Phosphorylation of the HCN channel auxiliary subunit TRIP8b is altered in an animal model of temporal lobe epilepsy and modulates channel function

Temporal lobe epilepsy (TLE) is a prevalent neurological disorder with many patients experiencing poor seizure control with existing anti-epileptic drugs. Thus, novel insights into the mechanisms of epileptogenesis and identification of new drug targets can be transformative. Changes in ion channel function have been shown to play a role in generating the aberrant neuronal activity observed in TLE. Previous work demonstrates that hyperpolarization-activated cyclic nucleotide-gated (HCN) channels regulate neuronal excitability and are mislocalized within CA1 pyramidal cells in a rodent model of TLE. The subcellular distribution of HCN channels is regulated by an auxiliary subunit, tetratricopeptide repeat–containing Rab8b-interacting protein (TRIP8b), and disruption of this interaction correlates with channel mislocalization. However, the molecular mechanisms responsible for HCN channel dysregulation in TLE are unclear. Here we investigated whether changes in TRIP8b phosphorylation are sufficient to alter HCN channel function. We identified a phosphorylation site at residue Ser237 of TRIP8b that enhances binding to HCN channels and influences channel gating by altering the affinity of TRIP8b for the HCN cytoplasmic domain. Using a phosphospecific antibody, we demonstrate that TRIP8b phosphorylated at Ser237 is enriched in CA1 distal dendrites and that phosphorylation is reduced in the kainic acid model of TLE. Overall, our findings indicate that the TRIP8b–HCN interaction can be modulated by changes in phosphorylation and suggest that loss of TRIP8b phosphorylation may affect HCN channel properties during epileptogenesis. These results highlight the potential of drugs targeting posttranslational modifications to restore TRIP8b phosphorylation to reduce excitability in TLE.


Temporal lobe epilepsy (TLE) is a prevalent neurological disorder with many patients experiencing poor seizure control with existing anti-epileptic drugs. Thus, novel insights into the mechanisms of epileptogenesis and identification of new drug targets can be transformative. Changes in ion channel function have been shown to play a role in generating the aberrant neuronal activity observed in TLE. Previous work demonstrates that hyperpolarization-activated cyclic nucleotide-gated (HCN) channels regulate neuronal excitability and are mislocalized within CA1 pyramidal cells in a rodent model of TLE.
The subcellular distribution of HCN channels is regulated by an auxiliary subunit, tetratricopeptide repeat-containing Rab8b-interacting protein (TRIP8b), and disruption of this interaction correlates with channel mislocalization. However, the molecular mechanisms responsible for HCN channel dysregulation in TLE are unclear. Here we investigated whether changes in TRIP8b phosphorylation are sufficient to alter HCN channel function. We identified a phosphorylation site at residue Ser 237 of TRIP8b that enhances binding to HCN channels and influences channel gating by altering the affinity of TRIP8b for the HCN cytoplasmic domain. Using a phosphospecific antibody, we demonstrate that TRIP8b phosphorylated at Ser 237 is enriched in CA1 distal dendrites and that phosphorylation is reduced in the kainic acid model of TLE. Overall, our findings indicate that the TRIP8b-HCN interaction can be modulated by changes in phosphorylation and suggest that loss of TRIP8b phosphorylation may affect HCN channel properties during epileptogenesis. These results highlight the potential of drugs targetingposttranslationalmodificationstorestoreTRIP8bphosphorylation to reduce excitability in TLE.
An estimated 3.4 million patients suffer from epilepsy in the United States, and temporal lobe epilepsy (TLE) 2 represents one of the most common drug-resistant forms of the disorder (1)(2)(3). Although many patients with TLE respond to anti-epileptic drugs, one-third of individuals experience seizures that are refractory to medical treatment (4). Changes in ion channel function that contribute to the epileptic state are observed in individuals suffering from inherited epilepsies and in rodent models of TLE (5)(6)(7)(8)(9)(10). Therefore, understanding ion channelopathy in the context of epilepsy may lead to new targets for drugs to treat this disorder.
Hyperpolarization-activated cyclic nucleotide-gatedtk; (HCN1-HCN4) channels are nonspecific cation channels that mediate the I h current and are highly expressed in the hippocampus (11). HCN1 and HCN2 channels are enriched in the distal dendrites of CA1 pyramidal cells, where they limit neuronal excitability and regulate temporal summation and dendritic integration (12)(13)(14)(15). The intracellular region of the HCN channel consists of a cyclic nucleotide-binding domain (CNBD) that binds cAMP and a C-terminal tail consisting of the amino acids serine, asparagine, and leucine (SNL), which are also involved in protein binding. Binding of cAMP to the CNBD speeds channel activation kinetics and depolarizes the activation voltage (16 -19).
Within CA1 pyramidal neurons, the auxiliary subunit tetratricopeptide repeat-containing Rab8b-interacting protein (TRIP8b) regulates the subcellular location and function of HCN channels (20 -25). TRIP8b is a cytosolic protein that directly binds to HCN channels at two distinct locations to hyperpolarize channel gating and facilitate dendritic trafficking (21,26). Within the HCN channel intracellular C-terminus, TRIP8b binds to the CNBD, where it acts as a competitive antagonist to inhibit cAMP binding (16,21,(27)(28)(29). The second TRIP8b-HCN interaction occurs between the TRIP8b tetratricopeptide repeat (TPR) domains and the HCN channel C-terminal tail (26,30). Disrupting the CNBD binding interaction prevents TRIP8b-mediated hyperpolarization of channel gating (21,26). Both binding sites are necessary for proper dendritic trafficking in vivo, and disruption of either the CNBD or C-terminal tail interaction precludes dendritic enrichment of HCN channels by TRIP8b (22). TRIP8b contains an alternatively spliced N terminus, and the isoforms are named according to included exons. Isoform 1a-4 enhances channel surface expression, dendritic trafficking, and current density (20,22). In TRIP8b knockout (TKO) mice, there is significant loss of dendritic HCN channels with a concomitant reduction in I h , highlighting the importance of TRIP8b for channel function (31).
Genetic screens of patients with epilepsy have identified several mutations in HCN channels and their auxiliary subunits (32)(33)(34)(35)(36)(37)(38). In rodents, reduced HCN1 expression results in enhanced neuronal excitability as well as increased seizure susceptibility and seizure-related death (15,(39)(40)(41). In animal models of TLE, there is loss of I h , reduced surface expression, and hyperpolarization of channel activation in chronic epilepsy (42)(43)(44)(45)(46). Prior work demonstrated mislocalization of the channels throughout CA1 pyramidal cell dendrites that correlated with disruption of the TRIP8b-HCN interaction (47). These results raise the possibility that loss of TRIP8b function is responsible for HCN channel dysregulation in TLE. However, it has been unclear what molecular mechanism might be responsible for regulating the TRIP8b-HCN interaction in vivo. We hypothesized that posttranslational modifications of TRIP8b would limit binding to HCN channels and affect channel function.
HCN channel trafficking and gating can be modulated by changes in kinase or phosphatase activity, including calcium/ calmodulin-dependent protein kinase II (CaMKII), p38 mitogen-activated kinase and PKC activity (48 -50). During epileptogenesis, several alterations in kinase and phosphatase activity have been observed (43,51,52). Thus, we sought to determine whether changes in TRIP8b phosphorylation would affect HCN channel function and whether TRIP8b phosphorylation is dysregulated in TLE.
Through phosphoproteomics analysis, we and others identified a phosphorylation site on TRIP8b at residue Ser 237 that is located within the TRIP8b domain critical for CNBD binding (53). We generated a phosphospecific antibody targeting residue Ser 237 on TRIP8b and determined that TRIP8b is phosphorylated at Ser 237 in areas of dendritic HCN channel enrichment. The phosphorylation of this site enhances TRIP8b binding to the HCN channel CNBD and is necessary for TRIP8b-mediated hyperpolarization of channel gating. Additionally, we determined that both phosphorylation of this residue and CaMKII␣ activity are significantly reduced during epileptogenesis in the kainic acid model of TLE. This work demonstrates that restoring TRIP8b phosphorylation may enhance HCN channel function in the distal dendrites of the CA1 and reduce neuronal excitability in epilepsy.

Identification of phosphorylation site Ser 237 on TRIP8b
A previous global phosphoproteome analysis identified several possible phosphorylation sites on TRIP8b, including on residue Ser 237 (53). We immunopurified TRIP8b from hippocampal lysates prepared from both a control rat and a rat subjected to kainic acid (KA)-induced seizures. We then subjected the immunopurified and gel-purified TRIP8b protein to LC-MS/MS. As shown in Fig. 1, we obtained extensive sequence coverage of the rat TRIP8b protein isoform 1a-4 (Uni-Prot Q925N3) from both samples (473/602 aa or 79% from control and 294/602 or 49% from KA-treated). Fig. 1C shows a representative MS/MS spectrum of the peptide NHSLEEEFER (aa 235-244) that demonstrates unambiguous identification of in vivo phosphorylation of TRIP8b at Ser 237 . Phosphorylation at Ser 237 was identified on TRIP8b purified from both control and KA-treated rats. The normalized spectral counts for the phosphopeptide containing Ser 237 were present at a lower level in the KA-induced seizure sample than in the control sample (data not shown), suggesting that phosphorylation at this site may be affected by acute seizures.
Ser 237 is located within a conserved "nano" domain of TRIP8b (residues 235-275) that is highly conserved across vertebrate species (Fig. 2, A and B). This previously identified nano domain directly binds to the CNBD of HCN channels and competes with cAMP binding (54). Additionally, removing a segment of TRIP8b that contains the nano domain prevented distal dendritic enrichment of HCN1 in vivo (22). Given that the Ser 237 site is included within this functional domain, we hypothesized that Ser 237 phosphorylation could play a key role in regulating the TRIP8b-HCN interaction.

Generation and validation of the phosphospecific antibody pSer 237
An antibody targeted against pSer 237 (the phosphorylated TRIP8b residue Ser 237 ) was generated using a 15-amino acid sequence that included the phosphorylated serine in the TRIP8b nano domain as the antigen (see "Experimental procedures"). The specificity of this antibody for the phosphorylated serine was confirmed in Western blots with native and dephosphorylated TRIP8b. Hippocampi from WT mice were incubated with or without the calf intestinal phosphatase (CIP), and the lysates were probed with antibodies targeting total-TRIP8b, pSer 237 , and ␣-tubulin (Fig. 2C). Although total TRIP8b was expressed in both samples, the signal from the pSer 237 antibody was largely reduced with addition of the phosphatase. Additionally, the pSer 237 antibody did not detect any protein in hippocampi from TKO mice (Fig. 2C). Thus, the pSer 237 antibody is specific for the phosphorylated serine residue on TRIP8b.
Based on the TRIP8b sequence and known protein kinase motifs, we predicted that CaMKII␣ and PKA would phosphorylate TRIP8b(Ser 237 ) (55). We performed in vitro protein kinase assays to investigate this possibility by incubating purified TRIP8b protein with CaMKII␣ or PKA, with or without ATP, at 30°C for 30 min. The reaction mixtures were immunoblotted using antibodies against pSer 237 and TRIP8b (Fig. 2D). Treatment of TRIP8b with CaMKII␣ or PKA in the presence of ATP TRIP8b phosphorylation modulates HCN channel function enhanced the pSer 237 signal compared with the nonphosphorylated protein without ATP. These data demonstrate that both CaMKII␣ and PKA can phosphorylate the Ser 237 residue of TRIP8b in vitro.

Localization of phosphorylated TRIP8b in the brain
To determine the subcellular distribution of the phosphorylated TRIP8b protein, we performed immunohistochemistry. In the mouse hippocampus, we found that phosphorylated TRIP8b is enriched in the CA1 in a pattern similar to HCN1 and TRIP8b. The data show that HCN1, total TRIP8b, and phosphorylated TRIP8b are expressed in distal dendrites of CA1 pyramidal cells in the stratum lacunosum molecular layer ( . TRIP8b isoform 1a-4, which enhances HCN channel dendritic trafficking in the hippocampus, has also been shown to be enriched in this location through use of an antibody targeting TRIP8b exon 4 (24) (Fig. 3, A and B) (exon 4 (WT: 1, TKO: 0.5 Ϯ 0.1, p Ͻ 0.001, n ϭ 6). Therefore, consistent with our LC-MS/MS results, we hypothesized that isoform 1a-4 in particular is phosphorylated at Ser 237 . We used an ␣-exon 4 antibody to immunoprecipitate TRIP8b isoforms containing exon 4 (the majority of which is isoform 1a-4 (23)) from mouse hippocampal lysate. The total TRIP8b antibody detected several TRIP8b isoforms that were present in the hippocampal input sample (Fig. 3C). However, the pSer 237 antibody primarily detected one band that was enriched following immunoprecipitation with the ␣-exon 4 antibody. The pSer 237 band corresponds to the isoform 1a-4 band and confirms that, indeed, isoform 1a-4 is significantly phosphorylated in the hippocampus (Fig. 3C).
Because HCN1 is also trafficked to the distal dendrites of the neocortex in a TRIP8b-dependent manner (31), we examined the retrosplenial and parietal association cortices (Fig. 3, D and E). We observed a similar pattern of pSer 237 staining with greater expression in the distal dendrites in layer I (HCN1 (WT: We then examined other areas with TRIP8b expression to determine whether the protein is phosphorylated at Ser 237 throughout the brain. TRIP8b is expressed in oligodendrocytes of hippocampal area CA3 (24) and the granular layer of the cerebellum. TRIP8b Ser 237 was not phosphorylated to a detectable degree in either region (Fig. 4, A and B). We noted that nonphosphorylated TRIP8b specifically colocalizes with HCN2 in cerebellar oligodendrocytes but not with HCN1 in the Purkinje cell layer (Fig. S1). The data suggest that TRIP8b phosphorylated at Ser 237 is neuron-specific in brain regions that exhibit distal dendritic enrichment of HCN channels.

Phosphorylated TRIP8b at Ser 237 enhances binding to HCN channels
We reasoned that phosphorylation of Ser 237 could regulate the strength of the TRIP8b-HCN interaction, given that the residue is located within the nano domain of TRIP8b responsible for CNBD binding. We investigated this possibility using coimmunoprecipitation. To isolate the effect of pSer 237 specifically on CNBD binding, TRIP8b constructs with an ablated C-terminal tail TRIP8b-HCN interaction were used. These constructs contain a mutation of a key residue in the third TPR domain (N382A, previously referred to as TPR3-N13A) which disrupts the downstream interaction with the HCN channel C-terminal tail (21). TRIP8b constructs containing the N382A mutation alone, a phosphoablative mutation (S237A,N382A), a kinase consensus sequence mutation (R234A,N382A), or a negative control that disrupts both interaction sites (⌬58,N382A) were analyzed (21) (Fig. 5A). Because the arginine located at position 234 on TRIP8b is a critical residue in the consensus motifs for both CaMKII␣ and PKA, the construct containing the R234A mutation is also predicted to prevent phosphorylation (phosphoablative) (56,57). WT TRIP8b was basally phosphorylated in HEK293T cells, whereas both phosphoablative constructs show reduced signal from the ␣-pSer237 antibody (Fig. 5B). These constructs were transfected into HEK293T cells along with HCN1, HCN1 protein was immunoprecipitated from the lysates, and the level of TRIP8b bound to the channel was quantified ( To confirm these results, we performed an additional in vitro fluorescence polarization binding assay to analyze the interaction between TRIP8b and the CNBD of HCN channels (58).

TRIP8b phosphorylation modulates HCN channel function
This assay utilizes the change in polarization of a small, fluorescently tagged ligand (cAMP) whose polarization increases after binding a much larger protein (CNBD) (59). cAMP displacement serves as a readout of the affinity of TRIP8b binding because cAMP and TRIP8b compete for CNBD binding. We utilized fixed concentrations of FITC-tagged cAMP (8f-cAMP, 10 nM) and the CNBD of HCN1 (CNBD 387-591 , 0.625 M) and titrated increasing concentrations of a TRIP8b fragment conserved across all isoforms that includes the full nano domain through the C terminus of the protein (TRIP8b 219 -602 ). Prior to titration, TRIP8b 219 -602 was incubated with CaMKII␣, with or without ATP, at 30°C for 30 min. Titrating increasing concentrations of phosphorylated TRIP8b increased the affinity of TRIP8b for the HCN1 CNBD (IC 50 There are no additional CaMKII␣ motifs near the CNBD-binding region of TRIP8b. Thus, TRIP8b phosphorylated at Ser 237 binds to the HCN1 CNBD to a greater degree than nonphosphorylated TRIP8b.

Phosphoablative TRIP8b constructs prevent hyperpolarization of HCN channel gating
The change in affinity for HCN1 resulting from TRIP8b Ser 237 phosphorylation could alter HCN channel gating or current density. To test this idea, we carried out patch clamp recordings from HCN channels. TRIP8b hyperpolarizes channel activation and enhances the current density of both HCN1 and HCN2 channels; hence, we used HEK293 cells that stably express HCN2 and exhibit robust currents. HEK293 cells (in which TRIP8b Ser 237 is phosphorylated under basal conditions) stably expressing HCN2 were transfected with GFP and either TRIP8b(WT), TRIP8b(S237A), TRIP8b(R234A), or a control

TRIP8b Ser 237 phosphorylation and CaMKII␣ activity are reduced in epilepsy
A prior report identified that total HCN1 protein is unchanged in chronic epilepsy; instead, the channels are mis-localized and the interaction with TRIP8b is disrupted (47). Using the KA model of TLE, we measured hippocampal HCN1 protein expression 1 h, 1 day, and 28 days post-status epilepticus (SE). Because seizures arise 2-4 weeks post-SE in this model, these time points correspond to the acute phase, the latent phase before spontaneous seizures arise, and the chronic phase exhibiting spontaneous seizures, respectively (60,61). We observed no change in total HCN1 protein in the hip-  Table S1).

TRIP8b phosphorylation modulates HCN channel function
We then measured the level of TRIP8b phosphorylation at Ser 237 in the hippocampus to determine whether changes take place within the epileptogenic zone. Immediately following SE, the level of phosphorylated TRIP8b at Ser 237 was reduced in KA-treated animals compared with saline-treated animals (pSer 237 /TRIP8b, 1 h post-SE: Sal (1.00 Ϯ 0.04, n ϭ 5), KA (0.81 Ϯ 0.04, n ϭ 6), p Ͻ 0.01) (Fig. 8A, i, and Table S3). To estimate CaMKII␣ activity, we used an antibody targeting

Discussion
In this study, we have shown that TRIP8b phosphorylation regulates HCN channel binding and function. TRIP8b binding results in a unique CNBD conformation, one that both hyperpolarizes gating and reduces cAMP affinity for the CNBD (20,21,26,28). The identity of the specific amino acid residues responsible for the TRIP8b interaction with the CNBD has been the subject of several recent in vitro studies, one of which identified the nano domain as necessary and sufficient for TRIP8b binding to the HCN CNBD (54). Residue Ser 237 is located within the nano domain, where it is predicted to influence the effect of TRIP8b on channel gating and trafficking (21, 22, 26, 29, 54).

TRIP8b phosphorylation modulates HCN channel function
We conclude that TRIP8b can be phosphorylated at residue Ser 237 by CaMKII␣ and PKA and that the presence of the phosphate group increases the affinity of TRIP8b for the HCN CNBD (Figs. 2D and 5). Our in vitro experiments using purified proteins produced results that were consistent with those seen in other expression systems, suggesting that additional proteins are unlikely to mediate the increased affinity. The negatively charged TRIP8b nano domain is thought to be drawn to the positively charged residues of the HCN CNBD. Immediately C-terminal to TRIP8b Ser 237 is a series of negatively charged residues, EEEFE (239 -243). Reversing the charge of this stretch has been shown previously to reduce TRIP8b binding to the HCN1 CNBD (64). Additionally, reversing the charge of two positively charged lysine residues (665 and 666) in the HCN2 CNBD reduces the binding affinity between the CNBD and TRIP8b (27). Thus, we reason that addition of a negatively charged phosphate group to Ser 237 contributes to stabilization of TRIP8b binding to the CNBD. In addition to the effect of electrostatic interactions, dephosphorylation at Ser 237 could interfere with the structural changes that naturally occur within the TRIP8b nano domain and the CNBD to reduce binding (54).
This work demonstrates that phosphorylation of Ser 237 contributes to the high-affinity TRIP8b-HCN binding that is essential for dendritic trafficking and surface expression of HCN channels in the distal dendrites of CA1 pyramidal neurons. Distal dendritic enrichment of HCN1 in the hippocampus depends on an intact CNBD interaction (involving a stretch of TRIP8b that includes Ser 237 ) as well as CaMKII␣ activity (22,48). The homeostatic increase in I h observed following long-term potentiation is also dependent on CaMKII␣ activity, which raises the possibility that Ser 237 phosphorylation may be enhanced and contribute to I h up-regulation in these instances (65). Because both TRIP8b-HCN binding sites are necessary for successful dendritic trafficking, there may be either a minimum strength of interaction required or some essential conformation of the proteins that is modulated by phosphorylation. The strength of the two TRIP8b-binding interactions involves allosteric modulation by TRIP8b, whereby binding to HCN at either site increases the affinity of the other (58). Thus, the loss of Ser 237 phosphorylation we observed in epilepsy may not only affect the CNBD interaction but propagate downstream to weaken the C-terminal tail interaction as well. The role of phosphorylated Ser 237 in dendritic channel enrichment is supported by data indicating that TRIP8b is not phosphorylated in regions containing somatic or presynaptic HCN channels but specifically phosphorylated in cells exhibiting dendritic channel enrichment. One possibility not addressed here is that Ser 237 phosphorylation could be a consequence rather than a cause of dendritic channel targeting. Additional experiments will be necessary to distinguish whether Ser 237 phosphorylation regulates both dendritic channel targeting and surface trafficking or only one aspect of TRIP8b function.
TRIP8b binding to the HCN channel CNBD generates hyperpolarization of channel activation, but the net result is enhanced channel trafficking and increased current density in CA1 (21,22). Our electrophysiology data demonstrate that TRIP8b phosphorylation at Ser 237 is necessary for hyperpolar-ization of channel gating. Phosphoablative TRIP8b constructs prevent this shift in activation voltage, which is consistent with our in vitro binding experiments that indicate that these constructs have a lowered affinity for the CNBD.

Implications for TLE
We observed that Ser 237 phosphorylation was altered in a rat model of TLE, raising the possibility that these changes could be involved in producing HCN channel dysfunction. Both pilocarpine and KA models of TLE demonstrate reduced HCN channel function and dendritic expression in chronic epilepsy (42,44,46,47). We noted that 1 h, 1 day, and 28 days post-SE, TRIP8b phosphorylation is significantly reduced. Given that Ser 237 is dephosphorylated within an hour of SE, it is possible that this abnormality is essential to epileptogenesis. Although the loss of phosphorylation 1 h post-SE cannot be explained by a reduction in CaMKII␣ activity, it may be due to increased calcineurin activity, which has been observed following SE (52). However, loss of TRIP8b phosphorylation during the latent and chronic phases correlates with reduced CaMKII␣ activity and may be maintained by dysregulation of this kinase pathway. The reduction of TRIP8b phosphorylation at early time points may initiate dissociation of the TRIP8b-HCN complex that is observed in chronic epilepsy, leading to reduced dendritic channel trafficking and I h (47). We identified reduced phosphorylation in the hippocampus, but interestingly, phosphorylation levels remained stable in the neocortex in chronic epilepsy. The data suggest that the reduction of Ser 237 phosphorylation is specific to the epileptogenic zone and may play a role in development of epilepsy in the hippocampus.
In summary, this report identifies a posttranslational modification of TRIP8b at residue Ser 237 as a regulator of HCN channel function under physiological conditions and TLE. Restoring TRIP8b phosphorylation at this site may strengthen HCN1 channel binding and dendritic trafficking to enhance I h in TLE.

Immunoprecipitation for mass spectrometry
All animal use procedures were in strict accordance with the Guide for the Care and Use of Laboratory Animals described by the National Institutes of Health and approved by the University of California Davis Institutional Animal Care and Use Committee. Adult female Sprague-Dawley rats (Charles River Laboratories, Wilmington, MA), were injected with either saline or kainic acid to induce acute seizures using a single i.v. dose of 10 mg/kg. For KA-treated rats, behaviors were scored according to the Racine scale, and rats were euthanized immediately after reaching level 5 seizures (66). Hippocampi from both treatment groups were harvested and homogenized on ice in 2 ml of ice-cold homogenization buffer (20 mM Tris-HCl (pH 8.0), 320 mM sucrose and protease inhibitors). The homogenate was then diluted to 0.5 mg/ml and solubilized in ice-cold buffer containing 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1 mM EDTA, 5 mM NaF, 1% Triton-X-100, 1 mM NaOV 4 , and protease inhibitors. Insoluble material was pelleted at 16,000 ϫ g for 10 min at 4°C and discarded. The resulting lysate was then subjected to immunoprecipitation overnight at 4°C with a rab-

TRIP8b phosphorylation modulates HCN channel function
bit TRIP8b polyclonal antibody (generated against a C-terminal His-tagged fragment corresponding to aa 396 -462 of TRIP8b). Immunopurified products were recovered by incubation with DynaBeads for 1 h at 4°C, followed by three washes using icecold solubilization buffer and 1 wash with ice-cold detergentfree solubilization buffer. The beads were then removed to a fresh microfuge tube for nine more additional washes with icecold detergent free buffer. The final immunoprecipitation products were then eluted by boiling in 50 l reducing SDS sample buffer.

Sample preparation for mass spectrometry
Immunoprecipitation products were resolved on a 7.5% SDS-PAGE gel and visualized by colloidal blue Coomassie stain. Bands corresponding to Ϸ66 kDa were excised for in-gel digestion for both saline-and KA-treated tissues as reported previously (66). Briefly, samples were destained with 50% acetonitrile and washed extensively with distilled water. Proteins were then reduced in 10 mM DTT at 56°C for 1 h and alkylated in 5 mM iodoacetamide at room temperature for 30 min. Gel pieces were then washed and dehydrated prior to rehydration with trypsin solution (Promega, Madison, WI) at 12.5 ng/ml in 50 mM ammonium bicarbonate and incubated overnight at 37°C. Digested peptide mixtures were extracted three times with 10% formic acid for 15 min on ice and 100% acetonitrile for 5 min at room temperature. Extracted peptides were desalted and stored on StageTip C18 at 4°C (or room temperature) until LC-MS/MS.

LC-MS/MS analysis
LC-MS/MS analysis was performed at the UC Davis Proteomics Core Facility. LC separation was performed on a Nano Acquity UHPLC (Waters Corp.) with a Proxeon nanospray source. The digested peptides were reconstituted in 2% acetonitrile/0.1% TFA, and roughly 3 g of each sample was loaded onto a 100 m ϫ 25 mm Magic C18 100 Å 5U reverse-phase trap column and desalted online before being separated on a 75 m ϫ 150 mm Magic C18 200 Å 3U reverse-phase column. Peptides were eluted using a gradient of 0.1% formic acid (A) and 100% acetonitrile (B) with a flow rate of 300 nL/min. A 120-min gradient was run with 5% to 35% B over 100 min, 35% to 80% B over 8 min, 80% B for 1 min, and 80% to 5% B over 1 min and finally held at 5% B for 10 min. Each of the gradients was followed by a 1-h column wash.
Mass spectra were collected on an Orbitrap Q Exactive Plus mass spectrometer (Thermo Fisher Scientific) in a data-dependent mode with one MS precursor scan followed by 15 MS/MS scans. A dynamic exclusion of 15 s was used. MS spectra were acquired with a resolution of 70,000 and a target of 1 ϫ 10 6 ions or a maximum injection time of 30 ms. MS/MS spectra were acquired with a resolution of 17,500 and a target of 5 ϫ 10 4 ions or a maximum injection time of 50 ms. Peptide fragmentation was performed using higher-energy collision dissociation with a normalized collision energy value of 27. Unassigned charge states as well as ϩ1 and ions Ͼϩ5 were excluded from MS/MS fragmentation.

Animals
All additional animal experiments were performed according to protocols approved by the Institutional Animal Care and Use Committees of Northwestern University and Vanderbilt University. Adult male and female C57BL/6 mice (2-12 months) were obtained from The Jackson Laboratory (Ellsworth, ME). Adult male Sprague-Dawley rats (2-3 months) were obtained from Charles River Laboratories (Wilmington, MA).

Phospho-specific antibody pSer 237
Phosphorylated TRIP8b antibody was generated by Yen-Zym Antibodies, LLC (South San Francisco, CA). The immunogen used was a 15-amino acid peptide spanning residues 234 -247 (RNHpSLEEEFERAKA) containing the phosphorylated serine residue. Rabbits were immunized with the synthesized peptide conjugated to keyhole limpet hemocyanin, and the elicited antiserum was subjected to peptide affinity absorption and purification to isolate the phosphospecific antibody components. The antibody is referred to as pSer 237 . All TRIP8b residues described in this manuscript use the Rattus norvegicus isoform 1a-4 (UniProt Q925N3) as a reference sequence.

Phosphatase treatment
Hippocampal lysates from WT mice were incubated with 1.5 units/50-l reaction of CIP (New England Biolabs) for 2 h at 37°C before Western blotting as above.

In vitro protein kinase assay
Purified TRIP8b (219 -602) (residue numbering based on mouse TRIP8b isoform 1a-4) was incubated with CaMKII␣ enzyme (New England Biolabs, P6060), CaCl 2 , calmodulin, and CaMKII␣ protein kinase buffer (according to the manufacturer's instructions) with or without 200 M ATP for 30 min at 30°C. Purified TRIP8b (219 -602) protein was also incubated with PKA (New England Biolabs, P6000) and protein kinase buffer with or without 200 M ATP for 30 min at 30°C. The reactions were terminated in Laemmli sample buffer (Bio-Rad) containing BME and boiled at 95°C for 5 min. The protein was resolved by SDS-PAGE and immunoblotted using the indicated antibodies.

Immunohistochemistry
Immunohistochemistry was performed as described previously (68). Mice were deeply anesthetized with isoflurane and transcardially perfused with ice-cold PBS followed by 4% paraformaldehyde. Brains were removed and post-fixed for 48 h at 4°C. Coronal sections (30 m) were made using a vibratome (Leica), and antigen retrieval was performed with 10 mM sodium citrate (pH 9.0) for 10 min at 80°C. Slices were allowed to cool, washed with PBS, and immersed in blocking solution (5% normal goat serum and 0.03% Triton X-100 in PBS) for 1 h at room temperature with gentle agitation. The following primary antibodies were diluted in blocking solution and applied overnight at 4°C: custom guinea pig anti-HCN1 (1:3000), mouse anti-TRIP8b (1:1000, NeuroMab, clone N212/17, RRID: AB_10698035), rabbit anti-pSer 237 (1:1000, YenZym), custom guinea pig anti-HCN2 (1:1000), mouse anti-Olig2 (1:1000, Millipore, MABN50, RRID: AB_10807410), and mouse anti-TRIP8b exon 4 (1:1000, NeuroMab, clone N212/3, RRID: AB_10671951). Sections were then washed three times for 5 min each and incubated with fluorescently labeled secondary antibodies (Invitrogen) in blocking solution for 1 h at room temperature. Sections were then washed an additional three times with 4Ј,6-diamidino-2-phenylindole included in the final wash. The tissue was mounted on glass slides using PermaFluor (ThermoFisher). Images were acquired on an Olympus Fv10i confocal microscope (FluoView Software) using a ϫ10 or ϫ60 oil immersion objective with or without ϫ2 digital zoom. Imaging work was also performed at the Northwestern University Center for Advanced Microscopy, generously supported by NCI, National Institutes of Health CCSG P30 CA060553, awarded to the Robert H. Lurie Comprehensive Cancer Center, on a Nikon A1R confocal microscope using a ϫ20 objective. Images were also acquired in part through use of the Vanderbilt Cell Imaging Shared Resource (supported by National Institutes of Health Grants CA68485, DK20593, DK58404, DK59637, and EY08126). WT and KO tissues used for comparison were prepared simultaneously, and images were acquired and presented with identical settings using Fiji software.

Coimmunoprecipitation
HEK293T cells were maintained and transfected as described above. 48 h post-transfection, cells were washed with PBS and lysed in 600 l of TEEN-TX buffer (50 mM Tris, 1 mM EDTA, 1 mM EGTA, 150 mM NaCl, and 0.1% Triton X-100) supplemented with Halt protease and phosphatase inhibitor mixture (ThermoFisher). For coimmunoprecipitation from hippocampal lysate, the hippocampus was dissected in ice-cold PBS and lysed in 20 mM HEPES (pH 7.4), 100 mM NaCl, 5 mM EDTA, and 5 mM EGTA with Halt protease and phosphatase inhibitor mixture (ThermoFisher) and 1% Triton X-100. Lysates were sonicated briefly on ice and centrifuged at 21,000 ϫ g for 10 min. A 50-l aliquot of the supernatant (labeled "input") was added to 12.5 l of Laemmli sample buffer (Bio-Rad) containing BME and boiled at 95°C for 5 min. Custom guinea pig anti-HCN1 antibody (4 -5 l for HEK293T samples) or mouse anti-TRIP8b exon 4 (5 l for hippocampal samples, NeuroMab, clone N212/3, RRID: AB_10671951) was added to 450 l of the remaining lysate and rotated at 4°C overnight. The next day, 40 l of protein A/G agarose bead slurry (ThermoFisher) was washed three times in TEEN-TX, added to the antibody-coupled sample, and nutated for 2 h at 4°C. The beads were then washed in TEEN-TX buffer five times, 5 min each time, nutating at 4°C, and then eluted by TRIP8b phosphorylation modulates HCN channel function boiling in 40 l of Laemmli sample buffer (Bio-Rad) containing BME at 95°C for 5 min. Samples were resolved by SDS-PAGE (HEK293T cells: 8 l of input and eluate; hippocampi: 10 l of input, supernatant, and eluate). Densitometry of band intensity was performed using Li-Cor ImageStudio software, and the level of TRIP8b elution was divided by HCN1 elution for each condition.

Fluorescence polarization
CNBD and 8-f-cAMP oligonucleotides were synthesized by Integrated DNA Technologies (Coralville, IA). The HCN1 CNBD fragment used spans residues 387-591 of the mouse HCN1 isoform and contains a maltose-binding protein tag to enhance protein purification yield. Fluorescence polarization experiments were performed as described previously with modifications (58). Prior to these experiments, TRIP8b protein was incubated with CaMKII enzyme (New England Biolab), CaCl 2 , CaM, and protein kinase buffer according to manufacturer's instructions with or without 200 mM ATP. A series of 2-fold TRIP8b dilutions was then titrated into 0.625 M HCN1 CNBD peptide and 10 nM 8-f-cAMP with sterile filtered PBS and 1 mM DTT. Samples were loaded in triplicate in black 384well microtiter plates (Corning, Corning, NY), and polarization measurements were obtained using a Tecan microplate reader (Tecan, Zurich, Switzerland) at the Structural Biology facility at Northwestern University. We acknowledge staff and instrumentation support from the Structural Biology Facility at Northwestern University, the Robert H. Lurie Comprehensive Cancer Center of Northwestern University, and NCI, National Institutes of Health CCSG P30 CA060553. Curve fitting and data analysis were performed using Igor Pro (WaveMetrics). Curves were fit by the Hill equation based on a previous report (58).

Electrophysiology
Electrophysiology was performed as described in Ref. 20 with modifications. HEK293 cells stably expressing HCN2 were generated by transfecting cells with mouse HCN2 in the pcDNA3 plasmid using TransIT-LT1 (MirusBio) transfection reagent according to the manufacturer's protocol. 24 h post-transfection, G418 (0.5 mg/ml final concentration, Sigma-Aldrich) was added to the culture medium. Cells were maintained in G418 solution for 6 weeks for selection of cells stably expressing the pcDNA-HCN2 construct, and then aliquots were frozen in culture medium with 5% DMSO (Sigma) and stored in liquid nitrogen.
Cells were plated on autoclaved 12 mm-diameter Germanglass coverslips coated with poly-L-lysine (0.1 mg/ml, Sigma) and washed with Hank's balanced salt medium (without calcium or magnesium). Cells were transiently transfected 18 -24 h later with EGFP plus a control vector, TRIP8b(WT), TRIP8b(S237A), or TRIP8b(R234A), using Lipofectamine 2000 (according to the manufacturer's instructions with modifications). Lipofectamine exposure was limited to 60 min, followed by a rinse and replacement using preconditioned culture medium. Whole-cell recordings were performed 24 -48 h after transfection with pipettes made from borosilicate glass using a vertical puller (Narishige) with a final resistance of ϳ2.5-3.5 megaohms. Cells were held at Ϫ40 mV in voltage clamp and progressively stepped (Ϫ10 mV) from Ϫ40 to Ϫ120 mV. Extracellular solution consisted of 145 mM NaCl, 10 mM KCl, 10 mM glucose, 10 mM HEPES, 2 mM CaCl 2 , and 1 mM MgCl 2 buffered to pH 7.4 and an osmolarity of 312-315 mOsmol. The intracellular pipette solution contained 135 mM potassium gluconate, 10 mM MgCl 2 , 0.1 mM CaCl 2, 1 mM EGTA, 10 mM HEPES, and 2 mM Mg-ATP buffered to pH 7.3 and an osmolarity of 295-305 mOsmol. Maximal tail current amplitude populations measured from each cell were fit to the Boltzmann equation as described in Ref. 20. Maximal I h current density was calculated by dividing the maximal tail current by the cell capacitance, as described in Ref. 24. Currents were recorded via WinWCP software (University of Strathclyde), a MultiClamp 700A amplifier (Molecular Devices), and a National Instruments USB6221 interface card. The sampling rate was 10 kHz. Voltage clamping and series resistance were monitored throughout each experiment, and cells were discarded if the series resistance rose during the experiment. Two-tailed t tests were used to compare each group with the TRIP8b(WT)-transfected condition to test hypotheses generated by a pilot experiment. Expression of TRIP8b(WT) and its aforementioned mutants during the electrophysiology experiments described above were confirmed by subjecting cell samples of the exact same transfected conditions described above to Western blotting using an ␣-TRIP8b antibody (data not shown).

Kainic acid injections
Rats were allowed to acclimate in the on-site animal facility for at least 1 week prior to the experiment with food and water ad libitum. SE was induced by repeated low-dose intraperitoneal injections of KA dissolved in sterile saline (5 mg/kg; Tocris, Ellisville, MO) every 30 -45 min. To reduce mortality, rats received full doses (5 mg/kg) or half-doses (2.5 mg/kg) of KA according to seizure progression. Control rats (no KA) were given equivalent injections of sterile saline. Seizures were scored using a modified Racine scale (69): stage 1, behavioral arrest with facial automatisms; stage 2, head nodding; stage 3, forelimb clonus; stage 4, forelimb clonus with rearing; stage 5, rearing and falling. Status epilepticus was terminated as described previously (47) using diazepam (10 mg/kg i.p. injection or up to 15 mg/kg as needed to terminate SE) (70). To prevent dehydration and minimize discomfort, animals were given subcutaneous injections of warmed lactated Ringer's solution as needed and monitored closely in the days following SE induction. For the 1 h and 1 day post-SE experiments, 11 rats were injected with saline, and 14 rats were injected with KA (3 of 14 died in SE). These rats were divided into five control rats and six KA rats for the 1 h post-SE experiment and six control rats and five KA rats for the 1 day post-SE experiment. For the 28 days post-SE experiment, five rats were injected with saline, and eight rats were injected with KA (one rat died in SE, and two rats died prior to the 28-day time point), leaving five KA rats for analysis.

Statistics
Data were analyzed using Student's t test for two groups or one-way analysis of variance with Tukey's post hoc test for pair-TRIP8b phosphorylation modulates HCN channel function wise comparisons among three or more groups. Electrophysiology data were analyzed using Student's t test for comparison between TRIP8b(WT) and each additional condition in a pairwise manner as indicated. Data are represented as mean Ϯ S.E. (*, p Ͻ 0.05, **, p Ͻ 0.01, ***, p Ͻ 0.001).