Detecting stoichiometry of macromolecular complexes in live cells using FRET

Abstract

The stoichiometry of macromolecular interactions is fundamental to cellular signalling yet challenging to detect from living cells. Fluorescence resonance energy transfer (FRET) is a powerful phenomenon for characterizing close-range interactions whereby a donor fluorophore transfers energy to a closely juxtaposed acceptor. Recognizing that FRET measured from the acceptor's perspective reports a related but distinct quantity versus the donor, we utilize the ratiometric comparison of the two to obtain the stoichiometry of a complex. Applying this principle to the long-standing controversy of calmodulin binding to ion channels, we find a surprising Ca²⁺-induced switch in calmodulin stoichiometry with Ca²⁺ channels-one calmodulin binds at basal cytosolic Ca²⁺ levels while two calmodulins interact following Ca²⁺ elevation. This feature is curiously absent for the related Na channels, also potently regulated by calmodulin. Overall, our assay adds to a burgeoning toolkit to pursue quantitative biochemistry of dynamic signalling complexes in living cells.

Figure 1. Theoretical scheme for deducing stoichiometry from FRET efficiencies.

(a) Left, cartoon illustrates 1:1 stoichiometry interaction of two binding partners tagged with donor and acceptor fluorophores. Right, conceptual scheme illustrates FRET between the excited donor (D*) and the bound acceptor (A*). E₁₁, the probability of energy transfer. (b) Both 3³-FRET and E-FRET methods report maximal FRET efficiency as E₁₁ yielding stoichiometry ratio ν=1. (c) Left, cartoon illustrates a multimeric complex containing n_D donor and n_A acceptor molecules. The probability of energy transfer from ith donor to jth acceptor is denoted E_ij. (d) For multimeric complexes, the maximal 3³-FRET (E_A,max) and E-FRET (E_D,max) may differ. The stoichiometry ratio ν=E_A,max/E_D,max reports the ratio of number of donors to acceptors in the complex (n_D/n_A).

Figure 2. Experimental validation for FRET-based stoichiometry assay.

(a) Bars depict average 3³-FRET (black) and E-FRET (red) efficiencies for various ECFP-EYFP concatemers as described in cartoon below (mean±s.e.m.; n, number of cells, as labelled for each bar). Note that the average 3³-FRET and E-FRET efficiencies are equal for ECFP-EYFP dimers but different for multimers. (b) The experimentally determined stoichiometry ratio, ν=E_A,max/E_D,max for each concatemer (grey symbol; mean±s.e.m.) follows the identity relation (black fit) with expected number of donor and acceptor molecules in the complex (n_D/n_A) confirming the theoretical relation equation 3.

Figure 3. Stoichiometry of calmodulin interaction with myosin Va neck domain.

(a) Schematic illustrates FRET binding pairs ECFP tagged CaM and EYFP tagged myosin Va peptide containing a single IQ domain (IQ₆). (b) Left, 3³-FRET efficiency (E_A) is plotted against estimated free donor concentration (D_free). Each black symbol corresponds to data from a single cell. Right, E-FRET efficiency (E_D) is plotted as a function of estimated free acceptor concentration (A_free). The maximal 3³-FRET efficiency (E_A,max) is approximately equal to maximal E-FRET efficiency (E_D,max). (c) Cartoon illustrates FRET pairs ECFP-CaM and EYFP-tagged full length myosin Va neck domain peptide containing six IQ domains (IQ_1-6). (d) Left, 3³-FRET efficiency as a function of free donor concentration (D_free). Right, E-FRET efficiency is plotted against free acceptor concentration (A_free). In both cases, each symbol corresponds to FRET measurement from a single cell. Notice that E_A,max is ∼6-fold larger than E_D,max suggesting 6:1 donor:acceptor stoichiometry. (e) Bar-graph summary depicts maximal 3³-FRET (black) and E-FRET (red) efficiencies for binding of ECFP-CaM with various YFP-tagged truncations of myosin Va neck domain containing varying number of IQ domains as shown in cartoon below (mean±s.e.m.; n, number of cells, as labelled for each bar). (f) Experimentally determined FRET-based stoichiometry ratio (ν) follow the identity relation with the number of IQ domains within each truncation shown in e. Each symbol, mean±s.e.m.

Figure 4. Stoichiometry of calmodulin binding to Ca_V and Na_V channels.

(a) Cartoon illustrates FRET pairs ECFP-CaM and Ca_V1.2 holochannel tagged with EYFP on its carboxy-terminus. The Ca_V channel auxiliary subunits β_2A and α₂δ subunits are coexpressed. (b) Bar-graph summary of maximal 3³-FRET (E_A,max; black) and E-FRET (E_D,max; red) efficiencies under basal (– Ca²⁺) and elevated Ca²⁺ conditions for CaM binding to Ca_V1.2 (mean±s.e.m.; n, number of cells, as labelled for each bar). E_A,max and E_D,max are approximately equal under resting Ca²⁺ conditions. With high cytosolic Ca²⁺ levels, E_A,max∼2 × higher than E_D,max. (c) Computing stoichiometry ratio (ν) shows that a single CaM binds to the holo-Ca²⁺ channels under low Ca²⁺ conditions while two CaM interact with the channel complex upon Ca²⁺ elevation (mean±s.e.m.). (d) Cartoon depicts FRET pairs ECFP-CaM and Na_V1.4 with EYFP fused to its carboxy-terminus. (e) Bar-graph summarizes maximal 3³-FRET (E_A,max; black) and E-FRET (E_D,max; red) efficiencies for CaM binding to Na_V1.4 (mean±s.e.m.; n, number of cells, as labelled for each bar). Notice that maximal 3³-FRET and E-FRET efficiencies are approximately equal to each when measured under both basal and elevated Ca²⁺ conditions. (f) Experimentally determined stoichiometry ratio (ν) shows that a single CaM interacts with Na_V channel complex under all Ca²⁺ conditions. Format as in c.