phloem as capacitor: radial transfer of water into xylem of tree

Phloem as Capacitor: Radial Transfer of Water intoXylem of Tree Stems Occurs via Symplastic Transport inRay Parenchyma[OPEN]

Sebastian Pfautsch*, Justine Renard, Mark G. Tjoelker, and Anya Salih

Hawkesbury Institute for the Environment, University of Western Sydney, Penrith, New South Wales 2751,Australia (S.P., M.G.T.); École Normale Supérieure, 75005 Paris, France (J.R.); and Confocal Bio-ImagingFacility, Hawkesbury Campus, Deputy Vice-Chancellor (Research and Development) Division, University ofWestern Sydney, Penrith South, New South Wales 1797, Australia (A.S.)

ORCID ID: 0000-0002-4390-4195 (S.P.).

The transfer of water from phloem into xylem is thought to mitigate increasing hydraulic tension in the vascular system of treesduring the diel cycle of transpiration. Although a putative plant function, to date there is no direct evidence of such watertransfer or the contributing pathways. Here, we trace the radial flow of water from the phloem into the xylem and investigate itsdiel variation. Introducing a fluorescent dye (0.1% [w/w] fluorescein) into the phloem water of the tree species Eucalyptus salignaallowed localization of the dye in phloem and xylem tissues using confocal laser scanning microscopy. Our results show that themajority of water transferred between the two tissues is facilitated via the symplast of horizontal ray parenchyma cells. Themethod also permitted assessment of the radial transfer of water during the diel cycle, where changes in water potentialgradients between phloem and xylem determine the extent and direction of radial transfer. When injected during themorning, when xylem water potential rapidly declined, fluorescein was translocated, on average, farther into mature xylem(447 6 188 mm) compared with nighttime, when xylem water potential was close to zero (155 6 42 mm). These findings provideempirical evidence to support theoretical predictions of the role of phloem-xylem water transfer in the hydraulic functioning ofplants. This method enables investigation of the role of phloem tissue as a dynamic capacitor for water storage and transfer andits contribution toward the maintenance of the functional integrity of xylem in trees.

Physiological and hydraulic functioning of the twolong-distance transport systems in trees, xylem andphloem, have intrigued plant researchers for more thana century. Since the pioneering work of Dixon and Joly(1895; cohesion-tension theory for xylem) and Münch(1930; pressure flow hypothesis for phloem), the ma-jority of studies have investigated these systems inde-pendently of each other. Although the work of Stoutand Hoagland (1939) as well as Biddulph and Markle(1944) laid the foundation for the physiological nexusbetween xylem and phloem, it is only recently that wehave begun to understand the importance of the hy-draulic nexus (Hölttä et al., 2006, 2009; Sevanto et al.,2011, 2014). Processes related to both nexus occur inparallel, and here the term physiological nexus coversall metabolite exchange, including the bidirectional flowof amino acids, minerals, and carbohydrates (Wardlaw,1974; Ferrier et al., 1975; Pfautsch et al., 2009, 2015; DeSchepper et al., 2013; for review, see van Bel, 1990,2003). The term hydraulic nexus refers to the function of

phloem as a capacitor, where water stored in phloemmoves into xylem vessels to maintain the integrity ofthe transpiration stream (Zweifel et al., 2000, and refs.therein). Throughout this article, we use the term phloemcollectively for cells that make up the transport phloemof woody plants (including sieve element/companioncell complexes, parenchyma cells, etc.), as opposed tocollection and release phloem tissue, which differ instructure and function. Transport phloem is character-ized by the retention of “high hydrostatic pressure byretrieval of leaked osmotica accompanied by water flux”(Patrick, 2013).

According to the cohesion-tension theory, water inxylem vessels is constantly under tension and moves ina metastable state from roots to leaves along a hydro-static pressure gradient. Depending on both the avail-ability of soil moisture and the vapor pressure deficit ofthe atmosphere, this tension can exceed the cohesiveforces that bind water molecules, resulting in the for-mation of a gas void that, after expanding, can lead torupture of the water column inside individual vessels(termed cavitation; Zimmermann, 1983). Once cavitated,vessels become dysfunctional, and the transport of waterand nutrients to leaves declines. However, water storedin woody tissues of trees can be mobilized to alleviatethe risk of cavitation, and recent theory suggests thatboth water and carbohydrates from phloem may aidin the reversal of vessel embolism (i.e. air intrusion),

* Address correspondence to [email protected] author responsible for distribution of materials integral to the

findings presented in this article in accordance with the policy de-scribed in the Instructions for Authors (www.plantphysiol.org) is:Sebastian Pfautsch ([email protected]).

[OPEN] Articles can be viewed without a subscription.www.plantphysiol.org/cgi/doi/10.1104/pp.114.254581

Plant Physiology�, March 2015, Vol. 167, pp. 963–971, www.plantphysiol.org � 2014 American Society of Plant Biologists. All Rights Reserved. 963 www.plantphysiol.orgon April 13, 2018 - Published by Downloaded from

Copyright © 2015 American Society of Plant Biologists. All rights reserved.

http://orcid.org/0000-0002-4390-4195

mailto:[email protected]

http://www.plantphysiol.org

mailto:[email protected]

http://www.plantphysiol.org/cgi/doi/10.1104/pp.114.254581


although the evidence is indirect (Salleo et al., 2009;Brodersen et al., 2010; Nardini et al., 2011).

All parts of plants have a water storage capacity(symplastic and apoplastic), and this capacitance in-creases with tree size and age (Phillips et al., 2003). Theability to mobilize stored water varies according to theforce required to drag it out of storage (Holbrook, 1995).One-half century ago, Reynolds (1965) highlighted theimportance of the volume of internally stored water tosupport the transpiration of trees. Since then, studieshave quantified the fraction of stored water in total dailytranspiration for a range of tree species. This fractionvaries between 2% and 20% (Tyree and Yang, 1990;�Cermák et al., 2007, and refs. therein; Barnard et al., 2011;Pfautsch and Adams, 2013) and is generally smallerin angiosperms compared with gymnosperms, wherea maximum fraction of 50% was reported for Pinussylvestris (Waring et al., 1979). Given that the daily wateruse of large adult trees can easily reach 260 to 380 L(�Cermák et al., 2007; Pfautsch et al., 2011), considerablevolumes of stored water must be mobilized from andrestored back into capacitors on a daily basis. Remobi-lization of stored water also can prolong stomatalopening and thus increase carbon gain (Goldstein et al.,1998).

The volume of stored water released depends on theelasticity of the storage tissue, its connectivity to xylemvessels, and the gradient of water potential (c) betweenthe storage tissue and vessels. Cells with elastic wallsrepresent ideal capacitors because they can change theirvolume as a consequence of small changes in c. Thus,phloem, cambium, and juvenile xylem cells are wellsuited for water storage and release (Yang and Tyree,1992; Zweifel et al., 2014). The magnitude of release andrefill of stored water in trees can be approximated byseparately measuring the change in thickness of phloemand xylem during a diel cycle using high-precisiondendrometers (Zweifel et al., 2014). Whitehead andJarvis (1981) have calculated that around 90% of thediurnal change in stem radius can be attributed tochanges in the water content of cambial and phloemtissues. To date, it is commonly accepted that tree bark,independent of the wood below, swells during the nightand shrinks during the day (Zweifel et al., 2000),reflecting the water flow dynamics that characterize thedynamic exchange of water between phloem and xylem.

Phloem and xylem tissues are separated by rows ofintermediary cambial cells. However, depending on thespecies, phloem and xylem are connected throughuniseriate or multiseriate strands of radially aligned rayparenchyma cells, commonly termed wood rays. Theserays have been shown to be capable of symplastic watertransport through plasmodesmata (Höll, 1975). Basedon measurements of radial conductance, Sevanto et al.(2011) suggested that aquaporins also might be in-volved in the radial transfer of water. Both theoreticaland experimental approaches have been developed tobetter understand the dynamic exchange of water be-tween xylem and phloem. Hölttä et al. (2006, 2009)developed a model based on Münch’s hypothesis and

included a term that represents the hydraulic connec-tion between the two tissue types. Through incorpo-rating this term, model outputs suggest the occurrenceof a constant exchange of water between xylem andphloem along gradients of c. However, some authorssuggested that changes in c of xylem alone might beinsufficient to account for the observed diurnal shrink-age and swelling of bark (Sevanto et al., 2003). Alongthis line of argument, loading and unloading of carbo-hydrates in phloem tissue has been suggested to furtherimpact the radial transfer of water and associatedchanges in bark thickness (Mencuccini et al., 2013).

Nevertheless, to date, all approaches remain indirect,and the routes of water transfer between phloem andxylem have yet to be determined. Here, we presenta technique that enables the visualization of watertransfer from phloem to xylem tissues and resolves theapoplastic and symplastic pathways and cell types. Themethod involves the injection of an aqueous solutionthat contains fluorescent dye into phloem followed byanalyses of woody tissues using confocal laser scanningmicroscopy. We assess the effectiveness of three differ-ent dyes to stain possible transfer pathways. We alsointroduce dye during different time intervals of the dieltranspiration cycle to test the effect of predicted dy-namic changes in c between phloem and xylem on thetransfer processes. We hypothesized that radial transferof water would be most pronounced during periodswhere conditions of the hydraulic nexus betweenphloem and xylem differ the most. These differences areexpected during high rates of transpiration that cause asteep decline in xylem c, commonly observed duringmorning hours. We use leaf water potential (cL) andhigh-precision dendrometer measurements to identifyrelevant time intervals. The simultaneous assessment ofcL and the independent diameter fluctuation of phloemand xylem may provide empirical evidence for the roleof phloem as a water storage capacitor that helps miti-gate increasing tension in the transpiration stream.

RESULTS

Tissue Anatomy

Unstained tissue samples of Eucalyptus saligna stemsclearly show uniseriate rays that extend from phloem,through cambial tissue, into xylem (Fig. 1A). The raycells are adjacent to both sieve element/companion cellcomplexes and parenchyma cells in the phloem tissue(Fig. 1B), while in the xylem they all traverse vesselwalls (Fig. 1, A and C). A close examination of the xy-lem vessels reveals that at the intersection of ray pa-renchyma and the vessel, the inner walls of vessels areperforated with bands of large nonvestured contact pits(Fig. 1D). Where rays are absent, only a small number ofvestured pits connect vessels with surrounding fibercells. These observations confirm that the transpirationstream in xylem vessels is directly connected to phloemtissue through strands of continuous ray parenchyma.

964 Plant Physiol. Vol. 167, 2015

Pfautsch et al.

www.plantphysiol.orgon April 13, 2018 - Published by Downloaded from Copyright © 2015 American Society of Plant Biologists. All rights reserved.


Movement of Fluorescent Dyes from Phloem into Xylem

The three water-soluble fluorescent dyes showed cleardifferences in their utility in tracing the pathway fromphloem into xylem tissue. When injecting fluoresceininto the bark of actively transpiring trees, tissue sur-rounding the point of injection absorbed the fluid thatcontained the dye and left the injection wound empty.This documents the uptake of fluorescein located in theapoplast of the phloem by sieve element/companioncell complexes and parenchyma cells. However, thisabsorption was locally confined to the point of injection,and the tracer did not spread further in these cell types(Fig. 2A). Small amounts of fluorescein also may havebeen absorbed directly by ray parenchyma cells thatwere severed during the drilling process. More impor-tantly, considerable amounts of fluorescein were detectedin ray parenchyma leading away from the point of in-jection, indicating that after uptake from the apoplast, thedye was conducted radially (i.e. assuming bulk flow) inthe symplast of rays.The dye clearly stained a great length of rays toward

the cambium and remained in ray parenchyma, leavingthe surrounding cell types unstained (Fig. 2A). Fluores-cein also was detected in rays that crossed the transitionfrom phloem into cambial tissue (Fig. 2B; see alsoSupplemental Fig. S2.) and was relocated in rays and asmall number of vessels in the xylem (Fig. 2C). In sometissue sections, fluorescein had stained ray parenchyma asfar as 800 mm into the xylem. Generally, when travelingsymplastically in the xylem section of rays, fluorescein was

not detected past the first ray-vessel contact zone.Assessment of discs above and below the point of injectionshowed no signs of vertical movement of fluorescein inphloem tissue. As evidenced by control experiments thatused stem sections, which were cut from freshly felled treesprior to injection of fluorescein, no radial transport of thedye occurred when the c in stem tissues had presumablyequilibrated (for detailed images, see SupplementalFig. S3). This observation is important, as it demonstratesthat the movement of fluorescein in rays requires agradient in c and associated flow of water.

Unlike fluorescein, neither eosin Y nor rhodamin Bwas transported in the symplast of ray parenchymacells. Eosin Y exclusively stained phloem parenchymaltissue adjacent to the point of injection (Fig. 2D). Thedye appeared to slowly diffuse toward the cambiumthrough cell walls of parenchymal tissue. Rhodamin Bremained mostly stationary in the area of injection,staining a few individual ray parenchymal cells, but didnot stain either sieve element/companion cell complexesor phloem parenchymal tissue (Fig. 2E; SupplementalFig. S7). Thus, neither eosin Y nor rhodamin B appears tobe an effective tracer to study symplastic or apoplasticradial transfer of water in tree stems.

Temporal Variation of Plant Hydraulic Traits and RelatedDye Transfer

The day of concurrent cL and dendrometer mea-surements was mostly sunny, with air temperature

Figure 1. Micrographs showing ana-tomical structures in woody tissues ofE. saligna. A, Intact free-hand transversesection showing phloem (right), cam-bium (c), and mature xylem tissue (left),including xylem parenchyma (xp) andvessels (v). Intact wood rays (r) traversefrom phloem through the cambium intothe xylem. Bar = 250 mm. B, Electronmicrograph showing the transverse viewof phloem, where rays extend throughphloem parenchyma (pp) and sieveelement/companion cell complexes (se/cc). Numerous calcium oxalate crystalsare visible. Bar = 100 mm. C, Detailedtransverse view of xylem vessels bor-dered by uniseriate rays surrounded byparenchymal tissue. Bar = 100 mm.D, Electron micrograph (radial plane)depicting vertical wood fibers (f) and thebroad bands of nonvestured contact pits(cp) connecting the vessel with ray cells.Bar = 100 mm. Additional micrographsof wood anatomy are provided inSupplemental Figure S1.

Plant Physiol. Vol. 167, 2015 965

Water Transfer from Phloem to Xylem


http://www.plantphysiol.org/cgi/content/full/pp.114.254581/DC1







reaching 30°C and photosynthetically active radiation(PAR) just below 2,000 mmol m22 s21 (Fig. 3A). Frommorning to midday hours, vapor pressure deficit(VPD) increased from around 0.2 to 2.1 kPa. Around2 PM, a small thunderstorm was responsible for atemporary decrease in air temperature, VPD, and PAR(Fig. 3A). Once the thunderstorm had passed, all weatherparameters resumed their common diel course, and bydusk, they had returned to values similar to those ofthe previous night.

cL at predawn remained stable around 20.19 6 0.06MPa (61 SD) until 9 AM, from where it declined con-tinuously, reaching 21.55 6 0.65 MPa at 12:30 PM (Fig.3B). Thereafter, cL became less negative, returningclose to predawn values after sunset. The gradual in-crease in cL during the afternoon was briefly interruptedduring the thunderstorm, when cL rose up to20.4660.18 MPa.

From midnight until 8 AM, the mean thickness ofphloem and underlying cambial tissue increased by57 6 7.13 mm for the three sampled trees (Fig. 3C). Atthe same time, xylem tissue contracted marginally by3 mm. During daytime, measurements of phloem thick-ness mirrored cL accurately, declining toward midday(Fig. 3C) and increasing during the afternoon. Duringthis time, the movement of xylem showed the exact

opposite trend, albeit at a much lower rate. The totaldaily amplitude of phloem movement spanned morethan 75 mm before reaching dimensions at 9 PM thatwere equal to those at maximum expansion during theearly morning. Based on measurements of cL, we as-sumed that at this time c differences between xylemand phloem were minimal and any further increase ofphloem thickness was mostly the result of cell growth(Fig. 3C). The amplitude of the movement of xylemtissue during daytime averaged 13 mm among trees.

It is noteworthy that the variation in the movementof xylem tissue among the measured trees was largerwhen compared with that of phloem (Fig. 3C). In ad-dition, both tissues responded immediately, and inopposite directions, during the thunderstorm, provid-ing further evidence for a temporally coupled hy-draulic interconnection of phloem and xylem (Fig. 3C).

Injections of fluorescein that specifically targeted thetwo time windows of bark shrinkage (morning) andswelling (evening) showed visually clear and statisti-cally significant differences (P , 0.001) in the distanceof migration of dye in rays (Fig. 4). For these time-of-day injection experiments, measurements of distancesthe dye had moved were pooled from multiple inde-pendent sample specimens (Table I). When injectingdye before 8 AM into phloem and collecting tissue

Figure 2. Staining characteristics (transverse view) several hours after injection of fluorescent dyes into phloem tissue ofE. saligna and emission spectra of tissues and dyes. Arrows indicate the direction of water flow from phloem into xylem tissue.A, Injection wound (iw) of the needle used to introduce fluorescein in phloem tissue and migration of the dye in ray parenchyma.B, Transition of fluorescein from phloem toward xylem tissue. C, Fluorescein passing cambial tissue, entering mature xylem,and staining vessels. D, Injection wound where eosin Y was introduced and spread of the dye into phloem parenchymal tissue.E, Injection wound where rhodamin B was introduced, showing that the dye remained mostly stationary. F, Relative fluores-cence emission spectra of woody tissues (phloem parenchyma and sapwood [peak 1]; autofluorescence), fluorescein (peak 2),eosin Y (peak 3), rhodamin B (peak 4), and chlorophyll (peak 5; autofluorescence). Micrographs are composite images ofbandwidths scanned for woody tissue, fluorescent dye, and chlorophyll. Individual images of all channels are provided inSupplemental Figures S4 to S6. Bars = 200 mm.


Pfautsch et al.



samples around midday, the dye had moved on av-erage 447 6 188 mm into the mature xylem, where itoften reached the first ray-vessel contact zone (Fig.4A). The distance that dye had been translocated inrays varied from 127 to 807 mm, indicating pronouncedvariability among individual rays. In contrast, wheninjecting the dye at dusk and collecting tissue sam-ples at predawn the following morning, the dye in-side ray parenchyma had migrated mostly into thezone of cambial and juvenile xylem tissue (Fig. 4B).Although this incubation period was three times lon-ger (12 h) compared with experiments during themorning hours (4 h), fluorescein had moved on aver-age just 155 6 42 mm. Injections at dusk producedrelatively uniform staining patterns in rays. Measuredtransfer distances ranged between 68 and 254 mm. In

none of the examined samples did the stain reach maturexylem and conducting vessels.

DISCUSSION

The combination of experimental (in situ and ex situ)and analytical techniques applied here provides directevidence for the radial transfer of water from phloemto xylem tissue in tree stems, as predicted from theory(Hölttä et al., 2006; Lacointe and Minchin, 2008) andempirical but indirect measurements (De Schepper andSteppe, 2010; Sevanto et al., 2011) of the hydraulicnexus of phloem and xylem. This transfer of water ispredominantly facilitated through the symplast inuniseriate ray parenchyma cells, driven by transpira-tion and associated gradients in c between the twotissue types. We do acknowledge that by injecting thedye we cannot separate diffusion from bulk flow offluorescein in the symplast of ray parenchyma and thatthe final distance the dye reached in rays is the resultof both movements. However, transfer of dye did notoccur due to diffusion, as documented by the results ofour ex situ control experiment using isolated, cut stemsections. Hence, this transfer must be mediated bybulk flow of both water and dye. It is reasonable tospeculate that water and stain entered the rays usingdifferent mechanisms (e.g. aquaporins versus plas-modesmata). Their exact routes of uptake remain to beconfirmed experimentally.

The initial uptake of water containing the tracermust have taken place from the apoplast into thesymplast of ray parenchyma cells. This indicates that cis slightly lower in the symplast of these cells com-pared with the apoplast and other cell types present inthe phloem. However, the fact that once the tracer wasinside the symplast of ray parenchyma cells no stain-ing of any surrounding cells was observed underpinsthe dominance and tight regulation of this transportpathway. Although we did not detect any fluorescencein the apoplastic space, we cannot entirely exclude thepossibility that some water molecules of the water-tracer solution followed this pathway. Once taken upinto living ray cells, the tracer moved inward by dif-fusion and/or bulk flow, passing through plasmo-desmata that connect ray cells. The importance andinvolvement of the plasmodesmatal pathway in thetranslocation of organic and inorganic substances havelong been known (Sauter and Kloth, 1986). As fluores-cein cannot diffuse through plasma membranes (andthus is a tracer specific to symplastic pathways), thetracer was retained in the region of contact pits whereonly small traces of the dye may have leaked out intothe adjacent xylem vessel.

Our results provide support for a functional role ofwater transfer in phloem during periods of increasingtension in the xylem (morning) compared with timeswhen differences of c in xylem and phloem are con-siderably smaller (night) or absent (ex situ controls). Inaddition, the movement of water necessarily implies

Figure 3. Environmental and physiological conditions during a dieltime course (March 5, 2014). A, PAR and VPD. B, Diurnal course of cL

collected from the upper canopy of three E. saligna trees (n = 9 leavesper time point; error bars depict SD). C, Differential of the movement inphloem (black line) and xylem tissue (blue line) in basipetal stemsections of E. saligna (n = 3 trees; shaded areas represent SD) zeroedaccording to measurements recorded at time 0. Numbers and arrowsindicate the phase of rapid shrinkage of phloem thickness duringmorning hours (peak 1) and growth during the night (peak 2); the graydotted line marks the point where phloem rehydration is complete andwood growth occurs (for details, see text).





differences in the c gradient between phloem andxylem, suggesting a greater complexity in internalwater transfer in trees. Once more, our ex situ controlexperiment, where equilibrium of c in phloem andxylem can be assumed, confirms that radial transferoccurs once a difference in c is present. We have foundno traces of the dye in rays when the transpirationstream had been disrupted, not even close to the pointof injection and regardless of the incubation time.

Our results highlight the function of phloem tissueas a potential contributor to stem water storage andthus capacitance in trees, as reported by Zweifel et al.(2000), Pfautsch et al. (2011), and others. Together withthe excellent work of Sokołowska and Zagórska-Marek (2012), which documents the radial transferof water from xylem to cambial cells, we now haveconclusive visual evidence for the occurrence of bi-directional flow of water between phloem and xylem.Their and our studies used fluorescent tracers. Thus,fluorescein-based dyes appear highly effective for usein studies of the hydraulic nexus. While fluorescein hasbeen shown to enter ray parenchyma cells when in-fusing angiosperm xylem vessels with a dye solutionex situ (Cirelli et al., 2008; Sokołowska and Zagórska-Marek, 2012), our method is suited for in situ appli-cations, where naturally occurring changes in osmoticand hydraulic pressures facilitate the distribution ofdye and the tracing of symplastic pathways.

The fast and highly efficient response to injuries ob-served in phloem tissue (Knoblauch and Mullendore,2013, and refs. therein) works to the advantage of ourmethod. Sieve elements contain large quantities ofP-protein that is mobilized immediately after an injury tosieve element/companion cell complexes, clogging sieveplates and closing the wound by the formation of callosetissue (Knoblauch and van Bel, 1998). This reaction isseen as an adaptation to the pressurized state whereuncontained wounds could cause extensive leaking oftransported assimilates (van Bel et al., 2002). We did notdetect any fluorescence in tissues sampled a few centi-meters above or below the point of injection and attri-bute this result to the effective clogging of sieve plates byP-protein and possibly also by the dislodgement of smallcellular structures during the drilling of injection holes(Knoblauch and Mullendore, 2013). We thus excludeany appreciable loss of the injected tracer due to axialtransport in the phloem.

Our study showed that not every water-soluble dye issuitable to trace the movement of water in living tissueof phloem, particularly the loading from the apoplastand symplastic transport in ray parenchyma cells. Themolar mass of sodium fluorescein is 376 g mol21,similar to that of Suc (342 g mol21). Suc is the primaryform for the transport of carbon in phloem of mostplants, and membrane-bound shuttles for this moleculeare in high abundance (van Bel and Thompson, 2013).Hence, using a fluorescent dye of comparable molecularmass perhaps increases the chances for staining ofsymplastic pathways. Contrary to fluorescein, themovement of eosin Y, and more so that of rhodamin B,in phloem and rays was very limited. The molar massof these two dyes is substantially greater compared withfluorescein or Suc (eosin Y, 691.85 g mol21; rhodamin B,479.02 g mol21). We acknowledge that other factors, suchas polarity or the affinity to chelate (Boens et al., 2007),could further influence the effectiveness of a dye totrace either apoplastic or symplastic pathways in woodytissues.

Observed differences in tracer localization among celltypes, however, raise an interesting question. Is the ef-ficiency of symplastic transport of water in rays superiorto that of the apoplast (i.e. cell walls and intercellularspaces)? Fluorescein, the smaller molecule, moved muchfarther when transported in ray parenchyma comparedwith eosin Y, the larger molecule, diffusing in the apo-plast. Also, other dye experiments have found that thesymplast is favored over the apoplast in transport routes

Figure 4. Contrasting day versus nighttime progression of fluorescein(transverse view) in ray parenchyma of E. saligna. Dashed lines indi-cate borders between tissue types, where p = phloem, c = cambialzone and juvenile xylem, and x = mature xylem. Yellow bars show themaximum distance that fluorescein had progressed. A, Injection of thedye at 8 AM, with tissue harvest at 12 noon. B, Injection of the dye at7 PM, with tissue harvest prior to 7 AM the following morning. Bars =200 mm.

Table I. Details of in situ dye injections into the phloem tissue of E. saligna tree stems during two different time intervals

Distance data are means and (SD).

Time of Injection No. of Trees Total Injections Sample SpecimensaNo. of Rays Containing

Fluorescein Measured

Distance of Fluorescein Spread

from the Phloem-Cambium Boundary

mm

Morning 6 12 9 49 447.1 (188.4)Evening 6 12 8 44 155.1 (42.0)

aNumber of sample specimens of intact bark-phloem-cambium-sapwood sequences at the point of injection used for image analysis.


Pfautsch et al.



(Sano et al., 2005; Sokołowska and Zagórska-Marek,2012). As shown in the scanning electron micrographs,the flow of large quantities of water could be accom-modated through the wide and numerous contact pitsthat connect rays with vessels. Such connections havebeen described for many woody species (Schoot andvan Bel, 1989) and are not present where rays connect tosurrounding phloem cells. These connections have yetto be characterized in detail (van Bel, 1990; Patrick,2013). It could be speculated that if rays are well con-nected at both ends of the pathway, a lower resistanceto symplastic flow and a good connection to xylemvessels would make this pathway superior to that of theapoplast. Experiments using dyes with different prop-erties from fluorescein (e.g. Texas Red, which is nottaken up by cells, and isotopic tracers like deuteratedwater) are needed to improve our understanding ofapoplastic and symplastic pathways and their quanti-tative importance.Further insight into plant hydraulic functioning

arises when tracing these transport pathways. Here,we showed that fluorescein did not migrate beyondthe point where the individual ray intersected the veryfirst vessel. Our anatomical observations support theconclusion that the bulk of the water transferred fromphloem to xylem was incorporated into the transpira-tion stream, as fluorescein never passed the most re-cently formed vessels in which c would be morenegative compared with that in the adjacent ray, fiber,or parenchymal cells.Differences in the distance that fluorescein had mi-

grated following injections during the morning andevening are in agreement with the theoretical work ofHölttä et al. (2009). c gradients are naturally greatestduring periods of high transpiration, reflected in ourmeasurements of declining cL and associated contrac-tion of phloem tissue during the morning hours. Theseconditions resulted in translocation of the dye fartherinto mature xylem. In contrast, the dye was onlytranslocated into the region of cambial and juvenilexylem tissue during the night, when transpiration isgenerally slow. The efflux of water from phloem dur-ing times of low transpiration is mostly a combinationof unloading of carbohydrates to fuel cell expansion inthe cambial zone and Münch counterflow (Hölttä et al.,2006). During our control experiments, where transpi-ration had ceased altogether in cut stem sections, thedye did not enter the rays at all.

CONCLUSION

By direct injection of fluorescein into phloem tissueof E. saligna, we have provided visual evidence thatradial transfer of water from phloem into xylem occurspredominantly via the symplast of ray parenchymacells. The distance of dye migration into xylem differedbetween day and night as a consequence of differencesin c between the tissue types. The results presentedhere show that injections of fluorescent dyes, together

with measurements that capture the hydraulic func-tioning of trees, can successfully be used to improve ourunderstanding of whole-tree physiology. Although animportant first step, our findings provide guidance forfuture research. Comparisons of apoplastic and sym-plastic pathways as well as the quantification and tim-ing of radial transfer of water appear to be the mostpressing issues.

MATERIALS AND METHODS

Plant Material

We used more than 30 stems of Eucalyptus saligna located on the premises ofthe Hawkesbury Forest Experiment, 60 km northwest of Sydney, Australia (25 mabove sea level; 33° 369 4099 S, 150° 449 26.599 E). The trees were between 6 and8 m tall and had regenerated from rootstocks that were row planted in a plan-tation. The diameter of the stems ranged between 5.1 and 7.2 cm, and thethickness of the smooth gum-type bark was 4 to 5 mm. The climate of the area istemperate subhumid, with an average annual temperature of 17°C. Mean annualprecipitation at the site is around 800 mm, with wet summers and relatively drywinters. We measured the cL and dynamic changes of bark and sapwoodthickness of three trees on March 5, 2014. Environmental conditions of that daywere recorded on site by a weather station (CR3000; Campbell Scientific), in-cluding air temperature (°C), relative humidity (%; HMP45C; Vaisala), and PAR(mmol m22 s21; SQ-110; Apogee Instruments). VPD (kPa) was calculated usingair temperature and relative humidity according to Snyder and Shaw (1984). Alldye injections and associated sample collections and analyses took place be-tween March 7 and May 16, 2014, during periods of generally sunny conditions.

cL

cL is a good indicator for hydraulic tension in xylem, and measurements ofcL were used as surrogates to characterize the daily evolution of tension in thexylem water of stems. For this purpose, three leaves were randomly sampledfrom the upper canopy of three E. saligna trees (n = 9) at hourly intervalsbetween predawn and postdusk (6 AM–7 PM). After excision, leaves wereplaced in a darkened bag that contained a moist paper towel to prevent fur-ther loss of water. Bags were immediately transferred to a nearby researchlaboratory, where cL was measured using a pressure chamber (1505D; PMSInstrument).

Dendrometer Measurements

The diurnal variation in the diameter of phloem and xylem was recordedusing high-precision point dendromenters (ZN11-O100-2WP; Natkon). Oneweek prior to measurements of cL , pairs of dendrometers were installed 70 cmabove the ground on three stems. A 3- 3 3-cm-wide bark window was re-moved, and the exposed sapwood was washed with distilled water to removecambial cells. The transducer of the dendrometer was placed onto the sap-wood, and the window was sealed using silicon grease. The transducer of asecond dendrometer was placed onto the bark surface. Both dendrometerswere mounted onto a carbon fiber ring that was affixed to the stem by drillingscrews into the sapwood. This experimental setup is described in detail byZweifel et al. (2014). Data were logged (CR1000; Campbell Scientific) at 5-minintervals. We subtracted values measured on sapwood from those measuredon bark to obtain values that represent the diameter variation of phloem andunderlying cambial cells through time.

Fluorescent Dyes

Fluorescein derivatives in aqueous solution have been shown to move wellin sieve tubes of the phloem and vessels and ray parenchyma of the xylem(Grignon et al., 1989; Cirelli et al., 2008; Sokołowska and Zagórska-Marek,2012), but care has to be taken, as emission characteristics may be similar tothose of woody tissues (Palmquist, 1938). This study used a 0.1% (w/w)solution of fluorescein (sodium fluorescein; Sigma-Aldrich) that exceedsthe autofluorescence of woody tissues, emitting maximally at 520 nm and





exceeding the more blue-shifted autofluorescent emissions of woody tissues.In addition, the mobility of two other fluorescent dyes, namely eosin Y andrhodamin B, also was assessed using solutions at 0.1% (w/w) strength. Eosinhas been shown to trace the uptake of water through twigs in Picea abies,where it was detected in wood and bark (Katz et al., 1989). A rhodamin dyehas been used successfully to trace water movement in leaf veins of maize (Zeamays) plants (Canny and McCully, 1986).

Dye Injections and Sample Collection

Small holes were drilled in a tangential direction into the phloem tissue withthe help of hypodermic needles (o.d. 0.64 mm) connected to 5-mL syringes.Each hole was filled with fluorescent dye using minimal force, ensuring that atleast 100 mL of the solution was released into each hole. During initial ex-periments, dyes were allowed to disperse for 2 d. After the assessment of theseinjections with three dyes, it became clear that only fluorescein could bedetected in xylem ray parenchyma. For this reason, fluorescein was used insubsequent experiments to examine the diurnal radial movement of water viashorter time intervals between injection and sample collection (morning =8 AM–12 noon; night = 6:30 PM–6:30 AM the following day). For each time interval,we used six tree stems. Fluorescein was injected at two opposing positions intothe phloem tissue at the basal segment of the stem. In addition to these in situinjections, we also included ex situ injections as a control experiment. For thisexperiment, we excised basal stem sections (30–35-cm length) of 10 trees be-tween 7 and 8 AM. Immediately after excision, both cut surfaces were sealed toprevent any evaporative water loss and associated movement of water. Thesections were left to rest for 1 h before fluorescein was injected following theidentical procedure described above. Stem sections were positioned uprightand left for a minimum of 6 h under ambient conditions in the laboratory.

A total of 42 stem discs (20-mm thickness) containing the point of injectionwere collected using a pruning saw. Adjacent discs from above and below thepoint of injection (n = 84) were also cut to assess possible vertical movement ofthe dye. Promptly after cutting the discs, samples were wrapped in moistpaper towels, placed in a plastic bag, and transferred to the laboratory, wherethe region of dye injection was extracted using a hammer and chisel, sepa-rating wood samples into stained and nonstained material. Nonstained ma-terial was used to prepare free-hand and microtomed sections. The samematerial was used for scanning electron microscopy, for which all sides ofsample blocks were trimmed using a microtome to provide clean and smoothsurfaces. Samples from the injection area were cut into cubes of approximately8 to 10 mm, and the radial surface of one side of such blocks was trimmedwith a razor blade to provide a clean and smooth surface for scanning laserconfocal microscopy. Extra care was taken during all processing steps to en-sure that the connection between bark and sapwood remained intact. Allsample blocks were kept at 4°C and processed within 24 h.

Scanning Electron and Optical Microscopy

Fresh sample blocks were mounted on aluminum stubs with double-sidedcarbon adhesive tape, allowing imaging of radial and transverse planes. Ascanning electronmicroscope (6510LV; JEOL) operated in low-vacuummode ata vacuum chamber pressure of 50 Pa, 20-kV accelerating voltage, and 15-mmworking distance was used for imaging the surface of the samples.

Free-hand sections of intact bark-phloem-cambium-sapwood sequenceswere prepared using a razor blade and transferred onto glass slides. Radialsections (25-mm thickness) of sapwood were prepared from 2-cm cubes usinga sliding microtome (Leica RM2255; Leica Microsystems). All microtomedsections were stained with safranin and washed clean before embedding onglass slides with a coverglass. Digital images of stained (103) and unstained(43) samples were taken using a transmission light microscope (Leica SM2010R; Leica Microsystems) equipped with a high-resolution (12 megapixel) digitalcamera (Leica DFC 500; Leica Microsystems).

Laser Scanning Confocal Microscopy

Samples stained with fluorescent dyes were analyzed using an invertedLeica TCS SP5 laser scanning confocal microscope (Leica Microsystems) suitedfor live-cell microscopy. This microscope uses the Acousto Optical BeamSplitter for excitation-emission separation, making it particularly suited forimaging and separating multiple fluorophores without the need to use specificfilters or beam splitters. Confocal imaging of the phloem and the xylem

components was by excitation with the 405-nm diode laser to detect auto-fluorescence at 410 to 470 nm. A 488-nm excitation line of the argon laser wasused for the detection of fluorescein and eosin, with the detection bandwidth setat 500 to 560 nm. The DPSS 561-nm laser was used for the excitation of samplesstained with rhodamin, and detection was set at 570 to 630 nm. Chlorophyllwas imaged with an emission channel set at 670 to 700 nm. Images of thedifferent tissues of the sample were collected using a confocal zoom magni-fication setting of 103. Scanning three to seven frames for each image achievedreduction of the signal-to-noise ratio. Serial optical sections in xyz mode werecollected to image samples in three dimensions, from surface to the requireddepths. The z series from the image stacks were reconstructed into three-dimensional views using Leica LAS software. Microspectral imaging was donewith the xyl scan mode using various laser lines for excitation of fluorophores,wood, and chlorophyll autofluorescence.

We tested extensively for spectral separation of tissue and chloroplastautofluorescence and fluorescence emission of the dyes and found goodseparation between tissues and compounds (Fig. 2F). Full details of the spectralscans shown as composite images in Figure 2, A, D, and E, are provided inSupplemental Figures S4 to S6.

The distance that fluorescein had traveled in individual rays during themorning and night was measured manually using the Leica Application SuiteAdvanced Fluorescence Software (version 4.2; Leica Microsystems). Thestarting point for these measurements was at the boundary of phloem andcambial cells. We measured only rays showing clear evidence of fluorescein forat least 30 mm past this border. Shapiro-Wilk’s test confirmed a normal dis-tribution of the measured distances of dye movement in the two samplepopulations. A statistical test of differences in measurements between themorning and night injection treatments was conducted using the Kruskal-Wallis test for nonhomogenous sample sizes between populations.

Supplemental Data

The following supplemental materials are available.

Supplemental Figure S1. Micrographs of phloem and xylem tissues ofE. saligna.

Supplemental Figure S2. Composite micrographs documenting the spreadof fluorescein in phloem tissue of E. saligna in situ.

Supplemental Figure S3. Composite micrographs documenting the spreadof fluorescein in phloem tissue of E. saligna ex situ.

Supplemental Figure S4. Detailed micrographs for Figure 2A.

Supplemental Figure S5. Detailed micrographs for Figure 2D.

Supplemental Figure S6. Detailed micrographs for Figure 2E.

Supplemental Figure S7. Detailed micrographs of rhodamin B in phloemtissue of E. saligna.

ACKNOWLEDGMENTS

We thank Liz Kabanoff (School of Science and Health, University ofWestern Sydney) for help in selecting the dyes and the site managers atthe Hawkesbury Forest Experiment and Craig Barton and Burhan Ahmiji(Hawkesbury Institute for the Environment, University of Western Sydney),for the provision of plant material and practical help in the field. We alsothank Richard Wuhrer (Advanced Materials Characterisation Facility at theUniversity of Western Sydney) for help with scanning electron microscopyimaging. Comments on the draft article by Maurizio Mencuccini (University ofEdinburgh) and two anonymous reviewers were highly appreciated.

Received November 28, 2014; accepted January 12, 2015; published January14, 2015.

LITERATURE CITED

Barnard DM, Meinzer FC, Lachenbruch B, McCulloh KA, Johnson DM,Woodruff DR (2011) Climate-related trends in sapwood biophysicalproperties in two conifers: avoidance of hydraulic dysfunction throughcoordinated adjustments in xylem efficiency, safety and capacitance.Plant Cell Environ 34: 643–654


Pfautsch et al.











Biddulph O, Markle J (1944) Translocation of radio-phosphorus in thephloem of the cotton plant. Am J Bot 31: 65–70

Boens N, Qin W, Basari�c N, Hofkens J, Ameloot M, Pouget J, Lefèvre JP,Valeur B, Gratton E, vandeVen M, et al (2007) Fluorescence lifetimestandards for time and frequency domain fluorescence spectroscopy.Anal Chem 79: 2137–2149

Brodersen CR, McElrone AJ, Choat B, Matthews MA, Shackel KA (2010)The dynamics of embolism repair in xylem: in vivo visualizations usinghigh-resolution computed tomography. Plant Physiol 154: 1088–1095

Canny MJ, McCully ME (1986) Locating water-soluble vital stains in planttissues by freeze-substitution and resin-embedding. J Microsc 142: 63–70

�Cermák J, Ku�cera J, Bauerle WL, Phillips N, Hinckley TM (2007) Treewater storage and its diurnal dynamics related to sap flow and changesin stem volume in old-growth Douglas-fir trees. Tree Physiol 27: 181–198

Cirelli D, Jagels R, Tyree MT (2008) Toward an improved model of maplesap exudation: the location and role of osmotic barriers in sugar maple,butternut and white birch. Tree Physiol 28: 1145–1155

De Schepper V, De Swaef T, Bauweraerts I, Steppe K (2013) Phloemtransport: a review of mechanisms and controls. J Exp Bot 64: 4839–4850

De Schepper V, Steppe K (2010) Development and verification of a waterand sugar transport model using measured stem diameter variations.J Exp Bot 61: 2083–2099

Dixon HH, Joly J (1895) On the ascent of sap. Philos Trans R Soc B 186:563–576

Ferrier JM, Tyree MT, Christy AL (1975) The theoretical time-dependentbehavior of Münch pressure-flow system: the effect of sinusoidal timevariation in sucrose loading and water potential. Can J Bot 53: 1120–1127

Goldstein G, Andrade JL, Meinzer FC, Holbrook NM, Cavelier J, JacksonP, Celis A (1998) Stem water storage and diurnal patterns of water usein tropical forest canopy trees. Plant Cell Environ 21: 397–406

Grignon N, Touraine B, Durand M (1989) 6(5)Carboxyfluorescein as atracer of phloem sap translocation. Am J Bot 76: 871–877

Holbrook NM (1995) Stem water storage. In BL Gartner, ed, Plant Stems:Physiology and Functional Morphology. Academic Press, San Diego, pp151–174

Höll W (1975) Radial transport in rays. In MH Zimmermann, JA Milburn,eds, Transport in Plants I. Springer, Berlin, pp 432–450

Hölttä T, Mencuccini M, Nikinmaa E (2009) Linking phloem function tostructure: analysis with a coupled xylem-phloem transport model. JTheor Biol 259: 325–337

Hölttä T, Vesala T, Sevanto S, Perämäki M, Nikinmaa E (2006) Modelingxylem and phloem water flows in trees according to cohesion theory andMünch hypothesis. Trees (Berl) 20: 67–78

Katz C, Oren R, Schulze ED, Milburn JA (1989) Uptake of water andsolutes through twigs of Picea abies (L.) Karst. Trees (Berl) 3: 33–37

Knoblauch M, Mullendore D (2013) Sieve element occlusion. In GAThompson, AJE Van Bell, eds, Phloem: Molecular Cell Biology, SystemicCommunication, Biotic Interactions. Wiley-Blackwell, Oxford, pp 141–153

Knoblauch M, van Bel AJE (1998) Sieve tubes in action. Plant Cell 10: 35–50Lacointe A, Minchin PEH (2008) Modelling phloem and xylem transport

within a complex architecture. Funct Plant Biol 35: 772–780Mencuccini M, Hölttä T, Sevanto S, Nikinmaa E (2013) Concurrent mea-

surements of change in the bark and xylem diameters of trees reveal aphloem-generated turgor signal. New Phytol 198: 1143–1154

Münch E (1930) Die Stoffbewegungen in der Pflanze. Gustav Fischer, Jena,Germany

Nardini A, Lo Gullo MA, Salleo S (2011) Refilling embolized xylem con-duits: is it a matter of phloem unloading? Plant Sci 180: 604–611

Palmquist EM (1938) The simultaneous movement of carbohydrates andfluorescein in opposite directions in the phloem. Am J Bot 25: 97–105

Patrick JW (2013) Fundamentals of phloem transport. In GA Thompson,AJE Van Bell, eds, Phloem: Molecular Cell Biology, Systemic Commu-nication, Biotic Interactions. Wiley-Blackwell, Oxford, pp 30–59

Pfautsch S, Adams MA (2013) Water flux of Eucalyptus regnans: defy-ing summer drought and a record heatwave in 2009. Oecologia 172:317–326

Pfautsch S, Bell TL, Gessler A (2015) Uptake, transport and redistributionof amino-N in woody plants. In JPF De’Mello, ed, Amino Acids inHigher Plants. CABI Publishing, Wallingford, UK, p 34

Pfautsch S, Gessler A, Adams MA, Rennenberg H (2009) Using amino-nitrogen pools and fluxes to identify contributions of understory Acacia

spp. to overstory Eucalyptus regnans and stand nitrogen uptake intemperate Australia. New Phytol 183: 1097–1113

Pfautsch S, Keitel C, Turnbull TL, Braimbridge MJ, Wright TE, SimpsonRR, O’Brien JA, Adams MA (2011) Diurnal patterns of water use inEucalyptus victrix indicate pronounced desiccation-rehydration cyclesdespite unlimited water supply. Tree Physiol 31: 1041–1051

Phillips NG, Ryan MG, Bond BJ, McDowell NG, Hinckley TM, Cermák J(2003) Reliance on stored water increases with tree size in three speciesin the Pacific Northwest. Tree Physiol 23: 237–245

Reynolds ERC (1965) Transpiration as related to internal water content.Nature 207: 1001–1002

Salleo S, Trifilò P, Esposito S, Nardini A, Lo Gullo M (2009) Starch-to-sugar conversion in wood parenchyma of field-growing Laurus nobilisplants: a component of the signal pathway for embolism repair? FunctPlant Biol 36: 815–825

Sano Y, Okamura Y, Utsumi Y (2005) Visualizing water-conductionpathways of living trees: selection of dyes and tissue preparationmethods. Tree Physiol 25: 269–275

Sauter JJ, Kloth S (1986) Plasmodesmatal frequency and radial transloca-tion rates in ray cells of poplar (Populus 3 canadensis Moench‘Robusta’). Planta 168: 377–380

Schoot C, van Bel AJ (1989) Architecture of the internodal xylem of tomato(Solanum lycopersicum) with reference to longitudinal and lateral trans-fer. Am J Bot 76: 487–503

Sevanto S, Hölttä T, Holbrook NM (2011) Effects of the hydraulic couplingbetween xylem and phloem on diurnal phloem diameter variation. PlantCell Environ 34: 690–703

Sevanto S, McDowell NG, Dickman LT, Pangle R, Pockman WT (2014)How do trees die? A test of the hydraulic failure and carbon starvationhypotheses. Plant Cell Environ 37: 153–161

Sevanto S, Vesala T, Perämäki M, Nikinmaa E (2003) Sugar transporttogether with environmental conditions controls time lags between xy-lem and stem diameter changes. Plant Cell Environ 26: 1257–1265

Snyder RL, Shaw RH (1984) Converting Humidity Expressions with Computersand Calculators. Cooperative Extension, Leaflet 21372, Division of Agricul-ture and Natural Resources, University of California, Davis

Sokołowska K, Zagórska-Marek B (2012) Symplasmic, long-distancetransport in xylem and cambial regions in branches of Acer pseudo-platanus (Aceraceae) and Populus tremula 3 P. tremuloides (Salicaceae).Am J Bot 99: 1745–1755

Stout PR, Hoagland DR (1939) Upward and lateral movement of salt incertain plants as indicated by radioactive isotopes of potassium, sodiumand phosphorus absorbed by roots. Am J Bot 26: 320–324

Tyree MT, Yang S (1990) Water storage capacity of Thuja, Tsuga and Acerstems measured by dehydration isotherms. Planta 182: 420–426

van Bel AJE (1990) Xylem-phloem exchange via the rays: the undervaluedroute of transport. J Exp Bot 41: 631–644

van Bel AJE (2003) The phloem, a miracle of ingenuity. Plant Cell Environ26: 125–149

van Bel AJE, Ehlers K, Knoblauch M (2002) Sieve elements caught in theact. Trends Plant Sci 7: 126–132

van Bel AJE, Thompson GA (2013) Phloem, the integrative avenue. In GAThompson, AJE Van Bell, eds, Phloem: Molecular Cell Biology, SystemicCommunication, Biotic Interactions. Wiley-Blackwell, Oxford, pp 3–29

Wardlaw IF (1974) Phloem transport: physical chemical or impossible.Annu Rev Plant Physiol 25: 515–539

Waring RH, Whitehead D, Jarvis PG (1979) The contribution of storedwater to transpiration in Scots pine. Plant Cell Environ 2: 309–317

Whitehead D, Jarvis PG (1981) Coniferous forests and plantations. In TTKozlowski, ed, Water Deficits and Plant Growth. Academic Press, NewYork, pp 49–152

Yang S, Tyree MT (1992) A theoretical model of hydraulic conductivityrecovery from embolism with comparison to experimental data on Acersaccharum. Plant Cell Environ 15: 633–643

Zimmermann MH (1983) Xylem Structure and the Ascent of Sap. Springer,Berlin

Zweifel R, Drew D, Schweingruber F, Downes GM (2014) Xylem as themain origin of stem radius changes in Eucalyptus. Funct Plant Biol 41:520–534

Zweifel R, Item H, Häsler R (2000) Stem radius changes and their relation tostored water in stems of young Norway spruce trees. Trees (Berl) 15: 50–57





phloem as capacitor: radial transfer of water into xylem of tree

Documents