By regulating carbon uptake and water loss by plants, stomata are not only responsible for productivity but also survival during drought. The timing of stomatal closure is crucial for preventing excessive water loss during drought, yet has high ecological variability between species. An aspect of stomatal response that remains disputed is the mechanism driving the trigger for stomatal closure during drought. We investigated this in a highly embolism resistant tree species Umbellularia californica. We tracked leaf endogenous abscisic acid (ABA) levels and determined the predawn and midday leaf water potential and gravimetric soil water content (gSWC) thresholds for stomatal closure and transpiration decline during a progressive drought. We found that while U. californica plants have a peaking type ABA dynamic where ABA levels rise early in drought then decline under prolonged drought conditions. The early increase in ABA levels corresponded to the closing of stomata and reduced transpiration. Furthermore, we found that the gSWC at which transpiration declines occurs before any large decreases in soil and predawn plant water status and could best be explained by declines in midday water potentials along with increased ABA levels. Our results indicate that ABA mediated stomatal regulation is an integral mechanism for delaying whole plant dehydration and can occur with minimal changes in bulk soil and plant water status.

The phytohormone abscisic acid (ABA) is synthesized by plants during drought to close stomata and regulate desiccation tolerance pathways. In conifers and a few angiosperms with embolism resistant xylem a peaking-type (p-type) response in ABA levels has been observed, in which ABA levels increase early in drought then decrease as drought progresses, declining to pre-stressed levels. The mechanism behind this dynamic remains unknown. Here we sought to characterize the mechanism driving p-type ABA dynamics in the conifer Callitris rhomboidea and the highly drought resistant angiosperm Umbellularia californica. We measured leaf water potentials (Ψ l), stomatal conductance, ABA, conjugates and phaseic acid (PA) levels in potted plants during a prolonged but non-fatal drought. Both species displayed a p-type ABA dynamic during prolonged drought. Measuring ABA levels in bench dried, rehydrated branches collected before and after the peak in ABA levels revelated that in both species ABA biosynthesis is deactivated in leaves that have been dehydrated beyond leaf turgor loss point. Considerable conversion of ABA to conjugates was found to occur during drought, but not catabolism to PA. The mechanism driving the decline in ABA levels in p-type species appears to be conserved across seed plants and is mediated by sustained conjugation of ABA and the deactivation of ABA biosynthesis as Ψ l becomes more negative than turgor loss.

Xylem embolism resistance varies across species influencing drought tolerance, yet little is known about the determinants of the embolism resistance of an individual conduit. Here we conducted an experiment using the optical vulnerability method to test whether individual conduits have a specific water potential threshold for embolism formation and whether pre-existing embolism in neighbouring conduits alters this threshold. Observations were made on a diverse sample of angiosperm and conifer species through a cycle of dehydration, rehydration and subsequent dehydration to death. Upon rehydration after the formation of embolism, no refilling was observed. When little pre-existing embolism was present, xylem conduits had a conserved, individual, embolism resistance threshold that varied across the population of conduits. The consequence of a variable conduit-specific embolism threshold is that a small degree of pre-existing embolism in the xylem results in an apparently more resistant xylem in a subsequent dehydration, particularly in angiosperms with vessels. While our results suggest that pit membranes separating xylem conduits are critical for maintaining a conserved individual embolism threshold for given conduit when little pre-exisiting embolism is present, as the percentage of embolized conduits increases, gas movement, local pressure differences, and connectivity between conduits increasingly contribute to embolism spread.

Embolism spreading in dehydrating angiosperm xylem is driven by gas movement between embolised and sap-filled conduits. Here, we examine how proximity to pre-existing embolism and hydraulic segmentation affect embolism propagation. Based on the optical method, we compared xylem embolism resistance between detached leaves and leaves attached to branches, and between intact leaves and leaves with cut minor veins for six species. Moreover, we directly compared the optical and pneumatic method on detached leaves. Embolism resistance of detached leaves was significantly lower than leaves attached to stems, except for two species with all vessels ending in their petioles. Cutting of minor veins showed limited embolism spreading in minor veins near the cuts prior to major veins. Moreover, there was strong agreement in embolism resistance between the optical and pneumatic method, with minor differences occurring during early stages of embolism formation. We conclude that embolism resistance may represent a relative trait, depending on the proximity and connectivity to pre-existing embolism as a gas source. Since embolism formation may not rely on a certain pressure difference threshold between functional and embolised conduits, we suggest that embolism is facilitated by pressure-driven gas diffusion, while hydraulic segmentation can prevent embolism propagation by reducing gas diffusion.

Embolism spreading in dehydrating angiosperm xylem is driven by gas movement between embolised and sap-filled conduits. Here, we examine how proximity to pre-existing embolism and hydraulic segmentation affect embolism propagation. Based on the optical method, we compared xylem embolism resistance between detached leaves and leaves attached to branches, and between intact leaves and leaves with minor veins cut (n = 6 species). Moreover, we directly compared the optical and pneumatic method on detached leaves. Embolism resistance of detached leaves was significantly lower than leaves attached to stems, except for two species with all vessels ending in their petioles. Cutting of minor veins showed embolism spreading in narrow vessels near the cuts prior to wide vessels in major veins. Moreover, embolism spreading between open and intact vessels occurred at largely similar xylem water potentials than embolism spreading between intact vessels, resulting in strong similarity between the optical and pneumatic method. We conclude that embolism spreading may depend on a direct connection to pre-existing embolism as gas source, is not exclusively pressure-driven, and indirectly related to conduit size. Hydraulic segmentation, however, can minimise embolism spreading due to confined and/or poorly interconnected conduits, which may increase hydraulic safety by slowing down gas diffusion.