Eur. J. Biochem. 203, 65 - 73 (1 992) C FEBS 1992

Localization of the intrachain disulfide bonds of the envelope glycoprotein 71 from Friend murine leukemia virus Monica LINDERl, Dietmar LINDER ’, Josef HAHNEN ’, Hans-Helming SCHOTT’ and Stephan STIRM’ I

Biochemisches Institut am Klinikum der Justus-Liebig-Universitat, Giessen, Federal Republic of Germany Institut fur Mikrobiologie und Molekularbiologie der Justus-Liebig-Universitiit, Giessen, Federal Republic of Germany

(Received July 11, 1991)

-

EJB 91 0902

Envelope glycoprotein 71 from Friend murine leukemia virus was purified to homogeneity by reversed-phase HPLC. It could be shown that all 20 cysteine residues of the molecule are linked by disulfide bonds. After complete tryptic digestion, peptides containing cystine were identified by comparison of the reversed-phase HPLC profile of the digest with that of a reduced aliquot which had been subjected to affinity chromatography on thiol-Sepharose. The locations of the 10 disulfide bonds were determined by isolation, further digestion and analysis of peptides containing cystine. The first cysteine residue of the sequence (Cys46) was shown to be coupled to the sixth (Cys98), leading to a large loop containing four additional cysteine residues. Computer model building and energy calculations led to the assignment of Cys72 to Cys87 and Cys73 to Cys83. The following four cysteine residues of the sequence also constitute a structural unit, with Cysl21 bonded to Cysl41 and Cys133 to Cys146, and the last two cysteine residues in the amino-terminal domain of glycoprotein 71 form a small loop (Cys178 to Cysl84). The first two cysteine residues of the carboxy-terminal domain produce a very small hydrophobic loop (Cys312-Cys315). Cys361 is bound to Cys373, Cys342 to Cys396 and Cys403 to Cys416. A model for the folding pattern of the viral glycoprotein is proposed. The spikes on the surface of the ecotropic Friend murine leukemia virions (F-MuLV) are composed of glycoprotein 71 (gp71) and the protein p15(E), both formed by proteolysis of a common precursor, gPr90, which is the glycosylated envelope ( e m ) gene product. Whereas p15(E) has been found to be anchored in the viral membrane (Pinter, 2989), the precise nature of the attachment of gp71 is not known. The viral glycoproteins mediate binding of virions to host cell receptors and penetration into the cell. They determine host cell specificity and play an important role in viral pathogenesis (Kabat, 1989). The env gene products of the Friend, Rauscher and Moloney murine leukemia viruses comprise a family of glycoproteins which are structurally and serologically related, but retain unique elements which provide the basis for discrimination between viral subgroups (Elder et al., 1977; Pinter and Honnen, 1984; Vogt et al., 1986; Trauger and Luftig, 1989). Comparison of amino acid sequences of these glycoproteins demonstrates the domain structure of the molecules, i.e. a differential domain at the amino-terminus, which has been shown to be responsible for the particular host ranges of these viruses (Vogt et al., 1986), a proline-rich hypervariable

region in the middle of the molecules and a constant domain at the carboxy-terminus (Koch et al., 1984). The positions of the cysteine residues are highly conserved, suggesting that they play an important role in the three-dimensional structure of the molecules. Furthermore, deletions in the amino-terminal domains of glycoproteins from dualtropic viruses, compared with those from ecotropic viruses, lead to the loss of a number of cysteine residues, indicating that structures stabilized by disulfide bonds could play a role in determining host cell specificity and pathogenesis. The primary structure of gp71 from F-MuLV has been studied extensively. In addition to the amino acid sequence (Chen, 1982; Koch et al., 1983), the structures of the carbohydrate moieties at each glycosylation site have been determined (Geyer et al., 1990). Precise data on the three-dimensional structure of this molecule, however, are not yet available. As a first step toward elucidation of the conformation of gp71 from F-MuLV, we have investigated the bonding state of its 20 cysteine residues and determined the positions of the disulfide bonds.

____

MATERIALS AND METHODS

Correspondence to S. Stirm, Biochemisches Institut am Klinikum der Juslus-Liebig-Universitat, Friedrichstrafle 24, W-6300 Giessen, Federal Republic of Germany Ahhrrviation.s. F-MuLV, Friend murine leukemia virus; F-MCF, Friend mink ccll focus-inducing virus; gp, glycoprotein. Enzymes. Chymotrypsin (EC 3.4.21 .I); endoproteinase Glu-C from Stuphylococcus uureus V8 (EC 3.4.21.19); proline-specific endopeptidasc from Flavohacterium meningosepticurn (EC 3.4.21.26); trypsin (EC 3.4.21.4). Note. These data are available from EMBL file server, accession number PO 3395 (Swissprot).

Materials Dulbecco’s modified Eagle’s medium was from Gibco, Paisley, Great Britain. Acetonitrile and 2-propano1, both HPLC grade, and scintillation cocktail were from Roth, Karlsruhe, FRG. Trifluoroacetic acid, sequencer grade, was from Rathburn, Walkerburn, Scotland. Trypsin from bovine pancreas, chymotrypsin from bovine pancreas and endoproteinase Glu-C from Staphylococcus aureus V8, all sequencing grade, were from Boehringer, Mannheim, FRG.

66 Proline-specific endopeptidase from Fluvobucterium meningo.septicurn was obtained from ICN ImmunoBiologicals, Costa Mesa, CA, USA. Activated thiol-Sepharose 4B was from Pharmacia, Freiburg, FRG, and [14C]iodoacetic acid from Amersham, Braunschweig, FRG. Solvents and reagents for sequencing were from Applied Biosystems, Foster City, CA, USA. Water purified by a Milli-Q Ultra-pure water system, Millipore, Eschborn, FRG, was used throughout, and all other chemicals used were of analytical grade. Virus The helper-independent component of the F-MuLV complex (Troxler et al., 1980), as produced by Eveline cells (Seifert et al., 1975), was used for the present studies. Eveline cells, obtained from W. Schdfer (Max-Planck-Institut fur Virusforschung, Tubingen, FRG), were propagated in suspension culture in Dulbecco's modified Eagle's medium, supplemented with 10% (by vol.) complement-inactivated (30 min, 56°C) fetal bovine serum. Virus particles were harvested by differential centrifugation (Moennig et al., 1974) and washed by repeated pelletting and resuspension. Viral glycoproteins The viral glycoproteins were released from the virus particles by freezing and thawing and were purified by chromatography through phosphocellulose, as described previously (Linder et al., 1982). The glycoprotein fraction containing predominantly gp71 from F-MuLV and glycoprotein 70 (gp70) from Friend mink cell focus-inducing virus (FMCF) was further purified by reversed-phase HPLC (Waters, Eschborn, FRG) on a Bakerbond column (Baker, GroI3Gerau, FRG). Enzymatic digests For tryptic digestion, gp71 was dissolved (1 mg/ml) in 0.1 M Tris/HCl, pH 8.5, containing 10% (by vol.) acetonitrile, and 2 - 5 pg trypsin/100 pg protein was added. Incubation was carried out under argon at 25 - 27°C for 45 h, with addition of fresh trypsin every 15 h. Chymotryptic digestion of peptides was carried out in 100 mM Tris/HCl buffer, pH 7.8, with 0.10 mM CaClz and 5% (by vol.) acetonitrile added. 2-10 pg peptide in 100 p1 buffer was incubated with 0.2 - 1.O bg chyinotrypsin under argon for 20-48 h a t 25-27°C. Digestion of peptides with endoproteinase Glu-C was carried out in 25 mM sodium phosphate buffer, pH 7.8, with 5% (by vol.) acetonitrile. 2- 10 pg peptide was dissolved in 100 pl buffer and incubated under argon with 0.1 -0.5 pg endoproteinase Glu-C for 8 - 10 h at 25 - 27 "C. Proline-specific endopeptidase was dissolved, 0.15 pg (5.25 mU)/pl, in 25 mM sodium phosphate buffer, pH 7.0. 1 - 2 pg peptide in 25 pl of the same buffer was incubated with 15 p1 of the enzyme solution under argon for 40 min at 37 "C.

ration system (Applied Biosystems, Foster City, CA, USA). For details see supplement. Reduction and affinity chromatography For reduction of an aliquot (100 pl) of the tryptic digest, trypsin was first inhibited with 0.5 p1 0.1 M phenylmethylsulfonyl fluoride in 2-propanol. After addition of 0.5 pl 2-mercaptoethanol, the sample was incubated at room temperature for 2 h, lyophilized, then redissolved in water and lyophilized twice more before it was dissolved in 0.1 M Tris/ HC1, pH 8.0, containing 1 mM EDTA (coupling buffer) for affinity chromatography. 300 mg activated thiol-Sepharose 4B was swollen and washed as recommended by the manufacturer and equilibrated with coupling buffer on a sintered-glass filter of I cm diameter, in a glass column. 100 p1 of the tryptic digest, reduced or non-reduced, in 1 ml coupling buffer was added, argon was blown over, and the column was sealed and rotated end over end for 20 h at 30°C. The filtrate was then aspirated, and the gel was washed with several column volumes of 0.2 M aqueous ammonium acetate, pH 8.6, containing 1 mM EDTA. A wash with 1 M NaCl was included to elute nonspecifically adsorbed peptides (Egorov et al., 1975). All buffers were purged with argon immediately prior to use. Combined filtrates were lyophilized, redissolved by adding 500 p1 water and injected in portions onto the reversed-phase HPLC column (see above), with buffer running at initial conditions until the salts had been washed through. Radiocarboxymethylation For carboxymethylation, 0.5 - 2.0 mg glycoprotein was dissolved (1 mg/ml) in 0.1 M Tris/HCl, pH 8.5, containing 6 M guanidine hydrochloride, previously purged with argon. 20 p1 50 mM ['4C]iodoacetic acid (50 Bq/nmol)/ml protein solution was added, argon was blown over and the tube was sealed and incubated in the dark at 37°C for 1-2 h with shaking (Allen, 1981). Alternatively, ['4C]iodoacetic acid (1 mM) was included in the tryptic digest of the glycoprotein. Samples were tested for incorporation of radioactivity by fluorography after SDSjPAGE (Laemmli, 1970) or by amino acid analysis (see below). In the latter case, fractions eluting from the analyzer at the position of carboxymethylcysteine were collected, and, after addition of scintillation cocktail, monitored for radioactivity with a model 4450 liquid scintillation counter (Packard, Downers Grove, IL, USA). Alkaline hydrolysis to release fatty acids 0.5 mg lyophilized glycoprotein was suspended in 2 ml 0.1 M KOH in methanol and incubated for 20 min at 22°C to release fatty acids (Schmidt et al., 1979). After neutralization with 0.1 M HC1, the sample was lyophilized then treated with [14C]iodoaceticacid as described above. Protein and peptide sequencing

Isolation of peptides Tryptic peptides were isolated by reversed-phase HPLC (Waters, Eschborn, FRG), employing a column filled with Hypersil (Shandon, Astmoor, UK). Peptides were further purified by chromatography through a microbore column (Vydac Hesperia, CA, USA) in a model 130A integrated sepa-

Proteins and peptides (20- 100 pmol) were sequenced by Edman degradation on an Applied Biosystems (Foster City, CA, USA) pulsed-liquid-phase sequencer, model 477 A, using the standard protocol recommended by the manufacturer (normal-I). Phenylthiohydantoin derivatives of amino acids were identified by an on-line analyzer, model 120 A (Applied Biosystems) with a repetitive yield of 92 - 95%.

67 b

k Da 2.0

100

3.

94 --c h

670

Zi.0

4 I

N

___-

0.0

0

0 7 -

1

20

10

30

30-

14-

40

Elution time ( m i d Fig. 1. Reversed-phase HPLC purification of F-MuLV gp71. (a) The glycoprotcin fraction obtained by chromatography on phosphocellulose was applied to a Bakerbond Wide-Pore C4 (5 pm) column (4.6 mm x 250 mm), and eluted in 0.1 % (by vol.) aqueous trifluoroacetic acid with a 2-propanol gradient (20-60% in 48 min) at 40°C with a flow rate of 1 ml/min. Proteins were monitored at 220 nm and peak fractions werc collected semiautomatically. Shaded fractions were pooled and used for the present studies. (b) SDSjPAGE of successive fractions indicated in (a).

Amino acid analysis

RESULTS AND DISCUSSION

Determination of amino acid composition and hexosamine content was performed on a Biotronik (Frankfurt, FRG) model LC 6001 analyzer, using o-phthalaldehyde as a postcolumn colouring agent. Prior to analysis, samples (0.5 5.0 pg) were hydrolyzed with 6 M aqueous HCI, containing 0.02% (by vol.) 2-mercaptoethanol, for 24 h at 110°C in evacuated and sealed tubes.

Purification of gp71 gp71 from F-MuLV was separated from gp70 from FMCF by reversed-phase HPLC on wide-pore material as shown in Fig. 1a. The glycoproteins were identified by SDS/ PAGE, Fig. 1b, and their identities confirmed by amino-terminal sequence analysis and comparison with the sequences found by Chen (1982) and Koch et al. (1983,1984). The purity of gp71 was shown by sequence analysis to be > 95%.

Computational procedures Random structures with defined disulfide bond patterns were generated using the distance geometry algorithm. The program package DG (de Vlieg et al., 1988) was used, and the bond lengths between sulfur atoms were incorporated as constraints. Dihedral angles (n))were set to 180" in order to obtain tram configuration for all peptide bonds. The generated geometries were optimized by performing a short (1 ps) molecular dynamics run at 300 K. The program package CHARMM was used for energy calculations (Brooks et al., 1983). Energy minimizations were performed in vucuo, and only polar hydrogens were treated explicitly, according to the United Atom Approach (McCammon et al., 1977). Conformational search was performed using the Boltzinann Jump procedure, which is implemented in the QUANTA molecular-modeling software (Polygen Corporation, Waltham, MA, USA). This procedure uses the Metropolis Algorithm to explore conformational space for local minima (Kirkpatrick et al., 1983). The temperature of the system was set to 300 K. In each step of the search, the structure was perturbed by random change of dihedral angles within a window of 10". Structures with a root mean square which differed from the initial structure by more than 0.5" in torsion space were sampled for further energy minimization. The search was stopped after 10 minimization steps.

Bonding state of the cysteine residues In order to detect free sulfhydryl groups, the non-reduced, denatured (6 M guanidine hydrochloride) glycoprotein was treated with ['"Cliodoacetic acid, according to the normal procedure for carboxymethylation, and an aliquot was subjected to amino acid analysis. No ['4C]carboxymethylcysteine could be detected, although, under the conditions used, about 10 Bq would have been expected for each free sulfhydryl group in the molecule. To rule out the presence of inaccessible sulfhydryl groups, ['4C]iodoacetic acid was added to a tryptic digest of the glycoprotein, and the incubation was repeated with fresh trypsin. Again no incorporation of radioactivity could be detected, further demonstrating that no disulfide exchange takes place under the conditions of enzymatic digestion. Viral glycoproteins often contain fatty acids which may be linked to cysteine as thio esters (Grand, 1989), e.g. in the case of glycoprotein G from vesicular stomatitis virus (Rose et al., 1984). We investigated this possibility by subjecting the glycoprotein to alkaline hydrolysis prior to radioactive carboxymethylation. Subsequent fluorography demonstrated that no radioactivity was incorporated into gp71. We thus conclude that all cysteine residues of the molecule are linked by disulfide bonds.

68 Table 1. Amino-terminal sequences of cystine-linked tryptic peptides from F-MuLV gp71. Double or triple sequences were detected parallel in equimolar amounts. (*) Cysteine, according to the sequence in Fig. 2. (+) Clycosylated asparagine, according to Geyer et al. (1990).

46

N H P L W T W W P V L T P D L C M L A L H G P P H W G L E Y 40

50

60

Peptide 7 2 73

83

Q A P Y S S P P G P P C C S G S G G S S P G C S R D C N E P 70

80

98

90

1 T2

T1.l

L T S L T P R C N T A W N R L K L D Q 110 V T H K I S S E G F Y1V 20

ITk3

loo 141 146 C P G S H R P R E A K S C G G P D S F Y C A S W G C E T T G

12 1

130

140

150

T.1.2

178

160

N K W C N P L A I O F T N A G R Q V T S W I T G H Y W G L R 200

230

240

S S E G F Y V * P G . .. . S*GGPDSPY*. . . .

T3

P S S S W D Y . .. . W*NPLAI... .

T4

TQE*WL*LVSGPPYYEGVAV . . . GSYYLVAPAGTTWA*NTGLT . . . TTDY*VLVELWPR

T5

GLXIGTVPK THQAL*+TTLK

5

P V L A D Q L S F P L P N P L P K P A K S P P V S / N S T l P T 250 260 270

5

M I S P S P T P T Q P P P A G T G D R L L N L V Q G A Y Q A

zao I ~4

290

10

T2.1lT2.2

210

L Y V S G Q D P G L T F G I R L K Y Q N L G P R V P I G P N 220

5 20 ETVWAISGNHPLWT . . . . D*NEPLTSLTPR *NTAWNR 5 10 ETVWAISGNHPLWT . . . . *NTAWNR 5

IT:~4a 190

Amino acid identified

87

300

10

15

20

5

G T T W A C N T G L T P C L S A T V L N R T T D Y C V L V E

L W P R V T Y H P P S Y V Y S Q F E K S Y R H K R 430

440

Fig. 2. Amino acid sequence of gp71 from F-MuLV according to Chen (1982). Tryptic peptides which were analyzed (TI -T5) are underlined, and numbers above the cysteine residues denote their position in the sequence. N-Glycosylation sites are enclosed in boxes (cf. Geyer ct al., 1990). (a) The amino acid at position 184 was identified as glycine by Chen. Our results show that this amino acid is cysteine, in accordance with the sequence found by Koch et al. (1983) for a molecularly cloned F-MuLV. All other amino acids detected by sequence analyses in the course of our studies were i n accordance with the sequence found by Chen.

Detection of tryptic peptides containing cystine An aliquot of a tryptic digest of gp71 was reduced with 2inercaptoethanol and subjected to affinity chromatography on thiol-Sepharose, which selectively binds free sulfhydryl groups. The reversed-phase HPLC elution profile of the filtrate was compared to that of the non-reduced digest. Results are summarized in Fig. S1. As expected, the elution profile of the filtrate of a non-reduced aliquot, which had been subjected to affinity chromatography, was identical to that of the untreated tryptic digest. Fractions (shaded in Fig. S1 a) of the non-reduced samples containing peptides TI - T5, lacking in the reduced sample, were isolated and further purified for the following investigations. Identification of peptides linked by cystine Peptides T I . Results of amino acid analysis of peptides T1.l and T1.2 (Fig. Sl) are given in Table S1. It is shown that both peptides have the same amino acid composition; both

contain amino acids 23 - 104 of the sequence of gp71 (Fig. 2). Amino-terminal sequence analysis (Table 1) revealed that T1.l had three amino-termini, beginning at amino acids 23, 86 and 98, whereas T1.2 had not been cleaved at Arg85 and hence showed only two amino-termini. T1.l and T1.2 were pooled, further cleaved with endoproteinase Glu-C and rechromatographed by microbore reversed-phase HPLC at pH 2. The elution profile obtained and the results of sequence analysis of the individual fractions are shown in Fig. S2. The equimolar parallel sequences found in fraction TI.I.I/T1.2.1 demonstrate that Cys46 and Cys98 are linked, leading to a large loop containing four additional cysteine residues. The peptides T1.1.2 and T1.2.2 (amino acids 60-97) with (Tl.I.2) and without (T1.2.2) cleavage at Arg85 were each further digested with proline-specific peptidase and rechromatographed (Fig. S3). Peak fractions were identified by sequence analysis, the results of which are included in Fig. S3. Parallel sequences found for peptide T1.1.2.1 (Fig. S3a) show that the cysteine residues at positions 83 and 87 are each bonded to one of the two adjacent cysteine residues at positions 72 and 73. This conclusion was confirmed by the parallel sequences found for peptide TI .2.2.1, shown in Fig. S3 b. Since further clarification was not possible by proteolytic methods, we applied computer model building and energy calculations to determine the more likely bonding pattern. It has been shown, through reduction/reoxidation experiments, that globular proteins are completely reconstituted to the most stable state with the same disulfide bond orientation as the native protein (Jaenicke, 1987). Relatively small differences in free energy suffice to guide the chain to a unique pattern of disulfide bonds (Pain, 1987). We thus assume that the correct orientation of the two disulfide bonds is that which leads to the most stable structure. A linear peptide comprising residues 69 - 90 of the sequence of gp71 was built by molecular modelling and folded, using a combined procedure of distance ge-

69

Fig. 3. Schematic representationof gp71 from F-MuLV showing disulfide bonds (this paper) and 0-glycosylation and N-glycosylation sites (Geyer et al., 1990). Black lines linking cysteines indicate disulfide bonds. Number in circles are glycosylation sites, the numbers indicate positions of glycosylated amino acids in the sequence of gp71. Glycosylation was found to be heterogeneous. Oligomannosidic oligosaccharides were found at Asn168, Asn334/341 and Asn410, hybrid species at Asnl2 and Asn3341341, and N-acetyllactosaminic glycans at Asn266, Asn302, Asn334/ 341, Asn314 and Asn410. Thr268, Thr 211, Thr279 and Thr3041309, as well as Ser273 and Ser275, were found to be 0-glycosylated. For details of oligosaccharide structures, see Gcyer et al. (1990).

ometry calculation and molecular dynamics simulation (de Vlieg et al., 1988). For each of the two alternative bonding patterns, (a) Cys72 - Cys83 and Cys73 - Cys87, and (b) Cys72- Cys87 and Cys73-Cys83, a set of 20 random structures was generated by folding the chain to attain a distance of 0.19-0.21 nm between the sulfur atoms of the disulfide bonds in question. Each of the 40 random structures was then further energyminimized using the CHARMM program. The mean potential energy for structures with bonding pattern (a) (-208 kJ/ mol) is slightly higher than that for the alternative (b) (- 236 kJ/mol), indicating that peptides with bonding pattern (b) are, on the average, less constrained than peptides with bonding pattern (a). In addition, the structure with the lowest potential energy found had bonding pattern (b), which also Favours this alternative. In order to elucidate this tendency, the best structure of each set was subjected to a conformational search, using the Boltzmann Jump procedure, which permits upward jumps in energy, overcoming energy barriers in the range of k . T (k, Boltzmann constant), and further explores conforinatioiial space for lower minima (Kirkpatrick et al., 1983). The search procedure reached a potential energy of -433 kJ/mol for alternative (a) and - 536 kJ/mol for alternative (b). Bearing in

mind that these calculations comprise rough approximations, they lead us to the conclusion that bonding pattern (b), Cys73 - Cys83 and Cys72 - Cys87, tends to the formation of lower-energy conformations and thus appears more likely than the other alternative. Peptides T2. The tryptic peptides T2.1 and T2.2 (Fig. S1) were each rechromatographed at pH 2 and identified by amino acid analysis (Table S l ) and sequence analysis (Table 1). T2.1 was found to consist of amino acids 114- 128 and 132-151 of the gp71 sequence, and T2.2 was shown to differ from T2.1 in that it had been further cleaved at Arg126. T2.1 and T2.2 were each further digested with chymotrypsin and rechromatographed. The reversed-phase HPLC elution profiles obtained, and the results of sequence analysis of the peak fractions, are shown in Fig. S4. Sequences found for peptide T2.1.1 (Fig. S4a) demonstrate that Cysl21 is linked to Cys141, and analysis of T2.1.2 (Fig. S4a) demonstrates that Cys133 is bonded to Cys146. Sequence analysis of peptides T2.2.1 and T2.2.2 (Fig. S4b) led to the same conclusions. Peptide T3. The tryptic peptide T3 (Fig. S l ) was repurified by inicrobore reversed-phase HPLC at pH 2. Amino acid analysis (Table S1) and sequence analysis (Table 1) showed that it contains amino acids 156-179 and 183-196, thus demonstrating that Cys178 is bonded to Cysl84.

70 Peptide T4. T4 (Fig. S l ) was isolated by rechromatography at pH 2 on wide-pore material and pooled from a number of fractions of the tryptic digest separation. It was shown by amino acid analysis (Table S l ) and sequence analysis (Table 1) to be homogeneous with respect to amino acid composition and amino-terminal sequence, and was found to comprise amino acids 309 - 349, 382 - 41 1 and 41 2 -424 of the gp71 sequence, including three glycosylation sites (cf. Geyer et al., 1990). Digestion of T4 with chymotrypsin, reversed-phase HPLC and sequence analysis of the individual fractions led to the results shown in Fig. S5. Peptide T4.1 contains only two cysteine residues, Cys312 and Cys315, demonstrating their linkage. Parallel sequences found for peptides T4.2 and T4.3 demonstrate that Cys342 is bonded to Cys396 and Cys403 is bonded to Cys416. Peptide T5. Glycopeptide T5 (Fig. S l ) was isolated from several fractions from the tryptic digest separation by rechromatography on microbore reversed-phase HPLC in trifluoroacetic acid. Amino acid analysis (Table S1) and sequence analysis (Table 1) confirmed that it comprises amino acids 359-367 and 368 - 378 of the gp71 sequence, thus demonstrating that Cys361 is bonded to Cys373. Conclusions A schematic representation of the folding pattern of gp71, based on its disulfide bonding, is presented in Fig. 3, which also takes into account the organization of the molecule into domains. It can be seen that there are no disulfide bonds between the amino- and carboxy-terminal domains, in accordance with previously reported findings (Pinter and Honnen, 1984). Comparison with the folding pattern of other env gene products, particularly with gp70 from the dualtropic F-MCF could lead to suggestions for the correlation of particular structural elements of the molecule with its known functions, such as determination of host range and pathogenesis. For this reason, we are currently investigating the disulfide linkages of PP70. The disulfide-bonding pattern of another retroviral envelope glycoprotein, gp120 from the spikes of the human immunodeficiency virus 1 has previously been determined (Leonard et al., 1990). All 18 cysteine residues in gp120 from this virus were also found to be involved in intramolecular disulfide bonds, with no linkage between the amino-terminal and carboxy-terminal regions of the molecule, showing a similarity to the results presented here. The spikes on the surface of F-MuLV are composed of gp71 and pl5(E), a transmembrane protein. It has been reported that gp71 is attached, at least partially, by means of disulfide bonds to pl5(E), to form a complex with an apparent molecular mass of 85-95 kDa (Gliniak and Kabat, 1989; Pinter, 1989; Trauger and Luftig, 1989). In some cases, however, only about 10% of the glycoprotein was found to be covalently linked to pl5(E) (Pinter et al., 1982), and the gp85 band found by Trauger and Luftig (1989) in SDSjPAGE under non-reducing conditions was not present in all virus preparations. Kilpatrick et al. (1989) found no disulfide-linked gp70-p15(E) complexes at all and concluded that these two proteins are associated by weak, non-covalent interaction. Our finding that all cysteine residues of gp71 are linked intramalecularly demonstrate that the glycoprotein which we separated from virus particles after freezing and thawing, with no reducing agent present throughout the entire separation

and purification procedure, is not linked to pl5(E) by covalent bonds. The exact nature of the association of gp71 with pl5(E) thus remains to be elucidated. It is perhaps also noteworthy that, as opposed to gp52 (gp55), the envelope glycoprotein from the replication-defective recombinant Friend spleen focus-forming virus, which has been shown to form disulfide-bonded oligomers as a prerequisite for processing, gp70 from the ecotropic Friend virus, does not oligomerize (Gliniak and Kabat, 1989; Kilpatrick et al., 1989). We thank Michael Dreisbach and Hans-GiinterWelker for excellent technical assistance. This project was supported by the Deutsche ~orschungsgemein.rchaft(SFB 272, TP 21 and 22) and was performed by M. L. in partial fulfillment of the requirements for the degree of Dr rer. nat. at Giessen University.

REFERENCES Allen, G. ( 3 981) Laboratory techniques in biochemistry and molecular biology (Work, T. S. & Burdon, R. H., eds) vol. 9, Elsevier,

Amsterdam. Brooks, B. R., Bruccoleri, R. E., Olafson, B. D., States, D. J., Swaminathian, S. & Karplus, M . (1983) J . Comp. Chem. 4, 187217. Chen, R. (1 982) Proc. Natl Acud. Sci.U S A 79, 5788 - 5792. De Vlieg, J., Scheek, R. M., van Gunsteren, W. F., Berendsen, H. J., Kaptein, R. & Thomasson, J. (1988) Proteins 3, 209-219. Egorov, T. A., Svenson, A., Ryden, L. & Carlsson, J. (1975) Proc. Natl Acad. Sci. U S A 72, 3029 - 3033. Elder, J. H., Jensen, F. C., Bryant, M. L. & Lerner, R. A. (1977) Nature 267, 23 -28. Geyer, R., Dabrowski, J., Dabrowski, U., Linder, D., Schliiter, M . , Schott, H.-H. & Stirm, S. (1990) Eur. J. Biochem. 187, 95-110. Gliniak, B. C. & Kabat, D. (1989) J . Virol. 63, 3561 -3568. Grand, R. J. A . (1 989) Biochem. J . 258,625 - 638. Jaenicke, R. (1987) Prug. Biophys. Ma/. Biol. 49, 117 - 237. Kabat, D. (1989) Curr. Top. Microbiol. Irnmunol. 148, 1-42. Kilpatrick, D. R., Srinivas, R. V. & Compans, R. W. (1989) J . Biol. Chem. 264,10732- 10737. Kirkpatrick, S., Geilat, C. D., J r & Vecchi, M. P. (1983) Science 220, 671 -680. Koch, W., Hunsmann, G. & Friedrich, R. (1983) J . Virol. 45, 1-9. Koch, W., Zimmermann, W., Oliff,A. & Friedrich, R. (1984)J. Virol. 49, 828 - 840. Laemmli, U. K. (1970) Nature 227, 680-685. Leonard, C. K., Spellman, M. W., Riddle, L., Harris, R. J., Thomas, J. N. &Gregory, T. J. (1990) J . B i d . Chem. 265, 10373-10382. Linder, D., Stirm, S., Schneider, J., Hunsmann, G., Smythers, G. & Oroszlan, S. (1982) J . Virol. 42, 352-355. McCammon, J. A , , Gelin, B. R. & Karplus, M . (1977) Nature 267, 585 - 590. Moennig, V., Frank, H., Hunsmann, G., Schneider, J. & Schafer, W. (1974) Virology61, 100-111. Pain, R. (1987) Trends Biochem. Sci. 12, 309-312. Pinter, A. (1989) in Retroviruses and disease (Hanafusa, H., Pinter, A. & Pullman, M. E., eds) pp. 20 - 39, Academic Press, New York. Pinter, A. & Honnen, W. J. (1984) J . Virol. 49,452-458. Pintcr, A., Honnen, W. J., Tung, J.-S., O’Donnell, P. V. & Hammerling, U. (1982) Virology 116, 499-516. Rose, J . K., Adams, G. A. & Gallione, C . J. (1984) Proc. Natl Acad. Sci. U S A 81, 2050-2054. Schmidt, M. F. G., Bracha, M. & Schlesinger, M. J. (1979) Proc. Natl Acad. Sci. U S A 76, 1687-1691. Seifert, E., Claviez, M., Frank, H., Hunsmann, G., Schwarz, H. & Schiifer, W. (1975) 2. Naturforsch. 30c, 698-700. Trauger, R. & Luftig, R. B. (1989) Intervirology 30, 137 - 147. Troxler, D. H., Ruscetti, S. K . & Scolnick, E. M. (1980) Biochim. Biophys. Acta 605, 305 - 324. Vogt, M., Haggblom, C., Swift, S. & Haas, M. (1986) Virology 154, 420-424.

71

Supplementary material to:

Localization Of the intrachain disulfide bonds of the envelope glycoprotein71 from Friend murine leukemia virus Monica LINDER, Dietmar LINDER, Josef HAHNEN, Hans-Hcnning SCHOTT and Stephan STIRM

a

0

2 0.5

a

-

0

90

60

30

Elution time (min) Fig. S1. Reversed-phase HPLC elution profiles of a tryptic digest of F-MuLV gp71 and the same digest after reduction and affinity chromatography on thiol-Sepharose. Peptides were separated on a Hypersil3 pm ODS column (4.6 inm x 250 mm) with a linear acetonitrile gradient (0 - 50% for 90 min) in 25 mM aqueous ammonium acetatc, pH 6.0, at 60°C with a flow rate of 1 ml/min. Peptides were monitored at 220 nm and collected semiautomatically. (a) Elution profile of the tryptic digest and (b) elution profile obtained after reduction of the tryptic digest with 2-mercaptoethanol and affinity chromatography. When a non-reduced aliquot of the digest was subjected to affinity chromatography, a profile identical to (a) was obtained. Fractions shaded in (a) contain peptides TI -T5 (not present in the reduced and cysteinc-peptide-free sample) which were pooled and purified for further analysis.

I

0

10

20

30

40

Elution time ( m i d Fig. S2. Elution profile and sequence data obtained after digestion of peptides TI.1 and T1.2 with endoproteinase Glu-C. Revcrsed-phase HPLC was carried out through a Vydac C4, 30-nm microbore column (2.1 mm x 250 mm) using 0.1 YO(by vol.) aqueous trifluoroacetic acid with an acetonitrile gradient (0 -60% in 45 min) and a flow rate of 200 pl/min at 45°C. Peptides werc monitored at 220 nm and fractions were collected manually. (*) Cysteine, according to the scquence in Fig. 2 (see main tcxt).

72 0.2

100

r -

1

100

0.2

2.1.1 V'PGSHRPR *ASW

GPP~~SOSGGSSP G*SR D*NEP

h

s .-WL -

h

s a, .-

VOAP-

Y

v

.z 4L

50

0 4a,

!I

N 0

"0.1

Q

S'CGPDSF GtEiTGR

5 0 .Z

c

J1,1.

2

1___ 0

T2.1.2

0

4-

a,

2

SSEGF-

c--

7 --

0

10

0

0

0

10

0

20

Elution time (min)

20

Elution time (min) 100

0.2

0.1

-T2.2.1 VIPOSHR

.*sw T1.2.2.1 GPPtkSGSGGSSP G*SRD*NEP

h

8

v

-T1.2.2.2

a, .-

GPP+rSGSGG

.= L

0 (Y

" 0.1

50

a

S

0 a,

-T 2 . 2 . 2 SSGOPDSF

0 N

GtETTGR

" 0.05 6

+-'

8 :.

+ SS EGr?

0

0 0

10

20

Elution time ( m i d Fig. S3. Elution profiles and sequence data obtained after digestion of peptides T1.1.2 (a) and T1.2.2 (b) with proline-specific peptidase. Conditions for reversed-phase HPLC were the same as in Fig. S2. (*) Cysteine, according to the sequence in Fig. 2 (see main text).

0 0

10

20

Elution time ( m i d Fig. S4. Elution profiles and sequence data obtained after digestion of peptides T2.1 (a) and T2.2 (b) with chymotrypsin. Conditions for revcrsed-phase HPLC were the same as in Fig. S2. (*) Cysteine, according to the sequcnce in Fig. 2 (sce main text).

73 Table Sl. Amino acid composition of tryptic peptides from F-MuLV gp71. (a) Uncorrected valucs as determined by amino acid analysis. (b) Values derived from amino acid sequence (Chen, 1982). n.d.. not determined; (+), not quantified.

Amino acid content of peptide T1 .I

_-___ Amino

TI .2

T2.1

(23- 104)

(23 - 104)

a

b

a

b

6 6 9 4 8 4 2 1 1 8 2 3 6 3 6 33 -

5.6 4.5 7.6 4.8 8.5 58 2.7 0.9 1.1 9.2 2.0 3.2

6 6 9 4 8 4 2 1 1 8 2 3 6

acid

T2.2

T3

_____

_____

T4

T5

.

____-

114-128

114-126

156- 179

309 - 349

(132-151) a b

(132-151) a b

(183-196) a b

(382-424) d b

-

a

b

1.4

1.4 2.4 5.4 2.5 5.6 1.3 1.5

1 2 6 2 6 1 1

4.6 9.1 6.8 6.1 7.7 5.8 7.3

1 4

-

-

0.9 3.7 1.4 2.2 1.1 1.3 0.9 3.2

(359-378)

niol/mol Asx Thr Ser Glx GlY Ala V d1 Met Ile Leu TYr Phe His TrP LYS '4%

CYS Pro G 1cN

6.1 4.7 8.3 4.8 8.0 5.5 3.0 1.1 1.3

8.5 2.0 3.2

+++ -

2.5 n.d. n.d. -

+- + +

1.9 5.5 2.7 6.1 1.2 1.4 -

1.9 2.2 1.3

+

-

3.2 n.d. n.d. -

3.1 n.d. n.d.

3 6 13 -

1 2 6 2 6 1 1 -

2 2 t 1 3 4 3

2.4 1.9 1.4

+

-

2.3 n.d. n.d.

2 2 1 1 2 4 2

I

6.8 3.7 2.8 3.6 1.5 3.6 3.4 1.9 2.8 0.8 1.1

4 3 3 1 3 3 2 2 1

-

0.8 8.6 5.5 -

3

2.7

++

++

2 1 t 2 2

0.7 0.9 n.d. n.d.

+

1 .l 2.2 n.d. n.d.

+

+

N 0 N

Q

0.1

0

6

0

7 -

0

10

0

-7---

20

30

Elution time ( m i d Fig. S5. Elution profile and sequence data obtained after digestion of peptide T4 with chymotrypsin. Conditions for reversed-phase HPLC wcre the same as in Fig. S2. (*) Cysteine, according to the sequence in Fig. 2 (see main text). (+) Glycosyldted asparagine, according to Geyer et al. (1990). (+) degradation product of chymotrypsin.

5 11

I 5 6 I 8

9 6 2 3 1 2 6 6

+

-

1 2 1 1 1 3

-

-

1.5 2.0

1

-

n.d. n.d.

+

-

2

2 1

+