Caspase-1 Inhibitors from an Extremophilic Fungus that Target Specific Leukemia Cell Lines

Abstract

Berkeley Pit Lake, Butte, Montana is a 540 m. deep abandoned open-pit copper mine filled with over 140 billion liters of acidic, metal-sulfate contaminated water. This harsh environment has yielded several microorganisms that produce interesting biologically active compounds. Several polyketide metabolites including the new berkazaphilones A (1) and B (2) and octadienoic acid derivatives berkedienoic acid (13) and berkedienolactone (15), as well as previously reported azaphilone 4, vermistatin (6), dihydrovermistatin (7), penisimplicissin (8), aldehyde 9 and methyl paraconic acid (11) were isolated from a culture broth of Penicillium rubrum taken from a depth of 270 m. The structures of these compounds were deduced by interpretation of spectroscopic data. The compounds were isolated either for their inhibition of the signal transduction enzyme caspase-1 or because of their structural similarity to these inhibitors. Selected compounds were further evaluated for their ability to inhibit interleukin-1β production by inflammasomes in induced THP-1 cells. Berkazaphilone B (2) and C (4) and vermistatin analogue penisimplicissin (8) exhibited selective activity against leukemia cancer cell lines in the National Cancer Institute 60 human cell line assay.

The Berkeley Pit Lake System is one of the largest contaminated sites in North America. The Pit itself is over 540 m deep with a surface area of 3.2 km² and is continually filling with metal-sulfate rich, acidic water (pH 2.5), at a rate of 10 million L/day. This represents roughly 140 billion L. of contaminated water and constitutes an important component of the largest EPA Superfund site in the United States.¹

In 1995 we began to study the microbes inhabiting the waters of this Pit Lake as if they were inhabitants of a new and exotic ecosystem. Over the past fifteen years we have studied the secondary metabolism of several microbes isolated from the water and sediments of this ecosystem under a variety of physicochemical conditions to determine whether or not they produce metabolites with desirable bioactivity. This approach has yielded interesting results.^2–8

Bioactivity is currently assessed using 96-well plate assays that demonstrate the ability of crude extracts, column fractions, and pure compounds to inhibit specific signal transduction enzymes. We routinely target the enzymes matrix metalloproteinase-3 (MMP-3), caspase-1 and caspase-3. These assays are proving to be effective tools for assessing the bioactivities of crude extracts and guiding isolation of pure enzyme inhibitors. The lead compounds presented in this manuscript were isolated because of their ability to inhibit caspase-1. After the structures of these compounds were elucidated, compounds with similar ¹H NMR spectral characteristics were also isolated to ascertain how subtle differences in structure might affect biological activity.

Caspase-1 was the first of a novel type of cysteine protease responsible for converting interleukin-1β to its mature form in monocytes. Caspase-1, also known as interleukin-1 converting enzyme (ICE), is responsible for the activation of IL-1β and IL-18 from precursor molecules.⁹ Caspase-1 is activated upon binding to the inflammasome, a multiprotein complex that plays a key role in innate immunity by activating the proinflammatory pleiotropic cytokines interleukin 1-β and IL-18.⁹ There is a strong correlation between dysregulated inflammasome activity and both inherited and acquired inflammatory diseases.⁹

Several researchers have demonstrated that caspase-1 inhibitors have shown promise in delaying the onset of Huntington's disease¹⁰ and amyotropic lateral sclerosis,¹¹ and in mitigating the effects of stroke¹² and multiple sclerosis.^{13, 14} All of these diseases exhibit autoimmune phenomena. Capsase-1 has also been implicated in the physiological production of interferon-gamma-inducing factor (IGIF). It therefore appears to play a critical role in the regulation of multiple proinflammatory cytokines.¹⁵

The up-regulation of caspase-1 and concomitant chronic inflammation has been associated with a number of different pathologies including the development of insulin resistance in obesity related diabetes,¹⁶ degeneration of retinal capillaries associated with diabetes and galactosemia,¹⁷ the demyelination of neurons in multiple sclerosis,^11,18 and the formation of amyloid plaques in Alzheimer's disease.¹⁹ High levels of caspase-1 and interleukin-1β have been found in certain cancers²⁰ by many different researchers: acute myelogenous leukemia,²¹ melanoma,^22–23 certain glioblastomas^24–25 and pancreatic cancers,^26–29 certain breast cancers,³⁰ and human cancer xenografts³¹ all of which may be exacerbated by chronic inflammation associated with activation of the inflammasome.

Caspase-1 inhibitors have been proposed as potential therapies for the above mentioned cancers, as well as osteoarthritis and rheumatoid arthritis,^32–33 Alzheimer's disease,¹⁹ amyotrophic lateral sclerosis⁹ and brain and nerve trauma.^34–35

Caspase-1 is also down-regulated in many solid tumor cancers and activation of caspase-1 in prostate cancer and ovarian cancer may be required for apoptotic breakdown of tumors. The development of new caspase-1 inhibitors will not only provide potential chemotherapeutics but will also provide tools for the investigation of the intricacies of signal transduction.

One of the first microbes to be studied from the Pit Lake was isolated from a depth of 270 m. and was subsequently identified as Penicillium rubrum Stoll based on rRNA sequencing alignment data (300 base pairs). The fungus was grown in acidified potato dextrose broth (pH 2.7) for 21 days as a still culture. At time of harvest the mycelium was removed by filtration and the broth was thoroughly extracted with chloroform. This organic extract was active in all three enzyme inhibition assays, although in this study we focused on the caspase-1 inhibitors. Size exclusion chromatography (LH-20) followed by HPLC yielded the new berkazaphilones A and B (1) and (2) as well as the new octadienoic acid derivatives berkedienoic acid (13) and berkedienolactone (15), and the previously reported azaphilone (4), vermistatin (6), dihydrovermistatin (7), penisimplicissin (8), aldehyde 9 and methyl paraconic acid 11.

HREIMS of 1 gave a molecular formula of C₁₃H₁₆O₃, corresponding to a molecule with six sites of unsaturation. The ¹³C NMR spectrum showed seven sp² hybridized carbons in the molecule, suggesting the presence of three double bonds and one carbonyl carbon. A bicyclic ring system accommodated the two remaining sites of unsaturation. The UV spectrum (λ_max 344 nm) indicated extended conjugation and the IR spectrum showed the presence of a dienone moiety (ν_max 1644 cm⁻¹)³⁶ that was supported by a carbon resonating at δ_C 198.2 ppm in the ¹³C NMR spectrum. The DEPT spectrum showed that fifteen protons were attached to carbons. The strong OH- stretch in the IR spectrum (ν_max 3415 cm⁻¹) provided sufficient information to assign the remaining proton to the hydroxy group. The ¹H NMR (Table 1) and ¹H-¹H COSY spectra indicated the presence of two spin systems: CH₃-CH-CHO-CH-CH₂O and a terminal CH=CH-CH₃, as well as two olefinic protons at δ_H 5.71 (d, J = 1.9 Hz) and δ_H 5.49 (s). The HMBC spectrum provided correlations to connect the first spin system to ketone C-6 (δ_C 198.2). Mutually ³J-coupled methine H-7 (δ_H 2.41, J = 10.3, 6.6) and methyl doublet H₃-12 (δ_H 1.25, J =6.6) showed strong HMBC correlations to carbonyl C-6 and to oxygen-bearing C-8 (δ_C 74.2 ppm). H₃-12 also showed correlations to methine C-7 (δ 50.0 ppm). ¹H-¹H COSY showed ³J- ppm coupling between H-7 and H-8 (δ_H 3.43, J = 10.3), between H-8 and methine H-8a (δ_H 2.84, J = 10.3), and between H-8a and oxygen-bearing methylene H-1 (1α = δ_H 3.72, J = 13.2; 1β = δ_H 4.79, J = 5.4). Methylene H-1β could be connected to the conjugated diene system through 3-bond HMBC correlations to oxygen-bearing C-3 (δ_C 159.9) and C-4a (δ_C 150.5). HMBC correlations from olefinic H-4 (δ_H 5.49) to C-5 (δ_C 118.1) established the diene backbone, and from H-4 to C-9 (δ_C 125.4) connected the diene system to the terminal propylene moiety. H-4 also showed a HMBC correlation to C-8a (δ_C 40.9), which allowed the α,β,γ,δ-unsaturated ketone bicyclic ring system to be established. The configuration of the propylene olefin could be established as E based on the magnitude of the 15.7 Hz coupling between H-9 and H-10.

Table 1.

¹³C and ¹H NMR Data for Compounds 1, 2 and 4, Berkazaphilones A-C, in CDCl₃ (δ ppm)^a

	1		2		4
atom	δC, type	δH, mult. (J in Hz)	δC, type	δH mult. (J in Hz)	δC, type	δH mult. (J in Hz)
1	68.6, CH2	β 4.79, dd (10.9, 5.4) α 3.72, dd (13.2, 10.9)	68.4, CH2	β 4.82, dd (11.1,5.3) α 3.78, dd (13.5, 11.1)	67.9, CH2	β 4.38, dd (10.8, 5.4) α 3.85, dd (13.0, 10.8)
3	159.9, C		160.8, C		160.7, C
4	102.8, CH	5.49, s	102.4, CH	5.51, s	102.8, CH	5.56, s
4a	150.5, C		152.5, C		151.7, C
5	118.1, CH	5.71, d (1.9)	115.4, CH	5.78, d (1.6)	115.7, CH	5.79, d (1.7)
6	198.2, C		190.5, C		194.8, C
7	50.0, CH	2.41, dq (10.3, 6.6)	89.3, C		74.1, C
8	74.2, CH	3.43, bt (10.3)	74.6, CH	3.58, bt (10.1)	74.7, CH	5.29, d (9.9)
8a	40.9, CH	2.84, dddd (13.2, 10.3, 5.4, 1.9)	37.8, CH	2.82, dddd (13.5, 10.1, 5.3, 1.6)	34.9, CH	3.44, dddd (13.0, 9.9, 5.4, 1.7)
9	125.4, CH	5.87, dq (15.7, 1.7)	125.2, CH	5.85, dd (15.3,1.5)	125.2, CH	5.87, dd (15.4, 1.7)
10	133.2, CH	6.40, dq (15.7, 6.8)	134.4, CH	6.45, dq (15.3,6.9)	134.5, CH	6.41, dd (15.4, 7.0)
11	18.4, CH3	1.84, dd (6.8, 1.7)	18.4, CH3	1.84, dd (6.9,1.5)	18.5, CH3	1.84, dd, (7.0, 1.7)
12	10.8, CH3	1.25, d (3H, 6.6)	15.4, CH3	1.77, s (3H)	20.4, CH3	1.38, s (3H)
1′			172.0, C		170.6, C
2′			104.7, C		104.4, C
3′			166.0, C		165.9, C
4′			101.5, CH	6.25, d (2.4)	101.6, CH	6.30, d (2.3)
5′			161.4, C		161.1, C
6′			112.1, CH	6.18, d (2.4)	112.0, CH	6.27, d (2.3)
7′			145.0, C		144.6, C
8′			24.7 q	2.28, s (3H)	25.0, CH3	2.58, s (3H)

^aAssignments of the ¹³C and ¹H signals were made based on HSQC spectral data.

The relative stereochemistry of 1 was established by interpretation of ¹H NMR and 1D NOE difference spectra. The broad triplet H-8 showed ax/ax coupling (J = 10.3 Hz) to both H-7 and H-8a, which also exhibited strong ax/ax coupling (J = 13.2 Hz) to H-1α. In the NOE difference spectrum irradiation of H-8a enhanced the resonance of H-1β, supporting the relative stereochemistry assignment as shown. These data were used to generate the proposed structure for berkazaphilone A (1).

Compound 1 belongs to the class of fungal metabolites known as the azaphilones. A comprehensive review of azaphilone analogues in 2010 listed over 170 compounds from 23 different fungal genera.³⁷ Most of the known azaphilones are oxygenated at both C-7 and C-8 and often form orsellinic or chlorinated orsellinic acid esters. Of the azaphilone analogues reviewed, only one other compound, pseudohalonectrin, was not oxygenated at C-7.³⁷

HREIMS established the molecular formula of compound 2 as C₂₁H₂₂O₇ which showed compound 2 to have eight more carbons than 1 and five additional sites of unsaturation, indicative of an aromatic moiety. There were many similarities between the ¹³C and ¹H NMR spectra of compounds 1 and 2 (Table 1). Two obvious differences were due to the replacement of the C-7 methine in compound 1 by a quaternary oxygen-bearing carbon (δ_C 89.3), leading to the replacement of the H₃-12 doublet of 1 with a singlet in compound 2 (δ_H 1.77). These data suggested that 2 had the same carbon skeleton as 1 with a substituent at C-7. HMBC and ¹H NMR spectral data provided the necessary information to support this deduction. A terminal E- configured propylene moiety, olefinic singlet H-4 (δ_H 5.51) and doublet H-5 (δ_H 5.78 ppm) were present as in 1. ³J-coupling data again showed ax/ax interactions between H-8 and H-8a (J = 10.1 Hz) and between H-8a and H-1α (J = 13.5 Hz), and ax/eq interaction between H-8a and H-1β (J = 5.3). Singlet H₃-12 showed 3-bond correlations to ketone C-6 (δ_C 190.5) and methine C-8 (δ_C 74.6 ppm), as well as 2-bond coupling to quaternary C-7. H-1β showed a 2-bond correlation to C-8a (δ_C 37.8) and 3-bond correlations to C-3 (δ_C 160.8) and C-4a (δ_C 152.5 ppm). Olefinic H-4 had 3-bond correlations to C-8a and C-5 (δ_C 115.4) again establishing the bicyclic diene, and to C-9 (δ_C 125.2), connecting the ring system to the terminal propylene moiety. H-4 also showed 2-bond correlations to C-3 and C-4a. H-5 (δ_H 5.78) showed 3-bond correlations to C-4 (δ_C 102.4) and to C-7, supporting the location of the oxy-substituent at C-7.

Eight carbons, seven hydrogens, three oxygens and five sites of unsaturation remained to be assigned and attached to the oxygen at C-7. A carbon resonating at δ_C 172.0 (C-1′) and an IR absorbance of 1708 cm⁻¹ showed the presence of an aromatic ester. The DEPT and ¹³C spectra showed the presence of five quaternary carbons, including the carbonyl carbon, two methines and a methyl carbon. Chemical shifts indicated that all but the methyl were sp²-hybridized, consistent with an aromatic ester, and that two carbons were oxygen-bearing. HMBC data showed 2-bond correlations from H-4′ (δ_H 6.25, J = 2.4 Hz) to both oxygen bearing C-3′ (δ_C 166.0) and C-5′ (δ_C 161.4). H-6′ (δ_H 6.18, J = 2.4 Hz) showed a 3-bond correlation to methyl C-8′ (δ_C 24.7), and H₃-8′ (δ_H 2.28) showed 3-bond correlations to both C-6′ (δ_C 101.5) and C-2′ (δ_C 104.7). These data established the substituent as an orsellinic acid moiety and were consistent with literature data.³⁸

Acetylation of compound 2 resulted in triacetate 3. The orsellinic acid moiety was diacetylated as expected and chemical shift data was consistent with the literature.³⁹ Acetylation of the hydroxyl group at C-8 induced a large downfield shift of H-8 from q 3.58 to δ5.00 (d, J= 9.9 Hz) and a downfield shift of H-8a from δ_H 2.82 to δ 3.34 ppm. H-8 was ³J-coupled to H-8a and showed HMBC correlations to C-8a, C-1, C-7 and to an acetate carbonyl. These data confirmed the structure proposed for compound 2, berkazaphilone B, with the placement of the orsellinic moiety at C-7.

Compound 4 was isomeric with 2, with a molecular formula of C₂₁H₂₂O₇ established by HREIMS. The ¹³C NMR spectra of 2 and 4 were very similar (see Table 1) but the ¹H NMR spectrum had two distinct differences. Methylene proton H-1β showed a marked upfield shift and methine H-8 showed a marked downfield shift in compound 4 when compared to 2. These differences indicated that the orsellinic acid moiety was at C-8 rather than C-7. HMBC correlations were observed from H-8 (δ_H 5.29) to C-1′ (δ_C 170.6), confirming this deduction. Acetylation of 4 gave the expected diacetate 5. The proposed compound has been previously reported and the NMR and mass spectral data of compound 4, berkazaphilone C, compared favorably to the data reported for azaphilone Sch 725680.³⁸ Unfortunately, the original authors did not provide the optical rotation for their compound so we cannot be sure if the two are the same stereoisomer.³⁸

Examination of mass spectral, NMR and optical rotation data indicated that the major cytotoxic compound in the extract was the known fungal metabolite vermistatin (6), which was previously reported as a metabolite of Penicillium vermiculatum.⁴⁰ The NMR data of compounds 7 and 8 were similar to that of vermistatin, indicating strong structural similarities. The HREIMS of compound 7 gave a molecular formula of C₁₈H₁₈O₆, with two more hydrogens than vermistatin. Indeed, the only major difference between the NMR spectra of vermistatin 6 and compound 7 were the peaks associated with the terminal propylene moiety. It was apparent from the NMR data that it was reduced to a n-propyl moiety, designating 7 as 14,15-dihydrovermistatin. Dihydrovermistatin was previously reported from broth cultures of Penicillium simplicissimum.⁴¹

The third vermistatin analogue, compound 8, had a molecular formula of C₁₆H₁₄O₆, with two less carbons than either vermistatin (6) or dihydrovermistatin (7). In this case, the ¹H NMR signals of both the terminal propylene and propyl moiety were absent and were replaced by a methyl singlet at δ 2.44. This compound was previously reported as penisimplicissin from broth cultures of Penicillium simplicissimum.⁴¹ Compounds 7 and 8 had the same sign and relative optical rotations as vermistatin (6), suggesting that they had the same configuration.⁴⁰

The molecular formula of compound 9 was C₁₃H₁₆O₅ as determined from HREIMS. ¹³C NMR and DEPT spectra indicated the presence of two carbonyl carbons: ketone C-8 at δ_C 208.8 (C) and aldehyde C-12 at δ_C 195.1 (CH). Six additional sp² hybridized carbons indicated the presence of an aromatic moiety with two oxygen-bearing carbons resonating at δ_C 165.2 and δ_C 164.6 ppm. These assignments accommodated all six sites of unsaturation associated with the molecular formula. Examination of the spectral data established the structure of compound 9 as shown, which was previously reported as a metabolite of Aspergillus versicolor.⁴² We prepared the Mosher ester of 9 and found the same absolute configuration (R) as reported.⁴³ The (S) stereoisomer of compound 9 has also been reported from a Pseudobotrytis sp.⁴⁴

EIMS established the molecular formula of compound 11 as C₆H₈O₄, associated with three sites of unsaturation. Its infrared spectrum indicated the presence of two carbonyl moieties: the broad –OH stretch (ν_max 3027 cm⁻¹) and carbonyl absorption at 1716 cm⁻¹ indicated the presence of a saturated carboxylic acid, while the carbonyl absorption at 1774 cm⁻¹ indicated the presence of a saturated γ-butyrolactone.⁴⁵ The presence of the acid was confirmed by methylation of compound 11 with diazomethane to yield the methyl ester, 12. Analysis of the spectral data established the structure of 11, which is the known compound α-methyl paraconic acid.⁴⁶

Compound 13 had a molecular formula of C₁₁H₁₆O₃ established by the HRESIMS spectrum, with four sites of unsaturation. The broad –OH stretch (ν_max = 3021 cm⁻¹) and the carbonyl absorption at 1683 cm⁻¹ in the infrared spectrum were indicative of an α,β-unsaturated carboxylic acid. Methylation of 13 with diazomethane yielded the methyl ester 14 which had a molecular formula of C₁₂H₁₉O₃, as determined by HREIMS.

The ¹H-¹H COSY spectrum of 13 provided connectivity data for two spin systems beginning with a terminal diene. The ³J-coupling of the diene protons clearly correlated H-8a (δ_H 5.20) and H-8b (δ_H 5.08) to H-7 (δ_H 6.28), H-7 to H-6 (δ_H 6.20), H-6 to H-5 (δ_H 5.70), H-5 to oxygen-bearing methine H-4 (δ_H 4.31), and H-4 to methylene H₂-3 (δ_H 2.57). The second spin system consisted of a terminal propylidene moiety with ³J-coupling from H-9 (δ_H 7.02) to methylene H₂-10 (δ_H 2.23) and from H-10 to the terminal methyl H₃-11 (δ_H 1.05). The HMBC spectrum provided the necessary information to connect these two spin systems to the carbonyl carbon with 3-bond correlations from H-9 and H₂-3 to C-1. H-9 also exhibited correlations to C-3 and C-10. Methylene H₂-3 provided a point of connectivity with 3-bond correlations to C-1, C-9 and C-5, and 2-bond correlations to flanking carbons C-2 and C-4. H-4 also afforded nice bilateral connectivity with 3-bond correlations to C-2 and C-6, and 2-bond correlations to C-3 and C-5. The chemical shift of H-9 indicated an E configuration for Δ2,9.⁴⁷ The magnitude of the coupling constant between olefinic H-5 and H-6 (J = 14.3 Hz) also indicated an E configuration. These data generated the structure proposed for compound 13, berkedienoic acid.

Compound 15 had a molecular formula of C₁₁H₁₄O₂ established by the HRESIMS [M+H]⁺ peak at m/z 179.1082. The carbonyl absorption at 1751 cm⁻¹ and accompanying C=C absorption at 1679 cm⁻¹ in the infrared spectrum indicated the presence of an αβ-unsaturated γ-lactone.⁴⁵ These data and the similarities in the NMR spectra suggested that compound 15 was the γ-butyrolactone of compound 13. The ¹H-¹H COSY and the ¹H NMR spectra provided connectivity data for a single extended spin system beginning with a terminal diene at one end. The ³J-coupling of the diene protons correlated H-8a (δ_H 5.28) and H-8b (δ_H 5.18) to H-7 (δ_H 6.30), H-7 to H-6 (δ_H 6.30), H-6 to H-5 (δ_H 5.69), H-5 to oxygen-bearing methine H-4 (δ_H 4.98), and H-4 to methylene H₂-3 (δ_H 2.58, 3.05 ppm).

The terminal propylidene moiety produced ³J-coupling from methyl H₃-11 (δ_H 1.07 ppm) to methylene H₂-10 (δ_H 2.17), and from H₂-10 to olefinic H-9 (δ_H 6.72). The coupling pattern for methylene H₂-10, however, was quite complex; the expected pentet pattern was actually a pentet of triplets and the ¹H-¹H COSY spectrum showed connectivity to methylene H₂-3, which required 5-bond coupling in our proposed structure. The coupling patterns for H₂-3a and b were also complex and both were doublets of doublets of pentets. Bearing in mind that a pentet is actually a dddd with equivalent coupling constants, then these patterns become complex indeed. These data suggest that these unusually complex patterns are the result of allylic and homoallylic coupling. Studies on 2-butene and compounds containing a butenyl moiety showed similar correlations.⁴⁸

Homoallylic coupling data from both the ¹H NMR and the ¹H-¹H COSY spectra provided the information to connect the two ends of the molecule. The coupling patterns for this molecule were complex but could be deconstructed to include homoallylic coupling between the methylenes H₂-10 and H₂-3.

The exocyclic double bond could be established as the E- isomer by chemical shift arguments. A series of E and Z isomers of α-alkylidene-γ-butyrolactones were synthesized and the chemical shifts of the olefinic protons analogous to H-9 were compared.⁴⁹ In all cases the chemical shift of H-9 in the Z-isomer approached δ_H 6.0 ppm while in the E-isomer it was closer to δ_H 6.6 ppm.⁴⁹ These data were used to generate the structure proposed for compound 15, berkedienolactone. A series of similar lactones including the tricyclic compound gallielalactone were isolated from an unidentified ascomycete.^50,51 Unfortunately, the reported monocyclic lactones most structurally similar to 15 were not isolated as pure compounds, so it was not possible to compare optical rotation data with that of berkedienolactone (15).

Compounds 1, 2, 4, and 6–9 were evaluated for their ability to inhibit caspase-1 in vitro, and the most active compounds and closely related analogues were evaluated for their ability to inhibit the production of interleukin 1-β in THP-1 cells (pro-monocytic leukemia cell line). THP-1 cells produce high levels of IL-1β when induced with titanium nanowires and bacterial lipopolysaccharide (LPS). Caspase-1 inhibition was determined in a fluorometric assay normalized to 1.00, where 0 is total enzyme inhibition and 1.00 is lack of enzyme inhibition. Both berkazaphilone B (2) and C (4) had IC₁₀₀ values of 25 μM against caspase-1, while berkazaphilone A (1), penisimplicissin (8), and compound 9 were completely inhibitory at a concentration of 250 μM. Vermistatin (6) and dihydrovermistatin (7) were not inhibitory at the concentrations tested.

Induced THP-1 cells were exposed to compounds 2, 4, and 6–8, and the concentrations of IL-1β post-exposure were determined. All of the compounds tested inhibited the production of IL-1β in THP-1 cells at a concentration of 250 μM. In dilution assays, however, only compounds 2 and 4 inhibited the production of IL-1β. Compounds 2 and 4 completely inhibited the production of IL-1β at concentrations of 5 μM and 50 μM, respectively.

Compounds 2, 4, 7, and 8 were tested in the National Cancer Institute (NCI) antitumor screen against 60 human cell lines.⁵² The compounds showed selective cytotoxicity towards leukemia cell lines only. Berkazaphilone B (2) exhibited a log₁₀ GI₅₀ of −5.67 against cell line RPMI-8226 and berkazaphilone C (4) exhibited a log₁₀ GI₅₀ of −6.42 against cell line SR. In the vermistatin family, penisimplicissin (8) exhibited a log₁₀ GI₅₀ of −6.70 against cell line CCRFCEM and −5.83 against HL-60(TB), and dihydrovermistatin (7) was inactive at the concentrations tested. Vermistatin (6) had been previously tested by the NCI and was also inactive. (Supplemental information)

The NCI Molecular Target database includes experiments that determine relative RNA levels for nearly 10,000 human clones, measured in microarray experiments for the NCI cell lines. It was interesting to note that in several microarray experiments caspase-1 was upregulated almost exclusively in different leukemia cell lines.^53–56

Experimental Section

General Experimental Procedures

¹H- and ¹³C-NMR spectra were run on Bruker DPX-300. Chemical shifts were recorded with respect to the deuterated solvent shift (CDCl₃, δ 7.24 for the proton resonance and δ 77.0 for the carbon). IR spectra were recorded on a Nicolet NEXUS 670 FT-IR spectrometer. Optical rotations were measured on a Perkin Elmer 241 MC Polarimeter using a 1 mL cell. Mass spectral data were provided by the Mass Spectrometry, Proteomics and Metabolomics Facility at Montana State University and the Mass Spectral Analysis laboratory at the University of Montana. All solvents used were spectral grade or distilled prior to use.

Collection, Extraction, and Isolation Procedures

The collection and isolation of the Berkeley Pit fungi has previously been described.^{7, 8} The fungus was identified as Penicillium rubrum by Microbial Identification, Inc. The fungus was grown at room temperature in 26 × 300 mL of DIFCO® potato dextrose broth (acidified to pH 2.7 with sulfuric acid) in 1 L Erlenmeyer flasks (shaken at 180 rpm for 6 days then still for 15 days). At time of harvest 50 mL of MeOH was added/flask. The combined cultures (7.8 L) were filtered through cheesecloth to remove the mycelia mat for a separate study. The filtrate from the combined cultures was extracted three times with 1L of CHCl₃ and the extract was reduced in vacuo to an oil (1.14 g). This extract demonstrated inhibition of caspase-1 and MMP-3, antimicrobial activity against Staphylococcus aureus and Escherichia coli, and brine shrimp lethality.

The CHCl₃ extract was fractionated using a flash Si gel column using hexanes, hexane/isopropyl alcohol mixtures to isopropyl alcohol /MeOH mixtures. The 50% IPA/hexane fraction was further fractionated by preparative HPLC on a Rainin 21 mm preparative Si gel column with a hexanes/isopropyl alcohol gradient. The 10% IPA fraction was further fractionated on Sigel to yield the three azaphilone derivatives 1 (5.1 mg), 2 (35.1 mg) and 4 (7.1 mg). The fraction that eluted with 50% IPA/hexane yielded the three cytotoxic compounds 6 (46.2 mg), 7 (4.0 mg), and 8 (18.0 mg). The 25% IPA/hexane fraction yielded aldehyde 9 (5.6 mg), lactone 11, berkedienoic acid 13 (1.5 mg) and berkedienolactone 14 (2.0 mg)

Berkazaphilone A (1): [α]²⁰_D +20.5° (c 0.0019, MeOH); UV (CHCl₃) λ_max (log ε) 344 (4.36), 241 (3.75) nm; IR (CHCl₃) ν_max 3415, 3020, 2934, 1644, 1592, 1384, 1063, 876 cm⁻¹; ¹H NMR and ¹³C NMR (CDCl₃) see Table 1 ; EIMS m/z 220 (50), 162 (48), 134 (100); HREIMS m/z 220.1099 [M+] (calcd for C₁₃H₁₆O₃, 220.1099).

Berkazaphilone B (2): [α]²⁰_D −55.2° (c 0.0025, MeOH); UV (MeOH) λ_max (log ε) 349 (4.11), 270 (4.05) nm; IR (neat) λ_max 3400, 2933, 1708, 1669, 1633, 1585, 1326, 1267, 1172, 1118 cm⁻¹; ¹H NMR and ¹³C NMR (CDCl₃) see Table 1; EIMS m/z 386 (15), 236 (20), 218 (38), 175 (83), 162 (94), 134 (100), 69 (85); HREIMS m/z 386.1363, [M]⁺ (calcd for C₂₁H₂₂O₇, 386.1365).

Acetylation of compound 2

Compound 2 (1.0 mg) was dissolved in pyridine (50 μL) and Ac₂O (50 μL) and stirred for 24 hours. After that time the solvents were removed to give 3 as an oil (0.9 mg).

Compound 3: ¹H NMR (CDCl₃) δ 6.82 (d, J = 2.0 Hz, H-6′), 6.77 (d, J = 2.0 Hz, H-4′), 6.38 (dq, J = 15.1, 6.9, H-10), 5.83 (dd, J = 15.1, 1.7 Hz, H-9), 5.82 (d, J = 1.6 Hz, H-5), 5.50 (s, H-4), 5.00 (d, J = 9.9 Hz, H-8), 4.26 (dd, J = 10.7, 5.1 Hz, H-1β), 3.71 (d, J = 13.2, 10.7 Hz, H-1α), 3.34 (dddd, J = 13.2, 9.9, 5.1, 1.6 Hz, H-8a), 2.33 (s, 3H, H-8′), 2.26 (s, 3H, OAc), 2.22 (s, 3H, OAc), 2.17 (s, 3H, OAc), 1.83 (dd, J = 6.9, 1.7 Hz, 3H, H-11), 1.61 (s, 3H, H-12); ¹³C NMR (CDCl₃) δ 190.4 (C-6), 170.2 (OAc), 168.8 (OAc), 168.5 (OAc), 164.1 (C-1′), 159.9 (C-3), 151.8 (C-5′), 149.8 (C-4a), 148.6 (C-3′), 138.4 (C-7′), 133.4 (C-10), 125.1 (C-9), 124.1 (C-2′), 121.0 (C-6′), 116.4 (C-5), 114.1 (C-4′), 102.4 (C-4), 83.3 (C-7), 74.2 (C-8), 67.8 (C-1), 35.8 (C-8a), 21.1 (OAc), 20.8 (OAc), 20.6 (OAc), 19.8 (C-8′), 18.4 (C-11), 17.2 (C-12).

Berkazaphilone C (4): [α]²⁵_D +54.3° (c 0.0030 CHCl₃); ¹H NMR and ¹³C NMR (CDCl₃) see Table 1; EIMS m/z 386; HREIMS m/z 386.1365 [M]⁺ (calcd for C₂₁H₂₂O₇, 386.1365).

Acetylation of compound 4

Compound 4 (1.0 mg) was dissolved in pyridine (50μ L) and Ac₂O (50 μL) and stirred for 24 hours. After that time the solvents were removed to give 5 as an oil (0.9 mg).

Compound 5: ¹H NMR (CDCl₃) δ 6.93 (d, J = 1.6 Hz, H-6′), 6.86 (d, J = 1.6 Hz, H-4′), 6.42 (dq, J = 15.2, 6.9 Hz, H-10), 5.88 (bd, J = 15.2, 1.5 Hz, H-9), 5.78 (bd, J = 1.9 Hz, H-5), 5.56 (s, H-4), 5.17 (d, J = 10.0, H-8), 4.51 (dd, J = 10.8, 5.4 Hz, H-1β), 3.85 (dd, J = 13.5, 10.8 Hz, H-1α), 3.34 (m, H-8a), 2.41 (s, 3H, H-8′), 2.29 (s, 6H, OAc), 1.85 (dd, J = 6.9, 1.5 Hz, 3H, H-11), 1.41 (s, 3H, H-12); ¹³C NMR (CDCl₃) δ 193.9 (C-6), 168.9 (OAc), 168.6 (OAc), 165.4 (C-1′), 160.4 (C-3), 149.6 (C-3′), 151.1, (C-5′), 152.1 (C-4a), 139.9 (C-7′), 134.3 (C-10), 122.5 (C-2′), 121.6 (C-6′), 125.2 (C-9), 116.0 (C-5), 114.7 (C-4′), 102.8 (C-4), 74.6 (C-7), 74.2 (C-8), 67.9 (C-1), 34.9 (C-8a), 21.1 (OAc, 2C), 21.1 (C-8′), 20.2 (C-12), 18.4 (C-11).

Berkedienoic acid (13): solid, IR (CHCl₃) ν_max 3021 (broad), 2971, 2877, 1683, 1637, 1419, 1004 cm⁻¹; ¹H NMR (CDCl₃) δ 7.02 (1H, t, J = 7.5 Hz, H-9), 6.28 (1H, ddd, J = 16.1, 10.4, 9.6 Hz, H-7), 6.20 (1H, dd, J = 14.3, 10.4 Hz, H-6), 5.70 (1H, dd, J = 14.3, 6.4 Hz, H-5), 5.20 (1 H, dd, J = 16.1, 1.4 Hz, H-8), 5.08 (1H, dd, J = 9.6, 1.4 Hz, H-8), 4.31 (1H, q, J = 6.4 Hz, H-4), 2.57 (2H, m, H-3), 2.23 (2H, quin, J = 7.5 Hz, H-10), 1.05 (3H, t, J = 7.5 Hz, H-11); ¹³C NMR (CDCl₃) δ 172.7 (C, C-1), 149.9 (CH, C-9), 136.2 (CH, C-7), 135.5 (CH, C-5), 131.0 (CH, C-6), 126.8 (C, C-2), 117.8 (CH₂, C-8), 71.8 (CH, C-4), 34.3 (CH₂, C-3), 22.6 (CH₂, C-10), 13.0 (CH₃, C-11); HRESIMS [M – H₂O + H]⁺ m/z 179.1076 (calcd for C₁₁H₁₅O₂, 179.1072).

Methylation of 13: Compound 13, (0.2 mg) was dissolved in Et₂O (100 μL) and a solution of CH₂N₂ in Et₂O was added drop wise until the yellow color persisted. The solution was stirred for 5 minutes and the solvent removed under a stream of N₂ to yield the methyl ester 14 (0.2 mg). HREIMS m/z [M+H]⁺ m/z 211.1325 (calcd for C₁₂H₁₉O₃, 211.1334. (calcd for C₁₁H₁₆O₃).

Berkedienolactone (15): solid, [α]²⁵_D −37.0° (c 0.0020, MeOH); IR (CHCl₃) ν_max 3023, 2971, 2877, 1751, 1679, 1301, 1016 cm⁻¹; ¹H NMR (CDCl₃) δ 6.72 (1H, tt, J = 7.5, 2.8 Hz, H-9), 6.30 (2 H, m, H-7, H-6), 5.69 (1H, dd, J = 14.3, 6.7 Hz, H-5), 5.28 (1H, dd, J = 15.1, 1.38 Hz, H-8), 5.18 (1H, dd, J = 9.5, 1.38 Hz, H-8), 4.98 (1H, ddd, J = 8.1, 7.6, 6.2 Hz, H-4), 3.05 (1H, dddd, J = 16.3, 8.1, 2.8, 1.4 Hz, H-3), 2.58 (1H, ddpentet, J = 16.3, 6.2, 1.5 Hz, H-3), 2.17 (2H, m, H-10), 1.07 (3H, t, J =7.5 Hz, H-11); ¹³C NMR (CDCl₃) δ 170.7 (C, C-1), 142.5 (CH, C-9), 135.4 (CH, C-6), 133.3 (CH, C-7), 131.0 (CH, C-5), 125.1 (C, C-2), 75.7 (CH, C-4), 119.6 (CH₂, C-8), 31.8 (CH , C-3), 23.6 (CH₂, C-10), 12.6 (CH₃, C-11); HRESIMS [M + H] m/z 179.1082 (calcd for C₁₁H₁₅O₂, 179.1072).

In Vitro THP-1 Assay

Human monocyte cell line THP-1 was purchased from ATCC (#TIB-202). The cells were suspended at (2–4) × 10⁵ viable cells/mL in RPMI media supplemented with 10% fetal bovine serum, 0.05 mM 2-mercaptoethanol, sodium pyruvate, and an antimycotic/antibiotic cocktail containing penicillin, streptomycin and amphotericin B (Mediatech,VWR). The cells were differentiated into macrophage-like cells by the phorbol ester, PMA (1 μg/mL, Sigma), 24 hour prior to experimentation. The transformed cells were removed from the flask by scraping, and centrifuged at 450 g for 5 min. The resulting cell pellet was suspended at 1.0 × 10⁶ cells/mL and exposed to caspase-1 inhibitors at concentrations described below (0.5%–0.005%), LPS [20ng/mL] and TiO₂ nanowires (100 μg/mL). Experiments were conducted in 96-well plates for 24 h in 37 °C water-jacketed CO₂ incubators (ThermoForma).

Toxicity Assay

Cell viability was determined by MTS reagent using the CellTiter96 assay (Promega), according to the manufacturer's protocol. The plate was read at 490 nm.

Cytokine Assays

Human IL-1β DuoSet was obtained from R&D Systems and ELISA assays performed according to the manufacturer's protocol. The plate was read at 490 nm.