Skip to main content
ARS Home » Pacific West Area » Davis, California » Crops Pathology and Genetics Research » Research » Publications at this Location » Publication #345065
Title: In-vivo quantification of plant starch reserves at micrometer resolution using X-ray microCT imaging and machine learning

Author
item EARLES, J - Yale University
item KNIPFER, THORSTEN - University Of California
item TIXIER, AUDE - University Of California
item OROZCO, JESSICA - University Of California
item REYES, CLARISSA - University Of California
item ZWIENIECKI, MACIEJ - University Of California
item BRODERSEN, CRAIG - Yale University
item McElrone, Andrew

Submitted to: New Phytologist
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/22/2018
Publication Date: 4/16/2018
Citation: Earles, J.M., Knipfer, T., Tixier, A., Orozco, J., Reyes, C., Zwieniecki, M.A., Brodersen, C., McElrone, A.J. 2018. In-vivo quantification of plant starch reserves at micrometer resolution using X-ray microCT imaging and machine learning. New Phytologist. 218(3):1260-1269. https://doi.org/10.1111/nph.15068.
DOI: https://doi.org/10.1111/nph.15068

Interpretive Summary: Starch is the primary energy storage molecule used by most terrestrial plants to fuel respiration and growth during periods of limited to no photosynthesis, and its depletion can drive plant mortality. Destructive techniques at coarse spatial scales exist to quantify starch, but these techniques face methodological challenges that can lead to uncertainty about the lability of tissue-specific starch pools and their role in plant survival. Here, we demonstrate how X-ray microcomputed tomography (microCT) and a machine learning algorithm can be coupled to quantify plant starch content in vivo, repeatedly and nondestructively over time in grapevine stems (Vitis spp.). Starch content estimated for xylem axial and ray parenchyma cells from microCT images was correlated strongly with enzymatically measured bulk-tissue starch concentration on the same stems. After validating our machine learning algorithm, we then characterized the spatial distribution of starch concentration in living stems at micrometer resolution, and identified starch depletion in live plants under experimental conditions designed to halt photosynthesis and starch production, initiating the drawdown of stored starch pools. Using X-ray microCT technology for in vivo starch monitoring should enable novel research directed at resolving the spatial and temporal patterns of starch accumulation and depletion in woody plant species.

Technical Abstract: Starch is the central energy storage molecule used by plants to fuel respiration and growth during periods of limited or no photosynthesis. Recently, the relative starch concentration of plants entering drought was linked to mortality probability, as the relative pools of starch and other non-structural carbohydrates (NSCs) are necessary for maintaining cellular functions, producing chemical defense compounds to prevent biotic attack, and growing new foliage. However, only destructive techniques at coarse spatial scales exist to quantify starch and NSCs, which face methodological challenges that can lead to uncertainty about the lability of tissue-specific starch pools and their role in plant survival. Here, we use X-ray microcomputed tomography and a machine learning algorithm for in vivo quantification of plant starch content. Starch estimates correlated strongly (R2 = 0.95; p << 0.01) with enzymatically-measured bulk-tissue starch concentration on the same stems. We then spatially mapped regions of high and low starch concentration in stems at micrometer resolution for the first time. Further, we detected regions of starch depletion within living stem tissue in plants that were withheld from light for thirty days to deplete these reserves showing unidirectional use of the starch reserves along radial coordinates. Such high spatial-resolution in vivo starch monitoring should enable novel research directions across the plant sciences.