# Natural Selection and the Concept of a Protein Space

> An annotated edition of this paper on Fermat's Library, where scientists and readers explain its ideas, equations and history in the margins. Annotated edition: https://fermatslibrary.com/s/natural-selection-and-the-concept-of-a-protein-space

- **Authors:** John Maynard Smith
- **Published:** 1970
- **Topic:** Biology
- **Original:** https://s3-us-west-1.amazonaws.com/fermatslibrary/uploads%2F1686340589248-rpjfyo66nj-9c82499895e3c983bf97896dd5cb21bb%2F225563a0.pdf
- **Pages:** 2
- **Annotations:** 5 comments in 5 threads
- **Featured:** Fermat's Library Journal Club
- **Annotated edition:** https://fermatslibrary.com/s/natural-selection-and-the-concept-of-a-protein-space

## Annotations

### Annotation 1 (page 1)

Beside the passage: “Natural Selection and the Concept of a Protein Space”

**Annotation** by Tymor Hamamsy (2023-07-02):

> > "Natural selection and the concept of a protein space" has been described as among the most influential analogies ever proposed in evolutionary genetics.
>
> Source: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7153927/

### Annotation 2 (page 1)

Beside the passage: “SALISBURY 1 has argued that there is an apparent contradict ion between two fundamental concepts of biology - the belief that the gene is a unique sequence of nucleotid es whose fun ct ion it is to det e rmine the sequence of aminoacids in a protein, an d the theory of ev olution by natural selection. In brief, ho …”

**Annotation** by Tymor Hamamsy (2023-07-02):

> > Salisbury wrote in 1969 in "Natural Selection and the Complexity of the Gene", "Modern biology is faced with two ideas which seem to me to be quite incompatible with each other. One is the concept of evolution by natural selection of adaptive genes that are originally produced by random mutations. The other is the concept of the gene as part of a molecule of DNA, each gene being unique in the order of arrangement of its nucleotides. If life really depends on each gene being as unique as it appears to be, then it is too unique to come into being by chance mutations. There will be nothing for natural selection to act on. "

### Annotation 3 (page 1)

Beside the passage: “SALISBURY 1 has argued that there is an apparent contradict ion between two fundamental concepts of biology - the belief that the gene is a unique sequence of nucleotid es whose fun ct ion it is to det e rmine the sequence of aminoacids in a protein, an d the theory of ev olution by natural selection. In brief, ho …”

**Annotation** by Tymor Hamamsy (2023-07-02):

> While it is difficult to estimate the overall potential protein-space, the amino acid vocabulary consists of 20 unique amino acids, and the longest known protein is Titin, which can be up to 35,000 amino acids long. While one can't extrapolate that the potential protein search space is therefore 20^35K or an even bigger number, it clearly is enormous...
>
> Titin is a protein that is encoded by the TTN gene in humans, and has a length that ranges from ~27,000 to ~35,000  amino acids depending on the splice isoform. It is the third most abundant protein in muscle.
> Titin background: https://en.wikipedia.org/wiki/Titin

### Annotation 4 (page 1)

Beside the passage: “SALISBURY 1 has argued that there is an apparent contradict ion between two fundamental concepts of biology - the belief that the gene is a unique sequence of nucleotid es whose fun ct ion it is to det e rmine the sequence of aminoacids in a protein, an d the theory of ev olution by natural selection. In brief, ho …”

**Annotation** by Tymor Hamamsy (2023-07-03):

> Here is a great article that articulates the significance of 'Natural Selection and the Concept of a Protein Space' on its 50th anniversary: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7153927/

### Annotation 5 (page 2)

Beside the passage: “JOHX MAYNARD SMfflI School of Biological Sciences, Universit,y of Sussex .”

**Annotation** by Tymor Hamamsy (2023-07-03):

> John Maynard Smith (1920-2004) was a British mathematical evolutionary biologist and geneticist.
>
> Source: https://en.wikipedia.org/wiki/John_Maynard_Smith

## Full text

Extracted from the PDF. [Annotation N] marks where each annotation above sits.

#### Page 1

© 1970 Nature Publishing Group

NATURE VOL 225 FEBRUARY 7 1970

••

Fig. 2. Cerebe llar cortex of the same rat as Fig. 1. Unsta in ed autoradiogram of a cryostat sectio n. M, Molecular layer; P, Purkinjc cells (nonreacting); G, granular fayer; \V, white n1atter.

Fig. 3. Lumbar spina l cord (a nt erior horn) of a rat, 45 min after a si ngle injection of TSC-"C. Rouin-fi xed and embedded specimen, counterstained with haematoxylin and eosin. i\fotoneuroncs (l\I) are non-rea cting; capsula r glial cells (arrow) and astrocytes in the grey matter (arrow with a circle) contain numerous grains.

binding by non-reacting structures might indicate thos e ce llular elements that do not exert GAD activity. Such an approac h is actually an adaptation of the principl e described by Ostrovsky and Barnard 5 • Rats weighing 160 g were inj ected intraperitoncally with 5-20 mg / kg thiosemicarbazide- 14 C (specific activity 10 mCi/mmole, obtained from the Central Isotope Institute , Budapest). Characteristic convulsions and lethal "jumps" ap peared 45-120 min after inje c tion. The animals were killed by decapitation and samples of tho central nervou s system (brain, medulla, cerebellum, spinal cord) as well as o th er tissues were eith er fixed on Bouin's solution and prepared for autoradiography in the usual way or were not fixed but frozen with dry ice, cut on a cryostat and applied to slides pre-coated with Kodak AR-10 stripping film, emu lsion side up. Th e exposure time was 3-8 weeks; autoradiograms were develop e d in Kodak autoradi og raphic d eve loper and some were countcrstained with ha cmatoxy lin and cosin. Fig. I s hows a coronal section of the brain of a rat given two intraperitoneal injection s of 10 mg/kg TSC- 14 C , with an interval of 40 min betw een them, a nd then killed 90 min after the first injection. The h eaviest r eactio n is confined to the hippocampus and fascia dentata chiefly in the layer of hippocampal pyramids (Fig. 1, in set); silver grains are, however, not found in the nerve cells themselves but are concentrated in the surrounding neuropil. Activity is high

563

also in the nucleus hab enu l aris medialis. There is a moderate reaction in the cortex, especially in the vicinity of the interhemispherical fissure . Both the molecular layer and the granular layer of the ce r ebe llum contain numerous silver grains (Fig. 2). The localization pattern of the reaction does not, however, conform to any of the neural e lem en ts but resembles mor e closely the glial structure. The reaction is slightly stronger around Purkinje cells, which are themselves devoid of any r eactio n both in pre-fixed and in non-fixed specimens. On the other hand, in deep cerebellar nuclei (and also in some of the brain stem nuclei, for example the substantia nigra and nucl ei ponti s) silver grains are located within the cyto plasms of nerve cells. Weak or virtually no activity can be seen in the white matter, although the glial cells in the white matter react if high er doses ofTSC are used or if the animals are killed after a shorter time interval. The nerve cells in the spinal cord do not contain silver grains, but capsular and other glial cells in both the grey and white matter contain numerous grains (Fig. 3). According to Curtis 2 , GABA is not involved in spinal inhibitory mechanisms. The gross distribution of s ilv er grains reduced by TSC- 14C is in accord with the distribution of GAD and GABA anticipated on the basis of earlier biochemical and pharmacological studies (cerebellar cortex 2 , hippocampus•, substantia nigra 6 ). Wl>jle the concentration of silver grains in n erve cells (dentate nucl e u s, substantia nigra) and in the neuropil surrounding them (hippocampus) is consistent with current views, it is striking that Purkinje cells do not exer t any reaction. Proteins sensitive to TSC are undoubted ly more widespread than GAD, so the apparent localization of TSC in glial cells of th e cere be liar cortex may be partly due to binding of the drug to other B 6 -dependent enzymes. Th e possibility cannot be excluded, however, that, at l eas t in this area, GABA is produced by glial cells, perhaps in order to be taken up by nerv e terminals and/or nerv e cells in a second step.

D epa rtment of Anatomy, University Medical School, Szeged, Hungary.

BERTALAN CsrLLIK ELIZABETH KNYIHAR

Received August 12; revised Octob er 6, 1969.

1 Roberts, E., in Structure and Function of Inhibitory Mechanisms, 401 (Pergamon Press, Oxford, 1968). 'C urtis , D. R., in Structure and Function of I n hibitory Mechanisms, 420 (Pergamon Press, Oxford, 1968). 'Van Gelder, N. M., J . Neurochem., 12 , 231, 239 (1965). • Killam, K. F., Dasgupta, S. R., and Killam, E. K., in Inhibiti on in the Nervous System and Gam,na-Aminobutyric Acid, 302 (Pergamon Pre ss, Oxford, 1960). ' Ostrovsky, K., an d Barnard, E. A., Exp. Cell Res., 25, 456 (1961). • Albers. R. W., in Inhibitio" in the Nervous Sys t em and Gamma-Aminobutyric Acid, 196 (P erga mon Press, Oxford, 1960).

Natural Selection and the [Annotation 1] Concept of a Protein Space

SALISBURY 1 has argued that there is an apparent con- [Annotation 2] tradict ion between two fundamental concepts of biology - the belief that the gene is a unique sequence of nucleotid es whose fun ct ion it is to det e rmine the sequence of aminoacids in a protein, an d the theory of ev olution by natural selection. In brief, ho calculated that the numb er of po ssible amino-acid sequences is greater by many orders of m agnitude than the number of proteins which could have [Annotation 3] existed on Earth since the origin of life, and hence that functionally effective proteins have a vanishingly small chance [Annotation 4] of arising by mutation. Natural selection is

#### Page 2

© 1970 Nature Publishing Group

564

therefore ineffective because it lacks the essential raw material--favourable mutations. I should like to look at the problem from a different point of view. I shall assume that mutations, while not random in a chemical sense, are random as far as their chances of improving the function of tho corresponding proteins are concerned. I shall also assume that evolution has occurred either by the natural selection of favourable mutations or by the chance fixation by genetic drift of selectively neutral mutations. Tho justification for making these assumptions is that no sensible alternatives have been suggested and that no evidence exists at the moment to invalidate them. If these assumptions are true, what can we say about tho frequency and distribution of aminoacid sequences which are functional, either as enzymes or in some other way ? The model of protein evolution I want to discuss is best understood by analogy with a popular word game. The object of the game is to pass from one word to another of the same length by changing one letter at a time, with the requirement that all the intermediate words are meaningful in the same language. Thus WORD can be converted into GENE in the minimum number of steps, as follows:

WORD WORE GORE GONE GENE

This is an analogue of evolution, in which the words represent proteins; the letters represent amino-acids; the alteration of a single letter corresponds to the simplest evolutionary stop, the substitution of one amino-acid for another; and the requirement of meaning corresponds to the requirement that each unit step in evolution should be from one functional protein to another. The reason for the last requirement is as follows: suppose that a protein A B C D ... exists, and that a protein a b C D ... would be favoured by selection if it arose. Suppose further that the intermediates a B C D ... and A b C D ... are non-functional. These forms would arise by mutation, but would usually be eliminated by selection before a second mutation could occur. Tho double step from a b C D ... to A B C D would thus be very unlikely to occur. Such double steps with unfavourable intermediates may occasionally occur, but are probably too rare to be important in evolution. This is a model of the way in which one gene may change into another. An increase in the number of different genes in a single organism presumably occurs by the duplication of an already existing gene foJiowed by divergence. If so, it remains true that new genes arise as modifications of pre-existing ones. It follows that if evolution by natural selection is to occur, functional proteins must form a continuous network which can be traversed by unit mutational steps without passing through nonfunctional intermediates. In this respect, functional proteins resemble four-letter words in the English language, rather than eight-lotter words, for the latter form a series of small isolated islands in a sea of nonsense sequences. Of course, this is not to deny the existence of isolated island proteins, analogous to tho four-letter words ALSO and ALTO. It is easy to state the condition which must be satisfied if meaningful proteins are to form a network. Let X be a meaningful protein. Let N be the number of proteins which can be derived from X by a unit mutational step, and f the fraction of these which are meaningful, in tho sense of being as good as or better than X in some environment. Then, if JN> 1, meaningful proteins will form a network, and evolution by natural selection is possible. In estimating Nit is necessary to distinguish two classes of .mutations: (i) substitutions of single amino-acids, and additions or deletions of small numbers of amino-acidR, making only a small change to tho protein; and (ii) mutations producing a major change in amino-acid sequence, such as frame shifts and intramolecular inversions.

NATURE VOL. 225 FEBRUARY 7 1970

Mutations of the former type are much more likely to give rise to meaningful proteins than the latter. In the same way, a single random letter substitution in a meaningful word is more likely to give rise to a meaningful word than the simultaneous alteration of all the letters. Although frame shift mutations are known to occur, it is not clear whether they have ever been incorporated in evolution. It is therefore better to take N as the number of possible substitutions of single amino-acids. If all substitutions were possible in a single mutational stop, N for a protein of 100 amino-acids would be 1,900. In practice the genetic code limits X to approximately 10 3 • Hence] must be greater than 1/1,000. It does not follow that the fraction of all possible sequences which are meaningful need be as high as 1/1,000. It is probably much lower. There is almost certainly a higher probability that a sequence will be meaningful if it is a neighbour of an existing functional protein than if it is selected at random. In fact, in treating X as the number of aminoacid substitutions rather than as the total number of possible mutational steps, it was in effect assumed that a random sequence has a negligibly small probability of being functional; this assumption will be confirmed if it turns out that frame shifts are rarely or never incorporated in evolution. Suppose now that we imagine all possible amino-acid sequences to be arranged in a "protein space", so that two sequences are neighbours if one can be converted into another by a single amino-acid substitution. Then the requirement that JN > I requires that the "density'' of functional proteins in certain n,gions of the space must be quite high-perhaps greater than 1/1,000. This agrees with Salisbury's conclusion that proteins, and hence the genes that determine thorn, cannot be as unique as aJI that. As a convinced Darwinist, I published• the conclusion that JN > I when little was known about the frequency of amino-acid substitutions in evolution. Since then evidence has accumulated (for a review, see King and Jukes 3 ) that many substitutions are either selectively neutral or at least make comparatively minor changes in the function of proteins. If JN > I, no quantitative difficulty arises in explaining the evolution of proteins by natural selection. A difficulty nevertheless remains in explaining the origin of life-that is, in explaining the origin of the first functional proteins together with the genetic mechanism for producing them. If it were true that only a minute fraction of possible amino-acid sequences have eYen the slightest enzymatic activity, it would be difficult to understand how the first proteins arose. I do not want to discuss the problem of the origin of life, but only to point out that it is a quite different problem from that of the mechanism of evolution. Some questions about molecular evolution can be formulated more clearly in terms of a protein space. For example: (i) Are all existing proteins part of the same continuous network, and if so, have they all been reached from a single starting point ? Possible alternatives are that there are two or more distinct networks, or that there is one network with multiple starting points. (ii) How often, if ever, has evolution passed through a non-functional sequence ? If so, has this been achieved by the random walk of genes rendered redundant by duplication, or by the chance concurrence of two or more mutations? (iii) What fraction of the functional network has already been explored in evolution? (iv) vVhat fraction of potentially useful proteins are inaccessible ?

JOHX MAYNARD SMfflI School [Annotation 5] of Biological Sciences, Universit,y of Sussex .

Received November 7, 1969. 1 Salisbury, F. n., Nature,224, 342 (1060). 'Maynard Smith, J., in Tlte Scientist Speculates (edit. by Good, I. J.) (Heinemann, London, 1961). 3 King, J. L., and ,Tnkes, T. H., Science, 164, 788 (1060).

---

Source: [Fermat's Library](https://fermatslibrary.com/) — https://fermatslibrary.com/s/natural-selection-and-the-concept-of-a-protein-space. More annotated papers: https://fermatslibrary.com/llms.txt
