Research in the Griffith lab

We are interested in elucidating the molecular and cellular processes that drive plastic behaviors. Our projects can be grouped into four categories: Sleep, Learning and Memory, Regulation of CaMKII and its Substrates, and Neuronal Homeostasis.

Research in the Griffith lab is constantly evolving, and we utilize a variety of techniques to address brain function at multiple levels: from single cell electrophysiology and functional imaging in reduced preparations to sleep, locomotor and learning assays in freely behaving animals. In all these approaches we enjoy the advantages of the amazing genetic tools available in Drosophila.


Sleep experiment setup on a shaker machine (for inducing sleep deprivation)

Sleep is a brain state common all animals, and is essential for mental and physical health. Although we typically spend 1/3 of our lives asleep, our understanding of this phenomenon is poor. In our lab, we aim to understand the molecular and cellular basis of sleep and elucidate the underlying circuitry regulating sleep/wake cycles and sleep homeostasis.

Regulation of sleep and sleep homeostasis by microRNAs

We are investigating how microRNAs (miRs) regulate sleep and sleep homeostasis (the ability to detect sleep loss and produce compensatory rebound sleep). MiRs are short, non-coding RNAs that bind to messenger RNAs and inhibit their translation into proteins. To identify miRs that play a role in sleep, we screened lines of flies that express inhibitors of specific miR function and assayed them for changes in sleep. From this screen, we have discovered roughly 30 miRs that affect sleep or sleep homeostasis. We are currently exploring when during the life cycle and where in the fly brain these miRs are required for their effects on sleep, and identifying which messenger RNAs they are targeting to have these effects.

Elucidating the circuitry controlling sleep

The ability of an animal to enter and maintain a sleep state is modulated by a variety of internal factors (e.g. hunger, arousal level, homeostatic drive and the circadian clock) and by external factors (e.g. temperature and light). Utilizing molecular-genetic tools that target and manipulate neural activity, we are interested in understanding the underlying circuitry controlling sleep.

To begin to understand the circuitry that allows the brain to sleep, we used genetic tools to alter the activity of candidate sleep circuit neurons. We found that, similar to mammals, GABAergic cells are involved in the wake-sleep transition (Agosto, et al., 2008). One of the targets of these sleep-promoting GABAergic neurons is a subset of clock cells known as the LNvs (Parisky, et al., 2008). LNvs are involved in setting morning activity and release a neuropeptide called Pigment Dispersing Factor (PDF). This peptide modulates activity of multiple cell types including clock cells and cells of the ellipsoid body, a motor control center. Mutant flies that do not make the peptide fall asleep more easily during the day. The strategy used by the fly brain to generate sleep thus appears to mirror that used by mammals. In both mammals and flies, GABAergic neurons induce sleep onset by inhibition of wake promoting neurons. In mammals, peptidergic wake promoting neurons which secrete orexin/hypocretin are critical for keeping the sleep switch stable--without them, the animal is unable to maintain a consolidated waking state. These neurons get input from the circadian clock. In flies, the modulatory wake-promoting peptidergic neurons are actually a subset of the clock neurons, but the function is similar: without the peptide, animals fall asleep more readily. Interestingly, the peptides PDF and orexin/hypocretin are NOT related--it is the strategy, not the molecule, that has been conserved. This is an important realization: we can get information about how sleep is generated by a circuit even without a direct molecular parallel.

In a broader screen of peptidergic neuron involvement in sleep (Shang et al., 2013) we found that sNPF, which is analogous to NPY in mammals, is a potent sleep-promoting peptide. One locus of action is on the small LNvs, though there are other brain areas that can respond to sNPF in a sleep-promoting manner. sNPF in the sleep system appears to couple to inhibitory G proteins and act as an inhibitory neuromodulator.

Learning and Memory

Mushroom bodies fluorescing under microscope in response to stimulus (40x magnification)
Setup for appetitive olfactory memory experiments

Interrogating functional connections in the memory system

The mushroom bodies have been shown by many groups to be critical for olfactory learning and memory, but the actual nature of the connections between mushroom body neurons and other memory circuit neurons is largely unknown. Using live functional imaging techniques, we are investigating the nature of connectivity within the mushroom body neuropil. The current focus of these studies is to ask how the DPM neurons interact with the mushroom bodies to consolidate memory. We recently found (Haynes et al., 2015) that the DPM neurons inhibit a subset of mushroom body neurons via GABA release onto GABAA receptors. We are now investigating what effect mushroom bodies have on the DPM neurons and the connectivity of dopaminergic neurons in the mushroom body neuropil.

Interactions between sleep and memory

Sleep and memory consolidation have been shown to be linked in humans, rodents and flies. Acute activation of DPMs promotes sleep, suggesting that experiences that are being encoded in long-term memory also cause a transient increase in sleep. The sleep-promoting effects of DPM activation are mediated by inhibition of wake-promoting α'/β' mushroom body neurons.

Elucidating the circuitry underlying learning and memory and how it is modulated by circadian rhythms

Circadian clock influence on cognition and performance have been documented in humans and other organisms, and are often apparent when subjects are jet-lagged. We are using associative olfactory conditioning paradigms and functional imaging to study the molecular and cellular basis of memory and how it is influenced by circadian rhythms.

Regulators and Substrates of CaMKII

Many neurological disorders such as Parkinson's disease and Huntington's disease are characterized by disruptive motor dysfunction, hyperexcitability (seizures), and in some cases, debilitating memory deficits. Our work in Drosophila melanogaster has helped identify calcium/calmodulin-dependent protein-dependent kinase II (CaMKII), its regulator, CASK, a MAGUK scaffolding protein, and its substrate, Eag, a voltage-gated potassium channel, as important to the maintenance of intrinsic membrane properties that underlie such memory networks and motor functions.

Eag: Regulation of excitability

Mutations in eag cause hyperexcitability and a failure to repolarize after repetitive stimulation. This phenotype is similar to that seen with inhibition of CaMKII with transgenic kinase inhibitors (Griffith et al., 1994). Eag can bind to CASK, and is phosphorylated by CaMKII. Eag is also capable of directly binding to activated CaMKII and locking it into a constitutively active state that persists even in low Ca2+ (Sun et al., 2004). This type of interaction provides a mechanism for perdurant regulation of a channel that outlasts an episode of neuronal activity. Eag also has CaM-binding domains (CaMBDs) on both its N- and C-termini, similar to hEag1 (Sun et al., 2004). The function of the N-terminal CaMBD is unknown. Expression of Eag with a mutant C-terminal CaMBD in HEK cells suggests that this is a low affinity inhibitory site. Ongoing work continues to investigate the cellular and behavioral effects of interactions between Eag, CaMKII, CASK and CaM.

CASK: Regulation of CaMKII

CASK is a MAGUK scaffolding protein with many ligands. In flies the CASK gene encodes two isoforms with separate transcriptional starts. CASK-β null flies have deficits in adult locomotion (Slawson et al., 2011) and sleep and place preference (Donelson et al., 2012). The locomotor defects map to dopaminergic neurons and we have shown that CASK regulates synaptic release via an interaction with Hsc4 (Slawson et al., 2015). CASK-β also directly binds to CaMKII and regulates its activity. In the presence of Ca2+/CaM, CaMKII complexed to CASK-β can autophosphorylate at T287 and become constitutively active. In the absence of Ca2+/CaM, ATP hydrolysis by CASK-bound CaMKII results in phosphorylation of T306, blocking subsequent activation of the kinase by Ca2+/CaM. Consistent with this, CASK-β overexpression suppresses CaMKII activity in transfected cells, and increases postsynaptic T306 phosphorylation. In animals lacking CASK, synapse-specific, activity-dependent autophosphorylation of CaMKII T287 is increased (Hodge et al., 2006). These results suggest that CASK-β, in the presence of Ca2+/CaM, provides a localized source of active kinase. When Ca2+/CaM or synaptic activity is low, CASK-β promotes inactivating autophosphorylation, producing CaMKII that requires phosphatase to reactivate. This interaction provides a mechanism by which the active postsynaptic pool of CaMKII can be controlled locally to differentiate active and inactive synapses.

Neuronal homeostasis in third instar larval motor neurons

Pre-synaptic motor neuron projections (red) and post-synaptic muscles (green) in the larval NMJ.

A human neuron can live over 100 years but components of a neuron constantly change on a timescale of seconds to weeks. In addition to this rapid turnover, connections between neurons within a circuit are plastic and change over the lifetime of an animal. How is a nervous system able to maintain stability and how are neurons able to maintain appropriate activity despite this changing activity? While not completely understood, there is strong evidence to suggest that calcium dependent sensors like the Ca2+/calmodulin-dependent protein kinases (CaMKs) are involved. CaMKs may be able to regulate both acute and long-term changes in homeostasis.

The third instar larval neuromuscular junction (NMJ) is a popular preparation that can be used to evaluate both pre- and post-synaptic changes. Combining the NMJ preparation with genetic tools that alter CaMK activity gives us a powerful way to investigate the role of CaMKs in sensing and changing neuronal activity. We performed sharp electrode recordings of the muscle and found that inhibition of CaMKII causes hyperexcitability (Wang et al., 1994, Griffith et al., 1994) while constitutively active CaMKII causes failures and delayed action potentials (Park et al., 2002). Because we are interested in how CaMKs may affect intrinsic properties of the motor neurons in response to a changing environment, we and others have developed the NMJ preparation for electrophysiological whole-cell patch clamp of motor neurons (Choi et al., 2004; Park and Griffith, 2006). Using dye-fills and electrophysiological recordings, we identified and characterized a cluster of motor neurons in the third instar larval ventral ganglion (Choi et al., 2004). Other work has focused on an afterhyperpolarization (AHP) of the 1st type of motor neuron (Pulver and Griffith, 2010). This AHP persists for several seconds forming a "memory" of spiking that holds the intrinsic properties of the motor neuron in the active state.

Current work is concerned with bringing together both the role of CaMKs in maintaining homeostasis and the effects of changing activity of the cell within a network. Using optogenetic tools we can artificially increase the activity of the motor neurons. We are using current clamp to measure the effects of this activity change. We hope to combine these activity tools with our genetic control of CaMKs to further investigate the role of CaMKs in homeostasis. In addition, we are interested in developing new tools to investigate CaMKI.