Biogeographical distribution of propionate syntrophs
A total of 6 204 541 high quality sequences of bacterial 16S rRNA genes were obtained from 113 paddy soil samples, which are clustered into 27 411 OTUs. The most dominant OTUs were affiliated to Proteobacteria, Firmicutes, Actinobacteria, Chloroflexi, Acidobacteria, Nitrospirae and Gemmatimonadetes (Additional file 1: Supplementary Figure 1). OTUs that were assigned to four genera of propionate-oxidizing syntrophs (i.e. Syntrophobacter, Pelotomaculum, Smithella and Desulfotomaculum) and those to Syntrophomonas were retrieved for analysis in this study (Additional file 1: Supplementary Figure 2). Syntrophomonas are not known to oxidize propionate but butyrate and up to C10 fatty acids [43]. The inclusion of Syntrophomonas in analysis was due to the fact that Smithella, which utilize the C6 dismutation pathway, can release butyrate as an intermediate product into environment, which is then metabolized by Syntrophomonas [44]. A total of 30 OTUs belonging to five syntroph genera were retrieved.
The total abundance of all five syntroph genera together (referred to synTotal) ranged from 0.01 to 0.34% across 113 soil samples. This low abundance certificates that syntrophs in paddy soils belong to a subset of “rare biosphere” community (defined as individual species or OTUs accounting for <0.1% of the total abundance). Albeit “rare” in relative abundance, the synTotal displayed a distinct geographical distribution, showing the highest abundance in the warm low latitude soils and gradually decreasing towards the high latitude regions (Fig. 1A). The conventional geographical division separates China into the south (low latitude) and the north (high latitude) regions according to the Qinling Mountains-Huaihe River Line (latitude ≈ 32°), a critical geographical boundary for climate, landform and soil conditions [45]. Accordingly, we summarized our soil samples into the low latitude and the high latitude groups, respectively. The mean relative abundance of the low latitude group is significantly greater (by 1.84-fold) than that of the high latitude group (bottom-left of Fig. 1A). In line with the synTotal, the α-diversity of syntrophs, estimated based on OTU richness, significantly decreased with the increase of latitude (Fig. 1B). We then identified the key environmental factors related to the synTotal distribution through Spearman correlation analysis. MAT and TS (total soil S) were identified as the two most important factors followed by NH4+-N, soil OM, MBC and TN (Fig. 1C). Soil CEC and pH showed the negative correlations. The importance index estimated based on Boruta algorithm reiterated that MAT and TS were the two most important factors linking to biogeographical distribution of synTotal (Fig. 1C).
To compare the distribution of different syntrophs, the relative abundances of five syntroph genera were individually analyzed (Additional file 1: Supplementary Figure 3). The distributions of Syntrophobacter, Pelotomaculum, Smithella and Syntrophomonas were in consistence with that of the synTotal (Additional file 1: Supplementary Figure 3A-D). Desulfotomaculum was an exception showing the highest abundance in the middle latitude regions (Additional file 1: Supplementary Figure 3E). Division of the low and high latitude groups revealed that the mean relative abundances of Syntrophobacter, Syntrophomonas and Pelotomaculum were significantly greater in the low latitude regions (Additional file 1: Supplementary Figure 3F), in consistence with the synTotal. Syntrophobacter showed the greatest mean relative abundance followed by Syntrophomonas and Smithella, while Pelotomaculum and Desulfotomaculum showed very low abundances (Additional file 1: Supplementary Figure 3F).
Biogeographical feature of syntroph functioning potentials
Next we evaluated the biogeographical feature of syntroph functioning potentials. For this purpose, fresh soil samples were incubated under anaerobic conditions with addition of 10 mM propionate. We observed the CH4 production until >90% of the added propionate consumed. Three distinct patterns of methanogenesis were identified according to the time lapse required for CH4 production from propionate oxidation (Fig. 2). The pattern I comprised 45 soil samples with the time lapse of 13 d to 27 d, representing the fast rate group of syntrophic metabolisms (Fig. 2A). A greater proportion of soil samples tends to be located at the low latitudes in this group. The pattern II, including 51 samples, represented the median rate of syntrophic metabolisms (28 d to 43 d), which did not show a distinct latitudinal tendency (Fig. 2B). The pattern III comprising 17 soil samples required 44 d to 82 d for CH4 production from propionate oxidation and hence represented the slow rate group. The soils of this group were distributed mainly in the cool high latitude regions (Fig. 2C). The analyses of propionate consumption confirmed the separation of soil samples into three groups (Fig. 3A). Acetate was a major intermediate detected in all samples (Fig. 3B). Butyrate was detected in 28.3% of soil samples (detection limit of 0.05 mM) (Fig. 3C), with the highest concentration occurring in the pattern III soils (Fig. 3C). Taking all soil samples together, the time lapse for methanogenesis from propionate degradation displays an explicit biogeographical pattern, being significantly slower in the high latitude soils than in the low latitude soils (Fig. 4A). The linear least squares regression revealed that the time lapse for methanogenesis was significantly negatively correlated with the relative abundance of synTotal in original soils, MAT, TS, and other edaphic factors including soil OM, MBC, available Fe, Cu and Mn (Fig. 4C). The functional potential of propionate degradation can also be inferred from the maximum rate of methanogenesis (Fig. 4B), which supports the biogeographical tendency revealed by the time lapse (i.e. the shorter the time lapse, the greater the maximum rate).
Shift of microbial community during the anaerobic incubation
At the end of anaerobic incubations, the structure of microbial community in soil samples was revisited. The community composition at the phylum level did not show significant changes between original soil samples and those after anaerobic incubations. The relative abundances of a few phyla, however, changed markedly (Additional file 1: Supplementary Figure 4A). Specifically, the relative abundances of Proteobacteria and Acidobacteria decreased over the incubations while those of Actinobacteria, Firmicutes and Chloroflexi increased. Notably, the α-diversity of bacterial communities at the OTU level showed a substantial decline at the end of incubation compared with original soils (Additional file 1: Supplementary Figure 4B). The β-diversity at the genera level showed the separation of soil samples into two clusters for the low latitude and high latitude soils, respectively (Additional file 1: Supplementary Figure 4C). The co-occurrence network analysis of the top 10% OTUs confirmed that the OTUs for the low and high latitude regions tended to group separately (Additional file 1: Supplementary Figure 4D). These results indicate that the bacterial community shifted significantly during anaerobic incubation, while the separation of metacommunities into the low latitude and the high latitude groups remained robust.
The relative abundance of syntrophs except Desulfotomaculum increased markedly after the incubation with the maximum abundance of synTotal reaching to 12.5%, indicating significant growth. We estimated the increase of individual syntrophs by calculating the logarithmic (log2) fold change (R) of the relative abundance over the incubation. The relative abundance of Desulfotomaculum, which was low at the beginning, did not change over the incubation, indicating no response to propionate. Other four genera showed significant increases but to different extents (Fig. 5). Pelotomaculum showed the greatest increase, followed by Smithella and Syntrophomonas, while Syntrophobacter exhibited the least increase (Fig. 5F). Strikingly, the relative increases of syntrophs showed an opposite geographic tendency compared with their abundances in original soils (Fig. 2). The log2 R values for Syntrophobacter, Syntrophomonas and Smithella increased with the increase of latitude (Fig. 5A,B,D). The values for Pelotomaculum also increased with latitude, though showing the highest value at the middle latitude (Fig. 5C). These results indicate that propionate syntrophs in anaerobic incubation grew to a greater extend in soil samples from the high latitudes than those from the low latitudes.