Blood Journal
Leading the way in experimental and clinical research in hematology

Impaired Cotranslational Processing as a Mechanism for Type I Antithrombin Deficiency

  1. Alison C. Fitches,
  2. Ruth Appleby,
  3. David A. Lane,
  4. Valerio De Stefano,
  5. Giuseppe Leone, and
  6. Robin J. Olds
  1. 1 From the Department of Pathology, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand; the Department of Haematology, Imperial College School of Medicine, Charing Cross Campus, London, UK; and the Department of Haematology, Catholic University, Rome, Italy.


Most secretory proteins, including antithrombin (AT), are synthesized with a signal peptide, which is cleaved before the mature protein is exported from the cell. The signal peptide is important in the process whereby nascent protein is recognized as requiring subsequent modification within the endoplasmic reticulum (ER). We have identified a novel mutation, 2436T→C L(-10)P, which affects the central hydrophobic domain of the AT signal peptide, in a proband presenting with venous thrombotic disease and type I AT deficiency. We investigated the basis of the phenotype by examining expression in mammalian cells of a range of variant AT cDNAs with mutations affecting the –10 residue. Glycosylated AT was secreted from COS-7 cells transfected with wild-type AT, –10L deletion, -10V or -10M variants, but not variants with P, T, R, or G at -10. Cell-free expression of wild-type and variant AT cDNAs was then performed in the presence of canine pancreatic microsomes, as a substitute for ER. Variant AT proteins with P, T, R, or G at residue –10 did not undergo posttranslational glycosylation, and their susceptibility to trypsin digestion suggested they had not been translocated into microsomes. Our results suggest that the ability of AT signal peptide to direct the protein to ER for cotranslational processing events appears to be critically dependent on maintaining the hydrophobic nature of the region including residue –10. The investigations have defined impaired cotranslational processing as a hitherto unrecognized cause of hereditary AT deficiency.

ANTITHROMBIN (AT) HAS a major role in maintaining blood fluidity by acting as the principal inhibitor of thrombin and other activated serine proteinases of coagulation. A conformational change occurs in AT when it interacts with the target proteinase, trapping it in a 1:1 complex.1 The inhibitory action of AT against thrombin is accelerated in vitro by the presence of heparin and evidence suggests that in vivo this function is provided by the proteoglycan heparan sulphate located on the endothelial cell surface.2 AT deficiency is associated with a predisposition to venous thrombosis.3 The prevalence of AT deficiency in patients presenting with thrombosis is 2% to 6%,4 while that of asymptomatic deficiency in the general population appears to be 1:600.5 Other interacting genetic or acquired risk factors play a role in the expression of the thrombotic phenotype in AT-deficient individuals.6

AT deficiency can be subdivided into two broad types based on immunological and functional activity levels of the protein in the plasma. Type I AT deficiency is characterized by a reduction to 50% of the normal level of immunologically and functionally detectable AT protein. Type II deficiency is associated with the presence of a variant protein leading to reduced functional activity, but normal immunological levels. AT deficiency is a heterogeneous disorder, as illustrated by the 106 distinct type I mutations identified to date,7 and this figure may be conservative, as haplotype analysis indicates that some apparently identical mutations may have arisen independently.8

Plasma AT is a glycoprotein with a molecular weight of 58,200, synthesized by hepatocytes as a 464 amino acid precursor from which a 32 amino acid signal peptide is cleaved. The signal peptide plays an important role in translation by ribosomes; nascent protein is bound via the signal peptide to the signal recognition particle (SRP), which guides the complex to the endoplasmic reticulum (ER). Translocation of the protein into the ER is followed by posttranslational processing, which includes disulphide bond formation, glycosylation, signal peptide cleavage, and folding.

The sequences of signal peptides are extremely heterogeneous, but three conserved features have been recognized and shown to be essential for protein export. The N terminal region of 5-8 amino acids is hydrophilic due to the presence of positively charged basic residues, and in prokaryotes, the charge affects the rate of protein translocation.9 A hydrophobic core of 7-15 amino acids is vital for cotranslational processing of the protein.10 The polar C terminal region of approximately six amino acids contains the signal peptide cleavage site. In this region positions -1 and -3 are usually occupied by small neutral residues, which are thought to fit the active site on the cleavage enzyme.11

We report the identification of a new AT mutation in an individual with type I AT deficiency and a history of venous thrombosis. The mutation results in an amino acid substitution in the hydrophobic domain of the AT signal peptide and blocks processing of the precursor AT protein.


Sample preparation.

Samples were collected from a proband with a family history of AT deficiency. DNA was extracted from peripheral blood leukocytes by standard proteinase K/phenol chloroform methods. Other family members were unavailable for testing.

DNA amplification and sequencing.

The seven exons comprising the AT gene, including flanking intron sequence, were amplified using the polymerase chain reaction (PCR). Amplification was performed in a 100 μL volume containing ≈200 ng DNA, 20 pmol of each oligonucleotide primer (as previously described),12 100 μmol/L of each deoxynucleotide triphosphate (dNTP), 50 mmol/L KCl, 10 mmol/L Tris HCl pH 8.3, 1.5 mmol/L MgCl2, and 2 U Taq polymerase (Boehringer Mannheim, Auckland, New Zealand). One of each primer pair was 5′-biotinylated to allow template preparation for single strand sequencing. The thermal cycling conditions were 30 cycles of denaturation at 94°C for 1 minute, annealing at 55° to 65°C for 1 minute, and extension at 72°C for 2 minutes with a final extension step of 5 minutes.

Amplified fragments were purified using streptavidin-coated magnetic beads (Dynabeads M280 Streptavidin, Dynal, Sydney, Australia) and a magnetic particle concentrator, and after incubation of the fragments with 0.15 mol/L NaOH to denature the double-stranded DNA, the biotinylated strand was isolated. Sequencing was performed by the dideoxy method using 35S-deoxyadenosine triphosphate (dATP) (Amersham, Auckland, New Zealand), Sequenase Version 2.0 (Amersham), and 5 pmol of nonbiotinylated primer. The products were electrophoresed on a 6% polyacrylamide:bis (19:1) gel containing 7 mol/L urea at 50 W for 2 to 5 hours. The gels were fixed in 10% methanol, 10% glacial acetic acid for 30 minutes and dried. Autoradiography was performed at room temperature for 48 hours.

Mutagenesis of AT cDNA.

Wild-type AT cDNA was cloned into the vector pCR II (Invitrogen, Bresatec, Adelaide, Australia). The construct contained the complete AT coding sequence, but none of the 5′ or 3′ untranslated region. Mutations were generated in the cDNA by inverse PCR using the ExSite site-directed mutagenesis kit (Stratagene, LabSupply Pierce, Christchurch, New Zealand). Primer sequences are contained in Table 1. After ligation with T4 DNA ligase, the variant cDNA-vector constructs were transformed in Epicurian Coli XL1-Blue supercompetent cells (Stratagene). DNA from the resulting colonies was sequenced to check that the mutation was present and no further changes had been introduced.

View this table:
Table 1.

Primers for Site-Directed Mutagenesis

Mammalian cell expression of AT constructs.

A 1.4-kb EcoRI fragment containing the AT cDNA was isolated from pCR II and cloned into the EcoRI site of the mammalian expression vector pcDNA3 (Invitrogen). After transformation intoDH5α competent cells, clones with the correct insert orientation were selected by screening with restriction enzyme digests.

COS 7 cells at approximately 50% confluence were transfected with 1 μg of the pcDNA3 constructs using Lipofectamine (Gibco BRL, Life Technologies, Auckland, New Zealand) for 6 hours as described by the manufacturer. Cells were then grown at 37°C in 5% CO2/air in Dulbecco’s modified Eagle’s medium (DMEM) with 2 mmol/L glutamine and 10% fetal bovine serum (FBS).

Approximately 48 hours after transfection, culture supernatant was discarded and the cells were washed in DMEM lacking FBS and methionine. After resuspension of the cell pellet in the same media, newly synthesized cellular proteins were radiolabelled by the addition of 50 μCi 35S-methionine (Amersham) to a 1.5 mL volume for 3 hours. Some cells were cultured in the presence of tunicamycin (Sigma, Auckland) for 1 hour before and during labelling of cellular proteins to inhibit glycosylation. Cells were then lysed in 10 mmol/L Tris pH 8.0, 140 mmol/L NaCl, 1 mmol/L EDTA, 5 mmol/L dithiothreitol (DTT), 1% NP40, 0.1% sodium dodecyl sulfate (SDS), with proteinase inhibitors (1 mmol/L phenylmethyl sulfonyl fluoride [PMSF], 2 μg/mL aprotinin, 0.5 μg/mL leupeptin, 0.7 μg/mL pepstatin). Cell lysates and culture supernatants were immunoprecipitated with rabbit polyclonal antisera raised against human AT (DAKO, MedBio, Christchurch), using formalin-fixedStaphylococcus aureus. The products were electrophoresed in 10% acrylamide-bis (37.5:1) protein gels for 1 hour at 100 V, followed by fixation and autoradiography.

Cell-free expression of AT constructs.

RNA from wild-type and mutant AT cDNA was generated using the SP6 promoter of the pCR II vector. Plasmid DNA was linearized withNot1 and incubated at 37°C for 30 minutes in a 50 μL reaction containing 10 μL 5 × SP6 buffer, 5 μL 0.1 mol/L DTT, 1 μL RNAguard (Amersham Pharmacia Biotech, Auckland, New Zealand), 10 μL guanosine triphosphate (GTP) mix (10 mmol/L ATP, cytidine triphosphate [CTP], uridine triphosphate [UTP] and 0.5 mmol/L GTP), 4 μL linear DNA (0.5 μg/μL), 2.5 μL Cap (New England Biolabs, Beverly, MA) and 1.5 μL SP6 polymerase (Amersham Pharmacia Biotech). After this, 5 μL of 10 mmol/L GTP was added and the reaction allowed to proceed for a further 30 minutes. After phenol chloroform extraction, RNA was precipitated with 3 mol/L sodium acetate and 100% ethanol.

Variant and wild-type AT proteins were produced using nuclease-treated rabbit reticulocyte lysate (Promega, Dade Diagnostics, Auckland, New Zealand) in a volume of 26 μL containing 17.5 μL lysate, 0.5 μL amino acids minus methionine, 5 μL 35S-methionine and 3 μL RNA in the presence (1.8 μL) or absence of canine pancreatic microsomal membranes (Promega) for 1 hour at 30°C. Some products were incubated with endoglycosidase H (New England Biolabs), 500 U at 37°C for 60 minutes, to remove any N-linked carbohydrate. Electrophoresis was as described above.


Clinical history.

The proband comes from a family in which three generations have been identified with type I AT deficiency. She suffered her first thrombotic episode at age 28 during the first trimester of pregnancy when she was diagnosed with a deep vein thrombosis of the left calf and a pulmonary embolism. She had previously undergone surgery twice without any problem. On investigation, her AT antigen level was 40% (normal range, 80% to 120%) and a diagnosis of type I AT deficiency was made. She received treatment as previously reported.13 After 5 years, she refused further treatment and 2 years later, during a period of bed rest, she suffered another thrombotic episode.

At the age of 52, her father was diagnosed with a deep leg vein thrombosis and a pulmonary embolism. Three years later he had a cerebral vein thrombosis, after which he received anticoagulant treatment. The proband’s son has also been diagnosed with type I AT deficiency.

AT gene sequence.

The seven exons and flanking intronic regions of the AT gene were sequenced in the proband. A mutation was identified in the region of exon 2 coding the signal peptide, a substitution of T→C at nucleotide 241814 in the second position of codon -10 converting the normal leucine (CTC) to proline (CCC) (Fig 1). The mutation was confirmed by repeat sequencing of independently amplified fragments. No other mutations were found in the exons or flanking intron sequences of the proband.

Fig. 1.

AT signal peptide mutation. Exon 2 PCR product was directly sequenced in the proband and a heterozygous single nucleotide substitution was identified at nucleotide 2418, leading to the amino acid substitution L(-10)P.

Expression of AT in mammalian cells.

Wild-type AT and the L(-10)P variant were expressed in cultured mammalian cells to confirm that the identified amino acid substitution was responsible for the phenotype of type I AT deficiency and to explore the mechanism by which the phenotype arose. In addition to the identified variant, another six substitutions at -10L were also created by site-directed mutagenesis. The additional mutations were a deletion of -10L and the replacement of –10L by V, M, T, R, or G. COS 7 cells were transiently transfected with wild-type and variant constructs. Cell lysates and culture supernatants were then examined by immunoprecipitation with polyclonal human AT antibody to determine the distribution of AT protein forms.

Culture supernatant from cells transfected with wild-type construct contained a single AT protein form of approximately 60 kD, while the cell lysate contained proteins of 56 kD and 47 kD, with little or no 52 kD form (Fig2). 52 kD is the predicted size of the unprocessed 464 amino acid AT protein, while the 47-kD band is a truncated AT protein resulting from internal initiation of translation (Fitches and Olds, unpublished and Sheffield and Blajchman15). The addition of tunicamycin, to inhibit N-linked glycosylation, resulted in AT protein forms of 50 to 52 kD in culture supernatant and both 50 to 52 kD and 47 kD in cell lysate (data not shown). 50 kD is the size expected of AT protein, which has undergone signal peptide cleavage, but has insignificant glycosylation. This suggests that the 60-kD exported protein represented processed and glycosylated AT, while the broad band of approximately 56 kD is incompletely and variably glycosylated intracellular protein. Cells transfected with wild-type AT construct, therefore, were able to posttranslationally modify the 52-kD, 464 amino acid precursor form of AT protein before export, mimicking the processing of AT expected in vivo by hepatocytes.

Fig. 2.

Expression of wild-type and variant AT cDNA in mammalian cells. COS 7 cells were transiently transfected with pcDNA3 vector constructs containing wild-type (WT), or variant AT cDNAs encoding P, M, V, T, G, or R at position –10, or with deletion (Del) of –10L. Cell lysate (l) and culture supernatants (s) were subsequently immunoprecipitated with polyclonal AT antibody. Mock-transfected COS 7 cells (COS) indicate nonspecific binding in the cell lysates. AT is exported to the supernatant only in cells transfected with the wild-type, -10M, -10V and –10L deletion constructs. The substitution of –10L by P, T, G, or R blocks processing and export of the variant AT proteins, with cell lysates showing the presence of only unprocessed and internally initiated protein. See Results for an interpretation of band sizes.

COS 7 cells also efficiently expressed the variant proteins (Fig 2). Deletion of -10L or substitution of –10L by M or V did not affect handling of the variant proteins by the COS 7 cells. In contrast, variants with P, T, R, or G at position -10 could not be found in culture supernatants, although cell lysates contained 52 kD and 47 kD AT forms. These observations suggested that the latter amino acid substitutions impaired processing and/or export of the AT protein.

Cell-free expression of AT protein.

A cell-free expression system using rabbit reticulocyte lysate was used to translate mRNA derived from the pCR II-AT constructs, to investigate the mechanism by which the L(-10)P substitution resulted in type I AT deficiency. Translation of mRNA from all of the constructs yielded two proteins of 52 kD and 47 kD (Fig 3), which immunoprecipitated with polyclonal human AT antibody (data not shown). Repeated experiments suggested that the efficiency of translation of the variants appeared similar to that of the wild-type protein. Further cell-free translation reactions were then performed in the presence of canine pancreatic microsomes to assess the efficiency of translocation and posttranslational processing of the normal and variant AT proteins. Wild-type protein clearly underwent modification, as indicated by higher molecular weight product of about 56 kD, in addition to the 52 kD and 47 kD bands (Fig 3). An increased molecular weight suggested the wild-type AT protein had undergone N-linked glycosylation within the lumen of the microsomes. This was confirmed by incubating protein generated in the presence of microsomes with endoglycosidase H, which resulted in the loss of the 56-kD product, but no change to the 52-kD and 47-kD bands (data not shown). A similar pattern of posttranslational modifications was discovered for variants with M and V at residue −10. A faint band of 56 kD was seen in some, but not all translations of the –10L deletion variant (compare Fig 3 and Fig 4), suggesting a low level of processing. Translation of the proteins with P, T, R, and G at residue −10 in the presence of microsomes did not result in observable modification of the proteins (Fig 3).

Fig. 3.

Cell-free translation of AT cDNA constructs performed in the presence (c) or absence of microsomes. Wild-type, -10M and -10V AT constructs undergo modification in the presence of microsomes to generate a 56-kD species, consistent with N-linked glycosylation, indicating that the protein has been translocated into microsomes.

In addition to becoming glycosylated, protein that is translocated into microsomes should be resistant to proteinase digestion because of the protected environment provided by the microsomes. Translated products were incubated with 0.1 mg/mL trypsin at 0°C for 30 minutes after exposure to microsomes. AT protein forms of 52 kD and 47 kD were digested, while the 56-kD bands observed in wild-type, −10M and –10V AT variants were relatively protected (Fig4). Permeabilization of the microsomes by incubation with 0.1% Triton for 1 hour at 30°C, after allowing processing to occur, resulted in a susceptibility to trypsin digestion of all protein forms (not shown). The apparent lack of posttranslational processing and the susceptibility to trypsin digestion of –10P and −10T (Fig 4), and –10R and –10G (not shown) variants is consistent with a failure of translocation of the translation products into the microsomes.

Fig. 4.

Trypsin protection assay. Wild-type, -10P, -10T, -10V, -10M, and –10L deletion variant AT proteins were translated by rabbit reticulocyte lysate in the presence of microsomes. The translation product was incubated with trypsin (t, see Materials and Methods) to digest product, which had not been translocated into the protected environment provided by the microsomal lumen. Relative protection from digestion of the 56-kD product is seen in translations of wild-type AT, -10V, and –10M variants. The 52-kD and 47-kD products from all translations were susceptible to trypsin digestion.


Proteins that are destined for membrane insertion or secretion, in both eukaryotes and prokaryotes, are characterized by the presence of a signal peptide. Signal peptides of different proteins display little sequence homology, but generally a central hydrophobic core of 7-15 residues is present16 (Fig 5). We have identified a mutation within the hydrophobic region of the AT signal peptide that leads to ≈ 50% reduction in levels of secreted AT. This is consistent with absence from the plasma of AT derived from the variant allele. Our data shows that although the protein is translated, it is unable to enter the ER to undergo posttranslational processing. Despite the diversity of mutations underlying type I AT deficiency, only two other mutations associated with the phenotype have been identified in the signal peptide, both of which produce stop codons, resulting in premature termination of translation.7 17

Fig. 5.

Cartoon of AT signal peptide illustrating the location of the central hydrophobic domain and the L(-10)P substitution.

Few other signal peptide mutations have been associated with human disorders, but three patterns of effect can be observed. Some mutations affect signal peptide cleavage, for example in the genes for preprovasopressin resulting in diabetes insipidus18 19 and coagulation factor X where amino acid substitution at codon -3 leads to a severe bleeding disorder.20 Interestingly, a substitution at codon -3 of AT apparently has no detrimental effect, although the site of signal peptide cleavage is altered.21 Several mutations resulting in a shift in the frame of translation have been described, affecting genes for biotinidase,22 apoC-II resulting in familial chylomicronemia syndrome,23 the insulin receptor gene causing insulin resistance,24 and AT.7 17 Three mutations that probably interfere with translocation have previously been identified within the signal peptide hydrophobic domains of amelogenin, resulting in x-linked hypoplastic amelogenesis imperfecta, a disorder characterized by defective dental enamel,25 preproparathyroid hormone causing familial isolated hypoparathyroidism,26 and bilirubin UDP-glucuronosyltransferase causing Crigler-Najjer syndrome.27

Our data provide some evidence of why substitutions affecting the –10 position of the signal peptide disrupt the usual fate of the AT protein. Total hydrophobicity of the signal peptide hydrophobic region appears to be a crucial factor, with the nature of the amino acids forming the hydrophobic region defining the length necessary for efficient cotranslational processing.28 It may therefore be possible to define upper and lower limits of signal peptide hydrophobicity, but these would vary widely for each protein.9 Assigning hydrophobicity to amino acid residues has been a task undertaken by numerous groups. A variety of tables have been produced based on experimental and statistical scales, combinations of the scales, and averages of scales. Two of the most widely used are those of Hopp and Woods,29 designed to identify antigenic determinants, and Kyte and Doolittle,30which is based on constituent parts of the amino acid side chains. Using either scale, the substitution of –10L by P, T, G, or R, which results in a failure of processing, decreases the overall hydrophobicity of the signal peptide hydrophobic region (Table 2). Processing was maintained with V or M at position –10 and also with deletion of the residue; these changes maintain the hydrophobic nature of the central region. An alternative explanation for the effects of the substitutions on processing relates to the possibility that the hydrophobic region of signal peptides may form an α-helix.31 P and G residues disrupt α-helices,32 so these amino acid substitutions may have a structural rather than a hydrophobicity-related effect.

View this table:
Table 2.

AT Protein Expression by Reticulocyte Lysate and COS 7 Cells

Posttranslational modification events, such as disulphide bond formation, protein folding, glycosylation, and cleavage of the signal peptide are integral to the production of functionally active mature proteins. A crucial aspect of generating functional protein, therefore, is the transfer of newly translated molecules from the ribosome into the ER. Transfer may either occur cotranslationally, the pathway that predominates in mammalian cells, or posttranslationally. In the former, the polypeptide crosses the ER membrane and enters the lumen while it is still being synthesized by membrane bound ribosomes. In general, the molecules involved in cotranslational processing of polypeptides are conserved from bacteria to mammals. In eukaryotes, this transfer is facilitated by a ribonucleoprotein complex, the signal recognition particle (SRP)10 and the SRP receptor located on the ER membrane. With several of the codon –10 variants we studied, including the L(-10)P variant identified in the patient, we showed a failure of protein to enter the ER. Whether this represents a failure of translocation of the variant across the ER membrane, a failure of the SRP-ribosome–nascent signal peptide complex to interact with the SRP receptor, or a failure of SRP to recognize the variant signal peptides is unknown. If the signal peptide emerging from the ribosome is not sufficiently hydrophobic, it may not be recognized by SRP54, and the variant nascent protein will not be maintained in a translocationally competent state. The identification here of the substitution within the AT signal peptide provides further support for the important role of the signal sequence in directing cotranslational processing events.


  • Address reprint requests to Professor Robin Olds, MBChB, PhD, Department of Pathology, Dunedin School of Medicine, University of Otago, PO Box 913, Dunedin, New Zealand; e-mailrobin.olds{at}

  • Supported by funding from the Health Research Council of New Zealand.

  • The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. section 1734 solely to indicate this fact.

  • Submitted May 4, 1998.
  • Accepted August 5, 1998.


View Abstract