THEJOURNALOF BIOLOGICAL CHEMISTRY Vol. 264. No. 6 , Issue of February 25, pp.

3066-3071,1983
01989 by The American Society for Biochemistry and Molecular Biology, Inc Printed in U.S.A.

CpG Mutations in the Reactive Site of Human C i Inhibitor*

(Received for publication, September 1, 1988)

Karen SkriverS, Elzbieta Radziejewska, Jane A. Silbermann, Virginia H. Donaldson$,and

Susan Clark Bockll
From the Departmentof Microbiologyllmmunology, Temple Uniuersity School of Medicine and the Thrombosis Research Center,
Philadelphia, Pennsylvania 19140 and the §Children’s Hospital Research Foundation, Cincinnati, Ohio45299

Ci inhibitor plays an important role in the regulation plexes with them. The reactive site of a serpin contains an
of vascular permeability through its ability to inacti- amino acid sequence which is an ideal substrate for its target
vate enzymes which release polypeptide kinins. Dys- protease(s). The reactive site is located on a loop protruding
functional Ci inhibitor molecules are present in the from the surface of the inhibitor molecule (21,22). In contrast
plasma of affected members of the Da and Ri heredi- to events associated with proteolysis of true substrates, serpin-
tary angioneurotic edema kindreds. We constructed protease complexes dissociate extremely slowly, and the pro-
genomic libraries from Da and Ripatient DNAs which tease remains inactive while engaged in the complex (6, 23-
had been cleaved with BclI to generate a fragment 26). The amino acid preceding the peptide bond of the serpin
containing 21 kilobases of the C1 inhibitor locus. Ci that is cleaved during complex formation iscalled the P1
inhibitor gene-containing recombinants originating residue and is an importantdeterminant of serpin target
from mutant Da and Ri alleles weredifferentiated from
those derived from normal alleles by linkage analysis protease specificity. In protease-inhibitor complexes, the P1
using the intragenic HgiAI restriction fragment length residue is probably ester-bonded to the active siteserine
polymorphism. Nucleotide sequencing of the complete residue of the target protease (27).
protein-coding regions of the mutant alleles identified Humans have a single gene for C i inhibitor which maps
two different mutations in a CpG dinucleotide corre- near the centromere on chromosome 11 (8, 9). C i inhibitor
sponding to the first two bases of arginine codon 444. deficiency is inherited as the autosomal dominant disorder
These single base mutations changed the identity of the hereditary angioneurotic edema (HANE)’ (28). HANE pa-
functionally critical P1 reactive site residue from ar- tients suffer from self-limited episodes of localized, enhanced
ginine to cysteine (Da) or histidine (Ri). The additional vascular permeability, particularly of subcutaneous tissue,
cysteine residue in C i inhibitor Da suggests how it is airways, and gastrointestinaltract (29). In addition to reduced
covalently bound to albumin in plasma. The presence levels of serum C i inhibitor, their Cz and Cx levels are also
of CpG dinucleotides in the codons specifying the P1 decreased due to consumption by uninhibited, activated C i
arginines of C i inhibitor and antithrombinI11 explains (30, 31). Individuals with HANE are heterozygous at the C i
the high incidence of histidine and cysteine substitu- inhibitor locus and have one normal and one abnormal C i
tions observed among dysfunctional mutants of these inhibitor gene. Although all persons with HANE are deficient
serine protease inhibitors. in C i inhibitor function, the molecular defects are different
in different families (32-34). Most kindreds exhibit parallel
reduction of Ciinhibitor functional activity and antigenlevels
(type I). However, in 15-25% of HANE pedigrees, decreased
The human plasma proteinase inhibitor C i inhibitor plays functional activity is found in thepresence of a dysfunctional
an important role in the regulation of vascular permeability inhibitor synthesized from the mutant allele (type 11). Dy?
through its ability to inactivate enzymes which release poly- functional gene products often represent the predominant C1
peptide kinins. It is the most important physiological inhibitor inhibitor species in type I1 patient plasma due to consumption
of plasma kallikrein and factor XIIa (1-5) and theonly known (complex formation and clearance) of the normal gene product
inhibitor of C i , the first component of complement (6).
(35-37). C i inhibitor genes encoding three type I1 dysfunc-
C i inhibitor is a member of the serpin (serine protease tional proteins did not exhibit major rearrangements com-
inhibitor) gene family (7-11) which also includes the genes pared to the normal C i inhibitor gene (8);however, electro-
for other plasma protease inhibitors such as al-antitrypsin, phoretic abnormalities and variability in the patterns of re-
antithrombin I11 (12, 13), heparin cofactor I1 (14), az-anti- sidual protease inhibitory activity of Ciinhibitor preparations
plasmin (15), protein C inhibitor (16), and the placental (17) from eight different type I1 families have been noted (38,39).
and endothelial (18-20) type plasminogen activator inhibi- Identification of the mutations responsible for type I1 hered-
tors. Serpins are suicidal proteins which inhibit their target itary angioneurotic edema will improve understanding of C i
proteases by forming stoichiometric protease-inhibitor cam- inhibitor structure/function relationships.
* This work was supported in part by United States Public Health Dysfunctional C i inhibitors Da and Ri have been studied
Service Grants HL-30712 (to S. C . B.) and HL-15690 (to V. H. D.). previously. Ci inhibitor Da binds to albumin (32) and con-
The costs of publication of this article were defrayed in part by the tains a free sulfhydryl group (40). C i inhibitor preparations
payment of page charges. This article must therefore be hereby from Da patient plasma had residual proteaseinhibitory
marked “aduertisement” in accordance with 18 U.S.C. Section 1734 activities against Cis, kallikrein, activated Hageman factor,
solely to indicate this fact. Hageman factor fragments, and plasmin which were 75, 50,
$ Supported by NationalInstitutes of Health-Fogarty Interna-
tional Fellowship TW-04025 and by the Danish Natural Science
Research Council and the Danish Research Academy. The abbreviations used are: HANE, hereditary angioneurotic
n Supported by American Heart Association Established Investi- edema; RFLP, restriction fragment length polymorphism; kh, kilo-
gatorship 880261. base(s); serpin,serine protease inhibitor.

3066
 CpG Mutations
Reactive
thein Site
70,194, and 20%, respectively,of the inhibitory activities
displayed by an equivalent amount of normal C i inhibitor.
Preparations of C i inhibitor from Ri patient plasma had
negligible inhibitory activity against any of these proteases
(38).
An HgiAI restriction fragment
lengthpolymorphism
(RFLP) in the C i inhibitor gene produces 0.7- and 0.4-kb
bands o n Southern blots (8). This polymorphic marker was
used to identify mutant C i inhibitor alleles inthe Da and R i
pedigrees and to distinguish them from the normal C i inhib-
itor alleles also present in recombinant librariesof heterozy-
gous patient DNA. Sequencing of the mutant C i inhibitor
genes identified single base changes relative to the normal
gene in both. These mutations, which have occurred i n a
single CpG dinucleotide encoding the first two positions of
arginine 444 codon, convert the reactive site P1 residue of C i
inhibitor to a cysteine in affected members of the Da family
a n d to a histidine in affectedmembers of the R i family.
MATERIALS AND METHODS~
Determ&ztion of C i Inhibitor Functional Activity and Antigen
Leuels-C1 inhibitor functional activity levels in serum samples were
determined by the method of Levy and Lepow (41) using the estero-
lytic substrate N-acetyl-L-tyrosine ethyl ester. Antigen levels were
measured by rocket immunoelectrophoresis (42) or radial immuno-
diffusion (43).
Preparation of Genomic DNA-Peripheral blood wasobtained from
several members of each family, and genomic DNA was isolated as
described previously (44).
Southern Blot Analysis-Three-pg samples of restriction endonu-
clease-treated genomic DNA were separated by agarose gel electro-
phoresis and transferred to nylon membranes (Gelman), which were
then hybridized to 32P-ni~ktranslated C i inhibitor cDNA probe
fragments (8).In order to avoid electrophoresis and blotting artifacts,
aliquots 0-f chicken genomic DNA (which does not hybridize to the
human C1 inhibitor cDNA probe) were added to samples of recom-
binant phage DNA used for side-by-side Southern blot analysis with
normal and patientgenomic DNAs.
Genomic Cloning-Family studies using the HgiAI restriction frag-
ment length polymorphism (8) were employed to identify and mark
mutant C1 inhibitor alleles in the Da and Ri families. Recombinant
phage carrying the mutant alleles were obtained as follows. Genomic
DNA samples (-45 pg) from HANE patients Da-11-2 and Ri-1-1, who
are heterozygous for the HgiAI RFLP, were digested to completion
with BclI. 18-23-kb fragments were isolated by centrifugation through
10-40% sucrose gradients, and one-fifth of the material from the
appropriate fractions was inserted into the BamHI site ofXDash
(Stratagene). In this manner, libraries of approximately 200,000
plaque-forming units were generated from size-fractionated material
obtained from an initial 9 pgof patient DNA. C i inhibitor gene-
containing recombinants were identified by hybridization to the 32P-
labeled C i inhibitor cDNA probe. HgiAI RFLP analysis of the posi-
tives indicated whether they originated from a normal or mutant
allele, as defined by linkage analysis.
Sequencing-The protein-coding portion of the human C i inhibi-
tor gene consists of seven exons interrupted by six introns (45).
Fragments of the mutant Da allele were subcloned into pUC vectors
(46). Fragments of the mutant Ri allele were subcloned into Genes-
cribe-Z vectors (United States Biochemicals Corp.). Linearized dou-
ble-stranded DNA was sequenced by the dideoxy chain termination
method (47) using Sequenase” (United States Biochemicals Corp.).
“Universal” primers (corresponding to vector sequences near poly-
linker insertion sites) and C i inhibitor-specific primers were used to
sequence the seven protein-encoding Exons in both directions. Se-
quences of 15-and 16-nucleotide-long_Clinhibitor primerswere based
on intron sequences of the normal C1 inhibitor gene (see Miniprint).
* Portions of this paper (including part of “Materials and Methods,”
part of “Results,” part of “Discusison,” and Figs. 4 and5) are
presented in miniprint at the end of this paper. Miniprint is easily
read with the aid of a standard magnifying glass. Full size photocopies
are included in the microfilm edition of the Journal thatis available
from Waverly Press.
of Human
Ci‘lnhibitor 3067
RESULTS
Identification of Alleles Coding for Dysfunctional Ci Inhibi-
tor Da and Ri Proteins-Ci inhibitor deficiency is inherited
i n an autosomal dominant manner,and hereditary angioneu-
rotic edema patientsare heterozygotes for a mutant (or null)
C i inhibitorgene.TheHgiAI RFLP was used to follow
inheritance of individual C i inhibitor alleles i n the Da and
R i HANE kindreds and to identify those carrying abnormal
genes. The studies shown inFig. 1 indicate that the dysfunc-
tional Da gene resides on a chromosome carrying the 0.7-kb
allele of the HgiAIRFLP, whereas the dysfunctional Ri gene
is associated witha chromosome carrying the 0.4-kb allele.
Cloning and Sequencing-Genomiclibrarieswere con-
structed from BclI digests of Da-11-2 or Ri-1-1 DNAs and
BamHI-digested XDash vector DNA. These two individuals
were chosen as the source of DNA for cloning experiments
because they have hereditary angioneurotic edema and are
heterozygotes for the HgiAI RFLP linkage marker (Fig. 1).
Three Da and six Ri recombinant phage carrying C i inhib-
itor gene inserts were isolated, and their HgiAI polymorphism
types were determined. Twoof the Da clones carriedthe 0.7-
kb HgiAI RFLP allele (which is associated with HANE i n the
Da kindred), and three of the R i clones carried the 0.4-kb
HgiAI RFLP allele (which is associated with HANE i n the R i
kindred). Side-by-side Southern blot analysis of cloned phage
DNA and patient and normal genomic DNAs reassured t h a t
gross rearrangement ofthe Ciinhibitor genehad not occurred
A
I I m-p1 2
11
1 2
B
DA RI
C 10.9 0 9.1 0 8.41.6 0
D 26 51 22 4 26.9
16.8 21.8
FIG. 1. Inheritance of hereditary angioneurotic edema and
HgiAI RFLP in Da and Ri kindreds. A, pedigrees of the Da and
Ri families. Solidsymbols represent HANE patients, and striped
symbols represent unaffected familymembers. B, Southern blots
prepared with HgiAI-digested genomicDNAhybridized to a 3zP-
labeled 500-base pair EcoRI fragment from the 3’-end of human C i
inhibitor cDNA. The region of the blot containing the 0.7- and 0.4-
kb HgiAI polymorphic fragments is shown. In these linkage studies,
the abnormal C i inhibitor gene segregates with the 0.7-kb HgiAI
marker in the Da family-and with the 0.4-kb HgiAI marker in the Ri
family. Dysfunctional C1 inhibitor genes were isolated from Da-11-2
and Ri-1-1,who are heterozygous for-the HgiAI marker polymor-
phism. C, functional activity levels of C1 inhibitor in seradetermined
by esterolytic assay with N-acetyl-L-tyrosine ethyl ester. Normal
range for pooled sera in this assay is 6-10 units/ml. D , C1 inhibitor
antigen levels in serum (mg/dl). Antigen levels weredetermined using
the method of Laurel1 (42) for the Da family and by the method of
Mancini et al. (43) for the Ri family. Normal range for pooled sera in
these assays is 11-22 mg/dl. Patient sera used for functional and
immunological assays were obtained prior to treatment with andro-
gens.
 3068 of Human Ci Inhibitor
CpG Mutations in the Reactive Site
E

u IV

FIG. 2. Subcloning and sequenc-

ing strategy for protein-encoding
exons 11-VI11 of Ci inhibitors Da
and Ri. Ayrowheads indicate universal
( U ) or C1 inhibitor gene-specific se-
quencing primers (seeFig. 5 for detailed
description of primers). The templates V VI
for universal primer sequencing of exon VI I
U*
deletions
V were SstI of the 4.5-kt, EcoRI -U U*
fragment shown. E, EcoRI; S , SstI; H , C 4 . b -U
HindIII; B, BanHI; N, NcoI; P, PstI. '1C5a
The scale bar indicates 1 kb.

P3 P2 P I " P1' PZ' P3' acid, GAG), whereas a C (glutamine,CAG) had been reported
in that position of the cDNA. Que and Petra (11) and Carter
Normal: GTG GCC ACC CTG
CTG
et al. (45) also report a glutamic acid (encoded by GAA and
Val - Ala - Arq - Thr ~ Leu - Leu GAG triplets, respectively) in this position.
DISCUSSION

Da: GTG GCC TGC ACC

CTG CTG An HgiAI restriction fragment length polymorphism, pro-
Val - Ala - cys - Thr ~ Leu - Leu ducing 0.7- and 0.4-kb hybridizing bands, was used to follow
the inheritance of dysfunctional C i inhibitor genes in two
unrelated type I1 HANE kindreds, the Da and Ri families.
The results from these studies are consistent with linkage of
a mutant gene to the 0.7-kb allele of the HgiAI RFLP in the
FLG.3. Nucleotide and derivedamino acid sequences of P1 Da family and with linkage of a mutant gene to the 0.4-kb
residue and surrounding reactive region site of normal human type HgiAI allele in the Ri family. This information was used
C i inhibitor and C i inhibitors Da and Ri. The CpG dinucleotide to isolate the mutant genes from hereditary angioneurotic
in the P1 arginine codon and thesingle base mutations and resulting
amino acid substitutions of the mutant inhibitors are indicated in edema patients Da-11-2 and R1-1-1.
boldface letters. The arrowhead shows the bondcleaved by Cls during In the course of this work, an efficient method for repetitive
complex formation. cloning of the human C i inhibitor gene was developed. An
ideal method for isolating mutant C i inhibitor genes should
during library construction or propagation of the DNA in satisfy several criteria. First, it should be efficient, so as to
bacteriophage X. require only small amounts of patient DNA. Second, recom-
Fig. 2 shows the strategy used for sequencing the seven binants should include the entire C i inhibitor gene locus, so
protein-encoding exons of the C i inhibitor Da and Rigenes. that linkage analysis with the HgiAI DNA polymorphism (the
C i inhibitor gene-specific primers were synthesized on the only known C i inhibitor RFLP) can be used effectively to
basis of intron sequence information from regions near intron- isolate the whole mutant gene. Finally, it is also desirable that
exon borders (see Miniprint) and used to generate sequences the break points of Ci inhibitor gene-containing inserts from
that could not be obtainedina simple manner from the different recombinants should be homogeneous in order to
universal priming sites of plasmid subclones. facilitate analysis of cloned Ci inhibitor genes from many
Identification of Da and Ri Mutations-The sequences of different kindreds. These criteria suggested that anapproach
the protein-encoding portionsof the dysfunctional Da andR 1 based on limitdigestion of genomic DNAmight be useful. We
genes were compared with the C i inhibitor cDNA sequence screened 11 restriction enzymes which produce four-base co-
from a normalindividual (8). Codon 444, encoding the reactive hesive ends that arecompatible with ends generated by diges-
site P1 residue of C i inhibitor, was different for the normal, tion of the polylinker from commercially available X phage
Da, and RI genes. In normal C i inhibitor, an arginine-encod- vectors for the ability to produce a single, packageable-sized
ing CGC triplet specifies residue 444. A single base change is hybridizing band on Southern blots probed with 32P-labeled
present in this codon in both of the dysfunctional genes we C i inhibitor cDNA. Bel1 generated a single band of about 21
studied TGC codes for a cysteine in the Da gene, whereas kb on Southern blots, and further mapping studies indicated
CAC codes for a histidine in the R1 gene (Fig. 3). that BclI sites are presentjust 5' and 3' to theportion of the
An additional difference between the sequences of both Ci inhibitor gene which contains theprotein-encoding exons
dysfunctional C i inhibitor genes and the normal cDNA was (see Fig. 4). Therefore, HANE patientlibraries were generated
also noted, but turns out to be an error inthe cDNA sequence by completely digesting patient DNAs with Bel1 and ligating
reported earlier by this laboratory (8). TheDa and Ri genes 18-23-kb fragments to BamHI arms of XDash, a bacteriophage
both have a G in the first position of codon 165 (glutamic X cloning vector.
 CpG Mutations
the
in Reactive Site of Human CiInhibitor
Universal andC i inhibitor-specific oligonucleotide primers
were used to sequence theprotein-encodingexons of the
dysfunctional inhibitorgenes. This representsa rapid method
for sequencing both strandsof relevant partsof the gene, and
oligonucleotides developed for this project can be usedreadily
t o identify the lesions present in other mutant C i inhibitor
genes as well.
Each of the dysfunctional inhibitorswe studied contains a
single base mutation which causes an aminoacid substitution
at the functionally important P1 residue. The P1 residue of
normal C i inhibitor, arginine 444, has become a cysteine in
C i inhibitor Da and a histidinein C i inhibitor Ri. As a
consequence of these substitutions, the physiologically rele-
vant target protease(s) of C i inhibitor may not be able to
recognize or form stable, inhibitory complexes with the dys-
functional proteins, leading ultimately t o a condition where
regulation of proteases which release vasoactive polypeptide
kinins becomes faulty, and an attack of edema occurs.
The specific P1 substitution which has occurred in C i
inhibitor Ri, arginine 444 to histidine, hasalso been observed
in a different, independently ascertained HANE family, C i
inhibitor At (48). However, C i inhibitor preparationsisolated
from the plasmaof patients from these two families displayed
variableamounts of residual activityagainsteach of five
different serine proteases inactivated by C i inhibitor, with
C i inhibitor Ri having negligible inhibitory activity against
any of them and Ciinhibitor At displayingsignificant residual
activity (40-50% range) against three of the five proteases
tested (38). Since the work reported here and the work of
Aulak et al. (48) indicate that C i inhibitors Ri and At have
the same P1 substitution, it is notable that their functional
properties were different(38). The apparently conflicting
observations may beexplained by phenomenarelatingto
contamination of some dysfunctional C i inhibitor prepara-
tions with 1) variable amounts of the patient’s normal C i
inhibitor gene product and 2) trace amounts of plasma pro-
teases, whichcould alterthe levels of residualinhibition
measured with the proteases testedfrom the parallel degrees
of inhibition expected for copurified, uncleaved normal C i
inhibitor. Alternatively, it is possible that C i inhibitor At
contains another mutation in addition to the P1 histidine
substitution since analysisof this mutant focused on a Pseu-
domonas elastase-generated C-terminalpeptide.
Like C i inhibitor Ri, antithrombin I11 Glasgow also con-
tains a histidine substitution at its P1 arginine residue (49).
Antithrombin I11 Glasgow has reduced abilitytoinhibit
thrombin, the protease target of antithrombin 111, but in-
creased affinity for heparin (50).
The mutationidentified by molecular cloningand sequenc-
ing of the C i inhibitor Da gene was the substitution of P1
arginine 444 by,cysteine. This result is consistent with pre-
vious work on the Da protein, in which it was shown to
possess a free sulfhydryl group (40) and was isolated largely
disulfide-bonded to albumin (32, 40). C i inhibitor antigen is
present in the seraof Da family patients at several times the
normal concentration (32). Itmay accumulate to abnormally
high concentrations in these individuals because the rate of
clearance for C i inhibitor Dacomplexed with albumin isvery
slow. Delayed clearance of another dysfunctionalC i inhibitor
protein ( C i inhibitor Ta) has been reported by Quastel et al.
(36). A phenotype similar to that of C i inhibitor Da has been
described for a Swedish hereditaryangioneuroticedema
kindred in which those affected have C i inhibitor antigen
complexed with albuminat three times the normal levels (51).
The dysfunctional C i inhibitor gene from thisSwedish family
could have the same mutation asC i inhibitor Da.
3069
The P1 arginine to cysteine mutation observed in C i inhib-
itor Da is also found in the Northwick Park variant of the
homologous serpin antithrombin I11 (49). As is the case for
C i inhibitor Da, antithrombin I11 Northwick Park is also
complexed with albumin (52); however, antithrombin I11 an-
tigen in Northwick Park patientsdoes not accumulate to the
very high levels observedfor C i inhibitorantigenin Da
patients (53).
A third P1 arginine to cysteine serpin mutant has been
studied. The gene for al-antitrypsin cysteine 358 was gener-
ated by i n vitro mutagenesistechniques,andthemutant
protein was expressed in yeast (54). Mutant protein which
had been preincubated with 1 mM dithiothreitol reacted rap-
idly with pancreaticelastaseandneutrophilelastase(the
physiological target of a,-antitrypsin), and the second-order
association rates were similar to thoseof the native inhibitor
containing a P1 methionine. Purified mutant protein which
was not pretreated with reducing agent migrated as a dimer
on sodium dodecyl sulfate-polyacrylamide gel electrophoresis
and was not an effective inhibitor. These findingssuggest
that angioedema in the Da family and thrombosis in the
Northwick Park family may result from complex formation
between the respective serpins and albumin, rather than from
an inherentlack of inhibitory activityin these serpin mutants.
Alternatively, the introductionof a fifth cysteinein C i inhib-
itor Da anda seventh cysteine in antithrombinI11 Northwick
Park may lead t o incorrect disulfide bondformationand
aberrant folding of these serpin variants.
A large number of P1 mutations are to beexpected among
serpin dysfunctional mutants.However, it is notable thatonly
two of the six possible amino acid substitutions which could
theoretically occur at the P1 arginines of C i inhibitor and
antithrombin I11 have actually been observed. The apparently
high incidence of P1 mutations andof recurring histidine and
cysteinesubstitutions in particular is consistentwiththe
presence of a CpG dinucleotide in the first two positions of
the codons specifyingarginine 444 (CGC) inC i inhibitor and
arginine 393 (CGT) in antithrombin 111. 5-Methylcytosine-
containing CpG dinucleotides are the major methylated se-
quences in vertebrates, anda large body of evidence supports
the hypothesis that they are hotspots for mutation viaa
mechanism involving spontaneous deamination of 5-methyl-
cytosine to thymine(55-57). For instance, CpG dinucleotides
are present muchless frequently than expected in sequenced
human genes (observed/expected = 0.37) (58). Also, restric-
tion endonuclease recognition sites containing CpG dinucle-
otides are known to exhibit increased rates of polymorphism
in human DNA (59). Finally, 35% of single base mutations
causing human genetic disease occur within CpG dinucleo-
tides (60).
Thus, we believe that themolecular mechanism underlying
the P1 arginine to cysteine mutations in C i inhibitor Da and
antithrombin I11 Northwick Park is deamination of the C in
the first position of CGY arginine codons to generate TGY
cysteine codons(whereY = pyrimidine). Similarly, deami-
nation of noncoding strand C residues in the second position
of the arginine-encodingCGY codons of C i inhibitors Ri and
At and antithrombin I11 Glasgow would result in the substi-
tution of A residues (and histidine-encodingCAY codons) on
the sense strand. The P1 residues of six human serpins are
arginines. In C i inhibitor, antithrombin 111, a2-antiplasmin,
and endothelialplasminogen activator inhibitor, theP1 argi-
nines are encoded by CGY codons (15, 18-20, 61). The P1
arginines of human protein C inhibitor (16) and placental
plasminogen activator inhibitor (17) are AGR (whereR =
purine). On the basis of the information discussed above, one
3070 CpG
the Mutations
in R e a hle Site of Human Cilnhibitor
would predict that a high incidence of cysteine and histidine
substitutions will be found at theP 1 arginines of C i inhibitor,
antithrombin 111, cwantiplasmin, and endothelial plasmino-
gen activator inhibitor and that the incidence of P1 substi-
tutions observed in naturally occurring protein C inhibitor
and placental plasminogen activator inhibitor mutants will
be much lower due to the absence of CpG in the AGR codons
specifying their reactive site arginines.
Acknowledgments-We thank members of the Da and Ri families
for their participation in this study and Dr. P. Model for helping us
to synthesize oligonucleotides.
REFERENCES
1. Ratnoff, 0. D., Pensky, J., Ogston, D., and Naff, G. B. (1969) J.
Exp. Med. 1 2 9 , 315-331
2. Gigli, I., Mason, J., Colman, R. W., and Austen, K. F. (1970) J.
Immunol. 104, 574-581
3. Forbes, C., Pensky, J., and Ratnoff, 0. (1970).J. Lab. Clin. Med.
76,809-815
4. Harpel, P. C., Lewin, M. F., and Kaplan, A. P. (1985) J . Biol.
Chem. 260,4257-4263
5. Pixley, R. A., Schapira, M., and Colman, R. W. (1985) J . Biol.
Chem. 260,1723-1729
6. Sim, R. B., Reboul, A,, Arlaud, G. J., Villiers, C. L., and Colomb,
M. G. (1979) FEBS Lett. 97, 111-115
7. Salvesen, G. S., Catanese, J. J., Kress, L. F., and Travis, J. (1985)
J. Bid. Chem. 2 6 0 , 2432-2436
8. Bock, S. C., Skriver, K., Nielsen, E., Thogersen, H. C., Wiman,
B., Donaldson, V. H., Eddy, R. L., Marrinan, J., Radziejewska,
E., Huber, R., Shows, T. B., and Magnusson, S. (1986) Bio-
chemistry 25,4292-4301
9. Davis, A. E., 111, Whitehead, A. S., Harrison, R. A., Dauphinais,
A., Bruns, G. A. P., Cicardi, M., and Rosen,F. W. (1986) Proc.
Natl. Acad. Sci. U. S. A. 83, 3161-3165
10. Tosi, M., Duponchel, C., Bourgarel, P., Colomb, M., and Meo, T.
(1986) Gene (Amst.) 42, 265-272
11. Que, B. G., andPetra,P. H. (1986) Biochem. Biophys. Res.
Commun. 137,620-625
12. Petersen, T. E., Dudek-Wojciechowska, G., Sottrup-Jensen, L.,
and Magnusson, S. (1979). in The Physiological Inhibitors of
Coagulation and Fibrinolysis (Collen, D., Wiman, B., and Ver-
straete, M., eds) pp. 43-54, Elsevier/North-HollandBiomedical
Press, Amsterdam
13. Carrell, R. (1984) Nature 3 1 2 , 14
14. Ragg, H. (1986) Nucleic Acids Res. 1 4 , 1073-1088
15. Holmes, W. E., Nelles, L., Lijnen, H. R., and Collen, D. (1987) J.
Biol. Chem. 2 6 2 , 1659-1664
16. Suzuki, K., Deyashiki, Y., Nishioka, J., Kurachi, K., Akira, M.,
Yamamoto, S., and Hashimoto S. (1987) J . Biol. Chem. 262,
611-616
17. Ye,R. D., Wun, T.-C., and Sadler, J. E (.1987) J . Biol. Chem.
262,3718-3725
15. Pannekoek, H., Veerman, H., Lambers, H., Diergaarde, P., Ver-
weij, C. L., van Zonneveld, A.-J., and van Mourik, J. A. (1986)
EMBO J . 5,2539-2544
19. Ny, T., Sawdey, M., Lawrence, D., Millan, J. L., and Loskutoff,
D. J. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,6776-6780
20. Ginsburg, D., Zeheb, R., Yang, A. Y., Rafferty, U. M., Andreasen,
P. A,, Nielsen, L., Dano, K., Lebo, R. V., and Gelehrter, T. D.
(1986) J . Clin. Inuest. 78, 1673-1680
21. Loebermann, H., Tokuoka, R., Deisenhofer, J., and Huber, R.
(1984) J. Mol. Biol. 177, 531-556
22. Carrell. R. W.. and Owen. M. C. (1985) Nature 317, 730-732
23. Reboul; A,, Arlaud, G. J.,'Sim, R. B., and Colomb, M. G. (1977)
FEBS Lett. 79,45-50
24. Hamel. P. C., and Cooper,- . N. R. (1975) J . Clin. Inuest. 5 5 , 593-
664 '
25. Sim, R. B., Arlaud,G. J., and Colomb, M. G. (1980) Biochim.
Biophys. Acta 6 1 2 , 433-449
26. Laskowski, M., Jr., and Kato, I. (1980) Annu. Reu. Biochem. 4 9 ,
593-626
27. Arlaud, G . J., Reboul, A., Sim, R. B., and Colomb, M. G. (1979)
Biochim. Biophys. Acta 5 7 6 , 151-162
28. Donaldson, V. H.,and Evans, R. R. (1963) Am. J. Med. 3 5 , 37-
44
29. Osler, W. (1888)Am. J. Med. Sci. 9 5 , 362-367
30. Austen, K. F., and Sheffer, A. L. (1965) N. Engl. J. Med. 272,
649-656
31. Donaldson, V. H., and Rosen, F. S. (1964) J. Clin. Inuest. 43,
2204-2213
32. Rosen, F. S., Alper, C. A., Pensky, J., Klemperer, M. R., and
Donaldson, V. H. (1971) J. Clin. Inuest. 50, 2143-2149
33. Cicardi, M., Igarashi, T., Rosen, F. S., and Davis, A. E., I11 (1987)
J . Clin. Inuest. 79, 698-702
34. Stoppa-Lyonnet, D., Tosi, M., Laurent, J., Sobel, A., Lagrue, G.,
and Meo, T. (1987) N. Engl. J. Med. 317, 1-6
35. Curd, J . G., Yelvington, M., Ziccardi, R. J., Mathison, D. A., and
Griffin, J . H. (1981) Clin. Erp. Irnmunol. 45, 261-270
36. Quastel, M., Harrison, R., Cicardi, M., Alper, C. A., and Rosen,
F. S . (1983) J. Clin. Inuest. 71, 1041-1046
37. Lachmann, P. J., and Rosen, F. S. (1984) Acta Pathol. Microbiol.
Immunol. Scand. Sect. C Suppl. 284,35-39
38. Donaldson, V. H., Harrison, R.A., Rosen, F. S., Bing,D.H.,
Kindness, G., Canar, J., Wagner, C. J., and Awad, S. (1985) J.
Clin. Inuest. 7 5 , 124-132
39. Donaldson, V. H., Wagner, C. J., Tsuei, B., Kindness, G., Bing,
D. H., Harrison, R. A., and Rosen, F. S. (1987) Blood 69,1096-
1101
40. Harrison, R. A,, and Rosen, F. S. (1982) Mol. Immunol. 19,1374
(abstr.)
41. Levy, L. R., and Lepow, I. H. (1959) Proc. SOC.Exp. Biol. Med.
101,608-611
42. Laurell, C. B. (1966) Anal. Biochm. 15,45-52
43. Mancini, G., Carbonara, A. O., andHeremans, J. F. (1965)
Immunochmistry 2 , 235-254
44. Bell, G. I., Karam, J. H., and Rutter, W. J. (1981) Proc. Natl.
Acad. Sci. U, S. A. 78,5759-5763
45. Carter, P. E., Dunbar, B., and Fothergill, J. E. (1988) Eur. J .
Biochem. 1 7 3 , 163-169
46. Viera, J., and Messing, J. (1982) Gene (Amst.) 1 9 , 259-286
47. Sanger, F., Nicklen, S., and Coulsen, A. R. (1977) Proc. Natl.
Acad. Sci. U. S. A. 74,5463-5467
48. Aulak, K. S., Pemberton, P. A., Rosen, F. S., Carrell, R. W.,
Lachmann, P. J., and Harrison, R. A. (1988) Biochem. J. 2 5 3 ,
615-618
49. Erdjument, H., Lane, D. A., Panico, M., DiMarzo, V., and Morris,
H. R. (1988) J. Biol. Chem. 263,5589-5593
50. Lane, D. A., Lowe, G. D. O., Flynn, A., Thompson, E., Ireland,
H., and Erdjument, H. (1987) Br. J. Haematol. 66, 523-527
51. Laurell, A.-B., and Martesson, U. (1971) Eur. J. Immunol. 1,
146-149
52. Erdjument, H., Lane, D. A., Ireland, H., Panico, M., Di Marzo,
V.. Blench. I.. and Morris, H. R. (1987) J. Biol. Chern. 262,
13381-13384 '
53. Howarth. D. J.. Samson. D.. Stirling, Y., and Seghatchian, M. J.
(1985) khrohb. Haemost6is5 3 , 314-319
54. Matheson, N. R., Gibson, H. L., Hallewell, R. A., Barr, P. J., and
Travis, J. (1986) J. Biol. Chem. 2 6 1 , 10404-10409
55. Salser, W. (1977) Cold Spring Harbor Symp. Quant. Biol. 4 2 ,
985-1002
56. Bird, A. P. (1980) Nucleic Acids Res. 8,1499-1504
57. Coulondre, C., Miller, J. H., Farabaugh, P. J., and Gilbert, W.
(1978) Nature 274, 775-780
58. Nussinov, R. (1981) J. Mol. Biol. 1 4 9 , 125-131
59. Barker, D., Schafer, M., and White, R. (1984) Cell 3 6 , 131-138
60. Cooper, D. N., and Youssoufian, H. (1988) Hum. Genet. 7 8 , 1 5 1 -
1 L.5
61. Bock, S. c . , W o n , K. L., Vehar, G. A., and Lawn, R. M. (1982)
Nucleic Acids Res. 1 0 , 8113-8125
62. Lawn. R. M.. Fritsch. E.F., Parker, R. C., Blake, G., andManiatis,
T. (1978) Cell 15,'1157-1174
63. Beaucage, S. L., and Caruthers, M. H. (1981) Tetrahedron Lett.
22,1859-1862
CpG Mutations
the
in Reactive Site of Human
Cilnhibitor 3071
FIGURE 5

111-R-10 20 30 40 50 60
TAATGGTCAG AGATTACAGA GTCCCTGACTATCCCTCATC TTCTGCAGAG ACATTCCTGT
80 70 90 100 110 120
GCACCCCCAC CCTCACCCTGTATTGCCCCT TCTCTGAGGA ATIAGTGGTG GTGGTTCTRA
130 150140 160 170 180
GACAGATTGC TCATCTGCTG CRCTGTCAGA AATTACTCTC TTGTACAGGA CATTTTCCAC
190 200 210 220 230 24C
IIICCACACCTTCTCTTCCIG CTTTGAGIAT TTTIGMGACTGTGCCICGI AGTMG-
""-C2b""".)
250 iII-R-260 268 lIII-L-10
R A A T G M C T CAGTTTCTTG AACCACAG, exon I 1 1 1499 b p ) lGTA4GACCCI
IIII-L-20
GCTTGMTTC . ................ .-1.2 k b ....................
,111-R-10 20 30 40 50 60
I'TIGGGAGGC AAGGCAGGAG TTCAAGACCA GCCTGGCCAA CATGTTGAAA CCCCATCTCT
'0 80 90 100 110 120
A C I W T G C W T T A G CCAGGCATAG TGATGCATGC TTATTGICCC AGCTACTTGG
130 140 150 160 170 Inn
GAGGCAGAGG TGGGAGGRTT GCTTGAACCT GGAGATTCAA GTGRGCTGAG AITGCACCAC
190 200 210 220 230 240
TGCATTCCAG CCTGGGCAAC AAAGCMGAC TCTGTCTCM RARAAARlIAA

250 260 270 280 290 300

AGAGAGAGAT TGAGAGAACA CTTCCAGCTC AGATGATCTG TGATCCCCTC CAAAGCAGGG

310 320 330 340 350 360

MTACCCTCCATTCCAGCCT GGTCCCCAAC CCTCATTCCC AAGGMGGCC CCCGRCTCAT
"."C3a""
170 380 iIII-R-390 395 4
CCTGCAAGTA TCTTTCATCT
CTGCCCTTTG TTGCAG exon IY (135 bpl 'GTGh
..~
10 IIv-L-20 30 40 50 60
GTGCCC AGGAATGGGC AGTGICTGCA GAGGAGGGTC CTGAGAGGAC TCTGAAGGGG
90 7 0 80 100 110 120
GACCAGCGCT GGGGRAAGAA AGGACAGAGG GAATGTTGGA GCTACAGIAT CAGGGATGGA
130 140 1bO I50
170 18c
CTGCAGACAG GGTGAAGACC TTGGCAGGAG CAITAGGTCA CTCCAGGAAC IAGACI'GTTC

190 200 210 210 230 I I ' I - L - 2 4 c

TTCTAATGAG ACCTTAGACA AGTCTCTGGC ITTCATCAAC TGCTTTAGAA W T M C T
........................ -4.0 kb ........................
lIV-R-10 20 30 40 50 60
ACATCTGTM GAGGAGTGGG CTGGACCGCG CTCCACCATG GCGTACACTA AGTGAGCAGA
'
0 80 90 100 110 120
TAGMCCATr\ GRAAGCRiGCTCACTCTCAA ATCGTGCTCA TGGAAAGMC GACGTGITCA
"".C4a"""")
130 140
LIV-R-150 151 I""
GGACTCATGC CTCCCTTTCT C M C A T A C C C CClG exon Y (201 bpl IGTAAGGG
10 20 30 40 50 60
TGC CCTTAGCCAG TTMTCTTCC CATTCTGGGT CCTTCTTCrC CTCCTGGCTTC
'1 81 91 1111 111 l/l
RAAGCCACTT M C C C C M G T TCTACAATCG GATCTCAATG TCCCTGCACTACTCITTCC?
131 111 151 1El 1
' 181
AACAAGGCTT TTAGCTCCTCTTCATCCrTTTCCTACCTGCATlAGAGCAACCClCCCACC
1"-191 IVL-L-lo 20 28
TCTTCCCTCT AGI exon V I ( 1 4 0 b p ) IGTMGTTCTT AACCTTTCCT
TCTCCTGT
30 40 5C 60
TI. GRAACCIACT TGAGTCTCCT GACTTTTTTT
<""".C5a.." .
70 80 90 100 110 120
CTGCTGGTAG TCCCATCATT TTGGGGTACA T G C T T A i M TTCATCACTT CTACTCCTTC
130 140 150 160 170 180
CATCTGTATT XCACCCTATCTTTTCTCCTTCTCCTTTCP CIAGCCTTGGC CAAG'ICRGRC
191 201 211 221 231 241
ATTGTATATT TTAGGCTGGA AACRGGAT'IC TCMGTTTCT TCATGCATCTCTXCTATAGT
251 261 271 281 291 301
GAATGGCAAA ATGTGAGTCG TGTTCCTATT CTACACATCT C T T T C T T T I T CCAGTCRACT
311 321 331 341 351 161
CATCAGMGA CTCTGGTGGC TTAGGCAAGT CCTGTGTGTR TGTGGGTACC TGTGTRCAGC
371 381 401 391
411 421
ATACATATAC ATGIGTAGAC GTGIATAAGT ATAICCCTCCTATACTGCAT ATATGTGGTA
131 441 451 461 471 481
AGTGCATGK TGCACACTGA GTTRAATGGC TTAAGTGCAT TAGCAGAAAT TGAACAACIT
491 511 501 5 2 1 ivl-L-531 5 3 4
TAGTITGATG AGTAGGGAGG GAGAGMGAC AGAATaTGGA GCCCAGIGGA TCC. ....
.................... - 5 5 kb . . . . . . . . . . . . . . . . . . . . . . . .
ivI-R-10 20 30 40 50 60
AAGCTIAGGTCTGACIGATGCTIGITGRAA TRCAGACTGT GGGAGCAGAC TGCAGGRCAG
70 80 100 90 110 120
CATTGTGACR GAGGGTGGGG C C X G A G A G A GATGCGGTAG GRAGLCTGTI MGGTGCATC
IVI-R-130 133 I"II-L-10 20
TCTTATTITC TAG; exon V I 1 1220 b p ) (GIGAGCXCTG GCAGCTTXGG GITACT
30 40 50 60
CCCA GGCCATCLGA GGAGRAAGGG GGGATCCCTA

70 80 90 100 110 120

AGATGTAGTT AGCATTCTCT AGAGTATITT TTRCATCCAT AATCICICTT TGTCCTGCM
130 110 150 I60 170 180
CCCTGCAkGT TAGAGGGGTA GGTGCTATTA TCCCATTGGA TCATTGTGGA AATGGAGGCT
IVII-L-I89
CAGAAGCII . . . . . . . . . . . . . . . . . . . . 1.9 kb . . . . . . . . . . . . . . . . . . . . . . .
lVII-R-IO 20 3" 40 50 tic
GAATTCAGGI CAAAGGTCIC CATCAGCTGA GGGTLICATG CTGGCTTCTG K T ' T G T I T T
I"II-R-70 '8
TCTCTGGTTT TGCCCTAGI exon V I 1 1 i518bp)