Pairwise comparisons of the ITS and LSU sequences belonging to the Japanese, Hawaiian and Texan specimens of E. melanoxanthus var. melanoxanthus showed they are nearly identical except for one C-T transition at position 421 of the ITS alignment in strains JCM 5156, JCM 5159, JCM 5161, JCM 7948 and NBRC 31299, and one G-A transition at position 396 of the LSU alignment in strain JCM 5159. Similarly, strains MFLUCC 15-0723A, B and C belonging to E. metroxyli collected in Thailand differed only by one T-C transition at position 49 of the ITS alignment whereas their LSU sequences were identical to those of the studied E. melanoxanthus var. melanoxanthus strains. In contrast, the strain Endocalyx sp. MAFF 244025 showed a total of 15 bp and one gap differences compared with the remaining ITS sequences. In the case of E. melanoxanthus var. grossus, all ITS sequences were identical except for the LSU of JCM 5167 which contained a C-T transition at position 430 of the alignment. The ITS and LSU sequences of E. cinctus strains each showed three nucleotide differences between them whereas they were identical between E. indumentum strains.
The combined ITS-LSU alignment consisted of 102 sequences and 1,683 positions including the outgroups, 897 from the ITS alignment and 786 from the LSU. The single best RAxML tree (ln = −17725.715920) resulted from the ML analysis is shown in Fig. 1 with E. melanoxanthus var. melanoxanthus, E. melanoxanthus var. grossus and E. metroxyli as E. melanoxanthus, “E. grossus” and E. melanoxanthus, respectively. The ML tree topology was similar in topology to the 50% majority rule consensus tree of the 6,688 sampled trees from the Bayesian analysis. Effective sample size values of all relevant parameters were >200 as verified in Tracer, indicating adequate sampling of the posterior distribution (Drummond et al. 2006; Drummond and Rambaut 2009). The twenty-six Endocalyx strains formed a strongly supported monophyletic group (97% BS, 1.0 BPP) within the family Cainiaceae. Each species and variety were well resolved, and they split in two subclades: one including E. melanoxanthus var. melanoxanthus, E. indumentum and E. cinctus and the other containing only E. melanoxanthus var. grossus. The Japanese, Hawaiian and Texan strains of E. melanoxanthus var. melanoxanthus formed a strongly supported monophyletic group (100% BS, 1.0 BPP). The three strains belonging to E. metroxyli from Thailand (MFLUCC 15-0723A, B and C, obtained from the same single specimen) and that one of Endocalyx sp. MAFF 244025 from Japan also clustered within this clade indicating they are conspecific with E. melanoxanthus var. melanoxanthus. They grouped sister to the two available strains of E. indumentum including an ex-holotype (JCM 5171) without significant BS support but showing significant BPP = 0.98. Strains of E. indumentum, E. cinctus and E. melanoxanthus var. grossus including an ex-holotype (JCM 5164) each formed highly supported clades (99% or 100% BS, 1.0 BPP). The four taxa of Endocalyx shown in Fig. 1 were clearly recognized at species level and not at variety level. Therefore, E. melanoxanthus var. melanoxanthus and E. melanoxanthus var. grossus should be treated as E. melanoxanthus and “E. grossus” and a new taxonomic treatment is proposed below for the latter. The Endocalyx lineage was sister to a moderately supported monophyletic group including Cainia graminis (Niessl) Arx & E. Müll., the type species of Cainia Arx & E. Müll., and other members of Cainiaceae (Xylariales) belonging to three different genera. The family was recovered as a highly supported monophyletic clade (99% BS, 1.0 BPP). Arthrinium and Nigrospora species grouped together in a highly supported clade (98% BS, 1.0 BPP) representing the family Apiosporaceae in the distant order Amphisphaeriales (Xylariales after MycoBank and NCBI Taxonomy).
Endocalyx cinctus Petch, Ann. Bot. (London) 22: 394 (1908). (Figs. 3F, G, 4A)
Materials examined: Japan, Tokyo, Ogasawara Islands, Anijima Island, on peduncle of dead inflorescence of Livistona chinensis var. boninensis, 18 Mar. 1990, leg. & det. G. Okada [TNS-F-91424 = G. Okada Ogasawara-191; strain JCM 7946; TNS-F-91425 (dried culture of JCM 7946 incubated on autoclaved petiole of L. chinensis var. subglobosa (Hassk.) Becc.)]; idem [ILLS 121502 = G. Okada Ogasawara-192 (together with E. melanoxanthus)].
Notes: Morphologically, the specimen studied (TNS-F-91424) agrees well with the description provided by Okada and Tubaki (1984). In culture, strain JCM 7946 produced conidiomata (Fig. 3G) and conidia (Fig. 4A) on autoclaved petioles of L. chinensis var. subglobosa after three month-incubation period that were almost the same as those on palm hosts (Fig. 3F). The conidial surface was slightly rough especially in young conidia (Fig. 4A). Endocalyx cinctus is the best species for inducing conidiomata and conidia on autoclaved petiole of palms (Okada and Tubaki 1984; this paper). After further prolonged incubation on autoclaved petiole of L. chinensis var. subglobosa, irregularly shaped conidiomata including subspherical ones were produced (TNS-F-91425).
Endocalyx grossus (G. Okada & Tubaki) G. Delgado & G. Okada, comb. et stat. nov. (Figs. 3B–D, 4G–I, 5A–E)
Basionym: Endocalyx melanoxanthus var. grossus G. Okada & Tubaki, Mycologia 76: 303 (1984).
Typification: Japan, Ibaraki, Tsukuba, Higashioka, 36°05'46.6"N 140°07'34.2"E, on dead petiole of Trachycarpus fortunei, 6 Aug. 1982, leg. G. Okada, TNS-F-18281 (TKBF 5027, holotype); ex-holotype strain NBRC 31308 = JCM 5164 = CBS 105.86 = G. Okada OFC 1116).
Materials examined: Japan, Saitama, Ogawa-machi, Otsuka, 36°03'28.4" N, 139°15'03.3" E, on dead petiole of T. fortunei, 2 Sep. 2017, leg. & det. G. Okada (TNS-F-91426 = G. Okada 1763; strain JCM 32411); idem (ILLS 121505 = G. Okada 1764); idem, Koshigoe, near Tatekawa Dam, 36°01'13.8" N, 139°12'40.0" E, on dead petiole of T. fortunei, 23 Jul. 2018, leg. & det. G. Okada [TNS-F-91427 = G. Okada 1766; strain JCM 32997; TNS-F-91428 (dried culture of JCM 32997 incubated on autoclaved petiole of T. fortunei)]; idem (ILLS 121504 = G. Okada 1767); Ibaraki, Bando, Oguchi, 36°02'26.5" N, 139°56'09.6" E, on dead petiole of T. fortunei, 23 Jul. 2018, leg. & det. G. Okada [TNS-F-91429 = G. Okada 1769; strain JCM 32998; TNS-F-91430 (dried culture of JCM 32998 incubated on autoclaved petiole of T. fortunei)]; idem (ILLS 121506 = G. Okada 1770); Saitama, Chichibu, Hinoda-machi, Business office of the University of Tokyo Chichibu Forest, 35°59'19.1" N, 139°04'48.3" E, on dead petiole of T. fortunei, 13 Sep. 2018, leg. & det. G. Okada (TNS-F-91431 = G. Okada 1772; strain JCM 33339).
Notes: Okada and Tubaki (1984) morphologically distinguished E. melanoxanthus var. grossus from E. melanoxanthus var. melanoxanthus based on the brown to yellowish brown colour of the peridial hyphae enclosing the conidial mass, which is vivid yellow to greenish yellow in the latter, and the verrucose conidial wall ornamentation. This is relatively smooth to tuberculate under light microscopy but scanning electron microscopy (SEM) clearly revealed a thin outer layer that cracks and forms coarse-relief islets evenly distributed over the conidial surface. Moreover, the colour of conidial mass in E. melanoxanthus var. melanoxanthus is somewhat glistening blackish than the variety grossus. Differences in conidiomata were less evident, except the colour of the peridial hyphae, but the variety grossus forms mostly cupulate, rarely cylindrical and usually smaller conidiomata, 0.2–0.3 mm and up to 3 mm high, while those of the variety melanoxanthus are 0.5–1 mm and can reach up to 7 mm high. Ecologically, this taxon is apparently restricted and known so far only from T. fortunei, the windmill palm, one of the best-known cold-hardiest species of Arecaceae. Conidia of variety grossus developed readily in culture compared to the variety melanoxanthus (Okada and Tubaki 1984). Clavate chlamydospore-like conidia were also formed in culture, but much obvious in the variety melanoxanthus, and growth was observed at 5 °C after three months whereas other Endocalyx taxa did not grow under these conditions. In addition to these differences, the eleven strains belonging to the variety grossus, including an ex-holotype strain, grouped together with high support in our concatenated analyses of ITS and LSU sequence data distant from those of the variety melanoxanthus. Considering all the morphological, ecological, cultural and molecular evidence based on multiple specimens and cultures (Seifert and Rossman 2010), the variety grossus is elevated here to species rank distinct from E. melanoxanthus and the remaining Endocalyx species.
In culture, strain JCM 32998 produced conidial columns (Fig. 4G arrows) on autoclaved petiole of T. fortunei. They consisted of almost rough conidia (Fig. 4I) and emerged from an annulus-like structure (Fig. 4G white arrowheads) associated with hemispherical to subspherical conidiomata (Fig. 4G black arrowheads). In the case of strain JCM 5164 (ex-holotype) and JCM 32997, many smaller conidiomata were produced on the autoclaved substrate under prolonged incubation. Moreover, it was observed that under the same conditions JCM 32997 developed a few conidial columns with brown peridial hyphae from an annulus-like structure (Fig. 4H). This conidiomatal structure is basically the same as that on the palm host. On the other hand, strain JCM 5166 abundantly developed similar subspherical conidiomata of various sizes on PDA under prolonged incubation and without any autoclaved substrate (Fig. 5A). Dark brown conidial mass appeared from broken conidiomata (Fig. 5B, C arrowheads), and conidial columns (Fig. 5D, E arrows) were produced from the lower part of broken conidiomata (Fig. 5D, E arrowheads), which were very similar to the annulus-like structure formed on autoclaved petiole of T. fortunei (Fig. 4G white arrowheads).
Endocalyx indumentum G. Okada & Tubaki, Mycologia 76: 305 (1984). (Figs. 3E, 4B–F, 5F)
Materials examined: Japan, Tokyo, Ogasawara Islands, Chichijima Island, Suzaki, on dead petiole of Phoenix canariensis H.Wildpret, 12 Mar. 1990, leg. & det. G. Okada [ILLS 121501 = G. Okada Ogasawara-80 (together with E. melanoxanthus, no isolate)]; idem, Anijima Island, on peduncle of dead inflorescence of Livistona chinensis var. boninensis, 18 Mar. 1990, leg. & det. G. Okada (ILLS 121503 = G. Okada Ogasawara-194); idem [TNS-F-91432 = G. Okada Ogasawara-200; strain JCM 8042; TNS-F-91433 (dried culture of JCM 8042 incubated on autoclaved petiole of L. chinensis var. subglobosa)].
Notes: Morphological details of the specimen studied (TNS-F-91432) agree well with the description of Okada and Tubaki (1984). In culture, strain JCM 8042 formed primordium-like structures of conidiomata on autoclaved petiole of L. chinensis var. subglobosa after three month-incubation period. Hemispherical to subspherical conidiomata were produced on the surface of the substrate after further prolonged incubation, and dark brownish conidial masses in columns or not emerged from inside of burst conidiomata (Fig. 4C, D arrow & arrowheads; TNS-F-91433). Conidia with hair-like projections (Fig. 4E, F) were the same as those on palm hosts (Okada and Tubaki 1984, Fig. 23). The strain JCM 5171 (ex-holotype) did not produce conidiomata in the same conditions this time. On the other hand, strain JCM 8042 developed subspherical conidiomata on PDA under prolonged incubation and without any autoclaved substrate (Fig. 5F). In the broken conidiomata (Fig. 5F arrowheads), small immature conidia without filamentous ornamentation were observed as reported by Okada and Tubaki (1984, Fig. 24).
Endocalyx melanoxanthus (Berk. & Broome) Petch, Ann. Bot. (London) 22: 390 (1908). (Figs. 2, 3A)
≡ Melanconium melanoxanthum Berk. & Broome, J. Linn. Soc. Bot. 14: 89 (1875).
= Endocalyx metroxyli Konta & K.D. Hyde, in Konta et al., Life 11(486): 18 (2021).
Conidiomata scattered or aggregated in small to large groups and emerging from an annulus-like, black, circular pustules, at first short-cylindrical or short-cupulate becoming long cylindrical, subcylindrical, long conical or cup-shaped after incubation for several days, reaching up to 3 mm high in well-developed fructifications and consisting of a black, glistening mass of conidia enclosed by a yellow to greenish yellow or orange yellow, annular mass of sterile peridial hyphae which remain at the base once the conidioma expands and grows upwards surrounding the column of conidia. Conidiophores micronematous, filiform, flexuous, hyaline, septate, smooth, anastomosing, 1–2.5(–3.5) µm wide. Conidiogenous cells holoblastic, monoblastic, integrated, terminal or intercalary, cylindrical, minutely denticulate. Conidia solitary, dry, globose, subglobose or broadly ellipsoidal, slightly polygonal and flattened in front view, (10–)12–16 × 9–12(–13) µm, ellipsoidal, lenticular or rarely oblong in lateral view, (7–)8–9(–10) µm thick, with a paler equatorial germ slit, aseptate, brown, dark brown or blackish brown, thick-walled, smooth to finely roughened, often with a central or nearly central attachment scar.
Colonies on MEA moderately fast growing reaching 25–35 mm diam. after 10 days at 25°C, cottony, white, up to 5 mm high at the center, flat and dull-white toward the edge, margin entire, sometimes with visible yellow bands or yellowish spots; reverse dull white, yellowish at the center. Sporulation not observed after 3 months. Chlamydospores present, abundant in gray or blackish patches immersed under the superficial mycelium, terminal or intercalary, solitary or catenate and in short chains of up to 6, globose, subglobose, ellipsoidal, long ellipsoidal, subcylindrical, long cylindrical or elongated, rarely narrowly clavate or pyriform when terminal, straight or flexuous to curved, pale brown to brown or dark brown, thick-walled, smooth, 0(–1)-septate, at times slightly constricted around the center, sometimes with a paler equatorial germ slit, 8–24(–29) × 5–9(–11) µm.
Materials examined: USA, Texas, Harris County, Spring, Meyer Park, 30°00'15.9" N, 95°31'35.7" W, 33 m a.s.l., on peduncle of dead inflorescence of Sabal minor, 20 Sep. 2020, leg. & det. G. Delgado (ILLS 121433; strain CBS 147393); idem, on spathe of dead inflorescence of S. minor, 20 Sep. 2020 (ILLS 121434; strain CBS 147394).
Other materials examined: Japan, Tokyo, Ogasawara Islands, Chichijima Island, Suzaki, on dead petiole of Livistona chinensis var. boninensis, 12 Mar. 1990, leg. & det. G. Okada (TNS-F-91434 = G. Okada Ogasawara-61; strain JCM 7948); USA, Hawaii, Hawaii Island, Hilo, around Liliuokalani Park and Gardens, on dead petiole of Cocos nucifera L., 2 Aug. 2005, leg. & det. G. Okada (ILLS 121495; strain JCM 13432); ILLS 121502 and ILLS 121501 (see the above sections of E. cinctus and E. indumentum for collection details).
Notes: Our specimens agree well with previous descriptions of the fungus in having distinct annular, vivid or greenish yellow fructifications surrounding a black mass of subglobose, more or less angular, dark brown to blackish brown, aseptate conidia with a paler germ slit. Conidiomata readily developed after incubation for a few days and the conidial mass together with the yellow peridial hyphae expand upward forming long cylinders up to 3 mm high in the longer fructifications. Okada and Tubaki (1984) also obtained morphologically similar, well-developed conidiomata in moist chamber that reached up to 7 mm high. The moisture conditions induced abundant sporulation that cannot be held by the outer layer of peridial hyphae and tears laterally in several places or apically, opening up and releasing spores on the substrate (Fig. 2D–F). Wall ornamentation of conidia was confirmed to be very finely roughened and more visible around the paler wall of the germ slits in agreement with Okada and Tubaki (1984), who reported a fine dust-like layer covering conidia which cracks and creates islets as seen under SEM. In culture, the Texas strains did not sporulate after three months incubation on MEA, but they produced abundant chlamydospores in the same time period (Fig. 2J–L). This also agrees with Okada and Tubaki (1984) who reported solitary or catenate, terminal or intercalary, dark brown, similar in size and shape chlamydospores, with a paler germ slit and superficially resembling conidia. They also obtained immersed or superficial, pycnidioid conidiomata on sterilized wooden chips that produced conidia similar to those on natural substrate, but this technique to induce sporulation was not attempted this time for the Japanese, Hawaiian and Texan strains.