Here, we have shown that our labelling chambers are air-tight, fast to attach and to remove, and therefore ideal for labelling experiments with single Arabidopsis leaves. During the experimental procedure only minimal amounts of labelled CO2 were released from the system, probably while attaching and removing the chamber, causing negligible incorporation of 14C by unlabelled leaves.
Feeding labelled carbon to individual leaves of intact plants has the advantage that carbon transport and carbon partitioning (in the specific source tissue) can be monitored simultaneously. It allowed us to determine, qualitatively as well as quantitatively, the amount of assimilated carbon which is transported form source leaves to sink leaves and other sink tissue such as the root in parallel. A defined pattern of carbon allocation from a mature leaf could be seen. A similar pattern, which is determined by the vascular system, was observed for other plants like tobacco[18, 19]. Following the vascular connections, carbon is transported from a mature leaf upwards to the leaves n + 3, n + 5, n + 8 and n + 10 in tobacco plants. For Arabidopsis, two vascular models have been described which either suggest connections from one leaf to the leaves n + 2 and n + 3 (from leaf 4 onwards;) or to the leaves n + 5 and n + 8. In the majority of autoradiograms performed, we could see carbon import into leaves n + 5, n + 8 and, if already emerged, in leaf n + 13 (which is linked to the fed leaf via leaf n + 5 and n + 8). This supports the vascular model proposed by Kang et al.. However, we occasionally observed carbon import into the leaves n + 2 and n + 3 (Figures 3,4 and5) which could be explained by stochastic irregularities during the formation of vascular connections.
A simultaneous transport from mature leaves towards developing leaves and towards the root was seen in the autoradiographs as well as in the quantitative measurements. It suggests that, in certain regions of the vasculature, bidirectional phloem transport occurs. Bidirectional transport within the phloem is long-known, but it is still unclear how it occurs[14, 18, 44]. According to our current understanding of the phloem, simultaneous transport within one sieve tube cannot occur. Therefore, a more complex phloem structure with separate cell files facilitating acropetal and basipetal transport seems likely[47, 48]. The labelling system presented in this work could help gain new insights into carbon transport and a better understanding of phloem structure/function.
The quantitative export data revealed that about 25% of the assimilated carbon is exported from the labelled source leaves within the same photoperiod, most of it already within the first hour after fixation. Higher export rates were previously described for sugar beet, where 24% of carbon was exported to sink leaves after a 1-h chase period. In Curcurbita pepo, 50% of label from a single leaf was found to be exported after a 2-h chase period, with 16% allocated to the roots[15, 49]. Export processes in the dark have not been measured in this study, but are likely to be significant. Experiments with whole Arabidopsis rosettes by Zeeman and ap Rees, using similar conditions to those described here (i.e. labelling in the middle of the photoperiod), revealed 6.9% export to the root by the end of the day, comparable to our results. However, by the end of the subsequent night, export from the rosette to the root had increased to 22.7%. This night-time export is driven by carbon released from starch during the night, significantly increasing the amount of carbon allocated to sinks during the whole day-night cycle. The time course analysis of the labelled leaf also showed that, as well as storage compounds like starch, carbon flows via intermediary metabolite pools towards the synthesis of end products like cell wall and proteins. These data suggest that leaf 8, despite being one of the most mature leaves on the rosette, is still growing, so not all assimilated carbon will be destined for export. This is consistent with sensitive growth analysis data which reveal a continuation of Arabidopsis leaf expansion for up to 3 weeks.
While the flow through acidic and basic fractions was very rapid after the labelling pulse, carbon flow through neutral compounds (e.g. sucrose, glucose and fructose) was slower and peaked well after the end of the labelling - between 15 to 30 min into the chase. Carbon incorporation into starch was fast, with the majority of carbon incorporation during the pulse period and the following 15-min chase. These data are broadly in line with data recently obtained in a detailed study using 13C labelling, where the degree of label saturation in the different metabolic intermediates followed quite different kinetics during one hour of labelling (presumably depending on both fluxes and intermediate pool sizes). In that study, sucrose labelling was relatively slow, with glucose and fructose labelling slower still. However, ADPglucose - the precursor for starch biosynthesis - was rapidly labelled. These data are also consistent with the dynamics we observed in label export from the leaf, which approached completion an hour after the pulse. It should be emphasized that, compared with the fairly crude separation of compound classes presented here, 13C labelling coupled with mass spectrometry gives excellent resolution in terms of the degree of labelling of intermediates over time. However, even with today’s technology it is still a major undertaking to measure metabolite fluxes with 13C time-course labelling.