Exceptional mechanical and structural stability of HSV-1 unveiled with fluid atomic force microscopy

Virus particles are known to undergo significant structural changes throughout the virus life cycle, and evidence is emerging that these alterations in structure are accompanied by marked changes in the mechanical properties of the virus particle. The biological significance of the mechanical alterations, however, is poorly understood. To give relevant insight into understanding the correlation between structural and mechanical changes of virus particles some demands must be met. Structural and mechanical investigations should preferably be performed simultaneously, at high resolution, and most importantly, under conditions closely mimicking those encountered by the virus particle in a host cell. To date, there is only one approach that meets all these challenging demands – atomic force microscopy (AFM). The invention of AFM (Binnig et al., 1986) opened up new realms for our perception of biology. AFM produces three-dimensional images of biological surfaces at molecular scale (Engel and Muller, 2000) in a physiologically relevant environment (Del Sol et al., 2007; Greenleaf et al., 2007). Beyond this one-of-a-kind capability, AFM can also be applied to biological samples for a variety of key mechanical investigations, such as volume and stiffness measurements at the nano-scale (Hillebrand et al., 2006; Schafer et al., 2007). It is therefore unsurprising that AFM qualifies as a suitable technique for uncovering physiologically relevant alterations of viral structural assemblies by modelling corresponding steps of the viral life cycle in vitro (Kuznetsov et al., 2003). Until recently, viruses were predominantly checked only for their mechanical properties when studied with AFM. A short time ago, however, two AFM studies dealing with murine leukemia and human immunodeficiency virus particles were published, which implied that changes in the mechanical properties of the virus particles during maturation are paralleled by structural changes of the virus particles, and that such combined changes are key to infectivity (Kol et al., 2006; Kol et al., 2007). In the present study, AFM was applied to investigate structure and mechanical properties of HSV-1 capsids in a physiologically relevant environment in situ. The choice of HSV-1 capsid for the present study was obvious for two reasons: (1) HSV-1 is one of the most widespread human viruses and (2) its structure has been described with considerable detail (Zhou et al., 2000). Thereby, the cryo-electron microscopy (cryo-EM) model on which much of our current knowledge of the structure of HSV-1 capsids rests provides a standard against which the AFM studies presented here can be compared. Despite being well studied structurally, the basic biology of HSV-1 still poses some unresolved questions. One particular step in the HSV-1 life cycle remains enigmatic. This is the process of viral genome release at the nuclear pore. In the present study, we investigated the behavior of HSV-1 capsids in terms of DNA release on various abiotic surfaces with different physical-chemical properties. Additional mechanical and chemical stimuli applied to the surface-bound capsids have further improved our understanding of a potential mechanism of the genome release.

HSV-1 capsids have been characterized structurally by AFM and electron microscopy (EM) (Fig. 1). EM analysis of negatively stained capsid preparations showed that treatment of HSV-1 virions with 0.2% Triton X-100 rendered capsids free of envelopes (Fig. 1A). AFM analysis of large areas (Fig. 1B) of capsid preparations adsorbed onto a poly-L-lysine-coated surface (in buffered solution) revealed that the particles were spherical and homogenous. The dimension of the capsids was determined by their height rather than their width. The width of the capsid is less accurate and appears to be larger than the height because of convolution between the AFM tip and the capsid. The height of HSV-1 capsids ranged from 120 to 130 nm (Fig. 1D). This finding is in close agreement with cryo-EM observations showing that the diameter of the viral capsid is approximately 125 nm (Zhou et al., 2000), and confirmed that the capsid preparation used in the present study was essentially free of debris and degraded subviral structures. At higher magnifications and lower scanning speed, the hexagonal outline of the HSV-1 capsids becomes apparent (Fig. 1C). Furthermore, and in good agreement with cryo-EM images of HSV-1, AFM revealed underlying patterning of the capsid substructures, such as apparent triangles forming the facets of the capsid icosahedron (Fig. 1E) as well as individual capsomeres (Fig. 1F) among which hexons (Fig. 1G) and pentons (Fig. 1H) can be distinguished.

Three distinct surfaces were examined: poly-L-lysine-coated glass, poly-L-lysine-coated mica and a highly oriented pyrolytic graphite (HOPG). Mica and HOPG were able to readily adsorb intact HSV-1 capsids (Fig. 2, top, middle), whereas glass predominantly induced capsid disassembly (Fig. 2, bottom), DNA release and adsorption to the surface (Fig. 3).

The mechanical properties of HSV-1 capsids were addressed using AFM as a single particle nano-indentation tool. After locating a suitable region of the specimen by tapping mode AFM imaging (as described in the Materials and Methods) the AFM tip was withdrawn and contact mode was applied for collecting the force-volume (FV) data sets (Heinz and Hoh, 1999). Such a data set is a two-dimensional array of force-versus-distance curves sampled over a surface of several capsids at a predefined loading force. In this case, the `height' image can be considered as an isoforce surface. Borders of capsids were characterized by unstable contact with the indenting tip. Previous studies with polymer microspheres (Tan et al., 2004) indicate that only force-distance curves sampled at the very top of the microspheres can be used for accurate quantification of the material response to the applied loading force. Therefore, we have confined the area of interest to the highest middle point of the capsid surface accessible to interaction with the AFM tip. Capsid spring constant was determined from the data obtained performing nano-mechanical probing over the surface of 40 viral particles. For the HSV-1 capsids, a loading force of 2.2 nN was shown to cause a linear elastic deformation of the capsids with indentations below the value of shell thickness (16 nm) (Saad et al., 1999). Results are summarized in Fig 4. In brief, the stiffness of the genome-filled HSV-1 capsid was found to be 523 pN/nm.

To probe the mechanical limits of HSV-1 capsid incremental loading forces (2.2-8.6 nN) were applied. Having shown that viral particles are capable of withstanding a loading force of 4.4 nN, a higher load of almost 9 nN was applied. Being above the capsid threshold, such force has caused mechanical failure of the capsid. This can be inferred by comparing force-distance curves obtained on the capsids with those obtained on the incompressible surface used for optical lever sensitivity calibration. Characteristic force-distance curves are presented in Fig 5. At the indentation values below viral shell thickness (15 nm), linear elastic response is observed (Fig. 5A). When the force approaches the threshold value of 7 nN, capsids sustain irreversible mechanical damage. This results in a dramatic increase of the indentation (∼80 nm) up to two thirds of the initial height of the capsid.

Alternatively, capsids immobilized on the surface of mica were treated with 1M guanidine hydrochloride solution (GuHCl) to investigate the process of chemically induced DNA release from the standpoint of capsid mechanics. GuHCl was previously shown to induce DNA extrusion out of the viral particles and transform the genome-containing C-capsids into the empty A-capsids (Newcomb and Brown, 1994). A loading force of 4 nN, which was shown to cause linear elastic response in the intact capsids, was applied during imaging of GuHCl-treated capsids. As shown in Fig. 5D and Fig. 6E GuHCl-induced DNA release results in a dramatic decrease in the stiffness of viral particle from 547±19 pN/nm down to 356±17 pN/nm.

Viral capsids carry out a subset of critical tasks including the protection, transport and delivery of the viral genome, and it seems that the mechanical properties of capsids are adapted to these tasks (for a review, see Roos et al., 2007). Capsids must be robust enough to protect the viral genome against physical-chemical assaults but flexible enough to release the infectious genomic material at a certain well-defined point in space and time. Also, many virions accommodate a maximum amount of genetic information in the minimum space, because the nucleic acid is packed to crystal densities (Booy et al., 1991). In bacteriophages, this leads to the internal spatial constraints and high pressurization of the capsid content (Smith et al., 2001), which further increases the stringency of the viral capsid design with respect to its mechanical properties. Unique pairing of high-resolution imaging with nano-mechanical probing under near physiological conditions provided by AFM allows us to directly correlate nano-scale structural alterations with changes in mechanics. With respect to the image resolution obtained, we conclude that AFM can detect structural changes of viral particles at a step of interest during the infection cycle in a host cell, for instance on binding of the capsid to the nuclear pore complex. The latter is just one of many steps crucial to infectivity; however, this step is far from being fully understood from the structural and mechanical point of view. Substitution of the biological surface (nuclear envelope in the case of HSV-1 DNA release) for an abiotic surface with well-defined physical-chemical properties leads to a simplified model in which mechanical and chemical induction of the DNA release can be studied. Unfortunately, glass, which has been shown to be the most suitable support for paired AFM-optical imaging, has proven to induce unspecific and full degradation of HSV-1 capsids (Fig. 3), whereas mica and HOPG enable adsorption of intact capsids. Therefore, subsequent measurements of capsid stiffness value were performed on poly-L-lysine-treated mica.

Stiffness value of HSV-1 capsids appeared to be in close agreement with previously reported values for other viruses (Carrasco et al., 2006; Kol et al., 2006; Ivanovska et al., 2004; Michel et al., 2006). Interestingly, it falls in the stiffness range of the DNA viruses while being higher than that for the RNA viruses. Roos et al. (Roos et al., 2007) hypothesize that the differences in stiffness among RNA and DNA viruses is due to the mode of the assembly of these viral particles. According to this hypothesis, large DNA viruses package their genomic material into preassembled capsids, applying considerable force, whereas the capsids of RNA viruses assemble around the prepackaged genomic material. The measured value of HSV-1 capsid stiffness seems to further support this hypothesis. However, the stiffness of the capsid in question is enormously high compared with the stiffness values of cellular structures known to interact with incoming virions, such as the plasma membrane (Hoh and Schoenenberger, 1994) and the nuclear pores (Schafer et al., 2007), which are below 10 pN/nm (Oberleithner, 2005). Moreover, the fairly high stiffness value of HSV-1 capsids raises the question as to how these structures are `cracked' for the release of the enclosed viral genome upon binding to nuclear pores. Subsequently performed nano-indentation experiments at different loading forces combined with chemical induction of DNA extrusion provide additional information on the mechanical properties of the HSV-1 capsids, as well as possible mechanisms and alterations of these properties during DNA uncoating. While exploring the process of mechanical induction of DNA release, several gradually increasing values of the loading force were applied until a certain threshold value approaching ∼7 nN was reached. At forces below this value, capsids showed linear elastic response to the stimulus. When the threshold was reached (Fig. 5B), capsids suffered mechanical failure. We presume that DNA extrusion occurs at this point. The internal pressure of the HSV-1 capsid is most likely to be very high owing to the packaging of the large viral genome to near-crystalline densities (Booy et al., 1991). Thus, the presence of the large HSV-1 genome inside the capsid will prevent the AFM tip from traveling 80 nm deep into the capsid, unless the DNA is partially or completely expelled. The results of the mechanical probing experiments show that a considerably high force is required to achieve DNA release from the HSV-1 capsid by applying the mechanical stimulus alone. This force, which is almost 6 nN, is currently by far the largest reported threshold for a mechanical breakdown of a virus; previously recorded values barely exceed 1-2 nN (Ivanovska et al., 2007; Ivanovska et al., 2004; Michel et al., 2006). If such a mechanical trigger for the release of the HSV-1 genome were adopted by the virus, it would require considerable `effort' from the host cell to break this tough shell. Therefore, the capsid mechanics are likely to be significantly altered inside the host cell prior to the genome release, possibly as a result of the recently demonstrated proteolysis of one of the tegument proteins VP1/2, which is intimately associated with the capsid and involved in the process of DNA release (Jovasevic et al., 2008; Batterson et al., 1983).

Chemical induction of DNA release and subsequent mechanical probing show that at the loading force of 4 nN, capsids still exhibit linear elastic behavior but the stiffness value is significantly reduced (Fig. 5D). This indicates that DNA does indeed largely contribute to the mechanical properties of the HSV-1 capsids. Intriguingly, neither the mechanically induced nor the GuHCl-induced release of the viral genome was paralleled by a breakdown of the HSV-1 capsid. This was confirmed by applying an initial low imaging force (2.2 nN) to the capsids that had been subjected to high loads and to chemical treatment. Isoforce images before and after applying critical load are shown in Fig. 6A and 6B, respectively. As is evident from the images, the majority of the capsids were able to largely retain their initial shape despite the high loading forces that were applied previously. No striking structural changes were detected in the case of the chemically treated capsids (Fig. 6C,D). At the resolution provided by these images, it is hardly possible to describe exactly the structural background of the dramatic fivefold and twofold increase in stiffness of mechanically and chemically treated capsids. The DNA release seems likely to occur either by means of chemical denaturing or by mechanical alteration of capsid components responsible for the DNA retention inside the capsid.

We show here that HSV-1 capsid stiffness is 523 pN/nm. The force required for changes of the mechanical properties is almost 6 nN. This change is structurally undetectable on the outer surface of the most of the capsids, which leads to the assumption that layers underlying the capsid surface, that is, viral genomic material, have been altered. This notion is further supported by the results of stiffness measurements on the GuHCl-treated empty capsids. We conclude that the genomic material densely packaged inside the HSV-1 capsid significantly contributes to the overall stiffness of the viral particle. We also believe that DNA was extruded upon application of critical mechanical load. The amount of this load seems to be quite high in comparison with currently known strengths of intracellular molecular motor proteins. Therefore, we presume that capsid mechanics undergo significant changes inside the host cell before genome release actually takes place. Finally, the exceptional structural and mechanical stability of the HSV-1 capsid might be the key determinant for its survival over long distances in the axonal cytoplasm where it is exposed to mechanical forces by molecular motors before it reaches the nuclear pore for the crucial genome uncoating.

Viral capsids from mature viruses were prepared as described previously (Ojala et al., 2000). Briefly, capsids were purified from virions collected from infected cell medium by centrifugation and resuspended in MNT buffer [30 mM MES (morpholineethanesulfonic acid), 100 mM NaCl, 20 mM Tris-HCl pH 7.4]. All steps were carried out at 4°C. The virions were stripped of their envelopes and some of the tegument components by incubation in a lysis buffer (500 mM NaCl, 20 mM Tris-HCl pH 7.4, 1% Triton X-100, 1 mM EDTA) for 30 minutes on ice in the presence of protease inhibitors [1 mM phenylmethylsulfonyl fluoride (PMSF) and 13 CLAP cocktail (chymotrypsin, leupeptin, aprotinin and pepstatin; 10 mg/ml each)]. After lysis, the sample was sonicated in a water bath (three times, 30 seconds each), layered onto a linear 20-45% sucrose gradient (in MNT supplemented with 400 mM NaCl, 1 mM EDTA and 0.5 mM dithiothreitol), and centrifuged in a Beckman SW50.1 rotor at 4000 g for 25 minutes. Capsids were collected as a light-scattering zone from the gradient and resuspended in nuclear isolation medium (NIM) (Shahin et al., 2001) and pelleted by centrifugation for 60 minutes at 4000 g in a Beckman SW28 rotor. The resulting capsid pellet was resuspended in NIM.

For the surface compatibility studies, surfaces of the materials chosen were prepared as follows. Freshly cleaved mica discs (11 mm diameter; Agar Scientific, UK) which were fixed by epoxy (Aron Alpha type 102, Agar Scientific) to 15 mm steel SPM specimen discs (Agar Scientific) and glass bottom Petri dishes (WillCo Wells BV, Amsterdam, The Netherlands) were incubated with 0.01% poly-L-lysine (Sigma) solution for 15 minutes. Following a brief rinse in deionised water and air drying of the surface, the HSV-1 capsid suspension was applied. After a 60 minute incubation at room temperature, samples were repeatedly rinsed with NIM to remove unadsorbed capsids. Similarly, capsid suspension was applied on freshly cleaved untreated surface of HOPG.

AFM imaging was carried out at room temperature using a Digital Instruments Multimode atomic force microscope equipped with a J-scanner and a Nanoscope V controller with an in-line electronics extender module (Veeco/Digital Instruments, Santa Barbara, CA) or a Bioscope II atomic force microscope (Veeco/Digital Instruments) for combination of AFM and fluorescent imaging. All topographical images were collected applying tapping mode in fluid, and using oxide sharpened silicon nitride probes (DNP-S; Veeco/Digital Instruments) with a spring constant of 0.26 N/m as determined using the thermal noise method (Hutter and Bechhoefer, 1993). Data analysis was performed with commercially available software (NanoScope 7.2 software, Digital Instruments; Scanning Probe Image Processor, Image Metrology, Lyngby, Denmark). Fluorescent labeling was performed by imaging in medium containing 0.1 μM YO-PRO 1 DNA intercalating dye (Invitrogen, Carlsbad, CA).

Force volume data was acquired in contact mode. At least two separate experiments for each treatment were performed and four force-distance curves on 10 particles in each experiment were taken for statistical analysis. Resulting isoforce topography images were processed by Scanning Probe Image Processor. Force data analysis was performed combining routine built-in NanoScope 7.2 software, Protein Unfolding and Nano-Indentation Software (Carl et al., 2001) and MS Excel.

This study was supported by grants from the Innovative Medizinische Forschung (SH-110315, SH-520404 and SH-120613) and the Deutsche Forschungsgemeinschaft (Graduate School Molecular Basis of Dynamic Cellular Processes) of the SFB629, International graduate school Interaction of pathogens with biotic and abiotic surfaces GRK1409 and OB 63/16-1 and Interdisciplinary Center of Clinical Research (IZKF) Münster, project no. Küh3/064/04. We thank Nelson Barrera from the Department of Chemistry, University of Cambridge, UK, for statistical analysis.