Aminooxypentane Addition to the Chemokine Macrophage Inflammatory Protein-1αP Increases Receptor Affinities and HIV Inhibition*

To enter its target cells, human immunodeficiency virus (HIV) must interact with CD4 and one of a family of chemokine receptors. CCR5 is widely used by the virus in this context, and its ligands can prevent HIV entry. Amino-terminal modified chemokine variants, in particular AOP-RANTES (aminooxypentane-linked regulated on activation normal T cell expressed and secreted), exhibit enhanced HIV entry inhibition. We have previously demonstrated that a non-allelic isoform of macrophage inflammatory protein (MIP)-1α, termed MIP-1αP, is the most active naturally occurring inhibitor of HIV entry known. Here we report the properties of a variant of MIP-1αP with an AOP group on the amino terminus. We show that, like RANTES, the addition of AOP to MIP-1αP enhances its interactions with CCR1 and CCR5, allows more effective internalization of CCR5, and increases the ligand's potency as an inhibitor of HIV entry through CCR5. Importantly, AOP-MIP-1αP is about 10-fold more active than AOP-RANTES at inhibiting HIV entry, making it the most effective chemokine-based inhibitor of HIV entry through CCR5 described to date. Surprisingly, the enhanced receptor interactions of AOP-MIP-1αP do not translate into increased chemotaxis or coupling to calcium ion fluxes, suggesting that this protein should be viewed as a partial, rather than a full, agonist for CCR1 and CCR5.

Chemokines are a large family of proteins that play central roles in the chemoattraction of leukocytes during immune surveillance, inflammation, and the establishment of immunity (reviewed in Refs. [1][2][3][4]. As such, they have been implicated in the pathogenesis of many autoimmune and chronic inflammatory diseases. Their biological effects are mediated by a group of heptahelical G-protein-coupled receptors expressed on the surface of their target cells. Interactions between chemokines and their receptors are complex, but in many cases the amino-terminal domain of the chemokine plays an important part in regulating affinity and receptor activation (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16) Thus, by modifying this region of the protein it is possible to generate variants that have enhanced receptor affinity and/or antagonistic activity. These variants have begun to reveal the nature of chemokine/receptor interactions and could be of use in the development of therapies targeting chemokine receptors. Indeed, such variants have been used experimentally to inhibit chemokine receptor activity in animal models of disease (17)(18)(19)(20).
Some chemokine receptors are unusual in that they are also able to interact with proteins outside the chemokine family (21)(22)(23)(24)(25). Of particular interest is the gp120 protein in the envelope of the HIV 1 virus, as this protein uses chemokine receptors, along with CD4, to mediate viral entry into target cells (reviewed in Ref. 25). In vitro, many chemokine receptors are able to act as HIV co-receptors, but in vivo two receptors, CCR5 and CXCR4, appear to be predominantly used. This is starkly demonstrated by the protective effects of polymorphisms in CCR5, its ligands, or the CXCR4 ligand stromalderived factor-1, on the transmission of the virus and/or the development of AIDS (26 -31).
The ligands for these receptors (MIP-1␣, MIP-1␤, monocyte chemoattractant protein-2 (MCP-2) and RANTES for CCR5; stromal-derived factor-1 for CXCR4) can inhibit HIV entry in vitro (25), thought to be due to steric interference and receptor internalization, and likely have a role in regulating HIV infection in vivo (32)(33)(34)(35)(36). Targeting these receptors may help in therapeutic regulation of HIV infection (25), and CCR5 in particular has been the focus of intensive research aimed at identifying small molecules or modified chemokines that can act as HIV entry inhibitors (10,11,37). Of particular relevance here, is an amino-terminal modified form of RANTES (aminooxypentane (AOP)-RANTES) that is markedly more potent than RAN-TES at inhibiting HIV entry through CCR5 (11). It has been proposed that this is due to decreased recycling of CCR5 following AOP-RANTES (compared with RANTES) binding and internalization, the consequence of which is more effective removal of CCR5 from the cell surface (38).
We have recently demonstrated that a non-allelic isoform of MIP-1␣ present in humans, MIP-1␣P (or LD78␤), is the most effective natural CCR5 agonist yet described, and moreover is the most potent natural inhibitor of CCR5-dependent HIV entry identified to date (39). Indeed, MIP-1␣P shows equivalent anti-HIV activity to AOP-RANTES in these assays. Here, we have generated an AOP variant of MIP-1␣P and examined its biological activity in comparison with MIP-1␣P, RANTES, and AOP-RANTES. Our results show that AOP addition to MIP-1␣P enhances its affinity for CCR1 and CCR5. Importantly, we also show that AOP-MIP-1␣P is a more potent inhibitor of HIV entry through CCR5 than either MIP-1␣P or AOP-RANTES. In fact, AOP-MIP-1␣P is the most effective chemokine-based inhibitor of CCR5-mediated HIV entry described to date. Additionally, we present data that suggests that the enhanced affinity of AOP variants for CCR1 and CCR5 is not consistently reflected by enhanced signaling and chemotactic activity, and in fact AOP-MIP-1␣P should be viewed as a partial receptor agonist.
Radioligand Displacement Assays-Chemokines were labeled with 125 I as described elsewhere (42) using IODO-GEN (Pierce, Rockford, IL). 300 nM stock solution of radiolabeled chemokine was stored at 4°C. The derivation of CHO cells stably transfected with human CCR1, CCR5, and D6, and the displacement assays have been described previously (43). Briefly, 2 ϫ 10 5 CHO cells were incubated overnight in each well of a 24-well plate, washed in PBS, then 250 l of complete medium was added containing 20 mM HEPES, 0.5% sodium azide, radioiodinated chemokine, and PBS or a given concentration of unlabeled chemokine in PBS. Final concentrations of 20, 20, and 3 nM 125 I-mMIP-1␣ were used for CCR1, CCR5, and D6 transfectants, respectively. After 1.5-2 h at room temperature, the cells were washed three times with ice-cold PBS, lysed with 0.1% sodium dodecyl sulfate, transferred to a counting vial and counted for 1 min in a Beckman Gamma 5500B counter (Beckman, High Wycombe, UK). For CEM-CCR5 and THP1 cells, 2 ϫ 10 5 cells were transferred to a 1.5-ml Eppendorf tube, washed, and incubated as above. For HEK293 transfectants, cells were harvested by trypsinization, washed, 2 ϫ 10 5 cells transferred to 1.5-ml Eppendorf tubes and incubated as above. For binding assays done in suspension, cells were counted directly without lysis. In all experiments, each point was done in triplicate and each experiment was done at least twice.
Chemotaxis Assay-The ability of THP1 or CEM-CCR5 cells to migrate in response to chemokines was analyzed using a Transwell assay. Cells were labeled for 30 min at 37°C in 1 M Calcein-AM (Molecular Probes, Leiden, The Netherlands) and washed twice in chemotaxis buffer (RPMI 1640, 1% bovine serum albumin, 20 mM HEPES) by centrifugation at 1000 ϫ g for 5 min. The cells were resuspended in buffer at 2 ϫ 10 6 /ml and 300 l was added to 3-m pore HTS Fluoroblok inserts (Becton Dickinson, Le Pont de Claix, France). The inserts were held in 24-well companion plates (Becton Dickinson) containing 700 l of the appropriate chemokine diluted in chemotaxis buffer, and incubated for 1 h 45 min at 37°C. The plates were read on a Fluostar fluorescent plate reader (BMG LabTechnologies, Offenburg, Germany). Chemotaxis seen in the presence of chemotaxis buffer alone was set at 1, and chemotaxis in chemokine-containing wells calculated relative to this control. Each point was done in triplicate, and experiments were performed at least twice.
Virus Internalization Assays-Viral internalization into CEMx174-CCR5 cells was examined using luciferase reporter viruses pseudotyped with envelope glycoprotein of HIV strains JR-FL, ADA, or JC2, or the amphotropic murine leukemia virus. Experiments were performed as described previously (39). Briefly, 3 ϫ 10 4 CEMx174-CCR5 cells were pretreated for 30 min with chemokine in medium containing 4 g/ml Polybrene. Reporter virus was then added (2-3 ng of p24). The next day, medium was changed without adding additional chemokine and 2-4 days later luciferase activity was measured using commercially available reagent (Promega, Inc., Madison, WI). The extent of inhibition of HIV entry was determined by luciferase activity of chemokine-treated cells, with untreated control cells.
Detection of Cell Surface CCR5-5 ϫ 10 5 cells were washed in ice-cold FACS buffer (PBS, 1% fetal calf serum, 0.25% sodium azide), then incubated for 30 min on ice in the presence of either phycoerythrin conjugated anti-human CCR5 (Pharmingen, San Diego, CA) or an appropriate isotype control (Pharmingen) according to the manufacturer's instructions. After two further washing steps the cells were fixed with 4% paraformaldehyde, PBS and analyzed on a FACScan (Becton Dickinson). Control experiments demonstrated that chemokine bound to surface CCR5 did not interfere with binding to the anti-CCR5 antibody (not shown).
The Internalization and Re-emergence of CCR5-To study the internalization of CCR5, CEM-CCR5 cells were incubated at 37°C in the presence of chemokine. At various times following addition of chemokine, 5 ϫ 10 5 cells were removed and processed for FACS as described above. To analyze the re-emergence of CCR5, CEM-CCR5 cells were incubated for 45 min in 100 nM chemokine. The cells were then washed four times in medium at room temperature and further incubated at 37°C. At the desired time points, 5 ϫ 10 5 cells were removed and processed for FACS as described above. As a negative control, cells were incubated for the required times in medium alone. Relative fluorescence intensity was calculated as (mean channel fluorescence [chemokine] Ϫ mean channel fluorescence [negative control])/(mean channel fluorescence [medium alone] Ϫ mean channel fluorescence [negative control]).

AOP Addition Enhances
Interaction with CCR1 and CCR5-To determine whether the addition of AOP to the NH 2 terminus of MIP-1␣P alters its interaction with chemokine receptors, we performed heterologous displacement studies using RANTES, MIP-1␣P, and their AOP derivatives. CHO cells expressing known human receptors for MIP-1␣P and RANTES (CCR1, CCR5, or D6) were incubated with a set concentration of radioiodinated murine (m) MIP-1␣ ( 125 I-mMIP-1␣), and a range of concentrations of unlabeled chemokine as competitor. For CCR1 and CCR5 binding, the addition of the AOP moiety to the NH 2 terminus of either chemokine dramatically decreased the IC 50 , i.e. the concentration at which 50% of the labeled protein was displaced (Fig. 1, A and B). With CCR1 transfectants, AOP-MIP-1␣P had an IC 50 of 7 nM, at least 20-fold lower than MIP-1␣P (Fig. 1A). Despite AOP-RANTES having a relatively high IC 50 (ϳ80 nM), this ligand was much more effective than RANTES in this assay which at the highest concentration tested (400 nM) only displaced 15% of the 125 I-mMIP-1␣ from CCR1. For CCR5 binding similar trends were seen, with the addition of AOP to MIP-1␣P decreasing the IC 50 for 125 I-mMIP-1␣ displacement by 10-fold (20 nM for MIP-1␣P versus 2 nM for AOP-MIP-1␣P). Likewise, AOP-RANTES exhibited more effective displacement of the labeled chemokine from CCR5 than RANTES (Fig. 1B). With D6, however, the AOP variants were less able to displace 125 I-mMIP-1␣ than their unmodified counterparts, reflected by an increase in their IC 50 values of 2-3-fold (Fig. 1C). Thus, AOP addition has a similar effect on RANTES and MIP-1␣P on all three receptors, enhancing displacement of 125 I-mMIP-1␣ from CCR1 and -5, but reducing it from D6. Similar effects were seen with MIP-1␣P or AOP-MIP-1␣P when these proteins themselves were radiolabeled and displaced from receptor by excess of the same, but unlabeled, ligand, i.e. homologous displacement assays (not shown).
Signaling Activity of AOP Chemokine Variants-Our previous work (39) has shown that for variants of MIP-1␣, greater affinity for CCR1 or CCR5 in heterologous displacement studies is associated with more potent chemokine signaling through these receptors, as assessed by the induction of calcium ion fluxes. Thus, it was anticipated that the AOP chemokines would elicit more robust fluxes into cells than the unmodified chemokines. To examine this, HEK293 cells expressing CCR1 or CCR5 were assessed for their ability to mobilize Ca 2ϩ in response to ligand stimulation. Initially, radioligand displacement experiments were done and importantly showed that, as with the CHO cell transfectants above, the AOP variants had enhanced affinity for CCR1 and CCR5 expressed in these cells (not shown). Curiously, however, when Ca 2ϩ flux assays were performed, AOP-MIP-1␣P signaling through CCR1 and CCR5 induced fluxes that were of similar magnitude to those seen with MIP-1␣P (Fig. 2, A and B, fluxes after arrow 1). Additionally, despite its enhanced ability to displace mMIP-1␣, AOP-RANTES showed weaker signaling through CCR1 than an equimolar concentration of RANTES (Fig. 2C). However, AOP- RANTES induced a more robust Ca 2ϩ flux through CCR5 than an equivalent concentration of RANTES (Fig. 2D). These observations were seen across a range of ligand concentrations (not shown). Also, the ability of AOP-MIP-1␣P to block a second signal induced with the unmodified protein was examined. For CCR5, a second flux induced with a 10-fold molar excess of MIP-1␣P was considerably smaller if AOP-MIP-1␣P had been used as the first stimulant (Fig. 2, A and B, fluxes after arrow 2). Likewise, with the CCR1 transfectants ( Fig. 2A), 5 nM AOP-MIP-1␣P completely prevents a detectable second flux induced with 50 nM MIP-1␣P, while 5 nM MIP-1␣P does not. Thus, while the addition of AOP to MIP-1␣P does not cause enhanced coupling to Ca 2ϩ flux induction, increased receptor occupancy appears to result in a more effective block of subsequent MIP-1␣P-induced signals.
Chemotactic Activity of Chemokines and Their AOP Variants-Signals induced by the AOP variants through CCR1 and CCR5 will regulate their chemotactic activity. To examine chemotaxis, MIP-1␣P, RANTES, and their AOP variants were used in Transwell chemotaxis assays to attract CEM-CCR5 cells, or THP1 cells (that express CCR1 but not CCR5) (Fig. 3). While they gave similar profiles, MIP-1␣P consistently appeared to be marginally more active than AOP-MIP-1␣P at attracting CEM-CCR5 cells (Fig. 3A). Thus, 0.1 nM MIP-1␣P in the bottom of the chemotaxis plate attracted significant numbers of cells through the filter, but the same concentration of AOP-MIP-1␣P was ineffective. However, equal numbers of CEM-CCR5 cells were attracted toward 0.3 nM of each chemokine. With THP1 cells, MIP-1␣P and AOP-MIP-1␣P produced near identical dose-response profiles (Fig. 3B). AOP-RANTES, on the other hand, was a more potent chemotactic agent for CEM-CCR5 cells than RANTES (Fig. 3C): considerably more cells migrated when 1 or 10 nM AOP-RANTES was added to the bottom well of the chemotaxis plate compared with the same concentration of RANTES. In contrast, AOP-RANTES was less effective at attracting THP1 cells through CCR1 (Fig. 3D) with virtually no cells being attracted toward 10 nM AOP-RANTES, while the same concentration of RANTES was an effective stimulus.
Thus, with MIP-1␣P, RANTES, and their AOP derivatives, the Ca 2ϩ flux data closely mirrors the chemotaxis data, suggesting that although AOP variants of MIP-1␣P and RANTES have enhanced interactions with CCR1 and -5, this does not consistently result in an equivalent enhancement of bioactivity.
D6-mediated signaling or chemotaxis was not tested. This receptor has shown no signaling or chemotactic activity when expressed in heterologous cells under assay conditions in which other chemokine receptors show robust signals. 2 AOP-MIP-1␣P Is the Most Potent Chemokine-based Inhibitor of HIV Entry through CCR5-We have shown previously (39) that, in vitro, MIP-1␣P is a potent inhibitor of HIV entry through CCR5, and is consistently as effective as AOP-RAN-TES in this context. Thus, it was anticipated that AOP-MIP-1␣P may show even greater potency at inhibiting viral entry due to its enhanced ability (compared with MIP-1␣P) to bind to CCR5. To test this, the ability of AOP-MIP-1␣P to prevent pseudotyped reporter virus entry into CEMx174-CCR5 cells was compared with MIP-1␣P and AOP-RANTES. Remarkably, 0.1 nM AOP-MIP-1␣P was able to almost completely inhibit entry of JR-FL and ADA pseudotyped virus (Fig. 4). AOP-RANTES and MIP-1␣P were less effective than AOP-MIP-1␣P, and only reduced viral entry by 60 -70% at 0.1 nM. Indeed, a 10-fold higher concentration of these ligands was required before complete inhibition of entry of JR-FL pseudotyped virus was seen, while some viral entry (5-10%) was still seen with the ADA-pseudotyped virus at this concentration. Infection by viruses pseudotyped with envelope from a CXCR4-tropic virus, JC2 (which uses CXCR4 on CEMx174 cells to gain entry) or an amphotropic murine leukemia virus, was unaffected by all the chemokines tested (data not shown). These assays show AOP-MIP-1␣P to be the most effective chemokine-based inhibitor of CCR5-mediated HIV entry currently known.
Internalization and Recycling of CCR5-CCR5 internalization and recycling may contribute to the HIV entry inhibitory properties of chemokines (38). Thus, we wished to compare the ability of MIP-1␣P and RANTES, and their AOP derivatives, to internalize and recycle CCR5 in CEM-CCR5 cells using FACS analysis to detect cell surface CCR5. First, cells were incubated with 5 nM of each chemokine and CCR5 expression determined over time. As seen in Fig. 5A, AOP-MIP-1␣P more rapidly internalizes CCR5 from the surface of transfected CEM cells, than MIP-1␣P, RANTES, or AOP-RANTES. With AOP-MIP-1␣P, nearly 60% of detectable cell surface CCR5 had been removed within 5 min, with a further 20% internalized by 30 min. When MIP-1␣P or AOP-RANTES were used, 50% of the surface CCR5 was removed after 30 min, while RANTES hardly internalized any of the CCR5 (Fig. 5A). However, by 45 min, equivalent levels of CCR5 remained on the surface of the CEM cells treated with MIP-1␣P or AOP-MIP-1␣P. Thus, the AOP-coupled chemokines more rapidly cause the internalization of CCR5 into these cells than the unmodified chemokines, with AOP-MIP-1␣P being more effective than AOP-RANTES. With 100 nM MIP-1␣P, AOP-MIP-1␣P, or AOP-RANTES, CCR5 internalization was broadly similar with more than 80% of cell surface CCR5 internalized by 15 min (Fig. 5B). However, RAN-TES removed only about 50% over 45 min. Next, we wished to examine whether CCR5 was able to recycle to the cell surface after internalization by MIP-1␣P or AOP-MIP-1␣P. CEM-CCR5 cells were treated for 45 min with 100 nM chemokine, washed thoroughly, reincubated in medium, and the re-emergence of CCR5 assessed by FACS analysis (Fig. 5C). After 2 h, cells treated with MIP-1␣P have 65% of the control levels of CCR5 back on the surface while those treated with AOP-MIP-1␣P have just above 30%. Therefore, AOP-MIP-1␣P brings about a more persistent down-regulation of cell surface CCR5 than an equimolar concentration of MIP-1␣P. There are also reproducible differences in the rate of re-emergence between the two proteins, particularly over the first 30 min after washing away the chemokine. Very little AOP-MIP-1␣P associated CCR5 returns to the surface over this time period, while cells treated with MIP-1␣P double their detectable cell surface CCR5 protein levels. Similarly, AOP-RANTES more effectively prevented the re-emergence of internalized CCR5 compared with RANTES in these assays (not shown). DISCUSSION Here we have described the biological properties of a form of MIP-1␣P modified at the NH 2 terminus by the addition of an aminooxypentane group. AOP-MIP-1␣P has enhanced affinity for CCR1 and CCR5, but surprisingly, does not show dramat-ically altered Ca 2ϩ flux signaling or chemotactic potency compared with MIP-1␣P. Importantly, AOP-MIP-1␣P is markedly more potent than MIP-1␣P and AOP-RANTES in inhibiting entry through CCR5 of viruses pseudotyped with the envelope protein of R5 HIV strains, and therefore, appears to be the most potent chemokine-based inhibitor of CCR5-mediated HIV entry currently known.
Modifying the amino-terminal domains of chemokines can dramatically alter receptor affinity, and change the ability of the ligated receptor to couple to the signaling machinery of the cell (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16). How AOP addition alters the CCR1 and CCR5 binding properties of MIP-1␣P is uncertain, but this hydrophobic chain may become embedded more readily in the hydrophobic core of the receptor to increase receptor affinity. However, AOP-MIP-1␣P should be viewed as a partial, rather than a full, agonist of these receptors with respect to coupling to Ca 2ϩ fluxes and chemotaxis, because increased receptor occupancy at a given concentration by AOP-MIP-1␣P results in little change in activity in these assays. Some, or all, of these AOP-MIP-1␣P-ligated receptors must not be contributing as fully to chemotaxis and Ca 2ϩ flux induction as when an equivalent number of receptors are bound to MIP-1␣P. Nonetheless, this increased receptor occupancy results in AOP-MIP-1␣P and MIP-1␣P being distinct in their ability to prevent subsequent Ca 2ϩ fluxes induced by MIP-1␣P. Thus, the use of AOP-MIP-1␣P as the first stimulus is a more effective inhibitor of subsequent stimulation because the amount of remaining receptor available is reduced.
AOP-RANTES, like AOP-MIP-1␣P, showed an enhanced ability to displace mMIP-1␣ from CCR1 compared with RAN-TES, yet this increased receptor occupancy is not effectively coupled to Ca 2ϩ fluxes and chemotaxis. This suggests that AOP-RANTES could also be classified as a partial agonist of CCR1. This is supported by the observation that, while treatment of CCR1-expressing cells with RANTES or AOP-RANTES gives starkly different Ca 2ϩ fluxes, it can still bring about comparable reductions in a second Ca 2ϩ flux induced with RANTES (not shown). On CCR5, AOP addition to RANTES increases mMIP-1␣ displacement, but also markedly increases its chemotactic potency and its ability to couple to Ca 2ϩ fluxes. We are currently uncertain as to why the AOP moiety only increases the activity of RANTES in these assays, and only through CCR5. However, this emphasizes the complexity of chemokine/receptor interactions, and demonstrates the difficulty in predicting the outcome of amino-terminal modification. Our results with AOP-RANTES are, to some extent, in contrast to published observations that, despite showing similar changes in Ca 2ϩ flux induction through CCR1 (44) or CCR5 (38) by AOP addition to RANTES, revealed little change in binding affinity (10,44). We are uncertain why these discrepancies exist, but they could be explained by the different protocols, radiolabeled ligands, and/or cell lines used for analyzing receptor affinity.
It is of note, that in contrast to CCR1 and CCR5, D6 has a reduced affinity for AOP variants of MIP-1␣P and RANTES compared with their unmodified counterparts. It is tempting to speculate that this is indicative of an atypical receptor conformation of D6, perhaps linked to the current "non-signaling" status of this receptor. It will be interesting to see if mutants of D6 that exhibit weak signaling ability 3 show altered interaction with AOP variants.
The increased receptor occupancy seen on addition of AOP to MIP-1␣P, while not changing chemotaxis and Ca 2ϩ flux induction, increases the rate of CCR5 internalization and slows 3 R. J. B. Nibbs, unpublished data. receptor re-emergence. Similarly, AOP-RANTES induces more rapid and effective CCR5 internalization, and slower recycling, than RANTES. Chemokine receptor internalization is believed to be controlled by signals emanating from the ligated receptor that alter the receptor phosphorylation status and allow its recognition by the endocytotic machinery (45,46). Our results suggest that CCR5 internalization uses signal transduction pathways that, unlike those coupled to Ca 2ϩ fluxes and chemotaxis, are more effectively stimulated by CCR5 bound to AOP-MIP-1␣P than the unmodified counterpart. It has been shown that AOP-RANTES triggers more rapid CCR5 phosphorylation and association with GRK2 and ␤-arrestin1 (proteins involved in receptor desensitization and endocytosis) (47), and similar pathways may be involved in mediating the rapid AOP-MIP-1␣P-induced CCR5 internalization. An alternative possibility is that AOP-linked chemokines cause CCR5 to cluster more effectively thereby encouraging internalization, and indeed AOP-MIP-1␣P does form higher order aggregates more readily than MIP-1␣P. 4 Once inside the cell, the slowed CCR5 recycling seen with AOP variants may be a consequence of greater receptor affinity, or perhaps increased aggregation tendency, preventing receptor-ligand dissociation and subsequent CCR5 recycling. As proposed by Mack and colleagues (38), such receptor complexes may be more readily targeted for degradation.
The primary aim at the outset was to generate a molecule with enhanced potency at inhibiting HIV entry through CCR5. This has proved successful: AOP-MIP-1␣P is markedly more active (approximately 10-fold) than MIP-1␣P and AOP-RAN-TES at preventing the entry of R5-pseudotyped reporter viruses into CEMx174-CCR5 cells. A simple explanation for this is that enhanced affinity for CCR5 causes more receptor to become internalized or blocked, leaving less available to be used by the virus. Indeed, this view is widely held as the mechanism by which chemokines prevent HIV entry. However, closer examination of the data suggests that this may not be an adequate explanation, since the concentration of AOP-MIP-1␣P (and MIP-1␣P and AOP-RANTES) required to block Ͼ95% of viral entry into CEMx174-CCR5 cells is so low. Indeed, in most reports (10, 11, 48 -51), concentrations of chemokine that have been used to prevent HIV entry are insufficient to result in many CCR5 receptors becoming occupied or internalized. For example, in our hands, the use of 0.1 nM AOP-MIP-1␣P (that blocks Ͼ95% of viral internalization) is not able to cause any detectable internalization of CCR5 from the surface of CEM-CCR5 cells (not shown), although chemotaxis is stimulated at this concentration showing some CCR5 molecules are occupied (Fig. 3A). However, from the heterologous displacement studies in recycling (all of which have been proposed to play a role in inhibiting HIV entry), are unlikely to be of much importance at the low chemokine concentrations that effectively inhibit viral entry.
These observations lead us to propose that an alternative mechanism is at play, and we are currently investigating whether signals emanating from CCR5 at low receptor occupancy are involved in chemokine-mediated inhibition of HIV entry through this receptor. While signaling appears not to be required for HIV entry through CCR5 (52)(53)(54), it may be important for the inhibition of viral entry seen at low chemokine concentrations. While we have shown here that MIP-1␣P and AOP-MIP-1␣P show similar potency in the Ca 2ϩ flux and chemotaxis assays with CCR5-expressing cells, nonetheless other signaling pathways coupled to CCR5 may be more potently stimulated by AOP-modified, rather than wild-type, MIP-1␣P and contribute to the enhanced inhibition of viral internalization seen with this protein. Indeed, as we have discussed above, the faster rate of CCR5 internalization upon binding AOP-MIP-1␣P, rather than MIP-1␣P, may reflect enhanced signaling activity by AOP chemokines. Also, it is worth considering that in the viral internalization assays chemokine stimulation is over a longer time period than in the Ca 2ϩ flux and chemotaxis assays. Such chronic, low-level CCR5 stimulation may elicit quite distinct intracellular signals. Experiments are underway to attempt to answer some of these questions.
In summary, the addition of aminooxypentane to MIP-1␣P enhances interaction with CCR1 and CCR5, increases CCR5 internalization, and creates a more potent inhibitor of HIV entry through CCR5 without affecting its chemotactic potency or its ability to induce Ca 2ϩ fluxes. These studies suggest that an alternative mechanism, other than steric interference or receptor internalization, may be involved in chemokine inhibition of viral entry. Finally, the properties of AOP-MIP-1␣P may make it useful experimentally to down-regulate MIP-1␣P receptors, and potentially therapeutically applicable in the treatment of HIV infection and inflammatory disease.