The Pseudo-Natural Product Rhonin Targets RHOGDI

Abstract

For the discovery of novel chemical matter generally endowed with bioactivity, strategies may be particularly efficient that combine previous insight about biological relevance, e.g., natural product (NP) structure, with methods that enable efficient coverage of chemical space, such as fragment-based design. We describe the de novo combination of different 5-membered NP-derived N-heteroatom fragments to structurally unprecedented "pseudo-natural products" in an efficient complexity-generating and enantioselective one-pot synthesis sequence. The pseudo-NPs inherit characteristic elements of NP structure but occupy areas of chemical space not covered by NP-derived chemotypes, and may have novel biological targets. Investigation of the pseudo-NPs in unbiased phenotypic assays and target identification led to the discovery of the first small-molecule ligand of the RHO GDP-dissociation inhibitor 1 (RHOGDI1), termed Rhonin. Rhonin inhibits the binding of the RHOGDI1 chaperone to GDP-bound RHO GTPases and alters the subcellular localization of RHO GTPases.

Keywords: Inhibitors; Liposomes; Osteogenesis; Proteins; Pseudo-Natural Products; RHOGDI; Small Molecules.

Figure 1

Design of a pseudo‐NP collection. a) Representative natural products embodying 5‐membered N‐heterocycles. b) Tandem catalysis sequence for the synthesis of a pseudo‐NP collection containing 5‐membered N‐heterocycles in different connectivities.

Figure 2

Synthesis of pseudo‐NPs that occupy a distinct portion of chemical space. a) Reaction conditions: A) 1 a–d (3 equiv), DBU (0.5 equiv), THF, rt; B) 1 a–d (1.5 equiv), Et₃N (1.5 equiv), DCM, rt; C) 1 e (1.5 equiv), DBU (0.5 equiv), DCM, rt. For 7 a, the yield represents the epimeric mixture of the phenylethyl ketone. ODA: Osteoblast differentiation assay. RGA: Reporter gene assay. All ODA and RGA data are mean values of three independent experiments (n=3). b) NP‐likeness score comparison of NPs represented in ChEMBL (dashed curve), Drugbank (dotted curve) and succinimide‐pyrroline pseudo‐NPs (solid curve). c) PMI plot for succinimide‐pyrroline pseudo‐NPs. The average of PMI coordinate distribution is shown by a cross. d) ALogP vs MW plot of succinimide‐pyrroline pseudo‐NPs.

Figure 3

Compound 7 a inhibits Hh‐induced osteogenesis. a) Structure of 7 a. b) Hh‐induced osteogenesis in C3H/10T1/2 cells. Cells were treated with 1.5 μm purmorphamine and compound 7 a for 96 h prior to detection of alkaline phosphatase activity (mean±SD, n=3). c) C3H/10T1/2 cells were treated with purmorphamine (1.5 μm) and of 7 a or DMSO for 96 h prior to detection of the expression of Ptch1, Gli1, Ap3d1 and Gapdh by means of RT‐qPCR. (mean± SD, n=3). d) Detection of SMO in cilia in NIH/3T3 cells. Blue: nuclei; red: SMO; green: acetylated tubulin. Insets: representative single cilia. Scale bar: 10 μm.

Figure 4

Compound 7 a is a RHOGDI1 inhibitor. a) Structure of the affinity probes 8 and 9. b) Affinity‐based enrichment of RHOGDI1 from NIH/3T3 lysates by probe 8 as compared to probe 9 and detection using a RHOGDI1 antibody. c) Competition pulldown was performed as in b in presence of S10a as a competitor. d) Structure of the fluorescent derivative 10. e) Binding of derivative 10 to RHOGDI1‐3. K _D (RHOGDI1Δ15): 8.51 μm, K _D (RHOGDI1Δ25): 3.09 μm; K _D (RHOGDI2): 9.08 μm; K _D (RHOGDI3): 11.45 μm. Fluorescence polarization measurements using 10 and RHOGDI1‐3. Representative data (mean values±SD, n=3). f) Displacement of prenylated GDP‐bound RAC1 from liposomes by GST‐RHOGDI1 in the presence or absence of 50 μm 7 a or inactive derivative 7 d as determined using a liposome sedimentation assay. Representative data (n=3). P: pellet; S: supernatant. g) Competition of derivative 10 with RAC1. Fluorescence polarization measurements after adding 2 μM prenylated RAC1 or non‐prenylated RAC1 to 2 μm compound 10 and 5 μm RHOGDI1. Representative data (n=3). h) Fluorescence polarization measurements upon titration of 7 a into a mixture of 5 μm FITC‐labelled GerGer‐Rab1 peptide and 50 μm RHOGDI1.

Figure 5

Rhonin (7 l) inhibits osteogenesis and binds to RHOGDI1. a) Structure of compound 7 l termed Rhonin. b) C3H/10T1/2 cells were treated with 1.5 μm purmorphamine and compound 7 l for 96 h prior to determination of alkaline phosphatase activity. Data are mean values±SD, n=3. c) C3H/10T1/2 cells were treated with purmorphamine (1.5 μm) and 7 l or DMSO for 96 h prior to detection of the expression levels of Ptch1, Gli1, Ap3d1 and Gapdh using of RT‐qPCR (mean values±SD, n=3). d) SMO binding assay upon treatment of cells with BODIPY‐cyclopamine followed by addition of 7 l, Vismodegib or DMSO and quantification of SMO‐bound BODIPY‐cyclopamine using flow cytometry. e) Ciliary localization of SMO in NIH/3T3 cells. Representative results; each data point represents the intensity value of one single cilium. Statistical significance was evaluated using an unpaired t‐test with a confidence interval of 95 % (p≤n.s.). f) Influence of 7 l on osteogenesis in presence of 1 μm or 0.1 μm of SAG and of 7 l (mean values±SD, n=3). g) Displacement of prenylated GDP‐bound RAC1 from synthetic liposomes by GST‐RHOGDI1 in the presence or absence of 50 μm 7 l as determined using a liposome sedimentation assay. Representative data (n=3). For uncropped blot see Figure S14. h) Limited proteolysis of RHOGDI1 in presence of 100 μm 7 l. Volcano plot (FDR=0.05, S₀=0.1) of the identified and quantified peptides of RHOGDI (≈95 % sequence coverage). i) Mapping of proteinase K‐protected peptides (amino acids 179–199) in the amino acid sequence of RHOGDI1. Protected lysines detected using the STPyne probe are shown in blue. j) and k) Mapping of proteinase K‐protected peptides in the structure of RHOGDI1 with the bound geranylgeranyl group (j) and a computationally predicted model of the RHOGDI1‐7 l complex (k). Red coloration: region protected from proteinase K‐mediated proteolysis in presence of compound 7 l. The structures were prepared based on the PDB entry 1HH4.

Figure 6

RHOGDI1 is a negative regulator of osteogenesis. a) and b) Influence of RHOGDI1 knockdown. a) Osteogenesis assay upon RHOGDI1 knockdown. NT: control siRNA (mean values±SD, n=3). Knockdown efficiency: 88 %. See also Figure S12a. b) Ptch1 and Gli1 expression upon RHOGDI1 knockdown in C3H/10T1/2 cells (mean values±SD, n=3). c) Influence of RHOGDI overexpression on osteogenesis (mean values±SD, n=3). See also Figure S12b. d) Detection of GTP‐bound RHO GTPases by means of G‐LISA upon treatment with 10 μm Rhonin for 24 h.—Control: lysis buffer;+Control: respective constitutively active GTPase (mean values±SD, n=3). e) and f) Influence of Rhonin (10 μM) on the total cellular levels of RHO GTPases upon treatment for 24 h detected using immunoblotting (e). Quantification of band intensities in relation to the loading control tubulin is shown in f (mean values±SD, n=3). g) Distribution of RHO GTPases in different cellular fractions upon treatment with Rhonin (10 μm). On a separate gel, calnexin and E‐cadherin were detected as markers for ER and plasma membrane, respectively. For uncropped blots see Figure S14.