Enhancement of Enveloped Virus Entry by Phosphatidylserine

PS treatment of target cells enhances infection by multiple enveloped viruses when functional virus receptors are present.

We treated a variety of cell types with PS and exposed them to several different viral vectors to examine the specificity and magnitude of transduction enhancement by PS (Table 1). All cells expressed functional receptors for the virus vectors used. Viruses carried the LAPSN retroviral vector (27) that encodes human placental alkaline phosphatase (AP) or the LNCG retroviral vector (32) that encodes green fluorescent protein (GFP) and were made using the following packaging cell lines that express the indicated Env proteins: PA317 (amphotropic murine leukemia virus [MLV] strain 4070A) (24); FlyRD (RD114) (11); PJ4 (jaagsiekte sheep retrovirus [JSRV]) (30); and PG13 (gibbon ape leukemia virus [GALV]) (25). VSV-G pseudotype vectors were made by transfection, as described previously (10). Treatment of the target cells with PS increased transduction by these viruses by 2- to 20-fold. We also examined the relationship of this enhancement to the one provided by Polybrene and determined that the treatments had synergistic effects. For example, JSRV infection of Rat-2/Hyal2 cells was enhanced 4-fold by Polybrene alone, 6-fold by PS alone, and 25-fold by the two combined (means of results of two experiments).

We attempted to use PS treatment to increase infection of HT-1080 cells by two nonenveloped viruses, an adenovirus type 5 vector (Ad.RSV-βgal) (42) and an adeno-associated virus type 2 vector (CWCZn) (15), both of which encoded β-galactosidase. In these cases, PS treatment actually reduced virus infection by 75% and 25%, respectively. These results show that PS does not stimulate infection by these nonenveloped virus vectors and suggest that PS enhancement of infection may take place at the fusion step of virus entry, which is not required by nonenveloped viruses.

PS treatment of target cells does not allow entry of viruses in cases where a functional receptor is not present.

To test whether the enhancement of virus infection was receptor independent, we exposed a variety of cell lines to viruses that ordinarily cannot infect these cells due to lack of functional receptors. We tested a Moloney MLV pseudotype LAPSN vector made by using PE501 retrovirus packaging cells (26) on HT-1080 and HEK 293 human cells; a JSRV pseudotype LAPSN vector on the rat cell lines 208F, NRK, and Rat-2; a GALV pseudotype LAPSN vector on NIH 3T3 mouse cells; and a xenotropic MLV (AKR6) pseudotype LAPSN vector on CHO-K1 hamster cells. The LAPSN(AKR6) xenotropic virus was produced by infection of Mus dunni tail fibroblasts with LAPSN and replication-competent AKR6 virus (8). Transduction was undetectable (<1 AP⁺ focus-forming unit/ml) with or without PS in two independent experiments for all of these virus-cell combinations. Positive-control infections showed that all of these cell types were appropriately infectible by retroviral vectors that recognized receptors expressed by the cells (data not shown).

PS treatment temporarily increases the amount of cell-surface PS on target cells.

To further characterize the effect of PS treatment, we investigated whether the exogenously added PS liposomes were incorporated into the plasma membrane of the target cells. The increase of cell-surface PS levels measured in NIH 3T3, ZF4, HT-1080, and Rat-2 cells after PS treatment is depicted in Fig. 1. Both the background PS levels and the magnitude of the increase after the addition of PS liposomes vary in these cell types. However, there was no obvious correlation between the increase in cell-surface PS levels (Fig. 1) and the magnitude of the enhancement of virus infection (Table 1). Expression of various viral receptors on NIH 3T3 and Rat-2 cells did not affect PS levels or the response to PS liposomes in this assay (<5% change in duplicate experiments [data not shown]). Within 72 h of PS treatment, the PS levels of treated cells were back to normal, and the cells were infectible to the same level as untreated cells (data not shown).

The dose response of target cells to PS treatment varies between cell types.

In addition to PS having various effects on virus infection, the different cell lines had varied responses to the addition of PS. A few cell lines exhibited moderate death and an inhibition of growth, whereas most cells lines were less affected. Variation between cell types in cell-surface PS levels after the addition of 400 μM PS liposomes is indicated in Fig. 1.

To examine the dose response of infection to different amounts of PS, HT-1080 and Rat-2/Hyal2 cells were incubated with various levels of PS for 24 h, and the rate of infection following exposure to LAPSN(JSRV) virus was determined (Fig. 2). For Rat-2/Hyal2 cells, the optimum concentration of PS was 80 μM, whereas the optimum concentration for HT-1080 cells was 320 μM. At the highest levels of PS, the Rat-2/Hyal2 cells began to die, whereas the HT-1080 cells were much more resistant to this toxicity (data not shown). Because the dose response of cells to PS varies between cell types, it will be important to determine the optimal concentration of PS for a given cell type. Our results indicate that even higher levels of enhancement might be achieved in some cell types by treating cells with lower concentrations of PS than those used in Table 1.

Addition of other phospholipids does not affect enveloped-virus infection.

To address the possibility that nonspecific disruption of the cell membrane by the addition of phospholipids was responsible for the enhancement of infection by PS, we examined the effects of liposomes composed of other phospholipids. Our previous work had shown that phosphatidylcholine (PC) liposomes had no effect on VSV-G and RD114 pseudotype vector infection in ZF4 cells (10). In addition to PS and PC, Rat-2/Hyal2 cells were treated with phosphatidylglycerol (PG), and phosphatidylethanolamine (PE). These lipids were chosen because of their similarity to PS and their potential roles in membrane fusion; PG may interact with the human immunodeficiency virus type 1 fusion peptide to play a role in lentiviral fusion (28, 29), and PE has the ability to promote the bilayer- to hexagonal-phase transition that may facilitate membrane fusion (34). The addition of PC, PG, or PE did not affect the PS levels in the target cells as measured by annexin-V staining (<5% change in duplicate experiments). None of these three phospholipids had any noticeable effect on virus infection, while the addition of PS resulted in a large boost in infection by LAPSN(JSRV) virus (Fig. 3). Similar results were found for LAPSN(RD114) virus infection of HT-1080 cells (data not shown).

PS treatment has little effect on virus receptor levels or Env binding to cells.

To examine whether PS treatment increased virus infection by increasing virus receptor levels or by increasing virus binding to cells, we measured cell-surface receptor levels and JSRV Env binding to Rat-2/Hyal2 cells. The Rat-2/Hyal2 cells contain an amino-terminal FLAG-tagged Hyal2 protein which functions as a receptor for JSRV (31). These cells were treated with or without 80 μM PS for 24 h, and receptor levels were measured using an anti-FLAG antibody (Fig. 4). This concentration of PS was used because it maximizes the increase in infectivity for this cell type (Fig. 2) and therefore should maximize changes in factors that mediate the increase. In the same assay, the amount of JSRV Env binding was measured by using a JSRV Env surface domain-human immunoglobulin G fusion protein (JSU-IgG) as previously described (39) (Fig. 4). Both the levels of Hyal2 protein and JSU-IgG binding show <2-fold increases after PS treatment. In addition, the background fluorescence of the cells alone increases with PS treatment, reducing the measured effect even further. These small changes are unlikely to account for the up-to-sevenfold increase in JSRV infection observed in these cells upon PS treatment. The fact that neither receptor levels nor virus binding was significantly affected by PS treatment further supports the hypothesis that PS treatment enhances the fusion step of virus entry.

PS liposomes are easily generated in large batches and can be frozen without loss of activity.

For all experiments described so far, fresh PS was produced before each experiment in order to maintain consistency. To be useful as a general tool to enhance virus infection, it would be more convenient to produce large quantities of PS liposomes and freeze them in aliquots. We found that PS liposomes had the same effects on cells whether they were made in large or small batches (data not shown). Therefore, we made PS in a few large batches and kept aliquots at either 4°C or −20°C. In two experiments, the liposomes kept at 4°C lost approximately 25% of their activity within 3 days. However, aliquots stored at −20°C retained their effectiveness for at least 30 days (data not shown). In fact, there was an additional twofold increase in the effectiveness of PS liposomes after being frozen, which persisted for the 30-day time course (results of two experiments [data not shown]).

PS treatment most likely affects virus fusion. This conclusion is supported by the fact that receptor levels and virus binding are minimally affected by PS treatment in a case where a large increase in infection is noted. Furthermore, the enhancement of infection occurred only for enveloped viruses and not for the two nonenveloped viruses tested. We do not believe that this enhancement is a charge effect relating to virus adsorption, such as that of Polybrene, since PS is negatively charged and anions have repeatedly been shown to reduce virus infection (4, 22, 38). There are many facets to virus fusion that could be influenced by a large increase in PS concentration in the plasma membrane, such as lipid packing, alterations in bilayer curvature, changes in membrane fluidity, and locally induced changes in the bilayer phase (for a review, see reference 33). Further biophysical studies will be required to elucidate the particular mechanism by which PS enhances virus fusion.

Our results demonstrate that an increase in cell surface levels of PS allows a corresponding increase in virus infection. Interestingly, PS is normally found primarily on the inner leaflet of the plasma membrane, and there are relatively low levels of PS found on the outside of cells (6, 41, 44). It is possible that the typical levels of PS in a cell are a rate-limiting factor affecting normal virus fusion.

Virus-cell fusion is a complex and incompletely understood process (for reviews, see references 5, 9, and 33). No role for PS in the fusion of enveloped viruses has been proposed so far, although there is evidence for postentry requirements of PS in Sindbis virus replication (18) and in mRNA capping of Semliki Forest virus (1). The presence of PS in the virion may play a supporting role in human immunodeficiency virus infection (7). Similarly, PS inclusion into Sendai virus virosomes allows an increased level of virosome/cell fusion (17). However, the level of phosphatidylserine in the target cell membrane has not been previously shown to play any role in virus entry. Our results indicate that PS plays a role in facilitating enveloped virus fusion and can be used as a general tool to increase infection rates.