Isolation of two cDNAs encoding novel α1-antichymotrypsin-like proteins in a murine chondrocytic cell line

Gene, 106 (1991) 213-220 0 1991 Elsevier Science Publishers

GENE

B.V. All rights reserved.

213

0378-I 119/91/$03.50

06043

Isolation of two cDNAs encoding novel a,-antichymotrypsin-like line (Recombinant

DNA;

serpin;

acute-phase;

cartilage;

development;

proteins in a murine chondrocytic cell

teratocarcinoma)

John D. Inglis, Muriel Lee, Duncan R. Davidson and Robert E. Hill MRC Human Genetics Unit, Western General Hospital, Edinburgh EH4 2XiJ Scotland (U.K.) Received by J.-P. Lecocq: 6 February 1991 Revised/Accepted: 21 May/24 May 1991 Received at publishers: 10 July 1991

SUMMARY

We have isolated two novel serpin-encoding sequences from EB22, a chondrocytic cell line derived from a mouse teratocarcinoma. Both sequences fall within the Spi-2 sub-family, and are related to the gene encoding human CY l-antichymotrypsin (ACT), a major acute-phase reactant. Considerable amplification of the Spi-2 gene family in the mouse has occurred, hindering the identification of a functional equivalent of the human gene. However, one of the sequences described here, EB22/4, exhibits several features which indicate that it may represent the physiological rodent equivalent of ACT. The sequence is expressed in the liver, as expected, and is induced several-fold during the acute-phase response. The Pl amino acid residue, which is primarily responsible for inhibitor specificity, is Met rather than the human Leu, most probably a functionally conservative substitution. Analysis of the orthologous sequence in related rodents demonstrates conservation of the predicted reactive centre-encoded specificity. The second isolated cDNA, EB22/3, encodes an unexpected Cys residue at the Pl position in the reactive centre, and represents a novel sub-class of the Spi-2 serine proteinase inhibitor (serpin)encoding gene family. At least one of the sequences appears to be expressed at sites of skeletal deposition during the later stages of mouse foetal development, indicating a role for serpins during development.

INTRODUCTION

The serine proteinase inhibitor superfamily includes an increasing

(serpin)-encoding gene number of mammalian

Correspondence to: Dr. J.D. Inglis, MRC Human

Genetics

Unit, Western

Generai Hosp., Crewe Rd., Edinburgh EH4 2XU Scotland Tel. (44-3 1)332-2471; Fax (44-31)343-2620. Abbreviations: encoding

aa, amino acid(s); ACT, al-antichymotrypsin;ACT,

ACT;

strand(ed); charide;

kb,

bp, base

pair(s);

kilobase

nt, nucleotide(s);

or

DEX,

1000 bp;

encoding

dexamethasone; LPS,

bacterial

oligo, oligodeoxyribonucleotide;

reading frame; Pl, see INTRODUCTION; PCR, polymerase chain reaction; serpin, Spi, gene(s)

(U.K.)

serpins;

gene ds, double

lipopolysacORF,

open

Pl ‘, see INTRODUCTION; serine proteinase inhibitor(s);

SSC, 0.15 M NaCl/O.OlS

M Na’citrate

(pH 7.0); TBG, thyroxine-binding globulin; UTR, untranslated UWGCG, University of Wisconsin Genetics Computer Group.

region;

serum inhibitors involved in the control of proteinases central to the coagulation cascade, fibrinolysis, the complement cascade, and the acute phase response (Travis and Salvesen, 1983; Carrel1 et al., 1989). The precise structure of a proteinase-serpin complex has not yet been determined but the inhibitor appears to interact with its target proteinase in an equimolar ratio, forming a highly stable intermediate structure. This complex can gradually resolve into free proteinase and inactivated inhibitor, accompanied by cleavage of the inhibitor between the Pl and Pl’ residues, in a domain termed the reactive centre, close to the C terminus. Several recent reports have demonstrated the importance of aa within the reactive centre, and in particular the nature of the Pl residue, in determining the specificity and kinetics of the proteinase-inhibitor interaction (Owen et al., 1983; Courtney et al., 1985; Jallat et al., 1986;

214 Holmes 1990).

et al., 1987; Stephens

et al., 1988; Rubin

et al.,

1

3

2

4

5

6

7

8

9101112

A rodent multigene family of serpin-encoding genes, the Spi-2 multigene locus, has previously been reported (Hill et al., 1985; Hill and Hastie, 1987). These sequences have resulted from amplification events relative to the homologous single-copy human ACT gene to yield at least twelve tightly clustered genes in the mouse and at least four in the rat. We have shown previously that the reactive centre domain, far from being conserved as would be expected for a functionally important coding region, is hypervariable at both the DNA and protein levels in the rodent Spi-2 genes (Hill et al., 1984; Hill and Hastie, 1987; Inglis and Hill, 1991). Human ACT has been characterised at the molecular level and encodes a 66-kDa glycoprotein, primarily synthesised in the liver (Travis et al., 1978; Chandra et al.. 1983; Hill et al., 1984; Rubin et al., 1990). Its precise biological function is unknown, but is thought to include a role in the control of the acute-phase response (Travis and Salvesen, 1983). The proteinases cathepsin G and the mastcell chymases have been proposed as physiological targets of ACT, due to its rapid and efficient inhibition of the former, and likely affinity for the latter group (Travis and Salvesen, 1983). Since the acute-phase response is not specific to man, it is likely that a requirement for an ACT activity exists in other mammals, including rodents. However, despite the considerable amplification of Spi-2 subfamily sequences in the mouse, a transcript likely to perform the role of the human protein has not been described. The identification of such a gene in the mouse is hindered by the high level of expression of the anti-tryptic Spi-2 product, contrapsin, in the liver (Hill et al., 1984). We have examined the distribution of expression of Spi-2 transcripts in other mouse tissues and have determined a possible murine equivalent of the human ACT gene, as well as a second novel member of this gene family.

RESULTS

AND DISCUSSION

(a) Analysis of Spi-2 transcript tissue distribution in mouse To determine the extra-hepatic distribution of Spi-2 expression in the mouse, RNA from several tissues was analysed by Northern blot and hybridised to a contrapsin cDNA probe (Fig. 1). This confirmed as expected that by far the highest levels of Spi-2 expression in the mouse are in the liver. The contrapsin probe hybridises to several size classes of RNA in the liver, at 1.8 kb, at 2.1 kb and a high-M, form at about 4.8 kb. Low but detectable levels also occurred in submaxillary gland, spleen, and testes, at a level lOOO-fold lower than that in the liver. Of particular interest was the level of expression of a 2.1-kb transcript in the teratocarcinoma cell line EB22, at a level some five fold

Fig. I. Northern-blot analysis of Spi-2 sequences in mouse tissues. BalbiC total RNAs (10 ng) were hybridised to an .Spi-2-specific probe, pLv54 (Hill et al., 1984)

radiolabelled

800 Ci,‘mol), by nick translation

(Barth et al., 1982) and Northern-blot as previously (containing

described.

with [r-‘ZP]dTTP analysis (Meehan

RNA was electrophoresed

2.2”,, formaldehyde)

in 10 mM Na

at 60 V for 6 h, and transferred Hybridisation

to “P-labelled

filter washed

to a stringency

(Amersham;

(Rigby et al., 1977). Isolation

in 10

x

of RNA

et al., 1984) were

on a I.5 O0 agarose phosphate

SSC to a nitrocellulose

pLv54 insert was carried

gel

buffer pH 6.5 filter.

out for I6 h, the

of 2 x SSC, 6X’C, and exposed

to Kodak

XAR-5 film at -7O’C

for 72 h. Lanes: 1 and 2, liver RNA; 3 to 6, liver RNA serially diluted IO ‘, IO ‘, 5 x 10 ‘, 10 ‘; 7, submaxillary gland;

8, spleen; 9, kidney; Size

markers

10, brain; 11,testes; 12, EB22 cells (DEX induced).

indicate

transcript

size

(IXS = 1.X kb,

20s = 2.1 kb,

28s = 4.8 kb).

higher than that observed in other nonhepatic tissues. EB22 cells exhibit a chondrocytic phenotype, and may indicate a second major site of Spi-2 expression in the mouse. (b) Isolation and characterisation of Spi-2 cDNAs from the EB22 cell line To determine the nature of the Spi-2 transcript(s) cxpressed in the EB22 cell line, a cDNA library of EB22 poly(A) + RNA was constructed in the vector %gt 10. Duplicate filters containing lo5 clones were hybridised to a rat Spi-2 probe, Spi-2.1. A full-length rat cDNA was used to prevent the enhanced detection of contrapsin sequences which occur with the partial mouse cDNA probe. Five plaques hybridised to the probe in the primary screen, confirming the relatively low level of expression in this tissue. Plaque-pure Spi-2 positive clones were isolated for restriction analysis, which indicated the existence of two classes of transcript. The two largest representative clones EB22/4 and EB22/3 were subcloned into the plasmid pBluescribe for DNA sequencing. (c) A novel distinctive Spi-2 sequence from EB22, EB22/3 The EB22/3 cDNA is 1781 bp in length and contains most of the coding region and entire 3’-UTR of a novel @i-2 sequence (Fig. 2A). Comparison of the 5’ regions of the cDNA with the EB22/4 cDNA and with other fulllength @i-2 cDNAs indicate that the EB2213 cDNA lacks 250 bp of the most 5’ coding sequence, including the puta-

215 tive AUG start codon. A 50-bp region with considerable similarity to the 5’-UTR of EB22/4 is still present. This cDNA may have arisen from an alternative splicing event, since we have identified additional coding sequences in a genomic clone for this sequence (unpublished results). EB22/3 shows most similarity to a full-length published Spi-2 sequence, the rat Spi3 gene (Pages et al., 1990), with SO:‘;, nt and 68 “/, aa identity. As expected, the closest human equivalent is ACT (66”;, nt identity, 59”~~ aa identity). Strikingly, the reactive centre of the deduced protein contains the Pl-PI’ residues Cys-Cys, which have not been observed in any other serpin to date. In addition to the novel reactive centre sequence, the cDNA also encodes an unusual C-terminal extension of 28 aa not seen in other Spi-2-encoded proteins. It is possible that this sequence encodes an inhibitor specific for an as yet uncharacterised proteinase type. However, the EB22/3 product may have acquired a non-inhibitory function following the amplitication of the Spi-2 genes in rodents. Precedents for this include thyroxine-binding globulin (Flink et al., 1986) which has a defined physiological function, although no obvious anti-proteinase activity. Alternatively, the C-terminal extension encoded by EB22/3 could exhibit chemoattractant properties for neutrophils following proteolytic modification, as has been demonstrated for other serpins (Banda et al., 1988; Hoffman et al., 1989). (d) A second novel full-length Spi-2 cDNA, EB22/4 The EB22/4 cDNA is 2035 bp in length and contains the entire coding region for a novel mouse Spi-2 sequence, with a complete 3’-UTR and 60 bp of 5’-UTR (Fig. 2B). Translation of the largest ORF predicts a protein with an N-terminal signal peptide of 22 aa followed by a mature polypepFig. 2. Nucleotide cDNAs.

sequences

An EB22 cDNA

the cloning vector

IgtlO.

and

deduced

aa

library was prepared Screening

sequences

for

by standard

and characterisation

EB22

methods

in

of clones were

also standard (Maniatis et al., 1982). Subcloned Spi-2 fragments in the plasmid pBluescribe (Stratagene, La Jolla, CA) were sequenced using a dsDNA

sequencing

Sequenase

technique

II modified

cals). The UWGCG used for sequence

package

and deduced

and

Seeburg,

(United

1985)

States

et al., 1984) mounted

and analysis.

locally was

(A) The sequence

of EB22/3.

begins at nt 55, this nt corresponding

(see B). The nt from 1-53 represent

exon, and the junction

between

site is indicated

by a vertical

Cys”‘” (circled).

(B) The sequence

double

this and a cryptic

line. The predicted

and

Biochemi-

splicing event (see section c), the predicted

aa sequence,

of the EB22/4 sequence 5’-UTR

(Devereux

alignment

Due to an alternative

(Chen

T7 DNA polymerase

ORF,

to nt 292

an upstream exon 2 splice Pl residue

is

of EB2214. Based on a comparison

with the rat Spi-2.3 sequence (Yoon et al., 1987), predicted translation begins at nt 61, with a 22 aa hydrophobic signal peptide preceding the mature

protein. The predicted

Pl residue is Metjx’ (circled). The reactive

centre is boxed for both cDNAs, and potential polyadenylation underlined. Outer numbering on left oftigures is for nt sequence, numbering

refers to deduced

submitted

to the GenBank-EMBL

Nos. M64085

and M64086.

aa sequence.

These sequences

sequence

database

sites are indented have been

with accession

216 tide of 396 aa. The predicted reactive centre domain contains a Met-Ser dipeptide in the Pl-Pl’ position. EB22/4 shows 84% nt identity and 77% aa identity to EB22/3, and 86% nt identity and 79”/, aa identity to both the rat Spi-2.2 and Spi3 cDNA sequences (Hill and Hastie, 1987; Pages et al., 1990). Interestingly, the two rat sequences differ by a single aa at the Pl residue (Val in Spi-2.2, Met in Spi3) and may represent either allelic variation in the rat population, or a recent gene duplication. Again, as expected for a Spi-2 sequence, the human serpin most similar to EB22/4 is ACT (683, nt identity and 61 Y;, aa identity). (e) The Spi-2.2-reactive

centre

is conserved

in several

rodents To determine whether the EB22/4 sequence was conserved in other rodents, a cDNA library from Apodemus sylvaticus liver was constructed and screened in the same way as the EB22 library. One Spi-2 cDNA, A12, bore a strong resemblance to both the rat Spi3 and mouse EB22i4 sequences, with the notable exception of a Leu in the Pl position. An alignment of the deduced aa sequences for the last coding exon of the cDNAs from mouse, rat and Apodemus is shown (Fig. 3A). These genes show X5-87”,> identity to one another at the DNA level, and 77-8 lo;, at

the protein level over this region which includes the reactive centre. In contrast, the EB22/3 sequence shows much less similarity (7 l-73 “/, nt identity and 60-66 :; aa identity) to the three Spi-2.2 sequences over the same region. This is primarily due to divergence within the EB22/3 reactive centre region. Other rodent Spi-2 sequences show an even more pronounced lack of relatedness to the Spi-2.2 group, and to each other, in the reactive centre region (Fig. 3B). The relative homogeneity of the EB22/4, rat Spi3, and Apodemus Al2 reactive centre sequences further supports the idea that these genes may be orthologous, and we suggest that each gene should be given the genetic designation SIG-2.2. It is of interest that the predicted Spi-2.2 Pl residue is variable (Leu, Met or Val) in rodents, although conserved in terms of predicted target proteinase specificity. However, recent analysis of recombinant human ACT involving sitedirected mutagenesis of the Pl residue indicated that a Leu to Met substitution caused little change in inhibitory specificity or efficiency (Rubin et al., 1990). (f) Hepatic expression of Spi-2.2 and its induction during acute phase To demonstrate hepatic expression of the mouse Spi-2.2 (EB22/4) transcript, and examine its behaviour during the

A m2.2 a2.2 r2.2

VVHKAVLDVAETGTEAAAATGV

FNRPFLIMIFDTETEIAPFIAKIANPK

__________-K----___--_________-------------

_-_-_-VA-Y------IY-LG--S---D-_--MI-S----A----L---F---

B EB/3 r2.1 m2.1 r2.3

_--------______-w-_--M -------__D____--T_____ ________----________-_ -----____D____-G----A_

----__M~_S__~H__L_M__VT__E -----MVF-T-MDSQSIL-V---TM-S ------FV-YH-SAQSIL-M--VN-------MLV-T-NNGQSVF-MG-VT--M

C 2.2~ hACT Fig. 3. Alignment sequences

VVHKAVLDVAEtGTEAAAATGV fVPmSAKLdPLtiy. ********* x **** *** * ~I****** x VVHKAVLDVFEEGTEASAATAV KITLLSALVETRTIVR

of C-terminal

Spi-2.2 sequences

from mouse (m), rat (r) and Apodemus

are shown aligned to the mouse Spi-2.2 (EB22/4)

sequence.

The reactive

FnRPFLxxIxDTETEIAPFIAKIxNPK * * * * * *** FNRPFLMIIVPTDTQNIFFMSKVTNPK (a). (A) Deduced

aa sequences

centre is boxed, and the Pl residue

from the three rodent is indicated

Spi-2.2

by an exclamation

mark. Dashes represent identical residues to the mouse sequence. (B) Other published rodent Spi-2 (but not Spi-2.2) sequences shown as a separate group aligned to the mouse sequence. The absence of reactive centre conservation in these sequences, relative to the Spi-2.2 sequences (A) is striking. Dashes represent identity and dots represent deletions relative to the mouse sequence. with human ACT. Asterisks represent identity with human ACT, the dot represents

(C) A consensus sequence for the three Spi-2.2 sequences, aligned a deletion, lower-case letters represent variable aa residues, and

x represents

EB/3 = EB22/3cDNA (all from present data); r2.3 = rat

no consensus.

m2.2 = mouse Spi-2.2 (EB22/4);

(Yoon et al., 1987); r2.2 = rat Spill (Pages 1987).

a2.2 = Apodemus Spi-2.2;

et al., 1990); r2.1 = rat Spi-2.1;

m2.1 = mouse contrapsin;

hACT = human

Spi-Z..1

ACT (all from Hill and Hastie,

217 acute-phase

response,

an EB22/4-specific

anti-sense

oligo

was employed. Northern-blot analysis was performed on liver RNA from mice that had been treated with bacterial lipopolysaccharide to induce the acute-phase response. RNA from EB22 cells treated with the inflammatory agent dexamethasone was also included (Fig. 4). The analysis showed pronounced induction of the EB22/4 transcript following inflammatory stimulus. Densitometric analysis (Elder et al., 1986) of the induction level revealed a six- to sevenfold increase of the S’G-2.2 transcript in mouse liver 24 h after induction of acute phase. A less dramatic, but distinct increase was also observed for the dexamethasonetreated EB22 cells. In man, serum ACT protein levels double within 8 h of inflammatory insult (Travis and Salvesen, 1983), and the rat Spi-2.2 mRNA level increases dramatically 24 h after acute-phase induction (Hill and Hastie, 1987). All other rodent Spi-2 sequences we, and others, have examined showed no increase, or a slight decrease during the acute-

12345678910 A 20s

I3

Fig. 4. Induction

of mouse Spi-2.2 by DEX and acute phase reactants.

(Panet A) Northern

blot probed

Lanes:

funinduced)

(induced

1, C5’7BL/6

with the oligo 224, specific for EB22/4. total

liver RNA

RNA (10 pg); 4, EB22 (DEX-induced) (uninduced)

poly(A)’

ly(A) + RNA

(2 pg);

8, C57BL/6 induced)

(LPS

RNA

induced)

poly(A)’

RNA

pGAPDHl0

(Edwards

RNA concentration

RNA

(5 pg);

10, C57BL/6

phosphate-buffered

induced)

po-

blot reprobed

with

to quantitate

mouse liver RNA was

injection of adult C57BL/6 animals with Esche(50 pg; serotype 0127 : B8; Sigma) in 50 ~1 saline (137 mM NaC1/2.7 mM KCl/iO mM

mM K. phosphate

cells in culture was achieved

pH 7.3). DEX

by addition

induction

using

[y-s’P]ATP

(Boehringer-Mannheim, gelj2.2%

formaldehyde)

(Amersham) U.K.).

RNA

and antoradio~aphy

to the nt was end-

and T4 polynucleotide electrophoresis

and Northern-blot

Fig. 1 legend, except for oligo 224 washing EB22/4 transcript.

of EB22

of DEX to 1 PM and harvesting

after 48 h. An IS-mer antisense ohgo, 224, complementary encoding the EB22/4 reactive centre residues VPMSAK, labelled

(un-

induced)

transcript,

Acute-phase

po(5 pg);

9, C57BL/6

(LPS

B) The same Northern

in each sample.

by subcutaneous

Na phosphate/2

poly(A)’

(LPS

poly(A)‘RNA

et al., 1985), a control

richia coli lipopolysaccharide

of

6, C57BL/6

(uninduced)

(10 pig);

total

total RNA (10 pg); 5, C57BL/6

(2 pg);

7, C57BL/6

ly(A)+ RNA (IO pg). (Panel

obtained

(10 pg); 2, C57BL/6

with LPS) total liver RNA (10 fig); 3, EB22 (uninduced)

analysis stringency

(nine days). 20s indicates

(1.2%

kinase agarose

were as described

in

(4 x SSC, 54”(Z),

the location

of the 2.1-kb

phase response (Hill et al., 1985; Hill and Hastie, 1987; Pages et al., 1990). Thus the rodent Spi-2.2 transcript is unique in resembling human ACT in tissue distribution, predicted reactive centre specificity, and inducibility following inflammation. (g) Spi-2

sequences are expressed

in cartilage

during

mouse development

As part of the analysis of Spi-2 expression in the mouse, in situ hybridisation of Spi-2 probes was carried out using 16-day-old foetuses. This was done to circumvent the high level of contrapsin expression in the liver, which is not detectable until 17 days of development (Meehan et al., 1984). Probes specific for individual Spi-2 family members were not technically possible at this stage, and so a conserved region of the contrapsin cDNA was employed at moderate stringency to detect all expressed Spi-2 transcripts. This analysis revealed that the main sites of Spi-2 expression in the l&day-old mouse foetus are in regions of skeletal cartilage deposition throughout the animal (Figs. 5 and 6). Expression is absent from adjacent regions where ossification appears to be more advanced. The expression pattern is consistent with the timing of ossification in the mouse foetus (Theiler, 19891, and correlates with the initial isolation of the EB22/3 and EB22j4 sequences from a chondrocytic cell line. Expression was also observed in certain tissues which remain cartilaginous, such as the nasal septum (not shown). This may indicate a role for the maintenance of cartilage in differentiated skeletal systems. Support for this idea comes from the observation that the proteinase cathepsin G, a probable target of human ACT, degrades proteoglycan and fibronectin (Travis and Salvesen, 1983). We have not distinguished between expression of EB22/3 and EB22/4 at this developmental stage, due to their high conservation outside the reactive centre, but both are expressed at equivalent low levels in the EB22 teratocarcinoma cell line. The future use of sequence-specific oligos for in situ analysis may reveal specific expression of one or other sequence in embryonic cartilage. (hf ~onelusions {I ) We report the isolation of two novel serpin sequences from a mouse chondrocytic cell line, EB22. The cell line was selected as a source of novel Spi-2 sequences as tissue surveys indicated that it represents a major site of Spi-2 gene expression in the mouse. (2) The first, EB22/3, is about 80% similar to other published rodent Spi-2 sequences, but contains the biochemically novel Cys-Cys dipeptide at the Pl-Pl ’ residues of the reactive centre. The potential in vivo targets of this inhibitor are not known, but the conservation of structural motifs flanking the reactive centre indicate that it may have a defined physiological role.

B

219 Courtney, M., Jallat, S., Tessier, L.H., Benavente, A., Crystal, R.G. and Lecocq, J.-P.: Synthesis in E. coli of al-antitrypsin variants of thera-

(3) The second cDNA, EB22/4, is a more orthodox Spi-2 sequence which may represent the rodent equivalent

peutic potential

of the human ACT gene. It is expressed in the liver in addition to EB22 cells, and is induced sevenfold during the inflammatory response. The sequence, which we have named Spi-2.2, is well conserved in other rodents, albeit with a conservatively altered Pl residue. Conservation of reactive centre residues flanking the Pl site suggest that these may be important in defining the proteinase target of

for emphysema

development ment

ofthe neural tube and vertebrae

J., Haeberli,

sequence

F. and Smithies,

analysis

programs

Y.H., Lloyd, J.C., McMillan,

zyme levels in developing sequence Nucleic

teases.

tally regulated

J. Biol. Chem. 263 (1988) 448 l-4485. L.C. and Hastie, N.D.: Developmen-

mRNAs

of multiple Hoffman,

adaption Chandra, S.L.C.:

Sequence

R., Kidd,

homology and

Fig. 5. In situ analysis

between

III.

K.J.H.

RNA transcripts.

Harbour

plasmid

sequencing:

DNA.

of Spi-2 expression

Swiss mouse embryos

with a synthetic

[a-‘5S]UTP

are magnified

in lh-day-old

were sectioned

is a general

mouse

an antisense

Spi-2 sequence

x 56. (Panel A) Light-field

foetus.

in the sagittal

products

Sections

(left) and dark-field

F.C.: Heparin

exhibit neutrophil

mutagenesis

of the reactive

site. Biochemistry

L.T. and Dayhoff, ovalbumin, Biophys.

S., Carvallo,

obtained

M.O.: A surprising antithrombin

Res. Commun.

D., Tessier,

cofac-

chemo-attrac-

R.G. and Courtney,

engineered

rl-antitrypsin

26 (1987) 5133-5140. superfamily

III, and r 1-proteinase

inhibitor.

95 (1980) 864-871. Spi-2 proteinase

reactive

inhibitor

site domain.

D., Roitsch,

M.: Altered specificities

variants.

E.S.: Amplification

ofrecombiby site directed

new protein

L.H., Roecklin,

F., Crystal,

This procedure

and transverse

pLv54 sequence probe.

locus

the level

Protein

Engin.

C., Ogushi, ofgenetically

1 (1986) 29-35.

of RNA. In: Innis, M.A., Gelfand,

J.J. and White,

T.J. (Eds.),

Applications.

PCR Protocols:

Academic

Press,

locus: EMBO

D.H.,

A Guide

San

Diego,

A., Le Cam, G., Leopold,

to

1990,

P., Benarous,

R. and Glaichenhaus, N.: Study of a growth hormone-regulated protein secreted by rat hepatocytes: cDNA cloning, anti-protease activity

DNA 4 (1985) 165-170.

RNA probe encoding

this sequence

a fast and simple

N.D.: A genetic

Blood 73 (1989) 1682-1685.

Methods and pp. 2 l-27.

(1983)

N.D.:

Mol. Cell. Biol. 5 (1985) 2114-2122.

and of mutants

Sninsky,

al-antichymotrypsin, 22

reaction

centre

within the

gene complex controls

a multigene family with a hypervariable J. 10 (1991) 255-261.

and Woo,

Biochemistry

H. and Hastie,

R.K. and Hastie, inhibitor

Le Cam, A., Pages, G., Auberger,

Sixteen-day-old

photographs

human

P.H.: Supercoil

for sequencing

(2 x SSC, 65”C),

Spring

326 (1987) 96-99.

&antiplasmin

Kawasaki,

V.J., Robson,

antithrombin

5055-5061. Chen, E.Y. and Seeburg, method

Cold

in the reactive

Nature

P.A., Baumann,

Inglis, J.D. and Hill, R.E.: The murine

Biol. LII (1987) 527-535.

T., Stackhouse,

rl-antitrypsin

inhibitors.

to the serine anti-pro-

evolution

nant human

Jallat,

E.:

binding globulin

W.E., Lijnen, H.R. and Collen, D.: Characterisation

Biochem.

P.A. and Boswell, D.R.: The serpins: evolution

scanner.

B.E. and Morkin,

M., Pratt, C.W., Brown, R.L. and Church,

comprising

in mouse liver. Proc. Natl. Acad. Sci. USA 79

in a family of protease

Symp. Quant.

Boyd,

P.H., Barth,

tor II-proteinase

(1982) 500-504. Carrell, R.W., Pemberton,

digital

thyroxine

close homology

inhibitors.

closely linked to a protease

inacti-

elastase.

Markham,

of human

N.D.: Accelerated

P.H.,

Hill, R.E., Shaw,

R.M.: xl-proteinase

K.W., Gremke,

reading of DNA

Plasma protease inhibitors in mouse and man: divergence reactive centre regions. Nature 3 11 (1984) 175-I 77.

Hunt,

vation by macrophage

levels and en-

Sci. USA 83 (1986) 7708-7712.

of serine protease

Hill, R.E., Shaw,

REFERENCES

Barth, R.K., Gross,

F.J.: Human

mRNA

E.M.: Automatic

Gustafson,T.A.,

from cloned cDNA:

tant activity.

proteolytic

S.L. and Benham,

using a large format

amino acid sequence

Proc. Natl. Acad.

regions

Holmes,

after

Res. 12

Acids Res. 14 (1986) 417-424.

Hill, R.E. and Hastie,

We thank Dr. John Ansell for assistance in obtaining Apodemus specimens, Dr. Martin Hooper for providing the EB22 cell line, Elizabeth Graham for assistance with in situ hybridisation, Dr. Daryll Green for quantification of Northern-blot signal, Dr. Richard Meehan for isolation of EB22 RNA, Norman Davidson, Sandy Bruce and Douglas Stuart for photographic work, and Dr. Nick Hastie for helpful discussions.

chemoattractant

set of

Acids

muscle. Mol. Cell. Biol. 5 (1985) 2147-2149.

gel autoradiographs

deduced

is a neutrophil

0.: A comprehensive

dehydrogenase:

Elder, J.K., Green, D.K. and Southern,

ACKNOWLEDGEMENTS

inhibitor

in the mouse. Develop-

for the VAX. Nucleic

glyceraldehyde-3-phosphate

Complete

G.L. and Senior,

313 (1985)

(1984) 387-395. Edwards,

Flink, I.L., Bailey,T.J.,

M.J., Rice, A.G., Grifftn,

Nature

104 (1988) 305-316.

Devereux,

this serpin. (4) Detection of Spi-2 expression at sites of ossification in the developing mouse skeleton suggests a potential role for this class of proteinase inhibitor in the control of chondrogenesis, and formation of skeletal elements.

Banda,

and thrombosis.

149-151. Davidson, D.R., Graham, E., Sime, C. and Hill, R.E.: A gene with sequence similarity to Drosophila engrailed is expressed during the

followed

the technique

planes to provide from nt 98-720

maximal

described

(Hill and Hastie,

were autoradiographed

et al. (1987).

Sections

were probed

1987). At the stringency

for two to three

(right) image of a section through

by Davidson

tissue coverage.

the developing

weeks

before

hind-paw

employed

developing.

All

region. The internal

skeletal structure is obvious. Grains indicating the location of Spi-2 transcripts are clearly located over regions internal to the developing bones and appear to correlate with the presence of a cartilaginous cell type. (Panel B) Light-field (left) and dark-field (right) image of a section through the mouse forelimb at the elbow joint. The junction

between

humerus

and lower limb bone can be clearly seen in the light-field

ends of the developing bones, but is absent at the periphery to that seen for the positive cells in panel A.

of the joint surfaces.

The loosely clustered

figure. Hybridisation

appearance

is most obvious

of the expressing

at the

cells is very similar

220 and regulation

of its synthesis

by various

hormones,

EMBO

J. 6

(1987) 1225-1232. Maniatis,

T., Fritsch,

tory Manual. NY, 1982.

E. and Sambrook,

Cold Spring Harbor

J.: Molecular Laboratory,

Cloning. A LaboraCold Spring Harbor,

D.P., Hill,R.E.,

Hogan,B.L.M.

Pattern of serum gene expression in murine foetal liver. EMBO J. 3 (1984) 1881-1885.

and Hastie, N.D.:

visceral

yolk sac and

Proc. Nat]. Acad.

Sci. USA 80 (1980) 2196-2200.

Owen, M.C., Brennan, S.O., Lewis, J.H. and Carrel], antitrypsin to antithrombin: rl-antitrypsin Met > Arg). A fatal bleeding disorder.

R.W.: Mutation of Pittsburgh (358

New Engl. J. Med. 309

J.L., Le Cam, G., Mariller,

characterisation

(1983)

M. and Le Cam, A.:

of three rat liver serine-protease

tors affected by inflammation

and hypophysectomy.

inhibi-

Eur. J. Biochem.

P.A., Stein, P.E., Pepys, M.B., Potter, J.M. and Carrell, R.W.: binding globulins

inflammation.

Nature

undergo

serpin conformational

change in

336 (1988) 257-258.

Rigby, P.W., Dieckmann, C., Rhodes, C. and Berg, P.: Labelling deoxyribonucleic acid to high specific activity in vitro by nick translation with DNA polymerase Rubin.

H.,

Finch, J.T., Turnell, W.G., McLaughlin,

R.W.: Crystal

centre of serpins. A.W.,

Siddiqui,

genesis ofthe reactive

Wang,

Z.,

J. Biol. Chem. 265 (1990)

structure Nature

of ovalbumin C.H.W.:

Nickbarg,

E.B.,

McLarney,

S., Naidoo,

N.,

Site-directed

centre (serine 394) ofantithrombin

muta-

III. J. Biol.

Chem. 263 (1988) 15489-15852. K.,

Deyashiki,

Yamamoto, human Takahori,

Y., Nishioka,

S. and Hashimoto,

protein

C inhibitor.

H. and Sinohara,

trypsin Theiler,

inhibitor.

K.: The

Springer-Verlag, action

J., Kurachi,

K.,

S.: Characterisation

Akira,

M.,

of a cDNA

for

J. Biol. Chem. 262 (1987) 61 l-616. H.: Mouse

plasma

of ccl-antitrypsin

trypsin

inhibitors:

and contrapsin,

iso-

a novel

J. Biol. Chem. 257 (1982) 2438-2446.

House

Mouse.

New York,

with chymotrypsin

5651-5665. Travis, J. and

Salvesen,

Annu. Rev. Biochem.

Atlas

of Embryonic

rl-antichymotrypsin:

like proteinases.

G.S.:

Development.

1989.

Human

plasma

Biochemistry proteinase

inter17 (1978) inhibitors.

52 (1983) 655-709.

Yoon, J.-B., Towle, H.C. and Seelig, S.: Growth mRNA

I. J. Mol. Biol. 113 (1977) 237-251.

P.J.

as a model for the

347 (1990) 99-102.

A. and Hirs,

Travis, J., Bowen, .I. and Baugh, R.: Human

190 (1990) 385-391. Hormone

c(l-antichymotrypsin.

lation and characterisation

G., Rouayrenc,

Pemberton,

reactive

Suzuki,

694-698. Molecular

human

and Carrel], Stephens,

Ohkubo, H., Kageyama, R., Ujihara, M., Hirose, T., Inayama, S. and Nakanishi, S.: Cloning and sequence analysis of cDNA for rat angio-

Pages,

and variant 1199-1207.

Stein, P.E.. Leslie, A.G.W.,

Meehan,R.R.,Barlow,

tensinogen.

Schoenberger, O.L., Johnson, J.L. and Cooperman, B.S.: Cloning, expression, purilication, and biological activity ofrecombinant native

species of the serine proteinase

hormone

inhibitor

liver. J. Biol. Chem. 262 (1987) 4284-4289.

induces

two

gene family in rat