Gene Families of Cuticular Proteins Analogous to Peritrophins (CPAPs) in Tribolium castaneum Have Diverse Functions

The functional characterization of an entire class of 17 genes from the red flour beetle, Tribolium castaneum, which encode two families of Cuticular Proteins Analogous to Peritrophins (CPAPs) has been carried out. CPAP genes in T. castaneum are expressed exclusively in cuticle-forming tissues and have been classified into two families, CPAP1 and CPAP3, based on whether the proteins contain either one (CPAP1), or three copies (CPAP3) of the chitin-binding domain, ChtBD2, with its six characteristically spaced cysteine residues. Individual members of the TcCPAP1 and TcCPAP3 gene families have distinct developmental patterns of expression. Many of these proteins serve essential and non-redundant functions in maintaining the structural integrity of the cuticle in different parts of the insect anatomy. Three genes of the TcCPAP1 family and five genes of the TcCPAP3 family are essential for insect development, molting, cuticle integrity, proper locomotion or fecundity. RNA interference (RNAi) targeting TcCPAP1-C, TcCPAP1-H, TcCPAP1-J or TcCPAP3-C transcripts resulted in death at the pharate adult stage of development. RNAi for TcCPAP3-A1, TcCPAP3-B, TcCPAP3-D1 or TcCPAP3-D2 genes resulted in different developmental defects, including adult/embryonic mortality, abnormal elytra or hindwings, or an abnormal ‘stiff-jointed’ gait. These results provide experimental support for specialization in the functions of CPAP proteins in T. castaneum and a biological rationale for the conservation of CPAP orthologs in other orders of insects. This is the first comprehensive functional analysis of an entire class of cuticular proteins with one or more ChtBD2 domains in any insect species.


Introduction
Chitin, a linear polymer of ß-1-4 linked N-acetyl-D-glucosamine, is an important component of the insect exoskeleton. It is present in epidermal and tracheal cuticles and in the peritrophic matrix (PM) that lines the midgut. Insect procuticle is composed mainly of chitin complexed with proteins that contain chitinbinding domains. Two major groups of chitin-binding domains found in insect cuticular proteins have either the Rebers & Riddiford Consensus sequence (R&R Consensus; pfam00379; [1]) that lacks cysteine residues, or the peritrophin-A motif with six distinctly spaced cysteine residues (ChtBD2 domain, pfam 01607). Proteins with the ChtBD2 domain were found initially in extracts from insect PMs [2]. Additional genes encoding proteins with the ChtBD2 domain were found to be expressed in cuticle-forming tissues from Drosophila melanogaster and Ctenocephalides felis [3,4,5], suggesting that the peritrophin-A motif is not restricted to proteins associated only with the PM.
By conducting a bioinformatics search of the Tribolium castaneum (red flour beetle) genome, we identified 49 putative genes capable of encoding 50 proteins with one or more ChtBD2s [6]. These include several enzymes of chitin metabolism as well as several other proteins with 1-14 ChtBD2 domains. Almost all of them are predicted to have a cleavable signal peptide, suggesting that they are secreted and potentially capable of interacting with extracellular chitin. Among these 50 proteins, 11 were expressed in gut tissues but not in the carcass, and therefore are likely to be associated with chitin in the PM. These have been classified as Peritrophic Matrix Proteins (PMPs) [6]. Included among the remaining 39 predicted proteins with ChtBD2 motifs were two distinct families of proteins that were expressed only in cuticleforming tissues but not in the midgut. We have named these proteins ''Cuticular Protein Analogous to Peritrophins'' (CPAPs).
Members of these two families of proteins have either one or three ChtBD2 domains but no other identifiable protein domains. We have classified them as belonging to the CPAP1 family (ten proteins encoded by ten genes) each with only one ChtBD2 domain, or the CPAP3 family (eight proteins encoded by seven genes) each with three ChtBD2 domains. The CPAP1 family of proteins are of variable length and show amino acid sequence conservation only in the ChtBD2 domain that typically constitutes only a small fraction of the total length of each protein. In contrast, each of the CPAP3 family proteins consists of three N-terminal tandem copies of the ChtBD2 domain, separated by small conserved linker regions and a short, but variable, C-terminal region.
Members of the CPAP3 family of proteins are encoded by orthologs of D. melanogaster genes collectively known as the ''gasp'' or ''obstructor'' family [3,5]. We have assigned names to individual T. castaneum members of this family that reflect their orthology to the corresponding D. melanogaster obstructor gene (e.g. T. castaneum CPAP3-A is the ortholog of D. melanogaster Obst-A). Homologs belonging to a separate subgroup (subgroup II) consisting of obstructors F, G, H, I and J were also identified in D. melanogaster [3], but these are absent from the genome of T. castaneum [6]. Thus, there are five fewer members of the CPAP3 family in T. castaneum than in D. melanogaster.
Phylogenetic analysis of the T. castaneum CPAP proteins revealed that CPAP1 and CPAP3 represent two distinct branches of a tree that also includes a large assortment of peritrophins that have related ChtBD2 domains [6]. Genes encoding orthologs of the eight members of the CPAP3 family of proteins (CPAP3-A1, -A2, -B, -C5a, -C5b, -D1, -D2 and -E) are found in insects belonging to several different orders whose genomes have been fully sequenced and annotated, suggesting that these proteins are essential for insect survival or fitness. On the other hand, orthologs of the T. castaneum genes encoding CPAP1 proteins have been found in only some of those insect orders, perhaps suggesting that they are less critical for insect survival ( Table 1).
All members of the CPAP1 and CPAP3 families of T. castaneum genes are expressed in cuticle-forming tissues, but not in the PMsecreting midgut [6]. Furthermore, there are differences in the relative expression levels of these genes in two closely related cuticle-forming tissues with different mechanical properties, namely the elytron and the hindwing, indicating tissue-specific regulation of expression of CPAP genes [6]. Corresponding differences in tissue-specific expression of the obstructor family of genes in D. melanogaster (orthologs of TcCPAP3 family) have been reported previously [3]. Recently, a detailed study on the functions of one member of this family, obstructor A, from D. melanogaster has been carried out [7]. That study has shown that this gene is essential for embryonic development and affects tracheal tubule formation by preventing premature turnover of lumenal chitin. In this study, we have carried out an extensive expression and functional analyses of T. castnaeum genes encoding each of the eighteen proteins belonging to the CPAP1 or CPAP3 families in order to explore their roles in cuticle formation and integrity, insect survival and fecundity at different developmental stages. The results of RNAi experiments indicate that many of these genes have distinct and non-redundant essential functions in development and in maintenance of the structural integrity of cuticles with a wide range of physical properties.
Representative members for different insect orders, sub-phyla and phyla:

Insect Cultures
The T. castaneum strains used were the standard wild-type GA-1 and pu11, an enhancer trap line in which the gene for -enhanced green fluorescent protein (EGFP) is expressed in the wing and elytral discs at the pharate pupal stage, but not in earlier stages [8]. In T. castaneum, the different larval instars cannot be reliably distinguished. To circumvent this problem, the pu11 strain was used to enable the identification of animals in the penultimate or earlier larval instar and facilitate RNAi experiments at desired developmental stages. Insects were reared at 30uC in wheat flour containing 5% brewer's yeast under standard conditions as described previously [9].

Developmental Expression Profiles of the CPAP Gene Families
The developmental expression patterns of the ten CPAP1 genes and seven CPAP3 genes were analyzed by RT-PCR. Total RNA was isolated by using the RNeasy Mini kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. RNA sources included whole insects at various stages of development including embryos, penultimate instar larvae, last instar larvae, pharate pupae, pupae, young adults (0-3 hours after adult eclosion) and mature adults (10 days-old), as well as pupae collected at day-0 to day-5 post-pupation. First strand cDNA synthesis was carried out using the Superscript III first-strand synthesis system (Invitrogen, Carlsbad, CA) utilizing 1 mg of total RNA for each reaction. This cDNA served as template for subsequent PCR reactions using the gene-specific primers listed in Jasrapuria et al., [6].

Double-stranded RNA Synthesis and Injection
Double-stranded RNAs (dsRNAs) were synthesized from the PCR-generated cDNA templates using the Ampliscribe TM T7-Flash TM Kit (Epicentre Technologies, Madison, WI) as per the manufacturer's protocol. Unique regions with greatest sequence divergence for each of the TcCPAP1 and TcCPAP3 genes were chosen as templates for synthesis of dsRNAs. Their positions in the target gene and sizes are indicated in Supplementary Table S1. In most cases, RNAi experiments were replicated by choosing two non-overlapping regions as targets. To assess the functional specificity of the two isoforms of proteins encoded by TcCPAP3-C, dsRNAs corresponding to the entire sequence of either alternative exon 5a or 5b of this gene were synthesized. Doublestranded RNAs (200 ng per insect in 0.1 mM potassium phosphate buffer containing 5 mM KCl, pH 7) corresponding to each target gene region were injected into either penultimate or last instar larvae (n = 40 for each stage) as described previously [10]. A dsRNA for the T. castaneum tryptophan oxygenase gene (TcVermilion, TcVer) was used as a positive control for effectiveness of RNAi as well as a negative control for the effects of a non-target dsRNA. DsRNA for chitin synthase A (Chs-A), which produces a lethal phenotype at the pharate adult stage, was used as a nontarget gene control to determine if transcripts for this gene were affected by the dsRNA treatments.
To observe any effects of RNAi on fecundity, oviposition behavior or embryonic development, dsRNAs for each of the ten TcCPAP1 genes and the seven TcCPAP3 genes were injected singly into adult females (n = 20). Two days after injection, dsRNAtreated females were mated with an equal number of untreated adult males. These insects were maintained under normal conditions as described [11]. Eggs were collected every 3 days for ,1 month for determining fecundity, hatchability and chitin content. Eggs were dechorionated by treating with 1:1 diluted 5% Clorox solution before recording images using a Leica MZFLIII microscope.

Measurement of Transcripts after RNAi for CPAP Families of Genes
RT-PCR experiments were carried out to monitor targetspecific transcript depletion. Total RNA was isolated from whole insects 4 days after dsRNA injection using the RNeasy Mini kit (Qiagen, Valencia, CA). Three insects were pooled for each RNA extraction. cDNA synthesis and RT-PCR were performed using gene-specific primer pairs. The specificity of RNAi for the targeted CPAP genes was confirmed by comparing the transcript levels of both the targeted gene, several closely related non-targeted TcCPAP1 or TcCPAP3 genes, as well as the Chs-A gene. PCR amplification products using the same cDNA template and a pair of primers for the gene encoding ribosomal protein S6 (RpS6) were used as internal loading controls.
Scanning Electron Microscopy (SEM) of Elytra after RNAi for TcCPAP3-D Co-injection of dsRNAs for the TcCPAP3-D1 and TcCPAP3-D2 genes (200 ng of each dsRNA per insect; n = 6) was done at the pharate pupal stage to maximize the effect of RNAi of these paralogous genes. Elytra were dissected 10-days after adult eclosion and fixed by passing through a series of washing solutions and subjected to critical point drying. The elytral samples were coated with a 60% gold/40% palladium mixture using a DESK II Sputter/Etch unit (Denton Vacuum, LLC, Moorestown, NJ). Images were taken using the S-3500N scanning electron microscope equipped with the S-6542 Absorbed Electron Detector (Hitachi Science Systems, Ltd., Hitachinaka, Japan).
Chitin Staining of Elytra with FITC-CBD after RNAi for CPAP Genes dsRNA injections into pharate pupae were repeated for each CPAP gene that was shown by RNAi experiments to be required for normal adult eclosion. These included TcCPAP1-C, TcCPAP1-H, TcCPAP1-J and TcCPAP3-C with TcChs-A and TcVer serving as positive and negative controls, respectively. Elytra were dissected from pharate adults, washed in 16PBS, pH 8, and then heated in 10 M NaOH for 6 hours at 95uC to remove the proteins. The resulting mixture with the insoluble cuticular material was neutralized in 16PBS, pH 8, and stained at room temperature for 1 hour with a chitin probe consisting of the fluorescein isothiocyanate-conjugated chitin-binding domain from chitinase of B. circulans W12 (FITC-CBD, New England BioLabs, Beverly, MA) at a 1:500 dilution in 16PBS, pH 8. After washing off the excess FITC-CBD probe with 16PBS, the green fluorescence was recorded using a GFP filter (470 nm excitation and 515 nm emission wavelengths) and a Leica MZFLIII microscope.

Chitin Content Analysis of Insects after RNAi for CPAP1 and CPAP3 Family Genes
Double-stranded RNA for each of the TcCPAP1 or TcCPAP3 genes was injected into insects at the pharate pupal stage. The treated insects were collected 5 days post-pupation at the pharate adult stage just prior to adult eclosion. Individual pharate adults were weighed and placed in separate 1.5 ml screw-cap microcentrifuge tubes. The total chitin content of each of these insects was measured using a modified Morgan-Elson method as described previously [12].

Confocal Microscopic Analysis of dsRNA CPAP1-H-treated Insects
TcCPAP1-H dsRNA-injected pharate pupae were collected at pupal day 5. The insects were fixed in 4% p-formaldehyde and passed through a sucrose gradient series. Cryosectioning was performed as described previously [13] using a Leica CM 1800 instrument. Twenty mm thick cross-sections of pharate adults (pupal day-5) were stained for chitin with the rhodamineconjugated chitin-binding probe (New England Biolabs, Beverly, MA). All incubations for immunofluorescence were performed at room temperature. The sections were rinsed three times for 5 min in 0.2% PBST buffer, incubated for 1 hour in PBST buffer containing 2% BSA, and then rinsed. The sections were incubated overnight with rhodamine-conjugated chitin binding probe (1:100 dilution) in 2% BSA. After this treatment, the sections were rinsed again with PBST buffer as described above and nuclei were stained with 4', 6-diamidino-2-phenylindole-dihydrochloride (DAPI, 1 mg/ml) in PBS buffer. After further washing with PBST buffer, the sections were covered with Fluromount (Sigma) and sealed with clear nail polish to prevent dehydration. Slides were visualized using a Zeiss Leica 510 confocal microscope. The instrument with a fully motorized stage, a Plan Apochromat objective (40 X/1.4 oil) and differential contrast interference (DIC) was used for taking images. Fluorescence emission imaging of the 543 nm line of the He-Ne laser was used to excite the rhodamineconjugated chitin-binding probe and the 405 nm line for DAPI staining. Analysis of the images was done using the NIH Image J software.

Results
As a first step in the functional analysis of individual members of the CPAP gene families and optimization of RNA interference experiments, the expression profiles of the genes belonging to the two CPAP families of T. castaneum were determined by RT-PCR using cDNA prepared from RNA isolated at different stages of development. The results are shown in Fig. 1 and described below.

CPAP1 Genes Exhibit Divergence in Expression Profiles
The individual genes of the CPAP1 family have divergent expression patterns (Fig. 1A). The most notable finding is that they are expressed at very low levels, if at all, during the embryonic stage. TcCPAP1-A, TcCPAP1-I and TcCPAP1-J are expressed throughout the larval and pupal stages, with little or no expression in the adult. TcCPAP1-B and TcCPAP1-C have the broadest range of expression, spanning from young larval to mature adult stages. Others have more restricted ranges of expression. For example, TcCPAP1-D, TcCPAP1-F and TcCPAP1-G are expressed only in young larval and pupal stages. TcCPAP1-E and TcCPAP1-H are expressed predominantly in the pupal stage and do not appear to be expressed in earlier developmental stages or at the mature adult stage.
As most of the CPAP1 genes show expression in the pharate pupal and pupal stages, more detailed expression patterns of CPAP1 genes were also analyzed by RT-PCR from the pharate pupal stage through the young adult stage. TcCPAP1-A, TcCPAP1-B, TcCPAP1-F, TcCPAP1-G and TcCPAP1-H transcripts were observed throughout the pupal stage. TcCPAP1-C was expressed similarly, except for pupal day-2 when it had low or undetectable levels of expression. TcCPAP1-D had low expression on pupal days 1-3. TcCPAP1-E and TcCPAP1-J showed late pupal expression, but the transcript levels were relatively low during pupal days 0-3. TcCPAP1-I had very low expression throughout, except for detectable transcripts on pupal day-3 and pupal day-4 (Fig. 1B).

CPAP3 Genes have Similar Expression Profiles Throughout Development
In contrast to the widely diverging patterns of expression among CPAP1 genes, many of the CPAP3 genes were expressed throughout all stages of development, from embryonic to young adult stages, at nearly the same levels, followed by low or undetectable levels at the mature adult stage (Fig. 1C). Notable exceptions were TcCPAP3-A1 and TcCPAP3-A2 with low levels of expression in the mature larval stage, and also TcCPAP3-E whose expression occurred mainly from the mature larval stage and continued into the mature adult stage. Expression analysis of the CPAP3 genes from the pharate pupal to young adult stages revealed that all of these genes were expressed throughout this period except for TcCPAP3-C5b and TcCPAP3-E, both of which exhibited low levels of transcripts on pupal days 1-2 and pupal days 2-3, respectively (Fig. 1D).

Specificity of dsRNA-mediated Transcript Depletion within Each Family
Sequence conservation among members of the CPAP1 family proteins is limited to the peritrophin-A domain, with the rest of the proteins being variable in sequence and in length. Therefore, there were no difficulties in locating regions in the cDNA sequences for the synthesis of dsRNAs unique for each CPAP1 gene. RT-PCR following CPAP1 RNAi indicated that there was no evidence of knockdown of non-targeted transcripts (see Fig. 2A for some examples). CPAP3 family proteins are more highly conserved, not only in the three peritrophin-A coding domains, but also in the two linker regions between these domains. However, we were still able to design dsRNAs unique to each member of the CPAP3 gene family. The specificity of knock-down of the targeted CPAP3 transcripts by dsRNAs for individual members of this family was confirmed by RT-PCR (Fig. 2B). Again down-regulation was confined to the targeted gene.

Some CPAP1 Genes are Required for the Pupal-to-adult Molt
To determine the roles of individual members of the two families of CPAP genes in insect development, we carried out RNAi experiments using dsRNAs specific for each gene. Our RNAi strategy was designed to down-regulate transcript levels for specific members of each CPAP gene family just prior to the stage(s) of peak expression of the target gene and/or prior to each molt. dsRNAs for each of the ten CPAP1 genes were injected into insects (n = 40 for each group) at multiple stages of development, namely, penultimate instar larval, last instar larval, pharate pupal or pupal day 0 stages depending on the developmental expression profile of the target gene. The phenotypes of the insects subjected to CPAP1 RNAi as well as insects injected with the dsRNA for Vermilion (dsVer), which served as positive and negative controls, were monitored daily. Of the ten CPAP1 family genes, TcCPAP1-C, TcCPAP1-H and TcCPAP1-J exhibited lethal phenotypes at the pharate adult stage, regardless of the stage at which dsRNA was injected (larval, pharate pupal and or pupal stages). At the time of developmental arrest, incomplete cuticle slippage was evident, and elytral expansion was incomplete, as shown in Fig. 3. The dsRNA TcVer-injected control insects developed into normal adults except for the lack of eye pigmentation [14]. Several members of the CPAP3 family of genes were found to be essential for normal insect development and/or survival. In contrast to the results for CPAP1 family genes, RNAi for many of the CPAP3 family genes led to a range of molting defects, abnormal cuticles and lethality of insects at different stages of development. These abnormalities are described below for each member of the CPAP3 family of genes of T. castaneum.

RNAi Phenotypes of Members of the CPAP3 Gene Family
TcCPAP3-A1. dsRNA-injected insects (n = 40; two dsRNA's targeting different regions of the gene) exhibited a lethal phenotype only at the mature adult stage approximately one week after adult eclosion, regardless of the stage at which dsRNA was injected (larval, pharate pupal or pupal). These adults were unable to dislodge fecal pellets from their anuses and had severely depleted fat bodies (Fig. 4A, 4B). No other visible abnormalities were detected. Approximately 500 eggs (14.9 mg) were collected and used for RNA isolation. Pools of four larvae, (young and old separately) pharate pupae, pupae from day-0 to day-5, young adults (,1-2 hour old) or mature adults (.2 weeks-old) were used for preparation of total RNA at these stages. cDNAs synthesized from total RNA using the oligo-(dT) primer and reverse transcriptase were used as templates for RT-PCR (28 cycles).The initial denaturation was at 94uC for 10 minutes, annealing at (56uC-60uC) for 1 minute and the final extension was at 72uC for 10 minutes, using gene-specific primers. RT-PCR products for the T. castaneum ribosomal protein S6 gene (RpS6) from the same cDNA templates served as an internal control for loading (24 cycles). (A) Expression profiles of CPAP1 genes at different stages of development: E, eggs; YL, penultimate instar or younger larvae; ML, mature (last instar) larvae; PP, pharate pupae; P, pupae; YA, young adult; 3 hours after eclosion; MA, mature adult; 10 days after eclosion. (B) Expression profiles of CPAP1 genes at different stages of pupal development. P [0][1][2][3][4][5] , pupae 0-day old to pupae 5-days old were collected at 1-day intervals. (C) Expression profiles of CPAP3 genes at different stages of development: E, eggs; YL, penultimate instar or younger larvae; ML, mature (last instar) larvae; PP, pharate pupae; P, pupae; YA, young adult; 3 hours after eclosion; MA, mature adult; 10 days after eclosion. (D) Expression profiles of CPAP3 genes at different stages of pupal development. P [0][1][2][3][4][5] , pupae 0-day-old to pupae 5-days-old were collected at 1-day intervals. doi:10.1371/journal.pone.0049844.g001  TcCPAP3-A2. RNAi for TcCPAP3-A2 did not result in any observable abnormality at any stage of development. Thus this gene appears to be dispensable, in contrast to its paralog, TcCPAP3-A1. Furthermore, co-injection of dsRNAs for TcCPAP3-A1 and TcCPAP3-A2 led to the same phenotypes as those observed after dsTcCPAP3-A1 injection alone, ruling out any synergism between these two genes (Fig. 4A).
TcCPAP3-B. Injection of dsRNA (n = 40) at the larval stage had no effect on pupation, adult eclosion or on adult mortality for up to two months. However, all resulting adults were permanently afflicted with a stiff, uncoordinated gait. The forelegs appeared to be least affected, whereas the second pair of legs showed normal tibial-tarsal joint movement but stiffness at the femoral-tibial joint (Fig. 4C). The hindlegs were most affected, with impaired articulation at both the femoral-tibial and tibial-tarsal joints, resulting in a severe walking disability as indicated by a wobbly gait (see Video S1). Nearly half of the eclosed adults also exhibited split elytra, which were rough and bumpy. The pronota were also dimpled (Fig. 4A).
TcCPAP3-C. Insects injected with dsRNA for TcCPAP3-C at either the larval or prepupal stage suffered developmental arrest and death at the pupal-to-adult molt. None of the insects (n = 40) were able to shed their pupal cuticle. On the other hand, injections of dsRNA specific for TcCPAP3-C5a or TcCPAP3-C5b exons corresponding to the two alternatively spliced forms of the fifth exon of the TcCPAP3-C gene [6] resulted in developmental arrest and death of only 5% and 10% of the insects, respectively. The moribund insects had a phenotype similar to that observed after administration of exon-non-specific dsRNA for TcCPAP3-C. This result indicated that these two alternatively spliced transcripts at least partially compensated for each other (Fig. 4A). Coadministration of splice variant-specific dsRNAs for both alternative exon resulted in a phenotype identical to that seen with RNAi using a dsRNA for a common exon (data not shown).
TcCPAP3-D1 and TcCPAP3-D2. Injection of dsRNAs for either TcCPAP3-D1 (n = 40) or TcCPAP3-D2 (n = 40) had no effect on larval-larval, larval-pupal or pupal-adult molts. However, there were visible differences in the appearance of elytra compared to dsRNA TcVer-treated controls. Insects injected with dsRNA for Figure 4. Summary of phenotypes after RNAi of CPAP3 genes. (A) Insects injected with dsRNA at the penultimate and/or last instar larval stages for many of the CPAP3 genes showed abnormal phenotypes (except for TcCPAP3-A2 and TcCPAP3-E). RNAi for TcCPAP3-A1 resulted in adult lethality (indicated by a red slash across the figure). dsRNA TcCPAP3-A1+ dsRNA TcCPAP3-A2: the insects did not show additive mortality or additional abnormalities compared to animals injected with dsRNA TcCPAP3-A1 alone; dsRNA TcCPAP3-B: adults show walking defects. Black arrows indicate dimpled pronotum. dsRNA TcCPAP3C: Insects showed pupal to adult molt arrest and 100% lethality. dsRNA TcCPAP3-D1: Insects had rough elytra, which did not completely cover the adult abdomen. dsRNA TcCPAP3-D2: The phenotype was similar to animals treated with dsRNA TcCPAP3-D1, but was less severe. dsRNA TcVer-injected insects (200 ng per insect; n = 40) were used as controls (white eye is shown in inset). (B) dsRNA TcCPAP3-A1 treatment at the larval stages resulted in fat body depletion in one week-old adults compared to dsRNA TcVer-treated controls; shown by white arrows. Also the insects were unable to completely expel the fecal pellet; marked with blue arrow. (C) dsRNA TcCPAP3-B treated adults show a wobbly, stiff-jointed gait, resulting from abnormal articulation of the tibial-tarsal (shown in blue circle) and femoral-tibial joints for the metathoracic legs and of the femoral-tibial joint for the mesothoracic legs (shown as white arrows). (D) dsRNAs for TcCPAP3-D1 and TcCPAP3-D2 were co-injected into pharate pupae (n = 40). The resulting insects molted to otherwise normal adults with abnormal elytra. The elytra were wrinkled and alligator skinlike, and did not completely cover the abdomen, but none of the insects died. dsRNA TcVer was injected as a control and the elytra from these insects were normal (top panel). (E) Defects in a single elytron following RNAi for TcCPAP3-D1 or TcCPAP3-D2 in comparison to the TcVer dsRNA treatment. The differences are also shown in the form of the schematic line diagram on the right, where the lines are wavy in comparison to the control, TcVer. doi:10.1371/journal.pone.0049844.g004 TcCPAP3-D1 or/and dsRNA for TcCPAP3-D2 failed to fully expand their elytra and the elytral surface was uneven and wrinkled (Fig. 4A, 4D and 4E). Also the pronota of dsRNA TcCPAP3-D2-treated insects were dimpled. Co-injection of dsRNAs for the paralogous TcCPAP3-D1 and TcCPAP3-D2 genes resulted in adult elytra that were more severely affected than animals injected with dsRNA for either gene alone, indicating synergism between these two genes (Fig. 4D), but the adults were still viable. The elytra had alligator skin-like wrinkles and creases, and were shorter in length compared to the dsRNA TcVer-injected controls that developed normally. RT-PCR tests to check transcript down-regulation after RNAi showed significant depletion of the TcCPAP3-D1 and TcCPAP3-D2 transcripts at the pharate adult stage relative to controls that were injected with dsRNA for TcVer (data not shown).
Treatment with dsRNA for TcCPAP3-D1 or TcCPAP3-D2 Affects Morphology of Elytral Cuticle The structural defects of the elytra were more severe when the insects were co-injected with dsRNAs for both TcCPAP3-D1 and TcCPAP3-D2 compared to animals injected with dsRNA for either gene alone (see Fig. 4A, 4D, 4E). Therefore, we analyzed these insects to further probe the structural details of these defects. Elytra were isolated from 10-day-old adults following co-injection with dsRNAs for TcCPAP3-D1 and TcCPAP3-D2 at the pharate pupal stage. These elytra were analyzed by scanning electron microscopy (SEM) and compared with elytra from animals injected with dsRNA for TcVer. Scanning electron microscopic analysis showed that the elytra from insects treated with both dsRNAs were not properly expanded and had creases and folds on their dorsal surfaces. The ventral surfaces of elytra from the treated animals were also rough and displayed unorganized fibrous structures (Fig. 5). The chemical composition of these fibers is unknown, but it is presumed that they are chitinous.

CPAP Genes and Embryonic and Early Larval Development
To study the roles of CPAP genes in post-eclosion adult survival and reproduction, as well as on embryonic and early larval development, one-month-old adult females (n = 20 per group) were injected with 200 ng of dsRNA for each CPAP1 or CPAP3 gene. Adult mortality, fertility (as measured by number of eggs laid) and percent of eggs that hatched into larvae were monitored.
There was no observable effect on adult morphology or survival when dsRNA for TcVer was injected into adult females as a control. There was no adult mortality following adult injections of dsRNAs for any of the CPAP genes, with the sole exception of TcCPAP3-A1. Within two weeks of injection of dsRNA for TcCPAP3-A1, all adult females had died, exhibiting severely depleted fat bodies and abdominal tips with adhering fecal pellets (Fig. 4A, 4B).
The fertility of females after injection of dsRNAs for each of the CPAP family genes was followed by mating a pool of twenty dsRNA-treated females with an equal number of untreated adult males, followed by collection of batches of eggs laid during several consecutive 3-day intervals. There were no significant differences in the number of eggs laid following adult RNAi for any of the CPAP genes with the sole exception of TcCPAP3-D2, which led to a drastic reduction in the number of eggs laid (Fig. 6A, 6B). We subsequently observed that the ovaries of TcCPAP3-A1 and TcCPAP3-D2 dsRNA-treated females were abnormal in comparison to dsRNA TcVer controls (Fig. 6C). The ovarioles were underdeveloped and deficient in eggs. However, the fat bodies of TcCPAP3-D2 dsRNA-treated females were normal in appearance.
The few eggs that were laid failed to hatch, and the chorion layer was fragile, easily becoming detached upon mild treatment with a 1:1 diluted solution of 5% Clorox. The eggs contained yolk, but there was no evidence of the presence of an embryo (Fig. 6D).
Females injected with dsRNA for TcCPAP1-J had normal ovaries and oviposited normally, but the eggs failed to hatch. Upon dechorionation, fully developed larvae, normal in appearance and with red eye-spots, were visible under the microscope (Fig. 6D). However, germ band retraction appeared to be incomplete, because the dsRNA TcCPAP1-J-injected embryos placed their heads in an erect position, rather than tilted forward as observed in dsRNA TcVer-injected controls.

No Phenotypes Observed after Knockdown of Other CPAP Family Genes
Injection of dsRNAs corresponding to TcCPAP3-A2, TcCPAP3-E, TcCPAP1-A, TcCPAP1-B, TcCPAP1-D, TcCPAP1-E, TcCPAP1-F, TcCPAP1-G or TcCPAP1-I at different stages of development (penultimate instar larvae, pharate pupae and pupae day-0) did not result in any observable morphological abnormalities (data not shown). Injection of dsRNAs for any of the above genes into young adult females also had no observable effect on their survival, fecundity or egg hatch.

Chitin Analysis
Biochemical analysis. To determine whether the phenotypes observed after RNAi for any of the CPAP family genes were due to alterations in chitin content either by promoting or inhibiting chitin synthesis or by protecting chitin from degradation by chitinases, we compared chitin levels of whole animals after injection of dsRNA for the genes of the CPAP families that produced observable morphological defects. Chitin content of dsRNA-treated pupae collected 5 days after pupation was determined using a modified Morgan-Elson method [12,15,16]. When compared to the dsRNA TcVer-treated negative control, there was some evidence for a decrease in chitin content of whole animals following RNAi for TcCPAP1-H, but this reduction was not as drastic as in the dsRNA TcChsA-treated insects, which served as positive controls for chitin depletion (Fig. 7A, 7B). However, there was no significant difference in the chitin content of whole insects collected on pupal day 5 (the stage when molt arrest occurs) treated with dsRNA for the other CPAP1 genes, including TcCPAP1-C and TcCPAP1-J, which also yielded lethal phenotypes. Biochemical analysis of the chitin content of animals after RNAi for the CPAP3 genes also did not show any reduction in chitin content, unlike the case for animals treated with dsRNA for TcChs-A (Fig. 7A, B).
FITC-CBD staining. To further determine whether the reduction in chitin was confined to specific parts of the insect that are most actively synthesizing chitin, pharate adult elytra were excised on pupal day-5 and stained with a fluorescein isothiocyanate (FITC)-conjugated chitin-binding protein from Bacillus circulans WL-12. Before staining, the elytra were heated in 10 M NaOH at 95uC to dissolve most of the proteins, leaving a matrix consisting mostly of chitin [12]. Elytra of dsRNA TcVer-injected insects were used as a negative control and those from dsRNA TcChs-A-treated animals as a positive control for depletion of chitin. Elytra of dsTcCPAP1-H-treated animals were totally disintegrated after NaOH treatment, losing structural integrity, as were elytra from dsRNA TcChs-A-treated insects. Elytra of dsRNA TcCPAP1-C or dsRNA TcCPAP1-Jtreated insects did not show any visible reduction in chitin staining. However, elytral morphology was altered after such treatment (Fig. 7C).
Confocal microscopy. To localize the sites of chitin depletion, confocal microscopy was performed on cryosections of dsRNA TcVer and dsRNA TcCPAP1-H-treated insects at the pharate adult stage after staining with the rhodamine-conjugated chitin-binding probe (chitin stains red). In comparison to the dsRNA TcVer control, we observed in dsRNA TcCPAP1-H-treated insects a reduction in chitin staining in the new elytral cuticle. The old pupal cuticle, however, did not exhibit a significant difference in chitin staining (Fig. 7D).

Conclusions CPAP Proteins are Involved in Cuticle Organization
The main focus of this investigation has been the elucidation of the functions of members of the CPAP1 and CPAP3 families of chitin-binding proteins in T. castaneum. Even though the D. melanogaster ''obstructor'' family of genes (orthologs of the CPAP3 family) and their tissue-specificity of expression have been described previously [3], there have been no detailed studies on their biological functions until recently. This situation was partially remedied by Petkau et al., [7] who carried out a detailed study of one member of the Drosophila obstructor family of genes, namely obstructor-A (presumptive ortholog of TcCPAP3-A1; A detailed  following dsRNA injection for each of the 10 CPAP1 genes and then mated with 20 males. dsRNA TcVer was injected as a positive control. When compared to control insects, dsRNA CPAP1-injected animals did not exhibit significant reduction in the number of eggs. The X-axis shows the gene target for dsRNA-treatment. Y-axis: the number of eggs laid was plotted as the mean egg number per batch +/2SD during a 3-day period. (B) Comparison of the numbers of eggs laid by a pool of adult females (n = 20) following dsRNA injection for each of the seven CPAP3 genes. When compared to dsRNA TcVer-treated control insects, dsRNA TcCPAP3-D2-injected animals exhibited a significant reduction in the number of eggs. X-axis shows the targeted gene. Y-axis is the number of eggs laid plotted as the mean egg number per batch with SD. The asterisks indicate insufficient egg numbers due to adult lethality. (C) Dissected ovaries of dsRNA TcCPAP3-A1-injected females lacked proper differentiation into ovarioles and did not contain eggs. In animals treated with dsRNA for TcCPAP3-D2, the ovaries had ovarioles but did not contain properly developed eggs. After TcCPAP1-J dsRNA-treatment, the adult females had normal ovaries with eggs, similar to dsRNA TcVer-injected animals. (D) dsRNA for TcVer-injected 2-weeks-old adult female controls laid normal number of eggs and the embryos had normal body segmentation and visible appendages just before hatching. dsRNA for TcCPAP3-A1: In rare cases where eggs could be collected, the body segmentation was not visible presumably due to failure of embryonic development. dsRNA TcCPAP3-D2-treated embryos also lacked body segment differentiation and the chorion was fragile and was detached upon Clorox treatment. dsRNA for TcCPAP1-J treatment resulted in embryos with visible appendages and the red eye spot, but the head was not bent down like in the TcVer control. Instead, they held up their heads. All of these embryos showed embryonic arrest and never hatched. A red slash line indicates lethal phenotypes. doi:10.1371/journal.pone.0049844.g006 comparison of their results with those reported here is presented elsewhere in this section). Jasrapuria et al., [6] were the first to describe the CPAP1 family of genes in insects, but there have been no reports on their physiological functions. The results presented here are from a comprehensive analysis of the functions of all members of both families of CPAP proteins, which contain either one or three peritrophin-A domains and presumably interact with chitin in vivo.
Several lines of evidence suggest a role for the CPAP1 and CPAP3 families of proteins in the assembly of epidermal cuticle and maintenance of its structural integrity. First, all of the CPAP proteins are predicted to have a cleavable leader peptide and lack membrane-spanning helices, and thus are expected to be secreted proteins, consistent with a role in shaping the structure of the extracellular chitin-protein matrix. Second, CPAP1 and CPAP3 proteins have one and three peritrophin-A domains, respectively, which have been implicated in chitin binding. Third, CPAP family genes are expressed only in epidermal tissues that synthesize cuticular chitin, and notably not in gut tissue that produces chitin required for PM assembly [6]. Finally, the CPAP3 genes are highly expressed at developmental stages when the rate of cuticle deposition and/or tissue remodeling is elevated. This timing includes embryonic, larval and pupal stages as well as the young adult stage (Fig. 1). In addition, their expression levels generally decline in mature adult stages when new cuticle synthesis is essentially completed while PM synthesis continues unabated.

CPAP3 Genes have Diverse Roles in Molting, Locomotion and Egg Hatching
In this work we have used RNAi, which is systemic and splicevariant-specific in T. castaneum, to investigate the roles of individual CPAP1 and CPAP3 genes in this species at multiple developmental stages and in different cuticle-forming tissues. These RNAi experiments have revealed that many of the CPAP3 genes are indeed indispensable for insect survival or maintenance of normal morphology and cuticle integrity. Among the seven genes in the T. castaneum CPAP3 family, RNAi for five genes (including the two alternative exons of TcCPAP3-C), individually or in some combinations, yielded a large spectrum of phenotypes including molt arrest and lethality, elytral abnormalities, defects in articulation of joints, abnormal gait, reduction in fecundity, fat body depletion, underdeveloped ovaries, reduced egg hatching and adult mortality. Table 2 summarizes the results of RNAi for TcCPAP3 genes, including many that yielded observable phenotypes. Of the seven genes, RNAi for only two genes resulted in mortality either at the pharate adult stage (TcCPAP3-C) or at the young adult stage (TcCPAP3-A1). RNAi for three other CPAP3 family genes (TcCPAP3-B, TcCPAP3-D1 and TcCPAP3-D2) led to cuticle abnormalities, the most obvious ones being wrinkled elytra and a dimpled pronotum. The finding that orthologs of all of these CPAP3 genes are found in insect genomes of different orders [6,7] further strengthens our conclusion from RNAi studies that many members of the CPAP3 family of proteins serve distinctly different but essential functions. Expression data for the Obstructor (CPAP3) family of genes of D. melanogaster also indicate that many of these genes are expressed in larval tracheae and carcass (e.g. http:// flybase.org/reports/FBgn0031097.html;/FBgn0027600.html;/ FBgn0026077.html). The ortholog of TcCPAP3-A1 in D. melanogaster has been shown to be essential for larval growth, molting, tracheal tube expansion and wound healing [7]. The protein product of this gene is indeed detectable at high levels in larval and pupal stages, and at lower levels in adults.
The finding that down-regulation of the expression of CPAP3-C leads to molting arrest and death at the pharate adult but not during earlier stages of development of T. castaneum suggests that this protein may be indispensable only during pupal-adult metamorphosis. The absence of any detectable phenotype following RNAi for CPAP3-C at earlier developmental stages, when this gene is indeed expressed, suggests that other CPAP3 proteins may compensate for the absence of this protein in the formation of the larval and pupal cuticles. Parental RNAi experiments indicated that down-regulation of this gene also did not appear to affect embryonic development, egg hatch or larval viability, indicating that this gene is not essential during embryonic development. TcCPAP3-C gene is the ortholog of the ''gasp'' or ''obstructor C'' gene previously reported in D. melanogaster [3,5]. Obstructor-C, which is expressed in larval trachea, carcass and hindgut is among the proteins that are secreted into the tracheal lumen and reabsorbed by lumenal cells during tubule clearance and air-filling [17]. However, no mutant phenotypes have been reported for this gene in D. melanogaster or any other insect. The finding that among the seven members of the T. castaneum CPAP3 gene family, RNAi for this gene alone leads to developmental arrest and death prior to the adult stage indicates an essential role for this gene in development.
In contrast to the results of RNAi for TcCPAP3C, which led to molting arrest at the pharate adult stage, RNAi for TcCPAP3-A1 at larval or prepupal stages did not affect molting, larval growth or survival, cuticle morphology or adult emergence, but it did lead to mortality one week after adult eclosion. RNAi of adult females also resulted in death within two weeks after dsRNA administration. The observation that these animals had depleted fat body tissue and were unable to defecate normally suggested that a cuticle lining the digestive system such as that of the hindgut or foregut might have been affected in these animals. Knocking down transcripts of both TcCPAP3-A1 and TcCPAP3-A2 simultaneously did not further enhance mortality or result in molting defects. RNAi for TcCPAP3-A2 alone did not result in any abnormal Figure 7. Analysis of elytra for chitin levels following RNAi for CPAP genes. (A and B) Biochemical analysis of chitin content of 5 day-old pupae was carried out using a modified Morgan-Elson method (n = 5) [11]. Values of mean and standard errors are shown. dsRNA for TcVer was injected as a control that has a normal chitin content and dsRNA for TcChs-A served as a control that was depleted of chitin. When compared to control insects, the dsRNA TcCPAP1-H-treated pupae had a slightly reduced chitin content. GlcNAc = N-acetylglucosamine. (C) FITC-CBD staining of elytral cuticle after digestion with NaOH to remove cuticular proteins. Pharate pupae were treated with dsRNAs for TcCPAP1-C, TcCPAP1-H, TcCPAP1-J or TcChs-A. The elytra were dissected out from 5-day-old pupae. The elytra from insects injected with dsRNA for TcCPAP1-H or TcChs-A were fragile and became disintegrated by the alkali treatment (10 M NaOH at 95uC for 6 hours). Elytra from dsRNA TcCPAP1-C-and TcCPAP1-J-treated insects were also fragile and lost their original shape after NaOH treatment, but they remained intact and showed less staining after treatment with FITCconjugated chitin-binding protein (FITC-CBD) when compared with elytra from dsRNA TcVer-treated control insects. Elytra from dsRNA TcCPAP3-C treated animals did not show any effect, compared to Ver controls even though these insects were arrested at the adult molt. (D) Confocal microscopic analysis of pharate adult cuticle from dsRNA TcCPAP1-H and dsRNA TcVer-treated (control) insects that were stained with a rhodamineconjugated chitin-binding probe (red). An apparent decrease in chitin staining of elytral cuticle (green arrows) following RNAi for TcCPAP1-H compared to RNAi for TcVer (control) was observed. Scale bar = 10 mm. P, pupal cuticle; A, adult elytral cuticle. Plan Apochromat objective (40 X/1.4 oil). doi:10.1371/journal.pone.0049844.g007 phenotype, indicating that TcCPAP3-A2 may not be an essential gene. However, its higher expression in the elytron relative to the wing tissue during the pupal stages [6] suggests that it may have a role in shaping the elytral cuticle at the ultra-structural level. This remains a subject for future study. It is interesting to point out that null mutants of Drosophila obstructor-A, the ortholog of TcCPAP3-A, exhibited greatly retarded larval growth, multiple cuticular abnormalities, convoluted tracheal tubules and defective chitin removal from tracheal lumen. Such insects died in early larval stages in spite of consuming food [7]. In T. castaneum, we did not find any obvious cuticular abnormalities by carrying out RNAi of TcCPAP3-A1 at the fourth and fifth larval instars and prepupal stages. The larvae (or prepupae) developed into mature adults, but within a week after adult eclosion, they had depleted fat bodies, and died showing obvious defecation problems precluding studies of effects on oogenesis, oviposition, embryonic and/or early larval development. The reason for the difference in the effects of loss of function of this gene in these two insect species is unclear. One possible explanation that the presence of the paralogous gene, TcCPAP3-A2, in Tribolium and other insect species [6,7] (unlike Drosophila, which has only a single TcCPAP3-A-like gene), has redundant and compensating functions seems unlikely as downregulation of transcripts of both of these genes does not result in lethality at earlier developmental stages. We have not performed RNAi using combinations dsRNAs for multiple CPAP3 genes, which may have redundant functions. However, two transcripts of Obstructor-A that differ only in the length of the 39-UTR region have been described in D. melanogaster [7]. Whether or not these two alternatively polyadenylated transcripts of Obstructor A serve distinct functions in different tissues or developmental stages of D. melanogaster has not been determined. The protein products of this gene are present at high levels in late embryos, and pupae and persist in the adult stage in D. melanogaster suggesting that this gene is functional through all stages of insect life [7]. Similar experiments have not been conducted with T. castaneum due to the non-availability of specific antibodies to address the differences in the requirements for these gene products during development in Tribolium and Drosophila.
While TcCPAP3-D1 and TcCPAP3-D2 are dispensable for molting and adult survival of T. castaneum, insects subjected to RNAi for either of these genes have clearly visible deformities in the elytron or pronotum, suggesting that some aspect of cuticular structure has been altered. RNAi using a mixture of dsRNAs for TcCPAP3-D1 and TcCPAP3-D2 resulted in synergistic effects (more pronouncedly wrinkled elytra), indicating these may be recently duplicated genes with similar functions. Nonetheless, after RNAi for TcCPAP3-D1 and TcCPAP3-D2, we could observe defects in the elytral cuticle with the severity increasing when transcript levels of both of these genes were down-regulated, indicating that these two CPAP proteins may be important for maintaining the structural integrity of the elytra. SEM experiments provided strong evidence that the organization of the elytral cuticle is affected following RNAi for these two genes because we observed creases and unorganized fibers in the elytra compared to dsRNA Ver-treated control insects.
Similarly, down-regulation of the levels of TcCPAP3-B transcripts did not affect molting or even adult survival for at least a month, but these animals had split elytra and dimpled pronota. Most remarkably, they exhibited defective locomotor ability, possibly indicating a failure of normal articulation of the femoraltibial and tibial-tarsal leg joints. The wobbly gait observed after RNAi for TcCPAP3-B suggests that this protein may influence the flexibility of these joints. Thus, a subset of TcCPAP3 family genes is nonessential for survival but is critical for normal functioning of specific types of cuticles.
Only Three Members of the CPAP1 Family are Essential for Development and Cuticle Integrity The functions of the CPAP1 family of genes are not obvious because there is no information in the databases about the nature of the domains of these proteins except for the peritrophin-A domain. Three genes belonging to the CPAP1 family, TcCPAP1-C, TcCPAP1-H and TcCPAP1-J, appear to be essential, based on lethality at the pharate adult stage following RNAi. Confocal microscopy confirmed that RNAi for at least one of these, TcCPAP1-H, is associated with a detectable loss of chitin in the newly formed adult elytral procuticle but not in the pupal cuticle. The causes of death at the pharate adult stage following RNAi of TcCPAP1-C and TcCPAP1-J are unclear, but there is extensive tissue remodeling occurring at this stage and some of these processes could require specific CPAP1 genes. While the functions of the other CPAP1 family proteins could not be demonstrated by RNAi, they may have important functions because orthologs for these genes are present in several orders of insects including D. melanogaster (Table 1).

Matrices with Different Compositions of CPAP Proteins Display Divergent Physicochemical Properties
It is likely that the nature of the chitin-binding proteins as well as the cross-linking of these proteins to chitin and other proteins influence the properties of the cuticle and PM, which range from membranous (PM) to flexible (e.g. wing) to rigid (e.g. larval head capsule and elytron). In the cuticle, these variations are likely to result in part from different CPAP proteins being incorporated into different cuticles. It has been proposed that hard cuticle is more generally associated with proteins that contain the RR2 motif, whereas soft cuticle is associated with proteins that contain the RR1 motif [1,18,19]. Knock-down of transcripts for two cuticular proteins that are abundant in elytra of T. castaneum belonging to the RR2 family has been shown to result in less rigid elytra with altered mechanical properties and also in mortality of insects a week after adult eclosion [20].
The present study has established that specific members of either the CPAP1 or CPAP3 family influence the physicochemical properties of the cuticle. Elytra isolated from some of the RNAi experiments described in this work (e.g. TcCPAP1-H) are fragile and deficient in chitin as reported for elytra from animals depleted of chitin by treatment with dsRNA for TcChs-A [12] or dsRNA for Knickkopf, (Knk), a protein essential for chitin organization in the cuticle [21]. Elytra from animals with reduced expression of TcCPAP1-J or TcCPAP1-C were also fragile and had lost most of their structural integrity, whereas down-regulation of transcripts for other genes such as TcCPAP3-C did not result in such a phenotype, even though all three of these phenotypes had nearly the same amount of chitin as control animals. Thus, many and perhaps all CPAP proteins may play unique roles in constructing specific types of cuticle at specific anatomical locations such as joints and head skeleton, which in turn may influence the unique mechanical properties of each type of cuticle.

Some CPAP Proteins Affect Cuticle Integrity by Affecting Content and/or Organization of Cuticular Chitin
The most common defect seen in the RNAi experiments reported here was developmental arrest at the time of adult molting when there is extensive tissue remodeling and cuticle deposition. Rough and shortened elytra as well as a defective gait might be secondary consequences of the cuticular defects. We suspect that even though the chitin content of many cuticles was unaffected, one or more of other structural parameters including the antiparallel organization of the chitin microfibers, the organization of these fibers into macrofibers and laminae, and the stacking of these laminae into the helicoidal or pseudoorthogonal structures as described for other arthropods [22] may have been affected in our RNAi experiments. Loss of laminar architecture by depletion of the cuticle-organizing protein Knk, has been demonstrated in D. melanogaster and T. castaneum [21,23]. This protein also protects chitin from degradation by chitinases [21,24]. Whether CPAP proteins play similar roles in protecting or maintaining exoskeletal chitin in specific cuticular structures remains to be determined.
The requirement for expression of some CPAP genes for survival or fecundity during the adult stage when most of the body wall and appendage-associated cuticle is fully assembled indicates that there is turnover of some cuticular chitin, perhaps involving the digestive tract or the reproductive tissues. Two genes, TcCPAP1-J and TcCPAP3-D2, which are essential for survival at the pharate adult stage, affect fecundity or egg hatch. The defective egg maturation that was seen in ovarioles following RNAi for TcCPAP3-D2 may be due to some structural abnormality of the ovaries and/or eggshell. Indeed eggs, ovaries and egg shells of Aedes aegypti have been shown to contain chitin-like material [24]. Following parental RNAi for TcCPAP1-J, the females oviposit normally and the embryos are fully developed but fail to hatch. The larval head is tilted upward rather than tucked down ventrally as in normal or dsRNA Vertreated controls, possibly indicating incomplete retraction of the germ band. This phenotype is different from that reported following RNAi for TcChs-A, in which the embryos are misshapen and have a significant reduction in chitin content [11]. The TcCPAP3-D2 gene also appears to be involved in some early developmental events, because there is no evidence of embryonic development after parental RNAi for this gene.
In conclusion, differences in the developmental expression profiles of individual CPAP3 and CPAP1 genes and differences in expression in different cuticle-forming tissues (e.g. elytron versus hindwing), as well as the diverse phenotypes resulting from downregulation of individual genes of the CPAP family, suggest that many of these genes have distinct, non-redundant functions. This hypothesis is further supported by our finding that downregulation of individual transcripts or splice variants results in lethality and/or morphological defects, which, in most cases, cannot be compensated for by other members of the same family. Since all of the encoded proteins probably bind to chitin, they could interact with this matrix polymer and/or other cuticular proteins in distinctly different ways. One can speculate that specific CPAP proteins are involved at different levels of organization of chitin into nanofibrils, larger microfibrils of chitin-protein complexes, laminae and stacked laminae. The presence of three ChtBD2 domains in the CPAP3 family of proteins makes them particularly attractive candidates to be involved in non-covalent and/or covalent interactions with chitin and other proteins in nearby fibrils. The CPAP1 family proteins, which have other domains in addition to a single peritrophin-A domain, appear to be ideal candidates for either protein-protein interactions or for other uncharacterized enzymatic functions in a chitin-rich environment as has been shown for many chitinases and chitin deacetylases, some of which have one or more peritrophin-A domains [25]. In these enzymes, the chitin-binding domain probably helps to anchor the enzymes onto the insoluble chitin matrix, facilitating performance of their enzymatic functions by preventing premature dissociation from the chitin substrate. It is conceivable that uncharacterized proteases or other enzymes may utilize their built-in chitin-binding domains to immobilize themselves in the chitinous cuticle.

Supporting Information
Table S1 Summary of properties of dsRNAs used for RNAi studies. The table provides the nucleotide position, the length in base pairs of the amplicon and the number of dsRNAs targeted for each of the TcCPAP1 and TcCPAP3 genes. (XLSX) Video S1 dsRNA TcCPAP3-B-injected insects show uncoordinated gait, stiff-joints and split elytra. The pronota were also dimpled. (MOV)