Envelope glycoproteins from biologically diverse isolates of ...

Proc. Nati. Acad. Sci. USA Vol. 88, pp. 512-516, January 1991 Biochemistry

Envelope glycoproteins from biologically diverse isolates of immunodeficiency viruses have widely different affinities for CD4 (acquired immunodeficiency syndrome/human immunodefidency virus/simian immunodeficiency virus/gpl20/heterologous gene expression)

MONA IVEY-HOYLEtt, JEFFREY S. CULPt, MARGERY A. CHAIKINt, BRIAN D. HELLMIGt, THOMAS J. MATTHEWS§, RAYMOND W. SWEETt, AND MARTIN ROSENBERGt tDepartments of Gene Expression Sciences, Molecular Genetics, and Protein Biochemistry, SmithKline Beecham Pharmaceuticals, King of Prussia, PA 19406; and §Department of Surgery, Duke University Medical Center, Durham, NC 27710 Communicated by Aaron J. Shatkin, October 15, 1990 (received for review August 22, 1990)

ABSTRACT The envelope glycoprotein gp120 of primate immunodeficiency viruses initiates viral attachment to CD4' cells by binding to the CD4 antigen on host cell surfaces. However, among different CD4' cell types, different viruses display distinct host cell ranges and cytopathicities. Determinants for both of these biological properties have been mapped to the env gene. We have quantitatively compared the CD4 binding affinities of gp120 proteins from viruses exhibiting different host cell tropisms and cytopathicities. The viral proteins were produced by using a Drosophila cell expression system and were purified to >90% homogeneity. Drosophilaproduced gp120 from T-cell tropic human immunodeficiency virus type 1 (HIV-1) BH10 exhibits binding to soluble recombinant CD4 (sCD4) and syncytia inhibition potency identical to that of pure authentic viral gpl20. Relative to the affinity of HIV-1 BH10 gp120 for sCD4, that of dual tropic HIV-1 Ba-L is 6-fold lower, that of restricted T-cell tropic simian immunodeficiency virus mac is 70-fold lower, and that of noncytopathic HIV-2 ST is >280-fold lower. Thus, viruses that utilize CD4 for infection do so by using a remarkably wide range of envelope affinities. These differences in affinity may play a role in determining cell tropism and cytopathicity.

Substitution of a single amino acid in gpl20 or of entire env gene sequences has been shown to alter viral tropisms (9, 10). The envelope gene has also been implicated in the tropism of viruses that show a restricted host range among CD4' T-cell lines. For example, SIVmac infects H9 and HUT 78 cells efficiently and SupT1 cells poorly but is unable to infect Molt-4, CEM, or Jurkat cells (11, 12). The fact that human cells expressing recombinant SIVmac envelope protein on their surface show the same selective ability to fuse with these T-cell lines suggests that this form of cellular tropism is directed by the viral envelope protein (12). The envelope protein is also a determinant of the cytopathic properties of HIV and SIV. Expression of recombinant envelope proteins from cytopathic viruses on human cells induces CD4-dependent cell fusion and cell death similar to that observed with virus-infected cells (12, 13). In contrast, human cells expressing recombinant envelope protein from noncytopathic, nonfusogenic HIV-2 ST fail to fuse with CD4' cells (13). Determinants of cytopathicity have also been mapped to the env gene by exchanging equivalent regions between cloned HIV-1 isolates having different cytopathic properties (14). The above observations identify a central role for the envelope protein in determining the tropic and cytopathic properties of HIV and SIV. To date, however, the aspects of envelope function responsible for these different biological properties have not been elucidated. One potential difference among the envelope proteins of these viruses is the recognition and interaction of the gp120 subunit with the CD4 receptor. It has been demonstrated that differing amounts of a soluble recombinant CD4 protein are required to inhibit infection by various HIV-1, HIV-2, and SIV isolates (15, 16). One interpretation of these results is that the envelope proteins of these viruses have different intrinsic affinities for human CD4. However, the variation in sensitivity to soluble CD4 (sCD4) may also reflect different densities of gpl20 expressed on the viral surfaces or differences in the ability of the envelopes to participate in additional cell-surface interactions. We wished to isolate one aspect of envelope protein function-namely, the binding of gpl20 to CD4-and determine quantitatively whether there are differences in the CD4 binding affinities of gpl20s from biologically diverse viral isolates. We expressed gpl20 proteins in a heterologous system, isolated these proteins at >90% purity, and measured their binding to purified sCD4. We found that the gp120 proteins from the T-cell tropic HIV-1 BH10, the dual tropic

The envelope glycoproteins of primate immunodeficiency viruses are derived from a precursor that is cleaved to form the external envelope glycoprotein gpl20 and a transmembrane glycoprotein gp4l (1). The gp4l protein is thought to play a role in membrane fusion in addition to its role as an anchor for gp120 at the surface of the virus or infected cell (2). The gpl20 protein is involved in attachment of virus to CD4' cells by virtue of its ability to bind the CD4 receptor on host cell surfaces (3-5). The CD4 molecule serves as the primary receptor for human immunodeficiency virus (HIV)-1, HIV-2, and simian immunodeficiency virus (SIV), despite variations as high as 70%o in the gp120 amino acid sequences of different isolates (3-5). The presence of several conserved regions interspersed throughout the gpl20 coding regions is thought to contribute to the maintenance of a discontinuous binding site for CD4 (6). Mutational analyses have implicated a number of these conserved regions in the recognition of CD4 (2, 6, 7). The role of gpl20 in determining cellular tropism is more complicated than simply targeting virus to CD4+ cells, as evidenced by the fact that viruses display different host cell ranges among CD4+ cell types. For example, the HIV-1 isolate Ba-L infects T cells as well as monocytes and macrophages efficiently, whereas HIV-1 BH10 can only infect T cells efficiently (8). Determinants of tropism for T cell vs. monocytic cell types have been mapped to the env gene.

Abbreviations: HIV-1 and HIV-2, human immunodeficiency virus types 1 and 2, respectively; SIV, simian immunodeficiency virus; sCD4, soluble CD4; Mt. Drosophila metallothionein promoter; tPA, tissue plasminogen activator. tTo whom reprint requests should be addressed.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

512

Biochemistry: Ivey-Hoyle et al. HIV-1 Ba-L, the noncytopathic HIV-2 ST, and the selective T-cell tropic SIVmac differ markedly in their binding affinity for sCD4. The potential relationships between these affinities and the different biological properties of the viruses are discussed.

MATERIALS AND METHODS Vector Construction. Plasmids are pBR322 based and contain the Drosophila metallothionein promoter (Mt), the signal sequence from the human tissue plasminogen activator (tPA) gene fused in-frame with gp120 sequences, and the simian virus 40 early polyadenylylation signal (17). For pMtBH10, gp120 sequences begin with amino acid residue 32 of mature HIV-1 BH10 gp120 and end at the cleavage site between gp120 and gp41 by the introduction of a stop codon (nucleotides 5986-7335 of ref. 18). Cleavage of the tPA signal sequence results in a secreted gpl20 molecule whose N-terminal 31 amino acids are replaced by 4 amino acids from tPA (see Fig. 1). pMtBaL contains gp120 coding sequences from an envelope clone of HIV-1 Ba-L (M. Popovic, unpublished sequence). pMtST contains HIV-2 ST envelope sequences (nucleotides 6293-7663 of isolate JSP4-27; ref. 19). pMtSIV contains envelope sequences (corresponding to nucleotides 6237-7673 of ref. 20) from SIVmac (isolate BK-28; ref. 21). Cell Culture. Drosophila melanogaster Schneider 2 cell culture and generation of stably transformed cell lines have been described (17). The Mt promoter was induced by the addition of 0.5 mM CuSO4 to the medium for 7 days. Protein Analysis and Purification. Proteins were separated on SDS/10% polyacrylamide gels, and gp120 proteins were detected by Western blot analysis as described (17). Silver staining was performed with the Accurate silver stain kit (Accurate Chemicals, Westbury, NY). Viral HIV-1 BH10 gpl20 was purified from HTLV-IIIBinfected H9 cells by immunoaffinity chromatography as described (22). HIV-1 BH10 gpl20 produced in Drosophila was also purified by immunoaffinity chromatography as described (26). HIV-1 Ba-L gp120 was purified by immunoaffinity chromatography (22) or by chromatography using S-Sepharose, lentil lectin Sepharose, Q-Sepharose, and Superose 12. HIV-2 ST and SIVmac gpl20s were purified by the latter method only. Protein concentration was determined by BCA protein assay (Pierce). N-terminal sequence analysis was performed with an Applied Biosystems gas phase sequencer. Syncytia Assay. The ability of purified gpl20s to block cell fusion was tested as described (22), except that 7 x 104 Molt-4 cells were used instead of 7.5 x 104 CEM cells. sCD4 Binding Analysis. Reactions were performed in triplicate and the data were averaged. For direct binding assays (see Fig. 3A), 125I-labeled sCD4 (107 dpm/,ug; ref. 23) was incubated with 0.1 pmol of gpl20 in 0.1 ml of buffer (0.1% Nonidet P-40/0.25% nonfat dry milk in phosphate-buffered saline) for 3 hr at 4°C. Complexes were immunoprecipitated with anti-gpl20 monoclonal antibody 178.1 and protein ASepharose (CL4B; Pharmacia). Alternatively (see Fig. 4), 0.07 pmol of 1251-labeled sCD4 was incubated with 0.01-10 pmol of gpl20 for 3 hr at 22°C and immunoprecipitated with rabbit polyclonal antibody to HIV-1 gpl20. For competition assays (see Fig. 5A), 0.24 pmol of 1251-labeled BH1O gpl20 (107 dpm/,ug), various amounts of gpl20 competitors, and 0.30 pmol of sCD4 were incubated for 3 hr at 22°C. Complexes were immunoprecipitated with OKT4 antibody and protein A-Sepharose. The Kd of binding was determined as described (24). lodinations were carried out with BoltonHunter reagent (ICN).

Proc. Natl. Acad. Sci. USA 88 (1991)

513

RESULTS Comparison of HIV-1 BH10 gpl20 Produced in Drosophila with Viral gp120. We wished to obtain sufficient quantities of gpl20 from a number of viral isolates to compare their interaction with the human CD4 receptor. Previous results showed that a secreted form of HIV-1 BH10 gpl20 could be efficiently expressed in Drosophila cells by using the Mt promoter to control transcription and the tPA signal sequence to effect secretion (26). The construct used, pMtBH10, contains envelope sequences encoding amino acid residue 32 (Asp) through the C terminus of mature HIV-1 BH10 gpl20 fused in-frame to the tPA signal sequence (Fig. 1). Drosophila cells stably transfected with pMtBH10 secreted a 110-kDa envelope protein, which we designate gp120* (we use the term gpl20* as a designation for the recombinant envelope proteins shown in Fig. 1, rather than a description of the protein's molecular weight), at 1-2 mg per liter of medium upon induction of the Mt promoter (Fig. 2A). To determine whether our recombinant system affected the activity ofBH10 envelope protein, the recombinant Drosophila BH10 gpl20* was purified to >90% homogeneity (Fig. 2B) and compared to purified authentic viral BH10 gpl20 in two assays. In the first assay, saturation binding analysis was performed with sCD4 (23). A fixed amount of purified gpl20 was incubated with increasing amounts of 1251-labeled sCD4, and bound complexes were immunoprecipitated by using an antibody to gpl20 (Fig. 3A). The two preparations of BH10 envelope proteins gave identical results. Scatchard analysis of the binding data (Fig. 3A Inset) revealed that Drosophila gpl20* bound to sCD4 with the same affinity (Kd 2.4 nM) as did viral gpl20 (Kd 2.9 nM). In a second assay, the ability of the envelope proteins to inhibit syncytia formation between HIV-1 IIIB-infected cells and CEM cells was measured. The Drosophila gpl20* inhibited syncytia formation with the same efficiency as the viral material (Fig. 3B). These results indicate that the gp120* produced in Drosophila cells recognizes CD4 in a manner analogous to that of native viral gpl20. Thus, the first 31 amino acids of gpl20 are dispensable for sCD4 binding, consistent with previous reports (7). Expression and Purification of Other gp120 Proteins. Since the Drosophila system is capable of expressing envelope protein that exhibits appropriate receptor recognition, we utilized the system to produce a number of gp120* proteins from distinct viral isolates for quantitative comparisons of CD4 binding. We selected HIV-1 Ba-L, which differs from BH10 in its cellular tropism (8), HIV-2 ST, which is noncytopathic and infects only a subset of T-cell lines (19), and a

|MtjtP/

gpl2O

lAni

pMtBH1 0 GARS/DTEV-442 a. a. -REKR pMtBaL GARSY/AYDTEV-441 a. a. -REKR pMtST GARS/DTWG-449 a. a. -RNKR pMtSIV GARS/DTWG-470 a . a. -RNKR FiG. 1. Constructs for gpl20 expression. The expression unit for gpl2O* is shown. Diagonal line represents an in-frame fusion between the first 36 amino acids of human tPA (signal sequence) and gpl20 sequences beginning at or near a conserved Asp-Thr (DT) sequence (underlined). A,, simian virus 40 early polyadenylylation signal. Secreted products of the tPA-gpl20 fusions encoded by the plasmids listed are represented by the single-letter amino acid code. N-terminal amino acids shown were determined by sequence analysis of the purified proteins and agree with the predicted sequence. In each plasmid, a stop codon was introduced to generate a C terminus identical to that found for the corresponding viral gpl20. The number of amino acids not shown is indicated. A slash separates amino acids derived from the human tPA gene from those derived from the gpl20 genes.

Biochemistry: Ivey-Hoyle et al.

514

A

/co

sIV

HIV-2

HIVCI -40

I

Proc. Natl. Acad Sci. USA 88 (1991)

AC,

- 210

-107

AM

_m

-69

-46

-29

B 0-4

0

0-4

CO

mbc

u 0

fV /

C-

f I21 Z

0

Cs,

--.I A-

' M

-107 - 69

46 -29

FIG. 2. (A) Detection of gpl2O* proteins by Western blot analysis. Culture supernatants from Drosophila cells transfected with pMtBH10 (BH10), pMtBaL (BaL), pMtST (ST), pMtSIV (mac), or nontransfected cells (-) were analyzed by using rabbit polyclonal antibody to BH10 gpl2O* (HIV-1) or to HIV-2 ST gp120 (HIV-2) or serum from SIV-infected macaques (SIV). (B) Silver stain of purified Drosophila gpl2O*sO(O.5 ,ug).

SIVmac isolate, which shows restricted tropism among CD4+ T-cell lines (21). Vectors analogous to pMtBH10 were constructed (Fig. 1) by using the gp120 coding regions from isolated clones of HIV-1 Ba-L (pMtBaL), HIV-2 ST (pMtST), and SIVmac (pMtSIV). Drosophila cells were stably transfected with these constructs, and upon induction of the Mt promoter, the appropriate gpl2O* was produced and secreted at levels ranging from 5 to 35 mg per liter of medium (Fig. 2A). Each gp120* was purified by standard chromatographic techniques involving ion exchange, lentil lectin, and size-exclusion chromatography to >90o homo-

geneity (Fig. 2B). The Ba-L gp120* was also separately purified by immunoaffinity chromatography (BaL-I; Fig. 2B) as was the HIV-1 BH10 gp120* to compare the effects of the different purification procedures on binding activity. N-terminal sequence analysis of the purified proteins confirmed the predicted primary structure in all cases (Fig, 1)_ of the HIV-1 BHIO and Ba-L gpl2W* Proteins. The purified BH10 and Ba-L gp120* proteins were compared for their ability to bind sCD4. In one assay, a fixed amount of '25I-labeled sCD4 was incubated with increasing amounts of gpl2O*, and bound complexes were immunoprecipitated with antibody reactive to both gp120*s (Fig. 4). Ba-L gp120*, purified either by immunoaffinity chromatography (BaL-I) or by standard chromatographic procedures (BaL-C), did not bind sCD4 as efficiently as BH10 gpl2O*. This reduction in binding did not result from inactivation of Ba-L gp120* during purification since similar results were observed when conditioned medium samples were used as a source of gpl2O* (data not shown). Thus, the affinity of Ba-L gp120* for sCD4 is significantly lower than that of BH10 gp120*. However, this reduced affinity precluded a determination of the Kd from the direct binding data shown in Fig. 4. The affinity of the Ba-L gpl2O* for sCD4 was quantitated by measuring its ability to compete with the binding of 125I-labeled BH10 gpl2o* to sCD4 (Fig. 5A). Consistent with the direct binding experiments described above, Ba-L gp120* did not compete as efficiently as BH10 gp120* itself. The Kd values for the interaction of the gp120*s with sCD4 were calculated from the IC50 values measured in the competition assay (Table 1). The Kd determined for BH10 gp120* by this method, 5 nM, is in excellent agreement with that determined from direct saturation binding experiments (Fig. 3A; ref. 7). The Kd determined for Ba-L gpl20*, 30 nM, reflects a 6-fold lower affinity than that of BH10 gpl20*. The purified BH10 and Ba-L gp120*s were also compared in a biological assay that measured the ability of the gp120*s to inhibit syncytia formation between HIV-1 IIIB-infected cells and CD4' T cells (Fig. 5B). Consistent with the 6-fold lower affinity of Ba-L gp12O* seen in the competition assay (Fig. 5A), we find that purified Ba-L gpl2o* was 5- to 10-fold less effective than BH10 gp120* in blocking cell fusion. Characterization of HIV-2 ST aid SlVmac gpl2W* Proteins. Attempts to immunoprecipitate the recombinant ST and SIV gpl2O* proteins from crude medium samples with excess sCD4 suggested a low binding affinity for SIV gpl2O* and revealed almost undetectable binding for ST gp12O* (data not shown). Thus, the affinities of these gpl2O* proteins for sCD4

0. 12

A

B

-4 -

a) ,-4

0 04

-H4 >1

.0 .0 0 Ci)

U)

0

z

0 0

2

4

6 8 sCD4 (nM)

10

12

1

2

gpl2O

3

4

5

(gg/ml)

FIG. 3. Comparison of purified Drosophila and viral BH10 gpl20s. (A) Saturation binding of 1251-labeled sCD4 to purified gpl20s. (Inset) Scatchard plot of the binding curve. *, Drosophila BH10 gpl2O*; *, viral BH10 gp120. (B) Inhibition of syncytia formation by Drosophila BH10 gpl2O* (A) or viral BH10 gpl20 (i).

Proc. Natl. Acad. Sci. USA 88 (1991)


Table 1. Determination of dissociation constants Viral isolate IC50, nM Kd, nM Relative Kdt HIV-1 BH10 7 5 1 HIV-1 Ba-L 45 30 6 SIVmac 500 350 70 HIV-2 ST >2000 >1400 >280 tThe Kd for BH10 gpl2O* is assigned a value of 1.

0. 04 -

n 0.03

~0 i : 0.02 I

0

IQ u

0.01

0.00 0

20

40 60 80 gpl20 (nM)

100

FIG. 4. Saturation binding of 1251I-labeled sCD4 to purified Drosophila BH10 and BaL gp120*s. *, BH10 gpl20*; *, BaL-I gp120*; *, BaL-C gpl20*.

were also quantitated in the competition experiment described above (Fig. SA). The SIV gp120* displayed a weak ability to compete with the binding of BH10 gp120* to sCD4, and the ST gpl20* was an extremely poor competitor. Relative to the affinity of BH10 gpl20* for sCD4, the affinities of the SIVmac and HIV-2 ST proteins were about 70-fold and 280-fold lower, respectively (Table 1). Our results clearly demonstrate a wide range of sCD4 binding abilities among the envelope proteins of these HIV and SIV isolates.

DISCUSSION Using a Drosophila cell expression system, we have produced recombinant envelope proteins (gpl2o*) from HIV-1, HIV-2, and SIV isolates and compared the affinities of the purified proteins for a purified recombinant form of the human CD4 receptor. The Drosophila-produced gpl20* from the T-cell tropic isolate HIV-1 BH10 bound sCD4 with an affinity equivalent to that of authentic viral gpl20 and in good agreement with other reports (7, 25). The gp120* of HIV-1 Ba-L, which exhibits a dual tropism for T cells and monocytes-macrophages, bound with a 6-fold lower affinity than BH10 gpl20*. In addition, SIVmac gpl20* binding was -70-fold weaker, and HIV-2 ST gpl20* binding was at least 280-fold weaker. We conclude that the affinity for sCD4 A

varies markedly for the envelope glycoproteins of these diverse primate immunodeficiency viruses. Although the glycosylation of gpl2o* produced in Drosophila is primarily of the high mannose type, in contrast to the complex glycans observed in mammalian systems, the Drosophila-produced BH10 gpl2O* showed the same activity as viral BH10 gpl20 in both sCD4 binding and syncytia inhibition assays. We cannot rule out that this altered glycosylation did affect the activity of the other gpl2O*s. However, we have recently found that HIV-2 ST gpl20 produced in mammalian cells, like that produced in Drosophila, has a low affinity for sCD4 (M.A.C., R.W.S., and Mark Mulligan, unpublished data). Previous studies have indicated that isolates of HIV-1, HIV-2, and SIV differ in their sensitivity to inhibition by soluble CD4 proteins (15, 16). Of particular relevance to the present study, the level of sCD4 required to inhibit SIVmac was 5- to 20-fold greater than that required to inhibit HIV-1 BH10. This differential sensitivity to sCD4 likely results from the low affinity of SIVmac gpl20 for sCD4. These results do not rule out a contribution by other factors, such as the density of gpl20 on the viral surface. Nonetheless, the viral envelope affinity for sCD4 correlates well with sensitivity to inhibition by sCD4. These results have important consequences for proposed HIV therapies based on interfering with the interaction of envelope and CD4. Therapeutic dose levels predicted on the basis of inhibition of isolates with high-affinity envelopes (e.g., BH10) are likely to be ineffective against isolates having lower CD4 affinities. How do the differences in sCD4 binding affinity of the envelope proteins in our study relate to the different biological properties of the viruses? Among the isolates examined, there is a correlation between the affinities of the viral envelopes for sCD4 and the cytopathic properties of the viruses. HIV-1 BH10, whose envelope exhibits the highest affinity for sCD4, consistently induces cytopathic effects and cell fusion in the T-cell lines that it infects. In contrast,

0.6' 0

0.5'

0.

0.4

10

0 .3 -

515

B

80

'-4

'-4

60

-

".14 >1

U

U '-4

0. 2 o

Ba-L

0

z

0 . 1i

0 .0

0

I1000 gpl2O (nM)

2000

0

10

20

30

gpl20 (gg/ml)

FIG. 5. Comparison of gpl2O*s by blocking experiments. (A) Binding of 1251-labeled BH10 gpl2O* to sCD4 in the presence of gpl2O* competitors. *, BHIO gpl2O*; *, BaL-I gpl2O*; *, SIV gp120*; m, ST gpl2O*. (B) Inhibition of syncytia formation by BH1O gp120* (A) or Ba-L

gpl2O* (m.)

516


SIVmac, whose envelope shows a 70-fold lower affinity, exhibits reduced or no cytopathic effects on some T-cell lines (11, 12). Moreover, HIV-2 ST, which shows the lowest CD4 affinity, is noncytopathic and nonfusogenic (19). For these as well as other isolates, determinants of cytopathic and fusogenic properties have been mapped to their envelope genes (12-14). Our data suggest that one aspect of envelope function involved in determining the cytopathic properties of immunodeficiency viruses may be the affinity of the viral envelope protein for CD4. The envelope gene has also been identified as a determinant of host cell tropism (9, 10, 12). We find that envelope proteins from viruses that are restricted in their host range among CD4' T-cell lines exhibit lower affinities for sCD4. In contrast to HIV-1 BH10, which has a wide host range among different T-cell lines and a high sCD4 affinity, SIVmac and HIV-2 ST infect only a subset of T-cell lines and have considerably lower affinities. The restricted host tropisms of the latter viruses and the observation that cell hybrids between nonpermissive cell types can be infected (11) suggest that accessory molecules in addition to CD4 are present on the susceptible cells and are important for efficient penetration by these viruses (11, 12, 19). If these viruses do require a second interaction with the cell surface, such an additional interaction may alleviate the need for high-affinity binding to CD4. Similarly, HIV-1 Ba-L may be able to infect monocytes-macrophages by participating in a secondary surface interaction crucial for infection of these cells, and this additional interaction may compensate for the lower affinity of the Ba-L envelope for CD4. Envelope proteins from T-cell tropic viruses such as BH10 would presumably lack determinants for this critical interaction on monocytes-macrophages. Although the high affinity of the BH10 envelope for CD4 may not compensate for the lack of monocytemacrophage-specific determinants, it may give the virus a selective advantage for growth on T cells. Thus, distinct determinants within envelope protein for CD4 binding and for additional secondary interactions at the cell surface may play a role in defining host cell range. The existence of such distinct determinants is suggested by mutations in the gp120 region of a HIV-1 isolate that altered tropism for monocytic cells with little effect on CD4 binding (10). In summary, our results reveal a wide range of affinities for sCD4 among envelopes of biologically diverse immunodeficiency viruses that use the CD4 receptor for infection. These results raise the possibility that reduced affinities are related to the involvement of secondary interactions or receptors in the infection of certain cell types, thereby affecting host cell tropism. These differential affinities may also affect the fusogenic and cytopathic properties of viruses. Studies of gp120s from isogenic viral derivatives with altered biological properties should further elucidate the role of CD4 affinity in these processes. We thank Mikulas Popovic for the gift of the HIV-1 Ba-L env gene, Beatrice Hahn for the HIV-2 ST env gene, James Mullins for the SIVmac env gene, Mark Mulligan for antibody to HIV-2 ST envelope protein, Ronald Desrosiers for antisera from SIV-infected macaques, Clotilde Thiriart and Claudine Bruck for monoclonal antibody 178.1, and Lynette Miles for excellent assistance in protein sequencing. This work was supported by Grant A12484502 from the National Institutes of Health.

Proc. Natl. Acad. Sci. USA 88 (1991) 1. McKeating, J. A. & Willey, R. L. (1989) AIDS 3, Suppl., 35s-41s. 2. Kowalski, M., Potz, J., Basiripour, L., Dorman, T., Goh, W. C., Terwilliger, S., Daxton, A., Rosen, C., Haseltine, W. & Sodroski, J. (1987) Science 237, 1351-1355. 3. Dalgleish, A. G., Beverly, P. C. L., Clapham, P. R., Crawford, D. H., Greaves, M. F. & Weiss, R. A. (1984) Nature

(London) 312, 763-766. 4. Klatzmann, D., Champagne, E., Chamaret, S., Gruest, J., Guetard, D., Hercend, T., Gluckman, J.-C. & Montagnier, L. (1984) Nature (London) 312, 767-768. 5. Maddon, P. J., Dalgleish, A. G., McDougal, J. S., Clapham, P. R., Weiss, R. A. & Axel, R. (1986) Cell 47, 333-348. 6. Cordonnier, A., Riviere, Y., Montagnier, L. & Emerman, M. (1989) J. Virol. 63, 4464-4468. 7. Lasky, L. A., Nakamura, G., Smith, D. H., Fennie, C., Shimasaki, C., Patzer, E., Borman, P., Gregory, T. & Capon, D. J. (1987) Cell 50, 975-985. 8. Gartner, S., Markovits, P., Markovitz, D. M., Kaplan, M. H., Gallo, R. C. & Popovic, M. (1986) Science 233, 215-219. 9. Fisher, A. G., Ensoli, B., Looney, D., Rose, A., Gallo, R. C., Saag, M. S., Shaw, G. M., Hahn, B. H. & Wong-Staal, F. (1988) Nature (London) 334, 444-447. 10. Cordonnier, A., Montagnier, L. & Emerman, M. (1989) Nature (London) 340, 571-574. 11. Hoxie, J. A., Haggarty, B. S., Bonser, S. E., Rackowski, J. L., Shan, H. & Kanki, P. J. (1988) J. Virol. 62, 2557-2568. 12. Koenig, S., Hirsch, V. M., Olmsted, R. A., Powell, D., Maury, W., Rabson, A., Fauci, A. S., Purcell, R. H. & Johnson, P. R. (1989) Proc. Natl. Acad. Sci. USA 86, 2443-2447. 13. Mulligan, M. J., Kumar, P., Hui, H., Owens, R. J., Ritter, G. D., Jr., Hahn, B. H. & Compans, R. W. (1990) AIDS Res. Hum. Retroviruses 6, 707-720. 14. York-Higgins, D., Cheng-Mayer, C., Bauer, D., Levy, J. A. & Dina, D. (1990) J. Virol. 64, 4016-4020. 15. Clapham, P. R., Weber, J. N., Whitby, D., McIntosh, K., Dalgleish, A. G., Maddon, P. J., Deen, K. C., Sweet, R. W. & Weiss, R. A. (1989) Nature (London) 337, 368-370. 16. Looney, D. J., Hayashi, S., Nicklas, M., Redfield, R. R., Broder, S., Wong-Staal, F. & Mitsuya, H. (1990) J. Acquired Immune Defic. Syndr. 3, 649-657. 17. Johansen, H., van der Straten, A., Sweet, R., Otto, E., Maroni, G. & Rosenberg, M. (1989) Genes Dev. 3, 882-889. 18. Muesing, M. A., Smith, D. H., Cabradilla, C. D., Benton, C. V., Lasky, L. A. & Capon, D. J. (1985) Nature (London) 313, 450-458. 19. Kumar, P., Hui, H., Kappes, J. C., Haggarty, B. S., Hoxie, J. A., Arya, S. K., Shaw, G. M. & Hahn, B. H. (1990) J. Virol. 64, 890-901. 20. Chakrabarti, L., Guyader, M., Alizon, M., Daniel, M. D., Desrosiers, R. C., Tiollais, P. & Sonigo, P. (1987) Nature (London) 328, 543-547. 21. Kornfeld, H., Riedel, N., Viglianti, G. A., Hirsch, V. & Mullins, J. I. (1987) Nature (London) 326, 610-613. 22. Matthews, T. J., Weinhold, K. J., Lyerly, H. K., Langlois, A. J., Wigzell, H. & Bolognesi, D. P. (1987) Proc. Natl. Acad. Sci. USA 84, 5424-5428. 23. Deen, K. C., McDougal, J. S., Inacker, R., Folena-Wasserman, G., Arthos, J., Rosenberg, J., Maddon, P. J., Axel, R. & Sweet, R. W. (1988) Nature (London) 331, 82-84. 24. DeBlasi, A., O'Reilly, K. & Motulsky, H. J. (1989) Trends Pharmacol. Sci. 10, 227-229. 25. Moore, J. P. (1990) AIDS 4, 297-305. 26. Culp, J. S., Johansen, H., Melling, B., Beck, J., Matthews, T. J., Delers, A. & Rosenberg, M. (1991) BiolTechnology, in press.