A peer-reviewed version of this preprint was published in PeerJ on 14August 2019.

View the peer-reviewed version (peerj.com/articles/7356), which is thepreferred citable publication unless you specifically need to cite this preprint.

Bai Y, Huang X, Zhou X, Xiang Q, Zhao K, Yu X, Chen Q, Jiang H, Nyima T, GaoX, Gu Y. 2019. Variation in denitrifying bacterial communities along a primarysuccession in the Hailuogou Glacier retreat area, China. PeerJ 7:e7356https://doi.org/10.7717/peerj.7356

Variation in the nitrous oxide reductase gene (nosZ)-

denitrifying bacterial community in different primary

succession stages in the Hailuogou Glacier retreat area, China

Yan Bai 1 , Xiying Huang 1 , Xiangrui Zhou 1 , Quanju Xiang 1 , Ke Zhao 1 , Xiumei Yu 1 , Qiang Chen 1 , Hao

Jiang 2 , Nyima Tashi 3 , Xue Gao 3 , Yunfu Gu Corresp. 1

1 Department of Microbiology/ College of Resources/Sichuan Agricultural University, Sichuan Agricultural University, Chengdu, Sichuan, China

2 Department Institute of Chengdu Mountain Hazards and Environment/Chinese Academy Sciences, Chinese Academy Sciences, Chengdu, Sichuan, China

3 Institute of Agricultural Resources and Environmental Science/Tibet Academy of Agricultural and Animal Husbandry Sciences, Tibet Academy of

Agricultural and Animal Husbandry Sciences, Lhasa, Tibet, China

Corresponding Author: Yunfu Gu

Email address: guyf@sicau.edu.cn

Background: The Hailuogou Glacier in the Gongga Mountain region (SW China), on the southeastern

edge of the Tibetan Plateau, is well known for its low-elevation modern glaciers. Since the end of the

Little Ice Age (LIA), the Hailuogou Glacier has retreated continuously due to global warming, primary

vegetation succession and soil chronosequence have developed in this retreat area. The retreated area

of Hailuogou Glacier has not been strongly disturbed by human activities, thus it is an ideal models for

exploring the biological colonization of nitrogen in the primary successional stages of ecosystem. The

nosZ gene encodes the catalytic center of nitrous oxide reductase and is an ideal molecular marker in

studying the variation in the denitrifying bacterial community.

Methods: Soil properties as well as abundance and composition of the denitrifying bacterial community

were determined via chemical analysis, quantitative polymerase chain reaction (qPCR), and terminal

restriction fragment length polymorphism (T-RFLP), respectively. The relationships between the nosZ

denitrifying bacterial community and soil properties were determined using redundancy analysis (RDA).

Soil properties, potential denitrify activity (PDA), and the nitrous oxide reductase gene (nosZ)-denitrifying

bacterial communities significantly differed among successional stages.

Results: Soil properties, potential denitrify activity (PDA), and the nitrous oxide reductase gene (nosZ)-

denitrifying bacterial communities significantly differed among successional stages. Soil pH in the topsoil

decreased from 8.42 to 7.19 in the course of primary succession, while soil organic carbon (SOC) and

total nitrogen (TN) gradually increased with primary succession. Available phosphorus (AP) and available

potassium (AK), as well as potential denitrify activity (PDA), increased gradually and peaked at the 40-

year-old site. The abundance of the nosZ denitrifying bacterial community followed a similar trend. The

variation in the denitrifying community composition was complex; Mesorhizobium dominated the soil in

the early successional stages (0-20 years) and in the mature phase (60 years), with a relative abundance

greater than 55%. Brachybacterium was increased in the 40-year-old site, with a relative abundance of

62.74%, while Azospirillum dominated the early successional stages (0-20 years). Redundancy analysis

(RDA) showed that the nosZ denitrifying bacterial community correlated with soil available phosphorus

and available potassium levels (P < 0.01).

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.27481v1 | CC BY 4.0 Open Access | rec: 12 Jan 2019, publ: 12 Jan 2019

1 Variation in the nitrous oxide reductase gene (nosZ)-

2 denitrifying bacterial community in different primary

3 succession stages in the Hailuogou Glacier retreat

4 area, China5

6 Yan Bai1†, Xiying Huang1†, Xiangrui Zhou1†, Quanju Xiang1, Ke Zhao1, Xiumei Yu1, Qiang

7 Chen1, Hao Jiang2, Nyima Tashi3, Xue Gao3, Yunfu Gu1*

8

9 1 Department of Microbiology, College of Resources, Sichuan Agricultural University, Chengdu

10 611130, China

11 2 Institute of Chengdu Mountain Hazards and Environment. Chinese Academy Sciences,

12 Chengdu 610041, China

13 3 Institute of Agricultural Resources and Environmental Science, Tibet Academy of Agricultural

14 and Animal Husbandry Sciences, Lhasa 850002, China

15

16 † Equal contributor

17 *Corresponding author

18 E-mail address: guyf@sicau.edu.cn

19 Abstract: 20 Background: The Hailuogou Glacier in the Gongga Mountain region (SW China), on the

21 southeastern edge of the Tibetan Plateau, is well known for its low-elevation modern glaciers.

22 Since the end of the Little Ice Age (LIA), the Hailuogou Glacier has retreated continuously due

23 to global warming, primary vegetation succession and soil chronosequence have developed in

24 this retreat area. The retreated area of Hailuogou Glacier has not been strongly disturbed by

25 human activities, thus it is an ideal models for exploring the biological colonization of nitrogen

26 in the primary successional stages of ecosystem. The nosZ gene encodes the catalytic center of

27 nitrous oxide reductase and is an ideal molecular marker in studying the variation in the

28 denitrifying bacterial community.

29 Methods: Soil properties as well as abundance and composition of the denitrifying bacterial

30 community were determined via chemical analysis, quantitative polymerase chain reaction

31 (qPCR), and terminal restriction fragment length polymorphism (T-RFLP), respectively. The

32 relationships between the nosZ denitrifying bacterial community and soil properties were

33 determined using redundancy analysis (RDA). Soil properties, potential denitrify activity (PDA),

34 and the nitrous oxide reductase gene (nosZ)-denitrifying bacterial communities significantly

35 differed among successional stages.

36 Results:Soil properties, potential denitrify activity (PDA), and the nitrous oxide reductase gene

37 (nosZ)-denitrifying bacterial communities significantly differed among successional stages. Soil

38 pH in the topsoil decreased from 8.42 to 7.19 in the course of primary succession, while soil

39 organic carbon (SOC) and total nitrogen (TN) gradually increased with primary succession.

40 Available phosphorus (AP) and available potassium (AK), as well as potential denitrify activity

41 (PDA), increased gradually and peaked at the 40-year-old site. The abundance of the nosZ

42 denitrifying bacterial community followed a similar trend. The variation in the denitrifying

43 community composition was complex; Mesorhizobium dominated the soil in the early

44 successional stages (0-20 years) and in the mature phase (60 years), with a relative abundance

45 greater than 55%. Brachybacterium was increased in the 40-year-old site, with a relative

46 abundance of 62.74%, while Azospirillum dominated the early successional stages (0-20 years).

47 Redundancy analysis (RDA) showed that the nosZ denitrifying bacterial community correlated

48 with soil available phosphorus and available potassium levels (P < 0.01).

49 Keywords: Hailuogou glacier retreat area; primary succession; nosZ denitrifying bacterial

50 community; T-RFLP; qPCR

51 Introduction52 Primary successional ecosystems, such as glacier forelands and volcanoes, are ideal models

53 for exploring the biological colonization of various substrates. Since the ice covers of many

54 glaciers have receded over the past century, glacier forelands have released substrates for soil

55 formation and development (Rangwala et al., 2012; Anesio et al., 2012). In this process, new

56 bare lands have emerged because of glacial retreats, starting ecological succession towards a

57 zonal ecosystem. Autotrophic microbes play a critical role in the initial stages of primary

58 community assembly (Cavicchioli et al., 2002). However, heterotrophic colonizers, decomposing

59 organic material, are also important in the initial establishment of functional communities

60 (Hodkinson et al., 2003). Previous studies in this field have mainly focused on the diversity and

61 composition of the soil microbial community in the primary succession of receding glaciers

62 (Ambrosini et al., 2017; Fernández-Martínez et al., 2017). Only a few studies have employed

63 molecular tools to understand the diversity and composition of the functional microbial

64 community along the forelands of receding glaciers (Kandeler et al., 2006; Töwe et al., 2010).

65 For example, in a study about the activity and composition of the denitrifying bacterial

66 community at the Rotmoosferner Glacier, Austria, the successional age shaped the N cycling

67 processes as well as the microbial community composition (Nicol et al., 2010).

68 The Gongga Mountain is a typical mountain glacier in southwest China and extremely

69 sensitive to climate change. The Hailuogou Glacier, located on the Gongga Mountain, has been

70 retreating continuously since the end of the Little Ice Age (LIA) (Li et al., 2018) with the

71 development of a soil and primary vegetation succession. Previous studies in this area have

72 investigated specific processes such as pedogenesis (He et al., 2008; Zhou et al., 2013), plant

73 succession (Yang et al., 2014), and microbial community changes (Sun et al., 2016). However, to

74 date, there is no information on the density of the functional community involved in the

75 denitrifying process in the primary succession of the Hailuogou Glacier.

76 Studying the dynamics of nitrogen cycling in the glacier retreat area is critical to evaluate

77 the responses and feedbacks of biogeochemistry cycles to global climate change. Inorganic

78 nitrogen compounds are primarily transformed through biochemical reactions mediated by

79 microbial communities, and their activities regulate the input or output of nitrogen from

80 ecosystems (Hallin et al., 2018). Nitrous dioxide (N2O) is a potent greenhouse gas and generated

81 via the microbially mediated processes nitrification and denitrification. Its impact on global

82 warming is about 300 times higher than that of CO2, and it´s half-life in the atmosphere is about

83 120 years; most atmospheric N2O is the result of soil processes (Solomon et al., 2007; Wuebbles,

84 2009). Increasing denitrification is the main pathway to reduce N2O production, and this process

85 can be mediated by bacteria, archaea, or fungi. More than 50 genera of bacteria have been

86 identified as denitrifying bacteria; hence, given the high phylogenetic diversity among

87 denitrifying bacteria, detection methods based on gene assessment, such as 16S rRNA, are not

88 adequate and require the use of functional genes which encode for enzymes directly involved in

89 the denitrification process (Gu et al., 2017).

90 The enzyme nitrous oxide reductase catalyzes the final step of the denitrification process,

91 namely the conversion of nitrous oxide (N2O) to nitrogen (N2), and the nosZ encodes the

92 catalytic center of nitrous oxide reductase; therefore, it is generally used as a key functional gene

93 of molecular markers to detect whether denitrification is complete (Philippot, 2002). Wang et al.

94 (2012) have shown that in the case of nosZ gene abundance, N2O reduction both in the surface

95 soils of the land area and in the soil core of the transition site was the dominant process.

96 According to a previous study (Brankatschk et al., 2011), nosZ gene copy numbers were lowest

97 in the early successional stages of a glacier retreat and significantly correlated with the potential

98 enzyme activities, but not with the abundances of nirK and nirS. In another study (Kandeler et

99 al., 2006), the nosZ gene was an ideal molecular marker in studying the variation in the

100 denitrifying bacterial community throughout primary succession in a glacier retreat. We

101 therefore infer that nosZ gene abundance and the dynamics of the denitrifying bacterial

102 community of glacier retreat areas are important factors to understand the ecological role of

103 bacteria involved in denitrification processes and, subsequently, in soil development

104 (Brankatschk et al., 2011; Kandeler et al., 2006).

105 In this study, in combination with the measurement of soil properties, potential

106 denitrification activities as well as the abundance and structural diversity of nosZ denitrifying

107 bacteria in four successional stages (time of exposure after ice melting of 0, 20, 40, and 60 years,

108 respectively) were assessed by chemical analysis, quantitative polymerase chain reaction (qPCR),

109 and terminal restriction fragment length polymorphism (T-RFLP), respectively. The objectives

110 of this study were (i) to investigate the variation in the nosZ denitrifying bacterial community in

111 the four different successional stages found in the Hailuogou Glaicer retreat and (ii) to

112 understand the relationships between the nosZ denitrifying bacterial community and soil

113 properties. We hypothesized that the variation in the activity and composition of the nosZ

114 denitrifying bacterial community is strongly correlated with the successional stages.

115 Material and methods116 Study area

117 The Gongga Mountain is located on the southeastern edge of the Tibetan Plateau, with a

118 summit elevation of 7,556 m above sea level (ASL) (Liu et al., 2010). The Hailuogou Glacier

119 (29° 34′ 07.83′′ N, 101° 59′ 40.74′′ E, Sichuan Province, China) is the single most developed

120 glacier on the eastern slope of the Gongga Mountain. Against the background of a warming

121 climate, the Hailuogou Glacier is shrinking continuously and is melting more rapidly now. This

122 glacier retreat area currently covers an elevation range of 2,855 to 2,982 m above sea level.

123 Average annual temperature is 4.2°C, with an average annual precipitation of 1,947 mm. The

124 retreat area of the Hailuogou Glacier provides an excellent opportunity to study the relationship

125 between vegetation succession and soil development, as its relatively mild and humid climate

126 allows for rapid moraine colonization by plants and promotes fast ecosystem development.

127 Along the approximately 2-km-long belt, there is a complete primary succession series from bare

128 land to climax vegetation communities (Mapelli et al., 2018; Jiang et al., 2018). In the area with

129 time of exposure after ice melting of about 60 years, a community resembling a mature phase

130 community has established, dominated by Abies fabri (Zhou et al., 2018). Primary succession

131 was marked by the following stages: (i) bare land for 8-15 years, (ii) Willow-Hippophae

132 rhamnoides-Populus purdomii, with a duration of about 35 years, (iii) Populus purdomii-

133 Betulautilis-Abies fabri, about 60 years after glacier retreat, (iv) Populus purdomii-Betulautilis-

134 Abies fabri community was gradually replaced by the Abies fabri community, with high biomass

135 productivity of the plant communityy, (v) Picea brachytyla, Abies fabri, mature phase species

136 (Zhou et al., 2018). The soil parent material is mainly composed of biotite schist, granodiorite,

137 and quartzite, with small amounts of phyllite, slate, and chlorite schist (Zhou et al., 2013). Table

138 1 provides information about the different sampling points.

139 Soil sampling

140 Soil samples were collected from four sites, representing four successional stages (0, 20, 40,

141 and 60 years) (Fig. 1). Site T0 is a bare land and contain with gravel, without any vegetation;

142 successional time is below 5 years. Site T1 represents a bare-dwarf vegetation stage with the

143 pioneer species Astragalus Membranaceus, Hippophae rhamnoides and willow; successional

144 time is about 20 years, and the horizontal distance from the end of the glacier is about 300 m.

145 Site T2 represents the stage of the Willow-Hippophae rhamnoides-Populus purdomii community,

146 and the pioneer species are Willow, Hippophae rhamnoides and Populus purdomii; successional

147 time is about 40 years, and the horizontal distance from the end of the glacier is about 600 m.

148 Site T3 represents the stage of the Populus purdomii-Betulautilis-Abies fabri community, where

149 the dominant plant species are Populus purdomii and Betulautilis, and some pine species (Abies

150 fabri) begin to appear. Succession age is about 60 years, and the horizontal distance from the end

151 of the glacier is about 1,200 m (Table 1). To avoid erosion and deposition effects, all sampling

152 plots were located in the middle of gentle slopes, far from the small streams inside the primary

153 succession areas. Moreover, all plots (except those in the T0 site) were located under the

154 canopies of the dominant plant species.

155 For each successional stage, five sampling plots of at least 20 × 20 m2, with a minimum

156 distance of 100 m between plots, were selected (Fig. 1). From each plot, we collected four soil

157 samples, using a 6.5-cm diameter steel auger, from the surface layer (0-20 cm, roots and residue

158 were removed and the samples were mixed to obtain one composite sample for each plot.

159 Subsequently, the samples were stored in plastic bags on ice and divided into two portions: one

160 was air-dried and sieved through a 2-mm mesh screen for physicochemical analyses, while the

161 other one was stored at -80°C for soil total DNA extraction.

162 Soil properties and denitrifying enzymatic activity analysis

163 Soil pH was measured potentiomectrically. Soil total nitrogen was measured using the semi-

164 micro Kjeldahl method, while soil organic carbon was determined via the potassium dichromate

165 oxidation-external heating method. Soil available phosphorus (AP) and available potassium (AK)

166 were assessed via the sodium bicarbonate extraction-Mo-Sb colorimetry methods and

167 ammonium acetate extraction-flame photometry, respectively (Bao, 2007). Soil potential

168 denitrifying activity (PDA) was determined using the C2H2 inhibition method (Dambreville et al.,

169 2006).

170 Total soil DNA extraction

171 Total soil DNA was extracted from 0.5 g of fresh soil using a FastDNA SPIN Kit for Soil

172 (MP Biomedicals, Solon, OH, USA) according to the manufacturer’s instructions. The

173 concentration and quality of extracted DNA were measured with a Nano-200 spectrophotometer

174 (Aosheng, Hangzhou, China); DNA samples were stored at -20°C for further analysis.

175 Quantitative PCR

176 Quantitative PCR (qPCR) of the nosZ gene was carried out using the ABI7500 sequence

177 detection system (Bio-Rad) in 25 μL of the reaction mixture containing 12.5 μL of ABI Power

178 SybrGreen qPCR Master Mix (Applied Biosystems, USA), 1 μM of each primer, 1.25 μL of

179 template DNA, adjusted to a concentration of 20 ng μL−1; sterile distilled water was used to bring

180 the final volume up to 25 μL. The nosZ gene was amplified with primer pairs used for qPCR of

181 nosZ gene were nosZ2F (5’-CGYTGTTCMTCGACAGCCAG-3’) and nosZ2R (5’-

182 CGSACCTTSTTGCCSTYGCG-3’) (Henry et al., 2006). For nosZ, the qPCR program consisted

183 of an initial denaturation step at 95°C for 10 min, 40 cycles of 95°C for 15 s, 60°C for 60 s, and

184 72°C for 30 s. The nosZ gene was quantified via qPCR with six technical replicates per sample.

185 Standard curves to assess nosZ gene abundance consisted of a 10-fold serial dilution of a plasmid

186 containing the nosZ gene fragments.

187 The PCR amplification and T-RFLP of the nosZ gene

188 Terminal restriction fragment length polymorphism (T-RFLP) analysis of the nosZ gene

189 was carried out using the primer pair nosZ1126F (5’-GGICTBGGICCRYTGCAYAC-3’) and

190 nosZ1884R (5’-CATYTCSAKR- TGCAKGGCRTG-3’); the forward primer was labeled with 6-

191 carboxyfluoroscein (FAM) (Chen et al., 2012). The PCR reaction mixture volume was 50 μL, 192 which consisted of 60 ng of template DNA, 0.4 μM of each primer, 25 μL of PCR master mix

193 (Tiangen, China), and sterile distilled water to obtain a final volume of 50 μL. The PCR was

194 performed with the following thermal profile: 95°C for 3 min, followed by 30 cycles of 95°C for

195 30 s, 58°C for 1 min, and 72°C for 1 min, with a final extension step at 72°C for 5 min. The PCR

196 products were checked via 1.0% agarose gels electrophoresis and visualized with a UV

197 Transilluminator Model M-26 (UVP, USA) after ethidium bromide staining. The amplified nosZ

198 gene fragments were digested with BstuI and HhaI endonucleases (NEB, Ipswich, Massachusetts,

199 USA) at 37°C for 6 h, followed by further purification with a Tiangen DNA purification kit

200 (Tiangen, China). The T-RFLP profiles were then generated by capillary electrophoresis using an

201 ABI Prism 3100 Genetic Analyzer at Sangong Corporation (Shanghai, China).

202 Statistical analysis

203 The means and standard deviations of the soil physicochemical parameters and nosZ gene

204 diversities were calculated and tested for normality and homogeneity of variances. One-way

205 analysis of variance (ANOVA) was applied to test for differences between sites, followed by

206 Tukey’s multiple comparison tests. In the two-way tests, the significance level was p ≤ 0.05.

207 Pearson’s correlation analysis was conducted to examine the relationships between soil

208 properties and nosZ gene diversity, using the software package SPSS 21.0. Analysis of the T-

209 RFLP profiles was performed using the software package Peak Scanner version 1.0 (Applied

210 Biosystems, Inc.). The peak heights of terminal restriction fragments (T-RFs) with size

211 differences ≤ 2 bp in an individual profile were combined and considered to be one fragment.

212 The T-RFs with a relative abundance < 1% were excluded from further analysis (Zhang et al.,

213 2017). Principal components analysis (PCA) was performed to compare the dissimilarities

214 among the denitrifying communities across different successional stages, based on Bray-Curtis

215 dissimilarity. Correlation best-of-fit model for nosZ gene abundance and soil properties was

216 carried out by curve estimation in SPSS. Redundancy analysis (RDA) was performed to assess

217 the relationships between the nosZ denitrifying bacterial community composition and the soil

218 properties, using the software package CANOCO 5.0.

219 Results220 Soil properties and denitrifying activities in different successional stages

221 Based on our results, the soil nutrient status changed substantially with primary succession

222 (Table 2). Soil pH gradually decreased from the T0 (8.42) to the T3 site (7.19). Soil organic

223 carbon (SOC) was significantly higher in the T3 site (30.77 g·kg-1) than in the other sites. Soil

224 total nitrogen (TN) ranged from 0.29 to 1.11 g kg -1. The siteT3 site had the highest

225 concentrations of soil total nitrogen (TN), while the lowest level was found in the site T0. Soil

226 available phosphorus (AP) varied from 0.98 to 5.54 mg kg -1, with the highest value observed in

227 the T2 site and the lowest in the T0 site. Soil available potassium (AK) ranged from 13.60 to

228 120.80 mg kg-1; the highest concentration was observed in the T2 site and the lowest in the T0

229 site. Potential denitrifying activities ranged from 0.08 to 0.31 μg N2O-N g -1 dry soil h-1.

230 Principal components analysis (PCA) was based on the analyzed soil properties of the four

231 different successional stages. The PCA ordinated samples along PCA1 (89.51% of variance) and

232 PCA2 (7.74% of variance), based on the successional stages. The soil micro-environment of the

233 sites T0 and T1 was similar, with significant differences compared to the other stages (Fig. 1).

234 The nosZ denitrifying bacterial community abundance and diversity in different

235 successional stages

236 Soil nosZ gene copy numbers ranged from 4. 47 log10 per gram dry soil in T0 to 5.86 log10

237 per gram dry soil in T2, showing that the nosZ denitrifying bacterial community abundance

238 changed significantly throughout primary succession (Fig. 2). The nosZ gene abundance was

239 negatively correlated with soil pH and positively correlated with available soil potassium (P <

240 0.05) (Fig. 3).

241 The diversity of the nosZ denitrifying bacterial community varied significantly among the

242 four different successional stages (P < 0.05) (Table 3). Shannon-Weiner index (H), richness (S),

243 and evenness (Eh) ranged from 1.82 to 2.54, 8.99 to 18.41, and 1.03 to 1.30, respectively. The

244 T2 site had the highest Shannon-Weiner index (H) and richness (S) values, while the lowest

245 values were found in T1. Evenness (Eh) increased gradually throughout primary succession, with

246 the highest values observed in the T3 site.

247 Pearson´s correlation analysis revealed that Shannon-Weiner index (H) was significantly

248 linked to soil AP (P < 0.01) and AK (P < 0.05), while richness (S) was positively correlated with

249 AP, AK, and PDA (P < 0.05) and evenness(Eh) was negatively correlated with soil pH (P < 0.01)

250 (Table 4). the abundance of nosZ gene was significantly correlated with PDA (P <0.01) richness

251 (S) (P <0.01)and AK (P <0.05).

252 The nosZ denitrifying bacterial community composition in different successional

253 stages

254 In the nosZ gene T-RFLP analysis, the dominant T-RFs were not significantly affected by

255 the composition of the nosZ denitrifying bacterial community; however, some minor T-RFs were

256 impacted (Fig. 3). In the T0 stage, 20- and 25-bp T-RFs were dominant nosZ T-RFs, with

257 relative abundances of 62.6 and 10.5%, respectively. At T1, 15-, 20-, and 25-bp T-RFs

258 dominated, with relative abundances of 22.8, 43.4, and 10.6%, respectively. In the site T2, 10-,

259 15-, and 20-bp T-RFs were most abundant, with relative abundances of 10.7, 41.7, and 24.7%,

260 respectively. In the site T3, the dominant nosZ T-RFs had 15 and 20 bps, with relative

261 abundances of 55.2 and 12.3%, respectively. Across all successional stages, the 40-bp T-RFs was

262 present, with a relative abundance of 2.2-2.6%. All sites also showed specific minor nosZ T-RFs.

263 For example, the 50-bp T-RFs were found in the sites T0, T1, and T2, and their relative

264 abundances gradually decreased over time. The 45- and 70-bp T-RFs were found at the site T1,

265 while 55- and 95-bp T-FRs were present in the successional stages T2 and T3. At the sites T0

266 and T1, the 90-bp T-RFs were dominant. Redundancy analysis (RDA) showed that denitrifying

267 bacterial communities in the different successional stages formed separate clusters, expect for the

268 sites T0 and T1 (Fig. 4).

269 In the in-silico T-RFLP analysis of the nosZ sequences obtained in this study and from the

270 NCBI database, the nosZ-denitrifying bacterial community in the clone library was generally

271 similar to those detected in the actual T-RFLP. The dominant 15-bp T-RF was consistent with

272 the T-RFs of Brachybacterium, the 20-bp T-RF was consistent with Mesorhizobium, and the T-

273 RFs with 25, 40, 50, and 55 bp were consistent with Azospirillum, Sulfuritalea, Pseudomonas,

274 and Nitrospirillum, respectively (Fig. 3).

275 Redundancy analysis of nosZ denitrifying bacterial community composition and

276 soil properties

277 The relationships between soil properties and nosZ denitrifying bacterial community were

278 analyzed via redundancy analysis (RDA). Based on the RDA pattern, the nosZ denitrifying

279 bacterial communities from the soils of the four different successional stages could be partitioned

280 into four separate groups (Fig. 4), indicating that the nosZ denitrifying bacterial communities

281 from these four different soils related to primary succession were significantly different (P <

282 0.05). The RDA results further confirmed that the nosZ denitrifying bacterial community

283 composition in the young soils (0-20 years) differed from that in the older sites (40-60 years).

284 Among the measured soil properties, soil AP and AK appeared to be the major factors in

285 determining the nosZ denitrifying bacterial community composition during primary succession

286 of the Hailuogou Glacier (P < 0.01). Soil nosZ denitrifying bacterial community variance was

287 also significantly linked to soil pH and PDA (P < 0.05).

288 Discussion289 Soil properties at different successional stages

290 Soil properties reflect the primary succession of the glacier retreat areas, with significant

291 consequences for soil formation, soil microbial community development, and plant growth

292 (Rangwala et al., 2012; Ambrosini et al., 2017). In the course of primary succession, soil is

293 generally considered as a nutrient pool for plant an microbial growth (Roubíčková et al., 2009).

294 In this sense, soil development and soil microbial community dynamics affect the species

295 replacement during primary succession, with feedback mechanisms for plants (Frouz et al.,

296 2016). On the other hand, plants directly affect soil properties and have an indirect effect via

297 modifying the soil microbial community, which in turn may affect soil formation (Frouz et al.,

298 2008). Liter decomposition redistributes organic matter, increases soil porosity, and enhances

299 soil aggregate formation, which greatly alters soil sorption capacity, water holding capacity, and

300 other soil properties (Ponge et al., 2003; Six et al., 2004). Roots and associated soil organisms

301 also affect the formation and development of soil, which in turn affects subsequent generations

302 of plants and other soil organisms, a phenomenon referred to as “plant-soil feedback” (Six et al.,

303 2004). In this study, SOC, TN, AP, and AK increased in the course of primary succession.

304 During primary succession, pioneer plants such as herbs initially colonize the bare land, followed

305 by shrubs; such processes facilitate colonization by soil microorganisms and fauna, ultimately

306 resulting in altered soil properties (Frouz et al., 2008). Vegetation growth leads to an increase in

307 litter residue and root exudate secretion, contributing to soil nutrient cycling (Cotrufo et al.,2013;

308 Schmidt et al., 2011) and increased soil nutrient levels. However, later-successional species

309 require higher amounts of nutrients (Püschel et al., 2007), leading to decreased soil nutrient

310 levels. In addition, loss through runoff may also lead to a decrease in soil nutrients. Soil pH is an

311 important indicator reflecting soil quality, and its value indicates the degree of acidity and

312 alkalinity of the soil and its effect on the survival of plants and animals (Li et al., 2017). In this

313 study, soil pH decreased in the course of succession, which was not in line with some previous

314 studies that found no correlation between soil pH and the distance from the glacier (Zeng et al.,

315 2016; Strauss et al., 2009).

316 Potential denitrifying activity in different successional stages

317 Potential denitrifying activity is significantly affected by a variety of environmental factors

318 such as plant species richness, soil moisture, soil carbon, and soil nitrogen (Leloup et al., 2018).

319 Zeng et al. have indicated that soil carbon is the key factor affecting soil potential denitrifying

320 activity. Besides, soil pH, available phosphorus, and available potassium also significantly

321 influenced soil activity such as urease, deaminase, and protease activities; these soil enzyme

322 activities were grouped together for newly exposed soils on young moraines (0-5 years), while

323 older soils (17-44 years) were well-separated from each other (Zeng et al., 2015). In another

324 study, soil potential denitrifying activity gradually increased with the succession of salt

325 marshland and reached the peak in the middle of the succession period. Soil potential

326 denitrifying activity was negatively correlated with soil pH and positively correlated with soil

327 organic carbon, nitrate, and soil moisture (Salles et al., 2017). Similarly, in our study, soil

328 potential denitrifying activity gradually increased and peaked in T2. The variation in soil

329 parameters such as plant litter, soil microbial activity, and plant species could lead to an increase

330 in PDA (Zeng et al., 2015). Plant litter decomposition increases soil organic carbon levels,

331 thereby facilitating soil microbial growth and survival and increasing soil metabolism and

332 enzyme activity (Ponge et al., 2003). In addition, a rich vegetation cover also positively affects

333 the activity of denitrifying enzymes (Means et al., 2017).

334 The nosZ denitrifying bacterial community abundance response to different

335 successional stages

336 Denitrifying bacteria are widely distributed in different ecosystems and strongly shaped by

337 the ecological conditions; soil depth and seasonal changes will significantly affect the

338 composition and abundance of denitrifying bacterial communities (Henry et al., 2006; Wu et al.,

339 2017; Tang et al., 2016). On the contrary, the abundance of functional genes (nirS/K, nosZ)

340 could also indirectly reflect the denitrification activity of the soil(Wu et al., 2017). Therefore,

341 quantitative analysis of denitrifying functional genes is often used to study denitrification

342 changes in a given ecosystem. In previous studies, nosZ gene abundance gradually increased and

343 peaked at the early successional stages, with a subsequent decreased in the mature stage

344 (Kandeler et al., 2006; Henry et al., 2006). Similarly, in our study, nosZ gene abundance reached

345 the peak at T2 and subsequently decreased. Soil formation and the accumulation of soil nutrients,

346 processes which occur in primary succession, provide a stable micro-habitat for soil

347 microorganisms. Plant species, richness, and litter also lead to variations in soil microbial

348 community dynamics and abundance (Zhang et al., 2017).

349 Previous studies have revealed that the abundance of functional genes differed among

350 different ecosystems. For example, the high soil pH of salt marshes is the most important factor

351 limiting microbial growth, and nosZ gene abundance was more or less constant throughout the

352 succession of such ecosystems(Salles et al., 2017). In forest ecosystems, plant species and plant

353 litter are abundant, facilitating the survival and growth of soil microorganisms, and the

354 functional genes gradually increased throughout succession(Meng et al., 2017). Generally, the

355 abundance of denitrifying bacterial communities is affected by soil properties and vegetation

356 factors such as soil organic carbon, total nitrogen, and the appearance of plant patches

357 (Brankatschk et al., 2011; Frouz et al., 2016).

358 Variation in nosZ denitrifying bacterial communities in response to primary

359 succession

360 We hypothesized that the variation in the composition of the nosZ denitrifying bacterial

361 community would strongly correlate with the successional stage. Although we did find that the

362 soils from the different sites, representing different successional stages, hosted distinct

363 denitrifying bacterial communities, this hypothesis was not clearly supported. Based on the

364 terminal restriction fragments (T-RFs), the microbial communities of the sites T0 and T1 stages

365 were similar and differed from those of the sites T2 and T3. In addition, more minor T-RFs

366 differed between the four successional stages, with each stage having its specific T-RFs. This

367 leads us to infer that the communities were relatively stable in the early stages of primary

368 succession, and only minor T-RFs contributed to the structural differences. Similar results have

369 been found in previous studies, suggesting that successional patterns of the compositions and

370 abundances of different denitrifying bacteria provide evidence that different denitrifying

371 bacterial communities develop in the course of primary succession. The relative abundance of

372 the denitrifying bacterial community increased with the development of the newly formed soils,

373 but significant differences were recorded only for older sites (Sigler et al., 2002; Kandeler et al.,

374 2006). The minor T-RFs contributed to the community structural differences that were tested in

375 different ecological systems, and Baldrian et al. (2013) and Gu et al. (2017) indicated that the

376 core denitrifying bacteria may not be sensitive to wetland degradation, whereas some minor

377 denitrifying bacteria were sensitive to primary succession (environmental changes).

378 Ollivier et al. (2011) have indicated that N cycling processes start from diazotrophs at the

379 initial succession sites of the glacier retreat areas, which can directly transform N2 into

380 ammonium which is sufficient to activate the nitrification process and further influence

381 denitrification. Our findings show that denitrification processes were initiated in the T0 site,

382 representing the starting point of primary succession, as evidenced by the predominance of

383 Mesorhizobium in the early stages (0 and 20 years); this genus has a strong nitrogen fixation

384 ability and promotes plant growth (Velez et al., 2017). At the T2 site, Brachybacterium was the

385 dominant genus, with a high nitrogen fixation and denitrification capacity; it mostly grows in

386 acidic-neutral soil(Saeki et al., 2017). Pseudomonas was found in the three sites T0, T1, and T2;

387 it has the ability to degrade cellulose, produce amylase, and dissolve calcium phosphate, and

388 based on its antibacterial properties, it has a certain certain inhibitory effect on soil diseases

389 (Meyer et al., 2017; Ramette et al., 2006; Almario et al., 2014). Sulfuritalea was stable

390 throughout succession; it is a facultative anaerobic bacterial genus that is autotrophic under

391 denitrification conditions and still abundant in low-oxygen nitrate-free environments (Watanabe

392 et al., 2017). In our study, those common nosZ denitrifying bacterial taxa were highly adapted to

393 the environmental conditions. The minor T-RFs differed among the four stages, and 45- and 70-

394 bp T-RFs were mainly found in the T1 site. Nitrospirillum and 95-bp T-RFs were found in T2

395 and T3, while the 90-bp T-RF was specific for the T0 and T1 sites. Those specific T-RFs were

396 sensitive to any changes in the course of primary succession, and the denitrification mechanism

397 of those specific T-RFs in the primary succession of glacier retreat areas deserves further

398 investigation.

399 Relationships between nosZ denitrifying bacterial communities and soil

400 properties

401 Environmental factors can significantly affect both the denitrification process and the

402 functional genes involved in this microbially mediated pathway, and factors such as carbon

403 substrate availability, moisture, pH, and other edaphic nutrients play crucial roles(Wallenstein et

404 al., 2006). A previous study has shown that organic carbon is the limiting factor in microbial

405 processes and activities in the early successional stages without vegetation cover, and carbon

406 accumulation affects the availability of soil nitrogen (Wardle et al., 2004), in this sense, soil

407 organic carbon and nitrogen levels in retreating glaciers can become the main driving factor in

408 shaping the composition of nitrogen-related microbial communities, and soil pH is the most

409 important factor impacting denitrification processes (Zeng et al., 2016; Čuhel et al., 2011 ; Kim

410 et al., 2015). Jha et al. (2017) have pointed out that the nosZ gene community structure had the

411 strongest correlation with soil moisture content and available phosphorus. In a similar study,

412 Tang et al. (2016) showed that phosphorus had a strong influence on the abundance of nitrogen-

413 cycling microorganisms in forest plantations, with a strong positive correlation between soil

414 available phosphorus and nitrogen cycling-related microbial functional gene abundance. Our

415 study also revealed that soil pH, AP, AK, and PDA significantly affect the composition of the

416 nosZ denitrifying bacterial communities. Soil pH was significantly negatively correlated with

417 denitrifying bacterial evenness (Eh), while soil AP and AK were positively correlated with the

418 Shannon diversity index (H). Soil AP, AK, and PDA were positively linked to richness (S).

419 Consistent with previous studies, AP and AK were the most important factors affecting the

420 composition of nosZ denitrifying bacterial communities (Tang et al., 2016). According to a

421 previous study, soil AP limits plant productivity and indirectly affects soil microbial

422 communities (Rangwala et al., 2012) Variations in the soil mineral elements stimulate the

423 microbial metabolic rates and activities, consequently changing the microbial community

424 composition. Soil microorganisms also significantly impact nutrient levels via nutrient cycling

425 (Gaimster et al., 2017). However, in forest ecosystems, carbon and nitrogen are generally

426 abundant, while phosphorus is the main factor affecting the abundance of microbial functional

427 genes. The different environmental factors had different influences on the microbial communities

428 and abundances. Nitrogen and carbon accumulation and nutrient availability shape the soil

429 microenvironment, although the geographic location of glaciers significantly affects microbial

430 succession (Zeng et al., 2015). However, microbial activities also influence and shape the soil

431 microenvironment through feedback mechanisms, regulating the stability and diversity of

432 ecosystems.

433 Conclusions434 We could show that the variation in soil properties, potential denitrifying activity,

435 abundance, and composition of the nosZ denitrifying bacterial community differed among

436 different successional stages. Soil pH decreased gradually in the course of succession, while

437 SOC and TN significantly increased with primary succession, and AP and AK increased

438 gradually and peaked at T2. The potential denitrifying activity and the nosZ gene copy numbers

439 followed the same trend as AP and AK, while the composition of the nosZ denitrifying bacterial

440 community differed among successional stages. Overall, 13 nosZ T-RFs were detected in the

441 primary succession stages; Mesorhizobium was the dominant genus across all stages, followed

442 by Brachybacterium and Azospirillum. The minor T-RFs were sensitive to the environmental

443 variation and contributed to the structural differences in the microbial community. Among the

444 measured soil properties, AP and AK were the key factors affecting the composition of the

445 denitrifying bacterial community (P < 0.01). The results of this study could contribute to support

446 a theoretical foundation for predicting the soil denitrifying bacterial community structure

447 variations in the primary succession of Hailuogou Glacier retreat areas.

448 Acknowledgments449 This study was supported by the Science and Technology Department of the Tibet and

450 Sichuan Agricultural University.

451 References452 Almario J, et al., 2014. Rhizosphere ecology and phytoprotection in soils naturally suppressive

453 to Thielaviopsis black root rot of tobacco. Environmental Microbiology 16:1949–1960 DOI:

454 10.1111/1462-2920.12459.

455 Ambrosini R, et al., 2017. Diversity and assembling processes of bacterial communities in

456 cryoconite holes of a karakoram Glacier. Microbial Ecology 73:827-837 DOI:

457 10.1007/s00248-016-0914-6.

458 Anesio AM, Laybourn-Parry J, 2012. Glaciers and ice sheets as a biome. Trends in Ecology &

459 Evolution 27:219-225 DOI: 10.1016/j.tree.2011.09.012.

460 BaldrianP,et al., 2013. Responses of the extracellular enzyme activities in hardwood forest to

461 soil temperature and seasonality and the potential effects of climate change. Soil Biology

462 and Biochemistry 56:60-68 DOI: 10.1016/j.soilbio.2012.01.020.

463 Bao SD, 2007. Soil Agro-Chemical Analysis. Beijing: China Agriculture Press.

464 Brankatschk R, et al., 2011. Abundances and potential activities of nitrogen cycling microbial

465 communities along a chronosequence of a glacier forefield. The ISME Journal 5:1025-1037

466 DOI: 10.1038/ismej.2010.184.

467 Cavicchioli R et al., 2002. Low-temperature extremophiles and their applications. Current

468 Opinion Biotechnology 13:253-261 DOI: 10.1016/S0958-1669(02)00317-8.

469 Chen Z,et al., 2012. Differentiated response of denitrifying communities to fertilization regime

470 in paddy soil. Microbial Ecology 63:446-459 DOI: 10.1007/s00248-011-9909-5.

471 Cotrufo MF, et al., 2013. The microbial efficiency-matrix stabilization (MEMS) framework

472 integrates plant litter decomposition with soil organic matter stabilization: Do labile plant

473 inputs form stable soil organic matter? Global Change Biology 19:988-995 DOI:

474 10.1111/gcb.12113.

475 Čuhel J, Šimek M, 2011. Effect of pH on the denitrifying enzyme activity in pasture soils in

476 relation to the intrinsic differences in denitrifier communities. Folia Microbiologica

477 56:230-235 DOI: 10.1007/s12223-011-0045-x.

478 Dambreville C, et al., 2006. Structure and activity of the denitrifying community in maize-

479 cropped field fertilized with composted pig manure or ammonium nitrate. FEMS

480 Microbiology Ecology 56:119-131 DOI: 10.1111/j.1574-6941.2006.00064.x.

481 Fernández-Martínez MA et al., 2017. Microbial succession dynamics along glacier forefield

482 chronosequences in Tierra del Fuego (Chile). Polar Biology 40:1939-1957 DOI:

483 10.1007/s00300-017-2110-7.

484 Frouz J, et al., 2016. Effects of soil substrate quality, microbial diversity and community

485 composition on the plant community during primary succession. Soil Biology and

486 Biochemistry 99:75-84 DOI: 10.1016/j.soilbio.2016.04.024.

487 Frouz J, et al., 2008. Interactions between soil development, vegetation and soil fauna during

488 spontaneous succession in post mining sites. European Journal of Soil Biology 44:109-121

489 DOI: 10.1016/j.ejsobi.2007.09.002.

490 Gaimster H, et al., 2017. Transcriptional and environmental control of bacterial denitrification

491 and N2O emissions. FEMS Microbiology Letters 365:277-297 DOI: 10.1093/femsle/fnx277.

492 Gu YF, et al., 2017. Implications of wetland degradation for the potential denitrifying activity

493 and bacterial populations with nirS, genes as found in a succession in Qinghai-Tibet plateau,

494 China. European Journal of Soil Biology 80:19-26 DOI: 10.1016/j.ejsobi.2017.03.005.

495 Hallin S, et al., 2018. Genomics and ecology of novel N2O-Reducing microorganisms. Trends in

496 Microbiology 26:43-55 DOI: 10.1016/j.tim.2017.07.003.

497 He L, Tang Y. 2008. Soil development along primary succession sequences on moraines of

498 Hailuogou Glacier, Gongga Mountain, Sichuan, China. Catena 72:0-269 DOI:

499 10.1016/j.catena.2007.05.010.

500 Henry S, et al., 2006. Quantitative detection of the nosZ gene, encoding nitrous oxide reductase,

501 and comparison of the abundances of 16S rRNA, narG, nirK, and nosZ genes in soils.

502 Applied and Environmental Microbiology 72:5181-5189 DOI: 10.1128/AEM.00231-06.

503 Hodkinson I D, Coulson S J, Webb N R, 2003. Community assembly along proglacial

504 chronosequences in the high Arctic:vegetation and soil development in north‐west Svalbard.

505 Journal of Ecology 91:651-663 DOI: 10.1046/j.1365-2745.2003.00786.x.

506 Jha N, et al., 2017. Soil properties impacting denitrifier community size, structure, and activity

507 in New Zealand dairy-grazed pasture. Biogeosciences Discussions 14:4243-4253 DOI:

508 10.5194/bg-14-4243-2017.

509 Jiang YL et al., 2018. Divergent assemblage patterns and driving forces for bacterial and fungal

510 communities along a glacier forefield chronosequence. Soil Biology and Biochemistr

511 118:207-216 DOI: 10.1016/j.soilbio.2017.12.019.

512 Kandeler E, et al., 2006. Abundance of narG, nirS, nirK, and nosZ Genes of denitrifying

513 bacteria during primary successions of a glacier foreland. Applied and Environmental

514 Microbiology 72:5957-5962 DOI: 10.1128/AEM.00439-06.

515 Kim M, et al., 2015. Highly heterogeneous soil bacterial communities around Terra Nova Bay

516 of Northern Victoria Land, Antarctica. PloS One 10:e0119966 DOI:

517 10.1371/journal.pone.0119966.

518 Leloup J, et al., 2018. Unravelling the effects of plant species diversity and aboveground litter

519 input on soil bacterial communities. Geoderma 317:1-7 DOI:

520 10.1016/j.geoderma.2017.12.018.

521 Li JB, et al., 2018. Stair-step pattern of soil bacterial diversity mainly driven by pH and

522 vegetation types along the elevational gradients of Gongga Mountain, China. Frontiers in

523 Microbiology 9:569-579 DOI: 10.3389/fmicb.2018.00569.

524 Li S, Peng M, Liu Z, Shah SS. 2017. The role of soil microbes in promoting plant growth.

525 Molecular Microbiology Research 7:30-37 DOI: 10.5376/mmr.2017.07.0004.

526 Liu XZ, Wang G, 2010. Measurements of nitrogen isotope composition of plants and surface

527 soils along the altitudinal transect of the eastern slope of Mount Gongga in southwest China.

528 Rapid Communications in Mass Spectrometry 24:3063-3071 DOI: 10.1002/rcm.4735.

529 Mapelli F et al., 2018. The stage of soil development modulates rhizosphere effect along a High

530 Arctic desert chronosequence. ISME Journal 12:1188-1198 DOI: 10.1038/s41396-017-

531 0026-4.

532 Means MM, Ahn C, Noe GB, 2017. Planting richness affects the recovery of vegetation and

533 soil processes in constructed wetlands following disturbance. Science of the Total

534 Environment 579:1366-1378 DOI: 10.1016/j.scitotenv.2016.11.134.

535 Meng H, et al., 2017. A comparison of denitrifying bacterial community structures and

536 abundance in acidic soils between natural forest and re-vegetated forest of Nanling Nature

537 Reserve in southern China. Journal of Environmental Management 198:41-49 DOI:

538 10.1016/j.jenvman.2017.04.066.

539 Meyer G, et al., 2017. Gross phosphorus fluxes in a calcareous soil inoculated with

540 Pseudomonas protegens CHA0 revealed by 33P isotopic dilution. Soil Biology and

541 Biochemistry 104:81-94 DOI: 10.1016/j.soilbio.2016.10.001.

542 Nicol GW, 2010. Primary succession of soil Crenarchaeota across a receding glacier foreland.

543 Environmental Microbiology 7:337-347 DOI: 10.1111/j.1462-2920.2005.00698.x.

544 Ollivier J, et al., 2011. Nitrogen turnover in soil and global change. FEMS Microbiology

545 Ecology 78:3-16 DOI 10.1111/j.1574-6941.2011.01165.x.

546 Philippot L, 2012. Denitrifying genes in bacterial and archaeal genomes. Biochimica et

547 Biophysica Acta-Gene Structure and Expression 1577:355-376 DOI: 10.1016/S0167-

548 4781(02)00420-7.

549 Ponge J F, 2003. Humus forms in terrestrial ecosystems: a framework to biodiversity. Soil

550 Biology and Biochemistry 35:935-945 DOI: 10.1016/S0038-0717(03)00149-4.

551 Püschel D, Rydlová J, Vosátka M, 2007. Mycorrhiza influences plant community structure in

552 succession on spoil banks. Basic and Applied Ecology 8:510-520 DOI:

553 10.1016/j.baae.2006.09.002.

554 Ramette A., Moënne-Loccoz Y, Défago G, 2006. Genetic diversity and biocontrol potential of

555 fluorescent pseudomonads producing phloroglucinols and hydrogen cyanide from Swiss

556 soils naturally suppressive or conducive to Thielaviopsis basicola-mediated black root rot of

557 tobacco. FEMS Microbiology Ecology 55:369–381 DOI: 10.1111/j.1574-

558 6941.2005.00052.x.

559 Rangwala I, Miller JR, 2012. Climate change in mountains: a review of elevation-dependent

560 warming and its possible causes. Climatic Change 114:527-547 DOI: 10.1007/s10584-012-

561 0419-3.

562 Roubíčková A, Mudrák O, Frouz J, 2009. Effect of earthworm on growth of late succession

563 plant species in postmining sites under laboratory and field conditions. Biology and Fertility

564 of Soils 45:769-77 DOI: 10.1007/s00374-009-0386-7.

565 Saeki Y, et al., 2017. Effect of flooding and the nosZ gene in bradyrhizobia on bradyrhizobial

566 community structure in the soil. Microbes and Environments 32:154-163 DOI:

567 10.1264/jsme2.ME16132.

568 Salles JF, et al., 2017. Successional patterns of key genes and processes involved in the

569 microbial nitrogen cycle in a salt marsh chronosequence. Biogeochemistry 132:185-201

570 DOI: 10.1007/s10533-017-0296-y.

571 Schmidt MW, et al., 2011. Persistence of soil organic matter as an ecosystem property. Nature

572 478:49–56 DOI: 10.1038/nature10386.

573 Sigler W V, Zeyer J, 2002. Microbial diversity and activity along the forefields of two receding

574 glaciers. Microbial Ecology 43:397-407 DOI: 10.1007/s00248-001-0045-5.

575 Six J,et al., 2004. A history of research on the link between (micro) aggregates, soil biota, and

576 soil organic matter dynamics. Soil and Tillage Research 79:7-31 DOI:

577 10.1016/j.still.2004.03.008.

578 Solomon S, Qin D, Manning M, et al. 2007. IPCC Fourth Assessment Report: Climate change,

579 Cambridge 4:213-252

580 Strauss SL, Ruhland CT, Day TA, 2009. Trends in soil characteristics along a recently

581 deglaciated foreland on Anvers Island. Polar Biology 32:1779-1788 DOI: 10.1007/s00300-

582 009-0677-3.

583 Sun HY et al., 2016. Variations of bacterial and fungal communities along a primary

584 successional chronosequence in the Hailuogou glacier retreat area (Gongga Mountain, SW

585 China). Journal of Mountain Science 13:1621-1631 DOI: 10.1007/s11629-015-3570-2.

586 Tang YQ, Et al., 2016. Impacts of nitrogen and phosphorus additions on the abundance and

587 community structure of ammonia oxidizers and denitrifying bacteria in Chinese fir

588 plantations. Soil Biology and Biochemistry 103:284-293 DOI:

589 10.1016/j.soilbio.2016.09.001.

590 Töwe S et al., 2010. Abundance of microbes involved in nitrogen transformation in the

591 rhizosphere of Leucanthemopsis alpina (L.) Heywood grown in soils from different sites of

592 the Damma Glacier Forefield. Microbial Ecology 60:762-770 DOI: stable/40926479.

593 Velez PA, et al., 2017. Synergistic effect of chickpea plants and Mesorhizobium as a natural

594 system for chromium phytoremediation. Environmental Technology 38:2164-2172 DOI:

595 10.1080/09593330.2016.1247198.

596 Watanabe T, et al., 2017. Dominance of Sulfuritalea species in nitrate-depleted water of

597 astratified freshwater lake and arsenate respiration ability within the genus. Environmental

598 Microbiology Reports 9:522-527 DOI: 10.1111/1758-2229.12557.

599 Wallenstein MD, et al., 2006. Environmental controls on denitrifying communities and

600 denitrification rates: insights from molecular methods. Ecological Applications 16:2143-

601 2152 DOI: 10.2307/40061947.

602 Wardle DA, Walker LR, Bardgett RD, 2004. Ecosystem properties and forest decline in

603 contrasting long-term chronosequences. Science 305:509–513 DOI:

604 10.1126/science.1098778.

605 Wang CX, et al., 2012. Nitrous oxide reductase gene (nosZ) and N2O reduction along the littoral

606 gradient of a eutrophic freshwater lake. Journal of Environmental Sciences 25:44–52 DOI:

607 10.1016/S1001-0742(12)60005-9.

608 Wuebbles DJ, 2009. Nitrous Oxide: No Laughing Matter. Science 326:56-57 DOI:

609 10.1126/science.1179571.

610 Wu X, et al., 2017. Effects of land-use change and fertilization on N2O and NO fluxes, the

611 abundance of nitrifying and denitrifying microbial communities in a hilly red soil region of

612 southern China. Applied Soil Ecology 120:111-120 DOI: 10.1016/j.apsoil.2017.08.004.

613 Yang Y et al., 2014. Dynamics of carbon and nitrogen accumulation and C:N stoichiometry in a

614 deciduous broadleaf forest of deglaciated terrain in the eastern Tibetan Plateau. Forest

615 Ecology and Management 312:10-18 DOI: 10.1016/j.foreco.2013.10.028.

616 Zeng J, et al., 2015. Primary succession of soil enzyme activity and heterotrophic microbial

617 communities along the chronosequence of Tianshan Mountains No. 1 Glacier, China.

618 Antonie van Leeuwenhoek 107:453-466 DOI: 10.1007/s10482-014-0343-9.

619 Zeng J, et al., 2016. Primary succession of nitrogen cycling microbial communities along the

620 deglaciated forelands of Tianshan Mountain, China. Frontiers in Microbiology 7:1353-1365

621 DOI: 10.3389/fmicb.2016.01353.

622 Zhang MM, Luo P, Liu F, 2017. Nitrogen removal and distribution of ammonia-oxidizing and

623 denitrifying genes in an integrated constructed wetland for swine wastewater treatment.

624 Ecological Engineering 104:30–38 DOI: 10.1016/j.ecoleng.2017.04.022.

625 Zhou J, et al., 2013. Changes of soil phosphorus speciation along a 120-year soil

626 chronosequence in the Hailuogou Glacier retreat area (Gongga Mountain, SW China).

627 Geoderma 195-196:251-259 DOI: 10.1016/j.geoderma2012.12.010.

628 Zhou J et al., 2018. Weathering of primary mineral phosphate in the early stages of ecosystem

629 development in the Hailuogou Glacier foreland chronosequence. European Journal of Soil

630 Science 69:450-561 DOI: 10.1111/ejss.12536.

631 Tables632 Table 1 Details of the sampling sites within the Hailuogou Glacier retreat area.

633 Table 2 Soil properties of the different sampling sites

634 Table 3 The nosZ gene diversity along successional stages.

635 Table 4 Pearson´s correlations between soil properties, potential denitrifying activity, and nosZ

636 gene diversity.

637 Figure legends638 Fig.1 Soil samples were collected in sites representing a chronosequence of primary succession

639 in the Hailuogou Glacier area. Triangles correspond to Site 1- 0 years, Site 2 - 20 years, Site 3 -

640 40 years, and Site 4 - 60 years.

641 Fig.2 Principal components analysis of the soil properties in different soil successional stages. T0:

642 Soil successional age ≤ 5 years; T1: Soil successional age ≤ 20 years; T2: Soil successional

643 age ≤ 40 years; T3: Soil successional age ≤ 60 years.

644 Fig.3 The nosZ gene copy numbers of different soil successional stages.T0: Soil successional age

645 ≤ 5 years; T1: Soil successional age ≤ 20 years; T2: Soil successional age ≤ 40 years; T3: Soil

646 successional age ≤ 60 years. Error bars indicate±SD.

647 Fig. 4 The nosZ-denitrifying bacterial community composition of different soil successional

648 stages. T0: Soil successional age ≤ 5 years; T1: Soil successional age ≤ 20 years; T2: Soil

649 successional age ≤ 40 years; T3: Soil successional age ≤ 60 years.

650 Fig.5 Redundancy analysis ordination diagram of the nosZ-denitrifying bacterial community

651 composition associated with environmental variables.T0: Soil successional age ≤ 5 years; T1:

652 Soil successional age ≤ 20 years; T2: Soil successional age ≤ 40 years;T3: Soil successional

653 age ≤ 60 years. SOC: soil organic carbon; TN: total nitrogen; AP: available phosphorus; AK:

654 available potassium; PDA: Potential denitrifying activity.

Table 1(on next page)

Details of the sampling sites within the Hailuogou Glacier retreat area.

Note: Horizontal distance refers to the distance from the glacier.

Table 1 Details of the sampling sites within the Hailuogou Glacier retreat area.

Site codeSuccessional time

(years)Longitude Latitude

Horizontal distance

(m)Elevation (m)

T0 ≤5 101°59′32.8″ 29°34′03.6″ 100 2,956

T1 ≤20 101°59′40.2″ 29°34′04.1″ 300 2,948

T2 ≤40 101°59′43.7″ 29°34′05.7″ 600 2,940

T3 ≤60 101°59′48.9″ 29°34′07.9″ 1,200 2,926Note.

Horizontal distance refers to the distance from the glacier.

Table 2(on next page)

Soil properties of the different sampling sites

Note: Values represent mean ± standard error (n = 5). Values within the same column

followed by the same letter do not differ at P < 0.05. SOC: soil organic carbon; TN: total

nitrogen; AP: available phosphorus; AK: available potassium. PDA: Potential denitrifying

activity.

Table 2 Soil properties of the different sampling sites

Sampling code pH SOC (g/kg) TN (g/kg) AP (mg/kg) AK (mg/kg) PDA (µg/g/h)

T0 8.41 ± 0.08a 4.87 ± 0.14d 0.29 ± 0.01d 0.98 ± 0.07c 13.60 ± 0.45d 0.08 ± 0.00d

T1 7.85 ± 0.03b 9.36 ± 0.35c 0.49 ± 0.02c 1.24 ± 0.04c 40.45 ± 1.78c 0.10 ± 0.01c

T2 7.31 ± 0.05c 16.19 ± 0.63b 0.76 ± 0.04a 5.54 ± 0.34a 120.80 ± 3.36a 0.31 ± 0.01a

T3 7.19 ± 0.15c 30.77 ± 1.93a 1.11 ± 0.09b 3.70 ± 0.15b 73.59 ± 2.23b 0.13 ± 0.00bNote.

Values represent mean ± standard error (n = 5). Values within the same column followed bythe same letter do not differ at P < 0.05. SOC: soil organic carbon; TN: total nitrogen; AP:available phosphorus; AK: available potassium. PDA: Potential denitrifying activity.

Table 3(on next page)

The nosZ gene diversity along successional stages

Note: Values represent mean ± standard error (n = 5). Values within the same column

followed by the same letter do not differ at P < 0.05.

Table 3 The nosZ gene diversity along successional stages

Sample codes Shannon-Wiener index (H) Richness (S) Evenness (Eh)

T0 1.82 ± 0.03c 8.99± 0.30c 1.03 ± 0.01c

T1 1.87 ± 0.03c 10.04 ± 0.38c 1.13 ± 0.04b

T2 2.54 ± 0.05a 18.41 ± 0.36a 1.28 ± 0.01a

T3 2.32 ± 0.07b 11.57 ± 0.47b 1.30 ± 0.02aNote.

Values represent mean ± standard error (n = 5). Values within the same column followed bythe same letter do not differ at P < 0.05.

Table 4(on next page)

Pearson´s correlations between soil properties, potential denitrifying activity, and nosZ

gene diversity

Note: SOC: Soil organic carbon; TN: Total nitrogen; AP: Available phosphorus; AK: Available

potassium; PDA: Potential denitrifying activity. **P < 0.01, *P < 0.05.

Table 4 Pearson´s correlations between soil properties, potentialdenitrifying activity, and nosZ gene diversity

pH SOC TN AP AK PDA nosZ geneabundance

Shannon-Weiner index (H) -0.879 0.684 0.776 0.995** 0.968* 0.871 0.901

Richness (S) -0.658 0.292 0.428 0.932 0.953* 0.999** 0.995**

Evenness (Eh) -0.995** 0.879 0.944 0.872 0.872 0.647 0.728

nosZ gene abundance -0.709 0.335 0.473 0.935 0.970* 0.990** 1.000

Note.SOC: Soil organic carbon; TN: Total nitrogen; AP: Available phosphorus; AK: Available

potassium; PDA: Potential denitrifying activity. **P < 0.01, *P < 0.05.

Figure 1

Soil samples were collected in sites representing a chronosequence of primary

succession in the Hailuogou Glacier area.

Triangles correspond to Site 1- 0 years, Site 2 - 20 years, Site 3 - 40 years, and Site 4 - 60

years.

Figure 2(on next page)

Principal components analysis of the soil properties in different soil successional stages.

T0: Soil successional age ≤ 5 years; T1: Soil successional age ≤20 years; T2: Soil

successional age ≤ 40 years; T3: Soil successional age ≤ 60 years.

-2 -1 0 1 2

PCA1 (89.51%)

-2-1

01

2

PC

A 2

(7

.74

%)

T0 T1 T2 T3

Figure 3(on next page)

The nosZ gene copy numbers of different soil successional stages.

T0: Soil successional age ≤5 years; T1: Soil successional age ≤ 20 years; T2: Soil

successional age ≤ 40 years; T3: Soil successional age ≤ 60 years. Error bars indicate±SD.

T0

T1

T2

T3

0 1 2 3 4 5 6Log10 number of nosZ gene copies g

-1 dry soil

Su

ccessio

nal s

tag

es

Figure 4(on next page)

The nosZ-denitrifying bacterial community composition of different soil successional

stages.

T0: Soil successional age ≤ 5 years; T1: Soil successional age ≤ 20 years; T2: Soil

successional age ≤ 40 years; T3: Soil successional age ≤ 60 years.

T0 T1 T2 T30

20

40

60

80

100R

ela

tive

ab

un

da

nce

(%

)

Successional stages

95bp

90bp

70bp

55bp

50bp

45bp

40bp

35bp

30bp

25bp

20bp

15bp

Figure 5(on next page)

Redundancy analysis ordination diagram of the nosZ-denitrifying bacterial community

composition associated with environmental variables.

T0: Soil successional age ≤5 years; T1: Soil successional age ≤ 20 years; T2: Soil

successional age ≤ 40 years;T3: Soil successional age ≤ 60 years. SOC: soil organic carbon;

TN: total nitrogen; AP: available phosphorus; AK: available potassium; PDA: Potential

denitrifying activity.

-1.0 -0.5 0.0 0.5 1.0

RDA1 (53.90%)

-1.0

-0.5

0.0

0.5

1.0

RD

A2 (

10.4

3%

)

pH

SOC

TN

APAK

PDA

T0 T1 T2 T3