Genetic regulation of sex differences in songbirds and lizards

Abstract

Sex differences in the morphology of neural and peripheral structures related to reproduction often parallel the frequency of particular behaviours displayed by males and females. In a variety of model organisms, these sex differences are organized in development by gonadal steroids, which also act in adulthood to modulate behavioural expression and in some cases to generate parallel anatomical changes on a seasonal basis. Data collected from diverse species, however, suggest that changes in hormone availability are not sufficient to explain sex and seasonal differences in structure and function. This paper pulls together some of this literature from songbirds and lizards and considers the information in the broader context of taking a comparative approach to investigating genetic mechanisms associated with behavioural neuroendocrinology.

1. Introduction

Sex differences in neuroanatomy and behaviour exist across vertebrate species. In many cases, these differences are due to availability of steroid hormones during development, which can permanently organize the structure and/or function of brain areas into masculine or feminine forms. These same hormones modulate the regions in adulthood, activating behaviour and in some cases affecting the magnitude of the sex difference on a seasonal basis [1]. While numerous factors downstream of gonadal hormones play critical roles across the life cycle, in rodents, the majority of sex differences in brain and behaviour are initially organized by postnatal oestradiol (E2) following metabolism from testosterone (T), with a limited number controlled by androgen directly [2].

Studies in mice using molecular manipulations including knocking out the receptors required for steroid hormones to act and separating the presence of male and female gonads from sex chromosome complement (the four core genotype model) have identified several sex differences in the brain that exist independently of gonadal hormone effects [3]. The ability to manipulate genetic factors in mice makes them a powerful model for investigation of molecular players. Additional model organisms provide the opportunity to assess the ubiquity and uniqueness of processes, as well as to uncover novel mechanisms with potential biomedical relevance. Some avian and reptilian systems indicate that differences in hormone availability can be insufficient to explain changes in structure and function of neural and peripheral structures important for the regulation of reproductive behaviours. Research in my laboratory has focused on two of these models—zebra finches for understanding developmental mechanisms and anole lizards primarily for investigating adult plasticity. Sections 2 and 3 describe this work in the context of information from other songbirds and lizards.

2. Songbirds

Songbirds are particularly useful for understanding sexual differentiation of brain and behaviour. Across these oscine species, great diversity exists in the characteristics of song and the magnitude of sex differences in structure and function [4]. In some species these songs are modified in adulthood, whereas in others the song does not change after sexual maturity [5]. Song learning and production are controlled by a well-identified set of structures in the forebrain. In the pathway critical for song production, HVC (proper name) projects to the robust nucleus of the arcopallium (RA), which innervates motoneurons in the brainstem that control muscles of the vocal organ (figure 1). Area X and the lateral magnocellular nucleus of the anterior nidopallium (LMAN) are components of the anterior forebrain pathway and are more involved in song learning and plasticity [7,8]. The vast majority of research on mechanisms regulating sexual differentiation of the song circuit has been conducted in zebra finches. In this species, song is learned from tutors (commonly a bird's father), and is performed exclusively by males [9]. In parallel, the HVC and RA are far larger in males than females due to the presence of both more and larger cells [8]. Area X is only visible in males, and while morphology of LMAN itself is similar between the two sexes, its projection to RA is enhanced in adult males compared with females [10].

Figure 1.

Sexually dimorphic song system in zebra finches. The top left panel depicts a simplified version of the circuit in the sagittal plane adapted from [6], with the motor pathway indicated with thicker blue lines and the anterior forebrain pathway noted with thinner red lines. The three photographs, all at the same magnification, show coronal sections of RA, indicated with an arrow in a control male, control female and a female treated with oestradiol (E2) beginning at post-hatching day 3. These images were taken from 25-day-old birds; the sex difference is substantially larger in adult birds. HVC is the proper name; RA, robust nucleus of the arcopallium; nXIIts, tracheosyringeal portion of the hypoglossal nucleus; DLM, dorsal lateral nucleus of the medial thalamus; LMAN, lateral magnocellular nucleus of the anterior nidopallium; Cb, cerebellum.

Like the song of male zebra finches, which is stable by three months of age, the sex differences in these brain areas are permanently organized prior to maturity. In HVC and Area X, the primary mechanism influencing sexual differentiation is the addition of more cells in males. In RA, cells are similar in number in the two sexes for the first three weeks after hatching and then more die in females [11]. Early investigations of the mechanisms regulating sexual differentiation focused on E2; treating hatchling females with this hormone can both masculinize the anatomy of their song circuits and allow them to produce some song in adulthood (figure 1). However, complete masculinization does not occur [12]. If E2 does normally masculinize males, then some mechanism should exist to increase its availability or activity in this sex, but oestrogen receptors in the song system appear limited overall and comparable in males and females. Similarly, consistent sex differences in plasma E2 and brain oestrogen synthesis during development have not been detected [13]. Interestingly, post-hatching E2 can have opposite, demasculinizing, effects on tissues outside the neural song circuit in zebra finches. For example, when administered to post-hatching males, this hormone feminizes muscles of the vocal organ (syrinx) [14] and inhibits mounting behaviour [15].

Attempts to block masculinization of most sexually dimorphic characteristics by castrating juvenile males or treating them with anti-oestrogens or oestrogen synthesis inhibitors have been unsuccessful [13]. The one exception is the projection from HVC to RA, in which growth in vitro is both enhanced by E2 treatment of female brain slices and inhibited by limiting E2 availability and action in male tissue [16]. Other evidence that factors beyond steroid hormones are important comes from treating genetically female embryos with fadrozole, an inhibitor of oestrogen synthesis. This manipulation produces substantial amounts of functional testicular tissue, but has no masculinizing effects on audible song production or song system morphology [17,18]. Data from a naturally occurring gyandromorphic zebra finch, in which the left side of the body was genetically female and the right was genetically male [19], confirm the idea that factors intrinsic to cells within the brain are critical to masculinization (or defeminization), particularly for HVC. However, diffusible factors also probably influence sexual differentiation of RA and Area X.

Male birds are homogametic (ZZ) and females are heterogametic (ZW). Because dosage compensation in birds is limited [20,21], such that many Z-genes exhibit enhanced expression in males compared with females, we have been investigating the possibility that Z-chromosome genes are involved in masculinization. Our early efforts, some coordinated with the Songbird Neurogenomics Initiative led by David Clayton, involved cDNA microarray screens for differential gene expression in the forebrains of developing zebra finches [22,23]. We identified a set of candidates and have studied a few of them in some depth. In each case, these genes are not only located on the Z-chromosome, but expression is increased in one or more song control nuclei in males compared with females. Of particular interest is that in no case is this increased expression consistent across all song nuclei. That is, even though the gene may be rather widely expressed in the brain, the sex difference is localized to specific components of the song system.

Examples of candidate Z-genes include ribosomal proteins 17 and 37 (RPL17 and RPL37). At post-hatching day 25, which is during a period of heightened sexual differentiation of the song circuit, the mRNA for both of these genes is increased in both Area X and RA. These sex differences are far smaller or absent in adulthood [24]. Sexually dimorphic protein expression is also detected in HVC. The number of cells expressing RPL17 and RPL37 increases from day 25 to adulthood in the HVC and Area X of males, while the number of these cells declines in females in HVC and RA [25]. Little is known about the function of these two genes, but levels of ribosomal mRNAs are at least in part determined by the rates of growth and proliferation of cells [26]. The patterns of expression in some song nuclei suggest they are consequences of changes in overall cell number, but in other cases, the patterns are consistent with specific roles of the ribosomal proteins in masculinization [25]. Another such protein, RPL7 which is not on the Z-chromosome, functions as a co-activator of the oestrogen receptor and appears important for masculinizing several aspects of song system morphology [27]. The authors of this work also highlight the potential for other steroid receptor coactivators, including steroid receptor coactivator 1 (SRC-1) and CREB-binding protein (CBP), to be involved in sexual differentiation of the songbird brain. More work should be done to investigate these steroid-related molecular mechanisms across vertebrates; we have done some in green anoles (see §3).

Other Z-genes we have investigated include secretory carrier membrane protein 1 (SCAMP1). Increases in its mRNA in males compared with females are specific to HVC and RA within the song system [28]. The number of cells expressing SCAMP1 protein in these regions is also sexually dimorphic, and the sex difference increases with age: males show a gradual increase between post-hatching day 25 and adulthood while females show a decrease during the same period [29]. These cells also increase in the Area X of males as they mature. SCAMP1, along with related genes in the same family, encode proteins involved in vesicle trafficking [30]. They function as carriers in the cell surface recycling system [31], and have been implicated in both exocytosis and endocytosis [30–33]. Another candidate Z-gene is sorting nexin 2 (SNX2). Its mRNA is increased specifically in the HVC and Area X of males at post-hatching day 25 [22], data paralleled by sex differences in the number and/or density of cells in these regions expressing the protein [34]. SNX2 is involved in recycling of membrane proteins [35]. 17β-Hydroxysteroid dehydrogenase IV (HSD17B4), which is a steroid inactivating enzyme involved in the conversion of E2 to oestrone, also exhibits increased expression in the HVC of developing male compared with female zebra finches [36]. At this time, no information regarding specific roles of these genes in masculinization of the song system is available. However, each of these Z-genes (RPL17, RPL37, SCAMP1, SNX2 and 17HSDB4) exhibits substantial co-expression with androgen receptors (AR) within song control regions. It is unclear whether this relationship is related to sexual differentiation, but it warrants further investigation in the context of the idea that steroid hormones and Z-genes are both important for masculinization. While the role of androgen acting at AR in song system development appears limited [8], one study suggests that an E2-induced increase in AR may be important for masculinization of morphology [37], and androgen is important for the development of stereotyped song [38].

Our more recent work has focused on tubulin chaperone protein A (TBCA). This gene is interesting because, while widely expressed, among the song nuclei investigated male-biased increases in mRNA and protein are detected only in LMAN [39]. This region is not particularly sexually dimorphic in adult zebra finches, but its projection to RA is (see above). Because TBCA is critical for microtubule formation, we developed hypotheses related to increased TBCA expression in developing males enhancing the structure or function of the LMAN to RA pathway. Consistent with this idea, TBCA mRNA is increased in males compared with females in the juvenile LMAN from at least post-hatching day 12 through day 45, but it is not different between the sexes in adulthood. The pattern of protein detected in LMAN via western blot analysis is similar, with a large sex difference at day 25 and none in adulthood [40]. The total number of cells expressing TBCA protein declines with age, similar to the overall levels of mRNA and protein expression, however it is not sexually dimorphic. These data suggest that the quantity of protein per cell is what differs between juvenile males and females. The majority of the TBCA⁺ cells in LMAN are neurons that project to RA [40].

TBCA is the one gene for which we have completed studies manipulating expression to determine effects on masculinization. We are particularly interested in not just the role of Z-genes themselves, but also potential interactions with E2. For example, E2 and Z-genes might have additive or interactive effects, such that both adequate E2 and Z-gene dosage are required for full masculinization. In addition, E2 might increase the expression of particular Z-genes, or a Z-gene might influence how the brain responds to the hormone. TBCA mRNA and protein in LMAN are not affected by exposure to E2 [41]. However, inhibiting TBCA in LMAN using site-specific injections of siRNA in juvenile males and females decreases the size of the projection between LMAN and RA, soma size and cell number in RA, as well as the volume of RA overall. It also decreases LMAN volume in females specifically. These studies were completed in males exposed to the oestrogen synthesis inhibitor fadrozole and in females treated with E2 (as well as appropriate controls). As no interactions with these endocrine manipulations were detected, it is clear that additive and interactive effects of TBCA and E2 do not exist. However, the data show that increased TBCA in the juvenile LMAN is important for masculinization of the projections to its target RA and various aspects of RA morphology [41].

It is worth noting that fadrozole treatment in this study caused some hypermasculinization of the song system in juvenile males [41]. Similar effects have been detected with manipulations that inhibit E2's ability to act [42,43]. In addition, treatment of males with E2 causes effects opposite to the same manipulation in females in some cases, including demasculinization of HSD17B4 expression [36] and reduced incorporation/survival of new cells in the developing song system [44]. The mechanisms and functional consequences of these effects of exogenous E2 manipulations in males are unclear, but they document instances of differential responsiveness to this steroid between the sexes. The genetic mechanisms associated with these effects should be investigated.

Several laboratories have also become interested in brain-derived neurotrophic factor (BDNF) as a critical mechanism associated with both sexual differentiation and adult plasticity in songbirds. BDNF is not on a sex chromosome, but its high affinity receptor tyrosine kinase B (trkB) is Z-linked. Thus, increased expression of trkB in males could enhance the action of BDNF. This neurotrophin facilitates cell survival in the developing zebra finch song system [45]. The trkB expression is increased in the song system of developing males compared with females [46]. Like trkB, BDNF is expressed in song control nuclei, and developmental changes in the number of cells expressing this protein in HVC and RA are consistent with roles in later stages of song learning and the production of mature song [47]. In addition, the HVC and RA of juvenile males contain more cells expressing BDNF protein than females, and E2 increases this measure in females [48]. E2 also increases BDNF mRNA in juvenile males and females [49]. Thus, one way that E2 and Z-genes may work together to enhance cell survival could involve E2 upregulating BDNF which acts on the increased level of trkB receptor in males. Expression of other Z-genes may facilitate this process, for example, with TBCA increasing BDNF transport and SCAMP1 increasing its release. Recent data involving experimental upregulation of BDNF indicate that enhanced expression in HVC increases copying of song syllables [50]. In parallel, prevention of the development of normal song via transection of the nerve that innervates the syrinx results in reduced BDNF expression specifically in HVC and RA, as well as a diminished projection from HVC to RA in juvenile zebra finches [51]. These results indicate the importance of the BDNF forebrain motor pathway for song development.

BDNF expression is also important in functional and morphological change in adulthood. In some songbirds, new neurons are incorporated into the HVC on a seasonal basis, associated with a rise in T. In canaries, trkB is present in the HVC of both sexes, but BDNF protein is only detectable in this region in males. However, infusion of BDNF into this region in females facilitates the recruitment and survival of new neurons. T-treatment of females also increases BDNF expression [52]. In adult male canaries, singing increases the survival of these new cells within HVC, as well as expression of BDNF in RA-projecting neurons within HVC [53]. Similar to canaries, conditions typical of the breeding season (increased T and a long photoperiod) upregulate BDNF expression in the HVC of white-crowned sparrows. The results of manipulating availability of BDNF and trkB in the RA of this species are consistent with the idea that BDNF is transported from HVC to RA, where it induces growth typical of the breeding season [54]. Collectively, the data suggest a suite of relationships among T, BDNF and singing behaviour, all facilitated by BDNF acting at the Z-gene trkB.

3. Lizards

Compared with mammalian and avian systems, relatively little work has been done on the mechanisms regulating sexual differentiation of behaviour and related structures in reptiles. However, a few studies provide intriguing evidence that genetic mechanisms, perhaps related to sex chromosome complement, may be important.

One of the most striking examples involves the copulatory system of leopard geckos. These animals have no sex chromosomes; their gonadal sex is determined by the temperature of egg incubation [55]. Male lizards possess paired copulatory organs (hemipenises), which are controlled by muscles in the tail, including the transversus penis (TPN) and the retractor penis magnus (RPM). In green anoles, which have genetic sex determination (see below), these structures are completely absent in adult females [56]. While these structures are present in both sexes early in development, they regress in females prior to hatching [57], and they cannot be resurrected either in juveniles [58] or adults [59] with T-treatment. While not investigated in depth in other reptiles, this situation is similar to the rodent copulatory neuromuscular system in which the structures are absent in adult females [60]. By contrast, adult female leopard geckos possess vestigial hemipenises and copulatory muscles. Interestingly, these structures are highly sensitive to androgens in adulthood, such that the muscle fibres approximately double in size and hemipenis cross-sectional area increases by an order of magnitude in females treated with T [61]. Certainly, we cannot draw conclusions based on one species without sex chromosomes, but the unusual presence of intersex characteristics and substantial steroid-induced morphological plasticity in adulthood suggests that the absence of sex chromosomes may result in less sexual differentiation of these peripheral structures. This idea warrants further investigation with a broader selection of species.

Wennstrom et al. [62] used a parthenogenetic species of whiptail lizard, Cnemidophorus (now Aspidoscelis) uniparens, to tease apart the importance of sex chromosomes from gonadal tissue on sex differences in the brain. These animals are genetically female (triploid—XXX), but as in zebra finches, treatment of embryos with fadrozole results in the development of testes rather than ovaries (as well as formation of other masculine reproductive structures dependent on androgen). In the ancestral C. inornatus, which has individuals of both sexes, the anterior hypothalamus/preoptic area (AH-POA) is larger in males and the ventromedial hypothalamus (VMH) is larger in females [63]. These sex differences parallel behaviour, as the AH-POA is important for male sexual displays [64] and the VMH is critical for female receptivity [65]. However, in the parthenogens, both brain regions are equivalent in volume in fadrozole-treated individuals with testes and unmanipulated lizards with ovaries [62]. These results raise the possibility that sex differences in the brains of whiptail lizards are controlled by sex chromosome complement rather than organization by gonadal steroids. Data from adult manipulations of environmental condition and T availability across the two whiptail species suggest that the sizes of these structures are in fact not permanently organized in development and that both this steroid hormone and factors intrinsic to the tissues are important. Specifically, while the AH-POA and VMH do not change in size across reproductive states in females of the sexual species (C. inornatus) or the parthenogen (C. uniparens), both are demasculinized/feminized in male C. inornatus following gonadectomy or removal to winter hibernation conditions of shorter days and colder temperatures [66]. In parallel, the AH-POA increases and the VMH decreases in size in adult castrated T-treated males compared with those without hormone replacement, whereas this hormone treatment does not affect the morphology of these brain regions in female C. inornatus or C. uniparens [67]. This variability in responsiveness to T suggests a sex difference at the genetic level and parallels a variety of pieces of data collected from green anoles.

Male anolis lizards exhibit a highly stereotyped set of reproductive behaviours, which include extension of a coloured throat fan (dewlap) used in courtship. Once a receptive female is attracted, the male mounts, holds the skin on the back of her neck between his teeth, and intromits one of the two hemipenises. Hundreds of species exist within this genus, although much of the work on behavioural neuroendocrinology has been conducted in green anoles (Anolis carolinensis). The sequence of male sexual behaviours is in part controlled by a seasonal rise in T; green anoles breed in the field from approximately April through July [68]. We have pursued investigations at three levels of the nervous system required for the suite of these behaviours to be displayed (figure 2). As in other vertebrates, regions of the forebrain are important for masculine and feminine sexual behaviours—the POA and ventromedial nucleus of the amygdala (AMY) are critical for male behaviour, and the VMH is involved in female receptivity (see above and [70]). The volume of the POA is greater in males than females [71], and across these three forebrain regions several effects on morphology in adulthood have been detected that are related to season, hormone exposure and/or presence of the gonads [68]. While the individual effects are somewhat complicated and difficult to interpret without further work, this adult plasticity suggests that the brain regions are not permanently organized in development. Clearer relationships between structure and function are present in the two neuromuscular systems regulating male reproduction.

Figure 2.

Structures important for the suite of sexual behaviours displayed by male anole lizards. The photograph is of a male Anolis bahorucoensis in the Dominican Republic, taken by Ariel Kahrl (Department of Biology, University of Virginia). This species is particularly interesting because its dewlaps are more similar in size between the sexes than in many species within this genus. In these tiny animals, the dewlap is quite small in both males and females; its use is relatively infrequent and sexually monomorphic. In parallel, ceratohyoid fibre size does not differ between the sexes, although the length of the dewlap cartilage is somewhat larger in males than females [69]. POA, preoptic area; AMY, ventromedial amygdala; VMH, ventromedial hypothalamus; AmbIX/VIImv, the combined region consisting of the glossopharyngeal portion of nucleus ambiguus and the ventral motor nucleus of the facial nerve; AmbX, vagal component of nucleus ambiguus; T17, caudal segment of the trunk portion of the cord; S1, rostral portion of the sacral cord; TPN, transversus penis; and RPM, retractor penis magnus.

Dewlap use is controlled by contraction of the ceratohyoid muscles in the throat, which results in extension of cartilage that lies underneath this skin. The motoneurons for these muscles are located in the vagal component of nucleus ambiguus (AmbX) and the region containing the glossopharyngeal portion of nucleus ambiguus and the ventral motor nucleus of the facial nerve (AmbIX/VIImv) [72–76]. Overall, dewlaps are much larger in male than female green anoles, and males extend theirs far more often, even when circulating T is controlled. In parallel, cartilage length, and the sizes of motoneuron somas, muscle fibres and neuromuscular junctions are all greater in males than females [68]. At least for muscle fibre and cartilage size, these sex differences in morphology exist across a variety of anole species in which males extend their dewlaps more frequently than females [69]. As indicated above, control of the copulatory organs involves a relatively simple neuromuscular system with muscles and peripheral organs present only in males. The motoneurons innervating the TPN and RPM are located in the trunk and sacral portions of the spinal cord (T17-S1) and, parallel to the sexual dimorphism of the muscles, these cells are larger and more numerous in adult males than females [56,77].

Members of my laboratory have discovered an array of effects of T on behaviour and morphology, and are particularly intrigued that responsiveness to this hormone changes seasonally. For example, the same dose of T administered to castrated males under non-breeding (decreased temperature and photoperiod) compared with breeding conditions has reduced effects on courtship and copulation, hemipenis and RPM fibre size, soma size in the AMY, and brain oestrogen synthesis (activity of the aromatase enzyme) [59,78,79]. The effect of T on aromatase activity occurs only in males; the hormone does not increase it in females [79]. Differences also exist between the dewlap and copulatory neuromuscular systems in that no effects of T have been detected on ceratohyoid muscle fibre size, whereas T increases RPM fibre size (with a greater effect in the breeding than non-breeding season) [78]. These results suggest that T levels are not the sole factor controlling changes in structure and function across the areas responsible for sexual behaviour in anoles.

Differential expression of steroid metabolizing enzymes in specific brain regions could affect local availability of E2 and dihydrotestosterone (DHT) that might influence behaviour. Aromatase regulates the synthesis of E2 from T, and 5α-reductase metabolizes T into DHT. T itself is the primary activator of male sexual behaviour in green anoles, with DHT playing little if any role; E2 is critical for reproduction in females. However, T in females appears to facilitate receptivity, and E2 in males may increase motivation to copulate [68]. Consistent with this latter role, males have a greater total number of cells that express aromatase mRNA in the POA compared with females. The density of aromatase-expressing cells is also higher in the breeding compared with non-breeding season in the POA [80]. Two forms of 5α-reductase are present in the anole brain. 5α-Reductase I mRNA is detected only in the brainstem, whereas 5α-reductase II is detected in discrete regions throughout the brain. Parallel to the lack of effect of DHT in activating masculine behaviour [68], the number of 5α-reductase II expressing cells is equivalent across the sexes and seasons in the POA and AMY [81].

Differential expression of receptors could mediate responsiveness to a hormone. Oestrogen receptor alpha (ERα) mRNA is expressed in the POA, VMH and AMY of green anoles across both sexes and seasons. However, expression is approximately threefold greater in the VMH compared with the POA and AMY. In the POA and VMH, expression is higher in females compared with males, independent of season [82]. ERβ has also been detected in discrete areas throughout the anole brain, with high levels in limbic regions. Females have a greater density of ERβ-expressing cells in the AMY and VMH than males [83]. Collectively, these data are consistent with the prominent role for E2 in female sexual behaviour. However, combined with the fact that T acting as an androgen is the primary mediator of male sexual behaviour, these data are not consistent with a role for ER in seasonal responsiveness to T in males. AR are more likely to play that role.

AR are expressed in the forebrain (POA, AMY and VMH), as well as in the muscles and motoneurons of both the dewlap and copulatory neuromuscular systems [84–86]. However, patterns of expression are generally not consistent with the idea that differences in AR availability regulate seasonal changes in responsiveness to T. For example, among gonadectomized animals in which some were treated with T, greater levels of AR protein in POA/hypothalamic dissections were detected in males compared with females, specifically within the breeding season. However, no significant effects of sex or season were detected in intact animals on either this measure or on the number of AR-expressing cells in the POA, AMY or VMH [87]. The number of nuclei positive for AR protein also does not vary across season in gonadally intact animals in either copulatory or dewlap muscles [86]. However, qPCR on the RPM indicates an increase in AR mRNA in the intact males from the breeding compared with non-breeding season [84]. When greater differences in circulating T are created by castrating males and providing either a T-filled or empty implant, T increases the percentage of AR-positive nuclei in the copulatory muscles of males from both the breeding and non-breeding seasons, but does not affect this measure in the dewlap (ceratohyoid) muscle [86]. This pattern of effects of T on AR protein suggests that upregulation of the receptors may be important for enhancing muscle morphology. However, the fact that T has a greater effect on increasing RPM fibre size in the breeding than non-breeding season (see above) without a differential effect on increasing AR expression, suggests that other mechanisms must be responsible for this seasonal change in responsiveness to T in this tissue. In sum, because these neural and muscular structures involved in green anole reproduction broadly express AR, and they do so without a pattern that is generally consistent with differences in responsiveness to T across tissues, sexes and seasons, it seems likely that other mechanisms are responsible for this feature.

Interestingly, within the hemipenis itself, transcriptional activity appears generally diminished in males from the non-breeding compared with breeding season. These results from multiple genes, including AR and housekeeping genes, suggest that the organs may decrease numerous functions when they are not needed for copulation. T-treatment increases AR protein extracted from the hemipenis, and to a greater extent in the breeding than non-breeding season [84]. These data suggest that T-induced changes at this most peripheral level critical for copulation in males could influence seasonal responsiveness in the display of sexual behaviour.

Our more recent experiments have focused on steroid receptor coactivators: SRC-1 and CBP are both important for increasing AR action [88]. Experiments using homogenates of peripheral copulatory system tissues generally do not support the idea that seasonal changes in the expression of either gene or the complementary proteins specifically cause an increase in responsiveness to T in the breeding season [84]. However, in situ hybridization studies on the brains of adult green anoles reveal more SRC-1 expressing cells in the POA of gonadally intact males compared with females. In the VMH, males also have more cells that express SRC-1 than females, and males exhibit an increase in the breeding compared with the non-breeding season. T-treatment of gonadectomized animals increases the number of cells expressing this coactivator in the POA and AMY, although no differences due to sex or season are detected in intact animals in the AMY. These results suggest that T selectively regulates SRC-1, and that this coactivator may influence the activity of these brain regions important for the display of sexual behaviour. However, changes in SRC-1 expression are probably not responsible for the seasonal change in responsiveness to T [89]. CBP is also expressed at relatively high levels in steroid-sensitive brain regions. In the AMY, mRNA levels are nearly twice as high in gonadally intact females compared with males. By contrast, CBP expression does not differ across seasons or hormone manipulation in this brain region. This pattern suggests that CBP might influence female-biased functions controlled by the AMY, but is not consistent with a role in mediating seasonal differences in responsiveness to T in these areas associated with reproductive function [90].

To date, our investigations into steroid-related molecular mechanisms (receptors and co-activators) have not provided many satisfying explanations for differences in morphology and behaviour related to sex or breeding season. Other avenues should be pursued, and the search is easier now that the green anole genome has been sequenced [91]. However, genomic information for this species lags behind data available for rodents and songbirds. In fact, prior to fluorescent in situ hybridization experiments conducted in conjunction with the genome sequencing, it had even been difficult to identify sex chromosomes in a variety of anole species. Sex chromosomes in green anoles and a number of other species within this genus appear homomorphic, although karyotyping indicates that other species appear to have varying numbers of X chromosomes in addition to a Y [92,93]. Now, it is clear that two copies of an X chromosome are present in female green anoles and one copy is present in males [91]; the Y has yet to be conclusively identified. Early work on gene dosage indicates male : female ratios of unity for autosomal genes and 0.5 for those quantified from the X-chromosome [92,94], suggesting limited dosage compensation. These data raise the possibility that the presence of a Y-chromosome in only males or two X-chromosomes in females might be important for sexual differentiation of structure or function. These ideas have yet to be pursued.

4. Genetic mechanisms and sex chromosome effects in reproductive structure and function

Steroid hormones play critical roles in reproductive structure and function in vertebrates. They exert a diverse array of effects on anatomy and behaviour across the life cycle. However, the research on birds and lizards described in §§2 and 3 highlights an array of examples in which responses to steroid hormones differ: between the sexes, across seasons and even between tissues within the same organism. In addition, predicted opposing effects do not always occur (e.g. increasing E2 masculinizes the female song system, but decreasing E2 in males does not inhibit masculinization). Some parallels between expression of steroid-related molecules (receptors, metabolizing enzymes and co-activators) and variability in hormone responsiveness have been detected. Overall, however, the evidence from birds and lizards suggests that other mechanisms must be involved to fully explain the differences in steroid function. Identifying critical factors across seasons and tissues may be more challenging, but sex chromosome genes offer logical candidates for differences between the sexes. Results from my laboratory and those of several other investigators have identified some promising avenues for future research on song system masculinization, including TBCA and trkB/BDNF.

In addition, taking a broader comparative approach to investigate the roles of sex chromosomes in the generation of sex differences in structure and function should have substantial value. Lizards could be particularly useful in this regard, as examples of both XX/XY and ZZ/ZW systems exist, as well as species lacking in sex chromosomes altogether. However, it will be challenging to do the types of morphological and behavioural experiments necessary across a sufficient number and diversity of species to draw strong conclusions. One idea is that the homogametic sex tends to be ‘neutral’ and more easily reversed both in terms of reproductive morphology and behaviour: the heterogametic sex produces the hormones that organize these sex differences [95]. Analyses across species, particularly now that molecular data can provide more concrete information on evolutionary relationships, should prove quite useful in understanding the breadth of this concept.

The type and/or presence of sex chromosomes can provide important clues as to genetic mechanisms, but it is identification of specific, critical molecules that will be of most use from a biomedical perspective. The fact that the primary difference between male and female zebra finches is their sex chromosomes allows us to identify candidate genes; for the purposes of elucidating molecular mechanisms, it does not matter that genes of interest are on a sex chromosome in birds and not in humans. Understanding the roles of particular genes, and how they may function in conjunction with factors, such as the ligand BDNF, will help us to identify mechanisms that contribute to development irrespective of chromosome location.

Sex chromosome genes have evolved independently multiple times [96], and the avian Z is not homologous to the human X or Y [97]. Sex chromosomes in reptiles, mammals and birds are derived from different ancient autosomes [98]. Thus, while comparative approaches to understanding the role of sex chromosome complement are interesting and important from a basic perspective, these types of data from a variety of model systems will probably not generalize to specific effects in humans. However, there should be substantial gains from using the power of particular model systems to identify candidate genes critical to development and plasticity of brain and behaviour that will be broadly applicable. As genomic data from diverse organisms are becoming more available, this approach of identifying and characterizing candidate genes will become all the easier. It is also important to analyse gene networks and patterns of change on larger scales [99].

Competing interests

I have no competing interests.

Funding

Research in my laboratory on zebra finches and green anole lizards has been supported by the National Institutes of Health (currently R01-MH096705) and National Science Foundation (most recently IOS-0742833).