Complete ¹H, ¹³C, ¹⁵N resonance assignments and secondary structure of the Vpr binding region of hHR23A (residues 223-363)

Abstract

Comprehensive resonance assignments and delineation of the secondary structure elements of the C-terminal Vpr-binding region of hHR23A, residues 223–363, were achieved by triple-resonance NMR experiments on uniformly ¹³C,¹⁵N-labeled protein. Assignments are 100% and >95% complete for backbone and side-chain resonances, respectively. This data constitutes important complementary information for our ongoing structure determination of the Vpr-hHR23A(223–363) complex. At high concentrations, severe line-broadening was observed for several residues in the ¹H-¹⁵N HSQC spectrum, most likely resulting from inter-molecular interactions.

Biological context

The human homolog of the yeast DNA repair protein RAD23, hHR23A, was identified to interact with the human immunodeficiency virus type 1 (HIV-1) accessory protein Vpr by yeast two-hybrid assay more than two decades ago (Withers-Ward et al. 1997). Like other HIV-1 accessory proteins, Vpr repurposes the cellular ubiquitination machinery to remove restriction factors or cellular proteins that interfere with successful viral replication (Arora et al. 2014). For example, Vpr hijacks the CRL4^DCAF1 E3 ubiquitin ligase to antagonize the DNA repair enzyme UNG2 (Ahn et al. 2010). hHR23A and hHR23B, together with xeroderma pigmentosum group C (XPC) protein, recognize UV-damaged DNA and initiate global genome nucleotide excision repair (NER) (Masutani et al. 1994). hHR23A comprises 363 amino acids, folded into four independently structured domains, with a ubiquitin-like (UBL) domain at its N-terminus followed by a ubiquitin-associated (UBA1) domain, an XPC-binding (XPCB) domain and another UBA domain (UBA2). Several years ago, we showed that the XPCB and UBA2 domains in the C-terminal region of hHR23A interact with Vpr (Jung et al. 2014). To assist in the assignment and three-dimensional structure determination of the hHR23A(223–363)/Vpr complex, we completed ¹H, ¹³C, ¹⁵N resonance assignments of the free hHR23A(223–363). Although the NMR structure of full-length hHR23A has been reported (Walters et al. 2003), NMR assignments were not deposited in the BMRB. NMR structures of the XPCB domain (Kamionka and Feigon 2004) and the UBA2 domain (Dieckmann et al. 1998) are also available but only the NMR assignments of the hHR23A UBA2 domain, residues 319–363, are available (BMRB accession 4757). In the latter construct, three N-terminal residues (V316-P318) were missing, which we found to interact with the rest of UBA2 structure. Here we report complete backbone assignments and >95% side-chain assignments, respectively, of hHR23A(223–363) that includes XPCB and UBA2.

Methods and experiments

The expression construct encodes the C-terminal region (residues 223–363) of hHR23A, inserted into the pET21 vector (EMD Chemicals), containing a C-terminal His₆ tag (LEHHHHHH). The protein was produced in E. coli Rosetta 2 (DE3) cells (EMD Chemicals), using 0.5 mM isopropyl 1-thio-β-D-galactopyranoside for induction over 16 h at 18 °C. Uniform ¹⁵N- and ¹³C-labeling was carried out by growth in modified minimal medium, containing ¹⁵NH₄Cl (1 g/L) and ¹³C₆-glucose (2 g/L) as sole nitrogen and carbon sources. Cells were harvested by centrifugation at 6000g, resuspended in the lysis buffer containing 50 mM sodium phosphate, pH 7.5, 10 mM imidazole and 200 mM NaCl and lysed using a microfluidizer (Microfluidics, MA). Cell debris was removed by centrifugation at 40000g and the supernatant was applied to a 5-mL HisTrap (GE Healthcare) affinity column, equilibrated in lysis buffer. The bound protein was eluted using a linear gradient of 0.01–0.50 M imidazole and further purified over a HiPrep Superdex 75 (2.6 cm × 60 cm, GE healthcare) gel filtration column equilibrated in 25 mM sodium phosphate buffer (pH 7.5), 150 mM NaCl, 2 mM DTT and 0.02% sodium azide. Protein fractions were pooled, concentrated and buffer-exchanged in an Amicon Ultra concentrator (Millipore) into NMR buffers: 25 mM sodium phosphate (pH 7.2), 10 mM HEPES (pH 7.2) or 10 mM MES (pH 6.5), each containing 50 mM NaCl, 1 mM ZnSO₄, 1 mM EDTA, 1 mM TCEP and 0.02% sodium azide. The molecular mass of the protein was confirmed by LC-ESI-TOF mass spectrometry (Bruker Daltonics, Billerica, MA).

NMR spectroscopy.

NMR data were collected on Bruker 600, 700, 800 and 900 MHz AVANCE spectrometers. All spectrometers were equipped with z-axis gradient, triple resonance cryoprobes. Experiments were performed at 291 K and 298 K. Total sequence-specific assignments were obtained from the following experiments (Bax and Grzesiek 1993; Sattler et al. 1995; Clore and Gronenborn 1998): heteronuclear 2D ¹H-¹⁵N HSQC and ¹H-¹³C HSQC and 3D HNCACB, CBCA(CO)NH, HNCA, HN(CO)CA, HNCO, HN(CA)CO, CC(CO)NH, H(CCCO)NH, HBHA(CO)NH, HCCH-TOCSY (mixing time 10.9 ms) and simultaneously ¹³C/¹⁵N-edited NOESY (mixing time 100 ms) experiments. All data were processed with NMRPipe (Delaglio et al. 1995) and TopSpin 3.1 (Bruker), and analyzed with CCPN Version 2.3.1 (Vranken et al. 2005). Proton chemical shifts were referenced directly to internal 2,2-dimethyl-2-silapentane-5-sulfonic acid (DSS) and ¹³C and ¹⁵N chemical shifts were referenced indirectly to DSS using the ¹³C/¹H and ¹⁵N/¹H frequency ratios (0.251449530 and 0.101329118, respectively) at zero-point (Wishart et al. 1995). Secondary chemical shifts, ΔCα and ΔCβ (and ΔCα-ΔCβ) are calculated by subtracting random coil values (Wishart et al. 1995) from the Cα and Cβ shifts of each hHR23A(223–363) residue. The ϕ and ψ backbone torsion angles were obtained from TALOS+ Version 3.80F1 (Shen et al. 2009).

Total assignments and data deposition, secondary structures and intra- and inter-molecular interaction

The ¹H-¹⁵N HSQC spectrum of hHR23A(223–363) at 0.6 mM exhibits well dispersed ¹H-¹⁵N resonances with narrow lines (Fig. 1), similar to full-length hHR23A that exists as a monomer at concentrations from 0.1 to 0.6 mM (Jung et al. 2014). In total, complete backbone N, HN, C’, Cα and Hα (100 %) and nearly complete side-chain (>95 %) assignments were achieved. The assignments were deposited in the BioMagResBank (www.bmrb.wisc.edu) under accession number 27978. Based on secondary chemical shifts, ΔCα-ΔCβ, ΔCα and −ΔCβ (Spera and Bax 1991), the XPCB and UBA2 domains possess 5 and 3 helical regions, respectively (Fig. 2). Together with the ϕ and ψ backbone torsion angles obtained from TALOS+, helical regions were delineated: residues 233–236, 239–250, 252–264, 267–275, 277–284 (in XPCB), 318–330, 334–343 and 348–357 (in UBA2). The two residues V316 and T317 immediately prior to the first helix (318–330) of UBA2 exhibit sizeable (|ΔCα–ΔCβ|>2.5 ppm) negative and positive secondary chemical shifts (Fig. 2), suggesting a non-random structure. In addition, V316 and T317 exhibit numerous long-range (|i–j|≥5) NOEs in the 3D ¹³C/¹⁵N-edited NOESY spectrum, e.g., with K321, E322 and I324 in helix 1 and I338, A340, Y341 and F342 in helix 2 (Fig. 3a), indicating that these residues are packed against helices 1 and 2 in UBA2. Similar structural features were seen in full-length hHR23A (Walters et al. 2003), confirming that the structure of the XPCB-UBA2 construct preserves the structure of the full-length protein. However, in the isolated UBA2 domain (residues 319–363) studied previously, three N-terminal residues (316–318) are missing (Dieckmann et al. 1998). As a result, their chemical shifts differ significantly. For example, the ¹H/¹⁵N amide resonances of E320 and K321 of hHR23A(223–363) appear at 7.72/122.66 ppm and 8.57/119.36 ppm, respectively, while in the isolated UBA2 domain (BMRB accession 4757) they are reported to resonate at 8.47/120.31 and 7.78/118.06 ppm, respectively. We therefore suggest that the correct domain boundaries are 316–358 for the UBA2 domain. For the XPCB domain, the domain comprises residues 230–286. The remaining parts of the polypeptide chain (residues 223–229, 288–315 and 359–363) are linkers that precede, connect or follow the folded domains in hHR23A(223–363). Their resonances exhibit small secondary shifts (Fig. 2) and mostly strong ¹H-¹⁵N HSQC signals (Electronic supplementary material Fig. S1), apart for residues 310–315 (see below), indicating that these regions are flexible and possess random coil conformations.

Fig. 1. 900 MHz ¹H-¹⁵N HSQC spectrum of hHR23A(223–363).

0.6 mM protein in 25 mM sodium phosphate buffer (pH 7.2), 50 mM NaCl, 1 mM ZnSO₄, 1 mM EDTA, 1 mM TCEP and 0.02% sodium azide at 298 K. The assignments are given by residue number and name. Asn δ and Gln ε side-chain resonances are connected by lines. An expansion of the crowded central spectral region is shown in the bottom-right corner.

Fig. 2. Secondary chemical shifts of hHR23A(223–363).

Secondary chemical shifts, ΔCα-ΔCβ, ΔCα and −ΔCβ, are plotted along the linear amino acid sequence. ΔCα and ΔCβ are calculated by subtracting random coil values (Wishart et al. 1995) from the Cα and Cβ shifts. Helical regions in the XPCB and UBA2 domains are depicted at the top.

Fig. 3. Intra-molecular NOEs and ¹H-¹⁵N HSQC resonance broadening involving the V316-T317 region of hHR23A(223–363).

(a) Strips from the 900 MHz 3D ¹³C-edited NOESY spectrum involving NOEs from the V316 γ methyl group (top) and from the 900 MHz 3D ¹⁵N-edited NOESY spectrum involving the T317 backbone amide NH (bottom) in a 0.6 mM protein sample (identical sample as used for Fig. 1). (b) Superimposition of the ¹H-¹⁵N HSQC spectra (900 MHz, 298 K) of 0.6 mM (black) and 2.7 mM (red) hHR23A(223–363), respectively. Resonances that broaden beyond detection at 2.7 mM are labeled. Several residues in the His6 tag (E365, H366 and H371) and Q277 located next to a histidine (H276) experience slight resonance shifts due to small pH differences in the three samples. These resonances are labeled in grey.

¹H-¹⁵N HSQC spectra at concentrations of 0.6, 1.2 and 2.7 mM were evaluated to assess the behavior of linker residues 310–315. Resonances from residues 312–317 exhibited severe line-broadening at 1.2 mM (Electronic supplementary material Fig. S2) and were broadened beyond detection at 2.7 mM (Fig. 3b). Concomitant with the broadening observed for these linker residue resonances, several resonances associated with helix 1 (Q319, K321 and R326) and helix 2 (I338, Q339, Y341, F342 and A343) residues also exhibit line-broadening. Since at low concentration (0.6 mM), where the protein exists as a monomer, NOEs are observed between V316/T317 and helices 1 and 2 residues (Fig. 3a), the simplest explanation for the line broadening is that an inter-molecular interaction occurs at high concentrations (1.2 and 2.7 mM) and exchange on the intermediate chemical shift scale between a monomer and homodimer would cause the line broadening. In this respect, it is worth pointing out that RAD23, the yeast homolog of hHR23A, has been reported to form predominantly a homodimer by gel-filtration column chromatography (Bertolaet et al. 2001).

Supplementary Material

12104_2019_9913_MOESM1_ESM

Fig. S1. 900 MHz ¹H-¹⁵N HSQC spectrum of hHR23A(223–363) at 298 K. Only the most intense resonances are displayed. The spectrum is identical to the one in Fig. 1, however, a different contour level was chosen to selectively display only intense resonances that are associated with residues in flexible linkers. The spectral region enclosed by a rectangle contains mostly Asn δ and Gln ε side-chain resonances.

Fig. S2. Superimposition of the 0.6 mM (black) and 1.2 mM (green) 900 MHz ¹H-¹⁵N HSQC spectra (298 K). Resonances that are broadened at 1.2 mM concentration are labeled. Resonances affected by the slight pH difference in the two samples are labeled in grey.