Arbuscular Mycorrhizal Fungi Increase Organic Carbon Decomposition Under Elevated CO2

See allHide authors and affiliations

Science  31 Aug 2012:
Vol. 337, Issue 6098, pp. 1084-1087
DOI: 10.1126/science.1224304


The extent to which terrestrial ecosystems can sequester carbon to mitigate climate change is a matter of debate. The stimulation of arbuscular mycorrhizal fungi (AMF) by elevated atmospheric carbon dioxide (CO2) has been assumed to be a major mechanism facilitating soil carbon sequestration by increasing carbon inputs to soil and by protecting organic carbon from decomposition via aggregation. We present evidence from four independent microcosm and field experiments demonstrating that CO2 enhancement of AMF results in considerable soil carbon losses. Our findings challenge the assumption that AMF protect against degradation of organic carbon in soil and raise questions about the current prediction of terrestrial ecosystem carbon balance under future climate-change scenarios.

Arbuscular mycorrhizal fungi (AMF), which form associations with roots of ~80% of land plant species, obtain carbon (C) from their host plants in return for mineral nutrients (1, 2). AMF utilize a large proportion (up to 20%) of net plant photosynthates under ambient atmospheric CO2 (aCO2) (3, 4), deposit slow cycling organic compounds such as chitin and glomalin (1, 5), and protect organic matter from microbial attack by promoting soil aggregation (6). AMF thus play a critical role in the global C cycle. Atmospheric CO2 enrichment often increases plant photosynthate allocation to AMF and stimulates the growth of AMF (3, 79), leading to a proposition that global soils may sequester more C through mycorrhizal symbioses under future scenarios of elevated CO2 (eCO2) (3, 5, 712). This hypothesis, however, does not consider the effect of AMF on decomposition under eCO2. Indeed, AMF growth can result in enhanced decomposition of complex organic material and alter plant N uptake (1315).

We conducted four independent but complementary experiments to investigate how CO2 stimulation of AMF affects organic C decomposition in soil and the subsequent N dynamics in the plant-soil system by combining dual 13C/15N labeling and hyphae-ingrowth techniques (16). We first ascertained the effect of eCO2 [main plot, n = 4; ambient at 380 versus elevated at 580 parts per million by volume (ppmv)] and N addition (subplot; control at 0 versus added at 5 g N m−2) on mycorrhizal mediation of decomposition in a N-poor soil, using a model mycorrhizal plant community consisting of AMF growing on roots of Avena fatua (14) in microcosms (fig. S1). The high levels of CO2 and N used in our experiment correspond to projected atmospheric CO2 concentrations and N deposition rates in North America during the 21st century (17). We chose A. fatua, an annual C3 grass native to Eurasia, because it has invaded many temperate grasslands and is considered one of the worst weeds in agricultural fields in North America.

After incubation for 10 weeks, AMF enhanced decomposition within hyphae-ingrowth bags (P < 0.001, Fig. 1A; also see 13C in fig. S2A). eCO2 had no impact on total soil C in the absence of AMF (NAMF) (P > 0.1, Fig. 1A), but significantly reduced it by 9% in the presence of AMF (P < 0.01, Fig. 1A; see 13C in fig. S2A), consistent with the CO2 stimulation of AMF infection of plant roots (P < 0.05, fig. S3A). Notably, the CO2 effect on AMF-mediated decomposition mainly occurred under the N amendment, with a reduction in total C in hyphae-ingrowth bags of 19% in soil (AMF+S) and 10% in quartz sand (AMF+Q) (Fig. 1A; see 13C in fig. S2A).

Fig. 1

The effect of arbuscular mycorrhizal fungi on organic C decomposition. (A) C remaining (%) within hyphae-ingrowth bags after 10 weeks of incubation under different CO2 and N concentrations. +S and +Q refer to autoclaved sandy loam soil (S) and quartz sand (Q) in hyphae-ingrowth bags, respectively. Blank and gray bars denote ambient CO2 without and with added N, respectively; hatched and black bars denote elevated CO2 without and with added N, respectively. Data shown (means ± SEM) are based on the fitted mixed model. The main effects of N, and CO2 × N and CO2 × N × AMF interactions were not significant (P > 0.05). (B and C) C remaining (%) within hyphae-ingrowth cores after 10 weeks of incubation under different CO2 and AMF species treatments (B) and within hyphae-ingrowth bags after 5, 10, and 15 weeks of incubation under different CO2 concentrations in the field (C). Full AMF species name and assemblage composition are in table S1. Gray bars, ambient CO2; black bars, elevated CO2. Data shown (means ± SEM) are based on the fitted mixed model.

Emerging evidence shows that AMF species may differ in their capabilities in acquiring N from decomposing residues (13). However, it is unknown whether the nature of AMF species or communities influences the CO2 effect on residue decomposition. We investigated the effect of three individual AMF species and two AMF assemblages (subplot) on residue decomposition with their host plant A. fatua exposed to two atmospheric CO2 levels (main plot, n = 4; 380 versus 580 ppmv) (16). One AMF assemblage consisted of three species and the other a total of eight species (table S1).

AMF enhanced decomposition in hyphae-ingrowth cores in comparison with the NAMF (P < 0.001, Fig. 1B; see 13C in fig. S2B), particularly under eCO2. Across five AMF treatments, eCO2 on average increased AMF infection of plant roots by 28% (P < 0.05, fig. S3B) and reduced total C by 15% within hyphae-ingrowth cores (P < 0.05, Fig. 1B; see 13C in fig. S2B). The magnitude of the CO2 effect on decomposition differed among the three individual AMF species (P < 0.05), with the high effect found for both Gigaspora margarita and Glomus clarum and the low for Acaulospora morrowiae, but was comparable between the two AMF assemblages (P > 0.1). Taken together, these microcosm experiments indicate that CO2 stimulation of AMF in general enhances organic C decomposition in soils with low N availability.

We also conducted a field study to examine the AMF effect on decomposition in a long-term CO2 (380 versus 560 ppmv) and O3 [20 versus 60 parts per billion by volume (ppbv)] experiment (2 × 2 factorial, n = 4) in a no-till wheat-soybean system (16, 18). We initiated the long-term experiment in May 2005 and carried out the decomposition study in the wheat season of 2008. There were no significant O3 or CO2 × O3 effects on any soil microbial parameter (e.g., biomass C and N, fungi/bacteria ratio, and heterotrophic respiration) (18), AMF biomass and infection of roots, or organic C decomposition within hyphae- and root-ingrowth bags (P > 0.05). However, eCO2 significantly increased both AMF colonization of fine roots collected from root-ingrowth bags (P < 0.001, fig. S3C) and the external AMF biomass as indexed by the biomarker fatty acid 16:1ω5c in the bulk soil (P < 0.05, fig. S3D). Concurrently, eCO2 significantly increased total C losses within hyphae-ingrowth bags across the three sampling points (P < 0.01, Fig. 1C; see 13C in fig. S2C). The instantaneous fractional loss rates for C (k = 1 – Xt/X0, where Xt and X0 are the organic C content at time t and time 0, respectively) induced by the hyphae-ingrowth effect under eCO2 were 29, 41, and 80% higher than those under aCO2, respectively, at weeks 5, 10, and 15 (Fig. 1C), indicating that the CO2 effect on AMF-mediated decomposition did not diminish over time.

To examine whether CO2 enhancement of AMF-mediated decomposition was accompanied with increased plant uptake of N released from decomposing residues, we determined 15N both in plants and hyphae-ingrowth bags and cores. eCO2 substantially reduced the total 15N within hyphae-ingrowth bags and cores in the presence of AMF in all three experiments (fig. S4) and increased AMF-mediated plant 15N uptake in the microcosms (fig. S5). These results provide direct evidence of CO2 enhancement of mycorrhizal N transfer from decomposing organic material to host plants.

We also examined the effect of eCO2 on soil available N pools [ammonium (NH4+) and nitrate (NO3)]. In microcosms where N was limiting and AMF were present, eCO2 reduced soil NH4+ in both experiments (P < 0.01, Fig. 2A; P < 0.05, Fig. 2B), but did not affect levels of soil NO3 (P > 0.1, fig. S6D; P > 0.1, fig. S6E). In the field where soil N was ample (mainly NO3, fig. S6F), eCO2 did not affect soil NH4+ (P > 0.1 for each of three soil layers, fig. S6C) but significantly increased both potential N mineralization (18) and soil NO3 (P < 0.05 for each of three soil layers, Fig. 2C). These results suggest that eCO2 may differentially affect plant acquisition of soil NH4+ and NO3.

Fig. 2

Differential CO2 effects on soil ammonium (NH4+) and nitrate (NO3) and on plant NH4+ and NO3 uptake. (A to C) Net CO2 effect (%) on soil NH4+ under different AMF and N concentrations (A) and different AMF species and assemblages (B) in microcosms, and on soil NO3 of three soil layers in the field (C). (D) A meta-analysis of net CO2 effects (%) on soil NH4+ (n = 44) and NO3 (n = 30), and on plant NH4+ (n = 71) and NO3 (n = 61) uptake. Error bars, 95% confidence intervals. The elevated CO2 effect on a response variable was considered significant if the 95% confidence interval did not overlap with 0.

We subsequently conducted a meta-analysis (16) of 38 studies that quantified the concentrations of soil NH4+ and NO3 and/or the capacity of plant use of NH4+ and NO3 under eCO2 (table S2). These studies encompassed more than 58 species of crop, grass, and tree species (16). eCO2 reduced the capacity of plant NO3 use by 16.2% and increased soil NO3 by 26.7% (Fig. 2D). By contrast, it had no impact on the capacity of plants to use NH4+ but decreased soil NH4+ by 7.9% (Fig. 2D). These differential CO2 effects on soil NH4+ and NO3 agreed with our results and were consistent qualitatively with recent discoveries of eCO2 effects on plant N utilization (19, 20). Together, these results suggest that plants under eCO2 may have to rely more on soil NH4+ for N nutrition, and a high demand for NH4+ may play a major role in mediating the AMF effect on organic C decomposition.

If CO2-induced high-plant demand for NH4+ is a primary driver in mycorrhizally mediated decomposition, high soil NH4+ may partially offset this effect. To test this possibility, we assessed the effect of AMF on decomposition by manipulating soil N transformations with a nitrification inhibitor (dicyandiamide) (21) in our long-term field CO2 and O3 study in the wheat season of 2011 (16). Dicyandiamide had no effect on plant growth and AMF infection of roots (P > 0.1). In the no-dicyandiamide control, eCO2 significantly increased AMF-mediated decomposition (P < 0.05, Fig. 3), consistent with the previous field experiment (Fig. 1C). In the dicyandiamide treatment, however, eCO2 did not affect organic C decomposition in the hyphae-ingrowth bag (P > 0.1, Fig. 3), indicating that the nitrification inhibitor largely offset the impact of eCO2 on AMF-mediated organic C decomposition. These results provide supporting evidence that enhanced plant demand for soil NH4+ may be the primary driver for CO2 enhancement of AMF-mediated decomposition.

Fig. 3

A nitrification inhibitor (dicyandiamide) offset the CO2 effect on organic C decomposition within hyphae-ingrowth bags after 10 weeks of incubation in the field. Gray bars, ambient CO2; black bars, elevated CO2. Data shown (means ± SEM) are based on the fitted mixed model. The letters a and b represent a significant difference between two CO2 levels under the no dicyandiamide treatment. The main O3 effect and the CO2 × O3 interaction were not significant in both dicyandiamide and no-dicyandiamide addition treatments (P > 0.05).

Based on this set of investigations, we therefore propose that eCO2 enhancement of plant N demand prompts plants to invest more C and energy to structures (mainly roots and their associated mycorrhizae) that best garner NH4+ from soil (22), while stimulating NH4+ release from organic materials and reducing NH4+ substrate for nitrification (Fig. 4). Two unique AMF properties enable host plants to compete better against nitrifying microbes for NH4+ in the fine, discrete decomposing hotspots: (i) external AMF hyphae are at least two orders of magnitude longer and three orders of magnitude thinner than roots (1, 15) and can exploit a much larger soil volume and finer soil microsites; and (ii) AMF possess a special N transfer pathway (22, 23) that can transport soil N from external to internal hyphae and to their hosts preferentially as NH4+ with minimal C loss (23). Because AMF generally lack saprotrophic capability (1), CO2 enhancement of AMF for N scavenging likely increases decomposition by stimulating (i.e., priming) saprotrophs in soil through three potential mechanisms. First, AMF likely grow preferentially toward (15), and thus facilitate saprotrophs’ access to, new organic patches (24). Second, AMF slowly release labile C for saprotrophs at relatively low concentrations (3), likely engendering a larger priming effect on decomposition than roots (fig. S7) (2527). And third, rapid removal of newly released NH4+ by AMF likely releases saprotrophs from metabolic repression (28).

Fig. 4

A conceptual framework of AMF-mediated decomposition driven by CO2 enhancement of plant N acquisition. CO2 enhancement of AMF primes residue decomposition and ammonium (NH4+) release and optimizes NH4+ acquisition while reducing nitrification. CO2 inhibition of nitrate (NO3) photo-assimilation constrains the capacity of plant NO3 uptake, prompting plants to rely more on the AMF-mediated pathway of NH4+ (and possibly some simple organic N compounds) acquisition. Solid and dashed arrows represent positive and negative CO2 effects, respectively.

Our findings indicate that CO2 enhancement of AMF may alter terrestrial ecosystem C dynamics by stimulating decomposition of soil organic C in AMF-active zones. This effect will likely occur in its interplay with other controlling factors such as temperature and plant species composition (29). In many agro- or grassland ecosystems where AMF dominate (1), but no aboveground C pool with an annual incremental increase exists, CO2 stimulation of AMF and organic C decomposition will mainly facilitate C turnover belowground, rather than ecosystem C sequestration (30). Even in forests with abundant AMF (e.g., tropical forests) (1), eCO2 stimulation of AMF, although creating a transient C sink in plant biomass by facilitating N transfer from soil to plants and partially alleviating N limitation on plants (31), is likely to reduce the largest carbon stocks (soil C) in the system. Also, our results suggest that the form, rather than just the total amount, of soil N might play a major role in mediating belowground C turnover and plant N acquisition under eCO2, thus offering a theoretical foundation for management of microbial N transformations in soil and plant N utilization to facilitate ecosystem C sequestration under future CO2 scenarios.

Supplementary Materials

Materials and Methods

Figs. S1 to S7

Tables S1 and S2

References (3279)

References and Notes

  1. See supplementary materials on Science Online.
  2. Acknowledgments: We thank F. Chapin III, D. Coleman, Y. Luo, R. Miller, and M. Rillig for valuable comments; M. Gumpertz for advice on statistical analyses; J. Barton, W. Pursley, and E. Silva for technical assistance; and D. Watson and J. Morton for providing mycorrhizal inoculum. L.C. was primarily supported by a fellowship from U.S. Department of Agriculture (USDA)–Agricultural Research Service Plant Science Research Unit (Raleigh, NC) and in part by a USDA grant to S.H. (2009-35101-05351). S.H., L.C. and C.T. conceived experiments 1 to 4. K.O.B. and F.L.B. designed and maintained the long-term CO2 and O3 study. H.D.S and T.W.R. contributed to design of experiments 1 and 2. L.C. performed experiments 1 to 3 and the meta-analysis study; and C.T., F.L.B., and L.Z. performed experiment 4. L.C. and S.H. analyzed the data and mainly wrote the manuscript with inputs from all coauthors. The data reported in this paper are deposited in the Dryad Repository (

Stay Connected to Science

Navigate This Article