Imaging Structural and Functional Dynamics in Xenopus Neurons

Abstract

In vivo time-lapse imaging has been a fruitful approach to identify structural and functional changes in the Xenopus nervous system in tadpoles and adult frogs. Structural imaging studies have identified fundamental aspects of brain connectivity, development, plasticity and disease and been instrumental in elucidating mechanisms regulating these events in vivo. Similarly, assessment of nervous system function using dynamic changes in calcium signals as a proxy for neuronal activity have demonstrated principles of neuron and circuit function, and principles of information organization and transfer within the brain of living animals. Because of its many advantages an as experimental system, use of Xenopus has often been at the forefront of developing these imaging methods for in vivo applications. Here we present protocols for in vivo structural and functional imaging, including cellular labeling strategies, image collection and image analysis that will expand the use of Xenopus to understand brain development, function and plasticity.

Morphological analysis and time-lapse imaging in frog tadpoles and embryos have been used extensively to study cellular rearrangements during embryogenesis and development (Keller et al., 1989). A major advance occurred when O’Rourke and Fraser used laser scanning confocal microscopy to collect time-lapse images of DiI labeled retinal ganglion cell axons in the optic tectum during initial stages of topographic map formation (O’Rourke and Fraser, 1990). Complementing static images of retinotectal axons in previous studies, daily in vivo time-lapse imaging of single arbors over 4–5 days directly revealed dynamic changes in axon arbor structure that can only be inferred by comparing static images from different animals. Subsequent studies using a variety of time-lapse imaging protocols with inter-imaging intervals ranging from minutes to hours showed that new branches are constantly added and retracted in axon arbors and the net elaboration of the arbor over days occurs by the stabilization of a surprisingly small fraction of the total added branches (Cohen-Cory and Fraser, 1995; O’Rourke et al., 1994; Rajan et al., 1999; Witte et al., 1996). Iontophoresis of DiI into single optic tectal neurons allowed high resolution in vivo time-lapse imaging of individual tectal cell dendritic arbors, revealing dynamic rearrangements of dendritic branches that generated complex dendritic arbors (Wu and Cline, 1998, 2003; Wu et al., 1999). While rapid acoustico-optical device (AOD)-mediated confocal laser scanning which minimized phototoxicity of the fluorescent dyes was essential for the success of these initial in vivo imaging experiments, the combined use of less phototoxic fluorescent proteins, such as green fluorescent protein (GFP) and custom built 2 photon microscopes (Denk and Svoboda, 1997) significantly improved imaging capacity, demonstrated in vivo in the Xenopus tadpole visual system (Ruthazer et al., 2003; Sin et al., 2002). Importantly, these advances permitted frequent images of neuronal structure to be collected without photodamage, increasing the temporal and spatial resolution of studies of structural dynamics.

The capacity to combine in vivo manipulations of visual experience, as well as molecular and cellular components in both presynaptic retinal axons and postsynaptic tectal neurons, with in vivo time-lapse structural imaging ushered in a heyday of mechanistic studies on neuronal development and connectivity. Studies demonstrated that the stabilization of newly added axon and dendritic branches is regulated by glutamate receptors and mediated by establishment of synaptic contacts (Chen et al., 2010; Haas et al., 2006; O’Rourke et al., 1994; Ruthazer et al., 2006; Sin et al., 2002; Wu and Cline, 1998) and the participation of key signaling mechanisms, for instance involving TrkB and BDNF (Cohen-Cory and Fraser, 1995; Marshak et al., 2007), CaMKII (Wu et al., 1996; Wu and Cline, 1998; Zou and Cline, 1996a, b, 1999) and transcription factors (Chen et al., 2012). These studies also demonstrated fascinating retrograde regulation of axon arbor structure by manipulation of postsynaptic proteins. For instance, viral expression of constitutively active CaMKII in tectal neurons strengthened retinotectal synapses (Wu et al., 1996), acted as a stop-growing signal in tectal cell dendrites (Wu and Cline, 1998) and retrogradely limited elaboration of presynaptic retinotectal axons (Zou and Cline, 1996b, 1999). Coordinated control of retinotectal axon and tectal cell dendrite structure was also seen in experiments in which tectal cell expression of the visual activity regulated protein CPG15 increased tectal cell dendritic arbor growth, strengthened retinotectal synapses, and increased axon arbor complexity (Cantallops et al., 2000).

Detailed protocols describing electroporation, morpholino-mediated knockdown, in vivo imaging and accurate reconstruction of axonal and dendritic arbors (Bestman and Cline, 2020; Bestman et al., 2006; Bestman et al., 2012; Haas et al., 2002; Hewapathirane and Haas, 2008; Ruthazer et al., 2013a, b, c, d; Saenz de Miera et al., 2018) have been instrumental in expanding the use of Xenopus tadpoles as an experimental system that is widely valued for studying mechanisms regulating neuronal development and plasticity (Ghiretti and Paradis, 2011, 2014). Only recently has 4D analysis of arbor dynamics been automated (He and Cline, 2011; He et al., 2016; Lee et al., 2013), facilitating the analysis of time-lapse structural plasticity data by speeding up analysis and increasing experiment reproducibility by decreasing inter-investigator influence on outcome measures. This chapter includes the protocol “In Vivo Time-Lapse Imaging and Analysis of Dendritic Structural plasticity in Xenopus laevis Tadpoles” by Lin, He and Cline, describing tectal neuronal electroporation, 2 photon in vivo time-lapse imaging of complete dendritic arbors in isolated tectal neurons, reconstruction from the complete image stacks and the 4D automated analysis and code. This flexible analysis facilitates quantitative comparisons between arbor branch dynamics over a range of time periods and a range of experimental conditions. The dynamic analysis of the 4D time-lapse data uses customized C++ software, 4DSPA, to analyze the branch dynamics, such as branch additions, branch growth and retractions in reconstructed dendritic arbors across different time points. The protocol describes how arbor reconstruction data, using either Imaris or Neurolucida, are prepared and imported into 4DSPA, followed by the pairwise matching analysis of reconstructions from sequential timepoints and the serial dynamic analysis. Output files provide data on branch categories and can be used to generate images with branches color coded according to their branch dynamic categories.

Studies mentioned above in which pre- and postsynaptic neuronal compartments are visualized by expression of targeted proteins tagged with fluorescent reporters have been instrumental in dissecting axonal and dendritic plasticity mechanisms related to synaptogenesis and synapse maturation. Tagging other organelles such as ribonucleoprotein particle (RNPs) (Bestman and Cline, 2008), cytoskeletal proteins and mitochondria (Plucinska et al., 2012) has also demonstrated fundamental elements regulating neuronal development and structural plasticity. Here, we provide the protocol “Imaging Mitochondrial Dynamics in Xenopus CNS” by Feng and Bestman. The protocol describes animal husbandry and bulk electroporation of Xenopus tadpoles to express cytosolic fluorescent protein (FP) to reveal cell morphology and a compatible FP targeted to mitochondria (Weber and Koster, 2013). Multiple proteins can be co-expressed by separating the individual proteins within a single polypeptide with a self-cleaving peptide (Donnelly et al., 2001). The protocol also describes screening animals, in vivo confocal imaging of FP-labeled cells containing FP-tagged mitochondria, and image analysis using open source software Fiji/ImageJ (Schindelin et al., 2012) and Simple Neurite Tracer (Longair et al., 2011) to identify mitochondrial locations within individual cells. This protocol drives FP reporter expression in neural progenitor cells and their neural progeny by using a promotor that requires binding of endogenous Sox2 (Bestman et al., 2012), allowing comparative analysis of mitochondrial dynamics in neural progenitor cells and their neuronal progeny. Although neural progenitor cells don’t use mitochondria for their primary energy source, they do contain many mitochondria, which recent studies suggest may be critical for neural progenitor cell fate and neurogenesis (Bhaskar et al., 2020; Khacho et al., 2017; Khacho and Slack, 2018; Su et al., 2020). Recent studies have used time-lapse in vivo imaging to determine cell lineages of radial glial neural progenitor cells in the optic tectum (Bestman et al., 2012), to screen for genes that regulate tectal neural progenitor cell proliferation and differentiation (Bestman et al., 2015) and to examine interactions between radial glial endfeet and the vasculature (Lau et al., 2017). Imaging mitochondrial dynamics in neural progenitors in Xenopus CNS where effects on cellular structure, cell proliferation and cell fate can be easily assessed under a variety of experimental conditions are likely to reveal elemental functions of mitochondria in neural progenitor cells.

Neuronal function and plasticity can be captured by imaging calcium dynamics in neuronal somata, dendrites and axons, as well as subcellular compartments. Bulk or single cell loading of chemical calcium indicators and expression of genetically encoded calcium indicators are widely used to monitor neuronal activity based on transient elevations of intracellular calcium levels (Garaschuk et al., 2006; Grienberger and Konnerth, 2012; Paredes et al., 2008). Genetically encoded calcium indicators continue to be developed and improved (Chen et al., 2013; Inoue, 2020; Salgado-Almario et al., 2020; Shemetov et al., 2020; Shen et al., 2020). Early studies in isolated Rana pipiens frog brains used membrane permeable fura-2 AM to examine and characterize calcium transients in retinotectal presynaptic compartments in response to optic nerve shock (Feller et al., 1996). Inspired by Feller et al (1996), subsequent studies imaged visually-evoked calcium dynamics in Xenopus tadpole retinotectal axons labeled by bulk delivery of calcium green 1 dextran (CaGD) into the retina (Edwards and Cline, 1999). A key advantage of CaGD for this study was that CaGD delivered into the retina was transported into retinotectal axon arbors, which could then be imaged without interference from the labeling site. In addition, CaGD has a relatively large change in fluorescence emission in response to calcium as well as significant emission under resting conditions, which allowed imaging and reconstruction of axon arbor morphology under resting conditions and with visual stimulation of the eye. Visual stimulation in intact animals increased calcium levels in retinal axon arbors and even in fine filopodia and growth cones, previously seen only in studies in vitro. Correlative studies of visual stimulation induced calcium transients and axon structural dynamics imaged at 3-minute intervals showed that new branches emerged from axons at sites of punctate calcium transients, inspiring later in vivo structural imaging studies investigating the relation between presynaptic sites and axon branch dynamics (Javaherian and Cline, 2005; Ruthazer et al., 2006). More recently, calcium dynamics in populations of neurons in the optic tectum and olfactory system in Xenopus tadpoles have helped identify circuit-based computations and principles of information processing and plasticity in response to sensory inputs. This chapter includes a protocol entitled “Bulk Dye Loading for in vivo Calcium Imaging of Populations of Xenopus Tectal Neuron Visual Responses” by Hogg and Haas. The protocol includes micropipette preparation and injecting the tectum with chemical calcium indicator. Bulk loading of the membrane permeable AM ester of Oregon Green 488 BAPTA (OGB-1) permits in vivo 2 photon time lapse imaging of large populations of neurons, here applied to the optic tectum and combined with in vivo presentations of visual stimuli. Protocols for 2 photon imaging and image analysis are available in Sakaki et al (Sakaki et al., 2020). This strategy has been valuable for imaging neuronal network activity in healthy animals (Dunfield and Haas, 2009; Podgorski et al., 2012) and in models of neurological disease (Hewapathirane et al., 2008). Alternate strategies to image calcium dynamics in individual neurons or populations include genetically encoded calcium indicators, such as the GCaMP family of indicators (Tian et al., 2009), which can be delivered by bulk or single electroporation and can be targeted to genetically defined cell types based on promoter driven expression (He et al., 2016).

Extensive studies in the Manzini and Schild labs have generated fundamental insights into the transformation and organization of odor representations in the vertebrate olfactory system. Here, Offner and colleagues from the Manzini lab provide a protocol “Whole-Brain Calcium Imaging in Larval Xenopus” for calcium imaging in the Xenopus tadpole olfactory system using local injection of calcium indicator dyes or a transgenic Xenopus line expressing GCaMP6s in neurons, which is available from the National Xenopus Resource (NXR) at the Woods Hole Marine Biological Lab. They describe fast volumetric multiphoton imaging strategies which reveal spatiotemporal dynamics of brain activity in different brain regions, indicative of stages of sensory information processing. The protocol describes preparation of the Fluo-4 AM calcium indicator, injecting the indicator into an experimental preparation consisting of the peripheral olfactory system and the brain, followed by 3D multiphoton calcium imaging of odor-evoked calcium responses and analysis of the imaging data using published algorithms and toolkits (Friedrich et al., 2017; Giovannucci et al., 2019; Pnevmatikakis and Giovannucci, 2017; Ramirez et al., 2014).

These diverse protocols will enable further investigation into molecular, cellular and circuit level mechanisms underlying brain development and function in Xenopus, extending the discovery power of this organism in both basic and disease-related research.