Research Article
A bacteriophage mimic of the bacterial nucleoidassociated protein Fis
Soumyananda Chakraborti1,*, Dhanasekaran Balakrishnan1,2,*, Alexander J. Trotter3, William H. Gittens3,4 , Ally W.H. Yang5, Arttu Jolma5, Joy R. Paterson3, Sylwia Światek1, Jacek Plewka1, Fiona A. Curtis3, Laura
Y. Bowers3, Lars-Olof Pålsson4, Timothy R. Hughes5, Michał Taube6, Maciej Kozak6,7, Jonathan G. Heddle1,* and
Gary J. Sharples3
1Malopolska Centre of Biotechnology, Jagiellonian University, Krakow, Poland; 2Postgraduate School of Molecular Medicine, Zwirki i Wigury 61, 02-091 Warsaw, Poland; 3Department of Biosciences, Durham University, Stockton Road, Durham DH1 3LE, U.K.; 4Department of Chemistry, Durham University, Stockton Road, Durham DH1 3LE, U.K.; 5Donnelly Centre, University of Toronto, 160 College Street, Toronto, ON M5S 3E1, Canada; 6Department of Macromolecular Physics, Faculty of Physics, Adam Mickiewicz University, Uniwersytetu Poznańskiego 2, PL 61-614 Poznań, Poland; 7National Synchrotron Radiation Centre SOLARIS, Jagiellonian University, Czerwone Maki 98, 30-392 Krak, Poland
Correspondence: Gary J. Sharples (gary.sharples@durham.ac.uk) and Jonathan G. Heddle ( jonathan.heddle@uj.edu.pl)
We report the identification and characterization of a bacteriophage λ-encoded protein, NinH. Sequence homology suggests similarity between NinH and Fis, a bacterial nucleoidassociated protein (NAP) involved in numerous DNA topology manipulations, including chromosome condensation, transcriptional regulation and phage site-specific recombination. We find that NinH functions as a homodimer and is able to bind and bend double-
stranded DNA in vitro. Furthermore, NinH shows a preference for a 15 bp signature sequence related to the degenerate consensus favored by Fis. Structural studies reinforced the proposed similarity to Fis and supported the identification of residues involved in DNA binding which were demonstrated experimentally. Overexpression of NinH proved toxic and this correlated with its capacity to associate with DNA. NinH is the first example of a phage-encoded Fis-like NAP that likely influences phage excision-integration reactions or bacterial gene expression.
Introduction
Efficient packaging is essential to accommodate the genome within a cell or cellular compartment while at the same time allowing access for the replicative, transcriptional and regulatory machinery. In bacteria, a group of histone-like proteins (e.g. Fis, IHF, H-NS and HU), collectively termed nucleoid-associated proteins (NAPs), fulfill a vital role in chromosome organization and gene expres-sion [1,2]. Each of the NAPs perform discrete roles in intra-and intermolecular nucleoprotein assemblies and are often temporally regulated to suit the requirements of the cell cycle. The diverse functions of Fis include transcriptional regulation of a large number of genes at various stages of growth [3], including those involved in virulence [4].
Fis was originally identified due to its involvement in the topology manipulations necessary to drive
*These authors contributed site-specific inversion reactions involving the Hin and Gin recombinases [5]. Similarly, Fis works with
equally to this work.
Xis to direct excision of the λ prophage
from
the
host
chromosome
[6–8] and can also stimulate phage integration reactions [9–11]. Fis condenses DNA by binding 21 bp segments via its non-specific
Received: 18 February 2020
Revised: 19 March 2020 DNA binding and bending activities [12]. However, Fis is also capable of forming stable complexes
Accepted: 24 March 2020 with multiple, specific DNA sequences found in regulatory areas of the genome [3]. High-affinity Fis binding sites display a degenerate consensus with a G/C dyad separated by a central AT-rich core of
Accepted Manuscript online:
13 bp [13–15]. Fis appears to initially select a DNA target based upon the narrow minor groove found
24 March 2020 Version of Record published: at AT-rich sites, followed by sequence-specific contacts with the adjacent major grooves [14]. The Fis
17 April 2020 homodimer maintains its structural integrity while it draws the duplex over its two helix-turn-helix
©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
1345
(HTH) motifs, stabilizing the association and inducing the DNA to bend by ∼65° [14,16,17]. Each HTH is held in place by multiple side-chain contacts with the phosphodiester backbone stabilized by interactions with selected bases in the major groove [14].
In this study, we identify the previously uncharacterized phage λ NinH protein as a Fis-like homolog. The ninH gene is located immediately downstream of rap (ninG) as part of the ninR region situated between the genes
for
replication
proteins
O
and
P
and
the
Q
antiterminator
and
adjacent
lysis
cassette
[18]. We demon-strate that NinH has double-stranded DNA binding and bending activities that resemble those found with Fis. The expression of NinH confers significant negative effects on bacterial growth and this was used to monitor DNA binding in vivo. The results are consistent with NinH affecting both cellular and viral DNA processing and also gene expression, with potentially advantageous effects on phage multiplication and dissemination.
Materials and methods
NinH cloning
The ninH gene was amplified by PCR from λ genomic DNA (cI857 Sam7) with oligonucleotides 50-AAGTGAGGCCATATGACGTTCTCAG-30 and 50-AAATATTTCGGTATAAGCTTCCATC-30. PCR products were inserted into pT7-7 (pFC109) and pET14b (pFC110) following digestion with NdeI and HindIII (underlined) to generate overexpression constructs of native NinH and N-His6-NinH (His-NinH), respectively. Nucleotide sequencing of these constructs revealed a guanine to adenine change at position 43 of the coding sequence relative to the published genome sequence of λ, although it matches that derived from λ cI857 Sam7. The alteration produces an alanine to threonine substitution at position 15 of NinH; both amino acids are represented among phage NinH family members at this position. Synthetic genes encoding His-NinH and sitedirected mutants with flanking NdeI and XhoI restriction enzyme sites inserted into pET28a(+) were purchased from Twist Bioscience (San Francisco). The sequence of this codon-optimized synthetic gene and its mutant derivatives are listed in Supplementary Figure S1A. The ninH gene was synthesized and inserted into pTH6838
[19] by Genscript to generate an N-terminal GST-NinH fusion. The sequence of this insert is shown in Supplementary Figure S1B.
Purification of NinH
Escherichia coli BL21-AI (Invitrogen; F− ompT gal dcm lon hsdSB(rB − m−B) araB::T7RNAP-tetA) cells carrying pFC109 (native NinH; 2 l) or pFC110 (His-NinH; 500 ml) were grown in LB broth supplemented with 150 mg/ ml ampicillin. NinH expression was induced with 0.5 mM IPTG and 0.2% (w/v) L-arabinose when the culture reached an A650nm of 0.5 and incubation continued for 3 h at 37°C. For NinH, harvested cells were resuspended in 20 mM Tris–HCl pH8, 1 mM EDTA, 0.5 mM DTT (Buffer A), lysed by sonication and the cleared lysate applied to a 10 ml Q-sepharose fast flow column (GE Healthcare); the majority of the NinH protein appeared in the flow through and wash fractions. The combined flow through and wash samples containing NinH were loaded onto a 3 ml dsDNA cellulose column (Sigma) and bound proteins eluted in Buffer A with a gradient of 0.2–1 M KCl. NinH eluted between 0.35 and 0.5 M KCl and pooled fractions were applied to a 1 ml heparin agarose column (Sigma). NinH eluted from the column between 0.4 and 0.8 M KCl and peak fractions were dialyzed in Buffer A containing 50% (v/v) glycerol and stored in aliquots at −80°C.
Cells containing overexpressed His-NinH were resuspended in 20 ml 50 mM NaH2PO4 pH 8, 300 mM NaCl, 10 mM imidazole and lysed by sonication. The cleared lysate was applied to a 1.5 ml His-Select™ Nickel affinity gel (Sigma) column and bound proteins eluted in buffer containing 250 mM imidazole. His-NinH subsequently precipitated, eliminating some additional minor contaminants, and was resuspended in 20 mM Tris–HCl pH8, 1 mM EDTA, 0.5 mM DTT, 1 M KCl. Purified protein was dialyzed against 20 mM Tris–HCl pH8, 1 mM EDTA, 0.5 mM DTT, 0.5 M KCl, 50% (v/v) glycerol and stored in aliquots at −80°C. Additional, His-NinH and mutant proteins were isolated from E. coli BL21-Star (Invitrogen; F− ompT gal dcm lon hsdSB(rB − mB −) rne131 (DE3)) cells carrying pET14b and pET28a plasmid constructs. Cells were grown in LB broth supplemented with 100 mg/ml ampicillin and induced with 1 mM IPTG at an A600nm of 0.6 and incubation continued for 4 h at 37°C. Harvested cells were resuspended in 20 ml column buffer (50 mM NaH2PO4 pH8, 300 mM M NaCl, 1 mM EDTA) containing 10 mM imidazole and lysed by sonication. The cleared lysate was applied to a 5 ml His Trap column (GE Healthcare) and bound proteins eluted in buffer containing a 0–500 mM gradient of imidazole. Purified His-NinH was dialyzed (3 kDa cut-off) against the column buffer. Contaminating DNA
1346 ©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
was removed using a 5 ml HiTrap Heparin column (GE Healthcare), with bound His-NinH eluted using a gradient of 0.3–1.0 M NaCl. Amounts of NinH are expressed as moles of dimer.
In silico NinH structure prediction
The structure of the full-length NinH dimer was validated by fitting an in silico model to data generated from small angle X-ray scattering (SAXS). The NinH monomer structure was homology-modeled using E. coli Fis as a template. The sequences of NinH (UniProt ID P03771) and Fis (UniProt ID P0A6R3) protein were aligned using Clustal Omega by creating two separate sequence alignments (aligning NinH residues 1–44 to the C-terminal region of Fis and NinH residues 45–69 to the N-terminal region of Fis). Two different templates were generated based on the same Fis protein. These templates were used as the input for building NinH monomer models. A multi-template modeling technique was implemented in Modeller9v21 [20] with the default parameters for building the monomer NinH models. A total of 125 NinH monomer models were built and the 15 best models shortlisted based on alignment to Fis. The criteria used for model selection were:
(a) minimizing the RMSD of the structural alignment between the DNA binding region of predicted NinH monomer models residues 29–35 and Fis protein and (b) evaluation of the overall similarity of the C-terminal region between the predicted monomer models and Fis. The NinH homodimer was then modeled using the HADDOCK online web tool [21,22]. The monomer models shortlisted previously were used as inputs and two different sets of DNA binding or dimerization (either His45, His55, Arg56 or His55, Arg56) domain residues were used as constraints for docking the two monomers. A total of 1200 homodimer structures were generated from the 15 monomer models and RMSD was calculated between the residues in the DNA binding region of the predicted models and Fis. A final refinement step was carried out on the favored model using the GalaxyRefineComplex online web tool [23,24] where the model was relaxed and 10 refined models generated. All structures were validated for Ramachandran outliers using the RAMPAGE server [25] and the best model (98.5% of residues in the favored region) was chosen for SAXS validation.
SAXS data collection and modeling
To determine the 3D structure of NinH protein under native conditions we performed SAXS analysis using a laboratory SAXS system Xeuss 2.0 (XENOCS, Grenoble, France) equipped with a MetalJet D2 microfocus X-ray generator (0.134 nm wavelength). The His-NinH solution at 4 mg/ml was centrifuged to remove any potential aggregates and injected into the low-noise liquid sample cell. SAXS data were collected in four sequential 10 min long frames using a Pilatus3R 1 M hybrid photon counting detector (Dectris, Switzerland). Data reduction and buffer subtraction procedures were performed using the Foxtrot package [26]. Resulting scattering profiles were checked for radiation damage and averaged. All structural parameters were calculated using Primus software from the ATSAS package [27]. The distance distribution function was calculated using GNOM [28]. The low-resolution molecular envelope was calculated using 10 rounds of DAMMIF [29] analysis followed by clustering and filtering. Resulting SAXS data were used to validate the quality and probability of the in silico model of the NinH dimer. Since SAXS measurements were performed using His-NinH dimer, flexible overhangs were added to the model in PyMOL to compensate for their electron density.
Quaternary analysis
The oligomeric states of His-NinH wild-type and mutant proteins were estimated by analytical ultracentrifugation (AUC, Beckman) at a concentration of ∼0.6 mg/ml. Sedimentation velocity measurements were performed in a Beckman-Coulter ProteomeLab XL-I analytical ultracentrifuge, equipped with AN-50Ti rotor (8-holes) and 12 mm path length. Experiments were conducted at 4°C at 50 000 rpm in continuous scan mode. Scans were collected at 8 min intervals at both 260 and 280 nm. Fitting of absorbance versus cell radius data was carried out using SEDFIT software. Oligomer formation by His-NinH was also examined using boiled and un-boiled protein samples separated by 15% SDS–PAGE and visualized by staining with either InstantBlueTM (Expedeon) or Coomassie blue. A pre-stained protein molecular mass standard (Bio-Rad or ThermoFisher) was used as a size marker.
DNA binding assays
Oligonucleotides used to make the DNA substrates used in this study are listed in Supplementary Table S1. Electrophoretic mobility shift assays (20 ml) using 32P-labeled DNA substrates were conducted in 50 mM Tris– HCl pH 8.0, 5 mM EDTA, 1 mM dithiothreitol, 5% (v/v) glycerol, 100 mg/ml BSA. Samples were incubated on
©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
1347
ice for 15 min prior to separation on 4% (w/v) PAGE in 6.7 mM Tris–HCl pH 8.0, 3.3 mM sodium acetate, 2 mM EDTA at 160 V. Gels were dried on filter paper and analyzed by autoradiography and phosphorimaging. Additional gel shift assays were performed in PBS (137 mM NaCl, 2.7 mM KCl, 8 mM Na2HPO4 and 2 mM KH2PO4 pH 7.4) with His-NinH and mutant derivatives incubated with Cy5-labeled dsDNA (Cy520) for 30 min at room temperature. In all cases, the DNA concentration was kept constant at 10 nM while protein concentrations varied. DNA–protein mixtures were separated by electrophoresis on 7% native PAGE (6.7 mM Tris–HCl pH 8.0, 3.3 mM sodium acetate, 2 mM EDTA) at 150 V for 35 min. Gels were scanned using a Typhoon scanner (GE Healthcare) and images analyzed and quantified by Image Quant software (GE Healthcare).
Steady-state fluorescence anisotropy measurements were taken at 25°C on a Carey Eclipse Spectrophotometer (Agilent Technologies). A sample (900 ml) containing fluorescein-labeled 10 nM dsDNA (o-FLU20 annealed to o-SS20) in 6.7 mM Tris–HCl pH 8.0, 3.3 nM sodium acetate, 2 mM EDTA was added to a 1 ml quartz cuvette. Fluorescein anisotropy was determined from λem (520 nm) with λex (495 nm). Samples were excited with a PMT voltage of 850 V and the excitation and emission slits at 5 nm. The G-factor was mea-sured at the outset and each anisotropy titration calculated using WinFLR software (Agilent Technologies).
Microscale thermophoresis (MST) was also employed to monitor binding to the Cy520-labeled dsDNA sub-strate in PBS with 0.05% (v/v) Tween-20 (PBST). Proteins were serially diluted (0.03–1000 nM) in PBST buffer in PCR tubes. The Cy5-labeled DNA (2.5–5 nM) was incubated with His-NinH for 30 min at RT, prior to measurement. After incubation, samples were introduced to premium capillaries (Catalogue # MO-K025, Monolith) and run on a Monolith NT.115. Sample measurement was carried out at 22°C and the determination of KD values was performed using proprietary software (Nanotemper).
GST-NinH produced in vitro using T7 RNA polymerase from the ninH gene cloned in pTH6838 was employed in a protein binding microarray (PBM) assay as described previously [30]. Preferential sequence binding with the fusion protein was performed in duplicate on two microarrays with differing probe sequences (denoted ME and HK; [31] in PBS, 2% (w/v) skimmed milk, 0.2 mg/ml BSA, 50 μM zinc acetate, 0.1% Tween-20; 8-mer Z and E scores were calculated as described by Berger et al.[30]. GST-NinH protein was also tested with a bead-based HT-SELEX (high-throughput-systematic evolution of ligands by exponential enrichment) method essentially as described [32,33]. The SELEX process was repeated three times in 140 mM KCl, 5 mM NaCl, 1 mM K2HPO4, 2 mM MgSO4, 100 mM EGTA, 1 mM ZnSO4, 20 mM HEPES-HCl pH 7 and aliquots of the enriched DNA were sequenced at each cycle.
DNA bending assays
Steady-state Fster resonance energy transfer (ssFRET) and time-resolved Fster resonance energy transfer (trFRET) were used to estimate the amount of DNA bending. Steady-state FRET was measured using a Perkin– Elmer LS 55 Fluorescence Spectrometer at room temperature in a final volume of 50 ml with 50 nM DNA. Assays used 1 μM protein and were incubated for 10 min at room temperature prior to analysis. Donor (fluorescein) was excited at 494 nm and emission spectra collected from 490 to 650 nm. Acceptor (tetramethylrhodamine) was directly excited at 558 nm and emission spectra collected from 558 to 650 nm. Excitation and emission bandwidths were 5 nm. Spectra were collected for both donor and donor–acceptor labeled dsDNA and FRET calculated from the increase in acceptor emission using the ratioA method [34]. RatioA was used to calculate energy transfer efficiency (E) using the equation E =(εA[558]/ = εD[494]) × ratioA − (εA[494]/ = εA[558]) where εA[λ] and εD[λ] are the acceptor and donor extinction coefficients provided by the supplier. Dye–dye distances (R) were calculated using the equation E = R06/(R06+R6)[34]. A Fster distance (R0)of 50Å was used in agreement with previous studies [35,36].
The trFRET was assessed using the time-correlated, single-photon counting technique (TCSPC). Oligonucleotides and sample preparation were as for steady-state FRET. Time-resolved fluorescence spectroscopy was performed using a PicoQuant pulsed diode laser LDH-P-C-485 (485 nm, 70 ps pulses full width half maximum (FWHM) at 20 MHz). The emission was detected in a right-angle geometry using suitable band pass/long pass filters (Comar Instruments) with a photon counting module Idquantic (id100-20) in combination with a Becker–Hickl SPC-130 time-correlated single-photon counting module. The data were subse-
P
quently fitted to a sum of exponentials; F(l, t) ¼ Ai exp( kit), by deconvolution with the instrument i
response function (IRF) obtained from a light scattering solution of Ludox particles. The IRF has a FWHM of ∼250 ps which provides a temporal time-resolution of ≥50 ps in the time-resolved fluorescence spectroscopy experiments. Fluorescence decays were collected for both donor and donor–acceptor labeled dsDNA with or
1348 ©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
without protein. Data were analyzed by the Grinvald–Steinberg method [37] to obtain the fluorescence lifetime for the donor and acceptor (τDA) and donor only (τD) labeled oligonucleotides. Energy transfer efficiency was
tDA
calculated using the equation E ¼ 1.
tD
Bacterial growth assays
BL21-AI cells transformed with plasmids carrying the ninH wild-type gene and substitution mutations and the relevant vector controls, pET28a(+) or pET14b, were cultivated at 37°C in LB broth supplemented with 50 mg/ ml kanamycin or 200 mg/ml ampicillin to an A650nm of 0.4. Ten-fold serial dilutions were performed in cold LB broth and 10 ml aliquots spotted onto agar plates containing antibiotic, arabinose (0.2%) and IPTG (1 mM) as appropriate. Plates were incubated at 30°C for 24–36 h and the extent of growth visualized.
Results
NinH resembles Fis
A search for previously uncharacterized λ genes likely to participate in phage DNA metabolism uncovered the 68-residue ninH gene product as a likely candidate due to the presence of a putative HTH motif. Residues 17– 38 in NinH were predicted to form a HTH using two different search methodologies [38,39]. NinH belongs to the
Protein
Family
PF06322
[40] as a domain of unknown function incorporating 346 entries. Members of the family are exclusively linked to phage and prophage sequences, with the former found primarily among the Siphoviridae and Podoviridae. Position-specific iterative BLAST analysis [41], alongside Phyre2 [42] and I-TASSER [43] structural homology searches with NinH, uncovered significant matches with three helices spanning the HTH motif in the E. coli Fis protein (Figure
1A
and Supplementary Figure S2) and other transcriptional regulators, including homologs from the NtrC and LysR families. The ancestral relationship between Fis and NtrC has been noted previously [44]. Using Phyre2, Fis yielded the highest match with 91.9% confidence over 32 amino acids at the N-terminus of NinH (residues 6–37). This region overlaps with the HTH helix motif
in
Fis
that
spans
helices
C
and
D
and
corresponds
to
the
DNA
binding
site
[14]. E. coli Fis shares 40% identity and 67% similarity with NinH over 30 residues in this region (Figure
1A). Strikingly, seven out of eight residues that are known to contact DNA from the crystal structure of the bipartite complex [14] are con-served in λ NinH (Figure
1A–D). The single exception is Asn84 in E. coli Fis which corresponds to Ser29 in NinH; although this residue is dispensable for recognition of high-affinity sites [45]. NinH lacks the N-terminal section present in Fis and has an extended C-terminus (Figure
1A). Residues 44–48 and 51–55 are predicted by JPred4
[46] and other secondary structure prediction tools to form β sheets and these could be analogous to those found at the N-terminus of Fis (Figure
1C). Overall, the bioinformatic analysis is consistent with NinH being a short, DNA binding protein related to E. coli Fis.
NinH structure
While the bioinformatic analysis outlined above supports structural homology between the C-terminal DNA binding region of Fis and the N-terminal portion of NinH, questions remain over the structure and role of the NinH C-terminus and its potential involvement in dimerization. To address this, we attempted to produce an in silico structural model of a NinH dimer which was further validated using SAXS.
For the in silico model, based on biochemical evidence presented below that the C-terminal region contributes to dimer formation, we chose a model that assumed the involvement of His55 and Arg56 and used these residues as a constraint in HADDOCK. The resulting model (Figure
2A,B) had a ‘closed’ conformation and an RMSD of ∼5.6 Å when compared with the Fis structure and ∼4.8 Å when only the residues of DNA binding region and the chosen model are considered. From the library of potential structures obtained, we noted that nearly all gave a similar structure to the αD helix in Fis corresponding to the DNA binding region (NinH residues 29–35). The C-terminal regions of the NinH protein were consistently predicted to form a β sheet structure, as found in the N-terminal region of Fis. In Fis, these regions are adjacent to the dimer interface formed by the interlocking of helices A and B [14], i.e. are in a ‘closed’ conformation. Our modeling generated a range of conformations from wide open to closed states (6.7 Å between C-terminal loop tips for closed states and
11.2 Å for open; Supplementary Figure S3A). It should be noted that NinH residues 44–48 predicted to form a β sheet (Supplementary Figure S2A) are not part of the hairpin structure in the modeled dimer.
To see if the model fitted experimentally obtained data, we carried out a structural analysis of purified His-NinH using SAXS. SAXS is a powerful structure determination method that does not require labeling or
©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
1349
Figure 1. Sequence and structural similarities between NinH and Fis.
(A)
Sequence homology between NinH and the Fis DNA binding domain. The position of homologous regions is depicted on the 98-residue Fis and 68-residue NinH proteins. The sequence alignment highlights identical residues (purple), functionally equivalent amino acids (blue) and others (gray). Residues in the Fis-DNA structure that participate in DNA binding are indicated by a filled circle. With the exception of Asn84 (orange circle), all of these are conserved in λ NinH and labeled above the primary sequence. Helices are labeled αA, αB and αC in NinH and correspond to helices αB, αC and αD in Fis, the latter two helices comprise the helix-turn-helix. (B) The NinH structure modeled by Phyre2 [42] on the apo Fis structure (PDB 1FIP). Residues are colored as in (A) and those matching the key residues in Fis that contact DNA are shown in stick format and labeled. (C) Crystal structure of the Fis–DNA complex [14]; PDB 3IV5) also showing residues conserved between NinH and Fis.
(D)
Detail of the Fis–DNA structure with residues involved in DNA contact depicted in stick format. Residues are colored as in (A), with N84 (orange) being replaced by serine in the NinH sequence. The orientation of the complex differs slightly from (C) to make it easier to visualize the contacts with DNA. Key guanosine and cytosine base pairs recognized by Fis are indicated at −7 and 7.
1350 ©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
Figure 2. Modeling and structural analyses of NinH.
(A) Primary sequence of λ NinH protein highlighting the predicted secondary structure elements. (B) Modeled structure of the NinH dimer with secondary structure features indicated. (C) Top, front and side views of the NinH dimer model (green) with N-terminal His-tags (blue) fitted into the low molecular mass envelope generated from the SAXS profile. The high-resolution model fits the triangular prism shape well with some empty space suggesting flexibility of the N-terminal His-tags. (D) Overlay of SAXS scattering profile of His-NinH (black dots) and putative scattering profile generated from the high-resolution model using Crysol software (blue). The overall quality of the fit is high (χ2-test value of 1.03). (E) Front and side views of the structural alignment of the NinH modeled complex with Fis (PDB: 3IV5).
crystal formation, but the resulting structural information is of low resolution. It is, however, an excellent validation tool for high resolution in silico modeling. The structural properties resulting from SAXS analysis con-firmed NinH to be a globular dimer with a molecular mass of ∼20 kDa, as expected (Supplementary Figure S3B). To generate such a model, we firstly built the NinH dimer de novo as described in the methods section. The chosen model fits well into the low molecular envelope that was generated solely from SAXS data (Figure
2C), further validating the generated model. It also fits well in the Pearson’s χ2-test between the putative scattering curve generated from this model and SAXS experimental data as calculated from Crysol software [47]; Figure
2D
and Supplementary Figure S3B,D). Moreover, the alignment of this monomer on the structure of the Fis dimer (Figure
2E) shows a good overall fit suggesting that NinH resembles Fis in quaternary structure.
NinH dimer formation
Fis is a dimer in solution allowing the HTH domain from each subunit to dock within the major groove of duplex DNA [14]. Both native and N-His-tagged NinH proteins were purified to homogeneity (Supplementary Figure S4A,B). The histidine tag could be readily removed using thrombin protease (Supplementary Figure S4B). Stable protein dimers can survive separation in SDS-polyacrylamide gels if reducing agents are omitted and the sample is not boiled [48]. His-NinH was analyzed using this approach and a dimer band of ∼20 kDa was evident in Coomassie-stained gels (Supplementary Figure S4B), which was confirmed by Western
©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
1351
blotting using anti-His antibodies (Supplementary Figure S4C). Dynamic light scattering was also performed with His-NinH, which showed an increase in particle size with increasing protein concentration, also consistent with dimerization (Supplementary Figure S4D). Finally, His-NinH was subjected to AUC to further help determine its quaternary structure. The His-NinH protein sedimented in a single peak at 20.2 kDa (Supplementary Figure S4E), matching the predicted molecular mass of a homodimer (20 kDa).
NinH DNA binding
The predicted interaction of NinH with DNA was initially evaluated in an electrophoretic mobility shift assay with short 50-32P-radiolabeled substrates (Figure
3A). Purified NinH bound to a 60 bp dsDNA (DS60; Supplementary Table S1) forming two complexes (Figure
3A, lanes 1–3; Supplementary Figure S5), confirming that it is capable of forming a stable association with DNA. To quantify this binding affinity of NinH for dsDNA we utilized fluorescence anisotropy of a 30-fluorescein-tagged 20 bp duplex (DS20; Figure
3B), which yielded an estimated KD of 57 nM for a NinH dimer. Interestingly, the best fit for the data is produced with a Hill slope coefficient of 2.23, implying cooperativity of binding to DNA. Similar results were obtained with His-NinH binding to DS20 in yielding a KD of 43 nM and a Hill slope coefficient of 1.985. DNA substrates that resemble branched molecules that might occur as intermediates during λ DNA metabolism, a replication forklike structure and a four-stranded Holliday junction were also tested, but no significantly improved His-NinH binding was detected with complexes appearing simply to assemble on available duplexes (Supplementary Figure S5A). Similarly, no enhanced affinity was observed with bubble DNA substrates containing centrally located mismatches of 1, 5, 13 and 20 nt (Supplementary Figure S5B).
Figure 3. DNA binding by NinH protein.
(A) Gel mobility shift assays contained 50 or 500 nM NinH protein and 0.15 nM of 32P-labeled 60 bp (DS60) dsDNA (lanes 1–3) and bent DNA (BT1–3) containing 1 (lanes 4–6), 2 (lanes 7–9) or 3 (lanes 10–12) additional adenine residues in one of the DNA strands. (B) NinH and His-NinH binding to 10 nM fluorescein-labeled 20 bp dsDNA (DS20) as determined by fluorescence anisotropy. Data are the mean and standard deviation of two (His-NinH) and three (NinH) independent experiments. (C,D) NinH binding to linear dsDNA and bent DNA. Binding reactions contained 0.15 nM DNA with 0.781, 1.562, 3.125, 6.25, 12.5, 25, 50 and 100 nM protein. Data are the mean and standard deviation of three independent experiments. The relative proportion of the two protein–DNA
complexes
was
determined
by
ImageJ
[68] analysis of phosphorimaged gels.
1352 ©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
Since Fis is known to bend duplex DNA [14,16,17], we explored the possibility that NinH might bind with a higher affinity to a duplex that already contained a kink. Three bent dsDNA substrates (BT1–3) were generated by the addition of extra adenine residues at the center of one strand of the duplex (Supplementary Table S1), causing it to kink to accommodate the extra nucleotide. Adenine insertions of 3 nt are known to introduce a bend of ∼60° in duplex DNA [49]. NinH bound to all three of the bent DNA structures tested, forming two discrete complexes (Figure
3A, lanes 4–12). There was little difference in apparent binding affinity between DS60 and BT1–3, although the bent DNA did allow more of the second complex to form (Figure
3C,D). The complexes observed in these gel shift assays most probably represent the initial assembly of a single NinH dimer followed by the cooperative loading of a second dimer, a feature apparently favored when the duplex already contains a bend.
NinH DNA bending
In keeping with its similarities to Fis, we next examined the capacity of NinH to induce a bend in linear duplex DNA. A 20 bp substrate (DS20), matching that employed in fluorescence anisotropy binding assays, was generated by annealing oligonucleotides labeled with either tetramethylrhodamine or fluorescein at the 30 end. ssFRET was determined from the increase in sensitized acceptor emission. The energy transfer efficiency increases upon the addition of NinH (Figure
4A) consistent with a shortening of the inter-fluorophore distance. No significant difference in FRET was observed between controls without protein or with BSA. The divergence between NinH and the controls is particularly evident in the raw data for fluorescence intensity (Supplementary Figure S6). Energy transfer was independently assayed by time-resolved fluorescence spectroscopy. As with the steady-state assay, energy transfer efficiency also increased upon the addition of NinH, with a more substantial increase noted
(Figure
4B) compared with the BSA and protein-free controls. From the energy transfer efficiencies, donor to acceptor distances were calculated (Supplementary Table S2). These data clearly show a decrease in donor to acceptor distance in the presence of NinH. Accordingly, the mean percent-age change in energy transfer distance was calculated as −4.63 for the steady-state assay and −8.03 for the time-resolved. The discrepancy in the absolute values of R calculated for the two methodologies has been noted previously
and
many
explanations
have
been
proposed
[50,51]. Both steady-state and time-resolved data were employed in calculations to estimate the extent of bending induced by NinH binding to duplex DNA (Supplementary Table S3).
Figure 4. Bending of dsDNA by NinH. (A,B) Energy transfer was determined in steady state and lifetime FRET assays. Reactions contained 1 μM NinH protein or controls without protein or addition of 3 mM BSA with 50 nM DNA and were incubated for 10 min at room temperature before analysis. Experiments were performed in triplicate; error bars indicate the standard error of the mean. Negative controls have been used previously [69].
©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
1353
The calculations based on the two methods yielded values of 46° and 35° for steady-state and time-resolved FRET, respectively (Supplementary Table S3). Averaging these values gives a bend of 41° which is somewhat lower than the 50–90° estimated for Fis in gel and footprinting experiments [16,52] or 65° from the crystal structure of Fis bound to a high-affinity DNA site [14]. Thus NinH resembles Fis in being capable of bending dsDNA, although the reduced bend with NinH suggests a difference in the arrangement of the dimer in its association with duplex DNA.
NinH sequence specificity
To determine whether NinH displays any preference for particular nucleotide sequences we first conducted universal PBM assays, which comprehensively survey 10-base sequences and provide robust estimates for binding to all possible 8-mers, since each is measured in multiple contexts [30]. An N-terminal GST-NinH fusion was constructed under the control of the vector T7 promoter and protein produced in vitro. Preferential sequence binding with the fusion protein was performed in duplicate on two microarrays with differing dsDNA probe sequences (denoted ME and HK; [31]). The results on both arrays were consistent with established metrics for sequence-specific binding ([30]; Supplementary Figure S7), but the highest-scoring sequences were dissimilar between the two arrays. A position-specific scoring matrix (PSSM) sequence logo of the top ten 8-mers recovered from each array is shown in Figure
5A.
The differing results from the two arrays suggest that each array samples from a larger binding site: while the arrays comprehensively sample 10-base sequences, their sampling of larger sequences (typical of bacterial transcription factors [53]) is much sparser and may not capture full binding sites; the Fis footprint, for example, is 19–21 base pairs [14]. High-throughput selection (HT-SELEX) was, therefore, used in an effort to more clearly define the sequence composition of any NinH DNA binding motif. The results revealed a degenerate motif with a clear 50-G-N13-C-30 signature (Figure
5B). The palindromic nature of the motif fits with the assembly of a NinH homodimer with each HTH domain engaging with the major groove. Moreover, the larger motif detected by HT-SELEX can also explain the PBM data, revealing that the ME and HK motifs correspond
Figure 5. Identification of the NinH DNA binding motif.
(A) The GST-NinH binding site obtained by PBM using HK and ME microarrays. (B) The GST-NinH binding motif determined by HT-SELEX. (C) Published Fis DNA binding site based on SELEX, ChIP and EMSA data [3,13–15]. Residues below the motif correspond to those that exert a negative effect on Fis binding. Motif residues in (A) and (B) are numbered and aligned according to the Fis recognition sequence shown in (C) and are also highlighted in Figure
1D. Motif alignments were performed manually.
1354 ©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
to the left and right ends of the NinH recognition sequence (Figure
5A). Remarkably, in several aspects, the motif closely resembles the Fis DNA binding site (Figure
5C) identified by a combination of SELEX, CHiP and gel shift experiments [3,13-15]. Both NinH and Fis binding sites have G and C separated by 13 nt, plus additional matches at −8/8, −4/4 and −3/3 (Figure
5B,C). Thus the homology model of NinH based on Fis and supported by SAXS data (Figure
2) fits with a similar arrangement in the two proteins with specific contacts with guanines at each end of the motif (Figure
1D). However, there are differences in the motifs between Fis and NinH in the core 5 bp from −2to 2 (Figure
5B,C). In the Fis motif, this region is AT-rich [13,14], whereas NinH shows a distinctly different signature, notably with a G or C at the −1 position and no preference for an A or T at position 2 (Figure
5B). Given the differences in the central core recognition sequences, it is possible that NinH and Fis do not assemble preferentially at precisely the same sites although it is likely that there will be overlap in those they favor.
NinH site-directed mutants
To further probe the similarities between NinH and Fis we selected many residues for replacement by alanine to investigate regions of NinH implicated in DNA binding and dimerization (Figure
6A). We initially selected conserved residues corresponding to the HTH of Fis, namely N18, R30, T32, R34 and K35 matching Fis N73, R85, T87 R89 and K90, respectively (Figures
1
and 6A). S29 was also chosen as it corresponds to Fis N84, the only residue involved in DNA contact that is not conserved between Fis and NinH (Figure
1). Additional sub-stitution mutants (H45A, H55A and R56A) were generated in the predicted β sheets at the C-terminus of
©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
1355
NinH proposed to be involved in dimerization. A C-terminal deletion (ΔC25, removing residues 44–68) and two double mutants (S29A R30A and H55A R56A) were also constructed (Supplementary Figure S1). All of the N-His-tagged NinH proteins were purified to homogeneity, with the exception of R34A and ΔC25 which could not be retrieved due to insolubility.
DNA binding mutants
N-terminal mutants affecting the predicted DNA binding region of NinH were initially checked for their ability to
form
homodimers
in
solution
using
AUC
(Figure
6B). All of these mutants form dimers, although N18A shows an additional peak at 8–10 kDa consistent with the formation of a small proportion of monomer. We proceeded to investigate whether any of these mutants were impaired in DNA binding. A gel shift assay with a Cy5-labeled 20 bp dsDNA sequence (Cy520) that matches the NinH consensus sequence derived from HT-SELEX experiments (Figure
5B) was employed. The His-NinH wt protein bound this DNA with a KD in the sub-micromolar range (Figure
6C, lanes 1–4). This binding affinity was further quantified using MST, which yielded a KD of 5.6 nM (Table
1
and Supplementary Figure S8). Both gel shift and MST experiments were repeated with the NinH site-directed mutants and key residues involved in DNA binding identified (Table
1, Figure
6D
and Supplementary Figure S9). The most important residue for DNA binding appears to be T32 (Table
1, Figure
6C
and Supplementary Figure S8C,D), equivalent to T87 in Fis (Figure
1D) where it is known to contact a phosphate on the DNA backbone and is similarly critical for DNA binding [14,45]. The T32A mutant also exhibited a defect in the bending of the DS20 substrate (Supplementary Figure S10). Other NinH mutants with significant deficiencies in DNA binding were R30A (Fis R85), K35A (Fis K90) and the double mutant S29A R30A, all increasing the KD by
at
least
three
orders
of
magnitude
(Table
1). The N18A (Fis
N73)
had
an
intermediate
effect
on
DNA
binding
(Figure
6D). Mutation of S29 (Fis N84) to alanine had only a minor effect on DNA binding (Table
1). A Fis N84A mutant shows reduced binding to λ attR, but is largely unaffected in binding to fis P II and hin distal
sites
[45]. All of these NinH mutants exhibited DNA deficiencies that broadly match the effects conferred by their corresponding mutants in the αC and αD helices of the Fis HTH (Table
1;[45]).
Mutants affecting dimerization
Three substitution mutants, H45A, H55A and R56A, predicted to be involved in dimerization were also purified and analyzed by AUC (Figure
6B). H45A and H55A produced a single peak at 8–10 kDa consistent with
Table 1. DNA binding of His-NinH mutants
NinH mutant KD on Cy520 Fis mutant KD on λ attR
wt 5.6 ± 1.5 wt 4.1 ± 0.4
N18A >220 N73A 91 ± 34
S29A 13 ± 2.7 N84A 157 ± 15
S29A R30A >5500 — —
R30A >9900 R85A >2200
T32A No binding T87A 1300 ± 545
K35A No binding K90A >2200
H45A >350
H55A >2500
R56A No binding
H55A R56A No binding
Dissociation constants (KD) in nM were determined by microscale thermophoresis and are the mean and standard deviation of three independent experiments. Binding mixtures contained 10 nM Cy520 DNA. No DNA binding was detected for mutants T32A, K35A and R56A under these conditions. Similarly, only weak binding was found with H55A R56A and the binding isotherm could not be used to estimate a KD. The λ attR binding data on Fis mutants equivalent to residues mutated in NinH are taken from Feldman et al. [45].
1356 ©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
the mass of a NinH monomer. These results are supportive of the structural model proposed for NinH whereby the C-terminus contributes to intersubunit interactions. In contrast, the R56A mutant located at the distal end of β1/β10 (Figure
2B) was able to form a stable dimer (Figure
6B), but was unable to associate with DNA (Figure
6D). It may be that the R56A mutation affects the optimal positioning of the two HTH domains. H55A also bound poorly to DNA, but H45A, which is a monomer in solution, retained significant binding capability (Table
1; Figure
6D; Supplementary Figure S9). It may be that the presence of DNA brings individual monomers in close enough proximity to promote dimerization. The H55A R56A double mutant displayed very poor DNA binding as with the two individually mutated NinH proteins (Figure
6D).
NinH toxicity
We next examined the effect of His-NinH wt and mutants on E. coli BL21-AI by plating dilutions of exponentially growing cells on media containing both arabinose and IPTG. Under conditions which induced overexpression, both His-NinH wt constructs severely inhibited bacterial growth when compared with either the pET28a or pET14b vector controls (Figure
7). In the absence of the two inducer molecules, cells carrying each of the plasmids yielded a normal growth phenotype (Figure
7). We considered it likely that NinH binding to
©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
1357
bacterial DNA was responsible for toxicity by interference with normal chromosome duplication and segregation. If this is correct, then bacterial growth inhibition could be used to monitor DNA binding of the NinH mutants in vivo. High-level production of all of the mutant NinH proteins was confirmed following the activation of gene expression (Supplementary Figure S11). In these toxicity assays, we found that mutants defective in DNA binding exerted little or no detrimental growth effect (T32A and K35A) whereas those that retained DNA binding, such as S29A, behaved like the wt when induced (Figure
7A). Similarly, N18A which has reduced DNA binding activity displayed an intermediate growth defect (Figure
7A). R34A, which could not be purified, resembled the wt suggesting that this protein retains some DNA binding activity, which would be in keeping with the equivalent Fis mutant, R89A [45]. R30A allowed improved growth but when combined with S29A exhibited a more severe growth effect, which is surprising given that R30A and the double mutant, S29A R30A, both bind DNA poorly. There may be a difference in specific versus non-specific binding in some of these mutants that confer differing in vivo effects [45]. Finally, dimerization mutants that were unable to bind DNA, R56A and H55A R56A, did not cause a growth defect (Figure
7B). H45A, which exhibits an intermediate level of DNA binding was toxic. H55A does not bind DNA well and did show some improved growth, though it was still relatively weak. The mutant lacking the C-terminus (ΔC25) showed no toxicity consistent with an expected
inability
to
dimerize
or
associate
with
DNA
(Figure
7B). Overall, the toxicity assays align well with the DNA binding results, linking affinity with DNA substrates in vitro and chromosomal DNA in vivo.
Discussion
The function of the previously uncharacterized phage λ NinH protein was investigated and our results reveal that it resembles bacterial Fis in predicted structure, ability to form homodimers and in bending double-stranded DNA. The conservation of primary sequence between NinH and Fis DNA binding domains suggests that both bind and bend DNA in an analogous fashion. Non-specific DNA binding was apparent with NinH, a feature shared by Fis and other NAPs [14,15], while no increased affinity for bubble, fork or Holliday junction DNA structures was detected. Dual assembly of NinH on duplexes that already contain a kink was favored, potentially overcoming an energetic barrier to bending; it is possible that these substrates more closely resemble the end product of conventional binding by the phage protein. Intrinsic DNA bending does not appear to con-tribute significantly to the recognition of preferred sites by Fis [14], though flexibility in residues flanking the 15 bp consensus is important [15]. Other NAPs, notably IHF, HU and H-NS, are known to preferentially associate with curved, bent and distorted DNA structures [54,55]. Using a FRET assay, the extent of DNA bending induced by NinH on a non-cognate site was found to be ∼41°, 24° less than that imposed by Fis on its preferred consensus sequence [14].
NinH shared a preference with E. coli Fis for a pseudo-palindromic consensus binding site containing G and C residues flanking a 13 bp core. Although within the central region their motifs differ slightly, there is a clear pyrimidine–purine pairing at positions −4/4 and −3/3
(Figure
5) that is similar to that also seen with Fis and helps binding by its enhanced propensity for bending [14]. Fis prefers sites that have an AT-rich core, with the highest affinity for a central A4–6 tract [3,14]. In contrast, the NinH motif has G/C at −1 and no preference for any residue at
position
2
(Figure
5). This variation in the central motifs probably reflects preferences for sequences that affect localized DNA bending and minor groove width, known to influence Fis binding affinity [15]. It is possible that NinH makes base-specific contact with residues in this region, although this would mean that the phage protein has a radically different configuration relative to Fis. The differences in Fis and NinH motifs may not be significant and both proteins could potentially compete for the same nucleotide sequence targets [14,15]; Asn84 in Fis (Ser29 in NinH) makes base-specific contacts at −4/4 but the consensus sequences match at this position (Figure
5B,C); neither this residue nor other internal nucleotide sequences are critical for Fis binding at high-affinity sites [14,45]. In keeping with this, a NinH S29A mutant displayed only a modest reduction in DNA binding to a sequence matching the NinH consensus. Given that NinH and Fis must differ in the way that they assemble as homodimers, it is remarkable that the signature sequences they recognize are not more divergent.
Sequence alignments and structural predictions combined with DNA binding analyses to support a model whereby the N-terminal domain of NinH consists of three α helices, analogous to αB, αC and αD in Fis that comprise the HTH domain (Figure
1). The structural model of the predicted NinH DNA binding region was confirmed experimentally using site-directed mutants which confirmed the importance of conserved residues equivalent to those known to be responsible for DNA binding by Fis [14,45]. These results also correlated with in vivo toxicity, with mutants defective in DNA binding lacking the growth inhibition phenotype associated
1358 ©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
with NinH overexpression. The ensemble in silico models we generated show a similarly conserved secondary structure. The tertiary structure of α helical regions is also retained but the positioning of the C-terminal regions on one monomer relative to the DNA binding regions and relative to the same regions of the partner monomer were found to vary model-to-model. While we chose a ‘closed’ conformation of the C-terminal regions consistent with it forming a dimer interface and found that this fits the SAXS data well, alternative con-formations cannot be ruled out. Furthermore, higher resolution data will be required to place the C-terminal regions unambiguously. Biophysical studies confirmed that NinH is a homodimer in solution and that residues outside of the DNA binding region appear to contribute directly or indirectly to dimerization. The structure of the C-terminal region of NinH implicated in dimer formation has proven more challenging to predict, particularly in its spatial relationship with the N-terminal region. All predictive models suggest that this region con-tains β sheets. Mutation of conserved residues predicted to lie in these β sheets (residues 45 and 55) resulted in a shift in the monomer-dimer equilibrium towards monomer consistent with a role in the dimer interface as suggested by its position in our favored model. A similar shift is seen for the protein with a mutation at residue 45 but on our model this does not appear to be at the dimer interface, implying an indirect effect.
The remarkable similarities in NinH and Fis DNA binding and bending, mean that they have the potential to be functionally equivalent factors affecting several aspects of λ and E. coli biology. Fis fulfills an accessory role in excision of the λ prophage by manipulating attL and attR into an appropriate position for cleavage by integrase (Int) in complex with Xis bound to attR[6,7,11,56]. Fis binding at this site also contributes to phage integration [9–11]. NinH is expressed early, along with the other ninR genes exclusively during the λ lytic cycle under PR transcriptional control, with trace levels appearing 2 min after prophage induction and high-level expression from 5 to 20 min [57]. This pattern of expression mirrors that of int and xis which are transcribed from PL [57], meaning that NinH could participate in the excision process in addition to integration. It is plausible that NinH and Fis can compete for binding at attRtoinfluence the efficiency of these site-specific recombination reactions. It may, therefore, be significant that the Fis binding site in attR contains an A6-tract (position -3 to 2) rather than the 50-AAGAAN-30 signature found with NinH (Figure
5). Following phage induction, Fis might be expected to predominate as it is abundant in E. coli during early exponential growth and is already available in the cell. However, the physiological condition of the cell might be a contributing factor since Fis expression is barely detectable during the stationary phase [58]. It is also feasible that NinH exerts influence on bacterial cells by interfering with the expression of genes, especially since Fis is a major factor in global transcriptional regulation, affecting the expression of 923 genes [3]. NinH could compete with Fis to affect specific regulatory pathways, or even nucleoid architecture, with specific advantages for phage productivity. Interestingly, a precise ninH gene deletion results in a phage that forms small plaques when grown on E. coli [59], implying an important role in maximizing phage yield or efficiency of lysis.
Several examples of NAP-like functions have previously been found associated with extrachromosomal elements [60], including Bacillus subtilis phage f29 p6 [61] and homologs of H-NS present in phage EPV1 [62] and conjugative plasmids R446 [63]
and
pSf-R27
[64]. H-NS is primarily a repressor of gene expression [65] and selfish DNA elements carrying versions of the protein may be harnessed to down-regulate bacterial defense genes to promote their persistence [64]. For example, the plasmid-encoded Sfh protein resembles H-NS and StpA [66]. The occurrence of phage-encoded anti-repressors of H-NS [67] also supports the notion that phages undertake a variety of approaches to modulate host NAPs and subvert host regulatory pathways to promote their own survival. NinH is the first example to be identified of a Fis-like NAP carried by bacteriophages. Further work is needed to determine the precise role of NinH in both phage and bacterial metabolism.
Competing Interests
The authors declare that there are no competing interests associated with the manuscript.
Funding
This work was supported by Biotechnology and Biological Sciences Research Council grants BB/C514523/1 and BB/FO20503/1 to GS. SC was funded by the Homing programme of the Foundation for Polish Science co-financed by the European Union under the European Regional Development Fund, grant No. Homing/2017-3/
22. SS was funded by the TEAM programme of the Foundation for Polish Science co-financed by the European Union under the European Regional Development Fund (TEAM/2016-3/23). JGH was funded by the National Science Centre (NCN, Poland) grant no. 2016/21/B/CC1/00274 (OPUS-11). AWHY was funded by CIHR Foundation Grant (FDN-138403) to TRH. AJ was supported by Swedish Research Council ‘‘Vetenskapsrådet’’
©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
1359
postdoctoral grant (2016-00158). MK was supported by the National Science Centre (NCN, Poland) grant no. 2017/27/B/ST4/00485.
Author Contributions
G.J.S. and J.G.H. designed and guided experiments and G.J.S. performed initial bioinformatics analysis.
S.C. and S.S. purified NinH proteins and carried out biochemical experiments, D.B. carried out in silico modeling, J.P., M.T. and M.K. carried out SAXS experiments and analyzed SAXS results. F.A.C. and
L.Y.B. cloned and purified NinH proteins. A.J.T. and W.H.G. performed DNA binding experiments and
W.H.G. performed DNA bending experiments. W.H.G. and L.-O.P. analyzed and interpreted the DNA bending data. A.W.H.Y. and A.J. performed PBM and HT-SELEX experiments and analyzed results with T.R.H. J.R.P. carried out NinH toxicity and overexpression experiments. All authors contributed to writing the manuscript.
Acknowledgements
The authors are grateful to the core instrumentation facility of the International Institute of Molecular and Cell Biology, Warsaw and Roman Szczepanowski for their generous help in analytical ultracentrifugation measurements.
Abbreviations
AUC, analytical ultracentrifugation; FWHM, full width half maximum; HTH, helix-turn-helix; IRF, instrument response function; NAPs, nucleoid-associated proteins; PBM, protein binding microarray; SAXS, small angle X-ray scattering; ssFRET, steady-state Fster resonance energy transfer; trFRET, time-resolved Fster resonance energy transfer.
References
1 Schneider, R., Lurz, R., Luder, G., Tolksdorf, C., Travers, A. and Muskhelishvili, G. (2001) An architectural role of the Escherichia coli chromatin protein Fis in organising DNA. Nucl. Acids Res. 29, 5107–5114 https://doi.org/10.1093/nar/29.24.5107
2 Dillon, S.C. and Dorman, C.J. (2010) Bacterial nucleoid-associated proteins, nucleoid structure and gene expression. Nat. Rev. Microbiol. 8, 185–195 https://doi.org/10.1038/nrmicro2261
3 Cho, B.K., Knight, E.M., Barrett, C.L. and Palsson, B.O. (2008) Genome-wide analysis of Fis binding in Escherichia coli indicates a causative role for A-/ AT-tracts. Genome Res. 18, 900–910 https://doi.org/10.1101/gr.070276.107
4 Duprey, A., Reverchon, S. and Nasser, W. (2014) Bacterial virulence and Fis: adapting regulatory networks to the host environment. Trends Microbiol. 22,92–99 https://doi.org/10.1016/j.tim.2013.11.008
5 Johnson, R.C. (2015) Site-specific DNA inversion by serine recombinases. Microbiol. Spectr. 3,1–36 https://doi.org/10.1128/microbiolspec.
MDNA3-0047-2014
6 Ball, C.A. and Johnson, R.C. (1991) Multiple effects of Fis on integration and the control of lysogeny in phage lambda. J. Bacteriol. 173, 4032–4038
https://doi.org/10.1128/JB.173.13.4032-4038.1991
7 Thompson, J.F., Moitoso de Vargas, L., Koch, C., Kahmann, R. and Landy, A. (1987) Cellular factors couple recombination with growth phase: characterization of a new component in the lambda site-specific recombination pathway. Cell 50, 901–908 https://doi.org/10.1016/0092-8674(87)
90516-2
8 Hancock, S.P., Cascio, D. and Johnson, R.C. (2019) Cooperative DNA binding by proteins through DNA shape complementarity. Nucl. Acids Res. 47, 8874–8887 https://doi.org/10.1093/nar/gkz642
9 Ball, C.A. and Johnson, R.C. (1991) Efficient excision of phage lambda from the Escherichia coli chromosome requires the Fis protein. J. Bacteriol. 173, 4027–4031 https://doi.org/10.1128/JB.173.13.4027-4031.1991
10 Esposito, D. and Gerard, G.F. (2003) The Escherichia coli Fis protein stimulates bacteriophage lambda integrative recombination in vitro. J. Bacteriol. 185, 3076–3080 https://doi.org/10.1128/JB.185.10.3076-3080.2003
11 Landy, A. (2015) The lambda integrase site-specific recombination pathway. Microbiol. Spectr. 3, MDNA3-0051-2014 https://doi.org/10.1128/
microbiolspec.MDNA3-0051-2014
12 Skoko, D., Yoo, D., Bai, H., Schnurr, B., Yan, J., McLeod, S.M. et al. (2006) Mechanism of chromosome compaction and looping by the Escherichia coli nucleoid protein Fis. J. Mol. Biol. 364, 777–798 https://doi.org/10.1016/j.jmb.2006.09.043
13 Shao, Y., Feldman-Cohen, L.S. and Osuna, R. (2008) Functional characterization of the Escherichia coli Fis-DNA binding sequence. J. Mol. Biol. 376, 771–785 https://doi.org/10.1016/j.jmb.2007.11.101
14 Stella, S., Cascio, D. and Johnson, R.C. (2010) The shape of the DNA minor groove directs binding by the DNA-bending protein Fis. Genes Dev. 24, 814–826 https://doi.org/10.1101/gad.1900610
15 Hancock, S.P., Stella, S., Cascio, D. and Johnson, R.C. (2016) DNA sequence determinants controlling affinity, stability and shape of DNA complexes bound by the nucleoid protein Fis. PLoS One 11, e0150189 https://doi.org/10.1371/journal.pone.0150189
16 Pan, C.Q., Finkel, S.E., Cramton, S.E., Feng, J.A., Sigman, D.S. and Johnson, R.C. (1996) Variable structures of Fis–DNA complexes determined by flanking DNA–protein contacts. J. Mol. Biol. 264, 675–695 https://doi.org/10.1006/jmbi.1996.0669
17 Perkins-Balding, D., Dias, D.P. and Glasgow, A.C. (1997) Location, degree, and direction of DNA bending associated with the Hin recombinational enhancer sequence and Fis-enhancer complex. J. Bacteriol. 179, 4747–4753 https://doi.org/10.1128/JB.179.15.4747-4753.1997
1360 ©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
18 Krer, M. and Hobom, G. (1982) A chain of interlinked genes in the nin R region of bacteriophage lambda. Gene 20,25–38
https://doi.org/10.1016/0378-1119(82)90084-1
19 Ray, D., Ha, K.C., Nie, K., Zheng, H., Hughes, T.R. and Morris, Q.D. (2017) RNAcompete methodology and application to determine sequence preferences of unconventional RNA-binding proteins. Methods 118–119,3–15 https://doi.org/10.1016/j.ymeth.2016.12.003
20 Webb, B. and Sali, A. (2014) Comparative protein structure modeling using MODELLER. Curr. Protoc. Bioinformatics 47, 5.6.1–5.6.30 https://doi.org/10.1002/0471250953.bi0506s47
21 Dominguez, C., Boelens, R. and Bonvin, A.M. (2003) HADDOCK: a protein-protein docking approach based on biochemical or biophysical information. J. Am. Chem. Soc. 125, 1731–1737 https://doi.org/10.1021/ja026939x
22 van Zundert, G.C., Rodrigues, J.P., Trellet, M., Schmitz, C., Kastritis, P.L., Karaca, E. et al. (2016) The HADDOCK2.2 web server: user-friendly integrative modeling of biomolecular complexes. J. Mol. Biol. 428, 720–725 https://doi.org/10.1016/j.jmb.2015.09.014
23 Heo, L., Lee, H. and Seok, C. (2016) Galaxyrefinecomplex: refinement of protein–protein complex model structures driven by interface repacking. Sci. Rep. 6, 32153 https://doi.org/10.1038/srep32153
24 Heo, L., Park, H. and Seok, C. (2013) Galaxyrefine: protein structure refinement driven by side-chain repacking. Nucl. Acids Res. 41, W384–W388 https://doi.org/10.1093/nar/gkt458
25 Lovell, S.C., Davis, I.W., Arendall, W.B., de Bakker, P.I., Word, J.M., Prisant, M.G. et al. (2003) Structure validation by Cα geometry: f, ψ and Cβ deviation. Proteins 50, 437–450 https://doi.org/10.1002/prot.10286
26 Girardot,
R.,
Viguier,
G.,
Ounsy,
M.
and
Pérez,
J.
(2010-2017)
Foxtrot,
https://www.synchrotron-soleil.fr/en/beamlines/swing
27 Konarev, P.V., Volkov, V.V., Sokolova, A.V., Koch, M.H. and Svergun, D.I. (2003) PRIMUS: a Windows PC-based system for small-angle scattering data analysis. J. Appl. Cryst. 36, 1277–1282 https://doi.org/10.1107/S0021889803012779
28 Svergun, D.I. (1992) Determination of the regularization parameter in indirect-transform methods using perceptual criteria. J. Appl. Cryst. 25, 495–503 https://doi.org/10.1107/S0021889892001663
29 Franke, D. and Svergun, D.I. (2009) DAMMIF, a program for rapid ab-initio shape determination in small-angle scattering. J. Appl. Cryst. 42, 342–346
https://doi.org/10.1107/S0021889809000338
30 Berger, M.F., Philippakis, A.A., Qureshi, A.M., He, F.S., Estep, III, P.W. and Bulyk, M.L. (2006) Compact, universal DNA microarrays to comprehensively determine transcription-factor binding site specificities. Nat. Biotechnol. 24, 1429–1435 https://doi.org/10.1038/nbt1246
31 Badis, G., Chan, E.T., van Bakel, H., Pena-Castillo, L., Tillo, D., Tsui, K. et al. (2008) A library of yeast transcription factor motifs reveals a widespread function for Rsc3 in targeting nucleosome exclusion at promoters. Mol. Cell 32, 878–887 https://doi.org/10.1016/j.molcel.2008.11.020
32 Jolma, A., Kivioja, T., Toivonen, J., Cheng, L., Wei, G., Enge, M. et al. (2010) Multiplexed massively parallel SELEX for characterization of human transcription factor binding specificities. Genome Res. 20, 861–873 https://doi.org/10.1101/gr.100552.109
33 Jolma, A., Yan, J., Whitington, T., Toivonen, J., Nitta, K.R., Rastas, P. et al. (2013) DNA-binding specificities of human transcription factors. Cell 152, 327–339 https://doi.org/10.1016/j.cell.2012.12.009
34 Clegg, R.M. (1992) Fluorescence resonance energy transfer and nucleic acids. Methods Enzymol. 211, 353–388 https://doi.org/10.1016/0076-6879
(92)11020-J
35 Stuhmeier, F., Welch, J.B., Murchie, A.I., Lilley, D.M. and Clegg, R.M. (1997) Global structure of three-way DNA junctions with and without additional unpaired bases: a fluorescence resonance energy transfer analysis. Biochemistry 36, 13530–13538 https://doi.org/10.1021/
bi9702445
36 Lorenz, M., Hillisch, A., Payet, D., Buttinelli, M., Travers, A. and Diekmann, S. (1999) DNA bending induced by high mobility group proteins studiedby fluorescence resonance energy transfer. Biochemistry 38, 12150–12158 https://doi.org/10.1021/bi990459+
37 Grinvald, A. and Steinberg, I.Z. (1974) On the analysis of fluorescence decay kinetics by the method of least-squares. Anal. Biochem. 59, 583–598 https://doi.org/10.1016/0003-2697(74)90312-1
38 Dodd, I.B. and Egan, J.B. (1990) Improved detection of helix-turn-helix DNA-binding motifs in protein sequences. Nucl. Acids Res. 18, 5019–5026
https://doi.org/10.1093/nar/18.17.5019
39 Narasimhan, G., Bu, C., Gao, Y., Wang, X., Xu, N. and Mathee, K. (2002) Mining protein sequences for motifs. J. Comput. Biol. 9, 707–720
https://doi.org/10.1089/106652702761034145
40 Finn, R.D., Bateman, A., Clements, J., Coggill, P., Eberhardt, R.Y., Eddy, S.R. et al. (2014) Pfam: the protein families database. Nucl. Acids Res. 42, D222–D230 https://doi.org/10.1093/nar/gkt1223
41 Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W. et al. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucl. Acids Res. 25, 3389–3402 https://doi.org/10.1093/nar/25.17.3389
42 Kelley, L.A., Mezulis, S., Yates, C.M., Wass, M.N. and Sternberg, M.J. (2015) The Phyre2 web portal for protein modeling, prediction and analysis. Nat. Protoc. 10, 845–858 https://doi.org/10.1038/nprot.2015.053
43 Yang, J., Yan, R., Roy, A., Xu, D., Poisson, J. and Zhang, Y. (2015) The I-TASSER suite: protein structure and function prediction. Nat. Methods 12, 7–8 https://doi.org/10.1038/nmeth.3213
44 Morett, E. and Bork, P. (1998) Evolution of new protein function: recombinational enhancer Fis originated by horizontal gene transfer from the transcriptional regulator NtrC. FEBS Lett. 433, 108–112 https://doi.org/10.1016/S0014-5793(98)00888-6
45 Feldman-Cohen, L.S., Shao, Y., Meinhold, D., Miller, C., Colon, W. and Osuna, R. (2006) Common and variable contributions of Fis residues to high-affinity binding at different DNA sequences. J. Bacteriol. 188, 2081–2095 https://doi.org/10.1128/JB.188.6.2081-2095.2006
46 Drozdetskiy, A., Cole, C., Procter, J. and Barton, G.J. (2015) JPred4: a protein secondary structure prediction server. Nucl. Acids Res. 43, W389–W394 https://doi.org/10.1093/nar/gkv332
47 Svergun, D., Barberato, C. and Koch, M.H. (1995) CRYSOL: a program to evaluate X-ray solution scattering of biological macromolecules from atomic coordinates. J. Appl. Crystallogr. 28, 768–773 https://doi.org/10.1107/S0021889895007047
48 Curtis, F.A., Malay, A.D., Trotter, A.J., Wilson, L.A., Barradell-Black, M.M., Bowers, L.Y. et al. (2014) Phage ORF family recombinases: conservation of activities and involvement of the central channel in DNA binding. PLoS One 9, e102454 https://doi.org/10.1371/journal.pone.0102454
49 Gohlke, C., Murchie, A.I., Lilley, D.M. and Clegg, R.M. (1994) Kinking of DNA and RNA helices by bulged nucleotides observed by fluorescence resonance energy transfer. Proc. Natl. Acad. Sci. U.S.A. 91, 11660–11664 https://doi.org/10.1073/pnas.91.24.11660
©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).
1361
50 Vamosi, G., Gohlke, C. and Clegg, R.M. (1996) Fluorescence characteristics of 5-carboxytetramethylrhodamine linked covalently to the 50 end of oligonucleotides: multiple conformers of single-stranded and double-stranded dye-DNA complexes. Biophysical. J. 71, 972–994 https://doi.org/10.1016/
S0006-3495(96)79300-1
51 Parkhurst, K.M. and Parkhurst, L.J. (1995) Kinetic studies by fluorescence resonance energy transfer employing a double labeled oligonucleotide: hybridization to the oligonucleotide complement and to single-stranded DNA. Biochemistry 34, 285–292 https://doi.org/10.1021/bi00001a035
52 Perkins, T.T., Dalal, R.V., Mitsis, P.G. and Block, S.M. (2003) Sequence-dependent pausing of single lambda exonuclease molecules. Science 301, 1914–1918 https://doi.org/10.1126/science.1088047
53 Wunderlich, Z. and Mirny, L.A. (2009) Different gene regulation strategies revealed by analysis of binding motifs. Trends Genet. 25, 434–440 https://doi.org/10.1016/j.tig.2009.08.003
54 Yamada, H., Muramatsu, S. and Mizuno, T. (1990) An Escherichia coli protein that preferentially binds to sharply curved DNA. J. Biochem. 108, 420–425 https://doi.org/10.1093/oxfordjournals.jbchem.a123216
55 Swinger, K.K. and Rice, P.A. (2007) Structure-based analysis of HU-DNA binding. J. Mol. Biol. 365, 1005–1016 https://doi.org/10.1016/j.jmb.2006.
10.024
56 Papagiannis, C.V., Sam, M.D., Abbani, M.A., Yoo, D., Cascio, D., Clubb, R.T. et al. (2007) Fis targets assembly of the Xis nucleoprotein filament to promote excisive recombination by phage lambda. J. Mol. Biol. 367, 328–343 https://doi.org/10.1016/j.jmb.2006.12.071
57 Liu, X., Jiang, H., Gu, Z. and Roberts, J.W. (2013) High-resolution view of bacteriophage lambda gene expression by ribosome profiling. Proc. Natl. Acad. Sci. U.S.A. 110, 11928–11933 https://doi.org/10.1073/pnas.1309739110
58 Azam, T.A., Iwata, A., Nishimura, A., Ueda, S. and Ishihama, A. (1999) Growth phase-dependent variation in protein composition of the Escherichia coli nucleoid. J. Bacteriol. 181, 6361–6370 https://doi.org/10.1128/JB.181.20.6361-6370.1999
59 Tarkowski, T.A., Mooney, D., Thomason, L.C. and Stahl, F.W. (2002) Gene products encoded in the ninR region of phage λ participate in Red-mediated recombination. Genes Cells 7, 351–363 https://doi.org/10.1046/j.1365-2443.2002.00531.x
60 Flores-Rios, R., Quatrini, R. and Loyola, A. (2019) Endogenous and foreign nucleoid-associated proteins of bacteria: occurrence, interactions and effects on mobile genetic elements and host’s biology. Comput. Struct. Biotechnol. J. 17, 746–756 https://doi.org/10.1016/j.csbj.2019.06.010
61 Gonzalez-Huici, V., Alcorlo, M., Salas, M. and Hermoso, J.M. (2004) Binding of phage Phi29 architectural protein p6 to the viral genome: evidence for topological restriction of the phage linear DNA. Nucleic Acids Res. 32, 3493–3502 https://doi.org/10.1093/nar/gkh668
62 Skennerton, C.T., Angly, F.E., Breitbart, M., Bragg, L., He, S., McMahon, K.D. et al. (2011) Phage encoded H-NS: a potential achilles heel in the bacterial defence system. PLoS One 6, e20095 https://doi.org/10.1371/journal.pone.0020095
63 Nieto, J.M. and Juarez, A. (1999) The putative Orf4 of broad-host-range conjugative plasmid R446 could be related to the H-NS family of bacterial nucleoid-associated proteins. Plasmid 41, 125–127 https://doi.org/10.1006/plas.1998.1388
64 Doyle, M., Fookes, M., Ivens, A., Mangan, M.W., Wain, J. and Dorman, C.J. (2007) An H-NS-like stealth protein aids horizontal DNA transmission in bacteria. Science 315, 251–252 https://doi.org/10.1126/science.1137550
65 Dorman, C.J. (2004) H-NS: a universal regulator for a dynamic genome. Nat. Rev. Microbiol. 2, 391–400 https://doi.org/10.1038/nrmicro883
66 Beloin, C., Deighan, P., Doyle, M. and Dorman, C.J. (2003) Shigella flexneri 2a strain 2457T expresses three members of the H-NS-like protein family: characterization of the Sfh protein. Mol. Genet. Genomics 270,66–77 https://doi.org/10.1007/s00438-003-0897-0
67 Ali, S.S., Beckett, E., Bae, S.J. and Navarre, W.W. (2011) The 5.5 protein of phage T7 inhibits H-NS through interactions with the central oligomerization domain. J. Bacteriol. 193, 4881–4892 https://doi.org/10.1128/JB.05198-11
68 Abramoff, M.D., Magalhaes, P.J. and Ram, S.J. (2004) Image processing with ImageJ. Biophotonics Int. 11,36–42
https://imagej.nih.gov/ij/docs/pdfs/
Image_Processing_with_ImageJ.pdf
69 Fenyk, S., Townsend, P.D., Dixon, C.H., Spies, G.B., Campillo,.S.E., Slootweg, A. et al. (2015) The potato nucleotide-binding leucine-rich repeat (NLR) immune receptor Rx1 is a pathogen-dependent DNA-deforming protein. J. Biol. Chem. 290, 24945–24960 https://doi.org/10.1074/jbc.M115.672121
1362 ©
2020
The
Author(s).
This
is
an
open
access
article
published
by
Portland
Press
Limited
on
behalf
of
the
Biochemical
Society
and
distributed
under
the
Creative
Commons
Attribution
License
4.0
(CC
BY).