Location of Post-translational Cleavage Events Within F and FIN Glycoproteins of Newcastle Disease Virus

Transcription

1 J. gen. ViroL (I980), 47, Printed in Great Britain I9 Location of Post-translational Cleavage Events Within F and FIN Glycoproteins of Newcastle Disease Virus By A. C. R. SAMSON, P. CHAMBERS AND J. HELEN DICKINSON* Department of Gene tics, University of Newcastle upon Tyne, Newcastle upon Tyne NEt 7RU, U.K. (Accepted 25 September 1979) SUMMARY The biologically active form of the fusion glycoprotein F from Newcastle disease virus (NDV) comprises two polypeptides, Fj and F 2 (derived from a precursor polypeptide F0 by a post translational cleavage event), which are covalently linked together (F1,2) by disulphide bonds. This feature was exploited in a twodimensional SDS-polyacrylamide gel electrophoretic analysis to orientate the position of the cleavage event within F0. Separation of proteins from NDV-infected CEF in the first dimension in the absence of reducing agent resolved F1. 2 protein from all NDV-induced proteins other than F0. Reduction of the first dimension gel with 2-mercaptoethanol, followed by electrophoresis in the second dimension, resolved F~ (55IO, F2 02"5K) and F 0 (64K) proteins. The only polypeptides other than F1 and Fz which fell below the diagonal, indicating the positions of the polypeptides from infected cells, were two minor glycoproteins designated HN~ (5 I-5K) and HN2 (27"5K) which took up positions vertically beneath the major haemagglutinin-neuraminidase glycoprotein HN (75K). Dual isotope labelling experiments with NDV-infected chick embryo fibroblasts, which had previously received a salt shock to effect synchronization of polypeptide initiation upon release of salt shock, revealed the following orientations within the parent molecules: NH2--Fz--Ft--COOH and NH2--HN1--HN2--COOH. The existence of intermolecular disulphide bonds, orientation and relative lengths of the two NDV HN fragments is analogous to the HAx and HA2 proteins of influenza virus haemagglutinin. INTRODUCTION It has been known since I973 that the smaller NDV glycoprotein (now called F1) is derived by post-translational cleavage of a larger precursor glycoprotein (now called F0; Kaplan & Bratt, I973; Samson & Fox, ~973, t974). Subsequently it was demonstrated that this cleavage event had considerable biological significance for various paramyxoviruses in that virions which contained F0 (but not F0 could not infect host cells, haemolyse erythrocytes or mediate cell-cell fusion. However, F0-containing virions could be activated in vitro with the proteolytic enzyme trypsin, concomitant with cleavage of F0 to F~ (Homma & Ohuchi, I973; Scheid & Choppin, 1974; Nagai et al. I976). More recently it was shown that the smaller cleavage product (now called F2) remains intact and covalently linked to F~ protein by disulphide bonds (Scheid & Choppin, 1977)- This feature of the F1.2 protein is exploited here to achieve a complete separation of F~ * Present address: Department of Biology, University of York, England. oo /8o/oooo-383! ~o2.oo (~) 198o SGM

2 20 A.C.R. SAMSON, P. CHAMBERS AND J. H. DICKINSON protein from another NDV virion protein NP (nucleocapsid protein) which has an apparent tool. wt. (54K) very similar to that of F1 (55K). The previous inability to completely separate F1 glycoprotein from NP protein using SDS-PAGE has hampered the use of pactamycin (Taber et al. 197 l) or salt shock techniques (Schochetman et al. I977) to orientate the FI+F2 cleavage event with respect to the NH2 terminus of the F0 glycoprotein. The orientation of F1 and F~ within the precursor protein Fo has a direct bearing upon the insertion of the protein into membranes during translation and its retention within the lipid bilayer of the mature virion (Katz et al. ~977). The presence of disulphide bonds linking F1 and F2 is exploited here to separate F1 and F2 proteins from all other proteins found in NDV-infected CEF cells on a single SDS slab gel. To do this, a non-reduced/reduced two-dimensional SDS-PAGE method was employed. The resolved proteins were quantified using a dual-isotope labelling protocol involving salt induced synchronization of polypeptide initiation. This allows a distinction to be made between the NH2 or COOH terminus location for a polypeptide cleaved from a precursor protein. METHODS Virus. The Beaudette C strain of NDV was grown in ovo, purified and assayed as previously described (Samson & Fox, 1973). Chemicals. All chemicals were of highest reagent grade. Polypeptide tool. wt. standards bovine serum albumin (66-2K), ovalbumin (45K), chyraotrypsinogen A (25"1K) and cytochrome c ( K) were obtained from Sigma Chemical Company, London, U.K., flgalactosidase (J I6.248K ) from Boehringer, Lewes, U.K. Diallyltartardiamide was obtained from Aldrich, Gillinghara, U.K., periodic acid from Sigma. All other chemicals were obtained from BDH, Poole, Dorset. Isotopes. L-4,5-3H-leucine (57 Ci/mraol), and L-U-14C-leucine (352 raci/mmol), were obtained from the Radiochemical Centre, Araersham. Radioactive labelling of polypeptides in ND V-infected chick embryo fibroblasts. Secondary cultures of CEF cells were grown as previously described (Samson & Fox, I973) at 37 C in specially formulated medium 199 lacking the amino acid leucine (Flow Laboratories, lrvine, Scotland) to which was added non-radioactive leucine at l #g/ml and dialysed calf serum to 5 %. When confluent (2.8 lo 6 cells per 3o mm diam. dish), monolayers were mock infected or infected with NDV at an ra.o.i, of ~o to ao per cell and allowed to absorb for 3o rain at 37 C. The inoeulura was replaced with 2. 7 ml per dish of fresh medium and incubated at 37 C in a 5 % COJ95 % air incubator for 5 h before any hypertonic salt treatment or isotope incorporation. To monolayers which were to receive a high salt shock, 0"3 ml of 3 M-NaCI solution was added to the liquid medium for 3o rain. This was followed by a io s wash with isotonic medium (without the added salt) to effect release of the saltinduced block in polypeptide initiation. Immediately after this wash, o.2 ral of medium containing 3H-leucine (IOO #Ci/dish) was added to the monolayers o to 2, o to 5 and o to ~o rain post release from salt shock. To a fourth monolayer isotonic medium was present [o rain before the addition of isotope for a 5 rain period (Io to 15 rain). Non-salt-shocked raonolayers were pulse labelled with '~H-leucine for a 15 rain period. Three infected monolayers were incubated with 14C-leucine (25 #Ci total) for a period of I5 rain, 5 h p.i. One infected and one uninfected monolayer were also incubated with 3H-leucine (2o #Ci/dish) for 15 rain, 5 h p.i. All labelled monolayers were subsequently incubated for 45 rain in the absence of isotope in the presence of medium I99 containing IOO/zg/ral leucine. Following this chase period, each monolayer was disrupted with 0"5 ml of boiling z % SDS gel sample buffer containing J o% glycerol but no mercaptoethanol. Extracts were boiled for 2 rain before SDS-PAGE.

3 Cleavage off and HN glycoproteins in ND V Sodium dodecyl sulphate polyacrylamide gel electrophoresis. The discontinuous ph system of Laemmli (Laemmli, I971) was used throughout. Stock solutions for acrylamide gels crosslinked with diallyltartardiamide (DATD) contained 4o g acrylamide and 2"4 g DATD per IOO ml solution. Stock solution for acrylamide gels crosslinked with methylenebisacrylamide (BIS) contained 4o g acrylamide and I.o8 g BIS per IOO ml solution. Gels were polymerized using ammonium persulphate and NNN'N'-tetramethylethylenediamine according to the method of Laemmli 097I). Samples of all-labelled SDS extracts were individually mixed with 14C-labelled SDS extracts from infected non-synchronized cells. Dual-labelled samples were electrophoresed in a Io-well SDS: IO ~ acrylamide BIS crosslinked resolving gel with a 4 ~o stacking gel for 6 h at 2o ma/slab. Individual channels (I'5 mm x 8 mm 9 mm) were cut from the slab and stored overnight in screw-capped tubes at -20 C. Each slice of gel (which represents a first-dimension gel to be used in a subsequent second-dimension gel separation) was equilibrated for 30 rain at 25 C with Lo ml of Laemmli's sample buffer containing lo %, v/v, 2-mercaptoethanol. The gel slice was then quickly rinsed in sample buffer lacking 2- mercaptoethanol and placed on top of a I cm 4% stacking gel above a IO cm long I5% acrylamide: DATD crosslinked SDS resolving gel slab. Six gel slabs were set up and electrophoresed at the same time using three identical sets of slab gel apparatus. Reduced samples of purified NDV virions and mol. wt. standards were applied to each slab in side wells. The first-dimension gels were sealed on to the stacking gels with 1% agarose. Each gel slab was electrophoresed at 3o ma for approx. 5"5 h, stained with Coomassie blue as previously described (Samson & Fox, I973) and photographed. Gels were then dried directly on to dialysis sheeting beneath which was placed Whatman 3 mm filter paper. Gels were overlaid with a sheet of Roastabag and dried to a clear crisp using a BioRad gel drier. The transparent dried gel was tagged with x~c-inked labels (to re-locate and identify the gel with the resulting autoradiogram) and autoradiographed using Kodak X-Omat RP X-ray film for 2o to 3o days. Quantification of 3H and 14C isotopes in NDV-induced proteins. NDV-induced proteins detected on autoradiograms were identified and distinguished from host proteins on the basis of their mol. wt. Regions of the gel corresponding to HN1, HN2, F1 and F2 were cut out of the gel using a 3"5 mm internal diam. cork borer. Discs of uniform size (still apposed to the dialysis sheet) were transferred to glass scintillation vials containing I ml of 5 %, w/v, periodic acid solution. The gel discs were solubilized at 6o C for 3o min in a water bath. Ten ml of scintillation fluid (5oo ml Triton X-Ioo, ~5oo ml toluene, 8 g 2-(4-tertbutylphenyl)-5-(4-biphenyl)-I,3,4-oxadiazole ) were mixed with each extract and samples monitored in a Beckman 3 channel liquid scintillation spectrometer. Two channels were set up to count ~H and ~4C in the presence of the other isotope. A double spillover correction was applied for the measured counts and a corrected 3H: ~ac ratio computed. 2 I RESULTS Resolution of F polypeptides on non-reduced." reduced polyacrylamide slab gels Fig. I illustrates that 14C-labelled F1 and F2 polypeptides can be clearly resolved from all other virus-induced and host polypeptides when the non-reduced/reduced SDS-PAGE procedure is employed. Using this electrophoretic procedure polypeptides which do not contain inter-molecular disulphide bonds take up a position on the diagonal. This position is determined by the mol. wt. of the polypeptide. Any protein comprising two or more polypeptides covalently linked together only by disulphide bonds will take up a position below the diagonal. The number and positions of these polypeptides reveal the number of different polypeptides so linked and their individual

4 22 A. C. R. S A M S O N, P. C H A M B E R S AND J. H. D I C K I N S O N I iii Fig. i. Autoradiograph (25 days exposure) of 14C-labelled proteins (15 rain pulse, 45 min chase) from NDV-infected CEF cells separated on non-reduced I o ~ acrylamide:bis-reduced 15 ~ooacrylamide:datd SDS-PAGE slab. Each virus induced protein is identified. tool. wt. (A true diagonal is not obtained in the slab gel profile shown in Fig. I because different acrylamide gel concentrations, I o % followed by 15 %, were employed to achieve acceptable resolution of F1.2 and NP in the first dimension and to resolve F2 from the tracking dye front in the second.) In Fig. i, F1 polypeptide lies vertically above F2 polypeptide and vertically beneath a position on the diagonal which would be associated with a protein of slightly higher tool. wt. than F0, a position found for F1,2 protein in non-reduced gels. Neither F~ nor F2 nor any other virus-induced protein is found in uninfected cell extracts (not shown). Subsequent labelling experiments with glucosamine hydrochloride confirm that Fx and F2 are both glycosylated and that F2 polypeptide contains a greater fraction of this label than F1 polypeptide (Scheid & Choppin, I977). The sum of the mol. wt. of F~ and F 2 (66K) is close to that determined for F1, 2 protein on non-reduced gels (65K) and for F0 protein (64K) - see Table I. Characterization of non-reduced: reduced polyacrylamide slab gels Examination of Fig. t reveals that proteins when boiled with SDS are rendered less soluble in the absence of than in the presence of reducing agent. The large NDV protein L, for example, cannot be discerned using non-reduced gels although it is readily seen on

5 Cleavage off and HN glycoproteins in NDV 9 3 Table t. Molecular weight determinations for NDV-induced polypeptides and cleavage products Apparent mol. wt.* Protein Symbol x to -3 Large L 180 Haemagglutinin HN 75 Cleaved fusion Fx, ~ 6T]" Uncleared fusion F0 64 Larger cleaved fusion F1 55 Nucleocapsid protein NP 54 Larger cleaved haemagglutinin HN~ 51.5 Matrix M 39 Smaller cleaved haemagglutinin HNz z7"5 Smaller cleaved fusion Fz HN1 + HN~ F~+ F~ 67"5 * Determined from IO % and 13 ~ acrylamide:bis-reduced SDS-PAGE calibrated with mol. wt. markers (fl-galactosidase, bovine serum albumin, ovalbumin, chymotrypsinogen A and cytochrome c). "}" Determined from ~o ~ acrylamide:bis-non-reduced SDS-PAGE calibrated internally against HN, NP and M proteins. conventionally-run reduced gels. Significant amounts of high mol. wt. proteins appear to be trapped near the origin of the first-dimension gel. This material also hinders the mobility of some lower mol. wt. proteins in the first dimension. These then appear as horizontal streaks in the second dimension which emanate from the region of the origin of the first dimension gel. This streaking is also manifest in stained gels and is not therefore an artefact of radioactively labelled proteins. A small spot of radioactivity above the F0/F1.2 region at the origin of the second dimension gel is seen in Fig. I. In other gel runs this radioactivity may appear as a vertical streak above the F region or may not be seen at all. A spot or streak in this region is not found when uninfected cell extracts are analysed and may represent aggregates of F proteins. In early attempts to increase the solubility of proteins separated in the first-dimension gel prior to the second-dimension run, the first dimension gel strips were boiled for 2 rain in the presence of SDS and mercaptoethanol. This resulted in an increase in the apparent mol. wt. of many (stained) proteins in the second dimension over the first dimension. That is, many proteins now lay neither on nor below but above the diagonal. This artefact, which is aggrevated by boiling, results from polypeptide self-aggregation. Following the first-dimension separation, all polypeptides of the same size (and hence identical polypeptides) will be stacked together at abnormally high local protein concentrations. When these proteins are analysed in the second dimension any dimers, trimers or higher oligomers formed and sustained following this close association will be revealed by their anomalous position in the second dimension. Closer examination of the original autoradiographs and the stained gels revealed that the majority of proteins (which lay on the diagonal) were in fact distributed about the diagonal. Minor differences in electrophoretic mobility between the non-reduced and reduced version of the same polypeptide account for this and may be due to the effect of intramolecular disulphide bonds upon protein folding and therefore binding of the dodecyl sulphate anion (Swank & Munkres, I97I). The F0 protein, for example, runs slightly ahead of its cleaved counterpart F1.2 under non-reduced conditions, perhaps indicating that FI,~ is free to fold more and bind less dodecyl sulphate than its uncleared precursor.

6 24 A. C. R. SAMSON, P. CHAMBERS AND J. H. DICKINSON Evidence for a new post-translational cleavage event within haemagglutinin-neuraminidase glycoprotein A most unexpected finding was the existence of two 14C-labelled polypeptides which lay beneath the diagonal in addition to the F1 and F2 polypeptides (Fig. l). These two polypeptides lay vertically beneath the virus-induced haemagglutinin-neuraminidase protein, HN, and were not found in uninfected celt extracts. Subsequent labelling experiments with 3H-glucosamine hydrochloride reveal that both HN1 and HN2 are glycosylated (not shown). The sum of the mol. wt. of these two disulphidelinked polypeptides is 79K, quite close to that determined for HN glycoprotein (75K). For these reasons the two polypeptides are considered to be fragments of the haemagglutinin protein and are dubbed HN~ (5I"5K) and HN2 (27"5K). It must be pointed out, however, that whereas F0 is very efficiently cleaved to Ft and F2, HNt and HN., together represent only approx. 2o % of the radioactive material found in HN protein after a 45 rain chase. The biological significance of this HN post-translational cleavage event is unclear; however, neither HN1 nor HNz are found in 35S-methionine labelled virions grown in ovo and analysed by the non-reduced: reduced gel technique (data not shown). Synchronous polypeptide initiation and labelling of polypeptides Proteins were radioactively labelled for short periods with ~H-leucine, following release from a high salt shock. A high salt level allows polypeptide chain elongation but prevents initiation; restoration to normal levels of salt after elongation has been completed allows synchronous re-initiation of polypeptide synthesis (Schochetman et al. 1977). The gradient of incorporated radioactive leucine along the length of each polypeptide is determined by the onset and duration of the pulse label post salt shock release. Proteins labelled during an early pulse are preferentially labelled towards the amino terminus, whereas proteins labelled following a later pulse are labelled more towards the carboxy terminus. Preliminary experiments had shown that the addition of 3oo mm-salt to normal growth medium for a period of 3o rain reduced the rate of protein synthesis to 1% of normal. This addition of salt for a 3o rain period served as the high salt shock in these experiments. Proteins were also labelled for [5 min with 3H-leucine in parallel with infected and uninfected monolayers which had not received any salt shock (unsynchronized controls). In addition, a parallel set of infected, unsynchronized monolayers was labelled for t5 min with 14C-leucine. All monolayers received a 45 rain chase period in the absence of isotope and in the presence of excess leucine before solubilization with SDS gel sample buffer. In dual-labelled samples the l~c-labelled proteins from infected monolayers served to (i) locate, following detection by autoradiography, the exact position of all 14C-labelled protein (and hence locate any 3H-labelled proteins which were co-electrophoresed) and (ii) provide an internal reference quantity of uniformly labelled polypeptides for 3H:14C ratio measurements. Inclusion of the 1~C material also obviated any problems arising from (i) gel to gel variations in solubility and electrophoretic mobility of proteins in either the first or the second dimension, (ii) gel to gel, and within gel variability in recovery of labelled proteins excised from the gel slab. Orientation of post-translational cleavage events within ND V glycoproteins Discs from 15 % acrylamide: DATD crosslinked dried down gel slabs were punched from regions corresponding to HN1, HN2, F1 and F2 polypeptides. The amounts of 3H and 14C isotopes in each disc were determined and 3H : 14C ratios for these polypeptides used together with their tool. wt. to calculate the fraction of 3H-isotope found in each of the pairs of

7 Cleavage off and HN glycoproteins in ND V (a) I I 1 I I I I I (b) ~_ "~ ~ ; HN1 ~HN2 0.2 r~ I I I! I I I I Duration of 3H-leucine pulse post salt shock (min) Fig. 2. Proportion of ah-leucine in (a) F1 polypeptide ( -- ), F2 polypeptide (A--A) and (b) HN1 polypeptide ( --O), HN2 polypeptide (~--A) calculated using ah:14c ratios and mol. wt. obtained from SDS gels. [Proportion of 3H-leucine for a polypeptide ~ = R~Mc~/Rc~Mc~+ RflM,8 where R~ and Rd = 3H: 14C ratio for polypeptides c~ and/? respectively, Ms and Md = mol. wt. of polypeptides a and/? respectively. Polypeptides a and l? refer to the pairs (a) F1 and F, and (b) HN1 and HN2.] Pointers F1, HN1, HN2 and F2 indicate proportions predicted for uniformly duallabelled (unsynchronized) polypeptides. Dotted line (... ) proportion found for unsynchronized (a) F1, (b) HN~. Dashed line (--) proportion found for synchronized (a) F2, (b) HNs. polypeptides (see legend to Fig. 2). These data were plotted as a function of the time and duration of the ah pulse, after release from the salt shock. It can be seen from Fig. 2 (a) that the proportion of 3H in polypeptide F 1 increases as the duration of the pulse increases and as the onset of the pulse is delayed. This shows that F1 is located towards the COOH terminus. A reciprocal relationship for F~ is necessarily found. In Fig. 2 (b), however, the proportion of ZH in polypeptide HN~ decreases as the duration of the pulse and the delay in the pulse increases. This shows that HN1 is located towards the NH2 terminus, and HN2 towards the COOH terminus. The dotted lines (... ) in Fig. 2 indicate the proportion of 3H label in the larger fragments (F1 or HN0 and the dashed lines (--) the smauer fragments (F 2 or HN2) found in unsynchronized (uniformly labelled) polypeptides. DISCUSSION To investigate the relative orientation of F1 and F2 polypeptides within F0 protein, a non-reduced/reduced two-dimensional SDS-PAGE system was used in conjunction with a salt shock procedure which synchronizes polypeptide chain initiation. Polypeptides such as F1 and F2 take up positions below the diagonal in these gels because during the first-dimension separation they are held together by one or more disulphide bonds; these bonds are cleaved using a reducing agent, allowing the polypeptides to migrate at a faster rate in the second dimension. F1 and F2 polypeptides are therefore separated from host and other host proteins which do not have intermolecular disulphide bonds. From the results presented here, the orientation of NDV fusion glycoproteins is as follows: NH~---F~--F1--COOH. Moreover, the sum of the mo[. wt. of F1 and F2 is consistent with the assumption that F1 and F2 are the only post-translational cleavage products off0 protein. The above sequence, which depends upon radioactive polarity labelling of polypeptides, is the same as that suggested by Scheid & Choppin (1977) for Sendal virus F proteins based upon N-terminal amino acid determinations.

8 26 A.C.R. SAMSON, P. CHAMBERS AND J. H. DICKINSON Given the above sequence for NDV fusion protein it would be expected that the F2 region would be the first to enter and proceed through the lipid bilayer followed by the F1 region (Katz et al. I977). A hydrophobic carboxy terminal region in F1 would remain embedded in the lipid bilayer and anchor the F 0 or F1.2 protein. This orientation together with recent amino acid sequence analysis of the N-terminus of F' 1 polypeptide from NDV (Scheid et al ) suggests a possible explanation for the requirement of the Fo ~ FI, ~ cleavage event for biological activity (fusion, haemolysis and infectivity) within the paramyxoviruses. At least the first six amino terminal residues of F~ polypeptides from Sendai, SV5 and NDV are hydrophobic. A hydrophobic tail for F1 could be important in the infection (and fusion) processes in that it may be capable of insertion into the plasma membranes of potential host cells. If F1 protein were synthesized de novo and not derived from the posttranslational cleavage event, this N-terminal hydrophobic region could become immediately embedded in the membrane vesicles of the infected host during protein synthesis. F 1 protein would eventually become anchored in the host plasma membrane by both amino- and carboxy-terminal hydrophobic residues. A hairpin conformation for FI in virions would leave no hydrophobic tail free to interact with other potential host plasma membranes. It is therefore suggested that a function of the F2 polypeptide is to mask the potential hydrophobic amino terminal region of F~ in the form of uncleaved precursor F0. This masking allows F1 (as part of F0) to pass through the lipid bilayer during protein synthesis, only the hydrophobie carboxy terminal region serving to anchor the protein in the bilayer. Once F2 and most of the F~ regions have passed through the bilayer, cleavage of F0 to FL 2 unmasks the hydrophobic N-terminal region of F1 now topologically equivalent in position to the outside of the virion. The continued linkage in virions of F2 to F~ by disulphide bonds may be fortuitous, or serve to hold Fz (and F~) in some active conformation on the surface of the virion. The entirely unexpected finding that NDV haemagglutinin-neuraminidase protein (HN) is also specifically cleaved, to some extent, in infected CEF cells to HN~ and HN2 polypeptides bears a striking similarity to the situation for haemagglutinin proteins HA1 and HA2 in influenza virus. From results described here, the following parallels between HN and HA proteins of NDV and influenza virus respectively can be drawn: sequences NHz--HNt--HN2--COOH and NH2--HA1--HA2--COOH; mol. wt. HNj (5I'5K), HN2 (z7"5 K) and HA~ (45 to 5oK), HA2 (25 to 3oK) (strain dependent); both pairs of polypeptides were glycosylated and derived from respective common precursor glycoprotein; both pairs of polypeptides are linked together by disulphide bonds. However, in NDV the cleaved products are not found in purified egg-grown virions. It has also been noted that in influenza virus (which does not possess a fusion protein), HA2 amino terminal residues bear a striking resemblance to F1 amino terminal residues in their sequence and hydrophobic nature (Gething et al. 1978; Waterfield et al. I979). The hydrophobic tail of HAs may again be important in the infection process for influenza virus (Lazarowitz & Choppin, I975). It is tempting to speculate that in paramyxoviruses this role in infection, which formerly resided in HN protein, has been superseded by F protein whereas HA protein in the orthomyxoviruses retains this role. REFERENCES GETHING, M. J., WHITE, J. M. & WATERrIELD, M. D. (I978). Purification of the fusion protein of Sendal virus: analysis of the NH2-terminal sequence generated during precursor activation. Proceedings of the National Academy of Sciences of the United States of America 75, IqOMMA, M. & OHOgm, i. 0973). Trypsin action on the growth of Sendai virus in tissue culture cells. IlL

9 Cleavage off and HN glycoproteins in ND V 2 7 Structural difference of Sendai viruses grown in eggs and tissue culture cells. Journal of Virology x2, I KAI~LAN, S. & BRATT, M. A. (1973). Synthesis and processing of Newcastle disease virus polypeptides. Abstracts of the American Society for Microbiology, pp KATZ, F. N., ROTHMAN, J. E., L1NGAPPA, V. R., BLOBEL, G. & LODISH, H. F. (1977). Membrane assembly in vitro: synthesis, glycosylation and asymmetric insertion of a transmembrane protein. Proceedings of the National Academy of Sciences of the United States of America 74, LAEMMLI, U. K. (197 D. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, London z27, LAZAROWITZ, S. G. & CHOPPIN, P. W. (I975). Enhancement of the infectivity of influenza A and B viruses by proteolytic cleavage of the hemagglutinin polypeptide. Virology 68, NAGAI, Y., KLENK, H.-D. & ROTT, R. 0976). Proteolytic cleavage of the viral glycoproteins and its significance for the virulence of Newcastle disease virus. Virology 72, SAMSON, A. C. R. & FOX, C. V. (I973). A precursor protein for Newcastle disease virus. Journal of Virology x2, SAMSON, A. C. R. & EOX, C. F. (I974). Selective inhibition of Newcastle disease virus-induced glycoprotein synthesis by D-glucosamine hydrochloride. Journal oj Virology I3, SCHEID, A. & CHOPI~IN, P. W. (1974). Identification of biological activities of paramyxovirus glycoproteins. Activation of cell fusion, haemolysis and infectivity by proteolytic cleavage of an inactive precursor protein of Sendai virus. Virology 57, o. SCHEID, A. & CHOPPIN, P. W. (1977). Two disulfide-linked polypeptide chains constitute the active F protein of paramyxoviruses. Virology 8o, SCHEID, A., GRAVES, M. C., SILVER, S. M. & CHOPPIN, P. W. (I978). Studies on the structure and functions of paramyxovirus glycoproteins. In Negative Strand Viruses and the Host Cell, pp. 18I. Edited by B. W. J. Mahy & R. D. Barry. London: Academic Press. SCHOCHETMAN, G., OROSZLAN, S., ARTHUR, L. & FINE, D. (1977). Gene order of the mouse mammary tumor virus glycoproteins. Virology 83, SWANK, R. "r. & MUNKRES, K.D. 0971)- Molecular weight analysis of oligopeptides by electrophoresis in polyacrylamide gel with sodium dodecyl sulphate. Analytical Biochemistry 39, TABER, R., REKOSH, D. & BALTIMORE, D. (I971). Effect of pactamycin on synthesis of poliovirus proteins: a method for genetic mapping. Journal of Virology 8, 395-4oi. WATERFIELD, M. D., ESPELIE, K., ELDER, K. & SKEHEL, J. J. (I 979). Structure of the haemagglutinin of influenza virus. British Medical Bulletin 35, (Received 28 June I979)