The influences of thorny bamboo growth on the bacterial community in badland soils of southwestern Taiwan
Abstract
The badland soils originated from mudstone in southwestern Taiwan are inhospitable for plant growth because of their high-salinity and poor physicochemical properties. Thorny bamboo is one of the few plants that can survive in these soils. To investigate the responses of the soil bacterial communities to plant cover, bare soils and soils from thorny bamboo plantations were analyzed with the barcoded pyrosequencing technique. Actinobacteria and Proteobacteria predominated in the bare soil communities, but Acidobacteria, Actinobacteria, and Proteobacteria were the most abundant groups in the thorny bamboo soils. Environmental stress may have selected for Actinobacteria in the bare community. Canonical correspondence analysis of the distributions of abundant operational taxonomic units also revealed consistent differences between the communities in the bamboo and bare soils. The bacterial diversity in the bamboo soils was also higher than that in the bare soils. The soluble organic carbon and nitrogen in bamboo soils were significantly higher, but electrical conductivity was lower than that in the bare soils. These soil properties, as well as soil pH, were related to the structure and diversity of the bacterial communities. Statistical analyses also indicated that these factors affected the distribution of Acidobacteria, Actinobacteria, Alphaproteobacteria, and Gammaproteobacteria. Analysis of thiocyanate oxidation activity revealed higher activity in the bare than in the bamboo soils, further suggesting differences in structure and metabolic activity between bamboo and bare soil microbial communities. Apparently, growth of thorny bamboo in the badland soil changed soil properties, which in turn directly and/or indirectly affected soil bacterial structure and diversity.
1 INTRODUCTION
In southern Taiwan, soils composed of Plio-Pleistocene mudstone form barren badlands devoid of vegetation. During the dry season, the mudstone becomes as hard as rock; but during the rainy season, the soil swells and softens upon water absorption, leading to slaking and rapid surface erosion. These badlands feature a series of ridges and gullies and are found throughout other arid or semiarid areas, including North America (Kasanin-Grubin & Bryan, 2007), southern Italy (Clarke & Rendell, 2006), and northern Spain (Desir & Marín, 2013). The badlands in southwestern Taiwan are formed from marine sedimentary rock (Lee, Chen, Wu, & Yang., 2013) and cover more than 1,000 km2. This area, often called a ‘Moonscape World,’ is representative of badlands in other parts of Monsoon East Asia. The average annual precipitation is about 2,000 mm (Lee, Lin, & Wu, 2007), much higher than the 645 mm found in more arid badlands (Piccarreta, Faulkner, Bentivenga, & Capolongo, 2006). Moreover, the average annual temperature, precipitation, and the intensity of heavy rainfall in this region have increased in recent years (Lee et al., 2007).
Plant cover is a crucial factor for soil microbial communities (Oh et al., 2012). Soil temperature and moisture are more stable under plant canopies in arid soils, such as those found in deserts, and this may lead to higher nutrient abundance. As a result, bacterial abundance and composition may be significantly different between shrub-covered and barren soils (Bachar, Soares, & Gillor, 2012). Microorganisms also have the ability to survive in arid soils with high levels of salt (Mondini, Cayuela, Sanchez-Monedero, Roig, & Brookes, 2006). Their responses to environmental fluctuation could be an important biological indicator in arid ecosystems, including badland regions.
Thorny bamboo (Bambusa stenostachya) is a dominant species for forestation of badland soils in Taiwan. As with other bamboo plantations, it is an important forest resource that is versatile and widely used for paper and pulp production as well as building construction. However, many thorny bamboo plantations have been abandoned for economic reasons, and management of the remaining plantations and their contributions to water and soil conservation in the badlands are receiving more attention. In the badlands, soil erosion is a serious problem, resulting in an irregular, wave-like landscape. The south-facing slopes receive strong sunshine, and soil moisture is usually low. Comparatively, north-facing slopes receive less sunshine and possess more soil moisture. The thorny bamboo prefers the higher soil moisture on the north-facing slopes. Consequently, bamboo covers the north-facing slopes, whereas south-facing slopes remain bare, as seen in Zuozhen, one of the field sites in this study (Figure 1a and 1b of this study).
Based on our soil survey, soil crusting is also common in the bare soil but not in the bamboo plantation soil surface. Soil crusting increases bulk density and penetration resistance (PR; Edwards, Burney, Richter, & MacRae, 2000). Bécel, Vercambre, and Pagès (2012) indicated that the PR value should be between 7.0 and 14 kg cm−2 to maintain good seed-to-soil contact; after 14 kg cm−2, root growth begins to be inhibited. In the bare soil near the sites studied here, the PR values are 14.5 ± 1.47 kg cm−2 and high enough to inhibit plant growth (Hseu, Jien, Chien, & Liou, 2014). The high values of PR and electrical conductivity (EC) could be the critical factors causing sparse vegetation in these badlands. Furthermore, crusting also seals the soil surface, reducing infiltration and increasing runoff on hillslopes. The results are very dry and hard surfaces for the bare badland soils (Singer & Shainberg, 2004).
Considering the critical functions that microbes play on organic matter decomposition and nutrient cycling in the soil ecosystems, the soil bacterial communities were analyzed to clarify their roles in the bare and bamboo plantation soils in the badland ecosystem. Here, soil samples were obtained from three study sites of bare and thorny bamboo plantations in badlands in southwestern Taiwan. One might hypothesize that the bacterial communities would be distinct between bare and bamboo plantation soils because differences in plant cover are critical in altering soil properties and the nutrient availability. To test this hypothesis, we applied barcoded pyrosequencing methods to elucidate the bacterial structure and diversity in the bare and bamboo plantation soil communities. This study will help us to understand how bacterial communities respond to thorny bamboo growth in badland soils and the relationship between soil properties and bacterial community structure and diversity in this extreme environment.
2 MATERIAL AND METHODS
2.1 Site description and soil sampling
This study was conducted at Zuozhen (S1; 22°85′N, 120°39′E) and Longci (S2; 22°90′N, 120°39′E), in Tainan City, and Tianliao (S3; 23°00′N, 120°43′E) in Kaohsiung City, located in southwestern Taiwan (Figure 1c). The locations of three sampling sites are owned by the Forestry Bureau and accessible by the public. At S1, some regions were developed as orchard at the valley. Some of the thorny bamboo was replanted with Taiwan Giant bamboo at S2. The elevation was less than 300 m asl, and the mean annual temperature was about 25 °C. The mean annual precipitation was about 2,000 mm, which mainly occurs during the rainy season from June to October. Thorny bamboo was the predominant vegetation in this area.
Soil samples were collected at each site from bare areas and thorny bamboo plantations. Three transect lines from bamboo plantation to bare soils were established at each study site (nine transect lines in total) in June 2014. Eight to 10 subsamples were retrieved from each transect line using a soil auger 4 cm in diameter and 10 cm deep. Equal volumes of subsamples from bare and thorny bamboo plantation soils were combined separately to compost as one soil sample. Visible detritus in soil, including roots and litter, were manually eliminated by hand, and then the soil was passed through a 2-mm sieve. Samples were then reserved at −20 °C, and DNA from the soil community was extracted within 2 weeks.
2.2 Soil analysis
A glass electrode (McLean, 1982) was used to examine soil pH at 1:1 soil/water slurry. Soil was oven dried at 105 °C for 24 h and weighed for water loss to determine the soil moisture. Soil organic carbon (C) and total nitrogen (N) were analyzed using the combustion method with an NCS elemental analyzer (Model NA1500 Fisons, Italy). Soil texture was analyzed with a pipette method (Gee & Bauder, 1986). EC was determined with saturation paste extract method (Rhoades, 1996). The sample was saturated with distilled water and mixed to a paste consistency. After standing for 1 hr, water was extracted from the paste and electrodes were used to measure its dissolved salts and the EC. Soil sulfate was extracted following a previous study (Rehm & Caldwell, 1968) except that samples were shaken for 1 h and activated charcoal was not added. Then, we used the turbidimetric method at 420 nm (American Public Health Association et al., 1998) and a spectrophotometer (SP-8001, Metertech Inc., Taiwan) to determine the extracted solution's content.
Cellulase (EC 3.2.1.4) and xylanase activity (EC 3.2.1.8) were analyzed according to Schinner and von Mersi (1990). Urease activity (EC 3.5.1.5) was determined using Kandeler and Gerber's method (1988). Phosphatase activity (EC 3.1.3.2) was determined using Tabatabai and Bremner's (1969) method. Protease activity was assayed following Ladd, Brisbane, Butler, and Amato (1976). β-Glucosaminidase was investigated in accordance with Parham and Deng (2000). We followed the description in Anderson and Domsch (1990) to determine soil basal respiration (respiration rate).
2.3 Barcoded pyrosequencing of the 16S rRNA genes
Soil community DNA was extracted, polymerase chain reactions were performed, barcodes were added, and polymerase chain reaction products were purified as described previously (Lin et al., 2015). To allow sequences to be grouped from one pyrosequencing run, each sample was tagged with unique, error-correcting barcodes (Hamady, Walker, Harris, Gold, & Knight, 2008). Mission Biotech (Taipei, Taiwan) performed amplicon pyrosequencing using the 454/Roche GS-FLX Titanium Instrument (Roche, NJ, USA). The accession number of the obtained pyrosequences submitted to the Short Read Archives is SRS1539383.
2.4 Sequence analyses
The Ribosomal Data Project (RDP) pyrosequencing pipeline was used to analyze pyrosequences (http://pyro.cme.msu.edu; RDP Release 11.4; release date: 2015.05.26). Sequences were sorted to each sample based on the tag file's barcode and then trimmed by removing the bar codes, primers, and linker. The pyrosequences were screened, and sequences that did not have Ns, were longer than 200 bp, and had quality scores >25 were used for further analyses. Taxonomic analyses were performed with the naïve Bayesian rRNA classifier of the RDP (Wang, Garrity, Tiedje, & Cole, 2007). Rarefaction curves and diversity indices were calculated using Complete Linkage Clustering data for operational taxonomic units (OTUs) with an evolutionary distance of 0.03. For community comparisons, the smallest number (2,900) of pyrosequences in a sample was applied to randomly select sequences in each sample to avoid measurement differences from sample size. The Mothur program (Schloss, Westcott, Ryabin, & authors, 2009) was used to derive the distribution of shared OTUs among soil communities, and this distribution was used to plot the heat map with Heatmapper (Babicki et al., 2016). The relationships between bacterial communities, phylogenetic groups, and soil properties were analyzed using PRIMER V6 software (Clarke & Gorley, 2006). To determine the effects of soil properties on bacterial community, canonical correspondence analysis (CCA) was performed using the Vegan package (Oksanen, Kindt, Legendre, O'Hara, 2017) in R v.3.2.1 program (R Development Core Team 2015).
2.5 Activity of thiocyanate oxidation
Three replicates (10 g wet weight for each replicate) from S1 and S3 study sites were incubated at 25 °C with 90 ml basic mineral medium. The medium contained 0.1-M NaCl, 1.7-mM potassium thiocyanate, 5-mM NH4Cl, and 25-mM NaHCO3 per liter. The pH was adjusted to 7.3. Potassium thiocyanate was used as electron donor, and the concentration of thiocyanate was analyzed periodically as described previously (Sörbo, 1957). The significance of changes in concentration of thiocyanate among soil samples were tested by using one-way analysis of variance.
3 RESULTS
3.1 Soil properties
Table 1 lists some basic properties of the mudstone soils. The bare soil was silt loam to silty clay loam and characterized by very high pH and EC values. Both soil organic C and total N were relatively low. The soil moisture of bare soil was also all lower than that in bamboo soils. The composition of the soils under the bamboo plantations was similar to that of the bare soils. However, the bamboo soils were characterized by lower pH, higher organic C and total N, and very low EC, in comparison with the bare soils. Based on the U.S. Soil Taxonomy (Soil Survey Staff, 2014), these bare mudstone and the bamboo soils are classified as Typic Eutrustept and Typic Dystrudepts, respectively (Shiau et al. 2017).
| Site | Vegetation | pH | Soil moisture (%) | Organic C (mg kg−1) | Total N (mg g−1) | SO42− (mg L−1 g−1) | EC (ms cm−1) |
|---|---|---|---|---|---|---|---|
| S1 | Bare | 8.6 ± 0.2 a | 7.2 ± 1.6 d | 34 ± 4.6 c | 7.0 ± 0.8 d | 42.5 ± 5.2 b | 7.6 ± 2.1 a |
| Bamboo | 7.5 ± 0.1 b | 15.9 ± 2.8 b | 90 ± 36 b | 13.0 ± 4.9 bc | 5.6 ± 0.3 d | 0.8 ± 0.1 b | |
| S2 | Bare | 8.4 ± 0.1 a | 8.7 ± 0.1 cd | 45 ± 2.4 c | 8.0 ± 1.2 d | 75.1 ± 5.1 a | 12.6 ± 4.4 a |
| Bamboo | 6.2 ± 0.6b | 20.5 ± 3.0 a | 177 ± 18 a | 21.0 ± 4.4 a | 7.3 ± 0.5 c | 0.7 ± 0.2 b | |
| S3 | Bare | 8.3 ± 0.2 a | 10.0 ± 0.6 c | 47 ± 1.6 c | 10.0 ± 1.4 cd | 45.6 ± 7.8 b | 13.3 ± 5.0 a |
| Bamboo | 7.4 ± 0.1 b | 17.0 ± 0.4 b | 111 ± 15 b | 14.0 ± 0.0 b | 5.9 ± 0.4 d | 0.7 ± 0.2 b |
- Note. pH, soil moisture, EC, organic C and total N contents were from Shiau et al. (2017). Results are means (SD) of three replicates. Values that are significantly different (p < .05) are identified with different letters. Sites: S1, Zuozhen in Tainan City; S2, Longci in Tainan City; S3, Tianliao in Kaohsiung City. CL = clay loam; EC = electrical conductivity.
The soil enzyme activities were affected significantly by bamboo plantation. Soil xylanase and proteinase activities in the bamboo plantations were significantly higher than those in bare soils, with the highest value in bamboo soil at S2 (Table 2). Soil cellulase, urease, acid phosphatase, and β-glucosaminidase did not show any significant difference between bamboo plantation and bare soils, with the exception of S2. Additionally, the bamboo soil at S2 exhibited the highest activities of these enzymes among the locations tested.
| Site | Vegetation | Cellulase (mg glucose g−1 day−1) | Xylanase (mg glucose g−1 day−1) | Urease (mmol NH4+-N g−1 hr−1) |
Acid phosphatase (μg nitrophenol g−1 hr−1) |
Proteinase (μg tyrosine g−12 hr−1) | β-Glucosaminidase (μg nitrophenol g−1 hr−1) |
|---|---|---|---|---|---|---|---|
| S1 | Bare | 94 c | 340 c | 0.02 b | 0.01 b | 0.02 c | 1.06 c |
| Bamboo | 94 c | 1,551 b | 0.64 b | 0.02 b | 60.0 b | 0.53 c | |
| S2 | Bare | 258 b | 346 c | 0.26 b | 6.57 b | 0.01 c | 14.8 b |
| Bamboo | 991 a | 2,793 a | 4.59 a | 51.1 a | 191 a | 93.0 a | |
| S3 | Bare | 295 b | 547 c | 0.07 b | 0.01 b | 0.01 c | 11.1 b |
| Bamboo | 270 b | 2,126 b | 0.51 b | 0.02 b | 93.8 b | 14.6 b |
- Note. Values in each column with different letters indicate significant differences between treatments (Duncan's multiple range test, p < .05). Site abbreviations are described in Table 1.
3.2 Structure of soil bacterial communities
In this study, about 2,900–7,600 pyrosequences were obtained from each of the three replicates, and 9,600–16,700 pyrosequences of average length of 274 bp were obtained from each soil type at each site (Table S1). The major phylogenetic groups represented in the pyrosequence libraries differed greatly between the bare and bamboo plantation soils. The phylum Actinobacteria was the most abundant group and comprised more than 40% of all pyrosequences from bare soils, although it was only about 9–25% in bamboo communities (Figure 2a). In contrast, the Acidobacteria accounted for 18–27% of the reads in the bamboo plantation soils but less than 1% in the reads from the bare soil communities. Within the Acidobacteria, Subgroups 1, 4, and 6 were the most abundant groups in the communities of bamboo communities (Figure 2b). Although the proportion of Proteobacteria was relatively constant in both soil types, 33–40% and 26–36% of pyrosequences in the bare and bamboo communities, respectively, there were large differences in the abundances of specific classes. The Gammaproteobacteria was the most abundant class (20–30%) in the bare soils and much less in the bamboo plantation soils (<5%). In contrast, Alphaproteobacteria was much more abundant in the bamboo plantation soils. Although much less abundant, large differences were also discovered for the relative abundances of the Betaproteobacteria and Deltaproteobacteria. The proportions of Bacteroidetes, Firmicutes, and Gemmatimonadetes were also higher in the bamboo plantation communities. Other phyla, such as Chloroflexi, Cyanobacteria, Nitrospira, and Planctomycetes, were all less than 2% in both bare and bamboo soil communities. These large differences between the soils were consistent in all three sites examined.
3.3 Diversity of the soil bacterial communities
Based on the genetic distance at 0.03, 863–902, and 4,301–4,724, OTUs were identified in the bare and bamboo plantation soils, respectively. The Shannon diversity index and species richness suggested that the bamboo plantation soils were more species-rich than the bare soils (Table 3). The rarefaction curve analyses also supported this conclusion (Figure S1). Thus, soils from the bamboo plantations were more diverse than those from the bare soil communities.
| Site | Vegetation | Species richness | Evenness | Shannon | Chao 1 |
|---|---|---|---|---|---|
| S1 | Bare | 698 b | 0.73 b | 4.77 b | 1,129 b |
| Bamboo | 2,948 a | 0.92 a | 7.33 a | 5,151 a | |
| S2 | Bare | 632 b | 0.67 b | 4.33 b | 886 b |
| Bamboo | 2,928 a | 0.93 a | 7.39 a | 5,334 a | |
| S3 | Bare | 633 b | 0.63 b | 4.04 b | 939 b |
| Bamboo | 2,990 a | 0.93 a | 7.44 a | 5,181 a |
- Note. Different letters in each row reveal the significant differences between values at p < .05.
3.4 Bacterial community comparisons
For community comparison, we performed the analysis of distribution of abundant genera of badland soil communities. Only the 70 most abundant genera accounted for more than 0.1% of all sequences and were used for further analysis. With a few exceptions, genera that were abundant in one soil type were either absent or present in only low amounts in the other soil type. Thus, within the Acidobacteria, the GP1, GP4, GP5, and GP6 were abundant genus-level groups found in the bamboo plantation soils, and Thiohalophilus, belonging to the Gammaproteobacteria, was very abundant in the bare soils. Within the Actinobacteria, genera Actinophytocola and Amycolatopsis were mainly retrieved from bare soils. In contrast, both Gaiella and Ohtaekwangia were moderately abundant in both soil types. The relative abundance of Gemmatimonas, affiliated to the phylum Gemmatimonadetes, was also higher in the bamboo communities than that in bare soils (Figure 3, Table S2). CCA also showed that the structures of the bare area and bamboo plantation soils were distinct (Figure 4). Bacterial communities were clustered based on the vegetation difference across replicate samples. The abundant phyla and OTUs also showed similar trends. The replicates from the communities at the bamboo plantation soils for each site had more variance within and between the sites (Figure 4). The analysis also revealed a positive correlation between bamboo community structure and soil organic C and TN, whereas soil pH and EC values positively correlated with the bare soil communities. The community structure based on the pyrosequences from each site was mainly affected by the soil properties, including soil pH, soil organic C and total N, and EC (Table 4). Bacterial Shannon diversity index was also influenced by these factors, except soil pH and total N. In addition, the distribution of Actinobacteria and Alphaproteobacteria positively interacted with soil pH, soil organic C and total N, EC. The distribution of Acidobacteria was negatively correlated with soil pH.
| pH | Organic C | Total N | EC | Clay | |
|---|---|---|---|---|---|
| Shannona | 0.088 | 0.356 | 0.212 | 0.680 | 0.026 |
| All communityb | 0.291 | 0.610 | 0.347 | 0.404 | −0.147 |
| Acidobacteria | −0.507 | 0.778 | 0.577 | 0.492 | −0.110 |
| Actinobacteria | 0.168 | 0.397 | 0.171 | 0.528 | −0.112 |
| Alphaproteobacteria | 0.309 | 0.528 | 0.301 | 0.507 | −0.018 |
| Gammaproteobacteria | 0.002 | 0.120 | 0.087 | 0.593 | 0.069 |
- Note. Values in bold are p ≤ .05. EC = electrical conductivity.
- a Shannon diversity index.
- b Analyzed with all pyrosequences from each site.
3.5 Activity of thiocyanate oxidation
Because thiocyanate-oxidizing bacteria of the genus Thiohalophilus were abundant in bare soils, thiocyanate oxidation activity of soil slurries was examined. The concentration of thiocyanate in S1 and S3 bare soils decreased rapidly with initial rates of 9.0 and 7.1 μmol g of soil−1 day−1, respectively, to less than 0.1 mM after 28 days of cultivation (Figure 5). However, in the S1 and S3 bamboo soils, the initial rates were much lower, 1.4 and 3.1 μmol g of soil−1 day−1, respectively, and the concentration of thiocyanate remained greater than 0.4 mM throughout the incubation. Therefore, the bare soils possessed much higher levels of thiocyanate oxidation activity. The pathway of thiocyanate oxidation would form SO42− (Sorokin, Tourova, Lysenko, & Kuenen, 2001). The concentration of SO42− of bare soils was also significantly higher than that in the bamboo soils, as might be expected if this process was common in these soils (Table 1).
4 DISCUSSION
The bacterial community of thorny bamboo plantation soils had far higher diversity than the bare soils in the badland ecosystem. The structures of the bacterial soil communities also differed along with major soil characteristics, including organic C, total N, and EC, between bare and bamboo forest soils. Previous studies revealed that vegetation could be an important factor to alter bacterial communities. Vegetation influences soil properties, microbial communities, and enzyme activities (Sauheitl, Glaser, Dippold, & Weigelt, 2010; Ushio, Kitayama, & Balser, 2010). These effects are largely the result of root systems releasing plant organic matter into soil (Berg & Steinberger, 2008; Haichar et al., 2008; McCaig, Glover, & Prosser, 1999). For instance, under bamboo trees, the surface soils contained much higher organic matter (SOM), presumably derived from plant exudates and litter. Formation of organic acids from the SOM would then be expected to reduce the soil pH. Similarly, the inhibition of soil crusting by SOM would allow rain to eluviate cations from the soils and further reduce the pH and EC values. Plant cover also forms patchy habitats and “resource islands” in such resource-limited environments (Bachar et al., 2012). In the arid and semiarid soils, bacteria are also more abundant in these plant patches (Bachar et al., 2012). In this study, CCA also indicated the positive correlation between soil organic C, total N, and bacterial communities of soils covered with bamboo. The plant canopy also provides protection from high radiation and wind velocity, increasing the water availability under the canopy (Berg & Steinberger, 2008). Without vegetation, bare surface soils are subjected to large temperature and moisture fluctuations and high ultraviolet (UV) light exposure. They are also low in organic matter. The study of Shiau et al. (2017) indicated that the thorny bamboo plantation in the same badland regions helped to improve soil physicochemical properties (e.g., increased organic C and water holding capacity and decreased in soil bulk density). Bamboo plantation also resulted in increasing microbial C and N and soluble organic C. The microbial biomass C to total organic C ratio was higher in bamboo plantations than bare land soils, suggesting that microbes utilize soil organic matter more efficiently in bamboo soils (Shiau et al., 2017). Moreover, the soil respiration to microbial biomass C ratio was higher in bare soils, indicating that microbes use energy less efficiently there (Anderson, 2003) and therefore may be under higher stress in bare soils than in bamboo soils (Shiau et al., 2017). This environmental stress in bare soils may explain the low bacterial diversity. Hence, the difference in bacterial diversity and structure could also result from changes in soil properties from vegetation cover and microclimate.
Several studies have shown that enzyme activities depend on the quantity of SOM and the microbial biomass (Chang, Chung, & Wang, 2008; Kanchikerimath & Singh, 2001). In this study, the enzymatic activities were highly correlated with SOM and microbial biomass. Thus, the differences in enzyme activities between bamboo plantation and bare soils may have been due to different SOM content. The enzyme activity was also positively correlated with the content of SOM and negatively correlated with pH, which agrees with the results of Safari and Sharifi (2006).
The thorny bamboo plantation soil community was also different from other bamboo soils. Previous studies showed that the Acidobacteria accounted for 30–50% of all pyrosequences in moso bamboo plantations (Lin et al., 2014; Lin et al., 2015), whereas they comprised less than 30% in this study. Soil pH affects Acidobacteria distribution, and their relative abundance decreases with increasing soil pH (Lauber, Hamady, Knight, & Fierer, 2009). Because the soil pH values in moso bamboo plantation soils were about 4.0, the more neutral pH in thorny bamboo soils could explain the lower abundance of Acidobacteria.
Compared with bare communities, the Acidobacteria were far more abundant in the thorny bamboo soils in this badland ecosystem. As oligotrophs (Nemergut, Cleveland, Wieder, Washenberger, & Townsend, 2010) and versatile heterotrophs, they could have low rates of metabolism when nutrients are low (Ward et al. 2009). Considering their ability to survive in low-nutrient environments, the low abundance of Acidobacteria in bare soils indicates that high levels of environmental stress, such as desiccation and radiation, negatively affect members of the phylum.
In this study, the relative abundance of Actinobacteria in the bare communities was quite high, and they comprised nearly 50% of all pyrosequences. This phylum dominates the bacterial communities of many arid soils. They comprised more than 80% of Atacama Desert communities (Crits-Christoph et al., 2013). Actinobacteria was also the dominant phylum in soil from Israel's Negev desert, which has low levels of available soil water and organic matter (Saul-Tcherkas & Steinberger, 2011). Members of this phylum have been found at high abundances in other soils with low moisture (Alvarez, Silva, Cesari, & authors, 2004; LeBlanc, Gonçalves, & Mohn, 2008). In this study, the genus Amycolatopsis was abundant in bare soils. It was also common in arid soils suffering from the physicochemical stress of desiccation (Okoro et al., 2009). In the badland ecosystem, rapid drying and lack of water infiltration make the bare soils similar to arid desert soils during the dry season, which may explain the abundance of this phylum. Another abundant genus, Gaiella, was mainly present in the bamboo community. This genus is sensitive to high doses (1 kGy) of γ-ionizing radiation (Albuquerque et al., 2011), which is often correlated to sensitivity to UV radiation and desiccation. Because the bare soils were subjected to large fluctuations in moisture and UV radiation, the development of Gaiella may have been precluded.
Compared with the bare soils, Alphaproteobacteria were more abundant in the bamboo plantation soil communities. This group also dominated soils beneath shrubs in arid and semiarid ecosystems (Bachar et al., 2012). Members of this group, including Bradyrhizobium and Rhodoplanes, are able to form associations with plants (Yarwood, Myrold, & Högberg, 2009), and this might explain their abundance in the thorny bamboo plantation soils. This group was also abundant in the moso bamboo plantation soil community (Lin et al., 2014).
It is unusual for a single OTU to make up more than 10% of soil microbial communities. Therefore, the high proportion of gammaproteobacterial sequences related to Thiohalophilus retrieved from bare soils is remarkable. This autotrophic bacterium is moderately halophilic and capable of growing on reduced sulfur compounds including hydrogen sulfide, thiosulfate, elemental sulfur, and thiocyanate (Sorokin, Tourova, Bezsoudnova, Pol, & Muyzer, 2007). Sulfur oxidation uses O2 as an electron acceptor. Reduced sulfur compounds are oxidized to sulfate, which is highly mobile and may be leached to the bamboo plantation's deep soils, which are very porous. This may result in less sulfate accumulating in the bamboo plantation soils than the bare soils. The higher concentration of SO42− and thiocyanate oxidation activity in slurries from the bare soils suggested that these bacteria were metabolically active in this soil. These results indicate that bare soils in the badland ecosystem have another supply of soil organic C, because the lack of plant cover and root exudates in the bare region. Further studies will be necessary to test this hypothesis.
A group of pyrosequences were related to the genus Gemmatimonas of Gemmatimonadetes. They were more abundant in bamboo communities and had low abundance in bare soils. This phylum is found in a variety of environmental 16S rRNA gene libraries. It is an abundant group in soils: High throughput sequencing studies have shown relative abundances of 0.2–6.5%, with a mean of 2.2% (DeBruyn, Nixon, Fawaz, Johnson, & Radosevich, 2011). Analysis of Gemmatimonadetes 16S rRNA gene sequences from various ecosystems indicates that they are adapted to low soil moisture, which could explain their presence in this harsh badland ecosystem (DeBruyn et al., 2011). In addition, Park, Kim, and Yoon (2017) revealed a unique regulatory mechanism of N2O reduction in an obligate aerobic Gemmatimonas strain, suggesting that this strain may play a role in nitrogen cycling in this ecosystem.
5 CONCLUSION
Bacterial communities showed distinct differences in diversity and structure between bare and thorny bamboo plantation soils. Vegetation cover likely altered the soil properties, such as organic C, total N, and EC, and directly and indirectly affected the soil bacterial community. The high abundance of Actinobacteria in the bare soil community indicates that environmental stress played a crucial part in selection of the badland soil bacterial community. Higher activity of thiocyanate oxidation, which is associated with Gammaproteobacteria, in the bare soils also showed the difference in bacterial structure and metabolic activity and could be helpful to the survival of microbes in this ecosystem. Further research about their function in the ecosystems with such low soil moisture and organic matter availability are needed to fully investigate the ecological effects of badlands to the terrestrial habitats.
ACKNOWLEDGMENTS
We thank Ms. Yu-Shiuan Huang and Ms. Pei-Yi Yu from the Biodiversity Research Center, Academia Sinica, Taipei, Taiwan, for laboratory work.
CONFLICT OF INTEREST
The authors declare no conflict of interest.
