Structural studies of binding site tryptophan mutants in the high-affinity streptavidin-biotin complex.

Previous thermodynamic and computational studies have pointed to the important energetic role of aromatic contacts in generating the exceptional binding free energy of streptavidin-biotin association. We report here the crystallographic characterization of single site tryptophan mutants in investigating structural consequences of alterations in these aromatic contacts. Four tryptophan residues, Trp79, Trp92, Trp108 and Trp120, play an important role in the hydrophobic binding contributions, which along with a hydrogen bonding network and a flexible binding loop give rise to tight ligand binding (Ka approximately 10(13) M-1). The crystal structures of ligand-free and biotin-bound mutants, W79F, W108F, W120F and W120A, in the resolution range from 1.9 to 2.3 A were determined. Nine data sets for these four different mutants were collected, and structural models were refined to R-values ranging from 0.15 to 0.20. The major question addressed here is how these mutations influence the streptavidin binding site and in particular how they affect the binding mode of biotin in the complex. The overall folding of streptavidin was not significantly altered in any of the tryptophan mutants. With one exception, only minor deviations in the unbound structures were observed. In one crystal form of unbound W79F, there is a coupled shift in the side-chains of Phe29 and Tyr43 toward the mutation site, although in a different crystal form these shifts are not observed. In the bound structures, the orientation of biotin in the binding pocket was not significantly altered in the mutant complex. Compared with the wild-type streptavidin-biotin complex, there were no additional crystallographic water molecules observed for any of the mutants in the binding pocket. These structural studies thus suggest that the thermodynamic alterations can be attributed to the local alterations in binding residue composition, rather than a rearrangement of binding site architectures.

X-ray crystallographic studies of streptavidin mutants binding to biotin.

On the basis of high resolution crystallographic studies of streptavidin and its biotin complex, three principal binding motifs have been identified that contribute to the tight binding. A flexible binding loop can undergo a conformational change from an open to a closed form when biotin is bound. Additional studies described here of unbound wild-type streptavidin have provided structural views of the open conformation. Several tryptophan residues packing around the bound biotin constitute the second binding motif, one dominated by hydrophobic interactions. Mutation of these residues to alanine or phenylalanine have variable effects on the thermodynamics and kinetics of binding, but they generate only small changes in the molecular structure. Hydrogen bonding interactions also contribute significantly to the binding energetics of biotin, and the D128A mutation which breaks a hydrogen bond between the protein and a ureido NH group results in a significant structural alteration that could mimic an intermediate on the dissociation pathway. In this review, we summarize the structural aspects of biotin recognition that have been gained from crystallographic analyses of wild-type and site-directed streptavidin mutants.

Development of new biotin/streptavidin reagents for pretargeting.

The high affinity of biotin for streptavidin has made this pair of molecules very useful for in vivo applications. To optimize reagents for one potential in vivo application, antibody-based pretargeting of cancer, we have prepared a number of new biotin and streptavidin derivatives. The derivatives developed include new radiolabeled biotin reagents, new protein biotinylation reagents, and new biotin multimers for cross-linking and/or polymerization of streptavidin. We have also modified streptavidin by site-directed mutation and chemical modification to improve its in vivo characteristics, and have developed new reagents for cross-linking antibody fragments with streptavidin. A brief overview of these new reagents is provided.

Streptavidin-biotin binding energetics.

The high affinity energetics in the streptavidin-biotin system provide an excellent model system for studying how proteins balance enthalpic and entropic components to generate an impressive overall free energy for ligand binding. We review here concerted site-directed mutagenesis, biophysical, and computational studies of aromatic and hydrogen bonding interaction energetics between streptavidin and biotin. These results also have provided insight into how streptavidin builds a large activation barrier to dissociation by managing the enthalpic and entropic activation components. Finally, we review recent studies of the biotin dissociation pathway that address the fundamental question of how ligands exit protein binding pockets.

Molecular engineering of proteins and polymers for targeting and intracellular delivery of therapeutics.

There are many protein and DNA based therapeutics under development in the biotechnology and pharmaceutical industries. Key delivery challenges remain before many of these biomolecular therapeutics reach the clinic. Two important barriers are the effective targeting of drugs to specific tissues and cells and the subsequent intracellular delivery to appropriate cellular compartments. In this review, we summarize protein engineering work aimed at improving the stability and refolding efficiency of antibody fragments used in targeting, and at constructing new streptavidin variants which may offer improved performance in pre-targeting delivery strategies. In addition, we review recent work with pH-responsive polymers that mimic the membrane disruptive properties of viruses and toxins. These polymers could serve as alternatives to fusogenic peptides in gene therapy formulations and to enhance the intracellular delivery of protein therapeutics that function in the cytoplasm.

Energetic roles of hydrogen bonds at the ureido oxygen binding pocket in the streptavidin-biotin complex.

The high-affinity streptavidin-biotin complex is characterized by an extensive hydrogen-bonding network. A study of hydrogen-bonding energetics at the ureido oxygen of biotin has been conducted with site-directed mutations at Asn 23, Ser 27, and Tyr 43. A new competitive biotin binding assay was developed to provide direct equilibrium measurements of the alterations in Kd. S27A, Y43F, Y43A, N23A, and N23E mutants display DeltaDeltaG degrees at 37 degrees C relative to wild-type streptavidin of 2.9, 1.2, 2.6, 3.5, and 2.6 kcal/mol, respectively. The equilibrium-binding enthalpies for all of the mutants were measured by isothermal titration calorimetry, and the Y43A and N23A mutants display large decreases in the equilibrium binding enthalpy at 25 degrees C of 8.9 and 6.9 kcal/mol, respectively. The S27A and N23E mutants displayed small decreases in binding enthalpy of 1.6 and 0.9 kcal/mol relative to wild-type, while the Y43F mutant displayed a -2.6 kcal/mol increase in the binding enthalpy at 25 degrees C. At 37 degrees C, the Y43A and N23A mutants display decreases of 7.8 and 7.9 kcal/mol, respectively, while the S27A, N23E, and Y43F mutants displayed decreases of 4.9, 3.7, and 1.2 kcal/mol relative to wild-type. Kinetic analyses were also conducted to probe the contributions of the hydrogen bonds to the activation barrier. Wild-type streptavidin at 37 degrees C displays a koff of (4.1 +/- 0.3) x 10(-5) s-1, and the conservative Y43F, S27A, and N23A mutants displayed increases in koff to (20 +/- 1) x 10(-5) s-1, (660 +/- 40) x 10(-5) s-1, and (1030 +/- 220) x 10(-)5 s-1, respectively. The Y43A and N23E mutants displayed 93-fold and 188-fold increases in koff, respectively. Activation energies and enthalpies for each of the mutants were determined by transition-state analysis of the dissociation rate temperature dependence. All of the mutants except Y43F display large reductions in the activation enthalpy. The Y43F mutant has a more positive activation enthalpy, and thus a more favorable activation entropy that underlies the overall reduction in the activation barrier. For the most conservative mutant at each ureido oxygen hydrogen-bonding position, bound-state alterations account for most of the energetic changes in a single transition-state model, suggesting that the ureido oxygen hydrogen-bonding interactions are broken in the dissociation transition state.

Atomic resolution structure of biotin-free Tyr43Phe streptavidin: what is in the binding site?

The streptavidin-biotin system is an example of a high-affinity protein-ligand pair (Ka approximately 10(13) mol-1). The thermodynamic and structural properties have been extensively studied as a model system for protein-ligand interactions. Here, the X-ray crystal structure of a streptavidin mutant of a residue hydrogen bonding to biotin [Tyr43Phe (Y43F)] is reported at atomic resolution (1.14 A). The biotin-free structure was refined with anisotropic displacement parameters (SHELXL97 program package). The high-resolution data also allowed interpretation of side-chain and residue disorder in 41 residues where alternate conformations were refined. The Y43F mutation is unambiguously observed in difference maps, although only a single O atom per monomer is altered. The atomic resolution enabled the identification of 2-methyl-2, 4-pentanediol (MPD) molecules in the biotin-binding pocket for the first time. Electron density for MPD was observed in all four subunit binding sites of the tetrameric protein. This was not possible with data at lower resolution (1.8-2.3 A) for wild-type streptavidin or mutants in the same crystal form using MPD in the crystallization. The impact of MPD binding on these studies is discussed.

Protein electrostatic surface distribution can determine whether calcium oxalate crystal growth is promoted or inhibited.

Acidic proteins found in mineralized tissues act as nature's crystal engineers, where they play a key role in promoting or inhibiting the growth of minerals such as hydroxyapatite and calcium oxalate. Despite their importance in such fundamental physiological processes as bone and tooth formation, however, there is remarkably little known of the protein structure-function relationships that govern crystal recognition. We have taken a model system approach to elucidate some of the relationships between protein surface chemistry and secondary crystal growth of biological minerals. We show here that the distribution of electrostatic surface charge on our model protein, Protein G, determined whether the secondary growth of calcium oxalate, the principal mineral phase of kidney stones, was promoted or inhibited when the proteins were preadsorbed at low and equivalent surface coverages of <10%. The native Protein G, which contains 10 surface carboxylates, increased the rate of calcium oxalate growth from aqueous solution under constant composition conditions up to 97%, whereas a site-directed mutant with six of the surface charges removed inhibited the growth rate by 60%. The adsorption isotherms of both proteins were determined and suggested that the differences in electrostatic surface properties also lead to differences in protein orientation on the crystal surface. These results demonstrate that differences in electrostatic surface potential of proteins can directly determine whether secondary calcium oxalate growth is promoted or inhibited, and a model is proposed that suggests the distribution of carboxylate residues determines the interrelated binding orientation and exposed surface chemistry of the adsorbed Protein G.

Streptavidin in antibody pretargeting. Comparison of a recombinant streptavidin with two streptavidin mutant proteins and two commercially available streptavidin proteins.

In this investigation, a comparison of wild type recombinant streptavidin (r-SAv) with two genetically engineered mutant r-SAv proteins was undertaken. The investigation also included a comparison of the r-SAv with two streptavidin (SAv) proteins from commercial sources. In vitro characterization of the SAv proteins was conducted by HPLC, SDS-PAGE, IEF, and electrospray mass spectral analyses. All SAv proteins studied appeared to be a single species by size exclusion chromatography (HPLC) and SDS-PAGE analyses, but multiple species were noted in the IEF and MS analyses. In vivo comparisons of the SAv proteins were accomplished with dual isotope-labeled SAv in athymic mice. In an initial experiment, tissue localization of r-[131I]SAv directly radiolabeled using chloramine-T was compared with r-SAv radiolabeled with the N-hydroxysuccinimidyl p-iodobenzoate conjugate ([125I]-PIB), a radioiodination reagent that has been shown to result in iodine-labeled proteins which are stable to in vivo deiodination. The data obtained indicated that there is little difference in the distribution (except kidney localization) when r-SAv labeled by the two methods. Data obtained from comparison of r-[131I]SAv with a disulfide-stabilized r-SAv mutant (r-SAv-H127C), a C-terminal cysteine-containing r-SAv mutant (r-[125I]SAv-S139C), and two 125I-labeled SAv proteins obtained from commercial sources indicated that their distributions were quite similar, except the kidney concentrations were generally lower than that of r-[131I]SAv. On the basis of the similar distributions of the SAv proteins studied, it appears that the r-SAv mutants may be interchanged for the (wild type) r-SAv in pretargeting studies. Further, the similarity of distributions with two commercially available SAv proteins suggests that the results obtained in our studies and those of other groups may be directly compared (with consideration of animal model, sacrifice time, etc.).

Antibody fragments in tumor pretargeting. Evaluation of biotinylated Fab' colocalization with recombinant streptavidin and avidin.

An evaluation of the use of a biotinylated monoclonal antibody Fab' fragment in tumor pretargeting was conducted. As a model system, tumor colocalization of avidin or recombinant streptavidin (r-streptavidin) and the biotinylated Fab' fragment (Fab'-S-biotin) of A6H, an antirenal cell carcinoma antibody, was evaluated in athymic mice bearing human renal cell carcinoma xenografts. A new water soluble sulfhydryl reactive biotinylation reagent, N-(13-N-maleimdo-4, 7,10-trioxatridecanyl)-biotinamide, was synthesized and used for biotinylation of Fab'. A biodistribution of ChT-labeled A6H Fab'-S-biotin was conducted. Data from that distribution indicated that the Fab'-S-biotin localized well (i.e. 28% ID/g at 24 h) to human tumor xenografts in athymic mice. Subsequently, a biodistribution study involving pretargeting radioiodinated A6H Fab'-S-biotin to tumor xenografts, followed by administration of r-streptavidin at 4 or 20 h, was conducted. Specific colocalization of r-streptavidin to tumors containing the A6H Fab'-S-biotin was evident from the data obtained. In a similar biodistribution study, specific colocalization of avidin to tumors pretargeted with A6H Fab'-S-biotin was also observed. The avidin used in the study was radioiodinated with the N-hydroxysuccinimidyl ester of p-[125I]iodobenzoate ([125I]PIB-NHS). Very low concentrations (e.g. 0.35% ID/g) of avidin colocalized at the tumor. To further show that specific colocalization within the tumor xenografts had occurred with biotinylated A6H Fab', radioiodinated avidin and r-streptavidin were co-injected into athymic mice bearing tumor xenografts to obtain their distributions without having biotinylated Fab' present. At 20 h postinjection, only small differences in the blood and tumor concentrations of either protein were observed, indicating that the specific tumor colocalization seen in the previous two biodistributions must have been due to the presence of Fab'-S-biotin. Calculations were conducted to estimate how much r-streptavidin (as a molar ratio) was colocalized. From the data obtained it was estimated that 36-61% of the tumor-localized Fab'-S-biotin molecules were bound with r-streptavidin and 4-23% bound with avidin, under the conditions studied.

A structural snapshot of an intermediate on the streptavidin-biotin dissociation pathway.

It is currently unclear whether small molecules dissociate from a protein binding site along a defined pathway or through a collection of dissociation pathways. We report herein a joint crystallographic, computational, and biophysical study that suggests the Asp-128 --> Ala (D128A) streptavidin mutant closely mimics an intermediate on a well-defined dissociation pathway. Asp-128 is hydrogen bonded to a ureido nitrogen of biotin and also networks with the important aromatic binding contacts Trp-92 and Trp-108. The Asn-23 hydrogen bond to the ureido oxygen of biotin is lengthened to 3.8 A in the D128A structure, and a water molecule has moved into the pocket to replace the missing carboxylate interaction. These alterations are accompanied by the coupled movement of biotin, the flexible binding loop containing Ser-45, and the loop containing the Ser-27 hydrogen bonding contact. This structure closely parallels a key intermediate observed in a potential of mean force-simulated dissociation pathway of native streptavidin, where the Asn-23 hydrogen bond breaks first, accompanied by the replacement of the Asp-128 hydrogen bond by an entering water molecule. Furthermore, both biotin and the flexible loop move in a concerted conformational change that closely approximates the D128A structural changes. The activation and thermodynamic parameters for the D128A mutant were measured and are consistent with an intermediate that has traversed the early portion of the dissociation reaction coordinate through endothermic bond breaking and concomitant gain in configurational entropy. These composite results suggest that the D128A mutant provides a structural "snapshot" of an early intermediate on a relatively well-defined dissociation pathway for biotin.