Aminotryptophan-containing barstar: Structure–function tradeoff in protein design and engineering with an expanded genetic code

https://doi.org/10.1016/j.bbapap.2006.04.012Get rights and content

Abstract

The indole ring of the canonical amino acid tryptophan (Trp) possesses distinguished features, such as sterical bulk, hydrophobicity and the nitrogen atom which is capable of acting as a hydrogen bond donor. The introduction of an amino group into the indole moiety of Trp yields the structural analogs 4-aminotryptophan ((4-NH2)Trp) and 5-aminotryptophan ((5-NH2)Trp). Their hydrophobicity and spectral properties are substantially different when compared to those of Trp. They resemble the purine bases of DNA and share their capacity for pH-sensitive intramolecular charge transfer. The Trp → aminotryptophan substitution in proteins during ribosomal translation is expected to result in related protein variants that acquire these features. These expectations have been fulfilled by incorporating (4-NH2)Trp and (5-NH2)Trp into barstar, an intracellular inhibitor of the ribonuclease barnase from Bacillus amyloliquefaciens. The crystal structure of (4-NH2)Trp-barstar is similar to that of the parent protein, whereas its spectral and thermodynamic behavior is found to be remarkably different. The Tm value of (4-NH2)Trp- and (5-NH2)Trp-barstar is lowered by about 20 °C, and they exhibit a strongly reduced unfolding cooperativity and substantial loss of free energy in folding. Furthermore, folding kinetic study of (4-NH2)Trp-barstar revealed that the denatured state is even preferred over native one. The combination of structural and thermodynamic analyses clearly shows how structures of substituted barstar display a typical structure–function tradeoff: the acquirement of unique pH-sensitive charge transfer as a novel function is achieved at the expense of protein stability. These findings provide a new insight into the evolution of the amino acid repertoire of the universal genetic code and highlight possible problems regarding protein engineering and design by using an expanded genetic code.

Introduction

One notable property of protein translation that did not escape the attention of many researchers is that it is performed with an invariant set of 20 amino acids prescribed by the universal genetic code [1]. The chemical diversity encoded in this way includes a defined number of side chain functional groups: carboxylic acids (Glu, Asp), amides (Gln, Asn), thiol (Cys) and thiolether (Met), alcohols (Ser, Thr), amines (Lys), guanidine (Arg), aliphatic (Ala, Val, Leu, Ile) and aromatic side chains (Phe, Tyr, His, Trp), and even a cyclic imino acid (Pro). On the other hand, a remarkable tolerance of the ribosome (i.e., ribosome plasticity) towards many amino acid types ranging from derivatives of biotin to polycyclic aromatic systems is now well documented [2]. A vast number of amino acids can be translated into protein sequences. Examples include amino acids with increased side chain bulk, novel chemical groups and functionalities (e.g., amino, cyano, nitro, nitroso, keto, azido, alkene, alkyne) and atoms such as various halogens or chalcogens and even novel backbone chemistries [3], [4], [5], [6]. Obviously, ribosome-mediated synthesis of proteins provides a sufficient room for increasing the chemical diversity of the genetic code vocabulary. Proteins generated in this way often possesses unique properties transcending those found in nature and might serve as useful tool for understanding the precise molecular basis of protein structure–function relationships, ligand binding, conformational change and protein–protein interactions [7].

The expansion of the amino acid repertoire in protein biosynthesis is possible since numerous noncanonical analogs and surrogates of canonical amino acids can act as substrates for translation machinery. This requires their intracellular accumulation by the cellular uptake systems. These systems and the aminoacyl-tRNA synthetases (AARS), crucial enzymes in the interpretation of the genetic code, often cannot distinguish between structurally or chemically similar amino acids [8]. In the experimental approach termed as selective pressure incorporation (SPI), these features are exploited. The basic principle of this method is that the choice of amino acids for protein synthesis can be conditioned by the control of environmental factors by changing the fermentation parameters. In this way, a controlled expression of a target protein in bacterial cells which are auxotrophic for particular amino acids is accompanied by ‘mis-incorporations’ of the related analog. Such a reprogrammed translation yields proteins with a changed chemical composition. In summary, SPI is now a well-established methodology that essentially requires the availability of metabolically engineered cells (auxotrophs) as expression hosts and controllable expression conditions [9]. This provides a basis for efficient amino acid replacement by codon reassignment on the level of a single protein [10].

The canonical amino acid Trp possesses at least three salient features: (i) low natural abundance in proteins (only ∼ 1%), (ii) unique biophysical and chemical properties, and (iii) its involvement in a variety of biological interactions and functions [11]. In addition, Trp encoded by a single UGG triplet is the main reporter of intrinsic fluorescence in most proteins and peptides. Thus, it is an attractive target for tailoring the spectral properties of proteins. In our previous studies, it has been shown that the spectral function of proteins can be engineered by the incorporation of various hydroxy-, amino-, halogen-, and chalcogen-containing Trp analogs and surrogates [12].

Aminotryptophans have a special place among all these substances. From a chemical point of view, they represent a class of aromatic amines that are much weaker bases than aliphatic amines. Further, (4-NH2)Trp and (5-NH2)Trp are also interesting due to their structural similarity to the natural DNA base adenine. In adenine, the exocyclic aromatic amino group is often a key recognition element in the formation of nucleic acid structures and in intermolecular interactions of DNA/RNA and proteins. In addition, the fluorescence properties of all four DNA bases are characterized by the existence of an excited state with charge transfer character [13]. During translation with (4-NH2)Trp and (5-NH2)Trp, these properties can be transmitted into the resulting protein structure as a response to all UGG triplets present in target sequence. Proteins synthesized in this way are endowed with a pH-sensitive fluorescence which results from charge transfer [14]. In other words, a variation in solvent pH causes a predictable change in the optical properties of these protein variants which might be an interesting practical application. In classical chemistry, aromatic amines are basic substrates for the synthesis of new pigments and dyes. Similarly, the incorporation of aminotryptophans into the “cyan” mutant of the green fluorescent protein (GFP) has yielded a protein with a novel chromophore. This specific engineering approach has given rise to a unique “gold” class of autofluorescent proteins with unique chromophore photophysics [15].

However, an important issue to be answered is about the thermodynamic price that must be paid in order to generate such proteins with tailored functions. Although rarely appreciated in current studies [16], this issue is highly relevant for protein design and engineering with an expanded genetic code. In order to address this problem, we extended our previous spectroscopic studies [14] on barstar as model protein by additional thermodynamic, X-ray and modeling investigations of (4-NH2)Trp- and (5-NH2)Trp-barstar. To dissect the influence of the amino group on the determined thermodynamic features of these protein variants, the incorporation of methylated Trp analogs, like 4-, 5-, 6-, and 7-methyltryptophan ((4-Me)Trp, (5-Me)Trp, (6-Me)Trp and (7-Me)Trp), into barstar has been attempted as well (Fig. 1). Indeed, the tryptophan → aminotryptophan → methyltryptophan replacement is a model for a –H → –NH2 → –CH3 isosteric substitution in proteins. Experiments reported in this study are expected to provide important insights into the fundamental roles of aromatic amines in structural stability and functions of tailor made proteins synthesized by the ribosome machinery.

Section snippets

Chemicals, expression host, fermentation, protein labeling, expression and purification

The noncanonical Trp-analogs (4-NH2)Trp and (5-NH2)Trp were synthesized enzymatically as described elsewhere [14]. All other chemicals were purchased from Sigma-Aldrich unless stated otherwise. Tryptophan auxotrophic bacterial expression host E. coli ATCC49980 (genotype: WP2 uvrA derivative of E. coli K-12 W3110) was used for expression, fermentation and incorporation experiments. In brief, transformed host cells were grown in New Minimal Medium (NMM) which contains 22 mM KH2PO4, 50 mM K2HPO4,

Translational activity of methyl- and aminotryptophan analogs

Barstar pwt, widely used as a model protein for folding studies, contains three Trp residues: Trp38 and Trp44 which are fully or partially solvent exposed and Trp53 which is completely buried inside the hydrophobic core of the protein [30]. Recently, we have discovered that all these three Trp residues of pwt-barstar can be substituted by (4-NH2)Trp and (5-NH2)Trp but not with (6-NH2)Trp and (7-NH2)Trp using the SPI method with the Trp-auxotrophic E. coli ATCC49980 strain as expression host [14]

Barstar folding and the functional role of its Trp residues

In principal, the stabilizing element in all globular proteins is the hydrophobic core. This core in the N-state structure, shielded from the solvent, is probably the most tightly packed form of organic matter [23]. Its close packing provides a unique and constrained conformation, whereas loose packing is often associated with flexible hinges and conformational changes. This high degree of side chain packing in a rigid core of proteins must be established from an unfolded conformation. Two

Acknowledgements

We thank Robert Huber and Luis Moroder for providing us with infrastructure and laboratory space. We are grateful to Peter Göttig and M. Kamran Azim for their help in crystallographic and modeling experiments and Mekdes Debela for help in manuscript preparation. Special thanks from R. G. go to his internship students Franziska Seifert and Erik Hinze. This work was supported by Deutsche Forschungsgemeinschaft (SFB533) and BMBF “BioFuture” program.

References (63)

  • D.B. Cowie et al.

    Amino acid analog incorporation into bacterial proteins

    Biochim. Biophys. Acta

    (1959)
  • K.B. Wong et al.

    Cold denaturation of barstar: H-1, N-15 and C-13 NMR assignment and characterisation of residual structure

    J. Mol. Biol.

    (1996)
  • L. Pradeep et al.

    Osmolytes induce structure in an early intermediate on the folding pathway of Barstar

    J. Biol. Chem.

    (2004)
  • S.K. Burley et al.

    Weakly polar interactions in proteins

    Adv. Protein Chem.

    (1988)
  • M.C. Shastry et al.

    The folding mechanism of barstar: evidence for multiple pathways and multiple intermediates

    J. Mol. Biol.

    (1995)
  • K. Sridevi et al.

    The slow folding reaction of barstar: the core tryptophan region attains tight packing before substantial secondary and tertiary structure formation and final compaction of the polypeptide chain

    J. Mol. Biol.

    (2000)
  • W. Kauzmann

    Some factors in the interpretation of protein denaturation

    Adv. Protein Chem.

    (1959)
  • R. Wolfenden et al.

    How hydrophilic is tryptophan

    Trends Biochem. Sci.

    (1986)
  • N. Budisa
  • N. Budisa

    Prolegomena to future efforts on genetic code engineering by expanding its amino acid repertoire

    Angew. Chem., Int. Ed. Engl.

    (2004)
  • L. Wang et al.

    Expanding the genetic code

    Chem. Commun.

    (2002)
  • N. Budisa
  • N. Budisa et al.

    Toward the experimental codon reassignment in vivo: protein building with an expanded amino acid repertoire

    FASEB J.

    (1999)
  • J.H. Bae et al.

    Crystallographic evidence for isomeric chromophores in 3-fluorotyrosyl-green fluorescent protein

    ChemBioChem

    (2004)
  • V. Shafirovich et al.

    Spectroscopic investigations of charge transfer in DNA

  • N. Budisa et al.

    Global replacement of tryptophan with aminotryptophans generates non-invasive protein-based optical pH sensors

    Angew. Chem., Int. Ed. Engl.

    (2002)
  • J.H. Bae et al.

    Expansion of the genetic code enables design of a novel “gold” class of green fluorescent proteins

    J. Mol. Biol.

    (2003)
  • T.T.F. Wong

    Evolution of the genetic code

    Microbiol. Sci.

    (1998)
  • N. Budisa et al.

    High-level biosynthetic substitution of methionine in proteins by its analogs 2-aminohexanoic acid, selenomethionine, telluromethionine and ethionine in Escherichia coli

    Eur. J. Biochem.

    (1995)
  • B. Nölting et al.

    The folding pathway of a protein at high resolution from microseconds to seconds

    Proc. Natl. Acad. Sci. U. S. A.

    (1997)
  • R. Golbik et al.

    Folding of barstar C40A/C82A/P27A and catalysis of the peptidyl-prolyl cis/trans isomerization by human cytosolic cyclophilin (Cyp18)

    Protein Sci.

    (1999)
  • Cited by (26)

    View all citing articles on Scopus
    1

    Both authors equally contributed to this work.

    View full text