Previous Article |
Table of Contents
| Next Article 
BIOLOGICAL SCIENCES / BIOCHEMISTRY
A genetically encoded fluorescent amino acid
*Department of Chemistry, The Scripps Research Institute, 10550 North Torrey Pines Road, SR202, La Jolla, CA 92037; Beth Israel Deaconess Medical Center, Division of Signal Transduction, Harvard Medical School, 77 Avenue Louis Pasteur, Boston, MA 02115; Department of Chemistry, North Carolina State University, Campus Box 8240, Raleigh, NC 27695; and Division of Protein and Nucleic Acid Chemistry, Medical Research Council Laboratory of Molecular Biology, Hills Road, Cambridge CB2 2QH, United Kingdom
Contributed by Peter G. Schultz, May 12, 2006
 |
Abstract |
|---|
 Top Abstract Results and DiscussionConclusionMethodsAcknowledgementsReferences |
|---|
The ability to introduce fluorophores selectively into proteins provides a powerful tool to study protein structure, dynamics, localization, and biomolecular interactions both in vitro and in vivo. Here, we report a strategy for the selective and efficient biosynthetic incorporation of a low-molecular-weight fluorophore into proteins at defined sites. The fluorescent amino acid 2-amino-3-(5-(dimethylamino)naphthalene-1-sulfonamide)propanoic acid (dansylalanine) was genetically encoded in Saccharomyces cerevisiae by using an amber nonsense codon and corresponding orthogonal tRNA/aminoacyl-tRNA synthetase pair. This environmentally sensitive fluorophore was selectively introduced into human superoxide dismutase and used to monitor unfolding of the protein in the presence of guanidinium chloride. The strategy described here should be applicable to a number of different fluorophores in both prokaryotic and eukaryotic organisms, and it should facilitate both biochemical and cellular studies of protein structure and function.
molecular evolution | fluorescent probes | genetic code expansion | protein design | unnatural amino acids
The use of small synthetic dyes allows investigation of local changes in distance or polarity with high precision and without significant structural perturbation of the protein. Nevertheless, most reactive dyes must be introduced into the protein in vitro, they exhibit relatively low chemoselectivity, and they are often limited to accessible positions at the protein surface (9). Moreover, mutagenesis of the target protein is often required to generate unique sites for derivatization. Recently, strategies have been developed to label proteins with small fluorescent dyes in vivo. One of these methods makes use of cell-permeant, biarsenic dyes that bind to their target protein through two arsenic–dithiol interactions (10). Another approach exploits a biotin ligase that recognizes a hydrazide-reactive ketone substrate that can be attached to a 15-mer acceptor peptide and subsequently linked to a hydrazide-derivatized fluorophore (11). It has also been shown that a highly promiscuous O6-alkylguanine-DNA-alkyltransferase, fused to a protein of interest, can be labeled by fluorescent O6-benzylguanosine derivatives (12). However, all of these approaches rely on the introduction of specific dye-acceptor motifs, peptides or proteins that restrict the sites of modification and can introduce undesired structural perturbations into the protein to be analyzed. Finally, in vitro mutagenesis with suppressor tRNAs chemically modified with fluorescent amino acids can be used to label proteins with fluorescent probes site-specifically, but this method is largely limited to in vitro systems affording small quantities of proteins (13–16).
Many of the challenges involved in the generation of fluorescently labeled proteins would be overcome if one could genetically encode fluorescent amino acids directly in prokaryotic or eukaryotic organisms. To this end, we report the efficient and selective biosynthetic introduction of a dansyl-containing amino acid into proteins in Saccharomyces cerevisiae in response to the amber nonsense codon, TAG.
|
Results and Discussion |
|---|
|
TopAbstractResults and DiscussionConclusionMethodsAcknowledgementsReferences |
|---|
|
107 mutants was generated by randomizing residues Met-40, Leu-41, Tyr-499, Tyr-527, and His-537 in the leucine-binding site. These amino acids form a hydrophobic pocket around one 
-methyl group of leucine in the x-ray crystal structure of the homologous Thermus thermophilus LeuRS (Fig. 1B) (20). Because this binding pocket largely consists of side-chain atoms, mutations might be expected to accommodate novel amino acid substrates without significant perturbation of the LeuRS polypeptide backbone. To evolve a LeuRS specific for 1, a selection scheme was used in which the codons for Thr-44 and Arg-110 of the gene for the transcriptional activator GAL4 were both converted to amber nonsense codons (TAG) (21). Suppression of these amber codons in the MaV203:pGADGAL4(2TAG) yeast strain leads to production of functional full-length GAL4, which drives expression of genomic HIS3 and URA3 reporter genes. These genes complement histidine and uracil auxotrophy, allowing clones harboring active synthetase mutants that aminoacylate endogenous amino acids or 1 to be positively selected on synthetic dropout (SD) medium (containing dextrose) supplemented with 1 mM 1. Negative selection of synthetases that accept endogenous amino acids is carried out by growth on SD medium lacking 1 but containing 0.1% 5-fluoroorotic acid, which is converted into a toxic product by the URA3 gene product (21, 22).
After three rounds of selection (positive, negative, positive), the growth rates of 96 clones were assayed individually on selective media in the presence or absence of 1. Three different clones were isolated that are dependent on 1 for growth. The ability of these clones to incorporate 1 into proteins selectively was tested by suppression of an amber mutant (Trp-33 
TAG) of His6-tagged human superoxide dismutase (hSOD-33TAG-His6), and subsequent SDS/PAGE analysis with either GelCode Blue (Pierce) staining or fluorescence imaging. LeuRS mutant B8 in which all five active site residues were mutated (M40A, L41N, Y499I, Y527G, and H537T) afforded the most protein in the presence of 1; however, significant amounts of hSOD-33TAG-His6 were also produced in the absence of 1. Analysis of hSOD-33TAG-His6 expressed in the presence of 1 by MALDI-TOF MS indicated that leucine or isoleucine was also being incorporated at position 33.
Redesign of the Editing Domain and Enhancement of tRNA Expression.
E. coli LeuRS has an editing site to enhance selectivity toward structurally similar amino acids, e.g., isoleucine, valine, and methionine. Recent structural and biochemical studies have shown that hydrolysis of a cognate activated or charged leucine at this site is prevented by steric repulsion between one or more active-site amino acids and a -methyl group of leucine. Mutation of one such residue, Thr-252, to alanine (T252A) results in efficient hydrolysis of aminoacylated tRNALeu (23, 24). Consequently, to enhance fidelity of LeuRSB8, the T252A mutation was introduced (LeuRSB8T252A). Analysis of the suppression efficiency with hSOD-33TAG-His6 revealed relatively low yields of protein in the presence of 1 [0.29 mg/liter hSOD-33TAG-His6 harboring 1 after nickel-nitrilotriacetic acid (Ni-NTA) purification]. However, the T252A mutation resulted in a marked reduction of hSOD expression in the absence of 1. Previous studies of the suppression efficiency of an amber suppressor E. coli tRNATyr/TyrRS pair in yeast indicated that suppressor tRNA levels can limit overall protein yields (S.C., unpublished results). To increase tRNA expression, three copies of tRNACUALeu5 with flanking regions from yeast suppressor tRNA SUP4, which are known to contain an RNA polymerase (pol) III promoter, were inserted into the plasmid pLeuRSB8T252A (25, 26). These multiple tRNA genes were placed under the control of an additional phosphoglycerate kinase 1 (PGK1) promoter, resulting in an 4-fold enhancement (1.11 mg/liter) of expression of hSOD harboring 1 at position 33 (Fig. 2A). A high degree of fidelity for incorporation of 1 was confirmed by MALDI-TOF MS {calculated mass, m/z, 16,802; observed mass, 16,802 ([M+H]+); Fig. 2B}. No peaks corresponding to incorporation of endogenous amino acids were observed. Moreover, no protein was observed in the SDS/polyacrylamide gel in the absence of 1 (Fig. 2A).
|
-barrel fold with an external loop and short helical regions at one barrel face near the active site, which contains the copper and zinc ions (27). Loop IV of this region is additionally connected by a disulfide bond (Cys-57 and Cys-146) to -strand 8 of the barrel (Fig. 3). Monomeric hSOD has been shown to undergo reversible and complete unfolding in the concentration range 0–3.5 M guanidinium chloride (GdmCl) (28). To analyze local changes in the barrel core and outer rim regions during unfolding, 1 was incorporated at two different positions, Gln-16 and Trp-33 (Fig. 3). Position 33 lies on the barrel at the center of -strand 3, the first N-terminal strand of Greek-key 1 (27); surface residue Gln-16 forms the N-terminal tip of -strand 2 adjacent to Greek-key -strand 3.
|
max of 1 also blueshifted compared with the free amino acid (14 nm and 27 nm for positions 16 and 33, respectively). These spectral changes likely reflect partial shielding by surface side chains or dielectric differences between bulk water and the protein hydration shell. GdmCl denaturation only weakly affected the emission intensity and wavelength of the Gln-16 1 mutant (Fig. 4 A, C, and D). However, increasing the GdmCl concentration from 0 to 5 M caused a significant redshift in the max (528–544 nm) and decrease in the fluorescence intensity of the Trp-33 1 mutant (Fig 4 B–D). The most striking effects were observed at low GdmCl concentrations (0–1.5 M) followed by moderate changes from 1.5 M to 3.5 M GdmCl and a relatively small change between 3.5 M and 5 M GdmCl (Fig. 4). These results suggest that relatively large environmental changes occur at position 33 during protein unfolding. Previous NMR studies have revealed transitions from the native state to intermediate states of similar structure at 0–0.5 M GdmCl and a subsequent complete transition to an unfolded state at 3.5 M. At 2 M GdmCl, the intermediate states and the unfolded state are present in a 1:1 ratio, whereas only the unfolded state is present at >3.5 M GdmCl. Changes in 1H–15N heteronuclear single-quantum correlation NMR spectra indicate that the loop regions and -strand 3 are the first to be affected at low concentrations (0.5 M), whereas other -strands are affected only at higher concentrations (28). The observed spectral response of 1 at position 33 corresponds well with these findings. In contrast, the fluorescence of mutant protein with 1 at position 16 did not show a large response to unfolding, indicating low changes in local dielectric during structural transitions near this side chain. Removal of copper and zinc ions or reduction of the disulfide bridge resulted in similar unfolding behavior at sites 16 and 33. These studies demonstrate that unnatural amino acid 1 provides a useful probe of structural transitions at defined sites in proteins.
|
|
Conclusion |
|---|
|
TopAbstractResults and DiscussionConclusionMethodsAcknowledgementsReferences |
|---|
|
Methods |
|---|
|
TopAbstractResults and DiscussionConclusionMethodsAcknowledgementsReferences |
|---|
H (ppm) = 2.81 (6H, s, CH3), 3.0 (1H, dd, 2J = 13.2, 3J = 8.4, CH2a), 3.11 (1H, dd, 2J = 13.2, 3J = 5.2, CH2b), 3.31 (1H, dd, 3J = 8.4, 3J = 5.2, CH), 7.21 (1H, d, 3J = 7.2, Ar), 7.53 (1H, "t," 3J = 7.6, Ar), 7.61 (1H, dd, 3J = 8.4, 3J = 7.2, Ar), 8.12 (1H, dd, 3J = 7.2, 4J = 1.2, Ar), 8.25 (1H, d, 3J = 8.8, Ar), 8.46 (1H, dd, 3J = 7.6, Ar). 13C NMR (100 MHz, DMSO): C (ppm) = 43.6 (CH2), 45.0 (CH3), 53.5 (CH), 115.2, 118.9, 123.5, 128.0, 128.5, 128.9, 129.0, 129.6, 135.0, 151.3 (Ar), 168.3 (COOH). Electrospray ionization MS: m/z = 338.2 [M+H]+, 675.1 [2M+H]+. Evolution of Aminoacyl-tRNA Synthetases for Incorporation of Dansylalanine. Selection was carried out in the presence of 1 mM 1 as described in ref. 19. Three different clones whose growth rate depended on the presence of 1 emerged after three rounds of selection (B8: M40A, L41N, Y499I, Y527G, and H537T; C6: M40G, L41N, Y499I, Y527G, and H537G; and D12: M40G, L41L, Y499V, Y527G, and H537E. B8 additionally had an S497C mutation, presumably introduced by PCR during library design. The T252A mutation was introduced by using the QuikChange method (Stratagene) with primers LRST252AFwd (5'-ACA CTA CCC GCC CGG ACG CGT TTA TGG GTT GTA CCT A-3') and LRST252ARev (5'-TAG GTA CAA CCC ATA AAC GCG TCC GGG CGG GTA GTG T-3').
Vector Construction.
Vector pLeuRSB8T252A' was constructed as follows. E. coli tRNACUALeu5 and its flanking sequences were assembled by overlap PCR with the following primers (1Fwd, 5'-ATC CCG ACC GGT AAG CTG CTA GCC TCT TTT TC-3'; 1Rev, 5'-GAA GAA AGA GTA TAC TAC ATA ACA CAT ATA CAA TTG AAA AAG AGG CTA GC-3'; 2Fwd, 5'-GTA TAC TCT TTC TTC AAC AAT TAA ATA GCC CGG ATG GTG GAA TCG GTA GA-3'; 2Rev, 5'-CAG CGC GAA CGC CGA GGG ATT TAG AAT CCC TTG TGT CTA CCG ATT CCA CC-3'; 3Fwd, 5'-TCG GCG TTC GCG CTG TGC GGG TTC AAG TCC CGC TCC GGG TAT TTT TTT GT-3'; 3Rev, 5'-CGG CTC TAG ACA TAA AAA ACA AAA AAA TAC CCG-3'). The 5' and 3' sequences are identical to the flanking sequences of yeast suppressor tRNA gene SUP4, which is known to contain a pol III promoter (25, 26). This combined tRNACUALeu5-SUP4 sequence was digested with AgeI and XbaI and repetitively inserted into the AgeI/NheI sites of vector pLeuRSB8T252A to yield three copies of this gene in one direction in the same vector. Subsequently, a PCR DNA fragment including the 650-bp upstream sequence of yeast PGK1 ORF was inserted into the AgeI/NheI site to introduce an additional pol II promoter for transcription of the tRNA region. Vectors pC1SOD-33TAG-His6 and pC1SOD-16TAG-His6 were constructed as follows. S. cerevisiae strain 208708 was obtained from American Type Culture Collection. This strain produces Cu/Zn SOD from a 2-µm leu2-d plasmid. This plasmid was extracted, and a BamHI restriction fragment of 2 kb containing the SOD gene was gel purified and inserted into the BamHI site of pBKJYRS; the larger BamHI fragment was recircularized to produce plasmid pC1. A gene encoding His6-tagged hSOD with or without an amber codon at position 33 was used to replace the NcoI/SalI fragment of SOD in pBKJYRS. The amber mutation 16TAG was introduced by using the QuikChange method with primers Q16TAGF (5'-GAA GGG TGA CGG CCC AGT TTA GGG TAT TAT TAA CTT CGA GCA GAA G-3') and Q16TAGR (5'-CTT CTG CTC GAA GTT AAT AAT ACC CTA AAC TGG GCC GTC ACC CTT C-3'). The resulting pBKSOD plasmids were digested with BamHI and inserted into pC1 to give pC1SOD-33TAG-His6 and pC1SOD-16TAG-His6. These plasmids were transformed into the MJY125-derived strain SCY4 (MATa, ade2-101, ura3-1, leu2-3,112, trp1, his3-11, 
CYB2::kan [cir0]) by a lithium acetate method and plated on SD medium (–leucine). Plasmids pLeuRSB8, pLeuRSB8T252A, and pLeuRSB8T252A' were transformed into the obtained strains by using a lithium acetate method and plated on SD medium (–leucine, –tryptophan).
Expression, Purification, and Characterization of hSOD-His6 Mutants. Expression was performed with strain SCY4 transformed with the hSOD expression plasmid pC1SOD-33TAG-His6 or pC1SOD-16TAG-His6 and the respective tRNACUALeu5/LeuRS-encoding plasmid (pLeuRS derivatives). Cultures were grown to saturation at 30°C in SD medium lacking leucine and tryptophan and diluted 50-fold into SD medium containing dextrose, lacking leucine and tryptophan, and containing 5 mM 1. This culture was incubated for 16 h at 30°C, and cells were harvested. Cells were lysed with Y-PER lysis reagent (Pierce) containing 1x complete protease inhibitor mix (–EDTA; Roche) and centrifuged, and buffer was exchanged with native purification buffer A [50 mM Tris·HCl, pH 7.8/100 mM NaCl/5% (vol/vol) glycerol] by repeated ultrafiltration (Amicon; 10-kDa molecular mass cutoff, Millipore). For 1-liter culture volumes, protein was bound to 500 µl of Ni-NTA, washed three times with 1 ml of buffer A containing 25 mM imidazole, eluted three times with 1 ml of buffer A containing 500 mM imidazole, and dialyzed against water by using Slide-A-Lyzer cassettes (Pierce) with 5-kDa cutoff. MALDI-TOF MS measurements were performed by using a Voyager STR 2 spectrometer (Applied Biosystems). Yields were routinely quantified by using a BCA assay (Pierce).
Unfolding Experiments. Measurements were performed on a Fluoromax-2 (Instruments SA, Edison, NJ) equipped with a 150-W continuous xenon lamp at 22°C with excitation at 340 nm and a bandpass of 5 nm for both excitation and emission. Fluorescence of hSOD-His6 with 1 at position 16 or 33 at a concentration of 3 µM was measured in 50 mM sodium phosphate (pH 7.2)/100 mM NaCl/6 µM ZnCl2/6 µM CuSO4/GdmCl concentrations of 0, 0.5, 1.5, 2.0, 3.5, and 5.0 M. Samples were incubated for 1 h at 22°C after the addition of GdmCl before measurements. Demetallated protein was analyzed after the addition of 2.5 mM EDTA; reduced and demetallated protein was analyzed after the addition of 1 mM DTT and 2.5 mM EDTA. Emission was recorded in steps of 1 nm in the range of 500–600 nm, and the results are uncorrected. Blank measurements with buffer at the respective GdmCl concentrations with or without EDTA and/or DTT were subtracted from protein measurements to correct for nonspecific background fluorescence.
|
Acknowledgements |
|---|
|
TopAbstractResults and DiscussionConclusionMethodsAcknowledgementsReferences |
|---|
|
Footnotes |
|---|
Abbreviations: GdmCl, guanidinium chloride; hSOD, human copper/zinc superoxide dismutase; LeuRS, leucyl-tRNA synthetase; NTA, nitrilotriacetic acid; pol, RNA polymerase; SD, synthetic dropout; T252A, Thr-252 Ala; tRNACUALeu5, mutant amber suppressor tRNALeu5 of E. coli.
¶To whom correspondence should be addressed. E-mail: schultz{at}scripps.edu
Author contributions: D.S., S.C., N.W., and P.G.S. designed research; D.S., S.C., and N.W. performed research; A.D. and J.W.C. contributed new reagents/analytical tools; D.S., N.W., and P.G.S. analyzed data; and D.S. and P.G.S. wrote the paper.
Conflict of interest statement: No conflicts declared.
© 2006 by The National Academy of Sciences of the USA
|
References |
|---|
|
TopAbstractResults and DiscussionConclusionMethodsAcknowledgementsReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg What's this?
This article has been cited by other articles in HighWire Press-hosted journals:
|
|
 |
|
  Y. Lu, B. Bulka, M. desJardins, and S. J. Freeland Amino acid quantitative structure property relationship database: a web-based platform for quantitative investigations of amino acids Protein Eng. Des. Sel., July 1, 2007; 20(7): 347 - 351. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
, 0 M; 
, 0.5 M; 
, 1.5 M; 
, 2.0 M; 
, 3.5 M; 
, 5.0 M). Samples were measured in 50 mM sodium phosphate (pH 7.2) and 100 mM NaCl containing 3 µM hSOD and 2 eq of ZnCl2 and CuSO4, respectively. Spectra were recorded in 1-nm steps at 22°C by using excitation at 340 nm and a bandpass of 5 nm for both excitation and emission. (B) Fluorescence spectra of hSOD with 1 at position 33 conducted under conditions identical to those in A. (C) Fluorescence intensity of 1 at positions 16 and 33 of hSOD at unfolding states of the protein present at different GdmCl concentrations. The fluorescence intensity of spectra shown in A (
) and B (