Gigobolins A-C, New Ophiobolins with Anticancer Activity from the Phytopathogenic Fungus .

Results and Discussion
The EtOAc crude extracts from D. gigantea culture filtrate were submitted to sequential
fractionation steps
involving both crystallization and chromatographic methods yielding
the new gigobolins A–C (1–3) along with six known ophiobolins (4–9) as fully described in the Experimental Section. The six known compounds were identified as OpA (4)22, 6-epi-OpA (5), 3-anhydro-6-epi-OpA (6), ophiobolin B (OpB, 7), ophiobolin I (OpI, 8), and maydispenoid A (9) by comparison of their MS and
NMR data with those of the literature. The planar structures of the
new gigobolins 1–3 were established
by analyzing the 2D NMR spectra in combination with HRESIMS, whereas
the relative stereochemistry was assigned by the examination of proton
and carbon resonance values and correlations in the NOESY spectrum.
Gigobolin A (1), as described below, is structurally
similar to 6-epi-OpA (5), specifically
sharing the identical ring ABC framework, with the only structural
difference being the tertiary alcohol at C-15 in 1. Compound 1 gave a [M + Na]+ ion peak at m/z 439.2466 in HRESIMS corresponding to the molecular
formula C25H36O5Na requiring eight
indices of hydrogen deficiency. Both 1H and 13C NMR spectra of 1 strongly resembled those of the co-occurring
6-epi-OpA (5). In particular, similar
to 6-epi-OpA (5), the proton spectrum
of gigobolin A (1) displayed the characteristic signals
of an unsaturated aldehydic group (δH 9.19, s, 1H,
H-21), an olefinic proton (δH 6.84, dd, J = 6.6, 2.5 Hz, 1H, H-8), four sp3 methines [δH 4.50 (td, J = 8.8, 4.1 Hz, 1H, H-17), δH 3.34 (d, J = 10.6 Hz, 1H, H-6), δH 2.64 (app dd, J = 15.0, 3.5 Hz, 1H, H-10);
δH 2.14 (m, 1H, H-2)], and an isolated methylene
group at δH 3.07 (d, J = 16.6 Hz,
1H, H-4α) and at δH 2.43 (dd, J = 16.6, 1.5 Hz, 1H, H-4β). The presence of an additional alcohol
group in 5, as suggested by its molecular formula, was
corroborated by the distinct, downfield methyl singlet (δH 1.37, s, H3-23) and the absence of the methyl
doublet (δH 1.02, d, J = 6.7 Hz,
3H, H3-23) observed on the spiro tetrahydrofuran ring of
6-epi-OpA (5). Furthermore, the 13C NMR spectrum of 1 showed an oxygenated quaternary
carbon signal resonating at δC 80.6 (C-15, C) replacing
the methine carbon at δC 35.8 (C-15, CH) observed
in 5. As expected, this quaternary carbon showed a correlation
with H3-23 in the HMBC spectrum (Figure
a).
The relative configuration of gigobolin A (1) was
elucidated by analyzing both NOESY correlations and carbon and proton
resonances. The trans ring junctions of the 5/8/5
carbon framework in 1 was inferred by correlations in
the NOESY spectrum between H-6 (δH 3.34) and H-10
(δH 2.64), and between Me-22 (δH 0.89) and H-2 (δH 2.14). The latter resonance also
showed a cross peak with Me-20 thus indicating the α-orientation
of the hydroxy group at C-3 (Figure
b). The proposed C-2/C-6 trans-junction
was also supported by the resonance value of C-1 (δC 41.2) that is downfield shifted with respect to the ophiobolins
with a H-2/H-6 cis-junction (δC 35.0),
and by the proton values of H-2 and H-8 as well as of H-4β (see Table
). It has been noted
that ophiobolins exhibiting a trans C-2/C-6 junction
with H-6 in the α-configuration (such as 1) typically
show shielded H-2 and H-8 proton values by about 0.2–0.3 parts
per million (ppm). This trend diverges from that of the H-6 β-series,
where the C-4 β-proton is instead deshielded. The downfield
shift of H-9a at δH 3.75 (usually resonating at ∼δH 2.80) strongly suggests the α-orientation of the hydroxy
group at C-15, whereas the NOESY cross peak correlations of H-16a
(δH 2.31) with both H3-23 and H-17, which
in turn had a correlation with H2-13 (δH 1.57–1.48), imply the α-orientation of the side chain
at C-17 (Figure
b).
Additionally, the absolute configuration of gigobolin A (1) is proposed to be the same as that of 6-epi-OpA
(5) by comparison of their ECD curves (Figure
). Furthermore, the analysis
of HSQC and HMBC spectra aided the assignments of all proton and carbon
resonances, as reported in Table
.
Gigobolin B (2) is structurally similar
to ophiobolin
B (OpB, 7), with the differences being the trans vs cis AB-ring junction (C-6) and an apparent dehydration
of 7 at C-10/C-14 to give gigobolin B (2). The molecular formula C25H36O3Na+ of 2, derived from the sodiated ion peak
at m/z 407.2544 [M + Na]+, implies dehydration and an additional hydrogen deficiency index
compared to the molecular formula of 7 (C25H38O4). HMBC correlations (Figure
a), between H-9 and H-15 with
C-10 (δC 139.1), as well as H-13 with C-14 (δC 143.9) indicate the presence of a tetrasubstituted olefin
in the C-ring of the ophiobolins. Furthermore, only three methine
correlations are observed in the HSQC spectrum, supporting that the
methine at C-10, which is typically present in the ophiobolins, is
absent in this new congener.
The C-2/C-6 trans-junction of 2 was
evident by NOESY correlations between H-2 (δH 2.15–2.05,
m) and both Me-20 (δH 1.45, s, 3H) and Me-22 (δH 0.79, s, 3H), and between H-6 (δH 3.36–3.31,
m, 1H) and both H-1 (δH 1.78–1.73, m, 1H)
and H-9b (δH 3.11–3.07, m, 1H) (Figure
b). In addition, some significant
differences were also observed in the chemical shift of protons H-2
and H-8 which were both shifted upfield with respect to 7 that suggested a 6-epi configuration. The configuration
at C-15 was not fully established, but by analogy, it was proposed
to be the same as that in the co-occurring OpB (7). All
proton and carbon resonances are reported in Table
.
The minor ophiobolin in D. gigantea extract named gigobolin C (3) has a sodiated ion peak
of m/z 407.2552 corresponding to
the same molecular formula of C25H36O3Na+ as ophiobolin X2 (OpX) and 6-epi-ophiobolin M (Figure
). The resonances of both the carbonyl lactone
at C-21 (δC 172.0) and the methine at C-5 (δC 81.6; δH 4.98, dd, J =
7.1, 5.7, 1H) were almost superimposable with the same signals in
OpX (10) (Figure
and Table
). On the other hand, the olefin C-13 (δC 120.8)
and C-14 (δC 150.3) in 3 also aligned
with the reported resonances of carbons C-13 and C-14 for 6-epi-OpM (11) (Figure
) as shown in Table
. HMBC correlations as depicted in Figure
a also corroborated
the configuration of C-13 and C-14. Unfortunately, full stereochemical
determination was hampered by degradation that occurred in the NMR
tube during experiment acquisitions. However, based on the comparison
of NMR resonances of 3 and those of published ophiobiolin
family members containing C-21-lactone groups,
,
we
propose the relative stereochemistry at C-5, C-6, and C-2 as all cis.
We evaluated gigobolins A (1) and B (2) and maydispenoid A (9), which have not been
previously
tested for anticancer activities, in our models of GBM and breast
cancer with a focus on studying the effects of these compounds on
diverse breast cancer subtypes, including those with and without epithelial-mesenchymal
transition (EMT) features, HMLE versus HMLE-TWIST (Table
).
We included in our cell line panels GBM cell lines
T98G and U87
in addition to patient-derived glioblastoma stem cell lines, GSC 040815
and GSC 082209, grown as neurospheres that were previously analyzed
for stemness markers and characterized for consistency by histology
with the primary tumor.
,
In addition, given
OpA’s enhanced activity against mammary epithelial cells induced
to undergo EMT and thus against breast cancer cell lines with CSC
properties, we selected EMT+ and EMT– epithelial (HMLE) cell lines as models for breast cancer stem cells,
in addition to multiple breast cancer subtypes,
,
as indicated in Table
. As expected based on our previous studies, OpA (4) displayed pronounced selectivity toward growth
inhibition of GSCs over transformed GBM cells registering nanomolar
GI50 values against patient-derived GSC neurospheres (Table
). Similarly, although
with a lower level of selectivity, OpA (4) effectively
inhibited the growth of mammary epithelial cells induced to undergo
epithelial-mesenchymal transition (EMT) via Twist expression (HMLE-TWIST
vs HMLE), despite their acquired resistance to multiple chemotherapeutic
agents. It was also highly active against multiple subtypes of breast
cancer cells.
,

In general, it is well
recognized that variations in the tetrahydrofuran
part of the ophiobolin skeleton, i.e., ring D, do not have a pronounced
effect on cytotoxic potency in this family of natural products. For
example, OpA (4), with an intact tetrahydrofuran ring,
and OpB (7), with an open tetrahydrofuran ring D, are
both equally low nanomolar against CLL leukemia cells. Therefore, the reduced potency of gigobolins
A (1) and B (2) is unlikely due to the hydroxylation
at C-15 (in 1) and/or elimination of ring D (in 2). The important factor here is the trans AB-ring fusion in 1 and 2 compared to cis in 4 and 7, i.e., the configuration
at C6. Indeed, the change of this ring fusion from cis to trans, as is found in 6-epi-OpA (5), leads to a drop in cytotoxic potency by ∼1
to 2 orders of magnitude. It is unfortunate
that we have run out of material and could not complete the evaluation
of gigobolin B (2) against GBM cells, but its low micromolar
potency against patient-derived GSC cells is an interesting finding
for future research toward therapies directed toward addressing cancer
stems cells.
Maydispenoid A (9) is the only ophiobolin
family member
in which the C-21-aldehyde is closed into a dihydrofuran ring; thus,
this structure represents a new chemotype within the ophiobolin family.
Notably, this congener exhibited cytotoxicity toward GSCs with a GI50 an order of magnitude lower than that toward transformed
GBM. This compound was also active against HMLE cells, although no
selectivity toward HMLE-EMT+ cells was observed here.

Conclusion
In summary, our ongoing investigation into
the diverse chemical
profile of the phytopathogenic fungus D. gigantea led to the successful isolation and structural characterization
of three new sesterterpenoids, which we named gigobolins A–C
(1–3). Their distinctive structural
variations in the C and D rings and the oxygenation patterns compared
to known congeners further expand the structural diversity of the
ophiobolin family.
The newly identified compounds were evaluated
for their therapeutic
potential against two highly aggressive and drug-resistant cancer
types: glioblastoma (GBM) and triple-negative breast cancer. Gigobolins
A (1) and B (2), along with the known analog
maydispenoid A (9), demonstrated significant antiproliferative
activity, with gigobolin B and maydispenoid A exhibiting promising
properties against GBM stem cells (GSCs), the subpopulation responsible
for tumor recurrence and chemoresistance. For maydispenoid A, this
is the first demonstration of promising anticancer activity of a congener
possessing a C-21-dihydrofuran in lieu of the ketoaldehyde moieties.
These findings validate D. gigantea species as a useful source of bioactive ophiobolins and reinforce
the potential of this sesterterpenoid family as anticancer agents,
especially for recalcitrant tumors driven by stemness and drug resistance.
Future work will continue to focus on isolation strategies for bioactive
ophiobolins, comprehensive structure–activity relationship
(SAR) studies, and in-depth mechanistic investigations to elucidate
the specific molecular targets and modes of action, paving the way
for the clinical development of these new anti-CSC agents.

Experimental Section

General Experimental Procedures
UV spectra were measured
in MeOH on a Jasco V-530 spectrophotometer in MeOH. CD curves were
recorded in MeOH on a JASCO F815 spectropolarimeter (Jasco, Lecco,
Italy). 1D and 2D NMR spectra were acquired in CDCl3 on
a Bruker Avance III HD 400 MHz spectrometer equipped with a CryoProbe
Prodigy, and on on a DRX 600 spectrometer equipped with a three-channel
inverse (TCI) CryoProbe. Chemical shift values were reported in parts
per million (ppm) referring to CHCl3 (δH 7.26 for proton and δC 77.0 for carbon). High-resolution
mass spectra (HRESIMS) were acquired on a Q-Exactive hybrid quadrupole-orbitrap
mass spectrometer (Thermo Scientific, San Jose, CA). NMR experiments
were recorded at the ICB-NMR Service Centre. Analytical TLC was performed
on silica gel (Kieselgel 60, F254, 0.25), and the spots
were visualized by exposure to UV radiation (253 nm), or iodine vapor,
or by spraying first with 10% H2SO4 in MeOH
and then with 12% sulfuric acid and CeSO4 in water, followed
by heating.

Fungal Strain Growth
The strain of D.
gigantea was isolated in Florida from naturally infected
large crabgrass (Digitaria sanguinalis). It was stored and growth in flasks (1 L) containing a mineral-defined
medium detailed as previously reported.

Extraction and Purification of Gigobolins
The lyophilized
fungal culture filtrates (10.9 L) was suspended in distilled water
(1.5 L, final pH 4.5) and extracted with ethyl acetate (3 × 500
mL). The organic extracts were combined and evaporated under reduced
pressure, giving a brown oily residue (3.0 g). Mycelium was also extracted
with ethyl acetate (2 × 500 mL) using ultrasound to facilitate
cell disruption. The organic residues were filtered, and the solvent
was evaporated to obtain 3.0 g of blackish gum. After TLC comparison,
the two extracts were combined and loaded onto a silica gel column,
using chloroform/isopropanol (95:5, 8:2) first, and then flushed with
methanol. A total of eight fractions (F1–F8) were collected
and analyzed by TLC and 1H NMR.
Fraction F2 (2.96
g), which appeared to contain mainly OpA (4), was repeatedly
treated with n-hexane/EtOAc at −20 °C
giving 1.3 g of a white precipitate. Proton spectrum of this precipitate
confirmed the presence of pure OpA. Fraction F3 (1.3 g) was purified
via flash chromatography using a gradient of diethyl ether in petroleum
ether (7:3, 6:4, 1;1, 4:6, 3:7, 1:9) and finally with MeOH yielding
30 subfractions (F3-1–F3-30). Fractions F3-12/F3-13/F3-14 (430.0
mg) were analyzed by 1H NMR showing to contain 6-epi-OpA
(5). Fractions F3-15/F3-16/F3-17/F3-18 (160.0 mg) contained
ophiobolin I (8). Fractions F3-20/F3-21/F3-22 (20 mg)
analyzed by TLC were shown to contain ophiobolin I and a more polar
UV absorbing compound. Purification of this fraction by SPE-RP18-cartridge
in MeOH/H2O 1:1 gave 17 subfractions that after NMR analysis
were identified as the new gigobolin A (1, 3.4 mg) and
additional amount of ophiobolin I (2.0 mg). Fraction F3-7/F3-8/F3-9
(19.0 mg) contained less polar ophiobolins, which were purified by
HPLC on an RP-amide semipreparative column using a gradient of CH3CN/H2O as the eluent (from 40:60 to 100% in 30
min, flow 2.0 mL/min). Two main peaks were collected and subjected
to NMR and were identified as 3-anhydro-6-epi-OpA
(6) and maydispenoid A (9). Fraction F3-10
(11.3 mg) was also subjected to HPLC purification using the same semipreparative
column and eluting with CH3CN/H2O (from 70:30
to 100% in 30 min, flow 2.0 mL/min). Two main peaks were collected
that after NMR analysis have been identified as gigobolin B (2, 1.0 mg) and gigobolin C (3, 0.5 mg). The more
polar fraction F3-27 (33.0 mg) was subjected to further purification
on a silica gel column using a mixture of petroleum ether and diethyl
ether to give 2.0 mg of OpB (7).

Gigobolin A (1)
White powder; [α]D = +17.4 (c 0.03, CHCl3); UV (MeOH) λmax (log ε) 276 (1.03), 225 (1.84), 203 (1.87) nm; ECD
(MeOH) θ235 + 45,736, θ281 + 3982,
θ318 – 1265; 1H and 13C NMR; see Table
; HR-ESI-MS m/z 439.2466 (calcd.
for C25H36O5Na, 439.2455).

Gigobolin B (2)
Pale yellow oil; [α]D = +54.8 (c 0.09, CHCl3); UV (MeOH) λmax (log ε) 277 (1.01), 224 (1.60), 202 (1.72) nm; ECD
(MeOH) θ228 + 1680, θ259 –
306, θ296 + 635; 1H and 13C
NMR; see Table
; HR-ESI-MS m/z 407.2544 (calcd. for C25H36O3Na, 407.2557).

Gigobolin C (3)
White amourphous powder; 1H and 13C NMR; see Table
; HR-ESI-MS m/z 407.2552 (calcd. for C25H36O3Na,
407.2557).

Maydispenoid A31 (9)
[α]D = +54.8 (c 0.04, CH3OH), lit. [α]D = +43.3 (c 0.3, MeOH); ECD (MeOH)
θ211 + 6631; NMR data for compound 9 in CDCl3. 1H NMR δ: 6.13 (s, H-21),
5.16 (br d, 8.4 Hz, H-18), 4.41 (ddd, 8.4, 7.2, 1.5, H, H-17), 4.20
(app d, 7.0 Hz, H-5), 3.47 (dd, 10.0, 6.2 Hz, H-8), 3.24 (s, OCH3-8), 2.51 (dd, 15.2, 7.0 Hz, H-4a), 2.43 (app d, 3.8 Hz, H-6),
2.18 (dq, 7.0, 7.0 Hz, H-15), 2.12 (ddd, 7.8, 4.2, 3.8 Hz, H-2), 2.00
(app dd, 13.7, 6.2 Hz, H-9a), 1.80 (m, H-13a), 1.75 (m, H-16a), 1.70
(s, H3-24), 1.67 (s, H3-25), 1.66 (m, H2-12), 1.65 (m, H-4b), 1.58 (m, H-9b), 1.56 (m, H-1a), 1.45
(ddd, 12.6, 8.4, 3.0 Hz, H-13b), 1.41 (app d, 10.2 Hz, H-10), 1.33
(ddd, 11.7, 7.8, 3.0 Hz, H-16b), 1.26 (s, H3-20), 1.23
(m, H-1b), 0.98 (d, 7.1 Hz, H3-23), 0.91 (s, H3-22); 13C NMR δ: 142.0 (C-21, CH), 134.7 (C-19,
C), 126.5 (C-18, CH), 109.6 (C-7, C), 96.6 (C-14, C), 86.3 (C-3, C),
81.9 (C-8, CH), 75.6 (C-5, CH), 70.8 (C-17, CH), 55.3 (C-10, CH),
55.0 (OCH3-8), 50.6 (C-4, CH2), 43.8 (C-6, CH), 42.6 (C-12, CH2), 42.5 (C-16,
CH2), 38.3 (C-1, CH2), 36.9 (C-2, CH), 36.6
(C-15, CH), 29.1 (C-9, CH2), 28.9 (C-13, CH2), 25.9 (C-24, CH3), 21.1 (C-20, CH3), 18.8
(C-22, CH3), 18.2 (C-25, CH3), 16.8 (C-23, CH3). HR-ESI-MS m/z 439.2828
(calcd. for C26H40O4Na, 439.2824).

Biological Testing

Cells
MCF7, MDA-MB-231, BT-20, SK-BR3, MDA-MB-468,
T98G, and U87 GBM cells were obtained from ATCC. T98G cells were maintained
in Eagle’s Minimum Essential Medium (Mediatech Inc., Corning,
cat no. 10-010-CV, Manassas, VA) supplemental with 10% Fetalgro bovine
growth serum (FBS, Rocky Mountain Biologicals, LLC, Missoula, MO)
and 1% Penn/Strep (Mediatech, Inc., Corning). U87 were maintained
in Dulbecco’s Modified Eagle Medium containing Ham’s
F-12 50/50 mixture (88%, obtained from Mediatech, Inc., Corning),
FBS (10%), Penn/Strep (1%), and nonessential amino acids (1%, obtained
from Gibco, cat no. 11140-050, Waltham, MA). MCF7, MDA-MB-231, SK-BR3,
and MDA-MB-468 cells were maintained in Dulbecco’s Modified
Eagle’s Media (Corning, cat. no. 10-013-CM) supplemented with
10% Fetal bovine serum (ThermoFisherScientific, cat. no. A5256801)
and 1% Penicillin–Streptomycin (ThermoFisherScientific, cat
# 15140122). BT-20 cells were maintained in Minimum Essential Medium
Eagle (MilliporeSigma, cat no. M4655) supplemented with 10% fetal
bovine serum (ThermoFisherScientific, cat no. A5256801) and 1% Penicillin–Streptomycin
(ThermoFisherScientific, cat # 15140122). HMLE and HMLE-TWIST cells
were derived and cultured as previously described. The patient-derived primary glioblastoma stem cells (GSCs),
GSC 040815 and GSC 082209, were developed and cultured as previously
described.
,
The cells were maintained as
neurospheres in neurobasal medium (Invitrogen, Carlsbad, CA), which
was supplemented with B27 serum-free supplement, 1% penicillin–streptomycin,
1% sodium pyruvate, EGF (20 ng/mL), bFGF (20 ng/mL), LIF (10 ng/mL),
and heparin (5 μg/mL).

Cell Viability Assays
The GBM cells were prepared by
trypsinizing each cell line and seeding 1 × 104 cells
per well into microtiter 96-well plates. All compounds were dissolved
in DMSO and diluted in 1% DMSO in full media at concentrations ranging
from 1 mM to 0.1 nM in 10-fold increments prior to cell treatment.
The cells were grown for 24 h in their respective media before the
media was removed and replaced with 100 μL of media plus the
respective drug concentration. After 48 h, the medium was removed
and replaced with 100 μL of MTT reagent in phenol red- and serum-free
medium (0.5 mg/mL) and incubated further for 2 h. Media was removed,
and the resulting formazan crystals were resolubilized in 100 μL
of DMSO. A550 (peak) and A700 (Baseline) were
measured using a BioTek Synergy H4 plate reader. The baseline absorbance
was subtracted from all peak absorbances to obtain the viability.
Cells treated with 1% DMSO were used as a control.
The mammary
and breast cancer cells were prepared by trypsinizing each cell line
and seeding 2 × 103 cells per well into microtiter
96-well plates. All compounds were dissolved in DMSO and diluted in
full media at concentrations ranging from 0.1 mM to 100 nM in 4-fold
increments prior to cell treatment. The cells were grown for 24 h
in their respective media before the media was removed and replaced
with 100 μL of media plus the respective drug concentration
or equivalently diluted DMSO in duplicate wells. After 72 h, the medium
was removed and replaced with 80 μL of medium plus 20 μL
of CellTiterBlue reagent and incubated further for 3 h. Fluorescence
was detected using Ex560 (excitation) and Em590 (emission) wavelengths using a Varioskan LUX Multimode Microplate
Reader plate reader (Thermo Scientific). The baseline fluorescence
was subtracted from all wells to obtain viability, and treated wells
were normalized to the average of the signal from wells with volume-equivalent
DMSO.
The effect of compounds on GSC viability was assessed
using the
CellTiter-Glo 2.0 Cell viability assay (Promega catalog no. G9241,
Madison, WI) following the manufacturer’s protocol. GSCs (4
× 103 cells) were treated for 7 days with either a
vehicle (DMSO 0.1% v/v), OpA (4), or new analogs at specified
doses. Luminescence levels were measured by using a Promega Glomax
Luminometer.

Statistical Analysis
Six replicates were used per condition.
IC50 values and standard deviations were calculated on
GraphPad Prism v9 using (log) inhibitor vs normalized responsevariable
slope analysis using a four parameter mode.

Supplementary Material