THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1986 by The American Society of Biological Chemists, Inc

Vol. 261, No. 33, Issue of November 25, pp. 15740-15745.1986 Printed in U. S. A.

Transcription and Post-transcriptional Regulation of Avian HSP70 Gene Expression*

(Received for publication, July 2, 1986)

Sunandita S. BanerjiS, Lee Berg, and Richard I. Morimotog From the Department of Biochemistry, Molecular Biology and Cell Biology, Northwestern University, Euanston, Illinois 60201

Transient incubation of chicken lymphoblastoid (MSB) cells at elevated temperatures induces the syn- thesis of three heat shock proteins of 89,000 Da (HSP89), 70,000 Da (HSP70), and 23,000 Da (HSP23). We have examined the effects of heat shock on the transcription and post-transcriptional regula- tion of the chicken HSP70 and &actin genes. The rate of HSP70 transcription is rapidly induced by heat shock, reaches maximal levels by 60 min, and there- after decreases. The level of HSP70 mRNA increases 20-fold by 60 min and remains constant through 6 h of heat shock. Upon return of heat-shocked cells to normal growth temperatures, the level of HSP70 mRNA rapidly decreases to pre-heat shock levels, These results suggest that HSP70 mRNA is stably maintained and translated during heat shock, but rap- idly degraded during recovery from heat shock. The effect of heat shock on &actin mRNA is opposite to the apparent stabilizing effects of elevated temperatures on HSP7O mRNA.

Heat shock has a multitude of effects on transcription and protein synthesis in eukaryotic cells (1, 2). Studies in Droso- phila have shown that the rate of heat shock gene transcrip- tion and the level of the heat shock messenger RNAs increase rapidly following temperature elevation. The activation of a small number of heat shock genes occurs while most other RNA polymerase I1 transcribed genes are repressed (3-5). At the level of protein synthesis, mRNAs that encode heat shock proteins are preferentially translated, whereas messenger RNAs for non-heat shock proteins are poorly translated (6- 9).

During heat shock of chicken lymphoblastoid cells, HSP89, -70, and -23 become the predominantly synthesized proteins. The initial burst of HSP70 synthesis is due to the de nouo transcription and accumulation of HSP7O mRNA (10). It is unclear whether the continued synthesis of HSP70 during heat shock is due to the continued transcription of the HSP70 gene or due to increased stability of HSP70 mRNA. In this paper, we examine the transcriptional and post-transcrip- tional regulation of heat shock mRNAs and non-heat shock mRNAs using the cloned chicken HSP70 and /?-actin genes as specific hybridization probes. We show that transcription of the HSP70 is rapidly but transiently induced during con-

* These studies were supported in part by grants from the National Institutes of Health and the Leukemia Foundation, Inc. The costs of publicat.ion of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduer- tisrrnent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Supported by a fellowship from the AMOCO Foundation. 0 To whom correspondence and reprint requests should be ad-

dressed.

tinuous heat shock. The newly transcribed heat shock mRNAs are stably maintained and actively translated in cells kept a t heat shock temperatures. Upon the return of cells to the normal growth temperature, the level of HSP70 mRNA rap- idly decreases. This is in contrast to the effects of heat shock on the synthesis of control proteins such as actin. During heat shock, actin synthesis is repressed, in part due to the turnover of actin mRNA in the cytoplasm. As heat-shocked cells re- cover, both the levels of actin mRNA and actin protein synthesis increase.

MATERIALS AND METHODS

Cell Growth and Heat Shock-MSB cells (11) were maintained in midlog phase (2-3 X lo6 cells/ml) by continuous culture in RPMI 1640 medium containing 5% fetal calf serum and 1% chicken serum in an atmosphere of 5% COP at 37 "C. The cells were heat-shocked by submerging aliquots in 1.5-ml Eppendorf tubes in a 45 "C water bath with shaking. The cultures reached 45 "C within 5 min, which was taken as zero time of heat shock. Larger volumes of cells were heat-shocked by resuspension into prewarmed media. Cell viability was determined by staining cells with 1% crystal violet.

Analysis of Protein Synthesis-The cells were heat-shocked as described for each experiment, washed in phosphate-buffered saline, and resuspended in methionine-free Dulbecco's modified Eagle's me- dia supplemented with 50 pCi/ml [35S]methionine (1075 Ci/mmol, Amersham). I n vivo labeling was carried out at 37 "C for 1 hr, after which the cells were washed in phosphate-buffered saline and frozen at -70 "C or immediately lysed in SDS' sample buffer. Patterns of protein synthesis were analyzed by SDS-10% polyacrylamide gel electrophoresis (12), stained with 0.25% Coomassie Brilliant Blue in 50% methanol, 10% acetic acid, and processed for fluorography (13). Analysis of the radioactivity associated with specific proteins was accomplished by densitometric scans of autoradiograms. The total amount of [35S]methionine incorporation was determined by trichlo- roacetic acid precipitation of SDS-solubilized cell lysates and liquid scintillation counting using a xylene-based fluor (OCS, Amersham).

Transcription and mRNA Levels-The relative amounts of specific messenger RNAs from control and heat-shocked cells were deter- mined by Northern blot analysis using "P-labeled (nick-translated) plasmids containing either the chicken HSP70 gene, PC1.8 (14), or the chicken @-actin gene, pAl (15), as hybridization probes. In order to determine the kinetics of HSP70 mRNA accumulation, total RNA was isolated by a rapid lysis procedure' from control and rapidly heat- shocked cells. After the desired time of incubation, the cells were pelleted in an Eppendorf microfuge for 10 s, and the media were removed. Fifty microliters of freshly prepared ice-cold lysis buffer (10 mM Tris-HCI, pH 7.5, 1.5 mg/ml heparin, 10 miu vanadyladenosine, and 1.5 mM MgCI,) was added, the tube was vortexed vigorously for 5 s, and the contents were quickly frozen on powdered dry ice. The cell lysates were thawed, RNase-free DNase was added, and the lysates were incubated a t room temperature for 15 min. Following incubation, 200 p1 of 1.2 X NaPF (1.2 M NaC1, 48 mM sodium phosphate buffer, pH 7.0, 7.2% formaldehyde) was added, and the contents were vortexed for 5 s and pelleted in a Microfuge for 2 min a t room temperature. The supernatant was removed and stored at 20 "C. Aliquots of supernatants were dilut,ed to 200 pl with 1 X NaPF,

__"____"______" ' The abbreviation used is: SDS, sodium dodecyl sulfate. ' Z. Krawczyk and C. Wu, personal communication.

15 740

Avian HSP70 Transcription 15741

heated to 65 "C, cooled on ice, and filtered onto nitrocellulose using a Schleicher & Schuell Minifold. The filter was washed with 1 M phosphate buffer, pH 7.0, dried, and baked at 80 "C for 1.5 h under vacuum. Filters were hybridized to:"P-labeled (nick-translated) pC1.8 and pAl in a buffer containing 6 X SSC (0.9 M NaCI, 90 mM sodium citrate), 50% formamide, 0.1% SDS, 5 X Denhardt's solution, and 50 pg of E. coli tRNA per ml a t 42 "C. After hybridization, blots were washed at 65 "C in 6 X SSC + 0.1% SDS followed by a wash in 1 X SSC + 0.1% SDS.

Northern blot analysis (16) was performed on RNA isolated from cell lysates (17) and purified by oligo(dT) chromatography (18). The RNA was denatured with glyoxal and dimethyl sulfoxide (19). sepa- rated on a 1.4% agarose-phosphate gel, and transferred to nitrocel- lulose. Filters were baked and hybridized as described.

(10, 20). RNA runoff transcripts were labeled with [ T J U T P (ICN, In oitro runoff transcription was performed in isolated MSB nuclei

Irvine, CA) and hybridized to filters spotted with plasmid DNAs (21) pAl (actin) and pC1.8 (HSP70). The chicken @-globin gene, pBlRR15 (22), and the plasmid vector, pAT153, were used as negative controls. The filters were hybridized with the labeled transcripts a t 42 "C and washed as described. The hybridization intensities were quantified by densitometry.

RESULTS

Changes in the Patterns of Protein Synthesis following Heat Shock-Transient incubation of chicken MSB cells at ele- vated temperatures alters both the patterns and the rates of protein synthesis. To examine the effects of heat shock, we incubated MSR cells at 45 "C for up to 5 h. Aliquots were removed at various times and pulse-labeled with [%]methi- onine and protein synthesis was analyzed by SDS-polyacryl- amide gel electrophoresis. The most dramatic effect of a shift from normal growth temperatures (37 "C) to heat shock tem- peratures (45 "C) is the increased synthesis of three proteins of molecular sizes 89,000 Da (HSP89), 70,000 Da (HSP70), and 23,000 Da (HSP23) (Fig. 1, Refs. 23-30). The induced svnthesis of HSP89, -70, and -23 is detected within 30 min of temperature shift (Fig. 1, lane 2). During this same period, the level of proteins normally synthesized (control proteins) is repressed well below its level of synthesis in control cells

1 2 3 4 5 6

Actin-

"HSP 23

FIG. 1. The effect of elevated temperatures on protein syn- thesis in chicken MSR cells. Cultures of MSR cells were heat- shocked and maintained at 45 "C for up to 5 h. At each time point (0, 0.5, 1, 2, 3 , and 5 h corresponding to lanes 1 to 6, respectively), aliquots were removed, and the cells were centrifuged and labeled in [""Slmethionine. Fluorograms are of SDS-polyacrylamide gel analysis o f equal cell lysates (2 X lo5). The heat shock proteins (HSP 89, -70, "23) and the control proteins, 94-kDa and actin, are labeled.

(Figs. 1 and 2). T h e repressive effects of heat shock on control protein synthesis are detected after a 60-min incubation a t 45 "C. When the cells are incubated for up to 5 h at 45 "C, the synthesis of control proteins is significantly reduced, while heat shock proteins continue to be synthesized at high levels.

We quantified the levels of synthesis of HSP89, -70, and -23 and control proteins of 94,000 Da and 43,000 Da (actin) by densitometry of the corresponding region from the auto- radiogram. All three heat shock proteins are induced by 30 min of heat shock, reach maximum levels of synthesis by 2 h, and continue to be synthesized at high levels through 5 h of heat shock (Fig. 2). While all three HSPs are induced with similar kinetics, HSP70 synthesis is induced 30-fold by 2 h, HSP89 is induced 3-5-fold, and HSP23 is induced 14-fold. During heat shock, the overall level of control protein synthe- sis gradually declines. The syntheses of both actin and a 94- kDa protein decrease within 30 min of heat shock and con- tinue to decline through 5 h at 45 "C to 20% of control levels of actin synthesis and 13% of control levels of 94-kDa syn- thesis.

I t is likely that the elevated synthesis of heat shock proteins * is due to the increased transcription and accumulation of heat shock mRNAs. The decreased translation of the mRNAs that encode control proteins could be due either to translational control in which non-heat shock mRNAs are sequestered in a translationally repressed state or to the turnover of control mRNAs.

Kinetics of HSP70 Transcription and Messenger RNA Ac- cumulation-We examined the kinetics of HSP70 mRNA accumulation in heat-shocked cells by measuring messenger RNA levels with the cloned chicken HSP70 gene as a hybrid- ization probe. Total RNA was isolat,ed from aliquots of cells removed after various times of incubation a t 4.5 "C, bound to nitrocellulose filters, and hybridized with either :'2P-labeled pC1.8, a plasmid containing the chicken HSP70 gene, or pAl , a plasmid containing the chicken @-actin gene (Fig. 3A). T h e level of HSP70 mRNA increases within 5 min of heat shock and accumulat.es rapidly to 15-fold higher levels over a 2-h

2 0

z 3

Y U W a

-

W

I-

-I W n

1 40 a

FIG. 2. Synthetic levels of selected control and heat shock proteins. Autoradiograms of the experiment descrihed in Fig. 1 were quantified hy densitometric scans. Several different exposures of the x-ray film were analyzed to ensure linearity of the autoradiographic response. Quantitation of 94-kDa (0- - -0), actin (A- - -A), HSP2.7 (A-A), and HSP89 (-) synthesis was determined I'rom a 3-h exposure (scale is shown on the left). Synthetic levels of HSP70 (X-X) was determined from a 1-h exposure of the autoradiogram, and the results were normalized to the levels of HSP89 (scale is shown on the right).

Avian HSP70 Transcription 15742

(A) ( C ) Time of Incubation (rnin)

0 5 10 45 30 60 90 4 2 0 1 8 0 240 C HS C HS

45 10 0 0 0 0 0 * . I p A l

37 y j P c l . 6

pCl.8 pA1 probe

(B)

8.0 1 HSP 70

/

x-x-x X'

W

X i

1 I 1 1 1 1 1 1 1

60 90 120 180 240 Time of Heat Shock (min)

FIG. 3. The effect of heat shock on the synthesis and accu- mulation of HSP7O and actin mRNA. A , MSR cells were heat- shocked for 0, 5, 10, 15, 30, 60, 120, 180, and 240 min a t 45 "C, aliquots were removed a t each time point, and total cellular RNA was prepared as described under "Materials and Methods." Dilutions of RNA corresponding to 5 X lo5 cells were spotted onto nitrocellulose and hybridized t.o :'2P-labeled pC1.8 (HSP70 gene) or the chicken actin gene (pAl). RNA was also prepared from MSB cells maintained a t 37 "C for the same time points and hybridized with "2P-labeled pC1.8. HSP70, X-X; actin, M. R, quantification of relative levels of HSP70 and actin mRNAs, scanning densitometry of auto- radiograms in pnnd a. C, Northern blot analvsis of poly(A)' mRNA isolated from MSR cells maintained a t 37 "C or 45 "C (60 min) and hybridized with pC1.8 or pAl.

incubation, whereas the levels of actin mRNA initially in- creases by 25% and then decreases to 50% of preshock levels over 2 h a t 45 "C (Fig. 3, A and 23). Similarly, by Northern blot. analysis of poly(A)+ RNA from control and heat-shocked cells, we see that the 2.6-kilobase pair HSP70 mRNA is detected at low levels in control cells and becomes abundant after heat shock, whereas the abundance of the 1.8-kilobase pair actin mRNA decreases following heat shock (Fig. 3C).

To determine whether the increase in HSP70 mRNA levels corresponds to a change in the rate of transcription of the HSP70 gene, in vitro runoff transcription in isolated nuclei was performed. This procedure allows a direct measure of the rate of transcription for a specific gene by hybridization of the '"P-labeled RNA in vitro transcribed to cloned gene se- quences. Nuclei were isolated from cells heat-shocked for 5 min or up to 2 h, and the preinitiated transcripts were elon- gated for 15 min in reaction mixtures containing "?P-labeled UTP. Total RNA was isolated and hybridized to the plasmid vector pAT153, the chicken 13-actin gene (pAl), the chicken HSP70 gene (pC1.8), or the chicken globin gene (pPlBR15). The results were quantified and are presented in Fig. 4. The rate of HSP70 transcription is rapidly induced between 5 and 15 min of heat shock and reaches maximal levels of synthesis a t 60 min. The rapid 20-fold increase in the rate of transcrip- tion of the HSP70 gene accounts for the 15-20-fold accumu- lation of HSP70 mRNA. Between 60 and 120 min after the temperature shift, the rate of transcription decreases signifi-

(A) Time of Heat Shock (rnin)

0 5 15 30 60 (20

pAT153 - - pA1 - -

- . " - .* - 0 6 -

nC1.8 - - - - I " .- I

, O O t f /

n 500 t I

2oo/// 100

30 60 90 120 I80 240

Time of Heat Shock (min)

FIG. 4. Kinetics of transcription of the HSP70 gene during heat shock. MSB cells were heat-shocked for 0, 5, 15, 30, 60, and 120 min, and nuclei were prepared. In vitro transcription was carried out, and '"P-labeled RNA was isolated and hybridized for 72 h to filters containing denatured DNA from plasmids, pAT153, pAl, PBlBR15, and pC1.8. The two concentrations of plasmid DNAs correspond to 1 pg and 0.5 pg of DNA. A , autoradiogram of RNA- DNA hybrids. R, densitometric scans of synthesis (X-X) (super- imposed on the accumulation curve (0- - -0) shown in Fig. 3H).

cantly. We have performed nuclei run-on experiments with nuclei from cells heat-shocked for periods up to 5 h. In these experiments, the level of HSP70 transcription increases through 1 h then decreases over a 5-h period to levels ap- proaching pre-heat shock transcription rates (data not shown). Together, these results reveal that transcriptional activation during heat shock is transient, although the cyto- plasmic level of HSP70 mRNA remains high. The transcrip- tion of an actin gene also appears to be transiently induced by heat shock. We do not understand the significance of this observation.

The kinetics of transcription of the HSP70 gene and accu- mulation of HSP70 mRNA reveals that the primary regula- tory control for HSP70 expression is at the level of transcrip- tion. The elevated synthesis of HSP70 as measured by [,'''SI methionine incorporation corresponds to an increase in the level of HSP70 mRNA. Continued incubation a t 45 "C reduces the level of both cytoplasmic actin RNA (Fig. 3) and actin synthesis. Thus, we conclude that the heat shock-induced increase in HSP70 synthesis and repression of actin synthesis reflects changes in t.he cytoplasmic levels of the respective mRNAs.

The Leveb of Actin and HSP70 Synthesis and Messenger RNA Levels during Heat Shock and Recovery at Normal Growth Temperatures-The results on transcription and ac- cumulation of HSP70 mRNA show that the initial burst in HSP70 synthesis is due to the rapid increase in HSP70 mRNA

Avian HSP70 Transcription 15743

levels. Thereafter, the continued and predominant synthesis of HSP70 appears to be due to the stability of HSP70 mRNA in the cytoplasm. The decreased synthesis of actin may in part be due to the turnover of actin mRNA. We examined the synthesis of HSP70 and actin and the levels of their corre- sponding cytoplasmic mRNAs under conditions of heat shock or during recovery from heat shock. MSB cells were heat- shocked for 2 h a t 45 "C, after which an aliquot was returned to 37 "C and the remaining cells were maintained at the heat shock temperature. Aliquots of cells were removed after var- ious periods of continued heat shock or recovery and pulse- labeled with [:"S]methionine, and the pattern of protein syn- thesis was examined by polyacrylamide gel electrophoresis. During continual exposure to 45 "C, synthesis of all three HSPs increases such that the synthesis of the three HSPs, 89, 70, and 23, corresponds to approximately 90% of total protein synthesis (Fig. 5 ) . During recovery, the synthesis of HSP89 and -70 decreases, while the synthesis of HSP23 increases 10-fold (Fig. 5 B ) . The increased synthesis of HSP23 during recovery parallels the overall increase in protein syn- thesis as cells return to the pre-heat shock state. These data show that the regulation of HSP23 synthesis during the recovery from heat shock is distinct from the regulation for HSP89 and -70. Whereas HSP89 and -70 are barely detected in cells recovered from heat shock for 8 h, HSP23 has become the major synthesized protein (Fig. 5R) . After 8 h a t 45 "C, the synthesis of control proteins is significantly repressed, whereas HSP synthesis continues at near maximal levels (Fig. 5). During a 4-h recovery a t 37 "C, HSP7O synthesis is de- creased 50%, and, after 8 h of recovery, HSP70 synthesis is no longer detected. The most striking effect of recovery is that the synthesis of control proteins, initially repressed by heat shock, increases rapidly-up 4-fold above non-heat shock levels after 8 h a t 37 "C.

We examined mRNA levels in heat shock and recovered cells to determine whether the decreased synthesis of HSP70 during recovery a t 37 "C was due to decreased levels of cyto- plasmic HSP70 mRNA, or due to translational repression of HSP70 mRNA. Total cytoplasmic mRNA from control, heat-

hours a? 45% C 2 6 8 1 0 -"

,

HSP 70 - 1

i ACTIN - 1

I

HSP23 -

recovery hours a? 45OC recovery 4 6 8 2 6 8 1 0 4 6 8

FIG. 5. HSP70 and act in synthesis during condit ions of heat shock at 45 "C and recovery at 37 "C. Cells were placed at 45 "C for 2 h, after which the culture was split and one-half was placed at 37 "C to recover and the other half was left for continued heat shock at 45 "C. At 0, 2, 6, and 8 h of recovery, an aliquot of cells was removed from each culture and labeled by ["S]methionine for anal- ysis by polyacrylamide gel electrophoresis and fluorography. A, 60-h exposure; H, 24-h exposure of the same fluorogram.

shocked, and recovered cells was isolated, absorbed on nitro- cellulose, and hybridized to the cloned chicken HSP70 gene (pC1.8) and the chicken P-actin gene (pAl) (Fig. 6). As previously shown, HSP70 mRNA accumulates rapidly during heat shock and remains a t high levels, corresponding with the increased synthesis of HSP70 protein (Figs. 3 and 6). Upon the shift back to normal growth temperatures, the level of HSP70 mRNA decreases rapidly (Fig. 6A). After 8 h a t 37 "C, recovered cells contain 15% of the maximal levels of HSP70 RNA. Thus, the decreased levels of HSP7O mRNA corre- sponds closely to the decreased synthesis of HSP7O during recovery.

The levels of actin mRNA and actin synthesis decrease during heat shock. However, following a shift to normal growth temperatures, the level of actin mRNA increases only slightly and actin protein synthesis increases significantly (Fig. 6). These results suggest that the pattern of protein synthesis during heat shock and recovery from heat shock reflects both the abundance of cytoplasmic mRNAs and the ability to translate these mRNAs.

DISCUSSION

Heat shock alters the pattern of protein synthesis in chicken lymphoblastoid cells (MSB) by the increased synthe- sis of three heat shock proteins (HSP89, -70, and -23) and the decreased synthesis of proteins normally synthesized a t 37 "C. The number and molecular sizes of the heat shock proteins correspond to the proteins induced in other avian

36 t /x\

k 2 4 6 8 IO

TIME OF INCUBATION (HRS)

(B) hours 01 45OC recovery

C 2 6 0 1 0 4 6 8 PC 1.8 ..e. 0 . 1

I , , I 1 I O I I

I I O I l l

PA^ e... 0 . . FIG. 6. HSP70 and act in mKNA levels during conditions of

heat shock and recovery. Cytoplasmic RNA was prepared from heat-shocked or recovered MSR cells as discussed in Fig. 5. A, quantitation of the RNA dot blots shown in H. HSP70 heat shock, X-X; HSP70 recovery, X- - -X; actin heat shock, o"--o, actin recovery, 0- - -0. R, dot blot of HSP50 and actin mRNAs in heat shock and recovered cells. Formamide-denatured RNA was spotted onto nitrocellulose and hybridized with ."P-labeled pC1.8 or pAl (%-h exposure for pC1.8, 4-h exposure for pAl probed filter).

15744 Avian HSP70 Transcription

cells (23-30). HSP89 and HSP70 are synthesized at low levels in cells maintained at normal growth temperatures, whereas HSP23 is detected only following heat shock. The kinetics and rates of synthesis differ for each HSP. The increased synthesis of heat shock proteins requires the transcription of the heat shock mRNAs (1-9, 23, 24,31,32). In this study, we have measured the kinetics of accumulation of HSP70 mRNA and show that increased levels are detected by 5 min after the temperature shift. The maximal level of HSP70 mRNA is obtained over a 2-hour period, after which the mRNA level remains high through 8 h of heat shock. The level of @-actin mRNA increases slightly following heat shock, then begins to decline after 6-8 h at 45 "C to 50-75% of preshock levels. Thus, changes in the pattern of protein synthesis after heat shock can be partially explained by changes in the relative levels of the corresponding cytoplasmic mRNAs (33). If com- ponents of the translational machinery are rate-limiting, the preferential translation of the newly transcribed heat shock mRNAs over other mRNAs may account for their increased stability by indirectly protecting the heat shock mRNAs against degradation. This mechanism could also render con- trol mRNAs susceptible to nucleases. Changes in the levels of cytoplasmic mRNAs can also be due to the generalized block in RNA polymerase I1 transcription (3-5) and the inability to process pre-mRNAs in the nuclei of heat-shocked cells (36). In Drosophila where the heat shock response is controlled at the level of transcription and translation (31), heat. shock mRNAs are preferentially translated over the control mRNAs, while control mRNAs are maintained in a translationally repressed state in heat-shocked cells. The mechanism for such preferential control of different classes of mRNAs remains poorly characterized (9, 34, 35).

We have shown that the rate of transcription of the chicken HSP70 gene is rapidly induced upon temperature elevation and attains maximal levels between 30 and 60 min of heat shock. With continued heat shock, the rate of transcription decreases to preshock levels, although the level of HSP70 mRNA remains high. This suggests that the level of HSP70 mRNA in the cytoplasm reaches a critical level, which directly or indirectly affects the rate of transcription of the HSP70 gene. Since HSP70 mRNA is more stable during heat shock as compared to the other messenger RNAs, continued high rates of HSP70 transcription are not necessary. Transient induction of HSP70 transcription has been observed in Esch- erichia colt (37) and Drosophila (4, 5 , 8, 34). In E. coli, the heat shock induction is transient; however, in dnaK (HSP70) mutants, the heat shock genes are induced and continue to be expressed at high levels (37). These results suggest a possible feedback between the dnaK product and transcrip- tional regulation of the bacterial heat shock response. A similar mechanism of autoregulation has been proposed for Drosophila HSP70 regulation (34). In amino acid analog- treated Drosophila cells, the increased transcription of the HSP70 genes is not curtailed, suggesting that proteins syn- thesized during heat shock may regulate HSP70 transcription. It, is clear from the studies by Wu (38,39) and Parker (40,41) that the heat shock transcription factors (HSTF or HAP) associate with the heat shock element following heat shock. However, there is no indication from these studies whether these effects are transient and if other factors interact with the HSP70 promoter to shut off transcription.

The mechanism that regulates the recovery of protein syn- thesis in heat-shocked cells returned to 37 "C appears to be opposite the regulation of the heat shock response. Upon return to the normal growth temperatures, the levels of HSP70 mRNA are reduced by 50% after 4 h relative to cells

L.

3.

4. 5. 6.

7.

8.

9. 10.

11. 12. 13.

14.

15.

kept at 45 "C and continue to drop during recovery. Interest- ingly, synthesis of HSP23 continues at even higher levels during the recovery period. Similar observations for HSP23 expression during recovery have been observed in Drosophila (34,42), chicken myoblast cells (21), and rat cells (43). Upon the return of heat-shocked cells to 37 "C, there is a rapid recovery in the level of @-actin synthesis, perhaps as a mech- anism to rapidly replenish the depletion of proteins whose synthesis was blocked during heat shock. Recently, it has been shown that heat shock blocks mRNA processing (36). Thus, the burst in actin protein synthesis, and the correspond- ing increase in the levels of actin mRNA in the cytoplasm, may reflect the rapid processing of the accumulated precursors of actin mRNA following a return to normal growth temper- atures. This suggests that eukaryotic cells may have mecha- nisms for the selective maintenance or degradation of certain messenger RNAs. Under the conditions of heat shock, HSP70 mRNA remains at a constant high level in the cytoplasm, although the rate of HSP70 gene transcription is very low. Then, upon return to normal growth temperatures, HSP70 mRNA degrades.

While expression of HSP7O is under transcriptional control following the shift to elevated temperatures, there appear to be other forms of control which determine the level of HSP70 and actin mRNAs. Post-transcriptional regulation may take the form of translational selection as discussed above, which allows for one class of mRNAs to be efficiently translated while the other class, no longer protected by its association with ribosomes, is subject to turnover. Such a mechanism acting early during the heat shock may result in the degra- dation of control mRNAs during heat shock. Alternatively, heat shock mRNAs may contain sequence information that regulates its half-life or translational selection. It is also possible that heat shock mRNAs are in a specialized cyto- plasmic compartment which dictates its translational selec- tion during heat shock or regulates its availability to degra- dation under various cellular states.

Acknowledgments-We thank N. G. Theodorakis and B. W u for their suggestions during these studies.

REFERENCES

1. Schlesinger, M. J., Ashburner, M., and Tissieres, A. (eds). (1982) Heat Shock: from Bacteria to Man, Cold Spring Harbor Labo- ratory, Cold Spring Harbor, NY

Craig, E. A. (1985) CRC Crit. Reu. Biochem. 18. 239-280 Spradling, A,', Pardue, M. L., and Penman, S. (1977) J. Mol. Biol.

Miller, D., and Elgin, S. A. (1981) Biochemistry 20,5033-5041 Findley, R. C., and Pederson, T. (1981) J. Cell Biol. 88, 323-328 McKenzie, S. L., Henikoff, S., and Meselson, M. S. (1975) Proc.

McKenzie, S. L., and Meselson, M. S. (1977) J. Mol. Biol. 177,

109,559-587

Natl. Acad. Sci. U. S. A. 72, 1117-1121

279-28.1 DiDomenico, B. J., Bugaisky, G. E., and Lindquist, S. (1982) Cell

_" 3 1,593-GO3

Ballinger, D. G., and Pardue, M. L. (1983) Cell 33, 103-114 Banerji, S. S., Theodorakis, N. G., and Morimoto, R. I. (1984)

Akiyama, Y., and Kato, S. (1974) Biken J . 17, 105-116 Laemmli, U. K. (1970) Nature 227, 680-685 Laskey, R. A,, and Mills, A. D. (1975) Eur. J . Biochem. 56, 335-

Morimoto, R. I., Hunt, C., Huang, S.-Y., Berg, K. L., and Banerji,

Mol. Cell. Biol. 4, 2437-2447

341

S. S. (1986) J. Biol. Chem. 261. 12692-12699 Cleveland, D. W., Lopata, M. A,, McDonald, R. J., Cowan, N. J.,

16. Thomas, P. (1980) Proc. Natl. Acad. Sci. U. S. A. 77, 5201-5205 17. Collins, P., Hightower, L., and Ball, L. (1978) J. Virol. 28, 324-

18. Aviv, H., and Leder, P. (1972) Proc. Natl. Acad. Sci. I/. S. A. 69,

Rutter, W. J., and Kirchner, M. W . (1980) Cell 20, 95-105

336

Avian HSP70 Transcription 15745

1408-1412 19. McMaster, G. K., and Carmichael, G. G. (1977) Proc. Natl. Acad.

20. Groudine, M., Peretz, M., and Weintraub, H. (1981) Mol. Cell.

21. Kafatos, F. C., Jones, C. W., and Efstradiadis, A. (1979) Nucleic

22. Dolan, M., Dodgson, J. B., and Engel, J . D. (1980) J. Biol. Chem.

23. Kelley, P. M., Aliperti, G., and Schlesinger, M. J. (1980) J. Biol.

24. Kelley, P. M., and Schlesinger, M. J. (1978) Cell 15, 1277-1286 25. Morimoto, R. I., and Fodor, E. (1984) J. Cell Biol. 99, 1316-1323 26. Wang, C., Gomer, R. H., and Lazarides, E. (1981) Proc. Natl.

27. Voellmy, R., and Bromley, P. A. (1982) Mol. Cell. Biol. 2, 479-

28. Voellmy, R., Bromely, P., and Hocher, H. P. (1983) J. Biol. Chem.

29. Bag, J. (1983) Eur. J . Biochem. 135, 187-196 30. White, C., and Hightower, L. (1984) Mol. Cell. Biol. 4, 1534-1541

Sci. U. S. A . 74,4835-4838

Biol. 1, 281-288

Acids Res. 7, 1541-1551

258,3983-3990

Chem. 255,3230-3233

Acad. Sci. U. S. A. 78, 3531-3535

483

258,3516-3522

31. Lindquist, S. (1980) Nature 293, 311-314 32. Wu, B. J., Hunt, C., and Morimoto, R. I. (1985) Mol. Cell. Bid.

33. Ray, B. K., Brendler, T. G., Adya, S., Daniels-McQueen, S., Miller, J . K., Hershey, J . W. B., Grifo, J . A., Merrick, W. C., and Thach, R. E. (1983) Proc. Natl. Acad. Sci. U. S. A. 80,

34. DeDomenico, B. J., Bugaisky, G. E., and Lindquist, S. (1982)

35. Scott, M. P., and Pardue, M. L. (1981) Proc. Natl. Acad. Sci. U.

36. Yosa, H. J., and Lindquist, S. (1986) Cell 45, 185-193 37. Tilly, K., McKittrick, N., Zylicz, M., and Georgopoulos, C. (1983)

38. Wu, C. (1980) Nature 306,854-860 39. Wu, C. (1984) Nature 309, 229-234 40. Parker, C. S., and Topol, J. (1984) Cell 36, 357-369 41. Parker, C. S., and Topol, J . (1984) Cell 39, 273-382 42. Lindquist, S. (1980) Deu. Biol. 77, 463-479 43. Welch, W. J., Garrels, J. I., Thomas, G. P., Lin, J. J.-C., and

5, 330-341

663-667

Proc. Natl. Acad. Sei. U. S. A. 79, 6181-6185

S. A. 78,3353-3357

Cell 34, 641-646

Feramisco, J. R. (1983) J. Biol. Chem. 258, 7102-7111