Isolation and Structure Determination of a Heat Shock Protein Inducer,Asparagus-Derived Proline-Containing 3-Alkyldiketopiperazines (Asparaprolines),From a Standardized Extract of Asparagus officinalis Stem

Abstract

ETAS^® has been developed from the stems of Asparagus officinalis L. as a functional ingredient for nutraceuticals. ETAS possesses heat shock protein 70 (HSP70) induction activity and may contribute to maintenance and improvement of health. Here, 3 compounds (1, 2, 3) were isolated from ETAS. The structures of 1, 2, and 3 were deduced by HREIMS and NMR spectroscopic data, and the compounds were identified as cyclo(l-Phe-l-Pro), cyclo(l-Tyr-l-Pro), and cyclo(l-Leu-l-Pro), respectively. Each compound contained a diketopiperazine ring derived from proline with an alkyl group at C-3; thus, we termed them asparagus-derived proline-containing 3-alkyldiketopiperazines (Asparaprolines). In an HSP70 mRNA induction assay in HL-60 cells, Asparaprolines significantly enhanced the expression of HSP70 mRNA compared with a control. To our knowledge, these results demonstrate for the first time that proline-containing diketopiperazines derived from natural amino acids exhibit HSP70 mRNA induction activity.

The heat shock response, a highly conserved mechanism in organisms ranging from yeast to human, is induced by extreme proteotoxic insults such as heat, oxidative stress, ultraviolet radiation, heavy metals, toxins, and bacterial infection. ^1,2 The heat shock response is mediated at the transcriptional level by heat shock elements (HSEs) present in the upstream region of genes encoding heat shock proteins (HSPs). Expression of an HSP is induced when a heat shock factor (HSF) binds to the HSE. Members of the HSP70 family, which represent some of the most ubiquitous molecular chaperones, are known to exert cytoprotective effects by enhancing cell viability and promoting protein damage repair. ³ Owing to their antiapoptotic and anti-inflammatory effects, HSPs are therapeutic targets; for example, geranylgeranylacetone, an inducer of HSP70, is used as antiulcer drug, ^4,5 while the induction of chaperones (including HSP70) may have therapeutic application in the prevention and treatment of disorders due to protein misfolding (eg, Alzheimer’s disease, Parkinson’s disease, Huntington’s disease, and cystic fibrosis). ^6,7

The heat shock pathway, incorporating HSP70 and HSF-1, also has a role in regulating sleep and circadian rhythm. ^{8 -11} As a result, we previously developed ONR-8, a dietary supplement containing a standardized extract of Asparagus officinalis stem (ETAS^®; Amino Up Co, Ltd, Sapporo, Japan), which is derived from stems of asparagus (A. officinalis L.) and has been demonstrated to have HSP70 induction activity. ^12,13 Furthermore, we demonstrated that ONR-8 improves sleep and modifies social jetlag among healthy volunteers. ^14,15 Ito et al ¹² reported that ETAS contains 5-hydroxymethyl-2-furfural (HMF) and its derivative, (S)-asfural, as active ingredients that induce HSP70 expression. In the process of ONR-8 development, however, we identified 3 further compounds with potential HSP70-inducing activity. Here, we report the isolation and structure determination of these compounds.

Initially, ETAS powder was dissolved in water and extracted with ethyl acetate. The extract was fractionated using HPLC to yield 3 compounds (compounds 1, 2, and 3).

Their molecular formulae were assigned as C₁₄H₁₆N₂O₂ (compound 1) based on the [M]⁺ ion peak at m/z 244.1205 (calcd 244.1212) in high-resolution electron impact mass spectrometry (HR-EI-MS); C₁₄H₁₆N₂O₃ (compound 2), based on the [M]⁺ ion peak at m/z 260.1159 (calcd 260.1161); and C₁₁H₁₈N₂O₂ (compound 3), based on the [M]⁺ ion peak at m/z 210.1360 (calcd 210.1368). The specific rotations of compounds 1, 2, and 3 were $[\alpha {]}_{\mathrm{D}}^{25}$ –113° (c = 0.326, CHCl₃/MeOH), $[\alpha {]}_{\mathrm{D}}^{25}$ –96° (c = 0.143, CHCl₃/MeOH), and $[\alpha {]}_{\mathrm{D}}^{25}$ –93° (c = 0.378, EtOH), respectively. The ¹H and ¹³C NMR spectral data of these compounds are summarized in Table 1.

Table 1

Spectrometric Data of Compounds 1, 2, and 3.

Compound 1	Compound 2	Compound 3
1H NMR (400 MHz, CD3OD)	1H NMR(400 MHz, CD3OD)	1H NMR (400 MHz, CDCl3)
δ 1.23 (1H, m)	δ 1.23 (1H, m)	δ 0.95 (3H, d, J = 6.8 Hz)
δ 1.79 (2H, m)	δ 1.80 (2H, m)	δ 1.00 (3H, d, J = 6.4 Hz)
δ 2.09 (1H, m)	δ 2.09 (1H, m)	δ 1.53 (1H, ddd, J = 4.8, 9.6, 14.4 Hz)
δ 3.16 (2H, d, J = 4.8 Hz)	δ 3.03 (1H, dd, J = 4.8, 14.0 Hz)	δ 1.89 (1H, m)
δ 3.36 (1H, td, J = 6.0, 12.4 Hz)	δ 3.08 (1H, dd, J = 4.8, 14.0 Hz)	δ 1.91 (1H, m)
δ 3.53 (1H, td, J = 8.4, 12.4 Hz)	δ 3.35 (1H, td, J = 6.8, 12.0 Hz)	δ 2.01 (1H, m)
δ 4.05 (1H, ddd, J = 2.0, 6.4, 10.8 Hz)	δ 3.55 (1H, td, J = 6.8, 12.0 Hz)	δ 2.06 (1H, m)
δ 4.43 (1H, dt, J = 2.0, 4.8 Hz)	δ 4.04 (1H, ddd, J = 1.6, 6.4, 10.8 Hz)	δ 2.12 (1H, m)
δ 7.20–7.30 (5H, m)	δ 4.36 (1H, dt, J = 1.6, 4.8 Hz)	δ 2.35 (1H, m)
	δ 6.70 (2H, d, J = 8.4 Hz)	δ 3.52–3.62 (2H, m)
	δ 7.04 (2H, d, J = 8.4 Hz)	δ 4.02 (1H, dd, J = 3.6, 9.6 Hz)
		δ 4.12 (1H, t, J = 8.4 Hz)
13C NMR (100 MHz，CD3OD)	13C NMR (100 MHz，CD3OD)	13C NMR (100 MHz，CDCl3)
δ 22.8	δ 22.8	δ 21.2
δ 29.4	δ 29.4	δ 22.7
δ 38.2	δ 37.7	δ 23.3
δ 46.0	δ 46.0	δ 24.6
δ 57.2	δ 57.9	δ 28.1
δ 60.1	δ 60.1	δ 38.6
δ 128.1	δ 116.2	δ 45.5
δ 129.5	δ 127.7	δ 53.4
δ 131.0	δ 132.2	δ 59.0
δ 137.4	δ 157.7	δ 166.2
δ 166.9	δ 167.0	δ 170.3
δ 171.0	δ 170.8

The NMR spectrum of compound 1 was similar to that of commercial cyclo(l-Phe-l-Pro). ^16,17 Long-range correlations in the heteronuclear multiple bond correlation (HMBC) spectrum (Figure 1) supported compound 1 as cyclo(l-Phe-l-Pro).

Figure 1

Key HMBC (arrows) correlations of compound 1.

In addition, the retention time of compound 1 in analytical HPLC (t _R 4.6 minutes) matched that of commercial cyclo(l-Phe-l-Pro) (t _R 4.6 minutes); thus, we finally identified compound 1 as (3S,8aS)-3-benzylhexahydropyrrolo[1,2-a]pyrazine-1,4-dione [cyclo(l-Phe-l-Pro)] (Figure 2).

Figure 2

Chemical structure of compounds 1, 2, and 3.

By HR-EI-MS, the elemental composition of compound 2 was determined to be C₁₄H₁₆N₂O₃, which has one more oxygen atom than compound 1, and that of compound 3 was determined to be C₁₁H₁₈N₂O₂. Their NMR data (Table 1) and [α]_D were in good agreement with the data of Wattana-Amorn et al, who analyzed cyclo(l-Tyr-l-Pro), ¹⁸ and the data of commercial cyclo(l-Leu-l-Pro). ¹⁹ The retention times of compound 2 (t _R 1.2 minutes) and compound 3 (t _R 2.9 minutes) matched those of cyclo(l-Tyr-l-Pro) and cyclo(l-Leu-l-Pro), respectively. Thus, we finally identified compound 2 as (3S,8aS)-3-(4-hydroxybenzyl)hexahydropyrrolo[1,2-a]pyrazine-1,4-dione [cyclo(l-Tyr-l-Pro)] and compound 3 as (3S,8aS)-3-isobutylhexahydropyrrolo[1,2-a]pyrazine-1,4-dione [cyclo(l-Leu-l-Pro)] (Figure 2).

All 3 compounds contain a diketopiperazine ring derived from proline and have an alkyl group at C-3; thus, we collectively named them asparagus-derived proline-containing 3-alkyldiketopiperazines or Asparaprolines.

We evaluated the biological function of Asparaprolines in terms of their ability to induce HSP70 mRNA expression in HL-60 cells. Ito et al evaluated HSP70 mRNA induction activity of HMF at micromolar order, ¹² and the molar ratio of compounds 1, 2, and 3 in ETAS powder (Lot #HSP1201S) was 1:1.5:3.1 as measured by LC-MS system, so we set the final concentration of Asparaprolines (ie, sum of compounds 1, 2, and 3) as 1 µM. The cells were treated with Asparaprolines (final concentration 1 µM), HMF (final concentration 1 µM), or control in prewarmed growth medium for 4 hours, and the levels of HSP70 mRNA were evaluated by real-time PCR. Compared with the control, Asparaprolines significantly increased HSP70 mRNA expression (stimulation index 5.75 ± 1.45; P = 0.022), whereas the effect of HMF was not significant (stimulation index 1.88 ± 0.67; P = 0.334; Figure 3).

Figure 3

HSP70 mRNA induction activity of Asparaprolines and HMF. HSP70 mRNA induction activity was evaluated by real-time PCR, as described in the experimental section.

To explore the structure-activity relationship, diketopiperazines, in which the proline of Asparaprolines was substituted by glycine, were tested in the assay (n = 5 or 6). Compared with the control (treated with growth medium), cyclo(l-Tyr-l-Pro) and cyclo(l-Tyr-l-Gly) significantly increased HSP70 mRNA expression [cyclo(l-Tyr-l-Pro): stimulation index 1.66 ± 0.14; P = 0.006; cyclo(l-Tyr-l-Gly): stimulation index 1.48 ± 0.14; P = 0.034]; cyclo(l-Phe-l-Pro) and cyclo(l-Leu-l-Pro) tended to increase HSP70 mRNA expression [cyclo(l-Phe-l-Pro): stimulation index 1.41 ± 0.19; P = 0.084; cyclo(l-Leu-l-Pro): stimulation index 1.49 ± 0.22; P = 0.094]; but cyclo(l-Phe-l-Gly) and cyclo(l-Leu-l-Gly) did not induce HSP70 expression [cyclo(l-Phe-l-Gly): stimulation index 0.83 ± 0.17; P = 0.378; cyclo(l-Leu-l-Gly): stimulation index 0.75 ± 0.18; P = 0.301]. Faden et al ²⁰ previously reported that synthetic diketopiperazines containing proline induce HSP70 expression. Consistent with this, our results demonstrate that proline-containing diketopiperazines derived from natural amino acids also exhibit HSP70 induction activity.

Cyclo(l-Phe-l-Pro), cyclo(l-Tyr-l-Pro), and cyclo(l-Leu-l-Pro) have been identified as bacterial metabolites ^{16 -19} and are found in common foods. ^{21 -26} Owing to their relatively simple structure, diketopiperazines are among the most common peptide derivatives found in nature. Some diketopiperazines are thought to result from nonenzymatic cyclization of dipeptides because they are often formed during chemical and thermal processes, as well as during storage of peptides and proteins. ²⁷ In this study we identified cyclo(l-Phe-l-Pro), cyclo(l-Tyr-l-Pro), and cyclo(l-Leu-l-Pro) in ETAS, which is obtained by thermal processing of asparagus stem. On the other hand, we hardly detected Asparaprolines in the ethyl acetate extract of boiled asparagus (data not shown). Enzyme treatment of asparagus releases intracellular proteins and peptides efficiently into the reaction solution, and the manufacturing process of ETAS involves many thermal processes (ie, extraction in hot water, enzyme deactivation by heating, and spray drying). The formation of diketopiperazine in ETAS is presumed to be due to a combination of these factors.

In the current study, we demonstrated that Asparaprolines induce HSP70 expression. Buhr et al ⁸ showed that the heat shock response plays an essential role in resetting of the circadian clock due to thermal stimuli and in temperature compensation of the circadian period. We have previously shown that the dietary supplement ONR-8 containing Asparaprolines modifies social jetlag; thus, we presume that the relationship between improved social jetlag and resetting or compensation of the circadian rhythm occurs via the heat shock response.

In summary, we have identified 3 constituents, collectively termed asparagus-derived proline-containing 3-alkyldiketopiperazines or Asparaprolines, in ETAS powder. Their structures were elucidated by spectroscopic analyses, including optical rotation, HR-EI-MS, and NMR. Among the ingredients identified in ETAS powder to date, Asparaprolines were shown to be an important inducer of HSP70 in an mRNA induction assay using HL-60 cells. To our knowledge, our study is the first to indicate that the presence of proline in diketopiperazines derived from natural amino acids can affect HSP70 induction activity.

Experimental

Production of ETAS

Asparagus (A. officinalis L.) stems were collected in Hokkaido, Japan, in 2012, and the fresh stalks (900 kg) were extracted with hot water (1800 L) at 121°C for 45 minutes. The extract was cooled to 60°C and treated with cellulase and hemicellulase (0.5% w/v each) for 24 hours to avoid clogging during production. After inactivation of the enzymes (121°C, 20 minutes), the extract was separated by centrifugation, concentrated in vacuo, and mixed with dextrin (45.0 kg, Pinedex; Matsutani Chemical Industry Co, Hyogo, Japan) as a filler. The mixture was sterilized at 121°C for 45 minutes and then spray-dried to produce ETAS powder (94.0 kg; Lot #HSP1201S), consisting of 52.1% ETAS and 47.9% dextrin. Component analysis showed that the ETAS powder comprised 82.2% carbohydrates, 10.3% proteins, 3.8% ashes, 0.9% lipids, and 2.8% moisture. The powder was manufactured by Amino Up Co, Ltd, and the manufacturing process was conducted in accordance with good manufacturing practice standards for dietary supplements and ISO9001:2008 and ISO22000:2005 criteria.

Isolation of Compounds 1, 2, and 3

ETAS powder (100 g; Lot #HSP1201S) was dissolved in 500 mL of water and extracted 4 times with 500 mL of ethyl acetate. The ethyl acetate extraction procedure was repeated for 10 batches of ETAS powder, and the combined extract was evaporated to dryness under reduced pressure (18.2 g). A portion of the extract (345 mg) was subjected to preparative HPLC using an Agilent 1260 system (Agilent Technologies Inc., Santa Clara, CA, USA; Wakosil-II 5C18 [5 µm, 10 × 250 mm, Wako Pure Chemical Industries, Ltd, Osaka, Japan]; mobile phase: solvent A, 0.1% formic acid; solvent B, 0.1% formic acid in acetonitrile; 0-20 minutes, 10%-90% B; flow rate 3.5 mL/min) equipped with a Corona charged aerosol detector (CAD; Thermo Fisher Scientific Inc., Waltham, MA, USA) to give a fraction (158.5 mg, t _R 8-18 minutes) containing compounds 1, 2, and 3. This fraction was confirmed to have HSP70 induction activity in a preliminary in vitro experiment. The fraction (4.2 g) accumulated from 30 HPLC batches was then separated by the same HPLC system (Synergi 4 µm Polar-RP 80A [4 µm, 10 × 150 mm, Phenomenex Inc., Torrance, CA, USA]; mobile phase: solvent A, 0.1% formic acid; solvent B, 0.1% formic acid in acetonitrile; 0-30 minutes, 10%-50% B; 30-31 minutes, 50%-90% B; and 31-40 minutes, 90% B; flow rate 3.5 mL/min) to obtain 2 further fractions: fraction I (876 mg, t _R 4.0-11.8 minutes) and fraction II (1158 mg, t _R 11.8-18.0 minutes). Compound 2 (5.2 mg) was isolated from fraction I (746.5 mg) by using HPLC (Synergi 4 µm Polar-RP 80A [4 µm, 10 × 150 mm]; mobile phase: solvent A, 0.1% formic acid; solvent B, acetonitrile; 0-25 minutes, 5%-20% B; 25-30 minutes, 20%-90% B; and 30-40 minutes, 90% B; flow rate 3.5 mL/min). Fraction II (914.8 mg) was purified by HPLC (Synergi 4 µm Polar-RP 80A [4 µm, 10 × 150 mm]; mobile phase: solvent A, 0.1% formic acid; solvent B, acetonitrile; 0-5 minutes, 10% B; 5-30 minutes, 10%-20% B; and 30-40 minutes, 20% B; flow rate 3.5 mL/min) to obtain compounds 1 (5.9 mg) and 3 (4.3 mg). Fractionation was based on the shape of HPLC chromatogram: fractions and peaks were selected with emphasis on those that were common to multiple lots of ETAS powder.

Analytical HPLC and LC-MS

The purity of compounds 1, 2, and 3 was assessed by analytical HPLC using the same system used for preparative HPLC (Synergi 4 µm Polar-RP 80A [4 µm, 4.6 × 150 mm]; mobile phase: solvent A, 0.1% formic acid; solvent B, 0.1% formic acid in acetonitrile; flow rate 1.0 mL/min) with a Corona CAD and the following solvent gradient: 0 to 40 minutes, 2% B (1, t _R 36.3 minutes); 0 to 25 minutes, 15% to 35% B (2, t _R 8.4 minutes); and 0 to 5 minutes, 10% B; 5 to 5.1 minutes, 10% to 15% B; 5.1 to 25 minutes, 15% B (3, t _R 17.7 minutes). Final identification of the compounds (1, 2, and 3) was conducted by analytical LC-MS using an Agilent 6470 Triple Quad LC/MS system (Kinetex 5 µm EVO C18 100A [5 µm, 2.1 × 100 mm]; column temperature 40°C; injection volume 2 µL; mobile phase: solvent A, 0.05% formic acid; solvent B, acetonitrile; flow rate 0.6 mL/min; solvent gradient: 0-4.5 minutes, 9% B [compound 2, t _R 1.2 minutes; compound 3, t _R 2.9 minutes]; 4.5-6 minutes, 9%-95% B [compound 1, t _R 4.6 minutes]; 6-8.5 minutes, 95% B; Figure 1) . The general source settings in positive ionization mode were as follows: ion mode, positive ion mode dual AJS ESI; drying gas temperature 300°C; drying gas flow 10 L/min; nebulizer gas 50 psi; sheath gas temperature 300°C; sheath gas flow 11 L/min; capillary voltage 3.5 kV; fragmentor voltage 90 V; collision energy 25 eV (compound 3), 21 eV (compounds 1 and 2); precursor ion m/z 211.1 (compound 3), m/z 245.1 (compound 1), and m/z 261.1 (compound 2); qualifier ion m/z 72.1 (compound 3), m/z 154.0 (compound 1), and m/z 107.0 (compound 2).

Structure Determination

Optical rotation was measured with a P-2300 polarimeter (JASCO Corporation, Tokyo, Japan). NMR spectra were recorded in CD₃OD or CDCl₃ with a JNM-400AL spectrometer (JEOL Ltd, Tokyo, Japan) at 400 MHz for ¹H NMR and 100 MHz for ¹³C NMR. HR-EI-MS data were obtained by using an Agilent 6550 iFunnel Q-TOF LC/MS system (Agilent Technologies Inc.).

Cell Culture

A human promyelocytic leukemia cell line (HL-60) was obtained from ECACC (via DS Pharma Biomedical Co., Osaka, Japan), and the cells were incubated in growth medium comprising RPMI-1640 (Thermo Fisher Scientific Inc.), 10% heat-inactivated fetal bovine serum, 100 unit/mL of penicillin, and 100 µg/mL of streptomycin at 37°C in 5% CO₂.

Real-Time PCR Assay for HSP70 mRNA

The real-time PCR assay was performed as described by Ito et al ¹² with minor modifications. In brief, HL-60 cells were cultured in growth medium and the cell density was adjusted to 5 × 10⁵ cells/mL by adding prewarmed growth medium. Next, 0.9 mL of cell suspension was transferred to a 1.5 mL microtube, and 0.1 mL of test sample dissolved in growth medium containing 1% (v/v) DMSO (Wako Pure Chemical Industries) was added. The final test concentration of Asparaprolines [cyclo(l-Tyr-l-Pro), Bachem, Bubendorf, Switzerland; cyclo(l-Phe-l-Pro), Bachem; and cyclo(l-Leu-l-Pro), Bachem] and HMF (Sigma-Aldrich Co., St Louis, MO, USA) was 1 mM. After incubation at 37°C for 4 hours in 5% CO₂, the cells were centrifuged and the supernatant was removed by aspiration; total RNA was extracted from the cell pellets by using TRIzol reagent (Thermo Fisher Scientific Inc.) and Direct-zol RNA MiniPrep kit (Zymo Research Corp., Irvine, CA, USA).

RNA concentration was measured spectrophotometrically, and 300 ng of total RNA was used to synthesize cDNA by ReverTra Ace qPCR RT Master Mix with gDNA Remover (Toyobo Co., Osaka, Japan). The resulting cDNA was diluted and subjected to real-time PCR to measure the levels of HSP70 and β-2 microglobulin (B2M) mRNAs on a StepOnePlus Real-Time PCR system (Thermo Fisher Scientific Inc.) using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad Laboratories, Inc., Hercules, CA, USA) and the following primers: HSP70 forward, 5′-GCATTTCCTAGTATTTCTGTTTGT-3′; HPS70 reverse, 5′-AATAGTCGTAAGATGGCAGTATA-3′; B2M forward, 5′-TAGCTGTGCTCGCGCTACT-3′; B2M reverse, 5′-AGTGGGGGTGAATTCAGTGT-3′. The reaction conditions were an initial denaturation step at 95°C for 3 minutes, followed by 40 cycles of denaturation at 95°C for 10 seconds and annealing at 57.8°C for 30 seconds. Expression of the HSP70 and B2M mRNAs was evaluated by the relative standard curve method using standard RNA extracted from heat-treated HL-60 cells (42°C for 30 minutes); HSP70 mRNA data were normalized to B2M mRNA expression. Data were expressed as a stimulation index (ie, the expression levels of HSP70 mRNA were normalized by those of B2M mRNA; then, the ratio of the relative expression level of HSP70 mRNA was calculated, when the level in the control [treated with prewarmed growth medium] was set as 1). Assessment of the structure-activity relationship was performed as mentioned above using cyclo(l-Tyr-l-Gly) (Toronto Research Chemicals, Toronto, Canada), cyclo(l-Phe-l-Gly) (Bachem), and cyclo(l-Leu-l-Gly) (Bachem).

Statistical Analysis

The results of HSP70 induction activity were reported as mean ± S.E. (n = 4-6). SAS software (R9.3, SAS Institute Japan Ltd, Tokyo, Japan) was used for statistical analysis. Unpaired t-test was used to compare values. The level of significance was set at a P value of 0.05.