Assessing and avoiding C isotopic contamination artefacts in mesocosm-scale 13CO2/12CO2 labelling systems: from biomass components to purified carbohydrates and dark respiration.

Abstract

Background: Quantitative understanding of plant carbon (C) metabolism by ¹³CO₂/¹²CO₂-labelling studies requires absence (or knowledge) of C-isotopic contamination artefacts during tracer application and sample processing. Surprisingly, this concern has not been addressed systematically and comprehensively yet is especially crucial in experiments at different atmospheric CO₂ concentrations ([CO₂]), when experimental protocols require frequent access to the labelling chambers. Here, we used a plant growth chamber-based ¹³CO₂/¹²CO₂ gas exchange-facility to address this topic. The facility comprised four independent units, with two chambers routinely operated in parallel under identical conditions except for the isotopic composition of CO₂ supplied to them (δ¹³C_CO2 -43.5‰ versus -5.6‰). In this setup, dδ¹³C_X (the measurements-based δ¹³C-difference between matching samples X collected from the parallel chambers) is expected to equal dδ¹³C_Ref (the predictable, non-contaminated δ¹³C-difference ), if sample-C is completely derived from the contrasting CO₂ sources. Accordingly, contamination (f_contam) was determined as f_contam = 1- dδ¹³C_X/dδ¹³C_Ref in this experimental setup. Determinations were made for biomass fractions, water-soluble carbohydrate (WSC) components and dark respiration of Lolium perenne (perennial ryegrass) stands following growth for ∼9 weeks at 200, 400 or 800 µmol mol^- 1 CO₂, with a terminal two weeks-long period of extensive experimental disturbance of the chambers.

Results: Contamination was small and similar (average 3.3% ±0.9% SD, n = 18) for shoot and root biomass and WSC fractions (fructan, sucrose, glucose, fructose) at every [CO₂] level. [CO₂] had no significant effect on contamination of these samples. There was no evidence for any contamination of WSC components during extraction, separation and analysis. At 200 and 400 µmol mol^- 1 CO₂, contamination of respiratory CO₂ was close to that of biomass- and WSC-C, suggesting it originated primarily from in vivo-contaminated respiratory substrate. Surprisingly, we found no evidence of contamination of respiratory CO₂ at 800 µmol mol^- 1 CO₂. Overall, contamination likely resulted overwhelmingly from photosynthetic fixation of extraneous contaminating CO₂ which entered chambers primarily during daytime experimental activities.

Conclusions: The labelling facility enables months-long, quantitative ¹³CO₂/¹²CO₂-labelling of large numbers of plants with accuracy and precision across contrasts of [CO₂], empowering eco-physiological study of climate change scenarios. Effective protocols for contamination avoidance are discussed.