Crystal structure of NFAT bound to the HIV-1 LTR tandem kappaB enhancer element

Abstract

The host factor, nuclear factor of activated T-cells (NFAT), regulates the transcription and replication of HIV-1. Here, we have determined the crystal structure of the DNA binding domain of NFAT bound to the HIV-1 long terminal repeat (LTR) tandem kappaB enhancer element at 3.05 A resolution. NFAT binds as a dimer to the upstream kappaB site (Core II), but as a monomer to the 3' end of the downstream kappaB site (Core I). The DNA shows a significant bend near the 5' end of Core I, where a lysine residue from NFAT bound to the 3' end of Core II inserts into the minor groove and seems to cause DNA bases to flip out. Consistent with this structural feature, the 5' end of Core I become hypersensitive to dimethylsulfate in the in vivo footprinting upon transcriptional activation of the HIV-1 LTR. Our studies provide a basis for further investigating the functional mechanisms of NFAT in HIV-1 transcription and replication.

Figure 1

Overall structure of the NFAT/HIV-1 LTR tandem kappaB complex (a) the Rel homology region (RHR) of three NFAT molecules bind to the tandem kappaB sites of the HIV-1 LTR. The proteins are in ribbon style, the DNA is represented by the stick model (deep cyan). The NFAT monomer bound to the 3’ end of core I is colored in ruby red, the NFAT dimer bound to core II is colored in deep purple (3’ end) and deep blue (5’end), respectively. The N-terminal immunoglobulin domain (RHR-N) and the C-terminal immunoglobulin domain (RHR-C) are labeled for all three molecules. The DNA sequence used in the crystallization is listed below. Core I is from Gua15 to Cyt24, whereas core II is from Gua1 to Cyt10. (b) The complex is rotated along the DNA axis by 90 degrees with respect to (a).

Figure 2

Orientation of the HIV-1 LTR tandem kappaB DNA in the crystal A 5-BrdU base analogue is engineered at the Thy8 position (magenta) in core II. The difference fourier map (contour level 3 sigma) shows clearly the extra density (yellow) corresponding to the heavier bromine atom. This region of DNA is occupied by the NFAT dimer.

Figure 3

Two distinct conformations of the NFAT RHR on the HIV-1 LTR tandem kappaB sites (a) Superposition of the NFAT monomer on the HIV-1 LTR core I (protein: ruby red; DNA: dark cyan) and the NFAT monomer in the NFAT/Fos-Jun/ARRE2 complex (yellow). The Fos-Jun complex and the NFAT dimer on core II are omitted for clarity. It is clear that the NFAT has a very similar structure in the two distinct complexes, including the relative orientation of the RHR-N and RHR-C. But the DNA in the HIV-1 LTR complex (dark cyan) has a greater bend than that in the NFAT/Fos-Jun/ARRE2 complex (yellow) (discussed further in the text). (b) Superposition of the NFAT dimer bound to core II of the HIV-1 LTR tandem kappaB sites (dark blue) and the NFAT dimer bound to an isolated kappaB site (green). Both the protein and DNA superimpose very well in this region. The NFAT monomer bound to core I is omitted.

Figure 4

DNA conformation of the HIV-1 LTR tandem kappaB sites (a) A simulated omit map (contour level 2 sigma) showing well-defined density of the HIV-1 LTR tandem kappaB sites and the distorted DNA helix. The sequence of the top stand is labeled. (b) Superposition of the DNA in the crystal (dark cyan) and an ideal B-DNA containing the sequence of the HIV-1 LTR tandem kappaB sites (magenta) shows that the DNA bend is located at the 5’ end of core I.

Figure 5

Protein-protein interactions in the NFAT/HIV-1 LTR complex (a) The NFAT dimer bound to core II is shown in surface representation superimposed on the ribbon structure. The arrow indicates the major dimer interface formed by RHR-C. The E’F loop also makes a dimer interface on the opposite side of the DNA. As a result, DNA is encircled by the NFAT dimer. In this view, the NFAT monomer is omitted for clarity. (b) Side view of the NFAT/HIV-1 LTR complex showing the small contact between the RHR-C of the NFAT monomer bound to core I and the RHR-N of the NFAT molecule bound to the 5’ end of core II. The DNA bend is also clearly visible, which may facilitate the interactions between NFAT molecules bound to Core I and Core II.

Figure 6

Schematic of interactions between NFAT and the HIV-1 LTR tandem kappaB sites DNA is represented as a ladder with bases as ovals and labelled according to the text and Figure 1a. The backbone phosphates are represented as circles with the letter P inside. “Monomer” represents the NFAT molecule bound to the 3’ end of core I, whose residues are coloured in ruby red. “Dimer 2” represents the NFAT molecule bound to the 3’ end of core II, whose residues are coloured in deep purple. “Dimer 1” represents the NFAT molecule bound to the 5’ end of core II, whose residues are coloured in deep blue. Hydrogen bonding interactions are solid arrows while van der Waals interactions are dashed arrows. The guanine nucleotides protected in the in vivo DMS footprinting are highlighted by bold-dashed circle, whereas guanine residues with enhanced reactivity toward DMS are highlighted by bold-red circles. For clarity, only representative protein-DNA contacts are shown in the figure.

Figure 7

Electrophoresis mobility shift assay of NFAT1 RHR binding to the HIV-1 LTR tandem kappaB sites DNA containing the wild type HIV-1 LTR tandem kappaB sequences (lanes 1–3), or the mutant modified at the 5’ end of Core I (lanes 4–6), or the mutant modified at both the 5’ end and 3’ end of Core I (lanes 7–9) are used in three sets of binding reactions. The detailed sequence changes in both mutants are described in materials and methods. The DNA concentration was kept at 1nM. For each set of DNA (lanes 1–3, lanes 4–6, or lanes 7–9), the concentration of NFAT RHR varies 0, 4nM to 40nM. DNA: unbound free DNA; Dimer: two NFAT molecules bound to DNA; Trimer: three NFAT molecules bound to DNA.

Figure 8

Isothermol titration calorimetry analysis of NFAT1 RHR binding to the HIV-1 LTR tandem kappaB sites Top panel: heat effects associated with the injection of DNA into the solution of NFAT1 RHR. Bottom panel: DNA concentration dependence of the heat released upon DNA binding to NFAT1 after normalization and correction for the heats of dilution. The horizontal axis refers to NFAT vs DNA molar ratio. The DNA sequence used in this experiment is shown below (see Methods for further details).

Figure 9

Effect of the spacing between the two kappaB sites in the HIV-1 LTR (a) Electrophoresis mobility shift assay of NFAT1 RHR binding to the wild type HIV-1 LTR (WT) (Lanes 1–6) or the insertion mutant (Insert) (Lanes 7–12). The sequences of both probes are listed below the gel, with the inserted bases italicized. The DNA was kept at 0.1nmoles. For each set of DNA probe (lanes 1–6, or lanes 7–12), the molar ratio of protein (NFAT RHR) to DNA varies from 0, 0.5, 1, 2, 4, and 6. The gel was stained for DNA. (b) The same gel was stained for protein.

Figure 10

Detailed view of the DNA bending region A simulated omit map (contour level 3 sigma) showing well-defined density of Lys482 of the NFAT molecule bound to the 3’ end of core II. This lysine residue inserts deeply into the minor groove opposing Gua14 and Gua15. Although at current resolution the exact conformation of individual bases cannot be refined without B-DNA constraints, a pronounced density is observed in the major groove of Gua14 and Gua15, suggesting that these two bases are distorted, perhaps becoming extra helical. These two bases showed enhanced reactivity toward DMS in the in vivo footprinting (see text for further details).