The bacterial strains and plasmids used in this study are listed in Table 1. The wild-type and mutant strains of A. media strain WS were grown at 30°C in Luria–Bertani (LB) medium supplemented with chloramphenicol (30 μg/mL) or ampicillin (100 μg/mL) when necessary. Escherichia coli were routinely grown at 37°C in LB medium and supplemented with kanamycin (50 μg/mL) or chloramphenicol (30 μg/mL) when they were harboring plasmids. Bacterial strain E. coli S17-1 (λpir) and plasmid pDM4 were donated by Prof. Shiyun Chen of Wuhan Institute of Virology, Chinese Academy of Sciences, Wuhan, China. Plasmid pBBR1MCS-5 was kindly donated by Prof. Zhixiong Xie of Wuhan University, Wuhan, China.

Bacterial culture was taken 24 h after the start of the cultures and centrifuged at 13,000 × g for 10 min to remove bacterial cells. The supernatant was immediately acidified with 0.1 volume of glacial acetic acid. After centrifugation, the sample was diluted threefold with 10 mM acetic acid and filtered through a 0.45-μm filter (Millipore, MA, United States). Chromatographic separation was achieved on an Agilent Eclipse column, which was fitted with C18 reverse phase column (250 mm × 4.6 mm i.d., 5 μm). Peaks eluting from the column were detected with a photodiode array detector (G1315B) at 260 nm, the absorption maxima of L-DOPA. L-DOPA (Sigma, United States) was used as the standard.

The melanin was prepared from A. media strain WS. Synthetic DOPA-melanin (Sigma, United States) was used as the standard. Bacteria were cultivated in LB medium at 30°C with shaking at 200 rev min^-1. After 72 h, the culture was centrifuged at room temperature for 15 min at 13,000 × g and the supernatant was filtered through a 0.45-μm filter (Millipore, MA, United States). The pH of the cell-free filtrate was adjusted to 3.0 with HCl and shaken at 25°C to hydrolyze the proteins bound with melanin. The filtrate was centrifuged at 13,000 × g and the pellet, containing the melanin, was dissolved in NaOH solution (pH 9.0). The precipitation process was repeated for three times. The melanin deposition was washed with double-distilled water and dried under a vacuum.

The melanin samples for Fourier transform infrared spectroscopy (FT-IR) analysis were prepared by mixing the melanin with potassium bromide powder. Analysis was carried out on a Nicolet 5700 FT-IR Spectrometer (Madison, WI, United States). The spectra recorded over a range of 4,000–400 cm^-1 with a resolution of 4 cm^-1.

The melanin samples (5 mg) were dissolved in 4 mL of methyl alcohol:acetonitrile (v/v 3:1). The electrospray ionization mass spectrometry (ESI-MS) of melanin samples were performed on LCQ Advantage instrument with an ESI source (4.8 KeV). The m/z range scanned in the MS measurements was from 50 to 1,000 or 2,000.

The PPO activity assay in this study was performed on native-PAGE. Briefly, bacterial strains were cultivated for 18 h at 30°C. The cells were harvested by centrifugation at 13,000 × g for 10 min, washed and resuspended in 50 mM Tris-HCl buffer (pH 7.0) and lysed by sonication. Cell lysates were centrifuged at 13,000 × g for 20 min at 4°C to remove cell debris. Electrophoresis was performed at 4°C on polyacrylamide gels, without sodium dodecyl sulfate (SDS) and β-mercaptoethanol to retain the activity of enzymes, at a constant current of 20 mA for about 8 h. PPO activity was assessed after incubating the gel at 30°C in 50 mM Tris-HCl buffer (pH 8.0) plus 5 mM L-DOPA for 30 min with gentle shaking.

Based on the draft genome sequence of A. media strain WS, the whole genome was sequenced using a combination approach of Roche 454 GS-FLX Titanium platform and PacBio RS II system. De novo genome assembly of PacBio reads was performed by Velvet 1.2.03 and Illumina reads sequences were used to estimate gap sizes of adjacent contig pairs and to evaluate correct assembly throughout the project. Open reading frames (ORFs) were predicted with Glimmer and annotated using KEGG, UniProt, and COG (Clusters of Orthologous Group) databases.

The conserved domains of tyrosinase or laccase, which are essential to DOPA melanin formation in bacteria, were screened online at http://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml. The potential proteins were predicted by searching the whole genome against these conserved domains.

Two PPO activity bands were observed by native-PAGE assay. After the identification of one PPO (PPO1) in A. media strain WS via bioinformatic method, the PPO activity of the other PPO (PPO2) was performed on native-PAGE after precipitation with ammonium sulfate at 35% saturation. The staining band of PPO2 was then cut from native-PAGE gel and characterized by liquid chromatography using matrix-assisted laser desorption/ionization/time-of-flight (MALDI-TOF/TOF) mass spectrometer at HuaDa (Guangdong, China).

Peak lists from LC-MS/MS experiments were generated by the LCQ software Bioworks Browser via TurboSEQUEST. The MASCOT search engine¹ was used for the protein database search using the completed genomic sequence of A. media strain WS. The output of the MASCOT search results was extracted to an Excel file to provide a list of proteins identified by the peptides. Finally, candidate proteins exhibiting PPO2 activity were manually examined by reaction analysis of each protein using KEGG database.

The deletion mutants of PPO activity proteins, designated as WS ΔyfiH, WS ΔkatE, and WS ΔyfiHΔkatE, were made by homologous recombination. For all constructions, a suicide vector pDM4 containing fusion PCR fragments located upstream and downstream of yfiH or katE (primer pairs for PCR; see Supplementary Table S1) with in-frame deletions were used. To introduce these plasmids into A. media, conjugal matings were done as follows: the knock-out vectors pDMM1, pDMM2, and pDMM3 were introduced into E. coli S17-1λ (pir) and the resultant strains mated with A. media. Transconjugants were selected by utilizing the chloramphenicol resistance gene located on the suicide plasmid and confirmed by PCR. A single transconjugant of each strain was grown on 15% sucrose plate without antibiotics, allowing the occurrence of the second cross that replaced the wild-type allele with the mutant one. The transconjugants were sensitive to sucrose due to the synthesis of levans (toxic compounds to Gram-negative bacteria) catalyzed by levansuerase, the product of sacB located on pDM4. Thus, only the clones that exchanged for the second time and excised the plasmid-borne sequence could survive on these plates. The resulting clones that could not grow on plates with chloramphenicol were selected and confirmed by PCR and DNA sequencing.

To complement the function of CatA in WS ΔkatE, the intact katE gene was amplified from chromosomal DNA of A. media strain WS using the primer pairs listed in Supplementary Table S1. The PCR product was cloned into the shuttle vector pBBR1MCS-5 (Kovach et al., 1995) and the resulting plasmid pBBR1MCS-5-katE was introduced by conjugation into the mutant strain WS ΔkatE. The shuttle vector pBBR1MCS-5 was employed as well, to construct the complementary strain of hpii and hpi in WS ΔkatE.

Catalase (CAT) activity was measured using an assay kit (Jiancheng, Nanjing, China), according to the manufacturer’s instructions. This kit utilizes the peroxidatic reaction of CAT for determination of enzyme activity. The hydrogen peroxide decomposition catalyzed by CAT was performed at 37°C for 1 min and then ammonium molybdate was added to terminate the reaction immediately. The residual hydrogen peroxide was monitored at 405 nm, the absorption maxima of the pale yellow complex produced by ammonium molybdate and hydrogen peroxide. CAT activity (U/mL) was defined as the consumption of H₂O₂ (μmol/sec) in buffer catalyzed by CAT at 37°C.

To construct the expression vector, the ORFs of CatA (1,449 bp), Hpi (2,181 bp) and Hpii (2,262 bp) were amplified from the genomes of A. media WS and E. coli BL21 (DE3) using primer pairs listed in Supplementary Table S1, respectively. The PCR products were inserted into the appropriate restriction enzyme sites of pET-26b (+) to yield pET-CatA, pET-Hpi, and pET-Hpii. E. coli BL21 (DE3), transformed with these plasmids were grown in LB medium containing 50 μg/mL kanamycin at 28°C and induced with 1 mM isopropyl-β-D-thiogalactoside at the exponential phase of growth (OD₆₀₀ = 0.6). Purification of His₆-tagged enzymes was carried out with Ni-NTA His●bind resins (Novagen). Bound proteins were eluted with a linear gradient from 20 to 500 mM imidazole. Fractions containing His₆-tagged enzymes were determined by SDS-PAGE. Fractions containing the expected proteins were collected and dialyzed against 50 mM Tris–HCl buffer (pH 7.5). The concentration of purified proteins was determined using the Bradford assay (Bradford, 1976). The enzymatic analysis for PPO activity and CAT activity of active His₆-tagged CatA, Hpi, and Hpii were performed as described above.

Production of melanin in A. media strain WS and its PPO mutants (WS ΔyfiH, WS ΔkatE, and WS ΔyfiHΔkatE) was monitored, using the method of Gibson (Gibson and George, 1998). To estimate the final melanin production, bacterial cultures were filtered through a 0.45-μm filter (Millipore, MA, United States) after 72 h of cultivation and the absorbance at 400 nm of the supernatants was measured. Growth of each strain was assessed by its OD at 600 nm. The capacity for melanin formation of wild-type and mutant strains was calculated using the OD₄₀₀/OD₆₀₀ value (Chatfield and Cianciotto, 2007); melanin production at various periods was recorded within 48 h of cultivation.

Cell viability of A. media wild-type strain WS and its catalase-deficient strain, WSΔkatE, was assessed with 2,3,5-triphenyltetrazolium chloride (TTC) staining (Towill and Mazur, 1975). After 8 h in culture, equal inocula of each strain were transferred to LB medium grown at 30°C. Cells were collected at various times by centrifugation and suspension in 50 mM Tris-HCl buffer (pH 8.0) containing 0.05% TTC (Sigma, United States). Solutions were shaken in the dark for 12 h at 30°C and centrifuged at 13,000 × g. The precipitate was treated with 85% dimethyl sulfoxide (DMSO) to extract the triphenylformazan (TPF), the red TTC reduction product. After centrifugation, the absorption maximum of TPF was monitored at 485 nm. Growth of each strain was evaluated at each time point by measuring OD at 600 nm. The cell viability was calculated using the OD₄₈₅/OD₆₀₀ value. All assays were repeated for three times.

The fluorimetric probe dichloro-dihydrofluorescein diacetate (DCFH-DA, Sigma–Aldrich) was used for detection of reactive oxygen species (ROS). An RF-5301PC spectrofluorometer (Shimadzu, Inc., Kyoto, Japan) equipped with 1-cm quartz cell was used for fluorometric measurement. Bacterial cells were collected before and after pigmentation. Each sample was loaded with DCFH-DA to a final concentration of 20 μM and incubated in dark at 30°C for 40 min before the fluorometric measurement. Qualitative oxidative stress assay was determined by microscopy image analysis and quantitative determination was measured by fluorescence intensity of dichlorofluorescein (DCF) at an excitation wavelength of 499 nm and an emission wavelength of 521 nm. All assays were repeated at least four times.

Our previous work shows that HppD in the HGA pathway is crucial in melanin synthesis of A. media strain WS, a finding consistent with the observation that the WS ΔhppD mutant is completely incapable of pigment formation (Wang et al., 2015). However, melanin formation in this mutant was partially restored when the culture medium was supplemented with 2 mg/mL tyrosine (Supplementary Figure S1). We hypothesized that A. media strain WS is capable of synthesizing melanin via the DOPA-based pathway, using tyrosine as a substrate.

We tested this hypothesis in several ways. First, we purified the melanin produced by a culture of A. media strain WS using the purification procedure for bacterial melanin, and analyzed it with FT-IR. The absorption spectra are shown in Supplementary Figure S2A. The main absorption peaks of the purified melanin are quite similar to those of synthetic DOPA-melanin, which led us to suggest that DOPA-melanin was present in the sample from A. media strain WS.

We next employed mass spectrometry to compare the structure of the acid hydrolysis products of melanin from A. media strain WS (referred to here as WS melanin) and synthetic DOPA-melanin (obtained from L-DOPA). Both of these were analyzed by ESI-MS, and yielded similar spectra. Both samples displayed a protonated molecule [M+H]⁺ at m/z 147 (Figure 1), which corresponds to the characteristic monomer of DOPA-melanin, indole-5, 6-quinone (Costa et al., 1992). The ions at m/z 274, 318, and 717 in both MS spectra probably resulted from the fragmentation of a melanin polymer. These results provide further evidence supporting our hypothesis that DOPA-melanin was produced in A. media strain WS. Another observation from these mass spectrometry results is of interest. Pyomelanin, synthesized via the HGA pathway, is believed to be the main component of WS melanin (Wang et al., 2015). The protonated molecule [M+H]⁺ of HGA at m/z 169 was present in the spectra of WS melanin but not the spectra of synthetic DOPA-melanin, consistent with the presence of pyomelanin in the WS pigment sample also.

Our final test of the hypothesis was to perform HPLC analysis of pigment from A. media strain WS. The results are presented in Supplementary Figure S2B and show the presence of the precursor of DOPA-melanin, L-DOPA. This result is consistent with previous reports (Gibson and George, 1998; Wan et al., 2007). These results, together, lead to our conclusion that A. media strain WS is capable of synthesizing DOPA-melanin under certain culture conditions.

In bacteria, PPO, are usually responsible for DOPA-melanin synthesis. The PPOs are usually tyrosinases, but are occasionally laccase (Fairhead and Thöny-Meyer, 2012). We previously identified the PPO TyrA by native-PAGE at pH 9.0 from A. media strain WS (Wan et al., 2009). However, in this study, we noted that deletion of tyrA in the WS ΔhppD strain did not affect pigment production by the mutant when it was cultured in LB media supplemented with 2 mg/mL tyrosine. Further, the bands exhibiting PPO activity were still detected on native-PAGE (data not shown). From these observations, we conclude that TyrA is not a key PPO in A. media strain WS. Therefore, we turned our attention to identification of the PPOs that are responsible for the conversion of L-DOPA to DOPA-melanin in this strain. To this end, we performed a PPO activity assay in a gradient pH staining buffer. Cell lysate of A. media strain WS was subjected to native-PAGE staining with 5 mM L-DOPA as substrate. Two bands exhibiting PPO activity were detected at pH 7.5 or 8.0 (Figure 2A). We conclude that there are at least two PPOs that can convert L-DOPA to DOPA-melanin in WS, which we designated PPO1 and PPO2 here (Figure 2B).

The complete genome of A. media strain WS has been sequenced (GenBank sequence: CP007567.1), but no protein has been annotated as a PPO. It has been known for several decades that tyrosinases from bacteria, including Streptomyces, have the ability to convert L-DOPA into melanin (Fairhead and Thöny-Meyer, 2012), but no protein with a typical bacterial tyrosinase domain has been found in A. media strain WS by searching for the common central domain of tyrosinase (Pfam00264; Finn et al., 2006). Besides tyrosinase, proteins belonging to the multi-copper oxidase family are also involved in DOPA-melanin formation in bacteria, for example, laccase in Sinorhizobium meliloti (Claus and Decker, 2006). These proteins often contain multiple conserved domains of Cu-oxidase enzymes, notably, pfam00394, pfam07731, pfam07732, and pfam02578 (Messerschmidt and Huber, 1990; Messerschmidt et al., 1992; Page et al., 1999; Beloqui et al., 2006). One protein classified as K05810 in the KEGG Orthology (KO) database (protein SeqID: B224_5050) stood out when we screened all the predicted proteins against four conserved domains of multi-copper oxidase. Protein blast analysis of B224_5050 (gene name: yfiH) revealed that it belongs to the Cu-oxidase_4 family, which contains the pfam02578 conserved domain homologous to laccase.

To determine whether yfiH encodes a PPO, an in-frame deletion mutant of yfiH was constructed. Cells of both A. media strain WS (wild type) and mutant WS ΔyfiH were collected at the same period, and the cell lysates were subjected to native-PAGE staining with 5 mM L-DOPA to detect the diphenol oxidase activity. As shown in Figure 3, the band corresponding to PPO1 disappeared in WS ΔyfiH, indicating that the laccase-like protein YfiH of A. media strain WS (B224_5050) was a PPO.

Although we concluded that the identity of the gene responsible for PPO1 activity had been established, we were unable to identify the gene responsible for PPO2 activity by bioinformatic analysis. Thus, we employed LC-MS/MS to determine the identity of PPO2.Crude extracts of A. media strain WS were precipitated by addition of ammonium sulfate gradually until it was 80% saturated (Ito, 2000). Sediments at 45% of ammonium sulfate displayed the highest PPO2 activity, so we selected these for dialyzation and native-PAGE staining with 5 mM L-DOPA. The band exhibiting PPO2 activity was cut and characterized by LC-MS/MS.

Peptides from LC-MS/MS were searched against the complete genomic sequence of A. media strain WS and 55 proteins were identified in the band with PPO2 activity (Supplementary Table S2). All of these had been annotated in the genome database. From a literature search, we found that one protein (protein SeqID: B224_4951) annotated as “catalase” (CatA) may be involved in melanin formation. Catalase affects melanin formation in eukaryotes (Skamnioti et al., 2007; Maresca et al., 2008). It has also been demonstrated that hydrogen peroxide (H₂O₂) is generated in a series of chemical reactions coupled to the enzymatic formation of o-quinones by tyrosinase acting on monophenols and o-diphenols, and during the auto-oxidation of the o-diphenols and other intermediates in this pathway (Munoz-Munoz et al., 2009). Notably, catalase, superoxide dismutase, or peroxidase help remove the H₂O₂ generated by reaction R00009 (conversion of hydrogen peroxide to oxygen and water) and R00602 (conversion of methanol and hydrogen peroxide to formaldehyde and water; Mithani et al., 2011). Moreover, we found that the enzymatic reaction catalyzed by PPO in DOPA-melanin formation is very similar to R02670 (the conversion of 3-hydroxyanthranilate to cinnavalininate), which corresponds to catalase in the context of the KEGG tryptophan metabolism map (MAP00380; shown in Supplementary Figure S3). This analysis led us to suggest that catalase is involved in melanin formation, via PPO activity.

To determine whether CatA is important for PPO activity, we deleted the CatA-encoding gene katE in the wild type A. media strain WS and the ΔyfiH mutant strain and tested whether deletion of katE affected the PPO2 activity. As shown in Figure 3, the band corresponding to PPO2 disappeared in both the ΔkatE and ΔyfiHΔkatE mutant, indicating that CatA is critical for PPO2 activity. Meanwhile, the PPO2 activity of both the mutant strains could be restored by supplying the katE gene in trans on the plasmid pBBR1MCS-5. Taken together, these results confirmed that CatA in A. media strain WS is responsible for the PPO activity of PPO2.

From these results, we hypothesized that YfiH and CatA have PPO activity and are important in the synthesis of DOPA-melanin in A. media strain WS. We next investigated their effect on melanin production.

The first step was to assess the importance of YfiH and CatA in the residual pigmentation in WS ΔhppD supplemented with tyrosine. yfiH and katE were deleted individually or in combination in the ΔhppD mutant of A. media strain WS to yield the mutants WS ΔyfiHΔhppD, WS ΔkatEΔhppD, and WS ΔyfiHΔkatEΔhppD. The production of melanin was measured in these mutants. Supplementary Figure S4 shows that the residual pigmentation was still present in these mutants, leading us to conclude that YfiH and CatA are not correlated with the residual pigmentation of A. media strain WS lacking HppD.

We next examined the role of YfiH and CatA on the entire, intact pigmentation process in A. media strain WS. Deletion mutants of yfiH, katE, or of both genes were cultured in LB medium and the pigmentation process was compared to the wild-type strain. While deletion of yfiH did not affect pigmentation at all, deletion of katE caused about a 55% decline of pigmentation (Figures 4A,B). Melanogenesis began about 9 h earlier in the mutant than in the wild-type strain (Figure 4C).

Given that secondary metabolism (including pigment metabolism) may be repressed when cell viability is reduced, we thought it wise to assess cell viability in the CatA-deficient mutants. We measured the cell viability of both the wild-type strain and the CatA-deficient strain with the TTC method. The reduction of TTC to TPF occurs in living but not dead cells. As shown in Supplementary Figure S5, the curves of cell viability between the wild-type strain and WS ΔkatE were similar, with viability declining in both after 20–22 h in culture. When the cells had been in culture from 14 to 20 h, the production of pigmentation in WS ΔkatE increased minimally; although, the cells were clearly viable.

Expression of CatA from a plasmid pBBR1MCS-5-katE (containing the entire katE from A. media strain WS) in the ΔkatE mutant restored pigmentation to the wild-type level (Figures 4A,B). From this series of experiments, we suggest that CatA is not important for the residual pigmentation in WS ΔhppD. However, it may regulate pigmentation in A. media strain WS via other mechanisms independently of its PPO activity.

We next wished to elucidate the role of CatA in pigmentation. The first step was to confirm that CatA does, indeed, have catalase activity. Catalases have been reported in Aeromonas (Barnes et al., 1999; Rio et al., 2007). Sequence alignment showed that CatA from A. media strain WS had 98% amino acid sequence homology to catalases from other Aeromonas species. We assessed the catalase activity of CatA in the wild-type strain and the mutant strain deficient in CatA, using a catalase assay kit (Wrześniok et al., 2013). The results in Table 2 show strong catalase activity (3.68 U/mL) in the cell lysates of the wild-type strain, but not in cell lysates of the ΔkatE mutant. No catalase activity was detected in the cell supernatant from either strain (Table 2), indicating that CatA is an endoenzyme. These results confirmed that CatA has catalase activity.

In human melanocytes and cells of the rice blast fungus Magnaporthe grisea, both eukaryotic, the effect of catalase on hydrogen peroxide (H₂O₂) decomposition is well understood and correlates with melanin formation (Skamnioti et al., 2007; Maresca et al., 2008). Hydrogen peroxide generation has been observed in both the enzymatic reaction phase and the auto-oxidation phase of melanin biosynthesis (Munoz-Munoz et al., 2009). This information led us to speculate that production of hydrogen peroxide regulates pigment synthesis, and must be removed for maximum pigment production. To test this idea, we asked if expression of catalases from other bacteria in the WS ΔkatE mutant could restore pigmentation to the wild-type level. Catalases Hpi and Hpii from E. coli BL21 (DE3), which show 47% amino acid sequence similarity with CatA from A. media strain WS, were chosen for this experiment (Supplementary Figure S6). As shown in Figure 5, expression of Hpi or Hpii from plasmids pBBR1MCS-5-hpior pBBR1MCS-5-hpiiin strain WS ΔkatE restored pigmentation to the wild-type level. We interpreted these results to support our suggestion that oxidative stress regulates pigmentation of A. media strain WS.

The results above led us to hypothesize that oxidative stress affects pigmentation in A. media strain WS, but the mechanism was unknown. To investigate the mechanism, we assessed the production of ROS during pigmentation in A. media strain WS, using the fluorimetric probe assay (DCFH-DA). Non-ionic DCFH-DA crosses cell membranes and is enzymatically hydrolyzed by intracellular esterases to non-fluorescent DCFH, which is then oxidized to fluorescent dichlorofluorescein (DCF) by hydrogen peroxide (LeBel et al., 1992). As showed in Figure 6, formation of fluorescent DCF was almost undetectable in the wild-type strain before pigmentation, suggesting that there was very little H₂O₂ produced. However, in the pigmented culture, a high level of DCF was detected in cells, indicating that H₂O₂ was generated during the pigmentation process. Unlike the wild-type strain, strong fluorescence of DCF was observed in the CatA deficient-mutant WS ΔkatE before pigmentation, indicating that in the absence of CatA, ROS accumulated (Figure 6A). Interestingly, we found that the fluorescence signal of DCF in the WS ΔkatE cells was higher before pigmentation than after pigmentation, suggesting that the ROS level of WS ΔkatE before pigmentation (15 h after the start of the culture) was even higher than in the pigmented period (at 24 h). Oxidative stress-induced melanogenesis has been explored in some organisms such as in human B16F10 melanoma cells (Kim, 2012). The accumulation of ROS in WS ΔkatE may explain the 9-h advance in production of melanin of the WS ΔkatE mutant (Figure 4C). There are reports that melanin synthesis caused oxidative stress, for example, generation of H₂O₂ (Munoz-Munoz et al., 2009). Consistent with this, the fluorescence signal of DCF in the wild-type cells was similar to the WS ΔkatE cells after pigmentation, which produced less than half of the melanin. These results suggest that accumulation of ROS induced production of melanin, but it may have negatively affected melanin synthesis at later stages.

Earlier in this study, we demonstrated that CatA of A. media strain WS was important for PPO2 activity. We next wished to extend this finding, and confirm that the purified CatA protein has polyphenol activity as well as test catalases from other bacteria for PPO activity. To this end, CatA from A. media strain WS and catalases Hpi and Hpii from E. coli, were tagged with a His6 tag, purified, and analyzed by gel filtration. The purified proteins had a single peak with a molecular mass of 55, 80, and 84 kDa, respectively (Figure 7A). All His₆-tagged proteins exhibited both catalase and PPO activity as we predicted (Table 2 and Figure 7B). It appears that catalases are bi-functional enzymes and may constitute a new class of PPOs.