ASPERGILLUS LUCHUENSIS, AN INDUSTRIALLY IMPORTANT BLACK ASPERGILLUS IN EAST ASIA

1Several species belonging to Aspergillus section Nigri are associated with food fermentations in East Asia. For example, A. luchuensis and A. awamori (black-koji molds) are linked with the production of awamori, a distilled alcoholic beverage made on Okinawa island in Japan, A. kawachii and A. coreanus (white-koji molds) with the making of shochu and makgeolli,,,,,. The black and white-koji molds are used to make koji (moldy material) for awamori, shochu and makgeolli fermentations, which provides various enzymes for maceration and saccharification of raw materials such as rice, barley, and sweet potatoes, as well as a large amount of citric acid for maintaining the fermentation mash at low pH to prevent from contamination of wild microorganisms.

2The taxonomy of section Nigri is revised various times,, and western taxonomists have accepted other species in this section than mycologists in East Asia. For example, Sakaguchi et al. and Murakami accepted A. luchuensis, and recently, Yamada et al. proposed that industrial black-koji molds, including A. kawachii (an albino mutant of black-koji mold) be named A. luchuensis. In contrast, western mycologists have not accepted A. luchuensis because there has been confusion about the validity of the name. Inui, described A. luchuensis as a mold used for the production of awamori on the Okinawa islands of Japan, but it is considered doubtful because it was described with uniseriate conidial heads, while biseriate heads can be observed in the ex-type culture. Samson et al. and Varga et al. re-described A. acidus for strains used for awamori fermentations. This name is based A. aureus var. acidus Nakazawa et al., isolated from awamori-koji in Okinawa in 1936. Mogensen et al. reported that A. acidus is the dominant microorganism in the Puerh tea which is a variety of post-fermented tea produced in Yunnan province in China.

3Raper and Fennell used CBS 557.65 ( = NRRL 4948), which does not originate from awamori Koji, for their description of A. awamori, and subsequently Al-Musallam designated the strain as neotype of A. niger var. awamori. Based on this neotype, Perrone et al. re-established A. awamori as a phylospecies in Aspergillus section Nigri based on multigene sequencing of isolates found on grapes. However, the name, A. awamori implies that this species is associated with black koji fermentations and awamori production. In fact, Perrone et al. demonstrated in his study that the A. awamori strains used in the Japanese koji fermentation do not belong to his proposed A. awamori phylospecies.

4The taxonomic position of species used in black- and white-koji fermentations is re-investigated. We studied the original descriptions of A. acidus, A. kawachii, A. luchuensis and A. coreanus and compared the available ex-type strains of these species with other species belonging to section Nigri using ITS, β-tubulin and calmodulin gene sequences and RAPD profiles. Furthermore, the taxonomic position of Aspergillus awamori was re-investigated.

5Ex-types of A. luchuensis, A. acidus and A. kawachii were obtained from NBRC (NITE Biological Resources Center, Japan) and compared with cultures deposited at CBS (CBS-KNAW Fungal Biodiversity Centre, the Netherlands), KACC (Korean Agricultural Culture Collection, Korea) and IBT (Center for Microbial Biotechnology, Denmark). A strain of A. kawachii (KACC 46516) which is used for rice-koji for making makgeolli in Korea, and the ex-type strain of A. coreanus (KACC 41731T) were obtained from the Chungmu company (Korea, Ulsan) and Dr Tae Shick Yu (Keimyung University, Taegu, Korea) respectively. Additionally, two Aspergillus strains were isolated from traditional nuruk (KACC 46420) and meju (KACC 46490) in Korea. Detailed information of the strains is given in Table 1.

6The morphological characters were examined by the methods described in Varga et al. [2011]. In addition, the ex-type culture of A. luchuensis was grown on steamed rice for microscopy to compare the observations by Inui, who used this medium to describe the species. The RAPDs in this study were performed according to Hong et al.. In order to determine phylogenetic relationship of the strains, partial fragments of the β-tubulin, calmodulin and ITS were sequenced by the methods of Peterson and combined. For comparing with the other species in Aspergillus section Nigri, sequences were obtained from Varga et al. and combined. Combined DNA data were analyzed using Tamura-Nei parameter distance calculation model, which was then used to construct the Neighbor-Joining (NJ) tree with MEGA version 5. Newly generated sequences were deposited in GenBank under accession numbers JX500062 - JX50090 (Table 1).

7Sequence comparison between A. awamori sensu Perrone et al., A. welwitschiae and A. niger strains was conducted on the basis of beta-tubulin, calmodulin, translation elongation factor-1 alpha data from Perrone et al., and by adding RNA polymerase II sequences amplified using primers 5F and 7CR. The analysis was conducted on a total of 100 strains grouped in A. awamori clade and 20 strains grouped as A. niger [Perrone, unpublished data]. The position of the sequence difference were fixed by comparison to the complete sequences of each gene recovered from A. niger CBS 513.88 genome; for calmodulin gene the NCBI ID = NT_166539, for β-tubulin NCBI ID = AM270165.1, for translation elongation factor-1 NCBI ID = NT166533 and for RPB2 NCBI ID = XM_001395124.2.

8For the extrolite analysis cultures were grown on the agar media CYA and YES for 7 days at 25°C prior to extraction. Extrolites were analyzed by HPLC using alkylphenone retention indices and diode array UV-VIS detection as described by Houbraken et al..

9The morphology on agar media of the ex-type of A. luchuensis (KACC 46772 T = NBRC 4281 T), A. kawachii (KACC 46771T = NBRC 4308T, KACC 46516) and A. acidus (KACC 45131T, KACC 41731) proved to be similar (Table S1, Figure S1), although the colony colors varied. The above five strains have shorter conidiophores and less roughened conidia than A. niger (KACC 45072T) and A. tubingensis (KACC 46805T). All strains including the ex-type of A. luchuensis have biseriate conidial heads on Malt Extract Agar (Figure S1), although aberrant conidial heads could be observed. No morphological differences were observed between the ex-types of A. luchuensis and A. acidus (Table S1). A detailed morphological study of the ex-type strain of A. luchuensis grown on steamed and sterilized rice for three days at 25°C, showed that the conidial heads were often uniseriate (Fig. 1). These conidial heads structures strongly resemble those illustrated by Inui,. After five days, the conidial heads were more frequently biseriate although irregular formation of metulae could be observed.

10In our RAPD experiments with the primers PELF and URP1F, the ex-type strain of A. luchuensis (KACC 46772 T), two strains of A. kawachii (KACC 46771 T and 46516), two strains of A. acidus (KACC 45131 T, 41731) and two strains of an Aspergillus sp. (KACC 46420, 46490) showed identical band patterns (Fig. 2). The band patterns of them were similar to those of two strains of Aspergillus sp. (KACC 45132, 45133), but were different from those of A. niger (KACC 45072T) and A. tubingensis (KACC 46805T).

11The combined analysis of β-tubulin, calmodulin and ITS sequences show that the strains of A. luchuensis, A. kawachii, A. acidus and Aspergillus sp. have 100% sequence homology and they are positioned distantly from A. tubingensis and A. niger (Fig. 3). The neotype of A. awamori CBS 557.65NT is on well-supported branch (100% bootstrap value) with the type strain of A. niger CBS 554.65T.

12A. awamori strains isolated from oriental food fermentation process could be accommodated in four groups. Of 38 A. awamori strains, 24 strains clustered into A. niger group, 10 into A. luchuensis group and two into A. tubingensis. Two isolates could be classified as A. awamori sensu Perrone et al. (Table 2). These two strains did not originate from awamori-koji.

13Most isolates of Aspergillus luchuensis specifically produce antafumicin and/or a partially characterized metabolite provisionally called “luchuensin”, in addition to asperazine, they often produce atromentin, funalenone, pyranonigrin A, and occasionally tensidol B (Table S2).

14Inui, described A. luchuensis as a major fermentation agent in awamori-koji in Okinawa. He reported that the species had blackish brown, uniseriate conidial heads and finely roughened conidia. Nakazawa, also isolated molds from awamori koji and he found black and golden yellow cultures and named them A. awamori and A. aureus, respectively. He examined the A. luchuensis strain obtained from T. Inui, and reported that the strain had biseriate conidial heads and had similar morphological features as A. awamori. He rejected A. luchuensis because it was described as a uniseriate species,,. However, Thom and Raper accepted A. luchuensis and accommodated the species in the Aspergillus luchuensis series. They described it as an Aspergillus with uniseriate heads with occasionally biseriate sterigmata. In contrast, Nehira considered A. luchuensis a synonym of A. awamori. Raper and Fennell reported that A. luchuensis was accepted by Japanese workers as a biseriate species but based on the original description of A. luchuensis with uniseriate conidial heads they rejected this taxon and accepted A. awamori with NRRL 4948 ( = CBS 557.65) as representative culture. Nevertheless, A. luchuensis has been used by Japanese mycologists, and Murakami suggested that if A. luchuensis was corrected as a biseriate species, it was reasonable that A. luchuensis was used as the main awamori koji mold. Recently, Yamada et al. proposed that industrial black-koji molds should be classified as A. luchuensis.

15There has been confusion about the existence of the ex-type culture of A. luchuensis, but a study of the data of strain deposition in NBRC showed that Inui, sent his strain to K. Sakaguchi at the ACTU (now ATU) collection. Our observations of this ex-type of A. luchuensis (NBRC 4281T) grown on steamed rice confirm the original description of Inui (Fig. 1). In our opinion, Inui, described the species at a young stage. This can be seen from his illustrations, where young, not fully developed conidiophores were depicted. Secondly, Inui,, in the description in the Japanese article, mentioned the single seriation of the conidial head, but he did not place importance to this feature. In his German article he did not mention the seriate structure of the conidial head either. An indication that the conidial heads could be biseriate was, that Inui, described that the phialides (sterigmata) were longer than other Aspergillus species. Based on the data mentioned above, A. luchuensis should be accepted as a valid species.

16Aspergillus kawachii, a white-koji mold that has been used widely in shochu making on the Kyusyu in Japan, which was probably formed by the mutation of a certain black Aspergillus species, is considered similar with A. luchuensis,. Brewers have used the name A. kawachii, but Japanese mycologists have used the name, A. luchuensis mut. kawachii,,, because A. kawachii was not validly published. Yamada et al. proposed that A. kawachii should be classified as A. luchuensis based on their multilocus sequence typing of ITS, D1D2 of LSU, histon3, β-tubulin and cytochrome b sequence data. Our results of RAPDs and MLST of ITS, β-tubulin and calmodulin also support the findings of Kitahara and Kurushima and Yamada et al.. Like A. luchuensis, also A. kawachii was described with uniseriate conidial heads. In our opinion, the same situation and observations occurred here as with the description of A. luchuensis. In conclusion, we propose to place A. kawachii in synonymy with A. luchuensis.

17Nakazawa et al. isolated black-koji molds from various awamori fermentations and proposed A. awamori with five varieties, A. aureus with five varieties and A. miyakoensis. One of the five varieties of A. aureus is variety acidus, which was renamed as A. foetidus var. acidus by Thom and Raper [21 Later the variety was raised to species level by Kozakiewicz based on morphological features including the conidial ornamentation observed by scanning electron microscopy. Varga et al. re-validated the species. Our RAPD data and multigene sequence data show that A. acidus shows no difference with A. luchuensis and A. kawachii, and based on priority, the name A. luchuensis is proposed as the correct name for these strains.

18From our study we can conclude that A. luchuensis is the correct name for the fungus used in awamori fermentations and that A. acidus and A. kawachii are synonyms. Since there has been confusion about the uni- or biseriate feature of the conidial heads of this species, A. luchuensis is re-described below.

19Aspergillus luchuensis Inui,.

20= Aspergillus perniciosis Inui,

21= Aspergillus awamori Nakazawa pro parte

22= Aspergillus aureus var. acidus Nakaz., Simo & A. Watan.

23= Aspergillus foetidus var. acidus (Nakaz., Simo & A. Watan.) Thom & Raper

24= Aspergillus kawachii Kitahara & Yoshida

25= Aspergillus acidus Kozak.

26= Aspergillus coreanus Yu et al.

27Colony diameters at 7 days, in mm: CYA, 37; DG18, 33; CREA, 29; MEA, 53; MEA37°C, 50.

28On CYA, sterile mycelium white, conidial areas gray to black, radially sulcate, floccose; reverse cream to brown, radially sulcate. On MEA, obverse, similar to CYA. On CREA, poor growth but with strong acid production.

29Conidial heads, radiate; stipes length up to 1.5 mm, width 10–13 µm, walls thick, smooth and hyaline; Vesicles, 20–40 µm wide, nearly spherical; when young uniseriate with irregular formation of metulae, later predominantly biseriate, metulae covering almost entire surface of the vesicle 17.0–26.1 × 4.5–8.1 µm, phialides ampulliform 5.6–8.4 × 3.5–4.9 µm; Conidia, globose, smooth, 3.5–4.5 µm, dark brown.

30Type strain: NBRC 4281 = KACC 46772 = CBS 205.80 = IFM 47726 = RIB 2642, isolated from awamori-koji in Okinawa, Japan, sent by T. Inui to K. Sakaguchi (University of Tokyo).

31Note: The description above is based on the ex-type strain. As shown in Table S1 and Figure S1 A. luchuensis strains show variable morphological characteristic. The colony color is white to gray, or brown to black; stipes have a width 8–30 µm and length up to 1.5 mm; conidial heads are predominantly biseriate, but aberrant uniseriate heads, vesicles are 15–90 µm; metulae 5.0–26.1 µm; phialides are 5.4–12.5 µm; conidia are 3.0–4.5 µm, smooth, finely rough, or rarely rough.

32The morphological characteristics are similar with those produced by A. niger and A. tubingensis in Aspergillus section Nigri, and it is difficult to differentiate the species based on only morphology. DNA sequences of β-tubulin and calmodulin are useful to differentiate the species from the others, and GenBank JX500062 for β-tubulin and JX500071 for calmodulin could be good molecular markers for the species. Chemically, the extrolites antafumicins and/or “luchuensin” sets A. luchuensis apart from all other black Aspergillus strains. Antafumicins were first reported from A. niger NH-401. We were not able to study this strain, but it is probably A. luchuensis.

33Inui, also described A. perniciosis as a contaminant found during awamori-koji fermentation. He distinguished this species by yellow green mycelium with larger conidial heads similar to A. niger. No type culture is available, but the description of this taxon resembles A. coreanus on the basis of yellow green mycelium. Our molecular analysis shows that A. coreanus is a synonym of A. luchuensis(Fig. 3). Therefore we consider A. perniciosis a further synonym of A. luchuensis.

34Most isolates of Aspergillus luchuensis specifically produce antafumicin and/or a partially characterized metabolite provisionally called “luchuensin”, in addition to asperazine (shared with A. tubingensis and A. vadensis), they often produce atromentin, funalenone, pyranonigrin A, and occasionally tensidol B [13, 29, 30, Table S2]. None of these extrolites are considered toxic to vertebrates. Furthermore none of these strains produced fumonisin B2, B4 or B6 or any ochratoxins. A total of 52 strains from tea, coffee, and other substrates in addition to industrial strains were analyzed chemically, and none of them produced fumonisins or ochratoxins [13, 32, see also Table S2).

35Aspergillus luchuensis is the representative fungus of industrial black- and white-koji mold in Japan, and it is also the dominant microorganism of Puerh tea in China. The species is also known from nuruk (KACC 41731, 46420), Korean traditional fermentation starter for makgeolli, and meju (KACC 46490), a starter for soybean paste and soy sauce. In the case of meju, out of 54 strains of black Aspergillus, 15 strains were A. luchuensis (unpublished data). These data suggest that A. luchuensis is common in the food fermentation environment and the species play an important role in industry of East Asia.

36The taxonomic position of A. awamori Nakazawa proved to be complex. This is mainly caused by the fact that no type of A. awamori exists. CBS 557.65 ( = NRRL 4948), which does not originate from awamori koji and received from Instituto Ozwaldo Cruz, Brazil, was designated as neotype of A. niger var. awamori by Al-Musallam. The selection of this neotype was based on the fact that Raper and Fennell used this isolate for their description of A. awamori. In this study, the neotype of A. awamori (CBS 557.65) is located near to A. niger (CBS 554.65T), but distant from Inui’s A. luchuensis (NBRC 4281) (Fig. 3).

37Based on the neotype (CBS 557.65), Perrone et al. regarded A. awamori as a cryptic phylogenetic species in section Nigri. In our opinion, the taxonomic position of A. awamori is not clear, because no ex-type culture is known while the majority of strains found in culture collections under this name belong to A. niger (n = 24), A. luchuensis (n = 10), A. tubingensis (n = 2) or A. awamori sensu Perrone et al. (n = 2) (Table 2). According to Yamada et al., 14 of 20 strains of A. awamori were A. niger and six strains were A. luchuensis. Out of six of Nakazawa’s strains of A. awamori, three strains were the A. niger and the other three were A. luchuensis. This indicates that Nakazawa’s species concept is not clear and that he dealt with both A. luchuensis and A. niger. A. awamori as described by Nakazawa can be best treated as a doubtful synonym of A. niger or A. luchuensis.

38The strains named as A. awamori by Perrone et al. based on the incorrect selected neotype (CBS 557.65NT) are identical with CBS 139.54, a strain deposited at CBS from Welwitschia in Namibia. Figure S2 shows two phylogenetic trees produced from the combined sequence data of two loci (CaM, benA) of 30 taxa belonging to A. niger “aggregate” group. The epitype of A. welwitschiae is accommodated in the group of A. awamori sensu Perrone et al..

39Isolates of A. welwitschiae have been reported on various substrates as mycotoxin producers or plant pathogens including grapes, onions as causative agents of black mold rot and fumonisin contamination and Welwitschia mirabilis seeds causing seedling rot (van Diepeningen et al. unpublished data). In addition, this species was also identified in various otomycosis cases both in Iran and Hungary,. Although the name A. awamori was used taxonomically correctly by Perrone et al., this currently used name is misleading, as this species is rarely identified in awamori fermentation processes, where mostly A. niger and A. luchuensis could be isolated. It should also be mentioned that isolates of A. awamori sensu Perrone are able to produce ochratoxins and/or fumonisins;, making the application of the name of an oriental fermentation process as the basis of the name of this fungus is inadequate.

40Aspergillus welwitschiae (Bresadola) Hennings apud Wehmer.

41= Ustilago welwitschiae Bres. in Saccardo, Bolm Soc. Broteriana 11∶9–90. 1893 = Sterigmatomyces welwitschiae (Bres.) Henn. in H. Baum Kunene-Zambesi Expedit., p. 168, 1903 = Aspergillus welwitschiae (Bres.) Henn. apud Wehmer in Centrbl. Bakteriol. ParasitK. II, 18∶294, 1907

42= Aspergillus awamori sensu Perrone et al..

43Epitype: CBS 139.54, isolated by H.J. Swart from female inflorescence of Welwitschia mirabilis collected in Namibia, stored for 2 years (sample no. 236).

44Colonies on CYA, MEA, YES grow similar as A. niger. Vesicles 45–85 µm in diameter. Conidia globose, finely to distinctly roughened, brown to dark brown, 3.5–5.5 µm.

45Isolates of Aspergillus niger and A. welwitschiae have overlapping features concerning conidium size, ornamentation, stipe ornamentation, stipe length, and conidium colour. Both species contain mutant isolates that have more brownish conidium colours. A. welwitschiae produced conidia which were mostly globose and finely to distinct roughened. The conidial dimensions vary from 3.5–5.5 µm. Vesicles were 45–85 µm in diameter. These morphological characters are identical with the morphological structures in typical A. niger strains, A. welwitschiae has the same ranges of growth rates of A. niger on the media CYA at 5, 25 and 37°C, on G25N, CZA, MEA at 25°C in the dark, and also at reduced water activity on CZ20, M40Y at 25°C. It had also a strong acid production on CREA substrate.

46Notes: The name Aspergillus welwitschiae was first mentioned by Wehmer who reported that this name was proposed by P. Hennings in a written communication. This species has been reported from Welwitschia mirabilis by several authors and is mostly called A. niger or A. niger var. phoenicis –. It frequently infects and destroys germinating seeds which has an impact on a Welwitschia colony which can sometimes go many years without successfully reproducing.

47Aspergillus welwitschiae is morphologically indistinguishable from A. niger but some fixed nucleotide differences between these two species could be useful for their identification. Sequence positions and differences of A. welwitschiae compared to A. niger in the four regions sequenced are: calmodulin: 442 (T), 465 (C), 486–87 (CT), 493–494 (TT), 518 801 (T); β-tubulin: 173339 (T), 173346 (A); - translation elongation factor-1∶665 (G), 669 (A); RNA polymerase II subunit RPB2∶1281 (C), 1449 (C), 1689 (A), 1692 (C), 1719 (C), 1770 (T), 1947 (C), 2103 (T), 2220 (T).

48Besides its specific occurrence on Welwitschia plants strains of A. welwitschiae have been found on grapes, dried fruits, coffee, cocoa, and other sources and it has also a worldwide distribution from all the continents. In particular A. welwitschiae resulted to grow very poorly or no growth in the majority of the strains tested on to 2-deoxy-D-glucose substrate, respect to A. niger which grows well on this substrate.

49Most isolates of A. welwitschiae produce large amounts of fumonisins and occasionally ochratoxins. The best producer of fumonisin B2, 4 & 6 found yet is NRRL 567, a strain identified by us as A. welwitschiae and used for citric acid production.