Universal Distribution in Smooth Muscle
Abstract Immunohistochemical analysis of surgically obtained porcine tissue samples reveals ubiquitous staining for prothrombin in organs rich in smooth muscle content and universal staining of smooth muscle in tissue vasculature. The native character of tissue prothrombin is verified first by chromogenic substrate hydrolysis and hirudin inhibition after incubation of tissue extracts with taipan snake venom and phospholipid. Western analysis of tissue extracts confirms the native zymogen molecular weight. In addition, prothrombin purified in good yield from porcine uterus is activated by Echis carinatus venom which, like taipan venom, is 4-carboxyglutamic acid-sensitive. After correction for blood (gross heme) and interstitial fluid (albumin), excess functional prothrombin is observed in extracts of tissues having abundant smooth muscle. In contrast with factor X, the yield of prothrombin purified from porcine uterus greatly exceeds that attributable to contamination by whole blood. Northern blot analysis from selected bovine tissues extracted for polyadenylated messenger RNA is equivocal for prothrombin mRNA with the exception of liver, which is positive. It is concluded that functionally intact prothrombin is widely distributed among tissues owing to smooth muscle content, although the mechanism of emplacement and physiologic significance of prothrombin in these tissues remains unclear.
- Received May 6, 1997.
- Accepted August 5, 1997.
The liver is known to be the major site of biosynthesis of the vitamin-K-dependent proteins and accounts for their plasma concentrations. Although other tissues have been inferred to biosynthesize prothrombin, extrahepatic sources are unlikely to exert a significant influence on any but local interstitial prothrombin concentrations. Prothrombin mRNA has been observed in assorted fetal and pregnant rat organs1 and in the central nervous system,2,3 and prothrombin antigen has been recovered from estrogen-primed rat uterus.4 A physiologic significance for tissue prothrombin has not been established.
In the course of observing thrombin distribution in experimental arterial thrombosis, we observed thrombin antigen by immunohistologic analysis in uninjured arterial smooth muscle remote from the site of injury and thrombus. In light of accounts of extrahepatic prothrombin and the diverse activities of thrombin,5,6 we undertook an analysis to determine if prothrombin might be more widely distributed in tissues other than liver. We report that, owing to its occurrence in vascular and other smooth muscle cells, prothrombin has essentially universal organ distribution.
Materials and Methods
Chromogenic substrate for thrombin (Spectrozyme TH) was purchased from American Diagnostica. Recombinant desulfatohirudin (Revasc) was a gift from Dr Robert B. Wallis, Research Center, Ciba-Geigy Pharmaceuticals.7 Second antibody conjugates and Naja nigricollis n and Oxyuranus scutellatus scutellatus venoms were purchased from Sigma Chemical Co. Phospholipids were purchased from Avanti Polar Lipids. BALB/c mice were purchased from The Jackson Laboratory, Bar Harbor, ME. A mouse monoclonal antibody against human thrombin (EST-4) was obtained from American Diagnostica. Human tissue factor was a gift from M. Getz, Department of Biochemistry and Molecular Biology, Mayo Clinic. The cDNA for bovine prothrombin (PBII-III) was a gift from R. MacGillivray, Department of Biochemistry, University of British Columbia.8
Porcine prothrombin was purified as described previously.9 Published methods were used for preparation of prethrombin 1, and fragment 1.10 Thrombin was obtained by activation with taipan snake (O scutellatus s) venom.11
Monoclonal Antibodies to Porcine Prothrombin
Murine monoclonal antibodies, designated 50-H11 and 87-F8, were prepared by immunizing BALB/c mice with purified porcine prothrombin (100 μg/dose) in Freund’s adjuvant at 2-week intervals. Three days after the third injection, spleen cells were fused with NS-1 mouse myeloma cells. Culture supernatants from wells showing growth were assayed for the presence of prothrombin-specific antibodies. Positive hybridomas were identified with a solid-phase radioimmunoassay using 125I-porcine prothrombin, cloned by limiting dilution, expanded in culture, and grown as ascites tumors. Monoclonal Igs were purified on protein A-Sepharose. By enzyme-linked immunosorbent assay and Western analysis, both 50-H11 and 87-F8 bind to prothrombin and fragment 1·2, but not to thrombin.
Immunochemical Staining of Tissues
Fresh tissue samples were obtained surgically under general anesthesia from 4-month-old female pigs. These tissues were rinsed with saline, flash-frozen in liquid nitrogen, and stored at −70°C. Frozen sections were stained with prothrombin-specific antibodies to determine the distribution of prothrombin antigen in tissues. After fixation in cold acetone, the sections were incubated with 0.3% H2O2 and 0.1% NaN3 to block endogenous peroxidase. Nonspecific binding of secondary antibody conjugates was blocked by incubating sections with 5% nonimmune serum of the same species as the labeled secondary antibodies. The presence of prothrombin was detected by incubation with mouse monoclonal antibodies having specificities for thrombin or fragment 1·2 followed by anti-mouse IgG-biotin conjugates. The sections were then probed with streptavidin coupled to horseradish peroxidase. Slides were rinsed thoroughly with distilled H2O between incubations. Red staining was achieved by using aminoethylcarbazole as a peroxidase substrate. Sections were counterstained with hematoxylin. Negative controls were performed using nonimmune murine immunoglobulin to delineate background staining.
Polyclonal Antibody to Porcine Prothrombin Fragment 1·2
Prothrombin fragment 1·2 was purified from digests of porcine prothrombin (10 mg in 10 mL of 100 mmol/L NaCl, 10 mmol/L CaCl2, and 20 mmol/L Tris, pH 6) with black-necked spitting cobra (N nigricollis n) venom (10 μg/mL, room temperature, 144 hours). The products were separated by ion-exchange chromatography on a column (1×15 cm) of QAE-Sephadex developed with a linear gradient of 0.1 to 0.6 mol/L NaCl. Purified fragment 1·2 (30 μg) was emulsified with 800 μL of Freund’s complete adjuvant and injected subcutaneously (three sites) in chickens. The birds were boosted at 14 days with 15 μg of purified fragment 1·2 emulsified with 800 μL of incomplete Freund’s adjuvant. Antisera were collected 7 days after each boost. The immunoglobulin fractions were prepared from sera by precipitation three times with 40% saturated ammonium sulfate. The preparations were dissolved in the original serum volumes and then dialyzed in buffered saline. Preimmune immunoglobulins were prepared from all birds.
Tissue Extracts for Prothrombin Assay
Fresh surgically obtained tissue samples (80 to 250 mg) were rinsed with saline, flash-frozen in liquid nitrogen, and stored at −70°C. The frozen tissue samples were pulverized in an anvil at −196°C and then homogenized (close-fitting Dounce, 10 strokes, 22°C) in 1 mL of buffered saline solution (100 mmol/L NaCl and 20 mmol/L NaOH-HEPES, pH 7.5) containing 0.1% PEG 8000. The supernatants from a 10 minute centrifugation at 80 000 rpm (Beckman Airfuge) were archived at −70°C. For those samples to be used for Western transfer analysis, 1/10 vol of 10% NaDodSO4 solution was added before storage.
Prothrombin in tissue extracts was assayed as thrombin generated with taipan snake venom.11 Samples (100 μL), liposomes (5:1 dioleoylphosphatidylcholine:dioleoylphosphatidylserine [PCPS], 4 mg/mL, 50 μL) and thrombin chromogenic substrate (10 μL) were added to 800 μL of buffered saline solution, and then taipan snake venom (1 mg/mL, 5 μL) was added to initiate the reaction. When the rate of substrate hydrolysis became constant, the reaction was back-titrated with recombinant hirudin until chromogenic substrate hydrolysis was halted. Typically, accurate titrations entailed three or more iterations. The thrombin concentration was estimated both by comparison to authentic porcine thrombin and by hirudin stoichiometry, which were congruent.
Hemoglobin concentration was measured in each of the tissue extracts to estimate the maximum extent of whole blood contamination. Hemoglobin content for the samples (diluted with 1 vol of buffered saline solution) were estimated by dividing the absorbance of the solution at 419 nm by the extinction coefficient of hemoglobin (194 mmol/L−1 · cm−1). The hemoglobin concentration was transformed to blood content by assuming an average hemoglobin concentration of 10 g/dL. Because 10 g/dL lies at the lower end of normal pig hemoglobin concentration and the myoglobin and other heme content was ignored, these assays are susceptible to overestimation of blood contamination.
Albumin concentration was measured in tissue extracts to estimate the extent that extracellular fluid (blood plus interstitial) contributed to total prothrombin content. Samples were diluted with 1 vol of buffered saline solution containing 2% NaDodSO4 and then analyzed by polyacrylamide gel electrophoresis in NaDodSO4 (Pharmacia Phast system, 12.5% homogeneous gels). Albumin was estimated by densitometry (NIH Image) of gels stained with Coomassie blue and standardized with bovine albumin. Without correction for contaminating plasma, the contribution of interstitial fluid plus plasma to the prothrombin yield was calculated by using a prothrombin:albumin ratio of 1:400, which variably exceeds that published for lymph.12 Tissue albumin values, typically near or at background, in turn overestimate interstitial fluid contamination due to comigration of intracellular proteins with similar electrophoretic properties.
Western Blot Analysis
Additional analysis of antibody specificity was accomplished by Western blotting, where reactivity is directed to proteins of known molecular weight. Purified prothrombin, thrombin or an activation mixture of prothrombin [with taipan snake venom and 5 mmol/L Ca(II)] were run on a 10% to 15% gradient NaDodSO4 gel on a PhastGel system (Pharmacia Biotech). The proteins were allowed to diffuse overnight onto a membrane of nitrocellulose by constructing a tightly compressed sandwich of gel, membrane, and filter paper (wet with Tris-buffered saline, pH 7.4). The membranes were blocked with 5% nonfat dry milk (blotto) for 30 minutes and incubated with antibody (5 μg/mL in 1% blotto) for 120 minutes. After three washes of 5 minutes in 1% blotto, the transfers were probed with a second antibody conjugated to alkaline phosphatase (500 ng/mL, 1% blotto) for 90 minutes. Again the membranes were washed and then incubated with an NBT/BCIP substrate (Bio-Rad) until sufficiently developed. The membranes were washed a final time with Tris-buffered saline. All reactions were done at 22°C.
Prothrombin in porcine tissues was compared structurally to that purified from plasma by Western transfer analysis. Extracts of tissues, along with prothrombin, thrombin, prethrombin 1, and prothrombin fragment 1·2, were separated by electrophoresis in NaDodSO4/10% polyacrylamide gels. The high protein concentrations in extracts yielded some nonspecific staining, so for each analysis companion samples were supplemented with authentic prothrombin to provide an internal standard. Proteins were electrotransferred to nitrocellulose membranes at 30 V, overnight at 4°C, in 0.19 mol/L glycine, 25 mmol/L Tris, and 20% methanol. The nitrocellulose membranes were blocked in 5% (w/v in Tris-buffered saline) blotto for 30 minutes, incubated in chicken anti-porcine fragment 1·2 (10 μg/mL in 1% blotto) for 120 minutes at 22°C, and then washed (three times for 5 minutes) in 1% blotto. Membranes were then incubated with 125I-labeled rabbit anti-chicken IgG for 18 hours at 22°C. After washing (three times for 5 minutes) in 1% blotto, the membrane was exposed to X-ray film for 24 to 96 hours at −70°C.
Tissue Prothrombin/Factor X Isolation
Frozen pig uterus (110 g) was thawed, sliced, and suspended in 200 mL of buffered hypotonic solution (0.01% NaN3, 50 mmol/L Tris, and 10 mmol/L benzamidine, pH 7.5) and stirred for 24 hours at 4°C. The solution was then filtered and centrifuged at 14 000 rpm for 15 minutes. Prothrombin and factor X were purified by the method of Owen et al10 and concentrated by dextran sulfate chromatography. The prothrombin was activated by E carinatus venom, and the thrombin concentration was estimated by chromogenic substrate hydrolysis. Factor X was assayed by the two-stage Russell’s viper venom coagulation assay.13 Both prothrombin and factor X activity were quantified by comparison with porcine plasma. Each milliliter of plasma was assigned an arbitrary value of 100 U for both prothrombin and factor X activities.
Northern Blot Analysis
Fresh bovine tissue for Northern blot analyses was obtained at local slaughterhouses, washed with normal saline, and flash-frozen in liquid nitrogen. After pulverization under liquid nitrogen, polyadenylated messenger RNA was extracted from 1 to 10 g (dry weight) tissue with a commercially available kit (RNA STAT-60, Tel-TEST “B,” Inc) and stored at −70°C in diethylpyrocarbonate-treated water until analysis.
The 2043-bp sequence for bovine prothrombin was inserted into plasmid vector pBR322 (Promega Corp) and incorporated into a competent Escherichia coli cell line (CSH-50) for expression. Modified alkaline lysis DNA purification (Qiagen, Inc), in vitro transcription in a Sp6/T7 system (Promega), and subsequent nick translation14 gave a 32P labeled RNA probe with 50 000 cpm for hybridization with tissue mRNA. Bovine, rather than porcine, tissue was used for this analysis due to the availability of species-specific mRNA probe.
Poly(A) mRNA (2 to 4 μg) from tissue extracts were electrophoretically separated in 1% agarose at 80 V/25 mA for 1 hour, transferred to nitrocellulose overnight, and prehybridized using standard conditions. Standards included 16 S/23 S and 18 S/28 S mRNA (Sigma Chemical Co), and human tissue factor mRNA. The membrane was hybridized overnight with the 32P labeled RNA probe at 42°C in 50% deionized formamide, 250 μg/mL of herring sperm DNA, 10× Denhardt’s solution 50 μg/mL of poly(A), 0.1% NaDodSO4, and 5× SSC, and washed under stringent conditions (0.1× SSC/0.1% NaDodSO4, 60°C). Radioautography was carried out at −70°C for 80 to 150 hours. Blots were then stripped twice for 20 minutes each in 0.1× SSC/0.1% NaDodSO4, probed with a 32P-labeled GAPDH standard, and redeveloped.
As anticipated in liver, hepatocytes stain diffusely with monoclonal and polyclonal immunoglobulins against thrombin, prothrombin fragment 1·2, and prothrombin 4-carboxyglutamic acid (Gla) domain with a cytosolic distribution that shows no obvious subcellular localization (Fig 1⇓). Not anticipated is the staining seen in small blood vessels, where the vascular smooth muscle appears as positive as hepatocytes, yet endothelium appears negative. Other cell types within the liver, Kupffer, bile duct epithelium, and stromal cells, including those of adventitia, are negative. An exception is the tissue surrounding the portal triad, in which myocytes stain positive.
A survey of other tissue reinforces the finding that prothrombin occurs ubiquitously in smooth muscle (Fig 2⇓). In femoral artery (Fig 2A⇓), prothrombin stain in smooth muscle clearly demarcates the boundary between media and adventitia, while the endothelium likewise appears negative. Juvenile uterus (Fig 2B⇓), a tissue rich in smooth muscle, stains positive for antigen. Small intestine (Fig 2C⇓) shows staining patterns comparable with the circular and longitudinal smooth muscle. The staining pattern of skeletal muscle (Fig 2D⇓) is typical of most other tissues, in which the myocytes are negative, but the blood vessels are nonetheless highlighted by carbazole stain in the smooth muscle layers. The preferential distribution of tissue prothrombin in the esophagus (not shown) shows the transition from negative skeletal type fibers to smooth muscle.
Prothrombin in tissue extracts was assayed as chromogenic activity of thrombin generated with taipan snake venom and PCPS vesicles. A representative tracing of substrate hydrolysis and back-titration with recombinant hirudin is shown in Fig 3⇓. To preclude possible reactions with endogenous protease inhibitors, the venom and PCPS concentrations were configured for maximum yield within 2 minutes, and titrations were completed before 10% of the substrate had been consumed. Prothrombin activity is distributed broadly among tissues with several, including uterus, esophagus, renal medulla and myocardium, having as much or more prothrombin per gram of tissue than liver after correction for extracellular fluids (Table 1⇓). Corrections for blood were calculated from gross heme and those for blood plus interstitial fluid by gross density underlying the albumin zone on the gels (Fig 4⇓). Selected extracts assayed after activation with E carinatus venom, also Gla-dependent, likewise yielded activity (not shown).
On Western blots, the chicken anti-porcine fragment 1·2 Ig binds isolated prothrombin, prethrombin-1, and fragment 1·2 but not thrombin or fragment 1 (Fig 5⇓). Blots of myocardial extract show multiple bands, one of which intensifies with addition of internal standard prothrombin (Fig 5e⇓). None of the other bands appears to be a known fragment of prothrombin. Blots obtained from femoral artery, uterus, liver, myocardium, skeletal muscle, and esophagus show a band at the position of prothrombin, which intensifies without broadening after addition of internal standard prothrombin (Fig 6⇓). These extracts all yield additional bands of various positions and intensities that likewise appear on blots developed with preimmune Ig (Fig 6⇓, right panel). None of the blots developed with preimmune Ig shows a band at the prothrombin position.
The yield of prothrombin purified from porcine uterus greatly exceeded that attributable to whole blood and interstitial fluid contamination (Table 2⇓). Thus, 110 g of frozen pig uterus yields prothrombin equivalent to 28 mL of plasma (2750 U/100 mL) or approximately 55 mL of blood. In contrast, the yield of factor X was equivalent to that from about 3 mL of blood, which can be accounted for by the maximum estimate of 8 mL whole blood contamination calculated from heme.
Northern analysis showed hybridization between the full-length prothrombin mRNA probe and a 2.0-kb mRNA from bovine liver (Fig 7⇓); mRNA from uterus, myocardium, and skeletal muscle yielded a weak signal at 1.9 kb. The molecular size of the full-length prothrombin mRNA is 2043 bp1 and is distinct from both the 18 S (1789 bp) and 28 S (3392 bp) ribosomal RNA, which at heavy loads cross-reacted with the probe. Rehybridization with an mRNA probe for the constitutive glyceraldehyde-6-phosphate dehydrogenase (not shown) confirmed the integrity of the mRNA signal.
Extrahepatic tissues have been observed to contain structurally and functionally intact prothrombin. Henrickson et al4 found prothrombin activity associated with intact zymogen in extracts of rat uterine myocytes, and Cunningham and associates2,3 have localized prothrombin mRNA to neural structures within the central nervous system. The finding of functional prothrombin within vascular and visceral smooth muscle now provides specific cellular localization and implies that owing to the universal vascularity of tissues prothrombin is universally distributed.
Smooth muscle prothrombin antigen appears to be cytosolic; however, subcellular localization, ie, small secretion granules or caveoli, is beyond the resolution of light immunohistologic analysis. Although immunohistochemical analysis is susceptible to false positivity, including cross-reactivity, staining was obtained with four independent Ig preparations: one polyclonal and two monoclonal Ig preparations that react within the fragment 1·2 portion of prothrombin, and a third monoclonal antibody specific for thrombin. Extraneous monoclonal Ig and preimmune chicken Ig were negative. Specific staining in smooth muscle that spared adjacent cell types (eg, the media-adventitia boundary or myocytes within the portal triad) excludes artifact from microtomy as an explanation for its presence.
The prothrombin antigen extracted from tissues appears structurally and functionally identical to that of plasma origin. Activation of the zymogen with Gla-dependent O scutellatus s11 and E carinatus venoms, and recovery of prothrombin after a standard isolation procedure including Gla-dependent barium adsorption, provide two independent criteria of structural integrity and γ-carboxylation.15–17 Western analysis further indicates that the zymogen is full-length and glycosylated. In every tissue, whether yielding a faint band (eg, femoral artery) or dense multiple bands (eg, myocardium), only that ascribed to prothrombin intensified with addition of internal standard prothrombin. Absence of band broadening with addition of authentic prothrombin points to complete processing of the extract prothrombin. Existence of a minor variant specific for tissues cannot be excluded, but the intensities of bands on Western blots (Fig 6⇑ and others not shown) suggest a proportionality (semiquantitative) with activities in the same extracts. Specific inhibition with r-hirudin confirms the structural integrity of the active enzyme.
Prothrombin could arrive in smooth muscle cells either by biosynthesis or by importation from surrounding interstitial fluid, known to contain coagulation factors.12 In fetal rats, prothrombin mRNA was found in diaphragm, stomach, spleen, and adrenal extracts and persisted in the postnatal stomach.1 Cunningham and associates2,3 observed prothrombin mRNA expression in rat neural cell lines at various developmental stages. In our study, Northern analysis of tissues included poly(A)-RNA to minimize spurious hybridization. In contrast with those from liver RNA, Northern blots from extrahepatic tissues, like those reported previously for stomach and diaphragm in the embryonic rat,1 yielded sporadic signals that, when positive, lie near background, so extrahepatic biosynthesis must proceed relatively slowly, if at all. However, the faint signals arising from uterine, skeletal, and cardiac muscle appear about 100 bp smaller than prothrombin mRNA from liver. There is only one gene for (human) prothrombin,18 which is biosynthesized with both signal (pre) and carboxylase-recognition (pro) sequences. The zymogen in tissue extracts appears mature by both Western and functional assays, therefore, not the product of a partial gene, so without direct evidence to the contrary, a smooth muscle source of prothrombin biosynthesis becomes less plausible with the likelihood that the slightly smaller mRNA reflects a false-positive hybridization. Owing to the weakness of the signal from extrahepatic tissues previously reported,1–3 an alternative explanation could include spurious transcriptional leakage even if the signal does arise from prothrombin mRNA.19
Differentiated smooth muscle cells, nonsecretory in nature, lack the well-developed endoplasmic reticulum/Golgi apparatus and related organelles required for protein biosynthesis and secretion, so prothrombin uptake from the interstitial fluid might be anticipated. Nonetheless, the disparity in the yields of factor X and prothrombin from uterus and of albumin and prothrombin from uterus and other tissues implies that acquisition of tissue prothrombin is specific, as simple bulk endocytosis would not be expected to discriminate prothrombin from structurally similar proteins such as albumin and factor X. Specific membrane adsorption, either at the cell surface or in abundant surface caveoli, in principle could account for the findings; a cytosolic distribution, although not excluded, would need a pathway commensurate with endoplasmic reticulum translocation. Henrickson et al20 originally ascribed elevated concentrations of prothrombin in estrogen-primed immature rat uterus to transudation. The isolated zymogen was fully γ-carboxylated and warfarin sensitive and yielded α-thrombin and fragment 1·2 on activation with factor Xa in the presence of calcium and phospholipid. However, they later observed4 that myometrial cells isolated with trypsin yielded full-length prothrombin antigen on Western blots of cytosolic extracts, which implies also that the protein is not simply adsorbed to the plasma membrane.
A function for tissue prothrombin in hemostasis seems implausible, as plasma prothrombin is abundant and destined to arrive at any injury site accessible to platelets and fibrinogen. Cunningham and associates2,3 postulated that neural thrombin in concert with the glial-derived inhibitor, protease nexin 1, governs neurite outgrowth and astrocyte proliferation and stellation. On the basis of paradoxical phenotypes arising from mouse gene knockouts, Erickson19 has suggested that unexpected tissue distribution of a protein arises from superfluous gene leakage, which might now be extended to include spurious, albeit specific, uptake. However, the diversity of biologic activities expressed by thrombin apart from coagulation, especially mitogenic and chemotactic activities,5,6 raises the prospect that smooth muscle prothrombin functions in differentiation or cellular reactions to injury and invites consideration as a factor in vascular disease.
Weinstein JR, Gold SJ, Cunningham D, Gall CM. Cellular localization of thrombin receptor mRNA in rat brain: expression by mesencephalic dopaminergic neurons and codistribution with prothrombin mRNA. J Neurosci. 1995;15:2906–2919.
Henrickson KP, Hall ES, Lin Y. Cellular localization of tissue factor and prothrombin in the estrogen-treated immature rat uterus. Biol Reprod. 1994;50:1143–1150.
Fager G. Thrombin and proliferation of vascular smooth muscle cells. Circ Res. 1995;77:645–650.
Banfield DK, MacGillivray RT. Partial characterization of vertebrate prothrombin cDNAs: amplification and sequence analysis of the B chain of thrombin from nine different species. Proc Natl Acad Sci U S A.. 1992;89:2779–2783.
Lollar P, Knutson GJ, Fass DN. Stabilization of thrombin-activated porcine factor VIII:C by factor IXa phospholipid. Blood. 1984;63:1303–1308.
Owen WG, Esmon CT, Jackson CM. The conversion of prothrombin to thrombin. I. Characterization of the reaction products formed during the activation of bovine prothrombin. J Biol Chem. 1974;249:594–60.
Olszewski WL. Coagulation and fibrinolysis in lymph. In: Peripheral Lymph: Formation and Immune Function. Boca Raton, Fla: CRC Press; 1985:73–74.
Babson AL., Flanagan ML. Quantitative one-stage assays for factors V and X. Am J Cllin Pathol. 1975;64:817–819.
Birnboim HC, Doly J. A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucl Acids Res. 1979;7:1513–1522.
Nelsestuen GL., Broderius M, Martin G. Role of gamma-carboxyglutamic acid: cation specificity of prothrombin and factor X-phospholipid binding. J Biol Chem. 1976;251:6886–6893.
Morita T, Iwanaga S, Suzuki T. The mechanism of activation of bovine prothrombin by an activator isolated from Echis carinatus venom and characterization of the new active intermediates. J Biochem. 1976;79:1089–108.
Erickson HP. Gene knockouts of c-src, transforming growth factor beta 1, and tenascin suggest superfluous, nonfunctional expression of proteins. J Cell Biol. 1993;120:1079–1081.