The CREB leucine zipper regulates CREB phosphorylation, cardiomyopathy, and lethality in a transgenic model of heart failure

Abstract

Signaling through cAMP plays an important role in heart failure. Phosphorylation of cAMP response element binding protein (CREB) at serine-133 regulates gene expression in the heart. We examined the functional significance of CREB-S133 phosphorylation by comparing transgenic models in which a phosphorylation resistant CREB-S133A mutant containing either an intact or a mutated leucine zipper domain (CREB-S133A-LZ) was expressed in the heart. In vitro, CREB-S133A retained the ability to interact with wild-type CREB, whereas CREB-S133A-LZ did not. In vivo, CREB-S133A and CREB-S133A-LZ were expressed at comparable levels in the heart; however, CREB-S133A markedly suppressed the phosphorylation of endogenous CREB, whereas CREB-S133A-LZ had no effect. The one-year survival of mice from two CREB-S133A-LZ transgenic lines was equivalent to nontransgenic littermate control mice (NTG), whereas transgenic CREB-S133A mice died with heart failure at a median 30 wk of age (P < 0.0001). CREB-S133A mice had an altered gene expression characteristic of the failing heart, whereas CREB-S133A-LZ mice did not. Left ventricular contractile function was substantially reduced in CREB-S133A mice versus NTG mice and only modestly reduced in CREB-S133A-LZ mice (P < 0.02). When considered in light of other studies, these findings indicate that overexpression of the CREB leucine zipper is required for both inhibition of endogenous CREB phosphorylation and cardiomyopathy in this murine model of heart failure.

Fig. 1

A: diagram of cAMP response element binding protein (CREB)-S133A and CREB-S133A-leucine zipper (LZ) protein structure. The amino acid location of each mutation is shown in relationship to the kinase inducible domain (KID), the DNA binding domain (DBD), and the LZ. B: CREB-S133A-LZ does not associate with the CREB LZ. Twenty percent of the [³⁵S]-methionine-labeled in vitro translation product used for binding is shown in the first lane. After the radiolabeled proteins had been incubated with immobilized glutathione-S-transferase (GST) or GST-CREB, the beads were washed and the bound proteins were eluted and separated by 10% SDS-PAGE. CREB-S133A was selectively retained by the CREB LZ GST fusion protein and not by GST alone (B, top). The GST-CREB protein, which includes only the CREB LZ, did not retain the CREB-S133A-LZ mutant protein (B, bottom). Shown are representative results from 3 experiments.

Fig. 2

A: cardiac expression of hemagglutinin (HA)-CREB mutant proteins. Equivalent amounts of whole heart lysates were separated and probed using an anti-CREB antibody (A, top) and an anti-HA antibody (A, bottom). Endogenous CREB protein was evident as 2 molecular species (top, solid arrows). The CREB transgene, which migrates more slowly because of the HA tag, was visible above the endogenous protein (top, open arrowhead). CREB-S133A expression was detected as a weak band by HA staining (bottom, second lane), whereas 2 lines of CREB-S133A-LZ mice showed more abundant transgene expression. B: CREB-S133A-LZ mice have preserved serine-133 phospho-CREB levels. Whole heart lysate immunoblots following SDS-PAGE were probed for serine-133 phospho-CREB levels. The separation of equivalent amounts of whole heart lysate is shown by Ponceau red staining (B, bottom). Abundant serine-133 phospho-CREB was found in nontransgenic (NTG) and transgenic CREB-S133A-LZ mice but not in CREB-S133A mice. Shown is a representative experiment of 3 from the LZ mutant line 1.

Fig. 3

CREB-S133A-LZ mice survive normally. Graph shows Kaplan-Meyer curves of CREB-S133A (n = 118), CREB-S133A-LZ line 1 (n = 88), CREB-S133A-LZ line 2 (n = 93), and NTG (n = 149) littermate control mice. Although early death of CREB-S133A mice became apparent by 20 wk of age, no significant death of CREB-S133A-LZ mice was evident by 1 yr.

Fig. 4

CREB-S133A-LZ hearts do not express markers of heart failure. Bars represent (means ± SE) brain natriuretic peptide (BNP; left), β-myosin heavy chain (β-MHC; middle), and sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA2; right) transcript abundance relative to GAPDH. Whereas expression of the BNP and β-MHC transcripts was significantly different in CREB-S133A mice, the LZ mutant mice showed no difference compared with NTG controls. *P < 0.05 compared with NTG.