Aihua Liang, Key Laboratory of Chemical Biology and Molecular Engineering of Ministry of Education, Institute of Biotechnology, Shanxi University, Taiyuan 030006, China. E-mail: email@example.com
Phytases are a group of enzymes capable of releasing phosphates from phytates, one of the major forms of phosphorus (P) in animal feeds of plant origin. These enzymes have been widely used in animal feed to improve phosphorus nutrition and to reduce phosphorus pollution in animal waste. This review covers the basic nomenclature and crystal structures of phytases and emphasizes both the protein engineering strategies used for the development of new, effective phytases with improved properties and the potential biotechnological applications of phytases.
Phytases (myo-inositol hexakisphosphate phosphohydrolases) catalyse the partial or complete hydrolytic removal of orthophosphates from phytates (myo-inositol hexakisphosphates). Phytate is the principal form in which phosphorus and inositol are stored in cereals, legumes used in commercial animal feeds and oilseeds; phytates constitute c. 60–90% of the total phosphorus content in plants (Reddy et al. 1982). Phytate is hydrolysed by phytase into one molecule of inositol and six molecules of inorganic phosphate (Fig. 1). Unfortunately, phytates are not hydrolysed in the monogastric gut, and the phytate-associated phosphates remain unabsorbed, requiring the exogenous addition of phosphate to avoid phosphorus deficiency. In addition to increasing basic food costs, the phosphates in the excreted phytates are available to microbial soil consortia, which break them down to create phosphate run-off and serious environmental pollution (Greiner and Konietzny 2006). Furthermore, phytates are a strong chelator of cations and bind minerals such as Ca2+, Zn2+ and Fe2+ (Lopez et al. 2002; Vats and Banerjee 2004), making them unavailable for absorption in the intestine of the monogastric animal. In addition, phytates are known to form complexes with proteins under both acidic and alkaline pH conditions. These interactions were found to affect the proteins’ structure, thus decreasing the enzymatic activity, protein solubility and proteolytic digestibility (Kies et al. 2006) (Fig. 1). Supplemental microbial phytases in corn–soybean meal diets for monogastric animals can reduce these problems, improve the animals’ utilization of the phytate phosphorus and reduce their faecal phosphorus excretion by up to 50% (Leytem et al. 2008; Kim et al. 2010).
So far, phytases have been mainly – if not solely – used as a feed supplement in diets for swine and poultry, and to some extent for fish. The inclusion of phytases in animal feed is attracting, increasing more attention internationally. Laboratory experiments and field trials have repeatedly demonstrated that 500–1000 units of phytase can replace c. 1 g of inorganic phosphorus supplementation and reduce total phosphorus excretion by 30–50% (Yi et al. 1996; Kemme et al. 1997). Thus, phytases perform a double duty, conserving expensive and nonrenewable inorganic phosphorus resources by reducing the need for their inclusion in animal feed, while also protecting the environment from pollution resulting from excessive manure phosphorus run-off. For these reasons, phytases are increasingly used worldwide as a phosphate-mobilizing feed supplement in the diets of swine and poultry (Ketaren et al. 1993).
Although phytase was first shown to hydrolyse phytate phosphorus in diets for chicks 40 years ago, its commercial application was not feasible for a long time because of low yield, high costs and low enzymatic activity after pelleting. Challenges in the above three areas have prompted the rapid emergence of the field of phytase science and biotechnology. Phytase optimization by genetic and protein engineering is actively being pursued as no known phytase fulfils all properties for the ideal additive. Here we highlight the current knowledge on phytases with regard to their three-dimensional structure, their enzyme mechanism, improvements in their bioavailability to support the wider application of phytases in animal and crop production.
Depending at which carbon in the myo-inositol ring of phytate dephosphorylation is initiated, phytases can be grouped into 3-phytases (myo-inositolhexakisphosphate-3-phosphohydrolase, EC 184.108.40.206), 6-phytases (myo-inositol hexakisphosphate 6-phosphohydrolase, EC 220.127.116.11) and 5-phytases (myo-inositol hexakisphosphate 5-phosphohydrolase, EC 18.104.22.168). 3-Phytases, in fact, include fungal and bacterial representatives that dephosphorylate phytates at either C1 or C3. (Sajidan et al. 2004). 6-Phytases are frequently found in grains and oil seeds of higher plants, while 5-phytases have been isolated from alfalfa, beans, peas and Selenomonas ruminantium (Chu et al. 2004). Recently, a new protein tyrosine phosphatase (PTP)-like inositol polyphosphatases (IPPases) phyAme was identified that defines a new class of IPPase based on its stereospecific reaction of cleaving InsP6 at positions 1D-3 or 1D4 (Puhl et al. 2009).
On the basis of their pH optima, phytases can be broadly divided into two major classes: acid and alkaline phytases. Acid phytases include those enzymes belonging to the histidine acid phosphatases (HAPs), purple acid phosphatases (PAPs) and PTP-like class of phosphatases (the latter of which was identified more recently). To date, the β-propeller phytases (BPPs) from Bacillus are the only extensively characterized class of alkaline phytase (Kerovuo et al. 1998; Tye et al. 2002). Phytases also exhibit variations in their catalytic mechanism and consequently have been categorized into HAPs, BPPs, PAPs or cysteine phytases (Mullaney and Ullah 2003). Several fungal, bacterial and plant phytases belong to the HAPs class of enzymes (Wyss et al. 1999). All of these phytases share a conserved active site hepta-peptide motif RHGXRXP and the catalytically active dipeptide HD, unique to this class of enzymes (Etten et al. 1991). This group of enzymes catalyses the phytic acid hydrolysis in a two-step process: a nucleophilic attack on the phosphorous atom by the histidine in the active site, followed by hydrolysis of the resulting phospho-histidine intermediate (Vincent et al. 1992). Phytases from Bacillus species constitute an exception: These enzymes share a sequence identity of 90–98% each other but are unrelated to HAPs and other phosphatases. Unlike other HAPs, they require Ca2+ for activity and show a different pH optimum of 7·0–8·0 (Kerovuo et al. 1998; Kim et al. 1998). Meanwhile, a phytase isolated from soybean was found to be unrelated to previously characterized microbial or maize (Zea mays) phytases, classified as HAPs. This soybean phytase is a PAP, characterized by seven conserved residues (bold) in the five conserved motifs –DXG, GDXXY, GNH(D/E), VXXH and GHXH– involved in the coordination of the dimetal nuclear centre (Hegeman and Grabau 2001; Li et al. 2002). In contrast, S. ruminantium phytase neither contains the conserved RHGXRXP motif nor is affected by divalent metal ions. The active site is located near a conserved cysteine-containing (Cys241) P loop (Chu et al. 2004).
Phytase crystal structures
Crystal structure analyses of a number of phytases have revealed a range of distinct folds for these enzymes and have allowed their biophysical properties to be rationalized in terms of their structure. The crystal structure of Aspergillus ficuum phytase at 2·5 Å resolution revealed three distinct domains, including a large α-helical domain and β-sheet domain, and a small α-helical domain. The large α-helical domain and small α-helical domain contain five α-helixes and four α-helixes, respectively, and the β-sheet domain contains eight β-sheets (Kostrewa et al. 1997). Crystal structure analysis of Escherichia coli phytase with a resolution of 2·2 Å also showed two domains. One contains five α-helixes and two β-sheets, and the other includes six α-helixes and nine β-sheets (Lim et al. 2000). The crystal structure of phytases from Debaryomyces castellii (PhytDc) was determined at a resolution of 2·3 Å. The PhytDc structure is very similar to that of A. ficuum phytases and can be divided into two parts: a large α-helical/β-sheet domain with a six-stranded β-sheet, and a small α-helical domain (Ragon et al. 2009). The crystal structures of the phytases mentioned above belong to one type of protein folding, which has an α-helical domain and a conserved α-helical/β-sheet domain, with two helices on each side of the seven-stranded sheet. The active site is located at the interface between the two domains. These structures closely resemble the overall folding in other HAPs. A three-dimensional model of A. ficuum phytase (1IHP) from the National Center for Biotechnology Information’s (NCBI) website is shown in Fig. 2a.
The crystal structure of Bacillus amyloliquefaciens phytase (TsPhy) at 2·1 A resolution revealed a six-bladed β-propeller in which each blade consists of a four- or five-stranded antiparallel β-sheet (Fig. 2b, PDB code 1H6L). The enzyme binds seven Ca2+: two near the periphery, one in the central channel and four near the ‘top’ of the molecule. Unlike other β-propeller structures, it does not show any conserved sequence repeats in the β-sheet. The crystal structures of TsPhy at 2·1 Å resolution in both the partially and the fully Ca2+-loaded states were determined. And the dependence of thermostability of TsPhy on Ca2+ was assessed by differential scanning calorimetry. The binding of two Ca2+ to high-affinity Ca2+-binding sites results in a dramatic increase in thermostability (with an increase of as much as c. 30°C in the melting temperature), because of the joining of loop segments remote in the amino acid sequence. Three Ca2+ bind to the active Ca2+-binding sites and create an ideal conformation and charge distribution for the substrate. Substrate binding to the active site would appear to be followed by occupation of the fourth Ca2+ site to offset the negative charge of the substrate phosphate group already coordinated by lysine and arginine (Fu et al. 2008b).
Selenomonas ruminantium phosphatase (SrPhy) represents a third, dual-specificity phosphatase type with a conserved cysteine (C241) in its so-called P loop. Two distinct crystal packing arrangements have been observed of the complex of SrPhy with the inhibitor myo-inositol hexasulfate. The inhibitor is bound to both ‘standby’ and ‘inhibited’ conformations. In a pocket slightly away from the conserved P loop Cys241 and at the substrate binding site, the phosphate group to be hydrolysed is held close to the -SH group of Cys241. Further, mutagenesis studies verify that the P loop-containing phytase attracts and hydrolyses the substrate (phytate) sequentially via a complicated mechanism (Chu et al. 2004). Figure 2c shows 1U26 to underscore the structural differences in these three classes of enzyme.
Development of effective phytases
Identification of novel phytases
Phytases are produced in a wide range of plant, bacterial, fungal and animal tissues. Ever since the discovery of phytase in rice bran (Suzuki et al. 1907), a variety of phytases with different properties have been identified from organisms (Table 1). Most scientific work has, however, been performed on microbial phytases, particularly those from filamentous fungi such as A. ficuum (Gibson 1987), Aspergillus fumigatus (Pasamontes et al. 1997) or Mucor piriformis (Howson and Davis 1983), Rhizopus oligosporus (Casey and Walsh 2004) and Cladosporium species (Quan et al. 2004). The search for phytases with higher thermostability resulted in the cloning of the phytase gene from A. fumigatus (Pasamontes et al. 1997), the purified enzyme of which retains 90% of its initial activity after being maintained at 100°C for 20 min. By contrast, the phytase from Aspergillus niger remains only 30% active after being heated to 70°C for 20 min (Pasamontes et al. 1997; Wyss et al. 1998). The thermostability of A. fumigatus phytase, A. niger phytase and A. niger acid phosphatase was assessed by protein unfolding and refolding experiments using circular dichroism spectroscopy and protein fluorescence. Only A. fumigatus phytase could refold into a fully active, native conformation even after heat denaturation at 90°C. In feed pelleting experiments performed at 85°C, the recovery of enzyme activity was 51% for A. fumigatus phytase but only 31% for A. niger phytase and 14% for A. niger acid phosphatase (Wyss et al. 1998). In contrast to most other phytate-degrading enzymes, the enzyme from A. fumigatus has a broad pH optimum; at least 80% of the maximal activity was observed at pH values between 4·0 and 7·3 (Wyss et al. 1999).
Table 1. Sources and properties of phytases, published during the last 10 years
Fungal phytases from Peniophora lycii, Agrocybe pediades, Ceriporia sp., and Trametes pubescens showed a high degree of refolding after thermal unfolding, as evidenced by differential scanning calorimetric studies (Lassen et al. 2001). The thermal stability of fungal phytases is often attributed to highly reversible thermal unfolding, rather than an intrinsic thermostability (Wyss et al. 1998). Fungal phytase have also been isolated from thermophilic Rhiomucor pusillus and Thermocyces lanuginosus. The former has a temperature optimum of 70°C, while the latter retains its activity up to 75°C, has a higher catalytic efficiency at 65°C than other fungal phytases and has a comparably broad pH optimum as A. fumigatus phytase (Chadha et al. 2004; Berka et al. 1998). Recently, a novel phytase gene from A. niger N-3 was cloned and expressed in Pichia pastoris. The purified enzyme of which retains 45% of its initial activity after being maintained at 90°C for 5 min. It showed a greater affinity for sodium phytate than for p-nitrophenyl phosphate. Dual optimum pH values were obtained at 2·0 and 5·5. The activity at pH 2·0 was about 30% higher than that at pH 5·5, which is more similar to conditions in the stomachs of monogastric animals (Shi et al. 2009). Two novel thermostable genes were identified in Aspergillus japonicus BCC18313(TR86) and BCC18081(TR170), respectively. The thermostable nature of this phytases gives it valuable potential for applications (Promdonkoy et al. 2009).
Apart from the phytase genes identified in fungi, others have been cloned and identified in other microbes, motivated by their potential for applications. To find a phytase with high activity at low temperature and neutral pH, two phytases have been isolated from Pedobacter nyackensis MJ11 CGMCC 2503 and Erwinia carotovora var. carotovota ACCC 10276. The Pedobacter phytase belongs to the BPP family and shares very low identity (approximately 28·5%) with Bacillus subtilis phytase. Compared with the major commercial phytases and B. subtilis phytase, the purified recombinant enzyme from E. coli displayed higher activity and hydrolysed phytate from soybean meal with better efficiency at neutral pH and 25°C. These characteristics suggest that this phytase has a great potential as an aquatic feed additive in the rapidly developing aquaculture industry. The Erwinia phytase contains a conserved active site hepta-peptide motif RHGXRXP and the catalytically active dipeptide HD that are typical of HAPs and shares a 50% amino acid identity to the Klebsiella pneumoniae phytase (Huang et al. 2009a,b). And except for potential application in aquaculture, the latter is also attractive for food processing by avoiding damage to the food in gradients at low temperatures (Greiner and Konietzny 2006). Moreover, owing to its typical properties as a low-temperature-active enzyme, it could be a good model protein to study the relationship between structure and function. The gene appA, encoding a phytase from Yersinia kristeensenii, was cloned and heterologously expressed in P. pastoris. The data show that the Y. kristeensenii phytase is highly pH stable at pH 1·5–11·0 and thermostable, providing significant advantages for processing, transportation, storage and application. Comparison of r-APPA with other well-known phytases suggested that the Y. kristeensenii phytase would be an attractive enzyme for feed industry use (Fu et al. 2008a). A list of bacterial phytases with considerable variations in biochemical properties is presented in (Table 1).
In addition, phytases from yeast have also been identified and characterized (motivated by their potential as a feed additive for improving the phytate-phosphorus digestibility in monogastric animals), such as the marine yeast Kodamaea ohmeri BG3 (Li et al. 2008, 2009b), Pichia anomala (Kaur et al. 2010; Vohra et al. 2010) and wastewater treatment yeast Hansenula fabianii J640 (Watanabe et al. 2009).
Protein engineering of phytases
Although properties of phytases vary, there is no single wild-type enzyme that is perfect or ideal for field applications. Theoretically, an ‘ideal’ phytase should be catalytically efficient, proteolysis-resistant, thermostable and cheap (Lei and Stahl 2001). In reality, phytases possessing all of these qualities may never be found or generated. To obtain enzymes with modified and desired properties, two different strategies are used: rational protein design and directed (molecular) evolution, which are increasingly applied in a synergistic manner to tailor-design the enzyme for a given process (Chica et al. 2005; Bottcher and Bornscheuer 2010) (see Fig. 3). One recent example of the improvement of phytase activity was shown for the phyI1s from A. niger 113, where two single mutant phyI1s Q53R and K91D were obtained via a semi-rational site-directed mutagenesis strategy. None of the single amino acid residues in the two mutants was in a position reported to be important for catalysis or substrate binding. Kinetic analysis of the phytase activity of the two mutants (Q53R and K91D) showed a respective 2·2- and 1·5-fold increase in specific activity, and a 1·47- and 1·16-fold increased affinity for sodium phytate. In addition, the overall catalytic efficiency (kcat/Km) of the two mutants was improved 4·08- and 2·84-fold compared with that of the wild type (Tian et al. 2010). A semi-rational protein engineering strategy based on 3D structure and sequence alignment was used to take advantage of the desirable characteristics of A. niger NRRL 3135 phytase (Anp) and A. fumigatus ATCC 13073 phytase (Afp); these were chosen because they are quite different but possess many mutually complementary properties (Bei et al. 2009). Another example for engineering of the thermostability of phytases is given by A. niger PhyA phytase; crystal structure comparisons with its close homolog, the thermostable A. fumigatus phytase (Afp), suggested thermostability associations with several key residues (E35, S42, R168 and R248) that formed a hydrogen bond network in the E35-to-S42 region, and ionic interactions between R168 and D161, and R248 and D244. Finally, four mutants showed improved thermostability; the best response came from the quadruple mutant (A58E P65S Q191R T271R), which retained 20% greater (P < 0·05) activity after being heated at 80°C for 10 min and had a 7°C higher melting temperature than that of wild-type PhyA (Zhang et al. 2007).
The E. coli phytase appA gene product was chosen as a candidate for increasing thermal tolerance to promote the survival of the enzyme during pelleting, because of its high specific activity and specificity and its favourable pH profile for gastric activity. To create an optimized phytase gene, site saturation mutagenesis (GSSM) technology was employed to identify mutations that increased enzymatic performance. The technology, which generates a library of all possible single-site mutations in an enzyme, can target individual aspects of a phenotype in association with an appropriate high-throughput screening assay. In this case, the dual goals of increased thermal tolerance and maintenance of high turnover were established. Combining GSSM technology with an assay that challenged all of the GSSM constructs with a heat step revealed mutations that increased thermal tolerance. Combining these mutations led to a phytase Phy9X with an optimal phenotype for economic use as an animal feed supplement (Garrett et al. 2004).
Another E. coli phytase (AppA2) was evolved to a thermostable phytase via epPCR. After a mutant library of AppA2 was generated via error-prone polymerase chain reaction, variants were expressed initially in Saccharomyces cerevisiae for screening and then in P. pastoris for characterization of thermostability. Compared with the wild-type enzyme, two variants (K46E and K65E/K97M/S209G) showed an improvement of over 20% in thermostability (80°C for 10 min) and 6–7°C increases in melting temperatures (Tm) (Kim and Lei 2008). To highlight the rapidly growing number of successful phytase engineering studies using rational protein design and directed (molecular) evolution, previously reported examples are summarized in Table 2.
Table 2. Previously reported examples of phytase engineering
N/A, not applicable.
Aspergillus niger phytase (PhyA)
Mutant (A58E P65S Q191R T271R) showed a higher thermostability
Hydrogen bond network and ionic interactions play an important role
Potential biotechnological applications of phytases
Because of the potential value of phytases for improving the efficiency of phosphorus use, biotechnology has led the rapid development of the field to its current stage. With the development of heterologous gene expression, large amounts enzymes could be produced at relatively low cost.
Microbial expression systems
A transgenic approach has proved to be a powerful tool for expressing specific phytase genes on a large scale in yeast strains, which have been verified to be an elite expression system for heterologous genes. Several studies have already investigated the use of various yeast expression systems as an alternative to the current phytase production method using overexpression in filamentous fungi. It was found that phytases used in ecotopic expression could be derived from a host of micro-organism sources (Xiong et al. 2006; Lei et al. 2007). They created a P. pastoris that expressed the modified phytase gene (phy-pl-sh), with the MF4I sequence producing 12·2 g of phytase per litre of fluid culture and a phytase activity of 10540 U ml−1. Using the phytase gene (phytDc) from D. castellii and an alpha-amylase gene (AMY) from Debaryomyces occidentalis as the target genes, Lim et al. (2008) developed an industrial strain of S. cerevisiae. Phytase has also been expressed in Stretptomyces lividans (Stahl et al. 2003) and Lactobacillus plantarum (Kerovuo et al. 2000b). The latter expression system offers the possibility of combining phytase with beneficial probiotic lactic acid bacteria. Recently, a great research effort has been made towards the use of transgenic rhizobacteria overexpressing citrobacter braakii appA on phytate-P availability to mung bean plants. This is the first report of the overexpression of phytase in rhizobacterial strains and its exploitation for plant growth enhancement (Patel et al. 2010).
Transgenic plants and animals
Transgenic plants might contain sufficient phytase activity that they could replace additional supplementation of feed and food with microbial phytases, thereby reducing the downstream processing and formulation costs involved in the commercial production of phytases. Alternatively, transgenic plants could be used as bioreactors for the production of phytase as a supplement. Meanwhile, transgenic animals producing endogenous phytase along with other digestive enzymes are known to increase the bioavailability of plant phytate, which in turn leads to reduced phosphorus output in manure.
Phytases have been expressed in several dicotyledonous plants like tobacco (Pen et al. 1993), Arabidopsis thaliana (Richardson et al. 2001; Xiao et al. 2005), alfalfa (Ullah et al. 2002), canola (Ponstein et al. 2002), soybean (Chiera et al. 2004) and so on. Aspergillus phytase expressed in maize seeds exhibited an increase in iron bioavailability, as evidenced by in vitro digestion and Caco-2 cell model studies (Drakakaki et al. 2005). Aspergillus fumigatus phytase expressed in tobacco exhibited high thermostability and retained 28·7% of the initial activity, even after incubation at 90°C for 15 min (Wang et al. 2007). The expression of B. subtilis phytase in tobacco indicated that the Bacillus phytase transgene could only improve the phytate-phosphorus uptake by transgenic plants under sterilized conditions, and its effectiveness might be limited under natural conditions because of microbial decomposition and mineral fixation. The microbial community in the rhizosphere appears to be resistant to the impact of single-gene changes in plants designed to alter rhizosphere biochemistry and nutrient cycling (Kong et al. 2005; George et al. 2009). Overexpression of heterologous phytases in transgenic potato not only offers an ideal feed additive for improving phytate-P digestibility in monogastric animals, but also improves tuber yield, enhances P acquisition from organic fertilizers and has the potential for phytoremediation (Hong et al. 2008). Transgenic soybean plants expressing Arabidopsis (Arabidopsis thaliana) PAPs (AtPAP15) exhibited enhanced bioavailability of phosphorus (Wang et al. 2009). In plants, the phosphate-absorption sites are in the root system, especially the root hairs. It is therefore better for the phytases ectopically expressed in transgenic plants to be secreted in the rhizosphere where the phytate and its derivatives are degraded by the biochemical reactions involved in the encoded phytases. Therefore, in transgenic soybean plants in which the A. ficuum phytase (AfPhyA) was integrated, a promoter from the Arabidopsis Pky10 gene and the carrot extensin signal peptide were used to drive the root-specific and secretory expression of the AfPhyA gene. The phytase activity and inorganic phosphate levels in transgenic soybean root secretions were 4·7 U mg−1 protein and 439 μmol l−1, respectively, compared with 0·8 U mg−1 protein and 120 μmol l−1 in control soybeans, suggesting that the transgenic techniques could be of great value for the generation of crop varieties with high P-use efficiency in the future (Li et al. 2009a). Expression levels of heterologous phytases in diverse transgenic plants are presented in Table 3.
Table 3. Expression of phytase in transgenic plants
Transgenic mice and pigs have been generated by overexpressing phytase in their salivary glands (Golovan et al. 2001a,b). A transgenic mouse model has been developed with an appA phytase gene from E. coli, driven by the upstream promoter of a pig parotid secretory protein gene. Expression of salivary phytase reduced faecal phytates by 8·5 and 12·5% in two transgenic mouse lines (Yin et al. 2006). The transgenic Enviropig, expressing E. coli appA phytase, could secrete active phytase into its saliva and showed a substantial reduction (60%) in the excretion of phosphorus compared with nontransgenic animals (Forsberg et al. 2003).
Conclusions and future perspectives
Along with the motivation provided by increasing concerns relating to phosphorus pollution in the areas of intensive livestock, developments in the phytase field are driven by their considerable potential in commercial and environmental applications. However, a limited number of phytases have been reported and studied, and our scientific knowledge of phytases has yet to yield a solution to meet the nutritional and environmental requirements that a real-world solution demands. Further research into identifying new phytases, engineering better phytases and developing more cost-effective expression systems should be continued. With the collaborative efforts of scientists from different fields, effective solutions for the biotechnological development of phytases for mineral nutrition and environmental protection should be available in the near future.
This project was supported by grants from the ‘National Natural Science Foundation of China’ (no. 31071924), the ‘Natural Science Foundation of Shanxi Province’ (2010011040-1) and the ‘Shanxi Scholarship Council of China’.