Fluorescence Spectroscopy Reveals a 2:1 Stoichiometry for Dimeric Cardiac Calcium Pump (SERCA2a) Bound to a Phospholamban Monomer

Abstract

Previous in vitro studies have suggested that the cardiac calcium ATPase (SERCA) can form functional homo‐dimers. To determine whether SERCA dimerizes in live cells, we fused Cerulean (Cer) or yellow fluorescent protein (YFP) to the N‐terminus of canine SERCA2a. We measured 18% maximum fluorescence resonance energy transfer (FRETmax) from Cer‐SERCA to YFP‐SERCA, suggesting an interaction between SERCA protomers. To determine whether the measured FRET was due to a specific interaction, we co‐transfected cells with unlabeled SERCA to compete for binding with Cer‐SERCA and YFP‐SERCA. FRETmax progressively decreased with increasing expression of unlabeled SERCA to a minimum of 5%. Competition with increasing molar ratios of unlabeled phospholamban (PLB) did not reduce the measured FRET or the apparent affinity of the SERCA dimer. In order to determine whether PLB phosphorylation status alters SERCA dimerization, we induced phosphorylation of PLB at serine 16 with PKA activation by Forskolin. PLB phosphorylation or phosphomimetic mutation (S16E) did not alter the maximum SERCA‐SERCA FRET or the apparent dissociation constant (Kd) of the SERCA dimer. The apparent affinity of the dimer did not vary with the presence or absence of calcium, ATP, or with the addition of the SERCA inhibitor, thapsigargin. The SERCA‐SERCA interaction was further supported by co‐immunoprecipitation experiments in cells co‐transfected with YFP‐SERCA and myc‐tagged SERCA. Immunoprecipitation with an anti‐mycTag antibody co‐immunoprecipitated YFP‐SERCA, as detected by western blot. Rabbit left ventricular myocytes were isolated and permeabilized in the presence of a benzophenone‐4‐maleimide crosslinker. Western blot analysis revealed two bands at approximately 110 and 200 kD, consistent with the expected molecular weights of a monomer and dimer, respectively. PLB was also detected in these higher molecular weight bands. We also quantified FRET from GFP‐SERCA to mCherry‐PLB using time correlated single photon counting to determine the GFP excited state lifetime. We observed a biexponential fluorescence decay consistent with two populations of donors with distinct lifetimes. The long lifetime was similar to that of a GFP‐alone control. The shorter lifetime was consistent with a high FRET population. Interestingly, the FRET efficiency of this species was 45%, approximately double the FRETmax value obtained from steady‐state FRET measurements. The population of the high FRET species approached, but did not exceed, 50% of the total, even at PLB to SERCA ratios as high as 40:1. This observation is consistent with a complex of two SERCA protomers in which only one of the pumps engages in FRET with a bound PLB.