Influence Sulfur Nutrition on Developmental Patterns of ...seeds showed that a high relative level of total vicilin (vicilin plus convicilin) synthesis was maintained throughout the

Plant Physiol. (1984) 75, 651-6570032-0889/84/75/0651 /07/$0 1.00/0

Influence of Sulfur Nutrition on Developmental Patterns of SomeMajor Pea Seed Proteins and Their mRNAs

Received for publication December 28, 1983

PETER M. CHANDLER, DONALD SPENCER,* PETER J. RANDALL, AND THOMAS J. V. HIGGINSDivision ofPlant Industry, Commonwealth Scientific and Industrial Research Organization, Canberra,A.C.T. 2601, Australia

ABSTRACI?

In addition to the marked reduction in legumin synthesis and leguminmRNA levels reported earlier (Chandler, Higgins, Randall, Spencer 1983Plant Physiol 71: 47-54), pulse labeling of S-deficient Pisum sativum L.seeds showed that a high relative level of total vicilin (vicilin plusconvicilin) synthesis was maintained throughout the entire phase ofprotein accumulation, whereas in nondeficient seeds vicilin synthesis islargely confined to the first half of this phase. Fractionation of pulse-labeled proteins on Na-dodecylsulfate-polyacrylamide gels showed thatthe synthesis of the Mr 50,000 family of vicilin polypeptides was in-creased and greatly extended in S-deficient seeds whereas that of convi-cilin was slightly reduced. Other changes apparent from pulse-labelingexperiments include a depression, to different degrees, in the synthesisof three major albumin polypeptides.The level of the mRNAs for seven major seed proteins was followed

throughout development of control and sulfur-deficient seeds. In all cases,the changes in each mRNA closely reflected the pattern of synthesis ofits corresponding polypeptide seen by pulse labeling. S-deficient seedsshowed an elevated level of Mr 50,000 vicilin mRNA which remainedhigh throughout seed formation, whereas legumin mRNA levels weregreatly reduced at all stages of development.When S-deficient plants were given an adequate supply of sulfate

midway through seed development, there was a shift toward the proteinsynthesis profile characteristic of healthy plants. The synthesis of legu-min and two albumins rapidly increased and the synthesis of Mr 50,000vicilin declined more slowly. Similar responses were seen in detached, S-deficient seeds supplied directly with adequate sulfate.

It has been shown in a range of legume and cereal species thatplants grown at suboptimal levels of S supply produce seed witha modified storage protein composition. Such seeds consistentlyshow a decrease in the proportion of the more S-rich proteins;for example, there is a selective reduction in the proportion oflegumin-type globulins in lupins (2) and peas (15), in the high-mobility gliadins in wheat (21), and in the B and D hordeins inbarley (17). In all cases there was an accompanying increase inthe proportion of the less S-rich seed protein components. In S-deficient peas, the reduction in the proportion of legumin wasassociated with a reduced rate of legumin synthesis and withreduced levels of legumin mRNA (5).To gain a better understanding ofthe way in which S deficiency

regulates the proportions of the major seed proteins, we haveexamined its effect on the synthesis and mRNA levels of a rangeof proteins during pea seed development. S deficiency resultedin both increases and decreases in the level of synthesis ofdifferent seed proteins and, in the case ofMr 50,000 vicilin there

was a change in the temporal pattern of its synthesis. Thesechanges were closely paralleled by changes in the level of thecorresponding mRNAs. Restoration of an adequate S supply toplants during seed development led to a return to a normalpattern of protein synthesis.

MATERIALS AND METHODS

Plant Material. Peas (Pisum sativum L.) line PI/G 086, se-lected from cv Greenfeast, were grown in artificially lit cabinetsat 20C with a 16-h photoperiod as described (13) and pods of aknown age, specified as DAF,' were used. The nutrient regimefor growing S-deficient plants and the characteristics of S-defi-cient seeds were the same as described earlier (5).

Recovery Experiments. S-deficient plants were watered with acomplete nutrient containing 1 mm MgSO4. S-deficient, detachedseeds were supplied with 12 mM MgSO4 via the funicle.Measurement of Vicilin Synthesis. Immunoaffinity chroma-

tography was carried out as described elsewhere (1). The extentof vicilin synthesis during pulse-labeling was expressed as theproportion of total extracted radioactive protein (TCA-insolubleprotein) which bound to IgG specific for the total vicilin fraction(vicilin plus convicilin) from mature peas.

Construction and Characterization of cDNA Clones. Poly (A)RNA from young (15 DAF), mid-stage (22 DAF), and old (27DAF) cotyledons was copied into double-stranded DNA andcloned into the Pst 1 site of pBR322 as previously described (5).Transformants were screened by colony hybridization (8) usingcDNA from each of the three developmental stages as well asinserts from selected recombinant plasmids obtained in an earliercloning. Nine different clones were eventually selected from atotal of 1600 clones with inserts, and these represent nine majorhybridization families in this population. The selected plasmidshave all been characterized by hybrid release translation andNorthern hybridization, and most ofthem by complete or partialDNA sequencing. They are described in Table I.The plasmids for legumin (pPS15-75) and lectin (pPS 15-104)

have been described in detail previously (5, 9). Single-strandedphage derivatives of pPS 15-112 have also been described, as wellas derivatives of two other plasmids which are homologous totwo plasmids in Table I. Thus, fdlO3-96B (4) was derived frompPS 15-96 which cross-hybridizes with pPS 15-21. Similarly,fdlO3-6T (4) was derived from pPS15-6 which cross-hybridizeswith PS 15-28.On the basis ofDNA and protein sequencing studies, three of

the remaining plasmids (pPS 15-24, pPS 15-71, and pPS 15-84)encode vicilin Mr 50,000 polypeptides (18). The inserts in theseplasmids differ however: each shows 75 to 85% nucleotide se-quence homology with the other two, so that under the usual

'Abbreviations: DAF, days after flowering; IgG, immunoglobulin G.

651 www.plantphysiol.orgon February 24, 2020 - Published by Downloaded from Copyright © 1984 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org

Plant Physiol. Vol. 75, 1984

Table I. Characterization ofSelected PlasmidsPlasmid Hybrid Release Corresponding mRNAa Reference

Designation Translation Product Seed Protein

pPS15-21 22k Albumin 1050 This studypPS15-24 50kb Vicilin 1600 This studypPS15-28 76kb Convicilin 2000 This studypPS15-71 50kb Vicilin 1600 This studypPS15-75 60k Legumin 1700 (5)pPS15-84 50kb Vicilin 1550 (18)pPS1 5-91 13k Albumin 660 This studypPS15-104 25k Lectin 1050 (9)pPS15-112 70k Unknown 1650 (4)

a Length ofmRNA (in nucelotides) to which plasmid hybridized.bIn hybrid release translation experiments plasmids for vicilin Mr

50,000 polypeptides also select convicilin mRNA and vice versa. Thesequence relationship between these plasmids is currently being investi-gated (P. M. Chandler, unpublished observations).

conditions of hybridization (Tm-20) they do not cross-hybridize(P. M. Chandler, unpublished observations). Whether the poly-peptides encoded by these plasmids remain as M, 50,000 poly-peptides in the mature seed, or are processed to the smallervicilin polypeptides (1 1, 18) is not yet known.The remaining plasmid, pPS15-91, selects an mRNA for an

Mr 13,000 polypeptide synthesized in vitro which corresponds toan Mr 11,000 polypeptide synthesized in vivo. This is processedto form an M, 8,000 albumin (T. J. V. Higgins, unpublishedobservations).Other Procedures. Procedures for in vivo labeling of cotyle-

dons, preparation of total extracts from cotyledons, as well aspolypeptide fractionation by SDS-PAGE and fluorography wereall described elsewhere (20). The '4C-amino acid mixture (Amer-sham) used in pulse-labeling detached cotyledons contained nomethionine or cysteine. Hybridization of labeled plasmids tosize-fractionated RNA and corrections for different loading levelsor efficiency of transfer were performed as previously described(5).

RESULTSVicilin Synthesis in Sulfur-Deficient Plants. Vicilin and legu-

min are the two major storage proteins of pea seeds. Since thelevel of the more S-rich legumin and albumin fractions is pref-erentially reduced by S deficiency ( 15, 16), it was of interest tomeasure the synthesis of the S-poor, vicilin fraction under thesesame conditions. Cotyledons were taken from S-deficient andcontrol plants during seed development and pulse labeled with'4C-amino acids. Proteins were then extracted and the contribu-tion of vicilin to total radioactive protein synthesis was deter-mined by immunochromatography using Sepharose 4B to whichwas covalently coupled IgG to a total vicilin (vicilin plus convi-cilin) fraction. Under our growing conditions, the active phaseof storage protein synthesis extends from approximately 11 to28 DAF. Pulse labeling during this period showed that at 16DAF the contribution of vicilin to total protein synthesis wassimilar in both control and S-deficient plants (Fig. 1). In cotyle-dons from control plants this level then declined sharply while,in marked contrast, it remained high in S-deficient seeds through-out the remainder ofthe protein accumulation phase. This phasewas extended considerably in S-deficient seeds in this experimentby approximately 8 d. Clearly, a suboptimal supply of nutrientS has a dual effect resulting in a decrease in legumin synthesis(5, 15) and the maintenance ofa high relative rate of total vicilinsynthesis right through the second half of seed development.

Sulfur Deficiency and the Developmental Pattern of MajorSeed Polypeptides. The contrasting effect of S deficiency on thetwo major storage protein fractions seen above prompted a more

40

0BCD

0

'4-

-

z~14-

0elz

301-

20 F

10 F

0

o. -S+s~~-+S

I I aIL I

16 20 24 28 32 36 40

DAYS AFTER FLOWERINGFIG. 1. Vicilin synthesis during development of S-deficient and non-

deficient seeds. Cotyledons were taken from S-deficient (-S) and controlplants (+S) at intervals during seed development and pulse labeled for 2h with a mixture offourteen '"C-amino acids. The contribution of vicilinto total protein synthesis was estimated by reating total extracts of thelabeled cotyledons with Sepharose 4B to which was covalently coupledIgG to a total vicilin fraction from mature seeds.

detailed examination of its effect on the developmental patternof all major seed polypeptides. At intervals during seed devel-opment, cotyledons from control and S-deficient plants werepulse-labeled for 2 h and the resultant radioactive proteins werefractionated by SDS-PAGE (Fig. 2). Each lane of the gel wasloaded with equal amounts of incorporated radioactivity. Inten-sity of radioactive bands, therefore, gives a relative measure oftheir synthesis at each harvest time during seed development,but is no indication of the absolute level of synthesis. Becausethe pulse-labeling period was only 2 h, the smaller subunits ofvicilin (Mr < 50,000) are not radioactive. These subunits ariseby delayed posttranslational processing of precursor moleculesfrom the Mr 50,000 family of polypeptides (7, 18). This process-ing takes place 6 to 20 h after the precursors are synthesized (6).Legumin subunits of Mr - 40,000 and 19,000 also arise fromprecursor molecules (Mr - 60,000) but in this case processingcommences about 1 h after precursor synthesis (19) with theresult that both precursors (Mr - 60,000) and processed subunits(Mr - 40,000 and 19,000) are detected after a 2-h labeling period.

In seeds from healthy plants, the Mr 50,000 family of vicilinpolypeptides was the dominant radioactive product early indevelopment and its contribution to total protein synthesis de-clined rapidly after 17 DAF (Fig. 2). Legumin polypeptides, (Mr60,000, 40,000, and 19,000) were mainly synthesized in thesecond half of seed development from 20 DAF onward. Sulfurdeficiency resulted in a number of changes in the polypeptidedevelopmental pattern. Consistent with the results in Figure 1,the high level of synthesis of vicilin Mr 50,000 polypeptides seenearly in development was maintained throughout the second halfof development, contrasting sharply with its pattern in healthyseeds. The other major component of the vicilin fraction (con-vicilin, Mr 75,000) makes a far smaller contribution at earlystages of development to total vicilin synthesis. At later stages ofdevelopment, however, Mr 75,000 is the major component ofthe vicilin fraction being synthesized. In healthy (control) seedsits developmental pattern covers almost the entire period of

|

652 CHANDLER ET AL

www.plantphysiol.orgon February 24, 2020 - Published by Downloaded from Copyright © 1984 American Society of Plant Biologists. All rights reserved.


SYNTHESIS AND mRNA LEVELS OF PEA SEED PROTEINS

G_ 4 m - -_ _-*

.:.. irZ W w l v_

t;----A _gI r!"

_ * . .'mi1mp 0-me ~

t'. ...lw q

___ - --g

-_-my

+ - +

I

10 13I . _.. . . ..

15 17 20 22 25 27

+ - _ _

30 L3 L330 33 36

FIG. 2. The patterns of polypeptide synthesis during seed development in S-deficient (-) and nondeficient (+) pea plants. Cotyledons were takenfrom seeds at intervals during development and pulse labeled for 2 h with a mixture of fourteen '4C-amino acids. Total extracts of the cotyledonswere then fractionated by SDS-PAGE and radioactive polypeptides were detected by fluorography. With two exceptions, all lanes were loaded withequal amounts of TCA-insoluble radioactivity, lanes containing 33- and 36-day samples had half the standard amount of radioactivity. Numbers on

vertical axis indicate the Mr X IO-' of relevant polypeptides.

protein accumulation from 13 to 27 DAF. This accounts for therelatively high levels of total vicilin being synthesized (12% oftotal protein synthesis) by control cotyledons 28 DAF (Fig. 1),at a stage when there is very little synthesis ofvicilin M, = 50,000polypeptides (Fig. 2). S deficiency resulted in a slight but con-sistent decrease in the relative level of convicilin synthesis. Theseresults indicate that, in S-deficient seeds, the enhanced andextended synthesis of total vicilin measured by immunochro-matography (Fig. 1) largely reflects the altered developmentalpattern of synthesis of the Mr 50,000 vicilin polypeptides.Other changes in the polypeptide spectrum include the previ-

ously-reported, severe depression oflegumin synthesis (5, 15). Atthe level of sensitivity in this experiment, legumin synthesis wasnot detected in S-deficient seeds during the second half of seeddevelopment (Fig. 2). S deficiency also reduced the contributionof several albumin polypeptides, namely, two polypeptides ofapproximate M, 92,000 and 100,000, a doublet of Mr - 22,000and a polypeptide ofMr 1 1,000. From labeling studies with [35S]sulfate, all these polypeptides are known to contain S-aminoacids (16). A radioactive polypeptide ofMr - 35,000, which wasnot detected in control seeds, was consistently present in seedsfrom S-deficient plants between 13 and 22 DAF. These resultsshow that S deficiency results in extensive qualitative and quan-titative changes to the developmental patterns of synthesis of awide range of seed proteins, the most striking of which concernits contrasting effects on synthesis of the vicilin Mr 50,000polypeptides compared with the polypeptides of legumin.mRNA Levels during Development of Healthy and Sulfur-

Deficient Seeds. It was previously shown that the greatly reducedlevel of legumin synthesis in S-deficient seeds was correlated withreduced levels of legumin mRNA at 22 and 25 DAF (5). In viewof the striking effect of S deficiency on the temporal pattern ofsynthesis of a number of seed proteins, a parallel study wasconducted on the relative level of a range of mRNAs to see

whether these also correlated with the observed changes in pro-tein biosynthesis. Total RNA was extracted from cotyledons atintervals throughout seed formation and equal amounts ofRNAwere fractionated by electrophoresis on agarose. The fractionatedRNA was blotted onto derivatized cellulose and hybridized withradioactive cDNAs representing seven different pea seed pro-teins. In healthy seeds each of these mRNAs displayed a char-acteristic developmental pattern (Fig. 3). It should be noted thatintensities of the bands reflect relative RNA levels within thetime series for any given probe. Differences in band intensitybetween different probes do not reflect the relative levels of theircorresponding RNAs but other factors such as specific activityof the labeled DNA sequences and length of exposure of theautoradiograph.The patterns in Figure 3 show that in healthy seeds some

mRNAs, such as those coding for the Mr 50,000 vicilin and Mr25,000 lectin polypeptides reach their maximum level early indevelopment and then decline rapidly, while others such as thosecoding for the albumin of Mr 22,000 and the polypeptide of Mr70,000 are predominant during the latter stages of development.The mRNAs for legumin (Mr 60,000) and the small albumin(Mr 13,000) reach their highest levels midway through develop-ment (20 to 22 DAF). Convicilin (Mr 75,000) mRNA is presentat relatively constant levels from 17 to 25 DAF.

In general there is a good correlation in healthy plants betweenthe temporal pattern of the mRNAs (Fig. 3) and that of thecorresponding polypeptides detected by pulse-labeling detachedcotyledons (Fig. 2), strongly suggesting a large measure ofcontrolof seed protein synthesis through regulation of mRNA levels.The same conclusion applies to the change in relative mRNAlevels in developing, S-deficient seeds, some examples of whichare shown in Figure 4. Once again equal amounts of total RNAwere electrophoresed in all lanes, representing both healthy anddeficient material. These data confirm that, relative to healthy

- 100- 75- 60- 50- 40- 35

- 22- 19

- 11

653

If,'




4_ __--MEWW I_W

- up MO: .'

MIM 4E1 -

-..-.......I

........ _

_*. amamsFIG. 4. The effect of S deficiency on mRNA levels for four major

seed proteins. The procedure was essentially as described for Figure 3,but using cotyledons from S-deficient as well as control plants. Labelingof figure as in Figure 3.

FIG. 3. Changes in relative level of mRNAs for seven major seedproteins during seed development on control (nondeficient) plants. TotalRNA was isolated from cotyledons at intervals during seed development,fractionated by electrophoresis on agarose, transferred to derivatizedcellulose filters, and hybridized with each one of eight 32P-labeled plas-mids containing DNA complementary either to the mRNA for one ofseven major seed proteins or to 18S ribosomal RNA. Equal amounts oftotal RNA were loaded on all tracks, and uniformity of transfer is shownby the probe for ribosomal RNA. The left hand and right hand letteringsindicate, respectively, the plasmid designation and the size of the poly-peptide (Mr x 10-3) corresponding to its cDNA insert (determined byhybrid release translation). pTA250. 10 is a plasmid with an insert com-plementary to 18S ribosomal RNA from wheat (R. Appels, personalcommunication).

plants the level of legumin mRNA remains low throughout seeddevelopment. There was also a marked effect of S deficiency onthe mRNA levels for the Mr 70,000 polypeptide and the Mr13,000 polypeptide. In contrast, the relative level of mRNA forthe Mr 50,000 vicilin subunits was increased by S deficiency andthis high level was maintained almost to the end of the proteinaccumulation phase.

Sulfur deficiency resulted in a reduction in mRNA levels forconvicilin (Mr 75,000), lectin (Mr 25,000), and the Mr 22,000albumin (data not shown). The extent of this reduction wasvariable with Mr 22,000 mRNA being unaffected in some cases.The reason for this variability is not known but it may reflectdifferences in the degree of sulfur stress from one experiment toanother. The same variability was found in the in vivo labelingpattern for these polypeptides. These results show that the effectof S deficiency is not restricted to legumin mRNA, but rather itresults in quantitative and qualitative changes to the develop-mental patterns of a range of seed mRNAs.Recovery Experiments. It was shown earlier that S-deficient

plants responded to increased S supply even when this was givenas late as 20 DAF. Increased levels of legumin mRNA andlegumin synthesis were detected after 2 d of adequate S supply(5). In view of the marked enhancement of Mr 50,000 vicilinmRNA levels and synthesis in S-deficient seeds, it was of interestto follow the changes in these parameters under recovery condi-

tions. S-deficient plants at 20 DAF were restored to an adequateS supply and the level of synthesis ofMr 50,000 vicilin and of itsmRNA were monitored during recovery. Pulse-labeling experi-ments showed that after 3 d the contribution of Mr 50,000synthesis to total protein synthesis in recovering seeds had de-clined slightly (Fig. 5A), while in the same period synthesis oflegumin Mr 60,000 increased sharply, as previously reported (5).After 6 d of recovery, M, 50,000 synthesis had dropped to a lowlevel and after 9 d there was no detectable synthesis of the majorM, 50,000 polypeptides at this level of sensitivity (Fig. 5A).Changes in mRNA levels paralleled both this slow decline insynthesis of M, 50,000 vicilin and the more rapid increase inlegumin synthesis seen by pulse-labeling (Fig. 5B). In otherexperiments (not shown), an increase in the relative level oflegumin mRNA was detected 12 h after restoring an adequatesulfur supply to the whole plant. In addition to increasedamounts of legumin mRNA following restoration of S supply,there were marked increases in amounts of the mRNAs for theMr 13,000 and Mr 70,000 polypeptides (data not shown). Thisindicates that the effects of S deficiency on mRNA levels allappear to be reversed by restoration of S supply.A similar rapid recovery of legumin synthesis and mRNA

levels was seen earlier when detached pods from S-deficientplants were supplied with adequate sulfate through the pedicel(5). Early attempts to obtain recovery in an even simpler tissuesystem, such as detached whole seeds or isolated cotyledons, wereinitially unsuccessful. It was then found that detached seeds fromS-deficient plants would respond to extra sulfate if this wassupplied at relatively high concentration (6 to 12 mM) throughthe funicle. The effect of restored sulfate supply on the level oflegumin and vicilin mRNAs is shown in Figure 6. Both mRNAsbehaved in a manner similar to that seen in the recovering wholeplant; legumin mRNA increased sharply in 24 h approachingthe level found in nondeficient seeds, and Mr 50,000 vicilinmRNA declined slightly in the same period. Recovery ofleguminmRNA levels in S-deficient seeds has also been found aftersupplying them with methionine, cysteine, glutathione, or mer-captoethanol rather than sulfate (data not shown). We have notyet been successful in obtaining a similar recovery pattern indetached cotyledons.

654 CHANDLER ET AL




pPS1 5-75

pPS1 5-84

__ _ _~~

fi~*~~- 50

40., 60

50

1 2 3

FIG. 6. Changes in levels of legumin (Mr 60,000) and vicilin (M,50,000) mRNAs following supply of sulfate to detached sulfur-deficientseeds. Lane 1, RNA from seeds from nondeficient, noncultured seeds;lanes 2 and 3, seeds from S-deficient plants which were cultured eitherwithout sulfate (lane 2) or with 12 mM MgSO4 (lane 3) for 24 h prior toextraction of total RNA and estimation ofmRNA levels as described forFigure 3. The lettering on the left hand and right hand sides indicates,respectively, the plasmid used and the size of polypeptide (M, x 10-3)corresponding to its cDNA insert.

-11

-S

0

IS -S

3

+s -S

6

+S -S

9

DAYS

b.

pPS1 5-75

pPS15-84 I _0

-S I+S -S+- -S + -So 3 6 9

DAYS

FIG. 5. Polypeptide synthesis and mRNA levels in developing peaseeds during recovery from S deficiency. S-deficient plants were suppliedwith optimum levels of sulfate at 20 DAF and cotyledons were harvestedeither immediately (0) or 3, 6, and 9 d later. The '4C-amino acid labelingpattern (2-h labeling) and mRNA levels for legumin (M, 60,000) andvicilin (M, 50,000) were compared with those of cotyledons taken fromequivalent untreated, S-deficient plants over the same time period. A, Invivo labeling patterns of polypeptides following SDS-PAGE and fluorog-raphy of cotyledon extracts; B, relative mRNA levels detected with 32P-

labeled plasmids containing DNA complementary to either legumin or

Mr 50,000 vicilin mRNA. Each lane was loaded with equal amounts oftotal RNA. Numbers at the right hand side show the approximate M, x

10-3 of relevant radioactive polypeptides.

DISCUSSION

This study compares the patterns of synthesis of the majorproteins of pea seeds and changes in their mRNA levels duringdevelopment of healthy and S-deficient plants. S deficiencyresulted in a number of major changes in addition to the previ-ously reported reduction in legumin synthesis and leguminmRNA levels (5). Both the relative level and the duration ofsynthesis of vicilin Mr 50,000 polypeptides were markedly in-creased (Figs. 1 and 2) and these changes were paralleled byincreases in the relative level of mRNA for these polypeptides

and the persistence of this high level through the second half ofseed development. This is in sharp contrast to the nondeficientseed in which Mr 50,000 vicilin synthesis and its mRNA arelargely confined to the first half of seed development. Anotherchange resulting from S deficiency was the greatly reduced levelof synthesis of the albumin polypeptide of M 11;000. Thischange was correlated with a marked reduction in the level of itsmRNA. The synthesis ofother albumin polypeptides, (Mr 22,000and -95,000) was also decreased by S deficiency, albeit lessdramatically. These changes in synthesis of the major albuminsare consistent with an earlier report on the effect of S deficiencyon the level of these albumin polypeptides accumulated in themature seed ( 16).

It should be stressed that data in this report are on a relativebasis and do not necessarily reflect changes in absolute levels ofprotein synthesis and mRNA. Changes in absolute level varywith the degree of S deficiency. For example, at a moderate levelof S stress, the total amount of vicilin per seed increased by 51%while legumin decreased to <5% ofthe amount in control seeds.Under more severe S deficiency conditions, absolute vicilin levelsalso declined but to a much smaller extent than legumin (datanot shown).Three related families of vicilin M, 50,000 mRNAs have been

identified from a study ofcDNA clones (P. M. Chandler, unpub-lished results). The inserts in the three respective clones (TableI) show sequence divergence of 15 to 25% and as a result theydo not cross-hybridize under the conditions used in this study.During development of normal cotyledons a pattern of mRNAaccumulation similar to that shown in Figure 3 was observed forall three families. In S-deficient cotyledons all three mRNAclasses reached higher relative levels and persisted much longerduring development (data not shown). It, therefore, appears thata common regulatory factor controls expression of these threefamilies and that S deficiency perturbs this control.

It was reported previously that a minor component ofthe totalvicilin fraction was reduced in S deficiency. This component,detected by immunoelectrophoresis using antiserum to totalvicilin and referred to as peak 3, consisted mainly of a S-containing Mr 50,000 polypeptide (14). The relationship of thisM, 50,000 polypeptide to the major M, 50,000 vicilin compo-nents, all of which are greatly increased in S deficiency and twoof which have been sequenced and found to lack cysteine andmethionine (1 1, 18), is not clear. It is possible that the Mr 50,000polypeptide ofpeak 3 corresponds to the minor, S-containing Mr50,000 polypeptide derived from a Mr 70,000 precursor whichwas detected in earlier pulse-chase labeling experiments using

a.

655




I35Sjmethionine (6).Mechanism of Regulation. The mechanism by which the S

status of the seed can exercise such varied controls on mRNAlevels and protein synthesis is not understood. The ability ofisolated S-deficient seeds to respond to restored S supply in thesame way as seeds on the intact plant (Fig. 6) indicates that nofactors from the remainder of the plant are needed to mediatethe response to changing S status. The fact that, in addition tosulfate, compounds such as cysteine, methionine, glutathione,and mercaptoethanol all elicited increased legumin mRNA levelsin detached S-deficient seeds makes it likely that a reduced Scompound is involved in the response.The various effects of S deficiency and of restored S supply on

the synthesis of individual seed proteins was closely correlatedwith changes in their mRNA levels. The same is true for devel-opmentally regulated patterns of protein synthesis during seeddevelopment in nondeficient plants. The simplest conclusionfrom this close relationship is that altered rates of protein syn-thesis result from modulation ofmRNA levels. This modulationcould occur at the transcriptional level, during posttranscrip-tional processing, or at the level of mRNA stability. Control ofmRNA abundance solely at the level of transcription wouldimply the existence of one or more agents which are sensitive toS supply and which mediate altered rates and duration of tran-scription of a number of major seed protein genes. On the otherhand, control during posttranscriptional processing, or ofmRNAstability, may indicate that transcription of these genes continuesat the same rate in both healthy and deficient seeds, and that inthe nucleus or in the cytoplasm there is selective degradation ofsome mRNAs in response to sulfur stress. In both these modelsit is mRNA levels which are modulated and these levels in turncontrol protein synthesis. We cannot exclude a third possibility,namely, that other factors, for example the availability ofS aminoacids (see below), may exert direct translational control on thesynthesis of particular polypeptides in S-deficient cotyledons.The observed changes in mRNA levels could then result fromthe preferential degradation of those mRNAs which are nottranslated at normal rates. In this model the changed mRNAlevels would be a consequence rather than a cause of alteredpatterns of protein synthesis.S deficiency in peas results in reduced synthesis of a number

of proteins which contain S amino acids (legumin and the Mr11,000, 22,000, and 95,000 albumins) and increased synthesis of50,000 vicilin polypeptides which lack both cysteine and methi-onine (1 1, 18). A similar preferential reduction in the amount ofthe more S-rich seed proteins has been reported in a number ofother species (2, 15, 17, 21). This immediately suggests thepossibility of a translational control mechanism exercisedthrough the availability of methionine and cysteine for polypep-tide synthesis. A shortage of these amino acids could lead todiscrimination against the synthesis of the more S-rich proteins.There are, however, several observations which argue against thisas the sole explanation for the changing pattern of proteinsynthesis in response to S deficiency.

First, Macnicol (12) has measured the level of all amino acidsin the aminoacyl-tRNA pool of S-deficient and control seeds andfound that, although total tRNA levels are reduced by S defi-ciency, there was no difference in the relative amounts ofcysteineand methionine in the aminoacyl-tRNA pools in the two situa-tions. Since these amino acids are the immediate substrates forpolypeptide synthesis, this argues against translational control atthis level. Second, the selection is not consistently against trans-lation of mRNAs coding for S-rich proteins. Jakubek and Przy-bylska (10) have determined the amino acid composition of theMr 22,000 albumin of pea seeds and found that it contains 0.9%cysteine and 1.6% methionine. This protein is richer in S amino

acids than is legumin (1.0% cysteine and 0.7% methionine, [3]),

yet its accumulation and mRNA level are affected much less byS deficiency. Third, it is apparent from DNA sequencing studiesthat the mature vicilin Mr 50,000 polypeptides contain few, ifany S amino acids. Nevertheless, cysteine and methionine arerequired for synthesis of these proteins as their signal peptidesare relatively rich in S amino acids: for example, the 27 aminoacids of the signal peptide deduced from the sequence of pPSl 5-84 include four methionines and one cysteine residue. A finalrelevant observation concerns the pattern of protein synthesis incultured, detached cotyledons. When detached cotyledons fromnondeficient plants are cultured for 24 h they synthesize arelatively large amount of a polypeptide which appears to beunique to cultured cotyledons (A. Millerd, unpublished obser-vations). This protein, which is known from labeling data tocontain S amino acids, was synthesized in similar quantity incotyledons from S-deficient plants which were cultured in amedium lacking S (P. M. Chandler, unpublished observations).This result indicates that S-deficient seeds are metabolicallycapable of responding to a particular stimulus by synthesis of anew, S-containing protein (and presumably its mRNA). Takentogether, these observations argue against a gross level of trans-lational control which discriminates against all S-containingproteins.

It is of interest to note that the supply of other nutrientelements also has a marked effect on the composition of the seedprotein fraction and these effects are different from that of Sdeficiency. Thus, a deficiency of either K or P in peas results inseeds with a greatly enhanced proportion of legumin in themature seed ( 15). Current experiments are examining the changesin protein synthesis and mRNA levels in these seeds duringdevelopment. These widely varying responses ofdeveloping seedsto different nutrient stresses, including the apparently compen-satory increase in synthesis of specific proteins, points to a highdegree of phenotypic plasticity in the developing seed system.This is further indicated by the ability of S-deficient seeds torespond rapidly to restored sulfur levels more than halfwaythrough seed development. A better understanding of the molec-ular events involved in these situations will provide more insightinto some of the regulatory mechanisms which determine thecomposition of the seed protein fraction.

Acknowledgments-We thank Zufi Ariffin, Edward Newbigin, and Linda Rob-erts for their skilled technical assistance and Denese McCann for typing themanuscript.

LITERATURE CiTiD

1. BADENOCH-JONES J, D SPENCER, TJV HIGGINS, A MILLERD 1981 The role ofglycosylation in storage-protein synthesis in developing pea seeds. Planta153: 201-209

2. BLAGROVE RJ, JM GILLESPIE, PJ RANDALL 1976 Effect of sulphur supply onthe seed globulin composition of Lupinus angustifolius. Aust I Plant Physiol3:173-184

3. CASEY R, MN SHORT 1981 Variation in amino acid composition of leguminfrom Pisum. Phytochemistry 20: 21-23

4. CHANDLER PM 1982 The use of single-stranded phage DNAs in hybrid arrestand release tnslation. Anal Biochem 127: 9-16

5. CHANDLER PM, TJV HIGGINS, PJ RANDALL, D SPENCER 1983 Regulation oflegumin levels in developing pea seeds under conditions of sulfur deficiency.Rates of legumin synthesis and levels of legumin mRNA. Plant Physiol 71:47-54

6. CHRISPEELS MJ, TJV HIGGINs, D SPENCER 1982 Assembly of storage proteinoligomers in the endoplasmic reticulum and processing of the polypeptidesin the protein bodies of developing pea cotyledons. J Ceil Biol 93: 306-313

7. GATEHOUSE JA, CW LYcET, RRD CROY, D BOULTER 1982 The post-trnsla-tional proteolysis of the subunits of vicilin from pea (Pisum sativum L.).Biochem J 207: 629-632

8. GRUNSTEIN M, DS HOGNESS 1975 Colony hybridization: a method for theisolation of cloned DNAs that contain a specific gene. Proc Natl Acad SciUSA 72: 3961-3965

9. HIGGINS TJV, PM CHANDLER, G ZURAWsKI, SC BurroN, D SPENCER 1983The biosynthesis and primary structure of pea seed lectin. J Biol Chem 258:9544-9549

10. JAKUBEK M, J PRZYBYLSKA 1979 Comparative study of seed proteins in the

CHANDLER ET AL656




genus Pisum III. Electrophoretic patterns and amino acid composition ofalbumin fractions separated by gel filtration. Genet Pol 20: 369-380

11. LYcErr GW, AJ DELAUNEY, JA GATEHOUSE, J GILROY, RRD CROY, DBOULTER 1983 The vicilin gene family ofpea (Pisum salivum L.): a completecDNA coding sequence for preprovicilin. Nucleic Acids Res 11: 2367-2380

12. MACNICOL PK 1983 Differential effect of sulfur deficiency on the compositionof aminoacyl-tRNA and free amino acid pools of the developing pea. FEBSLett 156: 55-57

13. MILLERD A, D SPENCER 1974 Changes in RNA-synthesising activity andtemplate activity in nuclei from cotyledons of developing pea seeds. Aust JPlant Physiol 1: 331-341

14. MILLERD A, JA THOMSON, PJ RANDALL 1979 Heterogeneity of sulfur contentin the storage proteins of peas. Planta 146: 463-466

15. RANDALL PJ, JA THOMSON, HE SCHROEDER 1979 Cotyledonary storage pro-teins in Pisum salivum IV. Effects of sulfur, phosphorus, potassium andmagnesium deficiencies. Aust J Plant Physiol 6: 11-24

16. SCHROEDER HE 1984 The major albumins of Pisum cotyledons. J Sci FoodAgric 35: 191-198

17. SHEWRY PR, J FRANKLIN, S PARMAR, SJ SMITH, BJ MIFLIN 1983 The effectsof sulfur starvation on the amino acid and protein compositions of barleygrain. J Cereal Sci 1: 21-31

18. SPENCER D, PM CHANDLER, TJV HIGGINS, AS INGLIS, M RUBIRA 1983Sequence relationships ofthe subunits of vicilin from pea seeds. Plant MolecBiol 2: 259-267

19. SPENCER D, TJV HIGGINS 1980 The biosynthesis of legumin in maturing peaseeds. Biochem Int 1: 502-509

20. SPENCER D, TJV HIGGINS, SC BUTrON, RA DAVEY 1980 Pulse labeling studieson protein synthesis in developing pea seeds and evidence of a precursorform of legumin small subunit. Plant Physiol 66: 510-515

21. WRIGLEY CW, DL DUCROS, MJ ARCHER, PG DOWNIE, CM ROXBURGH 1980The sulfur content of wheat endosperm proteins and its relevance to grainquality. Aust J Plant Physiol 7: 755-766

657



Documents

Influence Sulfur Nutrition on Developmental Patterns of ...seeds showed that a high relative level of total vicilin (vicilin plus convicilin) synthesis was maintained throughout the