In addition to requiring U1 snRNP particles, pre-mRNA splicing re quires at least three other snRNPs, U2, U5, and U4/U6, as well as a number of soluble proteins. Together these form a large complex that can be observed in the electron microscope, and which can be biochemically purified. The complex forms in the nucleus even while the RNA is being elongated, and exons near the 5’ end of the RNA can be removed even before synthesis of the RNA is complete. Formation of the complex requires that the regions to which both U1 and U2 bind be present. Scanning by the splicing apparatus from 5’ to the 3’ end may help explain the paradoxically high degree of specificity to splicing. The donor and acceptor splice sites contain only two essential nucleotides, too few to ensure specificity in RNA of random sequence. It is likely that the spacing between introns and exons also helps the splicing apparatus choose sites appropriately. One purification method of spliceosomes is to synthesize substrate RNA in vitro. This RNA is synthesized with ordinary nucleotides plus biotin-substituted uridine. After the RNA has been added to a splicing extract, the biotin can be used to fish out this RNA selectively with streptavidin bound to a chromatography column. Along with the RNA are found the U1, U2, U4/U6, and U5 snRNAs.

The reaction of the mammalian splicing components is at least partly ordered. U1 binds by base pairing to the 5’ splice site and U2 binds by base pairing to a sequence within the intervening sequence containing a nucleotide called the branch point that participates in the splicing reaction. The RNAs of U4 and U6 are extensively base paired whereas a shorter region of base pairing is formed between the U6 and U2 RNAs (Fig. 1). In the course of the splicing reaction the U4 particle is released first. Splicing in yeast is similar to that found in mammals, but differs in many small details. The same snRNPs are involved, but most of the U RNAs are considerably larger than their mammalian counter parts. Only U4 and U6 are closely homologous in both organisms.

Fig1. Base pairing among U1, U2, U4, U6, and the pre-mRNA showing the branch site and the 5’ cut site.

Sometimes extensive splicing is required to regenerate a single intact mammalian gene. For example, there are genes containing one million bases and 60 splice sites. Only a few genes in yeast are spliced, and they contain only a few introns. Great specificity is required in order that all of the splicing reactions proceed with sufficient fidelity that most of the pre-messenger RNA molecules ultimately yield correctly spliced messenger RNA. In part, we still do not know the reasons for the high fidelity. Although a consensus sequence at the 5’ and 3’ splice sites can be derived by aligning many splice sites, there are only two invariant and essential nucleotides present in each of the sites. This hardly seems like enough information to specify correct splicing.

Once in vitro splicing reactions could be performed with unique substrates, it was straightforward to examine the products from the reactions. Amazingly, the sizes of the products as determined by electrophoretic separations did not add up to the size of the substrate pre-mRNA. The structures were then determined by chemical means and by electron microscopy of the resultant RNA molecules. The excised RNA was found to be in a lariat form.

This results from the reaction of the nucleotide at the branch point within the intron attacking the phosphodiester at the 5’ splice site. Subsequent attack of the 3’-OH at the 5’ splice site on the phosphodiester bond at the 3’ splice site releases the intron in a lariat form and completes the splicing process. The freed introns in lariat form are rapidly degraded within the nucleus.