- Email: [email protected]
Proteolysis in the maturation of avian retroviruses does not require calcium
VIROLOGY
189, 77 l-774
(1992)
Proteolysis
in the Maturation VOLKER
‘Section
of Biochemistry,
Molecular Fox Chase
M.
VOGT,“,’
of Avian Retroviruses HAIM BuRsTEIN,t
AND
Does Not Require ANNA
Calcium
MARIE SKALKAt
and Cell Biology, Cornell University, Ithaca, New York 14853; and tlnstiture Cancer Center, 770 1 Burholme Avenue, Philadelphia, Pennsylvania 19 111 Received January 27, 1992; accepted May 2 I, 1992
for Cancer
Research,
After budding from the plasma membrane, retrovirus particles undergo a process of maturation, which includes changes in morphology caused by several proteolytic cleavages of the precursor of the internal structural proteins, products of the gag gene. Cleavage is mediated by the viral protease, PR. The fact that in most systems cleavage appears to occur only after assembly is complete, suggests that PR may become enzymatically active as a consequence of release of the virion from the cell. Using avian leukosis virus as a model system, we tested the hypothesis that leakage of calcium ions into newly budded virions plays a role in their maturation. We found that in both quail Qt35 cells and monkey COS-1 cells, maturation occurred normally in calcium-free medium and in the presence of EGTA. A calcium ionophore also did not affect maturation. We conclude that calcium influx does not act as a trigger for PR-mediated maturation. 0 1992 Academic Press. Inc.
The internal structural proteins of retroviruses (products of the gag gene) and the viral enzymes required for replication (products of the pal gene) are formed by proteolytic cleavages of the Gag and Gag-Pol precursor polypeptides, respectively. A protease encoded by the virus, PR, is responsible for these cleavages. PR belongs to the phylogenetically widespread class of aspartate proteases, which includes enzymes like renin and pepsin. In most retroviruses PR is encoded as a polypeptide domain of the GaggPol precursor, located between the C-terminal Gag protein, nucleocapsid (NC), and the N-terminal Pol protein, reverse transcriptase (RT). In a few retroviruses, such as the avian sarcoma and leukemia virus group (ASLV), PR is in the same genetic location but is expressed as the C-terminal domain of the Gag protein. The mature PR, for which the three-dimensional structure is known in the case of Rous sarcoma virus (a member of the ASLV) (7) and of human rmmunodeficiencyvirus (HIV-l) (2) functions as a dimer, with each subunit contributing to the active site. By contrast, the cell-derived enzymes are single polypeptides containing two structurally similar domains which function in an analogous fashion to that of the retroviral PR subunits. Proteolytic processing of retroviral proteins (3) and function of retroviral PRs (4, 5) have been reviewed. Although some exceptions have been reported (e.g., 6), proteolytlc cleavages of the viral Gag and Gag-Pol proteins normally take place after the virus particles have been assembled. In most retrovirus systems, “immature” viral cores, which are defined by their characteristic electron-lucent center in thin section electron microscopy, form at the inner surface of the plasma membrane (reviewed in 7). They become enveloped by a portion of the membrane and are released as ’ To whom
correspondence
should
free virus particles from the cell by a process called budding. After release, proteolysis separates the several domains of Gag and Gag-Pol precursors. This leads to collapse of the core, thereby giving the virion cores their “mature” or condensed morphology. For example, in the ASLV system rapidly harvested virus was found to have an immature morphology (8) and to contain the unprocessed gag precursor Pr76, which underwent spontaneous cleavage with a half time of about 3 min (9). In some retroviruses, exemplified by Mason-Pfizer monkey virus (MPMV), immature cores assemble in the cytoplasm of the infected cell, and migrate from there to the plasma membrane where they become enveloped and bud into the medrum. Again, proteolysis does not take place until the last stages of assembly or after the virus particles have been released. Mutations in MPMV that block myristoylation of the N-terminus of Gag and thereby prevent stable interaction of the preassembled cores with the membrane also block proteolytic cleavage of Gag (10). This is the best evidence that maturation requires that the nascent core particle be separated from the cell. If the cleavages of Gag and Gag-Pot precursor polypeptides occur only after formation of the virion is complete, some mechanism must exist to delay proteolysis until that time. One model to explain this delay proposes that proteolysis is initiated when the local concentration of the Gag or Gag- Pol precursors in the nascent virions reaches a level that permits dimerization of the PR domain. In both the ASLV and the HIV-1 system viral gag or gag-pal DNAs have been constructed to encode a “linked drmer” domain of PR--i.e., two identical subunits linked covalently in a manner similar to that of cellular aspartate proteases (1 l- 73). When these constructs are expressed in cells, premature proteolysis occurs, inhibiting virus assembly. These results can be interpreted as evidence for the concentra-
be addressed. 771
0042.6822/92
$5.00
CopyrIght 0 1992 by Academic Press, Inc All rights of reproducilnn I” any form reserved
772
SHORT
I???
.
5678
.
9
COMMUNICATIONS
”
- Pr76
-CA -MA
.
. . .
..--
-
FIG. 1. Effect of lack of calcium and EGTA in turkey and quail cells. Turkey fibroblast cells infected with the PrA strain of RSV, and quail Qt35 cells infected with a strain of RSV in which the viral SK was replaced with the bacterial neo gene, were grown in MEM medium (GIBCO) with 10% fetal calf serum as described (7 7). The quail cells had been selected in medium with the drug G418 to insure that all cells were infected. A plate of quail cells cultured in parallel expressed the same virus carrying the active site PR mutation D37l (17). Confluent 60-mm plates of cells were washed three times with MEM lacking both calcium and serum and then refed either with normal medium or with MEM lacking calcium but containing 10% dialyzed fetal calf serum and various concentrations of EGTA. After incubation of the cells at 39” for 6 hr, all the media were removed and immediately chilled and clarified by centrifugation for 8 min at 6000 g and then made 2 mM in EGTA. Virus particles were collected under a 10% sucrose cushion containing EGTA, after centrifugation for 20 min at 75,000 rpm in a TLl 00 centrifuge. The virus proteins were dissolved in 4% SDS sample buffer, electrophoresed on a 15% polyacrylamide gel, and then electroblotted to an Immobilon-P membrane and visualized by probing with rabbit antibodies to viral MA and CA proteins followed by 1251-A protein as described (77). Normal medium (lanes 1 and 5); calcium-free medium (lanes 2 and 6); calcium-free medium plus 0.5 mM EGTA (lanes 3 and 7); calciumfree medium plus 1 mM EGTA (lanes 4 and 8); 12 ng purified CA protein standard (lane 9); calcium containing medium from cells expressing virus with mutant PR (lane 10). Turkey cells, lanes 1-4; quail cells, lanes 5-8 and 10.
tion model, since covalent linkage of two PR coding regions bypasses the need to form intermolecular dimers. The results would appear to exclude models in which some cellular inhibitor of PR, which is not packaged into virions, prevents PR from functioning prematurely. However, the concentration model cannot easily account for the observation that myristoylation-deficient MPMV cores form quite normally but remain immature. Nor can it explain the common observation that in the final stages of budding, when only a narrow stalk still connects the virion to the cell, all retrovirus particles still look immature by electron microscopy. Although only sparse data are available, it has been suggested that PR is enzymatically inactive when it is
embedded in its Gag or Gag-Pol precursor polypeptide. For example, in the ASLV system, mutations at the NC-PR boundary block most or all of the cleavages of Gag protein in viva, suggesting that the PR domain must be liberated as a separate protein in order for normal proteolysis to occur (14, 15; G. Schatz and V. M. Vogt, unpublished data). Consistent with this inference, preliminary results suggest that the purified gag fragment NC-PR is inactive as a protease in vitro (M. Sellos-Moura and V. M. Vogt, unpublished). An analogous observation has been made for HIV-1 PR embedded in short stretches of flanking N-terminal and C-terminal PO/-derived amino acid residues, all expressed as a bacterial ma/A fusion protein (16). Constructs in which mutations prevented cleavage at the N- or C-terminus of PR yielded proteins with specific activities two orders of magnitude lower than that of PR itself. We considered the hypothesis that changes in the calcium ion concentration might provide a regulatory signal to trigger the initiation of proteolysis. This hypothesis is based on the likelihood that the concentration of free calcium in a virion would rise rapidly after its release from the cell. Before the membrane fusion event that leads to envelopment and release of the virus, the concentration of Ca*+ should be the same as that in the cytoplasm of the cell, about lop7 n/l. After this fusion, in the absence of a renewable source of ATP to drive any calcium pump that might be included in virions, and in the face of a large surface-to-volume ratio, calcium ions should leak into the virion, eventually approaching the medium concentration of about 1 O-3 M. It is not possible to predict the time course of this ion flow, since nothing is known about the permeability properties of virions. Calcium would leak into phospholipid vesicles of the size of virions in a matter of hours, but it may well be that virions are more permeable. Although aspartate proteases are not known to require metal ions as cofactors, one might imagine that a calcium influx could lead to a conformational change of the Gag or Gag-Pol precursor polypeptides, which in turn could activate PR, for example by allowing dimerization of the PR domains. Alternatively, it seemed conceivable that a small amount of a calcium-requiring protease like calpain could be incorporated into virions and that activation of this hypothetical cellular protease could trigger a cascade by which PR was liberated and active dimers were formed. We tested the hypothesis that calcium plays a role in triggering proteolysis after virion assembly, as it applies to ASLV. Two different expression systems were employed, turkey or quail cells shedding intact and infectious particles and monkey COS-1 cells transfected with a vector that overexpresses the ASLV gag gene.
SHORT
COMMUNICATIONS
773
12 . .
3.
4.
96
7s
‘?1~!‘121~14 .
.
1.5
-Pr76 -Pr76 69-
30-
*
* -CA
-CA
-MA
-MA
-PR
MEDIUM
-PR
LYSATE
FIG. 2. Effect of lack of calcium and calcium ionophore on monkey COS-1 cells. Cells were either mock transfected (lanes 2, 4, 5, 10, and 1 1) or transfected with pSV.Myrl (lanes 3, 6-9, and 12-l 5) as previously described (15). Forty-eight hours later cells were washed twice with PBS and then incubated with DMEM (GIBCO) lacking methionine (lanes 2 and 3) or with DMEM lacking both methronine and calcium, in the presence or absence of 25 plLl A231 87 (lanes 4-9 and 1 O-l 5, respectively) for 2 hr. Cells were then labeled with [?S]methronrne tn 0.8 ml of the same medium for 2.5 hr (lanes 2, 3, 4, 6, 10, and 12), 4 hr (lanes 7 and 13) 6 hr (lanes 8 and 14), and 8 hr (lanes 5, 9, 1 1, and 15). The culture medium was removed and centrifuged to concentrate virus-like particles. The particles and harvested cells were lysed as previously described (15) and gag proteins were immunoprecipitated from the lysate with anti-RSV antiserum. lmmunoprecipitated proteins were separated on SDS1 2% polyacrylamide gels, the proteins were fixed, and the gel was treated with Amplify, and radioactrve bands were detected by fluorography. (A) Medium-derived particles; (B) Cell lysates. Lane 1, molecular weight markers. The positrons (in kDa) of the markers are indicated on the left of each panel. The posrtion of the full-length precursor protein Pr76 and the cleavage products capsrd (CA), matrrx (MA), and protease (PR) are indicated at the right of each panel
We determined the effects on virus assembly of withdrawing calcium ions from the medium and of permeabilizing membranes to calcium with an ionophore. Figure 1 shows the results obtained in the avian cell system. Turkey fibroblast cells infected with RSV, and cloned quail Qt35 cells expressing the RCASneo provirus as described previously (17) were grown in complete medium containing calcium ions. The cells then were washed three times in calcium-free medium and incubated for 6 hr in calcium-free medium containing dialyzed serum and varying concentrations of EGTA. At the highest concentration of EGTA, distinct morphology changes and subsequent cytotoxicity were observed within 2 hr, as expected for cells growing in the complete absence of calcium ions. Virus particles released into the medium over a 6-hr period were collected by centrifugation. Then viral polypeptides were analyzed by SDS-PAGE and immunoblotting with anti-
serum to the gag proteins CA and MA. Proteolysis of the gag proteins in virions released from turkey (lanes 2-4) and quail (lanes 6-8) cells growing in calcium-free medium with (lanes 3, 4, 7, 8) or without (lanes 2, 6) EGTA was indistinguishable from the proteolysis of viral polypeptides in virions released from cells grown in calcium-containing medium (lanes 1 and 5). Proteolysis was dependent on viral PR as expected, since virions released from quail cells expressing a PR-defective mutant contained only the uncleaved gag precursor polypeptide Pr76 (lane 10). Similar results were obtained in the mammalian cell expression system. COS-1 cells were transfected with an SV-40-based plasmid carrying the RSV gag gene derivative Myr 1, which is altered so that the N-terminus of Pr76 has the myristoylation site of the oncoprotein ~60s~ (18). It has been shown previously that the Myr 1 particles generated in this system are indis-
774
SHORT
COMMUNICATIONS
tinguishable in their physical and biochemical characteristics from those generated in infected avian cells. Two days after transfection the cells were washed, incubated for 2 hr in calcium-free medium with or without the Ca*+ ionophore A231 87, and then labeled for 2 to 8 hr in presence of [35S]methionine in the same medium. Virus-like particles were centrifuged from the medium. Lysates of the cells were prepared, and the gag proteins therein were collected by means of immunoprecipitation. The proteins in the virus-like particles and the lysate fractions were analyzed by SDS-PAGE and fluorography. Figure 2 displays the outcome of these analyses. After 2 hr of labeling, particles released into calcium-free medium (lane 3) had the same amounts of radioactive mature CA, MA, and PR polypeptides as particles released into normal medium (lane 12), with no evidence of Pr76 or of any intermediate cleavage products. The calcium ionophore, which was added to trigger release of internal stores of Ca*+, reduced the overall level of labeling of cellular proteins. However, in this case as well, almost exclusively mature proteins were observed (lanes 6-9). Given the toxicity of the drug, we do not attribute significance to the partial inhibition of the appearance of the mature MA protein in these experiments (e.g., compare lanes 6 and 7 with lanes 12 and 13). In sum, neither absence of calcium nor presence of ionophore led to changes in the relative amounts of the Pr76 and mature proteins in the cell lysate fractions. We conclude from these results that avian sarcoma and leukemia viruses do not use calcium influx as a signal to trigger proteolytic cleavages of Gag and GagPol precursor proteins. This conclusion must be tempered by the absence of any direct measurement of Ca*+ concentrations in maturing particles. It cannot be excluded rigorously that enough calcium was present in the calcium-free medium to enter the particles and initiate the hypothetical proteolytic cascade. We consider this alternative unlikely, however, because even at a concentration of EGTA that was sufficient to induce cytopathic effects in the cells, no partially cleaved intermediates were evident. Although it is conceivable that internal cellular stores of calcium might have led to a burst of calcium ions at the time of virus release, the source of such a hypothetical burst is unclear; retroviruses budding at the plasma membrane are not observed to be in the proximity of intracellular vesicular compartments. Taken together, these results from very different cell expression systems argue that proteolysis during retroviral maturation does not involve calcium. Thus the mechanism underlying the delay of proteolysis remains unknown. Other events that might occur after release of the virus particle and that might
play a role in this delay include changes in other inorganic ions (e.g., K+ and Na+), changes in the redox potential, changes in small organic molecules (e.g., ATP), and changes in covalent modifications (e.g., phosphorylations or dephosphorylations). Perhaps the cellular ubiquitin that is packaged by an unknown mechanism into virions plays a role in proteolysis (19). Alternatively, conformational changes brought about by the closure of the core could trigger this event. Additional experiments will be required to resolve this interesting question.
ACKNOWLEDGMENTS We thank J. Wills for the infected turkey cells and G. Schatz for infected quail cells and for performing the Western blotting. This work was supported by National Institutes of Health Grants CA20081 to V.M.V. and CA47486, CA06927, and RR-05539 to A.M.%, a grant from the Pew Charitable Trust and an appropriation from the Commonwealth of Pennsylvania also to A.M.S.
REFERENCES 1. JASKOLSKI, M., MILLER, M., RAO, J. K. M., LEIS, J., and WLODAWER, A., Biochemistry 29, 5889-5898 (1990). 2. NAVIA, M. A., FITZGERALD, P. M. D., MCKEEVER, B. M., LEU, C.-T., HEIMBACH, J. C., HERBER, W. K., SIGAL, I. S., DARKE. P. L., and SPRINGER, J. P.. Nature 337, 615-620 (1989). 3. DICKSON, C., EISENMAN, R., FAN, H., HUNTER, E., ~~~TEIcH, N., In “RNATumor Viruses” (R. Weiss, N. Teich, H. Varmus, and J. Coffin, (eds.), Vol. 1. pp. 513-648. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1984. KR&USSLICH, H.-G., and WIMMER, E., Annu. Rev. Biochem. 57, 701-754 (1988). SKALKA, A. M., Cell56, 911-913(1989). KAPLAN, A. H., and SWANSTROM, R., Proc. Nat/. Acad. Sci. USA 88,4528-4532 (1991). BEARD, J. W., In “Ultrastructure of Animal Viruses and Bacteriophages” (A. J. Dalton and F. Haguenau, Eds.), pp. 262-282. Academic Press, San Diego, 1973. 8. KORB, J., TRAVNICEK, M., and RIMAN, J. Infervirology 7, 21 l-224 (1976). 9. MOELLING, K., Scorr, A., DITTMAR, K. E. J., and OWADA, M., J. Viral. 33, 680-688 (1980). 10. RHEE, S. S., and HUNTER, E., 1. viral. 61, 1045-1053 (1987). 11. BIZUB, D., WEBER, I. T.. CAMERON, C., LEIS, J. P., and SKALKA, A. M., J. Biol. Chem. 266, 4951-4958 (1991). 12. BURSTEIN, H., BIZUB, D., and SKALKA, A. M., /. Viral. 65, 61656172 (1991). 13. KR&JSSLICH, H.-G., Proc. Nat/. Acad. Sci. USA 88, 3213-3217 (1991). 14. OERTLE, S., and SPAHR, P.-F., /. Viral. 64, 5757-5763 (1990). 15. BURSTEIN, H., BIZUB. D.. KOTLER, M., SCHATZ, G., VOGT, V. M., and SKALKA, A. M., J. Viral. 66, 1781-1785 (1992). 16. LOUIS, J. M., MCDONALD, R. A., NASHED, N. T.. WONDRAK, E. M., JERINA, D. M., OROSZLAN. S., and MORA, P. T., Eur. f. Biochem. 199, 361-369 (1991). 17. STEWART, L., SCHATZ, G., and VOGT. V. M., J. Viral. 64, 50765092 (1990). 18. WILLS, J. W., CRAVEN, R. C., and ACHACOSO, J. A., J. Viral. 63, 4331-4343 (1989). 19. PUTTERMAN, D.. PEPINSKY, R. B., and VOGT, V. M., virology 176, 633-637 (1990).
189, 77 l-774
(1992)
Proteolysis
in the Maturation VOLKER
‘Section
of Biochemistry,
Molecular Fox Chase
M.
VOGT,“,’
of Avian Retroviruses HAIM BuRsTEIN,t
AND
Does Not Require ANNA
Calcium
MARIE SKALKAt
and Cell Biology, Cornell University, Ithaca, New York 14853; and tlnstiture Cancer Center, 770 1 Burholme Avenue, Philadelphia, Pennsylvania 19 111 Received January 27, 1992; accepted May 2 I, 1992
for Cancer
Research,
After budding from the plasma membrane, retrovirus particles undergo a process of maturation, which includes changes in morphology caused by several proteolytic cleavages of the precursor of the internal structural proteins, products of the gag gene. Cleavage is mediated by the viral protease, PR. The fact that in most systems cleavage appears to occur only after assembly is complete, suggests that PR may become enzymatically active as a consequence of release of the virion from the cell. Using avian leukosis virus as a model system, we tested the hypothesis that leakage of calcium ions into newly budded virions plays a role in their maturation. We found that in both quail Qt35 cells and monkey COS-1 cells, maturation occurred normally in calcium-free medium and in the presence of EGTA. A calcium ionophore also did not affect maturation. We conclude that calcium influx does not act as a trigger for PR-mediated maturation. 0 1992 Academic Press. Inc.
The internal structural proteins of retroviruses (products of the gag gene) and the viral enzymes required for replication (products of the pal gene) are formed by proteolytic cleavages of the Gag and Gag-Pol precursor polypeptides, respectively. A protease encoded by the virus, PR, is responsible for these cleavages. PR belongs to the phylogenetically widespread class of aspartate proteases, which includes enzymes like renin and pepsin. In most retroviruses PR is encoded as a polypeptide domain of the GaggPol precursor, located between the C-terminal Gag protein, nucleocapsid (NC), and the N-terminal Pol protein, reverse transcriptase (RT). In a few retroviruses, such as the avian sarcoma and leukemia virus group (ASLV), PR is in the same genetic location but is expressed as the C-terminal domain of the Gag protein. The mature PR, for which the three-dimensional structure is known in the case of Rous sarcoma virus (a member of the ASLV) (7) and of human rmmunodeficiencyvirus (HIV-l) (2) functions as a dimer, with each subunit contributing to the active site. By contrast, the cell-derived enzymes are single polypeptides containing two structurally similar domains which function in an analogous fashion to that of the retroviral PR subunits. Proteolytic processing of retroviral proteins (3) and function of retroviral PRs (4, 5) have been reviewed. Although some exceptions have been reported (e.g., 6), proteolytlc cleavages of the viral Gag and Gag-Pol proteins normally take place after the virus particles have been assembled. In most retrovirus systems, “immature” viral cores, which are defined by their characteristic electron-lucent center in thin section electron microscopy, form at the inner surface of the plasma membrane (reviewed in 7). They become enveloped by a portion of the membrane and are released as ’ To whom
correspondence
should
free virus particles from the cell by a process called budding. After release, proteolysis separates the several domains of Gag and Gag-Pol precursors. This leads to collapse of the core, thereby giving the virion cores their “mature” or condensed morphology. For example, in the ASLV system rapidly harvested virus was found to have an immature morphology (8) and to contain the unprocessed gag precursor Pr76, which underwent spontaneous cleavage with a half time of about 3 min (9). In some retroviruses, exemplified by Mason-Pfizer monkey virus (MPMV), immature cores assemble in the cytoplasm of the infected cell, and migrate from there to the plasma membrane where they become enveloped and bud into the medrum. Again, proteolysis does not take place until the last stages of assembly or after the virus particles have been released. Mutations in MPMV that block myristoylation of the N-terminus of Gag and thereby prevent stable interaction of the preassembled cores with the membrane also block proteolytic cleavage of Gag (10). This is the best evidence that maturation requires that the nascent core particle be separated from the cell. If the cleavages of Gag and Gag-Pot precursor polypeptides occur only after formation of the virion is complete, some mechanism must exist to delay proteolysis until that time. One model to explain this delay proposes that proteolysis is initiated when the local concentration of the Gag or Gag- Pol precursors in the nascent virions reaches a level that permits dimerization of the PR domain. In both the ASLV and the HIV-1 system viral gag or gag-pal DNAs have been constructed to encode a “linked drmer” domain of PR--i.e., two identical subunits linked covalently in a manner similar to that of cellular aspartate proteases (1 l- 73). When these constructs are expressed in cells, premature proteolysis occurs, inhibiting virus assembly. These results can be interpreted as evidence for the concentra-
be addressed. 771
0042.6822/92
$5.00
CopyrIght 0 1992 by Academic Press, Inc All rights of reproducilnn I” any form reserved
772
SHORT
I???
.
5678
.
9
COMMUNICATIONS
”
- Pr76
-CA -MA
.
. . .
..--
-
FIG. 1. Effect of lack of calcium and EGTA in turkey and quail cells. Turkey fibroblast cells infected with the PrA strain of RSV, and quail Qt35 cells infected with a strain of RSV in which the viral SK was replaced with the bacterial neo gene, were grown in MEM medium (GIBCO) with 10% fetal calf serum as described (7 7). The quail cells had been selected in medium with the drug G418 to insure that all cells were infected. A plate of quail cells cultured in parallel expressed the same virus carrying the active site PR mutation D37l (17). Confluent 60-mm plates of cells were washed three times with MEM lacking both calcium and serum and then refed either with normal medium or with MEM lacking calcium but containing 10% dialyzed fetal calf serum and various concentrations of EGTA. After incubation of the cells at 39” for 6 hr, all the media were removed and immediately chilled and clarified by centrifugation for 8 min at 6000 g and then made 2 mM in EGTA. Virus particles were collected under a 10% sucrose cushion containing EGTA, after centrifugation for 20 min at 75,000 rpm in a TLl 00 centrifuge. The virus proteins were dissolved in 4% SDS sample buffer, electrophoresed on a 15% polyacrylamide gel, and then electroblotted to an Immobilon-P membrane and visualized by probing with rabbit antibodies to viral MA and CA proteins followed by 1251-A protein as described (77). Normal medium (lanes 1 and 5); calcium-free medium (lanes 2 and 6); calcium-free medium plus 0.5 mM EGTA (lanes 3 and 7); calciumfree medium plus 1 mM EGTA (lanes 4 and 8); 12 ng purified CA protein standard (lane 9); calcium containing medium from cells expressing virus with mutant PR (lane 10). Turkey cells, lanes 1-4; quail cells, lanes 5-8 and 10.
tion model, since covalent linkage of two PR coding regions bypasses the need to form intermolecular dimers. The results would appear to exclude models in which some cellular inhibitor of PR, which is not packaged into virions, prevents PR from functioning prematurely. However, the concentration model cannot easily account for the observation that myristoylation-deficient MPMV cores form quite normally but remain immature. Nor can it explain the common observation that in the final stages of budding, when only a narrow stalk still connects the virion to the cell, all retrovirus particles still look immature by electron microscopy. Although only sparse data are available, it has been suggested that PR is enzymatically inactive when it is
embedded in its Gag or Gag-Pol precursor polypeptide. For example, in the ASLV system, mutations at the NC-PR boundary block most or all of the cleavages of Gag protein in viva, suggesting that the PR domain must be liberated as a separate protein in order for normal proteolysis to occur (14, 15; G. Schatz and V. M. Vogt, unpublished data). Consistent with this inference, preliminary results suggest that the purified gag fragment NC-PR is inactive as a protease in vitro (M. Sellos-Moura and V. M. Vogt, unpublished). An analogous observation has been made for HIV-1 PR embedded in short stretches of flanking N-terminal and C-terminal PO/-derived amino acid residues, all expressed as a bacterial ma/A fusion protein (16). Constructs in which mutations prevented cleavage at the N- or C-terminus of PR yielded proteins with specific activities two orders of magnitude lower than that of PR itself. We considered the hypothesis that changes in the calcium ion concentration might provide a regulatory signal to trigger the initiation of proteolysis. This hypothesis is based on the likelihood that the concentration of free calcium in a virion would rise rapidly after its release from the cell. Before the membrane fusion event that leads to envelopment and release of the virus, the concentration of Ca*+ should be the same as that in the cytoplasm of the cell, about lop7 n/l. After this fusion, in the absence of a renewable source of ATP to drive any calcium pump that might be included in virions, and in the face of a large surface-to-volume ratio, calcium ions should leak into the virion, eventually approaching the medium concentration of about 1 O-3 M. It is not possible to predict the time course of this ion flow, since nothing is known about the permeability properties of virions. Calcium would leak into phospholipid vesicles of the size of virions in a matter of hours, but it may well be that virions are more permeable. Although aspartate proteases are not known to require metal ions as cofactors, one might imagine that a calcium influx could lead to a conformational change of the Gag or Gag-Pol precursor polypeptides, which in turn could activate PR, for example by allowing dimerization of the PR domains. Alternatively, it seemed conceivable that a small amount of a calcium-requiring protease like calpain could be incorporated into virions and that activation of this hypothetical cellular protease could trigger a cascade by which PR was liberated and active dimers were formed. We tested the hypothesis that calcium plays a role in triggering proteolysis after virion assembly, as it applies to ASLV. Two different expression systems were employed, turkey or quail cells shedding intact and infectious particles and monkey COS-1 cells transfected with a vector that overexpresses the ASLV gag gene.
SHORT
COMMUNICATIONS
773
12 . .
3.
4.
96
7s
‘?1~!‘121~14 .
.
1.5
-Pr76 -Pr76 69-
30-
*
* -CA
-CA
-MA
-MA
-PR
MEDIUM
-PR
LYSATE
FIG. 2. Effect of lack of calcium and calcium ionophore on monkey COS-1 cells. Cells were either mock transfected (lanes 2, 4, 5, 10, and 1 1) or transfected with pSV.Myrl (lanes 3, 6-9, and 12-l 5) as previously described (15). Forty-eight hours later cells were washed twice with PBS and then incubated with DMEM (GIBCO) lacking methionine (lanes 2 and 3) or with DMEM lacking both methronine and calcium, in the presence or absence of 25 plLl A231 87 (lanes 4-9 and 1 O-l 5, respectively) for 2 hr. Cells were then labeled with [?S]methronrne tn 0.8 ml of the same medium for 2.5 hr (lanes 2, 3, 4, 6, 10, and 12), 4 hr (lanes 7 and 13) 6 hr (lanes 8 and 14), and 8 hr (lanes 5, 9, 1 1, and 15). The culture medium was removed and centrifuged to concentrate virus-like particles. The particles and harvested cells were lysed as previously described (15) and gag proteins were immunoprecipitated from the lysate with anti-RSV antiserum. lmmunoprecipitated proteins were separated on SDS1 2% polyacrylamide gels, the proteins were fixed, and the gel was treated with Amplify, and radioactrve bands were detected by fluorography. (A) Medium-derived particles; (B) Cell lysates. Lane 1, molecular weight markers. The positrons (in kDa) of the markers are indicated on the left of each panel. The posrtion of the full-length precursor protein Pr76 and the cleavage products capsrd (CA), matrrx (MA), and protease (PR) are indicated at the right of each panel
We determined the effects on virus assembly of withdrawing calcium ions from the medium and of permeabilizing membranes to calcium with an ionophore. Figure 1 shows the results obtained in the avian cell system. Turkey fibroblast cells infected with RSV, and cloned quail Qt35 cells expressing the RCASneo provirus as described previously (17) were grown in complete medium containing calcium ions. The cells then were washed three times in calcium-free medium and incubated for 6 hr in calcium-free medium containing dialyzed serum and varying concentrations of EGTA. At the highest concentration of EGTA, distinct morphology changes and subsequent cytotoxicity were observed within 2 hr, as expected for cells growing in the complete absence of calcium ions. Virus particles released into the medium over a 6-hr period were collected by centrifugation. Then viral polypeptides were analyzed by SDS-PAGE and immunoblotting with anti-
serum to the gag proteins CA and MA. Proteolysis of the gag proteins in virions released from turkey (lanes 2-4) and quail (lanes 6-8) cells growing in calcium-free medium with (lanes 3, 4, 7, 8) or without (lanes 2, 6) EGTA was indistinguishable from the proteolysis of viral polypeptides in virions released from cells grown in calcium-containing medium (lanes 1 and 5). Proteolysis was dependent on viral PR as expected, since virions released from quail cells expressing a PR-defective mutant contained only the uncleaved gag precursor polypeptide Pr76 (lane 10). Similar results were obtained in the mammalian cell expression system. COS-1 cells were transfected with an SV-40-based plasmid carrying the RSV gag gene derivative Myr 1, which is altered so that the N-terminus of Pr76 has the myristoylation site of the oncoprotein ~60s~ (18). It has been shown previously that the Myr 1 particles generated in this system are indis-
774
SHORT
COMMUNICATIONS
tinguishable in their physical and biochemical characteristics from those generated in infected avian cells. Two days after transfection the cells were washed, incubated for 2 hr in calcium-free medium with or without the Ca*+ ionophore A231 87, and then labeled for 2 to 8 hr in presence of [35S]methionine in the same medium. Virus-like particles were centrifuged from the medium. Lysates of the cells were prepared, and the gag proteins therein were collected by means of immunoprecipitation. The proteins in the virus-like particles and the lysate fractions were analyzed by SDS-PAGE and fluorography. Figure 2 displays the outcome of these analyses. After 2 hr of labeling, particles released into calcium-free medium (lane 3) had the same amounts of radioactive mature CA, MA, and PR polypeptides as particles released into normal medium (lane 12), with no evidence of Pr76 or of any intermediate cleavage products. The calcium ionophore, which was added to trigger release of internal stores of Ca*+, reduced the overall level of labeling of cellular proteins. However, in this case as well, almost exclusively mature proteins were observed (lanes 6-9). Given the toxicity of the drug, we do not attribute significance to the partial inhibition of the appearance of the mature MA protein in these experiments (e.g., compare lanes 6 and 7 with lanes 12 and 13). In sum, neither absence of calcium nor presence of ionophore led to changes in the relative amounts of the Pr76 and mature proteins in the cell lysate fractions. We conclude from these results that avian sarcoma and leukemia viruses do not use calcium influx as a signal to trigger proteolytic cleavages of Gag and GagPol precursor proteins. This conclusion must be tempered by the absence of any direct measurement of Ca*+ concentrations in maturing particles. It cannot be excluded rigorously that enough calcium was present in the calcium-free medium to enter the particles and initiate the hypothetical proteolytic cascade. We consider this alternative unlikely, however, because even at a concentration of EGTA that was sufficient to induce cytopathic effects in the cells, no partially cleaved intermediates were evident. Although it is conceivable that internal cellular stores of calcium might have led to a burst of calcium ions at the time of virus release, the source of such a hypothetical burst is unclear; retroviruses budding at the plasma membrane are not observed to be in the proximity of intracellular vesicular compartments. Taken together, these results from very different cell expression systems argue that proteolysis during retroviral maturation does not involve calcium. Thus the mechanism underlying the delay of proteolysis remains unknown. Other events that might occur after release of the virus particle and that might
play a role in this delay include changes in other inorganic ions (e.g., K+ and Na+), changes in the redox potential, changes in small organic molecules (e.g., ATP), and changes in covalent modifications (e.g., phosphorylations or dephosphorylations). Perhaps the cellular ubiquitin that is packaged by an unknown mechanism into virions plays a role in proteolysis (19). Alternatively, conformational changes brought about by the closure of the core could trigger this event. Additional experiments will be required to resolve this interesting question.
ACKNOWLEDGMENTS We thank J. Wills for the infected turkey cells and G. Schatz for infected quail cells and for performing the Western blotting. This work was supported by National Institutes of Health Grants CA20081 to V.M.V. and CA47486, CA06927, and RR-05539 to A.M.%, a grant from the Pew Charitable Trust and an appropriation from the Commonwealth of Pennsylvania also to A.M.S.
REFERENCES 1. JASKOLSKI, M., MILLER, M., RAO, J. K. M., LEIS, J., and WLODAWER, A., Biochemistry 29, 5889-5898 (1990). 2. NAVIA, M. A., FITZGERALD, P. M. D., MCKEEVER, B. M., LEU, C.-T., HEIMBACH, J. C., HERBER, W. K., SIGAL, I. S., DARKE. P. L., and SPRINGER, J. P.. Nature 337, 615-620 (1989). 3. DICKSON, C., EISENMAN, R., FAN, H., HUNTER, E., ~~~TEIcH, N., In “RNATumor Viruses” (R. Weiss, N. Teich, H. Varmus, and J. Coffin, (eds.), Vol. 1. pp. 513-648. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1984. KR&USSLICH, H.-G., and WIMMER, E., Annu. Rev. Biochem. 57, 701-754 (1988). SKALKA, A. M., Cell56, 911-913(1989). KAPLAN, A. H., and SWANSTROM, R., Proc. Nat/. Acad. Sci. USA 88,4528-4532 (1991). BEARD, J. W., In “Ultrastructure of Animal Viruses and Bacteriophages” (A. J. Dalton and F. Haguenau, Eds.), pp. 262-282. Academic Press, San Diego, 1973. 8. KORB, J., TRAVNICEK, M., and RIMAN, J. Infervirology 7, 21 l-224 (1976). 9. MOELLING, K., Scorr, A., DITTMAR, K. E. J., and OWADA, M., J. Viral. 33, 680-688 (1980). 10. RHEE, S. S., and HUNTER, E., 1. viral. 61, 1045-1053 (1987). 11. BIZUB, D., WEBER, I. T.. CAMERON, C., LEIS, J. P., and SKALKA, A. M., J. Biol. Chem. 266, 4951-4958 (1991). 12. BURSTEIN, H., BIZUB, D., and SKALKA, A. M., /. Viral. 65, 61656172 (1991). 13. KR&JSSLICH, H.-G., Proc. Nat/. Acad. Sci. USA 88, 3213-3217 (1991). 14. OERTLE, S., and SPAHR, P.-F., /. Viral. 64, 5757-5763 (1990). 15. BURSTEIN, H., BIZUB. D.. KOTLER, M., SCHATZ, G., VOGT, V. M., and SKALKA, A. M., J. Viral. 66, 1781-1785 (1992). 16. LOUIS, J. M., MCDONALD, R. A., NASHED, N. T.. WONDRAK, E. M., JERINA, D. M., OROSZLAN. S., and MORA, P. T., Eur. f. Biochem. 199, 361-369 (1991). 17. STEWART, L., SCHATZ, G., and VOGT. V. M., J. Viral. 64, 50765092 (1990). 18. WILLS, J. W., CRAVEN, R. C., and ACHACOSO, J. A., J. Viral. 63, 4331-4343 (1989). 19. PUTTERMAN, D.. PEPINSKY, R. B., and VOGT, V. M., virology 176, 633-637 (1990).