Connectomic reconstruction of a female Drosophila ventral nerve cord.

A deep understanding of how the brain controls behaviour requires mapping neural circuits down to the muscles that they control. Here, we apply automated tools to segment neurons and identify synapses in an electron microscopy dataset of an adult female Drosophila melanogaster ventral nerve cord (VNC)1, which functions like the vertebrate spinal cord to sense and control the body. We find that the fly VNC contains roughly 45 million synapses and 14,600 neuronal cell bodies. To interpret the output of the connectome, we mapped the muscle targets of leg and wing motor neurons using genetic driver lines2 and X-ray holographic nanotomography3. With this motor neuron atlas, we identified neural circuits that coordinate leg and wing movements during take-off. We provide the reconstruction of VNC circuits, the motor neuron atlas and tools for programmatic and interactive access as resources to support experimental and theoretical studies of how the nervous system controls behaviour.

Synaptic architecture of leg and wing premotor control networks in Drosophila.

Animal movement is controlled by motor neurons (MNs), which project out of the central nervous system to activate muscles1. MN activity is coordinated by complex premotor networks that facilitate the contribution of individual muscles to many different behaviours2-6. Here we use connectomics7 to analyse the wiring logic of premotor circuits controlling the Drosophila leg and wing. We find that both premotor networks cluster into modules that link MNs innervating muscles with related functions. Within most leg motor modules, the synaptic weights of each premotor neuron are proportional to the size of their target MNs, establishing a circuit basis for hierarchical MN recruitment. By contrast, wing premotor networks lack proportional synaptic connectivity, which may enable more flexible recruitment of wing steering muscles. Through comparison of the architecture of distinct motor control systems within the same animal, we identify common principles of premotor network organization and specializations that reflect the unique biomechanical constraints and evolutionary origins of leg and wing motor control.

Synaptic architecture of leg and wing motor control networks in Drosophila.

Animal movement is controlled by motor neurons (MNs), which project out of the central nervous system to activate muscles. Because individual muscles may be used in many different behaviors, MN activity must be flexibly coordinated by dedicated premotor circuitry, the organization of which remains largely unknown. Here, we use comprehensive reconstruction of neuron anatomy and synaptic connectivity from volumetric electron microscopy (i.e., connectomics) to analyze the wiring logic of motor circuits controlling the Drosophila leg and wing. We find that both leg and wing premotor networks are organized into modules that link MNs innervating muscles with related functions. However, the connectivity patterns within leg and wing motor modules are distinct. Leg premotor neurons exhibit proportional gradients of synaptic input onto MNs within each module, revealing a novel circuit basis for hierarchical MN recruitment. In comparison, wing premotor neurons lack proportional synaptic connectivity, which may allow muscles to be recruited in different combinations or with different relative timing. By comparing the architecture of distinct limb motor control systems within the same animal, we identify common principles of premotor network organization and specializations that reflect the unique biomechanical constraints and evolutionary origins of leg and wing motor control.

Reconstruction of motor control circuits in adult Drosophila using automated transmission electron microscopy.

To investigate circuit mechanisms underlying locomotor behavior, we used serial-section electron microscopy (EM) to acquire a synapse-resolution dataset containing the ventral nerve cord (VNC) of an adult female Drosophila melanogaster. To generate this dataset, we developed GridTape, a technology that combines automated serial-section collection with automated high-throughput transmission EM. Using this dataset, we studied neuronal networks that control leg and wing movements by reconstructing all 507 motor neurons that control the limbs. We show that a specific class of leg sensory neurons synapses directly onto motor neurons with the largest-caliber axons on both sides of the body, representing a unique pathway for fast limb control. We provide open access to the dataset and reconstructions registered to a standard atlas to permit matching of cells between EM and light microscopy data. We also provide GridTape instrumentation designs and software to make large-scale EM more accessible and affordable to the scientific community.

Dense neuronal reconstruction through X-ray holographic nano-tomography.

Imaging neuronal networks provides a foundation for understanding the nervous system, but resolving dense nanometer-scale structures over large volumes remains challenging for light microscopy (LM) and electron microscopy (EM). Here we show that X-ray holographic nano-tomography (XNH) can image millimeter-scale volumes with sub-100-nm resolution, enabling reconstruction of dense wiring in Drosophila melanogaster and mouse nervous tissue. We performed correlative XNH and EM to reconstruct hundreds of cortical pyramidal cells and show that more superficial cells receive stronger synaptic inhibition on their apical dendrites. By combining multiple XNH scans, we imaged an adult Drosophila leg with sufficient resolution to comprehensively catalog mechanosensory neurons and trace individual motor axons from muscles to the central nervous system. To accelerate neuronal reconstructions, we trained a convolutional neural network to automatically segment neurons from XNH volumes. Thus, XNH bridges a key gap between LM and EM, providing a new avenue for neural circuit discovery.

A size principle for recruitment of Drosophila leg motor neurons.

To move the body, the brain must precisely coordinate patterns of activity among diverse populations of motor neurons. Here, we use in vivo calcium imaging, electrophysiology, and behavior to understand how genetically-identified motor neurons control flexion of the fruit fly tibia. We find that leg motor neurons exhibit a coordinated gradient of anatomical, physiological, and functional properties. Large, fast motor neurons control high force, ballistic movements while small, slow motor neurons control low force, postural movements. Intermediate neurons fall between these two extremes. This hierarchical organization resembles the size principle, first proposed as a mechanism for establishing recruitment order among vertebrate motor neurons. Recordings in behaving flies confirmed that motor neurons are typically recruited in order from slow to fast. However, we also find that fast, intermediate, and slow motor neurons receive distinct proprioceptive feedback signals, suggesting that the size principle is not the only mechanism that dictates motor neuron recruitment. Overall, this work reveals the functional organization of the fly leg motor system and establishes Drosophila as a tractable system for investigating neural mechanisms of limb motor control.

In the body, spindly nerve cells called motor neurons connect the brain to the muscles. Their role is to control movement, as they translate the electrical signals from the brain into instructions to the muscles. In humans, it takes over 150,000 motor neurons to control the movement of one leg; in contrast, fruit flies only need 50 neurons to operate a leg, despite also executing a variety of movements. Fruit flies are commonly used in laboratories to study an array of biological processes, yet little is known about how their motor neurons direct movements. In particular, it was unclear whether the same principles that control how muscles contract in mammals also applied in the tiny fruit fly. To begin investigating, Azevedo et al. mapped out the arrangement of motor neurons that control muscles in the fruit fly leg. As the leg moved, the activity of both the neurons and the muscles they controlled was recorded, as well as the force that had been generated. The experiments showed that each motor neuron controls a certain range of leg force and speed: some produced small, slow motion important for posture and dexterity, while others created large, fast movements essential to running or escape. In addition, the neurons activate in a particular order ­ cells that control slow movements fire first, and those that direct fast maneuvers later. These processes are also found in other organisms, but the difference is that flies have so few neurons, allowing scientists to reliably identify each motor neuron. Future experiments will therefore be able to test how flies recruit the right neurons to create specific movement sequences. Fruit flies are often used to research human illnesses that affect movement, such as motor neuron disease. A better understanding of the way their neural circuits coordinate the body could help reveal how these conditions emerge.

Range, routing and kinetics of rod signaling in primate retina.

Stimulus- or context-dependent routing of neural signals through parallel pathways can permit flexible processing of diverse inputs. For example, work in mouse shows that rod photoreceptor signals are routed through several retinal pathways, each specialized for different light levels. This light-level-dependent routing of rod signals has been invoked to explain several human perceptual results, but it has not been tested in primate retina. Here, we show, surprisingly, that rod signals traverse the primate retina almost exclusively through a single pathway - the dedicated rod bipolar pathway. Identical experiments in mouse and primate reveal substantial differences in how rod signals traverse the retina. These results require reevaluating human perceptual results in terms of flexible computation within this single pathway. This includes a prominent speeding of rod signals with light level - which we show is inherited directly from the rod photoreceptors themselves rather than from different pathways with distinct kinetics.

Active Mechanisms of Vibration Encoding and Frequency Filtering in Central Mechanosensory Neurons.

To better understand biophysical mechanisms of mechanosensory processing, we investigated two cell types in the Drosophila brain (A2 and B1 cells) that are postsynaptic to antennal vibration receptors. A2 cells receive excitatory synaptic currents in response to both directions of movement: thus, twice per vibration cycle. The membrane acts as a low-pass filter, so that voltage and spiking mainly track the vibration envelope rather than individual cycles. By contrast, B1 cells are excited by only forward or backward movement, meaning they are sensitive to vibration phase. They receive oscillatory synaptic currents at the stimulus frequency, and they bandpass filter these inputs to favor specific frequencies. Different cells prefer different frequencies, due to differences in their voltage-gated conductances. Both Na+ and K+ conductances suppress low-frequency synaptic inputs, so cells with larger voltage-gated conductances prefer higher frequencies. These results illustrate how membrane properties and voltage-gated conductances can extract distinct stimulus features into parallel channels.

C-terminal threonines and serines play distinct roles in the desensitization of rhodopsin, a G protein-coupled receptor.

Rod photoreceptors generate measurable responses to single-photon activation of individual molecules of the G protein-coupled receptor (GPCR), rhodopsin. Timely rhodopsin desensitization depends on phosphorylation and arrestin binding, which quenches G protein activation. Rhodopsin phosphorylation has been measured biochemically at C-terminal serine residues, suggesting that these residues are critical for producing fast, low-noise responses. The role of native threonine residues is unclear. We compared single-photon responses from rhodopsin lacking native serine or threonine phosphorylation sites. Contrary to expectation, serine-only rhodopsin generated prolonged step-like single-photon responses that terminated abruptly and randomly, whereas threonine-only rhodopsin generated responses that were only modestly slower than normal. We show that the step-like responses of serine-only rhodopsin reflect slow and stochastic arrestin binding. Thus, threonine sites play a privileged role in promoting timely arrestin binding and rhodopsin desensitization. Similar coordination of phosphorylation and arrestin binding may more generally permit tight control of the duration of GPCR activity.

Experimental protocols alter phototransduction: the implications for retinal processing at visual threshold.

Vision in dim light, when photons are scarce, requires reliable signaling of the arrival of single photons. Rod photoreceptors accomplish this task through the use of a G-protein-coupled transduction cascade that amplifies the activity of single active rhodopsin molecules. This process is one of the best understood signaling cascades in biology, yet quantitative measurements of the amplitude and kinetics of the rod's response in mice vary by a factor of ∼ 2 across studies. What accounts for these discrepancies? We used several experimental approaches to reconcile differences in published properties of rod responses. First, we used suction electrode recordings from single rods to compare measurements across a range of recording conditions. Second, we compared measurements of single-cell photocurrents to estimates of rod function from in vitro electroretinograms. Third, we assayed the health of the post-receptor retinal tissue in these different conditions. Several salient points emerge from these experiments: (1) recorded responses can be altered dramatically by how the retina is stored; (2) the kinetics of the recovery of responses to bright but not dim flashes are strongly sensitive to the extracellular concentration of magnesium; (3) experimental conditions that produce very different single-photon responses measured in single rods produce near identical derived rod responses from the electroretinogram. The dependence of rod responses on experimental conditions will be a key consideration in efforts to extract general principles of G-protein signaling from studies of phototransduction and to relate these signals to downstream mechanisms that facilitate visual sensitivity.

Arrestin competition influences the kinetics and variability of the single-photon responses of mammalian rod photoreceptors.

Reliable signal transduction via G-protein-coupled receptors requires proper receptor inactivation. For example, signals originating from single rhodopsin molecules vary little from one to the next, requiring reproducible inactivation of rhodopsin by phosphorylation and arrestin binding. We determined how reduced concentrations of rhodopsin kinase (GRK1) and/or arrestin1 influenced the kinetics and variability of the single-photon responses of mouse rod photoreceptors. These experiments revealed that arrestin, in addition to its role in quenching the activity of rhodopsin, can tune the kinetics of rhodopsin phosphorylation by competing with GRK1. This competition influenced the variability of the active lifetime of rhodopsin. Biasing the competition in favor of GRK1 revealed that rhodopsin remained active through much of the single-photon response under the conditions of our experiments. This long-lasting rhodopsin activity can explain the characteristic time course of single-photon response variability. Indeed, explaining the late time-to-peak of the variance required an active lifetime of rhodopsin approximately twice that of the G-protein transducin. Competition between arrestins and kinases may be a general means of influencing signals mediated by G-protein-coupled receptors, particularly when activation of a few receptors produces signals of functional importance.