3.1 Environmental conditions
During the curing process where tobacco pole rot occurs in the curing chamber, relative humidity declined from 84-91.5% to 8.5-22%, lamina wetness declined from 63-67% to 4-14%, and petiole wetness declined from 73-78% to 5-20% (Table 1). In contrast, air temperature increased from 37-38 to 67-68 oC, and lamina temperature increased from 35-36 to 65-66 oC. Wind speed was only higher during the color-fixing stage. This shows that tissues are drying with increasing temperatures over time. Position in the chamber was correlated with relative humidity, decreasing at progressively lower positions from 91.5 to 84%, 73 to 57%, and 22 to 8.5% in the yellowing, color-fixing and stem-drying curing stages, respectively. In contrast, air temperature increased at progressively lower positions from 37 to 38 oC, 47 to 48 oC, and 67 to 68 oC in the yellowing, color-fixing and stem-drying curing stages, respectively.
3.2 Culture-based fungal diversity and abundance
A total of 108 fungal isolates belonging to 5 genera and 10 species were obtained by tissue isolation method (Table 2). More species were isolated on PDA (R. oryzae, E. thailandicum, D. ovalispora, A. oryzae, A. fumigatus, A. flavus, A. aflatoxiformans, A. alternata) than on AEA (R. oryzae, E. sorghinum, C. tenuissimum, A. fumigatus, A. flavus). The most abundant species was R. oryzae comprising 33 of the 54 isolates on PDA and 35 of the 54 isolates on AEA, followed by A. flavus with 10 of the isolates on PDA and 9 of the isolates on AEA and A. fumigatus with 6 isolates on PDA and 6 on AEA. The rarest species were E. thailandicum, D. ovalispora, A. oryzae, A. aflatoxiformans, A. alternata, E. sorghinum and C. tenuissimum.
When analyzed for curing stage combining lamina and petiole samples, diversity was highest during the color-fixing curing stage with 10 species, and lowest at the stem-drying curing stage with 2 species (Table 2). The relative abundance (% of isolates of a species out of the total number of isolates) during the yellowing curing stage showed that R. oryzae was most abundant (12.04% on AEA and 13.89% on PDA), during the color-fixing stage showed that A. flavus (5.56% on AEA and 7.41% on PDA) and A. fumigatus (3.70% on AEA and 3.70% on PDA) were most abundant, and during the stem drying stage showed that R. oryzae (16.67% on AEA and 15.74% on PDA) was most abundant. Isolates of E. sorghinum (1.85% on AEA and 0% on PDA), C. tenuissimum (1.85% on AEA and 0% on PDA), E. thailandicum (0% on AEA and 0.93% on PDA), D. ovalispora (0.93% on AEA and 0% on PDA) and A. oryzae (0% on AEA and 0.93% on PDA) were only found during the color-fixing curing stage.
When analyzed for tissue type combining curing stages, diversity was highest for lamina with 9 species compared to petioles with 5 species (Table 2). The relative abundance showed that A. flavus (4.63% on AEA and 5.56% on PDA) and A. fumigatus (3.70% on AEA and 2.78% on PDA) were higher in petiole, whereas R. oryzae was highest in both petiole (14.81% on AEA and 15.74% on PDA) and lamina (17.59% on AEA and 14.81% on PDA).
3.3 Sequence-based fungal diversity
There was a total of 4,373,082 high-quality sequences across the 27 petiole and 27 lamina samples. A total of 2,238 OTUs at ≥ 97% nt identity were obtained from the 54 samples after the removal of low quality, chimeric and rare sequences resulting in an average number of sequences per sample of 80,983. When the number of sequences reached approximately 40,000, the rarefaction curves for all 54 samples revealed that they approached the plateau phase (Fig.1), suggesting that there was sufficient sequence coverage to describe the fungal composition.
Comparing between curing stages showed that the total number of OTUs progressively increased with significantly more OTUs at the stem-drying stage than the yellowing stage (Table 3). For petioles, the number OTUs increased between curing stages, with significant differences between the yellowing stage and the stem-drying stage. For leaf laminas, the number OTUs increased between curing stages, with significant differences between the yellowing stage and the stem-drying stage.
Comparing between positions showed that during the yellowing stage, the number of OTUs for petioles were not significantly different, but the number of OTUs for leaf laminas were significantly greater in the upper position than the lower position (Table 3). During the color-fixing stage, there were significantly higher numbers of OTUs at the middle compared to the upper position for petioles, but no significant differences for leaf lamina. During the stem-drying stage, no significant differences were found for petioles based on position, but there were significantly lower OTU numbers in the upper and middle positions compared to the lower position for leaf lamina (Table 3).
OTUs on petioles at the stem drying stage and on lamina at the stem drying stage, regardless of position, had highest diversity based on the average Shannon, Simpson, Chao1 and ACE values (Table 3). The next highest diversity of the OTUs was at the color-fixing stage for both petiole and lamina samples, regardless of position based on the average Shannon, Simpson, Chao1 and ACE values. The lowest diversity index values were found for OTUs at the yellowing stage for both petiole and lamina samples, regardless of position based on all four diversity indices.
3.4 Taxonomic composition of the OTUs
The distribution of phyla for the OTUs showed that 88.29% of the clean sequence reads could be classified in the Ascomycota, Basidiomycota, Mucoromycota, Glomeromycota, Mortierellomycota, Blastocladiomycota, Olpidiomycota and Chytridiomycota. Members of the other phyla were unassigned, and were likely not true fungi (Fig 2). The fungal communities were dominated by the Ascomycota (47.36%), followed by the Basidiomycota (5.88%) and the Mucoromycota (0.60%). Combined together, the Glomeromycota, Mortierellomycota, Blastocladiomycota, Olpidiomycota and Chytridiomycota comprised only 0.01% of the reads.
Combining position and tissue type, the average number of reads per curing stage for the Ascomycota was 56.10%, 47.17%, 38.80% in the yellowing stage, color-fixing stage and stem-drying stage, respectively, indicating a decline as curing occurred. In contrast, the relative abundance for all samples for the Basidomycota was 4.12%, 7.11%, 6.41% in the yellowing stage, color-fixing stage and stem-drying stage, respectively, indicating a peak in the color-fixing stage. The relative abundance for all samples for the Mucoromycota was 1.01%, 0.15%, 0.65% in the yellowing stage, color-fixing stage and stem-drying stage, respectively, indicating a decline during curing. Combining position and curing stage, the relative abundance of samples based on tissue type showed that Ascomycota were 46.68% and 48.03%, Basidiomycota were 5.57% and 6.18%, Mucoromycota were 0.96% and 0.24% for petioles and lamina, respectively, indicating that there was no significant difference based on tissue type.
A total of 35.97% of the OTUs could be classified at the genus level. The 30 most common genera are shown in Fig 3. Among those, the 10 highest number of reads were for Alternaria, Phoma, Cercospora, Aspergillus, Cladosporium, Symmetrospora, Boeremia, Stagonosporopsis, Epicoccum and Hannaella. However, Rhizopus, which would include the pole rot pathogen, R. oryzae, had only 0.60% of the reads. Thus, the reads were dominated by OTUs for saprophytic, rather than pathogenic fungi.
A maximum likelihood tree of the 100 most abundant fungal genera showed that the most dominant fungi were in the Ascomycota, followed by Basidiomycota, and the least common fungi were in the Mucoromycota (Fig 4). For the Ascomycota, the dominant genera were Alternaria, Aspergillus, Cerospora, Cladosporium, Phoma, Boeremia, Golovinomyces, Stagonosporopsis and Epicoccum. For the Basidiomycota, the dominant genera were Symmetrospora, Hannaella, Golubevia and Rhodotorula. For the Mucoromycota, the dominant genera was Rhizopus, which would include R. oryzae.
The abundance of the top 10 genera varied considerably among the samples (Table 4). Alternaria was the highest at B21 followed by A31 and A21, indicating that it was more common in the upper position and somewhat more common in petiole than the lamina. Phoma was the highest at B13 followed by A12 and B11, indicating that it was also favored during the leaf yellowing stage. Boeremia was most common in B1, followed by B23 and B13, indicating that it has most abundant for lamina, curing stage or position. Cercospora was most abundant B31 followed by B22 and A22, indicating that it was more common in the upper position and somewhat more common in petiole than the lamina. Aspergillus was most common in A32, B32 and A33, indicating that the stem-drying stage at the middle and lower positions favored it. Rhizopus was the most dominant genus in A11, B32 and A31, indicating it was most common in petioles during the yellowing stage.
3.5 The relationship to environmental parameters
Spearman correlation analysis of the most abundant genera was made between air temperature, relative humidity, curing stage, position, wind speed and wetness of leaf and petiole (Fig. 5). Air temperature significantly affected the abundance of Golovinomyces. Relative humidity significantly affected the abundance of Alternaria, Phoma, Trichoderma, Leptosphaerulina, Gibellulopsis, and Candida. Curing stage significantly affected the abundance of Golovinomyces, Golubevia, Strelitziana, Dioszegia and Pestalotiopsis. Sample position significantly affected the abundance of Alternaria, Aspergillus, Rhizopus and Leptosphaerulina. Wind speed significantly affected the abundance of Golovinomyces, Golubevia, Septoriella, Strelitziana, Dioszegia and Pestalotiopsis. Wetness of leaf and petiole significantly affected the abundance of Alternaria, Stagonosporopsis, Trichoderma, Leptosphaerulina, Gibellulopsis and Candida. In general, Alternaria, Phoma, Golovinomyces, Strelitziana, Leptosphaerulina and Pestalotiopsis were the genera affected by the most environmental factors. Rhizopus, which would include the pole rot pathogen R. oryzae was only significantly affected by position.
3.6 Spatial distribution of microbial communities
PCA showed that the first two PCs accounted for 4.4% and 8.12% of the total variance in the fungal communities of the 18 sample groups (Fig. 6). All of the fungal communities overlapped with each other, except for three distinctive fungal communities, which were for the upper petiole samples at the yellowing stage (B11), the lower lamina samples at stem-drying stage (B33) and the lower petiole samples (A33).
3.7 Functional guilds analysis
FUNGuild database was used to classify the fungi in present study by ecological guild (Fig. 7). Members of the pathotroph-saprotroph-symbiotroph were the most common at 13.68% of sequences, pathotroph was the second most common at 13.37% of sequences, pathotroph-saprotroph was the third most common at 10.56% of sequences, saprotroph was the fourth most common at 6.81% of sequences, pathotroph-symbiotroph was the fifth most common at 4.22% of sequences, and the least common were symbiotroph, saprotroph-symbiotroph and pathogen-saprotroph-symbiotroph at 0.20, 0.10 and <0.10% of the sequences, respectively. However, the unassigned sequences was the largest group at 51.77%. For the average based on sample type, pathotroph-saprotroph-symbiotroph was the most abundant in petioles and lamina. For the average based on curing stage, pathotroph was most abundant in leaf yellowing, color fixing and stem drying stages, respectively.