Visualization of Calcium Ion Loss from Rotavirus during Cell Entry

Nonenveloped viruses penetrate into the cytosol of the cells that they infect by disrupting the membrane of an intracellular compartment. The molecular mechanisms of membrane disruption remain largely undefined. Functional reconstitution of infectious rotavirus particles (TLPs) from RNA-containing core particles (DLPs) and the outer layer proteins that deliver them into a cell makes these important pediatric pathogens particularly good models for studying nonenveloped virus entry. We report here how the use of a fluorescent Ca2+ sensor, covalently linked to one of the viral proteins, allows us to establish, using live-cell imaging, the timing of Ca2+ loss from an entering particle and other molecular events in the entry pathway. Specific Ca2+ binding stabilizes many other viruses of eukaryotes, and Ca2+ loss appears to be a trigger for steps in penetration or uncoating. The experimental design that we describe may be useful for studying entry of other viral pathogens.

and localization of changes in Ca 2ϩ concentrations and hence considerable subtlety in Ca 2ϩ -based signaling mechanisms. It also permits entering molecules to detect when they have crossed a cellular membrane, either from the cell surface or from an internal, membrane-bound compartment, and arrived in the cytosol.
Various nonenveloped viruses have evolved to depend on bound Ca 2ϩ for particle stability and hence to respond to its loss by conformational changes that lead to uncoating or disassembling. Examples include the polyomaviruses, most T ϭ 3 plant viruses, and rotaviruses (2)(3)(4)(5)(6)(7). The infectious form of rotavirus, the so-called triplelayered particle (TLP) (Fig. 1A), contains an icosahedral (T ϭ 13) double-layered particle (DLP) core, surrounded by a shell of viral protein 7 (VP7) trimers. The VP7 shell anchors 60 trimeric VP4 spikes onto the particle surface (8). Ca 2ϩ ions bound specifically at the subunit interfaces hold the VP7 trimers together (9). Chelating these Ca 2ϩ ions strips away both VP7 and VP4, leaving an intact DLP. A related process occurs during cell entry. The TLP binds ganglioside receptors on the cell membrane (10,11) and generates, by invagination and budding, a tightly fitting vesicle (12). Progressive loss of VP7 and VP4 leads, after several minutes, to the release of the DLP into the cytosol. The transcriptionally active DLP contains the 11 segments of the viral double-stranded RNA (dsRNA) genome, each associated with a copy of the viral polymerase VP1. DLP release is thus the critical step for initiating infection (13)(14)(15)(16).
Various experimental approaches have shown that VP4 is the key agent of vesiclemembrane disruption and DLP release. Proteolytic cleavage to VP8* and VP5* activates VP4 (17,18), enabling VP5* to undergo membrane-linked conformational changes (19)(20)(21). In previous work, we monitored the entry events outlined above by generating FIG 1 Rotavirus structure, reconstitution, location of the S469C mutation in VP4, and structures of the Ca 2ϩ sensor components. (A) Rotavirus structural components, including the outer coat proteins VP4 (red) and VP7 (yellow) and the DLP proteins VP6 (green) and VP2 (cyan). The double-layer particle (DLP) of rotavirus can be reconstituted into an infectious triple-layer particle (TLP) by the addition of VP4 and VP7 in the presence of Ca 2ϩ (19). Chelating Ca 2ϩ ions dissociates VP7 trimers, stripping off the outer layer of the TLP. The blue box indicates the VP4 spike, magnified in the larger box, showing the location of the S469C mutation (yellow and cyan spheres) and the VP5* "body" and VP8* "lectin" subdomains of VP4 (the "foot" subdomain of VP8* plus VP5* is not shown [see reference 8]  reconstituted TLPs (rcTLPs) from DLPs and recombinant VP7 and VP4 (12,22). Distinct fluorescent markers on the DLP and on VP7 allowed us to monitor directly, by the use of live-cell imaging, the loss of VP7 and release of the DLP. Installation of a third fluorophore on VP4 can then permit us to monitor the timing of other features of the entry process, with respect to VP7 uncoating and DLP release. We have now taken advantage of this scheme to monitor the local drop in Ca 2ϩ concentrations by using a conjugatable Ca 2ϩ sensor, rhodamine 5F (Rhod5F) azide ( Fig. 1C) (23). We generated a mutant VP4 with a cysteine at a position unlikely to affect the stability of known conformers of VP5* and developed a scheme for installing Rhod5F at that position. We show here that there is a relatively slow (ca. 1-min) loss of Ca 2ϩ from the vesicular compartment surrounding a virion. This event precedes VP7 decapsidation by about 2 min and DLP release by about 7 min. The results constrain models for the molecular mechanism of membrane disruption and illustrate the utility of directly monitoring Ca 2ϩ levels during uptake of macromolecules into cells.

RESULTS
Reconstitution and imaging of Atto 565-labeled rcTLPs. We prepared wild-type (wt) recombinant VP4 (rVP4) as described previously (22) and labeled it with Atto 565 N-hydroxysuccinamide (NHS) ester (VP4-565) at a ratio of ϳ1 dye molecule per protein monomer. Atto 647N NHS ester-labeled DLPs (DLP-647N) were then recoated with VP4-565 and Atto 488 NHS ester-labeled VP7 (VP7-488), as described in Materials and Methods. The specific infectivity of these Atto 565 rcTLPs was 113 particles per focus-forming unit (P/FFU) ( Fig. 2A), slightly higher than the 161 P/FFU for the native TLP sample tested on the same day, a relationship that we have found previously (16).
Atto 565 rcTLPs were added to BSC-1 cells at a multiplicity of infection (MOI) of 50 and imaged for 30 min, as described in Materials and Methods. Tracking of the labeled DLPs by their intensity in the 642-nm channel yielded 127 total uncoating events ( Fig.  2B and C). As in previous work (12), we defined uncoating as a particle trajectory in which the VP7 intensity (488-nm channel) was reduced to background levels while the DLP intensity remained roughly constant. After uncoating, we monitored the DLP until the particle began to move rapidly away from the site of uncoating ("release").
Analysis of these data showed that initiation of VP7 signal loss in the 488-nm channel (t 1 ) (Fig. 3) coincided with initiation of VP4 signal loss in the 560-nm channel (t VP4-1 ) ( Fig. 3B and C and Table 1). The average interval between t 1 or t VP4-1 and subsequent DLP release (t 3 ) was 4.0 Ϯ 0.4 min. The average interval (t 3 Ϫ t 2 ) between the endpoint of VP7 loss (t 2 ) and release (t 3 ) was 2.9 Ϯ 0.4 min.
Reconstitution and imaging of pHrodo Red-labeled rcTLPs. We labeled wild-type VP4 with pHrodo Red (Invitrogen) succinimidyl ester (VP4-pHrodo Red) at a ratio of 1 dye molecule per 2 VP4 monomers and recoated DLP-647N with VP4-pHrodo Red and VP7-488, as described in Materials and Methods. The specific infectivity of these pHrodo Red rcTLPs was 165 P/FFU (Fig. 4A), which was again higher than that of native TLPs tested on the same day with a specific infectivity of 477 P/FFU.
We verified the response to changes in pH of pHrodo Red-labeled rcTLPs by imaging the reconstituted particles on glass in Fluorobrite Dulbecco's modified Eagle's medium (DMEM). The pHrodo Red rcTLPs emitted no fluorescent signal in the 562-nm channel in medium alone (Fig. 5A), but the addition of sodium acetate (pH 5.2) to the sample gave rise to a signal in the 562-nm channel that colocalized with both the DLP (640-nm) and VP7 (488-nm) signals of individual particles (Fig. 5B).
pHrodo Red rcTLPs were added to BSC-1 cells at an MOI of 50 and imaged for 30 min, as described in Materials and Methods. Tracking of the labeled DLPs by their intensity in the 642-nm channel yielded 253 total uncoating events, none of which was accompanied by the appearance of a signal in the 561-nm channel ( Fig. 4B and C). Those particles that emitted a signal in the pHrodo Red channel had not uncoated, even at the end of the acquisition time ( Fig. 4D and E). Treatment of cells with bafilomycin A1 (Baf A1), a known inhibitor of the vacuolar-type H ϩ -ATPases, did not change the infectivity of the pHrodo Red rcTLP (Fig. 6A), nor did Baf A1 have any effect on the pH response of the labeled rcTLPs on glass ( Fig. 5C and D). These observations are consistent with our previous studies showing that rhesus rotavirus (RRV) uncoats before it reaches an acidic endosome and that particles reaching such endosomes never uncoat (12).
Live-cell imaging of Baf A1-treated BSC-1 cells infected with pHrodo Red-labeled rcTLPs at an MOI of 50 yielded 81 uncoating events. As in Baf A1-free samples, particles that uncoated did so in the absence of a colocalized 561-nm signal ( Fig. 6B and C). No particles that remained as intact rcTLPs developed a 561-nm signal in these experiments (data not shown), consistent with the inhibition of vacuolar-type H ϩ -ATPases.
The times between the initiation of VP7 loss and DLP escape (t 3 Ϫ t 1 ) are 5.9 Ϯ 0.3 min and 5.8 Ϯ 0.6 min in the absence and presence of Baf A1, respectively ( Fig. 7 and Table 2). The t 3 Ϫ t 2 intervals for these data were 3.4 Ϯ 0.3 min and 4.4 Ϯ 0.6 min, respectively.
Rhod5F labeling of VP4 469C and formation of VP5 CT digestion fragments. We expressed VP4 mutated at position 469 from serine to cysteine (VP4 469C) and labeled the protein sequentially with dibenzocyclooctyne-maleimide (DBCO-maleimide) and Rhod5F azide ( Fig. 8A; see also Materials and Methods) at a final ratio of 1 Rhod5F molecule to each VP4. We refer to the product as VP4 Rhod5F.
The relatively large combined size of the DBCO linker and Rhod5F azide might in principle hinder the VP4 conformational change proposed to be necessary for cell entry. The final, folded-back conformation is represented by the structure of the sequentially chymotrypsin-trypsin (CT)-digested VP5 CT fragment (21). We therefore subjected VP4 Rhod5F to sequential chymotrypsin and trypsin digestion. Apparent molecular masses (from SDS-PAGE) of the VP5 CT fragment were approximately 60 kDa for a reduced but unboiled trimer of roughly 60 kDa and about 30 kDa for a reduced and boiled monomeric form. Digestion of VP4 Rhod5F yielded essentially the same fragments under these conditions, with the full-length protein, digested trimer, and digested monomer all containing the fluorescent Rhod5F label (Fig. 8A).
Infectivity and binding of Rhod5F-labeled rcTLPs. Because VP4 is the limiting reagent in rcTLP reconstitution (22), we generated Rhod5F-labeled rcTLPs by first producing particles with unlabeled VP4 469C, VP7-488, and DLP-647N, purifying the reconstituted particles on a CsCl gradient to eliminate excess VP4 and VP7, and then labeling these 469C rcTLPs sequentially with DBCO-maleimide and Rhod5F azide. This procedure allowed us to compare the infectivities of the rcTLPs with and without the Rhod5F label from the same preparation. As shown in Fig. 8B, the 469C rcTLPs had a specific infectivity (407 P/FFU) similar to that of native TLPs (452 P/FFU). After labeling with the Ca 2ϩ sensor, the Rhod5F rcTLPs had a specific infectivity of 9,290 P/FFU, roughly a 20-fold increase (Fig. 8B). We also observed reduced infectivity when we labeled VP4 469C Rhod5F before rcTLP reconstitution (data not shown). The bulky Rhod5F modification, which includes both the DBCO linker and the Ca 2ϩ indicator itself, together with its proximity to the receptor binding domain VP8*, probably results in steric interference with the sialylated viral receptors on the surface of the cell, thus reducing the likelihood of binding by the Rhod5F rcTLPs.
To define the source of this reduced infectivity, we infected BSC-1 cells at a viral load of about 10,000 particles per cell with Rhod5F rcTLPs (MOI ϭ 1) and VP4 469C rcTLPs (MOI ϭ 20). After a 1-h incubation, cells were washed to remove unbound particles, fixed, and observed by spinning-disk confocal fluorescence microscopy. Of the 10,000 VP4 469C rcTLPs added per cell, an average of 204 particles bound, corresponding to about 2% of the added particles ( Fig. 9A and B and Table 3). An average of 94 of the bound particles, about 46%, lacked a signal in the 488-nm channel and had thus uncoated (see Materials and Methods). An average of 78 Rhod5F rcTLPs had bound per cell (Fig. 9C and D and Table 3), corresponding to 0.78% of the added particles; of these, an average of 26 (33%) had uncoated. The loss of infectivity upon labeling with Rhod5F is thus a result of poor cell attachment; once particles have attached, efficient entry and uncoating appear to be unaffected.
Live-cell imaging of Rhod5F-labeled rcTLPs. We infected BSC-1 cells at an MOI of 10 with Rhod5F rcTLPs and imaged them for 30 min. We tracked the DLP signal in the 642-nm channel, recording 222 uncoating events in 35 cells. Despite sufficient signals for both VP7 and DLP to track particles and classify uncoating events, the relatively weak fluorescence intensity of Rhod5F caused the signal to fall below noise if a particle moved slightly out of the plane of focus during the experiment. For this reason, only 134 of the catalogued Rhod5F rcTLP-uncoating events had enough signal in the 561-nm channel for further analysis of the Ca 2ϩ sensor (Fig. 10).
For these 134 uncoating events, we found that the loss of the Rhod5F signal (i.e., loss of Ca 2ϩ from the vesicle surrounding the particle) began 1.8 Ϯ 0.1 min before the initiation of VP7 signal loss (t 1 Ϫ t c1 ) and 3.3 Ϯ 0.2 min before the end of uncoating (t 2 Ϫ t c1 ) ( Fig. 11 and Table 4). The duration of the entire process, from the initiation of Ca 2ϩ signal loss to DLP release (t 3 Ϫ t c1 ), was 6.7 Ϯ 0.4 min.

DISCUSSION
The results that we report here relate the timing of Ca 2ϩ loss to other molecular events during rotavirus cell entry. In our previous work, we showed that an entering  Visualization of Rotavirus Ca 2ϩ Loss during Entry Journal of Virology particle becomes inaccessible to the external medium within about 5 min of attachment and that release occurs 3 to 5 min later. We have now analyzed in detail the relationship between uncoating (loss of outer layer proteins VP7 and VP4) and release (abrupt onset of rapid diffusional motion of the uncoated DLP) while monitoring directly the leakage of Ca 2ϩ from the vesicle surrounding an engulfed particle. The order of events is as follows: (i) sequestration of the virion from the external medium; (ii) a drop in Ca 2ϩ concentrations to levels well below the K d (dissociation constant) of Rhod5F for Ca 2ϩ (ϳ1 M) and, hence, probably to the concentration of Ca 2ϩ in the cytosol (ϳ0.01 M); (iii) onset of VP7 and VP4 dissociation; (iv) complete loss of detectable outer layer proteins from the vicinity of the DLP; and (v) particle release. We discuss molecular mechanisms for each of these in turn.
(i) The RRV virion elicits the formation of an engulfing vesicle, which closes off about 5 min after initial attachment (12). Our previous work showed not only that penetration is pH independent but also that particles that reach the reduced-pH environment of early endosomes never uncoat. Our experiments confirm these results: pHrodo Red  does not fluoresce at any point along the traces of particles that lose VP4 and VP7, while those particles that show pHrodo fluorescence do not subsequently shed their outer layers, presumably because of their localization in an early endosome. Given that the fluorescence of Rhod5F can be enhanced by the addition of protons near the Ca 2ϩ binding site of the fluorophore at lower pHs, this control is important for confirmation that the Rhod5F signal coincident with an uncoating virion never experiences a decrease in pH. Therefore, the addition of a wavelength to correct for a potential pH-induced Rhod5F signal enhancement can be circumvented in this case, and the three available wavelengths on the spinning-disk confocal microscope can be used to monitor dynamics of Ca 2ϩ , VP7, and DLP.
Although impaired in infectivity, the Rhod5F particles are defective principally at the attachment step. When 10,000 Rhod5F particles/cell were added, we detected 78 associated with the cell after 1 h, of which 26 (33%) had uncoated. The corresponding numbers for unmodified 469C particles were 204 and 94 (46%). Both of these uncoating efficiencies are comparable to those that we have determined previously. Moreover, our recent work has shown that about 10% of uncoated particles become transcriptionally active, at a level detectable by fluorescence in situ hybridization, yielding MOI estimates of 2.6 and 9.4 for Rhod5F and wt particles, respectively. In view of the variation found previously (16), these numbers agree adequately with the directly determined MOIs of 1 and 20. Thus, once attached, Rhod5F and wt particles appear to have indistinguishable properties. The position of the bulky DBCO linker plus the Rhod5F sensor is close to the VP5*-VP8* interface; inefficient attachment is a reasonable consequence of its size and location.
(ii) We estimate from electron micrographs that the inner diameter of the engulfing (comparable to the kinetics of subsequent VP7 loss) probably reflects the dissociation time(s) and not the conductance of the Ca 2ϩ leak in the vesicular membrane.
(iii) Our results show that loss of Ca 2ϩ always precedes loss of the outer layer proteins, an inference previously derived only from biochemical properties of purified virions. They also show that VP4 and VP7 leave the particle together, ruling out any contribution of VP4 dissociation to the loss of signal from Rhod5F. The Rhod5F signal reached baseline approximately 0.5 min before the apparent onset of VP7 dissociation. Thus, whatever compromise in membrane permeability that allows Ca 2ϩ to pass from the vesicle into the cytosol is probably insufficient to allow proteins to leave, implying that Ca 2ϩ loss induces some further, membrane-disrupting steps.
We have suggested previously that the interaction of the hydrophobic loops on VP5* with the vesicular membrane might create an initial Ca 2ϩ leak, loosening the Ca 2ϩ -dependent VP7 trimer contacts and allowing VP5* to undergo a membranedisrupting conformational rearrangement (12). This proposed sequence of molecular events is consistent with the order of fluorescence changes detected in the experiments reported here. Our ongoing structural studies of the VP5* conformational change on the virion surface promise to provide more-detailed, three-dimensional (3D) views of these conformational transitions.
(iv and v) Loss of the outer layer protein fluorescent signal in the diffraction-limited volume surrounding a DLP implies a disruption of the vesicle membrane extensive enough that relatively large proteins can diffuse away, but the mechanism of this disruption, by membrane coupling of the VP5* conformational change, awaits full elucidation. Equally puzzling is the delay between full uncoating and free diffusion of the uncoated DLP. The mean times between complete loss of VP7 and DLP release varied between 2.8 and 3.4 min for the various rcTLPs studied here. Indeed, even the vesicle-enclosed particle, while apparently fully pinched off from the plasma membrane, does not move rapidly away from the cell surface. Thus, the timing of full rupture of the surrounding vesicle is not necessarily defined by the release time of the DLP. Electron microscopy has not yet suggested a mechanism for this retention. If the vesicle or the initially liberated DLP were entrapped by cortical actin, we might expect that escape would require capture of a cytoskeletal motor, but our previous analysis of release showed a trajectory characteristic of free diffusion. New correlated light and electron microscopy studies may help resolve the puzzle.
Many other nonenveloped viruses appear to have evolved mechanisms for using a steep Ca 2ϩ gradient from the outside to the inside of a cell. For example, in polyomavirus particles, both Ca 2ϩ and disulfides lock the interactions between the pentameric assembly units (3,4). Disulfide exchange in the endoplasmic reticulum, an intermediate compartment en route to nuclear entry, is thought to release the disulfide constraints to uncoating; the timing and significance of Ca 2ϩ release have not yet been determined (24)(25)(26). During entry of adenoviruses, signal transduction from a Ca 2ϩ "leak" at the cell surface or from endosomes may facilitate viral uptake and infection (27). The use of a Ca 2ϩ sensor, such as Rhod5F, attached to the virion and hence reporting a relevant local Ca 2ϩ concentration, might clarify the role of Ca 2ϩ in controlling these entryrelated processes.
TLP, DLP, and recombinant protein purification. TLPs, DLPs, VP7, and VP4 were purified as previously described (9). For TLP and DLP production, MA104 cells were grown in 850-cm 2 roller bottles (Corning), and confluent monolayers were infected with rhesus rotavirus (RRV) (G3 serotype) at an MOI of 0.1 focus-forming units (FFU)/cell in M199 medium supplemented with 1 mg/ml porcine pancreatic trypsin (Worthington Biochemical). Cell culture medium was collected at 24 to 36 h postinfection, when cell adherence was Ͻ5%. TLPs and DLPs were purified by freeze-thawing, ultracentrifugation, Freon-113 extraction, and separation on a cesium chloride gradient. TLPs were desalted with a 5-ml Zeba spin column (Thermo Fisher) into a solution containing 20 mM Tris (pH 8.0), 100 mM NaCl, and 1 mM CaCl 2 (TNC). DLPs were desalted into a solution containing 20 mM HEPES (pH 7.5) and 100 mM NaCl (HN). Infectivity of native virions was determined as previously described (22,28), and specific infectivities were determined using concentrations based on densitometry measurements of an SDS-PAGE gel.
VP7 and VP4 were expressed in Sf9 cells infected with a baculovirus vector. VP7 was purified by successive affinity chromatography on concanavalin A and monoclonal antibody (mAb) 159, specific for the VP7 trimer (elution by EDTA). Purified VP7 was desalted into a solution containing 2 mM HEPES (pH 7.5), 10 mM NaCl, and 0.1 mM CaCl 2 (0.1 HNC).
For VP4, harvested cells were lysed by freeze-thawing and clarified by centrifugation after the addition of a cOmplete EDTA-free protease inhibitor (Roche). VP4 was precipitated by the addition of ammonium sulfate to 30% saturation, pelleted, and resuspended in a solution containing 25 mM Tris (pH 8.0), 10 mM NaCl, and 1 mM EDTA (T10NE), which was matched to the conductance of Phenyl HP start buffer (25 mM Tris [pH 8.0], 3.5 M NaCl, 1 mM EDTA), and the solution was loaded onto a Phenyl HP column (GE Healthcare). Following elution with T10NE, fractions containing VP4 were pooled, dialyzed against the same buffer, loaded onto a HiTrap Q column (GE Healthcare), and eluted in Phenyl HP start buffer. Pooled fractions containing VP4 were then concentrated to 600 l with a Centriprep 50 concentrator (Millipore) and subjected to a final purification on an S200 size exclusion column (GE Healthcare) in a solution containing 20 mM HEPES (pH 7.5), 100 mM NaCl, and 1 mM EDTA (HNE).
Fluorescent labeling of DLP, VP7, and VP4. 50 g DLP was brought to a volume of 100 l in HN, to which 11.1 l 1 M NaHCO 3 (pH 8.3) was added. This solution was then added to 1.11 l of 500 g/ml Atto 647N NHS ester dye. The reaction proceeded for 1 h at room temperature before quenching with 12 l of 1 M Tris (pH 8.0). The sample was then desalted through a 0.5-ml Zeba spin column into a solution containing 20 mM Tris (pH 8.0) and 100 mM NaCl (TN).
VP7 was brought to 1.08 mg/ml in a total volume of 75.5 l using 0.1 HNC, and 8.4 l of 1 M NaHCO 3 (pH 8.3) was added. This solution was mixed into 0.84 l of 340 g/ml Atto 488 NHS ester. The reaction proceeded at room temperature for 1 h before quenching with 9 l of 1 M Tris (pH 8.0). The labeled VP7 was then desalted into a solution containing 2 mM Tris, (pH 8.0), 10 mM NaCl, and 0.1 mM CaCl 2 (0.1 TNC).
For Atto 565 labeling, wt VP4 was brought to 5.0 mg/ml in a total volume of 85.8 l with HNE, and 9.6 l of 1 M NaHCO 3 (pH 8.3) was added. This solution was mixed with 0.7 l of 10 mg/ml Atto 565 NHS ester. The reaction proceeded at room temperature for 1 h before quenching with 10 l of 1 M Tris (pH 8.0). The labeled VP4 was then desalted into a solution containing 20 mM Tris (pH 8.0), 100 mM NaCl, and 1 mM EDTA (TNE). The degree of labeling of the protein was found to be 1 Atto 565 molecule to 1 VP4 molecule based on absorbances at 280 nm and 560 nm, using a 280-nm correction factor of 0.164 and an extinction coefficient of the label of 12,000 M Ϫ1 cm Ϫ1 , as described by the manufacturer.
For pHrodo Red labeling, wt VP4 was brought to 5 mg/ml in a total volume of 85.8 l with HNE, and 9.6 l of 1 M NaHCO 3 (pH 8.3) was added. This solution was mixed with 0.44 l of 6.7 mg/ml pHrodo Red. The reaction proceeded at room temperature for 1 h before quenching with 10 l of 1 M Tris (pH 8.0). The labeled VP4 was then desalted into TNE. The degree of labeling of the protein was found to be 1 pHrodo Red molecule to 2 VP4 molecules based on absorbances at 280 nm and 560 nm, using a 280-nm   correction factor of 0.12 and an extinction coefficient of the label of 65,000 M Ϫ1 cm Ϫ1 , as described by the manufacturer. Rhodamine 5F (Rhod5F) azide labeling of monomeric VP4 469C began by reacting the protein with dibenzocyclooctyne-maleimide (DBCO-maleimide; Sigma). A total of 250 l VP4 469C at 3.2 mg/ml in HNE was added to 4 l of 4.9 mg/ml DBCO-maleimide dissolved in dimethyl sulfoxide (DMSO). The reaction was allowed to proceed at room temperature for 1 h before quenching with 1.25 l of 1 M dithiothreitol (DTT) in HNE.
DBCO-labeled VP4 469C was then desalted into HNE and diluted to 1 mg/ml with the same buffer. This protein was then added to enough 1 mM Rhod5F dissolved in DMSO to achieve a final Rhod5F concentration of 57.5 M (about a 5-fold excess of label over protein; roughly 5.7% DMSO in the final reaction mixture). The Rhod5F reaction was allowed to proceed at room temperature for 3 h, followed by overnight incubation at 4°C. The following day, the protein was desalted, using a 5-ml Zeba spin column, into HNE and concentrated with a 50-kDa-cutoff Microcon instrument (Millipore). Labeling of the protein achieved a ratio of 1 Rhod5F molecule to 1 VP4 469C molecule based on absorbances at 280 nm and 548 nm, using a 280-nm correction factor of 0.274 and an extinction coefficient of the label of 78,000 M Ϫ1 cm Ϫ1 .
Labeling of VP4 469C on rcTLPs was performed after the recoating reaction and is outlined below. Sequential chymotrypsin-trypsin digestion of VP4 469C-Rhod5F. Ten micrograms of VP4 469C-Rhod5F was brought to a final volume of 10 l by the addition of a solution containing 20 mM Tris (pH 8.0), 100 mM NaCl, and 0.1 mM EDTA (TN0.1E). A total of 1.26 l of 50 g/ml tosyl-L-lysyl-chloromethane (TLCK)-treated chymotrypsin (Worthington Biochemical) in TNC was added, and the sample was incubated at 37°C for 30 min. A total of 1.16 l of 50 g/ml tosyl phenylalanyl chloromethyl ketone (TPCK)-treated trypsin (Worthington Biochemical) in TNC was then added, and the sample was incubated at room temperature for 1 h. The VP4 CT samples were then quenched by the addition of phenylmethylsulfonyl fluoride (PMSF) to a final concentration of 1 mM. Both samples were mixed with a reducing sample buffer containing a final concentration of 200 mM DTT and either boiled at 95°C for 3 min or incubated at room temperature for 3 min. A 5-to-20% gradient gel was then run with both the boiled and unboiled CT samples as well as 2.8 g of undigested labeled protein for comparison.
Preparation of fluorescently labeled rcTLPs. Recoating of rcTLPs containing Atto 565-labeled wt VP4 was performed according to previously described protocols (22), using the labeled VP7 and DLP prepared as outlined above. Briefly, 1 M sodium acetate (pH 5.2) was added to a volume of Atto-labeled DLP resulting in a final concentration of 100 mM sodium acetate. Atto 565-labeled VP4 was then added to a final concentration of 0.9 mg/ml (ϳ33-fold excess), and the mixture was incubated at room temperature for 1 h. Labeled VP7 was then added in a 2.3-fold excess along with a final addition of sodium acetate and CaCl 2 to maintain their respective concentrations of 100 mM and 1 mM. The mixture was incubated at room temperature for 1 h. Recoated particles were separated from excess labeled components by cesium chloride gradient centrifugation and desalted with a 5-ml Zeba spin column into TNC. Recoating of rcTLPs containing pHrodo Red-labeled wt VP4 performed according to the same protocol, with the particles being desalted with a 5-ml Zeba spin column into a solution containing 20 mM HEPES (pH 7.5), 100 mM NaCl, and 1 mM CaCl 2 (HNC).
rcTLPs containing Rhod5F-labeled VP4 469C were generated by first recoating with unlabeled VP4 469C according to the same protocol as the one outlined above. The purified, unlabeled rcTLPs in TNC were added to DBCO-maleimide such that the final concentration of DBCO-maleimide was 9.5 M, corresponding to a final DMSO volume equal to 1% of the total reaction mixture volume. The reaction was allowed to proceed at room temperature for 2 h, followed by desalting with a 5-ml Zeba column into TNC.
The DBCO-labeled rcTLPs were then added to a volume of 1 mM Rhod5F such that the final label concentration was 20 M, corresponding to a final DMSO volume equal to 2% of the total reaction mixture volume. The reaction mixture was incubated at 4°C overnight, after which the sample was desalted using a 5-ml Zeba column into HNC.
Titers of recoating reaction mixtures were determined by infectious focus assays as previously described (22,28), and specific infectivities were derived through concentration measurements based on densitometry of a Western blot using a primary antibody specific for VP6 (antibody 2B4; Santa Cruz Biotechnology).
For infectivity measurements in the presence of Baf A1, cells were incubated for 1 h at 37°C with unsupplemented DMEM containing 1% DMSO and 100 nM Baf A1. Control cells were similarly incubated in unsupplemented DMEM or DMEM with 1% DMSO for 1 h. pHrodo Red rcTLPs were then added in DMEM containing 100 nM Baf A1 and 1% DMSO, 1% DMSO, or only DMEM, and infection was allowed to proceed for 1 h at 37°C before processing as for the assays described above. 469C and Rhod5F rcTLP binding assay. Number 1.5 round glass coverslips (25 mm; Warner Instruments) were sonicated for 30 min in 3 M NaOH, rinsed thoroughly in water, and sonicated again for 20 min in 70% ethanol (EtOH). Each coverslip was then placed into a well of a 6-well culture plate (Corning) and rinsed twice in phosphate-buffered saline (PBS), followed by two rinses in unsupplemented DMEM. BSC-1 cells were plated in these wells and allowed to grow overnight to 50% confluence in supplemented DMEM.
On the day of the experiment, the coverslips were washed twice with prewarmed DMEM and infected with a viral load of about 10,000 particles per cell, corresponding to an MOI of 20 for 469C rcTLPs and an MOI of 1 for Rhod5F rcTLPs. Infection proceeded for 1 h at 37°C. The cells were then washed three times in HNC and fixed for 10 min at room temperature in HNC with 4% paraformaldehyde. Samples were then washed three times for 5 min in HNC prior to visualization.
Confocal imaging. Fixed-cell images were acquired with a Mariana system (Intelligent Imaging Innovations, Denver, CO) based on a Zeiss Axio-Observer inverted microscope (Carl Zeiss Microimaging Inc., Thornwood, NY) equipped with a CSU-X1 spinning-disk confocal unit (Yokogawa Electric, Tokyo, Japan), a piezo-driven z-translation, and linear encoded x and y translations and controlled with SlideBook V6.0 (Intelligent Imaging Inc., Denver, CO). Excitation wavelengths were 488, 561, and 640 nm (lasers were obtained from Cobolt, Solna, Sweden); the emission filters were 525/50-, 607/36-, and 680-nm long-pass filters (Semrock, Rochester, NY). z-series covering the full volume of the cells were acquired with a z-step of 0.5 m. All wavelengths were acquired with 100-ms exposure times at 100% laser power.
For live-cell imaging, BSC-1 cells were grown on 8-chambered number 1.5 cover glass slides (Cellvis) in a volume of 400 l supplemented DMEM and allowed to grow overnight to 50% confluence. Prior to

FIG 12
Probability densities of 640-and 488-nm channel amplitudes from 469C and Rhod5F rcTLP binding assays. (A and B) Distribution of 640-nm amplitudes derived from particle detections as described in Materials and Methods for all 469C (A) and Rhod5F (B) binding assay samples. Blue and red arrows indicate the upper and lower limits for detection to be considered a single particle; signals weaker than those indicated by the blue arrow were taken as noise, and signals stronger than those indicated by the red arrow were considered aggregates or multiple unresolved particles. (C and D) Distribution of 488-nm amplitudes derived from particle detections for all 469C (C) and Rhod5F (D) binding assay samples (see Materials and Methods). The red arrow points to the cutoff, above which a particle was taken to be a VP7-488-coated rcTLP and below which it was considered uncoated.