OUP user menu

Decreased protein and phosphorylation level of the protein phosphatase inhibitor-1 in failing human hearts

Ali El-Armouche, Torsten Pamminger, Diana Ditz, Oliver Zolk, Thomas Eschenhagen
DOI: http://dx.doi.org/10.1016/j.cardiores.2003.11.005 87-93 First published online: 1 January 2004

Abstract

Objective: The protein phosphatase inhibitor-1 (I-1) is a highly specific and potent inhibitor of type 1 phosphatases (PP1) that is active only in its protein kinase A (PKA)-phosphorylated form. I-1 ablation decreases, I-1 overexpression sensitizes β-adrenergic signaling in the heart. It is controversial whether I-1 expression is altered in human heart failure (HF), likely because its detection in heart is difficult due to its low abundance. Methods and results: I-1 was >500-fold enriched from left ventricular myocardium (LVM) from patients with terminal HF (n = 16) and non-failing controls (NF, n = 5) and quantified with an affinity-purified I-1 and a I-1 phosphospecific antiserum. In non-failing I-1 protein levels amounted to 126 fmol/mg protein. In failing hearts, I-1 protein levels were reduced by 58% and I-1 phosphorylation by 77% (P<0.001 vs. NF). I-1 phosphorylation correlated well with serine-16 phosphorylation of phospholamban (PLB) in the same hearts (P<0.001). In contrast, PLB, troponin I (TnI) and PP1 protein and TnI phosphorylation levels did not differ between HF and NF. Conclusions: The results suggest that the reduction in I-1 protein and phosphorylation in failing human hearts leads to increased phosphatase activity which in turn may result in reduced phosphorylation of cardiac proteins such as PLB.

Keywords
  • Heart failure
  • Protein phosphatases
  • Phospholamban
  • Troponin I
  • Adrenergic signal transduction

1 Introduction

The protein phosphatase inhibitor-1 (I-1) is an interesting potential modulator of the phosphorylation/dephosphorylation balance. It acts as a potent and highly specific inhibitor of protein phosphatase type 1 (PP1) only when phosphorylated by the cAMP-dependent protein kinase A (PKA). As such it could provide amplification of PKA-mediated signals [1].

Despite its early discovery and its fairly defined role in β-adrenergic inhibition of glycogen synthesis [1] the role of I-1 in the heart remained obscure for long. Recent studies now show that targeted ablation of I-1 increased PP1-activity, reduced isoprenaline-stimulated phospholamban (PLB) phosphorylation and impaired β-adrenergic contractile responses in mouse heart [2]. Conversely, adenoviral overexpression of I-1 or a constitutively active I-1 mutant sensitized cardiac myocytes to β-adrenergic stimulation [2,3]. Overexpression of I-1 was associated with increased isoprenaline-stimulated PLB phosphorylation [3]. These data substantiated an amplifier role of I-1 in β-adrenergic signaling of cardiac myocytes.

Uncertainty remains as to the expression level of I-1 in human heart failure (HF). One study concluded that I-1 protein levels were unchanged, but I-1 phosphorylation was reduced in samples from patients with HF [2]. Our own data showed reduced I-1 transcript levels and failed to detect a specific I-1 protein signal in standard homogenates [3]. The present study reevaluated this question by utilizing an extraction protocol and an affinity-purified I-1 antiserum.

2 Methods

2.1 Recombinant I-1 protein

Recombinant I-1 protein was generated by cloning the complete cDNA of rat heart I-1 into the pGEX-λT expression vector (Stratagene) as described previously [3]. In short, the plasmid was transformed into competent Escherichia coli BL21 (Novagen) to express I-1/glutathione-S-transferase (I-1–GST). I-1–GST was purified from clarified bacterial lysate by affinity purification with glutathione agarose beads. Loaded beads were washed three times, and recombinant I-1 was released by thrombin-cleavage.

2.2 Human myocardial tissue

Failing hearts were obtained from patients undergoing heart transplantation due to terminal HF (n = 9 dilated cardiomyopathy, DCM, and n = 7 ischemic cardiomyopathy, ICM). LV ejection fraction was 16–25%, cardiac index 1.7–2.7 l/min × m2. Most patients received ACE inhibitors, diuretics and cardiac glycosides, seven received calcium channel blockers, 13 nitrates and three antiarrhythmic drugs in addition. No patient received β-blockers. Five non-failing donor hearts (NF) that could not be transplanted for technical reasons were used for comparison. Donor patient histories or echocardiography revealed no signs of heart disease (for detail see Table 1). The study conforms with the principles outlined in the Declaration of Helsinki and was reviewed and approved by the Ethical Committee of the University Hospital Hamburg (Az. 532/116/9.7.1991).

View this table:
Table 1

Patient data

Patient #AgeGenderDiagnosisNYHA classLVEF (%)CI (l/min × m2)Drugs
144MNF (SAB)n.d.
252MNF (ICB)n.d.
350MNF (SAB)n.d.
442FNF (CIC)n.d.
519MNF (SAB)n.d.
647MDCMIV162.1DGNCA
765MDCMIV17n.d.DNR
856MDCMIII–IV251.7DGNCA
944MDCMIVn.d2.7DGNCAR
1059MDCMIIIn.d.1.7DGNCA
1132MDCMIV403.5DGNA
12n.d.n.d.DCMn.dn.d.n.d.n.d.
1345MDCMIVn.d.1.3NC
1448MDCMIII–IVn.d.1.4DGNC
1566MICMIII–IV161.9DNCAR
1664MICMIV20–30n.d.DNA
1746MICMIII–IVn.d.n.d.n.d.
1854MICMIII–IVn.d.n.d.DGNR
1964MICMIII–IV221.8DGAO
2057MICMIII–IVn.d.n.d.DGN
2165MICMIV20n.d.DGNR
  • LVEF, left ventricular ejection fraction; CI, cardiac index. Diagnosis: ICB, intracerebral bleeding; SAB, subarachnoidal bleeding; CIC, cerebral ischemia; DCM, idiopathic dilated cardiomyopathy; ICM, ischemic cardiomyopathy; NF, non-failing donor. Drugs: A, angiotensin converting enzyme inhibitors or angiotensin II receptor antagonists; C, calcium channel blockers; D, diuretics; G, cardiac glycosides; N, nitrates; R, antiarrhythmics (except β-AR blockers); O, dopamine/dobutamine; n.d., unknown.

2.3 Enrichment of I-1

I-1 was enriched from mouse skeletal muscle (wild-type [WT] and I-1 knockout [I-1(−/−)]), rat and rabbit skeletal muscle (500 mg each) and human left ventricular myocardium(LVM, 1 g) by an optimized extraction procedure [3–6]. In short, frozen tissue was pulverized with a mortar in liquid nitrogen, homogenized with a teflon/glass potter in ice-cold 1.5% trichloroacetic acid (TCA [wt./vol.], 4 mmol/l EDTA) and centrifuged at 20,000 × g for 30 min. The supernatant was adjusted to 19% TCA, incubated at 4 °C for 12 h, and centrifuged as above. The pellet was resuspended in 500 mmol/l Tris (pH 8.0), boiled for 10 min and centrifuged as above. The supernatant was dialyzed for 16 h against water at 4 °C, centrifuged as above and stored at −80 °C for further use. Protein was determined according to Bradford.

2.4 Western blot

SDS-PAGE and blotting were carried out as described [3]. Membranes were blocked with 5% (w/v) dried milk in 100 mmol/l Tris, pH 7.5, 0.1% (v/v) Tween 20 and 150 mmol/l NaCl (TBST) for 1 h prior to overnight incubation at 4 °C with the primary antibodies. Primary antibodies were a custom-made (Eurogentec, Brussels) rabbit polyclonal affinity-purified antibody against recombinant full-length rat I-1 that cross-reacts with mouse, rabbit and human I-1 (FB 70, 1/1000), and antibodies against rabbit skeletal muscle I-1 (1/200, kindly provided by P. Greengard, G 184), phospho-DARPP-32 (1/2000, Cell Signaling Technology, Beverly, MA), calsequestrin (CSQ, 1/2500; Dianova, Hamburg, Germany), PP1 (1/500, Upstate, Lake Placid, NY), cardiac troponin I (TnI, 1/30,000, Chemicon, Temecula, CA), phospho-TnI (1/30,000, HyTest, Turku, Finland), total PLB and Ser16-phosphorylated PLB (both 1/5000; PhosphoProtein Research, Bardsey, UK). Immunoblots were developed with anti-rabbit or anti-mouse IgG-horseradish peroxidase, subjected to Enhanced Chemiluminescence PLUS detection reagents (Amersham) and exposed to film for appropriate times. Densitometric signals on X-ray films were evaluated with GelDoc (Biorad).

2.5 Statistical analysis

Data are mean±S.E.M. Statistical analysis was performed using ANOVA or Students t-test, correlation analysis with Spearman–Rank correlation. P<0.05 was considered significant.

3 Results

3.1 Detection and quantitation of I-1 protein in immunoblots

To identify a specific “I-1 band”, immunoblots of standard homogenates for both WT and I-1(−/−) heart (30 μg) were probed with G 184 I-1 antibody (against rabbit skeletal muscle I-1) and FB 70 I-1 antibody (against rat heart I-1) and for negative control with G 184 incubated with 37.5 μg recombinant I-1 protein (for preadsorption) or FB 70 preimmune serum. As shown in Fig. 1A, there was no difference in the pattern of the immunoreacting bands. This confirmed our previous report that I-1 protein is not detectable in standard homogenates [3] and other published data showing that TCA-extracts of at least one whole guinea-pig heart were necessary for detection of I-1 by Western blots [6,7]. In contrast, immunoblots of TCA-extracts from WT and I-1(−/−) skeletal muscle (10 μg, ∼25 mg wet weight) probed with FB 70 I-1 antibody showed a strong immunoreactive band at ∼29 kDa in skeletal muscle from WT but not from I-1(−/−) mice (Fig. 1B). To further confirm the validity of this extraction method and the specificity of detecting I-1, immunoblots of TCA-extracts from rat skeletal muscle, rabbit skeletal muscle and human LVM were probed with the FB 70 I-1 antibody or preimmune serum. As shown in Fig. 1C a band of ∼29 kDa (rat) or ∼26 kDa (rabbit, human) was detected by the I-1 antibody, but not by its preimmune serum. The species-dependent difference in size of I-1 and the apparent molecular mass are in accord published data [6,8,9].

Fig. 1

(A) Immunoblots of standard heart homogenates (30 μg) for WT and I-1(−/−) were probed with G 184 I-1 antibody (against rabbit skeletal muscle I-1) and G 184 incubated with 37.5 μg recombinant I-1 (for preadsorption, top) as well as with the FB 70 I-1 antibody (against rat heart I-1) and its preimmune serum (bottom). (B) Immunoblot of TCA-extracts for WT and I-1(−/−) skeletal muscle (10 μg, ∼25 mg wet weight) were probed with FB 70 I-1 antibody. Note, the immunoreactive signal at ∼29 kDa in WT skeletal muscle is absent in the I-1(−/−) tissue. (C) Immunoblot of TCA-extracts enriched in I-1 isolated from rat skeletal muscle (5.25 μg), rabbit skeletal muscle (3.3 μg) and human LVM (6.0 μg) were probed with the FB 70 I-1 antibody or for control with preimmune serum. Note, a protein band of ∼29 kDa (rat) and ∼26 kDa (rabbit, human) were detected by the I-1 antibody, but not by the preimmune serum. (D) Immunoblot to quantitate absolute I-1 in TCA-extracts from non-failing human LVM (range of 3.5–14 μg) using the FB 70 I-1 antibody and increasing amounts for recombinant I-1 for calibration (range of 6.25–25 ng).

To get an estimate of the absolute level of I-1 in human heart tissue, a dilution of recombinant rat I-1 (6.25–25 ng) was probed with the FB 70 antiserum in parallel with a dilution of TCA-extract from a non-failing heart (3.5–14 μg; Fig. 1D). The intensity of the immunoreactive band at ∼29 kDa of 12.5 ng recombinant I-1 was almost identical to the band at ∼26 kDa of 10.5 μg TCA-extract protein (according to 34.2 mg wet weight tissue and 5.3 mg protein in the initial homogenate). Under the assumption that the antiserum detects human I-1 with the same affinity as rat I-1 this experiment finds I-1 levels in non-failing human hearts at 126 fmol/mg protein or 126 nM. This concentration is well above the IC50 of PP1 inhibition (1 nM).

3.2 PP1, I-1 protein and I-1 phosphorylation levels in failing human hearts

Protein levels of protein PP1 were analyzed in total homogenates and normalized to CSQ (20 μg per lane). No differences were observed (Fig. 2A). I-1 protein and phosphorylation levels were determined in TCA-extracts. 1 g LV-tissue yielded 120±4 mg protein in the initial homogenate and 234±13 μg protein in the TCA-extract (n = 21), with no differences between the groups. Immunoblots of these extracts (7 μg per lane≅30 mg wet weight) demonstrated that the signal intensity was significantly smaller in failing than in non-failing hearts, independent of the etiology with a mean reduction of 58% (Fig. 2B, left panel). Immunoblots of the same homogenates were probed with an I-1 phosphospecific antibody [10] and demonstrated a mean reduction of phospho-I-1 in failing hearts by 77% (Fig. 2B, right panel). Normalization to internal control proteins was impossible after TCA extraction, but Ponceau S staining of blots and Coomassie-stained gels run in parallel confirmed equal protein loading (Fig. 2B, top). The lack of systematic differences in the band pattern argues against confounders due to TCA extraction.

Fig. 2

(A) PP1 protein levels in homogenates from hearts with dilated (DCM) or ICM or donor hearts (NF) normalized to CSQ. (B) I-1 and phospho-I-1 in TCA-extracts from the same hearts. Coomassie-stained gels run in parallel demonstrates equal protein loading and similar band pattern. *P<0.05 vs. NF. #Samples and order are identical to (A).

3.3 Protein and phosphorylation levels of PLB and TnI in failing human hearts

PLB and TnI protein and phosphorylation levels were analyzed in standard homogenates (20 μg per lane) from the same samples as above and normalized to CSQ (Fig. 3A). PLB protein, TnI protein and TnI phosphorylation were similar in the groups (NF vs. HF: PLB/CSQ 5.58±0.60 vs. 5.73±0.25; TnI/CSQ 0.63±0.02 vs. 0.64±0.03; Phospho-TnI/TnI 1.14±0.05 vs. 1.13±0.04). However, phospho-PLB/PLB ratio was significantly reduced in HF (NF 1.11±0.32, HF 0.51±0.10, P<0.05). Phospho-PLB positively correlated with phospho-I-1 (Spearman r = 0.70, P<0.001, Fig. 3B).

Fig. 3

(A) Representative blots of total TnI, phospho-TnI, total PLB and phospho-serin-16-PLB in standard homogenates of the same hearts as in Fig. 1. The blot for CSQ, PP1 (Fig. 1A), TnI and phospho-TnI is identical, the blot for PLB and phospho-PLB (and CSQ) was run in parallel. (B) Correlation of phospho-I-1 with phospho-TnI and phospho-PLB, respectively.

4 Discussion

This study demonstrates that I-1 is markedly decreased and dephosphorylated, and therefore, inactive in failing human hearts. The reduction in I-1 correlates well with dephosphorylation of PLB. The study took advantage of the fact that I-1 is heat stable and, in contrast to most other proteins, not precipitated by low concentrations of TCA [4]. TCA-extracts were prepared under conditions where I-1 phosphorylation state is preserved [4,5]. Immunoblots of these extracts demonstrated a prominent band at ∼26 kDa that fulfilled the necessary criteria of specificity (see above). We believe that the band, we and others have seen before in standard homogenates [2,3] represents a non-specific cross-reacting protein. This is supported by the following arguments. (i) The band migrates at slightly higher molecular weight (∼29 vs. 26 kDa), (ii) is seen with preimmune serum, (iii) is not blocked by preincubation of the antiserum with recombinant I-1, and (iv) is seen in cardiac homogenates from I-1 knockout mice (Fig. 1). One could argue that a protein that is seen in immunoblots only after 500-fold enrichment is unlikely to play an important role. However, phosphorylated I-1 is highly potent (IC50 1 nM; [11]) suggesting that it does not need to be expressed at high concentrations. Our present analysis now shows I-1 levels in non-failing human heart to amount to 126 nM (Fig. 1D). This is clearly above the published IC50 and not far from estimates made for rat heart (∼500 nM; [4,12]).

Compared to other alterations in protein expression in human HF, the magnitude of the reduction in I-1 protein levels by 58% is large. Other well-studied proteins have been found to be downregulated by only 25–50% (e.g. β1-adrenergic receptors, SR Ca2+ ATPase, PLB, potassium channels, myosin light chains). We have found previously that I-1 transcript levels in failing human hearts were reduced by 50% [3]. In addition, isoprenaline-infusion in rats led to a 30% downregulation of I-1 mRNA levels [13]. These data suggest that gene expression of I-1, similar to β-adrenoceptors, is reduced by chronically elevated catecholamine levels in HF. I-1 phosphorylation was even more drastically diminished (−77%). This observation most likely reflects desensitization of β-adrenergic signaling with decreased cAMP levels and PKA activation in human HF (for review see [14]). Another contributor to diminished I-1 phosphorylation could be increased calcineurin activity in human HF [15], a phosphatase that dephosphorylates I-1.

Evaluation of further downstream elements of the β-adrenergic signaling cascade revealed interesting results. In the identical hearts in which I-1 was downregulated and dephosphorylated and PLB was dephosphorylated, PP1 protein and TnI protein/phosphorylation levels were unaffected. Unchanged PP1 is in accordance with unchanged global activity [16], but unchanged TnI phosphorylation was somewhat counterintuitive because TnI is a substrate for PKA just like PLB. Moreover, its phosphorylation level can be experimentally increased by PP inhibition [17,18] suggesting a role for PP1, and therefore, also I-1. Indeed, studies in failing human hearts found basal TnI phosphorylation to be reduced [19,20], but others found it unchanged in human HF [21] and in animal models [22]. Basal phosphorylation is notoriously prone to experimental conditions, which may explain the discordant results. Nevertheless, the experiments have been performed in identical samples, showing that phosphorylation of TnI is indeed less affected by HF than that of PLB. The observation that isoprenaline-stimulated TnI phosphorylation with a 20-fold higher sensitivity than that of PLB [23] could offer an explanation in the sense that basal phosphorylation of TnI is high, despite β-adrenergic desensitization. In addition, intracellular compartmentalization of PP1 and I-1 may play a role. Marks and colleagues [24] have proposed differential anchoring of PP1 to different compartments to explain differences between ryanodine receptor and PLB phosphorylation in HF and it is tempting to speculate that I-1 participates in this process. The marked reduction in I-1 identified here will cause less inhibition of PP1. This does not appear to translate into a globally increased PP1-activity [16], but goes along with a 2.5-fold increased in SR-associated PP1-activity [16], suggesting that I-1 preferentially localize to the SR. The good correlation between PLB and I-1 phosphorylation supports the notion of a causal relationship and argues for preferential affection of the free SR (PLB) compared with the junctional SR (ryanodine receptors) or the myofilaments (TnI). Thus, the reduction in I-1 could be one of the factors that explain differential affection of TnI and PLB phosphorylation in HF.

PLB is a major regulator of cardiac contractility [25] and accelerated PP1-mediated dephosphorylation of PLB will impair Ca2+ resequestration into the SR, reduce systolic Ca2+ release, and elevate diastolic Ca2+. The latter would activate calcineurin, which dephosphorylates I-1 and thereby accelerates a vicious cycle. Thus, the reduction in I-1 is a newly recognized factor that, besides changes in the adrenergic signaling cascade and sarcoplasmic reticulum function, can contribute to reduced systolic and diastolic function and to the blunted inotropic response to β-adrenergic stimulation in human HF.

Acknowledgements

We thank Jutta Starbatty for providing expert technical assistance. We also like to thank Dr. Paul Greengard and Peter Ingrassia for I-1 knockout mice. This work was supported by the Deutsche Forschungsgemeinschaft (Es 88/8-2, GRK 750) and the Johannes und Frieda Marohn-Stiftung (Arm/00).

Footnotes

  • 1 Both authors contributed equally to this work.

  • Time for primary review 26 days

References

  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
View Abstract