Diversity and distribution of bacterial phytases

Phytases: The stepwise dephosphorylation of phytate is catalyzed by phytases, a group of enzymes found mainly among bacteria and plants. Bacterial potential: We identified an unexpectedly widespread and previously underappreciated bacterial potential for phytate dephosphorylation, suggesting that microbial phytases play crucial roles in the environmental and host-associated phosphorus flux. Diversity: We analyzed 115,379 validated bacterial genomes to resolve the taxonomic distribution and diversity of potentially phytase-producing bacterial species, and identified phytase genes in 6281 out of 25,105 species (25%). Prevalence: Phytases were predominantly found in members of the phyla Pseudomonadota, Bacteroidota, and Cyanobacteriota, but were comparatively rare in Bacillota and Actinomycetota, with alkaline β-propeller phytases as the most prevalent and widespread category. Phytate, or inositol hexakisphosphate [InsP6], is the primary storage form of phosphorus and minerals in plants. It is particularly present in plant tissues such as bran and seeds of legumes and cereals, oil seeds, and nuts. Phytate and other inositol phosphates account for up to 65–80% of the total phosphorus in seed and for up to 80% of total phosphorus in the manure of monogastric animals [1]. Phytate-hydrolyzing phosphatases, termed phytases, release phosphorus during food processing and in the gastrointestinal tract, thereby increasing the bioavailability of phosphorus and helping to mitigate environmental eutrophication [2]. Additionally, phytases catalyze phosphate release in soil, where phytate represents the most recalcitrant phosphorus-containing compound [3]. The stepwise hydrolysis of phytate produces myo-Ins phosphate esters (IP1 − IP5), ultimately yielding free myo-Ins [4]. In the stomach and small intestine of humans, phytases are estimated to hydrolyze 37–66% of all dietary phytate when a diet is rich in plant-based foods containing active phytases [5]. This contrasts with Western diets that are typically poor in plant food and often lack active phytases due to heating during food processing. Phytases are mainly of plant (endogenous) or microbial (exogenous) origin. Because plant-derived phytases do not fully hydrolyze phytate, microbial phytases are frequently added to animal feed, particularly for monogastric species, to mitigate phytate-bound phosphate limitations in animal production [6]. Alternatively, inoculation with phytase+ strains into gastrointestinal tracts of animals has been proposed, though concerns have been raised regarding colonization stability and potential environmental contamination through fecal shedding by those probiotics [6]. Food supplementation can decrease fecal phosphorus to below 1% [7], a significant reduction given that a single pig excretes 1.23 kg of phosphorus over its lifetime in the United States [6]. Dephosphorylation of phytate not only enhances phosphorus bioavailability in livestock [8], but also that of essential dietary minerals [9], as phosphate groups increase the dietary zinc and iron binding capacity of phytate. On the other hand, such an intervention in human nutrition remains controversial because phytate acts as an antioxidant of food components by chelating free iron, thus preventing iron-driven hydroxyl radical formation [10]. Phytases are commonly classified by their optimal pH, their three-dimensional structures, and their catalytic mechanisms into categories such as histidine acid phytases (HAPhys), alkaline β-propeller phytases (BPPhys), and protein tyrosine phytases (PTPhys) [6]. All three groups are produced by bacteria [6], whereas purple acid phytases (PAPhys) are plant-specific. The multiple inositol polyphosphate phosphatase (MINPP), a subclass of the HAPhy family, has been found in plants and animals [11, 12]. Cysteine phytases, mainly produced by ruminal bacteria, belong to the PTPhy category [13]. Phytate-degrading microorganisms have been isolated from diverse environments, including marine ecosystems, soils, and the gastrointestinal tract of animals [14, 15]. Despite the critical roles of phytate and phytases in human and animal health and in the global phosphorus cycle, the distribution of bacterial phytases is largely unknown. Since a pilot bioinformatic study from 2007 was based on just 80 bacterial genomes [16], a comprehensive genome-based bioinformatic approach was performed to identify phytase gene-positive bacteria across all phyla. To investigate the presence and distribution of phytase genes in bacteria, we performed a comprehensive data mining approach and compiled a set of 34 non-redundant amino acid sequences of bacterial phytases with experimentally verified phytate-dephosphorylating activity in the literature (Table S1). These phytases originate from 28 distinct bacterial species, as well as four uncultured and uncharacterized microorganisms [13, 17]. To identify phytase orthologs, we queried the 34 protein sequences against the NCBI non-redundant protein database. In total, we screened 115,379 bacterial genome assemblies representing 25,105 species, for homologs exhibiting ≥20% sequence identity (e-value ≤ 10−06) and 50% query coverage to at least one of the 34 phytases. A species was classified as phytase-positive (phytase+) if at least one genome within that species encoded a phytase homolog. Overall, we identified putative phytases in 6281 bacterial species, corresponding to approximately 25% of all species examined (Table S2). To explore the clustering of phytase types, we calculated the Jaccard coefficient for all pairwise BLAST hit combinations among the 34 reference phytases. This analysis identified five clusters, while three enzymes appeared unique within the reference set. The largest cluster included 15 HAPhys represented by the bifunctional 4-phytase AppA from E. coli. A taxonomic analysis showed that HAPhy homologs were present in approximately 7.3% of all species tested here. EEG70762 and ACJ51391 (Table S1) did not exhibit significant similarities to AppA-like phytases, but were identified as histidine-phosphatases (HPs) based on BLAST analysis. Homologs of these two HPs were detected in 3.5% of all 25,105 species. Overlap between the two HAPhy groups was rarely found, as of the 1824 and 884, respectively, species harboring the two groups, only 10 species, including 6 Gilliamella strains, a honey bee gut symbiont, encoded both types (Table S2). This limited co-occurrence suggests that AppA-like phytase and the two HPs may serve similar functions. Given their potential relevance as a novel subgroup of phytase-like enzymes, we concluded HPs in subsequent analyses. We also identified bacterial homologs of MINPPs classified as a subclass of HAPhys [12]. For all but two of the 34 phytases, a signal peptide sequence was predicted with a probability >0.75, indicating that these enzymes are likely secreted (Table S1). The second largest cluster consisted of BPPhys, a group of alkaline phosphatases represented by PhyC from Bacillus amyloliquefaciens. BPPhy homologs were present in 14.7% of all species screened. Four PTPhys, homologous to the Selenomonas ruminantium alkaline phosphatase, formed another distinct cluster, with homologs identified in 1.2% (308) of all species. Five of the identified phytases (AKS25181, AAC05186, ADX99261, ADY15023, and ADY15024) exhibited homology to proteins with unspecific or poorly characterized annotations. AKS25181 is related to a protein tyrosine phosphatase-like phytase but is distinct from the PTPhy group. AAC05186 and ADX99261 could not be assigned to any of the three major bacterial phytase categories, and ADX99261 was identified in 95.4% of all species analyzed. ADY15023 and ADY15024 showed no significant similarity to known phosphatase families and are members of the widespread major facilitator superfamily [18]. To prevent misattribution, these five proteins were excluded from subsequent analyses. The majority of species (5866) encode (one or more) phytases from a single cluster. In contrast, the genomes of 400 species code for phytases belonging to two distinct clusters, and only 15 species from the genera Xanthomonas and Pseudomonas carry phytases from three different categories (HAPhy, BPPhy, and PTPhy) (Figure 1A). The largest group among species with phytase from two clusters comprises 254 species that carry both HAPhy- and BPPhy-like phytases. Of these, 196 belong to ubiquitous genera such as Sphingomonas (64 species), Pseudomonas (50), Stenotrophomonas (30), Xanthomonas (27), and Caulobacter (25), which are commonly associated with soil, plant, or aquatic environments (Table S2). Additionally, 15 species encode both a HAPhy and a HP phytase, while 34 species from the soil-dwelling genus Janthinobacterium carry genes encoding BPPhys as well as HP phytases. We also analyzed the number of species that harbor more than one phytase gene encoding enzymes from the same cluster (Figure 1B). Among the 1824 species with at least one HAPhy gene, 558 (30.6%) contain two or three homologs, which primarily represent AgpE-like glucose-1-phosphatases and AppA-like phytases. Similarly, two or three PTPhy genes were found in 10.7% of all species containing PTPhys. The corresponding percentages for BPPhys and HPs were 2.5% and 1.7%. Table S3 shows that a notable subset of bacteria, particularly those from ecologically diverse and plant-associated and soil-dwelling genera, possess multiple and phylogenetically distinct phytase types. To explore the taxonomic prevalence of phytase genes, we analyzed all 6281 bacterial species identified as phytase+ (Table S2, Figure 2A). Among the major phyla Pseudomonadota, Actinomycetota, Bacillota, Bacteroidota, and Cyanobacteriota, and Mycococcota, BPPhys were the most prevalent (Figure 2B). With a focus on families comprising over 100 species, BPPhy-positive genomes were highly abundant in Xanthomonadaceae (86% phytase+ genera), Caulobacteraceae (82%), Sphingobacteriaceae (64%), Erythrobacteraceae (62%), Sphingomonadaceae (61%), Pseudoalteromonadaceae (56%), and Weeksellaceae (53%). AppA-like HAPhys were predominantly present in Erwiniaceae (98% phytase+ genera), Enterobacteriaceae (98%), and Yersiniaceae (96%), all belonging to Pseudomonadota, but largely absent in most other phyla. HP homologs are less widely distributed but were notably enriched in Moraxellaceae (44% phytase+ genera), Oxalobacteraceae (32%), Burkholderiaceae (23%), and Bifidobacteriaceae (16%). The less common PTPhys were primarily associated with the phylum Bacillota, especially in Clostridiaceae (30%). To illustrate the distribution of phytases at the genus level, separate cladograms were generated for HAPhys and BPPhys (Figure 2C,D). The HAPhy cladogram encompasses six phyla, 27 families, and 84 genera, while the BPPhy cladogram depicts five phyla, 26 families, and likewise 84 genera. Details are given in Table S4. To summarize, phytase genes are broadly distributed throughout bacterial taxa, with BPPhys emerging as the most widespread in diverse phyla and families, while HAPhys, HPs, and PTPhys exhibit a more taxon-specific enrichment, thus stressing distinct evolutionary patterns and ecological adaptations of bacteria carrying phytase genes. The large-scale analysis performed here provides insights into the taxonomic breadth and diversity of phytase-gene harboring bacteria in all taxa. The extensive catalog of phytase+ bacterial species possibly producing phytases serves as a valuable resource for identifying novel phytases, potentially tailored for species and diets, and phytases with novel functions in metabolism and health [6]. Further functional assessment, including heterologous gene expression and metatranscriptomics, however, is needed, extending beyond enzymatic properties to encompass ecological roles and interspecies interactions [14]. Moreover, experimental studies using defined microbial consortia will be essential for uncovering yet unknown synergistic phytase activities within environmental and gut habitats. For example, multiple myo-Ins cross-feeding was recently observed between a piglet-derived bacterial consortium of seven phytase+ strains to Brevibacillus agri and S. enterica serovar Typhimurium [19]. Taken together, our findings suggest that an unexpectedly large portion of bacteria may contribute to the metabolization of phytate both in the gut and in biogeochemical systems. Thilo M. Fuchs: Conceptualization; supervision; writing—original drafts. Michael Weber: Bioinformatic analysis; visualization. Lena-Sophie Paul: Data curation; formal analysis. All authors have read the final manuscript and approved it for publication. We apologize for not being able to cite additional work owing to space limitations. The graphical abstract was created by Marcus Pfau, Friedrich-Loeffler-Institut, and the microbes by Anne Bleicher. Photos of nuts were created by Adobe Firefly, and the 3-D representation of phytases by I-Tasser. We apologize for not being able to cite additional work owing to space limitations. The authors declare no conflicts of interest. No animals or humans were involved in this study. The data that support the findings of this study are available in the supplementary material of this article. The datasets supporting the conclusions of this article are included within the article and its additional files. The data used are saved in GitHub https://github.com/mweberr/phytases-data. Supplementary materials (methods, tables, graphical abstract, slides, videos, Chinese translated version, and updated materials) may be found in the online DOI or iMetaOmics http://www.imeta.science/imetaomics/. Table S1: List of phytases. Table S2: Taxon list of 6,281 species. Table S3: More than 1 phytase per category. Table S4: Phytase gene prevalence across all taxa. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.