Sarcolipin in atrium-specific gene targeting

Department of Life Science and Medical Bioscience, Waseda University, TWIns, Tokyo, 162-8480 Japan Department of Molecular Cell and Developmental Biology, University of California Los Angeles, CA, 90095 USA Eli and Edythe Broad Center of Regenerative Medicine and Stem Cell Research, University of California Los Angeles, CA, 90095-7357 USA Jonsson Comprehensive Cancer Center, University of California Los Angeles, CA, 90024 USA Molecular Biology Institute, University of California Los Angeles, CA, 90095-1570 USA Department of Cell Physiology, The Jikei University School of Medicine, Tokyo, 105-8461 Japan


Introduction
Spatiotemporal regulation of gene expression is critical for proper heart development.The analysis of a specific gene's function in tissues is essential to understanding the role of a gene of interest.The Cre/loxP site-specific recombination system from the bacteriophage P1 enables the generation of tissue-specific gene-targeting mice by the recombination of the loxP-flanked gene [1,2] .In the field of cardiovascular research, a number of cardiac cell-type-specific Cre mice lines have been generated and have contributed to the analysis of the functions of genes of interest in the heart [3] .The α-myosin heavy chain (α-MHC) promoter Cre mouse line is the most popular Cre mouse line used to generate REVIEW cardiomyocyte-specific gene-targeting mice [3] .The atrium and the ventricle of the heart have chamber-specific functions, as well as distinct structure, gene expression, and even metabolic profiles [4][5][6][7] .Therefore, to understand the role of the heart, it is important to distinguish between the atrium and the ventricle and to analyze the characteristics of each chamber individually.To date, previous studies have focused on the entire heart and/or the ventricle that acts as a pump using a ventricle-specific Cre mouse line [8] .The characteristics of the atrium remain poorly understood.Many questions, especially those regarding the mechanisms of atrial development and the causes of congenital heart diseases such as atrial septum defects, remain unanswered.
Since sarcolipin (SLN) is specifically expressed in the atrial muscle of the heart, we generated a SLN-Cre knock-in (KI) mouse line (Sln Cre/+ mouse) by inserting Cre recombinase gene into the endogenous SLN locus [9,10] .Cre recombinase is restrictively expressed in the atrial muscle of Sln Cre/+ mice.In this review, we summarize the role of SLN in the atrium based on previous studies and describe the advantages and future outlook of Sln Cre/+ mice as atrium-specific gene-targeting mice.

Characteristics of sarcolipin and interaction with SERCA
SLN is a sarco(endo)plasmic reticulum (SR) membrane protein that is composed of 31 amino acids [11] .The SLN gene is conserved in vertebrates [12] and, especially in mammals, expressed in the atrial muscle and the skeletal muscle [9] .In small animals such as rodents, SLN is predominantly expressed in the atrium and is expressed to a lesser extent in the skeletal muscle; the expression is the reverse in large animals such as rabbits and humans [9,13] .Although the SLN expression pattern differs between small and large animals, it is of note that SLN is undetectable in the ventricle of the mammalian heart [9,13] .Accordingly, SLN has been used as a marker of human-induced pluripotent stem cell-derived atrium-like cardiomyocytes [14] .
SLN has an α-helix single transmembrane domain, and phospholamban (PLN) is also known as an SR membrane protein and a homologue of SLN [15,16] .In contrast to SLN, PLN is expressed in both the atrium and the ventricle, but the expression is predominant in the ventricle [9,15] .SR Ca 2+ release and uptake are important in muscle contraction and relaxation.After muscle contraction, most Ca 2+ in cytosol is taken to the SR lumen through SERCA [17] .Both SLN and PLN interact with SERCA and inhibit the SR Ca 2+ uptake function [15,18,19] .Although some of the binding sites in SERCA are shared between SLN and PLN, the C-terminus of SLN has a unique amino acid sequence that it is conserved among mammals and is considered to be important for the strong interaction with the aromatic residue in SERCA [16,[20][21][22][23][24] .Although most SLN loses its binding ability to SERCA under a high cytosolic Ca 2+ level, Sahoo et al. reported that although part of SLN can bind to SERCA in any kinetic state (even under a high Ca 2+ level), PLN does not [23] .Therefore, the ATPase activity of SERCA is not altered by the existence of SLN, and this promotes the uncoupling of SERCA [23] .Although the uncoupling of SERCA is considered to be involved in muscle thermogenesis (described below), it is unclear if this effect occurs in the cardiomyocytes that continuously contract.Moreover, the function of PLN is inhibited by the phosphorylation of Ser-16 and Thr-17 through the β-adrenergic pathway [25,26] .In SLN, no definitive phosphorylation site has been identified, although Thr-5 is a candidate through the β-adrenergic pathway to inhibit the SLN function [27] .

Regulation of SLN expression under a physiological or pathological condition
In mouse embryos, SLN starts being expressed around embryonic day (ED) 10.5 and is fully expressed on ED 12.5 in the atrium [9,10] .After ED 16.5, SLN expression is increased and maintained through the life span [9] .The transcription of SLN is regulated by thyroid hormone and pathological conditions.In mouse hearts with hyperthyroidism, SLN mRNA expression is decreased by post-transcriptional downregulation whereas hypothyroidism does not affect the SLN mRNA expression level [28] .In addition, effects similar to those of the thyroid hormone on SLN expression were observed in neonatal rat atrial myocytes [28] .In the atrium of the hypertrophic heart induced by overexpressing activated H-ras gene in SERCA2 b/b mice, SLN expression is decreased, but not induced in the ventricle [9, 13]   .We previously reported that mechanical stress downregulates SLN expression.Mice with transverse aortic constriction (TAC; left-side pressure overload) showed a decrease in SLN and SERCA expression levels in the left atrium [29] .In addition, monocrotaline-induced right-side pressure-overloaded rat heart exhibited a decrease in these expression levels in the right atrium [29] .SLN expression is decreased in the atrium of patients with chronic atrial fibrillation (AF) and in ischemic dog hearts [30,31] .Taken together, these findings indicate that SLN expression is mostly downregulated in the atrium under pathological conditions and is not induced in the ventricle.There are several exceptions, however; Zheng et al. reported that SLN expression is increased in the left ventricle of the human heart with mitral regurgitation [32] .It was also reported that the ventricles of dystrophin-deficient mice and ventricle-specific Nkx2.5 null mice express SLN [33,34] .Although the direct regulators of SLN transcription remain unknown, Nkx2.5 is considered to suppress SLN expression in mouse ventricle.Similarly, mice with the cardiomyocyte-specific deletion of Hrt2, which is known to be expressed in the ventricle and not in the atrium, showed ectopic expression of atrial genes, including SLN, in the ventricle whereas cardiomyocyte-specific COUP-TFII deletion resulted in ventricularized atria [35,36] .

Genetic modulation of SLN expression (overexpression and knock-out study)
Several studies and our unpublished data have demonstrated that overexpression of SLN decreased the Ca 2+ affinity of SERCA, Ca 2+ transient, muscle peak tension, and rate of contraction and relaxation [37][38][39] .In contrast to PLN, SLN overexpression decreases the V max of SERCA, even at high Ca 2+ concentration.Moreover, although SLN overexpression reduces the fractional shortening and induces mild hypertrophy, it does not affect the heart rate [37,39] .These SLN inhibitory effects on the Ca 2+ uptake function of SERCA are reproducible in HEK-293 cells co-expressing SLN and SERCA [23] .HEK-293 cells co-expressing SLN, PLN, and SERCA showed super-inhibition of the SERCA function [19] .Previous studies have reported that SLN shares the binding site in SERCA with PLN, and forms the ternary SLN/PLN/SERCA complex, consequently superinhibiting the SERCA function [19,20,40] .It remains unclear if they form the ternary complex in vivo.
Babu et al. generated an SLN knock-out (KO) mouse line by the homologous recombination of a neomycin-resistant gene to the endogenous SLN locus [41] .SLN KO mice show no abnormalities in heart morphology and have a normal life span.In terms of Ca 2+ homeostasis, the rate of Ca 2+ uptake and Ca 2+ transient is increased in SLN null hearts and there are no changes in the expression level of Ca 2+ handling proteins.Moreover, the isomeric force generation and maximum rate of contraction are higher in SLN KO atria whereas the maximum rate of relaxation and systolic and end-diastolic pressure in the hearts are not significantly altered [41] .Shanmugam et al. reported that double KO mice for SLN and PLN activated the SERCA function markedly and resulted in cardiac hypertrophy [42] .These results suggest that SLN and/or PLN are necessary for maintaining normal Ca 2+ homeostasis in mouse atria.In addition, Xie et al. recently reported an increased amount of interstitial fibrosis in SLN null mouse atria and the susceptibility to atrial arrhythmias in aged SLN null mice [43] .Although the precise mechanisms of fibrosis and arrhythmias in SLN null mice remain unclear, they found change in the gene expression related to fibrosis.In aged SLN null mice, prolongation of the action potential duration with increased I NCX and increased ryanodine receptor (RyR) phosphorylation were also found [43] .

SLN-Cre knock-in mouse (heterozygous deletion of SLN)
SLN overexpression and knock-out studies have been well documented and previously reviewed [44] .Taking advantage of the atrium-specific expression of SLN, we generated an SLN-Cre KI (Sln Cre/+ ) mouse line by inserting the Cre recombinase gene into the endogenous SLN exon2 locus (Figure 1) [10] .Sln Cre/+ mice enable us to delete a gene of Genomic southern blot analysis and PCR for mouse genotyping (B and C).X-gal staining for in vivo tracing Cre expression using SlnCre/+;R26R (D).DTA, diphtheria toxin A cassette.Reprinted with permission [10].
interest in an atrium-specific manner using the Cre/loxP site-specific recombinase system.Prior to generating an atrium-specific conditional KO mouse using the Sln Cre/+ mouse line, the effects of heterozygous SLN deletion on the atrium had not yet been evaluated, because one copy of the endogenous SLN gene is deleted in Sln Cre/+ mice in response to Cre recombinase insertion.Thus, we investigated the effects of SLN heterozygous deletion on atrial function for the first time [45] .Sln Cre/+ mice showed normal cardiac function that was the same as that of wild-type (WT) mice, even though the expression levels of SLN mRNA and protein were decreased by half.As physiological functions, we simultaneously measured the atrial peak tension and intracellular Ca 2+ concentration with the aequorin method.There was no significant difference in physiological functions between Sln Cre/+ and WT.In addition, the fibrosis and electrical pacing-induced atrial fibrillation that were observed in SLN null atria (mentioned above) were not induced in Sln Cre/+ mouse atria.Therefore, these data indicate that the Sln Cre/+ mouse line is appropriate for the atrium-specific gene-targeting strategy.
Interestingly, we found that Sln Cre/+ mouse atria showed less fibrotic change caused by chronic β-adrenergic stimulation of isoproterenol (ISO).It is known that chronic ISO stimulation induces fibrosis [46][47][48] .Accordingly, ISO stimulation increased fibrosis in both WT and Sln Cre/+ hearts compared with the control (saline-stimulated WT).The fibrosis level in Sln Cre/+ -ISO mouse atria was significantly lower than in WT-ISO mouse atria (Figure 2) whereas a similar grade of fibrosis was found in both WT and Sln Cre/+ ventricles.Further investigation is required to analyze the mechanisms of the protective effect of SLN deletion using other stress models such as pressure overload and/or volume Ratio of positive-to-total area was significantly increased after ISO stimulation in both WT-ISO and Sln Cre/+ -ISO mouse atria compared with WT-saline.The ratio in Sln Cre/+ -ISO mouse atria was significantly lower than WT-ISO atria.*P < 0.05, **P < 0.01, and ***P < 0.001 compared with saline-stimulated WT-saline mice; † †P < 0.01 compared with WT-ISO mice.Reprinted with permission [45].
overload and to find the signal pathway involved in fibrosis.

Atrium-specific gene targeting
The gene-targeting approach is an extremely important and essential method to modify gene expression and analyzing gene function.Although many types of heart-specific conditional KO mice have been generated, very few atrium-specific conditional KO mice have been developed, as discussed in the Introduction section [3] .de Lange et al. generated an atrium-specific Cre expression mouse line using the natriuretic peptide precursor type A gene (Nppa) promoter (Nppa-Cre mouse) [49] .The authors isolated and mutated the murine Nppa gene from -3 kbp to +4 kbp relative to the transcription site that contains the nlsCre gene and the neuron-restrictive silencer element (NRSE).Nppa expression is repressed in the ventricle by the NRSE/NRSF system [49,50] .The NRSE is known as a negative-acting DNA regulatory element that prevents the expression of neural genes in non-neuronal cells.The NRSE is located in the 3' untranslated region of the Nppa gene.The neuron-restrictive silencer factor (NRSF) formed a complex with mSin3 and the histone deacetylase to mediate the repression of Nppa promoter activity in ventricular myocytes [50]   .Therefore, the authors inserted Cre under the Nppa promoter construct to the atrium-specific Cre expression.The authors mated Nppa-Cre mice with R26R LacZ reporter mice to trace Cre expression and found that Cre expression was consequently induced in the atrium [49] .Nppa expression is not homogenous in the early developmental stage, however, and is partially induced in the ventricle.The Nppa gene is known as the biomarker of cardiac hypertrophy and heart failure, and its expression is induced in the ventricles [51,52] .Therefore, the reduction in the atrial specificity of Nppa is disadvantageous to the study of atrium-selective gene deletion under a pathological condition.
The expression of myosin light chain 2a (MLC2a) is also known to be restricted in the atrium around ED 12 in mice embryos [53] .Huang et al. inserted Cre into the MLC2a locus [54] and Wettschureck et al. used the MLC2a promoter [55] to generate the MLC2a-Cre mouse line.MLC2a null mice showed severe embryonic lethality although MLC2a-heterozygous mutants were viable [54] .Therefore, the MLC2a-Cre mouse line could be useful for atrium-specific gene targeting.It is of note, however, that MLC2a is expressed in both the atrium and ventricle before ED 12 (as mentioned above), although several previous studies have already used the MLC2a-Cre mouse line for cardiomyocyte-specific gene targeting [55][56][57] .
Our Sln Cre/+ mouse line was stable and sufficient for atrium-specific Cre/loxP recombination from the early embryonic stage to the full life span [10] .Using the Sln Cre/+ mouse line and tracing the cells that expressed SLN, we previously reported that Isl1+/SLN+ double-positive cells have the potential to give rise to both cardiac and smooth muscle cells in the cardiac inflow tract [10] .We also generated atrium-specific Nkx2.5 conditional KO mice [58] .The Nkx2.5 mutants died soon after birth at P1 or P2 and showed enlarged atria compared to the control.Moreover, we found a marked hyperplasia sinoatrial node and an atrioventricular node that expressed HCN4 and consequently caused abnormal conduction.Transcriptome analysis revealed activated Notch signaling, which suggests that Nkx2.5 has a suppressive role in atrial cardiomyocyte proliferation through Notch signaling [58] .GATA binding protein 4 (Gata4) is considered an important transcriptional factor for heart development and its mutation was found in human atrial septum defect patients [59] .Since Gata4 KO in the whole heart causes embryonic lethality before atrium formation [60] , the role of Gata4 in atrial development remains unclear.Therefore, we generated atrium-specific Gata4 KO mice and they did not show embryonic lethality (unpublished data).Analysis of the phenotype of atrium-specific Gata4 KO mice is currently underway.

SLN in the skeletal muscle
It is known that SLN is also expressed in the skeletal muscle, in addition to the atria [13] .Several previous studies indicated that mouse SLN mRNA is expressed in both fastand slow-twitch skeletal muscle, such as the extensor digitorum longus muscle (EDL) and soleus, but to a lesser extent than in the atrial muscle; rabbit and human SLN mRNA is much more abundant in fast-twitch skeletal muscle than in slow-twitch and atrial muscle [9,13,16] .It was also reported that rat diaphragm expresses SLN and that the expression is increased by corticosteroid treatment [61] .Periasamy's group reported that SLN in skeletal muscle is involved in nonshivering thermogenesis and whole body metabolism by promoting the uncoupling of the SERCA pump [23,62] .SLN null mice could not maintain their core body temperature after an acute cold challenge and were sensitive to diet-induced obesity.In addition, SLN overexpression in mouse skeletal muscle was resistant to fatigue and had increased twitch force, mitochondrial number and size, and muscle metabolism [63,64] .In terms of energy metabolism, the role of SLN in the heart has not yet been investigated.
In skeletal muscle pathology, it has been reported that SLN expression is increased in Duchenne muscular dystrophy and Srpk3 KO mouse skeletal muscle [65,66] .These data suggest that SLN may play a role in these abnormalities of mouse skeletal muscle or in compensating for the metabolic function.Future research is needed to clarify these issues.Notably, in terms of SERCA regulation, a micropeptide called myoregulin (MLN) that encodes skeletal muscle-specific putative long noncoding RNA has been recently discovered [67] .MLN shares structural and functional similarity with PLN and SLN, and is expressed in a skeletal muscle-specific manner but not in the cardiac muscle.Accordingly, MLN is considered to be a key regulator of SERCA in skeletal muscle, although the relationships and interaction between MLN and SLN in skeletal muscle remain unclear.If MLN works sufficiently in Sln Cre/+ mouse skeletal muscle, the effect of SLN heterozygous deletion on skeletal muscle function would be negligible.Further research may reveal the distinguished role of MLN and SLN in skeletal muscle.

Conclusion
SLN is stably expressed in an atrium-specific manner from the early embryonic stage to the full life span.We generated a Sln Cre/+ mouse line by inserting Cre recombinase into the endogenous SLN locus and revealed that the atrial function in Sln Cre/+ mice was largely normal, indicating that Sln Cre/+ mice have a great advantage for atrium-specific gene deletion.Many questions related to the development of the atrium remain to be answered.Investigation of these questions will aid in our understanding of the pathogenesis of congenital heart disease as well as the mechanism of development of mature atrial myocytes from stem cells in regenerative medicine.It is our hope that the Sln Cre/+ mouse line will be utilized in this future research.

Conflicting interests
The authors have declared that no conflict of interests exist.

Figure 2 .
Figure2.Fibrosis level after chronic ISO stimulation.Sirius Red staining (A) and Elastica van Gieson staining (B) of atria after chronic ISO stimulation.Ratio of positive-to-total area was significantly increased after ISO stimulation in both WT-ISO and Sln Cre/+ -ISO mouse atria compared with WT-saline.The ratio in Sln Cre/+ -ISO mouse atria was significantly lower than WT-ISO atria.*P < 0.05, **P < 0.01, and ***P < 0.001 compared with saline-stimulated WT-saline mice; † †P < 0.01 compared with WT-ISO mice.Reprinted with permission[45].