A Synthetic Model for Triple-Helical Domains in Self-Splicing Group I Introns Studied by Ultraviolet and Circular Dichroism Spectroscopy

Abstract

Structural studies were performed on synthetic oligonucleotides with sequences corresponding to the P4/P6 and J3/4, J6/7 regions of the self-splicing group I intron of the bacteriophage T4 nrdB pre-mRNA, which correspond to the proposed triple-helical domain in the Tetrahymena thermophila intron. A 23-mer RNA was synthesized as a mixed ribo-deoxyribo oligonucleotide, modeling an expected base-paired region of P4 along with the J3/4 and P6 (5'-end bases of P6) regions. strand modeling the 3'-end bases of P6 and J6/7 regions, with which a triple helix may form, was synthesized as a pure oligoribonucleotide (7-mer RNA). The interactions of these oligonucleotides have been characterized by UV and circular dichroism (CD) spectroscopy. The results show that the 23-mer RNA forms a stable hairpin modeling the P4 base-paired region. Triple helix association between the 23-mer RNA hairpin and the 7-mer RNA single strand was detected by CD in the presence of Mg2+ (>5mM) but not in presence of a monovalent cation like Na+ (up to 500 mM). Studies on selected variants of both 7-mer and 23-mer RNAs were carried out. The results show that for the association of the two partner strands not only the formation of P6 helix but also triplet interactions between two strands are required. The association of the two strands in general follow a pattern predicted by comparative sequence analysis. Parallel studies on pure oligoribonucleotides having base sequence corresponding to those of the oligoribonucleotides showed no evidence of association under similar conditions, which could indicate that the 2'-hydroxyl groups of the riboses might play an important role in hydrogen bonding to form the required nucleoside triples. Molecular modeling studies on the proposed "plaited triple helix" formed by the association of the 23-mer RNA hairpin and 7-mer RNA single strand showed that the structure is sterically and energetically feasible.

Specific metal-ion binding sites in a model of the P4-P6 triple-helical domain of a group I intron

“…The two-component model system has proven useful for probing the role of potential specific magnesium-ion sites that might be crucial for the formation of the triplehelix motif+ Added to previous results on base and sugar substitutions (Sarkar et al+, 1996(Sarkar et al+, , 1997, the present results show in detail how certain local interactions in the RNA contribute to the formation of the correct structural motif+ We were able to separate the effects of the metal-ion switch (Mg 2ϩ r Mn 2ϩ ) on the CD spectra from the effects of a true structural rescue by Mn 2ϩ in the phosphorothioate substituted systems+ …”

“…Here we have studied the formation of the core triple helix motif in a two-component model system+ We have used difference circular dichroism (CD) spectroscopy in a novel approach to probe structurally important magnesium-ion binding sites within the motif+ This structural approach is in contrast to many other phosphorothioate experiments, which use activity as an assay (Dahm & Uhlenbeck, 1991;Chowrira & Burke, 1992;Christian & Yarus, 1993;Piccirilli et al+, 1993;Scott & Uhlenbeck, 1999)+ Difference CD spectroscopy emphasizes the spectral differences of an associated system as compared to its nonassociated state (Johnson, 1985;Gray et al+, 1992)+ It can also be used to compare a modified system to its unmodified counterpart+ By combining these two approaches, we were able to separate the effects of the metal-ion switch (Mg 2ϩ r Mn 2ϩ ) on the CD spectra from the effects of a true structural rescue by Mn 2ϩ + Our model system consists of a 7-mer ribonucleotide (7-mer RNA) and a 23-mer mixed ribo-deoxyribo oligonucleotide (23-mer RNA) (Fig+ 1B,C)+ These have been shown by us, in earlier studies using CD and FTIR spectroscopy (Sarkar et al+, 1996(Sarkar et al+, , 1997, to associate in the presence of Mg 2ϩ , modeling the interactions in the P4-P6 triple-helical domain of a self-splicing group I intron from bacteriophage T4 nrdB pre-mRNA (Fig+ 1A), which has a core sequence analogous to the Tetrahymena ribozyme (Sjöberg et al+, 1986)+ The assembly process of this two-strand model system may in certain aspects better resemble the native folding as compared to a single-stranded model, as it is not constrained by the covalent connectivity in a short strand+ The model has previously been shown to be sensitive to effects of base substitutions (Sarkar et al+, 1996) and 29-OH substitutions (Sarkar et al+, 1997) on triple helix formation+ In the present study, the stereospecific effects of Rp and Sp sulfur substitutions at the 59 end of nucleotides U452 and U453 are evaluated+ U452 and U453 are positioned at the junction of helix P6 and helix P4 in the triple-helical model system (Fig+ 1C), and correspond to nucleotides U258 and U259 of the Tetrahymena ribozyme+…”

“…The results from the wild-type system show that the metal-ion switch (Mg 2ϩ r Mn 2ϩ ) give different effects on the observed (associated) and the calculated (nonassociated) spectra+ The decrease in CD intensity seen upon changing the ionic environment from 100 mM Mg 2ϩ to 100 mM Mn 2ϩ is larger in the calculated spectrum than in the observed spectrum (6% as compared to 2+5%; data not shown), which results in an enhancement of the difference spectrum in Mn 2ϩ + A similar effect is seen for the phosphorothioate system that is not affected by the sulfur substitution (U452 Rp, Fig+ 4A(ii-iv))+ Thus the possibility exists that the enhancement of the amplitudes seen in the difference spectra of U452 Sp and U453 Sp in 3:1 and 0:1 Mg 2ϩ :Mn 2ϩ (Fig+ 4B,D(iii-iv)) does not show regained RNA-metal-ion interactions, but might only reflect the different effects of the metal-ion identity on the associated and nonassociated states+ To elucidate this possibility, we also compared the observed spectra of the phosphorothioate substituted systems in the different ionic environments to the triple-helical structure of the wild-type system under identical salt conditions by constructing the difference between observed substituted and observed wildtype spectra (obs-wt)+ The results presented here (Fig+ 5) were obtained at 50 mM divalent concentration, where the association of the wild-type strands in Mg 2ϩ reaches saturation (Sarkar et al+, 1996)+ Very similar results were obtained also at 100 mM and 200 mM ion concentrations (data not shown)+ Figure 5 shows difference spectra at 1:0, 3:1, and 0:1 ratios of 50 mM Mg 2ϩ :Mn 2ϩ , where differences between observed and calculated spectra [(obs-calc) Rp, the effect of metal-ion identity is clearly seen in the (obs-calc) spectra (Fig+ 5A(i)), where the addition of manganese causes the amplitude of the difference spectra to increase, whereas for U453 Rp, the metal-ion identity effect is seen in the (obs-wt) spectra (Fig+ 5B(ii))+ For U452 Sp (Fig+ 5C(ii)), we do not observe the increasing amplitudes with increasing Mn 2ϩ in the (obswt) spectra+ This shows the opposing effect of increased association added to the metal-ion identity effect on the nonassociated strands, that is, a true Mn 2ϩ rescue+ For U453 Sp (Fig+ 5D(ii)) there is an increase in the amplitude of the difference (obs-wt) in the 3:1 Mg 2ϩ :Mn 2ϩ environment indicating that no true Mn 2ϩ rescue is involved+ However, there may be a certain rescuing effect in the 0:1 Mg 2ϩ :Mn 2ϩ environment+…”

“…The nonmodified 7-mer oligoribonucleotide (7RNA ns) was prepared using the H-phosphonate method as described previously (Sarkar et al+, 1996, and references therein)+ The 23-mer mixed oligoribo/oligodeoxyribonucleotide (23-mer RNA) was purchased from Dharmacon Research (Boulder, Colorado, USA) and purified as described (Sarkar et al+, 1996)+ The phosphorothioate substituted 7-mer ribonucleotides containing a single phosphorothioate linkage with fixed stereochemical configuration were synthesized in a fashion similar to the nonmodified 7-mer RNA but with incorporation of a fully protected dimeric building block+ The four sequences are as follows: 59-A-G-G-U-U-C-A-39 having a phosphorothioate with either Rp or Sp configuration between the two underlined uridines, that is, 59 to the nucleoside that corresponds to U453 in the wild-type sequence (referred to as U453 Rp and U453 Sp); 59-A-G-G-U-U-C-A-39 having a phosphorothioate with either Rp or Sp configuration between the underlined G and U, that is, 59 to the nucleoside that corresponds to U452 in the wild-type sequence (referred to as U452 Rp and U452 Sp)+ The U-U dimeric building blocks were the Rp and Sp isomers of 59-O-monomethoxytrityl-29-O-tert-butyldimethylsilyluridine 39-(S-o-nitrobenzyl 29-O-o-chlorobenzoyluridine 59-phosphorothioate) with an H-phosphonate moiety at the 39 end of the dimer+ For the G-U dimeric building blocks, we used the tert-butyldiphenylsiloxymethylbenzoyl group (DreefTromp et al+, 1990a(DreefTromp et al+, , 1990b The protected phosphorothioate dimers were produced by first condensing the 29,39-dibenzoyluridine-or 29,39,N 3 -tribenzoyluridine 59-H-phosphonate with either of the protected G or U 39-OH components (Almer et al+, 1992)+ After chromatographic separation of the resulting H-phosphonate diastereomers, each isomer was sulfurized stereospecifically with elemental sulfur in pyridine (Almer et al+, 1992)+ Removal of the benzoyl groups that were used as temporary protection was achieved using either ammonia or sodium methoxide and the resulting thiophosphate was protected at the S atom by alkylation with o-nitrobenzyl chloride+ The dimers were finally converted into a suitable H-phosphonate oligonucleotide synthesis building blocks following a published one-pot procedure (Rozners & Strömberg, 1997)+…”

“…phage T4 nrdB (nt 449-455 in the original sequence), and a 23-mer RNA having the same base sequence as the J3/4-P4 (59-end bases of P4, nt 42-51), followed by three deoxythymidines (dT 3 ) replacing the P5 loop, and the P4-P6 (39-end bases of P4 and 59-end bases of P6, respectively, nt 70-79)+ The P5 loop in the original sequence is replaced by dT 3 , as they form a stable hairpin loop that should not be involved in RNA tertiary interactions (Sarkar et al+, 1996) (Sarkar et al+, 1996), having the overall features characteristic of an A-helix, with a large positive band around 250-300 nm and a narrower but negative band positioned around 210 nm (Johnson, 1985;Gray et al+, 1992)+ The singlestranded nonsubstituted 7-mer RNA (7RNA ns) and phosphorothioate substituted 7-mer RNAs (U452 Rp, U452 Sp, U453 Rp, and U453 Sp) all showed CD spectra characteristic of non-base-pairing oligonucleotides (Johnson, 1985; data not shown)+ Triple helix formation of 23-mer RNA and nonsubstituted 7-mer RNA CD spectroscopy is very sensitive to conformational changes of nucleic acids+ If a triple helix is formed in a mixture of 23-mer RNA and one of the single-stranded 7-mer RNAs, the resulting CD spectrum differs from that of a nonassociated system (Gray et al+, 1992)+ The spectrum of nonassociated 23-mer RNA and 7-mer RNA at a specific divalent metal-ion concentration is equal to the sum of the individual spectra of 23-mer RNA and 7-mer RNA recorded under identical salt conditions+ Utilizing this fact, we recorded a set of CD spectra of 23-mer RNA, 7-mer RNA, and a 1:1 mixture (in single strands) of 23-mer RNA and 7-mer RNA+ Each set was recorded at five different ionic strengths (0, 5, 50, 100, and 200 mM) and in three different molar ratios (1:0, 3:1, and 0:1) of Mg 2ϩ :Mn 2ϩ + We added the appropri- (Michel & Westhof, 1990; in the shallow groove of helix P6 and in the deep groove of helix P4 are indicated by dotted bonds+ The 2 nt in question for phosphorothioate substitutions (U452 and U453) are marked with slanting lines and dots, respectively+ The substitutions are made 59 to U452 and U453, as shown by arrows+ ately weighted spectra of 23-mer RNA and 7-mer RNA (corresponding to a nonassociated triple-helical system), and compared this calculated spectrum to the one observed for the 1:1 mixture of 23-mer RNA and 7-mer RNA+ If no association takes place in the 1:1 mixture, the observed and calculated spectra overlap as for the wild-type system (23-mer RNA and nonsubstituted 7RNA ns) in 0 mM Mg 2ϩ (Fig+ 2A), and the difference spectrum (observed Ϫ calculated) follows the baseline (Fig+ 2C)+ When association takes place, the observed and calculated spectra differ, as for the wild-type system in 100 mM Mg 2ϩ (Fig+ 2B)+ The observed spectrum of the 1:1 mixture of 23-mer RNA and 7RNA ns at 100 mM Mg 2ϩ has a more intense and slightly narrowed positive band around 250-300 nm than the calculated (nonassociated) spectrum recorded under identical salt conditions+ As a result, a 269-nm positive peak and two zero-crossover points around 263 and 278 nm arise in the difference spectrum, due to the more intense and slightly narrowed 250-300 nm band of the associated system+ The enhancement of the spectral features characteristic of triple helix formation upon raising the Mg 2ϩ concentration reflects an increased degree of association between the RNA oligomers (Fig+ 2C)+ The isolated triple-helical model system requires high divalent ionic strengths to associat...…”