The pale spear-nosed bat: A neuromolecular and transgenic model for vocal learning

Abstract

Vocal learning, the ability to produce modified vocalizations via learning from acoustic signals, is a key trait in the evolution of speech. While extensively studied in songbirds, mammalian models for vocal learning are rare. Bats present a promising study system given their gregarious natures, small size, and the ability of some species to be maintained in captive colonies. We utilize the pale spear-nosed bat (Phyllostomus discolor) and report advances in establishing this species as a tractable model for understanding vocal learning. We have taken an interdisciplinary approach, aiming to provide an integrated understanding across genomics (Part I), neurobiology (Part II), and transgenics (Part III). In Part I, we generated new, high-quality genome annotations of coding genes and noncoding microRNAs to facilitate functional and evolutionary studies. In Part II, we traced connections between auditory-related brain regions and reported neuroimaging to explore the structure of the brain and gene expression patterns to highlight brain regions. In Part III, we created the first successful transgenic bats by manipulating the expression of FoxP2, a speech-related gene. These interdisciplinary approaches are facilitating a mechanistic and evolutionary understanding of mammalian vocal learning and can also contribute to other areas of investigation that utilize P. discolor or bats as study species.

Keywords: MRI; Phyllostomus discolor; bats; genome; language; speech; tracing; vocal production learning.

FIGURE 1

Improved gene annotations of the P. discolor genome. UCSC genome browser screenshots show examples of loci with various improvements, including annotation of (A) a gene previously missing from the annotations; CNTNAP2, (B) new exons; FOXP2, (C) improved UTRs; THSD1, and (D) alternative isoforms; GABRP. In each panel, the top track (light blue) indicates the previous annotation as reported in Jebb et al. 2020, and the second track (in black) reports the updated annotation from the current study. Additional tracks in blue and red depict experimental evidence to support the current annotation. Horizontal lines indicate the predicted or observed genetic locus. Vertical lines or thick rectangles indicate the exons identified via predictions or functional data. Thinner rectangles indicate untranslated regions (UTRs) that extend out from the first exon (5’UTR) or the last exon (3’UTR). Arrows indicate a noncoding sequence (introns) between coding regions (exons) and the direction of coding in the genome. Scale bars are indicated below each gene in kilobases (kb).

FIGURE 2

Annotation of miRNAs in the P. discolor genome. (A) In total, 2105 miRNAs were identified, of which 1572 were known miRNAs and 533 were private. (B) Genomic location of known and private miRNAs. The vast majority in both categories were encoded within intergenic and intronic regions. (C,D) Expression of miRNAs in the brain (cortex + striatum), liver, kidney, and testes from five adult P. discolor bats displayed as (C) Log reads per million (RPM) represented as box plots or (D) density plots. The horizontal lines in the box plots indicate the median expression of miR‐337‐3p, boxes extend between the first and third quartile, while whiskers extend by 1.5 times the interquartile range as per the default setting in R. In general, known miRNAs are more highly expressed and have more miRNAs in the high expression range than private miRNAs. (E,F) UpSet plots demonstrate the tissue‐specific expression pattern of (E) known and (F) private miRNAs. The vast majority of miRNAs are expressed in all four tissues tested. Known miRNAs also had large numbers of tissue‐specific miRNAs in the liver, testis, and brain.

FIGURE 3

Neuroimaging data provide anatomical information of the P. discolor brain. (A) An adapted plate of a previously published histological coronal P. discolor brain atlas. Modified from atlas plate 16 (combined Nissl and acetylcholine stain), published in Ref. . (B) Coronal T2‐weighted MR image of a female adult P. discolor bat brain (voxel size 0.1 × 0.1 × 0.1 mm). (C) Coronal, sagittal, and axial T1‐weighted MR images of the same brain (voxel size 0.1 × 0.1 × 0.1 mm). (D) PLI dispersion image of a matching coronal slice of the same brain. (E) PLI fiber orientation map of a matching coronal slice of the same brain. (F) Coronal, sagittal, and axial color orientation maps based on diffusion tensor imaging (DTI) of the same brain (voxel size 0.15 × 0.15 × 0.15 mm). Red annotations indicate white matter structures and blue annotations indicate gray matter structures. Green crosshairs refer to the same location across the different viewing planes and indicate slice depth. Abbreviations: ac, anterior commissure; cc, corpus callosum; Cd, caudate nucleus; ci, capsula interna; Pu, putamen. S, I, R, and L refer to the following orientations: S, superior; I, inferior; R, right side of the brain; L, left side of the brain. The scale bar represents 1 cm.

FIGURE 4

Auditory inputs into the frontal auditory field (FAF) and its descending projections to a possible motor pathway. (A) Summary diagram of ascending auditory pathways to the cortex in the bat and projections from a frontal cortical region to a possible motor pathway. Auditory input via the VIII nerve (not shown) enters the cochlear nucleus (CN). The extralemniscal pathway consists of the NCAT (nucleus of the central acoustic tract), followed by either the suprageniculate body (SG) and/or the superior colliculus (SC). Projections from the SG connect to either the auditory cortex (AC) or the frontal acoustic field (FAF). Alternatively, auditory information can be relayed from the CN to the IC (inferior colliculus) either via the SOC (superior olive complex), NLL (nucleus of the lateral lemniscus), or via direct projection. The IC projects either to the SC or the MGB (medial geniculate body), which itself projects to the AC. The AC projects to the FAF while also receiving FAF projections. Finally, the FAF projects to both the SC and the pyramidal tract (py) via the cerebral peduncle (CP). Arrows in black are known connections in either P. discolor or closely related species, and red connections are based on evidence from this study. Diagram based on work by Pollak and Casseday. (B) The injection site of the tracer into the frontal cortex (Plates #7 and 8 in the reference atlas ⁵⁹ ). (C) Labeled neurons (see arrow in highlighted insert C1) in AC and strong labeling of fibers (star) in the capsula interna (Plate #17). (D) Strong labeling of fibers (see highlighted insert D1) in the cerebral peduncle (Plate #26). (E) Strong labeling of fibers (see highlighted insert E1) in the pyramidal tract, note switching of hemispheres (Plate #34 in the reference atlas ⁵⁹ ).

FIGURE 5

Sagittal view of the adult P. discolor bat brain via histology. (A) Schematic representation indicates key anatomical structures used for protein expression analysis. The main anatomical structures are written in black, while subregions marked out by protein expression are in gray. (B) Nissl stain of the sagittal slice. Immunohistochemical detection of protein expression of (C) FoxP1, (D) FoxP2, (E) parvalbumin, and (F) GluR1. Foxp2 N16 antibody produces high levels of fiber staining that can be seen in panel D in the corpus collosum and in the white matter of the cerebellum and brainstem. For a clearer representation of the staining, the image brightness of all the stained images was altered using a linear adjustment of the mid‐tones of the levels parameter in Photoshop (mid‐tones of Image of Nissl stains were adjusted to 0.5, and images of protein expression were adjusted to 0.3). The scale bar indicates 500 μm. Abbreviations: Acb, nucleus accumbens; cc, corpus callosum; Cd, caudate nucleus; Fr, frontal cortex; Hip, hippocampus; IC, inferior colliculus; Occ, occipital cortex; Md, medulla; Par, parietal cortex; RT, reticular thalamic nucleus; SC, superior colliculus; Temp, temporal cortex; VPL, ventral posterolateral thalamic nucleus; VPM, ventral posteromedial thalamic nucleus.

FIGURE 6

Representation of the human FOXP2 protein displaying known functional domains (in red) and amino acid differences found in P. discolor bats (above) and in mice (below). No changes are found in the functional domains of the protein. Some further variability is detected in the low complexity Q‐rich tract as shown in File S2 (clustal alignment), although this is hard to resolve. Abbreviations: FOX, forkhead‐box DNA binding domain; LZ, leucine zipper domain; Q‐rich, glutamine‐rich region; ZF, zinc finger domain.

FIGURE 7

Design and in vitro testing of FOXP2 transgenic constructs. (A) Schematic of the expression construct used to overexpress FoxP2. We expressed the P. discolor FoxP2 under the control of the human ubiquitin promoter (hUBC). To facilitate detection, we fused FoxP2 with a series of small peptidic tags (HIS‐tag, FLAG‐tag, T7, and Xpress®), and we used a T2A system to separately express the FoxP2 protein from the same promoter driving GFP expression. The GFP protein acted as a marker of the region of infection in the brain. To stabilize the transcript and enhance expression, we used the termination and polyadenylation signal of the human growth hormone (hGH). A premade virus coding for GFP under the control of a CMV promoter was purchased from Virovek (Hayward, California) (AAV5‐CMV‐GFP) to be used as a control. (B) In‐vitro testing of the expression cassette. We transfected HEK293T/17 cells (ATCC, CRL‐11268) with the FoxP2 expression cassette from panel (A) and detected strong overexpression of FoxP2 via western blot compared to untransfected HEK293T/17 cells. The full membrane image is shown in Figure S2. (C) Subcellular localization of the ectopically expressed FoxP2. We transfected HEK293T/17 cells (left panel, DAPI stain) with the FoxP2 expression cassette from panel (A) and detected strong overexpression of FoxP2 (middle panel) in IF using a FoxP2 antibody. GFP (right panel) indicates the presence of transgenic rather than endogenous protein in these cells. FoxP2 expression in the nucleus shows the expected localization of the transgenic protein.

FIGURE 8

In vivo validation of transient transgenic bats. (A) IF of the P discolor brain 10 days after the injection of the viral constructs shows overexpression of FoxP2 in the left hemisphere (injection with the UBC‐GFP‐FoxP2 AAV5 virus) compared to the right hemisphere (injection with the control CMV‐GFP AAV5 virus). (B,C) Zoom in of the injected area. DAPI staining indicates cellular integrity in both hemispheres. (D,E) Strong overexpression of FoxP2 in the left hemisphere was detected with an antibody against FoxP2 (MABE415; Table S5). (F) To quantify the overexpression of FoxP2, we measured the intensity of the signal in IF using MetaMorph (Molecular Devices). The knockin hemisphere showed an approx. three‐fold increase in FoxP2 median expression compared to the control hemisphere (t‐test, p < 2.2e‐16). (G) Following infection with the UBC‐GFP‐FoxP2 AAV5 cassette, we recorded an increase in the number of cells positive for FoxP2 in the knockin hemisphere compared to the AAV5‐CMV‐GFP control hemisphere. Overall, 4515 cells (of 15,691 total cells detected in the region) were found to express FoxP2 in the knockin hemisphere compared to 1657 in the control hemisphere (of 14,449 total cells detected in the region). (F,G) Represent the combined data from two separate brain slices analyzed in the same way (see Table S6 for individual values).