Moderate to severe water limitation differentially affectsthe phenome and ionome of Arabidopsis

Lucia M. Acosta-GamboaA, Suxing LiuA, Erin LangleyA, Zachary CampbellA,Norma Castro-GuerreroB, David Mendoza-CozatlB,D and Argelia LorenceA,C,D

AArkansas Biosciences Institute, Arkansas State University, PO Box 639, State University, AR 72467, USA.BDivision of Plant Sciences, Christopher S Bond Life Sciences Center, University of Missouri, 1201 Rollins Street,Columbia, MO 65211, USA.

CDepartment of Chemistry and Physics, Arkansas State University, PO Box 429, State University, AR 72467, USA.DCorresponding authors. Emails: alorence@astate.edu; mendozacozatld@missouri.edu

Abstract. Food security is currently one of the major challenges that we are facing as a species. Understanding plantresponses and adaptations to limited water availability is key to maintain or improve crop yield, and this is even morecritical considering the different projections of climate change. In this work, we combined two high-throughput -‘omic’platforms (‘phenomics’ and ‘ionomics’) to begin dissecting time-dependent effects of water limitation in Arabidopsisleaves and ultimately seed yield. As proof of concept, we acquired high-resolution images with visible, fluorescence, andnear infrared cameras and used commercial and open source algorithms to extract the information contained in those images.At a definedpoint, sampleswere also taken for elemental profiling.Our results show that growth, biomass andphotosyntheticefficiency were affected mostly under severe water limitation regimes and these differences were exacerbated at laterdevelopmental stages. The elemental composition and seed yield, however, changed across the different water regimestested and these changes included under- and over- accumulation of elements compared with well-watered plants. Ourresults demonstrate that the combination of phenotyping techniques can be successfully used to identify specific bottlenecksduring plant development that could compromise biomass, yield, and the nutritional quality of plants.

Additional keywords: Arabidopsis thaliana, drought stress, high-throughput plant phenotyping, ionomics, phenomics.

Received 6 May 2016, accepted 5 September 2016, published online 20 October 2016

Introduction

Water availability is oneof themajor factors limiting plant growthand yield worldwide and global climate change is expected tocompromise water resources around the globe even further(Chaves et al. 2003). At the same time, world population isgrowing and the demand for food is also expected to rise overthe next decades. Thus, the challenge ahead of us is to developcrop varieties able to endure prolonged periods of stress withoutexperiencing significant loses in nutritional quality or yield.Significant advances have been made at the physiological andmolecular level to understand how some plant species areable to thrive in places where water is scarce (Yang et al.2015). In addition, plant responses to water limitation havebeen documented extensively in model and crop plantsusing growth chambers, greenhouses, field conditions andeven different growth media (e.g. agar, soilless systems andsoil with different degrees of water saturation). Morphologicaland physiological changes that occur during water limitationstress include decrease in transpiration and photosynthesis, andreduced biomass (Boyer 1970; Tardieu et al. 1999; Farooqet al. 2009). Metabolic changes that have also been reported inresponse to water limitation include the reduction of net carbon

assimilation rate due to stomatal closure and low CO2 diffusion,leading to a downregulation of the photosynthetic machinery(reviewed by Farquhar and Sharkey 1982; McDowell et al. 2008and Farooq et al. 2009). For many crop species, limited wateravailability impacts yield and the severity of yield loss dependson the level and duration of the stress (Farooq et al. 2009). Sinceyield integrates many phenotypical and physiological processesin a complex way, it is often difficult to predict what type ofchanges and adaptations during the life cycle of the plantultimately will led to the observed yield penalty when plantsexperience water limitation stress. All these studies haveprovided valuable insights into the different responses andadaptations that plants have evolved when water is limiting.These studies have also made clear that the timing of treatment,the plant developmental stage, and the severity of the treatmentare critical to define the type and magnitude of the plant response(for a detailed review about physiological and molecularresponses to drought see Claeys and Inzé 2013).

Over the last decade, significant technological advanceshave made possible the tracking of plant growth anddevelopment in a high-throughput and automated manner.This approach, termed high-throughput plant phenotyping

CSIRO PUBLISHING

Functional Plant Biology, 2017, 44, 94–106http://dx.doi.org/10.1071/FP16172

Journal compilation � CSIRO 2017 www.publish.csiro.au/journals/fpb

(HTPP or ‘phenomics’), allows the acquisition and quantificationof phenotypes from hundreds and even thousands of plantsin a short period of time. HTPP platforms typically usehigh-resolution cameras to capture images in the visible,fluorescence, and infrared ranges to quantify plant size,architecture, colour, in planta chlorophyll content, in plantawater content, and leaf temperature among other readoutsof interest (Fahlgren et al. 2015a). Benefits from HTPPapproaches include the fact that these methods are fast, non-invasive, unbiased and accurate. In addition, if documentationof those HTTP experiments is done properly, the capturedimages can be reanalysed and additional information can beextracted if novel algorithms or additional questions emergeafter the experiments have been completed (Fahlgren et al.2015a; Junker et al. 2015). Additional high-throughputtechniques, often called –‘omic’ approaches, have also beenestablished to identify and quantify metabolites (‘metabolomics’),the elemental composition or ions (‘ionomics’), gene expression(‘transcriptomics’) and protein abundance (‘proteomics’) inbiological samples. In principle, these techniques can be usedindependently of each other towards a full of understanding ofa biological system; however, data integration and interpretationof separate high-throughput approaches has proven to bechallenging. This is particularly true when these techniquesare applied in different laboratories, under slightly differentexperimental settings, thus altering the timing and magnitudeof the plant response, which complicates later the ability tointegrate -omics data seamlessly. To address this issue, wehave begun the systematic combination of high-throughputtechnologies to first identify critical points during plantdevelopment where additional -omic techniques can be broughtin to increase the depth of analysis in a more meaningful way.

Plants affected by water limitation need to undergo aperiod of osmotic adjustment and this process is necessary tomaintain water retention and turgor (Wang et al. 2013), thusallowing the cells to tolerate better the drought stress. Thisis possible in part by a variety of modifications that includechanges in ion uptake, distribution and growth arrest.‘Ionomics’, or the study of the elemental composition of anorganism or tissue, has been successful at identifying genesresponsible for controlling the accumulation of one or a groupof elements in different organisms (Lahner et al. 2003; Eideet al. 2005). One area that has received much less attention,despite its physiological relevance, is the effect of waterlimitation on the plant ‘ionome’. Here, as proof-of-concept,we first used HTPP to document the ‘phenome’ of Arabidopsisgrown under different water availability regimes. From thesedata, we selected a developmental stage where an additionalhigh-throughput technology (‘ionomics’) was used to monitorthe elemental composition of plants grown under differentwaterregimes. Our data support the idea that the sequential use ofhigh-throughput technologies may be beneficial to increase theanalytical depth at critical points during plant development.Future work could definitely be escalated to include as manytechniques and points as the researchers consider necessary tounderstand a biological system better, but the systematic useof these technologies will likely help with the analysisand data interpretation, which is still a limiting step of high-throughput technologies.

Materials and methodsPlant growth

Arabidopsis thaliana (L. Heynh.) (Col-0, CS-60000) seeds wereobtained from the Arabidopsis Biological Resource Centre (TheOhioStateUniversity, Columbus,OH,USA). Seedswere surfacesterilised sequentially with 70% ethanol, 50% bleach, 0.05%Tween 20 and finally rinsedwith sterile water before being platedon MS media (Murashige and Skoog 1962) supplemented with3% sucrose. Seeds were vernalised for 3 days at 4�C before beingtransferred to an environment controlled chamber (Conviron) at22� 1�C, 65� 5%RH, and 160–200mmolm–2 s–1 light intensityon a short day photoperiod (10 h day,14 h night). After true leavesformed (12 days after sowing), the most vigorous seedlings weretransferred into 85� 73mm Quick Pot 15 RW trays (HerkuPlastKubern GmbH) containing Arabidopsis plant growing media(Lehle Seeds). Blue mesh (Kittrich Corporation) was placed ontop of the soil mixture to prevent algae growth and to improve theobject segmentation during image analysis (Fig. 1). Plants weregrown to maturity under these conditions until the start of thewater limitation treatments (16 days after germination).

Water treatments

Quick Pot 15 RW trays were filled with dry Arabidopsisplant growing media and their weight was recorded. Measuredamounts ofwaterwere added to the dry soil and allowed to absorbfor 1 h until soil reached full water saturation (100% full watercapacity, FC), after which trayweightswere once again recorded.After seedling establishment (~16 days after germination), soilwas allowed to reach four different levels of water saturation:100 (control), 50, 25 and 12.5%. The weight of the trays waschecked daily and water was uniformly added to all wells untilthe target weight was reached.

Soil water potential meter

Soil water potential (Y MPa) was measured with a soil waterpotential meter (WP4C, Decagon). All measurements wereconducted at the same time of the day. Soil samples weretaken from five wells within the Quick Pot 15 RW tray, thenplaced in a round sample cup (4 cm in diameter and 1 cm tall)and set it in the sample drawer. This instrument measuredwater potential by determining the RH of the air above ofthe sample in a closed chamber. Once the sample reachedequilibrium with the vapour in the sealed chamber, theinstrument calculated the RH by using the chilled mirrormethod. At the dew point, the meter measured both mirrorand sample temperature with a 0.001�C accuracy deliveringwater potential readings with accuracy within the –0.1MPa to–300MPa range.

Image acquisition

Images of Arabidopsis plants were captured every 2 days;from seedling establishment to a full vegetative growth(developmental stage 6.1 as defined by Boyes et al. 2001)using a LemnaTec Scanalyzer HTS system controlled usingthe LemnaControl software. This automated imaging system isequipped with a robotic arm that holds three high-resolutioncameras that allows to image the top view on the visible(VIS), fluorescence (FLUO) and near infrared (NIR) spectra.

Arabidopsis phenome and ionome during drought Functional Plant Biology 95

The system also has a barcode reader placed in a cabinet withoptimal light conditions (Fig. 1). Images were captured duringa specific time window each day (3.5 h after onset of daylight�30min). VIS images were taken using a piA2400–17 gc CCDcamera (Basler) with a resolution of 2454� 2056 pixels.A scA1600–14 gc CCD camera (Basler) equipped with aresolution of 1624� 1234 pixels was used for the acquisitionof the FLUO images. NIR images were taken with a GoldeyeGIGE P-008 SWIR camera (Allied Vision Technologies)equipped with a resolution of 320� 256 pixels and withspectral sensitivity between 900 and 1700 nm. TheLemnaControl software allowed for detailed configuration ofthe pots, trays, barcode positions and camera settings (zoomfocus, aperture and shutter speed).

Image analysis

Images of A. thaliana plants (4320 images = 30 biologicalreplicates� four treatments� 12 time points� three cameras)were analysed by using the LemnaGrid software. Analysis ofthe images acquired with the VIS camera was done as previouslydescribed (Arvidsson et al. 2011). Once all the rosette leaveswere identified, multiple phenotypic parameters were calculatedfor each plant including: projected leaf area (cm2), convex

hull area (cm2), caliper length (mm), and compactness (theratio of projected leaf area to convex hull area, a measure ofthe ‘bushiness’ of the plant). The LemnaGrid software was alsoused to colour classify the images acquired with the VIS cameraand, based on this colour classification, calculate the relative leafarea with normal green colour versus the area with detectableyellow colour. A similar colour classification approach wasfollowed for the grey-scale images acquired with the NIRcamera (high water corresponds to darker tones while lowwater corresponds to lighter colours). The complete greyscale(pixels range in value from 0 to 255, where 0 is black and 255 iswhite) was divided into three bins with centres equally distributedover the whole range (histogram with three bins), and thesoftware used to calculate the relative leaf area with low,medium and high water content. Images acquired with theFLUO camera were analysed using the LemnaGrid software,in which the complete red scale was divided into 4 bins withcentres equally distributed over the entire range and thesoftware was programmed to calculate the relative leaf areawith zero, low, medium and high chlorophyll fluorescence.All raw images, analysed images and calculations were savedinto a PostgreSQL database (LemnaDB) for storage. Quantitativedata obtained from the images was exported as CSV files andanalysed in Excel (Microsoft Corporation).

(a)

(b)

(c)

(d) (e) (f) (g)

Fig. 1. Flow chart for high-throughput phenotyping experiments during water limitation in wild type Arabidopsis thaliana. (a) Seed germination,(b) seedling establishment and water application, (c) time line and data points from germination to harvesting.Water limitation treatment starts at day 16and samples for elemental profilingwere taken at day 29, (d) phenotyping using the ScanalyzerHTS, (e) soil water quantification using thewater potentialmeter, (f) photosynthetic efficiency using the MultispeQ fluorometer and (g) elemental profiling using the ICP OES Perkin Elmer Optima 8000.

96 Functional Plant Biology L. M. Acosta-Gamboa et al.

Leaf number analysisTo determine leaf number, VIS images were exported from theLemnaDB database and processed by using a desktop computerrunning the Ubuntu system with an algorithm developed in-house (available upon request). Briefly, the analysis includedfive steps: histogram equalisation, plant object segmentation,mask refinement, Euclidean distance map calculation, andmarker based watershed transformation. Leaf number estimationwas stopped at 29 days due to the overlapping nature ofArabidopsis leaves, which severely decreases the accuracy ofthis automated pipeline, particularly when plants have more thanfour levels of overlapping leaves.

Photosynthetic efficiency

Photosynthetic efficiency was measured using a MultispeQinstrument developed by Dr David Kramer Laboratory(Michigan State University). This instrument combines thefunctionality of a handheld fluorometer, a chlorophyll meterand a benchtop spectrometer. Non-destructive measurementswere taken every other day from 15 plants chosen randomlyat the same time of day. Data was visualised in an Androidtablet (Samsung Galaxy Tab 4) and analysed in the PhotosynQweb portal (www.photosynq.org, accessed 2 September)(documentation for the MultispeQ including concepts andtutorials can also be found here). Photosynthetic efficiencywas calculated by measuring chlorophyll fluorescence atdifferent light conditions as described in Baker et al. (2007).In addition, the following readoutswere determined: photosyntheticefficiency of PSII (FII), linear electron flow (LEF), non-photochemical quenching (NPQ), photon flux (vH+), conductivityof thylakoid membrane to protons (gH+), electrochromic shift(ECSt), and greenness (SPAD) of the leaves.

Seed yield

Seed number was calculated as follows: the weight of 100 seedsper treatment was determined multiple times and the averagewas used to determine the total number of seeds per treatment.

Elemental analysis

The elemental composition of plant samples was determinedusing inductively coupled plasma optical emission spectrometry(ICP-OES) as previously described (Mendoza-Cózatl et al.2014). Briefly, leaves or seeds from plants exposed to differentwater capacity regimes (100, 50, 25 and 12.5%) were harvestedand dried for 6 days in an oven at 60�C, then 20–40mg of DWwere re-suspended with 1mL of nitric acid (HNO3, trace metalgrade) and boiled at 90�C in a water bath for a total of 30min(3� 10min intervals) to ensure a complete digestion. Sampleswere further diluted with ICP-grade water to 10mL and theconcentration of macronutrients (Ca, K, Mg and Na) andmicronutrients (Cu, Fe, Mn and Zn) were determined by ICP-OES (Optima ICP-OES 8000 Spectrometer, Perkin Elmer). Forthis analysis, 15 biological replicates for each condition wereused.

Statistical analysis

For statistical evaluation of the experiments, the PROC GLMprocedure of the software SAS 9.4 (SAS Institute) was used.

The variation in projected leaf area, caliper length, compactness,convex hull area, colour classification, relative chlorophyllfluorescence, relative water content, soil water potential, linearelectron flow, leaf number and seed yield were analysed usingANOVA, General linear model analysis (GLM) and subsequentpost-hoc analysis least significant difference (L.S.D.) (Fisher’sleast significant difference range test) at a=0.05. Data presentedare means� s.e.

Results

Optimisation of plant growth for high-throughputphenotyping and ionomic assays

The first step in our studies was to optimise the growthconditions, image acquisition, image analysis and ionomicanalyses of A. thaliana plants grown under various regimesof water availability. As illustrated in Fig. 1, seeds weregerminated on MS plates and incubated under environmentallycontrolled conditions. Once robust and healthy seedlings wereobtained (12 days after sowing), they were transferred intotrays containing soil at 100% full-water capacity (FC). Afterseedling establishment, trays were divided into four waterregimes: 100, 50, 25 and 12.5% FC and plants weremaintained at that level of water saturation until they reachedmaturity. In all cases, humidity, light and temperature werekept at optimal conditions. Throughout the experiment, visible(VIS, also referred to as RGB), fluorescence (FLUO) and nearinfrared (NIR) images of plants were acquired every other dayby using a multi-camera digital imaging system (ScanalyzerHTS, Fig. 2). The target weight of the trays was checkedmanually on a daily basis and soil moisture was measuredusing a water potential meter. Photosynthetic efficiency ofthe plants was determined with a MultispeQ instrumentonce leaves were large enough for these non-destructivemeasurements. Leaves and seed samples were taken 29 daysafter germination to assess changes in the elementalcomposition (ionome) of plants due to the different wateravailability regimes as indicated in Fig. 1. Finally, seedswere collected and counted to determine yield.

Water status and phenotypic variation

To ensure that plants grown under 50, 25 and 12.5% FC wereunder true water limitation conditions, compared with plantsgrown on 100% FC, soil moisture measurements wereperformed by using a water potential meter. Fig. 3a showsthat water limitation regimes ranged from –0.5MPa (50% FC)to –1.6MPa (12.5% FC) and these ranges have previously usedin moderate to severe water deficit experiments (Bhaskaraet al. 2015; Durand et al. 2016). We noted that despite thefact that biomass was clearly affected under the two mostsevere water limitation regimes (12.5 and 25% FC), no sign ofchlorosis or necrosis was observed in any of the appliedwater treatments (Fig. 3b). Fig. 4 shows representativeimages acquired with the VIS, FLUO, and NIR cameras ofthe HTTP system. Also included in this figure are the imagesobtained after background removal (image segmentation) andanalysis using the commercial software LemnaGrid.

Arabidopsis phenome and ionome during drought Functional Plant Biology 97

Water limitation effect on plant size, growth, anddevelopment

Abiotic stresses often have variable and gradual effectsdepending on the specific developmental stage of the plant.Therefore, it is highly valuable to employ non-destructiveautomated phenotyping methods to fully assess the effects ofwater limitation on the growth and development throughout theentire plant life cycle. Fig. 5 shows the impact of water limitationconditions on four readouts related to plant size and architecturecalculated after VIS images were analysed: (1) projected leafarea, (2) convex hull area, (3) caliper length (rosette diameter)and (4) compactness, which is a measurement of the ‘bushiness’of the rosette. As illustrated in Fig. 5a–d, Arabidopsis plantsgrown under moderate water limitation (50% FC) displayed nopenalty in the vegetative growth compared with 100%FC. However, plants grown at lower water saturation (12.5and 25% FC) display significant changes. We noted that mostof these changes were significant only after day 27, which is10 days after the water limitation treatment started (Fig. 5a–c).Colour classification results showed no presence of yellowcolour, indicative of loss of chlorophyll or chlorosis, in theimages from plants grown under the various water regimes(data not shown).

Notably, the analysis of top view images yielded nodifferences in projected leaf area between plants grown at 50%FC compared with control plants (100% FC); however, visualinspection of the images indicated that plants grown between50–12.5% FC had fewer but larger leaves compared withcontrol plants. To determine whether water limitation had animpact on leaf number, we developed an algorithm (available

upon request) to count the total number of leaves per plant afterprocessing the VIS images. The analysis showed that 29 daysafter germination, and 2 weeks of water limitation regimes,plants in all treatments had significantly fewer leaves comparedwith control plants (Fig. 6).

Water limitation effect on leaf relative chlorophyllfluorescence

Chlorophyll fluorescence is used as an indicator to determineif a plant is tolerant to a particular abiotic stress. The FLUOcamera of the HTPP platform and the associated softwaregave us the ability to measure relative chlorophyll fluorescenceof Arabidopsis plants growing under various water regimes.Our results showed significant differences across treatmentsand in all of the fluorescence levels including the ‘no’, ‘low’,‘medium’ (P < 0.0001) and ‘high’ fluorescence (P < 0.0003)categories (Fig. 7). Normal and severe water limitationconditions showed the highest area of zero fluorescence(labelled as ‘no’ relative chlorophyll fluorescence) and thesewere significantly different from the moderate soil waterlimitation regimes, including 25 and 50% FC. In addition,plants grown on 12.5% FC showed the lowest area with ‘low’chlorophyll fluorescence, followed by plants grown under25, 50 and 100% FC. In addition, plants grown in 25 and50% FC had a larger proportion ‘medium’ relative chlorophyllfluorescence area compared with controls, while plants grownat 25% FC showed the largest area with ‘high’ chlorophyllfluorescence, which is an indicator of leaf senescence(Fahlgren et al. 2015b).

Fig. 2. Automated high-throughput system for small plants. The high-resolution cameras used in this study are: from left to right, (a) VIS, (b) FLUOand (c) NIR cameras.

98 Functional Plant Biology L. M. Acosta-Gamboa et al.

Water limitation effect on leaf relative water content

The NIR camera of the HTPP platform and LemnaGrid softwaregave us the ability to quantify in planta water content anddistribution. As shown in Fig. 8, all levels of leaf relativewater content showed significant differences across treatments(P < 0.0001). For the ‘low’ relative water content category,plants grown at 12.5 and 25% FC were statistically differentthan those grown at 50% FC and control. Plants grown at 100%FC showed the lowest value for ‘low’ relativewater contentwhencompared with all other treatments. Plants grown under severewater limitation (12.5 and 25% FC) were statistically different in‘medium’ relative water content compared with those grown at50 and 100% FC. Similarly, plants grown at 100% FC, showedthe lowest value for ‘medium’ relative water content comparedwith the rest of the treatments. As an overall trend, the greatestamount of relative water content was obtained for plants grownunder 100% FC whereas plants grown under severe waterlimitation showed the lowest values of leaf relative water content.

Water limitation effect on photosynthetic efficiency

Water limitation has been shown to negatively affectphotosynthetic efficiency (PE) and chlorophyll fluorescence

and linear electron flow (LEF) are parameters that can be usedto assess the plant PE. For instance, the efficiency of PSII isdirectly related to the rate of LEF and inversely related to thechlorophyll fluorescence (Baker et al. 2007). As illustrated inFig. 9a, plants grown under 100% FC had the highest linearelectron flow values compared with plants grown at 50, 25 and12.5% FC. However, we did not detect significant differencesfor any of the rest of photosynthetic parameters measured withthe MultispeQ (FII, NPQ, vH+, gH+, ECSt; data not shown).

Seed yield is affected during water limitation

It is well known that the stress caused by limited wateravailability induces yield reduction in many plant species,and this reduction depends on the severity and duration ofthe stress (Farooq et al. 2009). To assess the impact of waterlimitation on seed yield in Arabidopsis, we let the plantsgrown in all water saturation conditions to reach maturity andcomplete their life cycle. Seeds were then collected and countedon a per plant basis. We found that even under moderate waterlimitation conditions (50% FC) there is a significant reductionin the seed yield (Fig. 9b). Plants grown under the most severewater limitation treatment (12.5% FC) displayed the lowestseed yield of all.

(a) 0.5

0.0

–0.5

–1.0

–1.5

–2.0

–2.519 28 35

Days after germination

ψ s

oil (

MP

a)

(b)

Fig. 3. Different water availability regimes impact the soil water potential and plantgrowth to different degrees. (a) Effect of water deficit on soil water potential (MPa) forArabidopsis thalianaplants under four differentwater regimes.Data aremeans� s.e. (n= 5).Asterisks represent significant differences (P< 0.0001 at CI 95%). (b) Phenotypic variationof wild type Arabidopsis thaliana in response to four different water regimes.

Arabidopsis phenome and ionome during drought Functional Plant Biology 99

(a)

(b)

100%

50%

25%

12.5%

VIS FLUO NIR

High

High

MedMed

Green

Yellow Low

Low

No

(c)

(d)

(e)

(f )

(g)

(h)

Fig. 4. Representative images obtained using the Scanalyzer HTS (VIS, FLUO and NIR cameras) of Arabidopsis plantsgrowing under four water regimes 100, 50, 25 and 12.5% water saturation for 29 days. (a, c, e) and (g) represent the originalimages obtained from the Scanalyzer HTS and (b, d, f, h) the analysed images after using the LemnaGrid software. Colourscale from FLUO images represented by black no, red low, blue medium and green high fluorescence. Colour scale fromNIR corresponds to blue high, green medium and orange low water content.

100 Functional Plant Biology L. M. Acosta-Gamboa et al.

Elemental analysis under water limitation regimes

Macro- and micronutrients are critical for several metabolicprocesses and also to maintain ion and osmotic balance duringwater limiting conditions. To assess the impact of waterlimitation on the leaf ionome, leaf samples from plants grownunder the different water regimes were harvested at day 29 aftergermination to determine their elemental composition. At day29, but not before, significant differences were found in leafarea (Fig. 5a) and leaf number (Fig. 6); therefore, we reasonedthis could be a relevant point for elemental analyses. Theconcentration of Ca, K, Na, Mg, Mn, Fe, Zn and Cu in leafsamples was determined by ICP-OES as previously reported(Mendoza-Cózatl et al. 2014). The concentration of all elementswas originally calculated as micrograms of element permilligram of DW and these data can be found in Fig. S1,available as Supplementary Material to this paper; however, tofacilitate the analysis and include macro and micronutrients inthe same graph, the element concentration was normalised tothe values of plants grown at 100% FC (Fig. 10). Significantchanges were found in all water regimes tested and more

specifically three trends were identified. First, Ca, Mg, Nashowed a similar pattern where a decrease in their concentrationwas found at 50% FC but, surprisingly, the concentration ofthese elements was similar to control plants or even higherwhen the stress became more severe (25 and 12.5% FC). Thisis particularly notable for Na (with a net increase to 27% respectto control plants grown at 12.5% FC) (Fig. 10a). Second, forthe micronutrients Cu and Mn, we did not observe any decreasein their concentration in any condition tested. On the contrary,the concentration of both elements gradually increased with thereduction on water saturation reaching 30% (Cu) and 40% (Mn)higher concentrations, at 12.5% FC, compared with controlplants (Fig. 10b). Third, we detected a significant decrease at50 and 25% FC in the Fe concentration; however, at 12.5%FC Fe levels were not different that control plants. It should benoted that these data were originally calculated on a dry weightbasis before being normalised and that plants grown at lowerwater saturation regimes had a severe reduction in biomassand this has to be considered when interpreting changes in theelemental composition throughout the different water regimes

(a) (b)

(c) (d)

80 120

0.814

12

10

8

6

4

2

0

0.6

0.4

0.2

0

100

80

60

40

20

0

60

40

20

14 16 18 20 22 25 27

Days after germination

Cal

iper

leng

th (

cm)

Pro

ject

ed le

af s

urfa

ce a

rea

(cm

2 )

Con

vex

hull

area

(cm

2 )C

ompa

ctne

ss

29 31 33 35 37 14 16 18 20 22 25 27 29 31 33 35 37

0

Fig. 5. Water limitation regimes affects the rosette growth at advance stages of plant development. (a) Total projected leaf area (cm2), (b) convex hull area (cm2),(c) caliper length (cm) (rosette diameter), and (d) compactness (total leaf area/convex hull area). Data are means� s.e. (n= 30). Different letters representsignificant differences between treatments (P < 0.0001 at CI 95%).

Arabidopsis phenome and ionome during drought Functional Plant Biology 101

(i.e. metal mg–1 DW vs metal plant–1). Finally, the concentrationof K and Zn remained constant in all conditions tested (Fig. S1).

Discussion

Quantification of plant phenotypes by HTPP approaches isessential to take full advantage of natural variation, -omictechnologies and molecular breeding approaches with theultimate goal of crop improvement. The images acquired withHTPP platforms are rich on information in the sense that theycontain and describe all the major characteristics of plants andthe dynamics of their response to changes in environmentalconditions, including biotic and abiotic stresses (Junker et al.2015). In this study we describe the methodology to obtainHTPP images on an optimised water limitation assay usingthe reference plant A. thaliana. High-throughput phenotypingallows entire experiments to be accurately quantified in anon-destructive manner, meaning that changes in plant growththroughout the life cycle can be documented temporally.Moreover, by identifying critical shifts on plant growth anddevelopment patterns, additional high-throughput technologies

(a) (b)

(c) (d)

1.2

1.0

0.8

0.6

0.4

0.2

0

1.2

1.0

0.8

0.6

0.4

0.2

0

1.2

1.0

0.8

0.6

0.4

0.2

0

1.2

1.0

0.8

0.6

0.4

0.2

0

Rel

ativ

e ch

loro

phyl

l flu

ores

cenc

e

100% water saturation 50% water saturation

25% water saturation 12.5% water saturation

14 16 18 20 22 25 27

Days after germination

29 31 33 35 37 14 16 18 20 22 25 27 29 31 33 35 37

Fig. 7. Impact of water limitation on chlorophyll fluorescence. Chlorophyll fluorescence expressed as relative to projected leaf area as an indicator of tissuehealth in plants grown under 12.5% (a), 25% (b), 50% (c) and 100% (d) soil saturation conditions. Data represent the means of 30 replicates. Significantdifferences were found between 12.5, 25, 50 and 100% water saturation (P< 0.0001 at CI 95%).

25

100%

50%

25%

12.5%

20

15

10

5

0

Leaf

num

ber

14 16 18 20 22 25 27 29

Days after germination

Fig. 6. Water availability reduces the leaf number in Arabidopsis. Changesin leaf number under four differentwater regimes 12.5, 25, 50 and 100%.Dataare means� s.e. (n= 15). Different letters represent significant differencesbetween treatments (P< 0.0001 at CI 95%).

102 Functional Plant Biology L. M. Acosta-Gamboa et al.

such as elemental profiling (or ionomics), fluxomics,transcriptomics or proteomics can be implemented to gainfurther insight into the molecular mechanisms that led to theoriginal phenotype. These analyses may also shed light onprocesses that may have a delayed impact at later stages ofdevelopmental, including yield or nutritional compositionof grains. Here, we explored the combination of twohigh-throughput methods, phenomics and ionomics, to begindocumenting the effects of water limitation at a resolutionthat would have been difficult to obtain using stand-alone ormanual independent measurements (Figs 1, 2).

Soil water potential is a good indicator to understand fromthe plant physiology perspective how difficult it would befor a plant to take up water from their surroundings (Jones2007). Our measurements show that our water limitationregimes, expressed as a percentage of water holdingcapacity, correspond to water potential values that have beenpreviously considered as moderate to severe water deficitregimes (–0.5MPa to –1.6MPa; Fig. 3) (Bhaskara et al.2015; Durand et al. 2016). Having standardised water deficitregimes is important to obtain comparable results even ifexperiments are performed in different laboratories anddeviation from these values may impact the ionomic andphenomic data acquired through the experiments. At the

physiological level, it is accepted that water deficiency maylead to growth arrest, cavitation in the xylem vessels (Zuffereyet al. 2011), stomatal closure, cessation of water and nutrienttransport and carbon re-allocation and eventual starvation, allof which threaten plant survival (McDowell et al. 2008). Inagreement with these observations, our data shows that reducedwater availability has a negative impact on plant growth anddevelopment (Figs 4, 5). Notably, the negative effects weretreatment-dependent and time-dependent. For instance,differences in the projected leaf area were only significant10 days after the water limitation regime started and thesedifferences became more obvious between treatments as theexperiment progressed over time (Fig. 5a).

An additional indicator of the plant metabolic status ischlorophyll fluorescence. In optimal conditions, plants exhibita basal level of chlorophyll fluorescence from absorbed lightand the majority of this energy gets directed towardsphotosynthesis. When plants experience abiotic stresses, suchas water limitation, the balance between photosyntheticefficiency and energy dissipation as heat and fluorescenceincreases, resulting in particularly high levels of fluorescence(Muller et al. 2001). Our results show that medium and highrelative chlorophyll fluorescence are largest in plants grown inthe two lowest saturation regimes (25–12.5% FC), indicating

(a) (b)

(c) (d)

100% water saturation 50% water saturation

25% water saturation 12.5% water saturation

14

1.00

0.95

0.20

0

1.00

0.95

0.20

0

1.00

0.90

0.20

0

1.00

0.90

0.20

016 18 20 22 25 27

Days after germination

Rel

ativ

e w

ater

con

tent

29 31 33 35 3714 16 18 20 22 25 27 29 31 33 35 37

Fig. 8. Relationship between soil water saturation and plant water content. Water content results expressed as relative to projected leaf area as an indicatorof tissue health in health in (a) 12.5% (b) 25%, (c) 50% and (d) 100% water saturation treated wild type Arabidopsis thaliana plants. Data represent themeans of 30 replicates. Significant differences were found between 12.5, 25, 50 and 100% water saturation (P< 0.0001 at CI 95%).

Arabidopsis phenome and ionome during drought Functional Plant Biology 103

that water limitation negatively affected the photosyntheticefficiency of the plants (Fig. 7).

Photosynthetic activity can also be determined by measuringthe flow of available electrons used to produce ATP andNADPH(Foyer et al. 2012). When water is limited, the stress causesa reduction in carbon fixation due to stomatal closure, whichwill cause an over-reduction of the electron transport chainintermediaries (Golding and Johnson 2003). In addition,biomass accumulation can also be stunted when plants aresubjected to water limitation. Schuppler et al. (1998) showedthat leaf area and the number of leaves is reduced in responseto water deficiency, which in turn will reduce water availabilityand ultimately leading to biomass and yield loss. Accordingly,our data show that plants grown under control conditions (100%FC) produced more leaves than the plants growing under waterdeficit (Fig. 6). In addition, control plants had a higher linearelectron flow compared with those growing under waterlimitation stress, indicating that water availability allowedplants to thrive and support a larger number of leaves whilemaintaining a high photosynthetic efficiency (Fig. 9a). Waterlimitation has been shown to negatively impact crop yield.

In agreement, our data show that even moderate waterlimitation conditions (50% FC), led to a significant reductionon seed yield (Fig. 9b), decreasing by more than 3-fold theamount of seeds produced per plant.

As predicted from previous studies, water limitation in thesoil led to lower water content in leaves (Bartels and Sunkar2005). By visual inspection, water deficit symptoms such aschlorosis or necrosis may have been missed throughout thewater limitation experiment (Fig. 3b). However, the NIRcamera (Fig. 8) was able to detect changes in water content inplanta, and plants grown under normal conditions showed agreater amount of available water. In addition, Fig. 4 showsthe pattern in which water is distributed in the plants, includingthe fact that leaves had lower water content at the edges. Thisoccurs because these areas have the least amount of moisture inthe leaf.

As for the elemental composition of plant tissues duringwater limiting stress, it has been generally assumed that waterlimitation restricts the bulk flow movement of nutrients through

(a)

(b)

Water regimes

LEF

(µm

ol e

– m–2

s–1

)N

umbe

r of

see

ds/p

lant

0.6

0.5

0.4

0.3

0.2

0.1

0

250

100% 50% 25% 12.5%

150

200

100

50

0

Fig. 9. Effect of water limitation on photosynthetic efficiency and seedyield. (a) Linear electron flow as an indicator of photosynthetic efficiencyduring water limiting conditions in wild type Arabidopsis thaliana plants.Data represents the mean of 15 replicates� s.e. Different letters representsignificant differences (P< 0.0001 at CI 95%). (b) Water limitation reducesseed yield in Arabidopsis. Seed yield of wild type Arabidopsis thalianaunder four different water regimes. Different letters indicate significantdifferences between treatments (P< 0.001 at CI 95%).

(a)

(b)

Water regimes

Ele

men

tal c

onte

nt r

elat

ive

to w

ell-w

ater

ed p

lant

s

140

120

100

80

60

40

0

140

160

120

100

80

60

40

0100% 50% 25% 12.5%

Fig. 10. Water limitation affects the elemental composition of plants.(a) Macronutrient and (b) micronutrient concentrations in plants grownunder different water availability regimes (100, 50, 25 and 12.5%). Datawere normalised to the value of 100% water regime. Data points representnormalised % average� s.e. for 15 biological replicates. The concentrationof elements at 100% water regime used for normalisation was: Ca, 35.6;Mg, 11.8;Na, 1.7;Mn, 0.068;Fe, 0.11 andCu, 0.0081 (mgelementmg–1DW).Significant differences are indicated: *, P< 0.01; **, P< 0.05 wheren= 15� s.e.).

104 Functional Plant Biology L. M. Acosta-Gamboa et al.

the soil matrix into the roots, thus reducing the overall uptakeand accumulation of minerals in plant tissues (Kramer and Boyer1995). This effect may be counteracted by the fact that waterlimitation also restrict dry mass accumulation (i.e. shootgrowth) and this could in turn compensate for the reduceduptake of minerals into the plant (Kramer and Boyer 1995).Our elemental profiling results however, show that the waterlimitation treatment affects the elemental composition of leavesin an elemental- and treatment-specific manner (Fig. 10). At themacronutrient level, Ca, Mg and Na seem to decrease, on aDW basis, at the 50% FC (Fig. 10a). However, the valueswere similar, or in the case of Na, even higher comparedwith control plants at lower water availability regimes. InArabidopsis, Na is considered as a non-essential element butits over-accumulation has been associated with a highercapacity to sustain water retention, thus improving the abilityof plants to cope when water availability is scarce (Gaxiolaet al. 2001). Some micronutrients, on the other hand, showparticularly defined patterns across water limitation regimes.For instance, Mn and Cu remained at level similar to controlplant values at moderate FC levels and even increased at severewater limitation regimes (Fig. 10b). Fe levels on the other hand,were decreased at 50 and 25% FC but unexpectedly, the Felevels in plants grown under 12.5% FC were no different fromthose of control plants. Fe is critical for photosynthesis but athigh concentration it becomes toxic due to the generation ofreactive oxygen species (ROS) through the Fenton reaction(Matros et al. 2015). During water limitation photosynthesisis impaired and the accumulation of reducing intermediariescould become even more detrimental in the presence of Fe,leading to a higher rate of ROS production. Therefore, keepingFe at low levels during water limitation may be a plant responseto prevent further oxidative stress. Mn and Cu, on the otherhand, are critical for the water-splitting and electron transferreactions in photosynthesis (Suorsa and Aro 2007) and sincethe photosynthetic efficiency of plants is impaired during waterlimitation, upregulation of Mn and Cu uptake and accumulationsystems may be an attempt to maintain or improve photosyntheticefficiency. Over-accumulation of Mn during drought hasbeen previously observed in several Arabidopsis accessions(Ghandilyan et al. 2009). Moreover, Mn is a critical transitionmetal for ROS detoxification by Mn superoxide dismutases(Mn SOD); therefore, the additional Mn in plants experiencingwater limitation may be funnelled towards Mn SOD for a moreefficient ROS detoxification. In agreement with this hypothesis,Mn SOD has been found to be upregulated in drought conditions(Alscher et al. 2002). As a whole, these results also reinforcethe notion that plant responses are strictly stress-dependent and,more specifically, that plant responses depend on the severity ofthe stress imposed. Therefore, we should be cautious aboutgeneralising the effects of water limitation on the elementalcomposition of plants and other phenotypes.

Conclusions

Our HTPP platform allowed us to assess a water limitationprotocol that now can be used as an experimental pipeline tostudy plant responses to water limitation in a controlled andreproducible environmental setting. Reproducibility is critical for

future experiments that may include additional high throughputtechnologies at specific time points and gain further insight intomolecular and physiological responses of plants undergoingbiotic and abiotic stresses. Our data also show that plantsgrown under mild stress (50% FC) grew at the same rate thancontrol plants; however, a clear penalty was observed inphotosynthetic efficiency and seed yield. In addition, waterlimitation stress led to a reduced shoot growth and overallbiomass, determined as a reduced leaf number. When waterlimitation was more severe, all the physiological and metabolicparameters measured were negatively impacted even further. Wealso demonstrated the utility of the HTPP platforms to generatequick, accurate and large non-biased datasets that can be usedto determine the overall plant fitness and performance. Accuratequantification of subtle and novel physiological responses toabiotic stresses allow researchers to document time-dependentchanges that otherwise would be undetectable to the naked eyein a non-destructive and high-throughput manner. In addition,we also demonstrate the utility of HTPP platforms to identifycritical points, were additional high throughput technologiessuch as elemental profiling, can be implemented to get a betterunderstanding of the plant physiological responses to abioticstresses such as water limitation. Linking high throughputtechnologies in a systematic fashion will facilitate data integrationand interpretation, thus offering a better opportunity to unravelthe still largely unknown mechanisms that plants use to thriveon challenging environments. Understanding these mechanismsis critical to developing crops capable of sustaining growth andyield in times when global climate change poses a serious threatto food security worldwide.

Acknowledgements

This work was supported by the Plant Imaging Consortium (http://plantimaging.cast.uark.edu/) funded by the National Science FoundationAward Numbers IIA-1430427 and IIA-1430428. The HTPP platform wasacquired with funds from the Arkansas Centre for Plant Powered Production(http://www.plantpoweredproduction.com/) through the RII ArkansasASSET Initiative (Arkansas EPSCoR) by NSF grant # EPS –0701890.

References

Alscher RG, Erturk N, Heath LS (2002) Role of superoxide dismutases(SODs) in controlling oxidative stress in plants. Journal of ExperimentalBotany 53, 1331–1341. doi:10.1093/jexbot/53.372.1331

Arvidsson S, Pérez-Rodríguez P, Mueller-Roeber B (2011) A growthphenotyping pipeline for Arabidopsis thaliana integrating imageanalysis and rosette area modeling for robust quantification of genotypeeffects. New Phytologist 191, 895–907. doi:10.1111/j.1469-8137.2011.03756.x

Baker N, Harbinson J, Kramer D (2007) Determining the limitationsand regulation of photosynthetic energy transduction in leaves. Plant,Cell & Environment 30, 1107–1125. doi:10.1111/j.1365-3040.2007.01680.x

Bartels D, Sunkar R (2005) Drought and salt tolerance in plants. CriticalReviews in Plant Sciences 24, 23–58. doi:10.1080/07352680590910410

Bhaskara GB, Yang TH, Verslues PE (2015) Dynamic proline metabolism:importance and regulation in water limited environments. Frontiers inPlant Science 6, 484. doi:10.3389/fpls.2015.00484

Boyer J (1970) Differing sensitivity of photosynthesis to low leaf waterpotentials in corn and soybean. Plant Physiology 46, 236–239.doi:10.1104/pp.46.2.236

Arabidopsis phenome and ionome during drought Functional Plant Biology 105

Boyes DC, Zayed AM, Ascenzi R, McCaskill AJ, Hoffman NE, Davis KR,Görlach J (2001) Growth stage-based phenotypic analysis ofArabidopsis: a model for high throughput functional genomics inplants. The Plant Cell 13, 1499–1510. doi:10.1105/tpc.13.7.1499

Chaves M, Maroco J, Pereira J (2003) Understanding plant responses todrought – from genes to the whole plant. Functional Plant Biology 30,239–264. doi:10.1071/FP02076

Claeys H, Inzé D (2013) The agony of choice: how plants balance growthand survival under water-limiting condition. Plant Physiology 162,1768–1779. doi:10.1104/pp.113.220921

Durand M, Porcheron B, Hennion N, Maurousset L, Lemoine R, Pourtau N(2016) Water deficit enhances C export to the roots in Arabidopsisthaliana plants with contribution of sucrose transporters in both shootand roots. Plant Physiology 170, 1460–1479.

Eide DJ, Clark S, Nair TM, Gehl M, Gribskov M, Guerinot ML, Harper J(2005) Characterization of the yeast ionome: a genome-wide analysisof nutrient mineral and trace element homeostasis in Saccharomycescerevisiae. Genome Biology 6, R77. doi:10.1186/gb-2005-6-9-r77

Fahlgren N, Gehan M, Baxter I (2015a) Lights, camera, action: high-throughput plant phenotyping is ready for a close-up. Current Opinionin Plant Biology 24, 93–99. doi:10.1016/j.pbi.2015.02.006

Fahlgren N, Feldman M, Gehan MA, Wilson MS, Shyu C, Bryant DW,Hill ST, McEntee CJ, Warnasooriya SN, Kumar I, Ficor T, TurnipseedS, Gilbert KB, Brutnell TP, Carrington JC, Mockler TC, Baxter I(2015b) A versatile phenotyping system and analytics platform revealsdiverse temporal responses to water availability in Setaria. MolecularPlant 8, 1520–1535. doi:10.1016/j.molp.2015.06.005

Farooq M, Wahid A, Kobayashi N, Fujita D, Basra SMA (2009) Plantdrought stress: effects, mechanisms and management. Agronomy forSustainable Development 29, 185–212. doi:10.1051/agro:2008021

Farquhar G, Sharkey T (1982) Stomatal conductance and photosynthesis.Annual Review of Plant Physiology and Plant Molecular Biology 33,317–345. doi:10.1146/annurev.pp.33.060182.001533

Foyer C, Neukermans J, Queval G, Noctor G, Harbinson J (2012)Photosynthetic control of electron transport and the regulation ofgene expression. Journal of Experimental Botany 63, 1637–1661.doi:10.1093/jxb/ers013

Gaxiola RA, Li J, Undurraga S, Dang LM, Allen GJ, Alper SL, Fink GR(2001) Drought- and salt-tolerant plants result from overexpression ofthe AVP1H+-pump. Proceedings of the National Academy of Sciencesof the United States of America 98, 11444–11449. doi:10.1073/pnas.191389398

Ghandilyan A, Barboza L, Tisné S, Granier C, Reymond M, Koornneef M,Schat H, Aarts MG (2009) Genetic analysis identifies quantitative traitloci controlling rosette mineral concentrations in Arabidopsis thalianaunder drought.NewPhytologist 184, 180–192. doi:10.1111/j.1469-8137.2009.02953.x

Golding A, Johnson G (2003) Down-regulation of linear and activationof cyclic electron transport during drought. Planta 218, 107–114.doi:10.1007/s00425-003-1077-5

Jones HG (2007) Monitoring plant and soil water status: established andnovel methods revisited and their relevance to studies of droughttolerance. Journal of Experimental Botany 58, 119–130. doi:10.1093/jxb/erl118

Junker A, Muraya MM, Weigelt-Fischer K, Arana-Ceballos F, Klukas C,Melchinger AE, Meyer RC, Riewe D, Altmann T (2015) Optimizingexperimental procedures for quantitative evaluation of crop plantperformance in high throughput phenotyping systems. Frontiers inPlant Science 5, 1–21. doi:10.3389/fpls.2014.00770

Kramer PJ, Boyer JS (1995) ‘Water relations of plants and soils.’ (AcademicPress: New York)

Lahner B, Gong J, Mahmoudian M, Smith EL, Abid KB, Rogers EE,Guerinot ML, Harper JF, Ward JM, McIntyre L, Schroeder JI, Salt DE(2003) Genomic scale profiling of nutrient and trace elements inArabidopsis thaliana. Nature Biotechnology 21, 1215–1221. doi:10.1038/nbt865

Matros A, Peshev D, Peukert M, Mock HP, Van den Ende W (2015) Sugarsas hydroxyl radical scavengers: proof-of-concept by studying the fate ofsucralose in Arabidopsis. The Plant Journal 82, 822–839. doi:10.1111/tpj.12853

McDowell N, Pockman WT, Allen CD, Breshears DD, Cobb N, Kolb T,Plaur J, Sperry J, West A, Willians DG, Yepez EA (2008) Mechanismsof plant survival and mortality during drought: why do some plantssurvive while others succumb to drought? New Phytologist 178,719–739. doi:10.1111/j.1469-8137.2008.02436.x

Mendoza-Cózatl DG, Xie Q, Akmakjian GZ, Jobe TO, Patel A, Stacey MG,Song L, Demoin DW, Jurisson SS, Stacey G, Schroeder JI (2014) OPT3is a component of the iron-signaling network between leaves and rootsand misregulation of OPT3 leads to an over-accumulation of cadmiumin seeds. Molecular Plant 7, 1455–1469. doi:10.1093/mp/ssu067

Muller P, Li X, Niyogi K (2001) Non-photochemical quenching. A responseto excess light energy. Plant Physiology 125, 1558–1566. doi:10.1104/pp.125.4.1558

Murashige T, Skoog F (1962) A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiologia Plantarum 15, 473–497.doi:10.1111/j.1399-3054.1962.tb08052.x

Schuppler U, He PH, John PCL, Munns R (1998) Effect of water stresson cell division and cell-division- cycle 2-like cell-cycle kinase activityin wheat leaves. Plant Physiology 117, 667–678. doi:10.1104/pp.117.2.667

Suorsa M, Aro EM (2007) Expression, assembly and auxiliary functionsof photosystem II oxygen-evolving proteins in higher plants.Photosynthesis Research 93, 89–100. doi:10.1007/s11120-007-9154-4

Tardieu F, Granier C, Muller B (1999) Modelling leaf expansion ina fluctuating environment: are changes in specific leaf area aconsequence of changes in expansion rate? New Phytologist 143,33–43. doi:10.1046/j.1469-8137.1999.00433.x

Wang M, Zheng Q, Shen Q, Guo S (2013) The critical role of potassium inplant stress response. International Journal of Molecular Sciences 14,7370–7390. doi:10.3390/ijms14047370

Yang X, Cushman JC, Borland AM, Edwards EJ, Wullschleger SD, TuskanGA, Owen NA, Griffiths H, Smith JA, De Paoli HC, Weston DJ,Cottingham R, Hartwell J, Davis SC, Silvera K, Ming R, Schlauch K,Abraham P, Stewart JR, Guo HB, Albion R, Ha J, Lim SD, Wone BW,Yim WC, Garcia T, Mayer JA, Petereit J, Nair SS, Casey E, Hettich RL,Ceusters J, Ranjan P, Palla KJ, Yin H, Reyes-García C, Andrade JL,Freschi L,Beltrán JD,DeverLV,Boxall SF,Waller J, Davies J,BupphadaP, KaduN,WinterK, SageRF,Aguilar CN, Schmutz J, Jenkins J, HoltumJA (2015) A roadmap for research on crassulacean acid metabolism(CAM) to enhance sustainable food and bioenergy production in ahotter, drier world. New Phytologist 207, 491–504. doi:10.1111/nph.13393

Zufferey V, Cochard H, Ameglio T, Srping JL, Viret O (2011) Diurnalcycles of embolism formation and repair in petioles of grapevine (Vitisvinifera cv. Chasselas). Journal of Experimental Botany 62, 3885–3894.doi:10.1093/jxb/err081

106 Functional Plant Biology L. M. Acosta-Gamboa et al.

www.publish.csiro.au/journals/fpb