Elsevier

Biosensors and Bioelectronics

Volume 26, Issue 1, 15 September 2010, Pages 55-61
Biosensors and Bioelectronics

Site-selective dual modification of periplasmic binding proteins for sensing applications

https://doi.org/10.1016/j.bios.2010.05.012Get rights and content

Abstract

We have developed three sensitive and specific amino acid sensors based on bacterial periplasmic solute binding proteins. A site-specific amino-terminal transamination reaction provides a useful complement to cysteine chemistry for the covalent modification of biomolecules in this application. We demonstrate this combination to attach two different chromophores to a single biomolecule in two locations. The periplasmic glutamine binding protein from E. coli was modified with a pair of dyes suitable for fluorescence resonance energy transfer, and this conjugate exhibited an l-glutamine dependent optical response. Two periplasmic binding proteins from the thermophilic organism Thermotoga maritima, for arginine and aliphatic amino acids, were modified and evaluated similarly. All three conjugates manifested signal changes mediated by resonant energy transfer upon binding their respective ligands, with nanomolar dissociation constants and stereochemical specificity. This represents a readily generalizable method for construction of reagentless biosensors. The double-labeling strategy was also exploited for the surface attachment of a dye-labeled glutamine binding protein via a biotin–streptavidin interaction.

Introduction

Periplasmic binding proteins (PBPs) comprise a family of proteins involved in chemotaxis and uptake of nutrients and other small molecules from the extracellular space (Ames, 1986). A wide variety of PBP ligands are known, including amino acids, peptides, simple and complex sugars, inorganic ions and metals (Dwyer and Hellinga, 2004). Each member of this protein family binds its ligand or ligands with high affinity and specificity between two globular domains that enclose the binding site, Fig. 1a. Members of the family exhibit a conserved “venus fly-trap” conformational change upon ligand binding, in which a hinge-twist motion brings the two protein domains together (Quiocho, 1990). The ability to target a variety of diverse ligands, combined with the specificity and dramatic structural change associated with ligand binding, have attracted the attention of researchers interested in sensing devices for industry, biology, medicine, defense, and astrobiology (Brune et al., 1994, Dattelbaum and Lakowicz, 2001, de Lorimier et al., 2002, Fehr et al., 2003, Gilardi et al., 1994, Medintz et al., 2003). Homologs of these proteins, first identified in prokaryotes, have been observed in eukaryotes and archaea. Sensors based on PBPs are particularly promising for the detection of compounds like amino acids and carbohydrates that could serve as markers for the presence of living organisms in extraterrestrial environments. In this application the intrinsic stereochemical specificity of the PBPs is desirable, offering potential ways to distinguish amino acids produced by abiotic processes, which are enantiomeric mixtures, from those produced by biology, which are widely expected to be enriched for specific stereoisomers (Pace, 2001).

The construction of reagentless biosensors from PBPs requires coupling of conformational changes to change in an observable property (Dwyer and Hellinga, 2004). In particular, these proteins seem well-suited to exhibit changes in fluorescence resonance energy transfer (FRET) efficiency on ligand binding due to the large conformational changes they undergo (Heyduk, 2002). With the advent of modified versions of green fluorescent protein exhibiting distinct excitation and emission spectra, it became possible to create sensors based on energy transfer between a genetically encoded donor and acceptor, initially demonstrated in sensors for proteolytic cleavage (Heim and Tsien, 1996, Mitra et al., 1996). The first reversible sensors based on this principle were sensors for calcium ions based on calmodulin (Miyawaki et al., 1997). A variety of FRET sensors based on PBPs have since been created by Frommer and co-workers by fusing pairs of fluorescent proteins to various sites in or around maltose, ribose, glutamate, glucose and phosphate binding proteins (Deuschle et al., 2005, Fehr et al., 2002, Fehr et al., 2003, Fehr et al., 2005, Gu et al., 2006, Lager et al., 2003, Okumoto et al., 2005). These fusions can exhibit good signal to noise ratios and can be directly expressed in bacteria or eukaryotes to study the flux of small molecules in the cytoplasm or on the cell surface. While fluorescent proteins with a wide range of spectral properties are now available, obtaining a good signal change requires significant trial and error in the positioning of protein fusions (Deuschle et al., 2005). Some of the other myriad physical and biological considerations in creating FRET-based sensors have been articulated recently by van der Krogt et al. (2008). Extension of this work to small molecule fluorophores could allow further tuning of optical properties by exploiting dyes with a wider range of absorbance and emission spectra, larger Stokes shifts, more resistance to photobleaching and reduced sensitivity to pH and ion concentrations.

In order to use FRET to monitor protein conformational changes using non-natural dyes or to attach a chemically labeled protein to a solid support it is necessary to perform two different covalent modifications. Several approaches have been applied to site-selective attachment of multiple chemical labels to proteins, including non-natural amino acid mutagenesis (Kajihara et al., 2006), solid phase peptide synthesis (Marcaurelle et al., 1998), and semisynthesis through solid phase expressed protein ligation (Cotton and Muir, 2000). A cysteine protection strategy has also been employed, which allows the incorporation of thiol modifications at up to three specific sites in a protein using inserted sequences of 32 and 18 residues (Smith et al., 2005). This work resulted in FRET sensors based on maltose binding protein. We describe the use of two orthogonal chemical reactions to modify proteins at specific locations, an approach that can be applied to a wide variety of proteins easily and without changes to the sequence beyond a single cysteine point mutation. We then demonstrate application of double-labeled proteins from a mesophilic and a thermophilic organism to chemical sensing and protein immobilization.

Site-selective attachment of two different small molecules to the same protein has been a barrier to the development of organic dye FRET pairs for the detection of protein conformational changes. Precise labeling with one donor dye and one acceptor dye maximizes signal change and simplifies interpretation. Site-selective, covalent modification of native proteins at two specific sites can be difficult to achieve, due to the large number of reactive groups on protein surfaces. Reactions that target cysteine are widely used because cysteine is a relatively rare amino acid and its thiol group can be modified quite selectively (Hermanson, 1996), but site-specific placement of the second dye still constitutes a challenge.

As an alternative approach for the construction of FRET-based biosensors, offering potential advantages in terms of generality, we report the site-selective incorporation of two fluorophores into proteins purified from bacterial expression through the combination of traditional cysteine chemistry and an amino-terminal transamination reaction, Fig. 1a and b (Esser-Kahn and Francis, 2008, Gilmore et al., 2006, Scheck and Francis, 2007). The transamination, mediated by pyridoxal 5′-phosphate (PLP, 1), installs a uniquely reactive ketone or aldehyde group at the amino-terminus of the protein that can be further elaborated with alkoxylamine-bearing reagents (such as 2) through oxime formation. Several fluorophores with these functional groups are now commercially available, and thus can be installed at this location with high specificity. This strategy was used to convert a series of PBPs into small molecule sensors that report ligand-induced conformational movements via changes in FRET efficiency or to immobilize a fluorescently labeled protein.

Section snippets

Expression plasmid construction

A plasmid encoding the mature form of Gln-BP, lacking a secretion signal sequence and including an appended sequence of five histidines, was obtained in a pK223-2 expression plasmid (de Lorimier et al., 2002). Cysteine point mutations were introduced using the QuikChange mutagenesis method (Stratagene). See supporting material for primer sequences.

Wild-type genes for arginine (Arg-BP) and leucine–isoleucine–valine (LIV-BP) binding proteins were amplified by PCR from genomic DNA of Thermotoga

Preparation of protein bioconjugates

Our initial studies targeted the E. coli glutamine binding protein (GlnH), as crystal structures were available for both the open and closed forms, Fig. 1a. Based on these structures, a series of locations were chosen such that their distance from the amino-terminus might change significantly upon the closing of the binding pocket. Residues selected for mutation were at positions 99, 107, 110, 116, 118, 127, 131, 132 and 135 in the protein, as numbered in the crystal structures. Distances

Conclusion

We have created FRET-based chemical sensors from mesophilic and thermophilic proteins using two highly specific, orthogonal chemical modifications. These sensors exhibit strong chemical and stereochemical specificity and two sensors derived from a thermophilic organism have dissociation constants in the low nanomolar range. We anticipate that this double-labeling strategy will be applicable to many proteins that undergo conformational changes upon ligand binding, yielding new optical sensors

Acknowledgements

This work was supported by National Aeronautics and Space Administration, Astrobiology Science and Technology Instrument Development program grants ASTID04-0000-0108 and 07-ASTID07-0092.

References (32)

  • G.J. Cotton et al.

    Chem. Biol.

    (2000)
  • J.D. Dattelbaum et al.

    Anal. Biochem.

    (2001)
  • C.G. dos Remedios et al.

    J. Struct. Biol.

    (1995)
  • M.A. Dwyer et al.

    Curr. Opin. Struct. Biol.

    (2004)
  • M. Fehr et al.

    J. Biol. Chem.

    (2003)
  • H. Gu et al.

    FEBS Lett.

    (2006)
  • R. Heim et al.

    Curr. Biol.

    (1996)
  • T. Heyduk

    Curr. Opin. Biotechnol.

    (2002)
  • I. Lager et al.

    FEBS Lett.

    (2003)
  • L.A. Marcaurelle et al.

    Tetrahedron Lett.

    (1998)
  • R.D. Mitra et al.

    Gene

    (1996)
  • Q.C. Shen et al.

    J. Mol. Biol.

    (1989)
  • B.W. van der Meer

    Rev. Mol. Biotechnol.

    (2002)
  • G.F.L. Ames

    Annu. Rev. Biochem.

    (1986)
  • M. Brune et al.

    Biochemistry

    (1994)
  • R.M. de Lorimier et al.

    Protein Sci.

    (2002)
  • Cited by (0)

    View full text