A new activity for an old enzyme: Escherichia coli bacterial alkaline phosphatase is a phosphite-dependent hydrogenase

Abstract

Genetic analysis indicates that Escherichia coli possesses two independent pathways for oxidation of phosphite (Pt) to phosphate. One pathway depends on the 14-gene phn operon, which encodes the enzyme C-P lyase. The other pathway depends on the phoA locus, which encodes bacterial alkaline phosphatase (BAP). Transposon mutagenesis studies strongly suggest that BAP is the only enzyme involved in the phoA-dependent pathway. This conclusion is supported by purification and biochemical characterization of the Pt-oxidizing enzyme, which was proven to be BAP by N terminus protein sequencing. Highly purified BAP catalyzed Pt oxidation with specific activities of 62-242 milliunits/mg and phosphate ester hydrolysis with specific activities of 41-61 units/mg. Surprisingly, BAP catalyzes the oxidation of Pt to phosphate and molecular H2. Thus, BAP is a unique Pt-dependent, H2-evolving hydrogenase. This reaction is unprecedented in both P and H biochemistry, and it is likely to involve direct transfer of hydride from the substrate to water-derived protons.

Fig. 1.

Two pathways for Pt oxidation in E. coli. Growth on media containing Pt as the sole P source requires oxidation of Pt to phosphate. Strains differing only in the phn and phoA loci, which encode C-P lyase and BAP, respectively, were streaked on 0.4% glucose/Mops medium with either 0.5 mM phosphate or 0.5 mM Pt as sole P sources. All strains grew on phosphate medium (positive control). Strains that are phoA⁺ and/or phn⁺ are capable of growth on Pt medium, indicating that they can oxidize Pt; however, strains deleted for both phn and phoA cannot grow on Pt medium, indicating that they are incapable of Pt oxidation. Thus, two independent pathways for Pt oxidation exist in E. coli: one depends on phoA and the other depends on phn. Because the amount of P required for growth is relatively small, the contaminating levels of phosphate found in many media components allow a slight background level of growth of all strains on these media. To control for this variable, the strains in question were always compared with suitable positive and negative controls on the same plate. The following strains were used: BW13711 [ΔlacX74, phn(EcoB)], BW14894 (Δlac X74, Δphn 33–30), BW14322 [Δlac X74, ΔphoA532, phn(EcoB)], and BW14893 (Δlac X74, ΔphoA532, Δphn 33–30). The phn(EcoB) allele is phenotypically Phn⁺ (13).

Fig. 2.

The products of the BAP-catalyzed Pt oxidation are phosphate and molecular H₂.(A) The proton-decoupled ³¹P nuclear magnetic resonance spectra and peak positions are indicated for 5 mM Pt and 5 mM phosphate controls, as well as for an overnight reaction that initially contained 5 mM Pt plus 500 μg of purified BAP. A peak for unreacted Pt and a new phosphate peak are evident in the spectrum of the enzymatic reaction, but no other products were produced. The slight shift in the position of the phosphate peak between the control and the BAP-catalyzed reaction product is due to minor pH differences: addition of phosphate stock solution to the assay increased the height of the P_i peak but did not produce extra peaks. Proton-coupled spectra are completely consistent with this interpretation (data not shown). (B) BAP-catalyzed Pt oxidation produces stoichiometric amounts of phosphate and molecular H₂. In vitro reactions containing the indicated components were incubated overnight at 37°C in sealed vials. The headspace atmosphere was then assayed for H₂ by gas chromatography with thermal conductivity detection, whereas the aqueous phase was assayed for phosphate by using the malachite green assay. Reactions (3 ml) were conducted in 50 mM Mops (pH 7.0) with 50 mM Pt (pH 7.0) and 387 μg of purified BAP, as indicated. Triplicate controls containing either BAP or Pt alone were performed, but neither phosphate nor H₂ was detected under these conditions. In vivo results show the amount of H₂ produced during growth of WM3924 (Δlac X74, Δphn 33–30, ΔhypABCDE) in 0.4% glucose/Mops broth containing 2.5 μmol of the indicated P source. Because this level of P (500 μM) is growth limiting, the P source is expected to be completely assimilated when the cultures reach saturation. After overnight growth at 37°C in sealed vials, the headspace was assayed for H₂ by gas chromatography with thermal conductivity detection. Both the in vitro and in vivo experiments were conducted aerobically (i.e., O₂ was present during Pt oxidation).

Fig. 3.

Pt oxidation is unique to E. coli alkaline phosphatase. (A) The ability of B. subtilis and P. stutzeri to oxidize Pt, as demonstrated by growth on media with Pt as the sole P source, was tested, as described in Fig. 1. Neither organism grew on the Pt medium, whereas both organisms grew on phosphate medium. Thus, despite the fact that both organisms have active phosphatases, neither can oxidize Pt at rates sufficient to support growth. The P. stutzeri WM3617 (ΔptxA-htxP, Δphn) contains mutations that eliminate the two characterized Pt oxidation pathways of this organism but that do not effect phosphatase expression (9, 10). (B) BAP (E. coli alkaline phosphatase), SAP (shrimp alkaline phosphatase; Roche Applied Science, Mannheim, Germany) and CIP (calf intestinal phosphatase; Sigma) were assayed for Pt oxidation as described above. We used 56 μg of the indicated phosphatase in each assay, which corresponds to 3 BAP, 32 SAP, and 65 CIP phosphatase units, as measured by pNPP hydrolysis. Despite the much higher phosphatase activities of the eukaryotic enzymes, only BAP catalyzed Pt oxidation at significant rates. The average of two trials is plotted.

Fig. 4.

Pt oxidation by BAP may occur by means of hydrolysis with hydride anion as the leaving group. (A) The chemical mechanism for phosphate ester hydrolysis by BAP involves nucleophilic attack by an activated serine residue (Ser-102) on the phosphate ester to form a phosphoserine enzyme intermediate. The alkoxide leaving group rapidly acquires a proton from solution to form the corresponding alcohol. It seems to be likely that Pt oxidation occurs by means of a similar mechanism with hydride anion as the leaving group. (B) The role of Ser-102 in Pt oxidation was tested by examining whether a mutant carrying a Ser-102–Ala mutation in the phoA gene could grow on Pt media, as described in Fig. 1. The mutant failed to grow on Pt medium, demonstrating the requirement for the active-site Ser-102 in Pt oxidation. The host strain was BW14893 (Δlac X74, ΔphoA532, Δphn 33–30), WM3610 and WM3611 carry single copy integrants of plasmids pKY1 and pKY2, which encode the wild-type phoA gene (phoA+) or phoA-S102A mutant (Ser-102–Ala), respectively.