Planta (1985) 163:133-140 Planta �9 Springer-Verlag 1985

Xyloglucan (amyloid) mobilisation in the cotyledons of Tropaeolum majus L. seeds following germination

Mary Edwards 1, Iain C.M. Dea 2, Paul V. Bulpin 2 and J.S. Grant Reid 1 1Department of Biological Science, University of Stirling, Stirling FK9 4LA, and 2Unilever Research Laboratories, Colworth House, Sharnbrook, Bedford MK44 1LQ, UK

Abstract. The levels of cell-wall xyloglucan (amy- loid) in nasturtium (Tropaeolum majus L.) coty- ledons were monitored during a 28-d period cover- ing seed imbibition, germination and early seedling development. The activities of the following enzymes capable of hydrolysing the glycosidic link- ages in the xyloglucan were assayed in cotyledon extracts over the same period: endo - (1-+4)-fl- glucanase (EC 3.2.1.4), fl-glucosidase (EC 3.2.1.21), a-xylosidase and fl-galactosidase (EC 3.2.1.23). The endo-fl-glucanase was assayed viscometrically using xyloglucan as substrate, and the three glycosidases using appropriate p-nitrophenyl- glycosides. Alpha xylosidase and/~-galactosidase, the enzymes which would be expected to hydrolyse the side-chains from the xyloglucan molecule, were also assayed using xyloglucan as substrate. Under our culture conditions, xyloglucan levels remained constant at 30 mg per cotyledon pair for 7 d, that is until 3 d after germination: thereafter, the amount of xyloglucan diminished to zero in a 12-d period. The most rapid period of depletion was between days 9 and 13. The mobilisation of all reserve substances from the cotyledons resulted in a weight-loss of 92 rag: xyloglucan, therefore, is an important storage substance, representing 33% by weight of the seed's substrate reserves. It is a cell- wall storage polysaccharide. Xyloglucan mobilisa- tion was accompanied by a 17-fold increase in endo-j~-glucanase activity, a 7-fold increase in j%galactosidase and an 8-fold increase in a-xy- losidase activities, all determined using xyloglucan as substrate. All three activities began to increase at day 5, peaked at days 12-14 when the most rapid phase of xyloglucan breakdown was over, and had declined to zero by days 22-25. The levels of theses enzymes have been shown to be consistent with their being responsible for xyloglucan hydrolysis in vivo. Nitrophenyl-fl-galactosidase activity in-

creased up to day 3, remained constant and then increased again 2.5-fold from day 5, peaking at day 11. Nitrophenyl-fl-glucosidase remained relatively constant up to day 16 and then decreased to zero by day 25. Nitrophenyl-a-xylosidase activity was not detected.

Key words: Amyloid (seed) - Endo-/~-glucanase - fl-Galactosidase - Germination (seed) -/~-Glu- cosidase - Tropaeolum (amyloid mobilisation) - Xyloglucan - a-Xylosidase.

Introduction

As early as 1839, Schleiden (quoted by Vogel and Schleiden 1839) had noted that the thickened cell walls in some seeds could be stained blue with iodine, and he named the substance responsible for this starch-like reaction "Amyloid". Amyloids have since been shown to occur widely in seeds (Winterstein 1893; Kooiman 1960a) and to disap- pear following germination (Godfrin 1884). Win- terstein (1893) extracted the amyloid from nastur- tium (Tropaeolum majus) seeds and demonstrated that on acid hydrolysis it released glucose, xylose and galactose. The amyloids can be quantitatively extracted from seeds with dilute sodium-hydroxide solutions (Kooiman 1960b), but once isolated they are largely water-soluble, forming viscous solu- tions. Detailed stuctural studies have been carried out on amyloids from only four species: Tamarin- dus indica (Leguminosae-Caesalpinioideae; amy- loid stored in the cotyledons) (Kooiman 1961); Tropaeolum majus (Tropaeolaceae; amyloid in the cotyledons) (Le Dizet 1972); Impatiens balsamina (Balsaminaceae; amyloid in the cotyledons) (Cour- tois and Le Dizet 1974); and Annona muricata

1"34 M. Edwards et al.: Xyloglucan mobilisation in Tropaeolum seeds

OH OH

HOHzC /o oH / CH~OH C H~ CH~OH r

0 0 ~ .

OH OH OH OH

Fig. 1. Structural features of seed xyloglucans. A (1 - , 4)/?-lin- ked D-glucan backbone is substitued by D-xylopyranosyl and /?-D-galactopyranosyl (I ~ 2) D-xylopranosyl side chains. Both types of side- chain are linked 1 ~ 6-a to D-glucose. The distri- bution of side-chains is unknown

(Annonaceae; amyloid in the endosperm) (Kooi- man 1967). All four amyloids have the structural features illustrated in Fig. 1. A linear (1 ~ 4)-fl-D- glucan (cellulosic) backbone carries two types of substituents, both linked 1 ~ 6 to the backbone: single-unit a-D-xylopyranosyl groups and ~-D-ga- lactopyranosyl (1 --* 2) D-xylopyranosyl disaccha- ride side-chains. The ratio galactose: xylose: glu- cose in the polysaccharides from Tamarindus and Tropaeolum is 1 : 2: 3, whereas in those from Impa- tiens and Annona it is 1:2:4-5 and 1 : 1:4, respec- tively. It is generally assumed that all seed anayloids are galactoxyloglucans although they are usually termed "xyloglucans" (Reid 1984).

Apart from early reports by Heinricher (1888), Reiss (1889) and others that the amyloids in a variety of seeds, including Impatiens balsamina, Tropaeolum majus and Cyclamen europaeum, disap- pear following germination, there is little informa- tion in the literature relevant to the biochemistry of xyloglucan metabolism in seeds. Gould et al. (1971) observed that a "pectic" xyloglucan is mobilised from mustard cotyledons following germination; xyloglucan is, however, not a major component of the mustard seed. In the context of a detailed inves- tigation of the enzymology of xyloglucan mobilisa- tion in germinated seeds, we now report a cor- relative study of xyloglucan levels in TropaeoIum majus cotyledons and the activities in cotyledon extracts of enzymes capable of hydrolysing the gly- cosidic linkages in the polysaccharide.

Material and methods

Seed germination and growth conditions. Nasturtium (Tropaeo- lure majus L. seeds, tall single mixed), were purchased from Clause (UK), Charvil-Reading, UK. The batch supplied (No. 13875) had an average germination of 84% and a very high proportion of seeds without adhering fruit-coat tissue. A low proportion of "coated" seeds is advantageous since the coat has to be removed manually before homogenisation.

Dry seeds were set out (at time 0) in trays 21 - 35 cm z con- taining autoclave-sterilised wet "Vermiculite" ("Micafill" fibre- free insulation; Dupre Vermiculite Ltd., Hertford, Herts, UK) depth 4 cm. The seeds (150 to a tray) were placed on the surface of the Vermiculite, covered with a further thin layer of Ver- miculite and watered thoroughly. Trays were placed in a plant growth chamber with a 12-h day/12-h night regime; photon fluence rate 850 gmol m-2s -1 at tray level; temperature 15-20 ~ C. Trays were covered with transparent polythene sheeting until shoots appeared. Watering (with tap water) was at 2-d intervals.

Sampling of seeds and seedlings. Trays were selected at random. Seeds or seedlings for analysis were selected as quadruplicate samples fi'om different areas of a tray. Until day 5, when the seed population had completed germination, no seeds were rejected unless they were clearly damaged. After day 5 only germinated seeds were used.

Fresh and dry weights. Quadruplicate samples of 10 cotyledon pairs, with testae attached, were weighed immediately on har- vesting and after drying at 110~ for 48 h.

Xyloglucan levels. Quadruplicate samples of 10 cotyledon pairs, with testae attached, were each treated as follows. Thorough grinding in a mortar with sand (5 g) and water (10 ml) was followed by quantitative transfer to a 100-ml glass centrifuge tube containing NaBH4 (50 rag). Quantitative transfer was assured by rinsing the mortar with 2 M NaOH (90 ml).

The contents of the tube were stirred and alkaline extrac- tion allowed to proceed for 60 rain at 100~ (occasional stir- ring). Only one alkaline extraction was used; experiments with multiple extractions showed that over 94% of the xyloglucan was obtained at the first extraction. The sodium borohydride was used to convert all reducing end-groups of saccharides to alditols, thus preventing sequential stripping of monosac- charide residues from the reducing ends of polysaccharides (Aspinall et al. 1961). After centrifugation (1400 g; 10 min) the supernatant was poured into glacial acetic acid: ethanol 1:10, v/v; 200 ml). CAUTION: evolution of Hz and HzS, from NaBH4 and endogenous sulphur-compounds, respectively. The precipitate of crude xyloglucan was allowed to stand for 30 min and collected by centrifugation (1400 g; 10 min).

Purification was via the insoluble xyloglucan copper- complex (Rao 1959). The crude polysaccharide was dissolved or suspended in water (80 ml), and dissolution completed by adding 10 ml of the alkaline component of Fehling's solution (Fehling's B) and heating at 100~ for 30 min. On cooling, the copper-containing component of Fehling's solution (Fehling's A; 10 ml) was added and the mixture stirred vigorously. Any blue precipitate of xyloglucan copper-complex was collected by centrifugation (1400 g; 10 rain) and dissolved in an appropriate volume (20-150 ml) of a mixture (1:4, v/v) of glacial acetic acid and water. The decomplexed xyloglucan was recovered by add- ing two volumes of ethanol and collecting the colourless pre- cipitate by centrifugation (1400 g; 10 min). The precipitated xyloglucan was resuspended in 65% (v/v) ethanol, allowed to stand overnight, collected by centrifugation, dissolved or sus- pended in water, fi-eeze-dried and weighed.

Cotyledon extracts for enzyme assays. Each quadruplicate sample of 15 cotyledon pairs (with testae attached) was treated as follows. All operations were carried out at 0~4~ unless otherwise stated. The cotyledons were ground in a mortar with sand (1 g), insoluble polyvinylpolypyrrolidone (Sigma, Poole, Dorset, UK; 0.15 g) and 1 M sodium-chloride solution, unbuf- fered (15 ml). Homogenates were allowed to stand for 1 h and

M. Edwards et al. : Xyloglucan mobilisation in Tropaeolum seeds 135

spun (26 000 g; 30 rain). A sample of the clear liquid between the pellet and the surface layer of lipid (if present) was used for enzyme assays. A lipid layer was present only until day 14, indicating mobilisation of storage lipids.

Nitrophenylglycosidase assays. The standard assay mixture con- tained the appropriate p-nitrophenylglycoside (Sigma), 5 - 10 -2 M in water (0.1 ml); McIlvaine phosphate-citrate buffer, pH 4.5 for fl-galactosidase and pH 5.0 for fl-glucosidase (0.3 ml); and enzyme extract appropriately diluted with 1 M NaC1 (0.1 ml); substrate concentration in assay = 10 -2 M. Incubation (at 30~ was started by adding the substrate, allowed to proceed for 15 rain and stopped by the addition of 0.1 M NazCO3 (5 ml). Extinctions were read at 400 nm and activities calculated using a molar extinction coefficient of 1.84 - 104 for p-nitrophe- nol in 0.1 M NazCO3.

All nitrophenylglycosidase assays were carried out on the same working day as the extracts were prepared.

Beta-galaetosidase and a-xylosidase assays using xyloglucan as substrate. Tamarind seed xyloglucan was prepared by hot-wa- ter extraction of defatted commercial tamarind flour and puri- fied by ethanol-precipitation: copper-complexation was not used to avoid any possibility of enzyme inhibition by traces of copper ions.

A 1.2% (w/v) solution of tamarind seed xyloglucan dissoled in McIlvaine phosphate-citrate buffer pH 5.0 (5,0 ml) was mix- ed with enzyme (1.0 ml). After 5 min, a sample was removed and heated on a boiling-water bath for 10 min. After a further 4-h incubation the reaction was stopped by heating for 10 rain on the boiling-water bath. Beta-galactosidase and a-xylosidase activities were determined by assaying the D-galactose and pen- tose concentrations in the "5-rain" and "4-h 5-min" samples and calculating rates of release over 4 h. Separate control expe- riments had shown that D-galactose and pentose release were linear over 6 h, even at the highest activities encountered. Initial incubation and pentose determinations were carried out in the same working day as the enzyme had been isolated. Galactose levels were assayed on the following day.

D-Galactose was determined using fl-D-galactose dehydro- genase (Sigma) (Kurz and Wallenfels 1974). Boiled enzyme incubations were first clarified by centrifugation (11 600 g; 10 min). Galactose standards up to 1 mM were prepared freshly each day and assayed by the same method. No interference by D-xylose at the levels actually present was obtained.

Pentose was assayed by the p-bromoaniline method of Roe and Rice (1948), with a single modification. After addition of the chromogenic reagent and heating for 10 rain at 70~ the tubes were incubated for 60 min at 30~ in the dark before cooling to room temperature and reading extinctions at 520 nm. D-Xylose standards were prepared freshly each day in the same buffer as the enzyme incubation; high levels of NaCI affected the final optical-density values. Xyloglucan formed a stringy precipitate during this assay, but did not interfere provi- ded that none of the precipitate was transferred to the spectro- photometer cuvette. D-Galactose at the levels actually present did not interfere with the assay.

Viscometric assay of endo-(1 --* 4)-fl-D-glucanase. A 1.2% (w/ v) solution of tamarind seed xyloglucan was added to a thick- walled glass tube immersed in a water bath at 30~ J: 0.1~ Enzyme solution, diluted if necessary and pre-warmed to 30~ (0.4 ml), was added. Viscometric flow-times were determined using a graduated pipette (0.1 ml volume). An initial flow-time of about 30 s was suitable. Thereafter, further flow-time read- ings were taken at 5-rain intervals until a linear relationship between flow-time and elapsed time had been established (usu- ally 60 rain).

The activity of the enzyme, in arbitrary units, was calcula- ted from the plot of flow-time against elapsed time, extrapola- ted to give the flow-time at time 0 (time of mixing). The activity was defined as 100/t 0.8, where t 0.8 is the elapsed time taken for the flow-time to decrease to 0.8 of its value at time 0.

Provided that enzymes were diluted until a linear relation- ship between elapsed and flow-time was obtained, the activity was directly proportional to the amount of enzyme.

Protein was determined, after appropriate dilution, by the Coo- massie-Blue method of Sedmak and Grossberg (1977), using bovine serum albumin as standard.

D-Glucose was determined by the hexokinase/glucose-6-phos- phate dehydrogenase procedure (Bergmeyer et al. 1974).

Reducing power was determined by the direct ferricyanide me- thod (Halliwell and Riaz 1970).

Results and discussions

Xyloglucan levels in relation to irnbibition, germina- tion and seedling development. T h e n a s t u r t i u m seed (TropaeoIum majus) is n o n - e n d o s p e r m i c wi th h y p o g e a l ge rmina t ion . Fig. 2 i l lustrates the deve- l o p m e n t a l changes o c c u r r i n g over a 28-d pe r iod c o m m e n c i n g wi th the p lac ing o f d ry seeds on wet vermicul i te , and the a c c o m p a n y i n g changes in the fresh and d ry weights o f the c o t y l e d o n s are s h o w n in Fig. 3. W a t e r - i m b i b i t i o n was comple t e by d a y 3, whils t radicle p r o t r u s i o n o c c u r r e d be tween days 3 and 5. Mob i l i s a t i on o f bu lk reserves f r o m the co ty - ledons o c c u r r e d be tween days 7 and 22.

T h e a m o u n t o f xy log lucan in the c o t y l e d o n s (Fig. 4) r ema ined stat ic du r ing imbib i t ion and ger- m i n a t i o n , bu t b e g a n to d imin ish a b o u t day 8. Be- tween days 9 and 13, x y l o g l u c a n was very rap id ly depleted, and mob i l i s a t i on was comple t e by d a y 19.

The pe r iod o f xy log lucan dep le t ion corre- sponds overa l l to the pe r iod du r ing wh ich the co ty - l edon reserves are mobi l i sed ( c o m p a r e Figs. 3 a n d 4). F u r t h e r m o r e , the ave rage loss o f c o t y l e d o n a r y d ry we igh t per seed over t ha t pe r iod is 92 rag, and the average a m o u n t o f xy log lucan mobi l i sed per seed is 30 nag. The p o l y s a c c h a r i d e mus t , therefore , be cons ide red a quan t i t a t ive ly i m p o r t a n t subs t ra te reserve in the n a s t u r t i u m seed. W e suggest t ha t it be inc luded in the g r o u p o f m a j o r cell-wall s to rage po ly saccha r ide s (Meier a n d Re id 1982).

Hydrolytic activities in relation to xyloglucan deple- tion. T h e hyd ro ly t i c b r e a k d o w n o f n a s t u r t i u m seed x y l o g l u c a n (Fig. 1) w o u l d requi re three p r inc ipa l e n z y m a t i c activit ies: a f l -galac tos idase to r e m o v e s ide-chain ga lac tosy l units, an a -xy los idase to hy- d ro lyse s ide-chain xylosyl residues ei ther di rect ly o r af ter r e m o v a l o f ga lac tose , and an endo-

136 M. Edwards et al.: Xyloglucan mobilisation in Tropaeolum seeds

Fig. 2. Morphological changes associated with germination and seedling development of nasturtium (Tropaeolum majus). Arrow indicates protruded radicle

0 , 3

0 , 2

o,1

I I I I I 214 218 0 4 8 12 16 20

Days a f te r planting

Fi~. 3. Fresh ( � 9 and dry weights ( � 9 of cotyledons (+ testae) of nasturtium seeds/seedlings, g = completion of germination (radicle protrusion). Error bars represent 2 x SE

20

E

o

x 10

3O

. . . . i 0 4 8 12 16 20 24 28

D a y s a f t e r p lant ing

Fig. 4. Xyloglucan in cotyledons of nasturtium seeds/seedlings - Error bars represent 2 x SE

M. Edwards et al.: Xyloglucan mobilisation in Tropaeolum seeds 137

(1--, 4)-fl-i>glucanase to hydrolyse the internal linkages of the backbone. A fl-glucosidase capable of hydrolysing cello-oligosaccharides to glucose might also be of importance. All of these activities were detected in cotyledon homogenates.

Endo-(1 ~ 4)-/?-glucanases are generally assay- ed specifically in the presence of other hydrolytic activities by viscometric techniques, which are par- ticularly sensitive to endo- as opposed to exo- cleavage. Synthetic cellulose derivatives such as carboxymethyl- and hydroxyethylcelluloses are normally used as substrates, but they suffer from the disadvantage that the measured rate of reac- tion may bear little relation to the rate of hydroly- sis of the natural substrate. In these studies, tama- rind seed xyloglucan (Kooiman 1961), which has a structure almost identical to that of the nasturtium seed polysaccharide (Le Dizet 1972), was used as viscometric substrate.

Although fl-galactosidases capable of hydroly- sing phenolic glycosides and oligosaccharides have been reported relatively frequently in plant ex- tracts (for example Li et al. 1975), their action on xyloglucans has not generally been tested, fl-Ga- lactosidase was therefore assayed by two procedu- res - the conventional nitrophenylgalactosidase as- say, and by monitoring the release of i>galactose from xyloglucan.

The activity of a-xylosidase was similarly assa- yed using p-nitrophenyl-a-D-xylopyranoside, and by determining pentose-release from xyloglucan. Activity was detected only by the latter procedure (Fig. 5). It is noteworthy in this context that the pattern of breakdown of primary cell-wall xyloglu- can by wall-bound enzymes from soybean hypoco- tyls indicated the participation on an a-xylosidase (Koyama et al. 1981). The same enzyme prepara- tion broke down a heptasaccharide fragment (Xyl3Glc2) from xyloglucan by a mechanism which again indicated the cleavage of xylosyl units by an a-xylosidase. Yet the preparation hydrolysed neither isoprimverose (a-d-xyl 1 --* 6 i>glc) nor p- nitrophenyl-a-D-xylopyranoside (Koyama et al. 1983). The scarcity of information in the litera- ture concerning a-xylosidases may reflect the in- ability of many such enzymes to hydrolyse nitro- phenyl-a-xylosides. This is the first direct demon- stration of a-D-xylosidase activity in higher plants.

All of the enzyme assays were optimised with respect to pH, and the nitrophenylglycosidase as- says were carried out using saturating substrate concentrations. When xyloglucan was used as sub- strate the concentration used was limited by the solubility of the polysaccharide and the high visco- sity of its solutions.

. 36

�9 32

~= .28

== -~ .24 r

"p> �9 20

,16

~> .12 _=

. 08

'04

0 4 8 12 16 20 24

Days after planting

4

2

28

36

34

32

30

28

26

24

22 "i

,o~ v

12 I

8

Fig. 5. Endo-/~-glucanase (�9 ( � 9 and a-xylosi- dase ( � 9 activities in extracts of nasturtium cotyledons, assayed using xyloglucan as substrate. Error bars represent 2 x SE

6 O

50

40

| 30

o~ 2 0

to

Q .

0 4 8 12 16 20 24 28

Days after planting

Fig. 6. Nitrophenylglycosidase activities in extracts of nastur- tium cotyledons. �9 B,p-nitro- phenyl-/?-glucosidase. Error bars represent 2 x SE

The variations with time of all the enzymatic activities mentioned above are summarised in Figs. 5 and 6. Activity units are expressed per cotyledon pair; specific activities have not been used because of the steady decline with time of the protein con-

138 M. Edwards et al.: Xyloglucan mobilisation in Tropaeolum seeds

6.0

5.r

4,r

8 3"0

2.11

.E 1.1]

a .

I �9 i A d *

4 8 12 16 20 24 28 Days after planting

Fig. 7. Protein levels in extracts of nasturtium cotyledons. Error bars represent 2 x SE

tents of the extracts (Fig. 7). This decline is almost certainly due to the progressive mobilisation of storage proteins. The data in Fig. 7 allow specific activities to be calculated.

There is a clear correspondence of xyloglucan utilisation (Fig. 4) with the levels of endo- fl-glucanase, and of fl-galactosidase and a-xylosi- dase assayed using xyloglucan as substrate (Fig. 5). At day 0, all three activites are present at low, but detectable, levels. By day 7, just prior to any measurable xyloglucan depletion (Fig. 4), all three activities are above their base levels. Subse- quently they increase rapidly to peak at days 12 (a-xylosidase), 13 (endo-fl-glucanase) and 14 (fl-ga- lactosidase), declining rapidly thereafter to reach zero by day 25. The period of rapid increase in the enzymatic activities corresponds to that of rapid xyloglucan utilisation, and all the polysaccharide has disappeared from the cotyledons before the activities fall to zero. Endo-glucanase activity inc- reases 17-fold over its initial base level whilst fl-galactosidase and a-xylosidase increase 7- and 8- fold, respectively.

The time course of fl-galactosidase activity as- sayed using p-nitrophenylgalactoside as substrate (Fig. 6) does not exactly parallel that determined using xyloglucan as substrate (Fig. 5). During im- bibition (days 0 to 3), the level of nitrophenylga- lactosidase activity increases, but there is no corre- sponding increase in the release of galactose from the polymeric substrate (compare Figs. 5 and 6). Subsequently, both activities increase, and the ni- trophenyl-fl-galactosidase activity peaks somewhat earlier and more sharply than the xyloglucan-de- grading fl-galactosidase; both activities show a si-

milar decline from day 15. The results indicate the presence of multiple fl-galactosidases not all of which are capable of hydrolysing the fl-galactosidic link of the polymeric substrate.

fl-Glucosidase activity does not increase during the period of xyloglucan mobilisation (Fig. 6), and declines from day 15 onwards.

The close correlation between the mobilisation of xyloglucan (Fig. 4) and the activities of endo- fl-glucanase, a-xylosidase and fl-galactosidase (Fig. 5) leaves little doubt that they are responsible for xyloglucan hydrolysis in vivo. The status of the fl-glucosidase (Fig. 6) is less clear since its activity remains relatively constant up to day 15. Neverthe- less, it is probable that such an activity would hydrolyse cellooligosaccharides produced by the action of the endo-fl-glucanase.

The maximum level of fl-galactosidase activity in Fig. 5 (0.35 nkat per seed) is greater than that required to account for the maximum observed (Fig. 4) rate of xyloglucan breakdown. The maxi- mum level of a-xylosidase activity in Fig. 5 (0.14 nkat per seed) is 72% of the value required theoreti- cally. In calculating theoretical values a maximum breakdown rate of 30 mg in 108 h was read from Fig. 4, and a galactose: xylose: glucose ratio of 1 : 2: 3 for tamarind seed xyloglucan was used. Simi- lar calculations could not be done for the endo-fl- glucanase since the units obtained viscometrically are arbitrary.

A better assessment of whether or not the enzy- matic activities measured in vitro were sufficient to explain the observed rate of xyloglucan depletion in vivo was obtained using nasturtium xyloglucan as substrate. An amount equivalent to that from 10 ungerminated seeds was incubated with an amount of enzyme extract equivalent to 10 cotyledon pairs from 14-d-germinated seedlings (Fig. 8). Levels of galactose, xylose and total reducing power were monitored over 6 h and at 24 h. Under these condi- tions the rates of galactose and xylose production (7.4 mM and 6.9 mM, respectively, in 24 h) were sufficient to explain the rate of xyloglucan deple- tion calculated from Fig. 4 (2.7 mM and 6.6 mM, respectively, in 24 h).

It is interesting to note that the level of reduc- ing power can be almost completely accounted for by the sum of the galactose and xylose, although endo-fl-glucanase activity brings about a very ra- pid reduction in viscosity (Fig. 8). In a second, identical experiment glucose levels were measured over a 6-h period and were found to remain below 0.2 mM. These observations indicate that, in the initial stages of xyloglucan depletion at any rate, fl-glucosidase does not play an important role and

M. Edwards et al. : Xyloglucan mobilisation in Tropaeolum seeds 139

14-0

t 12.o / 12o

o / ,~ 10.0 100

o= / -

~-2 "- g g

u. _o 8 .~ 8o

": , o

3 , 0

2.0 20

1 ,0

0 1 2 3 4 5 8 24

Incubation T ime (h)

Fig. 8. Release of galactose ( � 9 xylose ([]) and reducing power (11) from nasturtium xyloglucan by enzyme extracted from 14- d nasturtium cotyledons. Enzyme equivalent to 10 cotyledon pairs was incubated with nasturtium xyloglucan (Le Dizet 1972) equivalent to that from 10 ungerminated seeds (300 rag). Total volume = 25 m[; substrate concentration = 12 mg mk I. A control incubation contained enzyme, but no substrate. Q, Viscometer flow time

that the oligosaccharide products of the endo-/?- glucanase are of relatively high degree of polyme- risation.

Our results confirm Heinricher's (1888) obser- vation that the amyloid or galactoxyloglucan pre- sent in the cotyledons of the nasturtium seed is mobilised following germination, and allow the conclusion that it is a major reserve in the seed, accounting for 33% by weight of the reserves sto- red in the cotyledons. The mobilisation of the poly- saccharide is further shown to be mediated by three hydrolytic activities: endo-~-glucanase,/Lgalacto- sidase, and a-xylosidase. It is not yet clear how these enzymes interact cooperatively to break down the polysaccharide, since this requires an exact knowledge of their modes of action and substrate specificities. Of interest here is our finding that a purified preparation of the nasturtium seed endo- /~-glucanase shows virtually no activity towards carboxymethyl or hydroxyethylcelluloses whilst retaining high xyloglucanase activity.

It is noteworthy that the storage of amyloids occurs only in dicotyledonous seeds (Meier and Reid 1982), and that the primary cell walls of di- cotyledonous plants contain a structural "hemicel- lulose" polysaccharide of the xyloglucan type (Dar- vill et al. 1980). The only difference in molecular structure between the "reserve" xyloglucans of di- cotyledonous seeds and the "structural" xyloglu- cans of the primary cell walls of vegetative tissues

in dicotyledons is the presence in the latter of a low percentage (6-8%) of I>fucopyranosyl units linked 1 ~ 2 to the side-chain D-galactose (Darvill et al 1980). The massive accumulation of cell-wall xyloglucan which must occur during seed matura- tion in nasturtium and other amyloid-storing seeds could be brought about by a relatively simple regu- lation of the turnover of xyloglucan which appears to occur naturally in growing primary cell walls (Labavitch and Ray 1974). This would certainly be consistent with our observation that the three key enzymes of xyloglucan breakdown are present in nasturtium seeds prior to xyloglucan mobilisation, although at very low levels. Terminal cotyledonary development following germination is accompa- nied by a large increase in these activities.

R e f e r e n c e s

Aspinall, G. O., Greenwood, C. T., Sturgeon, R. J. (1961) The degradation of xylans by alkali. J. Chem. Soc., 3667-3674

Bergmeyer, H.U., Berndt, E., Schmidt, F., Stark, H. (1974) D-glucose determination with hexokinase and glucose-6- phosphate dehydrogenase. In: Methods of enzymatic analy- sis, pp. 1196 1201, Bergmeyer, U., ed. Verlag Chemie, Wein- heim

Courtois, J. E., Le Dizet, P. (1974) l~tude de la structure de la galactoxyloglucane (amyloide) des graines de balsamine (Im- patiens balsamina). C. R. Acad. Sci. Set_ C 278, 81-83

Darvill, A., McNeil, M., Albersheim, P., Delmer, D. P. (1980) The primary cell walls of flowering plants. In: The biochemi- stry of plants, vol. 1: The plant cell, pp. 91-160, Tolbert, N.E., ed. Academic Press, New York London

Godfrin, M. J. (1884) L'anatomie compar6e des cotyl6dons et de l'albumen. Ann. Sci. Nat. 19, 5 158

Gould, S. E. B., Rees, D. A., Wight, N. J. (1971) Polysacchari- des in germination. Xyloglucans (amytoids) from the cotyle- dons of white mustard. Biochem. J. 124, 47-53

HalliweI1, G., Riaz, M. (1970) The formation of short fibres from native cellulose by components of Trichoderma konin- gii cellulase. Biochem. J. 116, 3542

Heinricher, E. (1888) Zur Biologie der Gattung Impatiens. Flora (Jena) 71, 163-185

Kooiman, P. (1960a) On the occurrence of amyloids in plant seeds. Acta Bot. Neerl. 9, 208-219

Kooiman, P. (1960b) A method for the determination of amy- loid in plant seeds. Rec. Trav. Chim. Pays-Bas 79, 675-678

Kooiman, P. (1961) The constitution of Tamarindus-amyloid. Rec. Trav. Chim. Pays-Bas 80, 849-865

Kooiman, P. (1967) The constitution of the amyloid from seeds of Annona muricata L. Phytochemistry 6, 1665-1673

Koyama, T., Hayashi, T., Kato, Y., Matsuda, K. (1981) Degra- dation of xyloglucan by wall-bound enzymes from soybean tissue. I. Occurrence of xyloglucan-degrading enzymes in soybean cell wall. Plant Cell Physiol. 22, 1191 1198

Koyama, T., Hayashi, T., Kato, Y., Matsuda, K. (1983) Degra- dation of xyloglucan by wall-bound enzymes from soybean tissue. II. Degradation of the fragment heptasaccharide from xyloglucan and the characteristic action pattern of the a-D- xylosidase in the enzyme system. Plant Cell Physiol. 24, 155-I62

Kurz, G., Wallenfels, K. (1974) D-Galactose. U.V.-Assay with galactose dehydrogenase. In: Methods of enzymatic analy-

140 M. Edwards et al.: Xyloglucan mobilisation in Tropaeolum seeds

sis, pp. 1279-1282, Bergmeyer, U., ed. Verlag Chemie, Wein- heim

Labavitch, J. M., Ray, P. M. (1974) Turnover of cell-wall poly- saccharides in elongating pea stem segments. Plant Physiol. 53, 669-673

Le Dizet, P. (1972) Quelques pr+cisions sur la structure de l'amylo'ide de capucine. Carbohydr. Res. 24, 505-509

Li, S.-C., Mazzotta, M., Chien, S.-E., Li, Y.-T. (1975) Isolation and characterization of jack bean ~-galactosidase. J. Biol. Chem. 250, 6786-6791

Meier, H., Reid, J. S. G. (1982) Reserve polysaccharides other than starch in higher plants. In: Encyclopedia of plant phy- siology, N. S., vol. 13A: Plant carbohydrates, pp. 418-471, Loewus, F. A., Tanner, W., eds. Springer, Berlin Heidelberg New York

Rao, P. S. (1959) Tamarind. In: Industrial gums, pp. 461-504, Whistler, R. L., ed. Academic Press, New York London

Reid, J. S. G. (1984) Cell wall storage carbohydrates in seeds

biochemistry of the seed "gum and hemicelluloses". Adv. Bot. Res. 11 (in press)

Reiss, R. (1889) Uber die Natur der Reservecellulose und fiber ihre Aufl6sungsweise bei der Keimung der Samen. Land- wirtsch. Jahrb. 18, 711-765

Roe, J. H., Rice, E. W. (1948) A photometric method for the determination of free pentoses in animal tissue. J. Biol. Chem. 173, 507-512

Sedmak, J. J., Grossberg, S. E. (1977) A rapid, sensitive and versatile assay for protein using Coomassie Brilliant Blue G250. Anal. Biochem. 79, 544-552

Vogel, T., Schleiden, M. J. (1839) Ober das Amyloid, eine neue Pflanzensubstanz. Poggendorf's Ann. Physik Chem., 46, 327-330

Winterstein, E. (1893) Uber das pflanzliche Amyloid. Hoppe- Seyter's Z. Physiol. Chem. 17, 353-380

Received 5 June; accepted 6 July 1984