High-Five Cells Research Articles

Improved expression and characterization of Ca 2+-ATPase and phospholamban in High-Five cells

The Ca 2+-ATPase accounts for the majority of Ca 2+ removed from the cytoplasm during cardiac muscle relaxation. The Ca 2+-ATPase is regulated by phospholamban, a 52 amino acid phosphoprotein, which inhibits Ca 2+-ATPase activity by decreasing the apparent affinity of the ATPase for Ca 2+. To study the physical mechanism of Ca 2+-ATPase regulation by phospholamban using spectroscopic and kinetic experiments, large amounts of both proteins are required. Therefore, we developed a Ca 2+-ATPase and phospholamban preparation based on the baculovirus-insect cell expression system using High-Five insect cells to produce large amounts of microsomal vesicles that contain either Ca 2+-ATPase expressed alone or Ca 2+-ATPase co-expressed with phospholamban. The expressed proteins were characterized using immunofluorescence spectroscopy, Ca 2+-ATPase activity assays, Ca 2+ uptake and efflux assays, and Western blotting. Our purification method yields 140 mg of microsomal protein per liter of infection (1.7 × 10 9 cells), and the Ca 2+-ATPase and phospholamban account for 16 and 1.4%, respectively, of the total microsomal protein by weight, yielding a phospholamban:Ca 2+-ATPase ratio of 1.6:1, similar to that observed in native cardiac SR vesicles. The enzymatic properties of the expressed Ca 2+-ATPase are also similar to those observed in native cardiac SR vesicles, and when co-expressed with phospholamban, the Ca 2+-ATPase is functionally coupled to phospholamban similar to that observed in cardiac SR vesicles.

Microcarrier culture of lepidopteran cell lines: implications for growth and recombinant protein production.

Several microcarrier systems were screened with Sf-9 and High-Five cell lines as to their ability to support cell growth and recombinant (beta-galactosidase) protein production. Growth of both cell lines on compact microcarriers, such as Cytodex-1 and glass beads, was minimal, as cells detached easily from the microcarrier surface and grew as single cells in the medium. Cell growth was also problematic on Cytopore-1 and -2 porous microcarriers. Cells remained attached for several days inside the microcarrier pores, but no cell division and proliferation were observed. On the contrary, insect cells grew well in the interior of Fibra-Cel disks mainly as aggregates at points of fiber intersection, reaching final (plateau) densities of about 4 x 10(6) (Sf-9) and 2.7 x 10(6) (High-Five) cells mL(-1) (8 x 10(6) and 5.5 x 10(6) cells per cm(2) of projected disk area, respectively). Their growth was described well by the logistic equation, which takes into account possible inhibition effects. Beta-Galactosidase (beta-gal) production of Sf-9 cells on Fibra-Cel disks (infected at 3.3 x 10(6) cells mL(-1)) was prolonged (192 h), and specific protein production was similar to that of high-density free cell infection. Cultispher-S microcarriers were found to be a very efficient system for the growth of High-Five cells, whereas no growth of Sf-9 cells took place for the same system. Concentrations of about 9 x 10(6) cells mL(-1) were reached within 120 h, with cell growth in both microcarriers and aggregates, appearance of cellular bridges between microcarriers and aggregates, and eventual formation of macroaggregates incorporating several microcarriers. Specific protein productions after beta-gal baculovirus infection at increasing cell concentrations were almost constant, thus leading to elevated volumetric protein production: final beta-gal titers of 946, 1728, and 1484 U mL(-1) were obtained for infection densities of 3.4, 7.2, and 8.9 x 10(6) cells mL(-1), respectively.

Supernatant proteolytic activities of High-Five insect cells grown in serum-free culture

The proteolytic activity of High-Five insect cell culture supernatants was analysed using substrate gel electrophoresis (zymography). During growth in serum-free media, High-Five cells constitutively expressed and secreted proteases that were active on casein gel but not on gelatin or bovine serum albumin gels. Two main protease bands were visible at about 41–42 kDa and 32–33 kDa. By addition of various protease inhibitors in the incubation buffer, the proteases were identified as metalloproteases as complete and specific inhibition of the proteolytic activities was only obtained by 1,10-phenanthroline.

Design of an efficient medium for insect cell growth and recombinant protein production.

We report the development of a new serum-free medium based on the use of factorial experiments. At first, a variety of hydrolysates were screened using a fractional factorial approach with High-Five cells. From this experiment yeastolate ultrafiltrate was found to have, by far, the most important effect on cell growth. Furthermore, Primatone RL was found to remarkably prolong the stationary phase of Sf-9 and High-Five cell cultures. The optimal concentrations for yeastolate and Primatone were determined to be 0.6 and 0.5%, respectively, on the basis of a complete factorial experiment. This new medium, called YPR, supported good growth of both Sf-9 and High-Five cells in batch cultures, with maximal densities of 5.4 and 6.1 x 10(6) cells/ml, respectively. In addition, both cell lines achieved good growth in bioreactor batch culture and had a prolonged stationary phase of 3-4 d in YPR medium compared to Insect-XPRESS medium. The ability of the new medium to support recombinant protein expression was also tested by infecting Sf-9 or High-Five cells at high density (2 x 10(6) cells/ml) with a baculovirus expressing secreted placental alkaline phosphatase (SEAP). The maximum total SEAP concentration after 7 d was about 43 lU/ml (58 mg/L) and 28 lU/ml (39 mg/L) for High-Five and Sf-9 cells, respectively.

Separation of Pure and Immunoreactive Virus‐Like Particles Using Gel Filtration Chromatography Following Immobilized Metal Ion Affinity Chromatography

A purification process was developed to obtain highly pure rVP2H particles, formed by a structural protein (VP2) of the infectious bursal disease virus (IBDV) with six additional histidine residues at its C-terminus. The ultimate goal was the development of an efficient subunit vaccine against IBDV infection. The particles within the infected High-Five (Hi-5) cell lysates were partially purified by employing immobilized metal ion (Ni(2+)) affinity chromatography (IMAC). The initial step could recover approximately 85% of immunoreactive rVP2H proteins but failed to separate the rVP2H particles from the free rVP2H proteins or its degraded products. To separate the particulate form from the free form of rVP2H, an additional step was added, which used either gel filtration chromatography or CsCl density gradient ultracentrifugation. Both were able to produce extremely pure rVP2H particles with a buoyant density close to 1.27 g/cm(3). However, the former method can process a larger sample volume than does the latter. By integrating IMAC and gel filtration chromatography, 1 mg of extremely pure rVP2H particles was routinely obtained from a 500 mL Hi-5 cell culture broth. The separation of the particulate form from the free form of rVP2H proteins exposes their respective immunogenicity to induce the virus-neutralizing antibodies and the ability to protect chickens from IBDV infection. Additionally, the abundant quantities of pure rVP2H particles coupled with their uniform dimensions facilitates an understanding of higher order structure of the immunogenic particles and can therefore result in improved vaccines against the virus.

High-level expression of biologically active human prolactin from recombinant baculovirus in insect cells.

We examined the feasibility of high-level production of recombinant human prolactin, a multifunctional protein hormone, in insect cells using a baculovirus expression system. The human prolactin cDNA with and without the secretory signal sequence was cloned into pFastBac1 baculovirus vector under the control of polyhedrin promoter. Prolactin was produced upon infection of either Sf9 or High-Five cells with the recombinant baculovirus containing the human prolactin cDNA. The production of recombinant prolactin varied from 20 to 40 mg/L of monolayer culture, depending on the cell types. The prolactin polypeptide with its own secretory signal was secreted into the medium. N-terminal amino acid sequence analysis of the recombinant polypeptide purified from the culture medium indicated that the protein was processed similar to human pituitary prolactin. Carbohydrate analysis of the purified protein indicated that a fraction of the recombinant prolactin made in insect cells appeared to be glycosylated. Also, both secreted and nonsecreted forms of the recombinant prolactin in insect cells were biologically equivalent to the native human prolactin (pituitary derived) in the Nb2 lymphoma cell proliferation assay.

Investigating the secretory pathway of the baculovirus-insect cell system using a secretory green fluorescent protein.

The secretory pathway is important in actively transporting proteins into the extracellular environment of eucaryotic cells. In this study a green fluorescent protein (GFP) mutant engineered to contain a secretion signal was used as a model protein in order to visualize the secretion process inside insect cells. Fluorescent microscopy indicated that significant amounts of secreted green fluorescent protein (sGFP) accumulated in High-Five, Trichoplusia ni, cells following infection with a baculovirus vector containing the gene under the polyhedrin promoter. Laser scanning confocal microscopy was used to reconstruct whole cell images of the infected High-Five cells at multiple days postinfection. While the protein was widely distributed at 2 days postinfection, certain intracellular regions appeared to contain higher or lower concentrations of the sGFP. A layer by layer examination indicated pockets in which sGFP was absent, and these appear to be vesicles that have recently released the sGFP or are not yet accumulating sGFP. By 3 days postinfection, the sGFP in some cells was concentrated in a number of widely dispersed globules, which may represent the vesicle remnants of a deteriorating secretory pathway. In contrast, nonsecreted GFP was more uniformly distributed in the cells than sGFP and did not accumulate in vesicles. In addition to GFP, the lectins wheat germ agglutinin (WGA) and concanavalin A (ConA), which have affinities for sugar residues, were used to examine the secretory pathway. The WGA, which is a Golgi marker, was distributed around the nucleus prior to infection but then was found to be polarized in one region of the cell following the baculovirus infection. The expansion of other cellular compartments following the baculovirus infection may have caused a change in intracellular distribution of the Golgi. While some of the sGFP was found to colocalize with the WGA label, much of the sGFP was outside this Golgi region. In contrast, ConA labeling, which was not as specific as WGA, was found throughout the cell both before and after infection similar to the sGFP distribution. These studies demonstrate that confocal visualization of fluorescent proteins can be used as an in vivo tool for examining secretory processing in insect cells.

Hybrid and complex glycans are linked to the conserved N-glycosylation site of the third eight-cysteine domain of LTBP-1 in insect cells.

Covalent association of LTBP-1 (latent TGF-beta binding protein-1) to latent TGF-beta is mediated by the third eight-cysteine (also referred to as TB) module of LTBP-1, a domain designated as CR3. Spodoptera frugiperda (Sf9) cells have proved a suitable cell system in which to study this association and to produce recombinant CR3, and we show here that another lepidopteran cell line, Trichoplusia niTN-5B1-4 (High-Five) cells, allows the recovery of large amounts of functional recombinant CR3. CR3 contains an N-glycosylation site, which is conserved in all forms of LTBP known to date. When we examined the status of this N-glycosylation using MALDI-TOF mass spectrometry and enzymatic analysis, we found that CR3 is one of the rare recombinant peptides modified with complex glycans in insect cells. Sf9 cells mainly processed the fucosylated paucomannosidic structure (GlcNAc)(2)(Mannose)(3)Fucose, although hybrid and complex N-glycosylations were also detected. In High-Five cells, the peptide was found to be modified with a wide variety of hybrid and complex sugars in addition to paucomanosidic oligosaccharides. Most glycans had one or two fucose residues bound through alpha1,3 and alpha1,6 linkages to the innermost GlcNAc. On the basis of these results and on the structure of an eight-cysteine domain from fibrillin-1, we present a model of glycosylated CR3 and discuss the role of glycosylation in eight-cysteine domain protein-protein interactions.

Baculovirus Expression, Purification and Evaluation of Recombinant Pneumococcal Surface Adhesin A of Streptococcus pneumoniae

Pneumococcal surface adhesin A (PsaA), with a molecular mass of ∼37 kD by SDS-PAGE, is a common surface protein expressed by all 90 serotypes of Streptococcus pneumoniae. S. pneumoniae serotype 6B genomic DNA was amplified to generate a DNA fragment carrying the full-length psaA sequence and was cloned into a baculovirus expression system. We expressed either cell-associated or cell-free nonfusion PsaA polypeptides using two insect cell lines, Spodoptera frugiperda (Sf9) and Trichoplusia ni 5B1-4 (High-Five). Recombinant PsaA (rPsaA) polypeptides were partially purified by partitioning in PBS/Triton X-114 buffers and by weakly basic ion exchange filter chromatography. Membrane-bound ‘hydrophobic rPsaA’ (hrPsaA) expressed by either Sf9 or High-Five cells had a molecular mass of ∼38 kD by SDS-PAGE and partitioned in a Triton X-114 phase, it reacted with both rabbit polyclonal and five monoclonal anti-PsaA antibodies by dot blot or Western blot analysis. High-Five-cell-expressed ‘soluble rPsaA’ (srPsaA) with a molecular mass of ∼37 kD by SDS-PAGE, was isolated from the serum-free culture medium and did not partition in the Triton X-114 phase; it reacted with anti-PsaA rabbit polyclonal and mouse monoclonal antibodies by ELISA and Western blot analysis. Both rPsaA polypeptide forms were immunogenic in Swiss-Webster adult female mice. In an infant mouse model of bacteremia, survival rates for mice given mouse anti-rPsaA immune serum (from mice immunized with High-Five-expressed srPsaA; 20 µl, 1:50,000 titer) 24 h before bacteremic challenge were greater than for the control group (48 h postchallenge, 20 vs. 90% survival rates) when challenged with S. pneumoniae serotype 6B. These results indicate that rPsaA is immunogenic and elicits protective antibody in mice similar to native protein.

Effect of modification of the β-hairpin and long loops simultaneously in both α- and β-subunits on the function of human choriogonadotropin: part II

According to the X-ray diffraction, human choriogonadotropin has four β-hairpin and two long loops, equally distributed in each of the α and β subunits. Radical mutations such as the replacement of α18Phe and α74Phe with Thr in the α 1 and α 3 loops respectively and the replacement of α45Lys with Asp in the α 2 loop in the α-subunit were introduced while the loop sequences in the β-subunit were replaced with the corresponding sequences in hFSH β. Nine different double mutants with simultaneous mutations in both the α and β loops including hCG α 1 β 1, hCG α 1 β 2, hCG α 1 β 3, hCG α 2 β 1, hCG α 2 β 2, hCG α 2 β 3, hCG α 3 β 1, hCG α 3 β 2 and hCG α 3 β 3 were partially purified from insect High-Five cells. As previously reported (Shao et al., 1996, Mol. Cell. Endocrinol. 122, 173–182), the mutation in the α 1 loop in the mutant, hCG α 1 β, the mutants hCG α 1 β 1 and hCG α 1 β 3 caused 200% increase in the receptor binding, cAMP and progesterone stimulation. The mutant, hCG α 1 β 2 and all other mutants behaved like the recombinant hCG (rehCG) in the receptor binding and post-receptor signaling activities. The molecular cause for this increase is probably due to a conformational change in the heterodimers caused by the mutation in the α 1 loop. This conclusion is based on the results of the dissociation studies of the mutants heterodimers which indicated a decreased affinity between the subunits. The first order rate constants for the dissociation of the mutants hCG α 1 β 1, hCG α1 β 2 and hCG α 1 β 3 were 3.7×10 −2 min −1, 1.4×10 −2 min −1 and 4.6×10 −2 min −1 respectively, as compared with 4.6×10 −3 min −1 for the rehCG. It seems from the data that α18Phe is located in, or in proximity to the receptor binding site and probably plays a critical role in maintaining either directly or indirectly its conformational integrity.

Effect of modification of all loop regions in the α- and β-subunits of human choriogonadotropin on its signal transduction activity

Human choriogonadotropin (hCG), according to its three dimensional structure as determined by X-ray diffraction, has three β-hairpin loops each in the α and β subunit designated as α 1, α 2 α 3 and β 1 β 2 and β 3, respectively. Since similar β-hairpin loops in NGF and TNFβ have been implicated in their direct interaction with the receptor, it prompted the present investigation to determine the role of such loops in receptor binding and post-receptor signaling events in hCG. Based on the three dimensional structure of hCG, radical mutations were introduced in the α loops by replacing hydrophobic α18Phe and α74Phe by hydrophilic Thr residues in the α 1 and α 3 loops, respectively, and positively charged α45Lys by negatively charged Asp in the helical segment in the α 2 loop. The β loops were mutated by replacement of the β 1, β 2 and β 3 sequences with the corresponding hFSH sequences. These replacements included β22Gly, β24Pro and β25Val with Glu, Arg and Phe in β 1, 45Leu Gln Gly Val Leu Pro Ala Leu Pro 53 with Tyr Lys Asn Pro Ala Arg Pro Leu Ile in β 2 and 73Pro Arg Gly with Ala His His in the β 3 loop. Six mutants, hCGα 1β, hCGα 2β and hCGα 3β and hCGαβ 1, hCGαβ 2 and hCGαβ 3, were obtained by co-infection of the insect High-Five cells with baculovirus containing mutant α or β cDNAs and that containing complimentary wild type β or α cDNAs. The mutants were almost completely secreted in the culture medium and were over expressed at levels ranging between 4.5 to 29 μg/ml indicating that mutations had no effect on the secretion or subunit assembly of hCG. In order to remove any contaminating β-subunit, the culture medium was passed through a column of a hCGβ-specific monoclonal antibody, B158. The receptor binding activity of the mutant hCGα 1β, in which α18Phe was replaced with Thr, increased almost 200% relative to rehCG. Similarly, increase in the cAMP and progesterone stimulation by the mutant ranged between 150 to 200%. This increase is believed to be due to a short range conformational change in the mutant as a result of the mutation rather than direct involvement of α18Phe in the receptor binding. The evidence in support of this was derived from the fact that the affinity or interaction between the two subunits was impaired as indicated by the first order rate constant of hCGα 1β ( k m = 4.1 × 10 −2 min −1) at pH 3.0 at 23°C which is one order of magnitude greater relative to rehCG ( k w = 4.6 × 10 −3 min −1). All other mutations had no effect on the receptor binding or signal transduction of hCG indicating that, unlike NGF or TNFβ, β-hairpin loops in hCG were not directly involved in rec binding or post-receptor signaling events. However, since the mutation in the α 1 loop affects the receptor binding site, its presence in the vicinity of the α 1 loop is highly likely.

Production of a recombinant imported fire ant venom allergen, Sol i 2, in native and immunoreactive form

Background: The complementary DNA encoding for the important imported fire ant venom allergen, Sol i 2, has previously been cloned. The binding of human IgE antibodies to Sol i 2 has been demonstrated to be conformation-dependent. Methods: A complete cDNA clone encoding the Sol i 2 protein sequence and its natural signal sequence has been produced by polymerase chain reaction. The clone was ligated into a pBluebac III transfer vector (Invitrogen Corp., San Diego, Calif.), and the recombinant baculovirus was isolated by plaque purification. The recombinant baculovirus was grown in Sf9 and High-Five cells (Invitrogen Corp.) in serum-free media. The recombinant Sol i 2 was isolated and characterized. Results: Recombinant (r) Sol i 2 was produced in microgram/per milliliter amounts in Sf9 cells and at 30 μg/ml in High-Five cells. It was isolated by ultrafiltration and reverse-phase chromatography. The rSol i 2 demonstrated similar binding to natural-Sol i 2 in both a conformation-dependent ELISA assay and in RAST with sera from patients allergic to Sol i 2. The N-terminal sequence of the rSol i 2 was identical to that of the natural molecule. No significant increase in binding activity was found after treatment of rSol i 2 with protein disulfide isomerase. The binding of rSol i 2 to a conformation-dependent monoclonal antibody was lost by heating in sodium dodecylsulfate and reduction. Conclusions: A recombinant Sol i 2 protein was produced at high yield in a baculovirus expression system by using serum-free medium with a sequence identical to that of the natural molecule. Conformation-dependent immunologic assays indicate that the recombinant protein is produced with the native conformation. (J A LLERGY C LIN I MMUNOL 1996;98:82-8.)

Effect of lipids on insect cell growth and expression of recombinant proteins in serum-free medium

The lipid emulsion components of a serum-free insect cell medium were varied and evaluated for effects on cell growth and recombinant protein expression. The growth of High-Five(TM) cells was significantly affected by polyol Pluronic F-68 and Tween-80, but not by lipids. Pluronic was essential for cell growth, while Tween-80 was required to achieve maximum cell densities. A dose response effect was observed for Tween-80 with optimal cell growth at a concentration of 25 mg/l. Cholesterol had a minor effect on cell growth, but was essential for the expression of recombinant proteins. The expression of β-galactosidase (β-gal) was directly affected by cholesterol with optimal expression at a concentration of 5.4 mg/l. Vitamin E, important as an antioxidant to stabilize lipids, did not directly affect recombinant protein expression. Although lipids were not required for cell growth, the presence of lipids were required during the cell growth phase in order to achieve efficient infection with baculovirus. These studies help to define the important components, and range of concentrations, for lipid emulsions which can effectively replace serum in insect cell culture.

Characterization of Recombinant Heparin Cofactor Ii Expressed in Insect Cells

Recombinant human heparin cofactor II (rHCII) was expressed as a fully active protein in the High-Five insect cell line. A maximal protein concentration of 6 μg/106cells was achieved 2 days postinfection. Approximately 40 μg of partially purified rHCII was routinely recovered from 50 ml of media after sequential heparin- and Q-Sepharose affinity adsorption. rHCII had a slightly lower apparent molecular weight than blood plasma HCII (pHCII) due to differences in N-glycosylation. Like pHCII, rHCII formed a stable bimolecular complex with thrombin when assessed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis. The thrombin and chymotrypsin inhibitory properties of rHCII and pHCII were quite similar. In the absence of glycosaminoglycan, the thrombin inhibition rate (k2× 10−4M−1min−1) was 2.29 ± 0.36 for rHCII and 3.38 ± 0.34 for pHCII. Chymotrypsin inhibition rates (k2× 10−5M−1min−1) were 6.2 ± 2.0 for rHCII and 8.0 ± 2.6 for pHCII. In the presence of glycosaminoglycans, the maximal thrombin inhibition rate (k2× 10−8M−1min−1) for rHCII was 10.4 ± 2.5 at 100 μg/ml heparin and 16.0 ± 4.3 at 1000 μg/ml dermatan sulfate compared to 9.0 ± 0.7 at 200 μg/ml heparin and 18.5 ± 5.3 at 1000 μg/ml dermatan sulfate for pHCII. HCII inhibition of thrombin was blocked by a synthetic sulfated hirudin peptide in both the presence and the absence of glycosaminoglycan. The present report describes for the first time the expression and characterization of HCII in a baculovirus system and demonstrates the feasibility of using this system to obtain adequate amounts of biologically active rHCII for future structure–function studies.

Horseradish Peroxidase Phe172→ Tyr Mutant: SEQUENTIAL FORMATION OF COMPOUND I WITH A PORPHYRIN RADICAL CATION AND A PROTEIN RADICAL

A gene coding for the F172Y mutant of horseradish peroxidase isozyme C (HRP) has been constructed and expressed in both Spodoptera frugiperda (SF-9) and Trichoplusia ni egg cell homogenate (HighFive) cells. Homology modeling with respect to three peroxidases for which crystal structures are available places Phe172 on the proximal side of the heme in the vicinity of porphyrin pyrrole ring C. The pH optimum and spectroscopic properties of the F172Y mutant are essentially identical to those of wild type HRP. Vmax values show that the mutant protein retains most of the guaiacol oxidizing activity. Stopped flow studies indicate that Compound I is formed with H2O2 at the same rate (kappa 1 = 1.6 x 10(7) M-1 s-1) at both pH 6.0 and 8.0 as it is with the wild type enzyme. This Compound I species decays rapidly at a rate kappa 2 = 1.01 s-1, pH 7.0, to a second two-electron oxidized species that retains the ferryl (FeIV = O) absorption. EPR studies establish that a ferryl porphyrin radical cation is present in the initial Compound I, but electron transfer from the protein results in formation of a second Compound I species with an unpaired electron on the protein (presumably on Tyr172). The presence or absence of oxidizable amino acids adjacent to the heme is thus a key determinant of whether the second oxidation equivalent in Compound I is found as a porphyrin or protein radical cation.

Mutagenesis of recombinant protein C inhibitor reactive site residues alters target proteinase specificity

Protein C inhibitor (PCI) is a heparin-binding plasma serine proteinase inhibitor (serpin) which is thought to be a physiological regulator of activated protein C. We are using recombinant PCI (rPCI) to study structural determinants of target proteinase specificity. A cDNA encoding full-length PCI has been expressed as a fully active proteinase inhibitor using Autographa californica nuclear polyhedrosis virus (baculovirus). rPCI was expressed maximally 4 days after infection and could be expressed either in Sf9 or High-Five cells. rPCI bound heparin and was conveniently purified with heparin-Sepharose (eluting > 0.5 M NaCl). The rPCI formed sodium dodecyl sulfate-polyacrylamide gel electrophoresis-stable complexes with thrombin and activated protein C (APC). The inhibitory properties of wild-type rPCI and plasma-derived PCI are essentially the same either in the absence or presence of heparin with thrombin, APC, trypsin, and urokinase. The residues Phe353-Arg354-Ser355 (P2-P1-P1') constitute part of the reactive site loop of PCI with the Arg-Ser peptide bond being cleaved by the proteinase. Using site-directed mutagenesis we studied the contribution of the reactive site FRS for proteinase inhibition in rPCI. Changing the P1 residue Arg354-->Met generated a reactive site similar to alpha 1-proteinase inhibitor which was a much poorer inhibitor of thrombin, APC, trypsin, and urokinase. Changing the P2 residue Phe353-->Gly generated a mutant with a reactive site like antithrombin which was better at inhibiting thrombin or urokinase, but was much less active with APC or trypsin. Changing the P1' residue Ser355-->Met generated a reactive site like plasminogen activator inhibitor-1 and this protein inhibits all the proteinases essentially like wild-type rPCI. These results show the importance of PCI's Phe353 (P2) and Arg354 (P1) in target proteinase specificity, and they further support the concept of reactive site sequences determining serpin function.