Figure 27.10 shows the reaction attaching an amino acid to a tRNA. There are 86 tRNAs in E. coli. Thus, there are more tRNAs than codons; that is, several tRNAs can obviously recognize the same codon. E. coli also has 20 aminoacyl-tRNA synthetases, one for each amino acid. Each synthetase recognizes one amino acid and one or more tRNAs (cognate tRNAs). The synthetases are divided into 10 class I and 10 class II enzymes based on structural features.
Once an amino acid is attached to a tRNA, the specificity for protein synthesis resides solely with the tRNA portion of the aminoacylated tRNA molecule (Figure TD1). For example, cysteine can be put on tRNACys (Cys-tRNACys), chemically modified to the amino acid alanine (Ala-tRNACys), and then used in a biochemical system for protein synthesis. The resulting polypeptides now made in the biochemical reaction have alanine at positions normally occupied by cysteine.
A general model for synthetase-tRNA binding suggests that the protein binds the tRNA along the "side" of the L-shaped molecule. The same general principle applies for all synthetase-tRNA binding: the tRNA is bound principally at its two extremities, and most of the tRNA sequence is not involved in recognition by a synthetase. However, the detailed nature of the interaction is different between class I and class II enzymes. The two types of enzyme approach the tRNA from opposite sides, with the result that the tRNA-protein models look almost like mirror images of one another.
A class I enzyme (Gln-tRNA synthetase) approaches the D-loop side of the tRNA. It recognizes the minor groove of the acceptor stem at one end of the binding site, and interacts with the anticodon loop at the other end. Figure Q1 is a diagrammatic representation of the crystal structure of the tRNAGln-synthetase complex. A revealing feature of the structure is that contacts with the enzyme change the structure of the tRNA at two important points. These can be seen by comparing the dotted and solid lines in the anticodon loop and acceptor stem:
Bases U35 and U36 in the anticodon loop are pulled farther out of the tRNA into the protein.
The end of the acceptor stem is seriously distorted, with the result that base pairing between U1 and A72 is disrupted. The single-stranded end of the stem pokes into a deep pocket in the synthetase protein, which also contains the binding site for ATP.
This structure explains why changes in U35, G73, or the U1-A72 base pair affect the recognition of the tRNA by its synthetase. At all of these positions, hydrogen bonding occurs between the protein and tRNA.
A class II enzyme (Asp-tRNA synthetase) approaches the tRNA from the other side, and recognizes the variable loop, and the major groove of the acceptor stem, as drawn in Figure Q2. The acceptor stem remains in its regular helical conformation. ATP is probably bound near to the terminal adenine. At the other end of the binding site, there is a tight contact with the anticodon loop, which has a change in conformation that allows the anticodon to be in close contact with the protein.
Many attempts to deduce similarities in sequence between cognate tRNAs, or to induce chemical alterations that affect their charging, have shown that the basis for recognition is different for different tRNAs, and does not necessarily lie in some feature of primary or secondary structure alone. We know from the crystal structure that the acceptor stem and the anticodon stem make tight contacts with the synthetase, and mutations that alter recognition of a tRNA are found in these two regions. (The anticodon itself is not necessarily recognized as such; for example, "suppressor" mutations change a base in the anticodon, and therefore the codons to which a tRNA responds, without altering its charging with amino acids.)
A group of isoaccepting tRNAs must be charged only by the single aminoacyl-tRNA synthetase specific for their amino acid. So isoaccepting tRNAs must share some common feature(s) enabling the enzyme to distinguish them from the other tRNAs. The entire complement of tRNAs is divided into 20 isoaccepting groups; each group is able to identify itself to its particular synthetase.
tRNAs are identified by their synthetases by contacts that recognize a small number of bases, typically from 1-5. Three types of feature commonly are used (Figure TD4):
Usually (but not always), at least one base of the anticodon is recognized. Sometimes all the positions of the anticodon are important.
Often one of the last three base pairs in the acceptor stem is recognized. An extreme case is represented by alanine tRNA, which is identified by a single unique base pair in the acceptor stem.
The so-called discriminator base, which lies between the acceptor stem and the CCA terminus, is always invariant among isoacceptor tRNAs.
No single one of these features constitutes a unique means of distinguishing 20 sets of tRNAs, or provides sufficient specificity.
Several synthetases can specifically charge a "minihelix" consisting only of the acceptor and TpseudoUC arms (equivalent to one arm of the L-shaped molecule) with the correct amino acid. The efficiency of aminoacylation of these substrates is much higher for the class II enzymes. For these tRNAs, specificity depends exclusively upon the acceptor stem. However, it is clear that there are significant variations between tRNAs, and in some cases the anticodon region is important. Mutations in the anticodon can affect recognition by the class II Phe-tRNA synthetase. Multiple features may be involved; minihelices from the tRNAVal and tRNAMet (where we know that the anticodon is important in vivo) can react specifically with their class I synthetases.
So recognition depends on an interaction between a few points of contact in the tRNA, concentrated at the extremities, and a few amino acids constituting the active site in the protein. The relative importance of the roles played by the acceptor stem and anticodon is different for each tRNA-synthetase interaction.
The nature of discriminatory events is a general
issue raised by several steps in gene expression. How do synthetases
recognize just the corresponding tRNAs and amino acids? How does
a ribosome recognize only the tRNA corresponding to the codon
in the A site?
Each case poses a similar problem: how to distinguish one particular
member from the entire set, all of which share the same general
features.
Probably any member initially can contact the active center by a random-hit process, but then the wrong members are rejected and only the appropriate one is accepted. The appropriate member is always in a minority (1 of 20 amino acids, 1 of 86 tRNAs), so the criteria for discrimination must be strict. We can imagine two general ways in which the decision whether to reject or accept might be taken:
The cycle of admittance, scrutiny, rejection/acceptance could represent a single binding step that precedes all other stages of whatever reaction is involved. This is tantamount to saying that the affinity of the binding site is sufficient to control the entry of substrate. In the case of synthetases, this would mean that only the cognate tRNAs could form a stable attachment at the site.
Alternatively, the reaction proceeds through some of its stages, after which a decision is reached on whether the correct species is present. If it is not present, the reaction is reversed, or a bypass route is taken, and the wrong member is expelled. This sort of postbinding scrutiny is generally described as proofreading. In the example of synthetases, it would require that the charging reaction proceeds through certain stages even if the wrong tRNA or amino acid is present.
Transfer RNA binds to synthetase by the two-stage
reaction depicted in Figure Q3. Cognate
tRNAs have a greater intrinsic affinity for the binding site,
so they are bound more rapidly and dissociate more slowly. Following
binding, the enzyme scrutinizes the tRNA that has been bound.
If the correct tRNA is present, binding is stabilized by a conformational
change in the enzyme. This allows aminoacylation to occur rapidly.
If the wrong tRNA is present, the conformational change does not
occur. As a result, the reaction proceeds much more slowly; this
increases the chance that the tRNAwill dissociate from the enzyme
before it is charged. This type of control is called kinetic proofreading.
Specificity for amino acids varies among the synthetases. Some are highly specific for initially binding a single amino acid, but others can also activate amino acids closely related to the proper substrate. Although the analog amino acid can sometimes be converted to the adenylate form, in none of these cases is an incorrectly activated amino acid actually used to form a stable aminoacyl-tRNA (Figure TD2). The presence of the cognate tRNA usually is needed to trigger proofreading, even if the reaction occurs at the stage before formation of aminoacyl-adenylate. (An exception is provided by Met-tRNA synthetase, which can reject noncognate aminoacyl-adenylate complexes even in the absence of tRNA.)
There are two stages at which proofreading of an incorrect aminoacyl-adenylate may occur during formation of aminoacyl-tRNA. Figure Q4 shows that both use chemical proofreading, in which the catalytic reaction is reversed. The extent to which one pathway or the other predominates varies with the individual synthetase:
The noncognate aminoacyl-adenylate may be hydrolyzed when the cognate tRNA binds. This mechanism is used predominantly by several synthetases, including those for methionine, isoleucine, and valine. (Usually, the reaction cannot be seen in vivo, but it can be followed for Met-tRNA synthetase when the incorrectly activated amino acid is homocysteine [which lacks the methyl group of methionine]. Proofreading releases the amino acid in an altered form, as homocysteine thiolactone. In fact, homocysteine thiolactone is produced in E. coli as a byproduct of the charging reaction of Met-tRNA synthetase. This shows that continuous proofreading is part of the process of charging a tRNA with its amino acid.)
Some synthetases use chemical proofreading at a later stage. The wrong amino acid is actually transferred to tRNA, is then recognized as incorrect by its structure in the tRNA binding site, and so is hydrolyzed and released. The process requires a continual cycle of linkage and hydrolysis until the correct amino acid is transferred to the tRNA.
A classic example in which discrimination between amino acids depends on the presence of tRNA is provided by the Ile-tRNA synthetase of E. coli. The enzyme can charge valine with AMP, but hydrolyzes the valyl-adenylate when tRNAIle is added. The overall error rate depends on the specificities of the individual steps, as summarized in Figure Q5. The overall error rate of 1.5 X 10-5 is less than the measured rate at which valine is substituted for isoleucine (in rabbit globin), which is about 2-5 X 10-4. So mischarging probably provides only a small fraction of the errors that actually occur in protein synthesis.
In the case of Ile-tRNA synthetase, accuracy is ensured by multiple mechanisms. The crystal structure of the enzyme shows that the activation site is too small to allow leucine (a close analog of isoleucine) to enter. Adjacent to the activation site is a second site, the hydrolytic site, which is too small to allow isoleucine itself to enter. However, valine can enter this site, and as a result an incorrect Val-tRNAIle is hydrolyzed. The activation and hydrolytic sites form a closed cavity within the enzyme. The combination of mechanisms provides a molecular sieve, in which size of the amino acid is important.
(See also Table 27.2 on page 1031)
|
|
||
|
|
|
|
| G | pairs with | C or U |
| C | pairs with | G |
| A | pairs with | U |
| U | pairs with | A or G |
| I (inosine) | pairs with | A, U, or C |
Figure AN shows the three E. coli genes for glycyl-tRNAs and their codon-anticodon interactions.
See page 1052 for a discussion about nonsense suppressors.