J Am Soc Nephrol 12: 726–735, 2001 Klf6 Is a Zinc Finger Protein Expressed in a Cell-Specific Manner during Kidney Development

EVELYNE A. FISCHER,* MARIE-CHRISTINE VERPONT,* LEE ANN GARRETT-SINHA,† PIERRE M. RONCO,* and JEROME A. ROSSERT* *INSERM U 489 and University of Paris VI, AP-HP, Paris, France; and †The University of Texas, M.D. Anderson Cancer Center, Houston, Texas.

Abstract. Molecular mechanisms that are responsible for the expressed in the Wolffian duct but not in the mesonephric development of the renal collecting duct system during embry- mesenchyme. Thereafter, Klf6 was expressed in the ureteric ogenesis are still poorly understood. A mouse cDNA encoding bud and its branches and in the collecting ducts, whereas it was a zinc finger protein, called Klf6, which is a member of the not expressed in tubular structures that derive from the meta- Kru¨ppel-like family of transcription factors, has been cloned. nephric mesenchyme. Glomeruli were not labeled during early Northern blot analyses showed that Klf6 was already expressed stages of differentiation, and it is only at the capillary stage that in 11.5-d postconception mouse embryos and that its expres- a staining of the mesangial area was observed, which persisted sion persisted after birth. They also disclosed that Klf6 had a after birth. This pattern of expression is strikingly similar to the restricted pattern of expression. In situ hybridization experi- one of GATA-3, which is another zinc finger protein. It sug- ments using mouse embryos showed that during kidney devel- gests that Klf6 may play a role during kidney development and opment, Klf6 was expressed selectively in the Wolffian duct in particular during the development of the renal collecting and in its derivatives. During mesonephros development, it was duct system, possibly in association with GATA-3.

In mammals, renal development proceeds in three stages (re- As for many other structures, the combinatorial action of viewed in reference 1). The first two stages lead to the forma- different cell-specific transcription factors is very likely to play tion of transient structures, the pronephros and the mesone- a critical role in the development of the Wolffian duct and of phros, and the third stage gives rise to the metanephros, which the ureteric bud. Different transcription factors such as Emx2, is the permanent kidney. During the development of the me- Pax-2, Lim-1, and GATA-2 are expressed in the ureteric bud sonephros, the Wolffian duct penetrates a nephrogenic mesen- and are involved in kidney development (reviewed in reference chyme and induces it to form nephric tubules. The formation of 2). Emx2 is expressed predominantly in the ureteric bud, and in the metanephros results from reciprocal inductive interactions mice that lack Emx2, the ureteric bud invades the metanephric between a mesenchymal structure, the metanephric blastema, mesenchyme but does not branch and never induces mesen- and an outgrowth of the Wolffian duct, the ureteric bud. The chymal cells to condense (3,4). Pax-2 is expressed both in the metanephric mesenchyme induces the ureteric bud to grow, ureteric bud and in the metanephric mesenchyme; mice that branch, and give rise to the collecting duct system. At the same lack Pax-2 do not have a ureteric bud, whereas hemizygous time, the ureteric bud induces the metanephric mesenchyme to mice have hypoplastic kidneys (5). Lim-1 is expressed in the condense around its tips and then to differentiate into epithelial ureteric bud and its derivatives and in the developing nephrons structures that ultimately will form the epithelial components (6). The very few Lim-1 null mice that survive until birth lack of the nephrons, through a multistep process. The condensed kidneys (7). GATA-2 is expressed in different tissues, includ- mesenchyme will successively differentiate into vesicles, com- ing the ureteric bud, and mice that do not express GATA-2 ma-shaped bodies, S-shaped bodies, and then nephrons. Paral- selectively in the developing kidney display an abnormal junc- lel to this differentiation process, the distal parts of the S- tion between the ureter and the bladder (8). A few other genes shaped bodies fuse with collecting ducts, and the proximal that encode transcription factors such as GATA-3 or Hoxb-7 parts of these structures become highly vascularized and form are also expressed in the developing ureteric bud (9,10), but glomeruli. their role in kidney development is still unknown. Zinc finger proteins have emerged as a major class of eukaryotic transcription factors. They are characterized by Received January 25, 2000. Accepted September 1, 2000. their DNA-binding domain containing cysteine or histidine Correspondence to Dr. Jerome A. Rossert, INSERM U489 and Department of residues that bind zinc atoms, and they can be divided into Nephrology, Tenon Hospital, 4 rue de la Chine, 75020 Paris, France. Phone: different subgroups depending on the amino acid residues that 33-1-56-01-60-29; Fax: 33-1-56-01-69-99; E-mail: [email protected] hop-paris.fr are important for zinc binding and on the spacing between 1046-6673/1204-0726 these amino acid residues. For example, factors that belong to Journal of the American Society of Nephrology the subfamily of Kru¨ppel-like transcription factors have two Copyright © 2001 by the American Society of Nephrology cysteine and two histidine residues that bind the zinc ion, and J Am Soc Nephrol 12: 726–735, 2001 Expression of Klf6 during Kidney Development 727

the consensus sequence of their zinc finger motif is Cys-X2Ϫ4- (Hybond-N, Amersham Pharmacia Biotech, Piscataway, NJ), and hybridized with a 32P-radiolabeled probe containing 192 bp of the Cys-X12-His-X3Ϫ4-His (reviewed in reference 11). This sub- family, which is itself part of the TFIIIA subclass (reviewed in coding sequence and 308 bp of the 3' untranslated region, following reference 12), includes ubiquitously expressed transcription standard procedures. After 18 to 20 h of hybridization, high-strin- factors but also transcription factors that have a restricted gency washes were performed, the last wash being done at 65°C for 10 min in 0.1 ϫ SSC (1 ϫ SSC is 0.15 M NaCl plus 0.015 M citrate pattern of expression and that can play important roles during Na), 0.1% SDS. To control for loading of RNA samples, each mem- organ differentiation, such as EKLF, LKLF, or GKLF/EZF brane was then stripped and reprobed using a 32P-radiolabeled probe (13–18). corresponding to the mouse glyceraldehyde-3-phosphate dehydroge- We report here the characterization of a mouse cDNA en- nase (GAPDH) gene. Labeling of the probes with [␣-32P]dCTP was coding a zinc finger protein, called Klf6, which belongs to the performed using a random priming labeling kit (Roche Diagnostics, Kru¨ppel-like family of transcription factors. Northern blot Basel, Switzerland) and following the manufacturer’s instructions. In analyses and in situ hybridization experiments showed that the some cases, the intensity of the signal was quantified using a Storm corresponding gene was expressed early during embryonic 860 PhosphorImager (Molecular Dynamics, Sunnyvale, CA). development and had a restricted pattern of expression. In particular, during kidney development, the expression of Klf6 In Situ Hybridization was restricted mostly to the collecting duct system, which The in situ hybridization study was performed using whole mouse suggests that this transcription factor may play a role during the embryos that ranged from 11.5 to 15.5 d p.c. and using different development of the renal excretory system. organs obtained from newborn mice and from 4-wk-old mice. Em- bryos and organs were collected under sterile conditions, fixed in Materials and Methods freshly made 4% paraformaldehyde at 4°C for 4 to 10 h, and embed- 35 Cloning of Klf6 ded in paraffin. Five to 7-␮ sections were made. S-radiolabeled Klf6 sense and antisense riboprobes were synthesized by in vitro transcrip- A partial cDNA encoding a protein called Klf6 was obtained using tion of a linearized pBluescript KS plasmid containing the 500-bp a method previously described for EZF (17). Briefly, poly(A)-RNA sequence that was used as a probe in Northern blot experiments. derived from 13.5-d postconception (p.c.) mouse embryos was used to Transcripts were synthesized using the maxiscript T3/T7 kit (Ambion, prepare first strand cDNA. These cDNA were then used as a template Austin, TX) and [35S]UTP, following the manufacturer’s instructions. in a PCR, with an oligo(dT) primer and two degenerate primers. These Before use, the RNA probes were incubated with1UofDNase I for primers (CACATCAGGACCCAC(C/T)ACIGG(A/G)GA and CA- 20 min at 37°C and purified using G50 columns (Amersham Phar- CATCCGIACCCA(T/C)ACIGG(T/C)GA) were homologous to an macia Biotech). Hybridization was performed as described by Sibony amino acid sequence that is conserved among several members of the et al. (19). Briefly, slides were deparaffinized, microwaved for 12 TFIIIA family of zinc finger proteins (HIRTHTGE). The PCR prod- min, refixed in 4% paraformaldehyde, digested with proteinase K (20 ucts were cloned into pBluescript KS (Stratagene, La Jolla, CA) and ␮g/ml), postfixed in 4% paraformaldehyde, dehydrated through in- hybridized with a second degenerate oligonucleotide (ACCGGC- creasing concentrations of ethanol, air dried, and incubated with the GA(A/G)AA(A/G)CCITT(T/C)G(A/C)TG) homologous to an over- probe (5 ϫ 104 to 5 ϫ 105 cpm/section) overnight at 50°C in a lapping region of the zinc finger domain (TGEKPFAC). cDNA from solution containing 2 ϫ SSC, 10% (wt/vol) dextran sulfate, 1 mg/ml 50 random positive clones were sequenced and compared with the denatured salmon sperm DNA (Roche Diagnostics), 70 mM DTT, and GeneBank database. One clone corresponded to a partial cDNA for a 50% formamide. Washes were then performed under high-stringency novel protein that contained at least two Cys -His zinc finger motifs 2 2 conditions, first in 5 ϫ SSC supplemented with 10 mM DTT, at 50°C, separated by seven amino acids and that was 81% homologous to second in 2 ϫ SSC supplemented with 50% formamide and 10 mM Wilms tumor-1 (WT-1). DTT, at 55°C, and then in TNE (10 mM Tris HCl [pH 7.5], 0.5 M A ␭gt10 library from 16.5-d p.c. mouse embryos (Clontech, Cam- NaCl, 5 mM EDTA [pH 8.0]) at 37°C. The third wash performed in bridge, UK) was screened using the cDNA obtained by PCR as a TNE was done at 37°C for 20 min in a solution that was supplemented probe. Among positive clones, two were cloned in pBluescript KS and with 20 ␮g/ml RNase A to minimize the background. After the entirely sequenced. A 290-bp fragment derived from the 5' end of one washing steps, the slides were dehydrated with graded ethanol solu- of these clones was then used to screen again the same library. Three tions, coated with NTB2 film emulsion (Kodak, New Haven, CT), and positive clones were isolated and sequenced. stored under desiccant at 4°C for 3 to 5 wk. After being photograph- ically developed and fixed, sections were counterstained with hema- Cell-Free Transcription and Translation Experiments toxylin (0.5%) and eosin (1%), observed, and photographed on an In vitro transcription and translation of cDNA cloned into pBlue- Axioplan 2 microscope (Zeiss, Go¨ttingen, Germany). script KS was performed using a reticulocyte lysate system (Promega, Madison, WI) according to the manufacturer’s instructions, in the presence of [35S]methionine. The radiolabeled peptides were resolved Results by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS- Cloning of Klf6 PAGE) and visualized by fluorography. To identify new zinc finger proteins of the TFIIIA subclass expressed during embryonic development, we performed PCR Northern Blot Analysis using an oligo(dT) primer, two degenerate oligonucleotides Total RNA was isolated from various mouse tissues using an homologous to a conserved amino acid sequence, and cDNA RNeasy kit (Qiagen, Valencia, CA) and following the manufacturer’s from 13.5-d p.c. mouse embryos. Fifty partial cDNA clones instructions. Fifteen ␮g of total RNA were then electrophoresed were isolated and sequenced. One of them encoded a novel through formaldehyde-agarose gels, transferred to nylon membranes zinc finger protein, which was 81% homologous to WT-1. 728 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 726–735, 2001

Because WT-1 plays a crucial role during kidney development stream ATG were not in a favorable context to serve as and because mutations in the WT-1 gene have been linked with initiation codons (20). By contrast, the second ATG codon, renal diseases, we decided to study this cDNA further. The which was located 99 bp downstream of the first one and 12 bp 534-bp PCR product was used to screen a ␭gt10 mouse library upstream of the third one, was in a favorable context for for clones containing the entire open reading frame. Sequenc- translation initiation according to Kozak (20). It is therefore ing of positive clones enabled us to identify a 1.5-kb cDNA likely that the major translation product initiates at this second with a single open reading frame, which encoded a protein ATG codon. This codon is followed by an open reading frame containing three Cys2-His2 zinc fingers (Figure 1). This open of 846 nucleotides capable of encoding a 282-amino acid reading frame contained a stop codon 4 bp downstream of the peptide containing three zinc fingers of the Kru¨ppel-like type third zinc finger, but it extended up to the 5' end of the cDNA. (Cys-X2Ϫ4-Cys-X3-Phe-X5-Leu-X2-His-X-Arg/Lys-X-His); To identify clones extending further upstream, we rescreened only one mismatch (Phe3Tyr) was present in the first zinc the same library using a 290-bp probe located at the 5' end of finger (Figure 1) (11). These zinc fingers are separated by the cDNA. Three additional clones were isolated and se- spacers of seven amino acid residues (Figure 1). The first quenced, but they did not extend further upstream. A comput- spacer corresponds to the consensus sequence Thr/Ser-Gly- er-based search of the dbest database also identified three Glu-Arg/Lys-Pro-Phe/Tyr-X, which has been described for mouse clones, but they extended at most 40 bp upstream of the Kru¨ppel-like proteins (11), and the second one adheres to this previous cDNA. consensus sequence with only one mismatch (Glu3Ala). A Analysis of the cDNA showed that it contained only one comparison of the zinc finger region with other available DNA open reading frame, with three ATG codons located at its 5' sequences confirmed the existence of strong homologies with end (Figure 1). The most upstream ATG and the most down- other zinc finger proteins of the Kru¨ppel-like class (Figure 2A). Besides the three zinc fingers, the protein contains an N- terminal domain (amino acids 29 to 118) that is rich in acidic residues (23.3% of glutamic acid and aspartic acid) and thus could be a transcription activation domain (Figure 2B). The central part of the protein also contains both a serine- and a proline-rich domain (42.2% of serine and proline residues between amino acids 82 and 152) and a glutamine-rich domain (30.3% of glutamine residues between amino acids 164 and 196) that could also act as transcription activation domains (Figure 2B). Just upstream of the first zinc finger motif, a sequence corresponds to a consensus core nuclear localization signal (Arg-Arg-Arg-Val-His-Arg). This protein also contains two potential protein kinase C phosphorylation sites (Thr-Thr- Lys at position 122 to 124 and Ser-Gly-Lys at position 180 to 182) (21), suggesting that it may undergo posttranslational modifications. Since we cloned this mouse cDNA, partial human and rat cDNA that code for proteins that are, respectively, 95% and 98% identical to our protein have been isolated (Genbank accession numbers U44975, AB017493, AF001417) (22–24). The rat protein has been called Zf9 (24) and the human one CPBP/GBF (22,23), but the Human Gene Nomenclature Com- mittee has recently suggested renaming the latter protein KLF6. Accordingly, we decided to call our protein Klf6, be- cause it is the mouse ortholog of KLF6.

In Vitro Transcription–Translation To confirm that the predicted protein can be synthesized in an eukaryotic system, we transcribed in vitro a cDNA contain- ing the entire open reading frame corresponding to Klf6 and cloned it into pBluescript KS and then translated it using a Figure 1. Nucleotide sequence and deduced amino acid sequence of reticulocyte lysate. By doing so, we isolated a protein with the the mouse Klf6 (GenBank accession number AY027436). The three predicted size of approximately 32 kD (Figure 2C), which is in zinc finger motifs are underlined, and the three in frame ATG codons good agreement with the predicted molecular mass for Klf6 are double underlined. (31.9 kD). J Am Soc Nephrol 12: 726–735, 2001 Expression of Klf6 during Kidney Development 729

Figure 2. Analysis of the mouse Klf6 protein. (A) Comparison of the amino acid sequence between Klf6 and other Kru¨ppel-like proteins. The sequences are those of the zinc finger regions. The cysteine and histidine residues binding a zinc atom are boxed. Dashes represent identities to the mouse Klf6 protein. M, mouse; h, human. (B) Schematic representation of the different domains of the Klf6 protein. NLS, consensus core nuclear localization sequence; P, potential protein kinase C phosphorylation site. (C) The Klf6 cDNA is translated into a polypeptide of approximately 32 kD. A pBluescript KS plasmid containing the Klf6 cDNA was linearized, and the sense or the antisense mRNA were transcribed in vitro using the T7 or T3 RNA polymerase, respectively. The cognate mRNA were translated in a cell-free reticulocyte lysate, in the presence of 35S-labeled methionine. The translation products were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and visualized by fluorography. The arrow indicates the migration of the labeled Klf6 polypeptide. Numbers on the right refer to the position of the molecular mass markers in kD.

Expression Pattern of Klf6 in Embryos and Adult Mice cisely by performing in situ hybridization experiments with Northern blot analyses using total RNA isolated from mouse 11.5-, 12.5-, 13.5-, and 15.5-d p.c. mouse embryos and with embryos ranging from 11.5 to 15.5 d p.c. and from newborn tissues from newborn mice (Figures 4 through 6). The cDNA mice showed that Klf6 was expressed as soon as 11.5 d p.c. and used to synthesize the 35S-radiolabeled sense and antisense that its expression persisted all through embryonic develop- riboprobes was identical to the one used to synthesize the probe ment (Figure 3A). At all stages of development, a single band for Northern blot experiments. In all of the cases, slides that of approximately 4.5 kb could be detected. were hybridized with the sense riboprobe did not show any As a first approach to determine which organs express Klf6, staining (data not shown). A labeling of the lung buds was we performed Northern blot analyses of different tissues in observed as soon as day 12.5 p.c. (Figure 4B), and from this newborn and adult mice. In both cases, Klf6 was expressed at stage until birth, Klf6 was expressed in the lung buds, in the high levels in lung and intestine and at lower levels in brain, bronchi, and in the layer of mesothelial cells that covers the heart, and kidney (Figure 3B, and data not shown). By contrast, lung buds and will form the pleura (Figure 4, B through D, and no expression could be detected in liver or spleen (Figure 3B, data not shown). The expression of Klf6 in the digestive tract and data not shown). was very weak at 11.5 d p.c., but it became stronger 1 d later Northern blot experiments were also performed using a and persisted until birth (Figure 4, and data not shown). It was PhosphorImager to analyze the levels of expression of Klf6 restricted to epithelial cells lining the lumen of the developing during metanephros development and afterward. They showed intestine (data not shown). In the nervous system, Klf6 was that Klf6 was expressed at similar levels in kidneys obtained expressed in some discrete areas of the brain, but it was also from 15.5-d p.c. mouse embryos, newborn mice, 10-d-old strongly expressed in ganglia such as root ganglia or the mice, and 4-mo-old mice (Figure 3C). trigeminal ganglion (Figure 4). Besides these organs, Klf6 was The pattern of expression of Klf6 was assessed more pre- expressed in the peritoneum and in the pericardium (Figure 4). 730 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 726–735, 2001

Expression Pattern of Klf6 during Kidney Development Because of sequence homologies between Klf6 and WT-1, the pattern of expression of Klf6 during kidney development was analyzed in detail. At 11.5 and 12.5 d p.c., the Wolffian duct penetrated the mesonephric mesenchyme, and this mes- enchyme was differentiated into epithelial structures that cor- respond to primitive nephrons. At these stages, Klf6 was ex- pressed in the Wolffian duct (Figure 5, A through F), but it was not expressed in the mesonephric mesenchyme (Figure 5, A through F). In particular, mesonephric tubules were negative (Figure 5F). At 12.5 d p.c., the ureteric bud sprouted from the distal portion of the Wolffian duct and penetrated into the metaneph- ric blastema, but only one or very few branchings occurred. At this stage, Klf6 was expressed in the ureteric bud not in the metanephric mesenchyme (data not shown). At 13.5 d p.c., the ureteric bud branched within the meta- nephric mesenchyme, and this mesenchyme gave rise to com- ma-shaped and S-shaped bodies. At this stage, Klf6 was ex- pressed in the ureter and in the ureteric bud (Figure 5, G and H, and data not shown). By contrast, it was not expressed in the metanephric mesenchyme (Figure 5, G and H). At 15.5 d p.c., nephrons at all developmental stages can be observed, the most mature ones being located in the juxtamed- ullary area. Klf6 was expressed in the ureteric bud, in the developing collecting ducts (Figure 5, I and J), but also in fully differentiated glomeruli (data not shown). By contrast, it was not expressed in tubular structures derived from the metaneph- ric mesenchyme or in noncapillarized glomeruli (Figure 5, I and J). Analysis of glomeruli showed that the labeling was restricted to cells located within the mesangial area, which correspond to mesangial cells and possibly to endothelial cells (data not shown). Klf6 was also expressed in the urogenital sinus, which will form the bladder, and in the urethra (Figure 4D). At birth, the pattern of expression was similar to the one Figure 3. Northern blot analyses of the expression of Klf6. Total observed at 15.5 d p.c. Klf6 was detected in mature glomeruli mRNA from whole mouse embryos and from newborn mice (A), or and in the collecting ducts but not in other tubular structures or from different organs obtained from 4-wk-old mice (B), or from in the interstitium (Figure 6, A through C, and data not shown). kidneys obtained from 15.5-d postconception (p.c.) mouse embryos In kidneys that were isolated from 4-wk-old mice, Klf6 was and from mice at different ages (C) were blotted to nylon membranes still strongly expressed in glomeruli and in the cortical portion and hybridized first with a probe corresponding to Klf6 and second of the collecting ducts (Figure 6D, and data not shown). with a probe corresponding to the mouse GAPDH gene. The mem- branes were then subjected to autoradiography (A, B) or were ana- Discussion lyzed using a PhosphorImager (C). dpc, days postconception; Nb, We isolated a mouse cDNA encoding a zinc finger protein newborn. (A) Klf6 was expressed in 11.5-d p.c. embryos, and its called Klf6, which belongs to the Kru¨ppel-like family of tran- expression persisted until birth. At all stages, a single band was detected. (B) In 4-wk-old mice, Klf6 was expressed at high levels in scription factors. In addition to three characteristic zinc fingers, lung and intestine and at lower levels in brain, heart, and kidney; no Klf6 contains an N-terminal domain that is rich in acidic expression could be detected in liver. (C) The levels of expression of residues and could be a transcription activation domain. Inter- Klf6 were similar in kidneys obtained from 15.5-d p.c. mouse em- estingly, the first 50 amino acids of Klf6 are 80% identical to bryos, newborn mice, 10-d-old mice, and 4-mo-old mice. those of UKLF, another Kru¨ppel-like protein, and the N- terminal domain of UKLF (amino acids 1 to 72) has been shown to be responsible for most of the transactivating prop- By contrast, no labeling could be detected in spleen, myocar- erties of this protein (25). The ability of Klf6 to enter the dium, muscles, or skeleton (Figure 4). The liver was almost nucleus and act as a transcription factor, i.e., to bind DNA and entirely negative, with the exception of a few cells scattered in modulate transcription, has been demonstrated using CPBP/ the organ (data not shown). GBF or Zf9, which are the human and the rat orthologs of Klf6, J Am Soc Nephrol 12: 726–735, 2001 Expression of Klf6 during Kidney Development 731

Figure 4. Embryonic expression of Klf6 mRNA, as detected by in situ hybridization. Dark-field views. Sections of 11.5-, 12.5-, 13.5-, and 15.5-d p.c. mouse embryos were hybridized with an 35S-radiolabeled antisense riboprobe. (A) At 11.5 d p.c., a signal was detected in some areas of the central nervous system, in root ganglia (small arrows), and in the Wolffian duct (large arrows). (B) At 12.5 d p.c., a labeling was observed in the lungs (L), in the peritoneum (P), in the pericardium (H), and in the Wolffian duct (arrows). (C) At 13.5 d p.c., a staining of the ureter (arrow) and of the intestine (I) was clearly visible. (D) At 15.5 d p.c., a staining of the kidney (K) and of the urogenital sinus (US) could be seen, as well as a labeling of the intestine (I), of the peritoneum (P), and of the trigeminal ganglion (T). Magnifications: ϫ30 in A and B; ϫ15 in C and D. 732 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 726–735, 2001

Figure 5. Expression of Klf6 mRNA in the kidney during embryonic development, as detected by in situ hybridization. On the left are dark-field views of tissue sections, and on the right are bright-field views of the same slides but at a higher magnification. Sections of 11.5-, 12.5-, 13.5-, and 15.5-d p.c. mouse embryos were hybridized with an 35S-radiolabeled antisense riboprobe and counterstained with hematoxylin-eosin. (A). (B) Section of an 11.5-d p.c. mouse embryo showing a labeling of the Wolffian duct (arrows). (C through F) Sections of a 12.5-d p.c. mouse embryo showing a staining of the Wolffian duct (arrows), whereas the mesonephric tubules are negative (arrowheads in F). (G and H) Section of a 13.5-d p.c. mouse embryo showing a staining of the developing ureter (arrow) but not of the metanephric mesenchyme (M). (I and J) Section of the metanephros of a 15.5-d p.c. mouse embryo disclosing a staining of the branches of the ureteric bud (arrowheads). By contrast, vesicles and comma-shaped bodies were negative. Magnifications: ϫ30 in A, C, and E; ϫ60 in B; ϫ85 in D; ϫ150 in F; ϫ45 in G; ϫ80 in H; ϫ125 in I; ϫ180 in J. J Am Soc Nephrol 12: 726–735, 2001 Expression of Klf6 during Kidney Development 733

Figure 6. Expression of Klf6 mRNA in the kidney at or after birth, as detected by in situ hybridization. Dark-field (A) and bright-field (B through D) views. Sections of kidneys obtained from newborn mice and from 4-wk-old mice were hybridized with an 35S-radiolabeled antisense riboprobe and counterstained with hematoxylin-eosin. (A through C) Kidney section from a newborn mouse showing a staining of the collecting ducts. (A) The lower box corresponds to Figure 6B and the upper one corresponds to Figure 6C. (B) The arrows indicate labeled medullary collecting ducts, and the inset corresponds to a higher magnification of a collecting duct. (C) Developing cortical collecting ducts are labeled, whereas a developing nephron (N) is negative. (D) Glomerulus from a 4-wk-old mouse disclosing a staining of glomerular cells that appear to be located in the mesangial area. Magnifications: ϫ62.5 in A; ϫ125 in B; ϫ375 in inset; ϫ250 in C;ϫ500 in D. respectively (22–24). First, experiments performed using a tein can bind G/C-rich double-stranded oligonucleotides (22– green fluorescence protein-GBF fusion protein or anti-Zf9 24); third, in transfection experiments, both the human and the antibodies have shown that these proteins can enter the nucleus rat proteins can enhance the transcriptional activity of promot- (23,24); second, in gel shift experiments, a glutathione-S- ers that contain G/C-rich sequences, such as the promoter of transferase–GBF fusion protein, a ␤-galactosidase–CPBP fu- the pregnancy-specific glycoprotein 5 or the proximal pro- sion protein, and a glutathione-S-transferase–Zf9 fusion pro- moter of the pro-␣1(I) collagen gene (22,24). 734 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 726–735, 2001

Few data are available regarding the expression pattern of GATA-3 is that its tissue-specific expression is controlled by a Klf6, and in particular its expression has not been studied modular arrangement of different cis-acting elements, some of during embryonic development. In human, Northern blot anal- them being located far from the coding sequence (28). In yses have shown that CPBP/GBF is expressed at various levels particular, the cis-acting element that is responsible for the in different malignant cell lines and that it is expressed at high expression of GATA-3 in the developing kidney is located levels in placenta, lung, and arguably pancreas (22,23). In rat, approximately 100 kb upstream of the transcription start site RNase protection assays using adult tissues have shown that (28). As shown by an analysis of sequence tag sites (sequence Zf9 is expressed at high levels in lung and intestine, as well as tag site WI-12084), the human KLF6 gene is located on chro- in different cell lines, including stellate cells (24). Our exper- mosome 10, less than 10 cm away from GATA-3. Thus, it is iments showed that Klf6 was already expressed in 11.5-d p.c. tempting to hypothesize that the same cis-acting element con- mouse embryos and that Klf6 had a restricted pattern of ex- trols the renal expression of both genes, as suggested for other pression. It was expressed at high levels in organs such as lung, genes (29,30). intestine, and brain, whereas it was not expressed in other In conclusion, we have isolated a mouse cDNA that encodes organs such as spleen or skeleton and was very faintly ex- a protein that belongs to the Kru¨ppel-like class of zinc finger pressed in liver. transcription factors and that we named Klf6. During kidney In situ hybridization experiments during kidney develop- development, Klf6 mRNA was expressed selectively in the ment showed that Klf6 was expressed during the development Wolffian duct, the ureteric bud, the collecting ducts, and the of the mesonephros and of the metanephros and that its ex- mesangium. This pattern of expression suggests that Klf6 may pression was mostly restricted to the excretory system: it was play a role in the development of the kidney, in particular of expressed in the Wolffian duct, in the ureteric bud, and in the the renal collecting duct system. Furthermore, strong similar- collecting ducts, whereas it was not expressed in the meso- ities between the patterns of expression of Klf6 and GATA-3 nephric mesenchyme or in the metanephric mesenchyme or in during kidney development suggest that these two proteins their derivatives. In addition to the excretory system, Klf6 was could interact to regulate renal development. expressed only in the mesangial area of glomeruli, this labeling appearing at the capillary loop stage and persisting thereafter. References The restricted pattern of expression of Klf6 suggests that it may 1. Ekblom P, Miettinen A, Virtanen I, Wahlstro¨m T, Dawnay A, play a role during kidney development, in particular in the Saxen L: In vitro segregation of the metanephric nephron. Dev differentiation of the ureteric bud system. A search for pro- Biol 84: 88–95, 1981 moters that contain a GGNGNGGGN consensus sequence and 2. Baker LA, Gomez RA: Embryonic development of the ureter. Semin Nephrol 18: 569–584, 1998 thus likely to bind Kru¨ppel-like factors (11) showed that this 3. Miyamoto N, Yoshida M, Kuratani S, Matsuo I, Aizawa S: sequence is present in the promoters of genes that are selec- Defects of urogenital development in mice lacking Emx2. De- tively expressed in collecting duct cells, such as the arginine velopment 124: 1653–1664, 1997 vasopressin type 2 receptor gene, the aquaporin 2 gene, and the 4. Yoshida M, Suda Y, Matsuo I, Miyamoto N, Takeda N, Kuratani aquaporin 3 gene, reinforcing the idea that Klf6 may play a role S, Aizawa S: Emx1 and Emx2 functions in development of in the differentiation of collecting duct cells. dorsal telencephalon. Development 124: 101–111, 1997 Most of the genes that are expressed in the ureteric bud 5. Rothenpieler UW, Dressler GR: Pax-2 is required for mesen- during kidney development, such as Emx2, Lim-1,orPax-2, chyme-to-epithelium conversion during kidney development. have a pattern of expression that is quite different from the one Development 119: 711–720, 1993 of Klf6; in particular, they are also expressed in tubular struc- 6. Karavanov AA, Karavanova I, Perantoni A, Dawid IB: Expres- tures that are derived from the metanephric mesenchyme sion pattern of the rat Lim-1 homeobox gene suggests a dual role during kidney development. Int J Dev Biol 42: 61–66, 1998 (3,5,6). By contrast, the pattern of expression of Klf6 in the 7. Shawlot W, Behringer RR: Requirement for Lim1 in head- kidney is strikingly similar to the one of GATA-3, which organizer function. Nature 374: 425–430, 1995 encodes a zinc finger protein of the Cys4 class (9). During 8. Zhou Y, Lim KC, Onodera K, Takahashi S, Ohta J, Minegishi N, mesonephros development, GATA-3 is expressed in the Wolf- Tsai FY, Orkin SH, Yamamoto M, Engel JD: Rescue of the fian duct and in its derivatives but not in the mesonephric embryonic lethal hematopoietic defect reveals a critical role for blastema (9). At the metanephric stage, GATA-3 is expressed in GATA-2 in urogenital development. EMBO J 17: 6689–6700, the excretory system and in mesangial cells but not in other 1998 structures derived from the metanephric mesenchyme. The 9. Labastie MC, Catala M, Gregoire JM, Peault B: The GATA-3 similarities between the patterns of expression of GATA-3 and gene is expressed during human kidney embryogenesis. Kidney of Klf6 in the developing kidney are reminiscent of previous Int 47: 1597–1603, 1995 observations made for two other genes of the same families: 10. Vogels R, de Graaff W, Deschamps J: Expression of the murine homeobox-containing gene Hox-2.3 suggests multiple time-de- GATA-1 and EKLF. Both genes are expressed in erythroid pendent and tissue-specific roles during development. Develop- cells, and both proteins cooperate to induce a normal differen- ment 110: 1159–1168, 1990 tiation of red blood cells, probably through direct protein– 11. Turner J, Crossley M: Mammalian Kru¨ppel-like transcription protein interactions (13,26,27). It is possible that Klf6 and factors: More than just a pretty finger. TIBS 24: 236–241, 1999 GATA-3 also directly interact to induce a normal differentia- 12. Shastry BS: Transcription factor IIIA (TFIIIA) in the second tion of the collecting duct system. One of the specificities of decade. J Cell Sci 109: 535–539, 1996 J Am Soc Nephrol 12: 726–735, 2001 Expression of Klf6 during Kidney Development 735

13. Southwood CM, Downs KM, Bieker JJ: Erythroid Kruppel-like 23. Suzuki T, Yamamoto T, Kurabayashi M, Nagai R, Yazaki Y, factor exhibits an early and sequentially localized pattern of Horikoshi M: Isolation and initial characterization of GBF, a expression during mammalian erythroid ontogeny. Dev Dyn 206: novel DNA-binding zinc finger protein that binds to the GC-rich 248–259, 1996 binding sites of the HIV-1 promoter. J Biochem 124: 389–395, 14. Perkins AC, Sharpe AH, Orkin SH: Lethal beta-thalassaemia in 1998 mice lacking the erythroid CACCC-transcription factor EKLF. 24. Ratziu V, Lalazar A, Wong L, Dang Q, Collins C, Shaulian E, Nature 3753: 18–22, 1995 Jensen S, Friedman SL: Zf9, a Kruppel-like transcription factor 15. Faustino P, Lavinha J, Marini MG, Moi P: beta-Thalassemia up-regulated in vivo during early hepatic fibrosis. Proc Natl mutation at -90C–ϾT impairs the interaction of the proximal Acad Sci USA 95: 9500–9505, 1998 CACCC box with both erythroid and nonerythroid factors. Blood 25. Matsumoto N, Laub F, Aldabe R, Zhang W, Ramirez F, Yoshida 88: 3248–3249, 1996 T, Terada M: Cloning the cDNA for a new human zinc finger 16. Kuo CT, Veselits ML, Barton KP, Lu MM, Clendenin C, Leiden protein defines a group of closely related Kruppel-like transcrip- JM: The LKLF transcription factor is required for normal tunica tion factors. J Biol Chem 273: 28229–28237, 1998 media formation and blood vessel stabilization during murine 26. Weiss MJ, Orkin SH: GATA transcription factors: Key regula- embryogenesis. Genes Dev 11: 2996–3006, 1997 tors of hematopoiesis. Exp Hematol 23: 99–107, 1995 17. Garrett-Sinha LA, Eberspaecher H, Seldin MF, de Crombrugghe 27. Merika M, Orkin SH: Functional synergy and physical interac- B: A gene for a novel zinc finger protein expressed in differen- tions of the erythroid transcription factor GATA-1 with the tiated epithelial cells and transiently in certain mesenchymal Kruppel family proteins Sp1 and EKLF. Mol Cell Biol 15: cells. J Biol Chem 271: 31384–31390, 1996 2437–2447, 1995 18. Segre JA, Bauer C, Fuchs E: Klf4 is a transcription factor 28. Lakshmanan G, Lieuw KH, Lim KC, Gu Y, Grosveld F, Engel required for establishing the barrier function of the skin. Nat JD, Karis A: Localization of distant urogenital system-, central Genet 22: 356–360, 1999 19. Sibony M, Commo F, Callard P, Gasc JM: Enhancement of nervous system-, and endocardium-specific transcriptional regu- mRNA in situ hybridization signal by microwave heating. Lab latory elements in the GATA-3 locus. Mol Cell Biol 19: 1558– Invest 73: 586–591, 1995 1568, 1999 20. Kozak M: An analysis of 5'-noncoding sequences from 699 29. Bradshaw MS, Shashikant CS, Belting HG, Bollekens JA, vertebrate messenger RNAs. Nucleic Acids Res 15: 8125–8148, Ruddle FH: A long-range regulatory element of Hoxc8 identified 1987 by using the pClasper vector. Proc Natl Acad Sci USA 93: 21. Woodgett JR, Gould KL, Hunter T: Substrate specificity of 2426–2430, 1996 protein kinase C. Use of synthetic peptides corresponding to 30. Le Beyec J, Chauffeton V, Kan HY, Janvier PL, Cywiner- physiological sites as probes for substrate recognition require- Golenzer C, Chatelet FP, Kalopissis AD, Zannis V, Chambaz J, ments. Eur J Biochem 161: 177–184, 1986 Pincon-Raymond M, Cardot P: The Ϫ700/Ϫ310 fragment of the 22. Koritschoner NP, Bocco JL, Panzetta-Dutari GM, Dumur CI, apolipoprotein A-IV gene combined with the Ϫ890/Ϫ500 apo- Flury A, Patrito LC: A novel human zinc finger protein that lipoprotein C-III enhancer is sufficient to direct a pattern of gene interacts with the core promoter element of a TATA box-less expression similar to that for the endogenous apolipoprotein gene. J Biol Chem 272: 9573–9580, 1997 A-IV gene. J Biol Chem 274: 4954–4961, 1999