¹³C isotope fractionation during rhizosphere respiration of C₃ and C₄ plants

Abstract

Stable carbon isotopes are used extensively to partition total soil CO₂ efflux into root-derived rhizosphere respiration or autotrophic respiration and soil-derived heterotrophic respiration. However, it remains unclear whether CO₂ from rhizosphere respiration has the same δ¹³C value as root biomass. Here we investigated the magnitude of ¹³C isotope fractionation during rhizosphere respiration relative to root biomass in six plant species. Plants were grown in a carbon-free sand-perlite medium inoculated with microorganisms from a farm soil for 62 days inside a greenhouse. We measured the δ¹³C value of rhizosphere respiration using a closed-circulation 48-hour CO₂ trapping method during 40~42 and 60~62 days after sowing. We found a consistent depletion in ¹³C (0.9~1.7‰) of CO₂ from rhizosphere respiration relative to root biomass in three C₃ species (Glycine max L. Merr., Helianthus annuus L. and Triticum aestivum L.), but a relatively large depletion in ¹³C (3.7~7.0‰) in three C₄ species (Amaranthus tricolor L., Sorghum bicolor (L.) Moench and Zea mays L. ssp. mays). Overall, our results indicate that CO₂ from rhizosphere respiration is more ¹³C-depleted than root biomass. Therefore, accounting for this ¹³C fractionation is required for accurately partitioning total soil CO₂ efflux into root-derived and soil-derived components using natural abundance stable carbon isotope methods.

Introduction

Stable carbon isotopes are commonly used in studying carbon cycles at various temporal and spatial scales (Bowling et al. 2008; Pataki et al. 2007). Studies that partition NEE (Net Ecosystem Exchange) into photosynthetic and respiratory components (e.g. Bowling et al. 2001), partition total ecosystem respiration into aboveground, root and microbial components (e.g. Tu and Dawson 2005), or partition total soil respiration into root-derived and soil-derived components (e.g. Lin et al. 1999) often use the δ¹³C values of organic carbon in total or each ecosystem component (leaf, stem, root, and litter) to estimate the δ¹³C values of respiration generated by total or each component. If the δ¹³C value of plant respiration is different than that of organic carbon substrate or bulk material, the results of these studies will need to be reconsidered and modified accordingly. Therefore, knowing carbon isotopic fractionation during major carbon cycle processes, such as rhizosphere respiration that includes both respiration of roots and respiration of rhizosphere microorganisms utilizing carbon substrates derived from live roots, is often crucial for the proper use and the reliability of isotope approaches (Werth and Kuzyakov 2010).

Carbon isotopic fractionation during plant respiration was initially assumed to be negligible (Flanagan and Ehleringer 1998; O’Leary 1981). Lin and Ehleringer (1997) cultured leaf mesophyl protoplasts of two plants (Phaseolus vulgaris and Zea mays) with carbohydrates of known δ¹³C values as the carbon source and found no significant differences in δ¹³C value between mitochondrial dark respiration and the substrates. Cheng (1996) grew Triticum aestivum in sand-vermiculite mixture inoculated with 1 g of soil and found that the δ¹³C value of CO₂ from rhizosphere respiration measured during 14~16 days after seedling emergence was virtually the same as the δ¹³C value of bulk root biomass. However, more recent studies suggested that CO₂ from plant respiration was remarkably ¹³C-enriched or ¹³C-depleted in comparison to bulk plant materials or assumed respiratory substrates (Bowling et al. 2008; Werth and Kuzyakov 2010). For example, Schnyder and Lattanzi (2005) used a special gas exchange system to measure the δ¹³C value of shoot or root respiratory CO₂ for two herbaceous species (Lolium perenne and Paspalum dilatatum) grown in quartz sand and found that shoot respiratory CO₂ was ¹³C-enriched relative to shoot biomass, while root respiratory CO₂ was ¹³C-depleted compared to root biomass. Moreover, relative to the number of studies on ¹³C fractionation during leaf respiration (e.g. Sun et al. 2009; Xu et al. 2004), there were much fewer studies on ¹³C fractionation during root respiration (e.g. Bathellier et al. 2009; Klumpp et al. 2005) and only one study on ¹³C fractionation during rhizosphere respiration (Cheng 1996), which is likely because the isotopic composition of root respiration and rhizosphere respiration is difficult to measure (Bowling et al. 2008; Werth and Kuzyakov 2010). Because additional ¹³C fractionation may occur during both rhizodeposition (i.e. rhizodeposits may have a different δ¹³C value than bulk root tissue, Werth and Kuzyakov 2005) and microbial respiration of rhizodeposits (i.e. microbial respired CO₂ may be different in δ¹³C value compared to rhizodeposits, Fernandez and Cadisch 2003; Mary et al. 1992), the overall ¹³C fractionation may be different between root respiration and rhizosphere respiration that includes both root respiration and rhizomicrobial respiration.

Most published studies on ¹³C fractionation during root respiration were based on snapshot measurements (minutes to hours) of the δ¹³C value of respiratory CO₂ from excised roots (Gessler et al. 2007; Wegener et al. 2010) or roots grown in sand or nutrient solution without the presence of soil microorganisms (Bathellier et al. 2008; Klumpp et al. 2005), and thus did not include the rhizomicrobial respiration by rhizosphere microorganisms utilizing materials released from live roots. In order to partition total soil respiration into rhizosphere respiration that includes both root respiration and rhizomicrobial respiration (root-derived, autotrophic respiration) and microbial decomposition of soil organic matter (soil-derived, heterotrophic respiration) using a two end-member isotope mixing model (Cheng 1996), we need to know δ¹³C values of rhizosphere respiration integrated over days or seasons. Furthermore, ¹³C fractionation during root respiration varies considerably between plant species (Klumpp et al. 2005; Schnyder and Lattanzi 2005). It remains unclear whether ¹³C fractionation during rhizosphere respiration also differs among species, particularly between C₃ and C₄ plants. Some previous studies have reported differences in ¹³C fractionation during root respiration between a C₃ plant and a C₄ plant (Schnyder and Lattanzi 2005), fungal respiration of sucrose derived from C₃ and C₄ plants (Henn and Chapela 2000), and microbial decompositions of residues of C₃ and C₄ plants (Fernandez et al. 2003; Schweizer et al. 1999). Undoubtedly, the issue of ¹³C isotopic fractionation associated with rhizosphere respiration requires further investigation.

In this study we grew three C₃ plants and three C₄ plants in carbon-free sand-perlite mixture inoculated with microorganisms from a farm soil inside a greenhouse for 62 days. We then measured δ¹³C values of bulk root biomass and CO₂ derived from roots and the associated microorganisms (i.e. rhizosphere respiration) during 40~42 and 60~62 days after sowing, using a closed-circulation 48-hour CO₂ trapping method (Cheng et al. 2003). Our primary goal was to answer two questions: (1) Is there a difference in δ¹³C value between bulk root biomass and rhizosphere respiration measured during a two-day period? (2) If the answer to (1) is yes, does the difference vary among species and with growth stages?

Materials and methods

Experimental setup

The experiment was conducted in a greenhouse at University of California, Santa Cruz. We made 36 polyvinyl chloride (PVC) pots (diameter 8 cm, height 15 cm). Each pot was closed at the bottom with a rubber stopper and had an air inlet and an air outlet consisting of clear plastic tubing. We filled each pot with 500 g burned and acid-washed carbon-free sand, 90 g carbon-free perlite, and 10 g soil as inoculant. The sand, perlite and soil were well mixed in each pot before planting seeds. The soil was a sandy loam (Mollisol) collected from a farm on the university campus, with 1.5% C, 0.14% N and δ¹³C value of −26.65‰. Various crops and vegetables (mostly C₃ plants, sunflower, soybean, strawberry, lettuce, etc.) have been grown in the farm since it was converted from a meadow dominated by C₃ annual grasses in 1974. Six pots were planted with each of the following six species: soybean (Glycine max L. Merr.), sunflower (Helianthus annuus L.), wheat (Triticum aestivum L.), amaranthus (Amaranthus tricolor L.), sorghum (Sorghum bicolor (L.) Moench), and maize (Zea mays L. ssp. mays). We planted four seeds and kept one plant per pot after seedling emergence for all species except wheat (planted 10 seeds, kept three plants). All pots were flushed with full-strength Hoagland solution every day. The volume of Hoagland solution increased from 20 ml per day initially to 100~150 ml per day at the end. The extra solution in excess of the holding capacity of growth medium and plant uptake drained out of the pot through the air outlet tube at the bottom of the pot. During the 62-day plant growth period, air temperature was maintained at 25°C during the day and 15°C during the night, relative humidity was maintained at 40%, and photoperiod was set as 14 h with supplemental lighting when needed.

Measurements

During 40~42 and 60~62 days after sowing (DAS), we measured rhizosphere respiration of each plant species in three randomly selected pots using a closed-circulation CO₂ trapping system (Cheng et al. 2003). Briefly, we sealed the pot at the base of the plant with non-toxic silicone rubber (GI-1000, Silicones Inc., NC) and removed CO₂ inside the pot by circulating the isolated air through a soda lime column for 1 h. Then CO₂ produced in the sealed pot was trapped in a 400 mL 0.5 M NaOH solution for 30 min every 6 h during the 48-h period. Three blanks were included to correct for possible contamination from carbonate in the NaOH stock solution and from sample handling. An aliquot of each NaOH solution was analyzed for total inorganic carbon using a Shimadzu TOC-5050A Total Organic Carbon Analyzer and another aliquot was precipitated as SrCO₃ and then analyzed for δ¹³C (relative to PDB standard) using a PDZ Europa ANCA-GSL elemental analyzer interfaced to a PDZ Europa 20–20 isotope ratio mass spectrometer at the Stable Isotope Facility at University of California, Davis (Harris et al. 1997). The δ¹³C values measured in SrCO₃ were corrected for a small amount of contamination from carbonate in the NaOH stock solution and from sample handling (Cheng et al. 2003).

Immediately after CO₂ trapping (42 and 62 DAS), the pots were destructively sampled. Shoots and roots were harvested, washed with deionized water, dried in an oven at 60°C for 48 h, and weighed. We then ground the dry shoot and root samples in a ball mill and analyzed them for C%, N%, δ¹³C and δ¹⁵N using a Carlo Elba 1108 elemental analyzer interfaced to a ThermoFinningan Delta Plus XP isotope ratio mass spectrometer at the Stable Isotope Laboratory at University of California, Santa Cruz.

Statistical analyses

Independent-samples t-test was used to test whether the variables (δ¹³C and rhizosphere respiration rate) were significantly different between the two trapping periods (Table 1). Paired-samples t-test was used to test whether the difference in δ¹³C values between shoot, root, and CO₂ from rhizosphere respiration was significantly different than zero (Fig. 1).

Table 1 δ¹³C values (‰) of shoot, root, and rhizospheric CO₂ (C_r), differences in δ¹³C values (‰) between shoot, root, and rhizospheric CO₂, and rhizosphere respiration rate (C_r, g C/g root N/day). Values represent means of three replicates with standard error in parenthesis. Different letters within each column for each species represent significant differences in mean value between two sampling periods (Independent-samples t-test, n = 3, P < 0.05); bold values of δ¹³C difference (Root—Shoot, CO₂—Shoot, and CO₂—Root) are statistically different than zero (Paired-samples t-test, n = 3, P < 0.05); DAS stands for days after sowing

Fig. 1

The difference in δ¹³C values between rhizospheric CO₂ (C_r) and root or shoot biomass (i.e. carbon isotopic fractionation during rhizosphere respiration relative to root or shoot biomass) in three C₃ plants (soybean, sunflower and wheat) and three C₄ plants (amaranthus, sorghum and maize) in this study. Error bars represent standard errors of the mean. Statistical comparisons were between the isotopic fractionation and the “0” line (Paired-samples t-test, n = 6), ***P < 0.001, **P < 0.01, *P < 0.05, ns-P > 0.05

Results

All plants were at vegetative stage during the first trapping period (40~42 DAS). During the second trapping period (60~62 DAS), sunflower and soybean were at flowering stage, wheat was at grain-filling stage, while the three C₄ species (amaranthus, sorghum and maize) remained at vegetative stage. Plant biomass (data not shown) was significantly higher during the second trapping period than during the first trapping period, while specific rhizosphere respiration rate (g C/g root N/day, Table 1) and shoot or root N concentration (data not shown) did not change significantly between the two trapping periods. Sunflower and wheat showed higher biomass than other species, while amaranthus and maize had nearly twice specific rhizosphere respiration rate as other species (Table 1).

Shoot and root δ¹³C values were approximately −14‰ in the three C₄ plants, except that amaranthus shoot δ¹³C value was −16‰ (Table 1). However, shoot and root δ¹³C values were about −28‰ in the three C₃ plants, except that wheat showed relatively large variations in δ¹³C value (−26.5‰ to 29.3‰) (Table 1). Most shoot and root δ¹³C values did not change significantly between the two trapping periods (P > 0.05), except that wheat (shoot and root) and amaranthus (shoot) was more ¹³C-depleted during the second trapping period than during the first trapping period (P < 0.05). Overall, root had similar δ¹³C values with shoot in soybean, sunflower, sorghum and maize (P > 0.05), but root was more ¹³C-enriched than shoot in wheat (1.0‰, n = 6, P < 0.01) and amaranthus (2.4‰, n = 6, P < 0.001) (Table 1).

We calculated the difference in δ¹³C value between root or shoot bulk organic matter and CO₂ from rhizosphere respiration. In general, the ¹³C fractionation during rhizosphere respiration relative to root biomass did not change significantly between the two trapping periods (P > 0.05), except that it was significantly higher (P < 0.05) during the second trapping period (4.5‰) than during the first trapping period (3.0‰) in maize (Table 1). Although we noticed the possible effect of plant phenology or growing period on ¹³C fractionation during rhizosphere respiration, particularly for maize, we used the average value during the two trapping periods (n = 6) to estimate an integrative ¹³C fractionation during rhizosphere respiration for each species (Fig. 1). Overall, CO₂ from rhizosphere respiration was slightly but significantly (n = 6, P < 0.05) more ¹³C-depleted than root biomass by 0.9 ± 0.2‰ in wheat, 1.0 ± 0.3‰ in sunflower, and 1.7 ± 0.2‰ in soybean. However, ¹³C fractionation during rhizosphere respiration was much higher in the three C₄ species. Compared to root biomass, rhizospheric CO₂ was remarkably (n = 6, P < 0.001) more ¹³C-depleted by 3.7 ± 0.4‰ in maize, 6.5 ± 0.2‰ in sorghum, and 7.0 ± 0.2‰ in amaranthus. Moreover, because shoot and root had similar δ¹³C values, the difference in δ¹³C value between shoot biomass and CO₂ from rhizosphere respiration was similar to or slightly lower than that between root biomass and CO₂ from rhizosphere respiration (Fig. 1)

Discussion

Comparison with previous studies

We summarized all available data in the literature on carbon isotopic fractionation during root respiration or rhizosphere respiration in Table 2. The difference in the measured ¹³C fractionation may result from three factors. First, all these studies except three from our lab (Cheng 1996; Dijkstra and Cheng 2007; and this study) excluded rhizosphere microorganisms and thus an important component of root-derived CO₂—rhizomicrobial respiration. If rhizomicrobial respiration accounts for a significant proportion of rhizosphere respiration and ¹³C fractionation differs between root respiration and rhizomicrobial respiration, then the measured ¹³C fractionation during root respiration (e.g. Klumpp et al. 2005; Werth and Kuzyakov 2005) would be different than that during rhizosphere respiration (Cheng 1996; Dijkstra and Cheng 2007; this study). However, simultaneous measurements of both root respiration (e.g. without inoculation, maintain a microbe-free growth medium) and rhizosphere respiration (e.g. inoculate with native soil microorganisms that consume rhizodeposits) of the same species, under the same environment and using the same method, are required to directly test this hypothesis.

Table 2 The magnitude of carbon isotope fractionation during root respiration (method 1–3) or rhizosphere respiration (method 4) relative to root biomass in different plant species by different studies. For the six species in this study, values represent means of six replicates with standard error in parenthesis

Second, many published studies conducted snapshot measurements (<30 min) of δ¹³C value of root respiratory CO₂ based on a gas exchange system (Badeck et al. 2005; Schnyder and Lattanzi 2005) or a direct gas sampling system (Gessler et al. 2009; Wegener et al. 2010). Both ¹³C values of respiratory substrates and the relative contribution of different metabolic pathways to CO₂ evolution may change within a day, therefore ¹³C fractionation during plant respiration can change significantly on a diurnal basis, as shown in previous studies on ¹³C fractionation during leaf respiration (Gessler et al. 2007; Sun et al. 2009; Wegener et al. 2010). Diurnal variations in ¹³C fractionation during root respiration and rhizosphere respiration are probable, but direct experimental evidence is not yet available. In this study we measured δ¹³C value of rhizospheric CO₂ integrated over a two-day period using a closed-circulation CO₂ trapping system (Cheng 1996; Cheng et al. 2003). The CO₂ trapping efficiency with this system was greater than 99%, eliminating preferential sorption of ¹³CO₂ vs. ¹²CO₂. The CO₂ from microbial respiration of the 10-g soil (with a δ¹³C value of −26.65‰) used to inoculate the sand-perlite mixture with microorganisms was less than 1% of the CO₂ from rhizosphere respiration. Therefore, the possible contamination of the δ¹³C value of root-derived CO₂ by the δ¹³C value of soil-derived CO₂ is negligible (<0.1‰). The circulation system was carefully maintained to avoid any leakage of ambient air, and blanks (the same circulation system in Fig. 1 without the pot) were included to correct for possible contamination from carbonate in the NaOH stock solution and from sample handling (Cheng et al. 2003). Therefore, unlike the gas exchange system or the direct gas sampling system which measured CO₂ produced during a short time period (<30 min) used in previous studies (Bathellier et al. 2008; Gessler et al. 2007; Klumpp et al. 2005; Moyes et al. 2010), our closed-circulation CO₂ trapping system rendered a more integrative measurement of ¹³C fractionation during rhizosphere respiration.

Third, different species may also contribute to the difference in measured ¹³C fractionation in addition to different CO₂ sources (root respiration vs. rhizosphere respiration) and methods of CO₂ measurement (snapshot measurement vs. integrative measurement) discussed above. Our results of ¹³C fractionation during rhizosphere respiration of three C₃ species (soybean 1.7‰, sunflower 1.0‰ and wheat 0.9‰) fall in the range of previous results of ¹³C fractionation during root or rhizosphere respiration of same or similar C₃ species (Phaseolus vulgaris 0~2‰, sunflower 0.5~2.0‰, wheat 0.2‰), but are significantly lower than previous results of ¹³C fractionation during root respiration of other C₃ species (Lolium perenne 3.5~5.4‰, Paspalum dilatatum 0.5~5‰ and Ricinus communis 2~4‰, Table 2). Note that two recent field studies (Gessler et al. 2007; Moyes et al. 2010) found ¹³C-enrichment (up to 9‰) of root respiration of two woody species (Eucalyptus delegatensis and Acer negundo) compared to root biomass, a result that differs markedly from those of laboratory studies. The short-term direct gas sampling method used in these two studies (incubation of excised roots in a root chamber for a few minutes) may partly contribute to this difference. In addition, our results of ¹³C fractionation during rhizosphere respiration of three C₄ species (maize 3.7‰, sorghum 6.5‰ and amaranthus 7.0‰) were significantly higher than those of three C₃ species in this study (0.9~1.7‰). Only two studies have reported data on ¹³C fractionation during root respiration in C₄ species (Table 2). Using a gas exchange system to measure the δ¹³C value of respiratory CO₂ from roots grown in quartz sand, Schnyder and Lattanzi (2005) reported 0.5~5.0‰ ¹³C-depletion (depending on growth temperature) in root respiration relative to root biomass in Paspalum dilatatum. However, using a closed-circulation CO₂ trapping system to measure the δ¹³C value of respiratory CO₂ from roots grown in nutrient solution, Werth and Kuzyakov (2005) reported a very small difference in δ¹³C value (<1‰) between root respiration and root biomass in maize. Therefore, although the relative ¹³C fractionation during rhizosphere respiration between C₃ plants and C₄ plants needs further investigation, our data clearly showed higher ¹³C fractionation during rhizosphere respiration in the three C₄ plants (amaranthus, sorghum, and maize) than in the three C₃ plants (soybean, sunflower, and wheat).

Possible mechanisms for ¹³C fractionation during rhizosphere respiration

Rhizosphere respiration has two components—root respiration and rhizomicrobial respiration. First, root respiration may be ¹³C-depleted relative to root biomass (Gessler et al. 2009; Klumpp et al. 2005; Schnyder and Lattanzi 2005; but see Gessler et al. 2007 and Moyes et al. 2010). A number of hypotheses have been advanced to explain this depletion in root respiration compared to likely respiratory substrates. The pentose phosphate pathway (PPP), the tricarboxylic acid cycle (TCA, or the Krebs cycle), and phosphoenolpuruvate carboxylase reaction (PEPC, or the refixation of CO₂ by phosphoenolpuruvate carboxylase) all release ¹³C-depleted CO₂ relative to the original substrate glucose, while PDC (pyruvate dehydrogenation complex) releases ¹³C-enriched CO₂ relative to glucose (Barbour and Hanson 2009; Wingate 2008). The relative allocation of carbon to the different pathways in roots may lead to different carbon isotope fractionation in root respiration among different species or under different environmental conditions (Bathellier et al. 2009; Gessler et al. 2009; Wegener et al. 2010).

Second, rhizomicrobial respiration, or the respired CO₂ from microorganisms utilizing root-derived substrates in the rhizosphere, may have different δ¹³C values than root biomass. The fractionation during rhizomicrobial respiration may occur during two processes: (1) root-derived carbon compounds in the rhizosphere (or rhizodeposition) may have different δ¹³C values than root biomass (Werth and Kuzyakov 2005); and (2) microbially respired CO₂ may be different in δ¹³C value with the root-derived carbon substrates taken up and utilized by them (Fernandez and Cadisch 2003; Henn and Chapela 2000; Mary et al. 1992). The first process is mainly determined by the chemical composition of rhizodeposition, which is a mixture of different compounds (Lynch and Whipps 1990) that may differ significantly in δ¹³C value (Bowling et al. 2008). Few studies have compared the ¹³C values of individual compounds or overall rhizodeposits with bulk root tissue. The second process is controlled by many factors: temperature, isotopic distribution within the substrates, chemical nature of the substrates, metabolic pathways of carbon, and physiological conditions of microbial growth (Fernandez et al. 2003; Werth and Kuzyakov 2010). Despite more studies on this process than the first process, the results remain inconsistent among studies (e.g. significant depletion in Blair et al. 1985 and Mary et al. 1992; not significant in Ehleringer et al. 2000 and Ekblad et al. 2002) and change over time within studies (e.g. Fernandez et al. 2003; Schweizer et al. 1999). There have been some studies on fractionation during root respiration (Table 2), but no studies are available on fractionation during rhizomicrobial respiration, mainly because it is difficult to collect root-derived carbon substrates for ¹³C measurement and concurrently measure the δ¹³C values of substrate and respired CO₂ for microbial respiration (Werth and Kuzyakov 2010). The ¹³C depletion in rhizosphere respiration relative to root biomass observed in this study may be partly caused by the ¹³C depletion in rhizomicrobial respiration, particularly during the microbial uptake and utilization of root-derived substrates, in addition to the ¹³C depletion in root respiration relative to root biomass or respiratory substrate as discussed above.

Why does the ¹³C fractionation differ significantly between the three C₃ plants and the three C₄ plants in this study? Although the exact mechanisms of this phenomenon are unknown at this point and need further investigation, here we point out two possible mechanisms. First, during root respiration, plants with different photosynthetic pathways (e.g. C₃ vs. C₄) may differ in the relative contribution of different metabolic pathways to root respired CO₂ (Bathellier et al. 2009; Gessler et al. 2009; Wegener et al. 2010), which can lead to different ¹³C values of CO₂ from root respiration. However, direct comparisons of carbon substrate allocation to different root metabolic pathways between C₃ and C₄ plants are lacking. Second, during rhizomicrobial respiration, differences in ¹³C distribution in sugars derived from C₃ vs. C₄ plants (Rossmann et al. 1991) and the possible different fractionation during fungi uptake of sugars derived from C₃ vs. C₄ plants (Henn and Chapela 2000), may further contribute to different ¹³C values of CO₂ from rhizomicrobial respiration. As only three C₃ and C₄ plants are included in this study, considering the large variation in ¹³C fraction during root respiration among species (Table 2), we urgently need more studies to determine whether the different ¹³C fractionation during rhizosphere respiration is species-specific or photosynthesis-pathway-specific.

Implications for partitioning soil respiration

Our results showed relatively small and consistent differences in δ¹³C values between root biomass and rhizospheric CO₂ in sunflower, soybean and wheat (0.9~1.7‰), but the differences were relatively large in maize, sorghum and amaranthus (3.7~7.0‰). This has important implications for partitioning total soil respiration (C_t) into root-derived (C_r) and soil-derived (C_s) components using isotope mixing models: C_r = C_t (δ¹³C_s – δ¹³C_t) / (δ¹³C_s – δ¹³C_r), C_s = C_t – C_r, where δ¹³C_t, δ¹³C_r, and δ¹³C_s are δ¹³C values of C_t, C_r, and C_s respectively (Cheng 1996). For example, if we grow a C₃ plant (e.g. sunflower) in a C₄ soil in an ambient air (−8‰), assume that soil-derived CO₂ has a δ¹³C value of −14‰ (δ¹³C_s = −14‰), root biomass has a δ¹³C value of −28‰, and total soil respiration has a δ¹³C value of −21‰ (δ¹³C_t = −21‰). If we use δ¹³C value of root biomass for δ¹³C value of root-derived CO₂ (δ¹³C_r = −28‰), rhizosphere respiration would be 50% of total soil respiration. However, if we consider the ¹³C-depletion in rhizosphere respiration relative to root biomass by 1~7‰ and use δ¹³C value of −29~−35‰ for δ¹³C_r, rhizosphere respiration would be 46.7~33.3% of total soil respiration (Fig. 2a). Therefore, for C₃ plants grown in C₄ soils, not accounting for the ¹³C-depletion in rhizosphere respiration relative to root biomass would slightly overestimate root-derived autotrophic respiration and underestimate soil-derived heterotrophic respiration. This suggests that previous estimates of positive rhizosphere priming effect for C₃ plants grown in C₄ soils (i.e. higher SOM decomposition rate in planted soil than in unplanted soil; Cheng et al. 2003) was conservative, because considering the ¹³C-depletion in rhizosphere respiration relative to root biomass would yield a higher rate of SOM decomposition in the presence of plants and thus a higher, positive rhizosphere priming effect (Fig. 2a). This situation is also applicable to the cases of C₃ plants labeled with ¹³C-depleted CO₂ and grown in native C₃ soils (Dijkstra et al. 2006; Zhu and Cheng 2010). The influence of the ¹³C-depletion of rhizospheric CO₂ on the CO₂ partitioning would be reduced if shoot δ¹³C value was used as the end-member for rhizospheric CO₂, because the isotopic fractionation between shoots and rhizospheric CO₂ was smaller, especially for wheat (Table 1, Fig. 1).

Fig. 2

A sensitivity analysis to show how the ¹³C-depletion during rhizosphere respiration relative to root biomass may affect the partitioning of total soil respiration (C_t) into root-derived (C_r) and soil-derived (C_s) components as well as the rhizosphere priming effect (RPE). C_r = C_t (δ¹³C_s – δ¹³C_t) / (δ¹³C_s – δ¹³C_r), C_s = C_t – C_r, RPE = [C_s(planted) – C_s(unplanted)] / C_s(unplanted) X 100%. a Grow a C₃ plant in C₄-plant-derived soil, assume δ¹³C(root) = −14‰, δ¹³C_s(unplanted) = −28‰, δ¹³C_t = −21‰, C_t = 100, and C_s(unplanted) = 40. b Grow a C₄ plant in C₃-plant-derived soil, assume δ¹³C(root) = −28‰, δ¹³C_s(unplanted) = −14‰, δ¹³C_t = −21‰, C_t = 100, and C_s(unplanted) = 60

Additionally, in the case of growing a C₄ plant (e.g. maize) in a C₃ soil, neglecting ¹³C fractionation during rhizosphere respiration will result in substantial underestimation of root-derived CO₂, and overestimation of soil-derived CO₂ in planted treatments. To illustrate this point, assume that δ¹³C_s = −28‰, δ¹³C_t = −21‰, and δ¹³C value of root biomass is −14‰. C_r (rhizosphere respired CO₂) would be 50% of C_t without figuring in the ¹³C fractionation, but 53.8~100% of C_t if 1~7‰ ¹³C fractionation is included in the calculation (Fig. 2b). For example, Fu and Cheng (2002) showed that the SOM decomposition rates in a C₃ grassland soil planted with sorghum and amaranthus were 9% and 5% lower than in the unplanted soil. If the ¹³C depletion in rhizosphere respiration relative to root biomass is used in a recalculation, the rate of SOM decomposition in the presence of plants would be much lower, and thus the negative rhizosphere priming effect would be much stronger (Fig. 2b).

The ¹³C depletion in rhizosphere respiration relative to root biomass also has implications in partitioning ecosystem respiration (R_eco) into different components (aboveground plant respiration—R_above, rhizosphere respiration—R_rhizo, and microbial respiration—R_mic). In natural systems without significant land use change, microbial respiration tends to be slightly ¹³C-enriched compared to rhizosphere and aboveground plant respiration, thus a three-source mixing model can be used to partition R_eco into R_above, R_rhizo and R_mic (Tu and Dawson 2005). If we account for the ¹³C-depletion in R_rhizo relative to root biomass, the fraction of microbial respiration (f_micro) in total ecosystem respiration will be higher than original estimate assuming no fractionation in rhizosphere respiration. This suggests that neglecting the ¹³C depletion in rhizosphere respiration relative to root biomass would underestimate the heterotrophic component of soil respiration (R_mic) and thus overestimate net ecosystem production (NEP = NPP (net primary production) – R_mic). Moreover, in studies that partition NEP into GPP (gross primary production) and R_eco at ecosystem (e.g. Bowling et al. 2001) and global (e.g. Fung et al. 1997) scales using carbon isotopes, the isotopic composition of NEP, GPP and R_eco are either directly measured by eddy covariance combined with flask sampling (for NEP/NEE) and Keeling plots (for R_eco) or indirectly estimated using theoretical models (for GPP). Therefore, accounting for ¹³C deletion in rhizosphere respiration relative to root biomass will not affect these results.

Conclusions

In conclusion, rhizospheric CO₂ was ¹³C-depleted (by up to 7‰) relative to root biomass in the six species studied. It appears that the ¹³C fractionation during rhizosphere respiration relative to root biomass (ΔC_r) did not change significantly between two measurement periods or growth stages, but varied significantly among species and particularly physiological groups (C₃ vs. C₄). Three C₃ plants (wheat, sunflower and soybean) showed small and consistent ¹³C-depletion in rhizosphere respiration (0.9~1.7‰), while three C₄ plants (maize, sorghum and amaranthus) had relatively large ¹³C-depletion in rhizosphere respiration (3.7~7.0‰) compared to root biomass. Whether the fractionation is photosynthetic-pathway-specific or simply species-specific should be tested with future research with more C₃ and particularly C₄ species. The mechanisms leading to ΔC_r remain unclear at this point. Further studies are needed to understand: (1) ΔC_r in other species, particularly woody species, and under field conditions; (2) ontogenetic, diurnal and seasonal changes in ΔC_r; and (3) responses of ΔC_r to changes in environmental and physiological factors (e.g. temperature, respiratory quotient). The new and emerging technologies (e.g. tunable diode laser absorption spectroscopy or TDLAS, and cavity ring-down spectroscopy or CRDS) that provide instantaneous and continuous field measurements of δ¹³C of CO₂ from plant (leaf, stem or root) or microbial respiration in combination with specially-designed chambers (e.g. Wingate et al. 2010) is particularly recommended. Nevertheless, our findings demonstrate that the ¹³C fractionation in rhizosphere respiration relative to root biomass should be seriously considered in future studies involving carbon flux partitioning using natural abundance carbon isotope methods (Baggs 2006; Kuzyakov 2006).