Solution NMR structures reveal a distinct architecture and provide first structures for protein domain family PF04536

Abstract

The protein family (Pfam) PF04536 is a broadly conserved domain family of unknown function (DUF477), with more than 1,350 members in prokaryotic and eukaryotic proteins. High-quality NMR structures of the N-terminal domain comprising residues 41–180 of the 684-residue protein CG2496 from Corynebacterium glutamicum and the N-terminal domain comprising residues 35–182 of the 435-residue protein PG0361 from Porphyromonas gingivalis both exhibit an α/β fold comprised of a four-stranded β-sheet, three α-helices packed against one side of the sheet, and a fourth α-helix attached to the other side. In spite of low sequence similarity (18%) assessed by structure-based sequence alignment, the two structures are globally quite similar. However, moderate structural differences are observed for the relative orientation of two of the four helices. Comparison with known protein structures reveals that the α/β architecture of CG2496(41–180) and PG0361(35–182) has previously not been characterized. Moreover, calculation of surface charge potential and identification of surface clefts indicate that the two domains very likely have different functions.

Introduction

684-residue protein CG2496 from Corynebacterium glutamicum (UniProt accession number Q6M3G5) and 435-residue protein PG0361 from Porphyromonas gingivalis (Q7MX54) contain N-terminally located domains, which belong to the Pfam [1] protein family PF04536 of unknown function (DUF477) (Fig. S1). This broadly conserved protein domain family contains currently 1,351 members from a wide range of different bacteria, eukaryotic organisms, and remarkably also one archaebacterium (crenarchaeota). The N-terminal domains CG2496(41–180) and PG0361(35–182), which exhibit very low ClustalW [2] pairwise sequence identity (<20%), were selected as targets of the Protein Structure Initiative and assigned to the Northeast Structural Genomics consortium (NESG; http://www.nesg.org) for structure determination (NESG Target IDs CgR26A and PgR37A, respectively), as part of the a cooperative intercenter effort aimed at providing structural coverage of large, uncharacterized protein domain families [3]. Initial structural representatives of such families exhibit high modeling leverage [4], expand our understanding of protein evolution [5], and generally expand our knowledge of fundamental relationships between protein sequences, three-dimensional structure, and protein function. The solution NMR structures of CG2496(41–180) and PG0361(35–182) presented here are the first atomic resolution structures for domains of Pfam family PF04536.

Methods

CG2496(41–180) and PG0361(35–182) were cloned, expressed, and purified following protocols [6–8] established by the NESG (see Supplementary Material for details; http://www.nmr2.buffalo.edu/nesg.wiki). The proteins included short C-terminal hexaHis tags (LEHHHHHH). The corresponding pET expression vectors (NESG CgR26A-41-180-21.3 and PgR37A-35-182-21.12), have been deposited in the PSI Materials Repository (http://psimr.asu.edu/). Protein samples were prepared at ∼0.9 mM concentration in 90% H₂O/10% D₂O, in a buffer containing 20 mM MES, 100 mM NaCl, 10 mM DTT, 5 mM CaCl₂, 50 μM DSS, 0.02% NaN₃ at pH 6.5. The [5% ¹³C; U-¹⁵N]-labeled samples enabled stereospecific assignment of the methyl groups of Val and Leu residues [9]. Isotropic overall rotational correlation times of about 9 ns were inferred from average ¹⁵N spin relaxation times for both CG2496(41–180) and PG0361(35–182) (Supplementary Material, http://www.nmr2.buffalo.edu/nesg.wiki), indicating that both protein domains are monomeric in solution. This finding was confirmed by analytical gel-filtration with static light scattering detection (Supplementary Figs. S2, S3).

NMR data were acquired at 25°C on Varian INOVA 600 and 750 MHz, and Bruker AVANCE 800 and 900 MHz spectrometers, each equipped with a cryogenic ¹H{¹³C,¹⁵N} probe. Total NMR measurement time for CG2496(41–180) and PG0361(35–182) was 150 h each. Nearly complete sequence-specific ¹H, ¹⁵N and ¹³C resonance assignments (Table 1; Supplementary Figs. S4, S5) were obtained from conventional triple-resonance NMR experiments (Supplementary Material) using the programs AutoAssign 2.3.0 [10, 11] and PINE [12], followed by manual assignment of side-chain resonances. Assignments were validated using the AVS software suite [13]. Chemical shifts, NOESY peak lists, and time domain NMR data have been deposited in the BioMagResBank (accession numbers 16569 and 16810 for CG2496(41–180) and PG0361(35–182), respectively).

Table 1. CG2496(41–180) and PG0361(35–182) structure statistics.

	CG2496(41–180)	PG0361(35–182)
Completeness of resonance assignmentsa (%)
Backbone/Side-chain	100.0/99.7	98.1/100.0
Completeness of stereospecific assignmentsb (%)
Val & Leu isopropyl/βCH2/αCH2 of Gly	100/35/50	100/25/30
Conformation-restricting distance constraintsc
Intraresidue (i = j)	432	593
Sequential (|i ‒ j| = 1)	567	894
Medium range (1 < |i ‒ j| < 5)	579	1,012
Long range (|i ‒ j| ≥ 5)	1,067	1,453
Total	2,645	3,952
Dihedral angle constraints (φ/ψ)	73/73	77/77
Distance constraints per residue (of those, long-range)	24.7 (9.4)	27.0 (9.6)
CYANA target function (Å2)	0.45 ± 0.16	1.85 ± 0.16
Average number of distance constraint violations per conformer
0.2‒0.5 Å	1.0	5.6
>0.5 Å	0.0	0.1
Average number of dihedral angle constraint violations per conformer
>10°	0.0	0.0
Average RMSD from mean coordinates (Å)
Backbone heavy atoms (all heavy atoms)d	0.4 (0.6)	0.5 (0.9)
Backbone heavy atoms (all heavy atoms)e	0.4 (0.7)	0.5 (0.9)
Heavy atoms of molecular coref	0.4	0.5
Global quality scoresc (raw/Z-score)
PROCHECK [32] G-factor(φ and ψ)	0.08/0.63	−0.04/0.16
PROCHECK [32] G-factor (all dihedral angles)	0.07/0.41	0.02/0.12
MOLPROBITY [33] clash score	19.81/−1.87	23.77/−2.55
Verify3D [34]	0.47/0.16	0.48/0.32
ProsaII [35]	0.98/1.36	0.68/0.12
RPF scores [23]
Recall/Precision/F-measure	0.98/0.93/0.96	0.97/0.92/0.95
DP-score	0.90	0.85
MOLPROBITY [33] Ramachandran summarye (%)
Most favored regions	98.9	97.2
Allowed regions	1.0	2.5
Disallowed regions	0.1	0.3

^aCalculated with the AVS suite [13] excluding low complexity regions (residues 172–180 in CG2496), as well as C-terminal tags, N-terminal and Lys and Arg side chain amino groups, hydroxyl of Ser, Thr and Tyr, carboxyls of Asp and Glu, and non-protonated aromatic carbons

^bRelative to pairs with non-degenerate chemical shifts

^cCalculated with PSVS 1.4 [22]

^dRegular secondary structure elements: residues 57–59, 66–82, 86–91, 100–110, 116–122, 127–132, 138–153, 157–169 in CG2496(41–180) and 54–56, 63–79, 83–89, 97–108, 118–124, 129–134, 144–161, 164–182 in PG0361(35–182)

^eOrdered residues: 56–60, 63–93, 98–111, 115–168 in CG2496(41–180) and 37–45, 54–92, 95–108, 118–134, 137–181 PG0361(35–182)

^fResidues 56–61, 63–72, 74–76, 79–83, 85–95, 98–100, 102–104, 106–111, 115–123, 127–128, 130–131, 134, 136–137, 139, 141–144, 146–149, 151–153, 155–158, 160–165, 167–168 in CG2496(41–180) and 36, 37, 39, 42, 43, 45, 53–55, 58, 60, 61, 64, 68, 69, 73, 75, 76, 78, 80–90, 92, 95–97, 100–102, 104, 105, 107, 108, 119–125, 130–134, 138, 141–143, 146–148, 150, 151, 155–159, 161, 163–165, 168, 169, 172–176, 178, 179 in PG0361(35–182). Includes best-defined side chains

Structure calculations were performed using standardized methods of the NESG consortium [14, 15] and consensus analysis of automated NOESY (mixing time 70 ms) cross peak assignments provided by the programs CYANA[ 16, 17] and AutoStructure 2.2.1 [18] based on ¹H–¹H NOE-derived upper limit distance constraints, and backbone dihedral angle constraints derived from chemical shifts using the program TALOS+ [19] for residues located in well-defined regular structure elements. Stereospecific assignments of methylene protons were performed with the GLOMSA module of CYANA and the final structure calculation was performed with CYANA followed by refinement of selected conformers in an ‘explicit water bath’ [20] using the program CNS 1.2 [21]. Validation of the resulting 20 refined conformers for each domain structure was performed with the Protein Structure Validation Software (PSVS) server 1.3 [22] and the agreement of structures and NOESY peak lists was verified using the AutoStructure/RPF 2.2.1 package [23].

Results and discussion

We obtained high-quality (Table 1) NMR structures of CG2496(41–180) and PG0361(35–182) (Fig. 1) and their coordinates were deposited in the Protein Data Bank [24] on 10/19/2009 (accession code 2KPT) and 03/31/2010 (accession code 2KW7), respectively. Both structures exhibit an α/β-fold (Fig. 1b,e) consisting of four α-helices and a four-stranded β-sheet with the topology A(↑)B(↑)C(↑)D(↓). α-Helices I, III and IV are packed against one side of the β-sheet, while helix II is located on the opposite side. The locations of regular secondary structure elements are: β-strands A (residues 57–59 in CG2496/54–56 in PG0361), B (86–91/83–89), C (116–122/118–124) and D (127–132/129–134), and α-helices I (66–82/63–79), II (100–111/97–108), III (138–153/144–161) and IV (157–169/164–182).

Fig. 1.

a Stereoview of the 20 conformers representing the solution structure of CG2496(41–180) obtained after superposition of the C^α atoms of the regular secondary structure elements for minimal RMSD. Residues 41–52 and 172–180 of the disordered N- and C-terminal polypeptide segments were omitted for clarity, and the termini are labeled as “N” and “C”. b Ribbon diagram of residues 53–171 of the lowest-energy conformer of CG2496(41–180): α-helices are shown in red and yellow, β-strands are depicted in cyan, other polypeptide segments are in gray. c Sausage representation of backbone and superposition of the conformation of the best defined side chains (Table 1). A spline curve was drawn through the mean positions of C^α atoms of residues 53–171 with the thickness proportional to the mean global displacement of C^α atoms in the 20 conformers superimposed in (a). d Same as a for PG0361(35–182) with residues 35–182 shown. e Same as b for PG0361(35–182) with residues 35–182 shown. f Same as (c) for PG0361(35–182) with residues 35–182 shown. g Surface and space-filling representations of the lowest-energy conformer of CG2496(41–180) colored according to the electrostatic potential and the degree of residue conservation, respectively. The default ConSurf color scheme for residue conservation is employed: burgundy for the strongest conservation, cyan for the highest variability, and yellow for residues with insufficient data. Only residues 53–171 are shown with the flexible terminal segments excluded. The structures shown on the left have the same orientation as Fig. 1, and those on the right are rotated by 180° around the vertical axis. Surface clefts identified by Mark-Us/SCREEN are labeled as C1 and C2. h Same as g for the lowest-energy conformer of PG0361(35–182) with residues 35–182 shown, The structures shown on the left are rotated around the vertical axis by 90° relative to the orientation in (d–f), and those on the right are rotated by a further 180°. Surface clefts identified by Mark-Us/SCREEN are labeled as C1, C2, C3 and C4

In spite of the very low sequence identity (18% inferred from structure, Fig. S1c), the three-dimensional structures of PG0361(35–182) and CG2496(41–180) are quite similar: the root mean square deviation (RMSD) calculated for the mean coordinates of the backbone heavy atoms N, C^α and C′ of regular secondary structure elements is 2.2 Å. Furthermore, α-helices III and IV exhibit the largest structural differences in terms of length and packing against the remainder of the protein molecule, and Pro 157 introduces a kink in α-helix III of PG0361(35–182) that is absent in CG2496(41–180). As a result, the corresponding RMSD value calculated for only for the β-sheet and α-helices I and II is much lower, that is, 1.0 Å. A rather distant homology is reflected by the fact that 25% of the residues of the molecular core are conserved between PG0361(35–182) and CG2496(41–180) (Fig. S1c).

A search of the PDB database for similar structures using the program DALI [25] identifies the C-terminal domain of alanyl-tRNA synthetase (named “C-Ala domain” in the following) from Aquifex aeolicus (PDB code 3G98) as the only highly significant hit (other hits had Z-scores < ∼6) for both CG2496(41–180) (best match with chain B of 3G98: Z-score 8.4, RMSD of C^α atoms = 2.3 Å for 85 aligned residues with 8% sequence identity) and PG0361(35–182) (best match with chain A of 3G98: Z-score 6.5, RMSD of C^α atoms = 2.5 Å for 89 aligned residues and 4% sequence identity). However, the comparably small number of aligned residues indicates that structural similarity with C-Ala domain is limited to segments of the protein molecules. This is confirmed by visual inspection (Fig. S6): C-Ala domain contains a β-sheet with topology A(↓)B(↑)C(↑)D(↓)E(↑)F(↓), but only β-strands B-D and α-helices II-IV align structurally with corresponding regular structure elements in CG2496(41–180) and PG0361(35–182). Moreover, the short β-strand A is arranged in opposite direction in C-Ala domain, α-helix I is absent, and the short polypeptide segment connecting α-helices III and IV in both CG2496(41–180) and PG0361(35–182) is replaced by the antiparallel β-strands E and F. Furthermore, the functionally important Arg 840 residue predicted to interact with the elbow of tRNA^Ala, which is located in the β-strand F [26], is not present in CG2496(41–180) and PG0361(35–182). Hence, CG2496(41–180) and PG0361(35–182) are quite likely functionally not similar to the C-Ala domain; the observed partial structural similarity may have emerged from convergent evolution. This view is further supported by the fact that the full length proteins CG2496 and PG0361 are certainly not tRNA synthetases, but have entirely different functions: they are predicted to contain (1) transmembrane segments (Figs. S7–S10) and (2) N-terminal signal sequences for translocation in the extracellular and periplasmic space (Figs. S7–S12). Taken together, the search for structurally similar proteins reveals that CG2496(41–180) and PG0361(35–182) exhibit a novel α/β architecture. So far, these structures have not yet yielded insights into their molecular functions.

Calculation of electrostatic surface potentials and identification of surface clefts, which are possibly of functional importance [27, 28], indicates that CG2496(41–180) and PG0361(35–182) actually have different functions in the context of the full-length proteins. CG2496(41–180) features a mostly negative electrostatic surface potential and analysis using Mark-Us/SCREEN2 [27, 29] reveals two adjacent surface clefts C1 and C2 (Fig. 1g) located between the β-sheet and α-helix II (C1, 33 Ų surface area, formed by Thr 58, Phe 88, Val 90, Trp 103, Ala 107, Asn 111 and Ile 188; C2, 25 Ų surface area, Tyr 60, Leu 92, Ser 93, Ser 94, Phe95, Asp 96 and Trp 103). In contrast, PG0361(35–182) exhibits a mixed charge surface potential and four clefts C1-C4 (Fig. 1h) in rather different locations (C1, 77 Ų surface area, Leu 64, Glu 65, Arg 96, Val 97, Arg 98, Ser 115, Ile 118, His 119, Ile 123; C2, 51 Ų surface area, Glu 65, Leu 69, Lys 81, Arg 98, Glu 100, Thr 101, Gly 102, Glu 106, Asp 111; C3, 34 Ų surface area, Arg 95, Arg 96, Ile 123, Phe 126, Arg 127; C4, 26 Ų surface area, Ile 59, Gly 60, Asp 61, Ala 62, Leu 64, Gln 94, Arg 96). The only common feature appears to be that cavities C2 are negatively charged in both proteins, and that they exhibit the highest degrees of conservation within their non-overlapping modeling families (see below).

Different functions for CG2496(41–180) and PG0361(35–182) are also suggested by the genomic context of the full-length proteins. Operon prediction [30] using the MicrobesOnline server (http://www.microbesonline.org) indicates that the gene encoding protein CG2496 is transcribed individually. In contrast, the gene of PG0361 is part of an operon also containing the genes pyrB and pyrI which encode subunits of an aspartate carbamoyl transferase catalyzing the first step of de novo pyrimidine biosynthesis.

Finally, identification of modeling families as was described previously [4, 31] reveals that the novel structural leverage, i.e., the number of protein structures that can be reliably modeled using the experimental structures presented here, is 7 and 368 for CG2496(41–180) and PG0361(35–182), respectively. However, structural leverage is dependent on the methods used for modeling [4, 31], and as homology modeling methods advance the leverage of these structures will also expand. Thus, considering that currently PF04536 contains 1145 non-redundant sequences, the two NMR structures presented here provide high leverage and conservatively ∼33% structural coverage for the very large Pfam family PF04536.

Abbreviations

C-Ala domain

C-terminal domain of alanyl-tRNA Synthetase

DSS

4,4-dimethyl-4-silapentane-1-sulfonate sodium salt

DTT

Dithiothreitol

MES

2-(N-morpholino)ethanesulfonic acid

NESG

Northeast structural genomics consortium

NOE

Nuclear overhauser effect

PDB

Protein Data Bank

RMSD

Root mean square deviation