March 19, 1964

Dr. Emil L. Smith

Department of Biochemistry

University of California School of Medicine
Los Angeles, California

Dear Emil:

I would like to repeat how very much I enjoyed your lecture on the
evolution of the cytochromes a couple of weeks ago, and particularly
your thoughtful rebhttals to a couple of my spontaneous questions.
The conservation of the cytochrome sequence between yeast and man is
absolutely staggering and {t has helped me to look at the iswues of
protein evolution in a completely different light, putting most of
the stress on the functional requirements of the sequence,

I thought you might be interested {n some versions of the “ Torah"

a little more compact than the rolls you displayed at the mesting,
which I have found very helpful in looking at polypeptide sequences,
These are off some of the computer programs we have been playing with
lately in a so far not very fruitful analysis of common information
in amino acid sequences, The enclosures on the cytochromes show the
amino acids in Sorm’s code, and with the amino acids separated into I
hope not enttmsely arbitrary groups, the dissection helping one to see
how these are organized along a sequence,

The Xerox copies also illustrate another matching program wherein the
computer Looks for the best match of an indicated pair of sequences

with respect to the number of spaces that one sequence has to be shifted
relative to another, and showing how far away one has to look to find a
matching element in the opposite string. So, for example, the consistency
of matching elements on line 5 for human heart cytochrome versus baker's
yeast cytochrome is an indication of what you already know very well, the
very close correspondence between these two strings subject to an extra

set of five elements at the initial of the yeast. The number at line ~20
refers to the position that the computer found most expeditious to initiate
the search from at each level. The other similar plot of human heart versus
pseudowonas cytochrome also shows what you very well know, nawely the almost
complete lack of homology between these two sequences,and the variety of
numbers geen at ~20 shows how far the computer program had to wander in
finding the best available match over any appreciable interval, as poor

as that was. These programs are organized to allow matching to be done

under a variety of conditions of equivalenge; the present display shows he

the condition of identify, We have spent some time in looking at other

wep ritchie sie Wiad bo ey Ge Lia. [{

pena nearest MEET OANA NE

on

bn

ww

Mf

ALT!

 
Dr. Emil L. Smith March 19, 1964
Page 2

more relaxes conditions, for example, functional equivalent similar to @&
those shown on the dissected plots, as well as various types of lists of
permitted mutations on either a theoretical or an experimental basis, ;
However, I must say that these relaxed conditions of equivalenps have not C<
shown anything especially interesting except that the functional similarity
does bring out very well the string of nine elements that begins at pesition
80 in human heart cytochrome, MWFVYWIII, which ts a pretty good match to the
sequence NLFYYLII in the baker's yeast cytochrome, and which also has a
corresponding sequence of nine elements in myoglobin, It is also pretty
plain that the string of hydrophobic elements followed by the polar amino
acids is relatively relaxed with regard to which polar amino acid follows
precisely which, just so long as the total charge in that region roughly
balances out. At least, this ie the impression that I get comparing the

myoglobins and the cytochromes together, But I certainly would not suggest

that there is a necessary phylogenetic relationship between these elements
of structure since it is such an obvious building block for a segmant of protein
to perform God knows what function, I have to state too that in a test of
randomized sequences we did find a match just as good as this in juat one out
of twenty comparisons of the sane protein strings, so it is difficult to
attribute even a statistical significance to this, though I find it hard to
avoid that there is some sense locked up in this group of nine elements, Since

we have reached this impasse, it {s not plain that there is very much more
constructive that we can do with these amino acid comparison programs, but I

would be delighted to have any suggestions from you. As more and more inforna~
tion does come out on sequences of a variety of proteins, I am quite confident

that this approach will be more and more essential,

 

Cordially, and with the best of luck (perhaps a little belatedly now) for
your new post,

Joshua Lederberg
Professor of Genatics