Effects of Field-Grown Transgenic Cry1Ah1 Poplar on the Rhizosphere Microbiome


 Background: Poplar (Populus) is a genus of globally important plantation trees used widely in industrial and agricultural production. However, poplar is easily damaged by Micromelalopha troglodyta and Hyphantria cunea, resulting in a decline in poplar quality. Due to their strong insect resistance, Bt toxin-encoded Cry genes have been widely adopted in poplar breeding; however, potential adverse effects of Cry1Ah1-modified poplars on the ecological environment have raised concerns. Results: In this study, we comprehensively analyzed the structural and functional composition of the rhizosphere microbiome in field-grown transgenic Bt poplar. Conclusions: Our analysis of soil chemistry patterns revealed that soil alkaline nitrogen, soil available phosphorus, and microbial biomass nitrogen and phosphorus levels were improved, whereas microbial biomass carbon declined in Cry1Ah1-modified poplar rhizosphere samples. We applied metagenomic sequencing of Non-Transgenic (NT) and Cry1Ah1-modified poplar rhizosphere samples collected from a natural field; the predominant taxa included Proteobacteria, Acidobacteria, and Actinobacteria. We also identified microbial functional traits involved in membrane transport, amino acid metabolism, carbohydrate metabolism, and replication and repair in NT and Cry1Ah1-modified poplars. Together, these results demonstrate that the NT and Cry1Ah1-modified poplar rhizosphere microbiomes had similar diversity and structure. These differences in relative abundance were observed in a few genera but did not affect the primary genera or soil.


1
Introduction 86 Poplar (Populus) is a genus of globally important plantation trees used widely in industrial and 87 agricultural production (Boerjan, 2005). However, with the deterioration of the global environment, 88 characterized by increasing salt, drought, pest, and disease stresses, the global production of poplar is 89 becoming challenging. One approach to address this challenge is genetic modification. The 90 manipulation of critical genes has been applied to alter poplar characteristics in transgenic lines,

Effects of CM varieties on M. troglodyta 220
The Bt-Cry1AELISA kit was used to identify the Cry1Ah1 expression level in NT and CM 221 varieties. The results showed that Cry1Ah1 was expressed in CM varieties, and lines A4-6 and A5-0 222 had the higher Cry1Ah1 expression level. In comparison, line A3-4 had the lower Cry1Ah1 expression 223 level (Supplemental Table1). In addition, the insecticidal activities of CM varieties were identified, 224 and the results showed that the CM varieties had higher insecticidal activity to M. troglodyta than NT 225 poplars (Supplemental Figure 2). Significantly, lines A4-6 and A5-0 with higher Cry1Ah1 expression 226 levels exhibited relatively more substantial insecticidal activity than M. troglodyta. 227

Effects of CM poplars on rhizosphere soil chemistry patterns 228
During the first three years of poplar establishment, the mean soil pH ranged from 7.73 to 8.23 in 229 rhizosphere soil ( Figure 1A). Also, there was no significant change in rhizosphere soil pH between NT 230 and CM varieties by sampling date. For the CM varieties (lines A5-0, A4-6, Z1-3, A5-23, and A3-4), 231 rhizosphere soil alkaline nitrogen ranged from 64.54 to 83.15 mg/kg, with similar values observed for 232 NT poplars, and no significant difference between NT and CM varieties ( Figure 1B). However, the 233 CM varieties had significantly lower rhizosphere soil available phosphorus in the field-grown stage 234 than NT poplars ( Figure 1C). The rhizosphere MBC contents of NT poplars ranged from 160 to 172 235 mg/kg and differed significantly from those of the CM varieties ( Figure 1D). In addition, CM varieties 236 had significantly lower MBN and MBP contents than NT poplars ( Figure 1E and F). 237

Data quality control and ASVs analysis 238
Using the Illumina hiseq, an average of 53,567 and 68,783 16S rDNA tags and 33,750 and 87,922 239 ITS1 tags were generated from the rhizosphere microbiome. Chimaeras and short tag sequences were 240 removed to obtain high-quality clean tags comprising an average of 33,390 and 86,787 16S rDNA tags 241 and 21,523 and 22,555 ITS1 tags (Supplemental Table 2). Also, clean tag distributions of rhizosphere 242 bacteria were visualized. The results showed that the lengths of clean tags ranged from 200 to 440 bp. 243 Clean tags with the lengths of 420-440 bp occupied the largest proportion (Supplemental Figure 3A), 244 while clean tag distributions of rhizosphere fungi ranged from 200-360bp, and clean tags with the 245 lengths of 200-260 bp occupied the largest proportion (Supplemental Figure 3B). Using the Qiime ver. 246 2.0 and Vsearch 2.7.1, the chimeric and organelle sequences were removed to produce 10787 247 rhizosphere bacterial community sequencing ASVs and 7,732 fungal community sequencing ASVs 248 (Supplemental Table 3). 249

Rhizosphere bacterial diversity 250
The mothur (Schloss et al., 2009) has been applied to perform ASV rarefaction analysis based on 251 ASV clustering results to construct alpha rarefaction curves and evaluate the putative differences in 252 the alpha diversity. The rhizosphere bacteria sample rarefaction curves illustrated that most NT and 253 CM varieties saturate around 6500-7000 ASVs, suggesting slight differences in the diversity of 254 rhizosphere bacterial community between NT and CM varieties (  The alpha diversity analysis was used to reflect the richness and diversity of rhizosphere bacteria. 262 The Chao1 indexes in NT poplars had highly similar results in CM varieties, including A5-23, Z1-3, 263 and A3-4. At the same time, the Cry1Ah1 expression might increase the community richness of 264 rhizosphere bacteria, which the analysis of Chao1 showed no dominant differences between NT and 265 CM varieties ( Figure 3A and Supplemental Table 4). Also, the observed species in NT poplars had no 266 dominant differences compared to that in CM varieties, except line A4-6 ( Figure 3B and Supplemental 267 Table 4). The PD whole tree in NT and CM varieties shared similar features with observed species 268 ( Figure 3C and Supplemental Table 4). There are no significant differences in the analysis of the 269 Shannon ( Figure 3C, D, and Supplemental Table 4), which suggested that the community diversity of 270 rhizosphere bacteria was similar between NT and CM varieties. Based on the alpha diversity, we 271 concluded that Cry1Ah1 expression slightly influences the rhizosphere bacterial richness but does not 272 affect the community diversity of rhizosphere bacteria. 12 A Bray-Curtis dissimilarity matrix was calculated on normalized and square-root transformed 274 read abundance data to compare the composition of rhizosphere bacterial members between NT and 275 CM varieties. Based on weighted UniFrac, beta diversity analysis with PCA was applied to analyze the 276 bacterial community structures among NT, A5-0, A4-6, Z1-3, and A5-23, and A5-0, A4-6, Z1-3, A5-277 23, A3-4, and NT were overlapped together and could not be separated (Figure 4), which indicated that 278 the community structures of NT and CM varieties were similar. The Cry1Ah1 expression did not affect 279 the bacterial community structures. To determine the effect of Cry1Ah1 expression on the rhizosphere bacteria, we investigated the 283 taxonomic distinctiveness of poplar rhizosphere bacteria. In addition, the DeSeq2 was used to select 284 the putative statistically differential rhizosphere bacteria. The relative abundances of rhizosphere 285 bacteria of NT and CM varieties at the phylum, class, order, family, and genus levels were identified. 286 At the phylum level, the Firmicutes, Myxococcota, Nitrospirota, Sva0485, Fibrobacterota, 287 Latescibacterota, Desulfobacterota, and Proteobacteria considered as the dominant bacteria were found 288 in the rhizosphere bacterial community ( Figure 5A). The DeSeq2 analysis found that the dominant 289 bacteria share similar abundances between NT and CM varieties. In contrast, the relative abundances 290 of Cyanobacteria and Methylomirabilota showed a significant difference between NT and line A3-4 291 ( Figure 6A). Also, the Methylomirabilota, Proteobacteria, Zixibacteria, MBNT15, Dadabacteria, 292 Thermoplasmatota, Cyanobacteria, Chloroflexi, Bacteroidota, Acidobacteriota, Firmicutes, and 293 Myxococcota were present at different abundances between NT and line A4-6 ( Figure 6B). In addition, 294 a few rhizosphere bacteria abundances were the difference between NT and A5-0, A5-23 or Z1-3 295 ( Figure 6C-E). According to the above evidence, we concluded that Cry1Ah1 expression has no 296 influence on most rhizosphere bacteria abundances and only changes a small part of rhizosphere 297 bacteria abundances. 298 At the class level, the rhizosphere bacterial community composition of NT and CM varieties was 299 similar ( Figure 5B). The relative abundances of rhizosphere bacteria had no significant difference 300 between NT and line A3-4 except for Cyanobacteria and Methylomirabilia ( Figure 7A). In native 301 fields, Alphaproteobacteria, Deltaproteobacteria, Betaproteobacteria, Subgroup6, Blastocatellia, and 302 Thermoleophilia, which account for approximately 60% of the total rhizosphere bacteria, were present 303 at similar relative abundances in NT and CM varieties ( Figure 5B). In addition, rhizosphere bacteria 304 with lower relative abundances, such as Actinobacteria, Gemmatimonadetes, Nitrospira, Holophagae, 305 and Acidimicrobiia, were significantly different between NT and line A4-6 ( Figure 7B). Compared 306 with NT, the relative abundances of rhizosphere bacteria in lines A5-0, A5-23, and Z1-3 were similar, 307 and only small parts of rhizosphere bacteria abundances displayed differences ( Figure 7C-E). 308 Furthermore, at the species level, a minor part of rhizosphere bacteria abundances from NT poplars 309 was slightly lower or higher than that from CM varieties, indicating that the CM varieties had little 310 influence on rhizosphere bacteria with lower relative abundances ( Figure 8). 311 We investigated the taxonomic distinctiveness of poplar rhizosphere soil fungi to determine 312 whether Cry1Ah1 expression affected rhizosphere fungus communities. The Chao1 analysis showed 313 that the community richness of rhizosphere fungi in CM varieties shares similar community richness 314 to NT, except line A5-0 (Supplemental Figure 4A). In addition, no significant difference was present 315 at the observed species, PD whole tree, and Shannon between NT and CM varieties. However, the 316 observed species and PD whole tree in line A5-23 had no slight difference compared to that in line Z1-317 3 (Supplemental Figure 4B  more than approximately 80% of the total sequence, whereas that of low-abundance phyla comprised 325 less than 20% of the entire sequence (Supplemental Figure 6A). Except for Ascomycota, there was no 326 significant difference among rhizosphere fungi between NT and CM varieties at the phylum level 327 (Table 1). Based on the analysis of rhizosphere fungal abundance in NT and CM varieties, we 328 concluded that Cry1Ah1 expression has no significant influence on the relative abundances of most 329 rhizosphere fungi and only affects a few rhizosphere fungal abundances. Besides, we filtered extremely 330 rare ASVs from the dataset to determine relative abundances at the class, order, family, and genus 331 levels. Similar relative abundances of most rhizosphere fungi were observed between NT and CM 332 varieties (Supplemental Figure 6B-E). However, the relative abundances of Archaeosporomycetes, 333 Agaricostilbomycetes, Lobulomycetes, Cystobasidiomycetes, Schizosaccharomycetes, 334 Sordariomycetes, and Ascomycota Incertae sedis at the class level were different between NT and CM 335 varieties (Supplemental Table 5). Compared with NT poplars, only 13 kinds of 148 rhizosphere fungi 336 relative abundances in CM varieties differed (Supplemental Table 6). Also, the major rhizosphere fungi 337 at the family level showed similar abundances between NT and CM varieties. In contrast, only 7.9% 338 of rhizosphere fungi with the lower abundance were present at differences between NT and CM 339 varieties (Supplemental Table 7). In addition, most rhizosphere fungi abundances were found to have 340 no differences between NT and CM varieties at the genus level (Supplemental Table 8    varieties at the class level. Data were analyzed using Kruskal-wallis comparison. "P < 0.05" indicates 720 a significant difference between NT and CM varieties. 721 Supplemental Table 6: The relative abundances of rhizosphere fungi at the order level between 722 NT and CM varieties. Data were analyzed using Kruskal-wallis comparison. "P < 0.05" indicates a 723 significant difference between NT and CM varieties. 724 Supplemental Table 7: The relative abundances of rhizosphere fungi at the family level between 725 NT and CM varieties. Data were analyzed using Kruskal-wallis comparison. "P < 0.05" indicates a 726 significant difference between NT and CM varieties. 727 Supplemental Table 8: The relative abundances of rhizosphere fungi at the genus level between 728 NT and CM varieties. Data were analyzed using Kruskal-wallis comparison. "P < 0.05" indicates a 729 significant difference between NT and CM varieties. 730 The English in this document has been checked by two professional editors, both native English 731 speakers (http://www.textcheck.com/certificate/dQEZQj). 732