Introduction

Arginases are trimeric metalloenzymes that catalyze the hydrolysis of L-arginine to form L-ornithine and urea, and play a key role in the urea cycle and the regulation of nitric oxide (NO) homeostasis in mammals. Arginases are present in two isoforms: arginase I, a cytosolic enzyme with highest abundance in the liver, and arginase II, a mitochondrial enzyme [1, 2]. However, the physiological functions of these two differently localized isoforms remain unknown [3]. Inhibition of human arginases has been proposed as a potential therapy for various illnesses, such as cardiovascular [4], anti-inflammatory [5], autoimmune [6], oncological [7], and infectious [8] diseases, characterized by abnormally high arginase activity or abnormally low NO synthase activity.

Several synthetic human arginase competitive inhibitors for clinical usage, such as L-nor-N-hydroxyarginine (nor-NOHA) [9, 10] and 2(S)-amino-6-boronohexanoic acid (ABH) [11], as well as CB-1158 (numidargistat), an ABH homologue, have been developed from L-N-hydroxyarginine (NOHA) [9, 10], an intermediate in NO synthesis [3]. NOHA [9], nor-NOHA [12], and ABH [13] exert competitive inhibition by binding to the substrate as well as the bimanganese-binding sites, at the active center of arginases [3]. Currently, CB-1158 is in Phase I/II clinical trials as a drug candidate for cancer therapy [14]. However, although such recognition mechanisms may be required for powerful arginase inhibitory activity, mechanistically distinct inhibitor(s) might also be necessary to develop anti-arginase drug(s) that may be used as drug therapies with fewer side effects.

Several arginase inhibitors from natural resources have been described. Piceatannol, a stilbene, was reported as a selective natural inhibitor for human arginase I [15]. Apart from human arginase inhibitors, flavonoids such as quercetin, fisetin, and kaempferol were reported as inhibitors of bovine liver and Leishmania arginases [3, 16]. In addition, obacunone [17], a limonoid, sauchinone [18], a lignan, and salvianolic acid B [19], a caffeic acid derivative, were reported to inhibit murine liver lysate arginases I and II. However, these natural inhibitors have not been successfully forwarded to clinical settings, due to their poor pharmacokinetics profiles [20]. In the course of our search for a natural arginase inhibitor, we found that a 70% ethanol extract of Curcuma comosa rhizomes from Myanmar exhibited this activity. C. comosa is an herb belonging to the Zingiberaceae family and is distributed in tropical and subtropical areas of Asia, such as Thailand, Indonesia, Malaysia, and Myanmar. The rhizome of C. comosa, known as Sa-nwin-ga in Myanmar, is used in Myanmar traditional medicine for treating headaches, diabetes mellitus, and hypertension [21]. The C. comosa rhizome extracts reportedly showed estrogenic [22], antioxidant [23], antiallergic [24], and anti-inflammatory [25] activities. Previous phytochemical investigations indicated that many interesting chemical constituents, including sesquiterpenes [26], labdane diterpenes [27], flavonoid glycosides [28], and diarylheptanoids [29], are produced by this plant. Assay-guided fractionation of the 70% ethanol extract of the C. comosa rhizomes led to the isolation of three known guaiane sesquiterpenoids, alismoxide (1), with selective arginase I inhibitory activity, as well as 7α,10α-epoxyguaiane-4α,11-diol (2) and guaidiol (3) (Fig. 1). Herein, we report the isolation and structural elucidations of 1‒3, as well as their inhibitory activities against human arginase.

Fig. 1
figure 1

Structures of 1‒3 isolated from C. comosa rhizomes

Full size image

Materials and methods

Chemicals and reagents

Nor-NOHA was purchased from Bachem AG (Bubendorf, Switzerland). Unless otherwise specified, all chemical reagents were purchased from Fujifilm Wako Pure Chemical (Osaka, Japan).

General experimental procedures

NMR spectra were recorded on an ECX400P spectrometer (JEOL, Tokyo, Japan). Chemical shift values are expressed in δ (ppm), based on the δ residuals of CDCl3 at 7.26 for 1H NMR and 77.0 for 13C NMR. HR–ESI–MS data were obtained with an LC–MS–IT–TOF spectrometer (Shimadzu, Kyoto, Japan). Open column chromatography was performed with normal-phase silica gel (silica gel 60N, spherical, neutral, 40–50 μm) (Kanto Chemical, Tokyo, Japan) and reverse phase silica gel (Cosmosil 75C18-OPN) (Nacalai Tesque, Kyoto, Japan). Analytical TLC was performed on pre-coated silica gel 60 F254 plates and RP-18 F254 plates (Merck, Billerica, MA, USA), with spot detection by visualization with a UV lamp (254 and 365 nm), as well as spraying with a p-anisaldehyde stain solution and heating at 120 °C for 10 min. An SH-1200 microplate reader (Corona Electric, Hitachinaka, Japan) was used to measure the absorbance of the urea produced in the arginase inhibition activity assay. Specific optical rotations were measured on a P2100 polarimeter (JASCO, Tokyo, Japan). A MiniAmp thermal cycler (Thermo Fisher, Waltham, MA, USA) was used to incubate protein assays.

Plant materials

The C. comosa rhizomes were collected from Maubin Township, Ayeyarwady Division, Myanmar, in 2019 and positively identified by Dr. New Ni Tun, a lecturer at the Department of Botany, University of Yangon (Myanmar). A voucher specimen (TMPW 31707) was deposited in the Museum for Materia Medica, Analytical Research Center for Ethnomedicines, Institute of Natural Medicine, University of Toyama, Japan.

Extraction and isolation

The dried C. comosa rhizomes (1.6 kg) were extracted with 70% aqueous EtOH (3.0 L × 5) by sonication (90 min each) to afford the extract. The extract (252.5 g) was evaporated under vacuum, and the remaining aqueous residue was successively partitioned into n-hexane (31.0 g), CHCl3 (67.0 g), MeOH (28.0 g), and H2O extracts (126.5 g). The n-hexane extract was separated on a Cosmosil 75C18-OPN column, using a H2O/MeOH (2:1, 1:1, 1:3, 1:5, and 1:7) solvent system, to give five fractions: F1 (5.2 g), F2 (8.7 g), F3 (2.7 g), F4 (11.0 g), and F5 (3.3 g). The F1 fraction was subjected to normal phase silica gel open column chromatography, using a CHCl3/MeOH solvent system (9.9:0.1, 4.9:0.1, 9.5:0.5, 9:1, 8:2, 7:3, 6:4, 5:5, 4:6, and 0:10) with increasing polarity to afford 1 (14 mg). Fraction F2 was chromatographed on an open silica gel column, which was eluted with a hexane/AcOEt solvent system (9.5:0.5, 7:3, 5:5, 3:7, and 1:9) to obtain 2 (8 mg). Fraction F3 was chromatographed on a Cosmosil 75C18-OPN with a H2O/MeOH (2:1, 1:1, 1:3, 1:5, 1:7, and 1:9) solvent system to afford 3 (35 mg).

Construction of expression plasmids, and expression and purification of human arginases I and II

The cDNAs encoding human arginases I (chromosome 6q23) and II (chromosome 14q24) [30, 31] were purchased from Integrated DNA Technologies (Coralville, IA, USA). The open reading frames (ORFs) of the human arginase I and II cDNAs were amplified by PCR, using the primer pairs shown in Table S3 (BamHI and SalI sites are underlined and bolded, respectively). Each resultant PCR amplification product was gel-purified and then inserted into the BamHI/SalI sites of the expression vector pQE-80L for production as an N-terminal hexahistidine-tag fusion protein, using an In-Fusion cloning kit (Takara Bio, Otsu, Japan). After sequence confirmation, the arginase I and II expression plasmids were each transformed into Escherichia coli M15 (pREP4) competent cells. The E. coli M15 (pREP4) cells harboring arginases I and II were grown at 37 °C in Luria–Bertani (LB) medium, supplemented with 100 μg mL−1 ampicillin and 25 μg mL−1 kanamycin. When the OD600 reached 0.6, recombinant protein expression was induced by adding isopropyl D-thiogalactopyranoside to a final concentration of 0.5 mM, and then each culture was incubated overnight at 16 °C. The cell pellets were collected by centrifugation at 6500 rpm for 15 min at 4 °C and stored at − 80 °C. The pellets were resuspended in 50 mM Tris–HCl (pH 8.0), containing 100 mM NaCl and 5% (v/v) glycerol (buffer A), placed on ice, and sonicated intermittently every 10 s, for a total of 30 min. The lysates were centrifuged at 9000 rpm for 20 min. The supernatants were collected and then loaded on a Ni Sepharose 6 Fast Flow open column (Cytiva, Tokyo, Japan) equilibrated with buffer A. The column was washed with buffer A containing 50 mM imidazole, and arginases I and II were eluted with buffer A containing 500 mM imidazole. The collected proteins were concentrated and then purified by gel filtration chromatography on a HiLoad 16/60 Superdex 200 (Cytiva) column. Finally, the purified enzymes were concentrated to 10 mg mL−1 in 20 mM Tris–HCl (pH 8.0), containing 100 mM NaCl, 5% (v/v) glycerol, and 1 mM DTT.

Arginase inhibitory activity

Arginase inhibitory activities were evaluated using a colorimetric assay to quantify urea production, according to the previously described method with slight modifications [32]. Compounds 1‒3 were dissolved in DMSO. In brief, arginase I and arginase II were incubated at 55 °C for 10 min for activation, and then 0.2 ng of arginase I or II was added into a final volume of 35 µL reaction buffer (28 mM Tris–HCl, pH 7.5) containing 4.3 mM MnCl2, 14.3 mM L-arginine, 0.015% bovine serum albumin (for arginase I), and inhibitors 1‒3 or DMSO as a blank. After an incubation for 30 min at 37 °C, the reaction was stopped by adding 60 µL of H2SO4/H3PO4/H2O (1:3:7) on ice, and then 5 µL of α-isonitrosopropiophenone (5% in absolute ethanol) was added. Subsequently, the resultant reaction mixture was heated at 95 °C for 45 min in the dark, and then centrifuged for 10 min to remove the precipitates. The supernatants were transferred to a 96-well plate, and after the microplate was shaken for 2 min, the absorbance at 550 nm was read. Arginase inhibitory activity is expressed as the percentage of inhibition relative to “100% arginase activity”. A Dixon plot analysis was performed to determine the type of inhibition of 1 against arginase I. Data were collected in triplicate, at concentrations ranging from 7.5 to 120 µM for 1 and 2.5 to 10 mM for L-arginine.

Docking simulation

The protein molecule used in the docking simulation was arginase I, PDB ID: 6Q9P. The three-dimensional model of 1 was generated using the Avogadro 1.2 program, and the compound was docked into the arginase I structure by Autodock Vina 1.0.2 [33]. The side chains of the Arg21, Asn130, Ser137, His141, Asp181, Asp183, Glu186, Thr246, and Glu277 residues were set as flexible. The binding affinity of 1 was calculated to be − 6.9 kcal mol−1.

Results and discussion

The 70% ethanol extract of C. comosa rhizomes showed inhibitory activities against human arginase I with an IC50 value of 61.4 μg mL−1 and arginase II with an IC50 value of 80.9 μg mL−1. Subsequently, the n-hexane-soluble fraction of C. comosa rhizomes was subjected to normal and reverse phase silica gel open chromatography to yield three guaiane sesquiterpenoids, 1‒3. Comparisons of spectroscopic data of 1‒3 (Tables S1 and S2 and Figs. S1‒S9) with those reported in the literature unambiguously identified the structure of 1 as alismoxide (1) with a trans-fused ring system [34, 35], 2 as 7α,10α-epoxyguaiane-4α,11-diol (2) [36], and 3 as guaidiol (3) [37] with a cis-fused ring system. Furthermore, the absolute configuration of 1 was determined as 1R,4S,5R,10R-1, based on the well-matched optical rotation value of 1 {\([\upalpha ]_{{\text{D}}}^{25}\) + 5.0° (c 0.9, MeOH)} with the previously reported values of 1R,4S,5R,10R-1 in the literature {\([\upalpha ]_{{\text{D}}}^{25}\) + 5.2° (c 0.5, MeOH)} [34, 38].

According to the previous report, the inhibitory activities of 1‒3 against human arginases I and II were assessed by a colorimetric method using α-isonitrosopropiophenone [32]. Nor-NOHA was used as the positive control, and showed IC50 values of 8.7 ± 0.6 μM and 12.8 ± 1.3 μM against human arginases I and II, respectively, consistent with the previously reported values of nor-NOHA in the literature [human arginase I (IC50 values 1.4‒6.0 μM) and human arginase II (IC50 value: 1.3 μM)] [12, 39]. Compound 1 inhibited human arginase I with an IC50 value of 30.2 ± 2.7 μM (Table 1), which was 3–20-fold weaker than nor-NOHA (IC50 value of 8.7 μM) and ABH (IC50 value of 1.45 μM) [11], while it was 2.3 times stronger than the natural product inhibitor, piceatannol (IC50 value: 69 μM) [15]. In contrast, neither 2 nor 3 was active against human arginase I, even at a 300 μM concentration, suggesting that 1 could be a valuable natural human arginase I inhibitor as compared with the previously reported ones. Further assays revealed no activities of 1‒3 against human arginase II, even at 300 μM concentrations, suggesting that 1 could be a selective inhibitor of human arginase I.

Table 1 Inhibitory activities of 1‒3 against human arginases I and II
Full size table

To the best of our knowledge, this is the first demonstration of human arginase I inhibitory activity by a sesquiterpenoid. To clarify the mode of inhibition by 1, we performed kinetic studies of human arginase I. First, we checked whether the human arginase I used in this experiment showed activity at a level consistent with previous reports. The Michaelis–Menten plot analysis revealed KM, kcat, and kcat/KM values of 11.6 mM, 20.9 min−1, and 1.8 × 103 M−1 min−1 against human arginase I with respect to L-arginine (Fig. S10), which were comparable to the previously reported kinetics parameters of human arginase I heterologously expressed in E. coli (KM = 2.3 mM, kcat = 5.0 min−1, and kcat/KM = 2.1 × 103 M−1 min−1) [40]. Subsequently, the Dixon plot analysis revealed that 1 is a competitive inhibitor of L-arginine with a KI value of 136.5 μM, which is roughly 85-fold higher than the substrate affinity (11.6 mM), suggesting that 1 selectively inhibited human arginase I by binding to the L-arginine binding site (Fig. 2). The KI value of 1 was lower than those of the synthetic inhibitors, nor-NOHA (KI = 0.50 μM for arginase I) and ABH (KI = 0.11 μM for arginase I) [11], while it was similar to that of the natural product inhibitor, piceatannol (KI = 136 μM for arginase I) [15].

Fig. 2
figure 2

Dixon plot analysis of the binding mode of 1 with arginase I

Full size image

A structural feature of the guaiane sesquiterpenoid is the presence of a bicyclo[5.3.0]decane skeleton. Comparisons of the structures of 1‒3 suggested that the significantly different functionalities between the active compound 1 and the non-active compounds 2 and 3 are the presence of the 7-isopropyl moieties attached to the olefinic groups at C-6 and C-7 in 1. These moieties in 1 may account for the acquisition of the selective inhibitory activity of 1 against human arginase I. However, comparisons of the three-dimensional structures of 1‒3 and enantiomers of 2 and 3 indicated that 1 possessed a flatter structure across the board, as compared with those of 2 and 3 and their ent-forms, due to the trans-fused ring system in 1 (Fig. 3). Thus, rather than the differences of the functionalities between 1 and 2/3, the totally different three-dimensional structure of 1 might be the major factor, leading to the selectively competitive inhibition against human arginase I in a competitive manner with respect to L-arginine.

Fig. 3
figure 3

Three-dimensional structures of 1, 2, 3, ent-2, and ent-3

Full size image

To further characterize the interactions between human arginase I and 1 in the L-arginine-binding site, we performed a docking simulation for 1 and human arginase I (PDB ID: 6Q9P) [14]. The docking simulation predicted that 1 is accommodated in an L-arginine-binding site by hydrogen bonds between the side chains of His141, Asp183, and Glu186 in human arginase I, with a binding affinity of − 6.9 kcal mol−1 (Fig. 4A, B). However, 1 did not indicate any interaction with manganese, in contrast with nor-NOHA and ABH, which exhibit potent inhibitory activities by coordinating with not only L-arginine, but also the two manganese atoms in the active center (Fig. 4C, D) [3]. The lower inhibitory activity of 1 than the synthetic inhibitors such as nor-NOHA and ABH would be caused by fewer interactions at the catalytic center, as compared to those of the synthetic inhibitors. On the other hand, the docking studies of 2 and 3 to human arginase I are not available, since these compounds were non-active for arginase I, which indicates that both compounds lack any interaction with this enzyme. However, considering the docking result of 1 and the three-dimensional structures of 2 and 3 (Fig. 3), totally different three-dimensional structures of 2 and 3 compared with that of 1 would prevent not only the hydrogen-bond formations with His141, Asp183, and Glu186 observed in the docking study between 1 and arginase I, but also their binding to the active-site cavity of human arginase I.

Fig. 4
figure 4

Docking model of alismoxide (1) to human arginase I (PDB ID: 6Q9P). Close-up view of the substrate binding site of human arginase I complexed with A 1, B 1 and L-ornithine (PDB ID: 3GMZ), C 1 and nor-NOHA (PDB ID: 3KV2), and D 1 and ABH. Compound 1 is shown as a green stick model. E Superimposition of human arginase I complexed with 1 and human arginase II (PDB ID: 4HZE). Arginases I and II are highlighted with white and pink colors, respectively

Full size image

One of the interesting points is that 1 showed tenfold stronger inhibitory activity against human arginase I (30.2 ± 2.7 μM) than human arginase II (> 300 μM). The detailed catalytic mechanisms for the distinct catalytic properties between arginases I and II, such as kinetic parameters, have not yet been determined. However, considering that the different residues in the active-sites between human arginases I and II are only Thr127 and Thr136 in human arginase I and Ala146 and Ser155 in human arginase II [41, 42], the selective inhibitory activity of 1 for human arginase I may be derived from the C-10 hydroxyl group of 1 and Glu186, which form a hydrogen bond network with Thr127 and Asp183 in human arginase I (Fig. 4E). However, to fully understand the distinct catalytic properties between arginases I and II, further investigations are still needed to elucidate the selective inhibitory mode of 1 against human arginase I.

In conclusion, we isolated three guaiane sesquiterpenoids, alismoxide (1), 7α,10α-epoxyguaiane-4α,11-diol (2), and guaidiol (3) from C. comosa rhizomes. Our study revealed that 1 is a selective arginase I inhibitor in a competitive manner with respect to L-arginine. Thus, this study provided new insights into a naturally occurring arginase I inhibitor.