Characterization and quality control of recombinant adenovirus vectors for gene therapy.

Highly purified recombinant adenovirus undergoes routine quality controls for identity, potency and purity prior to its use as a gene therapy vector. Quantitative characterization of infectivity is measurable by the expression of the DNA binding protein, an early adenoviral protein, in an immunofluorescence bioassay on permissive cells as a potency determinant. The specific particle count, a key quality indicator, is the total number of intact particles present compared to the number of infectious units. Electron microscopic analysis using negative staining gives a qualitative biophysical analysis of the particles eluted from anion-exchange HPLC. One purity assessment is accomplished via the documented presence and relative ratios of component adenoviral proteins as well as potential contaminants by reversed-phase HPLC of the intact virus followed by protein peak identification using MALDI-TOF mass spectrometry and subsequent data mining. Verification of the viral genome is performed and expression of the transgene is evaluated in in vitro systems for identity. Production lots are also evaluated for replication-competent adenovirus prior to human use. For adenovirus carrying the human IL-2 transgene, quantitative IL-2 expression is demonstrated by ELISA and cytokine potency by cytotoxic T lymphocyte assay following infection of permissive cells. Both quantitative and qualitative analyses show good batch to batch reproducibility under routine test conditions using validated methods.

Analysis of the lipidated recombinant outer surface protein A from Borrelia burgdorferi by mass spectrometry.

The outer surface protein A, OspA, from the spirochete Borrelia burgdorferi is a lipoprotein of 25 kDa. The recombinant OspA (rOspA) expressed in Escherichia coli has been purified and analyzed by electrospray mass spectrometry (ESMS). A heterogenous spectrum gave a measured mass of 28,462 +/- 9 Da for the major component compared to an expected mass of 28,445 Da (Deltam = +17 Da), and a measured mass of 28,228 +/- 7 Da for a minor component. The theoretical mass is based on the N-terminal being an S-[2,3-bis(palmitoyloxy)-(2R, S-[2,3-bis(palmitoyloxy)-(2R,S)-propyl]-N-palmitoylcysteine modification according to the model established by Hantke and Braun (Eur. J. Biochem. 34, 284-296, 1973) for bacterial lipoproteins. To determine whether rOspA conforms to this model, a complementary detailed analysis of this lipidation was necessary. The fatty acid content of the complete protein as analyzed by gas chromatography-mass spectrometry revealed saturated fatty acids ranging from C14 to C18 as well as C16 and C18 unsaturated fatty acids, with palmitate (C16:0) being the major component. Focusing on the lipid moieties, the N-terminal tryptic peptide was purified by normal phase HPLC using a silica column and a gradient of hexane in isopropanol. Analysis of the N-terminal peptide by ESMS and fast atom bombardment-mass spectrometry revealed a minor fraction of rOspA molecules which contained only two C16 residues and that in addition to partial oxidation, the major N-terminal peptide had a mass difference of -2 Da compared to a theoretical structure with three palmitate residues, indicating that one of the three fatty acid residues was unsaturated. Minor forms with mass differences of 28 Da were also observed, indicating that one of the three acyl residues was C14 in one case and C18 in the other, instead of C16 in the major form. Analysis of the rOspA peptide backbone revealed that the sole methionine at position 22 was partially oxidized to a methionine sulfoxide. Thus the mass analysis of the major mass is consistent with a mixed population of lipoprotein molecules containing variations not only in the lipid moiety contributing to an elevation in the mass of Deltam = 7 Da compared to the theoretical structure proposed, but also in the peptide chain. Partial oxidations at two points in the protein backbone (<30% of the population in each case) contribute to an additional augmentation in mass and thus can account for the remaining mass difference in the measured mass.

Analysis of the primary structure and post-translational modifications of the Schistosoma mansoni antigen Smp28 by electrospray mass spectrometry.

The Schistosoma mansoni glutathione-S-transferase with an apparent molecular mass of 28 kDa, Smp28, has a blocked N-terminus which has been elucidated with the aid of the cDNA sequence combined with mass spectrometry and amino acid composition analysis of the N-terminal tryptic peptide. The blocked N-terminal tryptic peptide (m/z 695.8) contained an equimolar ratio of E, G, H, A, I and K3 upon amino acid composition analysis in agreement with its expected sequence AGEHIK, and showed a delta m = +41.7 Da compared to the predicted mass, which is consistent with the N-terminal alanine being acetylated (delta m = +42.0 Da). The mass of the complete molecule (23,744.5 +/- 3.3 Da) determined by electrospray mass spectrometry showed a further mass increase of 14 Da with respect to Smp28 containing an N-acetylated alanine. This result is consistent with one of the seven methionines being present as a methionine sulfoxide in ca. 90% of the Smp28 molecules in this preparation. Tryptic mapping of Smp28 showed five of the seven methionines to be partially oxidized by mass spectrometry. This is indicative of the ease with which this modification occurs. Two minor components were detected along with the intact molecule, corresponding to modified forms of the molecule, originating from reaction of the only cysteine residue either with itself forming a covalent dimer or with glutathione. On-line liquid chromatography-mass spectrometry has been compared with the off-line complete tryptic map of Smp28 confirming 97% of the primary structure in less than 2 h.

Anti-hirudin monoclonal antibodies directed toward discontinuous epitopes of the hirudin amino-terminal and epitopes involving the carboxy-terminal hirudin amino acids.

A panel of eight monoclonal antibodies (MAbs) was obtained against recombinant hirudin variant 2 (rHV2). Specificities of the eight MAbs indicate that four of them recognize C-terminal amino acid residues (Group A) and four are directed against discontinuous epitopes and recognize a determinant (or determinants) within the 43 N-terminal residues (Group B). Using these antibodies recombinant hirudins missing one or more C-terminal amino acids can be distinguished from molecules with an intact C-terminus either in enzyme immunoassays (EIAs) or by immunoaffinity chromatography. A sandwich EIA using the combination of two antibodies, one from each group, can quantitate both recombinant hirudin variant 1 (rHV1) and rHV2 with a detection range from 1 to 10 ng/ml in either buffer or plasma. Using only one MAb a competitive antibody capture EIA can quantitate recombinant or natural hirudin variants 1, 2, and 3 with a detection range from 5 to 100 ng/ml for rHV2 with a lysine in position 47 (rHV2K47). None of the antibodies recognizes hirudin after it is complexed to alpha-thrombin. The ability of any one of these anti-rHV2 antibodies to interfere with hirudin binding to alpha-thrombin as measured by inhibition of thrombin's amidolytic activity correlates with the range of MAb affinity constants (KD = 3.5 x 10(-9) to 1 x 10(-6) M). Incubating hirudin with one antibody from Group A (KD = 3.5 x 10(-8) M) and one from Group B (KD = 6.0 x 10(-9) M) completely blocks the ability of hirudin to bind alpha-thrombin. This MAb panel is thus useful for probing the recombinant C-terminal integrity of hirudin, for sensitive free hirudin quantitations, and the combined use of two MAbs has potential applications as an antidote for hirudin in vivo.

Sequence-specific deamidation: isolation and biochemical characterization of succinimide intermediates of recombinant hirudin.

Natural hirudin variant 2 with a lysine residue in position 47 (rHV2-Lys47) was produced in a genetically engineered strain of Saccharomyces cerevisiae as a secreted protein of 65 amino acids and purified to greater than 99% homogeneity. Only reversed-phase high-performance liquid chromatography (RP-HPLC) using very shallow acetonitrile gradients indicated the presence of a component in the final product (approximately 1% of total protein) with a slightly increased retention time. Using successive RP-HPLC purification steps, this hydrophobic impurity was isolated and separated into two constituents defined as components A1 and A2 which differed from the parent molecule by mass reductions of 17.2 Da (A1) and 17.6 Da (A2), respectively, as determined by electrospray mass spectrometry (ESMS). Proteolytic digestion with endoprotease Glu-C from Staphylococcus aureus (V8 protease) and analysis of the peptide mixture by ESMS showed that the mass difference between rHV2-Lys47 and component A1 was due to a modification between amino acids 1 and 43, while the corresponding mass difference with component A2 was the result of a modification within the peptide fragment comprising residues 50-61. Further analyses using amino acid sequencing and ESMS in combination with collision-activated dissociation (CAD) detected modifications at residues Asn33-Gly34 in component A1 and at Asn53-Gly54 in component A2. Both of these sites were previously shown to be susceptible to spontaneous deamidation under slightly basic pH conditions. Thus, the mass reductions of approximately 17 Da and the fact that both asparagines, Asn33 in component A1 and Asn53 in component A2, proved to be resistant to Edman degradation provided strong support for them being stable succinimide intermediates of the corresponding deamidation reactions. Both intermediates were shown to have inhibition constants for human alpha-thrombin on the order of 1 pM, identical to that of rHV2-Lys47. The isoelectric point of component A2 was determined to be within 0.01 pH unit of that of the parent molecule by isoelectric focusing in an immobilized pH gradient.

Characterization of the deamidated forms of recombinant hirudin.

Recombinant hirudin variant rHV2-Lys 47 (MW = 6906.5) was intentionally deamidated by incubation in pH 9 phosphate buffer at 37 degrees C. Anion-exchange HPLC analysis showed that 11 forms could be generated. These were isolated and purified by combined anion-exchange and reversed-phase HPLC. Acid-catalyzed carboxyl methylation was used to introduce a mass shift of +15 amu per deamidated residue present in the molecule before analysis by liquid secondary ion mass spectrometry (LSIMS). Methylation enhanced, in particular, the abundance of the sequence ions in the LSIMS spectra. This permitted the determination of both the number (three) and the localization of the deamidated residues: Asn 52, Asn 53, and a residue located in the N-terminal 1-39 domain. Complementary sequencing techniques proved that the latter residue was Asn 33. Altogether four mono-, three di-, and four tri-deamidated forms were identified. The heterogeneity of the forms having identical deamidation positions but being chromatographically separable is thought to arise from the generation of alpha- and beta-aspartyl iso forms during the nonenzymatic deamidation process.

Analysis of recombinant proteins by isoelectric focusing in immobilized pH gradients.

Isoelectric focusing in immobilized pH gradients (IEF-IPG) was used to analyze three different recombinant proteins. Recombinant leech hirudin (65 amino acids, three disulfide bonds) expressed in Saccharomyces cerevisiae as a secreted protein and purified by anion-exchange and reversed-phase chromatography proved to be homogeneous with regard to its isoelectric point (pI). In addition, the theoretical pI, calculated on the basis of the primary structure, corresponded precisely to the measured pI of 4.30. IEF-IPG was further employed to follow the stability of recombinant hirudin at pH 9, indicating that deamidation occurred under these conditions. A variant of recombinant human alpha 1-antitrypsin (AAT) (389 amino acids, one cysteine residue) expressed in Escherichia coli and purified by anion-exchange, metal chelate and hydrophobic-interaction chromatography appeared to be homogeneous by polyacrylamide gel electrophoresis under reducing and denaturing conditions as well as by various high performance liquid chromatography methods. However, some heterogeneity was detected by IEF-IPG between pH 5-6. The measured pI values of 5.43-5.58 were slightly lower than the calculated pI based on the primary structure (5.72). This indicated deamidations of Asn or Gln residues. A recombinant Schistosoma mansoni parasite antigen, p28 (210 amino acids, one cysteine residue) obtained after intracellular expression in Saccharomyces cerevisiae and affinity purification on glutathione agarose was analyzed by IEF-IPG in a pH 7.3-8.3 gradient. It appeared to be heterogeneous with regard to its pI, with the major component having a pI of 7.81 compared to the calculated value of 7.17.(ABSTRACT TRUNCATED AT 250 WORDS)

Purification and biochemical characterization of recombinant alpha 1-antitrypsin variants expressed in Escherichia coli.

Site-directed variants of alpha 1-antitrypsin (alpha 1AT) expressed in a recombinant strain of Escherichia coli have been isolated with an overall process yield of 50% following tangential flow ultrafiltration, anion-exchange, immobilized metal affinity, and hydrophobic interaction chromatography. The primary structure of the purified variants including the integrity of the N- and C-termini has been verified by electrospray mass spectrometry of the intact molecules (44 kDa) for two of the variants (alpha 1AT Leu-358 and alpha 1AT Ala-357, Arg-358). Complementary classical peptide mapping and automated amino acid sequencing have verified 75% of the primary sequence of alpha 1AT Ala-357, Arg-358. Isoelectric focusing in an immobilized pH gradient revealed some microheterogeneity which proved to be reproducible from one purification batch to another. The isolated variants of alpha 1AT did not show any signs of proteolytic degradation during the purification process and proved to be fully active against their target proteases. The described process also allowed the complete removal of endotoxins from the preparations, opening the possibility to evaluate these novel protease inhibitors for their in vivo efficacy in different animal models of human disease.

Determination of disulfide bridges in natural and recombinant insect defensin A.

The primary-structure comparison of natural insect defensin A from Phormia terranovae and recombinant insect defensin A from Saccharomyces cerevisiae has been accomplished using a combination of Edman degradation and liquid secondary ion mass spectrometry. The natural and recombinant proteins have the same primary structure with identical disulfide-bond designations (formula; see text) as determined from the peptides obtained after thermolysin digestion. The combined use of Edman degradation and mass spectometry allowed the disulfide-bridge structure to be determined with a total of only 40 micrograms (9.9 nmol) natural peptide. Mass spectrometry provides a rapid means of disulfide-bridge verification, requiring not more than 20 micrograms recombinant insect defensin A, which is compatible with use in batch analysis.

Thrombin cleavage analysis of a novel antihaemophilic factor variant, factor VIII delta II.

Factor VIII delta II is a genetically engineered deletion variant of factor VIII expressed by recombinant Chinese hamster ovary cells, in which a major portion of the central (B) domain and a part of the light chain (Pro771-Asp1666) are missing. After immunoaffinity purification, the kinetics of thrombin cleavage of the novel molecule was analysed by SDS/PAGE, Western blotting and N-terminal amino acid sequencing. Thrombin first cleaves factor VIII delta II at Arg740-Ser741 to generate the 90-kDa heavy chain and an 80-kDa fusion polypeptide consisting of the remaining portion of the B domain and the 73-kDa light chain. The 90-kDa fragment is further cleaved, giving rise to 50-kDa and 40-kDa fragments while the 80-kDa fragment generates a 71/73-kDa doublet. The 71/73-kDa doublet, 50-kDa and 40-kDa fragments were further analysed by N-terminal amino acid sequencing and found to correspond to the predicted amino acid sequences. Our study shows that, in spite of the 900 amino acid deletion present in factor VIII delta II, the essential structural elements required for thrombin activation are conserved.

Application of electrospray mass spectrometry to the characterization of recombinant proteins up to 44 kDa.

Mass measurement by electrospray mass spectrometry (ESMS) is used as a rapid preliminary verification of the identity of various recombinant proteins ranging from 7 to 44 kDa with an accuracy of 0.01-0.03%. ESMS not only improves the speed but also the reliability of the protein structure determination when used in conjunction with other methods of protein analysis. Modifications of these large molecules, for example the loss of C-terminal amino acids, N-terminal acetylation, 2-mercaptoethanol addition to a cysteine, and trace formation of a covalent dimer (3%), are easily detected individually or in mixtures by mass measurement using ESMS; feats which would be very difficult to achieve using classical biochemical methods. As little as 1% of several structurally related protein contaminants have been identified in a 15 kDa recombinant protein preparation.

The structure of a complex of recombinant hirudin and human alpha-thrombin.

The crystallographic structure of a recombinant hirudin-thrombin complex has been solved at 2.3 angstrom (A) resolution. Hirudin consists of an NH2-terminal globular domain and a long (39 A) COOH-terminal extended domain. Residues Ile1 to Tyr3 of hirudin form a parallel beta-strand with Ser214 to Glu217 of thrombin with the nitrogen atom of Ile1 making a hydrogen bond with Ser195 O gamma atom of the catalytic site, but the specificity pocket of thrombin is not involved in the interaction. The COOH-terminal segment makes numerous electrostatic interactions with an anion-binding exosite of thrombin, whereas the last five residues are in a helical loop that forms many hydrophobic contacts. In all, 27 of the 65 residues of hirudin have contacts less than 4.0 A with thrombin (10 ion pairs and 23 hydrogen bonds). Such abundant interactions may account for the high affinity and specificity of hirudin.

Inhibition by recombinant hirudins of experimental venous thrombosis and disseminated intravascular coagulation induced by tissue factor in rats.

Antithrombotic potency of recombinant hirudins rHV2, rHV2-Lys47 and rHV2-Arg47 was studied in a model of experimental thrombosis induced by tissue factor in the rat. Venous thrombosis was induced by i.v. injection of 25 mg/kg tissue factor followed by stasis of the inferior vena cava. In this model natural recombinant hirudins, rHV2 and rHV2-Lys47 injected 5 min before thrombo-plastin totally inhibited thrombosis in the same micrograms range as heparin or natural hirudin extracted from leeches. However, the mutant variant rHV2-Arg47 gave a maximal 60% inhibition of thrombosis. Variants rHV2-Lys47 (30 micrograms/kg) and rHV2-Arg47 (157 micrograms/kg) injected 5 min before thromboplastin prevented by 90 to 100% the drop in platelet count observed during the disseminated intravascular coagulation induced by thromboplastin injection. Recombinant hirudins were less anticoagulant than heparin as measured by an APTT on rat plasma. After rat tail transection, rHV2-Lys47 caused a 2-fold smaller prolongation of the bleeding time than an equivalent antithrombotic dose of heparin. Plasmatic elimination of rHV2-Lys47 from rat plasma after i.v. injection had a fast distribution phase with a half-life of 3 min during which 90% of injected rHV2-Lys47 was lost and was followed by a slower elimination phase. Thus recombinant hirudin rHV2-Lys47 appears as a promising potent antithrombotic agent for the prevention of thrombin-dependent venous thrombosis and disseminated intravascular coagulation.

High level of expression of a protective antigen of schistosomes in Saccharomyces cerevisiae.

Strains of Saccharomyces cerevisiae expressing P28-I, an antigen inducing protection against schistosomiasis, have been constructed. Transformants containing a very high copy number of a P28-I expression vector were selected by genetic complementation involving deficient LEU2 or URA3 alleles carried by plasmids. Using the ura3 fur1 auto-selection system, constitutive and stable expression of P28-I could be obtained in cultures grown in rich medium. The accumulation of the foreign protein exceeds 25% of total yeast proteins when estimated by Coomassie Brilliant Blue staining of SDS-PAGE. Moreover, P28-I which was located intracellularly was soluble and biologically active.

Evaluation of the inhibition by heparin and hirudin of coagulation activation during r-tPA-induced thrombolysis.

Thrombin bound to a fibrin clot remains active and poorly accessible to heparin-AT III complex. During fibrinolysis, thrombin is released as thrombin-FDP complex and is inactivated by heparin-AT III. However, as successive fibrin layers are removed, inaccessible molecules of thrombin are exposed at the surface of the residual clot, possibly contributing to the occurrence during thrombolytic therapy of coagulation that is poorly controlled by heparin. We have investigated the accessibility of fibrin-bound thrombin to hirudin. The results clearly show that two recombinant hirudin variants neutralize thrombin both in solution and fibrin bound. Furthermore, we have found that in in vitro models, hirudin present in the surrounding medium of a clot under lysis is more efficient than heparin in preventing the activation of coagulation. This observation suggests that hirudin may be effective in the prevention of the rethrombotic process frequently encountered during thrombolytic therapy.

Isolation of recombinant hirudin by preparative high-performance liquid chromatography.

The purification of recombinant hirudin variant 2-Lys47 (rHV2-Lys47), produced by a genetically engineered yeast strain, is described. rHV2-Lys47 expressed and secreted into the culture medium was the starting material for the purification process of hirudin from the culture broth after cell harvesting by centrifugation. Initial purification of the product by preparative reversed-phase high-performance liquid chromatography (HPLC) using step-gradient elution, followed by precipitation of rHV2-Lys47 in the presence of acetone, removed most of the contaminants from the culture medium. The pure product was obtained by successive preparative anion-exchange and reversed-phase HPLC on silica based stationary phases. Characterization of the final product by analytical HPLC, isoelectric focusing gel electrophoresis, quantitative amino acid composition and sequence analysis did not reveal any contaminants. Liquid secondary ion mass spectrometry was used to confirm its primary structure. The isolated product was tested in an inhibition assay of human alpha-thrombin and proved to be fully active.

Isolation of recombinant partial gag gene product p18 (HIV-1Bru) from Escherichia coli.

The membrane-associated structural protein, p18, of the human immunodeficiency virus (HIV-1), has been expressed in Escherichia coli. The recombinant protein was purified by cation-exchange chromatography on S Sepharose followed by cation-exchange high-performance liquid chromatography (HPLC) on Sulfoethyl Aspartamide. The isolation of 28.7 mg of recombinant p18 from 16.71 of cell culture represents an overall yield of ca. 20%. Recombinant p18 was characterized by sodium dodecyl sulphate polyacrylamide gel electrophoresis, reversed-phase HPLC, amino acid composition and amino acid sequence analysis of the N-terminus. Edman degradation of peptides generated by trypsin or Staphylococcus aureus V8 proteolytic digestion, including the C-terminus, confirmed the amino acid sequence to be that predicted from the cDNA. A C-terminally cleaved form of recombinant p18, p18LM, was separated in the cation-exchange HPLC step and was partially characterized in parallel with the intact molecule. By Western blotting it was shown that recombinant p18 in addition to the cleaved form p18LM is recognized by a monoclonal antibody which was generated against the natural protein from HIV-1.

Mass spectrometry analyses of recombinant hirudins (7 kDa).

The use of liquid secondary ion mass spectrometry (LSIMS) in the characterization of related recombinant 7-kDa peptides illustrates the adequacy of average mass measurement by scanning at low resolution. The difficulty in using the high-resolution technique in the case of poor LSIMS sensitive peptides is discussed, as well as the fact that it does not give, for these molecular weights, any real advantage. The average (or chemical) molecular weights of three recombinant hirudin molecules, hirudin variant 2 (rHV2, 6892.4 Da), hirudin variant 2-Lys47 (rHV2-Lys47, 6906.5 Da), and hirudin variant 2-Arg47 (rHV2-Arg47, 6934.5 Da), less than or equal to 10 micrograms each, have been measured with an accuracy less than or equal to 0.3 Da in the narrow-scan mode and less than or equal to 0.5 Da (from the protonated molecular ion) in the wide-scan mode within 10-15 min; this allows easy distinction of the three 65 amino acid proteins, which differ by a single amino acid. These three molecules could also be distinguished from one another in a mixture. Mass spectrometry and limited sequence characterization of several minor, similarly isolated peptides identified them to be N-terminal additions and/or C-terminal deletions of rHV2-Lys47. LSIMS analysis is consistent with there being no covalent dimer of rHV2-Lys47 as a narrow scan of the 7-kDa molecular ion cluster at high resolution shows it not to be a doubly charged ion.

Purification and biochemical characterization of recombinant hirudin produced by Saccharomyces cerevisiae.

Recombinant hirudin was produced by the yeast Saccharomyces cerevisiae using the alpha-pheromone prepro sequence to direct its secretion into the culture medium. The secreted hirudin was isolated to greater than or equal to 95% purity as measured by 205-nm absorbance integration from a reverse-phase chromatogram. One major activity peak corresponding to the complete, correctly processed molecule and two minor activity peaks corresponding to C-terminally truncated forms were identified. The primary structure of the major peak, determined by N-terminal sequencing of tryptic peptides, was that predicted from the cDNA sequence, and the molecular mass analyzed by fast atom bombardment mass spectrometry (FAB-MS) was 6892.6 (calculated 6892.5). UV spectral analysis suggested that, in contrast to the natural molecule, recombinant hirudin produced by S. cerevisiae is not sulfated.