Molecular identification, heterologous expression and properties of light-insensitive plant catalases


J ürgen Feierabend. Fax: + 49 69 798 24822; e-mail:


Most catalases are inactivated by light in a heme-sensitized and O2-dependent reaction. In leaves of the alpine plant Homogyne alpina and in the peroxisomal cores of Helianthus annuus, light-insensitive catalases were observed. For the catalases Hacat1 of H. alpina and HnncatA3 of H. annuus, cDNA clones were obtained. Expression of recombinant active enzymes in insect cells confirmed that they coded for light-insensitive catalases. Kinetic and catalytic properties of light-sensitive or light-insensitive catalases did not differ substantially. However, the specific activity of the latter was markedly lower. The light-insensitive catalase HaCAT-1 was not resistant against inactivation by superoxide. Amino acid sequences of the light-insensitive catalases HaCAT-1 and HNNCATA3 were highly identical. They showed only a few exceptional amino acid substitutions at positions that are highly conserved in other catalases. These appeared to be localized mainly in a surface cavity at the entrance of a minor channel leading to the central heme, suggesting that this region played some, though yet undefined, role for light sensitivity. While the replacement of a highly conserved His by Thr225 was the most unique substitution, a single exchange of His225 by Thr in the light-sensitive catalase SaCAT-1 by mutagenesis was not sufficient to reduce its sensitivity to photoinactivation.


bovine liver catalase


photosynthetically active radiation


polymerase chain reaction


rapid amplification of cDNA ends


reactive oxygen species


reverse transcriptase polymerase chain reaction


sodium dodecyl sulfate polyacrylamide gel electrophoresis.


Catalases degrade H2O2 by dismutation to H2O and O2. Plants contain monofunctional, tetrameric and heme-containing catalases that are mostly localized in peroxisomes or glyoxysomes (Willekens et al. 1995; Heinze & Gerhardt 2002; Feierabend 2005). In photosynthesizing green leaves, H2O2 is produced mainly in the peroxisomes by the oxidation of glycolate during photorespiration or in the chloroplasts through the univalent reduction of O2 by photosystem I in the Mehler reaction and a subsequent dismutation of the superoxide. Both photorespiration and the Mehler reaction are enhanced by high light, particularly under conditions of limiting carbon assimilation (Foyer 1997; Foyer & Noctor 2003). While catalases are thus indispensable for the antioxidative protection of C3 plants in light (Willekens et al. 1997), catalases are themselves inhibited by ROS, such as superoxide (Kono & Fridovich 1982; Shimizu, Kobayashi & Hayashi 1984), and catalases are generally inactivated by UV or visible light and in the presence of O2. Bacterial, mammalian as well as plant catalases are photoinactivated in the presence of O2 by blue light which is absorbed by the prosthetic hemes (Cheng, Kellog & Packer 1981; Feierabend & Engel 1986; Shang & Feierabend 1999). According to Aubailly et al. (2000), the heme-sensitized inactivation of catalase by UV light is mediated by a photoreduction of the Fe(III) of the native ferricatalase with a subsequent binding of O2 and the generation of the unstable compound III state of catalase (oxyferrous catalase).

Catalases are not only in vitro but also in intact cells and tissues inactivated by light. In leaves, catalase appears to be also inactivated by oxidative damage that is initiated in the chloroplasts under high intensities of red light (Shang & Feierabend 1999). In green leaves, an apparently constant steady state level of catalase is maintained in visible light under favourable conditions, but only because the loss by photoinactivation is concomitantly replaced by de novo synthesis. Therefore, catalase exhibits a rapid turnover in leaves, which increases with the intensity of light, and the rate of new catalase synthesis must be flexibly attuned to fluctuating light conditions (Hertwig, Streb & Feierabend 1992; Shang & Feierabend 1999; Schmidt, Dehne & Feierabend 2002). Protein synthesis is repressed under unfavourable stress conditions, such as low temperature or chemical stresses. Therefore, apparent losses of catalase activity were observed in a wide variety of plants under such conditions and are generally accompanied by similar declines of photosystem II activity. This is because the reaction centre protein D1 of photosystem II also exhibits a rapid turnover in light and depends on repair (Feierabend, Schaan & Hertwig 1992; Mishra, Mishra & Singhal 1993; Streb & Feierabend 1996; Streb, Shang & Feierabend 1999).

However, plants living under extreme environmental conditions, such as high mountain plants frequently exposed to high irradiance at low temperatures, would not be able to survive when they are depleted of catalase and photosystem II activity. Indeed, in leaves of acclimated high mountain plants or hardened winter cereals, both catalase and photosystem II appeared to be largely resistant to photoinactivation (Streb, Feierabend & Bligny 1997; Streb et al. 1999). Nevertheless, after extraction from the leaves, the catalases of most acclimated plants were as light-sensitive in vitro as those from non-hardened plants. Only in vivo were they more efficiently protected. By contrast, leaves of the alpine plant Homogyne alpina appeared to contain a light-stable catalase (Streb et al. 1997). In order to elucidate the peculiar properties of a light-insensitive catalase, a cDNA for the H. alpina catalase was identified and cloned, and the recombinant enzyme was produced by heterologous expression in insect cells after infection with recombinant baculovirus. During the course of these investigations, sequences of catalase isozymes from sunflower became available. Interestingly, two of the known sunflower isozymes, HNNCATA3 and HNNCATA4, which are localized in the cores or the crystalline inclusions of sunflower peroxisomes (Tenberge & Eising 1995; Tenberge et al. 1997), showed very similar unusual amino acid substitutions, as observed in the catalase from H. alpina. Furthermore, it had been observed that sunflower catalases from the peroxisomal cores were also less sensitive to photoinactivation than those of the peroxisomal matrix (Grotjohann, Janning & Eising 1997; Eising et al. 1998). Therefore, sunflower HNNCATA3 was also expressed in insect cells and was included in a comparative investigation of the specific functional and structural properties of light-insensitive catalases.



Vectors were obtained from Pharmingen (San Diego, USA), Promega (Mannheim, Germany) and Stratagene (Heidelberg, Germany). Cell cultures of Spodoptera frugiperda were obtained from American Type Culture Collection (Rockville, Maryland, USA; ATCC no. CRL 17-11). For biochemical work, leaves of the alpine plants Soldanella alpina (L.) and Homogyne alpina (L.) were collected in the French alps between the Lautaret pass and the Galibier pass (see Streb et al. 1997). For electron microscopy, leaves of H. alpina were collected from an alpine meadow in the Central Tyrolean Alps near Obergurgl (Ötztal), Austria. Leaves of winter rye (Secale cereale L. cv. ‘Halo’) were grown under controlled conditions, and the catalase was prepared as described previously(Hertwig et al. 1992). Sunflower seedlings (Helianthus annuus L.) were grown in darkness in moist vermiculite for 3 d, and in continuous white light of 120 µmol m−2 s−1 PAR1 for subsequent 3 d, according to Tenberge et al. (1997).

Electron microscopy

Samples from leaves of H. alpina (approximately 1 mm2) were excised at the growth site and immediately fixed in a glutaraldehyde/cacodylic acid buffer system. Post-fixation and embedding were performed as described previously (Lütz 1987). Ultrathin sections were observed in a Zeiss TEM 902 electron microscope (Carl Zeiss AG, Ober-Kochen, Germany).

Construction of catalase clones

The cDNA for the leaf catalase of Secale cereale (L.), Sccat1a (EMBL accession number: Z54143), was cloned in Bluescript SK– as described previously (Schmidt & Feierabend 2000). A cDNA clone for S. alpina leaf catalase Sacat1 (accession number Z99633), with a calculated molecular mass of 56.9 kDa, was identified by screening a cDNA library in lambda ZAPII, which was constructed from a poly[A]+RNA preparation from leaves of the alpine plant S. alpina with a Sccat1a cDNA as probe. The cDNA was cloned in the pBK-CMV-vector. A cDNA sequence for a leaf catalase from H. alpina, Hacat1 (accession number: AJ 616025), was identified by PCR. From a poly[A]+RNA preparation of H. alpina leaves, a 512 bp fragment was obtained by RT-PCR with degenerate primers that were homologous to conserved regions of plant catalases (forward primer: 5′-CAYGCNTTNAARCCNAAYCC-3′; reverse primer: 5′-GGRCARAANGCYAAYTGYTCRT TYTCNGC-3′; the primers correspond to amino acids 156–162 and 317–326, respectively, of the Hacat1 sequence of Fig. 1). Using the 5′-3′-RACE Kit of Roche (Mannheim, Germany), two partial fragments for the 5′- and 3′-ends of the cDNA were obtained and amplified by PCR with anchor and internal primers homologous to sequences of the 512 bp fragment (5′-fragment: forward anchor primer 5′-GACCACGCGTATCGATGTCGACT15V-3′, reverse primer 5′-TAAACTCTCGGGATGGTG-3′; 3′-fragment: forward primer 5′-ATGGATCCTTACAAGCAGCG GCC-3′, reverse anchor primer 5′-GACC ACGCGTATCGATGTCGACT15V-3′). Based on this sequence information, a 1725 bp DNA fragment containing a complete open reading frame for a catalase polypeptide of 492 amino acids with a calculated molecular mass of 56.7 kDa was produced by PCR (forward primer: 5′-ATGGATCCTTACAAGCAGCGGCC-3′; reverse primer: 5′-GACCACGCGTATCGATGTCGAC-3′) and cloned in the pGEM-T vector (Promega, Mannheim, Germany). A 1488 bp DNA fragment coding for the complete open reading frame of the sunflower catalase isozyme HnncatA3 was produced by RT-PCR with specific primers deduced from the known cDNA nucleotide sequence (accession number AF 243518; forward primer 5′-CAATACCATGGATCCCTACAAGCAACG-3′; reverse primer 5′-AGTTCAGTAGTTGGGGCGCACATTGAG-3′) and a poly[A]+RNA preparation from cotyledons of sunflower seedlings. The DNA fragment was cloned in pGEM-T easy (Promega, Mannheim, Germany).

Figure 1.

Alignment of the deduced amino acid sequence of a catalase from leaves of Homogyne alpina CAT-1 (HaCAT-1) with catalase sequences from other dicotyledonous or monocotyledonous plants and with BLC. Amino acids that are identical to HaCAT-1 among plant catalases are marked in grey. Identity between HaCAT-1 and the other catalases including BLC is marked in black, allowing two exceptions at the most among the plant catalases. Exceptional substitutions in HaCAT-1 and Helianthus CATA3 (HNNCATA3) are marked by triangles pointing downwards.

Site-directed mutagenesis

The conversion of His225 of Sacat1 into Thr was performed by PCR (Imai et al. 1991) with Pfu polymerase and mutant oligonucleotide primers (5′-CGTGGAAACCTACAT GTGGAGTG-3′ and 5′-TGAACTTAACATAT TGAGCTTTTC-3′; mutated bases are underlined) using a circular Sacat1 containing pUC vector as template. The mutation simultaneously introduced a second cut for the restriction enzyme Eco72I which allowed a rapid identification of mutant clones. The mutations were verified by sequencing.

Sequence analysis

The cDNA fragments for Sacat1 and Hacat1 were sequenced by the dideoxynucleotide chain termination method of Sanger, Nicklen & Coulson (1977) using [35S]dATPαS and the T7-sequencing kit of Pharmacia Biotech (Freiburg, Germany). All other sequences were determined or verified by MWG AG Biotech (Ebersberg, Germany). Nucleic acid and protein sequences were analysed with the OMIGA 1.1 software package (Oxford Molecular Group, Cambridge, UK). Models of tertiary structures were analysed with the Swiss-PdbViewer (Swiss Institute of Bioinformatics, Geneva, Switzerland; Guex & Peitsch 1997;

Expression of recombinant catalases

Because attempts for heterologous expression of active plant catalases in Escherichia coli, which produced only inactive protein, and in yeast systems were not successful, recombinant catalases were produced in Sf9 insect cell cultures after infection with recombinant baculovirus. Recombinant transfer vectors were constructed by ligation of the 1.685 kbp EcoRI fragment from Sccat1a-clones in pYEX-BX, the 1.772 kbp EcoRI fragment from Sacat1-clones in pYEX-BX and the 1.725 kbp NotI/Apa fragment from Hacat1-clones in pGem-T into the appropriately digested transfer vector pVL1393. The 1.7 kbp EcoRI fragment from the HnncatA3 clone in pGem-T-Easy and the EcoRI fragment of the mutant Sacat1H225T clone in pUC18 were ligated into the transfer vector pVL1392. The constructs were verified by restriction analysis and DNA sequencing. Recombinant baculovirus was generated through homologous recombination by cotransfecting Spodoptera frugiperda insect cells (Sf9 cells) with 0.1 µg linearized BaculoGold-DNA (Pharmingen; modified Autographa californica nuclear polyhedrosis virus) and 0.4 µg of recombinant transfer vector DNA using cationic liposomes (Lipofektin, Gibco BRL Life Technologies, Karlsruhe, Germany) according to manufacturer's instructions. Recombinant baculoviruses were detected by a plaque assay (O’Reilly, Miller & Luckow 1992; Grünewald et al. 1996).

The ability of the recombinant baculoviruses to induce the expression of active recombinant catalases was examined by infecting Sf9 insect cell lines. For amplification, a 5 mL Sf9 culture was infected with 50 µL of primary recombinant baculovirus solution. After 7–10 d growth at 27 °C, cells were sedimented by 5 min centrifugation at 1500 g, and the supernatant was used as virus stock solution. For the production of recombinant catalases, Sf9 cells were grown in monolayer cultures in 175 or 500 cm2 NunclonTM (Fisher Scientific, Schwerte, Germany) cell culture flasks at 27 °C in TNM-FH medium containing 5% (v/v) fetal calf serum and 100 µg mL−1 gentamycin according to Summers & Smith (1987). Cells were infected with recombinant baculovirus at a density of (1.5–3) × 106 cells mL−1. For each recombinant baculovirus, the amount of virus needed to induce maximal production of recombinant catalase activity was determined empirically. For controls, insect cells were infected with baculovirus containing only vector sequences without catalase sequence insert. In order to optimize the expression of catalases, hemin was added to a final concentration of 20 µm (from a 10 mm stock solution obtained after solubilization with 1 m NaOH, addition of 0.2 m Hepes/KOH pH 8.2 and neutralization to pH 7.0 with 1 m HCl), and the protease inhibitor Pefa Block to a final concentration of 65 µg mL−1 at 48–50 h after infection. The optimal time of hemin addition after infection was empirically determined for each clone. Cells were harvested 96–108 h after infection for the extraction of recombinant catalases from S. cereale and S. alpina, and after 120–132 h after infection for the extraction of recombinant catalases from H. alpina and H. annuus. After sedimentation for 10 min at 2500 g, cells were either stored at −20 °C or suspended with 0.5 m K-phosphate, pH 7.5, containing 2% (v/v) Triton X-100 and 0.1 mg mL−1 DNase I (from bovine pancreas). After 1–2 h incubation on ice, cells were disrupted by sonification (5 × 10 s), and the homogenates were centrifuged for 15 min at 10 000 g and 4 °C. The supernatants were assayed for catalase activity and used for further purification of the recombinant enzymes.

The production of recombinant catalases was verified by Western blotting, and the identity of the selected recombinant baculovirus was verified by Southern analysis with specific cDNA probes. For the extraction of total DNA, Sf9 cells were sedimented and lysed for 12–18 h at 50 °C with 10 mm Tris/HCl, pH 8.0, 25 mm EDTA, 0.1 m NaCl, 0.5% (w/v) SDS and 0.1 mg mL−1 proteinase K. DNA was purified after extraction with 0.5 volume phenol and 0.5 volume chloroform/isoamyl alcohol (24:1) according to standard procedures (Sambrook, Fritsch & Maniatis 1989).

Enzyme purification

Catalase from rye leaves was extracted and purified as described by Hertwig et al. (1992). A peroxisomal core fraction was prepared from sunflower cotyledons according to Tenberge et al. (1997). The supernatant of a crude extract of sunflower cotyledons was used as a source of soluble (matrix) catalase. For the preparation, cotyledons were homogenized with 50 mm K-phosphate, pH 7.5, containing 10 mm Na2S2O4 and were centrifuged for 20 min at 20 000 g and 4 °C. The supernatant was passed over a NAPTM-10 column (Pharmacia Biotech) equilibrated with 50 mm K-phosphate, pH 7.5, in order to remove the Na2S2O4. The eluted solution was recentrifuged as before, and the supernatant was used for the assays.

For the individual catalase clones, different procedures had to be applied for the isolation of recombinant catalases from insect cells. Recombinant ScCAT-1 was purified from total cell extracts (see above) by precipitation with ethanol/chloroform according to Hertwig et al. (1992), except that silanized glass beads (0.18 mm diameter) were used instead of Celite to improve subsequent precipitation with ethanol. The resulting enzyme preparation was further purified by chromatography on Sephacryl S-300 HR (Pharmacia Biotech) and Phenyl-Sepharose CL-4B (Pharmacia Biotech) columns according to Hertwig et al. (1992). Recombinant catalases SaCAT-1 and HaCAT-1 were purified from total cell extracts by chromatography on Sephacryl S-300 HR and Phenyl-Sepharose CL-4B columns (Hertwig et al. 1992). For part of the preparations, the resulting protein fraction was further purified by chromatography on an immobilized metal ion affinity chromatography (IMAC) column (10 × 1 cm) according to Yang & De Pierre (1998), or chromatography on Phenyl-Sepharose was replaced by IMAC. The recombinant sunflower isozyme HNNCATA3 was partially purified by chromatography on Sephacryl S-300 HR and IMAC columns. Purity and identity of the recombinant catalases were examined by SDS–PAGE and immunoblotting. Usually, preparations of 50–70% purity were used.

Nucleic acid extraction, cDNA library construction and hybridization techniques

Standard molecular biological techniques were applied for the extraction of RNA from leaves, for the construction of cDNA libraries, the amplification of DNA fragments by PCR, the cloning of DNA fragments and for Southern analysis as described by Sambrook et al. (1989).


Samples of enzyme preparations were separated by electrophoresis in the presence of 0.4% (w/v) SDS on polyacrylamide slab gels consisting of a 10% (w/v) polyacrylamide resolving gel and a 5% (w/v) stacking gel in the buffer system of Laemmli (1970). Immunoblotting with rabbit antisera against catalase from rye leaves or against a recombinant H. alpina leaf catalase HaCAT-1 produced by heterologous expression in E. coli was performed as previously described (Hertwig et al. 1992).

Analytical methods

Catalase (EC was assayed spectrophotometrically at 240 nm according to the method described by Lück (1965). For the estimation of Km values, the exact H2O2 concentrations of the assays were determined spectrophotometrically. Km values were determined by linear regression from Lineweaver-Burk plots.

The peroxidatic activity of catalase was determined as oxidation of ethanol at a low steady state H2O2 concentration by the reduction of NAD in the presence of aldehyde dehydrogenase according to Oshino, Oshino & Chance (1973). The 1 mL assay contained 1 or 0.1 µkat catalase, 0.1 mm NAD, 50 mm ethanol, 0.5 units aldehyde dehydrogenase (from baker's yeast), 10 mm glucose and 0.5 m K-phosphate, pH 7.5. The reaction was started by adding 0.6 units glucose oxidase (from Aspergillus niger). The reduction of NAD was followed at 340 nm. The optimal H2O2 concentration was determined by varying the quantity of glucose oxidase. Protein content was estimated by the BCA-test according to Smith et al. (1985) or by the RotiR-Nanoquant kit (Carl Roth GmbH, Karlsruhe, Germany), a modification of the assay described by Bradford (1976).

For the estimation of specific activities, catalase preparations were separated by 10% (w/v) SDS–PAGE to remove contaminating proteins. The protein content of the catalase band was estimated by comparison with a standard concentration series of bovine serum albumin.

Experimental treatments

For the photoinactivation experiments, solutions of the catalase preparations (c. 1 µkat mL−1) were incubated in a water bath at 25 °C in white light of 800 µmol m−2 s−1 PAR, provided by Osram halogen lamps (75 W). Heat was absorbed by a Schott filter KG 3 (Mainz, Germany).

The inactivation of catalase in the presence of a superoxide generating system was investigated by a modification of the assay described by Kono & Fridovich (1982). In a total volume of 1 mL, 1 µkat catalase activity was incubated together with 30 µm acetaldehyde and 1 unit xanthine oxidase (from milk) in 50 mm K-phosphate, pH 7.5, and 0.1 mm NaEDTA at 25 °C in darkness. For controls, xanthine oxidase was omitted.

To analyse a potential binding of NADPH to catalases, 0.25–1 mg of catalase was incubated overnight at 4 °C in a total volume of 1 mL 25 mm K-phosphate, pH 7.5, with 1 mm NADPH and 250 µg Pefa Block. Subsequently, the mixture was separated by chromatography on a Sephacryl S-300 HR column (1.5 × 100 cm or 1 × 30 cm) and eluted with 25 mm K-phosphate, pH 7.5. Fractions were assayed for catalase activity, and the fluorescence at 455 nm after excitation at 340 nm was determined.


Cloning and heterologous expression of plant catalases

In order to compare the structural and functional properties of plant catalases, several cDNA clones were obtained: Sccat1a for the light-sensitive catalase of rye leaves (Schmidt & Feierabend 2000), Sacat1 for a catalase from S. alpina leaves which contain a light-sensitive activity (Streb et al. 1997) and Hacat1 for a catalase from H. alpina leaves which contain a light-stable activity (Streb et al. 1997). In Fig. 1, The deduced amino acid sequences are compared with those of other plant catalases and with that of the mammalian BLC, which is known to be light-sensitive (Cheng et al. 1981). A database search indicated that catalase polypeptides designated as HNNCATA3 and HNNCATA4, which are enriched in the peroxisomal cores of sunflower leaves (Heinze & Gerhardt 2002), were highly homologous (95.3% identity between HaCAT-1 and HNNCATA3) to HaCAT-1 of H. alpina. Interestingly, Eising et al. (1998) reported that catalases of the peroxisomal cores of sunflower were much less light-sensitive than the soluble catalases of the peroxisomal matrix (containing the closely related polypeptides HNNCATA1 and HNNCATA2). The identity between the core catalase HNNCATA3 and the peroxisomal matrix catalase HNNCATA1 from sunflower was only 81.3%. Therefore, a cDNA clone for the putative light-stable core catalase HNNCATA3 was produced and was also included in our comparative investigation.

The amino acid sequences of the putative light-stable catalases HaCAT-1 and HNNCATA3 showed high homologies to other catalases, particularly to those from dicotyledonous plants (Fig. 1). The 78.9% identity between HaCAT-1 and SaCAT-1 of S. alpina was, for instance, in a similar range as the homology between core catalases (HNNCATA3) and matrix catalases (HNNCATA1) of sunflower but was lower than that between HaCAT-1 and HNNCATA3.

The sequences of the putative light-stable catalases were distinguished from all other catalases shown in Fig. 1 by only a few unique amino acid substitutions which occurred in both HaCAT-1 and HNNCATA3. These were Thr124 instead of a Val (BLC: Val133), Ile135 instead of a Leu (BLC: Leu144), Ser206 instead of a Gly (BLC: Gly215), Thr225 instead of a His (BLC: His234) and Met291 instead of a Lys (BLC: Lys300). A further remarkable difference is that both HaCAT-1 and HNNCATA3 contain a Trp at position 189 which is occupied by Leu (BLC: Leu198) in most other catalases. Most substitutions are to be regarded as similar substitutions. The most remarkable substitution is that of a His by Thr225, because this His is conserved in all known eukaryotic catalases, except for HaCAT-1 and the core catalases of sunflower.

The striking homology of HaCAT-1 to the core catalases of sunflower strongly suggests that HaCAT-1 might also be localized in peroxisomal cores in the leaves of H. alpina, particularly because these presumably also express a second conventional catalase isozyme most closely related to the sunflower matrix catalases. In addition to the Hacat1 cDNA, a 533 bp fragment (not shown) of an apparently more rare catalase cDNA from H. alpina leaves was identified. This catalase cDNAcoded for a second catalase polypeptide with highest homology (94.3%) to the light-sensitive matrix catalase HNNCATA1 of sunflower. It is therefore likely that, similarly as HNNCATA3, the closely related isozyme HaCAT-1 is also localized in the cores of the peroxisomes of H. alpina leaves. Electron microscopy of sections from H. alpina leaves confirmed that marked crystalline cores occurred indeed in the majority of the peroxisomes (Fig. 2).

Figure 2.

Electron micrograph showing a peroxisome with crystalline core in a section of a Homogyne alpina leaf. Chl, chloroplast; V, vacuole. Scale bar represents 0.5 µm.

The plant catalases were expressed as active enzymes in insect cell cultures after infection with recombinant baculoviruses containing the catalase sequences. The presence of the appropriate catalase cDNA sequences in the infected insect cells was verified by Southern analysis of total DNA extracted from the cultures (Fig. 3a). Significant expression of the recombinant plant catalases was only observed when external hemin was added to the cultures. After infection with recombinant baculovirus and supplementation with hemin, the insect cells had catalase activities of 3–6 µkat mg−1 cells for ScCAT-1 or around 1 µkat mg−1 cells for HaCAT-1. The internal catalase activity of the insect cells was very low. It accounted for less than 1% of the recombinant catalases. Unfortunately, catalase constructs carrying a His-tag at either their 5′- or their 3′-end were virtually not expressed by the insect cells. Therefore, conventional purification procedures had to be applied (see Materials and methods). The production of recombinant plant catalases was verified by Western blotting with antibodies against catalase from rye leaves (Fig. 3b) or against a recombinant catalase HaCAT-1 (Fig. 3c). Extracts of the Spodoptera cells did not react with the antibodies. The antiserum against rye catalase detected all recombinant plant catalases, although its reaction with SaCAT-1 or HaCAT-1 was weaker. The antiserum against the recombinant H. alpina catalase detected the recombinant catalases of the dicotyledonous plants (HaCAT-1, SaCAT-1, HNNCATA3) but did not cross-react with the recombinant ScCAT-1. The recombinant enzyme ScCAT-1 had the same apparent molecular mass of 57 kDa as the authentic catalase from rye leaves. The apparent molecular masses of recombinant SaCAT-1, HaCAT-1 and HNNCATA3 were lower. In the SDS–PAGE system applied, the apparent molecular mass of the recombinant HNNCATA3 produced in insect cells was approximately 55.0 kDa. We did not observe the higher apparent molecular mass of 59 kDa that was reported previously(Tenberge et al. 1997; Heinze & Gerhardt 2002).

Figure 3.

Heterologous expression of plant catalases in insect cells. (a) Southern blot analysis after separation of total DNA from insect cells after infection with recombinant baculovirus containing sequences for leaf catalases from Secale cereale (AcSccat1a), Homogyne alpina (AcHacat1), Soldanella alpina (AcSacat1) or Helianthus annuus (AcHnncatA3). Corresponding cDNA or DNA probes were used for hybridization. Sacat1VL1393, pVL1393 vector containing the Sacat1 sequence after cleavage with EcoRI and hybridization with a Sacat1 cDNA probe. AcpVL1393, recombinant baculovirus containing pVL1393 without insert after hybridization with a pVL1393 probe. (b) Western blot analysis with an antiserum against catalase from rye leaves of purified catalase from S. cereale leaves (S.c. leaf), soluble matrix catalase (HNN Sol.) or peroxisomal core (HNN Core) catalase from H. annuus cotyledons and of partially purified recombinant catalases from S. alpina (rSaCAT-1), S. cereale (rScCAT-1), H. alpina (rHACAT-1) or H. annuus (rHNNCATA3). Recombinant catalases were produced in baculovirus-infected insect cells. Control, extract from insect cells infected with baculovirus containing pVL1393 without insert. (c) Western blot analysis with an antiserum against the recombinant catalase HaCAT-1 of peroxisomal core catalase (HNN Core) and soluble matrix catalase (HNN Sol.) from H. annuus cotyledons and of recombinant catalases from S. cereale (rScCAT-1), H. annuus (rHNNCATA3) and H. alpina (rHACAT-1). Control, extract from insect cells without baculovirus infection.

Kinetic and catalytic properties

Kinetic and catalytic properties of the recombinant plant catalases ScCAT-1, SaCAT-1 and HaCAT-1 are compared in Tables 1 and 2. The apparent Km value estimated for the putative light-insensitive H. alpina-catalase was in the same range as the apparent Km values observed for the light-sensitive catalases from S. cereale or S. alpina. HaCAT-1 was more sensitive to inhibition by 3-aminotriazole (Margoliash, Novogradsky & Schejter 1960) than the recombinant catalase from S. cereale or than BLC, which was assayed for comparison. The sensitivity of recombinant HaCAT-1 to inhibition by KCN or NaN3 did not differ substantially from that of the other catalases (Table 2). The peroxidatic activity of the recombinant catalases from both H. alpina and S. cereale was relatively low. The apparent peroxidatic activity determined for the catalase CAT-1 purified from rye leaves was even lower than that of the recombinant ScCAT-1. The most striking property of the recombinant HaCAT-1 was that its specific activity was low. It accounted for only 16% of that observed for the recombinant ScCAT-1 (Table 1).

Table 1.  Comparison of kinetic constants and peroxidatic activity (as percentage of catalatic activity) of catalase from rye leaves and of recombinant catalases of Secale cereale, Soldanella alpina and Homogyne alpina
Source of catalaseKm (H2O2) (mM)Specific activity (µmol s−1 mg−1)Peroxidatic activity (%)
  1. Specific activity is indicated as µmol·s−1 (mg protein)−1.

  2. Standard errors of the mean are indicated.

  3. Due to the catalatic reaction mechanism, data for Km represent apparent Km values. Peroxidatic activities were corrected for contamination by insect cell protein.

  4. n.d., not determined.

S. cereale, leafn.d.n.d.0.002 ± 0.0003
S. cereale, recombinant21.1.106.2 ± 10.10.071 ± 0.0045
S. alpina, recombinant42.0n.d.n.d.
H. alpina, recombinant33.117.1 ± 1.70.038 ± 0.0029
Table 2.  Effect of inhibitors on the activity of recombinant catalases of Secale cereale and Homogyne alpina and of BLC
Inhibitor ConcentrationRecombinant S. cereale catalase (%)Recombinant H. alpina catalase (%)Recombinant BLC (%)
  1. Residual activities (%) observed after 5 min pre-incubation with 3-aminotriazole or KCN or after 30 min pre-incubation with NaN3 are indicated.Standard errors of the mean are indicated.

8.5 mm 3-Amino-1,2,4-triazole42.0 ± 8.519.3 ± 0.972.2 ± 3.4
0.1 mm KCN36.1 ± 2.230.2 ± 2.2 9.1 ± 0.5
1.7 mm KCN 4.2 ± 0.4 6.1 ± 1.4 5.4 ± 0.4
0.1 mm NaN3 0.013.0 ± 0.01 0.0

Differential sensitivity of catalases to inactivation by light and superoxide

When exposed to white light of 800 µmol m−2 s−1 PAR, recombinant catalases of S. cereale or of the alpine plant S. alpina were inactivated and exhibited the usual light-sensitivity of natural catalases. The recombinant catalase HaCAT-1 was, however, not inactivated by light. These results confirm that a light-insensitive catalase was cloned from H. alpina leaves (Fig. 4a). The recombinant sunflower isozyme HNNCATA3 of H. annuus was as light-insensitive as HaCAT-1 (Fig. 4b). Photoinactivation of the native catalase activity of peroxisomal cores, which were shown to contain HNNCATA3 (Tenberge et al. 1997; Heinze & Gerhardt 2002), was also only half as fast as that of the soluble catalase from H. annuus leaves. However, catalase of natural peroxisomal cores was markedly more light-sensitive than the recombinant homotetrameric HNNCATA3 (Fig. 4b).

Figure 4.

Changes of the activities of isolated catalases during in vitro incubation in white light (800 µmol m−2 s−1 PAR). (a) recombinant catalases from S. cereale (○), S. alpina (□) and H. alpina (▵). Closed symbols, control incubations in darkness. (b) recombinant catalase HNNCATA3 from H. annuus (▵), soluble catalase from H. annuus (○), catalase of purified peroxisomal cores from H. annuus cotyledons (□). Closed symbols, control incubations in darkness. Incubations were performed at 25 °C.

Because O2 and possibly some reactive oxygen are involved in the inactivation by light, it was of interest whether the light-insensitive catalase was equally resistant to inactivation by superoxide. Xanthine oxidase with acetaldehyde as substrate was used as superoxide generating system, as in previous investigations with BLC (Kono & Fridovich 1982). The superoxide-induced inactivation of BLC was efficiently prevented by the peroxidatic substrate ethanol or by NADPH (Fig. 5), as previously described (Kono & Fridovich 1982; Kirkman et al. 1999). The recombinant light-stable catalase HaCAT-1 suffered from a rapid immediate inhibition by superoxide. During further incubation, it was much more strongly inactivated than the light-sensitive BLC. In contrast to BLC, HaCAT-1 was also inhibited by acetaldehyde in the absence of xanthine oxidase (almost 30% within 2 h). The superoxide-induced inactivation of HaCAT-1 was retarded by ethanol but was not affected by NADPH (Fig. 5). Notably, the sensitivity of the natural and recombinant light-sensitive leaf catalases of S. cerealeTo superoxide differed strikingly (Fig. 5). Because their primary structures were identical, differential post-transcriptional modifications might play some role. While the natural catalase from rye leaves was only slightly inhibited by superoxide, the recombinant ScCAT-1 was strongly inactivated. In contrast to the behaviour of BLC, this inactivation was not prevented by ethanol or NADPH and appeared to be caused by a different mechanism.

Figure 5.

Changes of catalase activities in the presence of a superoxide generating system. ●, complete system with xanthine oxidase and acetaldehyde as substrate; □, control without xanthine oxidase; ▵, complete system + 1% ethanol; ▿, complete system + NADPH (0.75 mm for BLC and HaCAT-1; 1 mm for catalases from S. cereale). Incubations were performed at 25 °C with BLC, recombinant catalase from Homogyne alpina (HaCAT-1), catalase from leaves of Secale cereale (ScCAT-1) and recombinant ScCAT-1.

While the peroxidatic substrate ethanol totally prevented the photoinactivation of BLC and of both the recombinant ScCAT-1 and the natural catalase from rye leaves, NADPH did not protect plant catalases (Fig. 6). Photoinactivation of BLC was partially retarded by NADPH. The results suggested that NADPH did not mitigate the light- or superoxide-induced inactivation of the plant catalases because they were unable to bind this cofactor, although Durner & Klessig (1996) claimed that catalase from tobacco leaves could bind NADPH. When a solution of BLC was separated by chromatography on Sephacryl S-300 HR after incubation with NADPH, the catalase fractions retained bound NADPH, as monitored by its fluorescence at 455 mn, but the unbound NADPH was clearly separated. However, neither the recombinant plant catalase ScCAT-1 (Fig. 7) nor the natural catalase from S. secale leaves (not shown) retained bound NADPH after chromatography on Sephacryl S-300 HR columns. Similarly, the recombinant HaCAT-1 did not bind NADPH (not shown).

Figure 6.

Comparison of the effects of ethanol and NADPH on the photoinactivation of catalases in vitro. BLC, catalase ScCAT-1 from leaves of Secale cereale and recombinant ScCAT-1 from S. cereale were incubated at 25 °C and 800 µmol m−2 s−1 PAR without further addition (○), with the addition of 1% ethanol (▵) or with 1 mm NADPH (□).

Figure 7.

Comparison of the binding of NADPH to BLC and to recombinant ScCAT-1 catalase from Secale cereale. After 24 h incubation with 1 mm NADPH, catalases were separated from unbound NADPH by chromatography on Sephacryl S-300 H columns. NADPH was detected by its fluorescence at 455 nm after excitation at 340 nm.

Mutagenesis of the light-sensitive catalase from S. alpina

Among the amino acid substitutions that appeared to be specific for the light-insensitive catalases HaCAT-1 and HNNCATA3, the replacement of a His by Thr225 (Fig. 1) was the most exceptional substitution because this position is strictly conserved in other eukaryotic catalases. In order to investigate whether the light sensitivity of catalases was related to the presence of this His, His225 was replaced by Thr through in vitro mutagenesis (see Materials and methods) in the light-sensitive catalase SaCAT-1 of S. alpina. In Western blots, the recombinant mutant SaCAT-1H225T was detected at the same position as the recombinant wild type SaCAT-1 (Fig. 8a). However, lower quantities of the mutant than of the wild-type enzyme protein were produced by the insect cells per equal cell mass. When a solution of recombinant SaCAT-1H225T was exposed to light, it was much more rapidly inactivated than the recombinant wild-type enzyme (Fig. 8b). The rate of inactivation during the first hour was increased by a factor of 2.8. A double mutant SaCAT-1H225T, K291M was expressed only in minute amounts by the insect cells and could therefore not be further analysed.

Figure 8.

Properties of the H225T mutant of the light-sensitive catalase SaCAT-1 of Soldanella alpina. (a) Heterologous expression of the recombinant catalase from S. alpina (rSaCAT-1) and its H225T mutant (rSaCAT-1H225T) in insect cells. Catalase polypeptides were detected in Western blots after SDS–PAGE with an antiserum against catalase from Secale cereale leaves. The recombinant catalase from S. cereale (rScCAT-1) is shown for comparison. (b) Changes of the activities of the recombinant catalase SaCAT-1 from S. alpina and of its mutant form SaCAT-1H225T during in vitro incubation at 25 °C in white light (800 µmol m−2 s−1 PAR). □, recombinant SaCAT-1; ○, recombinant SaCAT-1H225T. Closed symbols, control incubations in darkness.


Identification and occurrence of light-sensitive catalases

By cDNA cloning and heterologous expression of recombinant enzymes in insect cell cultures, the occurrence of light-insensitive catalases in the leaves of the highly light stress-resistant alpine plant H. alpina and of H. annuus was verified. The application of insect cell cultures supplemented with additional heme was most crucial because only these allowed the expression and characterization of active recombinant plant catalases. Although E. coli was frequently used for the expression of recombinant active catalases from other sources (e.g. Switala & Loewen 2002), it produced only inactive protein of plant catalases. To our knowledge, the production of active recombinant plant catalases has not been reported previously except in transgenic plants (e.g. Polidoros, Mylona & Scandalios 2001).

In sunflower leaves, the light-insensitive catalase isozymes, such as HNNCATA3 (which is 97.4% identical with HNNCATA4), are localized in the cores, the crystalline inclusions of the peroxisomes (Eising & Gerhardt 1986; Heinze & Gerhardt 2002). Because of the strikingly high identity of 95.3% of the sequences of HaCAT-1 of H. alpina and the peroxisomal core catalase HNNCATA3 of sunflower, it is quite likely that HaCAT-1 is also localized in the strong crystalline cores that were observed in the majority of the peroxisomes in H. alpina leaves. This is further supported by the fact that H. alpina and H. annuus belong to the same plant family of Asteraceae and contain an additional conventional catalase isozyme assigned to the peroxisomal matrix in sunflower leaves. The light-stable catalases of these Asteraceae might serve as a final reserve for extreme photooxidative stress conditions, and they seem to be particularly abundant in stress-adapted plants such as H. alpina.

Functional properties of light-insensitive catalase

In accord with the high general homology between the primary structures of the light-insensitive catalases HaCAT-1 and HNNCATA3 and all other catalases (Fig. 1), basic kinetic and catalytic properties, such as apparent Km or inhibitor sensitivities, of HaCAT-1 did not differ greatly from those of light-sensitive catalases (Tables 1 & 2). Furthermore, the peroxidatic activity, which was found to be enhanced in some catalase isozymes (Havir & McHale 1989), was low. The most marked difference was the low specific activity of HaCAT-1 (Table 1). Similarly, the specific activity of the core isozymes of sunflower catalases accounted for only 10% of that of the light-sensitive matrix isozymes (Eising & Gerhardt 1986; Heinze & Gerhardt 2002).

The observations that the inactivation of catalase by both light and O2·−were similarly suppressed by the peroxidatic substrate ethanol and that both appeared to depend on a formation of the reaction intermediates compound II or compound III (Cheng et al. 1981; Shimizu et al. 1984; Kirkman et al. 1999; Aubailly et al. 2000) suggested that both types of inactivation might follow identical or similar mechanisms. In light, O2·−might be formed as the effective injurious agent, as proposed for the inactivation of catalase by UV (Aubailly et al. 2000). However, the results of the present investigation that the light-insensitive HaCAT-1 was nevertheless strongly inactivated by O2·−(Fig. 5) while the native light-sensitive catalase from rye leaves was hardly affected by O2·−speak against the possibility that a common mechanism is mediating the inactivation of catalase by light or O2·−.

Molecular properties of light-insensitive catalase

Although they originate from different plant species, the light-insensitive catalases HaCAT-1 and HNNCATA3 are most closely related to each other by a remarkable 95.3% identity of their primary structures. Nevertheless, they are also highly homologous to all other catalases. Only six amino acid substitutions can be regarded as exceptional relative to the other catalases, and all residues that are essential for heme binding and catalytic activity (Fita & Rossmann 1985; Gouet, Jouve & Dideberg 1995; Zámocký & Koller 1999; Nicholls, Fita & Loewen 2001) are strictly conserved (Fig. 1) Furthermore, only those few unique amino acid substitutions that occur in otherwise strictly conserved positions in both HaCAT-1 and HNNCATA3 can be expected to contribute to the catalases’ extraordinary property of resistance to photoinactivation. In order to understand the potential significance of these specific amino acid substitutions, their location within the structural organization of the catalase molecule has been investigated. Crystal structures for plant catalases are not yet available. However, crystal structures that were determined for bacterial, yeast and mammalian catalases (Fita & Rossmann 1985; Gouet et al. 1995; Zámocký & Koller 1999; Nicholls et al. 2001) indicated that their three-dimensional structures are highly conserved, including that of the clade I catalase CatF of Pseudomonas syringae (Carpena et al. 2003). Plant catalases also belong to clade I of monofunctional heme-containing catalases (Nicholls et al. 2001). Based on this general structural similarity and on known X-ray structures, a model of the putative tertiary structure of the light-insensitive catalase HaCAT-1 was calculated and the putative locations of the specific amino acid substitutions were marked (Fig. 9).

Figure 9.

Model of the putative 3-dimensional structure of the light-insensitive catalase HaCAT-1 from Homogyne alpina leaves. Based the structural similarities and known crystal structures of catalases, the model was calculated by the Swiss-PdbViewer for the partial HaCAT-1 sequence between residues Pro15 and Leu454. Positions shown to be associated with the minor heme channel of BLC (Fita & Rossmann 1985) are marked in yellow. Unusual amino acid substitutions in the light-insensitive catalases HaCAT-1 and HNNCATA3, relative to light-sensitive catalases, are marked in green. The heme iron ligand Tyr348 is marked in red.

The model suggests that the amino acid substitutions that are specific for the light-stable catalases are localized in a restricted area of the molecule. All six unusual amino acid substitutions occur within the central antiparallel β-barrel domain comprising strands β1–β8 (Zámocký & Koller 1999; Nicholls et al. 2001). Thr124 (on strand β3), Ile134 (on strand β4) and Ser206 (between helix α5 and strand β5) are localized close to the central heme cavity. Thr124 and Ile135 are localized adjacent to presumptive proximal heme-binding domains, Ser206 is located within a presumptive proximal heme-binding domain (Gouet et al. 1995). The most exceptional amino acid substitutions of the light-stable catalases – Thr225 on strand β6, Met291 between strands β7 and β8 and Trp189 on helix α5 – are all localized at or near the surface of the molecule in a groove that binds NADPH in several catalases (e.g. BLC, yeast or Proteus mirabilis catalases) at the entrance of a minor lateral channel leading to the distal side of the central heme. Presumptive amino acids of the minor channel (according to Fita & Rossmann 1985) and the proximal heme-iron ligand Tyr348 are marked in Fig. 9. The functional significance of the minor channel is still under investigation. In the catalase HPII of E. coli, a mutation causing an enlargement of the lateral channel resulted in a marked increase of its specific activity (Sevinc et al. 1999). Conceivably, the remarkably low specific activity of the light-insensitive catalases HaCAT-1 and HNNCATA3 is therefore related to structural changes of the lateral channel restricting its accessibility. Post-translational modifications enabled by the specific amino acid substitutions of the light-sensitive catalases cannot be excluded. However, only Thr225, Met291 and Trp189 appear to be exposed to the surface. It is unlikely that the unique amino acid substitutions in the light-insensitive catalases affect the stability of the tetrameric quaternary structure of the enzyme. This is because the domains that are responsible for the interaction between subunits, the amino-terminal arm and the so-called wrapping loop, did not contain any conspicuous amino acid substitutions (Fig. 9).

The region of the peripheral cavity at the entrance of the minor channel which serves as binding site for NADPH in various catalases was previously shown to be quite essential for the stability of catalases (Zámocký & Koller 1999). Binding of NADPH further increased the enzyme's general stability and protected it from inactivation by O2·− (Kirkman et al. 1999; Zámocký & Koller 1999). While Durner & Klessig (1996) reported that tobacco catalase binds NADPH, several other plant catalases examined in this investigation did not bind NADPH. Similarly, Beaumont et al. (1990) found that catalase from potato tubers did not bind NADPH. Furthermore, amino acids that are involved in the binding of NADPH are not conserved in plant catalases. A most crucial residue for NADPH binding is His304 of BLC, which may be replaced by Gln in weakly NADPH binding catalases such as yeast catalase A (Fita & Rossmann 1985; Gouet et al. 1995; Zámocký & Koller 1999). However, all known plant catalases contain a Glu instead of His at this position which is considered unsuitable for NADPH binding (Carpena et al. 2003). In HaCAT-1, the Glu is replaced by Asp at position 295 (Fig. 1). Therefore, NADPH clearly does not represent a cofactor for plant catalases and is consequently unable to protect the plant enzyme from inactivation.

Even when it does not represent an NADPH binding site, the peripheral cavity at the entrance of the minor channel seems to be important for the stability of catalase (Zámocký & Koller 1999). For the light-insensitive catalases, the occurrence of the three most exceptional amino acid substitutions (Trp189, Thr225 and Met291) in this region further accentuates a special, though yet undefined, role for the catalases’ stability and resistance to photodamage. Most conspicuous is the substitution of His (234 in BLC) by Thr225. The observation of Nakatani (1961) that some His is oxidized during photooxidation of catalase raised the question of whether His234 (in BLC) was a main target of oxidative inactivation and whether its replacement by Thr might improve the stability of light-sensitive catalases. However, the replacement of the corresponding His225 by Thr in the light-sensitive catalase SaCAT-1 by in vitro mutagenesis even greatly enhanced the photooxidative inactivation of the enzyme. These results further emphasize that the structure at the entrance of the minor channel must be of crucial importance for the stability of catalases. However, the substitution of one single amino acid is obviously not sufficient to increase the catalases’ light-resistance, unless it is complemented by other accompanying changes.


Financial support by the Deutsche Forschungsgemeinschaft, Bonn, is greatly appreciated. We are very grateful to Prof. R. Bligny for his great hospitality and support at the Station Alpine du Lautaret of the Université Joseph Fourier of Grenoble, France, and to Dr H. Reiländer, Max-Planck-Institut für Biophysik, Frankfurt am Main, for introducing Nicole Engel into handling of baculovirus and insect cells. The technical assistance of Mrs Christel van Oijen is also greatly appreciated.