Carbon amendments in soil microcosms induce uneven response on H2 oxidation activity and microbial community composition

Abstract High-affinity H2-oxidizing bacteria (HA-HOB) thriving in soil are responsible for the most important sink of atmospheric H2. Their activity increases with soil organic carbon content, but the incidence of different carbohydrate fractions on the process has received little attention. Here we tested the hypothesis that carbon amendments impact HA-HOB activity and diversity differentially depending on their recalcitrance and their concentration. Carbon sources (sucrose, starch, cellulose) and application doses (0, 0.1, 1, 3, 5% Ceq soildw−1) were manipulated in soil microcosms. Only 0.1% Ceq soildw−1 cellulose treatment stimulated the HA-HOB activity. Sucrose amendments induced the most significant changes, with an abatement of 50% activity at 1% Ceq soildw−1. This was accompanied with a loss of bacterial and fungal alpha diversity and a reduction of high-affinity group 1 h/5 [NiFe]-hydrogenase gene (hhyL) abundance. A quantitative classification framework was elaborated to assign carbon preference traits to 16S rRNA gene, ITS and hhyL genotypes. The response was uneven at the taxonomic level, making carbon preference a difficult trait to predict. Overall, the results suggest that HA-HOB activity is more susceptible to be stimulated by low doses of recalcitrant carbon, while labile carbon-rich environment is an unfavorable niche for HA-HOB, inducing catabolic repression of hydrogenase.


Introduction
Molecular hydrogen (H 2 ) is an indirect greenhouse gas with an aver a ge atmospheric mixing ratio of 0.53 ppmv (Novelli et al. 1999 ).The biological sink of H 2 is mediated by bacteria possessing highaffinity [NiF e]-hydrogenase , catalyzing the oxidation of H 2 into two protons and two electrons conveyed to the r espir atory c hain for ATP generation (Grinter et al. 2023 ).In gener al, the ener gy potential of atmospheric H 2 is utilized in combination with organic carbon source for mixotrophic survival or growth (Constant et al. 2011, Greening et al. 2014, Liot and Constant 2016 ).Such div erse tr ophic ca pabilities ar e ac hie v ed thanks to the gr eat div ersity of H 2 oxidizing bacteria (HOB) with r epr esentativ es involv ed in v arious biogeoc hemical cycles (Sønder gaard et al. 2016 , Gr eening andGrinter 2022 ).High-affinity HOB (HA-HOB) ar e une v enl y distributed among different taxonomic groups with a prevalence of Actinomycetota according to metagenomic surveys and characterization of environmental isolates (Constant et al. 2011, Greening et al. 2016 ).The gr eat div ersity and v ersatility among HA-HOB makes the identification of hallmark conserved traits complicated.Distribution of HOB along organic concentration gradients in desert soil attributed them an oligotrophic life history strategy (Li et al. 2023 ).This is consistent with the involvement of HA-HOB in the degradation of recalcitrant organic matter in soil (Piché-Choquette and Constant 2019 ).
Despite the ubiquitous nature of HA-HOB in the environment (Ji et al. 2017, Jordaan et al. 2020, Bay et al. 2021, Lappan et al. 2023 ), land-use type is a significant driver of the biological sink of atmospheric H 2 , with lo w er activities in a gr oecosystems in comparison to forests (Ehhalt and Rohrer 2009 ).Soil water content, temper atur e , snow co ver, and net primary production were proven efficient to explain spatial and temporal variation of the biological sink of atmospheric H 2 at the global scale (Hauglustaine and Ehhalt 2002, Ehhalt and Rohrer 2009, Morfopoulos et al. 2012 ).At the laboratory scale, Khdhiri et al. ( 2015 ) proposed a linear model parameterized with soil total carbon content explained the variation of H 2 uptake rate measured in contrasting soil samples.Other variables including HA-HOB relative abundance, microbial community composition, alpha diversity and pH were correlated with H 2 oxidation rates (Gödde et al. 2000, Khdhiri et al. 2015 , Saav edr a- Lavoie et al. 2020, Baril et al. 2022 ).
The incidence of different carbohydrate fractions on H 2 oxidation activity and HA-HOB diversity has received little attention.Here, we tested the hypothesis that carbon amendments impact HA-HOB activity and diversity differentially upon their r ecalcitr ance and their concentration in an a gricultur al soil.The response of HA-HOB was examined along with compositional changes of bacterial and fungal communities to determine whether bacteria engaged in H 2 oxidation activity have distinctive fate and plasticity in the face of carbon amendment.Soil microcosms were subjected to amendment with sucr ose, starc h, or cellulose displaying different level of recalcitrance (Kögel-Knabner 2002 ).The microbial uptake of labile sucrose is mediated by membr ane tr ansport pr otein follo w ed b y the catabolism of the disaccharide into fructose and glucose (Reid and Abratt 2005 ).In contrast, cellulose and starch catabolism require extracellular enzymes, with the former being more recalcitrant than the second due to longer structure (Mizuta et al. 2015 ).The addition of cellulose was expected to promote the growth and H 2 -oxidation acti vity because Actinom ycetota ar e pr olific pr oducers of cellulases (Berlemont and Martiny 2013 ) and are among the most abundant HA-HOB in soil (Søndergaard et al. 2016, Ji et al. 2017, Piché-Choquette and Constant 2019 ).Sucrose was expected to inhibit the process in soil due to carbon r epr ession of hydr ogenase activity.

Soil microcosms
The soil for the experiments was collected at INRS, Centre Armand-Fr a ppier Santé Biotec hnologie (Lav al, Québec, Canada) as described by Agoussar et al. ( 2021 ).The soil was sampled in plots located outside ongoing experimental trials.Particle size was homogenized (2 mm sie v e) befor e tr ansferring 10 g soil (dry basis) into 125 ml nominal glass bottles.Each microcosm was amended with either sucrose (Fisher BioReagents FLBP220212), cellulose (Sigma-Aldrich C6429) or starch (J.T.Baker 4006) at differ ent concentr ations (0.1, 1, 3, 5%C eq soil dw −1 ).C eq refers to the proportion of C atoms added per gram of soil, without considering soil bac kgr ound carbon content.Carbon amendments increased by 0.27 to 13 folds the total carbon content in soil.Soil microcosms with no amendment were used as contr ol.Experiments wer e designed in three independent blocks (A, B, and C), resulting in 39 micr ocosms (3 bloc ks x 3 carbon sources × 4 doses + 1 contr ol r epeated three times).A foam plug was inserted to the aperture of microcosms to limit water loss due to evaporation while allowing air exchange during the incubation in the dark at 25 • C for 42 da ys .Soil w ater content w as fixed at 30% water holding capacity.Water loss was compensated on weekly basis by weighting microcosms and adding distilled water.The low water content in soil and the promotion of gaseous exchanges by a large headspace-to-soil volume ratio of 15.7 and a soil layer thickness of 0.5 cm, pr omoting aer obic conditions.H 2 oxidation rates were measur ed weekl y during the incubation period with a gas c hr omatogr a phic assay conducted with an initial H 2 concentration of 482 ± 142 ppbv (Baril et al. 2022 ).H 2 oxidation activity measurements were conducted by integrating the linear decay of H 2 concentrations in the static headspace of soil micr ocosms fr om 5 to 6 observ ations r ecorded within less than 20 min.A rate of 0 nmol H2 g dw −1 h −1 was assigned to non-significant H 2 concentration decay ( Supplementary materials ).Re presentati ve soil subsamples were collected in microcosms after 42 incubation days for total DNA extr action and physicoc hemical anal yses.DN A w as extr acted fr om soil samples (0.25 g) with the DNeasy Po w erLyzer Po w erSoil kit (Qiagen ®).Total C and N were determined using a Thermo Flash 2000 elemental Analyzer equipped with a thermal conductivity detector (Thermo Fisher Scientific, Pittsburgh, USA).Analyses were performed by the "services des Labor atoir es INRS-Centr e Eau Terre Environnement (Canada)".The calibration curve standar d w as generated with BBOT-2,5-Bis(5-tert-butyl-benzoxazol-2yl)thiophene (Thermo Fisher Scientific, Pittsburgh, USA).Soil pH (1 g soil suspension in 10 ml 0.01 M CaCl 2 solution) was measured with an Accumet ® model 15 pH-meter (Thermo Fisher Scientific, Pittsburgh, USA).

Abundance of taxonomic and functional genes
The abundance of bacterial 16S rRNA gene, fungal ITS, hhyL gene encoding the large subunit of group 5/1 h NiFehydrogenase were determined by droplet digital PCR (ddPCR) as described in Baril et al. ( 2022 ).Briefly, three runs of ddPCR were performed with the ddPCR Gene Expression EvaGreen ® Assays (Bio-Rad, Hercules, USA).A random distribution of DNA samples collected was applied and three negative controls with DNA-free sterile water were included.Manual threshold setting was used ( Table S2 ) and fixed to include rain in the positiv e fr action of the partitions (Huggett 2020 ).Only samples with more than 10 000 accepted droplets were used for subsequent analyses.Copy number concentrations were converted to gene copy per gram of dry soil.2011 ).Downstr eam sequence quality contr ol, amplicon sequence variant (ASV) denoising, and taxonomic assignation was done with the R pac ka ge "dada2" v ersion 1.16.0 with default parameters (Callahan et al. 2016 ).For the three genes, a minim um thr eshold of 0.005% r elativ e abundance was attributed to reduce noise in ASV tables (Bokulich et al. 2013 ).Following this cut-off, 1503 (1322438 reads), 293 (1919561 reads) and 648 (426297 reads) ASV were k e pt for 16S rRNA gene, ITS and hhyL anal yses, r espectiv el y.The taxonomic assignation of 16S rRNA gene and ITS region was based on SILVA version 138 (Quast et al. 2012 ) and UNITE database version 7.2 (UNITE Community 2017 ) databases, r espectiv el y.Taxonomic assignation of hhyL gene was not undertaken due to frequent lateral transfer events (Constant et al. 2011 ).Alpha div ersity of eac h gene was expr essed with the extr a polated v alue of the first thr ee Hill numbers r epresenting species richness (q = 0), the exponential function of the Shannon entropy index (q = 1), and the inverse of Simpson index (q = 2).For each microcosm, alpha diversity values wer e extr a polated with the pac ka ge "iNEXT" v ersion 2.0.20 (Chao et al. 2014 ).

Sta tistical anal ysis
Statistical analyses were performed using the software R version 4.1.2(R Core Team 2021 ).The variation in H 2 oxidation rate over the 6 weeks monitoring period was examined with simple linear r egr ession on Ln-tr ansformed data with the pac ka ge "stats" v ersion 4.1.2(R Cor e Team 2021 ).The median of H 2 oxidation r ates measur ed in eac h micr ocosm (8 observ ations ov er 42 days) was utilized to examine the relation between process rates and soil physicoc hemical pr operties as well as micr obial comm unity.The effect of carbon amendment treatments (carbon sources and doses), on median H 2 oxidation r ate, micr obial comm unities (gene copy number and alpha diversity) and soil physicochemical prop-erties (pH, total carbon and total nitrogen) was examined with tw o-w ay ANOVA follo w ed b y a Tuk e y post hoc test using the packa ge "stats" v ersion 4.1.2(R Cor e Team 2021 ).The potential block effect on these variables was tested with one-way ANOVA.Spearman correlation between H 2 oxidation rate and bacteria, fungi, HA-HOB abundance and alpha diversity were verified with the pac ka ge "corr plot" v ersion 0.92 (Wei and Simk o 2021 ).The effect of amendment treatments on beta diversity of soil microbial communities was tested with a Perm utational m ultiv ariate analysis of variance (PERMANOVA) on the Bray-Curtis distance matrix computed on Hellinger-tr ansformed r ead counts, gener ated with the pac ka ge "v egan" v ersion 2.6.4 (Anderson 2014, Oksanen et al. 2020 ).Multile v el pairwise comparisons wer e computed with the pac ka ge "pairwiseAdonis" v ersion 0.4.1 (Arbizu 2017 ).Principal Coor dinates Analysis (PCoA) w as executed with the pac ka ge "phyloseq" version 1.38.0 (McMurdie and Holmes 2013 ) and graphical outputs were generated with the package "ggplot2" version 3.3.6(Wickham 2016 ).The level of rejection was set at α = 0.05 for all statistical tests.

Assignation of carbon preference traits
Assignation of carbon pr efer ence tr aits (C-tr aits) was undertaken at the ASV le v el for the thr ee PCR amplicons utilizing data from soil microcosms exposed to 1% C eq sucr ose, starc h and cellulose.This concentration was chosen owing to the abatement of the H 2 o xidation acti vity by 50% compar ed to contr ol in soil amended with sucrose .T he classification fr ame work comprises thr ee axes, named C-ness , St-ness , and Su-ness defining the pr e v alence of each ASV as a function of cellulose, starch and sucrose soil amendments, r espectiv el y.ASV with less than 3 non-zero observations out of 9 samples were removed from the ASV tables to limit structur al zer o.Bias r egarding libr ary size between tr eatments was verified with Kruskal-Wallis test ( P > 0.11).Centered log-r atio (clr) tr ansformation was a pplied on the ASV tables to address the closed nature of the sequencing data (Quinn et al. 2018 ) and a pseudocount r epr esenting the absolute value of the smallest log-ratio was added to the transformed data to enable the computation of r elativ e abundance of ASV tables.Carbon preference tr aits wer e assigned according to the thr ee axes of the classification fr ame w ork.Those axes w ere obtained b y dividing the av er a ge of the three replicates of an ASV clr value in a treatment by the sum of the ASV clr av er a ge v alues computed in the three treatments .For instance , the C-ness of ASVx in soil amended with carbon i (C-ness (x, i) ) is computed with the following equation: Wher e x corr esponds to the av er a ge clr v alue of ASVx observ ed in the biological replicates, and the subscripts i , j , and k denote cellulose, starc h and sucr ose tr eatments, r espectiv el y.The v alue obtained for each trait were comprised between 0 and 1.The equation was applied for each ASV and each treatment (C-ness, St-ness, and Su-ness) to position ASV along the three axes of a ternary plot.The significance of the assigned traits was tested by comparing change in ASV log-ratio between 1% amendment and control with Analysis of Compositions of Microbiomes with Bias Correction (ANCOM-BC2) computed with the pac ka ge "AN-COMBC" version 1.6.3(Lin and Peddada 2020 ).R scripts for the assignation of carbon pr efer ence tr aits ar e av ailable on the GitHub project at https:// github.com/xbaril/ R _ script _ C-traits .

H 2 oxida tion r a tes
H 2 oxidation rate was measured weekly for 42 days to monitor change in activity in response to carbon amendment in soil (Fig. 1 ).
In gener al, measur ements sho w ed a net uptake of H 2 , whereas the absence of a significant uptake or e v en net emission were observed at certain measurement points for sucrose amendments.
The time series of the activity was in a steady state in control ( P > 0.5) and starch ( P > 0.4) microcosms .T he sole treatment that triggered a rise of H 2 oxidation rate during the incubation was 0.1% C eq soil dw −1 cellulose, whereas microcosms with the two highest doses of sucrose (3 and 5%) follo w ed the opposite trend (Fig. 1 ).Distinct temporal series noticed among treatments was supported by a compar ativ e anal ysis of median H 2 oxidation r ates measured in microcosms during the whole incubation.Carbon sources , carbon concentrations , and their interaction explained variation of H 2 oxidation rate (tw o-w ay ANOVA, P < 0.0001), without experimental block effect (one-way ANOVA, P > 0.6).Sucrose amendments had the greatest impact on oxidation rate, with a loss of activity proportional of the sucrose dose (Table 1 ).Median H 2 oxidation rate was significantly different between control and microcosms amended with 1, 3 and 5% C eq soil dw −1 sucrose (Tuk e y, P < 0.002).Under these conditions, up to 92% of the H 2 oxidation activity was loss.

Soil microbial communities
Diversity and abundance of bacterial 16S rRNA gene, fungal ITS and HA-HOB hhyL gene were analyzed to disentangle their contribution in explaining contrasting H 2 uptake activities among soil microcosms .T he variation of the three alpha diversity metrics w as explained b y amendments treatments and doses as well as their interaction for bacteria (tw o-w ay ANOVA, P < 0.0004) and fungi (tw o-w ay ANOVA, P < 0.03), whereas only the species richness of hh yL follo w ed the same pattern (tw o-w ay ANOVA, P < 0.011).Sucr ose tr eatments wer e the main driv er of these patterns, reducing the alpha diversity of bacteria and fungi.These reductions of microbial diversity were correlated with the loss of H 2 o xidation acti vity trigger ed by sucr ose amendments ( Fig. S2 ).For fungi, the loss of species richness was paralleled with a higher abundance of ITS gene copy number ( ρ < − 0.64, P < 0.0001), indicating the enrichment of fungal species under elevated sucrose concentr ations.In contr ast, neither the abundance of bacteria (tw o-w ay ANOVA, P > 0.4) nor the abundance of HA-HOB (tw o-w ay ANOVA, P > 0.06) was significativ el y r elated to amendment type or doses .T he abundance of HA-HOB ho w e v er declined with sucrose dose, with a one-log scale decrease from the 3% C eq soil dw −1 dose onw ar ds when compared to control (Table 1 ).The composition of micr obial comm unities w as dominated b y ASV encompassing the bacterial classes Actinomycetes (16%), Thermoleophilia (16%) and Alpha pr oteobacteria (15%), wher eas fungi wer e essentiall y dominated by Sordariomycetes (67%).The interaction between carbon type and dose explained variation of bacteria, fungi, and HA-HOB comm unity structur e (Table 2 ).The pattern was mainl y driv en by sucr ose tr eatments (Fig. 2 ).The composition of fungal and bacterial communities was distinguishable to a lesser extent between starch and cellulose treatments (PERMANOVA, R 2 > 0.07, P-adj < 0.04).Carbon sources, carbon concentrations, and their interaction explained total C accumulation and increase in C: N ratio in soil (tw o-w a y ANOVA, P < 0.0001).T he incr ease in C: N r atio was observed in cellulose and starch treatments (Tuk e y, P < 0.009) due to their lo w er mineralisation than sucrose.

Carbon preference traits
C-traits of ASV were determined based on their relative abundance across different carbon amendment treatments (Fig. 3 ).Despite their abatement in larger sucrose doses, HA-HOB were less responding to C amendments (6 ASV; 1%) than fungi (22 ASV; 8%) and bacteria (173 ASV; 12%).Bacteria ASV responding significantly to treatment based on Su-ness trait, either positively or negatively, accounted for 95% (165/173).Among them, 113 ASV were positiv el y impacted by sucrose, encompassing different phyla such as Pseudomonadota (66), Actinomycetota (20) and Bacteroidota (20).Five bacterial ASV were positively impacted by the addition of all three carbon sources, which were associated with phyla Pseudomonadota (3) and Actinomycetota (2).In contrast, the three carbon amendments had a negative impact on two Actinomycetota and one Pseudomonadota r epr esentativ e.In total, 52 bacterial ASV were negatively impacted by sucrose amendment with a pr e v alence of Actinomycetota ( 26) and Acidobacteriota (10).Moreov er, fungal comm unity tr aits show a continuum including both  2. Incidence of carbon sources and their doses on the composition of bacterial, fungal and HA-HOB communities .T he contribution of each factor and their interactions was examined with a PERMANOVA.PERMANOVA results for hhyL do not include the 3 and 5% C eq soil dw −1 sucrose tr eatments, as ther e was no amplification of this gene in these tr eatments.extreme and intermediate positions along Su-ness and C-ness axes.In total, 22 fungal ASV wer e significativ el y affected by treatment, almost all (95%) r epr esentant of the Ascomycota phylum (Fig. 4 ).Sucrose was the main treatment responsible for those changes in fungal ASV abundances (68%).Finally, HA-HOB C-traits were aligned along a continuum encompassing extreme and intermediate position along the three axes of the classification scheme (Fig. 4 ).Most significant responses of HA-HOB were related to Suness .Of these , tw o and four w er e negativ el y and positiv el y impacted by sucr ose, r espectiv el y.Significant hhyL ASV sequences were aligned against the NCBI type material database with the Basic Local Alignment Search Tool (BLAST) to identify their closest r elativ es.Four of the six significant ASV wer e closel y r elated to NiFe-hydrogenase genes found in Actinomycetota.

Direct and indirect effects of carbon amendments on H 2 oxidation activity
The addition of carbon in high amounts in soil microcosms is expected to hav e trigger ed a priming effect leading to enhanced degradation of the soil organic matter pool.The intensity of that priming effect varied among the different carbon sources depend-ing on their r ecalcitr ance and their bioavailability determined by their solubility (Blagodatskaya et al. 2009 ).Effects of carbon amendments on fermentation, on catabolic r epr ession of hydr ogenase and on microbial interactions are proposed to explain the variation of H 2 oxidation activity measured in soil microcosms.
H 2 ev olution w as higher than H 2 oxidation rate after one week in the 1 and 3% sucr ose tr eatments, r esulting in a net emission of H 2 (Fig. 1 ).While O 2 diffusion was pr omoted in soil micr ocosms, r a pid sucr ose-induced r espir ation ma y ha v e r esulted in anaer obic conditions within small ( < 2 mm) soil a ggr egates, leading to the production of H 2 by dark fermentation (Dunbar et al. 2023 ).The transient increase in H 2 production in the soil is typically offset by a concomitant increase in HOB activity (Piché-Choquette et al. 2018 ); this is the case for the high amount of H 2 emitted by nitr ogen-fixing nodules, whic h is consumed in the first few centimeters surrounding the nodule (La Favre and Focht 1983 ).The net H 2 production during the first week of incubation did not result in a sufficient stimulation of the H 2 uptake rate to compensate.A first explanation of the decr easing tr end in net uptake of H 2 with high doses of sucrose therefore is an elevation of the compensation point (H 2 concentration in the static headspace when both oxidation and production processes are in equilibrium) caused by fermentation metabolism (Conrad 1994 ).
As a labile source of carbon, sucrose was the most influential on microbial community composition and H 2 -oxidation activity.The disaccharide is readily available to microorganisms thr ough permeases, phosphotr ansfer ase and ABC tr ansporter systems channeling the sugar to subsequent gl ycol ysis metabolism (Reid and Abratt 2005 ).Energy generation changes the energy status in the cell, leading to pleiotr opic alter ation of gene expression, including an abatement of hydrogenase activity (Friedrich et al. 1981 , Eberz andFriedrich 1991 ).Such a catabolic repression has been observed in various HOB isolates, including the model Ralstonia eutropha H16 switching between lithoautrotrophic and heter otr ophic gr owth (Sc hwartz et al. 2009 ) and the HA-HOB Mycobacterium smegmatis using a mixotrophic growth strategy (Berney andCook 2010 , Greening et al. 2014 ).In Streptomyces spp., hydrogenase activity is expressed in spores generated for long-term persistence and facilitating dissemination (Constant et al. 2010 ).Catabolic r epr ession of hydr ogenase activity in HA-HOB affiliated to Streptomyces can be related to spore germination or F igure 2. Principal Coor dinates Analysis (PCoA) for bacterial, fungal and HA-HOB communities in soil amended with differ ent doses of sucr ose, starc h, or cellulose.sporulation inhibition following the increased availability of labile carbon (Ensign 1978 ).H 2 oxidation activities measured in soil micr ocosms potentiall y involv e v arious HA-HOB encompassing Actinomycetota, Pseudomonadota, Chloroflexota, Bacteroidota, and Acidobacteriota (Søndergaard et al. 2016 , Piché-Choquette and Constant 2019 ).The r elativ e contribution of each phylum to the biological sink of H 2 remains to be determined, but carbon repression is likely a conserved response among HA-HOB displaying mixotrophic metabolism.This response was unique to sucrose amendments, the most labile carbon source integrated in soil micr ocosms.As r ecalcitr ant carbohydr ates , cellulose , and starch exerted less incidence on microbial communities and H 2 oxidation activity.The metabolism of theses carbon sources r equir es extr acellular enzymes, hydr ol ysing pol ymers in r educed sugar tr anslocated into the cell by dedicated transport systems.On the basis of the sa pr ophyte metabolism of HA-HOB in soil (Piché-Choquette and Constant 2019 ), starch and cellulose were expected to promote H 2 oxidation activity.The process was only stimulated by 0.1% cellulose amendment and did not differ from the control after other starch or cellulose treatments (Fig. 1 ).Nitrogen limitation in soil amended with carbon may account for the lack of stimulation of H 2 uptake.
The presumed catabolic repression of H 2 oxidation activity induced by sucrose paralleled with the establishment of a suitable nic he for copiotr ophic micr oor ganisms.Suc h a tr ansition in ecological traits was reported in a previous work demonstrating the enrichment of fast growing organisms (r-strategists) to the detriment of K-strategists in glucose-amended soils (Blagodatskaya et al. 2009 ).The pr oportion of ASV r esponding to the tr eatment was higher for bacteria and fungi than for HA-HOB genotypes.Unsuccessful PCR amplification of hhyL gene for sequencing libr ary pr epar ation and the r eduction in hhyL gene copy number in the 3% and 5% C eq soil dw −1 sucr ose tr eatments indicate a less favor able nic he for HA-HOB.This was also observed in extreme oligotr ophic envir onment wher e labile carbon amendment favor ed copiotr ophic tr aits at the expense of HA-HOB activity (Li et al. 2023 ).Comm unity shifts favoring competitiv e tr aits with labile carbon availability could also trigger antagonistic interactions with HA-HOB or disrupt beneficial interactions (Wood et al. 2023 ).This inter pr etation is compatible with the decline of bacterial and fungal alpha diversity correlated with loss of H 2 oxidation activity in soil microcosms ( Fig. S2 ).The role of microbial interaction on the distribution and activity of trace gas oxidizers has ho w e v er r eceiv ed little attention.Supporting experimental evidence has There is no similar study for HA-HOB, impairing mechanistic insights driving their incompatibility with copiotrophic ecological niches.

Assignation of C-trait
The method used in this work allowed the assignation of C-traits to bacteria, fungi, and HA-HOB at the ASV le v el.A c hallenge experienced in the classification scheme was the inherent problem of the closed nature of the sequencing data (Quinn et al. 2018 ).
Coordinates of the ASV inside the fr ame w ork w ere calculated on the basis of transformed relative abundance, leading to potential inter pr etation bias to w ar d alteration of absolute abundance of genotypes .P osition of ASV along the thr ee axis is ther efor e a qualitative assignment that must be taken carefully.The application of ANCOM-BC integration estimation of absolute abundance allo w ed to test the significance of distribution patterns of ASV in microcosms amended with carbon compared to the control.This a ppr oac h enabled a more stringent assignation of carbon preference traits to a few ASV (from 1% to 12%).The centered position of the triple-positive ASV for cellulose, starch, and sucrose within the fr ame w ork sho ws a gr eement between the ANCOM-BC r esults and the fr ame work computation r esult (Fig. 3 ).The method pr esented in this study was assessed using specific C-traits as part of a case study for a single soil.Application of the method in more soils and as a fr ame work for examining other pertinent traits related to HA-HOB, (e.g .affinity for H 2 ) is expected to impr ov e pr edictability of H 2 oxidation process in response to environmental changes.
Classifying organisms according to their traits, or life strategies, has long been a goal for ecologist.The work of Ho et al. ( 2013 ) was pioneering by attributing the three life strategies developed for plants to methanotrophic bacteria on the basis of their distribution in various en vironments .T he ecological traits of HA-HOB are not extensiv el y documented.Ther efor e, observ ed bacterial and fungal C-tr aits wer e emplo y ed as a r efer ence for defining the sensitivity of HA-HOB related to the other members of soil microbial communities .T he elaboration of a C-trait classification framew ork w as not ac hie v ed at the taxonomic le v el.Despite the loss of H 2 o xidation acti vity with sucr ose, the most r esponsiv e HA-HOB ASV wer e pr edominantl y favor ed by the sugar.Such a decoupling between traits and process rates also applies C-trait at the taxonomic le v el.Nearl y all phyla wer e r epr esented b y ASV sho wing C-trait encompassing extremes and intermediate positions in the classification sc heme.Conserv ed C-tr aits at the phylum le v el was onl y observ ed for Acidobacteriodota and Basidiomycota that responded negativ el y to sucr ose .T his response is in line with their assignment of oligotrophic life strategies (Fierer et al. 2007, Yao et al. 2017 ).Ne v ertheless, the high le v el of v ariation of Su-ness of ASV encompassing Actinomycetota and Pseudomonodota reiterates the difficulty of generalizing ecological traits to the le v el of the entire phylum (Stone et al. 2023 ).The idiosyncratic responses of micr oor ganisms to carbon amendments observ ed her e, combined with the observation of higher species-specific variations of quantitativ e tr aits within a guild when compared to inter-guild (Westoby et al. 2021 ) question the r ele v ance of taxonomic or functional classification system to predict ecological traits of microorganisms.Flexible metabolism and ada ptation ca pabilities of mi-cr oor ganisms impair gener alizations in tr ait attribution.This implies that simplified models assuming monotonous response of HA-HOB to different carbon reservoir components (Li et al. 2023 ) must be taken with caution.Approaches directed to the genotype le v el (e.g.ASV, MAG) would be more efficient to examine adequation between ecological traits and distribution of microorganisms in contrasting environmental conditions.Inclusion of indirect effects triggered by microbial interactions governing the activity of HA-HOB also poses a significant c hallenge.Futur e ecological trait classification fr ame works integr ating abiotic and biotic featur es of extensive soil survey in machine learning environments would definitely contribute to address such a limitation.

Conclusion
This case study aimed to validate the effect of recalcitrant and labile carbon amendment on HA-HOB activity, community structure , and abundance .T he activity was lightl y stim ulated b y lo w doses of more recalcitrant carbon and more markedly inhibited by high doses of labile carbon.The abatement of activity was accompanied by a decrease in hhyL gene copy number and richness.Bacteria and fungi communities sho w ed pattern in their response to the treatments, but these were uneven across taxa.The incor por ation of m ultiple tr aits is the next step to decipher life history strategies and understand the variables governing the activity of HA-HOB in soil.Integration of more complex substrates such as compost and crop residues in the C-trait classification fr ame work and explor ation of nitr ogen limitation on H 2 oxidation activity is also recommended for future in vestigations .

Figure 1 .
Figure 1.H 2 oxidation rate ( u ) for the incubation period ( t ).Dash line r epr esent H 2 oxidation rates measured in the control, and solid line represent logarithmic model.Negative values indicate net emissions of H 2 .Regression equations are only shown for significative models .Circle , triangle and squar e r epr esent the A, B, and C replicates, respectively.

Figure 3 .
Figure 3. Assignation of C-trait for bacteria, fungi and HA-HOB.Point position in the plot refers to the ASV relative log-ratio value in the 1% treatments of sucrose, cellulose and starch, respectively.Significance of the assignation in the classification scheme was established with ANCOM-BC test.Light grey points were not significantly impacted by treatments compared to the control.Black point are ASV that were positively affected by all carbon amendments.Color ed point wer e significantl y impacted by the specified tr eatment compar ed to the contr ol.Negativ el y impacted ASV in all tr eatment (Su-, C-, St-) were not included in the plot.

Figure 4 .
Figure 4. Distribution of C-traits of bacterial and fungal ASV classified at the phylum level, and HA-HOB ASV.

Table 1 .
Median H 2 oxidation rate and soil biotic and abiotic properties for each treatment.: gene copies × 10 8 per gram of soil dw SR: Species richness a