Blaber Lab Recent Publications - (Updated 10/8/2021)

2021

The beta-trefoil protein architecture is characterized by three repeating �trefoil� motifs related by rotational symmetry and postulated to have evolved via gene duplication and fusion events. Despite this apparent structural symmetry, the primary and secondary structural elements typically exhibit pronounced asymmetric features. A survey of this family of proteins has revealed that among the most conserved symmetric structural elements is a ubiquitous buried solvent which participates in a bridging H-bond with three different beta-strands in each of the trefoil motifs. A computational analysis reported that these waters are likely associated with a substantial enthalpic contribution to overall stability. In this report, a Pro mutation is used to disrupt one of the water H-bond interactions to a main chain amide, and the effects upon stability and folding kinetics are determined. Combined with Ala mutations, the separate effects upon side chain truncation and H-bond deletion are analyzed in terms of stability and folding kinetics. The results show that these buried waters act to assemble a central folding nucleus, and are responsible for ~20% of the overall favorable enthalpy of folding.

Symmetric protein architectures have a compelling aesthetic that suggests a plausible evolutionary process (i.e., gene duplication/fusion) yielding complex architecture from a simpler structural motif. Furthermore, symmetry inspires a practical approach to computational protein design that substantially reduces the combinatorial explosion problem, and may provide practical solutions for structure optimization. Despite such broad relevance, the role of structural symmetry in the key area of hydrophobic core?packing cooperativity has not been adequately studied. In the present report, the threefold rotational symmetry intrinsic to the beta?trefoil architecture is shown to form a geometric basis for highly?cooperative core?packing interactions that both stabilize the local repeating motif and promote oligomerization/long?range contacts in the folding process. Symmetry in the beta?trefoil structure also permits tolerance towards mutational drift that involves a structural quasi?equivalence at several key core positions.

2020

Available high-resolution crystal structures for the family of ?-trefoil proteins in the structural databank were queried for buried waters. Such waters were classified as either: 1) unique to a particular domain, family, or superfamily, or 2) conserved among all beta-trefoil folds. Three buried waters conserved among all ?-trefoil folds were identified. These waters are related by the threefold rotational pseudo-symmetry characteristic of this protein architecture (representing three instances of an identical structural environment within each repeating trefoil-fold motif). The structural properties of this buried water are remarkable and include: residing in a cavity space no larger than a single water molecule, exhibiting a positional uncertainty (i.e., normalized B-factor) substantially lower than the average C? atom, providing essentially ideal H-bonding geometry with three solvent-inaccessible main chain groups, simultaneously serving as a bridging H-bond for three different ?-strands at a point of secondary structure divergence, and orienting conserved hydrophobic sidechains to form a nascent core-packing group. Other published work supports an interpretation that these interactions are key to the formation of an efficient folding nucleus and folded thermostability. The fundamental threefold symmetric structural element of the ?-trefoil fold is therefore, surprisingly, a buried water molecule.

The objective of this study was to characterize the therapeutic dose-response characteristics for topical FGF-1 in the full-thickness dermal healing of aged female BALB/cByJ mice. The approach utilized a splinted excisional model of dermal healing, and a novel fine-sampled photographic methodology, to quantify key wound healing parameters for different doses of topical FGF-1. The histology of healed wounds, representative of each dose cohort, was also evaluated by section and staining. The results show that topical FGF-1 pharmacotherapy for accelerating dermal healing in aged BALB/cByJ female mice yields a narrow dose-response curve, with diminished therapeutic effect at high concentration (i.e. �bell-shaped� dose-response). The physiological response of FGF-1 in wound healing involves a combination of cell types (including vascular endothelial cells, epidermal keratinocytes and dermal fibroblasts). These individual cells types in culture can have different FGF-1 dose-response curves; however, only the response of fibroblasts is bell-shaped. The bell-shaped dose-response in dermal healing therefore principally reflects the effect upon fibroblasts. A narrow bell-shaped dose-response requires precise dosing of FGF-1 for therapeutic benefit. The results identify the practical dose range to elicit such a benefit.

Gene duplication and fusion events in protein evolution are postulated to be responsible for the common protein folds exhibiting internal rotational symmetry. Such evolutionary processes can also potentially yield regions of repetitive primary structure. Repetitive primary structure offers the potential for alternative definitions of critical regions, such as the folding nucleus (FN). In principle, more than one instance of the FN potentially enables an alternative folding pathway in the face of a subsequent deleterious mutation. We describe the targeted mutation of the carboxyl-terminal region of the (internally located) FN of the de novo designed purely-symmetric ?-trefoil protein Symfoil-4P. This mutation involves wholesale replacement of a repeating trefoil-fold motif with a �blade� motif from a ?-propeller protein, and postulated to trap that region of the Symfoil-4P FN in a non-productive folding intermediate. The resulting protein (termed �Bladefoil�) is shown to be cooperatively folding, but as a trimeric oligomer. The results illustrate how symmetric protein architectures have potentially diverse folding alternatives available to them, including oligomerization, when preferred pathways are perturbed.

Many protein architectures exhibit evidence of internal rotational symmetry postulated to be the result of gene duplication/fusion events involving a primordial polypeptide motif. A common feature of such structures is a domain-swapped arrangement at the interface of the N- and C-termini motifs and postulated to provide cooperative interactions that promote folding and stability. De novo designed symmetric protein architectures have demonstrated an ability to accommodate circular permutation of the N- and C-termini in the overall architecture; however, the folding requirement of the primordial motif are poorly understood, and tolerance to circular permutation is essentially unknown. The ?-trefoil protein fold is a threefold symmetric architecture where the repeating ~42-mer �trefoil-fold� motif assembles via a domain-swapped arrangement. The trefoil-fold structure in isolation exposes considerable hydrophobic area that is otherwise buried in the intact ?-trefoil trimeric assembly. The trefoil-fold sequence is not predicted to adopt the trefoil-fold architecture in ab initio folding studies; rather, the predicted fold is closely related to a compact �blade� motif from the ?-propeller architecture. Expression of a trefoil-fold sequence and circular permutants shows that only the wild-type N-terminal motif definition yields an intact ?-trefoil trimeric assembly, while permutants yield monomers. The results elucidate the folding requirements of the primordial trefoil-fold motif, and also suggest that this motif may sample a compact conformation that limits hydrophobic residue exposure, contains key trefoil-fold structural features, but is more structurally homologous to a ?-propeller blade motif.

2018

The objective of this study was to determine quantitative parameters of dermal wound healing senescence in aged BALB/cByJ mice (an important animal model of aging) and evaluate the potential for therapeutic intervention by fibroblast growth factor-1 (FGF-1). The approach utilized a novel, non-invasive, fine-sampled photographic methodology to quantify wound healing parameters for healing phases from wounding through to wound closure. Parameters associated with key healing phases were quantified and compared for non-aged and aged cohorts of both sexes. The results identify a sexual dimorphism in dermal wound healing, with non-aged females exhibiting a greater overall healing efficiency compared to males. This enhanced healing in females, however, senesces with age such that healing parameters for aged males and females are statistically indistinguishable. Topical application of FGF-1 was identified as an effective therapeutic intervention to treat dermal healing senescence in aged females. Quantitation of the FGF-1 intervention was performed using a new, recently published model. The approach significantly increases the amount of pre-clinical animal data obtainable in wound healing studies, minimizes cohort number compared to (lethal) histological studies, and permits a direct statistical comparison between different healing studies. Quantitative parameters of dermal wound healing, obtained from non-invasive fine-sampled photographic data, identify topical FGF-1as an effective therapeutic to treat the senescence of dermal healing present in aged femaleBALB/cByJ mice.

We present a mathematical model to quantify parameters of mouse excisional wound healing from photographic data. The equation is a piecewise linear function in log scale that includes key parameters of initial wound radius (R0), an initial wound stasis phase (Ti), and time to wound closure (Tc); subsequently, these terms permit calculation of a latter active proliferative phase (Tp), and the healing rate (HR) during this active phase. A daily photographic record of wound healing (utilizing 6 mm diameter splinted excisional wounds) permits the necessary sampling for robust parameter refinement. When implemented with an automated nonlinear fitting routine, the healing parameters are determined in an operator-independent (i.e. unbiased) manner. The model was evaluated using photographic data from a splinted excisional surgical procedure involving several different mouse cohorts. Model fitting demonstrates excellent coefficients of determination (R2) in each case. The model thus permits quantitation of key parameters of excisional wound healing, from initial wounding through to wound closure, from photographic data.

In this study, we examined the local dynamics of acidic fibroblast growth factor (FGF-1) as well as the binding sites of various polyanions including poly-sulfates (heparin and low MW heparin) and poly-phosphates (phytic acid and ATP) using hydrogen-deuterium exchange mass spectrometry (HX-MS). For local dynamics, results are analyzed at the peptide level as well as in terms of buried amides employing crystallographic B-factors and compared with a residue level heat map generated from HX-MS results. Results show that strand 4 and 5 and the turn between them to be the most flexible regions as was previously seen by NMR. On the other hand, the C-terminal strands 8, 9 and 10 appear to be more rigid which is also consistent with crystallographic B-factors as well as local dynamics studies conducted by NMR. Crystal structures of FGF-1 in complex with heparin have shown that heparin binds to N-terminal Asn18 and to C-terminal Lys105, Tryp107, Lys112, Lys113, Arg119, Pro121, Arg122, Gln127 and Lys128 indicating electrostatic forces as dominant interactions. Heparin binding as determined by HX-MS is consistent with crystallography data. Previous studies have also shown that other polyanions including low MW heparin, phytic acid and ATP dramatically increase the thermal stability of FGF-1. Using HX-MS, we find other poly anions tested bind in a similar manner to heparin, primarily targeting the turns in the lysine rich C-terminal region of FGF-1 along with two distinct N-terminal regions that contains lysines and arginines/ histidines. This confirms the interactions between FGF-1 and polyanions are primary directed by electrostatics.

2017

We compare the thermal unfolding of fibroblast growth factor-1 (FGF-1), a naturally-evolved beta-trefoil protein which thermally aggregates, and Phifoil, a designed purely-symmetric protein which exhibits reversible thermal denaturation. These two proteins share a folding nucleus sequence. Detailed thermodynamic analysis of FGF-1 indicates that thermal aggregation initiates early in the unfolding pathway. Solid-state NMR of the FGF-1 aggregate identifies regions of structure consistent with the folding nucleus. The results show that design of an efficient folding nucleus, and the avoidance of aggregation in the folding pathway, are potentially separable design criteria � the latter of which should focus upon the physicochemical properties of primary structure outside the folding nucleus.

2016

Human FGF-1 has recently been shown to be a novel insulin-sensitizer. Position Serine 116 in FGF-1 is a site of phosphorylation. We report the properties of an Arginine mutation at this position (S116R) that is naturally present in bovine FGF-1 and cannot be phosphorylated. The S116R mutant is shown to have a reduced mitogenic activity, but an increased duration of insulin sensitization in ob/ob hyperglycemic mice. Thus, the mitogenic and insulin sensitizing functions of FGF-1 can potentially be differentially targeted in therapeutic design.

The folding nucleus (FN) is a cryptic element within protein primary structure that enables an efficient folding pathway - and is the postulated key heritable element in the evolution of protein architecture. However, almost nothing is known regarding how the FN structurally changes as complex protein architecture evolves from simpler peptide motifs. We report characterization of the FN of a designed purely symmetric ?-trefoil protein by ?-value analysis and compare it with the structure and folding properties of key foldable intermediates along the evolutionary trajectory of the ?-trefoil. The results show how novel turn structure created by gene fusion is subsequently incorporated into a larger and more efficient FN. Furthermore, the FN is adjusted by circular permutation in response to destabilizing functional mutation. This report provides the first insight into FN evolution. Selected for inclusion in the "Protein Evolution" special issue (June 2016).

Human FGF-1 has unique therapeutic potential in regenerative medicine and the treatment of metabolic disorder; however, biophysical properties intrinsic to FGF-1 are detrimental as regards drug development. FGF-1 contains three buried free thiols (cysteines) whose oxidation contributes to a short functional half-life. Simple mutational substitution without significant destablization is not possible. In this report a novel protein design strategy is employed whereby an additional cysteine is introduced that makes a stabilizing disulfide with one of the buried cysteines in FGF-1. The increase in stability of this novel disulfide almost exactly offsets the instability associated with mutational substitution of the remaining two cysteines - yielding a cysteine free form of FGF-1 with no overall stability perturbation. Cysteine free mutant forms of FGF-1 are shown to have substantially increased activity in functional assays of cell survival and mitogenicity.

2015

This report describes a detailed animal study, using a genetic mouse model of type II diabetes, for excisional wound repair and its acceleration by topical delivery of FGF-1. There are several important results in this report. The NONcNZO10/LtJ mouse is a newly-developed polygenic model of type II diabetes; however, no detailed description of its dermal wound healing has been reported. We show a delay in wound healing due to the diabetic condition; thus, this mouse is an excellent model with which to study delayed healing of diabetic ulcers. We also show that the rate of healing of splinted excisional wounds can be acclerated by topical application of FGF-1 formulated with heparin. We go on to show that engineered forms of FGF-1 (eFGF-1) that are stabilized in the absence of heparin, can equivalently accelerate dermal healing. Finally, to perform these experiments we had to develop a novel mouse jacket, and the details of this design are included in the report.

Protein therapeutics hold tremendous promise in treating human disease; however, proteins have unique problems as regards their choice as a drug. Principle among these is their ability to unfold and aggregate. A principle contributor to an irreversible unfolding pathway in proteins is the chemical modification of buried free thiols (i.e., cysteine residues). Mutation to eliminate such amino acids is an effective approach to removing one route of irreversible protein unfolding; however, exactly how to design such mutations is not clear. In this report we perform a comprehensive analysis of buried free cysteine mutations in FGF-1. We provide evidence for a key role in maintaining buried H-bonds that the wild-type cysteine engages in. There are two main structural effects in response to cysteine substitution - structural collapse and expansion. In the former case, serine substitution can maintain critical buried H-bonds despite its shorter H-bond distance compared to cysteine; in the latter case, alanine substitution can work best as it can permit room for inclusion of novel buried solvent that can similarly maintain critical buried H-bonds. The results show that simples rules for substitution of buried cysteines (e.g. substitution by the isosteric serine) are insufficient to deal with the structural complexity of the environment surrounding buried cysteines in proteins.

This report builds upon prior work showing that design of a beta-trefoil protein using an amino acid alphabet enriched for the "pre-biotic" set yields a halophile protein. This suggested a halophile origin of foldable proteins, followed by adaptation into mesophile conditions. In this report we show that such adaptation occurs upon a single amino acid subsitition to the non-pre-biotic amino acid Phenylalanine. Thus, upon emergence of a biosynthetic pathway for aromatic amino acids, a single substititon mutation in the core can provide the stability increase necessary for halophile-to-mesophile adaptation.

2014

One of the most important papers to come out of the lab. This research report describes successful symmetric protein design by first identifying the folding nucleus by phi-value analysis, followed by expansion of this sequence by the intrinsic structural symmetry to regenerate an intact protein. Experimentally we show that this design strategy yields a stable, foldable polypeptide for a purely-symmetric beta-trefoil architecture. This result provides an explanation as to why Top-Down Symmetric Deconstruction is successful, and also achieves a marked improvement in the design cycle efficiency. The results lead us to postulate that purely-symmetric proteins have redundant folding nuclei, and this provides a potential basis for selective advantage of symmetric protein folds in evolutionary processess involving gene duplication and fusion.

In this invited book chapter we describe the Top-Down Symmetric Deconstruction method of symmetric protein design. An image of a symmetric solution to the beta-trefoil fold was selected as the cover art for the book.

In this Opinion Article we argue in support of the halophile environment as the cradle of proteogenesis. The article covers: 1. Compelling data for a consensus prebiotic set of Ala, Asp, Glu, Gly, Ile, Leu, Pro, Ser, Thr and Val amino acids; 2. Favorable propensities for formation of helix, strand and reverse-turn secondary structure within this set; 3. Foldable potential of polypeptides with a prebiotic amino acid composition within the halophile environment; and 4. support for proteogenesis as a very early aspect of abiogenesis.

2013

Protein 3� structure symmetry is a defining feature of nearly one-third of protein folds and is generally thought to result from a combination of gene duplication, fusion, and truncation events. Such events represent major replication errors, involving substantial alteration of protein 3� structure and causing regions of exact repeating 1� structure, both of which are generally considered deleterious to protein folding. Thus, the prevalence of symmetric protein folds is counterintuitive and suggests a specific, yet unexplained, robustness. Using a designed ?-trefoil protein, we show that purely symmetric 1� structure enables utilization of alternative definitions of the critical folding nucleus in response to gross structural rearrangement. Thus, major replication errors producing 1� structure symmetry can conserve foldability. The results of this study provide an explanation for the prevalence of symmetric protein folds and highlight a critical role for 1� structure symmetry in protein evolution.

This report is an experimental test of our earlier hypothesis that the set of pre-biotic amino acids describes a foldable set, compatible with the halophile environment. In this study we enriched a de novo designed synthetic beta-trefoil protein (Symfoil)for the pre-biotic amino acid alphabet. This enrichment approached 80% pre-biotic composition. The resulting protein ("Primitive version #2")was fully foldable but only within a high salt (i.e., halophile) environment. Thus, the results suggest that proteogenesis was a potentially early abiogenic event, and likely occurred within a high-salt environment (e.g. evaporative lake). Both of these conclusions are counter to existing paradigms of hydrothermal vent origin, and RNA-first. This paper won the 2013 Kasha Award for best publication in molecular biophysics.

In this study we asked whether functional intersection is possible between the kallikrein-related peptidases (i.e., KLKs) and members of the matrix metalloproteinase (MMP) family by evaluating the ability of the MMPs to activate pro-KLKs. The results identify MMP-20 as a broad activator of pro-KLKs, suggesting the potential for intersection of the KLK and MMP axes under pathological dysregulation of MMP-20 expression.

2012

This review puts forth the argument that successful de novo protein design will bring about an economic windfall of a magnitude similar to that realized by the development of synthetic organic chemistry. Conversely, if such an economic benefit has not occurred, then the essential problem of de novo design has not yet been effectively solved. The review highlights developments in top-down protein design as a novel development in experimental de novo protein design.

This report describes a pharmacokinetic study (in rabbits) of wild-type and mutant forms of human fibroblast growth factor-1. Through the use of a mutant form of FGF-1 that lacks heparin-binding functionality, we show that heparan sulfat proteoglycan sequestration is the basis of the peripheral compartment in the PK profile. FGF-1 is typically formulated in the presence of heparin; however, this study shows that a consequence of this is a high initial systemic concentration, and low overall mean residence time. Conversely, stabilized mutant forms of FGF-1 that do not require heparin in the formulation are sequested efficiently after administration and do not exhibit high circulating levels. Furthermore, stabilizing mutants exhibit reduced redistribution kinetics (i.e., release from heparan sulfate proteoglycan); thus, a longer mean residence time. The results show how engineered forms of FGF-1 (for application in wound healing) can achieve benefits of efficacy, safety as well as cost.

A "stability/function tradeoff" hypothesis in proteins has been supported by a number of studies over more than a decade; however, a "foldability/function tradeoff" hypothesis is a recent development with little direct experimental support. In this report we describe a phi-value analysis of the FGF-1 protein that identifies turn regions essential to formation of the critical folding transition state. Although the overall fold of FGF-1 is symmetric (the beta-trefoil fold) the distribution of regions forming the folding transition state is shown to be highly-asymmetric. Furthermore, these regions are segregated from regions known to be responsible for various functionality of FGF-1 (e.g., heparin-binding, receptor-binding, and nuclear localization). These results provide essentially the first experimental support for the "foldability/function tradeoff" hypothesis. Importantly for protein design, the results identify that only a fraction of the structure is essential for formation of the important folding transition state, and that functional residues are a form of deviation from ideal symmetry within the overall fold. This report received the 2012 "Best Paper Award" from the Protein Society.

Symmetric elements within a protein fold argue for gene duplication and fusion in their evolution; however, purely-symmetric amino acid sequences are also reported by some investigators to contribute to folding frustration. This conundrum calls into question such evolutionary models as well as the utility of symmetry in de novo protein design (despite its potential to simplify the design problem substantially). This review covers this subject and argues that available data supports an evolutionary process involving a "conserved architecture" model - whereby, complex protein architecture was an early evolutionary accomplishment via oligomerization of simple peptide motifs. Subsequent duplication and fusion events produced complex proteins folds within a larger, single polypeptide chain. Some potential selective advantages of gene duplication and fusion events (i.e., resulting from what would otherwise be described as a major replication error) is discussed.

This review discusses the subject of proteogenesis (emergence of proteins from simple amino acids) within the larger context of abiogenesis (formation of living systems from simple non-living molecules). It is argued that a consensus is emerging that the alpha-amino acids Ala, Asp, Glu, Gly, Ile, Leu, Pro, Ser, Thr, and Val likely comprised the available pre-biotic amino acids, and thus, the building blocks for the earliest polypeptides. A key question is whether this set comprises a "foldable" set (i.e., can it provide a solution to the thermodynamic and kinetic requirements of protein folding)? If so, then complex protein architecture could have emerged early in abiogenesis (conversely, the emergence of complex protein architecture would have depended upon the prior emergence of (RNA-based?) biosynthetic pathways for the synthesis and accumulation of novel amino acids not present in this pre-biotic set). An analysis is presented that argues the pre-biotic set contains all fundamental properties required for foldable proteins; however, such proteins would likely be restricted to the halophile environment.

This review covers recent developments in 'top-down' approaches to protein design. Common among such efforts is the exploitation of symmetry in protein design, and the power of such symmetry to simplify de novo protein design.

This report describes a comprehensive biophyscial characterization, utilizing "empirical phase diagrams" of a series of "second-generation" mutants of fibroblast growth factor-1 (FGF-1). FGF-1 for therapeutic use in pro-angiogenic or acclerated wound healing therapies utilizes heparin as an additive to combat issues of poor thermal stability. Several mutant forms were designed to increase the intrinsic stability of FGF-1, and this study compares the muational effects to the wild-type protein formulated with heparin. The results identify several mutant forms with stability features that may obviate the need for heparin in their formulation. Thus, stability-based protein engineering may achieve results formerly achieved by formulation additives.

2011

This report describes in detail the development of the novel "top-down symmetric deconstruction" method of protein design and its application in the successful creation of a 42-residue polypeptide able to symmetrically self-assemble into a b-trefoil protein architecture.

This report is the culmination of almost 18 years of effort to develop a strategy to successfully design purely-symmetric protein architecture. The method is termed "top-down symmetric deconstruction". Existing paradigm posits that pure symmetry of amino acid sequence is not compatible with foldability; thus, it is not tenable as a de novo design principle, and logically calls into question the feasibility of gene duplication and fusion in the evolution of symmetric protein architecture. Our report disproves this paradigm, as we successfully demonstrate excellent foldability, thermostability and solubility for a purely-3-fold symmetric form of the b-trefoil protein fold. The results support one of two competing models for the evolution of this symmetric architecture (i.e. the "conserved architecture model") and also opens the door to a new approach in protein design - one that takes maximum advantage of symmetry in simplifying the design process.

2010

This report describes how a novel disulfide bond mutant of FGF-1 increases the in vitro functional half-life of the protein by 14-fold. This mutant design approach may be applicable to "second-generation" forms of different members of the fibroblast growth factor family.

This is a review article describing the known interactions between the Kallikrein-related proteases and those proteases of the thrombogenesis/thrombolysis pathways.

2009

Gold particle conjugates with proteins is a novel area of biotechnology with many unanswered questions. In particular, are such conjugates deleterious to protein conformation? This report shows that gold nanoparticle conjugates of FGF-1 are compatible with the native folded structure.

The rabbit hind limb model is the de facto animal model for studies of ischemia; furthermore, FGF-1 is emerging as the leading growth factor for pro-angiogenic therapy. Despite its relevance to these types of investigations, the cDNA sequence, X-ray structure, biophysical properties (including thermostability and receptor-binding affinity) and mitogenic activity of rabbit FGF-1 have never been reported; this publication presents such data.

Neutron diffraction studies of proteins requires large crystals; however, this is often difficult to achieve. FGF-1 crystals are large in two dimensions, but the third dimension is typically thin. This report describes mutations to the crystal contact in the thin phyisical dimension, whereby solvent-mediated contacts are replaced by direct protein-protein contacts. This approach improves the crystal growth along the thin dimension.

The FGF family of proteins contains an absolutely conserved Cys residue (at position 83 in the 140 amino acid form of FGF-1) that is present as a buried free-cysteine (in the majority of FGF proteins), or as a half-cystine involving adjacent position 66. We present evidence that mutation of position 66 to Cys can produce a stabilizing disulfide bond in FGF-1 (where Cys83 is a free cysteine), suggesting that the conserved free cysteine residue in the FGF family of proteins is actually a vestigial half-cystine.

This study identifies a cooperative interaction between buried free cysteines and the thermostability of a protein that effectively regulates the functional half-life. A protein design principle is outlined whereby the functional half-life of a mutant protein may be modulated by several orders of magnitude; furthermore, the design principle suggests a means by which a desired half-life can be achieved with limited immunogenic potential.

We complete our study of the KLK "activome" (i.e. the ability of mature KLK proteases to cleave the KLK pro-peptide and thereby activate the pro-KLKs) by reporting data for profiles of KLK9, 10 and 15. This report contains a now-complete matrix of the 215 possible pairwise activation combinations for the 15-member KLK family.

2008

The results of this study show that proteases of the thrombostasis family can efficiently activate specific pro-KLKs, demonstrating the potential for important regulatory interactions between these two major protease families.

Type 1 b-turns contain a consensus sequence of Asx-Pro-Asx-Gly, however, the details of the contribution of the residues in this motif to the turn structure, folding and stability are poorly understood. This study, involving 28 mutants of FGF-1, undertakes a comprehensive study of Asp, Asn and Ala mutations at each of the two canonical Asx positions. The results demonstrate a type of "Logical OR" interaction between the Asx residues that serve to stabilize the turn structure, and can even protect it from deleterious mutations.

2007

The ability of the different KLK proteases to cleave the 15 different pro-KLK peptide sequences was characterized using a fusion protein method. The results demonstrate the activation potential of the KLK's, and permit construction of hypothetical KLK activation cascades.

The first and last beta-strands in FGF-1 form a beta-sheet and are known to be a key step in the folding pathway. Two positions in these regions, Lys12 and Pro134, are shown to destabilize the structure. Mutation to Val provides a substantial 6 kJ/mol of additional stability at each position. The results suggest that targeting the termini beta-strand interactions in a beta-barrel might be a generally useful approach to increase protein stability. Furthermore, although FGF-1 is characterized as a marginally-stable protein, the results suggest that a handful of mutations might shift the stability into the "thermophile" regime.

This study quantifies the ability of KLK6 to hydrolyze its pro-peptide activation sequence and internal autolytic inactivating sequence. The results show that KLK6 activity is two-orders of magnitude greater for autolytic inactivation as compared to pro-sequence activation. Furthermore, several other proteases are compared to KLK6 and shown to exhibit much greater efficiency in activating pro-KLK6. Overall, the results show that KLK6 is unlikely to effectively self-activate, and requires a distinct protease for this purpose (thus forming the basis for a proteolytic activation cascade).

2006

This paper describes a new nomenclature for the human kallikrein-related peptidases, supported by a consortium of investigators led by Dr. Ake Lundwall.

FGF-1 contains fundamentally two types of beta-turns: type I 3:5 and type I 4:6. The role of the type I 4:6 turn was probed by mutations designed to convert the turn into type I 3:5. The results showed that a type I 3:5 turn can be implemented at each turn location in FGF-1 and yield a stable folded polypeptide. However, the mitogenic activity of such a mutant is substantially decreased, due to the functional requirement of a specfic type I 4:6 turn for receptor interaction. Thus, functional rather than folding considerations have resulted in the presence of a type I 4:6 turn in FGF-1.

2005

This is a contributed book chapter, detailing the solvent structure in the active site of human kallikrein 1, as presented at the International Workshop on "Hydrogen and Hydration in Proteins and Nucleic Acids", January 17-18, 2005, The University of Tokyo.

Symmetry-related hydrophobic packing groups in FGF-1 exhibit an asymmetric tolerance to side chain substitution. The packing environment of the positions in question is formed by specific surface loops. These loop regions vary in both sequence and length, and these differences dictate the asymmetric tolerance of the core substitutions. Thus, mutations in a specific surface loop region have "locked in" a requirement for a Cys residue at position 83. This residue cannot be mutated to a symmetrically-conserved Ile residue without first mutating the surrounding surface loop. Thus, this report highlights the co-evolutionary aspect of core and surface regions.

Only the second example of a human kallikrein structure to be deposited in the structural databank, we describe the x-ray structure of human apo- kallikrein 1. We detail the structural features, including solvent structure, of the S2 to S2' binding pockets. In comparison to porcine K1 with bound peptide inhibitors, the apo- kallikrein 1 structure exhibits a variety of induced-fit conformational changes upon substrate binding. These induced-fit changes include the active site serine, and indicate that the binding energy is utilized to position the active site serine for efficient catalysis.

Two b-hairpin structures in human fibroblast growth factor 1 are the same length but have different structures. Since they also differ in sequence, we thought this might be the reason for their different structures. However, sequence swapping mutations between the two turns did not result in a corresponding structural change. Thus, the turn structure is likely determined by the interaction of the turn residues with the local environment.

In this collaboration with Dr. Donald Caspar the structural and energetic consequences of mutations designed to alter the hydrophobic characteristics of the central cavity within interleukin 1-b are described. These mutations were designed to probe potential interactions with positionally-disordered solvent within the cavity (previously identified using low-resolution x-ray crystal data).

2004

This paper is really important for the lab. It shows that mutations designed to increase both the tertiary and primary structure symmetry of human fibroblast growth factor 1 also result in a mutant form that is dramatically more stable than the wild type protein. Thus, we believe that the b-trefoil (symmetric) superfold can be designed with a symmetric constraint upon the primary structure. There are several important ramifications of this work, related to protein evolution and de novo design. We hypothesize that a fundamental property of the symmetric protein superfolds is the ability to adopt a symmetric primary structure with an intrinsically high thermal stability.

This report describes a 1.10� resolution x-ray structure of human fibroblast growth factor 1. The data was of such high-resolution and quality that we were able to refine anisotropic thermal factors (i.e. the three-dimensional positional displacement for each atom). Using this information we asked whether certain regions of the protein exhibit correlated anisotropic motion (i.e. rigid body motion). We identified a region involving b-strands 5-12 that act as a rigid body distinct from the rest of the molecule. This demarcation also delineates heparin-binding from receptor-binding functionalities within the protein. It was selected as the cover article for the November 2004 issue of Proteins (the second cover article for the lab).

This paper is breakthrough work that shows that if the activity of kallikrein 6 (K6) is reduced (using a pre-immunization strategy) then animals that have been treated so as to develop a Multiple Sclerosis type of disease exhibit a delayed onset, and reduced severity. Thus, K6 represents a novel target in the treatment of inflammatory neurological diseases, like MS.

In this report we describe the x-ray structure of a mutant form of 2,5-diketo-D-gluconate reductase (the "vitamin C enzyme") which has an enhanced selectivity for NADH, versus NADPH, as a cofactor. The mutant was designed by our collaborator, Dr. Stephen Anderson at Rutgers University, and should prove useful in a new method for the industrial production of vitamin C.

This is a short invited review of a very interesting protein engineering study by Dr. Brian Matthews that investigates a possible mechanism for the evolution of allostery.

2003

This is a report that details our NMR studies of turn mutants in human fibroblast growth factor 1 that were performed at the National High Magnetic Field Laboratory. Some mutants do not readily crystallize, and Jihun Lee and Jaewon Kim, two graduate students in the lab, are spearheading efforts to utilize NMR to obtain structural information for such mutants.

This is another very important paper for the lab. It details the ability to redesign the core region of human fibroblast growth factor 1 using a symmetric design constraint. Despite the imposed design constraint, the repacked core is among the most successful published and let to our hypothesis (developed in our later publications) that perhaps symmetric superfolds can be successfully designed using a symmetric primary structure. This article was selected as the cover for the Dec 2003 issue of Protein Science - the first cover article for the lab.

This paper details one of the more successful discoveries of a design principle for turn structures. We show that a glycine amino acid at the i+3 position in a type I 3:5 and 4:6 b-turn plays a critical role in stabilizing such structures. This discovery allowed us to reconcile conflicting published reports regarding the thermodynamic consequences of glycine substitutions at positions that lie within the L-a region of the Ramachandran plot.

This is an invited review article that arose from my graduate student, Gulsah Sanli, giving a poster at a scientific conference. In addition to presenting a review of known information, Gulsah also included a structural analysis of the apo- and holo-forms of 2,5-diketo-D-gluconate reductase and identified a hinge-bending motion associated with cofactor binding. The hinge-bending motion delineates two halves of the symmetric structure of the enzyme, and suggests that it may have evolved from an ancient homo-dimer.

2002

This report describes the x-ray structure determination for human kallikrein 6 (K6) - the first human kallikrein structure deposited into the Protein Data Bank. Unknown to us at the time, another group also solved a structure for K6 (the inactive pro-form). Their report came out only two weeks after ours, and in the same journal! K6 is shown to be structurally related to the digestive enzyme trypsin, although it is expressed primarily in the brain and spinal cord.

In different crystal forms of human fibroblast growth factor a particular turn region exhibits distinctly different structures. Which one is correct, and which one may be a crystal packing artifact? This study describes a series of glycine mutations, in conjunction with stability studies, that identify one particular conformation as being likely populated in solution. The results also describe a statistical study of the structural data bank that suggests that a certain turn conformation may be present at higher than expected percentages due to crystal packing artifacts.

Our lab was the first to express recombinant kallikrein 6 (from the rat) and to determine its enzymatic properties. In this report, K6 is identified as a digestive-type protease, with broad substrate specificity. We also show that K6 has a built-in mechanism of autolytic inactivation, and this likely plays a key regulatory role.

2001

This is our first report on efforts to redesign the core region of fibroblast growth factor 1 using a symmetric design constraint. We introduce three mutations into the 15 amino acid core and show that the mutations are well-tolerated. This study paved the way for further symmetric redesign studies.

Two outstanding undergraduate students in the lab, Sasha Grek and John Davis, undertook the Herculean task of writing a Windows�-based software program to deconvoluted data from differential scanning calorimetry studies. The program was a success, and currently 15 copies are downloaded each month from the lab's website.

We solved the x-ray structures of 2,5-diketo-D-gluconate reductase (the "vitamin C enzyme") in both holo- (i.e. with NADPH cofactor) and apo- forms (i.e. without) and discovered that the active sites is in a non-catalytically competent conformation in the absence of bound cofactor. Thus, the NADPH cofactor not only provides a hydride for catalysis, but also serves to organize the active site. This was the first time that such a role for the cofactor in this family of enzymes had been reported.

Genes from Corynebacterium are GC rich and prove problematic for DNA polymerase-based manipulations (i.e. PCR and sequencing). We designed a couple of synthetic genes that retained the amino acid sequence, but substantially reduced the GC content (and also improving E. coli codon bias). These genes were shown to be amenable to polymerase-based manipulations, and could be readily mutated, sequenced and expressed in E. coli. This study laid the groundwork for further mutagenesis studies to improve the stability and catalytic features of this enzyme.

2000

Previous reports suggested that cysteine to serine mutations in human fibroblast growth factor 1 (FGF1) resulted in substantially longer half-life in cell culture mitogenic assays. This observation was postulated to be due to either an increase in stability or elimination of thiol-based modifications. We had developed the methodology for determination of the DG unfolding for FGF1 and decided to test the stability hypothesis. Our results showed that the serine mutations destabilized the protein. Thus, the increase in half-life was due to elimination of thiol-based modifications. We postulate that a mechanism to limit protein half-life is to introduce free cysteine residues within the core region of a protein. Such proteins cannot refold if the cysteine forms thiol adducts or are chemically modified by oxidation.

Using bound solvent as a guide, this report details the modeling of substrate into the active site of 2,5-DKG-reducates, the "vitamin C enzyme". Although this enzyme was originally identified from screening of soil bacteria capable of reducing 2,5-diketo-D-gluconate, this study shows that this compound is unlikely to be the natural substrate.

1999

This was an important publication for the lab. We detail the development of the methodology by which the reversible thermal denaturation of human fibroblast growth factor 1 can be achieved. The results allowed us to demonstrate the generally good agreement between isothermal equilibrium denaturation and differential scanning calorimetry, and showed that the 2-state models being invoked to analyze the data were valid.

This report was part of a collaboration with Dr. Don Caspar at Florida State University. Interleukin-1b has a large central cavity and there were conflicting reports in the literature on whether this cavity contained disordered solvent. By collecting low resolution data and scaling the electron density on an absolute scale, we were able to show that the central cavity contains approximately 1.8 water molecules worth of disordered electron density - suggesting that solvent is indeed rattling around within the core.

1998

This was the first x-ray structure for a prokaryotic member of the aldo-keto reductase family. The structure provided details of the structural "seat belt", a pair of loops that covered the cofactor binding site and contributed to the kinetics of cofactor binding.

1996

This is the first x-ray structure reported for human acidic fibroblast growth factor. A single crystal was used for both spacegroup determination (using precession photography) and data collection (using a multiwire detector). This was in the good old days when you needed to know the space group and cell dimensions prior to data collection. This is also the first crystal structure reported by our laboratory.