Pabak Sarkar

Fluorescence Polarization and Fluctuation Analysis Monitors Subunit Proximity, Stoichiometry, and Protein Complex Hydrodynamics

Förster resonance energy transfer (FRET) microscopy is frequently used to study protein interactions and conformational changes in living cells. The utility of FRET is limited by false positive and negative signals. To overcome these limitations we have developed Fluorescence Polarization and Fluctuation Analysis (FPFA), a hybrid single-molecule based method combining time-resolved fluorescence anisotropy (homo-FRET) and fluorescence correlation spectroscopy. Using FPFA, homo-FRET (a 1-10 nm proximity gauge), brightness (a measure of the number of fluorescent subunits in a complex), and correlation time (an attribute sensitive to the mass and shape of a protein complex) can be simultaneously measured. These measurements together rigorously constrain the interpretation of FRET signals. Venus based control-constructs were used to validate FPFA. The utility of FPFA was demonstrated by measuring in living cells the number of subunits in the α-isoform of Venus-tagged calcium-calmodulin dependent protein kinase-II (CaMKIIα) holoenzyme. Brightness analysis revealed that the holoenzyme has, on average, 11.9 ± 1.2 subunit, but values ranged from 10-14 in individual cells. Homo-FRET analysis simultaneously detected that catalytic domains were arranged as dimers in the dodecameric holoenzyme, and this paired organization was confirmed by quantitative hetero-FRET analysis. In freshly prepared cell homogenates FPFA detected only 10.2 ± 1.3 subunits in the holoenzyme with values ranging from 9-12. Despite the reduction in subunit number, catalytic domains were still arranged as pairs in homogenates. Thus, FPFA suggests that while the absolute number of subunits in an auto-inhibited holoenzyme might vary from cell to cell, the organization of catalytic domains into pairs is preserved.

Zinc-Induced Polymerization of Killer-Cell Ig-like Receptor into Filaments Promotes Its Inhibitory Function at Cytotoxic Immunological Synapses

The inhibitory function of killer cell immunoglobulin-like receptors (KIR) that bind HLA-C and block activation of human natural killer (NK) cells is dependent on zinc. We report that zinc induced the assembly of soluble KIR into filamentous polymers, as detected by electron microscopy, which depolymerized after zinc chelation. Similar KIR filaments were isolated from lysates of cells treated with zinc, and membrane protrusions enriched in zinc were detected on whole cells by scanning electron microscopy and imaging mass spectrometry. Two independent mutations in the extracellular domain of KIR, away from the HLA-C binding site, impaired zinc-driven polymerization and inhibitory function. KIR filaments formed spontaneously, without the addition of zinc, at functional inhibitory immunological synapses of NK cells with HLA-C(+) cells. Adding to the recent paradigm of signal transduction through higher order molecular assemblies, zinc-induced polymerization of inhibitory KIR represents an unusual mode of signaling by a receptor at the cell surface.

Covert Changes in CaMKII Holoenzyme Structure Identified for Activation and Subsequent Interactions.

Between 8 to 14 calcium-calmodulin (Ca(2+)/CaM) dependent protein kinase-II (CaMKII) subunits form a complex that modulates synaptic activity. In living cells, the autoinhibited holoenzyme is organized as catalytic-domain pairs distributed around a central oligomerization-domain core. The functional significance of catalytic-domain pairing is not known. In a provocative model, catalytic-domain pairing was hypothesized to prevent ATP access to catalytic sites. If correct, kinase-activity would require catalytic-domain pair separation. Simultaneous homo-FRET and fluorescence correlation spectroscopy was used to detect structural changes correlated with kinase activation under physiological conditions. Saturating Ca(2+)/CaM triggered Threonine-286 autophosphorylation and a large increase in CaMKII holoenzyme hydrodynamic volume without any appreciable change in catalytic-domain pair proximity or subunit stoichiometry. An alternative hypothesis is that two appropriately positioned Threonine-286 interaction-sites (T-sites), each located on the catalytic-domain of a pair, are required for holoenzyme interactions with target proteins. Addition of a T-site ligand, in the presence of Ca(2+)/CaM, elicited a large decrease in catalytic-domain homo-FRET, which was blocked by mutating the T-site (I205K). Apparently catalytic-domain pairing is altered to allow T-site interactions.

A Single Amino Acid Change in Inhibitory Killer Cell Ig-like Receptor Results in Constitutive Receptor Self-Association and Phosphorylation

Signaling by immunoreceptors is often initiated by phosphorylation of cytosolic tyrosines, which then recruit effector molecules. In the case of MHC class I–specific inhibitory receptors, phosphorylation of cytosolic tyrosine residues within ITIMs results in recruitment of a protein tyrosine phosphatase that blocks activation signals. Recent work showed that signaling by an HLA-C–specific killer cell Ig-like receptor (KIR) is independent of signaling by activation receptors. It is not known how ITIM phosphorylation is initiated and regulated. In this article, we show that substitution of His-36 in the first Ig domain of KIR2DL1 with alanine (KIR2DL1-H36A) resulted in constitutive KIR2DL1 self-association and phosphorylation, as well as recruitment of tyrosine phosphatase SHP-1. Furthermore, substitution of His-36 with a similar bulky amino acid, phenylalanine, maintained the receptor in its unphosphorylated state, suggesting that steric hindrance by the His-36 side chain prevents constitutive KIR2DL1 self-association and ITIM phosphorylation. The equally strong phosphorylation of KIR2DL1 and KIR2DL1-H36A after inhibition of tyrosine phosphatase by pervanadate suggested that KIR2DL1-H36A is selectively protected from dephosphorylation. We propose that KIR phosphorylation is controlled by the accessibility of ITIM to tyrosine phosphatases and that KIR binding to HLA-C must override the hindrance that His-36 puts on KIR2DL1 self-association. Expression of KIR2DL1-H36A on NK cells led to stronger inhibition of lysis of HLA-C+ target cells than did expression of wild-type KIR2DL1. These results revealed that ITIM phosphorylation is controlled by self-association of KIR and that His-36 serves as a gatekeeper to prevent unregulated signaling through KIR2DL1.

Deciphering CaMKII Multimerization Using Fluorescence Correlation Spectroscopy and Homo-FRET Analysis

While kinases are typically composed of one or two subunits, calcium-calmodulin (CaM)-dependent protein kinase II (CaMKII) is composed of 8-14 subunits arranged as pairs around a central core. It is not clear if the CaMKII holoenzyme functions as an assembly of independent subunits, as catalytic pairs, or as a single unit. One strategy to address this question is to genetically engineer monomeric and dimeric CaMKII and evaluate how their activity compares to the wild-type (WT) holoenzyme. Here a technique that combines fluorescence correlation spectroscopy and homo-FRET analysis was used to characterize assembly mutants of Venus-tagged CaMKIIα to identify a dimeric CaMKII. Spectroscopy was then used to compare how holoenzyme structure and function changes in response to activation with CaM in the dimeric mutant, WT-holoenzyme, and a monomeric CaMKII oligomerization-domain deletion mutant control. CaM triggered an increase in hydrodynamic volume in both WT and dimeric CaMKII without altering subunit stoichiometry or the net homo-FRET between Venus-tagged catalytic domains. Biochemical analysis revealed that the dimeric mutant also functioned like WT holoenzyme in terms of its kinase activity with an exogenous substrate, and for endogenous T286 autophosphorylation. We conclude that the fundamental functional units of CaMKII holoenzyme are paired catalytic-domains.

Positive Cooperativity and T286 Autophosphorylation is Observed in a Dimeric Mutant of Calcium-Calmodulin Dependent Protein Kinase II (CaMKII)

CaMKII is a multimeric enzyme that regulates long-term potentiation in the hippocampus. The assembly and organization of 8-14 subunits in a holoenzyme is thought to be required for transduction of calcium spike frequencies, cooperative binding to calmodulin, persistent activation by T286 autophosphorylation, and translocation into synaptic spines. CaMKII assembles into a holoenzyme by virtue of its unique C-terminal association domain (AD). ADs form a central hub-like structure from which regulatory and catalytic domains project. Each AD interacts tightly with three other subunits, two laterally and one transversely, to form a stable core composed of two stacked rings. FRET and analytical centrifugation has indicated that individual catalytic domains can form pairs by virtue of low affinity binding sites. But catalytic domains of the CaMKII mutant lacking AD do not form pairs in physiological conditions. We hypothesize that catalytic domain pairing is the fundamental structure that allows positive cooperativity and T286 autophosphorylation. Furthermore we postulate that high affinity AD core interactions lead to a high local concentration of catalytic domains that allows low-affinity catalytic domain pairing. To test this hypothesis, we mutated the lateral surface of the CaMKIIα AD. When expressed in HEK293 cells, the mutation generated a paired dimeric CaMKIIα. In support of the hypothesis, our study shows that the dimeric enzyme has the same affinity for calmodulin, a similar Hill coefficient for enzymatic activity, and T286 autophosphorylation comparable to that observed in the wild type holoenzyme. In contrast, the monomeric mutant CaMKII lacking AD has lower affinity, no cooperativity, and reduced T286 autophosphorylation. While catalytic domain pairing was observed in the dimeric mutant, biophysical analysis suggests that the nature of the pairing might be altered.