# Source paper archive

- Source title: Symbiotic and commensal root microbes are key determinants of plant responses to rising CO₂
- Author(s): Marina Soneghett Cotta, Paul Kajdewicz, Lysandros Anastasopoulos, Carlos Duran-Ballesteros, Jianbei Huang, Karl Kübler, Martin Göbel, Susan Trumbore, Paul Schulze-Lefert, Ruben Garrido-Oter
- Source landing URL: https://www.biorxiv.org/content/10.64898/2026.11.07.764780v1
- PDF URL: https://www.biorxiv.org/content/10.64898/2026.11.07.764780v1.full.pdf
- PDF SHA-256: 4103714421b8c7e90c6dde77e8555bb1de1126ed2b3c5c63e36efe45a6e894cf
- Converter: pymupdf4llm 0.0.27
- Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version.

---

[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


Symbiotic and commensal root microbes are key determinants of plant responses to rising


CO₂


Marina Cotta [1], Paul Kajdewicz [1], Lysandros Anastasopoulos [1], Carlos Duran-Ballesteros [1],
Jianbei Huang [2], Karl Kübler [2], Martin Göbel [2], Susan Trumbore [2], Paul Schulze-Lefert [1,3,] &

Ruben Garrido-Oter [1,3,4,]


1 Department of Plant-Microbe Interactions, Max Planck Institute for Plant Breeding
Research, 50829 Cologne, Germany


2 Department Biogeochemical Processes, Max-Planck-Institute for Biogeochemistry, Jena
07745, Germany


3 Cluster of Excellence on Plant Sciences, 40225 Düsseldorf, Germany


4 Earlham Institute, Norwich NR4 7UZ, UK


Corresponding authors. E-mail: schlef@mpipz.mpg.de and garridoo@mpipz.mpg.de


Abstract


Rising atmospheric CO₂ stimulates plant growth, but this effect is often constrained by nutrient
availability. While beneficial soil-derived microbial symbionts such as nitrogen-fixing rhizobia
and mycorrhizal fungi are known to mitigate this limitation, the role of plant commensal
microbiota remains unclear. Here, using a gnotobiotic growth system and the model legume
Lotus japonicus, we show that the magnitude of plant responses to elevated CO₂ is largely
explained by the composition of its root-associated microbial community. Axenic plants
exhibited limited physiological responses to increasing CO₂ but inoculation with symbiotic and
commensal microbes amplified this effect. Commensal bacteria alone, while not increasing
overall biomass, shifted plant carbon allocation belowground. Microbial abundances in roots
increased with atmospheric CO₂, even when normalized to plant DNA. The strongest plant and
microbial responses were observed when rhizobia, arbuscular mycorrhizal fungi, and the
commensal bacterial community were all present. These findings indicate that multi-partner
microbial associations will shape both plant growth and the allocation of carbon belowground
as atmospheric CO₂ concentrations continue to rise.


Introduction


The rising concentration of atmospheric CO 2 is altering the functioning of terrestrial ecosystems
and soil carbon cycles at a global scale (1–4). Elevated CO 2 is known to stimulate plant growth
and increase productivity, a phenomenon which is known as the CO 2 fertilization effect, and
which has long been recognized as a key mechanism by which land vegetation acts as a major
carbon sink (5–7). However, the magnitude of this effect is constrained by nutrient availability,
particularly nitrogen (N) and phosphorus (P), which are essential mineral macronutrients for
producing biomass and, consequently, carbon sequestration. Nutrient limitation can reduce or
even negate the CO₂ fertilization effect (8, 9), posing uncertainty in global carbon cycle models,
with major implications for global carbon budget predictions (10).


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


This macronutrient limitation can be alleviated by associations with symbiotic microbial
partners, notably those involving arbuscular mycorrhizal fungi (AMF) and nitrogen-fixing
rhizobia. Associations with AMF are widespread in land plants and result in increased P
acquisition, while rhizobia induce the formation of root nodules in legumes, where the bacterial
symbiont fixes atmospheric N and provides it to its host in the form of ammonia. The
effectiveness of these interactions, and how they modify the impact of increased CO 2 levels on
host growth varies with the plant nutritional status and with the symbiosis functional type. For
example, ectomycorrhizal (ECM) trees typically show strong CO 2 fertilization effects, even
under low N, while AMF-associated plants often remain N-limited (11). Nodule symbioses in
legumes can sustain a stronger growth response to elevated CO₂ by supplying fixed nitrogen
(12–15), although the stimulation of fixation varies with legume species and soil nitrogen, and
can fade over years (16, 17). These findings establish microbial functional types as critical
mediators of plant responses to rising CO₂ concentrations, with potential impact on terrestrial
carbon cycling.


However, while the role of nutrient limitation and its alleviation by AM fungi and N-fixing
rhizobia are well characterized, the influence of the taxonomically diverse community of
commensal microbes that colonize the root and rhizosphere, known as the root microbiota,
remains poorly understood. Such interactions could influence not only plant growth, but also
patterns of root exudation, rhizosphere colonization, and soil carbon fluxes (18), with cascading
implications for nutrient cycling and long-term carbon storage (19). In addition, microbial
uptake of plant-derived carbon increases microbial respiration, with variable effects on
microbial biomass, potentially tipping the balance between soil carbon sequestration and loss

–
(10, 20 23). Furthermore, the physiological responses and growth dynamics of soil microbes
are themselves shaped by changing CO₂ conditions, adding an additional layer of complexity
(24). Deciphering how these microbial responses are organized across functional groups
remains a major challenge for global change biology. Experiments under elevated CO₂ have so
far compared plants with and without a single symbiont in sterilized substrate (13, 15).
However, whether a taxonomically complex root microbiota alters the response of a plant to
rising CO₂ has never been tested against germ-free plants.


To address this, we implemented a reductionist approach to isolate microbial contributions to
plant carbon dynamics under varying CO₂. Using a gnotobiotic system with controlled
environmental conditions and automated gas exchange monitoring, we cultivated the model
legume Lotus japonicus under different atmospheric CO₂ concentrations. This system enabled
us to disentangle the roles of three microbial inputs: AM fungi, nitrogen-fixing rhizobia, and a
defined synthetic bacterial community of root commensals, alone and in combination, against
germ-free plants. We hypothesized that rising CO₂ alters the balance between photosynthesis
and respiration, and that microbial partners modulate plant responses by affecting nutrient
availability and carbon allocation between source and sink tissues. Our aim was to quantify how
single and combined microbial associations shape plant growth, carbon fluxes, and microbial
colonization along a CO₂ gradient. We show that the magnitude of the plant response to
elevated CO₂ depends on the composition of its root microbiota, and that the strongest
responses were observed when rhizobia, mycorrhizal fungi, and commensal bacteria were all

present.


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


Results


A reductionist system for gnotobiotic plant growth under varying CO₂ concentrations


To understand how changes in atmospheric CO₂ influence plant physiology and root microbiota
assembly, we required a system that combined precise environmental control with the capacity
to manipulate microbial composition. Natural soils and field experiments offer limited control
over abiotic variables and microbial inputs, making it difficult to isolate the specific impacts of
varying CO₂ concentrations. To overcome these limitations, we developed a custom gnotobiotic
growth platform, the EcoChamber, which allows plants to grow in sterile peat substrate under
controlled humidity, temperature, light, and defined CO₂ levels, while being colonized by
synthetic microbial communities (SynComs) of known composition.


The EcoChamber system is built around sealed plexiglass enclosures that house plastic pots
filled with a sterilized peat substrate that can be individually inoculated with defined microbial
SynComs (see materials and methods). Germ-free L. japonicus seedlings are then planted into
the inoculated substrate, ensuring that microbial colonization originates only from the defined
inputs. Each chamber contains 20 pots, all receiving the same microbial treatment to prevent
microbial cross-contamination. SynComs were derived from a previously established L.
japonicus root and nodule culture collection (25; Lj-SPHERE), and include taxonomically diverse
commensal bacteria representing 15 bacterial families spanning the four dominant phyla of the
root microbiota (Table S1). In addition to the SynCom, plants can also be inoculated with
nodule-forming, nitrogen-fixing rhizobia and/or AMF symbionts. The chambers are connected
to automated air and water circulation systems (Fig. 1A), and placed inside climate-controlled
growth cabinets where light intensity, humidity, and temperature can be adjusted. These
cabinets are also equipped with CO 2 scrubbers and connected to gas inputs (Fig. 1B), enabling
stable maintenance of CO₂ concentrations at defined levels.


Under standard operation, air from the growth cabinet circulates through each EcoChamber in
an open-loop configuration (Fig. 1C), enabling continuous supply of CO 2 . To measure plantdriven changes in CO₂ concentrations, the system can switch to a closed-loop operation mode
(Fig. 1D), where air from each chamber recirculates within a sealed circuit and passes through a
sensor array that monitors humidity, pressure, and CO₂ levels in real time. During these closed
cycles, measured changes in CO₂ concentrations reflect the balance of photosynthetic uptake
and respiratory release. Positive slopes in CO₂ concentrations, typically observed during the
night, correspond to net respiration, which includes contributions from plant shoots, roots, and
microbial activity in the soil. Conversely, negative slopes, observed during the day period, are
the consequence of net assimilation by the plant canopy. A limitation of the system is that it
does not separate above- and belowground compartments, so net CO₂ fluxes reflect
photosynthetic assimilation in the shoot combined with ongoing root and microbial respiration
in soil. Each growth cabinet accommodates up to six EcoChambers, which are connected to the
sensor array via automated valves. For 9 minutes every hour, air from each chamber circulates
through the sensor array in a closed-loop mode, enabling high-resolution monitoring of CO₂
fluxes across all chambers. All system operations, data logging, and flux calculations are
automated and managed by custom software (see materials and methods).


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


Three independent growth cabinets are used in parallel (Fig. 1B), each configured with identical
light, temperature, and humidity settings, but maintained at one of three distinct CO₂ levels:
200 ppm, representing pre-industrial conditions; 400 ppm, approximating current ambient
levels; and 800 ppm, simulating projected future atmospheric concentrations by the end of the
century (26). Each cabinet contains up to six gnotobiotic chambers, enabling parallel testing of
microbial treatments across CO₂ regimes. The EcoChamber platform thus provides a controlled
environment to quantify gas exchange dynamics and plant-microbe interactions under shifting
CO₂ regimes.


CO 2 fertilization effect on L. japonicus depends on microbial associations


To test how root-associated microbes modulate plant responses to rising atmospheric CO₂, we
grew L. japonicus in EcoChambers using three CO₂ concentrations (200, 400, and 800 ppm), in
the presence or absence of microbial partners for five weeks. We selected L. japonicus for its
capacity to form functional associations with both AMF (Rhizophagus irregularis) and nitrogenfixing rhizobia (Mesorhizobium loti), and for the availability of a collection of well-characterized
root- and nodule-derived bacterial cultures (25). This allowed us to design a modular
gnotobiotic experiment combining AMF, M. loti, and a bacterial SynCom composed of 15

diverse isolates derived from natural root communities.


Under axenic conditions, we observed a small but significant CO₂ fertilization effect across
nearly all plant traits measured, including shoot and root fresh weight, shoot length, chlorophyll
content, and total plant weight (Fig. 2A-E). Compared to ambient levels (400 ppm), shoot and
root fresh weight increased by 37% and 29%, respectively, under elevated CO₂ (800 ppm), and
decreased by 49% and 86% under low CO₂ (200 ppm; Fig. 2A, D). These results confirm that the
EcoChamber system captures expected CO₂-driven plant physiological responses and provides a
baseline against which the effects of different microbial associations can be compared (dashed
lines in Fig. 2).


Under ambient atmospheric conditions (400 ppm), inoculation with M. loti led to significant
improvements in all plant performance parameters relative to axenic controls (Fig. 2A-E).
Further increases were observed when M. loti was combined with the bacterial SynCom, or
when all three microbial treatments were present, including AMF. By contrast, treatments with
AMF or SynCom alone did not improve plant growth. The former result was expected as the
peat substrate used in these experiments is not limited in bioavailable P (Table S2) and thus
offsets the beneficial effect of AMF on plant growth. The absence of a growth benefit from the
SynCom alone contrasts with the growth promotion by Lj-SPHERE SynComs observed previously
in the presence of 1 mM nitrate (25), and is consistent with nitrogen being the limiting nutrient
in our substrate. Although AMF inoculation did not enhance growth, root colonization by R.
irregularis was detected across all CO₂ levels using qPCR targeting a region of the fungal
mitochondrial DNA (Fig. S1), indicating that AMF colonization alone was insufficient to boost
plant growth in the experimental conditions tested.


When plants were grown under elevated CO₂, the effect of microbial inoculation was markedly
increased. Plants inoculated with M. loti alone showed a 137% increase in shoot fresh weight
under 800 ppm compared to 400 ppm (Fig. 2A). Plants treated with M. loti plus SynCom, and
those combining all three microbial treatments, showed increases of 73% and 82%,


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


respectively, when comparing 800 ppm to 400 ppm conditions. These increases exceeded the
gains observed in axenic plants under elevated CO₂, indicating a synergistic interaction between
microbial treatment and plant physiological responses. Across all traits measured, the highest
average values for shoot biomass, chlorophyll content, and root biomass were observed when
M. loti, AMF, and the commensal SynCom were present together (Fig. 2A, C, D), although not
all differences were statistically significant. This suggests that the presence of multiple
microbial partners can produce additive or synergistic effects that enhance plant performance
beyond what is achieved by associations with either AMF or rhizobial symbionts alone.


As expected, nodulation was strictly dependent on the presence of M. loti. No nodules were
found in axenic, AMF-only, or SynCom-only treatments under any CO₂ condition (Fig. 2G-H).
Among M. loti-containing treatments, the number and total weight of nodules increased with
rising CO 2 levels, consistent with enhanced symbiotic nitrogen fixation under enriched carbon
conditions. Notably, plants co-inoculated with M. loti and the bacterial SynCom, with or without
AMF, showed significantly higher nodulation than those treated with M. loti alone (Fig. 2G). The
available data, however, are insufficient to determine if this increased nodulation is the
consequence of indirect effects of the bacterial SynCom on the plant metabolic status or the
result of direct microbial interactions with M. loti or with the plant symbiotic signaling pathway.


Unlike other growth traits, root-to-shoot ratios exhibited an unanticipated and markedly
distinct response pattern (Fig. 2F). The highest values were observed in plants inoculated with
the SynCom alone, followed by AMF-only, and axenic treatments. In contrast, plants associated
with M. loti showed lower root-to-shoot ratios at 400 and 800 ppm, consistent with greater
investment in shoot biomass when nitrogen is readily available. The higher ratios observed in
the SynCom-only treatment, despite limited impact on total plant biomass, suggest a
commensal influence on plant resource allocation. Whether this reflects a stress-induced shift
that selectively inhibits above-ground plant growth or an adaptive physiological response to
microbial signals that promote below-ground organ growth remains unclear. However, such
microbe-driven effects on biomass allocation have potential implications for belowground
carbon deposition under rising atmospheric CO₂.


Taken together, these findings show that microbial associations not only influence the
magnitude of the plant CO₂ fertilization response but also alter carbon allocation along the
root-shoot axis. In particular, the full benefit of CO₂ fertilization in L. japonicus required its
nitrogen-fixing symbiont, and was further enhanced when M. loti was combined with rootderived commensal communities and AMF. As global CO₂ levels rise, such interactions are likely
to affect not only plant growth and productivity but also have implications for soil carbon and
nitrogen fluxes in both managed and natural ecosystems.


Symbiotic and commensal microbes shape carbon assimilation responses to rising CO₂


To determine whether microbial associations modulate not only plant growth but also net
carbon fluxes, we used the EcoChamber system to follow CO₂ dynamics during approximately
35 days of L. japonicus growth under the three previously described atmospheric CO₂
concentrations (Fig. 3A-C). To estimate gas fluxes, each EcoChamber was periodically switched
to a closed-loop configuration allowing measurement of CO₂ concentrations with an accuracy of
±3 ppm, sampled once per second (Fig. 1A). During each closed-loop cycle, changes in CO₂


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


concentration were recorded and fitted with a regression model (Fig. 1C; see materials and
methods). The slope of the curve corresponds to the net CO₂ flux in the chamber. Negative
slopes reflect net CO₂ assimilation, and positive slopes indicate net CO₂ release. As both
photosynthesis and respiration occur simultaneously, and we do not separate above- and
belowground compartments, measured values reflect the net balance of shoot assimilation
minus respiration by shoots, roots, and soil microbes. These values are referred to as net fluxes
throughout.


Under ambient CO₂ conditions (400 ppm; Fig. 3B), fluxes remained positive during the day for
all treatments early in the experiment, indicating that respiratory outputs exceeded
assimilation. This process is likely attributable to the mobilization of seedborne carbon
resources by the plant, combined with microbial respiration in the soil substrate. From day 12
onwards, net fluxes became negative for most treatments, marking the onset of net daytime
carbon uptake. Starting at day 24, and until the end of the experiment, treatments including M.
loti, either alone, with SynCom, or with both SynCom and AMF, exhibited significantly larger
negative net fluxes on average (-230 to -331 μmol h [-1] ) than axenic controls, SynCom-only, or
AMF-only conditions (-41, -12, -58 μmol h [-1] on average, respectively). These results are
consistent with increased nitrogen availability supporting greater plant biomass and
photosynthetic capacity, and mirror the shoot fresh weight differences observed at harvest
(Fig. 2A).


At low CO₂ (200 ppm, Fig. 3A), daytime fluxes remained positive for longer, and net negative
fluxes only appeared later in the experiment for all treatments (days 27 to 32), except for
axenic and SynCom-only conditions. Overall, at 200 ppm net flux magnitudes were smaller than
those observed at 400 ppm, and only chambers with M. loti inoculated plants showed
significantly different patterns compared to axenic controls. When comparing ambient CO₂ and
elevated CO₂ conditions (400 and 800 ppm, respectively), net fluxes not only turned negative
earlier in the experiment (from day 12 compared to day 14 for axenic conditions) but also
reached greater magnitudes overall (Fig. 3B-C). Treatments containing M. loti again showed the
strongest negative fluxes, ranging from -534 to -627 μmol h [-1], approximately 13-fold greater
than axenic controls. While fluxes for AMF-only were similar to axenic conditions, SynCom-only
chambers displayed significantly weaker net assimilation (-27 μmol h [-1] ), suggesting either
increased respiration, decreased assimilation, or both.


To determine whether differences in gas fluxes could be attributed solely to differences in
aboveground plant biomass, we normalized final-day fluxes by shoot fresh weight at harvest
(Fig. 3D). This analysis revealed that under ambient conditions (400 ppm), all microbial
treatments showed stronger CO₂ assimilation per gram of plant shoot biomass compared to
their counterparts under 200 ppm, suggesting increased normalized photosynthetic
performance. However, no treatment showed higher normalized assimilation at 800 ppm than
at 400 ppm, except for SynCom-only, which exhibited a significantly stronger negative flux per
unit biomass under elevated CO₂ (Fig. 3D). Surprisingly, despite having lower shoot biomass
than axenic or symbiont-inoculated plants (Fig. 2A), SynCom-only plants achieved a higher
normalized CO₂ drawdown rate than axenic plants (Fig. 3D).


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


Several explanations could account for this result. It is possible that the bacterial SynCom
enhances the photosynthetic capacity of L. japonicus per unit of leaf tissue, for instance by
modulating leaf growth traits. Alternatively, the presence of bacterial commensals might
increase belowground carbon sink strength, stimulating photosynthesis via altered source-sink
dynamics. A third possibility is that SynCom-only treatments exhibit reduced microbial
respiration, resulting in a more negative net flux. These hypotheses are not mutually exclusive,
and further physiological and molecular analyses will be needed to disentangle their relative

contributions.


In sum, these findings indicate that microbial associations shape not only the magnitude of
differences in plant growth under changing CO₂ levels, but also the trajectory of net carbon
fluxes over time. The presence of symbiotic rhizobia accelerates both the onset and magnitude
of net carbon gain, particularly under elevated CO₂, while commensal microbes modulate
carbon assimilation in more subtle or indirect ways. Together, these results highlight a complex
and dynamic interaction between microbial association type, atmospheric CO₂ concentration,
and plant source-sink carbon balance.


Atmospheric CO₂ levels shape the composition and abundance of the root microbiota


To determine how rising atmospheric CO₂ concentrations affect the composition of the root
microbiota, we performed bacterial 16S rRNA gene amplicon profiling and absolute
quantification of bacterial abundance from L. japonicus plants grown under the same
conditions described above. Our analyses focused on both the root (root surface and interior)
and rhizosphere (surrounding peat substrate) compartments, and included comparisons across
microbial treatments and CO₂ concentrations.


Principal Coordinate Analysis (PCoA) of amplicon profiles revealed a clear and consistent
separation between microbial communities in the root and rhizosphere, after in silico depletion
of reads belonging to M. loti (Fig. 4A). This separation by compartment was evident across all
CO₂ conditions and independent of the microbial treatment applied, confirming that the plant
exerts strong spatial filtering over the SynCom inoculated in the growth substrate.


To assess how microbial treatments shaped bacterial community structure under each CO₂
condition, we constrained the principal coordinate analysis by treatment. This revealed
significant differences in community composition between treatments (SynCom only, SynCom +
M. loti, and SynCom + M. loti + AMF), with the magnitude of the separation increasing when
comparing 400 to 800 ppm (Fig. 4B, E). Permutational multivariate analysis of variance
(PERMANOVA) showed that microbial treatment explained 5.38%, 5.82%, and 9.76% of the
total variance in root community composition at 200, 400, and 800 ppm, respectively (Fig. 4B,
E). These results suggest that the impact of microbial input on the composition of the root
microbiota becomes stronger under elevated CO₂. A similar analysis for the rhizosphere
compartment (Fig. 4C, E) also showed treatment-dependent effects, which rose from 4.37%
and 2.87% of variance at 200 and 400 ppm to 9.45% at 800 ppm, close to the 9.76% explained
in roots at that level. The stronger treatment effect in roots confirms that selection by the plant
is a key determinant of root microbiota structure.


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


Analysis of the taxonomic composition of microbial communities further revealed that root
SynComs were characterized by increased relative abundances of strains belonging to the
Pseudomonadaceae, Microbacteriaceae, Hyphomicrobiaceae and Mycobacteriaceae families
compared to input samples, and decreased relative abundance of members of the
Flavobacteriaceae, Comamonadaceae and Alcaligenaceae families (Fig. 4D). In contrast,
rhizosphere communities showed higher average relative abundance of Flavobacteriaceae
compared to roots but a decrease in the abundance of members of the Microbacteriaceae,
Hyphomicrobiaceae and Mycobacteriaceae. These compartment-specific differences in
taxonomic profiles were consistent across CO 2 levels, as expected from the analyses of betadiversity (Fig. 4A). However, we also observed smaller but significant differences in the relative
abundance of a subset of community members across CO 2 conditions. The relative abundance
of Comamonadaceae and Bradyrhizobiaceae increased with rising CO 2 levels across all microbial
treatments, whereas the relative abundance of Xanthomonadaceae showed the opposite
pattern. For some taxa, the correlation with increasing CO 2 was dependent on the microbial
treatment: Mycobacteriaceae showed increased abundance in the presence of microbial
symbionts, but decreased abundance on SynCom-only treated plants, while the opposite trend
was observed for Rhizobiaceae (Fig. S2).


We next asked whether these CO₂- and treatment-dependent shifts in community composition
were accompanied by changes in total microbial load. We quantified bacterial abundance in
roots using qPCR of the bacterial 16S rRNA gene normalized to the L. japonicus Nfr5 single-copy
gene (Fig. 4F). Across all CO₂ conditions, plants inoculated with SynCom + M. loti or SynCom +
M. loti + AMF harbored significantly more bacteria per plant DNA than those treated with
SynCom or M. loti alone. This result indicates that co-inoculation with commensals and
symbionts leads to higher microbial loads in the root.


Bacterial abundances were higher under elevated CO₂ than under ambient conditions across all
microbial treatments (Fig. 4F), as previously reported for wheat roots under elevated CO₂ and
high nitrate supply (27). At 800 ppm, the same amount of plant DNA corresponded to
significantly higher bacterial DNA, suggesting that higher atmospheric CO₂ levels increase the
capacity of the root system to support microbial colonization. To directly probe whether
elevated CO₂ led to increases in plant carbon release belowground, we quantified the organic
carbon released over 24 h by the root systems of axenic, SynCom, and SynCom plus M. loti
inoculated plants grown at 400 and 800 ppm, in two independent experiments (see materials
and methods). Root systems of axenic plants released 1.6-fold more organic carbon at 800 ppm
than at 400 ppm, consistently in both experiments (Fig. 4G), showing that elevated CO₂
increases root carbon release in L. japonicus in the absence of microbial partners. This response
depended on the root microbiota. Plants colonized by the SynCom plus M. loti showed a
smaller increase (1.2-fold), and plants colonized by the SynCom alone released less carbon at
800 than at 400 ppm (0.55-fold), so that the CO₂ responses of the three treatments differed
from each other (Fig. 4G). These results show that the increase in root carbon release driven by
elevated CO₂ is reduced in plants colonized by the SynCom plus M. loti and abolished in plants
colonized by the SynCom alone.


Together, these findings reveal that both the composition and abundance of the root
microbiota are affected by atmospheric CO₂ concentration. While microbial composition


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


becomes increasingly responsive to rising CO₂ (Fig. 4E), the total microbial load also increases
(Fig. 4F), particularly when commensals and symbionts are combined. Under elevated CO₂, M.
loti alone supported higher microbial biomass than SynCom alone, but the highest bacterial
loads were found in treatments combining M. loti with the SynCom, with or without AMF.
These results parallel our earlier observations on increased nodule number and weight under
these same combinations (Fig. 2G-H), pointing to a functional interaction between symbionts
and commensals that becomes more pronounced under elevated CO₂. Integrating the carbon
release measurements with the microbial load data suggests that the root microbiota controls
the fate of the additional carbon fixed under elevated CO₂. In axenic plants, part of the
additional assimilate is released from the roots as organic carbon. In plants colonized by the
SynCom plus M. loti, carbon release rose only modestly while root systems, nodules and
bacterial loads increased (Fig. 2D, G-H, 4F), suggesting that the additional carbon is retained in
the root and by its associated microbes rather than released. Commensals alone, which supply
no nitrogen, reduced carbon release under elevated CO₂ and supported only modest bacterial
loads, in line with the limited growth response of these plants (Fig. 2). One possibility is that
axenic plants invest more carbon in exudation to recruit microbes for nutrient mobilization and
protection, a demand that is met once the root is colonized. Alternatively, part of the carbon
released by colonized roots may have been taken up by bacteria attached to the root during the
24 h collection, which we did not quantify. Given the strong link between microbial activity and
nutrient cycling in soil, these CO₂-dependent changes in carbon release and microbial biomass
have important implications for understanding how plant-microbiota associations will influence
soil carbon dynamics in future climate scenarios.


Microbial community composition shapes multi-trait plant responses to elevated CO₂


To compare how different microbial associations modulate plant and microbial responses to
rising atmospheric CO₂, we jointly analyzed the major physiological and microbial traits
measured (Fig. 5, Fig. S3). This analysis integrates ten key measurements across CO₂
concentrations, including plant growth parameters, nodulation patterns, CO₂ fluxes, and
microbial abundance, and provides an overview of how each treatment alters plant and
microbial traits in response to increasing CO₂ levels.


In axenic plants, we observed a small but significant CO₂ fertilization effect when comparing
400 and 800 ppm across most growth-related traits, including shoot fresh weight, shoot length,
and chlorophyll content (Fig. 5A). However, aside from these biomass gains, the overall shape
of the response remains relatively uniform across CO₂ concentrations, indicating limited
physiological plasticity in the absence of microbial partners. AMF-treated plants closely
resemble axenic controls (Fig. 5B), consistent with previous results and likely explained by the
non-limiting amount of phosphorus present in our growth substrate (Table S2).


The SynCom-only treatment exhibited a distinct signature (Fig. 5D). Across all CO₂
concentrations, these plants showed minimal changes in growth, nodulation, or chlorophyll
content, suggesting that commensals alone cannot support strong responsiveness to elevated
CO₂. However, SynCom-inoculated plants consistently displayed the highest root-to-shoot ratios
of any treatment, including under elevated CO₂. This suggests that interactions with bacterial
commensals strongly influence plant decisions on resource allocation to above- or below

[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


ground organs, even in the absence of nitrogen- or phosphorus-providing symbioses. Whether
this reflects a developmental response to microbial signaling or an adaptive reallocation of
resources under nutrient limitation remains unresolved, but the pattern is remarkably
consistent across independent replicates.


By contrast, plants inoculated with M. loti, with or without bacterial commensals (Fig. 5C, E-F),
displayed strong responses to elevated CO₂ across nearly all traits. In these treatments, gains in
shoot and root biomass, chlorophyll content, nodule number and weight, and microbial
abundance all increased with rising CO₂ levels. The combination of M. loti, SynCom, and AMF
produced the most substantial trait-level gains at 800 ppm across nearly all axes (Fig. 5F). In
addition, the impact of the SynCom treatment on root-to-shoot ratios was not observed when
in combination with M. loti and AMF, indicating that functional symbioses can override the
altered resource allocation patterns that we observed in the presence of commensal bacterial
communities alone. These patterns suggest that the full physiological benefit of elevated CO₂
requires simultaneous association between the plant host and all three types of microbial

partners.


These findings highlight that the root-associated microbiota shape plant CO₂ responses in a
treatment-specific manner. Functional associations with rhizobia and commensals together
enhance growth, microbial colonization, and net carbon uptake, particularly under elevated
CO₂, while AMF contribute less to plant performance in the tested conditions, unless paired
with other microbial partners. By contrast, under germ-free conditions, the CO₂ fertilization
effect on plant growth is small when CO₂ concentrations rise. Further, bacterial commensals
alone do not enhance growth but consistently shift biomass allocation toward belowground
compartments, underscoring their subtle but significant role in altering plant physiology. Taken
together, these data support a model in which microbial diversity provides complementary
functional benefits that increase under projected future atmospheric CO₂ conditions.


Discussion


How plants respond to rising atmospheric CO₂ is a central question for understanding future
terrestrial ecosystems. While the CO₂ fertilization effect on plant productivity is well established
(28), its interaction with the root-associated microbiota remains poorly resolved. Our study
reveals that both the magnitude and nature of plant responses to elevated CO₂ strongly depend
on microbial associations, with the strongest responses observed only when commensals,
symbiotic rhizobia, and mycorrhizal fungi were all present. Our results, obtained in a controlled
gnotobiotic system using the model legume L. japonicus, demonstrate that microbial diversity
plays a critical role in plant responses to rising atmospheric CO₂ conditions.


Using the EcoChamber platform, we were able to isolate and monitor plant physiological and
microbial parameters under pre-industrial (200 ppm), ambient (400 ppm), and elevated (800
ppm) CO₂ concentrations. The system enabled real-time quantification of net CO₂ fluxes,
revealing dynamic shifts in the balance between photosynthetic assimilation and plant and
microbial respiration over time. Across all conditions, inoculation with the nodule symbiont M.
loti significantly increased net CO₂ uptake rates, particularly under elevated CO₂, supporting the
view that nitrogen availability becomes increasingly limiting as carbon input increases. This is in
line with previous findings in Medicago, soybean and bean, where elevated CO₂ enhances


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


–
nodulation and the reliance on symbiotically fixed nitrogen (13 15). Notably, this stronger CO₂
drawdown is observed despite higher root-associated bacterial loads under elevated CO₂ (Fig.
4F), which would be expected to increase belowground respiration and counteract net
assimilation.Our results also raise the possibility that plant-microbial associations were
important for the survival and performance of C3 plants during past low atmospheric CO₂
periods, when carbon availability strongly constrained growth. Most low-CO₂ studies report
reduced growth, but rarely account for the root microbial community or for the ability of plants
to shape that community. Our results, although still consistent with the idea that symbioses
could be less important under low CO₂, nonetheless suggest that microbial associations can
remain a key determinant of plant performance even under carbon limitation.


Commensal bacteria alone had no detectable effect on plant growth but strongly influenced
shoot-to-root carbon allocation, indicating that commensals alter host resource allocation even
in the absence of canonical symbioses. When all three microbial groups (commensals, rhizobia,
and AMF) were present, plants showed the most pronounced responses across nearly all traits,
including biomass, nodulation, and CO₂ flux. This pattern suggests that to fully benefit from
elevated atmospheric CO₂, plants may require multi-partner microbial associations that span
trophic and functional niches. The importance of such synergistic interactions is underscored by
the observation that the response of N-fixation to elevated CO₂ is not universal, but can be
positive, neutral, or even negative depending on the specific legume species and background
soil nitrogen availability (17). Our work suggests that diverse microbial partnerships may help
buffer against this variability, leading to a more robust plant response.


These findings are especially relevant given that many ecosystem-scale models that predict
plant productivity and soil carbon cycling under future climate scenarios treat microbial activity
as a black box, rarely accounting for the structure or composition of microbial communities
(19). Our results point to a more complex interaction: not only do root-associated communities
change in abundance and composition under elevated CO₂, but the structure of microbial input
(i.e., which microbial functional guilds are present) modulates plant responses at multiple levels
and across multiple traits. Our measurements of root-derived carbon show that elevated CO₂
increases carbon release from root systems and that the root microbiota determines the size of
this response. In germ-free plants release rose 1.6-fold, with rhizobia and commensals present
the increase was reduced, and abolished with commensals alone (Fig. 4G). Together with the
increase in root bacterial loads under elevated CO₂ (Fig. 4F), these patterns suggest that
microbial partners determine the magnitude of carbon export under high CO₂, retaining it in
the root and associated microbes rather than releasing it, which in turn may amplify microbial
activity and nutrient mobilization. Such feedback loops could significantly alter belowground
carbon partitioning and nutrient fluxes, with consequences for carbon retention and release in
soil ecosystems, since greater carbon input to an active microbiota under elevated CO₂
accelerates the decomposition of soil organic matter and can limit soil carbon storage (29, 30).


Our data also highlight the unexpectedly strong influence of commensal microbes on plant
growth traits. While commensals alone did not enhance growth, they dramatically influenced
carbon allocation patterns, consistently driving higher root-to-shoot ratios, regardless of CO₂
levels. This suggests that even in the absence of direct growth promotion, commensals can shift
carbon allocation between root and shoot, possibly by preparing the root system for future


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


symbiotic interactions or improving plant foraging capacity under nutrient-limited conditions.
The underlying mechanism remains unclear but may involve the common symbiosis signaling
pathway. This hypothesis is consistent with previous findings indicating altered commensal
microbiota structures in symbiosis-impaired mutants in L. japonicus independently of nitrogen
supply (25, 31, 32). Whether this phenotype is driven by one or multiple members of the
bacterial community remains unknown, but it appears to be a dominant and reproducible
feature of commensal colonization.


Similarly, treatment with AMF showed significant interactions with other microbial partners.
While AM fungi are often viewed as key players in plant nutrient acquisition, particularly
phosphorus, their beneficial effects were negligible in our system, likely due to the high
phosphate content of the sterile peat substrate used. However, when AMF were added
alongside rhizobia and commensals, we observed further improvements across multiple traits,
suggesting that their contribution may also depend on interactions with other microbiota
members. This context-dependence of the beneficial impact of symbioses suggests that efforts
to engineer plant microbiomes for future agriculture will require an understanding of microbial
interkingdom interactions (33), rather than relying on single-strain inoculants.


Beyond its specific findings, our study provides a scalable framework for studying plantmicrobiota interactions under precisely controlled environmental conditions. The EcoChamber
system enables quantitative integration of physiological, microbial, and gas exchange data,
linking microbiota composition to plant carbon balance. Because the generation time of plantassociated microbes is orders of magnitude shorter than that of their hosts, the same system
can be used to ask whether microbial adaptation to rising CO₂ outpaces that of the plant over
ecological timescales.


Taken together, our results demonstrate that the root-associated microbiota is an integral
determinant of plant responses to altered CO₂. In our experiments, the strongest responses
across traits were observed when commensals, rhizobia, and AMF were all present under
elevated CO₂. Whether this pattern generalizes across plant species and soil types remains to
be tested. Nonetheless, our findings imply that rising CO₂ levels not only alter plant physiology,
but also modify the structure and function of their microbial partnerships.


Acknowledgments


This work was supported by the Max Planck Society and the Cluster of Excellence on Plant
Sciences (CEPLAS), funded by the Deutsche Forschungsgemeinschaft (DFG, German Research
Foundation) under Germany's Excellence Strategy (EXC 2048/1, project no. 390686111). R.G.-O.
acknowledges funding from the European Research Council (ERC) under the European Union's
Horizon Europe research and innovation programme (PHYCOSPHERES, grant agreement No.
101077231).


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


Main Figures


Figure 1. Design and operation of the EcoChamber platform. (A) The EcoChamber growth
system. Top, isometric and orthographic views of one unit holding 20 FlowPots. Bottom, air and
water circuits of a unit. Cabinet air enters through an air pump and a hydrophobic filter and
passes a sensor line that records CO₂ concentration, air flow, pressure, relative humidity, and
temperature before returning to the cabinet. Sterile water reaches the pots through a vent
port, a water vessel, a water pump, a non-return valve, and a hydrophilic filter. (B) Six
EcoChamber units inside each of three growth cabinets with controlled light, temperature, and
humidity, held at 200 ppm (CO₂ scrubber), 400 ppm (ambient atmosphere), and 800 ppm CO₂
(gas cylinders). An air supply module and a sensor line serve the six units of a cabinet. Each unit
operates independently, allowing full-factorial combinations of microbial treatments and
atmospheric CO₂ regimes. (C-D) Open-loop (C) and closed-loop (D) airflow configurations, each
with the CO₂ concentration of the air entering (solid line) and leaving (dotted line) the unit over
one 60-minute cycle. In open-loop mode, cabinet air flows through the unit. In closed-loop
mode, the air valve recirculates the air of the unit through the sensor line, and the CO₂
concentration falls when net assimilation exceeds respiration (C) or rises when net respiration
exceeds assimilation (D). The slope of the concentration over the closed-loop interval gives the
net CO₂ flux of the unit.


Figure 2. CO₂ fertilization effect on L. japonicus treated with different root-associated
microbial communities. L. japonicus plants were grown for five weeks under six microbial
treatments and three atmospheric CO₂ concentrations. (A) Shoot fresh weight, (B) shoot length,
(C) chlorophyll content, (D) root fresh weight, (E) plant total fresh weight, (F) root-to-shoot
ratio, (G) number of nodules, and (H) total weight of nodules. Nodule traits are shown for the
treatments containing M. loti. Boxes show the median and interquartile range, whiskers extend
to 1.5 times the interquartile range, and the dashed line marks the mean of the axenic
treatment at each CO₂ level. Each data point is the mean of the plants in one pot (mostly 3),
except in (B), where each plant was measured separately. Data are pooled from two to four
independent experiments per treatment and CO₂ level, with 20 pots per treatment and CO₂
level in each experiment, and symbols denote the experiment (A to D). Black letters indicate
differences between microbial treatments within a CO₂ level and grey letters differences
between CO₂ levels within a treatment. Statistical significance was determined by KruskalWallis or ANOVA followed by post hoc tests (see materials and methods). Treatment
abbreviations: SC, SynCom; SC + M, SynCom + M. loti; SC + M + A, SynCom + M. loti + AMF.


Figure 3. Net CO₂ fluxes during daytime across CO₂ and microbial treatments. CO₂ fluxes were
measured over approximately 35 days of L. japonicus growth under 200 ppm (A), 400 ppm (B),
and 800 ppm (C) atmospheric CO₂. Data points are daily means of the hourly daytime flux
measurements of individual chambers, from two to four independent experiments, and shaded
bands show the standard deviation of each treatment. The insets in (A) and (B) show days 24 to
32 at an expanded scale. (D) Final-day CO₂ flux measurements normalized by shoot fresh
weight. Each data point is one hourly flux measurement during the daytime of the final day (10
per chamber), from two to four independent experiments, with symbols denoting the
experiment (A to D). Positive values reflect net respiration, while negative values reflect net CO₂
assimilation. Black letters indicate differences between microbial treatments within a CO₂ level


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


and grey letters differences between CO₂ levels within a treatment. Statistical significance was
determined by Kruskal-Wallis or ANOVA. Treatment abbreviations: SC, SynCom; SC + M,
SynCom + M. loti; SC + M + A, SynCom + M. loti + AMF.


Figure 4. CO₂-dependent effects of microbial treatment on root- and rhizosphere-associated
communities. (A) Principal Coordinate Analysis (PCoA) of Bray-Curtis dissimilarities between
16S rRNA profiles from samples of root and rhizosphere compartments across all CO₂
concentrations. PCoA of bacterial community profiles constrained by microbial treatment at
each CO₂ level from root (B) or rhizosphere (C) samples. (D) Compositional profiles of SynCom
inputs and communities retrieved from roots and rhizosphere. (E) PERMANOVA analysis of
treatment effects on community composition, showing increased variance explained at
elevated CO₂. (F) Bacterial abundance in root systems quantified by qPCR of 16S rRNA
normalized to the L. japonicus Nfr5 gene. Each data point reflects qPCR results from the pooled
plants of one pot (mostly 3), from two to four independent experiments per treatment and CO₂
level, with 20 pots per treatment and CO₂ level in each experiment, and symbols denote the
experiment (A to D). Black letters indicate differences between microbial treatments within a
CO₂ level and grey letters differences between CO₂ levels within a treatment. Statistical
significance was determined by Kruskal-Wallis or ANOVA. Treatment abbreviations: SC,
SynCom; SC + M, SynCom + M. loti; SC + M + A, SynCom + M. loti + AMF. (G) Root-derived
organic carbon released over 24 h into sterile water by the root systems of axenic, SynCom, and
SynCom + M. loti plants grown at 400 or 800 ppm, in two independent experiments (triangles
and circles). The SynCom treatment was tested in one experiment. Each data point represents
one pot. Values are per pot and normalized to the median of the 400 ppm pots of the same
treatment and experiment (n = 19 to 40 pots per group). Asterisks mark differences between
800 and 400 ppm within each treatment (Wilcoxon rank-sum test, *, P < 0.05; ***, P < 0.001),
and letters mark treatments whose CO₂ responses differ from each other (P < 0.001).


Figure 5. Microbial associations shape multi-trait plant responses to rising CO₂. Spider plots
summarizing the CO₂-dependent response of L. japonicus to different microbial treatments
based on ten traits: shoot fresh weight, shoot length, chlorophyll content, root fresh weight,
plant total fresh weight, root-to-shoot ratio, nodule number, total nodule weight, bacterial
abundance, and normalized CO₂ fluxes. Each panel corresponds to one microbial treatment:
axenic (A), AMF only (B), M. loti only (C), SynCom only (D), SynCom + M. loti (E), and SynCom +
M. loti + AMF (F), with the three CO₂ concentrations shown as nested polygons. Each trait is
scaled from 0 (inner ring) to 1 (outer ring). Traits that do not apply to a treatment are absent
from its polygon.


References


1. M. Heimann, M. Reichstein, Terrestrial ecosystem carbon dynamics and climate feedbacks.
Nature 451, 289-292 (2008).


2. P. Ciais, C. Sabine, G. Bala, L. Bopp, V. Brovkin, J. Canadell, A. Chhabra, R. DeFries, J. Galloway,
M. Heimann, et al., Carbon and other biogeochemical cycles, in Climate Change 2013: The
Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change, T. F. Stocker et al., Eds. (Cambridge Univ. Press,
2013), pp. 465-570.


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


3. J. E. Campbell, J. A. Berry, U. Seibt, S. J. Smith, S. A. Montzka, T. Launois, S. Belviso, L. Bopp,
M. Laine, Large historical growth in global terrestrial gross primary production. Nature 544, 8487 (2017).


4. C. Le Quéré, R. M. Andrew, P. Friedlingstein, S. Sitch, J. Hauck, J. Pongratz, P. A. Pickers, J. I.
Korsbakken, G. P. Peters, J. G. Canadell, et al., Global Carbon Budget 2018. Earth Syst. Sci. Data
10, 2141-2194 (2018).


5. D. Schimel, B. B. Stephens, J. B. Fisher, Effect of increasing CO₂ on the terrestrial carbon cycle.
Proc. Natl. Acad. Sci. USA 112, 436-441 (2015).


6. T. F. Keenan, I. C. Prentice, J. G. Canadell, C. A. Williams, H. Wang, M. Raupach, G. J. Collatz,
Recent pause in the growth rate of atmospheric CO₂ due to enhanced terrestrial carbon uptake.
Nat. Commun. 7, 13428 (2016).


7. Z. Zhu, S. Piao, R. B. Myneni, M. Huang, Z. Zeng, J. G. Canadell, P. Ciais, S. Sitch, P.
Friedlingstein, A. Arneth, et al., Greening of the Earth and its drivers. Nat. Clim. Change 6, 791795 (2016).


8. P. B. Reich, S. E. Hobbie, T. Lee, D. S. Ellsworth, J. B. West, D. Tilman, J. M. H. Knops, S.
Naeem, J. Trost, Nitrogen limitation constrains sustainability of ecosystem response to CO₂.
Nature 440, 922-925 (2006).


9. C. Terrer, R. B. Jackson, I. C. Prentice, T. F. Keenan, C. Kaiser, S. Vicca, J. B. Fisher, P. B. Reich,
B. D. Stocker, B. A. Hungate, et al., Nitrogen and phosphorus constrain the CO₂ fertilization of
global plant biomass. Nat. Clim. Change 9, 684-689 (2019).


10. M. Jiang, B. E. Medlyn, J. E. Drake, R. A. Duursma, I. C. Anderson, C. V. M. Barton, M. M.
Boer, Y. Carrillo, L. Castañeda-Gómez, L. Collins, et al., The fate of carbon in a mature forest
under carbon dioxide enrichment. Nature 580, 227-231 (2020).


11. C. Terrer, S. Vicca, B. A. Hungate, R. P. Phillips, I. C. Prentice, Mycorrhizal association as a
primary control of the CO₂ fertilization effect. Science 353, 72-74 (2016).


12. A. Rogers, E. A. Ainsworth, A. D. B. Leakey, Will elevated carbon dioxide concentration
amplify the benefits of nitrogen fixation in legumes? Plant Physiology 151, 1009-1016 (2009).


13. H. Guo, Y. Sun, Y. Li, X. Liu, Q. Ren, K. Zhu-Salzman, F. Ge, Elevated CO₂ modifies N
acquisition of Medicago truncatula by enhancing N fixation and reducing nitrate uptake from
soil. PLoS ONE 8, e81373 (2013).


14. Y. Li, Z. Yu, X. Liu, U. Mathesius, G. Wang, C. Tang, J. Wu, J. Liu, S. Zhang, J. Jin, Elevated CO₂
increases nitrogen fixation at the reproductive phase contributing to various yield responses of
soybean cultivars. Front. Plant Sci. 8, 1546 (2017).


15. E. Bellido, P. de la Haba, E. Agüera, Responses in nodulated bean (Phaseolus vulgaris L.)
plants grown at elevated atmospheric CO₂. Plants 12, 1828 (2023).


16. B. A. Hungate, P. D. Stiling, P. Dijkstra, D. W. Johnson, M. E. Ketterer, G. J. Hymus, C. R.
Hinkle, B. G. Drake, CO₂ elicits long-term decline in nitrogen fixation. Science 304, 1291 (2004).


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


17. J. B. West, J. HilleRisLambers, T. D. Lee, S. E. Hobbie, P. B. Reich, Legume species identity
and soil nitrogen supply determine symbiotic nitrogen-fixation responses to elevated
atmospheric [CO₂]. New Phytol. 167, 523-530 (2005).


18. E. Paterson, J. Hall, E. Rattray, B. Griffiths, K. Ritz, K. Killham, Effect of elevated CO₂ on
rhizosphere carbon flow and soil microbial processes. Glob. Change Biol. 3, 363-377 (1997).


19. D. Naylor, N. Sadler, A. Bhattacharjee, E. B. Graham, C. R. Anderton, R. McClure, M. Lipton,
K. S. Hofmockel, J. K. Jansson, Soil microbiomes under climate change and implications for
carbon cycling. Annu. Rev. Environ. Resour. 45, 29-59 (2020).


20. J. Heath, E. Ayres, M. Possell, R. D. Bardgett, H. I. J. Black, H. Grant, P. Ineson, G. Kerstiens,
Rising atmospheric CO₂ reduces sequestration of root-derived soil carbon. Science 309, 17111713 (2005).


21. K. M. Carney, B. A. Hungate, B. G. Drake, J. P. Megonigal, Altered soil microbial community
at elevated CO₂ leads to loss of soil carbon. Proc. Natl. Acad. Sci. USA 104, 4990-4995 (2007).


22. K. J. van Groenigen, C. W. Osenberg, C. Terrer, Y. Carrillo, F. A. Dijkstra, J. Heath, M. Nie, E.
Pendall, R. P. Phillips, B. A. Hungate, Faster turnover of new soil carbon inputs under increased
atmospheric CO₂. Glob. Change Biol. 23, 4420-4429 (2017).


23. L. A. Domeignoz-Horta, S. L. Cappelli, R. Shrestha, S. Gerin, A. K. Lohila, J. Heinonsalo, D. B.
Nelson, A. Kahmen, P. Duan, D. Sebag, et al., Plant diversity drives positive microbial
associations in the rhizosphere enhancing carbon use efficiency in agricultural soils. Nat.
Commun. 15, 8065 (2024).


24. Y. Ruan, Y. Kuzyakov, X. Liu, X. Zhang, Q. Xu, J. Guo, S. Guo, Q. Shen, Y. Yang, N. Ling,
Elevated temperature and CO₂ strongly affect the growth strategies of soil bacteria. Nat.
Commun. 14, 391 (2023).


25. K. Wippel, K. Tao, Y. Niu, R. Zgadzaj, N. Kiel, R. Guan, E. Dahms, P. Zhang, D. B. Jensen, E.
Logemann, et al., Host preference and invasiveness of commensal bacteria in the Lotus and
Arabidopsis root microbiota. Nat. Microbiol. 6, 1150-1162 (2021).


26. IPCC, Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to
the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, T. F. Stocker et
al., Eds. (Cambridge Univ. Press, 2013).


27. A. Usyskin-Tonne, Y. Hadar, U. Yermiyahu, D. Minz, Elevated CO₂ and nitrate levels increase
wheat root-associated bacterial abundance and impact rhizosphere microbial community
composition and function. ISME J. 15, 1073-1084 (2021).


28. E. A. Ainsworth, S. P. Long, 30 years of free-air carbon dioxide enrichment (FACE): what
have we learned about future crop productivity and its potential for adaptation? Glob. Change
Biol. 27, 27-49 (2021).


29. L. Cheng, F. L. Booker, C. Tu, K. O. Burkey, L. Zhou, H. D. Shew, T. W. Rufty, S. Hu, Arbuscular
mycorrhizal fungi increase organic carbon decomposition under elevated CO₂. Science 337,
1084-1087 (2012).


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.07.764780; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.07.764780)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC 4.0 International license.](http://creativecommons.org/licenses/by-nc/4.0/)


30. K. J. van Groenigen, X. Qi, C. W. Osenberg, Y. Luo, B. A. Hungate, Faster decomposition
under increased atmospheric CO₂ limits soil carbon storage. Science 344, 508-509 (2014).


31. R. Zgadzaj, R. Garrido-Oter, D. B. Jensen, A. Koprivova, P. Schulze-Lefert, S. Radutoiu, Root
nodule symbiosis in Lotus japonicus drives the establishment of distinctive rhizosphere, root,
and nodule bacterial communities. Proc. Natl. Acad. Sci. USA 113, E7996-E8005 (2016).


32. T. Thiergart, R. Zgadzaj, Z. Bozsóki, R. Garrido-Oter, S. Radutoiu, P. Schulze-Lefert, Lotus
japonicus symbiosis genes impact microbial interactions between symbionts and multikingdom
commensal communities. mBio 10, e01833-19 (2019).


33. P. Durán, T. Thiergart, R. Garrido-Oter, M. Agler, E. Kemen, P. Schulze-Lefert, S. Hacquard,
Microbial interkingdom interactions in roots promote Arabidopsis survival. Cell 175, 973983.e14 (2018).


**B**





Air pump Open-loop


Hydrophobic

filter


Air

valve



Hydrophobic



**C**


**D**



Hydrophobic filter





Air

pump



filter





10 20 30 40 50 60


Time (minutes)



Air input

Air output



Non-return valve











Air pump Closed-loop


Hydrophobic

filter









Air

valve


|Col1|Col2|Col3|
|---|---|---|
||||
||||
||||
||||



|Water<br>supply|Col2|Col3|Col4|
|---|---|---|---|
|Water<br>supply||||
|Water<br>supply||||
|Water<br>supply||||
|Water<br>supply||||


CO 2
scrubber



Ambient
atmosphere





10 20 30 40 50 60


Time (minutes)



Respiration


Air input

Air output


**A**



100


75


50


25


0


−25


−50


−75


−100

24 25 26 27 28 29 30 31 32

Day


100

50

0

−50

−100

−150

−200

−250

−300

−350

−400

24 25 26 27 28 29 30 31 32

Day


Treatment


Axenic
AMF
_M. loti_
SynCom
SynCom + _M. loti_
SynCom + _M. loti_ + AMF


**a** Comparison between microbial treatments
**a** Comparison between CO 2 levels


Experiment A B C D



400


200


0


−200


−400


−600


−800


−1000


**B**

400


200


0


−200


−400


−600


−800


−1000


**C**

400


200


0


−200


−400


−600


−800


−1000


**D**


0.10


0.05


0.00


−0.05


−0.10


−0.15



CO 2 fluxes − day time − 200 ppm


0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32

Day


CO 2 fluxes − day time − 400 ppm


0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32
Day


CO 2 fluxes − day time − 800 ppm


0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32
Day


CO 2 flux normalized by shoot fresh weight


|200 ppm|400 ppm|800 ppm|
|---|---|---|
|**a**<br>**ce**<br>**de**<br>**ab**<br>**d**<br>**bc**<br><br>**b**<br>**b**<br>**c**<br>**c**<br>**c**<br>**b**<br>|**a**<br>**d**<br>**b**<br>**c**<br>**b**<br>**b**<br>**a**<br>**a**<br>**a**<br>**a**<br>**a**<br>**a**|**a**<br>**c**<br>**cd**<br>**c**<br>**d**<br>**b**<br>**a**<br>**a**<br>**b**<br>**b**<br>**b**<br>**a**|
||||


**A**


**B**


**C**



0.25


0.00


−0.25


−0.3 −0.2 −0.1 0.0 0.1 0.2 0.3
CPCoA 1 (59.5%)


CPCoA − 200 ppm − Rhizosphere
4.37% of variance; _P_ = 0.001


0.2


0.0


−0.2


−0.4


−0.4 −0.2 0.0 0.2
CPCoA 1 (66.62%)



−0.2 −0.1 0.0 0.1 0.2
CPCoA 1 (60.8%)



PCoA − 400 ppm


−0.3 −0.2 −0.1 0.0 0.1 0.2
PCoA 1 (64.31%)


CPCoA − 400 ppm − Roots
5.82% of variance; _P_ = 0.001


−0.2 −0.1 0.0 0.1 0.2
CPCoA 1 (76.27%)


CPCoA − 400 ppm − Rhizosphere
2.87% of variance; _P_ = 0.001


−0.2 0.0 0.2 0.4
CPCoA 1 (74.85%)



PCoA − 800 ppm


−0.3 −0.2 −0.1 0.0 0.1
PCoA 1 (57.9%)


CPCoA − 800 ppm − Roots
9.76% of variance; _P_ = 0.001


−0.1 0.0 0.1 0.2
CPCoA 1 (68.65%)


CPCoA − 800 ppm − Rhizosphere
9.45% of variance; _P_ = 0.001



0.1


0.0


−0.1


0.50



PCoA − 200 ppm


−0.2 −0.1 0.0 0.1
PCoA 1 (70.28%)


CPCoA − 200 ppm − Roots
5.38% of variance; _P_ = 0.001



0.2


0.1


0.0


−0.1


0.50


0.25


0.00


−0.25


0.2


0.0


−0.2


−0.4



0.2


0.1


0.0


−0.1


0.2


0.0


−0.2



Rhizobiaceae
Bradyrhizobiaceae
Mycobacteriaceae



Comamonadaceae
Oxalobacteraceae
Microbacteriaceae



Treatment Axenic
SynCom
SynCom + _M. loti_
SynCom + _M. loti_ + AMF



Compartment Roots
Rhizosphere Flavobacteriaceae
Pseudomonadaceae
Xanthomonadaceae



0.2


0.0


−0.2


Alcaligenaceae
Hyphomicrobiaceae
Caulobacteraceae



Intrasporangiaceae
Streptomycetaceae
Bacillaceae



**D**



1.00


0.75


0.50


0.25


0.00



1.00


0.75


0.50


0.25


0.00



1.00


0.75


0.50


0.25


0.00



Microbial profile composition −200 ppm



Microbial profile composition −400 ppm



Microbial profile composition −800 ppm



**E** **F** Bacterial absolute abundance **G**



Variance explained by
treatment (PERMANOVA)



Bacterial absolute abundance



18
16
14
12
10
8
6
4
2
0

200 400 800
CO 2 (ppm)


Compartment Roots Rhizosphere



75


50


25


0



4


3


2


1



Total organic carbon

released by roots


400 800 400 800 400 800

CO 2 (ppm)


**A** Axenic **B** AMF **C**



Axenic



CO 2 (ppm)



_M. loti_



SW CO 2 (ppm)

1 1



SW CO 2 (ppm) SW CO 2 (ppm) SW

1 1 1



CO 2 (ppm)



AMF


SW





200
400
800



200
400
800





R−S R


TPW





R−S R


TPW





CC


CO 2



CC


CO 2







CC


CO 2











WN



WN



**D** **E** _M. loti_ **F**



WN


SynCom + _M. loti_ + AMF



SynCom



SynCom + _M. loti_



SW CO 2 (ppm) SW

1 1



SW



CO 2 (ppm)



SW CO 2 (ppm) SW

1 1



CO 2 (ppm)





200
400
800



200
400
800



CO 2 (ppm)

200
400
800


R−S R


TPW


CO 2 (ppm)

200
400
800


R−S R


TPW





R−S R


TPW







R−S R


TPW



CC


CO 2



CC


CO 2











CC


CO 2







WN


Treatment Axenic
AMF
_M. loti_



SynCom
SynCom + _M. loti_
SynCom + _M. loti_ + AMF



WN - Total weight of nodules
NN - Number of nodules
CO 2 - Normalized CO 2 fluxes
CC - Chlorophyll content
SL - Shoot length



WN



WN


SW - Shoot fresh weight
RW - Root fresh weight
R-S R - Root-to-shoot ratio
TPW - Total plant fresh weight
BAB - Bacterial absolute abundance