Slc12a2 is a direct target of two closely related homeobox
proteins, Six1 and Six4
Zen-ichi Ando, Shigeru Sato, Keiko Ikeda and Kiyoshi Kawakami
Division of Biology, Center for Molecular Medicine, Jichi Medical School, Tochigi, Japan
The Six homeobox gene is characterized by the con-
served Six domain (SD) and homeodomain (HD),
both of which are required for specific DNA binding
[1,2]. The prototype of this gene family is Drosophila
sine oculis, which is essential for compound eye for-
mation [3,4]. Six members (Six1–Six6) of the Six
family gene have been identified in mouse and
human [1]. Six3 and Six6 are essential for forebrain
formation and eye development [5–11], whereas Six5
is involved in cataractogenesis and spermatogenesis
[12–14].
Among the Six family genes, Six1 and Six4 show
a remarkably similar expression pattern [1,15,16]. Both
Six1 and Six4 bind to the MEF3 site in the myogenin
promoter and positively regulate the activity of the
promoter in conjunction with their coactivator Eya
proteins [17,18]. Based on these observations, Six1 and
Six4 are thought to be functionally similar in vivo.
Six4
– ⁄ –
mice showed little anomaly in embryogenesis
including skeletal muscles [16]. This was explained by
the compensatory function of Six1 considering the
similar expression pattern of both genes and the func-
tional similarity in activating their target gene myo-
genin. In contrast, Six1
– ⁄ –

mice show
defective formation of various organs such as inner ear, nose, skeletal mus-
cle, kidney and thymus, whereas Six4
– ⁄ –
mice show little anomaly in organo-
genesis. To understand the molecular basis for the differential function of
Six1 and Six4 in vivo, we screened target genes of Six1 and Six4 and found
that Six1 and Six4 differentially regulated a set of target genes. Gel-retarda-
tion assays indicated that the promoter region of one of the targets, sodium–
potassium–chloride cotransporter 1 (Slc12a2), contains multiple Six1-binding
sites and one common binding site of Six1 and Six4, suggesting that the
DNA-binding specificity of Six1 is distinct from that of Six4. This underlies
the differential regulation of common target genes by Six1 and Six4. Fur-
thermore, in situ hybridization demonstrated that the expression of Slc12a2
was reduced in the developing dorsal root ganglia of Six1
– ⁄ –
⁄ Six4
– ⁄ –
mice,
suggesting that Six1 and Six4 regulate Slc12a2 in vivo.
Abbreviations
Atp1a1, sodium–potassium ATPase alpha 1 subunit; Clcn5, chloride channel 5; Coll2a1, procollagen type 2 alpha 1; Figf,c-fos induced
growth factor; Gas1, growth arrest-specific 1; GST, glutathione S-transferase; HD, homeodomain; MCK, muscle creatine kinase; SD, Six
domain; Slc12a2, sodium–potassium–chloride cotransporter 1; Trex, transcriptional regulatory element X.
3026 FEBS Journal 272 (2005) 3026–3041 ª 2005 FEBS
three putative categories of targets, Six1-specific
targets, Six4-specific targets and common targets. We
then focused our analysis on one of the typical com-
mon target genes, sodium-potassium-chloride cotrans-
porter 1 (Slc12a2) to understand the molecular basis

source of poly(A)
+
RNA to prepare hybridization
probes. Expression profiling was performed by hybridi-
zation of an oligo microarray containing 20 371 mouse
cDNAs. Scatter plot analysis of the microarray data
showed that most of the data points fell along the
diagonal, indicating that most of the genes were
equally expressed in the two samples (supplementary
Fig. S1). The genes that showed more than 1.5-fold
higher level of expression in cells infected with VP16–
Six1 adenovirus compared with cells infected with
VP16–Six1W171R were considered potential Six1 tar-
get genes, whereas 1.5-fold higher level of expression
in cells infected with VP16–Six4 compared with VP16–
Six4W263R were considered potential Six4 target
genes. A total of 363 Six1 target genes and 149 Six4
targets genes were identified. Of these, 63 were com-
mon target genes. The data were deposited in GEO
database GSE2043 (a complete list of these genes
appears in Table S1 and a partial list is shown in
Table 1). Because only a single microarray was used
for each condition, genes with a relatively low level of
expression were excluded. Examples of target genes
included the cyclin-dependent kinase inhibitor 1C
(Cdkn1c), a cell-cycle regulator expressed in the kidney
and cochlea, which controls the number of podocytes
and glomerular size in the kidney [26]. Another exam-
ple is the sodium–potassium–chloride cotransporter 1
(Slc12a2) that plays an important role in dorsal root

luciferase reporter constructs harboring the Gas1 pro-
moter region of )3408 to +19 (Gas1)3408Luc), the
Clcn5 promoter region of )1325 to +2529 (Clcn5)
1325Luc) and the Slc12a2 promoter region of )1938
to +149 (Slc12a2)1938Luc). Gas1 promoter showed
similar extent of activation by both Six1 and Six4 in
a dose-dependent manner (Fig. 1A), although it was
listed as a potential Six1-specific target gene. Clcn5 pro-
moter showed a moderate repression by Six1 and a
strong repression by Six4 in a dose-dependent manner
Z-i. Ando et al. Target genes of Six1 and Six4
FEBS Journal 272 (2005) 3026–3041 ª 2005 FEBS 3027
Table 1. A partial list of potential target genes regulated by Six1 and ⁄ or Six4 in mk4 cells. Expression profiling was performed using the
Mouse Development Oligo Microarray containing 20 371 mouse cDNAs (Agilent Technologies, Palo Alto, CA). Total RNA was prepared as
described previously [24] from mK4 cells infected with recombinant viruses. To prepare fluorescence-labeled cDNA probes, total RNA
(20 lg) was reverse transcribed using an oligo(dT) primer in the presence of aminoallyl dUTP and single-stranded cDNAs were coupled with
Cy3 (VP16–Six1wt and VP16–Six4wt infected samples) or Cy5 (VP16–Six1W171R and VP16–Six4W263R samples) dyes. Hybridization to the
microarray was carried out at 65 °C for 17 h according to the instructions provided by the manufacturer. The arrays were washed, dried and
scanned using ScanArray 5000 (GSI Lumonics Inc., ON, Canada). Cy3 and Cy5 intensities for each spot on the array were determined by
QUANTARRAY software (GSI Lumonics Inc.). The raw data were processed and the Cy3 to Cy5 ratios were calculated as follows: (a) subtrac-
tion of the fluorescence intensity of negative control spots as background from the intensity of each of the Cy3 and Cy5 spots, (b) normaliza-
tion of the entire data set using the global normalization method, (c) elimination of spots with high background intensity for either dye, (d)
determination of the Cy3 to Cy5 ratios. The microarray data were deposited in the GEO database under accession number GSE2043. Among
the genes with a Cy3 ⁄ Cy5 ratio of > 1.5, those with a relatively high level of expression (normalized signal value > 500 in both dyes) were
considered to be potential Six1- and Six4-target genes. Data are from a single microarray experiment and no dye-swap experiment was car-
ried out.
Symbol
VP16–Six1
a
wt ⁄ W171R

Cpd 2.65 Carboxypeptidase D
Gas1 2.17 Growth arrest specific 1
Laptm4a 2.12 Lysosomal-associated protein transmembrane 4 A
Fmr1 2.05 Fragile–mental retardation syndrome 1 homolog
Bmi1 1.96 B lymphoma Mo-MLV insertion region 1
Nrp2 1.86 Neuropilin 2
Irs1 1.85 Insulin receptor substrate 1
Mbnl1 1.84 Muscleblind-like 1 (Drosophila)
Clcn5 1.83 Yes Chloride channel 5 [55]
Cav2 1.73 Caveolin 2
Aqp1 1.66 Yes Aquaporin 1 [50,56]
Bmpr2 1.62 Yes Bone morphogenic protein receptor, type II [57]
Six4-specific target genes
Cyb561d2 2.29 Cytochrome b-561 domain containing 2
Ifitm3 2.27 Interferon induced transmembrane protein 3
Gpd2 2.24 Glycerol phosphate dehydrogenase 2, mitochondrial
Thbs1 2.21 Thrombospondin 1
Col2a1 2.20 Procollagen, type II, alpha 1
Matn2 2.05 Matrilin 2
Tmem2 2.01 Transmembrane protein 2
a
Ratio of expression level in VP16–Six1wt-expressing mK4 cells to expression level in VP16–Six1W171R-expressing cells.
b
Ratio of expres-
sion level in VP16–Six4wt-expressing mK4 cells to expression level in VP16–Six4W263R-expressing cells.
c
Genes involved in kidney devel-
opment or function.
Target genes of Six1 and Six4 Z-i. Ando et al.
3028 FEBS Journal 272 (2005) 3026–3041 ª 2005 FEBS

c
Six1–Six4 common targets
Ogn 1.91 (0.07) 1.88 (0.02) 20
Amot 1.99 (0.20) 2.02 (0.01) 20
Rhoe 2.71 (0.24) 2.59 (0.02) 20
Nid2 1.62 (0.06) 1.63 (0.04) 24
Acadm 1.67 (0.01) 2.04 (0.13) 24
Cdkn1c 2.09 (0.10) 3.10 (0.03) 24
Slc12a2 1.60 (0.04) 2.28 (0.11) 22
Six1-specific target genes
Figf 2.11 (0.08) 1.51 (0.15) 24
Gas1 2.15 (0.10) 1.28 (0.09) 20
Clcn5 1.86 (0.03) 2.03 (0.09) 20
Six4-specific target genes
Ifitm3 1.90 (0.02) 5.52 (0.27) 20
Thbs1 0.81 (0.05) 2.77 (0.12) 16
Col2a1 1.33 (0.06) 1.63 (0.04) 28
a
Ratio of expression level in VP16–Six1wt-expressing mk4 cells to
expression level in VP16–Six1W171R-expressing cells.
b
Ratio of
expression level in VP16–Six4wt-expressing mk4 cells to expres-
sion level in VP16–Six4W263R-expressing cells.
c
The number of
PCR cycles at which the relative amount of PCR products were
determined.
AB
10

0
Fold activation
Gas1-3408Luc Clcn5-1325Luc
Slc12a2-1938Luc pGL3MG-185
E
Six1
Slc12a2
pGL3MG
Gas1 Clcn5
Six4
Fig. 1. Differential regulations of the potential target genes by Six1
and Six4. Transient transfection assays with 175 ng of the indicated
promoter-luciferase reporter construct were carried out as des-
cribed in Experimental procedures. Luciferase activity was normal-
ized to the protein content and expressed relative to the value in
the presence of 75 ng pFLAG-CMV2 (white bars), which was set at
1. As effectors, increasing amounts (25 and 75 ng) of pfSix1 (black
bars) and pfSix4 (gray bars) were cotransfected into COS7 cells.
Data are mean ± SEM of three independent experiments (each per-
formed in duplicate). (A) Trans-activation of Gas1 promoter by Six1
and Six4. (B) Regulation of Clcn5 promoter by Six1 and Six4. (C)
Trans-activation of Slc12a2 promoter by Six1 and Six4. (D) Trans-
activation of myogenin promoter by Six1 and Six4. (E) Gel-retarda-
tion assays of Flag–Six1 and Flag–Six4 in nuclear extracts (800 ng
protein for Six1 and 860 ng protein for Six4) from transiently trans-
fected cells using 5 fmol of C3 oligonucleotide probe. Arrows indi-
cate the positions of specific retarded complexes and arrowhead
indicates the position of nonspecific complex.
Z-i. Ando et al. Target genes of Six1 and Six4
FEBS Journal 272 (2005) 3026–3041 ª 2005 FEBS 3029

ment ()97 to +77, Fig. 2B, probe A), Cfr10I–MspI
fragment (+2 to +103, Fig. 2B, probe B) and HinfI–
SacI fragment (+75 to +149, Fig. 2B, probe C). The
expressed glutathione S-transferase (GST) fusion pro-
tein of Six1 (GST–Six1) bound to probes A, B and C,
whereas GST fusion protein of Six4 (GST–Six4) bound
to probe C, but not to probes A and B (Fig. 2C).
These results suggest the existence of at least two Six1-
specific binding sites, and only one binding site com-
mon to Six1 and Six4.
To analyse the location of each binding site, we syn-
thesized the double-stranded competitor oligonucleo-
tides (oligo 1 to oligo 9) that covered various portions
of the promoter region (Fig. 3A). The formation of a
complex by Six1 was strongly competed by oligos 2, 3
and 6 for probe A, by oligos 6 and 7 for probe B and
by oligos 6, 7 and 9 for probe C (Fig. 3B). In contrast,
the formation of a complex by Six4 was competed only
by oligo 9 for probe C (Fig. 3C).
To precisely localize the binding element for Six1
and Six4, we generated 4-bp substitution mutations
in various regions of oligo 3, oligo 6 and oligo 9
(Table 3) and each mutated oligo was added to the gel
retardation assay mixture to examine the competition
to the binding of Six1 and probes A, B and C. Oligo
C
probe A B C A B C
GST-Six1 GST-Six4
123456
complex

proteins with each 5 fmol probe indicated in (B). Lanes 1–3 and
lanes 4–6 were from separated lanes in the same gel. Free probes
are indicated by the arrows and shifted complexes are indicated by
the bracket.
Target genes of Six1 and Six4 Z-i. Ando et al.
3030 FEBS Journal 272 (2005) 3026–3041 ª 2005 FEBS
3mut()13 ⁄ )10) showed the most reduced competition
among the mutated oligos examined (Fig. 4A, lane 7).
The reduction of Six1 binding was confirmed by com-
paring the binding of Six1 to the oligo 3wt probe and
the oligo 3mut()13 ⁄ )10) probe. The amount of the
retarded complex composed of Six1 and oligo
3mut()13 ⁄ )10) was approximately threefold less com-
pared with oligo 3wt (Fig. 4B, lanes 2 and 4). As for
probe B, all substitution mutations of oligo 6 showed
similar competition compared with oligo 6wt (data not
shown), suggesting the presence of multiple binding
sites for Six1 in oligo 6. We also tested deletion oligos
and found that oligo 6 with 10-bp deletion in +65 to
+74 [oligo 6del(+65 ⁄ +74)] slightly reduced the com-
petition compared with oligo 6wt (Fig. 4A, lanes 11
and 12). The double mutation oligo 6 containing
del(+65 ⁄ +74) and substitution mutation in +89 to
+92 [del(+65 ⁄ +74)⁄ mut(+89 ⁄ +92)] showed further
reduced competition against the probe B–Six1 protein
complex formation (Fig. 4A, lane 13). The reduction
in the binding was confirmed by comparing the bind-
ing of Six1 to the oligo 6wt probe and oligo
6del(+65 ⁄ +74) ⁄ mut(+89 ⁄ +92) probe and the bind-
ing was approximately threefold weaker in the latter

456 7
13141516
17
6789
probe
competitor
probe
competitor
12345678910
C
123456789
probe A
probe B
probe C
probe
oligo 1 -99 -62
oligo 2
-68
-31
oligo 3
-37
+3
oligo 4
-4
+34
oligo 5
+28
+65
oligo 6
+59

or absence (lane 1) of competitors; 500-fold molar excess compet-
itor oligonucleotides were added. The intensities of the retarded
bands relative to that in the absence of competitors, which was
set at 1(lane 1), are shown at the bottom. Free probes are indicated
by the arrow and shifted complexes by the bracket in (B) and (C).
Z-i. Ando et al. Target genes of Six1 and Six4
FEBS Journal 272 (2005) 3026–3041 ª 2005 FEBS 3031
mutation reporters, Slc12a2–97mut()13 ⁄ )10)Luc,
Slc12a2–97del(+65 ⁄ +74) ⁄ mut(+89 ⁄ +92)Luc, and the
combination of the two, Slc12a2–97mut()13 ⁄ )10) ⁄
del(+65 ⁄ +74) ⁄ mut(+89 ⁄ +92)Luc as Six1-binding muta-
tions. We also prepared Slc12a2–97mut(+135 ⁄ +138)Luc,
which abolished Six4 binding. We performed reporter
gene assays and analysed the effects of Six1 (Fig. 5A)
and Six4 (Fig. 5B). Cotransfection of pfSix1 showed
activation of Slc12a2–97Luc in a dose-dependent man-
ner to an  11-fold increase in the level of luciferase
activity. The activation level was decreased to approxi-
mately sixfold in Slc12a2mut()13 ⁄ )10)Luc, whereas
Slc12a2del(+65 ⁄ +74) ⁄ mut(+89⁄ +92)Luc showed com-
parable luciferase activity to the wild-type reporter
construct, Slc12a2–97Luc. In contrast, the combina-
tion of these two mutations, Slc12a2mut()13 ⁄ )10) ⁄
del(+65 ⁄ +74) ⁄ mut(+89 ⁄ +92)Luc, showed only 1.3-
fold activation by Six1 (Fig. 5A). These results
clearly indicate that the three Six1-binding sites
located at around )13 ⁄ )10, +65 ⁄ +74 and
+89 ⁄ +92 are the responsible element as a whole for
the activation by Six1. As for the effects of Six4,
cotransfection of pfSix4 showed 2.6- to 4.6-fold acti-

mice die soon after
birth and they show developmental defects in various
organs [23] (unpublished observation), we compared the
expression pattern of Slc12a2 in these embryos. The
expression level of Slc12a2 was too low to allow precise
comparison of its expression level in the nephrogenic
cord. Therefore, we analysed the expression of the gene
by in situ hybridization in the dorsal root ganglia, where
Six1 and Six4 were abundantly expressed and some
developmental abnormalities were observed in Six1
– ⁄ –
⁄
Six4
– ⁄ –
mice (K. Ikeda & K. Kawakami, unpublished
observation). The antisense probe detected expressions
of Slc12a2 in the dorsal root ganglia in E18.5 fetus
where significant expression was reported [38] (Fig. 6C),
12345678 1234
probe C
9wt
9mut(+117/+120)
9mut(+123/+126)
9mut(+129/+132)
9mut(+135/+138)
9mut(+141/+144)
9mut(+147/+150)
oligo 9wt
oligo 9mut(+135/+138)
CDE

6wt
6del(+65/+74)
6del(+65/+74)
/mut(+89/+92)
1.0 0.1 0.2 0.3
123456789
probe A
3wt
3mut(-37/-34)
3mut(-1/+3)
3mut(-7/-4)
3mut(-13/-10)
3mut(-19/-16)
3mut(-25/-22)
3mut(-31/-28)
1.0 0.2 0.2 0.1 0.2 0.1 0.3 0.2 0.2
AB
11121314
Z-i. Ando et al. Target genes of Six1 and Six4
FEBS Journal 272 (2005) 3026–3041 ª 2005 FEBS 3033
whereas the control sense probe gave only background
signals (Fig. 6D). The expression level of Slc12a2 was
apparently lower in Six1
– ⁄ –
⁄ Six4
– ⁄ –
embryo at E16.5
compared with the wild-type, whereas that in Six1
– ⁄ –
was similar to the wild-type (Fig. 6H–J). We observed

– ⁄ –
embryo were significantly reduced
(to 36.4% for Figf and to 58.1% for Col2a1 compared
with the wild-type), suggesting that these genes are also
regulated by Six1 and Six4 in vivo (Fig. 7).
Discussion
Screening of putative target genes of Six1 and
Six4 and effectiveness of the screening method
To understand the function of transcription factors
in organ development, it is essential to identify direct
Fig. 4. Identification of Six1- and Six4-binding sites in the Slc12a2 promoter by gel retardation assays. The competitor oligonucleotides har-
boring mutations are shown in Table 3. (A) Competition assays were performed with GST–Six1 protein in the absence (lanes 1, 10 and 14)
or presence of 500-fold molar excess of various competitor oligonucleotides (lanes 2–9, 11–13 and 15–21). Probes A, B and C indicated in
Fig. 2 were used. The protein amount of GST–Six1 was 20 ng for probe A (lanes 1–9), 15 ng for probe B (lanes 10–13), and 16 ng for probe
C (lanes 14–21). The intensities of the retarded bands relative to that in the absence of competitors for each probe, which was set at 1
(lanes 1, 10 and 14), are shown at the bottom. Shifted complexes were least competed by oligo 3mut()13 ⁄ )10) (lane 7), oligo 6-
del(+65 ⁄ +74) ⁄ mut(+89 ⁄ +92) (lane 13), and oligo 9mut(+135 ⁄ +138) (lane 19). (B) Comparison of binding of GST–Six1 to the wild-type (oli-
go 3wt, lane 2; oligo 6wt, lane 6; oligo 9wt, lane 12) and to the mutated probes [oligo 3mut()13 ⁄ )10), lane 4; oligo 6del(+65 ⁄ +74) and
oligo 6del(+65 ⁄ +74) ⁄ mut(+89 ⁄ +92), lanes 8 and 10, respectively; oligo 9mut(+135 ⁄ +138), lane 14]. Sixty-seven nanograms of GST–Six1
was used in the reactions. Binding of GST–Six1 to oligo 3mut()13 ⁄ )10) was  30% compared with that of oligo 3wt (lanes 2 and 4). Binding
of GST–Six1 to oligo 6del(+65 ⁄ +74) and oligo 6del(+65 ⁄ +74) ⁄ mut(+89 ⁄ +92) was  60 and 30%, respectively, compared with that of oli-
go 6wt (lanes 6, 8 and 10). Binding of GST–Six1 to oligo 9mut(+135 ⁄ +138) was  30% compared with that of oligo 9wt (lanes 12 and 14).
Lanes 5–6 and lanes 7–10, lane 11 and lanes 12–14 are from separate lanes of one gel. (C) Competition assays were performed with GST–
Six4 protein in the absence (lane 1) or presence of 500-fold molar excess of various competitor oligonucleotides (lanes 2–8). Probe C indica-
ted in Fig. 2 was used. Thirty nanograms of GST–Six4 was used in the reaction. The intensities of the retarded bands relative to that in the
absence of competitors for the probe, which was set at 1(lane 1), are shown at the bottom. Shifted complexes were most weakly com-
peted by the oligo 9mut(+135 ⁄ +138) (lane 6). (D) Comparison of binding of GST–Six4 to the wild-type (oligo 9wt, lane 2) and mutated probes
[oligo 9mut(+135 ⁄ +138), lane 4]. Thirty nanograms of GST–Six4 was used in the reactions. Binding of GST–Six4 to oligo 9mut(+135 ⁄ +138)
was  10% compared with that of oligo 9wt (lanes 2 and 4). (E) Schematic representation of binding elements of Six1 and Six4 in the
Slc12a2 promoter. Six1-specific binding elements reside in the region of )13 ⁄ )10, +65 ⁄ +74 and +89 ⁄ +92 (black ovals) and a bipartite Six1-

Target genes of Six1 and Six4 Z-i. Ando et al.
3034 FEBS Journal 272 (2005) 3026–3041 ª 2005 FEBS
target genes and recognize the gene cascade as well as
regulatory mechanisms involved in proper cell growth,
differentiation, cell movement and functional matura-
tion of the organ. In this analysis, we took advantage
of a model cell line that reflects a certain developmen-
tal stage of the kidney in order to identify the genes
N
HI J
Dorsal root
ganglion
Choroid
plexsus
Wild type
E
O
FG
K M
L
P
sp
drg
v
v
sp
drg
sp
drg
v

on the choroid plexus from E16.5 wild-type (K, N), Six1
– ⁄ –
(L. O) and Six1
– ⁄ –
⁄ Six4
– ⁄ –
(M, P) embryos stained with hematoxylin (K–M) and
with antisense probe (N–P) are shown. Similar expression levels of Slc12a2 were observed in all genotypes (N–P). Scale bar: 100 lm.
Z-i. Ando et al. Target genes of Six1 and Six4
FEBS Journal 272 (2005) 3026–3041 ª 2005 FEBS 3035
that are driven by Six1 and ⁄ or Six4. A similar strategy
was previously applied to identify target genes of Six5
using P19 carcinoma cells, and direct target genes
Igfbp5 and Igf2 were successfully identified [24]. By
overexpressing constitutively transcriptionally active
forms of VP16–Six1 and VP16–Six4, we expected that
genes accessible by Six1 and Six4 will be activated in
the cells. In this screening, we used mK4 cells, which
possess the characteristics of embryonic metanephric
mesenchyme [25]. As expected, we identified the poten-
tial target genes that might be involved in cell growth,
differentiation and specific ion transport functions in
the kidney. Of these, numerous terminal differentiation
genes were observed such as Aqp1, Atp1a1, Clcn5,
Npr3 and Slc12a2. Although Six1 is involved in the
early phase of kidney development, such as aggrega-
tion of metanephric mesenchyme around the ureteric
bud [21], it is not surprising that Six1 and Six4 pro-
teins also directly regulate the genes involved in the
terminal differentiation. Many terminal differentiation

We previously reported that Six2, Six4 and Six5 activa-
ted the myogenin promoter, and such activation was
enhanced by Eya proteins [18]. Furthermore, the extent
of activation was dependent on the combinations of
Six and Eya. In this study, we identified various types
of target genes that showed similar activation by Six1
and Six4 (Gas1), that were efficiently activated by Six1
compared with Six4 (Slc12a2) and that were repressed
by Six1 and even more strongly by Six4 (Clcn5). Such
differential regulation of each target gene by Six1 and
Six4 may reflect some aspects of the regulatory action
in vivo, which should be clarified in future analyses.
The presence of target genes activated and repressed
by Six1 is consistent with the recent finding that
Xenopus six1 affects ectodermal genes through both
transcriptional activation and repression [40].
To explore the molecular mechanism of the differen-
tial regulation by Six1 and Six4, we analysed the
responsive elements of Slc12a2 gene promoter. Our
results showed that the number of Six1-binding sites is
at least three, whereas there is one Six4-binding site,
indicating that the DNA-binding specificity of Six1 is
distinct from that of Six4. Another member of the
Six1 ⁄ 2 subfamily, Six2, showed similar features, i.e.
residual activation by Six2 even in a mutation myoge-
nin promoter construct of the MEF3 site, which abol-
ished Six4 and Six2 binding, due to the presence of
additional Six2-binding sites other than MEF3 that are
not identified in the myogenin promoter region [18].
Although we did not address the structural basis of

lier stages of otic and nasal development and in the
formation of branchial arch and some cranial ganglia,
which were not observed in Six4
– ⁄ –
mice and Six1
– ⁄ –
mice (K. Ikeda & K. Kawakami, unpublished observa-
tion). These observations indicate that the Six1 and
Six4 mutually compensate for their functions in the
early stage of development, suggesting a common role.
DNA-binding sequences of Six1 and Six4
As for the recognition sequence of Six1, which were
identified in this study, it is difficult to predict the con-
sensus sequence of the binding sites. The binding site
of the most similar protein Six2 has recently been
reported in Gdnf gene promoter [41]. It contains two
Six2-binding sites that show similarity to the homeo-
domain binding core sequence TAAT [41]. However,
this core sequence was not found in binding sequences
of Six1 identified in this study. The binding of Six1
may not depend on the exact DNA sequence but
rather some structural features like nonsequence-speci-
fic HMG protein DNA recognition [42].
A new binding site of Six4, the transcriptional regu-
latory element X (Trex), which is the positive control
element within the muscle creatine kinase (MCK)
enhancer, has been recently reported [43]. We found
(G)ACCCGAG, a single mismatch sequence of the
MCK Trex, in the +129 ⁄ +138 region of the Slc12a2
promoter.

– ⁄ –
⁄ Six4
– ⁄ –
embryos compared with wild-type.
This result suggests that these genes are upregulated
by Six1 and Six4 in the developing embryo.
Experimental procedures
Plasmid construction
Construction of expression plasmids and luciferase reporter
plasmids is described briefly here and in full in the supple-
mentary material.
Plasmids expressing FLAG-tagged Six1 and Six4 (pfSix1
and pfSix4) were constructed as described previously
[18,44]. To construct constitutively active Six1 and Six4, a
transcription activation domain of herpes simplex virus
virion protein 16 (VP16) [45] was fused to the N-terminal
end of full-length Six proteins [pCS2 + FLAG VP16–
Six1wt and pCS2 + FLAG VP16–Six4wt]. As a control,
mutated proteins harboring a tryptophan (W) to arginine
(R) substitution were used [pCS2 + FLAG VP16–
Six1W171R and pCS2 + FLAG VP16–Six4W263R]. The
corresponding mutation in the Six5 HD abolishes DNA-
binding activity [24].
Plasmids harboring the 5¢ upstream region of Slc12a2
were constructed by inserting appropriate fragments pre-
pared from pGL3-2065 [46] into pGL3-basic (Promega
Bioscience, San Luis Obispo, CA). To construct Clcn5)
1325Luc, a mouse Clcn5 promoter fragment ()1325 to
+2529) was amplified from C57BL ⁄ 6 genomic DNA and
ligated into pGL3-basic. To construct Gas1)3409Luc, a

VP16–Six1W171R (200), AxCAwt VP16–Six4wt (50) and
AxCAwt VP16–Six4W263R (200). Transfections were car-
ried out using CellPhect (Amersham Biosciences, Piscata-
way, NJ) or Lipofectamine 2000 (Invitrogen, Carlsbad, CA)
in 24-well plates. The cells were harvested after two days.
Luciferase activity was normalized for total protein content
in cell lysates. Data were shown as mean ± SEM.
Semiquantitative RT-PCR analysis
Total RNA used in the microarray analysis (10 or
100 ng) or total RNA prepared from mouse embryos
(100 ng) was subjected to RT-PCR using OneStep
RT-PCR kit (Qiagen, Hilden, Germany) and the PCR
primers listed in the supplemental Table S2. Each set of
PCR primers was derived from different exons except
that for the intronless Gas1. For RT-PCR analysis of
Gas1, RNA samples were treated with RNase-free DN-
aseI and checked for the absence of visible PCR products
even after 32 cycles of amplification as described pre-
viously [24]. Aliquots of PCR products were removed
from the thermal cycler at multiple cycle numbers, separ-
ated on a 5% acrylamide gel, stained with the fluorescent
dye Vistra Green (Amersham Biosciences) and scanned
using the STORM system (Amersham Biosciences). Quan-
titation of the amplified products was carried out using
the STORM system and imagequant software (Mole-
cular Dynamics, Sunnyvale, CA). Linearity of PCR
amplification was maintained over several cycles, and the
amount of PCR products at 16–28 cycles, depending on
the gene, was selected for comparison. Three independent
RT-PCR reactions were set up from the same RNA

cDNA probe and a sense cDNA probe (complementary to
the antisense) for mouse Slc12a2 mRNA were designed as
follows; Slc12a2 antisense, 5¢-ATCTTCACAAGAAAAAT
CACCTGGTACCAAGGATGT; Slc12a2 sense, 5¢-ACAT
CCTTGGTACCAGGTGATTTTTCTTGTGAAGAT.
All experimental protocols described in this study were
approved by the Ethics Review Committee for Animal
Experimentation of Jichi Medical School.
Acknowledgements
We thank S. S. Potter for providing mK4 cells, R. de
Martin for pUBT-luc ⁄ Gas1 3.5 k and E. Delpire for
pGL3-2065 plasmids. We also thank M. Nakamura,
Y. Goto and Y. Takano for the expert technical assist-
ance. This work was supported by grants from Minis-
try of Education, Culture, Sports, Science and
Technology of JAPAN (KK and SS), The Science
Research Promotion Fund from The Promotion and
Mutual Aid Corporation for Private Schools of Japan
(KK) and The Research Award to JMS Graduate
Student (ZA).
References
1 Kawakami K, Sato S, Ozaki H & Ikeda K (2000) Six
family genes – structure and function as transcription
factors and their roles in development. Bioessays 22,
616–626.
2 Kawakami K, Ohto H, Ikeda K & Roeder RG (1996)
Structure, function and expression of a murine homeo-
box protein AREC3, a homologue of Drosophila sine
oculis gene product, and implication in development.
Nucleic Acids Res 24, 303–310.

Solnica-Krezel L & Oliver G (2003) Six3 repression of
Wnt signaling in the anterior neuroectoderm is essential
for vertebrate forebrain development. Genes Dev 17,
368–379.
11 Lo
´
pez-Rı
´
os J, Tessmar K, Loosli F, Wittbrodt J &
Bovolenta P (2003) Six3 and Six6 activity is modulated
by members of the groucho family. Development 130 ,
185–195.
12 Sarkar PS, Appukuttan B, Han J, Ito Y, Ai C, Tsai W,
Chai Y, Stout JT & Reddy S (2000) Heterozygous loss
of Six5 in mice is sufficient to cause ocular cataracts.
Nat Genet 25, 110–114.
13 Klesert TR, Cho DH, Clark JI, Maylie J, Adelman J,
Snider L, Yuen EC, Soriano P & Tapscott SJ (2000)
Mice deficient in Six5 develop cataracts: implications
for myotonic dystrophy. Nat Genet 25, 105–109.
14 Sarkar PS, Paul S, Han J & Reddy S (2004) Six5 is
required for spermatogenic cell survival and spermio-
genesis. Hum Mol Genet 13, 1421–1431.
15 Oliver G, Wehr R, Jenkins NA, Copeland NG,
Cheyette BNR, Hartenstein V, Zipursky SL & Gruss P
(1995) Homeobox genes and connective tissue pattern-
ing. Development 121, 693–705.
16 Ozaki H, Watanabe Y, Takahashi K, Kitamura K,
Tanaka A, Urase K, Momoi T, Sudo K, Sakagami J,
Asano M et al. (2001) Six4, a putative myogenin gene

myotonic dystrophy type 1. Hum Mol Genet 11,
1045–1058.
25 Valerius MT, Patterson LT, Witte DP & Potter SS
(2002) Microarray analysis of novel cell lines represent-
ing two stages of metanephric mesenchyme differentia-
tion. Mech Dev 110, 151–164.
26 Tomari S, Nagahama H, Shu Y, Hoshi S, Nakayama
K, Nakayama KI & Nagata M (2002) Glomerular dif-
ferentiation in p27 and p57 double-mutant metanephroi.
Anat Embryol 206, 31–36.
27 Sung KW, Kirby M, McDonald MP, Lovinger DM &
Delpire E (2000) Abnormal GABA
A
receptor-mediated
currents in dorsal root ganglion neurons isolated from
Na-K-2Cl cotransporter null mice. J Neurosci 20,
7531–7538.
28 Pace AJ, Lee E, Athirakul K, Coffman TM, O’Brien
DA & Koller BH (2000) Failure of spermatogenesis in
mouse lines deficient in the Na
+
-K
+
-2Cl
–
cotransport-
er. J Clin Invest 105, 441–450.
29 Delpire E, Lu J, England R, Dull C & Thorne T (1999)
Deafness and imbalance associated with inactivation of
the secretory Na-K-2Cl co-transporter. Nat Genet 22,

¨
nther W, Piwon N & Jentsch TJ (2003) The ClC-5
chloride channel knock-out mouse – an animal model
for Dent’s disease. Pflugers Arch 445, 456–462.
37 Donahue LR, Chang B, Mohan S, Miyakoshi N,
Wergedal JE, Baylink DJ, Hawes NL, Rosen CJ,
Ward-Bailey P, Cheng QY et al. (2003) A missense
mutation in the mouse Col2a1 gene causes spondyloepi-
physeal dysplasia congenita, hearing loss, and retino-
schisis. J Bone Mineral Res 18, 1612–1621.
38 Hu
¨
bner CA, Lorke DE & Hermans-Borgmeyer I
(2001) Expression of the Na-K-2Cl-cotransporter
NKCC1 during mouse development. Mech Dev 102,
267–269.
39 Bergstrom DA, Penn BH, Strand A, Perry RLS,
Rudnicki MA & Tapscott SJ (2002) Promoter-specific
regulation of MyoD binding and signal transduction
cooperate to pattern gene expression. Mol Cell 9,
587–600.
40 Brugmann SA, Pandur PD, Kenyon KL, Pignoni F &
Moody SA (2004) Six1 promotes a placodal fate within
the lateral neurogenic ectoderm by functioning as both
a transcriptional activator and repressor. Development
131, 5871–5881.
41 Brodbeck S, Besenbeck B & Englert C (2004) The
transcription factor Six2 activates expression of the Gdnf
gene as well as its own promoter. Mech Dev 121,
1211–1222.

ture and regulation of the growth arrest-specific 1 (gas-
1) promoter. J Biol Chem 268, 22788–22793.
48 Kawakami K, Scheidereit C & Roeder RG (1988)
Identification and purification of human immunoglobu-
lin-enhancer-binding protein (NF-jB) that activates
transcription from a human immunodeficiency virus
type 1 promoter in vitro. Proc Natl Acad Sci USA 85,
4700–4704.
49 Toyoda H, Ohno K, Yamada J, Ikeda M, Okabe A,
Sato K, Hashimoto K & Fukuda A (2003) Induction of
NMDA and GABA
A
receptor-mediated Ca
2+
oscilla-
tions with KCC2 mRNA downregulation in injured
facial motoneurons. J Neurophysiol 89, 1353–1362.
50 Schwab K, Patterson LT, Aronow BJ, Luckas R, Liang
HC & Potter SS (2003) A catalogue of gene expression
in the developing kidney. Kidney Int 64, 1588–1604.
51 Miosge N, Ko
¨
ther F, Heinemann S, Kohfeldt E, Her-
ken R & Timpl R (2000) Ultrastructural colocalization
of nidogen-1 and nidogen-2 with laminin-1 in murine
kidney basement membranes. Histochem Cell Biol 113,
115–124.
52 Nagahama H, Hatakeyama S, Nakayama K, Nagata M,
Tomita K & Nakayama K (2001) Spatial and temporal
expression patterns of the cyclin-dependent kinase

aquaporin-1 water channels. J Biol Chem 273, 4296–
4299.
57 Martinez G, Loveland KL, Clark AT, Dziadek M &
Bertram JF (2001) Expression of bone morphogenetic
protein receptors in the developing mouse metanephros.
Exp Nephrol 9, 372–379.
Supplementary material
The following material is available from http://www.
blackwellpublishing.com/products/journals/suppmat/EJB/
EJB4716/EJB4716sm.htm
Table S1. A complete list of potential target genes
regulated by Six1 and ⁄ or Six4 in mk4 cells.
Fig. S1. Scatter plot analysis of microarray data. Log-
scale scatter plots of normalized fluorescent intensity
values (normalized data) obtained from Cy3 and Cy5
channels. The Agilent Mouse Developmental Oligo
Microarrays, which contains 20 371 genes, were
hybridized with cDNA probes prepared from mRNAs
derived from mk4 cells infected with recombinant
adenovirus overexpressing VP16–Six1 fusion proteins
(A) and VP16–Six4 fusion proteins (B). Normalized
data from Cy3 channel were plotted against Cy5
channel in log scale. (A) VP16–Six1wt (Cy3) vs. VP16–
Six1W171R (Cy5). (B) VP16–Six4wt (Cy3) vs. VP16–
Six4W263R (Cy5).
FEBS Journal 272 (2005) 3026–3041 ª 2005 FEBS 3041
Z-i. Ando et al. Target genes of Six1 and Six4