¹ Max-Planck-Institut für Molekulare Genetik, AG Ribosomen, Ihnestraße 73, D14195 Berlin , Germany and ² Department of Chemistry, Moscow State University, 119899 Moscow , Russia

Received March 29, 1996; Revised and Accepted May 6, 1996

ABSTRACT

The contacts of phosphate groups in mRNAs with ribosomes were studied. Two mRNAs were used: one mRNA contained in the middle two defined codons to construct the pre- and the post-translocational states, the other was a sequence around the initiation site of the natural cro-mRNA. Phosphorothioate nucleotides were randomly incorporated at a few A, G, U or C positions during in vitro transcription. Iodine can cleave the thioated positions if they are not shielded by ribosomal components. Only a few minor differences in iodine cleavage of ribosome bound and non-bound mRNA were observed: the nucleotide two positions upstream of the decoding codons (i.e. those codons involved in codon-anticodon interactions) showed a reduced accessibility for iodine and the nucleotide immediately following the decoding codons an enhanced accessibility in both elongating states. In initiating ribosomes where the mRNA contained a strong Shine-Dalgarno sequence, at least five phosphates were additionally slightly protected covering the Shine-Dalgarno sequence and nucleotides downstream including the initiator AUG in the P site (A1, G3, G-2, G-5 and A-7). The low contact levels of the phosphates in the mRNA with the elongating ribosome strikingly contrast with the pronounced contact patterns previously described for tRNAs. The data obtained in this study, as well as results of previous studies, suggest that mRNA regions downstream and upstream of decoding codons form only weak contacts with ribosomal components and that the mRNA thus is mainly fixed by codon-anticodon interaction on the elongating ribosome.

INTRODUCTION

The elongating ribosome from Escherichia coli covers a sequence of 39 +- 3 nucleotides (nt) of the mRNA ( 1 and references therein). The precise path of the mRNA through the ribosome is still a point of controversial discussion, although it is well documented that the decoding codons (the codons engaged in codon-anticodon interaction) and the Shine-Dalgarno sequence are located at the neck of the 30S subunit separating the head and the body of this subunit ( 2 , 3 , for review see ref. 4 ).

Various data are now available concerning the mRNA contacts with the ribosome, which were obtained by cross-linking ( 5 , 6 ), toe-printing ( 7 ) and chemical modification techniques ( 8 - 10 ). In most cases the interaction of mRNA bases with the ribosome was investigated. In particular cross-linking experiments showed that there are several specific contacts of this type between mRNA and 16S rRNA either upstream or downstream of the AUG initiation codon. However, since the translation of the mRNAs on the ribosomes occurs independently of their primary structure (although with different efficiency) one can assume that interactions of the sugar-phosphate backbone of the mRNA rather than those of the bases should play the more important role in holding the mRNA on the small ribosomal subunit during the translational process. Indeed, most of mRNA bases downstream from the initiation codon and some bases in the spacer region between the AUG codon and SD sequences were shown do be accessible to attack with different chemical probes ( 8 - 10 ). On the other hand mRNA sugar moieties in these regions appeared to be protected from chemical modification ( 8 , 10 ).

The aim of this study was to analyze whether tight contacts between phosphate residues of mRNA and the ribosomal components exist at different stages of the translational process. For this purpose we have used the procedure developed by Eckstein and co-workers ( 11 , 12 ) for the study of the interaction of tRNA phosphates with aminoacyl-tRNA-synthetase. Recently this method was also applied successfully to the identification of tRNA contacts with the translating ribosome ( 13 ).

The method is based on the statistical substitution of phosphate groups in RNA molecules by phosphorothioate groups. Such modified RNA molecule can be cleaved with iodine at the modified positions unless the access of iodine is prevented by another molecule, in our case by a ribosomal component. It has been shown that phosphorothioate-containing mRNAs have an activity in in vitro systems of protein synthesis comparable to that of unmodified mRNA ( 14 ).

In this study we have used two different mRNA sequences: one contained two codons (AUG-UUC) in the middle to model the elongating ribosome, and the other contained the sequence of the initiation site of the natural mRNA of the lambda phage Cro protein. The experiments have been performed in the initiation state of ribosomes and the two elongation states, viz . the pre- and the post-translocational states. Only weak protections could be found. In the elongation complexes the phosphate group two positions upstream of the decoding codons was protected, whereas in the initiation complex phosphate groups in the Shine-Dalgarno sequence were also protected. No protection could be seen downstream of the decoding codons.

MATERIALS AND METHODS

Ultrapure rNTPs and pre-packed spun columns were from Pharmacia; phosphorothio-NTPs-NEN from DuPont; alkaline phosphatase and T4 polynucleotide kinase from Boehringer-Mannheim; roman {{t R N A} sub f sup {m e t}} and tRNA ^Phe from Subriden; SSP1 restriction endonuclease from Biolabs; [[gamma]- ³² P]ATP from Amersham and cellulose nitrate filters (0.45 [mu]m) from Sartorius. Oligodeoxyribonucleotides used for T7 transcription were synthesized by Dr V. Tishkov at the Department of Chemistry, Moscow State University. 70S ribosomes, EF-G and S100 enzymes were prepared according to ( 15 ). The T7 RNA polymerase was isolated from E.coli strain BL21 containing the plasmid pAR1219 according to published procedures ( 16 ). Plasmid pT7 4.3 containing the gene for the MF-mRNA was a kind gift of Dr F. Triana.

Transcription of mRNAs

The T7 dependent transcription of the MF-mRNA followed the large-scale procedure described in ( 17 ) supplemented with phosphorothioate nucleotides: 40 [mu]g/ml of the plasmid pT7 4.3; 3.8 mM ATP, GTP and CTP each, 0.38 M UTP; in each sample 0.76 mM phosphorothiotriphosphate ([alpha]S-GTP, [alpha]S-ATP or [alpha]S-CTP) or 0.076 mM [alpha]S-UTP; 5 mM DTE, 104 [mu]g/ml BSA, 40 mM Tris-HCl, pH 8.8, 22 mM MgCl ₂ , 10 mM spermidine, 504 U/ml pyrophosphatase and 11 [mu]g/ml T7 polymerase. The reaction volume was 500 [mu]l and the mixture was incubated at 37^oC for 4 h. After phenol extraction the mRNA was purified by gel electrophoresis. The average yield of transcription was 8 [mu]g (500 pmol) from 1 [mu]g (0.6 pmol) of the plasmid.

In the case of the cro-mRNA an ss-DNA fragment (length 51 nt) which codes for the mRNA was annealed with a desoxy-primer (18 nt) as described in ( 5 ). Then the other components were added, establishing the final concentrations described for the transcription of MF-mRNA. Other procedures for isolation and purification of the cro-mRNA were as described above for the MF-mRNA. The yield of transcription was ~450 [mu]g (80.000 pmol) for 1 [mu]g of ssDNA, i.e . 640 pmol from 1 pmol ssDNA.

5 ' ³² P labeling of the mRNAs

For dephosphorylation 200 pmol mRNA in 12 [mu]l water was mixed with 2 [mu]l of a `10 * buffer' (500 mM Tris-HCl, pH 8.5, 1 mM EDTA) and 6 U (6 [mu]l) calf alkaline phosphatase. The reaction mixture was incubated at 25^oC for 2 h, then diluted with water to a volume of 150 [mu]l. After treatment with an equal volume of phenol-water the mRNA was precipitated by ethanol, washed with 70% ethanol, dissolved in 10 [mu]l water and used for ³² P labeling.

2.5 [mu]l kinase buffer (500 mM Tris-HCl, pH 8.2, 100 mM MgCl ₂ , 50 mM DTT, 1 mM EDTA, 1 mM spermidine) was added to 10 [mu]l dephosphorylated mRNA in water (see above), then 10 [mu]l of [[gamma]- ³² P]ATP (10 [mu]Ci/[mu]l, 5000 Ci/mMol) and 1 [mu]l T4-polynucleotide kinase (200 U/[mu]l) were added and the mixture was incubated at 4^oC for 12 h. The ³² P labeled mRNA was purified by gel electrophoresis.

Functional complexes of the ribosome

The formation of functional ribosomal complexes was as described ( 15 ) but with the following ionic conditions in all steps: 20 mM HEPES-KOH, pH 7.8 (0^oC), 6 mM MgCl ₂ , 150 mM NH ₄ Cl, 2 mM spermidine, 0.05 mM spermine and 4 mM 2-mercaptoethanol. In step one , 20 [mu]l 70S ribosomes (8 pmol/[mu]l) in tico-buffer [20 mM HEPES-HCl pH 7.6 (0^oC), 6 mM MgCl ₂ , 30 mM NH ₄ Cl and 4 mM 2-mercaptoethanol] was mixed with 5 [mu]l 4* tico-buffer and 20 [mu]l of an `adaptation-buffer 1' [60 mM HEPES-HCl pH 7.6 (0<sup>o</sup>C), 18 mM MgCl <sub>2</sub> , 690 mM NH <sub>4</sub> Cl, 10 mM spermidine, 0.25 mM spermine, 12 mM 2-mercaptoethanol], then 50 [mu]l mRNA in water (20.4 pmol/[mu]l with a specific activity of ~20 000 c.p.m./pmol) and 3.1 [mu]l roman {{t R N A} sub f sup {m e t}} in water (78 pmol/[mu]l) or 3.1 [mu]l water when indicated and 1.9 [mu]l water to adjust the appropriate buffer conditions were added. The mixture was incubated for 15 min at 37<sup>o</sup>C. An 18.7 [mu]l aliquot (initiation complex) was taken and kept on ice for 2 h for further analysis, the rest of the reaction mixture was kept on ice. Then 2.5 [mu]l of `adaptation-buffer 2' [100 mM HEPES-HCl, pH 7.6 (0 ^oC), 30 mM MgCl ₂ , 750 mM NH ₄ Cl, 10 mM spermidine, 0.25 mM spermine, 14 mM 2-mercaptoethanol] and 9.8 [mu]l N -Ac[ ¹⁴ C]Phe-tRNA ^Phe solution in water (19.5 pmol/[mu]l, specific activity 1208 c.p.m./pmol) were added. The reaction mixture was incubated for 30 min at 37^oC and divided in two equally-sized aliquots. One aliquot (pre-translocation complex) was kept on ice for further analysis. The second aliquot (43.2 [mu]l with 1.4 pmol/[mu]l) was used for the formation of the post-translocation complex: 8.6 [mu]l of the `adaptation-buffer 3' (0.75 mM GTP, 18 mM HEPES-HCl, pH 7.6 (0^oC), 5.4 mM MgCl ₂ , 136.4 mM NH ₄ Cl, 2.7 mM spermidine) and 2.5 [mu]l of an EF-G solution (9 pmol/[mu]l) in the factor-buffer (20 mM HEPES-HCl, pH 7.6 (0^oC), 6 mM MgCl ₂ , 150 mM KCl, 1 mM DTE, 0.01 mM GDP, 10% glycerol were added. The resulting mixture was incubated for 10 min at 37^oC and cooled in ice (post-translocational complex). Ribosomal complexes were purified from non-bound material either on spun columns or sucrose gradients.

Pre-packed spun columns filled with Sephacryl 300 were pre-washed three times with 2 ml of the binding buffer [20 mM HEPES-HCl pH 7.6 (0 ^oC), 6 mM MgCl ₂ , 150 mM NH ₄ Cl, 2 mM spermidine, 0.05 mM spermine, 4 mM 2-mercaptoethanol] within 2 h before use. The columns were dried by centrifugation at 2000 r.p.m. (400 g ) at 4^oC, then 50 [mu]l of the sample (one of the ribosomal complexes) was applied and the column centrifuged again. Then the column was washed 10 times with binding buffer, i.e. for each washing step 50 [mu]l of binding buffer was applied to the column and centrifuged as described. The fractions were collected and the radioactivity and optical density were measured. Fraction 2 contained the ribosomal complexes which were used for the iodine treatment and for functional tests.

In order to purify the complexes via sucrose gradients, 10-30% sucrose gradients in binding buffer were made. The samples (50 [mu]l) were centrifuged for 17 h at 4^oC at 22 000 r.p.m. (55 000 g ). After fractionating (fraction size ~500 [mu]l) the radioactivity and optical density were determined in each fraction and the fractions with the ribosomal complexes were combined and dialyzed against 200 vol binding buffer three times for 2 h in order to remove the sucrose. Then the samples were used for iodine treatment or functional tests.

The samples purified either by spun columns or sucrose gradients were used for iodine treatment. For that purpose 2 [mu]l 20 mM iodine solution in ethanol was mixed with 50 [mu]l of the sample (0.2 pmol/[mu]l of a complex or mRNA) incubated in ice for 3 min. The reaction was stopped by diluting the sample with 0.3 M NaAc, pH 5.5, up to a volume of 200 [mu]l followed by phenol extraction and ethanol precipitation. The RNAs were dissolved in the PAGE buffer and subjected to an electrophoresis in 12% PAGE (30 * 40 * 0.02 cm) as described above until the xylene cyanol dye migrated 10 cm (1/3 of the gel). The bands were visualized by autoradiography.

The residual samples (before iodine treatment) were used for functional tests. The quantity of mRNA which remained in the complex after purification was estimated by filtration through nitrocellulose filters. Two or three aliquots of the sample of 5 [mu]l each (1 pmol/[mu]l) were diluted up to 2 ml with the binding buffer and filtrated under low pressure. The filters were washed twice by 2 ml of the binding buffer. The filters were dried in air, dissolved in filter-count (Beckman), and the ( ³² P) radioactivity measured.

In the pre- or post-translocation complexes the amount of N-AcPhe-tRNA ^Phe was estimated by the determination of ¹⁴ C radioactivity of the same filters after more than six decay periods of ³² P had passed. In the case of roman {{t R N A} sub f sup {m e t}} the amount of this tRNA in the complex was determined in parallel experiments under the same conditions but with non-radioactive mRNA and 5'-end ³² P-labeled roman {{t R N A} sub f sup {m e t}}.

The peptidyltransferase activity was tested with the puromycin reaction. Purified sample (20 [mu]l) of the pre-translocation ribosomal complex (0.75 pmol/[mu]l) was mixed with 18 [mu]l binding buffer and 7 [mu]l `adaptation-buffer 3' yielding the concentrations of the binding buffer but with 0.75 mM GTP. The mixture was split into three equal aliquots (15 [mu]l each). To the first aliquot 1.2 [mu]l EF-G factor solution (0.9 pmol/[mu]l) was added, and the same amount of factor-buffer without EF-G was added to the two remaining aliquots. The samples were incubated at 37^oC for 10 min. Then 1.2 [mu]l of a puromycin solution [4.6 mg/ml in binding buffer with 1% (v/v) 1 M KOH added just before use] was mixed with the first and the second aliquot and the aliquots were incubated for 12 h at 0^oC. Then 16.2 [mu]l 0.3 M NaAc, pH 5.5, saturated with MgSO ₄ and 1 ml ethylacetate were added and mixed vigorously. The organic phase was used for the determination of ¹⁴ C radioactivity.

RESULTS

The mRNA sequences used in these experiments are presented in Table 1 . The MF-mRNA carries two codons: AUG-UUC in the middle, flanked by sequences containing repeats of (A) ₄ G. This mRNA is suitable for the formation of elongating states of the ribosome, since it does not contain a Shine-Dalgarno (SD) sequence. The cro-mRNA represents the initiation sequence of the natural mRNA of the lambda phage Cro protein and was used before for cross-linking experiments ( 5 , 6 ). This mRNA contains an 8 nt long SD-sequence separated from the initiation codon AUG by a short 3 nt long spacer.

Table 1 . mRNAs used in this study

MF-mRNA	GGG(A 4 G) 3 AAA- AUG - UUC -(A 4 G) 3 AAAU
cro-mRNA	GGG AAGGAGGU UGU- AUG -GAU-CAA-CGG-AAC-UGC-CAG

The initiation codon AUG is underlined with a double line, the codons involved in establishing the functional complexes are in bold letters, the Shine-Dalgarno sequence is underlined with a single line.

For the synthesis of both mRNAs one of the ribonucleoside triphosphates was substituted by a mixture of this triphosphate with the corresponding ribonucleoside 5'- O -(1-thiotriphosphate) in a molar ratio of 4:1 in the transcription mixture. The phosphorothioate groups were statistically incorporated into the mRNA chain. Four derivatives were synthesized for each mRNA with statistically distributed phosphorothioate groups present on the 5'-side of either A, U, G or C residues in the mRNA chain. Before use the mRNAs were labeled with ³² P at the 5'-end.

Three ribosomal complexes were formed with the MF-mRNA. A deacylated roman {{t R N A} sub f sup {m e t}} at the P site is similar to the initiating ribosome, which also carries only one tRNA at the P site. This state is called the P _i state, i for initiation. The pre-translocational state is established when an AcPhe-tRNA is bound to the adjacent A site. An EF-G dependent translocation transforms the complex into the post-translocational state. The quality of both states of the elongating ribosome, viz . the pre- and the post-translocational state, is documented in Table 2 . Around one quarter of the ribosomes carried an AcPhe-tRNA (~0.25 AcPhe-tRNAs were bound per ribosome). Under the conditions of the pre-translocational state 89-96% of the ribosomes carrying an AcPhe-tRNA were indeed in the pre-translocational state. After translocation most of the AcPhe-tRNA molecules reacted with puromycin indicating that under the conditions used almost all AcPhe-tRNA molecules were translocated from the A site to the P site, thus establishing the post-translocational state (for a detailed analysis of the extent of the translocation reaction see ref. 1 ).

Table 2 . Specificity of the pre- and the post-translocational states in the presence of MF-mRNA

Functional state	AcPhe-tRNA bound per ribosome (% at the A site) thioated nucleotides of the MF-mRNA
	U	C	G	A	N
Pre	0.26 (95%)	0.26 (96%)	0.26 (95%)	0.29 (89%)	0.25 (96%)
Post	0.33	0.20	0.35	0.29	0.25

The binding values of Ac[ ¹⁴ C]Phe-tRNA per ribosome are given (specific activity 1208 c.p.m./pmol); the background values (minus ribosomes, around 100 c.p.m.) were subtracted. The %-values in parenthesis give the fraction of AcPhe-tRNA present at the A site in the pre-translocational state. The A-site specificity was assessed by means of the puromycin reaction, background values (minus puromycin, in the range of 64-120 c.p.m.) were subtracted. The puromycin reaction is only a qualitative measure; after EF-G dependent translocation the puromycin values were significant and amounted to (68 +- 15)% of the bound AcPhe-tRNA, i.e. at least ~70% of the AcPhe-tRNA were translocated to the P site. U, C, G, A, the MF-mRNA was randomly thioated at statistically two phosphates at the respective nucleotide. N, non-thioated MF-mRNA. Pre, pre-translocational state. Post, post-translocational state. For more details see Materials and Methods.

The ribosomal complexes were purified from the non-bound components via gel-filtration (spun-columns) or sucrose-gradient centrifugations. Figure 1 demonstrates that the presence of cognate tRNA improves the binding of mRNA, even slightly in the case of cro-mRNA (Figs 1 C and D) where a long SD sequence supports the contacts of the mRNA with the ribosome. Most impressive is the effect of bound tRNAs on the binding of mRNA with no SD contacts in the gel filtration assays (Figs 1 A and B), where most of the mRNA bound in the absence of a cognate tRNA is lost, and in some assays almost all mRNA was removed upon gel filtration. This observation agrees well with previous results obtained with nitrocellulose filtration methods ( 18 ). It is clear that codon-anticodon interaction on the ribosome is important if not decisive for binding of the mRNA to the elongating ribosome. In the following experiments the functional complexes containing MF-mRNA were isolated via gel filtration.

Figure 1 . Purification of functional complexes of the ribosome. ( A and B ) Gel filtration with spun-columns of ribosomes in the presence of 5'-[ ³² P] MF-mRNA with or without fMet-tRNA ^Met _f , respectively. ( C and D ) Purification by sucrose gradient centrifugation of ribosomes in the presence of cro-mRNA with (C) or without fMet-tRNA ^Met _f (D). For experimental details see Materials and Methods.

Figure 2 A shows a representative protection experiment in the presence of MF-mRNA. The coding sequence AUG-UUC is marked in the figure. The conspicuous feature is the uniformity of the pattern, in that nearly all positions are accessible as if the mRNA were in solution (lane 0). All positions of the mRNA can be judged well except the three U nucleotides in the coding region, since a low accessibility of these positions is found with the mRNA in solution possibly due to formation of some tertiary contacts in the mRNA. A closer inspection reveals the following. (i) The phosphates from the A, G and C nucleotides of the coding region AUG-UUC are freely accessible. The apparent protection of the C phosphate in the pre-translocational state (lane 1) could not be confirmed in subsequent control experiments. The U of the AUG codon was always too weak in the control (lane 0) and could not be judged. The UU nucleotides in the UUC codon might be protected in the pre-translocational complex as compared to the P _i states (lanes 3 and 4), where only one tRNA was present on the ribosome (P site) similar to the initiating ribosome (P _i , i for initiation). (ii) Outside the coding region only two positions reproducibly deviated from the control (mRNA in buffer, lanes 0). The phosphate two positions upstream of the decoding codons (those codons which are engaged in codon-anticodon interactions) is significantly protected, and the phosphate one position downstream of the decoding codons shows an increased accessibility. A scan of the enboxed bands of lane 1 from the A block in Figure 2 A confirms the different pattern of these two bands (Fig. 2 B). A scan of some bands of the control lane (see enboxed region in Fig. 2 A, A block, lane 0) shows the uniformity of the respective bands.

Figure 2 . Contact patterns of various ribosomal functional complexes. ( A ) The sequence blocks indicated with U, C, G and A represent cleavage patterns derived from the MF-mRNA containing the respective thioated nucleotide. N, MF-mRNA without thioated nucleotides. A sequence ladder of MF-mRNA is shown on the right. Lanes 1 and 2, pre- and post-translocational complexes. Lanes 3 and 4, P _i complexes. Lane 0, cleavage pattern of the MF-mRNA in buffer. The coding sequences AUG-UUC are indicated. Open arrow, nucleotide protected in the functional complexes; closed arrow, nucleotide with increased intensity in the functional complexes. ( B ) Densitometer scans of the sequences enboxed in (A). The relative determination of the density of a band was reproducible within +-10% (see also ref. 13).

An evaluation of the scans of the bands shown in Figure 2 A and normalization of the data concerning the different input of radioactive material per lane demonstrates that the upstream position is protected to 30-43% as compared to the control lane, whereas the downstream position shows an increase in intensity of 40-45% over that of the corresponding control band, but only in elongating states (pre- and post-translocational states, lanes 1 and 2 of the A block); in the P _i state the increase is less significant and amounts to ~15-20%. The calculated effects in percent are minimal values and could be larger by up to one third, since up to one third of the amount of mRNA found on ribosomes in the presence of tRNAs was still observed in the absence of tRNAs (Fig. 1 A and B).

The position one nucleotide downstream of the decoding codons even shows some spontaneous breakage in non-thioated MF-mRNAs (see lanes 1 and 2 of the block N presenting non-modified mRNAs), which is also sometimes detected in the thioated MF-mRNAs (see, for example, lane 3 in the C block). This spontaneous breakage is observed in up to 25% of the experiments almost exclusively at this position and only with mRNA bound to ribosomes regardless whether or not iodine is present. In free mRNA this cleavage never occurred. In contrast, the enhanced iodine-dependent cleavage is always seen at this position, i.e. the enhanced accessibility observed at this position of bound mRNA is not an artifact provoked by a spontaneous cleavage.

As mentioned above, the P _i state in Figure 2 A mimics an initiating ribosome insofar that the ribosome contains only one tRNA. However, the established state does not truly reflect the initiation state since the mRNA does not contain an SD sequence. Therefore, in a control experiment we tested the P _i state again but now binding roman {{t R N A} sub f sup {m e t}} to the P site in the presence of cro-mRNA (Table 1 ), which contains a strong SD sequence. The cleavage pattern is shown in Figure 3 A. The three flanking lanes A, G and C represent cleavage patterns of the cro-mRNA in buffer. The functional complexes were purified via sucrose gradient centrifugation before the cleaving reagent I ₂ was added (lanes in the right half of Fig. 3 A), whereas similar aliquots were first phenolized before adding I ₂ to the naked RNAs (left half). In the latter case those positions protected by ribosomal components within the ribosome should be accessible again for the cleaving reagent. The sequence of the mRNA is also given in Figure 3 A, the nucleotides being numbered starting at the first nucleotide of the P-site codon.

Figure 3 . Contact patterns of functional complexes in the presence of cro-mRNA (Table 1) showing cleavages at A, G or C positions. Open arrows indicate protected positions, closed arrows increased intensities of the corresponding bands derived from functional complexes. ( A ) Free mRNA, cro-mRNA in buffer; next lane, alkaline ladder of the cro-mRNA. 70S + I ₂ , cleavage patterns of 70S complexes with and without tRNA ^Met _f after purification via sucrose gradient centrifugation. Phenolized 70S + I ₂ , 70S complexes deproteinized by a phenol treatment before the cleaving reagent I ₂ was added. The sequence of the cro-mRNA is given in the free lane. ( B ) Control assay with cro-mRNA thioated only at A positions. +I and -I, the samples were loaded on the gel with or without an iodine treatment. S-A, cro-mRNA with thioated A residues. N, cro-mRNA without thioated nucleotides. Lane 1, mRNA in buffer; lane 2, 70S plus cro-mRNA; lane 3, 70S plus cro-mRNA and tRNA ^Met _f . The A positions are indicated using the same numbering as in Figure A.

The phosphate at position C7 shows an enhanced accessibility (compare with C10 or C15, and note the equally strong cleavage at all three positions of the mRNA in solution) as observed already in the corresponding position (the first A after the UUC codon) with the MF-mRNA. If one compares G3 and G4 in the free mRNA (2. lane from the right) the band of G4 is clearly weaker than that of G3. The opposite is true for the mRNA within the ribosome (10. lane from the right). Since G4 appears not to be protected in similar experiments (not shown) the conclusion is that G3 is slightly protected when the mRNA is bound to the ribosome. Comparing the same lanes a slight protection of G-2 is directly obvious. With similar comparisons it is evident that also positions G-5 (compare G-5 with G-6 and G-8 in the 2nd and 10th lane from the right) and A-7 are protected. One has to take into account that in the presence of an SD sequence the mRNA also binds well to ribosomes in the absence of cognate tRNAs and that therefore the protections are probably more pronounced in the tRNA containing complexes.

A protection at position A1 cannot be excluded in Figure 3 A. Therefore, the experiment with the cro-mRNA containing phosphorothioated A's was repeated and the lanes aligned to facilitate a judgement (Fig. 3 B). It is clear that positions A1 and A-7 are protected whereas, for example, positions A-10 or A9 are not. Unspecific cleavage is seen at the C7 position with the transcript containing non-thioated nucleotides (N +- J) but not with the transcript containing thioated A's (thioA +- J).

DISCUSSION

Interpretation of the results obtained with the I ₂ -phosphorothioate approach used in this work has to be made with caution. Only R _p diastereomers of nucleotide residues are incorporated into the RNA chain in the course of T7 RNA polymerase transcription ( 19 ) and the presence of the sulfur atom disturbs in some cases both RNA-protein ( 20 ) and RNA-RNA interactions (see, for example, 21 ). However, as already mentioned in the Introduction Watanabe and co-workers have shown that phosphorothioate-substituted mRNAs were as active in in vitro protein biosynthesis as non-modified mRNAs ( 14 ). In addition, identification of protected phosphate groups in phosphorothioate-containing tRNA in the complexes with cognate aminoacyl-tRNA synthetases appeared to correlate well with X-ray analysis data ( 12 , 22 ).

Initiating ribosomes can bind mRNA containing a Shine-Dalgarno (SD) sequence even in the absence of tRNAs. The SD sequence significantly improves the binding of the mRNA and arranges the initiator codon AUG near the ribosomal P site ( 23 ), which is the decoding site during the initiation phase of protein biosynthesis. In accordance with this fact we find protections of the cro-mRNA scattered over a sequence which extends from the SD sequence up to the initiator AUG in the presence of roman {{t R N A} sub f sup {m e t}} (A-7, G-5, G-2, A1 and G3; the numbering begins at the first nucleotide of the P-site codon; the U residues were not analyzed for technical reasons). In addition, the phosphate at position 7, i.e. C7 in the cro-mRNA (Fig. 3 A) and the A following the UUC codon in the MF-mRNA (Fig. 2 A), showed an increased accessibility that correlates with the data by Murakawa and Nierlich ( 8 ) who observed the enhancement of the reactivity of both the sugar moiety and the base A7 in ribosome bound lacZ mRNA in the presence of roman {{t R N A} sub f sup {m e t}}. Furthermore, this position directly following the decoding codons at the 3'-side is the only position within an mRNA which shows a significant tendency (up to 25% of the cases) towards spontaneous cleavage. The spontaneous cleavage was exclusively found in mRNA bound to ribosomes regardless whether or not the mRNA contained thioated nucleotides (see, for example, Fig. 2 A, panel N or the position 7 in Fig. 3 B). Obviously, this position is characterized by some structural constraints which facilitate accessibility as well as spontaneous breakages. One explanation is that the mRNA makes a kink around position 7. This interpretation is supported by the fact that the nucleotide at position 6 could be crosslinked with nucleotide 1052 of 16S rRNA in the head of the 30S subunit and that at position 7 with nucleotide 1395 in the body of the subunit ( 5 ). Downstream of the initiating codon at the P site no protections were found. However, we cannot exclude interactions between negatively charged phosphate groups of mRNA and positively charged protein groups.

Elongating ribosomes are characterized by two tRNAs both linked to the mRNA via codon-anticodon interactions (for review see ref. 24 ). The tRNAs are either present at the A and P sites or the P and E sites depending on the elongating state of the ribosome, viz . the pre- or the post-translocational state. The mRNA is bound only very weakly to the ribosome without tRNAs and in the absence of a Shine-Dalgarno sequence. Most of the bound mRNA is lost from the ribosome during nitrocellulose filtration ( 18 , 23 ), gel filtration (Fig. 1 A and B) or toe-printing analysis ( 9 ) in the absence of tRNAs, whereas the mRNAs are tightly bound in the presence of cognate tRNAs ( 18 ). The interesting feature of the contact analysis of the mRNA presented here is that only one phosphate of the mRNA is protected outside the decoding codons in elongating ribosomes. One explanation could be that many phosphates of the mRNA are coordinated via magnesium ions, and that these phosphates cannot be detected with the method applied, since magnesium ions interact only poorly with sulfur atoms in thiophosphates. However, manganese ions can coordinate well with both oxygen and sulfur atoms ( 25 ). The presence of a system linked to a sulfur atom of a phosphate group would hinder an easy access of I ₂ molecules, since the octahedral coordination geometry defines a well organized structure around the phosphate. If directly coordinated the distance between Mg ²⁺ and an oxygen or nitrogen of the mRNA would be 2 +- 0.2 Å, if Mg ²⁺ would be coordinated to water oxygens which in turn form hydrogen bonds with hydrogen acceptors from mRNA the distance Mg ²⁺ -acceptor would be 4 +- 0.4 Å ( 26 ). Positions normally coordinated to a Mg ²⁺ but lacking this link, when a sulfur has replaced the oxygen, should be weakly or not protected in the presence of Mg ²⁺ and thioated mRNA and much stronger protected in the presence of Mn ²⁺ . However, when manganese ions were substituted for 80% of the magnesium ions the same poor protection pattern was found (data not shown). Therefore, it is likely that outside the decoding codons only one protected phosphate exists.

The poor signal of the contact analysis of the mRNA contrasts with the highly differentiated and distinct contact patterns found with both tRNAs on the ribosome ( 13 ). The importance of cognate tRNAs for mRNA binding, the poor contacts between mRNA and ribosome and the rich contact patterns found with tRNAs and ribosome strongly suggest that codon-anticodon interactions-probably together with the surrounding ribosomal components-are the main elements of mRNA fixation during the elongation phase of translation. This conclusion explains the necessity of codon-anticodon interactions for both tRNAs before and after translocation ( 27 - 29 ) and perfectly fits a recent observation that the minimal length of an mRNA is a hexanucleotide and that this minimal mRNA is essential and sufficient for both tRNA binding to A and P sites and the subsequent translocation of the tRNAs to P and E sites, respectively (D. Beyer and K. H. Nierhaus, unpublished observation).

The position of the protected phosphate remains the same relative to the two decoding codons before and after translocation, namely two nucleotides upstream of the two decoding codons, and the same is true for the phosphate with increased accessibility immediately downstream of the two decoding codons. The same changes at the two corresponding positions of the cro-mRNA were observed (G-2 and C7, closed arrow head in Fig. 3 A). This mRNA has a strong secondary structure, whereas the MF-mRNA was designed for a minimal secondary structure ( 17 ); this means that not the change per se of the secondary structure of an mRNA upon binding to the ribosome is responsible for the altered reactivity of these two position (since the reactivity of a position was compared with the corresponding one of the mRNA in solution), but rather the defined structure (conformation) and contacts of the mRNA at and immediately around the decoding codons causes the distinct reactivity. One gets the impression that the micro-topography of the decoding codons does not change during the translocation reaction. The same conclusion was reached when the contact patterns of both tRNAs were compared before and after translocation. The patterns of the tRNAs adjacent on the ribosome were radically different ( 13 ), but neither pattern changed upon translocation (M. Dabrowski, C. M. T. Spahn and K. H. Nierhaus, unpublished observation). It was inferred that a movable ribosomal domain exists which tightly binds two tRNAs and possibly the decoding codons and carries both tRNAs and the cognate codons, during the translocation reaction.

Several contacts of mRNA bases located either upstream or downstream of the codon in the ribosomal P-site have been described using cross-linking with the `zero-length' cross-linking reagent 4-thiouridine ( 4 - 6 ). It was shown that there are universal contacts between 16S rRNA and mRNA bases at positions +4, +6, +7 and +11 in the presence of P-site bound cognate tRNA. These crosslinks did not depend on the sequence of the mRNA. Further crosslinks were found between mRNA and ribosomal proteins S7, S11, S18, S21 as well as with proteins S5, S3, S2 with lower efficiency. For the upstream mRNA region, besides SD-anti-SD interactions some contact points of mRNA bases with the ribosome have also been observed, but in this case their position was dependent on the sequence of the mRNA. The crosslinked nucleotides of the rRNAs obtained with 4-thiouridine modified mRNA are important topographical markers. These bases are not however involved in Watson-Crick base pairing or strong stacking interactions with corresponding 16S rRNA bases since base-base contact of these types would prevent the cross-linking reaction ( 30 ). Instead these mRNA bases probably participate in specific but rather weak interactions with 16S rRNA bases due to overlapping of sulfur atoms in the 4-thiouracil residue in mRNA with double bonds of pyrimidine or purine bases in 16S rRNA ( 30 ). Therefore the sugar moieties and the phosphates of the mRNA were considered as prime candidates for a mRNA fixation. Here we show that at least the phosphates outside the decoding codons with one exception are not involved in any tight contacts with ribosome components.

Previously the accessibility of sugar moieties in the mRNA chain to 1,10-phenanthroline-copper ion oxidative complex (Cu-OP) and OH radicals have been studied. Since the mRNA sequence inside the ribosome could not be cleaved by these reagents, the authors concluded that all sugar residues in mRNA region interact with the ribosome and thus were protected by the ribosome ( 8 , 10 ). These findings are at variance with the results of this study. Using the OH radicals technique for a tRNA bound at the P site, again a complete protection of the bound tRNAs was reported ( 31 ). It is therefore likely that the generator of the OH radicals, namely a Fe ²⁺ -EDTA complex, as well as Cu-OP, cannot penetrate the ribosome due to its size and/or charge and therefore is not able to cleave the backbone of the bound tRNA or mRNA.

Our finding that mRNA lacks phosphates involved in any strong ribosomal interactions at least outside the two decoding codons in the main states of elongation of translation can have an important biological meaning: mRNA should not have strong interactions with the elongating ribosome via phosphates in order to have the freedom to move through the ribosome during the translation process. A consequence of this finding is that codon-anticodon interactions are mainly responsible for fixation and movement of the mRNA during the ribosomal elongation cycle.

ACKNOWLEDGEMENTS

We thank Dr R. Brimacombe and Christian M. T. Spahn for help and discussion. This work was supported by the Volkswagen-Stiftung grant I/69 527, the ISF grant M39000 and a grant from the Russian Foundation for Fundamental Research.

REFERENCES

1 Beyer,D., Skripkin,E., Wadzack,J. and Nierhaus,K.H. (1994) J. Biol. Chem. 269, 30713-30717.

2 Shatsky,I.N., Bakin,A.V., Bogdanov,A.A. and Vasiliev,V.D. (1991) Biochimie, 73, 937-945. MEDLINE Abstract

3 McKuskie-Olson,H., Lasater,L.S., Cann,P.A. and Glitz,D.G. (1988) J. Biol. Chem. 263, 15196-15204.

4 Bogdanov,A.A., Dontsova,O.A. and Brimacombe,R. (1993) In Nierhaus,K.H., Franceschi,F., Subramanian,A.R., Erdmann,V.A. and Wittmann-Liebold,B. (eds) The Translational Apparatus: Structure, Function, Regulation, Evolution. Plenum Press, New York, pp. 421-431.

5 Dontsova,O., Dokudovskaya,S., Kopylov,A., Bogdanov,A., Rinke-Appel,J., Junke,N. and Brimacombe,R. (1992) EMBO J. 11, 3105-3116.

6 Rinke-Appel,J., Jünke,B., Brimacombe,R., Lavrik,I., Dokudovskaja,S., Dontsova,O. and Bogdanov,A. (1993) Nucleic Acids Res. 22, 3018-3025.

7 Hartz,D., McPheeters,D.S., Traut,R. and Gold,L. (1988) Methods Enzymol. 164, 419-425.

8 Murakawa,G.J. and Nierlich,D.P. (1989) Biochemistry, 28, 8067-8072.

9 Balakin,A., Skripkin,E., Shatsky,I. and Bogdanov,A. (1990) Biochim. Biophys. Acta, 1050, 119-123.

10 Hüttenhofer,A. and Noller,H.F. (1994) EMBO J. 13, 3892-3901. MEDLINE Abstract

11 Schatz,D., Leberman,R. and Eckstein,F. (1991) Proc. Natl. Acad. Sci. USA, 88, 6132-6136.

12 Rudinger,J., Puglisi,J.D., Pütz,J., Schatz,D., Eckstein, F,. Florentz,C. and Giege,R. (1992) Proc. Natl. Acad. Sci. USA, 89, 5882-5886.

13 Dabrowski,M., Spahn,C.M.T. and Nierhaus,K.H. (1995) EMBO J., 14, 4872-4882.

14 Ueda,T., Tohda,H., Chikazumi,N., Eckstein,F. and Watanabe,K. (1991) Nucleic Acids Res. 19, 547-552.

15 Rheinberger,H.-J., Geigenmüller,U., Wedde,M. and Nierhaus,K.H. (1988) Methods Enzymol. 164, 658-670.

16 Davanloo,P., Rosenberg,A.H., Dunn,J.J. and Studier,F.W. (1984) Proc. Natl. Acad. Sci. USA 81, 2035-2039.

17 Triana-Alonso,F.J., Dabrowski,M., Wadzack,J. and Nierhaus,K.H. (1995) J. Biol. Chem. 270, 6298-6307.

18 Gnirke,A. and Nierhaus,K.H. (1989) Biochem. Int. 18, 551-559.

19 Griffiths,A.D., Potter,B.V.L. and Eperon,I. (1987) Nucleic Acids Res. 15, 4145-4162.

20 Ruffner,D.E. and Uhlenbeck,O.C. (1990) Nucleic Acids Res. 18, 6025-6029.

21 Hardt,W.-D., Warnecke,J.M., Erdmann,V.A. and Hartmann,R.K. (1995) EMBO J. 14, 2935-2944.

22 Biou,V., Yaremchuk,A., Tukalo,M. and Cusack,S. (1994) Science, 263, 1404-1410.

23 Calogero,R.A., Pon,C.L., Canonaco,M.A. and Gualerzi,C.O. (1988) Proc. Natl. Acad. Sci. USA 85, 6427-6431.

24 Nierhaus,K.H. (1990) Biochemistry 29, 4997-5008.

25 Pecoraro,V.L., Hermes,J.D. and Cleland,W.W. (1984) Biochemistry 23, 5262-5271.

26 Holbrook,S.R., Sussman,J.L., Warrant,R.W., Church,G.M. and Kim,S.-H. (1977) Nucleic Acids Res. 4, 2811-2820.

27 Rheinberger,H.-J. and Nierhaus,K.H. (1986) FEBS Lett. 204, 97-99.

28 Gnirke,A., Geigenmüller,U., Rheinberger,H.-J. and Nierhaus,K.H. (1989) J. Biol. Chem. 264, 7291-7301.

29 Triana-Alonso,F.J., Chakraburtty,K. and Nierhaus,K.H. (1995) J. Biol. Chem. 270, 20473-20478.

30 Favre,A. (1990) In Morrison,H. (ed.) Bioorganic Photochemistry. Willy and Sons, New York, pp. 379-425.

31 Hüttenhofer,A. and Noller,H.F. (1992) Proc. Natl. Acad. Sci. USA 89, 7851-7855.

---

Return

^* To whom correspondence should be addressed
CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				Z. Yang, A. M. Sismour, and S. A. Benner Nucleoside alpha-thiotriphosphates, polymerases and the exonuclease III analysis of oligonucleotides containing phosphorothioate linkages Nucleic Acids Res., May 14, 2007; 35(9): 3118 - 3127. [Abstract] [Full Text] [PDF]

				M. I. Zvereva, P. V. Ivanov, Y. Teraoka, N. I. Topilina, O. A. Dontsova, A. A. Bogdanov, M. Kalkum, K. H. Nierhaus, and O. V. Shpanchenko Complex of Transfer-messenger RNA and Elongation Factor Tu. UNEXPECTED MODES OF INTERACTION J. Biol. Chem., December 7, 2001; 276(50): 47702 - 47708. [Abstract] [Full Text] [PDF]

				M. Dabrowski, C. M. T. Spahn, M. A. Schafer, S. Patzke, and K. H. Nierhaus Protection Patterns of tRNAs Do Not Change during Ribosomal Translocation J. Biol. Chem., December 4, 1998; 273(49): 32793 - 32800. [Abstract] [Full Text] [PDF]

This Article Abstract Print PDF (125K) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (11) Request Permissions Commercial Re-use Guidelines for Open Access NAR Content Google Scholar Articles by Alexeeva, E. Articles by Nierhaus, K. Search for Related Content PubMed PubMed Citation Articles by Alexeeva, E. Articles by Nierhaus, K. Social Bookmarking          What's this?