Investigating sub-spine actin dynamics in rat hippocampal neurons with super-resolution optical imaging.

Cell Culture, Plasmids and Transfection The vector carrying dimer EosFP coding sequence was a generous gift from Dr. Wiedenmann (University of Ulm). The human actB cDNA sequence was amplified with PCR from a cDNA plasmid (OpenBiosystems, accession number BC004251 ), which was part of the IMAGE cDNA clone collection and inserted at C-term of the EosFP gene. Embryonic hippocampal neuron cultures were prepared from E17 Sprague–Dawley rat (Charles River Laboratories). Cells were plated on glass bottom culture dishes that were thoroughly cleaned by sonicating sequentially in 10% HCl, 20% NaOH and Millipore water and coated with poly-L-lysine (Sigma) overnight before use. In some cases cells were transfected by nucleofection (Amaxa-Germany) with EosFP-actin plasmid DNA (3-4 µg) before seeding at low-to-medium density. Cells were allowed to recover for 3 hrs in DMEM. After 3 hrs DMEM was replaced by Neurobasal media containing supplements. Cell cultures were maintained for up to 25 DIV in Neurobasal medium (Gibco) supplemented with B27 before switching to Hibernate E media (BrainBits) before the microscopy experiment. For FRAP experiments on filipodia and spines Lipofectamine 2000 (Invitrogen) was used to transfect neurons 8-16 DIV according to manufacturers recommendations. Xenopus embryonic fibroblast cells, and Xenopus melanophore was transfected via microinjecting plasmid DNA into the cells 24 hours prior to imaging. Before imaging the cells were transfered to hibernate E media. To inhibit actin filament dynamics cells were incubated in Hibernate E supplemented with 1 µM Jasplakinolide (Invitrogen) (stock at 1 mM dissolved in methanol and kept at -20°C).

Immunostaining Hippocampal neurons 20-22

DIV nucleofected with EosFP-Actin during plating were fixed in 4% paraformaldehyde in PBS for 20 minutes. Cells were washed 3 times (5 mins.)in PBS and incubated for 20-25 min in a solution containing 10% horse serum, 0.125% TritonX-100 in PBS. Next the cells were incubated with either mouse αPSD-95 antibody (1∶500) (a generous gift from Dr. Matt Rasband, Baylor college of Medicine, TX) or stained with Texas red labeled phalliodin (Invitrogen) for 1 hr and washed 3 times in PBS. Cells stained with mouse αPSD-95 antibody were incubated with α-mouse Alexafluor 647 secondary antibody (1∶400) (Molecular Probes) for 1 hr. Cells were fixed in Prolong Gold antifade (Invitrogen) and imaged under a Zeiss 510 confocal microscope with a 63×1.4 N.A. Oil immersion objective.

Single-Molecule Microscopy

Single-molecule fluorescence images were taken with a modified epi-fluorescence microscope (Olympus IX81) equipped with 60x microscope objective (NA = 1.45, Olympus) and a TE-cooled EM-CCD camera (PhotonMax, Roper Scientific, Trenton, NJ). To image single-molecules with PALM microscopy, a 405 nm diode laser (Cube laser system, Coherent Inc., Santa Clara, CA) was the light source for photo-activation. The activation pulse was controlled controlled by the computer through a D/A conversion interface (NI-USB-6251, National Instruments, Austin, TX). The green fluorescence of unactivated EosFP was excited with the 488 nm laser line from an argon ion laser (Melles Griot). For single-molecule imaging a 532 nm DPSS diode laser (Lambda) was used for excitation at a power density of ∼0.4 kW/cm. Unless otherwise stated, single-molecule time-lapse images were taken at 2 sec interval with 400 ms acquisition time for each image. Image acquisition software was built on top of the μManager platform ( http://micro-manager.org ).

Image Analysis

Single molecule trajectories (positions versus time) were constructed using standard particle tracking algorithm. From each data set, a region of interest that included only the dendritic spine region was manually defined based on the green-fluorescence signal. Single-molecule trajectories outside this region of interest was excluded from the dataset. So was trajectories shorter than 12 sec. Further analysis of the single molecule trajectories was carried out in Matlab.

FRAP and Confocal Microscopy

Neurons were transfected with the actin-Eos construct using lipofectamine2000 at 14-21DIV to observe actin dynamics in spines. 24-48 hours after transfection neurons were medium changed to low fluorescence hibernate E buffer and observed under a Zeiss LSM510 meta confocal microscopy. Spines were observed with a 40X, 1.2 numerical aperture(NA) water-immersion objective. All data was collected at room temperature. 20 frames were scanned at 5sec interval with fully opened pinhole. To photo-bleach spines 488nm laser was concentrated on a area of 1 µm×2 µm area (10 iterations). Images were collected before and after bleaching using low laser intensities to reduce bleach while monitoring and recovery was monitored for 2 min. EosFP-Actin Photobleaching Kinetics Single-molecule photobleaching kinetics of EosFP-actin is measured in xenopus fibroblast cells treated with jasplakinolide (see Figure S2 ). The actin molecules on stress fibers after the cells were treated are extremely stable and provide a good way to measure the photobleaching kinetics under imaging conditions that is close to those in the the neuronal cell experiments. We analyzed more than 800 single molecule trajectories to determine the time it took to photobleach single molecules. The histogram is best-fitted with a double exponential decay function, with the first decay time constant τ 1 = 12.3 imaging frames (24.6 sec) accounting for 97% of the molecules. A much smaller fraction (3%) of the molecules had a much slower photobleaching kinetics with a time constant τ 2 = 66.1 frames. Such a long tail in the photobleaching kinetics is common for fluorescent protein molecules. This second slow-bleaching population is particularly obvious in jasplakinolide treated cells and easy to see in time-lapse movies because they lasted for such a long time, but are almost never seen in the untreated neuronal cells, because the turnover of the filament is almost always faster. Resolution of Measuring Single-Molecule Centroid Positions We determined the resolution of single-molecule centroid positions by imaging EosFP-actin in neurons treated with jasplakinolide. Positions of individual actin molecules were repeated measured by the PALM microscope in 2 s intervals. The fluctuation in the measured positions were recorded and the distribution was plotted ( Figure S2 ). The distribution fit well with the theoretical prediction of probability density: assuming a symmetrical 2D random noise that is normally distributed. The parameter from the fitting indicated that the error in positional measurements was 28.5 nm. The average contribution of the measurement error to MSR is:

Supporting Information Figure S1 2D scatter plot of the Sv and MSR values from computer-simulated single-molecules trajectories. A1. Simulation of stationary molecules. The spread is due to measurement error. A2. Molecules under random walk with an equivalent diffusion constant of 100 nm2/s. The random motions mostly push the distribution into region III, but also “leaks” into region II because of the limited trajectory lengths. A3. Molecules undergoing only vectorial movement at 20 nm/s and without random motions. The distribution is clustered in region II. (0.24 MB TIF) Click here for additional data file. Figure S2 Characterization of the EosFP-actin single-molecule property using jasplakinolide stabilized stress fibers. A. Fluorescence image of a Xenopus fibroblast cell after jasplakinolide treatment showing clear stress fiber structures. B. A single molecule image of the boxed region in A. C. Histogram of the length of single-molecule trajectories before photo-bleaching. The data was fitted with double exponential function. D. Histogram of single-molecule positional measurement error obtained by repeatedly measuring the position of the same actin molecule on the stress fiber. (0.36 MB TIF) Click here for additional data file. Movie S1 Retrograde flow of actin network in lamellipodia region of a Xenopus fibroblast cell, visualized by single actin molecule imaging. The movie was sped up 40 times from real time. (1.85 MB AVI) Click here for additional data file. Movie S2 Retrograde flow of actin network in lamellipodia region of a Xenopus melanocyte cell, visualized by single actin molecule imaging. Beginning of the movie shows shows the overall fluorescence signals from the activated actin molecules. The red traces were the trajectories of individual molecules. The movie was sped up 80 times from real time. (10.32 MB AVI) Click here for additional data file. Movie S3 Single actin molecules in dendritic spines showing vectorial movements of various directions. The left side image shows the green fluorescence signal. The yellow pointers indicate retrograde moving molecules; the green pointers indicate anterograde moving molecules; and the red pointers indicate randomly moving molecules. The movie was sped up 40 times. (4.97 MB MOV) Click here for additional data file. Movie S4 Actin molecules in dendrites and spines after jasplakinolide treatment. Almost all movements stopped. The movie was sped up 80 times. (1.95 MB AVI) Click here for additional data file. Movie S5 Bidirectional movements of two actin molecules in the same spine. The left side fluorescence image shows the morphology of the dendritic process. The movie was sped up 20 times. (0.76 MB AVI) Click here for additional data file.