Discovery and Folding Dynamics of a Fused Bicyclic Cysteine Knot Undecapeptide from the Marine Sponge Halichondria bowerbanki

We describe the discovery and structure of an undecapeptide natural product from a marine sponge, termed halichondamide A, that is morphed into a fused bicyclic ring topology via two disulfide bonds. Molecular dynamics simulations allow us to posit that the installation of one disulfide bond biases the intermediate peptide conformation and predisposes the formation of the second disulfide bond. The natural product was found to be mildly cytotoxic against liver and breast cancer cell lines.


General Experimental Procedures
All chemicals, solvents, and media components were purchased from Sigma-Aldrich, Fisher Scientific, and VWR, and used without further purification.Optical rotation, circular dichroism, and UV spectra were measured on a JASCO J-815 spectropolarimeter (JASCO).One-dimensional (1D) and two-dimensional (2D) NMR spectra were acquired on Bruker AVIII-HD 800 and AVIII-HD 700 MHz NMR.The 1 H and 13 C{ 1 H} NMR chemical shifts were referenced to the solvent peaks for DMSO-d 6 at δ H 2.50 ppm and δ C 39.52 ppm.The chemical shift (δ) values are given in parts per million (ppm), and the coupling constants (J values) are in Hz.
High resolution electrospray ionization mass data were recorded on a 1290 Infinity II ultra-performance liquid chromatography (UPLC, Agilent Technologies) coupled to a Bruker ImpactII ultra-high-resolution Q-ToF mass spectrometer equipped with an electron spray ionization (ESI) source (Bruker Daltonics).A Kinetex C18 reverse phase UHPLC column (50 x 2.1 mm, 1.7 µm) was used for chromatographic separation.Data were acquired in positive ionization mode with m/z 50-2000 Da.Low resolution electrospray ionization mass data were recorded on a 1260 Infinity high performance liquid chromatography (HPLC, Agilent Technologies) coupled to a Bruker amaZon mass spectrometer equipped with an electron spray ionization (ESI) source (Bruker Daltonics).A Poroshell 120 EC-C18 reverse phase HPLC column (100 x 4.6 mm, 5.0 µm) was used for chromatographic separation.Data were acquired in negative ionization mode with m/z 50-2000 Da.Open column chromatography (CC) was performed over a C18 SPE cartridge (Phenomenex).Semi-preparative HPLC was performed on an Agilent 1260 Infinity II HPLC system equipped with a VWD detector, using a Luna C18 reverse phase column (250*10 mm, 5 µm).All solvents used in CC and HPLC were analytical (VWR) and HPLC grades (Fisher and Sigma-Aldrich).Synthetic peptide SCCPWIIWCCL was obtained from GenScript.

Sponge Material
The marine sponge Halichondria bowerbanki was collected from kelp forest habitat at a depth of 5 m at Coal Oil Point, Santa Barbara, California, USA.The species was identified by digesting subsamples of sponge tissue in bleach and examining spicules, and identification was confirmed with phylogenetic analysis of DNA sequences.Sequences are vouchered in GenBank under the numbers PP729067 and PP734164.

Reduction and Iodoacetamide Derivatization
To a solution of 0.5 mg of compound 1 in 1.0 mL water in a 4 mL vial was added 100 μL of 18 mM tris(2carboxyethyl)phosphine (TCEP).The mixture was stirred at room temperature and reduction of 1 was monitored by LC-MS until the disappearance of the starting material.Half of the reaction solution was transferred to a separate 4 mL vial and dried by lyophilization for further Marfey's analysis (see below).To the rest of the reaction mixture was added 20 μL of 1.0 M iodoacetamide and 50 μL of triethylamine (0.726 g/mL).This mixture was stirred at room temperature and monitored by LC-MS until the disappearance of the starting material.Compound 1, without reduction, was treated with iodoacetamide and trimethylamine in the same condition as a control.

Marfey's Analysis
The TCEP-reduced 1 (250 g) was hydrolyzed using 6 N HCl (1.0 mL) at 110 ˚C in a 4 mL vial.100 L reaction volume was taken at different reaction times (2 h, 3 h, 4 h, 5 h, 6 h, 7 h, 8 h, 9 h, and 24 h).The reaction aliquots were dried in vacuo and resuspended in dH 2 O (100 L), to which was added 10% w/v 1-fluoro-2-4dinitrophenyl-5-L-alanine amide (L-FDAA) in acetone (33 L) and 1 N NaHCO 3 in dH 2 O (50 L).These reactions were incubated at 40 ˚C for 1 h.The derivatization reactions were quenched by adding 1 N HCl (50 L), dried in vacuo.Residues were dissolved in MeOH (0.2 mL) prior to analysis by LC-MS.Amino acid standards were prepared and analyzed in an identical manner.
MestReNova (version 6.0) was used to perform NMR chemical shift assignment, and CYANA (version 2.1) was used for the structural calculation. 1 1H-1 H distance restraints, automatically assigned by CYANA and confirmed manually, were used as restraints in the structural calculation.A total of 200 structures were calculated, after seven iterations, using the software's simulated annealing protocol.Ten structures with the lowest energy were chosen to produce the final ensemble.Final structures were checked in MolProbity.Images of structures were generated with PyMOL (version 2.5.5).

Cell Culture
Cells were cultured in their respective media as follows: Hep-G2 cells in MEM; HuH-7 cells in low-glucose DMEM; and MDA-MB-231 and Vero cells in high-glucose DMEM.Media were supplemented with 10% fetal bovine serum, 1% Pen/Strep, and 1% L-glutamine (excluding media for MDA-MB-231 and Vero cells).

MTT Assay
Following the method of Yadava et al. 2 with minor modifications, cells, at a density of 4.5 x 10 3 cells/well and a volume of 100 μL, were seeded in 96-well tissue culture plates and allowed to attach for 24 h.Thereafter, the culture medium was removed, and the cells were treated for 72 h with 100 μL of varying concentrations (1 -200 μM) of molecule 1, dissolved in dimethyl sulfoxide but diluted with the respective culture media.Subsequently, 10 μL of MTT reagent (5 mg/mL) was added to the culture medium and the plates were incubated for 3 h, after which the MTT reagent/medium was carefully aspirated.Finally, formazan crystals formed were dissolved with molecular grade DMSO (100 μL/well) and absorbance was measured at 570 nm using multimode plate reader (Tecan Infinite M200 Pro, Männedorf, Switzerland).

Molecular Dynamics Simulations
All atom molecular dynamics (MD) simulations were performed explicit solvent using the solution NMR structure of halichondamide A (1, model 1 from PDB ID 9BHN).Simulations were performed on the wild-type halichondamide A (all cysteines oxidized), reduced C2/C9 halichondamide A (reduced disulfide bond between Cys2 and Cys9, while Cys3/Cys10 was oxidized), and reduced C3/C10 halichondamide A (reduced disulfide bond between Cys3 and Cys10, while Cys2/Cys9 was oxidized).For reduced structures, target disulfide bonds were broken in PyMOL using the unbond command.Inputs for MD were separately prepared in CHARMM-GUI Solution Builder Module in the Input Generator using the CHARMM36 force field and TPI3P water model. 3The halichondamide A peptides were solvated in an aqueous environment using default parameters in CHARMM-GUI, and sodium and chloride ions were added to the aqueous system at concentrations that naturally occur in the human body (150 mM).Default CHARMM-GUI periodic boundary conditions were used, and the thermodynamic ensemble was nPT.Temperature was set constant at 298.15 K and maintained using the Nosé-Hoover coupling method with a tau-t of 1 psec.An integration time step of 2 fsec was used with coordinates output every 10 psec.The LINCS algorithm was used to constrain H-bonds.An isotropic Parrinello-Rahman method with a tau-p of 5 psec and a compressibility of 4.5 × 10 -5 bar -1 was used for pressure coupling.Short range interactions were treated with a Verlet cutoff scheme with 10 Å electrostatic and van der Waals cutoffs and long-range electrostatics were treated using the particle mesh Ewald method.MD simulations were performed in GROMACS v2021.5 with NVIDIA Tesla V100 GPUs on the Phoenix Cluster provided to us by the Partnership for an Advanced Computing Environment at the Georgia Institute of Technology. 4 Simulations were run in independent duplicate replicates for a total of 1000 nanoseconds each.After the simulations concluded, the simulation trajectories were visualized and analyzed through GROMACS and the Visual Molecular Dynamics (VMD) program v1.9.3. 5Backbone RMSD for peptides were plotted in VMD.

Figure S10 .
Figure S10.EICs for the [M+2H] 2+ ions corresponding to compound 1 (top) and possible derivatives of 1 with different number of acetamide unit additions.Compound 1 was treated with iodoacetamide and the addition of different number of acetamide units was queried.

Figure S13 .
Figure S13.Comparison of retention time of TCEP-reduced halichondamide A (1) and synthetic SCCPWIIWCCL peptide.UV absorbance chromatograms were recorded at 280 nm wavelength.

Figure S16 .
Figure S16.Marfey's analysis to determine the absolute configuration of the Pro residue in TCEP-reduced 1. Extracted ion chromatograms (EICs) demonstrating the retention time of the DAA-derivatized Pro residue resulting from the acid hydrolysis of TCEP-reduced 1 (top), retention time of DAA-derivatized standard of D-Pro (middle), and the retention time of the similarly derivatized standard of L-Pro (bottom).Separation was achieved using the Agilent Poroshell EC-C18 (100×4.6 mm, 2.7 m) column.Mass spectrometry data were acquired in the negative ionization mode.

Figure S17 .
Figure S17.Marfey's analysis to determine the absolute configuration of the Ser residue in TCEP-reduced 1. EICs demonstrating the retention time of the DAA-derivatized Ser residue resulting from the acid hydrolysis of TCEP-reduced 1 (top), retention time of DAA-derivatized standard of L-Ser (middle), and the retention time of the similarly derivatized standard of D-Ser (bottom).Separation was achieved using the Agilent Poroshell EC-C18 (100×4.6 mm, 2.7 m) column.Mass spectrometry data were acquired in the negative ionization mode.

Figure S18 . 1 .
Figure S18.Marfey's analysis to determine the absolute configurations of the Leu and Ile residues in TCEPreduced 1. From top to bottom-EICs demonstrating retention time of DAA-derivatized Leu and Ile residues obtained by acid hydrolysis of TCEP-reduced 1, DAA-derivatized standard for L-Leu, DAA-derivatized standard for D-Leu, DAA-derivatized standard for L-Ile, DAA-derivatized standard for D-Ile, DAA-derivatized standard for L-allo-Ile, and DAA-derivatized standard for D-allo-Ile.By retention time matching, the Leu and Ile residues in TCEP-reduced 1 were determined to be L-Leu and L-Ile, respectively.Chromatographic separation was achieved using the Cosmosil Cholester (250×4.6 mm, 5 m) column.Mass spectrometry data were acquired in the negative ionization mode.

Figure S19 .
Figure S19.Marfey's analysis to determine the absolute configuration of the Trp residue in TCEP-reduced 1. EICs demonstrating the retention time of the DAA-derivatized Trp residue resulting from the acid hydrolysis of TCEP-reduced 1 (top), retention time of DAA-derivatized standard of L-Trp (middle), and the retention time of the similarly derivatized standard of D-Trp (bottom).Separation was achieved using the Agilent Poroshell EC-C18 (100×4.6 mm, 2.7 m) column.Mass spectrometry data were acquired in the negative ionization mode.

Figure S20 .
Figure S20.Marfey's analysis to determine the absolute configuration of the Cys residue in TCEP-reduced 1. EICs demonstrating retention time of DAA-derivatized standard for L-Cys spiked with the derivatized acid hydrolysate of TCEP-reduced 1 (top), and DAA-derivatized standard for D-Cys spiked with the derivatized acid hydrolysate of TCEP-reduced 1 (bottom).By retention time matching, the Cys residue in TCEP-reduced 1 was determined to be L-Cys.Chromatographic separation was achieved using the Cosmosil Cholester (250×4.6 mm, 5 m) column.Mass spectrometry data were acquired in the negative ionization mode.

Figure S21 .
Figure S21.Representative MD simulation derived principal component analysis (PCA) to study the folding dynamics of oxidized and reduced halichondamide A (1). Graphs are prepared form the collective motion of oxidized (green) and reduced halichondamide A (1, red or cyan) using projections of MD trajectories on two eigenvectors corresponding to the first two principal components.Example conformations are shown to represent different conformations observed in the simulations.

Table S2 :
NMR and refinement statistics for halichondamide A(1)

Table S3 :
Sequences and linkages of natural product peptides with disulfide bonds.MS 2 fragmentation spectra for four different molecular species detected in the organic extract of the marine sponge Halichondria bowerbanki.The spectrum on the top corresponds to 1. Fragment ions at m/z 86, m/z 159, m/z 136, among others, correspond to amino acid immonium ions.In all these spectra, neutral losses corresponding to proteogenic amino acids can be annotated.