This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.
Article pubs.acs.org/biochemistry
Conformational Heterogeneity and DNA Recognition by the Morphogen Bicoid Ramkrishna Adhikary, Yun Xuan Tan, Jian Liu, Jörg Zimmermann, Matthew Holcomb, Carolyn Yvellez, Philip E. Dawson, and Floyd E. Romesberg* Department of Chemistry, The Scripps Research Institute, La Jolla, California 92037, United States S Supporting Information *
ABSTRACT: The morphogenic activity of the Drosophila transcription factor bicoid (Bcd), the first morphogenic protein identified, is controlled by its DNA binding homeodomain. Homeodomains mediate developmental processes in all multicellular organisms, but the Bcd homeodomain appears to be unique as it can bind multiple DNA sequences and even RNA. All homeodomain proteins adopt a three-helix fold, with residues of the third helix mediating recognition of the nucleic acid target via interactions with the major groove. Interestingly, previous studies have revealed that conformational heterogeneity is present in the Bcd residues that interact with bound DNA, suggesting that it may underlie the morphogen’s unique polyspecificity. To begin to directly characterize the conformational heterogeneity in the homeodomain, we have introduced C− D bonds within each structural element and characterized their absorptions in the free and bound states, as well as during thermal denaturation. The data reveal that while residues within the first two helices experience unique environments, each environment is well-defined and similar in the presence and absence of bound DNA. In contrast, the data are consistent with residues within the recognition helix adopting multiple conformations, and while the binding of DNA does alter the environments, the conformational heterogeneity is similar in the bound and unbound states. Finally, thermal denaturation studies reveal that the conformational heterogeneity observed in this and previous studies results not from local instability and unfolding, as has been suggested for other transcription factors, but rather from the population of multiple stable conformations within the folded state of the protein. The results have important implications for how Bcd recognizes its different targets to mediate its critical developmental functions.
P
plays a role in Bcd function.11,12 With homeodomains in general, recognition helix residue 50 plays a key role in DNA binding,16−18 and DNA protection assays suggest that with Bcd, the side chain of Lys50 adopts different conformations when bound to different target sequences.12 Moreover, when the Bcd homeodomain is bound to its consensus recognition sequence, nuclear magnetic resonance (NMR) line broadening suggests that the Lys50 side chain adopts multiple conformations, and although they could not be resolved because of rapid exchange on the NMR time scale, their presence is consistent with observed NOEs and modeling studies.11 Such conformational heterogeneity could play an important role in allowing Bcd to recognize its different targets. If so, the mechanism of recognition depends on whether the heterogeneity is present in the free protein or only induced by DNA binding, and whether the heterogeneity is caused by fluctuations of the folded state or by local unfolding, as has been suggested for other transcription factors.19 Differentiating between these
rotein morphogens control tissue development, and the transcription factor bicoid (Bcd), which regulates the formation of the anterior−posterior axis in the developing Drosophila embryo,1−3 was the first morphogen discovered. Bcd mRNA is transcribed maternally and then deposited in the anterior end of the developing oocyte; upon fertilization, its translation produces an anterior−posterior concentration gradient of Bcd morphogen, which in turn dictates its position-dependent gene regulation. As with other morphogens, binding of Bcd to DNA is mediated by a single homeodomain, which is a conserved class of DNA binding domains that is found throughout multicellular organisms.4−9 Homeodomains adopt a three-helix fold, in which the second and third helices form a structure similar to the helix−turn− helix motif of prokaryotic transcriptional repressors10,11 (Figure 1). Bcd target recognition results from residues of the third helix, also known as the recognition helix, inserting into the major groove of DNA and making sequence-specific contacts with nucleobases therein. The consensus DNA recognition sequence of Bcd is d(TAATCC), but unlike other homeodomains, it also recognizes several additional, nonconsensus sequences,12,13 as well as RNA.14,15 This has led to speculation that flexibility © XXXX American Chemical Society
Received: March 20, 2017 Revised: May 5, 2017
A
DOI: 10.1021/acs.biochem.7b00255 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry
Val45 and Lys50, the spectral complexity is similar in the free and DNA-bound states. Despite this apparent heterogeneity, Val45, and presumably the entirety of the recognition helix, unfolds along with the rest of the homeodomain in a single, global transition, in the free and DNA-bound states. Thus, the heterogeneity results not from instability and local unfolding but rather from the population of distinct stable conformations within the folded state. Such heterogeneity could contribute to the unique ability of Bcd to recognize its range of different targets.
■
EXPERIMENTAL PROCEDURES Synthesis of the Bicoid Homeodomain. Previous studies of the Bcd homeodomain have employed a variant with an extra N-terminal Gly, which resulted from preparation via TEV protease cleavage of a His tag, and with the seven C-terminal amino acids included to increase solubility.11 To be consistent with these studies, we employed the same 68-amino acid protein. Briefly, synthesis of the C-terminal peptide (amino acids 32−67) and the N-terminal peptide (amino acids −1 to 31) was performed on 0.4 mmol of Boc-Ser(Bzl)-OCH2-Pam resin and Boc-Leu-Pam resin, respectively, using the following side-chain-protected amino acids: Arg(Tos), Asn(Xan), Asp(OcHx), Gln(Xan), Glu(OcHx), His(DNP), Lys(2-ClZ), Ser(Bzl), Thr(Bzl), and Tyr(2-BrZ). Proteo amino acids (2.2 mmol) were coupled to the resin using 2.0 mmol of HCTU (4 mL of a 0.5 M solution) and ∼4.0 mmol of DIEA (700 μL) for 20 min. Deuterated amino acids (0.6 mmol) were coupled for 1 h using 0.55 mmol of HCTU (1.1 mL of a 0.5 M solution) and ∼1.6 mmol of DIEA (280 μL). Flow washes between coupling and deprotection and between deprotection and coupling were performed with DMF except for Gln, for which an additional DCM flow wash was performed before and after Boc deprotection. To facilitate synthesis of the N-terminal peptide fragment with a C-terminal thioester, thioester resin was generated from Boc-Leu-Pam resin before coupling the first amino acid using the method developed by Dawson and coworkers.30 Briefly, 0.6 mmol of S-trityl-β-mercaptopropanoic acid was coupled with 0.4 mmol of Leu-Pam resin for 1 h using 0.55 mmol of HCTU and ∼1.6 mmol of DIEA, and the trityl group was removed by two 1 min incubations with a 95% TFA/ 2.5% triisopropylsilane/2.5% H2O (v/v/v) mixture. Peptides were cleaved from the resin and side-chain-deprotected via a 1 h treatment at 0 °C with anhydrous HF containing 10% anisole using a type I HF apparatus without a manometer manufactured by Peptide Institute, Inc. (Osaka, Japan). Cleaved peptides were precipitated with chilled diethyl ether, dissolved in 10% (v/v) aqueous AcOH, and lyophilized using a VirTis BenchTop Pro instrument (SP Scientific, Warminster, PA). Peptides were then purified or analyzed by HPLC using a Dynamax Solvent Delivery System SD-200 with a Dynamax Absorbance Detector UV-D II (Rainin) using a linear gradient of solvent A [0.1% (v/v) aqueous TFA] and solvent B [90% (v/v) acetonitrile, 10% (v/v) water, and 0.1% (v/v) TFA]. Lyophilized crude peptide fragments were purified by preparative HPLC on a Jupiter C18 reverse-phase column (10 μm particle size, 300 Å pore size, 250 mm length × 21.2 mm diameter, manufacturerd by Phenomenex, Torrence, CA) with a solvent gradient from 30 to 55% B over 50 min for the N-terminal peptide and from 25 to 55% B over 60 min for the C-terminal peptide with a constant flow rate 15 mL/min (monitoring A230). Pure fractions of peptide fragments were
Figure 1. Structure of the Bcd homeodomain−DNA complex with deuterated side chains shown (carbon atoms are colored orange, and the nitrogen of Lys50 is colored blue) (Protein Data Bank entry 1ZQ3).
possibilities requires an experimental method with high temporal and residue-specific resolution. Infrared (IR) spectroscopy is a valuable tool for the characterization of any molecule because the frequency, line shape, and number of unique absorptions of a given bond vibration can be related to the number and nature of environments the bond experiences. Unfortunately, the characterization of proteins by IR spectroscopy is challenging because the large number of overlapping absorptions precludes the observation and characterization of any single absorption. To circumvent this challenge, we have developed carbon− deuterium (C−D) bonds as spectroscopic probes that are incorporated site-specifically in place of nonexchangable C−H bonds in protein backbones and side chains.20 The isotopic substitution of D for H does not significantly alter the protein, but it does shift the stretching absorption into a region of the IR spectrum that is largely free of other absorptions and thus allows its largely background free characterization. We have used C−D bonds to characterize individual residues within the folded states and during unfolding of a variety of proteins, including cytochrome c,21,22 dihydrofolate reductase,23,24 and an SH3 domain.25−28 The inherently fast time scale of IR spectroscopy is of particular utility for the characterization of conformational heterogeneity, as even the fastest interconverting species are likely to be resolved.20,29 We now report an initial exploration of C−D bonds as probes of the Bcd homeodomain. In the free and DNA-bound forms, we find that residues within the first two helices show absorptions that are similar to those of the corresponding free amino acid, although they are narrowed and shifted because of their specific protein environments. In contrast, residues of the third helix, including Val45 and Lys50, show absorptions that are more complex than the corresponding free amino acids and are consistent with the presence of conformational heterogeneity. Moreover, while DNA binding induces a significant change in the frequency and line widths of the absorptions at B
DOI: 10.1021/acs.biochem.7b00255 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry
oligonucleotides in an equimolar ratio and then annealing them by slowly cooling them from 95 °C to room temperature in 10 mM sodium phosphate (pH 7.3). The resulting duplex DNA was desalted using a NAP-10 or NAP-25 column (GE Healthcare Life Sciences), eluted with deionized water, and quantified by UV−vis absorption. Aliquots containing 22 nmol of DNA were prepared; the solvent was removed with a Speedvac concentrator (Savant SC110A Plus with UVS400 SpeedVac System), and the dried duplexes were stored at 4 °C until they were used. Bcd−DNA complexes were prepared by combining DNA with protein in a 1:1.1 molar ratio, followed by dialysis in 10 mM sodium phosphate buffer (pH 7.3) and concentration in the dialysis bag (see above), before being stored at 4 °C until they were used. Circular Dichroism Spectroscopy. Far-UV circular dichroism spectra were recorded on an AVIV 62DS spectrometer with a 1 nm step size using a 2 mm path-length quartz sample cell. Each scan was performed with an integration time of 1 s at each step, and a total of two scans were averaged for each spectrum at 20 °C. Spectra were recorded with each deuterated Bcd homeodomain at 70 μM. FT IR Spectroscopy. Approximately 8 μL of 2 mM Bcd, free or bound to DNA, was loaded into a temperaturecontrollable demountable liquid cell with CaF2 windows (Harrick Scientific Products, Inc., Pleasantville, NY) and a 75 μm Teflon spacer, and spectra were recorded using a Bruker Equinox 55 FT IR spectrometer equipped with a liquid N2cooled mercury cadmium telluride (MCT) detector. A total of 8000 scans were averaged for each temperature with a resolution of 4 cm−1. For melting experiments, a series of FT IR spectra were collected in 5 °C intervals between 24 and 76 °C for free protein and between 24 and 93 °C for the DNAbound protein, but with 2.5 °C intervals used within the transition region (Supporting Information). At each temperature, the samples were equilibrated for 10 min before FT IR spectra were recorded. The temperature of the sample cell was controlled, and the FT IR spectrometer was automated using Labview (National Instruments Corp.). Proteo samples were collected under identical conditions and their spectra subtracted from those of a corresponding deutero sample in the region of 1900−2400 cm−1 using OPUS software (Bruker). The resulting spectra were further background corrected using a higher-order polynomial function and fit with pseudo-Voigt functions using Matlab (Mathworks, Inc.).26 The pseudo-Voigt function, I(ν), is defined as
analyzed by MS using a PE SCIEX API 2000 LC/MS/MS system (AB SCIEX), combined, and lyophilized. The N- and C-terminal peptide fragments were joined using native chemical ligation.31 Briefly, peptides were dissolved at a concentration of ∼5 mM in 200 mM sodium phosphate buffer (pH 8.0) containing 6 M guanidine hydrochloride and ligated using 1−2% thiophenol at 37 °C for 24 h. To facilitate complete removal of the DNP group from the ligated peptide, the reaction mixture was centrifuged and filtered through a 0.22 μm pore size filter (Millipore), and the filtrate was diluted with 100 mM Tris buffer (pH 8.5) containing 6 M guanidine hydrochloride and treated with 20% β-mercaptoethanol for 45 min at 37 °C. The ligated peptides were purified by preparative HPLC using a C18 reverse-phase column with a solvent gradient from 25 to 55% B over 40 min with a constant flow rate of 15 mL/min (monitoring A230). Pure ligated fractions were confirmed by MS, combined, and lyophilized. The ligation product contains an A32C mutation, which was converted to the fully native protein via desulfurization.32 Briefly, fresh nickel boride was generated from the reaction of aqueous nickel acetate with NaBH4. The lyophilized pure ligated peptide was dissolved in 20% (v/v) aqueous AcOH and mixed with freshly prepared nickel boride while being stirred at room temperature overnight. The desulfurized peptide was further purified by preparative HPLC with a solvent gradient from 25 to 55% B over 50 min. Pure desulfurized peptides were analyzed by MS, combined, and lyophilized. The purity of peptides was further assessed by analytical HPLC using a Jupiter C12 reverse-phase column (4 μm particle size, 90 Å pore size, 150 mm length × 4.6 mm diameter, manufactured by Phenomenex) with a solvent gradient from 20 to 70% B over 40 min and a constant flow rate of 1.5 mL/min (monitoring A220), as well as MS (Supporting Information). Sample Preparation. Lyophilized pure protein was dissolved in water at a concentration of ∼20 mg/mL. The aqueous protein solution was dialyzed using a Micro Float-ALyzer dialysis device (0.5−1.0 kDa molecular weight cutoff) (Spectrum Laboratories Inc., Rancho Dominguez, CA) in 10 mM sodium phosphate buffer (pH 7.3). Following dialysis, the device was buried in polyethylene glycol 8000, resulting in concentration to 2.5−3.5 mM. Protein samples were then diluted to 2.0 mM with 10 mM sodium phosphate buffer (pH 7.3) for FT IR measurements. The protein concentration was determined by ultraviolet−visible (UV−vis) absorption (ε280 = 6990 M−1 cm−1)33 using a Hewlett-Packard 8453 spectrophotometer. Concentrated protein samples were stored at 4 °C until they were used. For the characterization of the complex with DNA, a 13-mer DNA duplex that contains the Bcd consensus sequence and was used in previous studies of the homeodomain11 was employed (5′-GCTCT AATCC CCG-3′/5′-CGGG GATTAG AGC-3′), while to prevent melting during the thermal denaturation experiments, an analogous but longer (53-mer) duplex that encompasses the 13-mer used above and whose sequence was taken from the 3L chromosome of Drosophila melanogaster (GenBank entry AE014296.5; 5′-ACGGA GTTGC GCTGA TTGTA GCTCT AATCC CCGAA TCCTG ATCCC AGATC CCG-3′/5′-CGGGA TCTGG GATCA GGATT CGGGG ATTAG AGCTA CAATC AGCGC AACTC CGT3′, sequence corresponding to the 13-mer is underlined) was employed, with all oligonucleotides obtained from Integrated DNA Technologies (San Diego, CA). The 13-mer and 53-mer duplexes were prepared by first combining the two constituent
I (ν ) =
⎛ A fwhm 2 ⎜m m + (1 − m) π ln 2 ⎝ 4(ν − ν0)2 + fwhm 2
⎛ ⎛ ν − ν0 ⎞2 ⎞⎞ ⎟ ⎟⎟ + (1 − m) π ln 2 exp⎜ − 4(ln2)⎜ ⎝ fwhm ⎠ ⎠⎠ ⎝
(1)
where A is the amplitude at center frequency ν0, fwhm is the full width at half-maximum of the pseudo-Voigt profile, and m describes the Gaussian/Lorentzian character of the profile (when m = 0, the absorption band is purely Gaussian, and when m = 1, the absorption band is purely Lorentzian). Spectral deconvolutions into pseudo-Voigt functions for the free and DNA-bound homeodomain at 24 °C are provided in the Supporting Information, and fit parameters are listed in Table 1. Data are reported from at least triplicate runs with three independently prepared samples, and errors are standard deviations. To determine the melting temperature (Tm) at C
DOI: 10.1021/acs.biochem.7b00255 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry Table 1. Spectral Fit Parameters of C−D Labeled Free Amino Acids and the Bcd Homeodomain Free or Bound to DNAa symmetric stretch
asymmetric stretch
ν0 (cm−1)
fwhm (cm−1)
ν0(cm−1)
fwhm (cm−1)
(d7)Leu (d7)Leu16 (d7)Leu16− DNA (d7)Leu21 (d7)Leu21− DNA (d7)Leu31 (d7)Leu31− DNA (d8)Val (d7)Val45
2068.8 ± 0.1 2061.0 ± 0.9 2060.1 ± 0.2
12.2 ± 0.3 9.4 ± 1.0 8.2 ± 0.8
2223.4 ± 0.1 2212.5 ± 0.4 2211.9 ± 0.1
19.7 ± 0.2 15.1 ± 0.7 14.0 ± 0.1
2066.3 ± 1.0 2065.6 ± 1.3
11.4 ± 1.5 10.5 ± 1.1
2217.6 ± 0.4 2218.8 ± 0.2
17.9 ± 0.6 20.1 ± 0.7
2066.3 ± 0.6 2065.2 ± 0.4
8.2 ± 1.3 7.5 ± 0.9
2219.9 ± 0.7 2220.1 ± 0.4
17.9 ± 0.9 19.0 ± 0.3
2074.9 ± 0.1 2070.4 ± 0.7
12.8 ± 0.1 34.2 ± 0.7
(d7)Val45− DNA
2066.0 ± 0.5
21.7 ± 1.7
(d8)Lys (d8)Lys50
2112.5 ± 0.2 2107.8 ± 1.0
25.4 ± 0.4 17.5 ± 2.5
(d8)Lys50− DNA
2103.4 ± 0.7
12.8 ± 0.8
2113.6 ± 0.6
9.8 ± 0.2
2230.0 2229.9 2216.1 2229.3
± ± ± ±
0.1 1.7 1.3 0.9
17.9 17.0 20.0 17.5
± ± ± ±
0.2 1.3 1.6 1.4
2214.5 2223.4 2217.0 2119.7 2217.5
± ± ± ± ±
0.9 0.2 0.3 1.4 0.9
17.9 26.0 29.6 13.1 35.7
± ± ± ± ±
2.0 0.6 1.1 0.8 1.2
a
The m values for all absorptions were Lys) homeodomain-DNA complex at 1.9 A resolution: structural basis for enhanced affinity and altered specificity. Structure 5, 1047−1054.
(9) Treisman, J., Harris, E., Wilson, D., and Desplan, C. (1992) The homeodomain: a new face for the helix-turn-helix? BioEssays 14, 145− 150. (10) Aravind, L., Anantharaman, V., Balaji, S., Babu, M. M., and Iyer, L. M. (2005) The many faces of the helix-turn-helix domain: transcription regulation and beyond. FEMS Microbiol. Rev. 29, 231− 262. (11) Baird-Titus, J. M., Clark-Baldwin, K., Dave, V., Caperelli, C. A., Ma, J., and Rance, M. (2006) The solution structure of the native K50 Bicoid homeodomain bound to the consensus TAATCC DNAbinding site. J. Mol. Biol. 356, 1137−1151. (12) Dave, V., Zhao, C., Yang, F., Tung, C. S., and Ma, J. (2000) Reprogrammable recognition codes in bicoid homeodomain-DNA interaction. Mol. Cell. Biol. 20, 7673−7684. (13) Zhao, C., Dave, V., Yang, F., Scarborough, T., and Ma, J. (2000) Target selectivity of bicoid is dependent on nonconsensus site recognition and protein-protein interaction. Mol. Cell. Biol. 20, 8112− 8123. (14) Rivera-Pomar, R., Niessing, D., Schmidt-Ott, U., Gehring, W. J., and Jackle, H. (1996) RNA binding and translational suppression by bicoid. Nature 379, 746−749. (15) Chan, S. K., and Struhl, G. (1997) Sequence-specific RNA binding by bicoid. Nature 388, 634. (16) Hanes, S. D., and Brent, R. (1989) DNA specificity of the bicoid activator protein is determined by homeodomain recognition helix residue 9. Cell 57, 1275−1283. (17) Hanes, S. D., and Brent, R. (1991) A genetic model for interaction of the homeodomain recognition helix with DNA. Science 251, 426−430. (18) Trelsman, J., Gonczy, P., Vashishtha, M., Harris, E., and Desplan, C. (1989) A single amino acid can determine the DNA binding specificity of homeodomain proteins. Cell 59, 553−562. (19) Spolar, R. S., and Record, M. T., Jr. (1994) Coupling of local folding to site-specific binding of proteins to DNA. Science 263, 777− 784. (20) Adhikary, R., Zimmermann, J., and Romesberg, F. E. (2017) Transparent window vibrational probes for the characterization of proteins with high structural and temporal resolution. Chem. Rev. 117, 1927−1969. (21) Sagle, L. B., Zimmermann, J., Matsuda, S., Dawson, P. E., and Romesberg, F. E. (2006) Redox-coupled dynamics and folding in cytochrome c. J. Am. Chem. Soc. 128, 7909−7915. (22) Sagle, L. B., Zimmermann, J., Dawson, P. E., and Romesberg, F. E. (2006) Direct and high resolution characterization of cytochrome c equilibrium folding. J. Am. Chem. Soc. 128, 14232−14233. (23) Groff, D., Thielges, M. C., Cellitti, S., Schultz, P. G., and Romesberg, F. E. (2009) Efforts toward the direct experimental characterization of enzyme microenvironments: tyrosine100 in dihydrofolate reductase. Angew. Chem., Int. Ed. 48, 3478−3481. (24) Thielges, M. C., Case, D. A., and Romesberg, F. E. (2008) Carbon-deuterium bonds as probes of dihydrofolate reductase. J. Am. Chem. Soc. 130, 6597−6603. (25) Adhikary, R., Zimmermann, J., Dawson, P. E., and Romesberg, F. E. (2014) IR probes of protein microenvironments: utility and potential for perturbation. ChemPhysChem 15, 849−853. (26) Adhikary, R., Zimmermann, J., Liu, J., Dawson, P. E., and Romesberg, F. E. (2013) Experimental characterization of electrostatic and conformational heterogeneity in an SH3 domain. J. Phys. Chem. B 117, 13082−13089. (27) Adhikary, R., Zimmermann, J., Liu, J., Forrest, R. P., Janicki, T. D., Dawson, P. E., Corcelli, S. A., and Romesberg, F. E. (2014) Evidence of an unusual N-H···N hydrogen bond in proteins. J. Am. Chem. Soc. 136, 13474−13477. (28) Cremeens, M. E., Zimmermann, J., Yu, W., Dawson, P. E., and Romesberg, F. E. (2009) Direct observation of structural heterogeneity in a β-sheet. J. Am. Chem. Soc. 131, 5726−5727. (29) Horness, R. E., Basom, E. J., Mayer, J. P., and Thielges, M. C. (2016) Resolution of site-specific conformational heterogeneity in G
DOI: 10.1021/acs.biochem.7b00255 Biochemistry XXXX, XXX, XXX−XXX
