Structural basis for the interaction between the Drosophila RTK Sevenless (dROS1) and the GPCR BOSS

Zhang, Jianan; Tsutsui, Yuko; Li, Hengyi; Li, Tongqing; Wang, Yueyue; Laraki, Salma; Alarcon-Frias, Sofia; Stayrook, Steven E.; Klein, Daryl E.

doi:10.1038/s41467-025-55943-6

Download PDF

Article
Open access
Published: 18 January 2025

Structural basis for the interaction between the Drosophila RTK Sevenless (dROS1) and the GPCR BOSS

Nature Communications volume 16, Article number: 808 (2025) Cite this article

6202 Accesses
1 Altmetric
Metrics details

Subjects

Abstract

Sevenless, the Drosophila homologue of ROS1 (University of Rochester Sarcoma) (herein, dROS1) is a receptor tyrosine kinase (RTK) essential for the differentiation of Drosophila R7 photoreceptor cells. Activation of dROS1 is mediated by binding to the extracellular region (ECR) of the GPCR (G protein coupled receptor) BOSS (Bride Of Sevenless) on adjacent cells. Activation of dROS1 by BOSS leads to subsequent downstream signaling pathways including SOS (Son of Sevenless). However, the physical basis for how dROS1 interacts with BOSS has long remained unknown. Here we provide a cryo-EM structure of dROS1’s extracellular region, which mediates ligand binding. We show that the extracellular region of dROS1 adopts a folded-over conformation stabilized by an N-terminal domain comprised of two disulfide stapled helical hairpins. We further narrowed down the interacting binding epitopes on both dROS1 and BOSS using hydrogen-deuterium exchange mass spectrometry (HDX-MS). This includes beta-strands in dROS1’s third Fibronectin type III (FNIII) domain and a C-terminal peptide in BOSS’ ECR. Our mutagenesis studies, coupled with AlphaFold complex predictions, support a binding interaction mediated by a hydrophobic interaction and beta-strand augmentation between these regions. Our findings provide a fundamental understanding of the regulatory function of dROS1 and further provide mechanistic insight into the human ortholog and oncogene ROS1.

Structures of rhodopsin in complex with G-protein-coupled receptor kinase 1

Article 14 July 2021

Novel insight into mechanisms of ROS1 catalytic activation via loss of the extracellular domain

Article Open access 27 September 2024

Full-length RNA-Seq of the RHOH gene in human B cells reveals new exons and splicing patterns

Article Open access 16 November 2024

Introduction

Drosophila ROS1 (dROS1), also known as Sevenless, is a critical RTK required for R7 photoreceptor development¹. Sevenless has been a cornerstone in the study of Drosophila eye development and has significantly contributed to our understanding of cell differentiation and signaling pathways. Genetic screens involving Sevenless have led to the identification of key components of the Ras/MAPK signaling cascade, including Son of Sevenless (SOS), which is a guanine nucleotide exchange factor activating Ras1, and downstream kinases². Interestingly, rather than a soluble morphogen or growth factor, the ligand for dROS1 is a membrane-bound GPCR, which is an unusual ligand for an RTK³. The first reported physical interaction between dROS1 and BOSS was a heterotypic cell-aggregation assay involving dROS1-expressing and BOSS-expressing S2 cells, and further supported by co-immunoprecipitation^3,4. It has been proposed that the positional proximity of dROS1 on noncommitted R7 precursor cells with BOSS on adjacent differentiated R8 cells promotes binding through cell-cell contacts⁵. However, no quantitative in vitro study has been performed to confirm nor has any structural study been done to shed light on the details of this unusual interaction.

Like all other RTKs, the architecture of dROS1 is composed of an extracellular region (ECR), a single-pass transmembrane domain (TM), and an intracellular kinase domain (KD). Notably, ROS1 has the largest ECR of any described RTK. Human ROS1 and dROS1 both are predicted to have an ECR composed of nine fibronectin type III domains (FNIII) and three YWTD beta-propellers (Fig. 1a). Importantly, however, no structural studies have been done on ROS1 from any species to date. Therefore, there is no experimental evidence for the ECR architecture of ROS1. Interestingly, dROS1 contains a poly-arginine (9 consecutive arginines) sequence in the last FNIII domain just prior to the transmembrane domain. This sequence harbors multiple consensus furin proteolytic cleavage sites (i.e., RXRR) (Fig. 1a). Notably, cleavage occurs topologically in the same FNIII loop where insulin receptor is processed⁶. Previous literature suggests that the large ECR of dROS1 is cleaved at this position but remains non-covalently associated with the rest of the receptor⁷. This is in contrast to the insulin receptor which is cleaved but remains covalently associated through disulfide bonds to the kinase domain^8,9.

**Fig. 1: The N-terminal region of dROS1 interacts with BOSS’ ECR.**

Surprisingly, the soluble ECR of BOSS is insufficient to activate—and paradoxically inhibits—dROS1 even when presented as a preformed oligomer⁴. This is contrary to the ‘central dogma’ of RTK oligomer-induced activation^10,11. However, a dROS1 mutant that removes most of the ECR is constitutively active¹². While overexpression of many RTKs can lead to ligand-independent activation due to increased receptor dimerization, overexpression of dROS1 or a chimeric dROS1 with the kinase domain of dEGFR (Drosophila Epidermal Growth Factor Receptor) does not result in activation in the absence of ligand, revealing a potent inhibitory role of dROS1’s ECR¹. Together, this points to a unique regulatory mechanism employed by dROS1 unlike that of other well-studied RTKs. To further investigate this activation mechanism, we solved the structure of dROS1’s ECR and analyzed the biophysical basis for its interaction with BOSS.

Here we show, the extracellular region of dROS1 adopts a folded-over conformation stabilized by an N-terminal domain comprised of two disulfide stapled helical hairpins. Using hydrogen-deuterium exchange mass spectrometry (HDX-MS) we identified the interacting binding epitopes on both dROS1 and BOSS. An AlphaFold prediction of their complex, supported by HDX and mutagenesis studies, details the interaction between dROS1 and BOSS necessary for regulating dROS1’s activation. Together, our findings provide a model that sheds light on the regulatory mechanisms governing dROS1 activation.

Results

The N-terminal region of dROS1 interacts with the GPCR BOSS

To characterize the binding interaction between dROS1 and BOSS, we purified the full-length ECR of dROS1, including FNIII domains 1 through 9, construct (1-9) Fig. 1a, b—as well as smaller ECR fragments (Fig. 1b)—and examined their binding affinities towards the ECR of BOSS using surface plasmon resonance (SPR, Biacore) (Fig. 1c). We found that the full ECR of dROS1 binds to BOSS with a K_D of 4.4 µM (Fig. 1c). RTKs with soluble ligands typically exhibit nanomolar binding affinities or require co-receptors to capture dilute signals¹¹. In contrast, dROS1 binds to BOSS with a comparatively modest micromolar affinity. Since BOSS is also a membrane-bound receptor, this binding affinity is consistent with that of other cell-cell adhesion receptors (i.e. notch and delta¹³) where both ligand and receptor are spatially restricted. We then determined that the N-terminal region of dROS1 (constructs 1–4 and 2-3) also binds to BOSS with a similar affinity (Fig. 1c). Thus, these truncations narrow down the binding epitope on dROS1 to the region flanked by the second and third FNIII domains (construct 2-3) and includes the first YWTD propeller domain.

The N-terminal region of dROS1 adopts a folded-over conformation

To establish a biophysical understanding of dROS1 function we sought to obtain a 3-Dimensional structure of dROS1’s ECR. Notably, there is no published structural information on the ECR of ROS1 from any species. We expressed and purified the dROS1 ECR (construct 1–6) (Fig. 2a) and obtained a 3.94 Å structure using single-particle CryoEM (Fig. 2b; Supplementary Fig. 1). Surprisingly, model building only resolves the N-terminal region (equivalent to construct 1-4) starting from the N-terminus to the fourth FNIII domain (Fig. 2a, b). The C-terminal domains starting from the fifth FNIII domain to the sixth, including the intervening YWTD-C, were not observed in the map (Fig. 2a, b). The lack of density for the C-terminal domains is consistent with this region demonstrating greater flexible with respect to the N-terminal domains. The poor density and low resolution of the fourth FNIII domain also support this idea (Supplementary Fig. 1). Despite the flexibility, we were able to further improve the density of the fourth FNIII domain by masked local refinement (Supplementary Fig. 2).

**Fig. 2: CryoEM Structure of the N-terminal ligand binding region of dROS1.**

Interestingly, the model reveals that rather than having a linear arrangement of domains, the N-terminal region adopts a folded-over conformation stabilized by an interaction of an N-terminal cysteine-rich region with the second propeller (YWTD-B) and fourth FNIII domain (Fig. 2b, c). The N-terminal cysteine-rich region forms a previously unappreciated domain (Fig. 2a–c, yellow). The domain is comprised of two disulfide stapled helical hairpins that associate with one another through hydrophobic interactions (Fig. 2c, d). The two helical hairpins generate two loops (capping loops 1 and 2) which stabilize the N-terminal ECR compacted fold. Despite the rather limited resolution in this region, we observe clear density for bulky side-chains which assisted in model building (Supplementary Fig. 3).

In the loop generated by the first helical hairpin, capping loop 1, F235 sits between R1234 and L1236 of the fourth FNIII domain (Fig. 2e). The loop generated by the second helical hairpin, capping loop 2, also interacts with the fourth FNIII domain with residue F279 sitting between L1243 (FNIII-4) and R252 of the first hairpin (Fig. 2e). Using a similar interaction, loop 1 also makes contact to the YWTD-B domain. F231 sits between R1032 and L1034 of the second propeller. This interaction is further supported by a main chain hydrogen bond between L232 and R1032 (Fig. 2f). Notably, these surface exposed hydrophobic residues and basic residues are well-conserved across different insect species, suggesting the importance of the capping interactions and conservation of the folded-over conformation (Supplementary Fig. 4).

The disulfide stapled helical hairpin structure of the N-terminal cysteine rich region of dROS1 is unusual. However, it reminded us of a structure that we recently solved of the related RTK ALK (Anaplastic Lymphoma Kinase) bound to its ligand ALKAL (ALK And LTK Ligand)¹⁴. Indeed, the unique cysteine-rich region of dROS1 structurally resembles ALKAL (Supplementary Fig. 5). Both share the rare fold of a doubly disulfide-stapled helical hairpin. When the helical hairpins of dROS1 and ALKAL are structurally aligned, it is evident that they utilize similar surfaces to engage with their respective receptors (Supplementary Fig. 5b, d). This alignment highlights a convergent structural and functional strategy. Therefore, from this perspective, dROS1 could be envisioned as a self-ligated receptor through the N-terminal cysteine rich region, stabilizing a bent-over conformation. Given the relatedness of this region to ALKAL and its role in stabilizing the dROS1 N-terminal ECR, we designate it as a CATCH domain for (Cysteine-rich ALKAL-Type Coupled Helices).

Structural prediction of BOSS reveals a “Venus-flytrap” architecture similar to mGluRs

The evolution of the orphan GPCR BOSS diverged so rapidly that tracing homologs by sequence is incredibly difficult¹⁵. However, the sequence of BOSS’ transmembrane segments suggests homology to metabotropic glutamate receptors (mGluRs) among Class-C GPCRs¹⁵. An AlphaFold^16,17,18,19 prediction of BOSS’ ECR suggests it adopts a bi-lobed “Venus-Flytrap” domain (VFTD) architecture, also similar to mGluRs (Fig. 3a–c). A Dali^20,21 structural comparison search of the predicted structure shows the highest similarity to mGluR-2 (Supplementary Fig. 6).

**Fig. 3: BOSS’ ECD structure is predicted to have a Venus-flytrap fold similar to mGluRs.**

When compared with mGluR-2, BOSS’ ECR reveals several unique features. BOSS lacks a Cysteine-rich domain (CRD) “stem” region found in mGluR-2 that separates the VFTD and membrane region (Supplementary Fig. 7a). Instead, BOSS is predicted to have a configuration parallel to the membrane where after Lobe 2 a peptide leads to the transmembrane region (Fig. 3b, c). This includes a linker peptide (residues 480–500) following strand 7 of Lobe 2. A cysteine at 501 forms a disulfide bond restraining this peptide to Lobe 2. However, this peptide has a very low pLDDT (per-atom confidence estimate) score indicating a poor model—consistent with a dynamic and flexible peptide (Fig. 3b, c).

BOSS lacks glutamate binding and receptor dimerization motifs

Notably, while BOSS ECR shows structural similarity to the VFTD of mGluRs, it does not retain many of the functionally important residues. When the first Lobe of BOSS and mGluR-2 are aligned, it is apparent that the key residues for binding glutamate are not retained (Fig. 3d). Indeed, Helix-A of Lobe 1 does not extend far enough to reach the glutamate binding pocket in BOSS – lacking the critical arginine (R61) for glutamate binding (Fig. 3d, orange portion Helix-A). Additionally, BOSS completely lacks a structural loop feature (residues 166-186) in mGluRs that forms the glutamate binding pocket (including A166 and T168) (Fig. 3d, orange). Rather, in BOSS the first helix of Lobe 2 arises directly from the beta strand in Lobe 1 that precedes the glutamate binding structural loop. Importantly, the first helix of Lobe 2 positions the entire Lobe 2 (as a rigid body) with respect to Lobe 1. This helix moves after glutamate binding in mGluRs – moving from an “open” to “closed” state (Fig. 3d, black to gray). BOSS’ first helix of Lobe 2 is found in neither of these positions, and is >15 Å away, passing directly through the glutamate pocket (Fig. 3d). Therefore, BOSS has a unique orientation of Lobe 2 and could be considered as “pocketless”.

Furthermore, the residues that dictate dimerization of mGluRs are also not conserved in BOSS. In mGluR-2, dimerization is directed by Helix-B and Helix-C of Lobe 1 (Supplementary Fig. 7a, b). When structurally aligned, the hydrophobic nature of the mGluR-2 dimer interface is not maintained in BOSS (Supplementary Fig. 7b). Indeed, Helix-C is smaller in BOSS and lacks the terminal L157 and F158 equivalent residues that are crucial for dimerization (Supplementary Fig. 7c, orange). Together, this indicates that despite sharing a Venus-flytrap fold, BOSS’ function is distinct from that of mGluRs.

Mapping the binding interfaces in the dROS1-BOSS complex using HDX-MS

To find the binding interface for the interaction between dROS1 and BOSS, we used hydrogen-deuterium exchange and mass spectrometry (HDX-MS). HDX-MS probes protein conformation and protein-protein interactions by assessing deuterium uptake of the amide hydrogen atoms of the protein backbone²². By comparing deuterium uptake of peptide regions between each monomeric protein and the complex (ΔDs(i))²³, potential binding regions can be identified (Fig. 4). We have previously used this technique to map the binding epitope of antibodies targeting the RTK ALK¹⁴. For this experiment we used the dROS1 (1-4) construct and the full-length BOSS ECR. These constructs were stable at the concentrations necessary for the experiment and include the interaction sites (Fig. 1c). Given that the affinity between these proteins is relatively weak (~3–5 µM K_D), we developed a strategy to preserve the interaction by immobilizing His-tagged protein on Ni-NTA resin to increase the local concentration (details in Methods). To aid in visualization, regions protected from exchange (i.e. stabilized regions that are candidate binding epitopes) were mapped onto the structures of dROS1 and BOSS (Fig. 4b, d, Supplemental Fig. 8, Supplemental Table 2).

**Fig. 4: The dROS1-BOSS interaction probed by HDX-MS.**

Regions protected from deuterium exchange in dROS1 upon BOSS binding are found throughout dROS1 (Fig. 4a, b, blue). However, within the determined binding region of dROS1 (second through third FNIII domains) (Fig. 1c), significantly protected regions are localized in the third FNIII domain. This includes the first strand of the domain (residues 814–838), as well as the third and fourth strands (residues 876–896). Between these peptides, peptide 878-896 is more surface exposed and thus more likely to be a direct interaction site (Fig. 4a, b, blue).

HDX-MS results show that one of the significant regions in BOSS upon dROS1 binding includes the C-terminal beta-strand “FVTNVTTY” (residue 471–478, Fig. 4c, d, blue). This corresponds to strand 7 of Lobe 2 (Fig. 3c). Another stabilized region in BOSS includes the central linker that spans Lobe 1 and Lobe 2—beginning with Helix-B of Lobe 1 and passing through to the central first Helix of Lobe 2 (residues 220-239) (Fig. 4c, d, blue). This corresponds to the lobe linking peptide highlighted in Fig. 3d. As this peptide is mostly confined between lobes, it is unlikely to directly participate in dROS1 binding. Thus, based on our HDX results, aided with our structural prediction (Fig. 3), peptide 471–478 (C-terminal beta-strand “FVTNVTTY”) may represent a direct binding epitope for dROS1. Together, our HDX-MS data combined with our structure and binding data narrows down the likely complex binding epitope to dROS1’s third FNIII domain and a strand near BOSS’ ECR C-terminal loop.

Structural prediction of the dROS1-BOSS complex

In the absence of an experimentally determined dROS1-BOSS complex structure, we sought to generate a theoretical model of the complex using AlphaFold^16,17,18,19 (see Methods). Since our binding studies narrowed down the binding epitope to the N-terminal region of dROS1, we were able to apply this constraint when generating models. Initial predicted complex structures using AlphaFold 2 all narrowed in on third FNIII domain of dROS1. We then used AlphaFold 3¹⁹—which allows for post translational glycosylation—to generate a model of the target region. All AlphaFold 3 complexes converged with high confidence (ipTM=0.77, pTM=0.85) to a single model (Fig. 5a, b). Importantly, the predicted structure is consistent with our HDX-MS and binding experiments—the predicted complex interface includes the very epitopes identified in our HDX-MS (Fig. 5c). In the model, the interaction of the complex is mediated by parallel beta-strand augmentation between dROS1’s third FNIII domain (874-890) and a C-terminal strand of BOSS’ ECR (474-491) (Fig. 5d).

**Fig. 5: Structural model of the complex between dROS1 and BOSS.**

In the complex, strand augmentation builds on dROS1’s third FNIII’s strands (873-890) by coupling to two strands on BOSS (474-491), in parallel. Notably, this region of BOSS (ECR C-terminal peptide) had a low pLDDT score (<50, orange) in the unbound state—consistent with disorder (Fig. 3b, c). In the model complexed with dROS1, this peptide generates a new strand with high pLDDT scores (>90, dark blue), indicating high confidence in the structure (Fig. 5b). This is consistent with the C-term peptide of BOSS going from a disordered to ordered structure as it mediates binding to dROS1. The packing of the beta-stand augmentation is supported by an extension of dROS1’s FNIII’s hydrophobic core. This is coordinated by BOSS’ Leucine residues L489 and L490 making hydrophobic interactions with dROS1’s FNIII core (F876, I878, F886 and L888) (Fig. 5d). In addition to main chain beta-strand interactions, there are multiple predicted side chain interactions that involve basic residues from dROS1 interacting with acidic residues on BOSS. This includes: R879 on dROS1 interacting with E465 of BOSS; R882 bonding with D484 on BOSS, and R890 bonding with E490 on BOSS (Fig. 5d).

We wished to test the predicted complex model using site directed mutagenesis. However, site specific mutagenesis to disrupt the complex interaction is difficult to generate as the interaction is mediated mainly by main chain strand augmentation. Mutation or deletion that overtly disrupts the strand’s secondary structure would likely lead to misfolding of the domain itself. Therefore, we sought to disrupt the hydrophobic interactions mediated by dROS1’s third FNIII domain including the peptide (F876-S877-I878), “FSI”, with the Leucine residues on BOSS (L489-E490-L491), “LEL”. Notably, the “FSI” residues are not conserved between the second and third FNIII domains in dROS1, which could explain the specificity for the interaction with the third FNIII domain (Supplementary Fig. 9a). Rather than the hydrophobic residues “FSI”, the second FNIII domain has a hydrophilic sequence “SEQ”. Therefore, to disrupt the complex interaction by mutagenesis without disrupting the fold, we mutated the “FSI” of third FNIII domain into the corresponding residues “SEQ” found in the non-binding second FNIII domain. Binding results of dROS1_SEQ demonstrate no observable binding to BOSS (Fig. 5e). This result is consistent with the model. However, the dROS1_SEQ construct was less stable and less pure than the WT construct.

Therefore, to further validate the model, we sought to mutate the residues on BOSS that interact with the hydrophobic core of dROS1’s third FNIII domain – the reciprocal to the dROS1_SEQ experiment. Mutation of Leucine residues in BOSS (L489 and L491) to non-hydrophobic residues (“LEL” to “DEQ”; BOSS_DEQ) expressed as well as the WT indicating that the mutations did not affect protein stability. SPR analysis of BOSS_DEQ binding to the full ECR of dROS1 (Construct (1-9)) shows no detectable binding (Fig. 5e). This finding is consistent with the model of the interaction between dROS1 and BOSS (Fig. 5a and c).

Impact of binding site mutations on dROS1 cell activation

To assess the impact of binding site mutations on the signaling of full-length dROS1 in response to membrane-embedded full-length BOSS, we created a chimeric dROS1 receptor with a human ALK kinase domain. ALK was selected because it is the RTK with the highest kinase domain homology to ROS1, and both cluster within the same family²⁴. This chimera enables us to track the levels of total and phosphorylated dROS1 receptor using well established antibodies targeting the ALK kinase domain, as in our previous studies on human ALK activation¹⁴.

A stable NIH-3T3 cell line expressing the chimeric dROS1_ALK receptor was established. To evaluate the ability of wild-type BOSS and the binding site mutant BOSS_DEQ to stimulate dROS1_ALK phosphorylation, we transiently expressed either BOSS or its mutant counterpart in the stable dROS1_ALK cells. Cell lysates confirmed the expression of His-tagged BOSS in transiently transfected cells (Fig. 5f). BOSS was successfully pulled down using Ni resin, demonstrating enrichment of BOSS. The Ni-purified BOSS was probed for total and phosphorylated dROS1_ALK.

Wild-type BOSS was able to pull down dROS1_ALK, whereas the BOSS_DEQ mutant did not (Fig. 5f). Additionally, phosphorylated dROS1_ALK was detected only in the presence of wild-type BOSS but not with the mutant, consistent with the binding impairment caused by the DEQ mutation (Fig. 5f). Together, the binding and cellular studies support the structural model of the interaction between dROS1 and BOSS.

Discussion

dROS1 is homologous to the human RTK and oncogene ROS1, which is the least well-studied RTK¹⁰. Multiple ROS1 fusion proteins have been discovered to play an important role in a variety of human cancers including glioblastoma, non-small-cell lung cancer, and breast cancer^25,26. Due in part to the difficulty of purifying ROS1, no structural insight has yet been established to reveal the architecture of its ECR region. In addition, ROS1 is the last orphan human RTK, with its ligand remaining to be firmly established. Also, a mammalian BOSS homolog has not been identified⁵. Thus, ROS1 is similar to the RTK ALK, where the ligand controlling its activation is not structurally conserved between invertebrates and vertebrates^27,28,29,30.

We find here that dROS1 is regulated by relatively weak interactions with BOSS that are enhanced by the high local concentration when restricted to two dimensions between juxtaposed cell membranes. Our biophysical characterization narrows down the binding epitope between the interaction of dROS1 and BOSS. This could provide insight into the ligand binding region of mammalian ROS1. Indeed, for ALK, despite having structurally different ligands for different species, the binding regions are largely maintained¹⁴. However, alignment of the interreacting FNIII domain of dROS1 with mammalian ROS1 shows non conservation of the key hydrophobic residues in the peptide “FSI” (Supplementary Fig. 9b, c). This would indicate that if a ligand does bind at this site on ROS1, it likely uses a different set of interactions. Different ligands have been proposed to regulate mammalian ROS1 including NELL2³¹ and Ribonuclease-7³². Both are soluble factors and not membrane anchored like BOSS. Therefore, it is possible that if ROS1 responds to soluble factors, it may use a different mechanism for binding and regulation.

Earlier efforts to elucidate dROS1’s activation mechanism showed that neither a soluble monomeric ECR of BOSS nor a forced oligomer of BOSS’ ECR is sufficient to stimulate dROS1’s kinase activity^4,33. Instead, a membrane anchored BOSS is required to induce R7 differentiation³³. These findings, coupled with our AlphaFold predictions indicating BOSS is a monomeric protein, point to a mechanism whereby BOSS engagement of dROS1 from an opposing membrane may help coordinate a conformational change in dROS1 that is compatible with receptor dimerization and resulting activation (Fig. 6a, b). This would be similar to how EGF (Epidermal Growth Factor) regulates its receptor EGFR: a monomeric ligand induces receptor structural changes that relieve auto-inhibition and allow the receptor to dimerize³⁴. Our structure, of dROS1’s N-terminal ECR—being compact rather than linear, suggests that the bulky ECR could physically keep the kinase domain at distance and prevent oligomerization and signaling in the absence of ligand binding. An activation dependent conformational change in dROS1 may require tension afforded by both proteins being anchored in different membranes—and this tension is lost in soluble ECR constructs of BOSS whether monomeric or oligomeric, rendering them inactive^4,33.

**Fig. 6: Structural model of the dROS1 BOSS engagement complex.**

A Dali^20,21 structural comparison dROS1’s unique CATCH domain (including both helical hairpins) shows distant structural similarity to the complement component protein C3a (Supplementary Fig. 10a, b). However, the disulfide bond connections are largely not maintained (Supplementary Fig. 10c, d). Notably, in the structure of the unprocessed complement C3 (PDB:2A73)³⁵, the C3a peptide interacts with the complex via the loop connecting H3-H4 (Supplementary Fig. 10e). This is equivalent to capping loop 2 of dROS1’s CATCH domain. C3a also binds to and activates that C3a receptor (C3aR)—a GPCR^36,37. Interestingly, the ligand for dROS1, BOSS, is an orphan GPCR. Therefore, given that dROS1’s CATCH domain structurally mimics a GPCR ligand, this finding suggests that crosstalk between dROS1 and BOSS might be bidirectional. As the ECR of BOSS functions as a ligand for dROS1, so too the ECR of dROS1 might function as a ligand for BOSS. However, functional or evolutionary relatedness is not established, and further studies are needed to explore this possibility.

Future studies will focus on investigations with dROS1’s full-length receptor to determine whether dROS1 resembles other well-known RTKs and cytokine receptors in their self-assembly upon ligand binding. While BOSS does not have a direct full-length homolog in humans, its TM regions show similarity to GPRC5B, and both respond to glucose stimulation³⁸. However, GPRC5B completely lacks an ECR and could not function as a ROS1 ligand. More research should be done to determine whether dROS1 regulates this signaling function of BOSS. We acknowledge that during the review process, a related study by Cerutti et al. was published³⁹, and their Cryo-EM analysis identifies the same binding interface that we determined through HDX and modeling.

Methods

Recombinant protein expression and purification

BOSS ECR (residues 102-510) and dROS1 from D. melanogaster with various constructs dROS1 1-9 (residues 126-2113), 1–6 (residues 215-1795), 1–4 (residues 126-1298), 2-3 (residues 440-931), were incorporated into pFastBac plasmid (Invitrogen) with octa-histidine tag combined with a Factor Xa cut site at the N-terminus and Flag-tag at the C-terminus. The protein was then expressed in Trichoplusia ni (Hi5) cells driven by baculovirus infection at a density of 2 × 10⁶ cells/mL. Mutants of BOSS and dROS1 were generated by site-directed mutagenesis using QuickChange Kit (Agilent Technologies). Proteins were purified from the medium (Ex-Cell405, Sigma-Aldrich) 2 days post-infection by directly flowing over Ni-Penta™ Agarose-Base Resin (Marvelgent Biosciences). Resins were washed 3 times with 35 mL of 20 mM imidazole (pH 8.0), 100 mM NaCl, and proteins were eluted first with 18 mL of 200 mM imidazole (pH 8.0), 100 mM NaCl and then with 18 mL of 500 mM imidazole (pH 8.0), 100 mM NaCl. Eluted proteins were further purified by size-exclusion chromatography (HiLoad 26/600 superdex 200 pg, GE Healthcare Life Sciences) in 20 mM HEPES (pH 7.5), 150 mM NaCl. Corresponding peaks were checked by SDS-PAGE (sodium dodecyl sulfate-polyacrylamide gel electrophoresis) gel and concentrated by Amicon Ultra-4 10 kDa MWCO concentrator (Millipore).

The His-tags of BOSS proteins and the mutant were cleaved by Factor Xa protease (New England Biolab) with 1 mM CaCl₂ and flowed over Ni-Penta™ Agarose-Base Resin (Marvelgent Biosciences). The flow-through was further purified by size-exclusion chromatography (Superdex 200 Increase 10/300 GL size exclusion column GE Healthcare Life Sciences) in 20 mM HEPES (pH 7.5), 150 mM NaCl. dROS1 (1-4) without his-tag was generated following the same protocol (for the HDX-MS experiment).

CryoEM sample preparation and data collection

For cryo-electron microscopy, 3.5 µl of purified dROS1 (construct 1–6) at 500 nM were applied on glow-discharged Quantifoil Cu R1.2/1.3 300 mesh grids, and blotted for 5 s at 100% humidity and 16 °C. Samples were plunge-frozen into liquid ethane using Vitrobot Mark IV (FEI) and transferred into liquid nitrogen. Grids were imaged under the 300 kV Titan Krios electron microscope (FEI) with a K3 summit direct electron detector (Gatan). Images were recorded under super-resolution mode at a pixel size of 0.825 Å/pixel using SerialEM v3.8.6 and Digital Micrograph v3.31.2359.03. Data were collected with a dose rate of 17.5 e-/pix/s for 0.06 s/frame. A total of 7835 images were recorded and processed using CryoSPARC v4.4.1^40,41. Detailed parameters of data collection are summarized in Supplementary Table 1.

CryoEM data processing and 3D reconstruction

Raw movies were motion-corrected using patch motion correction and contrast transfer function (CTF) parameters were estimated using patch CTF⁴⁰. Bad micrographs were discarded by manual inspection. A small set of particles were manually picked, extracted and 2D classified as templates for autopicking. The template-picked particles were extracted with a box size of 786 Å and Fourier cropped to 192 Å, which was then used to generate 2D classes. Particles with good 2D classifications were re-cropped to 384 Å and re-extracted, which were then used for one round of ab initio reconstruction into 4 classes and iterative rounds of heterogeneous refinement into 5 classes. The best model (388,364 particles) was selected for non-uniform refinement. The overall resolution is 3.74 Å estimated by Fourier shell correlation (FSC) at 0.143 cutoff. These particle sets (388,364 particles) were then used in additional rounds of heterogeneous refinement into 4 classes, which gave us a 3.94 Å map (144,311 particles) with more density. These volumes were assessed by local resolution estimate using the CryoSPARC v4.4.1⁴⁰ program and the final resolution was colored using UCSF ChimeraX. Both volumes were used in model building. Local mask was created covering the CATCH domain, FNIII-1, FNIII-2 and FNIII-4, from the 3.94 Å map using UCSF ChimeraX and imported back into CryoSPARC v4.4.1⁴⁰. Local refinement was performed using created local mask and yield a slightly more density around the masked area.

CryoEM model building

The initial model of the structure was generated using AlphaFold2^16,17,18. The model was segmented into individual domains and fitted into a map using ChimeraX. Given that the 3.74 Å map has a higher resolution, FNIII-1 to YWTD-B were fitted using this map. The cysteine-rich domain and FNIII-4 were fitted into the 3.94 Å map. Domains were then merged and imported into Coot 0.9.6⁴² from SBGrid⁴³ along with the two maps for manual inspection. Real space refinement was done using Phenix 1.21-5207⁴⁴ from SBGrid⁴³. Detailed statistics and parameters for model building and refinement are provided in Supplementary Table 1. Figures prepared with Adobe Illustrator 27.1.1.

Surface plasmon resonance

Surface plasmon resonance (SPR) experiments of dROS1 BOSS interactions were all performed using BIAcore T200 in 20 mM HEPES (pH7.5), 100 mM NaCl. The Series S CM5 biosensor chip surface was activated with N- hydroxysuccinimide (NHS) and N-ethyl-N’-[3-(diethylamino)propyl] carbodiimide (EDC). dROS1 constructs at 30 ng/µL in 10 mM acetate pH 4.0–6.0 were flowed over the activated surface at 10 µL/min for 300 s for covalent coupling. The remaining active sites were blocked by 1 M ethanolamine pH 8.5. Purified ligand BOSS proteins at varied concentrations were flowed over the dROS1 surfaces at 10 µL/min for 180 s. The response unit (RU) values at the plateau, corrected by subtracting background binding from the control surface, were used to measure of binding. Binding affinities were calculated by plotting the plateau RU values against BOSS concentrations and fitting the data to a single-site specific binding model using GraphPad Prism9. All binding curves were then plotted as BOSS concentrations versus maximal binding (Bmax).

AlphaFold prediction

dROS1 model was searched from AlphaFold protein structure database (DeepMind and EMBL-EBI) and the construct 1–6 (residues 2156-1795) was used to generate the model for CryoEM map. AlphaFold 3¹⁹ (Alphafoldserver.com) was used for generating the prediction of full-length BOSS.

For the complex, initially, BOSS ECR (residues 102-510) and dROS1 construct 2-3 (residues 440-931) were used for heterocomplex prediction with AlphaFold 2 using the publicly available ColabFold v1.5.3⁴⁵. This construct was chosen as it included the determined binding epitope and was of a size capable of running on the available resources. MSA options were set to ‘mmseqs2_uniref_env’ and pair mode was ‘unpaired_paired’. All models converged on a binding epitope that only included the third FNIII domain of dROS1. However, the models exhibited a combination of parallel and anti-parallel binding modes. We then used AlphaFold 3 which allows for post translational modifications (glycosylations) to generate models of the complex between BOSS and dROS1’s third FNIII domain. We added a single NAG (N-Acetyl-beta-D-glucosamine) to the predicted sites in BOSS (including those near the interface, N474 and N485). All resulting models converged with high confidence to the single complex used in this study.

For the full-length complex model we docked a full-length dROS1 model from AlphaFold 3 onto the complex, also AlphaFold 3, of full-length BOSS bound to the third FNIII domain of dROS1 (ipTM = 0.8, pTM = 0.68).

Hydrogen-Deuterium Exchange and Mass spectrometry (HDX-MS)

Probing low-affinity interaction such as dROS1-BOSS poses a major challenge in HDX-MS experiments because dROS1-BOSS complex dilution with the deuterium labeling buffer (20 mM HEPES, 100 mM NaCl, pD 7.4) can lead to the complex dissociation. To maintain at least 80% of the complex after 10-fold dilution with the deuterium buffer, dROS1 at 0.6 mg/mL (4.5 μM) must be pre-incubated with a large excess amount of BOSS (210 μM, 11.8 mg/mL). Although such an experiment enables one to probe the complex, resulting mass spectra would be largely dominated by BOSS peptides due to the greater amount of BOSS, making dROS1 deuterium uptake assessment extremely challenging. Furthermore, protein sample overloading onto pepsin and UPLC columns can result in column overpressure and clogging. To circumvent these problems, we designed deuterium labeling experiment based on a previous study where the effect of weak EGFR kinase dimerization on the kinase activity was studied by immobilizing the his-tagged kinase domain on DOGS Ni-NTA lipid vesicles⁴⁶. The immobilization of the his-tagged protein increases its local concentration such that the kinase domain with weak propensity to dimerize could be dimerized (the kinase remains monomeric up to 50 μM).

In our new deuterium labeling protocol, his-tagged dROS1 (construct 1-4) or BOSS (0.6 mg/mL, 100 μL) was immobilized on 50 µL packed Ni-Penta™ Agarose-Base Resin (Marvelgent Biosciences) for 10 min. Next, either untagged protein was co-incubated with the immobilized dROS1 or BOSS followed by gentle centrifugation at 30 x g for 30 s to remove the unbound protein in flow-through. Immobilized protein concentration on Ni-NTA and appropriate binding partner concentration were determined and optimized by eluting the immobilized protein in a quench buffer (200 mM glycine, 100 mM TCEP, pH 2.4) before proceeding to deuterium labeling experiments. For deuterium labeling, either immobilized dROS1 or BOSS (0.6 mg/mL, 100 μL) was co-incubated with untagged protein to form the complex, followed by a brief centrifugation (30 g for 30 s) to remove the unbound protein. Subsequently, the deuterium labeling buffer (50 μL, 20 mM HEPES, 100 mM NaCl, pD 7.4) was added to the Ni-resin and incubated for 10 min before the addition of cold 50 μL quench buffer (200 mM glycine, 100 mM TCEP, pH 2.4) to stop the labeling and elute the complex off the Ni-resin. The labeled sample was immediately flash-frozen in liquid N₂ and stored at −80°C until mass spectrometry analysis. Similarly, control deuterium labeling experiments were carried out with either his-tagged dROS1 or BOSS only. A deuterium-labeled or -unlabeled protein was thawed quickly and immediately injected onto an Enzymate BEH pepsin column (Waters) maintained at 2°C to digest proteins for 3 min in UPLC solvent A (water + 0.1% formic acid) at 100 μL/min. Peptic peptides were trapped in BEH C18 pre-column (2.1 × 5 mm, 1.7 μm, Waters), were separated on BEH C18 analytical column (1.0 × 100 mm, 1.7 μm, Waters), and eluted from the analytical column by 5% to 45% linear solvent B (acetonitrile + 0.1% formic acid) gradient over 7 min. HD-MS^e data were acquired at 300–1200 m/z range using Synapt G2-Si mass spectrometer (Waters) using 0.4 s scan time in sensitivity ion mobility mode with Leu-Enk as a lock CCS and mass compound. Other instrument parameters were: capillary voltage, 3 kV; cone voltage, 10 V; source offset, 80 V; source temperature, 100 °C; desolvation temperature, 200 °C; cone gas flow, 0 L/Hr; desolvation gas flow, 700 L/Hr; nebulizer gas, 7.0 Bar. The low and high energy collision ramp was 5 to 10 V and 20 – 50 V, respectively.

HDX-MS data analysis

ProteinLynx Global Server 3.03 (Waters) was used to sequence each peptic peptide, and the deuterium uptake of peptic peptides was evaluated using DynamX 3.0 (Waters), with the minimum intensity cut-off set to 1000, minimum product per amino acid to 0.1, and maximum MH⁺ error set to 10 ppm. The deuterium uptake of both dROS1 and BOSS peptides is the average uptake of two biological samples (n = 2), with 3 technical repeats per biological sample. Peptides with a deuterium uptake standard deviation >0.5 Da from 2 biological repeat experiments were excluded from the analysis.

The HDX-MS data statistics, including the number of analyzed peptides, the peptide coverage, average peptide length/redundancy, and deuterium uptake standard deviations, as recommended by the HDX-MS guidelines⁴⁷, are presented in Supplemental Table 2 and the Source data file. To identify a binding region in dROS1 upon BOSS binding, the deuterium uptake of immobilized dROS1 in the presence of untagged BOSS was evaluated, and the uptake was compared to that of the his-tagged dROS1 without BOSS for obtaining the difference in the deuterium uptake (Δuptake). Similarly, to identify a binding region in BOSS upon dROS1 binding, the deuterium uptake of the immobilized BOSS in the presence of untagged dROS1 was assessed, and the uptake was compared to that of his-tagged BOSS without dROS1. The statistical significance for the sum of the Δuptake of each peptic peptide at 90% confidence limit was calculated as described previously²³. All data were collected and analyzed according to consensus HDX-MS guidelines⁴⁷.

Generation of Sevenless stable expressing cell line

The gene for the full extracellular domain of dROS1 (Sevenless) was fused with ALK’s transmembrane and kinase domains into the pcDNA3.1 plasmid using the NEBbuilder HiFi DNA Assembly Cloning Kit (New England Biolabs). An NIH 3T3 cell line (ATCC CRL-1658) was used to generate a stable cell line of dROS1_ALK. This plasmid was transfected into NIH 3T3 cells with Lipofectamine 2000 (Thermo Fisher Scientific). The NIH 3T3 dROS1_ALK stable cell line was selected with complete medium containing 400 µg ml⁻¹ G-418 (Santa Cruz Biotechnology).

Cellular pull-down assay

Full-length BOSS, with a C-terminal hexahistidine tag was subcloned into a pVRC vector for expression. The Phusion Site-Directed Mutagenesis Kit (Thermo Fisher Scientific) was used to generate the BOSS DEQ mutant (designated as: BOSS_DEQ). dROS1_ALK stable cells were seeded 2 days before transient transfection. The BOSS wild type plasmid or DEQ mutant plasmid was transfected into dROS1_ALK stable cells using Lipofectamine 2000 (Thermo Fisher Scientific) and the cells were starved overnight after 24 h of transient expression. The transfected cells were placed on ice and washed with ice-cold PBS twice. The cells were scraped and lysed with lysis buffer (20 mM HEPES pH 7.5, 150 mM NaCl, 1% NP-40) with 1× protease/phosphatase inhibitor cocktail (Cell Signaling Technology no. 5872) on ice. The lysed cells were then centrifuged at 15000 rpm at 4 °C for 15 min and supernatants were collected. 50 µL of Ni-Penta Agarose-Base Resin (Marvelgent Biosciences) was used for each pull-down sample. The resins were washed three times with 1 mL PBS and mixed with supernatant and incubated at 4 °C for 1 h with gentle rotation. The resin was then washed five times with 1 mL of lysis buffer for each round and bound proteins were eluted with lysis buffer containing 500 mM imidazole.

Western blot

The pull-down samples were boiled in 1 × Laemmli sample buffer (Bio-Rad) and then separated on 4–12% Tris-Glycine gradient gel. The separated proteins were then transferred to PVDF membrane in transfer buffer (25 mM Tris, 200 mM glycine, 10% methanol (v/v)) for 49 min on ice. The membranes were first blocked with 5% milk in TBST buffer (20 mM Tris pH 7.5, 150 mM NaCl, 0.1% Tween 20) for 1 h at room temperature and then incubated with primary antibodies overnight at 4°C. For detection of BOSS, a penta-his antibody (Qiagen, Cat. No. 34660) with 1:1000 dilution was used. The dROS1_ALK receptor signal was detected by anti-ALK antibody (1:1000, Cell Signaling Technology, no. 3633). The receptor’s phosphorylated signal was detected by anti-phospho ALK (Y1507) antibody (1:1000, Cell Signaling Technology, no. 14678). The membranes were then washed five times with TBST buffer and incubated with corresponding secondary antibodies. The HRP-linked anti-rabbit IgG (Cell Signaling Technology, no. 7074S) was used at a dilution of 1:4000. The HRP-linked anti-mouse IgG (Abcam, ab97046) was used at a dilution of 1:5000. The membranes were further developed with Amersham ECL Prime western blotting detection reagent (Cytiva Life Sciences) and imaged using a ChemiDoc (Bio-Rad).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The atomic coordinates for dROS1 have been deposited in the PDB with accession code 9BHX. The EM map has been deposited in the EMDB with accession codes EMD-44556. The monoisotopic mass (MH⁺), centroid mass, deuterium uptake, and retention time of all peptides are detailed in the Source Data file. The HDX-MS data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository, with the dataset identifier; PXD051929. All other data needed to support the conclusions in this manuscript can be found in the main text or supplementary materials. All materials are available from the corresponding authors upon request. Source data are provided with this paper.

References

Tomlinson, A., Mavromatakis, Y. E. & Arias, R. The role of sevenless in Drosophila R7 photoreceptor specification. Dev. Biol. 454, 181–189 (2019).
Article PubMed PubMed Central Google Scholar
Raabe, T. The sevenless signaling pathway: variations of a common theme. Biochim. Biophys. Acta 1496, 151–163 (2000).
Article PubMed MATH Google Scholar
Kramer, H., Cagan, R. L. & Zipursky, S. L. Interaction of bride of sevenless membrane-bound ligand and the sevenless tyrosine-kinase receptor. Nature 352, 207–212 (1991).
Article ADS PubMed Google Scholar
Sevrioukov, E. A., Walenta, J. H., Sunio, A., Phistry, M. & Kramer, H. Oligomerization of the extracellular domain of Boss enhances its binding to the sevenless receptor and its antagonistic effect on R7 induction. J. Cell Sci. 111, 737–747 (1998).
Article PubMed Google Scholar
Acquaviva, J., Wong, R. & Charest, A. The multifaceted roles of the receptor tyrosine kinase ROS in development and cancer. Biochim. Biophys. Acta 1795, 37–52 (2009).
PubMed Google Scholar
Bravo, D. A., Gleason, J. B., Sanchez, R. I., Roth, R. A. & Fuller, R. S. Accurate and efficient cleavage of the human insulin proreceptor by the human proprotein-processing protease furin. characterization and kinetic parameters using the purified, secreted soluble protease expressed by a recombinant baculovirus. J. Biol. Chem. 269, 25830–25837 (1994).
Article PubMed Google Scholar
Simon, M. A., Bowtell, D. D. & Rubin, G. M. Structure and activity of the sevenless protein: a protein tyrosine kinase receptor required for photoreceptor development in Drosophila. Proc. Natl Acad. Sci. USA 86, 8333–8337 (1989).
Article ADS PubMed PubMed Central Google Scholar
Sparrow, L. G. et al. The disulfide bonds in the C-terminal domains of the human insulin receptor ectodomain. J. Biol. Chem. 272, 29460–29467 (1997).
Article PubMed MATH Google Scholar
Schaffer, L. & Ljungqvist, L. Identification of a disulfide bridge connecting the alpha-subunits of the extracellular domain of the insulin receptor. Biochem. Biophys. Res. Commun. 189, 650–653 (1992).
Article PubMed MATH Google Scholar
Lemmon, M. A. & Schlessinger, J. Cell signaling by receptor tyrosine kinases. Cell 141, 1117–1134 (2010).
Article PubMed PubMed Central Google Scholar
Trenker, R. & Jura, N. Receptor tyrosine kinase activation: from the ligand perspective. Curr. Opin. Cell Biol. 63, 174–185 (2020).
Article PubMed PubMed Central MATH Google Scholar
Basler, K., Christen, B. & Hafen, E. Ligand-independent activation of the sevenless receptor tyrosine kinase changes the fate of cells in the developing Drosophila eye. Cell 64, 1069–1081 (1991).
Article PubMed Google Scholar
Luca, V. C. et al. Structural biology. Structural basis for Notch1 engagement of delta-like 4. Science 347, 847–853 (2015).
Article ADS PubMed PubMed Central MATH Google Scholar
Li, T. et al. Structural basis for ligand reception by anaplastic lymphoma kinase. Nature 600, 148–152 (2021).
Article ADS PubMed PubMed Central MATH Google Scholar
Bao, R. & Friedrich, M. Fast co-evolution of sevenless and bride of sevenless in endopterygote insects. Dev. Genes Evol. 218, 215–220 (2008).
Article PubMed MATH Google Scholar
Varadi, M. et al. AlphaFold protein structure database: massively expanding the structural coverage of protein-sequence space with high-accuracy models. Nucleic Acids Res. 50, D439–D444 (2022).
Article PubMed MATH Google Scholar
Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021).
Article ADS PubMed PubMed Central MATH Google Scholar
Evans, R. et al. Protein complex prediction with AlphaFold-multimer. bioRxiv https://doi.org/10.1101/2021.10.04.463034 (2022).
Abramson, J. et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature 630, 493–500 (2024).
Article ADS PubMed PubMed Central MATH Google Scholar
Holm, L. Dali server: structural unification of protein families. Nucleic Acids Res. 50, W210–W215 (2022).
Article PubMed PubMed Central MATH Google Scholar
Holm, L., Laiho, A., Toronen, P. & Salgado, M. DALI shines a light on remote homologs: one hundred discoveries. Protein Sci. 32, e4519 (2023).
Article PubMed PubMed Central Google Scholar
James, E. I., Murphree, T. A., Vorauer, C., Engen, J. R. & Guttman, M. Advances in hydrogen/deuterium exchange mass spectrometry and the pursuit of challenging biological systems. Chem. Rev. 122, 7562–7623 (2022).
Article PubMed Google Scholar
Houde, D., Berkowitz, S. A. & Engen, J. R. The utility of hydrogen/deuterium exchange mass spectrometry in biopharmaceutical comparability studies. J. Pharm. Sci. 100, 2071–2086 (2011).
Article PubMed Google Scholar
Chin, L. P., Soo, R. A., Soong, R. & Ou, S. H. Targeting ROS1 with anaplastic lymphoma kinase inhibitors: a promising therapeutic strategy for a newly defined molecular subset of non-small-cell lung cancer. J. Thorac. Oncol. 7, 1625–1630 (2012).
Article PubMed Google Scholar
Pal, P. & Khan, Z. ROS1 [corrected]. J. Clin. Pathol. 70, 1001–1009 (2017).
Article PubMed Google Scholar
Drilon, A. et al. ROS1-dependent cancers - biology, diagnostics and therapeutics. Nat. Rev. Clin. Oncol. 18, 35–55 (2021).
Article PubMed Google Scholar
Englund, C. et al. Jeb signals through the Alk receptor tyrosine kinase to drive visceral muscle fusion. Nature 425, 512–516 (2003).
Article ADS PubMed MATH Google Scholar
Lee, H. H., Norris, A., Weiss, J. B. & Frasch, M. Jelly belly protein activates the receptor tyrosine kinase Alk to specify visceral muscle pioneers. Nature 425, 507–512 (2003).
Article ADS PubMed Google Scholar
Ishihara, T. et al. HEN-1, a secretory protein with an LDL receptor motif, regulates sensory integration and learning in Caenorhabditis elegans. Cell 109, 639–649 (2002).
Article PubMed MATH Google Scholar
Fadeev, A. et al. ALKALs are in vivo ligands for ALK family receptor tyrosine kinases in the neural crest and derived cells. Proc. Natl Acad. Sci. USA 115, E630–E638 (2018).
Article PubMed PubMed Central MATH Google Scholar
Kiyozumi, D. et al. NELL2-mediated lumicrine signaling through OVCH2 is required for male fertility. Science 368, 1132–1135 (2020).
Article ADS PubMed PubMed Central MATH Google Scholar
Liu, C. et al. Ribonuclease 7-driven activation of ROS1 is a potential therapeutic target in hepatocellular carcinoma. J. Hepatol. 74, 907–918 (2021).
Article PubMed MATH Google Scholar
Hart, A. C., Kramer, H. & Zipursky, S. L. Extracellular domain of the boss transmembrane ligand acts as an antagonist of the sev receptor. Nature 361, 732–736 (1993).
Article ADS PubMed Google Scholar
Ferguson, K. M. Structure-based view of epidermal growth factor receptor regulation. Annu Rev. Biophys. 37, 353–373 (2008).
Article PubMed PubMed Central MATH Google Scholar
Janssen, B. J. et al. Structures of complement component C3 provide insights into the function and evolution of immunity. Nature 437, 505–511 (2005).
Article ADS PubMed MATH Google Scholar
Ames, R. S. et al. Molecular cloning and characterization of the human anaphylatoxin C3a receptor. J. Biol. Chem. 271, 20231–20234 (1996).
Article PubMed MATH Google Scholar
Wang, Y. et al. Revealing the signaling of complement receptors C3aR and C5aR1 by anaphylatoxins. Nat. Chem. Biol. 19, 1351–1360 (2023).
Article PubMed MATH Google Scholar
Kohyama-Koganeya, A., Kim, Y. J., Miura, M. & Hirabayashi, Y. A Drosophila orphan G protein-coupled receptor BOSS functions as a glucose-responding receptor: loss of boss causes abnormal energy metabolism. Proc. Natl Acad. Sci. USA 105, 15328–15333 (2008).
Article ADS PubMed PubMed Central Google Scholar
Cerutti, G. et al. Structures and pH-dependent dimerization of the sevenless receptor tyrosine kinase. Mol. Cell 84, 4677–4690.e6 (2024).
Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat. Methods 14, 290–296 (2017).
Article PubMed Google Scholar
Punjani, A., Zhang, H. & Fleet, D. J. Non-uniform refinement: adaptive regularization improves single-particle cryo-EM reconstruction. Nat. Methods 17, 1214–1221 (2020).
Article PubMed Google Scholar
Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr D. Biol. Crystallogr 60, 2126–2132 (2004).
Article ADS PubMed MATH Google Scholar
Morin, A. et al. Collaboration gets the most out of software. Elife 2, e01456 (2013).
Article ADS PubMed PubMed Central MATH Google Scholar
Liebschner, D. et al. Macromolecular structure determination using X-rays, neutrons and electrons: recent developments in phenix. Acta Crystallogr. D. Struct. Biol. 75, 861–877 (2019).
Article ADS PubMed PubMed Central MATH Google Scholar
Mirdita, M. et al. ColabFold: making protein folding accessible to all. Nat. Methods 19, 679–682 (2022).
Article PubMed PubMed Central MATH Google Scholar
Zhang, X., Gureasko, J., Shen, K., Cole, P. A. & Kuriyan, J. An allosteric mechanism for activation of the kinase domain of epidermal growth factor receptor. Cell 125, 1137–1149 (2006).
Article PubMed Google Scholar
Masson, G. R. et al. Recommendations for performing, interpreting and reporting hydrogen deuterium exchange mass spectrometry (HDX-MS) experiments. Nat. Methods 16, 595–602 (2019).
Article PubMed PubMed Central MATH Google Scholar
Monn, J. A. et al. Synthesis and pharmacological characterization of C4-(Thiotriazolyl)-substituted-2-aminobicyclo[3.1.0]hexane-2,6-dicarboxylates. Identification of (1R,2S,4R,5R,6R)-2-Amino-4-(1H-1,2,4-triazol-3-ylsulfanyl)bicyclo[3.1.0]hexane-2,6-dicarboxylic acid (LY2812223), a highly potent, functionally selective mGlu2 receptor agonist. J. Med Chem. 58, 7526–7548 (2015).
Article ADS PubMed Google Scholar
Du, J. et al. Structures of human mGlu2 and mGlu7 homo- and heterodimers. Nature 594, 589–593 (2021).
Article ADS PubMed MATH Google Scholar

Download references

Acknowledgements

We thank Claudio Alarcon, Yansheng Liu, Mark Lemmon, Kate Ferguson and their laboratories for valuable discussions. We additionally thank Yuhong Zuo and Long Han for their expertize in guiding the CryoEM processing. This work was supported by the NIH grants RM1GM149406 (D.E.K.) and NIH R03-CA259881 (Y.T.).

Author information

These authors contributed equally: Jianan Zhang, Yuko Tsutsui, Hengyi Li.

Authors and Affiliations

Department of Pharmacology, Yale University School of Medicine, New Haven, CT, 06520, USA
Jianan Zhang, Yuko Tsutsui, Hengyi Li, Tongqing Li, Salma Laraki, Sofia Alarcon-Frias, Steven E. Stayrook & Daryl E. Klein
Yale Cancer Biology Institute, Yale University, West Haven, CT, 06516, USA
Jianan Zhang, Yuko Tsutsui, Hengyi Li, Tongqing Li, Yueyue Wang, Salma Laraki, Sofia Alarcon-Frias, Steven E. Stayrook & Daryl E. Klein

Authors

Jianan Zhang
View author publications
Search author on:PubMed Google Scholar
Yuko Tsutsui
View author publications
Search author on:PubMed Google Scholar
Hengyi Li
View author publications
Search author on:PubMed Google Scholar
Tongqing Li
View author publications
Search author on:PubMed Google Scholar
Yueyue Wang
View author publications
Search author on:PubMed Google Scholar
Salma Laraki
View author publications
Search author on:PubMed Google Scholar
Sofia Alarcon-Frias
View author publications
Search author on:PubMed Google Scholar
Steven E. Stayrook
View author publications
Search author on:PubMed Google Scholar
Daryl E. Klein
View author publications
Search author on:PubMed Google Scholar

Contributions

D.E.K. designed the overall project. J.Z., Y.T. and D.E.K. wrote the manuscript with input from all authors. J.Z. and D.E.K. analyzed the structure assisted by S.E.S. J.Z. generated all materials and performed all the experiments assisted by H.L., T.L., and Y.W. Y.T. designed, conducted, and analyzed the HDX experiment. H.L. carried out the cellular activation assays assisted by J.Z., S.L. and S.A.F.

Corresponding author

Correspondence to Daryl E. Klein.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Luba Tchertanov and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Reporting Summary

Transparent Peer Review file

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Zhang, J., Tsutsui, Y., Li, H. et al. Structural basis for the interaction between the Drosophila RTK Sevenless (dROS1) and the GPCR BOSS. Nat Commun 16, 808 (2025). https://doi.org/10.1038/s41467-025-55943-6

Download citation

Received: 25 April 2024
Accepted: 03 January 2025
Published: 18 January 2025
Version of record: 18 January 2025
DOI: https://doi.org/10.1038/s41467-025-55943-6