The Raf/LIN-45 C-terminal distal tail segment negatively regulates signaling in Caenorhabditis elegans

Abstract

Raf protein kinases act as Ras-GTP sensing components of the ERK signal transduction pathway in animal cells, influencing cell proliferation, differentiation, and survival. In humans, somatic and germline mutations in the genes BRAF and RAF1 are associated with malignancies and developmental disorders. Recent studies shed light on the structure of activated Raf, a heterotetramer consisting of Raf and 14-3-3 dimers, and raised the possibility that a Raf C-terminal distal tail segment (DTS) regulates activation. We investigated the role of the DTS using the Caenorhabditis elegans Raf ortholog lin-45. Truncations removing the DTS strongly enhanced lin-45(S312A), a weak gain-of-function allele equivalent to RAF1 mutations found in patients with Noonan Syndrome. We genetically defined three elements of the LIN-45 DTS, which we termed the active site binding sequence (ASBS), the KTP motif, and the aromatic cluster. In the context of lin-45(S312A), the mutation of each of these elements enhanced activity. We used AlphaFold to predict DTS protein interactions for LIN-45, fly Raf, and human BRAF within the activated heterotetramer complex. We propose the following distinct functions for the LIN-45 DTS elements: (1) the ASBS binds the kinase active site as an inhibitor; (2) phosphorylation of the KTP motif modulates the DTS-kinase domain interaction; and (3) the aromatic cluster anchors the DTS in an inhibitory conformation. Human RASopathy-associated variants in BRAF affect residues of the DTS, consistent with these predictions. This work establishes that the Raf/LIN-45 DTS negatively regulates signaling in C. elegans and provides a model for its function in other Raf proteins.

Keywords: Extracellular signal regulated kinase (ERK); Raf; cell signaling; feedback phosphorylation; kinase autoinhibition.

Fig. 1.

LIN-45 protein domains and activity in C. elegans development. a) The conserved regions and phosphorylation sites are shown in the C. elegans Raf ortholog LIN-45, human BRAF, and human RAF1. Conserved region 1 (CR1) contains the Ras-binding domain (RBD) and cysteine-rich domain (CRD). Conserved region 2 (CR2) contains a 14-3-3-binding motif at phosphorylated S312 of LIN-45. Conserved region 3 (CR3) is the kinase domain. Sequences C-terminal to the kinase domain include a second 14-3-3 binding motif at phosphorylated S756 of LIN-45, followed by the distal tail segment (DTS), residues 760-813. b) At the L3 larval stage, the EGFR activity in vulval precursor cells (VPCs) is stimulated by an EGF ligand produced by the anchor cell. In wild-type development, LIN-45 activation occurs in a single VPC termed P6.p, resulting in 1° vulval fate. Notch activation in the VPCs P5.p and P7.p results in 2° vulval fate, while other VPCs adopt a non-vulval fate.

Fig. 2.

Deletion of the LIN-45 DTS enhances the LIN-45(S312A) activity. a) Adult animals carrying transgenes expressing yfp-lin-45(S312A) (left), yfp-lin-45(763stop) (middle), or yfp-lin-45(S312A, 763stop) (right). Asterisks indicate the normal vulva; arrows indicate ectopic pseudovulvae. Scale bar 50 μm. b–e) All panels are the percentage of adults displaying the Muv phenotype. Different, independent transgenic strains are listed as replicate genotypes. The number of adults scored (n) is shown in parentheses. b) Full-length LIN-45 transgenes yfp-lin-45(+) and yfp-lin-45(S312A). c) LIN-45 DTS truncations yfp-lin-45(772stop), yfp-lin-45(763stop), yfp-lin-45(S312A, 772stop), and yfp-lin-45(S312A, 763stop). The transgene yfp-lin-45(S312A, 763stop) was tested in the mutant lin-45(dx19). Groups expressing YFP-LIN-45 with truncations were compared to YFP-LIN-45(S312A) controls using the Fisher's exact test. **P-value < 0.001. d) LIN-45 transgenes with CRD mutations yfp-lin-45(R175E, T177E), yfp-lin-45(R175E, T177E, 772stop), yfp-lin-45(T177P), and yfp-lin-45(T177P, 772stop). e) Weak gain-of-function activity was tested in the lin-45(dx19) mutant for yfp-lin-45(R175E, T177E), yfp-lin-45(R175E, T177E, 772stop), yfp-lin-45(T177P), and yfp-lin-45(T177P, 772stop). Those expressing YFP-LIN-45(R175E, T177E, 772stop) were compared to YFP-LIN-45(R175E, T177E) controls using the Fisher's exact test. ns; not significant.

Fig. 3.

Negative-acting regulatory sequences within the LIN-45 DTS. a) Aligned C-terminal sequences of Raf proteins and accessions: C. elegansLIN-45 (NP_741430.3), D. melanogaster (NP_525047.1), D. rerio (NP_991307.3), X. laevis (XP_018109664.1), G. gallus isoform 2 (NP_001383912.1), and H. sapiens isoforms 2 (NP_001341538.1) and 1 (NP_004324.2). Dots in the alignment indicate gaps. Diamonds indicate stop codons introduced to truncate LIN-45. Asterisks indicate phospho-sites in human BRAF. Aromatic residues near the C-terminus are highlighted. b) Predicted structure of LIN-45 regions 756–813 from AlphaFold database model AF-Q07292. R group sidechains are shown for residues of the ASBS, KTP motif, and aromatic cluster. c–e) All panels are the percentage of adults displaying the Muv phenotype. Different, independent transgenic strains are listed as replicate genotypes. The number of adults scored (n) is shown in parentheses. c) DTS truncations in LIN-45(S312A) transgenes yfp-lin-45(S312A, 781stop), yfp-lin-45(S312A, 790stop), yfp-lin-45(S312A, 791stop), and yfp-lin-45(S312A, 795stop). d) Mutations in LIN-45(S312A) transgenes yfp-lin-45(S312A, Y783A, I784A), yfp-lin-45(S312A, del781–784), and yfp-lin-45(S312A, T791A). e) Weak gain-of-function activity was tested in the lin-45(dx19) mutant for yfp-lin-45(S312A), yfp-lin-45(S312A, Y783A, I784A), yfp-lin-45(S312A, del781–784), and yfp-lin-45(S312A, T791A). All groups were compared to the lin-45(dx19); yfp-lin-45(S312A) control using the Fisher's exact test. **P-value < 0.001.

Fig. 4.

A C-terminal aromatic cluster is critical for LIN-45 negative regulation. All panels are the percentage of adults displaying the Muv phenotype. Different, independent transgenic strains are listed as replicate genotypes. The number of adults scored (n) is shown in parentheses. a) DTS truncations in LIN-45(S312A) transgenes yfp-lin-45(S312A, 808stop), yfp-lin-45(S312A, 811stop). b) Mutations in LIN-45(S312A) transgenes yfp-lin-45(S312A, Y810A). The transgene yfp-lin-45(S312A, Y810A) was tested in the mutant lin-45(dx19). c) Tests of dpy-23/AP2M1 and gap-1/GAP mutants. Transgenes yfp-lin-45(+), yfp-lin-45(S312A), yfp-lin-45(S312A, 790stop) were introduced into dpy-23(e840) and gap-1(ga133) mutants. For all groups expressing the same transgene, dpy-23(e840) and gap-1(ga133) mutants were compared to wild-type controls using the Fisher's exact test. **P-value < 0.0001. d) Mutation of the RBD was tested in yfp-lin-45(Q95A,R119A,S312A,Y810A). e) The transgene yfp-lin-45(+), yfp-lin-45(S312A, 795stop) was introduced into ksr-1(ok786) mutants.

Fig. 5.

Predicted trans-interaction of the LIN-45 DTS and kinase domain. All panels show two orthogonal views of models predicted by AlphaFold for the heterotetramer containing two protomers of LIN-45 residues 470–813 with pS756 (LIN-45 A and B) and two protomers of PAR-5. For the LIN-45 DTS, panels a and b show five ensemble models superimposed and depicted as ribbons with the highest-confidence model in magenta. For clarity, PAR-5 is trimmed to exclude the C-terminal tail (residues 230–248). To indicate the location of the ATP binding pocket, ATP is superimposed on the kinase domain models. a) LIN-45 with phosphorylation at the activation loop (pT626, pT629) and unphosphorylated DTS. b) LIN-45 with phosphorylation at the activation loop (pT626, pT629) and the DTS (pT791). c) Higher-magnification views of model shown in panel b, with only the highest-confidence model depicted in ribbon (magenta). The locations of residues in the ASBS (W781), KTP motif (pT791), and aromatic cluster (Y810) are indicated.

Fig. 6.

Comparison of the predicted trans-interactions for the DTS and kinase domain of human, fly, and C. elegans Rafs. a) The DTS of human BRAF is encoded within exon 18 for isoform 1 or exons 18–19 for isoform 2. b–d) Panels show models predicted by AlphaFold for the heterotetramer containing two protomers of Raf kinase domain (magenta and blue), phosphorylations at the 14-3-3 binding motif and KTP motif of the DTS, and two protomers of 14-3-3 zeta (green). Top panels show five ensemble models superimposed and depicted as ribbons with the highest-confidence model in magenta. Bottom panels show the residue sidechains in the ASBS (red), KTP motif (blue), and aromatic cluster (yellow) depicted in spheres. The location of the ATP-binding pocket is indicated by superposition of ATP on the kinase domain. Models: b) Human BRAF isoform 1 (pS750, pT753). c) Human BRAF isoform 2 (pS750, pT753). d) D. melanogaster Raf (pS666, pT669) with 14-3-3zeta. e) C. elegansLIN-45 (pT791) with the 14-3-3 ortholog PAR-5. f) Variants reported in the ClinVar database that affected residues 723–766 of BRAF isoform 1 were non-synonomous and from patients with a RASopathy, unspecified (purple), or Cardiofaciocutaneous syndrome (orange). Positions marked with asterisks indicate residues also altered in three or more different tumor samples in the cBioportal database. Highlighted residues indicate the 14-3-3 binding site, ASBS, KTP motif, and aromatic cluster.