The overall goal of my laboratory is to gain a better understanding of signaling inside individual living cells, in neuronal networks, and in tumors. We design, synthesize, and use new molecules that detect or manipulate biochemical signals.
Genetically Targetable Fluorescent Indicators for Ca2+
Intracellular Ca2+ regulates numerous proteins and cellular functions and can vary tremendously over submicron and submillisecond scales, so precisely localized fast detection is desirable. Stephen Adams created a ~1-kDa biarsenical Ca2+ indicator, Calcium Green FlAsH (CaGF), to probe Ca2+ concentration surrounding genetically targeted proteins. CaGF attached to a tetracysteine motif becomes 10-fold more fluorescent upon binding Ca2+, with a dissociation constant of ~100 µM, <1 ms kinetics, and good Mg2+ rejection. Oded Tour showed that in HeLa cells expressing tetracysteine-tagged connexin 43, CaGF labels gap junctions and reports Ca2+ waves after injury. Total internal reflection microscopy of tetracysteine-tagged, CaGF-labeled L-type calcium channels shows fast-rising, depolarization-evoked Ca2+ transients, whose lateral nonuniformity suggests that the probability of channel opening varies greatly over micron dimensions. With moderate Ca2+ buffering, these transients decay surprisingly slowly, probably because most of the CaGF signal comes from closed channels feeling Ca2+ from a tiny minority of clustered open channels. With high Ca2+ buffering, CaGF signals decay as rapidly as the calcium currents, as expected for submicron Ca2+ domains immediately surrounding active channels. Thus CaGF can report highly localized, rapid Ca2+ concentration dynamics.
Genetically Encoded Indicators of Glutamate
The amino acid glutamate is the most important excitatory neurotransmitter in the mammalian central nervous system, so a genetically targetable indicator for glutamate should be of major usefulness in neurobiology. Andrew Hires has built and optimized such indicators, in which a bacterial glutamate-binding protein is genetically sandwiched between cyan and yellow fluorescent proteins. Glutamate binding to the bacterial receptor decreases fluorescence resonance energy transfer (FRET) between the fluorescent proteins. Hires has made quantitative optical measurements of the time course of synaptic glutamate release, spillover, and reuptake in cultured hippocampal neurons with near-millisecond temporal resolution or spine-sized spatial resolution. Enough spillover occurs from synapses onto extrasynaptic areas to cause significant activation of some receptors in the latter regions. Stimulation frequency–dependent modulation of spillover magnitude suggests a mechanism for nonsynaptic neuronal communication.
Detection of Protein-Protein Proximities over Tens of Nanometers
FRET is a popular and powerful method for monitoring the dynamics of protein conformations and interactions in intact cells, but FRET is limited to distances up to about 8 nm, which is not enough to span large proteins or complexes. FRET is also strongly dependent on the relative orientations of the donor and acceptor chromophores, so that protein-protein interactions can be missed if the orientations are unlucky. Diffusion of singlet oxygen (1O2) from a photosensitizer to an acceptor sensor can detect proximities up to 70–100 nm without any orientation dependence, but until now, the only available 1O2 sensors have been polymer beads of 250-nm diameter. Colette Dooley has developed green fluorescent protein (GFP) mutants that react rapidly, specifically, and irreversibly with 1O2 to produce ratiometric changes in excitation spectra. We have demonstrated detection of protein-protein interaction over a 25-nm distance, well beyond the range of FRET, using connexin-43 gap junctions as scaffolds to separate ReAsH, a genetically targetable 1O2 generator in the cytosol of one cell, from the 1O2-sensing GFP in the cytosol of a neighboring cell.
Turnover of Synaptic Protein Markers
Because many forms of long-lasting learning and memory involve growth or de novo formation of synapses, visualizing recently expanded or created synapses may indicate where memories are stored. Time-lapse microscopy can track synaptic growth in sparsely labeled, superficial neurons, but identifying growing synapses throughout a brain is currently not possible. A plausible strategy may be to selectively image newly synthesized proteins that preferentially accumulate in growing synapses. Varda Lev-Ram is imaging newly synthesized copies of synapse-specific cell adhesion proteins such as neuroligin, neurexin, and SynCAM, each tagged with either a tetracysteine motif or Dendra2. The tetracysteine motif is labeled in a pulse-chase protocol with the biarsenical dyes ReAsH followed by FlAsH. Dendra2, a photoconvertible fluorescent protein, irreversibly changes from green to red upon illumination with violet light. By either protocol, old protein molecules glow red and new copies glow green, which allows us to determine the dynamics of synaptic adhesion molecules and the appearance of new synapses. Lev-Ram's most striking current results are that repeated depolarizations of cultured cerebellar neurons with pulses of high K+ increase turnover of SynCAM within 12 hours, while glutamate pulses induce a slower onset of turnover but are more prominent than high K+ by 48 hours. These changes depend on new protein synthesis and the presence of extracellular Ca2+ during stimulation.
The FlAsH/ReAsH and photoconversion techniques are currently incompatible with deep tissues or freely behaving animals and have some toxicity and sensitivity limitations. Michael Lin has developed TimeSTAMP (time-specific tagging for age measurement of proteins), a new method in which a protein is fused via hepatitis C viral protease to an epitope tag. The protease constitutively removes itself and the accompanying tag until a small-molecule inhibitor is administered; subsequently all newly synthesized fusion proteins remain epitope-tagged. TimeSTAMP shows that new synapses in cultured hippocampal neurons preferentially contain new copies of a 95-kDa postsynaptic density protein (PSD95). In intact flies, TimeSTAMP reveals patterns of new synthesis of calmodulin-dependent protein kinase II (CaMKII) distinct from total CaMKII localization. TimeSTAMP should allow retrospective identification of new synapses and their key proteins anywhere in the nervous systems of freely behaving animals.
In Vivo Imaging of Protease Activity
The imaging reagents described above are all optical and genetically targeted, but for imaging disease processes in patients, one would prefer contrast mechanisms allowing nonoptical imaging in opaque tissues without requiring gene transfer. Tao Jiang has discovered such a general mechanism based on polycationic peptides already known to mediate cellular uptake of a wide variety of radioactive, magnetic, or optical contrast agents. He discovered that such uptake can be vetoed by certain tandemly fused polyanionic sequences and restored by cleavage of the linker between the cationic and anionic segments. Extracellular proteases such as matrix metalloproteinases can perform such cleavage and are known to be ubiquitously essential for tumor metastasis. This new molecular contrast mechanism incorporates enzymatic amplification, unlike previous mechanisms based on 1:1 binding of antibodies, aptamers, or receptor ligands to their targets. The new enzyme substrates work on a range of cancer models, including three-dimensional cultures of mammary carcinoma cells, xenografts of human tumors, and transgenic mice that spontaneously develop mammary tumors, as shown by Todd Aguilera, Emilia Olson, and Quyen Nguyen. Attachment of the substrates to macromolecular scaffolds of 50–70 kDa has improved contrast for tumors relative to liver and kidney, which are the normal organs showing the highest background uptake. Substrates labeled with a far-red dye or gadolinium can detect primary tumors and metastatic lymph nodes by, respectively, far-red fluorescence or magnetic resonance imaging. Our initial target for clinical application is likely to be intraoperative fluorescence guidance to help surgeons see tumor borders and residual malignancy.
Grants from the National Institutes of Health and the Breast Cancer Research Program provided support for the work of Adams, Aguilera, Dooley, Hires, Jiang, Lev-Ram, Lin, Nguyen, Olson, and Tour in the Tsien group. Lin also has a Jane Coffin Childs fellowship.
As of August 20, 2007