-* This work was supported in part by grants-in-aid for scientific research from the Ministry of Education, Science and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed: Institute of Biological Sciences, University of Tsukuba, Tsukuba-shi, Ibaraki 305, Japan. Tel./Fax: 81-298-53-6680; E-mail: kyamane@sakura.cc.tsukuba.ac.jp. 1 The abbreviations used are: SRP, signal recognition particle; scRNA, small cytoplasmic RNA; kb, kilobase(s); bp, base pair(s); PCR, polymerase chain reaction; IPTG, isopropyl-1-thio-␤-D-galactopyranoside; ds, double-stranded. mine the essential region of Ffh protein for the binding to scRNA of a homologue of SRP RNA, we constructed several mutant Ffh proteins of B. subtilis into which deletions and amino acid replacements were introduced. We then determined their RNA-binding activity using the filter assay (34). On the basis of the RNA-binding activity of wild-type and mutants of Ffh, we defined a minimal RNA-binding domain encompassing h2 to h3 in the M-domain. Moreover, in the 398 RRKRI-AKGSG 407 sequence located between h2 and h3, we found that Arg-401, Gly-405, and Gly-407 are essential for RNA binding.

EXPERIMENTAL PROCEDURES
Plasmid Construction for Large Deletion Mutants of Ffh-An E. coli plasmid expressing B. subtilis wild-type Ffh with a hexahistidine tag at its C terminus (Ffh-WT) was constructed as follows. A 1.3-kb NcoI-BglII fragment of pTUE815 (26) was inserted into NcoI-BglII sites of pQE60BS, in which BamHI and SmaI linker DNA fragments were introduced into HindIII and EcoRI sites of pQE60, respectively. The resulting plasmid was designated pTUE920.
The plasmid expressing Ffh-⌬G was constructed by digesting pTUE920 with HindIII to remove the 472-bp HindIII fragment encoding part of the G-domain, followed by re-ligation. The plasmid expressing Ffh-⌬C was constructed by digesting pTUE920 with NspI and BglII, to remove the 44-bp NspI-BglII fragment encoding the 14 C-terminal amino acids. Each cohesive end of the 4.7-kb NspI-BglII fragment was digested with S1 nuclease and re-ligated. The plasmid expressing Ffh-⌬h2 was constructed from pTUE920 by digestion with AvaI and NaeI to remove the 178-bp AvaI-NaeI fragment encoding amino acid positions 378 to 397 containing h2 region. Cohesive ends of the 4.6-kb AvaI-NaeI fragment were filled by T4 DNA polymerase and re-ligated.
The C-terminal part of Ffh was deleted further as follows. These DNA fragments were produced by the polymerase chain reaction (PCR) using four oligonucleotide primers and pTUE905 (23) as the template, which included the wild-type ffh gene. The synthetic oligonucleotide N (5Ј-CAAACCATGGCATTTCAA-3Ј) containing a novel NcoI site was used as one primer, and the synthetic oligonucleotides T2 (5Ј-CAG-ATCTCGGCTGTTCTTTTTCAA-3Ј), ⌬23 (5Ј-CAGATCTCAGGCCCT-TCATTTTAC-3Ј), C1 (5Ј-CAGATCTTTCCTGTACGGATGTCCCG-3Ј), and G1 (5Ј-CAGATCTGCCGAGAATCCTTGATGC-3Ј), all of which contained a novel BglII site, were used as the other primers with which to construct plasmids expressing Ffh-⌬M, Ffh-⌬h3, Ffh-⌬h3Ј, and Ffh-⌬h23, respectively. The resulting PCR fragments were digested with NcoI and BglII and inserted into the NcoI-BglII sites of pQE60BS. The plasmid expressing Ffh-⌬h1 was constructed using two DNA fragments. A DNA fragment was amplified by PCR using the synthetic oligonucleotides N and HO1 (5Ј-CAGGCCTGTGCTTTTTCAATC-3Ј) containing a StuI site, then digested with NcoI and StuI. Another DNA fragment was also amplified by PCR using the synthetic oligonucleotides O2 (5Ј-CAGGCCTTGAAAAACATCCAAGTTG-3Ј), which contained a novel StuI site, and C (5Ј-GGATCCAGATCTCATAAAAGG-TAGCTT-3Ј), which contained a novel BglII site, then digested with BglII and StuI. These two DNA fragments were ligated and inserted into the NcoI-BglII sites of pQE60BS. The plasmid expressing the Ffh-⌬h12 was constructed as follows. A DNA fragment of the 5Ј end of ffh was amplified by PCR using synthetic oligonucleotides N and HO1 and digested with NcoI and StuI. To prepare a DNA fragment of the 3Ј end of ffh, we introduced the novel StuI site on pTUE920. A HindIII-BamHI restriction fragment of plasmid pTUE920 was introduced into M13mp19, and a single stranded DNA was prepared as a template for site-directed mutagenesis, which proceeded using the "Mutan K" system (TAKARA SHUZO Co., Ltd., Tokyo) based on the method of Kunkel et al. (35,36). The nucleotide sequence of ACCGTC at position 1951 to 1956 in the ffh gene was changed to AGCGCT using the oligonucleotide STUI-A (5Ј-CTGCTTAAGCAAGCGCTTGACTTCCTG-3Ј). We then changed the nucleotide sequence of AGCGCT to AGGCCT, which was a novel StuI site, using the oligonucleotide STUI-B (5Ј-GCTTAAGCAAG-GCCTTGACTTCC-3Ј). The BstXI-BamHI fragment, which was introduced into the novel StuI site, was checked by sequencing and ligated into the BstXI-BamHI site of pTUE920. The resulting plasmid was designated pTUE921. A NcoI-StuI fragment from pTUE921 was then ligated with the PCR-amplified fragments of N and HO1.
Plasmid Construction for Ffh Mutants Using the Site-directed Mutagenesis-The plasmids expressing the Ffh derivatives carrying amino acid replacements and short deletions were constructed by site-directed mutagenesis as described above. The synthetic oligonucleotides used were: K364M (5Ј-TGGATGTTCATCAGGCCC-3Ј), K371M (5Ј-TTCAG- Using the oligonucleotides and the single-stranded DNA described above as the template, the respective replacements and a short deletion were introduced, then the BstXI-BamHI fragment was ligated into the BstXI-BamHI sites of pQE60BS. The plasmid expressing the three K404Q derivatives, RRKR-K404Q, RQQR-K404Q, and QQQR-K404Q, was constructed using the same oligonucleotide (5Ј-CCGCTTCCTTGT-GCAATCCGC-3Ј) and respective templates of single-stranded DNA. The single-stranded DNA templates for RRKR-K404Q, RQQR-K404Q, and QQQR-K404Q derivatives were prepared from M13mp19 which contained HindIII and BamHI fragments of the plasmid expressing Ffh-WT (pTUE920), RQQR, and QQQR, respectively. These plasmids were introduced followed by the BstXI-BamHI fragment ligated into BstXI-BamHI sites of pQE60BS.
Plasmid Construction to Prepare MH23 Polypeptide-The plasmid expressing the Ffh-MH23 was constructed as follows. A DNA fragment was amplified by PCR using the synthetic oligonucleotides C and N3 (5Ј-GCCATGGGTGCAGGTAAAATGAAGG-3Ј), the latter of which contained a novel NcoI site. This fragment was digested with BglII and NcoI and inserted into the BglII and NcoI sites of pQE60BS.
Synthesis of Peptide-A synthetic 24-mer peptide corresponding to the region between B. subtilis Ffh amino acid positions 392 to 415 (DIINASRRKRIAKGSGTSVQEVNR) were synthesized on BIOSYN 2000 peptide synthesizer (Lewisville, TX) by solid phase Fmoc (N-(9fluorenyl)methoxycarbonyl) chemistry (37). Upon completion of synthesis, the resin is treated with 95% trifluoroacetic acid, containing 1% thioanisole as a scavenger. The peptide are purified using the BIOCAD perfusion chromatography system from Perseptive Biosystems (Framingham, MA). The separation column used is the R2M 20-porous beads, also from Perseptive Biosystems. A gradient from 20% solvent A to 100% solvent B is used. Solvent A is 20% aqueous acetonitrile containing 0.1% trifluoroacetic acid. Solvent B is 100% acetonitrile containing 0.1% trifluoroacetic acid. The purity of the peptide was more than 95%.
Expression and Purification of Ffh and Its Derivatives with the Hexahistidine Tag-Each recombinant protein was expressed in E. coli M15 harboring pREP4 (Qiagen) grown for 5 h at 37°C in L-broth (150 ml) supplemented with 2 mM IPTG. Cells were harvested and suspended in 10 ml of sonication buffer (20 mM Tris-HCl (pH 8.0), 5 mM MgCl 2 , 150 mM NaCl, 10% glycerol, 0.05% Nonidet P-40, 0.1 mM phenylmethylsulfonyl fluoride). After freezing and thawing the suspension, a halfvolume of glass beads was added and the mixture was vigorously vortex-mixed for 10 min at 4°C. The supernatant was separated from the glass beads that went to the bottom and centrifuged at 10,000 ϫ g for 15 min at 4°C. The precipitate was dissolved in 10 ml of buffer A (0.1 M NaH 2 PO 4 , 0.01 M Tris-HCl (pH 8.0), 8 M urea) and centrifuged again at 12,000 ϫ g for 15 min at 4°C. Thereafter, 300 l of a 50% slurry of Ni 2ϩ -NTA agarose (Qiagen) equilibrated with buffer A was added to the supernatant and gently mixed for 1 h. The resin was washed three times with 10 ml of buffer B (same composition as buffer A but the pH was adjusted with 6.3), and the protein bound to Ni 2ϩ -NTA agarose was finally eluted with 500 l of buffer B containing 0.2 M imidazole. The eluted protein was sequentially dialyzed against buffer C (20 mM Tris-HCl (pH 7.5), 0.5 M NaCl, 50% glycerol, 0.1 mM phenylmethylsulfonyl fluoride) containing 6, 4, 2, and 0 M urea. The purity of each protein obtained was estimated to be between 67% and 93%, based on laser densitometry of the Coomassie Brilliant Blue-stained gel using a Bio-Image Analyzer (Millipore/MilliGen).

RESULTS
Binding of Wild-type Ffh with sc104 RNA in Vitro-To prepare the purified Ffh protein, we constructed the E. coli plasmid pTUE920 that encoded Ffh with a hexahistidine tag at its C terminus. We then examined the effect of adding the hexahistidine tag to Ffh in vivo. The ffh conditional mutant, B. subtilis NA20BF in which the expression of ffh is controlled by the Pspac-1 promoter (23), cannot grow in the absence of 1 mM IPTG. However, the growth in the absence of IPTG was complemented by the introduction of pTUBE911, in which the B. subtilis ffh gene was placed under the control of the protein A promoter, into the cells as well as by that of pTUBE917, in which a gene for ffh with a hexahistidine tag at its C terminus was placed under the control of the protein A promoter. Therefore, we constructed pTUE920 and prepared B. subtilis Ffh with a hexahistidine tag from an E. coli transformant as the wild-type Ffh of B. subtilis. The preparation was designated as Ffh-WT. The purity of the sample was over 85% according to laser densitometry of the Coomassie Brilliant Blue stained gel after SDS-polyacrylamide gel electrophoresis.
As shown in Fig. 1A, a distinct band migrated much more slowly than the free sc104 RNA in a nondenaturing gel when 32 P-labeled sc104 RNA, a derivative of scRNA, was incubated with Ffh-WT (lane 2). This band disappeared upon adding a 250-fold excess of non-radiolabeled scRNA (sc271) or sc104 RNA, but not 250-fold excess of tRNA as a competitor (lanes  3-5). Although a slowly migrating band was also detected using scRNA (sc271) as a probe in the presence of Ffh, it was closer to the band of the free probe than those of sc104 RNA (data not shown). Wild-type scRNA consisting of 271 nucleotides (sc271) was too large to be used as a probe in the gel retardation and filter binding assays. The Zuker minimal energy program (38) predicted that sc104 RNA could almost form a hairpin structure similar to that of wild-type of scRNA. Therefore, to monitor the formation of the complex between B. subtilis Ffh and scRNA in vitro, we used sc104 RNA corresponding to positions 116 to 219 of scRNA, which lacked helix 1ϳ4, and most of helix 5, but remained helix 8. The size of sc104 RNA is similar to that of E. coli 4.5 S RNA, which is a counterpart of scRNA. Furthermore, a scRNA of smaller size is functional in vivo, since the sc104 RNA complements the growth defect caused by depleting scRNA form B. subtilis (39). In this study, sc104 RNA was used for the gel retardation and filter assays instead of wild-type scRNA (sc271). The dose dependence of the binding of wild-type Ffh and sc104 RNA was also examined by a filter binding assay. Up to 10 M Ffh, the amount of sc104 RNA bound to Ffh increased as Ffh added (Fig. 1B). The binding curve of M. mycoides Ffh protein is also sigmoidal (34). The half-maximal binding (M1 ⁄2 ) was approximately 0.15 M.
scRNA Binding by the Large Deletion Mutants of Ffh-To search the RNA-binding region of Ffh protein, purified mutant proteins were used in the RNA-binding assay. Ffh-⌬G lacks most of the G-domain ( Fig. 2A), and it bound to sc104 RNA in the gel retardation (Fig. 2B) and filter assays (Fig. 2C). In contrast, Ffh-⌬M lacking the M-domain had no affinity for sc104 RNA in either assay (Fig. 2, B and C). Therefore, like mammalian SRP54, the RNA-binding domain of B. subtilis Ffh is also located in the M-domain. Since the binding activity data determined by the gel retardation and filter assays were consistent, we employed the filter assay to further studies as the quantitative analysis of the interaction between derivatives of Ffh and scRNA.
Using a series of large deletion mutant proteins of Ffh (Fig.  3A), we searched the region that binds to scRNA in Ffh (Fig.  3B). Neither Ffh-⌬h3, Ffh-⌬h3Ј (deleted amino acid positions 392 to 446 and 413 to 446, respectively), nor Ffh-⌬h2 (deleted amino acid positions from 338 to 397) bound to sc104 RNA. In contrast, Ffh-⌬h1 (deleted amino acid positions from 311 to 362) and Ffh-⌬C (deleted 14 amino acids at the C-terminal region) bound to sc104 RNA with the same affinity as that of wild-type Ffh. These results show that the region of Ffh spanning positions 364 to 432 and corresponding to the region from h2 to h3, is necessary for scRNA binding. Furthermore, the results using the other Ffh derivatives, Ffh-⌬h23 (deleted amino acids from position 364 to 446 including the h2 and h3 scRNA-binding Region of Bacillus subtilis Ffh regions) and Ffh-⌬h12 (deleted amino acids from position 311 to 415 including the h1 and h2 regions), were consistent with this conclusion, because these two Ffh derivatives could not bind sc104 RNA (Fig. 3B). When these mutant Ffh proteins were introduced into B. subtilis NA20BF, only Ffh-⌬C completely restored the growth defect of NA20BF in the absence of IPTG, whereas the others only partially did so (data not shown). All mutant Ffh protein except for Ffh-⌬G mutant was shown to hydrolyze GTP the same as Ffh-WT (data not shown).
scRNA Binding of Amino Acids Replaced and Short Deletion Mutant Proteins of Ffh-The results shown above indicate that the peptide region from position 364 to 432 of B. subtilis Ffh is necessary for its binding with sc104 RNA. The hydrophobic regions, h2 and h3, were highly conserved among identified SRP54 homologues. Furthermore, B. subtilis Ffh bound 4.5 S RNA and SRP RNA because several characteristic features caused by the depletion of scRNA in B. subtilis were functionally compensated by the expression of human SRP RNA or E. coli 4.5 S RNA (40). These facts suggested that the conserved amino acid residues in the region of amino acid positions 364 to 432 contributed to the RNA binding. Furthermore, there are positively charged and aromatic residues located on several RNA-binding motifs (41). Based upon a comparison of the amino acid sequences of B. subtilis Ffh, E. coli Ffh, and mammalian SRP54, conserved positively charged and aromatic amino acids in the hydrophobic h2 and h3 regions and the region between them were replaced by site-directed mutagenesis using the oligonucleotides described under "Experimental Procedures" (Fig. 4). The mutants proteins were designated as K364M (Lys-364 was replaced with Met), K371M (Lys-371 was replaced with Met), H375Q (His-375 was replaced with Gln), F428Q (Phe-428 was replaced with Gln), K428M (Lys-428 was replaced with Met), and K435M (Lys-435 was replaced with Met). The sequence RRKR (positions 398 to 401 in B. subtilis), a remarkable basic amino acid cluster located between the h2 and h3 regions was replaced with four glutamine residues, the mutant Ffh contained QQQQ. The sequence GSG (position at 405 to 407 in B. subtilis) was deleted and designed as ⌬GSG. The sequence GXG was highly conserved in the amino acid sequences among SRP54 and its homologues.
The RNA-binding activity of wild-type and mutant Ffh proteins with sc104 RNA was assayed using nitrocellulose filters (Fig. 5A), and the activity of each mutant protein is shown as a percentage of that of wild-type Ffh at a concentration of 1 M protein added to the reaction mixture (Fig. 5B). K364M, K371M, H375Q, and F420Q had activity similar to that of Ffh-WT, and the binding activities of K428M and K435M were reduced to 78 and 59% of that of Ffh-WT. In contrast, the QQQQ and ⌬GSG mutant proteins had no binding activity.
These results indicated that the shorter region between h2 and h3 of the Ffh M-domain is very important for the binding to sc104 RNA, and that the positively charged amino acids in h2 (amino acids position 346 to 391) and h3 (amino acids position 416 to 435) are less important than the shorter region. The short peptide sequence between h2 and h3 constitutes one of the most highly conserved regions in the M-domains of SRP54 proteins. Furthermore, the two arginine residues among sequences RRKR (italic), and GSG were especially highly conserved in the short region between the h2 and h3 regions among SRP54 homologues (Fig. 4).
RNA-binding Activity in Mutant Ffh Proteins at the RRKR Sequence-The cluster of four positively charged amino acids located between the hydrophobic region h2 and h3 has also been found in the SRP54 homologues of M. mycoides (17) and in chloroplasts of Arabidopsis thaliana (22) (Fig. 4). Furthermore, a similar structure, RXXR, has been identified in the h2 and h3 regions of Eukarya, Canis familiaris (10), Mus musculus (9), Lycopersicon esculentum (18), A. thaliana (21), Schizosaccharomyces pombe (19), and Saccharomyces cerevisiae (19,20), as shown in Fig. 4. Since the QQQQ derivative of B. subtilis Ffh completely lacked RNA-binding activity, we prepared three novel derivatives (RQQR, RQQQ, and QQQR). As shown in Fig.  6, the RQQR and QQQR derivatives had similar RNA-binding activities, whereas that of RQQQ was dramatically reduced. Therefore, the arginine residue at position 401 (R401) of 398 RRKR 401 in B. subtilis Ffh was essential for the RNA-binding activity.
RNA-binding Activity of Mutant Ffh Proteins at the GSG Sequence-The GSG sequence is highly conserved among SRP54 and its homologues. To determine whether or not the secondary structure of this region is closely involved in RNAbinding activity, ASA (two glycines were replaced with alanine) and PSP (two glycines were replaced with proline) derivatives were prepared. The RNA-binding activity of both ASA and PSP derivatives was reduced dramatically (Fig. 7). We then prepared the GSP and PSG derivatives to determine which glycine residue of the GSG sequence is more important for RNA binding. Neither PSG nor GSP derivatives bound to sc104 RNA (Fig. 7). These results indicated that the GSG sequence is crucial for the RNA-binding activity of Ffh.
Lys-404 Replacements of Ffh Protein Affected the RNA-binding Activity-A positively charged amino acid between RRKR and GSG sequences, such as Lys-404 of B. subtilis Ffh protein, was conserved among most SRP54 and its homologues, except  (23); Ec, E. coli (24); AtCP, A. thaliana chloroplast (22); Mm, M. mycoides (17); Cf, C. familiaris (10); At, A. thaliana (21); Sc, S. cerevisiae (19,20); and Sp, S. pombe (19) were aligned. Highly conserved residues with all SRP54 homologues are shown in the on the black boxes in reverse type. Only conserved residues with each group of Eubacteria or Eukaryotes are shown in the gray shaded boxes. On the top is a diagram of the residues replaced in this study.
RNA-binding Activity of MH23 Peptide and a Synthetic Peptide (a 24-Amino Acid Peptide)-Based upon the findings that the region from position 364 to 432 of Ffh is necessary for its binding with scRNA as shown in Fig. 3, we determined whether the peptide of this region was sufficient for the binding to scRNA. We constructed the MH23 peptide corresponding to amino acid positions 356 to 446, which included the peptide region from h2 to h3 (Fig. 9A) and assayed the RNA-binding activity using nitrocellulose filters. The MH23 peptide bound to sc104 RNA with the same affinity as Ffh-WT (Fig. 9B).
Furthermore, to determine whether or not the region between h2 and h3 regions is sufficient for the binding to scRNA, we synthesized a peptide consisting of 24 amino acid residues (amino acid positions 392 to 415) corresponding to the stretch between these regions (Fig. 9A). This synthetic peptide did not bind sc104 RNA according to the nitrocellulose filter assay (Fig.  9B). Additionally, under the conditions shown in Fig. 9B, a 1000-fold excess of the synthetic peptide did not compete with the RNA-binding activities of either Ffh-WT or MH23 in the nitrocellulose filter assay (data not shown).
In total, these results showed that the 24 amino acids contained a region that was essential, but not sufficient, for bind-ing to scRNA. In contrast, the amino acids positions spanning 354 to 446 were sufficient for the binding to scRNA. The data indicated that not only 24 amino acids but also the adjacent regions, h2 and h3, were necessary for the binding to scRNA. DISCUSSION The results of this study indicated that in B. subtilis Ffh-scRNA complex the peptide from amino acid position 364 to 432 of B. subtilis Ffh, including two conserved hydrophobic regions (h2 and h3), is necessary and sufficient for scRNA binding. Point mutations within the region between h2 and h3 dramatically affected the RNA-binding activity. Substitutions of Arg-401 and substitution of two glycine residues in the GSG sequence abolished the binding, suggesting that these amino acid residues bind directly to RNA. The RQQR derivative of B. subtilis Ffh showed the same binding activity as Ffh-WT (Fig.  5). The RXXR sequence in this region is found in SRP54 of C. familiaris in which the amino acid sequence of the corresponding region is RIQRVARGSG (the conserved amino acids are underlined). Moreover, mammalian SRP54 can bind bacterial SRP RNA (33). These data indicated that two conserved positively charged residues in this sequence are necessary for RNA binding. On the other hand, effect of the introduction of amino acid substitutions into the h2 and h3 regions is less than that positively charged residues. Computer-assisted secondary structure analysis of this region suggested that both hydrophobic regions (h2 and h3) fold into an amphipathic ␣-helix, with methionine and other hydrophobic residues clustered on one face (9). The positively charged amino acid residues clustered in the predicted boundaries between two helices. Therefore, it scRNA-binding Region of Bacillus subtilis Ffh is more plausible that h2 and h3 cannot bind directly to RNA, but that a secondary or tertiary structure formed by h2 and h3 plays a pivotal role in RNA binding. This model is supported by our data showing that a chemically synthesized oligopeptide (corresponding to amino acid positions 392 to 415) could not bind to RNA by itself (Fig. 9) and did not interfere with RNA binding of Ffh-WT (data not shown). Furthermore, Ffh-⌬h3, which contains the RRKRIAKGSG sequence but not the h3 region, had no binding ability.
The nucleotide sequence and secondary structure of helix 8 of SRP RNAs are highly conserved. This domain consists of an extended stem-loop structure. Moreover, highly conserved and invariant residues lie within the tetranucleotide loop (30). Mammalian SRP54 protein is homologous to bacterial Ffh protein, and it can bind 4.5 S RNA (33), suggesting that these conserved nucleotides within the stem-loop structure region are important for binding. Using an in vitro binding assay, Wood et al. (32) have shown that the evolutionarily conserved residues within the stem structure of 4.5 S RNA including the bulged region, are required for Ffh binding rather than conserved residues within the tetranucleotide loop. This suggests that the SRP54 family consists of double-stranded RNA (dsRNA)-binding proteins. Computer searches revealed a nonsignificant relatedness between Ffh and other cellular (RNAbinding) proteins. However, a reduced set of amino acids in the RRKRIAKGSG sequence of Ffh was similar to a consensus sequence (GXGXSKKXAK) found in dsRNA-binding proteins, such as E. coli RNaseIII, human immunodeficiency virus type I (HIV I) TRBP (TAR RNA-binding protein), Drosophila mRNA localization protein staufen, and human P1/dsI kinase (42). The tertiary structure of RNaseIII has been elucidated by NMR (43). In the RNA-binding domain of RNaseIII containing the consensus sequence, three-stranded antiparallel ␤-sheets are packed on one side against the two helices, and this structure exposes positively charged amino acids to facilitate RNA binding. On the other hand, a similar sequence has also been found in E. coli protein synthesis elongation factor G (EF-G). 2 This structural feature has been discovered in the C-terminal region of Thermus thermophilus EF-G by determination of the crystal structure (44). A deletion of the corresponding region of E. coli EF-G abolished the activity of ribosome binding (45). This suggests that this structure is essential for binding to 23 S rRNA (46,47), since EF-G interacts with the ribosome through the 23 S rRNA. These data indicate that not only the amino acid sequence, but also positively charged residues exposed by encompassing higher structures, are conserved among dsRNAbinding domains.
In this study, we concluded that amino acids 364 to 432, which include the hydrophobic h2 and h3 that will form an ␣-helix structure and the positively charged conserved region between these helices, are critical for RNA binding. Therefore, it would be useful to determine the crystal structure of the RNA-binding region of Ffh for further understanding the structure-function relationship of double-stranded RNA motifs.