RhoGEF Tiam2 Regulates Glutamatergic Synaptic Transmission in Hippocampal CA1 Pyramidal Neurons

Glutamatergic synapses exhibit significant molecular diversity, but circuit-specific mechanisms that underlie synaptic regulation are not well characterized. Prior reports show that Rho-guanine nucleotide exchange factor (RhoGEF) Tiam1 regulates perforant path→dentate gyrus granule neuron synapses. In the present study, we report Tiam1's homolog Tiam2 is implicated in glutamatergic neurotransmission in CA1 pyramidal neurons. We find that Tiam2 regulates evoked excitatory glutamatergic currents via a postsynaptic mechanism mediated by the catalytic Dbl-homology domain. Overall, we present evidence for RhoGEF Tiam2's role in glutamatergic synapse function at Schaffer collateral→CA1 pyramidal neuron synapses.

Previous reports have characterized Tiam2's role in cell migration and neurite extension in vitro (Matsuo et al., 2002;Yoshizawa et al., 2002;Chan et al., 2020).We find that shRNA-mediated depletion of Tiam2 in CA1 pyramidal neurons results in significant reductions in AMPA receptor (AMPAR)-and NMDA receptor (NMDAR)-mediated evoked currents.We establish that postsynaptic Tiam2 expression contributes to glutamatergic neurotransmission and that Tiam2's effect on neurotransmission at Schaffer collateral→CA1 pyramidal neuron synapses depends on its catalytic DH1 domain.Tiam1 shares significant DH1 domain homology with Tiam2, and we show that recombinant Tiam1 expression can readily substitute for Tiam2 in supporting glutamatergic neurotransmission at this synapse.Additionally, our data also suggest no role for Tiam2 in neurotransmission at perforant path→DG granule neuron synapses.Moreover, we find that Tiam1 and Tiam2 are individually dispensable for functional long-term potentiation (LTP) at Schaffer collateral→CA1 pyramidal neuron synapses.In summary, we find that Tiam2 regulates baseline glutamatergic neurotransmission at Schaffer collateral→CA1 pyramidal neuron synapses.

Materials and Methods
Experimental constructs.A previously characterized Tiam2 shRNA target sequence against rat Tiam2 was used (5 ′ -GGAGCTGCCTTTCTCACTTTA-3; Goto et al., 2011).The Tiam2 shRNA was subcloned behind the H1 promoter region of a GFP-expressing pFHUGW expression vector.A single-nucleotide mismatch was identified in the corresponding mouse Tiam2 target sequence.However, Western blot analysis confirmed effective and specific knockdown of Tiam2 expression by the rat-directed shRNA in both rat and mouse neurons (Fig. 1B and Extended Data Fig. 4-1B,C).
All cloning was performed using overlap-extension PCR followed by In-Fusion Cloning (Clontech Laboratories).The shRNA-resistant Tiam2 was generated by introducing five silent-point mutations within the RNAi target sequence in Tiam2 cDNA (GAAGTTGTCTATCACACTTTA). Tiam1 cDNA was from Horizon Discovery (catalog #MHS6278-211691099), and Tiam2 cDNA was from GenScript Biotech (catalog #ORa14059).cDNA sequences were cloned into pCAGGS-IRES-mCherry expression vectors.The Tiam2 ΔDH mutant was generated by deletion of the ∼200-residue DH1 domain (Arg1117-Met1312) using overlap-extension PCR followed by In-Fusion Cloning (Clontech Laboratories) into a pCAGGS-IRES-mCherry expression vector.Expression constructs were coexpressed with a pFHUG vector containing GFP, which also served as a control vector for spine imaging experiments.
Tissue harvest and immunohistochemistry.All animal procedures were performed in accordance with the University of South California animal care committee's regulations.Postnatal Day (P) 15 rats were perfused with 4% paraformaldehyde via transcardial perfusion following which hippocampi were dissected and fixed in 4% PFA overnight at 4°C and then transferred into 1XPBS solution.Fixed tissue was sliced to 100-150 µm slices using a MX-TS tissue slicer (Siskiyou).Tissue sections were probed with antibodies (mouse monoclonal anti-Tiam2, 1:100, sc-514090, Santa Cruz Biotechnology; sheep anti-Tiam1, 1:100 AF5038, R&D Systems; and anti-Prox1, 1:2000, ab5475, Merck Millipore).For peptide control experiments, antibody was incubated with a blocking peptide (sc-514090 P) at 1:1 ratio overnight at 4°C and applied to samples in parallel with antibody staining.Fluorophore-coupled secondary antibodies were used for antibody detection, and slides were imaged using a Zeiss LSM-780 inverted microscope equipped with a 10×/0.45M27 objective.Tiled 16-bit images were processed using Zen (Zeiss) to produce a single image file.
Organotypic slice preparation and electrophysiology.P6 to P8 Sprague Dawley rats of both sexes were used to prepare organotypic hippocampal slice cultures as previously described (Stoppini et al., 1991;Prang et al., 2001;Bonnici and Kapfhammer, 2009).Hippocampal tissue with and without attached entorhinal cortical tissue were harvested from six to eight rat pups, and an MX-TS tissue slicer (Siskiyou) was used to prepare 400 μm transverse sections.Tissue slices were placed on squares of Biopore Membrane Filter Roll (Merck Millipore, catalog #BGCM00010) and placed on Millicell Cell Culture inserts (Merck Millipore) in 35 mm dishes containing 1 ml culture media (MEM + HEPES, Invitrogen 12360-038, 25% horse serum, 25% HBSS, and 1 mM L-glutamine).Media were replaced on alternate days.
All electrophysiology experiments were performed on an upright Olympus BX50WI microscope.Differential interference phase contrast microscopy was used to identify untransfected DG granule neurons and CA1 pyramidal neurons, while GFP-transfected cells were identified using epifluorescence microscopy.Stratum radiatum pathway and perforant pathway afferents were stimulated with a monopolar glass electrode to produce simultaneous postsynaptic currents measured from a pair of untransfected and transfected CA1 pyramidal and DG granule neurons, respectively.This approach allows for a pairwise, internally controlled comparison of the consequences of acute genetic manipulations in the transfected postsynaptic neuron (Arnold and Heintz, 1997;McAllister, 2004;Pfeiffer et al., 2010;Rao et al., 2019).Membrane voltage was held at −70 mV to measure AMPAR-evoked excitatory postsynaptic currents (eEPSCs) and at +40 mV to measure NMDAR-eEPSCs.NMDAR current amplitudes were measured at 150 ms after stimulation to avoid contamination from AMPAR current.AMPAR-and NMDAR-mediated currents were typically recorded from the same cell.Paired-pulse facilitation was induced by delivery of two consecutive stimuli at intervals of 20 ms.No more than one pair was recorded from a single hippocampal slice.Membrane-holding current, pipette series resistance, and input resistance were monitored throughout recording sessions.Data were gathered through a MultiClamp 700B amplifier (Axon Instruments), filtered at 2 kHz, and digitized at 10 kHz.
Acute slice preparation and LTP.For acute slice experiments, in utero electroporation was performed as previously described (Elias et al., 2008).P18-P24 mice were subjected to isoflurorane exposure and verified to be under a surgical plane of anesthesia.Three hundred micrometer transverse slices were cut from a freshly harvested hippocampus using a ZERO1 vibrating microtome (Ted Pella).Slices were collected in cold high-sucrose low-sodium cutting solution containing the following (in mM), 2.5 KCl, 0.5 CaCl 2 , 7 MgCl 2 , 1.25 NaH 2 PO 4 , 25 NaHCO 3 , 7 glucose, 210 sucrose, and 1.3 ascorbic acid, bubbled with 95% O 2 and 5% CO 2 .Thereafter slices were transferred to aCSF but supplemented with CaCl 2 at 2.5 mM and MgSO 4 at 1.3 mM with osmolality adjusted to 290-295 mM.Slices were incubated in aCSF saturated with 95% O 2 and 5% CO 2 at 37°C and RT for 40 min each after which slices were transferred to the recording chamber maintained in oxygenated aCSF.Borosilicate recording pipettes were filled with the same internal solution as organotypic slices.To minimize runup of baseline responses during LTP, cells were held cell attached for ∼1-2 min before breaking into the cell.GFP-transfected neurons were held at −70 mV in whole-cell configuration, while a borosilicate pipette filled with aCSF was used to stimulate Schaffer collateral afferents in the stratum radiatum.Baseline individual synaptic responses were measured for ∼2 min.Thereafter, LTP was induced within 5 min of achieving the whole-cell configuration by holding neurons at 0 mV during a 2 Hz stimulation of Schaffer collateral afferents for 90 s.
Imaging and spine analysis.Cultured hippocampal slices were transfected with pFHUGW-GFP shRNA constructs alone or pFHUGW-GFP shRNA and pCAGGS-mCherry cDNA constructs ∼18-20 h after plating using biolistic transfection.The experimenter was blinded to the genotype during subsequent processing and imaging.Slices were fixed in 4% PFA, 4% sucrose in 1× PBS, and washed with 1× PBS and then cleared with an abbreviated SeeDB-based protocol and mounted on microscope slides (Ke et al., 2013).High-resolution confocal Z-stacks of spine-containing CA1 pyramidal neuron apical dendrites were acquired on a Zeiss 510 microscope using a Plan Apochromat 40×/1.4Oil DIC objective.NMDAR-eEPSC; Wilcoxon signed-rank test).F, Paired-pulse facilitation ratio (M ± SEM) for Tiam2 shRNA expressing CA1 pyramidal neurons and paired control neurons with no detectable difference in facilitation (p = 0.88; n = 5 pairs; Student's unpaired t test).Representative scaled current traces from control and transfected (yellow) neurons (scale bars: 20 pA, 20 ms).G, H, CV analysis of AMPAR-and NMDAR-eEPSCs from pairs of control and Tiam2 shRNA expressing CA1 pyramidal neurons.CV −2 values are plotted against corresponding ratios of M amplitudes within each pair (open circles) with M ± SEM (filled circle; for AMPAR-eEPSCs, n = 10 pairs; for NMDAR-eEPSC, n = 8 pairs,).*p < 0.05, n.s., not significant.See Extended Data Figure 1-1 for more details.
Z-stacks were collected at 2,048 px × 2,048 px, 16-bit X-Y pixel dimensions, with X-Y spatial resolution of 70 nm and axial resolution of 500 nm with a 488 nm laser excitation wavelength.Automated analysis of dendritic segments and spines was performed using the commercially available software Filament Tracer (Imaris 9.1.2,Bitplane).For each cell, an ∼60 μm dendritic segment was manually selected for analysis, and thresholds for dendritic surface and spine rendering were set (10 μm dendrite length; minimum spine diameter and maximum spine length were set to 0.2 and 2 μm, respectively).Data were exported into Microsoft Excel and analyzed and graphed using R Studio (Version 1.1.153).
Lentivirus preparation.Lentivirus was generated by transfecting HEK293T cells with plasmids pVSV-G and psPAX2 along with shRNA expressing constructs using FuGENE HD transfection reagent (Roche).Viral particle containing supernatant was collected 48 h later.Supernatant was filtered at 70 µm, concentrated using PEG-it (System Biosciences).Then viral particles were pelleted via centrifugation and resuspended in DMEM.
Experimental design and data analysis.For all experiments, at least three male and three female rat pups were used.Imaging analysis was performed blind to genotype.All electrophysiological data are expressed as mean (M ) ± standard error measurement (SEM).Data from simultaneous dual whole-cell recordings are shown within scatterplots, wherein each open circle represents one paired recording and a closed circle represents the average of all paired recordings.If the average data point falls above the diagonal line, it indicates that the eEPSC is lower in the control neuron and vice versa.Paired-pulse facilitation was calculated by dividing the peak response from the second stimulus by the peak response of the first stimulus.
Statistical significance was determined using Wilcoxon signed-rank test for paired dual whole-cell patch-clamp data, Wilcoxon rank-sum test for imaging data, and Student's t test for paired-pulse facilitation data.All p values <0.05 were considered significant and denoted with a single asterisk, p values <0.01 were denoted with a double asterisk, and p values <0.001 were denoted with a triple asterisk.All error bars represent standard error measurement.Sample sizes in the present study are as reported previously (Herring and Nicoll, 2016;Incontro et al., 2018).
Coefficient of variation (CV) analysis was performed by calculating the M and standard deviation (SD) of AMPAR-eEPSCs (Kokaia et al., 1998).At least 20 consecutively recorded current amplitudes for both control and transfected cells within a paired dual whole-cell patch-clamp recording were used to obtain the CV, calculated as SD / M. Theoretical and experimental work has shown that CV −2 (M 2 / SD 2 ) is invariant with changes in the quantal size (i.e., the number of AMPA receptors at all synapses) and that CV −2 varies predictably with changes in the quantal content (i.e., the number of functional synapses containing AMPA receptors) according to the following equation CV −2 = n × P r / (1 − P r ) where n is the number of vesicle release sites and P r is the probability of presynaptic release.To observe the eEPSC variance with changes in the mean amplitude, the CV −2 values for transfected and control cells were plotted on the y-axis, and the ratios of means for transfected and control cells were plotted on the x-axis.Values above the 45°(y = x) line indicate increases in quantal content, while values approaching the horizontal line (y = 1) indicate a change in the quantal size as ultimately responsible for the difference in AMPAR-eEPSC amplitude between the control and transfected cells.For LTP, individual experiments were normalized to the baseline (before stimulation), and 12 consecutive responses were averaged to generate 1 min bins, which were then averaged to generate summary graphs.Bar graphs of LTP are averaged eEPSC values for the first 2 min (prior to LTP induction) and last 2 min of LTP summary graphs.For immunoblotting experiments, images are representative of three experimental replicates.For quantification, proteins were transferred to the same nitrocellulose membrane and probed in parallel for technical consistency.Background-adjusted band intensities were measured on Image Lab (Bio-Rad Laboratories), graphed on GraphPad Prism, and analyzed for statistical significance using the Mann-Whitney test.
To examine whether Tiam2's impact on glutamatergic neurotransmission is the result of presynaptic mechanism, we performed an analysis of paired-pulse facilitation at Schaffer collateral→CA1 pyramidal neuron synapses.We found that loss of Tiam2 protein expression in CA1 pyramidal neurons had no effect on paired-pulse facilitation (Fig. 1F), indicating presynaptic neurotransmitter release mechanisms are unaltered and Tiam2's role in supporting glutamatergic synapse function is postsynaptic.

Tiam2 knockdown reduces spine density in CA1 pyramidal neurons
Next, we determined the synaptic alteration underlying reductions in AMPAR-and NMDAR-eEPSC amplitudes following Tiam2 knockdown.These alterations could arise from uniform reduction in postsynaptic AMPARs and NMDARs across all glutamatergic synapses or loss of a subset of functional glutamatergic synapses in CA1 pyramidal neurons.We performed CV −2 analysis, with a comparison of the normalized variance of eEPSC amplitudes from two neurons receiving the same stimulus to identify changes in the quantal size and/or quantal content (Del Castillo and Katz, 1954;Bekkers and Stevens, 1990;Malinow and Tsien, 1990;Gray et al., 2011;Levy et al., 2015).Data points that lie on the horizontal dotted line indicate changes in the quantal size and represent uniform changes in receptor function across existing synapses.Data points that fall on or near the diagonal dotted line indicate changes in the quantal content and represent changes in the number of functional glutamatergic synapses (Bekkers and Stevens, 1990;Malinow and Tsien, 1990).The averages for both AMPAR and NMDAR excitatory currents fell on or near the diagonal dotted line (Fig. 1G,H), showing that deficits in current amplitudes are explained by changes in quantal content.Thus, reductions in AMPAR-and NMDAR-eEPSC amplitude following Tiam2 knockdown result from a reduced number of functional glutamatergic synapses.
Tiam2 catalyzes small-GTPase Rac1 activity, which promotes actin-dependent formation of dendritic spines (Tashiro and Yuste, 2004;Goto et al., 2011).Similar reductions in AMPAR-and NMDAR-eEPSC amplitudes and CV −2 analysis suggest an underlying reduction in dendritic spine density.We compared Tiam2 shRNA-and GFP-transfected CA1 pyramidal neurons (Fig. 2A) and observed a dramatic loss in the spine number relative to control neurons (Fig. 2B,C).Spine length and other parameters remained unchanged (Fig. 2C; Extended Data Fig. 2-1A-D).We verified that reduced AMPAR and NMDAR function produced by Tiam2 knockdown could be rescued by recombinant Tiam2 expression.To do this we coexpressed Tiam2 shRNA and a recombinant shRNA-resistant Tiam2 cDNA in CA1 pyramidal neurons and recorded AMPAR-and NMDAR-eEPSCs.We found that AMPAR-and NMDAR-eEPSC amplitudes were similar between control and transfected neurons, demonstrating Tiam2 cDNA was sufficient to rescue Tiam2 knockdown (Fig. 2D,E).The expression of Tiam2 cDNA was also sufficient to rescue deficits in spine density in CA1 pyramidal neuron dendrites (Fig. 2F-H; Extended Data Fig. 2-1E-H).These results further demonstrate that Tiam2 shRNA is specific for endogenous Tiam2, and the loss-of-function phenotypes we observe are a result of targeted reduction in Tiam2 protein expression.Moreover, the reduction in dendritic spine density corroborates our previous functional analyses and is consistent with Tiam2's role in supporting glutamatergic synapse structure and function in CA1 pyramidal neurons.
Tiam2 has significant expression in DG granule neurons (Fig. 1A).Surprisingly, we found that Tiam2 depletion in DG granule neurons had no effect on AMPAR and NMDAR current amplitude (Fig. 3F,G), with no alterations in spine morphology (Extended Data Fig. 3-1A-C).This suggests a DH1-dependent role of Tiam2 at Schaffer collateral→CA1 pyramidal neurons and no role for Tiam2 in neurotransmission at perforant path→DG granule neuron synapses.
Having established Tiam2 is necessary for normal baseline neurotransmission at CA1 pyramidal neuron synapses; we tested whether Tiam2 affects mechanisms of synaptic plasticity.In acute mouse hippocampal slices where Tiam2 shRNA was delivered via in utero electroporation, we found Tiam2 depletion has no effect on the induction or maintenance of functional LTP at Schaffer collateral-CA1 pyramidal neurons (Fig. 4D).It is possible this lack of effect was due to a compensatory increase in Tiam1, but shRNA-mediated depletion does not affect expression of the either homolog in vitro (Extended Data Fig. 4-1B,C).Similarly, we found that Tiam1 knockdown produced no significant effect on LTP induction or maintenance in CA1 pyramidal neurons (Fig. 4E), suggesting that neither Tiam2 nor Tiam1 individually plays an essential role in this form of synaptic plasticity at Schaffer collateral→CA1 pyramidal synapses.

Discussion
Loss-of-function studies describe the contribution of differential gene expression towards input-specific regulation of synaptic transmission and plasticity (Arnold and Heintz, 1997;Elias et al., 2008;Westerink et al., 2012).Here, we found the RhoGEF Tiam2 is essential for normal glutamatergic neurotransmission at Schaffer collateral→CA1 pyramidal neuron  synapses.We show Tiam2 depletion in CA1 pyramidal neurons produces diminished synaptic currents conducted by postsynaptic glutamate receptors and that this depletion depends on the catalytic DH1 domain.Furthermore, Tiam2 depletion does not affect presynaptic glutamate release and reduces dendritic spine density, indicating that Tiam2 regulation of synapses is restricted to the postsynaptic compartment.Prior evidence indicates endogenous Tiam1 depletion does not affect baseline neurotransmission at Schaffer collateral→CA1 pyramidal neuron synapses (Rao et al., 2019).While we found recombinant Tiam1 rescues deficits from Tiam2 depletion at these synapses, differential ligand binding within other domains, PDZ or Pleckstrin homology coiled-coil extension (Zheng et al., 1996;Shepherd et al., 2011;Rao et al., 2019), may confer functional differences at other synapses.Tiam2 expression is evident within other hippocampal subregions.Surprisingly, Tiam2 depletion at perforant path-DG granule neurons produced no alterations in synaptic transmission.Overall, our analyses are consistent with necessary role of Tiam2 in the regulation of glutamatergic synapses in CA1 pyramidal neurons.
Several studies show Tiam1 is highly expressed in DG granule neurons (Ehler et al., 1997;Zhao et al., 2001;Lein et al., 2004Lein et al., , 2007;;Rao et al., 2019;Cheng et al., 2021).This is consistent with our immunostaining of hippocampal slices (Extended Data Fig. 4-1A); however there are reports of Tiam1 expression in CA1 as well (Sasaki et al., 2010).We cannot rule out that Tiam1 plays a functional role in CA1 pyramidal neurons, but we find Tiam1 and Tiam2 are dispensable for functional LTP at Schaffer collateral→CA1 pyramidal neuron synapses.RhoGEFs Trio and Kalirin are required for the expression of LTP in CA1 pyramidal neurons (Herring and Nicoll, 2016).Since our experimental paradigm produces a depletion and does not eliminate protein expression, it is entirely possible that low levels of Tiam1 or Tiam2 protein, together with Trio and Kalirin, contribute to LTP in CA1 pyramidal neurons.Recent studies report a Tiam1-mediated deficit in structural synaptic plasticity in CA1 pyramidal neurons (Kojima et al., 2019;Saneyoshi et al., 2019), and structural and functional LTP are well correlated (Matsuzaki et al., 2004).Our data support a dispensable role for Tiam1 and Tiam2 in LTP.This discrepancy may be explained by a partial contribution to plasticity by Tiam1, a possible cell autonomous role, or a role for Tiam1 at entorhinal→CA1 pyramidal neuron synapses.
Members of the RhoGEF protein family are implicated in glutamatergic synaptic deficits associated with neurodevelopmental disorders (Volk et al., 2015;Katrancha et al., 2017), and recent exome-sequencing studies have identified autism and schizophrenia-associated de novo missense mutations in Tiam1 and Tiam2 (Need et al., 2012;Takata et al., 2013;S. Lu et al., 2022).We show that Tiam2 is critical for glutamatergic synapse function at Schaffer collateral-CA1 pyramidal neurons.Additional RhoGEF proteins, such as Kalirin, Trio, and β-PIX, have been shown to have similar roles in supporting glutamatergic synapses in these neurons (Herring and Nicoll, 2016).This, along with the data from this study, suggests that Tiam2, Trio, and Kalirin form a cohort of RhoGEFs that are together responsible for normal glutamatergic neurotransmission at CA1 pyramidal neuron synapses.An improved understanding of their role in synapse development and function could contribute to therapeutics for neurodevelopmental disorders.