INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

NEPHROGENESIS is characterized by reciprocal induction between two distinct cell types, ureteric bud (UB) cells, which develop to the collecting duct system, and adjacent mesenchymal cells (26). As nephrogenesis proceeds toward the peripheral cortex, cells at the UB tip (tUB), i.e., cells facing condensing mesenchymal cells, repeat their part of the induction program many times. Gene expression patterns within the epithelial UB and mesenchymal cell populations change during these inductive interactions. Functional expression of ion channels is likely to differ, not only with organogenesis, but also between single cells along the longitudinal axis of both the branching ureter and the S-shaped body. Single cell techniques, therefore, are required to analyze the time- and localization-dependent expression.

Cells at the tUB apparently are nonpolar epithelia, as defined by the symmetric distribution of Na-K-ATPase -subunit in both plasma membranes (21). These cells have been demonstrated to polarize after completion of epitheliogenesis by the acquistion of different ion channel types in the apical plasma membrane (13), whereas postnatal maturation is characterized by changes in density of an existing channel type (25). However, the particular population of tUB cells in the key position for cell-to-cell signaling has not been characterized with regard to membrane-conductive properties.

Embryonic kidney and lung have branching morphogenesis and bud formation in common, and embryonic lung epithelia express the chloride channel ClC-2 (30). Specifically, ClC-2 mRNA shows maximum levels in fetal lung cells and is downregulated after birth, suggesting a particular role of ClC-2 in embryonic epitheliogenesis (22). ClC-2 is activated by strong hyperpolarization or by cell swelling (10). The hyperpolarization-evoked whole cell Cl currents generated by ClC-2 typically activate slowly (30), they are not directly dependent on cytoplasmic Ca or ATP concentration, and they have an anion permselectivity of Cl > I (15). By these properties and its voltage dependence, ClC-2 differs greatly from the volume-regulatory I_Cln Cl channel (24), from Ca-activated channels, and from secretory cystic fibrosis transmembrane conductance regulator channels, respectively (1). ClC-2 regulates cell volume, presumably in concert with volume-sensitive organic osmolyte and anion channel (29), and it probably contributes to Cl secretion in some epithelial cells (2, 7). To evaluate ClC-2 mRNA expression in tUB cells, the technique of single cell reverse transcription-polymerase chain reaction (RT-PCR) technique (20) was adapted for epithelial cells. The data indicate that ClC-2 mRNA is expressed in cells at the tUB.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Microdissection of embryonic ureteric buds. As shown in Fig. 1, branching ureteric buds and the attached mesenchyme were microdissected from the outermost cortex of embryonic rat kidney in Ca²⁺- and Mg²⁺-free phosphate-buffered solution at 4°C, explanted on coverslips coated with newborn rat tail collagen, and attached to the matrix at their ureteric trunk end. Coverslips had been glued to culture dishes (Nunc, 30 mm) with a central hole. The mesenchymal caps were removed in one or two of the bud tips by additional dissection to obtain direct access to the basolateral membrane (Fig. 4A). The freshly dissected tissue was kept in nephron culture medium (13) at 37°C for 30 min before analysis. All cells investigated by patch-clamp whole cell recording and RT-PCR were located within an ~20-cell area at the central circumference of the tUB.

View larger version (38K): [in this window] [in a new window]   Fig. 1.   Scheme of experimental protocol. mRNAs encoding glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and the Cl channel ClC-2 were harvested from single ureteric cells at the bud tip, reverse transcribed, and amplified by PCR.

Patch-clamp recordings. Branching UBs were rinsed with NaCl bath solution (in mM: 150 NaCl, 10 D-glucose, 10 HEPES, 5 KCl, 1.6 CaCl₂, and 0.8 MgCl₂, pH 7.2), and dishes were mounted on the stage of an inverted microscope equipped with differential-interference contrast optics (Zeiss, Oberkochen, Germany). Dishes were constantly superfused (1 ml/min) with NaCl bath solution or with a solution containing 180 mM N-methyl-D-glucamine (NMDG) titrated with 5 mM HEPES and ~130 mM HCl to pH 7.2 through a flow system inserted into the dish to reduce the bath volume to 50 µl. After an initial 15-min period of superfusion, patch-clamp experiments were performed, using 1.5-mm borosilicate glass pipettes with 2- to 5-M tip resistance (GC 150 TF-10; Clark Medical Instruments, Pangbourne, UK) and a WR-88 water hydraulic micromanipulator (Narishige, Tokyo, Japan). Pipettes were manufactured by a microprocessor-driven DMZ puller (Zeitz, Augsburg, Germany) and filled with 6.5 µl of pipette solution (in mM: 135 KCl, 5 HEPES, 4 MgCl₂, and 1 EGTA, pH 7.2). Currents were recorded in the whole cell and the outside-out mode and 1-kHz low-pass filtered by an Axopatch 200A amplifier (Axon Instruments, Foster City, CA). Whole cell voltage-clamp pulse protocols were applied, and data were acquired with a rate of 5 kHz by a microcomputer, using pCLAMP software and an TL1 DMA interface (Axon Instruments). Single cell data were stored on digital audio tape (DTR 1204 Recorder; Bio-Logic, Claix, France) and analyzed offline with a sampling rate of 10 kHz. Pipettes and Ag-AgCl wire were baked (220°C for 5 h), tubings and the dish insert were rinsed with 0.1% diethyl pyrocarbonate (DEPC) water before use, and 0.1% DEPC was mixed to bath and pipette solutions before autoclaving. For data analysis, applied voltages were corrected for estimated liquid junction potentials, as previously described (5). All data are expressed as mean ± SE.

Harvesting of cytoplasm. During formation of gigaohm seal and whole cell recording mode, large portions of cytoplasm were clearly visible as they were aspirated several micrometers into the pipette tip by negative pressure applied to the pipette lumen (Fig. 4A). Disrupting the aspirated membrane ensured electrophysiological access to the whole cell and, specifically, the dialysis of cytoplasm by pipette solution. This was monitored by the large increase in capacitive current evoked by a 5-mV square pulse. When the pipette was retracted, the aspirated cytoplasm was pulled from the cell, thus trapping a considerable fraction of cytoplasm within the pipette, whereas the nucleus always remained in the cell. At this point, the pipette was rapidly withdrawn from cell and bath, and the total pipette volume was directly expelled into a tube (filled with 3.5 µl of reverse transcriptase mixture) by applying gentle positive pressure and breaking the glass tip at the tube bottom. The whole cell current recording protocol was carried out in 35 out of 57 cells. In 22 experiments, the pipette was withdrawn immediately after formation of the whole cell recording mode, and its contents were expelled into the reverse transcriptase mixture, so that cDNA synthesis began already ~10 s after sealing and rupturing of the cell membrane.

Reverse transcription. The RT mixture contained 1 µl dithiothreitol (0.1 mM), 0.1 µl single-strand buffer (5-fold), 0.5 µl SuperScript RT (200 U/µl) (all Life Technologies, Eggenstein, Germany); 0.5 µl RNase inhibitor (RNasin, 40 U/µl) and 0.5 µl dNTP mixture (25 mM each) (both Promega, Ingelheim, Germany); and 1 µl random hexamer oligonucleotide primer (Boehringer, Mannheim, Germany). The RNA was transcribed for 1 h at 37°C. From the total volume of ~10 µl cDNA, a 2-µl aliquot was placed in a second tube, and both aliquots were stored at 25°C.

Polymerase chain reaction. cDNA specific for the Cl channel ClC-2 (GenBank accession no. X64139) and the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH; GenBank accession no. M17701) were amplified by 50 PCR cycles, using sequence-specific oligonucleotide primers with products of 375- and 340-bp lengths, respectively [ClC-2 sense: 5' CAA GCC TCC AGG AAG GTA C 3' (bp 2306-2324), ClC-2 antisense: 5' TCC CAA TGA GTC TGC CAA TG 3' (bp 2661-2680), GAPDH sense: 5' TCC GCC CCT TCC GCT GAT G 3' (bp 388-406), GAPDH antisense: 5' CAC GGA AGG CCA TGC CAG TGA 3' (bp 707-727); primers were purchased from Life Technologies]. The ClC-2 cDNA was amplified in the original RT tube by adding 48 µl of PCR reaction mixture to the remaining 8 µl cDNA. In parallel, GAPDH-PCR was started by adding 24 µl of PCR reaction mixture to the 2-µl aliquots. In 24 µl of the reaction mixture were 20.6 µl double-distilled H₂O, 0.2 µl dNTPs (25 mM, Promega), 0.25 µl sense primer, 0.25 µl antisense primer (both 10 µM), 0.125 µl AmpliTaq gold polymerase (5 U/µl), and 2.5 µl 10-fold buffer (15 mM MgCl₂) (both from Perkin-Elmer, Weiterstadt, Germany). Samples were incubated in a MJ Research thermal cycler (DNA Engine PTC 200; Biozym, Oldendorf, Germany), first for 10 min at 95°C to activate AmpliTaq gold polymerase (6), followed by 50 cycles for 45 s at 94°C, 1 min at 56°C, 1 min at 72°C, and a final extension at 72°C for 7 min. Starting at cycle 31, the extension time was prolonged by 3 s every cycle to compensate for degradation of the polymerase. Twelve milliliters of each PCR reaction were separated by electrophoresis in a nondenaturating 5% acrylamide gel. PCR product bands were stained with a fluorescence DNA dye (Vistragreen; Amersham, Braunschweig, Germany) and visualized with ImageQuant Software on a Storm Fluorophosphorimager (both from Molecular Dynamics, Krefeld, Germany). A specific ClC-2 product was not detectable after the first round of PCR amplification. Therefore, the products were purified by QIA quick-spin columns (Qiagen, Hilden, Germany), eluated in 30 µl double-distilled H₂O, and reamplified (15-µl PCR product, 48-µl PCR reaction mixture, 40 cycles), using a nested primer pair with a product length of 326 bp [ClC-2 sensenested 5' ATG GAA TCA GCA GGC ATT GC 3' (bp 2335-2354) and ClC-2 antisensenested 5' CTG GTG ACA TAA GCA TGG TC 3' (bp 2641-2660) (Life Technologies)].

PCR-negative controls. Aliquots of bath solution aspirated in the patch pipette just above the cells were processed identically to the cytoplasmic samples and served as extracellular controls, which were performed after every second cytoplasm harvested. Aliquots of pipette and bath solution were treated identically. For every PCR run, a water control was made by adding water instead of cDNA to the PCR reaction mixture. These solution controls, as well as the PCR water controls, were consistently negative. Fifteen cytoplasmic samples were processed for RT controls by incubating the RT step in the presence or absence of heat-inactivated reverse transcriptase enzyme.

PCR-positive controls. RNA from UB cultures was used after oligo-dT_12-16-primed reverse transcription and serial dilution of the cDNA for selection of oligonucleotide primer with maximal sensitivity, optimal annealing temperature, Mg²⁺ concentration, and pH. The cDNA also served as positive control for the PCR reaction in the single cell experiments. RNA was isolated (see Ref. 8) from primary monolayer cultures of embryonic day 17 (E17) rat UB (12), using a commercially prepared phenol guanidine isothiocyanate reagent (TRIsolve Reagent; Molecular Research Center, Cincinnati, OH).

Ultrastructure. For standard scanning electron microscopy, UBs were fixed at room temperature by modified Karnovsky reagent (2% paraformaldehyde and 2.5% glutaraldehyde in 80 mM phosphate buffer pH 7.4), postfixed by 1% OsO₄ (in 100 mM phosphate buffer, pH 7.4), dehydrated by increasing ethanol, and critical-point dried.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Ultrastructure. Scanning electron micrographs of microdissected UBs show the basolateral membrane of the cells at the tip (Fig. 2A). The basal lamina had been removed by dissection to provide free access for the patch pipette. Basolateral membranes of these tip cells, strikingly, have long cilia and are folded to short microvilli (Fig. 2B), as has been described previously for the apical membrane of cortical collecting duct (CCD) principal cells (9). The tip cells are small (~5 × 5 × 8 µm) and cubical to columnar in shape, and they enclose a small bud lumen (Fig. 4A).

View larger version (172K): [in this window] [in a new window]   Fig. 2.   A: scanning electron micrographs of a microdissected rat ureteric bud. B: high-magnification micrograph of same bud as in A. Basolateral aspect of cells at bud tip presents a cilium.

Electrophysiology. Gigaohm seal formation between the basolateral membrane of tip cells and the patch pipette occurred in ~1 of 10 approaches. Whole cell membrane capacity was low (6.62 ± 0.67 pF; n = 35), due to the small cell size. Whole cell conductance, in contrast, was very high (13.8 ± 2.1 nS/10 pF; n = 8; calculated for the outward currents in physiological NaCl bath solution, Fig. 3, A and B). The reversal potentials (V_rev) (KCl in the pipette) of the current-voltage relations, recorded with sodium or the impermeant NMDG as principal cation in the bath, were both near their Cl equilibrium potentials [E_Cl = 5 mV (NaCl) and E_Cl = 2 mV (NMDG-Cl); Fig. 3B)], suggesting Cl selectivity for the major fraction of whole cell conductance. In both bath solutions, whole cell currents rectified outwardly and activated time dependently at positive (more than or equal to +40 mV) and inactivated at negative voltages (less than or equal to 40 mV; Fig. 3, A and B), indicating Cl selectivity at least for the inactivating inward currents. Time constants, as calculated in the eight cells recorded in physiological NaCl bath solution, were  = 38 ± 9 ms for activation and  = 20 ± 2 ms for inactivation.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Electrophysiological characterization of cells at the ureteric bud tip (tUB). A: original whole cell current traces recorded by the patch-clamp technique with KCl in the pipette and NaCl in the bath solution. Currents were evoked by applying voltage steps (for 400 ms each) from the holding potential (0 mV) to voltages of 100, 80, 60, 40, 20, ±0, +20, +40, +60, +80, or +100 mV. (Potentials are related to pipette with respect to the bath. Traces of the single voltage steps are superimposed). Inward currents, i.e., flow of positive charge from the outside to the cytosolic membrane face, are negative currents and depicted as downward deflections of the current trace. B: mean current-voltage relation of tUB cells recorded as in A (, n = 8 ± SE) or with N-methyl-D-glucamine chloride bath solution (NMDG-Cl; , n = 27). Currents were analyzed at the 380-ms time point of each voltage step and normalized by membrane capacitance. C: original whole cell recording during voltage square pulse from 0-mV holding potential to 100 mV. Slowly activating inward current is indicated by the dotted line (current trace taken from A). D: slowly activating fraction of inward current (Iactivating) recorded with NaCl or NMDG-Cl bath solution (± SE, n = no. of cells). Iactivating was calculated by subtracting the current of the 100-mV square pulse at 60 ms from that at 380 ms. Data shown in D regrouped in those cells which were positive (n = 4) or negative (n = 16) for ClC-2 mRNA, as assessed by single cell RT-PCR. F: K-selective channels in the basolateral membrane of UB cells recorded in the outside-out mode. Original recording traces of two different channel types at different voltages (KCl pipette and NaCl bath solution). G: calculated current-voltage relations of 3 channels from 2 outside-out patches recorded as in E.

A second whole cell current component, which activated slowly at strong hyperpolarizing voltages, was seen in 20 of the 35 recorded cells (Fig. 3C). This slowly activating fraction (I_activating) of inward currents was identical in NaCl (5 cells) and NMDG-Cl bath solution (15 cells), indicating Cl selectivity (Fig. 3D). The time course of activation was  = 0.1 ± 0.008 s, as calculated in 10 cells, in which I_activating > 30 pA. The general characteristics of this current fraction, i.e., its strong hyperpolarization-induced activation, resembled that of ClC-2 Cl channels (30). To confirm the mRNA expression of this channel type, ClC-2 mRNA was evaluated in single cells at the tUB by RT-PCR (see below).

At the single channel level, two channel types with low (28 and 26 pS) and intermediate (63 pS) conductances, respectively, were identified in two outside-out patches of the basolateral membrane (Fig. 3F). The (extrapolated) V_rev of their current voltage relations was near the K equilibrium potential, which suggested K selectivity (Fig. 3G).

Single cell RT-PCR. As a prerequisite to interprete the negative controls, the highly abundant (and presumably high-yield) template encoding the GAPDH gene was amplified from a 2-µl aliquot of the 10-µl RT product. After 50 PCR cycles, the 340-bp GAPDH sequence-specific product was identified in 23 out of 57 cytoplasmic samples (Fig. 4B). Cells in which the whole cell recording protocol was run yielded fewer GAPDH products (11 out of 35), compared with those where RT started immediately after rupturing of the cell membrane (12 out of 22 cells). Contamination by genomic DNA could be excluded by identical processing of cytoplasmic samples (n = 15) without RT enzyme or with heat-inactivated enzyme (RT controls), which yielded no GAPDH product. Extracellular controls (n = 25) verified that GAPDH mRNA in the positive samples indeed originated from the cytoplasm of the aspirated cells and not from bath contaminated by leaky cells. With one exception, none of these negative controls (n = 40) yielded a GAPDH-specific PCR product. The only extracellular bath control positive for GAPDH was probably due to a contamination by mRNA of lysed cells in the bath and not from contaminating PCR products, as ruled out by the water and solution controls (see METHODS).

View larger version (103K): [in this window] [in a new window]   Fig. 4.   A: RNA harvesting by patch pipette; light microscopic view of a microdissected ureteric bud (bar, 10 µm). Cytoplasm had been aspirated several micrometers into the tip of the patch pipette (arrow) during formation of gigaohm seal and whole cell mode. Inset: entire ureteric tree within the mesenchyme. Only one ureteric bud has been dissected free of mesenchyme. B: acrylamide gel showing the GAPDH (top) and the ClC-2 (bottom) product amplified by single cell RT-PCR. Data from 8 different cytoplasmic samples and 6 controls are shown. Note characteristic ClC-2 double band in cytosol no. 6 lane, representing the predicted ClC-2 sequence and a shorter fragment in which 63 nucleotides are deleted.

mRNA of the volume-regulatory Cl channel ClC-2 was assumed to be much less abundant compared with GAPDH in tUB cells. Therefore, a nested PCR strategy was employed to detect the ClC-2 message. After amplification by 50 PCR cycles, purification of the PCR product, and reamplification by 40 cycles, ClC-2 sequence-specific PCR products were observed in 8 out of 57 cytoplasmic samples. No product could be detected in any of the extracellular controls (n = 25) or in the RT controls (n = 15; Fig. 4B).

Two PCR products were identified. Besides the expected 326-bp band seen in four samples, a 260-bp band was apparent in five experiments (in one sample, both products were amplified). The sequence of the longer PCR product was 99% identical to the published ClC-2 cDNA. In the shorter product, 63 bp (bp 2398-2460) were deleted by suspected alternative splicing, and the following CTG GCG (bp 2461-2466) was exchanged to TCA GAA, resulting in a loss of 21 amino acids and an exchange of two AA at the cytoplasmic COOH terminus of the channel between the putative transmembrane domain D12 and region D13.

In an attempt to simultaneously determine whole cell recordings and then do single cell PCR on these same cells, cytoplasm was aspirated from 35 cells after recording. Four of the 35 cells yielded ClC-2 PCR products. All four of these ClC-2-positive cells exhibited I_activating at 100 mV voltage. This current fraction, however, was not different from that of the 16 cells that expressed I_activating but happened to be negative for ClC-2 mRNA (Fig. 3E).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

This study presents for the first time a combined analysis by patch-clamp and single cell RT-PCR of isolated ureteric buds demonstrating ClC-2 mRNA expression in cells at the tUB. The tUB cell denotes a unique, complex, and transitory biological situation. 1) The cells maintain a high mitotic activity as long as branching morphogenesis (4) continues, i.e., until the distal end of the ureteric bud has fused with the mesenchyme-derived nephron, most likely in the region of the connecting tubule. 2) tUB cells represent the least differentiated progeny during branching morphogenesis. The tUB ("ampullary") cells in the rabbit have been shown to express the -subunit of the Na-K-ATPase in the entire plasma membrane, indicating that the pump is not yet distributed in the polar pattern characteristic for differentiated epithelial cells (21). 3) tUB cells face the cap of condensed mesenchymal cells across a small interspace (17), and they most likely produce two distinct signals to induce nephrogenesis (3). Moreover, the tUB cell itself is the target for mesenchyme-derived ligands of receptor tyrosine kinases (28).

In the present study, cells from the very tip of the ureteric bud were analyzed to ultimately compare this cell population with those located in other segments of the branching ureteric tree. The electrophysiological characterization of tUB revealed that the main fractional whole cell current rectified outwardly, and it is activated at depolarizing and inactivated at hyperpolarizing voltages (Fig. 3A). Conductance properties of rat UB cells (day E17) have been studied previously in monolayer cultures derived from microdissected UB (14). Whole cell currents of the tUB cells are identical to those of the cultured cells. The outwardly rectifying, depolarization-activated main current fraction of the latter has been demonstrated to be Cl selective (14) and resembles a Ca-activated type by its voltage and time dependence, which is typical for nonepithelial cells or for epithelia grown in monolayer on impermeable support but not for differentiated, i.e., highly polarized cells (1). Microdissected tUB cells, therefore, expressed a nonpolarized phenotype, as judged by their conductive properties. This nonpolarized phenotype is gradually downregulated with developmental differentiation of the early postnatal principal cell, whereas the epithelial phenotype is increasingly expressed by acquisition of the typical pattern of apical ion channels (13, 14). Taken together, these data suggest a specifically embryonic role for the outwardly rectifying, depolarization-activated conductance in tUB cells.

K channels with low and intermediate conductance were identified in two outside-out patches of tUB cells. Interestingly, their conductance resembled the basolateral K channels of mature CCD principal cells (11), suggesting that essential membrane proteins in the basolateral membrane are expressed earlier in epitheliogenesis than those in the apical membrane (13).

Another minor fractional whole cell current was measured in 20 out of 35 tUB cells. This current fraction, which activated slowly at strong hyperpolarizing voltages (Fig. 3C), was Cl selective, and, by these properties, it resembled a ClC-2-generated current expressed in oocytes (30) (except for the fact that activation kinetics of ClC-2 expressed in oocytes appear to be lower by more than a factor of 20). Importantly, activation kinetics of the slowly activating inward currents in tUB cells were quite similar to those of endogenous ClC-2-like whole cell currents in other epithelial cells (2, 7, 18). ClC-2 Cl channels were suggested to play an important role in lung development (22), and it is of interest that ClC-2 mRNA is differentially expressed early in collecting duct embryogenesis (12), as demonstrated in primary monolayer cultures of UB cells by semiquantitative RT-PCR.

Considering the data on single cell mRNA of the present work, it must be mentioned that the single cell RT-PCR technique (20) includes several steps, each with the potential for multiple errors. The consistently successful application of the method, therefore, requires a set of positive and negative controls that verify the variability and provide an estimate of relative efficiency. False-positive PCR amplifications were ruled out by special precautions (19). All solutions were negatively screened for template contamination by RT and PCR amplification. The cytoplasmic origin of a PCR product was proven by negative extracellular controls (see METHODS). The amplification of genomic DNA, which, per se, has been reported to be very unlikely (16), was ruled out in this study by not reverse-transcribed cytoplasmic samples (RT). Only one of a total of 40 negative controls was falsely positive for GAPDH, and none was falsely positive for ClC-2, whereas 23 and 8, respectively, out of a total of 57 cytoplasmic samples yielded positive PCR products. Thus the positive PCR products could be assigned to the cytoplasm-harvested tUB cells. The striking difference between the number of ClC-2- and GAPDH-positive cells might be due to the following. 1) RNA harvested from the cell might be in the range of detection for the highly abundant GAPDH but not so for ClC-2 mRNA. 2) Detection limits differ between ClC-2 and GAPDH due to a difference in efficiency of RT or PCR. Moreover, the protocol of harvesting cytoplasm appeared to be critical for the success of single cell RT-PCR. The highly abundant GAPDH mRNA was identified in ~55% of the cytoplasms sampled immediately after rupturing the cell membrane [which was within the range reported (27) for housekeeper mRNA detection by single cell RT-PCR in epithelial cells], but it was identified in only ~30% of the recorded cells. This might point to degradation of mRNA during the time span between rupturing the plasma membrane and the start of cDNA synthesis.

ClC-2 message was identified only in few cells, whereas slowly activating inward currents were detected in more than half of the recorded cells. For the reasons mentioned above, the absence of a PCR product in cytoplasmic samples from those cells cannot be interpreted as absence of template, i.e., the number of falsly negative cytoplasmic samples cannot be estimated (23). Nevertheless, the ratio between the cytoplasmic samples positive for a template and the total number of samples mirrors the abundance of template in the investigated cell population. This ratio, therefore, can be applied to compare the expression of a mRNA between two cell populations if identical experimental protocols have been used.

To summarize, this work presents functional characteristics of the cells at the tUB, and data suggest that they express some specific properties. 1) The tUB cell represents a Ca-activated-like Cl conductance, which is phenotypical for embryonic nonpolarized cells but not for differentiated polarized cells. 2) The single tUB cell expresses ClC-2 mRNA, consistent with the notion that this channel is widely expressed not only in epithelial but also in embryonic and in nonepithelial cells.