Editor's evaluation: SUMOylation of NaV1.2 channels regulates the velocity of backpropagating action potentials in cortical pyramidal neurons

Abstract

Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Voltage-gated sodium channels located in axon initial segments (AIS) trigger action potentials (AP) and play pivotal roles in the excitability of cortical pyramidal neurons. The differential electrophysiological properties and distributions of NaV1.2 and NaV1.6 channels lead to distinct contributions to AP initiation and propagation. While NaV1.6 at the distal AIS promotes AP initiation and forward propagation, NaV1.2 at the proximal AIS promotes the backpropagation of APs to the soma. Here, we show the small ubiquitin-like modifier (SUMO) pathway modulates Na+ channels at the AIS to increase neuronal gain and the speed of backpropagation. Since SUMO does not affect NaV1.6, these effects were attributed to SUMOylation of NaV1.2. Moreover, SUMO effects were absent in a mouse engineered to express NaV1.2-Lys38Gln channels that lack the site for SUMO linkage. Thus, SUMOylation of NaV1.2 exclusively controls INaP generation and AP backpropagation, thereby playing a prominent role in synaptic integration and plasticity. Editor's evaluation This fundamental study describes how the specific SUMOylation of Nav1.2 channels regulates neuronal function by slowing action potential backpropagation from the AIS. This compelling evidence breaks new ground in the role of SUMOylation in modulating synaptic plasticity and will be of interest to neuroscientists working on synaptic transmission and modulation of ion channel activity. https://doi.org/10.7554/eLife.81463.sa0 Decision letter Reviews on Sciety eLife's review process Introduction In cortical pyramidal cells, as in many central nervous system (CNS) neurons, action potentials (APs) initiate in the axon initial segment (AIS), the proximal part of the axon where the neuronal membrane is not covered with a myelin sheath. The AIS is characterized by a specialized assembly of scaffolding proteins and voltage-gated channels with distinctive biophysical properties (Bean, 2007; Rasband, 2010). Classically, APs propagate forward from the AIS into the axonal arbor, triggering neurotransmitter release from presynaptic terminals. APs can also propagate backward into the dendrites of cortical pyramidal cells, where they are proposed to play a role in synaptic plasticity by regulating synaptic strength and the coordination of synaptic inputs (Stuart and Sakmann, 1994; Markram et al., 1997). Both initiation and propagation of APs are critically dependent on the distribution and properties of voltage-gated Na+ (NaV) channels in specific neuronal compartments (Stuart et al., 1997; Hu et al., 2009; Baranauskas et al., 2013). Therefore, identifying signaling pathways that regulate the biophysical properties of neuronal NaV channels is key to understanding spike generation, propagation, and integration in cortical circuits (Cantrell et al., 1996; Cantrell and Catterall, 2001; Bender et al., 2012; Kole and Stuart, 2012; Yin et al., 2017). Central neurons express, to varying degrees, three primary NaV channels α-subunit isoforms. NaV1.1, NaV1.2, and NaV1.6 are found in mature neurons, while NaV1.3 channels are also found in the developing nervous system (Goldin et al., 2000). Each NaV isoform has a distinct spatiotemporal distribution and is subject to the activity of specific signaling pathways that regulate the biophysical properties and trafficking behavior of the channel. For example, in mature pyramidal neurons, the AP trigger zone, which is located in the distal AIS, contains almost exclusively NaV1.6 channels (Lorincz and Nusser, 2008; Hu et al., 2009; Lorincz and Nusser, 2010; Tian et al., 2014). These channels are also present in the nodes of Ranvier (Caldwell et al., 2000). In contrast, the proximal portion of the AIS, soma, and dendrites is believed to contain mostly NaV1.2 channels (Hu et al., 2009; Grubb et al., 2011). While this difference in the distribution has long made it tempting to posit that the initiation and forward propagation of APs are predominantly dependent on NaV1.6 channels and the backpropagation of APs is dependent on the activity of NaV1.2 channels, studies of the biophysical attributes of the two channels have not previously verified this hypothesis. While heterologous studies show that NaV1.6 channels activate at more negative voltages than other neuronal NaV isoforms and have a higher propensity to generate non-inactivating currents, differences in the gating behavior of NaV1.2 and NaV1.6 channels appear to be subtle within their native neuronal milieu (Smith et al., 1998; Zhou and Goldin, 2004; Rush et al., 2005; Chen et al., 2008). Indeed, using knockout mice, we found that NaV1.6 channels were not required to determine the initiation site for APs within the AIS, for backpropagation into the dendrites, or for the lower activation threshold voltage for APs that is commonly observed in pyramidal neurons (Katz et al., 2018). Although the biophysical differences between native NaV1.2 and NaV1.6 are subtle, each channel isoform is differentially regulated by neuromodulators. Thus, NaV1.6 channels are less sensitive to inhibition by cAMP-dependent protein kinase or protein kinase C-mediated phosphorylation (Chen et al., 2008). This finding explains divergent regulation of NaV channel subtypes following the activation of D1/D5 dopamine or 5-HT1A serotonergic receptors (Maurice et al., 2001; Yin et al., 2017). We have shown that ion channels are also subject to post-translational regulation by covalent linkage of small ubiquitin-like modifier (SUMO) proteins (Rajan et al., 2005; Plant et al., 2010; Plant et al., 2011; Plant et al., 2012; Plant et al., 2016; Xiong et al., 2017; Plant et al., 2020). Three SUMO isoforms (SUMO1–3) are operative in central neurons and can modulate the gating of specific channels following their conjugation to the ε-amino group of specific lysine residues on the intracellular termini or cytoplasmic loops of the channel subunits. Further, we found the enzymes required to activate, mature, and conjugate SUMO reside in the plasma membrane of Xenopus oocytes, tissue culture cells, and neurons (Rajan et al., 2005; Plant et al., 2010; Plant et al., 2011; Plant et al., 2016). Although SUMOylation is a covalent post-translational modification, it is a dynamic process subject to rapid reversal by the action of the SENP family of sentrin-specific cysteine proteases. Thus far, we have observed that SUMOylation increases excitability either by decreasing potassium flux through KV and K2P channels or by increasing the activity of NaV channels and that the opposite functional effects are mediated by the activity of SENPs. Further, we have observed that SUMOylation status varies among channel types at baseline and in response to environmental stimuli such as hypoxia (Plant et al., 2016; Xiong et al., 2017; Plant et al., 2020), and particularly germane to this study, that SUMO has differential effects on the principal NaV channel isoforms expressed in central neurons. Specifically, we found that SUMOylation modulates the voltage-dependent gating of NaV1.2 channels via linking to lysine 38, but SUMO does not interact with NaV1.6 (Plant et al., 2016). Here, we tested the hypothesis that SUMOylation regulates the excitability of cortical pyramidal neurons. By combining whole-cell recordings with high-speed fluorescence imaging to simultaneously monitor Na+ flux in different subcellular compartments of L5 cortical neurons, we found that SUMO1 increases excitability via a synergistic effect on subthreshold K+ and Na+ conductances. Thus, SUMOylation suppresses the open probability of K+ channels while concurrently increasing Na+ influx via a leftward shift in the steady-state activation of subthreshold persistent Na+ currents. These effects are absent in a CRISPR-generated mouse that constitutively expresses NaV1.2-Lys38Gln channels, a channel variant that cannot be SUMOylated. Confirming the long-held notion for their roles in the AIS based on distribution, and consistent with our previous report that SUMO regulates NaV1.2 but not NaV1.6 channels, we demonstrate that SUMOylation of NaV1.2 regulates the velocity of backpropagation in cortical pyramidal neurons independent of the speed at which AP propagate forward from the AIS. Results SUMO1 increases the excitability of layer 5 cortical neurons Previously, we showed that the effects of SUMOylation and deSUMOylation of neuronal ion channels underlying IDR, IKso, and INa could be assessed by including purified SUMO1 or SENP1 polypeptides, respectively, in the recording pipette solution (Plant et al., 2011; Plant et al., 2012; Plant et al., 2016). To characterize the effects of the SUMO pathway on the excitability of layer 5 cortical pyramidal neurons, we made whole-cell, current-clamp recordings from these cells using pipettes filled with a solution containing SUMO1 or SENP1 polypeptides at 1000 and 250 pmol/l, respectively. We have previously shown that polypeptides at these concentrations produce maximal effects on KV, K2P, and NaV channels in cultured rat hippocampal neurons, cerebellar granule neurons, human ventricular cardiomyocytes derived from iPS cells, and on channels expressed in heterologous cell systems (Plant et al., 2010; Plant et al., 2011; Plant et al., 2012; Plant et al., 2016; Plant et al., 2020). Passive neuronal properties and repetitive firing characteristics were assessed by examining the voltage responses to a series of prolonged hyperpolarizing and depolarizing current pulses delivered via the somatic pipette. Because cortical cells are geometrically complex, we compared data obtained 2 min and 35 min after breakthrough into whole-cell configuration to account for slow intracellular dialysis of the polypeptide into the neurons. As SUMO1 diffused into the cell, the frequency of spike firing in response to a given depolarizing suprathreshold current pulse increased (Figure 1a). Thus, the mean instantaneous firing frequency in response to a 0.3 nA current injection increased from 17.5 ± 3.6 Hz immediately after the break-in to 29.0 ± 5.4 Hz (n = 8, p=0.036, paired t-test) after 35 min of SUMO1 dialysis. In contrast, dialysis with the SENP1-containing solution caused a gradual decrease in the frequency of repetitive firing over time. The mean instantaneous firing frequency in response to a 0.3 nA current pulse decreased from 20.5 ± 3.4 Hz immediately after the break-in to 7.3 ± 3.5 Hz (n = 8, p=0.002, paired t-test) after 35 min of SENP1 dialysis. Figure 1 Download asset Open asset SUMOylation and deSUMOylation have opposing effects on the excitability of layer 5 pyramidal neurons. (a) Current-clamp, whole-cell recordings from L5 neurons 2 min after the break-in to whole-cell mode (black) and 35 min later demonstrate time-dependent effects of SUMO1 (blue) or SENP1 (green) dialysis on firing frequency. Voltage responses were elicited by injecting 400-ms-long current pulses, which started at –0.15 nA and incremented by 50 pA. (b) SUMO1 and SENP1 have opposite effects on passive membrane properties. Voltage responses to a small hyperpolarizing current pulse injection immediately after the break-in (black) and following SUMO1 (blue) or SENP1 (green) dialysis via the whole-cell pipette. Red dashed lines are the best exponential fits of the voltage responses. Notice that the amplitude of voltage deflection and the membrane time constant were enhanced by SUMO1 and decreased by SENP1 dialysis. (c) Apparent input resistance (Rin) increases in SUMO1 dialyzed neurons, whereas it decreases in SENP1 dialyzed cells. The lines connect the paired Rin values obtained from the same individual neuron at 2 min and 35 min of recording with SUMO1 (blue), SENP1 (green), and control solution-filled pipette (black). Box plots represent the 25–75% interquartile range of values obtained from neurons dialyzed with SUMO1 (n = 8), SENP1 (n = 6), and control (n = 11) solution; the whiskers expand to the 5–95% range. A horizontal line inside the box represents the median of the distribution, and the mean is represented by a cross symbol (X). p-Values were calculated using Student’s t-test for paired data. Examining the voltage responses to small hyperpolarizing current pulses before and following the SUMO1 and SENP1 dialysis revealed that the polypeptides elicited opposite effects on passive neuronal properties (Figure 1b). Thus, the apparent input resistance (Rin), calculated as a ratio of the steady-state amplitude of the voltage deflection to current amplitude, gradually increased when SUMO1 was included in the pipette, from 96.1 ± 12.7 MΩ at a time of break-in to the cell to 122.5 ± 13.7 MΩ (n = 8, p<0.002) at 35 min of recording (Figure 1c). In contrast, dialysis of the neurons with SENP1 caused Rin to decrease as a function of recording time from 130.0 ± 24.9 MΩ to 80.4 ± 14.6 MΩ (n = 6, p<0.01). In parallel, the membrane time constant (τm) obtained by fitting a monoexponential function to the voltage transient following the end of the hyperpolarizing current pulse was increased by SUMO1 application from 16.8 ± 1.9 ms at the time of break-in to 22.8 ± 2.7 ms (n = 8, p<0.01) and shortened by SENP1 application from 17.7 ± 0.7 ms to 11.2 ± 1.3 ms (n = 8, p<0.001). Recording of similar duration with control intracellular solution had no significant effect on Rin (113.7 ± 9.3 vs. 112.8 ± 8.9 MΩ, n = 11, p=0.77) and τm (24.3 ± 2.2 vs. 21.2 ± 1.9 ms, n = 11, p=0.09). These findings indicate that in L5 cortical neurons the SUMO pathway regulates potassium channels that determine the passive membrane properties. Furthermore, the relatively high effectiveness of SENP1 suggests that in L5 neurons, as in other cell types (Rajan et al., 2005; Plant et al., 2011; Plant et al., 2012; Plant et al., 2016; Xiong et al., 2017), a significant fraction of these channels are SUMOylated under control conditions. The effect of SUMO1 and SENP1 on repetitive firing may reflect the action of the polypeptides on passive neuronal characteristics or their influence on the ion currents underlying spike generation. Theoretical analysis revealed that while the former mechanism should elicit a parallel shift of the frequency–current (F-I) curve to the right or left along the current axis (Chance et al., 2002), the latter should alter the neuronal gain, that is, the steepness of the slope of the F-I characteristic. Comparing the linear fits of the mean F-I curves obtained immediately after the break-in and following the SUMO1 dialysis, we found that the curve steepness increased by ~75%, from 64 to 111 Hz/nA (n = 8) (Figure 2). Dialysis with control pipette solution had little to no effect on the neuronal gain (62 vs. 58 Hz/nA, respectively, n = 14). In contrast, in recordings with SENP1 containing pipette, the gain decreased from 88 to 44 Hz/nA (n = 8). Figure 2 Download asset Open asset The effects of SUMO1 and SENP1 on input/output gain. The frequency–current (F-I) characteristic of L5 pyramidal neurons, constructed by plotting the mean instantaneous spike frequency as a function of depolarizing current pulse amplitude, obtained immediately after the break-in (open black circles) and following 35 min of recording with SUMO1 (n = 8, blue), SENP1 (n = 8, green) and control solution (n = 14, closed black circles) containing pipette. Notice that the F-I curve was shifted to the left and became steeper in SUMO1 dialyzed neurons, whereas in SENP1 dialyzed cells the F-I characteristics were displaced to the right and its slope (dashed line) decreased. The F-I curve showed no significant change in control recordings. To test our hypothesis that SUMOylation of NaV1.2 channels can regulate the excitability of L5 cortical neurons, we used CRISPR/Cas9 to engineer a mouse model carrying NaV1.2-Lys38Gln, a mutation that removes the only SUMO-conjugation site in NaV1.2 channels (Plant et al., 2016). The genotype of the mice was verified by PCR screening and sequencing analysis (Figure 3—figure supplement 1). Comparison of passive and active electrophysiological characteristics of WT and NaV1.2-Lys38Gln mutant layer 5 pyramidal neurons revealed no significant difference (Table 1), indicating that the functional consequences of the mutation are largely compensated. Using whole-cell current-clamp recordings from L5 neurons from the NaV1.2-Lys38Gln mutant mice, we first sought to find out whether SUMO1 and SENP1 dialysis affect the F-I relationship (Figure 3a). As in WT neurons, SUMO1 dialysis enhanced the frequency of repetitive firing for a given amplitude of the current pulse, whereas the SENP1 dialysis had the opposite effect. Thus, the mean instantaneous firing frequency in response to a 0.3 nA current injection increased from 18.7 ± 2.6 Hz immediately after the break-in to 24.0 ± 2.4 Hz (n = 6, p=0.013, paired t-test) after 35 min of SUMO1 dialysis. In contrast, the mean instantaneous firing frequency decreased from 13.1 ± 1.9 Hz immediately after the break-in to 2.9 ± 1.7 Hz (n = 7, p=0.0005, paired t-test) after 35 min of SENP1 dialysis. Figure 3 with 1 supplement see all Download asset Open asset In L5 Nav1.2-Lys38Gln mutant neurons, SUMO1 and SENP1 do not affect the gain of the input–output curve. (a) Current-clamp, whole-cell recordings from L5 Nav1.2-Lys38Gln mutant neurons immediately after the break-in (black) and following SUMO1 (blue) or SENP1 (green) dialysis. Voltage responses were elicited by injecting 400-ms-long current pulses, which started at –0.15 nA and incremented by 50 pA. (b) The F-I characteristic of Nav1.2-Lys38Gln mutant neurons obtained immediately after the break-in (black open circles) and following SUMO1 (n = 6, blue) or SENP1 (n = 7, green) dialysis via the whole-cell pipette. Notice the opposite effects of SUMO1 and SENP1 on the position of the F-I curve over the current axis. Both treatments had little to no effect on the slope of the F-I curve. (c) The Rin increased over time in SUMO1 dialyzed Nav1.2-Lys38Gln mutant neurons, whereas it decreased in SENP1 dialyzed cells. The lines connect the paired Rin values obtained from the same individual neuron at 2 min and 35 min of recording with SUMO1 (blue), SENP1 (green). (c) Apparent input resistance (Rin) increases in SUMO1 dialyzed neurons, whereas it decreases in SENP1 dialyzed cells. The lines connect the paired Rin values obtained from the same individual neuron at 2 min and 35 min of recording with SUMO1 (blue) and SENP1 (green) containing solution. Box plots represent the 25–75% interquartile range of values obtained from neurons dialyzed with SUMO1 (n = 6) and SENP1 (n = 7) solution; the whiskers expand to the 5–95% range. A horizontal line inside the box represents the median of the distribution, and the mean is represented by a cross symbol (X). p-Values were calculated using Student’s t-test for paired data. Table 1 Comparison of electrophysiological characteristics of WT and NaV1.2-Lys38Gln mutant layer 5 pyramidal neurons. ParameterWild typeNaV1.2-Lys38Gln mutantDifferenceInput resistance (MΩ)127.4 ± 11.5(n = 28)134.9 ± 13.6(n = 13)NS, p=0.701Membrane time constant, τm (ms)20.4 ± 1.3(n = 30)24.0 ± 2.3(n = 13)NS, p=0.149Voltage threshold (mV)*–57.4 ± 1.4(n = 17)–57.5 ± 0.6(n = 13)NS, p=0.956Current threshold (pA)†448 ± 30(n = 18)514 ± 32(n = 12)NS, p=0.151AP peak (mV)+36.5 ± 1.4(n = 17)+37.1 ± 1.3(n = 13)NS, p=0.729AP dV/dtmax (V/s)268 ± 25(n = 17)271 ± 20(n = 13)NS, p=0.937AP half-width (ms)1.19 ± 0.13(n = 17)1.06 ± 0.10(n = 12)NS, p=0.486F-I characteristics slope (Hz/nA)91.0 ± 4.0(n = 30)94.5 ± 7.0(n = 13)NS, p=0.653 Data are presented as mean ± SE; WT and mutant neurons are compared using the Student’s t-test for unpaired data. * Voltages were corrected for liquid junction potential of –13 mV (recording temperature of 32°C). Data were collected within 2 min after breaking into the whole-cell configuration. † The current threshold was defined as the minimum amplitude of a 10-ms-long current step that elicited an AP. Both treatments, however, affected the position of the F-I curve relative to the current axis while little to no effect on its slope was observed (Figure 3b), consistent with the hypothesis that, in NaV1.2-Lys38Gln neurons, SUMOylation primarily affects passive neuronal properties. Indeed, in NaV1.2-Lys38Gln mutant neurons, SUMO1 dialysis increased the apparent Rin (from 144.2 ± 19.8 MΩ to 171.6 ± 21.1 MΩ, n = 6, p<0.005) whereas dialysis with SENP1 had an opposite effect (from 126.8 ± 19.6 MΩ to 79.8 ± 8.6 MΩ, n = 7, p<0.01) (Figure 3c). SUMO1 and SENP1 have the opposite effect on the voltage dependence of INaP In cortical pyramidal neurons, the persistent sodium current operates at a subthreshold range of voltages and is one of the main factors influencing the frequency of repetitive firing, thereby modifying the neuronal gain (Stuart and Sakmann, 1995; Astman et al., 2006). We have recently shown that in pyramidal cells, most of the whole cell INaP is generated by somatodendritic Na+ channels (Fleidervish et al., 2010; Shvartsman et al., 2021). However, because the steady-state activation curve of the AIS channels is shifted to the left by 7–9 mV, most of INaP at functionally critical subthreshold voltages is axonal. The immunohistochemical evidence indicates that soma, dendrites, and proximal AIS of L5 pyramidal neurons are populated predominately by the NaV1.2 channels whose activation and inactivation gating is sensitive to SUMOylation (Hu et al., 2009; Grubb et al., 2011; Plant et al., 2016; Liu et al., 2022). In contrast, the distal AIS membrane and the Ranvier nodes contain NaV1.6 channels, which are not subject to SUMOylation (Plant et al., 2016). To find out how SUMO1 and SENP1 affect the persistent sodium current in different neuronal compartments, we combined whole-cell, voltage-clamp recordings from L5 neurons with high-speed fluorescence imaging of a Na+ sensitive dye, SBFI. A comparison of the voltage ramp-elicited Na+ fluxes revealed that SUMO1 dialysis induces a left shift in the voltage dependence of INaP activation in soma, proximal apical dendrite, and in the AIS of L5 neurons (Figure 4a). Thus, at a voltage of –50 mV, the relatively small fluorescence change in the soma and apical dendrites was significantly increased by SUMO1 dialysis, whereas the amplitude of the Na+ signal in the AIS was less markedly increased (Figure 4b). Figure 4 with 2 supplements see all Download asset Open asset SUMO1 causes a leftward shift of INaP voltage dependence in pyramidal cells from wild-type but not from Nav1.2-Lys38Gln mutant mice. (a) Left: WT L5 pyramidal neuron filled with SBFI-containing, Cs+-based solution via a somatic patch pipette, as seen during the fluorescence imaging experiment with a NeuroCCD-SMQ camera. Right: INaP and normalized somatic (black), axonal (red), and dendritic (cyan) ΔF transients elicited by 2-s-long voltage ramp from –70 mV to 0 mV immediately after the break-in and following 10 min of dialysis with SUMO1. Notice the leftward shift in voltage dependence of INaP activation in soma, dendrite, and to a lesser extent, in axon initial segments (AIS). Capacitive and leakage currents were not subtracted. (b) Pseudocolor maps of the ramp elicited ΔF changes between the times marked by the arrowheads in (a). Top: voltage ramp from −70 to −50 mV produced Na+ elevation mostly in the AIS. Bottom: following the SUMO1 dialysis, voltage ramp from −70 to –50 mV elicited large Na+ signals also in the soma and dendrites. Measurements of half-activation voltage (V½) revealed that SUMO1 dialysis causes a significant leftward shift in the voltage dependence of activation of both somatic and axonal channels in WT neurons (Figure 4—figure supplement 1). However, the application of SUMO1 produced no effect on the voltage dependence of INaP in neurons from NaV1.2-Lys38Gln mice (Figure 4—figure supplement 2). Intracellular application of SENP1 resulted in an opposite effect on the voltage dependence of INaP. Thus, a small but significant rightward shift in the V½ of INaP was observed in the soma and AIS of neurons from WT but not NaV1.2-Lys38Gln mice (Figure 5). These findings indicate that in cortical neurons a portion of the NaV1.2 channels is SUMOylated under control conditions. Figure 5 Download asset Open asset SENP1 causes a rightward shift of activation kinetics of INaP in pyramidal cells from wild-type but not from Nav1.2-Lys38Gln mutant mice. Box plots represent the 25–75% interquartile range of INaP V1/2 activation values in the soma (black), and axon initial segments (AIS) (red) of WT (top, n = 9) and Nav1.2-Lys38Gln mutant (bottom, n = 7) pyramidal neurons immediately after the break-in and following 10 min of dialysis with SENP1; the whiskers expand to the 5–95% range. A horizontal line inside the box represents the median of the distribution, and the mean is represented by a cross symbol (X). p-Values were calculated using Student’s t-test for paired data. The lines connect the paired V1/2 values obtained from the same individual neuron after the break-in and following 10 min of dialysis with SENP1. V1/2 difference plots show a change in INaP half-activation voltage elicited by SENP1 in individual neurons (open circles); dashed lines show mean V1/2 values in WT (n = 9) and Nav1.2-Lys38Gln mutant (n = 7) cells. SUMOylation of Na+ channels affects voltage-dependent amplification of EPSPs in pyramidal neurons Changes in the amplitude of INaP at subthreshold voltages are expected to influence the spatial and temporal summation of synaptic potentials (Deisz et al., 1991; Stuart and Sakmann, 1995; Stuart, 1999). Therefore, we studied the effect of SUMOylation on the amplitude and duration of excitatory postsynaptic potentials (EPSPs) elicited in the pyramidal neuron by brief synaptic stimuli. The EPSPs were measured immediately after break-in to the whole-cell configuration and following 30 min of intracellular dialysis with SUMO1 in WT and NaV1.2-Lys38Gln neurons. SUMO1 did not change the duration of small EPSPs of less than 10 mV in amplitude (Figure 6a). In contrast, SUMO1 prolonged the decay time constant of EPSPs greater than 10 mV in amplitude in WT but not NaV1.2-Lys38Gln neurons. In pooled EPSPs obtained from six neurons in each experimental group, SUMO1 dialysis enhanced the steepness of the slope of EPSP integral-to-peak relationship (Figure 6b) in WT neurons, whereas SUMOylation had no effect on this relationship for NaV1.2-Lys38Gln cells. Figure 6 Download asset Open asset Effect of SUMO1 on voltage-dependent amplification of EPSPs in pyramidal neurons from wild-type and Nav1.2-Lys38Gln mutant mice. (a) Comparison of small and large EPSPs evoked in WT and Nav1.2-Lys38Gln mutant pyramidal neurons immediately after the break-in (black) and following the SUMO1 dialysis (blue). Notice the slower decay time constant of larger EPSP following SUMO1 dialysis in WT neuron. (b) The mean EPSP integral as a function of peak EPSP amplitude after the brake-in (black) and following the SUMO1 dialysis (blue) of the WT (n = 6) and Nav1.2-Lys38Gln mutant (n = 6) pyramidal neurons. Notice the amplification of larger EPSPs in SUMO1 dialyzed WT cells. SUMOylation differentially affects the speed of forward- and back-propagating action potentials In cortical pyramidal neurons, the NaV1.2 channels are predominantly localized in somatic, dendritic, and proximal AIS membrane, where they are responsible for the propagation of action potentials back into the dendritic tree (Hu et al., 2009; Grubb et al., 2011). The NaV1.6 channel subtype is present in the distal AIS and in the nodes of Ranvier, and it is responsible for the forward propagation of action potentials into the axonal arbor (Hu et al., 2009). Because NaV1.2 and NaV1.6 channels respond differentially to SUMOylation, with the former being susceptible and the latter resistant to SUMO1, we hypothesized that this neuromodulation could differentially affect the speed of forward and backpropagation of the spikes. Seeking to test this hypothesis directly, we measured the velocity of forward and backpropagation using paired, whole-cell, loose patch recordings to detect the times of the spike arrival from multiple sites along the axo-somatic axis in sequence (Figure 7; Baranauskas et al., 2013; Lezmy et al., 2017). In order to distinguish the axon from other thin processes emerging from the cell body and facilitate the distance measurements between the somatic and axonal pipettes, we filled the neurons for at least 15 min with the Na+-sensitive dye SBFI. Because of this and the relatively long time it takes to obtain action currents from multiple axonal locations, we were not able to measure the propagation velocity upon break-in to whole-cell configuration. Thus, we compared the propagation velocities in WT neurons dialyzed with control or SUMO1-containing intracellular solution. As an additional control, the same measurements were taken from the NaV1.2-Lys38Gln neurons dialyzed with SUMO1. Figure 7 with 1 supplement see all Download asset Open asset SUMOylation differentially affects the velocity of forward- and back-propagating action potentials (APs). (a) Left: normalized averaged action currents (n = 500) elicited by a single AP at the axonal regions indicated by arrows to demonstrate the difference in the delay of their onset. The dashed vertical line corresponds to the time of dV/dtmax of the somatic action potential. Right: distance from the edge of the soma as a function of delay of spike initiation plotted. Note th