An assessment of the biotechnological use of hemoglobin modulation in cereals

Physiologia Plantarum 150: 593–603. 2014 © 2013 Scandinavian Plant Physiology Society, ISSN 0031-9317

An assessment of the biotechnological use of hemoglobinmodulation in cerealsKim H. Hebelstrupa∗, Jay K. Shahb, Catherine Simpsonc, Jan K. Schjoerringd, Julien Mandone, SimonaM. Cristescue, Frans J. M. Harrene, Michael W. Christiansena, Luis A. J. Murc and Abir U. Igamberdievb

aDepartment of Molecular Biology and Genetics, Aarhus University, Flakkebjerg, DenmarkbDepartment of Biology, Memorial University of Newfoundland, St. John’s, NL, CanadacInstitute of Environmental and Rural Science, Aberystwyth University, Aberystwyth, UKdPlant and Soil Science Section, Department of Plant and Environmental Sciences, University of Copenhagen, Frederiksberg, DenmarkeLife Science Trace Gas Facility, Molecular and Laser Physics, Institute for Molecules and Materials, Radboud University, Nijmegen, The Netherlands

Correspondence*Corresponding author,e-mail: [email protected]

Received 11 July 2013;revised 30 September 2013

doi:10.1111/ppl.12115

Non-symbiotic hemoglobin (nsHb) genes are ubiquitous in plants, but theirbiological functions have mostly been studied in model plant species ratherthan in crops. nsHb influences cell signaling and metabolism by modulatingthe levels of nitric oxide (NO). Class 1 nsHb is upregulated under hypoxia andis involved in various biotic and abiotic stress responses. Ectopic overexpres-sion of nsHb in Arabidopsis thaliana accelerates development, whilst targetedoverexpression in seeds can increase seed yield. Such observations suggestthat manipulating nsHb could be a valid biotechnological target. We studiedthe effects of overexpression of class 1 nsHb in the monocotyledonous cropplant barley (Hordeum vulgare cv. Golden Promise). nsHb was shown tobe involved in NO metabolism in barley, as ectopic overexpression reducedthe amount of NO released during hypoxia. Further, as in Arabidopsis, nsHboverexpression compromised basal resistance toward pathogens in barley.However, unlike Arabidopsis, nsHb ectopic overexpression delayed growthand development in barley, and seed specific overexpression reduced seedyield. Thus, nsHb overexpression in barley does not seem to be an efficientstrategy for increasing yield in cereal crops. These findings highlight the neces-sity for using actual crop plants rather than laboratory model plants whenassessing the effects of biotechnological approaches to crop improvement.

Introduction

Non-symbiotic hemoglobin (nsHb) genes are ubiquitousin plants. Hemoglobin in plants was originally identifiedin symbiotic root nodules as early as 1939 and wasinitially believed to be specific for plants with this typeof organ (Kubo 1939). However, in the late 20th centuryhemoglobin genes were cloned from non-nodulatingplant species such as Trema tomentosa (Bogusz et al.1988) and barley (Taylor et al. 1994). Subsequent

Abbreviations – ICP-MS, inductively coupled plasma-mass spectrometry; nsHb, non-symbiotic hemoglobin; WT, wildtype.

phylogenetic analyses have suggested that hemoglobingenes are likely to exist in all plant species (Hebelstrupet al. 2007). Plant nsHbs constitute a diverse group ofhemeproteins belonging to three evolutionarily differentclasses: class 1, class 2 and class 3 (Hebelstrup et al.2007). Class 1 hemoglobins have an extremely highaffinity for oxygen, with a Kd on the order of 2 nM(Hargrove et al. 2000, Smagghe et al. 2009). A role forclass 1 nsHb in scavenging of nitric oxide (NO), bya dioxygenase mechanism, has been demonstrated in

Physiol. Plant. 150, 2014 593

different plant species and plant tissue systems, suchas alfalfa (Medicago sativa) roots (Dordas et al. 2003,Igamberdiev et al. 2004), maize (Zea mays) suspensioncell culture (Dordas et al. 2004) and Arabidopsis thaliana(Perazzolli et al. 2004, Hebelstrup et al. 2006). It hasbeen shown previously that this scavenging of NO byhemoglobin is facilitated by a monodehydroascorbatereductase which mediates ascorbate reduction ofmethemoglobin, formed when NO is oxidized to NO3

−

by oxy-hemoglobin (Igamberdiev et al. 2005). In otherstudies, it was demonstrated that overexpression orsilencing/knock-out of class 1 nsHb modulates severalbiological functions of NO.

Plant tissues produce a significant amount of NOduring responses to stress. During hypoxia, NO isoxidized to NO3

− by oxy-hemoglobin to contributeto increased homeostasis of cellular redox and energypotential during hypoxia in the biochemical cycleknown as the hemoglobin/NO cycle (Igamberdiev et al.2006). Simultaneously, the oxidation of NO can preventloss of nitrogen content from the plant by NO emission,which can be as high as 0.2 μmol N lost g−1 fresh weight(24 h)−1 at hypoxic concentrations below 0.1% O2

(Hebelstrup et al. 2012). Given this, it is relevant that theclass 1 nsHb promoter in maize is induced by submer-gence, high salt and drought stress (Zhao et al. 2008).Also, when the maize nsHb class 1 is overexpressed intobacco there is increased survival under submergenceand increased tolerance to salt stress. The expressionpattern of class 1 nsHb in maize roots also suggeststhat it is part of a control system, which regulates theconcentration of NO in response to changes in theNO3

− concentration (Trevisan et al. 2011). Overex-pression of nsHb in tomato slightly increases stomatalconductance and transpiration under waterlogging ofroots (Shi et al. 2012). NO is also an important signalingmolecule generated in response to both necro- andhemi(biotrophic) pathogens; and expression of class 1and to a lesser extent class 2 hemoglobins is able tomodulate NO levels in response to infection of suchpathogens and thereby susceptibility toward those (Muret al. 2012).

Considering NO/Hb control at a gross developmentallevel, expression of a hemoglobin gene from thebacterium Vitreoscilla (VHb) in Nicotiana tabacum wasfound to enhance growth and dryweight (Holmberg et al.1997). However, later studies were unable to confirmthis effect (Frey et al. 2004). Class 1 and 2 hemoglobinsare expressed in a spatial-specific pattern, suggestingthat nsHbs play roles in the control of local NO levels(Heckmann et al. 2006), and indeed overexpression ofeither class 1 or 2 nsHb in Arabidopsis is able tocontrol timing of flowering through local control of

NO in shoot meristems (Hebelstrup and Jensen 2008).A general comparison of the effect of endogenous NOand nsHb expression supports the hypothesis that planthemoglobins are important modulators of localizedNO effects during various developmental processes(Hebelstrup et al. 2013).

The biological function of class 2 and 3 plant nsHbsis not as well described as that of the class 1 type.Overexpression of class 2 nsHb reduces cellular NOlevels (Hebelstrup et al. 2006) and NO gas emission(Hebelstrup et al. 2012), demonstrating NO scavengingability. One developmental role for class 2 nsHb maybe in seed development. The endosperm of seeds,including that of the grains of barley, is hypoxicduring development (Borisjuk and Rolletschek 2009) soexpression of nsHb in seeds/grain could increase energymaintenance and oxidative potential. Indeed it has beenshown that overexpression of class 2 nsHb in Arabidopsisincreases seed yield and changes the lipid types inthe seeds toward a higher content of unsaturated lipids(Vigeolas et al. 2011). However, it was not clear whetherthis effect was also a result of increased NO turnover, ordue to a novel mechanism of plant hemoglobins. Whenclass 1 nsHb was overexpressed in Arabidopsis seeds,they were more pre-adapted to hypoxic stress and maturetransgenic seeds had a higher weight (Thiel et al. 2011).

It should be noted that effects of hemoglobin overex-pression have only been documented in dicotyledonouslaboratory model plants and not in any cereal cropplants. Given that overexpression of nsHb in plants canreduce time to flowering, increase stress tolerance, seedyield and possibly also biomass we suggest that thiscould be utilized as a biotechnological tool to improvecereal productivity. We here characterize the effect ofoverexpression of barley class 1 hemoglobin in barleyplants to assess the value of nsHb overexpression ina major cereal crop. Two approaches were used, onewhere the hemoglobin gene was expressed ectopicallyby a maize ubiquitin-2 promoter and one where thegene was expressed specifically in grain endosperm bya barley hordein-D promoter. We found that overex-pression of class 1 nsHb in barley delays development,decreases biomass, grain size and yield and results in apartial breakdown of basal resistance to fungal mildew(Blumeria graminis). Barley type 1 nsHb is an effectivescavenger of endogenous NO. Thus, the differingresponses of nsHb overexpression, when compared toprevious studies, reflect species-specific differences inthe effects of NO, rather than a difference in the abilityto scavenge NO by nsHb. This stresses the necessity forusing actual crop plants rather than laboratory modelplants when assessing biotechnological uses of genetechnology.

594 Physiol. Plant. 150, 2014

Materials and methods

Plant transformation

We generated two different types of transgenic barleylines (Hordeum vulgare var. Golden Promise) expressingcDNA of the barley hemoglobin gene HvHb1 (accessionnumber: U94968), donated as a plasmid (CA2α) fromDr Robert D. Hill, University of Manitoba (Duff et al.1998). The cDNA was cloned into either of the vectorspUCEHordD::USER::NOS or pUCEUbi::USER::NOS by USERcloning (Hebelstrup et al. 2010). Independent transgenicbarley lines (UHb) expressing HvHb1 under the maizeubiquitin-2 promoter were generated by a modifiedprotocol of Agrobacterium-mediated transformation,which is described in Carciofi et al. (2011). Thegeneration of independent transgenic lines (HHb) withoverexpression of HvHb1 by the barley hordein-Dpromoter is described in Hebelstrup et al. (2010).

Protein extraction, western blotting and qRT-PCR

Protein extraction and western blotting were conductedas described in Shah et al. (2013), with the exception thata polyclonal antibody against barley class 1 hemoglobinwas used (Dordas et al. 2003; donated by Dr Robert D.Hill, University of Manitoba). For RNA extraction, shootswere sampled in 15 ml falcon tubes, each having two1/4 inch ceramic beads. Roots were sampled in 20 mlplastic containers containing four 4 mm steel beads. Thesamples were immediately flash frozen by immersingthe tubes in liquid nitrogen. The tubes containing leaveswere vortexed 3 × 20 s. After each 20 s run, the tubeswere immersed in liquid nitrogen for 1 min. The rootswere disrupted in the Geno/Grinder 2000 (SPEX Cer-tiPrep, Metuchen, NJ) homogenizer at 1200 strokes permin over a period of 40 s for three times. After each run,the tubes were immersed in liquid nitrogen for 1 min.Total RNA from the samples was isolated using Fast RNAgreen kit (MP Biochemicals, Cedex, France) accordingto the manufacturer’s instructions. RNA concentrationand purity were checked on a NanoDrop 1000 Spec-trophotometer (Thermo Scientific, Hvidovre, Denmark).First-strand cDNA synthesis was performed on 1000 μgof RNA and a nonamer, random oligonucleotide primer(2.5 μM) by incubation at 65◦C for 5 min followed by10 min at room temperature in a volume of 18.4 μl.SuperscriptII (200 U; Invitrogen, Naerum, Denmark),RNAsin (40 U; Promega, Roskilde, Denmark), 1× FSBuffer (Invitrogen), 10 μM dithiothreitol (DTT) and 2 mMdNTPs (GE Healthcare, Broendby, Denmark) were addedto make a final volume of 30 μl which was incubated for1 h at 42◦C and 10 min at 70◦C, followed by the addi-tion of 70 μl of water. qRT-PCR was performed using

an ABI Prism 7900HT Sequence Detection System withPower SYBR Green PCR master mix (Applied Biosystems,Naerum, Denmark). The primers used for assessment ofHvHb1 and the housekeeping gene GADPH expressionare similar to those described in Hebelstrup et al. (2010).

Plant growth

Fifty wild type Golden Promise and UHb lines weregrown in soil each in a small square pot (6 × 6 × 8 cm)in a growth chamber under long day length (16 h dayat 15◦C and 8 h dark at 12◦C). The soil was Pindstrupsubstrate (Pindstrup Mosebrug A/S, Ryomgard, Den-mark). The pots were watered from the bottom withwater containing micro and macro nutrients (1 g l−1

Yara Liva Calcinit™, Yara Norge A/S, Oslo, Norway;1 g l−1 Pioner NPK Makro 14-3-23 + Mg, Brøste A/S,Lyngby, Denmark; 0.13 ml l−1 Pioner Mikro med jern,Azelis, Kgs. Lyngby, Denmark). At 6 weeks aftergermination, the plants were transferred to bigger roundpots (diameter: 15 cm; height: 13 cm) in a greenhousewith artificial lightning maintaining the 16/8 h long daylength. The growth stages of the plants were recordedfor 60 days according to the Zadoks decimal scalefor cereals (Zadoks et al. 1974). The Zadoks scale isbased on a 2-digit code which describes developmentalstages, where the first digit represents the principalgrowth stages: 1 – germination; 2 – seedling growth;3 – tillering. The second digit represents the number ofunfolded leaves during seedling growth or the number oftillers (in addition to the main shoot) during the tilleringphase. When plants were harvested for determination ofbiomass, roots were cleaned using a water spray. Plantswere harvested at the age of 6 months. At this timeall spikes were dry. The total yield (g grain per plant),number of spikes per plants, number of grains per spikeand average grain weight were determined.

NO emission analysis

NO emissions were measured online using a laser-baseddetector developed at Trace Gas Facility, Nijmegen, TheNetherlands by the method described in Cristescu et al.(2008) and Clarke et al. (2009). Leaves or roots from dif-ferent genotypes were measured in parallel experimentsplaced in dark sealed containers and flushed with gasmixtures at 2 l h−1. Normoxia was obtained by flushingwith atmospheric air. Anoxia was obtained by flushingwith N2 (>99.9995%). Hypoxia was obtained by flush-ing with mixtures of N2 and O2 (0.1, 0.3 and 1.0% O2).

Pathogen infiltration and analysis

Barley plants were infiltrated with spores of oat-adapted mildew (B. graminis f. sp. avenae) as described

Physiol. Plant. 150, 2014 595

previously (Prats et al. 2006). Three centimeter segmentswere cut from the infected portion of leaves collected 40and 88 h after infection and prepared for UV microscopy.Individual leaf samples were placed in glass universalbottles containing 5 ml 1 M KOH and autoclaved for2 min at 121◦C. The samples were then rinsed threetimes in deionized water (dH2O) and mounted onto76 × 26 mm glass microscopy slides. In order to preventleaves from shriveling up when placed on the slide,water was left in the bottle, slides were placed on apaper towel and the leaves coaxed out of the bottle ontothe slide. Excess water was removed with another papertowel before three to four drops of aniline blue stain wereadded. A cover slip was placed over the samples and theleaves examined under UV light at 200× magnification.The responses to the appearance of the AGT from 50germinated spores were scored for each sample. Picturesof the infection sites were taken.

Iron concentration analysis

Mineral element concentrations were determined usinginductively coupled plasma-mass spectrometry, ICP-MS (Agilent 7500c; Agilent Technologies, Manchester,UK). Digestion of samples was performed accordingto a micro-scaled procedure using microwave diges-tion (MULTIWAVE 3000, software version 1.24, AntonPaar GmbH, Graz, Austria; Hansen et al. 2009). Foreach digestion batch, the accuracy and precision of theICP-MS measurements were estimated using certified ref-erence material (durum wheat flour, NIST8436; NationalInstitute of Standards and Technology, Gaithersburg,MD). Data was accepted if the accuracy was above90% of certified reference values (see further details inHansen et al. 2012).

Results

The generation of transgenic barley plants

We developed two different types of barley plants withoverexpression of the class 1 type hemoglobin (HvHb1)from barley (Taylor et al. 1994). One type (UHb) hadan ectopic overexpression driven by the ubiquitin-2promoter from maize. The other type (HHb) had aspecific overexpression in developing endosperm ofgrains driven by the hordein-D promoter from barley.Twenty-one independent T0 plants of the UHb type weregenerated by Agrobacterium-mediated transformation.Eighteen T0 plants survived to make fertile T1 grains.The relative expression levels of HvHb1 in leaves weremeasured in the 18 T0 plants in comparison with theexpression level of wild-type (WT) plants (Fig. 1A).Three lines (UHb7, 8 and 9) showed expression levels of

Fig. 1. Gene and protein expression in UHb and HHb plants. (A)Expression of HvHb1 measured by qRT-PCR in leaves of 18 independentUHb T0 plants and a WT plant (Golden Promise). Bars indicate standarddeviations of three replicates. The lines 17–19 were not included. (B)Western blotting of HvHb1 in T0 UHb plants (lines 7–9) and WT leaves.(C) Western blotting of HvHb1 in T2 grains from T1 plants of HHb plants(lines 2, 6, 14 and 16) and WT. (D) Representative WT and UHb plantsat 60 days after germination.

HvHb1 that were 50- to 70-fold greater than those of theWT plants. The overexpression in these three lines wasfurther confirmed by western blotting (Fig. 1B). We havepreviously described the development of HHb plantswith an endosperm-specific overexpression of HvHb1that was up to 104 times higher than that of WT plants(Hebelstrup et al. 2010). Overexpression of HvHb1 inthe endosperm of four independent lines of HHb (HHb2,6, 14 and 16) was also confirmed by western blotting ofprotein purified from the endosperm (Fig. 1C).

Overexpression of nsHb delays plant developmentand reduces yield

The plants of T1 generation of the lines UHb7, 8 and 9all showed a delayed growth and flowering phenotypeand reduced yield (Fig. 1D). The phenotype was also

596 Physiol. Plant. 150, 2014

Fig. 2. Development of UHb plants in comparison with control WT (Golden Promise). (A–E) Developmental stage was scored on the Zadoks scale(Zadoks et al. 1974) at various days after germination. (F) Plant height during development. (G) Weight of shoots and (H) roots during development.Differences between UHb and WT that are significant according to Student’s t-test are marked with either *P < 0.05, **P < 0.01 or ***P < 0.001.Bars indicate standard errors.

observed in the next generation (T2) of the threeindependent lines and characterized quantitatively byscoring developmental stages according to the Zadoksscale for the line UHb8. The Zadoks decimal scale isused commonly for scoring growth stages of cereals(Zadoks et al. 1974). The delayed development (Fig. 1D)was evident in all of the three independent overexpres-sion lines UHb7, 8 and 9, where overexpression had

been confirmed. However, due to logistical limitations,only one line (UHb8) was carried on for quantitativeassessment in the following generation. UHb plantswere delayed in comparison with WT plants throughouttheir development (Fig. 2A–E). Similarly, the height(Fig. 2F) and fresh weight (Fig. 2G, H) of UHb plantswere lower than that of the WT plants throughoutdevelopment. At maturity, 6 months after sowing, the

Physiol. Plant. 150, 2014 597

Fig. 3. Yield (g grain per plant) and yield components (numberof spikes per plant, number of grains per spike and grain weight)of UHb plants in comparison with WT Control (Golden Promise).Significant differences according to Student’s t-test are marked witheither *P < 0.05, ***P < 0.001 or n.s. (not significant).

yield from the UHb plants was significantly lower thanthat of the WT plants (Fig. 3). The reduced yield wasdue to both a lower number of spikes per plant, grainsper spike and weight of the individual grains.

Overexpression (UHb) plants display increased NOscavenging

Class 1 hemoglobin gene expression is known to begenerally upregulated in flooded and/or hypoxic roots(Igamberdiev et al. 2005, Hebelstrup et al. 2012) wherethe Hb/NO cycle contributes to energy homeostasis inhypoxic root tissue. However, the Hb/NO cycle is alsoan important modulator of local NO levels controllingdevelopment and morphogenesis (Hebelstrup et al.2013). We have previously reported that hemoglobinoverexpression can significantly reduce NO emissionfrom leaves under severe hypoxia (Hebelstrup et al.2012). We compared NO emissions from roots andleaves of UHb and WT barley plants in air atdifferent concentrations of O2 (Fig. 4). UHb plantshad significantly lower NO emission levels than WTplants both from roots and leaves at hypoxia (0.1% O2),whereas NO emission under anoxia (N2) was not lowerin UHb than in WT plants. However, when leaves wererestored to normoxia after 65 min exposure to anoxia,NO emission remained at a high level in WT plantswhereas it fell to nearly zero in UHb plants after less

than 40 min. These data were consistent with UHb plantshaving higher NO dioxygenase activity than WT plants.

Overexpression (UHb) plants display a partial lossof basal resistance

It is well documented that overexpression of planthemoglobins reduces the NO bursts linked to pathogenattack and resistance is jeopardized (Mur et al. 2012).Thus, we tested the susceptibility of UHb and WTplants toward the agronomically important biotrophicpathogen, powdery mildew (B. graminis). Because ofthe already high susceptibility toward B. graminis of thebarley cultivar, Golden promise, of the UHb and WTplants, we used an oat-adapted mildew (B. graminis f.sp. avenae). As expected, conidial appressorial hyphaeof B. graminis f. sp. avenae failed to penetrate epidermalleaf cells on WT barley plants (Fig. 5A). In contrast, inthe UHb plants, a subpopulation of appressorial hyphaepenetrated epidermal cells and formed haustoria (Fig. 5B,C). These haustoria were at least initially functional asseen from the development of surface hyphae indicatingthe transfer of nutrients from the host to the pathogen.Quantitative assessments of the infection process inboth UHb and WT plants at 40 h after infection (hai)suggested that approximately 16% of infection sites hadcompromised non-host resistance against B. graminis f.sp. avenae (Table 1). However, by 88 hai the frequencyof infections sites with haustoria was reduced. Theseobservations are consistent with a delayed developmentof non-host resistance toward mildew in the UHb plants.

Characterizing the effects of endosperm-specificnsHb overexpression

HHb plants (which have overexpression of nsHbspecifically in the starchy endosperm) were scored foryield and yield components. HHb plants had a reducedyield compared to WT control plants, mostly due to areduced number of spikes (Fig. 6A–D). An additionalset of plants transformed with a DNA construct identicalto that used to generate the HHb plants, but where thensHb was replaced with green fluorescent protein (GFP)was included as an additional control of the effects oftransgenity in itself. The total yield, number of spikes,number of grains per spike and average grain weightof the GFP control line were all similar to that of theWT control line (Fig. 6A–D), demonstrating that thetransgenic construct had no effect.

Effects of endosperm-specific overexpressionof hemoglobin on grain iron concentration

Iron bound in (plant) hemoglobin represents a source ofnutritional iron with its high bioavailability for humans

598 Physiol. Plant. 150, 2014

Fig. 4. Nitric oxide emissions in nmol per hour (h) per gram fresh weight (g−1 FW) from roots (left panels) and leaves (right panels) of WT (opencircles) or UHb (closed circles) plants after transition to varying degrees of hypoxia in the dark. Bars indicate standard errors.

(Proulx and Reddy 2006). Overexpression of planthemoglobins in the starchy endosperm of cereal grainscould therefore be an ideal strategy for biofortifica-tion of iron toward dietary purposes. However, ironconcentration measurements showed that HHb grainshad a significantly (P < 0.02) lower concentration ofiron (51.7 ppm or 0.93 μmol g−1) as compared withWT grains (79.2 ppm or 1.42 μmol g−1). The starchyendosperm is a terminal tissue system, which goesthrough programmed cell death following accumulationof starch grains and storage protein. In this process, thecytosolic proteins are degraded by endogenous pro-teases, which could possible reduce the amount of nsHbin mature grains of HHb. Western blotting confirmedthat levels of endosperm hemoglobin in the HHb plantsstarted to decrease at the onset of grain drying in devel-oping spikes which occurs at 50–60 days post anthesis(Fig. 6E). Indeed, nsHb was only barely detectable inmature dry grains after harvest (Fig. 6E, lane 3).

Discussion

Overexpression of nsHb in plants has been associatedwith a number of physiological processes which maybe beneficial for improving crops. We previously

demonstrated in Arabidopsis that overexpression ofeither class 1 or 2 type nsHb resulted in earlier floweringand increased biomass (Hebelstrup and Jensen 2008).In other studies nsHb overexpression has resulted in anincreased tolerance toward hypoxic stress (Igamberdievet al. 2004, 2006, Hebelstrup et al. 2006). Other studiessimilarly indicate an increased growth and yield intobacco (Holmberg et al. 1997) or alfalfa root cultures(Dordas et al. 2003) and that specific overexpression ofeither type 1 nsHb or type 2 nsHb in seeds may resultin increased seed yield (Thiel et al. 2011, Vigeolas et al.2011). Barley is a major crop in the EU where productionin 2011 was approximately 52 Mtonnes (EUROSTAT2012). This species which also has several featuresassociated with a model plant (Varshney et al. 2007) andtherefore represents a good target with which to assessthe biotechnological effects of nsHb overexpression.Although a transgenic approach was followed in thiswork, our conclusions are likely to be equally validfor germplasm encoding a natural allele variant leadingto ectopic increased Hb expression. Thus, we assessedwhether any positive yield traits could be induced in thecereal crop barley by overexpression of the barley class1 nsHb identified by Taylor et al. (1994). However, it ispossible that natural alleles leading to nsHb expression

Physiol. Plant. 150, 2014 599

Fig. 5. Infection progress 40 hai of oat-adapted mildew (Blumeriagraminis f. sp. avenae) conidia in WT (Golden Promise) barley visualizedfollowing aniline blue staining and UV microscopy. (A) WT control line,infection is arrested very early by formation of papillae (P) and possiblecell death (CD) as visualized by red autofluorescence. (B and C) UHbline. The infection is much advanced with the formation of haustoria(Ha) and extensive surface hyphae (Hy). Bars indicate 100 μm.

Table 1. Classification of barley infection site responses to oat-adapted(Blumeria graminis f. sp. avenae) condia at 40 and 88 h after infection(hai). * indicates significant differences (P < 0.05) in the proportionof scored responses when compared to the WT according to Student’st-test (n = 50). aRed autofluorescence (RAF) indicates possible cell death.

Proportion of infection sites (%)

Papilla RAFa Haustorium

40 hai WT control 25.2 74.8 0.1UHb 33.1* 50.4* 16.6*

88 hai WT control 26.8 72.8 0.4UHb 22.4 72.7 5.0*

only in specific tissues, which is not examined in thiswork, may have positive effects that we did not observein this study.

In contrast to the effects on Arabidopsis, overexpres-sion of nsHb in barley had a negative effect on growthand delayed development and flowering. Crucially, seedyield in nsHb overexpressing plants was reduced bynearly 50%. The delayed development (Fig. 1D) wasevident in all of the three independent overexpressionlines UHb7, 8 and 9, where overexpression had been

Fig. 6. (A–D) Yield (g grain per plant) and yield components (numberof spikes per plant, number of grains per spike and grain weight) ofHHb plants in comparison with two control lines (WT, Golden Promise)and a vector control line (Golden promise) with overexpression of GFP.Significant differences between either of the control lines with the HHbline according to Student’s t-test are marked with either *P < 0.05,**P < 0.01 or n.s. (not significant). There were no significant differencesbetween the two control lines for any of the parameters. (E) Decayof nsHb content in WT and HHb barley grains during ripening. Tenmicrograms of protein purified from T3 HHb or WT grains of developingplants at 20, 50, 60 DAP or grains 3 months after harvest were analyzedby western blotting of HvHb1.

confirmed. However, due to space limitations, only oneline (UHb8) was carried on for quantitative assessmentin the following generation. We needed to confirm thatat least some of these effects could be due to altered NOproduction and therefore signaling. Some of the effectsof nsHb overexpression in plants, such as early flowering(Hebelstrup and Jensen 2008) and increased tolerancetoward hypoxia (Dordas et al. 2003, Igamberdiev et al.2004) have been demonstrated to be the result of NOscavenging by nsHb. Therefore, we measured NO emis-sions from WT control and plants with overexpression ofnsHb. Overexpression plants had a significantly lower

600 Physiol. Plant. 150, 2014

NO emission than WT plants. The effect was equiva-lent to previous observations in Arabidopsis (Hebelstrupet al. 2012) demonstrating that barley class 1 nsHb isas efficient as Arabidopsis class 1 nsHb in scaveng-ing of NO. This suggests that the differential effects ofnsHb overexpression between Arabidopsis and barleymay not be due to different abilities toward scavengingof NO by hemoglobin between the two species, butrather due to different effects of NO, different endoge-nous NO levels or different sensitivity toward NO. Wehave evidence from Arabidopsis that the effect of NO ongrowth is concentration-dependent, and that NO, whenprovided in the form of sodium nitroprusside (SNP, aNO releasing compound), is stimulatory or inhibitoryat concentrations below or above 50 μM, respectively(Hebelstrup and Jensen 2008). It may be that in barleythe thresholds of these stimulatory/inhibitory thresholdsare different than in Arabidopsis. In Arabidopsis high NOlevels delay flowering (He et al. 2004), whereas in otherspecies, such as the free-floating aquatic plant duckweed(Lemna aquinoctialis) NO induces flowering (Khuranaet al. 2011). In wheat, increased NO levels were detectedat the onset of flowering (Kolbert et al. 2011). Our UHbphenotype data are therefore consistent with the ideathat NO is required for flowering induction in cerealsand therefore nsHb overexpression should be consideredto be a deleterious trait as regards yield. However, asindicted by this study the effect of nsHb overexpressionvaries among species, and therefore results from barleycannot automatically be extrapolated to other crops.

Our previous studies have shown that plant nsHbsare effective modulators of pathogen susceptibility inArabidopsis (Mur et al. 2012). This effect was mediatedby modulation of host NO levels through scavengingby hemoglobin. NO plays a central role in many dif-ferent types of pathogenic responses: It is involved intriggering the hypersensitive response, a form of pro-grammed cell death, which occurs in plants to stopinfection progress of (hemi)biotrophic pathogens. How-ever, NO is also a central component in controlling theresponse toward necrotrophic types of pathogens. Wefound that hemoglobin overexpression increased suscep-tibility toward both types of pathogens (Mur et al. 2012).Crop diseases caused by microbial pathogens have amajor negative impact on crop yield. In estimating theperspectives of the biotechnological use of hemoglobinoverexpression in cereal crops, it is therefore of impor-tance to assess whether any beneficial effects in yield orincreased tolerance toward abiotic stress is compromisedby increased pathogen susceptibility. We therefore choseto test susceptibility in the hemoglobin overexpressionplants (UHb) in comparison with control WT plantstoward the economically important biotrophic pathogen

B. graminis, which causes the disease mildew in bar-ley and other cereals. Spores of oat-adapted mildew,B. graminis f. sp. avenae are usually arrested veryearly in infection in barley by basal response mecha-nisms such as formation of papillae in infected leaves.The nsHb overexpression plants (UHb) demonstrated apartial breakdown of basal resistance, where infectionprogressed and an initial infection with penetration ofhost cells and formation of haustoria was taking place(Fig. 5, Table 1). This shows, in line with previous studiesin Arabidopsis, that nsHb-overexpression in barley jeop-ardizes basal pathogen responses and indicates the riskthat cereal crops in the field overexpressing nsHb wouldbe more susceptible to yield losses caused by diseases.Thus, nsHb overexpression again is a deleterious trait.

Some of the complications of using ectopic expressionof nsHb may be avoided if nsHb is overexpressed inspecific tissues. We therefore tested the effects ofgrain endosperm-specific expression, where expressionwas driven by a hordein-D promoter (HHb plants).Earlier studies in Arabidopsis have shown that specificexpression of plant nsHbs in seeds increased seed yield(Thiel et al. 2011, Vigeolas et al. 2011). In addition,plant Hbs have been demonstrated to be an efficientsource of iron with high bioavailability for humans(Proulx and Reddy 2006), such specific overexpressionof nsHb in endosperm, could increase the bioavailabilityof iron content in cereal grains. In contrast to previousobservations in Arabidopsis, we did not see an increasedseed yield when nsHb was specifically overexpressed inbarley grains. On the contrary, HHb overexpressing bar-ley grains were smaller and the plants produced fewerspikes than WT control grains (Fig. 6). In this study weused a class 1 nsHb, similarly to the study in Arabidopsisreported by Thiel et al. (2011). However, in Arabidopsis,increased seed yield could also be achieved by overex-pression of a class 2 nsHb (Vigeolas et al. 2011), and sowe cannot exclude the possibility that overexpression ofa class 2 nsHb in barley could benefit grain yield or size.Surprisingly, we found that the increased nsHb contentin the endosperm had a negative effect of the total ironlevel in the grain. We saw that nsHb protein content inthe HHb overexpression grains was reduced at matura-tion of the grain; however, it is unlikely that this wouldactually remove any iron from the endosperm. NO isknown to be an important signal for the stimulation ofbiosynthesis of ferritin in grains (Arnaud et al. 2006)so nsHb overexpression in grains may have decreasedsuch stimulation by scavenging endogenous NO.

In conclusion, we observed several deleterious effectsof nsHb-overexpression in the cereal crop plant barleyby either an ectopic or an endosperm-specific promoter.In line with previous reports we confirmed that barley

Physiol. Plant. 150, 2014 601

class 1 nsHb scavenges endogenous NO, but the effectsof nsHb overexpression in barley on development andyield are opposite of earlier reports. This demonstratesby an example the necessity for using actual crop plantsrather than laboratory model plants when assessingbiotechnological uses of gene technology. However,nsHbs are effective modulators of cell-specific NO, andso it still remains to be assessed if root- or other types oftissue-specific overexpression or inducible expression ofany nsHb type could improve stress tolerance, in a waythat would improve crop yield.

Authors’ contributions

K. H. H. conceived of the idea, wrote the manuscriptand produced transgenic barley plants, confirmedoverexpression of nsHb by qRT-PCR and westernblotting, performed yield assessment and conductedNO measurements at Life Science Trace Gas Facility,Radboud University, with the assistance of J. M., S. M.C. and F. J. M. H. J. K. S. assisted on yield assessments andscored the development of plants. C. S. and L. A. J. M.assessed the interactions of plants with B. graminis andhelped edit the final manuscript. J. K. S. performed ironconcentration analysis of grains. A. U. I. supervised thedesign, execution and interpretation of the experiments.

Acknowledgements – We are much thankful to our field andgreenhouse manager Ole Brad Hansen for coordination andmaintenance of plant growth and to Giuseppe Dionisio fortechnical assistance on western blotting. The study wassupported by The Danish Research Council, Technologyand Production Sciences and EU-FP6-Infrastructures-5program, project FP6-026183 ‘Life Science Trace GasFacility’.

References

Appleby CA (1992) The origin and functions ofhemoglobin in plants. Sci Prog 76: 365–398

Arnaud N, Murgia I, Boucherez J, Briat JF, Cellier F,Gaymard F (2006) An iron-induced nitric oxide burstprecedes ubiquitin-dependent protein degradation forArabidopsis AtFer1 ferritin gene expression. J Biol Chem281: 23579–23588

Bogusz D, Appleby CA, Landsmann J, Dennis ES, TrinickMJ, Peacock WJ (1988) Functioning haemoglobin genesin non-nodulating plants. Nature 331: 178–180

Borisjuk L, Rolletschek H (2009) The oxygen status of thedeveloping seed. New Phytol 182: 17–30

Carciofi M, Shaik SS, Jensen SL, Blennow A, Svensson JT,Vincze E, Hebelstrup KH (2011) Hyperphosphorylationof cereal starch. J Cereal Sci 54: 339–346

Clarke SM, Cristescu SM, Miersch O, Harren FJM,Wasternack C, Mur LAJ (2009) Jasmonates act withsalicylic acid to confer basal thermotolerance inArabidopsis thaliana. New Phytol 182: 175–187

Cristescu SM, Persijn ST, Hekkert STL, Harren FJM (2008)Laser-based systems for trace gas detection in lifesciences. Appl Phys B-Lasers O 92: 343–349

Dordas C, Hasinoff BB, Igamberdiev AU, Manac’h N,Rivoal J, Hill RD (2003) Expression of a stress-inducedhemoglobin affects NO levels produced by alfalfa rootcultures under hypoxic stress. Plant J 35: 763–770

Dordas C, Hasinoff BB, Rivoal J, Hill RD (2004) Class-1hemoglobins, nitrate and NO levels in anoxic maizecell-suspension cultures. Planta 219: 66–72

Duff SMG, Guy PA, Nie XZ, Durnin DC, Hill RD (1998)Haemoglobin expression in germinating barley. Seed SciRes 8: 431–436

EUROSTAT (2012) Agricultural products. Available athttp://epp.eurostat.ec.europa.eu/statistics_explained/index.php/Agricultural_products (accessed 1 July 2013)

Frey AD, Oberle BT, Farres J, Kallio PT (2004) Expressionof Vitreoscilla haemoglobin in tobacco cell culturesrelieves nitrosative stress in vivo and protects from NOin vitro. Plant Biotechnol J 2: 221–231

Hansen TH, Laursen KH, Persson DP, Pedas P, Husted S,Schjoerring JK (2009) Micro-scaled high-throughputdigestion of plant tissue samples for multi-elementalanalysis. Plant Methods 5: 12

Hansen TH, Lombi E, Fitzgerald M, Laursen KH,Frydenvang J, Husted S, Boualaphanh C, ResurreccionA, Howard DL, de Jonge MD, Paterson D, Schjoerring JK(2012) Losses of essential mineral nutrients by polishingof rice differ among genotypes due to contrasting grainhardness and mineral distribution. J Cereal Sci 56:307–315

Hargrove MS, Brucker EA, Stec B, Sarath G,Arredondo-Peter R, Klucas RV, Olson JS, Phillips GN(2000) Crystal structure of a nonsymbiotic planthemoglobin. Structure 8: 1005–1014

He YK, Tang RH, Hao Y, Stevens RD, Cook CW, Am SM,Jing LF, Yang ZG, Chen LG, Guo FQ, Fiorani F, JacksonRB, Crawford NM, Pei ZM (2004) Nitric oxide repressesthe Arabidopsis floral transition. Science 305:1968–1971

Hebelstrup KH, Jensen EO (2008) Expression of NOscavenging hemoglobin is involved in the timing ofbolting in Arabidopsis thaliana. Planta 227: 917–927

Hebelstrup KH, Hunt P, Dennis E, Jensen SB, Jensen EO(2006) Hemoglobin is essential for normal growth ofArabidopsis organs. Physiol Plant 127: 157–166

Hebelstrup KH, Igamberdiev AU, Hill RD (2007)Metabolic effects of hemoglobin gene expression inplants. Gene 398: 86–93

Hebelstrup KH, Christiansen MW, Carciofi M, Tauris B,Brinch-Pedersen H, Holm PB (2010) UCE: a uracil

602 Physiol. Plant. 150, 2014

excision (USER (TM))-based toolbox for transformationof cereals. Plant Methods 6: 15

Hebelstrup KH, van Zanten M, Mandon J, Voesenek LACJ,Harren FJM, Cristescu SM, Møller IM, Mur LAJ (2012)Haemoglobin modulates NO emission and hyponastyunder hypoxia-related stress in Arabidopsis thaliana. JExp Bot 63: 5581–5591

Hebelstrup KH, Shah JK, Igamberdiev AU (2013) The roleof nitric oxide and hemoglobin in plant developmentand morphogenesis. Physiol Plant 148: 457–469

Heckmann AB, Hebelstrup KH, Larsen K, Micaelo NM,Jensen EO (2006) A single hemoglobin gene in Myricagale retains both symbiotic and non-symbioticspecificity. Plant Mol Biol 61: 769–779

Holmberg N, Lilius G, Bailey JE, Bulow L (1997)Transgenic tobacco expressing Vitreoscilla hemoglobinexhibits enhanced growth and altered metaboliteproduction. Nat Biotechnol 15: 244–247

Igamberdiev AU, Seregelyes C, Manac’h N, Hill RD (2004)NADH-dependent metabolism of nitric oxide in alfalfaroot cultures expressing barley hemoglobin. Planta 219:95–102

Igamberdiev AU, Baron K, Manac’h-Little N, StoimenovaM, Hill RD (2005) The haemoglobin/nitric oxide cycle:Involvement in flooding stress and effects on hormonesignalling. Ann Bot 96: 557–564

Igamberdiev AU, Bykova NV, Hill RD (2006) Nitric oxidescavenging by barley hemoglobin is facilitated by amonodehydroascorbate reductase-mediated ascorbatereduction of methemoglobin. Planta 223: 1033–1040

Khurana A, Khurana JP, Babbar SB (2011) Nitric oxideinduces flowering in the Duckweed Lemnaaequinoctialis Welw. (Syn. L. paucicostata Hegelm.)under noninductive conditions. J Plant Growth Regul30: 378–385

Kolbert Z, Peto A, Szollosi R, Erdei L, Tari I (2011) Nitricoxide (NO) generation during vegetative/generativetransition of the apical meristem in wheat. Acta BiolSzeged 55: 95–97

Kubo H (1939) Uber Hamoprotein aus denWurzelknollchen von Leguminosen. Acta Phytochim(Tokyo) 11: 195–200

Mur LAJ, Sivakumaran A, Mandon J, Cristescu SM, HarrenFJM, Hebelstrup KH (2012) Haemoglobin modulatessalicylate and jasmonate/ethylene-mediated resistancemechanisms against pathogens. J Exp Bot 63:4375–4387

Perazzolli M, Dominici P, Romero-Puertas MC, Zago E,Zeier A, Sonoda M, Lamb C, Delledonne M (2004)Arabidopsis nonsymbiotic hemoglobin AHb1 modulatesnitric oxide bioactivity. Plant Cell 16: 2785–2794

Prats E, Gay AP, Mur LAJ, Thomas BJ, Carver TLW (2006)Stomatal lock-open, a consequence of epidermal cell

death, follows transient suppression of stomatal openingin barley attacked by Blumeria graminis. J Exp Bot 57:2211–2226

Proulx AK, Reddy MB (2006) Iron bioavailability ofhemoglobin from soy root nodules using a Caco-2 cellculture model. J Agric Food Chem 54: 1518–1522

Shah JK, Cochrane DW, De Paepe R, Igamberdiev AU(2013) Respiratory complex I deficiency results in lownitric oxide levels, induction of hemoglobin andupregulation of fermentation pathways. Plant PhysiolBiochem 63: 185–190

Shi XZ, Wang XL, Peng FT, Zhao Y (2012) Molecularcloning and characterization of a nonsymbiotichemoglobin gene (GLB1) from Malus hupehensis Rehd.with heterologous expression in tomato. Mol Biol Rep39: 8075–8082

Smagghe BJ, Hoy JA, Percifield R, Kundu S, Hargrove MS,Sarath G, Hilbert JL, Watts RA, Dennis ES, Peacock WJ,Dewilde S, Moens L, Blouin GC, Olson JS, Appleby CA(2009) Correlations between oxygen affinity andsequence classifications of plant hemoglobins.Biopolymers 91: 1083–1096

Taylor ER, Nie XZ, Macgregor AW, Hill RD (1994) Acereal hemoglobin gene is expressed in seed and roottissues under anaerobic conditions. Plant Mol Biol 24:853–862

Thiel J, Rolletschek H, Friedel S, Lunn JE, Nguyen TH, FeilR, Tschiersch H, Muller M, Borisjuk L (2011)Seed-specific elevation of non-symbiotic hemoglobinAtHb1: beneficial effects and underlying molecularnetworks in Arabidopsis thaliana. BMC Plant Biol11: 48

Trevisan S, Manoli A, Begheldo M, Nonis A, Enna M,Vaccaro S, Caporale G, Ruperti B, Quaggiotti S (2011)Transcriptome analysis reveals coordinatedspatiotemporal regulation of hemoglobin and nitratereductase in response to nitrate in maize roots. NewPhytol 192: 338–352

Varshney RK, Langridge P, Graner A (2007) Application ofgenomics to molecular breeding of wheat and barley.Adv Genet 58: 121

Vigeolas H, Huhn D, Geigenberger P (2011)Nonsymbiotic hemoglobin-2 leads to an elevatedenergy state and to a combined increase inpolyunsaturated fatty acids and total oil content whenoverexpressed in developing seeds of transgenicArabidopsis plants. Plant Physiol 155: 1435–1444

Zadoks JC, Chang TT, Konzak CF (1974) Decimal code forgrowth stages of cereals. Weed Res 14: 415–421

Zhao L, Gu RL, Gao P, Wang GY (2008) A nonsymbiotichemoglobin gene from maize, ZmHb, is involved inresponse to submergence, high-salt and osmoticstresses. Plant Cell Tissue Organ Cult 95: 227–237

Edited by I. M. Møller

Physiol. Plant. 150, 2014 603