An androgenic gland membrane-anchored gene associated with the crustacean insulin-like androgenic gland hormone

Crustaceans, and particularly the Malacostraca (i.e. amphipoda, isopoda and decapoda), exhibit a unique masculine sexual differentiation and maintenance system (Ventura et al., 2011). Members of this group (~25,000 species) possess a unique male-specific androgenic gland (AG) (Charniaux-Cotton, 1962) that is known to induce masculine development. Many of the androgenic effects of the AG have been attributed to a secreted proteinaceous AG-hormone (Chang and Sagi, 2008) found to be encoded by a male-specific insulin-like gene (IAG) (Ohira et al., 2003; Manor et al., 2007; Ventura et al., 2009; Mareddy et al., 2011). Functionality and involvement of several IAGs in sexual differentiation have been demonstrated in different species by both gain and loss of function studies using a recombinant protein (Okuno et al., 2002) and RNAi (Rosen et al., 2010; Ventura et al., 2012), respectively.

While IAG-encoding transcripts in different species have only been identified in recent years, the nature of the AG as an endocrine tissue specializing in major protein production was determined as early as the second half of the twentieth century. Using electron microscopy, King (King, 1964) showed that the AG cells exhibit patterns characteristic of the intracellular structures of protein-secreting cells in vertebrates. These include a well-developed rough endoplasmic reticulum (rER), in which proteins are co-translated and translocated into the lumen and which symbolizes their entry into the secretory pathway. The course of a given protein to the intercellular space through the ER and later the Golgi apparatus includes comprehensive maturation processes (e.g. addition of functional groups, proteolysis, folding, etc.). These tightly regulated maturation steps require the involvement of designated enzymes that may be either restricted to a certain organelle (ER or Golgi) to which they are bound (via a signal anchor/transmembrane domain) or completely soluble.

With respect to the secretory pathway – and specifically, intracellular modification and processing – the insulin superfamily of peptides has attracted intensive research. This is not surprising given the family includes numerous members found to be crucial components across diverse phenomena in both vertebrates and invertebrates. Factors like insulin, IGF-I, IGF-II, relaxin, bombyxin, insulin-like peptides and others were found to be involved in growth, development, metabolism, reproduction, longevity and apoptosis, among others (Ishizaki and Suzuki, 1994; Katic and Kahn, 2005; Bernstein, 2010; Chaves and Saif, 2011). Known to be synthesized as preprohormones that contain a signal peptide (pre), B chain, C-peptide (pro) and an A chain, these peptides lose their signal peptide (resulting in the prohormone form) during translocation into the ER lumen. In addition, some are glycosylated (N-linked) prior to their packaging in vesicles (as prohormones) where further maturation takes place, possibly involving proteolytic cleavage (yielding a mature hormone and C-peptide (Steiner and Oyer, 1967; Itoh et al., 1996) and the subsequent trimming of the C-terminus (Chu et al., 2011). In contrast to the vast knowledge and understanding about the packaging and secretion of members of the insulin-like superfamily and their related interacting proteins, there are virtually no data on any other protein specifically interacting and/or co-localizing with the crustacean androgenic insulin-like hormone in its secreting cell. Furthermore, other than several IAGs in decapods and isopods, no other AG-specific transcript has been reported thus far in this subphylum. Such putative factors may be dependent on, and/or act independently of, the IAG hormone.

The aim of this study was to identify additional AG-specific transcripts in the crayfish. We used a previously constructed AG cDNA library (Manor et al., 2007) as a database for mining putative candidates. In the screening process, another AG-specific transcript was identified that was found to encode a deduced protein with a putative signal anchoring sequence, thus termed Cherax quadricarinatus membrane-anchored AG-specific factor (Cq-MAG, GenBank accession no. JX446634), and it was further studied.

Mature C. quadricarinatus (Von Martens 1868) males (40–70 g) were reared in 5 m³ tanks at Ben-Gurion University of the Negev. The water temperature was held at 28±2°C and water quality was assured by circulating the system's total volume through a biofilter. A photoperiod of 14 h:10 h L:D was used. Food, comprising shrimp pellets (30% protein) (Rangen Inc., Buhl, ID, USA), was supplied ad libitum three times a week. An endocrine manipulation causing hypertrophy of the AG (hAG) was performed by eyestalk ablation as previously described (Khalaila et al., 2001; Khalaila et al., 2002).

Briefly, as described previously (Rosen et al., 2010), colony hybridization was applied to an existing AG cDNA SSH library (Manor et al., 2007) to remove Cq-IAG-carrying plasmids. In this assay, negative colonies were considered non-Cq-IAG, and were selected and sequenced. These sequences were further used to design specific primers.

Total RNA isolation began by anesthetizing the crayfish on ice for about 10 min. Following the dissection of various tissues, RNA from endocrinologically manipulated males was extracted using the EZ-RNA Total RNA Isolation Kit (Biological Industries, Beit Haemek, Israel). cDNA was generated by an RT reaction containing 1 μg of total RNA, which was extracted from mature males (AG, muscle, testis, peripheral glands, hepatopancreas and thoracic ganglia) and mature females (ovary and hepatopancreas) using the M-MLV reverse transcriptase, RNase H minus (Promega, Madison, WI, USA), according to the manufacturer's instructions. The cDNA was then amplified by PCR. The amplification of Cq-MAG using AG cDNA was performed with the forward primer Cq-MAG-f: 5′-AAGGCTGTGCAGGTAGCAGT-3′ and the reverse primer Cq-MAG-r: 5′-CACCAGGAACCATGTCAGTG-3′. PCR products were separated through 1.3% agarose gel. Fragments were excised, purified (Qiagen QIAquick Gel Extraction Kit, Germantown, MD, USA) and cloned (pGEM-Teasy vector, Promega). Selected clones were isolated, and plasmid DNA was purified (Qiagen Miniprep kit) and sequenced.

cDNA was prepared by an RT reaction (Reverse-iTTM-1st Strand Synthesis Kit, ABgene, Tamar Laboratory Supplies, Jerusalem, Israel) using 1 μg of RNA that was extracted as above from mature endocrinologically induced males (hAG, N=5) and intact males (AG, N=4). Relative quantification (RQ) of Cq-MAG was performed using the forward primer qCq-MAG-f: 5′-CCAGACAGAAGACAGCAACAA-3′, the reverse primer qCq-MAG-r: 5′-AACGACAGGACGGAACGAT-3′, Universal ProbeLibrary probe no.84 (Roche Applied Science, Indianapolis, IN, USA) and the FastStart Universal Probe Master (ROX, Roche Applied Science). RQ values were normalized to Cq-18S, which was amplified using the forward primer q18s-f: 5′-CTGAGAAACGGCTACCACATC-3′ and the reverse primer q18s-r: 5′-GCCGGGAGTGGGTAATTT-3′ and then quantified with Universal ProbeLibrary probe no.74. RQ values were expressed as means ± s.e.m., and a t-test was used to determine whether there was a statistically significant difference.

Total RNA was isolated as above from the AG, muscle, hepatopancreas and testis of adult males. RNA (5 μg) from each sample was separated by electrophoresis through a 1% agarose formaldehyde gel, transferred to a nitrocellulose membrane and UV cross-linked. The blot was prehybridized overnight as described elsewhere (Shechter et al., 2005) and radiolabeled with an α-³²P probe prepared by adding dCTP and a cDNA template isolated from a clone containing a cDNA insert (Cq-MAG) to a random priming labeling mix (Biological Industries). The membrane was washed as previously described (Shechter et al., 2005) and exposed to BioMax MS Kodak film with intensifying screens at −70°C for 25 min. rRNA was visualized with ethidium bromide.

The sequences of both the 5′ and 3′ ends of Cq-MAG were obtained by 5′ and 3′ rapid amplification of cDNA ends (RACE). Both procedures were carried out with the Clontech SMART RACE kit (Clontech, Mountain View, CA, USA) following the manufacturer's protocol. PCR of the 5′ region was performed using the gene-specific reverse primer Cq-MAG RACE-r: 5′-AGCTGCTGCTGCCTAGCTGCTGCT-3′ and the Universal Primers Mix (UPM) provided in the kit. PCR of the 3′ region was performed with the UPM as a reverse primer and the gene-specific forward primer Cq-MAG RACE-f: 5′-TGAAGGGTAGCTGTCACGAGCCT-3′. The PCR products were cloned and sequenced as described above.

To confirm the identity of the PCR product, a forward primer from the 5′ end was designed based on the 5′ RACE results: Cq-MAG val.f: 5′-GTTATGGTGTGTCGCAGACAGCAA-3′. The forward primer was used for PCR with the reverse primer Cq-MAG val.r: 5′-TTGCTTATCCCAGGGGACAGAACT-3′, and a product measuring about 580 bp was amplified. This product was cloned and sequenced as described above, and the quality of the sequence was ensured by sequencing this region of the transcript five times.

Hypertrophied AGs attached to ~0.5 cm of the proximal vas deferens were fixed in modified Carnoy's II fixative and processed as described elsewhere (Ventura et al., 2009). Digoxygenin (DIG)-labeled oligonucleotides for antisense and sense probes, corresponding to nucleotides 140–863 of Cq-MAG cDNA, were synthesized using SP6 and T7 RNA polymerases. The probes were hydrolyzed to reduce their length to ~200 bases, as described in the DIG Application Manual (Roche Applied Science). Hybridization was carried out as described previously (Shechter et al., 2005), with the slight modification of adding 100 μg ml⁻¹ tRNA to the hybridization solution. DIG was visualized with colorimetric substrates NBT/BCIP (Roche Diagnostics, Indianapolis, IN, USA) according to the manufacturer's instructions. Slides were mounted in 15% glycerol and observed under a light microscope.

All non-Cq-IAG ESTs were initially processed as previously described (Manor et al., 2007). The full length cDNA sequence of Cq-MAG was computationally translated using the ExPASy Proteomics Server (http://web.expasy.org/translate/) and the most reasonable frame was selected (5′3′ Frame 1). The following structural analysis predictions for the Cq-MAG deduced amino acid sequence were applied: domains by SMART (http://smart.embl-heidelberg.de), signal peptide/anchor by signalP 3.0 (http://www.cbs.dtu.dk/services/SignalP-3.0/), and folding/unfolding probabilities by FoldIndex (http://bip.weizmann.ac.il/fldbin/findex).

All constructs were subcloned into the pUASp vector system, which is suitable for protein expression in Drosophila embryonic S₂R+ cells. The Cq-IAG prohormone was created by PCR amplification of AG cDNA with the forward primer IAG_Exp_For: 5′-GCGGCCGCATGTACAGAGTGGAAAACCTTCTG-3′ (including the Met start codon followed by Tyr at position 24 onwards) and the reverse primer IAG_Exp_Rev: 5′-GCGGCCGCTCAAGAACTGACGTAGATTCC-3′ (including a stop codon). Both primers contained the NotI restriction enzyme recognition sequence (underlined). Upon amplification, the PCR product was purified as described above, digested with NotI and subcloned into the pUASp plasmid (digested correspondingly) by ligation. In addition, two coding sequences of Cq-MAG were subcloned; the full length open reading frame (ORF) termed full Cq-MAG and a truncated form in which the first 42 amino acids of the N-terminus were omitted (Δ42Cq-MAG). Both constructs were initially PCR amplified as described above, using different primers – full MAG_For: 5′-TCTAGAATGGTGTGTCGCAGACAGCAACAGC-3′ (full construct) and Δ42 MAG_For: 5′-TCTAGACGTACCCTGGGTACGGAGAC-3′ (truncated construct). For both constructs, the reverse primer used was MAG_Rev: 5′-TCTAGATTATCCCAGGGGACAGAACTTC-3′. All primers started with the six nucleotide XbaI recognition site (underlined) for PCR product digestion and subsequent ligation at the C-terminus of GFP, which had previously been subcloned into the pUASp vector. All plasmids were validated and sequenced from both ends before the transfections were performed.

The expression of all pUASp cloned vectors (GFP, proCq-IAG, GFP::Cq-MAG and GFP::Δ42Cq-MAG) was carried out in S₂R+ cells, which were grown, maintained and transfected for 48 h (Dubin-Bar et al., 2008). Approximately 4×10⁶ cells were transfected with 1 μg pUASp-based expression vector and the Act5C-Gal4 driver using Escort IV (Sigma-Aldrich Israel Ltd, Rehovot, Israel). Expression included: GFP, GFP::Cq-MAG, GFP::Δ42Cq-MAG and proCq-IAG. Two co-expression experiments were conducted: GFP + proCq-IAG and GFP::Cq-MAG + proCq-IAG. After the transfections were completed, cells were washed, fixed and mounted as previously described (Dubin-Bar et al., 2008). For proCq-IAG detection, its specific antibody was used (1:10,000) followed by a goat anti-rabbit secondary antibody (1:500) (Cy3-conjugate AffinityPure Goat anti-rabbit IgG, Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA). Nuclei were stained with DAPI (1:1000), and cells were imaged with confocal microscopy for protein localization.

A specifically expressed AG transcript was identified and later termed the C. quadricarinatus membrane-anchored AG peptide (Cq-MAG). Its tissue-specific expression was demonstrated by RT-PCR (Fig. 1A), which showed that this transcript's cDNA is amplified only in the AG and not in any other representative tissues, i.e. muscle, testis, peripheral glands, hepatopancreas or thoracic ganglia from mature male crayfish and ovary or hepatopancreas from mature female crayfish. The positive control was based on the amplification of the crayfish housekeeping gene EFT-II (accession no. AI253924). Northern blot analysis revealed a Cq-MAG transcript of about 1000 bases long observed only in the lane containing RNA from the AG and not in the lanes containing samples from muscle, hepatopancreas and testis (Fig. 1B), despite the presence of rRNA in all the samples. RNA in situ hybridization confirmed that Cq-MAG-specific expression is restricted to the AG and not to neighboring tissues, such as the sperm duct or connective tissue (Fig. 1C). A specific signal in the cytoplasm of AG cells was clearly observed only when using the antisense probe, while no signal was detected when using the sense probe (Fig. 1C, bottom).

Cq-MAG was fully sequenced. Its cDNA was 920 nucleotides long, with 5′ and 3′ untranslated regions (UTRs) of 216 bases and 134 bases, respectively, and the 3′ UTR ended at the polyA tail (Fig. 2A). The ORF of 567 bases encoded a deduced 189 amino acid product. Bioinformatic analysis predicted a putative 42 amino acid signal anchor peptide at the N-terminus (SignalP, signal anchor probability: 0.912) and also showed that most of the newly identified putative protein was predicted to be intrinsically unfolded (fold index, 23/189 amino acids). However, positive folding probabilities were suggested for a few regions, the longest of which was from amino acids 19 to 43, where the majority of the residues constituting the predicted signal anchoring sequence are located (Fig. 2B).

The putative signal anchoring property of the N-terminus of Cq-MAG was examined based on cellular localization of recombinant fusion proteins in Drosophila Schneider cells (Fig. 3). The expression of GFP alone (Fig. 3A) demonstrated fluorescence spread evenly throughout the cell, including expression in both the nucleus and the cytoplasm, as observed by confocal microscopy. When expressing GFP fused to the wild-type Cq-MAG, including the first 42 amino acids containing the predicted signal anchor (Fig. 3B), the protein was mainly observed surrounding the periphery of the nucleus and in other parts of the cell, demonstrating an ER-like localization pattern. In contrast, when expressing GFP fused to Cq-MAG from which the first 42 amino acids at the N-terminus were deleted (Fig. 3C), the protein's localization was substantially altered, revealing a pattern similar to that of the GFP-only construct.

Our finding that Cq-MAG, similar to Cq-IAG, is an AG-specific transcript led us to examine the possible interaction between these two factors in Drosophila Schneider cells (Fig. 4). The sole expression of Cq-IAG showed this protein is localized throughout most of the cell but not in the nucleus (Fig. 4A,C). The protein was visualized as many spots, which suggests packed granules. In co-expression of Cq-IAG and GFP, the former showed the same localization pattern (Fig. 4D) as when expressed by itself, and the latter demonstrated the characteristic GFP fluorescence spread evenly throughout the cell (Fig. 4E). In contrast, co-expression of Cq-IAG and a GFP::Cq-MAG fusion protein substantially altered the localization pattern of Cq-IAG. Although it was still present in many parts of the cell, it formed an intense line around the periphery of the nucleus (Fig. 4G). The visualization of green fluorescence representing Cq-MAG revealed that this fusion protein has an ER-like localization pattern. Thus, it was highly evident both around the periphery of the nucleus and in a net-like spread in other parts of the cell (Fig. 4H), as shown earlier (Fig. 3C). An overlay image of the localization patterns of these two AG-specific proteins with the addition of DAPI stain (Fig. 4I) showed these two patterns overlap mainly around the exterior surface of the nucleus and slightly in distal regions of the cell.

The association between the two AG-specific transcripts was further examined with respect to their response to an endocrine manipulation. The X-organ sinus gland (XO–SG) negative effect over the AG was examined at the transcript level (Fig. 5). Real-time RT-PCR using AG cDNA samples of intact and induced crayfish showed that the expression of both Cq-IAG and Cq-MAG was stimulated (Fig. 5). Cq-IAG transcript levels were significantly elevated [RQ value of (59.53±3.05)×10³, P<0.0001, t-test] compared with the intact state [(1.38±0.59)×10³]. Cq-MAG levels showed a similar trend – a significant overexpression of this transcript in induced specimens was found [(4.99±0.53)×10³, P<0.005, t-test) compared with the control [(0.66±0.17)×10³].

Studies focusing on the AG and on the isolation of the AG hormone and its encoding transcript are accumulating, with ~20 AG-specific transcripts fully sequenced in Crustacea, over 10 of which were identified in decapods (Ventura et al., 2011; Li et al., 2012). Without exception, all of these transcripts belong to the insulin-like superfamily of peptides, which in decapods are termed insulin-like AG-specific factors (IAGs). In fact, until the current study there had been no reports in crustaceans of a different (non-IAG) AG-specific transcript. Thus, the identification of Cq-MAG, regardless of the nature and functionality of its encoded product, is exciting considering its validated tissue-specific expression in light of the AG's established, pivotal role in masculine sexual differentiation.

As the newly identified Cq-MAG protein seems to be mainly disordered and devoid of any conserved domains or motifs, it could not be assigned to a certain pathway or family of proteins, an outcome that eliminates a potential way to shed light on its putative functionality. The only positive prediction related to a structural region was of a signal anchor corresponding to the first 42 amino acids at the N-terminus, which integrates the protein into an intracellular membrane (as was established in the Schneider system). This property is usually characteristic of proteins that are associated with the ER or Golgi membrane (High et al., 1991; Schneider et al., 2000; Kida et al., 2009), which may enable them to participate in the modification and sorting of secreted proteins transported via this route. Thus, the mechanistic nature behind the co-localization of Cq-MAG and the IAG hormone may be rationalized based on known and characterized instances of processing of other insulin-like superfamily members. Several convertases have been shown to co-localize with their targeted insulin-like factors both at the transcript level (Renegar et al., 2000) and at the protein level in secreted granules (Itoh et al., 1996; Arvan and Halban, 2004; Chu et al., 2011). However, converting enzymes were not the only factors suggested to co-localize with, and participate in, the sorting and maturation of insulin-like bearing granules. Such is the unique case of ZnT8, an isoform of the zinc transporter (ZnT) family of peptides. This membrane-anchored protein specifically localizes to insulin granules and its encoding transcript was found to be exclusively expressed in β-cells of the Langerhans islets (Chimienti et al., 2004). Although Cq-MAG does not seem to include any conserved domain of a known enzymatic nature (using standard bioinformatic tools), it was shown to be AG-specifically expressed, membrane anchored and co-localized with Cq-IAG. These features of Cq-MAG are significant in light of the background of insulin-like factors, which may suggest this protein participates in the processing/maturation, sorting and secretion of the Cq-IAG hormone from its cell of origin. Moreover, an insulin secretion-related protein (SMP30/gluconolactonase) was found to bear some sequence homology and similarity to Cq-MAG. This protein was found to be crucial for proper insulin secretion (Hasegawa et al., 2010); however, unlike Cq-MAG it was not found to be an anchored protein.

Many studies have already addressed the issues of AG function (Khalaila et al., 1999; Khalaila et al., 2001; Barki et al., 2003; Manor et al., 2004; Barki et al., 2006) and its IAG hormone in C. quadricarinatus (Rosen et al., 2010). Upstream endocrine regulation of the AG has thus far been described in terms of AG inhibition by the neuroendocrine center located in the eyestalk within the XO–SG (Khalaila et al., 2002). In this respect, extensive AG hypertrophy and hyperplasia upon endocrine manipulation (elimination of this negatively regulating tissue) has previously been documented (Khalaila et al., 2002; Ventura et al., 2009). This general hypertrophy was further described in other species, as characterized by changes in the distribution of certain AG cell types (Phoungpetchara et al., 2011), increased numbers of IAG hormone-producing cells (Sroyraya et al., 2010), and IAG transcript overexpression (Chung et al., 2011). The induced overexpression of Cq-MAG and Cq-IAG under XO–SG deprivation provides important evidence for the specific negative control of neuroendocrine factors over the expression profile of AG activity-related factors in this crayfish. Moreover, it may present an additional level of association of these two factors. Still, the identification of AG intracellular components that affect the expression of its specific genes and the isolation of XO–SG factors bearing AG-inhibiting activity is essential to gain an initial molecular understanding of the complexity of endocrine control and sexual differentiation processes in crustaceans. Although we have been successful in implementing in vivo RNAi to knock down several transcripts involved in reproduction and molting of decapod crustaceans (Shechter et al., 2008; Ventura et al., 2009; Glazer et al., 2010; Rosen et al., 2010; Pamuru et al., 2012), Cq-MAG's transcript levels could not be reduced using standard dsRNA injections. Thus, the functionality of Cq-MAG and its specific involvement within the AG/IAG route remain to be directly elucidated.

We thank Mr Amit Savaia for his technical assistance with crayfish maintenance and the Fish and Aquaculture Research Station, Dor, Israel, for their help in maintaining and supplying the animals for this study. We would also like to thank Mr Patrick Martin for styling the manuscript and Rotem Kadir for his assistance with confocal microscopy.