Previous Article | Next Article ![]()
Journal of Virology, April 2008, p. 4169-4174, Vol. 82, No. 8
0022-538X/08/$08.00+0 doi:10.1128/JVI.01070-07
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Graham Bottley,1
Lucy P. Beales,2
Richard A. Killington,1
David J. Rowlands,1 and
Tobias J. Tuthill1*
Institute for Molecular and Cellular Biology, Faculty of Biological Sciences, The University of Leeds, Leeds LS2 9JT, United Kingdom,1 iQur Ltd., Faculty of Biological Sciences, The University of Leeds, Leeds LS2 9JT, United Kingdom2
Received 17 May 2007/ Accepted 21 January 2008
|
|
|---|
|
|
|---|
There are over 100 serotypes of HRV, which can be grouped according to receptor usage (32). Minor-group viruses such as HRV2 use the low-density lipoprotein receptor family as receptors. They are acid labile and have an absolute dependence on low pH for uncoating (24). During infection, the process of uncoating leads to the delivery of the HRV2 genome into the cytoplasm from within the intact endosome, through a size-selective pore in the endosome membrane (3, 25). However, the mechanism of the formation of this pore remains unknown.
Major-group viruses such as HRV14 and HRV16 use intercellular adhesion molecule 1 (ICAM-1) as their receptor. They are also generally acid labile but, unlike the minor-group viruses, do not have an absolute dependence on low pH for uncoating (1). For example, receptor binding at neutral pH is sufficient to initiate the uncoating of HRV14 in vitro (19), similar to the situation with the closely related poliovirus (PV). During infection, uncoating of HRV14 leads to the delivery of the genome to the cytoplasm by a mechanism that appears to completely disrupt the endosome membrane (28). The mechanism for producing such a membrane alteration remains unclear.
In addition to these findings with HRV, studies with PV have also provided information about early events in virus entry (4, 6, 18). During the uncoating of either HRV or PV, native virus (which sediments at
160S) is converted to an altered particle (
135S). For PV, VP4 has been released from the altered particle to become associated with the membrane (8, 31). The hydrophobic N termini of the VP1 proteins have also become externalized and allow the particle to interact directly with the membrane (12, 31). Further alterations in the particle ultimately result in an empty capsid (
80S) which in the case of HRV has lost some or all of its VP4 and contains pores at the fivefold axes through which the RNA genome is thought to have been released (17). The mechanisms of the release of RNA from the particle, the permeabilization of the endosome membrane, and the coordination of these events (to provide safe passage of RNA from the particle to the cytoplasm) remain unclear.
The interaction of VP4 with membranes following its release from the particle suggests a potential role for VP4 in the membrane alterations required for genome transfer. Indeed, a specific mutation in VP4 of PV prevents the formation of membrane-spanning channels in model membranes and also prevents productive genome delivery to the cytoplasm of the cell (8). There is therefore evidence for a link between VP4-mediated membrane permeability and genome delivery. However, the precise role of VP4 in picornavirus genome delivery has yet to be established.
We describe here the expression of recombinant HRV16 VP4 as a glutathione S-transferase (GST) fusion protein, VP4GST, and an investigation of its ability to interact with and induce permeability in lipid membranes in the form of liposomes.
Preparation of recombinant VP4. The sequence encoding HRV16 VP4 was amplified from an infectious clone of the virus (21) by PCR using primers VP4F (5'ATATGCTAGCCTCGCCATGGGCGCTCAAGTATCTAGACAGAATGTTG G3') and VP4R (5'GGATCCACGCGGAACCAGATGGTGATGGTGATGGTGTTGCAGAGTGGGTATGCCTTTCTC C3'). GST was amplified by PCR using primers GSTF (5'CA CCATCACCATCACCATCTGGTTCCGCGTGGATCCAT GTCCCCTATACTAGGTTATTGG3') and GSTR (5'TATAGCGGCCGCCTCGAGTCAATCCGATTTTGGAGGATG G3') from pGEX 4T-1 (Invitrogen). These PCR products were used as overlapping templates (underlined regions denote complementary primer sequences) in a further PCR to produce a cDNA encoding VP4GST, which was cloned into pET-23d (Novagen) using restriction sites NcoI and XhoI, producing pET-VP4GST. Unmyristylated VP4GST or GST was expressed in BL21(DE3) Rosetta Escherichia coli transformed with pET-VP4GST or pGEX 4T-1, respectively. Myristylated VP4GST was produced in the same host by cotransformation with a plasmid encoding N-myristyltransferase (pET-NMT) (9, 11). Expression was induced by the addition of 1 mM isopropyl β-D-1-thiogalactopyranoside. Myristic acid was added to a final concentration of 5 µg/ml to cells cotransformed with pET-NMT. Soluble VP4GST or GST was purified from lysates by glutathione affinity chromatography using standard methods and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Fig. 1A). The myristylation of VP4GST was confirmed using slow-crystallization matrix-assisted laser desorption ionization mass spectrometry (2). By this method, the mass of myristylated VP4GST was measured as 34,670 ± 29 Da and that of unmyristylated VP4GST as 34,472 ± 57 Da. The difference in molecular mass is consistent with the addition of myristic acid (205 Da).
![]() View larger version (47K): [in a new window] |
FIG. 1. (A) Coomassie blue-stained sodium dodecyl sulfate-polyacrylamide gel electrophoresis of purified unmyristylated VP4GST (lane 1), myristylated VP4GST (lane 2), and GST (lane 3). The positions of molecular mass markers (kDa) are indicated. The image was captured with a gel documentation system (Syngene). (B) VP4GST associates with liposomes. Lanes 1 to 10 represent fractions, from top to bottom, of Ficoll gradients, immunoblotted with GST-reactive serum. Liposomes banded at fraction 2. The positions of molecular mass markers (kDa) are indicated. Immunoblots were exposed to film and scanned (Epson 4870).
|
VP4GST induces membrane permeability in a concentration- and myristylation-dependent manner. There is existing evidence for the involvement of VP4 in the formation of membrane channels by PV (8). Having shown that VP4GST associated with liposomes, we therefore examined the effect of this interaction on membrane permeability by using a liposome dye release assay. Liposomes comprising a 1:1 molar ratio of PA and phosphatidylcholine with 0.5% (wt/vol) rhodamine B-labeled phosphatidylethanolamine and enclosing 50 mM (self-quenching) carboxyfluorescein (CF) were generated by rehydration in 50 mM CF-10 mM HEPES, pH 7.4 (PA liposomes were refractory to preparation with high concentrations of internal CF). After extrusion, CF-containing unilamellar liposomes were isolated from extraneous dye by washing four times in 107 mM NaCl-10 mM HEPES (pH 7.3) by pelleting at 100,000 x g and 20°C for 15 min. The level of rhodamine fluorescence in purified liposome preparations was used to estimate the lipid concentration relative to the initial rehydrated lipid.
Liposomes (25 µM lipid) were mixed at approximately 25°C with VP4GST or GST or mock treated. Membrane permeability was detected as the movement of quenched CF from inside the liposome to become dequenched on release to the outside, with a subsequent increase in fluorescent signal (485/520 nm) that was recorded using a 96-well fluorimeter (BMG FluoStar). At a concentration of 300 nM, both myristylated and unmyristylated forms of VP4GST induced the relatively rapid release of CF from liposomes (Fig. 2A), with the initial rates of release being approximately six- to sevenfold higher than the background level of release induced by GST (Fig. 2C). At this concentration, the presence of myristate therefore appeared not to be required for VP4 function in this assay. When used at a 10-fold-lower concentration (30 nM), myristylated VP4GST still induced a more rapid and extensive permeability than GST; however, unmyristylated VP4GST was no more effective than GST alone (Fig. 2B and C). Thus, at the lower concentration, myristylation of VP4 is required for optimal activity. When the concentration was further reduced to 3 nM, CF release could not be reproducibly detected. Similar overall results were obtained at 37°C (data not shown), suggesting that physiological temperature is not a requirement for membrane permeability in this assay. The kinetics of VP4-induced CF release appeared broadly comparable to that of CF release induced by the well-characterized pore-forming peptide melittin (Fig. 2B) (26).
![]() View larger version (21K): [in a new window] |
FIG. 2. VP4GST-induced membrane permeability measured by the release and fluorescent dequenching of CF from liposomes. (A) Time course of release induced by 300 nM myristylated (black circles) or unmyristylated (gray circles) VP4GST and by GST alone (open circles). The release induced by Triton X-100 (triangles) served as a control for the complete disruption of liposomes. The data shown are representative of multiple experiments. (B) Time course of release induced by 30 nM myristylated (black circles) or unmyristylated (gray circles) VP4GST and by GST alone (open circles). The release induced by 100 nM melittin (crosses) is shown for comparison. The data shown are representative of multiple experiments. (C) Relative initial rates of release for myristylated VP4GST (myr), unmyristylated VP4GST (un), and GST alone, as calculated by fitting a straight line to the initial four data points of the time course. Error bars indicate standard deviations (n = 3).
|
Liposomes (25 µM lipid) containing 1.6 mM (unquenched) CF were mixed at room temperature with VP4GST or GST or were mock treated. After 10 min, anti-GST antibodies (murine; Serotec) were added, followed after an additional 10 min by labeled secondary antibodies (anti-mouse Alexafluor 633; Invitrogen). After a further 10 min, liposomes were analyzed by flow cytometry (BD FACSaria and FACSDiva). Liposomes were detected by forward and side scattering of light and by the fluorescent signal derived from enclosed CF (488/525 nm). The interaction between proteins and liposomes was detected by the additional fluorescent signal from immunostaining with anti-GST (633/647 nm).
After the addition of VP4GST, liposomes were simultaneously found to have bound high levels of protein, increased in granularity, and lost the majority of their internal contents while remaining intact (Fig. 3). They could be physically sorted into distinct populations by, for example, their reduced internal fluorescence, and liposomes that had been permeabilized could still be sorted and reanalyzed, therefore confirming liposome integrity (data not shown). The formation of distinct populations of liposomes with either low or high levels of protein binding indicated that there may be a cooperative interaction of proteins with the membrane. The release of internal contents from within intact liposomes strongly suggests a controlled mechanism of permeability such as the formation of a membrane pore. At protein concentrations of 30 nM, typically 40% of intact liposomes had lost on average 90% of their internal fluorescence, compared with 10 to 15% of liposomes mixed with GST alone. The loss of the majority of CF from only a proportion of liposomes suggested that permeability was an all-or-nothing event, perhaps requiring a critical number of molecules per liposome. In order to calculate the VP4GST/liposome ratio, we first estimated the liposome concentration by calculating the number of lipid molecules per liposome based on the relative surface areas of liposomes (500 x 103 nm2) and lipid head groups (0.7 nm2). At a protein concentration of 30 nM, the VP4GST/liposome ratio was approximately 2,000:1. However, the number of functional pore-forming units per liposome would be only a fraction of this, because for pore formation to occur, a number of molecules would be required to specifically interact as a multimeric complex.
![]() View larger version (50K): [in a new window] |
FIG. 3. Liposomes permeabilized by VP4GST remain intact by flow cytometry. Analyses of liposomes alone (A) and after the addition of 30 nM myristylated VP4GST (B), unmyristylated VP4GST (C), or GST (D) are shown. Dot plots in the left column show detection of liposomes by forward and side light scattering; after addition of protein (B, C, and D), a proportion of liposomes (purple) have increased granularity (side scatter) relative to liposomes alone (A). Dot plots in the middle column show two populations of liposomes: low-level signal for protein binding with high-level CF internal fluorescence (orange) and high-level protein binding with decreased CF internal fluorescence (purple) as a result of induced permeability and release of CF. This analysis was not carried out with liposomes alone (A). The histograms in the right column show the distribution of CF fluorescence for the populations (orange and purple) described in the dot plots. The data shown are representative of three independent experiments.
|
At protein concentrations 10-fold higher (300 nM), two very similar populations of liposomes were also observed by flow cytometry (data not shown). However, at this increased concentration of protein, no significant additional effect was observed over that already seen at 30 nM. This was in contrast to the findings of the kinetic assay, where the higher protein concentration produced a more rapid and extensive permeability. It is possible that the increased rate of CF release seen at 300 nM in the kinetic assay was due to a large-scale membrane disruption, such that these liposomes were no longer intact and were below the size limit for detection by flow cytometry. In support of this hypothesis, we have observed the release of high-molecular-weight dextran molecules from liposomes together with electron microscopy evidence of their disruption in the presence of very high concentrations of VP4GST (data not shown).
It is tempting to speculate that permeability induced by low concentrations of VP4GST is due to the formation of discrete membrane pores, which may require myristylated VP4, while permeability at higher concentrations may be due to a less specific membrane disruption which is independent of myristylation. This hypothesis would also reconcile the apparently conflicting evidence for pore formation by HRV2 (and PV) and endosome disruption by HRV14 during the entry process. Existing analyses of several picornavirus empty particles (13, 14, 16-19) suggest that variable proportions of the 60 VP4 molecules present in virions are released during entry. This could influence the endosomal VP4 concentration, creating a preference for either pore formation or membrane disruption. Sixty VP4 molecules in a 350-nm-diameter endosome would generate an estimated concentration of 14 µM (or less if release of VP4 was incomplete), which is reasonably close to the concentrations of recombinant protein used in this study.
Alternatively, these different modes of membrane permeability may simply be the result of variation between different viruses. HRV2 and HRV16 are closely related and differ by only 5 amino acids in VP4; HRV2 and HRV14 are more distantly related and differ by 31 amino acids. However, the properties of variant amino acids are largely conserved, with the result that structure predictions for these proteins are indistinguishable. Despite this, there may be subtle functional differences between these VP4 proteins. The effects of sequence variation in VP4, including the effect of VP4 mutations known to influence the properties of PV VP4, are the subject of further investigation.
The results of this study suggest that VP4 has the intrinsic ability to bind lipid membranes, resulting in their permeabilization. When the N-terminal myristylation of VP4 was first demonstrated, it was suggested that this modification may be important for membrane association (7). More recently the myristylation of a recombinant VP4-green fluorescent protein fusion protein was shown to be required for membrane targeting within the cell (22). Our current findings provide the first confirmation of a specific functional role for the myristate in VP4 membrane activity. This report provides further evidence that VP4 is directly involved in mediating membrane penetration during cell entry, in agreement with previous hypotheses (5, 18). The possibility that VP4-liposome interactions may result in alternate, concentration-dependent mechanisms of membrane permeability provides a potential explanation for the differences observed between the entries of otherwise very similar viruses.
This work was funded by the Wellcome Trust and the Medical Research Council.
Published ahead of print on 6 February 2008. ![]()
Present address: Centre for Infectious Diseases, University of Edinburgh, Summerhall, Edinburgh EH9 1QH, United Kingdom. ![]()
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»