Báo cáo khoa học: Functional dissection of a small anaerobically induced bZIP transcription factor from tomato pdf

Functional dissection of a small anaerobically induced bZIP
transcription factor from tomato
Simone Sell and Reinhard Hehl
Institut fu
¨
r Genetik, Technische Universita
¨
t Braunschweig, Germany
A small anaerobically induced tomato transcription factor
was isolated from a subtractive library. This f actor, desig-
nated ABZ1 (anaerobic basic leucine zipper), is anaerobi-
cally induced in fruits, leaves a nd roots and encodes a
nuclear localized protein. ABZ1 shares close structural a nd
sequence homology with t he S-family of small basic leucine
zipper ( bZIP) transcription factors that a re implicated in
stress re sponse . N uclear localization o f A BZ1 is mediated by
the basic region and occurs under normoxic conditions.
ABZ1 binds to G-box-like target sites as a dimer. Binding
can b e abolished b y heterodimerization with a truncated
protein retaining the leucine zipper but lacking the DNA
binding domain. The protein binds in a sequence specific
manner to the C aMV 35S promoter which i s down regulated
when ABZ1 is coexpressed. This correlates with the anaer-
obic down regulation of the 35S promoter in tomato and
tobacco. These results may suggest that small bZIP proteins
are involved in the negative regulation of gene expression
under a naerobic c onditions.
Keywords: a naerobiosis; b ZIP; DNA b inding; Lycopersicon
esculentum; transcription factor.
Plant survival under adverse environmental situations is
largely dependent on their adaptation s trategies. Anaero-

biosis or low oxygen conditions occur when p lants are
subjected to flooding or to waterlogging of the soil. Under
these c onditions oxygen is rapidly consumed by micro-
organisms and plant roots [1]. Plants react to these
conditions with a variety of responses. To compensate the
decrease in energy production and the lack of NAD H
regeneration, the r ate o f glycolysis i s i ncreased and fermen-
tative pathways are i nduced [1]. Further more, plants can
respond to flooding with the induction of aerenchyma in the
root cortex and hyponastic growth to push their vital organs
above water level [2,3].
The reactions of plants towards a low oxygen environ-
ment en tail a significant reprogramming of gene expression
which comprises transcriptional induction and selective
translation of m RNAs for a naerobic proteins [ 4]. Plants
probably sense the lack of o xygen as an electron acceptor in
the mitochondria. Mitochondria are implicated in an early
response because they release calcium to the cytosol in
response to anaerobiosis [5]. The complete sign al transduc-
tion pathway has yet to be elucidated. Recent data suggest
that O
2
deprivation stimulates a G-protein signal transduc-
tion pathway that results in the induction of alcohol
dehydrogenase ( ADH) e xpression [6]. Other components of
the signal t ransduction pathway may c omprise 14-3-3
proteins, calcium dependent kinases and several transcrip-
tion factors [7–9]. ADH, one of the most e xtensively studied
genes that is induced during oxygen deprivation, is probably
induced by the t ranscription factor AtMYB2 in Arabidopsis

thaliana [10].
A c omprehensive analysis of low oxygen regulated gene
expression was recently reported for A. thaliana [11]. In a
microarray containing 3500 cDNA clones, 210 differentially
expressed genes were identified. Among these were 21
nonredundant down regulated genes. In contrast to tran-
scriptional induction and post-transcriptional regulation,
little is known about low o xygen mediated down regulation
of gene expression.
In the present work, a small basic leucine zipper (bZIP)
transcription factor (TF) designated ABZ1, was isolated
from a tomato cDNA library enriched for a naerobically
induced genes. This TF was stu died in detail. In addition to
binding site specificity, nuclear localization, iden tification of
DNA– a nd protein–protein binding domains it is shown
that efficient DNA binding of a h eterodimer requires two
DNA binding domains in the interacting proteins. The
putative role of ABZ1 in the anaerobic response p athway is
discussed.
Materials and methods
Tomato cDNA library construction and screening
A cDNA l ibrary from t omato, Lycopersicon e sculentum cv.
Micro-Tom [12], was generated in plasmid pSport1 using
the Invitrogen ( Karlsruhe, Germany) ÔSuperScript
TM
Plas-
mid System with GatewayÒ Technology for cDNA
Correspondence to R. Hehl, Institut fu
¨
r Genetik Technische

Universita
¨
t Braunschweig Spielmannstr. 7, D-38106 Braunschweig,
Germany. Fax: +49 531 391 5765, Tel.: +49 531 391 5772,
E-mail:
Abbreviations: ADH, alcohol dehydrogenase; as-1, activation
sequence-1; bZIP, basic leucine zipper; DAPI, 4¢-6-diamidino-2-
phenylindole; EMSA, electrophoretic mobility shift assay; GUS,
b-glucuronidase; LUC, luciferase; NLS, nuclear localization signal;
RBSS, random binding site selection.
Note: A webs ite is available at h ttp://www.tu-braunschweig.de/ifg/ag/
hehl
Note: The EMBL/GenBank accession number o f ABZ1 is AJ715788.
(Received 9 July 2004, revised 30 S eptember 2004,
accepted 4 October 2004)
Eur. J. Biochem. 271, 4534–4544 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04413.x
Synthesis and CloningÕ according to the protocol of the
supplier. Poly(A)+ RNA (5 lg) isolated from fruits, roots,
leaves, and stems of anaerobically induced tomato plants
was employed for cDNA synthesis. The plants w ere grown
in a greenhouse and were  3 m onths old. Anaerobic
incubations were carried out in an airtight glass container
(Merck, D armstadt, Germany) together with A naerocult A
(Merck) for 20 h in a light chamber (12 : 12 h light/
darkness). Screening of the library and all other recombin-
ant DNA work was done according to standard protocols
[13]. Sequence a nalysis of cDNA clones and other in vitro
constructs used in this work was performed by Seqlab
Company, Go
¨

ttingen, Germany. The sequences were ana-
lysed f or subc ellular l ocalization s ignals using
PSORT
at
Sequence comparisons
and alignments were performed at .
nih.gov/BLAST and />ClustalW.html. The phylogenetic tree was generated using
CLUSTALX
1.81 [14].
RNA isolation and Northern blot hybridizations
Total RNA was isolated according to a previously
published procedure [ 15]. To isolate poly(A)+ RNA,
1 m g total RNA was used with the O ligotex
TM
mRNA
Midi Kit (Qiagen, Hilden, Germany) according to the
manufacturer’s protocol. Northern blots w ere performed
according to standard protocols [13]. Total RNA (10 lg)
was used for RNA gel electrophoresis and radioactive
probes were generated usin g the HexaLabel
TM
DNA
Labeling Kit from MBI Fermentas (St. Leon-Rot,
Germany).
Recombinant protein production and purification
For the expression in and purification of recombinant
proteins from Escherichia coli the QIAexpressionist
TM
system from Qiagen was used. The ABZ1 c oding region
and two truncated derivatives from ABZ1 were PCR

amplified from a full size cDNA c lone using the following
primer pairs. Full size, ABZ1(1–138): 5¢-TATA
GGA
TCCATGTCACCTTTAAGGCAGAG-3¢ and 5¢-ATAT
CCCGGGTTAAAATTTAAACAATCCTG-3¢.N-ter-
minal deletion, ABZ1(47–138): 5¢-TAT
GGATCCATGC
TTCTGCAAGATTTGACAGG-3¢ and 5¢-ATAT
CCC
GGGTTAAAATTTAAACAATCCTG-3¢.C-terminal
deletion, ABZ1(1–100): 5¢-TATA
GGATCCATGTCACC
TTTAAGGCAGAG-3¢ and 5¢-A TA
CCCGGGTTATAA
ATACCTGAGCCTATCAGTC-3¢.
The BamHI and SmaI sites within the p rimers (under-
lined) w ere u sed t o directionally clone the amplified DNA
fragments into plasm id pQE-30. Prior to this, t he a mplified
DNA fragments w ere clon ed into pC RÒ2.1 (Invitrogen)
and sequenced. Recombinant pQE-30 plasmids were
transformed together with the repressor plasmid pREP4
(Qiagen) in BL21-CodonPlus(DE3)-RIL E. coli cells
(Stratagene, Amsterdam, t he Netherlands). Induction of
protein expression and purification over Ni-nitrilotriacetic
acid columns were performed according to The QIAex-
pressionist
TM
manual. Because the recombinant proteins
were mainly localized in the insoluble f raction, a previously
reported m ethod that includes steps for de- an d r e-naturing

of the protein was employed [16]. Protein c oncentrations
were determined according t o Bradford [ 17].
Electrophoretic mobility shift assays and random
binding site selection
Electrophoretic mobility shift assays (EMSA) were carried
out according to Ausubel et al. [18]. T he four probes used
for EMSA, RBSS1, 1.1, 5 and 6.1 are shown in Table 1.
These probes were used because they represent different
classes of binding sites. RBSS1-Mu, containing a mutation
in the A CGT core sequence (GGTTG
ATTAGGGAA),
was used a s a nonspecific competitor. All fragments are
bordered by primer binding sites 5¢-CAGGTCAGT
TCAGCGGATCCTGTCG-3¢ and 5¢-GCTGCAGTTG
CACTGAATTCGCCTC-3¢ that were also used for PCR
amplifications during the random binding site selection
(RBSS) assay. Random binding site selection was per-
formed according to Ausubel et al. [18]. Oligonucleotides
consisted of five rando m nucleotides 5¢ and 3¢ of the ACGT
core sequence bordered by the above mentioned primer
binding sites. Oligonucleotides were amplified in the pres-
ence of [
32
P]dCTP[aP]andincubatedwithrecombinant
ABZ1. After electrophoretic separation the bound oligo-
nucleotides were eluted from the gel and subjected to
another round of PCR amplification, ABZ1 binding, and
EMSA. Following five r ounds of selection the amplified
fragments were cloned into pCRÒ2.1 and sequenced.
Another f ragment used for EMSA was a 100 bp

fragment f rom the CaMV 35S promoter that was amplified
by PCR using the primers 5¢-TATGTCGACCGAG
GAACATAGTGGAAAAAG-3¢ and 5¢-ATAGTCGACT
GGGATTGTGCGTCATCCCTT-3¢.
Fragments for EMSA were amplified by PCR from
recombinant pCRÒ2.1 (RBSS1, 1.1, 5 6.1, and RBSS1-Mu)
and p RT103-GUS (b-glucuronidase) [19] in the presence of
[
32
P]dCTP[aP]. Binding reactions were carried out in 15 lL
10 m
M
Tris/HCl pH 7.5; 40 m
M
NaCl; 1 m
M
EDTA; 4 %
(v/v) glycerol; 1 0 m
M
2-mercaptoethanol; 10 m
M
dithio-
threitol; 5 m
M
phenylmethanesulfonyl fluoride; 2 lg/15 lL
poly(dI-dC) [20] for 30 min at room temperature. The
amounts of recombinant protein, radioactively labelled
Table 1. Sequences obtained from a random bi nding site selection
(RBSS) assay w ith ABZ1. RBSS1, 1.1, 5 a nd 6.1 are among the 17
selected bin ding sites. N ucleotide frequency at each position is shown

relative to the center o f the palindromic core. K ¼ G/T, N ¼ A/C/G/
T. Below the sequen ces the frequency of the nucleotides A, C, G and T
at each position of the 17 selected binding sites is shown. A co nsensus
sequence was derived from t hese frequencies.
–5 )4 )3 )2 )1 Core +1 +2 +3 +4 +5
RBSS1 G G T T G A C G T G G G A A
RBSS1.1 G G G C C A C G T G G T G T
RBSS5 T G C A C A C G T G T C G C
RBSS6.1 G G A T G A C G T G T A G G
A 2144–17–––1–364
C –1524 –17––– – 5 5 3
G 7 10 4 2 8 – – 17 – 16 12 6 5 6
T 85495 –––17– 5 3 1 4
Consensus K K N T K A C G T G G N N N
Ó FEBS 2004 Properties of a small bZIP protein from tomato (Eur. J. Biochem. 271) 4535
fragment and competitor DNA is given in the relevant
figure legends. A fter addition of 3 lL nondenaturing 5·
loading dye [50 m
M
EDTA pH 8.0; 50 m
M
Tris/HCl
pH 8.0; 50% ( v/v) glycerol; 1 2.5 mg/10 mL bromphenol-
blue; 12.5 mg/10 mL xylencyanol] the binding assay was
loaded on a native polyacrylamide g el (5–9%). Af ter
electrophoresis in 1· Tris/glycine the gel was dried under
vacuum and exposed to X-ray films.
Nuclear localization assays
Plasmid constructs for the analysis of nuclear localization
were generated by fu sing full size ABZ1 and t runcated

derivatives of ABZ1 in-frame upstream to the amino
terminal end of the uidA gene in pRT103-GUS [19]. The
ABZ1 coding region and three truncated derivatives fr om
ABZ1 were PCR amplified from a full s ize cDNA clone
using the f ollowing primer pairs. Abz1(1–138): 5¢-TA
TA
CTCGAGATGTCACCTTTAAGGCAGAG-3¢ and
5¢-ATAT
CCATGGAAAATTTAAACAATCCTGATG-3 ¢.
Abz1(1–44): 5¢-TATA
CTCGAGATGTCACCTTTAAG
GCAGAG-3¢ and 5¢-ATAT
CCATGGGCTTCTTCATC
CTCGATCGC-3¢. Abz1(1–24): 5¢-TATA
CTCGAGAT
GTCACCTTTAAGGCAGAG-3¢ and 5 ¢-ATAT
CCATG
GTCTCATCCATTCCTGCATAC-3¢. Abz1(45–138): 5¢-
TATA
CTCGAGATGCAGAAGCTTCTGCAAGATTT
GAC-3¢ and 5¢-ATAT
CCATGGAAAATTTAAACAA
TCCTGATG-3¢.
The XhoIandNcoI sites within the primers (underlined)
were used to directionally clone the amplified DNA
fragments into plasmid pRT103-GUS. The resulting clones
were subsequently sequenced.
Layers of onion epidermal c ells were placed on Mur-
ashige–Skoog media containing 3% (w/v) sucrose and
transformed by particle bombardment with recombinant

ABZ1-GUS constructs. Loading of the gold particles was
performed according to the CaCl
2
/spermidin protocol [21].
Particle bombardment of epidermal cell layers was per-
formed with 600–700 p.s.i. using t he Du Pont PDS-1000
particle delivery system [22]. After bombardment the tissue
was incubated at 24 °C for 24 h in a light chamber
(12 : 12 h light/darkness). The histochemical GUS assay
was performed by incubating the tissue i n 1 lgÆmL
)1
5-bromo-4-chloro-3-indolyl-b-glucuronic a cid (X-Gluc) in
50 m
M
NaPO
4
pH 7.0; 1 m
M
EDTA; 0.1% (v/v) Triton
X-100; 1 m
M
K-ferrocyanid; 1 m
M
K-ferricyanid for 4 h at
37 °C [ 23]. F or microscopy, tissues were transferred into
50 m
M
NaPO
4
pH 7.0. For staining of the nuclei, tissues

were treated by adding 1 lgÆmL
)1
4¢-6-diamidino-2-phen-
ylindole ( DAPI) and incubated for 15 min at room
temperature. Light microscopy was carried out with a
fluorescent microscope (Axioplan 2, Zeiss, Jena, Germany)
using a 360–370 nm filter for DAPI stained tissue.
Transient gene expression analysis in tobacco and
tomato leaves
For transient gene expression analysis four different effector
plasmids were constructed.
All c onstructs a re based on p lasmid pVKH-35S-pA1
(kindly provided by D. Großkopf, Max-Delbru
¨
ck-
Laboratorium, Ko
¨
ln, Germany). pVKH-35S-pA1 harbours
a CaMV 35S promoter and a poly(A)+ signal separated by
a cloning linker. The co ding region of ABZ1 was released
with BamHI and SmaI from a recombinant pCRÒ2.1
plasmid that harboured the complete ABZ1 coding
region (see above). This fragment was subcloned into
pVKH-35S-pA1 which was digested with Bam HI and
HindIII in which the HindIII site has been filled in. This
resulted in effector construct designated pVKH-35S-ABZ1-
pA1.
To generate pVKH-35S-AD-ABZ1-pA1 the activation
domain of GAL4 was amplified from p GADT7-Rec (Clon-
tech, H eidelberg, Germany). T he follow ing primers were

used to amplify the activation domain: 5¢-TATA
GGATCC
ATGGCCAATTTTAATCAAAGTGGGA-3¢ and 5¢-AT
AT
GGATCCCTCTTTTTTTGGGTTTGGTGGGGT-3¢.
The amplified fragment was cut with BamHI (restriction
site is underlined in the primers) and cloned into BamHI
digested pVKH-35S-ABZ1-pA1. The orientation of the
insert and the int egrity of t he construct was analysed by
restriction digest and sequencing.
The two effector plasmids pVKH-C1-ABZ1-pA1 and
pVKH-C1-AD-ABZ1-pA1 were generated b y replacing the
35S promoter with the C1 promoter from the sugar beet
cab11 gene. The 1097 bp long C1 promoter was released
with HindIII and BamHI from plasmid p C1L-1097 (kindly
provided by D. Stahl, Planta GmbH, Einbeck, Germany).
After HindIII digestion, the ends were filled in so that the
released fragm ent h arbours one blunt and one BamHI end.
This allowed the directional cloning of the C1 promoter into
pVKH-35S-ABZ1-pA1 and pVKH-35S-AD-ABZ1-pA1
from which the 35S p romoter was released with SacIand
BamHIandinwhichtheSacI site was filled in to generate a
blunt end.
As a reporter, a 35S-uidA construct w as generated by
removing the T ATA box fr om plasmid p BT10-TATA-
GUS [24] with NcoIandPstI and replacing i t with a 500 bp
NcoI/PstI CaMV 35S promoter fragment. The resulting
plasmid is designated pBT10-35S-GUS.
For transformation c ontrols, a luciferase gene was used
that was either expressed f rom t he 35S promoter in

pRT101-LUC [ 19] o r from the C1 promoter in pC1L- 1097.
For transformation, leaf discs with a diameter of 4 c m
were cut from tobacco leaves and placed on wet 3MM
Whatman paper. E quimolar amounts o f effector, reporter
and control plasmids were loaded onto gold particles
according to standard protocols [21]. Particle bombardment
was performed with 1100 p.s.i. using the Du Pont PDS-1000
particle delivery system [22]. After bombardment the t issue
was incubated at 24 °C for 24 h in a light chamber
(12 : 12 h light/darkness).
For protein extraction 0.3 g of tissue was homogenized
with liquid nitrogen a nd by addin g of 100 lLextraction
buffer (0.1
M
NaH
2
PO
4
pH 7.8; 1 m
M
dithiothreitol). After
centrifugationfor10minwith25500g at 4 °Cthe
supernatant was used for quantitative GUS and luciferase
assays. Protein concentrations were determined according
to Bradford [17]. The determination of GUS activity was
performed according to Jefferson & Jefferson et al. [25,26].
Four hundred and fifty microliters of GUS-reaction
buffer [1 m
M
4-methyl-umbelliferyl-b-

D
-glucuronide;
50 m
M
NaPO
4
pH 7.0; 10 m
M
EDTA; 0.1% (v/v) Triton
X-100; 0.1% (v /v) N-laurylsarcosine; 1 0 m
M
2-mercapto-
4536 S. Sell and R. Hehl ( Eur. J. Biochem. 271) Ó FEBS 2004
ethanol] were prewarme d to 37 °Cand50lLprotein
extract was added t o start the reaction. A blank value was
determined by transferring 50 lL of t he reaction after 1 min
into 950 lL stop-buffer (0.2
M
Na
2
CO
3
). Additional ali-
quots were transferred into stop-buffer after 20, 40 and
60 min, respectively. The extinction generated by the
reaction product 4-MU was measured in a spectral
photometer (Kontron Instruments, Eching, Germany,
SFM 25; excitation 365 nm; emission 455 nm) to determine
the relative fluorescence units of the generated 4-MU per
min and lg protein. These values were corrected with the

values obtained for the transformation control plasmid
expressing the luciferase gene [27]. For luciferase assays,
50 lL protein extract was transferred to 3 50 lLluciferase
buffer ( 25 m
M
glycylglycine; 15 m
M
MgSO
4
;1m
M
ATP
pH 7.8). After injecting 150 lL substrate s olution (0.2 m
M
luciferin in 2 5 m
M
gycylglycine pH 7.8) the emitted photons
were measured in a luminometer in a time interval of 10 s
(Berthold Lumat 9501; Bad Wildbad, Germany).
The determination of the relative GUS activity using the
luciferase values were performed as described previously
[24,28]. The resulting values were used to display the GUS
activity of the different transformations relative to the GUS
activity obtained without effector plasmids which was set to
100% (see below).
For the determination o f the relative expression strength
of the CaMV 35S promoter under aerobic and anaerobic
conditions in tobacco a nd tomato, a previously reported
approach was employed [29]. As a reporter a 35S-uidA
construct was generated by removing the TATA box from

plasmid pBT10-TATA-GUS [24] with NcoIandPstIand
replacing i t with a 500 bp NcoI/PstI CaMV 35S promoter
fragment. The resulting plasmid is designated pBT10-35S-
GUS. After bombardment of leaf discs w ith equimolar
amounts of pBT10-35S-GUS and p RT101-LUC [19], two
of the leaf discs from the s ame b ombardment were
incubated aerobically and the other two were incubated
anaerobically in an airtight glass container (Merck) together
with Anaerocult A (Merck). I ncubation was carried out for
24 h in a light chamber ( 12 : 12 h light/darkness). Luci-
ferase ac tivity, b-glucuronidase activity, and the d etermin-
ation of t he relative b-glucuronidase activity were
performed as described [29].
Results
ABZ1 belongs to a family of small bZIP transcription
factors and is anaerobically induced in fruits, roots,
and leaves
Suppression subtractive hybridization was employed for the
isolation of cDNA fragments for anaerobically induced
genes from tomato cv. Micro-Tom (S. Sell & R. Hehl,
unpublished observations). A f ull size clone was i solated for
a cDNA fragment that is homologous to bZIP transcription
factors. The cDNA is 1216 base pairs long, which includes a
poly(A) tail of 20 base pairs. Figure 1 shows the cDNA that
contains a 596 bp leader encoding a short 30 amino acid
long peptid e reminiscent of upstream open reading frames
found in other TFs [30,31]. The 596 bp leader is followed by
414bpcodingregionand186bp3¢ untranslated sequence.
The 4 14 bp coding region translates into a 138 amino acids
long protein with a proposed molecular mass of 15.4 kDa.

Figure 1 shows that t he protein harbours a basic r egion
followed by a leucin zipper consisting of five leucines and
one isoleucine that are spaced exactly by six amino acids.
The basic region also harbours a putative nuclear localiza-
tion signal (see below). The gene for t he small b ZIP protein
from tomato was designated ABZ1 for anaerobic basic
leucine bZIP.
Fig. 1. cDNA sequence and de duced amino
acid sequence of the ABZ1 gene. The138 amino
acid lo ng sequence of ABZ1 (nucleotide posi-
tions 597 –1010) harbours a basic leucine zip-
per domain ( bold ) that c on tains a putative
bipartite nuclear loc aliz ation signal (d ou ble
underlined). The single amino acids t hat are
part of the leucine zipper are underlined. The
30 amino a ci d long sequence e n coded by the
upstream open reading f ram e (nucleotide
positions 3 58–447) harbours several amino
acids (bold) t hat are conserved in o th er up-
stream open reading frames.
Ó FEBS 2004 Properties of a small bZIP protein from tomato (Eur. J. Biochem. 271) 4537
To investigate the spatial expression of A BZ1, RNA blot
hybridizations were carried out. Total RNA from aerobi-
cally and anaerobically treated organs from tomato was
hybridized w ith the ABZ1 cDNA fragment isolated in the
differential screen for a naerobically induced genes. Figure 2
shows that the gene is anaerobically induced in fruits, leaves,
and r oots of tomato. The ubiquitous nature of the anaerobic
induction of ABZ1 may suggest a more general role in
regulating anaerobic gene expression.

Phylogenetic analysis using the basic and leucine zipper
domains of 16 bZIP transcription factors shown in F ig. 3
indicates that ABZ1 is most closely related to BZI-4 from
tobacco [32], and belongs to a family of small bZIP proteins
that are o ften induced upon environmental stress [31]. For
example, maize LIP15 and rice LIP19 are low temperature
induced bZIP transcription factors that share 68.9%
sequence identity at the amino acid level [33,34]. BZI-4
from tobacco is transcribed specifically in the stamen, the
petals and the pistils of the tobacco flower [32].
The basic domain confers nuclear localization of ABZ1
For g ene expression re gulation ABZ1 needs to be imported
into the nucleus. Using bioinformatic tools it was found that
ABZ1 harbours a putative b ipartite nuclear localization
signal (NLS) i n its basic r egion between amino a cids 25 and
44 (Fig. 1). To determine w hether this region harbours a
functional NLS, nuclear localization was investigated with
fusion proteins using t he b-glucuronidase (uidA) reporter
gene. Fusion constructs were made with the whole 138
amino acid l ong protein, with the 24 and 44 a mino terminal
and 94 carboxy terminal amino acids, respectively. These
fusion constructs were transformed by p article bombard-
ment into onion epider mal cells. Figure 4 shows that the
majority of the GUS protein that is fused with the c omplete
138 or the amino terminal 4 4 amino acids of ABZ1 localizes
to the nucleus while the GUS protein alone or fused either
with the 24 amino terminal or with the 94 c arboxy terminal
amino acids of ABZ1 does not localize to t he nu cleus. T his
indicates that ABZ1 harbours the signal for nuclear
localization and that a functional NLS is localized within

the basic region between amino acids 25 and 44. Further-
more, nuclear localization was achieved under aerobic
conditions.
ABZ1 binding specificity and dimer formation
Transcription factors of the bZIP f amily are known to bind
to G-box like sequences [35]. T o investigate t he binding
specificity of ABZ1, the protein was expressed in E. coli as a
His-tag f usion protein. The purified protein was employed
in a binding site selection e xperiment i n which 17 putative
binding sites were i dentified. Table 1 shows the sequence of
four binding sites and the frequency of the nucleotides at
each position of the 17 selected binding s ites. Figure 5A
shows EMSA for the four individual binding sites RBSS1,
1.1, 5 and 6.1 (Table 1). These four sites are efficiently
and specifically bound by ABZ1 while the sequence
GGTTGATTAGGGAA that harbours a mutation in the
Fig. 2. Anaerobiosis specific e xpression of ABZ1, in t omato. Total
RNA from fruits (lanes 1 and 2), leaves (lanes 3 and 4), and roots (lanes
5 and 6) that were either prepared from aerobic organs (lanes 1, 3, and
5) or anaerobically incubated organs (lanes 2 , 4, and 6), w ere hybrid-
ized with a cDNA f ragment from the ABZ 1 gene. A single 1.2 k b
transcript hybridizes with th e probe. Staining o f the gels prior to
blotting indicates e qu al loading of t he RNA.
Fig. 3. Phylogenetic re lationship of ABZ1 with othe r bZIP proteins. A
phylogenetic tree was constructed by the neighbor joining method with
the basic leucine z ipper region of 1 6 different bZIP proteins. Zm, Zea
mays;Os,Oryza sativa;Nt,Nicotiana tabacum;Le,Lycopersicon
esculentum;Am,Antirrhinum majus;Pc,Petroselinum crispus;At,
Arabidopsis thaliana. The bZIP p roteins compared are B ZI-4, BZI-3,
and BZI-2 [32], tbz17 [53], bZIP911 and bZIP910 [ 30], CPRF6 [54],

mLIP15 [33], LIP19 [34], BZI-1 [55]. The five most similar Arabidopsis
proteins were also included and the Ar abidop sis geno me iden tificatio n
number provided. At3g62420 corresponds to AtbZIP53, At1g75390 to
AtbZIP44, At2g18160 to AtbZIP02, and At4g34590 to AtbZIP11 [31].
ABZ1 is underlined. The scale represents the frequen cy of amino acid
changes. Bootstrap values are indicated.
4538 S. Sell and R. Hehl ( Eur. J. Biochem. 271) Ó FEBS 2004
G-box core sequence (RBSS1-Mu) and was used as an
unspecific competitor in E MSAs i s not bound by ABZ1
(Fig. 5 A; u. Cmp). EMSAs with RBSS1 were subsequently
used to analyse specific binding conditions.
To confirm t hat the DNA bind ing d omain resides in the
basic r egio n, an N-terminally deleted p rotein was expressed
in E. coli. The deleted protein comprises amino acids 47–
138. Figure 5B shows that the wild type ABZ1 binds
efficiently to the RBSS1 sequence while the a mino termin-
ally deleted protein does not bind. This shows that the
binding domain of ABZ1 resides in the amino terminal 46
amino acids.
Because both proteins harbour the leucine zipper domain,
it was analysed if they interact with each other. If interaction
of the full size and the amino terminally deleted protein leads
to a heterodimer that binds DNA, a faster migrating
complex would be expected. Surprisingly, addition of the
truncated ABZ1 protein abolishs binding of full size ABZ1 in
a concentration d ependent manner (Fig. 5B). This indicates
that in vitro binding of a dimer requires the presence of a
DNA b inding domain in each i nteracting protein.
This was f urther investigated using a carboxy terminal
deletion of ABZ1. This protein harbours the first 100 amino

acids including basic a nd leucine zipper domains. Figure 5C
shows that binding of this protein yields a faster migrating
complex ( C2) compared to the full size ABZ1 (C1). W hen
both proteins are added in equimolar concentrations a third
complex i s observed that shows an intermediate migrating
behaviour (Fig. 5C; C1+2). This complex is interpreted to
be caused by heterodimer formation between full size ABZ1
and the carboxy terminally deleted ABZ1.
To summarize, ABZ1 binds to RBSS1 as a dimer and
efficient DNA binding requires a DNA binding domain in
both interacting proteins. This result may have important
implications for the regulatory properties of small bZIP
transcription factors.
ABZ1 binds to the CaMV 35S promoter which is
anaerobically down regulated in tobacco and tomato
One well known target of bZIP transcription factors i s the
CaMV 35S promoter [36]. Figure 6A,B s hows that recom-
binant ABZ1 bin ds to a 100 bp fragment from t he CaMV
35S promoter which harbours three potential binding sites
for ABZ1. One of the t hree putative binding sites w ithin
this fragment is the ac tivation s equence-1 ( as-1) b etween
positions )65 and )8 5 that consists of two imperfect
palindromes, with t he palindromic centers spaced by 12 bp
and which is known to b e bound by different tobacco bZIP
TFs [37]. The random b inding site selection e xperiment
indicates a high similarity between RBSS6.1 and the second
imperfect palindrome of as-1.
EMSA analysis with recombinant ABZ1 revealed three
shifted complexes of which two complexes can be com-
pletely competed w ith RBSS1 (Fig. 6A). The three shifted

complexes observed may be due to differential occupation
of ABZ1 binding sites and may represent different n umbers
of ABZ1 proteins bound to the promoter f ragment. These
results show that ABZ1 can also bind to the CaMV 35S
promoter.
The binding of the anaerobically induced ABZ1 tran-
scription factor to the CaMV 35S promoter indicates that
the 3 5S promoter may be r egulated under anaerobic
conditions. To investigate this proposal, transient gene
expression analyses were performed by transforming a
CaMV 35S promoter uidA reporter gene construct into
tobacco a nd tomato leaves. Subsequent to particle bom-
bardment, leaves w ere incubated under aerobic a nd anaer-
obic c onditions followed by a quantitative GUS assay. As
shown in Fig. 6C, in both host tissues, expression of the 35S
promoter is significantly lower under anaerobiosis when
compared with expression und er aerobic c onditions. In
tobacco, anaerobic expression is only 14% relative to
aerobic expression while in tomato the difference between
anaerobic and aerobic expression is less stringent (52%,
Fig. 6C).
ABZ1 down regulates the CaMV 35S promoter
in a transient gene expression assay
The effect of ABZ1 on gene expression of the CaMV 35S
promoter was analysed with transient expression assays
conducted by coexpressing the ABZ1 protein together with
the CaMV 35S driven uidA (GUS) gene in tobacco leaves.
As a transformation standard, a luciferase gene under the
control of t he sugar beet cab11 promoter was employed.
This promoter does not harbour G-box binding sites and

confers reporter gene expression in tobacco leaves (D. Stahl,
Fig. 4. The basic region of ABZ1 is required for nuclear localization. Fusion gene constructs expressing parts of or the whole ABZ1 protein fused to
the b-glucuronidase ( uidA) reporter g ene were transformed by particle bombardment into o nion epidermal cells. A s indicated, the c onstructs
express amino acids 1–24, 1–44, 1–138, and 45–138 from the ABZ1 protein fused in-frame with the uidA gene. Transformed cells were subjected to
histochemical GUS staining and t o a DAPI stain ing of the nuc leus.
Ó FEBS 2004 Properties of a small bZIP protein from tomato (Eur. J. Biochem. 271) 4539
personal communication). F igure 7 shows that the coex-
pression of ABZ1 with the 35S-uidA construct leads to
down regulation o f GUS expression compared to the 35S-
uidA construct alone (compare 35S-uidA with 35S-ABZ1/
35S-uidA, Fig. 7A). Expression of the 35S promoter is
about 40% reduced in the p resence o f A BZ1 than without.
When a f usion c onstruct b etween ABZ1 and the activation
domain of GAL4 is c oexpressed with the 35S-uidA
construct, expression is higher than ob served with ABZ1
(compare 35S-ABZ1/35S-uidA with 35S-AD-ABZ1/35S-
uidA, Fig. 7A).
To minimize possible a utoregulatory effects o f ABZ1 on
its own expression this experiment was repeated b y expres-
sing ABZ1 with the sugar beet cab11 promoter which is void
of putative ABZ1 binding sites. Figure 7B shows that
similar results were obtained compared t o ABZ1 expression
with the 35S promoter. Coexpression of ABZ1 with the 35S-
uidA construct leads t o down regulation of GUS expression
compared to the 35S-uidA construct alone (compare 35S-
uidAwithC1-ABZ1/35S-uidA, Fig. 7B). Expression of the
35S promoter is again about 40% reduced in the presence of
ABZ1 th an without. When a fusion construct between
A
B

C
Fig. 5. ABZ1 binding specificity and dimer formation. Electrophoret ic mobility s hift assays with recombinant full s ize A BZ1(1–138) and two
truncated derivatives harbouri ng amino a cids 47–138 a nd 1–100, respectively. Shifte d complexes (C) and free probe (P) a re indicated. (A) Four
sequences (probes) derived from a random binding site selection assay (Table 1) were radioactively labelled, and 0.1 ng (4 · 10
3
c.p.m.) were either
incubated with (+) or without ( –) ABZ1. As indicated (+/–) specific ( RBSS1) or unspecific (u . Cmp) competitor was ad ded in a 1 · 10
5
molar
excess and s ep arated on a n onde naturing polyacrylamide gel ( 10% ). (B) Increasing amounts of truncated derivative of A BZ1 harbouring amino
acids 47–138 interferes with DNA b inding of full size ABZ1. One microgram of protein (+) and increasing amounts of truncated ABZ1 (1.5 lg,
2.1 lg, and 2.5 lg, designated by the elongated triangle) was incubated with 0.05 ng (2 · 10
3
c.p.m.) radioactively labeled RBSS1 fragment (P) and
separated on a nondenaturing p olyacrylamide gel (9%). The complex C1 decreases when truncated A BZ1 lacking the first 46 amino acids is
incubated in i ncreasing concentrations t ogether with fu ll size ABZ1. (C) D imer formation b etween full s ize ABZ1 an d a truncated derivative o f
ABZ1 harbouring amino acids 1–100. One microgram of ABZ1 (+) and 770 n g truncated ABZ1 (+) was incubated w ith 0.1 ng (4 · 10
3
c.p.m.)
radioactively l abeled RBSS1 fragment and sep arated o n a nondenaturing polyacrylamide gel (9%). A novel c omplex (C1+2) is observed when
truncated and full size ABZ1 are incubat ed simultaneously with the radioactive probe.
4540 S. Sell and R. Hehl ( Eur. J. Biochem. 271) Ó FEBS 2004
ABZ1 and the activation d omain of GAL4 is coexpressed
with the 35S-uidA construct, expression is higher than
observed w ith A BZ1 (compare 35S-ABZ1/35S-uidAwith
C1-AD-ABZ1/35S-uidA, Fig. 7B).
In summary, these two independent sets of experiments
support t he notion t hat the anaerobically induced ABZ1
transcription factor contributes to the anaerobic down
regulation of the 35S CaMV promoter.

Discussion
Anaerobic gene expression regulation
The primary plant s tress in flooded or compressed soils is
conferred by oxygen limitation that i s most apparent in
below ground tissue. Plants respond to oxygen limitation
with a significant r eprogramming of gene expression. These
responses usually permit a p rolonged survival under t hese
adverse conditions. Gene expression regulation involves
various mechanisms. Many genes are induced by
transcription factors. In Arabidopsis thaliana for example,
this is achieved b y the low oxygen induction of the AtMYB2
transcription f actor which leads t o the enhanced expression
of the ADH1 gene [10]. The extent of post-transcriptional
regulation is best illustrated when t he limited number of
anaerobic proteins detected is compare d to the large number
of genes t hat a re still transcribed under low oxygen
conditions [4,11,38,39]. It has long been observed that low
oxygen conditions suppress t he tr anslation of the majority
of mRNAs and increase translation o f a particular subset
corresponding to anaerobic proteins. This may be related to
impaired ribosomal RNA transcription and ribosomal
protein synthesis under oxygen deprivation [40–42]. The
analysis of ribosome loading patterns indicated t hat trans-
lational control of anaerobic genes occurs at the initiation
and postinitiation phases in a message-specific manner [43].
Another post-transcriptional mechanism of low oxygen
regulated g ene expression is the increased splicing e fficiency
of specific introns [44,45].
In the present study a s mall anaerobically induced bZIP
transcription factor designated ABZ1 was identified from

tomato. ABZ1 binds to the CaMV 35S promoter (Fig. 6A).
AB
C
Fig. 6. Binding of ABZ1 to the CaMV 35S promoter and down regu-
lation under anaerobic conditions. (A) An electrophoretic mobility shift
assay with a radio actively labeled 100 bp fragm ent from the CaMV
35S promoter is shown. Lanes 1–4 harbour the radioac tive probe (P)
while in lane s 2–4, 500 ng recombinant ABZ1 was added to the probe
(0.3 ng, 2 · 10
3
c.p.m.) resulting in three shifted c omplexes (C1, C2,
and C3). Spe cific com petition w as ac hieved with a 4 · 10
3
molar exc ess
of fragment RBSS1 (lane 3). The un specific co mpetitor in which the
ACGT core sequence of fragment RBSS1 was altered to ATTA did
not compete for binding when added in a 4 · 10
3
molarexcess(lane4).
(B) The sequence of the 100 bp fragment from the 35S p romoter used
for EMSA is shown. Putative ABZ1 bind ing sites are underlined. The
as-1 element between positions )65 and )85 is indic ated. (C) Expres-
sion of a 35S-uidA promoter reporter gene c onstruct after transien t
bombardment of t obacco and t omato leaves under aero bic and
anaerobic conditions. The expression strength under anaerobic c on-
ditions is displayed relative to the e xpression under aerobic conditions
(100%). The me an values were derived from seven (tobacco) and s ix
(tomato) measurements.
Fig. 7. Transient gene expression analysis using reporter and effector
gene constructs in particle bombardments on tobacco l eaves. Reporter

construct 35S-uidA harbours the b-glucuronida se gene under the
control of t he CaMV 35 S promoter. (A) Relative e xpression s trength
of the 35S promoter in the presence a nd absence of effecto r c onstructs
under t he control o f the 35S promoter. Exp ression strength is shown
relative to the 35S-uidA e xpression (100 %). The mean value was
derived from 10 (35S-uidA), 10 (35S-uidA + 35S-ABZ1), and 12 (35S-
uidA + 35S-AD-ABZ1) measurements, respectively. (B) Relative
expression strength of the 35S promoter in the presence and absence o f
effector c onstruct s under the control o f the C1 promoter. Expressio n
strength is shown relative to the 35S-uidA expression (100%). The
mean value was derived from eight (35S-uidA), seven (35S-uidA+C1-
ABZ1), and e ight (35S-uidA + C1-AD-ABZ1) measurements,
respectively.
Ó FEBS 2004 Properties of a small bZIP protein from tomato (Eur. J. Biochem. 271) 4541
The p romoter activity is reduced under anaerobiosis in
tobacco and tomato (Fig. 6 C). A lthough m any b ZIP
transcription factors have been isolated from plants, their
role in gene expression regulation under low oxygen
conditions has not been analysed extensively. Previously,
de Vetten & F erl isolated a G-box binding protein f rom
maize, GBF1, which is anaerobically induced [46]. The main
structural differences b etween GBF1 and ABZ1 are t he size
(377 amino acids for GBF1 vs. 138 for ABZ1) a nd a proline
rich region at the N terminus of GBF1. The proline rich
region of GBF1 may i ndicate that this protein is a
transcriptional activator [47]. A second anaerobically
induced G-box binding factor from maize, mLIP15, is
structurally more similar to ABZ1 because i t is 135 amino
acids long and also lacks a p roline rich region a t its N
terminus [33]. Both maize G-box binding factors were

shown t o interact w ith the maize ADH1 promoter, which is
anaerobically induced [33,46]. I t may be conceivable that in
maize GBF1 acts as a t ranscriptional activator while
expression of mLIP15 may modulate or r epress anaerobic
expression by competing or i nteracting with GBF1. Small
bZIP proteins may be one of the c omponents of the cellular
machinery that contribute to t he low oxygen mediated
down regulation of gene expression.
Functional dissection of ABZ1
The ABZ1 transcr iption factor isolated i n this study was
extensively a nalysed u sing biochemical approaches.
Although the basic and leucine zipper domain are often
assumed to be the DNA binding and dimerization domains,
the p resent study confirms this experimentally (Fig. 5). The
nuclear localization signal r esides within the basic domain
required for DNA binding. Some bZIP factors are regulated
by a subcellular l ocalization mechanism i n response t o
environmental cues. For example nuclear import of the
parsley b ZIP factor C PRF2 is light mediated [48]. Cyto-
plasmatic retention of the bZIP factor RSG is mediated by a
14-3-3 protein which has been suggested to modulate the
endogenous amounts o f g ibberellins through t he control of
a gibberellic acid biosynthetic enzyme [49]. In the study
presented here no evidence for cytoplasmatic retention of
ABZ1 was found for the full size ABZ1 and the protein is
readily detected in the nucleus under aerobic conditions
(Fig. 4 ).
Interestingly, the binding of ABZ1 to its t arget sequence
can be a bolished with i ncreasing amounts of a truncated
ABZ1 that lacks the DNA binding domain (Fig. 5B).

Therefore, efficient DNA binding requires that the dimeri-
zation occurs with another bZIP factor h arbouring a basic
DNA binding domain. Whether heterodimerization of
ABZ1 to other bZIP factors occurs, has not been analysed
directly. However, because its closest relative BZI-4 hetero-
dimerizes with BZI-1 [32] it may be conceivable that ABZ1
can also f orm heterodimers. Remarkably, no other bZIP
transcription factor was isolated in a yeast two hybrid screen
(S. S ell & R. Hehl, unpub lished observations). This may
either relate to an insufficient number of primary clones or to
the fact that the mRNA u sed for constructing the prey library
was i solated f rom a naerobic t issue and may not contain
transcripts f or other bZIP factors because their exp ression
may be down regulated under anaerobic conditions.
Because ABZ1 is able to bind to the 35S pro moter which
is down r egulated under a naerobic conditions and in the
presence of ABZ1 in cobombardment analyses, this m ay
suggest t hat ABZ1 e ither competes with other bZIP factor s
for the same binding sites or that heterodimerization also
down regulates target gene expression. In mammalian
systems heterodimer formation of a bZIP factor with
another l eucine zipper containing transc ription f actor
results in down regulation of t arget gene e xpression [50].
To date several attempts to generate t ransgenic tomato or
tobacco plants that o verexpress ABZ1 have failed. Reverse
genetic approaches to analyse the role of small bZIP
proteins in anaerobic gene expression may be more readily
carried out in A. t haliana.
Therefore, a screen for A. thaliana homologs w as
performed using

TAIR BLAST
[51]. ABZ1 is closely related
to the four bZIP transcription factors AtbZIP53
(At3g62420; 53% identity), AtbZIP44 (At1g75390; 42%
identity), AtbZIP02 (At2g18160; 50% identity), a nd Atb-
ZIP11 (At4g34590; 58% identity). Recently, dat a on
AtbZIP02 and AtbZIP11 suggested that these small bZIP
factors bind to the sequence ACTCAT and may act as
transcriptional activators under hypoosmotic conditions
[52]. It may be very interesting t o learn how small b ZIP
proteins are involved in transcriptional activation. This may
relate to the position of the ci s-regulatory element in the
promoter or to the presence of interacting proteins that
contribute a transcription activation domain.
Acknowledgements
This work was supported by a grant t hrough t he ÔForschungsschwer-
punkt Agrarbiotechnologie d es Landes N iedersachsen (VW-Vo rab)Õ.
We are grateful to Ralf R. Mendel for critical reading of the manuscript
andtoRobertHa
¨
nsch for advice using the particle delivery system. We
would l ike to t hank Jo
¨
rn Petersen for h elp with t he phylogenetic
analysis.
References
1. Drew, M.C. (1997) Oxygen deficiency and root metabolism: I njury
and acclimation under hypoxia and anoxia. Annu.Rev.Plant
Physiol. Plant. Mol. Biol. 48 , 223–250.
2. Colmer, T .D. (2003) A erenchyma and a n inducible ba rrier to

radial oxygen loss facilitate root a eration in u pland, paddy
and deep-water rice (Oryza sativa L.). Ann. Bot. (Lon d.) 91, 301–
309.
3. Cox, M.C., Millenaar, F.F., Van Berkel, Y.E., Peeters, A.J. &
Voesenek, L .A. (2003) Plant movement. Subme rgence-induced
petiole elongation in Rumex palustris depends on hyponastic
growth. Plant Physiol. 132, 2 82–291.
4. Sachs, M.M., Freeling, M. & O kimoto, R. (1980) The anaerobic
proteins of maize. Cell 20, 761–767.
5. Subbaiah, C.C., Bush, D.S. & Sachs, M.M. (1998) Mitochondrial
contribution to the anoxic Ca
2+
signal in maize suspension-cul-
tured cells. Plant Physiol. 118, 7 59–771.
6. Baxter-Burrell, A., Yang, Z., Springer, P.S. & Bailey-Serres, J.
(2002) RopGAP4-dependent Rop GTPase rheostat control
of Arabidopsis oxygen deprivation tolerance. Science 29 6, 2026–
2028.
7. Lu, G., DeLisle, A.J., d e Vetten, N .C. & Ferl, R.J. (1992) Brain
proteins in plants: an Arabidopsis homolog t o n eurotran smitter
pathway activators is p art of a DNA binding complex. Proc. Natl
Acad. Sci. USA 89, 11490–11494.
4542 S. Sell and R. Hehl ( Eur. J. Biochem. 271) Ó FEBS 2004
8. Lu, G., Sehnke, P.C. & Ferl, R.J. (1994) Phosphorylation and
calcium binding properties of an Arabidopsis G F14 brain protein
homolog. Plant Ce ll 6, 501–510.
9. Ferl, R.J. (1996) 14-3-3 proteins and signal t ransduction. Annu.
Rev. Pla nt Physiol. Plant Mol. Biol. 47, 49–73.
10. Hoeren, F.U., Dolferus, R., Wu, Y., Peacock, W.J. & Dennis, E.S.
(1998) Evidence fo r a ro le for AtMYB2 i n the induction o f the

Arabidopsis alcohol dehydrogenase gene(ADH1)bylowoxygen.
Genetics 149 , 479–490.
11. Klok, E.J., Wilson, I.W., Wilson, D., Chapman, S.C., Ewing,
R.M., Somerville, S.C., Peacock, W.J., Dolferus, R. & Dennis,
E.S. (2002) Expression profile analysis of the l ow-oxygen response
in Arabidopsis root cultures. Plant Ce ll 14, 2 481–2494.
12. Meissner, R ., Jacobson, Y., Melamed, S., L evyatuv, S., S halev,
G.,Ashri,A.,Elkind,Y.&Levy,A.(1997)Anewmodelsystem
for tomato g enetics. Plant J. 12 , 1465–1472.
13. Sambrook, J ., Fritsch, E.F. & Maniatis, T. (1989) Molecular
Cloning: a L aboratory Manual, 2nd edn. Cold Spring Harbor
Laboratory Press, New York.
14. Thompson, J.D., Gibson, T.J., P lewniak, F., Jeanmougin, F. &
Higgins, D .G. (1997) T he CLUSTAL_X w indows interface: flex-
ible strategies fo r multiple sequence alignment aided by qu ality
analysis tools. Nucleic A cids Res. 25 , 4876–4882.
15. Meyer-Gauen, G., Herbrand, H., Pahnke, J., Cerff, R. & Martin,
W. (1998) Gene st ru ctur e, expression in Escherichia coli and bio-
chemical properties of the NAD+-d ependen t glyceraldehyde-3-
phosphate dehydrogenase from Pinus sylvestris chloroplasts. Gene
209, 167–174.
16. Schmitz, M.L. & Baeuerle, P.A. (1997) Bacterial e xpression, pur-
ification, and potential use of His-tagged G AL4 fusion proteins.
Methods Mol. Biol. 63, 129–137.
17. Bradford, M.M. ( 1976) A rapid and sensitive me thod for the
quantitation of microgram quantities of protein utilizing the
principle of p rotein -dye binding. Anal. B iochem. 7, 248–254.
18. Ausubel, F.M., Brent, R., Kingston, R.E., Moor e, D.D.,
Seidman, J.G., Smith, J .A. & St ruhl, K. ( 1988) Current P rotocols
in Molecular Biology, Greene and Wiley Interscience, New York.

19. To
¨
pfer, R., Maas, C., Horicke-Grandpierre, C., Schell, J. &
Steinbiss, H.H. (1993) Expression vectors for high-level gene
expression in dicotyledonous and monocotyledonou s plants.
Methods En zymol. 217, 67–78.
20. Schindler, U. & Cashmore, A.R. (1990) Photoregulated gene
expression may involve ubiq uitous DNA b inding pro teins. EMBO
J. 9, 3415–3427.
21. Klein,T.M.,Gradziel,T.,Fromm,M.E.&Sanford,J.C.(1988)
Factors influencing gene delivery into Zea mays cells by high-
velocity mi croprojectiles. Bi o/Technology 6, 559–563.
22. Sanford, J.C. (1988) The biolistic process. Trends Biotechnol. 6,
299–302.
23. Isono, K., Yamamoto, H., Satoh,K.&Kobayashi,H.(1999)An
Arabidopsis cDNA enc oding a DNA-bindin g protein that is highly
similar to the DEAH family of RNA/DNA helicase genes. Nucleic
Acids Res. 27, 3728–3735.
24. Sprenger-Haussels, M. & Weisshaar, B. (2000) Transactivation
properties of parsley proline-rich bZIP transcription factors. Plant
J. 22, 1–8.
25. Jefferson, R.A., Kavanagh, T.A. & Bevan, M.W. (1987) GUS
fusions: b-glucuronidase as a sensitive and versatile gene fusion
marker in h igh er plants. EMBO J. 6, 3901–3907.
26. Jefferson, R.A. (1987) Assaying chimaric genes in plants: the GUS
gene fusion system. Plant Mol. Biol. Report 5, 387–405.
27. deWet, J .R., Wood, K.V., DeLuca, M., Helinski, D.R. & Sub-
ramani, S. (1987) Firefly luciferase gene: structure and expression
in mammalian cells. Mol. Cell. Biol. 7, 725– 737.
28. Schledzewski, K. & Mendel, R .R. (1994) Quantitative transient

gene expression: c ompariso n of t he promoter for m aize
polyubiquitin1, rice actin1, maize derived Emu and CaMV 35S in
cells of bareley, maize and tobacco. Transgenic Res. 3, 249 –255.
29. Geffers, R., Sell, S., Cerff, R. & Hehl, R . (2001) The T ATA box
and a Myb binding site are essential for anaerobic expression of a
maize GapC4 minima l promoter i n toba cco. Biochim. Biophys.
Acta 1521, 120–125.
30. Martinez-Garcia,J.F.,Moyano,E.,Alcocer,M.J.&Martin,C.
(1998) Two bZIP proteins from Antirrhinum flowers preferentially
bind a hybrid C-box/G -box motif and help to d efine a new s ub-
family of bZIP transcription f actors. Plant J . 13, 489 –505.
31. Jakoby, M ., Weisshaar, B ., Dro
¨
ge-Laser, W ., Vicente-Carbajosa,
J., Tiedemann, J., Kroj, T. & Parcy, F. (2002) bZIP transcription
factors in Arabidopsis. Trends Plant Sci. 7, 106–111.
32. Strathmann, A ., Kuhlmann, M ., Heinekamp, T . & Dro
¨
ge-Laser,
W. (200 1) BZI-1 specifically heterodimerises w ith the tobacco
bZIP transcription f a ctors BZI-2, BZI-3/TBZF and BZI-4, and
is functionally involved in flower d evelopm ent. Plant J. 28, 397 –
408.
33. Kusano, T., Berberich, T. , Harada, M., Suzuki, N . & Sugawara,
K. (1995) A maize DNA-binding factor with a bZIP motif is
induced by l ow temperature. Mol. Gen. Genet. 248, 507–517.
34. Aguan, K ., Sugawara, K., Suzuki, N. & K usano, T. (1993) Low-
temperature-dependen t expression of a rice gene e ncoding a p ro-
tein with a leucine-zipper motif. Mol. Gen. Genet. 240, 1–8.
35. Izawa, T., Foster, R. & Chua, N .H. (1993) Plant bZIP protein

DNA binding spe cificity. J. Mol. Biol. 23 0, 1131–1144.
36. Lam, E. & L am, Y .K. ( 1995) B inding site re quirements and dif-
ferential representation of TGF factors in nuclear ASF-1 activity.
Nucleic A cids Res. 23, 3778–3785.
37. Krawczyk, S., Thurow, C., Niggeweg, R . & Gatz, C. (2002)
Analysis of the spacing between the two palindromes of activation
sequence-1 with respect to binding to different TGA f actors and
transcriptional a ctivation potential. Nucleic Acids Res. 30, 775–
781.
38. Russell, D.A. & Sachs, M.M. (1992) Protein synthesis in maize
during anaerobic and heat stress. Plant Physiol. 99, 6 15–620.
39. Germain, V., Ricard, B. , Raymond, P. & Saglio, P.H. (1997) The
role of sugars, hexokinase, and sucrose synthase in the determi-
nation of hypoxically induced tolerance to anoxia in tomato roots.
Plant Physiol. 114, 167–175.
40. Bailey-Serres, J. & Freeling, M. (1990) Hypoxic stress-induced
changes in ribosomes of maize seedling roots. Plant Phys. 94,
1237–1243.
41. Fennoy, S.L. & Bailey-Serres, J. ( 1995) Post-transcriptional reg-
ulation of gene e xpression in oxyg en-deprived roots of maize.
Plant J. 7, 287 –295.
42. Fennoy, S., Jayachandran, S. & Bailey-Serres, J. (1997) RNase
activities are reduced concomitantly with conservation of total
cellular RNA and ribosomes in O
2
-deprived seedling roots of
maize. Plant Physiol. 115, 1109–1117.
43. Fennoy, S.L., Nong, T. & Bailey-Serres, J. (1998) Transcriptional
and post-transcriptional pro cesses regulate gen e expression in
oxygen-deprived r oots of m aize. Plant J. 15, 727 –735.

44. Ko
¨
hler,U.,Donath,M.,Mendel,R.R.,Cerff,R.&Hehl,R.(1996)
Intron-specific stimulation of anaerobic g ene expression an d
splicing efficiency in maize cells. Mol. Gen. Genet. 251, 252–258.
45. Magaraggia, F., Solinas, G., Valle, G., Giovinazzo, G. & Cor-
aggio, I. (199 7) Maturation and translation mechanisms involved
in the expression of a myb gene of rice. Plant Mol. Biol. 35, 1003–
1008.
46. de Vetten, N.C. & Ferl, R.J. (1995) Characterization of a maize
G-box binding factor that is induced by hypoxia. Plant J. 7, 589–
601.
47. Siberil,Y.,Doireau,P.&Gantet,P.(2001)PlantbZIPG-box
binding factors. Modular struc ture and activation m echanisms.
Eur. J. Biochem. 268, 5 655–5666.
Ó FEBS 2004 Properties of a small bZIP protein from tomato (Eur. J. Biochem. 271) 4543
48. Kircher,S.,Wellmer,F.,Nick,P.,Ru
¨
gner, A., Scha
¨
fer, E. &
Harter, K. ( 1999 ) Nu clear import of t he pa rsley bZIP transcrip -
tion fac tor CPRF2 is regulated b y phytochrome ph otorecepto rs.
J. Cell Biol. 144, 2 01–211.
49. Igarashi, D., Ishida, S., Fukazawa, J. & Takahashi, Y. (2001) 14-3-
3 proteins r egulat e intracellular localization of the bZIP tran-
scriptional a ctivator RSG. Pl ant Cell 13 , 2483–2497.
50. Pognonec, P., Boulukos, K.E., Aperlo, C., Fujimoto, M., Ariga,
H., Nomoto, A. & Kato, H . (1997) Cross–family inte raction
between the bHLHZip USF and bZip Fra1 proteins results in

down-regulation of A P1 activity. Oncogene 14, 2 091–2098.
51. Rhee, S.Y., Beavis, W., Berardini, T.Z., Chen, G., Dixon, D.,
Doyle,A.,Garcia-Hernandez,M.,Huala,E.,Lander,G.,Mon-
toya, M., Miller, N., M ueller, L.A., Mundodi, S., Reiser, L.,
Tacklind, J., Weems, D .C., Wu, Y., Xu, I., Yoo, D ., Yoon, J . &
Zhang, P. (2003) The Arabidopsis Infor mation Resource (TAIR):
a model organism database providing a centralized, curated
gateway to Arabidopsis biology, research m aterials and commu-
nity. Nucleic A cids Res. 31 , 224–228.
52. Satoh, R., Fujita, Y., Nakashima, K., Shinozaki, K. & Yama-
guchi-Shinozaki, K. (2 004) A novel subgroup of bZIP proteins
functions as transcriptional activators in hypoosmolarity-
responsive expression of the P roDH gene in Arabidopsis. Plant
Cell Physiol. 45, 309–317.
53. Kusano, T., Sugawara, K ., Ha rada, M. & Berber ich, T. (1998)
Molecular cloning and partial characterization of a t obacco
cDNA encoding a small bZIP protein. Biochim. Biophys. Acta
1395, 171– 175.
54. Ru
¨
gner, A., Frohnmeyer, H., Na
¨
ke, C., Well mer, F., Kircher, S.,
Scha
¨
fer, E. & Harter, K . (2001) Isolation and characterization of
four novel parsley proteins that in teract with the transcriptional
regulators CPRF1 and CPRF2. M ol. Genet. Ge nomics 265,
964–976.
55. Heinekamp, T., K uh lmann, M., Lenk, A., Strathmann, A. &

Dro
¨
ge-Laser, W. (2002) The tobacco bZIP transcription factor
BZI-1 binds to G-box elements in th e p romoters o f phe nyl-
propanoid pathway genes in vitro, but it is not involved in their
regulation in vivo. Mol. Genet. Ge nomics 267, 16–26.
4544 S. Sell and R. Hehl ( Eur. J. Biochem. 271) Ó FEBS 2004