The HCN4 gene encodes a hyperpolarization-activated cation current contributing to the slow components of the pacemaking currents I(f) in the sinoatrial node and I(h) or I(q) in the thalamus. Heterologous expression studies of individual HCN channels have, however, failed to reproduce fully the diversity of native I(f/h/q) currents, suggesting the presence of modulating auxiliary subunits. Consistent with this is the recent description of KCNE2, which is highly expressed in the sinoatrial node, as a beta-subunit of rapidly activating HCN1 and HCN2 channels. To determine whether KCNE2 can also modulate the slow component of native I(f/h/q) currents, we co-expressed KCNE2 with HCN4 in Xenopus oocytes and in Chinese hamster ovary (CHO) cells and analysed the resulting currents using two-electrode voltage-clamp and patch-clamp techniques, respectively. In both cell types, co-expressed KCNE2 enhanced HCN4-generated current amplitudes, slowed the activation kinetics and shifted the voltage for half-maximal activation of currents to more negative voltages. In contrast, the related family members KCNE1, KCNE3 and KCNE4 did not change current characteristics of HCN4. Consistent with these electrophysiological results, the carboxy-terminal tail of KCNE2, but not of other KCNE subunits, interacted with the carboxy-terminal tail of HCN4 in yeast two-hybrid assays. KCNE2, by modulating I(f) or I(h) currents, might thus contribute to the electrophysiological diversity of known pacemaking currents in the heart and brain.

Voltage-operated Ca2+ channels are heteromultimeric proteins. Their structural diversity is caused by several genes encoding homologous subunits and by alternative splicing of single transcripts. Isoforms of alpha1 subunits, which contain the ion conducting pore, have been deduced from each of the six cDNA sequences cloned so far from different species. The isoforms predicted for the alpha1E subunit are structurally related to the primary sequence of the amino terminus, the centre of the subunit (II-III loop), and the carboxy terminus. Mouse and human alpha1E transcripts have been analysed by reverse transcription-polymerase chain reaction and by sequencing of amplified fragments. For the II-III loop three different alpha1E cDNA fragments are amplified from mouse and human brain, showing that isoforms originally predicted from sequence alignment of different species are expressed in a single one. Both predicted alpha1E cDNA fragments of the carboxy terminus are identified in vivo. Two different alpha1E constructs, referring to the major structural difference in the carboxy terminus, were stably transfected in HEK293 cells. The biophysical properties of these cells were compared in order to evaluate the importance in vitro of the carboxy terminal insertion found in vivo. The wild-type alpha1E subunit showed properties, typical for a high-voltage activated Ca2+ channel. The deletion of 43 amino acid residues at the carboxy terminus does not cause significant differences in the current density and the basic biophysical properties. However, a functional difference is suggested, as in embryonic stem cells, differentiated in vitro to neuronal cells, the pattern of transcripts indicative for different alpha1E isoforms changes during development. In human cerebellum the longer alpha1E isoform is expressed predominantly. Although, it has not been possible to assign functional differences to the two alpha1E constructs tested in vitro, the expression pattern of the structurally related isoforms may have functional importance in vivo.

The expression of Ca 2ϩ channel A1E isoforms has been analyzed in different cell lines, embryoid bodies and tissues. The comparison of the different cloned A1E cDNA sequences led to the prediction of A1E splice variants. Transcripts of two cloned A1E isoforms, which are discriminated by a carboxy terminal 129-bp sequence, have been detected in different cell lines and tissues. Transcripts of the shorter A1E isoform have been assigned to the rat cerebrum and to neuron-like cells from in vitro, differentiated embryonic stem cells. The shorter isoform is the major transcript amplified from total RNA by reverse transcription (RT)-PCR and visualized on the protein level by Western blotting with common and isoformspecific antibodies. Transcripts of the longer A1E isoform have been identified in mouse, rat and human cerebellum, in in vitro, differentiated embryoid bodies, in the insulinoma cell lines INS-1 (rat) and βTC-3 (mouse), in the pituitary cell line AtT-20 (mouse) when grown in 5 mM glucose, and in islets of Langerhans (rat) and kidney (rat and human). The detection of different isoforms of A1E in cell lines and tissues shows that the wide expression of A1E has to be specified by identifying the corresponding isoforms in each tissue. In islets of Langerhans and in kidney, a distinct isoform called A1Ee has been determined by RT-PCR, while in cerebellum a set of different A1E structures has been detected, which might reflect the functional heterogeneity of cerebellar neurons. The tissue-specific expression of different isoforms might be related to specific functions, which are not yet known, but the expression of the new isoform A1Ee in islets of Langerhans and kidney leads to the suggestion that A1E might be involved in the modulation of the Ca 2ϩ -mediated hormone secretion.Keywords : A1E isoforms ; Ca 2ϩ channels ; islets of Langerhans; kidney; single-cell RT-PCR.Low and high voltage-gated Ca 2ϩ channels are subdivided, according to their threshold of activation and their different single channel conductances [1]. The pore-forming subunit of the first low voltage-activated T-type Ca 2ϩ channel has been cloned and functionally expressed recently as A1G [2].The group of six different high voltage-activated Ca 2ϩ channels (A1S-, A1C-, A1D-, A1A-, A1B-, A1E subunits) has been investigated intensively during the last decade [3,4]. They are subdivided into two different families, based on their primary sequence. The A1S, A1C, and A1D subunits are sensitive against the classical blockers of L-type Ca 2ϩ channels, the dihydropyridines, the phenylalkylamines and the benzothiazepines. The subfamily of the non-L-type A1 subunits share, among each other, a functional interaction with inhibitory G proteins [5]. The Ntype Ca 2ϩ channels (A1B) are blocked by ω-conotoxin-GVIA, the P-type and Q-type Ca 2ϩ channels (A1A) are discriminated byCorrespondence to T. Schneider,

Ca2+-dependent regulation of L-type and P/Q-type Ca2+ channel activity is an important mechanism to control Ca2+ entry into excitable cells. Here we addressed the question whether the activity of E-type Ca2+ channels can also be controlled by Ca2+. Switching from Ba2+ to Ca2+ as charge carrier increased within 50 s, the density of currents observed in HEK-293 cells expressing a human Cav2.3d subunit and slowed down the inactivation kinetics. Furthermore, with Ca2+ as permeant ion, recovery from inactivation was accelerated, compared to the recovery process recorded under conditions where the accumulation of [Ca2+]i was prevented. In a Ba2+ containing bath solution the Ca2+-dependent changes of E-type channel activity could be induced by dialysing the cells with 1 micro m free [Ca2+]i suggesting that an elevation of [Ca2+]i is responsible for these effects. Deleting 19 amino acids in the intracellular II-III linker (exon 19) as part of an arginine-rich region, severely impairs the Ca2+ responsiveness of the expressed channels. Interestingly, deletion of an adjacent homologue arginine-rich region activates channel activity but now independently from [Ca2+]i. As a positive feedback-regulation of channel activity this novel activation mechanism might determine specific biological functions of E-type Ca2+ channels.

Multiple types of voltage-activated Ca(2+) channels (T, L, N, P, Q, R type) coordinate Ca(2+)-dependent processes in neurons and neuroendocrine cells. Expressional and functional data have suggested a role for Ca(v)2.3 Ca(2+) channels in endocrine processes. To verify its role in vivo, Ca(v)2.3(-/-) mutant mice were generated, thus deficient in alpha 1E/R-type Ca(2+) channel. Intraperitoneal injection of D-glucose showed that glucose tolerance was markedly reduced, and insulin release into plasma was impaired in Ca(v)2.3-deficient mice. In isolated islets of Langerhans from these animals, no glucose-induced insulin release was detected. Further, in stressed Ca(v)2.3-deficient mice, the rate of glucose release into the blood was only 29% of that observed for wild-type animals. Thus, the deletion of Ca(v)2.3 causes deficits not only in insulin release but also in stress-induced hyperglycemia. The complex phenotype of Ca(v)2.3-deficient mice has dual components related to endocrine and neurological defects. The present findings provide direct evidence of a functional role for the Ca(v)2.3 subunit in hormone secretion and glucose homeostasis.

Among voltage-gated Ca 21 channels the non-dihydropyridine-sensitive a1E subunit is functionally less well characterized than the structurally related a1A (v-agatoxin-IVA sensitive, P-/Q-type) and a1B (v-conotoxin-GVIA sensitive, N-type) subunits. In the rat insulinoma cell line, INS-1, a tissue-specific splice variant of a1E (a1Ee) has been characterized at the mRNA and protein levels, suggesting that INS-1 cells are a suitable model for investigating the function of a1Ee.In a1E-transfected human embryonic kidney (HEK-293) cells the a1E-selective peptide antagonist SNX-482 (100 nm) reduces a1Ed-and a1Ee-induced Ba 21 inward currents in the absence and presence of the auxiliary subunits b3 and a2d-2 by more than 80%. The inhibition is fast and only partially reversible. No effect of SNX-482 was detected on the recombinant T-type Ca 21 channel subunits a1G, a1H, and a1I showing that the toxin from the venom of Hysterocrates gigas is useful as an a1E-selective antagonist.After blocking known components of Ca 21 channel inward current in INS-1 cells by 2 mm (1/±)-isradipine plus 0.5 mm v-conotoxin-MVIIC, the remaining current is reduced by 100 nm SNX-482 from 212.4^1.2 pA /pF to 27.6^0.5 pA /pF (n 9). Furthermore, in INS-1 cells, glucose-and KCl-induced insulin release are reduced by SNX-482 in a dose-dependent manner leading to the conclusion that a1E, in addition to L-type and non-L-type (a1A-mediated) Ca 21 currents, is involved in Ca 21 dependent insulin secretion of INS-1 cells.

S U M M A R YThe calcium channel ␣ 1E subunit was originally cloned from mammalian brain. A new splice variant was recently identified in rat islets of Langerhans and in human kidney by the polymerase chain reaction. The same isoform of ␣ 1E was detected in rat and guinea pig heart by amplifying indicative cDNA fragments and by immunostaining using peptide-specific antibodies. The apparent molecular size of cardiac ␣ 1E was determined by SDS-PAGE and immunoblotting (218 Ϯ 6 kD; n ϭ 3). Compared to ␣ 1E from stably transfected HEK-293 cells, this is smaller by 28 kD. The distribution of ␣ 1E in cardiac muscle cells of the conducting system and in the cardiomyoblast cell line H9c2 was compared to the distribution of chromogranin, a marker of neuroendocrine cells, and to the distribution of atrial natriuretic peptide (ANP). In serial sections from atrial and ventricular regions of rat heart, co-localization of ␣ 1E with ANP was detected in atrium and with chromogranin A/B in Purkinje fibers of the conducting system in both rat atrium and ventricle. The kidney is another organ in which natriuretic peptide hormones are secreted. The detection of ␣ 1E in the distal tubules of human kidney, where urodilatin is stored and secreted, led to the conclusion that the expression of ␣ 1E in rat heart and human kidney is linked to regions with endocrine functions and therefore is involved in the Ca 2 ϩ -dependent secretion of peptide hormones such as ANP and urodilatin.

SUMMARY Polyclonal antibodies were raised against a common and a specific epitope present only in longer ␣ 1E isoforms of voltage-gated Ca 2 ϩ channels, yielding an "anti-Ecom" and an "anti-E-spec" serum, respectively. The specificity of both sera was established by immunocytochemistry and immunoblotting using stably transfected HEK-293 cells or membrane proteins derived from them. Cells from the insulinoma cell line INS-1, tissue sections from cerebellum, and representative regions of gastrointestinal tract were stained immunocytochemically. INS-1 cells expressed an ␣ 1E splice variant with a longer carboxy terminus, the so-called ␣ 1Ee isoform. Similarily, in rat cerebellum, which was used as a reference system, the anti-E-spec serum stained somata and dendrites of Purkinje cells. Only faint staining was seen throughout the cerebellar granule cell layer. After prolonged incubation times, neurons of the molecular layer were stained by anti-E-com, suggesting that a shorter ␣ 1E isoform is expressed at a lower protein density. In human gastrointestinal tract, endocrine cells of the antral mucosa (stomach), small and large intestine, and islets of Langerhans were stained by the anti-E-spec serum. In addition, staining by the anti-E-spec serum was observed in Paneth cells and in the smooth muscle cell layer of the lamina muscularis mucosae. We conclude that the longer ␣ 1Ee isoform is expressed in neuroendocrine cells of the digestive system and that, in pancreas, ␣ 1Ee expression is restricted to the neuroendocrine part, the islets of Langerhans. ␣ 1E therefore appears to be a common voltagegated Ca 2 ϩ channel linked to neuroendocrine and related systems of the body.