Sex chromosome gene expression associated with vocal learning following hormonal manipulation in female zebra finches

  1. Laboratory Language Neurogenetics, The Rockefeller University, New York, NY, USA
  2. Department of Molecular Genetics and Microbiology
  3. Department of Neurobiology, Duke Institute for Brain Sciences, Duke University School of Medicine, Durham, NC, USA
  4. Howard Hughes Medical Institute, Chevy Chase, MD


  • Reviewing Editor
    Yoko Yazaki-Sugiyama
    Okinawa Institute of Science and Technology Graduate University, Onna-son, Japan
  • Senior Editor
    Claude Desplan
    New York University, New York, United States of America

Reviewer #1 (Public Review):

Prior research demonstrated that vocal learning is sexually dimorphic in zebra finches; female song nuclei atrophy and fail to develop, but can be rescued with exogenous 17-𝛃-estradiol (E2) treatment. In previous research, the authors treated both male and female birds with exogenous E2. They laser-captured dissected tissue samples from the E2-treated individuals as well as untreated controls. They then extracted RNA and used RNA-seq to characterize the transcriptomes within and adjacent to four major song nuclei (HVC, LMAN, RA, Area X) in these birds. In this study, Davenport et al. remapped this massive amount of transcriptome data (n=3 birds per sex/treatment group) to fully resolve the genomic location of differentially expressed genes, which they assigned to several modules based on co-expression. Adequate read mapping to all chromosomes was previously impossible with zebra finch genome assemblies lacking W chromosome data. Using the high-quality zebra finch genome assembly with Z and W chromosomes (bTaeGut2.pat.W.v2), the authors were able to demonstrate the enrichment of certain modules on certain chromosomes; most interestingly, Z chromosome gene expression was increased in E2-treated females. This research greatly improves our understanding of the ontology and location of genes involved in song development in E2-treated females, providing insight into the development of vocal learning in the zebra finch.

The authors' main conclusions on the importance of certain gene modules in the vocal learning process are well warranted by their excellent data and thorough analyses, but should not be too broadly interpreted as necessarily applying to the gene expression involved in vocal learning in other species. While the data here further supports convergent evolution in vocal learning genes in humans and zebra finches, vocal learning is unusually sexually dimorphic in zebra finches compared to most other vocal learners.

The authors note the possibility of female haploinsufficiency of Z-linked genes such as the growth hormone receptor (GHR) and also imply there are potential effects of the fission of the ancestral chromosome into passerine chromosomes 1 and 1A impacting the typical development of male zebra finch song and the lack thereof in females. These thoughts are intriguing and should prompt further transcriptomic research in avian species with the same genomic features (ZW females, split 1 and 1A chromosomes) where females also learn song, i.e. female-singing passerine species. Currently, it is impossible to say if female-singing species are, as is likely with the E2-treated zebra finch females, using estrogen signaling pathways to regulate an increase in dosage of these genes. Alternately, these female-singing birds may be using different gene modules, which is also worthy of investigation. This research excels at elucidating the genomic underpinnings of vocal learning in a model organism; further research will demonstrate how broadly applicable these authors' findings are across other species.

Reviewer #2 (Public Review):

This work tried to identify genes involved in the song learning of zebra finches by looking at gene expression from individuals who could learn to sing (males and E2-treated females) or not learn to sing (untreated females). They use extensive RNAseq data from one of their previous publications, but this time align the reads to a female genome (from another of their previous publications). Here they use traditional Weighted Gene Correlation Network Analysis (WGCNA) to identify modules (sets of genes whose expression co-varies across all samples) and then find how these sets of genes collectively differ between brain regions involved in song learning and the surrounding tissues not involved in song learning. This approach identified modules that were significantly different in expression between males and females, and the authors interpret this as sex chromosomes being involved in song learning. However, this approach is highly skewed by unrelated patterns of gene expression from the sex chromosomes due to a lack of dosage compensation in birds. In short, by generating WGCNA modules from males and females, all sex chromosome genes will be expected to be artificially pulled into one module due to methodological artifacts and not true biologically relevant differences.

It's nice to see large datasets being reevaluated with updated genomes.

Zebra finches (like all birds) do not have XX/XY sex determination, but instead have ZZ/ZW, which means that males have two copies of the Z chromosome and females have one copy of the Z and one copy of the W. This is important because it means that if males and females express all their genes at the same rate, then expression of Z genes will always be twice as high in males relative to females. [While mammals have mechanisms to equalize expression of X chromosome genes between males and females (aka. dosage compensation), birds do not have such chromosome-scale mechanisms.] Therefore, the expression of genes on the sex chromosomes of birds will always differ dramatically between males and females, without necessarily indicating any biologically meaningful difference. WGCNA-based approaches (such as those used here) form modules based on differences in gene expression across all samples. Since this manuscript used all samples to generate their WGCNA modules then all (or nearly all) of the expressed genes on the sex chromosomes would be expected to be pulled into the same module - which is precisely what happened: the reported 'module E' contained 904 genes, while there are only 1,078 genes annotated on the Z chromosome of the reference genome used. Some of these genes may 'belong' in other modules if there are regional differences, but the dosage-driven-differences between sexes across all regions will overwhelm these signals. Therefore great care needs to be taken when interpreting the results of this study until such time as independent analyses can verify these results.

Reviewer #3 (Public Review):

Davenport et al have investigated how the administration of a masculinizing dose of estrogen changes the transcriptomes of several key song nuclei song and adjacent brain areas in juvenile zebra finches of both sexes. Only male zebra finches sing, learn song, and normally have a fully developed song control circuitry, so the study was aimed at further understanding how genetic and hormonal factors contribute to the dimorphism in song behavior and related brain circuitry in this species. Using WGCNA and follow-up correlations to re-analyze published transcriptome datasets, the authors provide evidence that the main variance of several identified gene co-expression modules shows significant correlations with one or some of the factors examined, including sex, estrogen treatment, regional neuroanatomy, or occurrence of vocal learning.

Among the main strengths are the thorough gene co-expression module and correlation analyses, and the inclusion of both song nuclei and adjacent areas, the latter serving as sort of controls for areas that are not dimorphic and likely broadly present in birds in general. The most relevant finding is arguably the identification of some modules where gene expression variation within song nuclei correlates with hormonal effects and/or gene location on sex chromosomes, which are present at different dosages between sexes. The study also shows how a published RNA-seq dataset can be reanalyzed in novel and informative ways.

Among its main weaknesses, the study relies entirely on one set of transcriptomic data and lacks effort to validate the inferred direction of regulation in the identified co-expression modules using other molecular methods or approaches on independent samples. The study shows that some representative and/or highly significant genes in some of the main modules that correlate with anatomical, sex, or hormone treatment group comparisons indeed differ in expression when comparing song nuclei vs surroundings, male vs female, or E2- vs VEH-treated tissues in independent samples by qPCR or in situ hybridization would provide important validation and enhance experimental rigor for the analyses presented. In the absence of this further validation, the WGCNA data need to be interpreted with caution.

The findings related to ex-chromosome genes (i.e. module E) are a significant strength of the study. Two points, however, need to be taken into account more closely. First, sex differences in gene expression in areas that are not song nuclei are likely related to functions other than song behavior or vocal learning, thus not related to the main question posed by the study. Furthermore, an alternative interpretation with regard to sex chromosome gene expression is that the higher male expression for a large number of Z chromosome genes may not significantly or fundamentally affect brain cell function and can be tolerated, thus not requiring active compensation. This alternative interpretation (mentioned for song nucleus RA in Friedrich et al, Cell Reports, 2022) suggesting that the higher male dosage of many of these genes might not affect or contribute to sex differences in brain function, cannot at present be discarded, and should at least be acknowledged.

Friedrich et al, Cell Reports, 2022 (Table S3 ) presented an extensive manual curation of W chromosome genes in zebra finches. BLAST alignments showed that a large proportion of W chromosome genes are also on the Z, noting that only a small subset of these are annotated as Z:W pairs. The genes that are truly W-specific and present at a higher dosage in females are thus only a fraction of W-chromosome genes. This creates a complication when examining the mapping of RNA-seq reads to sex chromosomes: to conclude about higher expression of W genes in female tissue samples one needs to take into account reads that may also map to the homologous genes on the Z, if that gene is present as a W:Z pair. Because Friedrich et al mapped reads to a male genome assembly, W genes were not assessed, thus the present study provides novel info. However, the issues above need to be acknowledged and taken into account to accurately assess sex differences in W chromosome gene expression.

  1. Howard Hughes Medical Institute
  2. Wellcome Trust
  3. Max-Planck-Gesellschaft
  4. Knut and Alice Wallenberg Foundation