HomeResearch Collaboration Serving Society Personnel News Contact Us

Thrust 2

The combination of biology with materials science is a powerful concept, which is only beginning to be explored. Achieving a fundamental understanding, both experimental and computational, of the molecular events that govern biological function and selectivity in nonbiological nanoscale environments is crucial in developing nanostructured biomolecule composite architectures.

The mission of Thrust 2 is to enable the efficient and selective interaction of biomolecules with synthetic nanoscale building blocks to generate functional assemblies. We have positioned the thrust squarely at the interface of the biological, chemical, and materials sciences, which closely integrates our expertise in biomolecular engineering; nanomaterial preparation, characterization, functional assembly, and theory/simulation; and applications at the meso to macro scales. We have continued our focus on both the fundamental properties and potential applications of hierarchical biomolecule/nanomaterial hybrid composites with tailored structures and functions. We have also begun to integrate more fully with Thrust 3 by investigating biologically-functionalized nanoscale materials and their impact on human biology and toxicology.



To that end, we continue to be guided by the following fundamental questions:

  • What are the key molecular events that govern biomolecule-nanomaterial interactions?
  • How do the unique structural and surface properties of nanoscale materials influence biomolecular structure, function, and stability?
  • How can these molecular events be controlled to the extent necessary to promote optimal structure and orientation of biomolecules on nanoscale surfaces?
  • How can the properties of proteins, for example, be engineered to promote functionality on nanoscale materials?
  • How can proteins, or other biomolecules, be used to assemble nanostructures into controlled and functionally useful structures?
  • What broad-based computational strategies can be used to guide experimental research to enable precise control of biomolecule-nanoscale interactions?

We are using these biomolecular-material hybrid building blocks to generate functional nanoscale assemblies. In this regard, we are specifically focused on biomolecular and natural product template-assisted design of nanostructures, which extend across a wide range of length scales from nm to μm. Connections to Thrust 3 exist for nanoscale architectures that influence cell growth and physiology. Recent results include:


Protein unfolding on nanoscale surfaces

We have studied the unfolding behavior of model proteins on the surfaces of silica nanoparticles (SNPs) (Siegel, Dordick).  Previously we have shown with ribonuclease A and cytochrome c (cyt c) that protein stability decreases upon adsorption to nanoparticles and that the decrease is a function of increasing nanoparticle size.  For cyt c we collected additional evidence that provides a specific structural basis to the changes in stability.  By taking advantage of the protein’s covalently bound catalytic heme group, buried within the interior of the protein, we were able to probe nanoparticle size effects on protein stability and on the influence of the surface on local changes to the enzyme’s active site. The buried heme group exhibits circular dichroism with a positive cotton peak around 408 nm and a negative cotton trough near 418 nm.  Upon adsorption to the SNP the positive cotton peak increases, while the negative cotton trough is eliminated, resembling the Soret spectra of a typical peroxidase.  Consistent with this spectral result is the finding that cyt c adsorbed to increasingly larger SNPs shows increasing peroxidatic activity (see figure below "a"). These results suggest that the larger nanoparticles open up the internal core of the cyt c, perhaps by a slight degree of unfolding on the larger (and flatter) SNPs.


Influence of nanomaterial geometry and crystal face properties on protein structure and function


Understanding how the structure and function of adsorbed proteins are affected by the surface geometry of nanomaterials requires the precise control of the synthesis of concentrated, monodisperse particles with desirable properties.  We (Siegel and Dordick) have undertaken a study of the influence of rod-like geometry vs. spherical geometry on protein structure and function, which raises the question of how crystal faces influence protein behavior. Electrochemical and template methods were used to generate gold nanorods following the method by Murphy et al. This approach produces nanorods of adjustable aspect ratio (3-20) and with a distinct crystal structure.  Small seeds of metallic gold are generated with citrate stabilization and form clustered atoms with very little long-range order. By taking newly formed seeds and immersing them in “growth” solutions of gold salt and a surfactant, cetyltrimethylammonium bromide (CTAB), it is possible to control the direction of particle growth and form rods. CTAB has an affinity for the {110/001} families of Au crystal planes; by binding to these facets, the growth in those directions is stunted and the particle elongates along the <111> direction (see figure below “b”). Producing these particles with monodisperse size and aspect ratio will now enable an investigation of how surfaces with variable curvature and defined features affect protein adsorption and orientation.


Nanomaterials employed in protein-binding studies: (a) Influence of SNP size on cyt c peroxidatic activity; higher activity results from increased exposure of the buried heme due to protein unfolding on the flatter surface. (b) TEM images of CTAB-stabilized gold nanorods (bar represents 62 nm).


Investigation of biomolecule adsorption and stability on raft-mimetic lipid domains


We (Kane, Kumar and Dordick) have demonstrated that controlling the heterogeneity of the reconfigurable surface – a soft liposomal membrane - can enhance the stability of adsorbed proteins.  Adsorbing enzymes onto “patchy” surfaces composed of adsorbing and non-adsorbing regions (see following figure “a”) can be used to reduce lateral interactions between adjacent enzymes and enhance enzyme stability.  To demonstrate the ability to pattern the adsorption of proteins, we adsorbed fluorescein-labeled soybean peroxidase (SBP) onto giant unilamellar vesicles (GUVs) composed of DPTAP, DOPC, and the fluorescent dye 1,1’-dioctadecyl-3,3,3’,3’-tetramethylin54 docarbocyanine perchlorate (DiI).  Phase separation was induced by heating the liposomes then allowing them to cool to room temperature. Characterization by confocal microscopy confirmed the highly selective binding of SBP onto the gel phase DPTAP-enriched domains (figure “b”).  We next made heterogeneous liposomes composed of a mixture of DPTAP and DOPC, and following heating and cooling to induce phase separation, SBP was allowed to adsorb. SBP-liposome conjugate activity was then measured in methanol solutions.  Characterization by CD spectroscopy revealed a slower rate of change of secondary structure for SBP adsorbed onto the heterogeneous liposomes as compared to SBP adsorbed onto homogenous DPTAP liposomes (figure c), and a ca. 3-fold increase in enzyme stability over that obtained on homogeneous DPTAP liposomes (figure “d”).  These results provide evidence that protein stability can be controlled through the heterogeneity of the underlying soft material.

Schematic of SBP adsorbed onto (i) homogeneous gel phase liposome; (ii) positively charged domains in a heterogeneous liposome (b) Confocal micrograph of (i) DiI partitioned into gel phase domains of a GUV ii) Fluorescein-labeled SBP adsorbed on the  same GUV and iii) merged image showing the patterned adsorption of SBP. (c) Percent  secondary structure retained vs. time and (d) Percent activity retained vs. time for SBP adsorbed on homogeneous (dark circle), and heterogeneous (open circle) liposomes.


Active and stable covalent enzyme-nanotube conjugates


Despite considerable progress in the preparation of nanotube-protein conjugates for numerous applications ranging from sensing to delivery and the design of functional composites, few studies have yielded detailed information on the structure and function of proteins covalently attached to nanotubes.  In partial collaboration with Genencor International (Palo Alto, CA), we (Dordick and Kane) have, therefore, evaluated in detail the structure, activity, and stability of enzymes covalently attached to carbon nanotubes.  Along these lines, we examined the enzyme perhydrolase S54V (AcT), which effectively catalyzes the perhydrolysis of acetate esters such as propylene glycol diacetate (PGD) to generate peracetic acid (PAA).  PAA is a potent oxidant increasingly used for sanitization, disinfection, and sterilization due to its broad effectiveness against bacteria, yeasts, molds, fungi, and spores. AcT is a large molecule formed through tight association of pairs of dimers that may negatively impact the activity of the enzyme on supports, and hence on MWNT-based conjugates.  Specifically, there are four insertions: residues 17-27; residues 59-69; residues 122-130; and residues 142-156 in the AcT structure, which form loops at the dimer interfaces and contribute to stabilization of the octameric structure. These loops enable formation of a hydrophobic channel that extends to the exterior of the octameric surface. The regions forming the hydrophobic channel lead to the active sites of the AcT being somewhat buried and thus having restricted substrate accessibility. Bioinformatic calculations revealed that ~60% of the amino acid residues that constitute the monomer are hydrophobic and the average hydropathicity of the monomer is 0.117 indicating a high degree of hydrophobicity. The large block-like structure and extensive hydrophobicity of AcT would presumably lead to substantial nonspecific hydrophobic interactions between the AcT surface and the non-functionalized hydrophobic regions of the MWNTs.  These nonspecific interactions, together with covalent attachment, determine close packing of AcT molecules onto the MWNT surface; as a comparison the bare acid-treated MWNTs are also shown). Consequently, the attached AcT molecules would have limited flexibility and their strong interaction with the nanotube would also reduce the substrate accessibility to the active sites.

Structure of AcT. (a) AcT octamer with catalytic triad Ser11, Asp192, and His195 shown in filled space, and all other residues shown with lines; colored residues, green: hydrophobic, pale blue: hydrophilic, dark blue: basic, and red: acidic. (b) Direct attachment of AcT onto MWNT. In addition to covalent binding,nonspecific hydrophobic interaction also exists due to the large size and hydrophobic nature of AcT. Inset:TEM image of AcT-MWNT conjugates. (c) Attachment of AcT onto MWNT using dPEG as spacer.

Enzyme flexibility can be improved by inserting a spacer between enzyme molecule and the attaching surface. To this end, a bifunctional amino-dPEG12-acid  spacer was first covalently attached to the acid treated MWNTs and subsequently AcT was attached to the free end of the spacer both via EDC/NHS amide formation. Both the free AcT and the AcT-dPEG-MWNT conjugates followed Michaelis-Menten kinetics, indicating that these conjugates possessed high intrinsic activity on the nanoscale support. The good kinetic properties of the AcT-dPEG-MWNT conjugates led us to use this formulation for preparation of the polymer and paint composites.


Molecular Modeling of Biomolecules in Nanoscale Environments


A fundamental understanding of biological self-assembly – e.g., protein folding and aggregation, micelle and membrane formation and molecular recognition – is expected to significantly improve potential applications in areas of health and medicine, and materials and device development.  Key aspects of this part of Thrust 2 are the elucidation of the role of water on biomolecule structure, function, and dynamics, particularly in environments that are substantially nonaqueous, and the influence of confined environments on biomolecule folding.  We (Garde) have focused on gaining a fundamental understanding of how interfaces affect conformational dynamics and thermodynamics (i.e., folding-unfolding processes) in biological systems.  This is central to protein-nanomaterial interfaces, and especially in understanding how the nanomaterial influences protein structure, dynamics, and function.  Extensive molecular dynamics simulations have been performed on the folding-unfolding of a hydrophobic polymer (a first step in mimicking a biomolecule) in bulk water and at two interfaces - a model hydrophobic solid-water interface and a vapor-liquid interface of water.  In bulk water, the model hydrophobic polymer collapses into globular folded structures (figure below “a”). When interfaces are present, we found that the polymer is driven to the interface by water-mediated interactions and polymer structure and thermodynamics of folding is significantly different than in the bulk. Polymer conformations are quasi 2-D (flat pancake like) at both interfaces (figure “b-d”). The driving force or the potential of mean force (PMF) for folding is also weaker at the interface compared to that in the bulk. At the vaporliquid interface, there is no stable minimum in the PMF as a function of the radius of gyration of the polymer, and correspondingly, the polymer samples a wide spectrum of conformations from compact to fully extended at that interface. These observations are consistent with observed binding of proteins and consequent unfolding at hydrophobic as well as vapor-liquid interfaces of water.  Interfaces are also highly dynamic at hydrophobic S-L or V-L interfaces.  Diffusivity of water as well as timescales for collapse transition or conformational transitions are faster at S-L interface compared to that in bulk and much faster at the V-L interface.

Representative conformations of the hydrophobic polymer in (a) bulk, (b) at a hydrophobic solidwater interface, and (c) at a vapor-liquid interface of water. Top plates are top views, and bottom plates are side views. (e) Free energy of the polymer as a function of radius of gyration. The minimum in bulk water at Rg = 0.5 nm indicates folding into globular states. Driving force for folding is weaker at the S-L and almost non-existent at the V-L interface.


Dynamical coupling between water and protein via thermal analysis


Using molecular dynamics simulations we (Keblinski and Garde) studied thermal energy flow between a green fluorescent protein and surrounding water to unravel the nature of dynamical coupling between biomolecules and their aqueous environment (see following figure). A novel characterization method was introduced allowing frequency dependent dynamical coupling to be determined at interfaces between biomolecules and their environment. Low frequency vibrations in the protein, which are thought to be critical for the protein function, are strongly coupled with water, whereas intermediate and high frequency vibrations are essentially decoupled with water, except for those present at the surface of the protein. These studies shed new light on the fundamental physical mechanism underlying the dynamical slaving of proteins to water, which is a topic of intense discussion by the scientific community. In future work, the applicability of this method will be assessed on systems where protein is exposed to a non-aqueous environment and ultimately to hydrophobic nanoscale materials.

Color-coded representation of the temperature of high (left panel) and low (right panel) frequency modes in green fluorescent protein subjected to the thermal flux to warmer surrounding water. High frequency modes are cold (blue color) in the protein center and even at the protein surface are colder than water (red color). Low frequency modes are warmer in the protein center and assume water temperature at the protein surface.


Self-Organization of Water Induced by Hydrated Electrons


Fundamental understanding of self-assembly in water requires knowledge of water structure and dynamics in the vicinity of solutes, specifically in the hydration shells. Mounting evidence suggests that qualitative water-structure making/breaking ideas are unhelpful, and quantitative knowledge of how water structure (packing, orientation, H-bonding) responds to solute shape and chemistry is critically important. Such information to date has come primarily from molecular simulations of classical point charge models of water. However, such simulations are complex, computationally intensive, and often depend on the precise choice of water models. Experimentalists who encounter new phenomena in aqueous media usually resort to qualitative arguments about hydration and the evolving structure of water.

We (Wong and Garde) have been able to bridge this gap. The goal is to efficiently reconstruct hydration behavior at Ångstrom lengthscales and femtosecond timescales.  Linear Response Imaging (LRI) has been employed to reconstruct the dynamical behavior of water from a library of dynamical structure factor S(q,ω) data measured at 3rd generation synchrotron X-ray sources. The density-density response function of water χ(q,ω) is extracted from S(q,ω), and subsequently used as a Green’s function to reconstruct the space- and time-dependent behavior of water.  LRI results were compared to those of existing scattering and spectroscopic experiments, as well as molecular dynamics simulations that we performed at RPI (see figure below).  The National Center for Supercomputing Applications (NCSA) at UIUC and the parallel computational platform at the Computational Center for Nanotechnology Innovations (CCNI) at RPI were used to perform the high-speed calculations required of this study.  To illustrate the potential of LRI, we reconstructed movies of the evolution of hydration structure around a charge distribution representative for hydrated electrons, by tracking the average oxygen density correlations at ~50 femtosecond temporal resolution and ~0.8 Å spatial resolution.  Instead of a spherical hydration shell, our results indicate that linear movement of the charge distribution induces an asymmetric ‘melting’ of the hydration structure. Depending on the velocity, the hydration ‘shell’ progressively evolves from closed ‘spherical’ shape to a hydration ‘bowl’ with its trailing edge open, to a cylindrical hydration ‘sleeve’, before the hydration structure is completely destroyed. These changes in the hydration structure potentially impact a broad range of aqueous phase chemistry.

The hydration structure around an ‘electron-like’ delocalized distribution of negative charge moving at (a) 500 m/s and (b) 1000 m/s. The equilibrated hydration shell melts in a position-dependent manner. The time stamp indicates the time in femtoseconds after the initiation of motion. After t~500 fs, the hydration structure is essentially the same as the infinite-time steady state structure as shown.


Impact of Negative Gaussian Curvature on Cell Membrane Penetration


Two self-assembly based approaches are being pursued to generate nanoparticle with negative Gaussian curvature.  We (Braun and Wong) have begun to investigate the effect of nanoparticle curvature on cell membrane penetration. The first negative curvature nanoparticle synthesis approach utilizes the internal void space of a self-assembled colloidal crystal to template the shape of electrodeposited nanoparticles. A polystyrene (PS) opal is grown on a conductive substrate and lightly sintered to create necks that interconnect adjacent spheres. Then, a layer of nickel is grown by electrodeposition. The amount of nickel plated is controlled so that the growth front stops at the height of neck.  Gold is then electrodeposited above the nickel layer.  The gold layer conformally wraps around the necks between the PS spheres and forms the negative Gaussian curvature shown in “a” of the figure below.  The thickness of the gold layer is controlled so that it does not merge together above the necks. The PS colloids are dissolved in tetrahydrofuran, resulting in gold islands sitting on a nickel substrate (figure “b”). After surface functionalization with a thiolated amine, the gold islands are removed from the supporting nickel structure, forming the desired nanoparticles (figure “c”).  Particles generated this way have a negative radius of curvature of ~50 nm along several sides, and a positive radius of curvature of ~5 nm at the points.  Initial cell penetration experiments are performed by mixing the nanoparticles with human HeLa cancer cells, followed by incubation at 37°C for 1 h. Then the cells are observed via dark field optical microscopy to determine the degree of particle uptake (figure “d”). In the image, HeLa cells appear in faint blue color. The negative curvature gold particles appear bright yellow. The particles indicated by dashed red circles are inside the cells.

(a) Schematic of the bilayer structure generated through electrodeposition through a colloidal template. The negative Gaussian curvature particles are formed from the gold residing above the nickel layer. (b) SEM image after removal of the colloidal template. (c) SEM image of an individual particle exhibiting negative Gaussian curvature. (d) Dark field optical microscopy of HeLa cells after incubation with negative Gaussian curvature nanoparticles.

The second approach to generate nanoparticles with negative Gaussian curvature takes advantage of the difference between conformal growth and conformal etching on a 3-D curved surface, for example at the contact between two spherical colloids (figure below “a”).  The result is a thin disk remaining in the area where colloids are in contact.  The rim of the disk has a strong negative Gaussian curvature.  Conformal growth of a variety of materials can be achieved by atomic layer deposition (ALD) or chemical vapor deposition (CVD).  For example, Al2O3 and HfO2 can be coated on a PS opal template by ALD and tungsten can be grown via CVD using a silica colloidal crystal template.  Conformal etching of the oxides is achieved by slow chemical etching, and the metal is etched via pulsed electrochemical etching (figure “b”). The Al2O3 particles generated here have a relatively large positive curvature (~370 nm) and a relatively small negative curvature (~80 nm); the results for the other materials are similar, and importantly, these dimensions scale with the colloid diameter, and thus a wide range of curvatures is experimentally accessible.

(a) Schematic of the difference between conformal deposition and conformal etching, resulting in disk-like objects with negative curvature.(b) SEM micrographs of Al2O3 nanodisks before and after release from the colloidal template, and tungsten nanodisks before release from the colloidal template.


Tubulin-Nanotube Conjugates

The active transport, assembly, and/or spatial organization of nanocomposites in a controlled manner in synthetic environments are key challenges in enabling nanocomposites application in fiber-reinforced products or electronic devices. Examples include the assembly of carbon nanotubes (CNTs) by directly embedding them into microscale polymer blocks to aid gas mixture transport and the large scale self-assembly of CNT-based circuits through use of organic molecular markers. However, CNTs are difficult to solubilize and organize into architectures, which limits their ultimate use in electronic displays, and nanoscale actuators, as well as in biological applications, e.g., protein-nanotube conjugates as sensor elements.  We (Dordick) have shown that biological macromolecules (cellular motor protein kinesin and the synthetically reconstituted microtubule) can be used for active transport of inorganic material such as multiwalled carbon nanotubes (MWNTs) in synthetic, non-physiological environments.  The method involved reconstitution of the cellular kinesin-microtubule transport system in a flow chamber and in a gliding geometry wherein surface-immobilized kinesins served as “conveyor belts“ for hybrids formed from microtubules loaded with MWNTs.  In order to form a stable microtubulenanotube hybrid, the tight and specific interaction between biotin and streptavidin, i.e., between biotinylated microtubules and streptavidin-MWNT conjugates, was exploited.  Mimicking active cellular transport in synthetic environments may facilitate the design of nano-electromechanical systems (NEMS) such as molecular motor actuators that behave as nanoscale robots or the bottoms-up assembly of functional polymeric materials.  Moreover, these results begin to establish a platform for assembling individually addressable MWNT nanostructures using microtubule templates.

Schematic representation of the strategy employed to transport MWNTs using a kinesin-microtubule system. (a) Streptavidin-functionalized MWNT is attached to a biotinylated microtubule and moved in the flow chamber by kinesin-adsorbed molecules. (b) Time-lapse fluorescence micrographs of gliding hybrids (yellow) formed from biotinylated microtubules (red, TRITC) and streptavidin-functionalized nanotubes (green, FITC) on kinesin-coated surfaces. Nanotubes are transported by individual microtubules. All the microtubules in the field of view are loaded with streptavidin-coated nanotubes.