Conceived and designed the experiments: HHJ PJP. Performed the experiments: HHJ PJP. Analyzed the data: HHJ PJP. Contributed reagents/materials/analysis tools: HHJ PJP. Wrote the paper: HHJ PJP.
The authors have declared that no competing interests exist.
Dipeptidyl peptidase-like protein 6 (DPP6) proteins co-assemble with Kv4 channel α-subunits and Kv channel-interacting proteins (KChIPs) to form channel protein complexes underlying neuronal somatodendritic A-type potassium current (ISA). DPP6 proteins are expressed as N-terminal variants (DPP6a, DPP6K, DPP6S, DPP6L) that result from alternative mRNA initiation and exhibit overlapping expression patterns. Here, we study the role DPP6 variants play in shaping the functional properties of ISA found in cerebellar granule (CG) cells using quantitative RT-PCR and voltage-clamp recordings of whole-cell currents from reconstituted channel complexes and native ISA channels. Differential expression of DPP6 variants was detected in rat CG cells, with DPP6K (41±3%)>DPP6a (33±3%)>>DPP6S (18±2%)>DPP6L (8±3%). To better understand how DPP6 variants shape native neuronal ISA, we focused on studying interactions between the two dominant variants, DPP6K and DPP6a. Although previous studies did not identify unique functional effects of DPP6K, we find that the unique N-terminus of DPP6K modulates the effects of KChIP proteins, slowing recovery and producing a negative shift in the steady-state inactivation curve. By contrast, DPP6a uses its distinct N-terminus to directly confer rapid N-type inactivation independently of KChIP3a. When DPP6a and DPP6K are co-expressed in ratios similar to those found in CG cells, their distinct effects compete in modulating channel function. The more rapid inactivation from DPP6a dominates during strong depolarization; however, DPP6K produces a negative shift in the steady-state inactivation curve and introduces a slow phase of recovery from inactivation. A direct comparison to the native CG cell ISA shows that these mixed effects are present in the native channels. Our results support the hypothesis that the precise expression and co-assembly of different auxiliary subunit variants are important factors in shaping the ISA functional properties in specific neuronal populations.
The somatodendritic A-type potassium current (ISA) regulates neuronal excitability, firing frequency, action potential back-propagation, and synaptic plasticity
ISA channels are multi-protein conglomerates composed of Kv4 type voltage-gated potassium channel subunits (Kv4.1, Kv4.2, Kv4.3) surrounded by modulatory subunits from two distinct gene families: Kv channel-interacting proteins (KChIPs) and dipeptidyl peptidase-like proteins (DPLPs)
Cerebellar granule (CG) cells are an ideal neuronal population to investigate the role of DPP6 variants in shaping native ISA properties. CG cells are electrophysiologically compact and express large ISA. DPP6 is the only DPLP that is expressed in CG cells
In this study, we sought to determine how expression of multiple DPP6 variants shapes the various functional properties of the native ISA in CG cells. Our results show that the DPP6K variant constitutes the largest percentage of total DPP6 expressed in CG cells, followed by DPP6a, DPP6S, and DPP6L. Because of its large contribution to DPP6 transcripts, DPP6K was re-examined in co-expression studies with Kv4.2 and KChIP3a, and we found that in the heterotrimeric complex DPP6K does generate unique functional effects that were not seen in previous co-expression studies with Kv4.2 alone
The procedures on animals conducted in this work were performed in strict accordance with Animal Welfare Act, the Public Health Services Animal Welfare Policy, and The National Institute of Health Guide for Care and Use of Laboratory Animals. The experimental protocol was approved by the Institutional Animal Care and Use Committees (IACUC) of Baylor College of Medicine (Protocol Number: AN-752). Following approved protocol, every effort was made to minimize suffering.
Sprague-Dawley rats were anesthetized by isoflurane inhalation (IsoFlo, Abbott Laboratories), rapidly killed by cervical dislocation, and decapitated using a guillotine. Brain tissues from P21–28 rats were used for synthesizing cDNA used in amplification PCR or qRT-PCR. For
For slice recordings, the whole brain from young P8–10 rats was divided using a blade into the two hemispheres and cerebellum. The blocks of tissue were immersed immediately in normal artificial cerebrospinal fluid (ACSF) (in mM: 124 NaCl, 44 KCl, 2 CaCl2, 1.2 MgSO4, 1 NaH2PO4, 2.6 NaHCO3, 1 glucose) saturated with 95% O2/5% CO2 at room temperature (RT, 21–23°C). After 30 min, a small block of cerebellar tissue is properly oriented and adhered to the cutting tray of the vibratome (VT1000S, Leica Microsystems Inc.) using cyanoacrylate glue (Vetbond, 3M), and the tray is submerged in ice-cold cutting solution (in mM: 182.6 sucrose, 30 NaCl, 4.5 KCl, 1 MgCl2, 1.2 NaH2PO4, 1 NaHCO3, 1 glucose). Cerebellar slices (300 µm) were cut in the coronal plane by the vibratome's automated function and stored at room temperature on a mesh submerged under normal ACSF saturated with 95% O2/5% CO2. After a 1 hr recovery period, a slice was transferred to the recording chamber perfused with normal ACSF saturated with 95% O2/5% CO2 at RT.
A 5′RACE-PCR was conducted on RNA isolated from the cerebellum of a p28 rat using the First Choice RLM RACE Kit (Ambion) as described previously
To generate the normalization controls for qRT-PCR, cDNA fragments covering appropriate intron/exon boundaries (DPLPs: Exons 1–3; GAPDH: Exons 5–8) were cloned in series into pBluescriptII KS(+). Plasmids containing rat Kv4.2, human KChIP3a, and human KChIP4bL cDNAs were obtained as previously described
Cerebellar samples (10 mg each) were collected from P21 rats, and total RNA was isolated using the RNAqueous-Micro kit (Ambion). Reverse transcription (RT) reactions were conducted with the Superscript III kit (Invitrogen) using random hexamers and following manufacturer's instructions.
Procedures for quantitative real-time PCR were adapted from Liss et al. (2001)
External primers:
GAPDH (362 bp): (f)
Kv4.2 (406 bp): (f)
DPP10a (270 bp): (f)
DPP10c (263 bp): (f)
DPP10d (246 bp): (f)
DPP6a (239 bp): (f)
DPP6K (257 bp): (f)
DPP6L (353 bp): (f)
DPP6S (165 bp): (f)
Internal primers:
GAPDH (221 bp): (f)
Kv4.2 (309 bp): (f)
DPP10a (140 bp): (f)
DPP10c (176 bp): (f)
DPP10d (110 bp): (f)
DPP6a (143 bp): (f)
DPP6K (178 bp): (f)
DPP6L (125 bp): (f)
DPP6S (123 bp): (f)
qRT-PCR was conducted using IQ SYBR Green Supermix (Bio-Rad) on a PTC-200 thermal cycler with a chromo-4 detection system (Bio-Rad) as described previously
The two-electrode voltage clamp technique was used to elicit whole-cell currents from injected oocytes. The microelectrodes had <1 MΩ tip resistance and were filled with 3 M KCl solutions. The voltage-clamp amplifier (Oocyte Clamp OC-725, Warner Instruments) was under the control of WinWCP software (John Dempster, University of Strathclyde, Glasgow, UK). The data were digitized and low-pass filtered (model 902, Frequency Devices) at various frequencies depending on the sampling rate. The capacitative transient and linear leak were subtracted off-line by scaling up transients at voltages without ionic currents (at −70 mV) and subtracting them from total currents. Recordings with offsets >3 mV were removed from data analysis, and the average leak current was <0.2 µA.
CG cells were visualized using a fixed-stage upright microscope (BX51WI, Olympus) equipped with dedicated water-immersion objectives designed with long working distances (LUMPlanFL, 60x/0.90 W). Imaging of neurons was conducted using infrared (IR) and differential interference contrast (DIC) techniques. Patch pipettes were fabricated from thin-walled borosilicate glass capillaries (TW150-4, World Precision Instruments, Inc.) using a Flaming/Brown micropipette puller (model P-97, Sutter Instrument Co.) and had a resistance of 1–2 MΩ. An Axopatch 200B integrating patch clamp amplifier (Axon Instruments) was used to direct voltage pulses, and the data were acquired using pClamp 7 (Axon Instruments) program on a desktop personal computer.
Pipettes were fire polished and had areas near their tips wrapped with parafilm to reduce electrode capacitance. Tight-seal whole-cell recordings were obtained using standard techniques. The patch electrodes were backfilled with solution containing (in mM): 120 K-gluconate, 20 KCl, 5 NaCl, 10 HEPES, 4 Mg2-ATP, 0.3 Tris-guanosine 5′-triphosphate (GTP), 14 phosphocreatine, pH = 7.25 after adjustment with KOH; 301 mOsm). Series resistance and capacitance were determined by optimal cancellation of the capacitative transient, and the series resistance was typically 3–5 MΩ. If the series resistance changed more than 10% of the initial value during recording, the recording was terminated or the data was discarded. The current signals were recorded using an Axopatch 200B and filtered at 5 kHz and sampled at 10 kHz. The liquid junction potential between the electrode and bath solutions was calculated to be ∼−14.2 mV, and the data has been corrected for this voltage. All experiments were conducted at RT (22–23°C). Ohmic leak and capacitative currents were subtracted off-line using scaled-up versions of null traces at low voltages.
Non-ISA currents were suppressed by modifications of the external bath solution and by pharmacological agents. The modified ACSF bath solution was low in Ca (in mM: 84 NaCl, 4 KCl, 0.1 CaCl2, 3.1 MgSO4, 1 NaH2PO4, 1 NaHCO3, 1 glucose; 300–320 mOsm) to reduce contamination by Ca currents and Ca-dependent K currents. Supplementation with 0.5 µM tetrodotoxin (TTX), 40 mM tetraethylammonium (TEA), and 100 µM 4-aminopyridine (4-AP) (Sigma-Aldrich) respectively block Na currents, non-inactivating K currents, and D-currents. TEA supplementation is accomplished by substituting TEA for equimolar amount of NaCl. In some experiments, 100 µM linopirdine (Sigma-Aldrich) and 100 µM ZD7288 (Tocris) were used to block M-current and H-current, respectively.
In addition, ISA was reliably isolated using a subtraction protocol that exploits the difference between the high- and low-threshold K currents. Currents were recorded in response to depolarizing steps to various voltages from a conditioning step at either −114 mV or −44 mV for 1 sec. The ISA is inactivated by prepulse to −44 mV, and the subtraction of the current after the −44 mV conditioning pulse from the current after the −114 mV conditioning pulse reveals the transient outward current during the test pulse. Tail reversal measurements gave ISA reversal potential at −80 mV. Calculated EK is −88.3 mV.
Data were analyzed with WinWCP (John Dempster, University of Strathclyde, Glasgow, UK), Clampfit 6 (Axon Instruments), and Origin softwares (OriginLab Corp.). Peak conductance (Gp) was calculated as Gp = Ip/(Vc−Vrev), where Ip is the peak current, Vc is the command voltage, and Vrev is the reversal potential (−90 mV in ND96). The Gp-V curves were described using the first-order Boltzmann function: Gp/Gp,max = 1/(1+exp(Vm−V0.5a)/Sa)), where Gp/Gmax is the fraction of maximal conductance, Vm is the membrane potential, Va is the potential for half-maximal activation, and Sa is the slope factor. Steady-state inactivation was also described assuming a simple Boltzmann distribution, with the corresponding parameters of I/Imax, V0.5i, and Si. Using Clampfit, the time courses of inactivation were fitted by a sum of two exponential terms, and the time courses of recovery from inactivation were described using a single exponential term initially and two exponential terms when the fitted curve deviated from experimental data significantly.
The number of DPP6 subunits per channel was examined by measuring the slowing of DPP6a-mediated fast inactivation via DPP6a∶DPP6K mixing and comparing it to the predicted inactivation slowing when only one DPP6a subunit is present. DPP6a-mediated fast inactivation is markedly faster than DPP6K-mediated inactivation; therefore, significantly reducing the DPP6a contribution to the mix should discernibly slow fast inactivation. Calculations for the predicted maximum amount of slowing in fast inactivation assume the use of simple single-step inactivation models for both DPP6a- and DPP6K-mediated inactivation:
Property | Kv4.2+KChIP3a | ||||
+DPP6K | w/o DPP6 | +DPP6a | +DPP6S | +DPP6K/ΔN16 | |
|
(3) | (10) | (3) | (6) | (3) |
V0.5a (mV) | −20.1±0.8 | −17.4±0.9 | −26.5±3.3 | −25.7±1.2 |
−15.8±0.2 |
Sa (mV/e-fold) | 18.9±0.2 | 20.4±0.8 | 17.9±1.0 | 19.8±1.0 | 18.2±0.5 |
|
(5) | (25) | (5) | (5) | (3) |
V0.5i (mV) | −72.4±0.6 | −61.2±0.5 |
−63.3±0.5 |
−64.9±1.3 |
−60.2±3.3 |
Si (mV/e-fold) | 3.7±0.1 | 3.6±0.1 | 4.3±0.1 |
4.7±0.1 |
3.4±0.0 |
|
(3) | (6) | (8) | (4) | (3) |
τ-1 (ms) | 35.0±1.1 | 42.2±0.8 |
6.0±0.4 |
39.4±5.1 | 70.1±8.0 |
τ-2 (ms) | 189±14 | 192±3 | 461±23 |
177±8.0 | 156±1.6 |
W-1 (%) | 74±1 | 24±2 |
71±1 | 63±2 |
40±10 |
W-2 (%) | 21±0 | 69±2 |
22±1 | 30±1 |
51±13 |
W-ss (%) | 4±1 | 6±1 | 7±1 |
7±2 | 9±3 |
|
(5) | (9) | (4) | (4) | (5) |
τ-1 (ms) | 43±4.4 | 76.7±1.0 | 15.3±1.6 |
24.4±4.1 |
32.5±5.2 |
τ-2 (ms) | 265±22 | ||||
W-1 (%) | 29±3 | ||||
W-2 (%) | 72±3 |
See “Methods” for descriptions of parameters measured. τ-n and W-n represent the corresponding time constants and relative weights, expressed as percentages, derived from exponential fittings of the 1-sec traces. W-ss represents the relative weight of the steady-state component.
#, recovery at −100 mV measured with a prepulse of 1 sec.
p<0.05, relative to Kv4.2+KChIP3a+DPP6K.
p<0.05, relative to Kv4.2+KChIP3a.
Since DPP6 proteins are known to form dimers
For modeling the effects of DPP6K on steady-state inactivation, we used a linear model based on an equal energetic contribution for each incorporated DPP6K subunit. Based on the magnitude of the shift produced by four DPP6K subunits, the model assumes a −2.3 mV shift in the steady-state inactivation curve for each DPP6K subunit incorporated in the channel. To confirm that this model gives a linear shift in the Boltzmann midpoint based on the DPP6K mole fraction, we constructed a set of Boltzmann curves based on the DPP6a curve with a −2.3×NK mV shift in the DPP6a curve, where NK is the number of DPP6K subunits in the mixed tetramer. These curves were then summed using the calculated mole fractions for the different subunit compositions based on the expression mix used and then the summed data was fit with a first-order Boltzmann distribution. For experimental averages, data are presented as mean ± standard error of the mean (SEM). Statistical significance was determined by comparing data sets using Student's two-tailed (independent)
To determine the relative expression of different DPP6 variants in CG cells, we quantified the mRNA expression levels for DPP6 variants using qRT-PCR. PCR primers were designed to amplify variant-specific DNA sequences for each of the four DPP6 N-terminal variants expressed in rat brain (
A) DPP6 gene shows a conserved set of four alternative first exons producing the protein variants DPP6a, DPP6K, DPP6L and DPP6S. B) qRT-PCR using SYBR Green Fluorescence for DPP6 variants from three brain regions: cortex, cerebellum and hippocampus. DPP6a and DPP6K show enhanced expression in cerebellum. C) Normalization controls used to correct for primer amplification efficiency differences. Amplification targets from GAPDH and two DPLP variants were diluted and used to construct amplification curves. Common GAPDH signal was used to ensure consistent dilution of standards. D) Relative expression levels of DPP6 variants in cerebellum following normalization. Due to high levels of expression in CG cells and the abundance of these neurons, these signals essentially report the relative expression of DPP6 variants in CG cells.
Therefore, we constructed a set of standardizing DNA constructs to precisely analyze the amplification efficiency of different primer sets. Each of the constructs consists of Exon 5–8 of GAPDH, followed in tandem by Exons 1–3 of two DPLP variants. For example, one construct contains the Exons of GAPDH, DPP10a, and DPP6a (
The high level of DPP6K in CG cells raised the prospect that this variant plays a special role in shaping ISA properties in CG cells. A previous study of DPP6K did not identify any significant unique functional effect when co-expressed with Kv4.2
The stabilization of inactivation in DPP6K channels could be due to favoring entry into the inactivated state or slower recovery from inactivation, or both. To measure the effects of DPP6K on recovery from inactivation, we examined recovery from inactivation at −100 mV using a two-pulse protocol and plotted the fractional recovery as a function of the interpulse duration (
Representative current traces generated during the two-pulse protocol used to measure recovery from inactivation for Kv4.2+KChIP3a (A), Kv4.2+KChIP3a+DPP6a (B), Kv4.2+KChIP3a+DPP6S (C), and Kv4.2+KChIP3a+DPP6K (D). A 1-sec depolarization to +50 mV was delivered to maximally inactivate the channels, followed by an increasing recovery interval at −100 mV before applying a 250-ms test pulse at +50 mV to check the degree of recovery from inactivation. (E) Fractional recovery was plotted as a function of the interval duration at −100 mV. The residual value at the end the first pulse was subtracted from the peak current values of the first and second pulses, and the fractional recovery was determined by dividing the peak value of the second pulse by that of the first pulse.
To determine whether it is the presence of the DPP6K N-terminus rather than the absence of DPP6a or DPP6S N-termini that is responsible for the DPP6K-differential effects, we deleted the DPP6K Exon 1K (E1K,
(A) Voltage dependence of steady-state inactivation of ternary complex containing DPP6K and DPP6K/ΔN16 deletion mutant. The protocol was the same as that of
Given that the DPP6K N-terminus is responsible for its unique functional effects, we sought to determine if it has been significantly conserved over evolutionary time. We examined DPP6 genes from a large number of species and found an obvious DPP6 Exon 1K sequence in vertebrate genomes from teleost and coelacanth fish to humans. The alignment of these sequences shows extremely high levels of sequence conservation over this long evolutionary time span with at most 2 non-homologous substitutions out of 17 positions for the most divergent teleost fish sequences (
Based on our qRT-PCR results, nearly 75% of all DPP6 transcripts in CG cells consist of either DPP6a or DPP6K, both of which separately confer distinct functional properties. An important question is: What are the functional consequences of forming channels with both DPP6a and DPPK in the same channel complex? Previous study has suggested that four DPP6 proteins may be co-assemble onto a Kv4-based channel, although DPP6 in solution tends to form dimers
(A) Outward currents expressed by oocytes co-injected by various combinations of cRNAs, as elicited by depolarization to +40 mV from holding potential of −100 mV. (B) Expected rise and decay of currents if DPP6a and DPP6K subunits do not co-assemble and produce segregated channel populations containing either one alone. (C) Slowing of the time constant of fast inactivation when DPP6a mRNA changes from 100% to 10% mixed with DPP6K mRNA. To get the average value for fast inactivation, the slow phase of inactivation and non-inactivating current were described by exponential fitting and subtracted from the total current. The remaining average fast inactivation time constant was measured by taking the peak current for the fast inactivating fraction divided by its area. The average time constant measured by this method was very similar to the time constant measured by the best single exponential fit to the fast inactivating component. The black and gray lines show the predicted maximal slowing of fast inactivation with four DPP6 and two DPP6 per channel, respectively, with only 1 DPP6a subunit per channel. (D) Recovery from inactivation at −100 mV after a 200 ms-long prepulse (symbols) as compared to predicted results assuming no co-assembly of DPP6a and DPP6K (dashes).
Property | CG cell ISA | Kv4.2+KChIP3a+ | Kv4.2+KChIP4bL+ | ||||
DPP6a∶DPP6K ratio | (1∶1) | (1∶2) | (1∶3) | (1∶1) | (1∶2) | (1∶3) | |
|
(7) | (3) | (4) | (5) | (3) | (3) | (4) |
V0.5a (mV) | −25±3 | −25.6±1.7 | −25.6±1.7 |
−22.9±0.9 | −31.0±1.7 | −30.9±0.1 |
−34±1.5 |
Sa (mV/e-fold) | 19±1.4 | 18.3±0.7 | 16.4±0.7 |
17.2±0.4 | 12.4±0.2 | 12.1±0.1 |
12.2±1.2 |
|
(9) | (3) | (6) | (11) | (3) | (7) | (9) |
V0.5i (mV) | −76±1 | −66.0±0.4 | −68.8±0.7 |
−69.6±0.5 | −64.4±0.2 | −70.6±0.7 |
−68.4±0.8 |
Si (mV/e-fold) | 8.4±0.5 | 4.5±0.0 | 4.5±0.1 |
4.5±0.1 | 4.5±0.2 | 4.5±0.2 |
4.2±0.2 |
|
(6) | (3) | (3) | (5) | (3) | (3) | (4) |
τ-1 (ms) | 11±1 | 7.4±0.2 | 6.9±0.4 |
8.5±0.1 | 6.9±0.1 | 5.1±0.2 |
5.5±0.2 |
τ-2 (ms) | 120±16 | 131±6 | 107±3 |
122±3 | 90±4 | 67±3 |
63±2 |
W-1 (%) | 72±2 | 69±0 | 70±2 |
65±1 | 62±1 | 71±1 |
70±0 |
W-2 (%) | 25±3 | 15±1 | 16±3 |
25±1 | 25±1 | 16±1 |
17±1 |
W-ss (%) | 3±1 | 16±1 | 13±1 |
10±0 | 12±2 | 12±0 |
12±1 |
|
(3) | (3) | (3) | (6) | (3) | (3) | (3) |
τ-1 (ms) | 8±1 | 11.0±0.1 | 10.6±0.4 |
14.5±0.9 | 11.3±0.9 | 14.3±0.9 |
12.7±0.5 |
τ-2 (ms) | 68±25 | 199±1.7 | 156±26 |
207±15 | 111±22 | 145±26 |
159±36 |
W-1 (%) | 80±7 | 91±1 | 88±3 |
79±2 | 76±3 | 63±3 |
78±1 |
W-2 (%) | 20±7 | 9±1 | 12±3 |
21±2 | 24±3 | 37±3 |
22±1 |
See “Methods” for descriptions of parameters measured. τ-n and W-n represent the corresponding time constants and relative weights, expressed as percentages, derived from exponential fittings of 200 ms-long pulse of both native and reconstituted currents. W-ss represents the relative weight of the steady-state component.
, recovery at −100 mV measured with a prepulse of 200 ms.
p>0.05.
p<0.05.
P-values were determined relative to CG cells.
To further investigate whether DPP6a and DPP6K co-assemble as four- or two-DPP6 subunits per Kv4.2 channel complex, simple kinetic models were developed to determine how much slowing of the fast inactivation time constant would occur if DPP6a expression was dramatically reduced in our DPP6a∶DPP6K co-expression studies. We showed previously that DPP6a inactivates by an N-type mechanism
With a 1∶9 ratio of DPP6a∶DPP6K, the resulting current exhibited a peak conductance-voltage relationship with a midpoint of −16.1±2.8 mV and slope factor of 20±1.1 mV/e-fold (n = 3). With the large excess of DPP6K, the time constant of the fast phase of inactivation was slowed to 10.8±0.5 ms (n = 4). With only one DPP6a subunit, Model #1 predicts a fast time constant of 11.6 ms for a four DPP6 per channel model and 8.7 ms for a two DPP6 per channel model. Model #2 predicts a fast time constant of 11.6 ms and 8.75 ms for the four DPP6 per channel and two DPP6 per channel with only one DPP6a subunit assembled. The observed slowing is significantly greater than the maximum amount of slowing predicted for both two DPP6 per channel models (P = 0.03). Compared to the four DPP6 per channel models, the measured time constant is slightly less than the predicted maximum but not significantly different (P = 0.14). Given that in this sort of mixing experiment some channels will have multiple DPP6a subunits, we would expect our measured fast time constant to be slightly less than the predicted 11.6 ms. In addition, we find that the amount of inactivation occurring by the fast N-type mechanism with 10% DPP6a cRNA is greater than predicted, with the observed relative weight of ∼46% significantly greater than the predicted values of 14% for heterodimer and 24% for heterotetramer. This suggests that channels with DPP6a are over-represented in our recordings by a factor of about two. Nevertheless, our data supports the model that there are most likely four DPP6 subunits per channel since this overrepresentation of DPP6a would only make the measured time constant additionally faster than predicted.
To complement our analysis of inactivation kinetics, an examination of the effect of mixing DPP6a and DPP6K on the kinetics of recovery from inactivation also clearly showed the effects of heteromultimeric assembly of DPP6 variants into the Kv4 channel complex. Kinetics for recovery from inactivation with mixed expression of DPP6a and DPP6K are dominated by the rapid recovery from DPP6a induced N-type inactivation (
The most consistent impact of DPP6K is on the steady-state inactivation properties for the channel. If we compare the maximum current in response to a depolarization to +40 mV, we see that holding at −65 mV reduces the amplitude of current evoked from DPP6a containing channel in half. As the ratio of DPP6K is increased, this steady-state inactivation at −65 mV increases until only about 10% current can be evoked with DPP6K alone (
(A) Representative traces for Kv4.2+KChIP3a channels co-expressed with DPP6a alone, DPP6K alone, or with a DPP6a∶DPP6K mixture at 1∶1 or 1∶2 ratios, showing changes in steady-state inactivation at −65 mV. For the colored traces, the channels were held for 30 sec at −65 mV before pulsing to +40 mV for 250 ms to test available current. The black traces show the total currents, from test pulses where the channels were held at −100 mV and experienced no inactivation. (B) Voltage dependence of steady-state inactivation of ternary complexes with homotetrameric and heterotetrameric DPP6 subunits. (C) Progressive shifting of inactivation midpoint with increasing DPP6K ratio. The V0.5i values were plotted against the calculated DPP6K mole fraction. Model assumes independent energetic effects for each DPP6K subunit incorporated into the channel, with symbols measured V0.5i values for summed Boltzmann curves from the model.
To better understand the role that co-assembly of DPP6a and DPP6K plays in shaping the native ISA from CG cells, we compared heterologously expressed currents in
In recordings using whole-cell patch clamp from acute cerebellar slices, we found that the native ISA of CG cells rises rapidly and decays with a clearly evident two exponential time course, in agreement with previous reports
(A) Outward transient currents elicited from CG cells and oocytes expressing Kv4.2, a mixture of DPP6a and DPP6K at 1∶2 ratio, and either KChIP3a or KChIP4bL. From a holding potential of −100 mV, either a 200-ms (CG cells) or 1-sec (oocytes) step depolarizations were made from −100 mV to +40 mV at 10 mV increments. (B) Overlapped normalized current traces at +40 mV from the indicated channels. (C) Time constants of inactivation at indicated membrane potentials for ISA from CG cells, Kv4.2+KChIP3a+DPP6a+DPP6K (1∶2), and Kv4.2+KChIP4bL+DPP6a+DPP6K (1∶2). (D) Recovery from inactivation at −100 mV, measured using the two-pulse protocol. (E) Normalized peak conductance-voltage relations (Gp/Gp,max) and steady-state inactivation curves (I/Imax) for ISA from CG cells and reconstituted channel complexes.
When Kv4.2 is co-expressed in oocytes with either KChIP3a or KChIP4bL and with DPP6a and DPP6K at a 1∶2 ratio, implementation of voltage protocols similar to the one used for CG cells elicited outward currents that activated and inactivated rapidly (
We next compared native ISA and reconstituted channels composed of Kv4.2, KChIP3a, and a mixture of DPP6a and DPP6K at a 1∶2 ratio for their kinetics of recovery from inactivation. As
For voltage-dependent gating parameters, we examined the peak conductance-voltage (Gp-V) relationship and the voltage dependence of steady-state inactivation (
Finally, we also examined the possibility that the native ISA channel may contain KChIP4 auxiliary subunits rather than KChIP3a. Therefore, KChIP4bL was co-expressed in oocytes with Kv4.2 along with DPP6a and DPP6K at 1∶2 ratio. Overall, the results showed that ternary complex channels formed with KChIP4bL are more different from the native ISA channels than those expressed with KChIP3a, but the exact kinetic and steady-state gating properties are sufficiently different from CG cell ISA to suggest that KChIP3a is the more likely auxiliary subunit for most native channels. In
We have determined the expression levels of DPP6 N-terminal variants in CG cells and examined their potential impact on the properties of native ISA. As a result, it was discovered that DPP6K is the most common DPP6 variant in these neurons, and that the DPP6K variable N-terminal domain has evolved to markedly slow down recovery from inactivation and leftward shift the steady-state inactivation of ISA. In terms of channel recovery, DPP6K appears to act as a competitive inhibitor of the accelerating effects of KChIPs and DPLPs on recovery from inactivation. Kv4.2+KChIP3a channels co-expressed with DPP6K recover more slowly than those with DPP6S or DPP6a, and moreover, they recover more slowly than a binary complex of Kv4.2+KChIP3a. However, DPP6K's effects are limited by the co-expression of DPP6a: there is less pronounced slowing of recovery from inactivation and a less pronounced shift in the steady-state inactivation curve. On the other hand, DPP6K still significantly accelerates a slow phase of inactivation as well as increases a slow component of recovery from ISA inactivation. By comparing reconstituted channels in heterologous expression systems with those of native ISA channels, our results suggest that the subunit composition of ISA channels is carefully controlled in order to finely tune the properties of neuronal ISA.
Previously, we demonstrated that the DPP6a variant, along with its paralog DPP10a, utilizes its N-terminus encoded in Exon 1a to confer ultra-fast inactivation via a pore-blocking N-type mechanism
In this paper we discovered that the N-terminal domain of DPP6K also confers unexpected unique functional properties. In the initial paper describing DPP6K, Nadal and colleagues (2006) described only moderate effects on Kv4.2 channels associated with the DPP6K variant that were not affected by truncation of the entire variable DPP6K N-terminus
In addition to slowing recovery from inactivation of Kv4.2+KChIP3a channels, DPP6K is also associated with a large hyperpolarizing shift in steady-state inactivation that is greater than those produced by DPP6S or DPP6a. Interestingly, the ΔN16 truncation that eliminates the DPP6K-mediated slowing of recovery also removes the hyperpolarizing shift in steady-state inactivation. This co-regulation by the DPP6K N-terminus is reasonable since slowing the rate of recovery from inactivation will favor the inactivated state and leftward shift the steady-state inactivation curve, assuming that the rate of low-voltage inactivation does not change. Further experiments will be required to elucidate the exact molecular mechanism by which DPP6K's modulatory domain disrupts recovery from inactivation and thereby hyperpolarizes the steady-state inactivation.
Our results showed that DPP6K and DPP6a are the two main DPP6 variants expressed in CG cells, with DPP6a constituting about 33% of total DPP6 transcripts. The finding that DPP6K is the most highly expressed DPP6 variant in cerebellum appears consistent with previous reports that show, among the four DPP6 variants expressed in CG cells, only DPP6K transcripts markedly exhibit a robust rostral-to-caudal gradient of decreasing level of expression
The reconstitution of ISA from CG cells in oocytes showed that in ternary Kv4 channel complexes when DPP6a is present, the DPP6a-mediated N-type inactivation dominates. The strength of N-type inactivation over overall inactivation kinetics was also tested previously when Kv4.2 and DPP10a were co-expressed with KChIP4a, a KChIP4 variant that dramatically suppresses inactivation and KChIP core-mediated acceleration of recovery from inactivation
Consequently, DPP6K controls the slower inactivation and slower recovery from inactivation processes in a manner consistent with the native properties of the native ISA from CG cells. Because the other DPP6 variants do not produce similar effects on slow inactivation and slow recovery (data not shown), these results suggest that DPP6K has a critical role in shaping the time- and voltage-dependent inactivation of native ISA in CG cells.
To the best of our knowledge, this study represents the first time a reconstituted ISA channel has been assembled using a combination of native subunits for the heterotrimeric complex found in a specific neuron type and compared side-by-side with its native counterpart. Previous reconstitution experiments had made comparisons to published results without focusing on a particular neuron
The values we obtained for various ISA functional properties are comparable to those reported in the literature. For activation, the midpoint and slope factor values are similar to the previously reported values, where V0.5a ranged from −15 to −27 mV and Sa ranged from 19 to 22 mV/e-fold, depending on the location of the cell
The functional properties of heterologous channel complexes composed of Kv4.2, KChIP3a, DPP6a, and DPP6K (DPP6a∶DPP6K at 1∶2 ratio) closely approximate those of native ISA channel. The most marked differences between the reconstituted and native channels are in the midpoint and slope factor of steady-state inactivation. One possible explanation for the detected difference is that the recorded native ISA was contributed mostly by channel complexes containing Kv4.3 instead of Kv4.2. Both Kv4.2 and Kv4.3 transcripts and proteins are expressed in the granule cell layer of the cerebellum but with reciprocal rostrocaudal gradients, where Kv4.2 level is highest at the anterior lobules and Kv4.3 level is highest at the posterior lobules
Alternatively, the differences in steady-state inactivation may be due to the influence of additional modulatory or associated proteins. A potential candidate is the T-type calcium channel, which according to published reports can interact with Kv4.2 ternary channel complexes and specifically modulate steady-state inactivation by rightward shifting the inactivation midpoint
Our results also show that the precise variant or type of KChIP present also has influence over the ISA channel properties. When KChIP3a is substituted by KChIP4bL in the complex, the reconstituted channel no longer resembles the native channel in all parameters examined, notably with the kinetics of recovery becoming significantly slower and the Gp-V curve shifting 10 mV in the hyperpolarizing direction. The reason behind the effects of different KChIP N-terminal variants in the ternary complex is unclear; however, it is likely to be complex and may derive from different N-terminal structures such as the presence of transmembrane segments and fatty acid modification (palmitoylation, myristylation)
In this paper, we have simplified the composition of expressed subunits down to Kv4.2, DPP6a, and DPP6K (at their experimentally determined ratio) with either KChIP3a or KChIP4bL, even though a large number of ISA subunits are reportedly expressed in CG cells, including Kv4.2, Kv4.3, KChIP3a, KChIP1, KChIP3a, KChIP3b (KChIP3x), KChIP4a, KChIP4b, KChIP4bL,KChIP4d, KChIP4e, DPP6a, DPP6S, and DPP6K
The kinetic and voltage-dependent properties of ISA are important determinants for the excitability and firing properties of neurons, and our study suggests that these ISA properties are dependent on the precise KChIP and DPLP variants that are expressed in specific neurons. In CG cells, multi-exponential inactivation and recovery from inactivation reflect the presence of DPP6a and DPP6K variants in the ternary channel complex. However, in other cell types where DPP6K is expressed without DPP6a, DPP6K may have even greater influence over ISA properties. For example, Nadal et al. (2006) reported that most of the DPP6 transcripts in the globus pallidus are contributed by DPP6K. The ISA from globus pallidus shows recovery from inactivation that is bi-exponential at −95 to −100 mV, with a fast time constant of ∼60 ms and a slow time constant of greater than 360 ms
Variants of auxiliary subunits are regulated by post-translational modifications such as phosphorylation and oxidation/reduction reactions, and the presence of specific variant may confer such regulatory effect. For example, Kvbeta1 confers rapid N-type inactivation to non-inactivating Kv1 channels, and oxidation and reduction of a cysteine at the Kvbeta1 N-terminus (C7) regulates this N-type inactivation
In conclusion, our results identify both DPP6a and DPP6K as important functional modulators of the native ISA channel in CG cells. For the initial phase of inactivation and recovery from short depolarizations, the rapid N-type inactivation domain of DPP6a is expected to dominate. However, for setting of the inactivation midpoint and inactivation responses to fluctuations in resting membrane potential, DPP6K is expected to play a critical role because the N-terminus of this variant contains a potent modulator of channel recovery from slower conformationally driven inactivated states. Finally, analysis of the specific effects of different KChIP proteins in the ternary channel complex suggests that KChIP3a is likely a key contributor in shaping the properties of the CG cell native ISA channel.
(TIF)
We thank Melissa K. Schaefer for her assistance in oocyte collection and other technical assistance.