Developmental expression of chicken antithrombin III is regulated by increased RNA abundance and intracellular processing

Biochimica et Biophysica Acta, 1171 (1993) 239-246

239

© 1993 Elsevier Science Publishers B.V. All rights reserved 0167-4781/93/$06.00

BBAEXP 92436

Developmental expression of chicken antithrombin III is regulated by increased RNA abundance and intracellular processing David L. Amrani a, Jonathan Rosenberg a,b, Fahumiya Samad a, Gerald Bergtrom b and David K. Banfield c a Department of Health Sciences and b Department of Biological Sciences, University of Wisconsin-Milwaukee, Milwaukee, W1 (USA) and c Department of Biochemistry, University of British Columbia, Vancouver (Canada)

(Received 16 April 1992) Key words: Antithrombin III; cDNA cloning; RNA nuclease-protection assay; In vitro translation We isolated and sequenced a 432 bp cDNA to cAT-III, that encoded 115 nucleotides of 5' untranslated sequence, a 17 amino acid long signal peptide and residues 1-88 of the mature protein, and used it to prepare a probe for measuring and correlating the developmental changes of steady-state cAT-III mRNA levels with known changes in antigen levels. Densitometric analysis of nuclease protection (n = 2), Northern blot (n = 4), and slot blots (n = 3) of total RNA from chick livers of 16-day-old embryos to 6-day-old chicks showed a 2.6 ___0.5-fold increase in steady-state cAT-III mRNA levels. Assay of functional mRNA levels by in vitro translation of poly(A) + RNA and specific immunoprecipitation of 35S-Met-labelled cAT-III was comparable to RNA analysis (16-day-old embyros vs. 10-day-old hatchlings). We evaluated whether there were developmental differences in post-translational secretion which may also contribute to the regulation of the circulating level of this protein. Pulse-chase studies of freshly-isolated hepatocytes from 16-day-old embryos and 10-day-old hatchlings maintained in suspension demonstrated a approx. 5.0-5.5-fold increase in cAT-III levels at steady-state secretion. The above findings indicate that changes in circulating cAT-III levels during late embryonic development are primarily due to increased abundance of cAT-III mRNA. In addition, we postulate that post-translational intracellular processing may account for further differences in circulating protein levels. Introduction Antithrombin III (AT-III) belongs to a family of serine proteinase inhibitors (Serpins) that play a critical role in maintaining normal hemostasis by acting as the major inhibitor of thrombin and factor Xa in the circulation of humans and other vertebrates [1-4]. ATIII from evolutionarily diverse vertebrates such as humans, rodents and aves are glycoproteins having a molecular size ranging from 57-62 kDa [4-8]. The complete chemical structure for human A T - I I I has been determined from both amino acid sequence anal-

Correspondence to: D.L. Amrani, University of Wisconsin-Milwaukee, Department of Health Sciences, P.O. Box 413, Milwaukee, WI 53201, USA. These results were presented in part at the Cold Spring Harbor Laboratory Meeting of the 'Regulation of Liver Gene Expression', May 3-7, 1989. The sequence data reported in this paper have been submitted to the EMBL/Genbank Data Libraries under the accession number L07842.

ysis [9] and c D N A sequence analysis [10-12]. A T - I I I inhibits thrombin by forming a covalent complex with thrombin in a 1:1 molar ratio [4,13]. The antiproteinase activity of A T - I I I is greatly accelerated when heparin is present [3,14]. All known A T - I I I molecules possess a high-affinity heparin binding domain [8], which may interact with heparan sulfate molecules on vascular cells or surfaces [15,16] to accelerate the inhibition of thrombin in vivo [17]. A T - I I I is produced in the liver [18,19]. Immunostaining and protein synthesis studies have clearly demonstrated that the hepatocyte is the cellular site of A T - I I I synthesis [20,21]. H u m a n hepatoma cells have also been shown to produce A T - I I I [22]. The hepatocyte is also the cellular site at which the A T - I I I thrombin complex is specifically cleared from the circulation [23,24]. The importance of A T - I I I in preventing thrombotic events is evidenced by the increased incidence of thromboembolic problems in patients with acquired or congenital deficiencies of this inhibitor [25,26]. Circulating levels of A T - I I I in humans [27-31] and chickens [21] are developmentally regulated, rising from 16-25%

240 of the adult level at approx. 7 months of human gestation (equivalent to the t6th day of chick embryonic development) to 50-80% after birth. In newborns who have pathophysiological complications or have catheters inserted, the development of a hypercoagulable state is often correlated with reduced AT-III levels [30-31]. The molecular basis for this developmental regulation is not known. Previously, we investigated AT-III development in the chicken model system and determined that antigen levels, degree of sialylation, and accompanying charge differences were under developmental control [21]. To further define the molecular mechanisms involved in chicken AT-III (cAT-lID, development, we isolated a cAT-II1 cDNA clone from a cDNA library prepared from chick liver mRNA in Agtll for assessing changes in cAT-III mRNA abundance. In this report, we evaluated changes in steadystate cAT-III mRNA levels. We correlated these results with changes in levels of functional cAT-III mRNA, and examined the kinetics of cAT-III secretion, and circulating cAT-III levels. Materials

and Methods

All chicks were white leghorns. Polyclonal antibodies against cAT-Ill were raised in rabbits as described by Koide et al. [32]. The antisera were made specific by absorption against chicken plasma which had been absorbed over a heparin-Sepharose (Pharmacia) column. The immunoglobulin fraction was purified over an Affi-Blue (Bio-Rad) gel column. A synthetic oligonucleotide (5'-GAG GAC ATC TGC ACT GCC AAG CCC ACT GAC ATC CCT GT-3) which corresponds to cAT-III amino acid residues 6-18 [32] was prepared using a 'best-fit' codon bias by Dana Fowlkes (University of North Carolina, Chapel Hill, NC). A cDNA synthesis kit using the protocol of Gubler and Hoffman [33] was obtained from Amersham (Arlington Heights, IL). DNA sequencing and micrococcal nuclease-treataed rabbit reticulocyte in vitro Translation kits were obtained from Promega (Madison, WI). All radioisotopes and GeneScreen Plus membranes were obtained from New England Nuclear (Boston, MA). Moloney Murine Leukemia Virus (M-MLV) reverse transcriptase as well as all other enzymes and ultrapure reagents were obtained from Bethesda Research Laboratories (Gaithersburg, MD).

Poly(A) + RNA isolation, in vitro translation, and immunoprecipitation Total RNA was prepared from chicken liver that had been placed in liquid nitrogen after removal of the bile duct and rinsed in Hanks' buffer (pH 7.4) or Williams E medium containing 20 mM Vanadyl Ribonucleoside Complex (BRL). Poly(A) + RNA was obtained by the guanidinium isothiocyanate/cesium chlo-

ride procedure of Chirgwin et al. [34] as described by Perbal [35] followed by oligo(dT) cellulose chromatography, or by the method of Badley et al. [36]. In vitro translation was initiated by the addition of e[35S]methionine (25 p~Ci, specific activity 1115 Ci/mM) and 5 ~g poly(A) + RNA, and followed by incubation at 35°C for 60 min. Incorporation of L-[3~S]methionine was measured by trichloroacetic acid precipitation [21] and total protein products were analyzed by electrophoresis on a 10% resolving polyacrylamide gel following the conditions of Laemmli [36]. For immunoprecipitation, 40 #1 of the translation mixture was diluted with equal volume of cell lysis buffer as described by Plant et al. [37], vortexed twice, and centrifuged at 10000 × g for 5 min. The supernatant was removed, 10 ~1 of rabbit anti-cAT-1II IgG (5 mg/ml) was added, and the mixture incubated at 37°C for 1 h followed by 4°C overnight. This antibody was originally made against the adult cAT-III; it was tested against purified adult, embryonic and desialyated adult [125I]cAT-III and showed equal ability to quantitatively precipitate these cAT-IIIs(data not shown). 50 ~1 of protein A-agarose (Sigma, St. Louis, MO) were added to the sample and the samples were subjected to gentle mixing for 3 h at 4°C. As previously described [21], immunoprecipitates were collected, eluted, radioactivity determined and samples analyzed on polyacrylamide gel electrophoresis. The gels were processed for fluorography by soaking in an radioautographic enhancer (Englighting, New England Nuclear) for 15 min, then dried and exposed to Kodak AR X-ray film at - 80oc.

Preparation of the cDNA library Preparation of cDNAs was by the procedure of Gubler and Hoffman [33]. Briefly, 15 ~g of chicken liver poly(A) + RNA was used as template for first strand cDNA synthesis by reverse transcriptase. The second strand was generated [33] and cDNAs were methylated. EcoR! linkers were added by ligation, and the dscDNAs were ligated and packaged into Agtll by Clontech (Palo Alto, CA). The chicken liver cDNA library contained 2.1.106 clear (recombinant) plaques with an average insert size of 1.1 (kilobases) kb with a range of 0.31 to 3.9 kb. Isolation, subcloning and sequencing of cAT-1H clones The Agtll cDNA library was grown in Escherichia coli Y 1090 cells. Recombinant plaques, (1-5000/100 mm plate) were transferred to nitrocellulose filters and were screened under standard conditions with the 38mer oligo probe labelled by C-tailing to a specific activity of (1-2). 109 dpm/~g. Several positive plaques were obtained after tertiary screening. DNA from isolated positive plaques was purified from Lambdasorb columns (Promega Biotec, Madison, WI). The DNA

241 was digested with EcoRI and the products electrophoresed on a 1% agarose gel. The liberated insert was isolated by gel elution and subsequently ligated to pUC 18, (restricted with EcoRI and dephosphorylated). Recombinant pUC was then transformed in competent E. coli DH5a (BRL) according to a modification of Hanahan [39] as recommended by BRL. Plasmid DNA (termed peAT clones) from isolated positive clones was prepared using the alkaline lysis method [40]. One clone was amplified and the insert isolated and subcloned into pGEM-4Z (Promega). The new plasmid, peAT-2 was sequenced by the modified SangerDideoxy method for sequencing double stranded plasmids [41] using [32p]dATP. The plasmid was sequenced in both directions using sequencing primers for both the SP6 and T 7 promoter regions on the vector and with an oligonucleotide (19-mer) synthesized to correspond to the coding strand sequence determined from nucleotide 168 to 186 (Fig. 1).

-160

-15O

Nuclease protection analysis Nuclease protection analysis was performed essentially as described by Norton and Hynes [42]. To obtain a RNA probe for these studies, the 432 bp cAT-Ill eDNA insert in peAT1 was ligated into pGEM4Z linearized with PuulI. A [32p]CTP-labeled cAT-Ill antisense mRNA strand was transcribed in vitro from the T7 promoter (Riboprobe, Promega). After purification of the RNA probe, the probe was denatured at 85°C and hybridized to chicken poly(A) ÷ RNA (0.5 /zg) from 16, 18 and 20-day-old embryonic livers and 10day-old Hatchling livers in 80% formamide, 40 mM Pipes, 400 mM NaC1, 1 mM EDTA overnight at 37°C. All samples contained an excess of probe and carrier yeast t-RNA (total 30 /zg RNA). Unhybridized probe was digested by the addition of 0.24 ml RNase A (0.2 /xg/ml) and RNase T1 (0.5/zg/ml) in 50 mM Tris-HCl buffer (pH 7.5) containing 500 mM NaC1 at 25°C for 40 min. Samples were phenol extracted, ethanol precipitated, and resuspended in 80% formamide dye mix,

-140

-130

-120

-110

I

AA

TIrC CAG CTG CTT GGG GCT GCT OGG GGC GTr - 100

CTG AGG ATe TGC ArT -50

-80

.90

•

TGG GTC TCA AGa CAC TGC CTG TCC ATC -70

-60

TAC ACC CAG AAC GCA GGA C.w3~ TGG AGG CAC CAG GCT G(3G ACA

-4O

-3O

-20

-10

1

ATG CAT CTC TTC ATC GGT GTG TCT CTl" CGG CCT CTG GGG CAT GGC ATC CCT GCA CCC M H k F I G V $ L R p L G H G I P A P 10

20

3O

TAT GCC GTG GAA GAC A l l " Y A V [E O I 70

4O

5O

TGC ACT GCA A,AG CCC CGT GAC ATC CCA GIG AAC CCC ATC C T A K P R D I P V] N P I

80

90

100

110

TGC ATC TAC CGC AAC CCT GAG AAG AAG CCC CAG GAA AGG ~ C I Y R N P E K K P Q E R R 130

6O

140

150

120

GGT GCT GGA GCC C-,-,-,-,-,-,-,-,~G G A G A G

160

1 70

GAP, GC~ CAG GAT CCC GGA GTT CAC AAA CCC CCG GICG TCT GGG AGC TGT CCA GGG CCA E G O D P G V H K P P A S G S C P G P 180

190

ACT ~ T R

200

GC.g l-rC GGC CGT CGG TCT TIC A F G R R S F

240

250

R

R

T

S

S

TIA L

220

230

CAA C-CA CCT GGC CoG ACT CCA AGG ACA O A P G P T P R T

260

ATG AGG AGA ACA TCT TCC TGT ~ M

210

C

CCC TCG G R

P

S

Fig. 1. Sequence of peAT-2 eDNA clone. The proposed signal peptide sequence is underlined and the sequence corresponding to the oligo (nucleotide sequence 16-54) used for screening is in brackets. The eDNA sequence obtained is presented together with corresponding amino acid sequence.

242 and run on a 6% denaturing polyacylamide gel at 45 mA for 5 h. Protected fragments were visualized by autoradiography.

Slot blot analysis Approx. 10 /xg of poly(A) + RNA from 16, 18 and 20-day-old embryonic livers and 10-day-old Hatchling livers was diluted in a final volume of 1 ml of sample buffer containing 6% formaldehyde/5 x S S C / 5 0 % formamide. The samples were heated at 50°C for 1 h and chilled on ice for 1 rain. The samples were then applied at 0.5, 1 and 3 txg/slot in duplicate slots (American Bionectics) to GeneScreen Plus membrane. The samples were drawn onto the membrane by vacuum (15" Hg) for 1 min and the slots were subsequently rinsed with 300/×l/slot of sample buffer. The blot was baked at 80°C under vacuum and the RNA was UV crosslinked to the blot. The blot was then prehybridized with 50% f o r m a m i d e / 5 x S S P E / 5 × Denhardt's r e a g e n t / l % SDS/200 /xg/ml salmon D N A / 5 0 txg/ml yeast t-RNA and hybridization was carried out in the same solution with the t-RNA deleted. Probes were either nick-translated chicken prothrombin cDNA (Irwin, D.M., Banfield, D.K., Walz, D. and MacGillivray, R.T.A., unpublished data) or labeled cAT-III RNA transcripts. Prehybridization/hybridization temperatures for probing with prothrombin cDNA or cAT-III RNA were 42°C and 55°C, respectively. Blots probed with prothrombin cDNA were washed sequentially in 6 x SSC/0.5% SDS at room temperature, 1 x S S C / I % SDS at 37°C, and in the same solution at 65°C for 15-30 min. Blots probed with labeled cAT-III RNA were washed three times in 1 x S S P E / I % SDS at 65°C with a final wash (two times) in 0.1 x S S P E / 2 % SDS at 65°C for 30 rain. The blots were exposed to X-ray film at - 7 0 ° C for up to 3 days. Relative changes in m R N A levels from Northern and slot blot data were determined from densitometric scans of autoradiograms using a Gilford response spectrophotometer with a densitometric scanning attachment at 500 nm.

Labelling of oligonucleotides, cDNA and RNA probes Oligonucleotide probes were labelled either with [T-32p]dATP using polynucleotide kinase reaction [43] or with [a-32p]dCTP tailing [44]. Oligonucleotide probes (10-30 ng) were generally labelled to (1-3)" 10 9 dpm//xg. Nick-translated cDNA to chicken prothrombin was prepared using a nick-translation kit (BRL) and [a32p]dCTP. The cDNA probe was generally labelled to a specific activity of (0.9-1•0). 108 dpm/p~g. ~2P-labelled beta-chicken actin was obtained from Oncor (Gaithersburg, MD). RNA transcripts, made from the SP6 (sense) o r T 7

(antisense) DNA stands from pCAT-2, were labelled using the Promega RNA transcript synthesis kit using the appropriate RNA polymerase. RNA probes were generally labelled to a specific activity of approx. (1-5) • l0 s dpm/ixg. Prior to use, the RNA probes were separated from free radioactive label by passage over a Sephadex G-25 spun column (Boehringer-Mannheim, Indianapolis, IN).

Pulse-chase studies of isolated heptocytes Metabolic labelling of secreted proteins was accomplished by pulse-labeling of hepatocytes freshly isolated from 16-day-old embryonic and 10-day-old hatchling livers• The isolation procedure was performed as previously described [21]. Hepatocytes were placed in suspension in 20 mM Hepes buffer (pH 7.5) containing 10% Williams E medium in a 37°C incubator for 15 rain. In this medium with a reduced methionine concentration, the cells were allowed to use their internal pool of methionine. The cells were then centrifuged at 50 x g, washed and resuspended in the same medium containing 500/xCi [35S]methionine (New England Nuclear) and incubated for 5 min. After incubation the cells were centrifuged, and resuspended in Williams E medium with 20 mM Hepes buffer (pH 7.5) containing unlabeled methionine. After one further resuspension, the cells were placed in suspension in a 37°C incubator for the chase period and duplicate 0.5 ml samples were removed at 15, 30, 60, 120, 240 and 360 min. Samples were centrifuged at 12000 x g for 1 min and the media transferred to a separate tube for immunoprecipitation analysis which was carried out as previously described [21]. In separate experiments, cells were collected after the pulse period, and the media was removed. In duplicate cells, the media was replaced as above and the labeled protein was chased for up to 6 h. All media samples were obtained for immunoprecipitation analysis. Results

Identification of a cAT-Ill clone We obtained a cDNA clone with a 432 bp insert (Fig. 1) having a sequence corresponding to the amino terminal end of mature cAT-III as determined from protein sequence analysis (Ref. 32, and D.L. Amrani, data not shown). The clone contains 115 nucleotides of 5' untranslated region, the signal peptide sequence and the sequence of the mature protein from amino acid 1 to 88. We subsequently used this cDNA to generate a probe for analysis of cAT-III mRNA developmental changes.

Developmental changes in cAT-III expression To determine the molecular mechanisms regulating developmental changes in circulating cAT-III levels,

243 we evaluated the role of steady-state mRNA levels, translational and secretory events between the 16th day of embyronic development and 10 days after hatching. Nuclease protection analysis of total RNA (0.5/zg from 16, 18 and 20-day-old embyronic livers and 6-dayold hatchling liver) for direct evaluation of cAT-III RNA levels revealed a clear change in the abundance of cAT-III mRNA. A single major mRNA band at approx. 432 bp was observed at all the developmental stages examined (Fig. 2). The partially protected fragments in the hatchling sample (lane 4) are not always present, and in any event are a minor component of the protected RNA, the majority of which migrates at about 432 bp. In two separate experiments, an approx. 2.8-fold increase in mRNA levels was revealed by densitometric scanning analysis. A similar change in mRNA levels was also found by Northern blot analysis (data not shown). We evaluated the amount of functional cAT-III mRNA by comparing the ability of poly(A) + RNA from 16-day-old embryos and 6-day-old hatchlings to synthesize cAT-III in an in vitro translation system. The addition of 5 /xg of poly(A) + RNA from each developmental stage initiated the synthesis of L-35Slabelled Met-cAT-IlI which was specifically demonstrated after immunoprecipitation followed by SDSPAGE analysis and autoradiography. A single major

1 2 3 4 5 6 7 ~=.,

432

bp--

--

--

--609

bp

m Q

Fig. 2. Nuclease protection analysis of poly(A) + RNA isolated from developing embryo and hatchling liver. Lanes: 1, 16-day-old embryonic; 2, 18-day-old embryonic; 3, 20-day-old embryonic; 4, 6 day old hatchling; 5, hybridization and nuclease digestion of probe in the presence of only yeast t-RNA, and 6 and 7, undigested probe (609 bp). Molecular weights of protected fragment were determined by running parallel sequencing lanes of M13mpl8.

A 1

2

i!ii~!i'iiiii~i~ii~ii~!i!iii!~i~iii!

ii~ ii!i~iiii!i!~ii!iiiiiii!~il !ii ~i ~ ~ ~i! ~ ~,~ ~i~ ~ ii~ i~ i~ ¸

~ ~i ~

!~I ~

!~,

Plasma cAT-III

52 KDa

ii~ i

~i!iiii~i ~i!ii ii~ii~i~iiill ~i~ ~i, ~ ~!ii~! ~i ~!i!~il !i !iiiii! iiii~i~ ! ii ~ iiii~

Fig. 3. In vitro translation and immunoprecipitation of cAT-11I. Immunoprecipitates from diluted original translation mixtures of 16-day-old chick embryonic and 6 day-old hatchling liver poly(A) + RNA are seen in lanes 1 and 2, respectively. The position of adult cAT-III is indicated by the arrow and the approximate molecular size of the immunoprecipitated cAT-III bands is 52 kDa. The apparent bands appearing above the 52 kDa band is only insoluble material at the resolving and stacker interfaces of the gel.

band migrating at approx. 52 kDa was observed with some fainter lower molecular weight bands and some insoluble material at the interfaces of the resolving and stacker gels (Fig. 3). Densitometric analysis revealed a mean 2.8-fold increase in cAT-III mRNA translated products. We also quantified by slot blot analysis cAT-III mRNA and prothrombin mRNA levels (Fig. 4). Previously, Kisker et al. [45] demonstrated that both sheep prothrombin protein and mRNA levels were developmentally regulated. We chose to compare the developmental changes in AT-III and prothrombin mRNA levels which have not been previously examined from any individual species. Slot blot analysis (normalized against chick/3-actin mRNA levels which do not change during this stage of development [46,47]) of RNA samples from livers at 16, 18 and 20 days of embryonic development and 10 days after hatching demonstrated parallel increases in cAT-III and chick prothrombin mRNA levels (Fig. 4). The differences between RNA levels of 16-day-old embryos and 10-day-old hatchings were 2.5 _+0.4-fold and 2.4 + 0.5-fold for normalized values of cAT-III and chick prothrombin, respectively. These increases were observed to be linear at each

244 16

18

20

H CAT-Ill

PROTHROMBIN

~-ACTIN

120 110 100 >

¢

90

<: Z n-

80

o~

60

"~

50

,-I

E c

1 Prothrombin [ " 7 C A T - III

70

o

40O

o~

30 20, 1016

18

20

H

Fig. 4. Histogram comparing the relative changes in cAT-III and chick prothrombin mRNA during development. The % of hatchling mRNA level is plotted versus the days before (16 to 20 days) and after hatching (10 days). Slot blots (see inset) of poly(A) + RNA (1 p,g/slot) were probed successively with cAT-III, chick prothrombin and/3-actin labels (n = 3, duplicate samples). All blots were hybridized to the same amount of labeled probe (1.106 cpm/ml) and blots were exposed to X-ray film for 18-20 h. Autoradiograms were scanned for relative changes in cAT-III and chick prothrombin mRNA levels at each developmental stage are expressed after normalization to/3-actin levels.

developmental stage at three different RNA concentrations (0.5, 1 and 3 /zg/slot) under conditions of probe excess (Amrani, D.L. and Rosenberg, J., unpublished results). Pulse-chase studies demonstrated differences in secretion kinetics and the relative amounts of newly synthesized cAT-III. Freshly-isolated hepatocytes from 16-day-old embryos and 10-day-old hatchlings were pulsed for 5 min and subsequently chased in label-free media for up to 6 h. Media samples (Fig. 5) were immunoprecipitated and analyzed on SDS-PAGE as previously described [21]. At steady-state secretion, which was reached between 3 and 4 h, de novo production of cAT-III was 5.0-5.5-fold higher in 10-day-old hatchlings (Fig. 5, middle panel) compared to 16-dayold embryos (Fig. 5, top panel). The embryonic form migrates in a position consistent with the previously determined [21] migration change for the desialyated CAT-III (see Fig. 5, top panel-position of purified adult CAT-III relative to immunoprecipitated embryonic cAT-III). The transit time for 50% of the newly synthesized cAT-III m o l e c u l e s (/1/2) at the two developmental periods was significantly different: tl/2 for 16-day-old embryonic cAT-III was 105 min, compared to approx. 63 rain for the hatchling. The secretion time

for cAT-III molecules synthesized in the hatchling is similar to that found for other plasma proteins synthesized by cultured chick hepatocytes [48,49]. At the same time, quantitative rocket immunoelectrophoresis, as previously described [21], revealed an increase in circulating cAT-III levels over the same developmental stages. Circulating antigen levels were approx. 5.3-fold higher at the 6th or 10th day after hatching than at 16 days of embryonic development. Discussion

Developmental changes in AT-III levels may result from modulation of transcription, post-transcriptional a n d / o r post-translational processing events. Our original studies [21] suggested that changes in circulating antigen levels and charge characteristics were in part due to transcriptional as well as post-translational events. Earlier studies [27-29] in human embryos, neonates, and premature infant-derived cord blood samples revealed that at the 7th month of human gestation, AT-III levels were 16-25% of adult levels and that adult levels were generally reached 3 months after birth. Based on radioimmunoassay of human liver lysates, Prochownik et al. [11] found that, compared

245 1

2

3

4

5

6

7

8

1

2

3

4

5

6

7

8

25 20

5

~i

60

I

120

I

180

I

240

r

300

360

Time (min.)

Fig. 5. SDS-PAGE analysis of the immunoprecipitated embryonic and adult cAT-III samples. Freshly-isolated hepatocytes from 16day-old embryos and 10-day-old hatchlings were maintained in suspension and pulse labeled for 5 rain with 500 ,~Ci of [35S]methionine. The de novo labeled cAT-III was chased into medium containing unlabeled methionine for the time periods shown. Lanes 1-7 in both autoradiograms correspond to 0, 15, 30, 60, 120, 240, 360 min of chase, respectively. Relative cAT-III levels in the immunoprecipitates were measured by densitometry. Lane 8 in both panels contain purified 125I-labeled adult cAT-III prepared as originally described [21]. Top panel: immunoprecipated embryonic cAT-III; middle panel: immunoprecipitated adult cAT-III; bottom panel: graphic representation of the time course of cAT-III secretion, integration of densitometric data for 16-day-old embryonic (lower curve) and 10-day-old hatchling (upper curve) cAT-III from pulse-chase labelling autoradiograms in the top and middle panels.

with adult liver, the relative abundance of AT-III was 10-fold and 5-fold less in the lysate from 12-14-weekold and 18-20-week-old fetuses, respectively. Amrani et al. [21] reported that chick AT-III levels rose from

26 tzg/ml in the 16-day-old embryo to 143 /~g/ml in the 5-10-day-old chick. These changes reflect a 5.3-fold increase in AT-III at comparable developmental stages and suggest similar regulation of AT-III levels in the two species. Hybridization or nuclease protection analyses of either total or poly(A) + liver RNA from 16-day-old embryos to 10-day-old hatchlings revealed the presence of a single AT-III mRNA species. The change in steady-state AT-III mRNA levels and in AT-III mRNA translational activity between 16-day-old embryos and 10-day-old hatchlings was 2.6 + 0.5-fold, or approx. one-third to half the difference found for the circulating antigen levels. Hassan et al. [50] examined the expression of several human blood coagulation factors including AT-III in 5 to 10-week-old human embryos and fetuses. While no protein or RNA analyses are actually shown, a graph of RNA levels from Northern analyses indicated that between the 6 and 10th week post-conception AT-III RNA levels rose to 50% of adult levels. These data support the conclusion that production of AT-III is dependent in part upon increased RNA levels. Developmental regulation of AT-III and prothrombin in chicks are at least partially controlled by either increased gene transcription, or enhanced mRNA stability. Previously, Kisker et al. [45] demonstrated that sheep prothrombin levels increase during development in parallel with changes in steady-state sheep prothrombin mRNA levels. Combined with our findings, it may be that circulating prothrombin levels may be solely dependent on increased mRNA levels. If so, then the loss of synthesized antithrombin III through as yet undefined intracellular mechanisms may produce a potential imbalance between the circulating levels of AT-III and prothrombin's active component, thrombin, leading to increased thrombosis. The report by Hassan et al. [50] that factor IX levels are relatively low in early embryonic and fetal life compared to factors X and VII suggested that while vitamin-K dependent factors are generally at a low level in the circulation [27-31], factors X and VII are significantly higher. This suggests that in embryonic and early fetal life, activation of blood coagulation via the extrinsic pathway might be favored. Since AT-III represents the major inhibitor of factor Xa and thrombin, a lower level of this inhibitor may contribute to increased risk of thrombosis. Therefore, the lower levels of AT-III may due to a combination of factors including lower mRNA and protein levels, increased intracellular turnover, a n d / o r clearance by the liver. The latter two possiblities may relate to the absence of terminal sialic acid residues on fetal AT-III [21], which would be consistent with it's known accelerated rate of clearance in newborn term infants [51]. We postulate that the difference between the net hepatic production of AT-

246 II1 and the lower circulating levels may yet be another contributing factor to increased occurrence of fetal thrombosis [52-54].

Acknowledgements We thank Angela Mallett for secretarial support in preparation of this paper and Karen Higgins for photographic services. This project was supported by National Institutes of Health Grant HL-39079, National Institutes of Health Grant RR-03326 to the ProteinNucleic Acid Shared Facility at the Medical College of Wisconsin, and a Grant from the Sinai-Samaritan Research Foundation.

References 1 Carrell, R.W. and Boswell, D.R. (1986) in Proteinase Inhibitors (Barrett, A.J. and Salvesen G., eds.), pp. 403-420, Elsevier, Amsterdam. 2 Abildgard, U. (1979) in Physiological Inhibitors of Blood Coagulation and Fibrinolysis (Collen, D., Wiman, B. and Verstraete, M., eds.), pp. 31-33, Elsevier/North Holland, Amsterdam. 3 Rosenberg, R.D. and Rosenberg, J.S. (1984) J. Clin. Invest. 74, 1-6. 4 Rosenberg, R.D. and Damus, P.S. (1973) J. Biol. Chem. 248, 6490-6505. 5 Kurachi, K., Schmer, G., Hermodson, M.A., Teller, D.C. and Davie, E.W. (1976) Biochemistry 15, 368-373. 6 Koide, T. (1979) J. Biochem. 86, 1841-1950. 7 Franzen, L.E. and Svensson, S. (1980) J. Biol. Chem. 255, 50905093. 8 Jordan, R.E. (1983) Arch. Biochem. Biophys. 2227, 587-595. 9 Petersen, T.E., Dudek-Wojciechowska, G., Sottrup-Jensen, S. and Magnusson, S. (1979) in Physiological Inhibitors of Blood Coagulation and Fibrinolysis (Collen, D., Wiman, B. and Verstraete, M., eds.), pp. 43-54, Elsevier/North Holland, Amsterdam. 10 Bock, S.C., Wion, K.L., Vehar, G.A. and Lawn, R.M. (1982) Nucleic Acids Res. 10, 8113-8125. 11 Prochownik, E.V., Markham, A.F. and Orkin, S.H. (1983) J. Biol. Chem. 258, 8389-8394. 12 Chandra, T., Stackhouse, R., Kidd, V.J. and Woo, S.L.C. (1983) Proc. Natl. Acad. Sci. USA 80, 1845-1848. 13 Owen, W.G. (1975) Biochim. Biophys. Acta 405, 380-387. 14 Abildgaard, U. (1968) Scand. J. Haematol. 5,440-453. 15 Lollar, P. and Owen, W.G. (1980) J. Clin. Invest. 66, 1222-1230. 16 Busch, C. and Owen, W.G. (1982) J. Clin. Invest. 69, 726-729. 17 Marcum, J.A., McKenney J.P. and Rosenberg, R.D. (1984) J. Clin. Invest. 74, 341-350. 18 Koj, A., Rgoeczi, E., Toews, C.J., Lveille, R. and Gauldie, J. (1978) Biochim. Biophys. Acta 539, 496-504. 19 Owens, M.R. and Miller, L. (1980) Biochim. Biophys. Acta 627, 30-39. 20 Leon, M., Aiach, M., Coezy, E., Guennec, J.Y. and Fiessinger, J.N. (1983) Thromb. Res. 30, 369-375. 21 Amrani, D.L., Mosesson, M.W. and Koide, T. (1985) Biochim. Biophys. Acta 848, 324-334.

22 Fair, D.S. and Bahnak, B.R. (1984) Blood 64, 194-204. 23 Shifman, M.A. and Pizzo, S.V. (1982) J. Biol. Chem. 257, 32433248. 24 Bauer, P.I., Mandl, J., Machovich, R., Antoni, F., Garzo, T. and Horvath, I. (1982) Thromb. Res. 28, 595-606. 25 Egeberg, O. (1965) Thromb. Diath. Haemorrh. 13, 516-530. 26 Bick, R.L. (1982) Sem. Thromb. Hemost. 8, 276-287. 27 Teger-Nilsson, A.C. (1975) Acta Paediatr. Scand. 64, 624-628. 28 McDonald, M.M., Hataway W.E., Reeve E.B. and Leonard, B.D. (1982) Thromb. Haemot. 47, 56-58. 29 Peters, M., Janen, E., Ten Cate, J.W., Kahle, L.H., Ockelford, P. and Breederveld, C. (1984) Br. J. Haemat. 58, 579-587. 30 Oski, F.A., Naiman, J.L., Stockman, J.A. and Pearson, H.A. (1982) in Major Problems in Clinical Pediatrics (Markowitz, M., ed.), pp. 137-174, W.B. Saunders, Philadelphia. 31 Schmidt, B. and Zipursky, A. (1984) Clin. Perinat. 11,461-487. 32 Koide, E., OIta, Y., Odani, S. and Ono, T. (1982) J. Biochem. 91, 1223-1239. 33 Gubler, U. and Hoffman, B.J. (1983) Gene 25, 263-269. 34 Chirgwin, J.M., Przybyla, A.E., MacDonald, R.J. and Rutter, W.J. (1979) Biochemistry 18, 5294-5299. 35 Perbal, B. (1984) in A Practical Guide to Molecular Cloning, pp. 385-421, Wiley, New York. 36 Badley, J.E., Bishop, G.A., St. John, T. and Frelinger, J.A. (1988) Biotechniques 6, 114-116. 37 Laemmli, U.K. (1970) Nature (London) 227, 680-685. 38 Plant, P.W., Deeley, R.G. and Grieninger, G. (1983) J. Biol. Chem. 258, 15355-15360. 39 Hanahan D. (1985) in DNA Cloning (Glover, D.M., ed.), Vol. I, pp. 109-136, IRL Press, Oxford. 40 Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982) Molecular Cloning, A Laboratoty Manual, p. 90, Cold Spring Harbor Laboratory Press, Cold Spring Harbor. 41 Chen, E.Y. and Seeburg, P.H. (1985) DNA 4, 165-170. 42 Norton, P.A. and Hynes, R.A. (1987) Mol. Cell. Biol. 7, 42694307. 43 Maxam, A. and Gilbert, W. (1980) Enzymology 65, 499-599. 44 Collins, M.L. and Hunsaker, W.R. (1985) Anal. Biochem. 151, 221-224. 45 Kisker, C.T., Perlman, S., Bohlken, D. and Wicklund, B. (1988) J. Lab. Clin. Med. 112, 407-412. 46 Ordahl, C.P., Tilghman, S.M., Ovitt, C., Fornwald, J. and Largen, M.T. (1980) Nucleic Acids Res. 8, 4989-5005. 47 Shani, M., Zevin-Sonkin, D. Carmon, Y., Czosnek, H., Nudel, U. and Yaffe, D. (1982) In Muscle Development (Pearson, M.L. and Epstein, H.F., eds.), pp. 189-200. 48 Grieninger, G. (1983) in Plasma Protein Secretion by the Liver (Glaumann, H., Peters, T. and Redman, C., eds.), pp. 155-193, Academic Press, London. 49 Nimmer, D., Bergtrom, G., Hirano, H. and Amrani, D.L. (1987) J. Biol. Chem. 262, 10369-10375. 50 Hassan, H.J., Leonardi, A., Chelucci, C., Mattia G., Macioce, G., Guerriero, R., Russo, G., Mannucci, P.M. and Peschle, C. (1990) Blood 76, 1158-1164. 51 Schmidt, B., Wais, U., Pringsheim, W. and Kunzer, W. (1984) Eur. J. Pediatr. 141,225-227. 52 Andrew, M., Mitchell, L., Berry, R., Schimdt, B. and Hatton, M.W.C. (1988) Pediatric Res. 23, 249-252. 53 Zhu, B.C.R. and Laine, R.A. (1987) Arch. Biochem. Biophys. 25, 1-7. 54 Andrew, M., Bhogal, M. and Karpatkin, M. (1981) N. Engl. J. Med. 305, 1130-1133.