Mobility-Based Quantification of Multivalent Virus-Receptor

Subscriber access provided by Iowa State University | Library

Communication

Mobility-Based Quantification of Multivalent Virus-Receptor Interactions: New Insights Into Influenza A Virus Binding Mode Matthias Müller, Daniel Lauster, Helen H. K. Wildenauer, Andreas Herrmann, and Stephan Block Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.8b04969 • Publication Date (Web): 05 Feb 2019 Downloaded from http://pubs.acs.org on February 5, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Mobility-Based Quantification of Multivalent Virus-Receptor Interactions: New Insights Into Influenza A Virus Binding Mode

Matthias Müller1, Daniel Lauster2, Helen H. K. Wildenauer1, Andreas Herrmann2, Stephan Block1,*

1

Department of Chemistry and Biochemistry, Emmy-Noether Group “Bionanointerfaces”,

Freie Universität Berlin, Takustr. 3, 14195 Berlin, Germany 2

Department of Biology, Molecular Biophysics, Humboldt-Universität zu Berlin,

Invalidenstr. 42, 10115 Berlin, Germany

*Corresponding author: Stephan Block, e-mail: [email protected]

Keywords: multivalent interactions, influenza A virus, single particle tracking, TIRF microscopy, Evans-Sackmann model, binding inhibition

1

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ABSTRACT Viruses, such as influenza A, typically bind to the plasma membrane of their host by engaging multiple membrane receptors in parallel, thereby forming so-called multivalent interactions that are created by the collective action of multiple weak ligand-receptor bonds. The overall interaction strength can be modulated by changing the number of engaged receptors. This feature is used by viruses to achieve a sufficiently firm attachment to the host’s plasma membrane, but also allows progeny viruses to leave the plasma membrane after completing the virus replication cycle. Design of strategies to prevent infection, for example by disturbing these attachment and detachment processes upon application of antivirals, requires quantification of the underlying multivalent interaction in absence and presence of antivirals. This is still an unresolved problem, as there is currently no approach available that allows for determining the valency (i.e., of the number of receptors bound to a particular virus) on the level of single viruses under equilibrium conditions. Herein, we track the motion of single influenza A/X31 viruses (IAVs; interacting with the ganglioside GD1a incorporated in a supported lipid bilayer) using total internal reflection fluorescence microscopy and show that IAV residence time distributions can be deconvoluted from valency effects by taking the IAV mobility into account. The so-derived off-rate distributions, expressed in dependence of an average, apparent valency, show the expected decrease in off-rate with increasing valency, but also show an unexpected peak structure, which can be linked to a competition in the opposing functionalities of the two influenza A virus spike proteins, hemagglutinin (HA) and neuraminidase (NA). By application of the antiviral zanamivir that inhibits the activity of NA, we provide direct evidence, how the HA-NA balance modulates this virus-receptor interaction, allowing us to assess the inhibition concentration of zanamivir based on its effect on the multivalent interaction.

INTRODUCTION Multivalent interactions (i.e., multiple, non-covalent ligand-receptor bonds acting in parallel) are typical for a multitude of biological processes.1-4 They are observed, for example, in the attachment of viruses to the membrane of their host cells, a dynamic process that eventually leads to internalization and infection.5 Design of strategies to prevent infection, for example, by inhibiting the attachment of viruses (or pathogens in general), requires quantification of the underlying multivalent interaction in absence and presence of antivirals.6 This is still an unresolved problem as the individual ligand-receptor association is typically weak in the 2


Page 2 of 21

Page 3 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

sense that multiple bonds are required to establish an association with sufficient residence time to enable cellular entry.7 Accordingly, each bond is maintained only transiently, causing temporal fluctuations of the valency (i.e., of the number of receptors bound to a particular virus).8 Quantification of multivalent interactions therefore requires methods offering single pathogen resolution, since only in this way heterogeneities and dynamics, which are typically hidden in conventional ensemble-averaging approaches, can be scrutinized. In recent studies, approaches combining total internal reflection fluorescence (TIRF) microscopy and single particle tracking (SPT) proved to be very powerful in characterizing the dynamics of virus-receptor interactions under equilibrium conditions.9-16 In such measurements, the receptor is located at an interface (e.g., at a glass surface or embedded within a supported lipid bilayer (SLB) serving as artificial cell membrane mimic). The attachment of viruses to and detachment from their receptors is monitored using TIRF microscopy. Although this concept has been successfully applied to probe, for example, the interaction kinetics of noroviruses binding to glycosphingolipids or to determine the interaction strength of influenza A viruses (IAVs) that bind to different sialic acid-containing glycolipids, the valency of the virus-receptor interactions could not be quantified from broad distributions in the measured residence times.9-10 The lack of knowledge about the distribution of the valency among viruses makes it impossible to judge, if changes in the binding kinetics of viruses (e.g., upon application of antivirals) are caused by changes in the valency distribution or in the properties of the single, monovalent virus-receptor interaction or both. Motivated by our previous work on SLB-tethered liposomes,4, 17 we estimate here the valency of single IAVs (interacting with SLB-embedded receptors) by measuring the IAV mobility using SPT, allowing the complex residence time distribution to be deconvoluted from valency effects. This enables us to extract the valency-dependent off-rate distribution of a virusreceptor interaction and to quantify, how off-rate distributions and on-rates are modified by application of an IAV antiviral (zanamivir). In particular, our results shed new light on the interplay between the two viral spike proteins hemagglutinin (HA) and neuraminidase (NA) in binding of IAV to the host cell surface. While HA with its sialic acid binding pocket is considered as the most important component for attachment, NA with its enzymatic activity of cleaving sialic acid moieties is regarded as the main factor for releasing of newly formed IAVs from the host cell. We now show that despite the much greater number of HA in the ratio of 5:1 to NA and the enzymatic activity of NA,18 NA plays an important role in attachment of virus to the host cell and that the binding of IAVs to sialic acid-containing glycolipids is strongly dependent on the functional balance of HA and NA. 3


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

RESULTS AND DISCUSSION In detail, the multivalent interaction of IAVs (strain X31) with sialic acid-presenting lipids is monitored in this study using TIRF microscopy and followed by SPT. All reported experiments have been conducted at room temperature (20 °C). In order to mimic the fluid environment that virus receptors experience within the plasma membrane of cells, a supported lipid bilayer (SLB), mainly consisting of POPC supplemented by a small fraction of the sialic acid-presenting ganglioside GD1a ( 1 mol%), is formed on top of a glass interface (Fig. 1a).11 TIRF illumination is used to excite only those (fluorescently labeled) IAVs that are bound to the SLB, while non-attached viruses (moving within the bulk solution) will not be excited, since they are outside of the excitation volume of the evanescent wave (Fig. 1a, green area).9 Hence, TIRF microscopy allows for probing the binding process of IAVs to their receptors in dependence of the receptor concentration and thus for quantifying IAV attachment and release kinetics at the single-virus level.14 In general, binding of single IAVs to the SLB are detected in SPT by a local increase of the fluorescence intensity (Fig. 1b). Some IAVs remain SLB-bound over the entire observation time, while others show only a transient interaction that leads, due to IAV release, to a decrease in the fluorescence intensity after a particular period (Fig. 1b, bottom) and thus to a limited IAV residence time (see Supporting Table S1 for further details on these different IAV populations). In addition, SPT also allows for tracking the lateral motion of SLB-bound IAVs. Quantification of the lateral diffusion coefficients D yielded broad distributions (Supporting Fig. S1a), ranging between fully immobile IAVs (Fig. 1b, black line) and IAVs showing 2D Brownian diffusion above the SLB (Fig. 1b, blue and red line). In a first set of experiments, the specificity of the IAV-SLB interaction was determined by quantifying the IAV attachment rate in dependence of the receptor concentration (i.e. the amount of GD1a lipids supplemented to the SLB). This was achieved by applying the socalled equilibrium fluctuation analysis (EFA) to the SPT data, which determines the number of newly arriving viruses between consecutive frames of the TIRF microscopy movies. Afterwards, the cumulative number of newly arriving viruses over time is calculated (Fig. 2a).9, 11 Only IAVs that remained SLB-bound for at least 3 frames have been included in this analysis. This corresponds to an IAV residence time of at least 0.5 seconds, which is much larger than the 10 ms needed for IAVs to diffuse through the evanescent wave of the TIRF illumination. Hence, only those IAVs are included in the EFA procedure that showed true SLB attachment. 4


Page 4 of 21

Page 5 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

As in related studies, the cumulative number of newly arriving viruses increased linearly with time for all measurements. The slope of these traces corresponds to an apparent on-rate kon of the IAV-SLB interaction.9, 11, 19 At a bulk IAV concentration of 0.7 pM (Supporting Fig. 2), kon increased from 0.01 to 0.17 IAVs per µm² and min for GD1a concentrations ranging between 0 and 0.075 mol% and leveled out at 0.17 IAVs per µm² and min for GD1a concentrations exceeding 0.075 mol% (Fig. 2a, inset). This saturation behavior is a typical hallmark of multivalent interactions and has been observed for a variety of virus species and measurement approaches.8, 12, 20 Furthermore, as the saturation of kon is already observed for GD1a concentrations exceeding ~0.1 mol%, no binding experiments were performed beyond a GD1a concentration of 1 mol%, although the thus achieved sialic acid surface densities (~0.03 SA moieties per nm² of the SLB, see section “Receptor Concentration within the SLBs” in the Supporting Information) were at least one order of magnitude below the values reported in literature (ranging between 0.2 and 2 SA per nm² for cell membranes).1, 21-22 It should be noted, however, that the latter values correspond to the total amount of sialic acids presented by the cell, which are only in part linked to glycostructures close to the cell’s plasma membrane (thereby promoting virus attachment), while the rest is linked to structures on top of the plasma membrane (e.g., to biological hydrogels such as the glycocalix, forming a permeation barrier for viruses).1 Nevertheless, kon in Fig. 2a levels out at a value being >10 times larger than that derived in absence of receptors in the SLB, indicating that the TIRFbased assay indeed probes the specific interaction between IAVs and sialic acid-presenting GD1a lipids. In addition, the EFA procedure also provides information about the off-rate of the probed interaction. This is typically derived by extracting the virus residence time tres for all binding events and by calculating the survival probability ps according to (Fig. 1c): 𝑡

𝑡

𝑝𝑠(𝑡res) = ∑∆𝑡𝑚𝑎𝑥= 𝑡 𝑁(∆𝑡)/∑∆𝑡𝑚𝑎𝑥= 0𝑁(∆𝑡).

(1)

res

In this equation, N(t) gives the number of IAVs being tracked for a time period of at least t and tmax denotes the maximum time lag included in the analysis.14 Equation 1 basically gives the probability that an IAV remains bound to the receptor-functionalized surface after residing for the residence time tres.11 For a monovalent interaction that dissociates spontaneously, it can be shown that the survival probability decays exponentially with a decay rate that corresponds to the off-rate koff of the interaction:19 𝑝𝑠(𝑡res) = 𝑒𝑥𝑝( ― 𝑘off ∙ 𝑡res) .

(2) 5


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The ps distributions derived for the IAV-sialic acid interaction show, however, a much more complex structure, which clearly deviates from a simple exponential decay (Fig. 2b). Most viruses (approximately 70 %) detach already within the first 10 s, while a minor IAV population (roughly 5 - 10 %) remains SLB-bound for residence times exceeding 1 min (Fig. 2b and Supporting Table S1). A qualitatively similar behavior has been reported in previous studies probing the interaction of IAVs with GD1a.11 A direct application of equations (1) and (2) to the SPT data apparently fails in deriving the off-rate kinetics, as the IAV-sialic acid interaction is multivalent.9, 11, 23 This means that each virus binds to the SLB by engaging a certain number of GD1a lipids, the value of which we will denote as the valency n(t) in the following. Owing to the transient nature of multivalent interactions,4 the valency n(t) of each virus changes over time t, since its value is 0 prior to virus attachment and after virus detachment, while various values may be adopted in between of these processes.8 As the off-rate koff of a multivalent interaction strongly depends on the valency n,24 any distribution of valency values among the IAV population will automatically lead to a broad IAV off-rate distribution.9 Thus, the survival probability ps is a superposition of multiple exponential decays, each of which contributes with a different decay rate (= off-rate) and magnitude to ps. Apparently, the usual approach of extracting off-rates using EFA fails here in absence of information about the IAV valency distribution. To overcome this problem, we investigated here the possibility to extract information about the IAV valency based on the lateral motion of SLB-bound viruses. The underlying rationale is motivated by our previous studies, in which we showed that the valency n of SLBbound particles (such as liposomes), which are linked to multiple lipids of the SLB, strongly influences the particle’s lateral diffusion coefficient D as long as the particles are smaller than 400 nm in diameter.4 This is a direct consequence of the fact that fluid-phase SLBs are highly viscous, so that the hydrodynamic friction of nanoscopic, SLB-bound particles within the bulk can be fully neglected in comparison to the hydrodynamic friction experienced by the associated lipids within the SLB. Hence, D is solely determined by the diffusion of the associated lipids in this case.25 Higher numbers of bound lipids (= higher valency) increase the overall friction experienced in the SLB and therefore decrease the diffusion coefficient D.26-29 For DNA-linked liposomes, the relationship connecting valency and diffusivity could be quantified, leading to D(n) = D0/n, in which D0 corresponded to the diffusion coefficient of a single cholesterol-equipped DNA-linker.4

6


Page 6 of 21

Page 7 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

In contrast to this previous work, the interaction of IAVs with the SLB was too fast to allow for extracting time-dependent diffusion coefficients D(t). Even using acquisition rates on the order of 20 fps did not allow to identify distinct mobility values of SLB-linked IAVs, so that for further analysis the average diffusion coefficient of the IAV, D=, was used, which is, nevertheless, expected to be indicative for the average valency experienced by the IAV during interaction with the SLB. These considerations suggest that the impact of multivalent effects on the analysis of IAV residence times can be simplified, if only those IAVs are included in the calculation of the residence time distribution, which are similar in their D value (and thus in their average valency n). This is in contrast to the typical EFA procedure, in which all observed IAVs are included in the calculation of ps. This procedure requires to dissect the measured D distribution of the entire IAV population into smaller IAV subpopulations (having D values ranging between particular boundary values; see Supporting Information section “Survival Probability and Off-Rate Analysis of IAV Subpopulations” for details) and to calculate the residence time distribution for each subpopulation independently. Applied to the IAV SPT data, this simple modification indeed yields residence time distributions that follow a single exponential decay (Fig. 3a). The decay rates of these traces (= absolute value of the slopes in Fig. 3a) correspond by definition to the off-rate koff of the IAV subpopulation having the indicated D value. Hence, despite the fact that the exact relationship connecting valency and diffusivity has yet to be resolved for IAVs, pooling only those IAV tracks that have a similar D value in the extraction of the residence time distributions indeed leads to very simple distributions that lack the typical signatures of multivalency (as in Fig. 2b). This feature is attributed to the fact that all models describing the motion of lipid oligomers within SLBs generally suggest a monotonously decreasing D value for an increasing number of associated lipids.25-28 Hence, all IAV tracks having a certain average D value are supposed to share the same average valency (experienced while being bound to the SLB) and therefore the same off-rate, leading to exponentially decaying tres distributions. As, according to these considerations, a lower D value corresponds to a larger valency, it is reasonable to plot the extracted off-rates in dependence of the inverse diffusion coefficient 1/D. This corresponds to evaluating the relationship between koff and an average, apparent valency n (Fig. 3b). In general, koff decreases with increasing n. This is the expected behavior, as an increase in n requires more bonds to be broken to allow for IAV release, thereby decreasing the rate of the overall process. Both features, that is the observation of exponentially decaying residence time distributions as well as the decrease of koff with 7


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

increasing 1/D value, indicates that the approximation of using average values for diffusion coefficient (and thus valency) instead of the time-resolved values is here sufficient to allow for extracting valency-dependent off-rate distributions of X31 IAVs. Surprisingly, even in absence of receptors in the SLB (Fig. 3b, POPC) some IAVs are able to interact sufficiently long with the SLB (> 0.5 s) to be included in the off-rate analysis. We attribute this feature to an unspecific interaction occurring between IAVs and the SLB, as the corresponding off-rate distribution increases much faster (i.e., for much smaller values of D) than the koff distribution extracted in presence of receptors (Fig. 3b, 1 mol% GD1a). Nevertheless, a comparison of the on-rates of these 2 cases (see Supporting Table S1) shows that these unspecific events occur only rarely: While the on-rate is 0.17 IAVs/µm2/min for a GD1a concentration of 1 mol%, the on-rate drops by a factor of 15 (0.011 IAVs/µm2/min) in absence of any GD1a in the SLB. This shows that the unspecific interaction observed here can be neglected when analyzing the binding dynamics of IAVs in presence of 1 mol% GD1a. Furthermore, it should be noted that the exact relationship connecting valency and D has still to be resolved for this interaction (and virus-receptor interactions in general). Hence, it is not yet possible to directly convert an observed D value into an absolute value of the average valency experienced by the corresponding IAVs. Despite this obstacle, the approach still allows for a qualitative assessment of the extracted valency-dependent off-rates, such as those shown in Fig. 3b, since all theoretical models agree that an increase in valency causes the D value to decrease monotonously. Hence, the shown koff(valency) plots may become stretched or squeezed along the valency-axis, once the exact valency-mobility relationship has been resolved, but their general shape will remain unaffected by these transformations. Interestingly, in presence of GD1a receptors koff does not decrease monotonously with apparent valency, but exhibits a peak at approximately (1/D =) 5 s/µm² (Fig. 3b, red trace; this peak is magnified in the inset of Fig. 3b). Hence, at 1/D = 5 s/µm² the IAVs detach with an off-rate, which is higher than expected from a monotonously decreasing koff(n) relationship. To our knowledge, this feature has not yet been reported in the literature. This observation is indicative for an interaction scheme that is more complex than just involving interactions between hemagglutinin and sialic acids. It is therefore straightforward to hypothesize that the peak in koff is caused by the activity of neuraminidase, which is known to cleave sialic acid moieties from membranes and therefore promotes release of (progeny) IAVs from cell membranes.30-31 In order to test the hypothesis that the peak in koff is caused by a competition between the opposing functionalities of the spike proteins hemagglutinin and neuraminidase (attachment versus release), a commercially available neuraminidase inhibitor (zanamivir) 8


Page 8 of 21

Page 9 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

was added to the IAV solutions and valency-dependent koff distributions were extracted in dependence of the zanamivir concentration, cinh. If the peak in koff is indeed caused by the activity of neuraminidase, then the off-rate is expected to decrease and the peak in koff is expected to vanish upon zanamivir addition. These expectations are indeed confirmed by the experiments, which show that koff in general decreases with increasing cinh (Fig. 4a and Fig. 4b). As the off-rate now depends on two parameters, 1/D and cinh, the impact of the neuraminidase inhibitor on the multivalent IAV-sialic acid interaction can be plotted in two different ways, either in dependence of 1/D for constant values of cinh (Fig. 4a) or in dependence of cinh for constant values of 1/D (i.e., at a constant average valency; Fig. 4b). Fitting a Langmuir-type inhibition curve 𝑘off(𝑐inh) = 𝑘off,0 +𝐴/(1 + 𝑐inh/IC50)

(3)

to the latter showed that the off-rate decreases in a dose-dependent manner at given average, apparent valency (Fig. 4b, solid lines), leading to zanamivir IC50 values being on the order of 2.5 µM (Fig. 4c). As the single zanamivir-neuraminidase interaction has a Ki value on the order of 1 nM,32 this relatively large IC50 value indicates that an effect on the off-rate is resolvable, once most of the IAV neuraminidase proteins are blocked by zanamivir. The observed decrease in koff with increasing cinh matches well to the commonly accepted mode of action of neuraminidase inhibitors, which are assumed to hinder the egress of progeny viruses by blocking the sialic acid-cleaving activity of neuraminidase.31 Based on this picture, one may assume that neuraminidase inhibitors such as zanamivir mainly influence the release process of IAVs. When evaluating the apparent on-rate kon of IAV attachment to the SLBs, we observed, however, that the inhibition of neuraminidase also has a very strong effect on kon (Fig. 4d). At a zanamivir concentration of 3 µM, for example, kon increased by a factor of 3 with respect to the on-rate measured in absence zanamivir. This result indicates that the mode of action of zanamivir is a combination of two different effects, an increase in kon and a decrease in koff, which together hinder IAV egress by shifting the IAV-sialic acid interaction to the attachment side. These considerations are in line with recent studies showing that neuraminidase is not solely responsible for cleaving of sialic acids to release progeny IAVs but also contributes IAV attachment and entry.33 The effect of neuraminidase inhibition on the molecular interaction of IAVs with glycostructures has been recently investigated using different biochemical and biophysical approaches. Using SPT, Sakai et al.15 studied the attachment and motion of IAVs being bound to sialic acids-functionalized glass slides as well as to live cells (the latter done in absence and 9


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

presence of a NA inhibitor). Despite using receptors that are laterally immobile at the glass surface, Sakai et al. still observed lateral diffusion of IAVs. This observation was attributed to the fact that NA cleaves sialic acids being engaged in IAV binding, leading to a clearance of receptors from the interface and thus allowing for a rolling of IAVs on the surface. In principle, a similar mechanism could also contribute to the lateral IAV motion observed in our assay. The contribution due to IAV rolling, however, is negligible in comparison to the lateral IAV motion that is caused by the diffusion of the receptor-presenting lipids within the SLB. This follows directly from the fact that the rolling-mediated diffusion coefficients, observed by Sakai et al., are at least one order of magnitude smaller than the lateral diffusion coefficients in our analysis. Hence, the lateral IAV motion observed on SLBs is dominated by the lateral diffusion of IAV-engaged lipid receptors, a finding which is in agreement with recent observations for the polyoma virus SV40.34 In a related set of experiments, Guo et al.35 use biolayer interferometric analysis (BLI) to probe the interaction of various IAV strains to sialic acid-functionalized glycostructures being fully immobile (that is, unable to diffuse laterally) at sensor surfaces. They show that neuraminidase activity strongly influences the dynamics of the IAV-sialic acid interaction and that virus dissociation was only possible after the receptor density at the surface had been sufficiently decreased by NA. While our findings are qualitatively in agreement with the former observation by Guo et al., they are in disagreement with the latter, as we always observe an IAV subpopulation that only transiently binds to the SLB. We attribute this discrepancy to deviations in the sensitivity levels of the different approaches (single-virus in SPT, while BLI is an ensemble-averaging approach). This interpretation is supported by a recent work of Lee et al.,11 who also observe a transiently binding population (indicated by a quick decay of the survival probability at low residence times). Unfortunately, Guo et al. did not report on changes in IAV attachment rate with increasing neuraminidase inhibition, so that a validation of our findings in Fig. 4d is not possible. They rather quantified the attachment rate of different IAV strains and receptor structures at a certain NA inhibitor concentration, followed by probing the dissociation of sensor-bound IAVs after removal of the neuraminidase inhibitor, which was mainly driven by NA-mediated sialic acid cleavage from the interface. Our assay, however, does not show any signs for significant sialic acid clearance from the SLB, which manifests itself, for example, in constant interaction properties even after measurements times exceeding 1 hour. We attribute this different behavior to the fact that SPT allows for probing IAV concentrations being as small as 0.7 pM, while most BLI experiments of Guo et al. were performed at 100 pM, which automatically leads to much 10


Page 10 of 21

Page 11 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

faster sialic acid cleavage. In this sense, our approach and those reported in literature are complementary and probe different aspects of the complex IAV-sialic acid interaction, thereby yielding different insights of this highly relevant system.

CONCLUSIONS In conclusion, TIRF microscopy combined with SPT was used to probe the interaction of IAVs (strain X31) with sialic acids exposed by GD1a-supplemented SLBs. As in related studies, the EFA procedure allowed for extracting apparent on-rates of the IAV-sialic acid interaction, but failed in generating residence time distributions being suitable for off-rate extraction. This problem is caused by pooling IAVs differing in their valency, creating multivalency-related convolution effects in the residence time distribution, which cannot be addressed in lack of information of the IAV valency. We circumvent this problem based on a new analysis scheme, orthogonally extending the conventional EFA procedure, which estimates the IAV valency distribution based on a measurement of IAV mobility on the SLB surface (i.e., its diffusion coefficient, D). By restricting the calculation of residence time distributions to IAVs sharing the same D value (i.e., by performing the EFA off-rate analysis to IAV subpopulations of given D instead of pooling all observed IAV), this problem is effectively addressed, allowing valency-dependent off-rates of the IAV-sialic acid interaction to be extracted. Using this new concept, it is shown that the detachment of IAV is ruled by a sensitive balance of the IAV spike proteins hemagglutinin (HA, binding to sialic acid moieties on cell membranes) and neuraminidase (NA, cleaving sialic acid moieties on cell membranes). The competition of HA and NA leads to a peak structure in the off-rate plots at intermediate values of the average, apparent valency 1/D. By addition of a NA inhibitor (zanamivir) that shifts the “functional balance” towards HA, we show how the IAV-sialic acid interaction is affected by changes in the HA-NA balance and how the induced modifications can be analysed in a valency-dependent manner. We observe that the peak structure in koff vanishes upon zanamivir application in a dose-dependent manner, allowing us to extract IC50 values based on the functional effect of zanamivir on the IAV-sialic acid interaction. This decrease in koff upon zanamivir application is in line with the established mode of action of this neuraminidase inhibitor, aiming to “glue” progeny IAVs to the plasma membrane of infected cells, which hinders further spreading of IAVs. However, application of zanamivir also caused a strong (and yet unreported) increase in the on-rate of the IAV-sialic interaction, 11


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

which was unexpected and suggests a strong impact of NA activity on IAV attachment. Hence, the neuraminidase inhibitor zanamivir shows in our SLB-based assay a Janus-like behavior, as it does “glue” progeny IAVs to cell membranes by decreasing the off-rate, but also promotes IAV attachment to membrane. As the relationship between virus attachment and uptake rate is highly complex, further experiments (for example, live-cell imaging of IAV uptake in absence and presence of neuraminidase inhibitors) will be necessary to determine the extent, by which the observed increase in attachment rate translates into an increase in cellular uptake rate of IAVs.

ACKNOWLEDGEMENTS This work was supported by the German Research Foundation (BL1514/1 and project C6 within the CRC 765) and the Focus Area Nanoscale (Freie Universtät Berlin).

ASSOCIATED CONTENT The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/xxx. Additional information and figures: Materials and methods, measurement statistics, distributions of diffusion coefficients and residence times, size distributions of IAVs in bulk and at the SLB, survival probability and off-rate analysis of IAV subpopulations, receptor concentration within the SLBs (PDF).

12


Page 12 of 21

Page 13 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

Mammen, M.; Choi, S. K.; Whitesides, G. M., Polyvalent interactions in biological systems: Implications for design and use of multivalent ligands and inhibitors. Angew. Chem. Int. Ed. Engl. 1998, 37 (20), 2755-2794. Barrow, E.; Nicola, A. V.; Liu, J., Multiscale perspectives of virus entry via endocytosis. Virol. J. 2013, 10, 177. Mercer, J.; Schelhaas, M.; Helenius, A., Virus entry by endocytosis. Annu. Rev. Biochem. 2010, 79, 803-33. Block, S.; Zhdanov, V. P.; Höök, F., Quantification of Multivalent Interactions by Tracking Single Biological Nanoparticle Mobility on a Lipid Membrane. Nano Lett. 2016, 16 (7), 43824390. Marsh, M.; Helenius, A., Virus entry: Open sesame. Cell 2006, 124 (4), 729-740. Fasting, C.; Schalley, C. A.; Weber, M.; Seitz, O.; Hecht, S.; Koksch, B.; Dernedde, J.; Graf, C.; Knapp, E. W.; Haag, R., Multivalency as a Chemical Organization and Action Principle. Angew. Chem., Int. Ed. Engl. 2012, 51 (42), 10472-10498. Sun, E.; He, J.; Zhuang, X. W., Live cell imaging of viral entry. Curr. Opin. Virol. 2013, 3 (1), 3443. Szklarczyk, O. M.; Gonzalez-Segredo, N.; Kukura, P.; Oppenheim, A.; Choquet, D.; Sandoghdar, V.; Helenius, A.; Sbalzarini, I. F.; Ewers, H., Receptor Concentration and Diffusivity Control Multivalent Binding of Sv40 to Membrane Bilayers. PLoS Comput. Biol. 2013, 9 (11). Bally, M.; Gunnarsson, A.; Svensson, L.; Larson, G.; Zhdanov, V. P.; Höök, F., Interaction of Single Viruslike Particles with Vesicles Containing Glycosphingolipids. Phys. Rev. Lett. 2011, 107 (18). Lee, D. W.; Allison, A. B.; Bacon, K. B.; Parrish, C. R.; Daniel, S., Single-Particle Tracking Shows that a Point Mutation in the Carnivore Parvovirus Capsid Switches Binding between HostSpecific Transferrin Receptors. J. Virol. 2016, 90 (9), 4849-4853. Lee, D. W.; Hsu, H. L.; Bacon, K. B.; Daniel, S., Image Restoration and Analysis of Influenza Virions Binding to Membrane Receptors Reveal Adhesion-Strengthening Kinetics. PloS One 2016, 11 (10). Parveen, N.; Block, S.; Zhdanov, V. P.; Rydel, G. E.; Hook, F., Detachment of Membrane Bound Virions by Competitive Ligand Binding Induced Receptor Depletion. Langmuir 2017, 33 (16), 4049-4056. Peerboom, N.; Block, S.; Altgarde, N.; Wahlsten, O.; Moller, S.; Schnabelrauch, M.; Trybala, E.; Bergstrom, T.; Bally, M., Binding Kinetics and Lateral Mobility of HSV-1 on End-Grafted Sulfated Glycosaminoglycans. Biophys. J. 2017, 113 (6), 1223-1234. Peerboom, N.; Schmidt, E.; Trybala, E.; Block, S.; Bergstrom, T.; Pace, H. P.; Bally, M., Cell Membrane Derived Platform To Study Virus Binding Kinetics and Diffusion with Single Particle Sensitivity. ACS Infect. Dis. 2018, 4 (6), 944-953. Sakai, T.; Nishimura, S. I.; Naito, T.; Saito, M., Influenza A virus hemagglutinin and neuraminidase act as novel motile machinery. Sci. Rep. 2017, 7. Goronzy, I. N.; Rawle, R. J.; Boxer, S. G.; Kasson, P. M., Cholesterol enhances influenza binding avidity by controlling nanoscale receptor clustering. Chem. Sci. 2018, 9 (8), 2340-2347. Block, S.; Fast, B. J.; Lundgren, A.; Zhdanov, V. P.; Hook, F., Two-dimensional flow nanometry of biological nanoparticles for accurate determination of their size and emission intensity. Nat. Comm. 2016, 7. Harris, A.; Cardone, G.; Winkler, D. C.; Heymann, J. B.; Brecher, M.; White, J. M.; Steven, A. C., Influenza virus pleiomorphy characterized by cryoelectron tomography. PNAS 2006, 103 (50), 19123-19127. Gunnarsson, A.; Dexlin, L.; Wallin, P.; Svedhem, S.; Jonsson, P.; Wingren, C.; Hook, F., Kinetics of Ligand Binding to Membrane Receptors from Equilibrium Fluctuation Analysis of Single Binding Events. JACS 2011, 133 (38), 14852-14855. 13


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

20.

21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

33. 34. 35.

Parveen, N.; Rimkute, I.; Block, S.; Rydell, G. E.; Midtvedt, D.; Larson, G.; Hytonen, V. P.; Zhdanov, V. P.; Lundgren, A.; Hook, F., Membrane Deformation Induces Clustering of Norovirus Bound to Glycosphingolipids in a Supported Cell-Membrane Mimic. J. Phys. Chem. Lett. 2018, 9 (9), 2278-2284. Seaman, G. V. F.; Knox, R. J.; Nordt, F. J.; Regan, D. H., Red-Cell Aging .1. Surface-Charge Density and Sialic-Acid Content of Density-Fractionated Human Erythrocytes. Blood 1977, 50 (6), 1001-1011. Mammen, M.; Dahmann, G.; Whitesides, G. M., Effective Inhibitors of Hemagglutination by Influenza-Virus Synthesized from Polymers Having Active Ester Groups - Insight into Mechanism of Inhibition. J. Med. Chem. 1995, 38 (21), 4179-4190. Sieben, C.; Kappel, C.; Zhu, R.; Wozniak, A.; Rankl, C.; Hinterdorfer, P.; Grubmuller, H.; Herrmann, A., Influenza virus binds its host cell using multiple dynamic interactions. PNAS 2012, 109 (34), 13626-13631. Evans, E., Probing the relation between force - Lifetime - and chemistry in single molecular bonds. Annu. Rev. Biophys. Biomol. Struct. 2001, 30, 105-128. Block, S., Brownian Motion at Lipid Membranes: A Comparison of Hydrodynamic Models Describing and Experiments Quantifying Diffusion within Lipid Bilayers. Biomolecules 2018, 8 (2). Saffman, P. G.; Delbruck, M., Brownian-Motion in Biological-Membranes. PNAS 1975, 72 (8), 3111-3113. Evans, E.; Sackmann, E., Translational and Rotational Drag Coefficients for a Disk Moving in a Liquid Membrane-Associated with a Rigid Substrate. J. Fluid Mech. 1988, 194, 553-561. Stone, H. A.; Ajdari, A., Hydrodynamics of particles embedded in a flat surfactant layer overlying a subphase of finite depth. J. Fluid Mech. 1998, 369, 151-173. Knight, J. D.; Lerner, M. G.; Marcano-Velazquez, J. G.; Pastor, R. W.; Falke, J. J., Single Molecule Diffusion of Membrane-Bound Proteins: Window into Lipid Contacts and Bilayer Dynamics. Biophys. J. 2010, 99 (9), 2879-2887. Das, K.; Aramini, J. M.; Ma, L. C.; Krug, R. M.; Arnold, E., Structures of influenza A proteins and insights into antiviral drug targets. Nat. Struct. Mol. Biol. 2010, 17 (5), 530-538. De Clercq, E.; Li, G. D., Approved Antiviral Drugs over the Past 50 Years. Clin. Microbiol. Rev. 2016, 29 (3), 695-747. von Itzstein, M.; Wu, W. Y.; Kok, G. B.; Pegg, M. S.; Dyason, J. C.; Jin, B.; Phan, T. V.; Smythe, M. L.; White, H. F.; Oliver, S. W.; Colman, P. M.; Varghese, J. N.; Ryan, D. M.; Woods, J. M.; Bethell, R. C.; Hotham, V. J.; Cameron, J. M.; Penn, C. R., Rational Design of Potent SialidaseBased Inhibitors of Influenza-Virus Replication. Nature 1993, 363 (6428), 418-423. Yang, J.; Liu, S. W.; Du, L. Y.; Jiang, S. B., A new role of neuraminidase (NA) in the influenza virus life cycle: implication for developing NA inhibitors with novel mechanism of action. Rev. Med. Virol. 2016, 26 (4), 242-250. Kukura, P.; Ewers, H.; Muller, C.; Renn, A.; Helenius, A.; Sandoghdar, V., High-speed nanoscopic tracking of the position and orientation of a single virus. Nat. Meth. 2009, 6 (12), 923-U85. Guo, H. B.; Rabouw, H.; Slomp, A.; Dai, M. L.; van der Vegt, F.; van Lent, J. W. M.; McBride, R.; Paulson, J. C.; de Groot, R. J.; van Kuppeveld, F. J. M.; de Vries, E.; de Haan, C. A. M., Kinetic analysis of the influenza A virus HA/NA balance reveals contribution of NA to virus-receptor binding and NA-dependent rolling on receptor-containing surfaces. PloS Pathog. 2018, 14 (8).

14


Page 14 of 21

Page 15 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 1. Scheme of the TIRF-based assay to probe multivalent interactions between influenza A viruses (IAVs) and sialic acids. (a) Supported lipid bilayers (SLBs) containing the sialic acid-presenting ganglioside GD1a were used as artificial cell membranes. Fluorescently labeled IAVs (R18 dye incorporated in the virus envelope) bind to sialic acids exposed by the SLB, a process which is followed by TIRF imaging (always performed at room temperature; 20 °C). Due to the evanescent excitation of TIRF (a, light green area; penetration depth of approximately 150 nm) only those of the IAVs are visible that are firmly bound to the SLB. (b) Single particle tracking (SPT) is used to follow the IAV-SLB interaction at the single-virus level. Individual IAVs are detected by a local increase in 15


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

fluorescence intensity (b, top panel, showing zoomed-in fluorescence images of 3 representative IAVs). SPT is applied to track the motion of SLB-bound IAVs and to determine their fluorescence intensity over time (b, bottom panel). Virus release is evidenced by a sudden decrease in the IAV fluorescence intensity, allowing the residence time, that is the period between attachment and release of a particular IAV, to be determined for each virus. While being bound to the SLB, IAVs may be fully immobile (b, black trace) or diffuse laterally (b, red and blue trace). Hence, SPT allows for determining the rate of IAV attachment and of diffusion coefficients and residence times of individual IAVs.

16


Page 16 of 21

Page 17 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 2. Equilibrium fluctuation analysis (EFA) confirms that IAVs bind specifically to GD1a-containing SLBs. (a) The cumulative number of newly arriving IAVs, calculated according to the EFA procedure, increases linearly with time, the slope of which is proportional to the on-rate kon of the IAV-GD1a interaction. kon strongly increases with GD1a concentration, confirming the specificity of the IAV-sialic acid interaction, and saturates for concentrations exceeding 0.075 mol % (a, inset). (b) This is also reflected when calculating the survival probability ps of the interaction (i.e., the probability to observe IAVs being SLBbound for the residence time tres as indicated; Eq. (1) main text) according to the EFA procedure, which increases roughly by one order of magnitude in presence of GD1a. Nevertheless, ps shows a very complex dependence of tres that cannot be directly translated into off-rates koff (further discussed within the main text).

17


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 3. Taking the IAV diffusion coefficient D into account allows for deconvoluting residence time distributions from multivalency-related effects. (a) Restricting the analysis of tres distributions to IAVs having a similar D value leads to distributions that follow a simple exponential decay (see a for three representative D values; extracted from an experiment conducted on a SLB containing 1 mol % GD1a). The decay rates of these traces (= absolute value of the slopes in a) correspond to the off-rate koff of the particular IAV subpopulation (sharing the same D value as indicated). All extracted distributions are given in Fig. S3 (Supporting Information). (b) According to theoretical considerations (detailed in the main text), it is expected that the IAV diffusion coefficient D decreases with increasing average valency (experienced during the interaction with the SLB). Hence, plotting the off-rates koff obtained from fitting the tres distributions (such as those in a) versus 1/D formally corresponds to correlate koff with an average, apparent valency of the IAV-sialic acid interaction, which in general shows the theoretically expected decrease in koff with increasing valency (b, blue data points). In presence of GD1a (b, red data points), however, this decrease is superimposed by a peak structure leading to elevated koff values at 1/D ~ 5 s/μm2. Data points and error intervals correspond to slope and standard error of the slope, respectively, obtained by fitting an exponential decay to the corresponding tres distributions (see Supporting Information for details). All experiments were performed at room temperature.

18


Page 18 of 21

Page 19 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 4. The peak structure in the valency-dependent koff is caused by neuraminidase. (a) Addition of the neuraminidase inhibitor zanamivir causes the peak of koff to vanish progressively with increasing zanamivir concentrations cinh (as indicated in the plot). Solid lines are polynomial interpolations of the data points and were added to guide the eye. The data was extracted from experiments conducted on a 1 mol % GD1a SLB (measured at room temperature). Data points and error intervals correspond to slope and standard error of the slope, respectively, obtained by fitting an exponential decay to the corresponding tres distributions (see Supporting Information for details). (b) The decrease in koff for increasing cinh was evaluated at 1/D values as indicated by the solid arrows and generally shows a dosedependent relationship between koff and cinh (b, solid lines showing fits using a Langmuir-type inhibition model, Eq. 3). To allow for a direct comparison of the traces, all off-rates have been normalized with respect to the koff value observed in absence of zanamivir, while the inset shows the same data using a semilogarithmic scaling. (c) Fitting Eq. 3 allows for extracting IC50 values of the inhibitor zanamivir, which show no significant dependence of the 1/D 19


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

values investigated. Data points and error intervals correspond to the fitted IC50 value and the standard error of the fit, respectively. (d) Zanamivir modifies not only the off-rate but also the on-rate distribution of the IAV-sialic acid interaction. The plot shows kon, quantified according to the EFA procedure, in absence of zanamivir for GD1a concentrations as indicated (d, bars left to the dashed line; same data as in Fig. 2a, inset) and at 1 mol% GD1a for zanamivir concentrations as indicated (d, bars right to the dashed line). All on-rates have been normalized with respect to the kon value observed at 1 mol% GD1a in absence of zanamivir. Data points and error intervals correspond to average value and standard deviation of kon derived from multiple replications of the measurement. Significance was tested using a t-test (**: p < 0.01; ***: p < 0.001).

20


Page 20 of 21

Page 21 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

for TOC only

21


Mobility-Based Quantification of Multivalent Virus-Receptor

Recommend Documents