Skip to Main Content

The Hilser Lab - Johns Hopkins University

Main Navigation::

117A Mudd Hall
Department of Biology
Johns Hopkins University
3400 N. Charles Street
Baltimore, MD 21218-2685

Office 410 516-6072
Lab 410 516-6757
Departmental fax 410 516-5213

Overall Research Goals

Although proteins are often depicted as static structures, it is well known that these molecules experience significant conformational fluctuations. As a result of these fluctuations, proteins in solution exist as ensembles of closely related, transient and interconverting conformational microstates, which on average, describe the crystal structure. Our lab's research examines the conformational ensembles of proteins to elucidate the complex interplay between local conformational fluctuations, global stability, and function in biology (for example, as in catalysis, allostery, and signal transduction).

We employ a diverse repertoire of experimental and computational approaches to address this complex problem. As such, the lab is engaged in projects that range from the development of atomistic, course-grain, and bioinformatics models of proteins to the experimental characterization of the effects of environmental perturbations and mutations on various biophysical and functional properties. Key to the lab’s approach is a close connection between the theoretical models and the experimental observations. Current projects in the lab fall into five main research areas.

Allostery in an Ensemble

The fields of structural biology and allostery developed in tandem with discoveries in each field informing the other. Indeed, the classic strategy for understanding allosteric mechanism focuses on examination of the structural changes between the allosteric protein with and without the allosteric effector bound. However, numerous examples exist that call into question this structural view. Namely, the observation of allostery without structural change and the ability of intrinsically disordered (ID) proteins to act as allosteric proteins calls into question the notion that allostery can be understood in the context of one (or even a few) structures.

To address this issue, we are interested in modeling allostery in terms of ensembles. We have two main research areas. First, we have been working over the past 15 years on developing and refining our structure-based ensemble model of proteins originally developed while a post-doc with Ernesto Freire [1, 2]. Using this approach, we have shown that allostery can be modeled as a shift in the population (or Boltzmann)-weighted distribution [3], that can be fine-tuned by modulating the structural stability of the protein [4] through modifications or binding at specific sites in the molecule [5]. Areas of interest include using our evolving COREX/BEST model to identify the most sensitive sites on proteins to which allosteric ligands can be designed.

In a second area of research we are interested in developing and using the Ensemble Allosteric Model (EAM) [6] to experimentally study allostery in ID proteins. Importantly, we have shown that ID can be used to optimize allosteric coupling in proteins and that the EAM can provide the organizing principles to understand many allosteric phenomena that are not readily reconciled in the context of the structural paradigm of allostery, such as allostery without a structural pathway [3], allosteric changes from surface exposed mutations [7], and switching between being a positive and negative effector for a given ligand [8]. We are currently using this model to investigate allostery in the steroid hormone receptor (SHR) family of transcription factors, using glucocorticoid receptor (GR) as a model system [9].

[1] Hilser and Freire, J. Mol. Biol. (1996).
[2] Hilser, et al., Proc. Nat. Acad. Sci. USA. (1998).
[3] Pan, et al., Proc. Nat. Acad. Sci. USA. (2000).
[4] Whitten, et al., Proc. Nat. Acad. Sci. USA. (2005).
[5] Liu, et al., Proc. Nat. Acad. Sci. USA. (2007).
[6] Hilser and Thompson, Proc. Nat. Acad. Sci. USA. (2007).
[7] Schrank, et al., Proc. Nat. Acad. Sci. USA. (2009).
[8] Motlagh and Hilser, Proc. Nat. Acad. Sci. USA. (2012).
[9] Li, et al., J. Biol. Chem. (2012).


Experimental Characterization of Conformational Fluctuations in Folded but Dynamic Systems

We use various biophysical techniques (including isothermal titration (ITC) and differential scanning calorimetry (DSC), fluorescence, circular dichroism, NMR chemical shift perturbation, 15N relaxation, relaxation dispersion, and hydrogen-deuterium exchange) to investigate the role of conformational fluctuations in determining stability and binding energetics in several model proteins in the lab. We currently have ongoing projects with SEM5 SH3 domain [1-6], Staphylococcal nuclease [7], and Escherichia coli adenylate kinase [8-10].

We have found in each of these systems that local folding/unfolding reactions (as opposed to conversion between two ostensibly folded structures) plays a major role in mediating the binding of ligands, the pH dependent shift in the protein ensemble, and allosteric signaling between distant sites in the molecule. Our strategy is to tie our experimental studies to the development and validation of theoretical models of function or of our evolving COREX/BEST structure-based model of the ensemble.

[1] Ferreon and Hilser, Prot. Sci. (2003A).
[2] Ferreon and Hilser, Prot. Sci. (2003B).
[3] Ferreon and Hilser, Biochemistry. (2004).
[4] Ferreon, et al., J. Am. Chem. Soc. (2004).
[5] Manson, et al., J. Am. Chem. Soc. (2009).
[6] Elam, et al. Biochemistry. (2013).
[7] Whitten, et al., Proc. Nat. Acad. Sci. USA. (2005).
[8] Schrank, et al., Proc. Nat. Acad. Sci. USA. (2009).
[9] Schrank, et al., Meth. Enzym. (2011).
[10] Schrank, et al., Top. Curr. Chem (2013).


Experimental Determination of Conformational Bias in Intrinsically Disordered and Denatured States

We are interested in the thermodynamic origins and consequences of conformational preferences in the denatured states of proteins, in particular the preference for the left-handed poly-proline II (PII) conformation. We harness the ability of SH3 domains to bind peptides that adopt the PII conformation in order to probe for residual PII content in unstructured peptides. By monitoring the effects of perturbants and mutations on the thermodynamics of binding, we can investigate the underlying thermodynamic bases for the observed preferences [1-6], and use this information to develop a more accurate model of the denatured states of proteins.

[1] Ferreon and Hilser, Prot.Sci. (2003).
[2] Ferreon and Hilser, J. Am. Chem. Soc. (2004).
[3] Hamburger, et al.,Biochemistry. (2004).
[4] Whitten, et al., Prot. Sci. (2008).
[5] Elam, et al., Prot. Sci. (2013A).
[6] Elam, et al., Biochemistry. (2013B)


Energetic Classification of Protein Fold Space

Using the ensemble-based model of proteins developed and validated in our lab, we are focused on characterizing proteins in energetic, rather than structural terms. We are interested in the underlying thermodynamic rules that relate sequence to fold, as well as thermodynamic homology between different folds. To date, we have shown that proteins can be represented in purely thermodynamic terms [1,2], that this representation is sufficient for fold recognition [3], and that energetic building blocks in proteins can be identified and used as the basis of a purely thermodynamic classification of protein fold space [4-6].

[1] Wrabl, et al., Prot.Sci. (2001).
[2] Wrabl, et al., Prot. Sci. (2002).
[3] Larson and Hilser, Prot. Sci. (2004).
[4] Vertrees, et al., Meth. Enzym. (2009A).
[5] Vertrees, et al., Biophys. J. (2009B).
[6] Wrabl and Hilser, PLoS: Comp. Biol. (2010).


Software Tools for Modeling and Comparing Protein Ensembles and Ensemble Properties

A. Development and Refinement of a Structure-based Model of the Protein Ensemble (COREX/BEST).
Model development, which is intimately tied to all of the experimental studies in our lab, is geared toward the design and refinement of our evolving structure-based course-grain model for conformational fluctuations originally introduced almost 20 years ago [1]. However, whereas classic model development approaches seek to quantitatively capture a particular phenomena of interest using complex models, our unique approach is targeted toward the development of a model that possesses the properties of being; 1) minimalistic, 2) experimentally verifiable, and 3) capable of unifying seemingly disparate observations within the context of a single formalism. In short, our approach is to identify the simplest possible model that can capture the phenomenon of interest, an approach that often requires unorthodox assumptions and the development of novel testing strategies. Using these approaches, we have been able to develop insight into, and unify the description of, a vast array of phenomena, ranging from site-site allosteric communication [2-4], to cooperativity in cold denaturation [5], to pH dependent fluctuations [6], and even hydrogen exchange protection factor patterns [7]. We maintain a web server [8] that allows researchers access to these approaches and we currently have over 1000 registered users from 30 countries on 5 continents. This site can be accessed here.

B. Energy Landscape Analysis of Protein Sequences (eScape).
Because the COREX/BEST algorithm is a structure-based method that uses the high-resolution PDB file as input, it cannot be applied to the analysis of proteins whose structures are unknown or proteins that are intrinsically disordered (ID). To overcome this limitation we have developed an approach that can compute the relative stability of a protein from its sequence alone. Using a machine-learning algorithm, the COREX/BEST profiles for a database of known protein structures was used as input to derive stability propensities that were mapped to individual tripeptides. The resultant eScape algorithm can compute the regional stability within each protein in a proteome [9], and can be used to compare evolutionary mechanisms for classes of proteins [10]. We maintain a web server that can be accessed here.

C. Horizontal Proteins Comparison and Alignment Tool (HePCAT).
Determining the evolutionary relatedness of two protein sequences is most successfully performed by amino acid sequence comparison. However, it is well known that structure can be preserved even when sequence has diverged past the point of amino acid similarity recognition, suggesting that sequences can bestow local, sub-global, and global properties to a protein that can be preserved in the absence of strict conservation of the side chain atoms. Thus approaches are needed that can augment sequence based analysis by matching patterns that may be independent of amino acid conservation at each position. To meet this challenge, we have developed a tool to compare the internal consistency of one-dimensional profiles defined by arbitrary sequences of numerical data [11]. The resultant algorithm, known as Horizontal Protein Comparison and Alignment Tool (HePCAT), can be used to compare energetic profiles (such as hydrogen exchange patterns, hydropathy, or any other type of property) that can be mapped over a sequence. The algorithm emphasizes the closeness in shape of the two data sets scanned over a horizontal range of positions, in contrast to the vertical position-by-position independent scoring of a standard amino acid substitution matrix. We maintain a web server that can be accessed here.

[1] Hilser and Freire, J. Mol. Biol. (1996).
[2] Pan, et al.,Proc. Nat. Acad. Sci. USA. (2000).
[3] Hilser, et al., Chem. Rev. (2006).
[4] Hilser, et al., Ann. Rev. Biophys. (2012).
[5] Babu, et al., Nat. Struct. Mol. Biol. (2004).
[6] Whitten, et al., Proc. Nat. Acad. Sci. USA. (2005).
[7] Liu, et al.,J. Am. Soc. Mass. Spectrom. (2012).
[8] Vertrees, et al., Bioinform. (2005).
[9] Gu and Hilser, Structure. (2008).
[10] Gu and Hilser, Mol. Biol. Evol. (2009).
[11] Hadzipasic et al., PLoS Computational Biol., In Press (2013).