Voltar
Artigo Acesso aberto Revisado por pares
BIBTEX RIS

Regulated Translation of Heparan SulfateN-Acetylglucosamine N-Deacetylase/N-Sulfotransferase Isozymes by Structured 5′-Untranslated Regions and Internal Ribosome Entry Sites

2002; Elsevier BV; Volume: 277; Issue: 34 Linguagem: Inglês

10.1074/jbc.m111904200

ISSN

1083-351X

Autores

Kay Grobe, Jeffrey D. Esko,

Tópico(s)

Carbohydrate Chemistry and Synthesis

Resumo

We report the full-length 5′-untranslated region (5′-UTR) sequences of the four vertebrate heparan sulfate/heparin GlcNAcN-deacetylase/N-sulfotransferases (NDSTs) and their role in translational regulation in vivo and in vitro. All four NDST 5′-UTR sequences are unusually long, have a high degree of predicted secondary structure, and contain multiple upstream AUG codons, which together impose a major barrier to conventional, cap-dependent ribosomal scanning. At least two alternatively spliced forms of NDST2 differing in their 5′-UTRs exist, and two forms of NDST4 arise from alternative transcriptional start sites. The 5′-UTRs do not show any significant sequence similarity between isozymes, but possess highly conserved regions between mouse and human orthologs, pointing toward evolutionarily conserved functions. Expression of bicistronic vector constructs showed that the 5′-UTRs of NDST1–4 are capable of regulating translation differentially in vivo dependent on cell type and culture conditions. In vitro translation of a reporter gene located downstream of the UTRs demonstrated the presence of internal ribosome entry sites, providing an additional, cap-independent step in fine-tuning NDST expression. Comparative studies of NDST1–3 mRNAs and protein expression in brain and embryonic extracts revealed striking differences in translational efficiency. Other genes necessary for glycosaminoglycan synthesis in addition to the NDST isozymes have long, structured 5′-UTRs. Because several growth factors and morphogens that bind heparan sulfate also contain structured 5′-UTRs, translational regulation may coordinate the action of these factors and their heparan sulfate co-receptors. We report the full-length 5′-untranslated region (5′-UTR) sequences of the four vertebrate heparan sulfate/heparin GlcNAcN-deacetylase/N-sulfotransferases (NDSTs) and their role in translational regulation in vivo and in vitro. All four NDST 5′-UTR sequences are unusually long, have a high degree of predicted secondary structure, and contain multiple upstream AUG codons, which together impose a major barrier to conventional, cap-dependent ribosomal scanning. At least two alternatively spliced forms of NDST2 differing in their 5′-UTRs exist, and two forms of NDST4 arise from alternative transcriptional start sites. The 5′-UTRs do not show any significant sequence similarity between isozymes, but possess highly conserved regions between mouse and human orthologs, pointing toward evolutionarily conserved functions. Expression of bicistronic vector constructs showed that the 5′-UTRs of NDST1–4 are capable of regulating translation differentially in vivo dependent on cell type and culture conditions. In vitro translation of a reporter gene located downstream of the UTRs demonstrated the presence of internal ribosome entry sites, providing an additional, cap-independent step in fine-tuning NDST expression. Comparative studies of NDST1–3 mRNAs and protein expression in brain and embryonic extracts revealed striking differences in translational efficiency. Other genes necessary for glycosaminoglycan synthesis in addition to the NDST isozymes have long, structured 5′-UTRs. Because several growth factors and morphogens that bind heparan sulfate also contain structured 5′-UTRs, translational regulation may coordinate the action of these factors and their heparan sulfate co-receptors. N-deacetylase/N-sulfotransferases eukaryotic initiation factor rapid amplification of cDNA ends 5′-untranslated region upstream AUG codon internal ribosome entry site(s) chloramphenicol acetyltransferase encephalomyocarditis virus enhanced green fluorescent protein Chinese hamster ovary glutathioneS-transferase Heparan sulfate and heparin bind a variety of growth factors, enzymes, and extracellular matrix proteins through specific arrangements of variably sulfated sugar residues (1****Google Scholar, 2Capila I. Linhardt R.J. Angew. Chem. Int. Ed. Engl. 2002; 41: 391-412Crossref PubMed Scopus (1524) Google Scholar). Several sulfotransferases guide the formation of these binding sequences. The initial reactions involve removal of acetyl groups from GlcNAc residues, followed by sulfation of the amino group catalyzed by one or more GlcNAc N-deacetylase/N-sulfotransferases (NDSTs)1 (3Hashimoto Y. Orellana A. Gil G. Hirschberg C.B. J. Biol. Chem. 1992; 267: 15744-15750Abstract Full Text PDF PubMed Google Scholar, 4Orellana A. Hirschberg C.B. Wei Z. Swiedler S.J. Ishihara M. J. Biol. Chem. 1994; 269: 2270-2276Abstract Full Text PDF PubMed Google Scholar, 5Eriksson I. Sandbäck D., Ek, B. Lindahl U. Kjellén L. J. Biol. Chem. 1994; 269: 10438-10443Abstract Full Text PDF PubMed Google Scholar, 6Aikawa J. Esko J.D. J. Biol. Chem. 1999; 274: 2690-2695Abstract Full Text Full Text PDF PubMed Scopus (130) Google Scholar, 7Aikawa J. Grobe K. Tsujimoto M. Esko J.D. J. Biol. Chem. 2001; 276: 5876-5882Abstract Full Text Full Text PDF PubMed Scopus (186) Google Scholar). Four isozymes exist in vertebrates, whereas only one is expressed inDrosophila melanogaster and Caenorhabditis elegans. The importance of NDST1 in normal physiology in mice has been established by targeted disruption of the gene, which results in embryonic and neonatal lethality due to multiple organ defects (8Ringvall M. Ledin J. Holmborn K. Van Kuppevelt T. Ellin F. Eriksson I. Olofsson A.M. Kjellén L. Forsberg E. J. Biol. Chem. 2000; 275: 25926-25930Abstract Full Text Full Text PDF PubMed Scopus (206) Google Scholar, 9Fan G. Xiao L. Cheng L. Wang X. Sun B. Hu G. FEBS Lett. 2000; 467: 7-11Crossref PubMed Scopus (143) Google Scholar). In contrast, disruption of NDST2 affects only connective tissue-type mast cells, resulting in a storage deficiency for proteases and histamine due to lack of heparin (10Forsberg E. Pejler G. Ringvall M. Lunderius C. Tomasini-Johansson B. Kusche-Gullberg M. Eriksson I. Ledin J. Hellman L. Kjellén L. Nature. 1999; 400: 773-776Crossref PubMed Scopus (403) Google Scholar, 11Humphries D.E. Wong G.W. Friend D.S. Gurish M.F. Qiu W.T. Huang C.F. Sharpe A.H. Stevens R.L. Nature. 1999; 400: 769-772Crossref PubMed Scopus (359) Google Scholar). The function of NDST3 and NDST4 are not yet established. As NDST1 and NDST2 show ubiquitous mRNA expression (7Aikawa J. Grobe K. Tsujimoto M. Esko J.D. J. Biol. Chem. 2001; 276: 5876-5882Abstract Full Text Full Text PDF PubMed Scopus (186) Google Scholar), the strikingly different phenotypes of the mutants point toward other modes of enzyme regulation, possibly at the post-transcriptional or post-translational level.The leader sequences of most mRNAs are usually short (50–70 nucleotides), contain 5′-m7G cap structures, and lack highly organized secondary structures (12Gray N.K. Wickens M. Annu. Rev. Cell Dev. Biol. 1998; 14: 399-458Crossref PubMed Scopus (447) Google Scholar, 13Cazzola M. Skoda R.C. Blood. 2000; 95: 3280-3288Crossref PubMed Google Scholar, 14van der Velden A.W. Thomas A.A. Int. J. Biochem. Cell Biol. 1999; 31: 87-106Crossref PubMed Scopus (311) Google Scholar). These features are compatible with efficient translation by the ribosomal scanning mechanism, which is considered valid for the majority of cellular and viral mRNAs (15Kozak M. Gene (Amst.). 1999; 234: 187-208Crossref PubMed Scopus (1121) Google Scholar, 16Pestova T.V. Kolupaeva V.G. Lomakin I.B. Pilipenko E.V. Shatsky I.N. Agol V.I. Hellen C.U. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 7029-7036Crossref PubMed Scopus (596) Google Scholar). Translational control of these transcripts is most commonly exerted at the rate-limiting initiation phase. The 5′-cap structure (m7GpppN) attracts the eukaryotic initiation factor (eIF) 4F complex, composed of the cap-binding protein eIF4E, the RNA-dependent ATPase (helicase) eIF4A, and the modular factor eIF4G. The small (40 S) ribosomal subunit binds to the 5′-end of the RNA·eIF4F complex along with eIF3 and the ternary complex of eIF2, GTP, and Met-tRNAi. The resulting 48 S complex then advances through the initiation cycle, resulting in lateral movement of the 43 S unit along the mRNA (scanning), usually to the first AUG triplet, which serves as the initiation codon. Upon dissociation of initiation factors, the large (60 S) subunit joins to form the 80 S ribosome. The eIF4E-eIF4G interaction with RNA is a limited controlled process, reduced by hypophosphorylation of eIF4E or the competitive inhibitors (eIF4E-binding proteins) or by phosphorylation of eIF2 (17Sonenberg N. Newgard C.B. Science. 2001; 293: 818-819Crossref PubMed Scopus (18) Google Scholar).Rapid amplification of the 5′-ends of murine and human NDST cDNAs (5′-RACE) revealed long, structured, and AUG-rich 5′-untranslated regions (5′-UTRs). These types of UTRs are incompatible with the described scanning mode for translation initiation due to secondary structure constraints and abortive initiation at upstream AUG codons (uAUGs) located 5′ to the actual initiation site for translation. Complex 5′-UTRs are found in ∼10% of mRNAs, often in those encoding regulatory proteins such as proto-oncogenes (18Willis A.E. Int. J. Biochem. Cell Biol. 1999; 31: 73-86Crossref PubMed Scopus (117) Google Scholar), growth factors (insulin-like growth factor II, platelet-derived growth factor 2, transforming growth factor-β, fibroblast growth factor 2, and vascular endothelial growth factor) and their receptors, and homeodomain proteins (14van der Velden A.W. Thomas A.A. Int. J. Biochem. Cell Biol. 1999; 31: 87-106Crossref PubMed Scopus (311) Google Scholar). In some of these cases, initiation is driven from internal ribosome entry sites (IRES) that allow for cap-independent translation, thus affording another level of control over protein expression. Specific sulfated sugar sequences in cell-surface heparan sulfate can bind to many of these growth factors and facilitate signaling (19Park P.W. Reizes O. Bernfield M. J. Biol. Chem. 2000; 275: 29923-29926Abstract Full Text Full Text PDF PubMed Scopus (305) Google Scholar). Thus, cap-independent translational regulation of NDST expression and the growth factors may be coordinated to fine-tune growth responses in target cells. Here, we report the comparison of the 5′-UTRs of the NDSTs, their regulatory properties with respect to translation, and the presence of complex UTRs in other enzymes involved in heparan sulfate biosynthesis. Our findings suggest the possibility of coordinated translational control over several components of growth factor signaling pathways.DISCUSSIONNDST1–4 are important enzymes in the formation of heparan sulfate because all other modifications of the chains (O-sulfation and epimerization of uronic acids) depend on the conversion of GlcNAc to GlcNSO3 residues (27Esko J.D. Lindahl U. J. Clin. Invest. 2001; 108: 169-173Crossref PubMed Scopus (782) Google Scholar, 28Esko J.D. Selleck S.B. Annu. Rev. Biochem. 2002; 71: 435-471Crossref PubMed Scopus (1226) Google Scholar). The four isozymes have different biochemical properties that manifest as variation in substrate preference and relative activity (7Aikawa J. Grobe K. Tsujimoto M. Esko J.D. J. Biol. Chem. 2001; 276: 5876-5882Abstract Full Text Full Text PDF PubMed Scopus (186) Google Scholar). The four isozymes are also differentially expressed in tissues, based on PCR analysis or Northern blots of mRNA (7Aikawa J. Grobe K. Tsujimoto M. Esko J.D. J. Biol. Chem. 2001; 276: 5876-5882Abstract Full Text Full Text PDF PubMed Scopus (186) Google Scholar, 23Kusche-Gullberg M. Eriksson I. Pikas D.S. Kjellén L. J. Biol. Chem. 1998; 273: 11902-11907Abstract Full Text Full Text PDF PubMed Scopus (78) Google Scholar). In this study, we have shown that the expression pattern of the NDST isozymes may also reflect differential translation in tissues dependent on their 5′-UTRs. The evidence is based on (i) the length and complexity of the 5′-UTRs from mice and human homologs, (ii) differential expression of reporter constructs in which the 5′-UTRs were ligated to EGFP, (iii) mutational analysis, (iv) the effects of serum and rapamycin on translation, and (v) differences in NDST protein expression relative to mRNA in brain and embryonic extracts. These findings demonstrate an important caveat in interpreting levels of enzyme expression from NDST mRNA levels. The data suggest that the level of enzyme activity contributed by the various isoforms will depend on both transcription and translation.The presence of regulatory elements in the 5′-UTRs of the NDSTs may explain some of the surprising results obtained in genetic studies of mice in which individual NDSTs have been inactivated (29Grobe K. Ledin J. Ringvall M. Holborn K. Forsberg E. Esko J.D. Kjellen L. Biochim. Biophys. Acta. 2002; (in press)PubMed Google Scholar). NDST1 and NDST2 mRNAs are abundantly expressed in nearly all tissues (7Aikawa J. Grobe K. Tsujimoto M. Esko J.D. J. Biol. Chem. 2001; 276: 5876-5882Abstract Full Text Full Text PDF PubMed Scopus (186) Google Scholar, 23Kusche-Gullberg M. Eriksson I. Pikas D.S. Kjellén L. J. Biol. Chem. 1998; 273: 11902-11907Abstract Full Text Full Text PDF PubMed Scopus (78) Google Scholar), and both enzymes can participate in heparan sulfate formation (6Aikawa J. Esko J.D. J. Biol. Chem. 1999; 274: 2690-2695Abstract Full Text Full Text PDF PubMed Scopus (130) Google Scholar, 30Ishihara M. Kiefer M.C. Barr P.J. Guo Y. Swiedler S.J. Anal. Biochem. 1992; 206: 400-407Crossref PubMed Scopus (16) Google Scholar). Inactivation of NDST1 has a profound effect on development (8Ringvall M. Ledin J. Holmborn K. Van Kuppevelt T. Ellin F. Eriksson I. Olofsson A.M. Kjellén L. Forsberg E. J. Biol. Chem. 2000; 275: 25926-25930Abstract Full Text Full Text PDF PubMed Scopus (206) Google Scholar, 9Fan G. Xiao L. Cheng L. Wang X. Sun B. Hu G. FEBS Lett. 2000; 467: 7-11Crossref PubMed Scopus (143) Google Scholar), suggesting a prominent role for this enzyme in many tissues. The restricted phenotype of NDST2 is surprising, however, given the widespread expression of its mRNA. The data presented here suggest that NDST2 enzyme expression may not occur even though abundant message is present. The lack of phenotype in those tissues containing both enzymes suggests that the two activities may be redundant in certain contexts.Having 5′-UTR sequences with embedded IRES motifs allows cells to regulate the level of translation, which is especially relevant for regulatory factors. One of the best studied systems involves cellular iron homeostasis, which is regulated by translational control of the transferrin receptor and ferritin expression (31Muckenthaler M. Gray N.K. Hentze M.W. Mol. Cell. 1998; 2: 383-388Abstract Full Text Full Text PDF PubMed Scopus (208) Google Scholar). Cytoplasmic RNA-binding proteins that bind iron (iron regulatory proteins) interact with specific stem-loop structures (iron response elements) in the 5′-UTRs of the corresponding mRNAs, thus regulating expression of the genes according to the iron status of the cytosol. Whether similar stem-loop structures and regulatory proteins exist for the UTRs of the NDSTs awaits further analysis.Interestingly, the cell type-specific translational regulation of NDST isozymes strongly resembles the pattern of regulation of growth factors, some of which bind heparan sulfate in a sulfation-dependent manner. For example, tissue-specific regulation of fibroblast growth factor 2 translation appears to be dependent on its 5′-UTR (32Creancier L. Morello D. Mercier P. Prats A.C. J. Cell Biol. 2000; 150: 275-281Crossref PubMed Scopus (118) Google Scholar). Similarly, vascular endothelial growth factor regulation by hypoxia (33Stein I. Itin A. Einat P. Skaliter R. Grossman Z. Keshet E. Mol. Cell. Biol. 1998; 18: 3112-3119Crossref PubMed Scopus (420) Google Scholar) and platelet-derived growth factor 2/c-sis expression that occurs during differentiation (34Bernstein J. Sella O., Le, S.Y. Elroy-Stein O. J. Biol. Chem. 1997; 272: 9356-9362Abstract Full Text Full Text PDF PubMed Scopus (153) Google Scholar) are regulated in this way as well. The 5′-UTRs of these and other growth factors are often long, have high folding energies, contain uAUGs, and are often equipped with an IRES, which together can contribute to precise expression levels during growth and differentiation (14van der Velden A.W. Thomas A.A. Int. J. Biochem. Cell Biol. 1999; 31: 87-106Crossref PubMed Scopus (311) Google Scholar, 18Willis A.E. Int. J. Biochem. Cell Biol. 1999; 31: 73-86Crossref PubMed Scopus (117) Google Scholar). The fact that many of these growth factors also bind heparan sulfate suggests the possibility of coordinate regulation of the growth factors and their heparan sulfate co-receptors.A survey of other enzymes involved in heparan sulfate formation indicates that 5′-UTR-dependent regulation may extend to other biosynthetic steps (Table III). Although clones containing the coding sequence for nearly all of the genes are now available (28Esko J.D. Selleck S.B. Annu. Rev. Biochem. 2002; 71: 435-471Crossref PubMed Scopus (1226) Google Scholar), only a few have published sequences that extend significantly into the 5′-UTRs. Notably, the available sequence for EXT1 and EXT2, which make up the copolymerase responsible for the formation of the polysaccharide backbone (35Zak B.M. Crawford B.E. Esko J.D. Biochim. Biophys. Acta. 2002; (in press)PubMed Google Scholar), contains an extended 5′-UTR sequence with high folding energy and numerous uAUGs. These genes are thought to be tumor suppressors based on the loss of heterozygosity in chondrosarcomas and hepatocellular carcinomas (36Hecht J.T. Hogue D. Strong L.C. Hansen M.F. Blanton S.H. Wagner M. Am. J. Hum. Genet. 1995; 56: 1125-1131PubMed Google Scholar,37Piao Z. Kim H. Jeon B.K. Lee W.J. Park C. Cancer (Phila.). 1997; 80: 865-872Crossref PubMed Scopus (83) Google Scholar). Interestingly, the glucosamine 3-O-sulfotransferases 1 and 3A also contain complex 5′-UTRs. These genes play pivotal roles in forming the antithrombin binding sequence and binding sites for herpes simplex virus, respectively (38Shworak N.W. Liu J. Fritze L.M.S. Schwartz J.J. Zhang L.J. Logeart D. Rosenberg R.D. J. Biol. Chem. 1997; 272: 28008-28019Abstract Full Text Full Text PDF PubMed Scopus (149) Google Scholar, 39Shukla D. Liu J. Blaiklock P. Shworak N.W. Bai X.M. Esko J.D. Cohen G.H. Eisenberg R.J. Rosenberg R.D. Spear P.G. Cell. 1999; 99: 13-22Abstract Full Text Full Text PDF PubMed Scopus (855) Google Scholar). The presence of complex 5′-UTRs in mRNAs of various enzymes involved in heparan sulfate biosynthesis suggests that translational control may be exerted at several steps. Translational regulation coupled with transcriptional controls provides a sophisticated way to control the expression of growth-promoting and growth-inhibiting heparan sulfate chains, consistent with the important roles these polysaccharides play in cell biology (19Park P.W. Reizes O. Bernfield M. J. Biol. Chem. 2000; 275: 29923-29926Abstract Full Text Full Text PDF PubMed Scopus (305) Google Scholar, 28Esko J.D. Selleck S.B. Annu. Rev. Biochem. 2002; 71: 435-471Crossref PubMed Scopus (1226) Google Scholar).Table III5′-UTR characteristics of vertebrate genes encoding proteins related to heparan sulfate biosynthesisVertebrate enzymeuAUGsFree energyLengthkcal/molbasesGal-transferase II (β3GalT6)AF092051(human)3−103.6236AK008674(mouse)7−26.9139Copolymerase EXT1S79639 (human)7−200.9652U78539 (mouse)4−90.5325Copolymerase EXT2NM000401 (human)5−199.3488α-GlcNAc-transferase EXT-L1XM001488 (human)7−374.8867α-GlcNAc-transferase EXT-L2AF000416 (human)4−55.4288NM021388 (mouse)0−61273α-GlcNAc-transferase EXT-L3XM027511 (human)10−146.5569NM018788 (mouse)9−128457GlcNAc NDST1U36600 (human)2−184.3424UTR (mouse)2−165.14203-a5′-UTRs confirmed by 5′-RACE.GlcNAc NDST2UTR (mouse)4−271.27203-a5′-UTRs confirmed by 5′-RACE.GlcNAc NDST3AF074924(human)8−93.6404UTR (mouse)5−68.02503-a5′-UTRs confirmed by 5′-RACE.GlcNAc NDST4AB036429(human)9/8−152.9679/3863-a5′-UTRs confirmed by 5′-RACE.AB036838 (mouse), long and short transcripts10/9−172.2/−95.7669/3773-a5′-UTRs confirmed by 5′-RACE.Glucuronyl C5-epimeraseAAG42004(mouse)2−79.72323-a5′-UTRs confirmed by 5′-RACE.Heparan-sulfate 2OST3-b2OST, 2-O-sulfotransferase; 6OST1, 6-O-sulfotransferase 1; 3OST, 3-O-sulfotransferase.AB007917(human)1−125.42383-a5′-UTRs confirmed by 5′-RACE.Heparan sulfate-GlcN 6OST13-b2OST, 2-O-sulfotransferase; 6OST1, 6-O-sulfotransferase 1; 3OST, 3-O-sulfotransferase.AB006179(human)1−57.5112AB024566(mouse)1−151.0280Heparan sulfate-GlcN 3OST13-b2OST, 2-O-sulfotransferase; 6OST1, 6-O-sulfotransferase 1; 3OST, 3-O-sulfotransferase.AF019386 (human)2−37.4119AF019385 (mouse)6−90.4323Heparan sulfate-GlcN 3OST2AF105375 (human)2−22.973Heparan sulfate-GlcN 3OST3AAF105376(human)5−346.9798Heparan sulfate-GlcN 3OST3BAF105377 (human)3−132.33313-a 5′-UTRs confirmed by 5′-RACE.3-b 2OST, 2-O-sulfotransferase; 6OST1, 6-O-sulfotransferase 1; 3OST, 3-O-sulfotransferase. Open table in a new tab Heparan sulfate and heparin bind a variety of growth factors, enzymes, and extracellular matrix proteins through specific arrangements of variably sulfated sugar residues (1****Google Scholar, 2Capila I. Linhardt R.J. Angew. Chem. Int. Ed. Engl. 2002; 41: 391-412Crossref PubMed Scopus (1524) Google Scholar). Several sulfotransferases guide the formation of these binding sequences. The initial reactions involve removal of acetyl groups from GlcNAc residues, followed by sulfation of the amino group catalyzed by one or more GlcNAc N-deacetylase/N-sulfotransferases (NDSTs)1 (3Hashimoto Y. Orellana A. Gil G. Hirschberg C.B. J. Biol. Chem. 1992; 267: 15744-15750Abstract Full Text PDF PubMed Google Scholar, 4Orellana A. Hirschberg C.B. Wei Z. Swiedler S.J. Ishihara M. J. Biol. Chem. 1994; 269: 2270-2276Abstract Full Text PDF PubMed Google Scholar, 5Eriksson I. Sandbäck D., Ek, B. Lindahl U. Kjellén L. J. Biol. Chem. 1994; 269: 10438-10443Abstract Full Text PDF PubMed Google Scholar, 6Aikawa J. Esko J.D. J. Biol. Chem. 1999; 274: 2690-2695Abstract Full Text Full Text PDF PubMed Scopus (130) Google Scholar, 7Aikawa J. Grobe K. Tsujimoto M. Esko J.D. J. Biol. Chem. 2001; 276: 5876-5882Abstract Full Text Full Text PDF PubMed Scopus (186) Google Scholar). Four isozymes exist in vertebrates, whereas only one is expressed inDrosophila melanogaster and Caenorhabditis elegans. The importance of NDST1 in normal physiology in mice has been established by targeted disruption of the gene, which results in embryonic and neonatal lethality due to multiple organ defects (8Ringvall M. Ledin J. Holmborn K. Van Kuppevelt T. Ellin F. Eriksson I. Olofsson A.M. Kjellén L. Forsberg E. J. Biol. Chem. 2000; 275: 25926-25930Abstract Full Text Full Text PDF PubMed Scopus (206) Google Scholar, 9Fan G. Xiao L. Cheng L. Wang X. Sun B. Hu G. FEBS Lett. 2000; 467: 7-11Crossref PubMed Scopus (143) Google Scholar). In contrast, disruption of NDST2 affects only connective tissue-type mast cells, resulting in a storage deficiency for proteases and histamine due to lack of heparin (10Forsberg E. Pejler G. Ringvall M. Lunderius C. Tomasini-Johansson B. Kusche-Gullberg M. Eriksson I. Ledin J. Hellman L. Kjellén L. Nature. 1999; 400: 773-776Crossref PubMed Scopus (403) Google Scholar, 11Humphries D.E. Wong G.W. Friend D.S. Gurish M.F. Qiu W.T. Huang C.F. Sharpe A.H. Stevens R.L. Nature. 1999; 400: 769-772Crossref PubMed Scopus (359) Google Scholar). The function of NDST3 and NDST4 are not yet established. As NDST1 and NDST2 show ubiquitous mRNA expression (7Aikawa J. Grobe K. Tsujimoto M. Esko J.D. J. Biol. Chem. 2001; 276: 5876-5882Abstract Full Text Full Text PDF PubMed Scopus (186) Google Scholar), the strikingly different phenotypes of the mutants point toward other modes of enzyme regulation, possibly at the post-transcriptional or post-translational level. The leader sequences of most mRNAs are usually short (50–70 nucleotides), contain 5′-m7G cap structures, and lack highly organized secondary structures (12Gray N.K. Wickens M. Annu. Rev. Cell Dev. Biol. 1998; 14: 399-458Crossref PubMed Scopus (447) Google Scholar, 13Cazzola M. Skoda R.C. Blood. 2000; 95: 3280-3288Crossref PubMed Google Scholar, 14van der Velden A.W. Thomas A.A. Int. J. Biochem. Cell Biol. 1999; 31: 87-106Crossref PubMed Scopus (311) Google Scholar). These features are compatible with efficient translation by the ribosomal scanning mechanism, which is considered valid for the majority of cellular and viral mRNAs (15Kozak M. Gene (Amst.). 1999; 234: 187-208Crossref PubMed Scopus (1121) Google Scholar, 16Pestova T.V. Kolupaeva V.G. Lomakin I.B. Pilipenko E.V. Shatsky I.N. Agol V.I. Hellen C.U. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 7029-7036Crossref PubMed Scopus (596) Google Scholar). Translational control of these transcripts is most commonly exerted at the rate-limiting initiation phase. The 5′-cap structure (m7GpppN) attracts the eukaryotic initiation factor (eIF) 4F complex, composed of the cap-binding protein eIF4E, the RNA-dependent ATPase (helicase) eIF4A, and the modular factor eIF4G. The small (40 S) ribosomal subunit binds to the 5′-end of the RNA·eIF4F complex along with eIF3 and the ternary complex of eIF2, GTP, and Met-tRNAi. The resulting 48 S complex then advances through the initiation cycle, resulting in lateral movement of the 43 S unit along the mRNA (scanning), usually to the first AUG triplet, which serves as the initiation codon. Upon dissociation of initiation factors, the large (60 S) subunit joins to form the 80 S ribosome. The eIF4E-eIF4G interaction with RNA is a limited controlled process, reduced by hypophosphorylation of eIF4E or the competitive inhibitors (eIF4E-binding proteins) or by phosphorylation of eIF2 (17Sonenberg N. Newgard C.B. Science. 2001; 293: 818-819Crossref PubMed Scopus (18) Google Scholar). Rapid amplification of the 5′-ends of murine and human NDST cDNAs (5′-RACE) revealed long, structured, and AUG-rich 5′-untranslated regions (5′-UTRs). These types of UTRs are incompatible with the described scanning mode for translation initiation due to secondary structure constraints and abortive initiation at upstream AUG codons (uAUGs) located 5′ to the actual initiation site for translation. Complex 5′-UTRs are found in ∼10% of mRNAs, often in those encoding regulatory proteins such as proto-oncogenes (18Willis A.E. Int. J. Biochem. Cell Biol. 1999; 31: 73-86Crossref PubMed Scopus (117) Google Scholar), growth factors (insulin-like growth factor II, platelet-derived growth factor 2, transforming growth factor-β, fibroblast growth factor 2, and vascular endothelial growth factor) and their receptors, and homeodomain proteins (14van der Velden A.W. Thomas A.A. Int. J. Biochem. Cell Biol. 1999; 31: 87-106Crossref PubMed Scopus (311) Google Scholar). In some of these cases, initiation is driven from internal ribosome entry sites (IRES) that allow for cap-independent translation, thus affording another level of control over protein expression. Specific sulfated sugar sequences in cell-surface heparan sulfate can bind to many of these growth factors and facilitate signaling (19Park P.W. Reizes O. Bernfield M. J. Biol. Chem. 2000; 275: 29923-29926Abstract Full Text Full Text PDF PubMed Scopus (305) Google Scholar). Thus, cap-independent translational regulation of NDST expression and the growth factors may be coordinated to fine-tune growth responses in target cells. Here, we report the comparison of the 5′-UTRs of the NDSTs, their regulatory properties with respect to translation, and the presence of complex UTRs in other enzymes involved in heparan sulfate biosynthesis. Our findings suggest the possibility of coordinated translational control over several components of growth factor signaling pathways. DISCUSSIONNDST1–4 are important enzymes in the formation of heparan sulfate because all other modifications of the chains (O-sulfation and epimerization of uronic acids) depend on the conversion of GlcNAc to GlcNSO3 residues (27Esko J.D. Lindahl U. J. Clin. Invest. 2001; 108: 169-173Crossref PubMed Scopus (782) Google Scholar, 28Esko J.D. Selleck S.B. Annu. Rev. Biochem. 2002; 71: 435-471Crossref PubMed Scopus (1226) Google Scholar). The four isozymes have different biochemical properties that manifest as variation in substrate preference and relative activity (7Aikawa J. Grobe K. Tsujimoto M. Esko J.D. J. Biol. Chem. 2001; 276: 5876-5882Abstract Full Text Full Text PDF PubMed Scopus (186) Google Scholar). The four isozymes are also differentially expressed in tissues, based on PCR analysis or Northern blots of mRNA (7Aikawa J. Grobe K. Tsujimoto M. Esko J.D. J. Biol. Chem. 2001; 276: 5876-5882Abstract Full Text Full Text PDF PubMed Scopus (186) Google Scholar, 23Kusche-Gullberg M. Eriksson I. Pikas D.S. Kjellén L. J. Biol. Chem. 1998; 273: 11902-11907Abstract Full Text Full Text PDF PubMed Scopus (78) Google Scholar). In this study, we have shown that the expression pattern of the NDST isozymes may also reflect differential translation in tissues dependent on their 5′-UTRs. The evidence is based on (i) the length and complexity of the 5′-UTRs from mice and human homologs, (ii) differential expression of reporter constructs in which the 5′-UTRs were ligated to EGFP, (iii) mutational analysis, (iv) the effects of serum and rapamycin on translation, and (v) differences in NDST protein expression relative to mRNA in brain and embryonic extracts. These findings demonstrate an important caveat in interpreting levels of enzyme expression from NDST mRNA levels. The data suggest that the level of enzyme activity contributed by the various isoforms will depend on both transcription and translation.The presence of regulatory elements in the 5′-UTRs of the NDSTs may explain some of the surprising results obtained in genetic studies of mice in which individual NDSTs have been inactivated (29Grobe K. Ledin J. Ringvall M. Holborn K. Forsberg E. Esko J.D. Kjellen L. Biochim. Biophys. Acta. 2002; (in press)PubMed Google Scholar). NDST1 and NDST2 mRNAs are abundantly expressed in nearly all tissues (7Aikawa J. Grobe K. Tsujimoto M. Esko J.D. J. Biol. Chem. 2001; 276: 5876-5882Abstract Full Text Full Text PDF PubMed Scopus (186) Google Scholar, 23Kusche-Gullberg M. Eriksson I. Pikas D.S. Kjellén L. J. Biol. Chem. 1998; 273: 11902-11907Abstract Full Text Full Text PDF PubMed Scopus (78) Google Scholar), and both enzymes can participate in heparan sulfate formation (6Aikawa J. Esko J.D. J. Biol. Chem. 1999; 274: 2690-2695Abstract Full Text Full Text PDF PubMed Scopus (130) Google Scholar, 30Ishihara M. Kiefer M.C. Barr P.J. Guo Y. Swiedler S.J. Anal. Biochem. 1992; 206: 400-407Crossref PubMed Scopus (16) Google Scholar). Inactivation of NDST1 has a profound effect on development (8Ringvall M. Ledin J. Holmborn K. Van Kuppevelt T. Ellin F. Eriksson I. Olofsson A.M. Kjellén L. Forsberg E. J. Biol. Chem. 2000; 275: 25926-25930Abstract Full Text Full Text PDF PubMed Scopus (206) Google Scholar, 9Fan G. Xiao L. Cheng L. Wang X. Sun B. Hu G. FEBS Lett. 2000; 467: 7-11Crossref PubMed Scopus (143) Google Scholar), suggesting a prominent role for this enzyme in many tissues. The restricted phenotype of NDST2 is surprising, however, given the widespread expression of its mRNA. The data presented here suggest that NDST2 enzyme expression may not occur even though abundant message is present. The lack of phenotype in those tissues containing both enzymes suggests that the two activities may be redundant in certain contexts.Having 5′-UTR sequences with embedded IRES motifs allows cells to regulate the level of translation, which is especially relevant for regulatory factors. One of the best studied systems involves cellular iron homeostasis, which is regulated by translational control of the transferrin receptor and ferritin expression (31Muckenthaler M. Gray N.K. Hentze M.W. Mol. Cell. 1998; 2: 383-388Abstract Full Text Full Text PDF PubMed Scopus (208) Google Scholar). Cytoplasmic RNA-binding proteins that bind iron (iron regulatory proteins) interact with specific stem-loop structures (iron response elements) in the 5′-UTRs of the corresponding mRNAs, thus regulating expression of the genes according to the iron status of the cytosol. Whether similar stem-loop structures and regulatory proteins exist for the UTRs of the NDSTs awaits further analysis.Interestingly, the cell type-specific translational regulation of NDST isozymes strongly resembles the pattern of regulation of growth factors, some of which bind heparan sulfate in a sulfation-dependent manner. For example, tissue-specific regulation of fibroblast growth factor 2 translation appears to be dependent on its 5′-UTR (32Creancier L. Morello D. Mercier P. Prats A.C. J. Cell Biol. 2000; 150: 275-281Crossref PubMed Scopus (118) Google Scholar). Similarly, vascular endothelial growth factor regulation by hypoxia (33Stein I. Itin A. Einat P. Skaliter R. Grossman Z. Keshet E. Mol. Cell. Biol. 1998; 18: 3112-3119Crossref PubMed Scopus (420) Google Scholar) and platelet-derived growth factor 2/c-sis expression that occurs during differentiation (34Bernstein J. Sella O., Le, S.Y. Elroy-Stein O. J. Biol. Chem. 1997; 272: 9356-9362Abstract Full Text Full Text PDF PubMed Scopus (153) Google Scholar) are regulated in this way as well. The 5′-UTRs of these and other growth factors are often long, have high folding energies, contain uAUGs, and are often equipped with an IRES, which together can contribute to precise expression levels during growth and differentiation (14van der Velden A.W. Thomas A.A. Int. J. Biochem. Cell Biol. 1999; 31: 87-106Crossref PubMed Scopus (311) Google Scholar, 18Willis A.E. Int. J. Biochem. Cell Biol. 1999; 31: 73-86Crossref PubMed Scopus (117) Google Scholar). The fact that many of these growth factors also bind heparan sulfate suggests the possibility of coordinate regulation of the growth factors and their heparan sulfate co-receptors.A survey of other enzymes involved in heparan sulfate formation indicates that 5′-UTR-dependent regulation may extend to other biosynthetic steps (Table III). Although clones containing the coding sequence for nearly all of the genes are now available (28Esko J.D. Selleck S.B. Annu. Rev. Biochem. 2002; 71: 435-471Crossref PubMed Scopus (1226) Google Scholar), only a few have published sequences that extend significantly into the 5′-UTRs. Notably, the available sequence for EXT1 and EXT2, which make up the copolymerase responsible for the formation of the polysaccharide backbone (35Zak B.M. Crawford B.E. Esko J.D. Biochim. Biophys. Acta. 2002; (in press)PubMed Google Scholar), contains an extended 5′-UTR sequence with high folding energy and numerous uAUGs. These genes are thought to be tumor suppressors based on the loss of heterozygosity in chondrosarcomas and hepatocellular carcinomas (36Hecht J.T. Hogue D. Strong L.C. Hansen M.F. Blanton S.H. Wagner M. Am. J. Hum. Genet. 1995; 56: 1125-1131PubMed Google Scholar,37Piao Z. Kim H. Jeon B.K. Lee W.J. Park C. Cancer (Phila.). 1997; 80: 865-872Crossref PubMed Scopus (83) Google Scholar). Interestingly, the glucosamine 3-O-sulfotransferases 1 and 3A also contain complex 5′-UTRs. These genes play pivotal roles in forming the antithrombin binding sequence and binding sites for herpes simplex virus, respectively (38Shworak N.W. Liu J. Fritze L.M.S. Schwartz J.J. Zhang L.J. Logeart D. Rosenberg R.D. J. Biol. Chem. 1997; 272: 28008-28019Abstract Full Text Full Text PDF PubMed Scopus (149) Google Scholar, 39Shukla D. Liu J. Blaiklock P. Shworak N.W. Bai X.M. Esko J.D. Cohen G.H. Eisenberg R.J. Rosenberg R.D. Spear P.G. Cell. 1999; 99: 13-22Abstract Full Text Full Text PDF PubMed Scopus (855) Google Scholar). The presence of complex 5′-UTRs in mRNAs of various enzymes involved in heparan sulfate biosynthesis suggests that translational control may be exerted at several steps. Translational regulation coupled with transcriptional controls provides a sophisticated way to control the expression of growth-promoting and growth-inhibiting heparan sulfate chains, consistent with the important roles these polysaccharides play in cell biology (19Park P.W. Reizes O. Bernfield M. J. Biol. Chem. 2000; 275: 29923-29926Abstract Full Text Full Text PDF PubMed Scopus (305) Google Scholar, 28Esko J.D. Selleck S.B. Annu. Rev. Biochem. 2002; 71: 435-471Crossref PubMed Scopus (1226) Google Scholar).Table III5′-UTR characteristics of vertebrate genes encoding proteins related to heparan sulfate biosynthesisVertebrate enzymeuAUGsFree energyLengthkcal/molbasesGal-transferase II (β3GalT6)AF092051(human)3−103.6236AK008674(mouse)7−26.9139Copolymerase EXT1S79639 (human)7−200.9652U78539 (mouse)4−90.5325Copolymerase EXT2NM000401 (human)5−199.3488α-GlcNAc-transferase EXT-L1XM001488 (human)7−374.8867α-GlcNAc-transferase EXT-L2AF000416 (human)4−55.4288NM021388 (mouse)0−61273α-GlcNAc-transferase EXT-L3XM027511 (human)10−146.5569NM018788 (mouse)9−128457GlcNAc NDST1U36600 (human)2−184.3424UTR (mouse)2−165.14203-a5′-UTRs confirmed by 5′-RACE.GlcNAc NDST2UTR (mouse)4−271.27203-a5′-UTRs confirmed by 5′-RACE.GlcNAc NDST3AF074924(human)8−93.6404UTR (mouse)5−68.02503-a5′-UTRs confirmed by 5′-RACE.GlcNAc NDST4AB036429(human)9/8−152.9679/3863-a5′-UTRs confirmed by 5′-RACE.AB036838 (mouse), long and short transcripts10/9−172.2/−95.7669/3773-a5′-UTRs confirmed by 5′-RACE.Glucuronyl C5-epimeraseAAG42004(mouse)2−79.72323-a5′-UTRs confirmed by 5′-RACE.Heparan-sulfate 2OST3-b2OST, 2-O-sulfotransferase; 6OST1, 6-O-sulfotransferase 1; 3OST, 3-O-sulfotransferase.AB007917(human)1−125.42383-a5′-UTRs confirmed by 5′-RACE.Heparan sulfate-GlcN 6OST13-b2OST, 2-O-sulfotransferase; 6OST1, 6-O-sulfotransferase 1; 3OST, 3-O-sulfotransferase.AB006179(human)1−57.5112AB024566(mouse)1−151.0280Heparan sulfate-GlcN 3OST13-b2OST, 2-O-sulfotransferase; 6OST1, 6-O-sulfotransferase 1; 3OST, 3-O-sulfotransferase.AF019386 (human)2−37.4119AF019385 (mouse)6−90.4323Heparan sulfate-GlcN 3OST2AF105375 (human)2−22.973Heparan sulfate-GlcN 3OST3AAF105376(human)5−346.9798Heparan sulfate-GlcN 3OST3BAF105377 (human)3−132.33313-a 5′-UTRs confirmed by 5′-RACE.3-b 2OST, 2-O-sulfotransferase; 6OST1, 6-O-sulfotransferase 1; 3OST, 3-O-sulfotransferase. Open table in a new tab NDST1–4 are important enzymes in the formation of heparan sulfate because all other modifications of the chains (O-sulfation and epimerization of uronic acids) depend on the conversion of GlcNAc to GlcNSO3 residues (27Esko J.D. Lindahl U. J. Clin. Invest. 2001; 108: 169-173Crossref PubMed Scopus (782) Google Scholar, 28Esko J.D. Selleck S.B. Annu. Rev. Biochem. 2002; 71: 435-471Crossref PubMed Scopus (1226) Google Scholar). The four isozymes have different biochemical properties that manifest as variation in substrate preference and relative activity (7Aikawa J. Grobe K. Tsujimoto M. Esko J.D. J. Biol. Chem. 2001; 276: 5876-5882Abstract Full Text Full Text PDF PubMed Scopus (186) Google Scholar). The four isozymes are also differentially expressed in tissues, based on PCR analysis or Northern blots of mRNA (7Aikawa J. Grobe K. Tsujimoto M. Esko J.D. J. Biol. Chem. 2001; 276: 5876-5882Abstract Full Text Full Text PDF PubMed Scopus (186) Google Scholar, 23Kusche-Gullberg M. Eriksson I. Pikas D.S. Kjellén L. J. Biol. Chem. 1998; 273: 11902-11907Abstract Full Text Full Text PDF PubMed Scopus (78) Google Scholar). In this study, we have shown that the expression pattern of the NDST isozymes may also reflect differential translation in tissues dependent on their 5′-UTRs. The evidence is based on (i) the length and complexity of the 5′-UTRs from mice and human homologs, (ii) differential expression of reporter constructs in which the 5′-UTRs were ligated to EGFP, (iii) mutational analysis, (iv) the effects of serum and rapamycin on translation, and (v) differences in NDST protein expression relative to mRNA in brain and embryonic extracts. These findings demonstrate an important caveat in interpreting levels of enzyme expression from NDST mRNA levels. The data suggest that the level of enzyme activity contributed by the various isoforms will depend on both transcription and translation. The presence of regulatory elements in the 5′-UTRs of the NDSTs may explain some of the surprising results obtained in genetic studies of mice in which individual NDSTs have been inactivated (29Grobe K. Ledin J. Ringvall M. Holborn K. Forsberg E. Esko J.D. Kjellen L. Biochim. Biophys. Acta. 2002; (in press)PubMed Google Scholar). NDST1 and NDST2 mRNAs are abundantly expressed in nearly all tissues (7Aikawa J. Grobe K. Tsujimoto M. Esko J.D. J. Biol. Chem. 2001; 276: 5876-5882Abstract Full Text Full Text PDF PubMed Scopus (186) Google Scholar, 23Kusche-Gullberg M. Eriksson I. Pikas D.S. Kjellén L. J. Biol. Chem. 1998; 273: 11902-11907Abstract Full Text Full Text PDF PubMed Scopus (78) Google Scholar), and both enzymes can participate in heparan sulfate formation (6Aikawa J. Esko J.D. J. Biol. Chem. 1999; 274: 2690-2695Abstract Full Text Full Text PDF PubMed Scopus (130) Google Scholar, 30Ishihara M. Kiefer M.C. Barr P.J. Guo Y. Swiedler S.J. Anal. Biochem. 1992; 206: 400-407Crossref PubMed Scopus (16) Google Scholar). Inactivation of NDST1 has a profound effect on development (8Ringvall M. Ledin J. Holmborn K. Van Kuppevelt T. Ellin F. Eriksson I. Olofsson A.M. Kjellén L. Forsberg E. J. Biol. Chem. 2000; 275: 25926-25930Abstract Full Text Full Text PDF PubMed Scopus (206) Google Scholar, 9Fan G. Xiao L. Cheng L. Wang X. Sun B. Hu G. FEBS Lett. 2000; 467: 7-11Crossref PubMed Scopus (143) Google Scholar), suggesting a prominent role for this enzyme in many tissues. The restricted phenotype of NDST2 is surprising, however, given the widespread expression of its mRNA. The data presented here suggest that NDST2 enzyme expression may not occur even though abundant message is present. The lack of phenotype in those tissues containing both enzymes suggests that the two activities may be redundant in certain contexts. Having 5′-UTR sequences with embedded IRES motifs allows cells to regulate the level of translation, which is especially relevant for regulatory factors. One of the best studied systems involves cellular iron homeostasis, which is regulated by translational control of the transferrin receptor and ferritin expression (31Muckenthaler M. Gray N.K. Hentze M.W. Mol. Cell. 1998; 2: 383-388Abstract Full Text Full Text PDF PubMed Scopus (208) Google Scholar). Cytoplasmic RNA-binding proteins that bind iron (iron regulatory proteins) interact with specific stem-loop structures (iron response elements) in the 5′-UTRs of the corresponding mRNAs, thus regulating expression of the genes according to the iron status of the cytosol. Whether similar stem-loop structures and regulatory proteins exist for the UTRs of the NDSTs awaits further analysis. Interestingly, the cell type-specific translational regulation of NDST isozymes strongly resembles the pattern of regulation of growth factors, some of which bind heparan sulfate in a sulfation-dependent manner. For example, tissue-specific regulation of fibroblast growth factor 2 translation appears to be dependent on its 5′-UTR (32Creancier L. Morello D. Mercier P. Prats A.C. J. Cell Biol. 2000; 150: 275-281Crossref PubMed Scopus (118) Google Scholar). Similarly, vascular endothelial growth factor regulation by hypoxia (33Stein I. Itin A. Einat P. Skaliter R. Grossman Z. Keshet E. Mol. Cell. Biol. 1998; 18: 3112-3119Crossref PubMed Scopus (420) Google Scholar) and platelet-derived growth factor 2/c-sis expression that occurs during differentiation (34Bernstein J. Sella O., Le, S.Y. Elroy-Stein O. J. Biol. Chem. 1997; 272: 9356-9362Abstract Full Text Full Text PDF PubMed Scopus (153) Google Scholar) are regulated in this way as well. The 5′-UTRs of these and other growth factors are often long, have high folding energies, contain uAUGs, and are often equipped with an IRES, which together can contribute to precise expression levels during growth and differentiation (14van der Velden A.W. Thomas A.A. Int. J. Biochem. Cell Biol. 1999; 31: 87-106Crossref PubMed Scopus (311) Google Scholar, 18Willis A.E. Int. J. Biochem. Cell Biol. 1999; 31: 73-86Crossref PubMed Scopus (117) Google Scholar). The fact that many of these growth factors also bind heparan sulfate suggests the possibility of coordinate regulation of the growth factors and their heparan sulfate co-receptors. A survey of other enzymes involved in heparan sulfate formation indicates that 5′-UTR-dependent regulation may extend to other biosynthetic steps (Table III). Although clones containing the coding sequence for nearly all of the genes are now available (28Esko J.D. Selleck S.B. Annu. Rev. Biochem. 2002; 71: 435-471Crossref PubMed Scopus (1226) Google Scholar), only a few have published sequences that extend significantly into the 5′-UTRs. Notably, the available sequence for EXT1 and EXT2, which make up the copolymerase responsible for the formation of the polysaccharide backbone (35Zak B.M. Crawford B.E. Esko J.D. Biochim. Biophys. Acta. 2002; (in press)PubMed Google Scholar), contains an extended 5′-UTR sequence with high folding energy and numerous uAUGs. These genes are thought to be tumor suppressors based on the loss of heterozygosity in chondrosarcomas and hepatocellular carcinomas (36Hecht J.T. Hogue D. Strong L.C. Hansen M.F. Blanton S.H. Wagner M. Am. J. Hum. Genet. 1995; 56: 1125-1131PubMed Google Scholar,37Piao Z. Kim H. Jeon B.K. Lee W.J. Park C. Cancer (Phila.). 1997; 80: 865-872Crossref PubMed Scopus (83) Google Scholar). Interestingly, the glucosamine 3-O-sulfotransferases 1 and 3A also contain complex 5′-UTRs. These genes play pivotal roles in forming the antithrombin binding sequence and binding sites for herpes simplex virus, respectively (38Shworak N.W. Liu J. Fritze L.M.S. Schwartz J.J. Zhang L.J. Logeart D. Rosenberg R.D. J. Biol. Chem. 1997; 272: 28008-28019Abstract Full Text Full Text PDF PubMed Scopus (149) Google Scholar, 39Shukla D. Liu J. Blaiklock P. Shworak N.W. Bai X.M. Esko J.D. Cohen G.H. Eisenberg R.J. Rosenberg R.D. Spear P.G. Cell. 1999; 99: 13-22Abstract Full Text Full Text PDF PubMed Scopus (855) Google Scholar). The presence of complex 5′-UTRs in mRNAs of various enzymes involved in heparan sulfate biosynthesis suggests that translational control may be exerted at several steps. Translational regulation coupled with transcriptional controls provides a sophisticated way to control the expression of growth-promoting and growth-inhibiting heparan sulfate chains, consistent with the important roles these polysaccharides play in cell biology (19Park P.W. Reizes O. Bernfield M. J. Biol. Chem. 2000; 275: 29923-29926Abstract Full Text Full Text PDF PubMed Scopus (305) Google Scholar, 28Esko J.D. Selleck S.B. Annu. Rev. Biochem. 2002; 71: 435-471Crossref PubMed Scopus (1226) Google Scholar).

Referência(s)