Soil microbial legacies differ following drying-rewetting and freezing-thawing cycles

Climate change alters frequencies and intensities of soil drying-rewetting and freezing-thawing cycles. These fluctuations affect soil water availability, a crucial driver of soil microbial activity. While these fluctuations are leaving imprints on soil microbiome structures, the question remains if the legacy of one type of weather fluctuation (e.g., drying-rewetting) affects the community response to the other (e.g., freezing-thawing). As both phenomenons give similar water availability fluctuations, we hypothesized that freezing-thawing and drying-rewetting cycles have similar effects on the soil microbiome. We tested this hypothesis by establishing targeted microcosm experiments. We created a legacy by exposing soil samples to a freezing-thawing or drying-rewetting cycle (phase 1), followed by an additional drying-rewetting or freezing-thawing cycle (phase 2). We measured soil respiration and analyzed soil microbiome structures. Across experiments, larger CO2 pulses and changes in microbiome structures were observed after rewetting than thawing. Drying-rewetting legacy affected the microbiome and CO2 emissions upon the following freezing-thawing cycle. Conversely, freezing-thawing legacy did not affect the microbial response to the drying-rewetting cycle. Our results suggest that drying-rewetting cycles have stronger effects on soil microbial communities and CO2 production than freezing-thawing cycles and that this pattern is mediated by sustained changes in soil microbiome structures.

For ITS2 region cDNA and DNA, raw sequence reads were treated as above. Trimmed sequences were processed using PIPITS workflow [17]. Primer-free sequences were merged using PEAR [18] and quality-filtered using FASTX-Toolkit (Hannon 2010). ITS2 sequences were extracted using ITSx [19] and clustered into OTUs using VSEARCH and a (97% sequence similarity). Representative OTU sequences were chimera-checked using UNITE-UCHIME reference dataset. Taxonomy of chimera-free OTUs was assigned with RDP

Fig. S2
CH4 uptake at 12 o C by soil samples during a week upon rewetting and thawing. A presents the CH4 uptake in Experiment 1, which tested the effect of drying-rewetting versus freezing-thawing. B presents the CH4 uptake in Experiment 2, which tested the legacy of freezing-thawing on drying-rewetting. C presents the CH4 uptake in Experiment 3, which tested the legacy of drying-rewetting on freezing-thawing. 12ºC indicates pre-incubation at 12 o C, while 21ºC indicates pre-incubation at 21 o C. DW indicates drying-rewetting cycle. FT indicates freezing-thawing cycle.

Fig. S3
N2O released by soil samples incubated at 12 o C during a week upon rewetting and thawing. A presents the N2O released in Experiment 1, which tested the effect of dryingrewetting versus freezing-thawing. B presents the N2O released in Experiment 2, which tested the legacy of freezing-thawing on drying-rewetting. C presents the N2O released in Experiment 3, which tested the legacy of drying-rewetting on freezing-thawing. 12ºC indicates pre-incubation at 12 o C, while 21ºC indicates pre-incubation at 21 o C. DW indicates drying-rewetting cycle. FT indicates freezing-thawing cycle. Different superscript letters denote significant different rates from the post hoc test at P < 0.05. Drying-rewetting versus freezing-thawing Supplemental information Meisner et al.

Fig S4
Abundance of 16S cDNA reads (A,C,E) and DNA reads (B,D,F) according to treatments as measured by qPCR. The treatments are described in fig. 1. A,B presents the abundance in Experiment 1, which tested the effect of drying-rewetting versus freezingthawing. C, D presents the abundance in Experiment 2, which tested the legacy of freezingthawing on drying-rewetting. E, F presents the abundance in Experiment 3, which tested the legacy of drying-rewetting on freezing-thawing. 12ºC indicates pre-incubation at 12 o C, while 21ºC indicates pre-incubation at 21 o C. DW indicates a drying-rewetting cycle. FT indicates a freezing-thawing cycle. Different superscript letters denote differences at P < 0.05 for a post hoc test for the mixed model. Drying-rewetting versus freezing-thawing Supplemental information Meisner et al.

Fig. S5
Abundance of ITS cDNA reads (A,C,E) and DNA reads (B,D,F) according to treatments as measured by qPCR. The treatments are described in fig. 1. A,B presents the abundance in Experiment 1, which tested the effect of drying-rewetting versus freezingthawing. C,D presents the abundance in Experiment 2, which tested the legacy of freezingthawing on drying-rewetting. E,F presents the abundance in Experiment 3, which tested the legacy of drying-rewetting on freezing-thawing. 12ºC indicates pre-incubation at 12 o C, while 21ºC indicates pre-incubation at 21 o C. DW indicates a drying-rewetting cycle. FT indicates a freezing-thawing cycle. Different superscript letters denote differences at P < 0.05 for a post hoc test for the mixed model. Drying-rewetting versus freezing-thawing Supplemental information Meisner et al.

Fig. S6
Richness of 16S rRNA (A,C,E) and ITS amplicon profiles (B,D,F) for cDNA according to treatments. The treatments are described in figure 1. A,B presents the richness in Experiment 1, which tested the effect of drying-rewetting versus freezing-thawing. C,D presents the richness in Experiment 2, which tested the legacy of freezing-thawing on dryingrewetting. E,F presents the richness in Experiment 3, which tested the legacy of dryingrewetting on freezing-thawing. 12ºC indicates pre-incubation at 12 o C, while 21ºC indicates pre-incubation at 21 o C. DW indicates a drying-rewetting cycle. FT indicates a freezingthawing cycle. Different superscript letters denote differences at P < 0.05 for a post hoc test for the mixed model.   figure 1. A,B presents the r Shannon diversity index in Experiment 1, which tested the effect of dryingrewetting versus freezing-thawing. C,D presents the Shannon diversity index in Experiment 2, which tested the legacy of freezing-thawing on drying-rewetting. E,F presents the Shannon diversity index in Experiment 3, which tested the legacy of drying-rewetting on freezingthawing. 12ºC indicates pre-incubation at 12 o C, while 21ºC indicates pre-incubation at 21 o C. DW indicates a drying-rewetting cycle. FT indicates a freezing-thawing cycle. Different superscript letters denote differences at P < 0.05 for a post hoc test for the mixed model.   figure 1. A,B presents the richness in Experiment 1, which tested the effect of drying-rewetting versus freezing-thawing. C,D presents the richness in Experiment 2, which tested the legacy of freezing-thawing on drying-rewetting. E,F presents the richness in Experiment 3, which tested the legacy of drying-rewetting on freezing-thawing. 12ºC indicates pre-incubation at 12 o C, while 21ºC indicates pre-incubation at 21 o C. DW indicates a drying-rewetting cycle. FT indicates a freezing-thawing cycle. Different superscript letters denote differences at P < 0.05 for a post hoc test for the mixed model. Drying-rewetting versus freezing-thawing Supplemental information Meisner et al.

Fig. S9
Shannon diversity index of 16S (A,C,E) and ITS amplicon profiles for DNA (B,D,F) according to treatments. The treatments are described in figure 1. A,B presents the r Shannon diversity index in Experiment 1, which tested the effect of drying-rewetting versus freezingthawing. C,D presents the Shannon diversity index in Experiment 2, which tested the legacy of freezing-thawing on drying-rewetting. E,F presents the Shannon diversity index in Experiment 3, which tested the legacy of drying-rewetting on freezing-thawing. 12ºC indicates pre-incubation at 12 o C, while 21ºC indicates pre-incubation at 21 o C. DW indicates a drying-rewetting cycle. FT indicates a freezing-thawing cycle. Different superscript letters denote differences at P < 0.05 for a post hoc test for the mixed model.

Shannon diversity index
Drying-rewetting versus freezing-thawing Supplemental information Meisner et al.

Fig. S10
Partial distance-based redundancy analysis of prokaryotes for DNA on Bray-Curtis dissimilarity using capscale ordination. A is Experiment 1 where we tested if a dryingrewetting (DW) or a freezing-thawing (FT) cycle leave different legacies in the prokaryote community. B is from Experiment 2 where we tested how the different legacies affected the microbial response to a drying-rewetting cycle. C is from Experiment 3 where we tested how different legacies affected the microbial communities after an additional FT cycle. 12ºC indicates pre-incubation at 12 o C, while 21ºC indicates pre-incubation at 21 o C (see Fig. 1). Significance of axes is tested with a permutation test by axis: ** P<0.01; ***p<0.001. Drying-rewetting versus freezing-thawing Supplemental information Meisner et al.

Fig. S11
Partial distance-based redundancy analysis of ITS for DNA on Bray-Curtis dissimilarity using capscale ordination. A is Experiment 1 where we tested if a dryingrewetting (DW) or a freezing-thawing (FT) cycle leave different legacies in the prokaryote community. B is from Experiment 2 where we tested how the different legacies affected the microbial response to a drying-rewetting cycle. C is from Experiment 3 where we tested how different legacies affected the microbial communities after an additional FT cycle. 12ºC indicates pre-incubation at 12 o C, while 21ºC indicates pre-incubation at 21 o C (see Fig. 1). Significance of axes is tested with a permutation test by axis: ** P<0.01; ***p<0.001. Drying-rewetting versus freezing-thawing Supplemental information Meisner et al.

Fig. S12
Response groups of 16S amplicons at DNA level. A) Heatmap of the z-score per OTU per treatment compared to the mean relative abundance. B) Barplot with the relative abundance per phylum per response group. Response groups are labeled with numbers and correspond to the groups in A, 1 is the bottom group, 3 is the top group, the rest is also shown in order. Drying-rewetting versus freezing-thawing Supplemental information Meisner et al.

Fig. S13
Response groups of ITS amplicons at cDNA level. A) Heatmap of the z-score per OTU per treatment compared to the mean relative abundance. B) Barplot with the relative abundance per phylum per response group. Response groups are labeled with numbers and correspond to the groups in A, 1 is the bottom group, 4 is the top group, the rest is also shown in order.