Topography of Estradiol-Modulated Genomic Responses in the Songbird Auditory Forebrain Sara E. Sanford, Henry S. Lange, Donna L. Maney Department of Psychology, Emory University, Atlanta, Georgia 30322
Received 12 August 2009; revised 12 September 2009; accepted 14 September 2009
ABSTRACT: Sex steroids facilitate dramatic changes in behavioral responses to sociosexual signals and are increasingly implicated in the sensory processing of those signals. Our previous work demonstrated that in female white-throated sparrows, which are seasonal breeders, genomic responses in the auditory forebrain are selective for conspecific song over frequency-matched tones only when plasma estradiol (E2) reaches breeding levels. Here, we sought to map this E2dependent selectivity in the best-studied area of the auditory forebrain, the caudomedial nidopallium (NCM). Nonbreeding females with low endogenous levels of E2 were treated with E2 or a placebo and exposed to conspecific song, tones, or no sound playback. Immunoreactive protein product of the immediate early gene zenk (egr-1) was then quantified within seven distinct subregions, or domains, of NCM. We report three main find-
ings: (1) regardless of hormone treatment, the zenk response is significantly higher in dorsal than in ventral NCM, and higher in medial than in lateral NCM; (2) E2-dependent selectivity of the response is limited to the rostral and medial domains of NCM; in the more caudal domains, song induces more zenk expression than tones regardless of hormone treatment; (3) even when no sound stimuli were presented, E2 treatment significantly increased zenk expression in the rostral, but not the caudal, domains of NCM. Together, the latter two findings suggest that E2-dependent plasticity in NCM is concentrated in rostral NCM, which is hodologically and neurochemically distinct from caudal NCM. Activity in rostral NCM may therefore be seasonally regulated in this species. ' 2009 Wiley Periodicals, Inc. Develop Neurobiol 70:
INTRODUCTION
songbirds, for example, females will perform a courtship display in response to the song of a conspecific male only during the breeding season, when plasma levels of estradiol (E2) are high. In captivity, songinduced display behavior can be activated in nonbreeding females by E2-treatment (Kern and King, 1972; Searcy and Marler, 1981; Moore, 1983). This extraordinary hormone-dependent behavioral plasticity suggests that E2 may alter the perceived behavioral context or the salience of the signal, perhaps by acting at peripheral or central auditory structures. Receptors for estrogen have been described in the songbird cochlea (Noirot et al., 2009) and auditory forebrain (Gahr et al., 1993; Gahr, 1996; Bernard et al., 1999; Saldanha and Coomaralingam, 2005), and the response properties of the auditory forebrain are modulated by season (Terleph et al., 2008), and by estradiol (Maney et al., 2006; Tremere et al.,
In many species, the process of locating, selecting, and courting a mate is strongly influenced by hormones. Not only is the timing of mating affected, particularly in seasonal breeders, but attraction to sociosexual traits can be altered as well (reviewed by Moffatt, 2003; Jones et al., 2008). In some cases, the effects of reproductive hormones on mating behaviors can be quite dramatic—in seasonally breeding
Correspondence to: D. Maney (
[email protected]). Contract grant sponsor: NSF; contract grant number: IBN 0346984. ' 2009 Wiley Periodicals, Inc. Published online 2 November 2009 in Wiley InterScience (www. interscience.wiley.com). DOI 10.1002/dneu.20757
73–86, 2010
Keywords: auditory; egr-1; estradiol; songbird; ZENK
73
74
Sanford et al.
2009). Collectively, this evidence suggests that in songbirds, auditory processing of conspecific signals may be seasonally regulated. Reproductive state and gonadal hormones can modulate auditory response properties in other taxa as well (reviewed by Arch and Narins, 2009; Miranda and Liu, 2009); thus this phenomenon may be widespread in vertebrates. Because we are interested in how hormones alter the salience of auditory signals, we have chosen to study a region of the auditory system that in songbirds responds selectively to sounds with high behavioral relevance—the caudomedial nidopallium (NCM). This large and complex region is analogous to the supragranular layers of mammalian auditory cortex and plays important roles in perceptual processing and auditory learning (reviewed by Mello, 2004; Pinaud and Terleph, 2008). The majority of research on song-induced activity in NCM has focused on the genomic response, or the expression of immediate early genes such as zenk (egr-1). Zenk expression in this region is greater in response to song than to synthetic tones, and greater to conspecific than heterospecific song (Mello and Clayton, 1994; Stripling et al., 2001). Social factors known to affect the magnitude of the behavioral response, such as song complexity, dialect, or familiarity to the listener, also affect the zenk response (Gentner et al., 2001; Sockman et al., 2002; Maney et al., 2003; Terpstra et al., 2006; cf Dong and Clayton, 2008; Woolley and Doupe, 2008). The response to a song can be enhanced by experimentally increasing its behavioral relevance, for example by pairing it with foot shocks or flashing lights (Jarvis et al., 1995; Kruse et al., 2004). Thus, the zenk response in NCM seems to be proportional to the salience of the signal, and NCM may be involved in encoding behavioral relevance (Pinaud and Terleph, 2008). For many female songbirds, male song is most behaviorally relevant during the breeding season when plasma E2 levels are peaking. In a previous study (Maney et al., 2006), we showed that in a seasonally breeding species, plasma E2 modulates the zenk response in NCM. In female white-throated sparrows, expression of the protein product of zenk (ZENK) was selective for conspecific male song over frequency-matched tones only when plasma E2 was at breeding levels. In nonbreeding females held on short day lengths, the zenk response to song was indistinguishable from that to tones. These results suggested that E2 may act within the auditory system, perhaps within NCM, to alter the processing and perception of reproductively relevant social signals during the breeding season. More recently, Tremere et al. (2009) showed in zebra finches that local blockDevelopmental Neurobiology
ade of estrogen receptors inhibits song-induced expression of ZENK, and that local infusion of E2 is as effective as song at inducing ZENK in NCM. Thus, it is clear that E2 acts directly within NCM to affect genomic responses. Neither of these studies, however, provided information about the distribution of these effects within NCM, which is a large and heterogeneous structure. Despite its heterogeneity, functional domains of NCM have not been clearly delineated. Throughout most of its extent, its responses to sound are organized tonotopically (Ribeiro et al., 1998; Terleph et al., 2006), and the zenk response to song is generally more robust in the dorsomedial aspect than the ventrolateral aspect (Gentner et al., 2001; Maney et al., 2003; Phillmore et al., 2003; Terpstra et al., 2004; Sockman et al., 2005; Avey et al., 2008; Lynch and Ball, 2008; Vignal et al., 2008; Sockman and Ball, 2009; cf. Eda-Fujiwara et al., 2003). Many researchers studying NCM have treated it as a single entity, sampling from a portion or randomly chosen portions within it (e.g., Mello et al., 1995; Duffy et al., 1999; Bolhuis et al., 2000; 2001; Whitney et al., 2003; Soderstrom et al., 2004; Vignal et al., 2004). Gentner et al. (2001) introduced the practice of dividing NCM into dorsal and ventral domains by bisecting it orthogonal to its dorso-ventral axis. Since that time, the majority of published studies have followed this convention (e.g., Eda-Fujiwara et al., 2003; Maney et al., 2003; Phillmore et al., 2003; Hernandez and MacDougall-Shackleton, 2004; Avey et al., 2005, 2008; Sockman et al., 2005; McKenzie et al., 2006; Tomaszycki et al., 2006; Lynch and Ball, 2008; Sockman and Salvante, 2008; Velho and Mello, 2008). Although many researchers have reported that dorsal and ventral NCM differ in their sensitivity to and selectivity for a variety of experimental stimuli, these domains do not appear to have a cytoarchitectonic, neurochemical, or hodological basis. Pinaud et al. (2005) argued that based on connectivity, electrophysiological responses and neurochemical markers, NCM could be divided into rostral and caudal domains. They demonstrated that in zebra finches, immunoreactivity for calbindin is concentrated along the caudal boundary of NCM, as is immunoreactivity for aromatase, an enzyme that converts testosterone into E2. Our own work has shown that the caudomedial aspect of NCM is particularly rich in TH-immunopositive fibers (Fig. 1; Matragrano, unpublished) as well as receptors that bind the avian homologues of vasopressin and oxytocin (Leung, unpublished). This area also has a high concentration of estrogen receptors in several species (Gahr et al., 1993; Gahr, 1996; Bernard et al., 1999;
Hormonal Modulation of Auditory Responses
75
ment differently than the rest of NCM? (3) Does the area dorsal to Field L2, which is typically excluded from analyses of NCM activity, behave like NCM and is it sensitive to E2?
METHODS Animals
Figure 1 Parasagittal view of tyrosine hydroxylase (TH) immunoreactive fibers in the caudomedial nidopallium (NCM) of a female white-throated sparrow. Rostral is to the left. The TH fibers are concentrated along the caudal edge. The cell-like shapes are actually baskets, or fibers that surround TH-negative cells. CMM, caudomedial mesopallium. Scale bar ¼ 300 lm. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]
Saldanha and Coomaralingam, 2005). Caudomedial NCM may thus represent a distinct and unique functional domain. Our goal in the present study was to map E2-dependent selectivity of the zenk response in NCM. Like Maney et al. (2006), we worked with whitethroated sparrows because they have a simple song consisting only of relatively pure-tone whistles, because they are highly seasonal breeders exhibiting large changes in hormonal profiles and behavior from fall to spring (Falls and Kopachena, 1994; Spinney et al., 2006), and because their behavioral responses to song are easily manipulated by E2-treatment in the lab (Maney et al., 2006, 2008, 2009). As is the case for many species, plasma E2 in wild-caught female white-throated sparrows does not reach breeding levels in captivity even under long day lengths; we therefore manipulated plasma E2 by administration of exogenous E2 in nonbreeding females held on a winter-like photoperiod. After E2 manipulation, females were presented with conspecific song, tones, or no playback and ZENK-IR in NCM was quantified in regions of interest chosen to address the following questions: (1) Do the effects of E2 on the zenk response differ between the dorsal and ventral domains of NCM, or between rostral and caudal domains, or both? (2) Does the neurochemically distinct caudomedial edge of NCM respond to E2 treat-
All procedures involving animals were approved by the Emory University Institutional Animal Care and Use Committee and were in keeping with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Forty-two female white-throated sparrows (Zonotrichia albicollis) were collected in mist-nets in Atlanta, GA on the campus of Emory University during November and December 2006. We determined each bird to be female by using PCR analysis of a blood sample (Griffiths et al., 1998) and sex was confirmed by necropsy at the end of the study. The birds were housed in the Emory animal care facility in walk-in flight cages and supplied with food and water ad libitum. Day length was kept at 10:14 h light–dark, which corresponds to the shortest day they would experience at their wintering grounds in Georgia. Prior to the experiment, the birds were moved to individual cages (38 3 38 3 42 cm3) inside large, identical walk-in soundattenuated booths (Industrial Acoustics, Bronx, NY) and kept under the same light conditions (10:14 h light–dark) throughout the experiment to prevent elevation of endogenous plasma E2 (Wolfson, 1958; Shank, 1959).
Hormonal Manipulation We manipulated E2 according to Maney et al. (2006). Briefly, each bird was implanted with one subcutaneous silastic capsule (length 12 mm, ID 1.47 mm, OD 1.96 mm, Dow Corning, Midland, MI) sealed at the ends with A-100-S Type A medical adhesive (Factor 2, Lakeside, AZ). Twenty-one birds received an empty capsule and 21 received a capsule packed with 17-beta-estradiol (Steraloids, Newport, RI). E2 implants of this size increase plasma E2 to physiological breeding levels within 2 days of administration and maintain these levels for at least the duration of this study (Moore, 1983; Maney et al., 2006, 2008). After receiving the implants, the birds were housed in individual cages in groups of four per sound-attenuated booth. These housing groups included birds from each hormone treatment.
Sound Stimuli The song and tone presentations have been previously described (Maney et al., 2006, 2007, 2008). Briefly, recordings of singing male white-throated sparrows were downloaded from the Borror Laboratory of Bioacoustics birdsong database. Stimulus presentations were constructed such that a unique male’s song was repeated for 3 min with Developmental Neurobiology
76
Sanford et al.
15 s of silence between each song. To prevent habituation to the stimulus (see Stripling et al., 1997), after 3 min the identity of the singer changed to a novel male. Each presentation contained 3-min segments from 14 unique males, so that the total presentation time was 42 min. We played seven unique presentations such that each female hearing song in the E2-treated group (n ¼ 7) heard the same presentation as one of the females in the blank-implanted group (n ¼ 7). Each tone presentation was based on a specific song presentation. For each of the 14 songs used to create the song presentations, the frequency of each whistle (note) was measured using AudioXplorer (Arizona Software, San Francisco, CA). Songs usually contained five distinct frequencies. For each song, eight sinusoidal tones were generated at these frequencies and arranged in a random order 200-ms apart. The resulting tone sequences matched individual songs in the average number of onsets and offsets as well as the total sound energy at each frequency. Tone sequences were arranged into presentations as for the song stimuli, with 15 s of silence between each tone sequence, in an order determined by a balanced Latin Square as above.
Stimulus Presentation Sound stimuli were presented 7 days following hormone manipulation. Late in the afternoon preceding stimulus presentation, each bird was isolated in an empty sound-attenuated booth. Each booth was equipped with microphone, speaker, and video camera. One hour after lights-on the following morning, the stimulus presentation was delivered via the speaker at 70 dB as measured at the bird’s cage. Birds were randomly assigned to stimulus groups such that one-third of the birds (E2-treated, n ¼ 7; blank implanted, n ¼ 7) heard song, one-third heard tones, and one-third heard no presentation (hereafter referred to as \silence"). Birds hearing songs or tones heard 18 min of silence following the end of the stimulus presentation so that the total time between presentation onset and tissue collection was 60 min. Using the video camera and microphone inside each booth, we made video and audio recordings of each bird beginning at least 30 min prior to presentation onset and ending 18 min after the presentation ended. On any given day, we presented stimuli to four birds, each housed in a different booth, with stimulus balanced such that each sound was represented each day (e.g., two song, one tones, and one silence). Stimuli were rotated among booths such that over the course of the study, each booth was used to present each type of stimulus an equal number of times.
Tissue Collection Sixty minutes after the onset of the stimulus presentation, each bird was deeply anesthetized with isoflurane (Abbott Laboratories, North Chicago, IL) and a blood sample (200 lL) was taken from the jugular vein. After isoflurane overdose, brains were harvested and immersion-fixed in 5% acrolein (Maney et al., 2003). Ovaries were inspected to Developmental Neurobiology
verify a regressed state. After 90 min of fixation, each brain was divided with a razor blade into left and right hemispheres, each of which was then returned to 5% acrolein for an additional 60 min. Hemispheres were then washed and cryoprotected in 30% sucrose and frozen at 208C.
Radioimmunoassay Plasma E2 was quantified by radioimmunoassay in the laboratory of Dr. John Wingfield at the University of Washington, where the assay has been fully validated. The methods were adapted from Wingfield and Farner (1975) and further described by Williams et al. (2004). We used an antibody (no. 1702, Arnel, New York, NY) directed against 17-betaestradiol. The lower limit of detection was 0.015 ng mL1. To verify that the E2 implants significantly elevated plasma E2 levels and to check for an effect of sound stimulus on plasma E2, we performed an ANOVA using hormone treatment and stimulus type as between-subjects factors.
Histology One hemisphere from each brain was used for this study. Whether the left or right hemisphere was chosen was balanced across treatment and stimulus groups. Every other 50-lm parasagittal section was immunolabeled for ZENK protein following procedures described by Maney et al. (2003). Briefly, sections were incubated in an antibody against ZENK (anti-egr-1; Santa Cruz Biotechnology, Santa Cruz, CA) at a dilution of 1:8,000 for 2 days. The specificity of this antibody has been established in songbirds (Mello and Ribeiro, 1998). The sections were then incubated in a biotinylated secondary antibody (BA-1000, Vector, Burlingame, CA) and then in an avidin-biotin complex (Vector), which amplified the signal. ZENK immunoreactivity (IR) was visualized using diaminobenzidine enhanced with nickel (see Shu et al., 1988; Maney et al., 2003). Sections were mounted onto microscope slides and coverslipped in DPX (Sigma, St. Louis, MO).
Acquisition of Digital Images Images of NCM (46 MB) were acquired using the 43 objective of a Zeiss Axioskop microscope (total magnification 403) and a Leica DC500 camera attached to a Macintosh G5 computer. The light level was kept exactly the same for the acquisition of all images. The region containing NCM was photographed in each section beginning at the midline and extending laterally *1 mm. Each image was converted to an 8-bit grayscale image in ImageJ 1.40g (National Institutes of Health). A representative image is shown in Figure 2(A).
Regions of Interest A primary goal of this study was to describe anatomical variation in E2-dependent selectivity in NCM. Functionally
Hormonal Modulation of Auditory Responses
77
markers (Gahr, 1996; Pinaud et al., 2005; Fig. 1) as well as to capture discernable clustering of ZENK-IR cells along the caudal aspect [see Fig. 2(A)]. We sampled ZENK-IR (see below) from within these four regions as well as from a fifth region, an apical domain (aNCM) located dorsal to Field L2. In the published literature, this region is usually considered part of NCM but may overlap the dorsal portion of Field L (Fortune and Margoliash, 1992). We quantified ZENK-IR in sections that clearly contained the five regions of interest and Field L2 (see Fig. 2) and for which the material from all of the birds could be easily aligned. These criteria were met in a series of four consecutive sections spanning 300 lm. Because the highest densities of estrogen receptor are most likely located in a narrow band of tissue adjacent to the midline (Gahr et al., 1993; Gahr, 1996; Bernard et al., 1999; Saldanha and Coomaralingam, 2005), we were interested in quantifying the effect of E2 on ZENK-IR as medially as possible. In addition to the four sections described above, therefore, we quantified ZENK-IR in the medialmost section in which NCM was clearly identifiable. Medially, the auditory forebrain is more round in shape and although the boundary between CMM and NCM is apparent, Field L2 is not usually discernable (see Fig. 3). Because of the small size of these sections, the rostral regions of interest (rdNCM and rvNCM) were combined into one rostral domain (mrNCM), and the caudal regions (cdNCM and cvNCM) were combined into one caudal domain (mcNCM). aNCM could not be delineated in medial sections.
Figure 2 Caudomedial nidopallium (NCM) in a parasagittal section *500 lm from the midline. Rostral is to the left. (A) Shows an example of ZENK-IR in this study (E2treated bird hearing song). Note the absence of labeling in Field L2 (see also Mello and Clayton, 1994), and clustered labeling caudally. (B) ZENK-IR was sampled in five domains (see text for rationale): apical (aNCM), rostrodorsal (rdNCM), rostroventral (rvNCM), caudodorsal (cdNCM) and caudoventral (cvNCM). CMM, caudomedial mesopallium. Scale bar ¼ 300 lm. distinct domains of NCM have not been clearly delineated. Despite evidence that the rostral domain of NCM may be neurochemically distinct from the caudal domain (see Pinaud et al., 2005; Fig. 1), in the majority of the published literature since 2000, NCM is divided into dorsal and ventral domains rather than rostral and caudal. We divided the region typically defined in parasagittal sections as NCM both rostro-caudally and dorso-ventrally into four primary domains [Fig. 2(B)] as follows: rostrodorsal (rdNCM), rostroventral (rvNCM), caudodorsal (cdNCM), and caudoventral (cvNCM). These domains were similar to those defined by Vignal et al. (2005), except that the boundary dividing the rostral from the caudal domains was located more caudally (*275 lm from the caudal boundary of NCM) to correspond more closely to neurochemical
Quantification of ZENK-IR We used ImageJ to quantify ZENK-IR within each region of interest in each of the five photos from each bird. We sampled ZENK in the rostral domains using the ImageJ circle tool. To sample the largest area possible within each rostral domain, the areas of the circles increased from the medial to the lateral sections as each domain increased in size (areas sampled were as follows: aNCM, 100–115 lm2; rdNCM, rvNCM, and mrNCM, 200–320 lm2). We traced the caudal domains (cdNCM, cvNCM, and mcNCM) with the ImageJ freehand tool. We quantified ZENK-IR according to methods described by Maney et al. (2003). Within each region of interest, ImageJ was used to calculate the percent area having an optical density higher than a threshold value. Because of variability in background staining among brains, the threshold was set manually for each photo such that clusters of pixels selected by the software program agreed with what an observer blind to treatment considered to be labeled nuclei. The value of the threshold was similar for most brains and did not vary according to stimulus treatment. We chose to quantify the percent area above threshold rather than the number of stained objects (cells) because ImageJ could not discriminate individual nuclei when they overlapped. As ZENK-IR increased and the number of overlapping nuclei increased, the number of objects counted by Developmental Neurobiology
78
Sanford et al. tested for effects of hormone treatment and sound stimulus on ZENK-IR by performing a repeated-measures ANOVA with treatment (E2 or blank) and sound stimulus (song, tones, or silence) as between-subjects factors, and region of interest (rdNCM, rvNCM, aNCM, cdNCM and cvNCM; see below for mrNCM and mcNCM) as the repeated measure (SPSS 16.0 for Macintosh, Chicago, IL). The medialmost portion of the auditory forebrain, which is quite small and only loosely connected to the rest of the brain in parasagittal sections, was not retained during sectioning and immunohistochemistry for several of the birds; to analyze the zenk response in mrNCM and mcNCM, we therefore ran a separate analysis in the subset of birds for which very medial tissue was available (Fig. 3; n ¼ 33 birds). For both analyses, when significant main effects or interactions were found, post-hoc unpaired t-tests were conducted to test for effects of hormone treatment within stimulus group and vice versa. We conducted paired t-tests to compare the levels of ZENK-IR among domains and between the medial and more lateral domains.
Analysis of Behavior
Figure 3 Caudomedial nidopallium (NCM) in a parasagittal section *100 lm from the midline. Rostral is to the left. (A) shows an example of ZENK-IR in this study (E2treated bird hearing silence). (B) ZENK-IR was sampled in two domains of medial NCM: rostral (mrNCM) and caudal (mcNCM). NCM, caudomedial nidopallium. CMM, caudomedial mesopallium. Scale bar ¼ 300 lm. ImageJ actually decreased. Quantifying percent area above threshold, rather than the number of labeled cells, allowed us to avoid this problem and to account for overlapping cells.
We scored the audio–video recordings for vocal behavior, including song and contact calls, and CSD behavior, including tail lifts, wing quivers, and trills (see Maney et al., 2003). We tested for effects of sound stimulus on the performance of CSDs by conducting a Kruskal-Wallis nonparametric ANOVA. To determine whether the birds vocalized more to a particular stimulus, we used Mann-Whitney U-tests to compare the number of vocalizations (trills, contact calls, and song) given during the tone presentation to the number given during the song presentation. To test whether E2 treatment caused an increase in vocal behavior, we compared the number of vocalizations given by E2treated birds hearing silence to the number given by blankimplanted birds hearing silence. Because we were interested in finding support for the null hypothesis (no effects), no corrections were made for multiple tests. We have never detected an effect of the birds’ own vocal behaviors on ZENK-IR in NCM (Maney et al., 2003, 2006, 2007).
RESULTS Analysis of ZENK-IR For each region of interest, we divided the total area above threshold (sum of area selected from all sections) by the total area measured (sum of area from all sections) to arrive at percent area above threshold. In some cases, data from one or more sections were not available due to tissue damage; for these, we used data from the remaining sections. Note that our analysis of mrNCM and mcNCM contained data from only one (the medial-most available) section from each bird. There were no usable sections from one brain in the E2-treated group hearing song; that bird was excluded from the ZENK-IR analysis. We then square-roottransformed the data to normalize their distribution. We Developmental Neurobiology
Plasma Estradiol Ovaries were regressed in all of the birds at the time of tissue collection, indicating that endogenous plasma E2 was low. Exogenous E2 treatment increased plasma E2 above the levels in the blankimplanted birds (E2-treated, 1.00 6 0.094 ng mL1; blank-implanted, 0.246 6 0.081 ng mL1; F1,36 ¼ 33.402; p < 0.0001). There was no effect of sound stimulus (F1,36 ¼ 0.249; p ¼ 0.781) and no interaction between stimulus and treatment (F1,36 ¼ 0.036; p ¼ 0.964) on plasma E2 levels.
Hormonal Modulation of Auditory Responses
79
active cells showed a striking clustering pattern along the caudal edge of NCM (cdNCM and cvNCM). Clustering was noted in all of the birds that heard song, and all but one of the birds that heard tones. In some cases, a band-like pattern appeared along the caudal edge [Fig. 4(C)]. We did not notice a predictable relationship between hormone treatment or stimulus type (song or tones) and the clustering patterns. Clustering was not observed in birds that heard silence. The effects of E2 treatment and sound stimulus varied according to the domain of NCM. The repeated-measures ANOVA revealed main effects of both hormone treatment (F1,35 ¼ 5.121, p ¼ 0.036) and sound stimulus (F2,35 ¼ 42.574, p < 0.0001) and a trend for an interaction between the two (F2,35 ¼ 3.049, p ¼ 0.060). There was also a highly significant main effect of domain (F14,140 ¼ 74.888, p < 0.0001); in general, ZENK expression was higher in the dorsal domains (aNCM, rdNCM, and cdNCM) than in the ventral ones (rvNCM and cvNCM). The interaction between stimulus and domain was highly significant (F14,140 ¼ 17.829, p < 0.0001). Although there was no interaction between domain and treatment (F14,140 ¼ 0.345, p ¼ 0.847, a significant interaction among domain, treatment, and stimulus (F14,140 ¼ 2.443, p ¼ 0.017) indicated that the effects of sound presentation and hormone treatment on ZENK-IR were not identical in the five domains of NCM.
rdNCM and rvNCM
Figure 4 Photomicrographs showing patterns of ZENKIR along the caudal boundary of the caudomedial nidopallium (NCM) in parasagittal sections. Rostral is to the left. Note the clustering and band-like patterns of ZENK-IR along the caudal boundary of NCM in birds that heard sound playback (A, B, C) but not in a bird that heard silence (D). These clustering or band-like patterns occurred in 96% of the birds that heard sound playback but there was no discernable association between ZENK-IR distribution and hormone treatment or sound stimulus (song or tones). (A) E2-treated, heard song; (B) blank-implanted, heard song; (C) E2-treated, heard tones; (D) blank-implanted, heard silence. Scale bar ¼ 300 lm.
ZENK-IR in NCM Examples of ZENK-IR labeling are shown in Figure 4. In the birds that heard silence, ZENK-IR was present but weak and relatively evenly distributed. In birds that heard sound playback (song or tones), immunore-
Although the overall level of ZENK expression was higher in rdNCM than rvNCM (p < 0.0001), the nature of its modulation in the two regions was similar [Fig. 5(A,B)]. In the blank-implanted birds, both song and tones induced ZENK-IR at similar levels; the response to song was not larger than the response to tones (rdNCM, p ¼ 0.204; rvNCM, p ¼ 0.632). In the E2-treated birds, however, the response to song was significantly larger than the response to tones (rdNCM, p ¼ 0.011; rvNCM, p ¼ 0.039). E2 treatment therefore seemed to promote selectivity for song over tones in these regions. E2 treatment also affected ZENK-IR independently of sound stimulus; within the group hearing silence, ZENK-IR was higher in E2-treated birds than in blank-implanted birds (rdNCM, p ¼ 0.018; rvNCM, p ¼ 0.030). E2 may therefore have affected basal levels of protein transcription in these regions.
aNCM The overall level of ZENK expression in aNCM was similar to the other dorsal domains (rdNCM, p ¼ Developmental Neurobiology
80
Sanford et al.
Figure 5 ZENK-IR in seven domains of the caudomedial nidopallium (NCM) in female whitethroated sparrows listening to silence (white bars), conspecific male song (black bars) or frequency-matched synthetic tones (gray bars). At *500–800 lm from the midline (A–E, see Fig. 2), E2-treatment induced selectivity for song over tones only in the rostral domains rdNCM and rvNCM. More medially, at *100 lm from the midline (F-G, see Fig. 3), E2-treatment induced selectivity in both the rostral (mrNCM) and caudal (mcNCM) domains. E2-treatment enhanced ZENK-IR in the absence of sound playback in the rostral regions aNCM, rdNCM, rvNCM, and mrNCM, but not in the caudal regions cdNCM, cvNCM, and mcNCM. Data from the medial domains of NCM (F and G) are presented separately above because they were analyzed separately. *significantly higher than silence, p < 0.02. **significantly higher than both silence and tones, p < 0.04. {significantly higher than silence in the blank condition, p < 0.04; {{significantly higher than song in the blank condition, p < 0.04.
0.706; cdNCM, p ¼ 0.487) and significantly higher than the two ventral domains (rvNCM, p
