Expression Profiling, Downstream Signalling and Inter-Subunit Interactions of GPA2/GPB5 in the Adult Mosquito Aedes aegypti

GPA2/GPB5 and its receptor constitute a glycoprotein hormone-signalling system native to the genomes of most vertebrate and invertebrate organisms, including humans and mosquitoes. Unlike the well-studied gonadotropins and thyrotropin, the exact function of GPA2/GPB5 is unclear, and whether it elicits its functions as heterodimers, homodimers or as independent monomers remains unknown. Here, the glycoprotein hormone signalling system was investigated in adult mosquitoes, where GPA2 and GPB5 subunit transcripts co-localized to bilateral pairs of neuroendocrine cells within the first five abdominal ganglia of the central nervous system. Unlike human GPA2/GPB5 that demonstrated strong heterodimerization between subunits, the GPA2/GPB5 subunits in A. aegypti lacked evidence of heterodimerization when heterologously expressed. Interestingly, cross-linking analysis to determine subunit interactions revealed A. aegypti and H. sapiens GPA2 and GPB5 subunits form homodimers, and treatments with independent subunits did not activate A. aegypti LGR1 or H. sapiens TSH receptor, respectively. Since mosquito GPA2/GPB5 heterodimers were not evident by heterologous expression of independent subunits, a tethered construct was generated for expression of the subunits as a single polypeptide chain to improve heterodimer formation. Our findings revealed A. aegypti LGR1 elicited constitutive activity that elevated levels of cAMP as determined by increased cAMP-dependent luminescence. However, upon treatment with recombinant tethered GPA2/GPB5 heterodimers, an inhibitory G protein (Gi) signalling cascade is initiated and forskolin-induced cAMP production is inhibited. These results provide evidence towards the functional deorphanization of LGR1 and, moreover, further support the notion that GPA2/GPB5 heterodimerization is a requirement for glycoprotein hormone receptor activation. Significance Statement The GPA2/GPB5 glycoprotein hormone signalling system is an evolutionarily ancient one that is involved in multiple physiological functions in diverse animal species, ranging from insects to worms to humans; however, its mode of signalling remains poorly characterized. To better understand how GPA2/GPB5 may exert its function, hormone subunit expression was localized in adult mosquitoes, and properties of subunit interaction along with receptor coupling for human and mosquito GPA2/GPB5 signalling was investigated. Our findings indicate certain aspects of GPA2/GPB5 signalling are unique from that of the classic glycoprotein hormones (i.e. the gonadotropins and thyrotropin) and could implicate independent physiological roles for heterodimers and homodimers in mosquitoes and humans.


3 8
However, similar inhibitory effects of GPA2 and GPB5 proteins were also observed with cells 2 3 9 not expressing LGR1 (Fig. 5D). were coapplied for receptor binding (Fig. 5A) and, unlike the heterodimerization of human 2 4 4 GPA2/GPB5 observed in our experiments, mosquito GPA2/GPB5 lacked evidence of 2 4 5 heterodimerization (Fig. 4). In light of these observations, it was hypothesized that the activation 2 4 6 of A. aegypti LGR1 also required subunit heterodimerization. To produce stable GPA2/GPB5 2 4 7 heterodimers using the heterologous expression system, both GPA2 and GPB5 mosquito 2 4 8 subunits were expressed as a tethered, single-chain polypeptide by fusing the C-terminus of the 2 4 9 GPB5 prepropeptide sequence with the N-terminus of the GPA2 propeptide sequence using a 2 5 0 tagged linker sequence composed of twelve amino acids, involving three glycine-serine repeats 2 5 1 and six histidine residues. fractions of mCherry transfected cells (Fig. 6A). However, a strong band was detected at 32-40 2 5 8 kDa in the lysates of cells transfected to express tethered GPA2/GPB5 (Fig. 6A). Moreover, two PNGase, the higher 40 kDa molecular weight band disappears and the 37 kDa band size 2 6 5 intensifies, which confirms the observed molecular weight shift results from removal of N-linked  The effects of tethered GPA2/GPB5 heterodimers on LGR1 activity was examined. Cell 2 7 0 lysates or secreted protein fractions collected from mCherry-or tethered GPA2/GPB5-2 7 1 transfected cells were incubated with HEK 293T cells co-expressing the cAMP luciferase 2 7 2 biosensor and either A. aegypti LGR1 or mCherry (i.e. not expressing LGR1). Whether tethered 2 7 3 GPA2/GPB5 proteins could elevate cAMP or inhibit a forskolin-induced rise in cAMP was 2 7 4 assessed and compared to control treatments with proteins harvested from mCherry-transfected 2 7 5 cells. Overall, the relative luminescent response was significantly greater in LGR1-transfected 2 7 6 cells compared to mCherry-transfected cells (Fig. 7). Unlike treatments with forskolin that 2 7 7 significantly increased cAMP-mediated luminescence, neither secreted protein fractions nor cell 2 7 8 lysates of tethered GPA2/GPB5-transfected cells elicited an increase in the cAMP-mediated 2 7 9 luminescent response relative to controls ( Fig. 7A-7B). Secreted protein fractions containing 2 8 0 tethered GPA2/GPB5 protein had no effect on the forskolin-induced cAMP-mediated 2 8 1 luminescence, compared to control treatments with mCherry secreted proteins in LGR1-2 8 2 expressing cells (Fig. 7C). Notably, however, treatments of LGR1-transfected cells, but not 2 8 3 15 mCherry-transfected cells or assay media, with cell lysates containing tethered GPA2/GPB5 2 8 4 protein significantly inhibited the forskolin-induced rise in cAMP relative to treatments with 2 8 5 mCherry-transfected cell lysates (Fig. 7D). The central nervous system (CNS) of adult mosquitoes is comprised of a brain and a 2 9 1 ventral nerve cord, consisting of three fused thoracic ganglia and six abdominal ganglia. Our 2 9 2 findings demonstrate that the GPA2 and GPB5 transcripts are significantly enriched in the 2 9 3 abdominal ganglia of adult mosquitoes relative to peripheral tissues and other regions of the 2 9 4 CNS, with no sex-specific differences. Although low levels of GPA2 and GPB5 transcripts were 2 9 5 detected in the thoracic ganglia and brain using RT-qPCR, fluorescence in situ hybridization 2 9 6 (FISH) techniques used to localize GPA2 and GPB5 transcripts, along with 2 9 7 immunohistochemical detection of GPB5, did not identify specific cells in these regions of the 2 9 8 nervous system. Instead, GPA2 and GPB5 transcripts, as well as GPB5 immunoreactivity was  where the lateral nerves emanate. In some abdominal ganlia preparations, a third bilateral pair of 3 0 8 cells immunoreactive for GPB5 protein was detected; however these additional cells were not 3 0 9 detected using FISH, which suggests GPB5 transcript may be differentially regulated between detected and these were more intensly stained compared to preparations treated with either probe 3 1 6 alone, which confirms that GPA2 and GPB5 are indeed co-expressed within the same that the mosquito GPA2/GPB5 subunits may be produced and released as heterodimers in vivo. To study the interactions of A. aegypti GPA2 and GPB5 subunits in vitro, hexa-histidine analysed under denaturing conditions after cross-linker treatments, which had been utilized protein samples were then deglycosylated to identify whether the removal of N-linked sugars GPA2/GPB5 protein were not able to confirm heterodimeriation since the detected band sizes 3 3 5 could also reflect GPA2 and GPB5 homodimeric interactions. Previous cross-linking studies that 3 3 6 demonstrated GPA2/GPB5 heterodimerization in human (3, 6) and fruit fly (29) did not provide 3 3 7 evidence on the interactions of each subunit alone to determine if homodimerization is possible.

8
In these earlier studies, molecular weight band sizes that were identified as heterodimers could have been the result of homodimeric interactions (3, 6, 29). As a result, the findings herein with  To clarify whether A. aegypti (mosquito) GPA2/GPB5 heterodimerize, each subunit was combinations of cross-linked subunits were probed with either anti-FLAG or anti-His antibody. to study unstable or weak protein-protein interactions (34). The incorporation of a linker 4 0 7 sequence between glycoprotein hormone subunits has been performed previously, and does not  with PNGase verified the tethered GPA2/GPB5 proteins are glycosylated, as observed for GPA2 expressing cells were individually tested for their ability to activate A. aegypti LGR1.

2 7
In humans, GPA2/GPB5-TSHR signalling stimulates adenylyl cyclase activity to increase 4 2 8 intracellular cAMP via interaction with a Gs protein (3, 6, 11), and these results were confirmed 4 2 9 in our studies. Comparatively, GPA2/GPB5 signalling was also shown to increase levels of been demonstrated to be stronger for the thyrotropin receptor than for the LH/CG receptor (40). Gi coupling for A. aegypti LGR1, given that GPA2/GPB5 heterodimers signficantly inhibited the  form an additional disulfide bridge which wraps around and "buckles" the alpha subunit (26).

7
Though heterodimerization can occur with mutated forms of this "seatbelt" structure, there is a projections that localized distinctly from organs that express the GPA2/GPB5 receptor (LGR1), 4 6 3 did support that this system is indeed endocrine in nature (16). Alternatively, the subunits could 4 6 4 function independently or regulate physiology as a heterodimer in a paracrine/ autocrine fashion. In rats, GPA2/GPB5 is expressed in oocytes and may act as a paracrine regulator of TSHR- to consider is that additional endogenous co-factors may be involved, but remain unidentified, 4 6 8 which help to strengthen interaction between the GPA2 and GPB5 subunits, since the tethered Our results establish that human and mosquito GPA2 and GPB5 subunits can weakly and 4 7 4 strongly, respectively, homodimerize. However, whether these homodimers, pertain to a 4 7 5 physiological function in vivo is unknown. Treatments of either mosquito or human GPA2 and 4 7 6 GPB5 subunits alone did not stimulate specific downstream signalling in LGR1 or TSHR-  In addition to the human/ mosquito GPA2/GPB5 homodimers observed in our studies, 4 8 7 human GPA2 was also shown to interact with the beta subunits of CG and FSH (3). Lastly, the localize, since GPA2 expression exhibits a much wider distribution and is expressed more 4 9 0 abundantly than GPB5 in a number of vertebrate and invertebrate organisms (7, 12, 23, 24, 49, 4 9 1 50). Taken together, this raises the possibility that GPA2 and GPB5 subunits may interact with  Although much is known about the classic vertebrate glycoprotein hormones including 4 9 7 LH, FSH, TSH and CG along with their associated receptors, little progress has been made thus interactions, particularly for the invertebrate organisms. To our knowledge, this is the first study 5 0 0 to demonstrate A. aegypti and H. sapiens GPA2 and GPB5 subunit homodimerization in vitro.

0 1
Our results also confirm that heterodimerization of A. aegypti and H. sapiens GPA2/GPB5 are 5 0 2 required for the activation of their cognate receptors LGR1 and TSHR, respectively. Unlike LGR1 couples to a Gi protein to inhibit cAMP levels following application of heterodimeric of fruit fly LGR1 (29, 51) as well as mammals including dog and human TSH receptor (52, 53).

1 0
In the mosquito nervous system, GPA2 and GPB5 subunits are co-expressed within the however, these results confirm GPA2 and GPB5 homodimers do not activate LGR1 and TSHR. is a research direction that should be addressed in future studies. All in all, this investigation has including a consensus Kozak translation initiation sequence, was amplified and a hexa-histidine 5 8 1 or FLAG tag sequence was incorporated on the carboxyl-terminus of subunits to produce the 5 8 2 following fusion proteins; A. aegypti GPA2-FLAG, H. sapiens GPB5-His (Table S2) Gradia (Addgene plasmid # 30125), was utilized to verify cell transfection efficiency.

8 5
Experiments also utilized previously prepared pcDNA3.1 + constructs with A. aegypti GPB5-His  Corp., Madison, WI), which were used for receptor activation and intracellular signalling assays. The ORFs of A. aegypti GPA2 and GPB5 sequences were tethered together in order to 5 9 6 promote heterodimer interactions for testing in receptor activity assays with mammalian cell 5 9 7 lines. A hexa-histidine tagged artificial linker sequence involving three glycine-serine repeats 5 9 8 was used to fuse the amino-terminus of A. aegypti GPA2 propeptide sequence to the carboxyl- terminus of A. aegypti GPB5 prepropeptide sequence, using multiple PCR amplifications with 6 0 0 several primer sets (Table S2) as performed previously using lamprey GPA2 and GPB5  At 48 hours post-transfection, serum-free culture media containing secreted proteins were 6 1 8 collected and concentrated in 0.5 mL 3-kDa molecular weight cut-off centrifugal filters (VWR 6 1 9 North America). In some experiments, cells were dislodged with PBS containing 5 mM sulfate, 1.5% protease inhibitor cocktail (v/v), and 1.5 mM dithiothreitol (DTT). For receptor 6 2 5 activity assays, in order to prevent carry-over of lysis buffer, cell lysates were concentrated in 3-6 2 6 kDa molecular weight cut-off centrifugal filters and re-constituted back to initial volumes with 6 2 7 serum-free media for a total of three repetitions. Proteins were then used for cross-linking 6 2 8 analysis, deglycosylation and immunoblotting or used as ligands for functional receptor 6 2 9 activation using the cAMP signalling biosensor assays. interactions, was employed to study GPA2 and GPB5 protein-protein interactions. A. aegypti and PNGase and subsequently cross-linked samples with DSS after deglycosylation . HCl, pH ~6.8, and resolved on 10% or 15% SDS-polyacrylamide gels under reducing conditions 6 5 2 at 120 V for 90-110 min. Using a wet transfer system, proteins were then transferred to Acad Sci USA 89:4304-4308. plasma half-life: role of carbohydrate in bioactivity and metabolic clearance.