INTRODUCTION

When organisms are subjected to the stress of high temperature, synthesis of a set of new proteins, the heat shock proteins (hsps),¹ is induced (for reviews see Refs. 1-4). In some organisms, notably Drosophila, heat shock also suppresses the synthesis of normal cellular proteins (5, 6). The selective translation of hsps which is seen in Drosophila, however, is not a general feature of the heat shock response in plants (7), although it has been reported in soybean and tomato (3).

In our studies on barley aleurone layers we have observed another response to heat stress which may have a profound effect on cellular metabolism. The aleurone layer of cereal grains is a specialized tissue whose main function in vivo is the secretion of hydrolytic enzymes during seed germination. In the barley aleurone layer system, heat shock has a specific effect on the mRNA levels for secreted proteins. In response to heat shock the normally stable message for -amylase is rapidly degraded (8). At 25 °C the half-life of the -amylase mRNA was estimated to be over 100 h (9), yet within 30 min at 40 °C over 60% of it was degraded (10). The same phenomenon has been observed for the mRNA levels for the secreted proteins endochitinase and protease, whereas the mRNA levels of the nonsecreted proteins actin and tubulin were not affected by heat shock (10). Through the use of RNA synthesis inhibitors we have determined that synthesis of the hsps is not required for this selective message degradation (11).

To better understand this phenomenon we are focusing on those features of protein synthesis unique to secreted proteins. A feature common to most secreted proteins that distinguishes them from cytoplasmically localized proteins is the presence of a signal sequence. The signal sequence at the amino terminus of the nascent peptide chain is bound by the signal recognition particle (SRP) that targets the nascent protein-mRNAribosome complex to the ER. As characterized in mammalian cells, the SRP is a complex of a single RNA molecule and six proteins (for reviews see Refs. 12, 13). The 54-kDa protein subunit (Srp54) is the subunit of the complex that binds the signal sequence and the RNA backbone of SRP (14-17). SRP-like complexes have been isolated from plants (18, 19), although the protein components have not been characterized. Previously we reported the characterization from Arabidopsis thaliana of the first plant genes to be described for srp54 (20, 21). srp54 genes from tomato (Lysopersicon esculentum) have also been reported (22).

Following heat shock, the ER membranes of aleurone layers are significantly changed biochemically (23, 24). In GA₃-treated aleurone layers the ultrastructure of the ER membranes is also changed by heat shock (8). These observations suggest that a heat shock-induced change in some functions of the ER membranes may have occurred. Our hypothesis explaining the observed heat shock-induced selective degradation of secretory protein mRNAs was that SRP binding to the signal peptide was occurring during heat shock thus selecting for mRNAs encoding secreted proteins. If a step in the secretory pathway subsequent to the binding of SRP, yet prior to the resumption of translation, was damaged by heat shock, the inhibition of continued translation could then result in the observed enhanced mRNA degradation. We have therefore investigated the effect of a heat shock on the binding of the SRP complex to the ER membranes.

In order to test our hypothesis we have cloned barley srp54 cDNAs and produced antibodies to an Srp54 fusion protein expressed in Escherichia coli. The antibody was used as a marker for SRP in cell fractions isolated from aleurone layers. Following heat shock there was more SRP bound to the ER membranes and less in the free particle fraction, when compared with the non-heat-shocked samples. This suggests heat shock inhibited the release of SRP from the membranes which would, in turn, inhibit further translation of associated mRNAs. The results presented here are consistent with the hypothesis that in barley aleurone layers heat shock results in damage to a component of the early steps in the secretory pathway which then results in enhanced degradation of mRNAs for secreted proteins.

EXPERIMENTAL PROCEDURES

Barley seeds (Himalaya, 1988 harvest) were obtained from the Department of Agronomy and Soils, Washington State University, Pullman, WA. For aleurone layer preparations, embryo-less half-seeds were surface-sterilized with 1.5% hypochlorite and imbibed for 4 days on filter paper overlaying vermiculite, which was saturated with 20 mM sodium succinate, pH 5.0, containing 20 mM CaCl₂. The aleurone layers were dissected from the starchy endosperm under aseptic conditions and incubated in imbibition buffer. In some samples 1 µM GA₃ was added.

Total RNA was isolated by using the guanidine HCl method described by Cox (25). For gel blot analysis, polyadenylated RNA was fractionated from total RNA using the Poly(A)Tract mRNA isolation system (Promega, Madison, WI).

For RNA gel blot analysis, 2 µg of poly(A)⁺ RNA were subjected to electrophoresis in formaldehyde agarose gels and transferred to nylon membranes (Magnagraph, Micron Separations, Inc., Westborough, MA) as described by Selden (26). The RNA gel blots were hybridized as described previously (21).

Probes used for the RNA gel blots were barley srp54-1 cDNA, barley -amylase high pI cDNA (provided by Dr. J. Rogers), A. thaliana -tubulin cDNA (provided by Dr. D. P. Snustad), barley endochitinase cDNA (provided by Dr. S. Muthukrishnan), wheat (Triticum aestivum) SRP RNA (provided by Dr. C. Zweib), and maize (Zea mays) hsp70 cDNA (provided by Dr. T-h. D. Ho).

A barley leaf  gt10 cDNA library, prepared from RNA extracted from 7- to 10-day old seedlings, was provided by Dr. R. Zielinski. The 921-bp SacI/KpnI fragment of the A. thaliana srp54-1 cDNA clone (21) was used to screen the cDNA library as described by Huynh et al. (27). The hybridization conditions were as for the DNA and RNA gel blots, except that the hybridization solution was 12.5% formamide instead of 50% formamide. The filters were washed twice in 2 × SSC, 0.5% SDS at 42 °C, followed by a wash in 1 × SSC, 0.25% SDS at 50 °C. In a screen of about 150,000 clones (amplified from 10,000 original plaques) two positive plaques were obtained. Both had an insert of approximately 1.1 kilobases which was sequenced and found to be homologous to A. thaliana srp54 but lacked about 550 bp at the 5-end of the coding sequence.

In order to obtain a full-length barley cDNA clone, we constructed a cDNA library using poly(A)⁺ RNA from 4-day, germinated barley seedlings. The poly(A)⁺ RNA was fractionated from total RNA by two rounds of oligo(dT)-cellulose chromatography as described previously (8). EcoRI-ended cDNA was produced using a commercial kit (Pharmacia Biotech Inc.), ligated into the EcoRI site of the phage cloning vector lambda ZAPII (Stratagene, La Jolla, CA), and packaged using commercial extracts (Gigapack II Gold, Stratagene). The primary library contained 1.2 × 10⁶ plaque-forming units. The library was screened with the A. thaliana srp54-1 cDNA fragment and, at high stringency, with the 1.1-kilobase barley clone. In a screen of 500,000 clones, eight positive plaques were identified. Three of these were further characterized. The cDNA inserts were excised from  ZAPII as recombinant pBluescript SK() plasmids according to the manufacturer's protocols.

For sequencing, restriction enzyme fragments were subcloned into both M13mp18 and M13mp19 (28) to obtain inserts in opposite directions. Dideoxynucleotide sequencing (29) of single-stranded templates with T7 DNA polymerase was performed by using a commercial sequencing kit (U. S. Biochemical Corp.) with 17-base oligonucleotides as primers.

In order to be ensured of producing antibodies to an Srp54 protein that is expressed in aleurone layers, an aleurone layer library (Stratagene) was screened with the barley cDNAs, srp54-1 and srp54-3, isolated from the seedling library. In a screen of about 500,000 clones, a positive clone with a 700-bp insert corresponding to the srp54-1 and srp54-2 3-ends was obtained. srp54-2 was chosen for antibody production. Polyclonal antibodies were produced against a fusion protein expressed in E. coli. A 1041-bp PstI fragment covering amino acids 40-386 from the srp54-2 cDNA clone was subcloned into pBluescript (Stratagene). The HindIII-SmaI restriction fragment which was in the desired orientation was excised from pBluescript and ligated into the HindIII and SmaI sites of the expression vector pFLAG.ATS supplied with a commercial kit (IBI, New Haven, CT).

We could not purify the fusion protein by using the anti-FLAG M1 or M2 affinity gels since the fusion protein was insoluble in TBS (50 mM Tris-HCl, 150 mM NaCl, pH 7.4). A Saccharomyces cerevisiae Srp54 fusion protein expressed in bacteria was also reported to be insoluble (30). The barley fusion protein was purified instead by cell fractionation and gel separation. Briefly, the cells were lysed in buffer A (50 mM Tris, pH 8.0, 5 mM EDTA, 0.25 mg/ml lysozyme) for 5 min at room temperature. One-tenth volume of buffer B (1.5 M NaCl, 100 mM CaCl₂, 100 mM MgCl₂, 0.02 mg/ml DNase I, 50 µg/ml ovomucoid protease inhibitor) was added and the extract incubated at room temperature for 5 min. The extract was then centrifuged at 15,000 × g for 15 min at 4 °C. The pellet was suspended in TBS containing 8 M urea and then centrifuged at 15,000 × g for 15 min at 4 °C. The supernatant, containing the solubilized protein, was subjected to SDS-PAGE. The gel was stained with cold 0.25 M KCl, 1 mM DTT (31), and the 40-kDa fusion protein band excised. The gel slice was minced finely in 0.5 ml of phosphate-buffered saline (50 mM phosphate, 150 mM NaCl, pH 7.2). The homogenized gel, containing about 100 µg of the fusion protein, was injected into the subscapular space of a New Zealand White rabbit. Injections were given four times at 3-4-week intervals. Serum was collected, and antibodies specific to the fusion protein were prepared by affinity purification. The fusion protein was excised from SDS-PAGE, and the excised band was transferred to nitrocellulose. The bound fusion protein was then used in affinity purification of the antibody as described previously (32).

SDS-PAGE was carried out using 10% polyacrylamide gels (33), and the proteins were then transferred to nitrocellulose (Nitro-Bind, Micron Separations, Inc.) (34). Membranes were blocked in 5% nonfat dry milk in TTBS (0.05% Tween 20 in 20 mM Tris, pH 7.5, 180 mM NaCl) for 1 h and then incubated overnight in the affinity purified antibody in TTBS. Membranes were washed four times in TBS, 15 min each wash, and incubated in a 1:10,000 dilution of goat anti-rabbit IgG horseradish peroxidase conjugate (Promega Corporation, Madison, WI) for 1 h. Membranes were washed three times in TTBS prior to chemiluminescent detection using a commercial kit (enhanced chemiluminescence Western blotting detection system, Amersham Corp.).

The immunoblot demonstrating the specificity of the antibody was detected using a goat anti-rabbit IgG alkaline phosphatase secondary antibody and 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium (35).

Two-hundred aleurone layers per sample were chopped in 7 ml of chopping buffer (25 mM Hepes-BTP, pH 7.4, 3 mM EDTA, 1 mM DTT, and 0.5% bovine serum albumin) using an electric carving knife retrofitted to hold single-edged razor blades (36). Lysates from the chopped aleurone layers were filtered through Miracloth (Calbiochem), and cell debris was pelleted at 1,000 × g. For samples fractionated on continuous sucrose gradients, the 1,000 × g supernatant was centrifuged through gradients with 18-45% sucrose (sucrose w/w, buffered with Hepes-BTP, pH 7.4, 1 mM DTT) at 70,000 × g for 14 h using SW27.1 buckets on an SW27 rotor (Beckman Instruments, Palo Alto, CA). One-ml fractions were collected using a Buchler Auto Densi-flow gradient fraction collector (Buchler Instruments, Fort Lee, NJ) and a Gilson Microfractionator (Gilson, Middletown, WI).

For samples fractionated on discontinuous sucrose gradients, the 1,000 × g supernatant was centrifuged through layered sucrose cushions of 12, 20, 30, 40, and 50% sucrose (w/w sucrose, 2 ml each; buffered with Hepes-BTP, pH 7.4, 1 mM DTT) at 70,000 × g for 2 h using the same buckets and rotor used for the continuous gradients. The turbid bands at the 12/20% and 30/40% sucrose interfaces were collected as free and ER-bound SRP-containing fractions, respectively. The free SRP fraction was centrifuged at 20,000 × g to remove any contaminating membrane debris.

ER-containing fractions were identified by cytochrome c reductase (ER-marker enzyme) activity and assayed according to the procedure of Jones (37). In discontinuous sucrose step gradients, cytochrome c reductase activity in the turbid band at the 30/40% interface was 6-8-fold higher than at the 12/20% interface. Protein concentrations were determined using the bicinchoninic acid assay from Pierce. Sucrose concentrations were measured by refractometry.

To inhibit the initiation of translation, aleurone layers were incubated in 50 µM T-2 toxin (4,15-diacetoxy-3-hydroxy-8-[3-methylbutyryloxy]-12,13-epoxytrichothec-9-ene) (Sigma) for the final 4.5 h of the 19-h incubation. Samples were either maintained at 25 °C during this period or shifted to 40 °C during the last 3 h of incubation.

One-hundred aleurone layers per sample were frozen in liquid nitrogen and then ground into powder with dry ice in an electric coffee mill. Polysomes were isolated using modifications in the procedure of Evins (38). Briefly, powdered samples were thawed in buffer (200 mM Tris, pH 8.5, 30 mM MgCl₂, 50 mM KCl, 350 mM sucrose, 6 mM 2-mercaptoethanol) containing an RNase/protease inhibitor mixture (30), filtered through Miracloth, then spun at 15,000 × g at 4 °C for 15 min to pellet cell debris. The supernatant was layered onto pads of 1 and 2 M ice-cold sucrose (in the above buffer) and centrifuged at 75,000 × g at 4 °C for 12 h in an SW27 rotor. The pellet, constituting free polysomes, was resuspended in 200 µl of diethyl pyrocarbonate-treated water and stored at 80 °C. The band at the 1/2 M sucrose interface was collected and diluted to 8% sucrose with diethyl pyrocarbonate-treated water. Nonidet P-40 was added to a concentration of 0.5%, and the sample was layered onto a 2 M sucrose pad and centrifuged at 90,000 × g at 4 °C for 1.5 h. The pellet, constituting membrane-bound microsomes, was resuspended in 200 µl of diethyl pyrocarbonate-treated water and stored at 80 °C.

To isolate RNA from the polysome preparations, samples were extracted twice with phenol/chloroform/isoamyl alcohol (25:24:1), ethanol-precipitated, and resuspended in water. Three µg of RNA extracted from polysomes was spotted onto GeneScreen Plus membranes (DuPont NEN) according to manufacturer's instructions. The blots were then prehybridized, hybridized, and washed according to the procedures of Church and Gilbert (39). Cloned cDNA for the high pI (pM/C) (40) isozyme of -amylase was nick-translated using a commercial nick-translation kit (Life Technologies, Inc.) in the presence of -[³²P]dCTP (specific activity >500 Ci/mmol, obtained from DuPont NEN). RNA blot analyses indicated that RNA isolated from the samples was the size expected for -amylase mRNA.

RESULTS

Barley srp54 cDNA clones were obtained by screening a cDNA library at reduced stringency with an A. thaliana srp54 cDNA fragment (21). Three different cDNA clones were obtained and were designated srp54-1, -2, and -3. All three contained the complete coding sequence. They had 76, 142, and 20 bp, respectively, of sequence 5 to the initiator Met codon. They contained 176, 312, and 215 bp, respectively, of 3-noncoding sequence. A comparison of the deduced amino acid sequences from the srp54-1 and srp54-3 cDNA clones is shown in Fig. 1.

srp54-1 and srp54-2 had nearly identical noncoding and coding DNA sequences, with only a single bp difference in the coding region. This results in an amino acid difference at position 96 in the protein sequence, threonine in srp54-1 and proline in srp54-2. The amino acid at the corresponding position in srp54-3 is a proline. The amino acid sequence of srp54-2 is not included in Fig. 1. DNA gel blot analysis (data not shown) indicated the presence of two genes or alleles suggesting the srp54-1 and srp54-2 cDNA clones do originate from distinct transcripts rather than from a cloning artifact.

The deduced amino acid sequences of the three clones predict proteins with molecular masses of 54,444, 54,448, and 53,761 Da, respectively. These sequences contain the conserved GTP binding motifs and the carboxyl-terminal, Met-rich domains (M-domains) found in other Srp54 sequences (21, 43). The M-domains of the barley sequences also contain the proposed RNA binding motif (42). The deduced amino acid sequences of the three barley srp54 cDNA clones are 79% identical. Amino acid sequences for the barley clones are 65-72% identical to the A. thaliana (21) sequences.

RNA gel blot analysis indicated the message level of srp54 in aleurone layers was slightly enhanced by GA₃ treatment (Fig. 2, lanes 2 and 3). The srp54 message levels at both 8 and 24 h of GA₃ treatment were approximately equal. This is in contrast to the dramatic effect of GA₃ on the message level of the secreted enzyme -amylase. As would be expected, srp54 message was also detectable in seedling tissue (Fig. 2, lane 4).

Effect of GA₃ on transcript levels for srp54 and -amylase.

The effect of a 3-h heat shock on the srp54 message level and the level of the RNA component of SRP (43) was investigated in aleurone layers incubated both in the absence or presence of GA₃ (Fig. 3). Neither was affected by the heat shock treatment. The level of the SRP RNA was also unaffected by GA₃. Heat shock did have the expected effect of dramatically reducing the steady state message levels of the secreted proteins -amylase and endochitinase and in inducing expression of Hsp70 (Fig. 3). Endochitinase is a secreted protein which is expressed both in the absence and presence of GA₃ (44).

Effect of GA₃ and heat shock on transcript levels for Srp54, -amylase, endochitinase, Hsp70, and on the level of SRP RNA.

In order to investigate the protein level and cellular localization of Srp54, polyclonal antibodies were prepared to a 40-kDa polypeptide of Srp54-2 expressed as a fusion protein in E. coli. The fusion protein consisted of the amino acids 40-386 encoded by the srp54-2 clone. Fig. 4 demonstrates the specificity of the antibody for Srp54. In the absence of IPTG induction of expression of the fusion protein, there was no antibody reaction (Fig. 4, lane 5). Following IPTG induction a single major protein at the expected size of the fusion protein was recognized by the antibody (Fig. 4, lane 6). The lower molecular weight immunoreactive bands are likely degradation products of Srp54 since they are not present in the absence of IPTG (Fig. 4, lane 5).

In protein blots of aleurone layer total protein extracts, Srp54 was barely detectable (data not shown) indicating it is not an abundant cellular protein. Concentration of SRP by cellular fractionation was required for good antibody detection of Srp54 (see below). Incubation with GA₃ did not result in any discernible difference in the total protein level of Srp54 (data not shown).

Based on our hypothesis regarding the effect of heat shock on the secretory system, we considered the possibility that heat shock may inhibit the binding of SRP to the ER membranes. If this occurred, it would be expected to extend the translational arrest imposed on binding of SRP to the signal peptide. Such an effect might potentiate the observed enhanced message degradation. To test this possibility, we compared the cellular localization of SRP in heat-shocked and non-heat-shocked aleurone layers.

Cell fractionation in continuous sucrose gradients followed by immunoblot analysis was carried out to localize Srp54, as a marker for SRP, in aleurone layers incubated for 16 h at 25 °C in the presence of GA₃ and in samples also subjected to a 3-h heat shock at 40 °C. For both samples, SRP had a bimodal distribution with peaks centered around 15-22 and 35-40% sucrose (Fig. 5B). The prominent protein band seen in the stained gels (Fig. 5A) is bovine serum albumin (66 kDa) which is a component of the extraction buffer. The 15-22% sucrose fractions represent the free SRP (18, 19, 32, 45). The SRP at 35-40% sucrose correlated with the ER enzyme marker activity, cytochrome c reductase (Fig. 5C), and represents the ER-bound SRP. These results indicate that for both non-heat-shocked and heat-shocked aleurone layers, Srp54 exists in the cell as part of a complex with migration in sucrose gradients as would be expected for free and ER-bound SRP. In this experiment the gels were loaded as equal volumes from each fraction so the 25 °C and heat-shocked samples cannot be directly compared. In order to load enough volume to detect Srp54 in the heavier fractions, the lightest fractions, which contained the bovine serum albumin, were necessarily overloaded, but this did not affect detection of Srp54. In some lanes, two protein species were detected. We do not yet know the reason for this, but may be due to degradation and/or protein modification.

In order to determine if heat shock had any quantitative effect on the distribution of SRP between the free and bound forms, sucrose step gradients were designed to separate the two SRP pools, and equal protein aliquots were subjected to immunoblot analysis. A 30-45% step was used to fractionate the ER-bound SRP from the free particle form found in the 12-20% step. When compared on the basis of equal protein, heat shock clearly had an effect on the partitioning of SRP between the free and bound forms. Heat shock resulted in an increase in the amount of SRP bound to the ER membrane fraction and a decrease in the amount in the free particle pool (Fig. 6). Similar results were obtained both in the presence and absence of GA₃. These experiments were done twice with similar results.

In order to obtain a quantitative measure of the differences, the signals on the immunoblots were digitized. The ratios of the Srp54 amount in the heat-shocked sample relative to the non-heat-shocked sample are dramatically different for the free versus bound forms (Fig. 6C). The quantitative data confirm the interpretation that there was more SRP in the bound form following a heat shock.

From these results it appeared that heat shock did not inhibit binding of SRP to the ER membranes. Rather, it appeared that heat shock inhibited the release of SRP from the membranes. This would also be expected to extend the translational arrest, since SRP release is required for translation to continue.

The heat shock-inducible increase in SRP binding may simply be due to nonspecific binding of protein to the ER. This possibility was investigated by incubating aleurone layers in the presence of T-2 toxin, an inhibitor of translation initiation in eukaryotes (46). Because SRP binding requires the translation of the signal peptide, T-2 toxin-induced inhibition of translation should prevent the binding of SRP to the ER during heat shock if the association is a specific binding to the SRP receptor rather than a nonspecific binding of heat-denatured proteins to the ER. Incubation in 50 µM T-2 toxin inhibited the incorporation of [³⁵S]methionine into proteins over 75,000-fold (data not shown).

Dot blot analysis of RNA extracted from free and ER-bound polysomes was used to determine the effect of heat shock and T-2 toxin on the distribution of -amylase mRNA. Immunoblot analysis using antibody to Srp54 was used to determine the proportions of free and ER-bound SRP. The signals were quantified using the program NIH Image. The results of these experiments are presented in Table I. This experiment was repeated twice with similar results. In these experiments the absolute pixil counts cannot be directly compared among samples, rather the ratios of bound:free are comparable.

Table I. Distribution of polysomes and Srp54 in barley aleurone layers incubated at normal and heat shock temperatures with and without the translation initiation inhibitor T-2 Barley aleurone layers were incubated in the presence of GA3 (GA) for 19 h; the final 3 h of each sample's incubation was either at 25 or 40 °C (HS). T-2 samples were incubated with 50 µM T-2 toxin during the final 4.5 h of incubation. -Amylase-translating polysomes were quantified by digitization measurements from dot blots. Srp54 was quantified by digitization measurements from Western blots. Numbers are relative densities measured by the pixil count method using NIH Image. Sample Amylase polysomes Srp54 Bound Free Bound:Free Bound Free Bound:Free Experiment 1   GA 964 378 2.55 21 66 0.32   GA/HS 421 289 1.46 121 32 3.78   GA and T-2 102 1,092 0.09 8 90 0.09   GA/HS and T-2 96 736 0.13 41 118 0.35 Experiment 2   GA 850 213 3.99 9 40 0.23   GA/HS 536 337 1.59 55 14 3.93   GA and T-2 115 921 0.12 <1 52 0.02   GA/HS and T-2 77 671 0.11 22 46 0.48

In the control sample (incubation with GA₃), -amylase message was found predominantly in the ER-bound polysome fraction, and SRP was found predominantly in the free particle fraction. Since SRP functions catalytically by cycling between the cytoplasm and ER, these results are as we expected. As was observed in the experiment depicted in Fig. 6, heat shock resulted in an increase in the proportion of the ER-bound form of SRP. Concomitantly, there was a reduction in the proportion of -amylase message detected in the ER-bound polysomes. These results are consistent with a heat-induced inhibition of SRP cycling.

T-2 toxin resulted in reductions in the proportions of both ER-bound SRP and -amylase message in the ER-bound polysomes. The inhibition of translation initiation prevents the synthesis of signal peptides that are required for SRP binding to the ER membranes. When a heat shock was applied to toxin-treated samples, there were modest increases in the proportions of both ER-bound amylase message and ER-bound SRP. These results indicate that there is some heat-induced nonspecific binding of SRP to the ER membranes. However, the proportion of heat-induced nonspecifically bound SRP (0.35) is only one-tenth that measured in the absence of the toxin (3.78). These results indicate that most of the heat-induced binding of SRP to the membranes is due to a specific interaction. The data using T-2 toxin support the hypothesis that heat shock results in a biologically meaningful inhibition of SRP cycling between the free and bound forms.

DISCUSSION

In this report we used antibodies to barley Srp54 to determine the effect of heat shock on SRP in barley aleurone layers. As a first step in this process we characterized barley srp54 cDNAs and determined the effect of GA₃ and heat shock on srp54 expression. GA₃ results in a dramatic change in the profile of secreted proteins synthesized by aleurone layers and a change in total amount of secreted protein. The overall increase in amount of protein secreted was determined to be 1.47-fold (24). Incubation of aleurone layers with GA₃ resulted in a slight increase in srp54 message level but no discernible increase in the total protein level of Srp54. Since the protein level was so low in a crude extract, however, it will be necessary to use a more sensitive technique to accurately determine the effect of GA₃ on protein level. There was also no detectable difference in the level of the RNA component of SRP in response to GA₃. Whether GA₃ results in an increase in the amount of functional SRP is not yet known.

Cell fractionation experiments revealed two pools of Srp54. One pool migrated as expected for free SRP and one was bound to the ER membranes. These results, in addition to the amino acid sequence homology, support the assignment of barley Srp54 as a component of a functional SRP. Binding of in vitro synthesized tomato Srp54 to tomato and human SRP RNA and to human SRP19 (22) also supports the expectation that plant SRPs are highly homologous to mammalian SRP.

Heat shock resulted in a marked increase in the amount of SRP bound to ER membranes, as assessed by immunoblot analysis using antibody to Srp54. There was a concomitant decrease in the amount of free SRP. These results suggest that the normal cycling of SRP between the free and bound forms was perturbed at the step of dissociation of SRP from the membrane. The use of T-2 toxin to inhibit translation initiation indicated that most of the observed heat-induced increase in binding was due to specific interactions between SRP and the membranes.

Based on our experimental data we can propose a model for the observed heat shock-induced selective degradation of messages for secreted proteins. Binding of SRP to the signal sequence of a nascent polypeptide chain results in a transient arrest in further translation (47). During this period of translation arrest the SRP-ribosome complex becomes bound to the SRP receptor of the ER membrane. The SRP then dissociates from the membrane, and translation of the message continues with the newly synthesized protein entering the ER lumen. Both Srp54 and the SRP receptor have GTP binding domains (41). GTP hydrolysis is required for dissociation of SRP from the receptor, and the dissociation is inhibited by a nonhydrolyzable GTP analog (48). A model for the order of events leading to SRP dissociation from the ER has been proposed (49). A heat-induced block at any of the proposed steps could result in the observed stabilization of the association of SRP with the membranes.

How could a block in SRP cycling result in increased mRNA degradation? The observed heat-induced stabilization of the SRP binding to the ER membranes would be expected to extend the translational arrest initiated by binding of SRP to the nascent signal sequence. In other situations, it has been established that messages on which translation has been prematurely terminated are highly susceptible to degradation. For example, frameshift null alleles of several seed storage proteins have been found to have normal levels of transcription but greatly reduced steady state message levels (50-52). This implies that the early termination of translation caused by the nonsense mutation leads to enhanced degradation of the message (51). The processes of translation and mRNA decay have been found to be linked, and translational pausing has been proposed to have a role in mRNA degradation of both inherently unstable mRNAs and in nonsense-mediated decay by exposing downstream nuclease recognition sites in the mRNA (53). The heat-shock induced inhibition of SRP dissociation from the membranes and the concomitant halt in translation may thus lead to the observed specific degradation of messages for secreted proteins.

Although the effect of heat shock on the mRNA levels for secreted proteins in the barley aleurone layer system is quite dramatic, in some systems messages that are translated in association with ER membranes are not degraded in response to heat shock. Message levels for soybean conglycinin and glycinin (54), Phaseolus vulgaris phytohemagglutinin (55), and maize 19-kDa zein² were not reduced by heat shock. These proteins are all storage proteins synthesized in developing seeds in association with ER membranes and ultimately deposited in protein bodies. Perhaps components of the secretory systems in the various tissues may have fundamental differences that result in different sensitivities to heat shock. In electroporated carrot (Daucus carota) protoplasts, heat shock increased stability of the reporter mRNA (56).

In other systems, however, heat stress has been found to inhibit various developmentally and environmentally induced processes that also rely on the synthesis of secreted proteins. In some cases degradation of messages for secreted proteins has also been observed. Wounded carrot root cells secrete the cell wall protein extensin that is believed to be involved in cell wall repair (57). In response to heat shock the induced, normally stable, message for extensin is degraded (58). Tomato fruit ripening is inhibited at 35 °C, and the message for the ripening induced secreted protein polygalacturonase shows a rapid decline (59). It is now known that polygalacturonase is not a major determinant of fruit softening, but other, as yet undescribed, secreted enzymes are likely to be important (60).

In some systems, heat shock has been shown to inhibit some functions that rely on the synthesis of secreted or ER membrane-bound proteins, but the effect on specific message levels has not been investigated. In general, heat stress has been found to delay or inhibit pathogen-induced responses associated with resistance, such as phytoalexin synthesis (61) and phenylalanine ammonia lyase and chalcone synthase enzyme induction (62). In some species a short high temperature treatment of styles can inactivate the self-incompatibility reaction (63) which is mediated through secreted proteins (64, 65). Secreted proteins are also required for somatic embryo development (66), and heat shock has been shown to arrest the further development of globular embryos (67). It would be interesting to determine if a heat-induced selective degradation of transcripts for secreted proteins is also occurring in these systems.