Adult cardiomyocytes isolation

First, inject mice with 0.1 ml heparin (2000IU/ml) and wait for 10 min. Then, inject mice with 0.2 ml urethane for anesthesia. After mice is fully anesthertized, open the chest and take heart out. Cannulate the heart and tie the aorta to the cannula with 5-0 silk thread. After that, perfuse the heart with Ca²⁺ free Tyrode buffer for 3 min at the rate 3ml/min. Then, switch to enzyme buffer (enzymes dissolve into perfusion buffer) and continue perfusion for 7 min at the same rate. Once the enzyme digestion is complete, trim the heart (keep the ventricles) and put the heart into Solution A buffer.

Then, separate and pipette the heart into pieces gently. Next, filter the cell suspension with 250 μm filter by gravity for 15 min. Transfer the cells to 0.06 mM, 0.24 mM, 0.6 mM, 1,2mM Ca²⁺ transfer buffer step by step every 10 min. For Ca²⁺ transient

measurement, transfer the cells into 1.2 mM Ca²⁺ contractility perfusion buffer. Load the fluorescent calcium indicator, Fura-2 AM (5 mM/ml), after 30 min wash the cells with 1.2 mM Ca²⁺ contractility perfusion buffer and wait for next 30 min. After that, measure the calcium transient with the cells. To measure the calcium transient, probenecid(100 μm) and glucose (10 mM) are freshly add into 1.2 mM Ca²⁺

contractility perfusion buffer for perfusion.

The following are buffers used in the experiment: Ca²⁺free Tyrode buffer (pH 7.4)

1.2 mM Ca²⁺contractility perfusion buffer (pH 7.4-7.55)

NaCl 137 mM

Ca²⁺free Tyrode buffer

Glucose 100 mM 1.2 mM Ca²⁺contractility

perfusion buffer

Chapter 3

RESULTS

3.1 Loss of NRIP leads to cardiac hypertrophy progressively.

To investigate insights into the in vivo function of NRIP, we have generated conventional NRIP^-/- mice as a genetic model. NRIP-null mice were observed no morphological abnormalities and appeared normal by the criteria including behavior, weight, and fertility. Intriguingly, NRIP mutant mice displayed significantly reduced exercise capacity according to our preliminary results of rotarod and treadmill tests (Chen HH, unpublished data). Compared with WT mice, NRIP^-/- mice showed considerably worse motor performance in rotarod and treadmill tests indicating that NRIP deficiency might lead to muscular weakness. Moreover, NRIP was found to be over-expressed in skeletal muscle and heart (Tsai et al, 2005) based on our previous study and these observations imply that NRIP might play functional roles in striated muscle such as skeletal and cardiac tissues. To assess the difference of muscle contractibility, the diaphragms excised from WT and KO mice were subjected to measure the muscle force in vitro. Indeed, NRIP deficient diaphragm showed weaker muscle strength than WT (Chen HH, unpublished data). Regarding that myocardium also belongs to striated muscle and the result of treadmill test indicates deficient

cardiopulmonary function (Bruce, 1974),we hypothesize that NRIP might also play a role in cardiac function. To elucidate our theory, we examined the cardiac function of WT and NRIP^-/- mice by echocardiography (Table 1) from young to middle aged stage.

According to the results, the measurements of ejection fraction (EF) and fraction shortening (FS) from NRIP^-/- mice are lower than WT implying that NRIP^-/- mice have cardiac functional defects. Additionally, the dimension of left ventricle, interventricular septum (IVSd), posterior wall (PW) and left ventricle mass were increased with aging showing that NRIP^-/- mice were on the process of cardiac hypertrophy to dilated cardiomyopathy. By comparing the hearts excised at 12 and that at 39 weeks, the sizes of ventricle and atrium were enlarged at elder stages (Figure 1). As shown in Figure 2A, the histological analysis using hematoxylin and eosin (H&E) staining indicates that left ventricle walls, such as IVS and PW, were thicken in the 12-week-old NRIP^-/- mice, but the chamber size in NRIP^-/- mice was similar with WT. Comparing with the data shown in Figure 3A, except the thickened IVS and PW, we also found the chambers of left and right ventricles were dilated in the 39-week-old NRIP^-/- mice which are consistent with conclusions of echocardiography. Besides, the sizes of cardiomyocytes were increased at 12 and 39-week-old NRIP^-/- mice comparing with that in WT (Figure 2B and 3B). Statistical analyses of cell sizes and number in WT and NRIP^-/- further

Accompanied with dilated cardiomyopathy, histological analysis with Massion’s trichrome staining revealed significantly fibrosis in the hearts of NRIP^-/- mice

compared to WT mice at 39 weeks (Figure 4). Taken together, NRIP deficiency in mice leads to progressively dilated cardiomyopathy and fibrosis in an age-dependent

manner.

3.2 NRIP interacts with a Z-disc protein, α-actinin-2, which is a major component of cardiac Z-disc apparatus maintaining the sarcomeric structure.

To investigate the molecular insights of cardiac hypertrophy in NRIP^-/- mice, a yeast two-hybrid screen was performed the full-length of NRIP as bait. Based on the results of two-hybrid protein-protein interaction assay, we speculated that the

interactive proteins of NRIP in a great majority belonged to the α-actinin family including α-actinin 1 to 4, each found within a specific tissue type and expression profile and these proteins can be grouped into two distinct classes: muscle (2 and 3) and non-muscle cytoskeletal (1 and 4) isoforms (Sjoblom et al, 2008). Both muscle isoforms (2 and 3) are commonly expressed in skeletal, cardiac, and smmoth muscle tissue and α-actinin-2 (ACTN2) is a major component of Z-disc in cardiac muscle (Sjoblom et al, 2008). Thus, we conjectured that NRIP might affect cardiac function through interacting with ACTN2. To confirm our suggestion plus the result of yeast

two-hybrid assay, we performed in vitro and in vivo protein-protein interaction assays.

Both recombinant GST-tagged ACTN1 or ACTN2 and His-tagged NRIP proteins expressed in bacteria were subjected to perform the in vitro binding assay (Figure 5).

As shown in Figure 5C (lane 6), NRIP was found to interact with ACTN2 in vitro, and the interactions between NRIP and ACTN1 or ACTN2 were calcium-enhanced (lane 5 and 7). As shown in Figure 5D (lane 2), NRIP was found to interact with ACTN2 in vitro. Furthermore, we examined whether NRIP associates with ACTN2 in vivo by

performing co-immunoprecipitation assay and expression vectors for EGFP-tagged NRIP and ACTN2-V5 were co-transfected into 293T cells. The protein lysates were than collected and immunoprtecipitated with anti-EGFP or anti-V5 antibody,

respectively (Figure 6). Figure 6C shows that the interaction between NRIP and ACTN2 also exists in mammalian cells (lane 5 and 8). In conclusion, we proved that NRIP interacts with ACTN2 both in vitro and in vivo.

3.3 The IQ motif of NRIP interacts with the CaM-like domain of ACTN2.

As described previously, NRIP has a calmodulin binding motif, IQ motif, which binds with calmodulin in a calcium-dependent manner (Chang et al, 2011). The IQ motif has been reported that it can associate with proteins containing EF-hand motif

proteins. As reported previously, the protein structure of α-actinin are composed of three domains, actin-binding domain (ABD), multiple spectrin repeats (SR) and calmodulin (CaM)-like domain (Sjoblom et al, 2008). Particularly, the carboxyl-terminal domain, calmodulin-like domain is composed of four EF-hand motifs. Therefore, we proposed that NRIP might interact with the CaM-like domain of ACTN2 through its IQ motif. To demonstrate that, we generated the CaM-like domain-truncated construct,

GST-ACTN2ΔEF-hand and the IQ motif-deleted construct (NRIPΔIQ), to perform the in vitro binding assay (Figure 7 A and C). Firstly, GST-ACTN2 protein and

GST-ACTN2ΔEF-hand protein purified separately from bacteria were incubated with His-MBP-NRIP protein and pulled down with GST-conjugated beads. The results show that ACTN2 binds to NRIP mainly through its C-terminal EF-hand motif (Figure 7B, lane 4). Reciprocally, to map which domain of NRIP is responsible for ACTN2 binding, the GST-ACTN2 protein expressed in bacteria was purified and incubated with various His-tagged NRIP fragments including full-length, N-terminus, C-termunus, and

IQ-deleted full-length to perform pull down assays. As described in Figure 7D, the ACTN2-binding activity of NRIPΔIQ was significantly reduced compared to the

C-terminal and full-length NRIP, indicating that NRIP interacts with ACTN2 through its C-terminal IQ motif. As the data shown in Figure 7, we defined that NRIP can directly interact with the CaM-like domain of ACTN2 through its IQ motif.

3.4 NRIP is a novel Z-disc protein and co-localized with ACTN2.

After a series of protein-protein interaction assays, we demonstrated that NRIP interacts with α-actinin-2 in vitro and in vivo, defined their interacting domain.

Therefore, it raised a question that NRIP is co-localized with ACTN2 in cardiac tissue.

To answer the question, the hearts tissues excised from WT and NRIP^-/- mice were fixed and co-stained with antibodies against NRIP, ACTN2, or myomesin to perform immunofluorescence assay (IF). IF with control antibodies such as α-actinin-2 and Myomesin labeling Z-disc and M-band, respectively indicated that both WT and NRIP deficient myocardium are well organized with regular cross-striations (Figure 8A).

Furthermore, NRIP was found to co-localize with ACTN2 and flanked by M-band labeled by myomesin. The subcellular distribution of NRIP was also confirmed within cardiomyocytes cultured from WT and NRIP-null mice at 12 weeks (Figure 8B).

According to the previous definition, a z-disc protein must conform to these

requirements. First, a suspected Z-disc protein must co-localize with a bona fide Z-disc protein, such as ACTN2, in immunofluorescence assay. Second, the Z-disc localization of the suspected protein should be detected by electron microscopy with immunogold labelling or biochemical evidences of direct protein–protein interaction with known

have demonstrated that NRIP interacts with ACTN2 in vitro and in vivo and these two proteins co-localize in Z-disc. Hence, we characterized that NRIP is a novel Z-disc protein.

3.5 Loss of NRIP reduces I-band width and widen the Z-disc of sarcomere.

Z-disc protein in the traditional concepts is a passive constituent scaffold of

sarcomeric structure. It cross-links with thin filament to stabilize the muscle contraction and transmit the force generated by the myofilaments. Many Z-disc proteins have proved that mutations or defects of these proteins disrupt cardiac cytoarchitectural

organization and lead to cardiomyopathy (Arber et al, 1997; Hassel et al, 2009; Knoll et al, 2011). Because NRIP has proved as a Z-disc protein and loss of NRIP leads to cardiac hypertrophy. Therefore, it raised a question whether lack of NRIP disrupts sarcommeric structure. To clarify the question, hearts were excised from adult mice, and fixed for transmission electron microscopy (TEM) analysis (Figure 9). As shown in Figure 9, deficiency of NRIP leads to narrower I-band and loose Z-disc. Because of the shortening of I-band, the boundary of H-zone is not clear in myocardium of NRIP -/-mice. To analyze the widths of sarcomeric structures (Figure 10 and Figure 11), the I-band width of NRIP^-/- mice was reduced 57.6% (0.11 microns short) and the Z-disc width was 15.2% widen (10.17 nm wide), but the A-band width was similar with WT.

The results of TEM and statistical analyses from hearts of embryonic day 17.5 (E17.5) and postnatal day 2 (P2) (Figure 12 and Figure 13) are corresponded with the

consequence from adult mice. The I-band width was reduced in NRIP^-/- mice at age E17.5 and P2. To further confirm the results of ultrastructural analysis, the frozen

sections of hearts from WT and NRIP^-/- were co-stained with phalloidin (F-actin) and anti-actinin antibody to perform immunofluorescence assay (IF) (Figure 14). As shown in Figure 14, the expression pattern of F-actin was more concentrative in Z-disc than the pattern of WT. This might implies deficiency of NRIP affects the expression pattern of F-actin and leads to narrower I-band. Following with the results above, we concluded that loss of NRIP affects the sarcomeric structure, especially the width of I-band. The narrower width of I-band might affect the contractility of hearts, and finally leads to dilated cardiomyopathy progressively.

3.6 Deficiency of NRIP decreases the amplitude of calcium transient.

According to the previous study, we know that NRIP is a calcium-dependent

calmodulin binding protein (Chang et al, 2011). Calmodulin (CaM) has been shown as a regulator of many ion channels, such as L-type Ca²⁺ channel, ryanodine receptor and IP3 receptor or as the Ca²⁺ sensor for signaling pathways in cardiac myocytes (Saucerman & Bers,

contractility of heart muscles (Marks, 2003). Hence, we interested to know whether loss of NRIP affects calcium transient of cardiomyocytes. Adult cardiomyocytes of WT and NRIP^-/- mice were isolated at 12 weeks and measured the calcium transient of

contractions (Figure 15). The preliminary result of calcium transient measurement shows that the calcium transient amplitude of NRIP^-/- mice is lower comparing with WT (Figure 15A). Further analyzed the ratiomeric calcium concentrations of base, peak and the calcium variation of an action, we found that the peak of calcium transient and the total calcium variation of an action are lower than WT (Figure 15B-D). Besides, the time to relaxation is longer in NRIP^-/- than in WT (Figure 15F). According the

investigation of calcium transient, we speculated NRIP might involve in regulating the calcium concentration in cardiomyocytes. Deficiency of NRIP decreases the variation of calcium concentration and impairs cardiac function.

Chapter 4

DISCUSSION

In summary, loss of NRIP defects cardiac function and leads to cardiac hypertrophy and fibrosis progressively. NRIP affects cardiac function might through interacting with a cardiac Z-disc protein, ACTN2. The IQ motif of NRIP associates with the CaM-like domain of ACTN2. Besides associates with ACTN2 biochemically, NRIP also co-localizes with ACTN2 in Z-disc histologically. Being a novel Z-disc protein, NRIP affects organizations of sarcomere, which narrows I-band and widens Z-disc.

Moreover, as a calcium-dependent calmodulin binding protein, deficiency of NRIP decreases the amplitude of calcium transient. According to these results, we speculated that the effects of sarcomeric structure and calcium concentration impair the

contractility of myocardium. To compensate the cardiac function, the hearts of NRIP -/-mice trend to hypertrophic cardiomyopathy progressively.

4.1 Deficiency of NRIP leads to cardiac hypertrophy.

According to our study, we know that deficiency of NRIP causes cardiac hypertrophy.

In most forms of cardiac hypertrophy, the expression of embryonic genes is increased, including the genes for natriuretic peptides and fetal contractile proteins (Hunter &

Chien, 1999). In our study, we have found the pathological morphology of hypertrophic hearts in NRIP^-/- mice, including the thickened ventricular walls and the enlarged cardiac myocytes. To further confirm the cardiac hypertrophy, we might investigate the gene expression of hypertrophic markers, such as atrial natriuretic factor (ANF) and

brain natriuretic peptide (BNP) (Kim et al, 2008; Vikstrom et al, 1998). Because we have found fibrosis in NRIP^-/- mice at elder stages, maybe the gene expression of some fibrosis markers, such as connective tissue growth factor (Ctgf), procollagen, type I, α2 (Col1a2), and procollagen, type III, α1 (Col3a1), could also be investigated.

4.2 NRIP reduces I-band width through affecting proteins involving in actin filament assembly.

Being a Z-disc protein, the defects of NRIP deletion are not similar with other Z-disc proteins, such as MLP or Nexilin, which deletion of MLP or Nexilin leads to Z-disc disarrangement(Arber et al, 1997; Hassel et al, 2009). Lack of NRIP mainly affects the length of actin filament, no matter in embryonic or in adult stages. Actin filament (F-actin) is organized from the polymerization of actin monomer (Ono, 2010; Taylor et al, 2011). Proteins involved in the process of actin polymerization such as CapZ, tropomodulin-1 (Tmod1) and nebulin have been reported that defects of these proteins

affects the length of actin filaments (Cooper & Schafer, 2000; Gokhin & Fowler, 2011;

Hart & Cooper, 1999; Littlefield et al, 2001; Littlefield & Fowler, 1998; Witt et al, 2006). From the study of Witt C. et al, deletion of nebulin lead to reduction of I-band length and ~15% extension of Z-disc, these phenotype are similar with our

observations. In the study of Witt C et al, they speculated that the loss of nebulin might affect actin filament stability by decelerating actin nucleation, affecting actin

termination. Therefore, lengths of actin filaments are reduced. According to the study of Witt c et al, we speculated that NRIP might affect the expression or localization of proteins involving in actin filament formations. For clarify the speculation, we can investigate the protein expression or expression pattern of these proteins in

cardiomyocytes of NRIP^-/- mice foremost.

4.3 NRIP disrupts myofibrilar arrangements through decreasing gene expressions of genes involving in actin filament formation.

Many Z-disc proteins have been reported as a mechanical or biochemical sensors, such MLP, zyxin and myopodin (Frank et al, 2006). MLP is a sensor of mechanical stimuli, cyclic stretch triggers MLP shuttles from Z-disc to nucleus (Boateng et al, 2007), and then induces downstream gene expression, such as MyoD (Frank et al, 2006).

stress of heat shock (Weins et al, 2001). According to our previous studies, we knew that NRIP is a ligand-dependent transcriptional co-activator of androgen receptor in prostate cancer cell lines. (Chen et al, 2008; Tsai et al, 2005). In addition, except seven WD-40 repeats and one IQ motif, NRIP also contains a nuclear localization sequence (NLS) (Tsai et al, 2005). Consistent with the study of Tsai et al, we found that besides localizing in Z-disc, NRIP is also expressed in nuclei of cardiomyocytes at P2 (data not shown) or adult stages. Furthermore, treated neonatal cardiomyocytes with A23187 (a divalent cation ionophore) (Reed & Lardy, 1972) enhanced the nuclear expression of NRIP (Data not shown). As described previously, NRIP interacts with calmodulin in calcium-dependent manner. In myocytes, calmodulin plays as a calcium sensor to regulate calcium-dependent signaling or activities the ion channels (Frank et al, 2006;

Ohrtman et al, 2008). In the study of Wyszynski et al, elevation of the cellular calcium concentration triggers calmodulin to compete the interactions between NMDA receptor and α-actinin, and then to release α-actinin from NMDA receptor(Wyszynski et al, 1997). Therefore, we speculated that in cardiomyocytes calmodulin senses the calcium signals and then associated with NRIP. The interactions of each other release NRIP to translocate into nucleus and to regulate the expressions of genes associated with actin filament formations and stabilizations. To clarify our speculations, the microarray data is essential to help us find out the downstream targets of NRIP. In addition, the

translocation of NRIP and the competition between calmodulin and ACTN2 must further confirm.

4.4 NRIP plays a role in regulating calcium homeostasis.

According to the calcium transient measurement of WT and NRIP^-/- mice, we found that loss of NRIP decreases the amplitude of calcium transient. But the

mechanism of NRIP regulating calcium homeostasis is still unknown. Many ion channel inhibitors or drugs are applied to investigate the calcium storage or the ability of

calcium removal of cells. For examples, caffeine is applied to measure the Ca²⁺ content of sarcoplasmic reticulum (SR) and thapsigargin (TG) or cyclopiazonic acid (CPA) are applied as SR/ER Ca²⁺ pump inhibitors (Campbell et al, 1991; Smith & Steele, 1998).

Therefore, ion channel inhibitors or drugs could be applied to further investigate which NRIP participate in which state of calcium regulation, calcium influx or removal.

REFERENCE

Arai M, Alpert NR, MacLennan DH, Barton P, Periasamy M (1993) Alterations in sarcoplasmic reticulum gene expression in human heart failure. A possible mechanism for alterations in systolic and diastolic properties of the failing myocardium. Circulation research 72: 463-469

Arber S, Hunter JJ, Ross J, Jr., Hongo M, Sansig G, Borg J, Perriard JC, Chien KR, Caroni P (1997) MLP-deficient mice exhibit a disruption of cardiac

cytoarchitectural organization, dilated cardiomyopathy, and heart failure. Cell 88:

393-403

Bahler M, Rhoads A (2002) Calmodulin signaling via the IQ motif. FEBS letters 513: 107-113

Barry WH, Bridge JH (1993) Intracellular calcium homeostasis in cardiac myocytes. Circulation 87: 1806-1815

Boateng SY, Belin RJ, Geenen DL, Margulies KB, Martin JL, Hoshijima M, de Tombe PP, Russell B (2007) Cardiac dysfunction and heart failure are associated with abnormalities in the subcellular distribution and amounts of oligomeric muscle LIM protein. American journal of physiology Heart and circulatory physiology 292: H259-269

Broderick MJ, Winder SJ (2005) Spectrin, alpha-actinin, and dystrophin.

Advances in protein chemistry 70: 203-246

Bruce RA (1974) Methods of exercise testing. Step test, bicycle, treadmill, isometrics. The American journal of cardiology 33: 715-720

Campbell AM, Kessler PD, Sagara Y, Inesi G, Fambrough DM (1991) Nucleotide

Campbell AM, Kessler PD, Sagara Y, Inesi G, Fambrough DM (1991) Nucleotide