genetical implications of the structure of deoxyribonucleic acid watson and crick nature 1953

7/30/2019 Genetical Implications of the Structure of Deoxyribonucleic Acid Watson and Crick Nature 1953

1/4

N A T

for South Africa has expanded enormously

ed a t all with the passing of years : "West," he said, "primarily to create and foster a

al work. TVe exist to provide a commonthe scientific angle." He defended the

e should pride ourselves on being parochial. I wouldaimyy.Having thus firmly and, most people would agree,ly placed the Association in its proper perspective,

South African Journal of Science shoulda series of semi-popular articles reviewing andwork and those who pay for it. This, heht, is the proper filnction of the Joz~rna l, ndbut one aspect of the Association's du ty, asresentative of all sections of scientific opinion in

Besides his plea that the Assooiation needs to form

n the creation of better facilities for advancedin South Africa. On this last-named point,

hed to and forming part of those univer-

NERAL educatiorl to-day should be plannedso as to enable the ordinary citizen to adaptThis was the t,heme of an international

Broadly, the Conference found that organizedf a country to prepare children for the type

at a substantial development of technical education*Education in a Technological Society : a Prelimiilary Inter-th e Xature and Efficacy of Technical Education.Pp. 76. (Paris : Unesco ; London :1952.) 200 francs; I s . ; 75 cents.

May 30, 1953 vOL. 71is required a t all levels : at present it is whollyinadequate for future needs, while the practicalcontent of general education is also inadequate forthe needs of future citizens of a t,echnological societ,y.The cultural cont,ent of technical education is alsogenerally inadequate ; technical education requiresspecial consideration, and training for adaptability isan outstanding requirement in an age of ultra-rapidtechnological change. The education of women andgirls also demands particular attention in view oftheir dual role as workers and home-makers, andimproved administratjive arrangements are essentialif education is t o fulfil its true funetaion n such asocietv.The report does not suggest t,hat all these pro-positions apply equally to every country, though theConference considered that, so far as its knowledgeextended, they are generally valid for the world asa whole. The stress is laid on the need for adaptingtechnology to man, not man to technology. Thequestions formulated in this report-and which meritattention in current discussions on the expansion ofboth technical and technological education in GreatBritain-are raised in the belief th at mastery of themachine by man is not an end in itself : it is a meansto the development of man and of the whole society .The distinction between technician and techno-logist is not always kept clear in this report, par-ticularly in the chapter on the content of technicaleducation. Nevertheless, the report directs atten tionto some fundamental issues which no sound policyfor either type of education can disregard. In bothfields it must be recognized that we are concernednot simply with the efficiency of production, but alsowith the fundamental attitude which the men andwomen of to-morrow will adopt m facing the problemsof a technological society. Both, too, in seeking tofoster flexibility, must recognize that flexibility isdetermined not only by education and training butalso by social, economic and technical conditions ;and the administrative measures required to ensurethat education becomes more adapted to the needsof a changing technological society are themselveslikely to be most effective when they are informaland varied rather than concentrated and uniform.The administrator, no less than the teacher andstudent, has need of frequent opportunities of contactwith the industrial world, and requires experience ofthe difficulties and problems created by technologicaldevelopment in society; just as the teacher andstudent should keep abreast of developments inresearch and of practical applications in industry.

GENETICAL IMPLICATIONS OFTHE STRUCTURE OFDEOXYRIBONUCLEIC ACIDBy J . D. WATSON and F. H. C. CRICK

Medical Research Council Unit for the Study of theMolecular Structure of Biological Systems, CavendishLaboratory, Cam bridgeimportance of deoxyribonucleic acid (DNA)within living cells is undisputed. I t is found inHEall dividing cells, largely if not ent,irely n the nucleus,where it is an essential constituent of the chromo-somes. Many lines of evidence indicate th at it is thecarrier of a part of (if not all) the genetic specificityof the chromosomes and t,hus of th e gene itself.


2/4

N ~ .36 1 May 30, 1953 N A T U R E0.N.A .

BASE- SUCARi\BASE- SUCAR

\B A S E - SUCAR

P H O S P H A T E

\RASE-SUGAR \ PHOSPHATEBASE -SUGAR /

\ PHOSPHATE/

Fig. 1. Chemical foruiiula of a Pig . 2 . This figure is pure ly/ .slngle chain of cleoxyribo- diagrammatic. The two ribbonsnucleic acid symbolize th e two phosphate-sugar chains, and the hori-zontal rods the pairs of basesholding the chains tosether.The vertical line marks th efibre axisIJnt il now, however, no evidence has been presentedto show how i t , might carry out the essentialoperation required of a genetic material, that ofexact self-duplication.We have recently proposed a structure1 for thesialt of deoxyribonucleic acid which, if correct,immediately suggests a mechanism for its self-cluplication. X-ray evidence obtained by the workersa t King's College, London2, and presented a t thesame time, gives qualitative support t o our struc tureand is incompatible with all previously proposedatructures3.' Though the structure will not be com-pletely proved until a more extensive comparison hasbeen made w ith the X-ray da ta, we now feel sufficientconfidence in its general correctness to discuss itsgenetical implications. In doing so we are assumingthat fibres of the salt of deoxyribonucleic acid arenot artefacts arising in the method of preparation,since it has been shown by Wilkins and his co-workersthat similar X-ray patterns are obtained from boththe isolated fibres and certain intact biologicalmaterials such as sperm head and bacteriophagepart icles2s4.The chemical formula of deoxyribonucleic acid isnow well established. The molecule is a very longchain, the backbone of which consists of a regularalternation of sugar and phosphate groups, as shownin Fig. 1 . To each sugar is attached a nitrogenousbase, which can be of four different types. (We haveconsidered 5-methyl cytosine to be equivalent tocytosine, since either can fit equally well into ourstructure.) Two of the possible bases-adenine andguanine-are purines, and the other two-thymineand cytosine-are pyrimidines. So far as is known,sequence of bases along th e chain is irregular.The monomer unit, consisting of phosphate, sugarand base, is known as a nucleotide.The first feature of our structure which is ofbiological interest is that it consists not of one chain,but of two. These two chains are both coiled around

a common fibre axis, as is shown diagrammaticallyin Fig. 2. It has ofton been assumed that since therem7as only one chain in the chemical formula therewould only be one in the s truc tura l unit. However,the density, taken with th e X-ray evidencez, suggestsvery strongly that there are two.The other biologically important feature is themanner in which the two chains are held together.This is done by hydrogen bonds between the bases,as shown schematically in Fig. 3. The bases arejoined together in pairs, a single base from one chainbeing hydrogen-bonded to a single base from the 'other. The important point is tha t only certain pairsof bases will fit into the structure. One member of apair must be a purine and the other a pyrimidine inorder to bridge between the two chains. I f a pairconsisted of two purines, for example, there wouldnot be room for it.We believe that the bases will be present almostentirely in their most probable tautomeric forms. Ifthis is true, the conditions for forming hydrogenbonds are more restrictive, and the only pairs ofbases possible are :

adenine with tJEiymino;guanine tvitli cyt,osine.Tlle way in which these are joined together is shownin Figs. 4 an d 5. A given pair can be either wayround. Adenine, for example, can occur on eitherchain ; but when it does, its partner on the otherchain must always be thymine.This pairing is strongly supported by the recentanalytical resultsb, which show that for all sourcesof deoxyribonucleic acid examined the amount ofadenine is close to the amount of thymine, and theamount of guanine close to the amount of cytosine,although the cross-ratio (the ratio of adenine toguanine) can vary from one source to ahother.Indeed, if the sequence of bases on one chain isirregular, it is difficult to explain these analyticalresults except by the sort of pairing we havesugq3sted.The phosphate-sugar backbone of our model iscompletely regular, but any sequence of the pairs ofbases can fit into the structure. It follows that in along molecule many different permutations arepossible, and it therefore seems likely that the precisesequence of the bases is the code which carries thegenetical information. If the actual order of the

. *PHOSPHATE

/\ UGAR-BASE

PHOSPHATE/\

PHOSPHATE/ u c A R - B A s E\/

PHOSPHATE\/ u G A R - a r s r

/'. - - . - - - . BASE-SUGAR \PHOSPHATE

- - . - -- - - - . A S E - UGAR/\ PHOSPtihTI

. . - . .---A S E - S U G A R /\

PHOSPHAT E/

. 7 .I'tg. 3. Chemical formula of a .pai r of dcoxyribonucleic .acidchains. The hydrogen bonding 1s synibollzed by dotted lines


3/4

N A T U R EADENINE THY MINE

Fig. 4 . Pairing of adenine and thymine. Hydrogen bonds areshown dotted. One carbon atom of each sugar is shownGUANlM CYTOSINE

Fig. 5. Pairing of ellanine and cytosine. Hydrogen bonds %reshown dotted. One carbon atom of each sugar is shownI

bases on one of the pair of chains were given, onecould write down the exact order of the bases on theother one, because of the specific pairing. Thus onechain is, as it were, the conlplement of the other,and it is this feature which suggests how the deoxy-ribonucleic acid molecule might duplicate itself.Previous discussions of self-duplication have usuallyinvolved the concept of a template, or mould. Eitherthe template was supposed to copy itself directly orit was to produce a 'negative', which in its turn wasto ac t as a template and produce the original 'positive'once again. I n no case has it been explained indetail how it would do this in terms of atoms andmolecules.hTow our model for deoxyribonucleic acid is, ineffect, a pair of templates, each of which is com-plementary to the other. We imagine th at prior toduplication the hydrogen bonds are broken, and thetwo chains unwind and separate. Each chain thenacts as a template for the formation on to itself of anew companion chain, so that eventually we shallhave two pairs of chains, where we only had onebefore. Xoreover, the sequence of the pairs of baseswill have been duplicated exactly.A study of our model suggests th at this duplicationcould be done most simply if the single chain (or therelevant portion of i t ) takes up th e helical con-figuration. We imagine that at this stage in the lifeof the cell, free nuo2eotides, strictly polynucleotideprecursors, are available in quanti ty. From time totime the base of a free nucleotide will join up by

hydrogen bonds to one of the bases on the chainalready formed. We now postulate tha t the polymer-ization of these monomers to form a new chain isonly possible if the resulting chain can form theproposed structure. This is plausible, because stericreasons would not allow nucleotides 'crystallized' onto the first chain to approach one another in such away that they could be joined together into a nemchain, unless they were those nucleotides whichwere necessary to form our structure . Whether aspecial enzyme is required to carry out the pol~mer-ization, or whether t,he single helical chain alreadyformed acts effectively as an enzyme, remains to beseen.Since the two chains in our model are intertwined,it is essential for them to untwist if they are toseparate. As they make one complete turn aroundeach other in 34 A., there will be about 150 turnsper million molecular weight, so that whatever theprecise structure of the chromosome a considerableamount of uncoiling would be necessary. I t is wellknown from microscopic observation that muchcoiling and uncoiling occurs during mitosis, andthough this is on a much larger scale it probablyreflects similar processes on a molecular level.Although it is difficult at the moment to see howthese processes occur without everything gettingtangled, we do not feel that this objection will beinsuperable.Our structure, as described1, is an open one. Thereis.room between the pair of polynucleotide chains(see Fig. 2) for a polypeptide chain to wind aroundthe same helical axis. I t may be significant th at t hedistance between adjacent phosphorus atoms, 7 . 1 A.,is close to the repeat of a fully extended polypeptidechain. We think it probable th at in the sperm head,and in artificial nucleoproteins, the polypeptide chainoccupies this position . The relative weakness of th esecond layer-line in the published X-ray pictures3apiis crudely compatible with such an idea. The functionof the protein might well be to control the coilingand uncoiling, to assist in holding a single poly-nucleotide chain in a helical configuration, or somoother non-specific function.Our model suggests possible explanations for anumber of other phenomena. For example, spon-taneous mutation may be due to a base occasionallyoccurring in one of its less likely tautomeric forms.Again, the pairing between homologous chromosomesat meiosis may depend on pairing between specificbases. We shall discl~ss hese ideas in detail else-where.For the moment, the general scheme we haveproposed for the reproduction of deoxyribonucleicacid must be regarded as speculative. Even if it iscorrect, it is clear from what we have said that muchremains to be discos~ercd efore the picture of geneticduplication can be described in detail. What are thepolynucleotide precursors ? What makes the pair ofchains unwind and separate ? What is the preciserole of the protein ? Is t he chromosome one long pairof deoxyribonucleic acid chains, or does it consist ofpatches of the acid joined together by protein ?Despite these uncertainties feel that our pro-posed structure for deoxyribonucleic acid may helpto solve one of th e fundamental biological problems-the molecular basis of the template needed for geneticreplication. The hypothesis we are suggesting is th atthe template is the pattern of bases formed by onechain of th e deoxyribonucleic acid and th at the geneco~ta inse complementary pair of such templates.


4/4

May 30, 1953 N A TOne of US (J .D. IV.) has been aided by a fellowshipfrom the National Foundation for Infant,ile Paralysis(U.S.A.).

\lTatson, J. D., and Crick, F. H. C. , Xature, 171, 737 (1953).a \FTjlkins,31. H . F., Stokes, A . R., and Wilson, H. R., iVatz~re, 71,738 (1958). Franklin, R. E., and Gosl~ng,li. G., JTature,171,740 (1953).(a )Astbury, 77'.T., Symp.No. 1 Soc. Exp. Biol., 66 (1967). ( b )Furberg .S., Acta Ch e ~ n . cand., 8, 634 ( 1 9 5 2 ) . ( c ) Pauling, I,., and Carey,R. B., Nature, 171, 346 (1963) ; Proc. U.S. Bat. Acad. S ICL . , 39 ,84 (1953). ( d ) Fraser, R. D. B. (in preparation).\J7ilkins,AT. H. I?., and Randall, J. T., Biochim. et Biophys. Acta, 10,192 (1953).

( l~argaff ,E., for references see Zamenllof, S . , Brawerntitn, G. , an dUhargaff, E., Biochim. et B i o j h ~ s . d a, 9, 402 (1952). I\ att,G. R., J . Gen. Physzol., 36 , 201 (1952).

GEOPHYSICAL ANDMETEOROLOGICAL CHANGES INTHE PERIOD JANUARY-APRIL 1949

Ia recent article1 Lewis and McIntosh haveconsidered the geophysical data for the period

January-April 1949, which we presented in anearlier communication2. On the basis of certainprobability criteria they appear to show that theapparent, regular variations in ionospheric andmeteorological phenomena. which occurred in thatperiod were not significant. We have studied theirarticle and made a separate statist ical analysis of theunsmoothed data, and conclude that in all respectsour original suggestions seem to be valid.I n our original article we presented graphs showingfive-day moving averages in four parameters :(a )ground pressure, p ; ( b )E-layer critical frequency,f E ; ( c ) F-layer crit'ical frequency, fF 2 ; and (d)K-index of geomagnetic activity. The connexion be-tween ionospheric and geomagnetic phenomena iswell known. Thus, Appleton and Ingram3 in 1935established the correlatio~l between geomagneticactivity and depressions in fP2. I t is worthy of notethat in the period under discussion the inversecorrelation between K and AfF2 is, as Lewis andMcIntosh point out, considerably less striking thanth at between p and AfE (cf. Figs. 1 and 2 in ouroriginal article) . It would seem, then, that if statisticalanalysis can be successfully applied to show tha t thereis no significance between the variations in p andAfE, it is, a fortiori, evident that a similar analysismight, in the present instance, be used for discreditingthe established relationship between K and AfB'2.Conversely, of course, the fact that a phenomenonappears to be statistically sign5cant over a shortperiod must likewise be treated with reserve. Theneed for the utmost care in the application andinterpretation of statistical analyses to such a limitedtime series is thus clear.From inspection of our graphs it seemed to usthat , so far a s p and AfE were concerned, the periodwas unusual in three respects : (i) there appearedto be four oscillatfons in ground pressure showing aprogressive diminution of amplitude, with an averageperiod of about 27 days ; (ii) in like manner thereappeared to be four marked oscillations of periodabout 27 days in AfE ; (iii) oscillations (i ) and (ii)appeared to be almost exactly out of phase. I naddition, we noted that the period was characterizedby a n unusual 27-day recurrence of great suddencommencement (S.C ) magnetic storms.In our original communication we merely directedattention to these matters, and suggested that there

U R Emight be some connoxion between them; We didnot then suggest, nor do we now suggest, th at froma period of length only four months any conclusionscan be drawn regarding the general behaviour overa long period of any of the geophysical parametersconsidered. The severely limited number of observa-tions available, together with the fact that there isconsiderable uncertainty about the correct statisticalapproa,ch to tim e series analysis, seemed to ussufficient reason for not entering-into an extendedstatistical analysis.However, the contrary conclusions reached byLewis and McIntosh (see below) have prompted usto re-examine the data . Briefly, their conclusionsare : (i ) the 27-day oscillation in ground pressureis of no significance, since the amplitude is no moreth an would be expected from mere chance considera-tions ; (ii ) the 27-day oscillation in AfE is probablysignificant ; (iii) oscillations (i) and (ii) a re exactlyin anti-phase ; (iv) there i$ no significant correlationcoefficient between the p and AfE data ; (v) 0111-conclusions arise from smoothing of-the data.We shall now outline our own analysis. I n variouscornrn~nications~-~,endall has made it abundantlyclear that most of the methods generally used forstudying periodicities in time series (for example,periodograms, Fourier analysis, etc.) may yield verymisleading results when applied to the k ind of timesaries with which we are here concerned. He hasalso questioned the reliability of the usual significancetests for periodicities when applied in time seriesanalysis. Kendall has shown th at the most reliableapproach is that of serial correlation coefficients asexhibited in the correlogram. He points out tha talthough the correlogram may be insensitive, it doesgive a ower limit to the oscillatory effects, and tha.$if it oscillates there is almost certainly some system-atic oscil!ation in t h e pr in~ ary serics explored.Pigs. 1 and 2 show the correlograms for Ap and AfErespectively for the period under consideration. Inboth of these the original unsmoothed data havebeen used.It is important to note that there is a markedtrend in the pressure data, and to eliminate this wehave dealt with values of pressure departures, Ap(as with the fE dat a), rather tha n w ith the absolutemagnitudes p. The oscillations in both correlogramsare clear, with a maximum a t 26-27 days in eachcase. These correlograms provide strong support forour original deductions (based, as they were, on simpIeinspection of g raphs) , and make i t essential for usto repeat Lewis and McIntosh's calculations.At the outset we must again stress that the pressuredata exhibit a marked downward trend (approx-imately linear), and it is imperative initially toeliminate this before proceeding with any numericalanalysis. I t appears that Lewis and McIntosh haveoverlooked this point, and as a resuIt have arriveda t quite contrary conclusions. This will be clearfrom an exam ination of Table 1, in which we presentthe results of calculations made by us using (i)pressure, p, (ii) pressure departures, Ap, and (iii) fEdepartures, AfE. The nomenclature employed( c , cp, a, etc.) is that used by Lewis and McIntosh.Without going into details, it can be stated thatthere is little significant difference between thepresent results using pressure, p, and those given byLewis and McIntosh. The slight differences in thevalues of amplitude c and first serial correlationcoefficient r , are of no signseance and can be ascribedto different ways of deducing the am plitude and phase

genetical implications of the structure of deoxyribonucleic acid watson and crick nature 1953

Documents