Tài liệu Báo cáo Y học: Structural diversity and transcription of class III peroxidases from Arabidopsis thaliana docx

Structural diversity and transcription of class III peroxidases from
Arabidopsis thaliana
Karen G. Welinder
1,2
, Annemarie F. Justesen
1
, Inger V. H. Kjærsga
˚
rd
1
, Rikke B. Jensen
1
,
Søren K. Rasmussen
3
, Hans M. Jespersen
1
and Laurent Duroux
2
1
Department of Protein Chemistry, University of Copenhagen, Denmark;
2
Department of Biotechnology, Aalborg University,
Denmark;
3
Plant Genetics, Risø National Laboratory, Denmark
Understanding peroxidase function in plants is complicated
by the lack of substrate specificity, the high number of genes,
their diversity in structure and our limited knowledge of
peroxidase gene transcription and translation. In the present
study we sequenced expressed sequence tags (ESTs) enco-

ding novel heme-containing class III peroxidases from
Arabidopsis thaliana and annotated 73 full-length genes
identified in the genome. In total, transcripts of 58 of these
genes have now been observed. The expression of individual
peroxidase genes was assessed in organ-specific EST libraries
and compared to the expression of 33 peroxidase genes
which we analyzed in whole plants 3, 6, 15, 35 and 59 days
after sowing. Expression was assessed in root, rosette leaf,
stem, cauline leaf, flower bud and cell culture tissues using
the gene-specific and highly sensitive reverse transcriptase-
polymerase chain reaction (RT-PCR).We predicted that 71
genes could yield stable proteins folded similarly to horse-
radish peroxidase (HRP). The putative mature peroxidases
derived from these genes showed 28–94% amino acid
sequence identity and were all targeted to the endoplasmic
reticulum by N-terminal signal peptides. In 20 peroxidases
these signal peptides were followed by various N-terminal
extensions of unknown function which are not present in
HRP. Ten peroxidases showed a C-terminal extension
indicating vacuolar targeting. We found that the majority of
peroxidase genes were expressed in root. In total, class III
peroxidases accounted for an impressive 2.2% of root ESTs.
Rather few peroxidases showed organ specificity. Most
importantly, genes expressed constitutively in all organs and
genes with a preference for root represented structurally
diverse peroxidases (< 70% sequence identity). Further-
more, genes appearing in tandem showed distinct express-
ion profiles. The alignment of 73 Arabidopsis peroxidase
sequences provides an easy access to the identification of
orthologous peroxidases in other plant species and will

provide a common platform for combining knowledge of
peroxidase structure and function relationships obtained in
various species.
Keywords: EST; expression analysis by RT-PCR; peroxi-
dase gene annotation; peroxidase structure; propeptides.
Peroxidase enzymes have challenged chemists and biologists
for more than 70 years and have been used in a great
number of analytical applications [1]. The majority of
peroxidases contain an extractable heme (Fe
3+
protopor-
phyrin IX) center, whereas others contain a cytochrome c
type heme, a selenium center or a vanadium center.
Peroxidases react first with a peroxide to yield highly
oxidizing intermediates with redox potentials up to
1000 mV and thereafter with a variety of organic or
inorganic reducing substrates, which are often oxidized to
form radicals. Peroxidase activity was detected early in
horseradish roots (reviewed in [1]), which is still the major
source of commercial heme peroxidases. In addition,
peroxidases have been isolated from a variety of plant,
animal, fungal and bacterial sources. The bacterium
Escherichia coli expresses a single intracellular heme peroxi-
dase with dual catalase–peroxidase activities [2], a finding
confirmed by its genome sequence [3]. Mitochondrial yeast
cytochrome c peroxidase, chloroplast and cytosol plant
ascorbate peroxidases are rather similar in amino acid
sequence to the bacterial enzymes, and they are collectively
referred to as class I peroxidases [4]. These intracellular
peroxidases appear to function as protective peroxide

scavengers and they constitute in plants a small family of
7–10 genes, encoding both soluble and membrane bound
enzymes [5]. White-rot fungi like Phanerochaete chrysospo-
rium and Trametes versicolor contain a small gene family
encoding approximately 10 different lignin-degrading or
Mn-dependent heme peroxidases. In contrast, the ink cap
fungus Coprinus cinereus contains only a single peroxidase
gene [6,7]. The extracellular fungal peroxidases (class II) can
participate in secondary metabolism under conditions of
limited nutritional supply [8]. The classical plant peroxidases
(class III) are targeted via the endoplasmic reticulum (ER)
to the outside of the plant cell or to the vacuole. They are
Correspondence to K. G. Welinder, Department of Biotechnology,
Aalborg University, Sohngaardsholmsvej 49, DK-9000 Aalborg,
Denmark. Fax: + 45 98141808, Tel.: + 45 96358467,
E-mail:
Abbreviations: AtP, transcribed A. thaliana (class III) peroxidase;
BP, barley peroxidase; dbEST, database of ESTs; ef-1a, elongation
factor-1a; EST, expressed sequence tag; HRP, horseradish
peroxidase; SBP, soybean peroxidase; TC, tentative consensus.
Notes: Equal contributions were made to this work by A. F. J., L. D.
and H. M. J. The GenBank accession numbers for the nucleotide
sequence data produced are listed in Table 1.
(Received 19 August 2002, revised 8 October 2002,
accepted 15 October 2002)
Eur. J. Biochem. 269, 6063–6081 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03311.x
ascribed a variety of functional roles in plant biology, which
include lignification, suberization, auxin catabolism, def-
ense, stress and developmentally related processes (reviewed
in [9,10]).

Prior to the present study it was known that
horseradish contained at least nine different genes for
class III peroxidases [11]. With this background, it seemed
ideal to study the entire repertory of plant peroxidase
genes in the model plant Arabidopsis thaliana,which
belongs to the same botanical family, taking advantage of
the expressed sequence tag (EST) sequencing programs in
progress [12–14], as well as the results of the Arabidopsis
genomic sequencing project [15]. Here we report the
complete sequencing and mRNA expression analyses of
class III Arabidopsis peroxidase transcripts mostly
obtained from the EST projects, and the predicted
protein structures derived from all 73 Arabidopsis peroxi-
dase genes [16].
MATERIALS AND METHODS
DNA sequencing and gene annotation
BLAST
and Entrez services at the National Center for
Biotechnology Information ()
[17,18] were used to search databases (nonredundant and
dbEST). EST clones were obtained from the Arabidopsis
Biological Resource Center, Ohio State University [12,13],
Genome Systems (Genome Systems Inc, St Louis, USA),
and the Kasuza Institute [14]. Plasmid DNA purification
and sequencing were performed as described previously [19]
and both strands were sequenced.
Genes encoding class III peroxidases in Arabidopsis
were searched for in the Munich Information Center for
Protein Sequences (MIPS) [20] and The Institute for
Genomic Research (TIGR) [21] annotated databases using

the keyword ÔperoxidaseÕ. Lists of genes were extracted
and those coding for class I peroxidases (ascorbate
peroxidases), glutathione peroxidases and catalases were
removed, leaving a set of 75 nonredundant acces-
sions. Predictions of intron splice-sites were done with
NETPLANTGENE
[22] ( />Putative transcriptional start sites and TATA-like boxes
were mapped in the 5¢-UTR with the eukaryotic neural net-
work promoter prediction server at itfly.
org/seq_tools/promoter.html, using human and fruit-fly
data. Predicted results were compared with known
5¢-UTRs from publicly available cDNA sequences. Nuc-
leotide compositions of the 5¢-UTRs were computed as
described in [23].
Protein sequence alignment
Amino acid sequences were derived from the coding regions
of the expressed genes using the program
NETSTART
for
plants [24] ( for
predicting initiating Met. The N-terminal signal peptides
were predicted with the
SIGNALP
program [25] (http://
www.cbs.dtu.dk/services/SignalP-2.0/) and checked with the
TARGETP
program [26] ( />TargetP/). The alignments were performed with the
CLUSTALX
program [27] using the
GONNET

substitution
matrices [28] on truncated sequences corresponding to
residues 1–305 of mature HRPC. A first alignment was
done with all sequences to obtain similarity clusters. An
improved alignment was built using the profile alignment
mode of
CLUSTALX
. First, a group of sequences highly
similar to horseradish peroxidase C (HRPC) was aligned
taking into account the secondary structure assignments for
HRPC (default settings in
CLUSTALX
). This group of aligned
sequences was then used as a core onto which clusters of
sequences were added sequentially. Finally, minor manual
adjustments were made to exclude an excessive number of
gaps.
In calculating the pairwise distances, the sequence length
was defined as all matched residues, not counting gaps.
Calculation of pairwise distances and isoelectric points
used only aligned full-length sequences, which were trun-
cated to start at the position corresponding to the
N-terminal pyroglutamate residue of mature HRPC, and
ending at the position corresponding to HRPC residue
N305 [29].
Plant material and RNA purification
A. thaliana seeds, ecotype Columbia were kindly provided
by F. Floto, and cell suspension culture by O. Mattsson,
both at the Department of Plant Physiology, University of
Copenhagen. Plants were grown in plastic containers on

Murashige and Skoog medium (catalog no. 2606, Betatech)
at 25 °C, 16 h light (3000 lux). Plants were harvested 3, 6,
15, 35 and 59 days after sowing. Plants older than 15 days
were dissected into roots, rosettes, cauline leaves, stems and
flower buds and the tissue was transferred immediately into
liquid nitrogen and ground in a mortar. Total RNA was
isolated using an RNeasy total RNA purification kit
(QIAGEN) according to the manufacturer’s instructions.
The quality of the RNA was evaluated by gel electrophor-
esis and by measuring A
260
/A
280
. Purified RNA was stored
at )80 °C.
RT-PCR analysis
The RT-PCR analyses were performed using the Perkin-
Elmer GeneAmpÒ RNA PCR kit. An oligo(d[T]
16
)
primer was used for the first strand synthesis. Primers
specific to each peroxidase gene were used for the second
strand synthesis and PCR amplification (Supplementary
material, Table S1). The specificity of each set of primers
was optimized using the corresponding cDNA clone.
Different combinations of annealing temperatures (60–
65 °C) and concentrations of MgCl
2
(1.0–2.0 m
M

)were
tested to find the optimal conditions at which the primers
were specific. When possible, the primers were designed to
anneal in the 5¢ sequence encoding the signal peptide or in
the 3¢-UTR. Primer sets were tested for specificity in a
PCR, performed on a mixture of cDNA clones encoding
all the peroxidases investigated, including and excluding
the clone encoding the peroxidase for which the primers
were designed. RT-PCR analyses were performed twice
for each peroxidase using two different reverse transcribed
reactions for each time point and organ. As a control of
the quality of the mRNA, RT-PCR was performed with
primers specific for the elongation factor-1a (ef-1a)[19].
The RT-PCR products were analyzed on a 1% (w/v)
agarose gel.
6064 K. G. Welinder et al.(Eur. J. Biochem. 269) Ó FEBS 2002
Digital expression analysis
Transcription profiles were inferred from peroxidase EST
counts, abstracted from TIGR A. thaliana Gene Index [30]
(AtGI release version 6, May 2001) using ÔperoxidaseÕ as a
keyword for the search. Each Tentative Consensus (TC)
accession was verified and assigned to a unique peroxidase
gene [15,20]. For each accession, the number of ESTs per
library was counted. EST libraries (TIGR codes indicated
by ¢#¢) were grouped according to organ: 1, root Columbia,
#5336 [14], root-1 and -2 Col0 Columbia, #2336 and #2337
(Genome Systems, Inc.); 2, seedling hypocotyl CD4-13, -14,
-15 and -16, #NH28, #NH25, #NH26 and #NH27 [12]; 3,
rosette-1, -2 and -3 Col0 Columbia, #2338, #2340 and
#2341 (Genome Systems, Inc.); 4, above-ground organs two

to six weeks-old, #4063, #5335 and #3792 [14], Ors-A green
shoot, #NH12, shoot 2-weeks old, #NH29; 5, flower bud
Columbia, #5337 [14], inflorescence-1 and -2 Col0 Colum-
bia, #2334 and #2335 (Genome Systems, Inc.), flower bud
Grenoble-A and -B, #NH08 and #NH09, inflorescence
young flower CD4-6, #NH36; 6, green silique Columbia,
#5339 [14], green silique GIF-Seed A, A + B and GIF-
Silique B, #NH05, #NH06 and #NH07, immature siliques,
#2369; 7, developing seeds, #5564 [31], early developing
seeds, #5576, germinating seed, #2370; 8, whole seedling
Versailles-VB, -VC and -VD, #NH18, #NH19 and #NH20;
9, various, consisting mainly of the mixed organs k-PRL2
library, #NH11 contributing 27 631 ESTs [12] as well as all
remaining EST libraries used in TCs by TIGR: #NH10,
#2339, #2342, #4924, #NH03, #NH39, #4921, #4932,
#5338, #NH02, #NH01, #NH13, #NH30, #6523, #6524,
#7052, #7053, #7054, #7055, #1725, #2373, #2741, #NH04,
#NH14, #NH15, #NH16, #NH17, #NH35, #NH44,
#NH31, #NH32, #NH34, #NH37, #NH38, #NH40,
#NH41, #NH43.
RESULTS AND DISCUSSION
cDNA and gene sequences
The total number of ESTs from Arabidopsis has recently
increased to 111 206, including 942 class III peroxidase
clones (TIGR release v 6.0), or 0.85% of the total. Genes
encoding class III peroxidases are easily identified by the
most conserved active site motif (Fig. 1), which is located
approximately 70 amino acids from the initiating Met
residue, or 210 nucleotides from the initiating AUG codon.
The selected clones were sequenced completely on both

strands and the putative peroxidases called AtP1 to AtP38.
The sequences have been deposited at GenBank or EMBL
databases under the accession numbers listed in Table 1.
Additional sequences of Arabidopsis peroxidase transcripts
were obtained from the literature and our own work,
AtPCa, -Cb, -Ea, -N, -A2, -RC (original names retained,
except for RCIIIa). Recent large-scale Arabidopsis cDNA
sequencing by the Riken Genomic Sciences Center, Yoko-
hama, Japan, and Ceres Inc., Malibu, California, has
currently brought the total of nonredundant peroxidase
transcripts up to 57, AtP39 to AtP51. These 57 transcripts
represent 58 genes, as two identical genes are represented by
AtP11 (Fig. 1; Table 1). The MIPS gene names are used for
the peroxidase genes for which no transcripts have been
observedsofar.
Analysis of the Arabidopsis genome [15] revealed a total
of 73 full-length class III peroxidase genes, two pseudo-
genes, and six fragments spread rather evenly on the five
Arabidopsis chromosomes[16;L.DurouxandK.G.
Welinder, unpublished observations]. Introns were localized
and their phase determined. Results are reported in Table 1,
and intron locations mapped to the protein sequences in
Fig. 1 (highlighted in reverse print). Introns 1, 2 and 3 are
predominant.
The peroxidase-encoding DNA sequences have been
analyzed thoroughly and annotated as in [23]. Table 1
provides an overview of all peroxidase genes and their
introns, the percentage adenine content of 5¢-UTRs,
predicted initiating Met, lengths of preproperoxidases and
ER-signal peptides, and calculated isoelectric points of the

putative mature polypeptides truncated to HRPC positions
1–305. The protein sequences predicted from the 73 genes
are aligned in Fig. 1 as a base for the comprehensive
structural characterization of the entire class III peroxidase
repertory of a flowering plant. Sites of initiating Met and
ER-signal cleavage were predicted using both hidden
Markov (scores reported in Table 1) and neural network
methods. Possible alternative sites are shown in Supple-
mentary material, Fig. S1. The nucleotide sequences, anno-
tation and percentage nucleotides of 5¢-UTRs of 73
peroxidase genes are given in the Supplementary material
accompanying this paper (Fig. S2, and Table S2).
Nucleotide differences have been observed between
similar cDNA clones, and between cDNA and the corres-
ponding gene. This can be ascribed to either allelic
variations or to different ecotypes despite the fact that all
were designated Columbia. Kjærsga
˚
rd et al. [19] described
Fig. 1. Alignment of the amino acid sequences of putative mature per-
oxidases predicted from the 73 class III Arabidop sis peroxidase genes.
The 58 transcribed genes are referred to by AtP# names; the rest by
MIPS gene numbers. The sequences are sorted according to similarity,
and peroxidases > 70% amino acid identity are boxed, alternating in
blue and grey. The Arabidopsis peroxidases are compared to horse-
radish peroxidase HRPC. The a-helices, A–J, observed in HRPC (top),
and residue or position numbers also refer to HRPC. Conserved res-
idues (bottom) include invariant (uppercase), and highly conserved
(lowercase). Active site residues are in red; side chain ligands to the
distal and proximal Ca

2+
ions are in blue; cysteine residues involved in
disulfide bridges 11–91, 44–49, 97–301 and 177–209 are in yellow; an
invariant ion-pair motif are on a grey background; and putative
N-glycosylated triplets are in green. Unusual residues are highlighted
on a yellow background. Residue 1 (Z) in HRPC is pyroglutamate, a
modification that is likely for all AtPs starting with glutamine
(Q). Predicted N-terminal ER-targeting signals have been removed
(Table 1; Supplementary material, Fig. S1) with alternative predic-
tions for AtP32 and AtP1 indicated in brackets. Some AtPs show
N-terminal extensions relative to HRPC residue 1, referred to as NX
propeptides in the text. C-terminal extensions, CX propeptides, are
shown in italics, and are not thought to be present in mature peroxi-
dase. Intron positions in the corresponding genes are indicated by
residues in reversed print, phase 0 introns between two marked resi-
dues, phase 1 and 2 introns within a single residue. Two genes marked
by (?) are unlikely to form stable proteins. At4g16270 ? encodes a
21-residue insert after intron 1 at HRPC position 48. At4g33870 ? has
an unusual intron 2 at position 122, and an extra intron at position
236, both of which give rise to abnormal sequences (marked in yellow).
Ó FEBS 2002 73 peroxidases from Arabidopsis (Eur. J. Biochem. 269) 6065
twosetsofcDNAsforAtP1,AtP1aandAtP1b,withthree
conserved nucleotide mismatches, and two sets for AtP2,
AtP2a and AtP2b, with 19 mismatches and three deletions.
AtP1b and AtP2a are identical in sequence to the genes
At4g21960 and At2g37130, respectively. The nucleotide
differences result in one amino acid substitution within the
6066 K. G. Welinder et al.(Eur. J. Biochem. 269) Ó FEBS 2002
Fig. 1. (Continued).
Ó FEBS 2002 73 peroxidases from Arabidopsis (Eur. J. Biochem. 269) 6067

Fig. 1. (Continued).
6068 K. G. Welinder et al.(Eur. J. Biochem. 269) Ó FEBS 2002
Fig. 1. (Continued).
Ó FEBS 2002 73 peroxidases from Arabidopsis (Eur. J. Biochem. 269) 6069
Fig. 1. (Continued).
6070 K. G. Welinder et al.(Eur. J. Biochem. 269) Ó FEBS 2002
putative mature AtP1, and three in AtP2. Differences
between transcripts and corresponding genes for AtP4,
AtP5, AtP7 and AtPN gave rise to one amino acid
substitution within the mature proteins. Two substitutions
were found for AtPCb and AtP6, and six for AtP14. Other
observed differences resulted from splice variants, for exam-
ple in AtP9, AtP15 [32], AtP36 (GenBank AF451952) and
AtPEa (TIGR TC115446 and TC115444).
Protein structure of 73 putative peroxidases
Figure 1 shows Arabidopsis peroxidases without their
predicted ER-signal peptides, sorted and aligned according
to similarity. The same similarity order is adopted in
Tables 1 and 2. The sequences are compared with the
classical HRPC which is 91% identical to AtPCb. The
atomic structure of HRPC has been solved at 2.15 A
˚
resolution by X-ray crystallography [33]. Moreover, HRPC
has been solved at 1.8 A
˚
resolution in complex with the
substrate analog benzhydroxamic acid [34], and at 1.45 A
˚
resolution in the ternary complex of HRPC–cyanide–ferulic
acid [35]. The structural elements of HRPC are shown in

Fig. 2 in the same color as in Fig. 1 for reference. The
structures of peanut peroxidase C1 [36], 67% identical to
AtP49, barley grain peroxidase BP1 [37], 56% identical to
AtP4, and recombinant mature AtPN [38], AtPA2 [39,40],
and soybean peroxidase SBP [41], 61% identical to AtPA2
and 60% identical to AtPEa, have also been determined by
X-ray crystallography. All showed the same active site
structure and very similar protein folds, except for BP1 that
is inactive above pH 5, and at pH 5.5, 7.5 and 8.5 has a
distorted loop of 21 residues [37]. This appears to be a
special feature of BP1.
Active site residues of the plant peroxidase superfamily
[4], shown in red in Fig. 1, include the catalytic distal Arg38,
and His42 hydrogen-bonded to Asn70. In addition, the
carbonyl of Pro139 accepts a hydrogen bond from reducing
substrates and thereby becomes a determinant of peroxidase
substrate specificity [39,40]. At the proximal site of the
heme, His170 is coordinated to heme Fe
3+
and hydrogen
bonded to Asp247 [42]. Many active site mutants have been
designed for HRPC with the purpose of studying the
function of the individual side chains (reviewed in [10,43]).
Proximal His and Asp are both invariant in Fig. 1. At the
distal site, the most significant substitutions occur in the
74% identical AtP50 and At5g24070 proteins, where Phe41-
His42 is replaced by Tyr-Ser. The substitution of distal
histidine will result in a different reaction mechanism. The
change of Asn70, found in seven peroxidases, can cause a
significant change in the enzyme kinetics [43].

Two stabilizing Ca
2+
ions are present in the structures of
all active class III peroxidases presently known. Figure 1
shows the predicted side chain ligands in blue, and
demonstrates that they are very well conserved. Main chain
carbonyl oxygen and a water molecule hydrogen-bonded to
the invariant Glu64 contribute other ligands. Each Ca
2+
Table 1. Annotation of the class III peroxidase gene family in Arabidopsis. Peroxidases are listed in the same similarity order as in Fig. 1, and referred
to by gene accession number at MIPS, AtP name and cDNA accession number at GenBank. Underlined cDNAs were sequenced in this work;
accession numbers from Ceres, Inc. are in parentheses. Positions of introns (1, 2, 3 and atypical n) and phases were predicted using the server at the
Technical University of Denmark ( and confirmed with available cDNA sequences.
NETSTART
and
SIGNALP
at this
server were used for predicting start methionine residues and N-terminal signal peptides. 5¢-UTR sequences were annotated with known cDNAs
and by using the
NNPP
program at University of California, Berkeley (itfly.org/seq_tools/promoter.html). The length and adenosine
contentof5¢-UTRs are given from observed and predicted (o/p) data. Predicted protein length is from the most likely start methionine. Score
corresponds to the maximum cleavage site probability predicted with the hidden Markov model. Underlined numbers indicate alternative
predictions. pI values were calculated from the putative mature proteins truncated to HRPC residues 1–305.
Peroxidase nomenclature Introns 5¢-UTR Protein Signal peptide
pI
Gene no.
MIPS Name
cDNA
acc. no. Name Phase

Length
(o/p)
A%
(o/p)
Start
Met score
Length
(aa) Length Score
At3g49120 AtPCb X71794 123 001 50/54 28/26 0.481 353 30 0.722 8.8
At3g49110 AtPCa AY049304 123 001 49/53 29/28 0.468 354 31 0.798 8.4
At3g32980 AtP16
X98777 123 001 44/48 39/35 0.478 352 29 0.727 7.7
At4g08770 AtP38
AF452387 123 001 11/51 55/49 0.682 346 22 0.899 8.1
At4g08780 123 001 –/52 –/46 0.534 346 22 0.900 8.1
At2g38380 AtPEa
AF452388 123 001 59/62 36/34 0.830 349 29 0.629 6.0
At2g38390 AtP34 AF452385 123 001 45/49 33/35 0.844 349 29 0.655 8.7
At5g06730 AtP29
Y11794 123 001 57/66 46/44 0.692 358 31 0.581 4.8
At5g06720 AtPA2
X99952 123 001 48/75 42/39 0.757 335 30 0.333 4.8
At5g19880 AtP42 (100990) 123 001 64/64 34/34 0.760 329 23 0.736 5.0
At5g19890 AtPN
X98453 123 001 67/69 45/45 0.647 321 21 0.978 6.4
At5g58390 AtP44 (124846) 12- 00- 81/83 43/42 0.293 316 19 0.957 9.9
At5g58400 12- 00- –/63 –/56 0.516 325 28 0.732 9.6
At5g05340 AtP49 AY065270 123 001 56/59 38/36 0.817 324 21 0.525 8.9
At1g14540 AtP46 AI996783
a

123 001 –/112 –/40 0.494 315 19 0.471 7.7
At1g14550 123 001 –/115 –/45 0.616 315 24 0.829 8.7
At5g66390 AtP6
X98774 123 001 40/66 50/47 0.575 336 23 0.763 8.6
At3g50990 123 001 –/52 –/42 0.457
336 21 0.764 4.7
At4g36430 AtP31
AF452384 123 001 49/52 27/27 0.608 331 22 0.982 8.8
Ó FEBS 2002 73 peroxidases from Arabidopsis (Eur. J. Biochem. 269) 6071
Table 1. (Continued).
Peroxidase nomenclature Introns 5¢-UTR Protein Signal peptide
pI
Gene no.
MIPS Name
cDNA
acc. no. Name Phase
Length
(o/p)
A%
(o/p)
Start
Met score
Length
(aa) Length Score
At2g18140 123 001 –/97 –/29 0.865 337 22 0.473 5.8
At2g18150 AtP36
AF451952 123 001 66/69 27/26 0.901 338 22 0.613 5.8
At1g44970 AtP18
X98804 123 001 25/25 52/52 0.894 346 23 0.691 7.0
At2g35380 AtP28 Y11793

a
12- 00- –/36 –/47 0.542 336 24 0.736 5.3
At2g22420 AtP25
Y11790 12- 00- 70/92 49/48 0.739 329 20 0.572 5.0
At1g49570 AtP5
X98809 123 001 33/54 33/35 0.706 344 21 0.464 5.6
At1g68850 AtP23
Y11789 123 001 56/74 43/39 0.613 336 20 0.929 5.1
At4g16270? 123 101 –/66 –/42 0.618 21 0.737
At1g71695 AtP4
X98773 12- 00- 54/54 65/65 0.779 358 31 0.515 8.4
At5g42180 AtP17
X99096 123 001 70/74 43/43 0.864 317 22 0.869 9.2
At5g51890 AtP27
Y11792 12- 00- 66/66 44/44 0.886 322 24 0.803 9.4
At4g33420 AtP32
AF451951 123 001 57/57 42/42 0.627 314 25 0.385 5.8
At4g33870? 1n3n 0202 –/79 –/42 0.452 24 0.207
At5g64100 AtP3
X98808 ) 23 ) 01 61/64 51/52 0.860 331 23 0.870 9.1
At5g64110 AtP45 AY065173 ) 23 ) 01 84/89 52/53 0.745 330 24 0.588 6.1
At5g64120 AtP15
X99097 ) 23 ) 01 56/61 45/43 0.740 328 23 0.618 8.2
At5g39580 AtP24
Y11788 ) 23 ) 01 52/83 50/47 0.920 319 22 0.755 8.7
At2g41480 123 001 –/39 –/49
0.415 328 26 0.233 6.6
At1g77100 123 001 –/26 –/27
0.249 319 22 0.990 5.0
At4g25980 1 – 0 – –/63 –/21 0.479

326 24 0.657 5.4
At5g17820 AtP13
X98776 1–n 0–2 61/66 49/47 0.811 313 22 0.952 10
At3g03670 AtP39 (41446) 1–n 0–2 28/67 46/42 0.418 321 21 0.976 4.9
At1g34510 1 – 0 – –/81 –/47 0.850 310 20 0.980 9.3
At4g26010 AtP35
AF452386 1 – 0 – 58/61 40/38 0.756 319 20 0.546 10
At5g22410 AtP14
X98803 123 001 22/38 55/47 0.778 331 26 0.645 7.0
At2g43480 AtP50 AY078928 123 001 13/60 15/25 0.817 335
25 0.542 8.7
At5g24070 123 001 –/161 –/34 0.772 340 25 0.638 7.1
At3g21770 AtP7
X98854 123 001 78/79 41/41 0.677 326 24 0.593 9.7
At1g05260 AtPRC U97684 123 001 59/59 51/51 0.767 326 24 0.771 8.8
At4g11290 AtP19
X98805 123 001 30/44 47/43 0.700 326 23 0.692 6.6
At1g05240\ AtP11
X98802 123 001 45/68 40/34 0.835 325 21 0.796 9.3
At1g05250/
At3g01190 AtP12
X98775 123 001 59/59 54/54 0.706 321 23 0.665 9.1
At5g15180 AtP33 AY072172 123 001 42/42 43/43 0.480 329 31 0.566 8.7
At2g39040 AtP47 AV554730 123 001 –/57 –/53 0.514
350 27 0.269 8.1
At4g37520 AtP9
X98314 123 001 42/44 40/41 0.622 329 25 0.938 9.0
At4g37530 AtP37
AF469928 123 001 34/37 38/38 0.762 329 25 0.947 8.4
At5g67400 AtP10

X98928 123 001 40/64 43/41 0.823 329 25 0.846 9.4
At3g49960 AtP21
X98807 123 001 48/78 23/24 0.793 329 25 0.732 9.4
At4g30170 AtP8
X98855 123 001 81/84 38/38 0.650 325 25 0.804 9.4
At2g18980 AtP22
Y08781 123 001 6/23 33/39 0.193 323 23 0.921 9.6
At5g14130 AtP20
X98806 12- 00- 39/97 59/42 0.726 330 30 0.828 4.9
At5g40150 AtP26
Y11791
a
– – –/151 –/21 0.881 328 27 0.909 8.6
At3g28200 AtP41 AY034973 – – 12/15 25/20 0.719 316 19 0.641 9.2
At5g47000 AtP43 AY093131 – – 167/170 26/26
0.782 331 25 0.824 6.8
At4g17690 – – –/99 –/32 0.821 326 20 0.987 8.4
At1g24110 – – –/362 –/39 0.160 326 20 0.453 6.1
At2g34060 AtP51 AY080602
a
12- 00- –/18 –/39 0.355 346 31 0.259 9.1
At3g17070 AtP40 (155041) 1–3 0–1 53/130 42/32 0.568 339 28 0.735 4.8
At1g30870 AtP30 AA067592 1 – 0 – 50/57 54/54 0.729 349 22 0.526 7.7
At2g24800 123 001 –/186 –/30 0.494 329 29 0.793 5.0
At4g31760 AtP48 AI999763
a
123 001 –/365 –/32 0.658 326 26 0.332 4.6
At4g21960 AtP1
X98189 123 001 70/76 39/38 0.905 330 27 0.387 8.1
At2g37130 AtP2

X98190 n123 2001 54/56 54/54 0.362 327 28 0.436 6.8
At1g34330 pseudogene
At3g42570 pseudogene
a
Nonfull-length cDNA.
6072 K. G. Welinder et al.(Eur. J. Biochem. 269) Ó FEBS 2002
Table 2. Amino acid sequence identity (%) of putative mature Arabidopsis peroxidase pairs.
Peroxi-
dase 1
Peroxidase 2
AtPCb
AtPCa 94 AtPCa
AtP16 91 91 AtP16
AtPEa 69 68 71 AtPEa
AtPA2 56 55 55 58 AtPA2
AtPN 52 51 51 51 56 AtPN
AtP6 49 49 49 46 52 50 AtP6
AtP31 50 50 50 47 52 53 70 AtP31
AtP18 50 49 50 50 53 50 63 62 AtP18
AtP5 47 46 47 45 51 50 50 49 50 AtP5
AtP23 43 44 44 43 42 42 43 44 43 43 AtP23
AtP4 44 43 44 44 42 44 40 43 42 41 36 AtP4
AtP17 42 41 4244 40 44444643383841AtP17
AtP32 41 40 4141 41 4441403941384147AtP32
AtP3 38 37 39 39 39 39 40 41 41 40 33 37 38 35 AtP3
AtP15 41 42 4244 47 45424243443839423958AtP15
AtP24 41 40 4041 48 4243444242363940395877AtP24
AtP13 41 41 4141 44 444342434036394143454647AtP13
AtP14 37 37 3736 37 39363536393535343842404047AtP14
AtP7 39 39 39 39 39 37 42 42 40 40 34 40 41 39 42 41 41 41 37 AtP7

AtP19 39 39 4041 42 404343414336394241414443433765AtP19
AtP11 38 38 3842 39 39404243403541433843424140364847AtP11
AtP12 38 37 3942 41 4141434541354242414544464439545654AtP12
AtP33 39 38 4039 40 424342424336444439404343403750545474AtP33
AtP9 39 40 40 41 42 37 38 38 35 37 33 39 43 37 38 39 37 42 38 40 38 41 37 35 AtP9
AtP10 38 38 3839 40 3938383736333941373637374037403840393974AtP10
AtP21 38 38 3840 41 413739363633394338373939423638383840377283AtP21
AtP8 38 38 39 41 41 38 40 39 38 38 32 41 39 35 37 40 42 43 37 41 40 41 39 37 65 65 64 AtP8
AtP22 40 40 3942 43 4041423938334241383740424336404141413963656484AtP22
AtP20 39 39 3938 39 403840353736424340384341423841424142405755545557AtP20
AtP26 38 37 3838 38 37393936343542364236373740333636383936444345434544AtP26
AtP1 34 34 34 34 36 34 36 36 34 32 31 36 33 34 31 34 33 37 31 32 32 33 35 34 34 33 33 33 34 35 37 AtP1
AtP2 33 33 33 32 33 32 39 38 36 32 33 34 32 34 28 32 32 38 31 32 32 33 34 34 32 32 32 33 35 35 39 57 AtP2
Ó FEBS 2002 73 peroxidases from Arabidopsis (Eur. J. Biochem. 269) 6073
site has two negatively charged aspartates and one or two
hydroxy side chains as ligands, except that AtP4 and AtP33
have one glutamate substituting for an aspartate. The Ca
2+
sites of the proximal domains of the AtP50 and At5g24070
proteins have only one negatively charged aspartate, and
might bind a monovalent cation similar to some class I
ascorbate peroxidases [5].
The presence of four disulfide bridges linking HRPC
cysteine residues, 11–91, 44–49, 97–301, and 177–209 are
conserved in class III peroxidases only, and highlighted in
dark yellow color in Fig. 1. The last Cys301 of AtP27 is
changed to a threonine. Therefore, only three disulfide
bridges can exist in this putative peroxidase, presumably
resulting in decreased stability.
A buried salt bridge motif, Asp99-Arg123, is an

invariantpartofallthreeclassesoftheplantperoxidase
superfamily [44]. This motif (shown in grey shading in
Fig. 1) includes Ser96 and Asp99 located at the beginning
of helix D and connecting to the following long loop in a
tight hydrogen bonding network with Gly122-Arg123. This
long loop continues around half the molecular surface to
the C-terminal domain in all peroxidases. Welinder and
Gajhede [45] proposed that the peroxidase superfamily
arose by an ancient gene duplication event, where helix D
(and exon 2, see Fig. 1) terminated the first part corres-
ponding to the present-day N-terminal domain. Conse-
quently, the salt bridge may play an important role in the
proper attachment of the N- and C-terminal domains
during protein folding.
Several other residues are either invariant or highly
conserved (Fig. 1) in class III peroxidases, Leu2, Tyr7,
Pro12, Ile17, Phe41, Gly48, Glu64, Gly76, Phe77, Lys84,
Glu88, Pro92, Val95, Ala98, Gly114, Pro115, Asp125,
Phe152, Asp161, Leu166, Gly168, Gly173, Arg183, Gly242,
Leu250, Phe273, Phe277, Met281, Gly295, Arg298 and
Asn305. Most of these conserved residues appear to be of
importance to the integrity of peroxidase structure. Several
aromatic positions, 6, 7, 45, 185, 201 and 233–234, are
conserved and might have special functions, for example in
electron-transfer reactions. On the other hand, hydrophobic
patterns within helices fulfil a standard role in protein
packing.
N-linked glycans were predicted for all putative peroxi-
dases and are indicated by green NXS/T triplets (Fig. 1).
Most surface turns or loops connecting helices are amenable

to glycosylation, in particular the highly variable loop
between helices F¢ and F¢¢. The eight N-linked glycans of
HRPC have been experimentally verified [46]. O-linked
glycans have never been seen in a plant peroxidase. Triplets
containing X ¼ proline, or followed by a proline residue,
are not glycosylated in peroxidases that we have sequenced
[29,47,48], or in other proteins [49]. The latter statistical
study also found a decrease in glycosylated triplets towards
the C-terminus. The only glycan seen in barley peroxidase
BP1 is found at the Asn-Cys-Ser triplet, residues 300–302,
near the C-terminus of the mature peroxidase, however,
31 amino acids before the C-terminus of the properoxidase
existing during glycan attachment in the ER [48]. This
glycan is most likely present in the similar AtP4 (Fig. 1).
Triplets overlapping with the distal active site residue
corresponding to Asn70 of HRPC, or with Ca
2+
ligands
that are buried in the folded structure, are unlikely to be
glycosylated in a functional peroxidase, and have been
excluded in Fig. 1. Again BP1 provides an experimental
example of a nonglycosylated triplet at position Asn70 [48].
The majority of the peroxidases carry one or two putative
glycans. Seven appear to be nonglycosylated. Therefore, the
high number of glycans found in HRP C, E and A types is
unusual among class III peroxidases. Since glycans are
large, those close to substrate-binding residues (near
Pro139) are likely to affect substrate access and reaction
dynamics, due to a dampening of backbone motion [40].
Half of the putative mature AtPs are likely to start with a

pyrrolidone carboxylyl residue (Z) formed from glutamyl
(Q). This has been shown experimentally for HRPC,
HRPE5 [50] (similar to AtP34 and AtPEa), HRPA2
(similar to AtPA2 [51]), and for peroxidases originating
from tobacco, turnip (reviewed in [11]), zucchini [52] and
soybean (K.G. Welinder, unpublished). Interestingly, fol-
lowing the predicted signal peptide, which varies in length
from 19 to 31 residues (Table 1), a few peroxidases have an
N-terminal extension, which we will refer to as the NX
propeptide (Fig. 1). The possible targeting potential of the
N-terminal sequences from the initiating Met was therefore
analyzed by
TARGETP
, which distinguishes between chloro-
plast, mitochondrial, secretory pathway and ÔotherÕ locali-
zations [26]. All AtPs were predicted for the secretory ER
pathway by
TARGETP
. However, the peculiar G, N or D/E
rich sequences of the NX propeptide might indicate a lack of
structure and susceptibility to proteolytic degradation.
Experimental work on secreted peroxidases isolated from
sweet potato tissue culture has demonstrated that their
N-termini (Asp-Glu-Ala-Cys-) start only four residues
before the first conserved Cys11 of Fig. 1 [53]. This is 66
amino acids after the initiating Met.
SIGNALP
predicted that
Fig. 2. Structural elements of HRPC. The central heme group is
sandwiched between the two protein domains. Helices A–J are

marked. The structural elements are highlighted in the same colors as
in Fig. 1. (By courtesy of A. Henriksen, Carlsberg Research Center,
Denmark.)
6074 K. G. Welinder et al.(Eur. J. Biochem. 269) Ó FEBS 2002
only the first 34 amino acids serve as the ER-signal.
Therefore, it appeared that more than 30 amino acids had
been removed by proteolysis. Whether this was an artefact
of tissue culture conditions, or can also occur in planta,is
unknown. In the case of BP1, purified from barley grain, the
mature protein is indeed extended by seven residues at the
N-terminus (Ala-Glu-Pro-Pro-Val-Ala-Pro-) compared to
HRPC [48]. In BP1 the prolines might protect against
proteolysis. Vacuolar targeting of an NX propeptide is
possible. Thus, the protein sporamin contains the vacuolar
targeting sequence Asn-Pro-Ile-Arg-Leu at its N-terminus, a
sequence that if moved to the C-terminus still provided
vacuolar localization [54]. The observation of an NX
propeptide in some peroxidases is novel, and its role has
not been analyzed experimentally.
Some peroxidases show a C-terminal extension, a CX
propeptide, indicated in italics in Fig. 1, which appears to
target for vacuolar import. The function of a CX propeptide
was first discussed for barley grain peroxidase BP1, because
the cDNA clone encoded an additional 22 residues
preceding the stop codon as compared with the amino acid
sequence of the mature protein [48]. The import is associ-
ated with removal of the propeptide since the purified
mature proteins HRPC [46], HRPE5 [50] and barley grain
peroxidase BP1 terminates before these propeptides. Vac-
uolar location of barley grain peroxidase BP2 has been

demonstrated by immuno-gold electron microscopy [55].
BP2 has a CX propeptide similar to that of BP1 but no NX
(K.G. Welinder, unpublished results). BP1 is most similar to
AtP4, which also has a CX as well as an NX (Fig. 1). The
propeptides indicated in italics for AtP29 and AtP25 are
predicted in analogy to peroxidases known to loose the CX.
Similar CX peptides have also been observed in chitinases,
lectins, proteases, protease inhibitors and others [56]. The
targeting features of the propeptide of barley lectin have
been analyzed by amino acid substitutions [57]. The unique
extension of AtP14 might indicate the presence of a short
cystine loop unique to AtP14. Its function, if any, is
unknown.
Functional proteins
We predict that all 58 AtPs of Fig. 1 can exist as functional
proteins, because (a) they are derived from cloned cDNA
and are therefore transcribed, and (b) contain the funda-
mental sequence characteristics of HRPC. Furthermore, we
predict that 13 additional gene products can fold into
functional proteins according to (b). Only At4g16270 ?
and At4g33870 ? code for abnormal sequences introduced
at intron 1 ( position 48) and intron 2 ( position 112),
respectively (marked in yellow in Fig. 1). The majority of
the 71 predicted mature peroxidases have rather high
isoelectric points (Table 1). In our experience [11], predicted
isoelectric points between 5 and 10 are generally two units
lower than the experimental values, because the two calcium
ions and the heme are not included in the calculation. Only
11 peroxidases have predicted pI values £ 5, among those
AtPA2 and AtP29 (AtPA1) putative orthologs to the well-

known anionic horseradish peroxidases HRPA2 and
HRPA1, respectively.
Table 2 shows the pairwise amino acid sequence identities
among the 33 putative mature peroxidases subjected to
RT-PCR expression analysis and illustrates the tremendous
evolutionary divergence among the Arabidopsis peroxidases.
(All pairs are shown in Supplementary Table S3). Only a
few clusters show greater than 70% amino acid sequence
identity, which might be considered as a lower limit for
potentially related biochemical function. Such clusters are
boxed in Fig. 1.
Expression of AtP transcripts
Peroxidases show very limited substrate specificity in
general. Therefore, the biological functions of Arabidopsis
peroxidases were approached by expression studies. The
temporal and spatial expression of mRNA coding for
Arabidopsis peroxidases was analyzed by RT-PCR. This
method can differentiate between similar genes contrary to
the methods based on hybridization. RT-PCR is very
sensitive, however, it is not a quantitative method. Speci-
ficity was obtained using unique primer sets designed to
discriminate known peroxidase genes in combination with
individually optimized annealing conditions as outlined in
the methods section. One primer was preferentially placed in
the poorly conserved 5¢-or3¢-UTR (Supplementary mater-
ial, Table S1). The primer specificity has been checked
against the 73 Arabidopsis peroxidase genes. Only the
primers for AtP11 cannot discriminate between the identical
AtP11.1 and AtP11.2 genes. The PCR fragment sizes were
generally between 450 and 950 bp. Experiments were done

twice using two different reverse transcribed reactions of
purified total mRNA from whole plant and dissected organs
at 3–59 days after germination (Table 3). The analyses were
standardized using ef-1a as a positive reference [19]. This
gene was found to be among the 10 most abundant
transcripts in Arabidopsis root, flower bud and open flower
[58], and can be considered as constitutively expressed in the
plant. This is supported also by its ubiquity in available EST
libraries. Table 3 shows all observations including weak
ones. Three examples are shown in Fig. 3.
The RT-PCR expression analyses were carried out for 33
Arabidopsis peroxidase genes, encoding AtP1–24, AtP26,
AtP31–33, AtPCa, AtPCb, AtPEa, AtPA2 and AtPN, as
only these were available at the time. The majority were
expressed from day 6–35 in whole plants (Table 3). Two
transcripts, for AtP14 and AtP19, could not be detected in
any organ at any time. The EST count (Table 4) supports
that AtP14 is rare, whereas AtP19 is present in root. A
sequence error in a primer for AtP19 may have occurred. It
appeared that all AtPs, except for AtP14 and AtPCa, are
expressed in root. Nine AtPs were found in root only.
Thirteen peroxidases were transcribed in all examined
organs, whereas nine showed a restricted expression in
certain organs.
Recently the Arabidopsis EST database exceeded
100 000 entries. Therefore, counting the ESTs for each
peroxidase (Table 4) may provide an independent esti-
mate of gene transcription that is also not biased by gene
similarity. On average, 0.85% of all Arabidopsis ESTs
presently known encode a class III peroxidase. Table 4

indicates that the level of transcription of peroxidase
genes varies tremendously, from zero to 181 ESTs.
Twenty-nine of the 49 different AtP ESTs have been seen
> 5 times. We consider libraries > 10 000 and EST
counts > 5 as significant semiquantitative indicators of
expression level [59,60], despite the fact that some of the
Ó FEBS 2002 73 peroxidases from Arabidopsis (Eur. J. Biochem. 269) 6075
libraries were normalized [14]. Groups 1, root, 4, Ôabove
groundÕ or aerial, 6, green silique, 7, germinating seed, 9,
ÔvariousÕ (dominated by the PRL2 library [12]), and ÔtotalÕ
shown in Table 4 were considered.
In general, the RT-PCR and EST counts agree well
regarding expression. However, the RT-PCR sensitivity was
optimized and very high. This can be seen by comparing the
expression results of whole plant and root in Tables 3 and 4.
Both methods support a constitutive expression of AtP1,
AtP2, AtP9, AtP17, AtP4 and AtPCb. It is noteworthy that
these AtPs represent highly diverse structures (Fig. 1,
Table 2) and properties. Three peroxidase transcripts,
AtP1, AtP16 and AtPCb, have been seen more than 100
times. RT-PCR did not differentiate expression levels
among organs of AtP16, while EST counts indicated a
high, but not exclusive, preference for root. AtP16 and
AtPCb are 91% identical in amino acid sequence (Table 2)
and probably catalyze identical reactions, if they are
expressed in the same cell type.
In root the AtPs accounted for an impressive 2.2% of
ESTs, and 35 different AtPs have already been seen in root
(Table 4). Both RT-PCR and EST counts (Tables 3 and 4)
demonstrated root preference of AtP22, AtP3, AtP11,

AtP12, AtP5, AtP6, AtP10 and AtP21. Again these AtPs
represent highly diverse structures (Fig. 1, Table 2), except
for AtP10 and AtP21 (83% amino acid identity). In general,
it appears that similar peroxidases (> 70% amino acid
identity) showed differential expression and therefore poss-
ibly fulfil different biological roles.
In the Standford microarray database, expression data
for about 8000 Arabidopsis ESTs, including 55 ESTs
representing 28 different peroxidase genes, are available
( />Expression levels in root, flower bud and leaf showed a clear
preferential expression of all peroxidases in root. High
expression levels of peroxidases in root have also been
demonstrated in rice, where 21 different peroxidase tran-
scripts were investigated [61]. Peroxidase ESTs were also
highly abundant in a root library of rice (Fig. 3 in [62]).
Therefore, all evidence supports that a variety of peroxidase
activities are needed in the development and maintenance of
root tissue.
Table 3. Temporal and spatial expression of Arabidopsis peroxidase mRNAs analyzed by RT-PCR. The majority of clones originated from a
kZipLox library designated PRL2, mixed from etiolated seedlings, roots, rosette leaves, stems, flowers and siliques to represent whole healthy plants
at all ages [12]. Whole plants were analyzed at day 3, 6, 15, 35 and 59 after sowing. Roots were analyzed at day 15, 35 and 59, rosettes at day 35 and
59, stems, cauline leaves, and flower buds at day 59. Cell culture was obtained as described previously [51]. Dry seeds gave a low signal for ef-1a and
no peroxidases were seen. (+) Indicates observed at all time points. An example experiment with AtP16, AtP13 and AtP22 compared to the control
ef-1a is shown in Fig. 3. EST counts of peroxidases in 111 206 Arabidopsis ESTs are from Table 4.
Peroxidase Clone origin Whole plant (day) Root Rosette Stem Cauline leaf Flower bud Cell culture EST counts
AtP1 PRL2 + + + + + + 181
AtP2 PRL2 + + + + + + 38
AtP7 PRL2 + + + + + + + 4
AtP9 PRL2 + + + + + + + 38
AtP16 PRL2 + + + + + + 129

AtP17 PRL2 + + + + + + 22
AtP32 PRL2 + + + + + + + 2
AtP15 PRL2 + 35 + + + + 18
AtP24 seedling + + + + + + 7
AtP4 PRL2 6,15,35 + + + + + 27
AtPCb cDNA 6,15,35,59 + + + + + + 101
AtP8 PRL2 + + + + + + + 37
AtP26 PRL2 6,15,35,59 35,59 + + + + 3
AtPCa gene 15,35,59 + + + 2
AtPEa PRL2 6,15,35,59 + + + 62
AtP13 PRL2 + + + + 13
AtP31 silique + + + + + 2
AtP23 PRL2 6,15,35,59 + + 7
AtPN PRL2 6,15,35,59 + + + 6
AtPA2 cell culture + + + 3
AtP18 PRL2 3,6,15,35 + + 1
AtP22 PRL2 6,15,35,59 + + 6
AtP3 PRL2 + + 53
AtP11 PRL2 + + 31
AtP12 PRL2 + + 13
AtP5 PRL2 35 35 8
AtP33 seedling 6,15,35 15,35 1
AtP6 PRL2 3,6,15 35 12
AtP10 PRL2 3,6,15,35 15,35 11
AtP21 PRL2 3,6,15,35 15,35 9
AtP20 PRL2 6,15,35,59 + 3
AtP14 PRL2 No transcipts detected 2
AtP19 PRL2 No transcipts detected 24
6076 K. G. Welinder et al.(Eur. J. Biochem. 269) Ó FEBS 2002
There is no simple correlation between gene transcrip-

tion, analyzed in this study, and translation into protein.
Protein extracts from Arabidopsis plants or cell suspen-
sion culture showed rather few peroxidase bands after
isoelectric focusing. Only the AtPA2 [51], AtPEa and
AtPCb proteins (R. Matthiesen and K.G. Welinder,
unpublished results) have been identified. It is plausible
that translation of existing peroxidase transcripts may be
triggered by stress. Østergaard et al. [23] found that 23 of
30 available AtPs had 5¢-UTRs containing 40–71%
adenine, a rare feature observed also in cDNAs that
predominantly encoded stress-induced proteins. The pre-
sent study extends this observation (Table 1; Supplement-
ary material, Table S2). Among the 57 observed AtPs, 29
clearly have adenine contents > 40% in the observed or
predicted 5¢-UTRs, and only 18 have < 40% A. Some
are close to 40% A, or the observed 5¢-UTR were
different from the predicted 5¢-UTR. The transcripts for
C-,E-andA-typesAtPCb,AtPCa,AtP16,AtPEaand
AtP34 all have adenine contents < 40% and are
therefore thought to be translated into proteins by the
standard mechanism. The 5¢-UTR analyses revealed that
very long transcripts have been found for AtPCb and for
AtPA2, but only once (not shown), whereas the majority
of transcripts were shorter and of lengths comparative to
those for the similar AtP16 and AtP29, respectively
(Supplementary material, Fig. S2). This indicates that
alternative sites for the start of transcription might be
functional.
AtP1 provided an unusual example in several respects. It
is the most highly expressed of all AtPs (181 ESTs), but it

has no consensus TATA-box. The AtP1 protein has not
been detected despite a 38–39% A content of the 5¢-UTR.
However, studies of recombinant AtP1 showed a very low
specific peroxidase activity and absorption spectra typical of
cytochrome b
5
-type heme coordination, independent of pH
(I.V.H. Kjærsga
˚
rd and K.G. Welinder, unpublished results).
Therefore, AtP1 will stain poorly for peroxidase activity and
might not be detected.
The A-, C-, and E-type peroxidase proteins are abundant
in horseradish root, and traditionally named according to
pI, with A being the most acidic and E the most basic
[1,50,63]. The transcripts for AtPC-type, AtP16 (129 ESTs)
and AtPCb (101 ESTs), are the most abundant of all in
Arabidopsis root, contrary to AtPCa (2 EST), AtP38 (6
ESTs) and At4g08780 (0 ESTs). The gene for AtPCa is
followed in tandem by AtPCb on chromosome 3, whereas
AtP16 is single (gene numbers and names in Table 1;
protein similarity in Fig. 1). The genes for AtP38 and
At4g08780 are also in tandem, however, in inverted
orientation. The two E-types of genes also appear in
tandem, AtPEa (62 ESTs) is followed by AtP34 (7 ESTs).
Botharewellexpressedinroot.AllC-andE-type
transcripts encode a vacuolar targeting CX signal (Fig. 1).
The biological functions of the abundant cationic C- and
E-type peroxidases are unknown.
AtP38 appears to be the ortholog of HRPC2 (91% amino

acid identity). Shinmyo et al. [64] have studied the promoter
activity and wound-induction of HRPC and -E types, and
of AtPCa and AtPEa. Most remarkably, C2, and only C2,
responded strongly to wounding. The AtP38 gene might
respond similarly.
The Arabidopsis genome encodes two 84% identical
anionic A-type peroxidases (Supplementary material,
Table S3). AtPA2 is followed in tandem by AtP29
(AtPA1, putative ortholog of HRPA1). Both AtPA2 (3
ESTs) and AtP29 (3 ESTs) showed low expression and
preference for root. It is interesting that only AtP29 has a
CX propeptide, which suggests a vacuolar location
(Fig. 1). This difference does not exist between any other
tandem pairs of peroxidase genes (L. Duroux and
K.G. Welinder, unpublished results). The role of the
extracellular AtPA2 in lignification has been demonstra-
ted recently [39].
Stress and peroxidase expression
Rather few transcripts were seen in cell suspension culture
derived from root tissue, despite the fact that nearly all AtPs
were observed in root (Table 3). AtPA2, however, consti-
tuted the major peroxidase of spent medium, and was
originally purified and cloned from this source [51]. The
result suggested that many of the peroxidase genes expressed
in root in standard growth conditions were down-regulated
when the organ was subjected to conditions of stress, except
for the extracellular lignifying AtPA2.
From the total of 942 AtPs encoding ESTs, 916
originated from healthy plants grown at normal conditions,
Fig. 3. RT-PCR analysis of temporal and spatial expression of mRNA

coding for AtP16, AtP13, and AtP22. Reaction products are compared
to a standard of DNA bands of known sizes (last lane) by agarose gel
electrophoresis. The predicted sizes of the amplification products are
560 bp for AtP16, 420 bp for AtP13, 574 bp for AtP22, and 474 bp for
ef-1a. The larger size products represent amplification of traces of
genomic DNA. Reactions including reverse transcriptase are indicated
by (+), controls without by (–); a control without RNA is negative in
all cases (penultimate lane). Expression of mRNA encoding ef-1a was
used as a control of the quality of the RNA preparation, of the reverse
transcription reaction and for expression reference.
Ó FEBS 2002 73 peroxidases from Arabidopsis (Eur. J. Biochem. 269) 6077
Table 4. Numbers of Arabidopsis EST sequences, AtP sequences and the percentages expressed in different organs. Arabidopsis EST libraries were
grouped according to organ as outlined in Materials and methods, and the occurrence of the different AtPs counted. Counts of peroxidase ESTs.
The TIGR TC accession numbers (release v 6.0) are given. Some (splice-) variant forms are included, which have separate TC numbers (not listed).
(–) indicates no name or TC number. The AtPs are sorted according to apparent abundance in the various tissues.
Name TC number
EST library group
Total
1
Root
2
Hypocotyl
3
Rosette
4
Above ground
5
Flower bud
6
Silique

7
Seed, germ.
8
Seedling
9
Various
EST sequences 18 740 2559 2234 15 954 7915 13 382 10 800 1634 37 988 111 206
AtP sequences 411 14 6 66 14 107 40 17 267 942
>AtP (%) 2.19 0.55 0.27 0.41 0.18 0.80 0.37 1.04 0.70 0.85
AtP1 TC121255 35 2 27 4 49 10 54 18
AtP16 TC103486 100 2 4 23 129
AtPCb TC121266 49 1 12 3 15 1 1 19 101
AtPEa TC115448 41 1 17 7 66
AtP3 TC115510 19 232 53
AtP2 TC109534 9 6 6 1 3 13 38
AtP9 TC115527 8 2 17 3 8 38
AtP8 TC115545 17 218 37
AtP11 TC109595 24 1631
AtP4 TC103612 1 1 4 6 5 9 1 27
AtP19 TC115663 15 2 7 24
AtP17 TC121524 2 2 13 5 22
AtP15 TC109700 8 1918
AtPRC TC109863 5 1 2 2 3 13
AtP12 TC121670 2 11 13
AtP13 TC103850 6 2513
AtP6 TC115900 5 2512
AtP35 TC109980 11 11
AtP10 TC109924 10 111
AtP21 TC104183 6 39
AtP5 TC104287 3 1 4 8

AtP34 TC115445 5 27
AtP23 TC116168 2 5 7
AtP24 TC122224 4 21 7
AtP27 TC122135 2 3 2 7
AtP22 TC104843 3 1 2 6
AtP41 TC122514 5 1 6
AtP38 TC116195 4 26
AtPN TC104410 4 26
AtP7 TC104292 1 34
AtPA2 TC122775 1 1 13
AtP29 TC104523 2 13
AtP20 TC110922 33
AtP26 TC105068 1 1 1 3
AtPCa TC121267 2 2
AtP32 TC117759 1 12
AtP31 TC123560 1 1 2
AtP14 TC111256 22
AtP46 TC112253 1 1
AtP30 – 11
AtP18 TC116977 11
AtP36 TC122998 1 1
AtP25 TC117275 11
AtP28 TC105819 11
AtP47 TC125273 1 1
AtP39 TC106126 1 1
AtP48 TC112957 11
AtP33 TC125109 11
AtP43 TC124099 1 1
6078 K. G. Welinder et al.(Eur. J. Biochem. 269) Ó FEBS 2002
and only 26 from liquid cultures or plants exposed to saline

stress (included with EST group 9 ÔvariousÕ,Table4).None
came from pathogen- or wound-challenged tissue. Recently,
we cloned a cDNA, AtP37, for gene At4g37530 from a
cDNA library of Erysiphe cichoracearum infected plants
(kindly provided by S. Somerville and I. Wilson, Standford
University, USA). The gene for AtP37 follows the consti-
tutively expressed AtP9 gene in tandem. The 3345 ESTs
from this library included only constitutively expressed
AtP1 (6 ESTs) and AtP4 (1 EST), and no other AtPs
(M. Lundsgaard, J. Emmersen, K. L. Nielsen and K. G.
Welinder, unpublished results).
Interestingly, turnip peroxidase TP7 [47] is 90% iden-
tical to AtP49 (0 ESTs). The TP7 protein is the major
isoperoxidase of turnip root, but present only during
winter. Therefore, AtP49 might function under similar
conditions.
The suberizing peroxidase genes induced in potato and
tomato upon wounding and pathogenic attack [65] do not
present close relatives in Arabidopsis as demonstrated by
sequence comparison to all 73 putative Arabidopsis peroxi-
dases. The protein predicted from AtP49 is the closest to the
suberizing peroxidases with 48% identity. Similarly, the
extensively studied anionic peroxidase from tobacco [66] has
no close relative in Arabidopsis with anionic AtPA2 (65%
amino acid identity) being the closest. We consider an AtP a
putative ortholog to a class III peroxidase from a different
genus of the Brassicaceae family, if the AtP is the closest
relative, and if their amino acid sequence identity is ‡ 88%.
For nonBrassicaceae flowering plants this cut-off is ‡ 80%
aminoacididentity.

In general we found that the Arabidopsis repertory of
class III peroxidase genes accounted very well for peroxid-
ases from other Brassicaceae, whereas Solanaceae, including
potato, tomato and tobacco, Fabaceae, including soybean
and peanut, and Poaceae, including barley and rice, have
additional groups and subgroups of paralogous peroxidase
genes, as illustrated by the few examples mentioned
throughout this paper. A phylogenetic analysis of class III
peroxidases will appear in a separate paper.
ACKNOWLEDGEMENTS
This research was supported by the Danish Agricultural and Veterinary
Research Council (5.23.26.10–1 to KGW and SKR), the Danish
Natural Science Research Council (9502825 to KGW), and the
European Commission (FMRX CT98 02000 to KGW). We are
grateful to Dr Anette Henriksen, Carlsberg Research Center, for
preparing Fig. 2.
REFERENCES
1. Paul, K.G. (1986) Peroxidase, Historical background. In
Molecular and Physiological Aspects of Plant Peroxidases (Grep-
pin, H., Penel, C. & Gaspar, T., eds), pp. 1–14. University of
Geneva, Switzerland.
2. Claiborne, A. & Fridovich, I. (1979) Purification of the o-diani-
sidine peroxidase from Escherichia coli B. Physicochemical char-
acterization and analysis of its dual catalatic and peroxidatic
activities. J. Biol. Chem. 254, 4245–4252.
3. Blattner, F.R., Plunkett, G., Bloch, C.A., Perna, N.T., Burland,
V.,Riley,M.,Collado-Vides,J.,Glasner,J.D.,Rode,C.K.,
Mayhew, G.F., Gregor, J., Davis, N.W., Kirkpatrick, H.A.,
Goeden, M.A., Rose, D.J., Mau, B. & Shao, Y. (1997) The
complete genome sequence of Escherichia coli K-12. Science 277,

1453–1474.
4. Welinder, K.G. (1992) Superfamily of plant, fungal and bacterial
peroxidases. Curr. Opin. Struc. Biol. 2, 388–393.
5. Jespersen, H.M., Kjærsga
˚
rd, I.V., Østergaard, L. & Welinder,
K.G. (1997) From sequence analysis of three novel ascorbate
peroxidases from Arabidopsis thaliana to structure, function and
evolution of seven types of ascorbate peroxidase. Biochem. J. 326,
305–310.
6. Baunsgaard, L., Dalbøge, H., Houen, G., Rasmussen, E.M. &
Welinder, K.G. (1993a) Amino acid sequence of Coprinus mac-
rorhizus peroxidase and cDNA sequence encoding Coprinus
cinereus peroxidase. A new family of fungal peroxidases. Eur.
J. Biochem. 213, 605–611.
7. Baunsgaard, L., Vind, J. & Dalbøge, H. (1993b) The sequence of
Coprinus peroxidase gene ctp1.InPlant Peroxidases Biochemistry
and Physiology (Welinder, K.G., Rasmussen, S.K., Penel, C. &
Greppin, H., eds), pp. 239–242. University of Geneva, Switzer-
land.
8. Reddy, C.A. (1993) An overview of recent advances on the
physiology and molecular biology of lignin peroxidases of Phan-
erochaete chrysosporium. J. Biotechnol. 30, 91–107.
9. Penel, C., Gaspar, T. & Greppin, H. (1992) Plant Peroxidases
1980–90. Topics and Detailed Literature on Molecular, Biochemi-
cal, and Physiological Aspects. University of Geneva, Switzerland.
10. Veitch, N.C. & Smith, A. (2001) Horseradish peroxidase. Adv.
Inorg. Chem. 51, 107–162.
11. Welinder, K.G. (1992) Plant peroxidases, Structure-function
relationships. In Plant Peroxidases 1980–90, Topics and Detailed

Literature on Molecular, Biochemica, and Physiological Aspects
(Penel, C., Gaspar, T. & Greppin, H., eds), pp. 1–24. University of
Geneva, Switzerland.
12. Newman,T.,deBruijn,F.J.,Green,P.,Keegstra,K.,Kende,H.,
McIntosh, L., Ohlrogge, J., Raikhel, N., Somerville, S. & Tho-
mashow, M. (1994) Genes galore, a summary of methods for ac-
cessing results from large-scale partial sequencing of anonymous
Arabidopsis cDNA clones. Plant Physiol. 106, 1241–1255.
13. Cooke,R.,Raynal,M.,Laudie,M.,Grellet,F.,Delseny,M.,
Morris, P.C., Guerrier, D., Giraudat, J., Quigley, F., Clabault,
G., Li, Y.F., Mache, R., Krivitzky, M., Gy, J., Kreis, M.,
Lecharny, A., Parmentier, Y., Marbach, J., Fleck, J., Clement,
B., Philipps, G., Herve, C., Bardet, C., Tremousaygue, D. &
Hofte, H. (1996) Further progress towards a catalogue of all
Arabidopsis genes, analysis of a set of 5000 non-redundant ESTs.
Plant J. 9, 101–124.
14. Asamizu, E., Nakamura, Y., Sato, S. & Tabata, S. (2000) A large
scale analysis of cDNA in Arabidopsis thaliana, generation of
12,028 non-redundant expressed sequence tags from normalized
and size-selected cDNA libraries. DNA Res. 7, 175–180.
15. The Arabidopsis Genome Initiative (2000) Analysis of the genome
sequence of the flowering plant Arabidopsis thaliana. Nature 408,
796–815.
16. Tognolli, M., Penel, C., Greppin, H. & Simon, P. (2002) Analysis
and expression of the class III peroxidase large gene family in
Arabidopsis thaliana. Gene 288, 129–138.
17. Altschul, S.F., Gish, W., Miller, W., Myers, E.W. & Lipman, D.J.
(1990) Basic local alignment search tool. J. Mol. Biol. 215, 403–
410.
18. Altschul,S.F.,Madden,T.L.,Schaffer,A.,Zhang,J.,Zhang,Z.,

Miller, W. & Lipman, D.J. (1997) Gapped BLAST and PSI-
BLAST, a new generation of protein database search programs.
Nucleic Acids Res. 25, 3389–3402.
19. Kjærsga
˚
rd, I.V., Jespersen, H.M., Rasmussen, S.K. & Welinder,
K.G. (1997) Sequence and RT-PCR expression analysis of two
peroxidases from Arabidopsis thaliana belonging to a novel evo-
lutionary branch of plant peroxidases. Plant Mol. Biol. 33, 699–
708.
Ó FEBS 2002 73 peroxidases from Arabidopsis (Eur. J. Biochem. 269) 6079
20. MIPS, Munich Information Center for Protein Sequences. http://
mips.gsf./proj/thal/
21. TIGR, The Institute for Genome Research. />tdb/agi/
22. Hebsgaard, S.M., Korning, P.G., Tolstrup, N., Engelbrecht, J.,
Rouze, P. & Brunak, S. (1996) Splice site prediction in Arabidopsis
thaliana pre-mRNA by combining local and global sequence
information. Nucleic Acids Res. 24, 3439–3452.
23. Østergaard, L., Pedersen, A.G., Jespersen, H.M., Brunak, S. &
Welinder, K.G. (1998) Computational analyses and annotations of
the Arabidopsis peroxidase gene family. FEBS Lett. 433, 98–102.
24. Pedersen, A.G. & Nielsen, H.H. (1997) Neural network prediction
of translation initiation sites in eukaryotes, perspectives for EST
and genome analysis. Proc. Int. Conf Intell. Syst. Mol. Biol. 5,226–
233.
25. Nielsen, H., Engelbrecht, J., Brunak, S. & von Heijne, G. (1997) A
neural network method for identification of prokaryotic and
eukaryotic signal peptides and prediction of their cleavage sites.
Int. J. Neural. Syst. 8, 581–599.
26. Emanuelsson, O., Nielsen, H., Brunak, S. & von Heijne, G.

(2000) Predicting subcellular localization of proteins based on
their N-terminal amino acid sequence. J. Mol. Biol. 300, 1005–1016.
27. Thompson, J.D., Gibson, T.J., Plewniak, F.F., Jeanmougin, F.F.
& Higgins, D.G. (1997) The CLUSTAL_X windows interface,
flexible strategies for multiple sequence alignment aided by quality
analysis tools. Nucl Acids Res. 25, 4876–4882.
28. Gonnet, G.H., Cohen, M.A. & Benner, S.A. (1994) Analysis of
amino acid substitution during divergent evolution, the 400 by 400
dipeptide substitution matrix. Biochem. Biophys. Res. Commun.
199, 489–496.
29. Welinder, K.G. (1976) Covalent structure of the glycoprotein
horseradish peroxidase EC 1.11.1.7. FEBS Lett. 72, 19–23.
30. Quackenbush, J., Liang, F., Holt, I., Pertea, G. & Upton, J. (2000)
The TIGR gene indices, reconstruction and representation of
expressed gene sequences. Nucl Acids Res. 28, 141–145.
31.White,J.A.,Todd,J.,Newman,T.,Focks,N.,Girke,T.,de
Ilarduya, O.M., Jaworski, J.G., Ohlrogge, J.B. & Benning, C.
(2000) A new set of Arabidopsis expressed sequence tags from
developing seeds. The metabolic pathway from carbohydrates to
seed oil. Plant Physiol. 124, 1582–1594.
32. Justesen, A.F., Jespersen, H.M. & Welinder, K.G. (1998) Analysis
of two incompletely spliced Arabidopsis cDNAs encoding novel
types of peroxidase. Biochim. Biophys. Acta 93193, 1–6.
33. Gajhede, M., Schuller, D.J., Henriksen, A., Smith, A.T. & Poulos,
T.L. (1997) Crystal structure of horseradish peroxidase C at 2.15 A
˚
resolution. Nat. Struct. Biol. 4, 1032–1038.
34. Henriksen, A., Schuller, D.J., Meno, K., Welinder, K.G., Smith,
A.T. & Gajhede, M.M. (1998) Structural interactions between
horseradish peroxidase C and the substrate benzhydroxamic acid

determined by X-ray crystallography. Biochemistry 37, 8054–8060.
35. Henriksen, A., Smith, A.T. & Gajhede, M. (1999) The structures
of the horseradish peroxidase C-ferulic acid complex and the
ternary complex with cyanide suggest how peroxidases oxidize
small phenolic substrates. J. Biol. Chem. 274, 35005–35011.
36. Schuller, D.J., Ban, N., Huystee, R.B., Mcpherson, A. & Poulos,
T.L. (1996) The crystal structure of peanut peroxidase. Structure 4,
311–321.
37. Henriksen, A., Welinder, K.G. & Gajhede, M. (1998) Structure of
barley grain peroxidase refined at 1.9-A
˚
resolution. A plant per-
oxidase reversibly inactivated at neutral pH. J. Biol. Chem. 273,
2241–2248.
38. Mirza, O., Henriksen, A., Østergaard, L., Welinder, K.G. &
Gajhede, M. (2000) Arabidopsis thaliana peroxidase N, structure
of a novel neutral peroxidase. Acta Crystallogr. D. Biol. Crystal-
logr. 56, 372–375.
39. Østergaard, L., Teilum, K., Mirza, O., Mattsson, O., Petersen, M.,
Welinder, K.G., Mundy, J., Gajhede, M. & Henriksen, A. (2000)
Arabidopsis ATP A2 peroxidase. Expression and high-resolution
structure of a plant peroxidase with implications for lignification.
Plant Mol. Biol. 44, 231–243.
40. Nielsen,K.L.,Indiani,C.,Henriksen,A.,Feis,A.,Becucci,M.,
Gajhede, M., Smulevich, G. & Welinder, K.G. (2001) Differential
activity and structure of highly similar peroxidases. Spectroscopic,
crystallographic, and enzymatic analyses of lignifying Arabidopsis
thaliana peroxidase A2 and horseradish peroxidase A2. Biochem-
istry 40, 11013–11021.
41. Henriksen, A., Mirza, O., Indiani, C., Teilum, K., Smulevich, G.,

Welinder, K.G. & Gajhede, M. (2001) Structure of soybean seed
coat peroxidase, a plant peroxidase with unusual stability and
haem–apoprotein interactions. Protein Sci. 10, 108–115.
42. Poulos, T.L. & Kraut, J. (1980) The stereochemistry of peroxidase
catalysis. J. Biol. Chem. 255, 8199–8205.
43. Smith, A.T. & Veitch, N.C. (1998) Substrate binding and catalysis
in heme peroxidases. Curr. Opin. Chem. Biol. 2, 269–278.
44. Welinder,K.G.,Mauro,J.M.&Nørskov-Lauritsen,L.(1992)
Structure of plant and fungal peroxidases. Biochem. Soc. Trans.
20, 337–340.
45. Welinder, K.G. & Gajhede, M. (1993) Structure and evolution of
peroxidases. In Plant Peroxidases, Biochemistry and Physiology
(Welinder,K.G.,Rasmussen,S.K.,Penel,C.&Greppin,H.,eds),
pp. 35–42. University of Geneva, Switzerland.
46. Welinder, K.G. (1979) Amino acid sequence studies of horseradish
peroxidase. Amino and carboxyl termini, cyanogen bromide and
tryptic fragments, the complete sequence, and some structural
characteristics of horseradish peroxidase C. Eur. J. Biochem. 96,
483–502.
47. Mazza, G. & Welinder, K.G. (1980) Covalent structure of turnip
peroxidase 7. Cyanogen bromide fragments, complete structure
and comparison to horseradish peroxidase C. Eur. J. Biochem.
108, 481–489.
48. Johansson, A., Rasmussen, S.K., Harthill, J.E. & Welinder, K.G.
(1992) cDNA, amino acid and carbohydrate sequence of barley
seed-specific peroxidase BP 1. Plant.Mol.Biol.18, 1151–1161.
49. Gavel, Y. & von Heijne, G. (1990) Sequence differences between
glycosylated and non-glycosylated Asn-X-Thr/Ser acceptor sites,
implications for protein engineering. Protein Eng. 3, 433–442.
50. Morita, Y., Mikami, B., Yamashita, H., Lee, J.Y., Aibara, S.,

Sato, M., Katsube, Y. & Tanaka, N. (1991) Primary and crystal
structures of horseradish peroxidase isozyme E5. In Biochemical,
Molecular, and Physiological Aspects of Plant Peroxidases
(Lobarzewski, J., Greppin, H., Penel, C. & Gaspar, T., eds), pp.
81–88. University of Geneva, Switzerland.
51. Østergaard, L., Abelskov, A.K., Mattsson, O. & Welinder, K.G.
(1996) Structure and organ specificity of an anionic peroxidase
from Arabidopsis. FEBS Lett. 398, 243–247.
52. Carpin, S., Cre
`
vecoeur,M.,Greppin,H.&Penel,C.(1999)
Molecular cloning and tissue-specific expression of an anionic
peroxidase in zucchini. Plant Physiol. 120, 799–810.
53.Huh,G.H.,Lee,S.J.,Bae,Y.S.,Liu,J.R.&Kwak,S.(1997)
Molecular cloning and characterization of cDNAs for anionic and
neutral peroxidases from suspension-cultured-cells of sweet potato
and their differential expression in response to stress. Mol. General
Genet. 255, 382–391.
54. Koide, Y., Hirano, H., Matsuoka, K. & Nakamura, K. (1997) The
N-terminal propeptide of the precursor to sporamin acts as a
vacuole-targeting signal even at the C terminus of the mature part
in tobacco cells. Plant Physiol. 114, 863–870.
55. Theilade, B., Rasmussen, S.K., Rosenkrands, I., Frøkjær, H.,
Hejgaard, J., Theilade, J., Pihakaski-Maunsbach, K. & Mauns-
bach, A.B. (1993) Subcellular localization of barley grain peroxi-
dase BP2 by immuno-electron microscopy. In Plant Peroxidases,
Biochemistry and Physiology (Welinder, K.G., Rasmussen, S.K.,
Penel, C. & Greppin, H., eds), pp. 321–324. University of Geneva,
Switzerland.
6080 K. G. Welinder et al.(Eur. J. Biochem. 269) Ó FEBS 2002

56. Nakamura, K. & Matsuoka, K. (1993) Protein targeting to the
vacuole in plant cells. Plant Physiol. 101, 1–5.
57. Dombrowski, J.E., Schroeder, M.R., Bednarek, S.Y. & Raikhel,
N.V. (1993) Determination of the functional elements within the
vacuolar targeting signal of barley lectin. Plant Cell 5, 587–596.
58. Ruan, Y., Gilmore, J. & Conner, T. (1998) Towards Arabidopsis
genome analysis, monitoring expression profiles of 1400 genes
using cDNA microarrays. Plant J. 15, 821–833.
59. Audic,S.&Claverie,J.M.(1997)Thesignificanceofdigitalgene
expression profiles. Genome Res. 7, 986–995.
60. Ewing, R.M., Kahla, A.B., Poirot, O., Lopez, F., Audic, S. &
Claverie, J.M. (1999) Large-scale statistical analyses of rice ESTs
reveal correlated patterns of gene expression. Genome Res. 9, 950–
959.
61. Hiraga, S., Yamamoto, K., Ito, H., Sasaki, K., Matsui, H.,
Honma, M., Nagamura, Y., Sasaki, T. & Ohashi, Y. (2000)
Diverse expression profiles of 21 rice peroxidase genes. FEBS Lett.
471, 245–250.
62. Yamamoto, K. & Sasaki, T. (1997) Large-scale EST sequencing in
rice. Plant Mol. Biol. 35, 135–144.
63. Shannon, L.M., Kay, E. & Lew, J.Y. (1966) Peroxidase
isoenzymes from horseradish root. J. Biol. Chem. 241, 2166–
2172.
64. Shinmyo, A., Fujiyama, K., Kawaoka, A. & Intapruk, C. (1993)
Structure and expression of peroxidase isozyme genes in horse-
radish and Arabidopsis.InPlant Peroxidases, Biochemistry and
Physiology (Welinder, K.G., Rasmussen, S.K., Penel, C. &
Greppin, H., eds), pp. 221–228. University of Geneva, Switzer-
land.
65. Mohan, R., Bajar, A.M. & Kolattukudy, P.E. (1993) Induction of

a tomato anionic peroxidase gene (tap1) by wounding in trans-
genic tobacco and activation of tap1/GUS and tap2/GUS chi-
meric gene fusions in transgenic tobacco by wounding and
pathogen attack. Plant Mol. Biol. 21, 341–354.
66. Klotz, K.L., Liu, T.T., Liu, L. & Lagrimini, L.M. (1998)
Expression of the tobacco anionic peroxidase gene is tissue-
specific and developmentally regulated. Plant Mol. Biol. 36, 509–
520.
SUPPLEMENTARY MATERIAL
The following material is available from http://www.
blackwell-science.com/products/journals/suppmat/EJB/
EJB3311/EJB3311sm.htm
Fig. S1. Annotation of the N-terminus of Arabidopsis
peroxidases.
Fig. S2. Annotation of Arabidopsis peroxidase 5¢-UTRs.
Table S1. Primer sets used for the RT-PCR analysis of
Arabidopsis peroxidases.
Table S2. Nucleotide composition of observed and predic-
ted Arabidopsis peroxidase 5¢-UTRs.
Table S3. Amino acid sequence identity (%) of all mature
Arabidopsis peroxidase pairs.
Ó FEBS 2002 73 peroxidases from Arabidopsis (Eur. J. Biochem. 269) 6081