Avery Gateway Dedication September 29, 1965

Dr. Robert W. Holley

Dr. Bronk, Dr. McCarty, ladies and gentlemen. It is a great honor to
participate in this program this afternoon. It seems to me that the
dedication of a memorial gate is especially fitting, and it's "particularly
appropriate symbol, a particularly appropriate way of conmemorating a man
whose work has opened up. such a vast world of new ideas and experimental
opportunities. It's difficult for me to think of any piece of work that
has been more important than that of Avery, MacLeod,and McCarty. There

are very few experiments that even come close to equalling the consequences
that this work has had in biochemistry, genetics,and biology in general.

I think that my view point is a little different than the earlier speakers
partly because I ama little younger and partly because of my ignorance as
an organic chemist for a while my knowledge of nucleic acids is post Avery,
MacLeod, and McCarty. enka at the time when I really became interested

in nucleic acids this was accepted, this was the truth, this was just back-
ground, something in history that everyone accepted. This makes me, as

Dr. McCarty has indicated, a member of the very large group of workers who
have come into the area of nucleic acids and related subjects ¢uetime, afters.
the period when the significance became clear. And I believe it is as a
representative of this group that I speak this afternoon. There isn't time
to consider all of the ways in which this work has been of tremendous
significance, pick just one area, that of the structure of nucleic acids,
and indicate the present status in the area with particular reference to
RNA, since that's the area in which the most of the work has gone recently.

I want to describe very briefly the({technical work /recenthy on the structure

of an RNA. The work I'm going to describe took a number of years, involved
Nor Sval
DE NAMES

Dr. Holley page 2

several people. Those particularly involved were Gene Apgar; B: P.' Doctor,
f araiee 0) —

George Everett, James Madison, Mark Marquise, susan fara, John Penswick and
D .
Odis_ and Mere. The nucleic acids that we have worked with on structure have
OO

rua TRANSPaR, RNAS,
been i , the relatively low molecular weight ribonucleic acids

that carry amime-eeidey-errrey specific amino acids to the site of protein

synthesis. As obtained from cells, one gets a mixture then of transfer

RNGSS carrying all of the different amino acids, that is,at least 20 different
RNAge. And the first problem, one that took us about half the time, actually,
was to separate the mixture obtained from the cells. And the first slide
indicates the results, using a technique, counter-current distribution, that
needs no description here at The Rockefeller Institute. This shows the
separation that is obtained in counter current distribution, starting with
bulk yeast transfer-RNA. The solid line indicates the spreading of the
material in the apparatus, the dash lines indicate positions of the different
individual ribonucleic acids. By picking the most active fractions, putting
them back in the apparatus, redistributing, it ts’ possible to obtain three
of these relatively homogenous. The next slide indicates the results with
the alanine transfer-RNA,using the counter-current distribution technique.
The solid line indicates RNA, the dash line indicates activity for alanine,
the dotted line is a calculated theoretical distribution curve. It is
material of this sort, then,that was used in structural studies. Although
this material had activity for no other amino acid,the possibility existed
that it was a mixture molecular species so it was a gamble to work on its
structure. We could be certain that if this was a mixture of molecular
species, that eventually we would find it was an insoluble problem. There

seemed to be no alternative but to try. By increasing the scale of procedures

we eventually obtained L@ts of milligrams rather than,as you see here,a peak

 
Dr. Holley page 3

of 2 milligrams;so we have tens of milligrams to work with. Even so, the
preparation was a continuing chore and we were always working with the
minimum amount of material. In determining the chemical structure of this
nucleic acid,we utilized about ] gram of purified RNA, obtained from about
200 grams of bulk yeast RNA, and that,in turn, from about 300 pounds of
commercial baker's yeast. The approach in determining the nucleotide set
quence was to use degradated enzymes, ribonucleases, to cleave the RNA chain.
The next slide summarizes the fragments obtained using pancreatic ribonuclease.
For the nonspecialists here,a nucleic acid can be considered as a message
written in English letters,if you please. The trans fer-RNage are the
smallest biologically active nucleic acids known; they contain rough# ly
80 units, i what we are trying to determine is a sentence of 80 letters.
the approach, then, is to break the sentence into words and determine the
nature of the words, hoping to eventually be able to put the words back
together into the sentence. These, then, would correspond to the words
Aimiaind
obtained by breaking the nucleic acid wherever peremidy nucleotides occur,
C+ vtddine, nicyweydiuradine, u-uridine, a-aden@ia and so forth. All
of these fragments or words were obtained in approximately 80 to 95%

yield,with the exception of two of them, actually one shown on here,

only about 50° of the theoretical amount.A e'll see this was obtained |
the methyl-i ps cudobase which 26 obtained i ya different word so we know it
ae

At,
actually exists. The approach was to identify each one of these fragments,
ant
these words ,¢determine a sequence of letters. Composition alone was
sufficient to determine sequence of the letters of the words containing
only two letters, because this enzyme cleaves next to the phramidine, the

Ce a
seed or the hue. In the case of a larger word three nucleotides, one

knows that pargmidine is at the right,as we're writing on here, the question

 
Dr- Holley page 4

then is whether this one, for example,is AGC or GAC. By complete degradation
with an enzyme called snake venom phosphodiesterase one obtains the letter
at the left end as a nucleoside, this one gave A,for example, and that gives
the sequence of a trinucleotide. You'll note a number of unusual letters,
complicated abbreviations methyLeommutte, aihyar See ino@ynic and so
forth, these are characteristic of the transfer ribonucleic acids. The
larger word, such as one composed of 4 letters, it's easy enough to get the
two ends; one has the question of the order of the letters in between. This
can be obtained, in this case anyway, by degradation with ribonuclease Ti
that cleaves next to G,so we had AC together we knew the sequence as GGAC.

A

One of these fragments is listed without a phosphate/ the p refers to phos-
~- hale, .

phate, cytadgne, that comes from the right end of the molecule that would

correspond to the period at the end of the sentence. As obtained froin

Apanyele
commerical baker's, yeast transfer- RNaWe mostly lack the terminal é

acid unityterminate “dacythdl Lic acid, that is the end of the molecule.

There is an extra p on this word; it's the only one that has a p on the left,
fhat would correspond to the capital letter at the beginning of the sentence,
that's the left end of the molecule. So we know this is at the left end of
the molecule. To determine the order of letters on anything as long as 8
nucleotides required a new method,which we developed as shown on the next
slide. What we did was treat with venom phosphodesterase to give a mixture,
in which successive letters had been removed from the right end of this
molecule. The presence of a phosphate on this end interferes with the enzyme
so that was removed first, by alkaline phosphptase. Now,this mixture can be

separated,as shown on the next slide. Where this is the starting material,

this is after removal of a phosphate and then removal of one, two, three,

 
wey

Holley - page 5

four, five nucleotide units. Now simply analysis for the terminal
group,which can be obtained by alkaline hydrolysis gives us UGAGA
corresponding to the sequence of those four, five groups. That gave us
a set of letters ,/set of words, corresponding to breaking sentence where-
ever a C or a U occu. Another enzyme was used ribonuclease TA, to break /A
the sentence wherever G occurs. These are shown on the next slide.
And you'll see each one of these ends in a G,with the exception of this
one piece. That,then,is the right end of the molecule. This is where
the period is again. And there is the left end as a guangg ine diphos-~
phate. Again, the identifies of all of these fragments were determined,
ab wee
nucleotide composition is analysis for the letters, with sufficient for
all of these. we had to determine the sequences of the others. The
methods I've described were sufficient for determining these words for
all except 3 of these. And a method indicated on the next slide was used
for those. The method I describe with partial degradation is not good
for the last 3 nucleotides. But nicrococcalmuclease %s shown by Dr. Las-
kowski , Sr-, to leave the last 2 nucleotides untouched, as a dinucleotide
is a combination of 2 letters. So that these remaining three sequences
were completed using this enzyme. That gave us two sets of words corres-
ponding to the sentence and now we wanted to put them together. Because
ofthe presonceof these minor nucleodtes occur only once in the molecule,
we had overlaps between the two sets of words, so we could put these to-
gether to some extent and this is shown on the next slide. Putting all
the information together, we could conclude there were 77 letters in the

sentence and they could be grouped to give us 14 sequences. Our problem
Holley - page 6

now was to arrange these 14, corresponding to the intact molecule. We
tried various approaches, the approach that worked was to get large
fragments that break the sentence into just a few pieces. The next slide
shows results obtained by John Penswick who did this. very brief treatment

Te pe :

of the RNA with ribonuclease TA 6* the presence of magnesia. The mixture
was chromatographed on aah reread celulose in the presence of 7

molar urea. This correspondes to the intact RNA untouched by the enzyme.
This is still active. These two peaks were inactive, and they look super-
ficially like the molecule has been cleaved in two. Each one of these
could be a mixture,however. This had to be determined. Each of these
peaks was purified and then analyzed,as shown on the next slide, BY com-
plete degradation to the words we'd already identified,using ribonuclease
Ti. This pattern corresponds to those words obtained with ribonuclease T1
where, egrhonmr3 with two or three exceptions, each one of these peaks
represents a different word. If the molecule is split into two pieces

then words should be found in one piece or the other and not in both, and
that is what is observed. One and two are dinucleotides, there in fraction 2.
Peak 3 is monoguand Lic acid that Wit? be expected to be in both parts.

Four is in 2, 5 and 1, 6 and 7 in 2, 8 and 2. Peak 9 is actually a mixture
of the two ends of the molecule, one end is found in 1], the other end in 2,
and so on. So we had the words separated into two groups, it's like ith
the-ercreward puzzle, with the jigzaw puzzle; if you could separate the

pieces that go together into two groups rather than having them all together.

LE ris

BD-- -*
With—ehis work to get two large pieces, it seemed reasonable it might work
GEr ‘ : oo :
to give several, as shown on the next slide,that was possible. This is again
the pattern on complete digestion to the words that 12 had on the slide

before. This is brief digestion with the enzyme; many of those peaks are
Holley - page 7

missing,and instead one has many large fragments containing combinations
of words. Each one of these now could be analyzed, put back through the
procedure to find which words were present in each. I'11 run through that
with just two of them, peak 16 and peak 18,and summarize the rest of the
results. Next slide shows; taking peak 16,retreating with the enzyme to
break to the words that we already knew, Me obtained peak 12 and peak 9,
traces of impurities. The identities are shown on the next slide. Mex
olide—pleasee Peak 9 is the end of the molecule, the one corresponding
to the period}; peak 12 can go then only to the left of it, and ye have two
of those words put together. The next slide. Peak 18 turned out to be a
mixture of things. They were separated by re-chromatography at 55° so we
had three large pieces to analyze, and these are shown on the following
slides. Next slide. 18A contained ] and 15. These are shown on the next
Slide. These two words - we knew that I was present in the sequence IGC,”~
pris fragment then could not go to the right here; there is no G to put in
between, it must be to the left and we know that sequence. The next slide.
18B had three components; these are shown on the next slide. We knew already
that the AG words were in this sequence, this word was not present in the
fragment, therefore that has to go to the left, and we know that sequence.
It's just putting pieces together one after another. The next slide.
Finally, peak 18C gave us these four pieces,as shown on the following slide.
We knew that end of the molecule that accounts for the PGP, the two chs; it
accounts for the only C present, Establishes that CG was there. We knew
that fmethy1-c was in this sequence, since there was no additional C,the UG
has to go to the left, we know that much. We have one G left. If that is

€
put there or there we create a GGU sequence, which was not present in the

molecule, that would show up in the pancreatic ribonuclease digest, therefore

 
Ws

Holley - page 8

the G has to go to the right of the methyl-G and we know that sequence.
The next slide summarizes all of the data. These are the four that I've
just indicated. We had additional large fragments corresponding to these
five sequences. Age the processes just like putting a jigsaw puzzle together;
put one more piece on, another two pieces on, eventually you have only a
piece or two left,and it is very easy to drop them into place. These
correspond then to the two halves that were obtained. We know the period
is over here, the capital letter over there, they can only go together in
one way to give us the nucleotide sequence of the alanine transfer RNA
isolated from yeast. Given the structure in these terms we're interested
in questions of biological interest. One is, where in this molecule is
the specificity for interaction with the messenger-RNA? it is the part
called the anticodon that corresponds to alanine. Our indications would
be it's not proved that it is the IGC sequence in the middle. One would

like to know what parts of the molecule interact with the ribosope when

 

T- pstvuoo-
the protein is formed, and we would suggest this region with the Feudere~
vainyeie, Gyridyeie DinyoRar- us C2)
é peychodyllic and the region with the dyhydrouse may well interact

in that way. One would like to know what this looks like in space, it isn't
just stretched out as a line in solution, how is it twisted around? “LE one
looks at sequences, you find that there are vers, only short regions that are
complementary in the Wates: and Crick sense. It is not possible to make long
double-stranded regions; various ways one can have short regions are shown

in the next slide. Where, attempt is made to line up one long double-stranded
region, one with two double-stranded regions, the one with three looked very
poor, the one with four looks quite good. My personal preference at this

point is something like this. Through the cooperation of Robert Harte of

the Americal Society of Biological Cuemists, we've gotten enough of the
Holley ~ page 9

Society's models, the Adams models, that they are producing,to build a
model of this structure. I'd hoped to be able to have a slide for it but
time ran out. What is clear from the model is that when you build this
structure you have something that is still very flexible. The different
parts here can be twisted around in many ways. The way that appeals to us
is to twist this part back over that one, it's sort of like shaking hands,
that two parts fit together, that would give a region for interaction with
the ribosofe. This would be the anticodon for interaction with the
messenger~RNA, and this is where the amino acid ie eee. This is a very
brief review of present status of structure of ribonucleic acids, the

&
reason for working with transfer-RNage is because they are relatively
small, 77 nucleotides, viral nage are much larger, and it's of course
a long way to a transforming DNA. However, considering the rate of
progress in the field of structure of nucleic acids, many laboratories
all over the world making rapid progress, I think it's a safe prediction
that in perhaps 10 years we will be able to write out the complete

chemical structure of something with transforming activity.

Thank you.