Erbin and ErbB2 Play Roles in the Sexual Differentiation of the Song System Nucleus HVC in Bengalese Finches (Lonchura Striata Var. domestica)

Yueliu Zhao,1† Xuebo Zhang,1,2† Rui Wang,1† Jie Bing,1 Fan Wu,1 Yitong Zhang,2 Jincao Xu,3 Zhongming Han,3 Xinwen Zhang,2* Shaoju Zeng 1* 1 Beijing Key Laboratory of Gene Resource and Molecular Development, Beijing Normal University, Beijing, 100875, China

2 College of Life Sciences, Hainan Normal University, Haikou, 571158, China

3 Department of Otorhinolaryngolgoy, The General Hospital of the PLA Rocket Force, Beijing, 100088, China

Received 11 April 2017; revised 20 September 2017; accepted 25 October 2017

ABSTRACT: Song control nuclei have distinct sex- injection of erbin-orerbb2-interfering lentiviruses into ual differences in songbirds. However, the mechanism the HVC and its overlying VZ at PHD 15, the cell prolif- that underlies the sexual differentiation of song nuclei is eration in the VZ at PHD 24, the number of the differen- still not well understood. Using a combination of anatom- tiated neurons (Hu1/BrdU1 or NeuN1/BrdU1)inthe ical, pharmacological, genetic, and behavioral HVC at PHD 31 or PHD 130, and the number of RA- approaches, the present study investigated the role of projecting cells at PHD 130 all decreased significantly. erbb2 (a homolog of the avian erythroblastic leukemia Additionally, the adult songs displayed serious abnormal- viral oncogene homolog 2) and the erbb2-interacting ities. Finally, 173 male-biased genes were expressed in gene, erbin, in the sexual differentiation of the song the developing HVC at PHD 15 using cDNA microarrays, nucleus HVC in the Bengalese finch. We first found that of which 27.2% were Z-linked genes and approximately both erbin and erbb2 were expressed in the developing 20 genes were involved in the Erbin- or ErbB2-related HVC at posthatch day (PHD) 15 in a male-biased fashion signaling pathways. Our results provide some specific using qRT-PCR and in situ hybridization. Following the genetic factors that contribute to neurogenesis and sex addition of a pharmaceutical inhibitor of the ErbB2 sig- differentiation in a song nucleus of songbirds. VC 2017 naling pathway to the culture medium, cell proliferation Wiley Periodicals, Inc. Develop Neurobiol 78: 15–38, 2018 in the cultured ventricle zone (VZ) that overlies the Keywords: Songbirds; cell proliferation; cell differentia- developing HVC decreased significantly. After the tion; sexual difference; Z-linked genes

Abbreviations: FC, fold changes; PHD, post-hatch day; RA, robust nucleus of the arcopallium; VZ, ventricle zone INTRODUCTION *Correspondence to: S.J. Zeng ([email protected]) and X.W. Zhang ([email protected]). †These authors contributed equally. In many oscine species, only male birds sing highly Contract grant sponsor: National Natural Science Foundation stereotyped songs. Accordingly, male birds possess of China to SJ Zeng; contract grant numbers: (No: 31372200, 31672283), XW Zhang (No: 31560275 and 31360517), ZM Han dimorphic and highly interconnected song control (No: 81371352) and XB Zhang (No: 31360243). nuclei (Nottebohm and Arnold, 1976; Tobari et al., Ó 2017 Wiley Periodicals, Inc. 2005). Therefore, songbirds provide a model to study Published online 30 October 2017 in Wiley Online Library (wileyonlinelibrary.com). the neural mechanism that underlies the sexual DOI 10.1002/dneu.22551 dimorphism of brain areas. 15 16 Zhao et al.

During mammalian embryonic development, the investigations, we wanted to know whether the two undifferentiated gonad develops into a testis rather genes are involved in the sexual differentiation of than an ovary under the action of the testis- HVC. determining gene, the sex-determining region of the To address the issue, we first investigated the sex- Y-chromosome (Sry) (Sinclair et al., 1990; Kashi- ual dimorphism in the expression of erbin and erbb2 mada and Koopman, 2010). The testes secrete steroid in the developing HVC of the Bengalese finch. We hormones, directing somatic cells to differentiate in a then examined whether erbin and erbb2 affect cell male fashion (McCarthy, 2009). However, this mode proliferation in the ventricular zone overlying the of sexual differentiation does not extend to all verte- developing HVC at PHD 15 using a pharmaceutical brates, including avian species with homogametic inhibitor for the ErbB family. Next, we injected males (ZZ) and heterogametic females (ZW). First, erbin- or erbb2-interfering lentiviruses into the devel- unlike in mammals, there is an incomplete sex- oping HVC and its overlying VZ and further investi- reverse of masculinization in female birds following gated how cell proliferation, neuronal differentiation estradiol treatment and an unsuccessful prevention of and adult song organization are affected. Finally, masculine development in male birds after blocking using cDNA microarrays, we examined sex-biased testicular hormones (Wade and Arnold, 1996). Sec- genes in the developing HVC, specifically those in ond, beyond the lack of a gene that is homologous to the Erbin- or ErbB2-involved signaling pathways. the mammalian Sry, chromosome-wide dosage com- pensation akin to mammalian X-inactivation is largely lacking in birds (Ellegren et al., 2007; Itoh METHODS et al., 2007; Naurin et al., 2011). Studies of gynan- dromorphic zebra finches and chickens reveal that Animals sexual differentiation of somatic cells (including cells in the song control nuclei) is the result of hormonal The Bengalese finches (Lonchura striata var. domestica) and genetic factors (Agate et al., 2003; Zhao et al., used in this study were originally purchased from a local 2010). Thus, the sexual differentiation of the song supplier (Beijing Guanyuan Flowers and Birds Market, Bei- jing Haidian District, Beijing, China) and then raised in the control nuclei seems to depend both on sex steroids breeding colony at Beijing Normal University. Nestling and on the steroid-independent differential expres- (PHD 15), young (PHD 45) and adult birds (PHD > 120) sion of genes. To our knowledge, the possible mecha- were used in this study. The birds were maintained in cages nisms that underlie the sexual differentiation of song (50 cm 3 62 cm 3 38 cm) equipped with perching sites control nuclei are still not clear. and nest boxes in a room under a 14/10 h light/dark cycle at We recently identified male-biased genes in the 20–308C. Seeds and fresh water were provided at all times, telencephalon of the Bengalese finch, including a Z while green vegetable supplements were provided occa- chromosome gene erbin (ErbB2 interacting protein) sionally. Each cage contained 4–7 adult birds or 3–4 nes- and ErbB2 (a homolog of the avian erythroblastic tling birds that lived with foster parents. All experiments leukemia viral oncogene homolog 2) (Sun et al., that were performed in this study were conducted in accor- 2010). Both Erbin and ErbB2 have been reported to dance with the Beijing Laboratory Animal Welfare & be involved in cell proliferation, survival, and differ- Ethics Review guidelines. Furthermore, all of the protocols entiation through signaling pathways mediated by were reviewed and approved by the Animal Management Erbin (mitogen-activated protein kinases, MAPK; Committee of the College of Life Sciences of Beijing Nor- transforming growth factor b, TGF-b; nuclear factor- mal University. jB, NF-jB) or by Erbb2 (extracellular regulated pro- tein kinases, ERK or Ras-Raf-Mek-Erk; protein Sex Genotyping, RNA Extraction, in Situ kinase C, PKC) (Borg et al., 2000; Rangwala et al., Hybridization and Quantitative Real-Time 2005; Citri and Yarden, 2006; Zhou et al., 2012; Ros- PCR (qRT-PCR) koski, 2014; Tao et al., 2014). Given the reported To identify the sex of nestling finches (PHD 15), genomic roles of erbin and erbb2, we further wanted to know: DNA was extracted from axillary venous blood using a (1) whether erbin and erbb2 expressions have sexual TIANamp Genomic DNA Kit (TIANGEN Biotech, Bei- differences in the developing HVC, a song control jing). The PCR primers used in sex determination were pre- nucleus in the Bengalese finch; (2) whether cell pro- viously described in detail (Wang and Zhang, 2009; Wang liferation, and neuronal differentiation are affected et al., 2010), and listed in Table 1. PCR products were elec- after injections of erbin- or erbb2-interfering lentivi- trophoresed in a 1.5% agarose gel. There was a 0.18-kb Z- ruses into the developing HVC and its overlying VZ linked fragment in both sexes and a 0.22-kb W-linked frag- (after erbin- or erbb2 knock-down). Based on these ment in the female.

Developmental Neurobiology Erbin and ErbB2 in the Sexual Differentiation of HVC 17

Table 1 Primers for PCR or qRT-PCR Use Gene name Primers (50! 30) For sex determination SS: CTCCCAAGGATGAGAAACTGTGCAAAACAG AS: CTTCACTTCCATTAAAGCTGATCTGGAATTTC For RNA probe synthesis erbin SS: CAAAGTCAGTCCAAGATC AS: GACAGGCTTCAGTGTTTA erbb2 SS: AAGATCTTTGGGAGCCTGGC AS: CGGGCCCCCAGCAGTGCC For qRT-PCR erbin SS: CAGTAAAGAATCAGTTCTGC AS: ATGTTCACATCACATGTTCAT erbb2 SS: ATCAAGTGGATGGCGCTGGA AS: TCGCCTTTCTCCAGCAGGT b-actin SS: TTGGCAATGAGAGGTTCAGGT AS: TACGGATGTCCACATCACACT SS: Sense sequence; AS: antisense sequence.

The Bengalese finches were anesthetized via intramuscu- manufacturer’s instructions. The qualified RNA was used lar injection of pentobarbital (Sigma, 0.2 g of pentobarbital in qRT-PCR or cDNA microarray analysis. in 20 ml saline). With reference to previous reports (Cardin The primers for the RNA probes used in the in situ and Schmidt, 2004; Schumacher et al., 2011; Pawlisch and hybridization protocol were designed with the Primer3 out- Remage-Healey, 2015), the following doses were used in put program and shown in Table 1. The Erbin probe length the present study: overdose (0.065 ml/each bird) for the was 234 bp, ranging from chromosome location 6300 to birds that were sacrificed immediately, and 0.03–0.05 ml/ 6533, with an 18-bp sense and anti-sense primer at the two each bird for surgical anesthesia (0.03 ml for PHD 15 and ends of zebra finch XM_004177159.1 (Chr. Z). The ErbB2 0.05ml for PHD > 120). For RNA extraction, the birds probe length was 466 bp, ranging from chromosome loca- were deeply anesthetized with an overdose of pentobarbital tion 1185 to 1650, with 20-bp sense and 18-bp anti-sense and sacrificed with a perfusion of ice-cold Hanks’ balanced primers at the two ends of chicken NM_001044661.1 (Chr. solution (HBSS, sigma) through the heart. The brains were 27). Reverse transcription was performed with M-MLV rapidly removed and placed in ice-cold HBSS. Then, the (Promega). The target bands were then cloned into the brain was split into two hemispheres along the midline. A pGEM-T Easy vector (Promega) and sequenced to confirm sagittal section of the hemisphere was glued to the stage of the accuracy of the probes. The Erbin fragment shared 99% a motorized advance vibratome (World Precision Instru- identity with zebra finch XM_004177159.1, and the ErbB2 ments Inc.) and cut parasagittally into 400-lm-thick slices. fragment shared 98% identity with chicken Based on the report showing the anatomical position of NM_001044661.1. The fragment of ErbB2 cloned from the HVC in the telencephalon, the brain sample that contained Bengalese finch was deposited into GenBank (KT932702). the developing HVC and its overlying VZ was obtained by The sense and anti-sense RNA probes were transcribed cutting along the dashed line located approximately at the using a digoxigenin (DIG) RNA labeling kit (Roche) end of the caudal hyperpallium lamina [LH, Fig. 1(E)]. according to the manufacturer’s protocol. The correspond- Under dark-field illumination, LH could be reliably deter- ing sense probes were used as negative controls. mined, although the boundaries of HVC, especially female For in situ hybridization, birds of both sexes at PHD 15, HVC, were difficult to determine. To know the expression PHD 45 and adult (PHD > 120) were euthanized as levels of the examined genes in the areas adjacent to the described above. The hearts were then perfused with ice- HVC (but not contain HVC), we also collected a long nar- cold 0.01 M phosphate-buffered saline (PBS, pH 7.4), fol- row brain belt [about 300 lm wide, delimited by the two lowed by ice-cold 4% paraformaldehyde (PFA, in 0.1 M dashed lines, Fig. 1(E)], just under the above brain sample PB, pH 7.4). After fixation with 4% PFA for 24 h and cryo- containing the developing HVC. protection in 30% sucrose (in 0.1 M PB, pH 7.4) at 48C, the Total RNA was extracted using the Trizol reagent (Invi- brains were cut into 10-lm-thick sections using a cryostat trogen). The RNA samples were treated with DNase I (Leica). Every sixth sagittal section was mounted on a (Takara Biotech Co., Dalian, China) for purification from poly-lysine-coated slide, and a total of six sets of sections DNA contamination. Sample quality was determined using were collected for each hemisphere. The sections were a formaldehyde agarose gel (no smearing of ribosomal stored at 2808C until use. bands was visible), and 260/280 ratios were calculated In situ hybridization was routinely carried out as described using an Ultraspec 3000 spectrophotometer (all values by Zhu et al. (2008). Briefly, after endogenous AP activity were between 2.0 and 2.1). Equal amounts of total RNAs (1 was quenched with 0.2M HCl, proteinase K (GIBCO) diges- lg) were reverse transcribed using the PrimeScriptTM II 1st tion was carried out at 378C, followed by postfixation in 4% Strand cDNA Synthesis Kit (TaKaRa) according to the freshly depolymerized paraformaldehyde. The sections were

Developmental Neurobiology 18 Zhao et al.

Figure 1 In situ hybridization and quantitative real-time PCR show sexual differences in expres- sion levels of erbin and erbb2 mRNA in the HVC of Bengalese finches. A1-D1, Nissl staining for male (A1 and C1) or female (B1 and D1) HVC at PHD 15 (A1 and B1), and 45 (C1 and D1). A2- D3, Labeling for erbin or erbb2 mRNA in male and female HVC at PHD15 (A2-B3) and 45 (C2- D3). E, Sagittal schematic of HVC in the brain. Dorsal is up, and caudal is to the left. Scale bar in D3 5 200 lm for A1-D3. F, G, Labeling for erbin or erbb2 sense probes in male (F) and female (G) HVC at PHD 45. H, I, Total numbers of labeled cells for erbin (H) and erbb2 (I) mRNA in HVC at PHD15, 45 and adult (>PHD120). *p < 0.05; **p < 0.01; ***p < 0.001. Error bars indicate SEM. J, K, Quantitative real-time PCR of erbin (J) and erbb2 (K) at PHD 15 in brain samples or explants that contained the developing HVC and its overlying VZ, which we obtained by cutting along the thick dashed line (shown in E) located at the caudal hyperpallium lamina, and in narrow belts that were outside the HVC (delimited by two dashed lines, E). Bars with different letters are signifi- cantly different. [Color figure can be viewed at wileyonlinelibrary.com] then acetylated with 0.25% (v/v) acetic anhydride (Sigma) in to the sections for incubation overnight (100ll for each sec- a 0.1M triethanolamine (Sigma), and incubated for 1 h with tion). The sections were then washed with 23 SSC and prehybridization solution (50% formamide, 2 3 SSC, 0.05 g/ 0.05% Tween-20, 0.1M Tris–HCl, and 0.15M NaCl. The sec- ml dextran sulfate, 1 3 Denhardt’s, and 0.1mg/ml salmon tions were incubated with a sheep polyclonal anti- sperm DNA). After digoxigenin-labeled RNA probes were digoxigenin antibody F(ab)2 fragment conjugated with alka- diluted in prehybridization buffer (1ng/ml), they were added line phosphatase (Roche Molecular Biochemicals,1:1000).

Developmental Neurobiology Erbin and ErbB2 in the Sexual Differentiation of HVC 19

After washed with TTN and Tris (pH 9.4), color was devel- (EGFR/ErbB2), a 4-(4-benzyloxyanilino)-6,7-dimethoxy- oped in the presence of nitroblue tetrazolium (NBT) and 5- quinazoline (Merck Millipore, 324673) stock solution in bromo-4-chloro-3-indolyl phosphate (BCIP) (Promega) for DMSO, and 50-bromo-20-deoxyuridine (BrdU, Sigma) to 48 h. Negative controls were prepared using the correspond- label proliferating cells were added to the culture medium ing sense probes. At least one set of sections was used for at a final concentration of 10 lM for both the inhibitor and Nissl-staining. BrdU and washed out after 24 h. According to the manufac- The primers for qRT-PCRs were designed following the turer’s protocol, 4-(4-benzyloxyanilino)-6,7-dimethoxyqui- aforementioned method, based on the sequences of the nazoline acts a potent, reversible, and ATP-competitive studied genes published in the NCBI GenBank: Erbin inhibitor of EGFR and c-erbB2. This drug can effectively (ERBB2IP, XM_004177159.1) and b-actin (ACT5C, inhibit the proliferation of cells expressing EGFR or c- XM_002192780.2) in zebra finch and ErbB2 erbB-2. (NM_001044661.1) in chicken (not available for zebra The chosen doses of the drug were determined in our finch). The designed primers were shown in Table 1, and preliminary experiments. To obtain the dose-response the expected fragments for erbin, erbb2 and b-actin were curve, final drug concentrations of 0, 1.25, 2.5, 5.0, 10.0, or 161 bp, 162 bp and 123 bp, respectively. The amplicons 20 lM were used in the cultured brain explants that con- were cloned into the pGEM-T Easy vector (Promega) and tained the developing male HVC at PHD 15, with reference sequenced to confirm the consistence with the correspond- to the protocol and a related report (Yu et al., 2011). The ing accession numbers in the NCBI GenBank mentioned concentration of DMSO in the media was adjusted to 0.1% above. Real-time PCR was performed with an ABI 7500 in all six groups. The numbers of proliferating cells Sequence Detection System (Applied Biosystems), using (labeled by BrdU1) per mm in the VZ that over-lay the the SYBR Green Mix kit with 10–20 nM cDNA template developing male HVC decreased with increased concentra- (2–4 ll), and 0.2 lM each primer (20 ll total volume). The tions from 1.25 to 20 lM and were examined 24 h after cycling conditions were 608C for 2 min, 958C for 10 min, inhibitor addition. However, the numbers of cells stained and 40 cycles at 958C for 15 s and 608C for 1 min. Each by cresyl violet within HVC (per mm2) were also signifi- sample was evaluated four times, and each had a single cantly decreased under the treatment with 20 lM inhibitor, peak of the melting curve produced at an acquired fluores- indicating a toxic effect, but there were no significant cence and melting temperature. The expression value of b- changes in the other groups (suggesting no significant toxic actin in each of examined brain sample was used as a nor- effect). Thus, the dose of 10 lM inhibitor for the ErbB fam- malization control: b-actin expression did not show signifi- ily was chosen. After culturing for 1 day at 378C and 5% cant differences between the male and female samples CO2, the explants were fixed, dehydrated and cut into 10- collected as described above, including the brain samples lm-thick sections according to the above described proto- containing HVC (t 5 3.25, p 5 0.41, n 5 5) and not contain- col. Every sixth sagittal section was mounted on a poly- ing HVC (t 5 2.33, p 5 0.37, n 5 5). Using the 2-DDCt lysine-coated slide and used for immunohistochemistry method, the normalized expression level of erbin or erbb2 analysis. gene relative to b-actin was calculated in each of examined brain sample. Relative expression values of erbin or erbb2 Construction of Lentivirus-Interfering were expressed as fold change over any assigned male sam- ple (assigned expression value 5 1.0). Vectors, Injections of Lentivirus- Interfering Vectors, Song Recording and Neural Tract Tracing Brain Slice Culture and the Effect of the ErbB2 Inhibitor on Cultured Brain Slices To obtain an effective interference fragment, two or three different short hairpin RNA (shRNA) fragments were After the nestling male Bengalese finches (PHD 15) were designed to target the erbin (XM_004177159.1) or erbb2 anesthetized, they were perfused with ice-cold HBSS con- (KT932702) mRNA of the Bengalese finch, respectively. taining 1% penicillin/streptomycin (Shanghai Weike) and These shRNA fragments were designed online (http:// 5.5 g/L D-glucose, which was sterilized by filtration using a www.genelink.com/sirna/shRNAi.asp). The following lin- 0.22-micron membrane. The explants that contained the ear DNA structure was used to encode the shRNA hairpins: developing HVC and its overlying VZ were obtained using sense-loop-antisense-TTTTT. The sequence of the loop the above-mentioned protocol. For each hemisphere, 2–3 was TTCAAGAGA. The constructs for the expression of explants containing HVC were obtained. The explants were the shRNA were cloned into a GU11 vector (pBS-U6- then transferred into cold HBSS, incubated for 90 min at RNAi-dual-CMV-GFP) in which the U6 promoter was 48C, and placed on 25-mm porous culture with a 0.4-lm used to drive expression of the shRNA. Each hairpin con- pore size (NUCN, 137060) inserted in a 6-well plate with struct was tested for knockdown efficiency in HEK293 cells six explants per well (1.8 ml Dulbecco’s Modified Eagle that were co-transfected with an erbin or erbb2 expression Media: Nutrient Mixture F-12 in a 1:1 ratio (Sigma), 1% plasmid and an shRNA GU11 vector. Here, we used the (v/v) penicillin/streptomycin, 1% (v/v) N2 (Sigma), 2% (v/ same fragments as in the in situ hybridization protocol, v) B27 (Sigma), 2 mM glutamine (Gibco) and 1% VC (v/v) which were not expressed in un-transfected HEK293 cells. (Sigma)). A highly selective inhibitor for the ErbB family The hairpins that had the best interference efficiency in

Developmental Neurobiology 20 Zhao et al.

HEK293 cells were synthesized (shErbin targeting cage of the same size as the housing cages. The songs of sequence: 50-CCTGAAGGACCAGCATCTA-30; shErbB2 each male were recorded using SAP software (Sound Anal- targeting sequence: 50-CCTGGCCACCTGACATGAA-30) ysis Pro 2011) and a condenser microphone (Behringer C- and inserted into a GV248 vector (phU6-MCS-Ubiquitin- 2) between 7:00 am and 7:00 pm after the bird had adapted EGFP-IRES-puromycin, Shanghai GeneChem) to generate to the new cage and began to sing. Recording was contin- a recombinant RNAi virus. The expression of shRNA for ued for at least 100 bouts for each bird. the target gene was driven by the human U6 promoter and After the song recording, the adult birds received the the enhanced green fluorescent protein (EGFP) gene using injections of a total of 500 nl retrograde tracer (dextran the Ubiquitin promoter. The recombinant RNAi virus was conjugated Alexa Fluor 647, 10,000 MW; Molecular packaged using the Lentivector Expression System (Shang- Probes; 5% solutions made in 0.01 M PBS) into the RA at hai GeneChem). The scramble sequence (50- four sites (medial/lateral: 2.2 mm, anterior/posterior: 0.1 TTCTCCGAACGTGTCACGT-30) proposed and provided and 0.4 mm, dorsal/ventral: 2.2 and 2.6 mm) in both hemi- by Shanghai GeneChem was used as a negative control. spheres using the same methods (including anesthetic 8 The titers were 6–8 3 10 TU/ml. method) as described for the injection of lentiviruses. The injection of lentiviral solution was referred to a pre- vious report (Haesler et al., 2007). Before surgery, the nes- tling birds (PHD 15) were deprived of food and water for Immunohistochemistry and Song one hour. They were then anaesthetized via intramuscular Analysis injections of sodium pentobarbital as described above, and The birds were deeply anesthetized, perfused and fixed fixed in a stereotaxic head holder. Lidocaine (2%) was using the above-described protocol. The brains were cut injected locally at the incision site over the bifurcation of into 10-lm-thick sections, and every sixth sagittal section the mid-sagittal sinus. The scalp was dissected along the was mounted on a poly-lysine-coated slide. For the BrdU midline, and the skull was opened by a small fenestra. The immunohistochemistry, the sections were incubated with 2 stereotaxic point 0.0 was defined as the branch point of the N HCl for 3 h and then incubated with Borax buffer sagittal sinus. To keep the developing HVC and its overly- (0.1 M, pH 8.4) for 30 min. All sections that were used for ing VZ to be well infected with lentiviral viruses, the lenti- immunohistochemistry were first incubated in 3% H O for viral solution was injected at four sites (medial to lateral: 2 2 15 min to quench the endogenous peroxidase activities (this 2.0 mm; anterior to posterior: 0.1 and 0.3 mm; dorsal to step was neglected for immunofluorescent staining) and ventral: 0.4 mm (to infect HVC) and 0.2 mm (to infect the then in a permeabilization/blocking buffer (5% normal don- VZ overlying HVC), by using a glass micropipette (internal key, rabbit or goat serum that was soluble in TritonX-100/ tip diameter, 15–25 lm) that was attached to a glass PBS) for 40 min at room temperature. The sections were syringe. To keep the same injection number of the three then incubated with the rat anti-BrdU antibody (AbD Sero- types of lentiviral viruses (erbin RNAi, erbb2 RNAi or neg- tec, 1:2000) in blocking buffer overnight at 48C. After the ative control) (TU/ml*Volume5 3 3 105) into a hemi- sections were washed, they were incubated with either fluo- sphere, 100–125 nl lentiviral solution were injected into rescent or biotinylated secondary antibody: Texas Red- each site over a course of 10 min. Such injections could cause the developing HVC and its overlying VZ to be well conjugated donkey anti-rat (Jackson; 1:400), Alexa Fluor infected with lentiviral viruses [Fig. 4(A1-B4)], and the 350-conjugated donkey anti-rat (Jackson, 1:600), or Biotin- infected cells were distributed around the injected areas. conjugated rabbit anti-rat (ZSGB-Bio, 1:400). For those For each of the birds, the recombinant viruses (erbin RNAi sections that required double labeling, the sections were or erbb2 RNAi) were injected into the developing HVC at incubated with additional primary antibodies (mouse anti- either side (right or left) of the hemisphere, while the con- HuC/D (molecular probes, 1:600), mouse anti-NeuN trol viruses (control RNAi) were injected into the other (Chemicon, 1:300), or rabbit anti-c-aminobutyric acid hemisphere. In the male birds that were used to study the (cGABA; sigma, 1:1000)) and an additional fluorescent effects of the lentiviral-mediated RNAi in song behavior, secondary antibody (Alexa Fluor 647-conjugated donkey each bird received an injection of the recombinant viruses anti-mouse (Jackson; 1:600); Alexa Fluor 594-conjugated (erbin RNAi or erbb2 RNAi) or the control viruses in both goat anti-mouse or rabbit (molecular probes, 1:400) accord- hemispheres of the brain. ing to the protocol described above. Those incubated with After recovery from the surgery, the birds were biotin-conjugated biotinylated secondary antibody were returned to the cage and lived with their foster parents. then reacted with ABC (Vector Labs; 1:200 in 0.5% Tri- On the seventh day following surgery, they received tonX-100/PBS) and visualized using the DAB reaction. intramuscular injection of BrdU (50 lg/g body weight in The specificity of antibodies in the brain of songbird has 0.75% NaCl; Sigma) twice daily (8:30 am and 6:30 pm) been verified in previous reports, including anti-HuC/HuD for two successive days to label the dividing cells. These (Louissaint et al., 2002), anti-NeuN (Scott and Lois, 2007; birds were killed at 24 h (PHD 24), 7 d (PHD 31) or Zeng et al., 2009), and anti-cGABA (Pinaud and Mello, adulthood (PHD 130) after the final injection of BrdU 2007). The antibody specificity was also confirmed from [Fig. 2(E)]. our staining patterns which were similar to related reports The song recording was performed in a small soundproof (Pinaud and Mello, 2007; Chen et al., 2014). To exclude room, and a single male bird (PHD > 120) was placed in a nonspecific staining, the primary or secondary antibody

Developmental Neurobiology Erbin and ErbB2 in the Sexual Differentiation of HVC 21

Figure 2 Effect of an ErbB family inhibitor (4-(4-benzyloxyanilino)-6,7-dimethoxyquinazoline) on cell proliferation within the VZ, interfering rates of lentiviruses and the protocol used in the in vivo study. A,B, BrdU-labeling within the cultured male VZ, 24 hours after addition of BrdU and an ErbB family inhibitor (10 lM) or DMSO (for control) into the culture medium at PHD 15. Scale bar in B 5100 lm for A and B, and 50 lm for inserts. C, A significant decrease in the number of BrdU-positive cells (per mm) within the VZ that just overlies the developing HVC. D, Quantitative real-time PCR indicates significant decreases in expression of erbin and erbb2 mRNA, seven days after injection of interfering lentiviruses into the male HVC at PHD 15 in contrast with injection of control lentiviruses. b-actin was used as an internal control to normalize the values for each reac- tion, and the gene expression level was quantified relative to one male sample receiving an injec- tion of interfering lentiviruses. **p < 0.01; ***p < 0.001. Error bars indicate SEM. E, In vivo study included short-, middle- and long-time survival groups. [Color figure can be viewed at wileyonline- library.com] was omitted, but all other immunohistochemical procedures (2) the phrase (or song motif) that appears repeatedly in dif- were the same as those described above. No staining was ferent songs and composed of serial syllables that always observed in these sections. appear in a relatively fixed sequential order, and (3) the The recorded songs were quantitatively analyzed using syllable (or note) that appears as a continuous, morphologi- SASLab Pro sound analysis software (Avisoft, Inc.) at the cally discrete trace on the sonogram (Catchpole and Slater, level of (1) the song produced repeatedly during a bout(s), 2008).

Developmental Neurobiology 22 Zhao et al. cDNA Microarray Analysis Affymetrix Lund-zf array. We also considered the follow- ing in our choice of microarray: 1) avian gene order, Four male or female brain samples that contained the devel- genome size, and chromosomal arrangements are highly oping HVC and its adjacent parenchyma were obtained as conserved among avian species (Derjusheva et al., 2004; described above [Fig. 1(E)] and combined for RNA Backstrom€ et al., 2006; Dawson et al., 2007), and 2) blast extraction. analyses indicate that the similarity of EST sequence was RNA quality and purity were assessed as described 89.5% between zebra finch and chicken and 92.8% between above. RNA was then reverse transcribed and labeled with zebra finch and house finch (Replogle et al., 2008). either a Cy3 (male) or a Cy5 (female) fluorochrome (GE healthcare Biosciences). The detailed methods for labeling using the two complementary dyes (Cy3 or Cy5) were Image Acquisition, Cell Counting and described in a previous report (Li et al., 2008). Four Statistical Analyses matched male and female sample pairs were hybridized Bright-field images were captured with a digital camera onto the Agilent Whole Chicken 60-mer oligonucleotide (Spot Enhance 2e; Diagnostic Instrument, Corp., USA) 44K microarray (Agilent Technologies). Each pair included attached to a BH-2 microscope (Olympus), and fluorescent RNA extracted from four male or female brains of nestling images were captured with a monochromatic digital camera birds at PHD 15. Each microarray contained 42,034 probes (AxioCam MRm, Zeiss). Adobe Photoshop was used to that were designed from the whole chicken transcriptome process TIFF files. All labeled cells in the examined areas (the February 2004 chicken v1.0 draft assembly), 150 were counted manually using Image J 1.44 software (NIH microRNAs (miRBase 8.1), the entire genome sequences of Image program). Marek’s disease virus, and two avian influenza virus The boundaries of song nuclei were outlined, and their genomes (Wallis et al., 2004; Griffiths-Jones, 2004; areas were obtained using Image-Pro Plus 5.2. The nuclei Griffiths-Jones et al., 2006). volumes were calculated by summing the areas in the series The slides were hybridized overnight at 428C using of sections in which song nuclei appeared and multiplying hybridization buffer (Ambion/Applied Biosystems) and by six (sampling interval) and 10 lm (section thickness). washed with a series of standard saline citrate solutions. Some examined measurements, such as the labeling for After the slides were centrifuged to dryness, all slide erbin and erbb2 mRNA in song nuclei, were assessed by images were scanned using an Agilent Microarray Scanner the total numbers of positive cells in the song nuclei, which and analyzed using Feature Extraction software 10.7 (Agi- were obtained using the densities of the labeled cells (the lent Technologies). The raw data were normalized using ratios of the total numbers of positive cells to the sizes of the Lowess (locally weighted scatter plot smoothing) algo- their examined areas, numbers/mm3) multiplied by the rithm, and log-transformed microarray data were analyzed nuclei volumes (mm3). with the Statistical t test after adjustments for multiple Song nuclei are generally characterized by some ana- comparisons. Biological processes and molecular functions tomical features including the relative sizes of cells and the were identified for the sex-biased genes via Gene Ontology density of cells to their adjacent areas. The outlines of song (GO; http://www.geneontology.org/). All sex-biased genes control nuclei studied in the present report could readily be found on chromosome Z were mapped along the chicken Z distinguished in Nissl-stained sections, including the devel- chromosome to determine how they were distributed oping HVC at PHD 15. According to our previous study to (http://www.ncbi.nlm.nih.gov/gene/). inject a neural tract tracer (Dil) into the Area X in 600-lm Only two types of microarrays were available for avian sagittal brain slices at posthatch day 15 (Chen et al., 2014), species: zebra finch (Affymetrix Lund-zf array), which con- the sizes of HVC determined by Area X-projecting cells tains approximately 15,800 genes with a 25-bp-long probe were largely consistent with those obtained by Nissl- (Naurin et al., 2008), and chicken, which contains approxi- staining. The borders of HVC were thus determined with mately 28,000 genes (covering almost all protein-coding the aid of Nissl-staining in the present study. However, the genes in the genome) with longer probes (60 bp) (Ellegren boundaries of some song nuclei, including male/female et al., 2007). Although Affymetrix arrays for the zebra finch LMAN and female Area X at PHD 15 or HVC at earlier have been successfully used in several oscine species, we ages than PHD 15, could not be clearly identified in Nissl- noted that most previously reported genes that were sexu- stained sections. The volumes of these nuclei were thus not ally differentially expressed in oscine brains, including obtained. These nuclei were not included in this report. For erbin and erbb2, were not detected out (Wade et al., 2004; in situ hybridization sections for erbin and erbb2 mRNA, Li et al., 2008; Tomaszycki et al., 2009; Sun et al., 2010; the boundaries of some song nuclei such as female HVC Naurin et al., 2011). This is probably due to the shorter were not clearly identified, and the positions or sizes of gene probes and lower number of genes contained in the these nuclei were determined by referencing those in the arrays for zebra finch compared to those for chicken (Li interval sections stained for cresyl violet. et al., 2008). Considering that our main goal in using The present study chose 15 days old birds as subjects, as microarray analysis was to determine the genes that are this age is within the critical period for sexual differentia- expressed in a sex-biased fashion in HVC, especially those tion, which is generally during the first few posthatch in erbin- or erbb2-mediated pathways, we did not use the weeks, and not beyond 30 days after hatching (Konish and

Developmental Neurobiology Erbin and ErbB2 in the Sexual Differentiation of HVC 23

Akutagawa, 1988). At the age of 15 days, a significant However, the labeling was negative for the sense number of HVC cellular cohorts were reported to generate probes of erbin- or erbb2-mRNA [Fig. 1(F,G)]. At (Alvarez-Buylla et al., 1994; Burek et al., 1994). They are PHD 15, some weakly stained small cells were incorporated into HVC 1–3 weeks after they are produced observed to be scarcely distributed along the VZ within the VZ (Burek et al., 1994; Kirn et al., 1999). 1 overlying HVC [VZ is pointed by arrow heads, Fig. Once the outline of the HVC was determined, BrdU 1(A2-B3)]. The adult group was very similar to that cells along the VZ overlying the HVC (approximately at PHD 45 and the labeling is thus not shown in Fig- 500 lm for males, and 300 lm for females) were counted at PHD 15 and expressed as the number of proliferating ure 1. The labeled cells for erbin or erbb2 mRNA BrdU1 cells per mm. During the analyses, the experiment- within HVC were counted, and the density or the ers were blind to sex and experimental treatments. total number of the labeled cells were obtained and The birds that did not survive to the scheduled ages compared among the studied groups (the sizes of or whose injection sites were outside the targeted areas female HVC were determined with reference to those were excluded from the study (not included in the sample in the interval sections for Nissl-staining, and the data). data for the density were not shown). By using two- Statistical analyses were performed using the SPSS 11.5 way ANOVA with sex and age as factors, there was software package. Two-way ANOVA was used to examine significant interaction between the effects of sex and the effect of gender and different ages, and one-way age on the total number of the labeled cells for erbin ANOVA was conducted to compare the differences among mRNA and erbb2 mRNA within HVC [erbin mRNA: the groups exposed to several treatments. Student’s t-test F 5 35.45, p < 0.001; erbb2 mRNA: F was used to compare the differences between the two (2, 13) (2, 5 41.76, p < 0.001]. Differences in the total num- groups under the same experimental conditions. Before 14) ANOVA, the distributions of dependent variables were ber of the labeled cells in HVC were significant tested for normality, and homogeneity of variance was between the two sexes [erbin mRNA: t 5 50.34, assessed for equality of error variance (Levene’s test). The p < 0.001; erbb2 mRNA: t 5 56.71, p < 0.001; Fig. data were presented as the mean 6 SEM. Differences 1(H,I)]. between the groups of data were considered either nonsig- We also examined the densities (cell numbers per nificant (ns, p > 0.05) or significant at various levels (i.e., mm3) of the labeled cells for erbin mRNA and erbb2 *p < 0.05, **p < 0.01 and ***p < 0.001). mRNA in the area (200 lm 3 200 lm) just under- Because antibiotic damage followed a proximal-to-distal neath the ventral bottom of HVC in each section con- sequence, nearly all hair cells were damaged in the proxi- taining HVC, and no significant changes were found mal part at a relatively earlier time (within 24 h after genta- between the males and females (n 5 5 for each sex) micin treatment) than those in the distal part. We only at PHD 15 [erbin: males: 90.0 6 16.2/mm3; females: compared the number of regenerated hair cells (PV1), 80.8 6 8.47/mm3; t 5 1.13, p 5 0.29; erbb2: males: mitotic cells (BrdU1) and differentiated cells (BrdU1/ 3 3 Sox2 or BrdU1/Myosin 6a) in the proximal region among 111.2 6 22.6/mm ; females: 95.2 6 19.0/mm ; the studied groups. At least three nonoverlapping fields t 5 1.28, p 5 0.26] and PHD 45 [erbin: males: 3 3 (150 3 150 lm) were captured in the middle region of the 71.2 6 13.3/mm ; females: 67.8 6 10.17/mm ; 3 proximal part (the whole-mount cochlea was divided t 5 2.24, p 5 0.13; erbb2: males: 93.6 6 25.6/mm ; equally into proximal and distal parts) under a 403 oil females: 84.8 6 21.7/mm3; t 5 2.59, p 5 0.36]. objective. The data for labeling density were averaged for Using qRT-PCR, we compared sex differences in each cochlea. One-way ANOVA with Tukey’s post hoc test the expression levels of erbin- or erbb2-mRNA in the was used to compare the differences in the numbers of brain slice samples that contained the developing regenerating hair cells in the proximal regions among the HVC and its overlying VZ, and the areas adjacent to studied groups. Student’s t-test was used to compare the HVC (but not contain HVC), the long narrow brain differences between groups. The significance was set at belts at PHD 15 [approximately 300 lm wide and p < 0.05. just under the above collected brain samples contain- ing HVC, Fig. 1(E), delimited by two dashed lines]. RESULTS The results indicated that erbin and erbb2 mRNA expression was significantly higher in the collected In Situ Hybridization for Erbin and erbb2 brain samples that contained the male HVC than in and qRT-PCR in the HVC of Bengalese the collected brain samples that contained the female Finch HVC [for erbin, t 5 7.54, p 5 0.012; for erbb2, t 5 6.17, p 5 0.031], or than in the brain belts that did In situ hybridization indicated that erbin- or erbb2- not contain male or female HVC [for erbin, t 5 6.35, mRNA-positive cells were both located within the p 5 0.027; for erbb2, t 5 5.98, p 5 0.039]. There HVC in a male-biased way [Fig. 1(A2-D3)]. were no significant differences in the expression

Developmental Neurobiology 24 Zhao et al. levels of erbin- or erbb2-mRNA between the male types of lentiviruses had similar infection activities and female brain belts that did not contain HVC [for within the HVC. The schematic of studies performed erbin, t 5 0.86, p 5 0.754; for erbb2, t 5 0.64, after the injections of erbin-orerbb2-interfering len- p 5 0.351; Fig. 1(J,K)]. tiviruses into the male HVC at PHD 15 is shown in Figure 2E. Effects of ErbB Family Inhibitor on Cell Proliferation in the VZ Overlying the Effects of Erbin or erbb2 Knock-down on Developing HVC Cell Proliferation at PHD 24 Following the addition of the ErbB family inhibitor BrdU1 cells along the VZ overlying the developing 4-(4-benzyloxyanilino)-6,7-dimethoxyquinazoline, HVC were examined 24 h after the final injection of and BrdU into the culture medium (both 10 lM) at BrdU (at PHD 24) in birds previously injected with PHD 15, the number of proliferating cells (BrdU1) lentiviruses with the interfering or control constructs per mm in the VZ that overlies the developing male [as in Fig. 2(E)]. At this time point, the proliferating HVC was examined 24 h after inhibitor addition. The cells (BrdU1 labeled) do not migrate out of the ven- number of BrdU1 cells per mm decreased by 63.11% tricular zone and do not begin to differentiate (Chen compared to that observed for the culture medium et al., 2014). As shown in Figure 4A1-B4, the cells that contained only DMSO [t 5 34.61, p < 0.001; Fig. infected with lentiviruses (GFP1) were distributed 2(A-C)]. within the HVC, as well as its overlying VZ. How- ever, no GFP1 cells were found outside the injected 1 Effects of Erbin or erbb2 Knock-down on areas. The number of BrdU cells (per mm) along the Sexual Differentiation of the HVC and the VZ overlying the developing HVC significantly Adult Birdsongs decreased in males [Fig. 4(A1-B2), 4(C); erbin: t 5 8.72, p 5 0.007, erbb2: t 5 5.11, p 5 0.02] or Cultured brain slices have several limitations, includ- females [Fig. 4(A3-B4), 4(D); erbin: t 5 6.79, ing a short survival time (not beyond 7 days) and an p 5 0.01, erbb2: t 5 10.82, p 5 0.005] after injection inability to represent the in vivo brain completely. To of erbin-orerbb2-interfering lentiviruses into the resolve these problems, we constructed erbin- and HVC at PHD 15, compared to the injection of control erbb2-interfering lentiviruses, both of which con- lentiviruses. To assess the effect scope of lentivirus tained the enhanced green fluorescent protein (EGFP) injection to cell proliferation, we counted the num- reporter gene downstream of the interfering fragment. bers of BrdU1 cells along the VZ (for approximately Using qRT-PCR, we observed significant changes in 500 lm long, in which GFP1 cells were not seen) the interference rates (the decreased percentage of next to the injection regions (GFP1, overlying the the target mRNA) 7 days after erbin-orerbb2-inter- developing HVC) in males, and found that the num- fering lentiviruses were injected into the male HVC bers (cells per mm) did not differ significantly at PHD 15, compared to injections of control lentivi- between the groups receiving injections of erbin-or ruses [for erbin RNAi, the interference erbb2-interfering lentiviruses and the groups receiv- rates 5 46.03 6 5.78%, t 5 13.58, p 5 0.001; for ing injection of control lentiviruses [erbin vs control erbb2 RNAi, the interference rates 5 54.86 6 4.76%, RNAi: n 5 5, t 5 0.58, p 5 0.43; erbb2 vs control t 5 9.75, p 5 0.009; Fig. 2(D)]. As ERBIN or ERBB2 RNAi: n 5 5, t 5 0.78, p 5 0.51]. antibodies were not available for avian species, the changes in protein levels were not examined. In addi- Effects of Erbin or erbb2 Knock-down on tion, EGFP could be detected in all experimental Early Neuronal Differentiation at PHD 31 groups, including the adult group [PHD 130, Fig. 7], which indicated that lentivirus-mediated RNAi could The numbers of BrdU-labeled, Hu-labeled (Hu is a be persistently produced in the whole HVC (not only marker expressed in early differentiated neurons, limited to the adjacent areas of injection sites) in all Barami et al., 1995) or BrdU and Hu double-labeled the studied groups. In addition, we compared the den- cells were examined in the HVC at PHD 31 (7 d after sity of EGFP cells (per mm2) in the male HVC at the final injection of BrdU). As the HVC is quite PHD 24 (9 days after injection of erbin, erbb2 or con- small in female Bengalese finches, the following trol lentiviruses), and there were no significant differ- inhibition studies were only performed in males. As ences among the three groups (erbin, erbb2 and shown in Figure 5A1-B20, the cells double-labeled control RNAi) [n 5 5 for each group, F(2,12)53.78, for BrdU and Hu were observed in the developing p 5 0.43; Fig. 3]. These data suggested that different HVC after the injection of erbin or erbb2-interfering

Developmental Neurobiology Erbin and ErbB2 in the Sexual Differentiation of HVC 25

Figure 3 Density of infected cells (per mm2) in the developing HVC, nine days after injection of control, erbin or erbb2 lentiviruses into the HVC at PHD 15. A-C, Infected cells in groups receiving injection of control RNAi (A), erbin RNAi (B) and erbb2 RNAi (C). D, No significance (NS) in the density of infected cells in the developing HVC among the three groups. Scale bar in C 5 100 lm for A-C. Bars are mean 6SEM. [Color figure can be viewed at wileyonlinelibrary.com] lentiviruses into the HVC. When compared to the lentiviruses [n 5 4 for each group; for erbin vs 1 1 control groups, a significant decrease was observed control RNAi, BrdU: t 5 0.32, p 5 0.37; Hu : in the total numbers of BrdU1,Hu1 or BrdU1/Hu1 t 5 0.55, p 5 0.45; BrdU1/Hu1: t 5 0.72, p 5 0.57. 1 1 cells within the HVC [for erbin RNAi, BrdU : for erbb2 vs control RNAi: BrdU: t 5 0.51, t 5 43.87, p < 0.001. Hu1: t 5 23.11, p 5 0.024. p 5 0.49; Hu1: t 5 0.89, p 5 0.61; BrdU1/Hu1: BrdU1/Hu1: t 5 25.98, p 5 0.031. Fig. 5(A1-A20), t 5 0.91, p 5 0.77]. 5(D-F); for erbb2 RNAi, BrdU1: t 5 19.24, 5 1 5 < 1 1 p 0.035. Hu : t 37.65, p 0.001;BrdU /Hu : Effects of Erbin or erbb2 Knock down on 5 5 0 t 20.06, p 0.027. Fig. 5(B1-B2 ), 5(D-F)]. We Adult HVC and Birdsongs also examined the densities (cell numbers per mm3) of BrdU1,Hu1 or BrdU1/Hu1 cells in the area (200 As in the zebra finch, song nuclei and song behavior lm 3 200 lm) just under the bottom of the HVC (no are mature after PHD 120 in the Bengalese finch GFP1 cells were seen) in the groups receiving injections (Okanoya and Yamaguchi, 1997; Tobari et al., 2005). of erbin-orerbb2-interfering lentiviruses [for erbin- To study the effect of erbin or erbb2 knock-down on 1 4 3 1 RNAi, BrdU: 2.31 6 0.46 3 10 /mm ;Hu:7.65 6 the adult HVC sizes and birdsongs, we examined 1.98 3 104/mm3;BrdU1/Hu1:1.326 0.65 3 104/mm3; changes in the sizes of the HVC and one of its target- 1 4 3 for erbb2-RNAi, BrdU: 2.87 6 0.89 3 10 /mm ; ing nucleus-RA, the number of BrdU-labeled cells Hu1:7.13 6 1.75 3 104/mm3;BrdU1/Hu1:1.876 0.64 that also project to the RA and the number of double- 3 104/mm3] and the groups receiving control lentivi- labeled cells for NeuN and BrdU or GABA and BrdU 1 4 3 1 ruses [BrdU: 2.65 6 0.79 3 10 /mm ;Hu :7.44 6 1.79 in the HVC at PHD 130. 3 104/mm3;BrdU1/Hu1:1.756 0.56 3 104/mm3]. No After the injection of erbin-orerbb2-interfering significant changes were found between the groups lentiviruses into the HVC at PHD 15, the volumes of receiving injections of erbin-orerbb2-interfering the HVC significantly decreased by 40.74% for erbin lentiviruses and the groups receiving control RNAi [t 5 35.47, p < 0.001] and 49.42% for erbb2

Developmental Neurobiology 26 Zhao et al.

Figure 4 Cell proliferation within the VZ that overlies the developing HVC nine days after injec- tion of erbinor erbb2-interfering lentiviruses into the HVC at PHD 15. A1-B4, BrdU-positive cells (magenta) in VZ overlying the developing HVC. Green fluorescent protein (GFP) was produced by erbin- or erbb2-interfering lentiviruses. Scale bar in B4 5200 lm for A1-B4 and 100 lm for the amplified inserts. C, D, Significant decreases in the number of BrdU-positive cells (per mm) along the VZ that overlies the developing HVC in males (C) and females (D). *p < 0.05; **p < 0.01. Error bars indicate SEM. [Color figure can be viewed at wileyonlinelibrary.com]

RNAi [t 5 20.41, p 5 0.008] in contrast to the control significantly after injection of erbin- or erbb2-inter- groups (both decreases in the sizes of HVC sections fering lentiviruses into the HVC compared to the con- and the lengths along medial-lateral axis) [Fig. 6(A1- trol groups [erbin: t 5 14.54, p 5 0.007; erbb2: D1)]. Accordingly, the volumes of the RA were also t 5 9.32, p 5 0.03; Fig. 7(F)]. In addition, following significantly reduced following the injection of erbin the injection of erbin- or erbb2-interfering lentivi- or erbb2 interfering lentiviruses into the HVC [erbin ruses into the male HVC, we compared the total num- RNAi: t 5 8.21, p 5 0.02; erbb2 RNAi: t 5 13.78, ber [Fig. 7G,H] and the density (cell numbers per p 5 0.004. Fig. 6(A2-D2)]. mm3) of NeuN1/BrdU1 and GABA1/BrdU1 cells Following a neuronal tracer (dextran conjugated with those in the groups receiving control lentiviruses Alexa Fluor 647) injection into RA in which the neu- [for NeuN1/BrdU1 density, erbin RNAi: ronal tracer-filled cells were distributed in most of 1262 6 103.12; control RNAi: 2364 6 184.86; erbb2 RA (>2/3) [Fig. 7D] (if not covered most of RA, the RNAi: 1700 6 153.09; control RNAi: 2481 6 175.16; 1 1 cases were not included), HVCRA-projecting neurons GABA /BrdU density, erbin RNAi: 205 6 25.03; (HVCRA PNs) were observed in the HVC [Fig. 7(A1- control RNAi: 254 6 22.67; erbb2 RNAi: A3, B1-B2, C1-C2)]. Some of these neurons were 160 6 9.75; control RNAi: 179 6 13.41]. There were also labeled with BrdU (injected at PHD 22 and 23). significant decreases in the total number or density of We compared the percentages of BrdU-labeled cells NeuN1/BrdU1 [for total number, erbin: t 5 48.72, that also project to the RA (double-labeled cells) to p < 0.001; erbb2: t 5 36.78, p < 0.001; for density, the total number of retrograde-labeled RA-projecting erbin: t 5 76.79, p < 0.001; erbb2: t 5 68.45, cells in the HVC and found that they decreased p < 0.001; Fig. 7(G)] or in the total number, but not

Developmental Neurobiology Erbin and ErbB2 in the Sexual Differentiation of HVC 27

Figure 5 Cell differentiation within the developing HVC 16 days after injection of erbin- or erbb2-interfering lentiviruses into the HVC at PHD 15. A1-B20, Hu-positive neurons (blue) and BrdU-positive cells (red) in the developing HVC. Green fluorescent protein (GFP) was produced by the erbin- or erbb2-interfering lentiviruses. Double-labeled cells (Hu1 and BrdU1) are indi- cated by arrows in A10-B20. Scale bar in C 5100 lm for A1-B2 and C, 50 lm for A10-B20 and 25 lm for the amplified images in A10-B20. C, Nissl-stained HVC in a section adjacent to A1. The bor- der of the HVC is shown in dashed lines. D-F, Significant decreases in the total number of BrdU (D), Hu (E) and double-labeled cells for BrdU and Hu (F) in the developing HVC. *p < 0.05; **p < 0.01. Error bars indicate SEM. [Color figure can be viewed at wileyonlinelibrary.com] in the density, of GABA1/BrdU1 cells [for total group). However, three of these died from inadequate number, erbin: t 5 7.23, p 5 0.02; erbb2: t 5 29.76, surgical operations (significant blood loss), and four p < 0.001; for density, erbin: t 5 3.45, p 5 0.43; birds did not survive to the scheduled age (unex- erbb2: t 5 2.79, p 5 0.67. Fig. 7(H)] in the HVC, pected death). We analyzed at least 25 songs (com- compared to the groups receiving control lentiviruses. plete typical songs as described above for control For normally reared Bengalese finches, a typical birds, and relatively complete songs for treated birds complete song generally began with “introductory” with interfering lentiviruses). These birds included notes, followed by two phrases “A” and “B”, and normally reared birds (n 5 5), birds receiving the ended with another phrase “C” (phrases “A”, “B,” or injection of control lentiviruses (n 5 3), erbin- “C” each consist of syllables “a”, “b” or “c”, respec- (n 5 3) or erbb2- (n 5 2) interfering lentiviruses at tively) [Fig. 8(A)]. Introductory (i) notes and sylla- PHD 15. Analyzing more songs (>25) did not alter bles “a”, “b” or “c” could be distinguished on the the results significantly for each studied bird. The sonogram on the basis of various sound parameters, numbers of repeating syllables “a”, “b,” or “c” are including duration (“i”< “a”< “b”< “c”), harmonic shown in Table 2. In all the birds that received injec- waves (two main frequencies in “a”, more than two tions of control lentiviruses, song organization was frequencies in “b”, no dominant frequency in “c”), similar to that observed in the normally reared birds and frequency modulation (frequencies in “i” or “a” [Fig. 8(A,B)]. Additionally, the numbers of repeating are highly modulated). syllables “a”, “b,” or “c” were within the range of The total number of birds originally used for the normal birds. These data indicated that the surgical analysis of song behavior was 20 (n 5 5 for each operation and control lentiviruses did not

Developmental Neurobiology 28 Zhao et al.

Figure 6 Changes in the sizes of adult HVC and RA at PHD 130 after injection of erbin- or erbb2-interfering lentiviruses into the developing HVC at PHD 15. A1-C2 Nissl-stained HVC (A1- C1) and RA (A2-C2) from the male birds receiving injections of control lentivirus (A1 and A2), erbin-interfering lentivirus (B1 and B2) or erbb2-interfering lentivirus (C1 and C2). The dashed lines indicate the outline of the HVC or RA. Scale bar in C2 5 100 lm for A1-C2. (D1, D2), Signif- icant decreases in HVC (D1) or RA (D2) volumes. **p < 0.01; ***p < 0.001. Error bars indicate SEM. [Color figure can be viewed at wileyonlinelibrary.com] substantially affect song learning and production. protein-coding genes in the chicken genome), to However, no audible songs were heard from two examine the sex-biased genes during the sex differen- birds that received an injection of erbin- or erbb2- tiation of HVC, especially those in erbin- or erbb2- interfering lentiviruses, respectively. Song organiza- related signaling pathways. A total of 19,996 genes tion was abnormal in the other three birds that (including 569 on chromosome Z) were expressed in received the injection of erbin (n 5 2)- or erbb2 the developing HVC of nestling Bengalese finches at (n 5 1)- interfering lentiviruses, compared to the nor- PHD 15. Of these, 2,098 genes were found to be sig- mally reared birds or birds receiving control lentivi- nificantly sex-biased (p < 0.05, one sample t-test, ruses [erbin RNAi, Fig. 8(C); erbb2 RNAi, Fig. SPSS), including 1,672 genes that were located on 8(D)]. These abnormalities were similar among the chromosomes. These genes are summarized in Table three birds, including the deletion of “i” notes and the 3 with different fold changes (FC) in expression level syllable “a” (lost in all the recorded songs) as well as (FC >1.3, 1.5 or 2). A total of 103 sex-biased genes a decrease in the number of repeating syllables “b” or were mapped onto the chicken Z chromosome and “c” in both erbin- and erbb2- RNAi groups [Fig. did not show obvious valley or peak distributions 8(C,D)]. Before statistical analysis, the data from the along the chromosome [Fig. 9(A)]. above three birds were combined and compared with The distribution of 367 sex-biased genes the control lentivirus group. As shown in Table 2, the (FC > 1.5; males: 173, and females: 194) in the chro- numbers of syllables “b” and “c” both significantly mosome is shown in Figure 9B. Overall, 27.17% of decreased after the injection of erbin- or erbb2-inter- the male-biased genes and 5.15% of the female- fering lentiviruses into HVC at PHD 15 [for syllable biased genes were found on chromosome Z [Fig. “b”, t 5 27.42, p 5 0.02; for syllable “c”, t 5 18.92, 9(C)]. Of the sex-biased genes (FC > 1.5), 35 genes p 5 0.03]. were classified into the category of “metabolic proc- ess” (i.e., protein metabolic process, RNA biosyn- thetic process and small molecule metabolic cDNA Microarray to Identify Sex-Biased process), and 31 genes were classified into a “single- Genes in the Developing HVC of the organism process” (i.e., signal transduction, single- Bengalese Finch organism cellular process and system development) Finally, we used the Agilent Whole Chicken Genome [Fig. 9(D)]. In addition, 81 genes were classified into Oligo Microarray, which contains approximately the category of “binding” (i.e., ion binding, protein 40,000 probes or 28,000 genes (covering almost all binding and organic cyclic compound binding), 26

Developmental Neurobiology Erbin and ErbB2 in the Sexual Differentiation of HVC 29

Figure 7 Changes in double-labeled cells for BrdU and RA-projecting cells, BrdU and NeuN, and BrdU and GABA in the adult HVC at PHD 130 after injection of erbin- or erbb2-interfering lentiviruses into the developing HVC at PHD 15. A1-C1, BrdU-positive cells (white) and RA-projecting cells (HVCRAPNS, blue) in the adult HVC. Some RA-projecting cells were also labeled for BrdU (pointed by a small arrow). A2-C2, BrdU-positive cells (white), NeuN-positive neurons (red) and RA-projecting cells (blue) in the adult HVC. Cells that were labeled for NeuN and BrdU are indicated by small arrows. A3, BrdU-positive cells (white), GABA-positive neurons (red) and RA-projecting cells (blue) in the adult HVC. Green fluorescent protein (GFP) produced by the erbin- or erbb2-interfering lentivi- ruses can be observed in A1-A3. D, Injection site of neural retrograde tracer into RA. The dashed circle indicates the outline of RA, and the asterisk marks the approximate injection center. E, Schematic for the injection site of RA and retrograde-labeled cells in HVC. F-H, Changes in the percentage of cells labeled for both BrdU- and RA-projecting cells to RA-projecting cells (F), total numbers of cells labeled for both BrdU and NeuN (G) or GABA (H) in the adult HVC. *p < 0.05; **p < 0.01. ***p < 0.001. Error bars indicate SEM. Scale bar in D 5150 lm for A1-C1, 100 lm for A2-A3, 350 lm for D and 50 lm for inserts in A1-C1. [Color figure can be viewed at wileyonlinelibrary.com]

Developmental Neurobiology 30 Zhao et al.

Figure 8 Changes in adult sonograms after the injection of erbin- or erbb2-interfering lentiviruses into the developing HVC at PHD 15. A-D, Sonograms of a normally reared bird (A) a bird that received a control injection of interfering lentiviruses (B) and two birds that received an injection of erbin- (C) or erbb2-interfering lentiviruses (D). The normally reared birds and birds receiving a control injection of interfering lentiviruses have similar song organizations, starting with “introductory” (i) notes, sequentially followed by phrases “A”, “B” and “C”, which consist of sev- eral repeating syllables “a”, “b”, or “c”, respectively (A and B). “Introductory” notes and the phrase “A” were deleted, and the number of repeating syllables decreased significantly in birds receiving an injection of erbin- or erbb2-interfering lentiviruses (C and D). [Color figure can be viewed at wileyonlinelibrary.com] genes were classified into “catalytic activity”, and gene alterations were detected in the ErbB2- 8 genes were classified into “receptor activity”, interacting NF-jB pathway. These altered genes, according to the gene product’s molecular function. including their names, GenBank accession number, Twenty-one genes in Erbin- or ErbB2-located sig- chromosomal location, fold change in their expres- naling pathways were found to be expressed in a sex- sion levels between the two sexes and p-values (one biased fashion (Table 4). However, no significant sample t-test, SPSS), are listed in Table 5.

Table 2 The Numbers of Repeating Syllable “a”, “b,” and “c” in the Songs of Adult Bengalese Finches after the Injec- tion of the Lentiviruses into the Developing HVC at PHD 15 Syllable “a” Syllable “b” Syllable “c” Groups Mean 6 SEM Range Mean 6 SEM Range Mean 6 SEM Range Normal (n 5 5) 8.42 6 1.50 6–12 11.48 6 2.21 8–15 2.16 6 0.84 1–4 Control lentivirus (n 5 3) 8.70 6 1.01 7–10 12.19 6 1.24 10–14 2.01 6 0.71 1–4 erbin-interfering lentivirus (n 5 3) - - 3.15a 6 1.10 1–14 0.98a 6 0.24 1–4 erbb2-interfering lentivirus (n 5 2) - - 6.84a 6 2.30 2–15 1.48a 6 0.65 1–4 aThe mean difference is significant at the 0.05 level. The two groups of erbin or erbb2 -interfering lentivirus (n 5 5) were combined before statistical analysis with the control lentivirus group.

Developmental Neurobiology Erbin and ErbB2 in the Sexual Differentiation of HVC 31

Table 3 The Number of Sex-Biased Genes Expressing in the Developing HVC of Bengalese Finch at Different FC Levels (p < 0.05) Mapped genesa FC Sex-biased Male-biased Female-biased 1672 1.32 57 (3.41%) 36 (2.15%) 21 (1.26%) aAll the sex biased genes could be localized to the chromosomes.

DISCUSSION from the HVC and its overlying VZ but also from other areas in the samples. Notably, the HVC is In situ hybridization analyses indicated that the Z- larger in males than in females, and the relative con- linked gene erbin and its binding partner erbb2 were tributions of the HVC and its adjacent tissue to the expressed within the HVC from PHD 15 to adult- reported expression levels of genes including erbin or hood, and its overlying VZ at PHD 15, which was erbb2 mRNA vary between male and female sam- consistent with qRT-PCR analysis of the brain sam- ples. However, according to our qRT-PCR data from ples containing male and female HVC. Although the narrow brain belts [just under the chosen samples some genes have been reported to be expressed in a containing the HVC, but the HVC was not contained, sexually dimorphic manner in song nuclei through Fig. 1(F,G)], the expression levels of erbin and erbb2 experiments that included suppression subtraction mRNA did not differ significantly between the two hybridization and cDNA microarray (Wade et al., sexes. Thus, if the distribution of a studied gene, like 2004; Tomaszycki et al., 2009; Kato and Okanoya, erbin or erbb2, had no significant sex difference in 2010; Sun et al., 2010), the expression of erbin or the areas adjacent to the HVC, the reported variations erbb2 in male-biased ways is reported here for the in the expression levels in the examined brain sam- first time. ples could also reflect, to some degree, those existing Following the addition of the ErbB family inhibitor between male and female HVC. Actually, our in situ hybridization indicated that erbin- or erbb2-mRNA- to the culture medium, we found that cell prolifera- positive cells were both located within the HVC in a tion in the VZ that overlies the developing male male-biased way, reflecting by higher densities or HVC decreased significantly, which was further con- more total numbers of the labeled cells for erbin- or firmed by injecting erbin-orerbb2-interfering lenti- erbb2-mRNA in male HVC [Fig. 1]. In contrast, the viruses into the developing HVC. We also observed densities of the labeled cells for erbin mRNA and decreases in the number of differentiated neurons erbb2 mRNA in the area just underneath the ventral (Hu1/BrdU1 or NeuN1/BrdU1) in the HVC, includ- bottom of HVC did not show any significant changes ing cells that project to the RA. In addition, we found between the males and females at PHD 15. These an abnormal organization of adult songs following data strongly indicate that erbin or erbb2 mRNA the injection of erbin-orerbb2-interfering lentivi- expression in the HVC differ significantly between ruses into the developing HVC at PHD 15. Finally, the two sexes. using Genome Oligo Microarray, 19,996 genes were It is known that lentiviral vectors can be stably found to be expressed in the developing HVC of nes- incorporated into the genomic DNA of dividing or tling Bengalese finches, including 367 sex-biased nondividing cells (Park, 2007). This is very important genes and 173 male-biased genes (of which 27.2% for birds, as it is not as easy as to create transgenic were Z-linked genes, and 21 genes were located in birds compared to mammals. Although lentiviral Erbin- or ErbB2-related signaling pathways). knockdown is often used in mammals, it has also The following point is noteworthy: The boundaries been used in oscine species (Haesler et al., 2007; of HVC, especially female HVC, were difficult to Schulz et al., 2010). Following the injection of determine in brain slices (approximately 400 lm lentivirus-mediated RNAi for the Forkhead box pro- thick) without any staining, even under dark-field tein 2 (Foxp2) into Area X, both mRNA and protein illumination (however, LH could be reliably identi- levels are reduced without observed side effects in fied). For RNA extraction from HVC, we used sepa- the zebra finch (Haesler et al., 2007). Our study indi- rate male or female brain samples containing HVC cated that after transferring the erbin-orerbb2-inter- (but not exact HVC) in qRT-PCR and DNA microar- fering lentiviruses into the developing HVC, ray analysis. Thus, the extracted RNA was not only lentivirus-mediated RNAi could be persistently

Developmental Neurobiology 32 Zhao et al.

Figure 9 DNA microarray to identify sex-biased genes expressed in the developing HVC of Ben- galese finches at PHD 15. A, Male/female expression ratios as a function of Z chromosome position (3107 bp) in chicken. There are a total of 103 genes on the Z chromosome that are expressed in the developing HVC in a significant sex-biased fashion. B, Distribution of male- and female-biased genes (with fold-change in expression greater than 1.5, p < 0.05 significant level) in the chromo- somes. C, Percentages of sex-biased genes in autosomes and sex chromosomes (Z/W). D, Func- tional classification of sex-biased expression genes according to GO biological processes (left) and GO molecular function (right). [Color figure can be viewed at wileyonlinelibrary.com] produced in the HVC even at adulthood [Fig. 7A1- overlying VZ at PHD 24 [Fig. 4A1-B4]. However, A3]. Positive cells for GFP (reporter gene for lentivi- GFP1 cells were not seen in the areas outside the ruses) were seen not only within HVC but also in its injection regions. These data indicated that

Developmental Neurobiology Erbin and ErbB2 in the Sexual Differentiation of HVC 33

Table 4 Sex-Biased Genes Expressing in the Developing HVC of Bengalese Finch in Erbin Interacted Pathways at 1.3 FC level (p < 0.05) Signaling pathways Male-biased genes Female-biased genes PKC signaling pathway ESR1, DAG1 - PI3K-Akt signaling pathway FRAP1, GSK3B - MAPK signaling pathway MAP4K4, HSPA5, MAPK8IP3, CRK(PAK-JNKK-TNK); MAPK8(PAK-JNKK-JNK) HSPB, MAPK14 (P38); HRAS, ATF4(Ras-Raf-MEK-ERK) TGF-b signaling pathway INHBA(ligand); SPP1(repressor) SMAD1, SMAD2, SMAD3(R-Smads); CREBBP (coactivator family); ZNF423, RUNX3(cofactor) NF-jB signaling pathway - - lentiviruses were successfully infected the cells pathways. As shown above, erbin and erbb2 were within HVC and its overlying VZ but did not spread both expressed within the HVC from PHD 15 to to wider areas through the vessels or other unknown adulthood, and in the overlying VZ at PHD 15.

Table 5 Details of Sex-Biased Genes in the Developing HVC of Bengalese Finch in Erbin Interacted Pathways GenbankAccession Chromosom Fold Chang GeneName number allocation (M/F) p-value ErbB2 signaling pathway ESR1 estrogen receptor 1 NM_205183 chr3 2.17 0.0385 DAG1 dystroglycan NM_001097540 chr12 2.00 0.0271 FRAP1 FK506 binding protein 12-rapamycin XM_417614 chr21 1.73 0.0036 associated protein 1 GSK3B glycogen synthase kinase 3 beta XM_416557 chr1 1.58 0.0126 ErbB2 v-erb-b2 avian erythroblastic leukemia viral NM_001044661 Chr27 1.37 0.0276 oncogene homolog 2, neuro/glioblastoma derived oncogene homolog HRAS v-Ha-ras Harvey rat sarcoma viral NM_205292 chr5 0.55 0.0406 oncogene homolog MAPK signaling pathway MAP4K4 mitogen-activated protein kinase CR391561 chr1 2.50 0.0156 kinase kinase kinase 4 HSPA5 heat shock 70kDa protein 5 NM_205491 chr17 1.93 0.0259 MAPK8IP3 mitogen-activated protein kinase 8 XM_424591 chr14 1.86 0.0113 interacting protein 3 MAPK8 mitogen-activated protein kinase 8 BX934173 chr6 1.77 0.0155 DAG1 dystroglycan NM_001097540 chr12 2.00 0.0271 HSPB2 heat shock 27kDa protein 2 NM_001001527 chr24 0.64 0.0007 MAPK14 mitogen-activated protein kinase 14 CR388894 chr26 0.70 0.0031 CRK v-crk sarcoma virus CT10 oncogene NM_001007846 0.69 0.015 TGF-b signaling pathway INHBA inhibin, beta A NM_205396 chr2 2.62 0.0093 CREBBP CREB binding protein XM_414964 chr14 1.70 0.0137 (Rubinstein-Taybi syndrome) SMAD3 SMAD family member 3 NM_204475 chr10 1.64 0.0197 SMAD2 SMAD family member 2 NM_204561 chrZ 1.58 0.0004 ZNF423 zinc finger protein 423 BX935955 chr11 1.39 0.0220 SMAD1 SMAD family member 1 AY953143 chr4 1.37 0.0019 RUNX3 runt-related transcription factor 3 XM_001232977 chr23 1.35 0.0159 SPP1 secreted phosphoprotein 1 NM_204535 chr4 0.42 0.0350 (osteopontin, bone sialoprotein I, early T-lymphocyte activation 1)

Developmental Neurobiology 34 Zhao et al.

Following the injection of erbin-orerbb2-interfering chicken embryos, of which 246 genes are male- lentiviruses, the expression level of erbin or erbb2 biased (Zhang et al., 2010). In addition, some reports significantly decreased in the HVC and its overlying have shown that the male-to-female ratios of Z chro- VZ, which might cause changes in cell proliferation mosome gene expression are similar across tissues or cell differentiation (Olayioye et al., 2000; Nie and and ages between the two bird species (chicken and Chang, 2006; Roskoski, 2014; Tao et al., 2014). zebra finch) (Itoh et al., 2010), and male-biased or Z- However, our data also indicated that the cell prolif- linked genes are extensive in the brain of two species eration and differentiation towards Hu1 cells in the of songbirds, zebra finch (Taeniopygia guttata) and areas (no GFP1 cells) adjacent to the injection whitethroat (Sylvia communis) (Naurin et al., 2011). regions did not differ between the groups receiving However, unlike the chicken Z chromosome, in injections of erbin-orerbb2-interfering lentiviruses which there is a reduced male-to-female ratio of gene with the groups receiving injection of control lentivi- expression near MHM (male hypermethylated) (Mel- ruses. Thus, the effect scope of lentivirus injection amed and Arnold, 2007), no obvious valley values or was limited, and it might be judged by the presence peak values in the male-to-female ratio of gene of GFP1 cells. expression occur along Chromosome Z for the stud- Although the different types of lentiviruses (erbin, ied sex-biased genes in the Bengalese finch [Fig. erbb2 and control RNAi) had similar infection activi- 9A]or zebra finch (Itoh et al., 2010). These data sug- ties in the HVC, the total number of NeuN1/BrdU1 gest that the mechanisms for dosage compensation cells in HVC decreased more significantly than in may differ among avian taxa. A detailed study of dos- GABA1/BrdU1 cells after injection of erbin- or age compensation in the developing HVC of Bengal- erbb2-interfering lentiviruses into the male HVC at ese finch is underway in our lab. We are attempting PHD 15 (erbin-interfering lentiviruses: NeuN1/ to use RNA-seq analysis to study the male/female BrdU1 decreased by 72.84% vs GABA1/BrdU1 expression ratios of Z-linked genes to autosomal decreased by 37.96%; erbb2-interfering lentiviruses: genes, as shown in a previous report (Graves, 2014). NeuN1/BrdU1 decreased by 66.67% vs GABA1/ Erbin does not affect the total level of erbb2 BrdU1 decreased by 57.69%). This was caused by mRNA (Liu et al., 2008) or ErbB2 tyrosine phosphor- significant decreases in the density of NeuN1/BrdU1 ylation (Huang et al., 2003). However, Erbin inhibits cells in HVC, but no significant changes in the den- the ubiquitination of ErbB2. Its PDZ domain specifi- sity of GABA1/BrdU1 cells. As shown above, fol- cally binds to ErbB2 to increase its protein levels, lowing the injection of erbin- or erbb2-interfering and the expression of Erbin is elevated in some lentiviruses, the sizes of the HVC decreased by 40– tumors, activating the ErbB2-dependent ERK path- 50%, resulting in decreases in the total number of way (Yao et al., 2015). Deletion of the PDZ domain GABA1/BrdU1 cells in the HVC. However, we did (or Erbin silencing) hinders ErbB2-dependent tumori- not further examine whether the total number of cells genesis or tumor growth and decreases the differenti- in HVC also decreased (although neurons decrease, ation of mature myelinating Schwann cells (Tao various types of glia cells may increase). The above et al., 2009, 2014). In addition, by interacting with inconsistent results concerning GABA1/BrdU1 or Ras-Raf in the MAPK pathway, Erbin can suppress NeuN1/BrdU1 in the HVC might be ascribed to the ERK activation (Dan et al., 2010). Thus, Erbin can fact that these cells have different birth sites or times either promote or suppress ERK signaling through (Alvarez-Buylla et al., 1994; Scott and Lois, 2007; various mechanisms, resulting in varied effects such Schulz et al., 2010). For example, GABA cells origi- as the inhibition of NGF (nerve growth factor)- nate from the VZ in the subpallium, but other types induced neuronal differentiation in PC12 (Huang of cells originate in the VZ of the pallium (Cobos et al., 2003; Dai et al., 2006, 2007) and the promotion et al., 2001). As a result, there were different effects of keratinocyte differentiation (Harmon et al., 2013). of erbin- or erbb2-interfering lentiviruses on the pro- Through its involved signaling pathways (Ras-Raf- duction or differentiation of these cells. Mek-Erk, PI3K and PKC), ErbB2 plays roles in cell The data from gene microarrays were largely con- proliferation, migration and cell survival in cancer sistent with a previous report that used the Affyme- formation and in early vertebrate embryogenesis trix Chicken Genome Array, which includes 19,734 (Olayioye et al., 2000; Nie and Chang, 2006). Our genes expressed in the brain, heart and gonads of E18 results indicated that both cell proliferation in the VZ chicks, including 286 sex-biased genes in the brain overlying the developing HVC and the number of dif- (Ellegren et al., 2007). Using the same Genome ferentiated neurons decreased significantly in HVC Array, another report has shown that 14,548 probe after down-regulation of erbin or erbb2, consistent sets appear in the anterior primitive streak region of with the above-mentioned reports. However, we did

Developmental Neurobiology Erbin and ErbB2 in the Sexual Differentiation of HVC 35 not determine whether our results were caused by the females before the critical period (around 30 days interaction of Erbin and ErbB2 (through ErbB2- after hatching) (Kirn and DeVoogd, 1989; Burek dependent ERK pathway) or the signal pathways et al., 1994; Chen et al., 2014). Unlike HVC or Area directly mediated by Erbin (TGF-b, MAPK or NF- X, almost all the neurons have been generated before jB) or ErbB2 (Raf-Mek-Erk, PI3K or PKC). These hatching in other song nuclei such as RA and LMAN, issues should be further investigated in future studies. and the neurons are similar in males and females The results of the cDNA microarray indicated that, prior to the critical period. However, the numbers of of 21 sex-biased expression genes in Erbin and neurons and their volumes and sizes decrease after ErbB2 or their interacting signal pathways, most the critical period to a much greater extent in females were located in the MAPK (including Pak-Jnkk-Jnk than in males (Bottjer et al., 1986; Nixdorf- and Ras-Raf-Mek-Erk) and TGF-b signal pathways. Bergweiler, 1996). Although the mechanism of this In the MAPK pathway, several genes that are neuronal reduction remains unclear, unilateral HVC expressed in a male- or female-biased fashion could lesions at PHD 20 increased cell death and decreased be involved in the sexual differentiation of HVC neuron number and soma size within the ipsilateral through different mechanisms, as mentioned above. RA (Akutagawa and Konishi, 1994; Chen et al., However, we did not examine further whether phos- 2014). Thus more cell loss in females in other song phorylation levels changed in Erbin- or ErbB2- nuclei might be primarily caused by the sexual differ- involved signal pathways following the down- ences in cell production of newborn cells in HVC or regulation of erbin or erbb2. Area X. In addition, among song control nuclei, only Studies in breast cancer cells indicate that the HVC contains estrogen receptors, and HVC is thus nuclear estrogen receptor (ER) can be phosphorylated most probably involved in estrogen mediated sexual by the activated MAPK in Erbin- or ErbB2-mediated differentiation of song control system. Considering signal pathways, resulting in increased ER-related the above findings, we chose HVC to study the mech- transactivation of genes, including those playing roles anisms under which the song control nucleus is sexu- in cell proliferation and cell differentiation (Levin, ally differentiated. 2003). Notably, such action occurs independently of In the present study we first indicated that both estrogen. If the above action through MAPK pathway erbin and erbb2 have male-biased expressions in the is also confirmed to be involved in the sexual differ- developing HVC of the Bengalese finch by using entiation of the HVC, it will explain, to some degree, qRT-PCR and in situ hybridization. Following the why the sexual differentiation of song nuclei is not injection of erbin- or erbb2-interfering lentiviruses solely dependent on steroid hormones (Wade and into the HVC and its overlying VZ at PHD 15, the Arnold, 1996; Agate et al., 2003; Zhao et al., 2010). cell proliferation in the VZ at PHD 24, the number of the differentiated neurons (Hu1/BrdU1 or NeuN1/ No significant differences were found in the song 1 organization between normally reared birds and birds BrdU ) in the HVC at PHD 31 or PHD 130, and that received an injection of control lentiviruses, sug- the number of RA-projecting cells at PHD 130 all gesting that the surgical operation and control lentivi- decreased significantly. These data indicated that ruses did not substantially affect song learning and both erbin and erbb2 are involved in cell prolifera- production (Haesler et al., 2007). However, no audi- tion and neuronal differentiation during the devel- ble songs were produced, and song organization suf- oping HVC. These data combined with the male- fered from large changes in birds that received an biased expressions of erbin and erbb2 in the devel- injection of erbin-orerbb2-interfering lentiviruses oping HVC suggest that erbin and erbb2 are into the HVC. In the present study, some birds did involved in the sexual differentiation (including not survive to the scheduled age, and thus we did not cell proliferation and neuronal differentiation) of obtain enough data to address the mechanisms under- HVC. By using cDNA microarrays, we further lying changes in song behavior. However, there was found that approximately 20 genes were involved a significant decrease in the number of differentiated in the Erbin- or ErbB2-related signaling pathways. neurons in the HVC, including those projecting to the However, the details of Erbin- or ErbB2-involved RA, in birds receiving erbin-orerbb2-interfering len- signal pathways involved in the sexual dimorphism tiviruses. This might be one reasons for the changes of HVC are required to be further clarified in in song behavior. future study. It is needed to point out that among the song con- We thank Drs. Jin-Liu and Chao-Xi in the Experimental trol nuclei with sexually dimorphic sizes, only HVC Technology Center for Life Sciences and the College of or Area X neurons are produced in the ventricular Life Sciences, Beijing Normal University, for technological zone with markedly greater numbers in males than in help.

Developmental Neurobiology 36 Zhao et al.

CONFLICT OF INTEREST Catchpole CK, Slater PJB. 2008. The Study of Bird Song. Bird Song: Biological Themes and Variations. Cam- The authors declare that there are no conflicts of bridge: Cambridge University, pp 1–18. Cardin JA, Schmidt MF. 2004. Auditory responses in multi- interest. ple sensorimotor song system nuclei are co-modulated by behavioral state. J Neurophysiol 91:2148–2163. Chen Q, Zhang X, Zhao Y, Zhou X, Sun L, Zeng SJ, Zuo AUTHORS CONTRIBUTION MX. 2014. Sexual differences in cell proliferation in the ventricular zone, cell migration and differentiation in the All authors had full access to all data in the study and HVC of juvenile Bengalese finch. PLoS One 9:e97403. took responsibility for data integrity and data analy- Citri A, Yarden Y. 2006. EGF-ERBB signalling: Towards sis. Study concept and design: SJZ, XWZ, XJC, the systems level. Nat Rev Mol Cell Biol 7:505–516. ZMH. Acquisition of data: YLZ, XBZ, RW, FW, Cobos I, Puelles L, Martınez S. 2001. The avian telence- phalic subpallium originates inhibitory neurons that YTZ. Analysis and interpretation of data: YLZ, XBZ, invade tangentially the pallium (dorsal ventricular ridge RW. Drafting of the article: SJZ, YLZ. Critical revi- and cortical areas). Dev Biol 239:30–45. sion of the article for important intellectual content: Dai P, Xiong WC, Mei L. 2006. Erbin inhibits RAF activa- SJZ, YLZ. Statistical analysis: YLZ, XBZ, RW, FW, tion by disrupting the sur-8-Ras-Raf complex. J Biol YTZ. Obtained funding: SJZ, XBZ, XWZ. Adminis- Chem 281:927–933. trative, technical, and material support: SJZ, XWZ. Dai F, Chang C, Lin X, Dai P, Mei L, Feng XH. 2007. Study supervision: SJZ, XWZ. Erbin inhibits transforming growth factor beta signaling through a novel Smad-interacting domain. Mol Cell Biol 27:6183–6194. REFERENCES Dan L, Shi M, Duan H, Han C, Guo N. 2010. Erbin, a nega- tive regulator in diverse signal pathways. Curr Protein Pept Sci 11:759–764. Agate RJ, Grisham W, Wade J, Mann S, Wingfield J, Dawson DA, Akesson M, Burke T, Pemberton JM, Slate J, Schanen C, et al. 2003. Neural, not gonadal, origin of Hansson B. 2007. Gene order and recombination rate in brain sex differences in a gynandromorphic finch. Proc homologous chromosome regions of the chicken and a Natl Acad Sci USA 100:4873–4878. passerine bird. Mol Biol Evol 24:1537–1552. Akutagawa E, Konishi M. 1994. Two separate areas of the Derjusheva S, Kurganova A, Habermann F, Gaginskaya E. brain differentially guide the development of a song con- 2004. High chromosome conservation detected by com- trol nucleus in the zebra finch. Proc Natl Acad Sci U S A parative chromosome painting in chicken, pigeon and 91:2413–12417. passerine birds. Chromosome Res 12:715–723. Alvarez-Buylla A, Ling CY, Yu WS. 1994. Contribution of Ellegren H, Hultin-Rosenberg L, Brunstrom€ B, Dencker L, neurons born during embryonic, juvenile, and adult life to Kultima K, Scholz B. 2007. Faced with inequality: the brain of adult canaries: Regional specificity and Chicken do not have a general dosage compensation of delayed birth of neurons in the song control nuclei. sex-linked genes. BMC Biol 5:40–49. J Comp Neurol 347:233–248. Graves JA. 2014. Avian sex, sex chromosomes, and dosage Backstrom€ N, Backstrom€ M, Gustafsson L, Qvarnstrom€ A, compensation in the age of genomics. Chromosome Res Cheng H, Ellegren H. 2006. Genetic mapping in a natural 22:45–57. population of collared flycatchers (Ficedula albicollis): Griffiths-Jones S. 2004. The microRNA Registry. Nucleic Conserved synteny but gene order rearrangements on the Acids Res 32:109–111. avian Z chromosome. Genetics 174:377–386. Griffiths-Jones S, Grocock RJ, van Dongen S, Bateman A, Barami K, Iversen K, Furneaux H, Goldman SA. 1995. Hu Enright AJ. 2006. miRBase: microRNA sequences, tar- protein as an early marker of neuronal phenotypic differ- gets and gene nomenclature. Nucleic Acids Res 34:140– entiation by subependymal zone cells of the adult song- 144. bird forebrain. J Neurobiol 28:82–101. Haesler S, Rochefort C, Georgi B, Licznerski P, Osten P, Borg JP, Marchetto S, Le Bivic A, Ollendorff V, Jaulin- Scharff C. 2007. Incomplete and inaccurate vocal imita- Bastard F, Saito H, Fournier E, et al. 2000. ERBIN: A tion after knockdown of FoxP2 in songbird basal ganglia basolateral PDZ protein that interacts with the mamma- nucleus Area X. PLoS Biol 5:e321. lian ERBB2/HER2 receptor. Nat Cell Biol 2:407–414. Harmon RM, Simpson CL, Johnson JL, Koetsier JL, Bottjer SW, Miesner EA, Arnold AP. 1986. Changes in Dubash AD, Najor NA, Sarig O, Sprecher E, Green KJ. neuronal number, density and size account for increases 2013. Desmoglein-1/Erbin interaction suppresses ERK in volume of song control nuclei during song develop- activation to support epidermal differentiation. J Clin ment in zebra finches. Neurosci Lett 67:263–268. Invest 123:1556–1570. Burek MJ, Nordeen KW, Nordeen EJ. 1994. Ontogeny of Huang YZ, Zang M, Xiong WC, Luo Z, Mei L. 2003. Erbin sex differences among newly-generated neurons of the suppresses the MAP kinase pathway. J Biol Chem 278: juvenile avian brain. Brain Res Dev Brain Res 78:57–64. 1108–1114.

Developmental Neurobiology Erbin and ErbB2 in the Sexual Differentiation of HVC 37

Itoh Y, Melamed E, Yang X, Kampf K, Wang S, Yehya N, Nottebohm F, Arnold AP. 1976. Sexual dimorphism in Van Nas A, et al. 2007. Dosage compensation is less vocal control areas of the songbird brain. Science 194: effective in birds than in mammals. J Biol 6:2. 211–213. Itoh Y, Replogle K, Kim YH, Wade J, Clayton DF, Arnold Okanoya K, Yamaguchi A. 1997. Adult Bengalese finches AP. 2010. Sex bias and dosage compensation in the zebra (Lonchura striata var. domestica) require real-time audi- finch versus chicken genomes: General and specialized tory feedback to produce normal song syntax. patterns among birds. Genome Res 20:512–518. J Neurobiol 33:343–356. Kashimada K, Koopman P. 2010. Sry: The master switch Olayioye MA, Neve RM, Lane HA, Hynes NE. 2000. The in mammalian sex determination. Development 137: ErbB signaling network: Receptor heterodimerization in 3921–3930. development and cancer. EMBO J 19:3159–3167. Kato M, Okanoya K. 2010. Molecular characterization of Park F. 2007. Lentiviral vectors: Are they the future of ani- the song control nucleus HVC in Bengalese finch brain. mal transgenesis? Physiol Genomics 31:159–173. Brain Res 1360:56–76. Pawlisch BA, Remage-Healey L. 2015. Neuroestrogen sig- Levin ER. 2003. Bidirectional signaling between the estro- naling in the songbird auditory cortex propagates into a gen receptor and the epidermal growth factor receptor. sensorimotor network via an ‘interface’ nucleus. Neuro- Mol Endocrinol 17:309–317. science 284:522–535. Kirn JR, DeVoogd TJ. 1989. Genesis and death of vocal Pinaud R, Mello CV. 2007. GABA immunoreactivity in auditory and song control brain areas of zebra finches. control neurons during sexual differentiation in the zebra J Chem Neuroanat 34:1–21. finch. J Neurosci 9:3176–3187. Rangwala R, Banine F, Borg JP, Sherman LS. 2005. Erbin Kirn JR, Fishman Y, Sasportas K, Alvarez-Buylla A, regulates mitogen-activated protein (MAP) kinase activa- Nottebohm F. 1999. Fate of new neurons in adult canary tion and MAP kinase-dependent interactions between high vocal center during the first 30 days after their for- Merlin and adherens junction protein complexes in mation. J Comp Neurol 411:487–494. Schwann cells. J Biol Chem 280:11790–11797. Konish M, Akutagawa E. 1988. A critical period for estro- Replogle K, Arnold AP, Ball GF, Band M, Bensch S, gen action on neurons of the song control system in the Brenowitz EA, Dong S, et al. 2008. The Songbird Neuro- zebra finch. Proc Natl Acad Sci USA 85:7006–7007. genomics (SoNG) Initiative: Community-based tools and Li X, Chiang HI, Zhu J, Dowd SE, Zhou H. 2008. Charac- strategies for study of brain gene function and evolution. terization of a newly developed chicken 44K Agilent BMC Genomics 9:131. microarray. BMC Genomics 9:60. Roskoski RJ. 2014. The ErbB/HER family of protein- Liu N, Zhang J, Liu S, Liu Y, Zheng D. 2008. Erbin-regulated tyrosine kinases and cancer. Pharmacol Res 79:34–74. sensitivity of MCF-7 breast cancer cells to TRAIL via Schulz SB, Haesler S, Scharff C, Rochefort C. 2010. ErbB2/AKT/NF-kappaB pathway. J Biochem 143:793–801. Knockdown of FoxP2 alters spine density in Area X of Louissaint A, Rao JS, Leventhal C, Goldman SA. 2002. the zebra finch. Genes Brain Behav 9:732–740. Coordinated interaction of neurogenesis and angiogenesis Schumacher JW, Schneider DM, Woolley SM. 2011. Anes- in the adult songbird brain. Neuron 34:945–960. thetic state modulates excitability but not spectral tuning Zhu N, Sun YY, Zeng SJ, Zhang XW, Zuo MX. 2008. Col- or neural discrimination in single auditory midbrain neu- lapsin response mediator protein-4 (CRMP-4) expression rons. J Neurophysiol 106:500–514. in posthaching development of song control nuclei in Scott BB, Lois C. 2007. Developmental origin and identity Bengalese finches. Brain Res Bull 76:551–558. of song system neurons born during vocal learning in McCarthy MM. 2009. The two faces of estradiol: Effects songbirds. J Comp Neurol 502:202–214. on the developing brain. Neuroscientist 15:599–610. Sinclair AH, Berta P, Palmer MS, Hawkins JR, Griffiths Melamed E, Arnold AP. 2007. Regional differences in dos- BL, Smith MJ, Foster JW, et al. 1990. A gene from the age compensation on the chicken Z chromosome. human sex-determining region encodes a protein with Genome Biol 8:R202. homology to a conserved DNA-binding motif. Nature Naurin S, Bensch S, Hansson B, Johansson T, Clayton DF, 346:240–244. Albrekt AS, von Schantz T, et al. 2008. TECHNICAL Sun L, Wang R, Zeng S, Geng Z, Zhang X, Zuo M. 2010. ADVANCES: A microarray for large-scale genomic and Identification of male-biased gene: Parvalbumin in song transcriptional analyses of the zebra finch (Taeniopygia control nuclei of the Bengalese finch. Neurosci Res 68: guttata) and other passerines. Mol Ecol Resour 8:275– 22–34. 281. Tao Y, Da P, Liu Y, Marchetto S, Xiong WC, Borg JP, Mei Naurin S, Hansson B, Hasselquist D, Kim YH, Bensch S. 2011. L. 2009. Erbin regulates NRG1 signaling and myelina- The sex-biased brain: Sexual dimorphism in gene expression tion. Proc Natl Acad Sci USA 106:9477–9482. in two species of songbirds. BMC Genomics 12:37. Tao Y, Shen C, Luo S, Traore W, Marchetto S, Santoni MJ, Nie S, Chang C. 2006. Regulation of early Xenopus devel- Xu L, et al. 2014. Role of Erbin in ErbB2-dependent opment by ErbB signaling. Dev Dyn 235:301–314. breast tumor growth. Proc Natl Acad Sci USA 111:4429– Nixdorf-Bergweiler BE. 1996. Divergent and parallel 4438. development in volume sizes of telencephalic song nuclei Tobari Y, Nakamura KZ, Okanoya K. 2005. Sex differ- in and female zebra finches. J Comp Neurol 375:45–456. ences in the telencephalic song control circuitry in

Developmental Neurobiology 38 Zhao et al.

Bengalese finches (Lonchura striata var. domestica). tumourigenesis and tumour growth in colorectal cancer Zoolog Sci 22:1089–1094. by preventing c-Cbl-mediated ubiquitination and down- Tomaszycki ML, Peabody C, Replogle K, Clayton DF, regulation of EGFR. J Pathol 236:65–77. Tempelman RJ, Wade J. 2009. Sexual differentiation of Yu Z, Cui B, Jin Y, Chen H, Wang X. 2011. Novel irrevers- the zebra finch song system: Potential roles for sex chro- ible EGFR tyrosine kinase inhibitor 324674 sensitizes mosome genes. BMC Neurosci 10:24. human colon carcinoma HT29 and SW480 cells to apo- Wade J, Arnold AP. 1996. Functional testicular tissue does ptosis by blocking the EGFR pathway. Biochem Biophys not masculinize development of the zebra finch song sys- Res Commun 411:751–756. tem. Proc Natl Acad Sci USA 93:5264–5268. Zeng SJ, Lin YT, Tian CP, Zhang XW, Zuo MX. 2009. Wade J, Peabody C, Coussens P, Tempelman RJ, Clayton Evolutionary significance of delayed neurogenesis in the DF, Liu L, Arnold AP, et al. 2004. A cDNA microarray core versus shell auditory areas of Mus musculus. from the telencephalon of juvenile male and female zebra J Comp Neurol 515:600–613. finches. J Neurosci. Methods 138:199–206. Zhang SO, Mathur S, Hattem G, Tassy O, PourquieO. Wallis JW, Aerts J, Groenen MA, Crooijmans RP, Layman 2010. Sex-dimorphic gene expression and ineffective D, Graves TA, Scheer DE, et al. 2004. A physical map of dosage compensation of Z-linked genes in gastrulating the chicken genome. Nature 432:761–764. chicken embryos. BMC Genomics 11:13. Wang N, Li JQ, Liu YY, Zhang ZW. 2010. Improvement Zhao D, McBride D, Nandi S, McQueen HA, McGrew MJ, on molecular sex identification primers for Passeriform Hocking PM, Lewis PD, et al. 2010. Somatic sex bird species. Chinese Birds 1:65–69. identity is cell autonomous in the chicken. Nature 464: Wang N, Zhang ZW. 2009. The novel primers for sex iden- 237–242. tification in the brown eared-pheasant and their applica- Zhou Q, Zeng R, Xu C, Liu L, Chen L, Kou P, Pei G, et al. tion to other species. Mo Ecol Resour 9:186–188. 2012. Erbin inhibits TGF-beta1-induced EMT in renal Yao S, Zheng P, Wu H, Song LM, Ying XF, Xing C, Li Y, tubular epithelial cells through an ERK-dependent path- et al. 2015. Erbin interacts with c-Cbl and promotes way. J Mol Med (Berl) 90:563–574.

Developmental Neurobiology