Biochimica et Biophysica Acta, 1090 (1991) 333-342 © 1991 Elsevier Science Publishers B.V. All rights reserved 0167-4781/91/$03.50

333

BBAEXP 92311

Developmental expression of glycogenolytic enzymes in rabbit tissues: possible relationship to fetal lung maturation Christopher B. Newgard !,2, Brian Norkiewicz 1,2, Steven D. Hughes 1,2, Rene A. Frenkel l, Ward S. Coats 1,2, Frank Martiniuk 4 and John M. Johnston 1,3 t Department of Biochonistry, " Clifford Laboratories, Center for Diabetes Research, .1Department of Obstetrics / Gynecology & Cecil H. and Ida Green Center for Reproductice Biology, Unicersity of Texas Southwestern Medical Center at Dallas, Dallas, TX (U.S.A.) and .t Department of Medicb~e, New York Unicersity School of Medicble, New York, NY (U.S.A.)

(Received 2 July 1991)

Key words: Glycogenolytic cm.yme; Gene expression; Lung maturation; (Rabbit)

Glycogen can be degraded in mammalian tissues by one of three iso~mes of glycogen phosphorylase, termed muscle (M), liver (L) and brain (B) after the tissues in which they are preferentially expressed iii adult animals, or by members of the family of a-glucosidases. In the current study, we have examined the developmental expression of these enzymes and their respective mRNAs in rabbit tissues, with particular emphasis on the developing lung, a tissue in which glycogen serves as an important source of carbon for surfactant phospholipid biosynthesis. Native gel activity assays and RNA blot hybridization analysis revealed that the B isoform of glycogen phosphorylase predominates in fetal and adult lung t~ssues, accompanied by a low level of expression of the M isoform. Total B and M phosphorylase activities increased during fetal lung development, with a peak at day 28 of gestation, then decreased to the adult level at term. This peak in activity coincided with the peak period of glycogen degradation in developing lung. While the increase in M isozyme activity was correlated with an increase in the level of its mRNA, B isoform mRNA showed no significant alteration during development, suggesting that the increase in B isoform activity is determined by a posttranscriptional mechanism. Analysis of phosphorylase mRNA levels in developing liver, skeletal muscle, brain and heart revealed a diverse expression pattern. The L isozyme mRNA was predominant at all time points in liver, the M isozyme was predominant at all time points in muscle, the B isozyme was predominant at all time points in brain, and heart contained a mixture of B and M mRNA in roughly equal ratios at all time points. Thus, our studies of phosphorylase mRNA in the rabbit provide no evidence for genera| predominance of the B isozyme in fetal tissues, or for isozyme 'switching' from the B to the L or M forms during development, as has been suggested by others. In addition to the increase in phosphorylase activity, acid, but not neutral a-glucosidase activity was found to increase significantly during fetal lung development, again with a peak at day 28 of gestation. Interestingly, RNA blot hybridization analysis with a probe for lysosomal a-glucosidase revealed no change in the level of expression of its 4 kb transcript in developing lung. Instead, we observed induction of a structurally related mRNA of 7.4 kb that peaked at day 28 of gestation. Hybridization with a sucrase/isomaltase-specific oligonucleotide excluded the possibility that the 7.4 kb transcript encodes this protein. The identity of the latye glucosidase-like transcript and its potential role in fetal lung glycogenolysis thus remains to be established. Introduction

Glycogen is the storage form of glucose, and in addition represents an important source of carbon for

Correspondence: C.B. Newgard, Department of Biochemistry & Gifford Laboratories, Center for Diabetes Research, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., Dallas, TX 75235, U.S.A.

energy production and the biosynthesis of a variety of compounds, including fetal lung surfactant. Cytosolic glycogen is broken down by one of the three isozymes of glycogen phosphorylase, termed muscle (M), liver (L) or brain (B) after the tissues in which they are preferentially expressed in adult mammals (reviewed in Ref. 1). Although previous studies have implied that the B isozyme is predominant in rat fetal tissues [2-9], this has not been carefully investigated in other species or at the mRNA level.

334 Glycogen can also be degraded by members of the family of a-glucosidases. Mammalian tissues have been shown to contain lysosomal acid a-glucosidase as well as at least two isoforms of neutral a-glucosidases, termed neutral glucosidases AB and C, that differ in substrate specificity, molecular weight and capacity to bind to the lectin concanavalin A, and that have been shown by genetic studies to be the products of distinct genes [10-12]. In addition, lysosomal a-glucosidase and intestinal sucrase/isomaltase are structurally related (approx. 42% sequence identity [13-15]), suggesting that the number of proteins with structural and/or functional similarity to a-glucosidase may be large, and may include variants that are as yet uncharacterized. in lung, three forms of glucosidase appear to be present [16]. Lamellar bodies contain two acid a-glucosidases, one that appears identical to the lysosomal form, and one that fails to bind to concanavalin A and is thus distinct from the lysosomal form. The former, but not the latter, is absent from lung samples taken from an individual with type II glycogen storage disorder, or Pompe's disease. In addition, microsomal fractions from lung appear to contain a neutral a-glucosidase activity which fails to bind to concanavalin A but that has not been further characterized as type AB or C [16]. Acarbose, an inhibitor of lysosomal a-glucosidase, has been shown to inhibit glycogenolysis by 40% in rat fetal lung explants [17]; the remaining activity is presumably due to phosphorylase and/or the alternate glucosidase activities described above. cDNA clones encoding the three mammalian glycogen phosphorylase isozymes [1,18-23], as well as the lysosomal acid a-glucosidase [ 14,24] have recently been isolated and characterized. In the present study, we have utilized these probes as well as assays of enzymatic activities to address the following questions: (1) what is the expression pattern of phosphorylase iso~me and lysosomal a-glucosidase in developing and adult rabbit tissues?; (2) is the burst of glycogenolysis that contributes carbon for surfactant biosynthesis in fetal lung associated with induced expression of any particular glycogen degrading enzyme? Materials and Methods

Animals and tissue isolation Pregnant female rabbits were suppaed 10 days after fertilization from Myrtle Rabbitry, Thompson Station, TN. Tissues were isolated from fetal animals at interpals between days 21 and 31 of gestation by anesthetizing the mother with nembutal, surgically removing the fetal animals and killing them by decapitation. Lung, liver, hind-leg skeletal muscle, brain and heart samples were rapidly excised and plunged into liquid nitrogen, then stored at - 8 0 ° C until processed. The entire procedure was performed in less than 1 rain for each

animal. Tissue samples were also obtained from newborn animals (less than 1 h old) and from adult female animals. Two independent litters (8-14 animals per litter) were analyzed for mRNA levels, glycogen content, and glycogenolytic enzyme activity at each time point and for each of the tissues referred to above.

Glycogen assay The procedure used was essentially that of Chan and Exton [25]. Briefly, 10% tissue homogenates were prepared in 5% trichloroacetic acid and spotted onto filter paper discs (grade, 3, Whatman Limited, U.K.). The filters were washed three times with 66% EtOH and allowed to dry. Glycogen precipitated onto the papers was digested to free glucose with 0.15 U/ml amyloglucosidase (Sigma) in 50 mM sodium acetate (pH 4.8). Glycogen-glucose was then assayed using a colorimetric kit (Sigma).

RNA isolation, probe preparation and Northern blot hybridization analysis Poly(A) + RNA was prepared by the method of Ashley and MacDonald [26], resolved on formaldehyde agarose gels and transferred to MSI nylon membranes as previously described [20]. Portions of the cDNAs encoding human liver glycogen phosphorylase [20], human brain glycogen phosphorylase [23], rabbit muscle phosphorylase [19], human lysosomal a-glucosidase [24] and human /3-actin [27] were subcloned into pGEM vectors (Promega) in order to be able to generate antisense RNA probes of high specific activity [28]. The specific fragments and pGEM vectors used were as follows: (1) the liver phosphorylase probe is a 1.42 kb EcoRl/Pst fragment, stretching from nucleotides 661 to 2085 (using the numbering system in Ref. 20), subcloned in pGEM 4Z; (2) the brain phosphorylase probe is a 1.24 kb Pst/Pst fragment, stretching from nucleotides 850 to 2090 (using the numbering system in Ref. 23), subcloned in pGEM 3Zf + ; (3) the muscle phosphorylase probe is a 1.3 kb Acc/HindIII fragment, stretching from nucleotides 1340 to 2634 (using the numbering system in Ref. 19), subcloned in pGEM 3Zf + ; (4) the lysosomal acid a-glucosidase probe is a 1.43 Sac l/Sacl fragment, stretching from nucleotides 200 to 1630 subcloned in pGEM 4Z (the numbering system is based on a larger, unpublished cDNA sequence than that previously reported by the Hirschhorn group in Ref. 24); (5) the fl-actin probe is a 410 bp Smal/Hpall fragment subcloned in pGEM 4Z [27]. For low stringency hybridization experiments, the purified 1.43 kb a-glucosidase insert was nick-translated and hybridized in 35% formamide hybridization solution, as previously described [20]. Probes were prepared by linearizing the vector with an appropriate restriction endonuclease and transcribing off the SP6 or T7 promoters of the pGEM vectors,

335 depending on which produces the antisense transcript, in the presence of [32 P]cytidine 5'-a-triphosphate (3000 Ci/mmol, ICN) [28]. All probes were prepared with specific activities of approx. 1 • l0 '~ cpm/ttg DNA. For each hybridization experiment, 5- 10~' cpm/ml of probe was added in high stringency hybridization buffer (consisting of 50% formamide, 5 × SSC, 50 mM NaH2PO 4 (pH 7.0), 10 × Denhardt's, 250 ~ g / m l salmon sperm DNA and 1% sodium dodecyl sulfate (SDS)) and incubated at a temperature of 65 °C for a period of 20 h. The blot was then washed four times in 0.1 x SSC, 0.1% SDS at 65°C for 20 min each and exposed to X-ray film for 1, 5 and then 17 h in the presence of an intensifier screen. Each blot was hybridized sequentially with all of the probes described above; blots were prepared for each new hybridization by stripping with two washes in 0.5 x SSC, 0.1% SDS at 90°C for 45 rain each. The blots were checked by autoradiography for complete removal of one probe before hybridization with the next. Quantitation of hybridization signals was performed using a Hoefer GS 300 scanning densitometer. The data were expressed as the ratio of the signal for the experimental probe to that obtained with the /3-actin probe. RNA size markers (Bethesda Research Laboratories; 0.24 to 9.5 kilobases) were included on all gels to allow sizing of particular mRNA species. The sucrase/isomaltase antisense oligonucleotide 5'-GCTGCAGGTCCAGCAAATI'CTGTCACAAATI'GATG-3' was synthesized using an Applied Biosystems 381A DNA synthesizer, and was prepared as a hybridization probe by end labeling with [),..a_,p] ATP and polynucleotide kinase. Oligonucleotide hybridization was carried out at a temperature of 50 °C in a buffer consisting of 6 x SSC and l0 x Denhardt's. Washing was in 2 x SSC, 0.1% SDS for 30 rain each at 50 o C and 60 ° C, respectively.

mM /3-glycerophosphate, 25 mM DTI', 5 mM EDTA, pH 6.8); buffer B, same as buffer A, but with 5 mM AMP added; buffer C, same as buffer B, but with 0.7 M Na2SO 4 added. The migration position of the various phosphorylase isoforms in ihe gel was visualized by staining newly formed glyco~e.n wilh an iodine solution (0.4% KI, 0.2% I,). Acid and neutral oe-glucosidase activities were determined in tissue homogenates prepared in 0.25 M sucrose, 20 mM Mops, pH 7.5. Reactions contained 20 /zl of a 5 mM solution of 4-methylumbelliferyl-oe-D-glucoside (Sigma) and 160 t~! of either 100 mM Tris-HC! (pH 7.5) (neutral a-glucosidase activity) or 100 mM pnta~sium acetate-acetic acid (pH 4.5) (acid a-glucosidase activity). Reactions were initiated by addition of 20 #i of appropriate dilutions of the tissue homogenate, and after incubation at 37 °C for 15 and 30 rain, reactions were stopped by the addition of 0.4 ml of 95% ethanol. The samples were then centrifuged at full speed in a Brinkman microcentrifuge for 5 min and aliquots of the supernatant were added to fluorometer tubes containing 2 ml of distilled water and 0.2 ml of 2 M K2CO 3. Fluorescence was measured with a Hoeffer fluorometer, using methylumbelliferone for the preparation of daily standards. Total protein was determined by the micro BCA (bicinchoninic acid) method, using a protocol and reagents supplied by Pierce (Rockford, IL). Results

Changes in tissue glycogen content during decelopnlent Glycogen content was measured in lung. heart, brain, skeletal muscle and liver taken from fetal, newborn A

o

20 4-

m m

Enzynlatic acticity assays Relative levels of glycogen phosphorylase enzymatic activity and identification of the protein isoform were determined by a native gel activity assay [9,29,30]. Briefly, tissue samples were homogenized with a polytron homogenizer in 20 volumes of ice-cold 5 mM EDTA, 75 mM NaF (pH 7.2), containing 0.2 mM phenyimethylsuifonyl fluoride, and centrifuged at 14000 × g for 30 min. After measurement of total protein concentration by the method of Bradford [31], using a kit and reagents from Sigma, 175 ~g of supernatant protein were loaded onto each of three identically prepared 5% acrylamide, 0.1% glycogen gels and electrophoresed for 6 h at 150 V, using a Tris-barbital running buffer (1 mM Tris, 26.5 mM barbital, pH 7.0). The gels were then soaked overnight in one of three reaction buffers: buffer A, 0.1% glycogen, 75 mM glucose-l-phosphate, in/3-glycerophosphate buffer (50

Oo

E 10

c o



o

o O

20

25

30

35

Days of Gestation

Adu!i Values

Newborn Fig. I. Glycogen content in fetal, newborn, and adult rabbit tissues. Tissues were sampled from litters consisting of 6-12 animals and extracted for glycogen assay as described in Materials and Methods. Adult values refers to glycogen content in tissue samples taken from the mother of the day 31 fetal animals. El, lung: 0, heart; zx brain; o, muscle; II, liver.

336 and adult animals, as shown in Fig. 1. In agreement with previous studies [32-37], we found the most dramatic changes in glycogen content to occur in liver and lung. In liver, glycogen was barely detectable at days 21 and 24 of gestation, but was seen to rise to a level of around 15 m g / g wet weight of tissue at day 28, staying at this high level until day 31, which is close to term. In the newborn animals, glycogen decreased to almost zero, but in the adult animal was again present at high levels (22 rag/g). This pattern is consistent with the need for deposition of glycogen stores late in fetal development in order to maintain circulating glucose levels in the newborn animal until effective feeding commences. In contrast, substantial glycogen was present at day 21 in lung (9 rag/g) and a further rise to a peak value of 16 mg/g was observed at day 24 of gestation. Thereafter, lung glycogen fell to a value of 1 mg/g at day 28, and then remained at this low level

through the remainder of fetal development and into the newborn and adult stages. The precipitous fall in glycogen levels is well correlated with the onset of surfactant production at day 27 in the fetal rabbit lung [37]. Glycogen was present at levels _< 1 mg/g at all time points in brain, and was generally present at a level of = 5 mg/g in heart and skeletal muscle.

Expression of mRNAs encoding glycogenolytic enzymes in developing lung Fig. 2 characterizes the expression of mRNAs encoding the glycogenolytic enzymes studied in developing lung. Fig. 2A shows the data obtained with the various antisense RNA probes. Fig. 2B depicts hybridization with the 1.43 kb Sacl/Sacl a-glucosidase fragment. In contrast to the hybridzations shown in Fig. 2A, the a-glucosidasc experiment in Fig. 2B was done at low stringency (35% formamide hybridization

Lung RNA Probes

1

2

3

4

5

6 Probes

Brain GP

I

1

2

3

1

2

3

4

5

6

7

8

(x Glu 28S - -

Liver GP

18S - -

d. x

Vm

LU

Muscle GP

Actin ecGLU

I

I

Actin ¢q

1

2

3

(x Glu

,4 X

u,l

Brain GP

Actin

I

I

Fig. 2. Expression of glycogen phosphorylase and a-glucosidase mRNA in fetal, newborn and adult lung. (A) A blot containing samples of lung RNA isolated from animals of different gestational ages was hybridized with the indicated antisense RNA probes at high stringency, as described in Materials and Methods. Two independent experiments were performed with the brain phosphorylase probe on separate RNA samples prepared from different litters (experiments 1 and 2). Each lane was loaded with 5/zg of poly(A) + RNA. Experiment 1: lane 1, day 21 of gestation; lane 2, day 24 of gestation; lane 3, day 28 of gestation; lane 4, day 31 of gestation; lane 5, newborn animal; lane 6, adult animal. Experiment 2: lane I, day 24 of gestation, lane 2, day 28 of gestation, lane 3, day 31 of gestation. (B) Blots were hybridized with a nick-translated 1.43 kb cDNA probe encoding lysosomal a-glucosidasc at low stringency, or with the antisense actin probe, as described in Materials and Methods. Experiment 1: lane 1, day 21 of gestation; lane 2, day 24 of gestation; lane 3, day 28 of gestation; lane 4, day 31 of gestation; lane 5, newborn; lane 6, adult lung; lane 7, adult lung; lane 8, adult brain. Experiment 2 (all lung samples): lane 1, day 21 of gestation; lane 2, day 28 of gestation; lane 3, adult. Exposure times (in the presence of an intensifier screen) were 5 h for actin and 17 h for all other probes. The reduced actin signal in lane 4 of experiment 1 is due to inefficient blot transfer in this region of the gel and not to unequal loading; to underscore this point, we show the ethidium bromide stained RNA after transfer to the blotting membrane in experiment 2 instead of actin hybridization.

337

buffer, 42°C) and using the eDNA fragment labeled with 3Zp by nick-translation, instead of the RNA probe, which in our hands was not useful for low-stringency screening. The critical observations contained in these data are as follows: (1) the B isoform phosphorylase mRNA is predominantly expressed at all developmental time points in lung, accompanied by a low level of expression of the M isoform and essentially no detectable expression of the L isoform. The mRNA encoding the B isozyme of phosphorylase has a size of 4.2 kb, which is significantly larger than the L (3.2 kb) or M (3.4 kb) isoforms, allowing unequivocal identification of the three transcripts [20,23]. (2) In two independent experiments, no change was noted in the levels of the B isoform mRNA during development (Fig. 2A, experiments 1 and 2). In particular, no change in expression is evident between days 24 and 28 of gestation, the time intervals between which glycogenolysis is activated and surfactant is produced. A small increase in the level of the M isoform mRNA was noted, with a peak at day 28. (3) An a-glucosidase mRNA of approx. 4 kb is detected at a low constant level at all developmental time points, using both the antisense RNA probe at high stringency and the cDNA probe at low stringency. A much larger message is also detected, but only under low stringency conditions (Fig. 2B). The band is more difuse than other mRNA species detected in our analysis; the center of the band runs at a position consistent with a size of approx. 7.4 kb. (4) Quantitation by densitometric scanning and normalization to the actin signal for the data in Fig. 2B, experiment I, reveals that the large glucosidase-related mRNA is increased at day 24 of gestation and reaches a peak at day 28, where it is 10-fold higher than at day 21 and 2.2-fold higher than in the adult. In a separate experiment, we present the ethidium bromide staining of the mRNA instead of actin hybridization, and confirm the increased expression of the 7.4 kb mRNA at day 28 relative to day 21 and adult lung (Fig. 2B, experiment 2). This induction correlates well with the time course of increased glycogenolysis in lung as shown in Fig. 1. Further experiments revealed that the 7.4 kb band is found in heart and brain in addition to lung, but not in liver or muscle (Fig. 2B and data not shown). Two recent papers report a remarkable 40% sequence identity between lysosomal a-glucosidase and intestinal sucrase/isomaltase [14,15]. The latter is a large protein (= 260000 Da) containing both sucrase and isomaltase active sites encoded by a single mRNA estimated to have a size of 6 kb [13]. To test the possibility that the large transcript observed in developing lung might be in fact be sucrase/isomaltase, we prepared a 38 residue antisense oligonucleotide, corresponding to a region of the sucrase/isomaltase eDNA sequence with no homology to a-glucosidase (nucleotides 558 through 520, using the numbering system and

0

c~

e,-

~

......................

4.4 K B 2.41.4-

Fig. 3. The large a-glucosidase-related mRNA in lung does not encode sucrase/isomaltase. 5 /~g of poly(A) + RNA isolated from adult intestinal mucosa (int. mucosa) or lung was probed with an antisense oligonucleotide derived from a region of the sucrase/ isomaltase sequence that is not related to a-glucosidase, as described in Materials and Methods. This particular figure represents a 6 h film exposure time in the presence of an intensifier screen. Longer exposure times (24 or 48 h) also failed to generate any signal in the lung RNA lane (data not shown).

the sequence provided in Ref. 13). This oligonucleotide was hybridized to day 28 lung RNA, and as a positive conti31, to RNA isolated from adult intestinal mucosa. As shown in Fig. 3, the probe hybridizes to the expected 6.0 kb message in mRNA isolated from rabbit intestinal tissue, but fails to hybridize to any large transcript in lung. Thus, we conclude that the large glucosidase-like transcript in lung does not encode sucrase/isomaltase.

Activities of glycogenolytic enzymes in developing tissues Fig. 4 shows an analysis of phosphorylase isoform expression by a native gel activity assay system. Phosphorylase isoforms can be identified in this system both on the basis of their reponsiveness to AMP and by their relative mobilities; incubations with 0.7 M Na2SO 4 give an estimate of total phosphorylase activity in a sample due to potent activation of all three phosphorylase b isoforms [9,29,30]. In this system, purified rabbit muscle phosphorylase a (Sigma) migrates relatively slowly and is active under all three assay conditions, including in the absence of AMP. Purified muscle phosphorylase b (Sigma) also migrates slowly, but differs from phosphorylase a in that it is inactive in the absence of AMP, but activated to near maximal activity by AMP alone. A single band of phosphorylase activity that migrates faster than purified muscle phosphorylase is seen in the rabbit liver extract; this activity is poorly activated by AMP, as has also been shown for the human liver phosphorylase protein expressed from its eDNA [30]. Adult lung extract contains two bands of activity, one that comigrates with muscle phosphorylase

338 Effectors Added

Muscle RNA 2

3

4

5

6

7

8

9

Probe Brain GP

None

1

2

3

4

5

~

Muscle GP 5mM AMP

qzGlu 5mMAMP +0,7 M Actin Fig, 4, Activity gel assay of phosphorylase activities in developing and adult rabbit tissues. Tissue extracts were prepared, resolved on three identically prepared native polyacrylamide gels and soaked in buffers lacking or containing various allosteric effectors as described in Materials and Methods, and as indicated to the left of each panel. Positions in the gels containing glycogen phosphorylase activity are resolved by staining with an iodine solution, lanes contained the following: lane 1, glycogen phosphorylasc a purified from rabbit muscle (Sigma); lane 2, glycogen phosphorylase b purified from rabbit muscle (Sigma); lane 3, adult liver extract; lane 4, adult lung extract; lane 5, adult brain extract; lane 6, newborn lung extract; lane 7, lung extract from day 28 of gestation; lane 8, lung extract from day 24 of gestation; lane 9, lung extract from day 21 of gestation. The gel is representative of two independent experiments.

b and that is maximally activated by AMP alone and a second band that migrates faster than either muscle or liver phosphorylase and that comigrates with the major band in an adult brain extract. Analysis of phosphorylase activities in developing lung shows a clear increase in both the faster migrating (brain) and slower migrating (muscle) isoforms of phosphorylase at days 24 and 28 of gestation relative to day 21, with a peak at day 28 and a reduction in activity in newborn and adult lung relative to day 28. Assays of a-glucosidase activity were also carried out at neutral (7.5) and acidic (4.5) pH values, as shown in Table I. No change was observed in neutral glucosidase activity as a function of gestational age, although the values obtained for fetal tissues were higher than in newbo~ or adult animals. Acid glucosidase activity, in contrast, increased significantly during

Fig. 5, Expression of glycogen phosphorylase and a-glucosidase mRNA in fetal, newborn, and adult skeletal muscle. The blot was hybridized with the indicated antisense RNA probes at high stringency, as described in Materials and Methods. 2 ~tg poly(A) + RNA was loaded in each lane, except for lane 5 (adult) which contains 0.5 ~g. RNA was prepared from hind leg skeletal muscle samples taken from animals of the following ages: lane I, day 24 of gestation; hme 2, day 28 of gestation; lane 3, day 31 of gestation; lane 4, newborn animals: lane 5, adult. The blot is representative of two independent experiments.

gestation with a peak at day 28 that was 2.6-fold higher than at day 21 (P = 0.02) and then declined in newborn and adult animals relative to the day 28 value. Parallel measurements of both acid and neutral glucosidase activities in developing liver showed a similar pattern of expression of a-glucosidases as observed in lung, except that the peak of acid glucosidase activity was found in newborn animals instead of at day 28 of gestation.

Expression of mRNAs encoding glycogenolytic enzymes in developing liver, brain, heart and skeletal muscle Liver, brain, heart and skeletal muscle tissues exhibit diverse profiles with regard to expression of mRNA encoding glycogenolytic enzymes during development. Previous suggestions that the B isozyme of phosphorylase is 0redominant in all deve|oping tissues [2-9] clearly do not apply at the level of mRNA in the rabbit. In skeletal muscle, for example, B phosphorylase mRNA is barely detectable, while the M isoform

TABLE ! Neutral and acid a.glucosidase acticities in extracts of deceloping rabbit lung and firer Glucosidase activities were determined as described in Materials and Methods and expressed as nmol/min per mg protein for neutral glucosidas¢ activity (pH 7.5) and as pmol/min per mg protein for acid glucosidase activity (pH 4.5). Data are expressed as the mean ± S.E. for three independent s, mples, except for the adult values which represent the mean of two samples. Tissue

pH

~ng Liver

7.5 7,5

Lung Liver

4.5 4,5

Day 21 4.3± 0.2 2.7± 0.6 152 +60 253 +98

Day 24 5.5± 0.7 4.5± 0 . ~ 233 +46 440 +45

Day 28 4.0± 0.4 4.7± 1.0 399 +27 467 _+32

Newborn 2.8± 5.0±

Adult 0.3 0.3

207 :l: 28 648 +142

1.3 4.0 154 458

339 Liver RNA Probe Brain G,°

1

2

3

4

Brain RNA 5

6

Probe Brain GP

Uver GP

Liver GP

Muscle GP

Muscle GP

ot Glu

et Glu

Actin ~ ~ ~ : ~ Fig, 6. Expression of glycogen phosphorylase and a-glucosidase mRNA in fetal, newborn and adult liver. The blot was hybridized with the indicated antisense RNA probes at high stringency as described in Materials and Methods. 5 gg poly(A) + RNA was loaded in each lane, isolated from liver samples taken from animals of the following ages: hme l, day 21 of gestation; lane 2, day 24 of gestation; lane 3, day 28 of gestation; lane 4, day 31 of gestation; lane 5, newborn animals; lane 6, adult. The blot is representative of two independent experiments.

is found to be present at day 24 of gestation and to increase during development, ultimately obtaining a level in the adult that is 61-fold higher than at day 24 (see Fig. 5). As shown in Fig. 6, the mRNA encoding the L isozyme of phosphorylase is by far the predominant message present at all developmental time points in liver, although some B isoform is also expressed. The data for phosphorylase mRNA express=on deHeart RNA

Probe Brain GP

1

2

3

4

5

Muscle GP

a Glu

Actin Fig. 7. Expression of glycogen phosphorylase and a-glucosidase mRNA in fetal, newborn and adult heart. The blot was hybridized with the indicated antisense RNA probes at high stringency as described in Materials and Methods. 5 /~g poly(A) + RNA was loaded in each lane, isolated from heart samples taken from anita ds of the following ages: lane 1, day 24 of gestation; lane 2, day 28 of gestation; lane 3, day 3! of gestation; lane 4, newborn animals; lane 5, adult. The blot is representative of two independent experiments.

1

2

3

4

5

6

Actin Fig. 8. Expression of glycogen phosphorylase and a-glucesidase mRNA in fetal, newborn and adult brain. The blot was hybridzed with the indicated antisense RNA probes at high stringency as described in Materials and Methods. 5 ~g poly(A)' RNA was loaded in each lane, isolated from brain samplrs taken from animals of the following ages: lane I, day 21 of gestation; lane 2, day 24 of gestation; lane 3, day 28 of gestation; lane 4, day 31 of gestation; lane 5, newborn animals; lane 6, adult. The blot is representative of two independent experiments.

scribed here for rabbit liver are consistent with those previously reported for single time points in the second trimester of human fetal development and adult human liver tissues [23,38]. Interestingly, heart muscle has an expression profile distinct from that described for skeletal muscle. Again, the L isozyme is absent (not shown), but in heart, there is significant expression of the B isoform, as shown in Fig. 7. The ratio of M:B signal measured by densitometry is approx. 1.5 during development, but climbs to 3.7 in the adult. Other than lung, the only tissue that predominantly expresses the B isozyme is brain. Very little L isoform i~ detectable in brain, but interestingly, the M isozyme is transiently expressed. As shown in Fig. 8, RNA encoding the M isozyme is present at significant levels at day 21 of development but disappears by day 24. M isozyme mRNA remains undetectable at the later time points in brain even after prolonged exposure of the blot to film (data not shown). The 4 kb lysosomal a-glucosidase message is found at intermediate levels, with no substantial changes noted as a function of ontogeny, in all of the tissues surveyed herein. Discussion

The present study was designed to evaluate the expression of glycogenolytic enzymes and m RNAs that encode them in developing rabbit tissues, with particular emphasis on the lung, a tissue in which glycogen is a

340 major precursor of surfactant phospholipid biosynthesis during fetal development. We found that both glycogen phosphorylase and acid a-glucosidase enzymatic activities increased during fetal lung development, achieving a peak at day 28 of gestation, thereafter declining to the adult levels. This increase in enzyme activity coincides with the burst in glycogen degradation that occurs in lung between days 24 arid 28. Our work shows that the predominant phosphorylase mRNA in developing and adult lung is the brain type, with a low level of expression of muscle phosphorylase mRNA and little detectable expression of liver phosphorylase message. These data are in agreement with the native gel activity assay, which shows that the predominant phosphorylase activity in developing and adult lung comigrates with the major band of phosphorylase activity in adult brain, in addition, lung extracts contain a secondary weaker band of activity that comigrates with purified muscle phosphorylase. There is a noticeable increase in both the brain and muscle isozyme activities in lung as a function of gestational age, with a peak at day 28. While the peak in muscle phosphorylase activity coincides with peak expression of its mRNA, two independent experiments failed to demonstrate any change in the level of the brain phosphorylase mRNA in developing lung. Previous investigators have shown glycogen phosphorylase activity to increase by approx. 2-fold in fetal rabbit [33,39] and rat [36] lung at time points coincident with the onset of increased glycogenolytic activity. In the latter study by Maniscalco et al. this increase was attributed largely to an increase in phosphorylase a activity. In the current study, an increase in brain iso~'yme phosphorylase a activity was indeed noted in developing lung with increasing gestational age, with a peak at day 28 (note the activity in fetal lung extracts assayed in the absence of effectors in Fig. 4). This activity, however, represents only a small percentage of the total phosphorylase activity, as shown by the intense bands generated by incubation with AMP + Na2SO4 (agents that activate phosphorylase b). Dephosphorylation of phosphorylase a during preparation of the tissue extracts for native gel activity assays is unlikely given that they were prepared in a NaF buffer in order to inhibit phosphatase activity (the same buffer used by Maniscalco et al. [33]). Thus, the increase in phosphorylase activity appears to be largely due to an increase in the level of the brain phosphorylase protein. in the absence of any noticeable increase in its mRNA, this induction would appear to be brought about by a posttranscriptional mechanism. Studies with the lysosomal a-glucosidase inhibitor acarbose have indicated that 40% of the glycogenolytic activity of fetal lung explants is contributed by this enzyme [17], although a recent study suggests that

acarbose may also be a weak inhibitor of muscle phosphorylase activity [40]. our studies indicate that there is a small but significant increase in acid but not neutral glucosidase activity in developing lung that peaks at day 28 of gestation. This increase in activity is not accompanied by any significant alteration in the levels of the 4.0 kb a-glucosidase mRNA that encodes the enzyme during fetal lung development. Instead, we find in two separate experiments that a large mRNA related to lysosomal a-glucosidase is induced in lung between the developmental time points (day 24 and 28) corresponding to the glycogenolytic burst that contributes to surfactant production. Although clearly related to a-glucosidase, the identity of the message is unknown. We have excluded the possiblity that it encodes sucrase/isomaltase, a protein encoded by a large mRNA that is similar in sequence to .-glucosidase. Subcellular fractionation and subsequent assay of glucosidase activity using the artificial substrate 4-methylumbelliferyl-a-D-glucoside has shown that lung contains lysosomal acid a-glucosidase, which is bound by concanavalin A, a lamellar body acid a-glucosidase activity which does not bind the lectin, and a neutral glucosidase activity associated with microsoma~ fractions that does not bind lectin and that is not inhibited by glycogen [16]. It is possible that the large glucosidase-like message described herein encodes one of the latter two proteins or an as yet uncharacterized glycogenolytic enzyme. The existence of three isozymes of glycogen phosphorylase in mammals, with distinct structural and kinetic properties, has been recognized for ,T~ny years. In addition, recent cloning and chromosome mapping studies have shown unequivocally that the liver (L), muscle (M) and brain (B) isoforms are products of distinct genes [1,18-23,41,42]. Studies on tissue distribution and developmental expr:~sion of phosphorylases have largely relied on protein electroph,::.-etic studies, and have usually been carried out in the rat. A number of investigators have c~:lcTaded that the B isozyme is predc:ninantly expressed m fetal rat tissues, being completely replaced in adult animals by the L and M isozymes in liver and muscle, respectively (iso.. zyme 'switching') [2-9]. in contrast, we have previously shown that the predominant phosphorylase mRNA in human fetal liver (24 weeks) is the L type; mRNA encoding the B isozyme is barely detectable at this stage of development, suggesting that isozyme switching does not occur in this tissue in humans [23]. In order to investigate this issue further, we decided to evaluate phosphorylase isozyme mRNA expression during development in a number of rabbit tissues other than lung. Analysis of mRNA expression in developing rabbit liver and muscle shows a pattern quite distinct from that described for expression of phosphorylase proteins

341 in liver and muscle of the rat. Our data show that the L and M phosphorylase mRNAs are by far predominant at all developmental time points in rabbit liver and muscle, respectively. B isozyme mRNA is barely detectable in muscle, and although present in liver, does not change significantly during the course of development. Thus, there is no evidence for isozyme switching (at the mRNA level) in developing liver and muscle tissues in the rabbit, as appears to occur at the protein level in the rat. We found a profound induction (25-fold) in the level of M isozyme mRNA in adult rabbit muscle relative to fetal or newborn muscle. Such an induction also appears to occur at the protein level in the rat [8]. Our data also are consistent with the protein data for heart, where others have noted coexpression of the B and M isozymes in fetal development, as well as in the adult. Heterodimers between B and M monomers (MB) have been shown to exist at significant levels in heart extracts from rats, humans and rabbits [8,43,44]. The selective advantage conferred by formation of MB heterodimers in heart is not obvious, since the M and B isozymes are similar in most of their kinetic and allosteric properties [30,45-49]. The B isoform is the predominantly expressed phosphorylase mRNA at all developmental time points in rabbit brain. In rat, the B isozyme was found to predominate during development in brain, but expression of other isozymes was observed, particularly in the adult, where a band comigrating with the M isoform was detected at significant levels in at least two different studies [8,9]. In the current study, we found abundant M isozyme mRNA in rabbit brain at the earliest time point examined (day 21), but no detectable expression at subsequent developmental times, including in the adult. The significance of this transient M isozyme expression in developing rabbit brain is currently unknown. In summary, developing rabbit tissues exhibit diverse patterns of phosphorylase mRNA isoform expression, with no single isozyme predominant during fetal development, and with no evidence of any isozyme switching. Isozyme switching could conceivably still occur at the protein level in the absence of changes in the relative levels of the respective mRNA species, via regulation of translation efficiency or protein stability, for example. Such mechanisms, however, are not supported by our activity gel analysis of selected fetal and adult tissues, which show that the abundance of the protein isoforms generally conforms to the relative abundance of their respective mRNAs.

Acknowledgements The authors thank Dr. Rochelle Hirschhorn for helpful advice and critical reading of the manuscript.

We are also grateful to Dr. Shuichi Miyaura and Ms. Karen Boykins for their help with the animal surgery and Mr. Lee Bryant for excellent technical assistance. Finally, we thank Dr. Peter Hwang, Dr. Micheile Browner and Dr. Robert Fletterick for providing us with the muscle phosphorylase probe. This work was supported by NIH grant R29-DK40734 (to C.B.N.) and 2-PO1-HDI3912 (to J.M.J.). B.N. was the recipient of a Chilton Foundation Fellowship.

References 1 Newgard, C.B., Hwang, P.K. and Fietterick, R.J. (1989) CRC Crit. Rev. Biochem. Mol. Biol. 24, 69-99. 2 Sato, K., Morris, H.P. and Weinhouse, S. (1972) Science 178, 879881. 3 Delain, D., Meienhofer, M.C., Proux, D. and Schapira, F. (1973) Differentiation I, 349-354. 4 Sato, K. and Weinhouse, S. (1973)Arch. Biochem. Biophys. 159, 151-159. 5 Sato, K., Satoh, K., Sato, T., lmai, F. and Mort'is, II.P. (197t~) Cancer Res. 36, 487-495. 6 DiMauro, S., Arnold, S., Miranda, A. and Rowland, L.P. (1978) Ann. Neurol. 3, 60-66. 7 Miranda, A.F., Somer, H. and DiMauro, S. (1979) in Muscle Regeneration (DiMauro, S., ed.), pp. 453-473, Raven Press, New York. 8 Richter, F., Bohme, H.-J. and Hofmann, E. (1983) Biomed. Biochim. Acta 10, 1229-1235. 9 David, E.S. and Crerar, M.M. (1986) Biochim. Biophys. Acta 880, 78-90. 10 Martiniuk, F. and Hirschhorn, R. (1981) Biochim. Biophys. Acta 658, 248-26 l. 11 Martiniuk, F., Smith, M., EIlenbogen, A., Desnick, R.J., Astrin, K., Mitra. J. and Hirschhorn, R. (1983) Cyogenel. Cell Genet. 35. ! 10-116.

12 Martiniuk, F., Ellenhogen, A. and Hirschhorn. R. (1985) J. Biol. Chem. 260, 1238-1242. 13 Hunziker, W., Spiess, M., Semenza, G. and Lodish, H.F. (1986) Cell 46, 227-234. 14 Hoefsloot, L.H., Hoogeveen-Westerveld, M., Kroos, M.A., Van Beeumen, J., Reuser, A.Jd. and Oostra, B.A. (1988) EMBO J. 7, 1697-1704. 15 t3arnes, A.K. and Wynn, C.H. (1988) Proteins Struct. Func. Genet. 4, 182-189. 16 De Vries, A.C.J, Schram, A.W., Tager, J.M., Batenburg, J.J. and Van Golde, L.M.G. (1986) Biochim. Biophys. Acta 878, 288-291. 17 Bourbon, J.R., Doucet, E. and Rieutort, M. (1987) Biochim. Biophys. Acta 917, 203-210. 18 Hwang, P.K., See, Y.P., Vincentini, A.M., Powers, M.A., Fletterick, R.J. and Crerar, M. M. (1985) Eur. J. Biochem. 152. 267-274. 19 Nakano, K., Hwang, P.K. apd Fletterick, R.J. (1986) FEBS Lett. 204, 283-287. 20 Newgard, C.B., Nakano, K., Hwang, P.K. and Fletterick, R.J. (1986) Proc. Natl. Acad. Sci. USA 83, 8132-8136. 21 Osawa, S., Chiu, R.H., McDonough, A., Miller, T.B. and Johnson, G.L. (1986) FEBS Lett. 202. 282-286. 22 Burke, J., Hwang, P., Anderson, L,, Lebo. R., Gorin, F, and Fletterick, R.J. (1987) Proteins Struct. Funct. Genet. 2, 177-187. 23 Newgard, C.B., Littman, D.R., Van Genderen, C.. Smith, M. and Fletterick, R.J. (1988) J. Biol. Chem. 263, 3850-3857. 24 Martiniuk, F., Mehler, M.. Pellicer, A., Tzall, S,, La Badie. G,. Hobart, C., Ellenbogen, A. and Hirschhorn, R. (1986) Proc. Natl. Acad. Sci. USA 83. 9641-9644.

342 3_5 Chan, T.M. and Exton, J.H. (1976) Anal. Biochem. 71, 96-105. 26 Ashley, P.L. and MacDonald, R.J. (1985) Biochemistry 24, 45124520. 27 Gunning, P., Ponte, P., Okayama, H., Engel, J., Biau, H. and Kedes, L. (1983) Mol, Cell. Biol. 3, 787-795. 28 Melton, D.A., Krieg, P.A., Rebagliati, M.R., Maniatis, T., Zinn, K. and Green, M.R. (1984) Nucleic Acids Res. 12, 7035-7056. 29 Hwang, P,K., Fletterick, R.J., Nakano, K, and Newgard, C.B. (1987) in Protein Structure, Folding and Design 2 (Oxender, D,, ed.), pp. 247-256, Alan R. Liss, New York. 30 Coats, W.S., Browner, M.F., Fletterick, R.J. and Newgard, C.B. (1991) J. Biol. Chem. 266, 16113-16119. 31 BTadford, M.M. (1976)Anal, Biochem, 72, 248-254, 32 Shelley, H.J. (1961) Br. Med. Bul. 17, 137-143. 33 Bandstrup, N. and Kretchmer, N. (1965) Dev. Biol. !1,206-216. 34 Sorokin, S, (1960) Acta Anat. 40, 105-119, 35 Williams, M,C. and Mason, R,J, (1977) Am, Rev. Resp. Dis. 115, 37-47, 36 Maniscalco, W.M,, Wilson, C.M,, Gross, I,, Gobran, L., Rooney, S.A. and Warshaw, J.B. (1978) Biochim, Biophys. Acta 530, 333-346, 37 Hoffman, D.R., Truong, C.T. and Johnston, J,M, (1986) Biochim. Biophys, Acta 879, 88-96,

38 Gorin, F.A., Mullinax, R.L., lgnacio, P.C., Neve, R.L. and Kurnit, D.M. (1987) J. Neurogenet. 4, 293-308. 39 Bourbon, J. and Jost, A. (1982) Pediatr. Res. 16, 50-56. 40 Goldsmith, E.J., Fletterick, R.J. and Withers, S.G. (1987) J. Biol. Chem. 262, 1449-1455. 41 Lebo, R.V., Gorin, F., Fletterick, R.J., Kao, F., Cheung, M., Bruce, B. and Kan, Y.W. (1984~ Science 225, 57-59. 42 Newgard, C.B., Fletterick, R.J., Anderson, L. and Lebo, R.V. (1987) Am. J. Hum. Genet. 40, 351-364. 43 Proux, D. and Dreyfus, J-C. (1973) Clin, Chim. Acta 48, 167-172. 44 Davis, C.H., Schliseifeld, L.H., Wolf, D.P., Leavitt, C.A. and Krebs, E,G. (1967) J. Biol. Chem. 20, 4824-4833. 45 Sutherland, E.W. and Wosilait, W.D. (1956) J. Biol. Chem. 218, 459-468. 46 Appleman, M,M., Krebs, E.G. and Fisher, E,H. (1966) Biochemistry 5, 2101-2107. 47 Maddaiah, V.T. and Madsen, N.B. (1966) J. Biol. Chem. 241, 3873-3881. 48 Guenard, D., Morange, M. and Buc, H. (1977) Eur. J. Biochem. 76, 447-452, 49 Kobayashi, M., Soman, G. and Graves, D.J. (1982) J. Biol. Chem. 257, 14041-14047,

Developmental expression of glycogenolytic enzymes in rabbit tissues: possible relationship to fetal lung maturation.

Glycogen can be degraded in mammalian tissues by one of three isozymes of glycogen phosphorylase, termed muscle (M), liver (L) and brain (B) after the...
1MB Sizes 0 Downloads 0 Views