Article ID: pjab.101.017
In Japan, serious food poisoning among individuals who took supplement tablets for lowering plasma cholesterol levels have been publicized since late March 2024. The tablets were prepared from red yeast rice (RYR), a product of Monascus pilosus. Puberulic acid (PA) was detected as an unexpected compound in tablets that caused food poisoning. We conducted an on-site investigation at the RYR production factory to determine the cause of PA contamination of the tablets. Fungi capable of producing PA were detected in wipe samples from the factory and were identified as Penicillium adametzioides. To understand the route through which P. adametzioides contaminated RYR and produced PA, coculture experiments with M. pilosus and P. adametzioides were performed. P. adametzioides grew on rice covered with M. pilosus and produced PA. These results suggest that PA-producing P. adametzioides inhabited the RYR production factory and accidently contaminated the culture of M. pilosus. Consequently, RYR tablets contaminated with PA were manufactured and caused the food poisoning outbreak.
In March 2024, health problems, including renal dysfunction, fatigue, and urinary abnormalities, were publicized in Japan as a result of the intake of red yeast rice (RYR) supplement tablets. RYR, called beni-koji in Japan, is a fermented product used as a food or food additive in East Asian countries.1),2) Monascus pilosus, a red pigment-producing fungus, was used to prepare the tablets. Because this fungus produces monacolin K, which inhibits an enzyme involved in the cholesterol biosynthesis pathway, the RYR tablets have been taken by patients with hypercholesterolemia to reduce plasma cholesterol levels.3),4) In October 2024, the Osaka City Government identified the health problems caused by the intake of RYR supplement tablets as food poisoning. Approximately 2,600 people received medical care by the end of November 2024, and approximately 540 required hospitalization.
Besides monacolin K, Monascus spp. produces citrinin, a mycotoxin mainly produced by some Penicillium species.5) Because citrinin causes nephrotoxic effects in experimental animals, taking a tablet contaminated with citrinin was considered to be the cause of renal dysfunction.6) The M. pilosus strain used to prepare RYR was derived from M. pilosus NBRC 4520. Genome analysis showed that the citrinin biosynthetic genes were incomplete in NBRC 4520, and citrinin was not detected in the culture broth of this strain.7) Consistent with this study, the RYR manufacturer did not detected citrinin in any of the RYR lots. To identify the causes of the food poisoning, the manufacturer conducted a nontargeted analysis and found an unexpected compound in some RYR lots in storage. Further analysis revealed that the RYR lots in which the compound was detected were consistent with those related to the food poisoning, and the compound was identified as puberulic acid (PA; Fig. 1A), a secondary metabolite produced by Penicillium species.8)
Chemical structures of metabolites produced by the isolated Penicillium adametzioides. (A) puberulic acid, (B) aspergillusol A (1), (C) lapatin A (2), and (D) glyantrypine (3). The absolute stereochemistry was not determined for 1, 2, and 3.
PA was first discovered in the culture of a corn pathogen, Penicillium puberulum, and its chemical structure was determined in subsequent studies.9),10) Some analogs of PA, namely stipitatic acid, puberulonic acid, and viticoloins, have been found in cultures of Penicillium species.11)-13) A research group in Japan, studying PA as a lead compound for an antimalarial drug, reported antimalarial activity of this compound, with an IC50 of 0.01 μg/mL, and showed that its cytotoxicity against human embryonic lung fibroblasts was relatively low (IC50 ＝ 57.2 μg/mL).13) When PA was administered to mice infected with malarial parasite at a subcutaneous dose of 5 mg/kg, four out of five mice died.14) Based on these results, chemical screening for less-toxic PA derivatives was performed. Except for these studies, no reports on the biological activity of PA are available, and its toxicity in humans is unknown. To assess the toxicity of PA, a 7-day repeated-dose oral toxicity study in rats was performed, and degeneration of proximal tubule epithelial cells was observed in the kidneys (Ministry of Health, Labour and Welfare of Japan, https://www.mhlw.go.jp/stf/seisakunitsuite/bunya/kenkou_iryou/shokuhin/daietto/index_00005.html). Based on this result, PA contamination in the RYR tablets was considered to be the cause of the food poisoning.
The manufacturer detected PA in some lots of RYR produced between February and August 2023 at the Osaka factory in Japan. The factory was closed in December 2023, and the manufacturing facilities were moved to the Wakayama factory in Japan. When the health problems caused by the intake of the RYR tablets were publicized in March 2024, the Osaka factory was not in the same condition as when RYR, which was contaminated with PA, was produced. To determine the cause of PA contamination in RYR, we conducted an on-site investigation at the Wakayama factory where the manufacturing facilities were in operation. The manufacturing area of the factory consists of six rooms: an incubation room for preculture, a preparation room for main culture, a culture tank room, a drying room, and inspection rooms I and II. The manufacturing process of RYR is summarized in Fig. 2. First, M. pilosus was inoculated in a liquid RYR powder medium in the inspection room I, and cultured in an incubation room for preculture. A solid rice medium in a large tank for main culture was prepared both in the preparation room for main culture and the culture tank room. The preculture broth was then inoculated into the solid rice medium to produce RYR. This process was performed in the culture tank room. Sampling and addition of water were performed during the main culture. At the end of the main culture, RYR was autoclaved, dried in the drying room, and stored. Then the RYR was grinded, bagged and shipped to another factory, where it was made into tablets. The active ingredients in RYR were analyzed in the inspection room II.
Manufacturing process of red yeast rice (RYR) and the contamination models reflecting each process. RYR was manufactured according to the process described on the left side. We assumed four routes of Penicillium adametzioides contamination and performed coculture experiments with Monascus pilosus and P. adametzioides. Model 1 is an assumption that contamination occurred at the start of the preculture. Spore suspensions of M. pilosus and P. adametzioides were inoculated in liquid rice and RYR powder media. The culture was then inoculated into the solid rice medium, which was used for the main culture. Model 2 represents a hypothesis that contamination occurred at the start of the main culture. The culture broth of M. pilosus in liquid rice medium was inoculated into solid rice medium with a spore suspension of P. adametzioides. Models 3 and 4 are suppositions that the contamination occurred during and after the main culture, respectively. The spore suspension of P. adametzioides was inoculated into untreated or dried RYR after autoclaving. The growth of M. pilosus and P. adametzioides was investigated in each model.
In this study, we aimed to clarify the mechanism of PA contamination in RYR tablets. For this, we performed two experiments. First, we collected wipe samples from each room in the RYR production factory to screen for PA-producing fungi. Second, to understand the contamination route, we cocultured M. pilosus and PA-producing fungus using four models that reflected the RYR manufacturing process (Fig. 2).
HPLC-grade acetonitrile and water, and special-grade TFA, guaranteed-grade methanol, acetic acid, hydrochloric acid, ethyl acetate, sucrose and polyoxyethylene (20) sorbitan monolaurate were purchased from FUJIFILM Wako Pure Chemical Corp (Osaka, Japan). Deuterated dimethyl sulfoxide-d6 and chloroform-d as NMR solvents were purchased from Merck (Darmstadt, Germany). Polished rice (japonica rice grains produced in Japan) was purchased from a local supermarket in Japan. M. pilosus NBRC 4520, Penicillium adametzioides (strains 11-1, and E-6-2) and RYR supplement tablets were received from Kobayashi Pharmaceutical Co., Ltd (Osaka, Japan). P. adametzioides IFM 68223 was purchased from Medical Mycology Research Center of Chiba University.
2.2. Isolation of Penicillium species in the RYR production factory.Wipe samples were collected from the factory in Wakayama Prefecture, Japan on April 15, 2024. A total of 29 samples from the incubation room for preculture, preparation room for the main culture, culture tank room, drying room, and analysis room were obtained using Wipe Check (Sato Kasei Kogyosho Co., Ltd, Tochigi, Japan). The list of the sampling points is shown in Table 1. Samples were collected by wiping an area of about 100 cm2 with a cotton swab wetted with PBS. Each swab sample was soaked into 10 mL of PBS and transported to our laboratory at room temperature. On the following day, an aliquot of each sample (10 and 100 μL) was spread onto DRBC plates (Thermo Fisher Scientific Inc., Waltham, MA, USA). After an incubation period of three days at 25°C, colonies of Penicillium spp. on the DRBC plates were selected based on macroscopic features. Each of colonies was picked and subcultured on a potato dextrose (PD) agar (Eiken, Tokyo, Japan) for isolation.
List of sampling points in the RYR production factory in Wakayama Prefecture
|
Wipe sample No. |
Sampling point | Room name |
|---|---|---|
| 1 | Top surface of an orbital shaker | Incubation room for preculture |
| 2 | Air conditioner vent | |
| 3 | Wiring on a wall | |
| 4 | Inner surface of a hose at a washing place | Preparation room for main culture |
| 5 | Side surface of a rack for a culture tank | |
| 6 | Top surface of a container | |
| 7 | Inner surface of a water supply hose | |
| 8 | Floor | Culture tank room |
| 9 | Water in a water bath for a culture tank | |
| 10 | Water in a water bath for a culture tank | |
| 11 | Back of a culture tank | |
| 12 | Outer surface of a tube connected with a culture tank | |
| 13 | Outer surface of a tube connected with a culture tank | |
| 14 | Outer surface of a rubber seal for a culture tank | |
| 15 | Inner surface of a washed culture tank | |
| 16 | Outer surface of a water filter for a culture tank | |
| 17 | Outer surface of an air filter for a culture tank | |
| 18 | Outer surface of an exhaust filter for a culture tank | |
| 19 | Outer surface of an exhaust filter for a culture tank | |
| 20 | Inner surface of an exhaust filter for a culture tank | |
| 21 | Inner surface of an air filter for a culture tank | |
| 22 | Inner surface of a drying machine | Drying room |
| 23 | Top surface of a refrigerator | Inspection room I |
| 24 | Bottom surface of a safety cabinet | |
| 25 | Top surface of a safety cabinet | |
| 26 | Top surface of a refrigerator | |
| 27 | Window frame | |
| 28 | Top surface of a safety cabinet | Inspection room II |
| 29 | Top surface of a storage shelf |
Both molecular phylogenetic analysis and morphological observation were performed, and the results were finally combined to identify the isolates. β-tubulin gene sequences were used for molecular phylogenetic analysis as previously described.15) The sequences determined in this study were deposited in GenBank (accession Nos. LC863941-LC863957). All sequences of isolates were automatically aligned using the MUSCULE program with several references downloaded from GenBank.16) Two strains of Penicillium citreonigrum NRRL 761 and Penicillium lividum CBS 347.48 (GenBank accession nos. EF198621 and KM088825, respectively) were used as outgroups. The phylogenetic relationships among Penicillium strains were inferred using the neighbor-Joining method.17) The nodal numbers indicate the bootstrap values based on 1000 replications.18) The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Maximum Composite Likelihood method.19) There were a total of 416 positions in the final dataset. Evolutionary analyses were conducted in MEGA11.20) Morphological observation was performed as follows; Penicillium isolates were inoculated onto Czapek yeast extract agar (35 g/L Czapek-dox broth (BD, NJ, USA), 5 g/L yeast extract (BD), and 20 g/L agar), malt extract agar (Thermo Fisher Scientific), and Czapek agar. After incubation at 25°C for 14 days, Penicillium isolates were identified by previously reported methods based on macroscopic and microscopic morphological features.21),22)
2.4. Analysis of PA production by the isolated Penicillium species.Each of the isolated Penicillium species was cultured in 3 mL of a liquid medium containing 20 g/L yeast extract and 100 g/L sucrose for 24 h at 25°C with shaking at 200 rpm. For preparing the solid rice medium, polished rice (10 g) was soaked in 3 mL of water in a 100 mL Erlenmeyer flask for 1.5 h and autoclaved. The culture broth of each strain and 4 mL water were added to each medium. After incubation for 7 days at 25°C in the dark, PA was extracted by homogenizing the culture with 40 mL of 75％ methanol. The extract was filtered through a filter paper, and 600 μL of the filtrate was dried using a centrifugal evaporator (Micro Vac MV-100, Tomy Seiko, Tokyo, Japan). The residue was suspended in 300 μL of aqueous hydrochloric acid (1 mol/L), and the suspension was extracted twice with the 300 μL of ethyl acetate. The ethyl acetate layer was evaporated to dryness and the residue was suspended in 150 μL of 30％ acetonitrile containing 1％ acetic acid. After centrifugation (12,000 × g, 5 min), 10 μL of the supernatant was subjected to HPLC analysis. The amount of PA produced was described as the amount per rice medium prepared from 10 g of polished rice. For the analysis of PA in the RYR tablets, a tablet was powdered in 2 mL of 75％ methanol, and the extract was purified as described above.
2.5. HPLC conditions for PA analysis.An LC-20A series HPLC system (Shimadzu Corp., Kyoto, Japan), equipped with a Triart C18 column (250 mm × 4.6 mm i.d., 5 μm particle size, YMC Co., Ltd., Kyoto, Japan), was used. Chromatographic separation was achieved at 40°C using two mobile phases: A (water containing 0.1％ TFA) and B (acetonitrile). A linear gradient of 10-80％ of solvent B was run for 25 min at a flow rate of 1.0 mL/min, and 80％ of solvent B was run for 5 min. The photodiode array detector was operated within a 190-500 nm range. The calibration curve for PA was obtained by injecting the standard (0.03, 0.1, 0.3, 1, 3, and 10 μg) and by plotting the peak areas measured at 358 nm.
2.6. Identification of metabolites produced by P. adametzioides.P. adametzioides strain 11-1 was cultured on 10 solid rice media (100 g of polished rice) according to the method described above. The culture was extracted with 400 mL of 75％ methanol, and the extract was filtered through a filter paper. The filtrate was evaporated to dryness using an evaporator and the residue was suspended in 30 mL aqueous hydrochloric acid (1 mol/L). The suspension was extracted three times with 30 mL ethyl acetate. The ethyl acetate layer was evaporated to dryness and the residue was suspended in 3 mL 75％ methanol. The suspension was divided into three equal parts, and each part was subjected to an HF Mega BE-C18 (5 g) cartridge (Agilent Technologies, Palo Alto, CA, USA), pre-equilibrated with 20 mL water containing 0.1％ TFA. Stepwise elution was successively performed using 20 mL water, 10％ acetonitrile, 25％ acetonitrile, and 50％ acetonitrile, each containing 0.1％ TFA. PA was eluted in the 10％ acetonitrile fraction, whereas the unknown metabolites were mainly collected in the 50％ acetonitrile fraction. Each fraction was separately evaporated to dryness. The residue of the 10％ acetonitrile fraction (74 mg) was subjected to analysis on an LC-20A series HPLC system, equipped with a Triart C18 column (250 mm × 10 mm i.d., 5 μm particle size; YMC Co., Ltd.). Isocratic elution with 15％ acetonitrile in water containing 0.1％ TFA at a flow rate of 4.0 mL/min and detection at 220 nm was used to obtain PA (RT, 8.2 min). The residue of the 50％ acetonitrile fraction (70 mg) was subjected to reverse-phase HPLC using the same column. Isocratic elution with 33％ acetonitrile in water containing 0.1％ TFA at a flow rate of 4.0 mL/min and detection at 220 nm was performed to obtain 1 (RT 8.0 min), 2 (RT 9.4 min), and 3 (RT 14.4 min). HR-ESI-TOF-MS data were obtained using a UHPLC system (Dionex Ultimate 3000 RS LC system, Thermo Fisher Scientific) interfaced with a Q Exactive hybrid quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific), according to a previous report.23) All NMR analyses (1H, 13C, 13C DEPT135, 1H-1H COSY, HSQC, HMBC) were performed using a JNM-ECZ600 spectrometer (JEOL Ltd., Tokyo, Japan) at 25°C operating at 600 MHz for 1H, and at 151 MHz for 13C.
2.7. Spore suspension preparation for coculture experiments.To prepare spore suspensions, M. pilosus NBRC 4520 and P. adametzioides strain 11-1 were cultured on a solid rice medium and a PD agar plate, respectively. After 2 weeks, 10 mL of 0.05％ polyoxyethylene (20) sorbitan monolaurate in water was added to each culture and spores were collected using a scraper. The spore suspension was filtered through a single layer of gauze to remove the mycelia. Serial dilutions of spore suspensions were prepared and spread onto PD agar plates. After four days, the colonies that appeared were counted to determine the cfu. Each spore suspension was diluted with 0.05％ polyoxyethylene (20) sorbitan monolaurate in water to prepare working spore suspensions of M. pilosus (3.3 × 102 and 3.3 × 103 cfu/μL) and P. adametzioides (3.3 cfu/μL).
2.8. Coculture experiment for the model assuming contamination at the start of preculture.To prepare the liquid medium, 1.5 g of ground polished rice or ground RYR was added to a 100 mL Erlenmeyer flask containing 30 mL water. After autoclaving, the spore suspensions of M. pilosus and/or P. adametzioides were added to the medium as shown in Fig. 7. The medium was incubated for 5d at 25°C with shaking at 200 rpm. An aliquot of 0.4 mL culture broth was transferred to the solid rice medium with 2 mL water or a mixture of hydrochloric acid (1％) and acetic acid (0.68％) in water to achieve neutral or acidic culture conditions, respectively. After 3 days of static incubation at 25°C, 1 mL water was added to each culture, and the incubation was continued.
2.9. Coculture experiment for the model assuming contamination at the start of main culture.M. pilosus spores (106 cfu) were added to the liquid rice powder medium. After 5 days of incubation at 25°C under shaking conditions (200 rpm), an aliquot of 0.4 mL culture broth was transferred to the solid rice medium with or without the spores of P. adametzioides (100 cfu). To achieve neutral or acidic culture conditions, 2 mL water or a mixture of hydrochloric acid (1％) and acetic acid (0.68％) was added to the medium. After 3 days of static incubation at 25°C, 1 mL water was added to each culture, and the incubation was continued.
2.10. Coculture experiment for the model assuming contamination during and after main culture.Polished rice (100 g) in a 500 mL Erlenmeyer flask was soaked in 30 mL water for 1.5 h and autoclaved. To prepare RYR, the M. pilosus spore suspension (106 cfu) and 20 mL of a mixture of hydrochloric acid (1％) and acetic acid (0.68％) in water was added to the rice medium. After 14 d of static incubation at 25°C, 10 g RYR was dispensed into 100 mL Erlenmeyer flasks. The P. adametzioides spore suspension (100 cfu) was added to the flask, which was untreated, autoclaved or dried at 90°C for 3 h after autoclaving. The static incubation was performed at 25°C.
We observed a major peak at a retention time (RT) of 9.8 min in the methanol extract of RYR tablets from the same lot of the product that caused food poisoning (Fig. 3A). This peak was not observed in the extract prepared from normal tablets (Fig. 3B). The UV spectrum of the peak in the HPLC diode array detector was identical to that previously reported for PA.24) A total of 29 wipe samples were collected from the manufacturing area in the RYR production factory, and each sample was applied to an agar plate. Environmental fungi, except for M. pilosus, were detected from fourteen samples. The representative culture plates are shown in Fig. 4, and other plates are shown in Supplementary Fig. 1. In a wipe sample collected from the preparation room for main culture (Sample No. 6), Penicillium-like colonies were dominant, and white colonies of M. pilosus were also observed (Fig. 4A). Many yeast-like colonies and some Penicillium-like colonies appeared on the plate sprinkled with the wipe sample collected from the inspection room I (Sample No. 26, Fig. 4B). Many Aspergillus- and Penicillium-like colonies appeared on the plate sprinkled with the wipe sample from the inspection room II (Sample No. 28, Fig. 4C). When Penicillium-like strains were picked and subcultured on new PD agar plates, some strains produced a yellow pigment (Fig. 4D), whereas others did not. Both yellow pigment-producing and non-producing strains were selected from each wipe sample, and 17 Penicillium-like strains were cultured on solid rice medium to confirm the production of PA. A peak with RT and UV spectrum consistent with those of PA in the RYR tablets was observed in the culture extract of 10 of the 17 strains. Each stain was classified based on both molecular phylogenetic analysis and morphological observations. The phylogenetic relationships between our sequence results and reference sequences downloaded from GenBank are shown in Fig. 5. Information regarding the isolated fungi is presented in Table 2. Ten strains, which produced yellow pigment on PD agar plates, also produced PA in rice medium at 1.1-7.8 mg/10 g of rice. All of the PA-producing strains form a monophyletic clade with the type strain of Penicillium adametzioides CBS 313.59. This result was supported by bootstrap value of 99％ (Fig. 5). Furthermore, several previously isolated strains of P. adametzioides including IFM 68223 are clustered in the clade. The RYR manufacturer collected wipe samples from the Osaka factory, and isolated a PA-producing fungus, E-6-2, from them. The strain is also clustered in the monophyletic clade which contains the PA-producing strains isolated from the Wakayama factory. The PA-producing strains isolated from the Wakayama and Osaka factories were all identified as P. adametzioides.
HPLC chromatograms of the extract from red yeast rice (RYR) tablets. (A) An RYR tablet from the same lot of the product that caused the food poisoning was extracted with 75％ methanol. The ethyl acetate layer after liquid-liquid extraction was subjected to HPLC analysis; the chromatogram obtained at 220 nm is shown. The UV spectrum of the main peak (*) on the HPLC diode-array detector is also shown. (B) The HPLC chromatogram of the extract from a normal RYR tablet.
Representative photographs of culture plates of fungi obtained from the production factory of the red yeast rice. The top surface of a (A) container in the preparation room for main culture (wipe sample No. 6), (B) refrigerator in the inspection room I (wipe sample No. 26), and (C) safety cabinet in the inspection room II (wipe sample No. 28) were wiped with cotton swabs. The suspensions obtained by rinsing the swabs were spread onto DRBC plates, and the plates were incubated at 25°C for 6 days. (D) The fungal colonies that appeared were isolated onto another culture plate. Penicillium-like fungi, which produced yellow pigment on the plate, were mainly obtained.
Phylogenetic tree among Penicillium strains based on partial sequences of β-tubulin. The bootstrap values are indicated at each node. Only values ＞50％ are shown. The strains marked by black circles were isolated from the wipe sample collected at the red yeast rice production factory in this study. The letter T indicates type strain. The out groups are Penicillium citreonigrum NRRL 761 and Penicillium lividum CBS 347.48.
Information about fungi isolated from wipe samples
| Strain ID | Yellow pigment | Puberulic acid productivity (mg/10 g of rice) | Taxonomy | Source | |
|---|---|---|---|---|---|
| Wipe sample No. | Room name | ||||
| N08-12 | ＋ | 5.6 | Penicillium adametzioides | 6 | Preparation room for main culture |
| N08-13 | － | ＜0.03 | Penicillium jugoslavicum | ||
| N08-14 | ＋ | 7.8 | Penicillium adametzioides | ||
| N10-16 | ＋ | 6.6 | Penicillium adametzioides | 8 | Culture tank room |
| N10-17 | － | ＜0.03 | Penicillium jugoslavicum | ||
| 11-1* | ＋ | 2.9 | Penicillium adametzioides | - | |
| S01-3 | ＋ | 2.9 | Penicillium adametzioides | 25 | Inspection room I |
| S01-4 | － | ＜0.03 | Penicillium section Citrina | ||
| S02-7 | － | ＜0.03 | Penicillium jugoslavicum | 26 | |
| S02-10 | ＋ | 4.2 | Penicillium adametzioides | ||
| S03-11 | ＋ | 4.5 | Penicillium adametzioides | 27 | |
| S03-12 | － | ＜0.03 | Talaromyces sp. | ||
| S04-15 | － | ＜0.03 | Penicillium section Citrina | 28 | Inspection room II |
| S04-16 | － | ＜0.03 | Penicillium section Exilicaulis | ||
| S04-18 | ＋ | 5.8 | Penicillium adametzioides | ||
| S05-22 | ＋ | 4.2 | Penicillium adametzioides | 29 | |
| S05-24 | ＋ | 1.1 | Penicillium adametzioides | ||
* This strain was provided by the RYR manufacturer.
Representative HPLC chromatograms of the metabolites produced by P. adametzioides are shown in Fig. 6. In addition to the PA peak, three peaks of unknown compounds 1 (RT, 16.7 min), 2 (RT, 17.2 min), and 3 (RT, 18.9 min) were observed in the metabolite chromatograms of P. adametzioides S02-10, N08-12 and S05-22 strains isolated from the Wakayama factory (Fig. 6A and Supplementary Fig. 2). The UV spectra of 1, 2, and 3 were similar for the three strains (Fig. 6B, 6C, 6D, and Supplementary Fig. 2). On the metabolite chromatograms of P. adametzioides strain E-6-2, which was isolated from the Osaka factory, and P. adametzioides IFM 68223, which is a stock culture preserved at Chiba University, Japan, peaks of PA and compounds 1, 2, and 3 were also observed (Fig. 6E, 6F). The unknown compounds were identified to determine the characteristics of these strains. P. adametzioides strain 11-1 was cultured on solid rice medium prepared from 100 g of polished rice for one week. The methanol-water extract of the culture was purified by liquid-liquid extraction, a C18 cartridge, and reverse-phase HPLC to isolate PA (18 mg), 1 (15 mg), 2 (18 mg), and 3 (11 mg). The Q-TOF LC-MS and NMR spectra of the purified compounds confirmed that 1, 2, and 3 were aspergillusol A, lapatin A, and glyantrypine, respectively, referring previous reports (Fig. 1B, 1C, 1D).25)-27) The LC-MS and NMR data are presented in Supplementary Table 1, 2, 3, 4 and 5. HPLC chromatograms of metabolites from other Penicillium strains are shown in Supplementary Fig. 3. Compounds 1, 2, and 3 were detected all P. adametzioides strains isolated in this study. To investigate the stability of these compounds, the culture of P. adametzioides was autoclaved, or dried after autoclaving. Signal intensities of compounds 1, 2, and 3 were reduced by autoclaving, while the effect of autoclaving and drying on PA was small as compared with the three metabolites (Supplementary Fig. 4).
HPLC chromatograms of the metabolites produced by the isolated Penicillium adametzioides strains. (A) P. adametzioides strain S02-10, isolated from the inspection room I in Wakayama factory, was cultured on a solid rice medium, and metabolites were extracted with 75％ methanol. The ethyl acetate layer after the liquid-liquid extraction was subjected to HPLC analysis; the chromatogram at 220 nm is shown. * indicates the peak of puberulic acid. The peaks of unknown compounds 1, 2, and 3 were also observed. UV spectra of 1 (B), 2 (C), and 3 (D). (E) HPLC chromatogram of the metabolites from P. adametzioides strain E-6-2 (an isolate from Osaka factory). (F) HPLC chromatogram of the metabolites from P. adametzioides IFM 68223 (a stock culture preserved in Chiba University). These two strains also produced puberulic acid (*), 1, 2, and 3.
The manufacturing process of RYR is shown in Fig. 2. In this process, four contamination routes of P. adametzioides were assumed: the start of preculture, start of the main culture, during the main culture, and after the main culture. To understand the contamination route of PA in RYR, coculture experiments with M. pilosus and P. adametzioides were performed using four models that reflected each route (Fig. 2).
First, the model assuming contamination at the start of preculture (Model 1) was tested. To determine whether P. adametzioides could grow in a liquid medium in the presence of M. pilosus, a spore suspension of M. pilosus and P. adametzioides was added to the medium simultaneously and the growth was investigated. In the manufacturing process of RYR, a liquid RYR powder medium was used for preculture. As shown in Fig. 2, the liquid RYR powder medium is cloud dark red, and it is difficult to monitor the fungal growth visually. We used not only a liquid RYR powder medium but also a liquid rice powder medium. As the fungi grow, pigments are produced in the medium. Because the liquid rice powder medium is white, fungal growth can be seen by color change of the medium. The liquid rice powder medium inoculated with 100 cfu of P. adametzioides or 104 cfu of M. pilosus turned from white to canary yellow or rose pink, respectively (Fig. 7A ID: I and II). When 100 cfu of P. adametzioides and 104 cfu of M. pilosus were inoculated, the medium turned chrome yellow (Fig. 7A ID: III). The rice powder medium inoculated with 106 cfu of M. pilosus turned ruby red, regardless of the presence or absence of P. adametzioides (Fig. 7A ID: IV and V). When the spore suspensions of P. adametzioides and/or M. pilosus were inoculated into the liquid RYR powder medium, the color of the medium did not change from the initial cloud dark red (Fig. 7B). Next, the culture broth in the liquid RYR powder medium was transferred to the solid rice medium to determine the dominant fungi in the medium. As the main culture in the manufacturing process of RYR was performed on acidified rice medium, we performed the coculture experiment on solid rice medium under both neutral and acidic conditions. The acidic condition was created by adding a mixture of hydrochloric acid and acetic acid (pH 0.5) to the solid rice medium. The neutral solid rice medium inoculated with the liquid RYR powder medium is shown in Fig. 7C. When the liquid RYR powder medium containing P. adametzioides was added, the rice was covered with green spores of P. adametzioides (Fig. 7C ID: I). RYR was produced when each of the two liquid RYR powder medium samples, in which only M. pilosus grew, was inoculated (Fig. 7C ID: II and IV). When each of the four coculture samples was inoculated, only P. adametzioides was observed to be growing, and the rice was covered with green spores, regardless of the cfu of M. pilosus (Fig. 7C ID: III and V). PA was detected in these rice media covered with green spores (Supplementary Fig. 5). In contrast to the results obtained using the neutral solid rice medium, P. adametzioides could not grow on the acidified solid rice medium when the liquid RYR powder medium sample with P. adametzioides was inoculated (Fig. 7D ID: I). Unlike P. adametzioides, M. pilosus grew on the acidified solid rice medium (Fig. 7D ID: II and IV). When each of the four coculture samples was added to the acidified solid rice medium, only M. pilosus grew over a 9-day culture period (Fig. 7D ID: III and V). Continuous incubation resulted in the growth of P. adametzioides, and the RYR was covered with green spores (Fig. 7E ID: III). PA was detected in these samples (3.8 mg/10 g of rice for ID: III left and 4.0 mg/10 g of rice for ID: III right, Supplementary Fig. 6). PA was also detected in the RYR which were partially covered with the green spores (0.4 mg/10 g of rice, Fig. 7E ID: V left), while PA was not detected in the RYR on which P. adametzioides slightly grew (Fig. 7E ID: V right, Supplementary Fig. 6). When the culture broth in the liquid rice powder medium was added to the solid rice medium, similar results were obtained as compared with the above-mentioned results of the liquid RYR powder medium (Supplementary Fig. 7).
Results of coculture experiment, model 1. Spore suspensions of Monascus pilosus and/or Penicillium adametzioides were inoculated in the liquid rice powder medium (A) and the liquid red yeast rice (RYR) powder medium (B). The spore numbers in each medium are mentioned in the table on the top. After 5-day incubation, the culture broth was observed. For sample IDs III and V, two samples each were prepared. (C) Each culture broth of the liquid RYR powder medium was inoculated in a neutral solid rice medium. After 7-day incubation, the culture media were observed. (D) Each culture broth of the liquid RYR powder medium was inoculated into the acidified solid rice medium. After 9-day incubation, the culture media were observed. (E) The incubation of the samples on acidified solid rice medium was continued until day 16, and the culture media were observed. Enlarged views are shown for two samples which were partially covered by the green spores. A white arrow indicates the RYR covered with P. adametzioides.
Second, the model assuming contamination at the start of the main culture (Model 2) was tested. To determine whether P. adametzioides could grow with M. pilosus in the solid rice culture medium, the culture broth of M. pilosus in the liquid rice powder medium and the spore suspension of P. adametzioides (100 cfu) were simultaneously inoculated into the solid rice medium and the growth was investigated. As in Model 1, the culture on solid rice medium was performed under both neutral and acidic conditions. Under the neutral condition, the rice medium in which 100 cfu of P. adametzioides was added was covered with green spores on day 7 (Fig. 8A ID: I), whereas RYR was produced from the rice medium in which the culture broth of M. pilosus was inoculated (Fig. 8A ID: II). In the cocultured rice medium, both fungi grew on the medium by day 7, but the spores of P. adametzioides covered the rice medium by day 14 (Fig. 8A ID: III). Under the acidic condition, the growth of P. adametzioides was not evident on day 7 (Fig. 8B ID: I), whereas M. pilosus extended its mycelia on the rice medium (Fig. 8B ID: II). Only the growth of M. pilosus was observed in the cocultured rice medium. The results on day 14 were similar to those on day 7 (Fig. 8B ID: III). This coculture experiment was conducted twice, and similar results were obtained (Supplementary Fig. 8).
Results of coculture experiment, model 2. (A) Spore suspension of Penicillium adametzioides and culture broth of Monascus pilosus in liquid rice powder medium were inoculated in neutral solid rice medium, in accordance with the table on the top. For sample ID III, three samples were prepared. After 7- and 14-day incubation, the culture media were observed. (B) Spore suspension of P. adametzioides and culture broth of M. pilosus in liquid rice powder medium were inoculated in acidified solid rice medium, in accordance with the table on the top. For sample IDs I and III, two and three samples were prepared, respectively. After 7- and 14-day incubation, the culture media were observed.
Third, the model assuming contamination during the main culture (Model 3) was tested. It was considered possible that P. adametzioides contaminated the culture at the time of sampling and addition of water during the main culture. When the P. adametzioides spore suspension (100 cfu) was inoculated on RYR, the growth of P. adametzioides was not observed during the 14-day culture period (Fig. 9A). After a 35-day culture period, a small number of rice grains with mycelium of P. adametzioides was found (Fig. 9B). This coculture experiment was conducted twice, and similar results were obtained. (Supplementary Fig. 9A).
Results of coculture experiment, models 3 and 4. (A) Spore suspension of Penicillium adametzioides was inoculated in red yeast rice (RYR). After 14-day incubation, the culture media were observed. After 35-day incubation, a small amount of rice grains with P. adametzioides mycelium was found on the RYR (B). Spore suspension of P. adametzioides was inoculated on dried RYR after autoclaving (C) and autoclaved, but not dried, RYR (D). After 14-day incubation, the culture media were observed. Each experiment was performed in triplicate.
Finally, the model assuming contamination after the main culture (Model 4) was tested. During the manufacturing of RYR, after the main culture, RYR was autoclaved, dried, grinded, bagged and stored until shipping. To deduce whether the contamination occurred after the main culture, the P. adametzioides (100 cfu) spore suspension was inoculated on the dried RYR after autoclaving and the growth was investigated. No growth of P. adametzioides was observed on the RYR after 14 days (Fig. 9C). In contrast, when P. adametzioides was inoculated on the autoclaved, but not dried RYR, the green spores of P. adametzioides covered the RYR (Fig. 9D), and PA (2.2-3.0 mg/10 g of RYR) was detected. This coculture experiment was conducted twice, and similar results were obtained (Supplementary Fig. 9B and 9C).
In Japan, foods with functional claims (FFCs), which are allowed to label specified health effects, have been on the market since 2015. Many FFCs, including fermented milk drinks, blue berry extract capsules, and RYR tablets, have been developed, and the size of the FFC market was approximately 700 billion yen in 2023.28) This is the first instance of serious food poisoning caused by an FFC reported since the inception of FFCs.
Because the causative substance of the food poisoning is considered to be PA, which was detected in the tablets, PA producers in the RYR production factories were searched. P. adametzioides strains that produce PA in rice medium were isolated from both the Osaka and Wakayama factories. The phylogenetic clade and main metabolites were common between the isolates from the two factories. As mentioned in the Introduction section, RYR contaminated with PA was manufactured in the Osaka factory, and then the manufacturing facilities were moved to the Wakayama factory. The P. adametzioides strain that inhabited the Osaka factory apparently caused PA contamination in the tablets and might move to the Wakayama factory with the facilities.
P. adametzioides is an environmental fungus that causes fruit rot. The first description of P. adametzioides was found in an article published in 1956 by a Japanese researcher.29) Subsequently, strains were isolated from decayed fruits, including grapes and pomegranates, by Korean, Italian, and Pakistani researchers.22),30),31) A research group in China analyzed the metabolites of P. adametzioides AS-53 isolated from a marine sponge.32),33) The strain produced indole diketopiperazine and quinazoline derivatives, including lapatin A and glyantrypine, which were also produced by the P. adametzioides strains in the present study. However, PA production by P. adametzioides AS-53 was not reported. The metabolite production by P. adametzioides in our study differed from that of the AS-53 strain. In contrast, P. adametzioides IFM 68223, a stock culture preserved at Chiba University, Japan, produces PA. This strain was isolated from a lung specimen in 2022. PA, aspergillusol A, lapatin A, and glyantrypine were detected in the IFM 68223 culture extract. The P. adametzioides strain that produces PA and three other compounds appears to be distributed in Japan. In vitro bioactivities of these three compounds were reported. Aspergillusol A is an inhibitor against α-glucosidase, and showed weak cytotoxic activity toward human cancer cell lines.25) Lapatin A exhibited anti-inflammatory activity against NO production induced by lipopolysaccharide.34) Glyantrypine showed antiviral activity against influenza virus A (H1N1).27) However, studies about in vivo bioactivity have not been conducted yet, and their effects on animals are unknown. Although the P. adametzioides strains isolated in this study mainly produced four compounds on the solid rice medium, the peaks of aspergillusol A, lapatin A, and glyantrypine were not observed in the HPLC chromatogram of the extract from the RYR tablets, which caused the food poisoning (Fig. 3A). The result in Supplementary Fig. 4 showed that PA was resistant to autoclave, and metabolites other than PA might been degraded by autoclaving. Other possibility is that differences in culture conditions between the laboratory and factory may have affected the production patterns of secondary metabolites produced by P. adametzioides.
To understand the contamination route of PA in RYR, coculture experiments were performed. P. adametzioides (100 cfu) was used in each experiment as a model of fungal contamination. In Model 1, the color of the liquid rice powder medium in which M. pilosus and P. adametzioides were cocultured differed from that of the medium in which only M. pilosus was cultured. When the coculture samples in both liquid rice and RYR powder media were inoculated in the neutral solid rice medium, P. adametzioides was predominant, and the growth of M. pilosus was inhibited. This indicated that P. adametzioides could exist together with M. pilosus in the liquid rice and RYR powder media, even if the initial cfu of P. adametzioides was much less than that of M. pilosus. In contrast to the results obtained under neutral conditions, M. pilosus grew predominantly on the acidified solid rice medium. In general, both Penicillium sp. and Monascus sp. tolerate moderately acidic conditions.35),36) However, when the pH of the rice medium is reduced by an acidic solution containing acetic acid, M. pilosus grows selectively on the medium.37) Contamination by other organisms during RYR production has been prevented by exploiting this property of M. pilosus. In acidified rice media inoculated with the coculture samples, the RYR was covered with green spores of P. adametzioides (Fig. 7E ID: III and V). After M. pilosus covered the acidified rice medium and pH of the medium might be increased, the surviving P. adametzioides grew on M. pilosus. and produced PA.
In Model 2, although only 100 cfu of P. adametzioides was inoculated into the neutral solid rice medium in which the broth of fully cultured M. pilosus was added, P. adametzioides grew predominantly on the medium. The growth rate of P. adametzioides on the neutral solid rice medium was much faster than that of M. pilosus. However, when 100 cfu of P. adametzioides was inoculated into the acidified solid rice medium, the growth of P. adametzioides was not observed, regardless of the presence of M. pilosus. This indicates that, compared to M. pilosus, survival under acidic conditions is quite difficult for P. adametzioides.
In Model 3, spore formation of P. adametzioides was not observed on raw RYR on the 14th day. In contrast, autoclaved RYR was covered by green spores of P. adametzioides for the same culture period. Live M. pilosus inhibited the growth of P. adametzioides, and slight mycelial growth of P. adametzioides was finally observed on the 35th day.
In Model 4, P. adametzioides grew on autoclaved RYR but not on dried RYR. This indicates that sufficient moisture is essential for the growth of P. adametzioides.
In conclusion, P. adametzioides grew on the rice medium with M. pilosus and produced PA in some culture conditions of our coculture experiments. This indicates that the cause of PA contamination in RYR was the presence of P. adametzioides in the RYR production factory. However, it is difficult to determine the contamination route of PA in the RYR manufacturing process only from our results. In the RYR production factory, the main culture was performed in a large tank. The culture condition in the tank is quite different from that in a small Erlenmeyer flask used in our experiment. For example, the pH of the rice medium in the tank might be inconsistent, and P. adametzioides might have grown in the neutral area. Therefore, the contamination of P. adametzioides in the RYR manufacturing process could have occurred at any points from the starting of preculture to after main culture.
To uncover the complete truth underlying the food poisoning outbreak, it is necessary to confirm whether PA can cause nephrotoxicity in humans. Furthermore, we are currently working on acquiring information on the distribution of PA-producing fungi to prevent the recurrence of food poisoning.
We would like to thank Editage (www.editage.jp) for English language editing.
T.Y. Conceptualization, Data curation, Investigation, Methodology, Writing - original draft, Writing - review and editing. M.W. Investigation, Formal analysis. W.A., S.T., and N.M. Investigation. M.I. Funding acquisition, Supervision. T.O. Project administration, Writing - review and editing. All authors have read and approved the final version of the manuscript.
The authors declare no conflicts of interest.
This study was supported by a research grant from the Ministry of Health, Labour and Welfare of Japan [grant number JPMH22KA2001].
Supplementary materials are available at https://doi.org/10.2183/pjab.101.017.
Edited by Sakayu SHIMIZU, M.J.A.
Correspondence should be addressed to: T. Yoshinari, Division of Microbiology, National Institute of Health Sciences, 3-25-26 Tonomachi, Kawasaki-ku, Kawasaki, Kanagawa 210-9501, Japan (e-mail: t-yoshinari@nihs.go.jp).
colony forming units
DRBCdichloran rose-bengal chloramphenicol
HPLChigh performance liquid chromatography
TFAtrifluoroacetic acid
NMRnuclear magnetic resonance
PBSphosphate buffered saline
PDpotato dextrose
RYRred yeast rice
