Why is Lariat-RNA formed in the RNA splicing reaction?

Both in the nuclear pre-mRNA- and in the Group II self-splicing reaction, introns splice out by way of two successive transesterification reactions (Fig. 1). The cleavage at the 5'-junction in splicing is due to an intramolecular nucleophilic attack by a 2'-OH from within the intron. This creates a lariat-RNA intermediate (Fig. 1) with the 5'-end of the intron attached through a 2'→5'-phosphodiester bond to a residue near the 3'-end of the intron in the penultimate step of the splicing reaction (Fig. 1). Subsequent cleavage at the 3'-junction results in the ligation of the exons and liberates the "free" intron in the form of a lariat. Group II splicing does not require any protein, or any external source of energy such as ATP or GTP. These lariat-RNA structures have adenosine as the "branch-point" residue, linked via a a 2'->5'   phosphodiester bond to a guanosine residue, and a 3'->5' phosphodiester bond to a pyrimidine residue. Studies on the specificity of the nucleobase requirement at the branch-accepting point have shown that the replacement of the central branch-accepting adenosine residue by any other nucleotide in mutants results in a considerable decrease in the efficiency of the splicing as compared to the wild type. Mutation experiments have shown that all four nucleotides can serve as the branch-acceptors, but it is adenosine which fulfills the structural criteria for the second step of splicing optimally, whereas cytidine residue is preferred to guanosine and uridine residues.
 

Figure 1. The two steps in Group II and Nuclear pre-mRNA Splicing. (Left) : Substrate (E1-IVS-E2), A = Branch-point Nucleotide. (Middle) : the 2'-OH of the branch-point adenosine (A) attacks the 5'-phosphate of G residue connecting the E1 exon giving the Lariat intron which is still covalently linked to the exon E2; (Right):  Upon formation of the Lariat intron, 3'-OH of E1 attacks the 5'-phosphate of E2 giving the ligated exons (E1 + E2) as products.. In Group II Splicing, no external source of energy such as ATP or protein required, only cofactor required is Mg²⁺ ion. Nuclear mRNA splicing reaction takes place within the spliceosome consisting of 50S to 60S complex of the pre-mRNA, four small ribonucleoprotein particles and as yet unknown number of associated protein factors, ATP and Mg²⁺ ion.

The above observations have prompted us to address to the following important questions regarding the mechanism of RNA splicing:

(1) Why is a lariat ubiquitously formed in RNA splicing reactions in eukaryotes?
(2) What are the unique structural features in the lariat that control the fidelity of the RNA splicing and, hence, protein biosynthesis in our cell?
(3) Since no ATP or GTP is required in the Group II splicing, and two covalent bonds are formed and broken, therefore the total energy input and output are balanced; these transesterification reactions should still need free energy of activation for covalent bond forming or -breaking reactions! Where does the free energy of activation come from in the transesterification reactions in the splicing? What is the energy inventory of the transesterification reactions involved in the splicing?
(4) Is there any structural and conformational basis for the evolutionary choice of adenosine as the branch-accepting nucleotide in the splicing reaction?
(5) How the conformation of the lariat is changed when the "branch-point" adenosine is replaced by cytidine, guanosine or uridine nucleotide?
(6) Does the chemospecific choice of adenosine nucleotide at the branch-point by Nature give the splicing reaction in vivo  a thermodynamic or a kinetic advantage that it can not have with any other nucleobase?

In order to answer these questions, we have developed synthetic methodologies to prepare model branched RNAs and lariat RNAs mimicking those of the branch-point of the natural lariat and have analyzed their tertiary structures by highfield isotope-edited NMR spectroscopy.

The summary of our results:

The conformation of the branched RNAs is dictated by the actual nucleobase composition. Thus, adenosine at the branch-point has a completely different tertiary structure than the counterparts with either guanosine, cytidine or uridine as the branch-point nucleotide. Second, the competing 2'->5' and 3'->5' stacking at the branch-point acts as an 'energy pump' since these two different stacked states have different free energies associated with their conformations. Thus a transition of the 2'->5' stacked conformation in a branch-RNA or lariat-RNA to 3'->5' stacked conformation releases a quantum of energy that can provide the free energy of activation for the transesterification reaction in the splicing [see our papers in the PDFs 121, 137, 145, 144, 150, 151, 153, 162, 163, 170, 183, 185, 195, 198, 202].

New chemical methodologies are being developed (both by solid and solution phase procedures) in my lab to synthesize some important model lariat-RNAs (Figure 2) that mimic that of the natural lariat-RNA of Fig. 1 more closely. Some of these efforts have resulted into the first unequivocal synthesis of lariat-RNAs [Figuire 2, see our papers in the PDFs 193, 206]. The most remarkable observation in the studies of these synthetic model lariat-RNAs is that the tetrameric and the pentameric loop in the lariat hexamer and heptamer self-cleaved to give corresponding branched RNA with a 5'-OH terminus and a 2',3'-cyclic phosphate, which is the hallmark of the first step of the catalytic self-cleavage reaction (the second being the catalytic turnover!) [see our papers in the PDFs 218, 225, 240].

    These synthetic lariat-RNAs are much smaller than the natural catalytic RNAs such as the hammerhead ribozyme (k = ~1 min^-1 at 37 °C), and their rates of the self-cleavage are also much slower (k = 0.25 x 10^-4 min^-1 for lariat hexameric lariat-RNA, and 0.16 x 10^-3 min^-1 for lariat heptamer at 22 °C).

    This slower self-cleavage rate of hexameric and hepatmeric lariat-RNAs, compared to the natural hammerhead have enabled us to perform their full conformational analysis by NMR spectroscopy [see our papers in the PDFs 218, 225, 240]. We have also shown that the trinucleotidyl loop in the tetrameric and pentameric lariat-RNAs [see our paper in the PDF 209] are completely stable.

    Our NMR studies have revealed the global as well as the local conformation of the self-cleavage site and the dynamics of the tetranucleotidyl and pentanucleotidyl loop in the hexameric or heptameric lariat-RNA [see our papers in the PDFs 218, 225], which has led us to understand, for the first time, the unique conformational requirements of the sugar-phosphate backbone in the lariat-RNAs that undergo the self-cleavage reaction akin to the natural ribozyme.

    In our search to explore the optimal structural requirement for the self-cleavage reaction of RNA, we have now synthesized a lariat-RNA analog in which the branch-point adenosine has a 2',5'-linked tetranucleotidyl loop and a 3'-ethylphosphate moiety mimicking the 3'-tail of the lariat-hexamer. What we found unique is that the 3'-ethylphosphate function at the branch-point in is the key structural feature that orchestrates its self-cleavage reaction (k = 0.15 x 10 ^-4 min^-1 at 19 °C) to give acyclic branched RNA, compared to the stable 2',5'-linked cyclic RNA or 3',5'-linked cyclic RNA  [see our paper in the PDF 240]. We have subsequently worked out [PDF 240] the detailed conformational features of the self-cleaving tetrameric lariat-RNA by 600/800 MHz NMR spectroscopy and molecular dynamics simulations in the aqueous environment, which we are now in a position to use in the design of a suitable ribozyme.
 

Interestingly, our NMR work on the self-cleavage and geometry of the self-cleaving center is gradually giving us a very good understanding of the dynamic geometry of the bent sugar-phosphate backbone both at the self-cleavage site as well of the other constituent stable internucleotidyl phosphates.

Efforts are now underway to synthesize other lariat-RNAs that would have larger loop (hexa- to octadecanucleotidyl) till we have seen the self-cleavage rate arriving at the plateau. Note that in order to reach the goal for the synthesis of larger lariat loop, we are trying to come up with new effective synthetic methodologies such as stereospecific phosphorylation techniques, orthogonal protecting group strategy, intramolecular transesterification reaction etc. We also aim to determine the conformation of each of these self-cleaving lariat-RNAs by 600/800 MHz NMR spectroscopy and figure out the optimal tertiary structural requirement for the self-cleavage reactions. These studies should allow us to examine the local geometry of the self-cleaving lariat-RNAs within a family of lariat-RNAs and observe how does the conformational constrain of the lariat-loop dictate their unique self-cleavage reaction. Clearly, the understanding of the hyperspace of the geometry of the self-cleaving sites would open some new exciting possibilities also for the design of the ribozyme.

Design of Ribozyme Mimics Basing on the Self-Cleavage  of Lariat-RNA and Branched-RNA
 

• First synthesis of lariat-RNA mimicking the Splicing intron has been achieved [Tetrahedron, 47, 9659 (1991), PDF 206;  Tetrahedron, 49, 649 (1993); PDF 193],and their conformational analysis have been performed using NMR spectroscopy in order to understand the structural uniqueness of lariat-RNA. The work is also in progress [PDF 209, 240] to understand the mechanism of the RNA Splicing reaction of Eukaryotes using synthetic chemistry to create model RNA systems and elucidating their solution conformations (NMR) both in the presence and absence of oligopeptides (reconstruction of splicesome and determination of its conformational dynamics and the interacting sites between its protein and RNA components).

 
• Studies on branched RNAs

Previously we have also carried out chemospecific synthesis of various branched RNAs [PDF 195, 193, 190, 167, 157, 139], mimicking those formed in Group II splicing. Detailed synthetic work leading to various branched RNAs have also enabled us to address the relative importance of 3' -> 5' vis-a-vis 2' -> 5' stacking in these molecules by configurational analysis (NMR) [See PDF 137, 145, 144, 150, 151, 153, 162, 163, 170, 183, 185, 195, 197]. In this context, the exact sequence around the branch-point is found to be very important in negotiationg whether 3' -> 5' or 2' -> 5' stacking is prefered.
 
• Discovery of the unique self-cleavage reaction of the lariat-RNA giving 2',3'-cyclic phosphate and a 5'-hydroxy termini mimicking the products that are formed in the catalytic cleavage of RNA and determination of the solution conformation of the unique self-cleavage site in RNA, for the first time, have opened possibilities to synthesize designed RNA catalyst [Tet. Lett., 34, 3929 (1993), PDF 218; Tetrahedron, 50, 1777 (1994), PDF 225; Tetrahedron, 50, 8711 (1994), PDF 240] against pathogen-specific RNA (gene-directed drug), which would specifically bind and cleave the pathogen-specific RNA. Design and synthesis of antisense DNA/RNA that bind to form a duplex or a triplex and act as a repressor at the transcriptional and translational level of expression of the pathogen-specific DNA/RNA is also in progress. NMR studies are being performed with the duplex and triplex DNA and their intercalated complex with the tethered cleaving agents. Development of new synthetic as well as separation methodologies for large scale production of large biologically functional oligo-DNA/RNA are being developed for structural studies by NMR.

 
• Studies on lariat RNAs , ribozymes, and ribozyme mimics

Self-cleaving of RNA occurs in several RNA molecules including the hammerhead, certain hairpins, the Delta ribozyme, the Neurospora mitochondrial domain and is also important in catalytic RNA processes such as splicing of group I nuclear pre-mRNA, Group II mitocondrial mRNA introns and RNAse P RNA promoted maturation of pre-tRNA. We are interested in understanding the structure requirements and mechanisms of such self-cleaving RNA domains and will conduct studies on RNA model systems. Remarkably simple synthetic lariat RNA models, mimicking those of the lariat intermediate formed in Group II pre-mRNA splicing, self-cleave in a manner similar to that of e.g. the self-cleaving catalytic hammerhead domain [PDF 218, 225, 240]. All of these self-cleavage reactions are site-specific and produce a 2',3'-cyclic phosphate and a 5'-hydroxyl terminus, analogous to other self-cleaving domains and in their corresponding artificial ribozyme-RNA substrate complexes. We have previously determined the structure of self-cleavage site in a lariat RNA[PDF 225, 240]. We will now attempt to use this information to design efficient ribozyme mimics which hybridize to and cleave the target RNA in presence of Mg²⁺. This reaction should occur in the absence of proteins or any external energy source such as ATP or GTP. The role of various metal ions as cofactors in various catalytic RNA processing reactions (metalloenzymes) will be also investigated.

• Uppsala University (UU)
• Biomedical Centre (BMC)
• Dep't of Bioorg. Chem.
• Research
• List of Research Projects