Quantification of desiccation resistance, performed through air exposure of differently aged eggs (Fig. 1, see items 2 and 3 of Methods for details), resulted in complete shrinkage of eggs up to 11 HAE (HAE = hours after egglaying) (Fig. 1A, B). In contrast virtually all the eggs 13 HAE or older remained intact (Fig. 1A, C). It should be note that in our conditions Ae. aegypti total embryonic development takes 61.5 hours at 28°C (Farnesi and Rezende, to be published elsewhere).

The serosal cuticle, but not mosquito chorion, is resistant to chlorine digestion 4613, and this approach was used to further evaluate 11 and 13 HAE eggs (item 5 of Methods). Eleven HAE eggs left in the chlorine solution (bleach) for 10–20 minutes start to disintegrate (Fig. 1D) and if left for a longer period (30 minutes) disintegrate completely, both eggshell and embryo (data not shown). In contrast a 13 HAE embryo exposed to bleach for 30 minutes remains intact and is clearly visible beneath the transparent serosal cuticle, after almost complete digestion of the chorion (Fig. 1E). Taken together, these results indicate the SC is associated to desiccation resistance acquisition that abruptly appears between 11 and 13 hours after egglaying.

To evaluate the physiological relevance of desiccation resistance acquisition, we quantified the viability of developing eggs transferred to dry conditions at different ages and for varying periods of time, from 25 to 120 hours (Fig. 2, see item 4 of Methods for details). No hatching was observed if eggs were exposed to dry conditions before SC formation (10 HAE, Fig 2A). In contrast, samples dried at 20 HAE were fully viable (Fig 2C). Intermediate values were obtained if eggs were transferred to dry conditions in a period close to SC formation (15 HAE, Fig 2B). In this case viability was inversely proportional to the period of time spent in dry conditions. Since 120 hours on dry, starting from 20 HAE, corresponds to approximately 3 days after the end of embryogenesis, we conclude that the fully formed SC at 20 HAE is competent to impede embryonic and larvae desiccation during and after completion of embryogenesis (see Discussion).

In order to determine the embryonic development stage related to SC formation, embryos were clarified as detailed in Methods, item 6 (Figure 3). Embryos at 8 HAE exhibited the beginning of a morphological differentiation of the serosa anlage, immediately after the cellular blastoderm stage (Fig. 3A). This serosa anlage was also deduced in An. gambiae embryos from the expression of serosa specific genes 3233 (see also Discussion). Subsequently (9-HAE embryos) germ band extension starts, and the serosa appears fully formed, completely surrounding the embryo (Fig. 3B, B'). At 13 HAE, the germ band extension is complete, as noted by the adjacent positions of the embryo head and its posterior end (at the dorsal embryo side), and the serosal cuticle is being formed (Fig. 3C). At 15 HAE the serosal cuticle is completely formed (Fig. 3D, see also Figures 1 and 4). The scheme in D' was deduced from the results shown in Figure 1 and 4 (see below).

Different methods were used to confirm the presence of chitin in mosquito SC, as previously stated by Beckel 4 and Harwood 13, see item 7 of Methods. The first approach consisted of labeling with Calcofluor White, a fluorescent molecule that binds β-1,4-linked D-glycopyranoside units 35. Since some mosquito embryonic structures have autofluorescence 36, controls were elaborated to confirm specific Calcofluor labeling. Figure 4 compares fluorescence of embryos before or after incubation with Calcofluor. In both cases fluorescence images are shown either alone or together with DIC microscopy. Clarified 9-HAE embryos incubated with Calcofluor display a very faint fluorescence in serosal cells (Fig. 4A–D) whereas 15-HAE embryos in the same conditions exhibit specific fluorescence in the SC (Figs. 4E–H). Moreover, isolated serosal cuticle derived from fully developed embryos exhibit intense fluorescence (Fig. 4I–L). Isolated SCs were also submitted to staining with WGA, a highly specific lectin for N-acetyl-D-glucosamine polymers 3738, showing uniform labeling (Fig. 4M–P).

Finally, quantitation of chitin content of isolated SCs was performed according to Lehmann and White 39. With this approach we detected 3.24 (± 0.08) ng of chitin, as corresponding glucosamine, for each isolated SC.

There are two CHS genes in the Ae. aegypti genome 40, AaCHS1 and AaCHS2 (see item 8 of Methods for details of nomenclature). The latter, a Class B CHS gene, is associated with chitin production in the midgut 41. Specific primers could not detect expression of AaCHS2 during Ae. aegypti embryogenesis (see Additional File 1, Figure 1)

Before the release of the Ae. aegypti genome 40, degenerated primers designed to amplify a conserved region of Dipteran CHS-A genes were applied on Ae. aegypti genomic DNA to look for a putative AaCHS1 gene. An AaCHS1 fragment was obtained and employed to design specific primers that detected the presence of both AaCHS1a and AaCHS1b mRNA in embryonic development (see Additional File 1, Figure 1). Afterwards, using Drosophila melanogaster kkv (Dm_kkv) sequence as the subject, the entire putative AaCHS1 was found in the Ae. aegypti genome project. Its deduced amino acid sequence was compared with An. gambiae CHS-A (AgCHS1) and Dm_kkv, demonstrating a high degree of similarity among dipterans (see Additional file 2, Figure 2). The schematic representation of the genomic structure of AaCHS1 is shown in Fig. 5A. The first intron contains an open reading frame similar to the pol-like MosquI-Aa2 protein, a non-LTR retrotransposon element 42 (data not shown).

Details of the AaCHS1 mutually exclusive exon 6a/6b organization, sequence and deduced transmembrane domains are displayed in Fig. 5B–D. The AaCHS1 protein can be subdivided into three domains: an N-terminal region containing eight transmembrane domains, a central cytosolic region with the catalytic domain, characterized by the QRRRW chitin synthase signature motif 19, and a C-terminal region containing seven additional transmembrane domains. Five transmembrane domains of the C-terminal region are clustered together close to the catalytic domain whereas the two other transmembrane domains are located 217 amino acids downstream. Both exons, 6a and 6b, translate 59 amino acids that encompass the penultimate transmembrane domain of the protein (Fig. 5C, D). The amino acid sequences encoded by exons 6a and 6b have 78% identity (Fig. 5C).

Messenger RNA abundance in differently aged eggs was analyzed through qPCR with primers for AaCHS1a or AaCHS1b. Embryos belonging to three groups, related both to the kinetics of desiccation resistance acquisition and SC formation, were evaluated before (6 and 9 HAE), during (12 HAE) and after (24, 31 and 52 HAE) the abrupt desiccation resistance acquisition, exhibited in Figure 1A. The AaCHS1 alternative spliced forms are differentially expressed during Ae. aegypti embryogenesis (Fig. 6). A slight expression of AaCHS1a is noted at 9 HAE, increasing significantly up to 24 HAE, a period that encompasses the complete serosa formation, at 9 HAE (Fig. 3B), and the abrupt desiccation resistance acquisition (11–13 HAE, Fig. 1A). Thus, AaCHS1a expression attains its higher rate of increase after the serosa has completely surrounded the embryo. This corresponds to the moment of serosal cuticle synthesis (see Fig. 4C and 4G). In contrast, there is no AaCHS1b expression up to 12 HAE, and its mRNA is only slightly expressed at 24 HAE, both isoforms being expressed at late embryogenesis, at 31 and 52 HAE, respectively at the moments of dorsal closure and embryo organogenesis (see Additional File 3, Figure 3).

Beckel 4 and Harwood 13 have previously pointed to the presence of chitin in Aedes SC. We have conclusively confirmed this issue through Calcofluor and WGA labeling and direct chitin quantitation. In a recent paper Moreira et al conclude that a chitin-like material is present in eggshells right from oviposition 18, possibly being part of the compound exochorion or the smooth endochorion 1214. As stated by Moreira et al, their data "do not rule out a subsequent synthesis of chitin by serosa" 18 which, according to the data presented here seems to be the case.

A fragment of the AaCHS1 genomic DNA was sequenced, and the whole putative AaCHS1 gene (class CHS-A) was subsequently obtained from Ae. aegypti genome according to alignment of predicted amino acids with dipteran CHS-A genes. CHS-A orthologues from T. castaneum and M. sexta possess two splice variants 242526, these same variants being expressed during Ae. aegypti embryogenesis. The difference between both is a region of 59 amino acids codified by different exons in each alternative mRNA.

There is no experimental evidence regarding the specific use of any of the two transcripts. However, specific RNAi for each of the two T. castaneum transcripts, TcCHS1a and TcCHS1b, display distinct phenotypes when dsRNA specific for each transcript are injected into larval and pupal stages 25. TcCHS1b is required only during pupal-adult molt where its absence leads to death at this stage; TcCHS1a is required both for larval-pupal and pupal-adult moults, its lack leading to death at both stages, indicating distinct roles for each transcript 25. Moreover, during both T. castaneum and M. sexta development the transcript containing the first alternative exon is far more expressed than the mRNA with the second one 242526, a trend that was also observed during Ae. aegypti embryogenesis, see below.

Similar to the T. castaneum TcCHS2 gene 24, synthesis of Ae. aegypti CHS-B, AaCHS2 41, was not found during embryogenesis. Accordingly, RNAi experiments with T. castaneum larvae demonstrated that the CHS-B TcCHS2 gene is required only for peritrophic matrix chitin synthesis 25.

The high increase of AaCHS1a expression at 9–12 HAE is concomitant with the appearance of the chitinized SC (compare Figs. 1, 4 and 6), pointing out a specific role of AaCHS1a in chitin synthesis for the serosal cuticle by serosal cells, as indicated. It is highly unlikely that AaCHS1a early expression, from 9 until 24 HAE, is related to larval cuticle chitin synthesis: 24-hour old embryos are at the germ band retraction stage (data not shown), and cuticle formation starts later. Raminani and Cupp 43 showed Aedes aegypti larval cuticle production starts at 47–52% of development. Accordingly, the later expression of both AaCHS1a and AaCHS1b at 31 HAE occurs at 52% of the total embryogenesis time period. This late expression of both AaCHS1a and AaCHS1b, but not the early AaCHS1a expression, would have a role in organogenesis and larval cuticle synthesis, as described for D. melanogaster 294445. In this later species, Dm_kkv (CHS-A gene) deletion is embryonic lethal 2930, and the expression of kkv is concomitant with chitin production for the larval cuticle 29. These data suggest a crucial role of CHS-A genes in Dipteran embryogenesis. In D. melanogaster the gene kkv is expected to participate only in chitin synthesis of the larval cuticle, since Cyclorrhapha flies do not have a serosa (see below), therefore lacking a serosal cuticle.

In insects, immediately after blastoderm formation, cells of the developing egg follow distinct fates, originating the embryonic (germ) rudiment and the serosa anlage where the embryonic rudiment will form the embryo itself and the amnion 8464748. The serosa originates from an anterior or anterodorsal portion of the egg 8324749. Following gastrulation the serosa completely envelopes the embryo and the yolk 28105051. Dipterans from Nematocera and Brachycera taxa also form separate amnion and serosa, with the exception of the Cyclorrhapha (Brachycera) flies, such as Musca (Muscidae, Calyptratae) and Drosophila (Drosophilidae, Acalyptratae), that do not form amnion and serosa, only a reminiscent amnioserosa 105152. It is argued that the serosa is important for developing insect embryos, protecting it against bacterial challenge 53 or insecticides 54. The absence of a functional serosa results in a complete everted (inside out) embryo as observed with RNAi for Tc-Zen2 and Of-Zen in T. castaneum and O. fasciatus, respectively 4855. Putative roles of serosa cells on yolk digestion 956 or in embryonic excretion 9 were also suggested.

In some mosquitoes, as in other insects (see below), after enclosing the entire embryo the serosa secretes the serosal cuticle (SC), a layer that is located above the serosa and below the endochorion 24, see Figure 7. In Aedes mosquitoes it has been previously shown that SC formation is related to egg desiccation resistance 2457. Orthoptera, Coleoptera and Lepidoptera insects also possess a SC 91157, but its function in these insects is unknown. Some findings indicate that in the Coleoptera Ocypus olens the serosal cuticle may function as a water barrier for developing embryos 58. Since the basal Hexapoda clade Archaeognatha also possesses a SC 46, it can be inferred that this is a primitive trait in insects. Taking this into account the absence of a serosa, and in consequence of a SC, would be a derived trait in Cyclorrhapha flies such as D. melanogaster.

Apart from its importance as the subject of study of the evolution of extra-embryonic membranes in insects 47, the serosa possesses the seminal role in Ae. aegypti of secreting the serosal cuticle. This layer is incriminated as important for desiccation resistance in Aedes embryos and probably has a similar role in other mosquito genera as well. Aedes eggs, when compared to other mosquito genera, can survive on dry for a much longer period of time 2. We therefore speculate that SC contributes with different desiccation resistance levels in eggs from distinct Culicidae genera. We are currently working to address the existence and role of SC in other mosquitoes species.

We have shown that the extra-embryonic serosa completely surrounds Ae. aegypti embryos at the beginning of germ band extension, at 9 HAE, followed by abrupt SC formation, between 11 and 13 HAE. This 2-hour interval corresponds to only 3.2% of total embryonic development.

The exposure of embryos to dry conditions for long periods before SC formation also results in complete blockage of hatching. By contrast, if the same procedure is performed after SC formation, all larvae are viable, even when embryos are left on dry from 20 to 140 HAE (i.e.: around 3 days after the end of embryogenesis, Fig. 2). Since after completion of embryogenesis the dormant larvae can resist many months on dry 3, it should be argued that the observed rate of survival after 3 days on dry indicates that after 20 HAE, a developing embryo is putatively able to survive also for many months on dry, as suggested before 1. The data presented strongly argues in favor of an essential role of SC in protecting the developing embryo, and further the larvae, from desiccation. Irrespective of a direct relationship between SC formation and desiccation resistance acquisition, early embryos are not resistant to desiccation neither water impermeable, as recently claimed 1718.

It is important to underlie that neither the SC nor the synthesized chitin from SC may be solely responsible for impermeability, where the serosal cuticle-endochorion bonding would be the functional unity that confers desiccation resistance 4. Additionally, it was claimed that a wax layer, as a part of the serosal cuticle, would be an important factor for desiccation resistance in Ae. aegypti eggs 124, in the same way wax is important for impermeability to the integument (exoskeleton) of insects 5960. Indeed, Slifer 11 observed two layers in the serosal cuticle of the grasshopper Melanoplus differentialis: the "yellow cuticle", a thin externalmost layer secreted first and a thicker layer, so-called "white cuticle" lying beneath. Chitin is present in the white cuticle and it was suggested that the yellow cuticle has wax in its composition 116162. In the Hemiptera Rhodnius prolixus, a wax layer is deposited around the embryo following serosa formation, insuring further resistance against desiccation 63.

We thus propose that the final, mature desiccation resistance will be achieved by the interaction of proteins from the endochorion with proteins, lipids and sugars present in the serosal cuticle. These interactions could be orchestrated by the process of sclerotization 64, which would be happening at the SC-endochorion bonding region. Obviously these interactions would occur after the SC was shed, and could be responsible for the observed differences in the kinetics of events related to desiccation resistance here described.

In the case of Culicidae embryonic development, the dynamics and the nature of the physiological processes that underlie eggshell impermeability acquisition are still poorly understood. Additional experiments will be needed to functionally confirm the putative SC role on desiccation resistance, the role of SC-endochorion interactions in the impermeable status of mosquito eggs as well as the presence of a SC wax layer in Ae. aegypti.

Aedes aegypti mosquitoes from the Rockefeller strain, continuously reared in the laboratory at 26°C and 70% r.h., were adopted. Larvae received cat food (Friskies^®, Purina, Camaquã, RS, Brazil), and adults were fed ad libitum with 10% sucrose solution. Blood feeding, necessary for egg production, was employed after female mosquitoes were sugar deprived for 24 hours, with anesthetized guinea pigs.

This method was adapted from Valencia et al 65. Three to four days after blood-feeding, Ae. aegypti females were placed in a tube, anesthetized in ice for 1 minute and quickly transferred to a Whatman No. 1 paper disk, fit in the lid of a Petri dish (90 × 15 mm or 150 × 15 mm). The lid was then covered with the base of the dish, mosquitoes being allowed to recover for about 5–10 minutes. With the aid of the tip of a pipet introduced between the base and the lid, the filter paper was soaked with dechlorinated water, and the Petri dish was put in a humid chamber inside an incubator at 28°C protected from light. With this procedure egglaying started immediately. In all cases oviposition was permitted during 20-minutes, when Petri dishes were opened inside a large cage to release the mosquitoes. In some cases egglaying could be induced twice with the same females in the same oviposition cycle. Eggs were kept at 28°C until the required age, the onset being considered the end of the 20-minute egglaying period. Embryonic age was assigned as hours after egglaying (HAE). At 28°C Ae. aegypti embryogenesis is completed in 61.5 hours (Farnesi and Rezende, to be published elsewhere).

Eggs at different HAE were treated according to a protocol adapted from Valencia et al 6566: 50 eggs were aligned on a polycarbonate filter (2.5 cm diameter, 8 μm pore, Poretics Corporation), deposited above a drop of distilled water. The filter was then blotted on a Whatman No. 1 filter paper to remove all water, eggs were air-exposed during 15 minutes and the number of shrunken eggs was confirmed with a Zeiss stereomicroscope. The 15-minute air-drying was performed at 24–27°C with 40–45% relative humidity.

Synchronized developing eggs obtained as indicated in section 2 above were transferred from wet to dry conditions at different ages: 10, 15 and 20 HAE. In each case, groups of 40 eggs were kept on dry for varying periods: 25, 48 and 120 hours (see Fig. 2 for the assay design). Egg viability was then quantified through L1 larvae counting, after a hatching stimulus, performed according to Farnesi and Rezende (to be published elsewhere). In some experimental conditions the total test interval ("wet plus dry") remained below the presumptive embryogenesis completion period (61.5 hours). In these cases eggs were returned to a moist Whatman n°1 filter paper up to that period. For each experimental condition a parallel control sample was kept in moist filter paper up to 61.5 hours and then, if necessary, remained in the dry for a period of time equivalent to its experimental counterpart. The experiments were performed ate least in triplicates inside an incubator at 28°C with 40–80% relative humidity.

Eggs aged 11 and 13 HAE were treated with 30% NaOCl (containing approximately 3.6% active chlorine at final concentration) for 15–30 minutes and observed under a Zeiss stereomicroscope.

Synchronously laid eggs of different ages were fixed and clarified according to Trpis 67. Embryos inside the resulting transparent eggshells were observed under an Axioskop 40 microscope (Zeiss) with phase contrast and a Stereo Discovery V.12 stereoscope (Zeiss). Embryonic stages were identified in compliance with previous authors 3643526869.

Attempts to observe serosa nuclei with DAPI were unsuccessful due to embryo autofluorescence and incompatibility of the fixation protocol with DAPI staining.

Calcofluor White M2R (Sigma, St. Louis, MO, also named Fluostain I or Fluorescent Brightener 28, Molecular Formula: C₄₀H₄₄N₁₂O₁₀S₂, Catalog # F-3543), already used to identify chitin in yeast cells 70, nematode eggs 71 and D. melanogaster embryos 29, was available in our laboratory only in the acid form which is not directly soluble in water. An aqueous 1 mg/ml Calcofluor solution was prepared by the addition of some drops of 2 N KOH, resulting in a light yellow solution.

Embryos of 9 and 15 HAE were clarified and fixed as mentioned above. Serosal cuticle (SC) isolation from dormant, fully developed embryos was adapted from Harwood 13 and Judson and Hokama 6. Complete digestion of exo and endochorion of fully developed eggs was attained with 30% commercial sodium hypochlorite (NaOCl) during 25–30 minutes inside a Falcon cell strainer (70 μm Nylon, Becton Dickinson). The remaining SCs, with larvae inside, were then thoroughly washed with water. This last procedure breaks the SC, often at the line of dehiscence 6. It should be noted that larvae that leave the SC remain in the sample and serve as a positive control for subsequent Calcofluor labeling.

Embryos of 9 and 15 HAE together with isolated SCs were incubated with Calcofluor during 10 minutes in the dark, thoroughly washed with 25 mM pH 6.2 sodium phosphate buffer and analyzed under a Zeiss LSM 510 META Confocal laser scanning microscope with a violet diode excitation laser (405 nm) and a bandpass BP 420–480 nm emission filter.

Isolated SCs were also incubated with 5 μg/ml WGA-FITC (EY Laboratories) in a solution of PBS with 2% BSA (w/v) for 1 hour at room temperature. Serosal cuticles were then thoroughly washed in this solution, mounted and analyzed with dark field or fluorescence microscopy.

Chitin content of isolated SCs was determined as described elsewhere 3972.

We adopted the CHS nomenclature proposed by Merzendorfer 27, who renamed Ae. aegypti and An. gambiae genes: class A and class B genes are termed respectively CHS-1 and CHS-2, regardless of the gene discovery chronology. Accordingly, Ae. aegypti CHS described earlier by Ibrahim 41, is a class B gene, and will be referred to as AaCHS2.

PCR with degenerated primers (CSA3Fdeg: ACNAAYCCNTAYTGGAT and CSA3Rdeg: GGCCAYTTNACRTGDTA) was performed with Ae. aegypti genomic DNA to amplify CHS-A fragments homologous to the D. melanogaster kkv (GenBank Accession No. NM_079509) and the putative An. gambiae AgCHS1 (previously referred to as AgCHS2, GenBank Accession No. XM_321336). The amplified fragments were cloned and sequenced at Instituto Oswaldo Cruz on an ABI 377XL with a BigDye Terminator V3.0 (Applied Biosystems). Sequence analysis was conducted with GCG software (Wisconsin Package Version 10.0), NCBI website http://www.ncbi.nlm.nih.gov/BLAST and ClustalW.

Complete AaCHS1 DNA sequence information was initially obtained from a genomic Blast http://www.ncbi.nlm.nih.gov/sutils/genom_table.cgi?organism=insects using TBLASTN with the kkv (D. melanogaster CHS-A) sequence against Ae. aegypti genome. The genomic region of AaCHS1 was located inside a whole genome shotgun (WGS) sequence (GenBank accession No. AAGE02004132). A partial cds of AaCHS1 is deposited in GenBank (accession No. XM_001662150). This partial mRNA lacks the first exon plus the last putative 30 bases, before the stop codon of the deduced protein, as interpreted by the WGS sequence and alignment (see Fig. 4). Multiple alignments of deduced amino acid sequences were assembled using ClustalW software. The protein sequence was analyzed for transmembrane helices with the TMHMM v.2.0 software available at http://www.cbs.dtu.dk/services/TMHMM-2.0/.

Total RNA was extracted with Trizol reagent (Invitrogen) according to manufacturer instructions and cDNA was synthesized with the TaqMan Reverse Transcriptase Kit (Applied Biosystems) with the Not I-d(T)₁₈ primer.

Quantitative Real-time PCR (qPCR) was carried out in ABI Prism 7000 Sequence Detection Systems (Applied Biosystems) with the Power SYBR Green PCR Master Mix (Applied Biosystems).

The primers utilized for specific AaCHS1 expression (AaCHS1a or AaCHS1b, using Exon 6a or 6b, respectively, see Fig. 5b) were CS9ExA (5' CAAGGACAATATCCACGTCAAG 3') and CS9R (5' AGGCCAAGATATGCGACAG 3') for AaCHS1a, and CS10ExB (5' ACCTATGACGAAGTGACAGCC 3') and CS10R (5' TTGCAACCCCAGTTCAGCTC 3') for AaCHS1b. Primers 5aeexpRP and 3aeaquaRP1b 73 were employed to amplify the constitutive gene rp49, used as an internal normalizer for qPCR. 24 and 52 HAE values were designated as calibrators for the expression levels of AaCHS1a and AaCHS1b, respectively.