High-density mutagenesis by combined DNA shuffling and phage display to assign essential amino acid residues in protein-protein interactions: application to study structure-function of plasminogen activation inhibitor 1 (PAI-I).

The identification of specific amino acid residues involved in protein-protein interaction is fundamental to understanding structure-function relationships. Supported by mathematical calculations, we designed a high-density mutagenesis procedure for the generation of a mutant library of which a limited number of random clones would suffice to exactly localize amino acid residues essential for a particular protein-protein interaction. This goal was achieved experimentally by consecutive cycles of DNA shuffling, under error prone conditions, each followed by exposure of the target protein on the surface of phages to screen and select for correctly folded, functional mutants. To validate the procedure, human plasminogen activator inhibitor 1 (PAI-1) was chosen, because its 3D structure is known, many experimental tools are available and it may serve as a model protein for structure-function studies of serine proteinases and their inhibitors (serpins). After five cycles of DNA shuffling and selection for t-PA binding, analysis of 27 randomly picked clones revealed that PAI-1 mutants contained an average of 9.1 amino acid substitutions distributed over 114 different positions, which were preferentially located at the surface of the protein. This limited collection of mutant PAI-1 preparations contained multiple mutants defective in binding to three out of four tested anti-PAI-1 monoclonal antibodies. Alignment of the nucleotide sequence of defective clones permitted assignment of single dominant amino acid residues for binding to each monoclonal antibody. The importance of these residues was confirmed by testing the properties of single point mutants. From the position of these amino acid residues in the 3D structure of PAI-1 and the effects of the corresponding monoclonal antibodies on t-PA-PAI-1 interaction, conclusions can be drawn with respect to this serpin-serine proteinase interaction.

Pharmacokinetic and thrombolytic properties of cysteine-linked polyethylene glycol derivatives of staphylokinase.

Recombinant staphylokinase (SakSTAR) variants obtained by site-directed substitution with cysteine, in the core (lysine 96 [Lys96], Lys102, Lys109, and/or Lys135) or the NH(2)-terminal region that is released during activation of SakSTAR (serine 2 [Ser2] and/or Ser3), were derivatized with thiol-specific (ortho-pyridyl-disulfide or maleimide) polyethylene glycol (PEG) molecules with molecular weights of 5,000 (P5), 10,000 (P10), or 20,000 (P20). The specific activities and thrombolytic potencies in human plasma were unaltered for most variants derivatized with PEG (PEGylates), but maleimide PEG derivatives had a better temperature stability profile. In hamsters, SakSTAR was cleared at 2.2 mL/min; variants with 1 P5 molecule were cleared 2-to 5-fold; variants with 2 P5 or 1 P10 molecules were cleared 10-to 30-fold; and variants with 1 P20 molecule were cleared 35-fold slower. A bolus injection induced dose-related lysis of a plasma clot, fibrin labeled with 125 iodine ((125)I-fibrin plasma clot), and injected into the jugular vein. A 50% clot lysis at 90 minutes required 110 microg/kg SakSTAR; 50 to 110 microg/kg of core-substitution derivatives with 1 P5; 25 microg/kg for NH(2)-terminal derivatives with 1 P5; 5 to 25 microg/kg with derivatives with 2 P5 or 1 P10; and 7 microg/kg with P20 derivatives. Core substitution with 1 or 2 P5 molecules did not significantly reduce the immunogenicity of SakSTAR in rabbits. Derivatization of staphylokinase with a single PEG molecule allows controllable reduction of the clearance while maintaining thrombolytic potency at a reduced dose. This indicates that mono-PEGylated staphylokinase variants may be used for single intravenous bolus injection.

Arginine 719 in human plasminogen mediates formation of the staphylokinase:plasmin activator complex.

Staphylokinase (Sak), a 16-kDa bacterial protein, forms a 1:1 stoichiometric complex with the serine proteinase domain of human plasmin, which in turn converts other plasminogen molecules into plasmin. To identify amino acid residues critical for generating the Sak:plasmin activator complex, alanine-scanning mutagenesis was performed on phage-displayed micro-plasminogen (microPlg). Substitution of Arg719 with Ala [microPlg(R719A)] disrupted complex formation, although the sensitivity of phage-displayed microPlg(R719A) to activation by urokinase and the amidolytic activity of the micro-plasmin derivative [microPli(R719A)] remained unaffected. Likewise, the soluble microPlg(R719A) molecule did not generate a functional activator complex with Sak, whereas quantitative activation into plasmin was obtained upon incubation with either urokinase or the Sak:plasmin complex. Real-time biospecific affinity measurements revealed that the Arg --> Ala substitution at position 719 increased the equilibrium dissociation constant between microPlg(R719A) and Sak from 46 nM to 1 microM, primarily by reducing the association rate constant. Arg719 has recently also been implied in the functional complex formation between human plasmin and streptokinase [Dawson, K. M., Marshall, J. M., Raper, R. H., Gilbert, R. J., and Ponting, C. P. (1994) Biochemistry 33, 12042-12047.], suggesting that both bacterial cofactors may share common structural and/or mechanistic aspects for plasminogen activation.

Dead-end based modeling tools to explore the sequence space that is compatible with a given scaffold.

The dead-end elimination algorithm has proven to be a powerful tool in protein homology modeling since it allows one to determine rapidly the global minimum-energy conformation (GMEC) of an arbitrarily large collection of side chains, given fixed backbone coordinates. After introducing briefly the necessary background, we focus on logic arguments that increase the efficacy of the dead-end elimination process. Second, we present new theoretical considerations on the use of the dead-end elimination method as a tool to identify sequences that are compatible with a given scaffold structure. Third, we initiate a search for properties derived from the computed GMEC structure to predict whether a given sequence can be well packed in the core of a protein. Three properties will be considered: the nonbonded energy, the accessible surface area, and the extent by which the GMEC side-chain conformations deviate from a locally optimal conformation.

Screening panels of monoclonal antibodies using phage-displayed antigen.

A procedure is described to screen panels of hybridomas or purified monoclonal antibodies using antigen displayed on the surface of filamentous bacteriophage. In this system, samples containing murine monoclonal antibodies are incubated with phage-displayed antigen in microtiter plates coated with rabbit anti-mouse IgG, and bound antibody-phage complex is detected with horseradish peroxidase-sheep anti-phage M13 conjugate. The assay has been validated with a panel of 16 monoclonal antibodies directed against human plasminogen, using phage-displayed miniplasmin-(ogen) (amino acids Ala444 through Asn791 comprising kringle 5 and the proteinase domain of plasminogen) or microplasminogen (amino acids Ala543 through Asn791 comprising the proteinase domain). Six monoclonal antibodies were identified directed against miniplasminogen and miniplasmin; this was confirmed using a microtiter plate coated with antigens. One of these monoclonal antibodies (MA-42B12) did not react with microplasminogen, suggesting that its epitope is comprised within the kringle 5 domain. This test is rapid and sensitive (detecting 10-20 ng/ml of monoclonal antibody), and screening can be performed using phage-displayed zymogens or active enzymes or selected domains thereof. The procedure eliminates the need for large amounts of purified antigen for screening. Furthermore, immunization can be performed with partially purified antigen because only antibodies raised against the antigen of interest will be identified with the use of phage-displayed antigen. Therefore, this test may offer distinct advantages over the classical one-site enzyme-linked immunosorbent assay using antigen-coated microtiter plates.

Epitope mapping by negative selection of randomized antigen libraries displayed on filamentous phage.

Since most antibodies directed against protein antigens recognize epitopes composed of several discontinuous segments of the polypeptide chain, attempts to delineate the amino acids constituting these epitopes with the use of linear peptides have generally been unsuccessful. Here, a method is described based on error-prone PCR, phage display and negative selection, whereby amino acid residues constituting the functional epitope are identified in the context of the native protein. First a library of randomized antigen variants containing most single, double and triple amino acid mutants generated by single nucleotide substitutions is produced by error-prone PCR amplification of the DNA sequence encoding the protein antigen. The phage-displayed library is then negatively selected for epitope loss mutants by passing through an affinity matrix derivatized with a specific antibody and positively selected for retention of function. This method was applied to the mapping of the epitopes of two murine monoclonal antibodies (MA-7H11 and MA-3G10) on staphylokinase, a 136 amino acid plasminogen activator secreted by some strains of Staphylococcus aureus. After two negative/positive selection cycles, DNA sequencing of several clones revealed preferential amino acid mutations at positions 35 and 130 (with MA-7H11), and at positions 62, 66 and 136 (with MA-3G10). Affinity measurements of staphylokinase variants carrying single amino acid mutations at these positions confirmed their contribution to the free energy of binding to MA-7H11 and MA-3G10. This approach may be useful for isolating mutants with altered antigenic or functional properties and in general to map critical regions for protein-protein interactions.

Enzymatic properties of phage-displayed fragments of human plasminogen.

Two low-molecular-mass forms of human plasminogen, plasminogen-(543-791)-peptide (micro-plasminogen), comprising the serine protease domain, and plasminogen-(444-791)-peptide (mini-plasminogen), which in addition contains kringle 5, were displayed on filamentous phage by fusion to the N-terminus of the minor coat protein pIII, to levels of 0.5 molecules micro-plasminogen-pIII/phage particle and 0.1 molecules mini-plasminogen-pIII/phage particle. The proenzymes, quantitatively activated by urokinase, showed catalytic efficiencies that were virtually identical to their soluble counterparts, and activity remained associated with the phage as demonstrated by phage ELISA and biopanning with human alpha2-antiplasmin or the inhibitor Phe-Pro-Arg-CH2Cl. Micro-plasminogen-pIII was activated by streptokinase and staphylokinase, two non-enzymatic plasminogen activators, to the same extent as by urokinase. Activated forms of mini-plasminogen-pIII micro-plasminogen-pIII and mini-plasminogen dissolved 125I-labelled fibrin films in a dose-dependent time-dependent manner, with 50% lysis in 20 h requiring 0.52, 3.2 and 0.46 nM active plasmin, respectively. Thus, proenzyme moieties derived from plasminogen can be successfully displayed on phage with maintenance of their enzymatic properties. The micro-plasminogen and mini-plasminogen phage-display systems may be useful to study mechanisms of plasminogen activation.

Computation of the binding of fully flexible peptides to proteins with flexible side chains.

Docking algorithms play an important role in the process of rational drug design and in understanding the mechanism of molecular recognition. An important determinant for successful docking is the extent to which the configurational space (including conformational changes) of the ligand/receptor system is searched. Here we describe a new, combinatorial method for flexible docking of peptides to proteins that allows full rotation around all single bonds of the peptide ligand and around those of a large set of receptor side chains. We have simulated the binding of several viral peptides to murine major histocompatibility complex class I H-2Kb. In addition, we have explored the limits of our method by simulating a complex between calmodulin and an 18-residue long helical peptide from calmodulin-dependent protein kinase IIalpha. The calculated peptide conformations generally matched well with the X-ray structures. Essential information about local flexibility and about residues that are responsible for strong binding was obtained. We have frequently observed considerable side-chain flexibility during the simulations, showing the need for a flexible treatment of the receptor. Our method may also be useful whenever the receptor side-chain conformation is not available or uncertain, as illustrated by the docking of an H-2Kb binding nonapeptide to the receptor structure taken from an octapeptide/H-2Kb complex.

Trematode myoglobins, functional molecules with a distal tyrosine.

The myoglobins of two trematodes, Paramphistomum epiclitum and Isoparorchis hypselobagri, were isolated to homogeneity. The native molecules are monomeric with Mr 16,000-17,000 and pI 6.5-7.5. In each species, at least four different globin isoforms occur. Primary structure was determined at the protein level. The globin chains contain 147 amino acid residues. Although major determinants of the globin fold are conserved, characteristic substitutions are present. A Tyr residue occurs at the helical positions B10 and E7 (distal position). This is confirmed by NMR measurements (Zhang, W., Rashid, K. A., Haque, M., Siddiqi, A. H., Vinogradov, S. N., Moens, L. & La Mar, G. N. (1997) J. Biol. Chem. 272, 3000-3006). A distal Tyr normally provokes oxidation of the iron atom and the inability to bind oxygen, whereas a Tyr-B10 is indicative for a high oxygen affinity. In contrast, trematode myoglobins are functional molecules with a high oxygen affinity. Molecular modeling predicts two possible positions for the aromatic ring of Tyr-E7: one being outside the heme pocket making it freely accessible to the ligand and one within the heme pocket potentially able to form a second hydrogen bond with the iron-bound oxygen. A hydrogen bond between Tyr-B10 and the bound oxygen as in the Ascaris hemoglobin is predicted as well. The predicted structure may explain the high oxygen affinity of the trematode myoglobins.

Recombinant staphylokinase variants with altered immunoreactivity. III: Species variability of antibody binding patterns.

BACKGROUND: The "charged cluster-to-alanine" substitution variants SakSTAR(K35A,E38A,K74A,E75A,R77A) and SakSTAR(K74A,E75A,R77A,E80A,D82A), previously identified as SakSTAR.M38 and SakSTAR.M89, respectively, induce less antibody formation in patients than wild-type recombinant staphylokinase (SakSTAR), but their specific activities are reduced by 50%. Therefore, the effect of the reversal of one or more of these substituted amino acids on the ratio of activity to antigenicity was studied. METHODS AND RESULTS: Fourteen mutants with one to four "alanine-to-wild-type" reversals were expressed in Escherichia coli and highly purified (> 95%). In rabbits immunized with wild-type SakSTAR, the combined K35,E38,K74,E75,R77 or K74,E75,R77,E80,E82 epitope accounted for only 30% of antibody absorption from plasma, and no clear immunodominant residue could be identified. In baboons immunized with SakSTAR, the K35,E38 and K74,E75,R77 epitopes or the K74,E75,R77 and E80,D82 epitopes contributed equally to account for 50% of total antibody binding, but no immunodominant residues were apparent. In pooled plasma from patients with peripheral arterial occlusion treated with wild-type SakSTAR, about 40% of the antibodies depended on K74 of epitope K74,E75,R77 for binding, whereas epitopes K35,E38 and E80,D82 had a negligible contribution toward antibody recognition. CONCLUSIONS: The recognition pattern by SakSTAR variants of antibodies induced with wild-type SakSTAR differs markedly among species. This implies that a systematic evaluation of reduced antigen recognition and antibody induction in humans will require the development of human or humanized systems. Surprisingly, SakSTAR(K74), with a single substitution of Lys74 with Ala, had an intact specific activity but did not absorb 40% of the antibodies induced in patients by treatment with wild-type SakSTAR.

Theoretical and algorithmical optimization of the dead-end elimination theorem.

The dead-end elimination theorem has proved to be a powerful method to reduce the theoretically accessible conformational space when modeling protein side chains by using a rotameric representation of possible conformations. In this work, theoretical details about variants to the original criterion are discussed. We also provide information on how the equations can be algorithmically implemented in such a way that both computational performance and structural accuracy are optimized. In addition, we discuss the theoretical and practical aspects of three new methods called the "bottom line theorem", dead-end elimination assisted by local modeling and a combinatorial search combined with conventional dead-end elimination. It is shown that the algorithm in its current from enables the determination of the global minimum energy side chain conformation of large proteins on a time scale of hours while for small proteins of up to 30 residues the calculations are done on a time scale of seconds. The latter opens a way to combine a main chain sampling algorithm with the dead-end elimination method to locally model entire fragments of a protein chain.

All in one: a highly detailed rotamer library improves both accuracy and speed in the modelling of sidechains by dead-end elimination.

BACKGROUND: About a decade ago, the concept of rotamer libraries was introduced to model sidechains given known mainchain coordinates. Since then, several groups have developed methods to handle the challenging combinatorial problem that is faced when searching rotamer libraries. To avoid a combinatorial explosion, the dead-end elimination method detects and eliminates rotamers that cannot be members of the global minimum energy conformation (GMEC). Several groups have applied and further developed this method in the fields of homology modelling and protein design. RESULTS: This work addresses at the same time increased prediction accuracy and calculation speed improvements. The proposed enhancements allow the elimination of more than one-third of the possible rotameric states before applying the dead-end elimination method. This is achieved by using a highly detailed rotamer library allowing the safe application of an energy-based rejection criterion without risking the elimination of a GMEC rotamer. As a result, we gain both in modelling accuracy and in computational speed. Being completely automated, the current implementation of the dead-end elimination prediction of protein sidechains can be applied to the modelling of sidechains of proteins of any size on the high-end computer systems currently used in molecular modelling. The improved accuracy is highlighted in a comparative study on a collection of proteins of varying size for which score results have previously been published by multiple groups. Furthermore, we propose a new validation method for the scoring of the modelled structure versus the experimental data based upon the volume overlap of the predicted and observed sidechains. This overlap criterion is discussed in relation to the classic RMSD and the frequently used +/- 40 degrees window in comparing chi 1 and chi 2 angles. CONCLUSIONS: We have shown that a very detailed library allows the introduction of a safe energy threshold rejection criterion, thereby increasing both the execution speed and the accuracy of the modelling program. We speculate that the current method will allow the sidechain prediction of medium-sized proteins and complex protein interfaces involving up to 150 residues on low-end desktop computers.

Anticoagulant repertoire of the hookworm Ancylostoma caninum.

Hookworms are hematophagous nematodes that infect a wide range of mammalian hosts, including humans. There has been speculation for nearly a century as to the identity of the anticoagulant substances) used by these organisms to subvert host hemostasis. Using molecular cloning, we describe a family of potent small protein (75-84 amino acids) anticoagulants from the hookworm Ancylostoma caninum termed AcAP (A. caninum anticoagulant protein). Two recombinant AcAP members (AcAP5 and AcAP6) directly inhibited the catalytic activity of blood coagulation factor Xa (fXa), while a third form (AcAPc2) predominantly inhibited the catalytic activity of a complex composed of blood coagulation factor VIIa and tissue factor (fVIIa/TF). The inhibition of fVIIa/TF was by a unique mechanism that required the initial formation of a binary complex of the inhibitor with fXa at a site on the enzyme that is distinct from the catalytic center (exo-site). The sequence of AcAPc2 as well as the utilization of an exo-site on fXa distinguishes this inhibitor from the mammalian anticoagulant TFPI (tissue factor pathway inhibitor), which is functionally equivalent with respect to fXa-dependent inhibition of fIIa/TF. The relative sequence positions of the reactive site residues determined for AcAP5 with the homologous regions in AcAP6 and AcAPc2 as well as the pattern of 10 cysteine residues present in each of the inhibitors suggest that the AcAPs are distantly related to the family of small protein serine protease inhibitors found in the nonhematophagous nematode Ascaris lumbricoides var. suum.

Enhanced dead-end elimination in the search for the global minimum energy conformation of a collection of protein side chains.

Although the conformational states of protein side chains can be described using a library of rotamers, the determination of the global minimum energy conformation (GMEC) of a large collection of side chains, given fixed backbone coordinates, represents a challenging combinatorial problem with important applications in the field of homology modelling. Recently, we have developed a theoretical framework, called the dead-end elimination method, which allows us to identify efficiently rotamers that cannot be members of the GMEC. Such dead-ending rotamers can be iteratively removed from the system under study thereby tracking down the size of the combinatorial problem. Here we present new developments to the dead-end elimination method that allow us to handle larger proteins and more extensive rotamer libraries. These developments encompass (i) a procedure to determine weight factors in the generalized dead-end elimination theorem thereby enhancing the elimination of dead-ending rotamers and (ii) a novel strategy, mainly based on logical arguments derived from the logic pairs theorem, to use dead-ending rotamer pairs in the efficient elimination of single rotamers. These developments are illustrated for proteins of various sizes and the flow of the current method is discussed in detail. The effectiveness of dead-end elimination is increased by two orders of magnitude as compared with previous work. In addition, it now becomes feasible to use extremely detailed libraries. We also provide an appendix in which the validity of the generalized dead-end criterion is shown. Finally, perspectives for further applications which may now become within reach are discussed.

NMR structure determination of tick anticoagulant peptide (TAP).

Tick anticoagulant peptide (TAP) is a potent and selective 60-amino acid inhibitor of the serine protease Factor Xa (fXa), the penultimate enzyme in the blood coagulation cascade. The structural features of TAP responsible for its remarkable specificity for fXa are unknown, but the binding to its target appears to be unique. The elucidation of the TAP structure may facilitate our understanding of this new mode of serine protease inhibition and could provide a basis for the design of novel fXa inhibitors. Analyses of homo- and heteronuclear two-dimensional NMR spectra (total correlation spectroscopy, nuclear Overhauser effect spectroscopy [NOESY], constant time heteronuclear single quantum correlation spectroscopy [CT-HSQC], and HSQC-NOESY; 600 MHz; 1.5 mM TAP; pH 2.5) of unlabeled, 13C-labeled, and 15N-labeled TAP provided nearly complete 1H sequence-specific resonance assignments. Secondary structural elements were identified by characteristic NOE patterns and D2O amide proton-exchange experiments. A three-dimensional structure of TAP was generated from 412 NOESY-derived distance and 47 dihedral angle constraints. The structural elements of TAP are similar in some respects to those of the Kunitz serine protease inhibitor family, with which TAP shares weak sequence homology. This structure, coupled with previous kinetic and biochemical information, confirms previous suggestions that TAP has a unique mode of binding to fXa.

The fuzzy-end elimination theorem: correctly implementing the side chain placement algorithm based on the dead-end elimination theorem.

Recently it has been shown that the dead-end elimination theorem is a powerful tool in the search for the global minimum energy conformation (GMEC) of a large collection of protein side chains given known backbone coordinates and a library of allowed side chain conformational states, also known as rotamers. A side chain placement algorithm based on this theorem iteratively applies this theorem to single as well as to pairs of rotamers leading to the identification of rotamers, single or pairs, that are incompatible with the GMEC and that can thus be qualified as 'dead-ending'. Here we formulate a theorem which proves that contrary to intuition dead-end rotamer pairs cannot simply be discarded from consideration in the iterative process leading to the further elimination of dead-end rotamers. We refer to this theorem as the fuzzy-end elimination theorem. We also describe how the obtained dead-end rotamer pairs can contribute to the search for the GMEC in the protein side chain placement problem. Hence the present work forms a theoretical basis for the correct implementation of a side chain placement algorithm based on the dead-end elimination theorem. In addition, possible future perspectives are presented.

Protein engineering of xylose (glucose) isomerase from Actinoplanes missouriensis. 1. Crystallography and site-directed mutagenesis of metal binding sites.

The structure and function of the xylose (glucose) isomerase from Actinoplanes missouriensis have been analyzed by X-ray crystallography and site-directed mutagenesis after cloning and overexpression in Escherichia coli. The crystal structure of wild-type enzyme has been refined to an R factor of 15.2% against diffraction data to 2.2-A resolution. The structures of a number of binary and ternary complexes involving wild-type and mutant enzymes, the divalent cations Mg2+, Co2+, or Mn2+, and either the substrate xylose or substrate analogs have also been determined and refined to comparable R factors. Two metal sites are identified. Metal site 1 is four-coordinated and tetrahedral in the absence of substrate and is six-coordinated and octahedral in its presence; the O2 and O4 atoms of linear inhibitors and substrate bind to metal 1. Metal site 2 is octahedral in all cases; its position changes by 0.7 A when it binds O1 of the substrate and by more than 1 A when it also binds O2; these bonds replace bonds to carboxylate ligands from the protein. Side chains involved in metal binding have been substituted by site-directed mutagenesis. The biochemical properties of the mutant enzymes are presented. Together with structural data, they demonstrate that the two metal ions play an essential part in binding substrates, in stabilizing their open form, and in catalyzing hydride transfer between the C1 and C2 positions.

Protein engineering of xylose (glucose) isomerase from Actinoplanes missouriensis. 2. Site-directed mutagenesis of the xylose binding site.

Site-directed mutagenesis in the active site of xylose isomerase derived from Actinoplanes missouriensis is used to investigate the structural and functional role of specific residues. The mutagenesis work together with the crystallographic studies presented in detail in two accompanying papers adds significantly to the understanding of the catalytic mechanism of this enzyme. Changes caused by introduced mutations emphasize the correlation between substrate specificity and cation preference. Mutations in both His 220 and His 54 mainly affect the catalytic rate constant, with catalysis being severely reduced but not abolished, suggesting that both histidines are important, but not essential, for catalysis. Our results thus challenge the hypothesis that His 54 acts as an obligatory catalytic base for ring opening; this residue appears instead to be implicated in governing the anomeric specificity. With none of the active site histidines acting as a catalytic base, the role of the cations in catalyzing proton transfer is confirmed. In addition, Lys 183 appears to play a crucial part in the isomerization step, by assisting the proton shuttle. Other residues also are important but to a lesser extent. The conserved Lys 294 is indirectly involved in binding the activating cations. Among the active site aromatic residues, the tryptophans (16 and 137) play a role in maintaining the general architecture of the substrate binding site while the role of Phe 26 seems to be purely structural.