Structure analysis of the 5' external transcribed spacer of the precursor ribosomal RNA from Saccharomyces cerevisiae.

Full-length precursor ribosomal RNA molecules were produced in vitro using as a template, a plasmid containing the yeast 35 S pre-rRNA gene under the control of the phage T3 promoter. The higher-order structure of the 5'-external transcribed spacer (5' ETS) sequence in the 35S pre-rRNA molecule was studied using dimethylsulfate, 1-cyclohexyl-3-(2-morpholinoethyl)-carbodiimide metho-p-toluenesulfonate, RNase T1 and RNase V1 as structure-sensitive probes. Modified residues were detected by primer extension. Data produced were used to evaluate several theoretical structure models predicted by minimum free-energy calculations. A model for the entire 5'ETS region is proposed that accommodates 82% of the residues experimentally shown to be in either base-paired or single-stranded structure in the correct configuration. The model contains a high degree of secondary structure with ten stable hairpins of varying lengths and stabilities. The hairpins are composed of the Watson-Crick A.T and G.C pairs plus the non-canonical G.U pairs. Based on a comparative analysis of the 5' ETS sequence from Saccharomyces cerevisiae and Schizosaccharomyces pombe, most of the base-paired regions in the proposed model appear to be phylogenetically supported. The two sites previously shown to be crosslinked to U3 snRNA as well as the previously proposed recognition site for processing and one of the early processing site (based on sequence homology to the vertebrate ETS cleavage site) are located in single-stranded regions in the model. The present folding model for the 5' ETS in the 35 S pre-rRNA molecule should be useful in the investigations of the structure, function and processing of pre-rRNA.     

Three-dimensional model of Escherichia coli ribosomal 5 S RNA as deduced from structure probing in solution and computer modeling.

The conformation of Escherichia coli 5 S rRNA was investigated using chemical and enzymatic probes. The four bases were monitored at one of their Watson-Crick positions with dimethylsulfate (at C(N-3) and A(N-1], with a carbodiimide derivative (at G(N-1) and U(N-3] and with kethoxal (at G(N-1, N-2]. Position N-7 of purine was probed with diethylpyrocarbonate (at A(N-7] and dimethylsulfate (at G(N-7]. Double-stranded or stacked regions were tested with RNase V1 and unpaired guanine residues with RNase T1. We also used lead(II) that has a preferential affinity for interhelical and loop regions and a high sensitivity for flexible regions. Particular care was taken to use uniform conditions of salt, magnesium, pH and temperature for the different enzymatic chemical probes. Derived from these experimental data, a three dimensional model of the 5 S rRNA was built using computer modeling which integrates stereochemical constraints and phylogenetic data. The three domains of 5 S rRNA secondary structure fold into a Y-shaped structure that does not accommodate long-range tertiary interactions between domains. The three domains have distinct structural and dynamic features as revealed by the chemical reactivity and the lead(II)-induced hydrolysis: domain 2 (loop B/helix III/loop C) displays a rather weak structure and possesses dynamic properties while domain 3 (helix V/region E/helix IV/loop D) adopts a highly structured and overall helical conformation. Conserved nucleotides are not crucial for the tertiary folding but maintain an intrinsic structure in the loop regions, especially via non-canonical pairing (A.G, G.U, G.G, A.C, C.C), which can close the loops in a highly specific fashion. In particular, nucleotides in the large external loop C fold into an organized conformation leading to the formation of a five-membered loop motif. Finally, nucleotides at the hinge region of the Y-shape are involved in a precise array of hydrogen bonds based on a triple interaction between U14, G69 and G107 stabilizing the quasi-colinearity of helices II and V. The proposed tertiary model is consistent with the localization of the ribosomal protein binding sites and possesses strong analogy with the model proposed for Xenopus laevis 5 S rRNA, indicating that the Y-shape model can be generalized to all 5 S rRNAs.     

An additional long-range interaction in human U1 snRNA.

We present evidence for the existence of an additional long-range interaction in vertebrate U1 snRNAs. By submitting human U1 snRNP, HeLa nuclear extracts, authentic human or X. laevis in vitro transcribed U1 snRNAs to RNase V1, a nuclease specific for double-stranded regions, cleavages occurred in the sequence psi psi ACC (positions 5-9) residing in the 5' terminal region of the RNA. The RNase V1 sensitive region is insensitive to single-stranded probes, something unexpected knowing that it was considered single-stranded in order to base-pair to pre-mRNA 5' splice site. We have identified the sequence GGUAG (positions 132-136) as the only possible 3' partner. Mutants, either abolishing or restoring the interaction between the partners, coupled to an RNase V1 assay, served to substantiate this base-pairing model. The presence of this additional helix, even detected in nuclear extracts under in vitro splicing conditions, implies that a conformational change must occur to release a free U1 snRNA 5' end.     

Proposed secondary structure of eukaryotic U14 snRNA.

U14 snRNA is a small nuclear RNA that plays a role in the processing of eukaryotic ribosomal RNA. We have investigated the folded structure of this snRNA species using comparative analysis of evolutionarily diverse U14 snRNA primary sequences coupled with nuclease digestion analysis of mouse U14 snRNA. Covariant nucleotide analysis of aligned mouse, rat, human, and yeast U14 snRNA primary sequences suggested a basic folding pattern in which the 5' and 3' termini of all U14 snRNAs were base-paired. Subsequent digestion of mouse U14 snRNA with mung bean (single-strand-specific), T2 (single-strand-preferential), and V1 (double-strand-specific) nucleases defined the major and minor cleavage sites for each nuclease. This digestion data was then utilized in concert with the comparative sequence analysis of aligned U14 snRNA primary sequences to refine the secondary structure model suggested by computer-predicted folding. The proposed secondary structure of U14 snRNA is comprised of three major hairpin/helical regions which includes the helix of base-paired 5' and 3' termini. Strict and semiconservative covariation of specific base-pairs within two of the three major helices, as well as nucleotide changes that strengthen or extend base-paired regions, support this folded conformation as the evolutionary conserved secondary structure for U14 snRNA.     

Characterization of an anti-RNA recombinant autoantibody fragment (scFv) isolated from a phage display library and detailed analysis of its binding site on U1 snRNA.

This is the first study in which the complex of a monoclonal autoantibody fragment and its target, stem loop II of U1 snRNA, was investigated with enzymatic and chemical probing. A phage display antibody library derived from bone marrow cells of an SLE patient was used for selection of scFvs specific for stem loop II. The scFv specificity was tested by RNA immunoprecipitation and nitrocellulose filter binding competition experiments. Immunofluorescence data and immunoprecipitation of U1 snRNPs containing U1A protein, pointed to an scFv binding site different from the U1A binding site. The scFv binding site on stem loop II was determined by footprinting experiments using RNase A, RNase V1, and hydroxyl radicals. The results show that the binding site covers three sequence elements on the RNA, one on the 5' strand of the stem and two on the 3' strand. Hypersensitivity of three loop nucleotides suggests a conformational change of the RNA upon antibody binding. A three-dimensional representation of stem loop II reveals a juxtapositioning of the three protected regions on one side of the helix, spanning approximately one helical turn. The location of the scFv binding site on stem loop II is in full agreement with the finding that both the U1A protein and the scFv are able to bind stem loop II simultaneously. As a consequence, this recombinant monoclonal anti-U1 snRNA scFv might be very useful in studies on U1 snRNPs and its involvement in cellular processes like splicing.     

Conservation of functional features of U6atac and U12 snRNAs between vertebrates and higher plants

Splicing of U12-dependent introns requires the function of U11, U12, U6atac, U4atac, and U5 snRNAs. Recent studies have suggested that U6atac and U12 snRNAs interact extensively with each other, as well as with the pre-mRNA by Watson-Crick base pairing. The overall structure and many of the sequences are very similar to the highly conserved analogous regions of U6 and U2 snRNAs. We have identified the homologs of U6atac and U12 snRNAs in the plant Arabidopsis thaliana. These snRNAs are significantly diverged from human, showing overall identities of 65% for U6atac and 55% for U12 snRNA. However, there is almost complete conservation of the sequences and structures that are implicated in splicing. The sequence of plant U6atac snRNA shows complete conservation of the nucleotides that base pair to the 5' splice site sequences of U12-dependent introns in human. The immediately adjacent AGAGA sequence, which is found in human U6atac and all U6 snRNAs, is also conserved. High conservation is also observed in the sequences of U6atac and U12 that are believed to base pair with each other. The intramolecular U6atac stem-loop structure immediately adjacent to the U12 interaction region differs from the human sequence in 9 out of 21 positions. Most of these differences are in base pairing regions with compensatory changes occurring across the stem. To show that this stem-loop was functional, it was transplanted into a human suppressor U6atac snRNA expression construct. This chimeric snRNA was inactive in vivo but could be rescued by coexpression of a U4atac snRNA expression construct containing compensatory mutations that restored base pairing to the chimeric U6atac snRNA. These data show that base pairing of U4atac snRNA to U6atac snRNA has a required role in vivo and that the plant U6atac intramolecular stem-loop is the functional analog of the human sequence.     

Higher order structure of chloroplastic 5S ribosomal RNA from spinach

The secondary and tertiary structure of chloroplastic 5S ribosomal RNA from spinach was investigated by the use of several chemical and enzymatic structure probes. The four bases were monitored at one of their Watson-Crick base-pairing positions with dimethyl sulfate [at A(N1) and C(N3)] and with 1-cyclohexyl-3-(2-morpholinoethyl)carbodiimide metho-p-toluenesulfonate [at G(N1) and U(N3)]. Position N7 of purines was probed with diethyl pyrocarbonate (adenines) and with dimethyl sulfate (guanines). Ethylnitrosourea was used to probe phosphate involved in tertiary interaction or in cation coordination. In order to estimate the degree of stability of helices, the various chemical reagents were employed under "native" conditions (300 mM KCl and 20 mM magnesium at 37 degrees C), under "semidenaturing" conditions [1 mM ethylenediaminetetraacetic acid (EDTA) at 37 degrees C], and under denaturing conditions (1 mM EDTA at 90 degrees C). Unstructured regions were also tested with single-strand-specific nucleases T1, U2, and S1 and double-stranded or stacked regions with RNase V1 from cobra Naja naja oxiana venom. The results confirm the existence of the five helices and the two external loops proposed in the consensus model of 5S rRNA. However, the regions depicted as unpaired internal loops appear to be folded into a more complex conformation. A three-dimensional model derived from the present data and graphic modeling for a region encompassing helix IV, helix V, loop D, and loop E (nucleotides 70-110) is proposed. Nucleotides in the so-called loop E (73-79/100-106) display unusual features: Noncanonical base pairs (A-A and A-G) are formed, and three nucleotides (C75, U78, and U105) are bulging out. This region adopts an unwound and extended conformation that can be well suited for tertiary interactions or for protein binding. Several bases and phosphates candidate for the tertiary folding of the RNA were also identified.     

Base pairing with U6atac snRNA is required for 5' splice site activation of U12-dependent introns in vivo.

The minor U12-dependent class of eukaryotic nuclear pre-mRNA introns is spliced by a distinct spliceosomal mechanism that requires the function of U11, U12, U5, U4atac, and U6atac snRNAs. Previous work has shown that U11 snRNA plays a role similar to U1 snRNA in the major class spliceosome by base pairing to the conserved 5'' splice site sequence. Here we show that U6atac snRNA also base pairs to the 5'' splice site in a manner analogous to that of U6 snRNA in the major class spliceosome. We show that splicing defective mutants of the 5'' splice site can be activated for splicing in vivo by the coexpression of compensatory U6atac snRNA mutants. In some cases, maximal restoration of splicing required the coexpression of compensatory U11 snRNA mutants. The allelic specificity of mutant phenotype suppression is consistent with Watson-Crick base pairing between the pre-mRNA and the snRNAs. These results provide support for a model of the RNA-RNA interactions at the core of the U12-dependent spliceosome that is strikingly similar to that of the major class U2-dependent spliceosome.     

cDNA cloning of U1, U2, U4 and U5 snRNA families expressed in pea nuclei.

Differences observed between plant and animal pre-mRNA splicing may be the result of primary or secondary structure differences in small nuclear RNAs (snRNAs). A cDNA library of pea snRNAs was constructed from anti-trimethylguanosine (m3(2,2,7)G immunoprecipitated pea nuclear RNA. The cDNA library was screened using oligo-deoxyribonucleotide probes specific for the U1, U2, U4 and U5 snRNAs. cDNA clones representing U1, U2, U4 and U5 snRNAs expressed in seedling tissue have been isolated and sequenced. Comparison of the pea snRNA variants with other organisms suggest that functionally important primary sequences are conserved phylogenetically even though the overall sequences have diverged substantially. Structural variations in U1 snRNA occur in regions required for U1-specific protein binding. In light of this sequence analysis, it is clear that the dicot snRNA variants do not differ in sequences implicated in RNA:RNA interactions with pre-mRNA. Instead, sequence differences occur in regions implicated in the binding of small ribonucleoproteins (snRNPs) to snRNAs and may result in the formation of unique snRNP particles.     

Chemical and enzymatic probing of spatial structure of the omega leader of tobacco mosaic virus RNA

The 5′-untranslated sequence of tobacco mosaic virus RNA — the so-called omega leader — exhibits features of a translational enhancer of homologous and heterologous mRNAs. The absence of guanylic residues, the presence of multiple trinucleotide CAA repeats in its central region, and the low predictable probability of the formation of an extensive secondary structure of the Watson-Crick type were reported as the peculiarities of the primary structure of the omega leader. In this work we performed chemical and enzymatic probing of the secondary structure of the omega leader. The isolated RNA comprising omega leader sequence was subjected to partial modifications with dimethyl sulfate and diethyl pyrocarbonate and partial hydrolyses with RNase A and RNase V1. The sites and the intensities of the modifications or the cleavages were detected and measured by the primer extension inhibition technique. The data obtained have demonstrated that RNase A, which attacks internucleotide bonds at the 3′ side of pyrimidine nucleotides, and diethyl pyrocarbonate, which modifies N7 of adenines not involved in stacking interactions, weakly affected the core region of omega leader sequence enriched with CAA-repeats, this directly indicating the existence of a stable spatial structure. The significant stability of the core region structure to RNase A and diethyl pyrocarbonate was accompanied by its complete resistance against RNase V1, which cleaves a polyribonucleotide chain involved in Watson-Crick double helices and generally all A-form RNA helices, thus being an evidence in favor of a non-Watson-Crick structure. The latter was confirmed by the full susceptibility of all adenines and cytosines of the omega polynucleotide chain to dimethyl sulfate, which exclusively modifies N1 of adenines and N3 of cytosines not involved in Watson-Crick interactions. Thus, our data have confirmed that (1) the regular (CAA)_n sequence characteristic of the core region of the omega leader does form stable secondary structure, and (2) the structure formed is not the canonical double helix of the Watson-Crick type.     

Structural changes in the 530 loop of Escherichia coli 16S rRNA in mutants with impaired translational fidelity.

The higher order structure of the functionally important 530 loop in Escherichia coli 16S rRNA was studied in mutants with single base changes at position 517, which significantly impair translational fidelity. The 530 loop has been proposed to interact with the EF-Tu-GTP-aatRNA ternary complex during decoding. The reactivity at G530, U531 and A532 to the chemical probes kethoxal, CMCT and DMS respectively was increased in the mutant 16S rRNA compared with the wild-type, suggesting a more open 530 loop structure in the mutant ribosomes. This was supported by oligonucleotide binding experiments in which probes complementary to positions 520-526 and 527-533, but not control probes, showed increased binding to the 517C mutant 70S ribosomes compared with the non-mutant control. Furthermore, enzymatic digestion of 70S ribosomes with RNase T1, specific for single-stranded RNA, substantially cleaved both wild-type and mutant rRNAs between G524 and C525, two of the nucleotides involved in the 530 loop pseudoknot. This site was also cleaved in the 517C mutant, but not wild-type rRNA, by RNase V1. Such a result is still consistent with a more open 530 loop structure in the mutant ribosomes, since RNase V1 can cut at appropriately stacked single-stranded regions of RNA. Together these data indicate that the 517C mutant rRNA has a rather extensively unfolded 530 loop structure. Less extensive structural changes were found in mutants 517A and 517U, which caused less misreading. A correlation between the structural changes in the 530 loop and impaired translational accuracy is proposed.     

The low-abundance U11 and U12 small nuclear ribonucleoproteins (snRNPs) interact to form a two-snRNP complex.

A novel small nuclear ribonucleoprotein (snRNP) complex containing both U11 and U12 RNAs has been identified in HeLa cell extracts. This U11/U12 snRNP complex can be visualized on glycerol gradients, on native polyacrylamide gels, and by selection with antisense 2'-O-methyl oligoribonucleotides. RNase H-mediated degradation of the U12 snRNA confirmed a direct interaction between the U11 and U12 snRNPs. This snRNP complex is the first to be identified involving low-abundance snRNPs. Selection of the U11/U12 snRNP complex is sensitive to high salt, suggestive of a protein-mediated interaction. Secondary structure analyses revealed several regions of the U11 snRNP accessible for interaction with other RNAs or proteins but no detectable difference between the accessibility of these regions in the U11 monoparticle compared with the U11/U12 snRNP complex. There are also several accessible single-stranded regions in the U12 snRNP, and oligonucleotide-directed RNase H digestion identified nucleotides 28 to 36 of U12 as containing sequences required for the U11/U12 interaction. Both the U12 snRNP and the U11/U12 snRNP complex can be disrupted without altering the cleavage/polyadenylation activity of a nuclear extract.     

Destabilizing mutations alter the hydrogen exchange mechanism in ribonuclease A

The effect of strongly destabilizing mutations, I106A and V108G of Ribonuclease A (RNase A), on its structure and stability has been determined by NMR. The solution structures of these variants are essentially equivalent to RNase A. The exchange rates of the most protected amide protons in RNase A (35°C), the I106A variant (35°C), and the V108G variant (10°C) yield stability values of 9.9, 6.0, and 6.8 kcal/mol, respectively, when analyzed assuming an EX2 exchange mechanism. Thus, the destabilization induced by these mutations is propagated throughout the protein. Simulation of RNase A hydrogen exchange indicates that the most protected protons in RNase A and the V108G variant exchange via the EX2 regime, whereas those of I106A exchange through a mixed EX1 + EX2 process. It is striking that a single point mutation can alter the overall exchange mechanism. Thus, destabilizing mutations joins high temperatures, high pH and the presence of denaturating agents as a factor that induces EX1 exchange in proteins. The calculations also indicate a shift from the EX2 to the EX1 mechanism for less protected groups within the same protein. This should be borne in mind when interpreting exchange data as a measure of local stability in less protected regions.     

Identification and assignment of base pairs in three helical stems of Torulopsis utilis ribosomal 5S RNA and its RNase T1 and RNase T2 cleaved fragments via 500-MHz proton homonuclear overhauser enhancements

Imino proton resonances in the downfield region (10-14 ppm) of the 500-MHz 1H NMR spectrum of Torulopsis utilis 5S RNA are identified (A X U, G X C, or G X U) and assigned to base pairs in helices I, IV, and V via analysis of homonuclear Overhauser enhancements (NOE) from intact T. utilis 5S RNA, its RNase T1 and RNase T2 digested fragments, and a second yeast (Saccharomyces cerevisiae) 5S RNA whose nucleotide sequence differs at only six residues from that of T. utilis 5S RNA. The near-identical chemical shifts and NOE behavior of most of the common peaks from these four RNAs strongly suggest that helices I, IV, and V retain the same conformation after RNase digestion and that both T. utilis and S. cerevisiae 5S RNAs share a common secondary and tertiary structure. Of the four G X U base pairs identified in the intact 5S RNA, two are assigned to the terminal stem (helix I) and the other two to helices IV and V. Seven of the nine base pairs of the terminal stem have been assigned. Our experimental demonstration of a G X U base pair in helix V supports the 5S RNA secondary structural model of Luehrsen and Fox [Luehrsen, K. R., & Fox, G.E. (1981) Proc. Natl. Acad. Sci. U.S.A. 78, 2150-2154]. Finally, the base-pair proton peak assigned to the terminal G X U in helix V of the RNase T2 cleaved fragment is shifted downfield from that in the intact 5S RNA, suggesting that helices I and V may be coaxial in intact T. utilis 5S RNA.     

Structural analyses of the 7SK ribonucleoprotein (RNP), the most abundant human small RNP of unknown function.

The human 7SK ribonucleoprotein (RNP) has been analyzed to determine its RNA secondary structure and protein constituents. HeLa cell 7SK RNA alone and within its RNP have been probed by chemical modification and enzymatic cleavage, and sites of modification or cleavage have been mapped by primer extension. The resulting secondary structure suggests that structural determinants necessary for capping (a 5' stem followed by the sequence AUPuUPuC) and nuclear migration (the sequence AUPuUPuC) of 7SK RNA may be similar to those for U6 small nuclear RNA (snRNA). It also supports existence of a 3' stem structure which could serve to self-prime cDNA synthesis during pseudogene formation. Oligonucleotide-directed RNase H digestion indicated regions of 7SK RNA capable of base pairing with other nucleic acids. Antisense 2'-O-methyl RNA oligonucleotides were used to affinity select the 7SK RNP from an in vivo 35S-labeled cell sonic extract and identify eight associated proteins of 83, 48, 45, 43, 42, 21, 18, and 13 kDa. 7SK RNA has extensive sequence complementarity to U4 snRNA, within the U4/U6 base pairing domain, and also to U11 snRNA. The possibility that the 7SK RNP is an unrecognized component of the pre-mRNA processing machinery is discussed.     

Solution structure of mRNA hairpins promoting selenocysteine incorporation in Escherichia coli and their base-specific interaction with special elongation factor SELB.

On the basis of chemical probing data, the solution structures of RNA hairpins within fdhF and fdnG mRNAs in Escherichia coli, which both promote selenocysteine incorporation at UGA codons, were derived with the help of computer modeling. We find that these mRNA hairpins contain two separate structural domains that possibly also exert two different functions. The first domain is comprised of the UGA codon, which is included within a complex and distorted double-stranded region. Thereby, release factor 2 might be prevented from binding to the UGA codon to terminate protein synthesis. The second domain is located within the apical loop of the mRNA hairpin structures. This loop region exhibits a defined tertiary structure in which no base is involved in Watson-Crick interactions. The structure of the loop is such that, following a sharp turn after G22 (A22 in fdnG mRNA), bases G23 and U24 are exposed to the solvent on the deep groove side of the supporting helix. Residues C25 and U26 close the loop with a possible single H-bonding interaction between the first and last residues of the loop, 04(U26) and N6(A21). The bulge residues U17 and U18 (in fdhF mRNA), or Ul7 only in fdnG mRNA, point their Watson-Crick positions in the same direction as loop residues G23 and U24 do, and at the same time open up the deep groove at the top of the hairpin helix. Chemical probing data demonstrate that bases G23 and U24 in both mRNA hairpins, as well as residues U17 and Ul7/U18 (for fdhF mRNA) located in a bulge 5' to the loop, are involved directly in binding to special elongation factor SELB in both mRNAs. Therefore, SELB recognizes identical bases within both mRNA hairpins despite differences in their primary sequence, consistent with the derived 3D models for these mRNAs, which exhibit similar tertiary structures. Binding of SELB to the fdhF mRNA hairpin was estimated to proceed with an apparent Kd of 30 nM.     

Direct binding of small nuclear ribonucleoprotein G to the Sm site of small nuclear RNA. Ultraviolet light cross-linking of protein G to the AAU stretch within the Sm site (AAUUUGUGG) of U1 small nuclear ribonucleoprotein reconstituted in vitro.

The major small nuclear ribonucleoproteins (snRNPs) U1, U2, U5 and U4/U6 participate in the splicing of pre-mRNA. U1, U2, U4 and U5 RNAs share a highly conserved sequence motif PuA(U)nGPu, termed the Sm site, which is normally flanked by two hairpin loops. The Sm site provides the major binding site for the group of common proteins, B', B, D1, D2, D3, E, F and G, which are shared by the spliceosomal snRNPs. We have investigated the ability of common snRNP proteins to recognize the Sm site of snRNA by using ultraviolet light-induced RNA-protein cross-linking within U1 snRNP particles. The U1 snRNP particles, reconstituted in vitro, contained U1 snRNA labelled with 32P. Cross-linking of protein to this U1 snRNA occurred only in the presence of the single-stranded stretch of snRNA that makes up the conserved Sm site. Characterization of the cross-linked protein by one and two-dimensional gel electrophoresis indicated that snRNP protein G had become cross-linked to the U1 snRNA. This was confirmed by specific immunoprecipitation of the cross-linked RNA-protein complex with an anti-G antiserum. The cross-link was located on the U1 snRNA by fingerprint analysis with RNases T1 and A; this demonstrated that the protein G has been cross-linked to the AAU stretch within the 5'-terminal half of the Sm site (AAUUUGUGG). These results suggest that the snRNP protein G may be involved in the direct recognition of the Sm site.