Evaluation of bactericidal effects of chlorogenic or hydroxycinnamic acid derivatives and soluble coffee under low pH or gastric acid conditions

Abstract

Soluble coffee and chlorogenic acid (CQA) showed bactericidal effects on Salmonella spp. and Escherichia coli under low pH conditions simulating gastric juice. However, soluble coffee contains various hydroxycinnamic acid derivatives other than CQA. The aim of this study was to evaluate the bactericidal activities of hydroxycinnamic acid derivatives. We prepared 12 kinds of hydroxycinnamic acid derivatives from green coffee beans and applied them to the bactericidal test using Salmonella spp. at pH 3.0. These compounds, including dicaffeoylquinic and caffeoyl-feruloylquinic acids, showed similar specific bactericidal activities, which were several times greater than that of soluble coffee. The sum of the activities of these compounds explained about 9% and 140% of the total activities of a soluble coffee and green coffee beans, respectively. Although the bactericidal activities and the contents of these compounds in retail soluble coffee samples were measured, no relationship between the activity and the content was apparent.

Introduction

Although we intake various kinds of bacteria, including pathogenic and food-borne bacteria through the mouth, these bacteria are exposed to strongly acidic conditions in the stomach and are typically killed at this acidity. Intragastric pH in healthy persons is about 1.4–2.1 in a fasted state (Dressman et al., 1990; Russell et al., 1993). However, some bacteria such as Salmonella spp. and Escherichia coli can survive at low pH such as pH 2–4 because they have acid tolerance mechanisms to excrete or reduce protons from cells (Beales, 2004; Foster, 2004; Lianou et al., 2017; Lund et al., 2014).

In our previous paper (Ueno et al., 2020), we assumed that some food components promoted the bactericidal effect of the strong acidity and showed that a soluble coffee sample and 5-caffeoylquinic acid (CQA) had bactericidal activity against Salmonella spp. and E. coli under low pH or gastric acid conditions. Green coffee beans and soluble or roasted coffee contain various kinds of hydroxycinnamic acid or CQA derivatives. Among them, 5-CQA, which is sometimes known as 3-CQA confusingly, is the major one (Clifford, 1985; Okada et al., 1997; Fujioka and Shibamoto, 2008). Although the bacteria showed similar growth in heart infusion (HI) broth with and without 0.5% 5-CQA at pH 7, 5-CQA showed bactericidal activities against Salmonella spp. at pH 3.0 and E. coli at pH 1.5 (Ueno et al., 2020). Further, we found that 3- and 4-CQAs also had similar bactericidal activity. The specific bactericidal activities of the three CQAs were several times higher than that of soluble coffee. Considering the content of the three CQAs in soluble coffee, these compounds had a partial contribution to the total bactericidal activity of soluble coffee (Ueno et al., 2020).

Soluble or roasted coffee contains various kinds of CQAs including dicaffeoylquinic acids (diCQAs), caffeoylferuloylquinic acids (CFQAs), and hydroxycinnamic acid derivatives other than 3-, 4-, and 5-CQAs (Okada et al., 1997). The aim of this study was to clarify the bactericidal activity of each CQA or hydroxycinnamic acid derivative in coffee (Fig. 1) and the contribution of these compounds to the bactericidal activity of coffee at low pH conditions.

Fig. 1.

Structure of 12 hydroxycinnamic acid derivatives in coffee (Compounds 1–12; A) and a typical chromatogram of green coffee bean on HPLC (B).

Materials and Methods

Coffee beans    Green coffee beans (Coffea canephora var. robusta) were supplied by All Japan Coffee Association (Tokyo, Japan) for preparation of CQAs or hydroxycinnamic acid derivatives.

Bacterial strains and conditions    Salmonella enterica subspecies enterica serovar Enteritidis NBRC 3313, S. Typhimurium DT104, E. coli NBRC 14237, and E. coli O157:H7 (strains C-13 and C-52; Fukuyama et al., 2009) held in our laboratory were used in this study.

Soluble coffee and reagents    Soluble coffee samples (Nescafe Gold Blend, Nescafe Gold Blend “Kokufukame”, Nescafe “Exella”, Nescafe “President”, Nestle Japan, Tokyo; MAXIM, Ajinomoto AGF, Tokyo, Japan; The BLEND 114, UCC Ueshima Coffee, Kobe, Japan) were purchased at a local market in Tokyo and used as coffee powder.

3-Caffeoyl quinic acid hemihydrate (Wako Fujifilm, Tokyo, Japan) was used as a standard CQA and abbreviated as 5-CQA. This compound has an ester bond between caffeic acid and a hydroxy group of 5-position of quinic acid in the IUPAC system (IUPAC, 1976). This numbering or abbreviation was recommended by Clifford (1985). Ferulic acid (Wako FUJIFILM, Tokyo, Japan), d (-)-quinic acid (Nacalai Tesque, Kyoto, Japan), HI broth (Eiken Chemical, Tokyo, Japan), and HI agar (Difco Laboratories, Detroit, USA) were commercially available.

Preparation of CQAs or hydroxycinnamic acid derivatives from green coffee beans    CQAs or hydroxycinnamic acid derivatives were prepared according to the method of Murata et al. (1995) with some modifications. About 100 g of green coffee beans was frozen with liquid nitrogen before being pulverized with a coffee mill (SKRM070-SF, Tiger, Osaka, Japan). About 2 L of 70% aqueous methanol was added to the powder and refluxed for 15 min. An extract was then obtained by centrifugation. This procedure was further repeated twice. The combined extracts were concentrated in vacuo, and the pH of the concentrate was adjusted to less than 2 prior to extraction of hydroxycinnamic acid derivatives with ethyl acetate. The ethyl acetate layer was dried with Na₂SO₄ and concentrated in vacuo. This crude paste (about 9.9 g) was applied to a reversed-phase open column (Chromatorex ODS DM1020T, i.d. 28 × 380 mm; Fuji Silysia Chemical, Kasugai, Japan), which was successively developed with a mixture of CH₃CN and 5% aqueous acetic acid (2 : 98, 5 : 95, 10 : 90, 15 : 85, 20 : 80, 25 : 75, and 30 : 70, v/v), each fraction being monitored by analytical HPLC as described below. Fractions containing the respective hydroxycinnamic acid derivatives were collected and concentrated in vacuo. The obtained fractions were then applied to preparative HPLC (column, YMC pack R&D ODS-A, i.d. 20 × 250 mm; YMC, Kyoto, Japan); eluent, a mixture of CH₃CN and 5% aqueous acetic acid (2 : 98, 5 : 95, 10 : 90, 15 : 85, or 20 : 80, v/v); flow rate, 9.99 mL; detection, 320 nm). Each fraction was concentrated in vacuo. Compounds 1–13 (Fig. 1) were obtained. Each compound was identified by comparison of the NMR data of authentic samples (5-CQA, quinic acid, caffeic acid, and ferulic acid) and the literature (Clifford and Jarvis, 1988; Clifford et al., 1989; Murata et al., 1995; Okada et al., 1997).

Compound 1 (3-CQA)    ESI-MS m/z calcd. for C₁₆H₁₇O₉, 353.0867, found 353.0880. ¹H-NMR δ_H (ppm); 2.15 (4H, m), 3.65 (1H, m), 4.16 (1H, m), 5.36 (1H, d, J = 15.9), 6.30 (1H, d, J = 15.9), 6.77 (1H, d, J = 7.9), 6.94 (1H, d, J = 7.9), 7.04 (1H, s), 7.58 (1H, d, J = 15.9).

Compound 2 (5-CQA)    ESI-MS m/z calcd. for C₁₆H₁₇O₉, 353.0867, found 353.0886. ¹H-NMR δ_H (ppm) 2.05 (1H, dd, J = 5.5, 14.0), 2.08 (1H, dd, J = 9.5, 13.2), 2.18 (1H, dd, J = 3.0, 14.0), 2.23 (1H, dd (J = 4.5, 13.2), 3.73 (1H, dd, J = 3.0, 8.7), 4.17 (1H, td, J = 3.0, 5.5), 5.33 (1H, ddd, J = 4.5, 8.7, 9.5), 6.26 (1H, d, J = 15.9), 6.77 (1H, d, J = 7.9), 6.94 (1H, d, J = 7.9), 7.04 (1H, s), 7.58 (1H, d, J = 15.9).

Compound 3 (4-CQA)    ESI-MS m/z calcd. for C₁₆H₁₇O₉, 353.0867, found 353.0877. ¹H-NMR δ_H (ppm); 2.15 (4H, m), 4.29 (1H, m), 4.30 (1H, m), 4.81 (1H, dd, J = 2.3, 9.4), 6.37 (1H, d, J = 15.9), 6.79 (1H, d, J = 7.8), 6.96 (1H, d, J = 7.8), 7.07 (1H, s), 7.64 (1H, d, J = 15.9).

Compound 4 (5-feruloylquinic acid; 5-FQA)    ESI-MS m/z calcd. for C₁₇H₁₉O₉, 367.1023, found 367.1049. ¹H-NMR δ_H (ppm); 2.05 (1H, d-like, J = 13.7), 2.09 (1H, dd-like, J = 10.0, 13.5), 2.18 (1H, d, J = 13.7), 2.24 (1H, d, J = 13.5), 3.73 (1H, d-like, J = 8.5), 3.88 (O-CH₃, 3H, s), 4.17 (1H, m), 5.34 (1H, ddd, J = 4.4, 8.5, 10.0), 6.35 (1H, d, J = 15.9)), 6.81 (1H, d, J = 8.1), 7.08 (1H, dd, J = 1.9, 8.1), 7.19 (1H, d, J = 1.9), 7.61 (1H, d, J = 15.9).

Compound 5 (3,4-diCQA)    ESI-MS m/z calcd. for C₂₅H₂₃O₁₂, 515.1182, found 515.1214. ¹H-NMR δ_H (ppm); 2.10 (1H, dd, J = 10.0, 13.2), 2.15 (1H, d-like, J = 14.6), 2.24 (1H, d-like, J = 13.2), 2.37 (1H, dd, J = 3.2, 14.6), 4.39 (1H, ddd, J = 4.2, 9.2, 10.0), 5.00 (1H, dd, J = 3.2, 9.2), 5.64 (1H, td, J = 3.2, 4.7)), 6.25 (1H, d, J = 15.9), 6.29 (1H, d, J = 15.9), 6.72 (1H, d, J = 8.2)), 6.78 (1H, d, J = 8.2), 6.87 (1H, dd, J = 1.6, 8.2), 6.93 (1H, dd, J = 1.6, 8.2), 7.02 (1H, d, J = 1.6), 7.04 (1H, d, J = 1.6)), 7.55 (1H, d, J = 15.9), 7.58 (1H, d, J = 15.9).

Compound 6 (3,5-diCQA)    ESI-MS m/z calcd. for C₂₅H₂₃O₁₂, 515.1182, found 515.1197. ¹H-NMR δ_H (ppm); 2.17 (1H, dd, J = 6.8, 13.8), 2.22 (1H, dd, J = 4.5,13.8), 2.25 (1H, dd, J = 8.5, 13.8), 2.33 (1H, dd, J = 3.4, 13.8), 3.99 (1H, dd, J = 3.2, 7.2), 5.39 (1H, ddd, J = 4.5, 7.2, 8.5), 5.44 (1H, ddd, J = 3.2, 3.4, 6.8), 6.27 (1H, d, J = 15.9), 6.35 (1H, d, J = 15.9), 6.78 (1H, d, J = 8.1), 6.79 (1H, d, J = 8.1)), 6.96 (1H, dd, J = 1.5, 8.1), 6.97 (1H, dd, J = 1.5, 8.1), 7.07 (2H, s-like), 7.58 (1H, d, J = 15.9), 7.62 (1H, d, J = 15.9).

Compound 7 (4,5-diCQA)    ESI-MS m/z calcd. for C₂₅H₂₃O₁₂, 515.1182, found 515.1204. ¹H-NMR δ_H (ppm) 2.11 (1H, dd, J = 3.2, 14.2)), 2.26 (2H, m), 2.31 (1H, d-like, J = 14.2), 4.37 (1H, m), 5.12 (1H, dd (J = 2.5, 8.9), 5.62 (1H, ddd, J = 4.6, 8.9, 9.4), 6.20 (1H, d, J = 15.9), 6.29 (1H, d, J = 15.9), 6.75 (2H, d, J = 8.2), 6.90 (1H, dd, J = 1.6, 8.2), 6.92 (1H, dd, J = 1.6, 8.2), 7.00 (1H, d, J = 1.6)), 7.03 (1H, d, J = 1.6), 7.03 (1H, d, J = 15.9), 7.52 (1H, d, J = 15.9).

Compound 8 (3,4-CFQA)    ESI-MS m/z calcd. for C₂₅H₂₃O₁₂, 529.1338, found 529.1350. ¹H-NMR δ_H (ppm) 2.09 (1H, dd, J = 10.5, 13.1), 2.15 (1H, d-like, J = 14.8), 2.24 (1H, d-like, J = 13.1), 2.37 (1H, dd, J = 3.3, 14.8), 3.78 (3H, s (OCH₃)), 4.39 (1H, ddd, J = 4.3, 9.6, 10.5), 5.00 (1H, dd, J = 3.5, 9.6), 5.64 (1H, ddd, J = 3.3, 3.5, 4.5), 6.30 (1H, d, J = 15.9), 6.35 (1H, d, J = 15.9), 6.76 (1H, d, J = 8.1), 6.77 (1H, d, J = 8.1), 6.92 (1H, dd, J = 1.6, 8.1), 7.01 (1H, dd, J = 1.6, 8.1), 7.04 (1H, d, J = 1.6), 7.08 (1H, d, J = 1.6), 7.57 (1H, d, J = 15.9), 7.59 (1H, d, J = 15.9).

Compound 9 (3,5-CFQA)    ESI-MS m/z calcd. for C₂₅H₂₃O₁₂, 529.1338, found 529.1365. ¹H-NMR δ_H (ppm); 2.21 (3H, m), 2.32 (1H, dd, J = 3.0, 13.2), 3.90 (3H, s (O-CH₃)), 3.98 (1H, dd, J = 3.1, 7.3), 5.39 (1H, dd-like, J = 3.1, 7.5), 5.44 (1H, dd-like, J = 3.0, 7.3), 6.35 (1H, d, J = 15.9), 6.36 (1H, d, J = 15.9), 6.78 (1H, d, J = 8.1), 6.81 (1H, d, J = 8.1), 6.96 (1H, dd, J = 1.6, 8.1), 7.06 (1H, s), 7.10 (1H, dd, J = 1.6, 8.1), 7.20 (1H, s), 7.61 (1H, d, J = 15.9), 7.63 (1H, d, J = 15.9).

Compound 10 (4,5-CFQA)    ESI-MS m/z calcd. for C₂₅H₂₃O₁₂, 529.1338, found 529.1375. ¹H-NMR δ_H (ppm); 2.12 (1H, dd, J = 3.3, 14.5), 2.25 (1H, d, J = 10.4, 12.4), 2.31 (2H, m), 3.84 (3H, s (O-CH₃)), 4.38 (1H, ddd, J = 3.1, 3.3, 7.5), 5.13 (1H, dd, J = 3.1, 9.0), 5.65 (1H, ddd, J = 4.6, 9.0, 10.4), 6.28 (1H, d, J = 15.9), 6.29 (1H, d, J = 15.9), 6.74 (1H, d, J = 8.2), 6.77 (1H, d, J = 8.2), 6.91 (1H, dd, J = 1.6, 8.2), 7.02 (1H, s), 7.02 (1H, d, J = 1.6, 8.2), 7.10 (1H, d, J = 1.6), 7.57 (1H, d, J = 15.9), 7.60 (1H, d, J = 15.9).

Compound 11 (caffeoyl-tryptophan)    ESI-MS (m/z): 365.1155 (M-H) ⁻, calcd. for C₂₀H₁₇N₂O₅, 365.1131, found 365.1155. ¹H-NMR δ_H (ppm) 3.31 (1H, dd, J = 7.3, 14.8), 3.45 (1H, dd, J = 5.2, 14.8), 4.98 (1H, dd, J = 5.2, 7.3), 6.41 (1H, d, J = 15.6), 6.80 (1H, d, J = 8.4), 6.88 (1H, dd, J = 1.5, 8.4), 7.02 (1H, t, J = 7.4), 7.04 (1H, d, J = 1.5), 7.10 (1H, t, J = 7.4), 7.12 (1H, s), 7.34 (1H, d, J = 8.0), 7.44 (1H, d, J = 15.6), 7.62 (1H, d, J = 8.0).

Compound 12 (p-coumaroyl-tryptophan)    ESI-MS (m/z): 349.1212 (M-H) ⁻, calcd. for C₂₀H₁₇N₂O₄, 349.1182, found 349.1212. δ_H (ppm); 3.28 (1H, dd, J = 7.4, 14.8), 3.42 (1H, dd, J = 5.5, 14.8), 4.92 (1H, dd, J = 5.5, 7.4), 6.42 (1H, d, J = 15.7), 6.79 (2H, d, J = 8.6), 7.00 (1H, ddd, J = 0.8, 6.8, 8.1), 7.08 (1H, ddd, J = 0.8, 6.8, 8.1), 7.11 (1H, s), 7.33 (1H, d, J = 8.1), 7.35 (2H, d, J = 8.6), 7.44 (1H, d, J = 15.7), 7.60 (1H, d, J = 8.1).

HPLC analysis of hydroxycinnamic acid derivatives    HPLC was performed as follows: column, YMC-Pack R&D ODS-A (i.d. 4.6 × 250 mm, YMC); eluent, solution A (CH₃CN : 5% aqueous acetic acid = 2 : 98, v/v) and solution B (CH₃CN : 5% aqueous acetic acid = 30 : 70, v/v), 0% B for 0–5 min, 0 to 100% B for 5–25 min, and 100% B for 25–30 min; flow rate, 1.0 mL/min; detection, Chromaster 5430 Diode Array Detector (HITACHI, Tokyo, Japan); wavelength, 250–400 nm; fixed wavelength, 320 nm.

Instrumental analyses    Spectroscopic measurements were made by using the following instruments: NMR (AVANCE 800, Bruker Biospin, Karlsruhe, Germany) and MS (Triple TOF 4600, AB Sciex, Framingham, USA). CD₃OD containing 0.03% TMS (Kanto Chemical, Tokyo, Japan) was used as a solvent and an internal standard for NMR. LC-MS was conducted as follows: column, Intersil ODS-3 (i.d. 4.6 × 150 mm, GL Science, Tokyo, Japan); column temperature, 40 °C; eluent, solution A (2% aqueous acetic acid) and solution B (methanol), 0% B for 0–5 min, 0–50% B for 5–50, 50% B for 50–55 min; flow rate, 0.2 mL/min; ionization, ESI (-); MS range, 100–1000.

Media for bactericidal test    HI broth was pH-adjusted with HCl, and then autoclaved at 121 °C for 20 min. Each sample (0–1.0% soluble coffee powder, 0–0.15% hydroxycinnamic acid derivatives) was added to the broth. The pH of each broth was adjusted to pH 1.5 ± 0.1 and 3.0 ± 0.1 after adding samples to the autoclaved media. Each broth medium (4.5 mL) was then dispensed aseptically into a screw-capped tube (i.d. 16.5 × 150 mm).

Inoculum preparation    Each inoculum was prepared according to a previous paper (Ueno et al., 2020). The two strains of Salmonella spp. and three strains of E. coli were mixed at a ratio of 1 : 1 and 1 : 1 : 1, respectively. Each mixed culture was appropriately diluted with HI broth and then inoculated into the HI broth.

Bactericidal test against Salmonella spp. and E. coli in HI broth under low pH conditions    A test solution (5 mL) containing HI broth, a sample and 10⁶ CFU/mL of Salmonella spp. or E. coli was respectively incubated in a test tube closed with a screw cap with shaking at 110 rpm, at 3 700 and pH 3.0 or 1.5 for 1.5 h. After incubation, the bacteria were washed with PBS containing 0.1% peptone, and the number of surviving bacteria was counted using the spread and pour plate method with HI agar. The detection limits of bacterial numbers were about 10⁰–10¹ CFU/mL.

One unit (U) of the bactericidal activity of each compound or coffee sample was defined as the reduction in viable bacterial number of 1/10 CFU/mL or 1-log reduction in a test solution from a control (no addition of samples) after 1.5 h of incubation. In a control, the bacterial number was reduced by 0.5- to 1.2-log after 1.5 h of incubation. The specific bactericidal activity (SBA; U/mg of compound or coffee sample) was calculated from the concentrations of an added compound or coffee sample. The bactericidal activity of a compound (TBA; U/g coffee sample) was calculated as the SBA (U/mg) × each content of a coffee sample (mg/g coffee sample).

Preparation of hot water extract from green coffee beans    Boiling reverse osmosis water (50 mL) was added to pulverized green coffee beans (2.5 g), which was stirred for 5 min at room temperature. After centrifuging at 2 800 × g for 10 min, the supernatant was freeze-dried (FREEZVAC 1C, Touzai Tuusho, Tokyo, Japan). About 0.71 g of yellowish powder was obtained.

Analysis of hydroxycinnamic acids in soluble coffee and green coffee beans    Aqueous methanol (methanol : water = 70 : 30, v/v; 20 mL) was added to a soluble coffee powder (0.5 g) or pulverized green coffee beans (0.5 g), then heated for 15 min at 70 °C. After centrifugation at 3 000 × g for 10 min at 4 °C, the supernatant was obtained. This procedure was further repeated three times. The four supernatants were combined, evaporated under reduced pressure, and filled up to 20 mL. The pH of part of the solution (10 mL) was adjusted to less than 2 by adding 6 M HCl, before extraction of cinnamic acid derivatives with ethyl acetate three times. The obtained ethyl acetate layers were evaporated under reduced pressure and dissolved in 10 mL with 70% methanol, which was passed through a Chromatodisk (0.45 µm; Juji Field, Tokyo, Japan) and then subjected to HPLC. The concentrations of hydroxycinnamic acid derivatives in soluble coffee were calculated from the peak areas of the sample as 5-CQA equivalence.

Preparation of ethyl acetate fractions from soluble coffee and green coffee beans    Ethyl acetate fractions were prepared from soluble coffee (Nescafe Gold Blend) and green coffee beans as described previously (Ueno et al., 2020). Green coffee beans were frozen in a deep freezer and pulverized with a mortar and pestle. Briefly, each powder (2.5 g) was successively extracted with ether, methanol, 80% methanol, and hot water. After removal of methanol in the 80% methanol extract by an evaporator, hydroxycinnamic acids were extracted with ethyl acetate. Each ethyl acetate fraction was obtained after removal of the solvent by evaporation under vacuum. SBA, TBA, and content of hydroxycinnamic acid derivatives were estimated as described above.

Statistical analyses    Statistical analyses were performed using Excel Mac 2011 (Microsoft, Redmond, USA) with the add-in software Statcel 3 (OMS, Tokorozawa, Japan). Data were assessed using Pearson's correlation coefficient and oneway analysis of variance followed by Tukey's multiple comparison test. The significance level was set at p < 0.05. All experiments were conducted in at least triplicate.

Results

Bactericidal effects of soluble coffee against Salmonella spp. and E. coli in HI broth under low pH conditions    In our previous study (Ueno et al., 2020) we showed that a brand of soluble coffee had bactericidal activity against Salmonella spp. at pH 3.0 and E. coli at pH 1.5. The pH was set at 3.0 for Salmonella spp., because viable cells of Salmonella spp. could not be detected after 15-min incubation at pH 1.5 without the coffee sample. At pH 3, coffee showed no bactericidal activity against E. coli. At first, to generalize the bactericidal activity of coffee, we examined the activities using five other brands of soluble coffee at these pH conditions. As shown in Fig. 2, all coffee samples we used showed bactericidal activities against Salmonella spp. and Escherichia coli. In a control, the bacterial number was reduced by 0.5- to 1.2-log from its initial number. The SBAs against Salmonella spp. were 0.53 to 0.63 U/mg (mean ± standard deviation (SD), 0.57 ± 0.06 U/mg) and those against E. coli were 0.45 to 0.51 U/mg (0.48 ± 0.06 U/mg). These activities meant that about 2 (= 1/0.57 or 1/0.48) mg/mL coffee powder reduced the viable cell numbers of Salmonella spp. and E. coli to 1/10 compared with each control. When the bacterial number in a control is reduced by 1-log from the initial value, this means a 1/100 reduction from the initial bacterial number. The SBA levels of all samples were similar and no significant differences among the coffee samples were apparent.

Fig. 2.

SBAs of soluble coffee samples against Salmonella spp. (A) and E. coli (B) under low pH conditions. The activities of retail 6 brands of soluble coffee were estimated using two strains of Salmonella spp. and three strains of E. coli in HI broth at pH 3.0 and pH 1.5, respectively. No significant differences were apparent among samples (n = 3).

Preparation and identification of hydroxycinnamic acid derivatives.    Twelve kinds of hydroxycinnamic acid derivatives (Fig. 1) were prepared from green coffee beans as described in the material and methods section.

Ten of them (Compounds 1–10) are CQA derivatives composed of quinic acid and caffeic or ferulic acids (Fig. 1). The position of the ester bond between a carboxy group of caffeic or ferulic acids and a hydroxy group of quinic acid was determined by the NMR data as shown in Table 1. Table 1 shows the differences in the chemical shifts of the protons between quinic acid (H-C-OH) and the part of quinic acid in CQAs (H-C-OH and H-C-OC=O). The chemical shifts of protons bound to the carbon carrying an ester bond or (H-COC=O) shifted more than 1 ppm downfield from those of free quinic acid (H-C-OH). Compounds 1–3 were 3-, 5-, and 4-CQAs. Compound 4 was 5-FQA. Compounds 5–7 were diCQAs. Compounds 8–10 were CFCQs. In CFQAs, although the position of a free hydroxy group in quinic acid was specified, the other two positions of the O-bond of feruloyl and caffeoyl esters could not be specified (Clifford et al., 1989).

Table 1. ¹H-NMR data of compounds 1–10 and their identification

Compound	Differences in chemical shifts from quinic acid (ppm)			Hydroxycinnamic acid	Identification
	C3-H	C4-H	C5-H
Quinic acid (standard)	0	0	0
5-CQA (standard)	0.07	0.34	1.33	Caffeic acid
1	1.27	0.26	0.16	Caffeic acid	3-CQA
2	0.08	0.34	1.33	Caffeic acid	5-CQA
3	0.20	1.42	0.30	Caffeic acid	4-CQA
4	0.08	0.34	1.34	Ferulic acid	5-FQA
5	1.55	1.61	0.39	Caffeic acid	3,4-diCQA
6	1.35	0.60	1.39	Caffeic acid	3,5-diCQA
7	0.28	1.73	1.62	Caffeic acid	4,5-diCQA
8	1.55	1.61	0.39	Caffeic acid/ Ferulic acid	3,4-CFQA
9	1.30	0.59	1.44	Caffeic acid/ Ferulic acid	3,5-CFQA
10	0.29	1.74	1.65	Caffeic acid/ Ferulic acid	4,5-CFQA

Values written in bold show larger differences from free quinic acid than the others. NMR was measured in CD₃OD with TMS. CQA, caffeoylquinic acid; FQA, feruloylquinic acid; diCQA, dicaffeoylquinic acid; CFQA, caffeoylferuloylquinic acid.

Compounds 11 and 12 (Fig. 1) were identified as caffeoyl-tryptophan (Morishita et al., 1987) and p-coumaroyl-tryptophan (Murata et al., 1995), respectively.

Bactericidal effects of hydroxycinnamic acid derivatives and soluble coffee against Salmonella spp. under low pH conditions    The bactericidal activities of 12 hydroxycinnamic acids shown in Fig. 1 were estimated using Salmonella spp. at pH 3.0. Table 2 shows the SBA of each compound. The values ranged from 1.8 to 3.3 U/mg, which are similar levels and significantly greater than that of the soluble coffee sample (0.54 U/mg). Next, from each content in the soluble coffee and specific activity, we estimated each contribution to the total activity of the coffee and the sum of them. As shown in Table 2, although 5-CQA was the major CQA or hydroxycinnamic acid derivative in soluble coffee, it explained only 2.3% (2.9 U/mg × 4.3 mg/g coffee/540 U/g × 100 = 2.31) of the activity of soluble coffee. The sum of each TBA of the 12 hydroxycinnamic acid derivatives was 48.3 U/g coffee, representing 8.9% of the activity of coffee (48.3/540 × 100 = 8.94).

Table 2. Bactericidal activities of hydroxycinnamic acid derivatives and a soluble coffee against Salmonella spp. in an acidic medium

Sample	Content in soluble coffee	SBA*	TBA**
	(mg/g coffee sample)	(U/mg)	(U/g coffee sample) (%)***
3-CQA	1.7	2.7 ± 0.8a	4.6
5-CQA	4.3	2.9 ± 0.2a	12.6
4-CQA	4.5	2.5 ± 0.9a	11.2
5-FQA	3.2	2.6 ± 1.0a	8.3
3,4-diCQA	0.5	2.1 ± 0.8a	1.0
3,5-diCQA	1.2	2.5 ± 04a	3.0
4,5-diCQA	0.9	2.6 ± 0.4a	2.3
3,4-CFQA	0.5	3.0 ± 0.7a	1.4
3,5-CFQA	0.9	2.0 ± 1.0a	1.8
4,5-CFQA	0.5	1.8 ± 0.5a	0.9
Caffeoyl-tryptophan	0.6	2.1 ± 0.9a	1.3
p-Coumaroyl-tryptophan	<0.1	3.4	<0.3
Sum of hydroxycinnamic acids	18.6		48.3 (8.9) ***
Coffee		0.5 4 ± 0.22b	540 (100) ***

Bactericidal activity was estimated using two strains of Salmonella spp. in HI broth (pH 3.0).

*  mean ± SD.

Different letters show a significant difference (n = 3 except for p-coumaroyl tryptophan, n = 2).

**  content × SBA

***  a soluble coffee was set to 100%.

CQA, caffeoylquinic acid; FQA, feruloylquinic acid; diCQA, dicaffeoylquinic acid; CFQA, caffeoyl-feruloylquinic acid.

Contribution of hydroxycinnamic acid derivatives to the bactericidal activity of green coffee bean extract    Next, we estimated the bactericidal activity of a hot water extract of green coffee beans. The SBA was 0.25 (U/mg), which was less than half of that of soluble coffee (0.54 U/mg). We subsequently estimated the contribution of hydroxycinnamic acid derivatives in the green coffee bean extract to its TBA using the same procedure as used for soluble coffee. As shown in Table 3, the TBA of the hot water extract of green coffee beans was 248 U/g, while the sum of each TBA of hydroxycinnamic acid derivatives was 347 U/g, which was about 140% of that of green coffee beans. Therefore, the TBA of green coffee beans was explained by the activities of hydroxycinnamic acid derivatives.

Table 3. Contents of hydroxycinnamic acid derivatives in a green coffee bean extract and their bactericidal activities against Salmonella spp. in an acidic medium

Sample	Content in green coffee bean extract	TBA*
	(mg/g green coffee bean extract)	(U/g green coffee bean extract) (%)**
3-CQA	13.4	37
5-CQA	42.1	124
4-CQA	19.5	48
5-FQA	23.0	61
3,4-diCQA	8.7	18
3,5-diCQA	7.0	18
4,5-diCQA	9.4	24
3,4-CFQA	1.8	5
3,5-CFQA	1.1	5
4,5-CFQA	1.7	3
Caffeoyl-tryptophan	1.7	4
p-Coumaroyl-tryptophan	1.0	3.7
Sum of hydroxycinnamic acids	130.4	347 (140)**
Green coffee bean extract		248 (100)**

*  content × SBA shown in Table 2

**  a green coffee bean extract was set to 100%.

CQA, caffeoylquinic acid; FQA, feruloylquinic acid; diCQA, dicaffeoylquinic acid; CFQA, caffeoyl-feruloylquinic acid.

Comparison of ethyl acetate fractions prepared from green coffee beans and soluble coffee    As described in our previous paper (Ueno et al., 2020), the ethyl acetate fraction of soluble coffee was the richest in hydroxycinnamic acid derivatives and showed the highest SBA among prepared fractions from soluble coffee. Thus, each ethyl acetate fraction was prepared from soluble coffee and green coffee beans, and then each bactericidal activity and content of hydroxycinnamic acid derivatives were compared. As a result, the SBA of the ethyl acetate fraction prepared from soluble coffee (7.8 U/mg) was 1.4 times greater than that prepared from green coffee beans (5.4 U/mg). On the other hand, the content of hydroxycinnamic acid derivatives in the ethyl acetate fraction prepared from green coffee beans was more than 2.6 times higher than that prepared from soluble coffee. We estimated that the contribution of hydroxycinnamic acids to the TBA of the fraction was about 6 times higher in soluble coffee than in green coffee beans.

Relationship between contents of hydroxycinnamic acid derivatives and the bactericidal activities of soluble coffee    The relationship between the contents of hydroxycinnamic acid derivatives and the bactericidal activities of the six soluble coffee samples was examined. However, no significant positive relationships between the SBAs of soluble coffee and the contents of hydroxycinnamic acid derivatives were apparent, as shown in Fig. 3A for Salmonella spp. and Fig. 3B for E. coli. Further, using each SBA of hydroxycinnamic acid shown in Table 2, the TBA of the 12 hydroxycinnamic acid derivatives in each coffee sample was estimated. However, no significant positive relationship between the TBAs of soluble coffee and the sum of each TBA of hydroxycinnamic acid derivatives was apparent (Fig. 3C). These results suggest the bactericidal activity of coffee was not only explained by CQAs or hydroxycinnamic acid derivatives.

Fig. 3.

Relationships between the SBAs (A, Salmonella; B, E. coli) of soluble coffee and the contents of hydroxycinnamic acid derivatives and between the TBAs of soluble coffee samples and the TBAs of hydroxycinnamic acid derivatives in soluble coffee samples (C; Salmonella spp.). The activities of retail 6 brands of soluble coffee were estimated using Salmonella spp. (A and C) and E. coli (B) in HI broth at pH 3.0 and pH 1.5, respectively. Twelve cinnamic acid derivatives in each sample were analyzed with HPLC, and the TBA of hydroxycinnamic acid derivatives (C) was calculated based on each SBA shown in Table 2.

Discussion

In our previous study (Ueno et al., 2020), we fractionated a soluble coffee sample and showed that the ethyl acetate fraction containing mainly hydroxycinnamic acid derivatives had a higher SBA than 5-CQA. These results strongly suggested the existence of another hydroxycinnamic acid derivative with a higher SBA than 5-CQA. Therefore, in the present study, we prepared 12 kinds of hydroxycinnamic acids, including diCQAs and CFQAs, from green coffee beans and estimated the bactericidal activity of each. However, we could not find any hydroxycinnamic acid derivatives having a significantly higher SBA than 5-CQA to explain the TBA of soluble coffee.

On the other hand, the SBA of a hot water extract of green coffee beans was less than half that of soluble coffee. The TBA of the green coffee bean extract was explained by hydroxycinnamic acid derivatives. These results show that any products formed during roasting have an effect on the activity of soluble coffee.

We have not yet been able to specify a principal factor explaining the bactericidal activity of soluble coffee. There might be a synergistic effect among hydroxycinnamic acids. However, we showed that the activity of green coffee beans is mainly attributed to hydroxycinnamic acid derivatives. Comparison of the ethyl acetate fractions of green coffee beans and soluble coffee showed that the content of hydroxycinnamic acid derivatives in green coffee beans was higher than that of soluble coffee, while the SBA of soluble coffee was higher than that of green coffee beans. Further, no positive relationships between the bactericidal activity of coffee and the content of hydroxycinnamic acid derivatives were apparent (Fig. 3). Considering these results together, there does not seem be a strong synergistic effect among hydroxycinnamic acid derivatives or CQAs for the activity of soluble coffee.

In general, CQAs or hydroxycinnamic acid derivatives are decomposed or polymerized during roasting. Various products are formed during roasting by the Maillard and other chemical reactions. Coffee melanoidins were reported to have antibacterial and bactericidal activities (Rufián-Henares et al., 2009). However, in our previous study (Ueno et al., 2020), there was no correlation between the color of coffee and bactericidal activity. Melanoidins or the colored substances of coffee themselves did not seem to contribute to intense bactericidal activities of coffee.

There are two possible explanations of the bactericidal activity of soluble coffee. One is that an unknown compound formed during roasting or through the Maillard reaction that has greater bactericidal activity. Unfortunately, we have not yet detected such a compound on HPLC. Another is that the synergistic effect of CQAs or hydroxycinnamic acid derivatives and melanoidins or other compounds formed through the Maillard reaction or roasting impacts the bactericidal activity. To clarify the bactericidal activity of soluble coffee or roasted coffee beans at the molecular level, it will be necessary to prepare roasted coffee samples from green coffee beans and examine changes in bactericidal activities during roasting. Further, it will be necessary to prepare a model coffee and examine the synergistic effects of the model substances and 5-CQA.

Another point is the acid resistance mechanism of Salmonella spp. and E. coli. In this mechanism, a few components such as glutamic acid and arginine play a crucial role (Lund et al., 2014). In the present study, we used HI broth, which contains various amino acids and proteins, to estimate the bactericidal activity. If a component disturbs the action of these amino acids in a sample, such a component may have a promotive effect on the bactericidal activity of hydroxycinnamic acids. It will be necessary to use a buffer of a specified composition as the assay medium to analyze the bactericidal effects of CQAs or hydroxycinnamic acid derivatives and coffee in detail.

In conclusion, each CQA or hydroxycinnamic acid derivative in soluble coffee and green coffee beans showed several times higher SBA than soluble coffee. However, there are unknown factors, such as degradation or polymerization products formed during roasting, in explaining the bactericidal effect of soluble coffee or roasted coffee at low pH.

References

•  Beales,  N. (2004). Adaptation of microorganisms to cold temperatures, weak acid preservatives, low pH, and osmotic stress: A Review. Comp. Rev. Food Sci., 3, 1-20.
•  Clifford,  M. (1985). Chlorogenic acids. In “Coffee Volume 1 Chemistry,” ed. by  R. J.  Clarke and  R.  Macrae. Elsevier Applied Science Publishers, Essex, UK, pp. 153-202.
•  Clifford,  M. and  Jarvis,  T. (1988). The chlorogenic acids content of green robusta coffee as a possible index of geographic origin. Food Chem., 29, 291-298.
•  Clifford,  M.,  Kellard,  B., and  Birch,  G. (1989). Characterization of caffeoylferuloylquinic acids by simultaneous isomerization and transesterification with tetramethylammonium hydroxide. Food Chem., 34, 81-88.
•  Dressman,  J.B.,  Berardi,  R.R.,  Dermentzoglou,  L.C.,  Russell,  T.L.,  Schmaltz,  S.P.,  Barnett,  J.L., and  Jarvenpaa,  K.M. (1990). Upper gastrointestinal (GI) pH in young, healthy men and women. Pharm. Res., 7, 756-761.
•  Foster,  J.W. (2004). Escherichia coli acid resistance: tales of an amateur acidophile. Nat. Rev. Microbiol., 2, 898-907.
•  Fujioka,  K. and  Shibamoto,  T. (2008). Chlorogenic acid and caffeine contents in various commercial brewed coffees. Food Chem., 106, 217-221.
•  Fukuyama,  S.,  Watanabe,  Y.,  Kondo,  N.,  Nishinomiya,  T.,  Kawamoto,  S.,  Isshiki,  K., and  Murata,  M. (2009). Efficiency of sodium hypochlorite and calcinated calcium in killing Escherichia coli O157:H7, Salmonella spp., and Staphylococcus aureus attached to freshly shredded cabbage. Biosci. Biotechnol. Biochem., 73, 9-14.
• IUPAC commission on the nomenclature of organic chemistry (CNOC) and IUPAC-IUB commission on biochemical nomenclature (CBN) (1976). Nomenclature of cyclitols. Biochem. J., 153, 23-31.
•  Lianou,  A.,  Nychas,  G.J.E., and  Koutsoumanis,  K.P. (2017). Variability in the adaptive acid tolerance response phenotype of Salmonella enterica strains. Food Microbiol., 62, 99-105.
•  Lund,  P.,  Tramonti,  A., and  Biase,  D.D. (2014). Coping with low pH: molecular strategies in neutralophilic bacteria. FEMS Microbiol. Rev., 38, 1091-1125.
•  Morishita,  H.,  Takai,  Y.,  Yamada,  H.,  Fukuda,  F.,  Sawada,  M.,  Iwahashi,  H., and  Kido,  R. (1987). Caffeoyltryptophan from green robusta coffee beans. Phytochemistry, 26, 1195-1196.
•  Murata,  M.,  Okada,  H., and  Homma,  S. (1995). Hydroxycinnamic acid derivatives and p-coumaroyl-(L)-tryptophan, a novel hydroxycinnamic acid derivative, from coffee beans. Biosci. Biotechnol. Biochem., 59, 1887-1890.
•  Okada,  H.,  Murata,  M.,  Ohyazu,  K. and  Homma,  S. (1997). Distinction between arabica and robusta coffee beans by hydroxycinnamic acid derivatives, especially by p-coumaroyltryptophan. Food Sci. Technol. Int. Tokyo, 3, 205-209.
•  Russell,  T.L.,  Berardi,  R.R.,  Barnett,  J.L.,  Dermentzoglou,  L.C.,  Jarvenpaa,  K.M.,  Schmaltz,  S.P., and  Dressman,  J.B. (1993). Upper gastrointestinal pH in seventy-nine healthy, elderly, North American men and women. Pharm. Res., 10, 187-196.
•  Ueno,  H.,  Tsutsuura,  S.,  Inoue,  A., and  Murata,  M. (2020). Bactericidal effects of coffee and chlorogenic acid on Escherichia coli and Salmonella spp. under low pH or gastric acid conditions. Food Sci. Technol. Res., 26, 247-256 (2020).