New Glutamine-Containing Azaphilone Alkaloids from Deep-Sea-Derived Fungus Chaetomium globosum HDN151398

Three new azaphilone alkaloids containing glutamine residues, namely N-glutarylchaetoviridins A–C (1–3), together with two related compounds (4 and 5) were isolated from the extract of Chaetomium globosum HDN151398, a fungus isolated from a deep-sea sediment sample collected in South China Sea. Their structures were elucidated on the basis of extensive 1D and 2D NMR as well as HRESIMS spectroscopic data and chemical analysis. N-glutarylchaetoviridins A–C (1–3) represent the first class of chaetoviridins characterized by embedded glutamate residues. Amino acids incubation experiments produced five azaphilone laden different amino acids residues (6–10) which indicated that this method can enhanced the structural diversity of this strain by culturing with amino acids. Cytotoxicity of the isolated compounds were evaluated against a panel of human cancer cell lines.


Introduction
Deep-sea-derived microorganisms have proven to be a prolific source of secondary metabolites with an ample variety of captivating chemical structures and diverse pharmacological properties [1,2]. In our recent search for bioactive secondary metabolites from marine-derived fungi, a fungal strain Chaetomium globosum HDN151398, isolated from a deep-sea sediment sample (depth 2476 m) collected from South China Sea, was selected for its intriguing HPLC-UV profile ( Figure S1) and significant crude extract cytotoxic activity (69% inhibition of K562 cells at the concentration of 100 µg/mL). Chemical investigation of the organic extract of the fungus led to the isolation of three new glutamine-containing azaphilone alkaloids, N-glutarylchaetoviridins A-C (1-3) together with two known chaetoviridins (4 and 5).
Azaphilones are a family of structurally erratic fungal pigmented polyketides with pyrone-quinone structures containing a highly oxygenated bicyclic core and a chiral quaternary center [3,4]. The oxygen atom in the pyran chromophore of the azaphilones could be biosynthetically replaced by nitrogen atom in the presence of primary amines and the colour of the pigment shifted to red accordingly [4]. Recently, azaphilones have been recognized as a unique family of secondary metabolites with diverse bioactivities including antimicrobial [5], cytotoxic [6], anti-inflammatory [7] and other activities [8][9][10], which have provoked enormous attention of scientists for biosynthesis [11] and chemical synthesis studies [12]. In the present work, we report the isolation, structure elucidation and biological activities of the previously unreported azaphilones (1-3) (Figure 1) from the strain Chaetomium globosum HDN151398 as well as the incubation experiments with different amino acids to produce more diverse analogues.
Mar. Drugs 2019, 17, x FOR PEER REVIEW 2 of 11 The oxygen atom in the pyran chromophore of the azaphilones could be biosynthetically replaced by nitrogen atom in the presence of primary amines and the colour of the pigment shifted to red accordingly [4]. Recently, azaphilones have been recognized as a unique family of secondary metabolites with diverse bioactivities including antimicrobial [5], cytotoxic [6], anti-inflammatory [7] and other activities [8][9][10], which have provoked enormous attention of scientists for biosynthesis [11] and chemical synthesis studies [12]. In the present work, we report the isolation, structure elucidation and biological activities of the previously unreported azaphilones (1-3) (Figure 1) from the strain Chaetomium globosum HDN151398 as well as the incubation experiments with different amino acids to produce more diverse analogues.

Results and Discussion
Compound 1 was isolated as a dark red powder with the molecular formula   1, 199.3). Careful comparison of the 13 C NMR data of 1 and those of chaetoviridin A [13] revealed that they share a similar pyrone-quinone-containing skeleton. The main differences between 1 and chaetoviridin A were the chemical shifts at C-1 (δC 136.9 versus δC 151.5) and C-3 (δC 148.4 versus δC 157.1), and a group of extra resonances in 1 which were attributed to a methylated glutamic acid moiety. The pyrone-quinone core structure was further verified by the 1 H-1 H COSY cross peaks from H-6″ to H-5″, from H-5″ to H-4″, from H-4″ to H-7″, from H-13 to H-12, from H-12 to H-11, from H-11 to H-10, from H-10 to H-9 and from H-14 to H-11 as well as HMBC correlations from H-1 to C-8 and C-3, from H-9 to C-3 and C-4, from H-4 to C-5 and C-4a, from H-15 to C-6, C-7, and C-8 and from H-4″ to C-3″and C-2″( Figure 2). The methylated glutamic acid moiety was deduced by the COSY correlations from H-2′ to H-3′and from H-3′ to H-4′ as well as the HMBC correlations from H-6′ and H-4′ to C-5′, from H-7′ and H-2′ to C-1′. Based on the variation between the chemical shifts at C-1 and C-3 and taking the molecular formula into account, a nitrogen atom, instead of an oxygen atom, was placed in position 2. Further HMBC cross peaks from H-2′ to C-1 (δ 136.9) and C-3 (δ 148.4) (Figure 2) attached the dimethylglutarate moiety to N-2 of 1,4-hydropyridinequinone scaffold moiety. As this compound has never been previously reported, it was named Nglutarylchaetoviridin A.   1, 199.3). Careful comparison of the 13 C NMR data of 1 and those of chaetoviridin A [13] revealed that they share a similar pyrone-quinone-containing skeleton. The main differences between 1 and chaetoviridin A were the chemical shifts at C-1 (δ C 136.9 versus δ C 151.5) and C-3 (δ C 148.4 versus δ C 157.1), and a group of extra resonances in 1 which were attributed to a methylated glutamic acid moiety. The pyrone-quinone core structure was further verified by the 1 H-1 H COSY cross peaks from H-6" to H-5", from H-5" to H-4", from H-4" to H-7", from H-13 to H-12, from H-12 to H-11, from H-11 to H-10, from H-10 to H-9 and from H-14 to H-11 as well as HMBC correlations from H-1 to C-8 and C-3, from H-9 to C-3 and C-4, from H-4 to C-5 and C-4a, from H-15 to C-6, C-7, and C-8 and from H-4" to C-3"and C-2" (Figure 2). The methylated glutamic acid moiety was deduced by the COSY correlations from H-2 to H-3 and from H-3 to H-4 as well as the HMBC correlations from H-6 and H-4 to C-5 , from H-7 and H-2 to C-1 . Based on the variation between the chemical shifts at C-1 and C-3 and taking the molecular formula into account, a nitrogen atom, instead of an oxygen atom, was placed in position 2. Further HMBC cross peaks from H-2 to C-1 (δ 136.9) and C-3 (δ 148.4) ( Figure 2) attached the dimethylglutarate moiety to N-2 of 1,4-hydropyridine-quinone scaffold moiety. As this compound has never been previously reported, it was named N-glutarylchaetoviridin A.   The geometrical configuration of the double bond between C-9 and C-10 was inferred to be trans from the coupling constants of the olefinic protons (J9,10 = 15.5 Hz). The relative configuration of 1 was determined based on a combination of NOESY correlations and comparison of its NMR data with those of chaetoviridin A. The stereochemistry of 1 established based on the NOESY correlations from H-1 to H-4″, from H-1 to H-6″, and from H-8″ to H-7″, the similar electronic circular dichroism (ECD) curves of 1 ( Figure 3) and chaetoviridin A [13,14], and the co-isolation of biogenetically related compounds 4 and 5 which also shared the same chiral centres. The absolute configuration of C-7 and C-11 were further confirmed by Steyn and Vleggaar's method [15] and the degradation of 1 [16], respectively. The ECD spectrum of 1 (Δε387 −10.1; Figure 3) revealed that the absolute configuration of C-7 is S according to Steyn and Vleggaar's CD method [15]. Compound 1 was degraded by 5% potassium hydroxide to afford a carboxylic acid ( Figure 4) which was identified as (4S)-2E-4methylhex-2-enoic acid by comparison of spectral data and specific optical rotation with the authentic sample. The configuration of the L-glutamate moiety in 1 was determined by the advanced Marfey's method [17] by comparison of the retention time and mass data of the hydrolysis product with those of D/L-glutamate standards by HPLC ( Figure S30). Accordingly, the absolute configuration of 1 was concluded to be 7S, 11S, 2′S, 4″S, and 5″R, respectively. The geometrical configuration of the double bond between C-9 and C-10 was inferred to be trans from the coupling constants of the olefinic protons (J 9,10 = 15.5 Hz). The relative configuration of 1 was determined based on a combination of NOESY correlations and comparison of its NMR data with those of chaetoviridin A. The stereochemistry of 1 established based on the NOESY correlations from H-1 to H-4", from H-1 to H-6", and from H-8" to H-7", the similar electronic circular dichroism (ECD) curves of 1 ( Figure 3) and chaetoviridin A [13,14], and the co-isolation of biogenetically related compounds 4 and 5 which also shared the same chiral centres. The absolute configuration of C-7 and C-11 were further confirmed by Steyn and Vleggaar's method [15] and the degradation of 1 [16], respectively. The ECD spectrum of 1 (∆ε387 −10.1; Figure 3) revealed that the absolute configuration of C-7 is S according to Steyn and Vleggaar's CD method [15]. Compound 1 was degraded by 5% potassium hydroxide to afford a carboxylic acid ( Figure 4) which was identified as (4S)-2E-4-methylhex-2-enoic acid by comparison of spectral data and specific optical rotation with the authentic sample. The configuration of the L -glutamate moiety in 1 was determined by the advanced Marfey's method [17] by comparison of the retention time and mass data of the hydrolysis product with those of d/l-glutamate standards by  Figure S30). Accordingly, the absolute configuration of 1 was concluded to be 7S, 11S, 2 S, 4"S, and 5"R, respectively.    13 C NMR data of 2 were very similar to those of 1, while the main differences were the absence of three methyl and one proton signals (δH 3.83, 3.72, 2.93, and 3.71, respectively) and the downfield shift of H-7″ (δ 1.87), H-6″ (δ 1.89) and H-5″ (δ 6.64), suggesting that the single bond between C-4″ (δ 138.0) and C-5″ (δ 146.7) was oxidized to a double bond. This postulation was confirmed by COSY correlations from H-5″ to H-6″ and HMBC cross peaks from H-7″to C-3″and C-4″, and from H-5″ to C-3″. Further 2D NMR analysis ( Figure 2) verified the planar structure as shown in Figure 1 and we named it N-glutarylchaetoviridin B.

Results and Discussion
The coupling constants of the olefinic protons (J9,10 = 14.2 Hz) indicated the trans configuration of the double bond (Δ9,10). The NOESY correlations between H-6″/H-7″ demonstrated that the double bond between C-4″ and C-5″ was in E configuration. The amino acid residue in 2 was identified as Lglutamate by the advanced Marfey's method [17]. The CD spectrum of 2 (Δε387 −10.8; Figure 5) revealed the 7S absolute configuration according to Steyn and Vleggaar's CD method [15]. The absolute configuration at C-11 was determined as S by the degradation of 2. Thus, the absolute configurations of C-7, 11 and 2′ of 2 were assigned as 7S, 11S, and 2′S.    8,190.6) carbons. The 1 H and 13 C NMR data of 2 were very similar to those of 1, while the main differences were the absence of three methyl and one proton signals (δH 3.83, 3.72, 2.93, and 3.71, respectively) and the downfield shift of H-7″ (δ 1.87), H-6″ (δ 1.89) and H-5″ (δ 6.64), suggesting that the single bond between C-4″ (δ 138.0) and C-5″ (δ 146.7) was oxidized to a double bond. This postulation was confirmed by COSY correlations from H-5″ to H-6″ and HMBC cross peaks from H-7″to C-3″and C-4″, and from H-5″ to C-3″. Further 2D NMR analysis (Figure 2) verified the planar structure as shown in Figure 1 and we named it N-glutarylchaetoviridin B.
The coupling constants of the olefinic protons (J9,10 = 14.2 Hz) indicated the trans configuration of the double bond (Δ9,10). The NOESY correlations between H-6″/H-7″ demonstrated that the double bond between C-4″ and C-5″ was in E configuration. The amino acid residue in 2 was identified as Lglutamate by the advanced Marfey's method [17]. The CD spectrum of 2 (Δε387 −10.8; Figure 5) revealed the 7S absolute configuration according to Steyn and Vleggaar's CD method [15]. The absolute configuration at C-11 was determined as S by the degradation of 2. Thus, the absolute configurations of C-7, 11 and 2′ of 2 were assigned as 7S, 11S, and 2′S.    13 C NMR data of 2 were very similar to those of 1, while the main differences were the absence of three methyl and one proton signals (δ H 3.83, 3.72, 2.93, and 3.71, respectively) and the downfield shift of H-7" (δ 1.87), H-6" (δ 1.89) and H-5" (δ 6.64), suggesting that the single bond between C-4" (δ 138.0) and C-5" (δ 146.7) was oxidized to a double bond. This postulation was confirmed by COSY correlations from H-5" to H-6" and HMBC cross peaks from H-7"to C-3"and C-4", and from H-5" to C-3". Further 2D NMR analysis (Figure 2) verified the planar structure as shown in Figure 1 and we named it N-glutarylchaetoviridin B.
The coupling constants of the olefinic protons (J 9,10 = 14.2 Hz) indicated the trans configuration of the double bond (∆ 9,10 ). The NOESY correlations between H-6"/H-7" demonstrated that the double bond between C-4" and C-5" was in E configuration. The amino acid residue in 2 was identified as l-glutamate by the advanced Marfey's method [17]. The CD spectrum of 2 (∆ε387 −10.8; Figure 5) revealed the 7S absolute configuration according to Steyn and Vleggaar's CD method [15]. The absolute configuration at C-11 was determined as S by the degradation of 2. Thus, the absolute configurations of C-7, 11 and 2 of 2 were assigned as 7S, 11S, and 2 S.  Mass spectrometric data as well as the key HMBC correlations from H-6′ (δ 3.67) to C-5′ (δ 172.1) and H-7′ (δ 3.77) to C-1′ (δ 168.3) confirmed that 3 is a 6′,7′-dimethoxyl analogue of 2, and was named Nglutarylchaetoviridin C. As the ECD curve of 3 is very similar to that of 2 ( Figure 5) it was concluded that 3 has the same stereochemistry as 2.
Two previously reported compounds, chaetomugilin A (4) [18] and chaetomugilin C (5) [18] were also isolated and their identity was proved by comparison of their NMR and MS data with those reported in the literature. Mass spectrometric data as well as the key HMBC correlations from H-6 (δ 3.67) to C-5 (δ 172.1) and H-7 (δ 3.77) to C-1 (δ 168.3) confirmed that 3 is a 6 ,7 -dimethoxyl analogue of 2, and was named N-glutarylchaetoviridin C. As the ECD curve of 3 is very similar to that of 2 ( Figure 5) it was concluded that 3 has the same stereochemistry as 2.
Two previously reported compounds, chaetomugilin A (4) [18] and chaetomugilin C (5) [18] were also isolated and their identity was proved by comparison of their NMR and MS data with those reported in the literature.  (Table S1). Compounds 3, 4, and 5 showed a broad spectrum of cytotoxic activity. Among them, 3 showed significant cytotoxic activity against MGC-803 and HO8910 with IC 50 values of 6.6 and 9.7 µM, respectively.
Inspired by the fact that azaphilones have a capacity to incorporate amino acids, five different amino acids (l-tryptophan, l-tyrosine, l-histidine, l-alanine, l-glycine) were added to the culture medium in order to produce more diverse analogues. All the molecular ion peaks of the proposed structures could be easily detected by LC-MS ( Figure 6). Furthermore, the structures of 6-10 ( Figure 7) were further confirmed by both (+)-HRESIMS and NMR data ( Table 3, Figures S35-S44). Consequently, these results further validate the property of azaphilones to combine with amino acids and to produce more diverse compounds.   (Table S1). Compounds 3, 4, and 5 showed a broad spectrum of cytotoxic activity. Among them, 3 showed significant cytotoxic activity against MGC-803 and HO8910 with IC50 values of 6.6 and 9.7 μM, respectively.
Inspired by the fact that azaphilones have a capacity to incorporate amino acids, five different amino acids (L-tryptophan, L-tyrosine, L-histidine, L-alanine, L-glycine) were added to the culture medium in order to produce more diverse analogues. All the molecular ion peaks of the proposed structures could be easily detected by LC-MS ( Figure 6). Furthermore, the structures of 6-10 ( Figure  7) were further confirmed by both (+)-HRESIMS and NMR data (Table 3, Figure S35-S44). Consequently, these results further validate the property of azaphilones to combine with amino acids and to produce more diverse compounds.      (Table S1). Compounds 3, 4, and 5 showed a broad spectrum of cytotoxic activity. Among them, 3 showed significant cytotoxic activity against MGC-803 and HO8910 with IC50 values of 6.6 and 9.7 μM, respectively.
Inspired by the fact that azaphilones have a capacity to incorporate amino acids, five different amino acids (L-tryptophan, L-tyrosine, L-histidine, L-alanine, L-glycine) were added to the culture medium in order to produce more diverse analogues. All the molecular ion peaks of the proposed structures could be easily detected by LC-MS ( Figure 6). Furthermore, the structures of 6-10 ( Figure  7) were further confirmed by both (+)-HRESIMS and NMR data (Table 3, Figure S35-S44). Consequently, these results further validate the property of azaphilones to combine with amino acids and to produce more diverse compounds.

Isolation
The whole fermentation broth (40 L) was filtered through muslin cloth to separate the supernatant from the mycelia. The supernatant was extracted with EtOAc (3 × 40 L), and the mycelia were homogenized and extracted with MeOH (3 × 10 L). The EtOAc and MeOH solutions of the supernatant and mycelia were combined and evaporated under reduced pressure to give a crude. The extract (30.0 g) was fractioned by VLC of silica gel using a step gradient elution DCM-MeOH (100:0 to 0:100) to give ten fractions (Fr.1 to Fr.10). Fr.

Absolute Configuration of Amino Acids
Compounds 1-3 were hydrolyzed in 6 N HCl at 60 • C overnight. The solution was dried under a stream of N 2 and dissolved in H 2 O (100 µL). The acid hydrolysates of 1−3 were dissolved in H 2 O (50 µL) separately, and then 0.25 µM FDAA in 100 µL of acetone was added, followed by 1 N NaHCO 3 (25 µL). The mixtures were heated for 1 h at 43 • C. After cooling to room temperature, the reaction was quenched by the addition of 2 N HCl (25 µL). Finally, the resulting solution was filtered through a small 4.5 µm filter and stored in the freezer until ready for HPLC analysis. Amino acid standards were derivatized with FDAA in a similar manner. The resulting FDAA derivatives of compounds 1−3, l-and d-glutamate were separately analyzed by reversed-phase HPLC (5 × 250 mm YMC C18 column, 5 µm, with a linear gradient of MeCN (A) and 0.05% aqueous TFA (B) from 5% to 55% A over 55 min at a flow rate of 1 mL/min, UV detection at 320 nm). Each chromatographic peak was identified by comparing its retention time with the FDAA derivatives of the l-and d-amino acid standards.  Figures S30 and S31), establishing the S configuration for all the glutamic acid residues [17,19].

Degradation of 1-3 by Potassium Hydroxide
Compounds 1-3 (3.0 mg) were separately dissolved in 5% aq. potassium hydroxide (5 mL) and the reaction mixture was stirred for 3 h at 100 • C. Then, the reaction mixture was extracted with CHCl 3 (5 mL). The water layer was adjusted to pH 3.0 with 9% sulfuric acid and re-extracted with petroleum ether (5 mL). The organic extract was concentrated to dryness in vacuo. The residue was purified by HPLC using MeCN-H 2 O gradient (30:70 to 100:0 in 45 min) as the eluent to afford (4S)-2E-4-methylhex-2-enoic acid (0.1 mg, t R = 15 min). The physicochemical properties of this carboxylic acid were identical to the authentic sample [16].

Amino Acid Incubation Experiment
The fungus was cultured and subjected to a large-scale fermentation under the same protocol, stated above in the fermentation section. The only difference is that monosodium glutamate was replaced by five different amino acids. Dried extracts were dissolved in 1 mL of MeOH and analyzed

Conclusions
In summary, a series of azaphilones (1)(2)(3)(4)(5) were isolated from the deep-sea-derived fungus C. globosum HDN151398. Distinguished from the previously reported azaphilone derivatives, 1-3 belong to a new class of chaetoviridins which is linked to a glutamate residue, indicating that unique geographical features of deep-sea environment may promote the unique biogenetic and metabolic pathways of the microorganisms in which they inhabit. Compounds 3, 4, and 5 showed a broad spectrum of cytotoxicity, among which, 3 was active against MGC-803 and HO8910 with the IC 50 values of 6.6 and 9.7 µM, respectively. Amino acids feeding experiment showed that it is an effective method to increase structural diversity of azaphilones.
Author Contributions: The contributions of the respective authors are as follows: C.S. drafted the work. C.S., X.G., S.M., and L.Z. performed the fermentation, extraction, isolation, and structural elucidation of the constituents. J.P. was performed the biological evaluations. G.Y., Q.C., Q.G., T.Z., G.Z., and D.L. contributed to checking and confirming all of the procedures of the isolation and the structural elucidation. D.L. and T.Z. designed the study, supervised the laboratory work, and contributed to the critical reading of the manuscript. All the authors have read the final manuscript and approved the submission.