Tài liệu Báo cáo khoa học: Arabidopsis thaliana BTB⁄ POZ-MATH proteins interact with members of the ERF⁄AP2 transcription factor family ppt

Arabidopsis thaliana BTB
⁄
POZ-MATH proteins interact
with members of the ERF
⁄
AP2 transcription factor family
Henriette Weber
1,2
and Hanjo Hellmann
1
1 Washington State University, Pullman, WA, USA
2 Freie University Berlin, Germany
Keywords
APETALA; BPM; cullin; proteasome;
ubiquitin
Correspondence
H. Hellmann, Washington State University,
Pullman, WA 99164, USA
Fax: +1 509 335 3184
Tel: +1 509 335 2762
E-mail:
(Received 17 July 2009, revised 7
September 2009, accepted 11 September
2009)
doi:10.1111/j.1742-4658.2009.07373.x
In Arabidopsis thaliana, the BTB ⁄ POZ-MATH (BPM) proteins comprise a
small family of six members. They have been described previously to use
their broad complex, tram track, bric-a-brac ⁄ POX virus and zinc finger
(BTB ⁄ POZ) domain to assemble with CUL3a and CUL3b and potentially
to serve as substrate adaptors to cullin-based E3-ligases in plants. In this
article, we show that BPMs can also assemble with members of the ethyl-

ene response factor ⁄ Apetala2 transcription factor family, and that this is
mediated by their meprin and TRAF (tumor necrosis factor receptor-asso-
ciated factor) homology (MATH) domain. In addition, we provide a
detailed description of BPM gene expression patterns in different tissues
and on abiotic stress treatments, as well as their subcellular localization.
This work connects, for the first time, BPM proteins with ethylene response
factor ⁄ Apetala2 family members, which is likely to represent a novel
regulatory mechanism of transcriptional control.
Structured digital abstract
l
MINT-7262792: BPM1 (uniprotkb:Q8L765) physically interacts (MI:0915) with RAP2-4 (uni-
protkb:
Q8H1E4)bytwo hybrid (MI:0018)
l
MINT-7262805: BPM1 (uniprotkb:Q8L765) physically interacts (MI:0915) with RAP2-13
(uniprotkb:
Q9LM15)bytwo hybrid (MI:0018)
l
MINT-7262749: BPM3 (uniprotkb:Q2V416) physically interacts (MI:0915) with RAP2-4 (uni-
protkb:
Q8H1E4)bytwo hybrid (MI:0018)
l
MINT-7262764: BPM3 (uniprotkb:Q2V416) physically interacts (MI:0915) with RAP2-13
(uniprotkb:
Q9LM15)bytwo hybrid (MI:0018)
l
MINT-7262838, MINT-7262882, MINT-7262898, MINT-7263072: RAP2-4 (uni-
protkb:
Q8H1E4) binds (MI:0407)toBPM1 (uniprotkb:Q8L765)bypull down (MI:0096)
l

MINT-7262911: RAP2-4 (uniprotkb:Q8H1E4) binds (MI:0407)toBPM2 (uni-
protkb:
Q9M8J9)bypull down (MI:0096)
l
MINT-7262935: RAP2-4 (uniprotkb:Q8H1E4) binds (MI:0407)toBPM3 (uniprotkb:Q2V416)
by pull down (
MI:0096)
l
MINT-7262945: RAP2-4 (uniprotkb:Q8H1E4) binds (MI:0407)toBPM4 (uni-
protkb:
Q9SRV1)bypull down (MI:0096)
l
MINT-7262970: RAP2-4 (uniprotkb:Q8H1E4) binds (MI:0407)toBPM5 (uni-
protkb:
Q1EBV6)bypull down (MI:0096)
l
MINT-7262992: RAP2-4 (uniprotkb:Q8H1E4) binds (MI:0407)toBPM6 (uni-
protkb:
A1L4W5)bypull down (MI:0096)
l
MINT-7263095: RAP2-4 (uniprotkb:Q8H1E4) binds (MI:0407)toRAP2-4 (uniprotkb:
Q8H1E4)bypull down (MI:0096)
Abbreviations
BPM, BTB ⁄ POZ-MATH; BTB ⁄ POZ, broad complex, tram track, bric-a-brac ⁄ POX virus and zinc finger; ERF ⁄ AP2, ethylene response
factor ⁄ Apetala2; GFP, green fluorescent protein; GUS, b-glucuronidase; MATH, meprin and TRAF homology; proBPM, promoterBPM;
RAP2.4, related to Apetala2.4; TRAF, tumor necrosis factor receptor-associated factor; Y2H, yeast two-hybrid.
6624 FEBS Journal 276 (2009) 6624–6635 ª 2009 The Authors Journal compilation ª 2009 FEBS
Introduction
In recent years, a novel superfamily of proteins has
been described in plants that contains a conserved pro-

tein–protein interaction motif named broad complex,
tram track, bric-a-brac ⁄ POX virus and zinc finger
(BTB ⁄ POZ) [1–5]. This protein family is highly diverse
with, for example, 80 members in Arabidopsis and 149
in rice [4,5]. The BTB⁄ POZ domain has a length of
around 116 amino acids and mediates homophilic and
heterophilic interactions between the same or different
proteins, respectively [6,7]. The BTB ⁄ POZ fold consists
of six a-helices and three b-sheets that form a tightly
interwound butterfly-shaped dimer with an extensive
hydrophobic interface [8,9]. BTB ⁄ POZ proteins are
often transcriptional regulators containing a C
2
H
2
domain for DNA binding, but can also be found in
combination with various other protein–protein inter-
action motifs, such as KELCH or meprin and TRAF
(tumor necrosis factor receptor-associated factor)
homology (MATH) motifs, indicating involvement in
various biological processes [5,10–13].
It has recently been demonstrated for animals and
plants that members of the BTB ⁄ POZ family use their
BTB ⁄ POZ domain to assemble with CUL3 proteins
[2–4,14–16]. Cullins are scaffolding subunits of multi-
meric E3-ligases that can polyubiquitinate their sub-
strates and thereby target them for degradation via the
26S proteasome [17]. Thus, plant BTB ⁄ POZ proteins
potentially serve as substrate adaptors for CUL3-based
E3-ligases. Furthermore, self-assembly of BTB ⁄ POZ

proteins has been established for Arabidopsis and rice
[2–4]. However, although plants encode for a large
number of BTB proteins, a functional role has only
been assigned for a few of them, including ETO1 (eth-
ylene biosynthesis [14]), NPH3 (blue light signal trans-
duction [18]), BOP1 (leaf development [19]), ARIA
(abscisic acid signaling [20]), NPY1 (auxin signaling
[21]) and NPR1 (salicylic acid signaling [22]).
Some BTB proteins from plants and animals contain
a secondary MATH domain which comprises around
150 amino acids forming eight b-sheets [23]. The motif
was noted on the basis of homology with the C-terminal
region of meprins A and B and the TRAF-C domain,
and, like the BTB domain, facilitates protein–protein
interaction [24]. Meprins are tissue-specific metalloen-
dopeptidases implicated in developmental and patho-
logical processes in animals by hydrolyzing a variety of
peptides and proteins [25–27]. In mammals, TRAFs
regulate cell growth signaling and apoptosis by interact-
ing with membrane-bound receptors through their
TRAF-C domains [28,29]. Although TRAFs and mep-
rins have not been described in plants, a variety of plant
proteins functionally unrelated to meprins and TRAFs
contain MATH domains [30], and proteins carrying
both BTB and MATH motifs are common in plants.
Arabidopsis, for example, expresses six members of this
BTB subfamily [referred to as the BPM (BTB ⁄ POZ-
MATH) family] [2,16], whereas, in rice, 74 members are
annotated [5]. Although it has been established that the
BTB domain is employed to facilitate assembly with

CUL3 and other BTB proteins [2,16], it remains unclear
what kind of interactions are mediated by the MATH
domain in plants.
In this article, we show that the MATH domain of
BPM proteins is used to assemble with members of the
ethylene response factor ⁄ Apetala2 (ERF ⁄ AP2) tran-
scription factor family. We also provide a detailed
description of the Arabidopsis thaliana BPM family
expression and subcellular localization profile, including
promoter:b-glucuronidase (promoter:GUS) and green
fluorescent protein (GFP) fusion protein studies for all
six members. Overall, the work demonstrates a novel role
for BPMs as potential regulators that affect transcrip-
tional activities of ERF ⁄ AP2 proteins in higher plants.
Results
BPM proteins assemble with members of the
ERF
⁄
AP2 transcription factor family
Because it has been shown previously that Arabidopsis
BPM proteins use their BTB ⁄ POZ domain to interact
with the cullins CUL3A and CUL3B [3,4,16], we inves-
tigated what kind of protein–protein interactions were
facilitated by their MATH domains by performing two
l
MINT-7262855: RAP2-13 (uniprotkb:Q9LM15) binds (MI:0407)toBPM1 (uniprotkb:
Q8L765)bypull down (MI:0096)
l
MINT-7263015: BPM1 (uniprotkb:Q8L765) binds (MI:0407)toAt4g13620 (uniprotkb:
Q9SVQ0)bypull down (MI:0096)

l
MINT-7263047: BPM1 (uniprotkb:Q8L765) binds (MI:0407)toAt4g39780 (uniprotkb:
O65665)bypull down (MI:0096)
H. Weber and H. Hellmann Arabidopsis thaliana BPM–ERF ⁄ AP2 interaction
FEBS Journal 276 (2009) 6624–6635 ª 2009 The Authors Journal compilation ª 2009 FEBS 6625
yeast two-hybrid (Y2H) screens on a root-specific
cDNA library. One screen was performed with a full-
length BPM3 (At2g39760), whereas, for the other, we
used a BPM1 (At5g19000) fragment that lacked the
BTB ⁄ POZ domain [denoted as BPM1(1–189); Fig. 1B].
As both the MATH and BTB domains mediate assem-
bly with other proteins, we speculated that this dual
approach would not only identify specific interactors
for the MATH domain, but would also provide infor-
mation on to what extent the two different MATH
domains of BPM1 and BPM3 target the same group
of proteins (see Table S1A,B for identity ⁄ similarity
comparisons of BPM proteins and their MATH
domains, respectively).
In total, 250 yeast clones were analyzed as primary
positives and, consistent with earlier studies [2], BPM4
was found using BPM3 as bait (data not shown). How-
ever, predominantly, we identified RAP2.4 (At1g78080;
related to Apetala2.4), which was found 15 times with
BPM3 and 18 times with the BPM1(1–189) fragment
(Fig. 1A). RAP2.4 belongs to the ERF ⁄ AP2 family of
transcription factors and contains a single AP2 domain.
The protein has been described previously in context
with abiotic stress tolerance, red light response and eth-
ylene signaling [31]. We also identified once At1g22190

in the BPM3 screen, which is the closest relative of
RAP2.4 [32]. It should be noted that both RAP2.4 and
At1g22190 were isolated as partial clones lacking the
first 60 amino acids (Fig. 1B), demonstrating that this
region is not essential for assembly with BPM proteins.
As 12 RAP2 proteins have been annotated previously
[33], we retained this nomenclature and denoted
At1g22190 as RAP2.13.
To further confirm the interaction of RAP2.4 and
RAP2.13 with BPM proteins, we cloned the corre-
sponding full-length cDNAs for both genes, generated
GST fusion proteins and tested these in pulldown
assays with BPM1(1–189). As shown in Fig. 1C,
BPM1(1–189) coprecipitated with both GST:RAP2
proteins, but not with GST alone, further corroborat-
ing the Y2H data. Because RAP2.4 and RAP2.13 are
closely related to each other, we mainly focused on
RAP2.4 as a representative example in subsequent
experiments. Here, RAP2.4 also interacted with a full-
length BPM1 (using a GST:BPM1 protein) and with
itself (using GST:RAP2.4) (Fig. 1D). Additional pull-
down assays positively confirmed binding to BPM2,
BPM4, BPM5 and BPM6 (Fig. 1E). To exclude non-
specific assembly with BPM proteins, we decided to
test At1g65050. This protein has no BTB motif, but
contains a MATH domain that is most closely related
to those from BPMs [30]. In these experiments,
RAP2.13
fragment
RAP2.4

fragment
33460
AP2
214150
261
60
AP2
81 142
MATH
BPM1
1–189
1 38 150 189
BPM1
1–189
RAP2.4
RAP2.13
Empty vector
BPM3
Input
GST
GST:RAP2.4
GST:RAP2.13
BPM1
1–189
BA
C
E
Input
GST
GST:RAP2.4

BPM1
BPM2
BPM3
BPM4
BPM5
BPM6
At1g65050
D
Input
GST
GST
GST:RAP2.4
GST:BPM1
RAP2.4
Fig. 1. BPM proteins interact with members of the ERF ⁄ AP2 family. (A) Y2H assays demonstrate the assembly of BPM1(1–189) and BPM3
with RAP2.4 and RAP2.13, respectively. (B) Schematic drawing of the BPM1 fragment used for the Y2H screen and the shortest RAP2.4
and RAP2.13 fragments retrieved from the screens. (C) In vitro-translated BPM1(1–189) coprecipitates with full-length GST:RAP2.4 and
GST:RAP2.13. (D) In vitro-translated RAP2.4 coprecipitates with GST:BPM1 and GST:RAP2.4. (E) In vitro-translated BPM1–6 co-precipitate
with GST:RAP2.4, but the MATH protein At1g65050 does not. In this and all subsequent figures, the input represents 1–3 lL of the coupled
in vitro transcription ⁄ translation reaction mixture, and GST was used as a negative control. All experiments in this and subsequent figures
were repeated at least three times.
Arabidopsis thaliana BPM–ERF ⁄ AP2 interaction H. Weber and H. Hellmann
6626 FEBS Journal 276 (2009) 6624–6635 ª 2009 The Authors Journal compilation ª 2009 FEBS
RAP2.4 did not interact with At1g65050 (Fig. 1E),
which is a critical finding as it suggests that RAP2.4–
BPM assembly is specific.
BPM1–RAP2.4 interaction requires a complete
MATH domain and the N-terminal region of
RAP2.4
The use of a truncated version of BPM1 in the Y2H

screens demonstrated that the BTB domain is not
involved in the assembly with RAP2.4. However,
BPM1(1–189) still contains nearly 80 amino acids that
are not part of the MATH domain and which could rep-
resent possible interaction sites for RAP2.4. To further
confirm that a full-length MATH domain is sufficient for
binding to RAP2.4, we generated a new truncated BPM1
version of 151 amino acids [BPM1(1–151)] comprising
the first 38 amino acids of BPM1 followed by the com-
plete MATH domain. As shown in Fig. 2A, BPM1(1–
151) is entirely capable of binding to GST:RAP2.4, mak-
ing it highly probable that only the MATH domain is
required for RAP2.4–BPM1 assembly.
Likewise, we were interested in the RAP2.4 region
that mediates the assembly with BPM1 proteins. Its
AP2 domain stretches from amino acid residue 150 to
214. To test whether a functional AP2 domain is criti-
cal for assembly with BPMs, we took advantage of an
earlier description of ap2-1 and ap2-5 mutants, in
which mutation of a glycine residue in the AP2
domain disrupts the protein’s DNA-binding affinity
[33,34]. This glycine is highly conserved and can also
be found in RAP2.4 at position 179 [35]. However, the
introduction of a point mutation that changed the gly-
cine residue to serine [RAP2.4(G179S)] did not affect
assembly with GST:BPM1, indicating that a functional
AP2 domain is not required for this type of interac-
A
B
347150381203

442
BTBHTAMBPM1
189
1–151
F
RAP2.4
1
150
214
AP2
334
60
125–251
C
D
BPM1
BPM1
1–151
BPM1
1–189
GST
GST:RAP2.4
––– – ––++ +
–––– –+–
–
+
BPM1
T7 input
BPM1
1–151

BPM1
1–189
–+––
+–
–––+
–+
GST
GST:BPM1
T7 input
RAP2.4
1–251
RAP2.4
1–251
RAP2.4
1–295
RAP2.4
1–295
E
RAP2.4
G179S
RAP2.4
134–END
RAP2.4
116–END
RAP2.4
125–END
Input
GST
GST:BPM1
RAP2.4

134–END
RAP2.4
125–END
RAP2.4
Input
GST
GST:RAP2.4
Fig. 2. Mapping of the interactive sites in
BPM1 and RAP2.4. (A) In vitro-translated
BPM1(1–151) can interact with GST:RAP2.4.
(B) Schematic drawing of BPM1. The trian-
gle indicates the fragment used for the Y2H
screen and for pulldowns in (A). (C) In vitro-
translated RAP2.4(1–251) is able to interact
with GST:BPM1. (D) GST:BPM1 can assem-
ble with in vitro-translated RAP2.4(G179S),
RAP2.4(116–END) and RAP2.4(125–END),
but not with RAP2.4(135–END). (E)
GST:RAP2.4 can interact with in vitro-
translated RAP2.4(125–END), but not with
RAP2.4(135–END). (F) Schematic drawing of
RAP2.4. Triangle indicates the fragment
found in the Y2H screen.
H. Weber and H. Hellmann Arabidopsis thaliana BPM–ERF ⁄ AP2 interaction
FEBS Journal 276 (2009) 6624–6635 ª 2009 The Authors Journal compilation ª 2009 FEBS 6627
tion. Next, we generated several truncated versions of
RAP2.4 that were translated in vitro and tested for
interaction with GST:BPM1. As we originally found
truncated versions of RAP2.4 and RAP2.13 that were
missing the first 60 amino acids in the Y2H screens,

we started out with further reduced versions that
lacked the first 116, 125 and 134 amino acids.
Although complete deletion of the first 116 and 125
amino acids [RAP2.4(125–END)] did not affect copre-
cipitation with GST:BPM1, we could not detect inter-
action with a truncated version that lacked the first
134 amino acids [RAP2.4(134–END)] (Fig. 2D). In
addition, deletion of amino acid residues C-terminal
from the AP2 domain [RAP2.4(1–251)] did not affect
the interaction with GST:BPM1 (Fig. 2C). We there-
fore conclude that a critical region for assembly with
BPM proteins is located within amino acid residues
125–251 of RAP2.4.
Detailed expression analysis of BPM and RAP2.4
shows distinct patterns for the different genes
Although the interaction studies presented demonstrate
that all BPM proteins can assemble with RAP2.4, and
even provide strong evidence for the assembly of the
transcription factor in planta, it is still unclear whether
BPM and RAP2.4 genes are expressed in the same tis-
sues. Consequently, we analyzed the tissue-specific
expression patterns of all BPM genes and RAP2.4 via
semiquantitative RT-PCR, and further described their
expression in greater detail using promoter:GUS lines
(referred to as proBPM:GUS and proPRAP2.4:GUS,
respectively).
The results from RT-PCR showed that BPM2 and
BPM5 were strongly expressed in all tested tissues
(roots, rosette and cauline leaves, stems and flowers)
(Fig. 3A). Although BPM6 was also strongly

expressed, its expression level was weaker overall in
comparison with BPM2 and BPM5. For BPM1 and
BPM3, we could hardly detect expression in the differ-
ent tissues, and had to load double the amount of RT-
PCR products on the gels to visualize any PCR prod-
ucts (Fig. 3A). BPM1 showed only slightly higher
expression levels in root and flower, and BPM3 expres-
sion levels showed little variation between the different
tissues (Fig. 3A). BPM4 also showed little variation,
and expression was lower than that of the other BPM
genes. In this case, we had to load triple the amount
of RT-PCR product relative to that used for BPM2
and BPM5. Finally, RAP2.4 was expressed strongly in
roots, rosette and cauline leaves, and flowers, with
slightly weaker expression levels in the stem (Fig. 3A).
Rosette leaf
Cauline leaf
Stem
Flower
Root
BPM1
Control
Control
NaCl
Sorbitol
BPM2
BPM3
BPM4
BPM5
BPM6

RAP2.4
actin2
Drought
CBA
*
*
(2x)
(2x)
(3x)
Fig. 3. Expression profiles of BPMs and RAP2.4 genes in Arabidopsis thaliana analyzed by semiquantitative RT-PCR. (A) Total RNA (100 ng),
extracted from roots, rosette and caulin leaves, sections of stems and open flowers of mature plants grown in soil, was used for RT-PCR.
The expression of all tested genes was detected in all tested tissues, but with considerable differences in expression strength ⁄ intensity: For
BPM1 and BPM3 twofold, and for BPM4 threefold, the amount of RT-PCR product was loaded (compared with actin2 control reaction).
(B) RT-PCR analysis showing BPM1, BPM2 , BPM5 and RAP2.4 up-regulated by salt (200 m
M NaCl for 6 h) and osmotic stress (200 mM sor-
bitol for 6 h). Sorbitol treatment also induced BPM3 and BPM4. (C) On drought treatment (drying for 6 h on a laboratory bench), only BPM1
and BPM4 showed up-regulation in expression. Numbers in parentheses indicate the fold amount of RT-PCR loaded in comparison with
actin2. Asterisks indicate correct RT-PCR products.
Arabidopsis thaliana BPM–ERF ⁄ AP2 interaction H. Weber and H. Hellmann
6628 FEBS Journal 276 (2009) 6624–6635 ª 2009 The Authors Journal compilation ª 2009 FEBS
Because RAP2.4 has been described previously to
play a role in abiotic stress tolerance, we tested
whether expression of the different BPM genes was
regulated by salt (NaCl), osmotic (sorbitol) and
drought stress. Treatment of Col0 wild-type plants
with 200 mm NaCl for 6 h resulted in a clear up-regu-
lation of BPM1 and BPM5 expression (Fig. 3B). We
also observed an up-regulation of BPM1 and BPM5
after treatment with sorbitol for 6 h, together with
increased BPM4 levels (Fig. 3B). RAP2.4 also

responded to both treatments with enhanced expres-
sion, which is in agreement with earlier findings from
Lin et al. [31] (Fig. 3B). Drought stress only induced
the expression of BPM1 and BPM4; all other BPM
genes and RAP2.4 remained unchanged (Fig. 3B).
Overall, these data indicate that BPM1, BPM4 and
BPM5 are involved in the abiotic stress response.
The analysis of transgenic plants carrying the differ-
ent promoter:GUS constructs showed, for proB-
PM1:GUS lines, GUS expression in pollen, but also in
stipules and leaf hydathodes (Fig. 4A). Rosette leaves
showed staining within the vascular tissue at the end
of the leaf blade, whereas the basal parts close to the
petiole remained almost unstained. We observed clear
GUS expression in the primary root of 7-day-old seed-
lings, strongest at the base of emerging lateral roots,
but no expression at detectable levels in the tips of pri-
mary and budding lateral roots. proBPM2:GUS lines
(Fig. 4B) showed strong expression in the vascular tis-
sue of cotyledons and rosette leaves, and in most parts
of the flower. Similar to proBPM1:GUS, strong stain-
ing was detectable in the stipules, pollen and at the
base of siliques. Expression in roots was detectable
along the primary but not lateral roots, with strongest
staining present at the budding lateral root primordia.
proBPM3:GUS lines showed clear GUS expression in
the root tips, but also in the stipules, anthers and in
the central veins and petioles of rosette leaves
(Fig. 4C). proBPM4:GUS showed GUS staining very
similar to that of proBPM3:GUS in the stipules, the

central veins of rosette leaves and in the anthers of dif-
ferentiated flowers. We also detected meagre expres-
sion along the root, with most obvious staining
present at the lateral root primordia and the base of
the lateral roots, and also in the columella (Fig. 4D).
Like proBPM2:GUS, proBPM5:GUS plants showed a
wide range of expression patterns in all tested organs
(Fig. 4E). Both cauline and rosette leaves showed
strong GUS expression, as did the primary root tips
and the stem, whereas, in the flower, expression was
detectable in the petals, stamen and stigmata. In proB-
PM6:GUS lines, we saw GUS expression in the vascu-
lar tissue of cotyledons and mature leaves, whereas, in
the flowers, the anthers, connectives and filaments and
the base and tip of the stigmata were stained. Similar
to BPM2, BPM3 and BPM5 promoter:GUS lines, the
root tips were strongly stained, with the exception of
columella cells which remained nearly white (Fig. 4F).
Finally, proPRAP2.4:GUS lines showed blue staining
in cotyledons of 3-day-old seedlings, but not in parts
of the hypocotyls (Fig. 4G). In rosette leaves, we
observed expression in the vascular tissue of the leaf
blade, whereas, interestingly, in older parts of the mid-
rib, no GUS expression was detectable. This was dif-
ferent from cauline leaves, in which all vascular tissue
was stained. In the flower, we detected GUS expres-
sion exclusively in the pollen. Siliques were stained at
the base and at the tip, with overall very weak staining
of the fused carpels. In the root, the central cylinder
was stained, whereas the primary root tips and tips of

emerging lateral roots showed no blue staining
(Fig. 4G, part f, marked by arrows).
Subcellular localization analysis of BPM proteins
and their interactors
Using a GFP:RAP2.4 fusion protein, it has recently
been established that RAP2.4 is primarily located in
the nucleus [31]. Accordingly, one would expect that
this organelle would be the most likely location for the
assembly of BPM proteins with RAP2.4. We generated
expression constructs for all BPM genes; however,
only for BPM4 were we able to obtain GFP:BPM4
overexpressing plants. As an alternative approach to
investigate the subcellular localization of the different
BPM proteins, we transiently expressed them in
tobacco leaves. We also included GFP:RAP2.4 and
GFP:CUL3a in these experiments to compare their
localization with that of BPM proteins.
Transient expression of GFP:BPM1 and GFP:BPM2
revealed that both proteins, like GFP:RAP2.4, were
primarily localized to the nucleus (Fig. 5 and Fig. S2).
The predominantly nuclear localization of BPM1 and
BPM2 GFP fusion proteins contrasted with all other
BPMs, as GFP:BPM3, GFP:BPM5 and GFP:BPM6
were found inside as well as outside the nucleus.
Remarkably, GFP:BPM4 was the only BPM protein
that was excluded from the nucleus, suggesting that
BPM4 and RAP2.4 are not present in the same cellular
compartments. We observed this in transient expres-
sion assays, but also in Arabidopsis plants that stably
expressed GFP:BPM4 (Fig. 5 and Fig. S3). Also note-

worthy was the observation that GFP:CUL3a showed
a subcellular localization pattern similar to
GFP:BPM3, GFP:BPM5 and GFP:BPM6. Overall,
these analyses revealed a very distinct and different
H. Weber and H. Hellmann Arabidopsis thaliana BPM–ERF ⁄ AP2 interaction
FEBS Journal 276 (2009) 6624–6635 ª 2009 The Authors Journal compilation ª 2009 FEBS 6629
proBPM6:GUS
proBPM1:GUS proBPM2:GUS
proBPM4:GUSproBPM3:GUS
proBPM5:GUS proRAP2.4:GUS
BA
DC
GFE
a
b
e
c
f
d
g
a
b
e
c
f
g
d
h
i
a

b
d
c
a
b
e
c
e
f
g
f
g
h
d
a
e
c
f
g
a
d
b
h
b
e
a
c
d
d
e

b
f
c
Fig. 4. Expression profile of proBPM:GUS and proRAP2.4:GUS in Arabidopsis thaliana. (A) proBPM1:GUS : hydathodes and stipules of 5-day-
old seedlings showed staining (a–c), as did fully developed siliques (d; base and stigma region) and vascular tissue of rosette leaves (f). In
flowers, expression was restricted to pollen and anthers (e). The primary root (g) was stained throughout, but the strongest expression was
detectable at the points of emerging lateral roots, indicated by the arrows. (B) proBPM2:GUS showed the strongest expression of all pro-
moter:GUS lines, detected in all tested tissues. Although cotyledons, hypocotyls and rosette leaves were strongly stained (a–c), GUS expres-
sion in cauline leaves was restricted to the base and apex of the lamina (c). Siliques showed staining at their bases and tips (stigma region)
(d). In flowers, the petals, stamens, receptacle and upper pistil were stained (e). Close-up of stigma with strong staining of the stigma’s
papillae (f). proBPM2:GUS plants showed GUS expression in the primary root (g, i) and lateral root primordia (h), but not in developed lateral
roots (g). (C) proBPM3:GUS lines showed altogether very weak expression. In seedlings, staining was only detectable in the stipules (a).
Rosette and cauline leaves showed good GUS expression in the central vein and petioles (b, c), as well as the anthers in flowers (d). proB-
PM3:GUS plants also showed expression in the root tips (e–g). (D) GUS expression under the BPM4 promoter was also weak, but with clear
expression in the stipules (a), midrib of rosette leaves (b), mature anthers and stigmata (c, d). In roots, faint expression was detected along
the primary root and its tip (g, h), whereas the lateral root primordia and base of the developed lateral roots showed stronger staining (e–g).
(E) GUS staining for proBPM5:GUS lines was strong in the vascular tissue of the cotyledons (a) and in mature leaves (b, rosette; c, caulin),
but also in the hypocotyl (a), young siliques (d) and flowers (g). (F) For BPM6, strong expression was observed in 3-day-old seedlings (a), as
well as in the entire lamina and petiole of rosette leaves, including vascular veins (b). In flowers, GUS expression started in the early stages
of the receptacles and stigmata (c) In older flowers, mature anthers and, later, filaments were also stained. In roots, GUS was expressed
only in the tips of primary roots (d), and at the base of differentiated siliques (e). (G) proRAP2.4:GUS lines showed GUS expression in 3-day-
old seedlings in cotyledons and in the central cylinder of the root, but not in the lower parts of the hypocotyl (a). Rosette leaves were
stained in the vascular tissue of the leaf blade, whereas the petiole and older parts of the midrib remained unstained (b). However, the cau-
line leaf blade was stained very evenly (c). In flowers, staining was only detectable in the pollen (d). Expression levels in siliques were very
low, with staining mainly present at the base and at the tip (e). A close-up of a root section showed that the central cylinder was stained (f),
whereas the lateral root primordia (marked by arrows) did not show any GUS expression.
Arabidopsis thaliana BPM–ERF ⁄ AP2 interaction H. Weber and H. Hellmann
6630 FEBS Journal 276 (2009) 6624–6635 ª 2009 The Authors Journal compilation ª 2009 FEBS
localization for the different BPM proteins, which
might reflect their diverse biological roles in the cell. In

addition, they indicate that, except for BPM4, all other
BPM proteins are potentially able to interact with
RAP2.4 in planta, and also that CUL3a can assemble
with other proteins either in the nucleus or the cyto-
plasm.
Discussion
This work provides novel information, including the
first description linking ERF ⁄ AP2 proteins to the
BPM family, but also a detailed report of BPM
expression and subcellular localization.
Our intensive interaction studies have identified the
domains required for RAP2.4 ⁄ BPM assembly to the
BPM MATH domain and a region of RAP2.4 that
stretches from amino acid residues 125 to 251, which
encompasses the AP2 domain. Although the point
mutation G179S in the AP2 domain of RAP2.4 does
not affect assembly with BPMs, this does not rule
out a possible role of the domain for this type of
interaction. However, to date, there is no evidence
that the AP2 domain mediates protein–protein inter-
action and, because of this, we consider it probable
that it is also not involved in assembly with the BPM
proteins.
Both RAP2.4 and RAP2.13 belong to a small sub-
group within the ERF ⁄ AP2 superfamily that comprises
eight members and is called the A-6 subfamily [32,36].
We tested more members of this subgroup, At1g36060,
At4g39780 and At4g13620, in pulldown assays with
GST:BPM1. Although At4g39780 and At4g13620 were
both able to assemble with GST:BPM1, At1g36060

was not (Fig. S1), indicating that assembly with
ERF ⁄ AP2 proteins is restricted to the A-6 subfamily
and, in this case, even to a subset of eight members.
The finding that, within the A-6 subfamily, not all of
its members interact with BPM1 indicates that BPM
proteins assemble only with a very limited set of
ERF ⁄ AP2 proteins, which potentially does not extend
beyond the A-6 subgroup. However, we experienced a
high degree of redundancy from the BPM site, as all
BPMs were able to interact with RAP2.4. It will be of
interest to determine whether this redundancy is also
present for all other ERF ⁄ AP2 proteins that bind to
BPM1.
Studies on gene expression showed that BPMs have
a widely overlapping pattern of expression. This was
further corroborated by the promoter:GUS lines,
GFP:BPM1
GFP:BPM2
GFP:BPM3
GFP:BPM4
GFP:BPM5
GFP:BPM6
GFP:RAP2.4
GFP:CUL3a
100 µm
100 µm
100 µm
100 µm
50 µm
50 µm

50 µm
50 µm
50 µm 50 µm
50 µm
50 µm
100 µm100 µm
200 µm
200 µm
Fig. 5. Localization of transiently expressed
GFP fusion proteins of BPM, RAP2.4 and
CUL3A in Nicotiana benthamiana leaves
was analyzed by confocal laser scanning
microscopy. GFP signals (green) and
merged images of GFP against the red chlo-
rophyll autofluorescence background signal
are shown. Like GFP:RAP2.4, GFP:BPM1
and GFP:BPM2 showed a predominant
localization to the nucleus, whereas
GFP:BPM4 showed an opposite pattern and
was detectable only outside the nucleus.
GFP:BPM3, GFP:BPM5, GFP:BPM6 and
GFP:CUL3A accumulated inside and outside
the nucleus (nuclei of single cells are indi-
cated by arrows).
H. Weber and H. Hellmann Arabidopsis thaliana BPM–ERF ⁄ AP2 interaction
FEBS Journal 276 (2009) 6624–6635 ª 2009 The Authors Journal compilation ª 2009 FEBS 6631
which showed that most BPMs were expressed in
anthers, root tips and rosette leaves. Although some
patterns were highly specific, such as, for example,
expression of BPM1 and BPM4 at the junction of

primary to lateral roots, or BPM3 expression specifi-
cally in root tips, overall our results indicated that
BPM proteins were functionally redundant. Conse-
quently, one would expect no obvious developmental
defects in plants affected in single BPM genes, which
is the case for available T-DNA insertion mutants (H.
Weber and H. Hellmann, unpublished work). How-
ever, on the basis of the expression patterns
described, it is predictable that mutants affected in
multiple BPMs will show aberrant flower, leaf and
root development. Likewise, the inducible expression
of BPM1, BPM4 and BPM5 on abiotic stress treat-
ment suggests that corresponding single or multiple
mutants will display an altered tolerance when
exposed to these stressors.
The widely overlapping expression patterns of BPMs
with RAP2.4 also suggest that the transcription factor
can potentially assemble with most BPM proteins. This
was further supported by our subcellular localization
studies, in which all of the BPMs, except BPM4, were
present in the nucleus. However, the specific nuclear
localization of BPM1 and BPM2 currently makes both
proteins the most favorable candidates for in planta
assembly with the transcription factor RAP2.4. As
BPM3, BPM5 and BPM6 were also present in the
cytosol, it is probable that they assemble with addi-
tional, yet unknown proteins in this cellular compart-
ment, and this is especially likely for BPM4, which
was never found in the nucleus. In this case, it is of
interest that CUL3a was also localized to the nucleus

and the cytosol. Because a proposed role of BPMs is
to function as substrate adaptors to a CUL3-based
E3-ligase, the current findings suggest that such an
assembly can occur within the nucleus and the cyto-
plasm and that, in both compartments, proteins can be
ubiquitinated and potentially marked for degradation
via the 26S proteasome.
Conclusion
In future work, it will be critical to verify our findings
on BPM–ERF ⁄ AP2 assembly in planta and to identify
the motif in RAP2.4 that mediates this type of interac-
tion, as this has the potential to predict possible pro-
tein binding to the BPMs. It will also be of importance
to define the functional impact of the BPM family on
the activity of ERF ⁄ AP2 transcription factors. For
example, do they affect the stability of these proteins
and does this require CUL3s? In this case, it is note-
worthy that we observed the instability of in planta-
expressed myc-tagged RAP2.4 in a 26S proteasome-
dependent manner (Fig. S4). We currently favor a
working model in which they bind to ERF ⁄ AP2s,
potentially interfering with their DNA-binding ability,
but which ultimately results in the degradation of
ERF ⁄ AP2 proteins ( Fig. 6). Independent of this hypo-
thetical role, the work described here opens up a novel
and important connection between two plant protein
families, and provides a forecast on a new regulatory
mechanism controlling ERF ⁄ AP2 transcription factor
activities.
Materials and methods

Plant growth conditions and transformation
Arabidopsis thaliana (ecotype Col0) and tobacco (Nicoti-
ana benthamiana) plants were grown under standard condi-
tions with 16 h : 8 h light : dark cycles in either soil or
sterile culture, using ATS medium [37] without supple-
mented sucrose. Arabidopsis floral dip transformation was
performed as described in [38].
Fig. 6. Schematic model for assembly and
functional impact of BPM–ERF ⁄ AP2 assem-
bly. BPM proteins function as substrate
adaptors to CUL3-based E3-ligases. They
also assemble with ERF ⁄ AP2 transcription
factors, and this interaction serves to bring
bound ERF ⁄ AP2 proteins to the core
E3-ligase. Docking of the BPM–ERF ⁄ AP2
complex to the E3-ligase results in ubiquiti-
nation and subsequent degradation of bound
transcription factor.
Arabidopsis thaliana BPM–ERF ⁄ AP2 interaction H. Weber and H. Hellmann
6632 FEBS Journal 276 (2009) 6624–6635 ª 2009 The Authors Journal compilation ª 2009 FEBS
Molecular cloning and mutagenesis
Full-length cDNAs of BPM genes were amplified from a
seedling-specific cDNA library [39]. The promoters of the
different BPM genes and RAP2.4 (for sizes, see Table S1),
as well as AP2 ⁄ ERF transcription factors, were amplified
directly from Col0 genomic DNA. In all cases, Pfu poly-
merase (Promega, Mannheim, Germany) was used and the
PCR products were controlled for correct sequences. The
cDNAs obtained were subcloned into pDONR221 (Invitro-
gen, Carlsbad, CA, USA) and shuffled into different desti-

nation vectors using GATEWAY technology (Invitrogen):
pACT2 and pBTM116 [2] for Y2H studies, pDEST15 (Invi-
trogen) for Escherichia coli expression and the binary vector
pK7FWG2 [40] for subcellular localization analysis. The
amplified promoters were subcloned into pCR2.1 by TOPO
TA reaction (Invitrogen). Afterwards, using the BamHI and
XbaI restriction sites, the promoters were fused to the GUS
gene in the binary vector pCB308 [41]. The primers used in
this and other sections are given in Table S2. Mutagenesis
was performed using a mutagenesis kit from Stratagene
(La Jolla, CA, USA) as described previously [2].
Expression analysis
Expression was studied by RT-PCR with gene-specific
primers, and histochemically using promoter-GUS fusions.
RT-PCR was performed on 100 ng of total RNA isolated
from different tissues of mature Arabidopsis ecotype Col0
plants grown on soil, or on total RNA extracted from
7-day-old seedlings grown on plates, respectively. For histo-
chemical analysis, promoters in the binary vector pCB308
were introduced into Arabidopsis plants. Transgenic plants
were selected by BASTA herbicide (Aventis Crop Science,
Leverkusen, Germany). GUS staining was carried out by
vacuum infiltration of plant material with staining solution
[2] and subsequent incubation at room temperature for up
to 24 h. For stress treatment, 7-day-old sterile-grown seed-
lings were transferred for 6 h into liquid ATS medium sup-
plemented with either 200 mm NaCl or sorbitol. To impose
drought stress on 7-day-old seedlings, the lids from culture
dishes were removed for 6 h before the samples were har-
vested.

Y2H assay
Screening for BPM-interacting clones was performed using
a root-specific suspension cell cDNA library in the prey
vector pACT2-GW [42]. The MATH domain of BPM1 and
full-length BPM3 were cloned into the bait vector
pBTM116-D9-GW [42]. Yeast transformation and testing
for interaction were performed as described in [2]. Clones
were transformed into yeast with an efficiency of 1.5 million
clones per transformation. All BPM-interacting clones were
tested for auto-activation and sequenced for the correct
open reading frame in pACT2.
Subcellular localization analysis
Fluorescent fusion proteins of the six BPM proteins, CUL3a
and RAP2.4 were transiently expressed in tobacco epidermal
cells using the method of Agrobacterium infiltration as
described in [43]. The bacterial attenuance (D) at 600 nm was
0.01–0.03 for all constructs. In addition, BPM4 localization
was also analyzed in stable transgenic Arabidopsis plants
expressing GFP:BPM4 fusion protein. In all cases, binary
GFP expression vectors obtained from [40] were used.
Transfected leaf sections were imaged using a Zeiss (Jena,
Germany) LSM 510 Meta confocal microscope.
In vitro transcription ⁄ translation assays
For interaction studies, full-length BPM proteins, fragments
of BPM1 and selected ERF ⁄ AP2 proteins were expressed in
the TNT-reticulocyte lysate system (Promega) as described
previously [2]. In vitro-translated proteins were labeled with
either [
35
S]methionine (Amersham, Chalfont St Giles, UK)

or
green
Lysine (Promega).
Acknowledgements
We thank Sutton Mooney for critical reading of the
manuscript. Financial support for this project was pro-
vided by the Deutsche Forschungsgemeinschaft (DFG)
grants HE3224 ⁄ 5-1 and 5-2 and Washington State
University to HH.
Accession numbers
BPM1 (At5g19000 ⁄ Q8L765); BPM2 (At3g06190 ⁄ Q9-
M8J9); BPM3 (At2g39760 ⁄ O22286); BPM4 (At3g-
03740 ⁄ Q9SRV1); BPM5 (At5g21010 ⁄ Q1EBV6); BPM6
(At3g43700 ⁄ A1L4W5); RAP2.4 (At1g78080 ⁄ Q8H1E4);
RAP2.13 (At1g22190 ⁄ Q9LM15).
References
1 Stogios PJ, Downs GS, Jauhal JJ, Nandra SK & Prive
GG (2005) Sequence and structural analysis of BTB
domain proteins. Genome Biol 6, R82.
2 Weber H, Bernhardt A, Dieterle M, Hano P, Mutlu A,
Estelle M, Genschik P & Hellmann H (2005) Arabidop-
sis AtCUL3a and AtCUL3b form complexes with
members of the BTB ⁄ POZ-MATH protein family.
Plant Physiol 137, 83–93.
3 Figueroa P, Gusmaroli G, Serino G, Habashi J, Ma L,
Shen Y, Feng S, Bostick M, Callis J, Hellmann H et al.
H. Weber and H. Hellmann Arabidopsis thaliana BPM–ERF ⁄ AP2 interaction
FEBS Journal 276 (2009) 6624–6635 ª 2009 The Authors Journal compilation ª 2009 FEBS 6633
(2005) Arabidopsis has two redundant Cullin3 proteins
that are essential for embryo development and that

interact with RBX1 and BTB proteins to form multi-
subunit E3 ubiquitin ligase complexes in vivo. Plant
Cell 17, 1180–1195.
4 Gingerich DJ, Gagne JM, Salter DW, Hellmann H,
Estelle M, Ma L & Vierstra RD (2005) Cullins 3a and
3b assemble with members of the broad complex ⁄ tram-
track ⁄ bric-a-brac (BTB) protein family to form essential
ubiquitin–protein ligases (E3s) in Arabidopsis. J Biol
Chem 280, 18810–18821.
5 Gingerich DJ, Hanada K, Shiu SH & Vierstra RD
(2007) Large-scale, lineage-specific expansion of a bric-
a-brac ⁄ tramtrack ⁄ broad complex ubiquitin-ligase gene
family in rice. Plant Cell 19 , 2329–2348.
6 Zollman S, Godt D, Prive GG, Couderc JL & Laski
FA (1994) The BTB domain, found primarily in zinc
finger proteins, defines an evolutionarily conserved fam-
ily that includes several developmentally regulated genes
in Drosophila. Proc Natl Acad Sci USA 91, 10717–
10721.
7 Bardwell VJ & Treisman R (1994) The POZ domain: a
conserved protein–protein interaction motif. Genes Dev
8, 1664–1677.
8 Ahmad KF, Engel CK & Prive GG (1998) Crystal
structure of the BTB domain from PLZF. Proc Natl
Acad Sci USA 95, 12123–12128.
9 Ahmad KF, Melnick A, Lax S, Bouchard D, Liu J,
Kiang CL, Mayer S, Takahashi S, Licht JD & Prive
GG (2003) Mechanism of SMRT corepressor recruit-
ment by the BCL6 BTB domain. Mol Cell 12, 1551–
1564.

10 Csankovszki G, Nagy A & Jaenisch R (2001) Synergism
of Xist RNA, DNA methylation, and histone hypoacet-
ylation in maintaining X chromosome inactivation.
J Cell Biol 153, 773–784.
11 Thomas JH (2006) Adaptive evolution in two large
families of ubiquitin-ligase adapters in nematodes and
plants. Genome Res 16, 1017–1030.
12 Tsukiyama T, Becker PB & Wu C (1994) ATP-depen-
dent nucleosome disruption at a heat-shock promoter
mediated by binding of GAGA transcription factor.
Nature 367, 525–532.
13 Huynh KD & Bardwell VJ (1998) The BCL-6 POZ
domain and other POZ domains interact with the co-
repressors N-CoR and SMRT. Oncogene 17, 2473–
2484.
14 Wang KL, Yoshida H, Lurin C & Ecker JR (2004)
Regulation of ethylene gas biosynthesis by the Arabid-
opsis ETO1 protein. Nature 428, 945–950.
15 Pintard L, Willems A & Peter M (2004) Cullin-based
ubiquitin ligases: Cul3–BTB complexes join the family.
EMBO J 23, 1681–1687.
16 Weber H, Hano P & Hellmann H (2007) The charming
complexity of CUL3. Int J Plant Dev Biol 1, 178–184.
17 Hellmann H & Estelle M (2002) Plant development:
regulation by protein degradation. Science (NY) 297,
793–797.
18 Pedmale UV & Liscum E (2007) Regulation of photo-
tropic signaling in Arabidopsis via phosphorylation
state changes in the phototropin 1-interacting protein
NPH3. J Biol Chem 282, 19992–20001.

19 Ha CM, Jun JH, Nam HG & Fletcher JC (2004)
BLADE-ON-PETIOLE1 encodes a BTB ⁄ POZ domain
protein required for leaf morphogenesis in Arabidop-
sis thaliana. Plant Cell Physiol 45, 1361–1370.
20 Kim S, Choi HI, Ryu HJ, Park JH, Kim MD &
Kim SY (2004) ARIA, an Arabidopsis arm repeat
protein interacting with a transcriptional regulator of
abscisic acid-responsive gene expression, is a novel
abscisic acid signaling component. Plant Physiol 136,
3639–3648.
21 Cheng Y, Qin G, Dai X & Zhao Y (2007) NPY1, a
BTB-NPH3-like protein, plays a critical role in auxin-
regulated organogenesis in Arabidopsis. Proc Natl Acad
Sci USA 104, 18825–18829.
22 Dong X, Li X, Zhang Y, Fan W, Kinkema M & Clarke
J (2001) Regulation of systemic acquired resistance by
NPR1 and its partners. Novartis Found Symp 236, 165–
173.
23 Sunnerhagen M, Pursglove S & Fladvad M (2002) The
new MATH: homology suggests shared binding surfaces
in meprin tetramers and TRAF trimers. FEBS Lett 530,
1–3.
24 Uren AG & Vaux DL (1996) TRAF proteins and mep-
rins share a conserved domain. Trends Biochem Sci 21,
244–245.
25 Bond JS & Beynon RJ (1995) The astacin family of
metalloendopeptidases. Protein Sci 4, 1247–1261.
26 Bertenshaw GP, Turk BE, Hubbard SJ, Matters GL,
Bylander JE, Crisman JM, Cantley LC & Bond JS
(2001) Marked differences between metalloproteases

meprin A and B in substrate and peptide bond specific-
ity. J Biol Chem 276, 13248–13255.
27 Bauvois B (2001) Transmembrane proteases in focus:
diversity and redundancy? J Leukoc Biol 70, 11–17.
28 Baker SJ & Reddy EP (1996) Transducers of life and
death: TNF receptor superfamily and associated
proteins. Oncogene 12, 1–9.
29 Arch RH & Thompson CB (1998) 4-1BB and Ox40 are
members of a tumor necrosis factor (TNF)-nerve
growth factor receptor subfamily that bind TNF
receptor-associated factors and activate nuclear factor
kappaB. Mol Cell Biol 18, 558–565.
30 Oelmu
¨
ller R, Peskan-Bergho
¨
fer T, Shahollaria B,
Trebickab A, Sherametia I & Varmac A (2005) MATH
domain proteins represent a novel protein family in
Arabidopsis thaliana, and at least one member is
modified in roots during the course of a plant–microbe
interaction. Physiol Plant 124, 152–166.
Arabidopsis thaliana BPM–ERF ⁄ AP2 interaction H. Weber and H. Hellmann
6634 FEBS Journal 276 (2009) 6624–6635 ª 2009 The Authors Journal compilation ª 2009 FEBS
31 Lin R-C, Park H-J & Wang H-Y (2008) Role of
Arabidopsis RAP2.4 in regulating light- and
ethylene-mediated developmental processes and drought
stress tolerance. Mol Plant 1, 42–57.
32 Nakano T, Suzuki K, Fujimura T & Shinshi H (2006)
Genome-wide analysis of the ERF gene family in

Arabidopsis and rice. Plant Physiol 140, 411–432.
33 Okamuro JK, Caster B, Villarroel R, Van Montagu M
& Jofuku KD (1997) The AP2 domain of APETALA2
defines a large new family of DNA binding proteins in
Arabidopsis. Proc Natl Acad Sci USA 94, 7076–7081.
34 Jofuku KD, den Boer BG, Van Montagu M &
Okamuro JK (1994) Control of Arabidopsis flower and
seed development by the homeotic gene APETALA2.
Plant Cell 6, 1211–1225.
35 Magnani E, Sjolander K & Hake S (2004) From endo-
nucleases to transcription factors: evolution of the AP2
DNA binding domain in plants. Plant Cell 16, 2265–
2277.
36 Sakuma Y, Liu Q, Dubouzet JG, Abe H, Shinozaki K
& Yamaguchi-Shinozaki K (2002) DNA-binding speci-
ficity of the ERF ⁄ AP2 domain of Arabidopsis DREBs,
transcription factors involved in dehydration- and cold-
inducible gene expression. Biochem Biophys Res
Commun 290, 998–1009.
37 Estelle MA & Somerville C (1987) Auxin-resistant
mutants of Arabidopsis thaliana with an altered mor-
phology. Mol Gen Genet 206, 200–206.
38 Clough SJ & Bent AF (1998) Floral dip: a simplified
method for Agrobacterium-mediated transformation of
Arabidopsis thaliana. Plant J 16, 735–743.
39 Minet M, Dufour ME & Lacroute F (1992) Comple-
mentation of Saccharomyces cerevisiae auxotrophic
mutants by Arabidopsis thaliana cDNAs. Plant J 2,
417–422.
40 Karimi M, Inze D & Depicker A (2002) GATEWAY

vectors for Agrobacterium-mediated plant transforma-
tion. Trends Plant Sci 7, 193–195.
41 Xiang C, Han P, Lutziger I, Wang K & Oliver DJ
(1999) A mini binary vector series for plant transforma-
tion. Plant Mol Biol 40, 711–717.
42 Dortay H, Gruhn N, Pfeifer A, Schwerdtner M,
Schmulling T & Heyl A (2008) Toward an interaction
map of the two-component signaling pathway of
Arabidopsis thaliana. J Proteome Res 7, 3649–3660.
43 Sparkes IA, Runions J, Kearns A & Hawes C (2006)
Rapid, transient expression of fluorescent fusion
proteins in tobacco plants and generation of stably
transformed plants. Nat Protoc 1, 2019–2025.
44 Campanella JJ, Bitincka L & Smalley J (2003)
MatGAT: an application that generates similarity ⁄
identity matrices using protein or DNA sequences.
BMC Bioinformatics 4, 29.
Supporting information
The following supplementary material is available:
Fig. S1. In vitro -translated A-6 ERF ⁄ AP2 proteins
At4g39780 and At4g13620 coprecipitate with
GST:BPM1, but not At1g36060.
Fig. S2. Verification of nuclear localization for
GFP:BPM1 and GFP:RAP2.4 in transient expression
experiments using Nicotiana benthamiana leaves.
Fig. S3. Generation of stable transgenic Arabidopsis
P35S:GFP:BPM4 plants.
Fig. S4. Instability of RAP2.4:myc in plants overex-
pressing myc-epitope-tagged RAP2.4 under the control
of a 35S promoter (pro35S:RAP2.4:myc).

Table S1. Comparison of whole BPM proteins (A) and
the BPM MATH domains only (B). Similarities and
identities between the amino acid sequences were deter-
mined by matgat 2.02 software using the default set-
tings (BLOSUM62, first gap 12, extending gap 2) [44].
Table S2. Overview of primers used in this work and
expected PCR products.
This supplementary material can be found in the
online version of this article article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.
H. Weber and H. Hellmann Arabidopsis thaliana BPM–ERF ⁄ AP2 interaction
FEBS Journal 276 (2009) 6624–6635 ª 2009 The Authors Journal compilation ª 2009 FEBS 6635