Bi-directional COS exchange in bryophytes challenges its use as a tracer for gross primary productivity

In this issue of New Phytologist, Gimeno et al. (pp. x-y) report results which challenge the main assumption underlying the concept of inferring plant gross primary productivity (GPP) from the uptake of carbonyl sulphide (abbreviated alternately as COS or OCS) by plants (Campbell et al., 2008; Wohlfahrt et al., 2012). Using two astomatous bryophyte species, Gimeno et al. quantified the concurrent carbon dioxide (CO₂) and COS gas exchange at different levels of hydration, light intensity, temperature and ambient COS mole fraction. In contrast to their hypotheses, they found (i) the largest COS uptake to occur during darkness and cooler conditions and (ii) lower COS uptake and, under low hydration levels, even release of COS during warmer daylight conditions.

Plant COS uptake during darkness is consistent with the view that the enzyme catalysing the irreversible hydration of COS, the carbonic anhydrase (CA), is light-independent (Protoschill-Krebs et al., 1996). Previous studies on COS uptake by vascular plants during darkness at leaf (Stimler et al., 2011) and ecosystem scale (Wehr et al., 2017), however, reported uptake rates to be quite small compared to daytime conditions. This is due to near-complete stomatal closure in darkness, which imposes a major resistance to the uptake. Using astomatous model organisms, Gimeno et al. were able to elegantly demonstrate that the COS uptake may be significant during darkness, if not curbed by diffusive limitations. While the COS uptake during darkness is not new, the comparably smaller uptake or even release of COS during daylight conditions in the presence of active photosynthesis contradicts our present understanding of the coupled CO₂ and COS uptake by plants (Seibt et al., 2010). Using a combination of COS, temperature and light response curves, Gimeno et al. were able to show that (i) concurrent production and consumption of COS must be at work inside the investigated bryophytes, causing a non-zero COS compensation point (i.e. the ambient COS concentration at which the COS exchange equals zero) and that (ii) the invoked COS production term was highly temperature-dependent. The lower COS uptake during light compared to dark conditions, could thus be explained by the associated higher temperatures when the light was on, which increased production of COS more than consumption and thus lowered the net uptake and eventually caused it to switch sign. Even though the authors were not able to unambiguously identify the production process, they hypothesised that COS may have been produced from sulphur-containing amino acids, which are expected to have degraded during desiccation as diagnosed from the observed reduction in protein and non-structural carbon contents.

The existence of a COS compensation point is at odds with most of the previous literature, which suggests that the leaf COS exchange is unidirectional, with a near-zero compensation point, due to CA’s strong affinity to COS (Protoschill-Krebs et al., 1996). Gimeno et al. determined a compensation point of 345 pmol mol^-1 in the light, which is lower than the global average COS mole fraction of ca. 500 pmol mol^-1, but may be reached close to and within plant canopies during active leaf gas exchange (Blonquist et al., 2011) and thus cause the COS exchange to cease or switch sign. The assumption of a unidirectional COS flux underlies the rationale of using COS as a tracer for GPP, which at the ecosystem scale is impossible to directly quantify. With a near-zero compensation point, the COS exchange must be directed into the leaf, driven by the ambient COS mole fraction and the series of resistances from the ambient air to the site of hydrolysis by CA. As the exchange of CO₂, even though it exhibits a non-zero compensation point, shares most of the resistances with COS (up to the hydrolysis by CA), measurements of the leaf COS uptake have been suggested to allow inferring GPP, which otherwise is confounded by concurrent respiration processes.

To this end a metric coined the leaf relative uptake rate (LRU) has been instrumental. The LRU represents the ratio of the leaf COS to CO₂ uptake rates normalised by the respective ambient concentrations and collectively literature data suggest it to converge to a value of around 1.7 during high radiation and in the absence of stress (Berkelhammer et al., 2014). With a known LRU and data on the COS flux and the COS and CO₂ mole fractions at hand, GPP can be inferred (Asaf et al., 2013). Previous criticism of the constant LRU concept focussed mainly on the fact that it does not account for changes in the internal to ambient CO₂ mole fraction (C_i/C_a) ratio (Wohlfahrt et al., 2012), which for example has been shown to cause the LRU (Stimler et al., 2011) or its ecosystem-scale analogue (ERU; Wehr et al., 2017) to deviate from its background value at low light.

If, on top of this, the COS exchange was, as Gimeno et al. demonstrate for bryophytes, dependent on a variable compensation point, the LRU concept would be rendered questionable and the usefulness of COS as a proxy for GPP (and stomatal conductance; Wehr et al., 2017) at stake. The key question in this context is whether the leaf internal production of COS (and thus a non-zero compensation point) is specific to bryophytes or whether this is a process that occurs in vascular plants as well and if so how important it is? Bryophytes are characterised by a specialised physiology and well adapted to rapid and recurring drying/rehydration cycles, which Gimeno et al. hypothesise to underlie the invoked COS production term. Stress-related protein turnover, which may cause COS to escape from sulphur-containing compounds, is however common in vascular plants as well, begging the question why available vascular plant LRUs converge to such a narrow range (Berkelhammer et al., 2014) and hardly any reports of leaf COS emission exist (but see Maseyk et al., 2014). Partially, the answer may lie in the fact that up to date leaf COS gas exchange data have been collected with an emphasis on understanding the basic principles and exploring the variability between species (e.g. C₃ vs. C₄ photosynthetic pathways) and thus mostly investigated experimental plants under near-optimal growth conditions. Possibly, there also might be a publication bias, results not conforming to the established LRU concept being less likely to be published. In other words – to date, we simply may lack the data to answer this question and thus targeted experiments exposing vascular plants to specific stressors are urgently needed. At present, it is thus also unclear whether stomatal closure in stressed vascular plants curtails any COS emissions, causing internally produced COS to be hydrolysed by CA. As suggested by Gimeno et al., experiments with mutants that lack stomatal control would be a promising pathway to pursue to this end. Testing the stress-response of COS exchange will also be beneficial for narrowing down the range of conditions under which the LRU may be treated as a constant.

Even though the findings of Gimeno et al. challenge the major assumption underlying the LRU, personally I am nevertheless confident that COS will develop into a sensible and useful constraint of ecosystem scale GPP (and stomatal conductance). Too convincing is the convergence of LRU across different vascular plant species (under high light and unstressed conditions) (Berkelhammer et al., 2014), the clear uptake of COS observed by the (few) available ecosystem scale flux measurements (e.g. Asaf et al., 2013; Commane et al., 2015), and the strong seasonal covariance between atmospheric COS and CO₂ across latitudes (Montzka et al., 2007). However, it also has become clear during the past few years that COS will not be the ‘silver bullet’, as a scientist recently put it in a COS meeting. On theoretical grounds, the LRU cannot be expected to be constant under all conditions (Wohlfahrt et al., 2012) and non-leaf, in particular soil (Whelan et al., 2015; Kitz et al., 2017) sinks and sources of COS complicate the attribution of the leaf sink, an issue to which the findings of Gimeno et al. contribute. The ‘new window into the carbon cycle’ (Berry et al., 2013), may thus not be transparent everywhere and all the time, but in a multiple-constraints approach in concert with flux partitioning based on CO₂ (Lasslop et al., 2010), the stable isotopes of CO₂ (Wehr et al., 2016), and proximal sensing-based light-use efficiency approaches (Migliavacca et al., 2017), I still expect COS to make an important future contribution to disentangling the ecosystem-scale CO₂ flux components.