Molecular Design of Inorganic-Binding Polypeptides

Controlled binding and assembly of peptides onto inorganic substrates is at the core of bionanotechnology and biological-materials engineering. Peptides offer several unique advantages for developing future inorganic materials and systems. First, engineered polypeptides can molecularly recognize inorganic surfaces that are distinguishable by shape, crystallography, mineralogy, and chemistry. Second, polypeptides are capable of self-assembly on specific material surfaces leading to addressable molecular architectures. Finally, genetically engineered peptides offer multiple strategies for their functional modification. In this article, we summarize the details and mechanisms involved in combinatorial-polypeptide sequence selection and inorganic-material recognition and affinity, and outline experimental and theoretical approaches and concepts that will help advance this emerging field. Molecular Design of Inorganic-Binding Polypeptides John Spencer Evans, Ram Samudrala, Tiffany R. Walsh, Ersin Emre Oren, and Candan Tamerler Introduction During the last two decades, combina torial-peptide selection methods1–3 have been used to generate sequence libraries that recognize and bind to different inorganic solids.4–12 Surface-exposed or displayed polypeptides produced by phage5,7–10,12 and bacteria4,6,10,11 have become the predominant in vivo techniques for material-specific peptide selection, and inorganic-binding peptides are quickly becoming molecular tools for biotechnological and nanotechnological applications. With 20 naturally occurring amino acids available for use, biological organisms can craft an extremely large and diverse set of linear sequences for a wide range of materials, including metals, oxides, semiconductors, and minerals. In addition to the numerous linear combinations, the potential twoand threedimensional configurations of these sequences add another dimension in that there are a number of polypeptide backbone geometries (i.e., secondary structures) that form from different amino-acid sequences of these peptides. Hence, factors such as polypeptide–material affinity and selectivity ultimately are chosen by the sequence and its molecular architecture as well as the chemical composition of the peptide.13,14 Thus, a successful design of polypeptide-inorganic materials is dependent upon our understanding of the molecular factors that govern sequence–structure–function selection. This article will summarize our current knowledge of phage and bacterialgenerated polypeptides directed against inorganic solids, using examples obtained from experiment and theory to define the molecular trends emerging from the screened polypeptide libraries generated against artificial and biological inorganic materials such as Pt, Au, hydroxyapatite, graphite, and quartz. First-Generation Peptides and Post-Selection Engineering A genetically engineered polypeptide for inorganics (GEPI) is defined as an aminoacid sequence that specifically and selectively binds to an inorganic surface.10 Bacterial-cell surface (BCS) and phage display (PD) libraries have been adapted to select for a variety of GEPIs.5–12 Typically, these libraries are generated by artificially inserting randomized nucleotides within genes specifying cell-surface or phage coat proteins. The host library (which typically consists of 109–1011 different members of either cells or phage) then is exposed to the desired substrate. The displayed surface-coat polypeptides on the hosts have different sequences that come in contact with the inorganic surfaces. Mild chemicalelution conditions remove weak or nonspecific binders, and strongly binding cells or viruses are recovered.15 This biopanning cycle is repeated a number of times to enrich specificity and highaffinity binders. Eventually, the amino-acid sequences of the inorganic-binding regions are deduced by DNA sequencing and cataloged. Once the first set of binding peptides is obtained, their affinity and specificity can be further “tuned” via the use of various molecular-tailoring strategies. For example, binding properties of a selected polypeptide can be tweaked by sitespecific change(s) of amino acids within the sequence. Molecular constraints and the use of multiple sequence repeats (i.e., multimerization) can also be used to tune the binding properties and conse quently the structural features of the initial polypeptide sequence.6,14 As an example, we found that the cyclic form of a Pt binding sequence exhibits a higher affinity than the linear version using surface plasmon resonance (SPR) spectroscopy (Figure 1).14 In the case of multimerization, we demonstrated that the affinity and selectivity for given inorganic materials improved as a function of the repeat number of the poly peptide sequence (e.g., 3-versus 1-repeat gold binding peptides).16 Hence, the initial sequences serve as a starting point for further improvement or modification. Although the combinatorial biology techniques are relatively straightforward, there are several important considerations that need to be addressed when combining inorganic materials with biological agents. First, the method for separating materialbound hosts from unbound ones may disqualify a particular display technology. For instance, phage particles are limited in size and thus are suitable for work with inorganic powders and enrichment by centrifugation separation techniques.15 The bacterial flagellin cell-surface-display system would not be amenable to this enrichment process since centrifugal forces would shear off the long flagella or tail from the bacterial cell.11 Second, the chemical and physical states of the inorganic surface itself may affect the efficiency of polypeptide binding. For example, many materials rapidly develop a surface-oxide layer when exposed to air or solution or may become modified when incubated in 514 MRS BULLETIN • VOLUME 33 • MAY 2008 • www.mrs.org/bulletin the biological media used during the panning process. Thus, it is important to characterize inorganic surfaces prior to and after panning procedures to determine if any alterations have taken place. Third, the compatibility of inorganic materials with biological buffers may need to be addressed. One may need to monitor the effect of wash or elution buffers on inorganic-surface integrity and optimize parameters to guard against surface modification, etching, or other forms of surface deterioration. Fourth, inorganic compounds come in a variety of forms, from polydisperse powders to single crystals. With diverse interfacial features available on different surfaces of the same solid, peptides may reconform to recognize different surface features. Thus, a different binding sequence could emerge depending on the nature of the surface topology. Finally, our expectations regarding sequence library convergence need to be re-examined when we use inorganic materials for selection. In traditional biological applications of peptide libraries, at the end of three to five biopanning cycles, the selected sequences typically converge toward a consensus consisting of identical sequences. However, this rule does not apply in the case of inorganic-binding sequences where similarities, rather than a strict consensus, are generally observed. This presumably reflects the heterogeneity of the inorganic substrate at the atomic, crystallographic, and morphological levels, as well as other, perhaps chemical, factors.15 Molecular Structures: Experimental Perspective Polypeptide structure influences function, and therein lies the challenge for GEPI research. Although it is relatively easy to generate sequence libraries against a given material, it is not so straightforward to wade through this expansive library, determine the individual poly peptide structures of this ensemble, and examine how these structures relate to function. Recent experimental studies that have been carried out with the GEPIs selected for a wide variety of materials such as Au,4,16,17 Pt,15,18 carbon nanohorns,19 and hydroxyapatite20 give us a glimpse of what may be general structural “rules” that exist within peptide sequences. The first trend is that both M13 phage pIII (7 or 12 amino acids (AA))8–10 and bacterial-cell receptor-generated11 polypeptides (14 AA, 42 AA) exhibit unfolded conformations that fall within two classifications (Figure 2). The first classification is predominantly randomcoil (RC) structures in equilibrium with other secondary structures such as alphahelix, beta-strand, and beta-turn.10,11 The second classification is polyproline Type II (PPII), an extended helical secondary structure common to sequences containing Pro, Ala, Gln, and other PPIIforming amino acids.14,20 This secondary structure is believed to exist in equilibrium with RC conformation but not with alpha helix or beta strand. What is the significance of unfolded structures that exist either as RC or PPII in material-selected polypeptide sequences? To answer this question, there are two hypotheses to consider. The first consideration is that both structures allow sidechain accessibility to the solvent and interfacial environments.14,21 This means that potential peptide–material interface interactions would be expected to be maximal for an unfolded polypeptide as opposed to a folded peptide that, due to internal contacts and folded topology, can offer only limited surface(s) for interaction. This phenomenon is also observed in biomineral-associated polypeptides.22–24 The second consideration is that labile, unfolded conformations are potentially better at adapting to irregular surface topologies at an inorganic interface than a polypeptide with an internal structure that is already stabilized and fixed.21–24 Thus, focusing on unstructured sequences appears to be the approach that nature uses for selection with inorganic materials. In addition to these structural considerations, there is also the fact that M13 phage pIII sequences are expressed in two geometric configurations: either as the 7 AA or 12 AA linear form or as a 7 AA cyclic (loop-constrained) form.14,20 Molecular Design of Inorganic-Binding Polypeptides MRS BULLETIN • VOLUME 33 • MAY 2008 • www.mrs.org/bulletin 515 7