The Schizophrenia-Associated Kv11.1-3.1 Isoform Results in Reduced Current Accumulation during Repetitive Brief Depolarizations

Recent genome wide association studies identified a brain and primate specific isoform of a voltage-gated potassium channel, referred to as Kv11.1-3.1, which is significantly associated with schizophrenia. The 3.1 isoform replaces the first 102 amino acids of the most abundant isoform (referred to as Kv11.1-1A) with six unique amino acids. Here we show that the Kv11.1-3.1 isoform has faster rates of channel deactivation but a slowing of the rates of inactivation compared to the Kv11.1-1A isoform. The Kv11.1-3.1 isoform also has a significant depolarizing shift in the voltage-dependence of steady-state inactivation. The consequence of the altered gating kinetics is that there is lower current accumulation for Kv11.1-3.1 expressing cells during repetitive action potential firing compared to Kv11.1-1A expressing cells, which in turn will result in longer lasting trains of action potentials. Increased expression of Kv11.1-3.1 channels in the brain of schizophrenia patients might therefore contribute to disorganized neuronal firing.


Introduction
Schizophrenia is a severe mental disorder that affects between 0.5% and 1% of the population [1]. It is associated with a myriad of symptoms including psychosis, motivational impairment, affective dysregulation as well as cognitive abnormalities [2,3]. The complexity and variability of the disease makes patient management very difficult. The causes of schizophrenia remain to be determined, however there is clearly a significant genetic contribution [4]. The gene candidates that have been identified to date [5] account for only a very small proportion of the known genetic contribution [6]. Consequently, there are significant ongoing efforts to identify new genes and to explore the possibility that the genes identified to date may open up novel targets for therapy.
In a recent study, single nucleotide polymorphisms (SNPs) in the second intron of the KCNH2 gene on chromosome 7q36.1 were shown to be significantly associated with an increased risk for development of schizophrenia [7]. The SNPs were replicated in a separate Caucasian case control population [7] and confirmed in Turkish [8] and Japanese [9] population studies. KCNH2 encodes for Kv11.1, a voltage-gated potassium channel, previously referred to as human ether-à-go-go-related gene (hERG). Kv11.1 channels have been extensively characterized, as they play a central role in repolarization of the cardiac action potential. Huffaker et al. showed that the SNPs, identified in the second intron of KCNH2 gene, promote transcription from an alternative transcription start site and the expression of a primate and brain specific Kv11.1 potassium channel isoform referred to as KCNH2-3.1 or Kv11.1-3.1. In the 3.1 isoform the first 102 amino acids of the full length Kv11.1-1A isoform are replaced with six unique amino acids [7].
Kv11.1 channels have unusual gating properties, most notably slow activation and deactivation kinetics but very rapid and voltage dependent inactivation and recovery from inactivation kinetics [10]. The amino-terminal region of Kv11.1 is crucial for determining the slow deactivation kinetics of the channel [11][12][13]. Unsurprisingly then, the 3.1 isoform shows faster deactivation than the full length Kv11.1-1A isoform. However, to appreciate the differences in the effect that the 3.1 and 1A isoforms will have on the electrical properties of neurons we also need to know how the N-terminal truncation affects activation, inactivation and recovery from inactivation as well as deactivation.
In this study we show, using the whole-cell voltage-clamp technique, that in addition to faster deactivation kinetics, Kv11.1-3.1 channels have a significant depolarizing shift in steady-state inactivation compared to Kv11.1-1A with heterotetrameric Kv11.1-1A/Kv11.1-3.1 channels having an intermediate phenotype. The altered gating of the Kv11.1-3.1 channels has minimal effect on single short depolarization pulses, however, it results in substantially less accumulation of current during repetitive stimuli, which would result in adaptation of action potential firing rates in response to prolonged stimuli. Thus expression of the Kv11.1-3.1 isoform will result in little difference in action potential firing patterns at rest but will have significant effects during repetitive firing.

Electrophysiology
Electrophysiological studies reported in the main manuscript were performed at 37uC. Additional experiments recorded at room temperature are described in the supporting information (Results S1, Figures S1-S4). Cells were trypsinized and plated onto glass cover slips. Glass capillary patch electrodes, with resistances of 2-4 MV when filled with internal solution, were made using a vertical two-stage puller (PP-830, Narishige, Tokyo, Japan). The internal solution contained (in mM): 120 potassium gluconate, 5 EGTA, 10 HEPES, 20 KCl, 1.5 Mg-ATP, pH 7.3 with KOH. Cells were superfused with external solution containing (in mM): 1 MgCl 2 , 1 CaCl 2 , 10 HEPES, 12.5 Glucose, 5 KCl, 130 NaCl, 0.1% dimethyl sulfoxide (DMSO), pH 7.4 with NaOH. The calculated junction potential for these solutions was 215 mV, which was corrected for in all experiments. Cells were voltage clamped in whole cell mode using an Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA). Current signals were digitized at 5 kHz and filtered at 1 kHz (except for inactivation protocols, which were digitized at 10 kHz and the short depolarization pulses, which were digitized at 100 kHz, and filtered at 5 kHz) and stored on an IBM-compatible PC interfaced with a Digidata 1440A analog-digital converter (Molecular Devices, Sunnyvale, CA). Series resistance was compensated by at least 80% in all experiments.

Voltage Protocols
All voltages protocols used are illustrated on top of each panel. Rates of activation were measured using an envelope-of-tail protocol [15]. Cells were depolarized from a holding potential of 280 mV to 0 mV for variable durations before stepping to 2120 mV where tail currents were recorded. To measure rates of deactivation, cells were first depolarized to +20 mV for 500 ms to fully activate the channels. Cells were then repolarized to voltages in the range 260 to 2130 mV for 5 s and exponential functions fitted to the decaying portion of the current traces. After each sweep, cells were hyperpolarized to 2120 mV for 100 ms to ensure full channel deactivation before reverting to the holding potential of 280 mV. This protocol was also used to measure the rates of recovery from inactivation although the voltage range was extended to voltages in the range 210 to 2130 mV. Rates of inactivation were measured during a triple pulse protocol [15]. Cells were depolarized from a holding potential of 280 mV to +40 mV for 500 ms, repolarized to 280 mV for 10 ms, followed by voltage steps to membrane potentials in the range of +60 to 230 mV.

Voltage-dependence of Steady-state Activation
Voltage-dependence of steady-state activation was measured using an isochronal tail current protocol [16]. Cells were depolarized from a holding potential of 280 mV, to voltages in the range of 2100 to +40 mV for 4 s before stepping to 260 mV to record tail currents. Between sweeps, cells were held at 2120 mV for 3 s to fully deactivate all channels. Steady-state activation curves were fitted with a modified Boltzmann equation: where V 0.5 is the half-activation voltage, V t is the test potential, k is the slope factor, g max is maximum conductance and g min is the minimum conductance.

Voltage-dependence of Steady-state Inactivation
Several different methods to measure steady-state inactivation have been described in the literature [17]. Most methods rely on a two pulse protocol and extrapolation of a deactivating Kv11.1 current back to the time of the voltage step. These methods work well for the 1A isoform of Kv11.1, but due to the much faster deactivation of the Kv11.1-3.1 isoform, these extrapolation methods do not provide very accurate estimates for Kv11.1-3.1. We therefore chose to use a method that relies on measuring the voltage dependence of the rates of inactivation and recovery from inactivation, and from these calculate the equilibrium constant for inactivation [18]. Specifically, a plot of the measured rates of inactivation and recovery from inactivation versus voltage gives rise to a chevron plot, which can be fitted with the equation: where k obs,v is the measured rate constant at any given voltage, k inact,V and k rec,v are the unidirectional forward and reverse rate constants at that voltage. The k inact,V and k rec,v values are then used to calculate the voltage dependent equilibrium constant for inactivation: with the midpoint for the voltage-dependence of steady-state inactivation occurring at the voltage where k inact,V~krec,V .

Data Analysis
Initial data analysis was performed using Clampfit 10.2 software (Molecular Devices, Sunny Vale, CA). All summary data are presented as mean 6 SEM. Statistical comparisons (performed using ANOVA, followed by Tukey's t-test) were carried out in Prism (GraphPad 5.04, La Jolla, CA,). A P value of ,0.05 was considered significant.

Results
Kv11.1 channels can exist in three different conformational states: closed, open and inactivated. The steady-state distribution and rates of interconversion between these three states were measured using the whole cell voltage clamp technique. We undertook a detailed analysis of the kinetics of the Kv11.1-3.1 and Kv11.1-1A isoforms at 37uC.
Typical examples of a family of Kv11.1-3.1 currents recorded at 37uC during an envelope-of-tails protocol to measure the rate of activation at 0 mV are shown in Figure 1. The time constant of activation for Kv11.1-1A (5865 ms, n = 5) was statistically indistinguishable from that for Kv11.1-3.1 (5765 ms, n = 4) ( Fig 1B). When rates of activation were measured at room temperature the small acceleration in the rate of activation for Kv11.1-3.1 compared to Kv11.1-1A became statistically significant (see Fig. S1). In addition, the Kv11.1-1A/Kv11.1-3.1 heterotetramer had rates of activation that were intermediate between those of the homotetrameric Kv11.1-1A and Kv11.1-3.1 channels (see Fig. S1 and Table S1).
It is apparent from the current traces shown in Fig. 2A that the rates of deactivation for Kv11.1-3.1 channels are markedly faster than for Kv11.1-1A. To further investigate the deactivation phenotype of the Kv11.1-3.1 channels, we measured rates of deactivation over the voltage range 2130 to 260 mV (Fig. 3A). A magnification of the tail currents recorded at 2120 mV, with the traces normalized to the maximum inward current, are shown in Fig. 3B. This clearly highlights the faster deactivation of Kv11.1-3.1 compared to Kv11.1-1A. Both currents show the characteristic hooked appearance reflecting recovery from inactivation followed by deactivation, which has been described extensively in the existing Kv11.1 literature [17] Fig. S3 and Table S2). Figure 4A shows typical current traces for Kv11.-1A (left) and Kv11.1-3.1 (right) recorded during a voltage protocol to monitor the rates of inactivation at potentials between 230 and +60 mV. A magnification of the +10 mV step is shown in Fig. 4B, highlighting slower rates of inactivation for Kv11.1-3.1 compared to Kv11.1-1A. Kv11.1-3.1 channels inactivated more slowly than Kv11.1-1A channels at all voltages in this range (Fig. 4C). The rates of recovery from inactivation, in the voltage range of 2130 to 210 mV (Fig. 4C), were measured from the protocol shown in Fig. 3A. Kv11.1-3.1 channels recover from inactivation more rapidly than Kv11.1-1A channels over the tested voltages. Furthermore, the values for the V 0.5 of steady-state inactivation (Fig. 4D, measured using Eq. 3, see material and methods) for Kv11.1-1A (25463 mV) and Kv11.1-3.1 (22563 mV), were statistically significantly different to each other (p,0.001). This shift in voltage-dependence of inactivation was also seen at room temperature with the heterotetramer Kv11.1-1A/Kv11.1-3.1 having an intermediate phenotype (see Fig. S4 and Table S5).
To examine the physiological relevance of the difference in voltage-dependent gating between the Kv11.1-3.1 and Kv11.1-1A isoforms, we next looked at the current responses to repetitive depolarization pulses to mimic trains of neuronal action potentials.

Discussion
The Kv11.

1-3.1 Isoform has Multiple Effects on Channel Gating
In the Kv11.1-3.1 isoform the first 102 amino acids are replaced with six unique amino acids compared to the more abundant Kv11.1-1A isoform. The N-terminal region is well known to affect the deactivation kinetics of Kv11.1 channels and so it was expected that the Kv11.1-3.1 isoform should have faster kinetics. This was demonstrated by Huffaker et al. in the first report describing the 3.1 isoform [7] and confirmed in this study (see Fig. 3 and Fig. S3). The rates of activation ( Fig. 1 and Fig. S1) and the voltagedependence of steady-state activation ( Fig. 2 and Fig. S2) were only minimally affected. Parameters for all gating variables are summarized in Table S1-S5.
Our data is consistent with numerous reports in the literature showing that various N-terminal truncations of the Kv11.1 channel selectively affect deactivation with no significant effects on rates of activation or steady-state activation [13,19]. In addition to the major effect on deactivation, we also show that the Kv11.1-3.1 isoform has a major effect on the kinetics and voltagedependence of steady-state inactivation ( Fig. 4 and Fig. S4). Specifically, the rates of inactivation are slower at depolarized potentials (Fig. 4C) whereas the rates of recovery from inactivation are faster at hyperpolarized potentials (Fig. 4C). As a consequence, the mid-point for the voltage-dependence of steady-state inactivation is shifted in the depolarized direction for Kv11.1-3.1 compared to Kv11.1-1A channels, with the heterotetrameric channel having an intermediate phenotype (Fig. S4). Consequently, at modestly depolarized potentials the steady-state level of current flow though Kv11.1-3.1 channels will be significantly greater than in Kv11.1-1A channels. For example, at 0 mV only 5% of Kv11.1-1A channels will not be inactivated, whereas for Kv11.1-3.1 20% of the channels will not be inactivated. Thus, at depolarized potentials Kv11.1-3.1 channels will have a ''gain of function'' compared to Kv11.1-1A channels which is in marked contrast to the ''loss of function'' seen after depolarization steps caused by the faster deactivation of Kv11.1-3.1 channels [7].

Physiological Implications of Altered Kv11.1 Gating
To investigate the physiological implication of the altered gating properties we investigated how the different isoforms would respond to repetitive short depolarization pulses in the voltage range typical of neuronal action potentials. Given that both isoforms have similar rates of activation and they both are largely inactivated at +40 mV, there was very little difference in the magnitude of the currents recorded during a single 5 ms depolarization step to +40 mV. However, due to the significant faster rates of deactivation ( Fig. 3 and Fig. S3) and rates of recovery from inactivation (Fig 4 and Fig. S4), Kv11.1-3.1 channels almost fully close during the 10 ms repolarization step to 270 mV, while most Kv11.1-1A channels remain in the open conformation. Consequently, Kv11.1-1A currents accumulate to a much higher extent than Kv11.1-3.1. This is consistent with the data reported by Huffaker and colleagues [7]. Measuring the currents after 3 ms indicates that the first pulses for Kv11.1-3.1 start at a higher level compared to Kv11.1-1A. This can be explained by the significantly slower rate of inactivation for Kv11.1.-3.1 (Fig. 4 and Fig. S4) compared to Kv11.1-1A channels. Nevertheless, after repeated depolarization pulses the slower deactivation of Kv11.1-1A means that the current magnitude measured after 3 ms of the depolarization pulses still accumulates to a larger extent than for Kv11.1-3.1 channels. Thus overall, the faster deactivation kinetics of Kv11.1-3.1 is more important than the slower inactivation, resulting in less accumulation of Kv11.1-3.1 current compared to Kv11.1-1A current in response to multiple short depolarization pulses. Consequently, neurones containing Kv11.1-3.1 channels rather than Kv11.1-1A channels should permit more rapid and longer lasting trains of action potentials.
The reduced inactivation would be associated with an increased current flow during the depolarization phase of the action potential, which in turn would be expected to result in a shorter action potential. This however, is unlikely to be of any consequence for single action potentials in neuronal cells as the action potential is so short that the channels will not have activated to any significant degree.

Relevance for Schizophrenia
The role of Kv11.1 channels in the cardiac repolarization is well studied [20] but the contribution of native Kv11.1 K + current to the intrinsic electrical properties of CNS neurons is for the most part poorly understood [21]. Kv11.1-3.1 channels are primate and brain specific and have never been studied in their native environment. Both Kv11.1-1A and Kv11.1-3.1 isoforms are expressed at comparable levels in the hippocampus and the prefrontal cortex [7]. However, in patients with schizophrenia, positive for the M17, M30, M31 and M33 SNPs, the ratio of Kv11.1-3.1 to Kv11.1-1A expression in the hippocampus is ,2.5fold higher than in controls [7]. These changes in expression are very likely to influence the electrical properties of the cells as discussed above. However, before we can determine the overall significance of this altered cellular electrical phenotype we will need to define the precise cellular localizations of these channels and whether they are expressed in inhibitory neurons and/or excitatory neurons.
Despite intensive research, the pathobiology of schizophrenia remains obscure and consequently there is no cure for the disease [1,22]. Therapeutics only rely on reduction of the symptoms. Without doubt, it is important to better understand the underlying mechanisms and how genes and their products impact brain function [23,24]. Given that Kv11.1-3.1 is also expressed in healthy controls, it is likely that the higher disease incidence in patients with increased expression of Kv11.1-3.1 cannot be attributed to the altered levels of Kv11.1-3.1 alone but rather it is acting in combination with other risk factors (genetic or environmental) that remain to be determined.      Results S1 In addition to characterising the channels at 376C we performed a detailed analysis at room temperature. In addition to the greater stability of the patch clamp recordings at room temperature the slowing of the kinetics at room temperature enabled us to investigate the subtle differences between Kv11.1-1A, Kv11.1-3.1, and the heterotetrameric Kv11.1-1A/ Kv11.1-3.1 channels. In all cases the phenotype of the heterotetrameric Kv11.1-1A/Kv11.1-3.1 channels was intermediate between that of Kv11.1-1A alone and Kv11.1-3.1 alone, although in some cases the differences were not statistically significant. Rates of activation at 0 mV are shown in Fig. S1, steady-state activation in Fig. S2 and the rates of deactivation in Fig. S3. All the inactivation properties (rates of recovery from inactivation, rates of inactivation and the V 0.5 of steady-state inactivation) are shown in Fig. S4. The data for all parameters at both room temperature and 37uC are also summarized in Tables S1-S5. (DOCX) Author Contributions