Portal de Periódicos da CAPES

Krüppel-like Factors: Three Fingers in Many Pies

2001; Elsevier BV; Volume: 276; Issue: 37 Linguagem: Inglês

10.1074/jbc.r100043200

1083-351X

James J. Bieker,

Genetic Syndromes and Imprinting

Krüppel-like factor erythroid lung basic gut intestinal ubiquitous kidney fetal basic transcription element binding zinc finger core promoter-binding protein TGF-β-inducible early gene transforming growth factor cAMP-response element-binding protein (CREB)-binding protein AP2 repressor embryonic stem interferon familial adenomatous polyposis Krüppel-like factors (KLFs)1 are DNA-binding transcriptional regulators that play diverse roles during differentiation and development. They form a subset of the broad class of proteins containing Cys2/His2 zinc fingers, a motif that is the second most abundant seen in the human genome and the most abundant in transcription factors, estimated to be present in just under 600 to over 700 proteins (1Tupler R. Perini G. Green M.R. Nature. 2001; 409: 832-833Crossref PubMed Scopus (311) Google Scholar, 2Lander E.S. et al.Nature. 2001; 409: 860-921Crossref PubMed Scopus (18333) Google Scholar, 3Venter J.C. et al.Science. 2001; 291: 1304-1351Crossref PubMed Scopus (10873) Google Scholar). The nomenclature is based on the homology of its founding member, EKLF (erythroid Krüppel-like factor; KLF1), to theDrosophila Krüppel protein (4Miller I.J. Bieker J.J. Mol. Cell. Biol. 1993; 13: 2776-2786Crossref PubMed Scopus (667) Google Scholar). There are numerous finger proteins with such homology; however, the fact that KLF proteins contain additional conserved residues between each finger, that their fingers are three in number, and that these are found at the extreme carboxyl end serves to further define KLF proteins as a separate family (Fig. 1). Phylogenetic analysis of the 15 human KLF members demonstrates that they form a clade distinct even from the closely related Sp1 and Krox zinc finger families. Structural considerations also place these members together, as three amino acids at specific locations (“XYZ” positions (5Klevit R.E. Science. 1991; 253: 1367-1395Crossref PubMed Scopus (108) Google Scholar)) adjacent to the coordinating histidines play a determining role in target site selection and high affinity binding. As a result, all members of the family bind very similar “GT-box” or “CACCC element” sites on DNA, although the configuration is such that the site tends to be C-rich on one strand and G-rich on the other. The homologies and relationships among KLF family members, as well as their biological properties, have been the subject of excellent comprehensive reviews in the past 2 years (6Turner J. Crossley M. Trends Biochem. Sci. 1999; 24: 236-240Abstract Full Text Full Text PDF PubMed Scopus (270) Google Scholar, 7Philipsen S. Suske G. Nucleic Acids Res. 1999; 27: 2991-3000Crossref PubMed Scopus (541) Google Scholar, 8Dang D.T. Pevsner J. Yang V.W. Int. J. Biochem. Cell Biol. 2000; 32: 1103-1121Crossref PubMed Scopus (380) Google Scholar). This discussion will focus on an update of the family and on recent experiments that demonstrate the wide and diverse range of cellular and molecular effects that are exerted by selected members (Fig.2). KLF6/ZF9/CPBP is rapidly induced in liver stellate cells that are activated after injury, leading to fibrogenesis and extracellular matrix formation (9Ratziu V. Lalazar A. Wong L. Dang Q. Collins C. Shaulian E. Jensen S. Friedman S.L. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 9500-9505Crossref PubMed Scopus (238) Google Scholar). Consistent with this, KLF6 activates the collagen α1(I) promoter and the TGF-β promoters, establishing a link between KLF6 activity and cytokine responsiveness to injury (10Kim Y. Ratziu V. Choi S.G. Lalazar A. Theiss G. Dang Q. Kim S.J. Friedman S.L. J. Biol. Chem. 1998; 273: 33750-33758Abstract Full Text Full Text PDF PubMed Scopus (231) Google Scholar). This response is also seen in injured arterial endothelial cells, in which induced KLF6 expression up-regulates the urokinase plasminogen activator, leading to proteolytic activation of latent TGF-β and subsequent tissue remodeling (11Kojima S. Hayashi S. Shimokado K. Suzuki Y. Shimada J. Crippa M.P. Friedman S.L. Blood. 2000; 95: 1309-1316PubMed Google Scholar). Vascular injury also plays a role in the induction of KLF5/BTEB2/IKLF expression in smooth muscle cells, and phorbol ester-sensitive protein kinase pathways (particularly mitogen-activated protein kinase induction of Egr1 binding to the KLF5 promoter) have been implicated in directing this response (12Kawai-Kowase K. Kurabayashi M. Hoshino Y. Ohyama Y. Nagai R. Circ. Res. 1999; 85: 787-795Crossref PubMed Scopus (84) Google Scholar). In a similar (but not identical) way, KLF5 is a downstream target of Wnt1 signaling as judged by its induction after Wnt1 infection of cultured epithelia and after comparison of KLF5 levels in transgenic Wnt1 mouse mammary cells to wild type controls (13Ziemer L.T. Pennica D. Levine A.J. Mol. Cell. Biol. 2001; 21: 562-574Crossref PubMed Scopus (58) Google Scholar). Induction is transcriptional and occurs via a β-catenin/T cell factor-independent mechanism that may involve protein kinase C activation. Cytokines have also been implicated in KLF induction. First, investigation of TGF-β-driven effects on pancreatic epithelia, prostate, and brain cell growth led to the identification of KLF10/TIEG1/EGRα and KLF11/TIEG2/FKLF (14Cook T. Urrutia R. Am. J. Physiol. 2000; 278: G513-G521Crossref PubMed Google Scholar). These two proteins are not only most homologous with each other (Fig.1, subgroup 2), but they are also immediate-early TGF-β-responsive genes that behave as potent repressors via three uniquely conserved repression motifs (15Cook T. Gebelein B. Belal M. Mesa K. Urrutia R. J. Biol. Chem. 1999; 274: 29500-29504Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar). Overexpression of KLF10 or -11 in a pancreatic cell line or in transgenic mice reveals that the functional effect of this repression is inhibition of cell growth (14Cook T. Urrutia R. Am. J. Physiol. 2000; 278: G513-G521Crossref PubMed Google Scholar) and induction of apoptosis via formation of reactive oxygen species (16Ribeiro A. Bronk S.F. Roberts P.J. Urrutia R. Gores G.J. Hepatology. 1999; 30: 1490-1497Crossref PubMed Scopus (143) Google Scholar). Second, KLF4/GKLF is directly induced by IFNγ in a human colon carcinoma cell line, as mRNA induction is rapid and occurs in the absence of protein synthesis (17Chen Z.Y. Shie J. Tseng C. FEBS Lett. 2000; 477: 67-72Crossref PubMed Scopus (58) Google Scholar). As more fully described below, this link may explain the antiproliferative effects of IFNγ. KLF4/GKLF has been most thoroughly investigated with respect to its role in cellular differentiation, initially within the gastrointestinal tract (18Shields J.M. Christy R.J. Yang V.W. J. Biol. Chem. 1996; 271: 20009-20017Abstract Full Text Full Text PDF PubMed Scopus (574) Google Scholar), where its expression is indicative of a growth-arrested state, and in the epidermis, where KLF4 is critical for late stage differentiation of skin cells (19Segre J.A. Bauer C. Fuchs E. Nat. Genet. 1999; 22: 356-360Crossref PubMed Scopus (659) Google Scholar). Three ways in which KLF4 accomplishes this have been proposed. First, it was noted that KLF4 and p21 levels increase upon induction of growth arrest by serum deprivation and that the kinetics of KLF4 expression slightly preceded that of p21 (20Zhang W. Geiman D.E. Shields J.M. Dang D.T. Mahatan C.S. Kaestner K.H. Biggs J.R. Kraft A.S. Yang V.W. J. Biol. Chem. 2000; 275: 18391-18398Abstract Full Text Full Text PDF PubMed Scopus (291) Google Scholar). Following the observation that the p21 promoter contains CACCC elements, KLF4 was shown to bind and directly transactivate the p21 promoter via these sites. This induction was dependent on p53 and thus also occurred after DNA damage with methyl methanesulfonate. In addition, KLF4 physically interacts with p53. The resultant synergistic induction by p53 and KLF4 of p21 then leads to its inhibition of cyclin-dependent kinases and subsequent growth arrest. The second set of data demonstrates that the minimal cyclin D1 promoter also contains multiple CACCC elements that bind KLF4in vitro and that KLF4 binding results in in vivorepression of the promoter, an effect not seen after transfection of Sp1 (21Shie J.L. Chen Z.Y. Fu M. Pestell R.G. Tseng C.C. Nucleic Acids Res. 2000; 28: 2969-2976Crossref PubMed Google Scholar). Finally, KLF4 activates late differentiation genes such as keratin 4 (22Okano J. Opitz O.G. Nakagawa H. Jenkins T.D. Friedman S.L. Rustgi A.K. FEBS Lett. 2000; 473: 95-100Crossref PubMed Scopus (59) Google Scholar). Together these data argue that KLF4 levels play a critical role in the decision between proliferation and cell cycle arrest/differentiation. A similar role has been postulated for KLF7/ULKF (23Matsumoto N. Laub F. Aldabe R. Zhang W. Ramirez F. Yoshida T. Terada M. J. Biol. Chem. 1998; 273: 28229-28237Abstract Full Text Full Text PDF PubMed Scopus (109) Google Scholar), particularly within developing and adult nervous systems (24Laub F. Aldabe R. Friedrich Jr., V. Ohnishi S. Yoshida T. Ramirez F. Dev. Biol. 2001; 233: 305-318Crossref PubMed Scopus (84) Google Scholar). Expression of KLF7 at specific phases of early development correlated with the time when neuronal precursors exit the cell cycle and differentiate. Two sets of data implicate KLF7 in this process. First, KLF7 can modulate cell cycle regulators, as its induced overexpression results in a decrease in DNA synthesis, induction of p21 protein, inhibition of cyclin D1, and G1 arrest (24Laub F. Aldabe R. Friedrich Jr., V. Ohnishi S. Yoshida T. Ramirez F. Dev. Biol. 2001; 233: 305-318Crossref PubMed Scopus (84) Google Scholar). Second, KLF7 may directly regulate expression of TrkA, which, as the receptor for nerve growth factor, is required for normal maturation and differentiation of sensory and sympathetic neurons (25Lei L. Ma L. Nef S. Thai T. Parada L.F. Development. 2001; 128: 1147-1158Crossref PubMed Google Scholar). The ability of molecules such as KLF4/GKLF to play critical roles in the proliferative state of the cell raised the issue of whether they play any role in the development of cancer. Intestinal samples from patients with familial adenomatous polyposis (FAP) and from multiple intestinal neoplasm (min) mice (a model of FAP) were monitored for expression of KLF4 (26Dang D.T. Bachman K.E. Mahatan C.S. Dang L.H. Giardiello F.M. Yang V.W. FEBS Lett. 2000; 476: 203-207Crossref PubMed Scopus (88) Google Scholar). Reverse transcription-polymerase chain reaction analysis revealed an inverse correlation between KLF4 levels and intestinal adenoma tumor size in min mice and decreased levels of KLF4 in colonic adenomas from FAP patients compared with neighboring normal mucosa. As the KLF4 promoter contains binding sites for the Cdx-2 protein (27Mahatan C.S. Kaestner K.H. Geiman D.E. Yang V.W. Nucleic Acids Res. 1999; 27: 4562-4569Crossref PubMed Scopus (64) Google Scholar), a model has been proposed whereby mutated adenomatous polyposis coli can no longer induce Cdx-2, leading to low levels of KLF4 and accelerated growth in FAP samples. Consistent with this idea, KLF4 levels remain very low in the RKO colon cancer cell line, which contains wild-type adenomatous polyposis coli but a mutated Cdx-2, a variant that also exerts a dominant negative effect on wild type Cdx-2 activation of the KLF4 promoter (28Dang, D. T., Mahatan, C. S., Dang, L. H., Agboola, I. A., and Yang, V. W. (2001) Oncogene, in pressGoogle Scholar). Although a similar negative correlation is seen between KLF4 level and growth in prostatic carcinoma, the role of KLF4 is not equivalent in all cancers; for example, KLF4 levels are up-regulated during progression of human oral/pharyngeal and breast carcinomas (29Foster K.W. Frost A.R. McKie-Bell P. Lin C.Y. Engler J.A. Grizzle W.E. Ruppert J.M. Cancer Res. 2000; 60: 6488-6495PubMed Google Scholar). The biological roles of three KLF family members have been tested by genetic ablation. Consistent with its restricted expression and molecular properties, disruption of KLF1/EKLF leads to a directed effect on β-globin expression (see below). However, other members, more generally expressed, exhibit very specific phenotypes upon their disruption. In addition to expression in various tissues within the gut (see above), KLF4/GKLF is also expressed at high levels in the epidermis. However, its ablation leads to a specific deficiency in the barrier function of the skin resulting in postnatal death (19Segre J.A. Bauer C. Fuchs E. Nat. Genet. 1999; 22: 356-360Crossref PubMed Scopus (659) Google Scholar). Although these data are consistent with KLF4 function in growth arrest and differentiation, the extent of normal development seen in its absence is paradoxical given the expression pattern and postulated role of KLF4 in the intestine and other endodermal tissues. Similarly, genetic disruption of KLF2/LKLF, a molecule originally named by virtue of its high level of expression in the lung (30Anderson K.P. Kern C.B. Crable S.C. Lingrel J.B. Mol. Cell. Biol. 1995; 15: 5957-5965Crossref PubMed Scopus (231) Google Scholar), led to defects in blood vessel organization and early embryonic death (31Kuo C.T. Veselits M.L. Barton K.P. Lu M.M. Clendenin C. Leiden J.M. Genes Dev. 1997; 11: 2996-3006Crossref PubMed Scopus (315) Google Scholar). Even though angiogenesis and vasculogenesis were normal, recruitment of pericytes and smooth muscle cells was deficient, leading to a vessel wall of low integrity and severe, lethal hemorrhage. The embryonic lethality of KLF2-null embryos made it difficult to analyze effects on other tissues. However, two other experimental approaches addressed this. To test the importance of KLF2 in lymphoid differentiation and at the same time avert embryonic lethality, KLF2-deficient embryonic stem (ES) cells were injected to recombinase-deficient blastocysts to generate chimeric mice (32Kuo C.T. Veselits M.L. Leiden J.M. Science. 1997; 277: 1986-1990Crossref PubMed Scopus (349) Google Scholar). B cell development was normal, but mature, single-positive T lymphocytes were susceptible to apoptosis and did not survive, implicating a role for KLF2 in quiescent T cells. To address the role of KLF2 in the lung, ES cells were again used to generate chimeric mice, and the contribution of the KLF2-deficient cells to a large number of tissues was determined (33Wani M.A. Wert S.E. Lingrel J.B. J. Biol. Chem. 1999; 274: 21180-21185Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar). In this case, KLF2-deficient cells contributed to all tissues except the lung, and histopathological analysis of lungs from highly chimeric animals that died at birth showed deficiency in late stages of development and an abnormal pathology. A striking consistency in the KLF1, -2, and -4 knockouts is that they affect late stages of differentiation within their respective cellular environments. KLF1/EKLF was originally isolated by a subtractive cloning approach to identify genes important for erythroid differentiation (4Miller I.J. Bieker J.J. Mol. Cell. Biol. 1993; 13: 2776-2786Crossref PubMed Scopus (667) Google Scholar). At present, it remains the best characterized member of the group (34Perkins A. Int. J. Biochem. Cell Biol. 1999; 31: 1175-1192Crossref PubMed Scopus (57) Google Scholar,35Bieker J.J. Ravid K. Licht J.D. Transcription Factors: Normal and Malignant Development of Blood Cells. Wiley-Liss, New York2000: 71-84Crossref Google Scholar), as its target binding sequence at the β-globin promoter and its transcriptional activation properties were quickly identified. Its expression pattern has not, however, been a paradigm for the family, as KLF1 is restricted in its expression only to blood-forming tissues during mammalian development (36Southwood C.M. Downs K.M. Bieker J.J. Dev. Dyn. 1996; 206: 248-259Crossref PubMed Scopus (92) Google Scholar). Its genetic ablation not only leads to the absence of adult β-globin expression and embryonic death (37Nuez B. Michalovich D. Bygrave A. Ploemacher R. Grosveld F. Nature. 1995; 375: 316-318Crossref PubMed Scopus (492) Google Scholar,38Perkins A.C. Sharpe A.H. Orkin S.H. Nature. 1995; 375: 318-322Crossref PubMed Scopus (540) Google Scholar) but also the loss of the chromatin hypersensitive site at the β-globin promoter and a diminution of another strong hypersensitive site (HS3) located more than 50 kilobase pairs away (39Wijgerde M. Gribnau J. Trimborn T. Nuez B. Philipsen S. Grosveld F. Fraser P. Genes Dev. 1996; 10: 2894-2902Crossref PubMed Scopus (184) Google Scholar). A molecular explanation for these results has followed from determining that KLF1/EKLF interacts with p300 and CBP (40Zhang W. Bieker J.J. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 9855-9860Crossref PubMed Scopus (328) Google Scholar), transcriptional coactivators that also acetylate histones and KLF1 itself. Such modifications appear to alter the ability of KLF1 to interact with other proteins, such as components of the SWI/SNF, ATP-dependent chromatin remodeling complex (41Armstrong J.A. Bieker J.J. Emerson B.M. Cell. 1998; 95: 93-104Abstract Full Text Full Text PDF PubMed Scopus (276) Google Scholar, 42Lee C.H. Murphy M.R. Lee J.S. Chung J.H. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 12311-12315Crossref PubMed Scopus (90) Google Scholar, 43Zhang W. Kadam S. Emerson B.M. Bieker J.J. Mol. Cell. Biol. 2001; 21: 2413-2422Crossref PubMed Scopus (159) Google Scholar). As a result, KLF1 has provided the clearest example of the role in transcriptional regulation and chromatin assembly that a KLF molecule can integrate. Interestingly, the basic region adjacent to the zinc finger domain contains one of the KLF1/EKLF acetylation sites, and this region is most highly conserved among KLF1, KLF2/LKLF, and KLF4/GKLF (43Zhang W. Kadam S. Emerson B.M. Bieker J.J. Mol. Cell. Biol. 2001; 21: 2413-2422Crossref PubMed Scopus (159) Google Scholar). Indeed, KLF4 has also been shown to interact with the p300 and CBP coactivators in vitro, and the residues required for this interaction are also required for KLF4 to exhibit its growth-suppressive effects in vivo (44Geiman D.E. Ton-That H. Johnson J.M. Yang V.W. Nucleic Acids Res. 2000; 28: 1106-1113Crossref PubMed Google Scholar). As KLF1, -2, and -4 form a closely related subgroup within the KLF family (Fig. 1,subgroup 3), KLF2 and -4 may also be targets for acetylation that alters their ability to interact with other proteins. Another characteristic common to this subgroup follows from deletion analysis of their transactivation regions. Not surprisingly, in each case it can be whittled down to a minimal activation module (44Geiman D.E. Ton-That H. Johnson J.M. Yang V.W. Nucleic Acids Res. 2000; 28: 1106-1113Crossref PubMed Google Scholar, 45Chen X. Bieker J.J. EMBO J. 1996; 15: 5888-5896Crossref PubMed Scopus (60) Google Scholar, 46Yet S.F. McA'Nulty M.M. Folta S.C. Yen H.W. Yoshizumi M. Hsieh C.M. Layne M.D. Chin M.T. Wang H. Perrella M.A. Jain M.K. Lee M.E. J. Biol. Chem. 1998; 273: 1026-1031Abstract Full Text Full Text PDF PubMed Scopus (171) Google Scholar, 47Conkright M.D. Wani M.A. Lingrel J.B. J. Biol. Chem. 2001; 276: 29299-29306Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar). However, a region adjacent to their zinc fingers behave functionally as inhibitory modules (45Chen X. Bieker J.J. EMBO J. 1996; 15: 5888-5896Crossref PubMed Scopus (60) Google Scholar, 46Yet S.F. McA'Nulty M.M. Folta S.C. Yen H.W. Yoshizumi M. Hsieh C.M. Layne M.D. Chin M.T. Wang H. Perrella M.A. Jain M.K. Lee M.E. J. Biol. Chem. 1998; 273: 1026-1031Abstract Full Text Full Text PDF PubMed Scopus (171) Google Scholar, 47Conkright M.D. Wani M.A. Lingrel J.B. J. Biol. Chem. 2001; 276: 29299-29306Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar), implying that the roles of KLF1, -2, and -4 in transcriptional control is complex and may be sensitive to selective modification and protein interactions. Recently, KLF13/FKLF2 has been shown to interact with coactivators to stimulate transcription of the human γ-globin gene (48Song, C. Z., Asano, H., and Stamatoyannopoulos, G. (2001) Mol. Cell. Biol., in pressGoogle Scholar). Similar to KLF1, KLF13 is a substrate for acetylation by CBP and p300. However, unlike KLF1, P/CAF also acetylates KLF13, and this enzymatic activity is required for enhancement of KLF13 transcription. Part of this activation likely follows from the strong stimulation of DNA binding by KLF13 in the presence of CBP or P/CAF. The cumulative data suggest that KLF factors are selective in their utilization of coactivators to stimulate transcription. A considerably different outcome arises from investigating the role of KLF3/BKLF, which primarily behaves as a strong transcriptional repressor. KLF3 was originally isolated by low stringency cDNA library screening with the KLF1 zinc finger region (49Crossley M. Whitelaw E. Perkins A. Williams G. Fujiwara Y. Orkin S.H. Mol. Cell. Biol. 1996; 16: 1695-1705Crossref PubMed Scopus (209) Google Scholar). It is highly expressed in hematopoietic cells and the developing central nervous system and to a lesser extent in many (but not all) adult tissues. It achieves repression by recruitment of CtBP2, a general corepressor protein that interacts with KLF3 by means of a Pro-X-(Asp/Asn)-Leu-(Ser/Thr) motif located in the KLF3 repression domain (50Turner J. Crossley M. EMBO J. 1998; 17: 5129-5140Crossref PubMed Scopus (281) Google Scholar). As preliminary data show that genetic ablation of KLF3 gives rise to a myeloproliferative disorder that infiltrates numerous tissues and effectively interferes with their normal growth (51Perkins A.C. Yang H. Crossley M. Fujiwara Y. Orkin S. Blood. 1997; 90: 575Google Scholar), it appears that its repression function may play a role in preventing unrestrained cell number expansion. KLF3 is not unique in its interaction with the CtBP2 corepressor, as KLF8/BKLF3 also contains a region of homology to KLF3 (in addition to the zinc finger domain) with the appropriate CtBP2 interaction motif that enables it to repress transcription (52van Vliet J. Turner J. Crossley M. Nucleic Acids Res. 2000; 28: 1955-1962Crossref PubMed Scopus (132) Google Scholar). KLF8 is broadly expressed, and endogenous target genes have not yet been determined for either KLF3 or -8. However, KLF12/AP2-rep, whose target promoter is known, contains the same interaction motif and functions as a repressor of the AP2α promoter in cotransfection assays (53Imhof A. Schuierer M. Werner O. Moser M. Roth C. Bauer R. Buettner R. Mol. Cell. Biol. 1999; 19: 194-204Crossref PubMed Google Scholar). Although KLF12 was originally isolated from a brain cDNA library, it is most highly expressed in the kidney, and induction of KLF12 expression correlates with down-regulation of AP2α gene expression during kidney development. Interestingly, KLF3, -8, and -12 are most closely related to each other simply by comparative sequence analysis in the absence of functional tests (Fig. 1, subgroup 1). As a result, it was not unexpected to find that KLF12 interacts with CtBP1 (54Schuierer M. Hilger-Eversheim K. Dobner T. Bosserhoff A.K. Moser M. Turner J. Crossley M. Buettner R. J. Biol. Chem. 2001; 276: 27944-27949Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar). However, the adenovirus E1A protein also contains the CtBP1 interaction motif, and recent experiments suggest that the ability of E1A to derepress the AP2α promoter results from its interaction with CtBP1, which prevents CtBP1 from productively interacting with KLF12, thus functionally inactivating KLF12 repression of AP2α (54Schuierer M. Hilger-Eversheim K. Dobner T. Bosserhoff A.K. Moser M. Turner J. Crossley M. Buettner R. J. Biol. Chem. 2001; 276: 27944-27949Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar). Given these examples of molecular repressors and activators within the KLF family, it may not be completely surprising to find examples of cross-regulation. Three are illustrative. The original screen that had isolated KLF12/AP2-rep (above) had also identified KLF9/BTEB1 binding to the AP2α promoter CACCC element (53Imhof A. Schuierer M. Werner O. Moser M. Roth C. Bauer R. Buettner R. Mol. Cell. Biol. 1999; 19: 194-204Crossref PubMed Google Scholar). Although KLF12 behaved as a repressor in transfection assays, KLF9 was a strong activator of the AP2α promoter. Their mutually exclusive binding leads to different reporter activities that are dependent on the relative levels of each protein. Unlike KLF12, KLF9 is expressed in many tissues. This provides an example of KLF factors whose differing modes of action (activationversus repression) may play a role in regulating target gene expression from the same site. KLF4/GKLF/EZF and KLF5/IKLF/BTEB2 have been implicated in the antagonistic regulation of two gene systems. As already described, KLF4 plays a role in reduced proliferation and increased differentiation of the small intestinal epithelium via its effects on cell cycle regulators. Conversely, KLF5 is primarily expressed in the proliferating cells of the crypt epithelium (55Sun R. Chen X. Yang V.W. J. Biol. Chem. 2001; 276: 6897-6900Abstract Full Text Full Text PDF PubMed Scopus (140) Google Scholar). KLF5 is induced by mitogens, and its overexpression in fibroblasts leads to their increased growth and a transformed phenotype. This leads to a model in which KLF4 and KLF5 play opposing roles in differentiation and proliferation in the intestine (55Sun R. Chen X. Yang V.W. J. Biol. Chem. 2001; 276: 6897-6900Abstract Full Text Full Text PDF PubMed Scopus (140) Google Scholar, 56Conkright M.D. Wani M.A. Anderson K.P. Lingrel J.B. Nucleic Acids Res. 1999; 27: 1263-1270Crossref PubMed Scopus (144) Google Scholar). The α-smooth muscle and SM22α promoters contain CACCC elements that were used in a yeast one-hybrid screen to isolate KLF4 (57Adam P.J. Regan C.P. Hautmann M.B. Owens G.K. J. Biol. Chem. 2000; 275: 37798-37806Abstract Full Text Full Text PDF PubMed Scopus (218) Google Scholar). However, KLF4 was found to repress TGF-β induction of the α-smooth muscle actin and SM22α promoters in transfection assays; indeed, TGF-β treatment of smooth muscle cells in culture leads to a decrease in KLF4 levels. However, KLF5/BTEB2/IKLF (but not KLF2) increased the activity of the α-smooth muscle and SM22α promoters in transient assays. As KLF5 is abundant in smooth muscle tissues and is preferentially activated during their proliferation (58Watanabe N. Kurabayashi M. Shimomura Y. Kawai-Kowase K. Hoshino Y. Manabe I. Watanabe M. Aikawa M. Kuro-o M. Suzuki T. Yazaki Y. Nagai R. Circ. Res. 1999; 85: 182-191Crossref PubMed Scopus (126) Google Scholar), the regulatory model again is based on opposing effects of KLF4 and KLF5, this time during muscle differentiation. A more subtle effect may be apparent during blood cell development, particularly with respect to regulation of globin gene switching (embryonic to fetal to adult) within the β-like cluster. As KLF1/EKLF appears dedicated to consolidating the switch from human fetal to adult β-globin expression, and because the other genes in the cluster also contain CACCC elements within their promoters, a search for related members that may play a role in embryonic and fetal β-like globin expression led to the identification of KLF11/FKLF (59Asano H. Li X.S. Stamatoyannopoulos G. Mol. Cell. Biol. 1999; 19: 3571-3579Crossref PubMed Google Scholar) and KLF13/FKLF2 (60Asano H. Li X.S. Stamatoyannopoulos G. Blood. 2000; 95: 3578-3584Crossref PubMed Google Scholar). KLF11 primarily activates the embryonic and KLF13 the fetal globin promoter, although KLF13 also stimulates non-globin promoters to a lesser extent. However, another factor highly expressed in erythroid cells is KLF3/BKLF. As discussed above, KLF3 primarily behaves as a repressor but in addition may be indirectly regulated by EKLF (61Turner J. Crossley M. Int. J. Biochem. Cell Biol. 1999; 31: 1169-1174Crossref PubMed Scopus (59) Google Scholar). These data raise the possibility that at different stages of red blood cell development, KLF11 and/or KLF13 compete with KLF1 to optimally activate their cognate high affinity targets in the β-globin cluster, and at the same time, expression of KLF3 may serve to repress the embryonic and fetal members. Clearly, because all of these CACCC binding factors are present together in the red cell in development when globin gene switching is occurring, the mechanism of how a particular target globin promoter is specifically induced at the correct time in the midst of so many effector molecules remains perplexing. This review has concentrated on specific properties of selected KLF family members; however, it is not meant to imply that their roles are thereby limited. For example, KLF1 has historically been characterized as a strong transcriptional activator with specific roles in β-globin transcription and β-locus chromatin integrity; however, recent data indicate that it may also play a role in erythroid cell proliferation (62Coghill E. Eccleston S. Fox V. Cerruti L. Brown C. Cunningham J. Jane S. Perkins A. Blood. 2001; 97: 1861-1868Crossref PubMed Scopus (73) Google Scholar) and may even function as a transcriptional repressor in specific contexts (63Chen X. Bieker J.J. Mol. Cell. Biol. 2001; 21: 3118-3125Crossref PubMed Scopus (68) Google Scholar). How do these proteins exert their particular effects, given that they contain such similar DNA binding regions that bind to virtually identical sequences? Studies with KLF1 are instructive. At one level, specificity can follow from tissue-restricted expression. However, not only are a number of these factors expressed in multiple tissues but most cells express more than one factor at any time. As a result, the second level of specificity is via their respective activation/repression domains, which are unique to each member, and thus likely determine their resultant protein/protein interactions. This has been directly tested for KLF1, where in vivo tests of β-globin promoter activation by transient transfection assays (64Bieker J.J. Southwood C.M. Mol. Cell. Biol. 1995; 15: 852-860Crossref PubMed Google Scholar) and in transgenic mice (65Gillemans N. Tewari R. Lindeboom F. Rottier R. de Wit T. Wijgerde M. Grosveld F. Philipsen S. Genes Dev. 1998; 12: 2863-2873Crossref PubMed Scopus (62) Google Scholar) demonstrate that the Sp1 transactivation domain cannot substitute for the KLF1 transactivation domain. Third, the zinc finger region can also play a role in KLF/protein interactions, whether bound to DNA (e.g. KLF1 interaction with DNA and SWI/SNF proteins (66Kadam S. McAlpine G.S. Phelan M.L. Kingston R.E. Jones K.A. Emerson B.M. Genes Dev. 2000; 14: 2441-2451Crossref PubMed Scopus (193) Google Scholar)) or not (e.g. KLF1 behavior as a repressor (63Chen X. Bieker J.J. Mol. Cell. Biol. 2001; 21: 3118-3125Crossref PubMed Scopus (68) Google Scholar)). Finally, although the KLF DNA binding domains interact with specific nucleotides within the CACCC element that directly affect their binding affinity, it is also clear that the overall architecture and context within which this element is located can have a dramatic effect on the ability of KLF factors (e.g. KLF1 (67Asano H. Stamatoyannopoulos G. Mol. Cell. Biol. 1998; 18: 102-109Crossref PubMed Google Scholar, 68Lee J.S. Lee C.H. Chung J.H. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 10051-10055Crossref PubMed Scopus (21) Google Scholar, 69Tanimoto K. Liu Q. Grosveld F. Bungert J. Engel J.D. Genes Dev. 2000; 14: 2778-2794Crossref PubMed Scopus (69) Google Scholar)) to bind and exert their transcriptional effects. Although this review has focused on the mammalian members of the KLF family, it is clear from accumulated sequence analyses that large numbers of Cys2/His2 zinc finger proteins are also encoded by the Drosophila, Caenorhabditis elegans, and Danio rerio genomes (1Tupler R. Perini G. Green M.R. Nature. 2001; 409: 832-833Crossref PubMed Scopus (311) Google Scholar, 2Lander E.S. et al.Nature. 2001; 409: 860-921Crossref PubMed Scopus (18333) Google Scholar, 3Venter J.C. et al.Science. 2001; 291: 1304-1351Crossref PubMed Scopus (10873) Google Scholar). Significant amino acid homologies among KLF family members beyond their DNA binding regions exist only between very closely related family members, and even then it is quite low. These families have expanded independently in different species, and thus direct functional analogies may be unobtainable by comparison of mammalian and non-mammalian KLF proteins. However, the localization of evolutionarily conserved domains and any proteins and pathways with which these domains may interact and link with in non-mammalian systems will prove useful for directing tests of homologous regions in their mammalian counterparts.