Transcriptional regulation of two genes specifically induced by ...

THEJOURNALOF BIOLOGICAL CHEMISTRY Vol. 258, No. 1, Issue of January 10, pp, 597-603, Printed IR U.S.A.

1983

Transcriptional Regulationof Two Genes Specifically Induced by Glucose Starvation in a Hamster Mutant Fibroblast Cell Line* (Received for publication, May 24, 1982)

Amy Shiu Lee+, Angelo M. Delegeane, Vicki Baker, andPeter C. Chow From the Departmentof Biochemistry, Uniuersity of Southern California Schoolof Medicine, Los Angeles, California 90033

This report concerns the characterization of the RNA transcripts encoded by two cDNA sequences p4A3 and p3C5, derived froma hamster temperature-sensitive mutant cell line K12. Using the two cDNA sequences as hybridizationprobes, we show that they occur as single copy genes in the hamster genome and encode for RNA transcripts which are highly inducible in K12 cells at 40.5 "C. Afterincubation at 40.5 "C for 16 h, there is a 10-fold increase in the p4A3 and p3C5mRNA levels, reaching a final concentration of about 1%of the cytoplasmic polyadenylated RNA. Wedemonstrate that the kinetics of transcription of p4A3and p3C5 directly parallel the accumulation of the mRNA levelsat 40.5 "C. Thus, our data indicate that the expression of these two genes are primarily regulated at the transcriptional level. In addition, there is a3- to 4-fold increase in the p4A3 and p3C5 mRNA levels when the cells are specifically starved of glucose. This implies that the expression of these two genes are stringently regulated by the availability of glucose in the culture medium.The relationship between the cDNA clones and two glucose-regulated proteins which are overproduced in K12 cells at 40.5 "C is discussed.

starved of glucose at 35 "C (Lee, 1981; Melero, 1981). These proteins arehighly conserved inanimal cells and aredifferent from the heat shock proteinspreviously described for animal cells (Lee et al., 1981). Investigation of the effect of actinomycin D on the94,000- and 78,000-dalton protein synthesized by K12 cells at 40.5 "C suggests that ts lesion is affecting new synthesis of the corresponding mRNA. Results obtained by in vitro translation of the mRNA isolated from K12 cells also imply that the synthesis of these proteinsis regulated at either the transcriptional or post-transcriptional level (Melero and Smith, 1978). The constructionof a cDNA library using RNA extracted from K12 cells incubated at 40.5 "C, as well as the identification of a cDNA clone, p3C5, possibly coding for the 78,000-dalton glucose-regulated protein, have been described (Lee et al., 1981). In the present study, we use two cDNA clones as hybridization probes to follow the induction of the corresponding genes in K12 cells when incubated a t 40.5 "C, and furnish evidence that the rapid accumulation of their mRNA transcripts is due to transcriptional regulation. In addition, we show that these two genesare specifically expressed when the cells are starvedof glucose. Thus, theK12 cells constitute an inducible system which may be useful for future identification of essential elements responsible for enhanced transcription of specific genes whose expression is stringently regulated by the availability of glucose in the culture medium.

The molecular mechanisms involved in the regulation of gene expression in mammalian cells remain elusive. Evidence from studieswith several gene systemssuggests that selective EXPERIMENTALPROCEDURES expression of a certain gene can be due to DNAamplification Cell Lines, Media, and Culture Conditions-The Chinese hamster (Brown and David, 1968; Alt et al., 1978; Wahl et al., 1979; fibroblast cell lines WglA and K12 have been described previously Beach and Palmiter,1981),to alternationsin the rate at which (Lee, 1981). The Chinese hamster ovary cell line was supplied to us et al., 1979, by Dr. R. Moran (University of Southern California). The cells are theDNAsequencesaretranscribed(Harpold Derman et al., 1981), or to a variety of post-transcriptional routinely maintained in DMEM' (Gibco Laboratories, Grand Island, and post-translational processing mechanisms (Roop et al., NY), supplemented with 10% cadet calf serum. To obtain synchronized cell cultures for RNA preparations, K12 cells were seeded at a 1978; Revel and Groner, 1978). In this report, we describe a mammalian mutant cell line density of 7 X IO3 cells per square centimeter in 150-mm diameter dishes and incubated at 35 "C in 25 ml of DMEM containing K12 which is a temperature-sensitive cell cycle mutant iso- culture 1 mg/ml of glucose and 10%calf serum. After 4 to 5 days of incubation lated from WglA, an established line of Chinese hamster without a change of the medium, most cells were synchronized in GI fibroblast (Roscoe et al., 1973). When K12 cells are incubated by serum deprivation. Upon addition of fresh medium, the arrested at the nonpermissive temperature (40.5 "C), the synthesis of cells were induced to proliferate. RNA samples were prepared from several cellular proteins are greatly enhanced. In particular, K12 cells at various times as they traversed through the cell cycle. the synthesis of two specific proteins of 94,000 and 78,000 Concomittantly, DNA synthesis was monitored by pulse labeling the cultures with 0.25 pCi/ml of [methyl-'Hlthymidine for 30 min, and daltons is increased 10- to 20-fold. We and others have further measuring the incorporation of labeled thymidine into trichloroacetic identified these proteins as glucose-regulated proteins in that acid precipitable material as described (Melero and Fincham, 1978). the same set of proteins is overproduced when the cells are Preparation of Nucleic Acids-High molecular weight DNA was * This investigation was supported by Public Health Service Grant No. CA-27607 awarded by the National Cancer Institute, Department of Health and HumanServices. The costs of publication of this article were defrayed in part by the payment of page charges. This art,icle musttherefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Recipient of a Junior Faculty Research Award from the American Cancer Society.

+

prepared by the method of Pellicer et al. (1978). For preparation of nuclear and cytoplasmic RNA, 1 to 4 X 10' cells were fractionated by lysis with 0.58 NP40 in isotonic high pH buffer (0.14 M NaCl, 0.01M Tris hydrochloride (pH 8.4), 0.0015 M MgC12). Total cytoplasmic RNA was extracted from the pooled NP40 wash supernatants (Harpold et al., 1979). To 7 ml of cell supernatant, the following were added at

' The abbreviations used are: DMEM, Dulbecco's modified Eagle's medium; SDS, sodium dodecyl sulfate; nt, nucleotide.

597

598

Transcription Control of Genes Induced by Glucose Starvation

4 "C: 1 ml of NETS buffer (0.1 M NaCI, 10 m~ Tris hydrochloride (pH 8.4), 1 mM EDTA, and 1%SDS);0.1 ml of 1 M Tris hydrochloride (pH 8.0); 0.2 ml of 0.5 M EDTA; 0.33 ml of 3 M NaCI; and 0.1 ml of 20% SDS. The final suspension was extracted 2 times with phenolchloroform (1:l) pre-equilibrated with NETS buffer. The aqueous phase was further extracted 2 times with chloroform, adjusted to 0.3 M NaOAc (pH 5.41, and the RNA was precipitated a t -20 "C by addition of 2 volumes of 95% ethanol. RNA preparations were routinely stored at -20 "C in 0.3M NaOAc, 67%ethanol. Aliquots were removed and centrifuged to recover the RNA pellet, which was resuspended in 10 mM Tris hydrochloride (pH 8.0), I mM EDTA. To prepare pulse-labeled nuclear RNA, K12 cells were seeded at cell density of 2 X lo4 cells/cm2 in 150-mm culture dishes and grown to about 90% confluency at 35 "C. After a change of fresh medium, the cells were incubated at either 35 or 40.5 "C. At various times after the temperature shift, the cells were labeled for 10 min with 0.75 to 1 mCi/ml of ['Hluridine (ICN, 45 Ci/mmol). The medium containing labeled uridine was prewanned to theappropriate temperature before addition to thecells. After the pulse period, the cells were immediately rinsed with several changes of ice-cold phosphate-buffered saline, scraped from the culture surfaces, and pelleted twice in cold phosphate-buffered saline. The cells were lysed with 0.5% NP40 in isotonic high pH buffer and nuclear RNA was extracted as previously described (Penman 1966; Soeiro and Darnell, 1969). RNA Gel Blots-The RNA gel blots were carried out according to a procedure developed by B. Seed and D. Goldberg.* The samples were run on a denaturing 1%agarose gel system containing 20 mM 3[N-morpholino]propanesulfonic acid, sodium salt, 5 mM NaOAc, 1 mM EDTA (pH 7.0), and 2.2 M deionized formaldehyde. Prior to loading, the cytoplasmic RNA samples (10-15 pg/lane) were denatured in a mixture of 50% deionized formamide, 2.2 M formaldehyde, and gel buffer by heating at 60 "C for 5 min. Bromophenol blue was added as a tracking dye. Electrophoresis was carried out at 15 V for 16 h a t room temperature with circulating electrophoresis buffer. Hamster 28 S and 18 S ribosomal RNA (4200 and 2000 nt) and Escherichia coli 23 S and 16 S ribosomal RNA (3000 and 1520 nt) were used as size standards. The size markers were visualized under UV light after soaking the gel in 5 X SSC (1 X SSC: 0.15 M NaCI, 0.015 M Na citrate) for 1 h and then staining with 1pg/ml of ethidium bromide. For blotting, the RNA gel was equilibrated in 20 X SSC for 20 min at room temperature, and the RNA was transferred onto nitrocellulose filters (Millipore) by the method of Southern (1975) except 20 X SSC was used as the transfer buffer. The transfer was usually carried out for 16 h. After baking under vacuum for 2-4 h at 80 "C, the fiiters were stored at 4 "C. In hybridization reactions, the filters were pretreated for 2-3 h at 42 "C in a mixture containing 10 X Denhardt (1 X Denhardt: 0.02% each of bovine serum albumin, PVP, Ficoll), 5 X SET (1 X S E T 0.15 M NaCl, 0.03 M Tris hydrochloride (pH 8.0), 2 mM EDTA), 0.05 M phosphate buffer (sodium phosphate buffer, pH 6.8), and 0.3% SDS. Thenthe RNA blot was prehybridized with a solution containing 50% deionized formamide, 5 x SET, 1 X Denhardt, 0.02 M PB, 0.1% SDS, and 50 p g / d of denatured salmon sperm DNA (Sigma, sonicated and phenol extracted) for 1 to 2 h at 42 "C. Labeled probes were denatured by boiling with salmon sperm DNA carrier in a small volume of hybridization buffer. Hybridization was carried out in the same solution as the prehybridization at 42 "C for 16 h. Filters were washed once at room temperature in 4 X SSC, 0.1% SDS, 0.1% Na pyrophosphate and 0.025 M PB; 3 times at 60 "C for 45 min each in 1 X SSC in the same buffer; and finally 2 times at room temperature for 30 min each in 0.3 x SSC in the same buffer. The filters were dried and autoradiographed. Kodak X-omat AR films were used and exposure was at -70 "C. Hybridization of RNA to Excess DNA-About 50-100 pg of plasmid DNA in 0.1 X SSC were boiled for 5 min in 0.1 N NaOH. After boiling, 15 volumes of 2 M NaCl was added and the denatured DNA was immediately loaded on a 25-mm Millipore HA nitrocellulose filter (0.45 pm, prewetted with water) placed inside a filter holder (Hoefer Scientific Instruments) which was attached to a water aspirator to provide a vacuum. The filters were rinsed in 10 ml of 6 X SSC, dried, and baked under vacuum at 80 "C for 3 h. The filters were stored at 4 " C . Blank filters used in the hybridization reactions to monitor background binding were treated identically except that DNA was omitted in the 0.1 X SSC solution. For each hybridization reaction described here, a quarter of the filter was sufficient to provide DNA excess. Just prior to hybridization, the filters were prehybridized at

' B. Seed and D. Goldberg, personal communications.

68 "C for 2 to 3 h in 0.5 ml of 2 X TESS (0.01 M 2-([2-hydroxy-l-lbis(hydroxymethyl)ethyl]amino)ethanesulfonic acid (pH 7.4), 0.3 M NaC1,O.Ol M EDTA, 0.2% SDS), containing 250 pg/ml of yeast RNA and 100 pg/ml of poly(rA) (Collaborative Research). Labeled RNA samples in 10 mM 2-([2-hydroxy-l-l-bis(hydroxymethyl)ethyl]amino) ethanesulfonic acid (pH 7.4) were denatured in 0.1 N NaOH for 10 at 4 "C. After denaturation, 1.5 vol of acid, pH 5.2, was 1 M 4-(2-hydroxyethyl)-l-piperazineethanesulfonic added. At this point, an equal volume of 4 X TESS was added, and the RNA samples were aliquoted to various hybridization reaction mixtures containing the DNA filters. Hybridization was performed at 65 "Cfor 40 h in capped glass scintillation vials with agitation. Following hybridization, the filters were washed 5 timeswith 10 ml of 2 X SCC at 65 "C, dried, and counted in 5 ml of toluene containing 0.4% (w/v) 2,5-diphenyloxazole.Under these conditions, the hybridization reaction was about 90% complete. RESULTS

Hamster cDNA Clones, p3C5 a n d p4A3"For the purpose of obtaining cloned probes which are preferentially expressed in K12 cells at 40.5 "C, a cDNAlibrary was constructed using poly(A+) RNA extracted fromK12 cells incubated at 40.5 "C (Lee et al.,1981). p3C5 and p4A3 were selected froma library of 500 cDNA clones because they preferentially hybridized with cDNA made from template RNA extracted from K12 cells a t 40.5 "C. Their restriction maps are shown in Fig. 1. Both inserts have been cloned into the BamHI site of the plasmidvector, pBR322, via thepoly(dA)poly(dT) tailing method (Zain et al., 1979). The sizes of the inserts for p3C5 2500 and 1400 nucleotides, and p4A3 are estimated to be respectively. By genomic blot analysis, we estimate that the gene number for p3C5 and p4A3 is one copy per haploid genome. There is no amplification nor rearrangement of the genes in WglA orK12 cells, incubated eitherat 35 or 40.5 "C." Gene Transcripts at Permissive a n d Nonpermissive Temperature-To follow the expression of the genes encoded by these two cDNA clones, cytoplasmic RNA extracted from K12 cells incubated either a t 40.5 "C or at 35 "C was size fractionated on denaturing formamide gels and hybridized with nick-translated p3C5 or p4A3. Increasing amounts of RNA were applied to each gel in order to quantitate the relative steady state concentrations of these transcripts a t both incubation temperatures. As shown in Fig. 2, both p3C5 and p4A3 hybridized very strongly to RNA extracted from K12 cells incubated a t 40.5 "C, although detectable amounts of the same transcripts were also observed with the 35 "C samples. By scanning the autoradiograms, we obtain peak areas linearly related to the amount of probe bound in each gel lane, which in turn is linearly related to the amount of RNA applied. By comparing the peak areas,we estimate that there is about a 10-fold increase in the steady state concentrations of both p3C5 and p4A3 transcripts in K12 cells after 16 h of incubation at 40.5 "C. In contrast, WglAcells at 40.5 "C show no increasein p4A3 mRNA levels and onlya 2-fold increase in p3C5 RNA levels.:' Size of Cytoplasmic Gene Transcripts-As shown in Fig. 2, p3C5 hybridized mainlyto an RNA species of 2700 nt, whereas p4A3 hybridized to RNA of 3200 nt. Upon overexposing of the autoradiographs of RNA blots, we have consistently observed other minor bands hybridizing with the probes. The sizes of these minor bands were about 3800 and 2100 nt for p3C5, and 4100 and 2000 nt for p4A3. It is possible that these RNA species are either unprocessed o r prematurely terminated products of the full length transcripts. Another possibility is that slight cross-hybridization of the labeled probe with other RNA sequences is unrelated to the major transcripts. The sharpnessof the hybridizing bands and the intact profile of the ribosomal RNA in the RNA samples would A. S. Lee, unpublished data.

Transcription Control of Genes Induced by Glucose Starvation 0

1

2

3

4

5

6

7

k

I

+

"P3C5 +

H;&

m

Sal I

..

m

Sa/ I

FIG.1. Partial restriction m a p s of cDNA clones p3C5 and p4A3. The cDNA insert (heavy line) was introduced into theunique BamHI site of pBR322. The methods used in the construction of the cDNA clones have been described (Lee et al., 1981).

p3c5

-

35"-

Insert Size

State poly

(ntlh

( A + )RNA'

(nt)"

-40.5'-

7 3 5 ' 7

3.8 ,

2.7, 2.1 '

Number of mole-

Steady

mRNA size

€COR I

by-b

Hindm Hind

-40.5'-

TABLE I Characterization of &A3 and ~ 3 C 5 mRNA Hamster cDNA clone

Eco R I

p4A3

b 1

Eco R I

EcoRI EcoRI

599

tules"

1 2250 3200 1400 p4A3 2700 2000 0.8 2550 p3c5 " The size of the insert (in nucleotides) was determined by digestion of the recombinant plasmid with restriction enzymes and electrophoresis on agarose gels (Lee et al., 1981). 'The size of the mRNAs were deduced from the gel blots as shown in Fig. 2, with ribosomal RNA as size markers. ' Poly(A') RNA was prepared and kinased in vitro by T4 polynucleotide kinase and hybridized to anexcess of plasmid DNA bound to Millipore filters. The fraction of total poly (A') RNA that remains bound to plasmid DNA after washing is expressed as percentage of steady state poly(A') RNA and calculated according to formula used by Harpold et al. (1979). The amount of radioactivity bound ranges from 825-4400 cpm above background binding, which is about 1 0 0 cpm. The values are average of three separate experiments. There is a linear relationship between the radioactivity bound and the amount of labeled RNA usedfor each hybridization, confvming that theDNA on the filter is in excess. These values represent the calculated number of specific mRNA molecules per cell, assuming 2.5 X 10" total mRNA/cell.

i A

B

FIG.2. Hybridization of p3C5 and p4A3 with hamster K12 cytoplasmic RNA. The probes used in these RNA gel blot hybridization experiments are p3C5 and p4A3 labeled by nick-translation to specific activity of about 5 X 10' cpm/pg. Total cytoplasmic RNA was extracted from K12 cells incubated at 35 "C or at 40.5 "C for 16 h. The amount of RNA (in micrograms) loaded per gel lane is indicated by numbers at the bottom of the autoradiogram. After hybridization with "'P-labeled plasmid DNA probes and extensive washing, the RNA blot was dried and autoradiographed. The sizes (in kilobases) of the most predominant transcripts as deduced from ribosomal RNA markers are indicated by arrows next to theautoradiograms. A shows hybridization pattern with p3C5 and B with p4A3. argue against random degradation during RNA extraction or gel electrophoresis. In either case, the fraction of the minor hybridizing species was at most 1%. In another experiment, we mixed the p3C5 and p4A3 probes and hybridized them to a RNA gel blot. The hybridization pattern obtained is a composite of that of the single probe results (data not shown). This result confms thenotion that p3C5 and p4A3 contain distinct nucleotide sequences and encode different RNA transcripts in K12 cells. Abundance of p3C5 and p4A3 Transcripts-We next estimated the number of RNA molecules hybridizable to p3C5 and p4A3 in the K12 cells after induction at 40.5 "C for 16 h. We fmt extracted total cytoplasmic RNA from K12 cells incubated at 40.5 "Cfor 16 h and thenisolated polyadenylated RNAby oligo(dT) column chromatography. The poly(A') RNA was uniformly labeled by in vitro kinasing of mildly nicked RNA. The labeled RNA was hybridized with excess p3C5 and p4A3 DNAbound to nitrocellulose filters. The fraction of the poly(A') RNA which hybridized to p3C5 and p4A3 sequences was determined from the amountof radioactivity bound to thefilter. Assuming that there were about 2.5 X 10% total mRNA moleculesper hamster cell (Harpold et al., 1979), the number of mRNA molecules complementary to

hr. 0

U

3

I4 0.5

1

2

3

7

13.5

0.5

1

2

3

7

13.5

v) c3

0

U

I/

hr. 0

it

g

-0

" "

0

2

3

I

1

""_"" 4

14

O-I

" " "

6

8 1 0 1 2

TIME ( hrs) FIG.3. Levels of RNA transcripts i n K12 cells. K12 cells were grown in 150-mm diameter culturedishes to about 90% confluency at 35 "C. After receiving a change of fresh medium, some cells were shifted to 40.5 "C, whereas other cells remained at 35 "C. At various times after the temperature shift (asindicated by numbers under the autoradiograms), total cytoplasmic RNA was extracted from the cells incubated either a t 35 or 40.5"C. 10 pg of each RNA sample was applied on denaturing RNA gel and, after electrophoresis, blotted onto nitrocellulose filter paper. The RNA gel blot was hybridized with '"P-labeled plasmid DNA probes prepared by nick translation. About 1 X 10' cpm of the probe was used for each hybridization, in a volume of about 10 ml.The hydridization patterns for p3C5 and p4A3 are shown. The autoradiograms were quantitated by densitometry to obtain the relative levels of these specific mRNA transcripts from K12 cells incubated a t 40.5 "C (M or) 35 "C (0- - -0).

600


A *

8

A

- G1 -IS-PG2+

0

35'

t

c

35' 40.5'

X

E

I

Q

2 4 n Lu

t-

a

[Y

0

& 4 0 v

p3c5

1

0

Z -

Lu

p4A3

7

0

x

I 0

2

4

6

12

14 16

18 19 2 0 2 2

16

16

TIME (hrs) I

4

I

1

1

1

1

8

12

16

20

24

HOURS POST STIMULATION FIG. 4. Gene transcript levels in synchronized K12 cells. K12 cells were synchronized by serum deprivation. Upon additionof fresh medium after 4 days, the K12 cells were induced to transverse the cell cycle. A, profile of of into trichloroacetic acid precipitable onset of DNA synthesis as monitored by incorporation[n~ethyZ-~H]thymidine material. B, the autoradiograms obtained from hybridization of p3C5 and p4A3 with cytoplasmic RNA at various points of the cell cycle. Total cytoplasmic RNA was extracted from synchronized K12 cells at various times after 10 pg of each RNA sample was applied to each gel lane and hybridization with the addition of fresh medium. About 32P-labeledp3C5 and p4A3 probes were as described in legend of Fig.3. The last two lanes contain RNA samples extracted from K12 cells which were incubatedat 35 "C, or had beenshifted to 40.5 "C for16 h.

p3C5 and p4A3 can be calculated (Table I). We estimated that after 16 h of incubation at 40.5 "C, the K12 cells accumulated close to 2000 mRNA molecules of p3C5 and p4A3 per cell. This represents about1%of the total poly(A+)RNA and falls into the category of moderately abundant mRNAs. Accordingly, it can be extrapolated that when the cells are incubated under normal growth conditions (35 "C), there are about 200 molecules of each of these transcripts percell. ut Accumulation of Cytoplasmic Gene Transcripts 40.5 "C-In order to follow the kinetics of accumulation of the p3C4 and p4A3 transcripts at 40.5"C,we prepared total cytoplasmic RNA from K12cells which had been shifted from 35 to 40.5 "C for various lengths of time. As a control, cytoplasmic RNA was also extracted from K12 cells which remained at 35 "C. Equal amounts of the RNA were applied onto denaturing formamide agarose gels, blotted, and hybridized with nick-translated p3C5 or p4A3. The autoradiograms are shown in Fig. 3. The level of hybridization at each time point was quantitated by scanning the autoradiograms. The relative peak area of the major hybridizing RNA species as a function of incubation time at 35 or 40.5 "C is shown inFig. 3. In both cases, there was a very rapid increase in the level of mRNA as the K12 cellswere shifted from 35 to 40.5"C. Within 4 h, the mRNA levelshad already reached the plateau level. At 35 "C, The mRNA levels remained constant. Consistent with earlier measurements presented in Fig. 2, there was a 10- to 15-fold increase in p3C5 and p4A3 mRNA levels at theplateau level. Gene Transcripts throughout the Cell Cycle-In addition to overproducing a specific set of proteins at 40.5 "C, K12was

originally characterized as a cell cycle mutant, that is, these cells become irreversibly blocked in mid-G1 when synchronizedcellswere incubated for longer than 5 h at 40.5 "C (Melero and Fincham, 1978). Therefore, it is of interest to examine the levels of the p3C5 and p4A3 transcripts throughout thecell cycle.We prepared synchronized K12 cells by the method of serum deprivation. Upon addition of fresh medium at the end of the fourth day, K12 cellswereinduced to proliferate. Duplicate dishes of the synchronized cells were pulse-labeled with [methyl-3H]thymidineto monitor the onset of DNA synthesis, and total cytoplasmic RNA was extracted from pardel cultures. Equal amounts of RNA were applied onto denaturing RNA gels, blotted, and hybridized with 32Plabeledp3C5 and p4A3. The results are shown in Fig.4. Transcripts for both p3C5 and p4A3 were present throughout the cell cycleand were detected in relatively constant amounts from late G1 through G2. However,when the cellswere blocked in G1 and during the 4 h immediately after the release, there was an increased level for p3C5 and p4A3 transcripts. Nonetheless, by 8 h, the concentration of the transcripts had returned to the normal level. There aretwo possibleexplanations for the higher levels of p3C5 and p4A3 mRNA observed in synchronized K12 cells in the early G1 phase. This may bedue to preferential transcription of these genes in early G1. Alternatively, these transcripts may have been synthesized while the cells were beingstarved to achieve synchronization and accumulated to high levels prior to change of fresh medium. Whenthe cells were released from serum deprivation, these transcripts turnedover and by 8 h, the normal level was restored.


601

Levels of Transcripts under Nutrient Starvation Conditions-To test the possibility that serum or glucose starvation may have caused the accumulation of p3C5 and p4A3 transcripts in the K12 cells, we performed the following experiment. Cytoplasmic RNA was extracted from 4 parallel cultures of K12 cells: l) cells grown at 35 "C with daily change of medium, 2) cells grown a t 35 "C without any medium change for 4 days but supplemented with freshglucose daily, 3) cells grown at 35 "C in glucose-free medium for 3 days, and4) cells grown at 40.5 "C with daily change of medium. Equal amounts of each RNA was applied on formaldehyde-formamide agaa b c d o b c d rose gels, blotted, and hybridized with nick-translated p3C5 FIG. 5. Effect of nutrient starvation on gene transcriptlev- or p4A3. The autoradiograms, as shown in Fig. 5, were quanels. K12 cells were seeded at 1 X IO4 cells/cm' in 150-mm diameter the and culture dish containing 25 ml of DMEM (4.5 mg/ml of glucose, Gibco) titated by densitometry. Our data indicated that p3C5 supplemented with 10% cadet calf serum. A t the end of the fourth p4A3 mRNA levels were 3- to 4-fold higher than the basal day, total cytoplasmic RNA wasextracted from parallel cellcultures level for cells starved of glucose. However, whenthe cells were grown under the following conditions. a,the cells were incubated at starved of all other nutrientsexcept glucose, the levels of the 35 "C, with a change of fresh, complete DMEM medium every day. two mRNA remained at basal level. This strongly suggests b, the cells were incubated at 35 "C for 4 days without anymedium change. At 24-h intervals, additionalglucose (4.5 mg/ml) was added that the levels of p3C5 and p4A3 transcripts are stringently to the culture medium. c, the cells were incubatedat 35 "C in glucose- regulated by the availability of glucose in the culturemedium. free medium for3 days. The cells received a change of fresh medium Furthermore, when the cells are incubated a t 40.5 "C in the every day, except the medium was completely devoid of glucose and presence of glucose and all other nutrients, by virtue of the the serum had been previously dialysed. d, the cells were incubated K12 ts mutation, p3C5 and p4A3 mRNA levels accumulated at 35 O C for 3 days, witha change of fresh, completeDMEM medium to about 10 times the basal level as observed in the previous every day. 16 h prior to extraction of RNA, the cells were changed to experiment (Figs. 2 and 3). complete DMEM (4.5 mg/ml of glucose) and shifted to 40.5 "C. Equal Relative Transcriptional Rates of the Gene Transcriptsamount of RNA was applied to agarose gels, blotted, and hybridized to p3C5 and p4A3as described in legend of Fig. 3.The autoradiograms It hasbeen demonstrated in Chinese hamster ovarycells and other systems that RNAlabeled by a very brief pulse repreare shown. sents primarily newly transcribed molecules. The relative transcriptional ratesof genes canbe compared by determining P3C5 the relative amountsof pulse label thatare incorporated into the gene transcripts (Harpold et al., 1979; Derman et al., 1981). Thus, to estimate the relative ratesof transcription of p3C5 and p4A3 at 35 and 40.5 "C, we pulse-labeled K12 cells with ["Hluridine for 10 min over a period of 8 h prior to extraction of nuclear RNA. The total amountof ['Hluridine incorporated per cell was the same irrespective of the incubation temperature, suggesting that general RNA synthesis was unaffected by incubation at 40.5 "C. T h e labeled RNA was hybridized with excess p3C5 and p4A3 bound tonitrocellulosefilter. The percentage of total pulse-labeled nuclear RNA hybridized to p3C5 and p4A3, as a function of the incubation time at 40.5 and 35 "C, is shown in Fig. 6. In both cases, there was a rapid increase in the transcription rate when the cells were shifted from 35 to 40.5 "C. By 3 h at 40.5 "C,there was about 15- to 20-fold increase in transcription rate for p3C5 and an 8-fold increase for p4A3. Since the amount of labeled RNA hybridizing to p4A3 in the 35 "C 1 1 2 3 4 5 6 7 8 control sample approaches the detection limit of this essay, TIME (hrs) we consider this to be a minimum estimate. FIG. 6. Relative rates of nuclear transcription. K12 cells incubated either at 35 or 40.5O C for various lengthsof time were pulseDISCUSSION labeled for 10 min with ['Hluridine. Total nuclear RNA was isolated The K12 cell system is both interesting and intriguing in at each time point and hybridized with p3C5 and p4A3 DNA bound on nitrocellulose Wters. To test for DNA excess in these reactions, that a temperature-sensitive mutation seems to causea varieach "H-labeled RNA sample was divided into three aliquots, con- ety of pleiotropic effects, including blocking the cells at G1, taining 1/6, 1/3, and 1/2 of the total radioactivity. Each aliquot was affecting synthesis of two enzymes related to DNA replication, hybridized with an identicalset of filters containinga fixed amount of and causing the overproduction of a distinct set of cellular DNA. After extensive washing, the amount of "H-labeled RNA bound proteins, in particular the 94,000- and 78,000-dalton glucoseper filter was determined. In all cases, there is a linear relationship between the amount of total cpm present in the reaction mixture and regulated proteins (Smith and Wigglesworth, 1973; Kit and the amount of cpm bound perfilter, confirming the presence of excess Jorgensen, 1976; Melero and Fincham, 1978; Lee, 1981). Animetallic DNA on the filters. Finally, in two cases where supernatant RNA was mal cells respond to adverse conditions such as heat, hybridized a second time with new DNA filters, the results indicated poisoning, or amino acid starvation by producing a set of that the hybridization was about 90%complete duringthe first round proteins generally known as the"heat-shock proteins" (Johnof hybridization. The backgroundhybridization to filters withno ston et al., 1980; Levinson et al., 1980; Hightower, 1980; Kelley DNA was about 20 cpm, while the experimental samples were from 50-650 cpm. The percentage of total radioactivity bound to p3C5 and and Schlesinger, 1982). However, when the cells are starved p4A3 was calculated and plotted against the time of incubation at with glucose, they induce a different set of proteins classified as "glucose-regulated proteins." The two majorandmost or 35 "C (0-- -0). either 40.5 "C1-(

SI

I

602

Transcription Control

of Genes Induced

by Glucose Starvation

commonlyobservedglucose-regulated proteins in human, levels of p3C5 and p4A3 mRNA are due to earlier synthesis mouse, rat, chicken, andhamsterhave molecular size of while the cells are starved of medium (including glucose) to 90-100,000 daltons and78,000 daltons (Pouysseguret al., 1977; achieve synchronization. Within 8 h after the addition of fresh McCormick et al., 1979; Lee, 1981; Melero, 1981). We believe medium, thelevels of the mRNA return to basal level. Neverthat thisspecific response of animal cells to glucose starvation theless, this issue can be best resolved by using a different is a common phenomenon and is important for cell survival method to obtain G1 cells, e.g. by mitotic shake off or by under this condition. The identification of a cell line, such as centrifugation elutriation,which does not involve glucose starK12, which overproduces these proteins and the possibility of vation and measuring directly the rate of synthesis of p3C5 isolating the correspondingcDNAclonescodingfor these and p4A3 mRNA in these cells. proteins will be extremely useful for physiological and molecUsing the two cDNA clonesp4A3 and p3C5 as hybridization ular studies. probes, we have demonstrated that the kinetics of transcripIn this report, we present evidence that the expression of tion of thetwo correspondinggenesdirectly parallelthe two genes, p4A3 and p3C5, is highly inducible in the hamster accumulation of specific mRNA levels in the K12 cells at ts mutant cell line K12. The expression of these genes caused 40.5 "C. This provides the strongest evidence that the two by the elevation in temperature is rapid and both the tran- genes are primarily regulated atthetranscriptional level. scription and accumulation of the RNA transcripts are maxi- What then is the molecular mechanism by which the K12 ts mal within a few hours. In addition,we show that theexpres- mutation blocks the cells at G1as well as inducing a 10- to 20sion of these twogenes arestringentlyregulated by the fold increase in transcription rate of these two genes? Our availability of glucose in the culture medium.A key question studies with a spontaneous revertant of K12 suggests that the then arises: what is the relationshipof these two cDNA clones K12 mutation exerts pleotropic effects on the cells and may to the 94,000- and 78,000-dalton glucose-regulated proteins? be of regulatory nature (Scharff et al., 1982). The processes Our previous work suggests thatp4A3 and p3C5 may contain involved in the regulated expression of these two glucosesequences coding for the94,000- and 78,000-dalton protein. In starvation specific genes are probably complex, but they are a previous report, we described using a hybrid-selection in amendable to present experimental techniques. By isolating vitro translation assay to identify the proteins coded for by the flanking sequences of these genes and using in vitro the cDNA clones (Lee et al., 1981). mRNA sequence comple- mutagenesis, it may be possible to identify DNA sequences mentary to the cDNA sequence was isolated by fiiterhybrid- which are important for the rapid and simultaneousincrease ization. Poly(A') RNA prepared from K12 cells incubated at in transcription of these genes. This approach, coupled with 40.5 "C washybridized to filter-bound plasmidDNA. The an in vitro transcriptional system which allows selective and hybridized mRNA was isolated by boiling and then translated accuratetranscription of eukaryotic genes (Manley et al., in a rabbit reticulocyte lysate cell-free translation system in 1980), may allow us to isolate and identify the regulatory the presence of ["'S]methionine. By two-dimensional gel elec- components interacting with the specific DNA sequence. trophoresis, we demonstrated that the p3C5 sequence selecAchnowledgments-We thank Drs. R. Stellwagen, R. E. K. Fourtively hybridized with amRNA species that codes for a protein similar in molecular size, as well as isoelectric focusing point, nier, and P.L. Lee for many helpful discussions and criticalreview of the manuscript. We are particularly grateful to Drs.M. Harpold and to the 78,000-dalton glucose-regulated protein. The in vivo T. Thomas for technical advice. 78,000-dalton proteinandthe in vitro translatedprotein yielded the same peptide maps. In the same way, p4A3 hyREFERENCES bridized to RNAwhich directs the synthesisof a protein with Alt, F. W., Kellems, R. E., Bertino, J. H., and Schimke, R. T. (1978) a molecular weight and isoelectric focusing point similar to J.Biol. Chem. 253, 1357-1370 those of the 94,000-dalton protein. Beach, L. R., and Palmiter, R. D. (1981) Proc. Natl. Acad. Sct. U. S. A . 78, 2110-2114 Our tentative conclusion is substantiated by our present results. (i) Thesize of transcripts encoded by p4A3 and p3C5 Brown, D. D., and David, I. B. (1968) Science 160,272-280 are 3200 and 2700 nt, respectively. These sizes are consistent Derman, L., Krauter, K., Walling, L., Weinberger, C., Ray, M., and Darnell, J. E., Jr. (1981) Cell 23, 731-739 for mRNAs coding for proteins of 94,000 and 78,000 daltons, Harpold, M. M., Evans, R. M., Salditt-Georgieff,M., and Darnell, J. respectively. (ii) The relative levels of the mRNAof p4A3 and E. (1979) Cell 17, 1025-1035 p3C5 present in K12 cells when the cells are starvedof glucose Hightower, L. E. (1980) J. Cell Physiol. 102, 407-427 or incubated at 40.5 "C directly parallel that of the levels of Johnston, D., Oppermann, H., Jackson, J., and Levinson, W. (1980) J. Biol. Chern. 255, 6975-6980 the 94,000- and 78,000-dalton glucose-regulated proteins accumulated in K12 cells under identical conditions (Lee, 1981; Kelley, P.M., and Schlesinger,M. J. (1982)Mol. Cell.Biol. 2,267-274 Scharff et al., 1982). (iii) In the revertant cell line where the Kit, S., and Jorgensen, G. N. (1981) J. Cell Physiol. 8 8 , 5 7 4 4 Lee, A. S. (1981) J. Cell Physiol. 106, 119-125 94,000 and 78,000 proteins were no longer overproduced at Lee, A. S., Delegeane, A., and Scharff, D. (1981) Proc. Natl. Acad. 40.5 "C (Scharff et al., 1982),we found that thelevels of p3C5 Sci. U. S. A. 78, 4922-4925 and p4A3 were also reverted to the basal level observed inthe Levinson, W., Oppermann, H., and Jackson, J. (1980) Biochim. Biononmutant parental cells.:' While the definitive answer to our phys. Acta 606, 170-180 question must await the direct comparison of the amino acid Manley, J. L., Fire, A., Cano, A.,Sharp, P. A., and Gafter,M. L. (1980) Proc. Natl. Acad.Sci. U. S. A . 77, 3855-3859 sequence of the proteins with the plasmid DNA sequence, McCormick, P. J., Keys, B. J., Pucci, C., and Millis, A. J. T. (1979) there is strong evidence that p4A3 and p3C5 contain sequences Cell 18, 173-182 coding for the 94,000- and 78,000-dalton protein. Melero, J. A. (1981)J. Cell Physiol. 109, 59-67 Since K12 is a cell cycle mutant blocked in mid-Gl when Melero, J. A,, and Fincham, V. (1978) J. Cell Physiol. 95,295-306 incubated at 40.5 "C, it is feasible that mRNAs which are Melero, J. A,, and Smith, A. E. (1978) Nature (Lond.)272, 725-727 overproduced at 40.5 "C are GI-specific transcripts. We have Pellicer, A.,Wigler,M., Axel, R., and Silverstein, S.(1978) Cell 14, 133-141 examined this issue by extracting RNA from synchronized Penman, S.(1966) J . Mol. Biol. 17, 117-130 K12 cells at various points of the cell cycle and analyzed the Pouyssegur, J., Shiu, R. P. C., and Pastan, I. (1977) Cell 11, 941-947 concentration of p4A3 and p3C5 throughout the cell cycle. Revel, M., and Groner, Y. (1978)Annu. Reu. Biochem. 47, 1079-1126 Although we observe a higher level of these mRNAsin starved Roop, D. R., Nordstrom J. L., Tsai, S . Y., Tsai, M. J., and O'Malley, and early G1 cells, we favor the interpretation that the higher B. W. (1978) Cell 15,671-685

Transcription Control

of Genes Induced

Roscoe, D. H., Read, M., and Robinson, H. (1973) J. Cell. Physiol82, 325-331 Scharff, D. J., Delegeane, A. M., and Lee, A. S. (1982) J. Cell Biol. 92,629-633 Smith, B. J., andWigglesworth, N. W. (1973) J . Cell PhysioE. 82, 339-347

by Glucose Starvation

603

Soeiro, R., and Darnell, J. E. (1969) J. Mol. Biol. 44, 551-562 Southern, E. M.Biol. (1975)J. Mol. 98,503-517 Wahl, G. M., Padgett, R. A,, and Stark, G. R. (1979) J. Biol. Chern. 254, 8679-8689 Zain, S., Sambrook, J., Roberts, R. J., Keller, W., Fried, M., and Dunn, A. R.(1979) Cell 16,851-861