Taxonomy and Phylogeny of Four New Species in Absidia (Cunninghamellaceae, Mucorales) From China

Four new species within the genus Absidia, A. globospora, A. medulla, A. turgida, and A. zonata, are proposed based on a combination of morphological traits, physiological features, and molecular evidences. A. globospora is characterized by globose sporangiospores, a 1.0- to 3.5-μm-long papillary projection on columellae, and sympodial sporangiophores. A. medulla is characterized by cylindrical to oval sporangiospores, a 1.0- to 4.5-μm-long bacilliform projection on columellae, and spine-like rhizoids. A. turgida is characterized by variable sporangiospores, up to 9.5-μm-long clavate projections on columellae, and swollen top of the projection and inflated hyphae. A. zonata is characterized by cylindrical to oval sporangiospores, a 2.0- to 3.5-μm-long spinous projection on columellae, and as many as eight whorled sporangiophores. Phylogenetic analyses based on sequences of internal transcribed spacer rDNA and D1–D2 domains of LSU rDNA support the novelty of these four species within the Absidia. All new species are illustrated, and an identification key to all the known species of Absidia in China is included.

The classification and circumscription of the Absidia have been debated since it was described. According to zygospore morphology, Hesseltine and Ellis (1964) divided Absidia into two Absidia sensu stricto, and consequently, a revised synoptic key to all the 13 known species of Absidia in China is provided.

Isolation and Strains
Strains were isolated from the soil collected in Hubei province, Shanxi province, Xinjiang province, and Yunnan province, China. Soil samples (1 g) were suspended in 100 mL sterilized water and shaken vigorously. Then, a 100 µL of the suspension was added onto a potato dextrose agar (PDA; Benny, 2008) plate with antibiotics streptomycin sulfate (100 mg/mL) and ampicillin (100 mg/mL). The plate was incubated at 27 • C and examined daily with a stereo microscope (SMZ1500, Nikon Corporation, Japan). Upon the presence of colonies, a single colony was picked and transferred to new PDA plates. Living cultures were deposited in the China General Microbiological Culture Collection Center, Beijing, China (CGMCC). Dried cultures were deposited in the Herbarium Mycologicum Academiae Sinicae, Beijing, China (HMAS).

Morphology and Growth Experiments
Pure cultures were established in triplicate, respectively, with malt extract agar (MEA; Benny, 2008), modified synthetic mucor agar (SMA; Zheng and Chen, 2001), and PDA plates. For morphological observation, they were incubated at 27 • C for 4-7 days and examined daily under a microscope (Axio Imager A2, Carl Zeiss Microscopy, Germany). For determining maximum growth temperatures, pure cultures were initially incubated at 32 • C for 4 days, and then the incubation temperature was adjusted until the colonies stopped growing. The color of colonies was designated according to Ridgway (1912).

DNA Extraction, Polymerase Chain Reaction Amplification, and Sequencing
Mycelia were grown at 27 • C for 5 days on PDA plates, and then cell DNAs were extracted using a kit (GO-GPLF-400, GeneOnBio Corporation, Changchun, China). The ITS and D1-D2 domain of LSU rDNA were amplified with primer pairs NS5M and LR5M (Wang et al., 2014). The polymerase chain reaction (PCR) procedure was as follows: an initial temperature at 95 • C for 5 min; then 30 cycles of denaturation at 95 • C for 20 s, annealing at 55 • C for 60 s, and extension at 72 • C for 60 s; and finally an extra extension at 72 • C for 10 min. PCR products were purified and then sequenced with primers ITS5 (White et al., 1990) and LR5M at BGI Tech Solutions Beijing Liuhe Co., Limited, Beijing, China. All newly generated sequences were deposited in GenBank and National Microbiology Data Center (NMDC, 2 Table 1).

Phylogenetic Analyses
The software platform Geneious 8.1 3 was used to assemble and proofread DNA sequences. All the sequences were realigned using AliView version 3.0 (Larsson, 2014). The sequence alignments and phylogenetic trees were deposited at TreeBase (submission ID 27734). Sequences of Cunninghamella elegans and Cunninghamella blakesleeana retrieved from GenBank were used as outgroups in the ITS and LSU analyses following Hoffmann et al. (2007). Phylogenetic analyses were carried out using maximum parsimony (MP), maximum likelihood (ML), and Bayesian inference (BI). MP phylogenetic analyses followed Zhao and Wu (2017), and the tree construction was performed in PAUP * version 4.0b10 (Swofford, 2002). All characters were equally weighted, and gaps were treated as missing data. Trees were inferred using the heuristic search option with TBR branch swapping and 1,000 random sequence additions. Max-trees were set to 5,000; branches of zero length were collapsed, and all parsimonious trees were saved. Clade robustness was assessed using a bootstrap analysis with 1,000 replicates (Felsenstein, 1985). Descriptive tree statistics tree length (TL), consistency index (CI), retention index (RI), rescaled CI (RC), and homoplasy index (HI) were calculated for each maximum parsimonious tree generated.
Maximum likelihood phylogenetic analyses were conducted with raxmlGUI 2.0 beta (Edler et al., 2020). A general time reversible model was used with a gamma-distributed rate variation (GTR + G) and 1,000 bootstrap replicates.
Bayesian inference phylogenetic analyses was calculated with MrBayes 3.2.7a by a general time-reversible model with an estimate of the proportion of invariant sites and a gamma distribution for variable rates across sites (GTR + I + G; Ronquist et al., 2012). Four Markov chains were run simultaneously from random starting trees, for 2,400,000 generations (ITS) or 300,000 generations (LSU). Trees were sampled every 100 generations. The chains stopped once the average standard deviation of split frequencies decreased lower than 0.01. The first one-fourth generations were discarded as burn-in. A majority rule consensus tree of all remaining trees was calculated. Branches were considered as significantly supported if they received ML bootstrap > 75%, MP bootstrap > 75%, or Bayesian posterior probabilities > 0.95.

Phylogenetic Analyses
The ITS dataset included sequences from 38 strains representing 33 species of Absidia and related genera. The dataset had an aligned length of 903 characters, of which 248 characters were constant, 136 were variable and parsimony-uninformative, and 519 were parsimonyinformative. MP analyses yielded two equally parsimonious trees (TL = 4163, CI = 0.3394, HI = 0.6606, RI = 0.3446, RC = 0.1170). At the end of the inference, the average standard deviation of split frequencies was 0.009990. All BI, ML, and MP phylogenetic trees resulted in similar topologies. The phylogram (Figure 1) consists of three clades, although with relatively low support values: (1) except the A. pararepens and A. bonitoensis, all members in the cylindrospora clade produce cylindric sporangiospores; (2) all members in the globospora clade produce globose sporangiospores; and (3) the Absidia cuneospora G.F. Orr & Plunkett separately groups as a cuneospora clade, forming conical sporangiospores (Orr and Plunkett, 1959).
The LSU dataset included sequences from 39 strains representing 34 species within Absidia. The dataset had an aligned length of 967 characters, of which 562 characters were constant, 117 were variable and parsimonyuninformative, and 288 were parsimony-informative. MP analyses yielded 20 equally parsimonious trees (TL = 1422, CI = 0.4339, HI = 0.5661, RI = 0.6443, RC = 0.2795). At the end of the inference, the average standard deviation of split frequencies was 0.009767. All BI, ML, and MP phylogenetic trees resulted in similar topologies. The phylogram (Figure 2) consists of three clades, similar to the ITS phylogram ( Figure 1) but with relatively high support values, in detail, clade cylindrospora, globospora, and cuneospora with a support of 80/-/0.99, 98/98/1.00, and 100/100/1.00, respectively.
Media and temperatures: Colonies on SMA, cottony, regularly zonate, attaining 70-mm diameter after 5 days at 27 • C, white at first and then pale olive-gray (R51). Colonies on PDA, regularly zonate, attaining 74-mm diameter after 5 days at 27 • C, white at first and then snuff brown (R29) to deep olive (R40) in center. No growth at 33 • C.  Etymology: turgida (Lat.) referring to the swollen hyphae and the inflate projection on columellae.
Media and temperatures: Colonies on SMA, rough around the edges, concentric ring-shaped, attaining 69-mm diameter after 8 days at 27 • C, white. Colonies on PDA, regularly wavy zonate, attaining 72-mm diameter after 7 days at 27 • C, white at first and then lime green (R31). No growth at 38 • C.

DISCUSSION
Phylogenetically, the ITS (Figure 1) and LSU (Figure 2) trees show that four new species cluster in different clades of Absidia. The A. globospora (100/100/1.00 for both ITS and LSU) is located in the globospora clade and most closely related to A. glauca Hagem (88/100/1.00 for ITS). Their sibling relationship is completely supported by ITS and LSU, with 99/100/1.00 and 100/100/1.00 support values, respectively. Physiologically, A. globospora is similar to A. glauca in heterothallism but differs in maximum growth temperature (37 vs. 29 • C). Morphologically, A. globospora is similar to A. glauca in forming green colonies and globose sporangiospores. However, A. glauca differs in its wider sporangiophores (up to 12 µm), glaucous stolons, and larger sporangia (mostly 50-to 60-µm diameter, Ellis and Hesseltine, 1965).
The other three new species are placed in the cylindrospora clade where A. zonata is most closely related to A. koreana (100/100/1.00 for ITS, 99/95/1.00 for LSU), which is strongly supported with a high value of 100/100/1.00 or 98/99/0.99. Both A. zonata and A. koreana are physiologically similar in maximum growth temperatures but morphologically differentiated by characteristics on SMA media ( Table 2; Ariyawansa et al., 2015). The width of sporangiospores in A. koreana are more narrow and uniform. In its protologue, A. koreana did not form projections, but the figure in the original article illustrated projections. A. turgida is basal to A. heterospora in ITS tree (Figure 1) or next to A. repens Tiegh. and A. pararepens in the LSU tree (Figure 2). A. medulla is closely related to A. repens in ITS tree (Figure 1) or A. saloaensis and A. ovalispora in LSU tree (Figure 2).
The two strains, CGMCC 3.16034 and CGMCC 3.16037, of A. medulla are similar in maximum growth temperature, micromorphology, and even colonies on MEA and PDA media, but slightly different in colonies on SMA media. The ex-paratype CGMCC 3.16037 is more floccose and thicker and grows more slowly than the ex-holotype CGMCC 3.16034 when they are incubated on SMA at 27 • C, and it lacks white concentric rings from the reverse side of the colony (Figure 7).

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ supplementary material.

AUTHOR CONTRIBUTIONS
T-KZ, C-LZ, and X-YL contributed to conception and design of the study. T-KZ wrote the draft of the manuscript. C-LZ and X-YL improved the manuscript. T-KZ, HZ, C-LZ, and X-YL observed and described the morphology. X-LL and L-YR collected the molecular data. All authors contributed to manuscript revision, proofread, and approved the submitted version.