Journal of Experimental Botany, Vol. 48, No. 312, pp. 1337-1356, July 1997

REVIEW ARTICLE

Hormonal signalling in cereal aleurone

Journal ofExperimentalBotany

Paul C. Bethke1, Robert Schuurink2 and Russell L. Jones1'3

1 Department of Plant and Microbial Biology, University of California, Berkeley, CA 94720-3102, USA2 Max Planck Institute, Carl Von Linneweg 10, D-50829 Kdln, Germany

Received 23 January 1997; Accepted 2 May 1997

Abstract

The cereal aleurone is an excellent model system forhormonal signalling in plants. When treated with gib-berellins (GAs), cereal aleurone layers and aleuroneprotoplasts initiate signalling cascades that result inthe synthesis and secretion of hydrolytic enzymes,most notably a-amylases. Abscisic acid (ABA) antagon-izes the effects of GA and stimulates the productionof ABA-up-regulated proteins. Receptors for GA andABA have been localized to the aleurone cell PM, andevidence suggests that another ABA receptor func-tions within the cytoplasm. Cytosolic and membrane-bound second messengers have been identified andsignal transduction pathways are beginning to beunderstood. Transacting factors that regulate the tran-scription of hormonally induced genes have beencloned and bring us closer to linking cytosolic signalsto changes in gene expression. In this review, recentdata pertaining to hormonal signalling in cereal aleur-one are summarized. Relationships between signallingand effector molecules are discussed, and models forhormone-induced signalling pathways are proposed.

Key words: Cereal aleurone, hormonal signalling, gibberel-lin, abscisic acid, gene expression.

Introduction

Research on the cereal aleurone spans at least twocenturies

The cereal aleurone is one of the most intensively investi-gated tissues in plants. Because of its importance inmalting and brewing, chemists and biologists have carriedout detailed studies of barley grain function for nearly200 years. a-Amylases from cereals were the first enzymesto be studied in detail. A factor responsible for modifying

starch was identified in wheat extracts in 1811 byKirchhoff (quoted in Fischer and Stein, 1960), and maltdiastase (a-amylase) was the first enzyme to be isolatedand partially purified (Payen and Persoz, 1833). Theorigin of a-amylase and other starch-modifying enzymeswithin the cereal grain was a matter of considerablecontroversy among botanists at the end of the last century.Haberlandt (1890) argued that the aleurone layer was thesource of diastase in germinating grain, but his viewswere disputed by many, including Brown and Morris(1890), who argued that digestive enzymes were producedby the scutellum epithelium. The role of the aleuronelayer in the production of endosperm-modifying enzymesremained controversial through the early part of the 20thcentury, and EC Miller in his plant physiology text (1931)stated, 'There is much evidence, however, that thealeurone layer has no secretory powers'.

The cereal aleurone responds to gibberellins and abscisic.acid

Interest in the cereal aleurone as an experimental systemwas rekindled in 1960 by publications from L Paleg inAustralia and H Yomo in Japan. These workers showedindependently that gibberellic acid (GA3) and gibberellinAl (GA,) stimulated the production of a-amylase byde-embryonated barley grain (Paleg, 1960a; Yomo, 1960).Based on his experiments Paleg (19606) speculated thatan endogenous gibberellin (GA) produced by the barleyembryo played a role in regulating the production ofenzymes by barley endosperm. He supported this hypo-thesis by pointing out that GA-like substances couldbe isolated from barley and malt (Radley, 1959).Subsequently, numerous researchers have used exogenousGAs to stimulate the production of hydrolytic enzymesfrom de-embryonated cereal grain, aleurone layers andaleurone protoplasts.

Many experiments support the view that endogenous

'To whom correspondence should be addressed. Fax: +1 510 642 4995. E-mail: rjones@nature.berkeley.edu

6 Oxford University Press 1997

1338 Bethke et al.

GAs play a role in regulating a-amylase production bythe aleurone layer of barley (Skadsen, 1993; Zwar andChandler, 1995), rice (Mitsunaga and Yamaguchi, 1993)and wheat (Lenton et al., 1994). In the case of barley,for example, a dwarf mutant of the Himalaya cultivarthat is deficient in endogenous GAs has reduced levels ofendosperm a-amylase as well as reduced rates of leafgrowth (Zwar and Chandler, 1995). Application ofexogenous GA3 to this dwarf mutant increases both a-amylase production and leaf growth, indicating that GAsplay a role in regulating aleurone activity in vivo.

The extent to which newly synthesized GAs are requiredfor regulation of the aleurone layer is not known.Gro/Jelindemann et al. (1991) could not inhibit a-amylaseproduction by aleurone layers of Himalaya barley withinhibitors of GA biosynthesis, even though leaf elongationin Himalaya seedlings was reduced in inhibitor-treatedplants. These experiments suggest that de novo GAbiosynthesis may not be required to stimulate a-amylaseproduction by the aleurone layer.

The cereal aleurone is also responsive to added abscisicacid (ABA) (Hetherington and Quatrano, 1991; Chandlerand Robertson, 1994). ABA both inhibits many GA-induced responses and promotes responses that are uniqueto this hormone. There is limited information, however,on the role of ABA as an endogenous regulator ofaleurone function in cereals. ABA is present in aleuronelayers and embryos of barley and wheat (Ried andWalker-Simmons, 1990), and in barley, the amount ofthis regulator is higher in aleurone layers of dormantcultivars (Schuurink et al., 1993). Although evidencelinking ABA directly to in vivo function is lacking, ABAis a useful tool for studying hormonal signalling in thealeurone cell.

This review focuses on recent advances in understand-ing the mechanisms of hormone perception and signaltransduction using the cereal aleurone as a model system.Recently published work is reviewed and an attempt ismade to integrate this information into models thatexplain hormone action in the cereal aleurone.

The cereal aleurone as an experimental system forstudying plant hormone action

The ultrastucture of aleurone cells reflects their function

Many aspects of the cereal aleurone make it a uniqueand versatile experimental system. The aleurone layerconsists of from one (oat, maize, rye, wheat) to several(rice, barley) layers of uniform, highly differentiated cells(Fig. 1A). AJeurone cells of mature, dry barley grain aresurrounded by a thick, hemicellulosic wall and do notundergo further cell division. The aleurone layer can beremoved from the adhering dead starchy endosperm,and a uniform population of protoplasts (Fig. IB) can

Fig. 1. Photomicrographs of a barley aleurone layer (A) and barleyaleurone protoplasts (B, C). (A) Transverse section through aleuronecells of barley. Each cell contains a prominent nucleus (N) andnumerous PSVs. (B) Early stage barley aleurone protoplasts, eachcontaining numerous PSVs. (C) Late stage barley aleurone protoplastwith a single PSV occupying most of the cell A phytin globoid (G) isindicated by the arrow.

be prepared from the isolated layer using enzymes(Hooley, 1982).

The cytoplasm of the aleurone cell is characterized bynumerous protein storage vacuoles (PSVs), often referredto as aleurone grains, that almost fill the cell (Fig. 1).This organelle plays a key role in the response of thealeurone to hormones. PSVs store proteins that becomehydrolysed following GA-treament to provide the aminoacids necessary for secretory protein synthesis (Bethkeet al., 1996). PSVs are also the principal store of mineralsin the cereal grain. Minerals are sequestered in PSVs asphytin, an insoluble crystalline complex of K., Mg, andCa with phytic acid (inositol hexaphosphate). This crys-talline inclusion is also referred to as the globoid (Fig. 1).X-ray microanalysis has shown that approximately 75%of barley grain P, K, Mg, and Ca is stored in the aleuronelayer (Stewart et al., 1988). The PSVs increase in sizeand coalesce with time of incubation of aleurone layersor protoplasts until one large central vacuole is formed(Fig. 1C). The process of vacuolation of aleurone layercells and protoplasts is hastened by the presence of GAin the incubation medium (Jones and Price, 1970; Bushet al., 1986).

Signalling in cereal aleurone 1339

Lipid bodies or oleosomes that store neutral lipids arealso prominent organelles in the aleurone cell, occupyingas much as 30% of the cell's volume (Jones, 1969a).Oleosomes consist of a triglyceride core surrounded by ahalf-unit membrane (Huang, 1992). In the aleurone celloleosomes remain attached to the endoplasmic reticulum(ER) from which they originate (S Hillmer and RL Jones,unpublished results), and to the surface of PSV(Fernandez and Staehelin, 1985). The observation thatoleosomes are attached to the surface of PSV supportsthe hypothesis that some types of PSVs originate directlyfrom the ER (Okita and Rogers, 1996). Stored neutrallipids provide the aleurone cell with a carbon source thatcan be used before endosperm mobilization begins as wellas fatty acids that can be used for membrane synthesis.That fatty acids are metabolized by aleurone cells issupported by several lines of evidence. Aleurone cellshave numerous glyoxysomes containing malate synthaseand isocitrate lyase (Jones, 1972; Schuurink et al., 1996),and isolated aleurone layers can synthesize sucrose, pre-sumably by gluconeogenesis (Chrispeels et al., 1973).

The endomembrane system is also a prominent featureof the aleurone cell following GA treatment (Jones, 1969ft;Jones and Price, 1970; Cornejo et al, 1988). The ER ofGA-treated barley aleurone is organized into flattenedstacks of rough surfaced membranes, a characteristic ofcells that are synthesizing large amounts of secretoryproteins (Chrispeels, 1991). The Golgi apparatus is alsoprominent in GA-treated aleurone cells, and the numberand size of cisternae and associated vesicles increasesfollowing exposure to GA (Cornejo et al., 1988).

Aleurone layers and protoplasts respond similarly to GA

and ABA

Since the aleurone layer is a secretory tissue, considerableattention has been paid to the effects of GA and ABAon the synthesis of secreted hydrolases (for reviews seeFincher, 1989; Jones and Jacobsen, 1991). This reviewfocuses on responses in the aleurone cell to GA and ABAthat precede the onset of hydrolase synthesis. Theseresponses include altered gene expression and changes inthe concentration of signalling molecules as well as higherorder effects requiring the integration of multiple signal-ling pathways or effectors. A partial list of these responsesis presented in Table 1, and many of them are describedin detail below. Figure 2 shows the times at which severalGA-induced responses of the aleurone layer occur.Clearly, the synthesis and secretion of hydrolases arerelatively late events when compared to GA-inducedchanges in putative signals such as the concentrationof cytosolic free Ca2+ ([Ca2+];), cytosolic pH (pH,), 3'5'cyclic guanosine monophosphate (cGMP), and calmodu-lin (CaM)(Fig. 2).

With the exception of timing, the response of barley

100

80

60

20

10 100 1,000

Time after GA treatment (min)

10,000

Fig. 2. Time-course of GA-induced responses of barley and wheataleurone layers. Data were extracted from (Chandler et al., 1984; Brownand Ho, 1986; Heimovaara-Dijkstra et al., 1994a; Gubler et al., 1995;Bush, 1996; Penson et al., 1996; Schuunnk et al., 1996).

aleurone protoplasts to GA and ABA is almost identicalto the response of aleurone layers. Aleurone protoplastsand layers secrete a similar spectrum of hydrolases inresponse to GA, but the onset of secretion is delayed andthe rate of secretion is reduced in protoplasts relative toaleurone layers (Fig. 3). Other differences in the biochem-ical and molecular responses of aleurone protoplasts andlayers to GA tend to be minor and are related to thewall-less nature of the protoplast. Cells of the intactaleurone layer, for example, are interconnected by a largenumber of plasmodesmata, an observation that was firstmade in the 19th century by Tangl (1885). The functionof plasmodesmata in aleurone cells is not understood,but it is unlikely that their disruption when makingprotoplasts (Taiz and Jones, 1973) can be without con-sequence to the aleurone cell. Another aspect of protoplastisolation and culture that is likely to affect their responseto stimuli is that they are cultured in a solution of highosmolarity (about 1000 mOsM), whereas aleurone layersare incubated in solutions of much lower osmolarity(about 50 mOsM). The difference in turgor pressure atthe plasma membrane (PM) in walled aleurone layer cellsis likely to be profoundly different from that in aleuroneprotoplasts and may affect the ability of these cells torespond to stimuli.

0 1 2 3 4 5Days after GA treatment

Fig. 3. Time course of a-amylase secretion by GA-treated barleyaleurone layers and barley aleurone protoplasts.

1340 Bethke et al.

Table 1. Hormone-induced responses in cereal aleurone

Parameter Response Reference

Responses to GASignalling molecules

pH,cGMPCaM

Gene expressionCaMa-AmylaseGAMYBMAP kinaseStorage globulin

Higher order effectsHydrolase secretionProtease activityCitrate and phosphate releaseVacuolationCell death

Responses to ABASignalling molecules

[Ca'1,PH,MAP kinase activityEmBP-1

Gene expressionEmASIRab-16PHAV1HVA1/HVA22Dehydrina-AmylaseGAMYB

Higher order effectsSecretion of hydrolasesMalate releaseCitrate and phosphate releaseVacuolationCell death

Rapid, sustained increaseTransient decreaseTransient increaseSlow, sustained increase

IncreaseIncreaseIncreaseDecreaseDecrease

IncreaseIncreaseIncreaseAcceleratesPromotes

DecreaseTransient increaseIncreaseUnknown

IncreaseIncreaseIncreaseIncreaseIncreaseIncreaseDecreaseDecrease

DecreaseIncreaseDecreaseSlowsRetards

(Bush, 1996)(Heimovaara-Dijkstra et al, 1994a)(Penson et al, 1996)(Schuurink et al, 1996)

(Schuurink et al, 1996)(Chandler et al., 1984)(Gubler et al, 1995)(Huttly and Phillips, 1995)(Heck et al, 1993)

(Jones and Jacobsen, 1991)(Bethke et al, 1996)(Drozdowicz and Jones, 1995)(Hooley, 1982)(Kuo et al, 1996)

(Wang et al, 1991)(Van der Veen et al, 1992)(Kiietsch et al, 1996)(Guiltinan et al, 1990)

(Marcotteef al, 1988)(Leah and Mundy, 1989)(Mundy and Chua, 1988)(Hong et al, 1988)(Shen and Ho, 1995)(Robertson et al, 1995)(Chandler et al, 1984)(Gubler et al, 1995)

(Jones and Jacobsen, 1991)(Drozdowicz and Jones, 1995)(Drozdowicz and Jones, 1995)(Hooley, 1982)(Kuo et al, 1996)

Although GA-treated aleurone protoplasts respondmore slowly than GA-treated aleurone layers, wall-lesscells offer several advantages that make them exception-ally useful experimental tools. The absence of the cellwall allows experimental manipulations such as patchclamping (Bush et al., 1988) and DNA transfection (Linet al., 1996). Protoplasts also facilitate procedures such asmicroinjection (Gilroy and Jones, 1994), in situ immuno-localization (Schuurink et al., 1996), and the monitoringof amylase secretion from single cells (Hillmer et al.,1993).

Kende and Gardner, 1976, for a review of earlier work).These approaches relied on cell fractionation techniquesand the use of equilibrium dialysis and affinity chromato-graphy with radiolabelled GA[ and GA, to detect binding,but these approaches proved unsuccessful (Kende andGardner, 1976). More recently, several novel approacheshave been used to localize the GA and ABA receptors inthe aleurone cell. These have included the use of cellimpermeant forms of GA, microinjection to introduceGA or ABA into the cell, photoaffinity labelling, andanti-idiotypic antibodies.

Perception of GA and ABA by cereal aleurone

Receptors for ABA and GA have yet to be isolated fromplants, but a great deal of experimentation has beencarried out with the cereal aleurone in attempts to isolatereceptors and to establish their location within the cell.Experiments to isolate cellular fractions from aleuronetissue that bound GAs were begun in the 1970s (see

GA is perceived at the plasma membrane of thealeurone cell

To localize the GA receptor in the aleurone of wild oat(Avenafatua), Hooley et al. (1991) carried out a series ofelegant experiments with a non-permeable form of GA4.Hooley et al. (1991) covalently coupled GA4 to Sepharose6B beads (approximately 120 /xm diameter) via a spacer

arm on carbon 17 of GA4 that allowed the GA4 moleculeto protrude about 2 nm from the surface of the Sepharosebead. Whereas the GA4-Sepharose complex was able toinduce the production of a-amylase in wild oat proto-plasts, it did not stimulate a-amylase production whenpresented to isolated wild oat aleurone layers (Hooleyet al., 1991). This experiment provides convincing supportfor the hypothesis that the GA receptor in the wild oataleurone cell is at the surface of the cell.

A different experimental approach has shown that theGA receptor is located at the surface of the barleyaleurone protoplast (Gilroy and Jones, 1994).Microinjection was used to introduce GA3 or inactiveGA8 into the cytosol of aleurone protoplasts. Theresponse of the cell to microinjected GAs was comparedwith the response of the cell to externally applied hormone(Gilroy and Jones, 1994). In these experiments theresponses monitored were a-amylase secretion (Hillmeret al., 1993), vacuolation (Bush et al., 1986), and expres-sion of the glucuronidase reporter gene fused to a hor-monally responsive a-amylase promoter (Lanahan et al.,1992). These microinjection experiments showed thataleurone protoplasts did not respond when GA3 wasintroduced into the cytosol. Only when cells were exposedto GA3 by adding it to the incubation medium did theyrespond to the hormone (Gilroy and Jones, 1994).

Photoaffinity labelling has been a powerful tool forinvestigating ligand interactions. Although earlierattempts to isolate GA-binding proteins by this techniquefailed (Kende and Gardner, 1976), more recent workwith improved photoaffinity probes has produced encour-aging results. Hooley and his colleagues have synthesizedphotoactivatable analogues of GA4 that retain biologicalactivity and have used these probes to isolate photoaffin-tiy-labelled proteins from wild oat aleurone as well asother GA-responsive tissues (R Hooley, personal com-munication). This group has isolated several proteinsfrom oat aleurone and pea shoots that are candidates forthe GA receptor.

Anti-idiotypic antibodies have proven to be powerfultools in investigations of hormone action in animal cells,and they have been used to a limited extent in investi-gations of plant hormone action (reviewed by Hooleyet al., 1992). Anti-idiotypic antibodies were prepared byimmunizing rabbits with a monoclonal antibody thatrecognized domains of the GA4 molecule that are thoughtto confer biological activity. These anti-idiotypic antibod-ies antagonized GA4 in wild oat protoplasts and causedagglutination of protoplasts (Hooley et al., 1992). Theseresults have been interpreted as showing that the anti-idiotypic antibodies interact with molecules on the surfaceof oat protoplasts and that these molecules play a role inthe response of the protoplast to GA4.

Other circumstantial evidence supports the concept thatthe GA receptor is located at the PM of the aleurone cell

Signalling in cereal aleurone 1341

(Goodwin and Carr, 1972; Musgrave et al., 1972; Hooleyet al., 1992). There appears to be little correlation betweenpresumed GA uptake rates and biological activity. Forexample, GA1; GA4 and 2,2-dimethyl GA4 differ inhydrophobicity, and hence in the rate at which they areexpected to cross the PM. Yet there was no difference inthe amount of a-amylase produced by Avena fatua proto-plasts incubated in these three GAs (Beale et al., 1985).Likewise, although GA3 is thought to enter cells in theprotonated form (O'Neill et al, 1986), external pH hasa inconsistent effect on the response of aleurone cells toGA3. Goodwin and Carr (1972) found that the responseof aleurone layers to GA3 (pATa 3.97) was not affectedby pH over the range of 4.4-5.2. Sinjorgo et al. (1993),however, found that the sensitivity of barley aleuronelayers to GA3 increased as external pH decreased frompH 6 to pH 3.7.

An alternative explanation for the lack of a correlationbetween GA uptake and the response of the aleurone isthat rapid metabolism of GA prevents the accumulationof the hormone in the cytosol. Studies of GA! metabolismhave shown that a steep gradient of GA concentration(high outside, low inside) is maintained accross the PM(Musgrave et al., 1972; Nadeau et al., 1972). GAt wasrapidly metabolized by barley aleurone layers to inactiveGAt glucoside or GA8, and estimates indicate that thecytosolic GAj concentration was as much as two ordersof magnitude lower than that in the external medium(Hooley et al., 1992). Rapid metabolism of GAs in thealeurone cell has important consequences for GA signal-ling. Hormone metabolism provides a mechanism forremoval of the GA signal. Signal propagation, therefore,is likely to continue only as long as active GAs areperceived by the cell. The existence of GA metabolizingenzymes in the cytosol, however, does not necessarily ruleout the existence of cytosolic receptors.

ABA is perceived by the aleuone cell both at the PM and atan internal site

Microinjection was used to localize the ABA receptor inbarley aleurone cells (Gilroy and Jones, 1994; Gilroy,1996). Because many responses of the aleurone cell toABA are to inhibit GA-stimulated events, experimentswere designed to ask whether GA-induced responses wereinhibited when ABA was microinjected into the cytosolor presented to the surface of the aleurone protoplast.The conclusion from these experiments was similar tothat drawn from the experiments using microinjection ofGA, namely that a receptor for ABA is localized to thePM of the barley aleurone cell (Gilroy and Jones, 1994;Gilroy, 1996). Using a similar experimental strategy,Schroeder and co-workers concluded that the ABA recep-tor in guard cells is also localized to the cell's surface(Anderson et al., 1994).

1342 Bethke et al.

Microinjection experiments also indicate that an addi-tional receptor for ABA is present within the aleuronecell (Gilroy, 1996). Gilroy used a reporter gene constructin which the promoter of the ABA-inducible wheat geneEm (Marcotte et al., 1988) was fused to GUS. When theEm-GUS construct and ABA were microinjected into thealeurone cytosol, Em-GUS expression was increased(Gilroy, 1996). These data led Gilroy to conclude thatthere are at least two sites of ABA perception in thebarley aleurone cell. One at the PM leads to down-regulation of GA-induced a-amylase gene expression; theother within the cell leads to up-regulation of the ABA-inducible gene Em (Gilroy, 1996).

Signal transduction from the plasma membrane tothe cytosol

As discussed above, compelling evidence suggests thatreceptors for both GA and ABA exist at the PM of cerealaleurone cells. How hormone binding to these receptorsis transduced into a cytosolic signal is unclear. Furthersignal propagation may utilize signalling molecules, suchas G-proteins and lipids, associated with the PM. Theactivity of PM ion channels or pumps may be stimulatedor inhibited. These changes in activity might couplehormone perception to cytosolic second messengers suchas Ca2+ or pH. Figure 4 presents models for transductionof ABA- (Fig. 4A) and GA-signals (Fig. 4B) in cerealaleurone. Although these models are supported by thedata discussed below, they will certainly be refined asadditional data become available.

G-proteins and lipids may link hormone perception to

cytosolic signals

G-proteins and lipids may link hormone perception atthe PM to cytosolic signalling pathways. A paradigm forthese initial links in the signal transduction chains leadingto cellular response is phosphoinositide signalling inanimals. In this signalling system, binding of an extracellu-lar signalling molecule to its receptor activates a hetero-trimeric G-protein. The Ga-subunit of this protein, inturn, activates a phospholipase that cleaves phosphotidylinositol-bisphosphate into inositol trisphosphate (IP3)and diacylglycerol (DAG). IP3 and DAG both act assecond messengers in signal transduction cascades.

The presence of GTP-binding proteins in the PM ofbarley aleurone cells has been shown directly by Wanget al. (1993). Following SDS-PAGE and protein blottingon to nitrocellulose, PM proteins with molecular weightsof 22-24 kDa and 16 kDa bound [a-32P]GTP. GTP bind-ing was competed by unlabelled GTP, GTPyS, GDP andGDP0S, but not ATP or ATPyS. Whether these proteinsfunction as signalling molecules in cereal aleurone isunknown. In vitro binding of GTP to PM proteins was

Phospho-lipase

GAreceptor

Antagonismof GA action

ATPase

Tyrosinephosphatase

ABA

Intern aiABA receptor

stimulated //amylase mRNA |transcription / / E r r i | R a b m R N A f

Nucleus

B GACa2+-channel

ABA Vreceptor o a

Vj*yGA Guanylyl

receptor cyclase

MAPK- 1 rrn2+i. A GA-stimulated ^MAPK t [Ca ], T t r a n s c r i p t i o n y ^

/ / C a M mRNA |

CaM f —csyyamyiase mRNA A

'fnternaT' / /ABA receptor

Nucleus

Fig. 4. Models of ABA- (A) and GA- (B) signalling pathways in cerealaleurone. (A) ABA binds to receptors at the PM and within the cellBinding to the receptor on the external face of the PM leads toreductions in [Ca2+]j and CaM, and antagonism of GA-action ThisABA signal may be transduced via a pathway involving G-proteins andDAG. ABA binding to an internal receptor stimulates transcription ofEm and other ABA-specific genes and inhibits transcription of CaMand a-amylase. A signalling pathway that includes a tyrosine phosphat-ase and a MAP kinase is also proposed. (B) GA binds to a receptoron the external face of the PM. This stimulates influx of Ca2+ from theapoplasm and increases [Ca2+]j. GA binding increases the amout ofcGMP and stimulates the transcription and translation of several genesincluding those for a-amylase and CaM. Unfilled arrowheads designatesignalling pathways Filled arrowheads indicate an increase or decreasein the adjacent parameter.

unaffected by coincubation of isolated PM with GTP andeither 100 ^M ABA or lOO^M GA3.

In the same report (Wang et al., 1993), protein blotswere probed with anti-ras Y13-259 and with antibodies

to a conserved sequence in animal G-proteins. Bothantibodies recognized proteins of 34-36 and 22-24 kDa.The anti-ras antibody recognized other proteins as well.It is possible, therefore, that barley aleurone containsproteins structurally similar to the subunits of heterotrim-eric G-proteins.

A different approach toward assessing the role ofG-proteins in aleurone cells has been taken by RichardHooley's group. They prepared degenerate oligonucleo-tide primers based on sequences conserved between plantand mammalian Ga and G/3 subunits. These primers wereused in the polymerase chain reaction (PCR) to amplifycDNAs prepared from GA^treated Avena fatua aleuronelayers (Jones et ah, 1996). The resulting PCR productswere cloned and found to have sequence similarity withGa and G/3 subunits. The transcripts for these productsappeared to be expressed at very low levels and may bepart of multigene families. It is not known if any of theseproteins are localized to the PM.

Although the data discussed above (Wang et ah, 1993;Jones et ah, 1996) are consistent with the hypothesis thatGTP-binding proteins play a role in signalling at the PMin cereal aleurone, further confirmation through biochem-ical or genetic characterization is required. This is particu-larly true in light of experiments done by Jones et ah(1995) and Kuo et ah (1996). Mastoparan, pertussistoxin (PTX) and cholera toxin (CTX) effect G-proteinaction in animal cells. Mas 7, an analogue of mastoparan,stimulated a-amylase production in aleurone protoplastsfrom Avena fatua (Jones et ah, 1995). This stimulationwas inhibited by PTX and CTX. PTX and CTX alsoreduced the inhibitory effect of ABA on GA-induced a-amylase production. PTX and CTX, however, did notinhibit GA-stimulated a-amylase secretion. In a secondreport (Kuo et ah, 1996), mastoparan, PTX and CTXhad no effect on secretion of a-amylase from GA-inducedwheat aleurone layers. These data provide some supportfor the hypothesis that G-proteins are involved in ABAsignalling, but little support for the hypothesis that theyplay a role in GA signalling.

To test for the involvement of DAG in aleurone signaltransduction, Ritchie and Gilroy (1996) applied1,2-dioctanoyl glycerol (1,2-DG), an active analogue ofDAG, to barley aleurone protoplasts. 1,2-DG mimickedthe effects of ABA. 1,2-DG prevented GA-induced a-amylase gene transcription and a-amylase secretion, butstimulated transcription of the ABA-inducible genesA SI and Em. Although [Ca2+]j increases followingGA-treatment in the absence of 1,2-DG (see below), itdoes not increase in its presence. These effects of 1,2-DGwere relatively specific, as they were not seen with 1,3-DG.In mammals DAG stimulates protein phosphorylation byactivation of protein kinase C. In aleurone cells a stimula-tion of protein kinase activity by 1,2-DG or DAG hasnot been demonstrated.

Signalling in cereal aleurone 1343

ABA treatment causes a transient hyperpolarization ofaleurone plasma membrane

ABA-induced signalling at the PM may not be restrictedto G-proteins and lipids. Changes in PM electrical poten-tial may also be involved. Such changes could be used tomodulate the activities of voltage-regulated ion channelsor H +-coupled transporters. Heimovaara-Dijkstra et ah(1994/)) used the lipophilic cation tetraphenylphosphon-ium to measure the membrane potential of barley aleuroneprotoplasts before and after application of ABA.Following ABA treatment (50 jiM), a rapid depolariza-tion of the PM was observed. From a mean resting valueof —53 mV, membrane potential decreased an averageof 9 mV, with peak hyperpolarization occurring 20 s afterABA addition. Membrane potential returned to restinglevels within 1 min despite the continued presence ofABA. This brief hyperpolarization occurred at ABAconcentrations of 10~9 to 10~4 M, with 10"8 M andgreater giving the maximal change in membrane potential.Since a small, statistically insignificant hyperpolarizationwas observed with the ABA analogues (±) cis/trans a-ionyledene acidic acid and (+) cis/trans ABA-glucoseester, the change in membrane potential was ratherspecific for ABA. Inhibitors of P-type ATPases, diethylst-ilbestrol (DES) and micanozole, both prevented ABA-stimulated hyperpolarization, suggesting that the protonpump was an early target of ABA signalling. In otherexperiments, ABA induced Rab gene expression, butABA in the presence of 100 ^M DES greatly reduced theamount of Rab mRNA that accumulated during a 1.5 hincubation period (Heimovaara-Dijkstra et ah, 19946).Whether this reduction in Rab mRNA was causedby loss of a hyperpolarization signal or by sustainedinhibition of P-type ATPases is not known.

Calcium fluxes through the aleurone plasma membrane arean essential component of hormone signalling

A critical component of hormonal signalling is regulationof [Ca2+]i (Bethke et ah, 1995; Bush, 1995). In plants[Ca2+]j acts as a second messenger for many environ-mental and hormonal signals (Bush, 1995). Increases in[Ca2+], can either activate enzymes directly (Roberts andHarmon, 1992) or indirectly through Ca2 +-binding pro-teins such as CaM. Spatial inhomogeneity in [Ca2+]jallows for targeting of the Ca2 + -signal to specific regionsof a cell or tissue. As shown in Fig. 4, [Ca2+], plays a keyrole in the response of the aleurone cell to hormonalsignals.

Data suggest that GA-treatment of aleurone cellsresults in calcium influx from the apoplast. Gilroy andJones (1992) and Bush (1996) demonstrated thatGA-induced increases in [Ca2+]; (see below) were depend-ent on extracellular Ca2+ ([Ca]o). At low [Ca]0, [Ca2+],no longer increased following GA-treatment. Further,

1344 BethkeelaL

when barley aleurone cells where imaged with a confocalmicroscope (Gilroy and Jones, 1992) or an intensifiedCCD camera (Bush, 1996), the highest [Ca2+]; wasobserved near the periphery of the cell. These observationsare consistent with GA-stimulated Ca2 + entry throughthe PM as a mechanism for elevating [Ca2+],. Calcium ispresumed to enter aleurone cells through an as yetunidentified PM Ca2 + channel. The existence of a Ca2 +

channel has been postulated on thermodynamic groundsand by analogy to hormonally-regulated, channel-mediated Ca2+ influx in stomatal guard cells (Schroederand Hagiwara, 1990).

Reduction of [Ca2+]j from elevated concentrations mayinvolve the activity of a PM Ca2+-ATPase. Bush andWang (1995) presented biochemical evidence for a PMCa2+-ATPase in wheat aleurone. These investigators usedisolated membrane vesicles to characterize the principalCa2 + transport activities in wheat aleurone cells. One ofthese, designated Type III, was found to be associatedwith PM marker enzymes. Transport was ATP-dependentand inhibited by the P-type ATPase blocker erythrosin B(0.1 jiM) and the CaM antagonist W7 (100 fjM). Activitywas not stimulated by bovine brain CaM. The A m forCa2 + was 2/xM, suggesting a role in reducing [Ca2+], toresting levels following a Ca2 +-increasing stimulus.

Signal transduction within the cytosol

In the aleurone cell, hormone perception and early signaltransduction events occur at the PM. These signals mustbe propagated throughout the cytosol, where biochemicalactivities must be co-ordinated in order to synthesize andsecrete hydrolases efficiently. [Ca2+],, acting in conjunc-tion with CaM, is thought to be essential for thisco-ordination (Bethke et al., 1995). Addtional signals,however, are also important. As illustrated in Fig. 4, thesemay include cGMP, cytosolic protons as well as proteinphosphorylation. Cross-talk between each of these signal-ling components is likely. Temporal or spatial convergenceof signalling pathways may encode information specifyingan appropriate cellular response.

Changes in the concentration of cytosolic calcium arecentral to hormone-mediated responses in the aleurone cell

As discussed above, Ca2+ is thought to enter the aleuronecell from the apoplasm following GA stimulation. Sincethe site of Ca2 + entry is close to the site of hormoneperception, Ca2 + entry may be an early event in thesignal transduction chains leading from GA binding tohydrolase secretion. Evidence for this comes from thework of Douglas Bush (1996). Bush loaded wheat aleur-one layers with the Ca2+-sensitive dye fluo-3 and mon-itored fluorescence in the cytoplasm before, during andafter perfusion with 5 M GA3. Within a few minutes of

GA3 addition, [Ca2+]; began to increase. The sustained[Ca2+], was 100-500 nM higher than the baseline concen-tration and was attained within 30-90 min of addingGA3. This increase in [Ca2+] ; was dependent on perceptionof an active GA. GA8, a biologically inactive GA did notincrease [Ca2+];, nor was an increase detected in aleuronecells from the GA-insensitive wheat mutant D6899. Asimilar, though slower increase in [Ca2+], followingGA-treatment was reported for protoplasts of barleyaleurone (Gilroy and Jones, 1992).

The increase in [Ca2+]j brought about by GA3 can beantagonized by ABA (Gilroy and Jones, 1992). Whenbarley aleurone protoplasts were first treated with GA3

and then with GA3 plus ABA, [Ca2+]j increased duringGA3 treatment, but decreased to the baseline concentra-tion in the presence of ABA. This effect of ABA on[Ca2+], is consistent with the hypothesis that ABA antag-onizes GA action by preventing a sustained increasein [Ca2+],.

When ABA was applied to non-GA-treated aleuronecells, a rapid decrease in [Ca2+], was observed (Wanget al., 1991). Fluorescence from the Ca2 + -sensitive dyeindo-1/AM decreased within seconds of ABA addition,reaching a steady intensity in about 10 s. Whether thisrepresents a new steady-state for [Ca2+], is not known asonly short term measurements were made. The magnitudeof the decrease in [Ca2+], depended upon ABA concentra-tion. A 20-70 nM decrease was observed with 10-50 /xMABA, and a decrease of about 150 nM was observed at500 fiM ABA (Wang et al., 1991). The importance of thelatter value is questionable given the unphysiologicallyhigh ABA concentration. Changes in [Ca2+]; were insens-itive to the Ca2 +-channel blockers nifedipine (35 ^M)and verapamil (150 /xM), but were sensitive to the Ca2 + -antagonists La3+ and Cd2 + . Wang et al. suggested thatABA-treatment may lower [Ca2+]; by activating a PMCa2 +-ATPase.

The calcium-binding protein calmodulin is an importantsecond messenger in cereal aleurone

Transduction of calcium signals is often accomplishedthrough the use of intermediary Ca2 +-binding proteins.Of these, the best known is CaM. The binding of Ca2 +

to CaM results in a conformational change that increasesthe affinity of CaM for CaM-activated enzymes. In thismanner, changes in [Ca2+]j may modulate changes inenzyme or transporter activity. In cereal aleurone, CaMis thought to be an essential component of signal transduc-tion pathways leading from GA-perception to hydrolasesecretion (Gilroy, 1996). A role for CaM in ABA-signalling has also been proposed (Hetherington andQuatrano, 1991).

In barley aleurone, the amounts of CaM mRNA andCaM protein are both under hormonal control. Schuurink

et al. (1996) found that physical stimulation of aleuronelayers increased the amount of CaM mRNA. Althoughthis increase in CaM mRNA occurred whether aleuronelayers were subsequently incubated in H2O, 10 mM CaCl2

or 5 fj.M GA3, it was prevented by incubation in 5 ^MABA. In contrast to CaM mRNA, CaM protein, asmeasured by densitometry of protein blots, did notincrease in aleurone layers treated with H2O, CaCl2 orGA3 (Schuurink et al., 1996). Hence the amount of CaMprotein in barley aleurone is under post-transcriptionalcontrol. Incubation in 5 M GA3 plus 10 mM CaCl2,conditions that lead to a-amylase secretion, doubled theamount of CaM protein. These data are consistent withprevious experiments of Gilroy and Jones (1993), whofound that treatment of barley aleurone layers with GA3

plus CaCl2 increased CaM protein by 50% as measuredby a radio-immunoassay.

The increase in CaM protein described by Schuurinket al. (1996) was approximately linear with time duringthe first 8 h of incubation with hormone. In the same setof experiments, GA3 plus CaCl2-treatment increased a-amylase mRNA after a lag of about 2 h. These data areconsistent with the hypothesis that CaM acts as anintermediate in the GA-induced signal transduction path-ways leading to hydrolase secretion (Figs 2, 4).GA-stimulated increases in CaM protein combined withGA-stimulated increases in [Ca2+]; would increase theamount of Ca2+-CaM several fold. This Ca2+-activatedCaM would then stimulate the activity of enzymes ortransporters required for hydrolase synthesis or secretion(Fig. 5).

In an elegant set of experiments, Simon Gilroy directlyexamined the roles of [Ca2+]i and CaM in GA- and ABA-signalling (Gilroy, 1996). By using microinjection tointroduce Ca2 + , Ca2 + chelators and CaM into barleyaleurone protoplasts, he has either mimicked or abolishedhormonally induced changes in [Ca2+]; and CaM (Gilroy,1996). In the absence of GA, mimicking GA-inducedchanges in [Ca2+], and CaM did not result in secretion ofa-amylase, vacuolation of protoplasts, or expression ofthe glucuronidase reporter gene fused to an a-amylasepromoter (Amy-GUS). When a sustained increase in[Ca2+]j was blocked by the Ca2+-chelator diazo-2, a-amylase secretion, but not Amy-GUS expression, wasprevented. This later observation is consistent with earlierexperiments demonstrating that GA in the absence ofexternal Ca2+ increases the transcription of a-amylasemRNA, but does not lead to a-amylase secretion(Deikman and Jones, 1985). When barley CaM wasmicroinjected into barley aleurone protoplasts, either inthe presence or absence of elevated [Ca2+]i, responsestypical of GA-treatment were not observed in the absenceof GA (Gilroy, 1996). Hence, although elevated [Ca2+]jis required for a-amylase secretion, neither elevated

Signalling in cereal aleurone 1345

[Ca2+], nor Ca2 + -CaM are sufficient for induction of a-amylase secretion or aleurone cell vacuolation.

Gilroy (1996) also treated aleurone protoplasts thathad been incubated in 5 M GA3 for 24 h with 5 /xMABA for 3 h. This ABA treatment caused [Ca2+],, Amy-GUS expression and a-amylase secretion to decrease(Gilroy, 1996). When [Ca2+]; was maintained atGA-stimulated levels, however, Amy-GUS expression anda-amylase secretion continued. Likewise, microinjectionof 1-10 ^M CaM also blocked the antagonistic effects ofABA on Amy-GUS expression and a-amylase secretion.These data suggest that ABA antagonizes GA action byreducing [Ca2+]j or the amount of Ca2+-CaM.

ABA induces a set of genes distinct from those inducedby GA (Table 1). Gilroy introduced Ca2 + and CaM intothe cytosol of barley aleurone protoplasts to see if lowlevels of [Ca2+]; or CaM were required for expression ofthe glucuronidase reporter gene fused to the wheat Empromoter (Em-GUS). Artificial elevation of [Ca2+]; orCaM in ABA-treated protoplasts did not reduce Em-GUSexpression. Since ABA-induced gene expression wasunchanged when [Ca2+]j was high, it is unlikely that allABA-induced signalling pathways are regulated by [Ca2+],or CaM. Significantly, microinjection of ABA into barleyaleurone protoplasts increased expression of Em-GUS.These data suggest that an internal receptor for ABA (seeabove) may be linked to a Ca2+-independent signallingpathway (Fig. 4).

Barley aleurone cells express genes encoding CaM-bindingproteins

To identify targets for Ca2+-CaM in barley aleurone cells,Schuurink and Jones (1995) used a molecular approach.Using Arabidopsis CaM-2 coupled to horse radish peroxi-dase as a probe, they were able to screen expressioncDNA libraries made of barley aleurone RNA for cDNAsencoding CaM binding proteins. Several cDNAs wereidentified, one of which was similar to a CaM-bindingprotein of unknown function from maize (Reddy et al.,1993).Their research focused on clone HvCBTl whichhad high homologuey to the inward rectifying potassiumchannel KATl from Arabidopsis and to catfish olfactorychannels (Anderson et al., 1992; Goulding et al., 1992).Overall similarity between HvCBTl and KATl and theolfactory channel was 46% and 47%, respectively. Theseion channels are characterized by six transmembranedomains, a pore between the fifth and sixth transmem-brane domain and a cyclic nucleotide binding domain inthe carboxyl-terminus. The olfactory channels have aCaM binding domain near the N-terminus. Homologueywith KATl and olfactory channels is limited in the poredomain. This suggests that HvCBTl may be a memberof a new class of ion channels from plants, but thishypothesis must be confirmed by a functional assay.

1346 Bethke et al.

Olfactory channels are able to respond to a wide concen-tration of cAMP and cGMP in vitro (between I fxM and1 mM), and their responsiveness to these two cyclicnucleotides decreases dramatically in the presence of CaM(Liu et al., 1994). This suggests a mechanistic basis forthe wide range of GA concentrations (10~12 to 10~7,Hooley, 1982) to which aleurone cells respond.

Cyclic GMP may be a transient second messenger inGA signalling

It is clear that signal transduction in the aleurone cell isnot dependent solely on [Ca2+], and the amount of CaM.Other signal transduction chains and other signallingmolecules are required. One of these signalling moleculesmay be cGMP. Penson et al. (1996) present compellingevidence that cGMP is required for barley aleurone layersto respond fully to GA. When aleurone layers weretreated with 5 fiM GA3, a transient increase in cGMPwas measured using a radio-immunoassay. MaximalcGMP content occurred 2 h after hormone treatment andwas about 3-fold higher than that of controls. cGMPlevels then decreased, so that 4 h after hormone treatmentthey were equal to those in untreated controls. An inhib-itor of guanylyl cyclase, LY-83583, reduced this transientincrease in cGMP when applied at 200 jiM. Accumulationof a-amylase and GAMYB mRNAs and secretion of a-amylase protein were inhibited by LY-83583. Inhibitionof a-amylase secretion by LY-83583 was observed atconcentrations greater than 100 ^M with an IC50 of120 fiM. The membrane permeant cGMP analogue dibu-tyryl cGMP at 400 / M nearly restored a-amylase secretionto control levels in the presence of 120/xM LY-83583.Dibutyryl-cAMP at the same concentration did notreverse the effects of LY-83583, suggesting a specificcGMP effect. Dibutyryl-cGMP in the absence of GA didnot induce the accumulation of a-amylase mRNA or thesecretion of a-amylase protein. ABA-treatment of aleur-one layers did not change the cGMP content of the layers,and LY-83583 had a relatively small effect on the accumu-lation of Rab21 mRNA in layers treated with ABA.Taken together, these data led the authors to concludethat cGMP plays a key role in GA- but not ABA-signalling in barley aleurone cells (Penson et al., 1996).cGMP-binding proteins involved in GA-signalling areunknown, although HvCBTl may bind cGMP near itscarboxyl-terminus.

Cytosolic pH may function as a second messenger forhormonal signals

The concentration of protons in the cytosol may functionas a second messenger in plant signal transduction inmuch the same way that the concentration of calciumions does. Cross-talk between signal transduction chainsutilizing pH ; and [Ca2+]i as second messengers has been

proposed (Felle, 1988). In barley aleurone, changes inpH, resulting from treatment with ABA or GA have beenreported. Van der Veen et al. (1992) determined pH, ofaleurone protoplasts using a null-point method. Inthis method, detergent permeabilization of protoplastsreleases cytosolic contents into weakly buffered media.The pH of the medium is measured before and afterdetergent addition. The procedure is repeated severaltimes using media buffered to pHs that span pH,. Whenmedium pH is unchanged by detergent permeabilizationof cells, the pH of the buffered medium is taken asrepresentative of pH,. Within 1 h of ABA-treatment(5^M) , pH; of barley aleurone protoplasts increased byapproximately 0.14 pH unit (Van der Veen et al., 1992).GA3 treatment (10 ^M) of protoplasts had the oppositeeffect on pH;. Within 1 h of GA3 addition, pHs decreasedapproximately 0.18 pH unit (Heimovaara-Dijkstra et al.,1994a). Addition of both ABA and GA3 to the protoplastincubation medium resulted in an intermediate change inpH; (Heimovaara-Dijkstra et al., 1994a). Changes in pH ;

were transient. Within 6 h of treatment with either ABAor GA, pHj returned to control levels.

Across a wide range of ABA concentrations, dose-response curves for cytosolic alkalization and Rab-16mRNA accumulation in aleurone cells were similar (Vander Veen et al., 1992). When protoplasts were treatedwith the weak acids 5,5-di-methyl-2,4-oxazolidinedioneor potassium propionate, treatments thought to decreasepHj, accumulation of Rab-16 mRNA was greatly reduced.Incubation with the weak bases methylamine or NH4C1had little effect. The authors suggest that alkalization ofthe cytosol is necessary but not sufficient for ABA-regulated gene expression. Confirmation of these datawith a less invasive method of pH measurement is eagerlyawaited. At present, no cytosolic targets for pH-mediatedsignal transduction have been identified in aleurone cells.

Protein phosphorylation is required for GA and ABAsignalling in aleurone cells

Regulation of cellular activities by reversible proteinphosphorylation is a ubiquitious process in animals andplants. In cereal aleurone, phosphorylation is importantin the signalling pathways linked to GA and ABA percep-tion. Kuo et al. (1996), found that okadaic acid (OA),an inhibitor of protein phosphatases types 1 and 2A,prevented GA-responses when applied to wheat aleuronelayers at 100 nM. In these experiments, [Ca2+], did notincrease, a-amylase secretion was inhibited and cell deathinduced by GA was prevented. OA decreased accumula-tion of a-amylase mRNA in GA-treated layers by 96%.In ABA-treated layers, OA decreased accumulation ofPHAV1 mRNA by 64%. In similar experiments with theserine/threonine protein kinase inhibitors staurosporine(50 jiM) and K252a (50/*M), no reduction in a-amylasesecretion was observed (Kuo et al., 1996).

The effects of protein phosphatase inhibitors on aleur-one cell signalling have also been characterized byHeimovaara-Dijkstra et al. (1996). Of the six inhibitorsthey tested, phenylarsine oxide (PAO), calyculin A (CA)and OA inhibited ABA-induced gene expression.Treatment with PAO, CA or OA resulted in increasedphosphorylation of two 40 kDa proteins. This hyperphos-phorylation coincided with an increase in tyrosinephosphorylation of two 40 kDa proteins as detected withanti-tyrosine antibodies.

Transduction of extracellular signals in higher eukary-otes is accomplished by phosphorylation cascades and byphosphorylation/dephosphorylation of effector molecules.Phosphorylation cascades, as exemplified by the Ras/MEK/MAPK kinase cascade, operate by sequentialphosphorylation of several protein kinases (Johnsonand Vaillancourt, 1994). Phosphorylation activates eachkinase and promotes phosphorylation of the next kinasein the chain. Each kinase may also phosphorylate effectormolecules. For example, MAP kinase phosphorylatesprotein kinases, transcription factors, components ofthe cytoskeleton and other enzymes (Johnson andVaillancourt, 1994). In this way, phosphorylation pro-motes branching of the signal transduction pathway andleads to co-ordination of the response throughout the cell.

Knetsch et al. (1996) demonstrated that ABA-treat-ment of barley aleurone protoplasts increased MAPkinase activity. Activity in immunoprecipitates, as meas-ured by the phosphorylation of myelin basic protein, wasstimulated more than 4-fold by ABA. Activity was max-imal about 3 min after hormone treatment, and returnedto basal levels after approximately 5 min. When anti-phosphotyrosine antibodies rather than antibodies to theMAP kinase ERK1 were used to immunoprecipitateTriton X-100- and SDS-soluble proteins, the time-coursefor MAP kinase activity in the precipitated material wassimilar. Experiments with the protein tyrosine phosphat-ase inhibitor PAO indicated that tyrosine dephosphoryl-ation was required for ABA-activation of MAP kinaseand for ABA-induced accumulation of rabl6 mRNA(Knetsch et al., 1996). The authors suggest that ABAinduces tyrosine phosphorylation of MAP kinase duringMAP kinase activation. They further propose that activityof a protein tyrosine phosphatase is part of the signaltransduction chain linking ABA preception to MAPkinase activation and ABA-inducible gene expression(Fig. 4). The effect of GA-treatment on MAP kinaseactivity is unclear. No effect was observed by Knetschet al. (1996), but GA reduced the accumulation of mRNAfor a MAP kinase homologue in oat aleurone layers(1995).

A putative ribosomal kinase (Askl 1) has been clonedfrom oat aleurone by Huttly and Phillips (1995).Accumulation of transcripts for this kinase in GA-treatedaleurone layers was substantially higher than in layers

Signalling in cereal aleurone 1347

incubated in the absence of hormone. The authors specu-late that Askl I may phosphorylate ribosomal proteinsand that this may play a role in GA-stimulated proteinsynthesis.

Endomembrane targets for hormonally inducedsignals

The evidence presented above suggests that [Ca2+],, CaM,cGMP, pHb protein kinases, and protein phosphatasesare all players in the regulation of hormonally inducedresponses within the aleurone cell. To bring about theappropriate response, however, each of these signal trans-duction intermediates must interact with one or moreeffector molecules. Because the primary function of thealeurone cell is the synthesis and secretion of hydrolases,processes which require co-ordination of activities withinthe PSV, ER and Golgi apparatus, some of the signaltransduction chains leading from GA- or ABA-perceptionare linked to the endomembrane system. As illustrated inFig. 5, transporters in the ER and PSV and PSV proteasesare likely targets for these signals.

Calcium transport into the endoplasmic reticulum is via a

CaM-stimulated Ca2+-ATPase

a-Amylase is a Ca-containing metalloprotein that is pro-duced at high rates on the rough ER of aleurone cells.

H+-PPase H+-ATPasePM

rotein kinase

Ca2 +-channel

. ^ A m y i a s e~C> secretion

Fig. 5. Endomembrane targets of hormonal signalling in cereal aleurone.GA stimulates acidification of the PSV lumen and increases the activityof PSV proteases. The second messenger Ca2 + -CaM stimulates theactivity of an ER Ca2 + -ATPase and regulates the activity of the SVchannel. Tonoplast-associated protein phosphatase and protein kinasealso modulate the activity of this channel.

1348 Bethke et al.

Replenishment of Ca2+ within the lumen of the ER,therefore, is required for continued amylase synthesis(Bush et al., 1989). Bush et al. (1993) measured rates ofCa2 +-transport (CaT) into microsomal vesicles frombarley aleurone layers and found that the rate of CaTwas greatly increased by GA3-treatment. Maximal ratesof transport were found in membrane fractions that hadthe highest rates of cytochrome c reductase activity,indicating that the transporter was likely to be in the ER.Calcium transport was significantly reduced by vanadate,but only minimally reduced by NO^~. This suggested tothe authors that the transporter was a P-type Ca2 +-ATPase. When layers were treated with ABA, there wasless CaT in all membrane fractions than when layers weretreated with GA3. For layers incubated in both 5 fiMGA3 and 5 MM ABA, or 5 ^M GA3 and then 5 ^M ABA,GA3-stimulated CaT was eliminated. This antagonisticeffect of ABA on CaT was not the result of reduced ERmembrane or ER protein. Since CaT increased for atleast 16 h following GA3-treatment, the authors proposedthat regulation was at the level of gene expression. Theynoted that their data were consistent with synthesis ofeither new Ca2 +-transporters or of a regulatory proteinsuch as CaM.

Subsequently, Gilroy and Jones (1993) showed thatCaT could be stimulated at least 5-fold by adding 500 nMCaM to ER microsomes isolated from non-GA-treatedbarley aleurone layers. No stimulation was observed whenCaM was added to microsomes from GA-treated layers.The CaM inhibitor W7, however, reduced the rate ofCaT into microsomes from GA-treated layers. This sug-gested that CaM or a CaM-like protein remained boundto ER microsomes and was sufficient for full activationof a Ca2 + -ATPase. A calcium transporter in wheat similarto the barley ER Ca2 + -ATPase has been characterized byBush and Wang (1995). As in barley, CaT into isolatedmembrane vesicles from GA-treated aleurone layers wasnot stimulated by exogenous CaM, but was inhibited byW7. Calcium transport activity in barley ER increased ata nearly constant rate for at least 16 h followingGA-treatement (Bush et al., 1993). The kinetics of thisrise in CaT activity are similar to those for CaM protein(Schuurink et al., 1996). It is likely, therefore, thatGA-stimulation of ER CaT is mediated by CaM.

Effector molecules that respond to hormonal signals arepresent on the tonoplast of aleurone PSVs

The PSV in cereal aleurone contains storage proteins andmineral nutrients. Following imbibition, protein reservesare hydrolysed by PSV proteases, and minerals such asK, Mg and P are released from phytin by phytases. Theproducts of these reactions are transported through thePSV tonoplast and are used for hydrolase synthesis andembryo nutrition. Evidence is accumulating that processes

within the PSV are co-ordinated with those in the cyto-plasm by hormonal signals (Fig. 5). The activity of cyst-eine proteases within the PSV, for example, is upregulatedby GA3-treatment of barley aleurone protoplasts (Bethkeet al., 1996). Using proteolytic activity gels, Bethke et al.showed that incubation of protoplasts in GA3 (5 pM) for18 h increased the activity of at least three cysteineproteases. Two of these activities were undetectable inABA-treated protoplasts and the third was present butreduced. No proteolysis was detectable when activity gelswere incubated in pH 6.5 buffer, but two of the cysteineproteases were active at pH 5.5 and all three at pH 4.5(Bethke and Jones, unpublished data).

Because these PSV proteases are inactive at neutralpH, acidification of the PSV lumen may be an importantregulator of storage protein mobilization. Swanson andJones (1996) measured the pH of barley aleurone PSVsusing the pH-sensitive dye BCECF-AM. They showedthat incubation of protoplasts with GA3 for 20 h promotesacidification of the PSV lumen. The pH of untreatedPSVs was 6.6, whereas that of PSVs in GA3-treatedprotoplasts was 5.8. This may underestimate the degreeof acidification as BCECF is insensitive to pH belowabout 5.5. The onset of vacuolar acidification beganshortly after GA3-treatment. Measurements of in vivovacuolar pH were made before and after GA3 additionto the media surrounding aleurone protoplasts. Within2 h the pH within PSVs had decreased. ABA treamentdid not result in vacuolar acidification. The pH of PSVsin protoplasts treated with ABA for 20 h was 6.7. Howthe process of vacuolar acidification is regulated is notknown. Immunoblot analyses suggested that differencesin abundance of tonoplast H +-pumps between PSVs inABA- or GA-treated protoplasts are unlikely to accountfor differences in vacuolar pH. Swanson and Jones (1996)demonstrated that approximately equal amounts ofV-type H+-ATPases and H + -PPiases were present in thetonoplast of PSVs from both ABA- and GA-treatedprotoplasts.

Another target of Ca2 + -CaM in the aleurone cell is theslow-vacuolar (SV) channel. The SV channel is an abund-ant cation channel that may play an important role inCa2+-signalling (Ward and Schroeder, 1994). In barleyaleurone PSVs, Bethke and Jones (1994) have shown thatthe activity of the SV channel is modulated by Ca2 +-CaM. Using the patch-clamp technique they demon-strated that SV channel activity was stimulated by increas-ing external free Ca2 + . Activity of the channel increasedas Ca2+ was increased above 600 nM. At high Ca2 +

(100fiM) the CaM inhibitors W7 and trifluoperizinereduced SV channel activity by more than 90%. Channelactivity inhibited by W7 could be partially restored byaddition of CaM. CaM on its own increased the sensitivityof the channel to Ca2 + . In the presence of CaM, specificcurrent through SV channels increased ~ 3-fold at 2.5 fxM

Ca2+ and over 13-fold at 10 MM Ca2 + . In the absence ofadded CaM, whole-vacuole SV currents in PSVs fromABA-treated protoplasts were 58% less than those inPSVs from GA-treated protoplasts. It is likely that thisrepresents differences in the regulatory environmentwithin the cells from which PSVs where isolated ratherthan differences in channel density within the tonoplast.

Activity of the SV channel is also regulated by reversibleprotein phosphorylation (Bethke and Jones, 1997). Whenisolated PSVs were treated with the protein phosphataseinhibitor OA (20 nM), whole-vacuole channel activitywas reduced by 60%. When the OA-insensitive proteinphosphatase calcineurin was added to PSVs in the pres-ence of OA, activity was restored to control levels. Thesedata suggested that protein dephosphorylation activatedthe SV channel. Bethke and Jones (1997) also found thatprotein phosphorylation could activate the channel.Purified CaM-like domain protein kinase (CDPK) stimu-lated channel activity at both low (200 ^M) and high(2mM) ATP concentrations. A model for regulation ofthe barley aleurone SV channel in which phosphorylationoccurs at two sites was proposed. Phosphorylation at onesite promotes channel opening and phosphorylation at asecond site promotes channel closure. At least one of theprotein phosphatases and one of the protein kinases thatmodulate SV channel activity were present in isolatedPSVs.

The aleurone cell nucleus contains CaM

Although honnonally-induced changes in mRNA syn-thesis are well characterized for the aleurone cell (seebelow), only tentative links between hormone perceptionand gene expression have been established. One of thesemay be CaM. Immunolocalization of CaM within barleyaleurone protoplasts suggests that the nucleus contains asignificant pool of CaM (Schuurink et al., 1996). Thefunction of this nuclear CaM is unknown, but work inRaymond Zielinski's laboratory suggests that it maymodulate gene expression. Szymanski et al. (1996) foundat least four DNA-binding proteins in cauliflower nuclearextracts that also bound CaM. Significantly, CaMenhanced the binding of these proteins to the C/G-boxin the Arabidopsis Cam-3 promoter. The Arabidopsistranscription factor TGA3 also bound CaM and thisenhanced its binding to the Cam-3 promoter. These datasuggest a mechanism whereby changes in Ca2+ or CaMcan be transduced into changes in gene expression.

Hormonal signalling results in GA- and ABA-specific gene expression

The responses of aleurone cells to GA and ABA describedabove precede increases in a-amylase mRNAs (see Fig. 2)and increases in ABA-up-regulated mRNAs. To address

Signalling in cereal aleurone 1349

the question of how mRNA levels of GA and ABAinducible genes are regulated in aleurone cells, their genesand promoters have been isolated. Promoter-reporterconstructs have been prepared and transiently expressedin aleurone protoplasts and aleurone layers. Transactingfactors that interact with specific elements in these pro-moters or that regulate the expression of these promotershave been identified. The data from these experi-ments using cereal aleurone have significantly increasedour understanding of phytohormone-regulated geneexpression.

Multiple regulatory elements within the promoters ofGA- and ABA-mducible genes confer hormone specificexpression

Of the genes that are regulated by GA and ABA, the bestcharacterized are those for a-amylases. A key regulatorystep in the production of a-amylase protein is synthesisof a-amylase mRNAs. Run-on transcription studies withnuclei from barley and wild oat aleurone protoplasts andtransient expression of promoter-reporter constructs inaleurone protoplasts and layers suggest that GA andABA control a-amylase synthesis at the transcriptionallevel (Jacobsen and Beach, 1985; Zwar and Hooley,1986). A role for additional translational control has notbeen excluded.

The promoters of hormone-inducible genes containmultiple sequence elements that act together to conferhigh levels of hormone-dependent transcription (Fig. 6).Regions essential for hormone regulated expression in thepromoters of barley and wheat a-amylase genes have beenidentified using deletion and linker-scan analyses (Gublerand Jacobsen, 1992; Rogers and Rogers, 1992; Tregearet al., 1995; Lin et al., 1996). In the promoters of a-amylase genes from wheat (Tregear et al., 1995), rice(Tanida et al., 1994) and barley (Gubler and Jacobsen,1992), three highly conserved elements are present within300 bp of the transcription start site. These three ele-ments, the pyrimidine (C/TCTTTT), TAACAA/GA, andTATCCAC/T boxes are sufficient for GA-induction andwere named the GA-responsive complex (GARC) byGubler and Jacobsen (1992). Experiments using ricetransformed with the GUS reporter gene driven by the— 232 to +31 region of the RAmylA promoter confirmedthat these boxes are sufficient for developmental, temporaland hormonal regulation of the a-amylase promotor ingerminating grain (Itoh et al., 1995).

A limited number of gain-of-function experimentshave been done using the boxes conserved in thepromoters of GA-inducible genes. Skriver et al.(1991) showed that multiple copies of the sequenceGGCCGATAACAAACTCCGGCC attached to a min-imal 35S cauliflower mosaic virus promoter were ableto confer proper GA-responsive and ABA-repressive

1350 Bethke et al.

expression. This 21 base sequence was called the GA3

responsive element (GARE). These experiments suggestthat there is functional redundancy between the bindingsites on the GARE and those on the GARC. Remarkably,it is still unclear which nucleotides within the GARE areresponsible for ABA-repressible expression. It is evident,however, that the ACGT core which plays a dominantrole in ABA up-regulated gene expression (see below) isabsent from the GARE (Giraudat et al., 1994). Othergain-of-function experiments (Lanahan et al., 1992;Rogers and Rogers, 1992) showed that a single element(GTAACAGAGTCTGG) was able to direct GA-regulated transcription of the low p/ a-amylase promoter,but only when a second short sequence identifiedas O2S (Opaque-2 similar) was also present. ThisGTAACAGAGTCTGG element, which contains theTAACAGA sequence of the GARC is similar to asequence motif found in the GARE. This motif,UTAACAUANTCYGG, where U is a purine, Y is apyrimidine and N is any nucleotide, is highly conservedin cereal a-amylase gene promoters (Huang et al., 1990).The O2S box was termed a coupling element (CE)(Fig. 6). O2S in conjunction with another CE (box 5) ofthe low p/ a-amylase promoter (Fig. 6) increased thetranscription of a high p/ a-amylase promoter containingthe three-element GARC. These experiments indicate thatunidentified CEs in high p/ a-amylase promoters mayenhance transcription. A low p/ a-amylase promoter fromwheat appears to have regions upstream of the pyrimidinebox other than OS2 that are required for high levels ofGA-induced expression (Huttly et al., 1992).

Transcription of several ABA-inducible genes requirespromoter sequences containing an ACGT core sequence.This sequence is present in the promoters of HVAl andHVA22, two genes that are up-regulated by ABA inbarley aleurone layers (Shen and Ho, 1995; Shen et al.,1996). Loss- and gain-of-function promoter studies usingGUS as the reporter have shown that boxes containingthe ACGT core (ABA responsive element, ABRE) inHVAl and HVA22 are necessary for ABA-inducible geneexpression. ABREs, however, are not capable of directing

transcriptionstartI

3CEo-

CE

CCTTTTJ-]|TAACAGA

1GARE

1 |-|TATCCAT

JH>

GARC

Fig. 6. Diagrammatic representation of the promoter of a low p/a-amylase gene. The shaded boxes are collectively refered to as theGARC. The GARE is composed of the TAACAGA box of the GARCand additional bases to either side of it. Two CEs are shown upstreamof the GARC. The size and spacing of the boxes in the figure are notto scale.

ABA up-regulation by themselves; CEs are also neededfor ABA responsiveness. Two different ABA responsecomplexes (ABRCs), each containing an ABRE and aCE, were sufficient to confer ABA responsiveness to theHVAl and HVA22 promoters (Shen et al., 1996). Thepromoter of dehydrin, another ABA-up-regulated barleygene, also contains an element with an ACGT core, buta CE has not been identified in this promoter (Robertsonet al., 1995). Inconsistencies between transcriptional activ-ity of the dehydrin promoter and the accumulation ofdehydrin mRNA indicate that experiments that measurerun-on transcription and mRNA stability need to be donewith ABA-up-regulated genes in barley aleurone.

Transacting factors control expression of a-amylase genes

Several approaches have been used to search for trans-acting factors interacting with c/s-elements in a-amylasepromoters. Classic DNAsel footprinting and gel retard-ation studies identified sequence-specific interactionsbetween nuclear proteins and promoter regions of rice,wheat and barley a-amylases (Rushton et al., 1992; Sutliffet al., 1993; Goldman et al., 1994; Mitsunaga et al.,1994). In general, the DNAsel footprinting and gelretardation assays are consistent with the promoter-reporter studies described above. Sutliff ef al. (1993), forexample, demonstrated that a fraction from a nuclearprotein extract bound to the promoter of a barley low p/a-amylase gene at regions corresponding to the GAREand the TATCCAT elements. Significantly, this bindingactivity was shown to be induced by GA.

Another method used to identify trans-acting factorsconferring hormone-specific gene expression is to screenexpression cDNA libraries with promoter sequences.Rogers et al. (1995) screened expression cDNA librariesmade of mRNA from barley aleurone cells with hexamersof the GARE (Fig. 6). A cDNA encoding a novel 60-kDaprotein (GAB1) with an unusual, repeated zinc fingerdomain and a nuclear localization signal was cloned usingthis strategy (J Rogers and J Mundy, personal commun-ication). Within the zinc finger domains, Cys and Hisresidues required for zinc and DNA binding were deter-mined by in vitro mutagenesis and binding assays. Thenuclear localization signal was shown to function in onioncells. When transiently overexpressed in barley aleuronecells, GAB1 repressed expression of the low p/ a-amylasepromoter (J Rogers and J Mundy, personal communica-tion; Rogers et al., 1995). A similar approach was usedby Rushton et al. (1995) to clone two cDNAs encodingproteins that bind to the OS2 motif in the promoters ofwheat and wild oat low p/ a-amylase genes. These proteinseach contain a conserved region of 56-58 amino acidsthat contain a putative zinc finger. Binding to OS2 regionswas abolished by the chelators 1,10-o-phenanthroline andEDTA, which supports the hypothesis that these proteinsbind zinc in vivo.

MYB proteins are transcriptional activators that arefound in plants as well as animals. These proteinscontain conserved DNA-binding domains and a carboxyl-terminal activator domain. Binding of MYB proteins toDNA is sequence specific. Gubler et al. (1995) observedthat the TAACAAA element in the GARC has twopossible c- and v-MYB binding sites. They cloned a MYBcDNA from barley aleurone cells by screening an expres-sion cDNA library with a 170-bp probe containing thesequence coding for the DNA-binding domain of themaize Cl MYB protein. mRNA abundance for thisGAMYB increased following GA-treatment of barleyaleurone prior to the increase in a-amylase mRNA(Fig. 2). These data suggested that GAMYB may regulatethe transcription of a-amylase. To test this hypothesis,Gubler et al. (1995) transiently overexpressed GAMYBin barley aleurone layers. In the absence of added GA,GAMYB was able to stimulate transcription of a high p/a-amylase promoter fused to the GUS reporter gene.A TAACAAA-like sequence is also present in theGA-regulated promoter of barley 1-3,1-4 /9-glucanaseisoenzyme II (Wolf, 1992), suggesting that its promotermay be regulated by GAMYB. Other GA-regulated genessuch as a cathepsin B-like gene from wheat (Cejudo et al.,1992), a carboxypeptidase gene from rice (Washio andIshikawa, 1992), a putative ubiquitin-conjugating enzymefrom rice (Chen et al., 1995), and two cysteine proteinasegenes from barley (Mikkonen et al., 1996), however, donot appear to have a TAACAAA box in their promoters.Transacting factors other than MYB, therefore, mayactivate these promoters. This would allow for temporalor stimulus-specific control of GA-regulated transcrip-tion. Alternatively, other related sequences may functionas GAMYB-binding sites.

Recently a putative ubiquitin-conjugating enzyme(UBC) from rice was shown to transactivate the riceOsamy-C promoter (Chen et al., 1995). Osamy-C isinduced by GA, but promoter deletions revealed a domainbetween -230 and - 158 that repressed this GA-inducibleactivity (Chen et al, 1995). Co-expression of Osamy-Cwith UBC overcame this repression, suggesting that the— 230/—158 domain of Osamy-C binds a repressor pro-tein that UBC targets for proteolysis. The role of UBCin the GA-inducibility of the full length Osamy-Cpromoter, however, has not been investigated. Thisrole might be minor since repression attributed to the— 230/—158 region was masked by upstream promoterdomains.

The region of the Osamy-C promoter between —230and —158 has been shown to interact with at least fiveproteins from imbibed rice grain (Kim et al., 1992). Twoof these are present in rice aleurone independent of GAtreatment. The —230/—158 region also contains asequence homologous to the cAMP responsive elementof cAMP regulated promoters in mammalian cells

Signalling in cereal aleurone 1351

(Deutsch et al., 1988). A similar element was found in ahigh p/ a-amylase promoter and shown to repressGA-inducible transcription in barley (Gubler andJacobsen, 1992). No proteins have been identified thatbind to either of these cAMP-like responsive elements.

Figure 7A shows a model of how the transacting factorsdescribed above might influence the transcription of a-amylase genes. In this model, GA increases the amountof GAMYB protein. GAMYB then binds to the GAREand increases transcription of a-amylase genes. GAB1,the other transacting factor that binds to the GARE,inhibits transcription of a-amylase genes in the absenceof GA. When present, VP1 increases the affinity of GAB1for the GARE and represses a-amylase gene expression(Hoeckerera/., 1995). How GAMYB and GAB 1 competefor the GARE and how this affects the activity of thea-amylase promoter remain unanswered.

Transacting factors have been identified that regulate theexpression of ABA-inducible genes

ABA may act on the ABRE in the GARE via a vivipar-ousl (VP1) homologue. VP1 is a transcriptional activatorof maize that prevents precocious germination. AVP1-homologue has been cloned from rice (OSVP1) andits transcript was detected in developing seeds andimbibed mature embryos (Hattori et al., 1994). Thepresence of OSVP1 in the aleurone layer, however, wasnot determined and this remains an important issue. VP1transiently expressed in barley aleurone cells repressedthe GA-stimulated up regulation of a 1.8 kbp high p/ a-amylase promoter (Hoecker et al., 1995). VP1 enhancedthe binding of many transcription factors to elementswith an ACGT core (Hill et al., 1996). It is suggested, asindicated in Fig. 7A, that a VPl-like transcription factormight increase the affinity of GAB 1 for the GARE.

Most efforts to clone transacting factors interactingwith cu-elements in ABA upregulated promoters havefocused on the Em gene of wheat. A cDNA encodingEmBP-1, a protein with affinity for the ABRE in the Empromoter, was cloned from a wheat embryo cDNA library(Guiltinan et al., 1990). The maize VP1 gene productenhanced binding of EmBP-1 to its binding site (Hillet al., 1996). This is consistent with the observations thatVP1 is able to activate the Em promoter in vivo. Recentlyit was shown that VP1 can also transactivate the ABRCin HVA1, but not the ABRC in HVA22, suggesting thatVP1 can differentiate between ABRCs in mediating anABA response (Shen et al., 1996). OSVP1 is able totransactivate OSEm, a rice gene homologous to the wheatEm gene (Hattori et al., 1995). Hatorri's results stronglysuggest that OSVP1 activated the OSEm promoter byinteracting with a protein, perhaps an EMBP-1 homo-logue, that binds to an ACGT containing ABRE. It isvery likely that homologues of EmBP-1 also exist in

1352 Bethke et al.

+GA

Rapid transcriptionof a-amylase genes

GARE

• GA

Little transcriptionof a-amylase genes

GARE

+GA, VP1 present

Little transcriptionof a-amylase genes

GARE

B+ABA

Transcription ofABA-upregulatedgenes

CE

ABRE

Fig. 7. Regulation of GA- (A) and ABA-inducible promoters (B) bytransacting factors. (A) Binding of GAB1 to the a-amylase GARErepresses expression of a-amylase genes in the absence of GA. GAreduces this repression by stimulating the production of the transactiv-ator GAMYB. Binding of GAB1 to the GARE may be enhanced byVP1. (B) Transcription of ABA-up-regulated genes such as the wheatEm gene is promoted by the transactivator EMBP-1. VP1 may promotetranscription by interacting with EMBP-1, and perhaps with CEs.

barley. One strategy for identifying such homologuesfrom barley aleurone would be to screen for proteins thatinteract both with the ACGT cores in HVA1 and withVP1. Figure 7B shows a model of how transacting factorsmight regulate the transcription of an ABA-up-regulatedgene such as the wheat Em gene. In this model, VP1binds to EMBP-1 and increases its affinity for the ABRE.Binding of EMBP-1 stimulates transcription of the gene.Interaction between VP1 and a CE may stabilize theVP1/EMBP-1/ABRE complex.

The mechanism of VP1 -mediated activation ofABA-inducible genes is clearly different from thatof VP1-mediated repression of a-amylase genes. TheN-terminal domain of VP1 is required for activation ofthe Em gene, whereas domains in the middle part of VP1are required for repression of a-amylase genes. Repressionof a-amylase expression is ABA-independent, but VP1and ABA have a synergistic effect on Em expression. Itwould be interesting to determine the regions inGA-regulated promoters that are required for regulationby VP1. Without doubt it is necessary to determinewhether VP1 functions in vivo both during seed develop-ment and during the transition to seed germination.

Enormous progress has been made in understandinghormone-regulated gene expression by using aleuronecells as a model system. It is clear that the timing andlevel of expression of GA-inducible genes is reflected inthe composition of their promoters. Some transactingfactors might be master regulatory proteins for entireclasses of GA-inducible genes. ABA-inducible promoterscontain their own specific promoter elements and severalappear to be differentially regulated by VP1. In parallelwith ongoing studies to identify additional transactingfactors, it is now possible to study interactions betweenpreviously identified transacting factors and other nuclearor cytosolic proteins. This may provide a link betweenthe signal transduction components described above andproteins which directly regulate transcription.

Concluding remarks

Two centuries of biochemical and ultrastuctural researchon the cereal aleurone have provided the framework forestablishing hormonal signalling pathways. Many of theresponses of the cereal aleurone to GA and ABA havebeen characterized. Receptors for GA and ABA havebeen localized, and efforts are underway to isolate them.Components of the signal transduction chains linkinghormone perception to cellular response have been identi-fied, and more are sure to be found. The effects ofhormones on gene expression are becoming well under-stood. Much has been learned about hormonal signallingin cereal aleurone, but much more remains unknown.Additional signalling molecules must be identified in orderto link hormone receptors to their downstream effectors

securely. The number of signalling pathways and theextent of cross-talk between them must be determined,and links must be found which join cytosolic signals totranscriptional events. Only when the interplay betweensignalling molecules, signalling pathways and downstreameffectors is known, will it be understood how the activitiesof the aleurone cell are differentially regulated in responseto hormone perception.

Acknowledgements

We would like to thank Eleanor Crump for assisting us in thepreparation of this manuscript. The work of the authors hasbeen supported by the National Science Foundation and theUnited States Department of Agriculture.

References

Anderson BE, Ward JM, Schroeder JI. 1994. Evidence for anextracellular reception site for abscisic acid in Commelinaguard cells. Plant Physiology 104, 1177-83.

Anderson JA, Huprikar SS, Kochian LV, Lucas WJ. 1992.Functional expression of a probable Arabidopsis thalkmapotassium channel in Saccharomyces cerevisiae. Proceedingsof the National Academy of Sciences, USA 89, 3736-40.

Beale MH, Hooley R, MacMillan J. 1985. Gibberellins:structure-activity relationships and the design of molecularprobes. In: Bopp M, ed., Plant growth substances 1985.Twelfth international conference on plant growth substances.Heidelberg: Springer-Verlag, 65-73.

Bethke PC, Gilroy S, Jones RL. 1995. Calcium and planthormone action. In: Davies PJ, ed. Plant hormones.Physiology, biochemistry and molecular biology. Dordrecht,The Netherlands: Kluwer Academic Publishers, 298-317.

Bethke PC, Hillmer S, Jones RL. 1996. Isolation of intactprotein storage vacuoles from barley aleurone. PlantPhysiology 110, 521-9.

Bethke PC, Jones RL. 1994. Ca2 + -calmodulin modulates ionchannel activity in storage protein vacuoles of barley aleuronecells. The Plant Cell 6, 277-85.

Bethke PC, Jones RL. 1997. Reversible protein phosphorylationregulates the activity of the slow-vacuolar ion channel. ThePlant Journal (in press).

Brown HT, Morris GH. 1890. Researches on the germinationof some of the Graminae. Journal of the Chemical Society,London 57, 458-528.

Brown PH, Ho T-HD. 1986. Barley aleurone layers secrete anuclease in response to gibberellic acid. Plant Physiology82, 801-6.

Bush DS. 1995. Calcium regulation in plant cells and its role insignaling. Annual Review of Plant Physiology and PlantMolecular Biology 46, 95-122.

Bush DS. 1996. Effects of gibberellic acid and environmentalfactors on cytosolic calcium in wheat aleurone cells. Planta199, 89-99.

Bush DS, Biswas AK, Jones RL. 1993. Hormonal regulation ofCa2+ transport in the endomembrane system of the barleyaleurone. Planta 189, 507-15.

Bush DS, Cornejo M-J, Huang C-N, Jones RL. 1986. Ca2 +-stimulated secretion of a-amylase during development inbarley aleurone protoplasts. Plant Physiology 82, 566-74.

Bush DS, Hedrich R, Schroeder JI, Jones RL. 1988. Channel-

Signalling in cereal aleurone 1353

mediated K+ flux in barley aleurone protoplasts. Planta176, 368-77.

Bush DS, Sticher L, Van Huystee R, Wagner D, Jones RL.1989. The calcium requirement for stability and enzymaticactivity of two isoforms of barley aleurone a-amylase. TheJournal of Biological Chemistry 264, 19392-8.

Bush DS, Wang T. 1995. Diversity of calcium-efflux transportersin wheat aleurone cells. Planta 197, 19-30.

Cejudo FJ, Ghose TK, Stabel P, Baulcombe DC. 1992. Analysisof the gibberellin-responsive promoter of a cathepsin B-likegene from wheat. Plant Molecular Biology 20, 849-56.

Chandler PM, Robertson M. 1994. Gene expression regulatedby abscisic acid and its relation to stress tolerance. AnnualReview of Plant Physiology and Plant Molecular Biology45, 113-41.

Chandler PM, Zwar JA, Jacobsen JV, Higgins TJV, Inglis AS.1984. The effect of gibberellic acid and abscisic acid on a-amylase mRNA levels in barley aleurone layers: studies usingan a-amylase cDNA clone. Plant Molecular Biology 3,407-18.

Chen X, Wang B, Wu R. 1995. A gibberellin-stimulatedubiquitin-conjugating enzyme gene is involved in a-amylasegene expression in rice aleurone. Plant Molecular Biology29, 787-95.

Chrispeels MJ. 1991. Sorting of proteins in the secretory system.Annual Review of Plant Physiology and Plant MolecularBiology 42, 21-53.

Chrispeels MJ, Tenner AJ, Johnson KD. 1973. Synthesis andrelease of sucrose by the aleurone layer of barley: regulationby gibberellic acid. Planta 113, 35-46.

Cornejo MJ, Platt-aloia KA, Thomson WW, Jones RL. 1988.Effects of GA3 and Ca2+ on barley aleurone protoplasts: afreeze-fracture study. Protoplasma 146, 157-65.

Deikman J, Jones RL. 1985. Control of a-amylase mRNAaccumulation by gibberellic acid and calcium in barleyaleurone layers. Plant Physiology 78, 192-8.

Deutsch P, Hoeffler JP, Jameson JL, Habener JF. 1988. CyclicAMP and phorbol ester-stimulated transcription mediated bysimilar DNA elements that bind distinct proteins. Proceedingsof the National Academy of Sciences, USA 85, 7922-6.

Drozdowicz YM, Jones RL. 1995. Hormonal regulation oforganic and phosphoric acid release by barley aleurone layersand scutella. Plant Physiology 108, 769-76.

Felle H. 1988. Cytoplasmic free calcium in Riccia fluitans L.and Zea mavs L.: interaction of Ca2+ and pH? Planta176, 248-55. '

Fernandez DE, Staehelin LA. 1985. Structural organization ofultrarapidly frozen barley aleurone cells actively involved inprotein secretion. Planta 165, 455-68.

Fincher GB. 1989. Molecular and cellular biology associatedwith endosperm mobilization in germinating cereal grains.Annual Review of Plant Physiology and Plant MolecularBiology 40, 305-46.

Fischer EH, Stein EA. 1960. a-Amylases. In: Boyer PD, LardyH, MyrbSck K, eds. The enzymes, 4. New York: AcademicPress, 313-43.

Gilroy S. 1996. Signal transduction in barley aleurone proto-plasts is calcium dependent and independent. The Plant Cell8, 2193-209.

Gilroy S, Jones RL. 1992. Gibberellic acid and abscisic acidco-ordinately regulate cytoplasmic calcium and secretoryactivity in barley aleurone protoplasts. Proceedings of theNational Academy of Sciences, USA 89, 3591-5.

Gilroy S, Jones RL. 1993. Calmodulin stimulation of unidirec-tional calcium uptake in the endoplasmic reticulum of barleyaleurone. Planta 190, 289-96.

Gilroy S, Jones RL. 1994. Perception of gibberellin and abscisic

1354 Bethke et al.

acid at the external face of the plasma membrane of barley(Hordeum vulgare L.) aleurone protoplasts. Plant Physiology104,1185-92.

Giraudat J, Parcy F, Bertaucbe N, Gosti F, Leung J, MorrisP-C, Bouvier-Durand M, Vartanian N. 1994. Current advancesin abscisic acid action and signalling. Plant Molecular Biology26, 1557-77.

Goldman S, Mawal YR, Tanida I, Wu R. 1994. Studies of agibberellin-dependent DNA-binding protein related to theexpression of a rice a-amylase gene. Plant Science 99, 75-88.

Goodwin PB, Carr DJ. 1972. The induction of amylase synthesisin barley aleurone layers by gibberellic acid. I. response totemperature. Journal of Experimental Botany 23, 1-7.

Goulding EH, Ngai J, Kramer RH, Colkos S, Axel R,Siegelbaum SA, Chess A. 1992. Molecular cloning and single-channel properties of the cyclic nucleotide-gated channelfrom catfish olfactory neurons. Neuron 8, 45-58.

Gro/3elindemann E, Graebe JE, Stockl D, Hodden P. 1991. ent-Kaurene biosynthesis in germinating barley (Hordeum vulgareL., cv. Himalaya) caryopses and its relation to a-amylaseproduction. Plant Physiology 96, 1099-104.

Gubler F, Jacobsen JV. 1992. Gibberellin-responsive elementsin the promoter of a barley high-p/ a-amylase gene. ThePlant Cell 4, 1435-41.

Gubler F, Kalla R, Roberts JK, Jacobsen JV. 1995. Gibberellin-regulated expression of a myb gene in barley aleurone cells:evidence for Myb transactivation of a high-p/ a-amylase genepromoter. The Plant Celll, 1879-91.

Guiltinan MJ, Marcotte WRJ, Quatrano RS. 1990. A plantleucine zipper protein that recognizes an abscisic acid responseelement. Science 250, 267-71.

Haberlandt G. 1890. Die KJeberschicht des Gras-Endospermsals Diastase ausscheidendes Drusengewebe. Benchte DeutscheBotanische Gesellschaft 8, 40-8.

Hattori T, Terada T, Hamasuna ST. 1994. Sequence andfunctional analyses of the rice gene homologous to themaize Vpl. Plant Molecular Biology 24, 805-10.

Hattori T, Terada T, Hamasuna ST. 1995. Regulation of theOsem gene by abscisic acid and the transcriptional activatorVP1: analysis of cw-acting promoter elements required forregulation by abscisic acid and VP1. The Plant Journal7,913-25.

Heck GR, Chamberlain AK, Ho T-HD. 1993. Barley embryoglobulin 1 gene, Beg]: Characterization of cDNA, chromo-some mapping and regulation of expression. Molecular andGeneral Genetics 239, 209-18.

Heimovaara-Dijkstra S, Heistek JC, Wang M. 1994a.Counteractive effects of ABA and GA3 on extracellular andintracellular pH and malate in barley aleurone. PlantPhysiology 106, 359-65.

Heimovaara-Dijkstra S, Nieland TJF, Van Der Meulen RM,Wang M. 1996. Abscisic acid-induced gene-expression requiresthe activity of protein (s) sensitive to the protein-tyrosinephosphatase inhibitor phenylarsine oxide. Plant GrowthRegulation 18, 115-23.

Heimovaara-Dijkstra S, Van Duijn B, Libbenga KR, HeidekampF, Wang M. 19946. Abscisic acid-induced membrane potentialchanges in barley aleurone protoplasts: a possible relevancefor the regulation of rab gene expression. Plant CellPhysiology 35, 743-50.

Hetherington AM, Quatrano RS. 1991. Mechanisms of actionof abscisic acid at the cellular level. New Phytologist 119,9-32.

Hill A, Nantel A, Rock CD, Quatrano RS. 1996. A conserveddomain of the viviparous-] gene product enhances the DNAbinding activity of the bZIP protein EmBP-1 and other

transcription factors. Journal of Biological Chemistry 271,3366-74.

Hillmer S, Gilroy S, Jones RL. 1993. Visualizing enzymesecretion from individual barley (Hordeum vulgare) aleuroneprotoplasts. Plant Physiology 102, 279-86.

Hoecker U, Vasil IK, McCarty DR. 1995. Integrated control ofseed maturation and germination programs by activator andrepressor functions of Viviparous-1 of maize. Genes andDevelopment 9, 2459-69.

Hong B, Uknes SJ, Ho T-HD. 1988. Cloning and characteriza-tion of a complementary DNA encoding a messenger RNArapidly-induced by ABA in barley aleurone layers. PlantMolecular Biology 11, 495-506.

Hooley R. 1982. Protoplasts isolated from aleurone layers ofwild oat (Avena fatua L.) exhibit the classic response togibberellic acid. Planta 154, 29-40.

Hooley R, Beale MH, Smitb SJ. 1991. Gibberellin perceptionat the plasma membrane of Avena fatua aleurone protoplasts.Planta 183, 274-80.

Hooley R, Beale MH, Smith SJ, Walker RP, Rushton PJ,Whitford PN, Lazarus CM. 1992. Gibberellin perception andthe Avena-fatua aleurone: do our molecular keys fit thecorrect locks? Biochemical Society Transactions 20, 85-9.

Huang AHC. 1992. Oil bodies and oleosins in seeds. AnnualReview of Plant Physiology and Plant Molecular Biology43, 177-200.

Huang N, Sutiiff TD, Litts JC, Rodriguez RL. 1990.Classification and characterization of the rice a-amylasemultigene family. Plant Molecular Biology 14, 655-68.

Huttly AK, Phillips AL. 1995. Gibberellin-regulated expressionin oat aleurone cells of two kinases that show homology toMAP kinase and a ribosomal kinase. Plant Molecular Biology27, 1043-52.

Huttly AK, Phillips AL, Tregear JW. 1992. Localisation of ciselements in the promoter of a wheat a-Amy2 gene. PlantMolecular Biology 19, 903-11.

Itoh K, Yamaguchi J, Huang N, Rodriguez RL, Akazawa T,Shimamoto K. 1995. Developmental and hormonal regulationof rice a-amylase (RAmylA)-gusA fusion genes in transgenicrice seeds. Plant Physiology 107, 25-31.

Jacobsen JV, Beach LR. 1985. Evidence for control oftranscription of a-amylase and ribosomal RNA genes inbarley aleurone protoplasts by gibberellic acid and abscisicacid. Nature 316, 275-7.

Johnson GL, Vaillancourt RR. 1994. Sequential protein kinasereactions controlling cell growth and differentiation. CurrentOpinion in Cell Biology 6, 230-8.

Jones HD, Desikan R, Plakidou-Dymock S, Hooley R. 1995.Ga and Gj3 subunit genes and signal transduction in A. fatuaaleurone. In: I5th International Conference on Plant GrowthSubstances, Minneapolis, Minnesota USA.

Jones HD, Smith S, Plakidou-Dymock S, Hooley R. 1996. Gaand G/3 subunit genes and signal transduction in Avena fatuaaleurone. Journal of Experimental Botany 47, Supplement 5.

Jones RL. 1969a. The effect of ultracentrifugation on finestructure and a-amylase production in barley aleurone cells.Plant Physiology 44, 1428-38.

Jones RL. 19696. Gibberellic acid and the fine structure ofbarley aleurone cells. I. Changes during the lag-phase ofa-amylase synthesis. Planta 87, 119-33.

Jones RL. 1972. Fractionation of the enzymes of the barleyaleurone layer: evidence for a soluble mode of enzyme release.Planta 103, 95-109.

Jones RL, Jacobsen JV. 1991. Regulation of synthesis andtransport of secreted proteins in cereal aleurone. InternationalReview of Cytology 126, 49-88.

Jones RL, Price JM. 1970. Gibberellic acid and the finestructure of barley aleurone cells. III. Vacuolation of thealeurone cell during the phase of ribonuclease release. Planta94, 191-202.

Kende H, Gardner G. 1976. Hormone binding in plants. AnnualReview of Plant Physiology 27, 267-90.

Kim JK, Cao J, Wu R. 1992. Regulation and interaction ofmultiple protein factors with the proximal promoter regionsof a rice high p / alpha-amylase gene. Molecular and GeneralGenetics 232, 383-93.

Knetsch ML, Wang M, Snaar-Jagalska BE, Heimovaara-Dijkstra S. 1996. Abscisic acid induces mitogen-activatedprotein kinase activation in barley aleurone protoplasts. ThePlant CellS, 1061-7.

Kuo A, Cappelluti S, Cervantes-Cervantes M, Rodriguez M,Bush DS. 1996. Okadaic acid, a protein phosphatase inhibitor,blocks calcium changes, gene expression, and cell deathinduced by gibberellin in wheat aleurone cells. The Plant Cell8, 259-69.

Lanahan MB, Ho T-HD, Rogers SW, Rogers JC. 1992. Agibberellin response complex in cereal a-amylase gene pro-moters. The Plant Cell 4, 203-11.

Leah R, Mundy J. 1989. The bifunctional a-amylase/subtilisininhibitor of barley: nucleotide sequence and patterns of seed-specific expression. Plant Molecular Biology 12, 673-82.

Lenton JR, Appleford NEJ, Croker SJ. 1994. Gibberellins anda-amylase gene expression in germinating wheat grains. PlantGrowth Regulation 15, 261-70.

Lin W, Gopalakrishnan B, Muthukrishnan S. 1996. Transcriptionefficiency and hormone-responsiveness of alpha-amylase genepromoters in an improved transient expression system frombarley aleurone protoplasts. Protopiasma 192, 93-108.

Liu M, Chen T-Y, Aharaed B, Li J, Yay K-VV. 1994. Calcium-calmodulin modulation of the olfactory cyclic nucleotide-gated cation channel. Science 266, 1348-54.

Marcotte WRJ, Bayley CC, Quatrano RS. 1988. Regulation ofa wheat promoter by abscisic acid in rice protoplasts. Nature335, 454-7.

Mikkonen A, Porali I, Cercos M, Ho T-hD. 1996. A majorcysteine proteinase, EPB, in germinating barley seeds: struc-ture of two intronless genes and regulation of expression.Plant Molecular Biology 31, 239-54.

Miller EC. 1931. Plant physiology. New York: McGraw-Hill.Mitsunaga S-I, Rodriguez RL, Yamaguch J. 1994. Sequence-

specific interactions of a nuclear protein factor with thepromoter region of a rice gene for a-amylase, RAmy3D.Nucleic Acids Research 22, 1948-53.

Mitsunaga S-I, Yamaguchi J. 1993. Induction of a-amylase isrepressed by uniconazole, an inhibitor of the biosynthesis ofgibberellin, in a dwarf mutant of rice, Waito-C. Plant andCell Physiology 34, 243-9.

Mundy J, Chua N-H. 1988. Abscisic acid and water-stressinduce the expression of a novel rice gene. EMBO Journal7, 2279-86.

Musgrave A, Kays SE, Kende H. 1972. Uptake and metabolismof radioactive gibberellins by barley aleurone. Planta 102,1-10.

Nadeau R, Rappaport L, Stolp CF. 1972. Uptake and metabolismof 3H -gibberellin A, by barley aleurone layers: response toabscisic acid. Planta 107, 315-24.

O'Neill SD, Keith B, Rappaport L. 1986. Transport of gibberellinA! in cowpea membrane vesicles. Plant Physiology 80, 812-17.

Okita TW, Rogers JC. 1996. Compartmentation of proteins inthe endomembrane system of plant cells. Annual Review ofPlant Physiology and Plant Molecular Biology 47, 327-50.

Signalling in cereal aleurone 1355

Paleg L. 1960a. Physiological effects of gibberellic acid. I. Onthe carbohydrate metabolism and amylase activity of thebarley endosperm. Plant Physiologv 35, 293-9.

Paleg L. 19606. Physiologicafeffects'of gibberellic aicd. II. Onstarch hydrolyzing enzymes of barley endosperm. PlantPhysiology 35, 902-6.

Payen A, Persoz J. 1833. Memoire sur la diastase, les principauxproduits de ses reactions et leurs applications aux artsindustriels. Annales de Chemie 53, 73-92.

Penson SP, Schuurink RC, Fath A, Gubler F, Jacobsen JV,Jones RL. 1996. cGMP is required for gibberellic acid-induced gene expression in barley aleurone. The Plant Cell8, 2325-33.

Radley M. 1959. The occurrence of gibberellin-like substancesin barley and malt. Chemistry and Industry Issue 27, 877-8.

Reddy ASN, Takezawa DHF, Fromm H, Poovaiah BW. 1993.Isolation and characterization of two cDNAs that encode forcalmodulin-binding proteins from corn root tips. PlantScience 94, 109-17.

Ried JL, Walker-Simmons MK. 1990. Synthesis of ABA-responsive, heat stable proteins in embryonic axes of dormantwheat grains. Plant Physiology 93, 662-7.

Ritchie S, Gilroy S. 1996. The regulation of the response ofbarley aleurone protoplasts to GA and ABA by elements ofthe phospholipid signal transduction pathway. PlantPhysiology 111, 153.

Roberts DM, Harmon AC. 1992. Calcium-modulated proteins:targets of intracellular signals in higher plants. Annual Reviewof Plant Physiology and Plant Molecular Biology 43, 375^414.

Robertson M, Cuming AC, Chandler PM. 1995. Sequenceanalysis and hormonal regulation of a dehydrin promoterfrom barley, Hordeum vulgare. Physiologia Plantarum 94,470-8.

Rogers JC, Rogers SW. 1992. Definition and functionalimplications of gibberellin and abscisic acid cw-acting hor-mone response complexes. The Plant Cell 4, 1443-51.

Rogers JC, Rogers SW, Mundy J. 1995. Biochemical andmolecular dissection of hormone regulated transcription frombarley. In: 15th International Conference on Plant GrowthSubstances, Minneapolis, Minnesota, USA. 23.

Rushton PJ, Hooley R, Lazarus CM. 1992. Aleurone nuclearproteins bind to similar elements in the promoter regions oftwo gibberellin-regulated a-amylase genes. Plant MolecularBiology 19,891-901.

Rushton PJ, Macdonald H, Huttly AK, Lazarus CM, Hooley R.1995. Members of a new family of DNA-binding proteinsbind to a conserved cw-element in the promoters of a-amy2genes. Plant Molecular Biology 29, 691-702.

Schroeder JI, Hagiwara S. 1990. Repetitive increases in cytosolicCa2+ of guard cells by abscisic acid activation of non-selective Ca2 + permeable channels. Proceedings of the NationalAcademy of Sciences, USA 87, 9305-9.

Schuurink RC, Chan PV, Jones RL. 1996. Modulation ofcalmodulin mRNA and protein levels in barley aleurone.Plant Physiology 111, 371-80.

Schuurink RC, Jones RL. 1995. A putative CaM and cyclic-nucleotide-regulated ion channel from barley. Journal ofCellular Biochemistry, Supplement 19A, 153.

Schuurink RC, van Duijn G, Heidekamp F. (1993). Is abscisicacid involved in barley aleurone dormancy. In: Walker-Simmons MK, Ried JL, eds. Pre-harvest sprouting in cereals.St. Paul: American Association of Cereal Chemists, 224-35.

Shen Q, Ho T-HD. 1995. Functional dissection of an abscisicacid (ABA)-inducible gene reveals two independent ABA-responsive complexes each containing a G-box and a novelcw-acting element. The Plant Cell 1, 295-307.

1356 Bethke et al.

Shen Q, Zhange P, Ho T-HD. 1996. Modular nature of abscisicacid (ABA) response complexes: composite promoter unitsthat are necessary and sufficient for ABA induction of geneexpression in barley. The Plant Cell 8, 1107-19.

Sinjorgo KMC, De Vries MA, Heistek JC, Van Zeijl MJ, VanDer Veen SW, Douma AC. 1993. The effect of external pHon the gibberellic acid response of barley aleurone. Journalof Plant Physiology 142, 506-9.

Skadsen R. 1993. Aleurones from a barley with low a-amylaseactivity become highly responsive to gibberellin whendetached from the starchy endosperm. Plant Physiology102, 195-203.

Skriver K, Olsen FL, Rogers JC, Mundy J. 1991. C/s-actingDNA elements responsive to gibberellin and its antagonistabscisic acid. Proceedings of the National Academy of Sciences,USA 88, 7266-70.

Stewart A, Nield H, Lott JNA. 1988. An investigation of themineral content of barley grains and seedlings. PlantPhysiology 86, 93-7.

SutliffTD, Lanahan MB, Ho T-HD. 1993. Gibberellin treatmentstimulates nuclear factor binding to the gibberellin responsecomplex in a barley a-amylase promoter. The Plant Cell5, 1681-92.

Swanson SJ, Jones RL. 1996. Gibberellic acid induces vacuolaracidification in barley aleurone. The Plant Cell 8, 2211-21.

Szymanski DB, Liao B, Zielinski RE. 1996. Calmodulin isoformsdifferentially enhance the binding of cauliflower nuclearproteins and recombinant TGA3 to a region derived fromthe Arabidopsis Cam-3 promotor. The Plant Cell 8, 1069-77.

Taiz L, Jones RL. 1973. Plasmodesmata and an associated cellwall component in barley aleurone tissue. American Journalof Botany 60, 67-75.

Tangl E. 1885. Studien flber das Endosperm einiger Gramineen.Sitzungsberichte der Akademie der Wissenschaften zu Wien92, 168-97.

Tanida I, Kim J-K, Wu R. 1994. Functional dissection of a rice

high-pi a-amylase promoter. Molecular and General Genetics244, 127-34.

Tregear JW, Primavesi LF, Huttly AK. 1995. Functionalanalysis of linker insertions and point mutations in the a-Amy2/54 GA-regulated promoter. Plant Molecular Biology29, 749-58.

Van der Veen R, Hclmovaara-Dijkstra S, Wang M. 1992.Cytosolic alkalinizati6n mediated by abscisic acid is necessary,but not sufficient, for abscisic acid-induced gene expressionin barley aleurone protoplasts. Plant Physiology 100, 699-705.

Wang M, Sedee NJA, Heidekamp F, Snaar-Jagalska BE. 1993.Detection of GTP-binding proteins in barley aleuroneprotoplasts. FEBS Letters 329, 245-8.

Wang M, Van Duijn B, Schram AW. 1991. Abscisic acid inducesa cytosolic calcium decrease in barley aleurone protoplasts.FEBS Letters 278, 69-74.

Ward J, Schroeder JI. 1994. Calcium-activated K+ channelsand calcium-induced calcium release by slow vacuolar ionchannels in guard cell vacuoles implicated in the control ofstomatal closure. The Plant Cell 6, 669-83.

Washio K, Ishikawa K. 1992. Structure and expression duringthe germination of rice seeds of the gene for a carboxypeptid-ase. Plant Molecular Biology 19, 631-40.

Wolf N. 1992. Structure of the genes encoding Hordeum vulgare(l-3,l-4)-)S-glucanase isoenzymes I and II and functionalanalysis of their promoters in barley aleurone protoplasts.Molecular and General Genetics 234, 33-42.

Yomo H. 1960. Studies on the a-amylase activating substance.IV. On the amylase activating action of gibberellin. HakkoKyokaishi 18, 600-2.

Zwar JA, Chandler PM. 1995. a-Amylase production and leafprotein synthesis in a gibberellin-responsive dwarf mutant ofHimalaya barley (Hordeum vulgare L.). Planta 197, 39-48.

Zwar JA, Hooley R. 1986. Hormonal regulation of a-amylasegene transcription in wild oat (Avena fatua L.) aleuroneprotoplasts. Plant Physiology 80, 459-63.