Search

Monitoring changes in molecular conformation is essential for studying protein interactions within and between complexes. Current methods such as FRET or FCS can only reveal limited aspects of these changes, which in many cases is insufficient for interpretation. Fluorescent Polarization and Fluctuation Analysis (FPFA), a time-correlated single-photon counting technique that combines homo-FRET and FCS was developed to address this problem. In FPFA, changes in mass and/or shape of a protein complex, the number of fluorescent subunits per complex, as well as subunit proximity (1 - 10 nm), are simultaneously detected. Here we have extended FPFA by incorporating simultaneous hetero-FRET analysis so that additional structural changes can be monitored concurrently. This multi-modal approach was then used to observe changes in CaMKII holoenzyme in response to activation and subsequent interactions with NR2B. Revised FPFA revealed that catalytic-domain pairing is the fundamental structural motif of the holoenzyme organization. A genetically engineered dimeric-CaMKII functioned like native holoenzyme (8-14 subunits) in terms of activity, T286 autophosphorylation, structural changes triggered by calcium-calmodulin, and those associated with interaction with NR2B.

Calcium/Calmodulin (CaM) dependent protein kinase (CaMKII) is an important regulator of learning and memory, and it is thought to play a key role in sensing neuronal calcium oscillations. In cells, CaMKIIα primarily forms dodecamers comprised of two stacked rings of 6 subunits, each of which has 3 domains: an N-terminal catalytic domain, a C-terminal oligomerization domain, and between them, a regulatory domain. This regulatory domain activates the enzyme upon binding CaM and harbors a T286 autophosphorylation site that is associated with calcium independent persistent activation. To form the holoenzyme, the oligomerization domain of each subunit interacts with three other oligomerization domains: two laterally to form a ring of association domains, and one transversely to stabilize the double ring structure. Projecting outwards from the rings are catalytic domains, which form pairs, as indicated by homo-FRET analysis. Assemblies comprised of 8-14 subunits are also thought to exist, but it is not known if the number of subunits in the holoenzyme alters the activity or behavior of the catalytic domain pairs. To study the impact of subunit stoichiometry on catalytic domain pairing and T286 autophosphorylation, we generated mutants by altering oligomerization domain residues at the lateral interaction sites. Using fluorescence polarization and fluctuation analysis (FPFA) we simultaneously monitored homo-FRET between Venus-tagged catalytic domains, lateral diffusion of the holoenzyme, and the number of subunits present in these altered assemblies. We find that when the holoenzyme has fewer than 8 subunits, the holoenzyme diffusion time changes and homo-FRET between catalytic domains begins to attenuate. Finally, T286 autophosphorylation in a dimeric mutant holoenzyme is approximately half of that observed in the native holoenzyme. The structural implication of these changes will be discussed.

FRET is a phenomenon where excited state energy of a fluorescent donor molecule is transferred, by a non-radiative dipole-dipole coupling mechanism, to a nearby acceptor. Because the efficiency of FRET provides information regarding the distance separating the donor and acceptor, FRET is used to study protein interactions in living cells. In conjunction with automated microscopy, FRET can also be used to screen for drugs that perturb specific protein-protein interactions. The utility of FRET, however, is limited by the possibility of both false positive (caused by over expression and molecular crowding), and by false negative signals (resulting from separation distances greater than 10 nm and from very low dipole orientation factors). To overcome these limitations we have developed Polarized Fluorescence Correlation Spectroscopy (pFCS), a single-molecule based method to characterize the interactions of proteins in complexes. Using pFCS, Homo-FRET (a 1-10 nm proximity gauge), brightness (a measure of the number of fluorescent subunits in a complex), and the lateral diffusion coefficient (an attribute sensitive to viscosity, mass, and the shape of a protein complex) can be simultaneously measured. With these measurements, the interpretation of FRET can be rigorously constrained thus reducing the likelihood of both false negative and false positive interpretations. Standards consisting of tandem covalently linked concatemers of between 1 and 6 Venus molecules were used to validate pFCS in both solution and in cells. The utility of pFCS was demonstrated by measuring differences in Homo-FRET, subunit stoichiometry and correlation time between Venus-tagged CaM-kinase-II holoenzyme (α or β).

The autoinhibited calcium/calmodulin-dependent protein kinases-II (CaMKII) assembles to form a holoenzyme structure with multiple catalytic domain pairs radiating out from a central core composed of oligomerization domains. Activation of CaMKII upon binding calcium2+/calmodulin (CaM) is known to initiate a biochemical cascade resulting in CaMKII translocation to synaptic spines and the induction of long-term potentiation. Little is known about the structural changes in the intact CaMKII holoenzyme responsible for this activation. Fluorescence polarization and fluctuation analysis (FPFA) microscopy, a method that combines time-resolved anisotropy and fluorescence correlation spectroscopy (FCS), was used to monitor Venus-CaMKIIα (where a Venus fluorescent-protein is tagged onto the N-terminus of CaMKIIα) before and after activation with CaM. FPFA simultaneously monitors brightness (a measure of the number of subunits in the holoenzyme), homo-FRET (a catalytic domain proximity detector) and diffusion time (a parameter sensitive to the mass and hydrodynamic volume of the holoenzyme). Together these measurements can potentially detect an array of different covert structural changes in the CaMKII holoenzyme triggered by CaM. In homogenates prepared from HEK cells expressing Venus-CaMKIIα, the addition of CaM triggered T286 autophosphorylation, an indicator of kinase activation. FPFA revealed a large increase in the holoenzyme diffusion time. Because changes in holoenzyme subunit stoichiometry were not observed, the change in diffusion time most likely indicates a large change in hydrodynamic volume, such as catalytic domain pair extension. This change was not observed in Venus-CaMKIIα-TT305/6DD, a mutant that cannot bind calmodulin. Holoenzyme activation by CaM did trigger a small, dose-dependent biphasic change in anisotropy (homo-FRET between Venus-tagged catalytic domains), but its magnitude was too small to be interpreted as catalytic domain un-pairing. Interpretation of these experiments will be discussed.

The calcium-calmodulin (CaM) dependent protein Kinase-II (CaMKII) regulates various cellular process and has been implicated in illness including schizophrenia, heart failure, arthritis, and even certain types of cancers. CaMKII subunits assemble to form a large dodecameric holoenzyme. The structure of a kinase subunit is organized into 3 domains, a catalytic domain, a regulatory domain, and an oligomerization domain. Each N-terminal catalytic domain harbors the kinase catalytic site, its ATP binding site, as well as the so-called ‘T-site’, thought to interact with the T286 autophosphorylation site in the regulatory domain. The T-site is responsible for mediating the interaction of the holoenzyme with other proteins, such as NMDA receptors, in response to activation. The catalytic domain is attached to an autoinhibitory regulatory domain that contains the T286 autophosphorylation site, and a CaM-binding site. In turn: the regulatory domain is itself tethered to a C-terminal oligomerization domain that organizes the enzyme into ring-shaped oligomers. We have previously shown in living cells using FRET analysis that catalytic domains form pairs in the intact auto-inhibited holoenzyme. In Hippocampal neurons, these pairs separate in response to calcium influx through NMDA-receptors, but it is not known if this is a direct response to CaM binding, or a down stream affect of kinase activation. Here using fluorescence polarization and fluctuation analysis (FPFA), a hybrid method that simultaneously measures homo-FRET and fluorescence correlation spectroscopy (FCS), we show that CaMKIINtide, a T-site ligand, disrupts catalytic domain pairing in the intact holoenzyme. This reaction is dose dependent but was only observed upon co-incubation with CaM to activate the enzyme. The holoenzyme can be activated without catalytic domain pair separation. We also show that T-site specific mutations at I205, stabilize catalytic domain pairing and counteract CaMKIINtide's ability to disrupt catalytic domain pairing.