Processing of precursor RNA

As well as the mature RNAs described above, cells also contain precursor molecules (see Figure 3.3). Many RNAs, especially in eukaryotes, are initially synthesized as precursor or pre-RNA, which has to be processed before it can carry out its function. The various processing events, all of which are described in detail in Chapter 10, include the following (Figure 3.7):

Figure 3.7

Schematic representation of the four types of RNA processing event. Not all events occur in all organisms - see the text for details.

• End-modifications occur during the synthesis of eukaryotic and archaeal mRNAs, most of which have a single, unusual nucleotide called a cap attached at the 5′ end and a poly(A) tail attached to the 3′ end. The cap and poly(A) tails are involved in assembly of the translation initiation complex on these mRNAs (Section 10.2.2).
• Splicing is the removal of introns from a precursor RNA. Many eukaryotic protein-coding genes contain introns and these are copied when the gene is transcribed. The introns are removed from the pre-mRNA by cutting and joining reactions. Unspliced pre-mRNA forms the nuclear RNA fraction called heterogenous nuclear RNA (hnRNA). Some eukaryotic pre-rRNAs and pre-tRNAs also contain introns, as do some archaeal transcripts, but they are extremely rare in bacteria.
• Cutting events are particularly important in the processing of rRNA and tRNA, many of which are initially synthesized from transcription units that specify more than one molecule. The pre-rRNAs and pre-tRNAs must therefore be cut into pieces to produce the mature RNAs. This type of processing occurs in both prokaryotes and eukaryotes.
• Chemical modifications are made to rRNAs, tRNAs and mRNAs. The rRNAs and tRNAs of all organisms are modified by addition of new chemical groups, these groups being added to specific nucleotides within each RNA. Chemical modification of mRNA, called RNA editing, is seen in a diverse group of eukaryotes.

Processing of mRNA has an important influence on the composition of the transcriptome. RNA editing, for example, can result in a single pre-mRNA being converted into two different mRNAs coding for quite distinct proteins (Figure 3.8A). This does not appear to be particularly common, but alternative splicing, in which one pre-mRNA gives rise to two or more mRNAs by assembly of different combinations of exons (Figure 3.8B), is fairly widespread. The mRNAs resulting from both editing and alternative splicing often display tissue specificity, the processing events increasing the coding capabilities of the genome without the requirement for an increased gene number.

Figure 3.8

Some RNA processing events change the coding properties of an mRNA. (A) RNA editing can change the sequence of an mRNA, resulting in synthesis of a different protein. An example occurs with the human mRNA for apolipoprotein B, as shown in Figure 10.29. (more...)

Studies of the yeast transcriptome

With less than 6000 genes, the yeast Saccharomyces cerevisiae is ideally suited for transcriptome studies, and many of the pioneering projects have been carried out with this organism. One of the first discoveries was that, although mRNAs are being degraded and re-synthesized all the time, the composition of the yeast transcriptome undergoes very little change if the biochemical features of the environment remain constant (DeRisi et al., 1997). When yeast is grown in a glucose-rich medium, which allows the cells to divide at their maximum rate, the transcriptome is almost completely stable, only 19 mRNAs displaying a greater than two-fold change in abundance over a period of 2 hours (Figure 3.9). Significant alterations to the transcriptome are seen only when the glucose in the growth medium becomes depleted, forcing the cells to switch from aerobic to anaerobic respiration. During this switch, the levels of over 700 mRNAs increase by a factor of two or more, and another 1000 mRNAs decline to less than half their original amount. The changing environment clearly results in a restructuring of the transcriptome to meet the new biochemical demands of the cell.

Figure 3.9

The yeast transcriptome is stable during growth of cells in a glucose-rich medium, but undergoes significant changes when the glucose is used up and the cells switch from aerobic to anaerobic respiration.

The yeast transcriptome also undergoes restructuring during cellular differentiation. This has been established by studying sporulation (spore formation), which is induced by starvation and other stressful environmental conditions. The sporulation pathway can be divided into four stages - early, middle, mid-late and late - on the basis of the morphological and biochemical events that occur (Figure 3.10). Previous studies have shown, not unexpectedly, that each stage is characterized by expression of a different set of genes. Transcriptome studies have added to our understanding of the sporulation process in several ways (Chu et al., 1998). Most significantly, the changes that occur to the composition of the transcriptome indicate that the early stage of sporulation can be subdivided into three distinct phases, called early (I), early (II) and early-middle. The levels of over 250 mRNAs increase significantly during early sporulation, and another 158 mRNAs increase specifically during the middle stage. A further 61 mRNAs increase in abundance during the mid-late period and five more during the late phase. There are also 600 mRNAs that decrease in abundance during sporulation, these presumably coding for proteins that are needed during vegetative growth but whose synthesis must be switched off when spores are being formed.

Figure 3.10

The sporulation pathway of Saccharomyces cerevisiae. The middle three drawings show the nuclear divisions that occur during sporulation. See Figure 5.15 for details of the events involved in meiosis I and meiosis II.

This work on yeast sporulation is important for two reasons. First, by describing the changes in genome expression that occur during sporulation, the transcriptome analyses open the way to studies of the interactions between the genome and the environmental signals that trigger sporulation. Studies of this type, in a relatively simple organism such as yeast, act as an important model for the more complex developmental processes that operate in higher eukaryotes, including humans. Secondly, several of the mRNAs whose levels change significantly during sporulation are transcripts of genes whose functions were previously unknown. Transcriptome studies therefore help to annotate a genome sequence, aiding identification of genes whose roles in the genome have not been determined by other methods. We will return to this issue in Chapter 7.

The human transcriptome

With six times as many genes, the human transcriptome is substantially more complex than that of yeast, and studies of its composition are still in their infancy. Some interesting results have, nonetheless, been obtained. For example, the transcriptomes of eight different cell types have been mapped onto the draft human genome sequence (Caron et al., 2001), resulting in global views of the pattern of gene expression along entire chromosomes (Figure 3.11). As well as providing a ‘blueprint’ for the transcriptomes of each of the eight cell types, this work has underlined the extent to which the transcriptome of a cancerous tissue differs from that of the normal version. Transcriptome restructuring as a result of cancer was first discovered several years ago, when it was shown that 289 mRNAs are present in significantly different amounts in the transcriptomes of normal colon epithelial cells compared with cancerous colon cells, and that about half of these mRNAs also display an altered abundance in pancreatic cancer cells (Zhang et al., 1997). It is hoped that by understanding the differences between the transcriptomes of normal and cancerous cells it will be possible to devise new ways of treating the cancers.

Figure 3.11

Comparison of the transcriptomes of different types of human cell. The diagram shows human chromosome 11 aligned vertically. The bar charts indicate the expression levels in different cell types of the genes on this chromosome. The lengths of the blue (more...)

Transcriptome studies also have applications in cancer diagnosis. The initial breakthrough in this respect came in 1999 when it was shown that the transcriptome of acute lymphoblastic leukemia cells is different from that of acute myeloid leukemia cells (Golub et al., 1999). Twenty-seven lymphoblastic and eleven myeloid cancers were studied and, although all the transcriptomes were slightly different, the distinctions between the two types were sufficient for unambiguous identifications to be made. The significance of this work lies with the improved remission rates that are achievable if a cancer is identified accurately at an early stage, before clear morphological indicators are seen. This is not relevant with these two types of leukemia because these can be distinguished by non-genetic means, but it is important with other cancers such as non-Hodgkin lymphoma. The commonest version of this disease is called diffuse large B-cell lymphoma, and for many years it was thought that all tumors of this type were the same. Transcriptome studies changed this view and showed that B-cell lymphoma can be divided into two distinct subtypes (Alizadeh et al., 2000). The distinctions between the transcriptomes of the two subtypes enable each one to be related to a different class of B cells, stimulating and directing the search for specific treatments that are tailored for each lymphoma.

The four levels of protein structure

Proteins are traditionally looked upon as having four distinct levels of structure. These levels are hierarchical, the protein being built up stage by stage, with each level of structure depending on the one below it:

1.

The primary structure of the protein is formed by joining amino acids into a polypeptide. The amino acids are linked by peptide bonds which are formed by a condensation reaction between the carboxyl group of one amino acid and the amino group of a second amino acid (Figure 3.14). In passing, note that, as with a polynucleotide, the two ends of the polypeptide are chemically distinct: one has a free amino group and is called the amino, NH ₂ -, or N terminus; the other has a free carboxyl group and is called the carboxyl, COOH-, or C terminus. The direction of the polypeptide can therefore be expressed as either N→C (left to right in Figure 3.14) or C→N (right to left in Figure 3.14).

2.

The secondary structure refers to the different conformations that can be taken up by the polypeptide. The two main types of secondary structure are the α-helix and β-sheet (Figure 3.15), both of which are stabilized by hydrogen bonds that form between different amino acids in the polypeptide. Most polypeptides are long enough to be folded into a series of secondary structures, one after another along the molecule.

3.

The tertiary structure results from folding the secondary structural components of the polypeptide into a three-dimensional configuration (Figure 3.16). The tertiary structure is stabilized by various chemical forces, notably hydrogen bonding between individual amino acids, and hydrophobic forces, which dictate that amino acids with non-polar (i.e. ‘water-hating’) side-groups must be shielded from water by embedding within the internal regions of the protein. There may also be covalent linkages called disulfide bridges between cysteine amino acids at various places in the polypeptide.

4.

The quaternary structure involves the association of two or more polypeptides, each folded into its tertiary structure, into a multi-subunit protein. Not all proteins form quaternary structures, but it is a feature of many proteins with complex functions, including several involved in genome expression. Some quaternary structures are held together by disulfide bridges between different polypeptides, but many proteins comprise looser associations of subunits stabilized by hydrogen bonding and hydrophobic effects, and can revert to their component polypeptides, or change their subunit composition, according to the functional requirements.

Figure 3.14

In polypeptides, amino acids are linked by peptide bonds. The drawing shows the chemical reaction that results in two amino acids becoming linked together by a peptide bond. The reaction is called a condensation because it results in elimination of water. (more...)

Figure 3.15

The two main secondary structural units found in proteins: (A) the α-helix, and (B) the β-sheet. The polypeptide chains are shown in outline with the positions of the α-carbons indicated by small dots. The R groups have been omitted (more...)

Figure 3.16

The tertiary structure of a protein. This imaginary protein structure comprises three α-helices, shown as coils, and a four-stranded β-sheet, indicated by the arrows. Redrawn from Turner et al. (1997).

Amino acid diversity underlies protein diversity

Proteins are functionally diverse because the amino acids from which proteins are made are themselves chemically diverse. Different sequences of amino acids therefore result in different combinations of chemical reactivities, these combinations dictating not only the overall structure of the resulting protein but also the positioning on the surface of the structure of reactive groups that determine the chemical properties of the protein.

Amino acid diversity derives from the R group because this part is different in each amino acid and varies greatly in structure. Proteins are made up from a set of 20 amino acids (Figure 3.17; Table 3.1). Some of these have R groups that are small, relatively simple structures such as a single hydrogen atom (in the amino acid called glycine) or a methyl group (alanine). Others are large complex aromatic side chains (phenylalanine, tryptophan and tyrosine). Most are uncharged, but two are negatively charged (aspartic acid and glutamic acid) and three are positively charged (arginine, histidine and lysine). Some are polar (e.g. glycine, serine and threonine), others are non-polar (e.g. alanine, leucine and valine).

Figure 3.17

Amino acid R groups. These 20 amino acids are the ones that are conventionally looked upon as being specified by the genetic code (Section 3.3.2). The classification into non-polar, polar etc. is as described in Lehninger (1970).

Table 3.1

Amino acid abbreviations.

The 20 amino acids shown in Figure 3.17 are the ones that are conventionally looked on as being specified by the genetic code (Section 3.3.2). They are therefore the amino acids that are linked together when polypeptides are assembled during the protein-synthesis phase of genome expression. However, these 20 amino acids do not on their own represent the limit of the chemical diversity of proteins. The diversity is even greater because of two factors:

• At least one additional amino acid - selenocysteine (Figure 3.18) - can be inserted into a polypeptide chain during protein synthesis, its insertion directed by a modified reading of the genetic code (Section 3.3.2).
• During protein processing, some amino acids are modified by the addition of new chemical groups, for example by acetylation or phosphorylation, or by attachment of large side chains made up of sugar units (Section 11.3.3).

Figure 3.18

The R group of selenocysteine. Selenocysteine is the same as cysteine but with the sulfur replaced with a selenium atom.

Proteins therefore have an immense amount of chemical variability, some of this directly specified by the genome, the remainder arising by protein processing.

Box 3.3

Non-covalent bonds in proteins. The secondary and higher levels of protein structure are stabilized by various types of non-covalent bonds and interactions: Hydrogen bonds are weak electrostatic attractions between an electronegative atom (such as oxygen (more...)

The genetic code specifies how an mRNA sequence is translated into a polypeptide

It was recognized in the 1950s that a triplet genetic code - one in which each codeword or codon comprises three nucleotides - is required to account for all 20 amino acids found in proteins. A two-letter code would have only 4² = 16 codons, which is not enough to account for all 20 amino acids, whereas a three-letter code would give 4³ = 64 codons. It was also assumed, as a working hypothesis, that mRNAs contain non-overlapping series of codons, and that these are colinear with the polypeptides that they encode (Figure 3.19). Although experimental proof of these assumptions was difficult to obtain, all three turned out to be correct. There are qualifications, such as the lack of strict colinearity displayed by many eukaryotic genes because of the presence of introns, but this complication was not appreciated until introns were discovered in 1977, long after the main work on the genetic code had been carried out. This work was completed in the mid-1960s when the meanings of all 64 codons were determined, partly by analysis of polypeptides resulting from translation of artificial mRNAs of known or predictable sequence in cell-free protein-synthesizing systems, and partly by determining which amino acids associated with which RNA sequences in an assay based on purified ribosomes, the protein-RNA complexes that carry out protein synthesis in the cell (Section 11.2). These experiments are described in more detail in Research Briefing 3.1.

Figure 3.19

Early assumptions about the genetic code. It was assumed that the genetic code is triplet, that codons do not overlap, and that there is a colinear relationship between the sequences of an mRNA and the polypeptide it encodes.

Box 3.1

Elucidation of the genetic code. Assigning amino acids to triplet codons was the main objective of molecular biology during the first half of the 1960s. Once it had been accepted that the genetic code is triplet and non-overlapping, and that genes display (more...)

When the work on the genetic code was completed it was realized that the 64 codons fall into groups, the members of each group coding for the same amino acid (Figure 3.20). Only tryptophan and methionine have just a single codon each: all others are coded by two, three, four or six codons. This feature of the code is called degeneracy. The code also has four punctuation codons, which indicate the points within an mRNA where translation of the nucleotide sequence should start and finish (Figure 3.21). The initiation codon is usually 5′-AUG-3′, which also specifies methionine (so most newly synthesized polypeptides start with methionine), although with a few mRNAs other codons such as 5′-GUG-3′ and 5′-UUG-3′ are used. The three termination codons are 5′-UAG-3′, 5′-UAA-3′ and 5′-UGA-3′; these are sometimes called amber, opal and ochre, respectively, these being the whimsical names given to the original Escherichia coli mutants whose analysis led to their discovery.

Figure 3.20

The genetic code. See Table 3.1 for the three-letter abbreviations of the amino acids.

Figure 3.21

The positions of the punctuation codons in an mRNA.

The genetic code is not universal

It was originally thought that the genetic code must be the same in all organisms. The argument was that, once established, it would be impossible for the code to change because giving a new meaning to any single codon would result in widespread disruption of the amino acid sequences of proteins. This reasoning seems sound, so it is surprising that, in reality, the genetic code is not universal. The code shown in Figure 3.20 holds for the vast majority of genes in the vast majority of organisms, but deviations are widespread. In particular, mitochondrial genomes often use a non-standard code (Table 3.2). This was first discovered in 1979 by Frederick Sanger's group in Cambridge, UK, who found that several human mitochondrial mRNAs contain the sequence 5′-UGA-3′, which normally codes for termination, at internal positions where protein synthesis was not expected to stop. Comparisons with the amino acid sequences of the proteins coded by these mRNAs showed that 5′-UGA-3′ is a tryptophan codon in human mitochondria, and that this is just one of four code deviations in this particular genetic system. Mitochondrial genes in other organisms also display code deviations, although at least one of these - the use of 5′-CGG-3′ as a tryptophan codon in plant mitochondria - is probably corrected by RNA editing (Section 10.3.2) before translation occurs (Covello and Gray, 1989; Gualberto et al., 1989).

Table 3.2

Examples of deviations from the standard genetic code.

Non-standard codes are also known in the nuclear genomes of lower eukaryotes. Often a modification is restricted to just a small group of organisms and frequently it involves reassignment of the termination codons (Table 3.2). Modifications are less common among prokaryotes but one example is known in Mycoplasma species. A more important type of code variation in nuclear genomes is context-dependent codon reassignment, which occurs when the protein to be synthesized contains selenocysteine. This applies to many organisms, both prokaryotes and eukaryotes, including humans, because selenoproteins are widespread (see Table 3.3). Selenocysteine is coded by 5′-UGA-3′, which therefore has two meanings because it is still used as a termination codon in the organisms concerned (Table 3.2). The 5′-UGA-3′ codons that specify selenocysteine are distinguished from those that are true termination codons by the presence of a hairpin loop structure in the mRNA, positioned just downstream of the selenocysteine codon in prokaryotes and in the 3′ untranslated region (i.e. the part of the mRNA after the termination codon) in eukaryotes. Recognition of the codon requires interaction between the hairpin and a special protein that is involved in translation of these mRNAs (Low and Berry, 1996).

Table 3.3

Examples of proteins that contain selenocysteine.

Box 3.4

The origin and evolution of the genetic code. The evolution of the genetic code has provoked argument ever since DNA was established as the genetic material back in the 1950s. Many geneticists support the ‘frozen accident’ theory, which (more...)

The amino acid sequence of a protein determines its function

The links between the different levels of protein structure are most clearly understood at the primary-to-secondary level. Because of the chemical properties of their R groups, certain amino acids are more commonly found in α-helices while others have a predisposition for β-sheets. A secondary structure therefore forms around a group of amino acids that favor that particular secondary structure and initiate its formation. It then extends to include adjacent amino acids that either favor the structure or have no strong disinclination towards it, and finally terminates when one or more blocking amino acids, which cannot participate in that particular type of structure, are reached (Figure 3.22). This process of nucleation, extension and delimitation is repeated along the polypeptide until each part of the chain has adopted its preferred secondary structure.

Figure 3.22

Formation of a secondary structure in a polypeptide. An α-helix is shown nucleating at a position containing amino acids that favor helix formation, and extending in either direction until groups of amino acids that block helix formation are reached. (more...)

By identifying which amino acids are most frequently located in which secondary structures, and by studying the structures taken up by small polypeptides of known sequence, biochemists have been able to deduce rules for this level of protein folding, and, to a certain extent, can predict which secondary structures will be adopted by a polypeptide simply by examining its primary sequence (Barton, 1995). It is less easy to predict the outcomes of the next two stages of protein folding, which result in the secondary structural units becoming arranged into the tertiary structure, and tertiary units associating to form quaternary multi-subunit structures. The tertiary structures of most proteins are made up of two or more structural domains, possibly with little interaction between them; these domains are thought to fold independently of one another. Understanding how this occurs is complicated by the fact that many domains include secondary structural units from quite different regions of a polypeptide. But the difficulty in identifying rules for predicting tertiary and quaternary structures does not detract from the fact that these higher levels of structure are determined by the amino acid sequence of the polypeptide. This is illustrated by the ability of proteins that have been unfolded in the test tube, for example by treatment with urea, to refold spontaneously into their correct structures when the treatment is reversed (Section 11.3.1).

Spontaneous refolding does not occur with all proteins and is particularly difficult to achieve with larger ones, the problem appearing to be that the protein can adopt alternative partially folded structures at various stages of the folding process, only one of which leads to the correctly folded tertiary configuration. If the protein makes the wrong ‘choice’ it can end up at a dead end in which it is partially folded in an incorrect manner but from which it cannot escape. In cells, proteins called molecular chaperones aid the folding of other proteins, probably by reducing the likelihood that the protein being folded adopts the wrong intermediate structure (Section 11.3.1). The existence of alternative folding pathways, and of proteins that aid the folding of other proteins, complicates the premise that a protein's folded structure is dictated by its amino acid sequence, but the premise still holds. The role of molecular chaperones is not to impose a new structure on a protein, but merely to increase the efficiency of the protein's natural, sequence-directed, folding pathway.

The multiplicity of protein function

The biological information encoded by the genome finds its final expression in a protein whose biological properties are determined by the spatial arrangement of chemical groups on its surface and within its folded structure. By specifying proteins of different types, the genome is able to construct and maintain a proteome whose overall biological properties form the underlying basis of life. The proteome can play this role because of the huge diversity of protein structures that can be formed, the diversity enabling proteins to carry out a variety of biological functions. These functions include the following:

• Biochemical catalysis is the role of the special type of proteins called enzymes. The central metabolic pathways, which provide the cell with energy, are catalyzed by enzymes, as are the biosynthetic processes that result in construction of nucleic acids, proteins, carbohydrates and lipids. Biochemical catalysis also drives genome expression through the activities of enzymes such as RNA polymerase.
• Structure, which at the cellular level is determined by the proteins that make up the cytoskeleton, is also the primary function of some extracellular proteins. An example is collagen, which is an important component of bones and tendons.
• Movement is conferred by contractile proteins, of which actin and myosin in cytoskeletal fibers are the best known examples.
• Transport of materials around the body is an important protein activity: for example, hemoglobin transports oxygen in the bloodstream, and serum albumin transports fatty acids.
• Regulation of cellular processes is mediated by signaling proteins such as STATs (Section 12.1.2) and by proteins such as activators that bind to the genome and influence the expression levels of individual genes and groups of genes (Section 9.3.2). The activities of groups of cells are regulated and coordinated by extracellular hormones and cytokines, many of which are proteins (e.g. insulin, the hormone that controls blood sugar levels, and the interleukins, a group of cytokines that regulate cell division and differentiation).
• Protection of the body and of individual cells is the function of a range of proteins, including the antibodies, and those proteins involved in the blood clotting response.
• Storage functions are performed by proteins such as ferritin, which acts as an iron store in the liver, and the gliadins, which store amino acids in dormant wheat seeds.

This multiplicity of protein function provides the proteome with its ability to convert the blueprint contained in the genome into the essential features of the life process.