Coronavirus (Covid-19): Latest updates and information
Skip to main content Skip to navigation

Spatiotemporal Localization of CamKII in Dendritic Spines

Abstract

The calcium calmodulin dependent kinase (CaMKII) is important for long-term potentiation at dendritic spines. Photo-activatable GFP (PaGFP) – CaMKII fusions were used to map CaMKII movements between and within spines in dissociated hippocampal neurons. Photo-activated PaGFP (GFP*) generated in the shaft spread uniformly, but was retained for about 1 s in spines. The differential localization of GFP*-CaMKII isoforms was visualized with hundred nanometer precision frame to frame using de-noising algorithms. GFP*-CaMKIIα localized to the tips of mushroom spines. The spatiotemporal profiles of native and kinase defective GFP*-CaMKIIβ, differed markedly from GFP*-CaMKIIα and mutant GFP*-CaMKIIβ lacking the association domain. CaMKIIβ bound to cortical actin in the dendrite and the stable actin network in spine bodies. Glutamate produced a transiently localized GFP*-CaMKIIα fraction and a soluble GFP*-CaMKIIβ fraction in spine bodies. Single molecule simulations of the interplay between diffusion and biochemistry of GFP* species were guided by the spatiotemporal maps and set limits on binding parameters. They highlighted the role of spine morphology in modulating bound CaMKII lifetimes. The long residence times of GFP*-CaMKIIβ relative to GFP*-CaMKIIα followed as consequence of more binding sites on the actin cytoskeleton than the post-synaptic density. These factors combined to retain CaMKII for tens of seconds, sufficient to outlast the calcium transients triggered by glutamate, without invoking complex biochemistry.

Publication

S Khan, TS Reese, NM Rajpoot, A Shabbir. Spatiotemporal Maps of CaMKII in Dendritic Spines, Journal of Computational Neuroscience (in press), December 2011.

[DOI for Full Version of the Paper]

Results

Spread of GFP* and GFP*-CaMKII proteins along dendrites

We first characterized the spread of GFP* (photo-activated GFP) and GFP*-CaMKII proteins along aspiny stretches of dendrite. PaGFP was photo-activated and RFP photobleached in the same region of interest (ROI) within one image scan. RFP photobleaching was used to identify the rradiated ROI. Substantial spread of GFP* along the dendrite occurred before acquisition of the first post-bleach scan was completed. The post-activation images showed uniform spread of the GFP* and GFP*-CaMKIIα from the irradiated ROI. GFP* fluorescence within the irradiated ROI decayed substantially between the first two post-activation scans indicating substantial diffusion within the time taken for image acquisition. In contrast, spread of GFP*-CaMKIIα. as markedly slower (Fig. 1(a)). A single exponential fit to the GFP*-CaMKIIα data, of form y0y0+a(exp)-bt)), had a correlation co-efficient R200.98, only marginally worse than a double exponential fit of form y0y0+a(exp(-bt))+c (exp(-ct)) (R200.994). We concluded that the spread of GFP*-CaMKIIα, as of GFP*, was dominated by a single process, consistent with diffusion of a homogenous species.

The membrane of the the dendritic shaft is lined with cortical actin (Kaech et al. 2001). CaMKIIβ bundles F-actin (O’Leary et al. 2006; Okamoto et al. 2007; Sanabria et al. 2009). GFP fusions of the native and mutant CaMKIIβ forms were expressed in a monkey kidney epithelial cell line (Cos-7) to validate phenotypes as assessed by interactions with the cell cortex (Shen et al. 1998). The association with the cell cortex of the native GFP- CaMKIIβ construct we used was shown previously (Khan et al. 2011). GFP-CaMKIIβΔ was uniformly distributed in the cell consistent with a weaker interaction with the actin cortex (Fig. 1(b)). In contrast, GFPCaMKIIβK43R formed string-like structures within the cortical actin (Fig. 1(c)), consistent with the documented stronger binder interaction of the non-phosphorylated form (Okamoto et al. 2007). When these constructs were expressed in hippocampal neurons, GFP*-CaMKIIβΔ* spread without forming congregations along the dendrite (Fig. 1(d)), but decay of its fluorescence in the irradiated ROI had a double exponential character in contrast to GFP*-CaMKIIα (Fig. 1(a)). GFPCaMKIIβK43R* spread more slowly (Fig. 1(a)) and sequestered strongly in dendritic spines relative to the shaft (Fig. 1(e)). We inferred, consistent with previous work, that GFP*-CaMKIIβ species bound F-actin in neurons and that the actin cytoskeleton was enriched in dendritic spines. We concluded that the relative mobility of the GFP*-CaMKIIβ species was consistent with their spread along the dendrite being dominated by interactions with cortical actin in addition to diffusion.

1a1.png 1b

Fig. 1 (a) Decay of fluorescence (± s. e) within irradiated ROI in dendritic shaft after photo-activation at time 0 s. GFP* (green), GFP*- CaMKIIα (orange), GFP*-CaMKβΔ (cyan), GFP*-CaMKβ (dark green), GFP*-CaMKβΚ43R (blue). Yellow line is a best-fit single exponential to the GFP*-CaMKIIα data. s.e0standard error. (b-c) Distributions of GFP*-CaMKIIβ species in Cos-7 cells co-expressing ER-RFP. Bar01 μm (b). Monomeric GFP-CaMKβΔ. (c) Nonphosphorylated GFP-CaMKβK43R. GFP-CaMKβΔ distributed uniformly. GFP-CaMKβK43R formed fibrous substructures in the cortex. (d-e) Averaged images constructed from the first 20 frames after photoactivation in hippocampal neurons co-expressing RFP. Red asterisks denote irradiated ROI. Bar05 μm (d) GFP*-CaMKIIβΔ (e) GFP*- CaMKIIβK43R. GFP*-CaMKIIβΔ was distributed similarly between dendrite and spine. GFP*-CaMKIIβK43R was sequestered in dendritic pines

A photo-activation protocol to study exchange between dendrite and spine

We next studied diffusion between dendrite and spine using photo-activated PaGFP (GFP*). The protocol involved irradiation of the dendritic shaft in a region adjacent to a spine (irradiated ROI). RFP was photo-bleached concurrently with PaGFP photo-activation in the irradiated ROI. Multiple pulses were applied in succession and the responses averaged. Two ROIs, equidistant from the irradiated ROI (Fig. 2(c)) monitored spread of GFP* along the dendrite and into the spine. Peak intensity of GFP* in the spine was attained before the first post-stimulus scan. GFP* was spread uniformly over the spine. GFP* efflux from the spine was slower than it’s loss from the irradiated ROI (Fig. 2(d)). This was evident in the time record of the GFP* fluorescence during the multiple pulse train as incomplete efflux of GFP* in the interval between pulses led to its gradual accumulation in the spine body (Fig. 2(d) inset). GFP* was also retained in the irradiated ROI, but to a lesser extent. The difference resulted in an elevated pre-stimulus GFP* spine body fluorescence in the averaged plot.We defined T1 and T2 as the half-rise time of GFP* fluorescence following photo-activation in the dendrite and the half residence time of GFP* respectively in the spine body. The RFP photobleaching marked the irradiated ROI since the irradiated ROI showed the largest drop in RFP fluorescence. The adjacent shaft also showed RFP fluorescence decrease due to diffusion of the bleached RFP molecules along the shaft. The spine showed a smaller, more gradual decrease due to permeation of the bleached RFP molecules (Fig. 2(e)). Since GFP* does not have binding targets within the spine, the difference between its influx and efflux kinetics must be due to constraints to diffusion set by spine viscosity or morphology or, alternatively, altered GFP* size due to conformational change or self-aggregation induced by the spine chemical environment.

2.png

Fig. 2 Single images of a dendritic spine in a hippocampal culture pre (a) and post (b) photo-activation of PaGFP. The neuron was cotransfected with PaGFP and RFP. Histograms of pixel fluorescence intensities of GFP* (green) and RFP (red) in (b) demonstrate uniform spread of GFP* within the duration of the first post-activation scan (77 milliseconds). Histogram left to right corresponds to fluorescence from spine tip to base along white arrowed line. (c) Averaged image of responses to the first 10 post-activation frames summed over four photo-activating pulses. Bar01 μm. Thirty frames (ca. 2 s) elapsed between activations. Circles denote ROIs from which are derived plots (below) showing the time courses of the intensity of GFP* and RFP fluorescence. Colored asterisks in each ROI match colors of symbols in plots: red0irradiated ROI; black0body of spine; green0dendritic shaft. (d) Averaged time plots of GFP* fluorescence (± s.e) Each plot is an average of 10 simultaneous photo-activation and photo-bleach pulses respectively. The pre-pulse fluorescence in the irradiated ROI was set to 1.0 in all cases. T10half-rise time and T20half-residence time in spine body (black). Inset: Time records (irradiated dendrite ROI (red), spine body (blue)) of GFP* fluorescence during the first 5 photo-activation pulses. Apparently elevated pre-stimulus value for spine body in averaged plot is due to greater accumulation of GFP* in the spine body ROI relative to the dendrite ROIs in intervals between pulses (see text). (e) Averaged time plots of RFP fluorescence (± s.e) from the same ROIs showing recovery after photobleaching

Spatiotemporal Maps of CaMKIIα in dendritic spines

The localization of GFP*-CaMKIIα in dendritic spines after photo-activation in the adjacent shaft was analyzed using de-noising algorithms based on the wavelet transform to achieve a resolved spatiotemporal readout in 6 mushroom spines. De-noising resolved the transiently localized GFP*- CaMKIIα at the spine tip from the soluble GFP*-CaMKIIα (Section 4). When irradiation was at the distal half of the same mushroom spine rather than in the adjacent shaft, a similar pattern emerged with the difference that permeation into the adjacent thin spine was negligible (Fig. 3(a)). It was possible that tightly bound PaGFP-CaMKIIα would make binding sites at the spine tip unavailable for the GFP*- CaMKIIα that diffused in from the dendrite. In such case, a substantial difference between the two protocols should have been observed in the amount of GFP*-CaMKIIα localized at the spine tip after photo-activation. This was not the case.

We also obtained spatiotemporal maps of GFP*-CaMKIIα in 5 thin (filipodial) and 5 stubby spines from two independent cultures. Thin spines were motile and frequently changed shape as these are immature spines (Bourne and Harris 2007). T2 times for GFP*-CaMKIIα in the spine bodies were similar, but discrete localization at the tip was not detected in any of the spines analyzed (Figs. 3(b), (d)).

We studied the effects of glutamate stimulation (five mushroom spines, each from a different culture) to resolve the “punctate” pattern documented by (Shen and Meyer 1999) in greater spatiotemporal detail. When photoactivation n the adjacent dendrite was triggered 3 to 5 min after glutamate stimulation, GFP*-CaMKIIα localized in spines for times long enough to be temporally separated from the soluble GFP*-CaMKIIα in the dendrite (Fig. 3(c)). The residence profiles in glutamate-stimulated spines had marked double-exponential character, as seen from comparison of the plot shown in Fig. 3(f) from those in Figs. 3(d),(e). Frame-by-frame analysis revealed that the major GFP*-CaMKIIα fraction localized in the spine body decayed more rapidly than the more persistent, minor fraction that localized to the spine tip (Fig. 4(a)). No substantial difference in either amount or residence of the GFP*-CaMKIIα at the tip was found when photo-activation was at the distal half of the spine, rather than the adjacent dendrite. We conclude that the GFP*-CaMKIIα sequestered in dendritic spines 3–5min after glutamate addition is composed of at least two fractions; a major fraction localized in the spine body and a minor, more persistent fraction localized at the spine tip. Formation within this period of a tightly bound GFPCaMKIIα fraction that would “persist” on the time scale of minutes (Bayer et al. 2006) was not observed.

3

Fig. 3 (a-c). GFP*-CaMKIIa in dendritic spines. Left panels show images averaged over the first 20 post-activation frames. Panels to right show de-noised single frames at 0 (pre-activation), 0.2, 1.2 and 2.4 s after activation. Asterisks mark irradiated ROIs as in Fig. 1. (a). Photo-activation at spine tip (orange asterisk). Compare with same spine after photo-activation in adjacent dendritic shaft (Methods). (b). Thin spine following photo-activation in shaft. (c). Mushroom spine following photo-activation in shaft 3–5 min after glutamate stimulation. M0Mushroom. T0Thin. (d–f). Averaged time plots of GFP* fluorescence (± s.e). Pre-stimulus values greater than 1.0 are due to GFP* molecules retained in the ROIs from preceding pulses as in Fig. 2d. Comparison of photo-activation in (d) dendritic shaft with (e) spine tip. Insets: Image averages of first 20 post-activation frames. (f) Time plot of image sequence shown in c. M-Tip0Mushroom spine tip. M-Body0mushroom spine body. Dendrite IROI0irradiated ROI at adjacent dendrite. M-Tip-IROI0irradiated ROI at mushroom spine tip. Symbols: Red0irradiated ROI in shaft; orange0irradiated ROI in spine tip; black0body of spine; yellow0distal spine tip; cyan0body of thin spine

Spatiotemporal maps of CaMKIIβ in dendritic spines

Spatiotemporal maps for GFP*-CaMKIIβΚ43R (9 spines/3 neurons) and GFP*-CaMKIIβ (4 spines/3 neurons) in dendritic spines were obtained as for GFP*-CaMKIIα and GFP*. Data for each species utilized two different cultures. Pronounced long-lived localization was documented for both constructs: T2 was>10 s for GFP*-CaMKIIβΚ43R fluorescence in the irradiated ROI had a double exponential character in contrast to GFP*-CaMKIIα (Fig. 1(a)). GFPCaMKIIβK43R* and 10±1.2 s for GFP*-CaMKIIβ. Both forms congregated in clusters centered in the spine body. Above background fluorescence was not detected in either the spine neck or regions adjacent to the spine membrane. The long T2 times enabled the clusters to be visualized readily in image maps over the post-activation sequence since the contribution of freely diffusing GFP*-CaMKIIβ species was averaged out (Figs. 5(a), (b), (d)). Frame by frame analysis of de-noised sequences revealed that both the cluster area as well as mean pixel intensity decreased in similar fashion with time (Fig. 4(b)). Experiments with GFP*-CaMKIIβΔ (4 spines/3 neurons) led to a different outcome (Fig. 5 (c), (f)). Neurons transfected with PaGFP-CaMKIIβΔ predominantly contained stubby spines (> 95% of 56 spines/2 independent cultures). The GFP*-CaMKIIβΔ fluorescence in the spine could not be temporally isolated from that in the shaft since T2 times for GFP*-CaMKIIβΔ in stubby spines were short (3.3±0.2 s).

The GFP*-CaMKIIβ generated by photo-activation in the dendrite 3–5 min after glutamate stimulation formed two populations in mushroom spines; a fast and a slow decay fraction (4 spines/4 cultures). The time constant of the fast fraction (1.8±0.3 s) was consistent with it being composed of soluble, freely diffusing GFP*-CaMKIIβ. The slowly decaying fraction in the spine body was stable for>10 s, analogous to GFP*-CaMKIIβΚ43R. We concluded that soluble GFP*-CaMKIIβ forms a large fraction of the total GFP*-CaMKIIβ in dendritic spines within 5 min after glutamate addition. Our observations were consistent with the intermediate stage of uniform localization reported by (Shen and Meyer 1999). Spine residence times for GFP* and all GFP*-CaMKII constructs are summarized in Fig. 4(c). T2 times for GFP*-CaMKIIα were greater than that for GFP*, lower than times for GFP*-CaMKIIβ or GFP*-CaMKIIβK43R, and comparable with those for GFP*-CaMKIIβΔ. Glutamate addition increased the amount, but not the lifetimes of the GFP*-CaMKIIα sequestered in spines.

5

Fig. 5 (a-d). Time averaged maps of GFP*-CaMKIIβ in dendritic spines after photo-activation in the shaft. Time windows; 1–4 s (a,d); 1–2 s (b); 0–1 s (c). Photo-activation at 0 s. Red asterisks denote irradiated ROIs as in Fig. 1. GFP*-CaMKIIβK33R in dendrites with 1.5 μm (a) and 0.75 μm (b) diameter shaft; spine in (b) is motile. Spines in (a) and (d) have prominent necks. (c.) GFP*-CaMKIIβΔ. d.) GFP*-CaMKIIβ 3–5 min after glutamate addition. Bar (a,b)0 1 μm. Bar (c,d)01 μm. (e, f, g). Averaged time plots of GFP* luorescence (± s.e) in sequences from which the averaged image maps shown in (a, c, d) respectively were obtained. Symbols: red0irradiated ROI in dendritic shaft; black0spine body. Photo-activation at 0 s. Prestimulus values greater than 1.0 are due to GFP* molecules retained in the ROIs from preceding pulses as in Fig. 2d

Different GFP*-CaMKII proteins manifest different spatiotemporal maps

GFP*-CaMKIIα congregated at the spine tip consistent with binding to targets on the PSD (Bayer et al. 2006; Okabe 2007; Otmakhov et al. 2004). GFP*-CaMKIIβ bound to cortical actin and congregated in the spine body. The GFP*-CaMKIIβK43R mutant allowed analysis uncomplicated by downstream phosphorylation reactions. GFP*- CaMKIIβ was absent from spine necks, indicating the absence of the actin cytoskeleton from the thin necks of mushroom spines, as reported previously (Honkura et al. 2008). It was also absent in regions adjacent to the spine membrane. CaMKIIβ would bind G-actin (Okamoto et al. 2007; Sanabria et al. 2009), but this binding interaction would not be detected in our experiments since the Gactin/ CaMKIIβ complex would diffuse almost as rapidly as CaMKIIβ alone.

Neurons bathed in glutamate for up to 5 min were used to study the early stimulus induced changes in spatiotemporal characteristics. The expectation was that basal calcium levels would have been restored after a transient spike over this period (Shen et al. 2000), after generating a persistently activated CaMKII fraction (Bayer et al. 2006). Our results were broadly in line with previous work, but with interesting anomalies. Glutamate stimulation generated a GFP*-CaMKIIα fraction in the spine body that transiently overwhelmed the fraction associated more persistently at the spine tip. Reversible CaMKII aggregates were found in spine bodies under excitotoxic conditions (Grant et al. 2008) consistent with this observation. Alternatively, GFP*-CaMKIIα could have bound to actin cytoskeleton associated proteins as discussed in the next section. Glutamate increased the soluble GFP*-CaMKIIβ fraction, consistent with the switch from an actin-binding pattern to a more uniform distribution recorded by (Shen and Meyer 1999). However, GFP*-CaMKIIβ localization at the spine tip expected from the subsequent transition to a punctate pattern indicative of PSD-bound CaMKII (Shen and Meyer 1999) was not observed. Nor did we observe a glutamate induced increase in the persistence of GFP*-CaMKIIα at the spine tip (Bayer et al. 2006). In contrast to (Bayer et al. 2006) we did not shift to glutamate free media after the 5-min incubation period. It is possible that additional reactions are triggered by the downshift, but further work will be needed to resolve these anomalies. The focus of the present study was on dendrite-spine exchange on the second timescale comparable to the lifetime of the glutamate induced calcium transient.

Photo-activation experiments reveal sub-spine architecture

This study adds to photo-activation measurements made previously on slower time scales. Whole spine photoactivation documented two time constants for efflux of GFP*-CaMKIIα; a fast time constant of about a minute and a slower time constant of about 20 min (Lee et al. 2009). Neither time constant would reflect the exchange of soluble CaMKII documented in our experiments and accounted for in our simulations based on plausible estimates of CaMKII oligomer size and spine viscosity. Whole spine photo-activation protocols do not allow visualization of transient sub-spine structures or a determination of whether efflux was limited by the spine neck (Byrne et al. 2011) versus dissociation from binding compartments within the spine head. Mixed PaGFP-CaMKIIα \ CaMKIIβ dodecamers were likely to be predominant in (Lee et al. 2009) since expression of PaGFP-CaMKIIα was comparable to endogenous CaMKII; whereas PaGFP-CaMKIIα expression in this study should be similar to that of GFPCaMKIIα by (Shen and Meyer 1999; Shen et al. 1998).

Photo-activation measurements of PaGFP-actin (Honkura et al. 2008) support the view that there are separate F-actin compartments within spines; a dynamic zone underneath the PSD and spine membrane and a stable domain centered in the core of the spine head. A third domain, stabilized by CaMKII, may form at the base of the spine after LTP induction. Regulation of the F-actin mesh at the spine base by cofilin has been proposed to gate protein transport into spines (Ouyang et al. 2005). Numerical simulations lend credence to this proposal (Byrne et al. 2011). Retrograde flow of GFP*-actin from tip to base due to treadmilling occurs within 5 min (Honkura et al. 2008). The GFP*-CaMKIIα at the PSD is limited by the number of binding sites, so excess GFP*-CaMKIIα may move with the retrograde actin flow to the spine body following glutamate stimulation. Actin binding proteins that also bind CaMKIIα, such as α-actinin (Okabe 2007), could be part of this flow from the PSD to spine body.

GFP*-CaMKIIβ localization should, in principle, provide a map of the stable actin cytoskeleton. Consistent with this view, our measurements together with the PaGFP-actin data imply that CaMKIIβ associates with the stable pool centered in the spine body. GFP*-CaMKIIβ binding to the dynamic actin pool identified by (Honkura et al. 2008) at spine tips is either negligible or too transient to be separated from the soluble pool.

The role of spine morphology in CaMKII binding site occupancy

The role of spine neck morphology as a regulatory element has been considered in some depth (Bloodgood and Sabatini 2007; Denk et al. 1996; Noguchi et al. 2005; Pi et al. 2010). The spines analyzed in the present study were selected by the dendrite photo-activation protocol to be open rather than closed as assessed by the fact that initial influx was too rapid (< 0.1 s) to be resolved. However, soluble CaMKII in the spine generated either by influx or dissociation of bound complexes was retained over many seconds.

Our experimental measurements and simulations, taken together, showed that these times were a consequence, most simply, of spine geometry coupled to binding site architecture. Non-mushroom, stubby spines that lacked a discernable neck retained bound CaMKII less effectively than mushroom spines. Narrow spine heads in thin spines were more effective in trapping CaMKII than the wider heads of stubby spines, but less so than mushroom spine heads coupled to long, narrow necks. Soluble CaMKII concentrations remained elevated for longer times in mushroom spines due to reduced permeation out into the dendrite, prolonging binding site occupancies. Thus, this study illustrates how sequestration of key signal molecules, like CaMKII, by binding compartments can be modulated by spine morphology over an order of magnitude, from seconds to tens of seconds. Increased binding affinity also increased residence times, but at the cost of response speed since turnover of the complexes was inhibited. A moderate binding affinity coupled to a wide head/long neck geometry, such as that found in mushroom spines provides the optimal solution.

References

Andrews, S. S., Addy, N. J., Brent, R., & Arkin, A. P. (2010). Detailed simulations of cell biology with Smoldyn 2.1. PLoS Computational Biology, 6, e1000705.

Bayer, K. U., LeBel, E., McDonald, G. L., O’Leary, H., Schulman, H., & De Koninck, P. (2006). Transition from reversible to persistent binding of CaMKII to postsynaptic sites and NR2B. Journal of Neuroscience, 26, 1164–1174.

Bloodgood, B. L., & Sabatini, B. L. (2005). Neuronal activity regulates diffusion across the neck of dendritic spines. Science, 310, 866– 869.

Bloodgood, B. L., & Sabatini, B. L. (2007). Ca(2+) signaling in dendritic spines. Current Opinion in Neurobiology, 17, 345–351.

Boulanger, J., Kervrann, C., Bouthemy, P., Elbau, P., Sibarita, J. B., & Salamero, J. (2010). Patch-based nonlocal functional for denoising fluorescence microscopy image sequences. IEEE Transactions on Medical Imaging, 29, 442–454.

Bourne, J., & Harris, K. M. (2007). Do thin spines learn to be mushroom spines that remember? Current Opinion in Neurobiology, 17, 381–386.

Bourne, J. N., & Harris, K. M. (2008). Balancing structure and function at hippocampal dendritic spines. Annual Review of Neuroscience, 31, 47–67.

Byrne, M.J., Waxham, M.N., Kubota, Y., (2011). The impacts of geometry and binding on CaMKII diffusion and retention in dendritic spines. Journal of Computational Neuroscience.

Cingolani, L. A., & Goda, Y. (2008). Actin in action: the interplay between the actin cytoskeleton and synaptic efficacy. Nature Reviews. Neuroscience, 9, 344–356.

Collins, T. J. (2007). ImageJ for microscopy. Biotechniques, 43, 25–30.

De Decker, A., Lee, J. A., & Verleysen, M. (2010). A principled approach to image denosing with similarity kernels involving patches. Neurocomputing, 73, 1198–1209.

Denk, W., Yuste, R., Svoboda, K., & Tank, D. W. (1996). Imaging calcium dynamics in dendritic spines. Current Opinion in Neurobiology, 6, 372–378.

Grant, P. A., Best, S. L., Sanmugalingam, N., Alessio, R., Jama, A. M., & Torok, K. (2008). A two-state model for Ca2+/CaM-dependent protein kinase II (alphaCaMKII) in response to persistent Ca2+ stimulation in hippocampal neurons. Cell Calcium, 44, 465–478.

Honkura, N., Matsuzaki, M., Noguchi, J., Ellis-Davies, G. C., & Kasai, H. (2008). The subspine organization of actin fibers regulates the structure and plasticity of dendritic spines. Neuron, 57, 719–729.

Huang, B., Bates, M., & Zhuang, X. (2009). Super-resolution fluorescence microscopy. Annual Review of Biochemistry, 78, 993–1016.

Hudmon, A., Lebel, E., Roy, H., Sik, A., Schulman, H., Waxham, M. N., & De Koninck, P. (2005). A mechanism for Ca2+/calmodulin-dependent protein kinase II clustering at synaptic and nonsynaptic sites based on self-association. Journal of Neuroscience, 25, 6971–6983.

Kaech, S., Brinkhaus, H., & Matus, A. (1999). Volatile anesthetics block actin-based motility in dendritic spines. Proceedings of the National Academy of Sciences of the United States of America, 96, 10433–10437.

Kaech, S., Parmar, H., Roelandse, M., Bornmann, C., & Matus, A. (2001). Cytoskeletal microdifferentiation: a mechanism for organizing morphological plasticity in dendrites. Proceedings of the National Academy of Sciences of the United States of America, 98, 7086–7092.

Kanaseki, T., Ikeuchi, Y., Sugiura, H., & Yamauchi, T. (1991). Structural features of Ca2+/calmodulin-dependent protein kinase II revealed by electron microscopy. The Journal of Cell Biology, 115, 1049–1060.

Khan, S., Zou, Y., Amjad, A., Gardezi, A., Smith, C. L.,Winters, C., & Reese, T. S. (2011). Sequestration of CaMKII in dendritic spines in silico. Journal of Computational Neuroscience, 31, 581–594.

Kolodziej, S. J., Hudmon, A., Waxham, M. N., & Stoops, J. K. (2000). Three-dimensional reconstructions of calcium/calmodulin-dependent (CaM) kinase IIalpha and truncated CaM kinase IIalpha reveal a unique organization for its structural core and functional domains. Journal of Biological Chemistry, 275, 14354–14359.

Lee, S. J., Escobedo-Lozoya, Y., Szatmari, E. M., & Yasuda, R. (2009). Activation of CaMKII in single dendritic spines during long-term otentiation. Nature, 458, 299–304.

Lippincott-Schwartz, J., Altan-Bonnet, N., Patterson, G.H., (2003). Photobleaching and photoactivation: following protein dynamics in living cells. Nature Cell Biology Suppl, S7–14.

Lisman, J. E., Raghavachari, S., & Tsien, R. W. (2007). The sequence of events that underlie quantal transmission at central glutamatergic synapses. Nature Reviews. Neuroscience, 8, 597–609.

Noguchi, J., Matsuzaki, M., Ellis-Davies, G. C., & Kasai, H. (2005). Spine-neck geometry determines NMDA receptor-dependent Ca2 + signaling in dendrites. Neuron, 46, 609–622.

Northrup, S. H., & Erickson, H. P. (1992). Kinetics of protein-protein association explained by Brownian dynamics computer simulation. Proceedings of the National Academy of Sciences of the United States of America, 89, 3338–3342.

Okabe, S. (2007). Molecular anatomy of the postsynaptic density. Molecular and Cellular Neuroscience, 34, 503–518.

Okamoto, K., Narayanan, R., Lee, S. H., Murata, K., & Hayashi, Y. (2007). The role of CaMKII as an F-actin-bundling protein crucial for maintenance of dendritic spine structure. Proceedings of the National Academy of Sciences of the United States of America, 104, 6418–6423.

Okamoto, K., Bosch, M., & Hayashi, Y. (2009). The roles of CaMKII and F-actin in the structural plasticity of dendritic spines: a potential molecular identity of a synaptic tag? Physiology (Bethesda, Md.), 24, 357–366.

O’Leary, H., Lasda, E., & Bayer, K. U. (2006). CaMKIIbeta association with the actin cytoskeleton is regulated by alternative splicing. Molecular Biology of the Cell, 17, 4656–4665.

Otmakhov, N., Tao-Cheng, J. H., Carpenter, S., Asrican, B., Dosemeci, A., Reese, T. S., & Lisman, J. (2004). Persistent accumulation of calcium/calmodulin-dependent protein kinase II in dendritic spines after induction of NMDA receptordependent chemical long-term potentiation. Journal of Neuroscience, 24, 9324–9331.

Ouyang, Y., Wong, M., Capani, F., Rensing, N., Lee, C. S., Liu, Q., Neusch, C., Martone, M. E., Wu, J. Y., Yamada, K., Ellisman, M. H., & Choi, D. W. (2005). Transient decrease in F-actin may be necessary for translocation of proteins into dendritic spines. The European Journal of Neuroscience, 22, 2995–3005.

Patterson, G. H., & Lippincott-Schwartz, J. (2002). A photoactivatable GFP for selective photolabeling of proteins and cells. Science, 297, 1873–1877.

Pi, H. J., Otmakhov, N., El Gaamouch, F., Lemelin, D., De Koninck, P., & Lisman, J. (2010). CaMKII control of spine size and synaptic strength: role of phosphorylation states and nonenzymatic action. Proceedings of the National Academy of Sciences of the United States of America, 107, 14437–14442.

Rajpoot, N. M., Wilson, R. G., Meyer, F. G., & Coifman, R. R. (2003). Adaptive wavelet packet basis selection for zerotree image coding. IEEE Transactions on Image Processing, 12, 1460–1472.

Rajpoot, N. M., Yao, Z., & Wilson, R. G. (2004). Adaptive wavelet restoration of noisy video sequenes. Proceedings of the International Conference Image Processing, 2, 957–960.

Sanabria, H., Digman, M. A., Gratton, E., & Waxham, M. N. (2008). Spatial diffusivity and availability of intracellular calmodulin. Biophysical Journal, 95, 6002–6015.

Sanabria, H., Swulius, M. T., Kolodziej, S. J., Liu, J., & Waxham, M. N. (2009). {beta}CaMKII regulates actin assembly and structure. Journal of Biological Chemistry, 284, 9770–9780.

Schaus, T. E., Taylor, E. W., & Borisy, G. G. (2007). Selforganization of actin filament orientation in the dendriticnucleation/ array-treadmilling model. Proceedings of the National Academy of Sciences of the United States of America, 104, 7086–7091.

Sharma, K., Fong, D. K., & Craig, A. M. (2006). Postsynaptic protein mobility in dendritic spines: long-term regulation by synaptic NMDA receptor activation. Molecular and Cellular Neuroscience, 31, 702–712.

Shen, K., & Meyer, T. (1998). In vivo and in vitro characterization of the sequence requirement for oligomer formation of Ca2+/calmodulin- dependent protein kinase IIalpha. Journal of Neurochemistry, 70, 96–104.

Shen, K., & Meyer, T. (1999). Dynamic control of CaMKII translocation and localization in hippocampal neurons by NMDA receptor stimulation. Science, 284, 162–166.

Shen, K., Teruel, M. N., Subramanian, K., & Meyer, T. (1998). CaMKIIbeta functions as an F-actin targetingmodule that localizes CaMKIIalpha/ beta heterooligomers to dendritic spines. Neuron, 21, 593–606.

Shen, K., Teruel,M. N., Connor, J. H., Shenolikar, S., &Meyer, T. (2000). Molecular memory by reversible translocation of calcium/calmodulin- dependent protein kinase II. Nature Neuroscience, 3, 881–886.

Sheng, M., & Hoogenraad, C. C. (2007). The postsynaptic architecture of excitatory synapses: a more quantitative view. Annual Review of Biochemistry, 76, 823–847.

Shroff, H., Galbraith, C. G., Galbraith, J. A., & Betzig, E. (2008). Livecell photoactivated localization microscopy of nanoscale adhesion dynamics. Nature Methods, 5, 417–423.

Yang, L., Parton, R., Ball, G., Qiu, Z., Greenaway, A. H., Davis, I., & Lu, W. (2010). An adaptive non-local means filter for denoising live-cell images and improving particle detection. Journal of Structural Biology, 172, 233–243.