Interaction of Virus-Like Particles with Vesicles Containing Glycolipids

Article pubs.acs.org/JPCB

Interaction of Virus-Like Particles with Vesicles Containing Glycolipids: Kinetics of Detachment Waqas Nasir,† Marta Bally,‡,§ Vladimir P. Zhdanov,‡,∥ Göran Larson,† and Fredrik Höök*,‡ †

Department of Clinical Chemistry and Transfusion Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden ‡ Department of Applied Physics, Chalmers University of Technology, SE-412 96 Gothenburg, Sweden § Institut Curie, Centre de Recherche, CNRS, UMR 168, Physico-Chimie Curie, F-75248 Paris, France ∥ Boreskov Institute of Catalysis, Russian Academy of Sciences, Novosibirsk 630090, Russia

Downloaded by STOCKHOLM UNIV on August 29, 2015 | http://pubs.acs.org Publication Date (Web): August 19, 2015 | doi: 10.1021/acs.jpcb.5b04160

S Supporting Information *

ABSTRACT: Many viruses interact with their host cells via glycosphingolipids (GSLs) and/or glycoproteins present on the outer cell membrane. This highly specific interaction includes virion attachment and detachment. The residence time determined by the detachment is particularly interesting, since it is directly related to internalization and infection as well as to virion egress and spreading. In an attempt to deepen the understanding of virion detachment kinetics, we have used total internal reflection fluorescence (TIRF) microscopy to probe the interaction between individual fluorescently labeled GSL-containing lipid vesicles and surface-bound virus-like particles (VLPs) of a norovirus genotype II.4 strain. The distribution of the VLP-vesicle residence time was investigated for seven naturally occurring GSLs, all of which are candidates for the not yet identified receptor(s) mediating norovirus entry into host cells. As expected for interactions involving multiple GSL binding sites at a viral capsid, the detachment kinetics displayed features typical for a broad activation-energy distribution for all GSLs. Detailed inspection of these distributions revealed significant differences among the different GSLs. The results are discussed in terms of strength of the interaction, vesicle size, as well as spatial distribution and clustering of GSLs in the vesicle membrane.

■

INTRODUCTION Viruses are small infectious particles, many of which with a diameter of only 20−300 nm, that carry either a DNA or RNA genome encapsulated in a helical or icosahedral protein capsid. Enveloped viruses are additionally surrounded by a lipid membrane.1 Many enveloped and nonenveloped viruses start the infection process by binding to specific carbohydrates carried on glycolipids (GLs) and/or glycoproteins on the surface of the host cell membrane.2−4 This initial binding is believed to be followed by internalization of the virus into the host cell, where it is uncoated or disassembled and the viral genome is released. After genome replication and synthesis of viral structural proteins, new virions reassemble and finally escape from the cell. Many of the experimental and theoretical studies carried out to understand the replication cycle of viruses have focused on intracellular tracking of single virus particles,5−10 identification of infection-related host factors,11 endocytosis,2−4,12 replication steps,13−17 kinetics of virus assembly,18−24 and inhibition of virion attachment.25−28 There are, however, very few studies specifically addressing the mechanisms governing the initial attachment of virions to the host cell membrane.3 Moreover, attempts to quantify the interaction kinetics between viruses and the host lipid membrane are rare,29−33 and complicated by the fact that the © XXXX American Chemical Society

strength of virus association to the membrane of their host cells is commonly modulated by dynamic engagement of multiple weak laterally mobile bonds. The distribution of the number of these bonds, combined with their strong impact on the residence time of the attachment, make their quantification very demanding. In particular, detachment of virions reversibly bound to GLs and the role of heterogeneity in types, sizes, and distribution of GLs in this process have been barely addressed in the literature.34,35 The mechanistic aspects of virion attachment and detachment are, however, of considerable intrinsic interest and may also be central in applications aimed at the design of entry inhibitors.25,26 Concerning the general importance of the corresponding quantitative data, we may, for example, refer to a recent discussion36 on a trade-off between the need for virions to efficiently attach to and enter uninfected cells and the need for newly generated virions to efficiently detach from the same, but then infected, cells. In order to deepen the understanding of attachment and detachment kinetics of viruses interacting with cell membranes, we have developed a binding assay based on total internal Received: April 30, 2015 Revised: August 10, 2015

A

DOI: 10.1021/acs.jpcb.5b04160 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article


The Journal of Physical Chemistry B

Figure 1. Scheme of the TIRFM-based binding assay used in the present study. The VLPs are immobilized on the lipid bilayer, containing GSLs (H1), and detected via attachment of fluorescently labeled vesicles, which also contain GSLs (top left). The blowup view of the crystal structure of Norwalk virus capsid47 (P domain, red; shell domain, gray; binding sites, yellow) is shown together with the parts of vesicular and planar lipid bilayers (middle). The bilayer image exhibits lipid molecules colored by cyan and water molecules as red balls (bottom left). The glycolipids can be seen as small protrusions (indicated by a small arrow) from the surface of the membrane. The relative molecular scales in this diagram and all reconstructions are based on the actual molecular structures. This setup is used to detect the bound VLPs as spots in TIRF images (right) where each spot represents a vesicle−VLP complex. The attachment and detachment kinetics are recorded on the basis of the residence time analysis (see the Materials and Methods in the Supporting Information for details).

reflection fluorescence microscopy (TIRFM), which makes it possible to study the interaction of individual vesicles containing glycosphingolipids (GSLs; a class of GLs) with single surface-immobilized unlabeled virus-like particles (VLPs).34,35 In this assay, the GSL-containing solid supported membrane is used to immobilize VLPs while the attachment of GSL-containing vesicles to and detachment from the VLPs are employed to mimic and quantify reversible multivalent virion− membrane association, as illustrated in Figure 1. In such experiments, the VLP surface concentration is low, the VLPs are immobile, the formation of VLP dimers or larger aggregates is negligible, the attachment of two or more vesicles to a VLP is negligible as well, and the focus is on single vesicle−VLP pairs.3 Specifically, we track the state of pairs as a function of time. The process of vesicle detachment in each pair is stochastic. This means that, at any given measurement time, each pair is either intact (if the detachment has not happened) or broken (if the detachment has already happened). Cumulatively, we measure the distribution of the vesicle−VLP residence times (see, e.g., Figure 1d in ref 3). Alternatively, the results of measurements can be expressed in terms of the dependence of the number of intact pairs over time. These two ways of presentation of the results are equivalent. Mathematically, the latter way is, however, preferable, because it allows straightforward treatment of the kinetics. Most of the nonenveloped viruses recognize GSLs via the establishment of multiple weak bonds between a virus capsid and cell-surface carbohydrates. One class of viruses of this type is the norovirus, which is a nonenveloped viral pathogen which causes acute gastroenteritis.37,38 To study these and other viruses, VLPs self-assembled from the capsid protein lacking the genetic material essential for infection are typically used, which is motivated by the fact that these VLPs do retain the morphology and binding characteristics of the infectious

virions.39 In the search for the receptors utilized during infection, it is relevant to recall that the norovirus may interact differently with different GSLs on the surface of the host-cell membrane. Some of them are weakly recognized by virions, while others bind with higher affinity. Motivated by this heterogeneity in vivo, the present report is aimed at the analysis of the differences in the interaction of human norovirus-like particles (NVLPs) with seven types of vesicles each containing specific GSLs. In particular, we quantify the detachment kinetics of these vesicles which is an appreciable extension compared to our previous studies on GSL−NVLP interactions.34,35 In previous work, we observed35 that NVLP−vesicle interaction is characterized by a broad distribution of the corresponding residence times and logarithmic time dependence of the detachment kinetics. In general, the kinetics of this special class are very different compared to more conventional exponential or power-law kinetics. Usually, logarithmic kinetics arise when one of the parameter values determining the overall reaction rate is broadly distributed. In our case, the logarithmic features were attributed to a distribution in the activation energy for detachment due to multiple weak bonds and a relatively broad vesicle size distribution. In particular, for the vesicles we use, the radius of the vesicle−VLP contact area was estimated to be between 5 and 12 nm (see ref 3); i.e., the number of bonds in large vesicles might indeed be appreciably higher than that in small vesicles. In our previous data analysis, the simplest rectangular desorption activation-energy distribution was used for vesicle binding to NVLPs.35 With the aim to provide further insights on the specifics of different GSLs, we have in this work employed an improved and more precise energy distribution function. The results are discussed with respect to effects of GSL clustering or the presence of GSL microdomains and lipid rafts, which have been previously B


Article



Figure 2. Fraction of bound vesicles, n(t)/n(0), as a function of time for the residence times from (a) 3 to 300 s with time lapses of 1 s and (c) from 30 to 1500 s with time lapses of 10 s. Panels b and d show the same data as a function of the detachment activation energy, E(t), corresponding to a given time. The activation energy is defined by eq 2 as E(t) = kBT ln(vt) with a pre-exponential factor of 1013 s−1. The measurements were performed at 295 K.

implicated40,41 as a potential factor influencing the entry of viruses into host cells.

was introduced in the early 1980s43 and reviewed in refs 44−46. Its advantage is in the possibility of selectively illuminating particles present in the evanescent field, with an exponentially decaying intensity profile which within ∼100 nm has decreased to 1/e of the intensity at the glass−water interface. This makes it possible to visualize surface-bound vesicles by discriminating them from the ones in solution (Figure 1). Time lapse movies were used to analyze the residence time of the vesicles and to estimate their activation energy of detachment, E(t), corresponding to vesicles leaving VLPs at a given time, t (see eq 2 below).

■

METHODS We have explored the interaction between vesicles containing different types of GSLs and unlabeled single NVLPs. The carbohydrates of seven GSLs used for vesicle preparations were already established as Ast6139 norovirus ligands:42 H type 1 (H-1), A type 1 (A-1), B type 1 (B-1), Lewis a (Lea), Lewis b (Leb), A Lewis b (ALeb), and Lewis y (Ley). NVLPs from the Ast6139/01/Sp GII.4 strain have been employed. The complete details of the measurements are described in the Supporting Information. Briefly, the NVLPs were immobilized on a 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) supported lipid bilayer (SLB) deposited at the glass bottom of a microwell. To ensure NVLP immobilization, the bilayer contained 5 wt % H-1 GSL. After NVLP binding to the SLB and rinsing the well, fluorescent vesicles containing 5 wt % GSLs mixed with POPC lipids were added for detection. The vesicles used for supported lipid bilayers and vesicles used for kinetics binding studies were made by hydrating the lipid film in buffer solution and extrusion through a 100 nm membrane. For detection of vesicles bound to NVLPs, these vesicles also contained an additional 2 wt % of Lissamine-Rhodamine B-1,2dipalmitoyl-sn-glycero-3-phosphoehanolamine (rhodamine-PE, Avanti Lipids, USA). On average, there were about 9000− 11000 dye molecules per vesicle. All vesicles were extruded using the same experimental parameters and in view of the structural similarity and of the low content of GSLs in the vesicles, the influence of GSL type on the vesicle size (diameter) is expected to be minor. The extruded vesicles typically had an average size of 172.8 nm and a full width at half-maximum (fwhm) of 173.5 nm (for example, see the data for H-1 in the Supporting Information) and differences in average size for vesicles containing different GSLs were less than 15%. Time lapse movies of individual events of vesicle attachment to and detachment from NVLPs were recorded at room temperature (295 K) using a TIRF microscope. This technique

■

EXPERIMENTAL AND THEORETICAL RESULTS As mentioned above, the kinetics under consideration are logarithmic due to a distribution in the activation energies for vesicle detachment, which is in turn related to a distribution of the vesicle−VLP binding energies. Physically, this means that for a chosen experimental time window there is a corresponding window of the detachment activation energies. In our experiments, we analyzed two different experimental time/energy windows: for the first set of movies, the vesicles with a residence time between 3 and 300 s were taken into account, whereas, for the second data set, the vesicle residence time was between 30 and 1500 s. The idea was to capture the single vesicle binding kinetics for both weaker and stronger interactions. The details of the analysis and evaluation of TIRFM data are presented in the Supporting Information, including data showing that no detectable interaction was observed for the negative control experiments performed without NVLPs or in the presence of NVLPs but without GSLs in the detecting vesicles. These observations confirm that the VLP binding observed is GSL specific. Basically, in our measurements, we track the ratio n(t)/n(0), where n(t) is the number of vesicles which are bound at t = 0 and remain bound at the time of measurement. This ratio can be shown as a function of t (see, e.g., Figure 2a,c). Although significant differences are observed for the different types of GSLs, it is clear that the corresponding curves display the shape far from those typical for conventional linear, power-law, or C


Article



Figure 3. Schematic of the potential energies along the coordinate for attachment or detachment of vesicles. Each vesicle has its own potential profile which is characterized by the energies of the unbound, transition, and bound states. In our case, the energy of the bound state is different for different vesicles. The energy of the transition state is in principle different as well, but taking into account that the activation barrier for attachment is low, the latter difference is relatively small and can be neglected (as done here). The difference of binding energies gives rise to the distribution of the detachment activation energies, f(E). For vesicles containing different GSLs, the shapes of f(E) are likely to be different. Earlier, f(E) was approximated using the simplest rectangular distribution function.35 In the current study, a more accurate representation is employed to accommodate the differences in the shapes of the distributions with improved accuracy.

vesicles detaching at time t by the condition k(E)t = 1, which yields

exponential kinetics. Our previous experience indicates that the shape becomes simple if the kinetics are exhibited as a function of ln(t), as shown in Figure 2b,d. As already noted, this means that the kinetics are logarithmic, and this feature is expected to be related to a relatively broad distribution, f(E), of the detachment activation energies of attached vesicles. Focusing on the distribution of activation energies, we consider that the detachment kinetics of each vesicle−VLP pair is exponential and do not describe in detail formation and rupture of multiple weak bonds in single pairs (for the model focused on the latter aspect, see, e.g., refs 29−33 and 48). Schematically, the distribution function is illustrated in Figure 3 for situations when its width is wider (Figure 3a,b) and more narrow (Figure 3c) than the interval of activation energies corresponding to the measured residence time interval. A complete determination of f(E) that could fully represent the heterogeneity of the detachment kinetics observed in our experiments requires a complete data set of single vesicle binding events including those with the shortest and longest residence times. Due to experimental limitations, we here cover residence times over 3 orders of magnitude (from 3 to 1500 s). Note that this interval is wide and its lower boundary is appreciably shorter than the resolution time typical for conventional ensemble averaging methods (see, e.g., the quartz crystal microbalance data in Figure S3). For this reason, compared to the latter methods, we are able to track the contribution to the kinetics of weakly interacting vesicle−virion pairs, with short residence times. Still, however, our singlevesicle measurements do not cover the entire range of the detachment times, and accordingly, we identify the main part of f(E) excluding the tails. To interpret the kinetics under consideration, we recall and extend the formalism used earlier.34,35 In particular, taking into account that the detachment kinetics of each vesicle is exponential, the ensemble of vesicles is described as n(t ) = n(0)

∫E

Emax min

f (E) exp[−k(E)t ] dE

E(t ) = kBT ln vt

(2)

Physically, it is clear that at a given time t the pairs with E < E(t) are almost completely depleted while the pairs with E > E(t) are nearly intact. In a very good approximation which has been widely used for logarithmic kinetics already a few decades (see, e.g., ref 49 and references therein), we consider that the pairs with E < E(t) are completely depleted while the pairs with E > E(t) are intact, and accordingly rewrite eq 1 as n(t ) ≅ n(0)

Emax

∫E(t)

f (E ) d E

(3)

For the simplest rectangular distribution, f(E) = 1/(Emax − Emin), eq 3 yields35 ⎛ W ⎞ ⎛ kBT ⎞ ⎟ − ln t ln⎜ ⎟ = ln⎜ ⎝ ΔE ⎠ n (0) ⎝ ⎠

(4)

where W = −dn/dt is the detachment rate. Equation 4 demonstrates that ln(W/n(0)) depends linearly on ln(t) with a slope of −1. Here we observed (see Figure 4) that, during our experiments with ligands of different composition and sizes, some of the GSL-containing vesicles showed deviations of 10− 20% from the slope value of −1 predicted by eq 4.

(1) Figure 4. Examples of logarithmic plots of detachment rates of GSLcontaining vesicles. Note the deviation from the slope of −1 especially for Leb. The data shown correspond to residence times from 3 to 300 s. The symbols x and y in the figure represent ln(t) and ln(W/n(0)), respectively.

where k(E) ≡ v exp(−E/kBT) is the detachment rate constant (v is the pre-exponential factor) and Emin ≤ E ≤ Emax is the whole range of activation energies (Figure 2). If f(E) is broad, one can specify the activation energy corresponding to the D


Article



Figure 5. Detachment activation energy plots ( f(E) vs E(t); left y-axes indicated by an arrow) of some of the GSL-containing vesicles bound to NVLPs obtained using eq 5 with parameters generated from fitting the experimental data to eq 6 (n(t)/n(0) vs E(t); right y-axes indicated by an arrow) for the data sets corresponding to the time frames from 3 to 300 s (red lines and “*”) and from 30 to 1500 s (blue lines and “+”) together with fits to eq 6 (solid black lines). Detachment is considered for only the vesicles within these time frames.

In these cases, the underlying distribution, f(E), is likely to be somewhat skewed and could not be properly approximated using the simplest rectangular function. To test this hypothesis, we suggest the following representation

plots for the data sets with the time windows from 3 to 300 s and from 30 to 1500 s. The distributions f(E) corresponding to these two time windows characterize fragments of the full distributions, and accordingly, they are expected to be proportional to each other in the common energy interval where they can be directly compared. Due to the statistical errors, this condition is fulfilled approximately. Finally, we note that that our treatment of the kinetics is based on the use of the ratio n(t)/n(0). The corresponding derivative, F(t) ≡ [dn(t)/dt]/n(0), represents the distribution of the residence times. Thus, n(t)/n(0) can easily be converted into F(t); i.e., as already noticed in the Introduction, n(t)/n(0) and F(t) can be employed interchangeably.

f (E) = a − b(E − E )2 (5) * where a and b are positive fitting parameters and E* is the most probable value of E. One of these parameters can be expressed via another one by normalizing f(E) to unity. If the information about a system is complete, the choice of f(E) is unique. In reality, however, the kinetics are measured in a certain time interval, and the construction of f(E) depends on the interval chosen (see the discussion below). Substituting eq 5 into eq 3, we obtain after integration:

■

DISCUSSION AND CONCLUSIONS In this study, we have investigated the detachment kinetics of individual NVLP−bound vesicles containing specific GSLs. In all cases, the NVLPs were immobilized using a supported lipid bilayer containing 5% H-1. In principle, release events of both vesicles and vesicle−VLP complexes could be contributing to the observed kinetics. In view of the fact that the bilayers were extensively rinsed after NVLP incubation and considering that the NVLPs are mostly irreversibly bound as verified by a realtime biosensing experiment (Supporting Information), the latter events are likely to be scarce. This thus strongly suggests that our TIRFM binding experiments represent vesicle release from single NVLPs firmly bound to the surface. We have compared vesicles exhibiting similar distributions in sizes but containing different types of glycosphingolipids at the same concentration. A large number of lipid types and analysis of the detachment behavior of the corresponding vesicles with a more accurate representation of the probability distribution function of detachment activation energies reveal differences in the binding behavior. An obvious difference between the binding systems stems from differences in the monovalent binding strength of the different GSLs under investigation. Even though there are no direct calorimetric data available on ΔG values, such differences have been predicted by molecular dynamics simulations.50,51 As a matter of fact, under the assumption that the size distribution, GSL distribution, and GSL concentration of different vesicles are similar, the

n(t ) b b ≅ E(t )3 − bE E(t )2 − (a − bE 2)E(t ) − Emax 3 * * n(0) 3 3 + bE Emax 2 + (a − bE 2)Emax (6) * * Equation 6 has been used to analyze the detachment kinetics of vesicles in our measurements. The examples of the fits to the data sets (n(t)/n(0) vs E(t) with R2 = 0.99 for H-1, Lea, and ALeb or 0.98 for Leb) are shown in Figure 5, together with the corresponding probability distribution function plots [f(E) vs E] for the parameters a, b, and E* obtained from the fit. The derivation of eq 6 assumes that the vesicle detachment can be tracked from the beginning to the end, i.e., from tmin = exp(Emin/kBT)/v to tmax = exp(Emax/kBT)/v. However, under experimental conditions, it is likely that tmin is shorter than the minimum experimental time. Similarly, tmax corresponding to the maximum desorption energy could be beyond the maximum time that was observed through our experiments. Therefore, in our case, Emin and Emax are identified with kBT ln vtmin = 0.78 and 0.84 eV and kBT ln vtmax = 0.90 and 0.95 eV, where tmin and tmax represent minimal (3 and 30 s) and maximum (300 and 1500 s) experimental times. This means that the vesicles taking part in the reaction before or after the minimal or maximal experimental times are excluded. The data from our measurements can be well represented by eq 6 for all the vesicle species, as shown in Figure 5 for H-1, Lea, Leb, and ALeb together with the corresponding f(E) vs E E


Article



being unrealistic. The biological cell membrane is a complex mix-up of GSLs and glycoproteins (GPs) of different chemical compositions and sizes.2−4 The fluidity of the cell membrane and the lateral movement of GSLs and GPs make the cellular membrane extremely heterogeneous. Moreover, the GSLs and GPs tend to cluster in the membrane in the form of microdomains or lipid rafts.52 These microdomains have been hypothesized to play a key role not only in cell signaling and protein sorting53 but also in the infection process of various viruses.54,55 Although lipid-domain formation on the membrane of vesicles with a diameter in the 100 nm regime is not straightforward to measure directly, it would be interesting to in forthcoming work investigate how the detachment kinetics depends on lipid compositions known to facilitate domain formation. Furthermore, since detachment kinetics is expected to be directly related to the probability of virus internalization and infection, such investigations, combined with addition of compounds designed to inhibit the interaction, could potentially aid the development of new antiviral drugs.

difference in detachment activation energies should reflect the difference in the binding strengths at the molecular level. It is, for example, interesting to note that theoretical binding data on NVLP−carbohydrate interactions predicts that ALeb binds strongly to NVLPs in line with our experimental observations where ALeb has the largest population of vesicles which did not desorb during the experimental time (Figure 2). Regarding detachment kinetics, it is instructive to recall that the simplest binding systems (e.g., a simple ligand−receptor interaction) typically exhibit unique and system-specific activation energy of detachment (Ed), as illustrated in Figure 3c. For such homogeneous cases, detachment kinetics data can be fitted with a single exponential function and the detachment rate constant (koff) can be directly estimated. The system under consideration in this work does not meet this criteria, and multiexponential detachment behavior is observed in the form of logarithmic kinetics. This can be conceptually explained by the multivalent character of the virus−membrane interaction associated with the broad size distribution of vesicles in our system. As a matter of fact, depending on the propensity of a vesicle to deform upon virus binding, and hence on vesicle size, between 6 and 12 bonds are likely to be formed, as previously deduced from simple geometrical considerations.35 A distribution in the number of GSL−NVLP bonds formed leads to a distribution in the vesicle’s activation energy of detachment, reflected in a distribution in the rates of detachment, as schematically illustrated in Figure 3a,b. The heterogeneity in vesicle sizes is not expected to considerably influence the attachment kinetics, though, because this is typically mediated by the formation of one or a few simultaneous weak bonds. The present work was therefore focused on estimating the probability distribution function of detachment activation energy for such a heterogeneous system. Previously, the simplest rectangular function was used to approximate the distribution for vesicles with GSLs. In this work, we employ a distribution function which allows one to take its different shapes into account (Figure 3). This enabled us to analyze the detachment kinetics of several GSL-containing vesicles of different types and sizes with improved accuracy. Figure 5 shows parts of the actual f(E) vs E plots corresponding to the two time windows, 3−300 and 30−1500 s, used in this work. The vesicles containing H-1 and Lea show the distribution plots for longer time window which are nearly parallel to the x-axis. This suggests that, for these GSLcontaining vesicles, we experimentally probe the vesicle population with activation energies near a broad nearly flat maximum of f(E); i.e., f(E) is nearly rectangular. This is not true for the rest of the GSL-containing vesicles, especially those with ALeb, where the experiments track the regions near one or both tails of the distributions in these cases. This piece-wise information on the distribution of detachment activation energies for the population of a certain type of GSL-containing vesicles gives new insights into the overall energetic landscape of the system under investigation. For example, it cannot be excluded that ligand clustering and hence the presence of GSL microdomains contribute to the observed heterogeneity and differences in detachment kinetics between the different GSL species. Considering, for instance, the case of Leb, it is clear that the distribution of detachment activation energies is more narrow than those of the other GSLcontaining vesicles, which might be indicative of a GSL clustering effect in this particular population of vesicles. In fact, the idea that glycolipids cluster into microdomains is far from

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.5b04160. Details of the materials and methods; data showing VLP specific interactions in the TIRFM setup; parameter values calculated by fitting data to eq 6 (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +46 31 772 61 30. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by grants from the Swedish Research Council (2013-8266 to G.L., 2014-5557 to F.H. and V.P.Z., 2012-5024 to M.B.), VINNOVA and Governmental Grants to the Sahlgrenska University Hospital (G.L.).

■

REFERENCES

(1) Cann, A. Principles of Molecular Virology; Academic Press: Oxford, UK, 2012. (2) Dimitrov, D. S. Virus entry: molecular mechanisms and biomedical applications. Nat. Rev. Microbiol. 2004, 2, 109−122. (3) Mercer, J.; Schelhaas, M.; Helenius, A. Virus entry by endocytosis. Annu. Rev. Biochem. 2010, 79, 803−33. (4) Ravindran, M. S.; Tanner, L. B.; Wenk, M. R. Sialic acid linkage in glycosphingolipids is a molecular correlate for trafficking and delivery of extracellular cargo. Traffic 2013, 14, 1182−1191. (5) Seisenberger, G.; Ried, M. U.; Endress, T.; Buning, H.; Hallek, M.; Brauchle, C. Real-time single-molecule imaging of the infection pathway of an adeno-associated virus. Science 2001, 294, 1929−32. (6) Endress, T.; Lampe, M.; Briggs, J. A.; Krausslich, H. G.; Brauchle, C.; Muller, B.; Lamb, D. C. HIV-1-cellular interactions analyzed by single virus tracing. Eur. Biophys. J. 2008, 37, 1291−301. (7) Brauchle, C.; Seisenberger, G.; Endress, T.; Ried, M. U.; Buning, H.; Hallek, M. Single virus tracing: visualization of the infection pathway of a virus into a living cell. ChemPhysChem 2002, 3, 299−303. (8) Ewers, H.; Smith, A. E.; Sbalzarini, I. F.; Lilie, H.; Koumoutsakos, P.; Helenius, A. Single-particle tracking of murine polyoma virus-like particles on live cells and artificial membranes. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 15110−5.

F


Article


The Journal of Physical Chemistry B (9) Lakadamyali, M.; Rust, M. J.; Babcock, H. P.; Zhuang, X. Visualizing infection of individual influenza viruses. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 9280−5. (10) Sivaraman, D.; Biswas, P.; Cella, L. N.; Yates, M. V.; Chen, W. Detecting RNA viruses in living mammalian cells by fluorescence microscopy. Trends Biotechnol. 2011, 29, 307−313. (11) Peng, X.; Chan, E. Y.; Li, Y.; Diamond, D. L.; Korth, M. J.; Katze, M. G. Virus−host interactions: from systems biology to translational research. Curr. Opin. Microbiol. 2009, 12, 432−438. (12) Zhdanov, V. P. Physical aspects of the initial phase of endocytosis. Phys. Rev. E 2013, 88, 064701. (13) Yin, J. Chemical engineering and virology: Challenges and opportunities at the interface. AIChE J. 2007, 53, 2202−2209. (14) Sidorenko, Y.; Schulze-Horsel, J.; Voigt, A.; Reichl, U.; Kienle, A. Stochastic population balance modeling of influenza virus replication in vaccine production processes. Chem. Eng. Sci. 2008, 63, 157−169. (15) Sardanyés, J.; Martínez, F.; Daròs, J.-A.; Elena, S. F. Dynamics of alternative modes of RNA replication for positive-sense RNA viruses. J. R. Soc., Interface 2012, 9, 768−776. (16) Timm, A.; Yin, J. Kinetics of virus production from single cells. Virology 2012, 424, 11−17. (17) Zhdanov, V. P. Stochastic kinetics of reproduction of virions inside a cell. BioSystems 2004, 77, 143−50. (18) Zandi, R.; van der Schoot, P.; Reguera, D.; Kegel, W.; Reiss, H. Classical nucleation theory of virus capsids. Biophys. J. 2006, 90, 1939− 1948. (19) Morozov, A. Y.; Bruinsma, R. F.; Rudnick, J. Assembly of viruses and the pseudo-law of mass action. J. Chem. Phys. 2009, 131, 155101. (20) Hagan, M. F.; Elrad, O. M. Understanding the concentration dependence of viral capsid assembly kineticsthe origin of the lag time and identifying the critical nucleus size. Biophys. J. 2010, 98, 1065−1074. (21) Dykeman, E. C.; Stockley, P. G.; Twarock, R. Building a viral capsid in the presence of genomic RNA. Phys. Rev. E 2013, 87, 022717. (22) Zlotnick, A.; Porterfield, J. Z.; Wang, J. C.-Y. To build a virus on a nucleic acid substrate. Biophys. J. 2013, 104, 1595−1604. (23) Mateu, M. G. Assembly, stability and dynamics of virus capsids. Arch. Biochem. Biophys. 2013, 531, 65−79. (24) Zhdanov, V. P. Viral capsids: Kinetics of assembly under transient conditions and kinetics of disassembly. Phys. Rev. E 2014, 90, 042721. (25) Law, M.; Hangartner, L. Antibodies against viruses: passive and active immunization. Curr. Opin. Immunol. 2008, 20, 486−492. (26) Tilton, J. C.; Doms, R. W. Entry inhibitors in the treatment of HIV-1 infection. Antiviral Res. 2010, 85, 91−100. (27) Magnus, C.; Regoes, R. R. Estimating the stoichiometry of HIV neutralization. PLoS Comput. Biol. 2010, 6, e1000713. (28) Zhdanov, V. P. Inhibition of the receptor-mediated virion attachment to a lipid membrane. Cent. Eur. J. Phys. 2012, 10, 1210− 1215. (29) Chou, T. Stochastic entry of enveloped viruses: fusion versus endocytosis. Biophys. J. 2007, 93, 1116−23. (30) Chou, T.; D’Orsogna, M. R. Multistage adsorption of diffusing macromolecules and viruses. J. Chem. Phys. 2007, 127, 105101. (31) Nowak, S. A.; Chou, T. Mechanisms of receptor/coreceptormediated entry of enveloped viruses. Biophys. J. 2009, 96, 2624−36. (32) Gibbons, M. M.; Chou, T.; D’Orsogna, M. R. Diffusiondependent mechanisms of receptor engagement and viral entry. J. Phys. Chem. B 2010, 114, 15403−12. (33) Zhdanov, V. P. Kinetics of virus entry by endocytosis. Phys. Rev. E 2015, 91, 042715. (34) Bally, M.; Dimitrievski, K.; Larson, G.; Zhdanov, V. P.; Hook, F. Interaction of virions with membrane glycolipids. Phys. Biol. 2012, 9, 026011. (35) Bally, M.; Gunnarsson, A.; Svensson, L.; Larson, G.; Zhdanov, V. P.; Hook, F. Interaction of single viruslike particles with vesicles containing glycosphingolipids. Phys. Rev. Lett. 2011, 107, 188103.

(36) Handel, A.; Akin, V.; Pilyugin, S. S.; Zarnitsyna, V.; Antia, R. How sticky should a virus be? The impact of virus binding and release on transmission fitness using influenza as an example. J. R. Soc., Interface 2014, 11, 20131083. (37) Rydell, G. E.; Kindberg, E.; Larson, G.; Svensson, L. Susceptibility to winter vomiting disease: A sweet matter. Rev. Med. Virol. 2011, 21, 370−82. (38) Donaldson, E. F.; Lindesmith, L. C.; Lobue, A. D.; Baric, R. S. Viral shape-shifting: norovirus evasion of the human immune system. Nat. Rev. Microbiol. 2010, 8, 231−41. (39) Jiang, X.; Wang, M.; Graham, D. Y.; Estes, M. K. Expression, self-assembly, and antigenicity of the Norwalk virus capsid protein. J. Virol. 1992, 66, 6527−32. (40) Chazal, N.; Gerlier, D. Virus entry, assembly, budding, and membrane rafts. Microbiol. Mol. Biol. Rev. 2003, 67, 226−237. (41) Sun, X.; Whittaker, G. R. Role for influenza virus envelope cholesterol in virus entry and infection. J. Virol. 2003, 77, 12543− 12551. (42) Fiege, B.; Rademacher, C.; Cartmell, J.; Kitov, P. I.; Parra, F.; Peters, T. Molecular details of the recognition of blood group antigens by a human norovirus as determined by STD NMR spectroscopy. Angew. Chem., Int. Ed. 2012, 51, 928−32. (43) Axelrod, D.; Burghardt, T. P.; Thompson, N. L. Total internal reflection fluorescence. Annu. Rev. Biophys. Bioeng. 1984, 13, 247−268. (44) Wazawa, T.; Ueda, M. Total Internal Reflection Fluorescence Microscopy in Single Molecule Nanobioscience. Microscopy Techniques 2005, 95, 77−106. (45) Schneckenburger, H. Total internal reflection fluorescence microscopy: Technical innovations and novel applications. Curr. Opin. Biotechnol. 2005, 16, 13−18. (46) Thompson, N. L.; Wang, X.; Navaratnarajah, P. Total internal reflection with fluorescence correlation spectroscopy: Applications to substrate-supported planar membranes. J. Struct. Biol. 2009, 168, 95− 106. (47) Prasad, B. V.; Hardy, M. E.; Dokland, T.; Bella, J.; Rossmann, M. G.; Estes, M. K. X-ray crystallographic structure of the Norwalk virus capsid. Science 1999, 286, 287−90. (48) Chesla, S. E.; Selvaraj, P.; Zhu, C. Measuring two-dimensional receptor-ligand binding kinetics by micropipette. Biophys. J. 1998, 75, 1553−1572. (49) Tachiya, M.; Mozumder, A. Decay of trapped electrons by tunnelling to scavenger molecules in low-temperature glasses. Chem. Phys. Lett. 1974, 28, 87−89. (50) Koppisetty, C. A.; Nasir, W.; Strino, F.; Rydell, G. E.; Larson, G.; Nyholm, P. G. Comput-ational studies on the interaction of ABOactive saccharides with the norovirus VA387 capsid protein can explain experimental binding data. J. Comput.-Aided Mol. Des. 2010, 24, 423− 31. (51) Nasir, W.; Frank, M.; Koppisetty, C. A.; Larson, G.; Nyholm, P. G. Lewis histo-blood group alpha1,3/alpha1,4 fucose residues may both mediate binding to GII.4 noroviruses. Glycobiology 2012, 22, 1163−72. (52) Simons, K.; Ikonen, E. Functional rafts in cell membranes. Nature 1997, 387, 569−72. (53) Fan, J.; Sammalkorpi, M.; Haataja, M. Formation and regulation of lipid microdomains in cell membranes: theory, modeling, and speculation. FEBS Lett. 2010, 584, 1678−84. (54) Veit, M.; Engel, S.; Thaa, B.; Scolari, S.; Herrmann, A. Lipid domain association of influenza virus proteins detected by dynamic fluorescence microscopy techniques. Cell. Microbiol. 2013, 15, 179−89. (55) Ono, A. Relationships between plasma membrane microdomains and HIV-1 assembly. Biol. Cell. 2010, 102, 335−50.

G


Interaction of Virus-Like Particles with Vesicles Containing Glycolipids

Recommend Documents