The effect of flue-curing procedure on the dynamic change of microbial diversity of tobaccos

Comparison of sampling methods for microbes on the surface of tobacco leaves

According to previous research, two sampling methods for microbes on the surface of tobaccos were selected to separately perform extraction and amplification of genome DNAs after sampling the microbes. As shown in Table 1, as for the first sampling method, the DNAs extracted from two tobacco leaf samples (fresh tobacco leaves and tobacco leaves in the later yellowing period) were both unqualified after subjected to amplification. Therefore, the first method was not suitable for extracting the microbes on the surface of tobacco leaves. The two tobacco samples extracted by using the second method both allowed favorably amplification and their amplification results were both proper. Thus, the second method was applied to sample the microbes on the surface of tobacco leaves subsequently.

OTU clustering analysis

To explore the species compositions of various samples, OTUs clustering was carried out on effective Tags of all samples based on 97% of identity; afterwards, species annotation was performed on the OTUs sequences. According to OTUs results obtained through clustering and research requirements, the common and specific OTUs among different samples (groups) were analyzed.

OTU clustering analysis of bacteria in tobacco leaves

The result is shown in Fig. 3. Each petal in the petal diagram represents a group (sample) and different colors mean diverse samples (groups); the number at the core stands for the total number of OTUs in all samples; the number in each petal denotes the number of OTUs specific in the sample (group). It can be seen from the figure that the numbers of the core microbial communities subjected to conventional flue-curing procedure and dry-ball temperature set and wet-ball temperature degradation flue-curing procedure were basically consistent, showing no great change.

OTU clustering analysis of fungi in tobacco leaves

The result is displayed in Fig. 4. As shown in the figure, the core microbial communities flue-cured by conventional procedure and temperature- and humidity-controlled procedure showed basically coincident numbers. The latter was only 5–10 core microbial communities more than the former. Similar to the core bacterial communities, the number of core fungal communities presented no great difference in the flue-curing process.

Analysis of relative abundances of species

According to the results of species annotations, the species of each sample or group with the maximum abundance ranking the top 10–30 at various classification levels were selected to generate the cumulative histogram of relative abundances of species. The species with a high relative abundance in various samples at different classification levels and their proportions can be intuitively found.

Analysis of relative abundances of bacteria in tobaccos

Based on the results of species annotations, the species of each sample or group with the maximum abundance ranking the top 10–30 at various classification levels were selected to generate the cumulative histogram of relative abundances of species. The species with a high relative abundance in various samples at different classification levels and their proportions can be visualized.

As shown in Fig. 5, at the level of phylum, the bacteria in tobaccos mainly contained Proteobacteria, Actinobacteria, Bacteroidetes, Firmicutes, Planctomycetes, Acidobacteria, Chloroflexi, unidentified bacteria, Thaumarchaeota and Gemmatimonadetes. It can be seen that the surface and endophytic bacterial communities of fresh tobacco leaves slightly differed at the level of phylum. Proteobacteria showed the largest content, followed by Actinobacteria and Bacteroidetes, and the contents of the other bacterial phyla were relatively low. By using conventional flue-curing procedure, the bacterial diversity on the surface of tobacco leaves progressively declined as the flue-curing continued, and the relative content of Proteobacteria rose at first and then reduced; the reduction amplitude of Actinobacteria was relatively stable in the flue-curing process while that of Bacteroidetes was relatively large. For the dry-ball temperature set and wet-ball temperature degradation flue-curing procedure, as the flue-curing proceeded, the relative content of Proteobacteria gradually increased and it did not greatly reduce until reaching the last flue-curing stage. However, its relative content was not significantly different from that in fresh tobacco leaves; similar to Proteobacteria, the relative contents of both Actinobacteria and Bacteroidetes also grew at first and then decreased; in terms of endophytic bacteria of tobaccos, as the flue-curing process continued, the relative contents of Proteobacteria and the other main bacterial communities rapidly dropped while those of the other communities sharply increased.

As shown in Fig. 6, at the level of genus, the main dominant bacterial communities in endophytic bacteria of fresh tobacco leaves included Pseudomonas, Sphingomonas, Ralstonia, Methylobacterium, Massilia, Sphingobacterium, Rhizobium, Halomonas, Serratia and Rickettsia.

When employing conventional flue-curing procedure, the bacterial communities on the surface of tobaccos were relatively marginally changed at the level of genus while the endophytic bacteria varied remarkably. As the flue-curing process proceeded, the relative content of 30 main endophytic bacterial communities found before the flue-curing had dropped to 2% even in the early flue-curing stage (35 ℃). In comparison, the relative content of the bacterial communities in the first two flue-curing stages under dry-ball temperature set and wet-ball temperature degradation flue-curing procedure was higher.

Under conventional procedure, the relative abundance of Pseudomonas on the surface of tobacco leaves increased at first and then decreased, so did that of Sphingomonas. Although no signs of Ralstonia solanacearum were visualized on the surface of the sampled tobacco leaf samples, Ralstonia was found in the analysis of bacterial communities. With the ongoing flue-curing process, the relative content of Ralstonia rapidly reduced; the relative content of Methylobacterium on the surface of fresh tobacco leaves declined to some extent in the flue-curing process, and accounted for a large proportion in bacterial communities on the surface of flue-cured tobacco leaves. The relative contents of the other main bacterial communities were all progressively lowered basically.

When implementing dry-ball temperature set and wet-ball temperature degradation flue-curing procedure, the relative contents of the main bacterial communities in the early flue-curing stage were higher than those in fresh tobacco leaves. The relative contents of them marginally differed from those in fresh tobacco leaves even though flue-curing process was ended; the relative content of Pseudomonas gradually increased in the flue-curing process. By contrast, the relative contents of Sphingomonas and Methylobacterium both grew at first and then declined. The relative contents of the other main bacterial communities relatively slowly varied in the flue-curing process and they did not greatly decrease until the flue-curing process was ended. Moreover, the relative contents of some bacterial genera, including Sphingobacterium and Rickettsia, had remarkably dropped in the early flue-curing stage.

Analysis of relative abundances of fungi in tobaccos

As shown in Fig. 7, at the level of phylum, fungi in tobaccos mainly covered Ascomycota, Basidiomycota, Mortierellomycota, Rozellomycota, Glomeromycota, Chytridiomycota, Kickxellomycota, Mucoromycota and Olpidiomycota. It can be seen from the figure that fungi in tobaccos mainly included Ascomycota and Basidiomycota; the other fungi (phylum) took up a relatively low proportion.

When being flue-cured by using conventional procedure, the relative abundance of Ascomycota on the surface of tobacco leaves gradually increased while those of Basidiomycotaand the other fungal phyla gradually decreased with the ongoing flue-curing process; under temperature- and humidity-controlled flue-curing, the evolution law of fungal communities was similar to that using conventional procedure at the level of phylum. To be specific, a trend was shown that the relative abundance of Ascomycota gradually rose while those of Basidiomycota and the other fungal phyla were lowered, which was basically similar to that under conventional flue-curing procedure.

Their proportions were higher than those on the surface of tobacco leaves.The change amplitude of the endophytic fungi of tobacco leaves was less significant than that of fungi on the surface of tobacco leaves. Either under conventional flue-curing or temperature- and humidity-controlled flue-curing, the relative abundance of Ascomycota basically increased at first and then declined while that of Basidiomycota reduced at first, then grew and finally dropped.

As shown in Fig. 8, the change trends of community compositions under conventional flue-curing and temperature- and humidity-controlled flue-curing at the level of genus were similar to those at the level of phylum. There was a great difference only in the relative contents of fungal communities. In terms of fungi on the surface of tobacco leaves, the relative content of Alternaria under conventional flue-curing greatly increased at 38.5 ℃ and 54 ℃, with increases at the same time points under temperature- and humidity-controlled flue-curing; however, the growth amplitude of the relative content was less significant than that under conventional flue-curing. Cladosporium was another main fungal community and its relative content slowly decreased in the later stage of temperature- and humidity-controlled flue-curing. The relative content of Symmetrospora progressively decreased when using conventional flue-curing procedure while its reduction rate slowed down under temperature- and humidity-controlled flue-curing. The relative content of Ophiocordyceps on the surface of tobacco leaves was relatively low and it both gradually reduced when using the two flue-curing technologies. Moreover, the relative contents of the other fungal genera also progressively declined as the flue-curing proceeded.

As for changes of endophytic fungi of tobaccos when using the two flue-curing technologies, the relative content of Alternaria under conventional flue-curing was higher than that under temperature- and humidity-controlled flue-curing. The result can be found even though the flue-curing process was ended. The relative content of Cladosporium marginally varied under conventional flue-curing and greatly increased at 35 ℃. After completing the flue-curing process, the relative content did not significantly differ from the value in fresh tobacco leaves. However, for dry-ball temperature set and wet-ball temperature degradation flue-curing procedure, the relative content of Cladosporium progressively reduced on the whole and the value after ending the flue-curing process was only about half of that in fresh tobacco leaves. The relative content of Symmetrospora showed a same change trend with Cladosporium under conventional flue-curing and temperature- and humidity-controlled flue-curing. Additionally, the reduction rate of the relative content of Symmetrospora was higher than that of Cladosporium under temperature- and humidity-controlled flue-curing. Relative to fungal communities on the surface of tobaccos, although the relative contents of the other endophytic fungal genera gradually dropped with the flue-curing.

Clustered heat maps of species abundances

According to species annotations and abundances of all samples at the level of genus, genera whose abundances ranked the top 35 were selected. Subsequently, based on the abundances of these genera in each sample, a heat map is drawn by conducting clustering from the two aspects: i.e. species and samples. By doing so, it is convenient to ascertain a species with a high abundance or low content and the sample from which it is found.

Clustered heat map of species abundances of bacteria in tobaccos

The result is displayed in Fig. 9. It can be seen from the clustered heat map that bacterial communities mainly reside on the surface and in the interior of the fresh tobacco leaves at first. As the flue-curing proceeded, the contents of the main bacterial communities were changed to some extent: the main bacterial communities reduced in the content and the tobacco leaves became light in color.

Clustered heat map of species abundance of fungi in tobaccos

According to species annotations and abundances of all samples at the level of genus, the genera whose abundances ranked the top 35 were selected. Based on the abundances of these genera in each sample, a heat map is drawn by conducting clustering from the two aspects, i.e. species and samples. It can be found from Fig. 10 that the distribution of fungi on the surface of tobaccos greatly differed from that of endophytic fungi. Moreover, the distribution of materials also presented a great difference when using the two flue-curing technologies.

Correlation with environmental factors

Correlation of bacteria with environmental factors

Temperature and humidity were mainly controlled in the flue-curing process of tobacco leaves; and the main environmental factors involved dry- and wet-bulb temperatures. It can be seen from Fig. 11 that Pantoea, Nesterenkonia, Staphylococcus, Variovorax, Chryseomonas, Rhodococcus, Paracoccus, Massilia, Serratia, Ralstonia and Pseudomonas were more likely to be affected by temperature and humidity. Pantoea and Variovorax exhibited a positive correlation with temperature and humidity; Nesterenkonia, Staphylococcus, Chryseomonas, Rhodococcus, Paracoccus, Serratia and Ralstonia presented a negative correlation with temperature and humidity. Actinomycetospora were negatively correlated with the dry-bulb temperature while positively correlated with the wet-bulb temperature; Stenotrophomonas, Cutibacterium and Sediminibacterium were all negatively correlated with both dry- and wet-bulb temperatures while they presented a higher negative correlation with the dry-bulb temperature.

Correlation of fungi with environmental factors

Similar to the analysis method of environmental factors of bacteria, the correlation between the fungi in tobaccos and environmental factors is displayed in Fig. 12. Temperature and humidity were mainly controlled in the flue-curing process of tobacco leaves; the main environmental factors were dry- and wet-bulb temperatures. As shown in the figure, different from the correlation of bacterial communities in tobaccos with environmental factors, the majority of fungal genera in tobaccos presented a negative correlation with temperature and humidity, for example, Rachicladosporium, Vishniacozyma, Symmetrospora, Sarocladium, Ascochyta, Wallemia, Colletotrichum, Fusarium, Claviceps, Cladosporium, etc. A small number of fungal genera (such as Ustilaginoidea, Septoriella and Alternaria) were positively correlated with temperature and humidity.

Function prediction of bacterial communities in tobaccos

According to the annotation result in the database, the functions with the maximum abundance ranking the top 10 in each sample or group at various layers of annotations were selected to generate the cumulative histogram of relative abundances of functions. Thus, it is convenient to check the functions with a high relative abundance in various samples at different layers of annotations and their proportions.

As shown in Fig. 13, bacterial communities with a half of relative abundances participated in the metabolism and a quarter of bacterial communities took part in the genetic information processing; the rest was engaged in the cellular processes and organismal systems and some bacterial communities were implicated to human diseases.

Function prediction of fungal communities in tobaccos

Based on amplificon analysis of 16S or ITS, the species classification and abundances of fungi present in the environment can be attained. In many cases, people will also concern what role these species found in the environment play in the ecological environment. By applying FunGuild tool, it is feasible to attain the ecological functions of corresponding fungi based on the species classification of fungi. It can be seen from the Fig. 14 that the fungal communities in tobaccos delivered relatively abundant functions, in which the three fungal communities with the highest relative abundances separately showed the following functions: plant saprophytes, plant pathogens and undefined functional fungal communities; the rest of fungal communities participated in plant parasitism,soil-borne plant pathogen, lichenization, dung saprotroph, etc.

Source: Ecology - nature.com