| Nucleic Acids Research | Pages |
Role of the conserved lysine 80 in stabilisation ofNF-[kappa]B p50 DNA binding
Introduction
Materials And Methods
Synthesis of oligonucleotides
Site-specific mutagenesis
Expression and purification of NF-[kappa]B p50
Analytical procedures
Partial proteolysis
Gel electrophoresis DNA binding assay
Surface plasmon resonance (SPR)
Molecular modelling
Sequence alignment
Results
Mutant p50 protein production
Chymotrypsin and proteinase K digestion of wild type and triple mutant p50
DNA binding properties of mutants
Discrimination between DNA binding activities of p50 mutants
Discussion
Acknowledgements
References
Role of the conserved lysine 80 in stabilisation ofNF-[kappa]B p50 DNA binding
ABSTRACT
INTRODUCTION
The DNA binding activity of the ubiquitous transcription factor NF-[kappa]B is activated, independently of protein synthesis, when cells are exposed to a variety of extracellular signals (1). NF-[kappa]B binding sites (termed [kappa]B motifs) are found in viral genomes and the promoter/enhancer regions of many cellular genes involved in the regulation of acute-phase response, inflammation, apoptosis, and cell growth or differentiation (2). NF-[kappa]B proteins are held in an inactive form in the cytoplasm by inhibitory proteins (I[kappa]Bs) that bind to the nuclear localisation signals (NLSs) (3-6) of the NF-[kappa]B proteins. After exposure of cells to signals which activate NF-[kappa]B, specific I[kappa]B family members are rapidly phosphorylated, ubiquitinated and degraded. Thus, the active DNA binding NF-[kappa]B is released and translocates to the cell nucleus (7). cDNAs encoding the NF-[kappa]B subunits have been isolated, and it is now clear that these genes belong to a large multigene family (Rel/NF-[kappa]B protein family). All members of this family possess a highly conserved region of ~300 amino acids in their N-termini which is responsible for DNA binding, dimerisation and nuclear localisation of these proteins (3,8) and is known as the NF-[kappa]B/Rel/Dorsal (NRD) region (9). These proteins can form different homodimers or heterodimers, which bind to specific [kappa]B motifs. The Rel/NF-[kappa]B protein family can be conveniently divided into two subgroups: p50 related and rel related. The proteins that belong to the p50 related subfamily are p50 and p52 (10,11). They are synthesised as precursor molecules (p105 and p100 for p50 and p52, respectively) which contain seven `ankyrin repeats' (an `ankyrin repeat' is a 33 amino acid motif which is present in erythrocyte ankyrin) in their C-terminus, are cytoplasmically located and inactive in DNA binding (12,13). The proteins that belong to the rel related subfamily are c-Rel (14,15), RelA (p65) (16,17), RelB (18) and the Drosophila proteins Dorsal (19) and Dif (20). Each of these proteins contains C-terminal regions that allow them to function as transcriptional activators (p50 related proteins lack such regions).
The ternary structure of (p50)2-DNA complex has a butterfly shape (21,22) in which protein domains, resembling the wings, clamp a cylindrical body of DNA. Each p50 subunit consists of two flexibly linked domains with a secondary structure similar to that of an immunoglobulin domain ([beta]-sandwich). As opposed to most DNA binding proteins, which use either small [alpha]-helices or [beta]-sheets to recognise their DNA targets, each p50 subunit contacts DNA through five flexible loops which connect the [beta]-strands. Two of these loops are in the N-terminal domain, two are in the C-terminal domain and the other one is the flexible linker connecting the domains. These flexible loops combine to form an extensive protein-DNA interface, interacting with the functional groups of the major groove and the bordering sugar-phosphate backbone, making ~38 individual contacts with target DNA, leaving only the minor groove free, so minor-groove-group proteins like HMG I(Y) can still bind to DNA along with NF-[kappa]B. The DNA of the complex is slightly unwound and bent, but not severely distorted by the interaction with the loops. These loops make base contacts with their target DNA bases, which determines p50 specificity and phosphate contacts which contribute to the binding energy of the complex. All these contacts give the complex an affinity which is higher than the affinity of most eukaryotic transcription factors. This strategy for DNA recognition, in which an immunoglobulin fold (23) acts as a scaffold for the DNA contacting flexible loops, is also employed by p53 (24), STAT-1 (25) and NFATC1 (26), even though there is little recognisable sequence homology.
The crystallographic data show that p50 interacts with DNA over a complete turn (10 bp). The most important NF-[kappa]B-DNA interaction is between the N-terminus of the L1 loop that connects A and B [beta]-sheets, and the bases of the major groove of target DNA. This part of the AB loop is termed the `recognition loop', because its interaction with DNA determines the sequence specificity of p50. In addition, protein footprinting by partial proteolysis (27,28) and chemical modification (29) suggested that the C-terminus of the L1-loops could also contact DNA. Although the lysine cluster (K77, K79 and K80) was implicated, the techniques employed did not have sufficient resolution to identify the lysine residue making the DNA contact (29). The objective of that study was therefore to use site-directed mutagenesis to investigate the role of each lysine in the sequence between 77 and 80. The DNA binding properties of wild type and mutant p50 proteins were studied by gel electrophoresis DNA binding assay and surface plasmon resonance (SPR). These experiments indicate that the conserved K80 participates in the binding of NF-[kappa]B p50 to DNA and is important for the stabilisation of the DNA-protein complex.
MATERIALS AND METHODS
Synthesis of oligonucleotides
Oligonucleotides were made on an Applied Biosystems 381A DNA synthesiser applying B-cyanoethyl phosphoramidite chemistry. Synthetic DNA was dissolved in water, ethanol precipitated, dried, and taken up in water. 3[prime] biotinylated oligonucleotides were produced with resins containing a biotinylated base. To estimate the concentration of single-stranded DNA oligonucleotides, the absorbance at 260 nm was measured using a spectrometer, and the concentration calculated based on the extinction coefficient of the oligonucleotide (30,31). Double-stranded oligonucleotides were produced by addition of equimolar amounts of complementary single-stranded oligonucleotides in TEN (100 mM NaCl, 10 mM Tris, pH 7.6, 1 mM EDTA, pH 8.0), boiling for 2 min and gradual cooling to room temperature (storage at 4 or -20°C).
Site-specific mutagenesis
The construction of an expression vector expressing p50 was described previously (32). For the construction of genes encoding mutated p50 molecules, a two-step PCR mutagenesis technique was employed (33), using a plasmid encoding the p105 precursor of p50 (kindly provided by A. Israël) as template. The products of the second PCR round were cleaved with BamHI and EcoRI and ligated into similarly cleaved pGEX-2T vectors (34). All the plasmids were used for the transformation of Escherichia coli to ampicillin resistance. The authenticity of the cloned products was checked by DNA sequencing (Sequenase 2.0, Amersham).
Expression and purification of NF-[kappa]B p50
Amino acids 35-381 of p50 were expressed in E.coli as part of a fusion protein with glutathione-S-transferase and purified by chromatography on glutathione agarose, thrombin cleavage and affinity chromatography on DNA-Sepharose as described previously (32).
Analytical procedures
Gel electrophoresis, DNA affinity chromatography and SDS-PAGE were carried out as described previously (32,35).
Partial proteolysis
To investigate whether the replacement of a positively charged hydrophilic amino acid residue (lysine) by a non-polar hydrophobic one (alanine) could induce a major conformational change, purified wild type p50 protein and p50 (K77A, K79A, K80A) were incubated with either proteinase K or chymotrypsin and the digestion products analysed by SDS-PAGE and Coomassie staining as described previously (27).
Gel electrophoresis DNA binding assay
The DNA binding properties of the different purified p50 proteins (wild type and mutants) were determined on 6% polyacrylamide gels (acrylamide:bis-acrylamide 55:1) in 0.5× TBE. Typically, the reaction (total volume 20 µl) contained 0.16 M NaCl, 27.5 mM Tris-HCl (pH 7.6), 10 mM DTT, 1 µg/µl BSA (protease free), 5% glycerol, 0.5% Nonidet P-40, p50 protein and 2 nM radiolabelled double-stranded, blunt-ended 17mer [kappa]B motif oligonucleotide (5[prime]-GCTGGGGATTCCCCATC-3[prime]) which was derived from the H-2K [kappa]B motif. After incubation for 30 min at room temperature, the samples were loaded in the gels to electrophoretically separate free DNA from the protein-bound DNA. The gels were run for 1.5 h at 150 V, placed on Whatman DE81 paper, dried and exposed to a PhosphorImager screen. Radioactive DNA was visualised using a FUJI PhosphorImager. Quantitation of the radioactivity in each DNA species was determined using MacBas. Gel electrophoresis DNA binding assays were complicated by the symmetrical nature of the [kappa]B motif oligonucleotides, as they had a propensity to form hairpin-like molecules. Thus, only a proportion of the DNA was fully double-stranded and competent for binding to NF-[kappa]B.
Surface plasmon resonance (SPR)
SPR detectors allow the direct visualisation of macromolecular interactions in real time, thus providing the data for the characterisation of the kinetics and thermodynamics (rate and equilibrium binding constants respectively) of these interactions (36).
The SPR detector which was used was a BIAcore-X system, run by a PC, containing an SPR monitor and an integrated microfluidic cartridge which, together with an autosampler, regulate the delivery of sample plugs into a running buffer that continuously passes over a sensor surface in a flow cell.
SA sensor chips were used (BIAcore) with streptavidin pre-immobilised on a dextran matrix. The removal of streptavidin, loosely bound to the chip, was performed by three applications of 0.07% SDS in water at 10 µl/min flow rate, for 1 min. Then, the surface was divided in two flow cells and double-stranded biotinylated DNA was applied to the second flow cell such that the capture would be 72 RU. The theoretical binding capacity of the chip is represented by an Rmax value of 500 RU. The dsDNAs (top strand shown) applied on the chip surfaces were MHC (5[prime]-GCTGGGGATTCCCCATC-3[prime]) derived from the H-2K [kappa]B motif, IRE (5[prime]-AAAGTGGGAAATTCCTCTG-3[prime]) derived from the interferon responsive element [kappa]B motif and NFIII (5[prime]-GAGTTAATATGCAAATAAG-3[prime]) derived from the octamer motif. One of the strands of each of these double-stranded oligonucleotides had a two-base (TT) protruding 3[prime] end which was biotinylated.
A 10 µl/min continuous flow of running buffer 250 mM NaCl, 10 mM HEPES pH 7.4, 3 mM EDTA, 2 mM DTT and 0.005% (v/v) surfactant P20 (BIAcore) was applied. The proteins were diluted in the running buffer (concentration 100 nM) and were injected to both flow cells, for 2 min (association step). Running buffer was then applied for 10 min (dissociation step). The surface was regenerated by the application of 0.07% SDS in water, for 1 min.
The kinetic data were produced by subtracting the SPR signals generated by passing the protein solutions across a flow cell containing streptavidin without DNA, from those obtained from a flow cell containing streptavidin with biotinylated double-stranded DNA. These data were evaluated with BIAevaluation 3.0 software package, as previously described (37,38).
Molecular modelling
The atomic coordinates of the human p50-DNA complex (22), were obtained from the Protein Data Bank at Brookhaven National Laboratory (PDB ID code: 1SVC). The protein fragment contained the residues 2-366 of the mutated p50 (C62A) bound to a DNA 19mer. These data were manipulated and processed by Insight II and RasMol 6 molecular modelling software. Atom distances were measured and figures were prepared with Insight II.
Sequence alignment
DNA and protein alignments were performed with the GCG software package (Genetics Computer Group) for UNIX.
Figure 1. Purification of wild type and mutant p50 proteins. (A) Coomassie Blue stained 10% polyacrylamide gel showing purified p50 and mutant p50 proteins (100 ng). The lower band corresponds to a C-terminal degradation product. (B) List of mutants containing the names of the purified proteins and the sequences of their C-terminal part of the AB loop. The residues within the box are the residues of the turn of the C-shaped loop. The point mutations introduced are in bold letters. To determine the importance in DNA binding of each lysine residue of the p50 sequence 77-80 wild type, single, double and triple mutants of p50, where lysines were replaced with alanines were purified (Fig. The replacement of positively charged hydrophilic residues with non-polar hydrophobic ones could introduce a change of the overall structure of (p50)2. Partial proteolysis experiments with chymotrypsin or proteinase K were performed with both wild type p50 and p50 (K77A, K79A, K80A) (triple mutant). The sites where chymotrypsin or proteinase K cleave wt p50 were identified previously (27). The two enzymes were selected because their cleavage specificities are such that K to A mutation in p50 would not alter substrate specificities and they could both cleave in the area of the C-terminus of AB loop (chymotrypsin cleaves after N78 and proteinase K after S74). The two enzymes could recognise many other cleavage sites in wild type p50 that were protected in the inner core of the protein. Thus, if a small conformational change of the loop or a major change of the overall structure occurred as a consequence of the mutations, then the sensitivity of some cleavage sites could be changed or completely abolished (especially the ones in close proximity to the mutations) or new cleavage sites could appear because of the exposure of protein sequences that were protected in the wild type p50. The digestion patterns of both wild type p50 and the triple mutant were almost identical (Fig. Figure 2. Proteolytic patterns of the wild type and the triple mutant p50. (A) Five micrograms of p50 and p50 (K77A, K79A, K80A) were digested with chymotrypsin at substrate:protease ratios of 2560:1, 1280:1, 640:1, 320:1, 160:1, 80:1, 40:1, 20:1 and 10:1 for 1 h at room temperature. Digestion products were resolved in a 10% polyacrylamide gel containing SDS and stained with Coomassie Blue. The masses of the molecular weight standards (M) are indicated in kDa. (B) Five micrograms of p50 and p50 (K77A, K79A, K80A) were digested with proteinase K at substrate:protease ratios of 20480:1, 10240:1, 5120:1, 2560:1, 1280:1, 640:1, 320:1, 160:1 and 80:1 for 1 h at room temperature. Digestion products were resolved in a 10% polyacrylamide gel containing SDS and stained with Coomassie Blue. The masses of the molecular weight standards (M) are indicated in kDa. Gel electrophoresis DNA binding assays were performed with all proteins (wild type and mutants). The data showed that the triple mutant was clearly more defective in binding to DNA than the other proteins (Fig. Figure 3. Gel electrophoresis DNA binding assays with all p50 proteins (wild type and mutants) using a 32P-labelled double-stranded 17mer oligonucleotide containing the [kappa]B motif from the H-2K enhancer: 5[prime]-GCTGGGGATTCCCCATC-3[prime] (MHC). (A) Aliquots of 1, 10, 100 and 1000 fmol of each protein were incubated for 30 min at room temperature with 20 fmol of 32P-labelled double-stranded MHC oligonucleotide. The samples were resolved in a 6% non-denaturing polyacrylamide gel (55:1 acrylamide:bis-acrylamide). The upper band corresponds to the DNA-protein complex. The lower band corresponds to protein free DNA. (B) DNA binding curves. Amount of bound to protein DNA over protein amount. To discriminate between the DNA binding activities of the p50 mutants SPR was employed. A biotinylated double-stranded DNA containing an NF-[kappa]B binding site was captured on the surface of a sensor chip by bound streptavidin. Wild type p50 and the various mutants were tested for DNA binding activity by passage over the sensor chip. DNA binding reactions were carried out under stringent conditions at 250 mM NaCl to eliminate non-specific interactions between the proteins and the sensor chip. Under these stringent conditions, wild type p50 bound efficiently to DNA containing a symmetrical [kappa]B motif (MHC), but bound less well to DNA containing an asymmetrical [kappa]B motif (HIV). Under these conditions binding to non-specific DNA sequence (NFIII) was negligible. Data were collected at a range of protein concentration between 2.5 and 1000 nM. At each protein concentration the analysis was carried out in triplicate. Although the data were highly reproducible it was not possible to derive association and dissociation rate constants for the interaction between p50 and DNA containing a [kappa]B motif, as the data could not be fitted to the theoretical binding models of the evaluation software. Thus it appears that interaction between p50 and DNA immobilised on the sensor surface did strictly follow the kinetics expected from a simple bimolecular reaction. This may be a consequence of conformational changes that accompany DNA binding, protein-protein interactions on the DNA (39) or a `cage' effect at the surface of the sensor chip. Although the quantitative evaluation could not be completed, the experimental data could be used to resolve the differences in the DNA binding of the wild type and mutant proteins by direct comparison of the binding curves, at a given protein concentration. Consistent with the data in Figure Figure 4. Typical sensogram (SPR), showing the association and the dissociation of the p50 proteins (wild type and mutants) to a double-stranded DNA containing the H-2K [kappa]B motif. In the association phase, 100 nM of each protein were injected for 2 min at 10 µl/min flow rate. In the dissociation phase, buffer containing no protein was injected for 10 min at 10 µl/min flow rate. Previous biochemical experiments suggested an interaction between the C-terminus of the AB loop of (p50)2 and DNA (27-29). Although residues K77, K79 and K80 were thought to be involved, the resolution of the techniques employed were such that the contribution of individual residues to the interaction could not be determined. Site-directed mutagenesis of this region was therefore employed to remove functional groups (K to A changes) and thus identify lysine residues in contact with DNA. The various mutants were placed in a rank order of DNA binding affinity using SPR. These experiments demonstrated that the residue involved in the DNA contact is predominantly K80, although K77 may also play a role. Partial proteolysis analysis of the wild type p50 and the triple mutant showed that the defect of the binding is not due to major conformational changes induced by introduction of the K to A changes. The crystallographic data are also consistent with a DNA contact beyond the consensus sequence. However this interaction was not clearly resolved because the C-terminal part of the L1 loop is not highly ordered, and because small double-stranded oligonucleotides (10mer, 11mer and 12mer) were used to form the co-crystals (21,22,40). Furthermore, it was shown that p50 affinity for a long oligonucleotide (16mer) was higher than of that for a shorter one (12mer), which corresponded to the oligonucleotide used in the crystallographic data (29). Analysis of the crystal structure of (p50)2-DNA (21,22), indicated that although the two crystal structures were quite similar, they had a fundamentally different relationship to their DNA targets (41), as a consequence of the flexibility of the NRD contacts with DNA. This flexibility is due to the adjustability of the loops and the hinge (which links the two domains of each subunit) which allows recognition of different but related DNA sequences. One explanation for the apparent differences between the two structures is the slight differences of the DNA sequences of the two oligonucleotides: 5[prime]-GGGAATTCCC-3[prime] (21) and 5[prime]-GGGGAATCCCC-3[prime] (22). Another reason is the slight difference of the length of the oligonucleotides: 10mer (21) and 11mer (22). This small difference of the size of the oligos played a dramatic role on the conformation of the AB (L1) loop in the two structures. While the 5[prime] end of the double-stranded 10mer is far from the C-terminus of AB loop, the 5[prime] end free phosphate group of the 11mer is in close proximity (2.87 Å) with K80 of the loop. This interaction affects the conformation of the lysine cluster of the loop. In the complex of p50 with the longer oligonucleotide the lysines of the loop have different orientations than in that with the shorter one. In both structures the side chain of K79 points away from the target DNA molecule and K80 is ideally positioned to directly contact it. While the K77 side chain is positioned facing DNA in the complex with the 10mer, it has the opposite orientation in the complex with the 11mer. An additional piece of evidence supporting the role of K80 in this interaction is that it is conserved among all of the NF-[kappa]B/Rel/Dorsal molecules. The only change is in human c-Rel where there is an arginine instead of a lysine, suggesting that the interaction is due to the positive charge of the amino acid (the chicken c-Rel contains lysine) (Fig. Figure 5. Conservation of K80 among NF-[kappa]B/Rel/Dorsal proteins. Multiple sequence alignment of p50, p52, c-Rel (human and chicken), v-Rel, RelA, RelB, Dorsal and Dif. The arrow points to the lysine residue in NF-[kappa]B/Rel/Dorsal proteins (K80 in p50). Molecular modelling studies in which four extra base pairs of B-form DNA were added to the oligo used in one of the crystallographic studies (22), clearly showed that the nitrogen atom of the positively charged K80-NH2 group is in close proximity (2.7 Å) to the oxygen atom of the negatively charged phosphate backbone of DNA. While K80 was ideally positioned to contact the phosphate backbone, K77 and K79 were pointing away from the DNA (Fig. Figure 6. Structural representation of the interaction of the p50 sequence 77-80 (lysine cluster of the C-terminal AB loop) with the 4 bp computer extended double-stranded DNA. Orientation of the lysine cluster: K80 is ideally positioned, because its [epsis]-NH2 group is 2.7 Å from the DNA phosphate backbone. K77 and K79 point away from the DNA backbone. While the p52 loop has a very similar structure (42) to its p50 homologue, the same loop in p65 homodimers bound to a palindromic DNA 20mer is at least 11 Å from the DNA backbone (43), although this may be a reflection of the complex and novel way that (p65)2 binds to DNA. This loop could have other functions that are not related to DNA binding but may serve as a target for I[kappa]B[alpha] recognition, in the same way that the p50 homologous loop serves as a target for I[kappa]B[gamma] recognition (29). We are extremely grateful to L. Torrance (Scottish Crop Research Institute) for the provision of the BIAcore-X system and J. Naismith for help with the molecular modelling. This work benefited by support from EU project ROCIO II.
RESULTS
Mutant p50 protein production
Chymotrypsin and proteinase K digestion of wild type and triple mutant p50
DNA binding properties of mutants
Discrimination between DNA binding activities of p50 mutants
DISCUSSION
ACKNOWLEDGEMENTS
REFERENCES
This article has been cited by other articles:
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 23 Dec 1998
Copyright©Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
M. R. Horton, S. Boodoo, and J. D. Powell
NF-kappa B Activation Mediates the Cross-talk between Extracellular Matrix and Interferon-gamma (IFN-gamma ) Leading to Enhanced Monokine Induced by IFN-gamma (MIG) Expression in Macrophages
J. Biol. Chem.,
November 8, 2002;
277(46):
43757 - 43762.
[Abstract]
[Full Text]
[PDF]
This Article 
Abstract
Print PDF (193K)
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (5)
Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Articles by Michalopoulos, I.
Articles by Hay, R. T.
Search for Related Content
PubMed
PubMed Citation
Articles by Michalopoulos, I.
Articles by Hay, R. T.
Social Bookmarking
What's this?