On the structure of yeast tRNA Phe . Complementary-oligonucleotide binding studies

Structural insights into translational fidelity

The underlying basis for the accuracy of protein synthesis has been the subject of over four decades of investigation. Recent biochemical and structural data make it possible to understand at least in outline the structural basis for tRNA selection, in which codon recognition by cognate tRNA results in the hydrolysis of GTP by EF-Tu over 75 A away. The ribosome recognizes the geometry of codon-anticodon base pairing at the first two positions but monitors the third, or wobble position, less stringently. Part of the additional binding energy of cognate tRNA is used to induce conformational changes in the ribosome that stabilize a transition state for GTP hydrolysis by EF-Tu and subsequently result in accelerated accommodation of tRNA into the peptidyl transferase center. The transition state for GTP hydrolysis is characterized, among other things, by a distorted tRNA. This picture explains a large body of data on the effect of antibiotics and mutations on translational fidelity. However, many fundamental questions remain, such as the mechanism of activation of GTP hydrolysis by EF-Tu, and the relationship between decoding and frameshifting.

Selection of tRNA by the ribosome requires a transition from an open to a closed form

A structural and mechanistic explanation for the selection of tRNAs by the ribosome has been elusive. Here, we report crystal structures of the 30S ribosomal subunit with codon and near-cognate tRNA anticodon stem loops bound at the decoding center and compare affinities of equivalent complexes in solution. In ribosomal interactions with near-cognate tRNA, deviation from Watson-Crick geometry results in uncompensated desolvation of hydrogen-bonding partners at the codon-anticodon minor groove. As a result, the transition to a closed form of the 30S induced by cognate tRNA is unfavorable for near-cognate tRNA unless paromomycin induces part of the rearrangement. We conclude that stabilization of a closed 30S conformation is required for tRNA selection, and thereby structurally rationalize much previous data on translational fidelity.

Determining the influence of structure on hybridization using oligonucleotide arrays

We have studied the effects of structure on nucleic acid heteroduplex formation by analyzing hybridization of tRNAphe to a complete set of complementary oligonucleotides, ranging from single nucleotides to dodecanucleotides. The analysis points to features in tRNA that determine heteroduplex yield. All heteroduplexes that give high yield include both double-stranded stems as well as single-stranded regions. Bases in the single-stranded regions are stacked onto the stems, and heteroduplexes terminate at potential interfaces for coaxial stacking. Heteroduplex formation is disfavored by sharp turns or a lack of helical order in single-stranded regions, competition from bases displaced from a stem, and stable tertiary interactions. The study is relevant to duplex formation on oligonucleotide microarrays and to antisense technologies.

Secondary structure dimorphism and interconversion between hairpin and duplex form of oligoribonucleotides

RNA hairpins can alternatively form a dimer with a bulged loop flanked by regularly base paired regions. [1H]NMR spectroscopy and native gel electrophoresis were used to study how the sequence of nucleotides in the loop of the hairpin affect the hairpin-duplex interconversion. As a model system, a hairpin containing 7 nucleotides in the loop and 5 base pairs in the stem was used. The loop size was gradually reduced from 7 to 4 nucleotides, yielding finally the stable UNCG tetraloop. Single nucleotide mutations were performed to investigate the influence of the self-complementarity of the loop sequence on the dimerization. The results demonstrate that (1) the initial fraction of hairpin is determined by concentration of the oligonucleotide, the annealing procedure, and the relative stability of the loop, (2) the degree of self-complementarity of the loop sequence of the hairpin governs the dimerization kinetics, and (3) oligonucleotides complementary to the loop sequence decrease the dimerization rate. We propose a secondary structure-based model for the dimerization reaction of RNA hairpins in which the formation of intermolecular base pairs between self-complementary nucleotides of the loops represents the nucleation step.

Coding coenzyme handles: a hypothesis for the origin of the genetic code

The coding coenzyme handle hypothesis suggests that useful coding preceded translation. Early adapters, the ancestors of present-day anticodons, were charged with amino acids acting as coenzymes of ribozymes in a metabolically complex RNA world. The ancestral aminoacyl-adapter synthetases could have been similar to present-day self-splicing tRNA introns. A codon-anticodon-discriminator base complex embedded in these synthetases could have played an important role in amino acid recognition. Extension of the genetic code proceeded through the take-over of nonsense codons by novel amino acids, related to already coded ones either through precursor-product relationship or physicochemical similarity. The hypothesis is open for experimental tests.

Nuclear magnetic resonance study of the codon-anticodon interaction in Bombyx mori tRNA(GCCGly)

NMR spectra of Bombyx mori tRNA(GCCGly) were recorded in the GCC absence and presence of the oligonucleotide, GGCUp, which contains the codon sequence, GGC. The difference between the spectra with and without the codon oligonucleotide indicates the appearance of five new imino proton peaks. For the assignment of these peaks, G*GCUp, in which the 5'-terminal G was enriched with 95% 15N, was prepared (G*, 15N-labeled guanosine). In the imino proton spectrum of B. mori tRNA(GCCGly) on the addition of G*GCUp, the peak at 12 ppm became a doublet due to coupling with 15N nuclei. In the two-dimensional 1H-15N heteronuclear multiple-quantum correlation (HMQC) spectrum, only the peak at 12 ppm was observed, and thus it was assigned to the imino proton of the 5'-terminal G of GGCUp interacting with tRNA(GCCGly). Judging from the temperature effect and chemical shifts, the five new imino proton peaks are presumed to be due to three G.C base pairs, induced by the codon-anticodon interaction, and one U.U base pair, induced by an interaction between the 3' terminal U of GGCUp and U33 neighboring the anticodon. The binding of three trinucleotides (GGCp, GGUp and GCUp) to B. mori tRNA(GCCGly) was also investigated. Ultracentrifugation analysis showed that tRNA(GCCGly) underwent dimerization through the anticodon-anticodon interaction, but the dimerization was broken on addition of GGCUp. On 1H-NMR and ultracentrifugation analysis, it was found that GCUp not complementary to the anticodon also binds to the anticodon loop.

Destabilization of codon-anticodon interaction in the ribosomal exit site

The affinities of the exit (E) site of poly(U) or poly(A)-programmed Escherichia coli ribosomes for the respective cognate tRNA and a number of non-cognate tRNAs were determined by equilibrium titrations. Among the non-cognate tRNAs, the binding constants vary up to about tenfold (10(6) to 10(7) M-1 at 20 mM-Mg2+) or 50-fold (10 mM-Mg2+), indicating that codon-independent binding is modulated to a considerable extent by structural elements of the tRNA molecules other than the anticodon. Codon-anticodon interaction stabilizes tRNA binding in the E site approximately fourfold (20 mM-Mg2+) or 20-fold (10 mM-Mg2+), corresponding to delta G degree values of -3 and -7 kJ/mol (0.7 and 1.7 kcal/mol), respectively. Thus, the energetic contribution of codon-anticodon interaction to tRNA binding in the E site appears rather small, particularly in comparison to the large effects on the binding in A and P sites and to the binding of complementary oligonucleotides or of tRNAs with complementary anticodons. This result argues against a role of the E site-bound tRNA in the fixation of the mRNA on the ribosome. In contrast, we propose that the role of the E site is to facilitate the release of the discharged tRNA during translocation by providing an intermediate, labile binding site for the tRNA leaving the P site. The lowering of both affinity and stability of tRNA binding accompanying the transfer of the tRNA from the P site to the E site is predominantly due to the labilization of the codon-anticodon interaction.

Complementary base-pairing properties of cyclized and linear oligonucleotides

Oligouridylates of varying chain lengths were synthesized by polynucleotide phosphorylase and cyclized by RNA ligase. Over chain lengths from 7 to 15, the bindings of the cyclized and linear oligomers to polyadenylate were measured on the basis of differential migration of bound and free oligomers on a gel exclusion column. Binding of the cyclized oligomers was found to be far weaker than that of their linear counterparts of equal length. Such a general reduction in base-pairing capacity due to the cyclized conformation, by limiting the strength of unintended base-pairing without obstructing the possible development of strong specific base-pairing, may represent an advantage important to the function and evolution of loop structures in tRNA and other RNA molecules.

Fluorine-19 nuclear magnetic resonance study of codon-anticodon interaction in 5-fluorouracil-substituted E. coli transfer RNAs

Codon-anticodon interaction was investigated in fully active 5-fluorouracil-substituted E. coli tRNAVal1 (anticodon FAC) by 19F NMR spectroscopy. Binding of the codon GpUpA results in the upfield shift of a 19F resonance at 3.9 ppm in the central region of the 19F NMR spectrum, whereas trinucleotides not complementary to the anticodon have no effect. The same 19F resonance shifts upfield upon formation of an anticodon-anticodon dimer between the 19F-labeled tRNA and E. coli tRNATyr2 (anticodon QUA). These results permit assignment of the peak at 3.9 ppm to the 5-fluorouracil at position 34 in the anticodon of fluorouracil-substituted tRNAVal1. The methionine codon ApUpG also causes a sequence-specific upfield shift of a peak in the central part of the 19F NMR spectrum of fluorinated E. coli tRNAMetm. However, ApUpG has no effect on the 19F spectrum of 19F-labeled E. coli tRNAMetf, indicating possible conformational differences between the anticodon loop of initiator and chain-elongating methionine tRNAs. 19F NMR experiments detect no binding of CpGpApA to the complementary FpFpCpG (replaces Tp psi pCpG) in the T-loop of 5-fluorouracil-substituted tRNAVal1, in the presence or absence of codon, suggesting that the tertiary interactions between the T- and D-loops are not disrupted by codon-anticodon interactions.

Reversibility of cyclization of the Tetrahymena rRNA intervening sequence: implication for the mechanism of splice site choice

The Tetrahymena rRNA intervening sequence (IVS) excises itself from the pre-rRNA and then mediates its own cyclization. We now find that certain di- and trinucleotides with free 3' hydroxyl groups reopen the circular IVS at the cyclization junction, producing a linear molecule with the oligonucleotide covalently attached to its 5' end. This linear molecule recyclizes with release of the added oligonucleotide. Thus the IVS RNA, like an enzyme, lowers the activation energy for both forward and reverse cleavage-ligation reactions. Certain combinations of pyrimidines are required for circle reopening. The most reactive oligonucleotide is UCU. This sequence resembles those preceding the major and minor cyclization sites in the linear IVS RNA (UUU and CCU) and the 5' splice site in the pre-rRNA (UCU). We propose that an oligopyrimidine binding site within the IVS binds the sequences upstream of each of these target sites for cleavage-ligation.

Anticodon-anticodon interactions in solution. Studies of the self-association of yeast or Escherichia coli tRNAAsp and of their interactions with Escherichia coli tRNAVal

The temperature-jump method was used to measure the thermodynamic and kinetic parameters of the yeast tRNAAsp (anticodon GUC) duplex, which involves a U/U mismatch in the middle position of the quasi self-complementary anticodon, and of the yeast tRNAAsp (GUC)-Escherichia coli tRNAVal (GAC) complex, in which the tRNAs have complementary anticodons. The existence of the tRNAAsp duplex involving GUC-GUC interactions as evidenced in the crystal structure has now been demonstrated in solution. However, the value of its association constant (Kass = 10(4)M-1 at 0 degrees C) is characteristic of a rather weak complex, when compared with that between tRNAAsp and tRNAVal (Kass = 4 X 10(6) M-1 at 0 degrees C), the effect being essentially linked to differences in the rate constant for dissociation. tRNAAsp split in the anticodon by T1 ribonuclease gives no relaxation signal, indicating that the effects observed with intact tRNA were entirely due to anticodon interactions. No duplex formation was observed with other tRNAs having quasi self-complementary GNC anticodons (where N is C, A or G), such as E. coli tRNAGly (GCC), E. coli tRNAVal (GAC) or E. coli tRNAAla (GGC). This is compatible with the idea that, probably as in the crystal structure, a short double helix is formed in solution between the two GUC anticodons. Because of steric effects, such a complex formation would be hindered if a cytosine, adenine or guanine residue were located in the middle position of the anticodon. Escherichia coli tRNAAsp possessing a modified G residue, the Q base, at the first position of the anticodon, showed a weaker self-association than yeast tRNAAsp but its complex with E. coli tRNAVal was found to be only 1.5 times less stable than that between yeast tRNAAsp and E. coli tRNAVal. Temperature-jump experiments conducted under conditions mimicking those used for the crystallization of yeast tRNAAsp (in the presence of 1.6 M-ammonium sulphate and 3mM-spermine) revealed an important stabilization of the yeast and E. coli tRNAAsp duplexes or of their complexes with E. coli tRNAVal. The effect is due exclusively to ammonium sulphate; it is entropy driven and its influence is reflected on the association rate constant; no influence on the dissociation rate constant was observed. For all tRNA-tRNA complexes, the melting temperature upon addition of ammonium sulphate was considerably increased. This study permits the definition of solution conditions in which tRNAs with appropriate anticodons exist mainly as anticodon-anticodon dimers.

Mechanism of codon recognition by transfer RNA studied with oligonucleotides larger than triplets

The binding of yeast tRNAPhe to UUCA, UUCC, UUCCC, UUCUUCU, U4, U5, U6 and U7 was analysed by fluorescence temperature jump and equilibrium sedimentation measurements. In all cases the two observed relaxation processes can be assigned to alpha) an intramolecular conformation change of the anticodon loop and beta) preferential binding of the oligonucleotides to one of the anticodon conformations. The anticodon loop transition is associated with inner sphere complexation of Mg2+ and proceeds with rate constants of about 10(3) s-1. The rate constants of oligonucleotide binding are between 4 and 10 X 10(6) M-1s-1 and reflect an increase of the association rate with the number of binding sites compensated to some degree by electrostatic repulsion in the preequilibrium complex. Neither temperature jump nor equilibrium sedimentation experiments provided evidence for UUCA or UUCC induced tRNA dimerisation, although UUC binding leads to strong tRNA dimerisation under equivalent conditions. The results obtained for the longer oligonucleotides are similar. In the case of UUCUUCU with its two potential binding sites for tRNAPhe there was no evidence for the formation of 'ternary' complexes. Apparently tRNAPhe binds preferentially to the second UUC of this 'messenger' and forms additional contacts with residues on either side of the codon. Some evidence for the formation of ternary complexes is obtained for U6 and U7, although the extent of this reaction remains very small. Our results demonstrate that the mode of tRNA binding to a codon is strongly influenced by residues next to the codon. The formation of cooperative contacts between tRNA molecules at adjacent codons apparently requires support by a catalyst adjusting an appropriate conformation of messenger and tRNA molecules.

Molecular model of ribosome frameshifting

Normal tRNAs can cause two- and four-base translocation errors by mistranslating certain noncognate codons. Several cases have been reported that reveal the identities of the frameshift-provoking tRNAs--i.e., the shifty tRNAs--and the mRNA sequence encompassing the site of frameshifting--i.e., the shifty codons. Here, a striking uniformity between the anticodon loops of the shifty tRNAs and their shifty codons is described. Stable "offset" anti-codon X codon pairs are postulated for each of the shifty tRNA-shifty codon combinations. This offset anticodon X codon pair leads to two- and four-base translocations when viewed in terms of the reciprocating ratchet mechanism of translocation.

Structure of the ribotrinucleoside diphosphate codon UpUpC bound to tRNAPhe from yeast. A time-dependent transferred nuclear Overhauser enhancement study

The structure of the ribotrinucleoside diphosphate UpUpC, the codon for phenylalanine, bound to yeast tRNAPhe in solution is elucidated using time-dependent proton-proton transferred nuclear Overhauser enhancement measurements to determine distances between bound ligand protons. The glycosidic bond and ribose conformations are low anti and 3'-endo, respectively, typical of an A-RNA type structure. The main chain torsion angles are all within the range of those expected for A-RNA but small differences from those in conventional A-RNA 11 result in a special structure with a larger rotation per residue (40 to 45 degrees compared to 32.7 degrees in R-RNA 11) and almost perfect stacking of the bases. These two structural features, which are similar to those found in the anticodon triplet of the monoclinic crystal form of tRNAPhe, can account for the known greater stability of the codon-anticodon complex relative to an equivalent double helical RNA trimer with a conventional A-RNA structure.

Codon-induced transfer RNA association. A property of transfer RNA involved in its adaptor function?

It is shown by equilibrium sedimentation that the binding of cognate codons to tRNAPhe (yeast), tRNAPhe (Escherichia coli), tRNALys, tRNAfMet and of the wobble codon UUU to tRNAPhe (yeast) induces dimerization of codon transfer RNA complexes. Analysis of the sedimentation profiles with a quantitative evaluation of the coupling between sedimentation and association equilibrium provides dimerization constants in the range from 1 X 10(4) to 6 X 10(4) M-1. These results on various tRNAs from different organisms suggest that the codon-induced tRNA association is a general phenomenon. Probably the codon-induced tRNA association facilitates the aminoacyl transfer reaction.

Codon:anticodon and anticodon:anticodon interaction: evaluation of equilibrium and kinetic parameters of complexes involving a g:u wobble

In order to learn about the effect of the G:U wobble interaction we characterized to codon:anticodon binding between triplets: UUC, UUU and yeast tRNAPhe (anticodon GmAA) as well as the anticodon:anticodon binding between Escherichia coli tRNAGlu2, E. coli tRNALys (anticodons: mam5s2UUC, and mam5S2UUU, respectively) and tRNAPhe from yeast and E. coli (anticodon GAA) using equilibrium fluorescence titrations and temperature jump measurements with fluorescence and absorption detection. The difference in stability constants between complexes involving a G:U pair rather than a usual G:C basepair is in the range of one order magnitude and is mainly due to the shorter lifetime of the complex involving G:U in the wobble position. This difference is more pronounced when the codon triplet is structured, i.e., is built in the anticodon loop of a tRNA. The reaction enthalpies of the anticodon:anticodon complexes involving G:U mismatching were found to be about 4 kcal/mol smaller, and the melting temperatures more than 20 degrees C lower, than those of the corresponding complexes with the G:C basepair. The results are discussed in terms of different strategies that might be used in the cell in order to minimize the effect of different lifetimes of codon-tRNA complexes. Differences in these lifetimes may be used for the modulation of the translation efficiency.

Conformational changes of yeast tRNAphe as monitored by 31P NMR

The 31P NMR spectra of tRNAs contain approximately 17 resonances resolved from the main resonance which consists of about 80% of the total resonance intensity arising from the sugar phosphate backbone. In the present paper we study the behavior of the 31P resonances of yeast tRNAPhe as a function of temperature and of solution conditions. By comparison with other melting experiments we show that three resonances (called c, e and j2) belong to phosphates in the anticodon loop, while the remaining resolved 31P resonances come from phosphates in specific conformations in the central part of the molecule imposed by the tertiary structure. These conformations are different from the normal g-,g- conformation found in A-RNA double helices. The assignments are in good agreement with those previously made on the basis of chemical and enzymatic modification experiments [P. J. M. Salemink, T. Swarthof & C. W. Hilbers (1979) Biochemistry, 18, 3477-3485]. AT high Mg2+ concentrations the anticodon loop is found to be present in two different conformations. For all solution conditions studied loss of the anticodon loop structure takes place before the tertiary structure is melted out. The melting of the tertiary structure is not strictly an all- or-none process. The lifetimes of phosphate conformations involved in the tertiary structure may differ by at least a factor of two. It can also be concluded that the range of chemical shifts observed for phosphodiesters cannot at the moment be accounted for by theoretical calculations.

The effect of chemical modification of the CCA end of yeast tRNAPhe on its biological activity on ribosomes

Yeast tRNAPhe containing 2-thiocytidine (s2C) at position 75 was alkylated specifically at this residue. The biological activities of alkylated and native tRNAPhe were compared in an Escherichia coli protein-synthesizing system in vitro. The alkylated tRNAPhe proved to be active in all steps involved in the elongation phase but the rate of the peptide transfer reaction was somewhat lower when the alkylated tRNAPhe acted as an acceptor of peptidyl residues as compared to native tRNAPhe. These results raise the possibility for attaching spectroscopic or affinity labels at the s2C-75 residue of tRNAPhe without impairing the activity of the tRNA.

The nucleotide sequence of two silk gland alanine tRNAs: implications for fibroin synthesis and for initiator tRNA structure

The nucleotide sequences of two major alanine tRNAs from the Bombyx mori posterior silk gland have been determined. One of these tRNAs appears to be specific to the silk gland, where its accumulation is associated with the rapid production of fibroin. Both sequences are identical, with the exception of a single nucleotide in the anticodon stem. A striking feature of both alanine tRNAs is that loop IV contains sequences previously believed to be restricted to initiator tRNA.

Comparison of the reactions of chemically reactive analogs of U-G-A and of A-U-G with ribosomes of Escherichia coli

The chemically reactive analog of U-G-A, 5'-(4-(Bromo-[2-14C] acetamido) phenylphospho) - uridylyl-(3'-5') - guanylyl-(3'-5') adenosine has a 20 fold lower affinity to 70S ribosomes than the corresponding analog of A-U-G though the U-G-A analog also preferentially reacts with protein S18 of 70S ribosomes. This reaction programs ribosomes for EF-T dependent Trp-tRNATrp-suIII binding. Therefore, it is concluded that this protein is part of the A'-site of the ribosomal codon binding site. Reaction of the U-G-A analog with 30S subunits lead to a predominant crosslinking of U-G-A to proteins S4 and S18. In contrast, a comparable reaction of the A-U-G analog with 30S subunits lead to a predominant crosslinking of A-U-G to proteins S4 and S12 (Pongs, O., Stoffler, G.A., Lanka, E., (1975) J. Mol. Biol. 99, 301). Since protein S12 is located at the 'P' site of the ribosomal codon binding site, it is proposed that the U-G-A analog does not bind at this site.

Codon-dependent rearrangement of the three-dimensional structure of phenylalanine tRNA, exposing the T-psi-C-G sequence for binding to the 50S ribosomal subunit

Codon-anticodon interaction induces an allosteric rearrangement of the three-dimensional structure of Phe-tRNAPhe that exposes the T-psi-C-G sequence for binding to the C-G-A-A sequence of the 5S rRNA within the 50S ribosomal subunit. The conformational change in the tRNAPhe structure was followed by the binding of C-G-[3H]A-[3H]A to the T-psi-C-G sequence, as measured by equilibrium dialysis at 10 mM Mg2+. C-G-A-A (14 pmol) was bound to tRNAPhe in the complete system containing elongation factor Tu-GTP-Phe-tRNA-(uridylyl-3',5')7-uridine-30S ribosomes (100 pmol). At a Mg2+ concentration lower than 5 mM the rearrangement was dependent on elongation factor-Tu, whereas GTP could be replaced by guanylyl imidodiphosphonate. In the absence of elongation factor-Tu-GTP a sigmoidal C-G-A-A binding curve with respect to Mg2+ concentration was obtained, showing half-saturation at 6 mM Mg2+. To achieve the change in the tRNAPhe structure in the absence of 30S ribosomes, a twofold higher concentration of (uridylyl-3',5')7-uridine had to be used. A sigmoidal curve was obtained again when the Mg2+ dependence of the C-G-A-A binding was followed, with 12 pmol of C-G-A-A being bound to 200 pmol of Phe-tRNA. Since T-psi-C-G exposure should influence the binding of Phe-tRNA to 70S ribosomes, Phe-tRNA binding to 70S ribosomes was examined. In the "nonenzymatic" binding (i.e., no elongation factor-Tu-GTP) of Phe-tRNA a sigmoidal Mg2+ dependence was found, whereas the "enzymatic" binding (elongation factor-Tu-GTP present) showed a hyperbolic curve. With 30S ribosomes as controls, only hyperbolic binding curves were found. The Mg2+ dependence of AA-tRNA binding thus reflects the rearrangement of the tRNA structure.

The binding of complementary oligoribonucleotides to yeast initiator Transfer RNA

Oligoribonucleotide binding to baker's yeast initiator tRNA was measured by equilibrium dialysis in order to determine which regions of the tRNA were free to bind complementary oligomers and which were involved in secondary and tertiary structure. Association constants of trinucleoside diphosphates and tetranucleoside triphophates complementary to the single-stranded regions of the cloverleaf structure of yeast tRNAfMet were measured at o degrees in 1.0 M NaCl, and 0.01 M MgCl2. The only regions of the tRNA whose complementary oligomers bound to the tRNA were the amino acid acceptor end and the five nucleotides at the 5' end of the anticodon loop. These results differ from those for the other tRNAs studied by this technique; usually oligomers complementary to the dihydrouracil loop bind to the tRNA. The sequence of yeast tRNAfMet and other eucaryotic initiators is unusual. The "TpsiC loop" contains the sequence A-U-C instead of T-psi-C, yet the binding pattern to the THE TpsiC LOOP IS LIKE THAT FOR OTHER TRNAs; no oligomers bind.

Complementary binding of oligonucleotides with 16S RNA and ribosomal ribonucleoproteins

The accessibility of single-stranded sequences in 16S RNA in free state and in ribonucleoprotein particles (RNP) to complementary binding with isoplith fractions of oligonucleotides was studied. RNP had different protein composition and corresponded to intermediate stages of E. coli 30S subunit assembly in vitro. Gel-filtration was used to detect the most strong binding. It was found that S4 essentially inhibited the hexamer binding to RNA. 'Core' proteins bound to 16S RNA strongly increased the shielding of single-stranded regions while 'split' proteins insignificantly changed the hexamer binding. Nevertheless evidence is presented that 'split' proteins might also interact directly with 16S RNA in the 30S subunit.