Plant hydraulics and measurement of vulnerability to embolism formation: a guide for beginners

Being the most commonly used method so far, destructive hydraulic measurements are performed on stressed samples gravimetrically under a water head (Sperry et al., 1988) (Figure 4). A shorter stem/root subsample (usually 4–15 cm long) is cut under water from the mid part of the segment subjected to dehydration, air injection, or centrifugation. The bark is removed and the sample trimmed several times with a razor blade to remove micro-bubbles (Beikircher and Mayr, 2016), gradually release tensions (Wheeler et al., 2013; Venturas et al., 2015), and minimize eventual artifacts due to embolism recovery under soaking (Trifilò et al., 2014; Hochberg et al., 2016). The sample is then connected to a tube and perfused with filtrated and degassed polyionic perfusion solution (sometimes distilled water; Sperry et al., 1988). The height of the water column (15–80 cm) generates a low pressure (1.5–8 kPa), which forces the water through the sample without displacing emboli (Trifilò et al., 2014; Venturas et al., 2015). The hydraulic conductivity of the sample can be measured a) with flow meters (measuring the water inflow), which can be external or internal to the hydraulic apparatus (Figure 4), or b) by collecting the water outflow from the distal end of the sample in the pan of a precision balance or in pre-weighed vials containing a piece of sponge (over 30- or 60-s intervals) (Figure 5).

Figure 4.

Figure 5.

The hydraulic conductivity (K) is defined as the flow rate of water through the sample cross section divided by the pressure gradient (P, height of the water head). A high-pressure flush (0.01–0.2 MPa) is then applied to the sample for about 3–30 min (Sperry et al., 1988; Beikircher and Mayr, 2016; Hochberg et al., 2016) to remove embolism from xylem conduits. Then, the maximum hydraulic conductivity (K_max) is measured at low pressure as described above. PLC is calculated as follows: PLC=[1−(K/Kmax)]×100 \text{PLC}=\left[ 1-\left( \text{K}/{{\text{K}}_{\max }} \right) \right]\times 100

Sometimes, longer or repeated flushing is applied until the K measurements show no further increase in conductivity (Venturas et al., 2015; Losso et al., 2016; Ganthaler and Mayr, 2021).

This method can be also used to measure hydraulic conductivity/PLC of petioles of bench-dehydrated leaves or those sampled directly from trees (e.g., maple, grapevine; Wheeler et al., 2013; Hochberg et al., 2016). In this case, Ψ_xylem is measured on two leaves closest to the one sampled for hydraulic measurements.

The hydraulic method has been considered as a “gold standard” for many years, before being questioned assuming an artificial induction of embolisms while cutting the sample under tension (Ennajeh et al., 2011a; Wheeler et al., 2013). This has triggered dozens of experiments aimed at verifying eventual artifacts and comparing the results obtained with the hydraulic method to those obtained with other techniques (Trifilò et al., 2014; Torres-Ruiz et al., 2015; Venturas et al., 2019). Although they are sometimes unclear or controversial (Torres-Ruiz et al., 2015), most results supported the validity of hydraulic measurements and did not highlight significant overestimation of K/PLC due to the “tension-cutting artifact,” provided standard protocols for samples preparation are followed (Venturas et al., 2015; Nolf et al., 2017; Savi et al., 2017b).

Hydraulic measurements are valuable to assess native embolism in the field/greenhouse or estimate the consequences of a target Ψ_xylem value (induced by air dehydration, air injection, or centrifugation; Savi et al., 2015; Beikircher and Mayr, 2016; Losso et al., 2016). However, several days (and even weeks in case of air dehydration) and a large number of samples (each of them used for one experimental point, see Figure 3a) are needed to generate a complete VC with a sufficiently wide range of Ψ_xylem (Cochard et al., 2013). In fact, the hydraulic measurements following embolism induction are destructive and prevent repeated measurements on the same sample which is discarded after the quantification of each K/PLC.

Thanks to slight adaptations, flow measurements through the branch/root segment (without cutting a central subsample) can be performed continuously while the segment is inserted a) in the pressure sleeve (Cochard et al., 1992) (Figure 6) or b) in the rotor of the centrifuge (Cochard, 2002) (Figure 7). This allows application of progressively lower Ψ_xylem target values on one single sample, which are always followed by relative measurements of K/PLC.

Figure 6.

Figure 7.

Specifically, in the air injection method with continuous hydraulic measurements, the basal end of the segment is connected to the water head with a tube large enough to favor the escape of air bubbles during pressurization (Cochard et al., 1992) (Figure 6). First, to calculate PLC for all points, maximum conductivity (Kmax) is initially measured on fully hydrated samples without pressure application (Ψ_xylem = 0, usually by outflow collection shown in Figure 5). Then, a first seeding injection is performed usually at 0.5 MPa. The pressure is released, the sample allowed to equilibrate in water for some minutes, and the K value remeasured. The pressure applied in subsequent air-injection cycles is gradually increased by steps of 0.5–1.0 MPa and the K measurements are performed at each step until PLC >90% is reached (species-specific value; Cochard et al., 1992; Savi et al., 2019). Similarly, according to the flow centrifuge technique (frequently called “cavitron”), the loss of conductivity can be determined many times on the same sample after progressive increase of the rotational speed of the centrifuge (thus, lowering the Ψ_xylem; Cochard, 2002; Paligi et al., 2023) (Figure 7). The upstream cuvette (in which the basal end of the sample is inserted) is periodically filled with water, thanks to a tube connected to a water reservoir external to the centrifuge. During spinning, beside lowering Ψ_xylem in the sample, the centrifugal force induces a positive hydrostatic pressure difference between the sample ends inserted in the cuvettes and, hence, water flow through the sample. This flow is determined optically during centrifugation, thanks to a camera focused on the water meniscus inside the upstream cuvette (Figure 7). In fact, with rotation speed and flow through the sample, the meniscus moves and the cuvette empties. Kmax is initially measured on fully hydrated samples by applying a very low rotational speed simulating Ψ_xylem or approximately −0.1 MPa. By increasing the rotation, embolisms build up, decreasing the flow through the sample and, hence, the speed of the meniscus (Cochard, 2002).

The air-injection and centrifugation methods with continuous flow measurements give real-time evaluation of water transport capacity loss during embolism formation in an excised sample. A complete VC is obtained using one single sample, usually in 45 min–2 h (Cochard et al., 2013). However, while they are generally accepted for conifers and short-vesseled angiosperms, both methods may be problematic for obtaining accurate VCs in long-vesseled angiosperm species (Torres-Ruiz et al., 2014; Wang et al., 2014; Savi et al., 2019). In particular, overestimation of vulnerability thresholds (i.e., P12, P50, and P88 with not enough negative values) has been reported when using xylem segments with a high proportion of vessels opened at both ends. This phenomenon is known as “open-vessels artifact” and, as its consequence, both air injection and flow centrifuge methods can sometimes give VCs with exponential shape, called “r-shaped” (Figure 3b) (Ennajeh et al., 2011b; Torres-Ruiz et al., 2014; Venturas et al., 2019). The r-shaped curves show high loss of hydraulic conductivity already at very low tension, leading to an overestimation of embolism vulnerability (higher P50 value; Figure 3b) (Wang et al., 2014). Such curves are difficult to interpret and often nonrealistic, compared to native field-measured embolism (Wang et al., 2014). Thus, the r-shaped curves could generally arise as a result of a) the plant anatomical characteristics (high density of long vessels) or b) the method used. To prevent this artifact, it is suggested to work with samples that are longer than 1.5-fold the maximum vessel length (Lmax) (Ennajeh et al., 2011b).

L_max can be easily estimated using air injection according to Wang et al. (2014). Briefly, the apical end of the branch/root is debarked and inserted in a silicone tube connected to a pump (or a Scholander chamber). The air is forced at low pressure (about 1.5–5 kPa) from the apical end of the sample, while its basal end is immersed in water. The apex is then progressively cut at about 2-cm intervals until the first air bubbles are observed escaping the cut section by using a magnifying lens. Although there have been several improvements to the air injection and centrifugation methods with continuous flow measurements, their reliability is still questionable and technical pitfalls persist (Venturas et al., 2019). Unfortunately, using samples that are 1.5-fold longer than the species maximum vessel length is often difficult, as for example in case of species with vessels longer than 40 cm (e.g., Quercus sp, Vitis sp; Ennajeh et al., 2011b; Savi et al., 2019). Furthermore, working with samples longer than Lmax is impossible when the measurements are limited by the size of the centrifuge rotor (usually 15 or 30 cm; Torres-Ruiz et al., 2014; Nolf et al., 2015; Brodribb et al., 2017; Venturas et al., 2019). Increasing the rotor diameter depends on the size of the centrifuge chamber (usually < 50 cm). Some huge set-ups exist, but are expensive and potentially very dangerous due to high speed rotation of big, custom-made metallic rotors.

The cavitation of water (change from liquid to vapor phase) in xylem conduits results in sudden pressure change (from sub-atmospheric to atmospheric) leading to vibration of rigid conduits’ walls. The latter produce a sound, that is, an acoustic emission event ranging from audible to ultrasonic (AE), which can be detected in a noninvasive manner (Milburn and Johnson, 1966). The rate of production of AEs indicates to what extent the decreasing Ψ_xylem is affecting the water transport capacity (Rosner et al., 2006; Nolf et al., 2015). Hence, rather than measuring the hydraulic consequences of embolism, the acoustic emission method generates VCs with monitoring of acoustic signals produced by a plant or an excised organ under air dehydration (laboratory, greenhouse, field).

In case of root and stem VCs, a small portion of the organ is debarked to remove the phloem and cambium, and an acoustic sensor is clamped directly to the xylem. A thin layer of petroleum jelly is smeared between the sensor head and the xylem to improve the acoustic contact and prevent local dehydration. In leaves, sensors are usually attached directly to the midrib. The recorded signals are amplified and logged automatically over time. The measurements of a complete VC may take from some hours to several days. Finally, the cumulative AEs (exceeding a defined detection threshold) or different AE rates (Rosner et al., 2006; Nolf et al., 2015) are plotted against Ψ_xylem, which is regularly measured during the experiment (see above).

The main advantages of the acoustic emission method are a) the nondestructive measurements and b) the in vivo continuous monitoring of embolism formation, both of which allow to measure a complete VC on a single sample. However, even if modern acoustic systems allow advanced analyses of signals, the method is considered more qualitative than quantitative. In fact, it gives only indirect evidence of xylem embolism, since there is no linear relation between AEs and hydraulic conductivity loss. Moreover, processes other than cavitation may be the source of large numbers of AEs within the xylem tissue (especially at the late stages of dehydration), for example, fibers’ cavitation, cell shrinkage, tissue ruptures, etc. (Rosner et al., 2006; Nolf et al., 2015).

Early attempts to directly visualize the functional status of xylem conduits under well-watered or drought-stressed conditions were based on cryo-scanning electron microscope (cryo-SEM) observations (Canny, 1997). The technique is based on snap freezing of samples (leaves, stems, roots) in liquid nitrogen, followed by sectioning and observation under a cryo-SEM to assess the presence of water or gas in the conduits.

In the last decade, modern optical measurements of embolism in the xylem network using light microscope, X-ray computed tomography (micro-CT; Brodersen et al., 2013) (Figure 8), magnetic resonance imaging (MRI; Holbrook et al., 2001; Zwieniecki et al., 2013), and high-definition scanners or cameras (Brodribb et al., 2017) have gained momentum for the evaluation of plants’ drought sensitivity. Briefly, all these methods allow to observe water in plants, revealing the location and formation of embolisms inside conduits. The loss of hydraulic conductivity is thus evaluated in vivo in intact plants or excised segments subjected to drought by visually assessing the functional status of the xylem conduits. Parallel measurements of Ψ_xylem, usually performed with the Scholander pressure bombe, are always necessary. Imaging can thus be applied to leaves, stems, or roots of both woody and herbaceous plants and relies on analyses of pictures which are gathered differently (see below; Brodribb et al., 2017; Scoffoni et al., 2017; Losso et al., 2019).

Figure 8.

MRI was the first method used to noninvasively monitor the status of xylem conduits in intact plants (Holbrook et al., 2001; Zwieniecki et al., 2013). Pixel brightness in the images (ranging from black to white) is related to the increasing concentration of water and is used for analyses of xylem water content (Zwieniecki et al., 2013; Meixner et al., 2020). Unfortunately, the imaging is rather challenging, since MRI scanners configured to accept intact plants are rare and expensive (the same physical limitation is valid for the micro-CT method, see below). The resulting pictures have typically low spatial resolution, and single vessels cannot be easily resolved on the image (Zwieniecki et al., 2013; Meixner et al., 2020). Thus, the method is mostly used to study species with wide conduit diameter.

With micro-CT, leaves, stems, or roots of both woody and herbaceous plants are pre-processed and fixed to the beamline sample holder (Figure 8). Bidimensional (2D) or three-dimensional (3D) images of the xylem network are reconstructed with resolution <1 μm (Savi et al., 2017b; Théroux-Rancourt et al., 2023). Gas-filled and water-filled conduits can be clearly distinguished, since the former appear electron transparent (gray) and the latter appear electron dense (black; Figure 9) (Torres-Ruiz et al., 2015; Losso et al., 2016; Nolf et al., 2017). Generally, in 2D images, the average diameter of each xylem conduit (derived from its area and assuming circular shape) is used to calculate the theoretical hydraulic conductivity with Hagen–Poiseuille equation (Tyree and Zimmerman, 2002; Savi et al., 2017b). The total xylem conductivity is assessed as the sum of all gas-filled and water-filled conduit conductivities. Finally, the theoretical PLC is calculated as the ratio between gas-filled conduits and total conductivity.

Figure 9.

Due to its supposed noninvasive nature and direct in vivo capacity to quantify xylem embolism, micro-CT has been suggested to represent a reference outperforming the hydraulic measurements (Torres-Ruiz et al., 2015; Venturas et al., 2017). However, this method is usually limited by the size and type of samples that can fit in the measurement chamber and is not currently usable in the field. Moreover, the scanned plant segment is rather small and the image analyses do not consider the radial flow movement and the resistance of interconduit pit membranes (Venturas et al., 2017). Caution is also needed in interpreting results, since vital samples are exposed to ionizing radiation and/or successive scans, which may cause harmful effects on the cells (Savi et al., 2017b; Secchi et al., 2021). Moreover, considering the limitations in terms of access to synchrotron facilities, which are few and expensive, it is improbable that this technique will become a routine for processing a large number of samples, at least not in the near future.

As an alternative to these hardware-expensive techniques, Brodribb et al. (2017) recently developed a new optical method for monitoring of embolizing bubbles’ propagation in the xylem during imposed water stress. The method is cost effective, since standard flatbed scanners, cameras, or light microscopes are used to image the vascular system (usually of leaves or stems). The image at full hydration is considered as the reference point (K_max /PLC = 0), and rapid changes in visual light transmission caused by embolism formation can be detected under progressive dehydration. PLC is, hence, assessed with image analyses as a proportion of observed events (optically detected cavitation; Brodribb et al., 2017; Venturas et al., 2019). The new optical method has several important advantages: a) it allows visualizing the spatial and temporal spread of emboli through leaf vein networks and b) it is a simple and portable tool which can be used to assess xylem vulnerability of both leaves and stems. However, it does not fully consider the three-dimensionality of the hydraulic pathway, as well as the radial water flow through interconduit pits.