research papers\(\def\hfill{\hskip 5em}\def\hfil{\hskip 3em}\def\eqno#1{\hfil {#1}}\)

Journal logoSTRUCTURAL
BIOLOGY
ISSN: 2059-7983

Structural basis of DNA binding by YdaT, a functional equivalent of the CII repressor in the cryptic prophage CP-933P from Escherichia coli O157:H7

crossmark logo

aStructural Biology Brussels, Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussel, Belgium, bVIB–VUB Center for Structural Biology, Vlaams Instituut voor Biotechnologie, Pleinlaan 2, 1050 Brussel, Belgium, cJean Jeener NMR Center, Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussel, Belgium, dDepartment of Physical Chemistry, Faculty of Chemistry and Chemical Technology, University of Ljubljana, Večna pot 113, 1000 Ljubljana, Slovenia, eDepartment of Chemistry, Universiteit Antwerpen, Groenenborgerlaan 171, 2020 Antwerpen, Belgium, and fResearch Group of Microbiology, Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussel, Belgium
*Correspondence e-mail: remy.loris@vub.be

Edited by A. Gonzalez, Lund University, Sweden (Received 26 December 2022; accepted 10 February 2023; online 27 February 2023)

YdaT is a functional equivalent of the CII repressor in certain lambdoid phages and prophages. YdaT from the cryptic prophage CP-933P in the genome of Escherichia coli O157:H7 is functional as a DNA-binding protein and recognizes a 5′-TTGATTN6AATCAA-3′ inverted repeat. The DNA-binding domain is a helix–turn–helix (HTH)-containing POU domain and is followed by a long α-helix (α6) that forms an antiparallel four-helix bundle, creating a tetramer. The loop between helix α2 and the recognition helix α3 in the HTH motif is unusually long compared with typical HTH motifs, and is highly variable in sequence and length within the YdaT family. The POU domains have a large degree of freedom to move relative to the helix bundle in the free structure, but their orientation becomes fixed upon DNA binding.

1. Introduction

Since the early days of molecular biology, Escherichia coli bacteriophage λ has served as a model organism and as a tool in molecular genetics. The `lysis versus lysogeny' decision of the phage has been studied in great detail and serves as a paradigm for regulatory gene circuits (Wegrzyn & Wegrzyn, 2005[Wegrzyn, G. & Wegrzyn, A. (2005). Prog. Nucleic Acid Res. Mol. Biol. 79, 1-48.]). The circuit, which depends on three transcription factors (CI, CII and Cro), results in bistability. The interplay between CI and Cro determines whether a λ prophage remains inserted in the host chromosome or start a lytic cycle (Ptashne et al., 1980[Ptashne, M., Jeffrey, A., Johnson, A. D., Maurer, R., Meyer, B. J., Pabo, C. O., Roberts, T. M. & Sauer, R. T. (1980). Cell, 19, 1-11.]; Johnson et al., 1981[Johnson, A. D., Poteete, A. R., Lauer, G., Sauer, R. T., Ackers, G. K. & Ptashne, M. (1981). Nature, 294, 217-223.]). CII, on the other hand, is essential in order to enter the lysogenic path immediately after infection (Chung & Echols, 1977[Chung, S. & Echols, H. (1977). Virology, 79, 312-319.]). CII is expressed after infection of E. coli with λ, with its expression level depending on both cellular and environmental factors that affect the half-life of this rather unstable protein (Kobiler et al., 2002[Kobiler, O., Koby, S., Teff, D., Court, D. & Oppenheim, A. B. (2002). Proc. Natl Acad. Sci. USA, 99, 14964-14969.]). When its expression level crosses a certain threshold, it directs the lysis/lysogeny decision towards lysogeny. CII activates three promotors on the phage genome: PI, PRE and PAQ. From PAQ an antisense RNA is transcribed that prevents production of the Q protein to reduce lytic activity until the repressor CI is sufficiently expressed (Ho & Rosenberg, 1985[Ho, Y. S. & Rosenberg, M. (1985). J. Biol. Chem. 260, 11838-11844.]). CI is transcribed from PRE, resulting in the shutdown of all lytic genes. Finally, from the PI promotor, the Int protein is produced that integrates the phage genome into the E. coli chromosome (Court et al., 2007[Court, D. L., Oppenheim, A. B. & Adhya, S. L. (2007). J. Bacteriol. 189, 298-304.]). Once lysogeny has been established, CII is no longer required and remains turned off.

Lambdoid phages are a family of bacteriophages related to coliphage λ, with which they can form viable recombinants. Most lambdoid phages grow on E. coli, but a few such as P22 come from Salmonella typhimurium. They are highly polymorphic in DNA sequence and biological specificity, with differences observed in receptor specificity, integration and the mechanism of packaging (Campbell, 1994[Campbell, A. (1994). Annu. Rev. Microbiol. 48, 193-222.]). Different E. coli strains typically contain several cryptic lambdoid phages in their genomes. For example, E. coli O157:H7 Sakai contains 18 such prophage genome elements that together constitute about 16% of its total genomic DNA content (Brüssow & Kutter, 2004[Brüssow, H. & Kutter, E. (2004). Bacteriophages: Biology and Applications, edited by E. Kutter & A. Sulakvelidze, pp. 91-128. Boca Raton: CRC Press.]).

Among the different defective prophages (that are missing one or more crucial genes to allow initiation of a lytic cycle) present in the genome of E. coli O157:H7 is CP-933P (Perna et al., 2001[Perna, N. T., Plunkett, G., Burland, V., Mau, B., Glasner, J. D., Rose, D. J., Mayhew, G. F., Evans, P. S., Gregor, J., Kirkpatrick, H. A., Pósfai, G., Hackett, J., Klink, S., Boutin, A., Shao, Y., Miller, L., Grotbeck, E. J., Davis, N. W., Lim, A., Dimalanta, E. T., Potamousis, K. D., Apodaca, J., Anantharaman, T. S., Lin, J., Yen, G., Schwartz, D. C., Welch, R. A. & Blattner, F. R. (2001). Nature, 409, 529-533.]). It contains a three-component version of the parDE type of toxin–antitoxin module termed paaR2–paaA2–parE2 next to the ydaST gene pair (Hallez et al., 2010[Hallez, R., Geeraerts, D., Sterckx, Y., Mine, N., Loris, R. & Van Melderen, L. (2010). Mol. Microbiol. 76, 719-732.]). The ydaS and ydaT genes were initially suspected (from an analysis using the RASTA-Bacteria algorithm) to constitute a toxin–antitoxin operon in E. coli O157:H7, with YdaT supposedly being the toxin (Sevin & Barloy-Hubler, 2007[Sevin, E. W. & Barloy-Hubler, F. (2007). Genome Biol. 8, R155.]). Experimental evidence nevertheless failed to prove this hypothesis (Christensen-Dalsgaard et al., 2010[Christensen-Dalsgaard, M., Jørgensen, M. G. & Gerdes, K. (2010). Mol. Microbiol. 75, 333-348.]). Indeed, their genetic context pinpoints ydaS and ydaT as equivalents of cro and cII, respectively (Casjens, 2003[Casjens, S. (2003). Mol. Microbiol. 49, 277-300.]; Jobling, 2018[Jobling, M. G. (2018). mSphere, 3, e00163-18.]; Jurėnas et al., 2021[Jurėnas, D., Fraikin, N., Goormaghtigh, F., De Bruyn, P., Vandervelde, A., Zedek, S., Jové, T., Charlier, D., Loris, R. & Van Melderen, L. (2021). mBio, 12, e02947-21.]). Homologs are found in a number of rac prophages, where their expression is repressed by RacR (Krishnamurthi et al., 2017[Krishnamurthi, R., Ghosh, S., Khedkar, S. & Seshasayee, A. S. N. (2017). mSphere, 2, e00392-17.]). Rac phages are a group of defective prophages that are found in various E. coli strains, with Rac itself being the first defective prophage to be discovered in E. coli K-12 (Kaiser & Murray, 1979[Kaiser, K. & Murray, N. E. (1979). Mol. Gen. Genet. 175, 159-174.]). The CI repressor is here typically referred to as `RacR'.

A schematic comparison between the immunity regions of bacteriophage λ and CP-933P is given in Fig. 1[link]. Similar to λ CII, which binds the PRE promotor located between the λ cro and cII genes and activates transcription from PRE, YdaT is predicted to bind at the interface between the ydaS and ydaT genes. This region contains the PRE993P promotor, the CP-933P equivalent of λ PRE. Although basal transcriptional activity from PRE993P is very weak, overexpression of YdaT increases this activity significantly (Jurėnas et al., 2021[Jurėnas, D., Fraikin, N., Goormaghtigh, F., De Bruyn, P., Vandervelde, A., Zedek, S., Jové, T., Charlier, D., Loris, R. & Van Melderen, L. (2021). mBio, 12, e02947-21.]).

[Figure 1]
Figure 1
Comparison of the immunity regions of bacteriophage λ and E. coli O157:H7 prophage CP-933P. Genes are shown as arrows, with the direction of the arrow indicating the direction of transcription. Promoters of λ and their CP-933P equivalents are indicated. The three repressors are coloured grey, while other neighbouring phage genes are white. The toxin–antitoxin genes parE2 and paaA2 in CP-933P are coloured black and replace rexB and rexA, respectively, in λ. The border sequence of the interface between the ydaS and ydaT genes is highlighted and the three YdaT binding sites OL, OM and OR are boxed in grey and labelled. The OM sequence also contains an alternative transcription start for the paaR2–paaA2–parE2 operon that is controlled via YdaT (Jurėnas et al., 2021[Jurėnas, D., Fraikin, N., Goormaghtigh, F., De Bruyn, P., Vandervelde, A., Zedek, S., Jové, T., Charlier, D., Loris, R. & Van Melderen, L. (2021). mBio, 12, e02947-21.]).

Although both YdaS and YdaT are predicted to contain a helix–turn–helix motif and serve a similar function as the λ Cro and CII proteins, respectively (Jurėnas et al., 2021[Jurėnas, D., Fraikin, N., Goormaghtigh, F., De Bruyn, P., Vandervelde, A., Zedek, S., Jové, T., Charlier, D., Loris, R. & Van Melderen, L. (2021). mBio, 12, e02947-21.]), the YdaS and YdaT proteins show no detectable sequence similarity to λ Cro or CII. YdaT proteins constitute a family of transcription factors that currently remain uncharacterized in terms of structure and DNA-binding activity. In order to better understand how YdaT functions at the molecular level, we determined the crystal structure of CP-933P YdaT and identified its exact binding site as three regions, OL, OM and OR, between the ydaS and ydaT genes. Of these, OM covers the alternative transcription start for the paaR2–paaA2–parE2 operon as identified by Jurėnas et al. (2021[Jurėnas, D., Fraikin, N., Goormaghtigh, F., De Bruyn, P., Vandervelde, A., Zedek, S., Jové, T., Charlier, D., Loris, R. & Van Melderen, L. (2021). mBio, 12, e02947-21.]). We furthermore created and validated a model for the interaction between YdaT and its operator. Together, our results paint a consistent picture of the functioning of YdaT repressors.

2. Materials and methods

2.1. Cloning, expression and purification

Plasmid pET-28b containing the open reading frame for YdaT from E. coli O157:H7 (UniProt ID A0A6M7H0F8) with an N-terminal His tag (GSSHHHHHHSSG) was transformed into competent E. coli BL21 (DE3) cells (Table 1[link]). Transformed cells were plated on agar plates supplemented with kanamycin (25 µg ml−1) and incubated at 37°C overnight. LB medium supplemented with kanamycin (25 µg ml−1) was inoculated with one colony and left to incubate overnight at 37°C while shaking at 130 rev min−1. 5 ml of the overnight culture was added to 500 ml LB medium (supplemented with 25 µg ml−1 kanamycin) in 2 l flasks and incubated at 37°C with shaking at 130 rev min−1. When the OD600 reached 0.6–0.8, protein expression was induced with 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG). Upon induction, the cultures were incubated at 37°C for 4 h, centrifuged at 5000 rev min−1 for 15 min, resuspended in lysis buffer [20 mM Tris–HCl, 500 mM NaCl, 20 mM MgCl2 pH 8.0, 0.1 mg ml−1 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride (AEBSF)] and stored at −80°C. To purify the protein, the frozen cells were thawed and DNase I was added (50 µg ml−1). The cells were lysed by sonication (5 min, 5 s on and 5 s off, 70% amplitude) and the lysate was centrifuged at 18 000 rev min−1 for 45 min. The supernatant was filtered (0.45 µm HAWP filter) and loaded onto a pre-packed HisTrap HP Ni2+-Sepharose column (GE Healthcare) pre-equilibrated in 20 mM Tris–HCl, 500 mM NaCl, 5 mM imidazole pH 8.0. The column was then washed with the same buffer until baseline stabilization. A linear gradient (0.0–1.0 M imidazole in 20 column volumes) elution was applied in 20 mM Tris–HCl, 500 mM NaCl, 1 M imidazole pH 8.0. Fractions containing the protein were concentrated and loaded onto a Superdex 200 16/90 size-exclusion chromatography (SEC) column (GE Healthcare) pre-equilibrated in 20 mM Tris–HCl, 500 mM NaCl pH 8.0. The purity of the protein was assessed by SDS–PAGE.

Table 1
Macromolecule-production information

Protein Wild-type YdaT YdaT1–96
Source organism Escherichia coli O157:H7 Escherichia coli O157:H7
DNA source Commercial gene synthesis by GenScript Commercial gene synthesis by GenScript
Cloning vector pET-28b pET-28b
Expression vector pET-28b pET-28b
Expression host E. coli BL21 (DE3) E. coli BL21 (DE3)
Complete amino-acid sequence of the construct produced MGSSHHHHHHSSGENLYFQGASKIKHEHIRMAMNVWAHPDGEKVPAAKITKAYFELGMTFPELYDDSHPEALARNTQKIFRWLDKDTPDAVEKMQALLPAIEKAMPPLLVARMRSHSSEYYREIVERRDRLVKDVDDFVASAVVLYDQMNRGGPAGNAVVMH MGSSHHHHHHSSGENLYFQGASKIKHEHIRMAMNVWAHPDGEKVPAAKITKAYFELGMTFPELYDDSHPEALARNTQKIFRWLDKDTPDAVEKMQALLPAIEKAMPPLLVARMRSHS
Molecular mass from chemical composition 18472.02 13348.27
Extinction coefficient (M−1 cm−1) 19940 15470

A truncated variant of YdaT lacking the 45 C-terminal residues (YdaT1–96) was created by replacing Ser97 with a stop codon (TAA) via PCR amplification of the whole plasmid using primers 1 and 2 (Table 1[link], Supplementary Table S1) and Q5 High-Fidelity 2× Master Mix (NEB). The unmodified plasmid was degraded by incubation with DpnI for 1 h at 37°C. The mutations were confirmed by sequencing and CaCl2-competent E. coli BL21 Star (DE3) cells were transformed with the mutated plasmids. Both proteins were expressed and purified as described for wild-type YdaT except for the use of a Superdex 75 16/60 SEC column (GE Healthcare) for the SEC purification step.

2.2. Concentration determination

The concentrations of the samples were determined by measuring the UV absorbance at 280 nm for the proteins and at 260 nm for the oligonucleotides. Extinction coefficients for the proteins were calculated using ProtParam (Gasteiger et al., 2005[Gasteiger, E., Hoogland, C., Gattiker, A., Duvaud, S., Wilkins, M. R., Appel, R. D. & Bairoch, A. (2005). The Proteomics Protocols Handbook, edited by J. M. Walker, pp. 571-607. Totowa: Humana Press.]; https://web.expasy.org/protparam/) and are 19 940 M−1 cm−1 for YdaT and 15 470 M−1 cm−1 for YdaT1–96. For DNA, extinction coefficients were calculated according to Tataurov et al. (2008[Tataurov, A. V., You, Y. & Owczarzy, R. (2008). Biophys. Chem. 133, 66-70.]).

2.3. X-ray crystallography

Crystals of YdaT were produced by a sitting-drop method using three-well Intelli-Plates and a Mosquito robot. 100 nl YdaT protein (13.23 mg ml−1 in 20 mM Tris, 200 mM NaCl pH 7.5) was mixed with 100 nl reservoir solution and equilibrated against 70 µl reservoir solution from various commercial crystallization kits at 19°C. Crystals grew in 0.1 M bis-Tris pH 5.5, 2.0 M ammonium sulfate. The crystals were harvested, dipped in cryoprotectant solution [0.1 M bis-Tris pH 5.5, 2.0 M ammonium sulfate, 22.5%(v/v) glycerol] and vitrified in liquid nitrogen.

Diffraction data were collected on beamline PROXIMA-1 at the SOLEIL synchrotron, Gif-sur-Yvette, France and were recorded on an EIGER X 9M photon-counting area detector. The data were indexed, scaled and merged using XDS (Kabsch, 2010[Kabsch, W. (2010). Acta Cryst. D66, 125-132.]). The data were further corrected for anisotropy with the STARANISO server (Tickle et al., 2018[Tickle, I. J., Flensburg, C., Keller, P., Paciorek, W., Sharff, A., Vonrhein, C. & Bricogne, G. (2018). STARANISO. Global Phasing Ltd, Cambridge, United Kingdom.]) using the `unmerged data' protocol. The data were limited to 240° total crystal rotation due to radiation damage.

The structure of YdaT was determined by molecular replacement using Phaser as implemented in the CCP4 package (McCoy et al., 2007[McCoy, A. J., Grosse-Kunstleve, R. W., Adams, P. D., Winn, M. D., Storoni, L. C. & Read, R. J. (2007). J. Appl. Cryst. 40, 658-674.]). The crystal structure of PDB entry 3c4r (sequence identity of 84%; annotated as an uncharacterized protein from a cryptic prophage in E. coli O6:H1; New York SGX Research Center for Structural Genomics, unpublished work) was used as the search model. The initial solution was refined with phenix.refine using a maximum-likelihood target against intensities (Afonine et al., 2012[Afonine, P. V., Grosse-Kunstleve, R. W., Echols, N., Headd, J. J., Moriarty, N. W., Mustyakimov, M., Terwilliger, T. C., Urzhumtsev, A., Zwart, P. H. & Adams, P. D. (2012). Acta Cryst. D68, 352-367.]) and was rebuilt with Coot (Emsley et al., 2010[Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. (2010). Acta Cryst. D66, 486-501.]). NCS restraints were applied throughout the refinement procedure and the final cycles involved refinement of TLS parameters. Data-collection and refinement statistics are summarized in Table 2[link] and full details are given in Supplementary Tables S2 and S3.

Table 2
Crystallographic data collection and refinement

Values in parentheses are for the outer shell. XDS was used for data collection and scaling.

Space group P1
a, b, c (Å) 45.57, 64.26, 63.91
α, β, γ (°) 66.83, 89.36, 77.67
Resolution range (Å) 44.37–2.40 (2.52–2.40)
Completeness (spherical) (%) 77.4 (32.5)
Rmerge 0.082 (0.386)
Rmeas 0.106 (0.552)
CC1/2 0.990 (0.797)
No. of reflections, working set 18719
No. of reflections, test set 979
Final Rcryst 0.2112 (0.2757)
Final Rfree 0.2658 (0.2783)
R.m.s. deviations
 Bond lengths (Å) 0.0088
 Bond angles (°) 0.959
Ramachandran plot
 Most favoured (%) 97.4
 Additionally allowed (%) 2.0
 Disallowed (%) 0.6
PDB code 8bt1

2.4. Circular-dichroism spectroscopy

Measurements were performed on a Jasco J-750 spectrophotometer (Jasco, Japan). Spectra were collected from 200 to 250 nm every 1 nm, with a bandwith of 1 nm, a scan rate of 50 nm min−1 and five acquisitions, in a quartz cuvette (1 mm optical path length). The protein concentration for all samples was 0.2 mg ml−1. Data were normalized to obtain the mean residue ellipticity ([θ] in deg cm2 dmol−1) using the formula

[[\theta] = {{\theta} \over {cl}}, \eqno (1)]

where θ is the ellipticity (raw data), c is the molar concentration and l is the optical path length. The buffer spectrum was subtracted from the sample spectra. Protein solutions were prepared in 20 mM Tris–HCl, 500 mM NaCl pH 8.0 buffer.

2.5. Size-exclusion chromatography and multi-angle light scattering (SEC-MALS)

YdaT was dialysed overnight against 20 mM Tris–HCl, 200 mM NaCl pH 7.5. The buffer was filtered through a 0.1 µm filter (Sartorius) three times before measurements. Protein samples were centrifuged at 17 000g for 10 min prior to injection. 20 µl of protein sample at a concentration of 1 or 10 mg ml−1 was injected into a Shodex KW402.5-4F column (Showa Denko K. K.) that had been pre-equilibrated with dialysis buffer on an Alliance e2695 XE HPLC System (Waters) connected to a TREOS II light-scattering detector (Wyatt Technology) and a Shodex RI-501 refractive-index detector (Showa Denko K. K.). The flow rate was set to 0.2 ml min−1. Data processing and molecular-weight calculations were performed using the ASTRA V software (Wyatt Technology). Bovine serum albumin at a concentration of 1 mg ml−1 was used as a standard for calibration.

2.6. Native mass spectrometry

YdaT was buffer-exchanged into 200 mM ammonium acetate pH 8 using Amicon Ultra 0.5 ml centrifugal filters (Merck Millipore) with a molecular-weight cutoff of 3 kDa. The concentrations of protein and oligonucleotide alone were 2.5 µM (tetramer concentration) and 5 µM (DNA duplex), respectively. The complexes between YdaT and oligonucleotide were prepared at different YdaT tetramer:DNA molar ratios (0.25:1, 0.25:1.5, 0.5:1, 0.75:1, 1:1, 1.25:1 and 1.5:1), keeping the protein concentration fixed at 2.5 µM YdaT tetramer. Native mass spectrometry was performed on a Synapt G2 mass spectrometer (Waters). The samples were introduced into the gas phase through nano-electrospray ionization with in-house-prepared gold-coated borosilicate glass capillaries. The settings were optimized for the analysis of larger structures as natively as possible. The critical voltages and pressures used were a sampling cone voltage of 50 V and a trap collision energy of 10 V, with pressures throughout the instrument of 6.18 and 2.42 × 10−2 mbar for the source and trap collision cell regions, respectively. Analysis of the acquired spectra was performed using MassLynx version 4.1 (Waters). Native MS spectra were smoothed (to an extent depending on the size of the complexes) and additionally centred to calculate the molecular weights to determine precise stoichiometries.

2.7. Isothermal titration calorimetry

Oligonucleotides were purchased from Sigma and were HPLC-purified. Double-stranded DNA fragments were prepared by mixing single-stranded oligonucleotides corresponding to the upper and lower strands of the operator DNA in a 1:1 molar ratio followed by incubation at 95°C for 5 min in a water bath. They were subsequently left to cool to room temperature. The annealing was checked by native PAGE. Proteins and oligonucleotides were dialysed against 10 mM NaH2PO4, 10 mM Na2HPO4, 100 mM NaCl, 50 mM glutamic acid, 50 mM arginine pH 7.5 with two buffer changes followed by overnight dialysis. Prior to measurements, the samples were spun down at 13 300 rev min−1 for 10 min and degassed on a degassing station (TA Instruments) for 15 min.

The titrations were carried out at 25°C. The concentrations of the oligonucleotides in the sample cell were 5 µM. The concentration of YdaT in the syringe ranged between 25 and 140 µM calculated for YdaT as a tetramer. When OM was titrated with YdaT, the concentration in the cell was 5 µM and that in the syringe was 100 µM. The measurements were performed on a VP-ITC microcalorimeter (MicroCal) or a MicroCal PEAQ-ITC microcalorimeter (Malvern Panalytical). The integration of thermograms was performed with NITPIC (Scheuermann & Brautigam, 2015[Scheuermann, T. H. & Brautigam, C. A. (2015). Methods, 76, 87-98.]).

For the binding of YdaT to OM, we assumed that YdaT contains two independent non-equivalent binding sites for OM. The corresponding binding reaction can be represented as

[\eqalignno {{\rm YdaT} + {\rm O_M} & \buildrel {K_1} \over \rightarrow ({\rm YdaT} {-} {\rm O_M}) \cr ({\rm YdaT} {-} {\rm O_M}) + {\rm O_M} & \buildrel {K_2} \over \rightarrow ({\rm O_M}{-} {\rm YdaT}{-}{\rm O_M}). & (2)}]

For the titrations of YdaT against OLM and OMR an additional binding step was included in the mechanism, since YdaT can also bind to the second site on the DNA,

[({\rm O_{LM}} {-} {\rm YdaT} {-} {\rm O_{LM}}) + {\rm YdaT} \buildrel {K_3} \over \rightarrow ({\rm O_{LM}} {-} {\rm YdaT} {-} {\rm O_{LM}} {-} {\rm YdaT}) \eqno (3)]

and

[({\rm O_{MR}} {-} {\rm YdaT} {-} {\rm O_{MR}}) + {\rm YdaT} \buildrel {K_3} \over \rightarrow ({\rm O_{MR}} {-} {\rm YdaT} {-} {\rm O_{MR}} {-} {\rm YdaT}), \eqno (4)]

The values of the equilibrium constants K1, K2 and K3 and the corresponding enthalpies of reaction were obtained by fitting the appropriate model equation to the ITC data similarly to as described previously (Vandervelde et al., 2017[Vandervelde, A., Drobnak, I., Hadži, S., Sterckx, Y. G., Welte, T., De Greve, H., Charlier, D., Efremov, R., Loris, R. & Lah, J. (2017). Nucleic Acids Res. 45, 2937-2950.]). A system of mass-balance equations was derived given the above reaction schemes. Given the total concentrations of reactants (YdaT and oligonucleotides) and the assumed values of the constants Ki, we calculate the equilibrium concentrations of all molecular species using a root-solving routine. By calculating the composition of the experimental system at each point during titration we then calculate the model-based value of the enthalpy as

[\Delta H = {\textstyle \sum\limits_i} \Delta {H_i} \left( {{\partial n_i} \over {\partial n_{\rm tit} }} \right)_{p,T}, \eqno (5)]

where ΔHi is the enthalpy of formation of the complex i, (∂ni/∂ntit)p,T is the corresponding partial derivative at the given pressure p and temperature T in which ni is the amount of complex i and ntit is the amount of added titrant (YdaT or oligonucleotide, depending on the titration). The parameters Ki and Hi were adjusted using the Nelder–Mead optimization algorithm to produce the best match between the model-calculated and experimental values of ΔH. The interactions of YdaT with OM in both direct (protein to DNA) and reverse titrations were fitted simultaneously (global fit), while for interactions of YdaT with OLM or OMR only direct titrations were used for fitting.

2.8. Electrophoretic mobility shift assay (EMSA)

For radioactive EMSA, a 206 bp segment of DNA was used that corresponds to the end of the ydaS gene and the start of the ydaT gene. The DNA was generated by PCR using primers 5 and 6 (Supplementary Table S1) and genomic DNA of E. coli O157:H7 strain EDL933 as a template. One of the primers was labelled with γ-32-P-ATP using T4 polynucleotide kinase. The PCR fragment was purified from polyacrylamide gel and concentrated by ethanol precipitation. In the same way, a non­specific DNA was generated from genomic DNA of Cupriavidus metallidurans NA4 containing the promotor region of prsQ2. Protein and DNA were mixed in 20 mM Tris, 200 mM NaCl pH 7.5 in a total volume of 20 µl at different protein concentrations and left to incubate for 30 min at room temperature. The protein concentrations (tetramer equivalents) were in the range 2–30 µM. DNA with 7500 cpm was used in each reaction. The samples were then loaded onto a 6% non­­denaturing polyacrylamide gel using 2 µl loading dye (25% Ficoll, 0.1% xylene cyanol, 0.1% bromophenol blue).

2.9. DNase I footprinting

For DNase I footprinting, a 236 bp segment of DNA corresponding to the end of the ydaS gene and the start of the ydaT gene was generated by PCR using primers 5 and 7 (Supplementary Table S1) as described above with genomic DNA of E. coli O157:H7 strain EDL933 as a template. One of the primers was labelled with γ-32P-ATP using T4 polynucleotide kinase. The PCR fragment was purified from polyacrylamide gel and concentrated by ethanol precipitation. The protein solution was mixed with labelled DNA at different concentrations of the protein for 30 min in 20 mM Tris, 200 mM NaCl pH 7.5 in a total volume of 20 µl before adding 0.2 µl DNase I. The reaction was stopped by the addition of 12.5 µl 3 M ammonium acetate, 0.25 M EDTA. The DNA was ethanol precipitated (3 µl of 3 mg ml−1 yeast tRNA was added to the precipitation solution for higher DNA recovery) and analysed on a 6% denaturing polyacrylamide gel using formamide dye (0.03% xylene cyanol, 0.03% bromophenol blue, 20 mM EDTA dissolved in formamide) as a loading buffer. Reference sequencing ladders were prepared from the same 32P-ATP labelled DNA using citrate or hydrazine following the Maxam–Gilbert sequencing method (Maxam & Gilbert, 1980[Maxam, A. & Gilbert, W. (1980). Methods Enzymol. 65, 499-560.]).

2.10. Small-angle X-ray scattering

Small-angle X-ray scattering (SAXS) data for YdaT and its mutants were collected in HPLC mode on the SWING beamline at the SOLEIL synchrotron, Gif-sur-Yvette, France. The data for YdaT in complex with the OM operator fragment were collected on beamline BM29 at the European Synchrotron Radiation Facility (ESRF), Grenoble, France, also in HPLC mode. Details of data collection and analysis are given in Supplementary Table S4. Samples were dialysed against 20 mM Tris–HCl, 200 mM NaCl pH 8 with three buffer changes, with the last buffer change left to dialyse overnight. The dialysis buffer was filtered through a 0.20 µm HAWP filter. For free YdaT and its mutants, a Shodex KW402.5-4F column (Showa Denko K. K.) was equilibrated with the dialysis buffer and 45 µl sample at 10 mg ml−1 was injected into the column. The sample was run at 0.3 ml min−1. Prior to measurements, the sample was centrifuged at 13 300 rev min−1 for 10 min. The YdaT–DNA complex for SAXS measurements was prepared by mixing YdaT with an excess of the oligonucleotide OM. The complex was left to incubate for 30 min and then injected onto an ENrich SEC 650 column (Bio-Rad). The peak corresponding to the complex was collected and concentrated to a final volume of 50 µl. The complex was then injected onto a pre-equilibrated AdvanceBio SEC 300 column (Agilent) and run at 0.16 ml min−1 at the beamline. Radiation damage was monitored by evaluating Rg values per frame during data collection. The data were normalized to the intensity of the transmitted beam and radially averaged. The contributions of the buffer to the scattering were measured at the beginning of the elution and were subtracted from the scattering of the protein. The data from the SWING beamline were processed with Foxtrot (David & Pérez, 2009[David, G. & Pérez, J. (2009). J. Appl. Cryst. 42, 892-900.]) and those from the BM29 beamline were processed with CHROMIXS (Panjkovich & Svergun, 2018[Panjkovich, A. & Svergun, D. I. (2018). Bioinformatics, 34, 1944-1946.]).

All simulations were performed in XPLOR-NIH version 2.49 (Schwieters et al., 2003[Schwieters, C. D., Kuszewski, J. J., Tjandra, N. & Clore, G. M. (2003). J. Magn. Reson. 160, 65-73.]). Refinement of the YdaT tetramer against the experimental SAXS data was carried out starting from the crystal structure presented in this work. Protons and atoms of the residues that were not resolved in the X-ray structure were added in XPLOR-NIH, followed by minimization of the energy function consisting of the standard geometric (bonds, angles, dihedrals and impropers) and steric (van der Waals) terms. The position of the C-terminal four-helix bundle was kept fixed and the N-terminal POU domains were treated as rigid-body groups, while the N- and C-terminal tails and the flexible hinges (residues Ser96–Tyr99) were given full torsional degrees of freedom. YdaT1–96 was refined as a monomer. The POU domain was kept fixed, while the N-terminal purification tag of YdaT1–96 was allowed to move. The computational protocol comprised an initial simulated-annealing step followed by side-chain energy minimization as described previously (Schwieters et al., 2003[Schwieters, C. D., Kuszewski, J. J., Tjandra, N. & Clore, G. M. (2003). J. Magn. Reson. 160, 65-73.]; Schwieters & Clore, 2014[Schwieters, C. D. & Clore, G. M. (2014). Prog. Nucl. Magn. Reson. Spectrosc. 80, 1-11.]). In addition to the standard geometric and steric terms, the energy function included a knowledge-based dihedral angle potential and a SAXS energy term incorporating the experimental data (Schwieters & Clore, 2014[Schwieters, C. D. & Clore, G. M. (2014). Prog. Nucl. Magn. Reson. Spectrosc. 80, 1-11.]; Schwieters et al., 2018[Schwieters, C. D., Bermejo, G. A. & Clore, G. M. (2018). Protein Sci. 27, 26-40.]).

For refinement of the YdaT–DNA complex, the atomic coordinates were taken from a model built from the crystal structure of YdaT solved in this work (PDB entry 8bt1) and that of mouse Oct-4 bound to its operator DNA (PDB entry 6ht5; J. Vahokoski, V. Pogenberg & M. Wilmanns, unpublished work). The position of the C-terminal four-helix bundle was kept fixed and pairs of YdaT DNA-binding domains and their respective DNA double-stranded helix were grouped as rigid-body units, while the N- and C-terminal tails and the flexible hinge (residues Ser96–Tyr99) were given full torsional degrees of freedom. Multiple copies of the molecular system (N = 1–5) were refined simultaneously in order to simulate molecular ensembles of multiple conformers (Schwieters & Clore, 2014[Schwieters, C. D. & Clore, G. M. (2014). Prog. Nucl. Magn. Reson. Spectrosc. 80, 1-11.]).

In each refinement run, 100 structures were calculated and the ten lowest-energy solutions, representing the best agreement with the experimental data, were retained for subsequent analysis. The agreement between the experimental and calculated SAXS curves (obtained with the calcSAXS helper program, which is part of the XPLOR-NIH package) was assessed by calculating χ2,

[{\chi^2} = {1 \over {n - 1}} {\textstyle\sum \limits_{i=1}^n} \left[{{I(q)_{{\rm calc},i} - I(q)_{{\rm exp},i}} \over {\delta I(q)_{{\rm exp},i}}} \right]^2, \eqno (6)]

where I(q)calc,i and I(q)exp,i are the scattering intensities at a given q for the calculated and experimental SAXS curves, respectively, δI(q)exp,i is the experimental error on the corresponding I(q)exp,i value and n is the number of data points defining the experimental SAXS curve.

3. Results

3.1. YdaT belongs to the POU family

YdaT was crystallized and its structure was determined at 2.4 Å resolution to Rwork = 0.195 and Rfree = 0.269 (Table 2[link] and Supplementary Tables S2 and S3). The crystal contained a tetramer in the asymmetric unit. The X-ray data are highly anisotropic, and despite anisotropy correction using STARANISO the R factors remained relatively high. While stereochemical parameters and density fit are good for chains A and B, they are significantly poorer for chain C and especially for chain D, despite careful model building combined with experimenting with different refinement strategies. The final structure was built for residues Lys2–Leu124. The N-terminal His tag as well as the C-terminal residues Tyr125–His141 are disordered and do not provide interpretable electron density.

The YdaT monomer consists of an N-terminal globular helix–turn–helix (HTH)-containing domain followed by a long 29-residue α-helix (Fig. 2[link]a and Supplementary Fig. S1a). The globular domain is all-α, with four longer four- to five-turn α-helices (α1–α4 in Fig. 2[link]a and Supplementary Fig. S1a) followed by a shorter two-turn α-helix (α5). A DALI search with residues Lys4–Met82 of the YdaT monomer picks up a whole series of DNA-binding proteins that all contain a helix–turn–helix structural motif. Next to the obvious YdaT from E. coli O6:H1 (PDB entry 3c4r; Z-score of 20.1 and 84% sequence identity), the best matches involve POU-domain transcription factors (Table 3[link]). POU domains are conserved domains that are found in eukaryotes, with the acronym referring to three transcription factors (Pit-1, Oct1/2 and Unc-86) in which the domain was first described (Phillips & Luisi, 2000[Phillips, K. & Luisi, B. (2000). J. Mol. Biol. 302, 1023-1039.]). The first bacterial match is the transcription regulator ClgR from Corynebacterium glutamicum (Russo et al., 2009[Russo, S., Schweitzer, J. E., Polen, T., Bott, M. & Pohl, E. (2009). J. Biol. Chem. 284, 5208-5216.]), while the first toxin–antitoxin-related match is HipB from Shewanella oneidensis (Wen et al., 2014[Wen, Y., Behiels, E., Felix, J., Elegheert, J., Vergauwen, B., Devreese, B. & Savvides, B. (2014). Nucleic Acids Res. 42, 10134-10147.]).

Table 3
Structural homologs of the N-terminal domain of YdaT picked up in a DALI search

The list shows the ten closest structures after removing duplicates.

Protein Organism PDB entry, chain DALI Z-score R.m.s.d. (Å) Sequence identity (%) No. of common Cα atoms Reference
YdaT Escherichia coli O6 3c4r, A 20.1 0.3 84 82 New York SGX Research Center for Structural Genomics (unpublished work)
Pit-1 Rattus norvegicus 1au7, B 4.9 3.0 11 61 Jacobson et al. (1997[Jacobson, E. M., Li, P., Leon-del-Rio, A., Rosenfeld, M. G. & Aggarwal, A. K. (1997). Genes Dev. 11, 198-212.])
Oct-4 Homo sapiens 6yov, K 4.9 2.1 11 57 Michael et al. (2020[Michael, A. K., Grand, R. S., Isbel, L., Cavadini, S., Kozicka, Z., Kempf, G., Bunker, R. D., Schenk, A. D., Graff-Meyer, A., Pathare, G. R., Weiss, J., Matsumoto, S., Burger, L., Schübeler, D. & Thomä, N. H. (2020). Science, 368, 1460-1465.])
Oct-1 Homo sapiens 1hf0, A 4.8 3.0 10 68 Reményi et al. (2001[Reményi, A., Tomilin, A., Pohl, E., Lins, K., Philippsen, A., Reinbold, R., Schöler, H. R. & Wilmanns, M. (2001). Mol. Cell, 8, 569-580.])
Oct-4 Mus musculus 6ht5, E 4.5 2.5 11 57 J. Vahokoski, V. Pogenberg & M. Wilmanns (unpublished work)
Oct-6 Mus musculus 2xsd, C 4.4 3.1 15 62 Jauch et al. (2011[Jauch, R., Choo, S. H., Ng, C. K. L. & Kolatkar, P. R. (2011). Proteins, 79, 674-677.])
Pit-1 Homo sapiens 5wc9, A 4.4 3.0 10 62 Agarwal & Cho (2018[Agarwal, S. & Cho, T. Y. (2018). Nucleic Acids Res. 46, 929-941.])
Brn-5 Homo sapiens 3d1n, P 4.1 2.9 10 60 Pereira & Kim (2009[Pereira, J. H. & Kim, S.-H. (2009). J. Struct. Biol. 167, 159-165.])
ClgR Corynebacterium glutamicum 3f51, C 4.0 2.4 10 50 Russo et al. (2009[Russo, S., Schweitzer, J. E., Polen, T., Bott, M. & Pohl, E. (2009). J. Biol. Chem. 284, 5208-5216.])
HipB Shewanella oneidensis 4pu7, B 3.6 2.5 8 52 Wen et al. (2014[Wen, Y., Behiels, E., Felix, J., Elegheert, J., Vergauwen, B., Devreese, B. & Savvides, B. (2014). Nucleic Acids Res. 42, 10134-10147.])
†N-terminal domain, residues 1–96.
[Figure 2]
Figure 2
Structure of YdaT. (a) Cartoon representation of the YdaT monomer (chain A from PDB entry 8bt1, this work). Secondary-structure elements are labelled except for helix α5, which is hidden behind helices α2 and α3. The helix–turn–helix motif is highlighted, with the recognition helix (α3) in red, helix α2 in orange and the connecting loop in yellow. (b) Cartoon representation of the YdaT tetramer with each chain in a distinct colour. The two subunits in shades of green form one functional DNA-binding dimer (chains A and C in PDB entry 8bt1), while the other two subunits in shades of blue (chains B and D) form the second functional DNA-binding dimer. (c) Superposition of the HTH motifs of YdaT [coloured as in (a)] and mouse Oct-4 (blue; PDB entry 6ht5) based on their POU domains (Phillips & Luisi, 2000[Phillips, K. & Luisi, B. (2000). J. Mol. Biol. 302, 1023-1039.]). The long loop between helices α2 and α3 is absent from Oct-4 and the recognition helix α3 is significantly longer in YdaT compared with Oct-4. (d) Tetramer formation though the creation of a four-helix bundle. In two subunits helix α6 is shown as a cartoon, while in the other two subunits this helix is shown as a grey Cα trace. Side chains that make up the hydrophobic core are shown as sticks. (e) Superposition of the four subunits in the crystal structure of the YdaT tetramer. The tetramerization helices α6 are superimposed, showing the variability in orientation of the corresponding POU domains. The POU domains of chains A and C are coloured green and oriented differently from the POU domains in chains B and D, which are coloured blue.

A long (14-residue) loop between helices α2 and α3 making up the HTH motif is unique to YdaT and is absent from all other POU-domain structures. In addition, the central α-helix α3 (Ala50–Asp63, which is the recognition helix in the HTH motif) of YdaT is longer than the corresponding helix in POU-domain structures such as Pit-1 or Oct-4 (PDB entries 1au7 and 6ht5), and α-helix α1 (Glu6-His17) is in a different relative orientation (Figs. 2[link]a and 2[link]b). A BLAST search (Altschul et al., 1997[Altschul, S. F., Madden, T. L., Schäffer, A. A., Zhang, J., Zhang, Z., Miller, W. & Lipman, D. J. (1997). Nucleic Acids Res. 25, 3389-3402.]) of our YdaT sequence against the UniProtKB reference proteomes and Swiss-Prot database resulted in 40 hits with an E-value smaller than 0.1. In these sequences the α2–α3 loop varies in length between nine and 23 residues and does not contain any obviously conserved residues (Supplementary Fig. S1b).

The POU domain of YdaT is highly rigid, with r.m.s.d. values between 0.4 and 0.5 Å for all backbone (N, Cα, C and O) atoms of residues 3–90. Only Asp44–Glu49, which corresponds to the C-terminal part of the long YdaT-specific loop between helices α2 and α3, shows some small variation.

3.2. YdaT is a tetramer in solution

YdaT forms a tetramer in the crystal (Fig. 2[link]c). The long C-terminal α-helix α6 serves as an oligomerization motif, with four such helices assembling into an antiparallel bundle, burying a total surface of around 2600 Å2. A hydrophobic core is formed via the side chains of Ile103, Leu110, Val111, Val114, Phe117, Val118 and Ala121 (Fig. 2[link]d). This C-terminal α-helix α6 can bend slightly, changing the direction of its N-terminus. Chains A and C adopt very similar conformations, as do chains B and D. When both pairs are compared and superimposed using the C-terminal helix α6 (residues Tyr100–Ser120), the orientation of the N-terminal domain rotates by almost 30° (Fig. 2[link]e). The loop Ser94–Tyr99 here serves as a hinge region. The POU-like DNA-binding domains themselves are highly similar, with only some small variations in backbone structure in the loop between helices α2 and α3.

To determine the oligomeric state of YdaT in solution, we performed SEC-MALS. The elution profile (injected monomer concentrations of 54.1 and 541.4 µM, corresponding to 1 and 10 mg ml−1, respectively) shows a single symmetrical peak corresponding to molecular weights of 69.4 or 72.9 kDa, respectively, in close agreement with the theoretical mass of 73.9 kDa for a tetramer (Figs. 3[link]a and 3[link]b). Native mass-spectrometry measurements confirm this result. At a concentration of 1.0 µM YdaT (monomer equivalents) is primarily a tetramer, but some monomer is also present (Fig. 4[link]a).

[Figure 3]
Figure 3
SEC-MALS. (a) SEC-MALS profile of YdaT at a concentration of 10 mg ml−1. (b) A similar profile but for a concentration of 1 mg ml−1. The molecular weight determined from both concentrations is very similar (72.9 and 69.4 kDa) and is within the margin of error of the technique. (c) SEC-MALS profile of YdaT1–96 at a concentration of 10 mg ml−1. With an experimental molecular weight of 11.1 kDa, this corresponds to a monomer.
[Figure 4]
Figure 4
Native mass spectrometry. (a) Native MS spectrum of YdaT at a concentration of 2.5 µM (tetramer concentration). Next to the expected tetramer, peaks corresponding to a YdaT monomer are also visible. (b) When an excess amount of the 30 bp OM operator fragment (15 µM of duplex) is added to YdaT (2.5 µM of tetramer), a single species of a YdaT tetramer bound to two OM duplexes is observed next to an excess of free DNA. (c) When the OM DNA duplex (1.67 µM) becomes saturated with YdaT (2.5 µM), the dominant species becomes a YdaT tetramer bound to a single OM duplex.

The tetramer observed in our crystal structure does not perfectly match 222 symmetry. Because of differences in bending and hinge conformation, the POU domains in chains A and C are significantly further from each other than those in chains B and D (10.9 versus 5.1 Å as their closest distance). Similar relative movements of the POU domains are also observed in the structure of the closely related uncharacterized protein in PDB entry 3c4r. In order to further characterize these inter-domain dynamics in solution, we performed SEC-SAXS (Figs. 5[link]a and 5[link]b). A good fit of the data (χ2 = 1.68) is obtained with an ensemble of ten structures where the N-terminal domains are allowed to move relative to the C-terminal four-helix bundle, confirming the structural variability observed in the crystal.

[Figure 5]
Figure 5
SAXS solution structure of YdaT. (a) Experimental SAXS data of YdaT (black) and the calculated curve (red) obtained from the best-fitting ensemble of ten conformers (χ2 = 1.684). (b) Superimposition of the ten conformers of the YdaT SAXS ensemble (black Cα traces) onto the YdaT crystal structure (cartoon representation coloured as in Fig. 2[link]b). In the SAXS ensemble, the POU domains adopt different orientations relative to the C-terminal four-helix bundle due to the flexibility of the loop between the C- and N-terminal domains (residues 96–99).

3.3. YdaT recognizes an inverted repeat located between the ydaS and ydaT genes

The structural organization of the prophage CP-933P suggests YdaS and YdaT as the equivalents of the Cro and CII repressors in λ, a hypothesis that was recently substantiated by in vivo experiments (Jurėnas et al., 2021[Jurėnas, D., Fraikin, N., Goormaghtigh, F., De Bruyn, P., Vandervelde, A., Zedek, S., Jové, T., Charlier, D., Loris, R. & Van Melderen, L. (2021). mBio, 12, e02947-21.]). As CII binds to the PRE promoter located between the cro and cII genes in λ, we tested in vitro whether YdaT can interact with a 206 bp segment covering the end of the ydaS gene and the start of the ydaT gene. EMSA experiments show concentration-dependent binding to this fragment (Fig. 6[link]a). Several bands can be observed with increasing YdaT concentration, indicating multiple binding events. In contrast, binding to the unrelated promotor/operon region of prsQ2 from Cupriavidus metallidurans NA4 requires roughly fourfold higher YdaT concentrations and does not lead to the observation of multiple species.

[Figure 6]
Figure 6
EMSA. (a) EMSA experiments performed using increasing concentrations of YdaT tetramers on either a 206 bp fragment of the ydaT promotor/operator region or a similar sized fragment of the Cupriavidus metallidurans prsQ2 promotor/operator region. (b) EMSA of YdaT1–96 on the 206 bp ydaT promotor/operator DNA fragment showing a lack of binding.

In order to pinpoint the exact binding site of YdaT, we turned to DNase I footprinting using a slightly longer 236 bp fragment (Fig. 7[link]). At the lowest protein concentration used (0.5 µM), protection was already observed on both strands for a 25 bp region containing a 5′-TTGATTN6AATCAA-3′ inverted repeat located at the end of the coding region of the ydaS gene and downstream from the PRE933P promotor that was proposed as an alternative transcription start for the paaR2–paaA2–parE2 operon and which is controlled via YdaT (Jurėnas et al., 2021[Jurėnas, D., Fraikin, N., Goormaghtigh, F., De Bruyn, P., Vandervelde, A., Zedek, S., Jové, T., Charlier, D., Loris, R. & Van Melderen, L. (2021). mBio, 12, e02947-21.]). This highly protected inverted repeat is referred to as OM (Figs. 1[link] and 7[link]). At higher protein concentrations, starting from 1.0 µM, this DNase I-protected area extends further on each side, resulting in a total region of protection of approximately 95 nt on both strands that is composed of zones of weaker and stronger protection. The latter contain sequences that show sequence similarity to part of the high-affinity binding site and may represent additional lower affinity binding sites, thus explaining the multiple bands observed in the EMSA experiments. These two incomplete inverted repeats are referred to as OL and OR (Figs. 1[link] and 7[link]).

[Figure 7]
Figure 7
DNAse I footprinting. DNase I footprint on a 236 bp operator fragment (bottom strand shown) using increasing concentrations of YdaT (in tetramer equivalents). The nucleotide sequence of the region that becomes protected is shown at the side with an indication of the regions of protection. The sequence of the 95 nt protected region is indicated on the right. At 0.5 µM YdaT, a 25 bp zone of protection (dark grey background) containing the inverted repeat 5′-TTGATTN6AATCAA-3′ (bold) can already be observed. At higher protein concentrations this zone is extended further on both sides of the primary binding site, resulting in a total footprint of 94–97 nt on both strands that consists of zones of stronger (light grey background) and weaker (white background) protection. The former overlap with two imperfect inverted repeats (bold). These inverted repeats are labelled OL, OM and OR as described in the text.

3.4. Thermodynamics of operator binding

Next, we turned to isothermal titration calorimetry (ITC) to understand how YdaT binds to its operator region. We therefore selected a set of fragments containing different potential binding sites (Table 4[link], Supplementary Fig. S5). Fragment OM contains the central inverted repeat that was identified as the main binding site for YdaT in the DNase I protection experiment. This fragment binds to YdaT with an affinity in the submicromolar range and two binding events can be discerned that differ in affinity by approximately a factor of four (Table 4[link], Fig. 8[link], Supplementary Fig. S5). This indicates that the YdaT tetramer binds two duplexes of OM with apparent negative cooperativity of enthalpic origin.

Table 4
Thermodynamics of operator binding obtained from ITC

The Kd and ΔH parameters were obtained from fitting the model equations to the ITC data and are reported at T = 298 K. In all cases index 1 refers to the binding of the central inverted repeat to the first YdaT binding site and index 2 refers to the binding of the same repeat from another DNA molecule to the second YdaT binding site. Index 3 refers to the binding of YdaT to the distant incomplete repeat (R or L), but it is not clear whether one or two YdaT binding sites are engaged in binding. The free energy of association ΔG and the entropic contribution TΔS were calculated using standard equations. Standard mean errors are obtained from the fitting procedure or are calculated through error propagation in the case of TΔS and ΔG. Kd values are given in µM, while ΔG, ΔH and TΔS values are in kcal mol−1.

DNA Kd1 ΔG1 ΔH1 TΔS1 Kd2 ΔG2 ΔH2 TΔS2 Kd3 ΔG3 ΔH3 TΔS3
OM 0.91 ± 0.15 −8.24 ± 0.10 −40.1 ± 1.7 −31.9 ± 1.8 3.73 ± 0.72 −7.40 ± 0.12 −30.2 ± 5.5 −22.8 ± 5.6
OLM 0.48 ± 0.09 −8.62 ± 0.11 −68.0 ± 1.5 −51.1 ± 1.8 1.03 ± 0.21 −8.17 ± 0.15 nd nd 53 ± 53 −5.84 ± 0.60 nd§ nd§
OMR 0.22 ± 0.11 −9.09 ± 0.25 −60.8 ± 2.3 −43.4 ± 2.6 0.89 ± 0.48 −8.25 ± 0.32 nd nd 1.52 ± 0.22 −7.94 ± 0.20 −86 ± 28 −78 ± 28
†Refers to the sum ΔH1 + ΔH2. In the case of OLM and ORM titrations it is not possible to reliably determine ΔH1 and ΔH2 separately as they are strongly correlated. The same holds for the respective entropic contributions. However, the total sum (ΔH1 + ΔH2) can reliably be obtained from fitting. Note that the obtained sums are comparable to that for the titrations with OM (ΔH1 + ΔH2 = −70.3 kcal mol−1), which correspond to the same binding events to the central inverted repeat via the first and second YdaT.
‡Refers to the sum TΔS1 + TΔS2.
§The affinity for the second operator binding site is too weak (Kd3 = 53 µM) to reliably determine ΔH3.
[Figure 8]
Figure 8
Isothermal titration calorimetry. (a) Binding between YdaT and OM. The integrated heat of binding per mole of injectant is plotted as ΔH as a function of the molar ratio r between the host and the ligand in the cell and the syringe, respectively, and fitted as described in Section 2[link]. The full line corresponds to a titration with 5 µM YdaT tetramer in the cell titrated with 100 µM OM duplex in the syringe. The dashed curve corresponds to a titration with 5 µM OM duplex in the cell and 25 µM YdaT tetramer in the syringe. A single global fit for titrations was used to obtain the thermodynamic parameters listed in Table 4[link]. (b) A similar titration for YdaT (100 µM tetramer) titrated into to the OMR duplex (15 µM). (c) A similar titration for YdaT tetramer (140 µM) titrated into the OLM duplex (15 µM). All molar ratios (r = [syringe]/[cell]) were calculated for YdaT as a tetramer and the DNA fragment as a duplex.

To further confirm this proposed binding model to OM, we turned to native mass spectrometry (MS). With an excess of OM, all YdaT tetramers are saturated with two such DNA fragments, leading to a species of 110.7 kDa. When YdaT is in excess, on the other hand, YdaT tetramers with both one and two OM fragments bound are observed (Figs. 4[link]b and 4[link]c, Supplementary Fig. S3), in agreement with the results from ITC.

In contrast, fragments containing the possible alternative incomplete inverted-repeat sequences OL or OR did not allow high-quality ITC data to be measured. Both fragments bind only weakly and no reliable thermodynamic parameters could be obtained (Supplementary Fig. S2). Similarly, longer fragments combining OM with OL or OR (OLM and OMR) show a combination of the two high-affinity events seen for OM as well as a lower affinity event (Table 4[link], Fig. 8[link]). Binding to these low-affinity sites increases the affinity for the main site by a factor of 2–4, indicating an interaction between the different sites. Given the distance between these sites, especially in the case of OLM, this communication is likely to propagate through the DNA rather than through direct contacts between two bound YdaT tetramers.

3.5. SAXS model of the YdaT–operator complex

The four DNA-binding domains of YdaT are oriented such that two surfaces are generated that face away from each other, and in principle each can accommodate an ∼30 bp DNA duplex. This is in agreement with the binding stoichiometry that is obtained from ITC and native MS. Using the structure of mouse octamer-binding protein 4 (Oct-4) in complex with a 21 bp DNA duplex (PDB entry 6ht5) as a guide, we built a model of YdaT bound to a 30 bp B-DNA duplex containing the OM sequence identical to that used for ITC (OM in Supplementary Table S4).

Next, this model was validated using SAXS (Fig. 9[link] and Supplementary Table S4). The molecular weight determined from the SAXS data is about 102 kDa, which is close to the theoretical molecular weight of 110.7 kDa for a complex consisting of one YdaT tetramer and two OM molecules. The central four-helix bundle was fixed, while the POU domains were allowed to reorient while remaining docked onto the DNA through variation of the hinge loop Ser96–Tyr99. The N-terminal His tag and the C-terminal 20 amino acids remain highly flexible. As a consequence, pairs of two POU domains (chains A and C or chains B and D) remain locked together via the bridging DNA molecule and their movements relative to the C-terminal helix α6 become highly restricted compared with those observed in the free structure (Figs. 9[link]a and 9[link]b). The best fit to the experimental data (χ2 = 1.418) was obtained with ten ensembles each containing two conformers. More conformers did not improve the fit.

[Figure 9]
Figure 9
SAXS solution structure of the YdaT–OM complex. (a) Experimental SAXS curve (black) with the theoretical curve for YdaT–OM (red) superimposed. The theoretical curve was calculated from the ten best-fitting ensembles (χ2 = 1.418) each consisting of two models. (b) YdaT tetramer bound to two OM duplexes. Ten superimposed conformers from the SAXS ensembles are shown. The DNA–POU-domain assemblies show rigid-body movement relative to the four-helix core of the YdaT tetramer. Within the POU domains, structural variations relative to the free structure are limited to the N-terminal His tag and the C-terminal tail as well as the loop (residues 96–99) between the POU domain and C-terminal helix α6. (c) Enlargement of the HTH motif bound to the OM duplex.

Compared with the crystal structure of the free form of YdaT, no conformational changes are required in YdaT except for some rigid-body movement of the different POU domains relative to each other and the reorientation of a few side chains. In particular, the AC pair of POU domains are oriented in our crystal structure of the free state such that they cannot correctly bind together to the same DNA duplex.

The residues of the Leu35–Glu49 loop fold over the ribose-phosphate backbone of one DNA strand, while the recognition helix α3 (Ala50–Asp63) of the HTH motif docks into the major groove of the DNA (Fig. 9[link]c). Arg60 is nicely positioned to make base-specific hydrogen bonds to Gua21. This arginine is highly conserved in all available POU-domain structures, where it makes similar contacts. It is also conserved in all sequences of YdaT homologs picked up by our BLAST search. Interestingly, in some of these sequences an insertion of a single amino acid between Phe59 and Arg60 is observed that is predicted by AlphaFold2 to result in a single turn of π-helix to allow the side chain of Arg60 to remain in the correct orientation and position. The side chain of Arg53 is likely to form a hydrogen bond to the neighbouring backbone phosphate group of Thy10. Residues of the His17–Gly20 loop together with the N-terminus of helix α2 (Glu21–Glu34) are in contact with the other DNA strand and a hydrogen bond between the backbone amide of Glu21 and an O atom from the ribose backbone is likely. This residue is also not conserved in the eukaryotic POU domains or in the bacterial YdaT homologs.

3.6. Oligomerization of YdaT is required for DNA binding

In order to understand the role of oligomerization in DNA binding, we created a mutant in which YdaT is truncated after His96 (YdaT1–96). This corresponds to a protein consisting of the N-terminal domain but lacking the C-terminal oligomerization helix α6. YdaT1–96 appears to be a well folded species in solution with a predominantly α-helical structure (Supplementary Fig. S4a), in agreement with the conformation of this domain in the crystal structure of the full-length protein. SEC-MALS and SAXS further support this conclusion (Supplementary Table S4, Fig. 3[link]c, Supplementary Figs. S5a and S5b). No measurable interaction between YdaT1–96 and the YdaT operator DNA was detected in EMSA experiments (Fig. 6[link]b), indicating that a monomeric POU domain is insufficient for effective DNA binding. The molecular weights determined from SEC-MALS and SAXS analysis are 11.1 and ∼12 kDa, respectively, and are in close agreement with the theoretical molecular weight of 13.3 kDa for a monomeric species. The theoretical scattering curve calculated for the ensemble of ten conformers fits the experimental curve well (χ2 = 0.983).

4. Discussion

YdaT was originally identified as one of two potential transcripts encoded by the ydaST operon in E. coli OH157:H7. This operon was originally suspected to encode a toxin–antitoxin pair (Sevin & Barloy-Hubler, 2007[Sevin, E. W. & Barloy-Hubler, F. (2007). Genome Biol. 8, R155.]), but was recently shown to be part of the cryptic prophage CP-933P, where the cognate proteins function as equivalents of λ Cro and CII (Jurėnas et al., 2021[Jurėnas, D., Fraikin, N., Goormaghtigh, F., De Bruyn, P., Vandervelde, A., Zedek, S., Jové, T., Charlier, D., Loris, R. & Van Melderen, L. (2021). mBio, 12, e02947-21.]). The cryptic prophage CP-933P has lost its ability to enter a lytic cycle. Its immunity region contains the paaA2–parE2 toxin–antitoxin gene pair that replaces the rexA and rexB genes and is preceded by the gene for the PaaR2 regulator that replaces λ CI.

CP-933P YdaT is a representative of a family of transcription regulators found in lambdoid phages and is functionally but not structurally related to λ CII. The DNA-binding domain of YdaT is a POU domain, with an unusually long loop between the two helices of the HTH motif (α2 and α3) as a defining structural feature. A sequence alignment of YdaT homologs shows that this long loop varies in length and sequence within the YdaT family (Supplementary Fig. S1b) and does not contain obvious conserved residues, despite being likely involved in operator binding. Equally, the recognition helix α3 contains only a single residue that is fully conserved in the YdaT family, Arg60, which according to our model is likely to be essential for operator binding. The evolutionary pressure that drives this variability even though the protein is still functional as a repressor in a cryptic phage such as CP-933P (Jurėnas et al., 2021[Jurėnas, D., Fraikin, N., Goormaghtigh, F., De Bruyn, P., Vandervelde, A., Zedek, S., Jové, T., Charlier, D., Loris, R. & Van Melderen, L. (2021). mBio, 12, e02947-21.]) is unclear. It is possible that it relates to a requirement to stably maintain cryptic prophages or segments thereof in the genome, similar to toxin–antitoxin pairs of the same family that are only found together on the same chromosome if they do not interact (Goeders & Van Melderen, 2014[Goeders, N. & Van Melderen, L. (2014). Toxins, 6, 304-324.] and references therein).

YdaT of the cryptic prophage CP-933P is functional as a DNA-binding protein, as illustrated by our EMSA, ITC and DNAse I protection experiments as well as by previously reported in vivo data (Jurėnas et al., 2021[Jurėnas, D., Fraikin, N., Goormaghtigh, F., De Bruyn, P., Vandervelde, A., Zedek, S., Jové, T., Charlier, D., Loris, R. & Van Melderen, L. (2021). mBio, 12, e02947-21.]). This functionality requires oligomerization. YdaT is a symmetric tetramer with two oppositely positioned sets of DNA-binding sites, meaning that it can recognize two 30 bp operator sequences simultaneously. Yet, the CP-933P prophage only contains a single strong binding site, possibly leaving one pair of POU domains unbound. Alternatively, YdaT has the potential to stabilize a DNA loop. However, the distance between the strong main binding site and the two flanking potential secondary sites is too short to allow the formation of a loop. Indeed, weak binding to these sites seems to occur independent of binding to the main site.

CII, the equivalent of YdaT in λ, is also a tetramer, but rather than forming a closed point group has an `unusual dimer-of-dimers' architecture in which two of the subunits (each from a different dimer) are correctly positioned relative to each other to recognize a direct repeat on the operator (Datta et al., 2005[Datta, A. B., Panjikar, S., Weiss, M. S., Chakrabarti, P. & Parrack, P. (2005). Proc. Natl Acad. Sci. USA, 102, 11242-11247.]; Jain et al., 2005[Jain, D., Kim, Y., Maxwell, K. L., Beasley, S., Zhang, R., Gussin, G. N., Edwards, A. M. & Darst, S. A. (2005). Mol. Cell, 19, 259-269.]). The other two monomers form a bridge between the `active' subunits but are not themselves involved in DNA binding. In this architecture the two DNA-binding subunits are oriented in the same direction, as would be required for recognition of a direct repeat. YdaT, on the other hand, forms a more classic type of tetramer with internal 222 point-group symmetry and is therefore suited to recognize an inverted repeat.

Supporting information


Acknowledgements

The authors thank Sarah Haesaerts for technical assistance, Gopinath Muruganandam for help with native mass spectrometry and the beamline staff of the SOLEIL macromolecular crystallography and SAXS beamlines for help during data collection and processing.

Funding information

The following funding is acknowledged: FWO-Vlaanderen (grant No. G.0226.17N to Remy Loris; grant No. G.0033.20N to Remy Loris); Slovenian Research Agency (grant No. P1-0201 to San Hadži).

References

First citationAfonine, P. V., Grosse-Kunstleve, R. W., Echols, N., Headd, J. J., Moriarty, N. W., Mustyakimov, M., Terwilliger, T. C., Urzhumtsev, A., Zwart, P. H. & Adams, P. D. (2012). Acta Cryst. D68, 352–367.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationAgarwal, S. & Cho, T. Y. (2018). Nucleic Acids Res. 46, 929–941.  CrossRef CAS PubMed Google Scholar
First citationAltschul, S. F., Madden, T. L., Schäffer, A. A., Zhang, J., Zhang, Z., Miller, W. & Lipman, D. J. (1997). Nucleic Acids Res. 25, 3389–3402.  CrossRef CAS PubMed Web of Science Google Scholar
First citationBrüssow, H. & Kutter, E. (2004). Bacteriophages: Biology and Applications, edited by E. Kutter & A. Sulakvelidze, pp. 91–128. Boca Raton: CRC Press.  Google Scholar
First citationCampbell, A. (1994). Annu. Rev. Microbiol. 48, 193–222.  CrossRef CAS PubMed Google Scholar
First citationCasjens, S. (2003). Mol. Microbiol. 49, 277–300.  CrossRef PubMed CAS Google Scholar
First citationChristensen-Dalsgaard, M., Jørgensen, M. G. & Gerdes, K. (2010). Mol. Microbiol. 75, 333–348.  Web of Science PubMed CAS Google Scholar
First citationChung, S. & Echols, H. (1977). Virology, 79, 312–319.  CrossRef CAS PubMed Google Scholar
First citationCourt, D. L., Oppenheim, A. B. & Adhya, S. L. (2007). J. Bacteriol. 189, 298–304.  CrossRef PubMed CAS Google Scholar
First citationDatta, A. B., Panjikar, S., Weiss, M. S., Chakrabarti, P. & Parrack, P. (2005). Proc. Natl Acad. Sci. USA, 102, 11242–11247.  Web of Science CrossRef PubMed CAS Google Scholar
First citationDavid, G. & Pérez, J. (2009). J. Appl. Cryst. 42, 892–900.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationEmsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. (2010). Acta Cryst. D66, 486–501.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationGasteiger, E., Hoogland, C., Gattiker, A., Duvaud, S., Wilkins, M. R., Appel, R. D. & Bairoch, A. (2005). The Proteomics Protocols Handbook, edited by J. M. Walker, pp. 571–607. Totowa: Humana Press.  Google Scholar
First citationGoeders, N. & Van Melderen, L. (2014). Toxins, 6, 304–324.  Web of Science CrossRef CAS PubMed Google Scholar
First citationHallez, R., Geeraerts, D., Sterckx, Y., Mine, N., Loris, R. & Van Melderen, L. (2010). Mol. Microbiol. 76, 719–732.  Web of Science CrossRef CAS PubMed Google Scholar
First citationHo, Y. S. & Rosenberg, M. (1985). J. Biol. Chem. 260, 11838–11844.  CrossRef CAS PubMed Google Scholar
First citationJacobson, E. M., Li, P., Leon-del-Rio, A., Rosenfeld, M. G. & Aggarwal, A. K. (1997). Genes Dev. 11, 198–212.  CrossRef CAS PubMed Web of Science Google Scholar
First citationJain, D., Kim, Y., Maxwell, K. L., Beasley, S., Zhang, R., Gussin, G. N., Edwards, A. M. & Darst, S. A. (2005). Mol. Cell, 19, 259–269.  Web of Science CrossRef PubMed CAS Google Scholar
First citationJauch, R., Choo, S. H., Ng, C. K. L. & Kolatkar, P. R. (2011). Proteins, 79, 674–677.  CrossRef CAS PubMed Google Scholar
First citationJobling, M. G. (2018). mSphere, 3, e00163-18.  CrossRef CAS PubMed Google Scholar
First citationJohnson, A. D., Poteete, A. R., Lauer, G., Sauer, R. T., Ackers, G. K. & Ptashne, M. (1981). Nature, 294, 217–223.  CrossRef CAS PubMed Google Scholar
First citationJurėnas, D., Fraikin, N., Goormaghtigh, F., De Bruyn, P., Vandervelde, A., Zedek, S., Jové, T., Charlier, D., Loris, R. & Van Melderen, L. (2021). mBio, 12, e02947-21.  PubMed Google Scholar
First citationKabsch, W. (2010). Acta Cryst. D66, 125–132.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationKaiser, K. & Murray, N. E. (1979). Mol. Gen. Genet. 175, 159–174.  CrossRef CAS PubMed Google Scholar
First citationKobiler, O., Koby, S., Teff, D., Court, D. & Oppenheim, A. B. (2002). Proc. Natl Acad. Sci. USA, 99, 14964–14969.  CrossRef PubMed CAS Google Scholar
First citationKrishnamurthi, R., Ghosh, S., Khedkar, S. & Seshasayee, A. S. N. (2017). mSphere, 2, e00392-17.  CrossRef CAS PubMed Google Scholar
First citationMaxam, A. & Gilbert, W. (1980). Methods Enzymol. 65, 499–560.  CrossRef CAS PubMed Google Scholar
First citationMcCoy, A. J., Grosse-Kunstleve, R. W., Adams, P. D., Winn, M. D., Storoni, L. C. & Read, R. J. (2007). J. Appl. Cryst. 40, 658–674.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationMichael, A. K., Grand, R. S., Isbel, L., Cavadini, S., Kozicka, Z., Kempf, G., Bunker, R. D., Schenk, A. D., Graff-Meyer, A., Pathare, G. R., Weiss, J., Matsumoto, S., Burger, L., Schübeler, D. & Thomä, N. H. (2020). Science, 368, 1460–1465.  CrossRef CAS PubMed Google Scholar
First citationPanjkovich, A. & Svergun, D. I. (2018). Bioinformatics, 34, 1944–1946.  Web of Science CrossRef CAS PubMed Google Scholar
First citationPereira, J. H. & Kim, S.-H. (2009). J. Struct. Biol. 167, 159–165.  CrossRef PubMed CAS Google Scholar
First citationPerna, N. T., Plunkett, G., Burland, V., Mau, B., Glasner, J. D., Rose, D. J., Mayhew, G. F., Evans, P. S., Gregor, J., Kirkpatrick, H. A., Pósfai, G., Hackett, J., Klink, S., Boutin, A., Shao, Y., Miller, L., Grotbeck, E. J., Davis, N. W., Lim, A., Dimalanta, E. T., Potamousis, K. D., Apodaca, J., Anantharaman, T. S., Lin, J., Yen, G., Schwartz, D. C., Welch, R. A. & Blattner, F. R. (2001). Nature, 409, 529–533.  CrossRef PubMed CAS Google Scholar
First citationPhillips, K. & Luisi, B. (2000). J. Mol. Biol. 302, 1023–1039.  CrossRef PubMed CAS Google Scholar
First citationPtashne, M., Jeffrey, A., Johnson, A. D., Maurer, R., Meyer, B. J., Pabo, C. O., Roberts, T. M. & Sauer, R. T. (1980). Cell, 19, 1–11.  CrossRef CAS PubMed Google Scholar
First citationReményi, A., Tomilin, A., Pohl, E., Lins, K., Philippsen, A., Reinbold, R., Schöler, H. R. & Wilmanns, M. (2001). Mol. Cell, 8, 569–580.  PubMed Google Scholar
First citationRusso, S., Schweitzer, J. E., Polen, T., Bott, M. & Pohl, E. (2009). J. Biol. Chem. 284, 5208–5216.  Web of Science CrossRef PubMed CAS Google Scholar
First citationScheuermann, T. H. & Brautigam, C. A. (2015). Methods, 76, 87–98.  Web of Science CrossRef CAS PubMed Google Scholar
First citationSchwieters, C. D., Bermejo, G. A. & Clore, G. M. (2018). Protein Sci. 27, 26–40.  Web of Science CrossRef CAS PubMed Google Scholar
First citationSchwieters, C. D. & Clore, G. M. (2014). Prog. Nucl. Magn. Reson. Spectrosc. 80, 1–11.  Web of Science CrossRef CAS PubMed Google Scholar
First citationSchwieters, C. D., Kuszewski, J. J., Tjandra, N. & Clore, G. M. (2003). J. Magn. Reson. 160, 65–73.  Web of Science CrossRef PubMed CAS Google Scholar
First citationSevin, E. W. & Barloy-Hubler, F. (2007). Genome Biol. 8, R155.  Web of Science CrossRef PubMed Google Scholar
First citationTataurov, A. V., You, Y. & Owczarzy, R. (2008). Biophys. Chem. 133, 66–70.  CrossRef PubMed CAS Google Scholar
First citationTickle, I. J., Flensburg, C., Keller, P., Paciorek, W., Sharff, A., Vonrhein, C. & Bricogne, G. (2018). STARANISO. Global Phasing Ltd, Cambridge, United Kingdom.  Google Scholar
First citationVandervelde, A., Drobnak, I., Hadži, S., Sterckx, Y. G., Welte, T., De Greve, H., Charlier, D., Efremov, R., Loris, R. & Lah, J. (2017). Nucleic Acids Res. 45, 2937–2950.  CrossRef CAS PubMed Google Scholar
First citationWegrzyn, G. & Wegrzyn, A. (2005). Prog. Nucleic Acid Res. Mol. Biol. 79, 1–48.  PubMed CAS Google Scholar
First citationWen, Y., Behiels, E., Felix, J., Elegheert, J., Vergauwen, B., Devreese, B. & Savvides, B. (2014). Nucleic Acids Res. 42, 10134–10147.  CrossRef CAS PubMed Google Scholar

This is an open-access article distributed under the terms of the Creative Commons Attribution (CC-BY) Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are cited.

Journal logoSTRUCTURAL
BIOLOGY
ISSN: 2059-7983
Follow Acta Cryst. D
Sign up for e-alerts
Follow Acta Cryst. on Twitter
Follow us on facebook
Sign up for RSS feeds