Endo-[beta]-Mannanase Activity Present in Cell Wall Extracts of Lettuce Endosperm prior to Radicle Emergence

Lettuce (Lactuca sativa L.) endosperm cell walls isolated prior to radicle emergence underwent autohydrolysis, the rate of which was correlated with whether radicle emergence would subsequently occur. Extraction of endosperm cell walls with 6 M LiCl suppressed autohydrolysis, and the desalted extract possessed activity that was capable of hydrolyzing purified locust bean galactomannan but not arabinogalactan, carboxymethylcellulose, glucomannan, polygalacturonic acid, tomato galactomannan, or native lettuce endosperm cell walls. Some hydrolytic activity was detected on endosperm cell walls if they were modified by partial trifluoroacetic acid hydrolysis or pretreatment with guanidinium thiocyanate. In extended incubations the cell wall enzyme extract released only large molecular mass fragments from locust bean galactomannan, indicating primarily endo-activity. Galactomannan-hydrolyzing activity in the cell wall extract increased as a function of imbibition time and was greatest just prior to radicle emergence. Thermoinhibition (imbibition at 32[deg]C) or treatment with abscisic acid at a temperature optimal for germination (25[deg]C) suppressed both germination and endosperm cell wall mannanase activity, whereas alleviation of thermoinhibition with gibberellic acid was accompanied by significant enhancement of mannanase activity. We conclude that a cell wall-bound endo-[beta]-mannanase is expressed in lettuce endosperm prior to radicle emergence and is regulated by the same conditions that govern germination.     

Temporal and spatial pattern of the biochemical activation of the endosperm during and following imbibition of tomato seeds

Electron microscopic observations of the endosperm of tomato ( Lycopersicon esculentum Mill.) seeds revealed that changes in the cell wall structures along with the vacuolation of protein bodies occurred in the micropylar portion of the endosperm prior to germination. No changes were detected at that time in the rest of the endosperm. Endo‐β‐mannanase activity was restricted to the micropylar region of the endosperm prior to germination. Cell wall digestion by this pregerminative mannanase seemed to be associated with the changes in cell wall structures occurring in the micropylar region prior to germination. The protein content in the micropylar part of the endosperm began to decrease shortly after imbibition and attained about 40% of the initial level by the time of radicle protrusion (38 h after imbibition). On the other hand, only slight changes in the content were detected in the lateral endosperm during the same time; the protein content in the lateral endosperm decreased only after germination started. In conformity with the results on protein contents, proteolytic activity began to develop first in the micropylar portion prior to germination, and then in the lateral portion after germination. Thus, the timing of the biochemical activation of the endosperm after imbibition differed between the micropylar and the lateral region. Some qualitative differences in patterns of polypeptides synthesized in vivo were detected, as analyzed by pulse‐labeling and fluorography, between the micropylar and the lateral portions of the endosperm of seeds imbibed for 25 h. This suggests that processes of the biochemical activation of the endosperm may be qualitatively, as well as quantitatively, different depending on the regions of the endosperm.     

Galactomannan hydrolyzing activity develops during priming in the micropylar endosperm tip of tomato seeds

Development of galactomannan hydrolyzing activity was followed in seeds of tomato [ Lycopersicon esculentum (L.) Mill. cv. Toyonishiki] during priming and germination. The activity developed in seeds that were being primed in polyethylene glycol (-0.8 MPa). The activity was detected exclusively in the endosperm portion just adjacent to the radicle tip. Part of the activity remained active after desiccation of the primed seeds. After transfer to water, the activity in the primed seeds immediately began to increase, while in unprimed seeds the beginning of the increase in activity was delayed by about 1 day. In scanning electron microscopy, the inner surface of the cell walls of the micropylar endosperm portion appeared eroded in primed seeds that had been imbibed in water for 16 h (before germination), but not in unprimed seeds imbibed for the same period. These results support the hypothesis that galactomannan hydrolyzing enzyme, which is believed to be responsible for breakdown of tomato endosperm cell walls and hence for the weakening of mechanical restraint against radicle growth, may be involved in the improved germination of primed tomato seeds.     

A Single-Seed Assay for Endo-[beta]-Mannanase Activity from Tomato Endosperm and Radicle Tissues

Completion of germination (radicle emergence) is an all-or-none developmental event for an individual seed. Variation in germination timing among seeds in a population therefore reflects variation among seeds in the rates or extents of physiological or biochemical processes prior to radicle emergence. For tomato (Lycopersicon esculentum Mill.) seeds, correlative evidence suggests that endo-[beta]-mannanase activity weakens the endosperm cap tissue opposite the radicle tip to permit radicle emergence. To test whether endo-[beta]-mannanase activity is causally related to germination rates, we have developed a sensitive assay suitable for use with individual radicle tips or endosperm caps. We show that endo-[beta]-mannanase activity varies at least 100-fold and often more than 1000-fold among individual inbred tomato seeds prior to radicle emergence. Other sources of variation (tissue size and experimental error) were evaluated and cannot account for this range of activity. Endo-[beta]-mannanase activity was generally 10-fold greater in leachates from endosperm caps than from radicle tips. Release of reducing sugars from individual endosperm caps also varied over a considerable (9-fold) range. These extreme biochemical differences among individual tomato seeds prior to radicle emergence indicate that results obtained from bulk samples could be misleading if it is assumed that all seeds exhibit the "average" behavior.     

The emergence of embryos from hard seeds is related to the structure of the cell walls of the micropylar endosperm, and not to endo-beta-mannanase activity

BACKGROUND AND AIMS: Seeds of carob, Chinese senna, date and fenugreek are hard due to thickened endosperm cell walls containing mannan polymers. How the radicle is able penetrate these thickened walls to complete seed germination is not clearly understood. The objective of this study was to determine if radicle emergence is related to the production of endo-beta-mannanase to weaken the mannan-rich cell walls of the surrounding endosperm region, and/or if the endosperm structure itself is such that it is weaker in the region through which the radicle must penetrate. METHODS: Activity of endo-beta-mannanase in the endosperm and embryo was measured using a gel assay during and following germination, and the structure of the endosperm in juxtaposition to the radicle, and surrounding the cotyledons was determined using fixation, sectioning and light microscopy. KEY RESULTS: The activity of endo-beta-mannanase, the major enzyme responsible for galactomannan cell wall weakening increased in activity only after emergence of the radicle from the seed. Thickened cell walls were present in the lateral endosperm in the hard-seeded species studied, but there was little to no thickening in the micropylar endosperm except in date seeds. In this species, a ring of thin cells was visible in the micropylar endosperm and surrounding an operculum which was pushed open by the expanding radicle to complete germination. CONCLUSIONS: The micropylar endosperm presents a lower physical constraint to the completion of germination than the lateral endosperm, and hence its structure is predisposed to permit radicle protrusion.     

Class I [beta]-1,3-Glucanases in the Endosperm of Tobacco during Germination

Rupture of the seed coat and rupture of the endosperm are separate events in the germination of Nicotiana tabacum L. cv Havana 425 seeds. Treatment with 10-5 M abscisic acid (ABA) did not appreciably affect seed-coat rupture but greatly delayed subsequent endosperm rupture by more than 100 h and resulted in the formation of a novel structure consisting of the enlarging radicle with a sheath of greatly elongated endosperm tissue. Therefore, ABA appears to act primarily by delaying endosperm rupture and radicle emergence. Measurements of [beta]-1,3-glucanase activity, antigen content, and mRNA accumulation together with reporter gene experiments showed that induction of class I [beta]-1,3-glucanase genes begins just prior to the onset of endosperm rupture but after the completion of seed-coat rupture. This induction was localized exclusively in the micropylar region of the endosperm, where the radicle will penetrate. ABA treatment markedly inhibited the rate of [beta]-1,3-glucanase accumulation but did not delay the onset of induction. Independent of the ABA concentration used, onset of endosperm rupture was correlated with the same [beta]-1,3-glucanase content/seed. These results suggest that ABA-sensitive class I [beta]-1,3-glucanases promote radicle penetration of the endosperm, which is a key limiting step in tobacco seed germination.     

Mechanisms of hormonal regulation of endosperm cap-specific gene expression in tomato seeds

The micropylar region of endosperm in a seed, which is adjacent to the radicle tip, is called the 'endosperm cap', and is specifically activated before radicle emergence. This activation of the endosperm cap is a widespread phenomenon among species and is a prerequisite for the completion of germination. To understand the mechanisms of endosperm cap-specific gene expression in tomato seeds, GeneChip analysis was performed. The major groups of endosperm cap-enriched genes were pathogenesis-, cell wall-, and hormone-associated genes. The promoter regions of endosperm cap-enriched genes contained DNA motifs recognized by ethylene response factors (ERFs). The tomato ERF1 (TERF1) and its experimentally verified targets were enriched in the endosperm cap, suggesting an involvement of the ethylene response cascade in this process. The known endosperm cap enzyme endo-β-mannanase is induced by gibberellin (GA), which is thought to be the major hormone inducing endosperm cap-specific genes. The mechanism of endo-β-mannanase induction by GA was also investigated using isolated, embryoless seeds. Results suggested that GA might act indirectly on the endosperm cap. We propose that endosperm cap activation is caused by the ethylene response of this tissue, as a consequence of mechanosensing of the increase in embryonic growth potential by GA action.     

Beta-mannosidase (EC 3.2.1.25) activity during and following germination of tomato (Lycopersicon esculentum Mill.) seeds. Purification,cloning and characterization

Beta-mannosidase, a high-salt-soluble enzyme, increases in activity in seeds of tomato prior to the completion of germination. This increase occurs in both the lateral and micropylar endosperm and becomes more evident during post-germinative seedling growth. The beta-mannosidase activity profile is similar to that of endo beta-mannanase although it is the first to increase in the lateral endosperm. Tomato seed beta-mannosidase was purified to homogeneity and its cDNA (LeMside1) obtained by 3'-RACE PCR using oligonucleotide sequences based on four peptide sequences obtained from the purified enzyme. The derived amino acid sequence of the tomato beta-mannosidase shows the enzyme is a member of the Glycosyl Hydrolases Family 1 (GHF1) but has a very low sequence identity with that of beta-mannosidases from non-plant sources; no other plant sequence for the enzyme is known. There appears to be only one gene encoding beta-mannosidase in tomato, the sequence of which has been determined (LeMSide2). Its expression occurs first in the micropylar endosperm, and then declines after germination. This is followed by an increase in its expression in the lateral endosperm, which precedes that of the gene for endo beta-mannanase. Expression of the beta-mannosidase gene increases appreciably in the growing seedling embryo. With this report, the cloning of all three of the enzymes involved in galactomannan mobilization (endo beta-mannanase, alpha-galactosidase and beta-mannosidase) in tomato seeds has now been achieved.     

Mechanism and control of Genipa americana seed germination

Genipa americana (Rubiaceae) is important for restoration of riparian forest in the Brazilian Cerrado. The objective was to characterize the mechanism and control of germination of G. americana to support uniform seedling production. Morphology and morphometrics of seeds, embryo and endosperm were assessed by light and scanning electron microscopy during germination. Imbibition and germination curves were generated and over the same time interval endosperm digestion and resistance were measured by puncture force analysis and activity assay of endo-β-mannanase (EBM) in water and in abscisic acid (ABA). The gene encoding for EBM was partially cloned and its expression monitored by quantitative real-time-polymerase chain reaction. Embryos displayed growth prior to radicle protrusion. A two-phase increase in EBM activity coincided with the two stages of weakening of the micropylar endosperm. The second stage also coincided with growth of the embryo prior to radicle protrusion. Enzyme activity was initiated in the micropylar endosperm but spread to the lateral endosperm. ABA completely inhibited germination by inhibiting embryo growth, the second stage of weakening and expression of the EBM gene, but EBM activity was not significantly inhibited. This suggests that a specific isoform of the enzyme is involved in endosperm weakening. EBM may cause a general 'softening' of micropylar endosperm cell walls, allowing the embryo to puncture the endosperm as the driving force of the decrease in puncture force.     

Relationship of Endo-[beta]-D-Mannanase Activity and Cell Wall Hydrolysis in Tomato Endosperm to Germination Rates

The endosperm tissue enclosing the radicle tip (endosperm cap) governs radicle emergence in tomato (Lycopersicon esculentum Mill.) seeds. Weakening of the endosperm cap has been attributed to hydrolysis of its mannan-rich cell walls by endo-[beta]-D-mannanase. To test this hypothesis, we measured mannanase activity in tomato endosperm caps from seeds allowed to imbibe under conditions of varying germination rates. Over a range of suboptimal temperatures, mannanase activity prior to radicle emergence increased in accordance with accumulated thermal time. Reduced water potential delayed or prevented radicle emergence but enhanced mannanase activity in the endosperm caps. Abscisic acid did not prevent the initial increase in mannanase activity, although radicle emergence was markedly delayed. Sugar composition and percent mannose (Man) content of endosperm cap cell walls did not change prior to radicle emergence under any condition. Man, glucose, and other sugars were released into the incubation solution by endosperm caps isolated from intact seeds during imbibition. Pregerminative release of Man was suppressed and the release of glucose was enhanced when seeds were incubated in osmoticum or abscisic acid; the opposite occurred in the presence of gibberellin. Thus, whereas sugar release patterns were sensitive to environmental and hormonal factors affecting germination, neither assayable endo-[beta]-D-mannanase activity nor changes in cell wall sugar composition of endosperm caps correlated well with tomato seed germination rates under all conditions.     

Developing tomato seeds when removed from the fruit produce multiple forms of germinative and post-germinative endo-β-mannanase. Responses to desiccation, abscisic acid and osmoticum

Several isoforms of endo-1,4-D-mannanase (EC3.2.1.78) are produced in the endosperm and embryo of tomato (Lycopersicon esculentum Mill.) seed prior to the completion of germination. Other isoforms appear in the embryo and in the lateral endosperm following germination. This occurs in seeds removed from the fruit prior to completion of development at 45 d after pollination and placed directly on water, or following drying. Hence desiccation is not required to induce either germination- or post-germination-related mannanase activity. Incubating seeds in abscisic acid or osmoticum results in a reduction of both germination and total mannanase activity, but the isoforms that are produced in the embryo and micropylar region of the endosperm are identical to those produced in water-imbibed seeds prior to germination. Incubation of seeds in a high concentration of abscisic acid prevents all enzyme production. Only after the completion of germination does mannanase increase in the lateral regions of the endosperm. In contrast, mannanase is produced in the micropylar region regardless of whether the seed germinates or not. The isoforms produced in the two regions of the endosperm are different, those in the lateral endosperm being more similar to those produced in the cotyledons and axes of the embryo. Embryos and endosperms dissected prior to completion of germination and incubated separately produce far fewer isoforms than when these parts are together in the intact seed.Abbreviations ABA cis-abscisic acid - DAP days after pollination - GA gibberellin - IEF isoelectric focusing - PEG polyethyleneglycol - pI isoelectric point This work was supported by Natural Sciences and Engineering Council of Canada grant A2210. B.V. received a fellowship from the Deutscher Akademischer Austauschdienst for her research at the University of Guelph. We are grateful to Dr. H.W.M. Hilhorst, Wageningen, for his critical comments.     

Development of beta-1,3-glucanase activity in germinated tomato seeds

Laminarin-hydrolysing activity developed in the endosperm of tomato (Lycopersicon esculentum) seeds following germination. The enzyme was basic (pI>10) and the apparent molecular mass was estimated to be 35 kDa by SDS-PAGE. It was specific for linear beta-1,3-glucan substrates. Laminarin was hydrolysed by the enzyme to yield a mixture of oligoglucosides, indicating that the enzyme had an endo-action pattern. Thus, the enzyme was identified as beta-1,3- endoglucanase (EC 3.2.1.39). The activity of the enzyme developed in the endosperm after radicle protrusion (germination) had occurred and the enzyme activity was localized exclusively in the micropylar region of the endosperm where the radicle had penetrated. When the lateral endosperm region, where no induction of the enzyme occurred, was wounded (cut or punctured), there was a marked enhancement of beta-1,3-glucanase activity. Thus the post-germinative beta-1, 3-glucanase activity in the micropylar endosperm portion might be brought about by wounding resulting from endosperm rupture by radicle penetration.     

A Role for the Stele in Intertissue Signaling in the Initiation of Abscission in Bean Leaves (Phaseolus vulgaris L.)

A combination of microdissection and viscometric endo-[beta]-1,4-glucanhydrolase assays was used to investigate if the early appearance of the abscission-related isoelectric point-9.5 endo-[beta]-1,4-glucanhydrolase in the stele of the pulvinus and abscission zone of the foliar abscission zone of Phaseolus vulgaris L. prior to cell separation (reported by E. del Campillo, P.D. Reid, R. Sexton, L.N.Lewis [1990] Plant Cell 2: 245-254) indicates that the vascular tissue of this region has a specific role in abscission. We find that no endo-[beta]-1,4-glucanhydrolase activity or cell separation is detectable in the abscission zone cortex if the abscission zone cortex is separated from the stele tissue. If the stele is separated from the abscission zone cortex after a lag period but again before any endo-[beta]-1,4-glucanhydrolase activity is present in the abscission zone cortex, then the enzyme is produced in the cortex and abscission ensues. We conclude that the cortex of the abscission zone is able to abscind independently of the vascular tissue only after the vascular tissue has begun to respond to abscission-promoting signals. We suggest that ethylene promotes formation of an abscission-permitting signal in the stele of the abscission zone and pulvinus, and that this signal is an essential elicitor for the synthesis of cell separation enzymes in the target cells of the abscission zone cortex.     

Peroxidase activity develops in the micropylar endosperm of tomato seeds prior to radicle protrusion

Peroxidase activity developed specifically in the micropylar region of the endosperm of imbibed tomato seeds prior to radicle emergence. The activity was first detected approximately 24 h after the start of imibibition (6 h before radicle emergence) and increased markedly thereafter. In the lateral portion of the endosperm, peroxidase activity was undetectable for the first 2 d after the start of imbibition. Although the activity in the lateral endosperm became detectable 3 d after imbibition, the extent of the development of the activity was slight. The localization of peroxidase activity in the micropylar endosperm 2 d after the start of imbibition was confirmed by tissue printing analyses. When the endosperm tissues were wounded, there was an enhancement of the enzyme activity at the wounded region. H2O2 was formed at the expense of NADH only in the presence of Mn2+ and dinitrophenol by the extract from the micropylar endosperm in which peroxidase activity was present. The presence of H2O2 in the micropylar portion of the endosperm was shown histochemically. The possible functions of the peroxidases that develop in the endosperm of tomato seeds are discussed.     

Involvement of Endomannanase in the Control of Tomato Seed Germination under Low Temperature Conditions

The cause of differences in germination rates in a cold-toleranttomato line (PI341988), a control line (UC82B), and six progenylines stemming from crosses and backcrosses between the twoparent lines was investigated. Pursuant to earlier work showingthat differences in germination ability at 12°C are dueto the barrier imposed by the endosperm layer, we analysed theactivity of cell-wall-hydrolysing enzymes extracted from theselines. A significant increase in endomannanase activity wasfound in plant line PI341988 prior to germination at 12°C.Extracts of PI341988 seeds that had imbibed at either 12 or25°C exhibited higher endomannanase activity than theircounterparts from plant line UC82B. Moreover, a positive relationshipwas found between germination ability at low temperature andendomannanase activity in the six progeny lines. Analysis ofendomannanase activity in sub-regions of the seed indicatedthat the increase in activity prior to germination was higherin the micropylar endosperm cap than in the rest of the seed.Exogenous application of mannanase originating from soil-bornebacteria increased germination rates under both moderate andlow temperature conditions. Cellulase (endo-1,4-ß-glucanase)activity was also found to be higher in plant line PI341988.However, the activity of this enzyme probably increases aftergermination and it is therefore not considered as a key enzymecontrolling germination at low temperatures.Copyright 1995,1999 Academic Press Lycopersicon esculentum, tomato, seed germination, cell wall     

The relationship between beta-mannosidase and endo-beta-mannanase activities in tomato seeds during and following germination: a comparison of seed populations and individual seeds

beta-Mannosidase and endo-beta-mannanase are involved in the mobilization of the mannan-containing cell walls of the tomato seed endosperm. The activities of both enzymes increase in a similar temporal manner in the micropylar and lateral endosperm during and following germination. This increase in enzyme activities in the micropylar endosperm is not markedly reduced in seeds imbibed in abscisic acid although, in the lateral endosperm, endo-beta-mannanase activity is more suppressed by this inhibitor than is the activity of beta-mannosidase. Gibberellin-deficient (gib-1) mutants of tomato do not germinate unless imbibed in gibberellin; low beta-mannosidase activity, and no endo-beta-mannanase activity is present in seeds imbibed in water, but both enzymes increase strongly in activity in the seeds imbibed in the growth regulator. For production of full activity of both beta-mannosidase and endo-beta-mannanase in the endosperm, this tissue must be in contact with the embryo for at least the first 6 h of imbibition, which is indicative of a stimulus diffusing from the embryo to the endosperm during this time. These results suggest some correlation between the activities of beta-mannosidase and endo-beta-mannanase, particularly in the micropylar endosperm, in populations of tomato seeds imbibed in water, abscisic acid and gibberellin. However, when individual micropylar endosperm parts are used to examine the effect of the growth regulators and of imbibition in water on the production of the two enzymes, it is apparent that within these individual seed parts there may be large differences in the amount of enzyme activity present. Micropylar endosperms with high endo-beta-mannanase activity do not necessarily have high beta-mannosidase activity, and vice versa, which is indicative of a lack of co-ordination of the activities of these two enzymes within individuals of a population.     

Class I beta-1,3-glucanase and chitinase are expressed in the micropylar endosperm of tomato seeds prior to radicle emergence.

beta-1,3-Glucanase (EC 3.2.1.39) and chitinase (EC 3.2.1.14) mRNAs, proteins, and enzyme activities were expressed specifically in the micropylar tissues of imbibed tomato (Lycopersicon esculentum Mill.) seeds prior to radicle emergence. RNA hybridization and immunoblotting demonstrated that both enzymes were class I basic isoforms. beta-1,3-Glucanase was expressed exclusively in the endosperm cap tissue, whereas chitinase localized to both endosperm cap and radicle tip tissues. beta-1,3-Glucanase and chitinase appeared in the micropylar tissues of gibberellin-deficient gib-1 tomato seeds only when supplied with gibberellin. Accumulation of beta-1,3-glucanase mRNA, protein and enzyme activity was reduced by 100 microM abscisic acid, which delayed or prevented radicle emergence but not endosperm cap weakening. In contrast, expression of chitinase mRNA, protein, and enzyme activity was not affected by abscisic acid. Neither of these enzymes significantly hydrolyzed isolated tomato endosperm cap cell walls. Although both beta-1,3-glucanase and chitinase were expressed in tomato endosperm cap tissue prior to radicle emergence, we found no evidence that they were directly involved in cell wall modification or tissue weakening. Possible functions of these hydrolases during tomato seed germination are discussed.     

Water stress and galactomannan breakdown in germinated fenugreek seeds. Stress affects the production and the activities in vivo of galactomannan-hydrolysing enzymes

Imposition of water stress on germinated fenugreek (Trigonella foenum-graecum L.) seeds and isolated fenugreek endosperms after the beginning of galactomannan mobilisation caused a reduction in the rate of breakdown of the polysaccharide relative to unstressed controls. The activities, measured in vitro, of the three hydrolytic enzymes involved in the breakdown process (-d-galactosidase, EC 3.2.1.22;endo--d-mannanase, EC 3.2.1.78;exo--d-mannanase, EC 3.2.1.25) were not decreased. Although there was some accumulation of galactomannan-hydrolysis products in endosperms under stress, there was no clear correlation between sugar levels and the inhibition of galactomannan breakdown. When water stress was applied to fenugreek seeds after germination but before the beginning of galactomannan hydrolysis, both galactomannan breakdown and the development of the hydrolytic enzyme activities were inhibited. Washing of newly germinated seeds for 2 h in water prior to the imposition of stress gave partial relief of the inhibition of galactomannan mobilisation, partial recovery ofendo--d-mannanase levels, and full recovery of -d-galactosidase levels. It is argued: 1) that water stress after germination but before the beginning of galactomannan hydrolysis inhibits the production of hydrolytic enzymes in the endosperm, probably via decreased removal at lowered water content of diffusible inhibitory substances; and 2) that water stress after the beginning of galactomannan hydrolysis decreases the rate of galactomannan breakdown in vivo principally via decreased diffusion at lowered water content of enzymes from the aleurone layer through the storage tissue of the endosperm.Abbreviation PEG polyethyleneglycol     

Effect of Salicylhydroxamic Acid on Endosperm Strength and Embryo Growth of Lactuca sativa L. cv Waldmann''s Green Seeds

Salicylhydroxamic acid (SHAM) stimulated germination of photosensitive lettuce (Lactuca sativa L. cv Waldmann's Green) seeds in darkness. To determine whether SHAM acts on the embryo or the endosperm, we investigated separately effects of SHAM on growth potential of isolated embryos as well as on endosperm strength. Embryo growth potential was quantified by incubating decoated embryos in various concentrations of osmoticum and measuring subsequent radicle elongation. Growth potential of embryos isolated from seeds pretreated with 4 millimolar SHAM was equal to that of untreated controls. Rupture strength of endosperm tissue excised from seeds pretreated with SHAM was 33% less than that of controls in the micropylar region. To determine if the embryo must be in contact with the endosperm for SHAM to weaken the endosperm, some endosperms were incubated with SHAM only after dissection from seeds. Rupture strength of SHAM-treated, isolated endosperms in the micropylar region was 25% less than that of untreated controls. There was no difference in rupture strength in the cotyledonary region of endosperm isolated from seeds treated with SHAM in buffer or buffer alone. SHAM therefore stimulates germination not by enhancing embryo growth potential, but by weakening the micropylar region of the endosperm enclosing the embryo.