THE GENETIC CODE

FHC. Crick

It gives me very great pleasure to be here to give
The Charles West Lecture, but it causes me a certain degree
of embarrassment because I am, in origin, a physicist and
this is strictly a medical lecture and a medical occasion.

My excuse is that the topic I have chosen - the genetic code -
is one which is of very great importance to all biological
organisms, ourselves included. If you were to press me and
ask exactly how our knowledge of the genetic code will be
important for medicine I could not easily, at this stage,

tell you precisely, but I can make the general forecast that
it will certainly have repercussions in medicine within our
lifetime.

The problem of the genetic code is basically speaking
the problem of protein synthesis. Proteins are important for
two main reasons. The major reason is that all the catalysts
that we find in living cells - all the enzymes - are made of
protein. In addition, in animals, but less so in plants,
proteins also fulfil a structural role. For a simple organism,
guch as a bacterial cell, it is almost true to say that once
you have specified the correct proteins you have done the job
of specifying the whole organism. Of course it is true that
you have got to start with a cell. ‘That is, you must have
pre-existing components of various kinds, but Lives impor-
tant thing is to get all the proteins right. We donee know
exactly how many different proteins there are in, for example,
Escherichia coli, but a reasonable guess would be perhaps two

thousand.
Let us consider briefly the elementary chemistry of pro-
teins, Proteins are an enormous family of molecules, but they
are all built on the same general plan. Each one consists of
a polypeptide chain, having a regular repeating sequence of
atoms, with side chains attached at regular intervals. You
could hardly have a simpler basic structure. However these
side chains are not all the same. Zach protein has a charac-
teristic length of chain, usually several hundred residues long,
but there are only twenty different kinds of side chains common-
ly found in proteins. A protein is in fact made by joining
together amino-acids with the elimination of water. The twenty
kinds of amino-acid are joined together - head to tail - to
give long polypeptide chains. “ach particular protein — your
own haemoglobin, for example - has a particular sequence of
these amino-acid residues along its backbone.

Proteins, therefore, from this point of view are remarkably
simple. How does it come about then that they are so impor-
tant? They are important because, as I have said, they act as
highly specific catalysts and they do this because they can
form intricate and subtle three-dimensional structures. I have
spoken as if the chain was fully extended, but this is not so.
In general, the chain folds up into helices which then fold on
themselves to make té8 cavities lined with active chemi-

feck boty spleber., dr Crdnyle,

cal groups. Bay has a complicated three-—dimensional structure
which holds the haem group, whose function is to carry the
oxygen molecule. It is an interesting fact that this enormous
structure is needed just to hold a small molecule of oxygen.
The reason for this is quite simple - proteins are the main
things the cell knows how to make. Everything else has to be
made by proteins. It is much more difficult to synthesize the

relatively small haem (because it requires several enzymatic
~3-

steps to do this, and for each one a special protein is required )
than to make the protein part of myoglobin.

We have now looked at proteins from two points of view.
Prom the point of view of their chemical structure we see that
although they are large molecules, they are simple. They are
made from a set of twenty standard types of unit, - twenty amino-
acids - joined together in a regular way, head-to-tail. On the
other hand from the point of view of their physical shape, they
are complicated and intricate molecules. ‘We believe that each
protein obtains its own particular shape by folding itself up,
and there is now suggestive evidence for this idea. One of the
great secrets of biochemistry is this ability of make complica-
ted catalysts by a relatively simple method.

The key problem, therefore, in obtaining a particular
protein is to join up the amino-acids in the correct order to
give a polypeptide chain which will then fold itself up into
the oomneot three-dimensional structure. We now know that this
is the most important function of the genetic material. Its
main function - certainly in simple organisms - is to determine
the structure of proteins. The old slogan of Beadle ‘one gene,
one enzyme’ expresses this particular point of view.

It is now reasonably sure that genes are made of nucleic
acid. We only know this quite certainly, I think, for a few
viruses, and for one or two bacteria. ‘Ye surmise that it is
true for higher organisms. Nucleic acid is of two sorts - DNA
(deoxyribonucleic acid) and RNA (ribonucleic acid). The main
genetic material is in fact DNA, Once again, when we look at
its chemical structure we see that it, too, is a polymer.’ It
consists of a very long, regular backbone which goes phosphate-
sugar-phosphate-suzar-phosphate-sugar and so on indefinitely.

To each sugar, at the same point, is attached a little side
group called a base, of which there are four sorts: adenine,
eytosine, guanine, and thymine. RNA is the genetic material
of certain viruses, such as poliomyelitis, and also has other
functions“important Gn the cell)for protein synthesis. It is
very similar to DNA. The main difference is that the sugar is
ribose rather than deoxyribose, and that it contains uracil
instead of thymine. The essential point is that for both DNA
and RNA the backbone is regular and is made by joining groups
together head-to-tail, though for nucleic an opposed to
protein, there are only four different types jcommonly found in
Nature.

How long, typically, is a molecule of nucleic acid?
This is not an easy question to answer. It suffices to say
that for genetic material it is usually at least several thou-
gand residues long and may be very much longer than this. Some-
times the molecules are in the form of a double helix, in which
two chains fit together in an interesting way. This is usual
for DNA. Sometimes, especially for small viruses, the moleculeg
consist‘of a single chain. The little RNA viruses - like polio-
myelitis - are usually single-stranded RNA, and are often about
five or six thousand residues long. Viruses with double-
stranded RNA also occur, and also ones with single-stranded DNA.
Some viruses are bigger than this. For example bacteriophage
T4 has double-stranded DNA as its genetic material. The DNA
is all in one piece and is about 200,000 residues long. It
probably has of the order of 100 separate genes.

We can go to a scale larger and ask about a bacterial cell,
such as & coli. It is one of the more recent and surprising
discoveries that the DNA of such a cell - the genetic material -

6

appears to be all in one long piece having about 3 x 10° base

pairs. The actual length of this DNA is about a millimetre.
~5-

This should be compared with the size of the bacterial cell,
which is about 1 or 2u. There is quite a lot to pack in,

since a millimetre of the genetic material has to go inside a
cell which is only about a thousandth of a millimetre in
length. For man the total length of the DNA is even greater.
If you were to take one of your own cells and put all the DNA
molecules in it end-to-end they would reach a distance of about
two metres.

What, then, is a gene? A gene, on this picture, is 4
certain length of a molecule of nucleic acid. We believe that
the genetic message is expressed by the sequence of the four
different bases. In the morse code you have a dot, a dash and
a space. In nucleic acid you have four symbols and the exact
order of these symbols we believe carries the genetic message.
There must be signals for saying what is the beginning and the
end of a gene but we believe that they, too, are written in
this language of the four letters. So a gene is a stretch of
nucleic acid, typically some thousand or so bases long.

The function of a gene is to instruct the cell how to
make a particular polypeptide chain of a particular protein.
The Sequence Hypothesis says that it does this because the
sequence of bases in the nucleic acid corresponds in some simple
way to the sequence of the amino-acid of the protein. Put
another way, we have to ask the question - how is it that a
sequence of four things can control the sequence of twenty quite
different things? This is like asking, for the morse code, how
ean the sequence of dots and dashes be translated into an
alphabet of twenty-six letters? The expression "The Genetic
Code" means not the message itself but the way that you trans-
late from one language to the other.

You might think that this would be a straightforward
problem to solve. For certain favourable proteins it is possible
to determine their amino-acid sequence. for example, the amino-
acid sequence of ribonuclease, the enzyme from the pancreas
which helps to digest the nucleic acid in the intestines, has
been determined. Another example is the protein of Tobacco
Mosaic Virus, which makes up the shell of the virus. Surely,
you might think, it is not too difficult to determine the amino-
acid sequence of a protein and in addition determine the base
sequence of the gene which controls it, and then compare the two.
The relationship between the two sequences would no doubt be
transparent.

Unfortunately it turns out one cannot easily do this at
this moment. In order to determine sequence you must have a
fairly large supply of whatever it is you are looking at - a
particular protein, for example. It must be pure, or at least
reasonably pure. There must be methods for breaking it down in
a controlled manner so that you can do your problem in bits.
For nucleic acid that is not easy. We do not have a simple
method yet of purifying a single gene. Until recently the best
we could do was to work on a small virus, which has between five
and ten genes, say, and which has about six thousand bases in its
RNA. That is quite different from working on a sequence with
only a few hundred residues. Moreover, nucleic acid chemistry
is rather more difficult than protein chemistry. It also turns
out that it is much more difficult to determine a sequence of
four things than a sequence of twenty things. In a sequence of
twenty things one or two of them are probably rather rare and
this makes it easier to determine the sequence. Ina sequence
of four things one does not have that help. ‘Consequently the

straightforward, direct anproach, outlined above, cannot be used.
We have another technique, however, which we can use.
This is the technique of genetic mapping. [From classical
genetics we know that in higher organisms genes are arranged
in a linear order along a chromosome and this order can be
determined by studying their rate of recombination. The
further two genes are apart, the more often they recombine,
and so by measuring this frequency of recombination one can
measure a genetic map distance. One of the things which one
would infer from what I have said of the genetic material is
that you would also expect to have genetic recombination
between two markers within a single gene, and would therefore
be able to make a genetic map of the sites within a gene.
Naturally these sites would be very close together, and conse-
quently recombination between them would be very rare. To
pick it up one has to use very large populations, and in fact
it is only technically possible in a worthwhile way using micro-
organisms, where one can handle very large populations and
where it is possible to use selective techniques to pick out the
rare recombinants one is trying to find.

By this method it has indeed proved possible to make a
map of the order of sites within a single gene. This has been
done by Benzer using the bacteriophage 74. He has found as
many as several hundred sites within the pair of genes on which
he works. (Whereas in the past people used to concentrate on
one organism and perhaps, if they got a little specialised, on
one chromosome, now people concentrate on one gene and I myself
only work on the left-hand cae of one particular gene). It
was also shown by Benzer (1959 and 1961) that all these sites
were arranged ina linear order.

It is convenient at this point to introduce the word
"eodon" ~ which means the group of bases on the nucleic acid
which codes for one particular amino-acid. As we shall see,
it is likely that each codon consists of three bases. If we
could write down the codons corresponding to each of the twenty
amino-acids such a list would define the genetic code, and
that would be the end of my lecture. Unfortunately although
molecular biology has advanced very rapidly it haswoh gone
quite as fast as that. We know something about the genetic
code, but we do not yet know it in full detail. So we have to
start off by asking a certain number of general questions of
which the first very pertinently is ~- what is the size of the
codon? Obviously it eannot be just a pair of bases. Ye have
four alternstives for each site, and this would give sixteen
combinations, which is not enough to code for twenty different
thines. In addition these must be a few symbols for spaces,
such as stop and start. So the smallest likely number, assuming
that it 1s an intever which is always the same, is three bases
per codon. This gives us sixty-four different possible triplets
and this, as you will see, we now think is the correct answer.

This leads us to a second question. 4111 there be more
than one codon standing for each amino-acid? You could imagine
that out of the sixty-four only twenty stood for amino-acids
and forty-four of the triplets stood for something else. ‘We do
not believe that this is so. The present evidenee suggests that,
in general, more than one triplet codes for a particular amino-
acid. The code, to use the technical term, is degenerate.

Now the four bises, with minor reservations, and the
twenty amino-acids, also with minor reservations, are the same
right throughout Nature. They are the same for animals, for

plants and for micro-organisms. For this reason we can attack

the problem of the genetic code (to start off with at any rate)
-9-

by studying which organism we please. Not unnaturally, we
choose the ones which we can handle most easily. It turns

out that the ones most used are, on the one hand, micro-
organisms, because one can grow them so rapidly, and their
nutritional requirements are easy; but also, I am glad to say,
human beings. This is because there are so many doctors look-
ing at so many human beings.

Let us assume for the moment that it is a triplet of
bases which makes a codon. There are two main possibilities
open to us. One is called an overlapping code, and the other
a non-overlapping code. The distinction is quite simple.

In the non-overlapping code the first amino-acid is coded by
the first three bases, the next by the next three, and s0 on.
In an overlapping code, on the other hand, the first amino-acid
is coded by base one, two and three but the second one by
numbers two, three and four, and third by numbers three, four
and five, ete. ‘“e are now reasonably certain that the non-
overlapping code is right. The reason is as follows: a genetic
mutation will in general change one base into another base. It
will alter the message at one particular point. vith a non-
overlapping code, this would mean that you would merely alter
one amino-acid. However, the typical change in an overlapping
code should be to several adjacent amino-acids. This is never
found. Our evidence comes mainly from the numerous abnormal
human haemoglobins and from the protein of Tobacco Mosaic Virus,
(Wittmann and Sittmann-Liebold, 1963; Tsugita, 1962).

‘e ask next: is the order of the codons along the nucleic
acid the same as the order of the corresponding amino actds in
the protein? In other words, are the two orders co-linear?
~xperimental evidence that this is so has been produced recently

by two groups of workers,
~10-

Yanofsky and his co-workers (1964) have studied one of the
proteins of the enzyme tryptophan synthetase of Escherichia
coli. They have determined part of the amino-acid sequence, by
the methods of protein chemistry. In addition they have picked
up a number of mutants, each of which, as I have already said,
nawe a change of one amino-acid. For example, one of them has
a cystine instead of a tyrosine at a certain point, while another
has an isoleucine irstead of a threonine, and so on. By genetic
methods (which have nothing to do with protein chemistry) they
have constructed a genetic map of these mutants, and put them in
a certain order by means of genetic crosses. The order of the
mutants on the genetic map is precisely the order of the corres-
ponding amino-acid changes in the protein. Moreover, where the
distance between two mutants is large, it is also large on the
genetic map; and conversely, when they are very close together
in the protein, they are also close together on the genetic map.

Secondly, some of my colleagues at Cambridge (Sarabhai,
Stretton, Brenner and Bolle, 1964) have come to a similar con-
clusion, again using a micro-organism, but this time the bacterio-
phage 74. They have studied the protein of the head of this
phage. They have picked up mutants of a particular type known
as amber mutants, and have proved that such mutants terminate
the polypeptide chain. You might wonder how can you work with a
phage if its head is incomplete. The answer is that there are
certain strains of E. coli (the host cell of the phage) which
suppress such mutants, so that one can use these strains to
grow a stock of the phage and then test how the mutant behaves
by growing it on the usual strain. By relatively simple methods,
using radioactive latelling, they have been able to estimate

On “Hoo
which of the different nolypeptide fragments is the longer. for

example, there is one particular peptide, containing cysteine,
-~ Jil -

which can be obtained by cutting up the protein with proto-
lytic enzymes. (There is only one cysteine residue in the
protein). One can ask, for cach one of the mutants, whether

it has got that cysteine peptide or not. The ones that have ub
will have lonser chains than the ones that lack it. This
meneral method can be applied to a11 the other pertides into
which the protein can be split. ‘The polypeptide chain that
gives the most fragments is clearly the longest and therefore
by lookine at the evidence in this way, without determining the
sequence precisely, they can determine the order of the mutants
purely by the methods of protein chemistry. They have also
obtained the order of the mutants by senetic methods, and, once
arain, the two orders are the same.

The next onestion is the followirs: since we know that we
are reading alons in croups of a certain number, which do not
overlap, how do we know how to cut the message up into the
richt sroups? For example, if you started reading in the middle
you misht begin out of phase- you might be one ahead or one
behind, and then you would read the whole of the message wrongly.
Tt turns out that we now have sood evidence that it is done in
the simplest way you could think of. The readin: berins at some
fixed point — you mivht say the besinning - and just goes on
reading from there in groups of three. I will not co into the
evidence in detail except to say that we have studied a deletion
that joins two feres torether and from this we can make a strong
case that the reading does start from a particular point, (Crick,
Barnett, Brenner and Watts-Tobin, 1961).

This has enabled us to tackle the question: is it really
a froup of three thet makes up a codon? The basic ides is the
followinr. We are able to rick up mutants which we believe
(from the way they behave in various contexts) are not merely

the change of one base into another, but are either the addition
- 12 ~

or a deletion of a base or bases. What happens when you have
a genetic message and you put in an extra base? The reading
starts from the beginning until it comes to that point and
from there onwards the whole of the message is read incorrectly,
because it is being read out of phase. In fact we find that
these mutants are completely inactive - this is one of our bits
of evidence that they are what we say they are. You can pick
up a number of such mutants and can put ther together, by genetic
methods, into the same gene. ‘For example you can put together
two of them. Such a gene would be read correctly until it
reached the first addition, and then it would be out of phase.
NYhen the reading came to the second addition it would read out
of phase again, and so the whole of the rest of the message
would be read incorrectly. Now it so happens that the left-hand
end of this gene is not terribly important for its function. ‘Ve
can actually delete it and the gene will work after a fashion.
In this region we have constructed, by senetic methods, a triple
mutant, using three mutants all of the same type, and we have
found that the gene will nevertheless function fairly normally.

This result is really very striking. Each of the three
different faults, used singly, will knock out the rene. You can
put them together in pairs in any combination you like, but then
the gene is still quite inactive. Put all three in the same gene
and the function comes back. We have been able to do this with
a number of distinct combinations of three mutants (Crick et al.,
1961).

The conclusion we draw from this is that there are indeed
three bases in a codon. I would emphasize that this is fenetic
evidence and it has to be confirmed by biochemistry before we

can be certain of it.
-~13-

Unfortunately I cannot go any further without saying
just a few words about protein synthesis. We have been talk-
ing up till now as if protein synthesis were a black box opera-
tion, but sooner or later it is obvious we have to come down to
biochemistry and ask what exactly is going on inside the cell.
It will be impossible for me in this lecture to make you under-
stand in any detail what our knowledge of protein synthesis
amounts to, because it is now quite considerable, (though still
a long way short of what we would like) but I will attempt to
describe it in outline. for details see the review by Watson
(1964).

The genetic material in most organisms is DNA. Proteins
are made on objects called ribosomes, which are small particles
(about the size of a polio virus) consisting of two unequal
parts. We believe that they are the reading mechanism. They
do not actually work on the DNA itself. A RNA copy is made of
one chain of the DNA. This copy is called messenger RNA (m-RNA).
The reading heads (the ribosomes) go onto the m-RNA and travel
along it, one after another, giving simultaneous readings (one
behind the other) of this working copy of the genetic message.
As they go along they construct the protein and when they come
to the end (assuming it is a simple message and only codes for
one protein) they come off, and release the protein, and then
go round again, not necessarily onto the same piece of messenger
RNA, The string of ribosomes attached to a piece of m-RNA is
known as a polyribosome or polysome for short. They can be seen
in the eleotron microscope.

How do the amino-acids, under the dictates of the m-RNA,
get into the right places? There is elaborate btochemtoal,
machinery to do this. Each amino-acid is carried in by, a special

species of molecule, known as soluble RNA (also called transfer RNA).
-~ 14 ~

There is a special enzyme for each amino-acid which joins it
specifically to its own soluble RNA and not to any of the
others. There are twenty different sorts of enzymes - one for
each amino-acid. However there are more than twenty kinds of
soluble RNA, and therefore there are several different sorts
for each amino-acid. ‘We think there is probably one for each
codon.

Yhen the amino-acid has been added to the polypeptide
chain the soluble RNA is then released, recharged with a new
copy of the amino-acid and used again. You can see that the
system is quite elaborate.

The important point I want to make is that a single-
stranded RNA can carry the instructions to make a protein. This
has been shown very convincingly in the case of RNA from a
smal] virus, by adding it to a cell-free system (Nathans et al.,
1962). This cell-free system is obtained by taking cells,
breaking them open, partially purifying the bits and then put—
ting them together again. To such a system you can add synthe-
tic nucleic acids and they, too, will act as messengers. The
recent work on the genetic code has gone very rapidly because of
this technique. The initial discovery was by Nirenberg and
Matthaei (1961) who added polyuridylic acid to such a system,
and found that it responded by making polyphenylalanine, showing
that one of the codons for phenylalanine consists of three
uracils. His group and Ochoa's sroup have done most of the
experimental work using this method. They have also shown that
polyadenylic acid will direct the synthesis of polylysine, imply-
ing that three adenines code for lysine. In the same way it is
likely that three cytosines code for proline, thu! las u lew Cer bom.

Tt is not too difficult, as it turns out, to make synthe-
tic messenger RNA of determined composition but with a random

sequence, because there is a special enzyme which will join the
~15-

precursors together. It is thus possible to synthesize an

RNA which has uracils and cytosines in an approximately random
order. This is found to icnorporate phenyalanine (UUU), pro-
line (CCT) and also leucine and serine. Thus some of the trip-
lets containing U's mixed with C's code for leucine and serine.
By these methods, without foing into details, they have been
able to obtain the composition (but not the order) of quite a
number of triplets. In certain cases more than one trinlet is
claimed for an amino-acid. We believe as a result of this, and
other evidence that it is very probable that more than one trip-
let codes for a particuler amino-acid, (see the review by Crick,
1953% and the arhtlts n CSHS, 1962),

Recently Leder and Nirenberg (1964) have found a way of
determining codons not by using long messages but just by adding
triplets (trinucleotides) to the system, by studying the binding
of s-RNA to ribosomes in the presence of a particular trinucleo-
tide. ‘They have been able to show that the triplet GUU (suanine,
uracil, uracil, in that order) codes for valine and that UGU and
UUG do not. Ineidertally this gives us direct biochemical evi-
dence that the codon is probably a triplet. There results will
have to be extended and confirmed, but it does look as if we
have a chance now of determining the order of all the trinlets
making up the code, though the results will need to be supple-
mented by other methods, of which mutarenesis is probably the
most important.

It is interesting that it is already possible to obtain
genetic recombination within a codon. For example, Yanofsky
(1953) has studied a particular protein, which has glutanic acid
at one place in the protein, in one mutant, and has argenine

(in the same place) in another mutant. [le has crossed these
- 16 —

together genetically, and has obtained glycine at that parti-

. (hth > lhe. dina cord a be cath hype
cular place, Genetic recombination his produced an amino-acid
which was not there before. This is not surprising if you think
of it in terms of the bases of the nucleic acid. The triplet

lke cath type

in one mutant has been altered in one place and in the other
mutant in a different place. The recombinant has produced the
original triplet again. This sort of result will erable us to
confirm the senetic code when we have a tentative version of it.

I have said that the four bases are universal (the same
throughout Nature), and that the twenty amino-acids are also
the same throughout Nature. One might guess from this that the
code (the relationship between them) might also be the same,
but this has to be proved. We now have preliminary evidence
about this. ‘Ye can make cell-free systems from different organ-
isms and add synthetic messengers to them and see if the same
polypeptide chains is made in each case. So far nobody has
shown a significant difference between one orsanism and another
(see,for example, Weinstein, 1963). Another technique is to
use mixed systems. It is possible to synthesize haemoglobin in
the test-tube, using a cell-free system from reticulocytes, In
addition it can be done using the polysomes from the reticulocytes,
and part of the soluble components - the s-RNA and the activating
enzymes ~ from ©. evli. In other words a mixed system from a
mammal and a micro-organism will synthesize a mammalian haemo-
globin (von Shrenstein and Lipmann, 1961).

The most dramatic case of a mixed system was discovered
recently by Abel and Trautner (1964), who succeeded in growing
vaccinia virus inside Bacillus subtilis. They infected the
bacterial cell with the DNA of the vaccinia virus, and the cell

became stuffed full of vaccinia. It is indeed a very remarkable

thing that you can take the genetic material of an animal virus
-17-

and use it to grow that virus inside a micro-organism. This
would again suggest that the genetic code of these two organisms
must be very similar.

In one's enthusiasm for the genetic code-one must realise
that, apart from such details as to how you start the polypep-
tide chain and how you finish it, there are very important prob-
lems which are not strictly touched on by the code, Genes have
to be turned on and off - they do not act all the time. This is
a problem which is not strictly part of the genetic code, althougn
it lies very close to it. We do not yet understand the control
mechanisms in bacteria. ‘We do not even know them in the case of
phage infection. In the case of phage T4 there are early pro-
teins which are made in the first ten minutes. A whole lot of
other proteins (the proteins which make up the virus) are made
later on. What controls the switching from early to late proteins
we do not yet know. Still less do we know about higher organisms,
or even if the control systems used in micro-organisms also exist
in man. It is very likely that there are in addition more ela-
borate and heirarchal controls which occur in higher organisms.

Leaving this aside and coming back to the code, what can
we say about it in summary? On the present evidence (which I
must emphasize is incomplete) it appears to be a non-overlapping
triplet code, with several triplets for each amino-acid. ‘The
code may not be identical in a1] living things but it is likely
to be fairly similar.

The final question I think you might ask me is, when are we
likely to know it in detail? This is always the most difficult
question to answer but I would be surprised if it had not‘ been
determined, at least for one organism, within the next couple of

years.