Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Hepatocyte nuclear factor 3 alpha belongs to a gene family in mammals that is homologous to the Drosophila homeotic gene fork head.

Hepatocyte nuclear factor 3 alpha belongs to a gene family in mammals that is homologous to the... Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Hepatocyte nuclear factor belongs to a gene family in mammals that is homologous to the Drosophila homeotic gene fork head Eseng Lai, 1 Vincent R. Prezioso, 2 Wufan Tao, 1 William S. Chen, z and James E. Darnell, Jr. 2'3 IDivision of Endocrinology and Program of Cell Biology and Genetics, Memorial Sloan-Kettering Cancer Center, New York, New York 10021 USA~ ~Laboratory of Molecular Cell Biology, Rockefeller University, New York, New York 10021 USA By analysis of cDNA clones that cross-hybridized with a portion of the cDNA encoding the recently described rat protein hepatocyte nuclear factor 3a (HNF-3oq previously called HNF-3A), we now describe two additional members, HNF-313 and HNF-3~,, of this gene family. A 110-amino-acid region in the DNA-binding domain of this family is not only very highly conserved in rodents (HNF-3% -313, and -3~ are identical in 93 of 110 amino acids in this region) but also in Drosophila where the homeotic gene fork head has 88 of the 93 residues that are identical in the three rat genes. The HNF-3 family in rodents is expressed in cells that derive from the lining of the primitive gut; some of the embryonic Drosophila cells in which fork head is expressed also give rise to gut and salivary glands. Thus, it appears that this gene family, the DNA-binding portion of which is unlike that of any previously recognized DNA-binding proteins, may contribute to differentiation of cells in internal organs in both vertebrates and invertebrates. [Key Words: Transcriptional control; HNF-3 family; mammalian development] Received November 27, 1990; revised version accepted December 28, 1990. whose eDNA we cloned recently, is transcriptionally ac- Using the techniques of molecular genetics to identify required DNA-binding sites for gene function and then tive in liver but not in a number of other tissues. Liver protein purification by oligonucleotide affinity chroma- extracts, however, interact with a labeled oligonucle o tography of proteins cognate with these sites, several lab- otide containing an HNF-3 site to produce several bands oratories including our own have identified transcription in a gel mobility-retardation assay (Costa et al. 1989; Lai factors that direct expression of genes specifically in he- et al. 1990), and several mRNAs were detected in North- patocytes (Johnson et al. 1987; Frain et al. 1989; Baum- ern blots with the cloned sequence of the HNF-3ot eDNA hueter et al. 1990; Lai et al. 1990; Sladek et al. 1990). (Lai et al. 1990). These results most likely indicate either Hepatocyte-specific gene expression of such characteris- a single transcriptional product that is differentially pro- tic products as albumin begins quite early (-11 or 12 cessed or the transcription of several related but distinct days of gestation) in the mouse embryo (Tilghman and genes. Belayew 1982), indicating the probable early activity of Characterization of additional HNF-3-1ike eDNA from transcription factors that are also active in the adult rodent liver has now revealed that HNF-3c~ belongs to a liver. Furthermore, some of these factors, for example, gene family with a tightly conserved -ll0-amino-acid hepatocyte nuclear factors 1 and 3 (HNF-1 and HNF-3) sequence within the region shown earlier to be required appear to be present in yolk sac cells (Costa et al. 1990), for DNA binding. This conserved DNA-binding domain which, like the hepatocyte, are derived from gut endo- is unlike that of any previously recognized family of derm. Because the genes for some of the liver-enriched transcription factors. A recently reported sequence com- transcription factors mentioned above are transcription- parison further heightens the interest in the HNF-3 fam- ally controlled themselves (Xanthopoulos et al. 1989, ily as consisting of important contributing members to and in prep.), a study of their regulation should provide early development (Weigel and J~ickle 1990). The protein access to genes that are important in early developmen- encoded by the Drosophila gene fork head, mutants of tal cascades. which fail to make proper terminal embryonic tissues Hepatocyte nuclear factor 3~ (HNF-3~, previously (the proctodeum and stomodeum, which contribute cells called HNF-3A), one of the factors purified from liver and to the foregut and hindgut), has an extensive amino acid sequence homology to the HNF-3A sequence described. 3Corresponding author. The greatest sequence conservation is in the DNA-bind- 416 GENES & DEVELOPMENT 5:416-427 © 1991 by Cold Spring Harbor Laboratory ISSN 0890-9369/91 $1.00 Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press HNF-3 gene |amily ing domain of all the HNF-3 family members. The ho- hybridize to either one and therefore represented poten- meo box proteins that participate in Drosophila embry- tial additional family members. Six of these clones were onic larval segmentation are segmentally expressed; and, shown by restriction mapping and partial sequencing to presumably, similarly functioning homologs in mam- be overlapping fragments of a single cDNA that was dif- mals (Hox genes) are also segmentally expressed (Akam ferent from the two cDNA clones isolated previously. A 1989; Graham et al. 1989). By analogy, it seems likely cDNA containing the complete coding sequence of this that the HNF-3 family and fork head (as well as any third family member was reconstructed from two over- lapping clones, L11 and L35. The joint L11/L35 sequence other Drosophila genes like fork head) may function in contains an ORF of 322 amino acids or -140 amino acids constructing internal organs from the embryonic gut. Thus, there may be an evolutionarily conserved genetic shorter than the proteins encoded by the first two pathway operating in the differention of cells of internal cDNAs (Fig. 2). The position of the translation initiation organs of both vertebrates and invertebrates. site in this third sequence was verified by use of a series In addition to delineating three distinct members of of 5' deletions that were transcribed in vitro with SP6 or the HNF-3 family, we also show that (1) the mRNAs for T7 RNA polymerase and translated in reticulocyte ly- these proteins in rats are distributed differently in differ- sates (Fig. 3B). Deletion of 311 bp of the 5' end of the ent organs that derive from the primitive gut, and (2) cDNA did not alter the size of the translated protein (d7). An additional deletion of 84 nucleotides (d8) resulted in HNF-3 proteins are positive activating proteins. The a shortened translation product (presumably initiating at study of the HNF-3 family whose members are differen- the ATG at position 537-539), confirming that the major tially expressed in different cell types promises to be of translation initiation site must be one of the two tandem considerable value in opening new insights into early ATGs at position 312-314 and 315-317 (Fig. 2). mammalian development. We renamed these three related HNF-3 cDNAs, HNF-3e~ (the original HNF-3A), HNF-3[3 (K2), and Results HNF-37 (Lll/L35). The nucleic acid sequence compari- son of HNF-3a, HNF-3[3, and HNF-37 suggests that these Isolation of cDNA clones for HNF-3[3 and HNF-37 cDNAs are products of three separate but related genes, rather than the products of alternative splicing. For ex- Liver extracts exhibit multiple DNA-binding activities ample, in the regions of greatest homology, the amino for a labeled double-stranded deoxyoligonucleotide con- acid sequence is more highly conserved than the nucleic taining an HNF-3 site, and Northern analysis of liver acid sequence. Genomic Southern analyses using probes mRNA with labeled DNA of the original HNF-3A cDNA specific to each cDNA showed hybridization to a differ- clone shows multiple mRNAs (Lai et al. 1990). There- ent pattern of genomic DNA fragments, thus supporting fore, we suspected that there might be several related genes that encode transcription factors sharing binding the existence of three separate and likely single-copy specificity for HNF-3 sites in DNA. When we selected genes. Autoradiographs from Southern analysis of rat ge- and sequenced a murine cDNA that cross-hybridized nomic DNA for HNF-3a and HNF-3[3 are shown in Fig- with an HNF-3A probe, we isolated a partial cDNA ure 4. clone, suggesting that more than one HNF-3-1ike gene existed. This clone, M2, was highly similar in one region HNF-3[3 and HNF-37, like HNF-3a, encode of the two sequences but diverged markedly outside of sequence-specific DNA-binding proteins this region. We used a probe derived from the divergent RNA transcripts from the HNF-3a, HNF-313, and sequence to identify the rat counterpart of the mouse M2 HNF-3~/clones were produced and translated in reticu- cDNA and obtained a clone, termed K2, which was locyte lysates, and the products were analyzed on SDS- >90% identical to the entire mouse M2 sequence and polyacrylamide gels {Fig. 3AI. The monomer molecular distinctly different from the original HNF-3A (rat) gene in most of its sequence. The K2 rat cDNA contained a masses of HNF-3a, HNF-313, and HNF-3~, respectively, single large open reading frame (ORF) that would encode are 50, 47, and 42 kD. The product of HNF-3a has been a protein of 458 amino acids (Fig. 1 ). The selection of an shown previously to bind specifically to the proximal initiator methionine was suggested by an adjacent se- HNF-3 site in the transthyretin (TTR} promoter {Lai et quence that conformed well to the consensus for eukary- al. 19901. We now show that the products of both HNF-313 and HNF-37 also bind specifically to an oligo- otic initiation {Kozak 1986) and the presence of a termi- nucleotide representing the original HNF-3-binding site nation codon immediately upstream. Comparison of the {Fig. 3C). For comparison we show the complexes formed amino acid sequence with that of HNF-3A indicated that with crude liver extract. Three of these complexes have these two proteins shared a highly homologous central been referred to previously as A, B, and C in decreasing domain but differed in most of the coding region. To obtain additional members of this family, we re- size {or increasing gel mobility}. screened 500,000 recombinants of the rat liver cDNA Because we purified and sequenced fragments of the library, using a probe to the conserved sequence, and HNF-3A protein, we can definitively state that it is iden- isolated 20 positive clones. We then used probes derived tical to the product of the HNF-3a gene. For the products from nonhomologous sequences in the HNF-3A and the of HNF-313 and HNF-3~, indirect evidence indicates that they are identical or closely related to the HNF-3B and K2 cDNAs to identify 7 clones of the 20 that did not GENES & DEVELOPMENT 417 Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Lai et al. TCTCcCGGGTATTGGCTGcAGCTAAGcGGGGcTTCCTGGGCCGACTGAGGTGGGTAGCCAGAAAAAGGCCTGAGGTGGCTGAcAACcAGGGcGGCCAGACAACGCGAGTCCTGCGcGCCTTCTGAGGCCTc 131 CCTGGGACTTAACTGTAACGGGGAGGGGCCTCTGGAGCAGCGGCCAGCGAGTTAAAGT ATG CTG GGA GCC GTG AAG ATG GAA GGG CAC GAG CCA TCC GAC TGG AGC AGC TAC 243 M L G A V K M E G H E P S D W $ S Y 18 TAC GCG GAG CCT GAG GGC TAC TCT TCC GTG AGC AAC ATG AAC GCC AGC CTG GGG ATG AAT GGC ATG AAC ACT TAC ATG AGC ATG TCC GCG GCT GCA ATG 342 Y A E P E G Y S S V S N M N A S L G M N G M N T Y M S M S A A A M 51 GGC AGT GGT TCC GGC AAC ATG AGC GCA GGC TCC ATG AAC ATG TCA TCC TAT GTG GGC GCT GGA ATG AGC CCG TCG CTG GCT GGC ATG TCC CCG GGC GCG 441 G S G S G N M S A G S M N M S S Y V G A G M S P S L A G M S P G A 84 GGC GCC ATG GCG GGC ATG AGC GGC TCA GCT GGG GCG GCC GGC GTG GCG GGC ATG GGA CCG CAC CTG AGT CCG AGT CTG AGC CCA CTC GGG GGA CAG GCG 540 G A M A G M S G S A G A A G V A G M G P H L S P S L S P L G G Q A 117 GCC GGG GCT ATG GGT GGC CTT GCT CCC TAC GCC AAT ATG AAC TCC ATG AGT CCT ATG TAC GGG CAG GCG GGC CTG AGC CGC GCT CGG GAC CCC AAG ACG 639 A G A M G G L A P Y A N M N S M S P M Y G Q A G L S R A R D P K T 150 TAC CGG CGC AGC TAC ACT CAC GCC AAG CCT CCC TAC TCG TAC ATC TCG CTC ATC ACC ATG GCC ATC CAG CAG AGC CCC AAC AAG ATG CTG ACG CTG AGC 738 Y R R S Y T H A K P P Y S Y I S L I T M A I Q Q S P N K M L T L S 183 GAG ATC TAT CAG TGG ATC ATG GAC CTC TTC CCT TTC TAC CGG CAG AAC CAG CAG CGC TGG CAG AAC TCC ATC CGT CAT TCT CTC TCC TTC AAC GAC TGC 837 E I Y Q W I M D L F P F Y R Q N Q Q R W Q N S I R H S L S F N D C 216 TTT CTC AAG GTG CCC CGC GCG CCA GAC AAG CCT GGG GGC TCC TTC TGG ACC CTG CAC CCT GAC TCT GGC AAC ATG TTC GAG AAC GGT TGC TAC CTG CGC 936 F L K V P R A P D K P G G $ F W T L H P D S G N M F E N G C Y L R 249 CGC CAG AAG CGC TTC AAG TGT GAG AAG CAA CTG GCG TTG AAG GAA GCA GCG GGT GCG GGC AGT GGC GGA GGC AAG AAG ACC GCT CCT GGG ACA CAG GCT 1035 R Q K R F K C E K Q L A L K E A A G A G S G G G K K T A P G T Q A 282 TCT CAG GTT CAG CTC GGG GAG GCC GCA GGC TCG GCC TCT GAG ACT CCG GCG GGC ACC GAG TCC CCC CAT TCC AGC GCT TCT CCG TGT CAG GAG CAC AAG 1134 S Q V Q L G E A A G S A S E T P A G T E S P H $ $ A S P C Q E H K 315 CGA GGT GGC CTG AGC GAG CTG AAG GGA ACA CCT GCC TCT GCG CTG AGT CCT CCG GAG CCG GCG CCC TCG CCT GGG CAG CAG CAG CAG GCT GCA GCC CAC 1233 R G G L S E L K G T P A S A L S P P E P A P S P G Q Q Q Q A A A H 348 CTG CTG GGC CCA CCT CAC CAT CCT GGC CTG CCA CCA GAG GCC CAC CTG AAG CCC GAG CAC CAT TAC GCC TTC AAC CAC CCC TTC TCT ATC AAC AAC CTC 1332 L L G P P H H P G L P P E A H L K P E H H Y A F N H P F S I N N L 381 ATG TCC TCC GAG CAG CAA CAT CAT CAC AGC CAC CAC CAC CAT CAG CCC CAC AAA ATG GAC CTC AAG ACC TAC GAA CAG GTC ATG CAC TAC CCT GGG GGC 1431 M S S E Q Q H H H $ H H H H Q P H K M D L K T Y E O V M H Y P G G 414 TAC GGT TCC CCC ATG CCA GGC AGC TTG GCC ATG GGC CCA GTC ACG AAC AAA GCC GGC CTG GAT GCC TCG CCC CTG GCT GCA GAC ACT TCC TAC TAC CAG 1530 Y G S P M P G S L A M G P V T N K A G L D A S P L A A D T S Y Y Q 447 GGA GTG TAC TCC AGG CCT ATT ATG AAC TCG TCC TAA GAAGATGGCTTTCAGGCCCTGCTAGCTCTGGCCACTGGGGAGAAGGACAAGGGAAATGACAGGCTGAGTGGAAACATTGGGGG 1649 G V Y $ R P I M N S S - 458 AACTTTGAGGAAAAAGTAGCCACCACACTTAAGGCCCCAAAGGAGCAGTTTCACCTGTCTGTGTCCCTAAATAGCTGGGCCACACTGATCTGTcATTCTAAATAGGGAAGGGAATGGAAATATATATGTAT 1780 ACACATAAACTTGTTTTAATGGAGGACCTTTGGTTCCCACTATGTAGACTACTGCTTCTCAAGGCACCTGCAGATTTTGATTTTTGTTCTTGATTCTCTCTCTATTGCTGTCGTTGACAGAGAAGTCTGAC 1911 TTAAAAAAAAAAAAACTAAACAAAAAAAAAAACTTTTGATGAGTGACTTGAGTGTAAAACATGTAGTTTTAAAGAAAACAGAGGGTTGTATTGATGTTTAAAAGAGAAAAAAAAATAATGATGTAAG 2042 AGTCTGGTATAAATGACAGGAGAAAGGACAAAAATGATCCATTCTGGACATGGTGAAATCCAGTCTCGGGTCTGATTTAATTTATGGTTTCTGCGTGCTTTATTTATGGCTTATAAATGTGTGTTCTG 2173 GCTAGAATGCCAGAGTTCCACAAATCTATATTAAAGTATTATTGC 2217 Figure 1. Nucleic acid sequence of HNF-313. The cDNA clone (K2) contains the complete protein-coding sequence but is missing the 3' end of the mRNA. The protein sequence is shown below beginning at the initiator methionine at base 190, which is preceded by an in-frame termination codon. HNF-3C proteins, respectively. Definitive identification sates {lane 2), resulting in a much slower mobility com- of the proteins that form each of the gel mobility-shift plex, labeled T, which presumably represents a complex of probe, DNA-binding protein, and antibody. The anti-a complexes would require their purification and direct sequencing. The gel mobility-shift complex formed with serum also reacts weakly with the HNF-313 product, the HNF-313-translated product migrates identically to shifting a small part of that complex to a slower mobility that of HNF-3B. The gel mobility-shift complex formed (lane 4). with the HNF-3v product migrates slightly faster than The anti--y antiserum reacts with the HNF3~- the complex labeled HNF-3C. However, we obtained translated product {Fig. 5A, lane 6). This antiserum ap- two antisera that aid in the identification of the proteins pears to disrupt the binding of the protein to the DNA in the liver extracts. The first antiserum, which was rather than to change the mobility of the gel-shift com- plex. The corresponding controls for the antibody reac- raised in response to a synthetic peptide corresponding tions with preimmune serum are shown in Figure 5, to the amino-terminal 15 amino acids of HNF-3a, reacts with the protein products of the HNF-3a and HNF-313 lanes 1 and 3. genes (anti-a). HNF-313 shares 12 of 15 amino acids at the With crude liver extracts, the anti-or serum shifts the A amino terminus with HNF-3a. The second antibody was complex and a small part of the B complex (Fig. 5B, lane raised to a bacterial fusion product containing 77 amino 3). The anti-~/serum disrupts complex formation of the C acids of the HNF-3~/ sequence and is specific to the protein without affecting the complexes formed with A HNF-3~ protein (anti--y). The activity of anti-a is shown and B (Fig. 5B, lane 4). The specificity of the anti-~/an- in Figure 5A. This antiserum binds to the HNF-3a pro- tiserum is shown more clearly using partially purified duced by translation of this mRNA in reticulocyte ly- HNF-3C; anti-c~ serum does not affect binding, whereas 418 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press HNF-3 gene family TGTCGAGTGACTCCTCCTCTTCAGAAGAGGGCAATCTAGGGATAGTGAAGCT 52 GAAACACAGTGGAACAAGTCCTTGAAcCACAAGCTCATCGAACAGTTTTCTCGTTCCCCTTTGAGATATCAAGGAAGCTGTCACAGTCACTGTCTCATTCAGATGGTATAGATGCAGGCAGATTTTCTCAG 183 GGGGTCTCGGTGTAGATCTGGGACGGACCAGCACCATATACTTCGAGTTTCCATTGAGCAAGGAAGCAAAAGGCAGGAAGGCAAGAAGAGCTGAAAAAAGGCACAGCTGAGCCTCTCTGTCCTTCTTCATG 314 ATG GAA GGA CAG GAG GAC AGG ACG ACA GTA AGA CAA GAG GAC AGG GCT TAT GCT GTC CTC TTC CTT AAC ACT GCC TCC CCT CTG CCT ACA GGA CCC CTG 413 M E G Q E D R T T V R Q E D R A Y A V L F L N T A S P L P T G P L 33 GCA CCC CCA GCC CCC ACC GCT CCC CTG GGG CCC ACC TTC CCA GGC CTG GGC GCA GGC AGC GGC ACC GGA GGC AGT GCT TCC GGG TAT GGC GCC CCA GGG 512 A P P A P T A P L G P T F P G L G A G S G T G G S A S G Y G A P G 66 CCC GGG CTT GTA CAC GGA AAG GAA ATG GCC AAG GGG TAC CGG CGG CCA CTG GCT CAT GCC AAA CCA CCT TAT TCC TAC ATC TCT CTC ATC ACC ATG GCT 611 P G L V H G K E M A K G Y R R P L A H A K P P Y S Y I S L I T M A 99 ATC CAG CAG GCA CCA GGC AAG ATG CTG ACC CTG AGT GAA ATC TAC CAA TGG ATC ATG GAC CTC TTC CCG TAC TAC CGG GAG AAC CAG CAA CGT TGG CAG 710 I Q Q A P G K M L T L S E I Y Q W I M D L F P Y Y R E N Q Q R W Q 132 GTG GCA CGC TCC CCA GAC AAA CCG GGC AAA GGC TCC TAC TGG GCC TTG CAT CCC 809 AAC TCC ATC CGG CAC TCG CTG TCC TTC AAT GAC TGC TTC GTC AAG V A R S P D K P G K G S Y W A L H P 165 N S I R H S L S F N D C F V K AAG CGC TTC AAG CTG GAG GAG AAG GCA AAG AAA GGA AAC AGT GCC ACA TCC GCC 908 AGC TCT GGG AAC ATG TTT GAG AAC GGC TGC TAT CTC CGC CGT CAG K R F K L E E K A K K G N S A T S A 198 S S G N M F E N G C Y L R R Q GCT GCC ACT GCA GTC ACT TCC CCG GCT CAG CCC CAG CCT ACG CCG CCG TCT GAG 1007 ACC AGG AAT GGA ACA GTG GGG TCA GCC ACC TCT GCC ACC ACT ACA A A T A V T S P A Q P Q P T P P S E 231 T R N G T V G S A T S A T T T CCC GAG GCC CAG AGT GGG GAA GAT GTG GGG GGT CTG GAC TGC GCC TCG CCT CCT TCA TCC GCA CCC TAT TTC ACT GGC CTG GAG CTC CCA GGG GAA CTG 1106 P E A Q S G E D V G G L D C A S P P S S A P Y F T G L E L P G E L 264 AAG TTG GAT GCG CCC TAC AAC TTC AAC CAC CCT TTC TCT ATC AAC AAC CTG ATG TCA GAA CAG ACA TCA ACA CCT TCC AAA CTA GAT GTG GGG TTT GGG 1205 K L D A P Y N F N H P F S I N N L M S E Q T S T P S K L D V G F G 297 GGC TAC GGT GCT GAG AGT GGG GAA CCT GCG GTC TAC TAC CAG AGC CTC TAT TCC CGC TCT CTG CTC AAT GCA TCC TAG CGGCGCAACTGGGAGAAATGCGGTGAT 1310 G Y G A E S G E P G V Y Y Q S L Y S R S L L N A S - 322 GGGGGTTTGCTGTGACAGATGACCGGTTCTTCGGCCCTGAT~TTTCTGGTTACACTCTGCTTATCCAGTTAATTAACATCGTATTTGGTCTATTACTGTGATATGACCCATTGGCTACTGTGGTAACTGCA 1441 TGGACTCTTCGGTGGGCCTAGGGTTGGGGTATTGTGAAGGCAGATGCATTTGGGAGTGTTACGAAGGTGGCCATGTCGGACATACGTGAAGGCAATTAGACTGGTGTACTATAAAAGCTGCATGTTAAGT 1572 GAGTGATCCACTGGGTGCCTGATGGCCGCGATGTCGGAGGACATATTGTTTGGCCCTTTGGATGCT 1638 Figure 2. Nucleic acid sequence of HNF-3~/. The L11/35 cDNA contains the complete protein-coding sequence but is missing the 3' end of the mRNA. The protein sequence is shown below, beginning with two potential initiator methionines at bases 312 and 315. The second conforms much better to the consensus translation initiation sequence (Kozak 1986). anti-7 serum completely disrupts DNA binding {Fig. 5B, sential for DNA binding. The HNF-3a and HNF-313 pro- lanes 5-8). These data indicate that the HNF-3~/product teins are similar throughout the amino-terminal half of and HNF-3C share epitopes in the region of the protein the protein and are particularly rich in methionine in this region (Lai et al. 1990; Fig. 1}. Most of the rest of the that is distinct from HNF-3a or HNF-313. The molecular HNF-3a and HNF-313 proteins have recognizable se- mass on SDS gels of HNF-3C purified from rat liver ex- quence similarity (-30% identity) but are considerably tracts was determined previously to be 43 kD (Lai et al. 1990). Because the translated product of HNF-3~/is dis- divergent. The HNF-3~/ protein is -140 amino acids tinctly smaller on SDS-polyacrylamide gels, we con- shorter than HNF-3a and HNF-3~3, does not have a large clude that the two proteins result from differential mod- number of methionines in its amino terminus, and is ifications of the same gene product or are very closely highly divergent outside of the three conserved regions. After we completed the sequence of HNF-3A we con- related but not identical in sequence. ducted computer searches for matches with other pro- teins and failed to discover any. Later, when both the The major conserved region in HNF-3 proteins is in sequences of HNF-3A and the Drosophila gene fork head the DNA-binding domain had been entered in the available data bases, Weigel and J~ickle (1990) found a remarkable amino acid identity in A comparison of the amino acid sequences for the HNF-3 these two sequences within the domain shown to be proteins revealed three regions Ilabeled I, II, and III) that important for DNA binding of HNF-3A (HNF-3et). We are highly conserved in all three proteins (Fig. 6). We compared the fork head sequence to each of the HNF-3 demonstrated previously that the region between amino family members and found that there is greatest similar- acids 124 and 288 {numbering refers to HNF-3~} is re- ity to HNF-313, where 100 of 110 amino acids are iden- quired for DNA binding. Within this sequence is found tical in the binding domain (Fig. 6). The homology region I, a l l0-amino-acid stretch that is remarkably among all four proteins extends to regions II and III, as highly conserved in all three proteins. Ninety-three of well as to the binding domain, region I. the 110 amino acids are identical in the three proteins, whereas only one position is not shared between at least two of the three proteins. Nearly all of the differences HNF-3 proteins have different affinities for different represent conservative changes. Two other short, but binding sites well conserved, regions, II and III, are found in the car- boxyl end of the proteins. As reported earlier (Lai et al. The promoter of the TTR gene contains two HNF- 1990), the amino acids in these two regions are not es- 3-binding sites, both of which appear necessary for max- GENES & DEVELOPMENT 419 Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Lai et al. in vitro (Fig. 7). Using equal amounts of in vitro-trans- (2 y Full d7 d8 lated protein, determined by [35S]methionine incorpora- tion, we performed gel mobility-shift assays with labeled -84kD -84kD probes containing either the strong site, oligo S ( - 111 to - 57 kD - 57 kD -85) or the weak site, oligo W (-151 to -130). The - 48.5 kD - 48.5 kD amount of oligo S probe bound to HNF-3c~ or to HNF-3~/ m Qo was greater than the amount bound to HNF-3f~ (Fig. 7, - 365 kD .... - 36.5 kD lanes 1-6}. In contrast, with the oligo W probe, the amount complexed with HNF-3a or HNF-3~/ was less than that bound to HNF-3B (Fig. 7, lanes 7-12). Thus, HNF-3~ and HNF-3v have a higher affinity for the strong site than does HNF-3f~, whereas HNF-3~ has a higher Laver affinity for the weak site than either HNF-3a or HNF-3~/. - ÷ - + - + - + - + ,o This difference was most striking between HNF-3cx and HNF-3B. While HNF-3c, has a much higher affinity for the strong site compared to the weak site, HNF-3B has similar affinities for the two sites. To illustrate these differences more specifically, com- petition experiments with both oligonucleotides were carried out. Using labeled oligo S probe, competition with 10-fold molar excess of unlabeled oligo S reduced binding by -90% for HNF-3a (Fig. 8A, lane 2). The same amount of competitor only reduced the binding of HNF-3B to labeled probe by -50% (Fig. 8A, lane 8). HNF-3a complex formation was not significantly re- duced in the presence of a 40-fold molar excess of oligo W and a 200-fold excess of oligo W only decreased the HNF-3a complex by -50% (Fig. 8A, lanes 5 and 6). In I 2 3 4 5 6 7 8 I0 9 contrast, HNF-3B complex formation was reduced at Figure 3. The HNF-3B and HNF-37 cDNAs encode sequence- least 75% with a 40-fold of excess oligo W and abolished specific DNA-binding proteins. {A1 Autoradiograph of the trans- by 200-fold excess competitor (Fig. 8A, lanes 11 and 12). lated products of the HNF-3~, HNF-3[3, and HNF-37 cDNA an- Competition studies with weak-site oligo W as the alyzed by SDS-PAGE on a 9% acrylamide gel. (B) Confirmation probe are consistent with the results using the strong- of site of translation initiation in the HNF-3~ cDNA. Autorad- site oligo S as probe. A labeled oligo W probe in the iograph of translation products of full-length template and tem- presence of a 10-fold excess of unlabeled oligo S effec- plates shortened by 5' deletions of L11/35, d7, and d8. (C) Gel mobility-shift assay of the translated products compared to liver nuclear extract. Each of the translated proteins binds to the HNF-3 oligonucleotide (- 111 to -85 of the TTR promoter) in HNF-3[~ HNF-3c~ the presence of a nonspecific oligonucleotide ( - 175 to - 151 of the TTR promoter, indicated by the ( - in ) lanes 1, 3, 5, 7, and B E t [ I' S X B Iq E P S X 9) and is competed specifically by unlabeled HNF-3 oligonucle- otide [indicated by the + in lanes 2, 4, 6, 8 and I0]. The three complexes, A, B, and C, observed with liver extract are indi- U ,, N cated. Two micrograms of crude liver nuclear extract or 1 Ixl of m .... reticulocyte lysate, containing the translation products of HNF-3a, HNF-313, or HNF-3% was assayed. U imal activity of transfected DNA constructs and both of which are occupied in vivo by protein, presumably one of , ! the HNF-3 family members, as detected by a new in vivo footprint method (Mirkovitch and Darnell 1991 ). The to- tal set of binding proteins present in liver cell extracts, however, exhibits different affinities for the two DNA sites. A strong site is located between -96 and -106, 1 2 3 4 5 6 7 8 9 10 11 12 and a weak site is located between - 131 and - 140 nu- cleotides upstream of the cap site (Costa et al. 1989). In Figure 4. Southern hybridization of rat genomic DNA with an effort to gain insight into which proteins might actu- specific probes to HNF-3cx and HNF-3B. Digestion of genomic ally function in vivo, we compared the relative affinities DNA was performed with restriction endonucleases BamHI (B}, of each of the three HNF-3 proteins for each binding site EcoRI (E), HindlII {HI, PstI {P), SstI (SI, and XhoI (X). 420 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press HNF-3 gene family Comp - + - Antiserum Pre Pre o~ I Pre Pre a Antiserum Pre ~ Pre a Pre T--~ * T-" ' ~:? • Figure 5. The HNF-3~ protein is closely re- lated to the HNF-3C protein. (A) Anti-~ serum ..... eO .... reacts with the HNF-3c~ protein (lane 2) and O ~a~P* weakly with the HNF-3B protein (lane 4). Anti-v serum disrupts binding to DNA of the HNF-3v protein (lane 6). Preimmune or im- mune serum (1 wl) was added to the standard gel mobility-shift reactions containing reticu- locyte lysate-translated protein. (B) Anti-~ se- rum specifically reacts with the A and B (weakly) complexes formed with liver nuclear extract. The C complex is not affected (lanes 3 and 7). The anti-y serum specifically disrupts the C complex formed with liver extracts (lane 4) and with purified HNF-3C (lane 8). The re- actions were performed in the presence of a nonspecific competitor (-) or a specific com- 1 2 3 4 I 5 6 7 petitor (+) oligonucleotide (see legend to Fig. ,o213 41s 6 Liver Ext J HNF-3C 3). tively abolishes complex formation with HNF-3c~ but sion of the reporter gene compared to the control expres- not with HNF-3B (Fig. 8B, lanes 5 and 11). Furthermore, sion construct, which in this experiment contained a a 200-fold oligo W excess is able to compete effectively partial HNF-3~ cDNA in the antisense orientation. Sim- for binding to HNF-3~ but not to HNF-3o~ (Fig. 8B, lanes ilar results were seen with the HNF-3~ expression con- 9 and 3). Thus, it is clear that HNF-3e~ has a higher af- struct (data not shown}. In addition, the expression of a finity for oligo S than HNF-3B, and the reverse is true for reporter gene with multiple HNF-3 sites ligated to a het- oligo W. Since both of those sites are occupied in mouse erologous promoter was stimulated by cotransfection with HNF-3 expression constructs, whereas the reporter liver it seems likely that HNF-3~ is bound to the - 111 to - 85 site (oligo S}, whereas HNF-3~ is bound upstream construct lacking the multiple HNF-3 sites did not re- at - 151 to - 130. Because the DNA-binding properties spond (not shown). Taken together, the data on cotrans- of HNF-3~ and HNF-3~ are more similar to each other fection of plasmids producing HNF-3 and bearing HNF-3 than to those of HNF-3f~, it is possible that HNF-3oL and binding sites clearly indicate that each of the three pro- HNF-3v are interchangeable. teins, HNF-3a, HNF-3[3, and HNF-3% can act as positive activators of transcription. HNF-3 family members are transcriptional activators HNF-3 family members are differentially expressed Because early studies showed that mutation of the HNF- 3 strong site markedly reduced transcription at the TTR In our initial report of the cloning of an HNF-3 cDNA, promoter (Costa et al. 1989), it was expected that this the Northem analysis had shown two prominent mRNA site would be bound by transcriptional activators. With bands at 2.0 and 2.2 kb, as well as a minor band at 3.4 kb the identification of a family of proteins, however, we (Lai et al. 1990). In those experiments, we used a probe wanted to test which of the proteins were positive acti- that we now know spanned the homologous region of vators of transcription. This was accomplished by the three genes HNF-3~, HNF-3[3, and HNF-3% Thus, we presumably detected all three of the mRNAs that we cotransfection of expression vectors containing the cDNA for each of the HNF-3 family members and a re- now know to exist. Using probes that are specific to each porter construct with the TTR promoter (-202 to + 9) HNF-3 cDNA, we repeated the Northern analyses and containing the two HNF-3-binding sites described above found that the HNF-3cx transcript is 3.4 kb, the HNF-3[3 fused to the chloramphenicol acetyltransferase (CAT) transcript is 2.2 kb, and the HNF-3~ transcript is 2.0 kb gene. Figure 9 shows that the cotransfection into HepG2 {Fig. 10AI. These studies also show that in the liver cells of either HNF-3oL or HNF-3~ stimulated the expres- HNF-3[3 and HNF-3~/mRNAs were both more abundant GENES & DEVELOPMENT 421 Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Lai et al. 1o 20 3O 40 HNF-3e S YPL~HAKP P Y S Y I S L I TMA I QQAP[S KMLT LS E I YQW I MD L F P YY RQNQQ R Fork head T[TYRRSYT HAKP P YS Y I SL I TMA I Q T M L T L S E I Y Q~I M D L F Y R Q N Q Q R 70 80 ~ I00 110 HNF-3y HNF-3a NSIRHSLSFN~CF KVARSPDKPGKGSYWTLHPDSGNMFENGCYLRRQKRFK NSIRHSLSFNDCF K R PDKPGKGS WTLHPDSGNMFENGCYLRRQKRFKCE I HNF-3~ NSIRHSLSFNDCFVK]IPL~_RJT[PDKPGKGS WTLHPDSGNMFENGCYLRRQKRFK~L~ Fork head HNF-3~ IF N H P F S I N N L M S S S HNF-3a HNF-3~ F~__~H P F S I N N L M S S HNF-3~ S Y Q[S L]G GT T L Fork head $ S[H P F $ I N~L P T E Fork head I 1I llI .~F-3~ I 77 1 1186 ~ HNF- 3Or. l 1601 i ~ H I~ HNF-3~ [ 1~1 ]238 ~ [~ Fork head I 2001 13°9 ~ Figure 6. Homology of amino acid sequence among HNF-3e~, HNF-3B, HNF-37, and the Drosophila nuclear protein fork head. Sequence comparison of the DNA-binding domain (I) and two short regions at the carboxyl end of the proteins (II and III). Shaded areas show sequence identity between at least two of the proteins. (Bottom) A schematic diagram of the proteins, showing the location of the homologous regions. The positions of residues at both ends of region I are indicated. than HNF-3oL mRNA. The stronger gel mobility-shift ac- each of the family members independently because the tivity of HNF-3A (HNF-3~) with crude liver extracts (Fig. transcription rate is low and the cDNA-derived probes 5A; Lai et al. 1990) probably reflects the higher affinity of necessary to separate signals from the HNF-3a, HNF-3B, HNF-3a for the binding site (Figs. 7 and 8) rather than a and HNF-3~/ genes are too short. Only when genomic larger amount of HNF-3oL protein. clones are available will we be able to determine Using gene-specific probes, we confirmed by Northern whether different tissue distributions reflect proportion- ally different transcription rates. However, no transcrip- analysis that all three mRNAs were expressed in liver tion signal was obtained in kidney or brain, whereas a (Fig. 10A) but were absent in brain, spleen (data not clear signal was obtained in liver and lung using the full- shown), and kidney (Fig. 10B). There was a small but length probe for HNF-3a (Lai et al. 1990; K.G. Xantho- definite amount of expression in the small intestine of all three mRNAs (Fig. 10A). Longer exposures of the polous et al., in prep.). Therefore, the absence of any of blots of Figure 10A showed definite signals for HNF-3o~ the HNF-3 mRNAs in a tissue likely indicates no tran- scription of that gene in that tissue. and HNF-3f~ mRNAs. A more extensive tissue survey revealed additional in- formation about differential expression. HNF-3B was ex- Discussion pressed in the lung at high levels, approximately twofold greater than in liver (Fig. 10A, B). In contrast, HNF-3~/ We originally identified HNF-3oL as a protein that was mRNA was absent in the lung but present in the testis important in the transcriptional activation of two genes, (Fig. 10B). The HNF-3~/-like mRNA in the testis is dis- TTR and ~l-antitrypsin, which are expressed mainly in tinctly larger than that found in the liver. These differ- hepatocytes (Costa et al. 1989). Our current studies show ences were not due to differences in the amount of intact that HNF-3~ is only one member of a gene family that mRNA, as shown by hybridization with a probe to the encodes at least three transcription factors in rats. Se- GAPDH mRNA, which is present in all tissues (Fig. quence comparison of the three proteins reveals a highly 10B). The distribution of HNF-&x was identical to that of conserved 110-amino-acid domain, which lies within the HNF-3f~ for all of the tissues examined to date. region defined previously as neccessary for DNA bind- We have not been able to score the transcription rate of ing, and all three proteins bind the same DNA sequence, 422 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press HNF-3 gene family homeotic gene fork head product. Since all three rat pro- Oligo W Probe Oligo S teins and fork head conserved the entire 110 amino acids (-151/-130) (-111/-85) of the DNA-binding domain, it seems likely that all of HNF-3 (Z ~ ~/ Protein these amino acids may be required for DNA binding. For -- ÷ -- 4- -- 4- example, the DNA-binding domain in homeo box genes Competitor - ÷ - + - ÷ is conserved in vertebrates and invertebrates over a 60- amino-acid region that is required and sufficient for DNA binding (Gehring 1987). Furthermore, in the Pou y--. homeo gene family, which has members identified in Drosophila, Caenorhabditis elegans, and humans, an ~50-amino-acid conserved domain (the Pou region) has greater identity than the homeo box region and is re- quired for DNA binding (Herr et al. 1988). The other carboxy-terminal conserved regions of the HNF-3 fam- ily, which are not required for DNA binding, may con- ceivably be related to activating functions and either ac- tivation or repression for the fork head gene. The DNA-binding domain does not resemble that of any previously identified protein and thus represents a novel DNA-binding motif. At present, we can only spec- 1 2 3 4 5 6 7 8 10 9 11 12 ulate on the nature of the HNF-3c~ binding. The proteins Figure 7. Relative affinity of HNF-3 proteins for two binding give no evidence of dimerization. For example, cotrans- sites in the TTR promoter. Gel mobility-shift assays with trans- lation of full-length and truncated versions of HNF-3a or lated proteins using the strong-site oligo S (lanes 1-6) or the mixtures of HNF-3e, HNF-313, or HNF-3~/proteins does weak-site oligo W (lanes 7-12} as labeled probe. The reactions not yield complexes of intermediate mobility compared were performed in the presence of a nonspecific competitor or a to the mobility of each complex alone. Within the 110- specific competitor (oligo S for lanes 2, 4, and 6; oligo W for amino-acid conserved region, there is a cluster of basic lanes 8, 10, and 12). Approximately equal amounts of translated amino acids near the carboxyl end (RRQKREK). Such HNF-3 protein were used in all assays as estimated from the regions are known to be required for DNA binding of amount of [3SS]methionine incorporated. This experiment is helix-loop--helix and leucine zipper proteins and are representative of several using products from three separate thought to represent DNA contact points, whereas the translation reactions for each HNF-3 protein. helical portions of these proteins direct dimerization (Kouzarides and Ziff 1988; Murre et al. 1989; Turner and albeit with different affinities. This conservation in the Tjian 1989). There is only one fairly long region (from DNA-binding region is striking compared to the varia- Y/F at position 47 to V at position 72) within the l l0- amino-acid conserved region of the HNF-3 family that tion in the remainder of the coding region of these pro- has characteristics associated with helix formation. Ex- teins. For example, HNF-3~/is ~ 140 amino acids shorter cept for this region, prolines and glycines, which are con- than HNF-3c~, and HNF-313 and has only two short sidered helix breakers, are liberally scattered through the stretches of conserved sequence (II and III in Fig. 6) in remainder of region I. Such an amino acid distribution addition to the DNA-binding domain. The strong con- probably does not rule out helix formation but also does servation of the DNA-binding and the two carboxy-ter- minal regions (II and III) in the three rat genes is mirrored not allow speculation of its existence. The actual struc- by conservation of the same regions in the Drosophila ture of the HNF-3 class of proteins is perhaps one that Figure 8. Competition of HNF-3et and HNF-313 Probe Oligo S (-111/-85) B Probe Oligo W (-151/-130) binding to two binding sites in the TTR pro- moter. (A) Binding to the strong-site oligo S of translated HNF-3a (lanes 1--6) or HNF-313 (lanes Mo, r 7-12) in the presence of varying amounts of un- 0 40 200 2 10 40 0 40 2001 2 I0 40 Excess 0 10 40~10 40 200~ 0 10 40 10 40 200 Excess labeled competitor oligo S (lanes 2, 3, 8, and 9) or oligo W (lanes 4--6 and 10--12). (B} Binding to the weak-site oligo W of translated HNF-3a (lanes 1-6) or HNF-3[3 (lanes 7-12) in the presence of ~ii ~i ~ ~i~i~ varying amounts of unlabeled competitor oligo W (lanes 2, 3, 8, and 9) or oligo S (lanes 4--6 and 10--121. Approximately equal amounts of trans- lated HNF-3 protein were used in all assays as I1 2 3 4 5 6117 8 9 10 11 121 I l 2 3 4 6117 ~ 8 9 10 I1 121 HNF-3 (~ A HNF-3 estimated from the amount of [3SS]methionine Protein I° Protein incorporated. GENES & DEVELOPMENT 423 ~i~!!!i: Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Lai et al. form the terminal embryonic structures of the larval em- Expression Control [ HNF-3ot [ HNF-3[~ Vector bryo that contribute to such tissues as the salivary glands, foregut, and hindgut (Hartenstein et al. 1985; Jfir- gens and Weigel 1988; Weigel et al. 1989). It is logical to "T TT propose that this family of genes has maintained a role from early in evolution in forming gut-related organs• We came to the study of HNF-3 by its apparent regula- tory action on the TTR gene and a~-antitrypsin gene in adult liver. Whether one of the presently characterized HNF-3 proteins or some other member of this family has a role to play in earlier steps of hepatocyte determination YYTTTT now seems an important question. Furthermore, we can ask what the function may be of 1 2 3 4 5 6 different proteins of similar but not identical binding Figure 9. Transcriptional activation by cotransfection in affinities such as the HNF-3 family in adult rodent tis- HepG2 cells. Autoradiograph of TLC plate measuring activity of sues. First, it is obvious that different HNF-3 family CAT produced from a reporter construct TTR-CAT when members have widely different affinities for different cotransfected with a control expression plasmid (lanes 1 and 2) binding sites on DNA. Thus, different HNF-3 proteins or a plasmid expressing HNF-3a (lanes 3 and 4) or HNF-313 (lanes may be required to act on various sites in different genes 5 and 6). or even in the same gene. From our analysis of cell dis- tribution, it is clear that different tissues may vary in amounts of the different HNF-3 proteins, with HNF-3~/ allows the protein to bind to DNA as a monomer, thus being absent from lung, for example. All cells deriving presenting several unique challenges to the crystallogra- from endoderm may share some HNF-3-1ike proteins, pher. but some family members could be more restricted in What can we say of the broader biologic role of the expression and serve more local functions than others. Further tissue analyses and genomic searches for more HNF-3 family of proteins? The identity of the DNA- binding region of HNF-3a, HNF-3B, and HNF-3~/ with HNF-3 family members are needed to illuminate these that of the homeotic gene fork head product suggests questions. that they comprise a third group of mammalian and in- Finally, we must search more widely in adults and vertebrate genes that encode proteins having both simi- embryos for members of this family to attempt to ex- lar DNA-binding regions and at least somewhat similar plore the hypothesis about the restricted cell lineage in determining roles in early development. One group, the which this gene family is active. For example, fork head homeo box genes that control the proper segment forma- is present in brain cells in Drosophila, and we did not tion in the early Drosophila embryo, have a counterpart in vertebrates in Hox genes that have the homeo box sequence (Akam 1989). Because of the segmental expres- sion of Hox genes, these genes are expected to participate in proper segmental division of vertebrate embryos {Gra- ham et al. 1989). A second group encodes the helix-loop- • 28S helix proteins, one of which, myoD, has a demonstrated role in muscle cell differentiation when overexpressed in IINF-3~-- ° I8S certain susceptible cells and is a normal product during myogenesis. Another helix-loop-helix protein, E47, is 28S thought to play a role in lymphocyte differentiation. -28S Drosophila homologs of the helix-loop-helix genes in- HNF-3T -~" ~. 18S clude daughterless (Caudy et al. 1988) and achaete-scute '0 HNF-3~-~ ~ {Cabrera et al. 1987; Villares and Cabrera 1987); daugh- - 18S terless affects sex determination and the formation of ° 28S sensory precursor cells, whereas achaete-scute operates - 28S in cuticle formation. Thus, this second group of genes .18S 4m' may be devoted mainly to cell specialization. CAPDH-p ~ ~ ~ aJl HNF-3y"~ We have shown here that the HNF-3 family in rodents - 18S is expressed in cell types {liver, lung, and intestine) that derive from the embryonic gut tube by the outpouching Figure 10. Northern analysis of mRNA levels for each HNF-3 of endodermal cells, which on correct contact with mes- gene. (A) Specific probes distinguish the major mRNA species enchyme, form structures such as salivary glands, lungs, for HNF-3a, HNF-313, and HNF-3~ in the liver. The blot for pancreas, liver, and small and large intestine. It hardly HNF-3a was exposed for 72 hr; the blots for HNF-313 and seems accidental that the Drosophila protein homolog of HNF-3-y were exposed for 24 hr each. (B) HNF-313 and HNF-3~/ the HNF-3 family, fork head, is required to correctly mRNA expression in several tissues. 424 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press HNF-3 gene family observe HNF-3 mRNAs in total brain tissue. However, teine residue was added to facilitate coupling to keyhole limpet hemocyanin. Injections of 100 ~.g of coupled peptide with we have not yet examined individual cell types with in Freund's adjuvant were made subcutaneously at 4-week inter- situ hybridization or with specific antisera. We already vals. High-titer antiserum was obtained after three injections. know that we only detected CAAT enhancer binding Anti-~ serum was prepared by immunization of two rabbits protein (C/EBP), another transcription protein present in with a fusion peptide produced in Escherichia coll. The XmnI- liver, by in situ hybridization in brain cells (Kuo et al. SstI fragment of the HNF-3~I cDNA was inserted in the T7 poly- 1990). Thus, a search for region-specific expression of merase expression plasmid pET-3a (Rosenberg et al. 1987), HNF-3 in the brain is strongly indicated. which yielded an insoluble fusion peptide containing 77 amino acids of the HNF-3~ protein that was distinct from HNF-3~ or HNF-3~. Bacterial cells were sonicated in 50 mM Tris-HC1 {pH Materials and methods 7.5), 2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, and 0.1 M NaC1. Insoluble protein was recovered by centrifugation Isolation of HNF-3fl and HNF-37 cDNA at 12,000g for 10 min. The pellet was resuspended in sonication A random primer-labeled probe was prepared from the H3 buffer plus 1% SDS and centrifuged for 10 min. SDS-PAGE anal- cDNA (HNF-3a) and used to screen a mouse liver eDNA library ysis of the supematant showed that it was comprised of 50% of (generously provided by Dr. K.E. Paulson). Two clones were a single polypeptide of 10 kD. This preparation {200 ~g) was isolated that proved to be homologous to the rat cDNA. The used for each injection. longer clone, M2, was sequenced and found to encode a protein of 329 amino acids that was missing the amino-terminal end. A Preparation of probes specific to HNF-3a, HNF-3fl, and highly homologous region of - 100 amino acids was followed by HNF-37 a divergent sequence. A 270-bp fragment from this region was used to prepare a probe for screening a rat liver cDNA library Oligonucleotide primers were synthesized corresponding to the (Stratagene). One clone was identified, K2 (HNF-3f~I, which was coding strand of nucleotides 808-827 and the noncoding strand homologous to the M2 clone. of nucleotides 1252-1233 from the HNF-3c~ sequence. These The HNF-3~/ eDNA was obtained by screening the rat liver sequences are identical in the HNF-3~ eDNA, except for one cDNA library with a random primer-labeled fragment of mismatch, and flank an -300-nucleotide nonhomologous re- HNF-3a. Twenty positive clones were isolated and excised as gion. These primers were used in the polymerase chain reaction plasmids (kZAP protocol, Stratagene). Southern hybridization with plasmids containing HNF-3a and HNF-3B cDNA as tem- to the plasmid DNA digested with EcoRI was performed with plates to prepare fragments that are specific to each gene, except probes specific to HNF-3a and HNF-3~ cDNA. Seven clones did for the primer sequences. The corresponding region of the not hybridize, and six were found by restriction mapping and HNF-3~ cDNA could be obtained by isolating the 234-bp frag- partial sequence determination to be overlapping fragments of a ment generated by restriction enzymes XmnI and SstI. Probes distinct cDNA. Two of these clones were ligated at a shared were prepared by random primer labeling of these fragments NcoI site to create Lll/35, which encodes the full-length after denaturation. HNF-3-y protein sequence. Sequence was performed by the dideoxy chain-termination method on double-stranded tem- Transfection constructs and assays plates. The K2 and L I 1/35 sequences were determined com- pletely on both strands with overlapping subcloned and exonu- The vectors used for the expression of HNF-3 cDNAs were de- clease-shortened templates, as well as with ITP-containing se- scribed previously (Chen et al. 1987). The H3 eDNA was used quencing mixes (U.S. Biochemical Sequenase Kit) to eliminate for HNF-3ct, the K2 cDNA was used for HNF-313, and the L11135 compression artifacts. cDNA was used for HNF-3~/. The control construct contained the murine HNF-3~ partial cDNA, M2, in the antisense orien- tation. Each of these was cloned into the 5' XbaI and 3' HindIII In vitro translation and gel mobility-shift assays sites of the expression vector. The XbaI and HindIII sites of each Linearized plasmids were used as templates for transcription cDNA were generated by cloning each EcoRI fragment into with T7 or T3 RNA polymerase (Melton et al. 1984). In vitro pBluescript KS in the appropriate orientation. The reporter con- translation was performed by use of nuclease-treated rabbit re- struct TTR-CAT was made by purifying the mouse TTR pro- ticulocyte lysate (Promega) according to the manufacturer's pro- motet sequence -202 to + 9 of the TTR minigene (Costa et al. tocol. [3SSlMethionine (translation grade) was purchased from 1986), adding HindIII linkers, and fusing to the HindIII site of New England Nuclear. The reticulocyte lysate containing the the CAT gene (Lew et al. 1991) without any other promoter translated protein was used directly in gel mobility-shift assays sequence. as described previously (Kovesdi et al. 1986) with end-labeled Transfection of human HepG2 cells was carried out by cal- double-stranded oligonucleotide probes. For the comparison of cium phosphate precipitation IWigler et al. 1979), except that DNA-binding affinities, the relative amount of translated pro- 5 mM sodium butyrate was used after glycerol shock. tein was estimated by SDS-PAGE and autoradiography of the pCMV-B-gal (MacGregor and Caskey 1989) was used as an in- dried gel. The relative amount of label incorporated was quan- ternal control for transfection efficiency. For each transfection, titated by densitometry and then adjusted for the number of 1 ~g of pCMV-~-gal, 2 ~g of TTR-CAT, and 10 ~g of each methionine residues in each HNF-3 protein being translated to expression vector were added to 5 x 106 HepG2 cells in 100- arrive at an estimate of the relative amount of protein. mm dishes. CAT assays and B-gal normalizations were carried out as described (Lew et al. 1991). Preparation of antisera to HNF-3 proteins RNA analysis Anti-a was obtained by immunization of two rabbits with a synthetic peptide MLGTVKMEGHESNDWC, corresponding to RNA was isolated from rat tissues by the acid guanidinium the 15 amino-terminal residues of HNF-3(~. The terminal cys- thiocyanate/phenol-chloroform extraction method (Chomczyn- GENES & DEVELOPMENT 425 Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Lai et al. ski and Sacchi 1987; Puissant and Houdebine 1990). Poly(A) + 6593. Frain, M., G. Swart, P. Monaci, A. Nicosia, S. Stampfli, R. Frank, RNA was selected by chromatography on oligo(dT)-cellulose. and R. Cortese. 1989. The liver-specific transcription factor Two micrograms of poly{A) + RNA from each tissue was sepa- LF-B1 contains a highly diverged homeobox DNA binding rated on a formaldehyde-l.2% agarose gel and transferred to domain. Ce11 59: 145-157. Zeta-probe nylon membranes (Bio-Radl by capillary action. Hy- Gehring, W.l. 1987. Homeo boxes in the study of development. bridization was performed in 0.5 M sodium phosphate (pH 7), 1 mM EDTA, and 7% SDS at 65°C for 18-24 hr. The high-strin- Science 236: 1235-1252. Graham, A., N. Papalopulu, and R. Krumlauf. 1989. The murine gency wash was performed in 0.1% SSC/0.1% SDS at 65°C for 10 rain. The blot was exposed to X-ray film at - 80°C with two and Drosophila homeobox gene complexes have common intensifying screens for 24-72 hr. features of organization and expression. Cell 57: 367-378. Hartenstein, V., G.M. Technau, and ].A. Campos-Orterga. 1985. Fate-mapping in wild type Drosophila melanogaster. Acknowledgments Wilhelm Roux's Arch. Dev. Biol. 194: 181-195. Herr, W., R.A. Sturm, R.G. Clerc, L.M. Corcoran, D. Baltimore, We thank Gabriela Balas and Oksana Litvin for excellent tech- P.A. Sharp, H.A. Ingraham, M.G. Rosenfeld, M. Finney, G. nical assistance, and C.F. Kuo, K.E. Paulson and F.M. Sladek for Ruvkun, and H.R. Horvitz. 1988. The POU domain: A large helpful discussions. This work was funded in part by National conserved region in the mammalian pit-l, oct-l, oct-2, and Institutes of Health grant CA16006-16A to I.E.D., an NRSA Caenorhabditis elegans unc-86 gene products. Genes & training grant (GM07982-09) to V.R.P., and a Merck Fellowship Dev. 2: 1513-1516. to W.S.C. Johnson, P.F., W.H. Landschulz, B.J. Graves, and S. McKnight. The publication costs of this article were defrayed in part by 1987. Identification of a rat liver protein that binds to the payment of page charges. This article must therefore be hereby enhancer core element of three animal viruses. Genes & marked "advertisement" in accordance with 18 USC section Dev. 1: 133-146. 1734 solely to indicate this fact. lfirgens, G. and D. Weigel. 1988. Terminal versus segmental development in the Drosophila embryo: The role of the ho- meotic gene fork head. Wilhelm Roux's Arch. Dev. Biol. Note added in proof 197: 345-354. Sequence data described in this paper have been submitted to the Kouzarides, T. and E. Ziff. 1988. The role of the leucine zipper EMBL/GenBank Data Libraries. in the fos-jun interaction. Nature 336: 646-651. Kovesdi, I., R. Reichel, and I.R. Nevins. 1986. Identification of a cellular transcription factor involved in E1A trans-activa- References tion. Cell 45: 219-228. Akam, M. 1989. Hox and HOM: Homologous gene clusters in Kozak, M. 1986. Point mutations define a sequence flanking the AUG initiator codon that modulates translation by eukary- insects and vertebrates. Cell 57: 347-349. Baumhueter, S., D.B. Mendel, P.B. Conley, C.I. Kuo, C. Turk, otic ribosomes. Cell 44: 283-292. M.K. Graves, C.A. Edwards, G. Courtois, and G.R. Crabtree. Kuo, C.F., K.G. Xanthopoulos, and I.E. Damell lr. 1990. Fetal 1990. HNF-1 shares three sequence motifs with the POU and adult localization of C/EBP: Evidence for combinatorial domain proteins and is identical to LF-B1 and APF. Genes & action of transcription factors in cell-specific gene expres- sion. Development 109:473-481. Dev. 4: 372-379. Cabrera, C.V., A. Martinez-Arias, and M. Bate. 1987. The ex- Lai, E., V.R. Prezioso, E. Smith, O. Litvin, R.H. Costa, and I.E. Darnell Jr. 1990. HNF-3A, A hepatocyte-enriched transcrip- pression of three members of the achaete-scute gene com- tion factor of novel structure is regulated transcriptionally. plex correlates with neuroblast segregation in Drosophila. Genes & Dev. 4: 1427-1436. Cell 50: 425-433. Lew, D.I., T. Decker, I. Strehlow, and J.E. Damell. 1991. Over- Caudy, M., H. Vassin, M. Brand, R. Tuma, L.Y. Jan, and Y.N. lapping elements in the guanylate-binding protein gene pro- Jan. 1988. daughterless, a Drosophila gene essential for both neurogenesis and sex determination. Cell 55: 1061-1067. moter mediate transcriptional induction by alpha and Chen, W.S., C.S. Lazar, M. Poenie, R.Y. Tsien, G.N. Gill, and gamma interferons. Mol. Cell Biol. 11: (in press). M.G. Rosenfeld. 1987. Requirement for intrinsic protein ty- MacGregor, G.R. and C.T. Caskey. 1989. Constructions of plas- rosine kinase in the immediate and late actions of the EGF mids that express E. coli 13-galactosidase in mammalian receptor. Nature 328: 820-823. cells. Nucleic Acids Res. 17: 2365. Chomczynski, P. and N. Sacchi. 1987. Single-step method of Melton, D.A., P.A. Krug, M.R. Rebaglicti, T. Maniatis, K. Zinn, RNA isolation by acid guanidinium thiocyanate- and M.R. Green. 1984. Efficient in vitro synthesis of biolog- ically active RNA and RNA hybridization probes from plas- phenol-chloroform extraction. Anal Biochem. 162: 156- 159. raids containing a bacteriophage SP6 promoter. Nucleic Ac- ids Res. 12: 7035-7056. Costa, R.H., E. Lai, and J.E. DameI1 Jr. 1986. Transcriptional Mirkovitch, J. and J.E. Damell Jr. 1991. Rapid in vivo footprint- control of the mouse prealbumin (transthyretin) gene: Both ing technique identifies proteins bound to the TTR gene in promoter sequences and a distinct enhancer are cell specific. liver. Genes & Dev. 5: 83-93. MoI. Cell. Biol. 6: 4697-4708. Murre, C., P.S. McCaw, and D. Baltimore. 1989. A new DNA Costa, R.H., D.R. Grayson, and I.E. Darnell lr. 1989. Multiple hepatocyte-specfic nuclear factors function in the regulation binding and dimerization motif in immunoglobulin en- of the transthyretin and al-antitrypsin genes. Mol. Cell. hancer binding, daughterless, MyoD, and myc proteins. Cell Biol. 9: 1415-1425. 56: 777-783. Costa, R.H., T. Van Dyke, C. Yan, C.F. Kuo, and I.E. Damell Jr. Puissant, C. and L. Houdebine. 1990. An improvement of the 1990. Similarities in transthyretin gene expression and dif- single-step method of RNA isolation by acid guanidinium ferences in transcription factors: Liver and yolk sac com- thiocyanate-phenol-chloroform extraction. BioTechniques pared to choroid plexus. Proc. Natl. Acad. Sci. 87: 6589- 8: 148-149. 426 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press HNF-3 gene family Rosenberg, A.H., B.N. Lade, D. Chui, S. Lin, I.J. Dunn, and F.W. Studier. 1987. Vectors for selective expression of cloned cD- NAs by T7 RNA polymerase. Gene 56: 125-135. Sladek, F.M., W. Zhong, E. Lai, and J.E. Damell Jr. 1990. Liver- enriched transcription factor HNF-4 is a novel member of the steroid hormone receptor superfamily. Genes & Dev. 4: 2353-2365. Tilghman, S.M. and A. Belayew. 1982. Transcriptional control of the murine albumin/a-fetoprotein locus during develop- ment. Proc. Natl. Acad. Sci. 79: 5254-5257. Turner, R. and R. Tjian. 1989. Leucine repeats and an adjacent DNA binding domain mediate the formation of functional cFos-cJun heterodimers. Science 243: 1689-1694. Villares, R. and C.V. Cabrera. 1987. The achaete-scute gene complex of D. melanogaster: Conserved domains in a subset of genes required for neurogenesis and their homology to myc. Cell 50: 415-424. Weigel, D. and H. J~ckle. 1990. Fork head: A new eukaryotic DNA binding motif? Cell 63: 455-456. Weigel, D., G. Jfirgens, F. Kuttner, E. Seifert, and H. J~ickle. 1989. The homeotic gene fork head encodes a nuclear protein and is expressed in the terminal regions of the Drosophila em- bryo. Cell 57: 645--658. Wigler, M., A. Pellicer, S. Silverstein, R. Axel, G. Urlaub, and L. Chasin. 1979. DNA-mediated transfer of adenylphosphori- bosyl transferase locus into mammalian cells. Proc. Natl. Acad. Sci. 76: 1373-1376. Xanthopoulos, K.G., J. Mirkovitch, T. Decker, C.F. Kuo, and J.E. Darnell Jr. 1989. Cell-specific transcriptional control of the mouse DNA-binding protein C/EBP. Proc. Natl. Acad. Sci. 86:4111--4121. GENES & DEVELOPMENT 427 Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Hepatocyte nuclear factor 3 alpha belongs to a gene family in mammals that is homologous to the Drosophila homeotic gene fork head. E Lai, V R Prezioso, W F Tao, et al. Genes Dev. 1991, 5: Access the most recent version at doi:10.1101/gad.5.3.416 This article cites 35 articles, 13 of which can be accessed free at: References http://genesdev.cshlp.org/content/5/3/416.full.html#ref-list-1 License Receive free email alerts when new articles cite this article - sign up in the box at the top Email Alerting right corner of the article or click here. Service Copyright © Cold Spring Harbor Laboratory Press http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Genes & Development Unpaywall

Hepatocyte nuclear factor 3 alpha belongs to a gene family in mammals that is homologous to the Drosophila homeotic gene fork head.

Lai, E; Prezioso, V R; Tao, W F; Chen, W S; Darnell, J E
Genes & Development – Mar 1, 1991
Download PDF
13 pages

Loading next page...
 
/lp/unpaywall/hepatocyte-nuclear-factor-3-alpha-belongs-to-a-gene-family-in-mammals-HWrml0DzCD

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Unpaywall
ISSN
0890-9369
DOI
10.1101/gad.5.3.416
Publisher site
See Article on Publisher Site

Abstract

Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Hepatocyte nuclear factor belongs to a gene family in mammals that is homologous to the Drosophila homeotic gene fork head Eseng Lai, 1 Vincent R. Prezioso, 2 Wufan Tao, 1 William S. Chen, z and James E. Darnell, Jr. 2'3 IDivision of Endocrinology and Program of Cell Biology and Genetics, Memorial Sloan-Kettering Cancer Center, New York, New York 10021 USA~ ~Laboratory of Molecular Cell Biology, Rockefeller University, New York, New York 10021 USA By analysis of cDNA clones that cross-hybridized with a portion of the cDNA encoding the recently described rat protein hepatocyte nuclear factor 3a (HNF-3oq previously called HNF-3A), we now describe two additional members, HNF-313 and HNF-3~,, of this gene family. A 110-amino-acid region in the DNA-binding domain of this family is not only very highly conserved in rodents (HNF-3% -313, and -3~ are identical in 93 of 110 amino acids in this region) but also in Drosophila where the homeotic gene fork head has 88 of the 93 residues that are identical in the three rat genes. The HNF-3 family in rodents is expressed in cells that derive from the lining of the primitive gut; some of the embryonic Drosophila cells in which fork head is expressed also give rise to gut and salivary glands. Thus, it appears that this gene family, the DNA-binding portion of which is unlike that of any previously recognized DNA-binding proteins, may contribute to differentiation of cells in internal organs in both vertebrates and invertebrates. [Key Words: Transcriptional control; HNF-3 family; mammalian development] Received November 27, 1990; revised version accepted December 28, 1990. whose eDNA we cloned recently, is transcriptionally ac- Using the techniques of molecular genetics to identify required DNA-binding sites for gene function and then tive in liver but not in a number of other tissues. Liver protein purification by oligonucleotide affinity chroma- extracts, however, interact with a labeled oligonucle o tography of proteins cognate with these sites, several lab- otide containing an HNF-3 site to produce several bands oratories including our own have identified transcription in a gel mobility-retardation assay (Costa et al. 1989; Lai factors that direct expression of genes specifically in he- et al. 1990), and several mRNAs were detected in North- patocytes (Johnson et al. 1987; Frain et al. 1989; Baum- ern blots with the cloned sequence of the HNF-3ot eDNA hueter et al. 1990; Lai et al. 1990; Sladek et al. 1990). (Lai et al. 1990). These results most likely indicate either Hepatocyte-specific gene expression of such characteris- a single transcriptional product that is differentially pro- tic products as albumin begins quite early (-11 or 12 cessed or the transcription of several related but distinct days of gestation) in the mouse embryo (Tilghman and genes. Belayew 1982), indicating the probable early activity of Characterization of additional HNF-3-1ike eDNA from transcription factors that are also active in the adult rodent liver has now revealed that HNF-3c~ belongs to a liver. Furthermore, some of these factors, for example, gene family with a tightly conserved -ll0-amino-acid hepatocyte nuclear factors 1 and 3 (HNF-1 and HNF-3) sequence within the region shown earlier to be required appear to be present in yolk sac cells (Costa et al. 1990), for DNA binding. This conserved DNA-binding domain which, like the hepatocyte, are derived from gut endo- is unlike that of any previously recognized family of derm. Because the genes for some of the liver-enriched transcription factors. A recently reported sequence com- transcription factors mentioned above are transcription- parison further heightens the interest in the HNF-3 fam- ally controlled themselves (Xanthopoulos et al. 1989, ily as consisting of important contributing members to and in prep.), a study of their regulation should provide early development (Weigel and J~ickle 1990). The protein access to genes that are important in early developmen- encoded by the Drosophila gene fork head, mutants of tal cascades. which fail to make proper terminal embryonic tissues Hepatocyte nuclear factor 3~ (HNF-3~, previously (the proctodeum and stomodeum, which contribute cells called HNF-3A), one of the factors purified from liver and to the foregut and hindgut), has an extensive amino acid sequence homology to the HNF-3A sequence described. 3Corresponding author. The greatest sequence conservation is in the DNA-bind- 416 GENES & DEVELOPMENT 5:416-427 © 1991 by Cold Spring Harbor Laboratory ISSN 0890-9369/91 $1.00 Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press HNF-3 gene |amily ing domain of all the HNF-3 family members. The ho- hybridize to either one and therefore represented poten- meo box proteins that participate in Drosophila embry- tial additional family members. Six of these clones were onic larval segmentation are segmentally expressed; and, shown by restriction mapping and partial sequencing to presumably, similarly functioning homologs in mam- be overlapping fragments of a single cDNA that was dif- mals (Hox genes) are also segmentally expressed (Akam ferent from the two cDNA clones isolated previously. A 1989; Graham et al. 1989). By analogy, it seems likely cDNA containing the complete coding sequence of this that the HNF-3 family and fork head (as well as any third family member was reconstructed from two over- lapping clones, L11 and L35. The joint L11/L35 sequence other Drosophila genes like fork head) may function in contains an ORF of 322 amino acids or -140 amino acids constructing internal organs from the embryonic gut. Thus, there may be an evolutionarily conserved genetic shorter than the proteins encoded by the first two pathway operating in the differention of cells of internal cDNAs (Fig. 2). The position of the translation initiation organs of both vertebrates and invertebrates. site in this third sequence was verified by use of a series In addition to delineating three distinct members of of 5' deletions that were transcribed in vitro with SP6 or the HNF-3 family, we also show that (1) the mRNAs for T7 RNA polymerase and translated in reticulocyte ly- these proteins in rats are distributed differently in differ- sates (Fig. 3B). Deletion of 311 bp of the 5' end of the ent organs that derive from the primitive gut, and (2) cDNA did not alter the size of the translated protein (d7). An additional deletion of 84 nucleotides (d8) resulted in HNF-3 proteins are positive activating proteins. The a shortened translation product (presumably initiating at study of the HNF-3 family whose members are differen- the ATG at position 537-539), confirming that the major tially expressed in different cell types promises to be of translation initiation site must be one of the two tandem considerable value in opening new insights into early ATGs at position 312-314 and 315-317 (Fig. 2). mammalian development. We renamed these three related HNF-3 cDNAs, HNF-3e~ (the original HNF-3A), HNF-3[3 (K2), and Results HNF-37 (Lll/L35). The nucleic acid sequence compari- son of HNF-3a, HNF-3[3, and HNF-37 suggests that these Isolation of cDNA clones for HNF-3[3 and HNF-37 cDNAs are products of three separate but related genes, rather than the products of alternative splicing. For ex- Liver extracts exhibit multiple DNA-binding activities ample, in the regions of greatest homology, the amino for a labeled double-stranded deoxyoligonucleotide con- acid sequence is more highly conserved than the nucleic taining an HNF-3 site, and Northern analysis of liver acid sequence. Genomic Southern analyses using probes mRNA with labeled DNA of the original HNF-3A cDNA specific to each cDNA showed hybridization to a differ- clone shows multiple mRNAs (Lai et al. 1990). There- ent pattern of genomic DNA fragments, thus supporting fore, we suspected that there might be several related genes that encode transcription factors sharing binding the existence of three separate and likely single-copy specificity for HNF-3 sites in DNA. When we selected genes. Autoradiographs from Southern analysis of rat ge- and sequenced a murine cDNA that cross-hybridized nomic DNA for HNF-3a and HNF-3[3 are shown in Fig- with an HNF-3A probe, we isolated a partial cDNA ure 4. clone, suggesting that more than one HNF-3-1ike gene existed. This clone, M2, was highly similar in one region HNF-3[3 and HNF-37, like HNF-3a, encode of the two sequences but diverged markedly outside of sequence-specific DNA-binding proteins this region. We used a probe derived from the divergent RNA transcripts from the HNF-3a, HNF-313, and sequence to identify the rat counterpart of the mouse M2 HNF-3~/clones were produced and translated in reticu- cDNA and obtained a clone, termed K2, which was locyte lysates, and the products were analyzed on SDS- >90% identical to the entire mouse M2 sequence and polyacrylamide gels {Fig. 3AI. The monomer molecular distinctly different from the original HNF-3A (rat) gene in most of its sequence. The K2 rat cDNA contained a masses of HNF-3a, HNF-313, and HNF-3~, respectively, single large open reading frame (ORF) that would encode are 50, 47, and 42 kD. The product of HNF-3a has been a protein of 458 amino acids (Fig. 1 ). The selection of an shown previously to bind specifically to the proximal initiator methionine was suggested by an adjacent se- HNF-3 site in the transthyretin (TTR} promoter {Lai et quence that conformed well to the consensus for eukary- al. 19901. We now show that the products of both HNF-313 and HNF-37 also bind specifically to an oligo- otic initiation {Kozak 1986) and the presence of a termi- nucleotide representing the original HNF-3-binding site nation codon immediately upstream. Comparison of the {Fig. 3C). For comparison we show the complexes formed amino acid sequence with that of HNF-3A indicated that with crude liver extract. Three of these complexes have these two proteins shared a highly homologous central been referred to previously as A, B, and C in decreasing domain but differed in most of the coding region. To obtain additional members of this family, we re- size {or increasing gel mobility}. screened 500,000 recombinants of the rat liver cDNA Because we purified and sequenced fragments of the library, using a probe to the conserved sequence, and HNF-3A protein, we can definitively state that it is iden- isolated 20 positive clones. We then used probes derived tical to the product of the HNF-3a gene. For the products from nonhomologous sequences in the HNF-3A and the of HNF-313 and HNF-3~, indirect evidence indicates that they are identical or closely related to the HNF-3B and K2 cDNAs to identify 7 clones of the 20 that did not GENES & DEVELOPMENT 417 Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Lai et al. TCTCcCGGGTATTGGCTGcAGCTAAGcGGGGcTTCCTGGGCCGACTGAGGTGGGTAGCCAGAAAAAGGCCTGAGGTGGCTGAcAACcAGGGcGGCCAGACAACGCGAGTCCTGCGcGCCTTCTGAGGCCTc 131 CCTGGGACTTAACTGTAACGGGGAGGGGCCTCTGGAGCAGCGGCCAGCGAGTTAAAGT ATG CTG GGA GCC GTG AAG ATG GAA GGG CAC GAG CCA TCC GAC TGG AGC AGC TAC 243 M L G A V K M E G H E P S D W $ S Y 18 TAC GCG GAG CCT GAG GGC TAC TCT TCC GTG AGC AAC ATG AAC GCC AGC CTG GGG ATG AAT GGC ATG AAC ACT TAC ATG AGC ATG TCC GCG GCT GCA ATG 342 Y A E P E G Y S S V S N M N A S L G M N G M N T Y M S M S A A A M 51 GGC AGT GGT TCC GGC AAC ATG AGC GCA GGC TCC ATG AAC ATG TCA TCC TAT GTG GGC GCT GGA ATG AGC CCG TCG CTG GCT GGC ATG TCC CCG GGC GCG 441 G S G S G N M S A G S M N M S S Y V G A G M S P S L A G M S P G A 84 GGC GCC ATG GCG GGC ATG AGC GGC TCA GCT GGG GCG GCC GGC GTG GCG GGC ATG GGA CCG CAC CTG AGT CCG AGT CTG AGC CCA CTC GGG GGA CAG GCG 540 G A M A G M S G S A G A A G V A G M G P H L S P S L S P L G G Q A 117 GCC GGG GCT ATG GGT GGC CTT GCT CCC TAC GCC AAT ATG AAC TCC ATG AGT CCT ATG TAC GGG CAG GCG GGC CTG AGC CGC GCT CGG GAC CCC AAG ACG 639 A G A M G G L A P Y A N M N S M S P M Y G Q A G L S R A R D P K T 150 TAC CGG CGC AGC TAC ACT CAC GCC AAG CCT CCC TAC TCG TAC ATC TCG CTC ATC ACC ATG GCC ATC CAG CAG AGC CCC AAC AAG ATG CTG ACG CTG AGC 738 Y R R S Y T H A K P P Y S Y I S L I T M A I Q Q S P N K M L T L S 183 GAG ATC TAT CAG TGG ATC ATG GAC CTC TTC CCT TTC TAC CGG CAG AAC CAG CAG CGC TGG CAG AAC TCC ATC CGT CAT TCT CTC TCC TTC AAC GAC TGC 837 E I Y Q W I M D L F P F Y R Q N Q Q R W Q N S I R H S L S F N D C 216 TTT CTC AAG GTG CCC CGC GCG CCA GAC AAG CCT GGG GGC TCC TTC TGG ACC CTG CAC CCT GAC TCT GGC AAC ATG TTC GAG AAC GGT TGC TAC CTG CGC 936 F L K V P R A P D K P G G $ F W T L H P D S G N M F E N G C Y L R 249 CGC CAG AAG CGC TTC AAG TGT GAG AAG CAA CTG GCG TTG AAG GAA GCA GCG GGT GCG GGC AGT GGC GGA GGC AAG AAG ACC GCT CCT GGG ACA CAG GCT 1035 R Q K R F K C E K Q L A L K E A A G A G S G G G K K T A P G T Q A 282 TCT CAG GTT CAG CTC GGG GAG GCC GCA GGC TCG GCC TCT GAG ACT CCG GCG GGC ACC GAG TCC CCC CAT TCC AGC GCT TCT CCG TGT CAG GAG CAC AAG 1134 S Q V Q L G E A A G S A S E T P A G T E S P H $ $ A S P C Q E H K 315 CGA GGT GGC CTG AGC GAG CTG AAG GGA ACA CCT GCC TCT GCG CTG AGT CCT CCG GAG CCG GCG CCC TCG CCT GGG CAG CAG CAG CAG GCT GCA GCC CAC 1233 R G G L S E L K G T P A S A L S P P E P A P S P G Q Q Q Q A A A H 348 CTG CTG GGC CCA CCT CAC CAT CCT GGC CTG CCA CCA GAG GCC CAC CTG AAG CCC GAG CAC CAT TAC GCC TTC AAC CAC CCC TTC TCT ATC AAC AAC CTC 1332 L L G P P H H P G L P P E A H L K P E H H Y A F N H P F S I N N L 381 ATG TCC TCC GAG CAG CAA CAT CAT CAC AGC CAC CAC CAC CAT CAG CCC CAC AAA ATG GAC CTC AAG ACC TAC GAA CAG GTC ATG CAC TAC CCT GGG GGC 1431 M S S E Q Q H H H $ H H H H Q P H K M D L K T Y E O V M H Y P G G 414 TAC GGT TCC CCC ATG CCA GGC AGC TTG GCC ATG GGC CCA GTC ACG AAC AAA GCC GGC CTG GAT GCC TCG CCC CTG GCT GCA GAC ACT TCC TAC TAC CAG 1530 Y G S P M P G S L A M G P V T N K A G L D A S P L A A D T S Y Y Q 447 GGA GTG TAC TCC AGG CCT ATT ATG AAC TCG TCC TAA GAAGATGGCTTTCAGGCCCTGCTAGCTCTGGCCACTGGGGAGAAGGACAAGGGAAATGACAGGCTGAGTGGAAACATTGGGGG 1649 G V Y $ R P I M N S S - 458 AACTTTGAGGAAAAAGTAGCCACCACACTTAAGGCCCCAAAGGAGCAGTTTCACCTGTCTGTGTCCCTAAATAGCTGGGCCACACTGATCTGTcATTCTAAATAGGGAAGGGAATGGAAATATATATGTAT 1780 ACACATAAACTTGTTTTAATGGAGGACCTTTGGTTCCCACTATGTAGACTACTGCTTCTCAAGGCACCTGCAGATTTTGATTTTTGTTCTTGATTCTCTCTCTATTGCTGTCGTTGACAGAGAAGTCTGAC 1911 TTAAAAAAAAAAAAACTAAACAAAAAAAAAAACTTTTGATGAGTGACTTGAGTGTAAAACATGTAGTTTTAAAGAAAACAGAGGGTTGTATTGATGTTTAAAAGAGAAAAAAAAATAATGATGTAAG 2042 AGTCTGGTATAAATGACAGGAGAAAGGACAAAAATGATCCATTCTGGACATGGTGAAATCCAGTCTCGGGTCTGATTTAATTTATGGTTTCTGCGTGCTTTATTTATGGCTTATAAATGTGTGTTCTG 2173 GCTAGAATGCCAGAGTTCCACAAATCTATATTAAAGTATTATTGC 2217 Figure 1. Nucleic acid sequence of HNF-313. The cDNA clone (K2) contains the complete protein-coding sequence but is missing the 3' end of the mRNA. The protein sequence is shown below beginning at the initiator methionine at base 190, which is preceded by an in-frame termination codon. HNF-3C proteins, respectively. Definitive identification sates {lane 2), resulting in a much slower mobility com- of the proteins that form each of the gel mobility-shift plex, labeled T, which presumably represents a complex of probe, DNA-binding protein, and antibody. The anti-a complexes would require their purification and direct sequencing. The gel mobility-shift complex formed with serum also reacts weakly with the HNF-313 product, the HNF-313-translated product migrates identically to shifting a small part of that complex to a slower mobility that of HNF-3B. The gel mobility-shift complex formed (lane 4). with the HNF-3v product migrates slightly faster than The anti--y antiserum reacts with the HNF3~- the complex labeled HNF-3C. However, we obtained translated product {Fig. 5A, lane 6). This antiserum ap- two antisera that aid in the identification of the proteins pears to disrupt the binding of the protein to the DNA in the liver extracts. The first antiserum, which was rather than to change the mobility of the gel-shift com- plex. The corresponding controls for the antibody reac- raised in response to a synthetic peptide corresponding tions with preimmune serum are shown in Figure 5, to the amino-terminal 15 amino acids of HNF-3a, reacts with the protein products of the HNF-3a and HNF-313 lanes 1 and 3. genes (anti-a). HNF-313 shares 12 of 15 amino acids at the With crude liver extracts, the anti-or serum shifts the A amino terminus with HNF-3a. The second antibody was complex and a small part of the B complex (Fig. 5B, lane raised to a bacterial fusion product containing 77 amino 3). The anti-~/serum disrupts complex formation of the C acids of the HNF-3~/ sequence and is specific to the protein without affecting the complexes formed with A HNF-3~ protein (anti--y). The activity of anti-a is shown and B (Fig. 5B, lane 4). The specificity of the anti-~/an- in Figure 5A. This antiserum binds to the HNF-3a pro- tiserum is shown more clearly using partially purified duced by translation of this mRNA in reticulocyte ly- HNF-3C; anti-c~ serum does not affect binding, whereas 418 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press HNF-3 gene family TGTCGAGTGACTCCTCCTCTTCAGAAGAGGGCAATCTAGGGATAGTGAAGCT 52 GAAACACAGTGGAACAAGTCCTTGAAcCACAAGCTCATCGAACAGTTTTCTCGTTCCCCTTTGAGATATCAAGGAAGCTGTCACAGTCACTGTCTCATTCAGATGGTATAGATGCAGGCAGATTTTCTCAG 183 GGGGTCTCGGTGTAGATCTGGGACGGACCAGCACCATATACTTCGAGTTTCCATTGAGCAAGGAAGCAAAAGGCAGGAAGGCAAGAAGAGCTGAAAAAAGGCACAGCTGAGCCTCTCTGTCCTTCTTCATG 314 ATG GAA GGA CAG GAG GAC AGG ACG ACA GTA AGA CAA GAG GAC AGG GCT TAT GCT GTC CTC TTC CTT AAC ACT GCC TCC CCT CTG CCT ACA GGA CCC CTG 413 M E G Q E D R T T V R Q E D R A Y A V L F L N T A S P L P T G P L 33 GCA CCC CCA GCC CCC ACC GCT CCC CTG GGG CCC ACC TTC CCA GGC CTG GGC GCA GGC AGC GGC ACC GGA GGC AGT GCT TCC GGG TAT GGC GCC CCA GGG 512 A P P A P T A P L G P T F P G L G A G S G T G G S A S G Y G A P G 66 CCC GGG CTT GTA CAC GGA AAG GAA ATG GCC AAG GGG TAC CGG CGG CCA CTG GCT CAT GCC AAA CCA CCT TAT TCC TAC ATC TCT CTC ATC ACC ATG GCT 611 P G L V H G K E M A K G Y R R P L A H A K P P Y S Y I S L I T M A 99 ATC CAG CAG GCA CCA GGC AAG ATG CTG ACC CTG AGT GAA ATC TAC CAA TGG ATC ATG GAC CTC TTC CCG TAC TAC CGG GAG AAC CAG CAA CGT TGG CAG 710 I Q Q A P G K M L T L S E I Y Q W I M D L F P Y Y R E N Q Q R W Q 132 GTG GCA CGC TCC CCA GAC AAA CCG GGC AAA GGC TCC TAC TGG GCC TTG CAT CCC 809 AAC TCC ATC CGG CAC TCG CTG TCC TTC AAT GAC TGC TTC GTC AAG V A R S P D K P G K G S Y W A L H P 165 N S I R H S L S F N D C F V K AAG CGC TTC AAG CTG GAG GAG AAG GCA AAG AAA GGA AAC AGT GCC ACA TCC GCC 908 AGC TCT GGG AAC ATG TTT GAG AAC GGC TGC TAT CTC CGC CGT CAG K R F K L E E K A K K G N S A T S A 198 S S G N M F E N G C Y L R R Q GCT GCC ACT GCA GTC ACT TCC CCG GCT CAG CCC CAG CCT ACG CCG CCG TCT GAG 1007 ACC AGG AAT GGA ACA GTG GGG TCA GCC ACC TCT GCC ACC ACT ACA A A T A V T S P A Q P Q P T P P S E 231 T R N G T V G S A T S A T T T CCC GAG GCC CAG AGT GGG GAA GAT GTG GGG GGT CTG GAC TGC GCC TCG CCT CCT TCA TCC GCA CCC TAT TTC ACT GGC CTG GAG CTC CCA GGG GAA CTG 1106 P E A Q S G E D V G G L D C A S P P S S A P Y F T G L E L P G E L 264 AAG TTG GAT GCG CCC TAC AAC TTC AAC CAC CCT TTC TCT ATC AAC AAC CTG ATG TCA GAA CAG ACA TCA ACA CCT TCC AAA CTA GAT GTG GGG TTT GGG 1205 K L D A P Y N F N H P F S I N N L M S E Q T S T P S K L D V G F G 297 GGC TAC GGT GCT GAG AGT GGG GAA CCT GCG GTC TAC TAC CAG AGC CTC TAT TCC CGC TCT CTG CTC AAT GCA TCC TAG CGGCGCAACTGGGAGAAATGCGGTGAT 1310 G Y G A E S G E P G V Y Y Q S L Y S R S L L N A S - 322 GGGGGTTTGCTGTGACAGATGACCGGTTCTTCGGCCCTGAT~TTTCTGGTTACACTCTGCTTATCCAGTTAATTAACATCGTATTTGGTCTATTACTGTGATATGACCCATTGGCTACTGTGGTAACTGCA 1441 TGGACTCTTCGGTGGGCCTAGGGTTGGGGTATTGTGAAGGCAGATGCATTTGGGAGTGTTACGAAGGTGGCCATGTCGGACATACGTGAAGGCAATTAGACTGGTGTACTATAAAAGCTGCATGTTAAGT 1572 GAGTGATCCACTGGGTGCCTGATGGCCGCGATGTCGGAGGACATATTGTTTGGCCCTTTGGATGCT 1638 Figure 2. Nucleic acid sequence of HNF-3~/. The L11/35 cDNA contains the complete protein-coding sequence but is missing the 3' end of the mRNA. The protein sequence is shown below, beginning with two potential initiator methionines at bases 312 and 315. The second conforms much better to the consensus translation initiation sequence (Kozak 1986). anti-7 serum completely disrupts DNA binding {Fig. 5B, sential for DNA binding. The HNF-3a and HNF-313 pro- lanes 5-8). These data indicate that the HNF-3~/product teins are similar throughout the amino-terminal half of and HNF-3C share epitopes in the region of the protein the protein and are particularly rich in methionine in this region (Lai et al. 1990; Fig. 1}. Most of the rest of the that is distinct from HNF-3a or HNF-313. The molecular HNF-3a and HNF-313 proteins have recognizable se- mass on SDS gels of HNF-3C purified from rat liver ex- quence similarity (-30% identity) but are considerably tracts was determined previously to be 43 kD (Lai et al. 1990). Because the translated product of HNF-3~/is dis- divergent. The HNF-3~/ protein is -140 amino acids tinctly smaller on SDS-polyacrylamide gels, we con- shorter than HNF-3a and HNF-3~3, does not have a large clude that the two proteins result from differential mod- number of methionines in its amino terminus, and is ifications of the same gene product or are very closely highly divergent outside of the three conserved regions. After we completed the sequence of HNF-3A we con- related but not identical in sequence. ducted computer searches for matches with other pro- teins and failed to discover any. Later, when both the The major conserved region in HNF-3 proteins is in sequences of HNF-3A and the Drosophila gene fork head the DNA-binding domain had been entered in the available data bases, Weigel and J~ickle (1990) found a remarkable amino acid identity in A comparison of the amino acid sequences for the HNF-3 these two sequences within the domain shown to be proteins revealed three regions Ilabeled I, II, and III) that important for DNA binding of HNF-3A (HNF-3et). We are highly conserved in all three proteins (Fig. 6). We compared the fork head sequence to each of the HNF-3 demonstrated previously that the region between amino family members and found that there is greatest similar- acids 124 and 288 {numbering refers to HNF-3~} is re- ity to HNF-313, where 100 of 110 amino acids are iden- quired for DNA binding. Within this sequence is found tical in the binding domain (Fig. 6). The homology region I, a l l0-amino-acid stretch that is remarkably among all four proteins extends to regions II and III, as highly conserved in all three proteins. Ninety-three of well as to the binding domain, region I. the 110 amino acids are identical in the three proteins, whereas only one position is not shared between at least two of the three proteins. Nearly all of the differences HNF-3 proteins have different affinities for different represent conservative changes. Two other short, but binding sites well conserved, regions, II and III, are found in the car- boxyl end of the proteins. As reported earlier (Lai et al. The promoter of the TTR gene contains two HNF- 1990), the amino acids in these two regions are not es- 3-binding sites, both of which appear necessary for max- GENES & DEVELOPMENT 419 Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Lai et al. in vitro (Fig. 7). Using equal amounts of in vitro-trans- (2 y Full d7 d8 lated protein, determined by [35S]methionine incorpora- tion, we performed gel mobility-shift assays with labeled -84kD -84kD probes containing either the strong site, oligo S ( - 111 to - 57 kD - 57 kD -85) or the weak site, oligo W (-151 to -130). The - 48.5 kD - 48.5 kD amount of oligo S probe bound to HNF-3c~ or to HNF-3~/ m Qo was greater than the amount bound to HNF-3f~ (Fig. 7, - 365 kD .... - 36.5 kD lanes 1-6}. In contrast, with the oligo W probe, the amount complexed with HNF-3a or HNF-3~/ was less than that bound to HNF-3B (Fig. 7, lanes 7-12). Thus, HNF-3~ and HNF-3v have a higher affinity for the strong site than does HNF-3f~, whereas HNF-3~ has a higher Laver affinity for the weak site than either HNF-3a or HNF-3~/. - ÷ - + - + - + - + ,o This difference was most striking between HNF-3cx and HNF-3B. While HNF-3c, has a much higher affinity for the strong site compared to the weak site, HNF-3B has similar affinities for the two sites. To illustrate these differences more specifically, com- petition experiments with both oligonucleotides were carried out. Using labeled oligo S probe, competition with 10-fold molar excess of unlabeled oligo S reduced binding by -90% for HNF-3a (Fig. 8A, lane 2). The same amount of competitor only reduced the binding of HNF-3B to labeled probe by -50% (Fig. 8A, lane 8). HNF-3a complex formation was not significantly re- duced in the presence of a 40-fold molar excess of oligo W and a 200-fold excess of oligo W only decreased the HNF-3a complex by -50% (Fig. 8A, lanes 5 and 6). In I 2 3 4 5 6 7 8 I0 9 contrast, HNF-3B complex formation was reduced at Figure 3. The HNF-3B and HNF-37 cDNAs encode sequence- least 75% with a 40-fold of excess oligo W and abolished specific DNA-binding proteins. {A1 Autoradiograph of the trans- by 200-fold excess competitor (Fig. 8A, lanes 11 and 12). lated products of the HNF-3~, HNF-3[3, and HNF-37 cDNA an- Competition studies with weak-site oligo W as the alyzed by SDS-PAGE on a 9% acrylamide gel. (B) Confirmation probe are consistent with the results using the strong- of site of translation initiation in the HNF-3~ cDNA. Autorad- site oligo S as probe. A labeled oligo W probe in the iograph of translation products of full-length template and tem- presence of a 10-fold excess of unlabeled oligo S effec- plates shortened by 5' deletions of L11/35, d7, and d8. (C) Gel mobility-shift assay of the translated products compared to liver nuclear extract. Each of the translated proteins binds to the HNF-3 oligonucleotide (- 111 to -85 of the TTR promoter) in HNF-3[~ HNF-3c~ the presence of a nonspecific oligonucleotide ( - 175 to - 151 of the TTR promoter, indicated by the ( - in ) lanes 1, 3, 5, 7, and B E t [ I' S X B Iq E P S X 9) and is competed specifically by unlabeled HNF-3 oligonucle- otide [indicated by the + in lanes 2, 4, 6, 8 and I0]. The three complexes, A, B, and C, observed with liver extract are indi- U ,, N cated. Two micrograms of crude liver nuclear extract or 1 Ixl of m .... reticulocyte lysate, containing the translation products of HNF-3a, HNF-313, or HNF-3% was assayed. U imal activity of transfected DNA constructs and both of which are occupied in vivo by protein, presumably one of , ! the HNF-3 family members, as detected by a new in vivo footprint method (Mirkovitch and Darnell 1991 ). The to- tal set of binding proteins present in liver cell extracts, however, exhibits different affinities for the two DNA sites. A strong site is located between -96 and -106, 1 2 3 4 5 6 7 8 9 10 11 12 and a weak site is located between - 131 and - 140 nu- cleotides upstream of the cap site (Costa et al. 1989). In Figure 4. Southern hybridization of rat genomic DNA with an effort to gain insight into which proteins might actu- specific probes to HNF-3cx and HNF-3B. Digestion of genomic ally function in vivo, we compared the relative affinities DNA was performed with restriction endonucleases BamHI (B}, of each of the three HNF-3 proteins for each binding site EcoRI (E), HindlII {HI, PstI {P), SstI (SI, and XhoI (X). 420 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press HNF-3 gene family Comp - + - Antiserum Pre Pre o~ I Pre Pre a Antiserum Pre ~ Pre a Pre T--~ * T-" ' ~:? • Figure 5. The HNF-3~ protein is closely re- lated to the HNF-3C protein. (A) Anti-~ serum ..... eO .... reacts with the HNF-3c~ protein (lane 2) and O ~a~P* weakly with the HNF-3B protein (lane 4). Anti-v serum disrupts binding to DNA of the HNF-3v protein (lane 6). Preimmune or im- mune serum (1 wl) was added to the standard gel mobility-shift reactions containing reticu- locyte lysate-translated protein. (B) Anti-~ se- rum specifically reacts with the A and B (weakly) complexes formed with liver nuclear extract. The C complex is not affected (lanes 3 and 7). The anti-y serum specifically disrupts the C complex formed with liver extracts (lane 4) and with purified HNF-3C (lane 8). The re- actions were performed in the presence of a nonspecific competitor (-) or a specific com- 1 2 3 4 I 5 6 7 petitor (+) oligonucleotide (see legend to Fig. ,o213 41s 6 Liver Ext J HNF-3C 3). tively abolishes complex formation with HNF-3c~ but sion of the reporter gene compared to the control expres- not with HNF-3B (Fig. 8B, lanes 5 and 11). Furthermore, sion construct, which in this experiment contained a a 200-fold oligo W excess is able to compete effectively partial HNF-3~ cDNA in the antisense orientation. Sim- for binding to HNF-3~ but not to HNF-3o~ (Fig. 8B, lanes ilar results were seen with the HNF-3~ expression con- 9 and 3). Thus, it is clear that HNF-3e~ has a higher af- struct (data not shown}. In addition, the expression of a finity for oligo S than HNF-3B, and the reverse is true for reporter gene with multiple HNF-3 sites ligated to a het- oligo W. Since both of those sites are occupied in mouse erologous promoter was stimulated by cotransfection with HNF-3 expression constructs, whereas the reporter liver it seems likely that HNF-3~ is bound to the - 111 to - 85 site (oligo S}, whereas HNF-3~ is bound upstream construct lacking the multiple HNF-3 sites did not re- at - 151 to - 130. Because the DNA-binding properties spond (not shown). Taken together, the data on cotrans- of HNF-3~ and HNF-3~ are more similar to each other fection of plasmids producing HNF-3 and bearing HNF-3 than to those of HNF-3f~, it is possible that HNF-3oL and binding sites clearly indicate that each of the three pro- HNF-3v are interchangeable. teins, HNF-3a, HNF-3[3, and HNF-3% can act as positive activators of transcription. HNF-3 family members are transcriptional activators HNF-3 family members are differentially expressed Because early studies showed that mutation of the HNF- 3 strong site markedly reduced transcription at the TTR In our initial report of the cloning of an HNF-3 cDNA, promoter (Costa et al. 1989), it was expected that this the Northem analysis had shown two prominent mRNA site would be bound by transcriptional activators. With bands at 2.0 and 2.2 kb, as well as a minor band at 3.4 kb the identification of a family of proteins, however, we (Lai et al. 1990). In those experiments, we used a probe wanted to test which of the proteins were positive acti- that we now know spanned the homologous region of vators of transcription. This was accomplished by the three genes HNF-3~, HNF-3[3, and HNF-3% Thus, we presumably detected all three of the mRNAs that we cotransfection of expression vectors containing the cDNA for each of the HNF-3 family members and a re- now know to exist. Using probes that are specific to each porter construct with the TTR promoter (-202 to + 9) HNF-3 cDNA, we repeated the Northern analyses and containing the two HNF-3-binding sites described above found that the HNF-3cx transcript is 3.4 kb, the HNF-3[3 fused to the chloramphenicol acetyltransferase (CAT) transcript is 2.2 kb, and the HNF-3~ transcript is 2.0 kb gene. Figure 9 shows that the cotransfection into HepG2 {Fig. 10AI. These studies also show that in the liver cells of either HNF-3oL or HNF-3~ stimulated the expres- HNF-3[3 and HNF-3~/mRNAs were both more abundant GENES & DEVELOPMENT 421 Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Lai et al. 1o 20 3O 40 HNF-3e S YPL~HAKP P Y S Y I S L I TMA I QQAP[S KMLT LS E I YQW I MD L F P YY RQNQQ R Fork head T[TYRRSYT HAKP P YS Y I SL I TMA I Q T M L T L S E I Y Q~I M D L F Y R Q N Q Q R 70 80 ~ I00 110 HNF-3y HNF-3a NSIRHSLSFN~CF KVARSPDKPGKGSYWTLHPDSGNMFENGCYLRRQKRFK NSIRHSLSFNDCF K R PDKPGKGS WTLHPDSGNMFENGCYLRRQKRFKCE I HNF-3~ NSIRHSLSFNDCFVK]IPL~_RJT[PDKPGKGS WTLHPDSGNMFENGCYLRRQKRFK~L~ Fork head HNF-3~ IF N H P F S I N N L M S S S HNF-3a HNF-3~ F~__~H P F S I N N L M S S HNF-3~ S Y Q[S L]G GT T L Fork head $ S[H P F $ I N~L P T E Fork head I 1I llI .~F-3~ I 77 1 1186 ~ HNF- 3Or. l 1601 i ~ H I~ HNF-3~ [ 1~1 ]238 ~ [~ Fork head I 2001 13°9 ~ Figure 6. Homology of amino acid sequence among HNF-3e~, HNF-3B, HNF-37, and the Drosophila nuclear protein fork head. Sequence comparison of the DNA-binding domain (I) and two short regions at the carboxyl end of the proteins (II and III). Shaded areas show sequence identity between at least two of the proteins. (Bottom) A schematic diagram of the proteins, showing the location of the homologous regions. The positions of residues at both ends of region I are indicated. than HNF-3oL mRNA. The stronger gel mobility-shift ac- each of the family members independently because the tivity of HNF-3A (HNF-3~) with crude liver extracts (Fig. transcription rate is low and the cDNA-derived probes 5A; Lai et al. 1990) probably reflects the higher affinity of necessary to separate signals from the HNF-3a, HNF-3B, HNF-3a for the binding site (Figs. 7 and 8) rather than a and HNF-3~/ genes are too short. Only when genomic larger amount of HNF-3oL protein. clones are available will we be able to determine Using gene-specific probes, we confirmed by Northern whether different tissue distributions reflect proportion- ally different transcription rates. However, no transcrip- analysis that all three mRNAs were expressed in liver tion signal was obtained in kidney or brain, whereas a (Fig. 10A) but were absent in brain, spleen (data not clear signal was obtained in liver and lung using the full- shown), and kidney (Fig. 10B). There was a small but length probe for HNF-3a (Lai et al. 1990; K.G. Xantho- definite amount of expression in the small intestine of all three mRNAs (Fig. 10A). Longer exposures of the polous et al., in prep.). Therefore, the absence of any of blots of Figure 10A showed definite signals for HNF-3o~ the HNF-3 mRNAs in a tissue likely indicates no tran- scription of that gene in that tissue. and HNF-3f~ mRNAs. A more extensive tissue survey revealed additional in- formation about differential expression. HNF-3B was ex- Discussion pressed in the lung at high levels, approximately twofold greater than in liver (Fig. 10A, B). In contrast, HNF-3~/ We originally identified HNF-3oL as a protein that was mRNA was absent in the lung but present in the testis important in the transcriptional activation of two genes, (Fig. 10B). The HNF-3~/-like mRNA in the testis is dis- TTR and ~l-antitrypsin, which are expressed mainly in tinctly larger than that found in the liver. These differ- hepatocytes (Costa et al. 1989). Our current studies show ences were not due to differences in the amount of intact that HNF-3~ is only one member of a gene family that mRNA, as shown by hybridization with a probe to the encodes at least three transcription factors in rats. Se- GAPDH mRNA, which is present in all tissues (Fig. quence comparison of the three proteins reveals a highly 10B). The distribution of HNF-&x was identical to that of conserved 110-amino-acid domain, which lies within the HNF-3f~ for all of the tissues examined to date. region defined previously as neccessary for DNA bind- We have not been able to score the transcription rate of ing, and all three proteins bind the same DNA sequence, 422 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press HNF-3 gene family homeotic gene fork head product. Since all three rat pro- Oligo W Probe Oligo S teins and fork head conserved the entire 110 amino acids (-151/-130) (-111/-85) of the DNA-binding domain, it seems likely that all of HNF-3 (Z ~ ~/ Protein these amino acids may be required for DNA binding. For -- ÷ -- 4- -- 4- example, the DNA-binding domain in homeo box genes Competitor - ÷ - + - ÷ is conserved in vertebrates and invertebrates over a 60- amino-acid region that is required and sufficient for DNA binding (Gehring 1987). Furthermore, in the Pou y--. homeo gene family, which has members identified in Drosophila, Caenorhabditis elegans, and humans, an ~50-amino-acid conserved domain (the Pou region) has greater identity than the homeo box region and is re- quired for DNA binding (Herr et al. 1988). The other carboxy-terminal conserved regions of the HNF-3 fam- ily, which are not required for DNA binding, may con- ceivably be related to activating functions and either ac- tivation or repression for the fork head gene. The DNA-binding domain does not resemble that of any previously identified protein and thus represents a novel DNA-binding motif. At present, we can only spec- 1 2 3 4 5 6 7 8 10 9 11 12 ulate on the nature of the HNF-3c~ binding. The proteins Figure 7. Relative affinity of HNF-3 proteins for two binding give no evidence of dimerization. For example, cotrans- sites in the TTR promoter. Gel mobility-shift assays with trans- lation of full-length and truncated versions of HNF-3a or lated proteins using the strong-site oligo S (lanes 1-6) or the mixtures of HNF-3e, HNF-313, or HNF-3~/proteins does weak-site oligo W (lanes 7-12} as labeled probe. The reactions not yield complexes of intermediate mobility compared were performed in the presence of a nonspecific competitor or a to the mobility of each complex alone. Within the 110- specific competitor (oligo S for lanes 2, 4, and 6; oligo W for amino-acid conserved region, there is a cluster of basic lanes 8, 10, and 12). Approximately equal amounts of translated amino acids near the carboxyl end (RRQKREK). Such HNF-3 protein were used in all assays as estimated from the regions are known to be required for DNA binding of amount of [3SS]methionine incorporated. This experiment is helix-loop--helix and leucine zipper proteins and are representative of several using products from three separate thought to represent DNA contact points, whereas the translation reactions for each HNF-3 protein. helical portions of these proteins direct dimerization (Kouzarides and Ziff 1988; Murre et al. 1989; Turner and albeit with different affinities. This conservation in the Tjian 1989). There is only one fairly long region (from DNA-binding region is striking compared to the varia- Y/F at position 47 to V at position 72) within the l l0- amino-acid conserved region of the HNF-3 family that tion in the remainder of the coding region of these pro- has characteristics associated with helix formation. Ex- teins. For example, HNF-3~/is ~ 140 amino acids shorter cept for this region, prolines and glycines, which are con- than HNF-3c~, and HNF-313 and has only two short sidered helix breakers, are liberally scattered through the stretches of conserved sequence (II and III in Fig. 6) in remainder of region I. Such an amino acid distribution addition to the DNA-binding domain. The strong con- probably does not rule out helix formation but also does servation of the DNA-binding and the two carboxy-ter- minal regions (II and III) in the three rat genes is mirrored not allow speculation of its existence. The actual struc- by conservation of the same regions in the Drosophila ture of the HNF-3 class of proteins is perhaps one that Figure 8. Competition of HNF-3et and HNF-313 Probe Oligo S (-111/-85) B Probe Oligo W (-151/-130) binding to two binding sites in the TTR pro- moter. (A) Binding to the strong-site oligo S of translated HNF-3a (lanes 1--6) or HNF-313 (lanes Mo, r 7-12) in the presence of varying amounts of un- 0 40 200 2 10 40 0 40 2001 2 I0 40 Excess 0 10 40~10 40 200~ 0 10 40 10 40 200 Excess labeled competitor oligo S (lanes 2, 3, 8, and 9) or oligo W (lanes 4--6 and 10--12). (B} Binding to the weak-site oligo W of translated HNF-3a (lanes 1-6) or HNF-3[3 (lanes 7-12) in the presence of ~ii ~i ~ ~i~i~ varying amounts of unlabeled competitor oligo W (lanes 2, 3, 8, and 9) or oligo S (lanes 4--6 and 10--121. Approximately equal amounts of trans- lated HNF-3 protein were used in all assays as I1 2 3 4 5 6117 8 9 10 11 121 I l 2 3 4 6117 ~ 8 9 10 I1 121 HNF-3 (~ A HNF-3 estimated from the amount of [3SS]methionine Protein I° Protein incorporated. GENES & DEVELOPMENT 423 ~i~!!!i: Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Lai et al. form the terminal embryonic structures of the larval em- Expression Control [ HNF-3ot [ HNF-3[~ Vector bryo that contribute to such tissues as the salivary glands, foregut, and hindgut (Hartenstein et al. 1985; Jfir- gens and Weigel 1988; Weigel et al. 1989). It is logical to "T TT propose that this family of genes has maintained a role from early in evolution in forming gut-related organs• We came to the study of HNF-3 by its apparent regula- tory action on the TTR gene and a~-antitrypsin gene in adult liver. Whether one of the presently characterized HNF-3 proteins or some other member of this family has a role to play in earlier steps of hepatocyte determination YYTTTT now seems an important question. Furthermore, we can ask what the function may be of 1 2 3 4 5 6 different proteins of similar but not identical binding Figure 9. Transcriptional activation by cotransfection in affinities such as the HNF-3 family in adult rodent tis- HepG2 cells. Autoradiograph of TLC plate measuring activity of sues. First, it is obvious that different HNF-3 family CAT produced from a reporter construct TTR-CAT when members have widely different affinities for different cotransfected with a control expression plasmid (lanes 1 and 2) binding sites on DNA. Thus, different HNF-3 proteins or a plasmid expressing HNF-3a (lanes 3 and 4) or HNF-313 (lanes may be required to act on various sites in different genes 5 and 6). or even in the same gene. From our analysis of cell dis- tribution, it is clear that different tissues may vary in amounts of the different HNF-3 proteins, with HNF-3~/ allows the protein to bind to DNA as a monomer, thus being absent from lung, for example. All cells deriving presenting several unique challenges to the crystallogra- from endoderm may share some HNF-3-1ike proteins, pher. but some family members could be more restricted in What can we say of the broader biologic role of the expression and serve more local functions than others. Further tissue analyses and genomic searches for more HNF-3 family of proteins? The identity of the DNA- binding region of HNF-3a, HNF-3B, and HNF-3~/ with HNF-3 family members are needed to illuminate these that of the homeotic gene fork head product suggests questions. that they comprise a third group of mammalian and in- Finally, we must search more widely in adults and vertebrate genes that encode proteins having both simi- embryos for members of this family to attempt to ex- lar DNA-binding regions and at least somewhat similar plore the hypothesis about the restricted cell lineage in determining roles in early development. One group, the which this gene family is active. For example, fork head homeo box genes that control the proper segment forma- is present in brain cells in Drosophila, and we did not tion in the early Drosophila embryo, have a counterpart in vertebrates in Hox genes that have the homeo box sequence (Akam 1989). Because of the segmental expres- sion of Hox genes, these genes are expected to participate in proper segmental division of vertebrate embryos {Gra- ham et al. 1989). A second group encodes the helix-loop- • 28S helix proteins, one of which, myoD, has a demonstrated role in muscle cell differentiation when overexpressed in IINF-3~-- ° I8S certain susceptible cells and is a normal product during myogenesis. Another helix-loop-helix protein, E47, is 28S thought to play a role in lymphocyte differentiation. -28S Drosophila homologs of the helix-loop-helix genes in- HNF-3T -~" ~. 18S clude daughterless (Caudy et al. 1988) and achaete-scute '0 HNF-3~-~ ~ {Cabrera et al. 1987; Villares and Cabrera 1987); daugh- - 18S terless affects sex determination and the formation of ° 28S sensory precursor cells, whereas achaete-scute operates - 28S in cuticle formation. Thus, this second group of genes .18S 4m' may be devoted mainly to cell specialization. CAPDH-p ~ ~ ~ aJl HNF-3y"~ We have shown here that the HNF-3 family in rodents - 18S is expressed in cell types {liver, lung, and intestine) that derive from the embryonic gut tube by the outpouching Figure 10. Northern analysis of mRNA levels for each HNF-3 of endodermal cells, which on correct contact with mes- gene. (A) Specific probes distinguish the major mRNA species enchyme, form structures such as salivary glands, lungs, for HNF-3a, HNF-313, and HNF-3~ in the liver. The blot for pancreas, liver, and small and large intestine. It hardly HNF-3a was exposed for 72 hr; the blots for HNF-313 and seems accidental that the Drosophila protein homolog of HNF-3-y were exposed for 24 hr each. (B) HNF-313 and HNF-3~/ the HNF-3 family, fork head, is required to correctly mRNA expression in several tissues. 424 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press HNF-3 gene family observe HNF-3 mRNAs in total brain tissue. However, teine residue was added to facilitate coupling to keyhole limpet hemocyanin. Injections of 100 ~.g of coupled peptide with we have not yet examined individual cell types with in Freund's adjuvant were made subcutaneously at 4-week inter- situ hybridization or with specific antisera. We already vals. High-titer antiserum was obtained after three injections. know that we only detected CAAT enhancer binding Anti-~ serum was prepared by immunization of two rabbits protein (C/EBP), another transcription protein present in with a fusion peptide produced in Escherichia coll. The XmnI- liver, by in situ hybridization in brain cells (Kuo et al. SstI fragment of the HNF-3~I cDNA was inserted in the T7 poly- 1990). Thus, a search for region-specific expression of merase expression plasmid pET-3a (Rosenberg et al. 1987), HNF-3 in the brain is strongly indicated. which yielded an insoluble fusion peptide containing 77 amino acids of the HNF-3~ protein that was distinct from HNF-3~ or HNF-3~. Bacterial cells were sonicated in 50 mM Tris-HC1 {pH Materials and methods 7.5), 2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, and 0.1 M NaC1. Insoluble protein was recovered by centrifugation Isolation of HNF-3fl and HNF-37 cDNA at 12,000g for 10 min. The pellet was resuspended in sonication A random primer-labeled probe was prepared from the H3 buffer plus 1% SDS and centrifuged for 10 min. SDS-PAGE anal- cDNA (HNF-3a) and used to screen a mouse liver eDNA library ysis of the supematant showed that it was comprised of 50% of (generously provided by Dr. K.E. Paulson). Two clones were a single polypeptide of 10 kD. This preparation {200 ~g) was isolated that proved to be homologous to the rat cDNA. The used for each injection. longer clone, M2, was sequenced and found to encode a protein of 329 amino acids that was missing the amino-terminal end. A Preparation of probes specific to HNF-3a, HNF-3fl, and highly homologous region of - 100 amino acids was followed by HNF-37 a divergent sequence. A 270-bp fragment from this region was used to prepare a probe for screening a rat liver cDNA library Oligonucleotide primers were synthesized corresponding to the (Stratagene). One clone was identified, K2 (HNF-3f~I, which was coding strand of nucleotides 808-827 and the noncoding strand homologous to the M2 clone. of nucleotides 1252-1233 from the HNF-3c~ sequence. These The HNF-3~/ eDNA was obtained by screening the rat liver sequences are identical in the HNF-3~ eDNA, except for one cDNA library with a random primer-labeled fragment of mismatch, and flank an -300-nucleotide nonhomologous re- HNF-3a. Twenty positive clones were isolated and excised as gion. These primers were used in the polymerase chain reaction plasmids (kZAP protocol, Stratagene). Southern hybridization with plasmids containing HNF-3a and HNF-3B cDNA as tem- to the plasmid DNA digested with EcoRI was performed with plates to prepare fragments that are specific to each gene, except probes specific to HNF-3a and HNF-3~ cDNA. Seven clones did for the primer sequences. The corresponding region of the not hybridize, and six were found by restriction mapping and HNF-3~ cDNA could be obtained by isolating the 234-bp frag- partial sequence determination to be overlapping fragments of a ment generated by restriction enzymes XmnI and SstI. Probes distinct cDNA. Two of these clones were ligated at a shared were prepared by random primer labeling of these fragments NcoI site to create Lll/35, which encodes the full-length after denaturation. HNF-3-y protein sequence. Sequence was performed by the dideoxy chain-termination method on double-stranded tem- Transfection constructs and assays plates. The K2 and L I 1/35 sequences were determined com- pletely on both strands with overlapping subcloned and exonu- The vectors used for the expression of HNF-3 cDNAs were de- clease-shortened templates, as well as with ITP-containing se- scribed previously (Chen et al. 1987). The H3 eDNA was used quencing mixes (U.S. Biochemical Sequenase Kit) to eliminate for HNF-3ct, the K2 cDNA was used for HNF-313, and the L11135 compression artifacts. cDNA was used for HNF-3~/. The control construct contained the murine HNF-3~ partial cDNA, M2, in the antisense orien- tation. Each of these was cloned into the 5' XbaI and 3' HindIII In vitro translation and gel mobility-shift assays sites of the expression vector. The XbaI and HindIII sites of each Linearized plasmids were used as templates for transcription cDNA were generated by cloning each EcoRI fragment into with T7 or T3 RNA polymerase (Melton et al. 1984). In vitro pBluescript KS in the appropriate orientation. The reporter con- translation was performed by use of nuclease-treated rabbit re- struct TTR-CAT was made by purifying the mouse TTR pro- ticulocyte lysate (Promega) according to the manufacturer's pro- motet sequence -202 to + 9 of the TTR minigene (Costa et al. tocol. [3SSlMethionine (translation grade) was purchased from 1986), adding HindIII linkers, and fusing to the HindIII site of New England Nuclear. The reticulocyte lysate containing the the CAT gene (Lew et al. 1991) without any other promoter translated protein was used directly in gel mobility-shift assays sequence. as described previously (Kovesdi et al. 1986) with end-labeled Transfection of human HepG2 cells was carried out by cal- double-stranded oligonucleotide probes. For the comparison of cium phosphate precipitation IWigler et al. 1979), except that DNA-binding affinities, the relative amount of translated pro- 5 mM sodium butyrate was used after glycerol shock. tein was estimated by SDS-PAGE and autoradiography of the pCMV-B-gal (MacGregor and Caskey 1989) was used as an in- dried gel. The relative amount of label incorporated was quan- ternal control for transfection efficiency. For each transfection, titated by densitometry and then adjusted for the number of 1 ~g of pCMV-~-gal, 2 ~g of TTR-CAT, and 10 ~g of each methionine residues in each HNF-3 protein being translated to expression vector were added to 5 x 106 HepG2 cells in 100- arrive at an estimate of the relative amount of protein. mm dishes. CAT assays and B-gal normalizations were carried out as described (Lew et al. 1991). Preparation of antisera to HNF-3 proteins RNA analysis Anti-a was obtained by immunization of two rabbits with a synthetic peptide MLGTVKMEGHESNDWC, corresponding to RNA was isolated from rat tissues by the acid guanidinium the 15 amino-terminal residues of HNF-3(~. The terminal cys- thiocyanate/phenol-chloroform extraction method (Chomczyn- GENES & DEVELOPMENT 425 Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Lai et al. ski and Sacchi 1987; Puissant and Houdebine 1990). Poly(A) + 6593. Frain, M., G. Swart, P. Monaci, A. Nicosia, S. Stampfli, R. Frank, RNA was selected by chromatography on oligo(dT)-cellulose. and R. Cortese. 1989. The liver-specific transcription factor Two micrograms of poly{A) + RNA from each tissue was sepa- LF-B1 contains a highly diverged homeobox DNA binding rated on a formaldehyde-l.2% agarose gel and transferred to domain. Ce11 59: 145-157. Zeta-probe nylon membranes (Bio-Radl by capillary action. Hy- Gehring, W.l. 1987. Homeo boxes in the study of development. bridization was performed in 0.5 M sodium phosphate (pH 7), 1 mM EDTA, and 7% SDS at 65°C for 18-24 hr. The high-strin- Science 236: 1235-1252. Graham, A., N. Papalopulu, and R. Krumlauf. 1989. The murine gency wash was performed in 0.1% SSC/0.1% SDS at 65°C for 10 rain. The blot was exposed to X-ray film at - 80°C with two and Drosophila homeobox gene complexes have common intensifying screens for 24-72 hr. features of organization and expression. Cell 57: 367-378. Hartenstein, V., G.M. Technau, and ].A. Campos-Orterga. 1985. Fate-mapping in wild type Drosophila melanogaster. Acknowledgments Wilhelm Roux's Arch. Dev. Biol. 194: 181-195. Herr, W., R.A. Sturm, R.G. Clerc, L.M. Corcoran, D. Baltimore, We thank Gabriela Balas and Oksana Litvin for excellent tech- P.A. Sharp, H.A. Ingraham, M.G. Rosenfeld, M. Finney, G. nical assistance, and C.F. Kuo, K.E. Paulson and F.M. Sladek for Ruvkun, and H.R. Horvitz. 1988. The POU domain: A large helpful discussions. This work was funded in part by National conserved region in the mammalian pit-l, oct-l, oct-2, and Institutes of Health grant CA16006-16A to I.E.D., an NRSA Caenorhabditis elegans unc-86 gene products. Genes & training grant (GM07982-09) to V.R.P., and a Merck Fellowship Dev. 2: 1513-1516. to W.S.C. Johnson, P.F., W.H. Landschulz, B.J. Graves, and S. McKnight. The publication costs of this article were defrayed in part by 1987. Identification of a rat liver protein that binds to the payment of page charges. This article must therefore be hereby enhancer core element of three animal viruses. Genes & marked "advertisement" in accordance with 18 USC section Dev. 1: 133-146. 1734 solely to indicate this fact. lfirgens, G. and D. Weigel. 1988. Terminal versus segmental development in the Drosophila embryo: The role of the ho- meotic gene fork head. Wilhelm Roux's Arch. Dev. Biol. Note added in proof 197: 345-354. Sequence data described in this paper have been submitted to the Kouzarides, T. and E. Ziff. 1988. The role of the leucine zipper EMBL/GenBank Data Libraries. in the fos-jun interaction. Nature 336: 646-651. Kovesdi, I., R. Reichel, and I.R. Nevins. 1986. Identification of a cellular transcription factor involved in E1A trans-activa- References tion. Cell 45: 219-228. Akam, M. 1989. Hox and HOM: Homologous gene clusters in Kozak, M. 1986. Point mutations define a sequence flanking the AUG initiator codon that modulates translation by eukary- insects and vertebrates. Cell 57: 347-349. Baumhueter, S., D.B. Mendel, P.B. Conley, C.I. Kuo, C. Turk, otic ribosomes. Cell 44: 283-292. M.K. Graves, C.A. Edwards, G. Courtois, and G.R. Crabtree. Kuo, C.F., K.G. Xanthopoulos, and I.E. Damell lr. 1990. Fetal 1990. HNF-1 shares three sequence motifs with the POU and adult localization of C/EBP: Evidence for combinatorial domain proteins and is identical to LF-B1 and APF. Genes & action of transcription factors in cell-specific gene expres- sion. Development 109:473-481. Dev. 4: 372-379. Cabrera, C.V., A. Martinez-Arias, and M. Bate. 1987. The ex- Lai, E., V.R. Prezioso, E. Smith, O. Litvin, R.H. Costa, and I.E. Darnell Jr. 1990. HNF-3A, A hepatocyte-enriched transcrip- pression of three members of the achaete-scute gene com- tion factor of novel structure is regulated transcriptionally. plex correlates with neuroblast segregation in Drosophila. Genes & Dev. 4: 1427-1436. Cell 50: 425-433. Lew, D.I., T. Decker, I. Strehlow, and J.E. Damell. 1991. Over- Caudy, M., H. Vassin, M. Brand, R. Tuma, L.Y. Jan, and Y.N. lapping elements in the guanylate-binding protein gene pro- Jan. 1988. daughterless, a Drosophila gene essential for both neurogenesis and sex determination. Cell 55: 1061-1067. moter mediate transcriptional induction by alpha and Chen, W.S., C.S. Lazar, M. Poenie, R.Y. Tsien, G.N. Gill, and gamma interferons. Mol. Cell Biol. 11: (in press). M.G. Rosenfeld. 1987. Requirement for intrinsic protein ty- MacGregor, G.R. and C.T. Caskey. 1989. Constructions of plas- rosine kinase in the immediate and late actions of the EGF mids that express E. coli 13-galactosidase in mammalian receptor. Nature 328: 820-823. cells. Nucleic Acids Res. 17: 2365. Chomczynski, P. and N. Sacchi. 1987. Single-step method of Melton, D.A., P.A. Krug, M.R. Rebaglicti, T. Maniatis, K. Zinn, RNA isolation by acid guanidinium thiocyanate- and M.R. Green. 1984. Efficient in vitro synthesis of biolog- ically active RNA and RNA hybridization probes from plas- phenol-chloroform extraction. Anal Biochem. 162: 156- 159. raids containing a bacteriophage SP6 promoter. Nucleic Ac- ids Res. 12: 7035-7056. Costa, R.H., E. Lai, and J.E. DameI1 Jr. 1986. Transcriptional Mirkovitch, J. and J.E. Damell Jr. 1991. Rapid in vivo footprint- control of the mouse prealbumin (transthyretin) gene: Both ing technique identifies proteins bound to the TTR gene in promoter sequences and a distinct enhancer are cell specific. liver. Genes & Dev. 5: 83-93. MoI. Cell. Biol. 6: 4697-4708. Murre, C., P.S. McCaw, and D. Baltimore. 1989. A new DNA Costa, R.H., D.R. Grayson, and I.E. Darnell lr. 1989. Multiple hepatocyte-specfic nuclear factors function in the regulation binding and dimerization motif in immunoglobulin en- of the transthyretin and al-antitrypsin genes. Mol. Cell. hancer binding, daughterless, MyoD, and myc proteins. Cell Biol. 9: 1415-1425. 56: 777-783. Costa, R.H., T. Van Dyke, C. Yan, C.F. Kuo, and I.E. Damell Jr. Puissant, C. and L. Houdebine. 1990. An improvement of the 1990. Similarities in transthyretin gene expression and dif- single-step method of RNA isolation by acid guanidinium ferences in transcription factors: Liver and yolk sac com- thiocyanate-phenol-chloroform extraction. BioTechniques pared to choroid plexus. Proc. Natl. Acad. Sci. 87: 6589- 8: 148-149. 426 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press HNF-3 gene family Rosenberg, A.H., B.N. Lade, D. Chui, S. Lin, I.J. Dunn, and F.W. Studier. 1987. Vectors for selective expression of cloned cD- NAs by T7 RNA polymerase. Gene 56: 125-135. Sladek, F.M., W. Zhong, E. Lai, and J.E. Damell Jr. 1990. Liver- enriched transcription factor HNF-4 is a novel member of the steroid hormone receptor superfamily. Genes & Dev. 4: 2353-2365. Tilghman, S.M. and A. Belayew. 1982. Transcriptional control of the murine albumin/a-fetoprotein locus during develop- ment. Proc. Natl. Acad. Sci. 79: 5254-5257. Turner, R. and R. Tjian. 1989. Leucine repeats and an adjacent DNA binding domain mediate the formation of functional cFos-cJun heterodimers. Science 243: 1689-1694. Villares, R. and C.V. Cabrera. 1987. The achaete-scute gene complex of D. melanogaster: Conserved domains in a subset of genes required for neurogenesis and their homology to myc. Cell 50: 415-424. Weigel, D. and H. J~ckle. 1990. Fork head: A new eukaryotic DNA binding motif? Cell 63: 455-456. Weigel, D., G. Jfirgens, F. Kuttner, E. Seifert, and H. J~ickle. 1989. The homeotic gene fork head encodes a nuclear protein and is expressed in the terminal regions of the Drosophila em- bryo. Cell 57: 645--658. Wigler, M., A. Pellicer, S. Silverstein, R. Axel, G. Urlaub, and L. Chasin. 1979. DNA-mediated transfer of adenylphosphori- bosyl transferase locus into mammalian cells. Proc. Natl. Acad. Sci. 76: 1373-1376. Xanthopoulos, K.G., J. Mirkovitch, T. Decker, C.F. Kuo, and J.E. Darnell Jr. 1989. Cell-specific transcriptional control of the mouse DNA-binding protein C/EBP. Proc. Natl. Acad. Sci. 86:4111--4121. GENES & DEVELOPMENT 427 Downloaded from genesdev.cshlp.org on November 25, 2021 - Published by Cold Spring Harbor Laboratory Press Hepatocyte nuclear factor 3 alpha belongs to a gene family in mammals that is homologous to the Drosophila homeotic gene fork head. E Lai, V R Prezioso, W F Tao, et al. Genes Dev. 1991, 5: Access the most recent version at doi:10.1101/gad.5.3.416 This article cites 35 articles, 13 of which can be accessed free at: References http://genesdev.cshlp.org/content/5/3/416.full.html#ref-list-1 License Receive free email alerts when new articles cite this article - sign up in the box at the top Email Alerting right corner of the article or click here. Service Copyright © Cold Spring Harbor Laboratory Press

Journal

Genes & Development – Unpaywall

Published: Mar 1, 1991

There are no references for this article.