Content advisory
This digitised material is free to access, but contains information or visuals that may:
- include personal details of living individuals
- be upsetting or distressing
- be explicit or graphic
- include objects and images of objects decontextualised in a way that is offensive to the originating culture.
Licence: In copyright
Credit: "Split genes and RNA splicing". Source: Wellcome Collection.
2/5
![1) The DNA in the cell producing the message might be rearranged to displace needed. On this hypothesis the DNA in 2) 8 The DNA would remain unaltered but the RNA polymerase, producing the primary RNA transcript, would skip across the introns on the DNA so that only the exons appear in the primary separately, and the separate pieces of RNA would then be joined together in the correct order to form the final mRNA. 4) The RNA polymerase would make a primary transcript of the whole region, both exons and introns. This transcript time the exons were all joined together in which is almost certainly the one that oc curs in the majority of cases, is now pop ularly known as splicing. suggest'.' It has been shown that the first mechanism—the rearrangement of DNA light chain of the immunoglobulin of the mouse (either k or X) is coded, in the germ line, on two widely separated stretches of DNA. These are found to be much closer together in the DNA of the somatic cells producing the protein (2,6, 7). This is a very important result but I shall not pursue it further. There are good reasons for suspecting that the im mune system may be a special case, al though not necessarily a unique one. So far there is no evidence at all suggesting that either the second or the third mecha nism listed above is actually used (8). On the other hand, the evidence (not de scribed in detail here) for the fourth mechanism, is now so widespread that there is little doubt that it, or something very like it, is actually happening. For the rest of this article, therefore, I shall ignore the first three mechanisms and concentrate on splicing. How Widespread Is Splicing? I have spoken as if splicing only oc curred in the processing of mRNA, but we already know that at least two other species of RNA are spliced. Indeed, one of the earliest discoveries was that some of the transfer RNA (tRNA) molecules in yeast are spliced, although their introns are fairly small (9, 10). More recently two groups of investigators have isolated a crude enzyme preparation that will per form that operation in the test tube (/1, 12). The single gene for ribosomal RNA (rRNA) in a yeast mitochondrion ap pears to contain an intron (13). Some genes for rRNA in Drosophila also ap- evidence suggests that these particular genes may not be transcribed (15). Whether the nuclear precursor of rRNA is ever spliced remains to be discovered. So far, there is no evidence at all to show whether or not other kinds of RNA molecules, such as the small RNA's found in the nucleus, are produced by splicing. Thus splicing is defined as the mechanism by which a single func tional RNA molecule is produced by stretches of RNA during the processing of the primary transcript. Where are split genes found? So far, karyotes (the bacteria and the blue-green algae), they would almost certainly have been discovered earlier. We cannot yet say categorically that they do not occur in prokaryotes but it certainly seems un likely that they do. They are common in eukaryotic viruses. Indeed, that is where their importance was first realized, but have only been found in DNA viruses that occur in the cell nucleus (2) or in RNA retroviruses which have a DNA nuclear phase (/ 6). Split genes have not so far been discovered in viruses that ex ist only in the cytoplasm of a cell. All this would suggest that the phe nomenon of splicing is correlated with This hypothesis would make very good sense. In a prokaryotic cell, which lacks a nuclear membrane, the translation of the message by ribosomes starts well be fore the transcription of the message from the DNA has finished. In a eu karyotic cell, by contrast, the process of transcription takes place in the nucleus, whereas the process of translation on the ribosomes takes place mainly if not en tirely in the cytoplasm. The two opera tions are separated by the nuclear mem brane, and this gives an obvious oppor tunity for additional processing to take place. Such a hypothesis would predict that split genes would not be found in mitochondria. Unfortunately, the experi mental evidence suggests that there, too, genes are split into pieces. In yeast mito chondria the single gene for the larger rRNA molecule is almost certainly split (13). The evidence that an mRNA is also split is not yet completely decisive, but it is certainly very suggestive (17). For two enzymes required for splicing will be available in the cell, so it would not be too surprising if they (or a close relative of them) penetrated into the mitochon drion. What is surprising is that there ap pears to be no evidence that there is any membrane separating the DNA of the mitochondrion from its ribosomes, al- problem is discussed again below. Splicing in Higher Organisms I now attempt to give a rapid and nec essarily incomplete summary of the dis tribution of split genes in higher orga nisms. In mammals, one or more of the fn several species (19-25), as have the genes for certain k and X light chains of immunoglobulin in the mouse (6, 7, 26- 28) and the heavy chain of a mouse im munoglobulin (29). As was mentioned bumin gene in chickens has been shown to be split into many pieces (30-38), and there is suggestive evidence that this is also true for the chick ovomucoid gene (39). So far, there is no evidence from other vertebrates, nor for any gene in a plant. There is a report (40) that the gene for silk fibroin in the silkworm is split but there is no definitive evidence for a split transcribed (15). If split genes do exist in Drosophila, one would expect that a case would be discovered fairly soon. In the fungi, the only example known is that of several tRNA's in yeast (9-12). It is obviously impossible to deduce much from such sparse experimental evidence, but results are likely to come in fast, and it may only be a year or two before we can begin to answer what is probably the most important question of this sort: are there any eukaryotes in which split genes When we come to consider the actual protein molecules whose genes have been shown to be split, we notice that they all have one thing in common. They are all molecules of terminal dif ferentiation. This is because they are technically the easiest to study. Nobody has yetdescribed or reported an example of a gene for a common-or-garden en zyme (such as one from the Krebs cycle) although such studies are in progress. The other thing that one cannot help no ticing is the high frequency of introns. There are two in certain immunoglobulin light chains (4), two in various hemoglo bin chains (19-24), at least four in the y, heavy chain (29), and no less than seven in chick ovalbumin (31, 32). Moreover, they are of considerable length, running from just under 100 base pairs to more than 1000. In the ovalbumin gene, the to tal length of the introns is at least three times that of the exons. If we average this small amount of data, we find that we might expect about one intron for that its average length would be greater than 600 base pairs. That is, on an aver age, the introns are longer than the exons. In a higher organism a gene has, if anything, more nonsense than sense in it. These preliminary estimates are nec essarily very insecure. The introns in yeast tRNA are much smaller. So far the lengths found are 14, 18, 19, and 34 bases (41). Whether there are introns in yeast mRNA is not yet Are there any proteins for which we not split? This appears to be the case for the sets of histone genes which have been studied both in a sea urchin (42) and in Drosophila (43). In both species, the genes are repeated many times in a tan dem arrangement. Unfortunately, there is reason to suspect that histone genes are not completely typical. They do not have polyadenylate [poly(A)] at the end of their mRNA's, for example, and may be designed to exit quickly from the nu cleus. It would obviously be interesting firm up the preliminary evidence which shows how the transcript of any particu lar gene is split. The study, by electron microscopy, of the hybridization of ge netic DNA with the related mRNA (or of nucleic acid clones derived from them) needs only small amounts of material and, in careful hands, gives reliable re sults. Historically, it was this method that first suggested that viral mRNA was not a simple colinear transcript of the viral DNA (2). Its resolving power is low, however, as is that of mapping by restriction enzyme digestion. For de tailed mapping it is essential to obtain the actual base sequences (44). Details and Generalizations Let us now consider in more detail the arrangement of introns and exons. The first thing we notice, from the very limit ed experimental data at present available produces a single protein (45), whereas a stretch of DNA in a virus may produce more than one protein, depending on which way the primary transcript is spliced (2). I adopt the attitude that in short of DNA and, by various devices, their limited amount of DNA is made to code for more proteins than would other wise be possible. We can see this even in prokaryotic viruses, such as t/>X174, where the same stretch of DNA can be read in one phase for one protein and in another phase for another protein (46). A typical example of a gene producing more than one protein is the early T-anti- gen region found in both SV40 (47, 48) and polyoma (49). It now seems certain that at least two proteins are produced by this region, each beginning with about 100 residues having exactly the same amino acid sequence. The remaining parts of their amino acid sequences seem to depend on exactly how the RNA tran script is spliced (50). Such cases are of interest because the favorable technical nature of the viral systems make it likely that many details will be worked out by studying them. However, such multiple- chromosomal genes although, as has already been argued (3, 4), they may be important as transitional stages in evolu- more DNA than they know what to do with. Should a chromosomal gene arise whose transcript was processed to make more than one protein, I would expect that in the course of evolution the gene would be duplicated, one copy sub sequently specializing on one of the pro teins and the other copy on the other. If this point of view is correct, then one would expect multiple-choice genes to occur only rarely in the chromosomes of eukaryotes (51). The other tentative generalization we can make from the present data is that the order of the exons on the DNA is the same as the order in which they are found in the final mRNA. There does not seem any strong reason why this should always be true. It is possible to devise mechanisms in which the order would sometimes be different. This colinearity of the exons probably reflects some im portant aspect of the origin of introns or of the splicing process and therefore shouid not be overlooked. It is, in ly within the coding region of a message since in the case of ovalbumin, for ex ample, one intron is found in the leader region of the mRNA before the coding Howls Splicing Done? What is the actual mechanism of splic ing? At the moment any ideas must nec essarily be largely speculative. One would certainly expect at least one en zyme to be involved, if not several. In the case of the tRNA from yeast, an en- though it has still to be purified (11, 12). It is not completely obvious that such a mechanism would require a source of en ergy since two phosphate ester bonds need to be broken whereas only one (or possibly two) have to be made. On bal ance, one would suspect that energy might be required if only because the process must be an accurate one. Prelim inary evidence indicates that the enzyme appears to need adenosine triphosphate (ATP) (11). Not all the different kinds of tRNA molecules found in yeast need to be processed by splicing, but so far the indications are that those that are spliced are processed by one and the same en zyme (41). There is still no evidence that this same enzyme can also process pre cursors of mRNA, and I argue that in any case this is unlikely. This brings us to one of the major un solved problems: how many different en zymes are involved in splicing? In other](https://iiif.wellcomecollection.org/image/b1817727x_PP_CRI_I_1_31_0002.jp2/full/800%2C/0/default.jpg)
