Bulk electroporation and population calcium imaging in the adult mammalian

19 20 The optical recording of light-evoked activity in populations of neurons in the 21 mammalian retina offers several benefits over the use of multi-electrode arrays. However, 22 population imaging has been hindered by the effective loading of synthetic fluorescent 23 indicators, especially in the mature tissue. We have therefore developed an 24 electroporation method to label the complete ganglion cell layer of the adult mammalian 25 retina. We optimized the protocol such that the retina recovers from electroporation and 26 generates responses to visual stimuli. The method can be used with a diverse set of 27 indicators with a range of affinities and emission wavelengths. It therefore can be 28 combined with transgenic animals expressing fluorescent markers to target specific 29 neuronal types. Importantly, the ganglion cell layer remains accessible for subsequent 30 intracellular recording and morphological identification. 31 32 Introduction 33 34 The recording of populations of neurons in the adult mammalian retina has been 35 realized by the use of planar multielectrode arrays (MEAs) (Meister et al., 1994; DeVries 36 and Baylor, 1997; Litke et al., 2003). MEAs record precise spike time information 37 simultaneously from dozens of retinal ganglion cells (RGCs) and allow the recording of 38 nearly complete mosaics from some subclasses of RGCs (DeVries and Baylor, 1997; 39 Shlens et al., 2006; Field et al., 2009). With state-of-the-art MEAs (512 electrodes) and 40 advanced analysis techniques, even detailed functional measurements of RGC receptive 41 field structure – down to the position of single cones – are possible (Field et al., 2007; 42 Field et al., 2010). A limitation of commercially available MEAs is the inability to 43 unequivocally identify/locate the somata generating the recorded spikes. While this 44 problem may in the future be ameliorated by the arrival of high-resolution MEAs 45 (Eversmann et al., 2003; Lambacher et al., 2004; Hutzler et al., 2006), other limitations 46 Articles in PresS. J Neurophysiol (February 23, 2011). doi:10.1152/jn.00722.2010 Copyright © 2011 by the American Physiological Society. 2 are more serious. Because recordings are made with the ganglion cell layer (GCL) 47 attached to the electrode, direct access for subsequent intracellular recording or 48 iontophoretic dye filling is lost. Furthermore, the isolation of spikes attributable to 49 individual neurons is probably biased toward neurons that generate large action potentials 50 (Segev et al., 2004), have larger somata, and/or are located closer to the MEA surface. 51 For instance, when reconstructed from MEA recordings, the mosaics of the smaller 52 midget RGCs tend to be incomplete compared to those from larger RGCs (e.g. (Gauthier 53 et al., 2009)). 54 These limitations of MEA recording can be overcome by using optical population 55 imaging, albeit usually at the expense of spike timing precision. Several protocols have 56 been developed to label the GCL of the vertebrate retina with synthetic fluorescent 57 calcium indicators, including optic nerve backfilling (Zhan and Troy, 1997; Behrend et 58 al., 2009), ballistic “gene-gun” delivery (Kettunen et al., 2002; Morgan and Wong, 2008), 59 and multi-cell bolus loading with membrane permeable indicators (Blankenship et al., 6