Metamagical Themas, page 90
FIGURE 27-4. A strand of messenger RNA, with some short double-helical regions where it is linked to itself by hydrogen bonds (indicated by dotted lines). [Drawing by David Moser. I
of thousands of nucleotides strung together like beads. (See Figure 27-4.) An mRNA chain generally codes for several proteins and has special markers along it telling where the stretches representing the various proteins begin and end. That is why there are three special codons that do not stand for any amino acid. They act a little bit like semicolons, in that they convey to the ribosome: "Cut this protein off right now; don't add a single amino acid more!"
We are coming to the crux of the matter. Where is the genetic code stored? I have made it sound as if ribosomes "know" the code, but they do not. Although ribosomes do the translating, they know neither language involved. How can this be?
* * *
Imagine yourself at the United Nations (see Figure 27-5a). An important speech is about to be given by Mr. Na, the flamboyant ambassador from Nucleotidia. A simultaneous interpreter of great skill, Meri Boso, is summoned. Unfortunately, Ms. Boso has no knowledge of either the language the speech will be in or the language it must be translated into. It looks bad! But at the last moment, just before the speech begins, the members of a rescue team rush into the translating booth, where they suspend from the ceiling a huge number of tiny flash cards. Each card has on its front a word of Nucleotidian (curiously, all the words consist of three letters) and on its back the word's translation into the target language, which happens to be Aminoacidian. Meri Boso is saved! All she must do is listen carefully to Mr. Na and then, for each word she hears, find with lightning speed its flash card. Having found the card, she deftly flips it around so that she can speak its Aminoacidian translation in the nick of time into the microphone before her. Next word, please!
FIGURE 27-5. Two views of the activity of translation. In (a), Meri Boso in her U.N. translating booth translates Mr. Na's speech from Nucleotidian into Aminoacidian, using the flash cards dangling all about her. In (b), a ribosome chugs down an mRNA strand, surrounded by tRNA molecules. As it reads each codon, it locates a nearby tRNA molecule with the matching anticodon; from that tRNA, it strips off the amino acid, attaching it to a growing protein that is emerging in its folded form into the cytoplasm. [Drawings by David Moser. I
It's no sweat, being a ribosome. All you need to do is find the right flash card in a jiffy. But where are the flash cards in the cell? Even more to the point, what are they? At this juncture, it seems that the genetic code has receded from view a little; it has become more decentralized, harder to localize. Whereas at first one might have guessed that the genetic code was somehow stored inside each ribosome in the chemical equivalent of a tablet or dictionary, now it seems to lie in those flash cards. So if we want to determine how arbitrary the genetic code is, we must determine whether the flash cards could be changed and, if so, how.
The cell's flash cards are tRNA's (that is, transfer RNA's). The term suggests that they are made out of the same stuff as mRNA is: A's, C's, G's, and U's. This is true, except that some nucleotides are occasionally modified by enzymes, but for our purposes we can ignore that fine detail. At birth, a tRNA is just an ordinary snippet of RNA, but it is much shorter than an mRNA train. Also, quite unlike mRNA, which stays long and snaky, a young tRNA folds up just like a protein, assuming a specific tertiary structure. This is in contrast to mRNA, which merely forms rather aimless curls over short stretches. The curling-up of mRNA is nonfunctional, whereas the curling of tRNA is functional. (Or rather, we don't yet know much about mRNA's curling; probably it is functional but in a more cryptic or subtle way.) All tRNA's fold up into roughly the same shape: a chubby `L', rather like the bent arm of Mr. America. At a more detailed level, however, the tertiary structures of tRNA's differ. In Figure 27-6, you will find a series of pictures of tRNA at various levels of abstraction.
FIGURE 27-6. Transfer RNA, viewed at three levels of abstraction. In (a), physically the most realistic, the three-dimensional structure as it has been revealed by X-ray diffraction techniques. In (b), a more schematic "cloverleaf " representation, showing the various hydrogenbonded loops, as well as the amino-acid attachment site and the anticodon. In (c), the most schematic representation of all, a tRNA molecule is portrayed in its barest functionality: a molecule labeled at one end by an anticodon and potentially carrying at its other end an amino acid. [Drawings by David Moser. ]
Once it is folded up, a tRNA acts like a flash card, in that it has an amino acid at one end of the'L', and a codon at the other. Actually, it is not a codon but an anticodon. An anticodon is to a codon as a photographic negative is to a positive, or an engraving to a bas-relief. To make one from the other, you merely interchange A with U, and C with G. (A and U are said to be complementary, as are C and G.) Therefore CUC and GAG are each other's anticodons. To be more explicit about tRNA, one end of it simply is an anticodon. The other end is a site where an amino acid can be attached. And if you're wondering who does the attaching, you'll find out soon enough.
* * *
In a nutshell, a ribosome is a translating mechanism between the two intracellular languages of Nucleotidian and Aminoacidian. The words of Nucleotidian are codons; the words of Aminoacidian are amino acids. The mRNA is a long speech whose sentences are written in Nucleotidian. The ribosome is a quick but ignorant simultaneous interpreter who, guided by tRNA molecules, assembles proteins, which are the word-by-word translations of the mRNA sentences into Aminoacidian. (By "quick" I mean the following. Under normal conditions, a ribosome in a bacterial cell can translate about twenty codons per second. In a rabbit cell, things are slower: a little better than one codon per second. I have no idea why rabbits are so much slower than -bacteria.)
As is shown in outline in Figure 27-5b and in more detail in Figure 27-7, an mRNA "speech" is constantly clicking through the ribosome, one codon at a time. On encountering a new codon, the ribosome must seek out a matching tRNA, one whose anticodon perfectly fits the codon. Of course a ribosome has no eyes, and cannot scan about as Meri Boso does. It must try one tRNA after another (again, think of Cinderella and her slipper). A mystery is how a ribosome can find a matching tRNA so quickly. In any case, having found one and clicked its anticodon into position against the mRNA codon, the ribosome snips off the tRNA's amino acid and snaps it onto the growing protein chain; then it releases the "nude" tRNA, which is free to pick up a new amino acid.
This is a salient difference between the metaphorical flash cards and tRNA molecules. Whereas flash cards can be used over and over again, each time a tRNA molecule gets used, it has to be "recharged" with the right amino acid. Just where and how does this take place? Which amino acid should it get charged with? How is this determined? Who determines it? All of a sudden, these questions loom large, because they have everything to do with the link between a codon and its amino acid. We shall return to them shortly.
It is now apparent that if the genetic code is stored anywhere, it is in a spread-about fashion, distributed among the thousands of tRNA's floating in suspension in the cell near the ribosomes. Could these tRNA's somehow be subverted? Could they falsely guide the translation process? Certainly we
FIGURE 27-7. The intracellular translation process, in more detail. Inside this ribosome, one can see the matching-up of one tRNA's anticodon (while) with an mRNA codon (black). The amino acid at the top of the tRNA has just been snapped onto the growing chain of amino acids with a 'peptide bond", symbolized by the curly link between rectangles. [Drawing by David Moser. ]
can imagine the UN rescue team rushing in with the wrong set of flash cards, hanging them all up in Ms. Boso's booth, and her then translating Mr. Na's speech into a completely inappropriate language. Could the analogue happen in a cell?. Could there conceivably be produced an entire set of "bad" tRNA's: tRNA's with wrong amino acids attached to them, tRNA's that would fool the ribosomes into manufacturing nonsensical proteins? Who could perpetrate such a nasty practical joke?
* * *
Well, this is the stage I was at when I started drawing pictures on the blackboard for my students. I drew a typical tRNA molecule and stated that at one end-its AA end-it would attract a particular amino acid. But why should it attract the right amino acid? Simple enough, I thought to myself. As with most chemical affinities in the cell, the AA end of the tRNA would simply have the right shape. Each tRNA would lure only the amino acid that (by the genetic code) corresponds to its anticodon. My supposition was that for each anticodon, the tRNA that carried it would be shaped differently at its AA end. And so that's what I drew on the board: a tRNA molecule with
FIGURE 27-8. My first guess at how tRNA molecules manage (nearly always) to have just the right amino acid attached at their AA ends. In this appealing but simplistic theory, which turned out to be completely wrong, the AA end and the desired amino acid are like lock and key. Thus, only the desired amino acid will fit a given tRNA's AA end. The truth of the matter is that the AA end of a tRNA is completely nonspecific: it will accept any of the twenty types of amino acid. [Drawing by David Moser. ]
a specific anticodon at one end and a specific "attractive shape" at the other end, a shape that would presumably combine with just one kind of amino acid. (See Figure 27-8.)
Here a good question arises. Why should each tRNA attract the right amino acid for its anticodon, "right" being the amino acid defined by the genetic code? Why couldn't some tRNA fold up in such a way as to attract some other amino acid? Or is there some intrinsic connection between the two ends of the tRNA? Does the anticodon, for instance, somehow "tell" the other end of the tRNA how to fold up? This was one thought Vahe had, and it would imply that codons and amino acids really had some direct chemical associations. But I didn't believe that for a moment.
I told my class that neither end of the tRNA could possibly know anything about the other. I insisted that you could surgically replace the anticodon with some other anticodon, and the AA end would not know the difference. Conversely, you could surgically lop off the specially shaped AA tip of the tRNA and graft on an alien AA tip, which would then lure the wrong amino acid, thereby making the tRNA embody a false piece of genetic code. I concluded by saying: "Since the two ends of any tRNA are independent, the genetic code can in principle be subverted and is therefore arbitrary. " Then I blew the chalk dust off my hands and turned to another topic.
Well, it turns out that this picture I had drawn was right in spirit but wrong in detail. Contrary to my supposition, all tRNA molecules have at their AA tip precisely the same structure! For instance, the last three nucleotides at the AA tip are always CCA (glance back at Figure 27-6). Thus, the site where the amino acid gets attached is completely nonspecific. There is no special chemical affinity between the AA tip of a tRNA and the amino acid that goes there! When I first found this out (after class was over), I was somewhat at a loss. How, I wondered, does the tRNA always end up with the right amino acid attached to it? What lures it there? Could it be the anticodon, even though it is at the other end of the tRNA? And if so, does that mean that there is, as Vahe surmised, some special and intrinsic link between the anticodon and its amino acid partner? Is the genetic code, after all, inevitable?
* * *
By talking with biologist friends and looking in books, I found the answer. In the end, it seemed to come out supporting my side, but matters turned out to be far subtler and murkier than I had suspected. Although the AA end of a tRNA molecule is indifferent to the amino acid that docks there, so that in principle it can accept any amino acid, under normal circumstances only one amino acid will get attached. However, this is due not to the anticodon but to the tertiary structure of another region of the tRNA: its DHU loop. ("DHU" stands for "dihydrouridine", in case you were curious.) This is a loop that every tRNA molecule has, and it bends around io a characteristic shape in each different kind of tRNA. It is therefore a kind of three-dimensional signature by which the tRNA's type can be recognized from the outside. (Actually, as it turns out, probably considerably more is involved in tRNA recognition than just the DHU loop, but for simplicity's sake, I will here continue to speak as if that were the entire story.)
But who could accomplish such recognition? Why, an. enzyme, of course -in fact, an aminoacyl-tRNA synthetase. (Sorry about that! But despite the strangeness of this name, you should try to remember it, because these molecules turn out to play, if not the starring roles in our saga, then certainly pivotal roles.) Such an enzyme has two active sites. One of them recognizes the tRNA's three-dimensional signature, and the other looks for an amino acid. That site, unlike the AA end of the tRNA, is not indifferent to the amino acid. It will bind one and only one amino acid-namely, the one coded for by the tRNA's anticodon. To be sure, the synthetase itself never looks at the anticodon. All it does is "sniff" the DHU loops of various tRNA's (and perhaps other substructures as well-as I said, this is still not entirely clear) and when it finds one it "likes", it fastens its amino acid tightly to the tRNA and bids it farewell. For each type of amino acid, there is at least one type of synthetase.
So here we have a funny thing. There are molecules floating around in the cell whose purpose it is to "instruct" the tRNA's in the genetic code. They load up each tRNA with an appropriate amino-acidic burden and then let it trudge off to encounter a ribosome somewhere. So ... Do the tRNA's know the genetic code? No; they have to be instructed. And who instructs them? The synthetases. Well, then ... Do the synthetases know the genetic code? No; they merely match up DHU loops of various shapes with amino acids. So in the end, we find out that nobody in the cell knows the genetic code!
Of course, that has to be an exaggeration. The truth, again, is simply that "knowledge" of the genetic code is extremely spread out. It is shared by the entire set of tRNA's and synthetases, and cannot be claimed by either one alone. And yet, there is one place where one might contend that the genetic code is stored all in one piece . . . "And where, pray tell, is that ?" you ask. Ah-it is the DNA. You might have been wondering when we would come to DNA, usually the star in tales of molecular biology. Well, this is the moment.
* * *
One might regard DNA as a big, fat, aristocratic, lazy, cigar-smoking slob of a molecule. It never does anything. It is the ultimate "lump" of the cell. It merely issues orders, never condescending to do anything itself, quite like a queen bee. How did it get such a cushy position? By ensuring the production of certain enzymes, which do all the dirty work for it. And how can it make sure that these desirable enzymes will get produced? Ah, that is the trick.
DNA is a set of blueprints for all kinds of cellular constituents, lumps and doers alike. If you want to know where something in a cell comes from, the chances are the answer is: It is coded for in the DNA. The piece of DNA that codes for some specific entity is that entity's gene. The entity may be a protein, it may be a tRNA molecule, or it may be some RNA that will eventually become part of a ribosome. Whatever the constituent is, there is a gene for it in the long, twisty DNA molecule. Indeed, that is why DNA is so long. The length of the DNA for a mere bacterium can be a million nucleotides-and for a human being, thousands of times longer than that! A DNA strand therefore consists of a sequence of thousands, millions, or even billions of codons, constituting anywhere from a handful of genes to many thousands of them, arranged sequentially, like sentences following one another in a book, or songs in the grooves of a record.
FIGURE 27-9. DNA at two levels of abstraction.
In (a), an architectonic view, showing the famous double helix. The outer winding staircases are formed of non-informational matter (sugars and phosphates), while the inner core, represented here by hydrogen-bonded spheres, is where all the genes are stored, defining the entire nature of the cell or organism within which all this is occurring.
In (b), the helices are uncoiled but no bonds are broken. This 'flattened" DNA is then spread out like a rug on a floor, allowing you to see exactly how the bases join up with each other in complementary pairs ("Watson-Crick bonding"). [Drawings by David Moser. ]
DNA, like RNA, is made up of nucleotides, but instead of U it uses T (which stands for "thymine"). In DNA, A and T, like C and G, are complementary. For every strand of DNA there is a complementary strand that twists around it, making the entire supermolecule look like a double vine (see Figure 27-9). The reason DNA does this while RNA does not is that A and U do not fit together as tightly as A and T do, and so the twists of any would-be RNA double helix are not as stable as those in DNA. Actually, RNA can form a double helix for short stretches, but not for long ones. That is also why tRNA's have short double-helical hairpin turns but are not double helices all the way.
I've said several times that an entity's gene is a coded version of the entity. Now where there is code, there must be decoding. But attention: There are two possible depths of decoding, for a stretch of DNA. (See Figure 27-10.) First of all, you can decode it into RNA. This is the shallower way, and it is done merely by complementation: A codes for U, T for A, and C and G for each other. Thus, DNA stretch "TCAT" becomes RNA stretch "AGUA", which can come in handy if ever you're parched in Paraguay. The deeper way of decoding DNA involves a second layer of decoding (shades of the two layers of decoding of Enigma messages, described in Chapter 21)-that is, one must go further, and decode the message contained in the RNA. That, of course, is the job of all the Meri Bosos and their tRNA flash cards.
