cytoplasmic distribution ofheatshock proteins in soybean' · cytoplasmic distribution...

Plant Physiol. (1988) 86, 1240-12460032-0889/88/86/1240/07/$01.00/0

Cytoplasmic Distribution of Heat Shock Proteins in Soybean'Received for publication July 9, 1987 and in revised form December 30, 1987

MICHAEL A. MANSFIELD*2 AND JOE L. KEYDepartment of Botany, University of Georgia, Athens, Georgia 30602

ABSTRACT

Previous analyses of the distribution of heat shock (hs) proteins insoybean (Glycine max L. Merr., var Wayne) have demonstrated that afraction of the low molecular weight hs protein associates with ribosomesduring hs. To more specificafy characterize the nature of this association,isokinetic centrifuption of ribosomes through sucrose gradients was usedto separate monosomes from polysomes. The present analysis demon-strated that hs proteins were bound to polysomes but not monosomes.Treatment of polysomes with puromycin, K+, and Mg2+, which causeddissociation of ribosomes into 40S and 60S subunits, also caused disso-ciation of the hs proteins. Using the procedure of Nover et al. (1983, Mol.Ceil Biol, 3: 1628-1655), a hs granule fraction was also isolated. As intomato cefls, hs granules from soybean seedlings contained the low mo-lecular weight hs proteins as a primary component and a number of othernon-hs proteins of relative molecular mass 30 to 40 kilodaltons and 70 to90 kilodaltons. On metrizamide gradients they exhibited a buoyant densityof 1.20 to 1.21 grams per cubic centhneter, typical of ribonucleoproteinparticles. Heat shock granules were characterized as unique cytoplasmicparticles based on protein composition and buoyant density. Isopycniccentrifugation of ribosome preparations demonstrated that they containedhs granules, but the hs proteins bound to polysomes were not released byKCI/EDTA treatment. Thus, the polysome-bound hs proteins and thegranule-bound hs proteins appear to represent two distinct populationsof hs proteins in the cytoplasm. Heat shock granules were not disinguish-able from ribosomes at the level of resolution used in transmission electronmicroscopy.

In tomato and a number of other plant species, hs proteinswere found in the form of cytoplasmic aggregates termed hsgranules (18, 20). These granules were distinct from ribosomesbased on several characteristics. First, they were found in twosize classes: 30 to 40 nm and 70 to 80 nm. Second, they containedapproximately half of the low mol wt hs proteins present in thecytoplasm. Third, they contained a number of minor proteinsapparently unrelated to ribosomal proteins. Formation of thesegranules was temperature-dependent, implying that they have aprotective function. If hs protein synthesis was preinduced, hsgranules formed when tomato cells were given a supraoptimalhs; but disaggregation during recovery was much slower than incells receiving a normal hs.Comparison of the isolation procedures for ribosomes (15) and

hs granules (20) suggested that there is a possibility for the con-tamination of ribosome preparations with hs granules. The majordifference between the two procedures is that, prior to the iso-lation of hs granules, the ribosomes are dissociated by KCI/EDTAtreatment. It was reported that hs proteins associated minimallywith intact ribosomes and not at all with ribosomal subunits (20).Because the association of hs proteins with ribosomes of soybean(15) has not been rigorously characterized, it is possible that hsproteins are bound in a granule-like structure. To examine thisquestion, the association of hs proteins with ribosomes has beenanalyzed by isokinetic and isopycnic centrifugation. A hs granulefraction has also been isolated and characterized. The resultsindicate that hs proteins are bound to polysomes rather thanmonosomes and that the polysome-bound and granule-bound hsproteins represent two distinct populations in the cytoplasm.

One approach to determining the function of hs3 proteins hasbeen to compare their localization patterns in cells during hs andrecovery. Specific associations of hs proteins have been reportedfor nuclei (15, 22, 23), nucleoli (22, 23), chloroplasts (10, 24),mitochondria (15, 22), the plasma membrane (14, 22), and ele-ments of the cytoskeleton (12, 22). In tomato (20) and soybean(15), a fraction of the hs proteins was also found bound to ri-bosomes in a temperature-dependent interaction. Most of theseproteins were represented by the low mol wt hs proteins (15-18kD in soybean and 17 kD in tomato), but a smaller contributionfrom the high mol wt hs proteins was evident. The functionalsignificance of this association is uncertain, but it may be pro-tective. In soybean, thermotolerant seedlings contained a greateramount of ribosome-bound hs proteins at the supraoptimal hstemperature of 45°C than at the normal hs temperature of 40°C(15).

1 Supported through Department of Energy Contract No. DE-AS09-80ER10678 to Joe L. Key.

2 Current address: MSU-DOE Plant Research Laboratory, MichiganState University, East Lansing, MI 48824-1312.

3Abbreviations: hs, heat shock; PMSF, phenylmethylsulfonyl fluo-ride; IEF, isoelectric focusing; RNP, ribonucleoprotein; scRNP, smallcytoplasmic ribonucleoprotein.

MATERIALS AND METHODS

Plant Material. Soybean seeds (Glycine max L. Merr. varWayne) were surface-sterilized in 10% Clorox, washed thor-oughly with cold tap water, and allowed to germinate in the darkfor 40 to 48 h (15).

Seedling Incubation. Prior to incubation, seed coats and co-tyledons were removed from the seedlings. The seedlings wereincubated with gentle agitation in the dark for 3 h at 40°C inbuffer containing 1 mM potassium phosphate (pH 6.0), 1% (w/v) sucrose, and 50 ,ug/ml chloramphenicol. L-[4,5-3H]Leucine(New England Nuclear) was included in the incubation buffer tolabel the hs proteins.

Isolation of Ribosomes. Ribosomes were isolated as previouslydescribed (15) with modifications. Briefly, at the end of theincubation period, seedlings were rinsed with distilled water andcoarsely chopped with a razor blade. The tissue was then ho-mogenized in buffer A (200 mM Tris-Cl [pH 8.8], 30 mM MgCl2100 mm KCI, 0.5 M sucrose, 1 mM DTT, 1 mM PMSF, and 0.1%[v/v] diethylpyrocarbonate). After centrifugation at 9,000g for10 min, the supernatant was filtered through Miracloth, and 0.05volumes of 20% (v/v) Triton X-100 was added to the filtrate.The postmitochondrial supernatant was obtained by centrifu-gation at 27,000g for 10 min and layered onto 7-ml pads of 1.7M sucrose in buffer A. Ribosomes were pelleted by centrifugation

240 www.plantphysiol.orgon November 11, 2017 - Published by Downloaded from

Copyright © 1988 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org

CYTOPLASMIC DISTRIBUTION OF HEAT SHOCK PROTEINS

in a Spinco 60 Ti rotor for 135 min at 55,000 rpm. Ribosomepellets were resuspended in the appropriate buffer and pooledfor further analysis. Prior to gradient analysis, nonresuspendablematerial was removed from the samples by centrifugation at9,000g for 10 min.

Preparation of Ribosomal Subunits. Ribosomal subunits wereprepared by treatment with puromycin and 500 mm KCl as de-scribed by Gantt and Key (6) except that the sucrose concen-tration in the dissociation buffer was changed from 250 mm to10% (w/v).

Isolation of Heat Shock Granules. Heat shock granules wereisolated according to Nover et al. (20) with minor modifications.After being coarsely chopped, seedlings were homogenized inbuffer B (50 mM Tris-Cl [pH 7.8], 10 mM MgCl2, 25 mM KCI,2 mm CaC12, 0.1% [v/v] Nonidet P-40, 150 mm sucrose, 20% [w/v] glycerol, 1 mm PMSF, 0.1% [v/v] diethylpyrocarbonate) witha Brinkman Polytron equipped with a PT20 probe. The homog-enate was filtered through Miracloth and 28 fim nylon mesh andthen centrifuged at 9,000g for 10 min. The supematant was son-icated for 10 s, adjusted to 250 mM KCI and 30 mM EDTA, andthen centrifuged at 27,000g for 10 min to remove starch grains.The resulting supernatant was then layered onto two 6-ml padsin modified buffer B. For the lower pad, CaCl2 was eliminatedand the sucrose concentration was increased to 350 mm. For theupper pad, MgCl2 was eliminated and concentrations of KCI,Nonidet P-40, and sucrose were adjusted to 250 mM, 1% (v/v)and 250 mm, respectively. The granules were pelleted by cen-trifugation in a Spinco 60 Ti rotor at 30,500 rpm for 60 min.Pellets were resuspended in the appropriate buffer and pooled.Prior to gradient analysis, nonresuspendable material was pel-leted by centrifugation at 9,000g for 10 min.

Sucrose Gradient Centrifugation. For ribosome preparations,10 to 35% (w/v) sucrose gradients were prepared in buffer C(200 mM Tris-Cl [pH 8.8], 30 mM MgCl2, 100 mM KCl, 1 mMDTT). For ribosomal subunits, 15 to 40% (w/v) gradients wereprepared in the same buffer. For hs granules, 10 to 35% (w/v)gradients were prepared in buffer D (50 mM Tris-Cl [pH 7.8],10 mM MgCl2, 25 mm KCl, 1 mm DTT). Ribosome pellets andhs granule pellets were resuspended in the appropriate buffersupplemented with 5% (w/v) sucrose. The ribosomal subunitpreparation was used directly. For each sample, the absorbanceat 260 nm was determined, and equal amounts of material (A260units) were loaded onto duplicate gradients. After centrifugationin a Spinco SW 27 rotor, 1.2-ml fractions were collected in anISCO model 640 fraction collector equipped with a type 6 UVmonitor. To determine the distribution of radioactivity, fractionsfrom duplicate gradients were pooled. Protein and RNA wereprecipitated by adding trichloroacetic acid to a final concentra-tion of 5% (w/v). The protein was pelleted by centrifugation,washed once with 70% (v/v) ethanol, washed once with acetone,and then dried under vacuum. The final pellets were resuspendedin 250 ,ul of Laemmli sample buffer (11). Radioactivity in 10-,ulaliquots was assayed as previously described (16).Metrizamide Gradient Centrifugation. Metrizamide (Sigma)

solutions were prepared in buffer C for ribosomes or buffer Dfor hs granules. Gradients were prepared as described in Ballin-ger and Pardue (3). Equivalent amounts of ribosomes or hs gran-ules were loaded onto duplicate gradients and centrifuged at35,000 rpm for 67 h in a Spinco SW 50.1 rotor. The gradientswere then fractionated, and 0.5-ml fractions from duplicate gra-dients were pooled. Fraction densities were calculated from therefractive indices. Protein was isolated and radioactivity was de-termined as described for sucrose gradients analysis.

Electrophoresis. One-dimensional SDS-PAGE was performedaccording to Laemmli (11) using 12.5% (w/v) acrylamide gels.Proteins were resolved in two-dimensions (21) by IEF in the firstdimension and SDS-PAGE in 12.5% (w/v) acrylamide gels inthe second dimension. Proteins were loaded on the acidic ends

of IEF gels. The effective pH range of the IEF gels was pH 4to 7. Stainable proteins were visualized with Coomassie brilliantblue R-250. Radioactive proteins were visualized by fluorogra-phy (15).

Electron Microscopy. At the end of the incubation period,apical 1 mm sections were excised from the primary roots andfixed for 1 h at 4°C in 2% glutaraldehyde, 50 mm potassiumphosphate (pH 6.8), 1 mM CaCl2. The root tips were washed fora total of 30 min in three changes of 50 mm potassium phosphate(pH 6.8), 0.33 M CaCl2, and then fixed for 1 h in 2% OS04, 50mm potassium phosphate (pH 6.8), 0.33 M CaCl2. After washingfor 2.5 h in two changes of 50 mm potassium phosphate (pH6.8), 1 mM CaC12, and 30 min in three changes of distilled water,the tissue was stained for 1 h in 0.5% uranyl acetate. The tissuewas then dehydrated in an ethanol series and embedded in Spurr'slow viscosity resin. These sections were stained with lead acetateand examined by electron microscopy.

RESULTS

Distribution of Heat Shock Proteins in Ribosome Preparations.The ribosome fraction prepared from soybean seedlings con-tained both monosomes and polysomes. Although association ofhs protein with ribosomes was reported previously (15, 20), thedistribution of hs proteins between monosomes and polysomeswas not determined. To separate these two components, ribo-some preparations were subjected to isokinetic centrifugation onsucrose gradients. By including [3H]leucine in the incubationmedium during hs, the hs proteins were preferentially labeledwhile most of the normal, cellular proteins remained unlabeled(1, 9). When gradients loaded with ribosomes were centrifugedfor 19.6 h at 72,000 rpm, the monosomes sedimented as a prom-inent peak in the lower third of the gradient (Fig. 1). The poly-somes were contained in the pellet. Radioactivity was restrictedto the top of the gradient and the pellet.To resolve the polysomes, a second set of gradients was cen-

trifuged for 2 h at 108,000g. The polysomes sedimented as abroad peak in the lower two-thirds of the gradient (Fig. 2).Consistent with the distribution of radioactivity in Figure 1, apeak of radioactivity was observed at the top of the gradient andno radioactivity above background levels was associated with themonosome peak. There was, however, a slight increase in the

1.1

1.0

0.9

d 0.8

O 0.7

0105

.0 0.4

C.

0* 0.3

0.1

1 5 10 15 20

Fraction number25 30

3.5-

-3.0 -_

2.5-I

C.2.0 0

1.5

1.0

0.5

FIG. 1. Isokinetic centrifugation of monosomes. Ribosomes were iso-lated from 40 g of heat-shocked soybean seedlings incubated in the pres-ence of 2 mCi of [3H]leucine. Sucrose gradients were loaded with 40 A2W,units (1.83 x 106 cpm) of ribosomes and centrifuged for 19.6 h at 72,000g.Gradients were fractionated and analyzed as described in "Materials andMethods." The top of the gradient is oriented to the left. Approximately20% of the loaded cpm was recovered.

1241

www.plantphysiol.orgon November 11, 2017 - Published by Downloaded from Copyright © 1988 American Society of Plant Biologists. All rights reserved.


MANSFIELDANDKEY~~~~PlantPhysiol. Vol. 86, 1988

1 5 10 15 20

Fraction Number

25 30

4.5

4.0

3.0

Sm:~.:m.I.

FIG. 2. Isokinetic centrifugation of polysomes. Ribosomes were iso-

lated from 40 g of heat-shocked soybean seedlings incubated in the pres-

ence of 2 mCi of [3HJleucine. Sucrose gradients were loaded with 45 A,,.,

units (6.89 x 105 cpm) of ribosomes and centrifuged for 2 h at 108,000g.

Gradients were fractionated and analyzed as described in "Materials and

Methods." The top of the gradient is oriented to the left. Approximately

30% of the loaded cpm was recovered.

2.0 6.5

~~~~~~~~~~~~~~~~~~~6.05.5

16-5.0

1.4- 4.5

C" 1.2- 4.0?

T11.0 ~~~~~~~~~~~~~~3.0

0.8 2.5

1.)

0.642.

0.42.10

5 10 15 20 25 30

Fraction number

FIG. 3. Isokinetic centrifugation of ribosomal subunits. Ribosomes

were isolated from 40 g of heat-shocked soybean seedlings incubated in

the presence of 3 MCi of I3Hlleucine. After incubating the ribosomes in

puromycin and 500 mm~KCI, sucrose gradients were loaded with 45 A260

units (1.71 x 106 cpm) of ribosomal subunits and centrifuged for 15.5

h at 72,000g. Gradients were fractionated and analyzed as described in

"Materials and Methods." The top of the gradient is oriented to the left.

amount of radioactivity associated with polysomes.The next question examined was whether hs proteins could

bind independently to ribosomal subunits. Prior to isokinetic

centrifugation, the ribosomes were dissociated into 40S and 60S

subunits by treatment with puromycin and 500 mm~KCI (6).

When resolved on sucrose gradients, no radioactivity above back-

ground levels was found associated with either subunit peak (Fig.

3). Greater than 95% of the radioactivity was found at the top

of the gradient and in the pellet. The pellet contained approxi-

mately 85% of the total radioactivity and a small amount of

ribosomal protein. The disproportionate size of the 60S peak can

be accounted for by the contribution from 40S dimers (6).

In each of the gradient analyses, positions of ribosomal pro-

teins and hs proteins were confirmed by one-dimensional SDS-

PAGE. The protein profiles from selected fractions are shown

in Figure 4. Whiile fractions at the tops of the gradients contained

no ribosomes (Fig. 4, lane a), they did contain hs proteins (Fig.

4, lane b). Similarly, fractions containing polysomes also con-

tained hs proteins (Fig. 4, lanes e and f). No hs proteins were

af

a b cd e f

FIG. 4. Protein content of selected ribosome fractions. Proteins in

equal volume aliquots from fractions prepared for Figures 1, 2, and 3

were resolved by one-dimensional SDS-PAGE. Selected fractions are

presented here: lanes a and b, nonribosomal material (fraction 3, Fig.

1); lanes c and d, monosomes (fraction 22, Fig. 1); lanes e and f, poly-

somes (fraction 17, Fig. 2): lanes g and h, 40S subunits (fraction 12, Fig.

3); lanes i and j, 60S subunits (fraction 17, Fig. 3). Proteins were visu-

alized by staining with Coomassie brilliant blue (lanes a, c, e, g, and i)

or fluorography (lanes b, d, f, h, and j).

detected in association with monosomes (Fig. 4, lanes c and d),

40S subunits (Fig. 4, lanes g and h), or 60S subunits (Fig. 4,

lanes i and j).

Preparation and Analysis of Heat Shock Granules. A hs granule

fraction was prepared from soybean seedlings, and the proteinswere resolved by two-dimensional electrophoresis (Fig. 5). The

hs granule fraction contained many of the low mol wt hs proteinsas determined by both staining (Fig. 5A) and fluorography (Fig.

5B). A few other proteins were also evident in this preparation,most notably in the 30 to 40 kD region and the 70 to 90 kD

region. Although hsp70 was associated with hs granules in tomato

(20), none of the high mol wt proteins observed in hs granules

from soybean corresponded to hs proteins. Prolonged exposure

of fluorograms, while resulting in overexposure of the low mol

wt hs proteins, gave no indication that any other radiolabeled

proteins were present in this preparation.Isokinetic Centrifugation of Heat Shock Granules. A hs granule

preparation was resolved by isokinetic centrifugation on sucrose

gradients to compare these particles with polysomes and mono-

somes. Because hs granules were characterized as large particlesin tomato cells (20), the gradients were centrifuged for a much

shorter period of time to keep the granules from pelleting. When

the gradients were analyzed, both monosomes (fractions 1-3)and polysomes (fractions 4-5) were present at the top of the

gradient (Fig. 6). Sedimenting in advance of the polysome peakwas a broad shoulder of absorptive material presumably repre-

senting the hs granules (fractions 6-14). This material contained

the majority of the radioactivity. The locations of monosomes,

polysomes, and radiolabeled hs proteins were confirmed by SDS-

12421.6

1.4

0 1.2

10

0

0.2

olIIL

FF

b.0

Io

MANSFIELD AND KEY

E




t * w;

+IEF -

B.

a

., s,% tm.ew

FIG. 5. Two-dimensional electrophoresis of heat shock granule pro-teins. Total protein was extracted from hs granules prepared from 40 gof soybean seedlings incubated at 40°C in the presence of 3 mCi of[3H]leucine. Proteins (5 x 105 cpm) were resolved by two-dimensionalelectrophoresis and visualized by staining with Coomassie brilliant blue(A) or fluorography (B). The acidic end of the gel is oriented to the left.

1243

PAGE (data not shown).Isopycnic Centrifugation of Heat Shock Granules. The data

presented in Figure 6 showed that the hs granule isolation pro-cedure did not completely exclude monosomes and polysomes.Thus, it cannot be stated unequivocally that the material pelletingin advance of the polysomes was not due to an artifactual ag-gregation between hs proteins, monosomes, and polysomes. Todemonstrate conclusively that polysome-bound hs proteins andgranule-bound hs proteins were distinct, isopycnic centrifugationwas used. Metrizamide was selected as the centrifugation me-dium because it is nonionic. Thus, the problems inherent inprotein/RNA cross-linking, as must be performed before cen-trifugation on cesium chloride gradients, were avoided (4). Con-ditions were selected that generated a sigmoidal density distri-bution so as to give maximum resolution ofRNP particles. Whenthe distribution of radioactivity was analyzed, about 40% of thelabeled protein was found in association with hs granule fractions3 and 4 (Fig. 7). The same percentage was associated with mono-some/polysome fractions 8 and 9. Monosomes and polysomeshave reported densities of 1.30 and 1.35 g/cm3, respectively, inmetrizamide gradients (4) and were not separable from one an-other under the conditions used (Fig. 8, lanes 8 and 9). Heatshock granules banded at a density of 1.20 to 1.21 g/cm3 (Fig.8, lanes 3 and 4) which is typical for RNP particles (4). Theseresults show that a structure analogous to hs granules is presentin soybean cells.

Isopycnic Centrifugation of Ribosomes. It was then necessaryto determine if the hs proteins found associated with polysomes(Fig. 2) were distinct from those found in hs granules. The majordifference between the two isolation procedures is that, in thehs granule protocol, ribosomes are dissociated with KCl andEDTA prior to ultracentrifugation. If polysomes are bound tohs granules, treatment of ribosomes with KCl and EDTA priorto ultracentrifugation should release the granules (20). One ri-bosome preparation was resuspended in buffer C, and a secondpreparation was resuspended in buffer C modified to 250 mmKCl and 30mm EDTA. The two preparations were then resolvedon metrizamide gradients. While KC1/EDTA treatment did causean increase in the percentage of radioactivity in granule fractions3 and 4 from 18 to 24% (Fig. 9), there was no dramatic loss ofradioactivity from the monosome/polysome fractions. There wasa shift of radioactivity from fraction 9 to fraction 8, but the total

13.0

12.0

11.0

10.0

0.11. 1.0

i0~~~~~~~~~~~~~~~~~~~~~~~~C.ai 0.9 *4

ed-~ ~ a7.0

6.0

0.7~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.0.5/.

0 %- 4.0

0.4-~~ ~ ~ ~ ~ ~~~',.3.00.3-.

0.1 1.0

0

1 5 10 15 20 25 30

Fraction Number

FIG. 6. Isokinetic centrifugation of heat shock granules. Heat shockgranules were isolated from 40 g of heat-shocked soybean seedlings in-cubated in the presence of 3 mCi of [3H]leucine. Sucrose gradients wereloaded with 30A260 units (1.13 x 106 cpm) of hs granules and centrifugedfor 50 min at 72,000g. Gradients were fractionated and analyzed as

described in "Materials and Methods." The top of the gradient is orientedto the left. Approximately 50% of the loaded cpm was recovered.

35 1-c0

:p 30U

zL 25

0

20Eo 15

0 10

0 50

0 _, II

1 2 3 4 5 6 7 8 9

Fraction Number

1.330

1.275 -

E

1.220 "2._fc

1.165 a

1.110

FIG. 7. Distribution of radiolabel in heat shock granules. Heat shockgranules were isolated from 10 g of soybean seedlings incubated in thepresence of 2 mCi of [3H]leucine. Metrizamide gradients were loadedwith 3.5 A2w. units (8.9 x 105 cpm) of hs granules. After centrifugationat 150,000g for 67 h, gradients were fractionated. Each fraction was

analyzed for [3H]leucine content, and the percentage of radiolabel ineach fraction was calculated. Fractions 3 and 4 contain hs granules;fractions 8 and 9 contain monosomes and polysomes. The solid lineindicates the gradient profile.

A.

a0CI

i.



Plant Physiol. Vol. 86, 1988

percentage of radioactivity in these two fractions did not changesignificantly. The distribution of radioactivity in the remainderof the gradient was not substantially altered. When analyzed bySDS-PAGE and fluorography (Fig. 10), the distribution of hsproteins was continuous throughout the gradient. Because therewas no major increase in the hs granule fractions after KCI/EDTA treatment, the hs proteins found associated with poly-somes are distinct from hs granules.

Examination of Cytoplasmic Ultrastructure. Based on resultsof Nover et al. (20), who reported that hs granules were con-

1 2 3 4

A.~~~~ ~ ~~~~_-_

1.112 1.172 1.206 1.214 1.217 1.222 1.231 1.263 1.331:rn3

FIG. 8. One-dimensional electrophoresis of heat shock granules afterresolution on metrizamide gradients. The proteins in equal volume ali-quots of each fraction isolated in Figure 7 were resolved by one-dimen-sional SDS-PAGE and visualized by staining with Coomassie brilliantblue. Fraction numbers are indicated above each lane with fraction 1representing the top of the gradient. The number below each lane is theaverage density of that fraction. The bracket on the left indicates theposition of the 15 to 18 kD hs proteins. B.

I _ ~

1 2 3 4 5 6 7 8 9Fraction Number

FIG. 9. Distribution of radiolabel in ribosomes treated with KCI andEDTA. Ribosomes were isolated from two 10-g lots of heat-shocked

soybean seedlings incubated in the presence of 2 mCi of [3H]leucine;12.5 A260 units (3.76 x 105) of ribosomes (0) or 10 A2,0 units (6.9 x

105 cpm) of ribosomes resuspended in buffer containing 250 mM KCIand 30 mm EDTA (0) were loaded on metrizamide gradients. Followingcentrifugation, the gradients were fractionated. Each fraction was ana-

lyzed for [3H]leucine content, and the percentage of total radiolabel in

each fraction was calculated. The density of each fraction correspondsto the values shown in Figure 6. Fractions 3 and 4 contain hs granules;fractions 8 and 9 contain monosomes and polysomes. The solid line

indicates the gradient profile.

FIG. 10. Effects of KCI and EDTA on the association between heatshock proteins and ribosomes after resolution on metrizamide gradients.Equal volume aliquots from each fraction prepared in Figure 9 were

analyzed for protein content by one-dimensional SDS-PAGE and fluo-

ro,graphy (A, control ribosomes; B, ribosomes treated with 250 mM KCIand 30 mm EDTA). Fraction numbers are indicated above each lane

with fraction 1 representing the top of the gradient. Brackets indicatethe position of the low mol wt hs proteins. Arrows indicate the positionof hsp70.

r f

_l

_

__r

-__

_L_|._E

F

t.! 7

lslI

g,:

*''.._ ::

.... '"- i:-.+¢ |

::;4 i [4sL__

..... .o,.

I.-

;I.

c0= 30

I. 250

f 20S.

0.Eo 15

Z 10

0 5

I.-

1.330

1.275 _CIO

1.220 -._

C0

1.165 0

1.110O0

35 -j 1

1244 .MANSFIELD AND KEY

I

.9ifw VP.i.. ob .. is.4m, 0




spicuous in the cytoplasm, cortical cells from roots of 2-d-oldsoybean seedlings were examined by electron microscopy for thepresence of hs granules (Fig. 11). In tissue incubated at 28°C,the ribosomes were densely packed in the cytoplasm with a fairlyeven distribution (Fig. 11). After 3 h at 40°C, however, theribosomes were clustered in irregular clumps throughout the cy-toplasm (Fig. 11). No electron-dense particles that could be equatedwith hs granules of tomato (23) were observed in any of thespecimens examined. This suggests that the nature of the hsprotein aggregation phenomenon in soybean is different fromthat in tomato cells. Further evidence for this conclusion comesfrom the failure to observe any novel electron-dense particles intissue receiving a supraoptimal hs at 45°C after pretreatment at

FIG. 11. Electron microscopy of root cell cytoplasm. Seedlings wereincubated for 3 h at 280C (A), 3 h at 400C (B), or 3 h at 40°C, 3 h at28°C, and 1 h at 450C in succession (C). Apical 1 mm sections of theroot tips were fixed, embedded, sectioned, and then examined by trans-mission electron microscopy. Cytoplasm from cortical cells is shown here.Arrowheads in C indicate endoplasmic reticulum (bar = 0.25 ,m).

40 and 28°C (Fig. 11). Similar hs regimes resulted in a higherfrequency of hs granules in tomato cells (20).

DISCUSSIONOne of the rapid responses to hs is translational readout of

polysomes engaged in the translation of normal mRNAs (9, 19).If the hs temperature is maintained, a fraction of the ribosomeswill again form into polysomes as hs mRNAs become availablefor translation. Through isokinetic centrifugation it was possibleto analyze the association of hs proteins with monosomes andpolysomes. Heat shock proteins were not associated with mono-somes (Figs. 1, 2, and 4) even though they comprised the bulkof the ribosome population in heat-shocked soybean seedlings(9). Rather, hs proteins were bound to polysomes (Figs. 2 and4). Dissociation of ribosomal subunits with puromycin and 500mM KCI effectively removed the hs proteins. This result is con-sistent with that reported for ribosomes in tomato (20).Heat shock proteins are also aggregated into particles analo-

gous to the hs granules described by Nover et al. (20). Eventhough crude hs granule preparations were contaminated withribosomes (Figs. 6-8), several observations demonstrated thaths granules were distinct from ribosome-bound hs proteins. First,hsp70 was not detected in hs granule preparations (Fig. SB).Ribosome preparations from heat-shocked tissue contain hsp70(15). Second, while most of the low mol wt hs proteins werecommon to hs granules and heat-shocked ribosomes (Figs. SBand 8; Ref. 15), there were differences in the relative amountsof specific polypeptides in the two preparations. Third, severalacidic polypeptides not corresponding to hs proteins were de-tectable by staining in hs granule preparations. Fourth, when hsgranules were separated from ribosomes by isopycnic centrifu-gation, they exhibited a buoyant density of 1.20 to 1.21 g/cm3,which is characteristic of RNP particles (4). Monosomes andpolysomes had buoyant densities greater than 1.27 g/cm3. Fifth,the hs proteins associated with polysomes are, for the most part,not bound in the form of hs granules. Although 18% of theradiolabel in a ribosome preparation banded in fractions cor-responding to hs granules (Fig. 10, lanes 3 and 4), dissociationof ribosomes with KCI and EDTA prior to isopycnic centrifu-gation increased the amount of radioactivity in these fractionsto only 24%. Thus, the polysome-bound hs proteins are distinctfrom the granule-bound hs proteins. Heat shock proteins did notassociate with any preexisting particles as no pellets were ob-tained from control seedlings incubated at 28°C. Nover et al. (20)also determined that synthesis of hs proteins was essential forgranule formation. Blocking hs protein synthesis with actino-mycin D prevented the formation of hs granules.

Unlike tomato cells (20), hs granules were not observable asdistinct particles in the cytoplasm of soybean root cells. Whenanalyzed by isokinetic centrifugation, however, the radiolabel inthe granule preparation sedimented in advance of the polysomes(Fig. 6) indicating that hs granules were very large and, basedon the broadness of their distribution, heterogeneous in size.This heterogeneity can be explained by nonspecific associationof hs proteins. One conserved feature of the low mol wt hsproteins is a strongly hydrophobic domain in the carboxyl endof the proteins (17). This domain has been implicated in theaggregation phenomena observed among hs proteins in both plantand animal systems (17, 19). Detection of ribosomes in hs granulepreparations from soybean seedlings only after hs suggests thatdisruption of the soybean cells results in higher order aggrega-tions of hs granules with themselves, with soluble hs proteins,or with ribosomes. Failure to observe hs granules in vivo asparticles distinct from ribosomes in heat-shocked tissue (Fig. 11)indicates that they are not aggregating into such comparativelylarge structures as seen in tomato cells (20).

Nonspecific associations of hs proteins would also account for

1245



Plant Physiol. Vol. 86, 1988

the wide density distribution of material in heat-shocked ribo-some preparations. As anticipated, the largest fraction of hsprotein was associated with the ribosome fractions (Fig. 10). Ifthe low mol wt hs proteins were associated with polysomes in astoichiometric fashion, hs proteins would have been restrictedto specific fractions. Detection of hs proteins in lower densityfractions (Fig. 9) was not expected. The simplest explanation forthis observation is that a subpopulation of the polysomes is as-sociated with a sufficiently large amount of hs proteins to reducetheir buoyant density from that considered normal for poly-somes.

Association of hs proteins with ribosomes may have a protec-tive function (15). When seedlings were allowed to accumulatehs proteins after induction by a brief, severe hs or arsenite treat-ment, the relative amount of hs protein that subsequently as-sociated with ribosomes increased with temperature (15). If thetemperature was increased gradually, hs proteins accumulatedprior to reaching the optimal hs temperature (2, 8). This accu-mulation permitted the persistence of both normal protein syn-thesis and hs protein synthesis into temperature ranges severaldegrees above those observed when the temperature was changedabruptly (1, 9). It is conceivable that the association of hs proteinswith polysomes protects the translational machinery from thedeleterious effects of hs. This speculation is also consistent withthe eventual, apparent recovery of normal protein synthesis dur-ing prolonged hs (5, 7).The function of hs granules in the cytoplasm, however, is

unclear. Nover et al. (20) postulated that hs granules are uniqueto plants as no structure analogous to hs granules has been de-tected in animal systems. Unlike the cell-wide distribution of thelow mol wt hs proteins in plants (19), the low mol wt hs proteinsof Drosophila are associated with elements of the cytoskeletonduring hs (13). During recovery, they are found in 20S RNPparticles similar to prosomes, a class of scRNP particles impli-cated in the negative control of mRNA translation (3). The ab-sence of hs granules in animals was attributed to the lack ofconservation among the low mol wt hs proteins of plants andanimals (20). Nevertheless, a comparison of hydropathy profilesfor low mol wt hs proteins from Drosophila, Caenorhabditis,Xenopus, and soybean demonstrated structural similarities amongthese proteins (15). Given the fact that many hs proteins aredistributed throughout the cell, it is likely that they have multiplefunctions (19). It is possible that the formation of hs granules inplants is not a function of the unique nature of plant hs proteinsbut rather of normal proteins unique to plant cells in general.Further analysis of the normal proteins and RNAs associatedwith hs granules may provide insights into the functional signif-icance of these particles.

LITERATURE CITED

1. ALTSCHULER M, JP MASCARENHAS 1982 Heat shock proteins and effects ofheat shock in plants. Plant Mol Biol 1: 103-115

2. ALTSCHULER M, JP MASCARENHAS 1982 The synthesis of heat-shock andnormal proteins at high temperatures in plants and their possible roles insurvival under heat stress. In MJ Schlesinger, M Ashburner, A Tissieres,eds, Heat Shock: From Bacteria to Man. Cold Spring Harbor Laboratory,Cold Spring Harbor, NY, pp 321-327

3. BALLINGER DG, ML PARDUE 1983 The control of protein synthesis duringheat shock in Drosophila cells involves altered polypeptide elongation rates.Cell 33: 103-114

4. BUCKINGHAM ME, F GRos 1975 The use of metrizamide to separate cyto-plasmic ribonucleoprotein particles in muscle cell culture: a method for theisolation of messenger RNA, independent of its poly(A) content. FEBS Lett53: 355-359

5. COOPER P, THD Ho 1983 Heat shock proteins in maize. Plant Physiol 71: 215-222

6. GANTT JS, JL KEY 1983 Auxit-induced changes in the level of translatableribosomal protein messenger ribonucleic acid in soybean hypocotyl. Bio-chemistry 2: 4131-4139

7. KEY JL, JA KIMPEL, CY LIN, RT NAGAO, E VIERLING, E CZARNECKA, WBGURLEY, JK ROBERTS, MA MANSFIELD, L EDELMAN 1985 The heat shockresponse in soybean seedlings. In JL Key, T Kosuge, eds, Cellular andMolecular Biology of Plant Stress. Alan R. Liss, New York, pp 161-179

8. KEY JL, J KIMPEL, E VIERLING, CY LIN, RT NAGAO, E CZARNECKA, FSCHOFFL 1985 Physiological and molecular analyses of the heat shock re-sponse in plants. In BG Atkinson, DB Walden, eds, Changes in EukaryoticGene Expression in Response to Environmental Stress. Academic Press,Orlando, PL, pp 327-348

9. KEY JL, CY LIN, YM CHEN 1981 Heat shock proteins of higher plants. ProcNatl Acad Sci USA 78: 3526-3530

10. KLOPPSTECH K, G MEYER, G SCHUSTER, I OHAD 1985 Synthesis, transportand localization of a 22-kd heat-shock protein in the chloroplast membranesof peas and Chlamydomonas reinhardii. EMBO J 4: 1901-1909

11. LAEMMLI UK 1970 Cleavgge of structural proteins during assembly of the headof bacteriophage T4. Nature 227: 680-685

12. LATHANGUE NB 1984 A major heat-shock protein defined by a mnonoclonalantibody. EMBO J 3: 1871-1879

13. LEICHT BG, H BIESSMAN, KB PALTER, JJ BONNER 1986 Small heat shockproteins of Drosophila associate with the cytoskeleton. Proc Natl Acad SciUSA 83: 90-94

14. LIN CY, YM CHEN, JL KEY 1985 Solute leakage in soybean seedlings undervarious heat shock regimes. Plant Cell Physiol 26: 1493-1498

15. LIN CY, JK ROBERTS, JL KEY 1084 Acquisition of thermotolerance in soybeanseedlings. Plant Physiol 74: 152-160

16. MANS R, GD NOVELLI 1961 Measurement of the incorporation of radioactiveamino acids into protein by a filter-paper disk method. Arch Biochem Bio-phys 94: 48-53

17. NAGAO RT, E CZARNECKA, WB GURLEY, F SCHOFFL, JL KEY 1985 Genesfor the low-molecular-weight heat shock proteins of soybeans: sequenceanalysis of a multigene family. Mol Cell Biol 5: 3417-3428

18. NEUMANN D, KD SCHARF, L NdVER 1984 Heat shock induced changes ofplant cell ultrastructure and autoradiographic localization of heat shock pro-teins. Eur J Cell Biol 34: 254-264

19. NOvER L, D HELLMUND, D NEUMANN, KD SCHARF, E SERFLING 1984 Theheat shock response of eukaryotic cells. Biol Zbl 103: 357-435

20. NOVER L, KD SCHARF, D NEUMANN 1983 Formation of cytoplasmic heatshock granules in tomato cell cultures and leaves. Mol Cell Biol 3: 1648-1655

21. O'FAIELL PH 1975 High resolutiof two-dimensional electrophoresis of pro-teins. J Biol Chem 250: 4007-4021

22. SINABALDI R, T TURPEN 1985 A heat shock protein is encoded within mito-chondria of higher plants. J Biol Chem 260: 15382-15385

23. VELAZQUEZ JM, S SONODA, G BUGAISKY, S LINDQUIST 1983 Is the majorDrosophila heat shock protein present in cels that have not been heat shocked?J Cell Biol 96: 286-290

24. VIERLING EV, ML MISHKIND,GW SCHMIDT, JL KEY 1985 Specific heat shockproteins are transported into chloroplasts. Proc NatI Acad Sci USA 83: 361-365

1246 MANSFIELD AND KEY



cytoplasmic distribution ofheatshock proteins in soybean' · cytoplasmic distribution...

Documents