Simultaneous RNA Sequencing and DNA Methylation Profiling Reveals Neural Mechanisms That Regulate Sensitive Period Behavioral Learning

Abstract

Developmental processes emerge from both maturational and experience-dependent mechanisms. Experience at the proper maturational stage is essential for the acquisition of many complex cognitive and behavioral processes. A striking example of this is a critical period, a restricted developmental phase during which experience is required for both behavioral acquisition and period closure. Juvenile male zebra finches (Taeniopygia castanotis) possess a critical period for song learning; hearing an adult "tutor's" song between posthatch days 30-65 is necessary for each male to produce a socially functional adult song. However, if tutor song is not experienced in this age range, juveniles can still learn beyond posthatch day 65. Our broad objective is to decipher the neurogenomic mechanisms that promote or limit the ability to learn, leveraging the known parameters of the critical period in the male zebra finch's sensory song learning ontogeny. Here, we manipulated juvenile males' tutor exposure and provided song playback experience at two ages, at the beginning or end of the critical period. We probed the relationship between DNA methylation and transcriptional profiles from the same individual and tissue samples to enhance interpretation across different levels of biological organization. Our findings uncovered specific genes and processes that may regulate aspects of critical period learning, as well as aspects of DNA methylation dynamics and how they correspond to RNA measures. Because we distinguished effects of age and experience, outcomes provide insight into fundamental links between epigenetic and molecular properties as the developing brain shifts its ability to learn.

Keywords: epigenetics; social ontogeny; songbird; song‐learning; zebra finch.

FIGURE 1

Timeline of critical period for male tutor song memorization with experimental details. Zebra finches hatch on Posthatch day 1 (P1) and are adults at P90. When males are tutored, the critical period for tutor song memorization spans P30–65 (red line), but it can be extended in birds isolated from hearing song during that phase (red dashed arrow). To assess how mechanisms of maturation and accumulated experience at the typical boundaries of the critical period affect the acute epigenetic and molecular response to hearing song, we assayed auditory forebrains from P32 and P67 males (blue arrows) after they were either placed in either Tutored or Isolate conditions at P30, and subsequently exposed to acute song playback or silence (music note and null speaker symbols, respectively).

FIGURE 2

Influence of age and tutoring experience on gene expression and methylation. (A–B) Lollipop plots displaying the number of differentially‐methylated CpG sites (red), predicted genes based on those sites (blue), and differential RNAs (yellow) in (A) the byAge comparison, considering rearing condition and song playback exposure; and (B) the byCondition comparison, considering age and song exposure. The number of genes that overlap between RRBS and TagSeq data are listed as italicized numbers above each comparison. (C–F) Volcano plots with select genes labeled showing distribution of differential RNAs in Tutored versus Isolate comparisons for (C) P32‐silence, (D) P32‐song, (E) P67‐silence, and (F) P67‐song groups. The 10 genes with the lowest adjusted p values and HGNC gene symbols (i.e., not LOC‐annotated genes) are labelled. All genes with adjusted p value < 0.05 are in red.

FIGURE 3

bySong analysis reveals effects of age and tutor rearing condition on RNA and DNA methylation profiles. (A) Lollipop plot showing the number of differentially methylated CpG sites (red), predicted genes based on those sites (blue), and differential RNAs (yellow) in the bySong comparison, considering age and rearing condition. (B) Normalized read counts showing a diversity of effects of age and rearing condition on the effect of song playbacks for IEGs Arc, egr‐1, fos, jun, junD, and NR4A3. Data for each gene is shown in two side‐by‐side panels, one for each age. p‐adjusted < 0.05, ^ p value < 0.05. (C) Scatterplot of fold changes between Song and Silence in Tutored versus Isolate auditory forebrains. Fold changes are plotted for genes significantly differentially expressed in the Tutored condition against the Isolate condition (where they are not necessarily differentially expressed). FOS and UROC1, for example, show a consistent response in both rearing conditions, whereas FAM151B, DUSP28 have opposite responses in Tutored compared to Isolate. Labeled genes have fold change > 1 in both rearing conditions, and have assigned gene names (i.e., not “LOC” genes). (D) Data from the top two genes showing a statistical interaction between acute song playback and rearing condition at P67. All box plots show median count and the 25th and 75th percentile; whiskers extend based on the inter‐quartile range in both directions. Circles depict individual data points. Please note different y‐axes.