Investigation of the ligand binding site in the glucuronide-H+ symporter, GusB, in Escherichia coli, using chemical and solid state NMR spectroscopy approaches

J. A. Rajakarier, H. Venter, A. R. Brough, F. Melvin, R. B. Herbert and P. J. F. Henderson

Introduction

Alkyl or aryl b -D-glucuronides (Figure 1) are important natural compounds, often being the conjugates by which xenobiotics or drugs are excreted from the human body. The drug connection provides a pharmaceutical interest for the synthesis of these compounds.

Figure 1: b -D-glucuronide where X= O, S or NH; R= alkyl or aryl group

The source of all glucuronides in the mammals is UDP-glucuronide, which possesses the capacity to conjugate with a wide range of molecular groups namely R-OH, R-COOH, R-NH2 or R-SH. Glucuronides are produced in the liver cells and are then excreted into the bile and removed by the urinary tract or large intestine. Escherichia coli in the human gut uses the glucuronides as a carbon source, assimilating them through the cell membrane by agency of membrane protein GusB, one of a family of important bacterial transport proteins.

We have cloned, sequenced and over expressed the GusB protein to high levels. The gusB gene encodes 457 amino acid residues with a theoretical molecular weight of 49,982Da. Based on the amino acid sequence, GusB is deduced to consist of 12 trans membrane a -helices with N- and C- terminals on the cytoplasmic side of the membrane. Each helix contains approximately 20-25 amino acid residues. GusB is a member of the galactose-pentose-hexuronide (GPH) transport family and shares significant homology with other members within this family.

Aim of the research

We are applying novel cross polarisation-magic angle spinning solid state NMR (CP-MAS NMR) methods to determine the 3-D structure of the binding site of GusB in its native membrane state. This technique is based on the interaction of stable isotope labels (13C, 15N, 19F) introduced at specific sites in the ligand (glucuronide) and the protein (GusB) by means of short-range (5-15Å) polarisation transfer. By making quantitative distance measurements a picture of the ligand-binding site can be constructed.

Chemical synthesis of substrates for CP-MAS NMR and bioassay

Starting with [1-13C]-labelled glucose (which is readily available commercially) we have developed a high yielding route to prepare methyl [1-13C]-b -D-glucuronide over five steps. The overall reaction yield is 70% (Figure 2).

 

 

Figure 2: Total synthesis of methyl [1-13C]-b -D-glucuronide and 1-(3’-trifluoroethyl)-b -D-glucuronide for carbon and fluorine solid state NMR experiments, respectively

The pivotal step in the synthesis is the oxidation of methyl [1-13C]-b -D-glucoside to methyl [1-13C]-b -D-glucuronide by tetramethylpiperidinyl-1-oxy (TEMPO) and t-butyl hypochlorite. This oxidation method is readily applicable to the preparation of a wide range of glucuronides from other glucoside derivatives. All the products obtained were spectroscopically characterised.

Similarly, various glucuronide derivatives such as fluoro-deoxy, epimers, deoxy and carbonyl protected glucuronides were chemically synthesised and tested in biological assays to establish the importance of the corresponding hydroxyl groups in the substrate recognition by GusB.

CP-MAS solid-state NMR experiments

CP-MAS solid-state NMR was carried out at 75.46MHz for 13C nuclei (300.13MHz for 1H) using a Bruker (Billerica, MA) MSL-300 spectrometer in Leeds.

The proton-decoupled 13C CP-MAS NMR spectra that were recorded on inner membrane vesicles containing 0.15m mol (8mg) of GusBH protein is shown in Figure 3A. The spectrum is dominated by the signal from the natural abundance 13C nuclei of the alkyl groups (d C 0– 50ppm) from the phospholipids, but the region expected to accommodate the resonance from the substrate (d C 103ppm) is free of any natural abundance contributions.

After methyl [1-13C]-b -D-glucuronide (6mM) was added to the membranes, the CP signal arising from a motion-restricted part of this substrate emerges in the anticipated spectral region as shown in Figure 3B. It could be shown that the signal at d C 103ppm arises from the specific binding of methyl [1-13C]-b -D-glucuronide to the protein, as opposed to non-specific binding to the lipid bi-layer, by competing out the methyl glucuronide with another substrate. p-Nitrophenyl b -D-glucuronide is such a substrate that binds to GusBH with a similar affinity to methyl b -D-glucuronide. The methyl b -D-glucuronide signal disappeared when unlabelled p-nitrophenyl b -D-glucuronide was added to the GusBH containing membranes in ten-fold excess (60mM) over the amount of methyl b -D-glucuronide added (Figure 3C).

 

Figure 3: Proton decoupled 13C CP-MAS solid-state NMR spectra from inner membranes containing 0.15m mol of GusB(His)6: without any substrate (A); with methyl [1-13C]-b -D-glucuronide (6mM) added (B) showing intensity from bound substrate (at d C 103ppm); and when the glucuronide was competed out by treatment with unlabelled p-nitrophenyl b -D-glucuronide (60mM) (C).

As further proof that the substrate signal arises from methyl [1-13C]-b -D-glucuronide motionally restricted in the binding site of GusB, negative control membranes were prepared. The plasmid for over-expression of GusBH (pWJL24H) is derived from the high copy number plasmid pTTQ18. A suitable negative control would therefore be membranes from NM554 (pTTQ18) prepared with similar conditions of growth and induction as for GusBH-containing membranes. The CP-MAS NMR spectrum recorded from inner membrane vesicles without any GusB, but with the same amount of total protein (35mg) as the GusBH containing membranes, is shown in Figure 4A. No substrate signal was observed at d C 103ppm when labelled methyl glucuronide was added to the negative control membranes (Figure 4B), showing that the signal observed in the GusBH containing membranes arises from substrate bound to GusB.

Figure 4: Proton decoupled 13C CP-MAS solid-state NMR spectra from inner membranes without GusB expressed. Membranes without any substrate (A) and after addition of methyl [1-13C]-b -D-glucuronide (6mM) (B).

The initial results showed that methyl b -D-glucuronide can be observed specifically bound to GusB and it is possible to displace the peak by competitive displacement. Furthermore, variable contact time (CT) and de-phasing delayed cross polarisation (DDCP) NMR techniques were used to confirm the peak observed at d C 103ppm was due to methyl [1-13C]-b -D-glucuronide bound to the GusB (experimental details are not included).

Biological assays

The structural requirements for glucuronide binding to the GusB protein were explored by testing various synthetic glucuronide derivatives as inhibitors for acetamidophenyl [1-14C]-glucuronide uptake in E. coli cells. It was found that all the substituents needed to be equatorial for efficient inhibition. Further and importantly, the C-3-OH is essential and is a H-bond donor; also the C-4-OH is important and acts as an acceptor; and the C-2-OH is not important. Naturally the carboxyl group (C-6) is essential.

Conclusions

CP-MAS solid state NMR experiments have been carried out in order to identify the specific binding of the methyl [1-13C]-b -D-glucuronide to GusB.

A variety of methyl b -D-glucuronide derivatives have been chemically synthesised and the importance of the hydroxyl groups, carboxyl group and stereochemistry for binding to GusB have been established in biological assays.

Acknowledgements

We wish to acknowledge BBSRC for financial support.