The gene phyA encoding phytase was isolated from Obesumbacterium proteus genomic library and sequenced. The cleavage site of the PhyA signal peptide was predicted and experimentally proved. The PhyA protein shows maximum identity of 53% and 47% to phosphoanhydride phosphorylase from Yersinia pestis and phytase AppA from Escherichia coli, respectively. Based on protein sequence similarity of PhyA and its homologs, the phytases form a novel subclass of the histidine acid phosphatase family. To characterize properties of the PhyA protein, we expressed the phyA gene in E. coli. The specific activity of the purified recombinant PhyA was 310 U mg−1 of protein. Recombinant PhyA showed activity at pH values from 1.5 through 6.5 with the optimum at 4.9. The temperature optimum was 40–45 °C at pH 4.9. The Km value for sodium phytate was 0.34 mM with a Vmax of 435 U mg−1.
Phytate (myo-inositol hexakisphosphate) is the main storage form of phosphorus in seeds . Monogastric animals are incapable of digesting phytate phosphorus due to lack or low level of the phytase activity in the intestine [2,3]. Phytates are regarded as decreasing feed quality because they bind proteins and mineral . Further, undigested phytates cause significant environmental pollution .
Phytases (EC 184.108.40.206) belong to the family of histidine acid phosphatases, a subclass of phosphatases  capable of hydrolyzing phytates and releasing phosphate groups. They are successfully used as a feed additive. The characteristics phytase that important in practical applications are high specific activity, pH optima corresponding to various departments of the animal digestive tract, resistance to stomach proteases, drying and high temperature. This creates a need for phytases with an optimal combination of various properties.
A number of microorganisms, mainly of fungal and bacterial origin, are capable of phytase synthesis [7–15]. Recently, artificial phytases with superior thermostability were designed using semi-rational sequence comparison methods based upon several mesophilic homologues phytases from fungi (with sequences identity of about 50–70%) . Phytase AppA from Escherichia coli is known to have the greatest specific activity  but it is inferior in terms of thermostability. In order to construct a more stable enzyme based on the E. coli phytase, it is necessary to have a set of homologs to E. coli phytase sequence. However, this protein has only a few homologs with low sequence identity not exceeding of 30% to other known phytases. We hypothesized that species from the Enterobacteriaceae family related to E. coli would have similar phytases. Indeed, a member of this family Obesumbacterium proteus was recently found to possess periplasmic phytase activity . This activity is expected to be determined by a phytase similar to AppA.
In this paper, the phytase gene from O. proteus was cloned, sequenced, the enzyme recombinantly produced, characterized and compared with phytase from E. coli. Our aim was to expand the range of sequences close to the E. coli phytase in order to study this enzyme group and to develop high production of phytases.
2Materials and methods
2.1Strains, plasmids and chemicals
The following strains were obtained from the Russian Collection of Industrial Microorganisms (VKPM) O. proteus B-4567, B-6897, B-6898. E. coli strains and plasmids: XL-1 Blue (Strategne, La Jolla, CA), BL21 (DE3), pET22b(+) (Novagen, Madison, WI) and pUC19 . All strains were grown at 37 °C in Luria–Bertani (LB) medium . All chemicals were of the analytical grade, commercially available.
2.2Cloning of the phytase gene
The O. proteus B-6898 genomic DNA was partially digested with Mph143I to obtain fragments of size 3–6 kb, and purified using DNA extraction Kit #K0513 (Fermentas AB, Vilnius, Lithuania). The DNA fragments were cloned into the BamHI site of pUC19 and transformed into E. coli XL-1 Blue. Transformants were tested for the phytase activity on LB plates by pouring over the Test Upper Layer cooled to 37 °C and containing (w/v) 1% agarose, 1% phytate, and 0.5% CaCl2. Positive clones caused clear zones around the colonies after 1–6 h of incubating with the upper layer. The clones were directly looped from beneath the upper layer in order to avoid the procedure of clone picking. The colonies harboring plasmids with DNA fragments encoding active phytase were named pPhyAn, where the last symbol n was the serial number. The size of insert was subcloned from the 4.2 kb insert by partial digestion with Mph143I to produce the final insert size of about 2.0 kb. The final pUC19 plasmid that harbored the subcloned DNA insert encoding the phytase activity, was named pPhyAmini. The insert DNA was sequenced by Syntol Co. (Moscow, Russia). The GenBank Accession No. is AY378096.
2.3Expression of O. proteus phyA and E. coli appA
The templates for PCR were the total genomic DNA of O. proteus B-6898 and E. coli XL-1 Blue. PCR amplification of DNA containing the phyA gene including the signal peptide was performed using primers PphyaECF (forward, 5′-CCCATATG ACAATTTCTCTGTTTACACA-3′) and PphyaECR (reverse, 5′-CCGAATTC TATTGGCACTCCACCAGTTCGT-3′) (the NdeI and EcoR1 restriction sites, respectively, are underlined). Amplified DNA was digested by NdeI and EcoRI and ligated into the respective sites of pET22b+ vector DNA. The obtained pETPhyA plasmid was used to transform E. coli BL21 (DE3). To express the E. coli appA gene, the same procedure was performed with different primers: PappaECF (forward, 5′-GCATATG AAAGCGATCTTAATCCCAT-3′) and PappaECR (reverse, 5′-GGGAATTC ATTACAAACTGCACGCCG-3′) (the NdeI and EcoRI restriction sites, respectively, are underlined).
To obtain high level expression of recombinant phytases, cells were incubated for 20 h in 500 ml flasks containing 100 ml of LB plus the following components (w/v): 0.2% glucose, 1.5% lactose or 1 mM IPTG and ampicillin (100 μg ml−1). Recombinant protein produced by appA in E. coli was named AppA-coli. Recombinant protein produced by phyA in E. coli was named PhyA-coli.
2.4Purification of the phytases
After expression of phytase, cells were harvested, resuspended in 50 mM Tris–HCl buffer, pH 7.0 and sonicated at 4 °C. Phytases were purified in two steps (Table 1). At the first step, the majority of extraneous proteins were precipitated at low pH (AppA-coli and PhyA-coli are resistant to low pH of 2.0) by adding to the crude extract sample an equal volume of 0.5 M glycine–HCl buffer (pH 1.8). Solution was incubated 30 min at 37 °C and centrifugured at 14000g. At the second step, the enzymes were purified by FPLC gel-filtration Superdex 75-HR column.
Table 1. Purification scheme
Total protein (mg)
U mg−1 of protein
Precipitation at pH 2.0
Precipitation at pH 2.0
2.5Determining of the N-terminal amino acid sequence of the protein
The phytase samples to be analyzed were prepared by SDS–PAGE followed by semi-dry electroblotting of the separated material onto Immobilon-P membrane. The obtained Immobilon filter strip was used as a carrier in the automated Edman protein sequencing on an Applied Biosystems Procise cLC instrument. The N-terminal sequence was determined using the protein purified by a single-stage FPLC on a Superdex-75HR column.
2.6Activity and properties of the phytase
The effect of pH on phytase activity was determined using the following buffer solutions: glycine–HCl, pH 1.5–3.5; Na acetate–acetic acid, pH 3.5–6.0; Tris–acetic acid, pH 6.0–7.0; and Tris–HCl, pH 7.0–8.0. Phytase activity was measured by accumulation of free phosphate in the reaction mixture detected according to Fiske–Subbarow . The activity unit (U) causes releasing of 1 μmol of phosphate per min.
Phytase and phosphatase pH optima were determined using the above buffers with the concentration of 0.1 M at 37 °C. The temperature optima were determined at the pH optimum (4.9) ranging the temperature values from 20 to 90 °C. The temperature stability was estimated by incubation of the enzyme sample in 50 mM buffers for 30 min at 50, 60, 70, or 80 °C, followed by the 30 min incubation at 5 °C before measuring. The concentration of protein was determined by the Bradford method , the Lowry method  and by measuring the UV absorbance at 280 nm. The Bradford and Lowry assay was not adopted because of very low estimation of the AppA  and PhyA concentration. UV absorbance method was adopted as the most suitable routine method.
3.1Isolation of the gene encoding phytase from the genomic library
Phytase activity of three O. proteus strains (VKPM B-4567, B-6897, B-6898) was measured at the pH optimum (4.9). O. proteus strain VKPM B-6898 (further referred to as O. proteus) has the highest phytase activity of 0.11 U mg−1 of total protein and was chosen to create a genomic library. The phytase activity was detected in 13 of 10,000 clones from the O. proteus genomic library. To compare the phytase activities, cells from stationary cultures of all 13 clones were subjected to sonication. The obtained values were about 3 U mg−1 of total protein with no significant variations for each of the clones. The clone harboring plasmid pPhyA8 with the shortest insert of 4.2 kb was selected and its insert was reduced to 2.6 kb to form the pPhyAmini. The latter insert was sequenced and the sequence deposited in GenBank under Accession No. AY378096. An ORF of 1.3 kb within this insert with homology to the phytase gene appA from E. coli was named phyA.
To detect the presence of the phyA gene in the remaining 12 active clones, PCR analysis of the internal portion of phyA (with primers 5′-GACGCAAACCATGCGCGACGTA-3′ and 5-CATCTGGCATGCCCTGCGCATA-3′) was carried out. This analysis demonstrated that all 13 clones contain the same gene phyA in the O. proteus phytase genomic library.
The orientation of the phyA gene in the plasmid pPhyA8 is in the direction opposite to pUC19 genes and thus cannot start at any of the plasmid promoters. However, the phytase activity of the strain that carries the phyA-containing plasmid considerably exceeds that of the E. coli and O. proteus wild-type cells. Therefore, we suppose that the phyA gene is expressed from a multi-copied vector, and the transcription initializes from the gene's own promoter which is contained within the inserted fragment. Indeed, this fragment contains several promoter-like sequences identified using the SignalX software . However, the specificity of the existing algorithms is insufficient to reliably predict bacterial promoters , and thus the identified candidate sites are only tentative (data not shown).
3.2Sequence and phylogenetic analysis
Several PhyA homologs were found in GenBank using BLAST and aligned (Fig. 1). Using this alignment we constructed a phylogenetic tree (Fig. 2). All these proteins were of bacterial origin. The closest homologs to PhyA were phosphoanhydride phosphorylase from Yersinia pestis (55% identity) and AppA from E. coli (48% identity). Other homologs have only about 30% identity to both O. proteus and E. coli phytases.
Positions of 12 out of 14 amino acid residues responsible for the enzyme–substrate binding in AppA  were identical to the corresponding residues in PhyA (Fig. 1). Positions of all 4 disulfide bonds were identical (Fig. 1).
3.3Overexpression of phyA and appA in E. coli
Cloning of phyA into the pET22b+ vector and expression in E. coli BL21 (DE3) enhanced the enzyme production by two orders of magnitude compared with the wild type O. proteus VKPM-6898. Transcription of the phyA gene in this system depends on the lactose (or IPTG) induction. A completely analogous scheme using the pET22b+ vector was applied to overexpress the appA gene. The phytase activity in E. coli BL21 (DE3) crude extracts was 9.6 and 17.1 U mg−1 of total protein for PhyA-coli and AppA-coli, respectively (Table 1).
3.4Properties of the purified enzymes
The activities of the purified enzymes were 310 U mg−1 and 1420 U mg−1 of total protein for PhyA-coli and AppA-coli, respectively (Table 1). Compared to other known phytases of bacterial and fungal origin, the specific activity of the O. proteus phytase was very high.
Samples of purified PhyA-coli displayed activity at acidic pH values (1.5–6.6) (Fig. 3(a)). The pH-optimum was observed at pH 4.9, and a local pH-optimum was also detected at pH 3.4. The maximum activity was observed at 40–50 °C (Fig. 3(b)). The stability of the enzyme at a short high temperature action is shown in Fig. 3(c). The enzyme is highly resistant to 30-min heating at 50 °C, but loses its activity very rapidly under further increase of the temperature. As a control, characteristics of AppA-coli were measured and plotted on the corresponding graphs (Fig. 3(a)–(c)).
PhyA-coli was very stable for long term incubation in 0.2 M buffer solutions at pH 1.5–8.0, 37 °C, and even after 24 h of incubation no significant changes in the activity could be detected. It is likely that disulfide bonds strongly contribute to the stability of the phytase molecules, which is clearly seen at low pH values. Most E. coli soluble proteins are denatured and precipitated in this range of pH values (Fig. 4).
The predicted mass of the peptide (45.2 kDa) was agreed with the SDS–PAGE value (about 45 kDa) (Fig. 4).
3.5Identification of N-termini position of PhyA
The most likely position of the cleavage site at the N-terminus, Ser33–Ala34, was determined using the SignalP program and multiple sequence alignment. The sequence of the N-end of the mature PhyA protein purified from the E. coli crude extract, SETEPSGYQLEKVVI, confirmed this prediction. Unlike the mature enzymes, the leader peptides of PhyA homologs from Y. pestis and E. coli show no sequences similarity, nor even conservation of the peptide length. This data on the localization of the N-terminus of the mature PhyA protein is a necessary step in development of the enzyme expression in heterologous systems, including yeast, to reach high level expression of industrial proteins.
3.6Substrate specificity of PhyA-coli
PhyA-coli was shown to cleave some phosphorus-containing organic substances other than phytate (Table 2). The kinetic parameters for the hydrolysis of phytate has a Km of 0.34 mM and a Vmax of 435 U mg−1 as determined from a Lineweaver–Burke plot. However, similarly to the E. coli phytase AppA , PhyA-coli degrades phytate with much higher rate than any other substrate under study, and thus it is a true phytase.
Table 2. Substrate specificity of PhyA-coli
U mg−1 of proteina
aValues represent the arithmetic average of two measurements.
The PhyA sequence from O. proteus showed high identity with the E. coli phytase. The high identity of E. coli AppA and O. proteus PhyA sequences, position of the active center and the cysteine residues allows us to suggest that the proteins may have similar three-dimensional structure and a similar mechanism of the enzyme action. The same similarity have been observed for both these phytases and their homolog with unknown function from Y. pestis. All these proteins form a particular branch in the phylogenetic tree. So it is possible that the protein from Y. pestis is also a phytase. Further, several bacterial species are known to have phytases with similar properties but unknown sequences. These bacteria are Enterobacter cloaceae, Citrobacter freundii and Citrobacter braakii, from the Enterobacteriaceae family. It is likely that their phytases also have similarity to the E. coli phytase AppA.
Note that homologous proteins aligned and depicted in the tree (Fig. 2) are produced from Gram-negative bacteria most of which (7 of 10) belong to the Enterobacteriaceae family. These proteins were found using BLAST and the identity between the members of the group is about 30–50% throughout the entire protein length. Other known phytases of the fungal origin demonstrated no significant similarity to the bacterial phytases except the conserved motif RHGXRXP of the histidine acid phosphatase EC 220.127.116.11 family . So, based on the protein sequence similarity, the studied proteins form a separate subclass of this protein family. Similarly, a separate subclass of fungal phytases was established earlier .
Unlike other phytases whose substrate specificity has been studied , the E. coli and O. proteus phytases show specificity to phytate that is 10- to 100-fold higher than specificity to other phosphorylated compounds. On the other hand, known phytases of Enterobacteriaceae are frequent in the animal intestine compartments (with neutral or weak alkaline pH of 7–8)  where the phytase activity toward phytate must be non-existent. Thus, the role of phytases in bacterial metabolism remains unclear. Nevertheless, the phytase activity at low pH values is exactly the feature required in agricultural manufacturing to use them as feed additives for farm animals with the stomach pH values of 2–6 .
We have obtained 100-fold increase in the PhyA synthesis during overexpression in E. coli cells. The recombinant enzyme PhyA-coli is not inferior to the wild type enzyme in such industrial characteristics as temperature resistance, pH optimum, as well as activity at low pH values of 1.5–6.5 , (Fig. 3). AppA-coli, the closest homolog of PhyA-coli, has somewhat higher specific activity. Although the O. proteus phytase is inferior to E. coli phytase in specific activity, it also can be used as feed additives as an alternative enzyme.
This study demonstrated that the phyA gene could be used for developing strains-producers of the enzyme with higher expression level. The high identity between PhyA and AppA may be helpful in investigation of the structure, evolution, and mechanisms of the enzyme action. It can also be used to compute the modifications of the enzyme in order to enhance its industrial potential using rational or semi-rational sequence comparison methods.
We are grateful to U.A. Rybakov, V.V. Samsonov, T.V. Uzbashev, and A.B. Rakhmaninova for useful discussion. The research was partially supported by grants from the Russian Fund of Basic Research (#03-04-48433), the Howard Hughes Medical Institute (#55000309), The Fund for Support of the Russian Science, and the Program “Molecular and Cellular Biology of the Russian Academy of Sciences”.