Research Strategy: The overarching goal of this proposal is to take a fundamentally different approach towards

studying viral infection at the cellular level. Instead of using live cell imaging methods to investigate viral dynamics
with low temporal precision, we will develop and implement virus-locked imaging methods to continuously
interrogate the internal and external dynamics of invading viral particles in living cells. This work will encompass
four major research directions. In Research Direction 1, a new method for locking onto diffusing viral particles
with high precision and sensitivity will be developed. Research Direction 2 will add local 3D imaging to the
target-locked virus to monitor its surroundings, allowing determination of the exact moment of viral landing as
well as the long-range interactions between viral particles and the cell surface. Research Direction 3 will
advance this imaging technique towards submillisecond 3D imaging around the locked viral particle, enabling
observation of membrane dynamics during viral internalization and transient ligand-receptor interactions.
Research Direction 4 will delve into the internal dynamics of the viral envelope during viral landing, receptor
binding, internalization and intracellular trafficking.
A. Research Direction 1: Virus-locked interrogation. The goal of this research direction is to develop and
implement a method to continuously monitor a diffusing viral particle in the live cell setting. This will be
accomplished by effectively locking the virus in the focal volume of a fluorescence microscope utilizing force-
free real-time 3D single particle tracking.
A.1. Significance. Phenomena which happen at high speeds in three dimensions are present in nearly all normal
biological function as well as disease states. These include the transient interactions of viral particles with a
dynamic and three-dimensional cell membrane, the interaction of cellular cargoes with the nuclear pore complex,
or simply the interaction of nanoscale objects with macromolecular complexes in solution. These processes
challenge current live cell imaging methods due to the involvement of diffusion of nanoscale complexes, which
occur on timescales much shorter than the exposure time of a typical CCD. As a result, phenomena which occur
on short time scales are blurred or overlooked. Imaging methods have undergone a revolution in both spatial
and temporal resolution. Methods such as PALM (1, 2), STORM (3), STED(4) and (S)SIM(5) have broken far
beyond the diffraction limit in all three dimensions. However, each of these methods requires a trade-off of
acquisition rate for spatial resolution. For example, in PALM and STORM, high precision localizations are made
by imaging small subsets of molecules in a field of view repeatedly until the entire image is filled in. This leads
to an enormous information overhead which slows acquisition. Other methods have shown the ability to rapidly
evaluate 3D scenes in tissue and live cells. These methods (DLSM (6), SPIM (7), SCAPE(8)) typically employ a
planar or quasi-planar illumination which can simultaneously illuminate an optical section for collection by a CCD.
However, in these methods, there is still a requirement to stack these optical sections to get a complete picture.
This again leads to an overhead which limits the overall temporal resolution.
A.2. Innovation. The critical limiting factor in these methods is the trade-off between spatial scale and imaging
rate. As an example, consider a scanning imaging method such as confocal laser scanning microscopy. If the
process is confined to just a single pixel, the temporal resolution becomes unlimited (such as in single molecule
spectroscopy (9, 10)), limited only by how many photons are available. Unfortunately, a molecule in a live cell,
as opposed to a contrived in vitro experiment where a molecule is tethered to a surface, is free to move in three
dimensions, rapidly diffusing out of the detection volume of a single pixel. One method which can translate the
photon limited temporal resolution of single molecule spectroscopy is real-time 3D single particle tracking. Here
we develop a new method of 3D single particle tracking which is optimized towards following objects at high
diffusive speeds and low photon count rates.
A.3. Approach: Real-time 3D single particle tracking. Typically, real-time 3D single particle tracking uses
optical feedback (fluorescence or scattering) to lock-on to a diffusing particle. The method has been applied in a
variety of configurations by Mabuchi (11), Werner (12-15), Gratton (16), Lamb (17) and Yang (18-22). Real-time
3D single particle tracking mechanisms can be broken into two subsets. One way is to have a set of detectors
or optical fibers which spatially separate the emission from the particle. A second common method is the lock-in
method, which uses a rotating laser spot to illuminate the particle. The emission from the particle is demodulated
through a lock-in amplifier. When the particle is centered, the amplitude from the lock-in will read zero. As the
particle moves away from the center of the observation volume, the frequency of the circular illumination pattern
will be encoded in the particle’s intensity time trace, yielding an actionable error readout. Unfortunately, both
methods have drawbacks which limit their ability to track fast moving and lowly emitting particles. In the detector
based method, the few collected photons must be split onto multiple detectors at very short bin times, leading to
many empty bins and tracking instability. The lock-in method is limited both by the time constant of the low pass
filter in the amplifier as well as Nyquist sampling, which requires that the photon collection rate must be larger
than the rate of the circular laser spot. To address these shortcomings, we propose a new method called 3D
dynamic photon localization tracking (3D-DyPLoT).
A.3.i Dynamically moving laser spot for high speed 3D tracking.
The dramatic difference in this new tracking scheme is a dynamically
moving laser spot (Fig. 1) inspired by the anti-Brownian
Electrokinetic (ABEL) molecular trap (23-26). Instead of using
separate detectors, we employ a moving laser spot which tags
incoming photons with location information. First, the laser beam is
deflected in the XY plane using a 2D electro-optic deflector in a 5x5
Knight’s Tour over a 1 µm x 1 µm grid in the XY plane, with a 20
µsec dwell time at each spot. The Knight’s Tour pattern provides
uniform illumination without the lag inherent in a raster scan pattern.
At each point in the scan, the number of photons collected on an
avalanche photodiode (APD) are counted and used to feed a
Kalman update filter which predicts the particle position and
Figure 1. Schematic of 3D-DyPLoT
associated uncertainty.
xˆk |k −1w2 + ck nkσˆ k2|k −1 σˆ k2|k −1w2
xˆk |k = σˆ k2|k =
w2 + nkσˆ k2|k −1 w2 + nkσˆ k2|k −1
The position at each point ( xˆk |k ) is determined using the previous particle position ( xˆk |k −1 ), laser spot covariance
( w2 ), current laser position ( ck ), number of photons collected ( nk ) and uncertainty of the previous position
measurement ( σˆ k2|k −1 ). To achieve axial deflection of the laser spot, a tunable acoustic gradient (TAG, Tag Optics)
lens is employed (27). This lens deflects the laser focal spot dynamically in a sine wave pattern at certain
resonance frequencies, the lowest of which (70 kHz) is applied to tag photon arrival times with axial positions.
Combined, these XYZ position estimates give a real-time estimate of the position of the particle relative to the
center of the scan range, which is 1 µm x 1µm x 2 µm. To restrict the particle to the focal volume of the objective
and track its position over large distances, these position estimates are fed into an integral controller which
controls the position of a 3D piezoelectric stage to lock the
position of the particle to the center of the scan range.
A.3.ii Preliminary results from 3D-DyPLoT. This
implementation has many advantages over other real-time
tracking methods resulting from the use of just one detector and
the large effective detection volume, which means the tracking
algorithm is able to recover from large random jumps in the
particle’s position during trajectories. As a result, preliminary
experiments have shown that 3D-DyPLoT is capable of tracking
single silica-coated non-blinking quantum dots (28) (~30 nm, D
~ 11 µm2/sec), with intensity ranging from 10-30 kHz (Fig. 2).
The ability to track at high speeds and low photon counts rates
has further allowed 3D-DyPLoT to move beyond artificial
constructs such as fluorescent beads and quantum dots. Figure
3 shows the 3D trajectory of a pVenus-VSV-G pseudotyped
lentivirus particle tracked by the 3D-DyPLoT microscope. This
illustrates the ability to track biologically relevant samples.
Using a system such as 3D-DyPLoT allows continuous
observation over long length and time scales for fast moving
objects with sampling of the position every 10 µsec. Thus, any Figure 2. 3D-DyPLoT applied to 30 nm silica coated
spectroscopic method that can be applied to a particle fixed on quantum dots in water
a coverslip can be applied to the locked probe. In other words,
the temporal resolution becomes unlimited. In addition to this continuous spectroscopic sampling, the localization
precision for slow moving particles is below 10 nm in X and Y and 15 nm in Z depending on the choice of
feedback parameters at count rates near 80 kHz.
A.3.iii Local virus-locked imaging. To correlate the viral dynamics captured by the 3D-DyPLoT method local
cellular information must be acquired to realize a complete picture of the viral lifecycle. As mentioned above,
imaging an entire 3D cellular structure places an enormous overhead burden on live cell imaging systems. To
image the entire process of a virus landing on a cell surface, interacting with receptors and eventually
internalizing would require observation over 10s of microns in three dimensions with spatial and temporal
precision of the viral particles motions in line with its typical diffusive behavior (10s of nanometers,
submillisecond). Here, the overhead associated with imaging the 3D cellular structure is discarded in favor of
high spatiotemporal measurement of the viral dynamics using 3D-DyPLoT. To recoup this lost observation of the
contextual environment, 3D-DyPLoT will be combined with two-photon laser scanning microscopy in a method
called 3D Multi-resolution Microscopy (3D-MM) (18, 29). 3D-MM works by using a two-photon laser scanning
module (2P-LSM) to generate two-photon fluorescence from the area around the target-locked probe (Fig. 4a).
As the particle moves, the second laser scans the X-Y plane containing the particle. As the particle moves
between focal planes, the imaging laser moves with it. Over time, these data can be integrated to generate
complete 3D picture of fast processes, as shown in Fig. 4b for a 100 nm TAT peptide modified fluorescent
nanoparticle transiently landing on the cell surface.
A.4. Potential Obstacles. In this implementation of 3D-MM, the
3D cell structure is only acquired as the probe travels to different
optical sections. As a result, there may be limited information on
what is directly above or below the probe at any given moment
during the measurement. To acquire full 3D images without
needing to integrate over the entire 3D viral trajectory, we will
expand the imaging planes of 3D-MM into cubes which image
rapidly around the particle using an electrically tunable lens
(ETL). The ETL will deflect the pulsed imaging laser into different
focal planes using a sine wave pattern near its resonant
frequency (~100 Hz). This will yield imaging voxels at z positions
which are different from that of the particle, giving a full 3D Figure 3. 3D-DyPLoT applied to pVenus-VSG-G
lentivirus
picture.
B. Research Direction 2: Interaction of viral particles with the cell surface
The goal of this aim is to understand the long and short range interactions of freely diffusing viruses with
the extracellular matrix and cell membrane. This aim seeks to understand the dynamics which underlie the first
stage of viral infection, which is the initial contact with the tissue to be infected from the extracellular space, via
the 3D-MM method described above.
B.1 Significance. Viruses are nanoscale particles which hijack native cellular machinery in order to replicate
their genetic material and propagate (30). At the cellular and single particle level, viral infection has been well-
studied by methods such as fluorescence microscopy (31), single-particle tracking (17, 32) and electron
microscopy (33). However, these studies address the behavior
of viral particles only after the first stage of infection (i.e. the
virus’s transition from the extracellular matrix to the cell surface)
is already complete.
Viruses have many different routes for infection at the
tissue level. Prior to receptor binding events, these viral
particles utilize Brownian motion to perform a random walk
search to find and bind to the cell surface. This process is
impossible to capture with conventional methods, due to the
high speed of the viral random walk (2-10 µm2/sec) and the
large length scales over which the virus interacts with and lands
on the cell surface. In the epithelium, where much of viral
infection occurs, the situation is even more complex. A particle
has to navigate a crowded extracellular environment containing
a complex mesh of polysaccharides, such as mucins (34-36),
and proteins, such as collagens (37). The particle then must
navigate towards the cell, which exists in a tightly packed 3D Figure 4. (a) 3D-MM tracks the high speed behavior
of nanoscale probes, using 2P-LSM to generate a 3D
arrangement with other epithelial cells. These extracellular landscape of the cell environment. (b) Transient
components and multicellular organization undoubtedly play a landing of 100 nm particle on the cell surface as
role in viral infection. measured by 3D-MM. (c) 3D structure of a
B.2 Innovation. The newly developed 3D-DyPLoT combined macropinosome traced out by an internalized
with 3D-MM as described in Research Direction 1 will be nanoparticle.
applied to tackle these poorly understood interaction dynamics. By locking onto a freely diffusing viral particle
and following it as it approaches and lands on the cell surface, high precision dynamics can be extracted from
the viral scale particle. At the same time, the 2P-LSM module yields the cellular contour and shows areas of
close approach between the particle and the cell. Using both the particle dynamics and approach information, a
correlation between diffusive behavior and distance from the cell surface can be observed. In previous studies
on Tat-functionalized nanoparticles (18) this correlation showed longer range interactions than predicted by
simple physical models (38, 39). This suggests that the extracellular material, as well as the biochemical nature
of both the particle and cell surface, affects the diffusive behavior which precedes membrane attachment. These
data provide a characteristic length scale for the interaction of a diffusing nanoscale particle with the cell surface.
B.3 Approach: Characteristic interaction length and frequency of attachment of viral particles. These
long-range interactions will be probed in two ways. First, the effect of receptor density on these transient
interactions between viruses and the cell membrane will be probed by introducing fluorescent virus particles into
live cell cultures. Initial experiments will be performed using the pVenus VSV-G pseudotyped lentivirus shown in
Figure 3 above and NIH-3T3 fibroblast cells. The VSV-G VLPs exhibit wide tropism and will provide a baseline
for the general interaction of viral particles with the cell membrane. Once these data are acquired, the effect
specific receptors on the viral landing dynamics will be probed. Here, we will replace the pVenus VSG-G with a
pVenus Env vector to mimic HIV-1. NIH-3T3 will then be transformed to express the HIV-1 cellular receptor
(CD4)(40) or critical chemokines (CXCR4 and/or CCR5)(41, 42), and then will be allowed to interact with the
fluorescent VLPs. Characteristic length scale of interaction as well as frequency and location of landing will be
recorded for varying levels of cell surface receptors. This will give direct insight into the role of the receptor level
on the early stage dynamics of viral infection.
The unique capabilities of the 3D-MM technique will further enable the study of the effect of the
extracellular environment on viral infection. The extracellular matrix may play a significant role in the chances of
viral infection, especially in the complex epithelia which is so often the target for viral infection. To test the effect
of extracellular matrix elements on infection dynamics, human lung adenocarcinoma cells (A549) which express
large quantities of mucins will be studied to mimic the bronchial epithelia, an infection site for many viruses(43,
44). These dynamics will be compared with mucin knock-outs to see how this mucous layer either aids or deters
viral infection, providing a direct glimpse into the dynamics of a virus before it interacts with the cell surface in a
complex environment.
B.4 Potential Obstacles. A potential hurdle in these studies is the difficulty in resolving the position of the viral
particle relative to the cell membrane. While the particle being tracking via 3D-MM yields high precision
localization (down to 10-40 nm XYZ), the image of the cell membrane generated by 2P-LSM will be limited by
diffraction (~300 nm lateral, ~800 nm axial). This mismatch may yield unworkably large experimental
uncertainties which preclude conclusions to be drawn. This can be tackled by replacing the 2P-LSM module with
a scanning STED module (4, 45-47). By combining 3D-MM with the power of STED, the resolution mismatch
can be eliminated, leading to more conclusive data.
B.5 Conclusions and Outlook. These studies aim to provide characteristic interaction lengths for a variety of
virus-to-cell interactions. The results of these experiments will be telling in which factors dictate the long range
interactions between viral particles and cells. This work will also evaluate particles which are likely to come close
or transiently interact with the cell surface without long term binding. These two factors (long range attraction
and long term binding) are not necessarily governed by the same causes. It may be that certain physiochemical
properties dictate whether a particle will come close to the cell surface, while more specific interactions (e.g.
ligand-receptor) are required for the particle to bind to the surface. Determining these factors would open the
possibility that these specific interactions could be blocked, paving the way to a new class of antivirals.
C. Research Direction. Beyond video-rate local imaging of the viral environment.
This research direction aims to provide sub-millisecond 3D imaging local to a viral particle local to critical
infection points, namely the association of the viral particle with critical receptors or chemokines and the
dynamics of the cell membrane during internalization of the viral particle.
C.1 Significance. The critical aspects of viral infection are dynamic in nature. Upon initial binding of the virion
to the cell membrane, there are multiple necessary interactions with receptors and chemokines which lead to a
chemical signaling which allow for the virus to internalize (30). Virus internalization requires significant
restructuring of the cell membrane, which occurs rapidly at small (nm) spatial scales in 3D. A full spatiotemporal
investigation of this internalization phenomena has eluded live cell imaging methods due to the transient
interactions between viral particles and cell receptors and the subtle and fast 3D motions of the cell membrane.
C.2 Innovation. The 3D-MM technique, while powerful, relies on integrating the 3D cell image (generated by the
2P-LSM module over the entire course of the particle’s motion. Figure 5 shows a 3D-MM measurement of the
internalization of a 100 nm fluorescent nanoparticle. As a result of this integration, no temporal information about
the cell is acquired. With 3D-DyPLoT, not only does the dynamically moving laser spot allow the potential to
follow a nanoscale probe wherever it goes, but it allows for the potential to image the 3D volume directly around
the probe at extremely high speeds to sample its environment. This can be achieved by adding a second laser
through the 2D-EOD and TAG lens which excites a feature of interest in the probes surroundings and a second
detector dedicated to collecting the emission from the surroundings. Due to the high speed of the scanned laser
spot, the 3D volume around the focal spot is sampled with high frequency. While small, the rapid local imaging
will capture interactions between the locked probe and other macromolecules and larger cellular structures.
C.3 Approach: Beyond video-rate local 3D imaging. This
rapid 3D imaging is dramatically different from other scanning
methods. First, there are no mechanical parts involved in
scanning, resulting in very high speed. Second, it is not a
traditional raster scan. The Knight’s Tour pattern is required
for high speed tracking of the locked probe. Further, the axial
scanning is asynchronous with the XY scan, leading to
Time uneven sampling of the imaging volume. To account for this,
the 2D Knight’s Tour of the EOD will be synchronized to the
Figure 5. (a) Edge on view of trajectory of 100 nm
period of the TAG lens. The TAG lens driver outputs a
nanoparticle diffusing on membrane and internalizing. synchronization pulse at the beginning of each cycle, which
(b) Side view of trajectory showing cell contour for GFP can be used to drive the EOD to the next step. At the 70 kHz
expressing NIH-3T3 cells. TAG lens resonance, this would result in a XY dwell time of
14.3 µsec. During this time, a full axial scan of the area is
made. As photons are collected on the imaging APD, they are stamped with XYZ positions based on the location
in the TAG lens period and the current EOD position. This leads to an even sampling in time.
This rapid 3D imaging volume can be tuned by adjusting the parameters of the EOD and TAG lens. In
the 3D-DyPLoT application above, the XY range is 1 µm x 1 µm. This can be expanded up to 4 µm x 4 µm by
increasing the voltage supplied to the EOD during scanning. The resolution of the image in the lateral plane is
dictated by the number of spots in the Knight’s Tour. For 3D-DyPLOT, a 5x5 grid is chosen to scan as quickly
as possible. The overall 3D imaging speed can be determined from the TAG lens frequency as each spot in the
Knight’s Tour is held for the entire period of the TAG lens. Therefore, the overall imaging frequency is simply the
number of points in the XY plane divided by the TAG lens frequency. For 5x5, the 3D imaging rate around the
particle is 357 µsec/volume. For 11x11, this increases to 1.7 msec/volume, while a 25x25 grid (which would give
40 nm XY pixels over a 1 µm x 1 µm grid) yields a rate of 8.9 msec/volume. These fast imaging rates will enable
visualization of previously impossible to capture dynamics.
C.3.i Imaging membrane dynamics during internalization A full 3D real-time view of particle internalization
is still the realm of modeling. This rapid local imaging can address these types of interactions, such as those
between a nanoscale particle or virus interacting with the cell membrane. The simultaneously imaging of a rapidly
diffusing viral particle and the fluctuations of a 3D membrane are lost in existing 3D imaging methods. Here we
will apply the 3D rapid local imaging described above to image dynamics of the cell membrane at the precise
moment of internalization. For these experiments, the pVenus VSV-G lentivirus VLPs will be tracked and the
membrane of target cells will be labelled with a red emitting lipophilic dye, such as 1,1′-dioctadecyl-3,3,3′,3′-
tetramethylindotricarbocyanine (DiR). Using two rapidly scanned beams (one at 488 nm to excite the virus, the
other at 640 to excite the membrane) and a dedicated APD for each spectral channel, a full 3D beyond-video-
rate observation of the internalization of a single virus particle will be made.
C.3.ii Imaging receptor dynamics and correlating with function Ligand-receptor interactions play a large role
in viral infection, where virus particles use a variety of cell surface receptors to gain entry to cells (48). For
internalization of viral particles to occur, it is generally required that the viral particle interact with a cell surface
receptor (e.g. EGF-EGFR for hepatitis B (49), Epstein-Barr (50) and pox viruses (51)). Here we will apply this
rapid volumetric imaging to monitor the interaction of pVenus VSV-G VLPs with RFP fusion of their cognate
receptor, LDLR (52). Taken with the above membrane imaging, these will provide a complete 3D spatiotemporal
map of viral particles on the cell membrane.
C.4 Potential Obstacles. A potential technical hurdle is the uneven excitation spot dwell time along the axial
direction. Since the TAG lens moves in a sine wave, the laser spot spends more time at the top and bottom focal
planes and less time in the middle (Fig. 6). To compensate, the laser power can be modulated using the absolute
value of the TAG lens signal, shifted 90 degrees out of phase. This will lead to uniform excitation throughout the
rapidly sampled volume.
C.5 Conclusion and Outlook. The fast local imaging described above will successfully combine continuous
observation of a target-locked virus with beyond video-rate 3D imaging of its environment. This highly local
and highly rapid probing of fast processes will provide new insights into many new areas in live cell studies.
D. Research Direction 4: Correlating viral envelope dynamics with the viral lifecycle. This research
direction will combine the methods described above to correlate the internal functions of the virus with stages in
the infectious cycle, including the first moment of attachment to the cell membrane, internalization, association
with receptors and intracellular trafficking.
D.1. Significance and Innovation. Understanding the viral lifecycle ultimately requires knowledge of how the
virus responds to different external stimuli at different infection stages. For instance, upon initial binding to the
cell membrane, how do the proteins responsible for envelope fusion respond? Does this response require the
presence of a receptor? Does it occur at the cell surface or within cellular
compartments? For some viruses (such as VSV-G), envelope fusion is
known to be pH dependent and triggered within low pH late endosomes
and lysosomes (30). For these systems, fusion of the viral capsid can be
studied by binding the viruses to an artificial membrane and changing the
pH (53). For others, such as HIV-1, the envelope fusion is pH-independent.
As a result, the envelope fusion cannot be triggered with a fixed viral
particle. Here, using virus-locked imaging, we will be able to continuously
monitor the viral envelope to correlate the viral envelope fusion mechanism
with the real-time biochemical environment in live cells. By locking on to the
Figure 6. Achieving uniform sampling.
virus as it invades the cellular environment we will continuously monitor the
internal dynamics of the virus particle as it negotiates different cellular
environments.
D.2 Approach. Here we will address the internal dynamics of invading viral particles by monitoring fusion of the
viral envelope. To do so, VLPs will be generated with a Vpr-YFP fusion (54) which will
enable the 3D-DyPLoT to lock on to the virus particle. Next, a lipophilic dye such as 1,1'-
dioctadecyl-3,3,3',3'-tetramethylindodicarbocyanine (DiD) will be used to treat the
envelope. These dyes will intercalate into the viral envelope and become fluorescent. As
the concentration of DiD in the viral envelope increases, the fluorescence intensity from
each viral particle increases linearly with dye molecule concentration. However, with
further increases in the dye molecule concentration, the distance between adjacent dye
molecules becomes decreased to the point that there is self-quenching and the
fluorescence intensity of the viral particle begins to decrease. This overpacking of the viral
envelope with these lipophilic dyes can then act as a readout of the integrity of the viral
envelope. Upon attachment of the virus to the cell surface, there should be no membrane
mixing. However, after activation of the ligand-receptor complex, internal viral proteins are
activated to initiate the viral fusion process. Once this happens, a tracked viral particle will
display an increase in fluorescence as the envelope merges with the larger cell plasma
membrane, resulting in an increase in fluorescence as DiD molecules move farther apart
and no longer self-quench. Combining this simple readout of viral fusion with the ability to Figure 7. Monitoring
pinpoint the virus in its local environment at an exact moment allowed by 3D-DyPLoT will envelope fusion.
enable measurement of the exact chemical environment needed for envelope fusion to
occur. This is particularly relevant for pH independent viruses, which are believed to fuse with the outer cellular
membrane since they do not require a change in pH induced by endosomal maturation. By simultaneously
imaging the membrane with an IR membrane dye such as DiR, we will determine whether this process happens
at the cell membrane or only after internalization. Further, by performing rapid 3D imaging around the particle as
in Research Direction 3 we can correlate this process with receptors or chemokines. This work will firmly
determine the biochemical environment of viral envelope fusion for pH independent viruses.
D.3 Potential Obstacles. The most likely obstacle to this measurement is the long-time scale associated with
the entire viral fusion process. While the fusion itself is rapid, transition from the cell membrane to an endosome
may take 10 minutes or more. In this case, upon binding of the virus to the outer cell membrane, the observation
time will be increased from 10 µsec to 1 msec and the laser intensity reduced. This will reduce bleaching and
increase observation time to include the eventual envelope fusion step should it occur after endocytosis.
E. Conclusions and Future Directions. By generating a suite of methods for locking on to virus particles in
real-time in the live cell setting, this proposed work will provide a full 3D spatiotemporal map of viral infection
starting from the extracellular space. Beyond this work, these methods can be applied to correlating steps in the
infection process with internal protein dynamics (using FRET labels) and can be easily extended to other 3D
processes, such as exocytosis or import at the nuclear membrane.