To help improve the functional annotation of the Anopheles gambiae genome we have generated the MozAtlas, a unified catalogue of tissue-specific gene expression from a single mosquito strain. In Drosophila melanogaster, cataloguing tissue expression patterns has been useful, especially for inferring biological functions, since the majority of genes encoded in the genome are not ubiquitously expressed . As with the fruit fly, Anopheles gene expression also exhibits substantial tissue specificity, with only a third of detectably expressed genes found in all tissues. Thus, the MozAtlas is a useful resource for better understanding the mosquito genome, providing direct evidence of genes with tissue restricted expression. Below we highlight the utility of MozAtlas for identifying classes of gene with tissue or sex-biased expression that may be exploited for vector control. Analysis of the MozAtlas also identifies gene expression features that are of interest from an evolutionary perspective, revealing both highly conserved and species-specific aspects of insect biology. Of particular interest, given that malaria parasites are only transmitted through female mosquitoes, we separately catalogued gene expression for each tissue in males and females, thus providing both tissue and sex-specific views of gene expression in the adult.
A major finding from our analysis is the substantial degree of sexually dimorphic gene expression we find at the tissue level: more than half of the genes for which we detect expression exhibit sexual dimorphism in terms of expression level. The head, in particular, has a significantly higher number of female-biased genes and of these, odorant receptors are significantly over-represented (Additional File 4). When searching for a blood-meal, female mosquitoes are attracted to odours emitted by humans, a behaviour mediated by receptors in the antennal sensilla . This activity is not exhibited by males, who feed entirely on nectar, and we presume that the female elevated expression of odorant binding molecules reflect this biology. The identification of molecules associated with female-specific aspects of odorant detection may provide targets for controlling malaria transmission .
We identified other sexually dimorphic expression signatures that appear to be associated with female characteristics, in particular, adaptation to hematophagy. For example in the female salivary gland we found an over-representation of genes with protein and lipid catabolic activity, ion transport and cellular homostasis functions. We suggest that these reflect the fact that, in females, the salivary gland produces compounds to disarm host hemostatic and immune responses, thus allowing mosquitoes to take a blood-meal. Similarly, many proteins found in the midgut are only synthesized by blood-feeding females [3, 21]: numerous digestive and proteolytic molecules implicated in blood digestion were identified as female elevated in our analysis.
In contrast, elevated male gene activity is largely associated with carbohydrate metabolism and ion transport activity. Since male mosquitoes feed entirely on sugar, these results were not surprising. However, somewhat more novel is that iron binding molecules are up-regulated in males. While in female mosquitoes iron is especially important for egg development and is strongly influenced by blood-feeding , iron metabolism has diverse physiological and developmental roles . Although females obtain iron from the blood meal, the sugar diet of males may necessitate more efficient iron uptake and up-regulation of genes that encode iron binding functions.
In both somatic and reproductive tissues, we identified genes with considerable specificity. Tightly controlled, tissue-specific expression is of interest for understanding the basic biology of a species, and is likely to be key in the development of next generation insect control agents. For example, genes uniquely expressed in particular tissues could be targets for inducing sterility or providing regulatory elements to drive localised expression of transgenes. In this respect, the highest proportion of Anopheles tissue-specific expression is in the testis, with approximately 10% of transcription uniquely detected in this tissue. Testis specific expression of genes with important roles in spermatogenesis, sperm competition or sperm-egg interactions present a set of targets with potential for inducing male sterility.
After mating, Anopheles females undergo distinct behavioural and physiological changes due to the transfer of both sperm and proteins produced in the male accessory glands : proteins secreted by males and passed to females in seminal fluid could provide a route for altering female fertility. Via specific expression profiling of accessory glands we have identified a new set of potential Anopheles Acp genes that will enable further investigation of sexual conflict within the mosquito. Sexual antagonism between males and females may be expected to cause rapid Acp sequence evolution . We find that among tissue-specific genes, those expressed in the accessory gland have a higher A/S ratio than in many tissues, including the testis. Slower evolutionary rates in the Anopheles testis might be explained, in part, by their mating behaviour: in polyandrous insects genes involved in spermatogenesis are often under strong positive selection as a result of post-copulatory male-male competition , whereas these pressures in the testis are expected to be absent from the largely monandrous Anopheles mosquitoes .
Genes with ovary specific expression provide potential targets for inducing female sterility in mosquitoes given that they are closely associated with egg formation. Chorion components of the fruit fly eggshell, for example, provide the embryo with protection from the physical environment, and disrupting their function causes female sterility . Recently, proteomic techniques have identified Anopheles eggshell constituents, several of which we find to be specifically expressed in the ovary, making them favourable candidates for use in population control .
In terms of genome structure, we show that genes with male-biased expression are non-randomly distributed around the Anopheles genome. Two mechanisms have been proposed to explain the disparity in chromosomal distribution of male expressed genes. First, during spermatogenesis the X chromosome of males becomes inactivated: since few testis genes are expressed post-meiotically, evidence suggests that chromosomal inactivation has promoted autosomal duplication events from X-linked genes [18, 29, 30]. There is compelling evidence that X-linked inactivation also occurs in nematodes  and mammals , however, an under-representation of male-biased somatically-expressed genes on the X chromosome indicates that other forces are also at work. Second, since males only have one X chromosome, polymorphisms beneficial to one sex may arise that are detrimental to the other sex. Such antagonistic sexual selection may eventually lead to sequence changes and demasculinization of the X chromosome , and consistent with this expectation, genes on the Anopheles
X chromosome have less sequence polymorphism than on the autosomes.
Identifying expression divergence within and between closely-related species provides important insights into the selective pressures underlying gene regulation [34, 35]. The opportunity to compare divergence between Drosophila and Anopheles, separated by some 250 million years of evolution, allows us to explore gene and tissue evolution over a considerable time scale. We find that expression similarity in one-to-one orthologues of the midgut, head, carcass and ovary expressed genes is well conserved in the Diptera and, as expected, genes in conserved co-expression clusters perform integral physiological functions.
In contrast, tissues such as the testis, often show considerable transcriptional variation between closely related species [36, 37]. It's been proposed that testis gene regulation plays a critical role in the initial formation of reproductive isolation . In addition to the Anopheles testis, expression in other tissues is also highly divergent: for example, expression in the Malpighian tubules is largely not conserved between Anopheles and Drosophila. As an organ with a key role in detoxification and osmoregulation, this divergence may reflect fundamental differences in the diet of each insect . In addition, salivary gland and male accessory gland expression cluster within rather than between species, evidence for a bout of simultaneous evolution since the last common ancestor was shared. Indeed, no significant co-expression was detected between species, indicating that secretory organ functions have diverged during the Dipteran split.
Recent Anopheles gene duplications are often expressed in the testis and, in Drosophila, extreme expansions also have spermatogenesis-related functions . As well as the testis, other tissues display narrow expression profiles of recent origin in Anopheles. Certainly, the blood meal imposes a range of challenges on the digestive system of mosquitoes and, in part, explains a predominance of gene duplications with salivary gland, Malpighian tubule or midgut expression. Even between members of the same mosquito subgenera, salivary proteins can diverge rapidly over time : our data suggests that this evolutionary pattern may also be common in Malpighian tubule proteins and, to a lesser extent, proteins within the midgut. However, specifically expressed genes in large families do not necessarily highlight unique functions, since homologues may perform the same or similar functions in a larger set of tissues. Gene families with single members are of interest for identifying unique processes, given that closely related homologues are not found within the genome. Narrowly expressed single-copy families were detected dating back to Metazoan and Hymenopteran clades, perhaps accompanying the emergence of differentiated organs. It will be of considerable interest for insect control programs to determine whether such proteins perform integral functions in their specific tissues, given that as single copies they should perform unique roles within the organism.