Characterization of the Escherichia coli pyridoxal 5'-phosphate homeostasis protein (YggS): Role of lysine residues in PLP binding and protein stability

Abstract

The pyridoxal 5'-phosphate (PLP) homeostasis protein (PLPHP) is a ubiquitous member of the COG0325 family with apparently no catalytic activity. Although the actual cellular role of this protein is unknown, it has been observed that mutations of the PLPHP encoding gene affect the activity of PLP-dependent enzymes, B₆ vitamers and amino acid levels. Here we report a detailed characterization of the Escherichia coli ortholog of PLPHP (YggS) with respect to its PLP binding and transfer properties, stability, and structure. YggS binds PLP very tightly and is able to slowly transfer it to a model PLP-dependent enzyme, serine hydroxymethyltransferase. PLP binding to YggS elicits a conformational/flexibility change in the protein structure that is detectable in solution but not in crystals. We serendipitously discovered that the K36A variant of YggS, affecting the lysine residue that binds PLP at the active site, is able to bind PLP covalently. This observation led us to recognize that a number of lysine residues, located at the entrance of the active site, can replace Lys36 in its PLP binding role. These lysines form a cluster of charged residues that affect protein stability and conformation, playing an important role in PLP binding and possibly in YggS function.

Keywords: COG0325 family; PLPHP; YggS; pyridoxal 5′-phosphate; vitamin B6.

FIGURE 1

Properties of recombinant wild‐type YggS. (a) Absorption spectrum of apo‐ (red line) and holo‐YggS (blue line) measured in 50 mM NaHEPES buffer, pH 7.6. (b) PLP‐binding curves were obtained with fluorimetric measurements using 100 nM apo‐YggS in 50 mM NaHEPES, pH 7.6, upon excitation at 280 nm. Fluorescence change is expressed as fractional variation as a function of total PLP concentration and analyzed with Equation (1) described in Section 4. The inset shows the binding stoichiometry analysis obtained with 1 μM apo‐YggS. Fluorescence change, expressed as fractional variation as a function of the [PLPtot]/[protein] ratio, is linear as shown by the thick continuous line, up to the stoichiometry point corresponding to the crossing with the horizontal line. (c) Variation of the melting temperature of YggS, obtained by DSF analysis, as a function of the PLP concentration. Inset: zoom of the same saturation curve in the range 0–20 μM PLP. (d) Melting temperatures of apo‐ (red) and holo‐YggS (black) as a function of pH (left panel) and in the presence of different concentrations of NaCl (right panel)

FIGURE 2

Crystal structure of WT YggS. (a) Superposition of holo‐YggS (green) with apo‐YggS (salmon). (b) Close‐up figure of the active sites (K36‐binding sites) of holo‐YggS (green) and apo‐YggS (salmon), the former with bound PLP. (c) Close‐up figure of the secondary PLP binding sites (K89‐binding sites) of holo‐YggS and apo‐YggS, the former with bound PLP. (d) Close‐up figure of the active sites of holo‐YggS (green) and YggS‐PNP (salmon)

FIGURE 3

Spectroscopic properties of lysine variants of YggS. Absorption (upper panels) and CD spectra (lower panels) of (a) K36A single variant; (b) K36A/K38A, K36A/K137A, and K36A/K233A/K234A double variants; (c) K36A/K38A/K137A, K36A/K38A/K233A/K234A, K36A/K137A/K233A/K234A triple and K36A/K38A/K137A/K233A/K234A quadruple variants; (d) K38A/K137A/K233A/K234A, K137A and ΔK233/K234 K36‐containing variants. Spectra of WT‐YggS (in blue) is reported in all panels as reference

FIGURE 4

Fragments observed in the MSMS spectrum. (a) MSMS spectrum of the WT tryptic peptide SPEEITLLAVSK*TKPASAIAEAIDAGQR, m/z = 755.1628 and z = 4, with a mass shift of +151.06 Da at K*, corresponding to reduced PLP after phosphatase treatment. Mascot Score 73; Sequest Xcorr 11.54. (b) MSMS spectrum of the K36A tryptic peptide SPEEITLLAVSATK*PASAIAEAIDAGQR, m/z = 740.8958 and z = 4, with a mass shift of +151.06 Da at K*, corresponding to reduced PLP after phosphatase treatment. Mascot Score 60; Sequest Xcorr 8.15

FIGURE 5

Limited proteolysis of YggS forms. SDS/PAGE analysis of YggS limited proteolysis time course, in the absence or presence of PLP. Apo‐forms of YggS (6.5 μM; the calculated molecular weight of the recombinant protein with the poly‐His tag is 26,853 Da) were exposed to trypsin digestion either in the presence (+PLP) or absence (Apo) of 20 μM PLP (see methods section for details). Arrows indicate the main fragments resulting from trypsin digestion (green arrow, 26 kDa band; red arrow, 21 kDa band; molecular weights were estimated using the MS Image Capture‐A software of the Major Science Digimage System), which were analyzed by mass spectrometry as explained in the text. Numbers on the right correspond to molecular weights (in kDa), as indicated by the protein molecular weight standard (PageRuler prestained protein ladder; ThermoFisher Scientific) that was used in the electrophoretic run, which for simplicity was not included in the figures. Some images were generated by splicing together different gels in order to allow the comparison of different data sets.

FIGURE 6

(a) Superposed close‐up figure of active site of holo‐YggS (green) and K36A YggS (magenta) showing covalently and non‐covalently bound PLP. (b) Superposition of holo‐YggS (green) with K137A YggS (yellow) showing the mutation site at residue 137. (c) Superposed close‐up view of the active site showing covalent PLP binding in holo‐YggS‐PLP (green), and non‐covalent binding in K36A/K137 (cyan) and K36A/K38A/K233A/K234A variants (amber)

FIGURE 7

PLP transfer from YggS to eSHMT. (a) The activity of eSHMT (expressed as percentage of the total activity as a function of time) was measured as the initial velocity of the conversion of l‐serine and tetrahydrofolate to glycine and 5,10‐methylene tetrahydrofolate. The reactivation of apo‐eSHMT in the presence of 100 μM PLP (black line) was used as the 100% activity reference, assuming that in these conditions SHMT was completely in the holo‐form. Transfer kinetics were measured in the presence of the indicated YggS variants. (b) The kinetics of the PLP hydrolysis was measured in order to compare k _off of PLP dissociation from WT and variant forms of holo‐YggS, as explained in the text. Data are represented as percentage of hydrolyzed PLP, with respect to total protein‐bound PLP, as a function of time

FIGURE 8

In vivo complementation of the 4‐deoxypyridoxine (4dPN) sensitivity phenotype by expression of YggS variants in trans. 4dPN sensitivity assays were carried out on M9 0.4% glucose minimal media supplemented with 1 μM PN (Amp 100 μg ml⁻¹, +/− Arabinose [inducer]) as described in methods. Plated bacterial cells (OD₆₀₀ = 0.006) were treated with 20 μl of 25 mM 4dPN in a central well, revealing a clear 4dPN sensitivity phenotype in E. coli ΔyggS that is complemented by expression of WT yggS in trans. Assays were repeated to investigate the ability of YggS lysine mutation variants to complement the 4dPN sensitivity phenotype, the figure shows that only the loss of K36 or K137 activity results in loss of complementation (all variant trials can be found in Figure S13). All plate sensitivity experiments were repeated at least three times and expression of each YggS variant was confirmed by western blot analysis (Figure S14)