The α1H Ca2+ channel subunit is expressed in mouse jejunal interstitial cells of Cajal and myocytes

T-type Ca2+ currents have been detected in cells from the external muscular layers of gastrointestinal smooth muscles and appear to contribute to the generation of pacemaker potentials in interstitial cells of Cajal from those tissues. However, the Ca2+ channel α subunit responsible for these currents has not been determined. We established that the α subunit of the α1H Ca2+ channel is expressed in single myocytes and interstitial cells of Cajal using reverse transcription and polymerase chain reaction from whole tissue, laser capture microdissected tissue and single cells isolated from the mouse jejunum. Whole-cell voltage clamp recordings demonstrated that a nifedipine and Cd2+ resistant, mibefradil-sensitive current is present in myocytes dissociated from the jejunum. Electrical recordings from the circular muscle layer demonstrated that mibefradil reduced the frequency and initial rate of rise of the electrical slow wave. Gene targeted knockout of both alleles of the cacna1h gene, which encodes the α1H Ca2+ channel subunit, resulted in embryonic lethality because of death of the homozygous knockouts prior to E13.5 days in utero. We conclude that a channel with the pharmacological and molecular characteristics of the α1H Ca2+ channel subunit is expressed in interstitial cells of Cajal and myocytes from the mouse jejunum, and that ionic conductances through the α1H Ca2+ channel contribute to the upstroke of the pacemaker potential. Furthermore, the survival of mice that do not express the α1H Ca2+ channel protein is dependent on the genetic background and targeting approach used to generate the knockout mice.

. These data are consistent with the presence of an unclassified non-selective cation conductance in mouse colonic myocytes [5].
Together with enteric nerves and myocytes, interstitial cells of Cajal (ICC) are required for normal gastrointestinal motility [10]. ICC generate the electrical slow wave; an oscillation in membrane potential that is required for normal phasic contractions of gastrointestinal smooth muscles [11,12] and a role for T-type Ca 2ϩ currents in the generation of slow waves has been proposed. Normal slow wave activity results from Ca 2ϩ influx through plasma membrane ion channels, Ca 2ϩ release from inositol 1,4,5trisphosphate sensitive Ca 2ϩ stores, and re-polarization dependent on a variety of ion channel types including non-selective cation channels and/or Ca 2ϩ activated Cl Ϫ channels [13,14]. The Ca 2ϩ influx that contributes to the upstroke of the electrical slow wave is sensitive to block by agents that alter TϪtype Ca 2ϩ channel activity. Intracellular recordings from the external muscle layers of mouse small intestine [15] indicate that a nifedipine-insensitive, Ca 2ϩ -permeable conductance is responsible for Ca 2ϩ influx during the electrical slow wave. In submucosal ICC from mouse colon, Ni 2ϩ (10-100 M) and mibefradil (3 M) application resulted in a reduced rate of rise of the upstroke of the electrical slow wave [16,17]. Detailed analysis of the pacemaker potentials and electrical slow waves recorded by impaling ICC and smooth muscle cells in mouse small intestine showed that mibefradil (Ն 10 M) reduced the rate of rise of the upstroke depolarization because of failure to entrain unitary potentials recorded from ICC [18]. Experiments using imaging of intracellular Ca 2ϩ transients to follow pacemaker activity in myenteric ICC of human [19] and mouse [20] small intestine, demonstrated that the upstroke phase of the transient depends on activation of dihydropyridine-resistant Ca 2ϩ influx. In mouse ileum, this Ca 2ϩ influx was blocked by low concentrations of mibefradil (0.1 M) and Ni 2ϩ (100 M) [20] whereas in human small intestine [19], higher concentrations of mibefradil (10-50 M) were required to inhibit the Ca 2ϩ influx. These observations indicate that the conductance responsible for the upstroke of the slow wave is probably a T-type Ca 2ϩ current in mouse but could be because of a different conductance in ICC of the human small intestine [21]. The genes that encode three types of T-type Ca 2ϩ channels have been cloned from mouse and human tissue. These proteins, Cav3.1, Cav3.2 and Cav3.3 (or ␣1G, ␣1H, and ␣1I, respectively) have the properties of T-type Ca 2ϩ channels in heterologous expression systems [22]. The messenger RNA (mRNA) for all three T-type Ca 2ϩ channels have been identified in mouse small intestine by PCR but ICC of the deep muscular plexus (ICC-DMP) do not express any T-type Ca 2ϩ channel mRNA [23]. In mouse colonic myocytes, ␣1G complementary DNA (cDNA) could not be amplified by reverse transcriptase polymerase chain reaction (RT-PCR) [5].
For this study, we have further investigated whether T-type Ca 2ϩ currents play a role in the electrical properties of the mouse jejunum based on the increased knowledge of the physiological and pharmacological properties of the T-type Ca 2ϩ current. We have identified the T-type Ca 2ϩ channel mRNA that is expressed in myocytes and in ICC from the external muscle layers as ␣1H and we attempted to study the effect of knocking out expression of this gene by gene-targeted mutagenesis.

Materials and methods
The Institutional Animal Care and Use Committee at Mayo Clinic, Rochester approved all animal handling procedures.

RNA isolation and reverse transcription and PCR amplification of cDNA
RNA was isolated from the external muscle coat of the adult mouse jejunum immediately after dissection as described previously for isolation of RNA from human jejunum smooth muscle [24].

Reverse-transcription PCR (RT-PCR)
PCR amplifications were performed using GeneAmp 2400 PCR Systems (PE Biosystems, Foster City, CA, USA) or icycler (Bio-Rad, Hercules, CA, USA) using standard procedures as previously published [24]. Reverse transcription (RT) was performed using a mixture of random hexamer and oligo dT primers following the instructions of the manufacturer (PE Biosystems). The product of the RT reaction was then amplified for T-type Ca 2ϩ channel ␣ subunits using gene specific primers that were specifically designed to flank regions containing introns in the genomic sequence (see Table 1 for details). All PCR products were purified and sent to the Mayo Molecular Core Facility for automated DNA sequencing.

Laser capture microdissection (LCM)
Sections of mouse jejunum, 6 m thick, were mounted on glass slides and fixed in ice-cold acetone according to the protocol described previously [24]. A number of spots of tissue containing about 1500 smooth muscle cells from the circular muscle layer or the longitudinal muscle layer were collected using the PIX II Cell LCM system (Arcturus Engineering Inc., Santa Clara, CA, USA) with the 7.5-m spot size. The caps with collected cells were then immediately placed into sterile 0.5 ml microcentrifuge

Single cell PCR amplification from identified ICC
Single ICC from 3-to 5-day-old mice were obtained as previously described [25] from a dissociation of jejunal smooth muscle that had been immunolabelled with an antibody directly conjugated to the fluorophore Alexa 546 (Invitrogen). 3-5 labelled cells and 3-5 un-labelled, spindleshaped cells were collected using a glass pipette and placed into an RNAse free tube. A sample of bath solution was collected as a negative control against possible contamination. RNA was extracted then reverse transcribed using the ViLo Superscript III kit from Invitrogen. The cDNA was then probed for the presence of the ␣1H Ca 2ϩ channel transcript and c-Kit by two-step nested PCR using the primers shown in Table 1.

Genotyping
Genotyping was done by Southern analysis using a probe (XP1) generated by PCR from genomic DNA. The genomic DNA was obtained from tails of the mice extracted using the Tissue Direct™ multiplex PCR system (GenScript, Piscataway, NJ, USA). EcoRI was used to digest 10 g of the genomic DNA for Southern blotting. The probe recognized a 11-kb fragment of wild-type DNA and a 9.5-kb fragment of DNA from the targeted allele. PCR genotyping was also used to distinguish the wild-type and targeted alleles. The primers JLR5 and JLR6 were used to identify a 487 nucleotide band in the wild-type alleles and primers Puro3a and KO37 identified a 520 nucleotide band in the knockout allele (see Table 1 for primer sequences).

Intracellular electrical recordings
The segments of small intestine were opened along the anti-mesenteric border and transferred to a Petri dish filled with fresh oxygenated normal Krebs solution. The mucosa was removed under direct vision by using a binocular microscope and muscle strips (

Whole-cell voltage clamp recordings
Currents were recorded from voltage clamped cells at 22ЊC using standard whole-cell techniques [26,27]

Results
The presence of mRNA for the T-type Ca channel ␣ subunits, ␣1G, ␣1H and ␣1I was investigated by reverse transcription and PCR of total RNA derived from the external muscle layers of adult mouse jejunum. Total RNA purified from mouse hippocampus was used to provide a positive control for the effectiveness of the primers. ␣1G, and ␣1H mRNA was amplified from the jejunum but although ␣1I transcripts were detected in the hippocampal samples, this was not detected in samples from jejunum (Fig. 1A). The identity of the products was confirmed by sequencing of the excised, purified products. We used the same primers to test whether any of the transcripts were specifically expressed in cells from the circular smooth muscle layers of the gastric fundus and jejunum by isolating cells using laser capture micro-dissection. ␣1H mRNA was amplified from all of the samples containing cells collected from the jejunum but the primers for ␣1G and ␣1I did not amplify products of the expected size in any of the jejunal LCM samples (Fig. 1B). ␣1H mRNA but not ␣1I mRNA was also detected in cells collected from the circular muscle layer of the gastric fundus. The presence of ␣1G mRNA in circular smooth muscle from the gastric fundus was indicated by a faint band in one of the four gastric fundus samples (Fig. 1B) Fig. 1C). c-Kit mRNA was amplified from tubes containing Kit immunoreactive cells but was not amplified from tubes containing Kit-negative, spindle shaped cells (Fig. 1C) confirming the identity of the collected cells.

Representative PCR results are shown for a sample containing Kitpositive cells (S1) and Kit-negative cells (S2) as well as results from a separate RT-PCR reaction using Kit-positive cells (S3) and bath solution. A product was not amplified from a tube containing the bath water.
The presence of ␣1H transcripts in cells from mouse jejunum smooth muscle suggested that cells from this region may express mibefradil-sensitive, nifedipine and Cd 2ϩ -resistant T-type Ca 2ϩ channels. To verify this, we did whole-cell voltage clamp recordings on freshly dissociated myocytes from the mouse jejunum (Fig. 2). In the presence of 1 M nifedipine and 80 mM Ba 2ϩ , an inward current was recorded that was inhibited by more than 90% with 2.7 M mibefradil ( Fig. 2A). This rapidly inactivating current activated at -50 mV with a peak inward current at -5 mV ( Fig. 2A(ii)). Cadmium (30 M), expected to block L-type but not T-type Ca 2ϩ channels, did not block the current and this Cd 2ϩ resistant current was also inhibited by more than 90% with 2.7 M mibefradil (Fig. 2B).
The molecular and functional evidence for T-type Ca 2ϩ channels in the mouse jejunum prompted us to investigate whether inhibitors of T-type Ca 2ϩ channels would affect the electrical slow wave in mouse jejunum. When studying balb/c wild-type mice, mibefradil, at a concentration that will predominantly block T-type Ca 2ϩ channels (2 M, [28]), did not completely block the slow wave. However, the time constant for the rising phase of the electrical slow wave was increased from 0.093 Ϯ 0.005 sec. (n ϭ 9 cells from five mice) to 0.171 Ϯ 0.015 sec. (n ϭ 10 cells from three mice, P Ͻ 0.01) and the frequency of electrical slow waves was reduced from 0.67 Ϯ 0.02 Hz to 0.59 Ϯ 0.02 (P Ͻ 0.05). Mibefradil had no significant effect on the resting membrane potential of the impaled smooth muscle cells (-64.7 Ϯ 2.9 mV in control, -69.2 Ϯ 2.6 mV in mibefradil, P ϭ 0.26) (Fig. 3).  Table 1

Fig. 2 Representative T-type whole-cell currents recorded in myocytes from the outer muscle layers of the mouse jejunum. (A) (i) Nifedipineresistant, mibefradil-sensitive Ba 2ϩ currents, (ii) Current-voltage relationship for the peak inward currents show in (A) (i). (B) (i) Mibefradil-sensitive Ba 2ϩ currents obtained in the presence of nifedipine and Cd 2ϩ , (ii) Current-voltage relationship for the peak inward currents shown in (B) (i). Currents were recorded by stepping the command voltage from -100 mV to between -80 and ϩ35 mV in 5 mV steps.
To further examine the contribution of ␣1H  Table 2). The foetuses did not have anatomical abnormalities that allowed us to distinguish between knockout and heterozygous or wild-type animals at embryonic age 9.5 to 10.5 (Fig. 5). The hearts were beating when the foetuses were removed and there were no clear neural tube defects. However, later during gestation (E17.5 to E19.5) reabsorbed foetuses were observed in the uteruses removed for genotyping of the foetuses. We were unable to obtain more than one mouse that was homozygous for the knockout allele so we compared the properties of the electrical slow wave in the circular smooth muscle layer from the jejunum of four mice heterozygous for the knockout with the properties in four wild-type siblings. Forty-one cells from heterozygous mutant mice and 39 cells from wild-type mice were studied and no differences were detected in the rate of rise of the electrical slow wave as (Time constant in heterozygous animals ϭ 0.143 Ϯ 0.007 sec, homozygous wild type ϭ 0.139 Ϯ 0.006 sec., P Ͼ 0.05). The electrical slow wave recorded from smooth muscle cells in the jejunum in the one mouse homozygous for knockout of the ␣1H gene appeared abnormal. The slope of the initial rising phase of the slow wave was lower than usual and the frequency of the slow waves was less than half the normal value (0.296 Ϯ 0.009 Hz, n ϭ 7 cells, Fig. 6 versus 0.67 Ϯ 0.02 Hz in 9 cells from 5 balb/c wild-type mice, see above) with no clear difference in the resting membrane potential of the cells (-56.1 Ϯ 2.3 mV, n ϭ 7 cells versus -64.7 Ϯ 2.9 mV in 9 cells from 5 balb/c wild-type mice, see above).

Discussion
Inward cationic conductances that are activated at comparatively negative membrane potentials (Ͻ -30 mV) and that are not selectively permeable, or are permeable to Ca 2ϩ or Na ϩ , have been demonstrated in gastrointestinal smooth muscle cells and interstitial cells of Cajal from many species [3-9, 24, 29]. As a result of recent pharmacological and technical advances, the biophysical and molecular identification of these conductances should be possible but so far few have been definitively identified [24]. In this study, we have demonstrated the expression of the ␣1H Ca channel subunit in the circular smooth muscle layer of the mouse jejunum and fundus, specifically in myocytes and ICC from the jejunum. We have also determined that inhibition of T-type Ca 2ϩ currents affects the electrical slow wave in this tissue. We did not directly record from identified ICC in this article and attempts to directly determine the contribution of the ␣1H subunit to gastrointestinal function were not successful because of the lethal effects of knocking out the cacna1h gene, which encodes this protein.
The problems with dissecting out Ca 2ϩ currents in myocytes and ICC are that some of the T-type Ca 2ϩ channels are quite resistant to the effect of Ni 2ϩ (Ͼ 200 M for ␣1I and ␣1G, [30]) and Ni 2ϩ has effects on other Ca 2ϩ permeable channels or transporters in the plasma membrane [31][32][33][34]. When used at the correct concentration, mibefradil is a fairly selective inhibitor of T-type Ca 2ϩ channels [35] with a 4-to 10-fold higher potency as an inhibitor of T-type Ca 2ϩ channels when compared to its effects on L-type Ca 2ϩ channels and the sodium channel, Nav1.5 [28] but it is often tested at concentrations that are not selective and its effects are diminished at positive membrane voltages [36]. Similarly, excluding the other Ca 2ϩ permeable conductances in the cells can be difficult because of the lack of selective inhibitors of non-selective cation conductances. When L-type Ca 2ϩ currents are present in the cells, these currents can be inhibited by dihydropyridine calcium channel blockers but the effectiveness of dihydropyridines is reduced at the more negative membrane    [37] where the low-voltage activated current is observed, possibly leading to the erroneous identification of residual L-type current as a low-voltage activated Ca 2ϩ current. In addition dihydropyridines inhibit some T-type Ca 2ϩ channels at concentrations greater than 5 M [38]. The molecular identification of the ␣1H subunit in mouse jejunal smooth muscle is consistent with the biophysical and pharmacological properties of the low-voltage activated Ca 2ϩ selective conductance reported herein in mouse jejunal myocytes and recorded by several groups in smooth muscle cells from rat small intestine [3] as well as myocytes from other gastrointestinal tissue [3,[6][7][8][9] [5,23]. In one study, mRNA for the ␣1G subunit was not detected in myocytes from mouse colonic circular muscle [5]. In the other study, the presence was demonstrated of all three T-type Ca 2ϩ channel ␣ subunits in the external muscle layers of mouse small intestine but this study did not determine if myocytes expressed the mRNA for any of the subunits. We could not detect the ␣1I subunit mRNA in our samples even though the primers worked in our positive control experiment. Three different primer sets were tested for the ␣1I subunit and none amplified a product of the expected size (data not shown). One possible explanation for the amplification of ␣1H Ca 2ϩ channel mRNA from the circular smooth muscle layer of the mouse jejunum is that the mRNA came from a cell type other than the myocytes or ICC. In human myometrium, ␣1H Ca 2ϩ channel immunoreactivity was detected in lymphocytes but not smooth muscle cells [1]. This is consistent with a role for T-type Ca 2ϩ channels in the migration of leucocytes [39]. However, the samples that we collected by laser capture microdissection were from histochemically stained sections and were taken from the circular smooth muscle layer in regions that did not show evidence of lymphocytic contamination. In addition, the single cells that we collected were either Kit-immunoreactive or had myocyte morphology, so we consider it unlikely that this could account for the results.
The reduction in the rate of rise of the electrical slow wave recorded in myocytes of mouse jejunum in response to mibefradil treatment is consistent with an effect of mibefradil on ICC as previously published in other tissues. The upstroke of electrical slow waves recorded in mouse small intestine involves Ca 2ϩ influx through a non-L type Ca 2ϩ channel [15], Ni 2ϩ (1-100 M) reduces the rate of rise of slow waves in guinea pig gastric antrum [40] [41]. Interestingly, slowing of the rise time for the electrical slow wave in mice in mibefradil and Ni 2ϩ is similar to the effects of inhibiting the tetrodotoxin (TTX)-resistant Na ϩ channel, Nav1.5 on the electrical slow wave recorded in circular muscle of human jejunum [29]. Also, mibefradil and Ni 2ϩ do not completely block the upstroke at concentrations where near complete inhibition of a T-type channel is expected [18] suggesting the possibility that more than one channel type may contribute to the upstroke. These pharmacological properties also raise the possibility that the observed current carried by non-L-type Ca 2ϩ channels is carried by T-type-like channels and the possibility that the current is carried by a combination of alpha subunits or a novel T-type alpha subunit cannot be excluded. The target for mibefradil is likely the ␣1H channel subunit identified in ICC by single cell PCR rather than in smooth muscle because the electrical slow wave is generated by ICC in the myenteric plexus region of the mouse jejunum, where the pacemaker potential is initiated. As the slow wave is generated by ICC, there should not be a need for a T-type Ca 2ϩ channel to participate in the generation of the upstroke in smooth muscle cells. It may be that Ca 2ϩ influx through T-type Ca 2ϩ channels in myocytes is more important for functions of Ca 2ϩ unrelated to contractility such as cellular proliferation or migration. In this respect, myocytes from mouse jejunum are similar to other smooth muscles [2,42] although there does appear to be role for ␣1H Ca 2ϩ channel subunits in controlling relaxation of coronary smooth muscle [43].
Studies in the mice heterozygous for the knockout of the ␣1H Ca 2ϩ channel subunit did not provide any further information because when compared with wild-type litter-mates, there were no differences in the properties of the electrical slow wave. This indicates that the partial knockout did not have a gene dosing effect.
The lethal effects of knocking out expression of the ␣1H Ca 2ϩ channel subunit in all but one of the many hundreds of mice studied were surprising given that another group has successfully generated a strain of mice by deletion of exon 6 in the cacna1h gene [43]. The reported effects of this knockout were cardiac injury because of coronary artery constriction and reduced body mass compared to wild-type litter mates. These are both phenotypes consistent with reduced survival but were clearly not sufficient to prevent survival to maturity of the knockout mice [43]. We were only able to obtain one adult mouse homozygous for the knockout allele over a period of 4 years studying these animals and we conclude that this very low survival rate reflects the genetic backgrounds of the strain of either embryonic stem cells or recipient blastocysts [44]. We could not identify the actual cause of death of the homozygous knockout foetuses but death of foetuses at ages around E10.5 is associated with cardiovascular abnormalities (e.g. [45]), and this is consistent with both the phenotype of the published ␣1H knockout [43] and the known role of T-type Ca 2ϩ channels in smooth muscle cell growth and proliferation [2]. Neural tube abnormalities, which also can cause embryonic lethality at around E10.5 [46], were not observed. Interestingly, the rate of rise and frequency of the electrical slow wave were lower than normal values in recordings from the jejunum of the one mouse that did survive.
In conclusion, the ␣1H Ca 2ϩ channel subunit is expressed in myocytes and ICC from the circular muscle layer of mouse jejunum. A T-type Ca 2ϩ current also appears to play a role in the generation of the upstroke of the electrical slow wave in mouse tissue and the pharmacological properties of this current are consistent with expression of the ␣1H Ca 2ϩ channel subunit in ICC contributing to this current. The phenotypic effect of knocking out the gene for the ␣1H Ca 2ϩ channel subunit is dependent on the genetic background of the mouse strain used for the construct and can be embryonic lethal in some mouse strains.