Contractile Behavior of Mouse Aorta Depends on SERCA2 Isoform Distribution: Effects of Replacing SERCA2a by SERCA2b

The Sarco(endo)plasmic reticulum Ca2+ ATPase (SERCA) actively pumps Ca2+ into the sarco/endoplasmic reticulum, thereby regulating intracellular Ca2+ concentrations and associated physiological processes. Different SERCA isoforms have been described (SERCA1, 2, and 3) with SERCA2 playing a pivotal role in Ca2+ homeostasis in cardiovascular tissues. In the heart, SERCA2a is the dominant isoform and has been proposed as therapeutic target in patients with heart failure. In the vasculature, both SERCA2a and SERCA2b are expressed with SERCA2b being the predominant isoform. The physiological role of SERCA2a in the vasculature, however, remains incompletely understood. In the present study, we used gene-modified mice in which the alternative splicing of the SERCA2-encoding gene (Atp2a2), underlying the expression of SERCA2a, is prevented and SERCA2a is replaced by SERCA2b. The resulting SERCA2b/b mice provide a unique opportunity to investigate the specific contribution of SERCA2a versus SERCA2b to vascular physiology. Aortic segments of SERCA2b/b (SERCA2a-deficient) and SERCA2a/b (control) mice were mounted in organ baths to evaluate vascular reactivity. SERCA2b/b aortic rings displayed higher contractions induced by phenylephrine (1 μM). Surprisingly, the initial inositol-3-phosphate mediated phasic contraction showed a faster decay of force in SERCA2b/b mice, while the subsequent tonic contraction was larger in SERCA2b/b segments. Moreover, in the presence of the calcium channel blocker diltiazem (35 μM) SERCA2b/b aortic rings showed higher contractions compared to SERCA2a/b, suggesting that SERCA2a (deficiency) modulates the activity of non-selective cation channels. Additionally, in endothelial cell (EC)-denuded aortic segments, the SERCA-inhibitor cyclopiazonic acid (CPA) caused markedly larger contractions in SERCA2b/b mice, while the increases of cytosolic Ca2+ were similar in both strains. Hence, aortas of SERCA2b/b mice appear to have a stronger coupling of intracellular Ca2+ to contraction, which may be in agreement with the reported difference in intracellular localization of SERCA2a versus SERCA2b. Finally, EC-mediated relaxation by acetylcholine and ATP was assessed. Concentration-response-curves for ATP showed a higher sensitivity of aortic segments of SERCA2b/b mice, while no difference in potency between strains were observed for acetylcholine. In summary, despite the relative low expression of SERCA2a in the murine aorta, our results point toward a distinct role in vascular physiology.

The Sarco(endo)plasmic reticulum Ca 2+ ATPase (SERCA) actively pumps Ca 2+ into the sarco/endoplasmic reticulum, thereby regulating intracellular Ca 2+ concentrations and associated physiological processes. Different SERCA isoforms have been described (SERCA1, 2, and 3) with SERCA2 playing a pivotal role in Ca 2+ homeostasis in cardiovascular tissues. In the heart, SERCA2a is the dominant isoform and has been proposed as therapeutic target in patients with heart failure. In the vasculature, both SERCA2a and SERCA2b are expressed with SERCA2b being the predominant isoform. The physiological role of SERCA2a in the vasculature, however, remains incompletely understood. In the present study, we used gene-modified mice in which the alternative splicing of the SERCA2-encoding gene (Atp2a2), underlying the expression of SERCA2a, is prevented and SERCA2a is replaced by SERCA2b. The resulting SERCA2 b/b mice provide a unique opportunity to investigate the specific contribution of SERCA2a versus SERCA2b to vascular physiology. Aortic segments of SERCA2 b/b (SERCA2a-deficient) and SERCA2 a/b (control) mice were mounted in organ baths to evaluate vascular reactivity. SERCA2 b/b aortic rings displayed higher contractions induced by phenylephrine (1 µM). Surprisingly, the initial inositol-3-phosphate mediated phasic contraction showed a faster decay of force in SERCA2 b/b mice, while the subsequent tonic contraction was larger in SERCA2 b/b segments. Moreover, in the presence of the calcium channel blocker diltiazem (35 µM) SERCA2 b/b aortic rings showed higher contractions compared to SERCA2 a/b , suggesting that SERCA2a (deficiency) modulates the activity of non-selective cation channels. Additionally, in endothelial cell (EC)-denuded aortic segments, the SERCA-inhibitor cyclopiazonic acid (CPA) caused markedly larger contractions in SERCA2 b/b mice, while the increases of cytosolic Ca 2+ were similar in both strains. Hence, aortas of SERCA2 b/b mice appear to have a stronger coupling of intracellular Ca 2+ to contraction, which may be in agreement with the reported difference in intracellular localization of SERCA2a versus SERCA2b. Finally, EC-mediated relaxation by acetylcholine and ATP was assessed. Concentrationresponse-curves for ATP showed a higher sensitivity of aortic segments of SERCA2 b/b INTRODUCTION Cytosolic calcium (Ca 2+ ) mobilization is a key event controlling multiple physiological processes in the blood vessel wall. Whereas in endothelial cells (EC) Ca 2+ activates NO-synthase (eNOS) to produce the relaxing factor nitric oxide (NO), Ca 2+ mobilization in smooth muscle cells (SMC) controls contraction and relaxation (Berridge, 2008). In both cell types, raised activator Ca 2+ has to be cleared from the cytoplasm upon subsequent inactivation. To ensure cytoplasmic Ca 2+ clearance and, hence, EC and SMC inactivation, the plasmalemmal Ca 2+ ATPases (PMCA), and the Na + /Ca 2+ exchanger cooperate with the mitochondria and the sarco/endoplasmic reticulum Ca 2+ ATPases (SERCA) (Berridge, 2008).
SERCA proteins are expressed mainly as three isoforms SERCA1, 2, and 3. SERCA2 is the most widely distributed isoform and plays a central role in Ca 2+ homeostasis in cardiovascular tissues. The ATP2A2 gene gives rises to different SERCA2 variants through alternative splicing. SERCA2a, often called the cardiac isoform, is expressed in cardiac muscle, slow-twitch skeletal muscle and SMC, while SERCA2b displays ubiquitous expression in muscle and non-muscle cells. SERCA2a and SERCA2b are produced by alternative splicing and only differ by replacement of the last 4 amino acids of SERCA2a by 49 additional amino acids in SERCA2b. However, both SERCA2 isoforms display different functionality with SERCA2b presenting a two times higher affinity for Ca 2+ , but a twofold lower catalytic turnover rate compared to SERCA2a (Lipskaia et al., 2013(Lipskaia et al., , 2014McIvor et al., 2018).
In the heart, SERCA2a is the prominent isoform and SERCA2a is down-regulated during pathological end stage heart failure. Down-regulation of SERCA2a is associated with impaired intracellular Ca 2+ handling in these pathologies, whereas restoration of SERCA2a expression by gene-transfer improved contractility in preclinical models of heart failure (Kho et al., 2012). Clinical trials in heart failure patients with SERCA2a adenoviral gene therapy initially showed promising results (CUPID) (Jessup et al., 2011), but follow-up trials (CUPID-2 and AGENT-HF) failed to demonstrate benefit of SERCA2a gene transfer, most likely due to inadequate gene transfer (Greenberg et al., 2016;Hulot et al., 2017).
In the vasculature, both SERCA2a and 2b are expressed. In vascular SMC from the rat aorta, SERCA2a and SERCA2b mRNA accounted for respectively 30 and 70% of total SERCA2 mRNA . Paradoxically, another study reported more abundance of the SERCA2a isoform in mouse aorta (Lipskaia et al., 2014). The specific roles of SERCA2a and 2b in vascular tissue, however, are not fully understood. There are indications that SERCA2b, among others, controls cell cycle and proliferation, whereas SERCA2a expression is more related to differentiation-specific contractile activity (Lipskaia et al., 2005(Lipskaia et al., , 2013. Some information on the physiological role of the different SERCA isoforms was obtained from gene ablation studies in mice. A homozygous knock-out of SERCA2 is embryonic lethal, whereas in heterozygous knock-out mice with only one functional SERCA2 allele, SMC contractile behavior was unchanged (Ishida and Paul, 2005). Additionally, a mouse model was developed in which the SERCA2-encoding gene (Atp2a2) was modulated such that the alternative splicing underlying the expression of SERCA2a is prevented. In these mice, the SERCA2a isoform is replaced by SERCA2b (referred to as SERCA2 b/b mice, as compared to wild-type SERCA2 a/b mice where both SERCA2 variants are normally expressed), which resulted in a clear cardiac structural and functional phenotype (Ver Heyen et al., 2001;Vangheluwe et al., 2006). Adult SERCA2 b/b animals developed a mild compensatory concentric cardiac hypertrophy with impaired cardiac contractility and relaxation, but preserved β-adrenergic response. Cardiac homogenates displayed reduced Ca 2+ uptake levels and in isolated cells, relaxation and Ca 2+ removal by the SR were significantly reduced. Because the SERCA2a to SERCA2b swap not only occurs in cardiac tissue, but probably also in other tissues, the present study investigated changes in vascular reactivity of aorta segments of SERCA2 a/b and SERCA2 b/b mice.

Mice
The studies were approved by the Ethical Committee of the University of Antwerp, and the investigations conform with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). Generation of the mouse strain is described in Ver Heyen et al. (2001). SERCA2 b/b mice and their corresponding wild-type strain SERCA2 a/b were used at the age of 4 to 5 months (SERCA2 a/b : n = 17, 13 females, 4 males, age 140 ± 6 days, body weight: 29.7 ± 1.1 g; SERCA2 b/b , n = 18, 13 females, 5 males, age 142 ± 5 days, body weight 28.1 ± 0.9 g).

Aortic Segments
After anesthesia (sodium pentobarbital, 75 mg kg −1 , i.p.), the thoracic aorta was carefully removed, stripped of adherent tissue and dissected systematically. Starting 3 mm from the origin of the left subclavian artery (where the azygos vein crosses the aorta) down to the diaphragm [11], the descending thoracic aorta was cut in segments of 1 mm (Ca 2+ -force studies) or 2 mm (vasomotor studies) width.
When SMC cytosolic Ca 2+ and force measurements were to be combined, ECs were removed by perfusion of the aorta with 3 ml 0.01% Triton X 100 in Krebs-Ringer solution, as described (Van Assche et al., 2007), or mechanically by rubbing the interior of the segment with a braided silk wax. Endotheliumdenuded segments were mounted in a wire (40 µm) myograph (Danish Myotechnology A/S, Aarhus, Denmark) above an inverted microscope, immersed in Krebs-Ringer solution (37 • C) and continuously aerated with 95% O 2 /5% CO 2 (pH 7.4). Only segments, in which endothelium-dependent relaxation by acetylcholine or ATP was absent, were considered as EC denuded. The segment was loaded in situ with 10 µM Fura-2 AM in Krebs-Ringer solution with 0.02% Pluronic for 120 min at room temperature. Then, the segment was set to its normalized diameter at 13.3 kPa according to Mulvany and Halpern (1977). The single emission (510 nm) ratio at dual excitation (340 and 380 nm) was used as a relative measure of free [Ca 2+ ] i (relative units, RU) and was analyzed with Felix software (PTI, United States). At the end of the experiment 2 mM MnCl 2 was added to the segment to determine background emission values, which were subtracted from the respective emission values during the experiment. Absolute levels of [Ca 2+ ] i are not reported due to the uncertainty of the conventional calibration method in intact tissues (Tosun et al., 1998;Lalli et al., 1999). Contractile force was measured simultaneously and reported in mN mm −1 .

Studies of EC Function
When only vasomotor responses were studied, aorta rings (2 mm) were mounted as described above, but the L-NAME was omitted and only indomethacin 10 µM was added to the Krebs-Ringer solution to avoid any interferences by prostanoids. When isometric force and EC Ca 2+ measurements were to be combined, the segments were mounted and loaded as described above, but without chemical or mechanical EC removal.

Western Blot
For each mouse genotype, five isolated aortic ring samples were combined. The frozen tissues were grinded with pestle and mortar in liquid nitrogen, then resuspended in 500 µl RIPA buffer (Pierce) supplemented with SIGMAFAST protease inhibitor (Merck). Further homogenization was done by douncing each sample 50 strokes on ice. The resultant suspension was rotated in head over head for 3 h at 4 • C, then centrifuged at 14000 rpm to remove the pellet. The protein concentration of the supernatant was determined by Bradford method.

Immunostaining
Aortic and left ventricular heart segments of SERCA2a/b and SERCA2b/b mice were collected immediately after dissection, embedded in NEG-50TM (Thermo Scientific) and kept at −80 • C. Cryosections of the aortic segments (± 6 µm) were air-dried for 20 min and, then, fixated with acetone (5 min). The cryostat sections were rinsed with phosphate buffered saline (PBS, 5 min) at room temperature. Following incubation overnight at 4 • C with the primary antibodies against SERCA2a (1:10000, R15-7, homemade), SERCA2b (1:10000, R5-24, homemade) and 3 times wash (10 min) with PBS, they were incubated for 2 h at 37 • C with secondary antibody (goat anti-rabbit Alexa green, Molecular Probes). The cryostat sections were placed again in washing buffer (15 min), washed 3 times (5 min) in PBS and mounted in VECTASHIELD Antifade Mounting Medium with DAPI (nuclear staining). Images were made on an Olympus BX40 epifluorescence microscope with a SensiCam 12 bit cooled CCD (PCO, Germany) and stored on computer for later analysis with Photoshop software.

Data Analysis
All results are expressed as mean ± SEM; n represents the number of mice, s the number of segments. Dose-response curves were fitted with sigmoidal dose-response equations with variable slope, which revealed maximum contraction or relaxation (E max ) and the logarithm of the concentration resulting in 50% of the maximal response (log EC 50 ) for each vessel segment. Twoway ANOVA with Bonferroni post-test and unpaired t-test were performed using GraphPad Prism, version 6 (GraphPad Software, San Diego, CA, United States) and were used to compare means of the different experimental groups. A 5% level of significance was selected.

SERCA2a/SERCA2b mRNA Transcripts and Protein Expression
Real-time RT-PCR on RNA isolated from aortic segments of SERCA2 a/b and SERCA2 b/b mice ( Figure 1A) revealed the presence of SERCA2a and SERCA2b mRNA transcripts in the aorta of SERCA2 a/b mice with about 20-fold higher SERCA2b levels compared to SERCA2a (assuming an amplification efficiency close to 100%). SERCA2a mRNA was undetectable in SERCA2 b/b mice, confirming successful prevention of alternative splicing of the modified SERCA2 allele to SERCA2a. SERCA2b transcripts were significantly lower in SERCA2 b/b than in SERCA2 a/b mouse aorta. Immunohistochemistry (IHC) using antibodies against SERCA2a and SERCA2b protein isoforms confirmed the mRNA data and revealed that in SERCA2 a/b mouse aorta the SERCA2a protein levels were much lower in aorta than in cardiac tissue ( Figure 1B). SERCA2a staining was clearly present in hearts of SERCA2 a/b mice, but was absent from hearts of SERCA2 b/b animals. SERCA2b staining was present in ECs and SMCs, and was similarly distributed in SERCA2 a/b and SERCA2 b/b animals. Western blot analysis (Figures 1C,D) confirmed the data of the RT-PCR and IHC in that the SERCA2a protein isoform was present in SERCA2 a/b mice, but absent from SERCA2 b/b animals. In line with previous observations in the heart of these mice (Ver Heyen et al., 2001), SERCA2b displayed about 1.6 times higher expression in SERCA2 a/b than in SERCA2 b/b aortic segments.

Contractile Responses
Contractions were elicited by means of two stimuli, which involve different mobilization of intra-and extracellular Ca 2+ . Depolarization with high external K + causes isometric contraction of aortic segments by opening voltage-gated Ca 2+ channels only . α 1 -Adrenoceptor stimulation with phenylephrine (PE), on the other hand, causes phasic contractions by release of intracellular Ca 2+ from the SR Ca 2+ stores and tonic contractions through Ca 2+ influx via voltage-gated Ca 2+ channels and non-selective cation channels . By comparing these contractile stimuli in SERCA2 a/b and SERCA2 b/b mice, the functional role of SERCA2a/b in aortic vascular smooth muscle could be explored.

Depolarization
Basal SMC Ca 2+ levels of 1.04 ± 0.04 RU for SERCA2 a/b and 1.02 ± 0.05 RU for SERCA2 a/b (n = 14) were not different. Depolarization of the aortic segments with stepwise elevation of K + concentrations from 5.9 to 10, 20, 30, 40, and 50 mM K + elicited parallel Ca 2+ signals and contractions, which were not significantly different between the strains with or without supplementation of L-NAME, an inhibitor of basal NO release (Figure 2). α 1 -Adrenergic Stimulation of Aortic Segments With 1 µM Phenylephrine In endothelium-intact segments, phenylephrine (PE, 1 nM-30 µM) caused significantly higher contractions in SERCA2 b/b than in SERCA2 a/b mice with a near-significant shift of the concentration response curve to the left (Figures 3A,B). After addition of 300 µM L-NAME to inhibit basal NO release, isometric force increased more in the SERCA2 a/b than in the SERCA2 b/b segments with significant shifts in PEsensitivity for both strains (Figure 3B). At steady-state the isometric contractions were not significantly different between the strains. When these experiments were repeated with a single concentration of PE (1 µM) in the absence (Figure 3C) or presence ( Figure 3D) of 1 µM levcromakalim, contractions were larger in SERCA2 b/b than of SERCA2 a/b segments both in the absence and presence of L-NAME ( Figure 3C). In the presence of levcromakalim, which prevents the activation of L-type Ca 2+ influx by hyperpolarization of the aortic segments, contraction of vascular SMC is mediated by non-selective cation influx only , suggesting increased Ca 2+ influx in SERCA2 b/b aortic segments stimulated with PE.

Phasic Contractions by PE in the Presence of 0 µM Ca 2+
Addition of 1 µM PE in the absence of external Ca 2+ caused phasic contractions which decreased with duration of the challenge with 0 mM Ca 2+ , suggesting spontaneous emptying of the contractile SR Ca 2+ stores in the absence of extracellular Ca 2+   (Figure 4). With each duration of the 0 mM Ca 2+ period (3, 5, 10, or 15 min) the AUC of the phasic contractions was consistently lower in SERCA2 b/b segments ( Figure 4B) and in these segments spontaneous emptying of the contractile SR Ca 2+ stores occurred slower (7.4 ± 1.9 versus 2.7 ± 0.6 min, p = 0.07, n = 5-4, Figure 4C) and was more complete. Subsequent phasic contractions by PE in other experiments were measured after a period of 3 min challenge with 0 mM Ca 2+ . Interestingly, and as shown in Figure 4A, at each interval after addition of 0 mM Ca 2+ , PE-induced phasic contractions showed a faster decay in SERCA2 b/b compared to SERCA2 a/b segments and at 3 min the time constants were 17 ± 1 s versus 39 ± 4 s, p < 0.001 (n = 5). The time constants of force development were not different (respectively 2.3 ± 0.1 and 2.1 ± 0.1 s).

Tonic Contractions by PE Upon Re-addition of 3.5 mM Ca 2+
Re-addition of external Ca 2+ after the phasic contraction induced a tonic contraction (Figure 5), which was slightly higher in SERCA2 b/b aortic segments. Following inhibition of L-type Ca 2+ channels with 35 µM diltiazem, the remaining contraction, which is now due to Ca 2+ influx via non-selective cation channels (NSCC) was significantly higher in SERCA2 b/b segments (Figures 5B,C).
Refilling of SR Ca 2+ Stores After α 1 -Adrenergic Stimulation With 1 µM PE During the Ca 2+ re-addition phase, SR Ca 2+ stores, which were emptied by admission of 1 µM PE, are re-filled, mainly by activity of L-type Ca 2+ channels . By measuring FIGURE 2 | Intracellular Ca 2+ signals and concomitant force development induced by stepwise increase of K + in aortic segments of SERCA2 a/b and SERCA2 b/b mice. Sensitivity of SERCA2 a/b (circles) and SERCA2 b/b (squares) aortic segments to depolarization with elevated external K + . (A) Ca 2+ signals and concomitant force development with increasing concentrations of extracellular K + , measured in Fura-2-loaded endothelium-denuded aortic segments mounted in a myograph (n = 4). Force at increasing K + for endothelium-intact aortic segments measured in an isometric organ bath in control (B) and after inhibition of eNOS with 300 µM L-NAME (C) (B,C: n = 5).
FIGURE 3 | Isometric contractions by PE are higher in aortic segments of SERCA2 b/b compared to SERCA2 a/b . Concentration-effect curves (A) and respective logEC 50 values (B) for contractions by PE (10 -9 to 3*10 -5 M) in the absence (open symbols) and presence (closed symbols) of 300 µM L-NAME in SERCA2a (circles) and SERCA2b (squares) aortic segments. Graphs show the contraction induced by 1 µM PE in the absence (C) and presence (D) of 1 µM levcromakalim. 10 Min after the addition of PE, 300 µM L-NAME was added in all conditions to inhibit release of basal NO. *, **p < 0.05, 0.01 SERCA2 b/b versus SERCA2 a/b (n = 5); #, ###p < 0.05, 0.001 + L-NAME versus -L-NAME.  Frontiers in Physiology | www.frontiersin.org the second phasic contraction during the re-filling of the stores, the amount of re-filling can be estimated. Figure 6A shows the experimental protocol, whereas Figure 6B shows the first and second PE-induced phasic contraction. In SERCA2 a/b aortic segments the second PE application caused a phasic contraction with an AUC which was about 20% of the first PE response ( Figure 6C). In SERCA2 b/b segments the contractile SR store was re-filled for only about 10%. When 35 µM diltiazem was added after PE 1 , phasic contractions were significantly reduced in SERCA2 a/b aortic segments (Figure 6D), but not in SERCA2 b/b segments (Figures 6E,F).

Contractions by SERCA Inhibition With CPA
The SERCA inhibitor CPA (10 µM) causes large intracellular Ca 2+ transients in SMC without producing substantial contractions . In the following experiments SERCA was inhibited in EC-denuded aortic segments of SERCA2 a/b and SERCA2 b/b mice to investigate intracellular Ca 2+ and concomitant isometric contractions. Addition of 10 µM CPA caused a large increase of internal Ca 2+ (0.27 ± 0.05 RU) in SERCA2 a/b aortic segments, which was paralleled by a small and transient increase of isometric force (E max 0.70 ± 0.23 mN/mm) (Figures 7A,B). In SERCA2 b/b segments, the Ca 2+ signal was not different (0.27 ± 0.03 RU), but was accompanied by a substantially higher isometric force development (3.60 ± 0.43 mN/mm) (Figures 7A,B). These data were confirmed in aortic segments, in which endothelial NO release was inhibited with 300 µM L-NAME ( Figure 7C). SERCA inhibition with 10 µM CPA caused significantly higher contractions in SERCA2 b/b than in SERCA2 a/b segments (Figures 7C,D). After inhibition of L-type Ca 2+ channels with 35 µM diltiazem, the contraction by CPA was inhibited but the difference between the two strains remained.

Relaxation of PE-pre-constricted Segments by NO and cGMP
To assess endothelium-dependent and endotheliumindependent effects of NO or cGMP in vascular SMCs, PE-constricted segments were exposed to sources of endogenous NO (acetylcholine, ATP, CPA), exogenous NO (NO − 2 , DEANO) and the membrane permeable cyclic GMP-analog 8-Br-cGMP (Table 1). Except for ATP, there was no difference between endothelium-dependent or endothelium-independent relaxation of pre-constricted aortic segments of SERCA2 a/b and SERCA2 b/b mice. Figure 8 shows representative examples of simultaneous endothelial Ca 2+ signals and isometric contraction/relaxation of aortic segments with intact endothelium. 1 µM PE caused a very small increase of intracellular Ca 2+ , which is probably due to the Ca 2+ increase in SMC and which suggests rather selective loading of EC with Fura 2-AM. CPA (0.3 µM and higher) caused a concentration-dependent increase of endothelial intracellular Ca 2+ , which was accompanied by relaxation of PE-pre-contracted segments in both strains. Only the highest concentration of 10 µM CPA caused contraction due to increase of Ca 2+ in the SMC. Also in this experimental setting, CPAinduced Ca 2+ (EC 50 −5.44 ± 0.17 logM to −5.75 ± 0.10 logM) and relaxation (EC 50 −6.32 ± 0.15 logM to −6.27 ± 0.26 logM) were unaltered in SERCA2 b/b segments.

DISCUSSION
According to several reports, vascular EC express two of the three SERCA genes, i.e., SERCA2 (SERCA2b) and SERCA3 (SERCA3a) (Mountian et al., 1999;Khan et al., 2000), whereas vascular SMC mainly express the spliced isoforms SERCA2a and 2b Clark et al., 2010;Lipskaia et al., 2014). In cardiac muscle, SERCA2a is the prominent SERCA isoform (95%). In SMC of pig aorta, 72% of the SERCA2 mRNAs encode the SERCA2b isoform (Eggermont et al., 1990). Also in rat aorta, SERCA2b is the more abundant SERCA isoform . Similarly, the real-time PCR data of the current study points toward a significantly higher mRNA expression of the SERCA2b isoform in the aorta of SERCA2 a/b mice. The mRNA data were confirmed by IHC and Western blot. SERCA2b was abundantly present in the aorta, while SERCA2a was hardly detected by IHC or Western blot as compared to cardiac tissue. Conversely, Lipskaia et al. (2014) reported a high mRNA expression of SERCA2a (91% of total SERCA) and low expression of SERCA2b (∼5%) in the aorta of C57Bl/6 mice . Possible explanations for this discrepancy are genetic background (C57Bl/6 versus 50/50 R1 129/Swiss), age of the mice (8 weeks versus 4/5 months), aortic segment origin and RNA quantification methods.
Here, we investigated vascular reactivity in the SERCA2 b/b mouse model. With this model, it has been shown that SERCA2a is essential for normal cardiac performance (Ver Heyen et al., 2001). SERCA2a-isoform deficiency leads to increased incidence of fetal/neonatal mortality and cardiac structural malformations. Hearts of SERCA2 b/b mice develop hypertrophy and display decreased total SERCA2 mRNA and protein levels (50%). Concomitantly, impaired cardiac contractility and relaxation occurred with reduced SR Ca 2+ -uptake activity (−50%). The apparent Ca 2+ affinity of the Ca 2+ uptake was increased in SERCA2 b/b , suggesting a higher apparent Ca 2+ affinity for SERCA2b in cardiac homogenates of SERCA2 b/b mice. A prominent role of SERCA2a in vascular tissue was not evident. In a heterozygous SERCA2 knock-out mouse (SERCA2 ± , 50% reduction of SERCA2), aortic contractility to PE or depolarization (KCl) was not changed, and there was no difference in relaxation to activation of the A-kinase pathway (forskolin) or G-kinase pathway (sodium nitroprusside) (Ishida and Paul, 2005). In the present study, however, despite the small amount of SERCA2a versus SERCA2b in the mouse aorta, functional effects of SERCA2a ablation were considerable. Importantly, also a decrease in SERCA2b mRNA and protein (about 60%) expression in the aorta of SERCA2 b/b mice was observed, which may, in part, contribute to the observed vascular phenotype.
Depolarization of aortic segments with increased extracellular K + induced similar Ca 2+ signals and isometric force in SERCA2 b/b and in SERCA2 a/b animals. We have previously shown that isometric contractions of C57Bl6 aortic segments FIGURE 6 | Amount of re-filling of the PE-sensitive contractile SR Ca 2+ stores in SERCA2 a/b and SERCA2 b/b segments. (A) Experimental protocol to investigate PE (1 µM)-sensitive phasic contractions in 0 mM Ca 2+ after a period of 5 min of Ca 2+ re-addition (+ Ca 2+ ) to refill the SR Ca 2+ stores. (B) Phasic contractions of the first (PE 1 ) and second (PE 2 ) phasic contraction by 1 µM PE in 0 mM Ca 2+ in SERCA2 a/b (black) and SERCA2 b/b (blue) aortic segments. (C) Ratio of AUC of the first and second PE application (PE1/PE2 in%) in both strains. When after PE1 35 µM diltiazem was added, refilling was attenuated in SERCA2 a/b segments (D), but was less affected in SERCA2 b/b segments (E). (F) Ratio of AUC of the first and second PE application (PE1/PE2 in%) in control (open bars) and after addition of 35 µM diltiazem (patterned bars) in both strains. *, **, ***p < 0.05, 0.01, 0.001 SERCA2 b/b versus SERCA2 a/b (n = 5), ##p < 0.01 diltiazem versus control (n = 5).
upon depolarization are completely due to fast and slow Ca 2+ influx via voltage-gated Ca 2+ channels (VGCC) . Hence, the results of the present study suggest that the contractile SR (SERCA2a and SERCA2b) does not contribute to depolarization-induced contractions in mouse aorta. Hence, in SERCA2 b/b mice, there seems to be no functional change of the L-type Ca 2+ channel mediated contractions as long as the SR is not involved in the contraction.
However, prominent differences in contractile behavior were observed with contractions elicited by α 1 adrenergic stimulation with PE. These contractions occur in two phases. First, PE induces a transient, phasic contraction, which is due to Ca 2+ release from IP 3 -sensitive SR Ca 2+ stores. This phasic contraction depends on the SR Ca 2+ store content and the ratio of Ca 2+ release from the stores and removal of Ca 2+ from the cytoplasm via plasmalemmal Ca 2+ pumps (PMCA) . The initial phasic contraction is followed by a persistent, tonic contraction, which is due to Ca 2+ influx via VGCC and via NSCC . Interestingly, when the VGCC Ca 2+ influx of the PE contraction was prevented by hyperpolarizing vascular SMCs with levcromakalim, which is an ATP-dependent K + channel agonist, the difference in contraction between SERCA2 a/b and SERCA2 b/b mice remained. This observation suggests an altered NSCC Ca 2+ influx.
Interestingly, the amplitude of the phasic contraction (performed in 0 mM Ca 2+ conditions) was not different between SERCA2 a/b and SERCA2 b/b segments. However, the area under the curve (AUC) was significantly lower in SERCA2 b/b segments. This was mainly due to the faster decay of force in SERCA2 b/b segments, which is not due to re-uptake of contractile Ca 2+ to the SR via SERCA, but rather to the extrusion of intracellular Ca 2+ via the plasmalemmal Ca 2+ pump (PMCA) . Furthermore, the spontaneous leakage of contractile Ca 2+ from the SR assessed by PE phasic contractions upon various duration of 0 mM Ca 2+ challenges consistently showed that the AUC of the phasic contraction was smaller in SERCA2 b/b mice suggestive for a lower contractile Ca 2+ content of the SR stores. Additionally, segments of SERCA2 b/b mice spontaneously lose their contractile Ca 2+ slower in function of the duration of the 0 mM Ca 2+ challenge, suggesting an altered ratio of Ca 2+ release and removal (Michiels et al., 2015).
Conversely, the tonic contraction induced by the re-addition of external Ca 2+ tended to be larger in SERCA2 b/b segments, especially after addition of diltiazem that blocks VGCC. This again suggests that the difference between both strains relates to NSCC Ca 2+ influx upon store emptying. When Ca 2+ was added again to the organ bath, the SR Ca 2+ stores refilled mainly through Ca 2+ influx via VGCC and subsequent re-uptake of Ca 2+ in the SR Ca 2+ stores via SERCA . The FIGURE 7 | SERCA Ca 2+ re-uptake inhibition with CPA caused similar intracellular Ca 2+ signals, but increased isometric force signals in SERCA2 b/b more than SERCA2 a/b aortic segments. (A) Intracellular Ca 2+ and (B) isometric force were simultaneously measured in endothelium-denuded (1 mm width) aortic segments (n = 5) of SERCA2 a/b and SERCA2 b/b mice when treated with 10 µM CPA as indicated by the horizontal bar. *, ***p < 0.05, 0.001 SERCA2 b/b versus SERCA2 a/b . (C,D) Tonic contractions by 10 µM CPA were measured in (2 mm width) aorta segments treated with 300 µM L-NAME to inhibit eNOS production. CPA-induced contractions were increased in SERCA2 b/b also after inhibition of L-type Ca 2+ influx with diltiazem. CPA was added during a 0 mM Ca 2+ challenge, while re-addition of 3.5 mM Ca 2+ (+ CPA solution) resulted in a contractile response, both in the absence and presence of 35 µM diltiazem. *, **, ***p < 0.05, 0.01, 0.001 SERCA2 b/b versus SERCA2 a/b , #, ##p < 0.05, 0.01 diltiazem versus control (n = 5).
refilling of the SR store was less in aortic segments of SERCA2 b/b mice, which is in line with the reported twofold lower catalytic turnover rate of SERCA2b compared to SERCA2a and the lower expression of total SERCA protein in SERCA2 b/b aortic segments (Ver Heyen et al., 2001;McIvor et al., 2018). As mentioned before, L-type (VGCC) Ca 2+ channels play a prominent role in refilling of the SR Ca 2+ stores , hence, when 35 µM diltiazem was added during the refilling period, phasic contractions by PE were significantly reduced. Moreover, this reduction was most pronounced in SERCA2 a/b segments, thereby confirming the putative role of VGCC in SR refilling.
We previously reported that the SERCA inhibitor CPA (10 µM) causes large intracellular Ca 2+ transients in SMC (i.e., in EC-denuded segments or segments treated with the eNOS inhibitor L-NAME), which were accompanied by relative minor contractions . In the current study, CPA caused similar effects in aortic segments of SERCA2 a/b mice. In SERCA2 b/b mice, however, CPA caused significantly larger contraction while the intracellular Ca 2+ response was of similar magnitude as in SERCA2 a/b mice. In other words, in the aorta of SERCA2 b/b mice there is a more efficient coupling of CPA-elicited intracellular Ca 2+ to contraction, which persisted after inhibition of L-type Ca 2+ channels with 35 µM diltiazem. This observation suggests that intracellular Ca 2+ responses induced by CPA are differently coupled to contraction in SERCA2 a/b and SERCA2 b/b aortic segments respectively. Functionally segregated SR Ca 2+ stores have been described in pulmonary arterial SMC (Clark et al., 2010). In these cells SERCA2b was located predominantly proximal to the plasma membrane while SERCA2a was located centrally. Remarkably, in these cells, the proximal SERCA (SERCA2b) was sensitive to the SERCA inhibitor CPA, whereas the central SERCA (SERCA2a) was CPA-insensitive, but thapsigargin-sensitive. The different localization of both SERCA2a/b Ca 2+ stores may explain the difference in contractile response between both mouse strains. Moreover, we have previously reported that inhibition of SR Ca 2+ stores with 10 µM CPA turns a PE-mediated contraction from mainly VGCC-mediated to mainly NSCC-mediated . Our data support the observation that inhibition of SERCA2b with CPA results in stronger coupling of NSCC Ca 2+ influx to contractile responses.
Relaxation of PE-induced contractions by increasing intracellular Ca 2+ in the EC as with CPA and ACh, were similar in SERCA2 a/b and SERCA2 b/b mice, suggesting equal Concentration-relaxation curves for endogenous NO release with ACh, ATP and CPA, exogenous NO supply with NO − 2 or DEANO and the membrane permeable cGMP analog, 8-Br-cGMP, revealed log(EC 50 ) and maximal relaxation (E max ) values as indicated in the table. For relaxation byNO − 2 , DEANO and 8-Br-cGMP segments were treated with 300 µM L-NAME to inhibit basal endothelial release of NO. * p < 0.05 SERCA2 b/b versus SERCA2 a/b . activation of eNOS and comparable NO release in both mouse strains. On the other hand, results with ATP that mainly activates purinergic (P2Y) receptors (Guns et al., 2005) suggest that in some conditions EC from SERCA2 b/b produce higher amounts of NO or that the SMC of the SERCA2 b/b mice, which only possess SERCA2b, are more sensitive to NO. Although EC 50 values of the exogenous NO donors, DEANO and nitrite, were not significantly different between both mouse strains, there was a tendency for a higher sensitivity of the SERCA2 b/b SMC to NO. Since cGMP relaxation is the same in SERCA2 a/b and SERCA2 b/b mice, this indicates that the direct effect of NO on Ca 2+ re-uptake (Cohen et al., 1999;Van Hove et al., 2009) via SERCA2a and SERCA2b SR Ca 2+ stores is different.
Finally, basal NO release, i.e., the difference between the amplitude of PE contractions in the absence and presence of the eNOS blocker L-NAME , was attenuated in SERCA2 b/b mice. Moreover, the fact that in the presence of levcromakalim, basal NO did not differ between SERCA2 a/b and SERCA2 b/b mice suggests that the difference in basal NO results in a differential regulation of the VGCC Ca 2+ influx (Van Hove et al., 2009). Previously, studies with SERCA3 gene-ablated mouse have demonstrated that SERCA3 plays a critical role in EC Ca 2+ signaling involved in NO release and NO-mediated relaxation of vascular SMC, but not in SMC, consistent with its known distribution (Liu et al., 1997;Paul, 1998). The lack of compensation by SERCA2b for the SERCA3 dysfunction further suggests that they are involved in restoring original Ca 2+ to specific subcellular compartments (Hadri et al., 2006) and that they serve functionally distinct Ca 2+ pools in EC. In salivary epithelial cells, SERCA3 and SERCA2b have indeed separate intracellular locations (Lee et al., 1997). Furthermore the apparent K m of SERCA3 for Ca 2+ is much higher than that of SERCA2b, indicating that SERCA3 might function at higher Ca 2+ concentrations than SERCA2b (Lytton et al., 1992;Arredouani et al., 2002;Hadri et al., 2006).

CONCLUSION
Despite the low incidence of SERCA2a in wild type (SERCA2 a/b ) mice, prevention of the alternative splicing underlying the expression of SERCA2a, resulted in a distinct cardiovascular functional phenotype, i.e., higher PE-induced contractions and increased NSCC Ca 2+ influx, slower re-filling of the SR Ca 2+ stores and stronger coupling of NSCC Ca 2+ influx with contraction, lower basal NO efficacy and slightly increased EC-dependent relaxation. Whether the altered vascular reactivity directly relates to SERCA2a depletion or to indirect compensatory mechanisms will require further investigation. Although in a heterozygous SERCA2 knock-out mouse (SERCA2 ± , 50% reduction of SERCA2), aortic contractility to PE or depolarization (KCl) was not changed (Ishida and Paul, 2005), it was observed that the combination of SERCA2a absence and SERCA2b bisection in the SERCA2b/b mouse model caused an altered vascular (aortic) functional phenotype. At least, the attenuated SERCA2b protein expression in the SERCA2 b/b mouse model as compared to "wild-type" SERCA2 a/b mice is not sufficient to explain the altered cardiovascular functional phenotype. Overall, given the promise of SERCA2a as therapeutic target in cardiac disease, it is important to carefully consider its role in the vascular system.

DATA AVAILABILITY STATEMENT
The datasets generated for this study are available on request to the corresponding author.

ETHICS STATEMENT
The animal study was reviewed and approved by the Ethical Committee for Animal Testing (ECD), University of Antwerp.
Written informed consent was obtained from the owners for the participation of their animals in this study.

AUTHOR CONTRIBUTIONS
PF performed the vascular reactivity experiments, P-JG performed the PCR experiments, JC contributed to the Western blots, and PV provided the mouse model and antibodies for IHC and Western blot. P-JG and PF have jointly written the manuscript, whereas JC and PV contributed to the revision of the text. All authors have read and approved the final version of the manuscript.