Effective Establishment of Donor Gut Microbiota in Gnotobiotic Mice

Home , Anaeroplasmatales

1 Effective establishment of donor gut microbiota in gnotobiotic mice

3 Jocelyn M. Choo1,2*, Geraint B. Rogers1,2

4 1. SAHMRI Microbiome Research Laboratory, Flinders University College of Medicine and

5 Public Health, Adelaide, SA, Australia

6 2. Microbiome and Host Health, South Australia Health and Medical Research Institute, North

7 Terrace, Adelaide, SA, Australia.

9 *Corresponding author

10 Dr Jocelyn Choo

11 5D305 Flinders Medical Centre, Flinders Drive, Bedford Park,

12 South Australia 5042

13 [email protected]

15 Running title: Faecal microbiota transfer for gnotobiotic mice models

16 Word count for text: 4053

17 ABSTRACT

18 Background: Determining whether associations between gut microbiota characteristics and

19 host physiology represent causal relationships is a fundamental challenge for microbiome

20 research. One way these associations can be investigated is to instil donor faecal material into

21 gnotobiotic mice and to assess the extent to which donor phenotype is recapitulated. However,

22 the manner in which this process is performed varies considerably between studies, and

23 assessment of microbiota re-establishment in recipient animals is not always carried out. We

24 report a detailed investigation of microbiome assembly in germ-free mice and compare the

25 effects of single and multiple rounds of faecal gavage, using both native and antibiotic-

26 disrupted donor material.

27 Results: Levels of bacteria within the faeces of recipient animals increased rapidly following

28 the instillation of donor material. However, considerable instability in microbiota composition

29 continued during the first two weeks post-gavage, with substantial changes in taxon relative

30 abundance occurring in parallel to declining faecal pH. Relative compositional stability was

31 not achieved until day 28 and persistent differences between recipient and donor microbiota

32 remained. These included an increased relative abundance of Bacteroidetes, and a reduced

33 relative abundance of Firmicutes. Of taxa detected in donor material, 52% were represented in

34 stable recipient microbiota following transplantation with native faecal material (single

35 gavage), increasing to 66% following three rounds of gavage. These taxa accounted for 95%

36 and 91% of total donor bacterial abundance, respectively. Performing multiple rounds of

37 gavage significantly increased microbiota similarity between donor and recipient, and

38 significantly reduced within-group dispersion (P<0.05). Instillation of antibiotic-associated

39 microbiota resulted in substantially lower temporal and inter-animal variance, with multiple

40 rounds of gavage providing no substantial benefit.

41 Conclusions: Microbiome assembly in recipient animals is not immediate and several weeks

42 are required for microbiota stability to be achieved. Multiple rounds of faecal gavage result in

43 greater similarity to donor microbiota and reduced inter-animal variance. The process of donor

44 microbiota re-establishment, and therefore the interval required prior to investigations using

45 recipient animals, is influenced by donor microbiota characteristics.

47 Key words: germ-free, faecal microbiota transfer, gut microbiota colonisation.

48 Introduction

49 Much of our understanding of relationships between the gut microbiome and human

50 physiology comes from observational studies. Having identified associations between gut

51 microbiome traits and host measures, the next challenge is to determine whether these

52 relationships are causal, secondary to observed physiological phenomena, or result from

53 parallel but independent processes.

54 One of the few experimental strategies available to assess causality in host-microbiome

55 interactions is the instillation of intestinal microbiota from human donors or animal models

56 into germ-free mice, and assessment of whether donor phenotype is recapitulated. This

57 approach has been used, for example, to demonstrate that gut microbiota contributes to obesity

58 in the context of a high fat diet [31], to underdevelopment in the context of a low nutrient

59 density/bioavailability diet [4], and to variation in treatment efficacy in recipients of cancer

60 immunotherapy [18, 34]. The employment of this strategy is becoming increasingly popular

61 due to growing access to gnotobiotic facilities, and evidence that antibiotic depletion of

62 recipient intestinal microbiota in conventional mice prior to faecal transfer represents a

63 relatively poor alternative [13, 37].

64 Despite the utility of gut microbiota transplantation as a means to understand microbiome-host

65 relationships, there is little consistency in the manner in which the technique is performed. In

66 particular, the number of rounds of gavage employed varies between studies, as does the period

67 allowed to elapse between the final gavage and the initiation of the experiment or assessment.

68 The extent to which donor microbiota are replicated within recipient animals is commonly

69 neglected, and where assessment is performed, it often focuses donor taxon presence/absence,

70 rather than microbiota structure or composition.

71 Many of the bacterial clades that are closely associated with the regulation of host physiology

72 are obligate anaerobes [33, 46], and are particularly susceptible to loss of viability during the

73 processing of material for transplant [26, 27]. Not only can failure to establish such taxa in the

74 gut lumen of gnotobiotic recipient animals lead to divergence in microbiota composition from

75 donor animals, but it can allow the proliferation of opportunistic facultative anaerobes [22, 42].

76 Such changes can have profound implications for the metabolic and immuno-regulatory

77 properties of the gut microbiome [1, 16].

78 A number of previous studies have aimed to describe the process of intestinal microbiota

79 assembly in germ-free mice [12, 17]. In particular, Gillilland and colleagues described

80 microbiota assembly at two mucosal sites, the caecum and the jejunum, during the first 21 days

81 following the instillation of caecal microbiota, reporting evidence of both ecological succession

82 and site-specific effects [17]. Aidy and colleagues described microbiome compositional

83 dynamics and metabolic function in germ-free mice that received a faecal suspension over a 16

84 day period [12], again with evidence of ecological succession. However, important knowledge

85 gaps remain, including what the effects of multiple rounds of microbiota instillation are, and

86 what considerations are necessary when attempting to establish substantially modified donor

87 microbiota.

88 We investigated the dynamics of donor microbiome assembly in the gut of recipient germ-free

89 mice, including a comparison of the effects of single and multiple rounds of gavage.

90 Assessment of microbiota transplantation was performed using both native and antibiotic-

91 disrupted gut microbiota, with the latter used to represent changes that are commonly described

92 in association with pathophysiology [25, 39].

94 RESULTS

95 Establishment of normal gut bacterial abundance

96 In mice that received native donor microbiota (approximately inoculum of 105 bacterial cells),

97 total bacterial load peaked between four days post initial gavage (D4) and seven days post

98 initial gavage (D7), before declining to D14 (Figure 1A). From D14, faecal bacterial load was

99 broadly stable at approximately 5x106 bacterial cells/mg. No significant differences were

100 observed between mice that received one or three rounds of gavage.

101 Compared to native microbiota, antibiotic-disrupted donor had reduced species richness

102 (Figure S2A-S2B), reduced relative abundance of Firmicutes, and increased relative abundance

103 of Verrucomicrobia and Bacteroidetes (Figure S2C). The pattern of an increase in bacterial

104 load following the instillation of antibiotic-disrupted microbiota was similar to that seen with

105 native microbiota (Figure 1B). The number of rounds of faecal gavage performed was not

106 associated with a significant difference in total bacterial load (Figure 2B).

107

108 Donor taxa representation within recipient faecal microbiota

109 Initial assessment of recipient faecal microbiota was based on the percentage of donor taxa

110 detected. Similar trends were observed for recipients of one and three rounds of gavage, with

111 detected taxa rising steeply from D4 (median, IQR: 1G = 43.1%, 32.8 - 44.0; and 3G= 44.8%,

112 40.5 - 44.8) but remaining variable throughout the 70-day study (Figure 2A). On average, the

113 percentage of donor taxa detected was higher in recipients of three rounds of gavage than

114 recipients of a single gavage at all time-points bar one (D56). This difference was significant

115 at the end of the study period (median, IQR: 1G= 51.7%, 46.6-60.4; 3G= 65.5%, 62.9-69.0; P

116 <0.05). In recipients of antibiotic-disrupted microbiota, donor taxa detected increased

117 significantly as a result of multiple gavages (Figure 3A). The difference between recipients of

118 single and multiple rounds of gavage remained significant throughout the 70-day study

119 (median, IQR: 1G= 33.3%, 33.3-37.5; 3G= 37.5%, 37.5-39.6; P < 0.05, except at D42).

120 The percentage of donor taxa detected in recipient animals fluctuated over the 70-day period

121 of assessment. As mice were housed in a controlled environment in which the only route of

122 bacterial acquisition was faecal gavage, intermittent detection of individual taxa was likely to

123 be due to sampling bias. Assessment of the relative abundance of taxa in relation to their

124 presence or intermittent absence at each timepoint supported this hypothesis (Figure S3), with

125 the latter taxa relative abundance being significantly closer to the threshold of detection

126 (Median, IQR: Present= 0.005, 0.001 - 0.013; Intermittent= 0, 0 - 0.001; Mann-Whitney test,

127 P< 0.05). As a result, we also assessed the percentage of donor taxa detected as a cumulative

128 measure. Again, median percentage of donor taxa detected was lowest at D4 (metrics as above;

129 median, IQR: G1 = 43.1%, 32.8-44.0; G3 = 44.8%, 40.5-44.8) and reached a maximum at D70

130 (G1 = 81.0%, 79.3 - 84.5; G3 = 91.4%, 88.8 - 94.0) (Figure 2B). The mean percentage of donor

131 taxa detected was significantly higher in recipients of multiple compared to single gavage at

132 all sample points between D21 and D70 (Mann-Whitney test, P< 0.05). Similarly, the median

133 percentage of donor taxa in recipients of single or multiple rounds of gavage with antibiotic-

134 disrupted microbiota was lowest from D4 to D14 (Figure 3B). Multiple gavage was associated

135 with a detection of a higher median percentage of donor taxa compared to single gavage at

136 D21, a difference that persisted throughout the 70 day study (1G= 37.5%, 35.4 - 37.5; 3G=

137 45.8%, 37.5 - 45.8; P<0.05).

138 Recipient microbiota was further assessed based on the proportion of donor bacterial

139 abundance represented. Between D4 and D70, bacterial taxa detected in recipient animals

140 represented 89.9 to 95.5% of donor relative abundance (single gavage) or 90.1 to 95.5%

141 (multiple gavage) (Figure 2A). In recipients of antibiotic-disrupted microbiota, detected

142 bacterial taxa represented 99.8 to 100% (single gavage) and 95.2 to 100% (multiple gavage)

143 between D4 and D70, respectively (Figure 3A).

144

145 Re-establishment of donor microbiota composition

146 Recipient-donor microbiota similarity was assessed based on weighted UniFrac distance.

147 Similarity increased in both groups to D35, after which no significant differences were

148 observed between consecutive time points (PERMANOVA P>0.05), with the exception of D70

149 in the multiple gavage group. From D14, recipients of a single gavage had lower similarity to

150 donor (higher mean weighted UniFrac distance) compared to recipients of multiple gavages

151 (mean ± SD: 0.19 ± 0.05 vs 0.17 ± 0.04, respectively), with the exception of D42 and D70

152 (Figure 4A). Differences between single and multiple gavage groups achieved significance at

153 D28, D56 and D70 (P< 0.05).

154 Microbiota composition between consecutive timepoints in animals that received antibiotic-

155 disrupted microbiota achieved stability by D35 in recipients of multiple gavage, and D56 in

156 recipients of single gavage (PERMANOVA P>0.05). Recipient-donor microbiota similarity

157 was lower (D14-D70, 0.34 ± 0.02 for single and multiple gavage groups; Figure 4B), although

158 variations in similarity scores across time were substantially less when compared to recipients

159 of native microbiota. No significant differences were observed between recipients of single or

160 multiple rounds of gavage.

161 Microbiota composition was also assessed based on distance to group centroid. In recipients of

162 native microbiota, distance to centroid increased from D4, peaking at D7, and declining

163 thereafter (Figure 5A). Distance to centroid was higher at all but one time-point (D49) in mice

164 that received one round of faecal gavage compared to those that received three, achieving

165 statistical significance at D14, D28, and D63 (Mann-Whitney test, P< 0.05). Distance to group

166 centroid in recipients of antibiotic-disrupted microbiota did not differ significantly between

167 single and multiple gavage groups (Figure 5B).

168

169 Temporal dynamics in taxon relative abundance

170 Temporal changes in taxon relative abundance differed substantially between phylogenetic

171 clades. The relative abundance of Bacteroidales increased substantially from donor values

172 following instillation. Both Bacteroidales and Lactobacillales remained high throughout the

173 study period (Figure 6A), while the relative abundance of Bifidobacteriaceae,

174 Verrucomicrobiales, and Erysipelotrichaceae increased sharply following instillation, peaked

175 at D4, and then declined to near donor levels (Figure 6A-B). Bacillales, Desulfovibrionales,

176 Saccharimonadales, and Anaeroplasmatales were not detectable at D4, but increased gradually

177 over time (Figure 6B), while Peptostreptococcaceae relative abundance declined to D21 and

178 was undetectable thereafter (Figure 6C). Less variation were observed for the relative

179 abundance of other taxa, including Ruminococcaceae. Bacterial family temporal dynamics

180 were broadly consistent between single and multiple gavage groups.

181 The variation in bacterial family relative abundance was significantly lower in recipients of

182 antibiotic-disrupted microbiota (Figure 7A-B). Dominant families (Lachnospiraceae,

183 Bacteroidales, and Verrucomicrobiales) were broadly stable, although a sustained increase in

184 Bacteroidales was observed between D28 and D35. In contrast to recipients of native donor

185 microbiota, the relative abundance of Ruminococcaceae fell substantially following instillation

186 until D7, before partially recovering. Again, no significant differences were observed between

187 single and multiple gavage groups.

188 At the phylum level, substantial differences between donor and recipient microbiota were

189 evident at D7, persisting to D70 (Figure S4). In particular, the relative abundance of the most

190 prevalent phylum, Firmicutes, was reduced by an average of 30.3% (3G, D70), while the

191 relative abundance of the second most prevalent phylum, Bacteroidetes, increased by an

192 average of 29.1% (3G, D70). Phylum-level differences between donor and recipient microbiota

193 were explained largely by the relative abundance of Lactobacillus and uncultured members of

194 Muribaculaceae, respectively.

195

196 Temporal dynamics in taxon absolute abundance and faecal pH

197 The absolute abundance of three genera, Blautia (Lachnospiraceae), Bifidobacterium

198 (Bifidobacteriales) and Akkermansia (Verrucomicrobiales), which were dominant members of

199 the bacterial families whose relative abundance changed substantially in recipients of native or

200 antibiotic-disrupted microbiota, were determined by qPCR (Figure S5). Absolute abundance

201 dynamics for these genera were broadly consistent with their relative abundances. Levels of

202 Bifidobacterium and Akkermansia declined from D4 and were almost depleted by D14 in

203 recipients of the native microbiota, whereas consistent levels of Blautia were observed across

204 the timepoints in recipients of the antibiotic-associated microbiota.

205 Temporal variation in faecal pH was observed in recipients of the native microbiota. Faecal pH

206 decreased following microbiota instillation, from pH 7.1 and pH 7.3 to pH 6.5 and pH 6.4 at

207 D4, in the single and multiple gavage groups respectively (Figure S6). These levels increased

208 by D7 and remained consistent for the multiple gavage group until D28, whereas larger

209 variation across subsequent timepoints were observed for the single gavage group.

210 DISCUSSION

211 Our goal was to characterise microbiota assembly in gnotobiotic mice following the instillation

212 of donor faecal material. In doing so, we aimed to address knowledge gaps relating to how

213 microbiota transplantation is performed in the investigation of causal relationships in host-

214 microbiome associations. Our focus, in particular, was the effect of multiple versus single

215 rounds of gavage, and differences in the dynamics of bacterial colonisation when native or

216 substantially disrupted donor microbiota are used.

217 In keeping with a previous investigation of microbiota establishment in gnotobiotic mice [17],

218 we observed the total bacterial levels in recipient animals to resemble donor animals rapidly.

219 The rate of microbiota expansion was not affected by the number of rounds of gavage

220 performed. Increases in bacterial diversity during the early stages of gut colonisation are

221 constrained by ecological succession rather than by rate of biomass increase, as described in

222 vaginally-born human infants [43]. Succession involves a process whereby early gut colonisers

223 facilitate the growth of other taxa through modification of growth substrates, production of

224 bioactive metabolites, and alteration of the physicochemical characteristics of the gut

225 environment.

226 There are many well-described examples of inter-species interactions influencing gut

227 microbiota assembly. For example, members of the Actinobacteria phylum, including

228 bifidobacteria, are able to hydrolyse host glycans to release products including glucose,

229 galactose, lactate and acetate, metabolites that are then utilised by members of the

230 Bacteroidetes and Firmicutes phyla [11, 14]. Antagonistic interactions are also common and

231 important in the suppression of pathogen growth. For example, bacterial fermentation of

232 complex carbohydrates to produce short-chain fatty acids alters intestinal pH and inhibits

233 replication of Enterobacteriaceae, including Escherichia coli and Salmonella [7], while

234 bacterial conversion of primary bile acids to secondary bile acids inhibits the endospore

235 germination and vegetative growth of Clostridioides difficile [36]. These processes are

236 particularly important in the re-assembly of gut microbiota following antibiotic disruption [6].

237 The changes in microbiota characteristics that we observed following instillation into germ-

238 free mice were consistent with ecological succession. In mice transplanted with native

239 microbiota, dissimilarity to donor microbiota and within-group dispersion were highest during

240 the initial two weeks period, with changes in keystone bacterial clades consistent with well-

241 described mechanisms of gut bacterial succession. For example, levels of Bifidobacteriaceae

242 were initially high, but became substantially depleted by D14, consistent with the role played

243 by members of this family as primary gut colonisers [21, 41]. In contrast, the relative abundance

244 of Desulfovibrionales increased substantially over the first two weeks post-instillation. Many

245 members of Desulfovibrionales are sulphate-reducers, utilising hydrogen gas produced in the

246 biosynthesis of SCFAs. Levels of Desulfovibrio, for example, have been reported to increase

247 with levels of butyrate [12]. The dynamic changes in microbiota structure during early

248 colonisation with native microbiota were also reflected by changes in faecal characteristics,

249 including in pH. The rapid decrease in pH at D4 post-colonisation followed by an increase

250 thereafter might be explained by the high levels of Bifidobacterium observed, a taxon capable

251 of the central hexose fermentation pathway (bifid shunt) to produce lactate, which reduces

252 faecal pH [2, 15].

253 While we did not encounter overgrowth of individual opportunistic taxa, such as Escherichia

254 coli, in recipient animals, as reported in previously [17], persistent differences between donor

255 and recipient microbiota composition were evident. Most notable was a reduction in the relative

256 abundance of Firmicutes, the predominant phylum, and an increase in the relative abundance

257 of Bacteroidetes. In addition, several bacterial genera were detected in donor microbiota but

258 not in transplant recipients, including Anaerofustis, Bacteroides and Erysipelothrix. Why such

259 differences should occur between mice of similar genetic backgrounds is not clear, however,

260 there are a number of potential contributory factors. For example, the donor transplant material

261 were derived from the caecum, while recipient microbiota composition was assessed based on

262 faecal pellets. Differences in gut physiology in gnotobiotic mice compared with conventional

263 mice have also been described, including in immune mechanisms that are involved in the

264 regulation of intestinal microbiology [19, 23, 28, 40, 45], which could influence colonisation.

265 Approaches used in the conventionalisation of germ-free mice as part of published studies vary

266 considerably. Some investigators have employed single rounds of gavage [4, 31], while others

267 have used multiple (typically three) rounds [13, 18]. Disappointingly, many other studies do

268 not describe donor material preparation or the number of instillations that were performed. In

269 our study, multiple rounds of gavage were associated with modest but consistent increases in

270 the similarity of recipient and donor microbiota, both where native or antibiotic-disrupted

271 microbiota were instilled. These differences are attributable to several bacterial taxa that were

272 detected in recipients of multiple rounds of gavage but not in recipients of single gavage,

273 including Jeotgalicoccus, Lachnospiraceae UC5-12E3 and Mollicutes RF39. Multiple gavage

274 was also associated with reduced inter-animal variance, an important factor when using such

275 animals to investigate host-microbiota interactions.

276 In addition to investigating microbiota assembly in mice transplanted with native microbiota,

277 we also assessed colonisation dynamics following the instillation of microbiota from antibiotic-

278 exposed animals. In doing so, our aim was to determine whether the dynamics of

279 conventionalisation with substantially-disrupted microbiota, as associated with a number of

280 pathophysiological contexts [4, 31], differs to those with intact gut microbiota. The donor

281 microbiota that resulted from antibiotic disruption exhibited substantially reduced taxa

282 richness, and an increased prevalence of facultative anaerobes. Compared to native microbiota,

283 instillation of antibiotic-disrupted microbiota was associated with greatly reduced temporal

284 variation. The early colonisation period (up to D7) was characterised by the rapid establishment

285 of predominant taxa (Lachnospiraceae, Bacteroidales, and Verrucomicrobiales). These rapidly

286 achieved levels comparable to those in donor material and showed little variation over time.

287 Divergence from donor composition was driven instead by a failure to recover less prevalent

288 taxa (including Staphylococcus and Clostridiales vadin BB60), which were undetectable by

289 D4. In addition, Alistipes and multiple members of the Lachnospiraceae and Ruminococcaceae

290 families, failed to colonise recipient animals (either single or multiple gavage), while the

291 Intestinimonas and Anaeroplasma genera were detected only in multiple gavage recipients.

292 The absence of several members of both the Lachnospiraceae and Ruminococcaceae families

293 can be attributed to competitive interaction between taxa in an altered bacterial community

294 [35]. Taxa in these families that successfully colonise recipient mice can also provide

295 compensatory function, including their role in producing important metabolites such as short

296 chain fatty acids by degrading complex polysaccharides [3].

297 In our study, caecal microbiota were harvested and processed under strict anaerobic conditions.

298 Alternative approaches that utilise stool, or that involve exposure to aerobic conditions, are

299 likely to be associated with lower levels of donor-recipient similarity. Furthermore, our

300 transplantation was between members of the same species, subject to similar environmental

301 and dietary exposures. A growing number of studies involve the instillation of human stool, or

302 stool-derived microbiota, into gnotobiotic mice. Substantial differences in genetics,

303 physiology, anatomy, and diet, would be expected to further reduce associated levels of

304 microbiota recapitulation.

305 Our study had a number of limitations. For example, we did not attempt to assess all possible

306 gavage schedules or donor microbiota variants. Microbiota assembly dynamics are likely to

307 change with donor microbiome characteristics, and are therefore vary with mouse genetic

308 background [20] and between populations housed at different research facilities [8, 30]. What

309 host measures are being assessed, and what the hypothesised mechanisms of host-microbiome

310 interaction are, will also be important considerations in the design of experiments involving

311 transplanted microbiota.

312

313 CONCLUSIONS

314 In summary, a substantial period (more than 4 weeks) may be required following faecal

315 instillation into germ-free mice to achieve microbiota stabilisation. Establishment of donor

316 microbiota occurs more rapidly in less diverse bacterial communities that result from antibiotic

317 exposure. While multiple rounds of faecal instillation result in greater similarity to donor

318 microbiota, the process of microbiota assembly differs considerably based on the complexity

319 and composition of donor bacterial communities. A failure to understand the extent to which

320 donor microbiota has been established in recipient mice, and the degree of community stability

321 achieved, could contribute to spurious findings.

322 METHODS

323 Preparation of donor faecal material for transplantation

324 Faecal material was obtained from seven week old C57BL/6J donor mice, bred at the SAHMRI

325 Bioresources animal facility, Adelaide, Australia. To induce antibiotic disruption of faecal

326 microbiota, mice received water containing erythromycin ethylsuccinate (20 mg/kg) for 90

327 days, as opposed to plain water. Mice were killed by carbon dioxide inhalation. Caeca were

328 harvested immediately and transferred to 15% anaerobic glycerol-phosphate buffered saline

329 (PBS). All subsequent processing was performed under anaerobic conditions (10% CO2, 10%

330 H2, 80% N2).

331 Caecal contents were removed from encasing tissue, weighed, resuspended in 4x (w/v)

332 anaerobic PBS, and homogenised by vortexing. The resulting suspension was passed through

333 a Falcon 100 μM nylon cell filter (Thermofisher Scientific, Waltham, USA) to obtain the non-

334 fibrous content. The supernatant was mixed with an equal volume of 30% anaerobic glycerol-

335 PBS, and stored in a Hungate tube at -80°C until required. Prior to oral gavage, caecal

336 supernatant was diluted with 2x volume of anaerobic PBS (pH 7.2) and sealed in a glass vial

337 within the anaerobic chamber.

338 Germ-free mice were obtained from the Translational Research Institute (University of

339 Queensland) and housed within germ-free isolators (Park Bioservices LLC, USA) at the

340 SAHMRI Preclinical Imaging and Research Laboratories (PIRL) germ-free facility (Gilles

341 Plains, Adelaide, Australia). All mice were maintained on autoclaved Teklad Global 18%

342 Protein Rodent Diet (Envigo, Huntingdon, UK) throughout the study.

343 After a 10 day period of acclimatisation, caecal suspensions were instilled into germ-free mice

344 under sterile conditions. Germ-free mice were inoculated with 150 μL of the appropriate donor

345 caecal suspension (containing 105-106 bacterial cells) via an oral gavage. Faecal sampling was

346 performed at baseline and throughout the experiment (Figure S1).

347

348 Faecal pellet collection, DNA extraction, and bioinformatic processing

349 Mice were placed in sterile individual cages for faecal pellet collection. Fresh faecal pellets

350 were transferred aseptically to 1.5 mL Eppendorf tubes and stored at -80°C prior to analysis.

351 Faecal pellets were resuspended in 300 μL of PBS by vortexing, and pelleted by centrifugation

352 at 13 000 × g for 10 min at 4°C. Supernatant was transferred to a sterile 2 mL screwcap tube

353 and stored at -80°C. Pellets underwent DNA extraction by a combination of mechanical and

354 chemical cell lysis methods using the DNeasy PowerSoil kit (QIAGEN, Hilden, Germany),

355 according to the manufacturer’s instructions and eluted in 100 μL of sterile DNase- and RNase-

356 free water.

357 Amplicon libraries of the V4 hypervariable region for 16S rRNA gene amplicon sequencing

358 were prepared from DNA extracts using modified universal bacterial primer pairs 515F and

359 806R [9]. Amplicon libraries were indexed, cleaned, and sequenced according to the Illumina

360 MiSeq 16S Metagenomic Sequencing Library Preparation protocol on a 2 x 300 bp Miseq

361 reagent kit v3 at the David R Gunn Genomics Facility, South Australian Health and Medical

362 Research Institute.

363 Paired-end 16S rRNA gene sequence reads were analysed using QIIME v2.0 [5]. Briefly, de-

364 noising was performed on de-multiplexed sequences using Dada2, with sequence reads

365 truncated at a specific length based on a quality filtering score of 30 to remove low quality

366 sequence region. Taxonomic classification of sequence variants were performed based on a

367 pre-trained classifier composing of the V4 hypervariable region sequences of the SILVA 132

368 16S rRNA reference database clustered at 97% similarity [29]. All samples were subsampled

369 to 4,859 reads. Sufficient coverage at this depth is confirmed by the rarefaction curve, which

370 reached an asymptote. Taxon relative abundance were used in downstream analyses. Sequence

371 data is available from the Sequence Read Archive (SRA) repository under the accession

372 number PRJNA592263.

373

374 Quantitative PCR

375 Previously described quantitative PCR (qPCR) assays were used to determine the abundance

376 of Akkermansia muciniphila [44], Bifidobacterium spp [32] and Blautia spp. [38], as well as

377 the 16S rRNA gene for total bacterial load [24]. SYBR-based qPCR assays were performed

378 using 0.2 nM of each primer, 1X PowerUP SYBR Green mastermix (Thermofisher Scientific,

379 Waltham, USA), 1 μL of DNA, and sterile DNase- and RNase-free water to make up to a total

380 reaction volume of 35 μL. Each reaction was divided to three, 10 μL, replicate reactions. Total

381 bacterial load was calculated using a standard curve generated from a known concentration of

382 Escherichia coli DNA.

383

384 Measurement of faecal pH

385 The pH levels of faecal samples were measured using the method described by Xie et al., 2016,

386 with modifications. Faecal samples from each mice were collected and pooled according to

387 cage to a total of approximately 50 – 100mg. Faeces were resuspended in 9x volume of

388 deionized water and the suspension vortexed for 1 min. The pH value of faecal suspension was

389 them measured on a FE20 FiveEasyTM pH meter (Mettler-Toledo AG, Schwerzenbach,

390 Switzerland).

391

392 Statistical analysis

393 Bacterial taxa that were not detected in donor material, or present as a single read in only one

394 sample, were removed from recipient microbiota data. Representation of donor taxa in recipient

395 mice was determined based on their presence in more than one mouse per group. Beta diversity

396 analysis were performed based on weighted UniFrac distances computed between samples

397 using QIIME. Compositional differences between groups and distance to the group centroid

398 were determined based on the PERMANOVA and PERMDISP test, respectively, using

399 PRIMER [10]. Comparison between single and multiple gavage groups was performed by

400 Mann-Whitney test using Graphpad PRISM (v8).

401 Ethics approval and consent to participate: All animal studies were performed in

402 accordance to comply with the relevant guidelines(Australian Code for the care and use of

403 animals for scientific purposes (8thEdition 2013)) and approved by the South Australian

404 Health And Medical Research Institute Animal Ethics Committee under the study reference

405 SAM378.

406

407 Consent for publication: Not applicable.

408

409 Availability of data and material: Sequencing data generated are available in Sequence Read

410 Archive (SRA) repository, PRJNA592263.

411

412 Competing interests: The authors declare that they have no competing interests.

413

414 Funding: GBR is supported by a Matthew Flinders Research Fellowship and a National Health

415 and Medical Research Council Senior Research Fellowship (GNT1155179).

416

417 Authors’ contributions: JMC and GBR designed the study. JMC performed laboratory work

418 and data analysis. JMC and GBR interpreted the findings and prepared the manuscript.

419

420 Acknowledgements: Not applicable.

421 REFERENCES

422 [1] Arthur, J.C., Perez-Chanona, E., Muhlbauer, M., Tomkovich, S., Uronis, J.M., Fan, 423 T.J., Campbell, B.J., Abujamel, T., Dogan, B., Rogers, A.B., et al. (2012). Intestinal 424 inflammation targets cancer-inducing activity of the microbiota. Science 338, 120-123. 425 426 [2] Baxter, N.T., Schmidt, A.W., Venkataraman, A., Kim, K.S., Waldron, C., and 427 Schmidt, T.M. (2019). Dynamics of Human Gut Microbiota and Short-Chain Fatty Acids in 428 Response to Dietary Interventions with Three Fermentable Fibers. mBio 10. 429 430 [3] Biddle, A., Stewart, L., Blanchard, J., and Leschine, S. (2013). Untangling the genetic 431 basis of fibrolytic specialization by Lachnospiraceae and Ruminococcaceae in diverse gut 432 communities. Diversity 5, 627-640. 433 434 [4] Blanton, L.V., Barratt, M.J., Charbonneau, M.R., Ahmed, T., and Gordon, J.I. (2016). 435 Childhood undernutrition, the gut microbiota, and microbiota-directed therapeutics. Science 436 352, 1533. 437 438 [5] Bolyen, E., Rideout, J.R., Dillon, M.R., Bokulich, N.A., Abnet, C.C., Al-Ghalith, 439 G.A., Alexander, H., Alm, E.J., Arumugam, M., Asnicar, F., et al. (2019). Reproducible, 440 interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol 441 37, 852-857. 442 443 [6] Buffie, C.G., Bucci, V., Stein, R.R., McKenney, P.T., Ling, L., Gobourne, A., No, D., 444 Liu, H., Kinnebrew, M., Viale, A., et al. (2015). Precision microbiome reconstitution restores 445 bile acid mediated resistance to Clostridium difficile. Nature 517, 205-208. 446 447 [7] Cherrington, C.A., Hinton, M., Pearson, G.R., and Chopra, I. (1991). Short-chain 448 organic acids at ph 5.0 kill Escherichia coli and Salmonella spp. without causing membrane 449 perturbation. J Appl Bacteriol 70, 161-165. 450 451 [8] Choo, J.M., Kanno, T., Zain, N.M., Leong, L.E., Abell, G.C., Keeble, J.E., Bruce, 452 K.D., Mason, A.J., and Rogers, G.B. (2017). Divergent Relationships between Fecal 453 Microbiota and Metabolome following Distinct Antibiotic-Induced Disruptions. mSphere 2. 454 455 [9] Choo, J.M., Leong, L.E., and Rogers, G.B. (2015). Sample storage conditions 456 significantly influence faecal microbiome profiles. Sci Rep 5, 16350. 457 458 [10] Clarke, K.R., and Gorley, R.N. (2006). PRIMER v6: User Manual/Tutorial. PRIMER- 459 E, Plymouth. 460 461 [11] De Vuyst, L., and Leroy, F. (2011). Cross-feeding between bifidobacteria and 462 butyrate-producing colon bacteria explains bifdobacterial competitiveness, butyrate 463 production, and gas production. Int J Food Microbiol 149, 73-80.

464 465 [12] El Aidy, S., Derrien, M., Merrifield, C.A., Levenez, F., Dore, J., Boekschoten, M.V., 466 Dekker, J., Holmes, E., Zoetendal, E.G., van Baarlen, P., et al. (2013). Gut bacteria-host 467 metabolic interplay during conventionalisation of the mouse germfree colon. The ISME 468 journal 7, 743-755. 469 470 [13] Ericsson, A.C., Personett, A.R., Turner, G., Dorfmeyer, R.A., and Franklin, C.L. 471 (2017). Variable Colonization after Reciprocal Fecal Microbiota Transfer between Mice with 472 Low and High Richness Microbiota. Front Microbiol 8, 196. 473 474 [14] Falony, G., Verschaeren, A., De Bruycker, F., De Preter, V., Verbeke, K., Leroy, F., 475 and De Vuyst, L. (2009). In vitro kinetics of prebiotic inulin-type fructan fermentation by 476 butyrate-producing colon bacteria: implementation of online gas chromatography for 477 quantitative analysis of carbon dioxide and hydrogen gas production. Applied and 478 environmental microbiology 75, 5884-5892. 479 480 [15] Fushinobu, S. (2010). Unique sugar metabolic pathways of bifidobacteria. Biosci 481 Biotechnol Biochem 74, 2374-2384. 482 483 [16] Galipeau, H.J., McCarville, J.L., Huebener, S., Litwin, O., Meisel, M., Jabri, B., Sanz, 484 Y., Murray, J.A., Jordana, M., Alaedini, A., et al. (2015). Intestinal microbiota modulates 485 gluten-induced immunopathology in humanized mice. Am J Pathol 185, 2969-2982. 486 487 [17] Gillilland, M.G., 3rd, Erb-Downward, J.R., Bassis, C.M., Shen, M.C., Toews, G.B., 488 Young, V.B., and Huffnagle, G.B. (2012). Ecological succession of bacterial communities 489 during conventionalization of germ-free mice. Applied and environmental microbiology 78, 490 2359-2366. 491 492 [18] Gopalakrishnan, V., Spencer, C.N., Nezi, L., Reuben, A., Andrews, M.C., Karpinets, 493 T.V., Prieto, P.A., Vicente, D., Hoffman, K., Wei, S.C., et al. (2018). Gut microbiome 494 modulates response to anti-PD-1 immunotherapy in melanoma patients. Science 359, 97-103. 495 496 [19] Husebye, E., Hellstrom, P.M., Sundler, F., Chen, J., and Midtvedt, T. (2001). 497 Influence of microbial species on small intestinal myoelectric activity and transit in germ-free 498 rats. Am J Physiol Gastrointest Liver Physiol 280, G368-380. 499 500 [20] Korach-Rechtman, H., Freilich, S., Gerassy-Vainberg, S., Buhnik-Rosenblau, K., 501 Danin-Poleg, Y., Bar, H., and Kashi, Y. (2019). Murine Genetic Background Has a Stronger 502 Impact on the Composition of the Gut Microbiota than Maternal Inoculation or Exposure to 503 Unlike Exogenous Microbiota. Applied and environmental microbiology 85. 504 505 [21] Le Roy, T., Debedat, J., Marquet, F., Da-Cunha, C., Ichou, F., Guerre-Millo, M., 506 Kapel, N., Aron-Wisnewsky, J., and Clement, K. (2018). Comparative Evaluation of 507 Microbiota Engraftment Following Fecal Microbiota Transfer in Mice Models: Age, Kinetic 508 and Microbial Status Matter. Front Microbiol 9, 3289.

509 510 [22] Lupp, C., Robertson, M.L., Wickham, M.E., Sekirov, I., Champion, O.L., Gaynor, 511 E.C., and Finlay, B.B. (2007). Host-mediated inflammation disrupts the intestinal microbiota 512 and promotes the overgrowth of Enterobacteriaceae. Cell Host Microbe 2, 204. 513 514 [23] Manca, C., Boubertakh, B., Leblanc, N., Deschenes, T., Lacroix, S., Martin, C., 515 Houde, A., Veilleux, A., Flamand, N., Muccioli, G.G., et al. (2020). Germ-free mice exhibit 516 profound gut microbiota-dependent alterations of intestinal endocannabinoidome signaling. J 517 Lipid Res 61, 70-85. 518 519 [24] Nadkarni, M.A., Martin, F.E., Jacques, N.A., and Hunter, N. (2002). Determination of 520 bacterial load by real-time PCR using a broad-range (universal) probe and primers set. 521 Microbiology 148, 257-266. 522 523 [25] Nobel, Y.R., Cox, L.M., Kirigin, F.F., Bokulich, N.A., Yamanishi, S., Teitler, I., 524 Chung, J., Sohn, J., Barber, C.M., Goldfarb, D.S., et al. (2015). Metabolic and metagenomic 525 outcomes from early-life pulsed antibiotic treatment. Nat Commun 6, 7486. 526 527 [26] Papanicolas, L.E., Choo, J.M., Wang, Y., Leong, L.E.X., Costello, S.P., Gordon, 528 D.L., Wesselingh, S.L., and Rogers, G.B. (2019). Bacterial viability in faecal transplants: 529 Which bacteria survive? EBioMedicine 41, 509-516. 530 531 [27] Papanicolas, L.E., Wang, Y., Choo, J.M., Gordon, D.L., Wesselingh, S.L., and 532 Rogers, G.B. (2019). Optimisation of a propidium monoazide based method to determine the 533 viability of microbes in faecal slurries for transplantation. Journal of microbiological methods 534 156, 40-45. 535 536 [28] Pollard, M., and Sharon, N. (1970). Responses of the Peyer's Patches in Germ-Free 537 Mice to Antigenic Stimulation. Infect Immun 2, 96-100. 538 539 [29] Quast, C., Pruesse, E., Yilmaz, P., Gerkin, J., Schweer, T., Yarza, P., Peplies, J., and 540 Glockner, F.O. (2013). The SILVA ribosomal RNA gene database project: improved data 541 processing and web-based tools. Nucleic acids research 41, D590-D596. 542 543 [30] Rausch, P., Basic, M., Batra, A., Bischoff, S.C., Blaut, M., Clavel, T., Glasner, J., 544 Gopalakrishnan, S., Grassl, G.A., Gunther, C., et al. (2016). Analysis of factors contributing 545 to variation in the C57BL/6J fecal microbiota across German animal facilities. Int J Med 546 Microbiol 306, 343-355. 547 548 [31] Ridaura, V.K., Faith, J.J., Rey, F.E., Cheng, J., Duncan, A.E., Kau, A.L., Griffin, 549 N.W., Lombard, V., Henrissat, B., Bain, J.R., et al. (2013). Gut microbiota from twins 550 discordant for obesity modulate metabolism in mice. Science 341, 1241214. 551

552 [32] Rinttila, T., Kassinen, A., Malinen, E., Krogius, L., and Palva, A. (2004). 553 Development of an extensive set of 16S rDNA-targeted primers for quantification of 554 pathogenic and indigenous bacteria in faecal samples by real-time PCR. J Appl Microbiol 97, 555 1166-1177. 556 557 [33] Riva, A., Borgo, F., Lassandro, C., Verduci, E., Morace, G., Borghi, E., and Berry, D. 558 (2017). Pediatric obesity is associated with an altered gut microbiota and discordant shifts in 559 Firmicutes populations. Environ Microbiol 19, 95-105. 560 561 [34] Routy, B., Le Chatelier, E., Derosa, L., Duong, C.P.M., Alou, M.T., Daillere, R., 562 Fluckiger, A., Messaoudene, M., Rauber, C., Roberti, M.P., et al. (2018). Gut microbiome 563 influences efficacy of PD-1-based immunotherapy against epithelial tumors. Science 359, 91- 564 97. 565 566 [35] Shashkova, T., Popenko, A., Tyakht, A., Peskov, K., Kosinsky, Y., Bogolubsky, L., 567 Raigorodskii, A., Ischenko, D., Alexeev, D., and Govorun, V. (2016). Agent Based Modeling 568 of Human Gut Microbiome Interactions and Perturbations. PloS one 11, e0148386. 569 570 [36] Sorg, J.A., and Sonenshein, A.L. (2008). Bile salts and glycine as cogerminants for 571 Clostridium difficile spores. J Bacteriol 190, 2505-2512. 572 573 [37] Staley, C., Kaiser, T., Beura, L.K., Hamilton, M.J., Weingarden, A.R., Bobr, A., 574 Kang, J., Masopust, D., Sadowsky, M.J., and Khoruts, A. (2017). Stable engraftment of 575 human microbiota into mice with a single oral gavage following antibiotic conditioning. 576 Microbiome 5, 87. 577 578 [38] Suchodolski, J.S., Markel, M.E., Garcia-Mazcorro, J.F., Unterer, S., Heilmann, R.M., 579 Dowd, S.E., Kachroo, P., Ivanov, I., Minamoto, Y., Dillman, E.M., et al. (2012). The fecal 580 microbiome in dogs with acute diarrhea and idiopathic inflammatory bowel disease. PloS one 581 7, e51907. 582 583 [39] Sun, L., Zhang, X., Zhang, Y., Zheng, K., Xiang, Q., Chen, N., Chen, Z., Zhang, N., 584 Zhu, J., and He, Q. (2019). Antibiotic-Induced Disruption of Gut Microbiota Alters Local 585 Metabolomes and Immune Responses. Front Cell Infect Microbiol 9, 99. 586 587 [40] Thompson, G.R., and Trexler, P.C. (1971). Gastrointestinal structure and function in 588 germ-free or gnotobiotic animals. Gut 12, 230-235. 589 590 [41] Turroni, F., Peano, C., Pass, D.A., Foroni, E., Severgnini, M., Claesson, M.J., Kerr, 591 C., Hourihane, J., Murray, D., Fuligni, F., et al. (2012). Diversity of bifidobacteria within the 592 infant gut microbiota. PloS one 7, e36957. 593 594 [42] Ubeda, C., Taur, Y., Jenq, R.R., Equinda, M.J., Son, T., Samstein, M., Viale, A., 595 Socci, N.D., van den Brink, M.R., Kamboj, M., et al. (2010). Vancomycin-resistant

596 Enterococcus domination of intestinal microbiota is enabled by antibiotic treatment in mice 597 and precedes bloodstream invasion in humans. J Clin Invest 120, 4332-4341. 598 599 [43] Wampach, L., Heintz-Buschart, A., Hogan, A., Muller, E.E.L., Narayanasamy, S., 600 Laczny, C.C., Hugerth, L.W., Bindl, L., Bottu, J., Andersson, A.F., et al. (2017). 601 Colonization and Succession within the Human Gut Microbiome by Archaea, Bacteria, and 602 Microeukaryotes during the First Year of Life. Front Microbiol 8, 738. 603 604 [44] Wang, L., Christophersen, C.T., Sorich, M.J., Gerber, J.P., Angley, M.T., and Conlon, 605 M.A. (2011). Low relative abundances of the mucolytic bacterium Akkermansia muciniphila 606 and Bifidobacterium spp. in feces of children with autism. Applied and environmental 607 microbiology 77, 6718-6721. 608 609 [45] Wichmann, A., Allahyar, A., Greiner, T.U., Plovier, H., Lunden, G.O., Larsson, T., 610 Drucker, D.J., Delzenne, N.M., Cani, P.D., and Backhed, F. (2013). Microbial modulation of 611 energy availability in the colon regulates intestinal transit. Cell Host Microbe 14, 582-590. 612 613 [46] Zeng, Q., Li, D., He, Y., Li, Y., Yang, Z., Zhao, X., Liu, Y., Wang, Y., Sun, J., Feng, 614 X., et al. (2019). Discrepant gut microbiota markers for the classification of obesity-related 615 metabolic abnormalities. Sci Rep 9, 13424.

616 FIGURES

617

618 Figure 1. Bacterial load in faecal samples following establishment of gut microbiota in

619 recipient germ-free mice. Determination of faecal total bacterial load in recipient mice that

620 received (A) intact native microbiota or (B) microbiota derived from antibiotic-exposed mice

621 as a single (1G) or three rounds (3G) of gavage. Bacterial load was determined using

622 quantitative PCR of the 16S rRNA gene. The top and bottom grey lines denote the maximum

623 and minimum bacterial load of donor faecal samples. Solid lines and dotted lines denote

624 median values and interquartile ranges, respectively. Statistical comparison between groups

625 were performed using the Mann-Whitney test at a level of P< 0.05 for significance.

626

A B 100 100 * 90 90 1G 3G 80 80

* * ) ) * *

70 70

(

60 e 60

a t

50 t 50

e c

40 c 40

e P 30 P 30 20 20 10 10 0 0 4 7 14 21 28 35 42 49 56 63 70 4 7 14 21 28 35 42 49 56 63 70 Day Day 627

628 Figure 2. Representation and relative abundance of donor bacterial taxa in mice that

629 received intact native microbiota. (A) Percentage of donor taxa represented in recipient

630 germ-free mice that received one (1G) or three (3G) rounds or faecal gavage with intact

631 microbiota. Representation of the donor taxa (solid line) and their total relative abundances

632 (dashed line) in the recipient mice at each timepoints were determined. (B) The cumulative

633 detection of donor bacterial taxa percentage) in recipient mice. Solid and dashed lines denote

634 median values, dotted lines denote interquartile ranges. Statistical comparison between

635 groups were performed using the Mann-Whitney test at a level of P< 0.05 for significance.

636

A B 100 100 * 90 90 1G 3G

80 80 )

) 70 70

(

e 60 60

a t

t 50 50 n

n * * * * e

e * * * c

c 40 40

e e

P 30 P 30 20 20 10 10 0 0 4 7 14 21 28 35 42 49 56 63 70 4 7 14 21 28 35 42 49 56 63 70 Day Day 637

638 Figure 3. Representation and relative abundance of donor bacterial taxa and in mice

639 receiving material from antibiotic-exposed mice. (A) Percentage of donor taxa represented

640 in recipient germ-free mice that received one (1G) or three (3G) rounds of faecal gavage with

641 antibiotic-exposed microbiota. Solid lines denote representation of the donor taxa and dashed

642 lines denote their total relative abundance. (B) Cumulative representation of donor bacterial

643 taxa (percentage) in recipient mice. Solid and dashed lines denote median values, dotted lines

644 denote interquartile ranges. Statistical comparison between groups were performed using the

645 Mann-Whitney test at a level of P< 0.05 for significance.

646

A B e

e 0.4 0.4 c

c *

n l

l *

i e

e 0.3 0.3 t

t *

c m

m *

i o

o 0.2 0.2

t m

m 0.1 0.1 h

h 1G o

o 1G

f i

i 3G e

e 3G W W 0.0 0.0 4 7 14 21 28 35 42 49 56 63 70 4 7 14 21 28 35 42 49 56 63 70 Day Day 647

648 Figure 4. Weighted UniFrac distances of the donor microbiota and recipient microbiota.

649 Weighted UniFrac distance between the microbiota of recipient mice that received (A) intact

650 native microbiota or (B) microbiota derived from antibiotic-exposed mice, and their

651 respective donor microbiota composition throughout the 70 day study period. Recipient mice

652 received either one round (1G) or three rounds (3G) of donor microbiota. Statistical

653 comparison between the groups at each timepoint was performed using the Mann-Whitney

654 test (P< 0.05).

A 0.20 B 0.20 1G 1G

3G 3G

i o

o 0.15 0.15

e e

* c

o t

t 0.10 0.10

* e

e *

s i

i 0.05 0.05

D D

0.00 0.00 4 7 14 21 28 35 42 49 56 63 70 4 7 14 21 28 35 42 49 56 63 70 Day Day 655

656 Figure 5. Distance to group centroid of the recipient microbiota. Compositional

657 dispersion of microbiota among recipient mice that received (A) intact native microbiota or

658 (B) microbiota derived from antibiotic-exposed mice were determined based on distances to

659 their respective group centroid throughout the 70 day study period. Recipient mice received

660 either one round (1G) or three rounds (3G) of donor microbiota. Statistical comparison

661 between the groups at each timepoint was performed using the Mann-Whitney test (P< 0.05).

A 0.7

0.6 )

n 0.5

p (

0.4 e

c Bacteroidales 1G

n a

d Bacteroidales 3G n

u 0.3 b

a Lactobacillales 1G

e Lactobacillales 3G

a l

e 0.2 Erysipelotrichaceae 1G R Erysipelotrichaceae 3G

Verrucomicrobiales 1G 0.1 Verrucomicrobiales 3G

0.0 4 7 14 21 28 35 42 49 56 63 70 Day

B 0.04

) 0.03 Coriobacteriales 1G

n o

i Coriobacteriales 3G

r o

p Desulfovibrionales 1G o

r Desulfovibrionales 3G

(

c Mollicutes RF39 1G n 0.02

a Mollicutes RF39 3G

d n

u Bacillales 1G

b a

Bacillales 3G

v i

t Bifidobacteriaceae 1G

a l

e Bifidobacteriaceae 3G R 0.01 Saccharimonadales 1G Saccharimonadales 3G

Anaeroplasmatales 1G Anaeroplasmatales 3G

0.00 4 7 14 21 28 35 42 49 56 63 70 Day

C 0.4 0.3 0.2 0.1

0.010

) Lachnospiraceae 1G

n o

i Lachnospiraceae 3G t

r 0.009

o Ruminococcaceae 1G

p o

r 0.008 Ruminococcaceae 3G

(

e Clostridaceae 1G

c 0.007

n Clostridaceae 3G

a d

n 0.006 Family XIII 1G u

b Family XIII 3G

e 0.005

v Peptococcaceae 1G

i t

a Peptococcaceae 3G l

e 0.004

R Peptostreptococcaceae 1G 0.003 Peptostreptococcaceae 3G Eubacteriaceae 1G 0.002 Eubacteriaceae 3G 0.001

0.000 4 7 14 21 28 35 42 49 56 63 70 Day 662 663 Figure 6. Relative abundance of donor taxa in recipient mice receiving the native

664 microbiota. Donor taxa observed in recipient that received one or three gavages of intact

665 native microbiota were plotted at the order level based on (A) high relative abundance taxa

666 (>0.03 relative abundance), (B) low relative abundance taxa (<0.03 relative abundance). (C)

667 Bacterial taxa within the Clostridiales order were plotted at the family level. Recipient mice

668 received either one round (1G) or three rounds (3G) of donor microbiota. Solid and dotted

669 lines denote the mean ± SEM values.

A 0.8 Lachnospiraceae 1G B 0.10 Peptostreptococcaceae 1G Lachnospiraceae 3G 0.08 Peptostreptococcaceae 3G Bacteroidales 1G 0.06 Anaeroplasmatales 1G Bacteroidales 3G

Anaeroplasmatales 3G )

) 0.04 n n 0.6 Verrucomicrobiales 1G

o 0.04

o i

i Ruminococcaceae 1G

t r

r Verrucomicrobiales 3G o

o Ruminococcaceae 3G

p (

( Clostridiales vadin BB60 1G

e 0.03 Clostridiales vadin BB60 3G

n n

a 0.4

a d

d Staphylococcus 1G

n u

u Staphylococcus 3G

0.02

l e

0.2 e

R R 0.01

0.0 0.00 4 7 14 21 28 35 42 49 56 63 70 4 7 14 21 28 35 42 49 56 63 70 Day Day 670 671 Figure 7. Relative abundance of donor taxa in recipient mice receiving the antibiotic-

672 exposed microbiota. Donor taxa observed in recipient that received one or three gavages of

673 the antibiotic-exposed microbiota were plotted at the order level, except for bacterial taxa in

674 the Clostridiales order, which were plotted at the family level. Bacterial taxa were plotted

675 according to (A) high relative abundance (>0.03 relative abundance) and (B) low relative

676 abundance taxa at the order level (<0.03 relative abundance). Recipient mice received either

677 one round (1G) or three rounds (3G) of donor microbiota. Solid and dotted lines denote the

678 mean ± SEM values.

679

680 Supplementary Figure 1. Study design for the establishment of gut microbiota in germ-

681 free mice. Germ-free mice (n=7 per group) received either one (1G) or three (3G) rounds of

682 anaerobically-prepared caecal suspension containing approximately 105 bacterial cells. Faecal

683 pellets were collected from individual mice throughout the experiment.

684

685 Supplementary Figure 2. Alpha diversities and phylum-level relative abundances of

686 donor microbiota. Alpha diversity measures of (A) observed OTUs and (B) Pielou’s

687 evenness of the native and antibiotic-exposed microbiota of the donor were determined using

688 QIIME2. (C) Relative abundances of phyla (relative abundance >1%) of the native and

689 antibiotic-exposed donor microbiota. The median values are plotted and the error bars

690 represent the interquartile ranges. Statistical comparisons were performed using the Mann-

691 Whitney test at a level of P< 0.05 (denoted as asterisk).

692

693 Supplementary Figure 3. Relative abundance of donor bacterial taxa in recipient mice

694 following faecal gavage. Relative abundance of donor taxa observed in recipient mice that

695 received three rounds of faecal gavage with intact native microbiota. Bacterial taxa detected

696 at all nine sampled timepoints following faecal gavage were categorised as consistent (C),

697 while the remaining taxa that were not observed at all timepoints were categorised as

698 intermittent (I). The middle and error bars indicate the median and interquartile ranges,

699 respectively. Statistical comparison between the relative abundance levels of consistent and

700 intermittent taxa were performed using the Mann-Whitney test at a level of P< 0.05 for

701 significance.

702

A B 1.0 1.0 0.8 Firmicutes 1G 0.9

0.6 Firmicutes 3G

) )

n 0.8

n 0.4 Bacteroidetes CON 1G

t r

r Bacteroidetes CON 3G

0.2 o

o 0.7 p

p 0.20

o r

r Actinobaacteria 1G

p p ( 0.6 ( 0.18 Firmicutes 1G

Actinobacteria 3G

e c

c 0.16 Firmicutes 3G n

n 0.5 a

a Proteobacteria 1G d

d 0.14 n

n Proteobacteria 3G Bacteroidetes 1G u u 0.4

0.12 b b

a Bacteroidetes 3G

Patescibacteria 1G e

e 0.10

v i

i Patescibacteria 3G 0.3 t

t Verrucomicrobia 1G a

a 0.08

l e e Verrucomicrobia 3G

Tenericutes CON 1G R 0.2 R 0.06 Tenericutes CON 3G 0.04 0.1 Tenericutes 1G 0.02 Verrucomicrobia CON 1G Tenericutes 3G 0.00 Verrucomicrobia CON 3G 0.0 4 7 14 21 28 35 42 49 56 63 70 4 7 14 21 28 35 42 49 56 63 70 703 704 Supplementary Figure 4. Relative abundance of donor taxa at the phylum level in

705 recipient mice. Relative abundances of bacterial phyla (present in donor at a relative

706 abundance of >0.01) in recipient mice that received one (1G) or three gavages (3G) of the (A)

707 native or (B) antibiotic-exposed microbiota. Solid and dotted lines denote the mean ± SEM

708 values.

A B

i h

h 0.20 0.24 1G

p i i 1G

n 0.22 n

0.18 i

i 3G c

c 3G 0.20 u u 0.16

0.18

]

)

) 0.14 a

a 0.16

s n

n 0.12 0.14

0.10 0.12

S r

r 0.10

e 6

e 0.08

0.08

- k

- 0.06 A

A 0.06

(

0.04 T

T 0.04 C

C 0.02 0.02



- [

[ 0.00 0.00

^ ^

2 4 7 14 21 28 2 4 7 14 21 28 Day Day

C D

. p

0.10 p 0.10 p

1G p 1G

3G m 0.08 3G m

u 0.08

]

)

n n

a 0.06 0.06

f i

0.04 6 0.04

(

T 0.02 T

C 0.02



[

^ ^

2 0.00 0.00 2 4 7 14 21 28 4 7 14 21 28 Day Day

E 0.50 F 1.0 1G 1G

0.45 .

. 0.9 3G p

3G p p

0.40 p 0.8

]

a )

0.35 )

a 0.7

n n

0.30 u 0.6

B 0.25

B 0.5

(

6 T

0.20 6

T 0.4

- -

 0.15 0.3



[ [

^ 0.10

^ 0.2 2 0.05 2 0.1 0.00 0.0 4 7 14 21 28 4 7 14 21 28 Day Day 709

710 Supplementary Figure 5. Absolute abundance levels of selected taxa in recipient mice.

711 The absolute levels of (A, B) Akkermansia muciniphila (C, D) Bifidobacterium spp. and

712 Blautia spp. (E, F) were determined by quantitative PCR (qPCR) in recipient mice that

713 received one (1G) or three gavages (3G) of the native or antibiotic-exposed microbiota,

714 respectively. Absolute abundances of each bacterial taxa were normalised against the total

715 bacterial load in the sample based on the delta CT values of the target gene and 16S rRNA

716 gene. Solid and dotted lines denote the median value and interquartile ranges, respectively.

717

718

719 Supplementary Figure 6. Faecal pH levels in recipient mice of the native microbiota.

720 Faecal samples of recipient mice of the native microbiota at each timepoint were pooled

721 according to their cages and the pH levels analysed. Recipient mice received either one round

722 (1G) or three rounds (3G) of donor microbiota. Solid and dotted lines denote the mean value

723 and standard deviation of pH levels, respectively.