Unexpected reversal of C₃ versus C₄ grass response to elevated CO₂ during a 20-year field experiment

Physiological responses of both C3 and C4 species to elevated CO2 (eCO2) have been studied extensively during the last ten years and the FACE facilities have improved significantly our understanding about how eCO2 affects plants from leaf to community level. Recently, Reich et al. (1) reported very interesting and unexpected findings about the long-term acclimation of both photosynthetic types under eCO2, although statistically perhaps not very convincing and without providing a complete explanation of the phenomena observed. In brief, C4 species became more responsive to eCO2 than C3 species and produced more biomass after long-term exposure. Such intriguing result was associated with increases in soil nitrogen mineralization after long exposure.
While these results are essential for revealing how plants react to eCO2, an on-going change in our planet, some points need further reflection. Reich and colleagues did not find any association between biomass production and light-saturated photosynthesis in C3 and C4 species, when considering punctual measurements on unit area bases. As biomass production is determined by the balance between photosynthesis and respiration (2), the analysis of both processes integrated at canopy level (total leaf area) and during the diurnal period would reveal why biomass was increased. In other words, the service done by each worker would be the same at eCO2 (as they found) but a bigger factory (higher leaf area per plant) has more workers and then a higher output. While enhanced C4 photosynthesis would be an unexpected result under eCO2 – but still possible as field-grown plants most likely face limiting conditions – there is no consensus about how respiration and its components (growth and maintenance) are affected by eCO2 (3).
Nitrogen dynamics in soil and plants is another point that should be further investigated. If more inorganic nitrogen is available and plants are able to take it up, one would expect higher leaf nitrogen concentration and higher photosynthesis. Reich et al. (1) report results obtained under no or relatively low level of nitrogen supply, do not support soil N mineralization with data on depletion of nitrogen and other resources in the soil and resource uptake by the plants, nor do they relate leaf nitrogen content to leaf photosynthesis. Their results suggest that nitrogen taken from soil was directed to other cellular components rather than to photosynthetic ones. When addressing nitrogen partitioning, we should consider that all studied species were perennials and then resources driven to root system are important to support re-growth after cutting or during the next growing season. Belowground changes were also found after long-term exposure of a coffee plantation to eCO2 (4). Elevated CO2 increased carbon concentration in the soil profile, suggesting that roots were acting as sinks for photo-assimilates.
Would we expect differential responses to eCO2 when annual and perennial species are compared? Following the results reported by Reich and colleagues, it seems that annual C3 species would be always favored by eCO2, while only perennial C4 species would take advantage under eCO2. At this moment, two points are clear about plant responses to increasing CO2 levels: we must consider long-term responses of perennial plants and also what happens below soil surface. As changes in soil quality would have a profound impact in agricultural systems and then affect food supply, a holistic view about the soil-plant-atmosphere system is needed for feeding an increasing population in a changing world. Unfortunately, this interesting paper from Reich et al. only shows a small part of that picture.

REFERENCES
1. P.B. Reich et al., Science, 360, 317 (2018).
2. X-G. Zhu et al., Annu. Rev. Plant Biol., 61, 235 (2010)
3. Z.Z. Xu et al., Front. Plant Sci., 6, 701 (2015)
4. M. Rakocevic et al., Ann. Bot., 121, 1065 (2018)

Unexpected results of a long-term experiment do not challenge the current paradigm of C3 vs C4 response to CO2
In their research (20 April 2018, p. 317-320), Reich et al. interpret data from a 20-year study. The initial response of C3 and C4 grasses to elevated CO2 (eCO2) is consistent with the theory of C3 and C4 photosynthesis physiology. A reversal of this pattern after ca 12 years of experimentation was interpreted to be contrary to theory. The authors claim the observed concomitant shift in soil fertility (nitrogen mineralization) responses between the C3 and C4 functional groups (their Fig. 2) could be the reason for this unexpected trend (see their Fig. 3). Similarly, in their Perspective (20 April 2018, p.263-264) on the Reich et al. paper, Hovenden and Newton state this experiment "overturned existing theory". There is no doubt that this paper is unique in terms of the unprecedented magnitude and the time span of experimentation under field conditions. Here we express our concern over the uncertainties associated with this long-term study and the over-interpretation of their results.
- Biomass of both species groups changed substantially over time (their Fig.1 and S1), with a 3-fold decrease for C3 grasses and an over 2-fold change for C4 grasses. However, the changes in soil N mineralization over time (their Fig.S5) does not agree with the trends in biomass. The down then up trends near the end for biomass for the C3 species and the more complicated pattern for the C4 species remain unaccounted for.
- To interpret plant responses to eCO2, one has to consider plant N status, in the context of understanding plant response to a long-term exposure to eCO2 (1). However, they did not measure plant N uptake, which would potentially explain the association shown in their Fig. 3. It is difficult to evaluate cause and effect when key photosynthesis controlling variables such as N concentration in the plant tissues and biotic stressors such as insects and plant diseases are not measured, particularly not knowing the level of soil N and trace minerals over time. Furthermore, the authors' data on leaf-level photosynthesis (their Fig. S4) does not indicate any switch between C3 and C4 groups in responding to eCO2.
- If N-mineralized "explained" the biomass response to CO2, then it would have been informative for the results of the two N-supply treatments to have been presented separately. The response to eCO2 must be different between the N treatments if N availability was behind the response.
- Regarding the differential changes to soil N mineralization possibly as a result of the changed soil microbial community, no information was given on why and how this happened, other than it might suggest a gradual decrease in critical soil nutrients throughout much of the 20-year experiment for the C3 group, but a different pattern for the C4 group.
- A possible reason for their results is the secondary effects of eCO2 on transpiration and canopy temperature. It is known, from both theory (2) and experimental evidence (3), that relative to C3 species, C4 species respond more to eCO2 in decreasing stomatal conductance and transpiration. This difference in their response can result in changes in soil moisture dynamics and soil temperature profile, and thus likely in soil microbial community and activity and soil N mineralisation (4), in a long term. Again, no data on soil moisture and temperature were presented.
The result shown in Fig. 3 is possibly just a simple correlation, which does not represent causality. If more background data and required explanatory parameters were measured or presented, one might analyze what would be needed to unravel the underlying bases for the strange patterns for the C3 and C4 groups over time. We don't believe the data presented by Reich et al. supports the claim that their results "challenge the current C3-C4 eCO2 paradigm". The paradigm for primary responses of C3 vs C4 plants is defined by the underlying physiology of C3 and C4 photosynthesis (i.e. the absence in C3 vs the presence in C4 leaves of the CO2-concentrating mechanism that suppresses photorespiration, 5,6), which remains intact. Note that secondary effects are often associated with plant response to eCO2 in a longer term (7). For example, response of leaf area index to eCO2, which is secondary, may overwhelm the effect of primary photosynthetic response to eCO2 (8,9). This does not mean a paradigm shift, but only stresses the necessity to improve models of simulating leaf area index in response to eCO2. In our opinion, if the authors' reported trends are general, real challenges are not to hypothesize a paradigm change, but to improve biogeochemical ecosystem models being able to mechanistically generate related secondary effects.

REFERENCES
1. X. Yin, Annals of Botany 112, 465-475 (2013).
2. G.D. Farquhar, T.D. Sharkey, Annual Review of Plant Physiology 33, 317-345 (1982).
3. E.A. Ainsworth, S.P. Long, New Phytologist 165, 351-372 (2005).
4. D.S. Jenkinson et al., Nature 351, 304-306 (1991).
5. M.D. Hatch, Biochimica et Biophysica Acta 895, 81-106 (1987).
6. X. Yin, P.C. Struik, New Phytologist 218, 986-998 (2018).
7. S.B. Gray et al., Nature Plants 2, 16132 | DOI: 10.1038/NPLANTS.2016.132 (2016).
8. F. Ewert, Annals of Botany 93, 619-627 (2004).
9. T. Hasegawa et al., Scientific Reports 7, 14858. DOI:10.1038/s41598-017-13582-y (2017).