Bacterial Diversity and Community Structure in Korean Ginseng Field Soil Are Shifted by Cultivation Time

Abstract

Traditional molecular methods have been used to examine bacterial communities in ginseng-cultivated soil samples in a time-dependent manner. Despite these efforts, our understanding of the bacterial community is still inadequate. Therefore, in this study, a high-throughput sequencing approach was employed to investigate bacterial diversity in various ginseng field soil samples over cultivation times of 2, 4, and 6 years in the first and second rounds of cultivation. We used non-cultivated soil samples to perform a comparative study. Moreover, this study assessed changes in the bacterial community associated with soil depth and the health state of the ginseng. Bacterial richness decreased through years of cultivation. This study detected differences in relative abundance of bacterial populations between the first and second rounds of cultivation, years of cultivation, and health states of ginseng. These bacterial populations were mainly distributed in the classes Acidobacteria, Alphaproteobacteria, Deltaproteobacteria, Gammaproteobacteria, and Sphingobacteria. In addition, we found that pH, available phosphorus, and exchangeable Ca+ seemed to have high correlations with bacterial class in ginseng cultivated soil.

Fig 1. Rarefaction curves depicting the effect of sequencing on the number of operational taxonomic units (OTUs) within 30 soil samples.

OTUs are shown at the 3% genetic distance levels.

Fig 2. Richness and diversity indices along a time gradient.

(A) Richness indices and (B) Shannon’s diversity index. 2-year-old (Y2, n = 6), 4-year-old (Y4, n = 12), 6-year-old (Y6, n = 9), and non-cultivated (Y0, n = 3) soil samples. All values are indicated as the mean ± SE. “a” and “b” indicate differences in alpha diversity indices values between 2-year-old and 6-year-old samples (one-way ANOVA by post hoc Tukey’s test). *, p<0.05; **, p<0.01.

Fig 3. Unweighted-pair-group method with arithmetic-mean (UPGMA) tree.

The UPGMA tress was generated from 1,000 jackknife iterations based on the non-normalized weighted UniFrac calculation. Dark blue, purple, grey, green, and yellow branches represent to soil samples from the Yulpori, Wondangri, Juwonri, Jajangri, and Jajakri areas, respectively. The circle, diamond, square, and triangle symbols indicate non-cultivated, 2-year-old, 4-year-old, and 6-year old soil samples, respectively. Black shapes represent non-cultivated samples, aqua is cultivated in the first round, and red is cultivated in the second round. Green font represents healthy samples and red represents unhealthy samples.

Fig 4. Relative abundance of bacterial phyla in 30 soil samples.

Relative abundances are reported as percent of total bacterial sequences observed per samples. The other category includes phyla showing a percentage of reads <2% of the total reads in all of the soil samples.

Fig 5. Heat map analysis showing the distributions of the core bacterial classes.

Fig 6. Network-inferred correlations of bacterial classes to soil parameters.

Soil parameters (yellow hexagons) included soil pH, available phosphorus (P₂O₅), and exchangeable Ca²⁺. Aqua V-shapes indicate the classes that decreased with cultivation time, and the red triangles indicate the classes that increased with cultivation time. The light green diamond (Actinobacteria) indicates changes associated with soil depths. Red and blue fonts indicate the classes that were more and less populous, respectively, in unhealthy samples compared to healthy samples. The solid green lines indicate positive correlations while dashed red lines indicate negative correlations. Edge width is the correlation value. All correlations are significant at p<0.05 (one-way ANOVA, post hoc Tukey’s test).

Fig 7. Heat map visualization of the distributions of core assigned families.

The assigned families composing the core microbiota of the ginseng soil samples, with up to 46.9% relative abundance.

Fig 8. Non-metric multidimensional scaling (nMDS) plots of the core bacterial families in 30 soil samples.

(A) nMDS plots show dissimilarities between cultivation years: non-cultivated (black), 2-year-old (aqua), 4-year-old (blue), and 6-year-old (red) samples. (B) nMDS plots show dissimilarities based on geography: Jajangri (JJG, black), Jajakri (JJK, pink), Wondangri (WD, darkgreen), Juwonri (JW, violet), and Yulpori (YP, orange).

Fig 9. Changes in bacterial populations according to rounds of cultivation.

(A) Boxplots indicate bacterial classes, (B) Boxplots indicate bacterial families. R0, non-cultivation; R1, first round of cultivation; R2, second round of cultivation. The p-values were generated by one-way ANOVA, post hoc Tukey’s test. “a” and “b” indicate significant differences between the round of cultivation (p<0.05).

Fig 10. Changes in bacterial populations over years of cultivation.

Bubbles represent the average relative abundances of bacterial populations at the class level (A) and the family level (B). 2-year-old (Y2, n = 6), 4-year-old (Y4, n = 12), 6-year-old (Y6, n = 9), and non-cultivated (Y0, n = 3) soil samples. All bacterial populations showed significant differences between cultivation times with p<0.05 (one-way ANOVA, post hoc Tukey’s test).

Fig 11. Healthy and unhealthy ginseng soil samples had significant differences in bacterial population.

(A) At the class level, (B) At the family level.