Knowledge of structure and function of microbial communities in different
successional stages of biological soil crusts (BSCs) is still scarce for
desert areas. In this study, Illumina MiSeq sequencing was used to assess the
compositional changes of bacterial communities in different ages of BSCs in
the revegetation of Shapotou in the Tengger Desert. The most dominant phyla
of bacterial communities shifted with the changed types of BSCs in the
successional stages, from Firmicutes in mobile sand and physical crusts to
Actinobacteria and Proteobacteria in BSCs, and the most dominant genera
shifted from
Biological soil crusts (BSCs) are assemblages of cryptogamic species and microorganisms, such as cyanobacteria, green algae, diatoms, lichens, mosses, soil microbes and other related microorganisms that cement the surface soil particles through their hyphae, rhizines/rhizoids and secretions (Eldridge and Greene, 1994; Li, 2012; Pointing and Belnap, 2012; Weber et al., 2016). Due to their specialized structures and complicated assemblages of their members, BSCs constitute one of the most important landscapes and make up 40 % of the living cover of desert ecosystems, even exceeding 75 % in some special habitats (Belnap and Eldridge, 2003).
It is well known that BSCs play critical roles in the structure and function of semiarid and arid ecosystems (Eldridge and Greene, 1994; Li, 2012). They provide ecological services such as soil stabilization, reduction of wind and water erosion, and facilitation of higher plant colonization (Belnap, 2003; Belnap and Lange, 2001; Maier et al., 2014; Pointing and Belnap, 2012). BSCs are functionally important and variable, and may be a useful model system for diversity-function research. Their functional attributes are relatively well known, and estimation and manipulation of biodiversity in experiments are feasible, at least within some groups of BSC biota (Bowker et al., 2010). This relationship is more easily interpreted in artificially constructed BSCs. There are primary successional stages for BSCs in desert ecosystems: mobile sand, algal crust, lichen crust and moss crust (Lan et al., 2012a; Liu et al., 2006). The different successional stages of BSCs vary in their ecological function (Belnap, 2006; Bowker et al., 2006b; Li, 2012; Moquin et al., 2012).
During BSC succession, physical crusts in mobile sand contain the lowest carbon (C) and nitrogen (N) contents (Zhang et al., 2009). Algal crust is the earliest stage; it has a thin surface layer composed of eolian-borne materials and an organic layer formed by filamentous cyanobacteria associated with sand particles (Housman et al., 2006; Zhang, 2005; Zhang et al., 2009). Lichen and moss appear following stabilization of the algal filaments on the soil surface. The C and N fixation rates are increased in lichen crust (Evans and Lange, 2003; Lan et al., 2012b; Zhang et al., 2010), and there is higher photosynthesis, exopolysaccharide and nitrogenase activity in moss crust than in the early successional crusts (Housman et al., 2006; Lan et al., 2012b). In the BSC successional process, the microbial composition and community structure change greatly (Hu and Liu, 2003; Zhang et al., 2009). Crust succession is positively correlated with phospholipid fatty acid content and microbial biomass (Liu et al., 2013). The microbial biomass of soils is the greatest driving force in most terrestrial ecosystems, largely due to control of conversion rates and mineralization of organic matter (Albiach et al., 2000; Baldrian et al., 2010).
Bacteria present the highest proportion of the microbial biomass in BSCs (Bates et al., 2010; Green et al., 2008; Gundlapally and Garcia-Pichel, 2006; Maier et al., 2014; Wang et al., 2015) and thus have important roles in the BSC successional process. They can decompose organic material and release nutrients, mediating geochemical processes necessary for ecosystem functioning in the persistence of BSCs (Balser and Firestone, 2005). Species composition and community structure of bacteria change greatly during the successional process of BSCs (Gundlapally et al., 2006; Moquin et al., 2012; Zhang et al., 2016). Most research on prokaryotic diversity of BSCs has focused on cyanobacteria-dominated biocrusts in arid and semiarid regions (Abed et al., 2010; Garcia-Pichel et al., 2001; Nagy et al., 2005; Steven et al., 2013; Yeager et al., 2004). Recent studies of the bacterial community structure of bryophyte- or lichen-dominated crusts indicate that lichen-associated communities encompass a wide taxonomic diversity of bacteria (Bates et al., 2011; Cardinale et al., 2008; Maier et al., 2014). Heterotrophic bacteria may perform a variety of roles such as nutrient mobilization and N fixation and could be of considerable importance for the stability of lichen-dominated soil communities. However, there have been few studies on changes of bacterial diversity and their function in BSCs during the development process in desert zones, and these have only focused on the Sonoran (Nagy et al., 2005) and Gurbantunggut deserts (Zhang et al., 2016). What changes occur in bacterial community composition and in their potential roles in improving soil properties for different BSC successional stages? What is the significance of these changes to BSC succession in the recovery process of desert revegetation in temperate zones?
Bowker (2007) examined the role of BSCs in primary succession (vs. secondary succession) during a time when resources were available (e.g., light); however, they became less important once higher vegetation took over. In some environments of high abiotic stress (e.g., deserts), BSCs play a role in succession but remain a permanent component. Bowker's review and discussion is supported by work performed in southern Africa (Büdel et al., 2009) in which different successional BSCs are described. Büdel et al. (2009) also describe in detail crust types that were representative of successional stages. Castillo-Monroy et al. (2011) showed few BSC effects on ecosystem function could be ascribed to bacteria.
A recent study on crusts in the Tengger Desert, China, showed that bacterial diversity and richness were highest after 15 years, and at least 15 years might be needed for recovery of bacterial abundance of BSCs (Liu et al., 2017). To better understand these questions, we must analyze in detail the bacterial community composition of BSCs at all levels of classification and their corresponding function in the recovery process of BSCs. In the present study, bacterial community composition and potential function were analyzed in BSCs along a chronosequence of over 50-year-old revegetation. We investigated the following questions. What are the drivers of bacterial composition over time? What are the micro-processes that drive bacterial composition and function? Do bacteria drive changes in soil physicochemical properties which in turn have a direct influence on bacterial composition and function?
Sand dune landscape before (MS,
The study site is located in Shapotou, at the southeast fringe of the Tengger
Desert, northwest China. The natural landscape is characterized by the
reticulated chains of barchan dunes with a vegetation cover of less than
1 %. The mean annual precipitation is about 180 mm with large seasonal
and interannual variation. The mean wind speed is 3.5 m s
The detailed sampling method is shown in Fig. 1c, and BSCs were sampled
individually using a sterile trowel. To decrease spatial heterogeneity, each
BSC sample was taken from six individual plots (at least 20 m between two
adjacent plots) from each revegetation time. Therefore, we obtained 30 BSC
samples in total (5 cores
Microbial DNA was extracted from BSC samples using E.Z.N.A Soil DNA (Omega
Bio-tek, Norcross, GA, USA) according to the manufacturer's protocols. The
extracted DNA was diluted in TE buffer (10 mM Tris-HCl and 1 mM EDTA at
pH 8.0) and stored at
Purified amplicons were pooled in equimolar and paired-end sequenced (
qPCR was performed to determine the absolute 16S rRNA gene abundance. We used
the primer sets of 515F (5
Raw FASTQ files were demultiplexed and quality-filtered using QIIME
(version 1.17) with the following criteria: (i) the 300 bp reads were
truncated at any site receiving an average quality score
Rarefaction results of the 16S rDNA libraries based on 97 % similarity in different age of BSCs. MS, 5YR, 15YR, 28YR, 34YR and 51YR represent mobile sand, 5-, 15-, 28-, 34- and 51-year-old BSCs, respectively.
Hierarchical clustering analysis and PCA of bacterial communities in six different ages of BSCs at OTU level based on 97 % similarity (triplicate samples for each age). MS, 5YR, 15YR, 28YR, 34YR and 51YR represent mobile sand, 5-, 15-, 28-, 34- and 51-year-old BSCs, respectively.
Operational taxonomic units (OTUs) were clustered with 97 % similarity
cut-off using UPARSE (version 7.1
Illumina MiSeq sequencing was used to assess the bacterial community composition and diversity of BSCs in successional stages for revegetation in Shapotou. In total, 18 libraries of bacterial 16S rRNA were constructed, and at least 37 332 effective sequences in each sample were obtained, with an average length of 437 bp. A total of 1197–2307 OTUs were generated using a threshold of 0.97 (Table S1 in the Supplement); 394 OTUs were shared and occupied a relatively high proportion among all samples (17.07–32.92 %) (Table S2), and these OTUs accounted for 41.96–84.88 % of the total sequences (Table S2). This indicated a high coherence of community among these soil crusts. Alpha-diversity analysis revealed the microbial richness and diversity. Rarefaction curves showed that the most bacterial OTUs were found in 51YR crust, whereas MS contained the fewest. The number of OTUs was almost the same from 15YR to 51YR (Fig. 2). Community richness estimation using ACE and Chao revealed a similar trend to that for community diversity, which was further supported by Shannon's indexes (Table S1). Hierarchical clustering analysis (Fig. 3a) and PCA (Fig. 3b) showed that the triplicate samples of each age of BSCs were clustered, verifying that the sequencing results were reliable and the samples were reproducible.
In the bacterial community, a total of 28 phyla were retrieved at genetic
distances of 3 % and clustered into four groups according to their
relative abundance (Fig. 4). Of the total sequences, 4.48 % were not
classified at the phylum level. The percentages of major phyla for each age
of BSCs are shown in Fig. 5. The most abundant phylum shifted from
Firmicutes (72.8 %) in MS and 5YR to Actinobacteria in BSCs (minimum of
27.4 % in 15YR and maximum of 30.7 % in 51YR). The following major phyla
were at high abundance (
Percentages of the major classes in each age of BSCs. MS, 5YR, 15YR, 28YR, 34YR and 51YR represent mobile sand, 5-, 15-, 28-, 34- and 51-year-old BSCs, respectively.
Heat map of bacterial communities in different ages of BSCs at phylum level. MS, 5YR, 15YR, 28YR, 34YR and 51YR represent mobile sand, 5-, 15-, 28-, 34- and 51-year-old BSCs, respectively.
Abundant phyla (
At the class level (Table 1), 95.61 % of sequences were assigned, and
there was considerable consistency in dominant classes among the crusts.
Bacilli was the largest class in MS and 5YR with sequence percentages of
68.73 and 32.62 %, respectively; Actinobacteria was the predominant
class from 15YR to 51YR. In addition to subdivisions of Proteobacteria, other
major classes included Acidobacteria, Cyanobacteria, Chloroflexi, Clostridia,
Cytophagia, Deinococci, Gemmatimonadetes, Ktedonobacteria, Sphingobacteria
and Thermomicrobia. The percentages of high-abundance (
Bacterial community composition in six different ages of BSCs at the genus level. Data are defined at a 3 % OTU genetic distance. MS, 5YR, 15YR, 28YR, 34YR and 51YR represent mobile sand, 5-, 15-, 28-, 34- and 51-year-old BSCs, respectively.
At the family level, there were 133 identified families (data not shown), with the most abundant families being Bacillaceae, Enterococcaceae and Streptococcaceae (Table S3). Other dominant families were Geodermatophilaceae, JG34-KF-161, JG34-KF-361, Methylobacteriaceae, Micromonosporaceae, Bradyrhizobiaceae and Enterobacteriaceae.
A large proportion of sequences were not assigned to any genera. Even for
genera with relative abundance
The phylogenetic relationships of the 30 most abundant genera are shown in
Fig. 7. They clustered into three groups at the phylum level: Actinobacteria
formed one group and included 10 genera; another group was Firmicutes and
Proteobacteria; and Cyanobacteria, Chloroflexi and Deinococcus–Thermus formed
the third group. The genera
Phylogenetic relationship of the 30 most abundant genera in bacterial composition of BSCs.
Abundant species (
RDA (Fig. 9) and hierarchical clustering analysis (Fig. 3) were used to
discern the correlations between bacterial communities and soil
physicochemical properties. Taking into account the likely changes in the
soil properties from samples with the same successional stages at the same
experimental site, we selected soil biogeochemical data collected from 2005
in the RDA (data from Li et al., 2007a; Table S5). The BSC grouping patterns
of bacterial communities at the phylum and genus levels were similar to the
OTU level, with all divided into two groups. Group I contained two members,
MS and 5YR, which dominated the physical crusts and algal crusts (Fig. 1a and
b) and had the lowest diversities with Shannon indexes of 3.3 and 4.61, and
Simpson indexes of 0.139 and 0.0531, respectively (Table S1). The remaining
BSCs comprised the largest branch of group II, which dominated BSCs composed of
algae, lichens or mosses (Fig. 1c–f) and had higher diversity with Shannon
indexes
Redundancy analysis (RDA) of bacterial community structures in relation to soil physiochemical properties. Arrows indicate the direction and magnitude of soil physiochemical index associated with bacterial community structures. The lengths of arrows in the RDA plot correspond to the strength of the correlation between variables and community structure. Each circle represents the bacterial community structure for each sample.
Absolute abundances of bacteria (copies of ribosomal genes per gram
of soil) in BSCs quantified by qPCR (means
Means with different letters are significantly different (
From Fig. 9, it can be inferred that BSC development was associated with soil
physicochemical properties. The development of microbial community structure
was positively correlated with the physicochemical index except for soil bulk
density. The total variation in OTU data explained by the first four axes in
the RDA (as constrained by the measured environmental variables) was
82.16 %, with the first axis explaining 75.27 % and the second axis
explaining 4.42 %. Of all the environmental factors, silt
The averaged bacterial abundance in MS was 1.12
On a landscape scale and in high-stress environments, the role of diversity hot spots of BSC microbes is crucial to establishing stability and regulating moisture and nutrient cycling (Bowker, 2007). Additionally, bacteria are the conduits between the larger BSC organisms and plants, facilitating micro-processes (Castillo-Monroy et al., 2011). Thus, bacteria are key contributors to the BSC primary succession process and no doubt also in terms of secondary succession.
In the present study, we gained information concerning the diversity of bacterial communities in BSCs of different ages in restored vegetation in Shapotou in the Tengger Desert. The 16S rRNA gene-based amplicon survey revealed the dominance of Actinobacteria, Proteobacteria, Chloroflexi, Acidobacteria and Cyanobacteria in all BSCs, with Firmicutes dominating MS (72.8 %) and decreasing to 3.05 % in 51YR, and Actinobacteria increasing from 15YR (27.4 %) to 51YR (30.7 %). Due to different arid conditions, comparisons with other studies of BSCs should be viewed with caution. Cyanobacteria, Actinobacteria, Proteobacteria and Acidobacteria are ubiquitous in soils and sediments everywhere, in arid as well as wet landscapes (Fierer et al., 2012), and Proteobacteria are very common and diverse among all BSCs. We observed that Actinobacteria were the most abundant phylum in the developing (15YR, 28YR and 34YR) and relatively developed (51YR) BSCs, similar to BSCs from the Colorado Plateau and the Sonoran Desert, where Actinobacteria were dominant (Gundlapally and Garcia-Pichel, 2006; Nagy et al., 2005; Steven et al., 2013). Actinobacteria and Proteobacteria are usually predicted to be copiotrophic groups which increase in high-C environments (Fierer et al., 2007). These results differ from those reported in BSCs from Oman and the Gurbantunggut Desert (Abed et al., 2010; Moquin et al., 2012; Zhang et al., 2016), and even from BSCs of natural vegetation at the edge of the Tengger Desert (Wang et al., 2015), where Proteobacteria were the most abundant phylum, followed by Cyanobacteria, Actinobacteria and Chloroflexi. Unexpectedly, Cyanobacteria had a high proportion in the developed BSCs, although they were prevalent in early successional stages of BSCs (5YR) and play crucial roles in initial crust development (Belnap and Lange, 2001). This is relatively similar to that in the natural habitat around the Tengger Desert, where Cyanobacteria (19.5 %) and Actinobacteria (19.4 %) were the most dominant phyla after Proteobacteria (25.0 %). Moreover, the results did not resemble those from arid Arizona soils (Dunbar et al., 1999) or the Gurbantunggut Desert (Zhang et al., 2016) due to the high proportion of Chloroflexi, an unexplained presence of thermophilic phyla (Gundlapally and Garcia-Pichel, 2006; Moquin et al., 2012; Nagy et al., 2005) that display good adaptation to drought conditions and the important roles in the development of BSCs in arid zones (Lacap et al., 2011; Wang et al., 2015).
More recent information about BSC bacteria has been reported with the
convenience of culture-independent sequencing methods, and studies of their
function and classification in BSCs are increasingly detailed. The main
function of these dominant bacteria involves the cycling and storage of C and
N in desert ecosystems, which is vital to the functioning of arid land (Weber et
al., 2016). Firmicutes are more frequently detected in below-biocrust soils
(1–2 cm depth) (Elliott et al., 2014) and dominated in MS and 5YR, with the
vast majority of abundant species being in Firmicutes in the Tengger Desert.
Cyanobacteria are the main contributors to C and N fixation in soils during
successional processes of BSCs (Belnap and Gardner, 1993). They are thought to
serve as pioneers in the stabilization process of soils (Garcia-Pichel and
Wojciechowski, 2009), of which the genus
Owing to limited culture collections and curated sequence databases of BSC
bacteria, most non-cyanobacterial sequences from DNA-based bacterial surveys
cannot be reliably named or taxonomically defined, especially in relatively
abundant genera in Actinobacteria and Proteobacteria, such as
PCA and RDA showed that bacterial community compositions of MS and 5YR significantly differed from those of BSCs of more than 15 years in age and were positively correlated with soil physicochemical properties. Combined with the results of alpha-diversity analysis and qPCR, this means that the species richness and abundance reached their highest levels at 15 years of BSC development and then maintained similar levels thereafter. Similar trends were found in recovery of soil properties and processes after sand binding at five different-aged revegetated sites – proportions of silt and clay, and organic C increased with years since revegetation (Li et al., 2007a, b). The annual recovery rates of soil properties were greater at the initial revegetated sites (0–14 years) than at the old revegetated sites (43–50 years) (Li et al., 2007a). These results suggest that bacterial communities of BSCs recovered quickly in the fastest recovery phase of soil properties (the initial 15 years), and the bacterial biomass increased with the improvement of soil texture and nutrients, especially silt, clay and total K content in the Tengger Desert. A significant positive correlation was found between silt and clay and the number of BSC types in southern Africa (Büdel et al., 2009), suggesting that fine grain size promotes BSC succession and their biomass content. This may be attributed to the diversity of BSCs, vegetation composition, soil temperature and soil moisture, because these are key factors regulating soil microbial composition and activity (Butenschoen et al., 2011; De Deyn et al., 2009; Sardans et al., 2008), soil nutrient uptake and release (Peterjohn et al., 1994; Rustad et al., 2001), especially in the BSCs of top soil. It would be good to understand more of the factors that together influenced the composition and function of BSC bacteria in long-term revegetation, including BSCs, plants, soil biochemical properties and climate conditions, and the microorganisms that in turn have the positive influence on soil improvement (Li et al., 2007b, 2010).
Many reports have interpreted correlations among soil properties and BSCs as an indicator that BSCs are drivers of soil fertility and development (Chamizo et al., 2012; Delgado-Baquerizo, 2013; Yu et al., 2014; Zhang et al., 2010); some have reported the opposite and suggest a direct influence of soil properties on BSC development (Belnap et al., 2014; Bowker et al., 2006a; Bowker and Belnap, 2008; Concostrina-Zubiri et al., 2013; Rivera-Aquilar et al., 2009; Root and McCune, 2012; Weber et al., 2016). These are important questions, and parsing out the interactions of BSCs and soil biogeochemical properties remains an important frontier in BSC research. However, further work to identify controlled experimental approaches is required because field correlations do not explain the directionality of causality over time.
In temperate desert regions, BSCs are not well investigated regarding community structure and diversity. Furthermore, studies on succession are rare (Langhans et al., 2009). Most evidence indicates that BSC facilitate succession to later series, suggesting that assisted recovery of BSCs could speed up succession (Bowker, 2007). Because BSCs are ecosystem engineers in high-abiotic-stress systems, loss of BSCs may be synonymous with crossing degradation thresholds. Whether BSCs are deemed facilitative or inhibitory for later successional vegetation may depend on how exhaustively the interaction between plants and BSCs is investigated. In fixed-sand areas, BSCs may in some cases reduce infiltration (inhibitory effect) (Mitchell et al., 1998), but they also increase soil stability and serve as an N source for surviving and recolonizing trees (facilitative effects) (Tateno et al., 2003; Uchida et al., 2000). The BSC bacterial communities in the successional stages may help establish stability and regulate nutrient and biogeochemical cycling. Castillo-Monroy et al. (2011) found that the BSC richness matrix had the greatest direct effect on the ecosystem function matrix. Despite this result, very few of the BSC effects on ecosystem function could be ascribed to changes within the bacterial community. This provides valuable insights concerning semiarid ecosystems where plant cover is spatially discontinuous and ecosystem function in plant interspaces is regulated largely by BSCs.
Illumina MiSeq sequencing showed that changes of BSC bacterial diversity and richness in BSC succession were consistent with the recovery phase of soil properties in vegetation succession of Shapotou in the Tengger Desert. The shift of bacterial community composition in BSCs at all levels of classification was related to their corresponding function in the BSC recovery process. BSC bacteria are crucial to establishing stability and nutrient cycling in desert ecosystem, and they are the conduits between the larger BSC organisms and plants facilitating micro-processes. These results have confirmed that bacteria are key contributors to the BSC succession process.
Raw data for Illumina MiSeq sequencing of 18 samples were
deposited in the NCBI Sequence Read Archive database
(
LL and YL designed the research. PZ, GS and RH collected samples from the field. YL and JW performed DNA extraction and quality detection. YL analyzed the high-throughput data and prepared the manuscript with consistent contributions from LL. ZW analyzed the soil biogeochemical data and made the RDA figure.
The authors declare that they have no conflict of interest.
This article is part of the special issue “Biological soil crusts and their role in biogeochemical processes and cycling”. It is not associated with a conference.
This work was financially supported by the Creative Research Group Program of the National Natural Science Foundation of China (grant no. 41621001) and the National Natural Science Foundation of China (grant nos. 41371100 and 41401112). Edited by: Anita Antoninka Reviewed by: two anonymous referees