Molecular Biology of the Cell - Volume 2 - Alberts B., Bray D., Lewis J., Raff M., Roberts K., Watson J. 1993

Control of Gene Expression
Post-transcriptional Control

Although The activity of most genes is controlled primarily at the level of METABOLISM/31.html">Transcription initiation, such control also occurs later at various stages during The transfer of information from RNA to protein. Moreover, it is quite likely that for certain genes every stage of their expression is under control. Most of our knowledge regarding post-Transcriptional Regulation has been acquired only recently, and the control mechanisms at this level are still far from fully understood.

This section discusses the Introduction/30.html">Regulation of Gene Expression after RNA polymerase has bound to its promoter and initiated RNA Synthesis. The topics are presented in the same chronological sequence in which regulatory mechanisms can affect the RNA molecule following the onset of transcription.

10.4.1. Transcription attenuation leads to premature termination of synthesis for certain RNA molecules [38]

The phenomenon of transcription attenuation has been studied predominantly in Bacteria, where this mechanism regulates the expression of many genes. For this type of regulation to occur, a specific nucleotide sequence must be present near the beginning of the RNA chain, allowing the RNA molecule to adopt one of two alternative Conformations. The more stable conformation features an RNA hairpin that acts as a termination signal for bacterial RNA polymerase, resulting in the premature arrest of RNA synthesis (see Fig. 5-6). However, if a regulatory molecule binds to a specific sequence within the growing RNA chain, the RNA adopts an alternative conformation in which the termination signal is not formed, yielding a long, functional RNA molecule (Fig. 10-50). In bacteria, this regulatory component is typically a ribosome that attaches to the nascent RNA chain during Translation.

Class="center">

Fig. 10-50. Regulation of Gene Expression by transcription attenuation in prokaryotes. As the RNA transcript grows through The addition of NUCLEOTIDES at the 3' end, it folds into one of two alternative conformations. Conformation A is the most stable, containing two double-helical hairpins formed by complementary base pairing. Because the hairpin formed by the pairing of sites 3 and 4 is followed by a series of U nucleotides, it serves as a transcription termination signal for bacterial RNA polymerase. Conformation B arises when a regulatory protein or ribosome binds to site 1 on the RNA transcript, causing free site 2 to pair with site 3 instead; this abolishes the termination signal and generates a long, functional RNA transcript that drives gene expression.

In eukaryotes, transcription attenuation is involved in The regulation of only a small number of genes. Because functionally active Ribosomes are absent from the Cell Nucleus, it is possible that regulatory molecules bind to specific RNA sequences, although The Mechanism of attenuation in Eukaryotic Cells remains poorly understood.

10-23

10.4.2. RNA splicing can be regulated so that the same gene directs the synthesis of different protein isoforms [39]

RNA splicing was originally discovered in Viruses. It was found that a single primary transcript can yield multiple mRNA molecules and, consequently, direct the synthesis of several different Proteins (see Section 9.4.12). Many higher eukaryotic genes produce diverse proteins precisely through alternative RNA splicing. If multiple splice sites exist within different regions of a transcript, a single gene can serve as a template for dozens of different proteins. Usually, however, splicing options are limited, and only a few distinct proteins are translated from each transcript.

In some cases, Alternative Splicing results from "intron ambiguity": the standard intron removal mechanism cannot clearly distinguish between two or more alternative pairs of 5' and 3' splice sites, leading to the random realization of different variants in different contexts. Such a constitutive form of alternative splicing is likely responsible for generating various aberrant mRNAs from the mutant β-globin gene in certain individuals suffering from β-thalassemia (see Fig. 9-86). For other genes, such ambiguity is also a normal feature, resulting in The production of different encoded versions of the same protein across all Tissues where the gene is expressed.

In many instances, it is specifically alternative RNA splicing, rather than constitutive splicing, that is regulated. The Selection of splice sites is determined by The Cell. Consequently, depending on the physiological needs of the Organism, different cells can translate distinct proteins (or sets of proteins) from the same primary RNA transcript. Thus, many proteins are produced in tissue-specific forms. These include components of 1) the Extracellular matrix (Fibronectin), 2) the Cytoskeleton (Tropomyosin), 3) The Plasma Membrane, 4) The Nucleus (see Table 10-1), and 5) Intracellular Signaling pathways (C-kinase and Tyrosine-protein kinase encoded by the src proto-oncogene; Fig. 10-51).

As a rule, the exon substitutions resulting from alternative RNA splicing do not generate entirely unrelated proteins. Instead, they produce a series of functionally analogous proteins known as isoforms. Protein isoforms are modified to suit the requirements of a specific tissue. These modifications may determine which other molecules the protein interacts with, while the catalytic or Structural domains remain unchanged. For example, in thyroid cells, the same transcription unit gives rise to Calcitonin, whereas in Nerve Tissue it yields CGRP (calcitonin gene-related peptide), a peptide hormone with vastly different properties.

Fig. 10-51. Regulatory alternative RNA splicing of the src gene yields two slightly different forms of tyrosine protein kinase. Exon A is incorporated into the sequence exclusively in neural tissues. Because this tissue-specific splicing pattern has been conserved throughout evolution (characteristic of both birds and mammals), it strongly suggests that the observed differences in src-encoded proteins are crucial for the biological activity of this regulatory protein. Only protein-coding exons are shown (exon 1 forms the 5'-leader sequence of the mRNA). (After J. B. Levy et al., Mol. Cell. Biol. 7: 4142–4145, 1987.)

10.4.3. Alternative RNA splicing can be used to switch genes on and off [40]

Certain Genes are transcribed constitutively in all cells; however, due to constitutive splicing, the resulting mRNA encodes a nonfunctional protein, and the gene is effectively expressed only in cells where a specialized splicing reaction takes place. This type of gene regulation is particularly well characterized in Drosophila. For instance, the ability of the P element (see Section 5.6.10) to transpose exclusively in Germ Cells is linked to its failure to produce active transposase in somatic cells. This phenomenon, in turn, is due to the presence of an intron in the transposase mRNA that appears to be removed only in germ cells. Genetic analysis has revealed another striking example: the sex of flies is determined by a cascade of gene activations, each responsible for synthesizing a protein that dictates the correct RNA splicing of the subsequent gene in the pathway (Fig. 10-52). The DNA regions encoding some of these sex-determining proteins have been cloned and sequenced, greatly facilitating The Study of mechanisms governing splice-site selection.

Fig. 10-52. Cascading Changes in the expression of sex-determining genes in Drosophila rely on alternative RNA splicing. When The ratio of the number of autosome sets to the number of X Chromosomes equals 1.0 (normally two sets of autosomes and two X chromosomes), the individual develops as a female; flies in which this ratio is 0.5 (normally two sets of autosomes and one X chromosome) develop as males. This ratio is determined early in development and subsequently maintained in every cell. The function of the genes shown in the diagram is precisely to transmit information about this ratio to other genes that establish the sex-specific phenotype. These genes function as two alternative sets defining female- and male-specific traits, respectively. The dsx (doublesex) gene owes its name to the fact that in mutants failing to express this gene, both the female- and male-specific sets are active. The diagram illustrates The sequence of events during female development. Arrows indicate the regulatory action of each gene involved. In males, the Sxl, tra, and tra2 genes are transcribed but yield only nonfunctional mRNA, whereas the dsx transcript is spliced to produce a protein that switches off genes specifying the female developmental pathway. In females, the Sxl transcript undergoes an entirely different splicing pattern, generating a splicing-control protein that, on the one hand, maintains its own synthesis and, on the other hand, turns on the two tra genes (indicated by arrows). In turn, the tra gene products act cooperatively to alter the splicing pathway of the dsx gene transcript. The resulting dsx mRNA produces a modified form of the DSX protein that turns off genes specifying the male developmental pathway.

Fig. 10-53. Four alternative RNA splicing pathways observed experimentally. In each case, a single transcribed RNA can be spliced in two alternative ways to generate two different mRNAs (1 and 2). Darker shading indicates RNA sequences present in both types of mRNA. Lighter rectangles correspond to sequences specific to only one mRNA type. Adjacent rectangles are connected by colored lines indicating intron sequences. Evidently, no simple mechanism or single rule can account for why a particular choice is made. (After A. Andreadis, M. E. Gallego, and B. Nadal-Ginard, Annu. Rev. Cell Biol. 3: 207–242, 1987.)

10-23

10.4.4. The mechanisms responsible for splice-site selection in regulated RNA splicing remain unknown

It is believed that regulated changes in RNA splice-site selection are mediated by the binding of tissue- and gene-specific proteins or RNA molecules to the nascent RNA transcript. Because both constitutive and regulated splice-site selection utilize the same standard consensus sequences, binding of a specific regulatory component must alter the conformation of the RNA transcript to mask or expose previously accessible splice sites. Complex mechanisms are likely at play here, as the simple assumption that protein binding merely blocks a splice site fails to fully explain the observed diversity of splicing products (Fig. 10-53). To elucidate the MOLECULAR MECHANISMS OF regulated splicing, it will be necessary to reconstruct the process in a cell-free system, which would allow the isolation of all required components and the analysis of each component's effect on the spliceosome.

10.4.5. Shifts in the site of RNA transcript Cleavage and polyadenylation can alter the carboxyl terminus of a protein [41]

In eukaryotes, the 3'-end of an mRNA molecule is determined not by the termination of RNA synthesis by RNA polymerase, but rather through an RNA cleavage reaction catalyzed by auxiliary factors during transcript elongation (see Section 9.4.5). The site of this cleavage can vary, thereby altering the carboxyl terminus of the resulting protein molecule, which is encoded by the 3'-end of the mRNA. In prokaryotes, the generation of a longer RNA transcript merely results in the addition of extra Amino Acids to the protein chain. In eukaryotes, however, alternative cleavage and splicing of a longer transcript can cause the original carboxyl terminus of the protein to be completely removed and replaced with a new one.

This type of alteration is responsible for the switch in antibody synthesis during B lymphocyte maturation from membrane-bound to secreted forms. In immature B cells, the Antibodies produced are anchored to the plasma membrane, where they function as antigen receptors. Antigen stimulation simultaneously triggers the division of these cells and the initiation of antibody secretion. The secreted form of the antibody differs from the membrane-bound form only in the terminal region of its carboxyl terminus: the membrane-bound form contains a long stretch of hydrophobic amino acids that spans The Lipid Bilayer, whereas the secreted form bears a much shorter segment of Water-soluble amino acids. Thus, switching from the synthesis of membrane-bound to secreted antibodies requires a different nucleotide sequence at the 3'-end of the mRNA.

Fig. 10-54. The choice of RNA cleavage and polyadenylation sites plays a crucial role in antibody production. In unstimulated B cells (left), long RNA transcripts are produced, and splicing removes an intron sequence near the 3'-end of the molecule. This generates an mRNA that encodes a membrane-bound antibody molecule. By contrast, following antigen stimulation (right), the primary RNA transcript is cleaved upstream of the acceptor splice site of the terminal exon. As a result, certain intron sequences that were excised from the long transcript are retained as coding sequences in the short transcript. These are the sequences that encode the hydrophilic carboxy-terminal portion of the secreted antibody molecule.

The membrane-bound form of the protein is produced when all DNA coding sequences are transcribed into a long transcript. The nucleotides encoding the long hydrophobic carboxyl terminus of the membrane-bound protein are localized in the final exon (Fig. 10-54, left). The intron preceding this exon contains the nucleotides that encode the water-soluble tail of the secreted molecule; these are removed during mRNA splicing. The secreted form of the molecule is generated from a shorter primary transcript that terminates before THE START OF the next exon. Consequently, this transcript lacks an acceptor splice site upstream of the nucleotides encoding the water-soluble tail that could otherwise interact with the existing donor site. As a result, these nucleotides are retained in the resulting mRNA molecule (Fig. 10-54, right).

However, it remains unknown how the cleavage reaction that dictates the switching of RNA transcript lengths is controlled.

10.4.6. The Discovery of Alternative Splicing Requires a Revision of THE CONCEPT OF the "Gene" [42]

Ever since it became clear that eukaryotic genes contain introns and that their coding sequences can be joined in various combinations, the question of what constitutes a "gene" has resurfaced. The first molecular-level definition of a gene was proposed in the early 1940s based on biochemical genetic studies of the fungus Neurospora. Prior to that time, a gene was considered to be a region of The Genome that segregates in Meiosis as a discrete unit and is responsible for the expression of a specific phenotypic trait, such as white versus red eyes in Drosophila or smooth versus wrinkled seeds in peas. Following the work on Neurospora, it became evident that a gene generally corresponds to the region of the genome that directs the synthesis of a single enzyme. This led to the hypothesis that one gene encodes one polypeptide chain. This hypothesis proved remarkably fruitful. During the investigation of expression mechanisms in the 1960s, a gene came to be defined as a segment of DNA that is transcribed into an RNA encoding a single polypeptide chain (or a single structural RNA, such as a tRNA or rRNA molecule). The discovery of interrupted eukaryotic genes in the late 1970s did not contradict these accepted views. However, we now know that in higher eukaryotic cells, many DNA sequences can encode two or more distinct proteins through alternative splicing. What, then, should be considered a gene?

In those relatively rare instances where two markedly different eukaryotic proteins are produced from a single transcription unit, these proteins are said to be encoded by separate genes that overlap on the chromosome. Nevertheless, defining the modified proteins generated by alternative RNA splicing as products of overlapping genes may seem unnecessarily complex. It is simpler to modify the original formulation and define a gene as any DNA sequence that is transcribed as a single unit and encodes a set of closely related polypeptide chains (protein isoforms).

10.4.7. RNA Export from the Nucleus Can Be Regulated [43]

A primary RNA transcript is, on average, about 10 times longer than the mature RNA molecule generated by splicing. Yet, according to some estimates, only about one-twentieth of all hnRNA leaves the cell nucleus (see Section 9.4.8). Thus, a significant fraction of primary transcripts (likely half) is completely degraded within the nucleus without ever forming an mRNA molecule destined for transport. Degradation affects both RNA molecules whose sequences cannot be converted into mRNA and those capable of undergoing proper Processing only in a different cell type.

The transport of RNA through nuclear pores is an active process (see Section 8.3.3). If this process depends on the specific recognition of transported RNA molecules (or RNA-bound proteins or RNAs) by a receptor protein belonging to the nuclear pore complex, then RNAs lacking such an identification tag will be selectively retained in the nucleus. Alternatively, RNA transport may not require recognition signals, in which case all RNA is exported automatically except for that which is selectively retained. A third possibility is that both selective export and selective retention are utilized. Because any RNA is retained in The Nucleus as long as spliceosome components remain bound to it, selective retention could be mediated by a mechanism that prevents the completion of splicing for a specific RNA molecule. At present, none of these hypotheses has received definitive experimental confirmation; moreover, it seems unlikely that nuclear RNA transport plays a major role in regulating the expression of most eukaryotic genes.

Because viruses rely on the host cell's machinery to replicate, studying their developmental cycle often provides valuable insight into cellular processes. For instance, the adenovirus genome consists of a double-stranded DNA molecule that replicates and is transcribed within the host cell nucleus. In the late stages of infection, the export of host RNA from the nucleus halts, with the result that most of the RNA entering the Cytoplasm belongs to the adenovirus. Genetic analysis has demonstrated that two adenovirus proteins produced early in infection are required for this shift in selectivity of RNA nuclear export. Thus, the interaction between an adenovirus and its host cell can serve as a promising model system for investigating the regulation of RNA transport.

10-24

10.4.8. Proteins binding to the 5' leader region of mRNA are involved in negative translational control [44]

Not all RNA molecules reaching the cytoplasm are translated into protein. The translation of some is blocked by specialized repressor proteins (Fig. 10-55) that bind near the 5' end, right where Translation initiation should take place. This type of regulation was first discovered in bacteria, where an excess of ribosomal proteins can suppress the translation of their own mRNAs.

In eukaryotic cells, a particularly well-characterized type of negative translational control allows for the rapid adjustment of intracellular ferritin synthesis in response to the concentration of free iron atoms within the cell. It has been shown that upon the addition of iron, ferritin mRNA in the cytoplasm shifts from an inactive ribonucleoprotein complex to a translationally active polyribosomal complex. Recombinant DNA studies have established that iron-mediated regulation involves a 30-nucleotide sequence located in the 5' leader region of the ferritin mRNA molecule. This iron-responsive element folds into a stem-loop Structure (see Fig. 10-60B) and binds to a regulatory protein provided that the protein is not complexed with iron. When the regulatory protein binds to the iron-responsive element, the translation of any downstream RNA sequence is repressed (Fig. 10-55). The addition of iron leads to the dissociation of the RNA-protein complex, resulting in a 100-fold increase in The rate of mRNA Translation.

Fig. 10-55. Negative translational control mediated by a site-specific DNA-binding protein (translational repressor). Binding of this protein to an mRNA molecule leads to a decrease in its translation level. Several Examples of this type of translational control are known. The diagram illustrates the mechanism of enhanced ferritin synthesis upon an increase in the concentration of free iron within the cell (see also Fig. 10-60).

10.4.9. The presence of a translational enhancer in certain viral mRNAs indicates the existence of positive translational control [45]

In principle, positive translational control can be mediated by a specialized "translational enhancer" region located on the mRNA, which has The ability to selectively recruit ribosomes. Certain RNA viruses (Picornaviruses) have been shown to contain such a region. Its presence causes translation to initiate at internal AUG sites that are otherwise not utilized for the initiation of Protein Synthesis in eukaryotic cells (Fig. 10-56).

Positive translational control has also been discovered in Yeast cells. Genetic approaches have identified specific proteins required to activate the translation of mRNA from the yeast GCN4 gene. In the absence of these proteins, the mRNA is not translated. The GCN4 mRNA resembles poorly translated mRNAs of higher eukaryotes, and translation of such mRNAs is presumed to be regulated similarly. (This class comprises about 5% of all mRNAs characterized to date). In these RNAs, the 5' leader sequence is unusually long and contains a series of AUG triplets that impede the Translation of the main coding sequence located downstream of the short peptide initiation site. A stop codon positioned prior to the main coding sequence prevents readthrough. Similarly, in picornaviruses, the translation of the main coding sequence of such an mRNA may depend on the binding of translational activator molecules to translational enhancer sequences located near the respective AUG codons, thereby driving reinitiation (Fig. 10-57). However, the exact mechanism of such translational activation remains unknown.

Fig. 10-56. Experimental scheme demonstrating the presence of an RNA sequence that acts as a translational enhancer in the genomes of certain RNA viruses. The genome of picornaviruses (which include the poliovirus) consists of a plus-strand RNA, meaning it can function directly as an mRNA for the synthesis of virus-specific proteins. These viruses lack a 5' cap, which is required for translation initiation on most mRNA molecules. By measuring the level of protein synthesis driven by various recombinant RNAs, translational enhancer sequences several hundred Base Pairs in length can be identified within the viral RNA molecule. As illustrated, moving this sequence into the middle of the mRNA chain allows a ribosome to initiate translation from a neighboring internal AUG codon, thereby bypassing the rules that normally restrict protein synthesis initiation exclusively to the first AUG codon.

Fig. 10-57. A model of positive translational control, proposing that high-level mRNA translation requires the binding of a specific protein (a translational activator). Although it is well established that the translation of specific mRNAs is under positive control, the exact mechanism remains elusive. Evidence indicates that positive control is linked to the synthesis of short Peptides whose translation is initiated upstream of the first AUG. This suggests a mechanism analogous to that described in picornaviruses (Fig. 10-56).

10-25

10.4.10. Many mRNAs are targets of translational control [46]

How widespread is translational control in higher eukaryotes? According to some estimates, the expression of roughly one in ten genes is regulated in this manner. Translational control allows cells to rapidly and reversibly modulate protein concentrations without suppressing the Synthesis of the underlying mRNA (see Section 12.4.7). The expression of certain Proto-oncogenes appears to be regulated in precisely this way.

Translational control plays an especially vital role in fertilized eggs, where it is necessary to switch protein synthesis from the repertoire characteristic of dormant oocytes to that required for rapid Cell Division. Such eggs contain a large reserve of mRNA produced during oocyte maturation. Many of these maternal mRNAs remain untranslated prior to Fertilization. Studies of bivalve Mollusks have shown that different sets of mRNAs are associated with ribosomes before and after fertilization. When these mRNAs are translated in a cell-free system, they yield proteins corresponding to specific stages of mollusk egg maturation; however, this only occurs when the mRNA is preserved as a ribonucleoprotein. If the RNA is stripped of its associated proteins prior to translation, the differences between proteins synthesized from mRNAs of various stages disappear. Consequently, the factor determining whether a given mRNA is translated must depend on how the RNA interacts with regulatory molecules.

5-8

10.4.11. Translational frameshifting generates two distinct proteins from a single mRNA molecule [47]

The translational control mechanisms described above influence the rate of initiation of new polypeptide chains on an mRNA molecule. Protein synthesis typically terminates automatically. In certain cases, however, the resulting protein can be altered during a process known as translational frameshifting.

Fig. 10-58. Translational frameshifting is required to produce retroviral Reverse Transcriptase. Viral reverse transcriptase and integrase are generated by the cleavage of a large gag-pol fusion protein, whereas capsid proteins result from the Cleavage of the abundantly expressed gag protein. Synthesis of both proteins begins at the same start site, but for gag, it terminates at a stop codon in the reading frame, whereas a -1 frameshift yields the chimeric protein. This frameshift is driven by local RNA structural features (such as the RNA stem-loop shown), which cause the tRNA^Leu attached to the C-terminus of the growing polypeptide chain to occasionally slip backward by one nucleotide within the ribosome, pairing with a UUU codon instead of the UTR-defining UUA. The sequence shown is from HUMAN IMMUNODEFICIENCY VIRUS type 1 (HIV-1). (After T. Jacks et al., Nature 331: 280-283, 1988.)

Translational frameshifting is widespread among Retroviruses, enabling them to produce varying amounts of two or more proteins from a single mRNA. These viruses typically synthesize a large polyprotein that is subsequently cleaved by a viral protease to yield several capsid proteins (gag proteins) along with viral reverse transcriptase and integrase (pol proteins). In many instances, the gag and pol genes reside in different reading frames; thus, a frameshift is required to produce the much less abundant pol proteins. Frameshifting occurs at a specific codon within the mRNA and requires the presence of specific cis-acting sequences located both upstream and downstream of this site (Fig. 10-58).

10-26

10-27

10.4.12. Gene expression can be controlled by altering mRNA stability [48]

Most bacterial mRNAs are highly unstable, exhibiting a half-life of approximately 3 minutes. Because bacterial mRNA is both rapidly synthesized and rapidly degraded, bacteria can swiftly adjust their GENE EXPRESSION IN response to environmental fluctuations.

Fig. 10-59. Specific signal sequences responsible for the instability (rapid turnover) of certain mRNAs. A. Three mRNAs with vastly different half-lives. The continuous decay of mRNAs encoding various growth factors allows for rapid fluctuations in their concentration in response to extracellular signals. Histones are required primarily for Chromatin formation during DNA Replication. Marked variations in the stability of their mRNAs restrict histone synthesis strictly to the S phase of the Cell Cycle. B. Specific sequences located within the 3' untranslated region of the molecule are responsible for unusually rapid mRNA decay. The results shown were obtained from experiments in which these altered RNA sequences were expressed in cells from genetically engineered genes.

Fig. 10-60. In response to elevated iron levels, the cell increases the synthesis of ferritin (to sequester excess iron) and decreases the synthesis of transferrin receptors (to limit iron uptake). Both responses are believed to be mediated by the same regulatory protein, which recognizes shared features in the mRNAs encoding ferritin and the transferrin receptor. A. The regulatory protein dissociates from the mRNA. B. The iron-responsive element (IRE) located in the 5' untranslated region of ferritin mRNA. Highlighted nucleotides are identical in mammalian, avian, and amphibian ferritin mRNAs. Placing this sequence into the 5' untranslated region of other mRNAs renders their translation likewise responsive to ambient iron levels. C. Two of the five identical sequences required for the iron-dependent response, located in the 3' untranslated region of transferrin receptor mRNA. Iron-dependent control of mRNA stability requires Multiple binding sites for regulatory proteins. The highlighted nucleotides, along with the stem-loop structure (whose precise stem sequence is less critical), are thought to play a decisive role in protein recognition. Because transferrin receptor and ferritin synthesis are controlled by different mechanisms, their expression levels respond inversely to changes in iron concentration, even though the same regulatory protein is involved. (After M. W. Hentze et al., Science 238: 1570-1573, 1987; J. L. Casey et al., Science 240: 924-928, 1988.)

In eukaryotic cells, mRNA is considerably more stable. For example, the mRNA encoding $eta$-globin has a half-life exceeding 10 hours, whereas other mRNAs have half-lives of merely 30 minutes or even less! Unstable mRNAs frequently encode regulatory proteins whose cellular concentrations must fluctuate rapidly (such as growth factors and the products of the proto-oncogenes fos and myc). The 3' untranslated region of many such unstable mRNAs contains lengthy A- and U-rich sequences that appear to dictate their instability (Fig. 10-59).

mRNA stability can be modulated in response to extracellular signals. For instance, Steroid Hormones act on cells not only by enhancing the transcription of specific genes (see Section 12.2.1) but also by increasing the stability of certain mRNAs transcribed from those genes. Conversely, the addition of iron to cells reduces the stability of transferrin receptor mRNA, leading to diminished production of the iron-binding protein. Interestingly, the destabilization of transferrin receptor mRNA appears to involve the same iron-sensitive RNA-binding protein that controls ferritin mRNA translation. In the case of the transferrin receptor, the protein binds to the opposite end of the mRNA (the 3' untranslated region), thereby enhancing rather than repressing protein synthesis (Fig. 10-60).

10.4.13. Continuous Protein synthesis is required for selective mRNA degradation [49]

The control of mRNA stability in eukaryotic cells is best understood for histone mRNAs. The half-life of these mRNAs during the S phase of the cell cycle (when new histones are in high demand) is roughly 1 hour, but it drops to a few minutes when DNA Synthesis is halted. If DNA synthesis is inhibited with Antibiotics during the S phase, histone mRNA becomes unstable; presumably, the accumulation of free histones in the absence of nascent DNA for them to bind specifically accelerates histone mRNA turnover.

The regulation of histone mRNA stability relies on a short stem-loop structure at the 3' end, which replaces the poly(A) tail found on most other mRNAs. This specialized 3' end is generated following the synthesis of histone mRNA by RNA polymerase II via a cleavage reaction that requires base-pairing with a small RNA component of a ribonucleoprotein particle (U7 snRNA). Engineering this 3' end onto other mRNAs makes them similarly unstable when DNA synthesis is arrested (see Fig. 10-59). Thus, as with other mRNA types, the rate of decay is tightly controlled by signals situated near the 3' end, which is believed to be the starting point for mRNA degradation.

Inserting a stop codon into the middle of the coding sequence of a histone mRNA abolishes its ability to undergo rapid degradation. This observation led to the hypothesis that the nuclease responsible for mRNA turnover is tethered to the ribosome, and that a significant portion of the histone mRNA must be translated before degradation can proceed. This hypothesis could explain why many unstable mRNAs are selectively stabilized when cells are treated with the protein synthesis inhibitor cycloheximide. Nevertheless, it remains unclear why the mRNA degradation machinery should be linked to the ribosome in this manner.

10.4.14. Some mRNAs are localized in specific Regions of the cytoplasm [50]

The in situ Hybridization technique makes it possible to localize specific mRNA molecules within a cell (see Section 4.6.11). As will be discussed in Chapter 16, certain mRNAs involved in body pattern formation during early development are concentrated in specific regions of the oocyte cytoplasm (see Figs. 16-22, 16-60). Although what they are attached to remains unknown, in many cases this regional distribution depends on the long 3' untranslated region of the mRNA.

Although localizing mRNA molecules is easiest in large oocytes, examples of specific mRNA localization are also known in somatic cells. In the future, Genetic Engineering techniques will likely make it possible to alter the intracellular positions of specific mRNA molecules to determine the functional consequences of such displacement.

10-28

10.4.15. mRNA editing changes the meaning of Genetic information [51]

The discovery of ever-new molecular mechanisms utilized by cells continually astonishes biologists. For example, it was recently discovered that all trypanosome mRNAs contain a common capped sequence at their 5' end, which is transcribed separately and then added to the 5' ends of pre-mRNA transcripts by splicing two previously unlinked molecules. Similar trans-splicing occurs in nematodes, where a 5' leader sequence is added to A number of mRNAs. It has also been discovered in plants via the combination of separate RNA transcripts to form the coding sequence of certain chloroplast proteins. The cutting and rejoining of transcripts may prove to be a shortcut to the evolution of new proteins, and the few known instances of such exon joining today may be remnants of processes that were once much more widespread.

Another fascinating phenomenon that significantly modifies RNA transcripts was discovered in trypanosome Mitochondria. Known as RNA editing, it involves the insertion or deletion of one or more U nucleotides in specific regions of a transcript, which alters the original reading frame and changes the information content. However, it remains unknown how this editing is controlled to consistently produce the sequence encoding the required protein.

RNA editing, albeit to a more limited extent, also occurs in mammals, where this mechanism is used in the apolipoprotein B gene to produce Two Types of transcripts: in one transcript, a DNA-encoded cytosine is replaced by uracil, creating a stop codon and leading to the synthesis of a shorter, tissue-specific variant of this large protein. Although this remains the only known example of RNA editing in mammals, it seems unlikely that the phenomenon is restricted to just a single gene.

10.4.16. RNA-catalyzed reactions probably have a very ancient origin [52]

All the post-transcriptional control mechanisms discussed in this section rely on the specific recognition of a particular RNA molecule; this is how molecules are selected for special transformations, such as splicing or degradation. Such recognition is made possible by the existence of A large number of site-specific RNA-binding molecules, most of which remain uncharacterized. The RNA sites with which these molecules associate typically contain a stretch of unpaired nucleotides within a single-stranded region (see Fig. 10-60). Consequently, this type of site-specific binding differs from DNA binding, where The nucleotide sequence is usually recognized via paired bases in The Double Helix. Moreover, all known molecules that bind to specific DNA sequences are proteins, whereas both proteins and RNA molecules are capable of binding to specific RNA sequences, relying in part on complementary RNA-RNA base-pairing for recognition. Thus, in attempting to understand post-transcriptional processes, we have entered the world of RNA. RNA-catalyzed reactions are much more difficult to study than protein-catalyzed ones. Due to trace amounts of ribonucleases, isolated large RNA molecules degrade easily and are difficult to purify to homogeneity while retaining activity. However, modern Genetic engineering Methods using purified RNA polymerases now make it possible to produce large amounts of pure RNA with any desired sequence in vitro (Fig. 9-81). This has made it possible to study in detail the chemistry of RNA-catalyzed self-splicing reactions (see Section 3.2.11) and to determine the minimum sequence sizes required for the self-cleavage of plant viroid RNA (Fig. 10-61). Another reaction in which RNA plays a catalytic role in both PROKARYOTES AND EUKARYOTES is the cleavage of tRNA precursors by a protein-RNA complex known as RNase P. The RNA component of spliceosomal snRNP particles likely also participates in forming and breaking covalent bonds (although this has yet to be proven). It is believed that the Active Site of the peptidyl transferase enzymatic complex, which is responsible for Amino Acid Polymerization by the ribosome, resides within rRNA (see Section 5.1.8).

Fig. 10-61. STRUCTURE OF THE active site of a plant viroid RNA. This short RNA molecule self-cleaves at the site indicated by the arrow. The highlighted nucleotides are identical in seven self-cleaving RNAs, six of which are found in plants and one in animals. Self-cleavage reactions convert tandemly repeated single-stranded RNA sequences (an intermediate product of viroid-like RNA) into a single-copy linear RNA molecule, which subsequently circularizes. (After A. C. Forster et al., Nature 334: 265-267, 1988.)

Fig. 10-62. Regulatory interaction between two RNA molecules helps maintain a constant copy number in the ColE1 family of bacterial DNA Plasmids. RNA 1 (about 100 nucleotides long) is a regulatory molecule that inhibits the activity of RNA 2 (about 500 nucleotides long), which is required to initiate plasmid DNA replication. RNA 1 is complementary to a sequence at the 5' end of RNA 2, and its concentration increases in proportion to the number of plasmid DNA molecules in the cell. In RNA 2, sequence 2 is complementary to both sequence 1 and sequence 3 (cf. Fig. 10-50) and can be displaced upon binding of RNA 1; this alters the conformation of sequence 4, inactivating RNA 2. (After H. Masukata and J. Tomizawa, Cell 44: 125-136, 1986.)

RNA molecules also perform regulatory Functions. Antisense RNA introduced into experimentally altered cells renders them unable to express a specific gene (a mechanism analogous to that which normally regulates the expression of certain bacterial genes). This mechanism may, in fact, be much more widespread. A particularly well-studied example of this is the feedback control of DNA replication initiation in a large family of bacterial plasmids. This control system limits plasmid copy number, thereby preventing plasmids from killing the host cell (Fig. 10-62).

The study of RNA-catalyzed reactions is of particular interest for understanding the course of evolution. As discussed in Chapter 1, the earliest cells apparently contained no DNA and had very few, if any, proteins. Many of the RNA-catalyzed reactions in modern cells may represent molecular fossils—that is, relics of the complex network of RNA-directed reactions that presumably dominated cellular metabolism 3.5 billion years ago. By unraveling these processes, biologists may ultimately be able to retrace the origins of the first living cell.

Summary

To control gene expression, cells act on many Stages of the RNA→protein pathway. Most genes are thought to be regulated at multiple levels, although transcriptional initiation control is widely considered predominant. Nevertheless, some genes are transcribed at a constant rate, and their activation and inactivation are achieved solely through post-transcriptional mechanisms affecting RNA. These post-transcriptional regulatory processes include: 1) transcription attenuation via premature termination, 2) alternative splice site selection, 3) 3'-end control via cleavage and polyadenylation, 4) translation initiation control, and 5) regulated mRNA degradation. Most of these control processes require the recognition of specific sequences or structures within the RNA molecule, which is carried out by either a regulatory protein or a regulatory RNA molecule.