Antibiotics Disturb the Microbiome and Increase the Incidence of

Subscriber access provided by Thompson Rivers University | Library

Article

Antibiotics Disturb the Microbiome and Increase the Incidence of Resistance Genes in the Gut of a Common Soil Collembolan Dong Zhu, Xin-Li An, Qing-Lin Chen, Xiaoru Yang, Peter Christie, Xin Ke, Longhua Wu, and Yong-Guan Zhu Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b04292 • Publication Date (Web): 29 Jan 2018 Downloaded from http://pubs.acs.org on January 31, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 38

Environmental Science & Technology

1

Antibiotics Disturb the Microbiome and Increase the Incidence of

2

Resistance Genes in the Gut of a Common Soil Collembolan

3 4

Dong Zhu,†,‡ Xin-Li An,†,‡ Qing-Lin Chen,†,‡ Xiao-Ru Yang, † Peter Christie,

§

Xin Ke,||

5

Long-Hua Wu, § Yong-Guan Zhu, *, †,

6

†

7

Chinese Academy of Sciences, 1799 Jimei Road, Xiamen 361021, China.

8

‡

University of the Chinese Academy of Sciences, 19A Yuquan Road, Beijing 100049, China.

9

§

Key Laboratory of Soil Environment and Pollution Remediation, Institute of Soil Science,

⊥

Key Laboratory of Urban Environment and Health, Institute of Urban Environment,

10

Chinese Academy of Sciences, Nanjing 210008, China.

11

||

12

Chinese Academy of Sciences, Shanghai 200032, China.

13

⊥

14

Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, China

Institute of Plant Physiology and Ecology, Shanghai Institute of Biological Sciences,

State Key Laboratory of Urban and Regional Ecology, Research Center for

15 16

ABSTRACT:

17 18

Gut microbiota make an important contribution to host health but the effects of

19

environmental pressures on the gut microbiota of soil fauna are largely uncharacterized. Here,

20

we examine the effects of norfloxacin and oxytetracycline on the gut microbiome of the

21

common soil collembolan Folsomia candida and concomitant changes in the incidence of

22

antibiotic resistance genes (ARGs) in the gut and in growth of the collembolan. Exposure to 1

ACS Paragon Plus Environment


23

10 mg antibiotics kg-1 for two weeks significantly inhibited the growth of the collembolan

24

with roughly a ten-fold decrease in 16S rRNA gene abundance. Antibiotics did alter the

25

composition and structure of the collembolan gut microbiome and decreased the diversity of

26

the gut bacteria. A decline in the Firmicutes/Bacteroidetes ratio in the antibiotic-treated

27

collembolans may be responsible for the decrease in body weight. Exposure to antibiotics

28

significantly increased the diversity and abundance of ARGs in the collembolan gut. The

29

Mantel test and Procrustes analysis both reveal that ARGs and gut microbiota were

30

significantly correlated with one another (P < 0.05). These results indicate that antibiotics

31

may induce a shift in the gut microbiota of non-target organisms such as soil collembolans

32

and thereby affect their growth and enrichment of ARGs.

33 34 35 36 37 38 39 40 41 42 43 44 2


Page 2 of 38

Page 3 of 38


45

TOC ART

46

47 48 49 50 51 52 53 54 55 56 57 58 59 3



60

Page 4 of 38

INTRODUCTION

61 62

Soil fauna play a key role in the soil ecosystem, influencing the maintenance of 1-3

63

biodiversity, litter decomposition, nutrient cycling and energy transfer

. Collembolans are

64

one of the most widespread and important groups of soil microarthropods and are often found

65

in areas of high organic matter content where organic particles and microorganisms are

66

ingested

67

extensively in soil ecology, ecogenomics and ecotoxicology5-11. However, to date, only a

68

study has described the collembolan gut microbiome using 16S rRNA gene high-throughput

69

sequencing 12. Gut microbiota make a vital contribution to many aspects of animal health 13-16,

70

including feeding, growth, pathogen defense, energy metabolism, reproduction, immunity

71

and aging. It has been reported that arthropods rely on their gut microbiota to obtain essential

72

nutrient resources

73

collembolan gut microbiome.

4, 5

. The collembolan Folsomia candida is a model organism that has been studied

17, 18

. We therefore urgently require more information regarding the

74 75

Soil pollution is a matter of great concern in many agricultural and industrial areas 19 and

76

soil fauna are direct witnesses of soil pollution and are often used as indicators of

77

environmental pollution20. Antibiotics are major soil contaminants that have recently given

78

rise to considerable concern because they (or their residues) usually have long half-lives and

79

high biological activity

80

evolve as a consequence of competition among bacterial communities, repeated applications

81

of large amounts of animal manure to agricultural soils have profoundly increased antibiotic

18, 21

. Although antibiotics are part of the natural soil ecosystem and

4


Page 5 of 38


21, 22

82

concentrations in the soils

. It is generally thought that the effects of antibiotics on

83

animals consist of direct damage to their physiology and disturbance of their gut microbial

84

communities

85

animal physiology (e.g. growth, reproduction and metabolism)

86

microbiota will be exposed to antibiotics when manure containing antibiotics are ingested by

87

soil biota. Studies on aquatic and aboveground terrestrial animals have shown that ingestion

88

of antibiotics can reduce their gut microbial diversity

89

exposure on the gut microbiota of non-target soil fauna remain largely unknown. At the same

90

time, it has been demonstrated that anthropogenic arsenic exposure can suppress the

91

microbiomes of several taxa of earthworm such as Verminephrobacter symbionts 27. A study

92

has also indicated that the viability of the earthworm gut microbiota has some potential in

93

reflecting the status of soil pollution 28. These investigations suggest that further studies are

94

required to explore the responses of soil animal gut microbial communities to antibiotics.

18

. Many previous studies have investigated the effects of antibiotics on soil 23-25

.

And soil animal gut

16, 26

. However, the effects of antibiotic

95 96

In addition, some bacteria can adapt rapidly to antibiotic exposure in the environment via

97

antibiotic-resistance mechanisms, and antibiotic resistance genes (ARGs) can transfer within

98

and between bacterial communities via numerous mobile genetic elements (MGEs) 18, 29. The

99

overuse or misuse of antibiotics has led to a significant increase in the abundance and 30-32

100

amounts of ARGs in the soil environment worldwide

, and this is now a topic of great

101

public concern. The gut is an important site for the accumulation of ARGs due to the

102

selective pressure of the ingested antibiotics 33, 34. Previous studies confirm that oral exposure

103

to antibiotics can increase the diversity and abundance of ARGs in the gut of mice and pigs 35, 5



Page 6 of 38

104

36

105

animals due to antibiotic exposure and this restricts our understanding of the spread of ARGs

106

in soil food webs. In addition, antibiotic biosynthesis genes have recently been found in F.

107

candida and these are believed to originate from microorganisms

108

microbiome and ARGs of F. candida may contribute to our understanding of gut microbial

109

and ARG diversity.

. However, there is virtually no information available on ARG enrichment in the gut of soil

37

. Thus, studying the gut

110 111

The aims of the present study were therefore to confirm the effects of exposure to

112

antibiotics on soil collembolan growth and gut microbiota using bacterial 16S rRNA gene

113

high-throughput sequencing, to characterize the abundance and diversity of ARGs in the

114

collembolan gut via high throughput qPCR (295 primer sets targeting almost all major classes

115

of ARGs and 10 MGE marker genes), and to explore the relationship between the

116

collembolan gut microbial community and ARGs. The results may help us to understand the

117

spread of ARGs in soil food webs.

118 119

MATERIALS AND METHODS

120 121

Test Animals and Antibiotics. The soil collembolan Folsomia candida (“Berlin strain”)

122

used originated from Aarhus University in Denmark and has been cultured in our laboratory

123

for > six years. The conditions of rearing and synchronization of the animals have been

124

described by Zhu et al. (2016)

125

individuals were exposed to food spiked with antibiotics.

38

. In the current experiment, synchronized F. candida

6


Page 7 of 38


126

Norfloxacin

(NFC;

127

-carboxylicacid;

128

10,12,12a-hexahydroxy-6-methyl-1,11-dioxo-monohydrochloride;

129

purity, >99%) were selected as the model antibiotics. Both are widely used in human and

130

veterinary medicine and are the main residual antibiotics in animal manure. Previous studies

131

shown that their concentrations are typically up to several mg kg-1 in contaminated manure39,

132

40

133

collembolans. Therefore, in the present study we selected an environmentally relevant

134

concentration (10 mg kg-1) of norfloxacin and oxytetracycline for exposure to the test

135

collembolan.

CAS

1-ethyl-6-fluoro-1,4-dihydro-7-(1-piperazinyl)-4-oxo-3-quinoline 70458-96-7,

purity,

>99%)

and

oxytetracycline CAS

(OTC;

2058-46-0,

. When these manure are applied to agricultural soils they are likely ingested by the soil

136 137

Antibiotic Exposure Treatment. Baker’s yeast (Angel Yeast Co., Ltd, Hubei, China)

138

was used as the diet. Yeast spiked with each antibiotic (10 mg antibiotic kg-1 dry weight food)

139

was obtained by thorough mixing of the yeast with an appropriate volume of antibiotic

140

solution. The same volume of deionized water was mixed with the yeast as a control. The

141

three mixtures were frozen at -20 oC, freeze-dried, ground and stored at -20 oC prior to use.

142

There were three treatments including the control and four replicates of each treatment,

143

giving a total of 12 samples. For each replicate, fifty synchronized F. candida individuals

144

(18-20 days) were transferred into a Petri dish (inner diameter 90 mm) with a layer (thickness

145

5 mm) of a moist mixture of activated charcoal and plaster of Paris (1:8 w/w).

146

Antibiotic-spiked and control yeasts (5 mg) were respectively added twice a week, and

147

unused old food was removed. Oral exposure of F. candida to the antibiotics took place for 7



148

two weeks in a 16:8 h (dark: light, 800 lux) light regime at 20 ± 2 °C. Distilled water was

149

supplied twice a week to maintain moist conditions. The number of collembolans was

150

counted in each replicate dish at the end of the feeding period and all the collembolans were

151

harvested. The guts of thirty of the harvested individuals were dissected. The surplus

152

collembolans were starved for 92 hours on moist filter paper to remove the gut contents and

153

then freeze-dried. The dry weight of the collembolans was determined using a Mettler Toledo

154

XS3DU automatic electronic microbalance (precision ± 1 µg).

155

A series of norfloxacin and oxytetracycline solutions were mixed with the yeast to

156

produce nominal concentrations of 0 (Control), 0.1, 1, 10, 100 and 1000 mg antibiotics kg-1

157

dry weight food. Ten synchronized F. candida (10-12d) were exposed to a series of

158

antibiotics-spiked yeast to preform a standard 28-day reproduction test according to ISO

159

guideline 11267 (ISO1999)

160

was counted via the number of juvenile.

41

. After exposure for 28 days, the reproduction of collembolan

161 162

DNA Extraction from the Collembolan Gut. The harvested collembolans were killed

163

immediately with chloroform to restrain their ingestion and excretion of the gut contents. The

164

dead collembolans were immediately introduced into 2% sodium hypochlorite (NaOCl)

165

solution for 10 s to prevent interference from the body surface microbiota. The collembolans

166

were then rinsed five times with sterilized water. After aseptic treatment of the body surface

167

the collembolans were dissected using sterile forceps under sterile conditions, and thirty

168

dissected guts were transferred into a 2-mL Eppendorf tube with 0.96 mL sterile phosphate

169

buffer solution. The resulting gut samples were stored at −80 °C until DNA isolation. 8


Page 8 of 38

Page 9 of 38


170 171

Total DNA was extracted from the 12 collembolan gut samples using a FastDNA Spin

172

Kit for soil (MP Biomedicals, Illkirch, France). The procedure of DNA isolation generally

173

followed the manufacturer’s instructions with slight modification as follows. Before the

174

bead-beating step, 20 µL proteinase-K (Thermo Fisher Scientific, Waltham, MA) was added

175

to the spin column and kept at 55°C for 3 h. The extracted collembolan gut DNA was eluted

176

using 50 µL DES solution. The concentration and quality of extracted DNA were checked

177

using 1.0% agarose gel electrophoresis and spectrophotometric analysis (Nanodrop ND-1000,

178

Thermo Fisher). Finally, the extracted DNA was preserved at -20 oC for further analysis.

179 180

Quantification of Bacterial Abundance using Real-Time Quantitative PCR. The total

181

bacterial abundance of the collembolan gut was estimated in triplicate with 16S rRNA gene

182

copy number using real-time quantitative PCR (RT-qPCR) with a Light Cycler 480

183

used the universal primers 5 ′ - GGGTTGCGCTCGTTGC -3 ′

184

ATGGYTGTCGTCAGCTCGTG -3′ to amplify the bacterial 16S rRNA gene. All assays

185

were conducted in a 20 µL qPCR reaction system consisting of 10 µL 2× TransStart® Top

186

Green qPCR SuperMix (AQ131, Transgen biotech, Beijing, China), 0.8 µL bovine serum

187

albumin (BSA, 20 mg mL-1), 0.25 µM each primer, and 2 µL gut DNA as a template. The

188

thermal profile of RT-qPCR was as follows: 5 min at 95 °C followed by 40 cycles of 15 s at

189

95 °C, 60 s at 60 °C and 20 s at 72 °C. Three non-template reactions were performed as

190

negative controls. Triplicate 10-fold serial dilutions of plasmid DNA with target-gene were

191

used to obtain a standard curve (amplification efficiency 96−104%, r2 > 0.99). 9


42

. We

and 5 ′ -


Page 10 of 38

192 193

High-Throughput Quantitative PCR for Analysis of Antibiotic Resistance Genes. All

194

high-throughput quantitative PCR (HT-qPCR) of ARGs were carried out in triplicate using

195

the Wafergen SmartChip Real-time PCR system to investigate the abundance and diversity of

196

ARGs in collembolan gut samples. We used a total of 296 primer (Table S1) sets to target 285

197

ARGs (major classes of ARGs), class 1 intergron, clinical class 1 integron, 8 transposases and

198

the 16S rRNA gene in one run. The 100 nL HT-qPCR reaction system per well consisted of

199

LightCycler 480 SYBR Green I Master mix, each primer, nuclease-free PCR grade water,

200

bovine serum albumin and gut DNA template. The reaction mixture was heated for 10 min at

201

95 °C and then 40 cycles of 30s at 95 °C and 30s at 60 °C. The SmartChip qPCR software

202

was used to analyze the outcomes. A threshold cycle (CT) of 31 was set up as the detection

203

limit. The ARGs were considered to be detected when the ARGs of three replicates were all

204

amplified. The normalized copy number of ARGs per cell was calculated using equations

205

below 43:

206

Relative gene copy number = 10^((31−CT)/(10/3))

207

Normalized ARG copy number = (Relative ARG copy number/ Relative 16S rRNA gene

208

copy number)*4.1. Because the average number of 16S rRNA gene per bacterial cell is currently estimated

209 210

at 4.1 based on the Ribosomal RNA Operon Copy Number Database 44.

211

16S rRNA Gene Amplification, Illumina Sequencing and Bioinformatic Analysis.

212 213

The

gut

DNA

samples

were

amplified

using

10


universal

primers

515F:

Page 11 of 38


214

GTGCCAGCMGCCGCGG

and

907R

labelled

with

unique

215

CCGTCAATCMTTTRAGTTT to target the hypervariable V4-V5 region of the bacterial 16S

216

rRNA gene in 50 µL reaction system. The conditions of PCR amplification and recovery of

217

purified products were performed as described previously

218

spectrophotometer was used to determine the concentration of PCR purified products. The

219

sequencing library was constructed via twelve barcoded samples of equal quality and then

220

high-throughput sequencing of the library was carried out on the Illumina Hiseq2500

221

platform (Novogene, Tianjin, China).

30,

barcodes:

31

. The NanoDrop

222 223

The high-throughput sequencing data were analyzed using Quantitative Insights Into 45

224

Microbial Ecology (QIIME, version 1.8.0) following the online instructions

225

reads were merged after the adaptor, and primer sequences, ambiguous nucleotides and

226

low-quality reads were removed to obtain clean combined reads. In QIIME,

227

pick_de_novo_otus.py was used to pick the operational taxonomic units (OTUs), and OTUs

228

were identified at 3% sequence difference by UCLUST clustering

229

analysis, OTUs with only one sequence (singletons) were discarded. The default method was

230

adopted to obtain a representative sequence of the OTU in each cluster. The taxonomy of

231

each representative sequence was aligned using PyNAST and assigned via the RDP Classifier

232

2.2 based on the Greengenes 13.8 16S rRNA gene database 47, 48. The Fast tree algorithm was

233

used to build a phylogenetic tree for downstream analysis

234

estimated using metrics observed species (OTU), and it was used to describe bacterial alpha

235

diversity. The microbial difference and beta-diversity of different samples were compared 11


. Paired-end

46

. Before downstream

49

. The Shannon index was


236

using the Adonis test and principal coordinate analysis (PCoA) based on the unweighted

237

Unifrac metric, respectively.

238 239

Statistical Analysis. All data are presented as mean value ± standard error (SE). The R 50

with vegan 2.4−3

51

240

version 3.4.1

was used to conduct the Adonis test and calculate the

241

Shannon index of the gut microbiota. The Procrustes test was used to describe the correlation

242

between ARGs and bacterial communities in R version 3.4.1 with vegan 2.4−3, and a

243

heatmap package 52 of R was selected to draw the heatmap. Differences among samples were

244

analyzed using single factor analysis of variance (ANOVA), LSD and t-tests using the IBM

245

SPSS version 22 statistical software package. The associated OTUs of different samples were

246

counted using ANOVA and Bayesian model-based moderated t-tests in R. Significant

247

differences were detected at the 0.05 level. The ternary plot was produced using OriginPro

248

9.1. In ternary plot, each individual OTU is depicted by one circle. The size of the circle

249

reflects the relative abundance of the OTU. The position of each circle is determined by the

250

contribution of the indicated compartments to the relative abundance. The associated OTUs

251

are based on a Bayes moderated t-test (P < 0.05, FDR-corrected). OriginPro 9.1 and

252

Microsoft Excel were used to generate other graphics.

253 254 255

Accession Numbers. All sequencing data are deposited in the National Center for Biotechnology Information Sequence Read Archive under the accession number SRP116006.

256 257

RESULTS 12


Page 12 of 38

Page 13 of 38


258 259

Change in Collembolan Growth and Gut Bacterial Abundance. The collembolan

260

body weights declined significantly in the norfloxacin and oxytetracycline treatments by 16.2

261

and 21.6%, respectively, compared to the control (P < 0.05) after 14 days of exposure to the

262

antibiotics (Figure 1a). Results of a standard 28-day reproduction test indicated that the

263

reproduction of collembolan was reduced after exposure to antibiotics at 10 mg kg-1

264

compared to the control, but no significant difference was observed (t-test, P > 0.05) (Figure

265

S1). Antibiotic exposure strongly inhibited the abundance of the gut microbiota (P < 0.05)

266

and the gut 16S rRNA gene copy number was reduced about 10-fold in the antibiotic

267

treatments compared with the control (Figure 1b). No significant differences in body weight

268

or gut microbial abundance were observed between norfloxacin and oxytetracycline

269

treatments (Figure 1) (P > 0.05).

270 271

Effects of Antibiotics on Collembolan Gut Bacterial Diversity.

Across all samples,

272

453195 non-singleton reads were calculated, minimum 19495 sequences per sample were

273

obtained and 4718 OTUs were identified. Actinobacteria (37.93%), Firmicutes (24.42%),

274

Proteobacteria (17.22%) and Bacteroidetes (14.57%) were the four predominant phyla in all

275

samples of collembolan gut (Figure 2a). The results of high throughput sequencing indicate

276

that exposure to norfloxacin and oxytetracycline significantly altered the collembolan gut

277

microbial community (Adonis test, P < 0.05). At the phylum level, the collembolans exposed

278

to antibiotics had higher abundance of Actinobacteria and Bacteroidetes (P < 0.05) and lower

279

abundance of Firmicutes (P < 0.05) in the gut compared to the control. Antibiotics exposure 13



Page 14 of 38

280

caused the significant reduction of Firmicutes abundance from 42.51 to 16.38%, and the

281

obvious

282

Firmicutes/Bacteroidetes (F/B) ratio of the antibiotic treatments was significantly lower than

283

that of the control (P < 0.05). Within the phylum Actinobacteria a significant increase in the

284

abundance of the family Streptomycetaceae was observed in the antibiotic exposure

285

treatments compared to the control (P < 0.05) (Figure 2a). Exposure to antibiotics led to a

286

significantly higher gut abundance of the family Sphingobacteriaceae (P < 0.05) (Figure 2a).

287

Compared with the control, the relative abundance of the genus Streptomyces increased

288

significantly (by 538 and 348%) in the norfloxacin and oxytetracycline treatments,

289

respectively (P < 0.05) (Figure 2b). However, the abundance of 16S reads at the family level

290

overall decreased in antibiotic-treated collembolan gut (Figure 2b), a result supported by

291

Figure S2 showing that only a few significantly higher OTUs were found in the antibiotic

292

treatments. Around 43% of norfloxacin-treated OTUs and 42% of oxytetracycline-treated

293

OTUs changed significantly in relative abundance due to antibiotic exposure (Figure S3).

294

Compared with the control, pronounced shifts in microbial OTU profiles were observed in

295

antibiotic-treated collembolan gut microbiota. Principal coordinates analysis based on

296

unweighted unifrac distance also shown a clear separation between the antibiotic treatments

297

and the control gut microbiota along with the X axis (Figure 3a). Significant differences were

298

observed between intra-treatment distances and inter-treatment distances (Figure 3b) (P
1% families) are

668

shown, and the data presented (mean, n = 4) were log-transformed.

669 670

Figure 3. (a) Principal coordinates analysis (PCoA) of collembolan gut samples using

671

unweighted unifrac distances. C, N and O indicate control, norfloxacin and oxytetracycline

672

treatments, respectively. Each point represents a microbial community of one sample. (b)

673

Similarity of gut microbiota between samples. Unweighted unifrac distances (mean ± SE)

674

between control gut microbiota (C-C), between norfloxacin-treated gut microbiota (N-N),

675

between

676

norfloxacin-treated gut microbiota (C-N), between control and oxytetracycline-treated gut

677

microbiota (C-O) and between norfloxacin-treated and oxytetracycline-treated gut microbiota

678

(N-O). Different letters indicate significant differences (p < 0.05) among different distances

oxytetracycline-treated

gut

microbiota

(O-O),

31


between

control

and


679

(LSD test). (c) Change in Shannon index (mean ± SE, n = 4) of the antibiotic-treated

680

collembolan gut microbiota. Different letters indicate significant differences (p < 0.05)

681

among different treatments (LSD test).

682 683

Figure 4. (a) Detected number per 30 animals and (b) normalized copy number of ARGs and

684

mobile genetic elements (MGEs) (mean ± SE, n = 4) in the collembolan gut microbiota

685

exposed to antibiotics.

686 687

Figure 5. Heat map describing enrichment of ARGs in the antibiotic-treated collembolan gut

688

compared with the control. The data presented (mean, n = 4) were log-transformed.

689 690

Figure 6. Procrustes test showing the significant correlation between ARG profile and

691

collembolan gut microbial composition (16S rRNA gene OTU data) on the basis of

692

Bray−Curtis dissimilarity metrics (sum of squares M2 = 0.27, r = 0.7965, P = 0.043, 9999

693

permutations). The triangles indicated 16S rRNA gene OTU data of collembolan gut

694

microbial composition, and the squares indicated experimental treatments.

32


Page 32 of 38

Page 33 of 38


Figure 1

a

b


Figure 2

a


b


Page 34 of 38

Figure 3

Page 35 of 38


a

b

c


c


Figure 4

a

b


Page 36 of 38

Page 37 of 38


Figure 5



Figure 6


Page 38 of 38

Antibiotics Disturb the Microbiome and Increase the Incidence of

Recommend Documents