Exploring modes of microbial interactions with implications for methane cycling

Abstract Methanotrophs are the sole biological sink of methane. Volatile organic compounds (VOCs) produced by heterotrophic bacteria have been demonstrated to be a potential modulating factor of methane consumption. Here, we identify and disentangle the impact of the volatolome of heterotrophic bacteria on the methanotroph activity and proteome, using Methylomonas as model organism. Our study unambiguously shows how methanotrophy can be influenced by other organisms without direct physical contact. This influence is mediated by VOCs (e.g. dimethyl-polysulphides) or/and CO2 emitted during respiration, which can inhibit growth and methane uptake of the methanotroph, while other VOCs had a stimulating effect on methanotroph activity. Depending on whether the methanotroph was exposed to the volatolome of the heterotroph or to CO2, proteomics revealed differential protein expression patterns with the soluble methane monooxygenase being the most affected enzyme. The interaction between methanotrophs and heterotrophs can have strong positive or negative effects on methane consumption, depending on the species interacting with the methanotroph. We identified potential VOCs involved in the inhibition while positive effects may be triggered by CO2 released by heterotrophic respiration. Our experimental proof of methanotroph–heterotroph interactions clearly calls for detailed research into strategies on how to mitigate methane emissions.


Introduction
Methane (CH 4 ) is Earth's second-most important greenhouse gas (GHG) after CO 2. The current atmospheric CH 4 concentration of 1.896 parts per million (ppm v ) is the highest in at least 80 000 years (Masson-Delmotte et al. 2021 ) and in recent years, the increase has even accelerated (Fletcher andSchaefer 2019 , Masson-Delmotte et al. 2021 ).Global CH 4 emissions contribute 15%-35% to total global r adiativ e forcing (Masson-Delmotte et al. 2021) of whic h the lar gest part originates fr om a gricultur al (i.e.rice paddies) and natural wetlands (Saunois et al. 2020 ).In fact, the recent increase in atmospheric CH 4 concentrations is likely associated to higher biological CH 4 production and emission from wetlands due to warmer and wetter climate (Peng et al. 2022 ).These biological feedbacks to global w arming, often inv olving micr obiall y driv en biogeoc hemical r eactions, ar e poorl y understood, necessitating further investigations of the controlling factors of the main pr ocesses involv ed (Cavicc hioli et al. 2019, Gedney et al. 2019 ).The only biological sink for CH 4 globall y ar e CH 4 -oxidizing micr obes (Bodelier et al. 2019 ), which exert a profound influence on CH 4 dynamics in v arious envir onments.Because of their specialized lifestyle, utilizing CH 4 as their primary carbon and energy source, they are called methanotrophs.
Methanotr ophs, encompassing div erse taxa ca pable of degr ading CH 4 with and without oxygen (O 2 ), hav e garner ed considerable attention due to their role in mitigating CH 4 emissions from natur al and anthr opogenic sources (Bodelier et al. 2019, Guerr er o-Cruz et al. 2021 ).These microbes employ unique enzymes, including CH 4 monooxygenase (MMO), to oxidize CH 4 to methanol, serving as the foundation for their metabolic activities .T her e ar e two types of MMO: the soluble CH 4 monooxygenase (sMMO) and the particulate CH 4 monooxygenase (pMMO).The sMMO is found in the cytoplasm of methanotrophic bacteria and is notable for its ability to oxidize a br oad r ange of substrates besides CH 4 .It contains a nonheme iron center and operates under a wide range of en vironmental conditions .T he pMMO is embedded in the cellular membr anes of methanotr ophic bacteria.It primaril y oxidizes CH 4 to methanol with high specificity and contains copper centres that are crucial for its catalytic activity .Concurrently , heterotrophic micr oor ganisms, comprising a v ast arr ay of bacteria and fungi, derive their energy from organic carbon substrates , displa ying remarkable metabolic versatility and ecological adaptability across diverse habitats.
Ho w e v er, r ecent r esearc h has unv eiled a complex inter play between methanotrophs and heterotrophic microorganisms.While shedding light on their intricate interactions and their broader ecological significance, this intriguing area of research revealed a spectrum of symbiotic , competitive , and syntrophic relationships.(Stock et al. 2013, Ho et al. 2014, 2017, P admav athy 2017 ).Earlier paradigms predominantly viewed these two functional guilds as operating independently, but mounting evidence suggests intricate metabolic exchanges and mutual dependencies among them (Stock et al. 2013, Ho et al. 2014, Veraart et al. 2018 ).For instance, heter otr ophs can utilize intermediates produced during CH 4 oxidation by methanotrophs or vice versa (e.g.cobalamin, organics, or CO 2 ), thereby fostering a dynamic network of crossfeeding interactions within microbial consortia (Stock et al. 2013, Ho et al. 2014, Yu et al. 2017 ).Understanding the mechanisms underpinning the mutualistic or antagonistic relationships between methanotr ophs and heter otr ophs holds immense significance in elucidating the biogeochemical cycles of carbon and energy fluxes within ecosystems .T hese inter actions potentiall y influence CH 4 consumption r ates, substr ate av ailability, and comm unity structur e, consequentl y impacting ecosystem resilience and functioning (Ho et al. 2017 ).
A potential way of communication between the methanotrophic and heterotrophic guilds, especially in case of environmental physical separation, could be via volatile organic compounds (VOCs) (Audrain et al. 2015, Schmidt et al. 2015, Tyc et al. 2015, Veraart et al. 2018 ).Plants and soil micr oor ganisms in the rhizosphere can produce and release various VOCs such as ethylene, CH 4 , and v olatile terpenes (Diy apoglu et al. 2022, Srikamwang et al. 2023 ).Microbial VOCs are known to act as signalling molecules, facilitating communication, and modulating microbial beha viour (K orpi et al. 2009(K orpi et al. , Weisskopf et al. 2021 ) ), which can even act over cm's distances in soil (Schulz-Bohm et al. 2018 ).T hese VOCs pla y important roles in plant-micr obe inter actions , defence mechanisms , and signalling processes (Ortíz-Castro et al. 2009, Bitas et al. 2013 ), but can also have ecological consequences for biogeochemical cycles (de la Porte et al. 2020 ).
In high CH 4 environments like rice pad d y soils , landfills , or chemoclines of lakes, with a high turnover of CH 4 , VOC-mediated inter actions between methanotr ophs and heter otr ophs can hav e profound influence on CH 4 cycling and emission to the atmosphere.Rice paddies or stratified lakes are major sources of CH 4 due to anoxic conditions that are conducive to the production of CH 4 (Minami and Neue 1994, Cao et al. 1996, Schubert et al. 2010, Fuchs et al. 2022 ).Methanotrophs can help to mitigate these emissions by consuming the CH 4 that is produced, thus reducing the amount that escapes into the atmosphere, but can also play a k e y role in the biogeochemical cycling of carbon and other nutrients (Guerr er o-Cruz et al. 2021, He et al. 2023 ).Ho w e v er, the extent to which interactions with other microbes modulate methanotrophic activity as well as the underlying mechanisms hav e sparsel y been inv estigated.Considering the crucial envir onmental role of methanotrophs, it is therefore necessary to assess and understand the modes of interaction involved.In this context, we aimed to assess the interplay between methanotrophs and heter otr ophs, integr ating v arious anal ytical tec hniques, including proteomics, volatile analysis, and measurements of bacterial growth and CH 4 oxidation.This multidisciplinary approach aims at identifying and disentangle the complex volatolomes and elucidate the impact of the volatolomes on the methanotroph activity (process measurements) and proteome.Volatile analysis was emplo y ed to investigate the production and consumption of specific VOCs during micr obial inter actions.Pr oteomics, a po w erful tool in systems biology, was employed to elucidate the protein expression patterns of methanotrophs and heterotrophs when con-fronted with each other.By comparing the proteomes of these organisms, we identified k e y pr oteins involv ed in CH 4 oxidation, carbon metabolism, and interspecies communication.This allo w ed for a better understanding of the molecular basis of the interactions between methanotrophs and heterotrophs and will shed light on our earlier, more descriptive findings (Veraart et al. 2018 ).The findings of this r esearc h will contribute to our knowledge of CH 4 cycling in various en vironments , ultimately helping to dev elop str ategies for mitigating CH 4 emissions and understanding the ecological roles of methanotrophs and heterotrophs in microbial communities.

Experimental design
In this study, we analysed the interaction of heter otr ophs with our model methanotr oph or ganism ( Methylomonas spp.LL1; de Assis Costa et al. 2021 ), as well as the effect of ele v ated CO 2 concentrations on the LL1 strain, on its activity, volatile production, and growth.In addition to Methylomonas spp.LL1, we used two heter otr ophs ( Microbacterium oxydans H1 and Exiguobacterium undae H5; details about isolation see Veraart et al. ( 2018 ) and Krause et al . ( 2015 ).The heter otr ophs wer e isolated fr om methanotr ophic enric hment cultur es, whic h wer e inoculated with wetland sediment and hence, they co-occur natur all y with methanotrophs.Meth ylomonas spp.LL1 w as also isolated from a w etland soil (Bodelier et al. 2012, 2013, de Assis Costa et al. 2021 ) and was used on the basis of environmental origin.Exiquobacterium was chosen because of the positive effect on methanotrophic growth and activity while Microbacterium was chosen from our collection of wetland enrichment isolates because these strains have been demonstrated in pilot incubation to have an inhibitory effect on methanotr ophic gr o wth (Cor dovez et al. 2018 ).Methylomonas sp.LL1 was incubated either alone or together with one of the heter otr ophs.In addition, we incubated Methylomonas sp.LL1 under ele v ated CO 2 (5%) atmospher e (Vor obe v et al. 2011 ).An ov ervie w is pr ovided in Table 1 .We used a similar incubation a ppr oac h as in a pr e vious study (Veraart et al. 2018 ) in which methanotrophs were grown in nitr ogen miner al salts (NMS) medium, while heter otr ophs wer e incubated in dilute tryptic soy broth (0.1% TSB) medium.In brief, pr eincubations wer e conducted at 25 • C with shaking at 145 r m −1 in the dark using 50 ml Erlenmeyer flasks with either NMS medium for the methanotroph (Whittenbury et al. 1970 ) or TSB medium for the two heter otr ophs (Johnson and Kelso 1983 ).From these pr ecultur es two par allel experiments wer e conducted, one for the measurement of CH 4 uptake rates and the other one for the measurement of volatile production, bacterial growth, and proteomics analyses of the methanotroph.In total the experiment was run for 2 weeks.

CH 4 oxidation assay
CH 4 oxidation was measured using two-compartment Petri dishes (Greiner, catalogue number 635102), one side containing 12.5 ml 0.1% TSB-agar, the other side containing 12.5 ml NMS-agar.Nutrient composition of the agar was as described in Veraart et al. ( 2018 ), but with 15 g l −1 agar added (Bacto agar, for NMS, Merck agar for TSB).In total we prepared 28 plates, with four replicates for each combination (Table 1 ).In treatments which contained heter otr ophs, 50 μl of the r espectiv e liquid pr ecultur e [diluted to optical density at 600 nm (OD 600 ) of 0.5] was spread on the TSB side of the plates.Following, 50 μl of the Methylomonas sp.LL1 precultur e was spr ead on the NMS side of the plates (diluted to OD 600 of  Thermo Fisher Scientific) Software was used to analyse the gas c hr omatogr ams obtained from the GC.CH 4 oxidation rates were calculated using linear r egr ession of headspace CH 4 concentrations r etrie v e betw een 2.5 and 28.5 h, to av oid variation in measur ements dir ectl y after starting incubation and subsequentl y ca pturing maxim um r ates.

Vola tile anal ysis
Volatiles were analysed for all three strains alone, the combinations of Methylomonas sp.LL1 with both heter otr ophs and the combination of Methylomonas sp.LL1 with additional CO 2 .The same treatments as for the CH 4 oxidation measurements were applied (Table 1 ), each with four replicates.Split plates as described above were inoculated with precultures growing in the mid-exponential phase, spread with 50 μl of each heterotroph on one side and spotted with se v en dr oplets of 8 μl of Methylomonas sp.LL1 on the other side.Plates were preincubated for 5 days in gas-tight jars containing 20% CH 4 in the headspace and additional 5% CO 2 in the Methylomonas sp.LL1 + CO 2 treatment.After preincubation, the plates were placed in volatile-tr a pping c hambers (Ho et al. 2011, Veraart et al. 2018 ), closed with a butyl septum through which CH 4 and CO 2 could be injected, and a steel volatile tr a p containing 150 mg T enax T A and 150 mg Carbopack B (Markes International, Ltd., Llantrisant, UK) was inserted (Tyc et al. 2015 ).
Plates containing tr a ps wer e incubated in the dark at 25  et al. 2010, Lankenau et al. 2015, Chong et al. 2018 ).Before the statistical analysis, the data was filtered using Interquartile range (IQR) and normalized by the log transformation with automatic scaling (Ossowicki et al. 2020 ).T he VOCs ra w data can be r eceiv ed upon request due to the size of the dataset ( ∼650 GB).The data displayed in Figs 4 and 5 are normalized and display r elativ e mass intensity values of compounds detected, relative to the empty plates.Compounds detected in the empty plates should also appear in the treatments.Ho w ever, since the 'contaminating/bac kgr ound' compounds are in most cases (see Supplementary Table 6 ) in low intensities present they will be either detected or not.Whether a nonmicr obiologicall y pr oduced compound is detected or not may depend on the experimental handling but also potential degradation by the microbes in the inoculated plates.Especially, when amounts are low this can lead to zer os whic h will not ha ppen in the empty plates .T hese inconsistencies ar e pr obabl y due to a combination of r eplication pr oblems (i.e.degree of contamination is difficult to control in our experimental handling), caused measuring around detection limits in combination with degradation by the microbes in the experiment.

Determination of bacterial growth and proteome analyses
Immediately following the incubation for the volatile experiment, cell material from the methanotroph and the two heter otr ophs wer e harv ested by flushing the plates with 3 ml of NMS medium or TSB medium, r espectiv el y.The cells wer e gentl y scr a ped off using a cell-scr a per, and an aliquot of the resulting cell suspension was diluted for OD 600 measurement.Afterwards the cell suspension of each sample was centrifuged, and the pellets were resuspended and tr ansferr ed to 2 ml vials .T hen all suspensions were centrifuged again and washed twice with 1x PBS after which the cell pellets were frozen using liquid N 2 and stored at −80 • C till further processing for proteomics analysis.Since the growth of Methylomonas sp.LL1 in the combinations with M. oxydans H1 was inhibited by the heter otr oph, this combination was excluded from the proteomics analyses.

Extr action, solubiliza tion, and digestion of proteins
The proteomics sample preparation was done according to a pr e vious study (Hakob y an et al. 2018 ).In brief, the frozen cell pellets were resuspended and sonicated (Hielscher Ultrasound Technology) in 2% sodium deoxycholate buffer (SDC, dissolved in 100 mM ammonium bicarbonate) in the presence of 5 mM tris[2carboxyethyl]phosphine and subsequently incubated at 95 • C for 60 min.The samples were subjected to additional 20 s rounds of sonication after 15 and 30 min of incubation.Upon the SDC treatment, all the protein samples w ere allo w ed to cool follo w ed b y incubation with 10 mM iodoacetamide at 25 • C for 30 min.The resulting crude lysates were used for further digestion as described below.The concentration of proteins in each sample was measured with BCA protein assay kit (Thermo Fisher Scientific) according to the manufacturer's instructions.
For protein digestion step, 50 μg of total solubilized protein fr om crude l ysate samples was used.For in-solution digestion, the samples were diluted to 0.5% of solubilizing (SDC) agent in 100 mM ammonium bicarbonate buffer.LysC (0.5 μg, Wako Chemicals GmbH) was added dir ectl y to the protein extract and incubated for 4 h at 30 • C, follo w ed b y trypsin (1 μg, Promega) digestion overnight at 30 • C. Before LC-MS analysis, traces of SDC were precipitated using 2% trifluoroacetic acid, and all protein digests were desalted using C18 microspin columns (Harvard Apparatus) according to the manufacturer's instructions.

LC-MS/MS analyses, peptide/protein identification, and LFQ quantification
The LC-MS/MS analysis of protein digests was performed on a Q-Exactive Plus mass spectrometer connected to an electr ospr ay ion source (Thermo Fisher Scientific).Peptide separation was carried out using the Ultimate 3000 nanoLC-system (Thermo Fisher Scientific), equipped with an in-house packed C18 resin column (Magic C18 AQ 2.4 μm, Dr. Maisch).The protein digest (1 μg) was first loaded onto a C18 pr ecolumn (pr econcentr ation set-up) and then eluted in backflush mode with a gradient from 98% solvent A (0.15% formic acid) and 2% solvent B (99.85% acetonitrile, 0.15% formic acid) to 25% solv ent B ov er 105 min, continued from 25% to 35% of solvent B up to 135 min.The flow rate was set to 300 nl/ min.The data acquisition mode was set to obtain one highresolution MS scan at a resolution of 60 000 (m/z 200) with scanning r ange fr om 375 to 1500 m/z follo w ed b y MS/MS scans of the 10 most intense ions (Top10 DDA).The dynamic exclusion duration was set to 30 s .T he ion accumulation time was set to 50 ms (both MS and MS/MS).The automatic gain control was set to 3 × 10 6 for MS survey scans and 1 × 10 5 for MS/MS scans.For label-fr ee quantitativ e anal ysis MS r aw files wer e imported into Progenesis (Nonlinear Dynamics, version 2.0) and the output data (MS/MS spectra) were exported in mgf format.MS/MS spectra w ere then sear ched using MASCOT (v.2.5, Matrix Science) against a decoy database of the predicted proteomes from Methylomonas sp .LL1, Exiguobacterium sp .RIT341, and M. oxydans .The protein databases w ere do wnloaded fr om unipr ot and e v aluated for pr otein sequence redundancy.The following searc h par ameters wer e used: full tryptic specificity r equir ed (cleav a ge after l ysine or ar ginine residues); two missed cleavages allo w ed; carbamidomethylation (C) set as a fixed modification; and oxidation (M) set as a variable modification.The mass tolerance was set to 10 ppm for precursor ions and 0.02 Da for fr a gment ions for high energy-collision dissociation (HCD).Results from the database search were imported back to Progenesis to map peptide identifications to MS1 features .T he peak heights of all MS1 features annotated with the same peptide sequence wer e summed, and pr otein abundance was calculated per LC-MS run.Next, the data obtained from Progenesis wer e e v aluated using SafeQuant R-pac ka ge v ersion 2.2.2 (Glatter et al. 2012 ).Hereby, 1% FDR of identification and quantification as well as intensity-based absolute quantification (iBAQ) v alues wer e calculated.
Significantly up-or downregulated proteins were defined as having values of the log 2 ratios of > 1.0 or < −1.0, respectively, with q -value < 0.05.The 75 proteins with the highest order of significance were also visualized on a heat map in which clustering was based on Euclidian distance, using W ard' s algorithm.For visualization of similarity between the different proteomics profiles we used a principal component analysis (PCA) and permutational m ultiv ariate anal ysis of v ariance (PERMANOVA) anal yses was performed to compare the differences between groups based on m ultiv ariate data ( Supplementary Table 7 ).ClusterPr ofiler v ersion 3.12.0 was used to perform functional enrichment analyses based on Ky oto Enc yclopedia of Genes and Genomes (KEGG) pathwa ys .Here , the gene set enrichment analysis (GSEA) function (Subramanian et al. 2005 ) was applied to determine whether an a priori defined set of genes based on KEGG shows statistically significant, concordant differences between the different treatments.

Sta tistical anal ysis of gro wth and CH 4 oxida tion
All statistical analyses were done using R version 4.2.0 (R Dev elopment Cor e Team 2008 ).The CH 4 uptake rate and bacterial gr owth fr om the differ ent tr eatments wer e tested for normality by K olmogoro v-Smirno v test and homogeneity of variance by Levene's test.If necessary, normal distribution was ac hie v ed by logtransformation of the data.Mean differences were assessed using one-way analysis of variance ANOVA follo w ed b y Tuk e y's post hoc test.All le v els of significance wer e defined at P < .05.
Statistical analysis of the volatile data was performed on the resulting peak intensity table using the MetaboAnalyst platform (Chong et al. 2018 ).We filtered peak intensities based on IQR, then log transformed and autoscaled them (mean-centred and divided by the standard deviation of each variable) to obtain a normal distribution.We used one-way ANOVA with Tuk e y's post hoc test, to identify significant differences in peak intensities between samples, for each observed compound.PCA was used to visualize maxim um separ ation between gr oups based on their peak intensities and permutational multivariate analysis of variance (PER-MANOVA) analyses was performed to compare the differences between groups based on multivariate data ( Supplementary Table 7 ).We constructed heat maps of peak intensities of the detected compounds in the samples to visualize differences between treatments, in which clustering was based on Euclidian distance, using W ard' s algorithm.

Data deposition
The MS proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD051964, which can be accessed with Username: r e vie wer_pxd051964@ebi.ac.uk and P assw or d: bkeJRW0q.

Bacterial growth and CH 4 oxidation
Methylomonas sp.LL1 performed best on plates with the addition of CO 2 (OD 600 = 5.3), while no growth was observed for Methylomonas sp.LL1 in combination with M. oxydans H1 (Fig. 1 ; Supplementary Table 1 ).Even though the growth of Methylomonas sp.LL1 together with Exiguobacterium sp.H5 (OD = 2.8) was ∼47% lo w er compared to its growth with CO 2 , it was still 71% higher than Methylomonas sp.LL1 (OD = 0.8) growing alone .T he two heter otr ophs sho w ed a similar gro wth performance, no matter whether they were incubated with or without the addition of Methylomonas sp.LL1.
F igure 1. Cell gro wth of the six differ ent tr eatments based on OD 600 measur ements.When a str ain gr e w alone on the two-compartment a gar plate, only its name is stated.If the strain grew with another strain, the name of the other strain is mentioned in parentheses below the name of the measur ed str ain.Differ ent letters indicate significant differ ences ( P < .05) in the mean cell gr owth of eac h sample.Mean ± SD ( n = 4).See Table 1 for further details on the six treatments.The highest CH 4 uptake rate (Fig. 2 ; Supplementary Table 2 ) of Meth ylomonas sp.LL1, w as observ ed when gr owing together with Exiguobacterium sp.H5 (mean: 275 ppm h −1 ), which was almost 30 times higher as compared to growing alone (mean: 15 ppm h −1 ).The CH 4 consumption rate with additional CO 2 added (mean: 180 ppm h −1 ) tended to be lo w er than the rate within the presence of Exiguobacterium sp.H5.Ho w e v er, due to the high variability between replicate samples this difference was not significant ( P > .05).

Proteomic analyses
Ov er all, we identified 2605 proteins for Methylomonas sp.LL1, 1586 proteins for M. oxydans H1, and 1590 proteins for Exiguobacterium sp.H5.For Methylomonas sp.LL1 we found 115 proteins that wer e upr egulated (log 2 > 1 and P -v alues < .05)when gr own together with Exiguobacterium , while 178 proteins were downregulated ( Supplementary Table 3 ).The incubation with ele v ated CO 2 led to the detection of 40 upr egulated pr oteins and 46 downregulated proteins.After the growth together with Methylomonas , the Figur e 3. PC A and compar ativ e heatma p visualization of the pr oteomics a ppr oac h.(A) PCA plots of the pr oteomics composition between thr ee differ ent tr eatments: (i) Methylomonas sp.LL1 (r ed and triangle symbols), (ii) Methylomonas sp.LL1 + Exiguobacterium H5 (blue and x), and (iii) Methylomonas sp.LL1 with the addition of 5% CO 2 in the headspace (green and cross).(B) Heatmap visualization and clustering analysis of the TOP 75 proteins that significantly differ in their expression between the three treatments shown in the PCA.proteome of Exiguobacterium sp.H5 only revealed eight upregulated and se v en downr egulated pr oteins.Similarl y, the pr oteome of Microbacterium only sho w ed seven upregulated proteins and no downr egulated pr oteins after gr o wth together with Meth ylomonas .Giv en these r esults, it was not possible to identify or predict any pr oteomic inter play betw een the tw o heter otr ophs and Methylomonas .( Supplementary Tables 4 and 5 ).All three Methylomonas sp.LL1 proteomes were clearly separated in the PLS-DA analyses (Fig. 3 A), with the proteome of Meth ylomonas sp.LL1 gro wn under ele v ated CO 2 being more similar to the proteome of Methylomonas sp.LL1 grown alone than in combination with one of the heter otr ophs.Since the incubation of Methylomonas sp.LL1 in combination with M. oxydans H1 sho w ed no gro wth, these samples could not be used for a pr oteomic anal ysis .T he heat maps (Fig. 3 B) of the TOP75 proteins significantly differed ( P < .05) between the treatments.In particular, the proteome of Methylomonas sp.LL1 grown in the presence of Exiguobacterium sp.H5 was remarkably different from those of Methylomonas sp.LL1 grown alone or with CO 2 amendment.A pathway analysis (GSEA) of the differ entiall y expr essed pr oteins sho w ed that in cultures of Meth ylomonas sp.LL1 grown together with Exiguobacterium sp.H5, the CH 4 , fatty acid and pyramidine metabolisms had been downregulated relative to cultures in which Methylomonas sp.LL1 was grown alone (Table 2 ).In Methylomonas sp.LL1 cultures supplemented with additional CO 2 , the expression of genes involved in the biosynthesis of valine, leucine and isoleucine, and the 2-Oxocarboxylic acid metabolism wer e significantl y downr egulated, while genes involv ed in CH 4 and glutathione metabolism, bacterial c hemotaxis, degr adation of aromatic compounds and two-component systems were upregulated r elativ e to cultur es in whic h Methylomonas sp.LL1 was grown alone (Table 2 ).

Negative effects of the volatolome on the growth of Methylomonas
The volatolome differed between all the treatments (Fig. 4 A, Supplementary Table 6 ), with 14 volatiles been identified as significantl y differ ent between tr eatments (Fig. 4 B).Dimethyl disulphide and dimethyl trisulphide wer e onl y pr esent in samples with M. oxydans H1, while cyclohexane was only detected when Methylomonas sp.LL1 w as gro wn alone.In addition, a phenol-like compound and an unknown compound were detected on plates where M. oxydans H1 and Meth ylomonas sp.LL1 w er e gr own together, while these compounds were not present when M. oxydans H1 and Methylomonas sp.LL1 were grown alone.

Positive effects of the volatolome on the growth of Methylomonas
The volatile compounds emitted from plates where the growth of Meth ylomonas w as stimulated w ere not as distinct as those from plates wher e gr o wth w as inhibited (Fig. 5 A).Still, w e observed 13 significantl y differ ent VOCs between the treatments (Fig. 5 B).We found an unknown compound that was only present when Methylomonas sp.LL1 was growing together with Exiguobacterium sp.H5.By contrast, 2,4-di-tert-butylphenol was only detected when the methanotroph was growing in the presence of additional CO 2 .

Discussion
This study explored the mechanisms and effects of the interaction between methanotrophic and heterotrophic bacteria, which occur natur all y together in the envir onment (Stoc k et al. 2013 , The coexistence and interactions of these microbial groups, once consider ed separ ate entities in ecological models , ha ve emerged as pi votal dri vers influencing the fate of CH 4 , a potent GHG, and the flow of carbon in diverse habitats .Methanotrophs , Figur e 5. PC A and heatmap analysis of the volatile analyses from Methylomonas sp.LL1 alone (blue and cross) and in combination with Exiguobacterium sp.H5 (green and triangle) and additional 5% CO 2 in the headspace (light blue and x) as well as from Exiguobacterium H5 alone (red and circle), and an empty Petri dish plate as a negative control (pink and diamond).(A) PCA plots of the volatile composition between the different samples.(B) Heatmap visualization and clustering analysis of all significantly different volatiles between the different samples.by virtue of their unique metabolic capability to oxidize CH 4 , interface dir ectl y with heter otr ophic comm unities r eliant on alternative carbon sources by exuding methanol and other metabolites (Stock et al. 2013 ).The resulting interactions, ranging from cooperativ e m utualism to competitiv e anta gonism, yield a complex network of metabolic exchanges and dependencies (Ho et al. 2016, Veraart et al. 2018 ).Unravelling the dynamics and mechanisms of these relationships holds profound implications for ecosystem functioning, carbon cycling, and global climate regulation.
Next to exchange of primary metabolites.Volatile compounds deriv ed fr om secondary metabolism hav e been shown to play a role in the interaction of several methanotrophic and heter otr ophic species (Veraart et al. 2018, Puri 2019, Kilic 2021 ).We aimed to underpin our earlier investigations (Veraart et al 2018 ) with more mechanistic information by exploring this interaction further using a proteomic approach to observe the impact of these VOCs on the metabolic activities of the methanotrophs and heter otr ophs.We also wanted to investigate whether the effect of an inter action between heter otr ophs and methanotr ophs is mainl y due to the additional CO 2 provided by the heter otr ophs or is actuall y trigger ed by the r elease of VOCs.We obtained v ery different results depending on the interacting species, but also a difference between the addition of solely CO 2 , as proxy for a respiring heter otr oph, and the presence of a real respiring microbe.While growth of Methylomonas sp.LL1 was inhibited in the presence of Microbacterium's v olatolome, gro wth and activity w er e stim ulated when exposed to E. undae .Addition of CO 2 caused a similar stimulation of Methylomonas sp.LL1 on growth and CH 4 uptake, but the accompan ying pr oteomics data of Methylomonas sp.LL1 differed significantly between addition of CO 2 and presence of Exiguobacterium , suggesting different underlying mechanisms of the stimulation observed.

Inhibition of growth of Methylomonas by Microbacterium
We observed a complete inhibition in growth of Methylomonas dir ectl y fr om the beginning of the incubation when exposed to the volatolome of a growing Microbacterium , without direct contact.Volatolome analyses led to the conclusion that sulphur compounds [dimethylsulphide (DMS) and dimethyldisulphide (DMDS)] were the agents inhibiting Methylomonas .Microbacterium strains have been observed to produce DMDS and DMS, eliciting e.g.plant growth promoting effects, like increased root biomass and a more extensive root system as well as increased shoot biomass (Cordovez et al. 2018, Ballot et al. 2023 ).It was suggested that the VOCs emitted by Microbacterium spp.act as signalling molecules, triggering physiological responses in plants that enhance growth and nutrient uptake.DMS and DMDS are w ell-kno wn and ubiquitous bacterial volatiles (Ryu et al. 2020 ) and have also been demonstrated to influence methanotrophic growth and activity, albeit with mixed effects on different methanotrophs (Veraart et al. 2018 ).While Methylobacter luteus showed a stimulation in CH 4 uptake, but no effect on growth, after growing together with the DMS and DMDS producing heterotroph Pseudomonas mandelii , Meth yloc ystis parvus exhibits a contrasting effect with a stimulation in growth, but a decrease in CH 4 uptake.Ho w ever, a complete inhibition of growth and activity as in our study was not observ ed (Ver aart et al. 2018 ).The volatile analyses in this study clearly sho w ed that specifically sulphur compounds like DMS and DMDS wer e pr oduced by P. mandelii and seem to play a role in the interaction.Methylophaga sulphidovorans is able to convert DMS to carbon dioxide and thiosulfate .T his heter otr ophic bacterium, which uses the ribulose monophosphate (RuMP) route for carbon assimilation, can also use sulphide as an additional energy source (Zwart et al. 1996 ).Furthermore, a large variety of bacteria are capable of oxidizing DMS to DMSO, provided an additional carbon source is present (Zhang et al. 1991 ).The oxidation can be carried out for example by the ammonium monooxygenases, with its similarity to the pMMO there could be potentially also methanotrophic oxidization of DMS (Fuse et al. 1998 ).While DMS might be beneficial for certain bacteria, our results show that, at least for Methylomonas , the presence of DMS and DMDS has negative effects.In a recent study analysing the methanotrophic community in landfill cover soils, a negative effect of higher DMS concentrations was observed (Wang et al. 2023 ).The relative abundance of Methylocaldum decreased and the CH 4 oxidation rate was reduced, while at the same time the relative abundance of Methylobacter and Crenothrix increased.In the same study, a metagenomic analysis of these soils demonstrated a reduction in the mmo and mxa genes involved in CH 4 uptake and CH 4 assimilation.Since the growth of Methylomonas was completely inhibited in our study, we could not perform a proteomic analysis to determine possible inhibition of specific metabolic pathways such as sulphur oxidation or reduction.To understand how DMS, DMDS or other sulphur compounds can interact with different bacteria strains needs to be further analysed to understand their impact in the environment.
Ho w e v er, important to k ee p in mind is that in the natural context, ther e ar e m ultiple micr obes ar ound pr oducing and possibl y consuming volatiles which makes that our results of one on one inter action hav e to be put in context.With respect to this, when looking at r ele v ant scales to microbes , e .gsoil pore or aggregate scale, than the numbers of cells r eall y inter acting with eac h other can be very small and e v en one to one (Amelung et al. 2024 ), str engthening the envir onmental r ealism of our results.On the other hand, volatiles in soil can tr av el ov er cm's distance and can e v en ob vious one to one interaction may be influenced from elsewher e (Sc hulz- Bohm et al. 2018 ).

Stimulation of activity of Methylomonas by Exiguobacterium and increased CO 2 concentr a tions
The results of our study clearly sho w ed an increase in growth and CH 4 uptake rates by Methylomonas sp.LL1 growing together with Exiguobacterium undea or by addition of CO 2 solely.To investigate the mechanisms behind these stimulatory effects we conducted a proteomics analysis.In a recent study, growth of Methylocystis parvus was also promoted by additional CO 2 and the presence of heter otr ophs, but not CH 4 uptake rates (Veraart et al. 2018 ).In the same study Methylobacter luteus sho w ed the complete opposite effect, with an increase in CH 4 uptake rates, but a constant gr owth compar ed to the contr ol.In both, Ver aart et al. ( 2018) and the current study the effects of the presence of a heter otr oph or the addition of CO 2 seem to be quite similar.Ho w e v er, the pr oteomics data in this study differ significantly when Methylomonas is grown alone, with the addition of extra CO 2 or together with Exiguobacterium .The latter results in a significantl y differ ent pr oteome of Methylomonas than the first two treatments, leading to the assumption that the stimulation by the volatolome of a heter otr oph is not solely caused by the presence of CO 2 , but probably by other VOCs.
Stimulation of growth by CO 2 maybe caused by fixation using the r eductiv e gl ycine c ycle, a pathw ay which in theory can also be used by some methanotr ophs (Tv eit et al. 2019, Claassens 2021, Nguy en et al. 2021 ).Ho w e v er, in our pr oteomics anal yses the enzymes involved in the reductive glycine cycle did not show differences between the different samples, but all involved enzymes were detected in all samples ( Supplementary Table 3 ).Ne v ertheless, ther e wer e two v ery clear differ ences in the pr oteome analyses in the CH 4 processing pathwa ys .While the sMMO of Methylomonas is upregulated after the addition of CO 2 , it is downregulated growing together with Exiguobacterium .An opposite trend is observed for the RuMP pathway, which is upregulated alongside Exiguobacterium and downregulated with additional CO 2 .This suggests a regulation of pMMO versus sMMO expression, possibly in response to higher CO 2 concentrations and changing redox conditions, similar to the regulation seen with copper.The downregulation of RuMP in the presence of Exiguobacterium might indicate a shift to directing more CH 4 carbon into the electron transport chain for additional energy generation, especially when essential compounds are provided by the heterotroph (Holmes et al. 2018 ).Additionall y, ther e is also a possibility that methanotrophs ( Verrucomicrobia ) can grow as autotrophs together with H 2 as a sole electron source (Mohammadi et al. 2019 ).It was shown that the NiFehydrogenase that is needed for this conversion is also present in our Methylomonas strain (de Assis Costa et al. 2021 ) .Ho w e v er, we did not detect an upregulation of the NiFe-hydrogenase in our proteomics data, even though it was detectable .T here are Methylomona s strains which are capable of fixing CO 2 through the PEP carbo xylase, pyruvate carbo xylase, and acetyl-CoA carbo xylase (Nguyen et al. 2018 ).Howe v er, it seems that our Methylomonas str ain onl y harbours the first two carboxylases but not the latter one.
Remarkabl y, ther e ar e m ultiple studies in rice paddies and grasslands showing that elevated CO 2 (eCO 2 ) concentrations have a positive impact on growth of methanotrophs and the rate of uptake of CH 4 (Yu et al. 2018, Qian et al. 2020, 2022 ).In these studies type I methanotr ophs wer e the benefactors of this eCO 2 compar ed to type II (Liu et al. 2022 ).The authors of these studies hypothesized that there is an indirect effect of CO 2 driving an increase in DOC and r oot gr owth, pr oviding a mor e friendl y envir onment for methanotrophs by an increase in available CH 4 and increase of oxic zones.Ho w e v er, our data point to a direct effect driving the observed results, of which the principal mechanisms of the direct effect of CO 2 on the sMMO expression in Meth ylomonas , w e cannot explain.Speculative, the effect of CO 2 could be through reducing equivalents that the sMMO needs to initiate the first step of CH 4 metabolism.To date the exact source of reducing equivalents for sMMO is still not fully understood, but it is thought to involve flavopr oteins and ir on-sulphur pr oteins that participate in electr on tr ansfer r eactions.It utilizes a cofactor called a dinuclear ir on center, often r eferr ed to as the diir on center, whic h plays a crucial role in the catalytic reaction (Whittington and Lippard 2001 ).The r educing equiv alents gener ated thr ough cellular metabolism ar e utilized by MMO to activate molecular O 2 and initiate the oxidation of CH 4 .There are some studies that show that formate dehydr ogenase (FDH) enzymes, cytoc hr ome c, or succinate dehydr ogenase flavoprotein subunit enzymes are important in the uptake and reduction of CO 2 in the cell (Wan et al. 2016, Ruiz-Valencia et al. 2020, Alothaim 2023 ).Even though we detect these enzymes in the proteome of Methylomonas , they do not show a specific regulation in the incubation with CO 2 .The handling of the samples can influence the proteomics analyses .T he setup of these experiments may introduce bias in the analyses due to the examination of cells in various growth stages within a single colony.So maybe it is a matter of timing, since for example for the FDHs it w as sho wn that this could be used as a precursor for methanol pr oduction fr om CH 4 by methanotr ophic bacteria (Ruiz-Valencia et al. 2020 ), which speeds up the whole first step in CH 4 oxidation, we may have missed this with our sampling a ppr oac h.Besides the indications from other bacterial species there are no studies for methanotrophs about transport or uptake of CO 2 , therefore more studies are needed that specifically focus on this interaction, especially since under natural conditions, CO 2 can reach up to 5% in in soils and especially in rhizosphere environments (Button et al. 2023, Bereswill et al. 2024 ).
The impact of Exiguobacterium on the function and growth of methanotr ophs a ppear ed to exhibit similarities to the effects of CO 2 , as pr e viousl y mentioned.While both stimulate the growth, this effect was more than doubled with the added CO 2 , with the same CH 4 uptake rates .T his also hint into the direction that the CO 2 was dir ectl y used as a carbon source .T he amount of CO 2 r espir ed by Exiguobacterium is pr obabl y nowher e near the amount of CO 2 added dir ectl y.Unfortunatel y, the CO 2 pr oduction and consumption was not measured during the course of the experiment, therefor, the sudden exposure to high and graduall y incr easing CO 2 is differ ent, whic h may lead to differ ent r esponses of which we do not known what consequences it may hav e on methanotr ophic gr o wth and activity.Ho w e v er, our analysis of volatolomics and proteomics data revealed also divergent indications from those associated with CO 2 .While we successfully detected specific VOCs in the volatolome analysis, these compounds were found exclusively in the interaction between Methylomonas and Exiguobacterium , yet their precise identification could not be determined.Ho w e v er, the pr oteomics anal ysis pr ovided insights into the upregulation observed in this interaction, revealing that it was not the sMMO like the effect with CO 2 , but rather the RuMP pathway that exhibited upregulation.The RuMP pathway r epr esents one of the three primary pathways employed by methanotrophic bacteria to assimilate CH 4 (Hanson and Hanson 1996, Chistoserdova and Lidstrom 2013, Chistoserdova and Kalyuzhnaya 2018 ).In this pathway, CH 4 is oxidized to formaldehyde, whic h is subsequentl y conv erted to ribulose-5-phosphate through a series of enzymatic reactions.Ribulose-5-phosphate can then enter the pentose phosphate pathwa y, pro viding the cell with energy and biosynthetic intermediates .T he RuMP pathwa y is widely utilized by methanotrophs and plays a critical role in the global carbon cycle by converting CH 4 into biomass.Ho w e v er, the mechanisms causal of the stimulation of the RuMP pathway thr ough the inter action with Exiguobacterium r emain uncertain.Additionall y, we observ ed an upr egulation of the pyrr oloquinoline quinone (PQQ)-dependent methanol dehydrogenase and an incr ease in PQQ pr oduction.This suggests that the pathway from methanol to formaldehyde, and subsequently to the RuMP pathwa y, ma y be influenced by the growth of both heter otr ophic bacteria and methanotrophs.It is possible that the heterotrophs provide reducing equivalents at an accelerated rate, enabling the methanotrophs to gain an adv anta ge in the initial uptake and oxidation of CH 4 , consequently leading to higher metabolic activity do wnstream.Ho w ever, further investigations are needed to elucidate the precise mechanisms involved, as the optimal sampling time or missing connecting information may have influenced the measurements.

Outlook
The current study explored the interaction between methanotrophic and heterotrophic bacteria, as well as high concentrations of CO 2 , and their effects on the functionality of the methanotr ophs (gr owth, CH 4 oxidation, and carbon cycling).Understand-ing the dynamics of the interaction between methanotrophic and heter otr ophic bacteria can have implications for ecosystem functioning, carbon cycling, and global climate regulation.Methanotrophs play a crucial role in mitigating the effects of CH 4 , a potent GHG, by oxidizing it.The study highlights that the interaction between these bacteria can influence the fate of CH 4 and the flow of carbon in diverse habitats .T he findings suggest that the release of VOCs by heter otr ophic bacteria can affect the growth and CH 4 uptake of methanotr ophs.Differ ent VOCs, suc h as dimethylsulphide and dimethyl-disulphide, were found to have varying effects on differ ent methanotr oph species.Some VOCs stimulated growth and CH 4 uptake, while others inhibited them.These findings indicate that the presence and composition of VOCs in the environment can impact the activity of methanotrophs and, consequentl y, the r ate of CH 4 oxidation.Mor eov er, the study highlights the role of CO 2 in the interaction between bacteria.Elev ated CO 2 concentr ations wer e found to have a positive impact on the growth of methanotrophs and their CH 4 uptake rates .T his suggests that incr eased CO 2 le v els, suc h as those associated with climate change, can potentially enhance the activity of methanotrophs and contribute to the reduction of CH 4 emissions.Overall, the study emphasizes the importance of understanding the interactions between methanotrophic and heterotrophic bacteria in order to comprehend their influence on CH 4 oxidation, carbon cycling, and ultimately, global climate regulation.These insights can inform strategies for mitigating GHG emissions and managing carbon in the environment.

Figure 2 .
Figure 2. CH 4 uptake rate for Methylomonas sp.LL1 incubated alone, in combination with the two heter otr ophs, and with the addition of 5% CO 2 in the headspace.Since the growth of Methylomonas sp.LL1 was inhibited by M. oxydans , no CH 4 uptake rate was measurable.Same letters indicate similarity in CH 4 uptake r ate, differ ent letters indicate significant differences ( P < .05).Mean ± SD ( n = 4.)

Figur e 4 .
Figur e 4. PC A and compar ativ e heatma p visualization of the volatile anal ysis r esults for Methylomonas sp.LL1 incubated (i) alone (blue and cr oss) and (ii) in combination with M. oxydans H1 (green and triangle), as well as for (iii) M. oxydans H1 alone (red and circle) and (iv) an empty Petri dish plate that acted as a negativ e contr ol (light blue and x).(A) PCA plots of the volatile composition between the different treatments.(B) Heatmap visualization and clustering analysis of all the volatiles that significantly differed between the treatments.

Table 1 .
Experimental setup with all used incubation combinations.

Plates compartment 1 NMS agar Plates compartment 2 TSB agar Headspace
with MZMine V 2.36(Copyright © 2005(Copyright ©  -2012MZmine De v elopment Team) to cr eate a m/z and peak intensity table that could be used as input file for MetaboAnalyst 4.0 softwar e ( http://www.metaboanal yst.ca/MetaboAnal yst )(Pluskal • C for 48 h, after which volatile traps were removed, rapidly capped, and stored at 4 • C until further analysis within 2 weeks.Volatile anal yses wer e performed on a Quadrupole Time of Flight GC/MS (hereafter GC-Q-TOF, Agilent 7890B GC, Agilent 7200A/B Q-TOF) as described inVeraart et al. ( 2018 ).For the volatolomics analysis, the acquired raw mass spectrometry (MS) data was extracted to m/z format using MassHunter Qualitative Analysis Software V B.07.00 (Ag ilent Technolog ies, Santa Clara, C A, USA).T he m/z data was processed
1 Upregulated pathways have a positive NES score, while downregulated pathways have a negative score in the first mentioned strain, respectively.Pathways with a P -value < .05 and an FDR q -value < 0.25 are shown. 2NES = normalized enrichment score; 3 NOM = nominal; 4 FDR = false discovery rate.