AIDS RESEARCH AND HUMAN RETROVIRUSES Volume 8, Number 4, 1992 Mary Ann Liebert, Inc., Publishers
and Kinetic Analysis of sCD4 Binding to HIV-1 Virions and of gpl20 Dissociation
Thermodynamic
JOHN P.
MOORE' and
PER JOHAN
KLASSE1"2
ABSTRACT Kinetic and thermodynamic aspects of the binding of sCD4 to intact virions of human immunodeficiency virus type 1 (HIV-1 RF), and of the subsequent induction of gpl20 dissociation were studied. sCD4 binding to virions at 4 and 37°C is half-maximal at approximately 40 and 10 nM, respectively. The transition between low-affinity and high-affinity binding of sCD4 to virions occurs over a narrow temperature range between 20 and 25°C. Shedding of gpl20 from virions after sCD4 binding is also temperature dependent, being initiated above approximately 20°C. The minimum temperatures for the sCD4 affinity transition and gpl20 shedding are, therefore, similar and we suggest how the two processes might be related mechanistically.
INTRODUCTION
ONEmunodeficiency
treating human im¬ virus (HIV) infection is to inhibit virus binding to CD4+ cells. Derivatives of the viral receptor, CD4,1"'3 rendered soluble by a truncation that removes the transmembrane and cytoplasmic region of the molecule (sCD4), have been expressed.48 Bivalent or multivalent immunoglobulin chimeras (sCD4-IgG and sCD4-IgM) of sCD4 also have been produced, with significantly improved antiviral properties in vitro.9~" sCD4 binds with high affinity to recombinant gp 120 or virion-derived soluble gp 120 (HIV-1 s SU glycoprotein) from laboratory-adapted isolates,4"l2~'4 and also potently neutralizes these strains of HIV-1 in vi¬ tro. 4-8-15-'7 However, isolates of HIV-1 not adapted to growth in transformed cell lines (primary isolates) require much higher concentrations of HIV for neutralization in vitro;18 this may compromise sCD4's therapeutic efficiency in HIV infec¬ pharmacological approach to
'
tion.19·20
It has been reported that neutralization of HIV-1 (IIIB and RF) in vitro occurred by a combination of mechanisms: competitive inhibition of virus cell binding at low sCD4 concentra¬ tions;16"17'21 and facilitated dissociation of gpl20 from virions at higher sCD4 concentrations,2123 where synergistic neutral¬ ization occurs.17'21"24 In principle, variation in the efficiency of either sCD4 binding to virions or in its induction of gpl20
dissociation could account for the relative resistance of primary isolates to sCD4.18 Indeed, sCD4-insensitive escape mutants have been isolated with both of these properties.25 Prior to analyzing primary isolates, we sought to increase our under¬ standing of the mechanisms by which sCD4 binds to a wellcharacterized tissue culture-adapted strain of HIV-1 (RF) and induces gpl20 dissociation. We reported previously that both sCD4 binding and gpl20 shedding were temperature-dependent reactions, in that there were major differences at 4°C compared with 37°C.21 These observations implied that conformational changes in the HIV envelope post-sCD4 binding were enthalpically driven. Such changes might be part of the mechanisms by which HIV fuses with CD4+ cells.26 To evaluate these conformational changes more thoroughly, we examined in detail how sCD4 binding and gpl20 shedding were influenced by temperature. We confirmed that the affinity of sCD4 for virions increased with temperature, but found that the greatest increase in affinity occurred over a narrow temperature range between 20 and 25°C. sCD4-facilitated dissociation of gpl 20 from virions was a function both of the sCD4 concentration and of temperature; the minimum temperature for dissociation at saturating sCD4 concentrations was approximately 20°C, similar to that for the sCD4 affinity change. We suggest that a larger CD4-interactive surface is exposed on gpl20 at the higher temperatures, and that this is concomitant with a decreased interaction of gpl 20 with gp41.
Chester Beatty Laboratories, The Institute of Cancer Research, 237 Fulham Road, London SW3 6JB, England. 2Section of Virology, Department of Medical Microbiology, University of Lund, Lund, Sweden.
443
444
MOORE AND KLASSE
MATERIALS AND METHODS Viruses Extracellular HIV-1 RF (RF) was prepared from H9 two viral harvests were used which did not vary significantly in their responses to sCD4. Most experi¬ ments have since been repeated on additional RF harvests with similar results. Harvest 1 contained 150 ng/ml gpl20 and 3,000 ng/ml p24, of which 70% and 74% were virion bound, the rest soluble, when analyzed by Sephacryl S-1000 gel filtration chromatography.21-23-27 For harvest 2, the gpl20 and p24 contents were 90 and 1500 ng/ml with 60% and 65% of the totals present as virion-bound antigen. Infectious virus titers (log TCID50/ml) for harvests 1 and 2 were 7.0 and 6.5, respectively, assessed by syncytium formation in C8166 cells.27
cells.21"27 In this study
Binding of sCD4
to
virions and
detergent-disrupted
virions RF (1-3 ng total gpl20) was incubated with sCD4 in 100 µ RPMI/10% fetal calf serum (FCS) for 1 h at 4 or 37°C (or as indicated in the text) before gel filtration using Sephacryl S-1000 as previously described.21"23'27 Elution was with ice-cold Trisbuffered saline (TBS) or RPMI-1640 culture medium, and the columns were precooled on ice. Similar data were obtained when the columns were run at room temperature. The fraction (550 µ ) containing virions was collected and disrupted with 60 µ of 10% Nonidet P40 (NP40) plus 10% fetal calf serum and immediately frozen unless indicated otherwise. In some exper¬ iments, the soluble antigen fraction (1.2 ml) was collected and processed in similar manner. For determination of sCD4 binding to detergent-disrupted virions,21 RF was mixed with 1% NP40 and incubated for 1 h at 37°C with sCD4 in a total volume of 100 µ , then diluted with 510 µ of 1% NP40/FCS to mimic the dilution of virions during gel filtration, and frozen. The amount of RF in the incubation was adjusted to give a gpl20 concentration that was the same as that in the detergent-solubilized virion fraction eluted from the columns (typically 1-3 ng/ml or 0.01-0.03 nM).
Determination
of sCD4-gpl20 complexes
µ ) of gel-filtered virion-sCD4 com¬ captured onto a solid phase by sheep antibody D7324 to the carboxyl-terminus of gp 120, in the presence of 2% nonfat milk powder.14,21 (Regions close to the carboxyl-termi¬ nus of gpl20 normally are involved in gp41 binding, ref. 28.) NP40 solutions (100
plexes
were
Bound sCD4 was detected with a rabbit anti-sCD4 antiserum (CBL-34) and a sheep anti-rabbit Ig alkaline phosphatase con¬ jugate. 14-21 NP40 solutions of virions and sCD4 (containing free sCD4) were captured onto the solid phase in the same way, except that after the 2 h capture stage unbound sCD4 was washed away with TBS and the wells were incubated for 2 h with 100 µ of TBS/1% NP40/1% FCS to allow dissociation of
sCD4-gpl20 complexes. This was negligible.21 sCD4-gpl20 complex data are presented as OD492
values derived from the above assay. This is because the level of the signal varies with both the gpl20 concentration and the amount of sCD4, complicating calibrations. However, in most experi¬ ments the gpl20 concentration in the immunoassay wells was
kept constant, so variation in signal represents variation in the amount of gpl20-bound sCD4. Under these conditions, the OD4y2 signal in the range we analyze is expected to be in direct proportion to the amount of complexed sCD4. Where the gpl20 concentration was varied (e.g., because of prior shedding), OD492 changes of 0.10 in the range 0.20-0.70 correspond to approximately 1.5-fold differences in gpl20 concentration. A plot of OD492 as a function of log(gpl20) is nearly linear in this OD4y2 range. All enzyme-linked immunosorbent assay (ELISA) determina¬ tions were carried out in triplicate, and data were presented as means ± SD. In figures where SD bars are shown, the omission of a bar means that it lies within the symbol. In figures where SD bars are omitted for clarity of presentation, their magnitude is similar to that shown in other figures describing experiments of similar design. Determination
of gpl20 and p24
Concentrations of gpl20 and p24 in virion extracts were determined by twin-site ELISA as previously described,21"27 using CHO-derived gpl20 (IIIB)29 and baculovirus-expressed p24 (IIIB)30 as calibrants. The volumes assayed were adjusted where possible to be on the linear portion of the calibration curves (typically 100 pgof gp 120 or p24 per well, giving OD492 signals of about 0.40). The sCD4 present in column-purified sCD4-gpl20 complexes did not interfere significantly with either capture or detection of gpl20 in the assays used. High concentrations of free sCD4 do cause a small, nonspecific assay interference that complicates precise determination ( ± 1.5-fold) of gpl20 in the soluble antigen fraction of gel filtration eluates. Within this factor, the sum of virion-bound and soluble gpl 20 in the column eluates was constant, indicating that NP40 had efficiently dissociated gpl20 from gp41. Unless otherwise indicated, virion p24 levels were not significantly affected by any of the treatments described, indicating that virion recovery
during chromatography was constant. RESULTS We have described the binding of sCD4 to HIV-1 virions and its induction of gpl20 dissociation.21"23 To analyze the thermo¬ dynamics of this process, it was necessary to obtain virus preparations with a high level of virion-bound gpl20, such that the amounts of gpl 20 that remained after incubation with sCD4 at 37°C gave significant signals in our immunoassay for gpl 20sCD4 complexes. We obtained virus preparations that contained up to threefold more gpl20 than previously,2'-23 yet still retained a high percentage of their gpl20. In our experience there is considerable variation in virion gpl20 retention in culture,27 but in general RF is the most stable of the commonly used laboratory-adapted HIV-1 isolates (cf., LAV-1 /IIIB, SF-2, MN). The particular harvests we analyzed in the present study retained their gpl20 extremely well. In the absence of sCD4 there was no loss of gp 120 over 4 h at 37°C within the detection limit of our assay systems (i.e., ± 15% approximately) (data not
shown). The binding of sCD4 to intact and detergent-disrupted RF virions differed significantly, as shown by the sCD4 titration curves (Fig. 1). sCD4 bound half-maximally to detergent-
THERMODYNAMICS AND KINETICS OF sCD4 BINDING TO HIV
sCD4-gp120 complexes
OD.92 0-3
y
0-2 0-1
./
.y
»«- "
FIG. 1. Binding of sCD4 to intact and detergent-disrupted RF virions. Intact RF virions were incubated with the sCD4 concen¬ trations indicated at 37°C (·) or 4°C (o) before isolation of the virion-sCD4 complexes by gel filtration and determination of sCD4-gp 120 complexes after disruption of the virions in NP40. For comparison, virions separated from soluble gpl20 by gel filtration ( ), or unfractionated virions ( ), were disrupted in 1 % NP40 then incubated with the sCD4 concentrations indicated before determination of sCD4-gpl20 complexes. No sCD4 assay background, common for all curves ( ). Note that for all the curves bar (·), the gpl20 concentration in the immunoassay wells was constant. In the latter curve, gpl20 shedding causes a reduction in the virion gpl20 concentration at high sCD4 concentrations.
disrupted RF at 1.1 nM, the same (± 2-fold) as to recombinant gpl 20 in the presence or absence of detergent (data not shown, but see our previous study).21 Because the gpl20 concentration in the assays is low (typically 0.03 nM) compared with the input sCD4 concentrations, the half-maximal binding values approx¬ imate well to Kdiss. sCD4 binding to intact virions at 37°C showed a maximum at 22 nM sCD4 and a subsequent decline at higher sCD4 concentrations as gpl20-sCD4 complexes dissoci¬
445
ated from the virions (there was a corresponding decrease in recovery of virion-bound gpl20 from these samples, data not shown but see Fig. 4b). Half-maximal binding was estimated to be at 10 nM sCD4 in the experiment shown, assuming saturation was at the same OD492 value as occurred at 4°C. When the binding reaction was carried out at 4°C where no gp 120 dissoci¬ ates from the virions, half-maximal binding was found at 40 nM (Fig. 1), consistent with our previous report.21 In many similar experiments on the same and different RF stocks, the halfmaximal binding concentration at 37°C varied from 7-15 nM and at 4°C from 35-55 nM (data not shown). Differences in sCD4 binding curves for detergent disrupted HIV-1 IIIB virions and intact IIIB virions at 4CC and 37°C were also clearly observable, with the shifts in the curves being similar to those described for RF (data not shown). Control experiments with known concentrations of sCD4 indicated that the extent of separation of free sCD4 from virion-bound sCD4 during gel filtration was about 200-fold. At input sCD4 concentrations greater than approximately 100 nM the free sCD4 concentration in the virion fraction therefore, is, sufficient to bind significantly to virion gpl20 after detergent disruption (see Fig. 1), although the extent to which this occurs is difficult to estimate. This factor cannot, however, account for the temperature dependence of the affinity of sCD4 for virions. The mechanism underlying the virion-sCD4 affinity differ¬ ence at 4°C and 37°C was investigated by comparing the rates of sCD4 binding to virions at the two temperatures. The initial rate of sCD4 binding to RF virions was approximately proportional to the initial concentration of sCD4 at both 4°C and 37°C (Fig. 2). This is compatible with pseudo-first-order kinetics, which would be expected since sCD4 is in high molar excess. At 37CC, maximal sCD4 binding to virions with subsaturating sCD4 concentrations (
and Kinetic Analysis of sCD4 Binding to HIV-1 Virions and of gpl20 Dissociation
Thermodynamic
JOHN P.
MOORE' and
PER JOHAN
KLASSE1"2
ABSTRACT Kinetic and thermodynamic aspects of the binding of sCD4 to intact virions of human immunodeficiency virus type 1 (HIV-1 RF), and of the subsequent induction of gpl20 dissociation were studied. sCD4 binding to virions at 4 and 37°C is half-maximal at approximately 40 and 10 nM, respectively. The transition between low-affinity and high-affinity binding of sCD4 to virions occurs over a narrow temperature range between 20 and 25°C. Shedding of gpl20 from virions after sCD4 binding is also temperature dependent, being initiated above approximately 20°C. The minimum temperatures for the sCD4 affinity transition and gpl20 shedding are, therefore, similar and we suggest how the two processes might be related mechanistically.
INTRODUCTION
ONEmunodeficiency
treating human im¬ virus (HIV) infection is to inhibit virus binding to CD4+ cells. Derivatives of the viral receptor, CD4,1"'3 rendered soluble by a truncation that removes the transmembrane and cytoplasmic region of the molecule (sCD4), have been expressed.48 Bivalent or multivalent immunoglobulin chimeras (sCD4-IgG and sCD4-IgM) of sCD4 also have been produced, with significantly improved antiviral properties in vitro.9~" sCD4 binds with high affinity to recombinant gp 120 or virion-derived soluble gp 120 (HIV-1 s SU glycoprotein) from laboratory-adapted isolates,4"l2~'4 and also potently neutralizes these strains of HIV-1 in vi¬ tro. 4-8-15-'7 However, isolates of HIV-1 not adapted to growth in transformed cell lines (primary isolates) require much higher concentrations of HIV for neutralization in vitro;18 this may compromise sCD4's therapeutic efficiency in HIV infec¬ pharmacological approach to
'
tion.19·20
It has been reported that neutralization of HIV-1 (IIIB and RF) in vitro occurred by a combination of mechanisms: competitive inhibition of virus cell binding at low sCD4 concentra¬ tions;16"17'21 and facilitated dissociation of gpl20 from virions at higher sCD4 concentrations,2123 where synergistic neutral¬ ization occurs.17'21"24 In principle, variation in the efficiency of either sCD4 binding to virions or in its induction of gpl20
dissociation could account for the relative resistance of primary isolates to sCD4.18 Indeed, sCD4-insensitive escape mutants have been isolated with both of these properties.25 Prior to analyzing primary isolates, we sought to increase our under¬ standing of the mechanisms by which sCD4 binds to a wellcharacterized tissue culture-adapted strain of HIV-1 (RF) and induces gpl20 dissociation. We reported previously that both sCD4 binding and gpl20 shedding were temperature-dependent reactions, in that there were major differences at 4°C compared with 37°C.21 These observations implied that conformational changes in the HIV envelope post-sCD4 binding were enthalpically driven. Such changes might be part of the mechanisms by which HIV fuses with CD4+ cells.26 To evaluate these conformational changes more thoroughly, we examined in detail how sCD4 binding and gpl20 shedding were influenced by temperature. We confirmed that the affinity of sCD4 for virions increased with temperature, but found that the greatest increase in affinity occurred over a narrow temperature range between 20 and 25°C. sCD4-facilitated dissociation of gpl 20 from virions was a function both of the sCD4 concentration and of temperature; the minimum temperature for dissociation at saturating sCD4 concentrations was approximately 20°C, similar to that for the sCD4 affinity change. We suggest that a larger CD4-interactive surface is exposed on gpl20 at the higher temperatures, and that this is concomitant with a decreased interaction of gpl 20 with gp41.
Chester Beatty Laboratories, The Institute of Cancer Research, 237 Fulham Road, London SW3 6JB, England. 2Section of Virology, Department of Medical Microbiology, University of Lund, Lund, Sweden.
443
444
MOORE AND KLASSE
MATERIALS AND METHODS Viruses Extracellular HIV-1 RF (RF) was prepared from H9 two viral harvests were used which did not vary significantly in their responses to sCD4. Most experi¬ ments have since been repeated on additional RF harvests with similar results. Harvest 1 contained 150 ng/ml gpl20 and 3,000 ng/ml p24, of which 70% and 74% were virion bound, the rest soluble, when analyzed by Sephacryl S-1000 gel filtration chromatography.21-23-27 For harvest 2, the gpl20 and p24 contents were 90 and 1500 ng/ml with 60% and 65% of the totals present as virion-bound antigen. Infectious virus titers (log TCID50/ml) for harvests 1 and 2 were 7.0 and 6.5, respectively, assessed by syncytium formation in C8166 cells.27
cells.21"27 In this study
Binding of sCD4
to
virions and
detergent-disrupted
virions RF (1-3 ng total gpl20) was incubated with sCD4 in 100 µ RPMI/10% fetal calf serum (FCS) for 1 h at 4 or 37°C (or as indicated in the text) before gel filtration using Sephacryl S-1000 as previously described.21"23'27 Elution was with ice-cold Trisbuffered saline (TBS) or RPMI-1640 culture medium, and the columns were precooled on ice. Similar data were obtained when the columns were run at room temperature. The fraction (550 µ ) containing virions was collected and disrupted with 60 µ of 10% Nonidet P40 (NP40) plus 10% fetal calf serum and immediately frozen unless indicated otherwise. In some exper¬ iments, the soluble antigen fraction (1.2 ml) was collected and processed in similar manner. For determination of sCD4 binding to detergent-disrupted virions,21 RF was mixed with 1% NP40 and incubated for 1 h at 37°C with sCD4 in a total volume of 100 µ , then diluted with 510 µ of 1% NP40/FCS to mimic the dilution of virions during gel filtration, and frozen. The amount of RF in the incubation was adjusted to give a gpl20 concentration that was the same as that in the detergent-solubilized virion fraction eluted from the columns (typically 1-3 ng/ml or 0.01-0.03 nM).
Determination
of sCD4-gpl20 complexes
µ ) of gel-filtered virion-sCD4 com¬ captured onto a solid phase by sheep antibody D7324 to the carboxyl-terminus of gp 120, in the presence of 2% nonfat milk powder.14,21 (Regions close to the carboxyl-termi¬ nus of gpl20 normally are involved in gp41 binding, ref. 28.) NP40 solutions (100
plexes
were
Bound sCD4 was detected with a rabbit anti-sCD4 antiserum (CBL-34) and a sheep anti-rabbit Ig alkaline phosphatase con¬ jugate. 14-21 NP40 solutions of virions and sCD4 (containing free sCD4) were captured onto the solid phase in the same way, except that after the 2 h capture stage unbound sCD4 was washed away with TBS and the wells were incubated for 2 h with 100 µ of TBS/1% NP40/1% FCS to allow dissociation of
sCD4-gpl20 complexes. This was negligible.21 sCD4-gpl20 complex data are presented as OD492
values derived from the above assay. This is because the level of the signal varies with both the gpl20 concentration and the amount of sCD4, complicating calibrations. However, in most experi¬ ments the gpl20 concentration in the immunoassay wells was
kept constant, so variation in signal represents variation in the amount of gpl20-bound sCD4. Under these conditions, the OD4y2 signal in the range we analyze is expected to be in direct proportion to the amount of complexed sCD4. Where the gpl20 concentration was varied (e.g., because of prior shedding), OD492 changes of 0.10 in the range 0.20-0.70 correspond to approximately 1.5-fold differences in gpl20 concentration. A plot of OD492 as a function of log(gpl20) is nearly linear in this OD4y2 range. All enzyme-linked immunosorbent assay (ELISA) determina¬ tions were carried out in triplicate, and data were presented as means ± SD. In figures where SD bars are shown, the omission of a bar means that it lies within the symbol. In figures where SD bars are omitted for clarity of presentation, their magnitude is similar to that shown in other figures describing experiments of similar design. Determination
of gpl20 and p24
Concentrations of gpl20 and p24 in virion extracts were determined by twin-site ELISA as previously described,21"27 using CHO-derived gpl20 (IIIB)29 and baculovirus-expressed p24 (IIIB)30 as calibrants. The volumes assayed were adjusted where possible to be on the linear portion of the calibration curves (typically 100 pgof gp 120 or p24 per well, giving OD492 signals of about 0.40). The sCD4 present in column-purified sCD4-gpl20 complexes did not interfere significantly with either capture or detection of gpl20 in the assays used. High concentrations of free sCD4 do cause a small, nonspecific assay interference that complicates precise determination ( ± 1.5-fold) of gpl20 in the soluble antigen fraction of gel filtration eluates. Within this factor, the sum of virion-bound and soluble gpl 20 in the column eluates was constant, indicating that NP40 had efficiently dissociated gpl20 from gp41. Unless otherwise indicated, virion p24 levels were not significantly affected by any of the treatments described, indicating that virion recovery
during chromatography was constant. RESULTS We have described the binding of sCD4 to HIV-1 virions and its induction of gpl20 dissociation.21"23 To analyze the thermo¬ dynamics of this process, it was necessary to obtain virus preparations with a high level of virion-bound gpl20, such that the amounts of gpl 20 that remained after incubation with sCD4 at 37°C gave significant signals in our immunoassay for gpl 20sCD4 complexes. We obtained virus preparations that contained up to threefold more gpl20 than previously,2'-23 yet still retained a high percentage of their gpl20. In our experience there is considerable variation in virion gpl20 retention in culture,27 but in general RF is the most stable of the commonly used laboratory-adapted HIV-1 isolates (cf., LAV-1 /IIIB, SF-2, MN). The particular harvests we analyzed in the present study retained their gpl20 extremely well. In the absence of sCD4 there was no loss of gp 120 over 4 h at 37°C within the detection limit of our assay systems (i.e., ± 15% approximately) (data not
shown). The binding of sCD4 to intact and detergent-disrupted RF virions differed significantly, as shown by the sCD4 titration curves (Fig. 1). sCD4 bound half-maximally to detergent-
THERMODYNAMICS AND KINETICS OF sCD4 BINDING TO HIV
sCD4-gp120 complexes
OD.92 0-3
y
0-2 0-1
./
.y
»«- "
FIG. 1. Binding of sCD4 to intact and detergent-disrupted RF virions. Intact RF virions were incubated with the sCD4 concen¬ trations indicated at 37°C (·) or 4°C (o) before isolation of the virion-sCD4 complexes by gel filtration and determination of sCD4-gp 120 complexes after disruption of the virions in NP40. For comparison, virions separated from soluble gpl20 by gel filtration ( ), or unfractionated virions ( ), were disrupted in 1 % NP40 then incubated with the sCD4 concentrations indicated before determination of sCD4-gpl20 complexes. No sCD4 assay background, common for all curves ( ). Note that for all the curves bar (·), the gpl20 concentration in the immunoassay wells was constant. In the latter curve, gpl20 shedding causes a reduction in the virion gpl20 concentration at high sCD4 concentrations.
disrupted RF at 1.1 nM, the same (± 2-fold) as to recombinant gpl 20 in the presence or absence of detergent (data not shown, but see our previous study).21 Because the gpl20 concentration in the assays is low (typically 0.03 nM) compared with the input sCD4 concentrations, the half-maximal binding values approx¬ imate well to Kdiss. sCD4 binding to intact virions at 37°C showed a maximum at 22 nM sCD4 and a subsequent decline at higher sCD4 concentrations as gpl20-sCD4 complexes dissoci¬
445
ated from the virions (there was a corresponding decrease in recovery of virion-bound gpl20 from these samples, data not shown but see Fig. 4b). Half-maximal binding was estimated to be at 10 nM sCD4 in the experiment shown, assuming saturation was at the same OD492 value as occurred at 4°C. When the binding reaction was carried out at 4°C where no gp 120 dissoci¬ ates from the virions, half-maximal binding was found at 40 nM (Fig. 1), consistent with our previous report.21 In many similar experiments on the same and different RF stocks, the halfmaximal binding concentration at 37°C varied from 7-15 nM and at 4°C from 35-55 nM (data not shown). Differences in sCD4 binding curves for detergent disrupted HIV-1 IIIB virions and intact IIIB virions at 4CC and 37°C were also clearly observable, with the shifts in the curves being similar to those described for RF (data not shown). Control experiments with known concentrations of sCD4 indicated that the extent of separation of free sCD4 from virion-bound sCD4 during gel filtration was about 200-fold. At input sCD4 concentrations greater than approximately 100 nM the free sCD4 concentration in the virion fraction therefore, is, sufficient to bind significantly to virion gpl20 after detergent disruption (see Fig. 1), although the extent to which this occurs is difficult to estimate. This factor cannot, however, account for the temperature dependence of the affinity of sCD4 for virions. The mechanism underlying the virion-sCD4 affinity differ¬ ence at 4°C and 37°C was investigated by comparing the rates of sCD4 binding to virions at the two temperatures. The initial rate of sCD4 binding to RF virions was approximately proportional to the initial concentration of sCD4 at both 4°C and 37°C (Fig. 2). This is compatible with pseudo-first-order kinetics, which would be expected since sCD4 is in high molar excess. At 37CC, maximal sCD4 binding to virions with subsaturating sCD4 concentrations (
