Heterogeneous Expression of Ca 2+ Handling Proteins in Rabbit Sinoatrial Node

Abstract

We investigated the densities of the L-type Ca²⁺ current, i_Ca,L, and various Ca²⁺ handling proteins in rabbit sinoatrial (SA) node. The density of i_Ca,L, recorded with the whole-cell patch-clamp technique, varied widely in sinoatrial node cells. The density of i_Ca,L was significantly (p<0.001) correlated with cell capacitance (measure of cell size) and the density was greater in larger cells (likely to be from the periphery of the SA node) than in smaller cells (likely to be from the center of the SA node). Immunocytochemical labeling of the L-type Ca²⁺ channel, Na⁺-Ca²⁺ exchanger, sarcoplasmic reticulum Ca²⁺ release channel (RYR2), and sarcoplasmic reticulum Ca²⁺ pump (SERCA2) also varied widely in SA node cells. In all cases there was significantly (p<0.05) denser labeling of cells from the periphery of the SA node than of cells from the center. In contrast, immunocytochemical labeling of the Na⁺-K⁺ pump was similar in peripheral and central cells. We conclude that Ca²⁺ handling proteins are sparse and poorly organized in the center of the SA node (normally the leading pacemaker site), whereas they are more abundant in the periphery (at the border of the SA node with the surrounding atrial muscle).

The Sinoatrial (SA) Node is a heterogeneous tissue, and the action potential and pacemaker activity change from the periphery (at the border of the SA node with the surrounding atrial muscle) to the center (normally the leading pacemaker site). The heterogeneity is essential for the normal functioning of the SA node. It helps the SA node drive the surrounding atrial muscle and not be suppressed by it, it results in multiple pacemaker mechanisms appropriate for different conditions, and it helps prevent re-entrant arrhythmias involving the SA node (Boyett et al. 2000). A breakdown of the normal heterogeneity may be responsible for the functionl deterioration of the SA node with age (Zhang et al. 1998). According to the gradient model, the heterogeneity is the result of a gradient in the intrinsic properties of SA node cells from the periphery to the center. Ca²⁺ is important in the SA node. The L-type Ca²⁺ current, i_Ca,L, supports the action potential and pacemaker activity, and there is evidence that Ca²⁺ release from the sarcoplasmic reticulum (SR) is involved in pacemaking, e.g., by activating inward Na⁺-Ca²⁺ exchange current (Hata et al. 1996; Rigg and Terrar 1996; Li et al. 1997; Satoh 1997; Ju and Allen 1998,1999; Rigg et al. 2000; Bogdanov et al. 2001). Models of the action potential in the periphery and center of the SA node predict that in the center the density of i_Ca,L should be ~30% of that in the periphery (Zhang et al. 2000). The aims of the study were to (a) determine whether the density of i_Ca,L and the expression of the L-type Ca²⁺ channel protein vary as predicted and (b) characterize the expression of other Ca²⁺ handling proteins (sarcolemmal Na⁺-Ca²⁺ exchanger, SR Ca²⁺ release channel, and SR Ca²⁺ pump), as well as a protein indirectly involved in Ca²⁺ handling (Na⁺-K⁺ pump). Evidence is presented for decreases in the density of all proteins, apart from the Na⁺-K⁺ pump, from the periphery to the center of the SA node.

Materials and Methods

SA node cells were isolated as described by Lei and Brown (1996) from 500–1000-g New Zealand White rabbits. For the experiments shown in Figure 1, cells were isolated from throughout the SA node: from a piece of tissue ~1 mm (parallel to the crista terminalis) × 2 mm (perpendicular to the crista terminalis) roughly from the region within the red outline in Figure 4. For experiments shown in the remaining figures, cells were isolated separately from the periphery and centre of the SA node. A piece of tissue (like that above) was divided into three strips [peripheral (closest to the crista terminalis, transitional, and central (most distant from the crista terminalis)] and cells were isolated from the peripheral and central strips after carefully removing subepicardial atrial muscle. Rabbit ventricular and atrial cells were isolated using the Langendorff procedure as previously described by Linz and Meyer (1998). Dissection of the intact SA node (shown in Figure 4) was carried out as previously described (Boyett et al. 1999). The intact SA node was embedded in 10% gelatin (diluted in 0.01 M PBS), frozen in isopentane using liquid N2, and stored at — 80C before sectioning.

The whole-cell patch-clamp technique was used for the recording of i_Ca,L from single SA node cells with amphotericin-permeabilized patches. An Axopatch-1C amplifier, a Digidata1200A, and pCLAMP software (Axon Instruments; Union City, CA) were used. Pipettes had a tip diameter of ~1–2 μm and a resistance of 3–8 MΩ. Pipette solution contained (in mM): 140 KCl, 1.8 MgSO₄, 5 HEPES, 1 EGTA, pH 7.4, KOH, amphotericin 200 μg/ml. C_m was obtained from the capacity compensation control of the amplifier. In a previous study (Honjo et al. 1996) the accuracy of this method was checked. The series resistance was electronically compensated (>80%) and the current signal was filtered by a low-pass Bessel filter with a cutoff frequency of 5 kHz (— 3 dB). Cells were superfused with Tyrode's solution (in mM: 140 NaCl, 5.4 KCl, 1.8 CaCl₂, 1 MgCl₂, 10 glucose, 5 HEPES, pH 7.4, NaOH) at ~1 ml/min at 35C. Three hundred μM lidocaine was added to Tyrode's solution to block the Na⁺ current, i_Na, and 5 μM E-4031 (Eisai Pharmaceuticals; Tokyo, Japan) was added to block the rapid delayed rectifier K⁺ current, i_K,r. It was confirmed that i_Ca,L could be blocked by 500 nM nisoldipine (Bayer; Newbury, UK). Lidocaine was dissolved in distilled water and ethanol (50:50), E-4031 was dissolved in distilled water, and nisoldipine was dissolved in ethanol to make 10 mM stock solutions.

Immunocytochemistry experiments were carried out using established methods. Single cells were plated on Bunsen burner flame-treated coverslips and allowed to settle for 30 min. From the intact SA node (Figure 4), 18–20 μm tissue sections were cut perpendicular to the crista terminalis through the crista terminalis and intercaval region. Cells and tissue were fixed in 2% paraformaldehyde and washed three times with 0.01 M PBS. Cells and tissue were permeabilized by incubating them in PBS containing 0.1% Triton X-100 for 30 min, washed with PBS, and then blocked in 10% normal donkey serum in PBS for 1 hr. After washing three times with PBS, cells and tissue were incubated with rabbit polyclonal anti-L-type Ca²⁺ channel α_1C-subunit (anti-CNC1 peptide from Snutch et al. 1991; 1:200; provided by Alomone Labs, Jerusalem, Israel), mouse monoclonal anti-Na⁺-Ca²⁺ exchanger (clone C2C12; Frank et al. 1992; 1:100; provided by Affinity Bioreagents, Golden, CO), mouse monoclonal anti-Na⁺-K⁺ pump α1-subunit (antibody α6F from Lebovitz et al. 1989; 1:20; provided by Developmental Studies Hybridoma Bank, University of Iowa, Ames, IA), mouse monoclonal anti-RYR2 (antibody AE 8.91 from Lewis Carl et al. 1995; 1:100; gift from G. Meissner), mouse monoclonal anti-SERCA2 (from Jorgensen et al. 1988; cells 1:6000, tissue 1:12000; provided by Affinity Bioreagents), or mouse monoclonal anti-connexin43 antibody (amino acids 252–270; Cx43 1:1000, provided by Chemicon, Harrow, UK) overnight at 4C. The slides were brought to room temperature, washed three times in PBS, and incubated with secondary antibody, donkey anti-rabbit IgG (1:100) or donkey anti-mouse IgG (1:100) conjugated to FITC (Chemicon) for 1 hr. After washing three times in PBS, coverslips were mounted on microscope slides, sealed with nailpolish, and stored in the dark at 4C for subsequent viewing in a Leica TCS SP laser scanning confocal microscope. All images presented are single optical sections. In the case of single cells, the images were taken approximately midway through the depth of the cell. In the case of tissue sections, neighboring pairs of sections were labeled, one for Cx43 and one for the protein of interest. The presence or absence of Cx43 label was used as a marker, as described in the Results. No labeling above background was obtained when the primary or secondary antibodies were omitted (data not shown). When primary antibody was replaced with normal mouse or rabbit serum (Burry 2000) there was also no labeling above background (data not shown). Image analysis was carried out using Scion Image (NIH; Bethesda, MD).

Figure 1

Correlation of i_Ca,L density and C_m. (A) Records of i_Ca,L from large and small cells (C_m of cells shown) elicited by depolarizing clamp pulses (bottom). (B) Current density-voltage relationships for i_Ca,L in groups of large (n = 6) and small (n = 6) cells (range of C_m of cells shown). Means ± SEM shown. (C) Plot of density of i_Ca,L at 0 mV against C_m for different cells. Filled circles, data from present study; open circles, data from rabbit SA node cells from Honjo et al. (1996); open squares, data from peripheral (C_m, 65 pF) and central (C_m, 20 pF) rabbit SA node cell models from Zhang et al. (2000).

Figure 2

Labeling of L-type Ca²⁺ channel. (A) Ventricular cell; (B) atrial cell; (C) peripheral SA node cell. (D,E) Central SA node cells. The nucleus is stained red by propidium iodide in some panels in this and other figures.

Figure 3

Labeling of Na⁺-Ca²⁺ exchanger. (A) Ventricular cell (B) atrial cell; (C) peripheral SA node cell (D) central SA node cell. (E-G) Sections through crista terminalis (E), periphery (F) and center (G) of the SA node. (H) Section cut perpendicular to the crista terminalis labeled for Cx43. RA, right atrial appendage; CT, crista terminalis; endo, endocardium; epi, epicardium; SEP, interatrial septum.

Figure 4

Photograph of SA node preparation showing positions of the Cx43 (red outline) and Na⁺-Ca²⁺ exchange (green outline) negative regions SVC, superior vena cava; SEP, interatrial septum; IVC, inferior vena cava; RA, right atrial appendage; CT, crista terminalis.

All results are presented as means ± SEM (number of cells). Statistical significance was determined by ANOVA. Linear regression analysis was used for correlations. p<0.05 was considered to indicate a significant difference. All statistical analysis was carried out using SigmaStat software (Jandel Scientific; Chicago, IL).

Results

Correlation Between Density of i_Ca,L and C_m

Experiments were carried out only on cells showing spontaneous activity. They were spindle- and/or spider-shaped, with no obvious or faint striations. Cells were isolated from the whole of the SA node and distinguished on the basis of size as measured by C_m (large cells tend to be from the periphery and small cells tend to be from the center). i_Ca,L was elicited by 500-msec depolarizing clamp pulses to various test potentials from a holding potential of —50 mV at 1 Hz (Figure 1A. Figure 1A shows records of i_Ca,L from large and small cells and Figure 1B shows current-voltage relationships for two groups of cells: large cells with C_m >30 pF and small cells with C_m <30 pF (we have classified cells in a similar manner previously (Honjo et al. 1996,1999; Lei et al. 2001). The mean density of peak inward current (peak inward current normalized by C_m for each cell) is plotted against the test potential. The density of i_Ca,L was significantly less in the smaller cells (p<0.01). The filled symbols in Figure 1C show the density of i_Ca,L in 12 cells at 0 mV plotted against C_m. There is a significant correlation between the density of i_Ca,L and C_m in the 12 cells (p<0.001; r² = 0.86). This suggests that in the center of the SA node the density of i_Ca,L is less than in the periphery. The other symbols in Figure 1C are considered in the Discussion.

Sarcolemmal Ca²⁺ Handling Proteins

The abundance of proteins in the periphery and center of the SA node was assessed by immunocytochemistry. Figures 2, 3, and 5–7 show labeling of various cell types (ventricular, atrial, peripheral SA node, central SA node) with various antibodies. Ventricular and atrial cells were labeled as a reference. Rather than isolating cells from the whole of the SA node, cells were separately isolated from tissue taken from the periphery and center of the SA node.

Figure 2 shows labeling with an antibody to the L-type Ca²⁺ channel α_1C-subunit. In ventricular cells there was labeling of the outer cell membrane in some cells and internal striated punctate labeling with a periodicity of 1.73 ± 0.04 μm (n=5) (Figure 2A, corresponding to L-type Ca²⁺ channels in the outer cell membrane and t-tubules, respectively. In atrial cells there was punctate labeling of the outer cell membrane (Figure 2B. In some atrial cells (Figure 2B there was limited punctate labeling in the cell interior, possibly corresponding to L-type Ca²⁺ channels in poorly developed t-tubules. Although atrial cells are widely considered not to have t-tubules, some rat atrial cells are reported to have well or poorly developed t-tubules (Forssman and Girardier 1970) and guinea pig atrial cells are reported to have short t-tubules (Forbes and Van Niel 1988). In peripheral SA node cells there was punctate labeling of the outer cell membrane and little or no internal labeling (Figure 2C, whereas in central cells there was little or no detectable labeling (Figures 2D and 2E).

Figure 3 shows labeling with an antibody to the Na⁺-Ca²⁺ exchanger. In ventricular cells there was labeling of the outer cell membrane as well as internal striated (periodicity 1.87 ± 0.02 μm; n=5) punctate labeling (Figure 3A, corresponding to Na⁺-Ca²⁺ exchanger in the outer cell membrane and t-tubules, respectively. In atrial cells there was labeling of the outer cell membrane only (Figure 3B. In peripheral SA node cells there was also labeling of the outer cell membrane (Figure 3C, whereas in central cells there was little or no labeling (Figure 3D.

The Na⁺-Ca²⁺ exchanger was also labeled in tissue sections. In the case of tissue sections, pairs of neighboring sections were labeled, one for the gap junction protein connexin43 (Cx43) and one for the protein of interest. Cx43 was used as a marker. Figure 3H shows a tissue section through the thick crista terminalis and thin intercaval region labeled for Cx43. The center of the SA node is known to be located in the intercaval region adjacent to the crista terminalis. In Figure 3H (arrow G) this region lacks green labeling; it is well known that the center of the SA node lacks Cx43 (e.g., Coppen et al. 1999). From the intercaval region, SA node tissue rises up the endocardial face of the crista terminalis, where it terminates (Kodama and Boyett 1985). The SA node tissue on the crista terminalis is the periphery of the SA node (Figure 3H, arrow F) (Kodama and Boyett 1985). The peripheral SA node tissue is labeled green in Figure 3H—the periphery is known to contain Cx43 (Coppen et al. 1999)—and it is separated by connective tissue from the bulk of the crista terminalis. The bulk of the crista terminalis is atrial muscle and in Figure 3H (arrow E) is also labeled green; atrial muscle also contains Cx43 (e.g. Coppen et al. 1999). Figures 3E-3G show labeling of the atrial muscle of the crista terminalis (Figure 3E and the periphery (Figure 3F and center (Figure 3G of the SA node by the antibody to the Na⁺-Ca²⁺ exchanger in the neighboring tissue section (Figures 3E-3G approximately correspond to the locations marked by the arrows E to G in Figure 3H). Whereas there was clear labeling of the outer cell membrane of the atrial cells (Figure 3E, there was little labeling of cells in the periphery or center of the SA node (Figures 3F and 3G). In two SA node preparations the extent of the region lacking labeling of the Na⁺-Ca²⁺ exchanger was mapped. Figure 4 summarizes data from one preparation. Figure 4 shows a photograph of a preparation. Multiple sections (like that in Figure 3H) were cut perpendicular to the crista terminalis and labeled for either Cx43 or the Na⁺-Ca²⁺ exchanger. The red outline in Figure 4 shows the extent of the Cx43-negative region, whereas the green outline shows the extent of the region lacking labeling of the Na⁺-Ca²⁺ exchanger. Similar data were obtained from the second preparation.

Figure 5A shows that in ventricular cells there was little labeling by an antibody to the Na⁺-K⁺ pump of the outer cell membrane, except at the intercalated discs. In addition, there was internal striated (periodicity 1.91 ± 0.03 μm; n=5) punctate labeling of t-tubules. Figure 5B shows that in atrial cells there was labeling of the outer cell membrane, particularly at intercalated discs. Whether the increase in labeling at the intercalated disc is the result of a high density of the Na⁺-K⁺ pump (per unit area of membrane) at this location or the membrane folding at the intercalated disc is not known. In peripheral (Figure 5C and central (Figure 5D SA node cells, the pattern of labeling was similar to that of atrial cells. Figures 5E-5G show that in tissue sections there was labeling of the outer cell membrane of atrial cells of the crista terminalis (Figure 5E and of peripheral (Figure 5F and central (Figure 5G SA node cells. The density of labeling was similar in the three regions, in accordance with the data from single cells. Figure 5H summarizes the labeling of Cx43 in the neighboring section. The approximate locations of the images shown in Figures 5E-5G are shown by arrows E-G.

Figure 5

Labeling of Na⁺-K⁺ pump. (A) Ventricular cell; (B) atrial cell; (C) peripheral SA node cell; (D) central SA node cell. (E-G) Sections through crista terminalis (E), periphery (F) and center (G) of the SA node. (H) Diagram of section cut perpendicular to the crista terminalis labeled for Cx43 (black, label; white, no label). Abbreviations as in Figure 3.

Figure 6

Labeling of RYR2. (A) Ventricular cell; (B) atrial cell; (C,D) peripheral SA node cells; (E) central SA node cell. (F-I) Sections through right atrial appendage (F), crista terminalis (G), and periphery (H) and center (I) of the SA node. (J) Diagram of section cut perpendicular to the crista terminalis labeled for Cx43 (black, label; white, no label). Abbreviations as in Figure 3.

Figure 7

Labeling of SERCA2. (A) Ventricular cell; (B) atrial cell; (C) peripheral SA node cell; (D) central SA node cell. (E-H) Sections through right atrial appendage (E), crista terminalis (F), and periphery (G) and center (H) of the SA node. Arrow in F highlights an example of labeling of the nuclear envelope. (I) Diagram of section cut perpendicular to the crista terminalis labeled for Cx43 (black, label; white, no label). Abbreviations as in Figure 3.

SR Ca²⁺ Handling Proteins

In ventricular cells, there was internal striated (periodicity 1.83 ± 0.03 μm; n=5) punctate labeling by an antibody to the SR Ca²⁺ release channel, anti-RYR2 (Figure 6A, corresponding to junctional SR (JSR-SR Ca²⁺ release site) adjacent to t-tubules. In atrial cells there was labeling adjacent to the outer cell membrane (Figure 6Bi), corresponding to JSR adjacent to the outer cell membrane, as well as internal striated (periodicity 1.94 ± 0.01 μm; n=5) labeling (Figure 6Bii), corresponding to extended JSR (or corbular SR). Extended JSR displays all the anatomic features of JSR but is not associated with the outer cell membrane or t-tubules and is organized at the level of the z-line (Forbes and Van Niel 1988; Lewis Carl et al. 1995). The internal labeling was separated by a gap from the labeling adjacent to the outer cell membrane. In some peripheral SA node cells there was a similar pattern of labeling as in atrial cells, with both labeling adjacent to the outer membrane (Figure 6Ci) and internal labeling (periodicity 1.86 ± 0.03 μm; n=5; Figure 6Cii). However, in other peripheral SA node cells, punctate labeling adjacent to the outer cell membrane (Figure 6Di) was observed together with random spots of internal labeling (Figure 6Dii). In central SA node cells, there was only punctate labeling adjacent to the outer cell membrane. There was no internal labeling above background intensity (Figure 6E. Figure 6 also shows labeling by anti-RYR2 in tissue sections. In atrial cells in longitudinal section (Figure 6F, from right atrial appendage) there was labeling adjacent to the outer cell membrane and internal striated (periodicity 1.87 ± 0.05 μm; n=5) labeling. In atrial cells in cross-section (Figure 6G, from the crista terminalis) a ring of labeling surrounding (but separated by a gap from) labeling of the cell interior was observed. In the middle of cells there was often no labeling (presumably corresponding to the nucleus). In the center of the SA node there appeared to be labeling adjacent to the outer cell membrane only. There was no evidence of internal labeling (Figure 6I. In the periphery of the SA node two types of labeling were observed: labeling similar to that of atrial cells (labeling adjacent to outer cell membrane and internal labeling; Figure 6Hi) and labeling similar to that of central SA node cells (labeling adjacent to outer cell membrane and no internal labeling; Figure 6Hii). These data are in accordance with the data from single cells. Figure 6J summarizes the labeling of Cx43 in the neighboring section; the approximate locations of the images shown in Figures 6G-6I are shown by arrows G-I.

In ventricular cells there was labeling by the antibody to the SR Ca²⁺ pump, anti-SERCA2, adjacent to the outer cell membrane, as well as internal striated (periodicity 1.89 ± 0.02 μm; n=5) labeling and rings of labeling around the nuclei (Figure 7A. This corresponds to network SR (SR Ca²⁺ uptake site, separate from JSR) adjacent to the outer cell membrane, network SR adjacent to the t-tubules, and the nuclear envelope, respectively. In atrial cells, a similar pattern of labeling was observed (Figure 7B. The internal striated (periodicity 1.93 ± 0.01 μm; n=5) labeling corresponds to network SR adjacent to the z-lines (site of extended JSR) in this case. In peripheral SA node cells there was internal labeling. However, the internal labeling was not striated. Instead, there was random and diffuse labeling of the cell interior (Figure 7C. There was also labeling of the nuclear envelope in peripheral cells. In central SA node cells there was a similar pattern of labeling to that in peripheral cells, but the density of labeling was less (Figure 7D. In tissue sections, in longitudinally or transversely sectioned atrial cells (Figures 7E and 7F), some labeling adjacent to the outer cell membrane, internal labeling (striated in longitudinally sectioned cells; periodicity 2.00 ± 0.03 μm; n=5) and labeling of the nuclear envelope (e.g., Figure 7F, arrow) could be discerned. In peripheral SA node tissue the pattern of labeling was similar to that in atrial tissue (Figure 7G, whereas in central tissue (Figure 7H less labeling was discernible (apart from clear labeling of the nuclear envelope). Figure 7I summarizes the labeling of Cx43 in the neighboring section. The approximate locations of the images shown in Figures 7F-7H are shown by arrows F-H.

Summary of Single Cell Labeling

With single cells, image analysis was used to measure the total area of labeling within an optical section, approximately midway through the depth of a cell, as well as the total area of the cell in the optical section, from which the density of labeling (percentage of cell area labeled) was calculated. Figure 8A shows that, as expected central SA node cells were significantly smaller in area than peripheral SA node cells. For all cell types, cell area of ventricular cells>atrial cells>peripheral SA node cells>central SA node cells. Figure 8B shows that the density of labeling by the antibody to the L-type Ca²⁺ channel α_1C-subunit was high in ventricular, atrial, and peripheral SA node cells, but significantly lower in central SA node cells. This finding concerning peripheral and central SA node cells is consistent with the electrophysiological data. Figure 8C shows that the density of labeling by the antibody to the Na⁺-Ca²⁺ exchanger of ventricular cells>atrial cells—peripheral SA node cells>central SA node cells. Figure 8D shows that the density of labeling by the antibody to the Na⁺-K⁺ pump was significantly higher in ventricular cells than in the other cell types, but not significantly different among atrial and peripheral and central SA node cells. Figure 8E shows that the density of labeling by anti-RYR2 was high in ventricular, atrial, and peripheral SA node cells and significantly less in central SA node cells. Figure 8F shows that the density of labeling by anti-SERCA2 in the central SA node cells was significantly less than in atrial and peripheral SA node cells.

Figure 8

Density of labeling of Ca²⁺ handling proteins. (A-F) Cell area (A) and density of labeling of L-type Ca²⁺ channel (B), Na⁺-Ca²⁺ exchanger (C), Na⁺-K⁺ pump (D), RYR2 (E), and SERCA2 (F) in ventricular, atrial, peripheral SA node, and central SA node cells. Means ± SEM shown. n numbers are shown below bars and statistical differences shown above bars. V, ventricular; A, atrial; P, peripheral SA node; C, central SA node; v, versus.

Discussion

The present data suggest that the L-type Ca²⁺ channel, Na⁺-Ca²⁺ exchanger, RYR2, and SERCA2 are less abundant in the center of the SA node than in the periphery, whereas the Na⁺-K⁺ pump is equally abundant in the two regions. The present data also show that the density of i_Ca,L is less in smaller cells. In the rabbit and other species, it is known that cells in the center of the SA node are small, whereas cells in the periphery are large (Boyett et al. 2000). Taken together, these findings suggest that the density of i_Ca,L is less in the center of the SA node than in the periphery. This is consistent with the evidence of a lower abundance of the L-type Ca²⁺ channel in the center than in the periphery.

Validity of Data

In the study of i_Ca,L, cells were isolated from the whole of the SA node and the cells were subsequently classified according to cell size (Figure 1). Previous studies suggest that the assumption that cell size is an indicator of the site of origin of a cell is reasonable (Honjo et al. 1996,1999; Boyett et al. 2000; Lei et al. 2000; Zhang et al. 2000). Classification of cells according to C_m has a number of advantages compared to isolating cells from the periphery and center of the SA node. For example, (a) it is technically straightforward, whereas the isolation of cells from the periphery and center is difficult, and (b) it is effectively continuous (depends only on resolution of measurement of C_m), whereas we have only managed to isolate cells from just two regions of the SA node.

In the present study, ventricular and atrial cells were labeled as a point of reference to check that the labeling pattern seen is consistent with previously published data. In the present study, labeling of the L-type Ca²⁺ channel and RYR2 in ventricular and atrial cells was similar to that of rabbit ventricle and atrium by Lewis Carl et al. (1995), and labeling of the Na⁺-Ca²⁺ exchanger in ventricular and atrial cells was similar to that of rabbit ventricular cells by Chen et al. (1995) and guinea pig atrial cells by McDonald et al. (2000). On the basis of Western blotting, the density of the Na⁺-Ca²⁺ exchanger has previously been reported in guinea pig right atrium to be 42% of that in right ventricle (McDonald et al. 2000), consistent with the data in Figure 8C (see also Wang et al. 1996). Lancaster et al. (2000a) (also M.K. Lancaster, personal communication) observed similar labeling of the Na⁺-K⁺ pump in rabbit ventricle to that observed in ventricular cells in the present study. The Na⁺-K⁺ pump in atrial cells has not been labeled previously. Other types of measurement suggest that the density of the Na⁺-K⁺ pump is less in the atria than in the ventricles (Wang et al. 1996), consistent with Figure 8D. SERCA2 in both ventricular and atrial cells has not been labeled previously. In the present study, single cells and tissues sections were labeled in most cases. The subcellular location of a protein can be better determined with single cells, whereas the position of cells within the SA node is known with greater certainty in the case of tissue sections. The labeling pattern of single cells and tissue sections was similar in the case of labeling of the Na⁺-K⁺ pump, RYR2, and SERCA2. In the case of labeling of the Na⁺-Ca²⁺ exchanger, the data from single cells and tissue sections agree in showing less labeling of central SA node cells than of atrial cells (Figure 3). However, in tissue sections no labeling of the Na⁺-Ca²⁺ exchanger was observed in either the periphery or the center of the SA node whereas, although no labeling was observed in single cells from the center, labeling was observed in single cells from the periphery (Figure 3). This discrepancy could be the result of better access of the antibody to the antigenic site in the case of single cells. In the present study there was little or no labeling of either the L-type Ca²⁺ channel or the Na⁺-Ca²⁺ exchanger in central SA node cells (Figures 2 and 3). This does not mean that the two proteins are absent from the center of the SA node. It is more likely that the density of the proteins in the center is below the detection limit of immunocytochemistry.

Comparison with Previous Studies

In a previous study of rabbit SA node cells, we failed to observe a correlation between i_Ca,L density and C_m (Honjo et al. 1996). The data from our previous study (with the exception of one point) are shown by the open circles in Figure 1C. However, when we subsequently developed models of action potentials in rabbit peripheral and central SA node cells, with C_m of 65 and 20 pF, respectively, the i_Ca,L density in the central model cell had to be set at ~30% of that in the peripheral model cell (Zhang et al. 2000). The i_Ca,L densities chosen for the peripheral and central model cells are shown by the open squares in Figure 1C. The discrepancy between experiment and model prompted the present re-examination of the relationship between i_Ca,L density and C_m. The new data (Figure 1C, filled circles) are consistent with the old data. There is a significant correlation between i_Ca,L density and C_m, regardless of whether the new data are considered alone (p<0.001) or whether the new and old data are considered together (with or without the point excluded in Figure 1C; p<0.001). The reason why no significant correlation was observed with the original data set is that the range of C_m investigated was not sufficient. The i_Ca,L densities observed in large and small cells in experiments are comparable to the predicted i_Ca,L densities from the peripheral and central models (Figure 1C. There are also significant correlations between the densities of i_Na, 4-AP-sensitive current, i_K,r, slow delayed rectifier K⁺ current (i_K,s), and hyperpolarization-activated current (i_f) and C_m in rabbit SA node cells (Honjo et al. 1996,1999; Lei et al. 2000, 2001).

In various species, including rabbit, in the center of the SA node the cells have been observed by electron microscopy to be “empty,” principally because they lack well-ordered myofilaments (Masson-Pévet et al. 1979; Boyett et al. 2000). In the rabbit at least, there is a progressive decrease in the density of myofilaments from the periphery to the center (Boyett et al. 2000). As well as lacking myofilaments, electron microscopy shows that cells in the center of the SA node in the rabbit also contain little SR (about a third of that in atrial cells) (Masson-Pévet et al. 1979). The data from the present study are consistent with these findings and provide further information on the organization of the SR in SA node cells. In ventricular cells, SR is known to be abundant and organized in phase with the t-tubules, and this is confirmed by the distribution of the two SR proteins, RYR2 and SERCA2, in this (Figures 6 and 7) and other studies (RYR2; Lewis Carl et al. 1995). In atrial cells, SR is also known to be abundant and well organized. Within atrial cells, despite the general lack of well developed t-tubules, SR is organized in a periodic fashion (in phase with the myofilaments) (Forbes and Van Niel 1988). RYR2 and SERCA2 labeling, in this (Figures 6 and 7) and other studies (RYR2; Lewis Carl et al. 1995), confirms the extended JSR and network SR to be organized in a striated fashion in atrial cells. In atrial cells, in addition to the transversely organized SR, there is SR adjacent to the outer cell membrane, as judged by RYR2 labeling (Figure 6); this is JSR. In contrast to the abundant well-organized SR in ventricular and atrial cells, the SR in SA node cells, as judged by RYR2 and SERCA2 labeling, is sparse and poorly organized (Figures 6 and 7). RYR2 labeling is principally found adjacent to the outer cell membrane (Figure 6), whereas SERCA2 is randomly scattered throughout the cytoplasm (Figure 7).

In contrast to cell size, densities of myofilaments, SR, i_Na, 4-AP-sensitive current, i_K,r, i_K,s, i_f, i_Ca,L, L-type Ca²⁺ channel, Na⁺-Ca²⁺ exchanger, RYR2, and SERCA2 (this study and Masson-Pévet et al. 1979; Bleeker et al. 1980; Honjo et al. 1996,1999; Lei et al. 2000,2001), the density of the Na⁺-K⁺ pump does not appear to decrease from the periphery to the center of the rabbit SA node (Figure 5). All of the features that decrease from the periphery to the center are associated with “excitability”—perhaps the Na⁺-K⁺ pump does not conform to the general rule because it is also required for vegetative functions of a cell.

Physiological Importance

In the center of the SA node, i_Ca,L is solely responsible for generation of the action potential, whereas in the periphery i_Na is principally responsible (although in the periphery i_Ca,L is responsible for the action potential plateau as it is in the center), and in this respect it is paradoxical that the density of i_Ca,L is less in the center than in the periphery. However, in the center various current densities are lower than in the periphery, as explained above, and there is no need for the density of i_Ca,L to be as high as in the periphery for i_Ca,L to be able to generate the action potential. In the models of action potentials in the periphery and center of the rabbit SA node (Zhang et al. 2000), if the density of i_Ca,L in the central cell model is set to be the same as in the peripheral cell model, the action potential amplitude and duration become too great and spontaneous activity becomes too slow (as a consequence of the increase in action potential duration).

The regional differences in Ca²⁺ handling proteins within the SA node are expected to result in regional differences in Ca²⁺ handling in the SA node. Prompted by the results from the present study, we have recently measured intracellular Ca²⁺ transients in rabbit SA node cells. In cells isolated from the whole of the rabbit SA node, there were significant correlations between the amplitude of the Ca²⁺ transient, the peak systolic Ca²⁺ concentration, the diastolic Ca²⁺ concentration, the duration of the Ca²⁺ transient, time to peak of the Ca²⁺ transient, and decay time of the Ca²⁺ transient with cell size (Lancaster et al. 2000b). Ca²⁺ concentrations were less and kinetic parameters were slower in small cells. In addition, ryanodine significantly increased the duration and reduced the amplitude of the Ca²⁺ transient in large cells but not in small cells (Lancaster et al. 2000b). These results are consistent with the results from the present study. The paucity of L-type Ca²⁺ channel, RYR2, and SERCA2 in smaller central cells is expected to result in a smaller Ca²⁺ transient. The paucity of RYR2 in smaller central cells is expected to result in a prolongation of the time to peak of the Ca²⁺ transient (because the Ca²⁺ transient will be the result of Ca²⁺ entry via the L-type Ca²⁺ channel). The paucity of SERCA2 and the Na⁺-Ca²⁺ exchanger in smaller central cells is expected to result in a prolongation in the relaxation of the Ca²⁺ transient. The paucity of RYR2 in smaller central cells is expected to result in a decrease in ryanodine sensitivity.

Various studies have shown that Ca²⁺ release from the SR plays an important role in pacemaking (by modulating i_Ca,L, Na⁺-Ca²⁺ exchange current, delayed rectifier K⁺ current, and if) in various species, including rabbit (Hata et al. 1996; Rigg and Terrar 1996; Li et al. 1997; Satoh 1997; Ju and Allen 1998,1999; Bogdanov et al. 2001). The present data suggest that intracellular Ca²⁺ may be handled differently from the periphery to the center and, therefore, the role of intracellular Ca²⁺ in pacemaking may also vary. In support of this, in the study of Lancaster et al. (2000b), ryanodine significantly slowed pacemaking in large but not in small cells.

Gradient vs Mosaic Models of SA Node

In addition to the gradient model of the SA node, a mosaic model has been proposed in which the regional differences in electrical activity of the SA node are not the result of a gradient in the intrinsic properties of SA node cells from the periphery to the center. Instead, they are the result of variation in the proportions of atrial and SA node cells from the periphery to the center (Verheijck et al. 1998). The present data are in accord with the gradient model rather than the mosaic model, in that they show that the density of i_Ca,L changes with C_m and in terms of cell size, and the densities of the L-type Ca²⁺ channel, Na⁺-Ca²⁺ exchanger, RYR2, and SERCA2 differing between peripheral and central SA node cells. Furthermore, the present data provide no evidence of two cell populations in the periphery or center of the SA node (with the exception of two patterns of RYR2 labeling in peripheral cells; Figure 6). We have recently shown that the gradient model, but not the mosaic model, can explain the electrical properties of SA node tissue (Zhang et al. 2001).