banner virusanimation

Below you will find a summary of our work on cell mechanics, virus mechanics and some technique development. More information can be found in the publications.

1) Mechanics of the cell

1.1) Basics
In our lab we have various tools to measure the mechanical properties of single cells, including AFM and optical trapping to indent cells but also to extract membrane tethers from the cells via pulling experiments. Such cell mechanical measurements are widely used as an indicator for differentiation or the response to drug treatment.
Because of the inhomogeneous organization of the cell, its mechanical response is expected to depend strongly on the measurement technique. To investigate this, we compared different techniques based on AFM and optical trapping (figure 1a) over a wide range of deformation time- and length-scales. Figure 1b shows that both the indentation rate or depth have a large effect on the measured response. These results reflect the heterogeneous structure of the cell and clearly show that a cell cannot be simply characterized by a single 'Young's modulus'. Instead, information can be obtained about the viscous nature of the cellular structures and by varying the measurement parameters, the experiments can even be tuned to be sensitive to different structural parts of the cell.

a b

Figure 1: Measuring cell mechanics at different length- and time-scales.
a) Cartoon of cell deformation experiments with the vertical optical trap (left) and AFM (right). With the optical trap we can exert much lower forces and deformations than with AFM.
b)The cell stiffness increases when the cell is indented at higher speeds, which indicates a viscous component. The contribution of this viscous component increases at larger deformations (black points). Graph from Nawaz et al. 2012.

1.2) Cell differentiation
The actin filament turnover at the leading edge of the myelin sheaths provides the driving force for myelination. Tether pulling experiments showed an increased membrane tension at the leading edges (figure 2). Stabilization of the sheaths is performed by the myelin basic protein (MBP). In figure 3 we identified a mechanical change in a mutated protein that turned out to inhibit sheath formation in affected cells.

Figure 2: Membrane tension depends on the local presence of an actin cortex.
To measure surface tension in oligodendrocytes, we used AFM to pull membrane tubes (tethers) in a vertical direction at different positions of the cell. The measured tether force is shown in relation to the distance from the cell body. The force map shows oligodendrocytes cultured for 5 days. The tether force was anisotropic in sheath forming cells with higher values in the outer cellular rim (the leading edge). The tether force is shown as mean ± SEM (n = 9–12 cells, with a total of 291–800 pulled tethers for each stage). Figure from Nawaz et al. 2015.

Figure 3: A single point mutation inhibits protein-protein interactions.
Comparison of interaction forces between wild-type myelin basic proteins (MBP) and the F→S mutant. The histogram shows the interaction forces between wild-type MBP (black), mutant-MBP (red), and without proteins (green). The wild-type proteins show the largest interaction forces. For the experiments the proteins were adsorbed to both the mica surface and AFM tip. Inset shows the schematic depiction of shape of the force-distance curve as AFM tip approaches the sample surface (1), as tip touches the surface (2), and as tip is retracted from the sample surface (3). Figure from Aggarwal et al. 2013. 

2) Mechanics of viruses

2.1) Influenza virus

Enveloped viruses consist of a protein shell that is enclosed by a lipid bilayer (figure 4). From previous deformation experiments on a variety of viruses the picture emerged that viruses pack their genome in stiff symmetric protein shells (k~0.3 N/m). Such a rigid design helps to withstand the high internal pressure that occurs in some viral strains due the densely packed genome, and is speculated to enhance the virus survival when it travels from host to host.
To measure the stiffness, we use AFM to make 'force maps' of single viruses (an array of force curves). From such a force map a height image can be reconstructed (figure 5a) and the stiffness for each pixel can be determined (figure 5b).
One of our surprising findings was that the influenza virus is about a 10 fold softer than all other viruses studied so far (Science magazine news, Schaap et al. 2012). Despite (or thanks to) this 'softness' it can sustain deformations up to almost 100% of its diameter, making the influenza virus as hard to rupture as a stiff protein shell. In collaboration with Andreas Herrmann (Humboldt University, Berlin), we have characterized the mechanics of the isolated viral envelope and found that i) the envelope is in the liquid state at room temperature, ii) it has a stiffness close to that of the intact virus, and iii) it does self-repair after rupture (Li et al. 2011).
As a next step we focused on the influenza M1 matrix. From electron microscopy data it has been shown that M1 coats the inside of the lipid envelope with a semi-continuous protein layer. Although we found that this layer has only a small role in reinforcing the virus (figure 6) it turned out have an essential role in regulating the unpacking of the virus and the efficient release of its genome (Li et al. 2014).
After the virus is taken up by the target cell via the endosomal route, lowering of the pH leads to the stripping-off of the M1 protein layer from the lipid bilayer. This step is likely essential for the release of the viral RNA before the viral membrane fuses with the endosome.

Figure 4: Cartoon of the influenza virus structure.

A lipid bilayer containing the spike proteins is coated from the inside by a layer of M1 proteins. 

a b
Figure 5: Measuring the virus' stiffness from AFM force maps.
a) Tapping mode image (left) and the reconstructed height image (right) from a force map obtained on the same particle. Force maps are obtained by collecting force vs. distance curves on an array of 24 x 24 points, covering an area of 300 x 300 nm
b) From the force maps a spatial distribution of the stiffness can be plotted. The stiffness of the particle is highest in its centre.

Figure 6: Two-step softening of the influenza virus.
The stiffness of the viruses decreased with pH (black line and gray arrows). After neutralizing the buffer from pH 6.0 to pH 7.4, the stiffness recovers (green line and arrow). However, after neutralizing the buffer from pH 5.0 or pH 5.5 to pH 7.4, the stiffness does not recover (red line and arrow). This indicates two different processes that are responsible for the softening.

2.2) Uncoating of the adenovirus
Collaboration with Pedro de Pablo
(Universidad Autónoma de Madrid) and Urs Greber (University of Zurich).
For this project we combined AFM with single molecule fluorescence microscopy. By using fluorescent labels, we can identify specific parts of the sample. Since the AFM tip and cantilever will cause lots of light scattering in a conventional fluorescence design we employed a total internal reflection fluorescence (TIRF) approach. With TIRF only the 100 nm close the coverslip are exited, such that the AFM cantilever and most of the tip are outside of the excitation field and do not contribute to the background fluorescence. Although such combined instruments are commercially available, none of them offer the possibility to operate both techniques simultaneously without reducing the quality of the AFM scans. We rigorously re-engineered the optical part of the microscope and succeeded in combining single molecule fluorescence with nm-accuracy AFM imaging, at the same time. In experiments with adenovirus, we opened up single capsids with the tip of the AFM and observed the expansion of the released genome with fluorescence. This allowed us to identify different expansion states that we speculate to be responsible for the lack of infectivity of a tested mutant strain of the virus (Ortega-Esteban et al. 2015).

Figure 7: AFM unpacking of adenovirus
A cartoon of the experiment shows the unpacking experiment. By using Yoyo-1 as fluorescent label we can track the release of the viral DNA. The full animation can be found here.

2.3) Lipid bilayer mechanics
Lysosomes, enveloped viruses, synaptic and secretory vesicles are all examples of natural nano-containers (diameter ~100 nm) which specifically rely on their lipid bilayer to protect and exchange their contents with the cell. We have developed methods primarily based on AFM and finite element analysis that allow precise characterization of the mechanical properties of the lipid bilayer.
First small spherical vesicles are formed from a lipid mixture of interest. These liposomes are then one by one indented by an AFM tip using forces up to 300 pN, which leads to an elastic deformation of about 10-20 nm. From the measured elastic response and a mechanical model we can estimate the bending rigidity of the membrane (Li et al. 2011).

Figure 8: The stiffness of individual liposomes with different diameters made out of viral lipids.
The stiffness scales with the diameter, which is expected from modelling (continuous lines). The liposomes get softer at increasing temperature, caused by a gradual increase of the lipid disordering. Graph from Li et al. 2011.


2.4) Mechanical modelling of protein shells
In contrast to the influenza virus most other viruses protect their genome with a stiff polyhedral protein capsid. Despite this minimalist design, such viruses often show a counter-intuitive response to mechanical stress. We develop mechanical models based on finite element methods to explain the response of these viruses and to learn more about their architecture. The advantage of finite element analysis over analytical methods is that more realistic experimental boundary conditions can be included. Thus we can predict the effects of the exact tip size and shape on the measured mechanical response (figure 9, Schaap et al. 2006).

Figure 9: Tips size and wall thickness affect the probed mechanical response.
a) When the radius of the indenting AFM tip increases, the reported stiffness will also increase.
b) Buckling of a structure (visible as a sudden softening) depends on the ratio between wall thickness and object radius, and the size of the indenting probe.

A large part of this work is performed in collaboration with Pedro de Pablo (Universidad Autónoma de Madrid), whose group is performing nano-indentation experiments on non-enveloped virus capsids. With our modelling we confirmed how a parvo-virus uses small parts of it DNA genome to specifically reinforce the protein shell to increase its global stiffness by a factor of two (figure 10).

Figure 10: Mechanics of a DNA reinforced virus capsid.
When an icosahedron is reinforced with patches of DNA (pink circles) this leads to an increase in stiffness. The reinforcement is highest when the virus is probed along the two-fold symmetry axis (blue) and lowest along the five-fold symmetry axis (red). Figure from Carrasco et al. 2006.

3) Development of biophysical methods

31) Vertical optical trap
The lowest force that can be accurately controlled with AFM is limited to ~30 pN. Although this does not seem much, it is actually enough the rupture protein bonds or to deform cells for hundreds of nm. To be able to perform mechanical measurements at lower forces we set out to explore the possibilities of optical trapping. Because we wanted to keep the contact mechanics similar to those in AFM experiments we developed a vertical optical trap (figure 11). The trapped bead can be moved up and down with respect to the microscope coverslip. The trapped bead can thus be used to indent samples, in a similar fashion as the AFM tip. Since a focused laser beam is used to trap the bead we had to minimize the influence of optical artefacts like the interferometric effect between the bead and the coverslip and a shift of focus. The full technical details of this instrument and its performance is described in Bodensiek et al. 2013. With a fast responding force feedback loop we can achieve deformation rates as high as 50 μm/s, which allow the investigation of the elastic and viscous components of cells. Figure 12 shows force vs. deformation curves that were obtained on single cells, the noise level is less than 1 pN.

Figure 11: Optical design of the vertical optical trap.
The trap is formed by the laser light (shown in red) coming from a single mode fibre that is coupled into the imaging path (blue) of an upright optical microscope. The vertical position of the trap is controlled by a z-piezo that moves the objective up and down. To monitor the displacement of the bead from the centre of the trap, the laser light is collected by the condenser and cast onto a photodiode. Image from Bodensiek et al. 2013.

Figure 12: Typical optical trap indentation curves on fibroblast cells.
The cells were indented between 100 and 250 nm at a force of 10 pN. The large variation between the curves is inherent to measurements on cells.
Graph from Bodensiek et al. 2013.

4) AFM imaging of enzyme dynamics

With AFM we can image our samples in liquids and are thus not limited to static snapshots. By recording sequential images we can actually 'see' single enzymes in action. Previously we applied this strategy to observe the motion of single kinesin motor proteins along microtubules (figure 13). We found that kinesin moved in a straight line along a single protofilament of the microtubule in multiples of 8 nm.

Figure 13: Kinesin moving along a microtubule.
Two frames that show a single kinesin motor making a step along the microtubule. Both 'heads' of the kinesin motor, just 8 nm apart, can be clearly distinguished. Movie from Schaap et al. 2011.