On-site inspection and sampling
The ancient memorial archway in Beijing was established more than 200 years ago. The two large, round wooden pillars buried in the ground were numbered No. 1 and No. 2. They were nearly 40 years old and seriously decayed and replaced in a later stage (Fig. 1). The wood species of No. 1 and No. 2 pillar were Pseudotsuga menziesii and Picea sp., respectively, according to maintenance records. The in-ground part was covered in cement, and the pillar base was submerged in underground water. The in-ground wood was seriously decayed. The sampling was conducted on in-ground, decaying wood (HUP); contact-ground, decaying wood (HC); and above-ground wood without progressing decay (HD). Samples from No. 1 pillar were coded as HC1, HUP, and HD1, and those from No. 2 pillar were HC2 and HD2. The buried part of No. 2 pillar had been cut off during the second maintenance, and the in-ground foundation was connected to the cement pillar. Thus, only the contact-ground and above-ground samples were available. The moisture content was measured at the corresponding sampling location. At least three points were measured in each location, and the average was obtained. All data are shown in Table 1.
Table 1 shows that the moisture content at the bottom of the wood pillar was extremely high (average moisture content: 56.3%). The contact-ground wood moisture content was also notably high and exceeded 20%, but that above-ground was less than 14%. The average moisture content of the control ancient wood (dry and rotten wood) was notably low at 8.9%. Mold infections start with 14% wood moisture content, whereas wood decaying fungi need more than 20% to initiate decay [27]. The moisture content in contact-ground and in-ground wood is suitable for the growth of decaying fungi and occurrence of wood decay.
What caused the high moisture content of the in-ground and above-ground wooden pillars? During on-site observation, the bottom of the wood pillar was soaked in ground water, whereas the in-ground pillar was encased in cement. The water at the bottom of pillar originated partly from underground water or accumulated rain, whereas the cement coating prevented water evaporation.
Fungal isolation
By traditional tissue isolation and purification, one species of mold Phoma sp. was isolated from HC1, Trichoderma atroviride from HUP, and Entoloma from HD1. Meanwhile, two species of Ascomycetes Alternaria and Phaeosphaeriaceae were isolated from HC2 of No. 2 pillar. However, no fungus was isolated from HD2 (Fig.2). Entoloma is a decaying fungus, whereas the rest are molds.
Microbial diversity estimation on five samples
α-Index diversity analysis
Table 2 sums up the richness and diversity estimation of fungal ITS sequencing libraries from the MiSeq sequencing test. The total numbers of fungal ITS reads obtained with clean tags from sample HC1, HUP, and HD1 were 39962, 31442, and 30700, respectively, which were clustered into 115, 122, and 95 OTUs, with almost 0.97 similarity in nucleotide identity. The total numbers of fungal ITS from samples HC2 and HD2 reached 28637 and 41029, which were clustered into 134 and 182 OTUs, respectively.
Good’s coverage of six samples were close to 100%, with the mean obtained from the full sequence. The community diversity indicator Shannon index was 1.9–3.9, indicating the relatively simple fungal community composition of the sample. The abundance index Chao1 was 103.02–294.51, which was used to evaluate the community richness, and the value of phylogenetic diversity (PD_whole tree) was 21.89–69.39 (Table 2).
Fungal composition analysis
The community structure and relative abundance of the five samples in different taxonomic levels (phylum, class, order, family, and genus) are shown in a colored column chart (Fig. 3). Fig. 3 and Table 3 show that the dominant fungi were Ascomycota, followed by Basidiomycetes. Basidiomycetes accounted for the largest proportion (37.86%) in HC2. The second highest proportion was observed in HD2 at 27.02%. Similar to the control G, the proportion of Ascomycota in HD1 reached more than 90%, whereas that of Basidiomycetes approximated 5%. Compared with the microbial diversity structure of the discolored wood, the overall proportion of Basidiomycetes in rotten wood pillars increased but was less than 10% in the discolored wood [22].
Basidiomycetes are the main degradation fungi of wood decay. The proportion of Basidiomycetes is a possible indicator to predict whether ancient wood decay progresses. Further research is needed to prove this assumption.
Figure 4 shows the microbial structure composition of samples from three different sampling locations at the phylum and genus levels. The microbial structure composition of contact-ground wood (C) is similar to that of in-ground wood (UP), whereas that of above-ground wood (D) differs from the other two positions. The proportions of Ascomycota in the contact-ground and in-ground woods were about 74%, whereas the proportion of Basidiomycota reached 25.08% in contact-ground wood and 24.03% in in-ground wood. At the genus level, Lecythophora, Phoma and Cryptcoccus were the main fungi, and their proportions slightly differed at 25.25%, 34.15%, and 10.77%, respectively, in the contact-ground wood, and 50.8%, 9.16%, and 3.32% in the above-ground wood.
Fig. 5 shows the fungal structure composition of two pillars at the phylum and genus levels. The wood species of No. 1 pillar was Pseudotsuga menziesii (Mirbel) Franco, and that of No. 2 pillar was Picea sp (Table 1). Both were coniferous wood. No. 1 pillar comprised 83.9% Ascomycota and 15.23% Basidiomycota. For No. 2 pillar, Ascomycota accounted for 67.4%, and Basidiomycota amounted to 31.88%. Among the fungal species that can be identified, Lecythophora was dominant at the genus level. The proportions reached 24.44% and 12% in pillars 1 and 2, respectively. Scedosporium accounted for the second highest content in pillar 1 (21.26%). The third was Cryptcoccus (3.18%), which accounted for 7.46% in No. 2. Postia (6.21%) was also observed at a high proportion in No. 2 pillar.
Principal component analysis (PCA) and cluster tree
The relationship among the samples could be analyzed by PCA (Fig. 6) and cluster tree analysis (Fig. 7) based on the composition of each sample OTU. The closer the position in the PCA figure, a more clustering relationship occurs, and the more similar composition of samples. Fig. 7 shows that PC1 factor reached 53.78%, and PC2 totaled 28.82%. Three sequencing tests have shown excellent identity (Fig. 7). The cluster tree showed the close relationship between HC1 and HC2, whereas HC (HC1+HC2) and HUP exhibited a closer relationship. HD (HD1+HD2) was distantly located from HC and HUP, consistent with fungal composition analysis (Fig. 4).
Venn diagram and core microbiome analysis
Venn diagrams showed the common and exclusive fungal OTUs of the different wood samples (Fig. 8). A total of 23 OTUs were observed in all wood samples. Meanwhile, 75 OTUs were detected in three different locations, and 168 OTUs were identified in different pillars.
Core microbiome analysis of five samples showed 23 OTUs (Table 4). Two OTUs were identified at the phylum level, in which OUT_4 belonged to Ascomycota and OUT_8 to Basidiomycota. OTU_70 was identified under order Pleosporales. Five OTUs were identified at the family level. Fifteen OTUs were identified at the genus and species levels. The OTUs comprised Lecythophora sp. W3a2, Aspergillus cibarius, A. subversicolor, A. caesiellus, Rhodotorula mucilaginosa, Aureobasidium pullulans, Cryptococcus albidus, Epicoccum sp. NFW7, Cladosporium, Alternaria, Fusarium, Cladophialophora immunda, Podospora ellisiana, Ilyonectria macrodidyma, and Kernia pachypleura. Two species belonged to Basidiomycota (R. mucilaginosa and C. albidus), and the other thirteen OTUs all belonged to Ascomycota. Aspergillus spp., Fusarium sp., Aureobasidium pullulans, Trichoderma sp., Alternaria sp., and Cladosporium sp. were the common mold fungi reported previously [22, 28-29].
Dominant fungal species analyzed by MiSeq sequencing and isolation
Table 3 shows the dominant fungal species analyzed with MiSeq sequencing and isolation. The dominant fungi in five wood samples were Ascomycota and Basidiomycota at the phylum level. For HC1, the fungal composition was 84.40% Ascomycota and 15.92% Basidiomycota. At the genus level, the fungi composition of HC1 was 52.21% Phoma, 22.46% Lecythophora, and 5.51% of Cryptococcus. One fungus (Phoma) was isolated by traditional method. Thus, Phoma is the dominant genus in HC1.
In sample HC2, the proportions of Ascomycota and Basidiomycota were 60.26% and 37.86%, respectively. At the genus level, HC2 was inhibited by Lecythophora (29.15%), Cryptococcus (18.11%), Dacryopinax (11.79%), Phoma (8.96%), and Sporidiobolaceae (7.72%). Two fungi (Alternaria and Phaeosphaeriaceae) were isolated from HC2. MiSeq sequencing revealed Alternaria accounted for 2.13% of the fungal proportion, and Phaeosphaeriaceae belonged to an unidentified fungi. The traditional isolation failed to determine the dominant fungus, indicating the need for high-throughput sequencing. However, several OTUs were insufficient to identify certain genus or species. Thus, the development of molecular identification of fungi is needed.
The fungal composition of HUP comprised 74.07% Ascomycota and 24.23% Basidiomycota. The fungal structure at the genus level was composed of 50.8% Lecythophora, 20.66% Sporidiobolales, 9.16% Phoma, 6.97% Cladophialophora, and 3.32% Cryptococcus. Trichoderma accounted for 0.006%, which had been obtained by isolation method. The dominant fungi (50.8% Lecythophora) was not obtained.
A high proportion (91.96%) of Ascomycota was observed in HD1. At the genus level, three genus were dominant: Scedosporium (70.71%), Sporothrix (10.53%), and Entoloma (4.5%). The Basidiomycota fungus Entoloma was obtained by isolation method. The strains obtained by isolation method were not always dominant.
In HD2, the proportion of Ascomycota and Basidiomycota reached 72.39% and 27.02%, respectively. Meanwhile, MiSeq sequencing identified 10.54% Postia (Basidiomycota), whereas 80.44% OTUs were unidentified at the genus level. Not fungus was obtained by isolation method.
Not only is fungal infection of timber an unsightly and potentially hazardous to human health, it can also adversely affect the structural integrity of timbers and disrupt the use of buildings [6]. The tests revealed numerous mold species at certain proportions in rotten pillars. Degradation substances by wood decay fungi provides most of the nutrients for the growth of saprophyte molds. Mold spores can cause a great risk to the human body. Therefore, timely repair of wood decay of ancient buildings is also important for tourist health.
On the other hand, the moisture content at the bottom of the wood pillar is extremely high; thus, the role of bacteria in decay formation should not be ignored based on the progressive infection mechanism of microorganisms on wood [30]. The bacterial diversity of rotten pillars should be further analyzed.