Article Images (0)
×![]()
featured_play_listSave article to folder
X
Create new folder
Full Text
PDF
Table of Contents
Abstract
The Deubiquitinating Enzyme USP5
Modulates Neuropathic And Inflammatory
Pain By Enhancing Cav3.2 Channel Activity
Agustin Garcı́A-Caballero,1,3 Vinicius M. Gadotti,1,3 Patrick Stemkowski,1 Norbert Weiss,1 Ivana A. Souza,1 Victoria Hodgkinson,1 Chris Bladen,1 Lina Chen,1 Jawed Hamid,1 Anne Pizzoccaro,2 Mickael Deage,2 Amaury FrançOis,2 Emmanuel Bourinet,2 And Gerald W. Zamponi1,*
SUMMARY
INTRODUCTION
RESULTS
Cav3.2 Ubiquitination In Neurons From Mouse DRG
Cav3.2 Surface Density Is Regulated By WWP1 And WWP2
The DUB USP5 Interacts With Cav3.2 Channels
USP5 Affects Cav3.2 Ubiquitin Modification And T-Type Channel Activity
USP5 Levels Are Upregulated In Response To Peripheral
Effect Of USP5 Knockdown On Inflammatory Pain
Effect Of USP5 Knockdown On Neuropathic Pain
Effect Of A Tat-Cav3.2-III-IV Linker Peptide On T-Type Channel Activity
Effect Of The Tat-Cav3.2-III-IV Linker Peptide On Chronic Pain Hypersensitivity
DISCUSSION
EXPERIMENTAL PROCEDURES
AUTHOR CONTRIBUTIONS
ACKNOWLEDGMENTS
Abstract
Agustin Garcı́a-Caballero,1,3 Vinicius M. Gadotti,1,3 Patrick Stemkowski,1 Norbert Weiss,1 Ivana A. Souza,1 Victoria Hodgkinson,1 Chris Bladen,1 Lina Chen,1 Jawed Hamid,1 Anne Pizzoccaro,2 Mickael Deage,2 Amaury François,2 Emmanuel Bourinet,2 and Gerald W. Zamponi1,* 1Department of Physiology and Pharmacology, Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada 2Laboratories of Excellence in Ion Channel Science and Therapeutics, Institut de Génomique Fonctionnelle, CNRSUMR5203, INSERMU661, IFR3 Universités Montpellier I&II, Montpellier, France 3Co-first authors *Correspondence: Zamponi@ucalgary.ca http://dx.doi.org/10.1016/j.neuron.2014.07.036
Article
The Deubiquitinating Enzyme USP5
Modulates Neuropathic and Inflammatory
Pain by Enhancing Cav3.2 Channel Activity
Agustin Garcı́a-Caballero,1,3 Vinicius M. Gadotti,1,3 Patrick Stemkowski,1 Norbert Weiss,1 Ivana A. Souza,1 Victoria Hodgkinson,1 Chris Bladen,1 Lina Chen,1 Jawed Hamid,1 Anne Pizzoccaro,2 Mickael Deage,2 Amaury François,2 Emmanuel Bourinet,2 and Gerald W. Zamponi1,*
1Department of Physiology and Pharmacology, Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada 2Laboratories of Excellence in Ion Channel Science and Therapeutics, Institut de Génomique Fonctionnelle, CNRSUMR5203, INSERMU661, IFR3 Universités Montpellier I&II, Montpellier, France 3Co-first authors *Correspondence: Zamponi@ucalgary.ca http://dx.doi.org/10.1016/j.neuron.2014.07.036
SUMMARY
T-type calcium channels are essential contributors to the transmission of nociceptive signals in the primary afferent pain pathway. Here, we show that T-type calcium channels are ubiquitinated by WWP1, a plasma-membrane-associated ubiquitin ligase that binds to the intracellular domain III-IV linker region of the Cav3.2 T-type channel and modifies specific lysine residues in this region. A proteomic screen identified the deubiquitinating enzyme USP5 as a Cav3.2 III-IV linker interacting partner. Knockdown of USP5 via shRNA increases Cav3.2 ubiquitination, decreases Cav3.2 protein levels, and reduces Cav3.2 whole-cell currents. In vivo knockdown of USP5 or uncoupling USP5 from native Cav3.2 channels via intrathecal delivery of Tat peptides mediates analgesia in both inflammatory and neuropathic mouse models of mechanical hypersensitivity. Altogether, our experiments reveal a cell signaling pathway that regulates T-type channel activity and their role in nociceptive signaling.
INTRODUCTION
Low-voltage-activated T-type calcium channels are important mediators of electrical signaling in nerve, heart, and smooth muscle (Chen et al., 2003; Perez-Reyes, 2003; Simms and Zamponi, 2014), while their dysfunction has been linked to conditions such as epilepsy and pain (Bourinet et al., 2014; Simms and Zamponi, 2014). Vertebrates express three different types of T-type channels (termed Cav3.1, Cav3.2, and Cav3.3) with specific expression patterns and unique functional and pharmacological profiles (Perez-Reyes, 2003). At the molecular level, T-type channels are formed by a pore-forming Cav3 subunit that is comprised of four transmembrane domains that are 1144 Neuron 83, 1144–1158, September 3, 2014 ª2014 Elsevier Inc. connected by cytoplasmic linkers (Perez-Reyes, 2003). These linker regions contain protein interaction sites and are the targets of second messengers that modulate T-type channel activity (Iftinca and Zamponi, 2009). T-type channels regulate neuronal excitability (Chemin et al., 2002), and they contribute to low-threshold exocytosis (Weiss et al., 2012). These important functions are readily apparent in the primary afferent pain pathway, where T-type channels not only shape the firing patterns of pain sensing neurons but also contribute to neurotransmitter release at dorsal horn synapses (Jacus et al., 2012). Primary afferent pain fibers have their cell bodies in the dorsal root ganglia (DRG) and express predominantly the Cav3.2 subtype (for review, see Bourinet et al., 2014). Knockdown of Cav3.2 via intrathecal (i.t.) delivery of antisense oligonucleotides, or their direct inhibition by small organic molecules, results in an increase of bothmechanical and thermal thresholds, thus validating T-type channels as a suitable pharmacological target (Bourinet et al., 2005; Francois et al., 2013). T-type currents in primary afferents are enhanced in various painful pathological conditions, including diabetic neuropathy, bowel inflammation, and nerve injury (Jagodic et al., 2007; Marger et al., 2011;Messinger et al., 2009). Themolecular mechanisms by which this upregulation occurs are unknown. It is also not known whether increased T-type channel activity in these conditions takes place along the afferent fiber, in dorsal horn synapses, or at both loci (Waxman and Zamponi, 2014). Nonetheless, preventing this type of enhancement could be a potential means of combating the development of pain hypersensitivity. It is well known that ion channel protein levels can be regulated through ubiquitination by E3 ubiquitin ligases (Altier et al., 2011; Staub et al., 1996, 1997; Younger et al., 2004). Akin to the relationship between protein kinases and phosphatases, ubiquitinspecific proteases (USPs) or deubiquitinases (DUBs) remove ubiquitin groups from proteins targeted for degradation, leading to an increase in protein stability (Clague et al., 2013; Komander et al., 2009). However, their role in ion channel stability has been virtually unexplored. Here we report the regulation of T-type channels by a USP, USP5, that binds to and modulates T-type channel expression. We show that the ubiquitination state of (A) Western blot of ubiquitinated Cav3.2 channels immunoprecipitated via the H-300 antibody (Santa Cruz) from cultured mouse DRG neurons under control conditions (‘‘C’’) or after preincubation with 5 mM MG132 (MG) or 100 mM chloroquine (Chl) overnight, as detected with an anti-ubiquitin antibody. (B) Cav3.2 immunoprecipitates from mouse DRG neurons probed with the H-300 antibody. The membrane from (A) was stripped and used for reprobing with the indicated antibody. The lower panel shows a control blot for a-tubulin using the same samples as in (A) and (B) but run on a separate gel. Note that the Cav3.2 band most likely contains a combination of mono-ubiquitinated and non-ubiquitinated channels as the addition of a single ubiquitin molecule adds only 8 kDa. The blots are representative examples from three different experiments. (C) Cav3.2 isoform1 and isoform 2 III-IV linker sequence alignment for both mouse and human channels. Arrows indicate lysines and underlined sequence indicates the PY motif. (D) Ubiquitin signal of Cav3.2 WT and different lysine (K1560R, K1576R, K1587R) mutant channels from tsA-201 cell immunoprecipitates; a representative blot is shown, n = 4. (E) Cav3.2 channel immunoprecipitates from tsA-201 cells of the total pool of channels are shown by western blot. The membrane from (D) was stripped and used for reprobing with an antiCav3.2 antibody. A blot for a-tubulin is shown as control (bottom panel). (F) Cav3.2 WT and mutant ubiquitinated signals were quantified by densitometry, normalized as a ratio of ubiquitinated-Cav3.2 versus a-tubulin, and expressed as percentage of the control (WT signal). See also Figure S1. All error bars reflect standard errors. Cav3.2 channels in the afferent pain pathway is regulated by an interplay of the HECT E3 ligase WWP1 and USP5 that fine tunes the stability of T-type channel protein in the plasma membrane. We show that disrupting USP5 regulation increases ubiquitination and decreases Cav3.2 channel activity, thereby leading to analgesia in mouse models of inflammatory and neuropathic pain.
RESULTS
Cav3.2 Ubiquitination in Neurons from Mouse DRG
To determine if T-type channels are subject to ubiquitination, cultured mouse DRG neurons were harvested in the presence or the absence of the proteasome inhibitor MG132. Cav3.2 channels were immunoprecipitated, and western blots were probed with a ubiquitin-specific antibody. We detected a Cav3.2 ubiquitination signal only from immunoprecipitates from cells treated Ne with 5 mMMG132 but not from cells treated with 100 mM chloroquine, a lysosomal inhibitor (Figure 1A), indicating that Cav3.2 channels are ubiquitinated and degraded in the proteasome. Higher molecular weight bands consistent with channel aggregates were also evident in some of the western blots probed with a Cav3.2 antibody (Figure 1B; Figure S1 available online), with these higher molecular weight bands appearing more abundant in MG132-treated cells. Ubiquitination of Cav3.2 channels was verified by using a different precipitating Cav3.2 antibody (Figure S1A) and by mass spectrometry analysis (Figures S1B and S1C). Ubiquitination was also observed with the two other members of the T-type channel family, Cav3.1 and Cav3.3 (data not shown). To identify the ubiquitination site in Cav3.2, we focused on the intracellular domain III-IV linker region of the channel, as this region includes a number of lysine residues (i.e., the residues modified by ubiquitination) as well as a putative Proline-Tyrosine uron 83, 1144–1158, September 3, 2014 ª2014 Elsevier Inc. 1145 A B C D E F G Figure 2. HECTE3LigasesWWP1andWWP2 Regulate Cav3.2 Channel Ubiquitination (A) Western blot of ubiquitinated surface Cav3.2 channels expressed in tsA-201 cells transfected with WWP1 or WWP2 shRNAs and pretreated with 5 mM MG132 overnight. After biotin labeling of cells, labeled surface proteins were enriched with neutravidin beads. Samples were eluted off the beads with RIPA buffer containing 4M guanidine at pH 1.6, and samples were tumbled for 30 min at 4 C before the samples were dialyzed overnight. Surface Cav3.2 channels were enriched by immunoprecipitation with an anti-Cav3.2 antibody and probed for ubiquitin. (B) Western blot of surface pool of Cav3.2 channels expressed in tsA-201 cells treated with WWP1 or WWP2 shRNAs. (C) Western blot of total pool of Cav3.2 channels expressed in tsA-201 cells treated with WWP1 or WWP2 shRNAs. An actin blot is shown as sample control for (A)–(C). (D) Quantification of ubiquitinated surface Cav3.2 channels by densitometry, normalized as a ratio of surface ubiquitinated-Cav3.2 versus surface Cav3.2 (Total Cav3.2), and expressed as percentage of the control relative integrated density values from western blots. Data from three experiments are included in the bar chart. (E) Quantification of surface Cav3.2 channels expressed as the ratio of surface versus total channels by densitometry to obtain relative integrated density values from western blots. Data from four to seven experiments are included in the bar chart. (F) Affinity purification assays were done using the following biotinylated peptides; human Cav3.21556–1602 (long III-IV linker, lane 1) and Cav3.21569–1586 (short III-IV linker, lane 2) peptides. Mouse DRG lysates (500 mg) were incubated with each peptide (50 mg) and neutravidin beads for 2 hr at 4 C. Western blot analysis was performed using rabbit anti-WWP1antibody. A representative immunoblot is shown (n = 3). (G) Western blot of Cav3.2 immunoprecipitates from mouse DRG lysates probed with an antiWWP1 antibody. See also Figure S2. All error bars reflect standard errors. (PY) HECT E3 ubiquitin ligase binding motif (Staub et al., 1996), which is conserved in all three Cav3.1, Cav3.2, and Cav3.3 channels in both human and mouse (Figure 1C). We mutated three individual lysine residues upstream from the PY motif (i.e., K1560R, K1576R, and K1587R) and carried out immunoprecipitations from transfected tsA-201 cells combined with western blot assays to analyze the effects of the mutations on the ubiquitination level. Ubiquitination of the K1560R and K1576R mutants decreased by 60% and 25%, respectively, when compared to wild-type (WT) channels, whereas the K1587Rmutant showed a similar ubiquitination signal as theWT channel (Figures 1D–1F). Altogether, these data strongly support the notion that Cav3.2 channels undergo ubiquitinmodification and that the intracellular III-IV loop is the target for E3 ubiquitin ligases. 1146 Neuron 83, 1144–1158, September 3, 2014 ª2014 Elsevier Inc.
Cav3.2 Surface Density Is Regulated by WWP1 and WWP2
We used shRNAs directed against a number of different HECT E3 ligases, including Nedd4, Nedd4-Long, Itch, Smurf1, Smurf 2, WWP1, and WWP2, and examined their consequences on ubiquitination of transiently expressed Cav3.2 channels (Figure 2). tsA-201cells (shRNA treated or untreated) were labeled with biotin, the labeled surface pool of channels enriched by immunoprecipitation using a Cav3.2 antibody, and ubiquitination determined by western blots. Among the various shRNAs examined, only shWWP1 and shWWP2 mediated a reduction in ubiquitinated Cav3.2 channels among the Cav3.2 cell surface pool (to 23% and 25% of control levels, respectively) (Figures 2A and 2D). The efficiency of WWP1 knockdown was verified using western blotting (Figure S2A). Concomitantly, we observed an 2-fold increase in Cav3.2 cell surface expression levels when cells were depleted of WWP1 and WWP2 (Figures 2B, 2C, and 2E) but not when cells were treated with shRNA against Nedd4, an HECT ligase implicated in regulation of channels such as ENaC (Abriel et al., 1999; Staub et al., 1996, 1997). Altogether, these data suggest that WWP1 and WWP2 regulate Cav3.2 cell surface density via ubiquitin modification. Next, we tested if WWP1 and WWP2 are able to bind to the III-IV linker region. We created a 46-amino-acid-long III-IV linker peptide (hCav3.21556–1602) containing the PYmotif or an 18 amino acid short peptide lacking the PY motif (hCav3.21569–1586), and these peptides were then incubated with mouse DRG lysates. The 46-amino-acid-long peptide encompassing almost the entire III-IV linker bound WWP1 ( 104 kDa) (lane 1 in Figure 2F). In the 18 amino acid peptide lacking the PYmotif, WWP1 binding was reduced, albeit not completely eliminated (lane 2 in Figure 2F), suggesting that the PY motif is not the sole determinant of WWP1 binding. This is confirmed by substitution of the two tyrosines of the PxYxxY motif with asparagine (Figures S2B and S2C), which resulted in a partial decrease in their ubiquitination level ( 50%) (Figures S2B–S2D). WWP1 coimmunoprecipitated with Cav3.2 channels in mouse DRG lysate (Figure 2G), whereas WWP2 did not (Figure S2E). Control experiments indicate that WWP2 is in fact capable of interacting with Cav3.2 channels (Figure S2F), but that is only weakly expressed in DRG neurons (Figure S2G). Altogether, these data suggest that WWP1 and WWP2 can bind to Cav3.2 and regulate its ubiquitination state.
The DUB USP5 Interacts with Cav3.2 Channels
To determine if the domain III-IV linker interacts with other components of the ubiquitination machinery, we used a biotinylated version of the highly charged region in the domain III-IV linker (i.e., residues F1558–R1586) as bait and carried out affinity precipitation assays followed by mass spectrometry. This yielded several candidate interacting partners (Figure 3A), including p97 (a protein involved in the proteasomal degradation pathway) and USP5 (or isopeptidase T), a DUB recently identified as part of the 26S proteasome (Besche et al., 2009). We confirmed the Cav3.2-USP5 interaction by means of affinity precipitation and coimmunoprecipitation assays. A biotinylated peptide (which corresponds to residues 1,556–1,602 of the III-IV linker of the human Cav3.2 isoform 2) showed robust binding to USP5 from mouse brain lysate, whereas a scrambled Cav3.21556–1602 sequence or a Cav3.2 C terminus peptide (corresponding to residues 1,860–1,884 of hCav3.2) did not (Figure 3B). To determine if USP5 binds to the III-IV linker directly, we carried out in vitro binding assays with human recombinant USP5 proteins. As there is evidence of long (858 amino acid) and short (835 amino acid) USP5 splice isoforms in human tissues (Falquet et al., 1995), we incubated both recombinant long (96 kDa) and short (93 kDa) hUSP5 proteins with hCav3.21556–1602 (III-IV linker) or hCav3.21860–1884 (CT) peptides. Both USP5 isoforms strongly bound to the Cav3.2 III-IV linker peptide but not to the Cav3.2 C terminus peptide (Figure 3C), indicating the Ne observed interactions occur independently of USP5 splice variation. PCR analysis (Figure S3A) reveals that both USP5 splice isoforms are expressed in various regions of the nervous system that are known to express Cav3.2 channels (Talley et al., 1999), including cortex, cerebellum, spinal cord, and DRG. DRG appear to show greater expression of the long form of USP5, but it is unclear which of the two isoforms is more important at the functional level (Figure 3D). Expression of USP5 in DRG neurons was also verified by immunostaining with a USP5 antibody (Figure 3E). USP5 expression was observed in myelinated (i.e., NF200 positive) and to a much lesser degree in unmyelinated (i.e., IB4 positive) neurons. There was overlap with markers of peptidergic neurons (Figure S3F). USP5 expression could also be visualized in dorsal horn slices (Figure 3F). Intense USP5 staining was evident in the nuclei in a number of neurons (Figures S3B and S3C), which fits with the well-documented nuclear expression of other types of DUBs (Clague et al., 2013). Cav3.2 could be immunoprecipitated with USP5 from mouse DRG neurons (Figure 3G). Control experiments included Cav3.2null mouse DRG tissue and using an additional precipitating Cav3.2 antibody (Figure 3H). Successful coimmunoprecipitations were also obtained from mouse dorsal horn lysate (Figure 3I). Coimmunoprecipitations from whole rat brain further confirmed the Cav3.2/USP5 interaction, whereas the channels did not coimmunoprecipitate with USP15, another DUB that is found in brain (Figures S3D and S3E). Experiments using brain lysate revealed that Cav3.3 channels weakly interact with USP5, whereas Cav3.1 channels do not (data not shown). Overall, these data demonstrate that USP5 binds to the III-IV linker of Cav3.2 in vitro and binds to the full-length channel in native tissue.
USP5 Affects Cav3.2 Ubiquitin Modification and T-Type Channel Activity
To determine if USP5 is able to regulate Cav3.2 channel ubiquitination, we transfected CAD cells (which express USP5 endogenously) (Figure S4A) with USP5 shRNA. Knockdownwas verified, and the level of Cav3.2 ubiquitination assessed. Cav3.2 channels immunoprecipitated from cells transfected with USP5-shRNA (and treated with 5 mMMG132) exhibited five times greater ubiquitination levels when compared to channels transfected with USP15-shRNA (Figure S4B). As a result, total Cav3.2 protein levels were significantly reduced by USP5-shRNA treatment (Figures S4C and S4D). We also examined the effects of USP5-shRNA on whole-cell T-type channel currents in CAD cells. Peak T-type currents were decreased in CAD cells transfected with USP5-shRNA (shUSP5: 1.4 ± 0.2 pA/pF, n = 18) when compared with peak currents obtained from control cells (GFP: 3.3 ± 0.4 pA/pF, n = 20, Students t test: p < 0.001) (Figures S4E and S4F). Application of nickel confirmed that the T-type current expressed endogenously in CAD cells was carried by Cav3.2 channels (Figure S4G). Taken together, these results indicate that ubiquitinated Cav3.2 channels are substrates for USP5 and that the association of the channel with USP5 results in altered Cav3.2 protein levels and whole-cell current densities. uron 83, 1144–1158, September 3, 2014 ª2014 Elsevier Inc. 1147 (A) Interacting proteins from mouse brain lysates that bind to the hCav3.21556–1602-III-IV linker peptide as identified by mass spectrometry. (B) Affinity purification assays from mouse brain lysate were done using the following biotinylated peptides: scrambled hCav3.21556–1602 (lane 1), human Cav3.21556–1602(III-IV linker, lane 2), or Cav3.21860–1884 (CT, lane 3) peptides. Western blot analysis was performed using a rabbit antibody to USP5. A blot for actin is shown as control (bottom panel). A representative immunoblot is shown (n = 3). (legend continued on next page) 1148 Neuron 83, 1144–1158, September 3, 2014 ª2014 Elsevier Inc.
USP5 Levels Are Upregulated in Response to Peripheral
Nerve Injury It is possible that the observed enhancement of T-type channel activity in various rodent pain models could be due to altered USP5 regulation.We thus examinedUSP5 levels in DRGand spinal cord tissue from mice that had received an injection of either PBS or complete Freund’s adjuvant (CFA) into the hindpaw. Tissue was then isolated from the contralateral and ipsilateral side and analyzed by western blots. As shown in Figure 4, there was an upregulation of USP5 levels in DRG (Figures 4A and 4B) and spinal dorsal horn (Figures 4C and 4D) tissue on the ipsilateral side in CFA-treated animals, whereas there was no change in USP5 levels in tissue isolated from PBS-treated animals. We also performed similar experiments in mice subjected to a chronic constriction injury (CCI) of the sciatic nerve. These experiments revealed that there was an enhanced association between USP5 and Cav3.2 in the ipsilateral dorsal horn in chronically (14 days after CCI) nerve-injured mice (Figures 4E and 4F). Along these lines, immunostaining experiments involving spinal cord slices from mice with a spared nerve injury (SNI) reveal an increase in USP5 levels in the spinal dorsal horn ipsilateral to the lesion (Figure 4G). Altogether, these data fit with a mechanism in which USP5mediated regulation of Cav3.2 can occur physically and functionally in both DRG neurons and dorsal horn synapses and is aberrantly enhanced in chronic pain conditions.
Effect of USP5 Knockdown on Inflammatory Pain
To determine if USP5 could modulate pain signaling via T-type channels, we treated mice via an i.t. injection of either USP5shRNA or USP15-shRNA (both delivered at a dose of 12.5 mg/ i.t. 1 day prior to pain testing). As shown in Figures 5A and 5B, i.t. delivery of USP5 shRNA resulted in a decrease in USP5 protein levels in both DRG and dorsal horn tissue and consequently prevented the upregulation of Cav3.2 channels in animals subjected to CCI (Figure 5C). To determine whether USP5 modulates afferent pain signaling under chronic inflammatory processes, we analyzed mechanical withdrawal threshold of USP5-shRNA-treated animals after CFA injection. As shown in Figure 5D, mice injected with CFA developed mechanical hyperalgesia, as indicated by a significant decrease of paw withdrawal thresholds (p < 0.001) when compared to pre-CFA baseline levels of the control group. Three (C) In vitro binding assay of purified recombinant USP5 (long isoform [left] and shor (CT, lane 3) peptides. The asterisk indicates a degradation product of the USP5 sh approximately 3 kDa and is not resolved on a nongradient gel. (D) RT-PCR analysis of USP5 mRNA in different neuronal tissues. The orange bo (528 bp) USP5 isoforms (n = 3). (E) Representative example of a mouse DRG slice immunostained for USP5 (1:5 (F) Representative example of USP5 staining of a mouse spinal cord slice colabe (G) USP5-Cav3.2 coimmunoprecipitation assay using mouse DRG tissue. Cav3.2 i different immunoprecipitate: the first lane is an IgG control, the second lane reflec USP5 antibody. A blot for a-tubulin is shown as control (bottom panel). The gel is ( 350 kDa) band in the coimmunoprecipitation lane likely reflects an IgG aggreg (H) Coimmunoprecipitation of USP5 and Cav3.2 from WT and Cav3.2 KO mouse N terminus (NT) of the channel. (I) USP5-Cav3.2 coimmunoprecipitation from mouse dorsal horn tissue. Each lan See also Figure S3. Ne days after CFA injection, i.t. treatment of mice with USP5-shRNA (12.5 mg/i.t.) reversed mechanical hyperalgesia induced by CFA from 9 hr up to 3 days after treatment. In contrast, similar treatment of mice with USP15-shRNAwas ineffective (Figure 5D). We also tested the role of USP5 in both phases of the formalin model (Hunskaar and Hole, 1987). A blind analysis demonstrated that upon USP5-shRNA treatment, the nocifensive response time (i.e., time spent licking and biting the injected paw) was reduced by 47%± 7% in the first phase (Figure S5A) and 57%± 5% in the second phase (Figure S5B). Collectively, these data indicate that USP5 is a regulator of chronic pain states accompanying inflammatory processes.
Effect of USP5 Knockdown on Neuropathic Pain
To establish if USP5 also contributes to the development of neuropathic pain, we examined the prophylactic (1 day before injury) and therapeutic (12 days after injury) action of USP5shRNA in CCI mice. Sciatic nerve constriction triggered both mechanical (Figure 5E) and thermal (Figure 5F) hyperalgesia, as indicated by a significant decrease of mechanical and thermal withdrawal thresholds when compared to pre-CCI baseline levels of control group (p < 0.01). Treatment of mice with spinal USP5-shRNA (12.5 mg/i.t., delivered 1 day before nerve injury) prevented mechanical and thermal hyperalgesia, consistent with the effects of USP5 shRNA on injury-induced upregulation of Cav3.2 channels (see Figure 5C). This effect lasted for several days before returning to control levels at about day 10. Importantly, a second injection on day 12 reestablished analgesia, indicating that USP5 knockdown can act both prophylactically and therapeutically. The second injection of USP5 shRNA appeared less effective than the prophylactic treatment (Figures 5E and 5F), perhaps due to the expression of other plastic (non-T-type channel-mediated) changes that could occur in response to injury. Altogether, these data indicate that USP5 is a major determinant of hypersensitivity in both inflammatory and neuropathic pain models.
Effect of a Tat-Cav3.2-III-IV Linker Peptide on T-Type Channel Activity
To determine if the effects of USP5 occurred via Cav3.2 T-type channels, we designed Tat peptides to competitively disrupt USP5 binding to the Cav3.2 domain III-IV linker (Figure 6A). We first verified in an in vitro affinity competition assay that the t isoform [right]) to humanCav3.21556–1602 (III-IV linker, lane 2) or Cav3.21860–1884 ort isoform. The predicted size difference between the long and short isoform is xes highlight the expected bands for the long (597 bp and 362 bp) and short 00 Ab) and either NCF200 or IB4. led for USP5 (red) and IB4 (green). mmunoprecipitates were probed for USP5 by western blot. Each lane reflects a ts an IP control with a Cav3.2 antibody, and the third lane reflects an IP with the a representative example from three experiments. The high molecular weight ate that occasionally appears in co-IPs. DRG neurons using a Cav3.2 antibody (N18, Santa Cruz) directed against the e reflects a different immunoprecipitate, as indicated by the small arrows. uron 83, 1144–1158, September 3, 2014 ª2014 Elsevier Inc. 1149 A B Tat-3.2-III-IV peptide was capable of interacting with USP5 (Figure S6A), and we verified that the Tat peptide did not directly block Cav3.2 currents in tsA-201 cells (data not shown). We then tested the effects of the Tat peptides on T-type channelmediated calcium entry into acutely dissociated DRG neurons that were loaded with the calcium indicator dye Fluo-4. After loading, high-voltage-activated calcium channels were blocked with 3 mM cadmium, and cells were depolarized with application of 25 mM KCl to trigger T-type channel-mediated calcium entry (verified by application of 100 mM nickel, which eliminated 75.3% ± 5.1% [n = 35] of calcium influx). Next, we examined T-type channel-mediated calcium entry before and after application of either the Tat-3.2-III-IV linker peptide or a control Tat peptide directed against the C terminus of the channel. As shown in Figures 6B and 6C, the former peptide attenuated calcium entry, whereas the control peptide did not. Altogether, these experi- 1150 Neuron 83, 1144–1158, September 3, 2014 ª2014 Elsevier Inc. cus et al., 2012). To test if these synaptic T-type channels are regulated by USP5, we performed miniature excitatory postsynaptic current recordings from spinal cord slices from the lumbar region (Figures 6D and 6E). Electrodes were placed in lamina II, which is known to receive inputs from Cav3.2 expressing afferents (Jacus et al., 2012) and includes nerve terminals from both myelinated and unmyelinated fibers. For each cell, miniature excitatory baseline synaptic activity was recorded, and then the slice was exposed to 100 mM nickel to determine if synaptic input in a given cell contained a contribution from T-type channels (Figures 6D and S6B). Nickel was then washed out and the effects of a subsequent application of the Tat-3.2-III-IV peptide (10 mg/ml) assessed (Figures 6D and S6C). For 12 out of 15 cells examined, the application of nickel resulted in a statistically significant increase (p < 0.05, Kolmogorov-Smirnov twosample test) in the interevent interval, consistent with an (A) In vivo knockdown of USP5 via i.t. delivery of shRNA (12.5 mg/i.t.) against USP15 (as a control) and USP5. L5 DRG were isolated and lysed, and USP5 immunoprecipitates were probed with an USP5 antibody. This experiment was repeated three times. The bar graph represents a quantification of the USP5 band intensity relative to the actin control for three different experiments. (B) In vivo knockdown of USP5 in dorsal horn tissue via i.t. delivery of shRNA against USP5. The asterisk denotes the putative USP5 proteolytic fragment. (C) Quantification of the effect of USP5-shRNA on Cav3.2 protein levels in L4–L6 dorsal horn tissue neurons from mice that were subjected to CCI of the sciatic nerve. The data were normalized to an actin control. (D) Time course of mechanical hyperalgesia of mice treated with USP5-shRNA or USP15shRNA (12.5 mg/i.t.) under CFA-induced chronic inflammatory pain. Each circle represents the mean ± SEM (n = 6–8) and is representative of two independent experiments. Two-way ANOVA revealed that paw withdrawal thresholds of mice receiving USP5-shRNA were significantly greater than those treated with USP15-shRNA (negative control). (E and F) Time course of mechanical (E) and thermal (F) hyperalgesia of mice treated with USP5-shRNA or USP15-shRNA (12.5 mg/i.t.) under CCI-induced neuropathic pain. Each circle represents the mean ± SEM (n = 5 to 6) and is representative of two independent sets of experiments. Two-way ANOVA revealed that spinal treatment of mice with USP5-shRNA significantly attenuated the mechanical hyperalgesia induced by CCI when compared with the CCI + USP15-shRNA (control) group. The dashed line and hashtag indicate the range of data points where injured animals differed from the sham treated group (p < 0.001). See also Figures S4 and S5. All error bars reflect standard errors. involvement of T-type channels. One third of the nickel-sensitive cells displayed a statistically significant increase in interevent interval in response to the Tat-3.2-III-IV peptide (which was quantified and compared to the nonresponders) (Figure 6E). In contrast, the Tat-3.2-III-IV peptide did not affect the nickelinsensitive cells. These data indicate that (1) dorsal horn neurons/synapses form a heterogeneous population that differs in their sensitivities to T-type channel inhibition and (2) there is a subpopulation of T-type channel-expressing synapses whose activity is reduced by disruption of the USP5/Cav3.2 channel interaction. To determine if USP5 regulation occurs in synaptic nerve terminals from afferent fibers rather than excitatory interneurons, we examined the effect of the Tat-3.2-III-IV peptide on T-type channel-mediated events in spinal cord slices evoked by paired pulse stimulation of the dorsal root entry zone. As shown in Fig- Ne ures 6F and 6G, the peptide resulted in a statistically significant increase in paired pulse ratio but not a statistically significant decrease in evoked EPSC amplitude. Qualitatively similar effects were observed upon application of 100 mM nickel (Figure S6D). Altogether, these experiments indicate that USP5 regulation of Cav3.2 channels occurs both in cell bodies and in synaptic nerve terminals arising from DRG neurons, which collectively would result in an enhancement of afferent pain signaling.
Effect of the Tat-Cav3.2-III-IV Linker Peptide on Chronic Pain Hypersensitivity
We then examined the effects of the III-IV linker Tat peptide (and as controls, a Tat-free III-IV linker peptide and a Tat peptide directed against the C terminus of the channel) (see Figure 6A) in inflammatory (CFA) and neuropathic (CCI) pain models. Once the reduced withdrawal thresholds to mechanical stimuli uron 83, 1144–1158, September 3, 2014 ª2014 Elsevier Inc. 1151 A B C D E F G Figure 6. USP5 Regulation of Cav3.2 Channels in DRG Neurons and Dorsal Horn Synapses (A) Sequence alignment of peptides used for in vitro and in vivo mouse studies. Underline indicates a helix prediction (http://bioinf.cs.ucl.ac.uk). (B) Examples of calcium fluorescence measurements in acutely dissociated mouse DRG neurons before and after application of either the Tat-3.2-III-IV or C terminus control peptide (Tat3.2-CT) in response to a 25 mM KCl-induced depolarization. (C) Bar graph representing a quantification of the effects of the Tat-3.2-III-IV linker and control peptides on T-type channel-mediated calcium rises. (D) Electrophysiological recordings of miniature excitatory synaptic events in spinal cord slice recordings from the Substantia gelatinosa in the absence or presence of either nickel or the Tat3.2-III-IV peptide. All traces were taken from the same nickel-sensitive cell. (E) Quantification of experiments such as those shown in (D). The bar graph depicts the effects of nickel and the Tat-3.2-III-IV peptide on interevent interval. Data were normalized to control and grouped based on statistical analysis (Kolmogorov-Smirnov two-sample test) of effects within a given recording (normalized to control). (F and G) Effect of the Tat-3.2-III-IV peptide (10 mg/ml) on evoked synaptic responses in dorsal horn neurons. Sample recordings of average EPSCs in response to paired-pulse stimulations (100 ms apart, evoked at a rate of 0.2 Hz) before (baseline; black) and after Tat-3.2-III-IV peptide (gray) are shown (F). To illustrate effect on paired pulse ratio (PPR), traces have been scaled for EPSCs evoked in pulse 1 to be equivalent in amplitude. Stimulus artifacts have been cropped for clarity, and scale bars represent 25 pA and 20 ms. Scatter plots (G) summarize the effect Tat-3.2-III-IV peptide on PPR (left) and eEPSC amplitude (right) in multiple cells. Tat-3.2-III-IV peptide significantly increased PPR (p = 0.02, paired t test) with no concurrent change in eEPSC amplitude (p = 0.103, paired t test). Black bars are means ± SEM, and black lines between points indicate paired experiments. All recordings were performed at 70 mV in the continuous presence of 10 mM Cd2+, bicuculline, and 1 mM strychnine. See also Figures S6 and S7. All error bars reflect standard errors. were firmly established in CFA-injected (p < 0.001, 3 days after CFA) or CCI-injuredmice (p < 0.01, 5 days after surgery), animals received a single peptide injection (10 mg/i.t.). i.t. delivery of the Tat-3.2-III-IV linker peptide significantly (two-way ANOVA) attenuated the mechanical hyperalgesia induced by CFA compared to animals injected with either the C terminus peptide or with III-IV linker peptides lacking the Tat epitope (Figure 7A). Remarkably, the effect of the Tat-3.2- III-IV linker peptide developed rapidly (over the time course of 30 min) and persisted for up to 6 hr (Figure 7A). Next, we determined if the Tat peptide altered endogenous Cav3.2 protein levels in DRG neurons from CFA-injected mice. Mice were treated with the Tat-3.2-III-IV linker or the Tat-3.2CT peptides 3 days after they had received CFA, and 90 min 1152 Neuron 83, 1144–1158, September 3, 2014 ª2014 Elsevier Inc. later their ipsilateral lumbar (L5) DRG were isolated and Cav3.2 levels analyzed by western blotting (Figures 7B and 7C). We found that total Cav3.2 levels were reduced by 50% in mDRGs from CFA-injected mice treated with the Tat-3.2-III-IV linker peptide (Figure 7C, lane 2) compared with those treated with Tat-3.2-CT peptide (lane 3) and to levels that were similar to those seen in altogether naive mice (lane 1). This robust reduction in total Cav3.2 protein in Tat-3.2-III-IV linker-treated DRG is consistent with the analgesic effect of the peptide. We note that there was no direct effect of the peptide on CFA-induced inflammation per se (i.e., we observed no changes in paw volume measured via a plethysmometer, paw diameter measured via a digital caliper, paw weight, and MPO activity; data not shown). (A) Blind analyses of the time course of mechanical hyperalgesia of CFA-injected mice treated with Tat-peptides or control peptides (10.0 mg/i.t.) compared to the control (intraplantar PBS). Ten micrograms of peptide correspond to 2.27, 2.22, and 3.5 nmoles for the Tat-3.2-III-IV, Tat-3.2-CT, and 3.2-III-IV peptides. (B) Cav3.2 immunoprecipitates from ipsilateral lumbar 5 (L5) DRGs from naive mice (lane1), or fromCFA-injectedmice treated with either Tat-3.2III-IV linker (lane2) or Tat-3.2-CT (lane3) peptides, analyzed bywestern blot 90min after i.t. treatment. A blot for a-tubulin is shown as control (bottom panel). (C) Quantification of Cav3.2 protein levels expressed as the ratio of Cav3.2 versus a-tubulin by densitometry. (D) Time course of mechanical hyperalgesia of mice treated with either theTat-3.2-III-IV linker or the Tat-3.2-CT peptides (10.0 mg/i.t.) under CCIinduced neuropathy. (E) Effect of the Tat-3.2-CT and Tat-3.2-III-IV peptides 15 min after i.t. delivery to animals that subjected to CCI 3 weeks prior to the experiment. The experimenter was blind to the treatment. In all panels, each circle/bar represents themean ± SEM (n = 5–8) and is representative of two independent sets of experiments. Statistical analyses were performed by two-way ANOVA followed by a Tukey test. The dashed line and hashtag indicate the range of data points where injured animals differed from the sham-treated group (p < 0.001). See also Figures S6 and S7. All error bars reflect standard errors. The Tat-3.2-III-IV linker peptide also produced rapid and long-lasting (up to 6 hr) antihyperalgesia when delivered intraperitoneally (i.p.) (15 mg/kg) compared to animals treated with the C terminus control (Figure S7A). The observed time course fits with observations by other groups that show uptake of Tat peptides into DRG and spinal cord tissue within 15 min of i.p. delivery (Brittain et al., 2011). No differences between Tat-3.2-III-IV linker and Tat-3.2-CT (control)-treated groups were found in naive (noninjured) mice (Figure S7B), indicating that the USP5mediated regulation of T-type channels occurs only during chronic pain states and not under basal conditions (in agreement with Figure 4). i.t. delivery (10 mg/i.t.) of the Tat-3.2-III-IV linker peptide also partially reversed mechanical hyperalgesia in CCI-neuropathic mice (Figure 7D). As in the CFA model, the effects of the peptide fully developed in less than 1 hr, although the effects did not last as long in the CCImodel. As shown in Figure 7E, the Tat-3.2-III-IV peptide was even able the reverse mechanical hyperalgesia under longer term chronic pain conditions (i.e., 3 weeks after CCI) within 15 min of delivery. To ensure that the effects of the Tat peptides were specific for Cav3.2, we carried out a series of experiments in Cav3.2-null mice. In response to CFA, these mice showed slightly higher Ne but not statistically significant mechanical withdrawal thresholds comparedwithWT animals (Figure 8A). However, the Cav3.2-null mice were completely insensitive to the Tat-3.2-III-IV linker peptide (Figure 8B). To ensure that the tested Tat peptides were in fact functional, their activity was verified in WT mice that were examined on the same day as the knockouts (in a blinded fashion). We also tested the effects of the Tat peptides in CFAtreated mice in the presence of the T-type channel blocking compound mibefradil (i.t.). We first assessed the dose and time dependence of the effects of mibefradil in CFA-treated mice to optimize the experimental conditions (data not shown). We then delivered a maximally effective dose of mibefradil together with either the Tat-3.2-III-IV or the Tat-3.2-CT peptide (10 mg i.t.) as a control (Figure 8C). While the Tat-3.2-III-IV peptidemediated a significant reversal of pawwithdrawal thresholds in vehicle-treatedmice, it did not augment the effects of mibefradil. Altogether, these data suggest that the USP5 mediates its effect on pain signaling via its action on Cav3.2 channels.
DISCUSSION
Here, we have presented evidence that ubiquitination and deubiquitination of Cav3.2 T-type channels is an important uron 83, 1144–1158, September 3, 2014 ª2014 Elsevier Inc. 1153 (A) Baselines andmechanical hyperalgesia ofWT or Cav3.2-null mice following CFA injection. Each column represents the mean ± SEM (n = 5 to 6) and are representative of two independent experiments. (B) Blind analyses of the time course of mechanical hyperalgesia of CFA-injected Cav3.2-null mice treated with the Tat-3.2-III-IV linker or the Tat-3.2-CT peptides (10.0 mg/i.t.). Each circle represents the mean ± SEM (n = 5–8), and data are representative of two independent sets of experiments. (C) Effects of Tat-3.2-III-IV (10.0 mg/i.t.) on mechanical withdrawal threshold in CFA-injected mice treated with mibefradil (10.0 mg/i.t.). Note the lack of Tat-3.2-III-IV effect in mibefradil-treated mice. Each circle represents the mean ± SEM (n = 9–13), and data are representative of two independent sets of experiments. The experimenter was blinded to the experimental conditions. 1154 Neuron 83, 1144–1158, September 3, 2014 ª2014 Elsevier Inc. mechanism for regulation of T-type channel activity in primary afferent neurons. Our data show that the Cav3.2 domain III-IV linker is a key regulatory site for channel stability. It interacts with the E3 ubiquitin ligases of the WWP family and the DUB USP5 and contains key ubiquitination sites. Hence, the domain III-IV linker of the channel is a hub for channel regulation that allows for control of channel activity/stability via ubiquitin modification (Figure S8). Our data show that WWP1 is able to ubiquitinate the channel, whereas USP5, but not a related DUB USP15, removes ubiquitin groups from the channel protein. This deubiquitination process then results in an increase in whole-cell current density, which is expected to be pronociceptive. Conversely, in vivo knockdown of USP5 with shRNA, or uncoupling USP5 from its binding site on the channel via delivery of Tat peptides, mediated analgesic effects in both inflammatory and neuropathic pain models. The Tat peptides did not affect basal nocifensive behavior in naive animals, indicating that under normal physiological conditions, Cav3.2 channels are either not regulated by USP5 or that T-type channels do not contribute significantly to basal nociception. It is likely that DUBs such as USP5 have many cellular targets (Komander et al., 2009), which could confound the interpretations of the in vivo studies, especially those involving USP5 knockdown. However, two lines of experiments support the idea that USP5 regulates pain via its modulation of T-type channels. First, uncoupling USP5 from Cav3.2 channels via Tat peptides directed to specific domain III-IV linker sequences mediated analgesia that was similar to that observed with USP5 knockdown. Second, the peptides were completely ineffective in Cav3.2 knockout mice and in mice treated with mibefradil. We thus conclude that the pronociceptive effect of USP5 is due to its regulation of T-type channels rather than another cellular target. Acute delivery of mibefradil was able to significantly attenuate mechanical hypersensitivity in the CFA model (see Figure 8C), consistent with a role of T-type channels in chronic pain conditions. On the other hand, Cav3.2 knockout mice showed a relatively normal withdrawal response to mechanical stimuli in the CFA model when compared to WT animals. The simplest explanation for this apparent discrepancy is the possible existence of compensatory mechanisms in the Cav3.2-null mice that maintain hypersensitivity in response to injury but which are insensitive to the USP5 interacting III-IV linker Tat peptide. In principle, T-type channels could contribute to signaling in afferent fibers via multiple mechanisms, including a contribution to neuronal firing properties (which fits with the Tat-3.2-III-IV sensitivity of T-type channel-mediated calcium entry into DRG cell bodies) and a participation in neurotransmitter release at dorsal horn synapses. Indeed, a recent study has revealed a contribution of T-type channels to glutamate release from synaptic terminals in dorsal horn lamina I and II and that this effect is enhanced in a model of diabetic neuropathy (Jacus et al., 2012). Our observation that the Tat-3.2-III-IV peptide inhibited Statistical analyseswere performed by two-way or three-way ANOVA followed by Tukey’s test. The dashed line and hashtag indicate the range of data points where injured animals differed from the sham-treated group (p < 0.001). All error bars reflect standard errors. spontaneous activity in a subpopulation of spinal cord synapses in lamina II fits with a mechanism involving a USP5 regulation of T-type channels at a subset of afferent terminals that contribute to hypersensitivity. The variability in the observed responses to nickel and the Tat-3.2-III-IV peptide is not surprising given that dorsal horn neurons form a highly heterogeneous population (Graham et al., 2007) that appears to receive inputs from both T-type channel-expressing and -nonexpressing afferent terminals. While we did not perform an in depth analysis of the colocalization of USP5 and Cav3.2 in specific subsets of neurons, our observation of preferential expression of USP5 in NF200-positive neurons (Figure 3E) could suggest that non-nociceptive fibers are able to trigger nocifensive responses via enhanced T-type channel activity during inflammatory and neuropathic pain conditions. Alternatively, it is possible that USP5 expression profiles in various DRG neuron populations may change in response to injury. At this point, we cannot rule out the possibility that Cav3.2 channels might contribute to synaptic inputs from excitatory interneurons in the dorsal horn and that these terminals are a target for the Tat-3.2-III-IV peptide. However, our data showing that evoked synaptic events were sensitive to the Tat-3.2-III-IV peptide support an involvement of synaptic terminals arising from primary afferents. The Tat-3.2-III-IV peptide also reduced T-type channel-mediated calcium entry into DRG neuron cell bodies. A USP5-mediated enhancement of Cav3.2 channels in cell bodies and along afferent axons would be predicted to mediate an enhancement of afferent fiber excitability (Yue et al., 2013), which could contribute to pain hypersensitivity. The stability of T-type channels is governed by a balance between ubiquitin ligase and DUB activity and the rate of degradation of ubiquitinated channels (Figure S8). Alterations of either one of these processes in response to nerve injury/inflammation could give rise to increased T-type channel protein levels in the plasma membrane. Indeed, knockdown of USP5 or treatment with the Tat-3.2-III-IV peptide is expected to shift this equilibrium to increase the fraction of ubiquitinated channels (note that the Tat-3.2-III-IV peptide sequence does not include the PY motif implicated in WWP1 interactions and should thus selectively uncouple USP5 from the channel). While little is known about the WWP family of ligases and their regulation, there is a growing body of literature showing that DUBs are under tight second messenger control and are activated by cellular pathways such as phosphatidylinositol 3-kinase (PI3K)/Akt that are upregulated during various pain states (Duan et al., 2012; Zhang et al., 2012). Our observation showing an enhanced expression of USP5 and an associated increase in its interaction with Cav3.2 channels in the ipsilateral dorsal horn of injured mice may hint at altered transcriptional regulation or trafficking of USP5 in the primary afferent pain pathway and perhaps in spinal cord interneurons. Nonetheless, our key finding is that T-type channel activity can be potently regulated by USP5 and that this regulation can be exploited to mediate analgesia in different forms of chronic pain. It is well established the ubiquitination and subsequent processing in the endoplasmic-reticulum-associated protein degradation (ERAD) complex is an important quality control mechanism that ensures expression of properly folded proteins (Römisch, 2005). This type of ubiquitination process is carried Ne out by a large family of endoplasmic reticulum (ER)-localized E3 ubiquitin ligases. For example, we have recently shown that Cav1.2 L-type channels are ubiquitinated in the ER and then processed in the ERAD complex (Altier et al., 2011). In contrast, several types of ion channels, including the epithelial sodium channel ENaC, are ubiquitinated by HECT ubiquitin ligases such as Nedd4 (Abriel et al., 1999; Staub et al., 1996, 1997) and then internalized and degraded. Our data show that the plasma-membrane-localized ubiquitin ligase WWP1 interacts with Cav3.2 channels in DRG neurons and that knockdown of WWP1 reduces ubiquitination and cell surface expression of these channels (see Figure 2). This suggests that Cav3.2 channels are regulated by ubiquitination directly at the cell surface similarly to previous observations with ENaC and Nav1.7. The observation that total protein levels were unaffected by depletion of WWP1 (Figure 2C) is thus likely due to the fact that less than 10% of Cav3.2 channels in tsA-201 cells are expressed at the cell surface (I.A.S and G.W.Z., unpublished data). This may also be true for cultured DRG neurons whose total Cav3.2 pool was only slightly increased by MG132 treatment (Figure 1B) even though MG132 stabilized ubiquitinated channels that are presumably at the plasma membrane. An effect on turnover of cell surface channels rather than ER export may explain the rapid time course of the physiological effects of i.t. delivered III-IV linker Tat peptides. This would not be without precedent, as USP2-45 has been shown to bind to and regulate ENaC, and USP10 has been reported to regulate the postendocytic sorting of CFTR channels (Bomberger et al., 2009; Fakitsas et al., 2007; Oberfeld et al., 2011; Ruffieux-Daidié et al., 2008). The particular significance of ion channel regulation by ubiquitination for pain signaling is underscored a recent study showing that reduced expression of Nedd4-2 in DRG neurons results in increased Nav1.7 expression and thereby increased pain hypersensitivity (Cachemaille et al., 2012). More than 90 different types of deubiquitinating enzymes have been identified in the humangenomewith varying subcellular distributions including nuclear, cytoplasmic, and endocytic compartments (Nijmanet al., 2005;Urbé, 2005). Their role in cell physiology has been widely recognized, and they are known to act on numerous cellular pathways to regulate protein stability and degradation. DUBs have been suggested as potential drug targets in a variety of pathological conditions such as cancer, inflammation, and neurological disorders (Daviet and Colland, 2008; Hussain et al., 2009; Shanmugham and Ovaa, 2008). Here, we present evidence that DUBsmay also be important regulators of pain processing in the primary afferent pain pathway by increasing T-type channel stability, which may open new therapeutic approaches for pain management. Disrupting the USP5/ T-type channel interaction for therapeutic purposes would specifically target a process that is involved in aberrant upregulation of channel activity, while presumably sparing normal channel function, thereby reducing the risk of adverse side effects.
EXPERIMENTAL PROCEDURES
Detailed methods for spinal cord slice recordings, calcium imagining, electrophysiology, tissue culture, RT-PCR, immunostaining, and in vivo pain models are provided as Supplemental Experimental Procedures. uron 83, 1144–1158, September 3, 2014 ª2014 Elsevier Inc. 1155 Drugs and Reagents The following drugs were used in the study: formaldehyde, complete Freund’s adjuvant (CFA), mibefradil (Sigma Chemical Company, Saint Louis), human biotin-Cav3.21556–1602, biotin-Cav3.21569–1586, scrambled biotinCav3.21556–1602-III-IV linker, biotin-Cav3.21860–1884-CT, Tat-Cav3.21569–1589-IIIIV, no-Tat Cav3.21569–1589-III-IV, Tat-Cav3.2-CT1860–1884, no-Tat-Cav3.2CT1860–1884 peptides (Genemed synthesis, Houston), and USP5 and USP15 shRNAs (Thermo Scientific, Open Biosystems). Recombinant long and short USP5 proteins where purchased from Enzo Life Sciences. When drugs were delivered by the i.p. route, a constant volume of 10 ml/kg body weight was injected. When drugs were administered i.t., volumes of between 2–10 ml were injected in order to adjust the appropriate dose of solution injected. Appropriate vehicle-treated groups were assessed simultaneously. Cav3.2 T-Type Channel Antibodies The following antibodies against Cav3.2 channels were used: H-300 (Santa Cruz Biotechnologies) is directed against amino acids 2174-2353 in the C terminus region and was used in the majority of our experiments, N-18 (Santa Cruz Biotechnologies) is directed against the N terminus region, and 555-10 (Novus Biologicals) is an antibody against the domain II-III linker region. No cross reactivity of the H-300 antibody was observed on western blots for tsA-201 cell lysate transfected with hCav3.1, and hCav3.3. Cav3.2 was also immunoprecipitated from WT and Cav3.2-null mouse tissue with the 555-10 antibody and western blots probed with H-300, revealing a specific ( 250 kDa) band in WT, but not null, mouse tissue. Plasmids Cav3.2 III-IV linker from Cav3.2 full-length EGFP-C1was subcloned into pBluescript II KS (+/ ) using Sal 1 ( 1.6kb)-BamHI. pBluescript II -Cav3.2 III-IV linker single mutants K1560R, K1576R, K1587, and Y1594N (human sequence) were generated by PCR mutagenesis. Cav3.2 III-IV linker mutants were subcloned into Cav3.2 isoform 2-pcDNA3.1 using Sal1-BamHI. WT Cav3.2 isoform 2 cDNA was kindly provided by Dr. T. Snutch. Nedd4, WWP1, and WWP2 shRNA plasmids were purchased from Santa Cruz Biotechnology, Inc. Immunoprecipitation and Coimmunoprecipitation Assays Cells were lysed in RIPA buffer (in mM: 50 Tris, 100 NaCl, 1% Triton X-100, 1% NP-40, 0.2% SDS, 0.1% Na Deoxycholate, 20 NaF, 10 Na4P2O7 pyrophosphate, and 10 EDTA + protease inhibitor cocktail [pH 7.5]). A modified RIPA buffer (in mM; 50 Tris, 100 NaCl, 0.2% [v/v] Triton X-100, 0.2% [v/v] NP-40, and 10 EDTA + protease inhibitor cocktail, with or without 5 mM MG132 [pH 7.5]) was used to coimmunoprecipitateCav3.2 channelswithUSP5protein. Lysates from tsA-201, CAD cells, and rat brain tissue were prepared by sonicating samples at 60% pulse for 10 s and by centrifugation at 13,000 rpm for 15 min at 4 C. Supernatants were transferred to new tubes, and solubilized proteinswere incubatedwith 50 ml of ProteinG/A beads (Pierce) and 1 mg of HA antibody (Roche) or 2 mg of anti-Cav3.2 (H-300, Santa Cruz Biotechnologies, Inc., unless stated otherwise), anti-USP5 (ProteinTech Group, Inc.), anti-WWP1, anti-WWP2 (Sigma), and anti-USP15 (Abnova) antibodies overnight while tumbling at 4 C. Total inputs were taken from whole-cell samples representing 4%of total protein and probed for actin. Immunoprecipitates werewashed twicewith (mM)500NaCl 50Tris (pH7.5) buffer andoncewith 150 NaCl 50 Tris (pH 7.5) buffer, and coimmunoprecipitates were extensively washed with modified RIPA buffer; beads were aspirated to dryness. Laemmli buffer was added and samples were incubated at 96 C for 7 min. Eluted samples were loaded on an 8% or 10% Tris-glycine gel and resolved using SDS-PAGE. Samples were transferred to 0.45 mm polyvinylidenedifluoride membranes (Millipore), and western blot analysis was performed using an anti-HA (Covance), anti-Cav3.2 (H-300, Santa Cruz Biotechnologies, Inc.), anti-ubiquitin (BD PharMingen), anti-actin (Sigma), anti-USP15 (Abnova), antiUSP5 (ProteinTech Group, Inc.), and anti-tubulin (Abcam) antibodies. Western blot quantification was performed using densitometry analysis (Quantity OneBioRad software). Student’s t tests for unpaired data were performed to determine statistical significance. In each coimmunoprecipitation experiment, beads-only and IgG controls were performed—for space reasons, such controls were not always included in the figures. 1156 Neuron 83, 1144–1158, September 3, 2014 ª2014 Elsevier Inc. Affinity Precipitation of Cav3.2-Interacting Proteins Total extracts from mouse DRG, dorsal horn, and rat brain were prepared by centrifugation at 16,100 g for 30 min in buffer containing the following (in mM): 50 Tris pH 7.6, 150 NaCl, 1% Triton X-100, 1% NP40, 10 EDTA, 10 EGTA, and protease inhibitor cocktail (Roche). Lysates were sonicated at 60% pulse for 25 s, and soluble proteins were collected by centrifugation at 16,100 3 g for 15 min at 4 C. Cav3.2-interacting proteins were collected by incubation with neutravidin-agarose beads (Thermo Scientific) while tumbling for 2 hr at 4 C with human Cav3.21556–1602, Cav3.21860–1884 or scrambled peptide for Cav3.21556–1602 covalently linked to aC-terminal biotin group (Genemed Synthesis Inc., San Francisco). Protein lysates were also incubated with neutravidin beads but no peptide as control. After extensive washing, bound proteins were analyzed by SDS-PAGE and visualized by Coomassie staining. Visible bands were excised and samples analyzed by MALDI/TOF-MS (Bruker Instruments Co., Bremen) and Nano-ESI-MS/MS on an API QSTAR-Pulsar (QSTAR, Applied Biosystems Div., Perkin-Elmer Corp., Foster City). Pain Models For the formalin test, animals received 20 ml of 1.25% formalin solution intraplantarly (i.pl.) in the ventral surface of the right hindpaw. Animals were immediately placed individually in observation chambers, and the time spent licking or biting the injected paw was recorded. A blinded experimenter observed animals individually from 0–5 min (neurogenic phase) and 15– 30 min (inflammatory phase). To induce inflammatory chronic pain, mice received 20 ml of Complete Freund’s Adjuvant (CFA) injected subcutaneously in the plantar surface of the right hindpaw (i.pl.). Control groups received 20 ml of PBS in the ipsilateral paw. Neuropathic pain was induced by loose ligatures of sciatic nerve. Mice were anaesthetized, and the right sciatic nerve was exposed at the level of the thigh. Proximal to the sciatic nerve trifurcation four loose ligatures were loosely tied around so that the epineural circulation was preserved. In sham-operated mice, the nerve was exposed but not injured. For additional detail, see Supplemental Experimental Procedures. Pain Behavioral Assays Mechanical hyperalgesia wasmeasured using a digital plantar aesthesiometer (DPA). Animals were placed individually in a small, enclosed testing arena on top of a wire mesh floor. The DPA device was positioned beneath the animal so that the filament was directly under the plantar surface of the ipsilateral hind paw. Each pawwas tested three times per session. Thermal hyperalgesia was examined by measuring the latency to withdrawal of ipsilateral hind paws on a focused beam of radiant heat of a Plantar Test apparatus. Animals were placed individually in a small, enclosed testing arena on top of a wire mesh floor. The device was positioned beneath the animal so that the radiant heat was directly under the plantar surface of the ipsilateral hind paw. For additional detail, see Supplemental Experimental Procedures. Data Analysis and Statistics For biochemical and electrophysiological analyses, data values are presented as mean ± SEM for n experiments. Statistical significance was determined using Student’s t test unless stated otherwise: *p < 0.05; ** p < 0.01; *** p < 0.001; NS, statistically not different. For behavioral analyses, data are presented as means ± SEM and evaluated by one-way, two-way, or three-way ANOVA followed by a Tukey’s test. A value of p < 0.05 was considered to be significant. (*p < 0.05; ** p < 0.01; *** p < 0.001). SUPPLEMENTAL INFORMATION Supplemental Information includes eight figures and Supplemental Experimental Procedures and can be found with this article online at http://dx.doi. org/10.1016/j.neuron.2014.07.036.
AUTHOR CONTRIBUTIONS
A.G.-C., V.M.G., and G.W.Z. designed the study and wrote the manuscript. G.W.Z. supervised the study. A.G.-C., V.M.G., P.S., N.W., I.S., V.H., C.B., J.H., L.C., A.P., M.D., A.F., and E.B. performed experiments and data analysis. C.B. contributed to molecular biology.
ACKNOWLEDGMENTS
This work was supported by grants to G.W.Z. from the Canadian Institutes of Health Research and to E.B. from the Agence Nationale de la Recherche (ANR-09-MNPS-037, ANR-08-NMPS-025) and the Association Francaise contres lesMyopathies (AFM). G.W.Z. is an Alberta Innovates-Health Solutions (AI-HS) Scientist and a Canada Research Chair. V.G., P.S., and N.W. were supported by AIHS Fellowships. A.F. is supported by an AFM Fellowship. We would like to thank N. Daniel Berger for help with the blinded studies, S. Laffray for the SNI animals used for the immunohistochemistry, and Dr. Peter Smith (University of Alberta) for helpful discussion on the spinal cord slice recordings. Accepted: July 10, 2014 Published: September 3, 2014
Article Images (0)
×![]()
Bioz is the world’s most comprehensive AI search engine for scientific experimentation. The Bioz search engine offers researchers billions of data-driven product, technique, and protocol recommendations. Bioz taps into the latest advances in Artificial Intelligence (AI) – including Natural Language Processing (NLP), Machine Learning (ML) and Generative AI to mine and structure hundreds of millions of pages of complex and unstructured scientific papers. By harnessing the power of AI, Bioz provides researchers with an unprecedented amount of summarized scientific experimentation knowledge right at their fingertips, ultimately helping to speed up drug discovery and the rate of success in finding cures for diseases. Bioz is used by over 15 Million researchers from over 17,000 different universities and companies in 196 countries.
