Unexpected diversity of dye-decolorizing peroxidases

Abstract

Dye-decolorizing peroxidase (DyP)-type peroxidases are a family of heme-containing peroxidases. Because DyP-type peroxidases can degrade recalcitrant anthraquinone dyes and lignin, their potential applications in the treatment of wastewater containing dyes and lignin degradation are expected. Although many DyP-type peroxidases have been characterized experimentally, most of the reported DyP-type peroxidases are from basidiomycetous fungi and bacteria. Therefore, the taxonomic distribution of the DyP-type peroxidases remains unclear. In this study, we analyzed the phylogenetic tree using all DyP-type peroxidase sequences available in the InterPro database. The findings mainly divided this family into three classes. Metazoa and Archaea also have the genes coding for DyP-type peroxidases, and the sequences belonging to two subclasses have the pyruvate formate lyase or cytochrome P450 domain in addition to the DyP domain. This study reveals differences in the conservation of important residues among classes. The findings will accelerate research on the DyP-type peroxidase family.

Highlights

• •DyP-type peroxidase family is mainly divided into three classes P, I, and V.
• •Obvious difference of three classes is the amino acid sequence lengths.
• •Metazoa and Archaea also have the genes coding for DyP-type peroxidases.
• •Subclasses with domains other than DyP also exist.

1. Introduction

Heme-containing peroxidases were among the first enzymes to be discovered. They reduce hydrogen peroxide to water and oxidize various substrates. The catalytic cycle consists of three steps [1]. The cycle starts from the reaction of the ferric resting state of the enzyme with hydrogen peroxide, producing water and a two-electron oxidized state of the enzyme known as compound I. Compound I is subsequently reduced to compound II by the one-electron oxidation of one substrate. Finally, compound II is reduced to its resting state by one-electron oxidation of the second substrate. The compound I and II intermediates oxidize a wide variety of organic and inorganic substrates. These include phenolic compounds, such as guaiacol, and persistent compounds, such as azo dyes, due to their high redox potential.

Dye-decolorizing peroxidase (DyP)-type peroxidases are members of a family of heme-containing peroxidases. DyP-type peroxidase was first isolated in 1999 from the basidiomycetous fungus Bjerkandera adusta Dec 1 strain as the enzyme capable of decolorizing the synthetic dyes, including azo and anthraquinone dyes. This activity prompted the DyP designation [2,3]. Since the DyP-type peroxidase family was proposed in 2007 [4], more than 50 types of DyP-type peroxidases have been characterized [5,6]. Although most are derived from basidiomycetous fungi and bacteria, DyPA from Dictyostelium discoideum (Eukarya) and XgrDyP from the ascomycetous fungus Xylaria grammica have also been reported recently [7,8].

DyP-type peroxidases are classified into four classes (A, B, C, and D) in RedoxiBase [9]. We have reclassified the DyP-type peroxidases into three classes, P (primitive), I (intermediate), and V (advanced), based on structure-based sequence alignment [5,10]. Here, classes P, I, and V correspond to B, A, and combined C and D, respectively. DyP-type peroxidases from basidiomycetous fungi belong to class V and DyP-type peroxidases from bacteria belong to classes P and I. The amino acid sequence identity between the classes is low, with values of 19% between DyPB from Rhodococcus jostii in class P [11] and DyP from B. adusta in class V, 24% between EfeB from Escherichia coli in class I [12] and DyP in class V, and 25% between DyPB in class P and EfeB in class I. However, the tertiary structures are very similar, with two ferredoxin-like folds and heme buried deep in their molecules [13]. On the heme proximal side, heme iron is coordinated by a conserved histidine residue. On the heme distal side, the four conserved residues (arginine, aspartate, and two hydrophobic residues) form the binding pocket of hydrogen peroxide [14]. Both the arginine and aspartate residues in the GXXDG motif are important for compound I formation [4,15,16]. The arginine residue is more important than the aspartate residue in DyPB from R. jostii [17].

The physiological roles of some DyP-type peroxidases have been previously elucidated. DtpA from Streptomyces lividans is the last gene of an operon involved in the copper (Cu) trafficking pathway, which consists of Cu chaperones Sco, Cu transporter ECuC, and DtpA. DtpA has a twin-arginine translocation (tat) signal sequence at the N-terminus and is thought to function extracellularly. DtpA is essential for the formation of aerial hyphae because the formation of aerial hyphae depends on the radical copper oxidase GlxA and its maturation depends on DtpA [18]. During the maturation step, DtpA is thought to oxidize Cu (I) to Cu (II) binding to GlxA. EfeB from Bacillus subtilis is the last gene of an operon involved in the iron (Fe)-transport system, which consists of Fe transporter EfeU, Fe-binding protein EfeO, and EfeB. EfeB has a tat signal sequence and likely functions extracellularly. EfeB oxidizes Fe (II) to Fe (III) for uptake by EfeU to promote growth under microaerobic conditions where Fe (II) is more abundant [19]. Moreover, EfeB removes reactive oxygen species that accumulate under such conditions to promote growth. DyP-type peroxidase from the human pathogen Mycobacterium tuberculosis forms an operon with an encapsulating protein. Encapsulin forms a virus-capsid-like nanocompartment and encloses DyP-type peroxidase. This nanocompartment system can withstand oxidative stress in low-pH environments, such as macrophage lysosomes, and thus is important for pathogenicity [20]. DyP from B. adusta is secreted extracellularly. Its expression is stimulated in the presence of the natural anthraquinone compound alizalin, which is an antifungal compound produced by the plant. Because alizalin is a good substrate for DyP, it was proposed that DyP degrades alizalin to help parasitism in plants [21]. Some DyP-type peroxidases from white-rot fungi are extracellularly secreted. EglDyP from the wood-degrading fungus Exidia glandulosa, MepDyP from the litter-degrading fungus Mycena epipterygia, and MscDyP from the litter-degrading fungus Mycetinis scorodonius degrade lignin model compounds. These findings suggest that they are part of a biocatalytic system degrading lignin [22]. DyP-type peroxidase (VNG0798H) from Halobacterium salinarum (Archaea), which has a tat signal sequence, is upregulated in the presence of hydrogen peroxide (H₂O₂) and functions as an additional protection against H₂O₂ when its production exceeds the detoxification capability of another peroxidase, PerA [23].

Although we have proposed reclassification of DyP-type peroxidases into classes P, I, and V based on structure-based sequence alignment, the analysis was based on only 12 structures available at that time [10]. This motivated us to verify our analysis using more data. Moreover, the taxonomic distribution of DyP-type peroxidases is unclear because the reported DyP-type peroxidases are limited to basidiomycetous fungi and bacteria, with the exception of DyPA from Eukarya [7] and XgrDyP from Ascomycota [8]. This further motivated us to comprehensively investigate the origin of DyP-type peroxidases.

In the present study, we performed a phylogenetic analysis of DyP-type peroxidases using all the sequences included in the InterPro entry, DyP-type peroxidase family (IPR006314), from the InterPro database. The analysis clearly shows that Metazoa and Archaea also contain genes coding for DyP-type peroxidases.

2. Methods

2.1. Sequence data

All amino acid sequences included in an InterPro entry, DyP-type peroxidase family (IPR006314), of the InterPro database were retrieved (25,475 sequences as of 17 February 2022). The tsv file containing the accession ID, amino acid sequence length, and region matching the entry of each protein was also retrieved. In addition, we used the data of the domain architectures found in the entry.

To perform phylogenetic analysis of DyP-type peroxidases, the regions matching DyP-type peroxidase family of the InterPro entry were extracted from each sequence. The extracted 25,475 sequences were processed using the CD-HIT clustering program to reduce the number of sequences by removing similar sequences [24]. In this program, the sequence identity threshold was set to 0.7, and a sequence was clustered to a cluster that met >70% sequence identity. This is because a higher threshold generates more clusters, which is computationally demanding and makes phylogenetic analysis using the maximum likelihood method impossible. The alignment coverage for the longer sequence was set to 0.9, and the alignment must cover 90% of the sequence to avoid clustering sequences whose lengths are completely different. Moreover, the length of the throwaway sequences was set to 270 for the following reasons. ElDyP from Enterobacter lignolyticus [16] and KpDyP from Klebsiella pneumoniae [25], which are the smallest DyP-type peroxidases whose structures were solved experimentally, have 299 amino acids, and the structures consist of a few loops and two ferredoxin-like folds conserved in DyP-type peroxidases. Moreover, the smallest DyP-type peroxidase characterized experimentally, PpDyP from Pseudomonas putida, has 287 amino acids [26]. The findings suggests that the sequence with a length much shorter than 288 does not function as a DyP-type peroxidase or is not able to fold correctly. Therefore, we excluded sequences with 270 amino acids or fewer. The CD-HIT program excluded 1259 sequences with 270 amino acids or less and finally generated 3027 clusters. The excluded 1259 sequences consisted of 391 sequences with 1–99 amino acids, 518 sequences with 100–199 amino acids, and 350 sequences with 200–270 amino acids. Most of the sequences with 1–99 amino acids were deposited as fragments rather than full length.

2.2. Phylogenetic analysis

Because the CD-HIT program produces one representative sequence from each cluster, 3027 sequences were used for subsequent analysis. Multiple sequence alignment of the 3027 sequences was performed using MAFFT version 7.503 with the L-INS-i strategy [27]. The alignment was trimmed using ClipKIT with the default setting of the gappy mode, which trimmed 92.3% of the sites (5792 sites/6317 sites) [28]. The histidine residue on the heme proximal side and the four residues (arginine, aspartate, and two hydrophobic residues) on the heme distal side are conserved in all DyP-type peroxidases characterized experimentally. Therefore, 108 sequences with gaps in the sites of these amino acid residues were manually excluded. A phylogenetic tree was reconstructed using the phylogenetic tool for maximum likelihood analysis, IQ-TREE ver 2.2.0, with model selection via the -m MFP option, 1000 ultrafast bootstrap replicates via the -bb option, and 1000 replicates to perform the SH-like approximate likelihood ratio test via the -alrt option [29]. The phylogenetic tree was visualized using iTOL version 6 [30].

3. Results and discussion

3.1. DyP-type peroxidases are mainly divided into three classes

We performed a comprehensive phylogenetic analysis using all amino acid sequences included in an InterPro entry, DyP-type peroxidase family (IPR006314) (25,475 sequences as of 17 February 2022). After extraction of only the DyP domain from each sequence, similar sequences were reduced by the clustering program. Sequences that were too short were removed. Finally, 3027 sequences were used for the phylogenetic analysis by the maximum likelihood method. The sequences included 1,007, 1,286, 150, 165, 51, and 44 from Proteobacteria, Terrabacteria, Ascomycota, Basidiomycota, Metazoa, and Archaea, respectively. The obtained phylogenetic tree showed that the DyP-type peroxidase family is mainly divided into three classes: P, I, and V (Fig. 1), which supports our previous results. We subdivided each class into four subclasses (1–4) to make the following results and discussion easier.

Fig. 1.

Phylogenetic tree. Each branch is color coded by organism of origin. Numbers in parentheses denote the branch numbers of each organism. SH-aLRT (>80%) and Ultrafast bootstrap support (>95%) values are indicated by circles on nodes. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

To verify the validity of combining two classes C and D into one class (V), we checked the distribution of the sequences of classes C and D deposited in RedoxiBase (24 and 138 sequences as of 08 October 2022, respectively) in this tree (Fig. S1). All the sequences of class D were located in a clade included in subclass V2, which consists of the branches of Basidiomycota and Ascomycota origin. The sequences of class C were located in both subclasses V2 and V3. Therefore, it seems reasonable to combine classes C and D into class V.

The obvious difference among the three classes is the amino acid sequence length. Class P is the shortest and class V is the longest (Fig. 2a and d). Classes P, I, and V contained approximately 300, 300–400, and 500 residues, respectively. Although DyP-type peroxidases that have been reported are distributed in most clades of the phylogenetic tree, subclasses P1, P2, I2, and V4 are unexplored. Subclass I1 has also been unexplored, except for DyP-type peroxidase (VNG0798H) from H. salinarum [23]. These subclasses exhibit the following characteristics. Subclass I1 is of Archaea origin. P1 is mainly of Metazoa origin. Subclasses P2 and V4 have pyruvate formate lyase and cytochrome P450 domains, in addition to the DyP domain, respectively. Subclass I2 is adjacent to subclass I1, and the species of origin varies.

Fig. 2.

Reformatted view of the tree in Fig. 1. (a) Each branch is color coded by organism of origin as in Fig. 1. Branch lengths are ignored. Circle I shows the existence of the domains except DyP domain. Representative domain structures are shown in (b). Circle II shows the classification as shown in (c). Circle III shows the amino acid sequence lengths of each sequence as concentrical bar graph as shown in (d). Circle IV is the known DyP-type peroxidase mapping. Each DyP name and organism of origin are listed in (e). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.2. Unexplored subclasses

The phylogenetic tree revealed five unexplored subclasses. Subclass P1 consists of sequences from Amoebozoa, Excavata, Sar, and Metazoa (Table 1). These have not been reported, except for DyPA from D. discoideum [7]. DyPA does not belong to subclass P1, but rather to P4, in which the sequences from Eukarya, except fungi and Metazoa, form a small clade including DyPA (Fig. 2a and e). The sequences from Metazoa of subclass P1 consisted of many sequences from Protostomia and a few sequences from Deuterostomia. The sequences from Deuterostomia consisted of sequences from the phyla Hemichordata and Chordata origins. The sequences from Chordata included only sequences from the subphylum Cephalochordata, and sequences from the subphyla Tunicata and Vertebrata were not found. These results suggest that DyP-type peroxidases of subclass P1 are distributed in eukaryotes, but not in fungi, Tunicata, and Vertebrata. Subclass I1 consists of sequences from the archaea class Halobacteria. The sequences of subclass P2 contain a pyruvate formate lyase (PFL) domain in addition to the DyP domain (Fig. 2a and b). PFL converts pyruvate and coenzyme A to acetyl-CoA and formate. PFL is one of the central enzymes in anaerobic metabolism in E. coli and other facultative anaerobes [31]. PFL is activated by a PFL-activating enzyme (PFL-AE) with an Fe–S cluster and S-adenosylmethionine, forming the glycyl radical [32]. As activated PFL and the Fe–S cluster of PFL-AE are sensitive to oxygen, the DyP domain in subclass P2 may protect the PFL domain and PFL-AE from H₂O₂. The sequences of subclass V4 have an N-terminal cytochrome P450 domain followed by some transmembrane regions in addition to the C-terminal DyP domain (Fig. 2a and b). It is unclear whether the P450 and DyP domains are located on the same or opposite sides of membrane because of the varying numbers of transmembrane regions. As both DyP and P450 can react with H₂O₂ [33], both domains may scavenge H₂O₂. Alternatively, a compound may be produced by reactions of both DyP and P450 domains, such as the precocious sexual inducer (psi) factor-producing oxygenases consisting of a fatty acid heme dioxygenase/peroxidase and P450 domains. The psi factor is produced by the oxidation reaction of the peroxidase domain and the isomerase reaction of the P450 domain [34].

Table 1.

Organisms of origins of branches consisting of subclass P1

-	kingdom	–	phylum	branch numbers
Amoebozoa	–	–	–	1
Excavata	–	–	–	3
Sar	–	–	–	3
Opisthokonta	Metazoa	Deuterostomia	Chordata	2
Opisthokonta	Metazoa	Deuterostomia	Hemichordata	1
Opisthokonta	Metazoa	Protostomia	Nematoda	3
Opisthokonta	Metazoa	Protostomia	Rotifera	5
Opisthokonta	Metazoa	Protostomia	Annelida	2
Opisthokonta	Metazoa	Protostomia	Brachiopoda	1
Opisthokonta	Metazoa	Protostomia	Mollusca	6
Opisthokonta	Metazoa	Protostomia	Platyhelminthes	16
Opisthokonta	Metazoa	–	Placozoa	1
Opisthokonta	Metazoa	–	Cnidaria	3

Next, we investigated the structures of unexplored subclasses. The predicted structures of all the unexplored subclasses also showed similar structures, with two ferredoxin-like folds conserved in the DyP-type peroxidases (Fig. 3a). We also investigated the conservation of important residues in unexplored subclasses. On the heme distal side, a H₂O₂ binding pocket is formed by the four well-conserved residues (shown as d1–4 in the DyP structure of Fig. 3a), which consist of catalytic aspartate and arginine residues and two hydrophobic residues [14]. In the heme proximal side, the histidine residue is the heme ligand, the acidic residue forms a hydrogen bond with the histidine ligand, and the arginine residue forms a hydrogen bond with the propionic acid of heme. These residues are also well-conserved (underlined p1–3 in the DyP structure of Fig. 3a). This arginine residue seems to be essential for the correct folding and/or incorporation of heme because the arginine mutant of DyP never displayed peroxidase activity and the correct Soret band (data not shown). While the distal arginine residue is conserved in all subclasses, the distal aspartate residue of subclass I1 is substituted with alanine or serine residues in the amino acid sequence alignment (Fig. 3b). To eliminate the possibility that the sequence alignment was incorrect, we checked the predicted structures. In DyP-type peroxidase (VNG0798H, UniProt ID: Q9HR97) from H. salinarum belonging to subclass I1, the distal aspartate residue is substituted with a serine residue in the sequence alignment. In the predicted structure, the serine residue formed an H₂O₂ binding pocket with the other three residues, suggesting that the amino acid sequence alignment is correct. This DyP-type peroxidase (VNG0798H) functions in vivo may be a true peroxidase, although in vitro characterization has not been performed [23]. Therefore, it appears that the distal arginine residue is essential for the formation of compound I in subclass I1. The aspartate residue is also occasionally substituted with glutamate residue in subclasses I3 and P2. Glutamate is also able to function as the catalytic residue for the compound I formation, although aspartate is better than glutamate [35]. In subclass I2, the proximal aspartate residue was not well-conserved. To eliminate the possibility that the sequence alignment was incorrect, we checked the predicted structures. In the subclass I2 DyP-type peroxidase (UniProt ID: K9TRM8), the aspartate residue is substituted with an asparagine residue. In the predicted structure, the asparagine residue is separated from the proximal histidine residue. Therefore, it is unclear whether the asparagine residue forms a hydrogen bond with histidine. Other important residues were also conserved in all subclasses.

Fig. 3.

Comparison of overall structures, heme active site residues, and active radical sites among subclasses. (a) Known structure and predicted structures of unexplored subclasses. Top: DyP name or UniProt IDs and the organisms of origins; Middle: overall structures. Heme active site residues and candidates of active radical sites are shown as stick and sphere models, respectively; Bottom: close-up views of heme active sites. DyP from B. adusta are shown in black box as representative known structure (PDB ID: 3afv). Residues “d1–4″ and underlined “p1–3″ denote the important residues of heme distal and proximal sides, respectively. In subclasses I1, I2, P1, and P2, the predicted structures by AlphaFold were downloaded from UniProt database. In subclass V4, the structure was predicted by ColabFold coupled with Google Colaboratory [39]. (b) Sequence logo representations of the occurrences of amino acid residues of heme active site. Active site residues are highlighted in grey and the adjacent residues are shown. Vertical axes show bits from 0 to 5. The sequence logos were generated using the WebLogo 3 program [40]. (c) Sequence logo representations of the occurrences of amino acid residues of candidates of active radical sites.

In addition, we investigated the radical sites of these unexplored subclasses (Fig. 3c). Large substrates, such as anthraquinone dyes, are thought to be oxidized via surface-exposed radical sites. Tryptophan and tyrosine residues are the most common candidates for radical sites, while phenylalanine residues are not. Five radical sites (1–5) have been reported in DyP-type peroxidases (Fig. 3a). In VcDyP from Vibrio cholerae, which belongs to the subclass P4, four active radical sites (sites 1–4) have been identified [36]. Sites 2 (Y129) and 4 (Y235) are essential, and sites 1 (W64) and 3 (W183) are important. In TcDyP from Thermomonospora curvata, which belongs to subclass I4, one active radical site 3 (W263) has been identified [37]. In AauDyP1 from Auricularia auricula-judae, which belongs to the subclass V2, one active radical site 5 (W377) has been identified [38]. Tyrosine residues at sites 2 and 4 and tryptophan residues at sites 1 and 3 of VcDyP were not well-conserved in subclasses P1–4. Therefore, each of class P DyP-type peroxidases may have different radical sites. The tryptophan residue at site 3 of TcDyP is conserved in subclasses I1–4, suggesting that class I DyP-type peroxidases use the site 3 tryptophan residue as the active radical site for large substrates. Interestingly, subclass I2 also contains a site 5 tryptophan residue. Because the site 3 tryptophan residue of subclass I2 is occasionally substituted with phenylalanine, the site 5 tryptophan residue may also function as a radical site. The tryptophan residue at site 5 of AauDyP1 is conserved in subclasses V1–4, suggesting that class V DyP-type peroxidases use the site 5 tryptophan residue as the active radical site.

3.3. DyP-type peroxidases from Dikarya

Many DyP-type peroxidases of class V from Basidiomycota have been investigated for their lignin-degrading ability [22]. However, the phylogenetic tree shows that DyP-type peroxidases from Basidiomycota are distributed in class P as well as in class V (Fig. 1, Fig. 2). Their origins were different. Class V (subclass V2) originates in Agaricomycetes, including white-rot and brown-rot fungi, and Pucciniomycetes. Class P (subclass P4) originates in Tremellomycetes and Exobasidiomycetes (Table 2). DyP-type peroxidases from Ascomycota are also distributed in class V and P. In class P, they are distributed in two subclasses, P2 and P4. Subclasses V2 and P4 originate in Dothideomycetes, Eurotiomycetes, Leotiomycetes, and Sordariomycetes. Subclass P2 originates in Leotiomycetes and Sordariomycetes. These differences in organisms of origin may indicate differences in physiological roles.

Table 2.

Detail of branches originating from Dikarya included in subclasses P2, P4, and V2.

phylum	subphylum	class	P2	P4	V2
			branch numbers
Ascomycota	Pezizomycotina	Dothideomycetes	0	4	23
Ascomycota	Pezizomycotina	Eurotiomycetes	0	11	53
Ascomycota	Pezizomycotina	Leotiomycetes	1	1	8
Ascomycota	Pezizomycotina	Sordariomycetes	5	15	7
Basidiomycota	Agaricomycotina	Agaricomycetes	0	0	127
Basidiomycota	Agaricomycotina	Tremellomycetes	0	6	0
Basidiomycota	Agaricomycotina	Pucciniomycetes	0	0	8
Basidiomycota	Agaricomycotina	Exobasidiomycetes	0	2	0
unclassified	–	–	1	1	1

4. Conclusion

Comprehensive phylogenetic analysis of the DyP-type peroxidase family revealed the existence of unexplored subclasses I1, I2, P1, P2, and V4, which show novel origins or have another domain in addition to the DyP domain. Moreover, the conservation of important residues revealed the similarities and differences among the classes.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: