Overexpression of Tropomyosin Isoform Tpm3.1 Does Not Alter Synaptic Function in Hippocampal Neurons

Tropomyosin (Tpm) has been regarded as the master regulator of actin dynamics. Tpms regulate the binding of the various proteins involved in restructuring actin. The actin cytoskeleton is the predominant cytoskeletal structure in dendritic spines. Its regulation is critical for spine formation and long-term activity-dependent changes in synaptic strength. The Tpm isoform Tpm3.1 is enriched in dendritic spines, but its role in regulating the synapse structure and function is not known. To determine the role of Tpm3.1, we studied the synapse structure and function of cultured hippocampal neurons from transgenic mice overexpressing Tpm3.1. We recorded hippocampal field excitatory postsynaptic potentials (fEPSPs) from brain slices to examine if Tpm3.1 overexpression alters long-term synaptic plasticity. Tpm3.1-overexpressing cultured neurons did not show a significantly altered dendritic spine morphology or synaptic activity. Similarly, we did not observe altered synaptic transmission or plasticity in brain slices. Furthermore, expression of Tpm3.1 at the postsynaptic compartment does not increase the local F-actin levels. The results suggest that although Tpm3.1 localises to dendritic spines in cultured hippocampal neurons, it does not have any apparent impact on dendritic spine morphology or function. This is contrary to the functional role of Tpm3.1 previously observed at the tip of growing neurites, where it increases the F-actin levels and impacts growth cone dynamics.


Introduction
Most synapses between excitatory neurons form on dendritic spines, which are protrusions from dendrites that consist of a large, actin-rich head separated from the dendrite shaft by a thin neck. Actin exists in multiple populations in dendritic spines, with different functions and rates of turnover [1]. Most actin present in spines is dynamic and can change its organisation very rapidly [2].
The process of learning involves structural changes to synapses. The development and plasticity of dendritic spines are reliant on the flexibility of the actin cytoskeleton. The most studied cellular model of synaptic plasticity is long-term potentiation (LTP). In LTP, high-frequency stimulation causes a signalling cascade that leads to insertion of AMPA receptors, enlargement of dendritic spines, an increase in synaptic strength [3], and polymerisation of actin [4].

The Expression of Tpm3.1 Is Highly Increased in Tpm3.1 Tg Mice and Is Enriched in Dendritic Spines
We first analysed the overexpression of Tpm3.1 in Tpm3.1-overexpressing transgenic mice (Tpm3.1 Tg). Brains were extracted from Tpm3.1 Tg and wild-type (WT) mice, prepared into a brain homogenate, and further processed to isolate the PSD. The expression level of Tpm3.1 was analysed using the LC1 antibody to detect exogenous human Tpm3.1 (hTpm3.1), and the γ 9d antibody to detect all Tpm3.1/2 [21]. As expected, WT mice did not express any hTpm3.1 ( Figure 1A). Tpm3.1 Tg mice had an approximately threefold increase in Tpm3.1 in the total brain homogenate ( Figure 1B,C), and a fivefold increase in the PSD-associated fraction ( Figure 1D,E).
To investigate the effect of Tpm3.1 overexpression, we used primary hippocampal neurons prepared from mouse embryos overexpressing Tpm3.1 and WT littermates. Cells were fixed and stained for hTpm3.1 to identify transgenic cells. We confirmed that, as with endogenous Tpm3.1 [12], hTpm3.1 is enriched in dendritic spines ( Figure 2). To investigate the effect of Tpm3.1 overexpression, we used primary hippocampal neurons prepared from mouse embryos overexpressing Tpm3.1 and WT littermates. Cells were fixed and stained for hTpm3.1 to identify transgenic cells. We confirmed that, as with endogenous Tpm3.1 [12], hTpm3.1 is enriched in dendritic spines ( Figure 2).   To investigate the effect of Tpm3.1 overexpression, we used primary hippocampal neurons prepared from mouse embryos overexpressing Tpm3.1 and WT littermates. Cells were fixed and stained for hTpm3.1 to identify transgenic cells. We confirmed that, as with endogenous Tpm3.1 [12], hTpm3.1 is enriched in dendritic spines ( Figure 2). Staining for hTpm3.1 shows that it is enriched in dendritic spines. From left to right: merged image with mCherry in magenta and hTpm3.1 in green, mCherry channel, and hTpm3.1 channel.

Figure 2.
Staining for hTpm3.1 shows that it is enriched in dendritic spines. From left to right: merged image with mCherry in magenta and hTpm3.1 in green, mCherry channel, and hTpm3.1 channel.

Tpm3.1 Overexpression Does Not Significantly Increase F-Actin in Dendritic Spines
To examine the effect of increased Tpm3.1 on actin in dendritic spines, we transduced neurons with cytosolic mRuby2 and F-tractin, a probe for F-actin fused with EGFP. Cells were fixed and immunostained for hTpm3.1 to identify transgenic cells, and spine F-tractin

Tpm3.1 Overexpression Does Not Significantly Increase F-Actin in Dendritic Spines
To examine the effect of increased Tpm3.1 on actin in dendritic spines, we transduced neurons with cytosolic mRuby2 and F-tractin, a probe for F-actin fused with EGFP. Cells were fixed and immunostained for hTpm3.1 to identify transgenic cells, and spine F-tractin fluorescence was measured. The was little or no difference in spine F-tractin in transgenic and wild-type neurons ( Figure 3C; WT: 1.0 ± 0.08; n = 14; Tpm3.1 Tg: 0.99 ± 0.14; n = 13, Mann-Whitney U = 79; p = 0.57).

The Effect of Tpm3.1 Overexpression on Dendritic Spine Morphology
We used a mixed culture system to study the effect of Tpm3.1 overexpression on spine morphology. Neurons were transfected to express fluorescent protein mCherry at 14 DIV, fixed at 17-18 DIV, and immunostained for hTpm3.1. Dendrite lengths of 20 µm were imaged, and individual protrusions from the dendrite were measured and counted.

The Effect of Tpm3.1 on Long-Term Potentiation in Brain Slices
To examine the effect of Tpm3.1 overexpression on synaptic plasticity, we prepared acute brain slices from Tpm3.1 transgenic mice and recorded extracellular field potentials from the hippocampus ( Figure 6). An input-output (IO) curve was generated, measuring the response size of the fEPSP slope and fibre volley at various levels of stimulation. From these IO curves, we measured the fEPSP slope, representing the synaptic current, and the fibre volley, representing the number of terminals activated by the stimulus. Plotted together, these show the synaptic efficacy, which is the postsynaptic response evoked per presynaptic stimulation. There were also no differences in the mEPSC rise time (WT: 0.82 ± 0.05 ms; Tg: 0.69 ± 0.04 ms; Mann-Whitney test p = 0.2) or decay (WT: 4.8 ± 0.470 ms; Tg: 4.0 ± 0.29 ms; Mann-Whitney test p = 0.32).

The Effect of Tpm3.1 on Long-Term Potentiation in Brain Slices
To examine the effect of Tpm3.1 overexpression on synaptic plasticity, we prepared acute brain slices from Tpm3.1 transgenic mice and recorded extracellular field potentials from the hippocampus ( Figure 6). An input-output (IO) curve was generated, measuring the response size of the fEPSP slope and fibre volley at various levels of stimulation. From these IO curves, we measured the fEPSP slope, representing the synaptic current, and the fibre volley, representing the number of terminals activated by the stimulus. Plotted together, these show the synaptic efficacy, which is the postsynaptic response evoked per presynaptic stimulation.  We observed no difference in the paired pulse ratio (PPR) at baseline ( Figure 6G; unpaired t-test; t = 0.31, p = 0.76), the fEPSP slope ( Figure 6A; two-way RM ANOVA interaction p = 0.25, genotype p = 0.71), or the fibre volley ( Figure 6B; two-way RM ANOVA interaction p ≥ 0.99, genotype p = 0.72). There was also no apparent difference in synaptic efficacy between the two groups ( Figure 6C). Overall, basal synaptic function was not affected by Tpm3.1 overexpression in the brain slices.
The response was then recorded for 60 min ( Figure 6D). We observed no difference between groups, both in the first 10 min and last 10 min of recording after LTP induction ( Figure 6E; two-way ANOVA interaction p = 0.33, genotype p = 0.51). We also saw no difference in the PPR at any time point ( Figure 6G; two-way RM ANOVA interaction p = 0.11, genotype p = 0.67). Together, these results suggest that overexpressing Tpm3.1 has no effect on synaptic plasticity.

Discussion
In this study, we examined the effect of Tpm3.1 overexpression on dendritic spine morphology and synapse function using transgenic mice that express hTpm3.1. Previous studies have shown that Tpm3.1 is enriched in dendritic spines [11,12]. Here, we show that hTpm3.1, expressed in transgenic cells, also localises to dendritic spines. Importantly, there was a fivefold increase in total Tpm3.1 in the PSD-associated fraction. These results suggest that overexpressed Tpm3.1 is properly trafficked.
In neurons, Tpm3.1 localises to several actin-rich compartments including the growth cone and axon initial segment [17,19]. Previous studies showed hTpm3.1 overexpression increases the size of actin-rich growth cones at the tip of growing neurites and the amount of F-actin present in growth cones [22]. While transgenic neurons had more Tpm3.1 associated with the postsynaptic specialisation, the amount of F-actin present in spines was not different from the control. There was also no apparent difference in the length and width of the spines. We observed a modest decrease in the number of thin spines in Tpm3.1 Tg cells. However, this decrease was not accompanied by a shift toward larger, more mature spine phenotypes, nor was there a difference in the spine density. Thus, unlike in growth cones, overexpression of Tpm3.1 in spines does not increase the F-actin pool or enlarge the spine. One possibility may be that the total actin filament level in dendritic spines is fixed and the excess Tpm3.1 remains in the soluble pool.
Despite its lack of effect on morphology, it is still possible for there to be an effect on synaptic function as actin dynamics are critical to receptor localisation [7]. In cultured neurons, Tpm3.1 overexpression did not alter the mean mEPSC amplitude or frequency. Spine volume is correlated with synaptic strength [23]. We noted a difference in the distribution of the inter-event intervals between Tpm3.1-overexpressing and wild-type cells; however, as there was no difference in the mean mEPSC frequency, this difference is difficult to interpret ( Figure 5). Similar results were obtained in the brain slices. There was no apparent difference in the stimulus response function or the PPR measures of synaptic strength, and the transmitter release probability. This lack of difference in synaptic transmission is consistent with the lack of morphological differences.
Tpm3.1 is thought to have an actin-stabilising function [10], and thus overexpression may modulate the remodelling of actin in dendritic spines [24] associated with the induction of LTP. LTP is associated with a rapid increase in the spine head size to accommodate more receptors [23]. Our data suggest that Tpm3.1 does not play a key role in this process, in either the breakdown of the cytoskeleton or stabilising the new structure. There were no differences in the magnitude or time course of synaptic potentiation observed.
It is unclear why we did not detect any effects, given that studies of modulating Tpm3.1 expression in other compartments show significant changes. A possibility is that hTpm3.1 may not function normally in mouse cells, but this is unlikely given that past studies using the same model have observed effects. Furthermore, hTpm3.1 differs from mTpm3.1 by only one amino acid [21].
Previous studies have suggested there are functionally distinct populations of F-actin in dendritic spines [1,2]. These different populations are involved in maintaining the spine head size [1] and anchoring receptors to the membrane [7], and some are involved in trafficking receptors between membrane compartments [7,25]. It is possible that Tpm3.1 regulates postsynaptic F-actin populations which are involved in processes distinct from synaptogenesis and clustering of ionotropic glutamate receptors. It is also possible that the tropomyosins are not the rate-limiting step in these processes. It is known that actin filaments are already largely saturated with tropomyosin [9]. As we did not observe a change in the levels of F-actin in the postsynaptic compartment of Tpm3.1-overexpressing neurons, Tpm3.1 may also have functions beyond actin stabilisation in this compartment.
A detailed analysis of isolated PSD fractions from mouse brain tissue has revealed the presence of several tropomyosin isoforms [12], with the Tpm4 gene product Tpm4.2 being the most abundant Tpm isoform. Therefore, future studies will need to determine whether Tpm4.2 has a functionally distinct role compared to that of Tpm3.1 in regulating the spine shape and synaptic function of central nervous system neurons.
We demonstrated in our study that the Tpm3.1 transgenic model strongly overexpresses Tpm3.1 in the postsynaptic compartment. In other compartments (including the axon initial segment, growth cones, and dendritic tree), altering of Tpm3.1 expression leads to pronounced phenotypes. Contrary to this, our results surprisingly suggest the presence of a pool of filamentous actin, associated with the PSD, and defined by the decoration with Tpm3.1, which is not involved in synaptic function, as tested in our functional assays. Alternatively, the endogenous levels of Tpm3.1 in spines are already at saturation, the increased expression of exogenous Tpm3.1 remains in the soluble pool, and the association of Tpm3.1 with the PSD is independent of Tpm3.1 s association with actin. Future studies will be important in determining the subspine localisation of Tpm3.1 to provide further insight into which filament populations in dendritic spines are decorated with Tpm3.1.

Animals
All animal studies were carried out in accordance with the New South Wales Animal Research Act and Regulation and approved by the Animal Ethics Committees of UNSW Sydney and Macquarie University. Mice were housed in a temperature-controlled facility (22-24 • C) on a 12 h light-dark cycle. The original Tpm3.1 transgenic mouse line was generated on an FVB/NJ background as described in [10] and was previously backcrossed to a C57BL/6 background as described in [26]. The mice express hTpm3.1 under the β-actin promoter, resulting in ubiquitous expression of the transgene.

Quantification of Tpm3.1 Expression
Quantification of Tpm3.1 expression was performed as previously described [12]. Whole brains were removed from mice and snap frozen. The brain homogenate was prepared by homogenising with buffer (0.1 mM EDTA, 0.1 mM EGTA, 0.25 mM PMSF, and 1 mM HEPES at pH 7.4). The homogenate was centrifuged twice at 1000× g for 7 min, and some supernatant was collected for the total brain homogenate measurement. The rest was further processed to collect synaptosomes, which were isolated as described in [27]. The synaptosome fraction was then used to prepare the PSD fraction as described by [28].
Western blotting was performed as previously described [12]. Membranes were blocked in 5% milk powder or BSA in 0.1% Tween-TBS and probed with primary antibodies overnight at 4 • C. The primary antibodies used were the LC1 antibody for hTpm3.1 and the γ 9d sheep polyclonal antibody for all Tpm3.1s, which have been previously characterised in detail in [21]. Membranes were then washed with 0.1% Tween in TBS, and HRP-conjugated secondary antibodies (GE Healthcare, Sydney, Australia) were applied. Membranes were developed using Luminata Crescendo Western HRP Solution (Millipore) and imaged using the GelDoc system (BioRad). Quantification was performed by densitometry using ImageJ (version 1.47).

Cell Culture
Mouse dissociated hippocampal cultures were prepared as previously described [29]. Heterozygous transgenic mice were mated. On embryonic day 16.5, pregnant mice were sacrificed by cervical dislocation. Embryos were harvested and placed in a 6 cm dish with ice-cold 1x HBSS to induce hypothermia. A cut was made along the midline of the head, and the brain was removed using forceps. The brain was hemisected, and the meninges were peeled off. The hippocampus was isolated using microscissors. Isolated hippocampal tissue was placed in 2 mL 1× HBSS with 250 µL trypsin (Sigma-Aldrich, Sydney, Australia) at 37 • C for 20 min. An amount of 250 µL Deoxyribonuclease (DNaseI, Sigma, Sydney, Australia) was added for 30 s before being diluted with DMEM with 10% FBS. The total volume of the tissue solution was adjusted to 1 mL, and the tissue was very slowly triturated with fire-polished glass Pasteur pipettes until homogenous. The volume of the solution was adjusted up to 10 mL with DMEM with 10% FBS, and debris was left to settle in the bottom of the tube. The supernatant was centrifuged at 0.3 RCF for 7 min, and the pellet was resuspended in DMEM and 10% FBS. Cells were plated on 12 mm coverslips (#1.5 thickness glass, Menzel) coated with 100 µg/mL poly-D-Lysine (Sigma-Aldrich) at a density of 14 × 10 4 cells/mL. Media were changed to 1 mL per well of Neurobasal (Life Technologies) with 2% B27 supplement and 0.25% Glutamax (Invitrogen) 2 h after plating. Cells were maintained at 37 • C with 95% oxygen and 5% CO 2 .

Viral Production and Viral Transduction of Cultured Neurons
AAV_CAG-mRuby2, AAV_F-tractin-EGFP, and AAV_hTpm3.1-IRES-mRuby2 were packaged into the PHP.B capsid, and adeno-associated virus production was performed as previously described [30]. Briefly, 293T cells were seeded at 70-80% confluence in complete DMEM (Sigma, Sydney, Australia) with 10% FBS. At 3 h before transfection, culture medium was replaced with IMDM (Sigma, Sydney, Australia). Transfection was performed with AAV genome-containing plasmid, pFdelta6, and AAV-PHP.B using polyethyleneimine-Max (PEI-Max, Polysciences) reagent. Cell harvesting was carried out 72 h after transfection, while clarification of the supernatant was conducted with 40% PEG8000/2.5 M NaCl to a final concentration of 8% PEG8000/0.5M NaCl and incubated at 4 • C for at least 2 h, after which it was centrifuged at 2000× g for 30 min. For treatment of the combined precipitate from the clarified supernatant and cell pellet, sodium deoxycholate (0.5% final concentration) and benzonase (500 U) were used at 37 • C for 40 min. Purification of supernatants was performed using ultracentrifugation in an iodixanol gradient (475,900× g for 2 h at 18 • C). Concentrating and exchanging of AAV particles were carried out in PBS in an Amicon 100 kDa 15 mL concentrator at 5000× g at 4 • C. Genomic titering was conducted by quantitative polymerase chain reaction (qPCR), and the following titres were obtained: AAV-PHP.B_mRuby2 (2.2 × 10 14 vp/µL) and AAV-PHP.B_F-tractin-EGFP (1.44 × 10 14 vp/µL). The viral stock solution was diluted in Neurobasal media, and 8 × 10 8 viral particles per 70,000 hippocampal cells were pipetted directly into the cell culture wells at 3-5 DIV. After five minutes, stable recordings were obtained from a cell, and the patch electrode was left on for as long as possible to allow the dye to enter the cell. An image was taken of the cell for identification after immunocytochemistry for hTpm3.1. After 1-2 h of recording, the coverslip was fixed in 4% PFA for 15 min at room temperature and then washed three times with 1× PBS. mEPSCs were detected and measured using AxoGraph (Sydney, Australia). A notch filter (49.9-50.1 Hz) was applied, and an event template (maximum 0.5 ms rise time, 3 ms minimum decay time) was used to collect mEPSCs [31]. Events outside of 5-150 pA were excluded, and detected events were manually verified. Event amplitudes and inter-event intervals were measured. Either 5 min of activity or 1000 events were analysed from each cell. Out of 72 cells that were recorded, 40 were recovered. A further 11 cells were excluded because the recordings were too unstable to be able to identify mEPSCs.

Transfection
Tpm3.1 Tg mixed cultures were transfected with pmCherry-C1 (Clontech #632524). Transfection of primary cultures was performed at 14 DIV using Lipofectamine 3000 reagent. Amounts of 1.32 µg plasmid DNA and 2.56 µL Lipofectamine 3000 were used per coverslip. The required amount of Lipofectamine was mixed with half of the total required Neurobasal media in one tube. Plasmid DNA and the other half of the media were combined in a second tube. The Lipofectamine/medium was then pipetted dropwise into the DNA tube and mixed gently by inverting. It was left to incubate at room temperature in the biosafety cabinet for 25 min. All media were collected from the wells and transferred to a 15 mL tube. An amount of 50 µL of the original medium was added back to each well, along with 100 µL of transfection mixture. The cells were placed in an incubator for 60 min, along with cultured media. After the incubation period, the transfection media were removed, and the collected cultured media were returned to each well. The cells were fixed for 15 min in 4% PFA, 4 days later at 18 DIV, to be immunostained for identification of cells from Tpm3.1 transgenic embryos.

Immunocytochemistry
Unstained coverslips were stored wrapped in Parafilm at 4 • C for no more than 2 weeks before being processed. Cells were permeabilised with 0.1% Triton X100 in PBS for 5 min and blocked with 2% FBS in PBS for 30 min. They were incubated with the primary antibodies mouse LC1 [21] (at 1:200 dilution in blocking solution) and rabbit GFP antibody (1:500 dilution in blocking solution; Abcam ab290), for 2 h at room temperature or overnight at 4 • C. Cells were then washed 5 times with 1× PBS and then incubated with secondary antibodies goat anti-mouse Alexa 488 (1:500 dilution in blocking solution, Life Technologies, A32723) and goat anti-rabbit Alexa 647 (1:250 dilution in blocking solution, Life Technologies, A32733) for 30 min at room temperature, protected from light. They were then washed again with 1× PBS 5 times and dipped in Milli-Q element purified water before being mounted on glass slides using Fluoromount (Sigma-Aldrich, Sydney, Australia) mounting medium. The slides were left to set in the dark overnight. Clear nail polish was then used to seal the edges of the coverslips.

Identifying Transgenic Cells
Coverslips were viewed on an LSM 710 confocal microscope (Zeiss) using a 20 × 1.0 NA water immersion objective or a 63 × 1.4 NA oil immersion objective and an Axio-scope (Zeiss). Alexa 488 was excited with a 488 nm argon ion laser, and emission was captured at 493-581 nm. For each batch of immunostained coverslips, we imaged an area that clearly contained both hTpm3.1-positive and negative cells so they could be distinguished. To preserve consistency, imaging conditions and exposure times were kept the same across all biological replicates. Patched cells were identified by referring to images taken at the time of patching.

Dendritic Spine Morphology Analysis
Immunofluorescence images were taken using an LSM 710 confocal microscope (Zeiss) with a 63 × 1.4 NA oil immersion objective. mCherry was excited with a 561 nm DPSS laser, and emission was captured at 578-696 nm. Two 20 µm length segments of secondary dendrites 50-75 µm away from the soma were selected from each neuron. A Z-stack image with a 0.21 µm interval and 0.04 × 0.04 µm pixel size was acquired from each selected dendrite.
Deconvolution was performed on images using the DeconvolutionLab2 plugin in ImageJ [32]. Images were processed with 5 iterations of the Richardson-Lucy algorithm using a point spread function generated by the Diffraction PSF 3D plugin (Optinav). Spine morphology was analysed as previously described [33]. Z-stack images were imported to RECONSTRUCT [34]. The length of the dendrites and the length and widths of spines were manually traced and measured. Spine density, average protrusion width, and average protrusion length for each cell were calculated.
F-actin fluorescence was quantified in ImageJ (v. 2.1.0). Cells transduced with both AAV_mRuby2 and AAV_F-tractin-EGFP were selected. F-tractin has been previously confirmed as a reliable reporter of F-actin levels in cultures of primary neurons without a significant effect of the reporter on cell function [35]. Tpm3.1 Tg cells were distinguished from WT cells based on the LC1-Alexa 647 fluorescence level. Polygon selections were conducted to outline dendritic spines, and the mean fluorescence intensity in the F-tractin-EGFP channel was measured. A nearby area of the background was measured and subtracted from the spine fluorescence. To control for differences in the fluorescence level between experimental sets, F-actin fluorescence, for each set, was normalised to the mean wild-type fluorescence of that set. More than 25 spines were sampled from at least 2 segments of the dendritic tree for each neuron.
stable baseline for 20-30 min, LTP was induced with 1 burst of high-frequency stimulation (100 Hz 1 s). Responses were then recorded for the following 60 min.

fEPSP Analysis
Electrophysiological data were analysed offline using AxoGraph (Sydney, Australia). Measures included the fibre volley amplitude, the fEPSP slope, and the paired pulse ratio. Fibre volley amplitude was defined as the amplitude of the negative peak preceding the fEPSP. The fEPSP slope was defined as the maximum slope during the initial 2.5 ms after the fibre volley. This was used to measure excitatory activity. The paired pulse ratio was calculated as the second fEPSP slope divided by the first, in order to provide a measure of the presynaptic transmitter release probability [37].

Statistical Analysis
Statistical analysis was carried out using GraphPad Prism (GraphPad Software, La Jolla, CA, USA). Differences between groups were tested as indicated with Student's t-test, the Mann-Whitney test, the Kolmogorov-Smirnov test, 2-way ANOVA with Šídák's multiple comparisons test, or mixed model analysis with α = 0.05. Box-whisker plots indicate the median, quartiles, minimum, and maximum. Data are shown as mean ± standard error of the mean. Data collection and analysis were performed blind to the genotype of the animals or cells.