Plant Physiol. 91, 1006-10130032-0889/89/91/1 006/08/$01 .00/0

Received for publication January 31, 1989and in revised form May 23, 1989

An Abundant, Highly Conserved Tonoplast Protein inSeeds1

Kenneth D. Johnson*2, Eliot M. Herman, and Maarten J. ChrispeelsDepartment of Biology, University of California, San Diego, La Jolla, California 92093 (K.D.J., M.J.C.); and Plant

Molecular Biology Laboratory, U.S. Department of Agriculture, Agricultural Research Service,Beltsville, Maryland 20705 (E.M.H.)

ABSTRACT

We have isolated the membranes of the protein storage vacu-oles (protein bodies) from Phaseolus vulgaris cotyledons andpurified an integral membrane protein with M, 25,000 (TP 25).Antiserum to TP 25 recognizes an abundant polypeptide in thetotal cell extracts of many different seeds (monocots, dicots, anda gymnosperm), and specifically labels the vacuolar membranesof thin-sectioned soybean embryonic axes and cotyledons. TP 25was not found in the starchy endosperm of barley and wheat or

the seed coats of bean but was present in all seed parts examinedthat consist of living cells at seed maturity. The abundance of TP25 was not correlated with the amount of storage protein in seedtissue, and the protein was not found in leaves that accumulateleaf storage protein. On the basis of its abundance, evolutionaryconservation, and distribution in the plant, we propose that TP 25may play a role in maintaining the integrity of the tonoplast duringthe dehydration/rehydration sequence of seeds.

Seed development is characterized by the accumulation oflarge amounts of protein in protein storage vacuoles (proteinbodies) in the parenchyma cells of seed storage tissues, suchas the cotyledons or endosperm. These modified vacuolesmay contain storage proteins, lectins, and enzyme inhibitorsas well as a complement of acid hydrolases similar to thatnormally found in the large, central vacuoles of leaf and rootparenchyma cells. Since these protein storage vacuoles con-

stitute the lytic compartment of the cell (17), their limitingmembranes (tonoplasts) must remain intact during the de-hydration/rehydration sequence that characteristically occurs

between developmental arrest and germination of seeds.Several biochemical studies ofthe membranes surrounding

protein storage vacuoles have been made (12, 18). Pusztai etal (14) were the first to identify a hydrophobic protein withMr 25,000 in the protein body tonoplast ofPhaseolus vulgariscotyledons. Maeder and Chrispeels (11) showed it to be an

integral tonoplast protein synthesized on the rough endo-plasmic reticulum. We have purified this protein, hereafter

' Supported by grants to M. J. C. from the National ScienceFoundation (Metabolic Biology) and to E. M. H. and M. J. C. fromthe U.S. Department of Agriculture (Genetic mechanisms for cropimprovement).2On leave from the Department of Biology, San Diego State

University, San Diego, CA 92182.

called TP 25,3 and with antibodies raised against it, studiedits intracellular and tissue distribution as well as the presenceofcross-reactive proteins in the storage tissues ofother species.We report here that TP 25 is an abundant, integral membraneprotein that is highly conserved among different species andis present only in seeds.

MATERIALS AND METHODS

Plant Material and Chemicals

Plants of Phaseolus vulgaris L. cv Greensleeves and Nico-tiana tabacum L. cv Xanthi were grown in a greenhouse.Because the conditions ofgrowth are not uniform throughoutthe year, we determine the stage of seed development of thebeans by the weight of the seed rather than DAF. Midmatur-ation embryos are about 12 to 14 mm long at 18 to 21 DAF.Seeds of mung bean (Vigna radiata), soybean [Glycine max(L.) Wilczek cv Berken], and pea (Pisum sativum L. cvAlaska) were obtained from Burpee Seed Co. Barley (Hor-deum vulgare L. cv Himalaya) seeds were a generous gift ofDr. Russell Jones, University of California, Berkeley. Zamiapumilo seeds were collected by R. Ornduff, University ofCalifornia, Berkeley; Macrozamia sp. (burrawong palm) werecollected by T. J. Higgins, CSIRO, Canberra. Wheat (Triticumaestivum L. em. Thell) spikes containing midmature seedswere a generous gift of Dr. Giles Waines, University of Cali-fornia, Riverside. Castor bean (Ricinus communis) and aleppopine (Pinus halpensis) seeds were collected locally in SanDiego. Seeds ofbean were germinated in moist vermiculite inthe light and the cotyledons and axes collected at intervals.

Triton X-1 14 was purchased from Boehringer Mannheim,FRG; ultrapure grade Tris and acrylamide were obtained fromBethesda Research Lab.; ultrapure SDS was purchased fromInternational Biotechnologies, Inc. Unless otherwise stated,all other organic chemicals were obtained from Sigma Chem-ical Co.; solvents and inorganic chemicals were supplied byFisher Scientific.

Isolation of Protein Body Membranes (Tonoplasts)

Protein bodies from 18-h imbibed, mature bean seeds wereisolated as described by Maeder and Chrispeels (11). Theprotein bodies were lysed by suspension in lysing medium(100 mm Tris-HCl [pH 7.4] containing 1 M NaCl and 1 mM

I Abbreviations: TP 25, integral tonoplast protein with Mr 25,000.1006

DISTRIBUTION OF SEED-SPECIFIC TONOPLAST PROTEIN

EDTA) and occasional mixing for 30 min on ice. The tono-plasts were sedimented by centrifugation at 12,000g for 20min and washed twice with lysing medium and once withdistilled water by resuspension and recentrifugation. The ton-oplast pellet was suspended in 10 mM Tris-HCl (pH 7.6)containing 150 mm NaCl (Tris/salt buffer) using a glasshomogenizer fitted with a Teflon pestle.

Triton X-1 14 Treatment

The resuspended tonoplasts (3 to 6 mg/mL protein) wereadjusted at 0°C to a final concentration of 2% (w/v) TritonX-1 14 using a 16% stock of precondensed detergent. Theprocedure described by Pryde and Phillips (13) was followedto yield the detergent-insoluble fraction (discarded) and thedetergent-rich and aqueous phases. The final detergent-richfraction was diluted by the addition of two volumes of Tris/salt buffer. Proteins were precipitated from the detergent-richand aqueous fractions by the addition of 15 volumes ofacetone:ethanol (1:1, v/v) followed by overnight storage at-20°C. The protein precipitates were recovered by centrifu-gation for 20 min at 1 5,000g, washed twice with 90% acetone,dried under vacuum, and finally dissolved in denaturingbuffer (60 mM Tris-HCl [pH 6.8], 15% glycerol, 4% [w/v]SDS, and 4% [v/v] f,-mercaptoethanol). Protein estimationwas carried out by a modification of the method of Lowry etal. (10). After TCA precipitation, the protein pellet was dis-solved in 2% (w/v) Na2CO3/0.1 M NaOH and incubated 1 to2 h at 37°C. After complete dissolution, aliquots were assayedfor protein.

PAGE and Immunoblotting

SDS-PAGE was carried out using 15% (w/v) polyacryl-amide gels according to Laemmli and Favre (8). Proteinsamples were incubated in denaturing buffer for 15 min at37°C (100°C treatment caused aggregation of TP 25) andcentrifugation before loading. For preparative gel purificationof TP 25, 1.2 mg of protein from the detergent-rich fractionof bean cotyledon tonoplasts was loaded onto a slab gel (0.2x 11 x 14 cm) and the proteins separated by electrophoresis.The TP 25 band was visualized by submerging the gel in cold0.25 M KCl, cut out, and electroeluted overnight at 350 V in2.5 mm Tris (pH 8.8) containing 19 mm glycine and 0.1%SDS. The electroeluted polypeptide was precipitated in 90%acetone, recovered by centrifugation, and dried under vac-uum. For immunoblotting, the polypeptides separated bySDS-PAGE were transferred to nitrocellulose as described byFaye and Chrispeels (2) and visualized as described in theBio-Rad (Richmond, CA) technical information sheet. Pri-mary antiserum was used at a 1:2000 dilution; goat anti-rabbitimmunoglobulin G conjugated to horseradish peroxidase(Bio-Rad) was used at 1:1000 dilution.

Preparation of Antiserum

Fifty ,g of gel-purified TP 25 were suspended in 0.4 mL ofphosphate-buffered saline (PBS = 10 mm Na-phosphate [pH7.4] containing 150 mM NaCl) and emulsified with 0.4 mLof Freund's complete adjuvant (Behring Diagnostics). The

emulsion was injected subcutaneously into a New Zealandwhite rabbit. Two booster injections (30 ,g protein in PBS/incomplete adjuvant per injection) were given at 2 weekintervals. After an additional 2 weeks, blood was withdrawnfrom an ear vein approximately every 2 weeks for 60 d. Thefresh blood was allowed to clot for 2 h at room temperature,then kept at 4°C overnight. The clot was removed by twocentrifugations and the serum was aliquoted into Eppendorftubes and stored at -70°C. Preimmune serum was preparedfrom the same rabbit.

Preparation of Total Protein Extracts

Total cellular protein was obtained by homogenizing theplant tissue (cotyledons, endosperm, etc.) directly in denatur-ing buffer (2-4 mL/g) at 4°C in a mortar and pestle, thenhomogenizing in a glass tissue homogenizer. The homogenatewas centrifuged twice for 15 min at 15,000g, and the finalsupernatant (total extractable protein) was stored at -20°C.

lmmunocytochemcial Localization

Thin sections of midmaturation soybean (Glycine max cv.Forrest) axis and cotyledon (5) were embedded in LR Whiteresin. For immunogold labeling, the tissue sections were ini-tially blocked in 10% normal goat serum in TBS Tween (50mM Tris-HCl [pH 7.4], 150 mm NaCl, 0.05% Tween-20) for10 min at room temperature. The grids were then labeled ina 1:50 or 1:100 dilution of TP 25 antiserum in the blockingbuffer for 30 min at room temperature. Parallel controls usingidentical dilutions of preimmune serum accompanied eachassay. The grids were then washed with TBS Tween andindirectly labeled with S nm colloidal gold-goat anti-rabbitIgG conjugate (Janssen Life Sciences Products) for 5 min atroom temperature. The unbound gold particles were thenwashed offby sequential immersion ofthe grids in TBS Tweenand distilled water, and the grids were stained in 5% (w/v)uranyl acetate for 20 min at room temperature. The gridswere examined and photographed with Hitachi H300 andH500 electron microscopes.

RESULTS

TP 25 Is An Abundant, Integral Tonoplast Protein inLegume Cotyledons

Protein body membranes (tonoplasts) were isolated fromcotyledons of late-maturation bean seeds (700 mg per seed or24 DAF) and extracted with 2% Triton X-114 to separate thepredominantly hydrophobic, integral membrane proteins (de-tergent-rich phase) from the more hydrophilic, peripheral, oradherent proteins (aqueous phase). Equal amounts of proteinfrom tonoplasts and the detergent fractions were then ana-lyzed by SDS-PAGE (Fig. 1). Most of the abundant polypep-tides associated with the tonoplast preparation partitionedinto the aqueous phase. A notable exception is the polypeptideat Mr 25,000 (arrowhead in Fig. 1), which became enrichedin the detergent phase. We refer to this protein as TP 25. Ourresults agree with earlier findings (11, 14) that TP 25 is ahydrophobic, integrl tonoplast protein.

1 007

Plant Physiol. Vol. 91, 1989

C.J._:

Q

W c0o1 32 1- tiirritid I P>A...

47 -

-30-

m.

p9: 3 1 1 1 ng: 25 50} 75 100

A

Figure 2. Visualization and quantitation of TP 25 with a rabbitpolyclonal antiserum. After separation by SDS-PAGE, the polypep-tides were transferred to nitrocellulose and immunostained usingantiserum against TP 25. (A) Immunoblot of indicated cell fractionsand Triton X-1 14 phases of a tonoplast extract prepared from mid-maturation (18 DAF) bean cotyledons. (B) Immunoblot of increasingamounts of gel-purified TP 25.

18

Figure 1. Separation of tonoplast polypeptides by SDS-PAGE beforeand after fractionation with Triton X-1 14. Polypeptides (40 jig/lane)were visualized with Coomassie blue. The positions of mol wt stand-ards and TP 25 (arrowhead) are indicated.

We purified TP 25 by preparative SDS-PAGE and obtaineda rabbit polyclonal antiserum against the electroeluted pro-tein. Figure 2A shows an immunoblot of various cell fractionsand Triton X-1 14 phases from midmaturation bean cotyle-dons using the antiserum against TP 25. Preimmune serum

elicited no immunostaining reaction (data not shown). It isclear that TP 25 became enriched in the tonoplast fraction,and virtually all of this protein partitioned into the detergent-rich phase. The mobility ofTP 25 in the enriched fractions isslightly greater than in the total protein extract. A lower bandat Mr 23,000 often appeared on immunoblots. To check thespecificity of our antiserum, we eluted antibodies from thesetwo polypeptide bands after visualization on an immunoblotaccording to the method of Smith and Fisher (16). The band-specific antibody fractions were then used to probe fresh blotstrips of total bean cotyledon proteins. Each set of antibodiesstained both polypeptides at the same relative intensitieselicited by the unfractionated antiserum. This finding indi-cates that TP 25 is antigenically related to the smaller poly-peptide. Figure 2B shows an immunoblot of varying amountsof gel-purified TP 25 immunostained with the antiserum. Bycomparing the intensity of the TP 25 band for 3 gg of totalextractable protein (Fig. 2A) against the intensities of knownamounts of pure TP 25 (25-100 ng), we estimate the abun-

dance of this protein to be approximately 2% of the totalextractable protein of mature cotyledons.

Figure 3 shows that the antiserum against bean TP 25recognizes polypeptides with similar Mr values from the ma-ture cotyledons of other legumes: soybean (lane 1), pea (lane2), and mung bean (lane 4). The cross-reactive protein frommung bean cotyledons copurifies with protein body mem-branes and partitions into the Triton X-1 14 detergent-richphase (data not shown). Because of this remarkable cross-reactivity (the polypeptides were similar in size and immu-noreactivity) and because thin sections of soybean cotyledonswere available, we examined the subcellular location of TP25 by immunocytochemistry using soybean seeds (Fig. 4). Inthe cotyledon parenchyma cells, the membranes surroundinga protein storage vacuole became abundantly labeled withgold particles when ultrathin sections were treated with TP25 antiserum followed by goat anti-rabbit immunoglobulinG coupled to colloidal gold (Fig. 4a). There was relativelylittle labeling of the vacuolar content (protein body matrix).The tonoplast of the axis cells also labeled intensely (Fig. 4b),while controls (Fig. 4c) showed very few gold particles. Similarresults were obtained with tobacco endosperm and embryo(data not shown).

TP 25 Expression Is Developmentally Regulated andOrgan Specific

Figure 5 shows immunoblots of total extractable proteinprepared from bean (P. vulgaris) cotyledons at various stagesof seed development. The blots were probed with TP 25antiserum (upper panel) or antisera against the two seedstorage proteins, PHA, and phaseolin (lower panel). TP 25 isnot evident at the earliest developmental stage tested (lane 1),and like the seed storage proteins, it accumulates steadily

c0~

0

100

65-

47. -

30--

1008 JOHNSON ET AL.

DISTRIBUTION OF SEED-SPECIFIC TONOPLAST PROTEIN

65-

47-

30-

U?_ -

18-

1 2 3 4Figure 3. Detection ofTP 25 in other legume seeds. Total extractableproteins (10 Ag/lane) from mature, 24-h-imbibed soybean (lane 1),pea (lane 2), bean (lane 3), and mung bean (lane 4) cotyledons wereimmunoblotted using TP 25 antiserum.

thereafter up to the stage just preceding seed dehydration(lane 5). The accumulation of TP 25 parallels the accumula-tion of the vacuolar proteins phaseolin and PHA.An immunoblot of total extractable proteins from bean

seed cotyledons (ct), embryonic axes (ax), and seed coats (sc)of midmaturation cotyledons (18-20 DAF), and from theroots (rt) and leaves (lJ) of 7-week-old bean plants is shownin Figure 6. Only the cotyledons and embryonic axes of seedsexpress TP 25 (upper band of doublet). Judging by the com-parable intensities ofthe bands, the cotyledons and embryonicaxes appear to contain similar amounts ofTP 25 on a percent-protein basis.

Because embryonic axes contain TP 25 but mature shootand root tissues do not, we examined the fate ofTP 25 duringgrowth of bean (P. vulgaris) seedlings (Fig. 7). Two dayspostimbibition the immunoblot signal for TP 25 from embry-onic axes is quite strong. At later stages of seedling develop-ment, the shoot and root systems were analyzed separately.Residual TP 25 was evident in both developing shoots androots after 4 and 6 d ofgrowth, but the signal strength declinedsteadily and became undetectable after eight days of growth.To examine the fate ofTP 25 in the cotyledons, we harvestedbean cotyledons at 2 d intervals after imbibition. TP 25disappears rapidly from the cotyledons after germination (Fig.8), although a possible breakdown product persists longer.

Both the developing cotyledons and embryonic axes oflegume seeds synthesize storage proteins and lectins. There is,however, a large difference in the amount of storage proteinaccumulated, and in the appearance of the protein storagevacuoles in these two organs. In cotyledons, phaseolin ac-counts for approximately 50% of the total protein present,while it comprises only about 1% of the total protein inembryonic axes (9). The level of PHA in the axis is about10% of that found in the cotyledons of beans (4). Further-more, while the protein bodies of cotyledons are filled withdensely packed proteins, those found in the axis tissues showa sparse, patchy network of electron-dense material (4) (alsocompare Fig. 4, a and b, for soybean). Does the quantitativedistribution ofTP 25 in these two tissues parallel that for thestorage proteins? Figure 9 shows the results of an experimentdesigned to assess the relative amounts ofTP 25 (lanes 1 and2) and phaseolin (lanes 3 and 4) in bean cotyledons (lanes 1and 3) and embryonic axes (lanes 2 and 4) obtained fromseeds 22 to 24 DAF. While the abundance of TP 25 relativeto total protein in both tissues is similar, the cotyledonscontain much more phaseolin. It appears, therefore, that thelevel of expression of TP 25 is not tied in a quantitative wayto the accumulation of storage protein.We used soybean leaves to examine whether the presence

of TP 25 is linked to storage protein deposition in otherorgans. A unique storage protein accumulates in the vacuolesof the paraveinal mesophyll of mature soybean plants, and isbroken down during pod-fill (3). The protein accumulates tohigh levels in all vegetative tissues when flowering plants arecontinuously depodded (16). Immunoblots of total proteinsfrom leaves of normal and depodded plants showed no evi-dence of TP 25 (data not shown), although TP 25 is presentin soybean cotyledons (Fig. 3). Thus, TP 25 appears to beseed-specific and is not linked to storage protein accumulationin either a quantitative or qualitative way.

TP 25 Is Present in Nonembryonic Seed Tissues inNonlegumesThe mature seeds of legumes consist entirely of an embryo

surrounded by a seed coat. Most of the storage proteinsaccumulate in the large cotyledons of these seeds. For manyother plants, however, a persistent endosperm is the majorsite of seed protein storage. To examine whether TP 25 ispresent in non-embryonic seed tissues, we extracted totalproteins from a variety of seeds or specific seed tissues andsubjected them to immunoblot analysis using TP 25 anti-serum. Figure 10 shows the results of such an analysis forbarley, castor bean, and tobacco seeds, all of which showcross-reactive bands at Mr 25,000. In imbibed castor bean andbarley, both the embryo and endosperm tested positive, butonly the aleurone portion (lane 2) of the barley endospermharbors a TP 25-like protein. The starchy endosperm (lane 3)lacks the typical band at Mr 25,000. We found a similardistribution of TP 25 in wheat (T. aestivum L. Cv Thell.)seeds, and used midmature wheat (a gift of Dr. G. Waines ofthe University of California, Riverside) to show that theabsence of TP 25 from cereal starchy endosperms is not dueto loss of this protein during the final stages of endospermmaturation (data not shown).

1 009

Plant Physiol. Vol. 91,1989

X

''"i;sts4Ad'

4|l4,1.;'.G'',, ':

*; PB

:A...;...,

'.4..... .

.m

al i

..................''>,.'.5x5:Q.,5'S-,;,, .inatio*'m'' X "" ;''~"'S '' - ':hi. A . is,.,.s",...............Sb.. 'C.. ,.-W

reSUS.....

Figure 4. Immunocytochemical localization of TP 25 in soybean cotyledons and axes. (a) Midmature cotyledon protein body membrane labeledwith TP 25 antiserum. Note that the gold label is primarily localized on the membrane (arrowheads) of the protein body (PB). Also shown are oilbodies (OB) (x57,000). (b) Midmature soybean axis protein body membrane labeled with TP 25 antiserum. Although the label is primarily localizedon the protein body membrane (arrows), the axis protein bodies (PB) exhibited higher levels of labeling on the protein body matrix than cotyledonprotein bodies (x49,000). (c) An immunological control of a midmature soybean cotyledon protein body in which preimmune serum was

substituted for the specific antiserum. Note the absence of gold particles along the protein body membrane (arrowheads) (x57,000).

We have also detected TP 25 cognates in coconut endo-sperm and whole seeds of aleppo pine (P. halepensis), but notin the seeds of two cycad species (Z. pumilo and Macrozamiasp.), nor in the yeast, Saccharomyces cerevisae (data notshown). With the exception of the cycads, TP 25 appears tobe a highly conserved seed protein.

DISCUSSION

The protein corresponding to TP 25 from mature P. vul-garis cotyledons has been previously characterized as a hydro-phobic, integral protein of the protein storage vacuoles (1 1,14). This characterization is based on its insolubility inaqueous buffers lacking detergent, its preponderance of hy-drophobic amino acids as shown by amino acid compositionanalysis (14), and its resistance to extraction from its mem-brane environment by 10 mm NaOH (1 1). Our finding thatTP 25 partitions into the detergent-rich phase of Triton X-1 14-extracted tonoplasts (Fig. 1) corroborated the earlier con-clusions, and made the task ofpurifying the polypeptide ratherstraightforward (TP 25 accounted for more than half of the

detergent-soluble protein from tonoplasts; see Fig. 1). Wewere thus able to obtain a sensitive and relatively specificpolyclonal antiserum against this protein, which led to theexperiments reported herein.

Specificity of the Antiserum

The antiserum raised against gel-purified TP 25 clearlyrecognizes this polypeptide in a bean cotyledon extract. Othercross-reactive polypeptides, however, were also evident on

most immunoblots, including a major antigen at Mr 23,000(see Fig. 2A, middle lanes; Fig. 6, lanes ct and ax; Figs. 7 and8; Fig. 9, lane 2). To determine whether TP 25 and the 23 kDpolypeptide are antigenically related, we used the Smith andFisher (15) procedure to elute specific antibodies bound toeach polypeptide on preparative immunoblots, then subjectedfresh blot strips containing total SDS-PAGE-separated coty-ledon protein to each purified antibody fraction. Because bothband-specific antibody eluates recognized both antigens witha staining pattern similar to that elicited by the unpurifiedantiserum, we conclude that TP 25 and the 23 kD polypeptide

1:;:.., :.

1010 JOHNSON ET AL.

DISTRIBUTION OF SEED-SPECIFIC TONOPLAST PROTEIN

47m.

30Qm

30-

-o inmi *1801-

ct ax scFigure 6. Organ-specific expression of TP 25 in beans. Total extract-able proteins (10 gg/lane) from mid-maturation (18-20 DAF) beancotyledons (ct), embryonic axes (ax), and seed coats (sc), and totalextractable proteins (20 Mg/lane) from roots (rt) and leaves (If) of 7-week-old plants were separated by SDS-PAGE, transferred to nitro-cellulose, and immunostained using TP 25 antiserum.

1 2 3 4 5

Figure 5. Expression of TP 25 in bean cotyledons during seeddevelopment. The developmental stages are defined by seed freshweight: 50 mg (lane 1), 170 mg (lane 2), 370 mg (lane 3), 710 mg(lane 4), and 1000 mg (lane 5). These weights correspond approxi-mately to 10, 14, 18, 24, and 30 DAF, respectively. Upper panel: 10Ag total cotyledon protein/lane; blot immunostained with TP 25antiserum. Lower panel: 2 tgg cotyledon protein/lane; blot immuno-stained with both anti-phaseolin (Phas) and anti-PHA antisera.

Axis I

65-

30-

18-

have similar epitopes (i.e., they are antigenically related).More importantly, these results suggest that the antiserumagainst TP 25 is specific. We believe that the 23 kD polypep-tide is a breakdown (processing?) product of TP 25. Thisconclusion is further supported by the fact that TP 25, butnot the 23 kD polypeptide, can be immunoselected from invitro translation products in a rabbit reticulocyte lysate forti-fied with poly(A)+ RNA isolated from midmature bean coty-ledons (KD Johnson, MJ Chrispeels, unpublished results).On several immunoblots, TP 25 antiserum elicited a diffuse

immunostaining pattern in the higher mol wt range, includinga discernible band at M, 42,000 (Fig. 3, lane 3; Fig. 5, lanes3-5; Fig. 6, lanes ct and ax-, Fig. 7, lane Axis). The intensitiesof these diffuse bands tended to increase with the amount ofTP 25 loaded on the gel and increased dramatically when theprotein samples were heated at 100°C rather than the usual37°C before loading. We believe that these higher-molecular-weight bands represent various levels of aggregation ofTP 25,a phenomenon not unusual for very hydrophobic proteins.This interpretation is supported by the observation that gel-purified TP 25 (the 25 kD band cut out of a preparative gel

2 4 6 8 4 6

Days of GrownthFigure 7. TP 25 disappearance from developing bean seedlings.Total extractable proteins (5 Mg/lane) from shoots (hypocotyl, epicotyl,and expanding plumule) and roots (primary and secondary roots)were separated by SDS-PAGE, transferred to nitrocellulose, andimmunostained using TP 25 antiserum.

and electroeluted), when rerun on SDS-PAGE, yielded notonly the major 25 kD polypeptide, but also the diffuse bandsat M, 42,000 and higher (data not shown).

TP 25 Is an Abundant and Highly Conserved Protein

TP 25 is surprisingly abundant in bean cotyledon cells. Bycomparing the signal strengths ofknown amounts ofpure TP25 on immunoblot with the band intensity of this proteinpresent in a 3-MAg sample of total extractable protein (Fig. 2),we estimate that TP 25 represents about 2% of the totalprotein present in mature bean cotyledons. Of course, there

65-

47-

18-

rt I f

47- EWUhahas4 4*

30-

Shoot Root

1011

I I

:-. s,. 4"f: VIIIIP r-,-If : PHA

Plant Physiol. Vol. 91, 1989

b a r I e ya stobear

47

30

_w__

18

1 2 3

DaYs of f3rowrt

Figure 8. Disappearance of TP 25 from bean cotyledons after ger-mination. Total extractable proteins from 0.00025 cotyledons per lanewere separated by SDS-PAGE, transferred to nitrocellulose, andimmunostained using TP 25 antiserum. Day 0 indicates cotyledonsfrom dry seeds.

47om..

30om

Figure 10. Detection and distribution of TP 25 in barley, castor bean,and tobacco seed tissues. Barley panel: 10 ug total protein/lane from3 d postimbibition shoots and roots (lane 1), aleurone layers (lane 2),and starchy endosperm (lane 3). Castor bean panel: 10 Mg totalprotein/lane from 24-h-imbibed embryo (lane 1) and endosperm (lane2). Tobacco (tob) panel: 20 Mug total seed protein/lane. The polypep-tides were separated by SDS-PAGE, transferred to nitrocellulose,and immunostained with TP 25 antiserum.

the potential errors inherent in this semiquantitative ap-proach, and feel that our estimate of TP 25 abundance isreasonable. Finally, the relatively high density of colloidalgold particles staining the protein body membranes (Fig. 4)also suggests an abundant protein.The antiserum against bean TP 25 cross-reacts with proteins

having similar Mr values (23,000-26,000) in seeds of otherplants, including soybean, mung bean, pea, tobacco, castorbean (embryo and endosperm), coconut (endosperm), barleyand wheat (embryo and aleurone), and aleppo pine. Thesimilarities in the strength of the signal and the MR of theprotein in each case lead us to suggest that TP 25 is a highlyconserved protein.

4

Figure 9. Comparison of TP 25 and phaseolin levels in bean cotyle-dons and embryonic axes. Total extractable proteins from cotyledons(600 mg/seed; 22-24 DAF) (lanes 1 and 3) and embryonic axes ofthe same seeds (lanes 2 and 4) were immunoblotted using TP 25antiserum (lanes 1, 2; 5 Mg protein/lane) or phaseolin antiserum (lanes3, 4; 2 Mg protein/lane).

are problems associated with comparing signal strengths ofimmunostained bands, particularly when comparisons are

being made between purified versus heterogeneous proteinsamples. To minimize the potential errors, we employed a

modification of the Lowry et al. (10) method, which ensuresthat proteins are solubilized before color development isinitiated. Furthermore, the mole percent of tyrosine in bovineserum albumin (3.3%), our protein standard, and TP 25(3.2%) (14) are similar. Thus, we have attempted to minimize

Accumulation of TP 25 Is Seed Specific and Is NotRelated to Storage Protein Deposition

Although TP 25 is present in most of the seed storagetissues examined, its accumulation does not appear to berelated to storage protein deposition. In bean the axis andcotyledons contain similar levels of TP 25 (expressed on a

per-mg-protein basis), yet the axis contains only about 2% as

much phaseolin and 10% as much phytohemagglutinin whencompared to the cotyledons (4, 9). In the leaves of soybean,the accumulation of a vegetative storage protein is not corre-lated with the presence of TP 25. Thus, TP 25 appears to beseed-specific, but does not appear in those tissues that dieduring seed maturation (seed coat of bean and the starchyendosperms of wheat and barley); it is present, however, inthe coconut endosperm, a tissue that does not undergo exten-sive dehydration.

30.

18-

1012 JOHNSON ET AL.a-r.I80-

DISTRIBUTION OF SEED-SPECIFIC TONOPLAST PROTEIN

Function of TP 25

We can only speculate as to the function of TP 25. Wepropose, on the basis of its distribution in plants, that it playsa role in stabilizing the tonoplast during the dehydration/rehydration sequence of seeds. Recent evidence indicates thatseeds synthesize a number of abundant proteins (so-called"late embryogenesis abundant," or Lea proteins) that are

hydrophilic (1). These proteins have been postulated to playa role in protecting the cytoplasm from desiccation damage.A similar role has been postulated for small molecules (e.g.,sucrose and raffinose) that accumulate to relatively high levelsin embryos (7). It is reasonable to expect that membranes(tonoplast and plasma membrane) that must withstand desic-cation and rehydration may have special lipids and proteins.Storage protein vacuoles have been shown to be unusuallyrich in sterols (6), but it is not known whether these lipidsperform a role in membrane stabilization. The presence of ahighly conserved protein in the vacuolar membranes of allseeds that we have examined (with the exception of cycads)leads us to propose a protective role for TP 25. Such a generalrole in seed physiology is consistent with its abundance andits presence in both the gymnosperms (pine) and angiosperms.We propose to test this hypothesis by expressing antisense TP25 behind a strong seed-specific promoter in transgenic to-bacco and examining the capacity of the seeds to withstanddesiccation and rehydration.

LITERATURE CITED

1. Baker J, Steele C, Dure LIII (1988) Sequence and chacterizationof 6 Lea proteins and their genes from cotton. Plant Mol Biol11: 277-291

2. Faye L, Chrispeels M (1985) Characterization of N-linked oli-gosaccharides by affinoblotting with concanavalin A-peroxi-dase and treatment of the blots with glycosidases. AnalBiochem 149: 218-224

3. Franceschi VR, Wittenbach VA, Giaquinta RT (1983) Paraveinalmesophyll of soybean leaves in relation to assimilate transferand compartmentation. Plant Physiol 72: 586-589

4. Greenwood JS, Keller GA, Chrispeels MJ (1984) Localizationof phytohemagglutinin in the embryonic axis of Phaseolusvulgaris with ultrathin cryosections embedded in plastic afterindirect immunolabeling. Planta 162: 548-555

5. Herman EM (1987) Immunogold localization and synthesis ofan oil-body membrane protein in developing soybean seeds.Planta 172: 336-345

6. Herman EM, Platt-Aloia KA, Thomson WW, Shannon LM(1984) Freeze fracture and filipin cytochemical observations ofdeveloping soybean cotyledons. Eur J Cell Biol 35: 1-7

7. Koster KL, Leopold AC (1988) Sugars and desiccation tolerancein seeds. Plant Physiol 88: 829-832

8. Laemmli UK, Favre M (1973) Maturation of the head of bacte-riophage T4. J Mol Biol 80: 575-599

9. Long SR, RMK Dale, IM Sussex (1981) Maturation and ger-mination of Phaseolus vulgaris embryonic axes in culture.Planta 153: 405-415

10. Lowry OM, Rosebrough MJ, Farr AL, Randall RJ (1951) Pro-tein measurements with the Folin phenol reagent. J Biol Chem193: 265-275

11. Maeder M, Chrispeels MJ (1984) Synthesis ofan integral proteinof the protein-body membrane in Phaseolus vulgaris cotyle-dons. Planta 160: 330-340

12. Mettler IJ, Beevers H (1979) Isolation and characterization ofthe protein body membrane of castor beans. Plant Physiol 64:506-511

13. Pryde JG, Phillips JH (1986) Fractionation of membrane pro-teins by temperature-induced phase separation in Triton X-114. Biochem J 233: 525-533

14. Pusztai A, Croy RRD, Stewart JS, Watt WB (1979) Proteinbody membranes ofPhaseolus vulgaris L. cotyledons: isolationand preliminary characterization of constituent proteins. NewPhytol 83: 371-376

15. Smith DE, Fisher PA (1984) Identification, developmental reg-ulation, and response to heat shock oftwo antigenically relatedforms of a major nuclear envelope protein in Drosophilaembryos: application of an improved method for affinity pu-rification of antibodies using polypeptides immobilized onnitrocellulose blots. J Cell Biol 99: 20-28

16. Staswick PE (1988) Soybean vegetative storage protein structureand gene expression. Plant Physiol 87: 250-254

17. Van der Wilden W, Herman EM, Chrispeels MJ (1980) Proteinbodies ofmung bean cotyledons as autophagic organelles. ProcNatl Acad Sci USA 77: 428-432

18. Weber E, Suess K-H, Neumann D, Manteufel R (1979) Isolationand partial characterization of the protein body membranefrom mature and germinating seeds of Vicia faba. BiochemPhysiol Pflanzen 174: 139-150

1013