Structure Validation

After three-fold refining of the SynWH_Zur1 using i3Drefine software (Bhattacharya & Cheng,

2013), the final refined model for SynWH_Zur1 was verified by the Structural Analysis and

Verification Server (SAVES) to evaluate its stereo-chemical quality

(http://services.mbi.ucla.edu/SAVES/). After optimization, most residues fall within the

allowed regions of the Ramachandran plot (Laskowski et al., 1993; Kleywegt & Jones, 1996) as defined by the PROCHECK24 program. PROCHECK showed that 92.2% of the residues

were present in the most favoured regions (A, B, L); 6.9% of the residues were in additional

allowed regions (a, b, l, p); and 0.9% of the residues were in generously allowed regions (~a,

~b, ~l, ~p); no residues were present in disallowed regions (see Appendix F, Figure F.02A).

ERRAT25 _{analysis showed the overall structural quality of the predicted structure to be 96.83,}

22_{C-score is a confidence score for estimating the quality of predicted models by I-TASSER. This score calculated}

based on the threading template alignments and the convergence parameters of the structure assembly simulations.

23_{TM-score is a proposed scale for measuring the structural similarity between two structures.}

24_{PROCHECK: Checks the stereo-chemical quality of a protein structure by analysing residue-by-residue}

geometry and overall structural geometry calculated by analysis of the phi (Φ) and psi (ψ) torsion angles as determined by Ramachandran plot statistics.

25_{ERRAT: Analyses the statistics of non-bonded interactions between different atom types and plots the value of}

the error function versus the position of a 9-residue sliding window calculated by a comparison with statistics from highly refined structures.

169 which is very good both experimentally and computationally (see Appendix F, Figure F.02B).

VERIFY_3D26 _{analysis showed that 85.10% of residues gave an average 3D-1D score >= 0.2,}

showing good primary sequence to tertiary structure compatibility (see Appendix F, Figure

F.02C).

In order to assess the reliability of the modelled structures, the root mean square deviation

(RMSD) for the two homology models and other known available structures for Zur proteins

were calculated. Although a range of structural data are available for the Fur family, only the

structures for three Zurs are known to date i.e., MtZur from Mycobacterium tuberculosis

(Lucarelli et al., 2007), ScZur from Streptomyces coelicolor (Shin et al., 2011) and EcZur from

E. coli (Gilston et al., 2014). To assess how close the topology of the initial structure for SynWH_Zur1 was to MTZur (33.3%), ScZur (28.6%) and EcZur (26.4%), the backbone of

SynWH_Zur1 was superimposed with three known Zurs (Figure 6.12, and see Appendix F,

Table F.01).

The superposition was performed using the Superpose program (Maiti et al., 2004) and Pymol software, in which a low RMSD value corresponds to similarity in the structure of the two

proteins. Use of the structural alignment program SuperPose (Maiti et al., 2004) indicates that if the superposition value is <1.6 Å, RMSD for the α-carbons, this provides an indication that

there is a strong agreement between the two selected structures, while if the superposition value

is RMSD >3.0 Å, this will indicate poorer agreement. RMSD values between SynWH_Zur1

and ScZur gave the smallest values ~ 1.1 Å and 2.97 Å in SuperPose and Pymol, respectively.

At the same time SynWH_Zur1 and EcZur gave ~ 2.1 Å and 3.17 Å in SuperPose and Pymol,

26 _{VERIFY_3D: Determines the compatibility of an atomic model (3D) with its own amino acid sequence (1D) by}

assigning a structural class based on its location and environment (alpha, beta, loop, polar, nonpolar etc) and comparing the results to good structures.

170 respectively, while SynWH_Zur1 and MtZur were ~ 6.2 Å and 5.55 Å. Remarkably,

SynWH_Zur2 and MtZur are ~ 1.93 Å in Pymol (see Appendix F, Table F.01).

Figure 6.12 Superimposition of SynWH_Zur1 with three known Zur proteins.

Ribbons represent the superposition of the C-terminal dimerization domain for homology model SynWH_Zur1 from Synechococcus sp. WH8102 (salmon) with (A) EcZur from E. coli ( (Gilston et al., 2014); pdb 4MTE, green), (B) ScZur from S. coelicolor ((Shin et al., 2011); pdb 3MWM, brown) and (C) MtZur from M. tuberculosis ((Lucarelli et al., 2007), pdb 2O03, grey).

A

B

171 These comparisons reveal that SynWH_Zur1 aligns with a reasonable degree of similarity to

EcZur and ScZur. Moreover, it was reported that ScZur from Streptomyces coelicolor (Shin et al., 2011) and EcZur from E. coli (Gilston et al., 2014) are found in the closed conformation, while MtZur from Mycobacterium tuberculosis (Lucarelli et al., 2007) is thought to correspond to the open conformation. We have compared SynWH_Zur1 to the MtZur open conformation,

which shows a good agreement across the C-terminal dimerization domains, but differs

significantly in the relative positioning of the N-terminal DNA binding domains (Figure 6.12C).

The “open” conformation is postulated to have a low affinity for DNA, as the N-terminal DNA- binding domains are too far apart to interact with a DNA molecule (Figure 6.12C). Structures

thought to be determined in the closed conformation, ScZur (Shin et al., 2011) and EcZur (Gilston et al., 2014), proposed to both correspond to the fully metal-loaded form of the repressor and to have a high affinity for DNA, showed good agreement across the C-terminal

dimerization domains and the N-terminal DNA binding domains (Figure 6.12A and B).

The overall structure of SynWH_Zur1 with low RMSD values corresponds to protein structures

in the ‘‘closed’’ state which may be capable of binding to their respective operator DNA. SynZur is a homodimer with a modular architecture similar to previously determined Fur-

family members, especially ScZur (Shin et al., 2011) and EcZur (Gilston et al., 2014). The three dimensional model for SynWH_Zur1 has a DNA-binding domain (residues 1–76) and a C-

terminal dimerization domain (residues 77–134). The DNA binding domain is composed of

three helices followed by a two-stranded antiparallel sheet (β1 and β2). This domain exhibits

the typical winged helix motif with three helices bundles (α1, α2 and α3), where the helix α1

and α3 may be the putative DNA recognition region (residues 9– 22 for α1 and 43–56 for α3). The dimerization domain comprises three antiparallel strands (β3, β4 and β5) and two helices

(α4 and α5) stabilized by one of the metal-binding sites. The DNA-binding domains of SynWH_Zur1 and (EcZur, ScZur) share the same basic fold, with one important difference

172 when superimposed. EcZur contains an additional N-terminal helix (residues 4-20) that may be

involved in DNA recognition (Figure 6.12A). Four cysteine residues (Cys-83 and Cys-86, Cys-

124, Cys-126) cluster to form a high-affinity Zn(II) binding site (Cys4Zn) this is the structural

site 1 (Figure 6.13). Another four residues (Asp-77, His-79, Cys-95, His-115) are candidate

ligands for the sensory site 2 (Figure 6.13).

Figure 6.13 Structural features for SynZur monomer form.

Homology model for the SynZur protein, encoded by the synw_2401 gene, from Synechococcus sp. WH8102; only one monomer is shown, with putative ligands from the structural site 1 (Green) and the sensory site 2 (Purple) highlighted. SynZur is shown in salmon and zinc ions as red spheres.

The electrostatic charge distribution over the SynZur monomer is shown in Figure 6.14, and

reveals a predominantly positive charge for the helical regions predicted to be involved in DNA

recognition, particularly helix α1 and α3.

Site 1: C83, C86 C123, C126 Site 2: D77, H79, C95, H115

173

Figure 6.14 Electrostatic potential map of the predicted SynZur protein.

The electrostatic surfaces for SynZur monomer calculated using Adaptive Poisson-Boltzmann Solver (APBS) (Baker et al., 2001) generated using the PyMOL software (DeLano, 2007). The electrostatic surface of the SynZur monomer (A) and with 40% transparency for the surface (B). A colour gradient from blue to white to red is used to colour the molecular surface where red, white and blue are for negative, neutral, and positive potentials, respectively. The arrow on the blue surface indicate the DNA binding domain.

6.6 Summary and Conclusion

Optimisation of conditions for protein NMR studies revealed that 150 mM NaCl at a

temperature of 308K gave the best conditions for SynZur. In addition, a significant

improvement in sensitivity and dispersion was achieved using the TROSY technique on the

SynZur protein rather than the conventional HSQC.

However, because a good homology model can be an important step towards determining

functional annotation for a protein, a 3D homology model for SynZur was created. Moreover,

it was found that SynZur is a homodimer with a modular architecture similar to that determined

for other Fur-family members.

α1

α2

α3

α1

α3

α2

A

_B

174