Ca-\(\alpha\)1T, a fly T-type Ca2+ channel, negatively modulates sleep

Kyunghwa Jeong1, Soyoung Lee2, Haengsoo Seo2, Yangkyun Oh1, Donghoon Jang1, Joonho Choe1, Daesoo Kim1, Jung-Ha Lee2,\(+\), Walton D. Jones1,\(\ast\)

1 KAIST, Department of Biological Sciences, Daejeon, 305-701, Republic of Korea
2 Sogang University, Department of Life Sciences, Seoul, 121-742, Republic of Korea
\(+\) E-mail: jhleem@sogang.ac.kr \(\ast\) E-mail: waltonjones@kaist.edu

Abstract

Mammalian T-type Ca2+ channels are encoded by three separate genes (Cav3.1, 3.2, 3.3). In mammals, T-type channels are reported to be sleep stabilizers that are important in the generation of the delta rhythms of deep sleep, but controversy remains. Progress in identifying the precise physiological functions of the T-type channels has been hindered by many factors, including possible compensation between the products of these three genes and a lack of specific pharmacological inhibitors. Invertebrates have only one T-type channel gene and its physiological functions are less well-studied. We cloned Ca-\(\alpha\)1T, the only Cav3 channel gene in the Drosophila melanogaster genome, expressed it in Xenopus oocytes or HEK-293 cells, and verified that it is capable of passing typical T-type currents. Voltage-clamp analysis revealed that the biophysical properties of Ca-\(\alpha\)1T show mixed similarity, sometimes falling closer to Cav3.1, sometimes to Cav3.2, and sometimes to Cav3.3. We found that Ca-\(\alpha\)1T is broadly expressed across the adult fly brain in a pattern vaguely reminiscent of mammalian T-type channels. In addition, flies lacking Ca-\(\alpha\)1T show an abnormal increase in sleep duration that is most pronounced during subjective day under continuous dark conditions despite normal oscillations of the circadian clock. Thus, our study suggests invertebrate T-type Ca2+ channels promote wakefulness rather than stabilizing sleep.

Introduction

T-type Ca2+ channels are a subfamily of voltage-dependent Ca2+ channels (VDCCs) that produce low-voltage-activated (LVA) Ca2+ currents implicated in NREM sleep in mammals(Lee 2004). Three different genes encode the pore-forming alpha subunits of mammalian T-type channels, Cav3.1, 3.2, and 3.3. Of these, Cav3.1 and 3.3 are highly expressed in the thalamus, where the oscillations required for NREM sleep are generated(Steriade 1991, Dossi 1992, Talley 1999). Mice lacking Cav3.1 show reduced delta-wave activity and reduced sleep stability, suggesting that mammalian T-type currents have a sleep-promoting or stabilizing function(Lee 2004).

Unlike mammals, Drosophila melanogaster has only one T-type Ca2+ channel, Ca-\(\alpha\)1T, which is also known as Dm\(\alpha\)G and Ca-\(\alpha\)1T. A recent study found that motor neurons in flies lacking Ca-\(\alpha\)1T show reduced LVA but also reduced high-voltage-activated (HVA) Ca2+ currents, suggesting that although Ca-\(\alpha\)1T seems to be a genuine T-type channel, it may have interesting biophysical properties(Ryglewski 2012). We therefore cloned a single isoform of Ca-\(\alpha\)1T, expressed it in Xenopus oocytes or HEK-293 cells, and compared its biophysical properties with those of the rat T-type channel Cav3.1. We also generated several Ca-\(\alpha\)1T mutant alleles and identified a defect in their sleep/wake cycles. Contrary to results in mammals, the fly T-type Ca2+ channel destabilizes sleep. We anticipate that our findings will help clarify species-dependent differences in the in vivo functions of T-type Ca2+ channels, particularly their role in sleep physiology.

Results

Ca-\(\alpha\)1T produces LVA currents in Xenopus oocytes

The fly T-type Ca2+ channel Ca-\(\alpha\)1T spans roughly 90 kilobases of genomic DNA and has five different annotated mRNA transcripts designated RB through RF. Because the smallest of these transcripts is still over 10 kilobases in length, we used a piece-meal approach to assemble a full-length cDNA. To do so, we isolated total RNA from fly heads and used reverse transcription to produce cDNAs. Using these cDNAs as a template, we amplified and then assembled partial clones to obtain full-length cDNAs for both the RB (NM_132068) and RC (NM_001103419) Ca-\(\alpha\)1T transcripts. After sequence verification, we used these Ca-\(\alpha\)1T cDNA clones to produce cRNAs for injection into Xenopus oocytes. We were able to confirm expression of the RC isoform, but not the RB isoform, by measuring robust inward currents using 10 mM Ba2+ as a charge carrier 4 days after cRNA injection. In all subsequent experiments performed with this RC isoform cDNA, we refer to it simply as Ca-\(\alpha\)1T.

We next compared the biophysical properties of Ca-\(\alpha\)1T with those of a mammalian T-type Ca2+ channel homolog, rat Cav3.1(Perez-Reyes 1998), using the Xenopus oocyte expression system. Both Ca-\(\alpha\)1T and Cav3.1 have low-voltage activation thresholds, but the threshold of Ca-\(\alpha\)1T (\(-\)60 mV) is slighty lower than that of the rat channel by 3\(\sim\)4 mV. Both Ca-\(\alpha\)1T and Cav3.1 produce current kinetics typical of T-type Ca2+ channels when subjected to a protocol of serial step pulses from a holding potential of \(-\)90 mV. Although the inactivation kinetics of Ca-\(\alpha\)1T are slightly slower than those of Cav3.1, both the activation and inactivation kinetics of currents through Ca-\(\alpha\)1T accelerate at higher step pulse values. This produces the criss-crossing pattern typical of T-type Ca2+ channels (Fig. \ref{fig:1}a). Together, these biophysical properties—an activation threshold of \(-\)60 mV, a potential of maximal current at \(-\)20 mV, transient current kinetics, a criss-crossing pattern in currents evoked by a step pulse voltage protocol—all of these properties mark Ca-\(\alpha\)1T as a typical T-type Ca2+ channel(Perez-Reyes 1998, Carbone 1984, Cribbs 1998, Lee 1999).

We next obtained activation curves for Ca-\(\alpha\)1T and Cav3.1 by fitting chord conductances with the Boltzmann equation. The potential for half-maximal activation (V50,act) of Ca-\(\alpha\)1T and Cav3.1 are \(-\)43.32 \(\pm\) 1.58 and \(-\)38.92 \(\pm\) 1.15 mV, respectively. This indicates that Ca-\(\alpha\)1T is activated at 4.4 mV lower test potentials than Cav3.1 (p\(<\) 0.05, Student’s t-test) (Fig. \ref{fig:1}b and Table \ref{tab:1}). During steady-state inactivation, the potentials of 50% channel availability (V50,inact) for Ca-\(\alpha\)1T and Cav3.1 are estimated to be \(-\)58.04 \(\pm\) 0.71 and \(-\)61.31 \(\pm\) 0.70 mV (p\(<\) 0.05, Student’s t-test). In other words, the V50,inact of Ca-\(\alpha\)1T is 3.3 mV more positive than that of Cavv3.1 (Fig. \ref{fig:1}b and Table \ref{tab:1}). An ion channel’s so-called “window current” is the range of overlap in its steady-state act