W"""'--

:T.

"".."

~.:j

j

==r.

I

j

680

demonstrated a shielding effect of the ribosome on nascentproteins; ribosome-boUnd nascent chains 30-35 residueslong were found to be protected from external proteolyticattack. Protamine, 31-32 residues long, may thereforebe protected from attack by a putative methionine-removing enzyme tmtil the chain is completed andreleased into the cytoplasm. Other contributory circum-stances may be the d.i:fficulty of cleavage of the Met-Propeptide bond; in general, X-Pro bonds are resistant tomany proteases with the possible exception of imidodi-pepti~~s~3O; - but this enzyme, is specific only for dipep-

. .--tidei,i': Also, as shown by Ling and Dixon2., at the stageof differentiation at which protamine synthesis is maximal,there has been a considerable reduction of the spermatid

. cytoplasm and ribosOJ;:ne content. The levels of cytoplas-mic enzymes; incl1J.ding the methionine-removing enzy.me,may therefoF6:.];>ecome limiting at this stage.

TJ;le :involvement of methionine in the initiation ofprotein synthesis in eukaryotic cells is not yet firmly

. est~blislied although the presence of :,i formylatablespecies of Met-tRNA with the propertie,{of an initiatortRNA in mouse ascites cells31,32. is: certainly consistentwith such a role. Our ,observations of methionine involve-ment in the sy.nthesls:.19f the very ullu8uaJ, sperm-specificpolypeptide, protamine, suggest .Pli?-t,)'-lliih a mechanismmay be of widesprea.d importance iii'::e~yotes, and inthe special circunistances of pr:otamine' biosynthesis,where removal of the. N-terminal methionyl residue maybecome limiting, the transient incorporation of this,amino-acid is readily observed.Received:May12; revisedJune 19,1970.

1 Marcker,K. A.,and Sanger,F., J. Mol.Bioi.,8, 835(1964)..Adams, J. M., and Capecchi, M. R., Proc. US Nat. Acad. Sci., 55, 147'{1966).

>.

-

NATURE VOL. 227 AUGUST 15

'Webster, R. E., Engelhardt, D. L., and Zinder X D piSci., 55,155(1966). ,-. .,. roc.VIS], aDld Spring Harbor Symp. Quant. Bioi., 31 (1966); 34 (19', Smith, A. E., and Marcker, K. A., J. Mol. Bioi 38 ~ ' I (69). '"

,. ", -~ 1968).

, Miller, R. L., and Schweet, R., .Dorch'..,Biochem. Biophys.,125 "7 Heywood, S. M., Nature, 225, 696 (19,0). ,632

'Caskey, C. T., Redfield, B., and Weissbach, H. 4J'eh B .120, 119 (1967). ' . !Dehom.

9 Takeishi, K., Ukita, T., and Nishimura, S., J. Bioi Chem(1968). . .,:10 Caskey, C. T., Beaudet, A., and Nirenberg, M., J. ],fDI.Bi I 3~11Ando, T., and Watanabe, S., Int. J. Protein Res 1 . 9~1 (

°19

"6' t, 9

" ,-- 9)"Ling, Y., Trevithick,J. R., and Dixon, G.'R., CanadJ B . .(1969). '. . 'Dth.:..

13Marushige, K., and.Di:r-°n, G. H., De~elop. Bioi., 19,397 (1969), '"Ingles, C. J., Trevlthick, J. R., SlIUth, j\>1.,and Dixon G R ',Biophys. Res. aDmmUn.,22, 627 (1966). . " . .,'

10 Ingles, C. J., and Dixon, G. H., Proc. US Nat. Acad. Sci.. 58 1016 Sung, Mo,and Smithies, 0., Biopolymers, 7, 39 (1969), '11,17 Marushige, K., Ling, Y., and DL-.:on,G. H., J. Bioi. Chern .(1969). , " .13JergiJ, B., and Dixon, G. H:, J. Bioi. ahem., 245, 425 (1970).19 Gray, W. R., in Methods ,n Enzymology, 11, 139 (Acadenuc p-York, 1967). ,

"Anderson, G. W., Zimmerman, J. E., and Callahan, F )f 'ahe:m.Soc., 86, 1839 (1964). . ."

" Sheehan, J. C., and Yang, D. R., J. Amer. ahem. Soc., 80, 115~"Baliga, B. S., Pronczuk, A. W., and Munro, H. N., J. Bioi C4480 (1969). . ,

23 Kaji, A., Kaji, H., and Novelli, G. D., J. Bioi. ahem., 240 1(1965). "

" Momose, ~., and Kaji, A., J. Bioi. ahe:m.,241,32940(1966).",Kaji, K., Novelli, G. D., and Kaji, A., Biochim. Biophys. Acta(1963)., '" Ling, Y., and :Dixon, G. H.; J. Bioi. ahem., 245, 3035 (1970)." Rich, A., Ei1,<:enberry,E. F.~ and Malkin, L. 1., aDld Spring H,Quant. Bioi., 31, 303 (1966).23 Malkin,L. 1., and Rich, A" J. Mol. Bioi.,26, 329(1967)." Schmidt, P. J., Mitchell, B. S., Smith, M., and Tsuyuki, R., Gen:~Endocrinol., 5, 197 (1965)..0 Davis N. C., and Smith, E. L., J. Biol. ahem., 224, 261 (1957).31 Smith, A. E., and Marcke!-, K. A., Nature, 226, 607 (1970).32Brown, J. C., and'Snlith, A. E., Nature, 226, 610 (1970).

~Ieavage of Structural Proteins during the Assembly of the.Head of Bacteriophage T4 c,

byU. K. LAEMMLIMRC Laboratory of Molecular Biology,Hills;Road, Cambridge

Using an improved method of gel electrophoresis, many hitunknown proteins have been found in bacteriophage'T4 andof these have been identified with specific gene products.major components of the head are cleaved during the procassembly, apparently after the precursor proteins have asselinto some large intermediate structure.

BACTERIOPHAGESof the T-even type are complex struc-tures containing many different proteins and specifiedby many genes. Using an iIp.proved technique of electro-phoretic separation I have'found that the phage particlecontains at least twep;ty -eight components, eleven of whichare in the head. In the course of identifying the genesspecifying these proteins I discover,ed that four majorcomponents of the head, the product of gene 22, 23, 24and a protein called IP of unknown genetic origin arecleaved during the process of assembly. The head ofbacteriophage T4 is therefore no longer a self-assemblysystem in the narrow sense, because the bonding propertiesof the various components become altered, during theassembly process.

The product of gene 23 is the principal protein com-ponent of the head of bacteriophage T4 (refs. 1-3). Twominor components of unknown genetic origin have alsobeen found in capsids.,3. Besides the product of gene'23, the products of genes 20, 21, 22, 24, 31, 40 and 66 arerequired to determine the size and shape of the head-shell(refs. 4 and 5 and unpublished work of F. A. Eiserling,

E. P. Geiduschek, R. H. Epstein and E. J. Metterseveral of these genes shape-specifying functionsb~en tentatively assigned 5. Gene 22 is associated widiameter selecting (initiation) process of head formgene 66 with the elongation of the particle and geland 40 with the formation of the hemisphericaGene 31 somehow modifies or activates the major Slfor. ordered assembly.. Ten more proteins, the pr(of genes 2; 4,13,14,16,17,49,50, 64 and 65, aretto control later steps in head formation 7.

Structural Components of the PhageMany phage proteins can be separated w:ith o~

proved metl10d of disk-electrophoresis in sodium.d<sulphate (SDS). This system, to be descriJ:>edIIIelsewhere (U. K. L. and J. V. Maizel), combmes tI;'resolution power of disk-electrophoresis8 with th;ec~ity of SDS to break'down proteins into their mdi'polypeptide chains9. The proteins are also se]according to their molecular weight as was first rtfor a continuous systemlo. All the proteins the

AUGUST 15 1970 681

Fig. 1. Autora,liogram of "C-Iabelled T4 phage proteins separated in acrylamide gels. "C-amino-acid-Iabelledpreparations were analysed in 10 per cent aorylamide gels containing SDS. a, Wild type lysate; b, purified heads;c, purified phage particles; d, purified" ghosted" phage particles; 8, supernatant of" ghosted" phage particles;f; "early labelled" phage particles. The prefix P is used to designate the protein of a particular gene: for example,P20 stands for the product of gene 20; the asterisk indicates that th;s protein is derived from a large precursor,

protein and has become modified during head assembly. ."G-Iabelled lysates. Ten ml. cultures of Escherichia coli Bb (the restrictive host for phage carrying amber' mutations)in 'M9' medium" grown at 37° C to 2 x 10' cells/ml. were infected with the various phages at a multiplicity of fiveand superinfected at 8 min with the same phage and the same multiplicity to ensure lysis.inhibition. Two ",Ci of"C-amino-acid mixture ('CFB 104', Radiochemical Centre, Amersham) with a specific activity of 45 mCi/m atomwas added to each c'ature 13 min after the first infection. Three mi. of 3 per cent 'casamino-acids' mixture (Difco)was added to eaci . sample at 30 min, and the infected cells were concentrated. by low speed centrifugation35 min following iLfection. ,The pellets were drained and directly resuspended in "final sample'! buffer (see gel

electrophoresis). :"G-labelled phage and head pm.ticles. Ten ml. cultures were grown and infected as described above ("C-labelledIysates). .The double matant (B255-E18) in genes 10 and 18 was used for production of tailless heads. Ten ",Ci ofthe "C-amino-acid mixture was added 13 min after the first infection to each culture. The infected cells wereconcentrated by a low speed centrifugation 35 min aftel: infection and the pellet was resuspended in 1 ml. neutralphosphate buffer containing 10'" M MgSO" 20 ",g deoxyribonuclease and a drop of chloroform. The pellet wasresuspended by repeated pipetting, and incubated for 15-30 min at room temperature before layering on a CsClstep gradient'. The latter was prepared in tubes for a Spinco 'SW50"rotDr, with 0'8 ml. layers, and the followingdensities starting at the bottom of the tube: 1'55,1'4&,1'38 and 1,29 g/cm3. Furthermore, a 10 per cent sucrosesolution (in neutral phosphate buffer and 10'" M MgSO,) w"!! layered 'on the last CsCl step to prevent precipitationof soluble proteins at the CsCl interface. Centrifugation' was for 1 h at 40,000 r.p.m. The phage and heads, whichform a sharp visible band two-thirds down the tube, were collected through the bottom of the tube and dialysedagainst water. The band containing the heads was always viscous, indicating that the heads lost their DNA in CsClalthough the DNA was still confined within the band. Occasionally the heads lost their DNA before centrifugationparticularly in more concentrated lysates, which had to be treated with deoxyribonuclease for a much longer time."Early labelled" phages were prepared identically, but the label was added 1 min and chased 6 min after infection

by the addition Or3 ml. of 3 per cent' casamino-acids'."Ghosted" phage particle. and supernatant of "ghosted" phage particles. Crystalline NaCl was added to "C-labelledpuri'lied phage particles in water to a final concentration of 5 M. The preparation was then repeatedly frozen in asolid CO,-acetone bath ana tbawed in warm water. This procedure releases the DNA and its internal proteins fromthe phage head. The sample was dialysed against neutral phosphate buffer containing 10-' M MgSO. and the DNAdigested by the addition of a small amount of orystalllne deoxyribonuclease. The ghost was finally separated fromthe soluble proteins (interIll\l..proteins) by centrifugation and layered on a step gradient identical to that alreadydescribed. Centrifugation wMat 35,000 r.p.m. for 1 h. The tDP of the gradient containing the internal proteins wascollected with a pipette and the ghosts, which band at an approximate density of 1'3 g/cm3, were collected throughthe bottom of the tube. SD!>at a final concentration of 2 per cent was added to both samples and the samples

were dialysed against water containing 2 per cent SDS.Gel electraphores'is. Gels containing 3 per cent (stacking gel), 8'0 per cent or 10 per cent acrylamide were preparedfrom a stock solution of 30 per cent by weight of acrylamide and 0'8 per cent by weight of N,N' obis-methyleneacrylamide. The final concentrations in the separation gel were as follows: 0,375 M Tris-HCl (pH 8,8) and 0.1 -percent SDS. The gels were polymerized chemically by the' addition of 0.025 per cent, by volume of tetramethyl-ethylenediamine (TEMED) and ammonium persulphate. Ten"em gels'were prepared in glass tubes of a total lengthof 15 cm and with an inside diameter of 6 mm. The stacking gels of 3 per cent acrylamide and a length of 1 cmcontained 0'125 M Tris-HCl (pH 6,8) and 0'1 per cent SDS and were polymerized chemically in the same way as forthe separating gel. The electrode buffer (pH 8'3) eontained 0'025 M Tris and 0-192 M glycine and 0'1 per cent SDS.The samples (0'2-{)'3 ml.) contained the final concentrations ("final sample buffer"): 0'0625 M Tris-HCl (pH 6'8),2 per cent SDS, 10 per cent glycerol, 5 per cent 2-mercaptoethanol and 0.001 per cent bromophenol blue as the dye.The proteins were completely, dIssociated by immersing the samples for 1'5 min in boiling water'. Electrophoresiswas carried out with a current of 3 mA per gel until the bromophenol blue marker reached .the bottom of the gel(about 7 h). The proteins were fixed in the gel with 50 per cent trichloroacetic acid (TCA) overnight, stained for 1 hat 37° C with a 0'1 per cent Coomassie brilliant blue solution made up freshly in 50 per cent TCA. The,gels werediffusion-destained by repeated washing in 7 per cent acetic acid. Autoradiograms of gels were prepared by amodified version (U. K. L. and J. V. Maizel, unpublished) of Fairbanks et aI." (autoriLdiograms are shown in

Figs. 1--7).

which was determined are labelled in Fig. 1,.molecular weights are listed in Table 1. Atands can be distinguished in the autoradiogram

~ctively labelled, purified phage particles (Fig. lc).'roteins found in dissociated phage particles do notproteins with molecular weights less than aboutThose proteins are not sieved in..gels of 10 p,er

'ylamide (unpublished results of U. K. L. .and. el) and migrate with th,e marker dye. In gels

acrylamide concentration another three lowweight proteins have been separated (resultsl). The largest protein in the phage has an

late molecular weight of at least 120,000 and no9"Sat the top of the gel, indicating completeon of the particles by the method used.of these 28 proteins are found in the purified

paration (Fig. Ib). The rema~g 17 proteins'om the head but pres~nt in the whole phagere presumably structural proteins of the tail ands.The complete absence of these proteins from. gel pattern indicates the high degree of purityeparation. The classificaj;ion of the proteins intohead components_may not be valid for proteinsup th~ head to tail junction.iWOIlliIlor proteins besides the major components, found by others2,3 in phage capsids plll'ified inway but fractionated in urea gel~- The larger

)f proteins found in SDS gels is to be expected,,east 46 genes are known to affect T4 morphogen-~thoug!1 the proteins of these genes may not all be~ated mto the particle. .

ge! pat~ern of a total lysate of wild type infectedlioactlvely labelled at late times is presenteda. Note that most of the resolved proteins that

are synthesized late in infection are structural phagecomponents. A few rather intense bands (Xl and X2),however, are completely missing in the phage particles,also demonstrating that the separation .of the phageparticles from the soluble proteins is complete.

When phages are subjected to osmotic shock a numberof proteins are released". The 'principal one is'IP*, aninternal protein (Fig. 1). It is present in complete phageparticles, is extracted from the phage particles by fioeezingand thawing in high salt concentrations and is quantita-tively recovered in the supernatant. The 1!p.perna~ant frac-tion also contains many minor components which are foUndboth in phages as well as phage ghosts, indicating that the

Table 1. MOLEOULARWEIGHTSOF PHAGEPROTEINSDETERMINEDBY OOM-PARINGTHEIR MOBILITYIN SDS GELSWITH THOSEOF lIiARKERPROTEINS

WITH KNOWNMOLECULARWEIGHTS"'"Gene product Observed value in SDS gels Published value

. P18 69,000 50,0003P20 63,000P23 56,000P23* 46,500P24 45,000P24* 43,500P22 31,000IP 23,500IP* 21,000]?19 18,000

The following marker protein,; were used: serum albumin (68,000),y-giobuIin, heavy chain (50,000), ovalbumin (43,000), y-giobuIin, light chain(23,500) and TMV (17,000), to calibrate 10 per cent aorylamide gels. In thisimproved gel system the dIstances of migration' of the various markerproteins relative to the dIstance of migration of bromophenol blue are also alinear function of the logarithm of the molecular weight of the markerproteins, as has been described".". Radioactively labelled phage proteinswere mixed with uniabelled marker proteins. befor& electrophoresis; tiledIstance of migration of the phage proteins was determined from the auto-radiogram and those of the marker proteins from the stained gel. It isassumed that the ph2ge proteins also separate in SDS gels solely accordingto their molecular weight.. .

46,000"

T

682

P18-

-IP"~

- P20

P23 -P24 - - P23"

- P24"

','

P22 -

lP-

P19-

a b e 9 hdc

Fig. 2. Identification of gene products on 10.per cent acrylamide gels.The lysates were prepared as described in Fig, 1 and analysed on 10 percent acrylamide gels. a, Gene 2o.-defective lysate, mutant N5o.; b, gene21-defective lysate, mutant N9o.; c, gene 22-defective lysate mutantB27o.; d, gene 23-defective lysate, mutant Hll; e, gene 31-defectivelysate, mutant N54; f, gene 24-defective lysate, mutant N65; g, lysatefrom wild type infected cell; h, gene 18-defective lysate, mutant E18;

'i, gene 19-defective lysate, mutant E1137. Mutants N9o. (gene 21), E18(gene 18) and E1137 (gene 19) carried second mutations in gene 10(mutant B255). The following amber ,fragments may be detected. Arather intense band just below P23 is consistently seen in 18-defectivelysates, and is presumed to be the amber fragment of mutant E18 ,ingene 18. Furthermore, the gel pattern of defective lysates of all doublemutants in genes 18-10., 21-10., and 19-10. possess a band just aboveP23..which is probably the amber fragment of mutant B255 in gene 1()..It is striking that the amber fragment of mutant ;B27o.gene 22 behavesanomalously in the gel: it migrates more slowly than the wild typeproduct, although SDS gels are known to separate on the basis ofmolecular weights. This anomalous behaviour of certain proteins will

be discussed elsewhere (U. K. L. and J. V. Maizel).,

separatioJ1. af the released prateins and the ghastedparticl~s by centrifugatian was nat camplete. Same afthese prateins, hawever, are extracted into. the supernatantquantitatively by the freezing and thawing pracedure.

Only the pratein IP*, and perhaps same law malecularweight prateins which migrate with the marker dye,are structural phage campanents synthesized at, earlytimes (early prateins). This can be seen in the gel patternaf purified phage labelled at early times anly (Fig. If).IP* is also. labelled at late times. All the ather structuralphage campanents are synthesized anly late in infectian.Same af the principal late prateins shaw up an the auta-radiagram, prabably because af residual incarparatian afradiaactive amina-acids fallawing the chase with un-labelled amina-acids.

Identification of Gene Products

The praducts af genes IS, 19, 20, 22, 23 and 24 wereidentified by camparing the gel pattern af extracts af cellsinfected with wild type phage with thase infected with

, amber mutants in variaus genes. The identificatian af thetail and tail fibre prateins will be described later (.T.Kingand U. K. L.). Amber mutatians praduce anly fragmentsaf the pratein chain af the mutant gene an infectian afrestrictive bacterial. These fragments J:nigrate differentlyin the gel fram the camplete prateins. The malecularweights af the prateins identified are listE)din Table 1.Fig. 2 is the autaradiogram fram dried and sliced gels far

.~=

NATURE VOL. 227 AUGUs,

variaus mutants. The praduct af

by its absence in the gel pattern af ag;~ed 20 is i<..~Figs. 2 and 3a), and the praduct af ge~e-2~fect~\"e~Ir: the gel pattern af a 22-defective lysate (F~Y Its a~,Nate the amber fragment af mutant B'>7 ~. 2 aiiIThis £:agment was also. identified by Ha; g 11l ge~4thaP2 m urea gels (see leg:nd to. Fig. 2). a a and~, The praduct af gene 23 ISeasily identified bv '

m the gel pattern af a 23-defective lysat (F,lts Ican also., be seen that P23 averlaps with etwaIg. .~campanents. If the 23-defective lysate is anal rnuaf lawer acrylamide cancentratian anath' YS,edabservatian' is made. A band, P23* whi ~"I!llt

in variable amaunts in ,the ather head de£c ,ISdeis ~ampletely missing in the 23~defective ly:~~.vel~This band, P23*, averlaps with P24 in Fi e ~Flg.better separated fram P24 in less cance~' -, ~u(Fig. 3). As with the praduct P23 the r ~rate1,was identified by its absence in th~ gel p~t~ Uct ofdefective lysate (Figs. 2f and 3e). ern ata

So. far; no. missing bands have been faund in ,:patterns af 31 ar 21-defective lysates (Fig. 2b dtlanalysis af the 21-defective lysate an gels af laan!

'

d '

(F' wer

amI e cance~tratIOn ,lg., 3b), which resalves oJmalec~ar ,w~lght pr~tems b~tter, clearly shaws tlband IS mlssmg. This pratem, hawever, is the n?f gene 10, a. baseplate gene. T~e mutant N90 in~em fact carrIes a secand mutatIOn in gene 10 (B255). II

In camparing the gel pattern af the head~def~ysates (Fig. ~ar-f) with that af wild type (Fig. 2g),lffipartant differences are abserved, which she(an the precursar-praduct relatianship af the heapanents.

The principal fractian af the gene 23 praductmal~cula~ weight af 56;000 in a~l the head-defective l,ybut.l~ Wild-type ar t~11-defectIve lysates it appears'pasltIOn af P23*, With a malecular weight of 4Small but significant amaunts af P23* are alSo.0.6;in head-defective lysates (Fig. 3a-f). In lysates pl'identically abaut 20 per cent af the tatal P23 is canto. P23* in the 20-defective lysate, 10 per cent in j2-3 per cent in 22, 24 and 31-defective lysates, asmined fram densitameter tracing af the autoradiagra

The bands P22 and IP are bath absent ar cansiderless intense in the wild type gel pattern (each awwith two ather prateins). A new band, IP*, is seen ai,battam af the gel, which is campletely missing in the!defective lysates (Fig. 2a-f).

The band P24, which is faund in all head-defeclysates is missing in the wild type pattern, but aband, P24*, which migrates slightly faster is abserThis is difficult to. visualize in Fig. 2, but will becevident in Fig. 6.

AlSo. included in Fig. 2 is the gel pattern af two.defective lysates. The praduct af gene 18 (!ll?l.69,000), the principal pratein af the tail sheathl:, ISId(fied by its absence in an IS-defective lysate (FIg. 2h)the praduct af gene 19 (mal. wt IS,OOO) by its absence19-defective lysate (Fig. 2i). P19 is thought to ~chief campanent af the tail tube". This demanthat the differences in the head prateins are nat J ,to. tail attachment, far the gel patterns afthe taIl.defe~lysates are identical with that af wild type:

Evidence will be presented that the pratelns P23,P24 and IP are cleaved in wild type infected cells ~(precursars to. prateins P23*, P24* and IP* faund9~final head structure. (The cleavage productaf P~'"nat detected.) This precursar-praduct conver:'tstrangly inhibited by mutatians in g,enes 20, 21,dJ24 and 31. Two. impartant canclUSIans can ?6 Ie(a) the aberrant head-related structures-smg

tr'

t. conditioRS wi".Henceforth, lysates of cells infected at res ICI~e. a "21-

mutants in various genes will be referred to as, for Instance,lysate", where the amber mutant used was in gene 21.

VOL. 227 AUGUST 15 1970

1~iIti-la.vered polyheads.,5,15, 't"-particles4,5,1. a.nd lumps'hnl Dl to be produced in these mutant infected cells-

i~flYconsist of th: precursor prot~in P23; (b) because9'1{S not cleaved III these mutant mfected cells, but is,,- ired for polyhead and 't"-particle forma.tion, as has

establ~shed by geI),etic means5, it is strongly suggested. P22 is incorporated into these structures as such.

etics of the Cleavage R.eactionshe following exPeriments were designed to study the

;cursor relationship of the proteins P23, P24 and IP ,

'thP23*, P24* and IP*, respectively. In this experi-l~nt ,the infected cells were pulse-labelled with radio-l've amino-acids for a short period (1 min) and the

difica.tion of the various proteins was then followed bylysis of the samples taken at intervals in SDS gels.e'results of autoradiography of the dried gels aresented in Fig. 4:IZeavageof P 23. It is readily seen in Fig, 4 that most of theprotein is at the position of P23 immediately' following the5e of the radioactive amino-acids. It then rapidly disap-Jis, and a new band, P23*, appears simultaneously. That

is cleaved and gives rise to P23* is suggested by the fact;'both are principal components and this is reinforced byabsence of P23 and P23: from the gel patterns of a .23,tive lysate (Fig. 3d). The kinetics of the cleavage reaction'+P23* are plotted in Fig. 5a. Cleavage is very r.apid:

;iit 50 per cent of the precursor is cleaved within the :6.rstlin following chase of the label. P23* appears at ab'out theie'rate, in a satisfactorily correlated way. The total labelled

~

,ein in P23 and P23* is als!) ,plotted in Fig. 5a. The totalel increases during the :6.rst minute, which reflects the com-ion'time of the chase of the labelled amino-acid, but :finally

~total falls off by 25-30 per cent. This :final decrease of the., labelled protein can be nicely explained, for a cleav,age

about 56,000 to 46,500 corresponds to a loss of about'I'cent by weight of protein.:avageof F22 and IF. In the pulse.chase experiment of

. 4" two other protein bands, P22 and IP, disappear withe. (The disappearance of P22 in wild type infected cells

20- 2f 22 23 24- 31-

P23

P23.P24

P22

, a b C 'd e f

.~r.ig.3. Identitication of geneproducts on 8 per cent acrylamidegels." '8C-labelledlysates were prepared as described (Fig. 1) and analysed on

b per cent acrylamidegels. a, Gene20-defectivelysate, mutant N50;ni tnet 21-defective lysate, mutant N90; c, gene 22-defective lysate,d u ~ B270; d, gene 23-defective lysate, mutant Hll; e, gene 24-efective lysate, mutant N65; J, gene 31-defective lysate, mutant N54.I'i

P23-

P23~P24~

'OfI ,

683

-P24-

P22-

IP--Ip.

15' 16' 17''18' 19' 20' 22' 24 26' 30'

Fig. 4. Cleavage of products of genes 22, 23, 24 and protein IP (10 percentacrylamide gels). A 10 ml. culture grown at 370 C was infected witha clouble mutant defective in genes 10 and 18 (mutant B255 and E18)as described (Fig. 1). The radioactive amino-acid mixture (10 I'Ci)was added 14 min after the first infection, and chased 1 min later with anexcess, of uulabelled amino-acids (final concentration 1 per cent). Thechase of the label was verified by measuring the counts in the total TCAprecipitable proteins.' One mI. samples were prepared at intervals afterthe chase and immediately frozen in a solid CO,-acetone bath. SDS wasadded after thawing to a final concentration of 2 per cent and the sampleswere carefully dialysed into 2 per cent SDS in water. The samples werefinally mixed with an equal volume of twice concentrated "final samplebuffer" and boiled for 1 min before electrophoresis. The sampling timeis, indicated at the bottom of the gels. All the preliminary experimentsw~re done with wild type phage, but this experiment was performed withthe double mutant in genes 10 and 18 in view of plans for future

experiments. .

J:!as also been observed by M. Showe, personal communica-tion.) A new band, IP *; appears at the bottom of the gelpattern. The kinetics of these cleavage reactions are plottedin Fig. 5b. P22 disappears with approximately the same initialrate as P23: about 50 per cent is cleaved 2-3 min followingchase of the labelled amino-acids. I have not found a bandin the gel pattern which may be derived from P22. Rosodaand Levinthal12 reported indirect evidence that P22 is a struc-tural phage component, but they considered the possibilitythat P22 might become altered during head formation.

Evidence for the precursor-product conversion IP---IP* is'provided by the observation that the disappearance of I;E' andthe appearance ofIP* is coordinated in time (Fig. 5b). Further-more., the total label lost at the IP position is recovered in the- :IP* 'band. The reaction is slower than that of P23 and P22:Only about 50 per cent of the total label at the IP positiondisappears. This could be explained either by another protein '

band overlapping with the IP band or synthesis in excess.IP* cannot arise from P22. IP* must be derived from a

precursor which is synthesized early, for I have shown (Fig. 1)that IP* is strongly labelled in the "early labelled" phagepreparation. P22, however, is reported to be synthesized atlate times only12, which I have confirmed. The precursorconversion of IP->-IP* has also been observed in a pulse-chaSeexperiment in which the label was added between 4 and 5 minfollowing infection, thus labelling only early proteins (resultsnot shown). The hand IPV[.as easily recognized in these gelsand the disappearance of IPand the appearance of IP * wereagain correlated in time. This quantitative agreement and theabsence of other unaccounted changing bands in the gel patternsupport the argument for the IP->-IP* relationship. Of course,a :final proof awaits chemical analysis.

It was also observed that, although the pulse was performedbetween 4 and 5 min after infection, cleavage of IP starts only

684.

2'0

)J..,/ ,I ,

I 0\0\

\,

\ ,.,. 0 0o>~'-1:f- - --0- - -~"U o,,

II

I,

a

""§~~,.Q..; ,

II.

II

'1I,

"I

Labell,,'

1.0

14 15 11; 2618 20 22 24

Time after infection (min)

b

""§

.1; Q-4'"~ . .

1)'3

0 0-----00.2

O'I~ ~5'-_P-- 0Labell,.tJ

J.

18 20 22 24 26Time after infection (min)

Fig. 5. Kinetics of cleavage of P23, P22 and IP. The kinetics of'cleavage were measured using a microdensitometer (double beam record-"jng microdensitometer, Joyce-Loebl) to record the autoradiogram. Theexposure of the autoradiogram was chosen so that the absorbance of theband to be measured did not.exceed 1 unit. -The abscissa represents theintegrated absorbance over the relevant peaks. a: e-e, integratedabsorbance over the P23 peak; 0 -.--0, integrated absorbance overthe P23* peak; 0 --- D. total absorbance in P23 and P23*.b: A-A. integrated absorbance over peak P22; e-'e, integratedabsorbance over peak IP; 0 ---0, integrated absorbance over peak

IP*.

14 15 Hi 28 30

,,II

at late times (after 17 min). Phage assembly starts at aboutthis time and it is therefore thought that the cleavage of 11'is linked to phage assembly. .

azea~age of P24. 1'24* (mol. wt 43,500) appears coordin-ately with 1'23* and 1'22* (Fig. 4) (the kinetics are not plotted).The precursor product relationship 1'24-+1'24* is more difficult.to demonstrate, because 1'24 migrates only slightly faster than1'23*, but the following experiment proves that 1'24 is missingin a pulse-chase wild type lysate. 1'24 separates somewhatbetter from 1'23* in a gel of lower concentration. The samplesof the pulse-chase exp~riment were analysed on 8 'per centacrylamide gels and four time points are presented in Fig. .6.1'24 is easily distinguished from the small amount of 1'23*existjng immediately after the chase of the radioaCtive label(Fig. 6, 15 min). 1'24 disappears at the same time as 1'24*appears while 1'23* increases. One might argue that 1'24 isobscured by the heavy 1'23* band. This possibility wasexcluded by adding a lysate (23-defec.tivelysate) containing1'24 to the final samples of.the pulse-chase experiment. 1'24 wasthen detected and it can be concluded that measurements of1'24 are reliable, thus showing that 1'24 mo~ likely gives riseto 1'24*. The integrated absorbance values- over these twobands are about equal, but the small loss of protein weight bythe 1'24-+1'24* reaction is not likely to be detected by thedensitometric measurements. 1'24 could not give rise to 11'*,because the total label in IP* is two or three times larger than.that of 1'24 and 1'24 is synthesized late. Moreover, the pre-cursor relationship of 1'24-+1'24* is considerably strengthenedby recent observations on head maturation genes (myunpuh.lished results). 1'24 does not seem to be cleaved at all in.50-

'"

""""",

NATURE VOL. 227 AUGUST 15 1970

defective cells and, indeed, no 1'24* is found. Cleavage of P931'22 and IP does occur in 50-defective cells, although at a red\].; ,rate. ~

Fate of the Small Fragments

Where are the small fragments of these cleavage reactioThe expected molecular weights for the small fragments st ns?ming from p~3, 11' and ~24 would. be .about 10,000, 2,500 :/!).1,500 respectIvely. Peptides of this SIze are not sieved on ~dper cent acrylamide gels and migrate with the marker d 0(unpublished results ofU. K. L. and.T. V. Maizel). Atte-'lpts {efind at least the 10,000 molecular weight fragment. from P'>3 0gels of higher acrylamide concentration have failed. Po;sib~nthe fragments are further broken down to undetectable size:Fragments of 1'22 also have not been detected. Acid-solubl'components whic? ~ deriyed from an acid-insoluble precurso~are known to eXIst m T4 infected cellsl? Two are associatewith the phage particle and they are released with the DNlfrom the head upon osmotic shockl? The genetic dete~ tof one of these internal peptides has been mapped recently a:'dlies in the neighbourhood of genes 20 and 21 (ref. 18). 11results definitely rule out gene 20, which is incorp<;)rate~unmodified into normal phage. Unfortunately, I have notdiscovered the product of gene -2~.in the gel pattern. Theseresults do not rule out the possIbility that one of the internalproteins is derived from, the small cle~vage fragment of P22,1'23 or 1'24. The appearance of the mternal peptides seerosindeed to be coordinated with the cleavage reactions of P231'22 and 1'24. Genes 20, 21, 22, 23, 24 and 31, which affectthe cleavage of 1'23, 1'22 and IP, are known also to affect theappearance of the internal peptidesl'.

It has been pointed out to me by S. Brenner and A. Strettonthat the cleavage point must occur at the N-terminal end of the1'23 protein. In establishing the co-linearity of gene 23 andits polypeptide!, they observed peptides in 23-defective lysateswhich contain the amoer fragments, but these are absent inwild type lysates or in purified phage particles. These peptidesmay be derived from the N -terminal end of the amber fragmentwhich is cleaved off'ln the protein 1'23*: These observation~also suggest that the small cleavage fragment, with an expectedmolecular weight of about 10,000, is fragmented to even smallerpieces.

P23 -

P24-- P23.- P24.

Fig. 6. Cleavage of the product of gene 24 (8 per cent acrylami.t!eg~I>JfSome samples (15, 16, 19 and 26 min) of the pulse chase exper1lXlenntFig. 4 were analysed on 8 per cent acrylamide gels. Only the reley'

part of the gel pattern is shown.

J....

VOL. 227 AUGUST 15 1970

leavage OCCUrS in a Large StructureTDeobservation that all the proteins P20, P21, P22, P23, P24d p31 are required for efficient cleavage of P23, P22, P24 andI nggeststhat these precursor proteins aggregate first to formsJigo:rneric structure, and are cleaved subsequently, rather,0 being cleaved first, and then'assembled. The following ex-

/::""~ent supports this'view.. The experime.nt is based on the:'" ~~rvation ~hat most of the precursor protems are soluble and".oonoJIleric ill SDS at :oom temp~rature, bu~ that phage

'Jes are not totally disrupted, as IS also true ill urea12, The~ttern of purified phage treated with SDS at room tem-

.Jure is compar~d with completely ,degraded p~age, boiledI JIlinin SDS (Frg. 7a and b). Only a small fractIOn of P23*

e:s:tractedfrom phage with SDS at room temperature. Most of.eproteins stay at the top of the gel or enter' the gel as hig~

,ular weight aggregates. Some proteins are not extracted. Note, however, that a few proteins are ahnost quanti-

,ely extracted from the phage particles. The samples frompulse-chase experiment treated in SDS at room tempera-ODlyare shown in Fig. 7. P23 disappears with time but

'P23* appears, suggesting that P23 enters an SDS-resistantilcture before being cleaved. Of course, these experiments"Illotrule out the possibility that P23* is converted to anS.resistant structure so rapidly that it is not detected. .Ah :molecular weight protein appeared at the' top of the gel,ich could be an aggregate of P23 but only accounts for partthe label which disappears from P23. Most of the label, 5at .the top of the gel. The molecular weight of this struc-

must therefore be greater than 300;000, for such a mole<fularIt is excluded from these gels (unpublished results of. L. and J. V. Maizel). It is possible that this stru~turecapsid-like shape. IP* and P24* are not resolved in

gels. This is because of the high salt concentration insamples ('M9' growth medium) which impairs the resolu-

of the gel in the low molecular weight region.

uration of the Head

~yexperiments demonstrate that the assembly of theof bacteriophage T4 is not a simple, straightforward,ssembly, because several. structural proteins areacally altered at some stage of .assembly. The

,leaved precursor protein P23 can, however, be poly-rized into 'single and multilayered polyheads and'mieles, if its eleavage is blocked as a result ofmutation..vestigations of genes 2, 4, 13, 14, 16, 17, 49, 50, 6465, which supposedly control late steps in head

'. .ation, are in progress. It is interesting that. the.,lavage of P23, P22, P24 and IP seems to be normal'§ells infected with phage carrying mutations in these,188with the exception of genes 2, 50 and 64 (my)ublished results). . '

hy are these structural head components cleaved?finding that IP 'gives rise to an internal IP* shedson a possible consequence of the clearvage reactions.

l~rnaI proteins are thol:lght to bind to DNA and theeavage reactions may possibly trigger the necessary'NA-protein interactions, which result in orderly packing~.the DNA within the shell of proteins. The following

del may be proposed. The proteins P20; P21,' P23,4, P31 and IP form an intermediate structure (SDS-. ;ant) which combines with an end of a DNA strand.

rage of P23, P22, P24 and IP proceeds from th{3 DNA,hment site~ During this process more and :rp.ore

binding sites may be formed at the inside of thisiure, perhaps by the formation of IP*, thus winding

"e DNA strand successively. The small acid-solubledes formed during this process may also interactthe packed DNA17 to neutralize charges. My

,ts do not decide whether DNA packing proceedsnultaneously with the polymerization. ,of the head

.lembrane or follows thereafter. '

'J thank Dr A. KIug for encouragement and facilities,Ly-colleaguesJ. King, J. Maizel and S. Altman for many

?,~luable suggestions (J. Maize! in parlicular for advice2: gel t~chniques and J. King for help in preparing the;i~~scnpt), and S. B~enner (among others) for critically'~, mg the. manuscnpt. U. K. L. holds an EMBO

'. oWShip.

685

!i

!1

1

i'1j

1~

1,i

P22

17' 18' 19' 20" 21' 23' 25' 21' 30' a bFig. 7. ChaSe of P23 into a product stable to SDS at room temperature.A culture was infected and pulse labelled as described in Figs. 1 and 3with the following changes. The culture was grown at 30° C, infected''with wild phage, and labelled fot 2 min from 1&-17 JIlin after infection.The samples were mixed at room t<>mperaturewith an equal volume oftwice concentrared "final sample buffer'" without preVious clialysis.'The samples were then directly applied to the geJSwithout being boiled.The sampling thne is indicared at the bottom of the gels. As a control apurified phage preparation is analyse<L a, Phage treated in SDS at room

remperature only; b, phage boiled in SDS for 1'5 min. "

Note added in proof. During the preparation of thismanuscript I was informed that the alteration of P23has al-sobeen observed by other workers: E. Kellenbergerand C. Kellenberger-van der Kamp, FEBS Lett., 8, 3, 140(1970); R. C. Dickson, S. L. Barnes and F. A. EiserIing,J. Mol. Biol. (in the press); and J. Hosoda and R. Cone,Proc.. US Nat. Ac.ad.Sc.i. (in the press).Received May 7, 1970.

1 Sarabhai, A. S., Stretton, A. O. W., Brenner, S., and Bolle, A.. NaJ.ure,201.13 (HJ64).

, Kellenberger, E., Virology, 34, 549 (1968).3 Baylor, M. B., and Roslansky, P. F., Virology, 40, 251 (1970)..Epsrein;R;H., Bolle, A., Sreinberg, C. H., Kellenberg, E., Boy de la Tour,'

E., Chevalley, R., Edgar, R. S., Susman, M., Denhardt, G. H., andLielausis, J., ColdSpring Harbor Symp. Quant. BiOl., 28, 375 (1963).

'Laemm!i, U. K., Molbert, E., Showe, M., and Kellenberger, E., J. Mol. Biol.,49, 99 (1970)..Laemmli, U. K., Beguln, F., and Gujer-Kellenberger, G., J. Mol. Biol., 47,69 (1970).

"Edgar, R. S., and Wt16d,W. B., Proc. US Nat. .Acad. Sci., 55, 498 (1966).. Davis, B. J., .Ann. NY .Acad.Sci., 121, 404 (1964)..Maizel, J. V., Fundamental Tec/m.iques of Virology (edit. by HabeL K., and

Salzman, N. P.), chap. 32, 334 (Academic Press, New York, 1969).10 Shapiro, A. L., Vifiuela, E., and Maizel, J. V., Biophys. Biochem. R£s.

Commun., 28, 815 (1967).11 Minigawa, T., Virowgy, 13, 515 (1961)."Hosoda, J" and Levinthal, C., Virology, 34, 709 (1968).10 King, J., J. Mol. Biol., 32, 231 (1968).11 King, J., FEBSSymp. (eclit. by Ochoa, S., Nachmansohn, D., Aseusio, C.,

and Heredia, C. F.), 21.156 (Academic Press, London, 1970).15 Favre, R., Boy de 180Tour, E., Segre, N., and Kellenberger, E., J. Ultra-

st1"l.lct.Res., 13, 318 (1965). '

" Kellenberger,. E.. Eiserling, F. A., and Boy de la Tour, E., J. Ultrastruct.R£s.,21, 335 (1968), .

" Eddleman, H. L., and Champe, S. P., Virowgy, 30, 471 (1966).10 Sternberg, N.., and ChaII\pe, S. P.. J. Mol. Biol., 46, 337 (1970).19 Laemm!i, U. K., and Eiserling, F. A., Mol~. Gen. Genet..101, 333 (1968).20Fairbanks. G., Levinthal. C., and Reeder, R. H., Biochem. Bi.ophys, Res.

Oommun.,20, 393 (1965)." Larcom, L., Bendet, r., and Mumma, S., Virology, 41, 1 (1970)." Weber, K., and Osborn, M.; J. Bioi. Chem.,244, 4406 (1969).