Upstream Elements Present in the 3′-Untranslated Region of Collagen Genes Influence the Processing Efficiency of Overlapping Polyadenylation Signals

Abstract

3′-Untranslated regions (UTRs) of genes often contain key regulatory elements involved in gene expression control. A high degree of evolutionary conservation in regions of the 3′-UTR suggests important, conserved elements. In particular, we are interested in those elements involved in regulation of 3′ end formation. In addition to canonical sequence elements, auxiliary sequences likely play an important role in determining the polyadenylation efficiency of mammalian pre-mRNAs. We identified highly conserved sequence elements upstream of the AAUAAA in three human collagen genes, COL1A1, COL1A2, and COL2A1, and demonstrate that these upstream sequence elements (USEs) influence polyadenylation efficiency. Mutation of the USEs decreases polyadenylation efficiency both in vitro and in vivo, and inclusion of competitor oligoribonucleotides representing the USEs specifically inhibit polyadenylation. We have also shown that insertion of a USE into a weak polyadenylation signal can enhance 3′ end formation. Close inspection of the COL1A2 3′-UTR reveals an unusual feature of two closely spaced, competing polyadenylation signals. Taken together, these data demonstrate that USEs are important auxiliary polyadenylation elements in mammalian genes. 3′-Untranslated regions (UTRs) of genes often contain key regulatory elements involved in gene expression control. A high degree of evolutionary conservation in regions of the 3′-UTR suggests important, conserved elements. In particular, we are interested in those elements involved in regulation of 3′ end formation. In addition to canonical sequence elements, auxiliary sequences likely play an important role in determining the polyadenylation efficiency of mammalian pre-mRNAs. We identified highly conserved sequence elements upstream of the AAUAAA in three human collagen genes, COL1A1, COL1A2, and COL2A1, and demonstrate that these upstream sequence elements (USEs) influence polyadenylation efficiency. Mutation of the USEs decreases polyadenylation efficiency both in vitro and in vivo, and inclusion of competitor oligoribonucleotides representing the USEs specifically inhibit polyadenylation. We have also shown that insertion of a USE into a weak polyadenylation signal can enhance 3′ end formation. Close inspection of the COL1A2 3′-UTR reveals an unusual feature of two closely spaced, competing polyadenylation signals. Taken together, these data demonstrate that USEs are important auxiliary polyadenylation elements in mammalian genes. downstream element cleavage and polyadenylation specificity factor cleavage stimulatory factor upstream sequence element bovine growth hormone Poly(A) tails are found on the 3′ end of nearly every fully processed eukaryotic mRNA. The poly(A) tail has been suggested to influence mRNA stability, translation, and transport (for review, see Refs. 1Lewis J. Gunderson S. Mattaj I.W. J. Cell Sci. Suppl. 1995; 19: 13-19Crossref PubMed Google Scholar, 2Jacobson A. Peltz S.W. Annu. Rev. Biochem. 1996; 65: 693-739Crossref PubMed Scopus (576) Google Scholar, 3Sachs A.B. Sarnow P. Hentze M.W. Cell. 1997; 89: 831-838Abstract Full Text Full Text PDF PubMed Scopus (588) Google Scholar, 4Wickens M. Anderson P. Jackson R.J. Curr. Opin. Genet. Dev. 1997; 7: 220-232Crossref PubMed Scopus (283) Google Scholar). Polyadenylation is a two-step process that first involves specific endonucleolytic cleavage at a site determined by binding of polyadenylation factors (for review, see Refs. 5Proudfoot N. Cell. 1996; 87: 779-781Abstract Full Text Full Text PDF PubMed Scopus (66) Google Scholar, 6Proudfoot N. Trends Biochem. Sci. 2000; 25: 290-293Abstract Full Text Full Text PDF PubMed Scopus (140) Google Scholar, 7Keller W. Minvielle-Sebastia L. Curr. Opin. Cell Biol. 1997; 9: 329-336Crossref PubMed Scopus (105) Google Scholar, 8Colgan D.F. Manley J.L. Genes Dev. 1997; 11: 2755-2766Crossref PubMed Scopus (629) Google Scholar, 9Zhao J. Hyman L. Moore C. Microbiol. Mol. Biol. Rev. 1999; 63: 405-445Crossref PubMed Google Scholar, 10Shatkin A.J. Manley J.L. Nat. Struct. Biol. 2000; 7: 838-842Crossref PubMed Scopus (254) Google Scholar). The second step involves polymerization of an adenosine tail to an average length of ∼200 residues. These steps are tightly coupled processes since reaction intermediates are not detectable under normal conditions. The vast majority of eukaryotic polyadenylation signals contain the consensus sequence AAUAAA between 10 and 35 nucleotides upstream of the actual cleavage and polyadenylation site. In addition, sequences 10–30 nucleotides downstream of the cleavage site are known to be involved in directing polyadenylation (Refs. 11Chen F. MacDonald C.C. Wilusz J. Nucleic Acids Res. 1995; 23: 2614-2620Crossref PubMed Scopus (171) Google Scholar, 12Chou Z.F. Chen F. Wilusz J. Nucleic Acids Res. 1994; 22: 2525-2531Crossref PubMed Scopus (56) Google Scholar, 13Graber J.H. Cantor C.R. Mohr S.C. Smith T.F. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 14055-14060Crossref PubMed Scopus (213) Google Scholar and references therein). These downstream elements (DSEs)1can be characterized as a block containing 4 of 5 uracil (U) residues. These two sequence elements recruit cleavage and polyadenylation specificity factor (CPSF) and cleavage stimulatory factor (CstF), respectively, to define the cleavage site; therefore, mutations within these sequences abolish polyadenylation. The intricate nature of this process implies that polyadenylation might be a useful mechanism to regulate gene expression. The efficiency of 3′ end processing is a level at which regulation can occur. Because most pre-mRNAs in the cell are not efficiently processed, even small changes in the overall processing efficiency of a particular pre-mRNA may have a substantial effect on gene expression. Experimental evidence has demonstrated that poly(A) signal strength directly influences the amount of mature, exported mRNA (14Denome R.M. Cole C.N. Mol. Cell. Biol. 1988; 8: 4829-4839Crossref PubMed Scopus (44) Google Scholar, 15Edwalds-Gilbert G. Prescott J. Falck-Pederson E. Mol. Cell. Biol. 1993; 13: 3472-3480Crossref PubMed Scopus (58) Google Scholar). Poly(A) signal strength is also directly correlated with the rapid assembly of the polyadenylation machinery on the nascent transcript (16Chao L.C. Jamil A. Kim S.J. Huang L. Martinson H.G. Mol. Cell. Biol. 1999; 19: 5588-5600Crossref PubMed Scopus (38) Google Scholar) and with transcription termination efficiency (17Osheim Y.N. Proudfoot N.J. Beyer A.L. Mol. Cell. 1999; 3: 379-387Abstract Full Text Full Text PDF PubMed Scopus (67) Google Scholar). Detailed mechanistic studies on regulation of polyadenylation are now emerging, revealing both cis- and trans-acting factors (for review, see Refs. 5Proudfoot N. Cell. 1996; 87: 779-781Abstract Full Text Full Text PDF PubMed Scopus (66) Google Scholar, 9Zhao J. Hyman L. Moore C. Microbiol. Mol. Biol. Rev. 1999; 63: 405-445Crossref PubMed Google Scholar, and 18Edwalds-Gilbert G. Veraldi K.L. Milcarek C. Nucleic Acids Res. 1997; 25: 2547-2561Crossref PubMed Scopus (454) Google Scholar). In addition to the sequence elements previously described, elements upstream of the AAUAAA sequence and downstream of the DSE (see Refs. 19Chen F. Wilusz J. Nucleic Acids Res. 1998; 26: 2891-2898Crossref PubMed Scopus (48) Google Scholar and 20Arhin G.K. Boots M. Bagga P.S. Milcarek C. Wilusz J. Nucleic Acids Res. 2002; 30: 1842-1850Crossref PubMed Scopus (77) Google Scholar and references therein) have been identified as auxiliary cis-acting polyadenylation efficiency elements. Such elements may play an important role in modulating the overall processing efficiency. Upstream elements (USEs) have been characterized in viral and cellular systems, including SV40 (21Schek N. Cooke C. Alwine J.C. Mol. Cell. Biol. 1992; 12: 5386-5393Crossref PubMed Scopus (71) Google Scholar), human immunodeficiency virus (22Brown P.H. Tiley L.S. Cullen B.R. J. Virol. 1991; 65: 3340-3343Crossref PubMed Google Scholar, 23Valsamakis A. Zeichner S. Carswell S. Alwine J.C. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 2108-2112Crossref PubMed Scopus (89) Google Scholar, 24Valsamakis A. Schek N. Alwine J.C. Mol. Cell. Biol. 1992; 12: 3699-3705Crossref PubMed Scopus (62) Google Scholar, 25DeZazzo J.D. Kilpatrick J.E. Imperiale M.J. Mol. Cell. Biol. 1991; 11: 1624-1630Crossref PubMed Scopus (55) Google Scholar, 26Gilmartin G.M. Fleming E.S. Oetjen J. EMBO J. 1992; 11: 4419-4428Crossref PubMed Scopus (70) Google Scholar, 27Gilmartin G.M. Fleming E.S. Oetjen J. Gravely B.R. Genes Dev. 1995; 9: 72-83Crossref PubMed Scopus (101) Google Scholar), adenovirus major late region (28DeZazzo J.D. Imperiale M.J. Mol. Cell. Biol. 1989; 9: 4951-4961Crossref PubMed Scopus (67) Google Scholar, 29Prescott J. Falck-Pedersen E. Mol. Cell. Biol. 1994; 14: 4682-4693Crossref PubMed Scopus (36) Google Scholar), cauliflower mosaic virus (30Sanfacon H. Brodmann P. Hohn T. Genes Dev. 1991; 5: 141-149Crossref PubMed Scopus (81) Google Scholar), ground squirrel hepatitis virus (31Russnak R. Ganem D. Genes Dev. 1990; 4: 764-776Crossref PubMed Scopus (98) Google Scholar,32Russnak R. Nucleic Acids Res. 1991; 19: 6449-6456Crossref PubMed Scopus (42) Google Scholar), and human C2 complement (33Moreira A. Wollerton M. Monks J. Proudfoot N.J. EMBO J. 1995; 14: 3809-3819Crossref PubMed Scopus (71) Google Scholar, 34Moreira A. Takagaki Y. Brackenridge S. Wollerton M. Manley J.L. Proudfoot N.J. Genes Dev. 1998; 12: 2522-2534Crossref PubMed Scopus (133) Google Scholar). Spacing between the AAUAAA and the USEs plays a significant role in that the USEs closest to the AAUAAA are most important (21Schek N. Cooke C. Alwine J.C. Mol. Cell. Biol. 1992; 12: 5386-5393Crossref PubMed Scopus (71) Google Scholar). Studies on human immunodeficiency virus suggest that definition of the polyadenylation site involves the recognition of multiple sequence elements, including the USE, in the context of the AAUAAA (27Gilmartin G.M. Fleming E.S. Oetjen J. Gravely B.R. Genes Dev. 1995; 9: 72-83Crossref PubMed Scopus (101) Google Scholar). Comparisons between the polyadenylation signals of the SV40 late mRNA and other cellular mRNAs revealed that three human collagen genes, COL1A1, COL1A2, and COL2A1, each possess elements similar in sequence (USE consensus UAU2–5GUNA) and position relative to the AAUAAA to the SV40 late USEs (21Schek N. Cooke C. Alwine J.C. Mol. Cell. Biol. 1992; 12: 5386-5393Crossref PubMed Scopus (71) Google Scholar)). 2C. S. Lutz, unpublished data. Collagens are extracellular proteins that are responsible for the strength and flexibility of connective tissue. They account for 25–30% of all proteins present in animals and are the major fibrous element of skin, bone, tendon, cartilage, blood vessels, and teeth (see Ref. 35Persikov A.V. Brodsky B. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 1101-1103Crossref PubMed Scopus (67) Google Scholar and references therein). In addition to their structural role, collagens have a directive role in tissue development. The basic structural unit of collagen consists of three polypeptide chains that are extensively covalently cross-linked to each other. The composition of the chains depends on the type of collagen. Type I collagen consists of two COL1A1 chains and one COL1A2 chain, whereas type II collagen consists of three COL2A1 chains. Interestingly, COL1A1, COL1A2, and COL2A1 each possess 3′-UTRs that are extremely conserved between human and other vertebrates, including mice, cows, chickens, and puffer fish (Refs. 36de Wet W. Bernard M. Benson-Chanda V. Chu M.L. Dickson L. Weil D. Ramirez F. J. Biol. Chem. 1987; 262: 16032-16036Abstract Full Text PDF PubMed Google Scholar, 37Sangiori F.O. Benson-Chanda V. de Wet W.J. Sobel M.E. Ramirez F. Nucleic Acids Res. 1985; 13: 2815-2826Crossref PubMed Scopus (23) Google Scholar, 38Sandell L.J. Prentice H.L. Kravis D. Upholt W.B. J. Biol. Chem. 1984; 259: 7826-7834Abstract Full Text PDF PubMed Google Scholar, 39Ninomiya Y. Showalter A.M. van der Rest M. M. B.R. 1984; 23: Scopus Google Scholar, W.B. L.J. N. Y. Acad. Sci. 1985; PubMed Scopus Google Scholar, M. D. de B. E. J. Biol. Chem. 1991; Full Text PDF PubMed Google Scholar, J.C. Dickson de Wet W. Bernard Chu M. G. F.O. Ramirez F. J. Biol. Chem. Full Text PDF PubMed Google Scholar, T. E. Nucleic Acids Res. 1987; PubMed Scopus Google Scholar, see also and human COL1A1 has a 3′-UTR in length with polyadenylation signals. The first containing the first polyadenylation are in by a block of and the including the second polyadenylation are A. P. 1991; PubMed Scopus (23) Google Scholar). high degree of evolutionary conservation suggests important regulatory of the collagen is important to that the of one polyadenylation signal the and this important regulatory elements. the sequences the polyadenylation signal are the is even in the of the polyadenylation signal (see for polyadenylation have been extensively for the and genes H. Mol. Cell. Biol. 1998; PubMed Scopus Google Scholar, Y. M.L. Manley J.L. Cell. 1996; 87: Full Text Full Text PDF PubMed Scopus Google Scholar, R. G. L. Milcarek C. Proc. Natl. Acad. Sci. U. S. A. 1998; PubMed Scopus (71) Google is not one collagen polyadenylation signal is conservation of collagen of polyadenylation to puffer COL1A1, COL1A2, and COL2A1 for 3′-UTR and of polyadenylation signals and for evolutionary conservation of the the polyadenylation signal of the gene by are as COL1A2 is as 5 polyadenylation signals one signal has the sequence J.C. Dickson de Wet W. Bernard Chu M. G. F.O. Ramirez F. J. Biol. Chem. Full Text PDF PubMed Google Scholar). in a COL1A1, COL1A2, and COL2A1 for 3′-UTR and of polyadenylation signals and for evolutionary conservation of the the polyadenylation signal of the gene by are as COL1A2 is as 5 polyadenylation signals one signal has the sequence J.C. Dickson de Wet W. Bernard Chu M. G. F.O. Ramirez F. J. Biol. Chem. Full Text PDF PubMed Google Scholar). of collagen gene expression in a of including and is this regulation of expression is E. J. N. 1996; 22: Full Text Full Text PDF PubMed Scopus Google Scholar, J.E. J. Res. 2000; PubMed Scopus Google Scholar, M. D. W. G. M. 9: Full Text PDF PubMed Scopus Google Scholar, J. E. L. J. Res. 2002; PubMed Scopus (58) Google Scholar). studies suggest in that collagen may be by mRNA transcription R. G. L. Milcarek C. Proc. Natl. Acad. Sci. U. S. A. 1998; PubMed Scopus (71) Google Scholar, R. A. J. 1998; PubMed Scopus Google Scholar), the expression may not be fully by changes in transcription and may be by as polyadenylation. the regulation of 3′ end in human collagen genes. The evolutionary conservation of the polyadenylation to that these regions key regulatory elements. We USEs as auxiliary elements to influence polyadenylation of the collagen these elements in a similar to other and these elements of polyadenylation signals. We determined that the USEs present in these collagen genes influence 3′ end efficiency in these genes. the of poly(A) signals in the COL1A2 gene suggests that assembly of 3′ end processing factors on the signal assembly of the processing on the suggests that between polyadenylation signal elements may a of polyadenylation signal for in vitro polyadenylation and cleavage by of to the in the of of of COL1A2 a transcription of COL2A1 a and transcription of COL1A1 a by in high at in and in cell as previously 1990; PubMed Scopus Google Scholar). In vitro polyadenylation a of and of in a reaction of Alwine J.C. Genes Dev. 1994; 8: PubMed Scopus Google Scholar, Schek N. Manley J.L. Alwine J.C. Genes Dev. 1996; PubMed Scopus Google Scholar). to at for with with and on containing in vitro cleavage a of and of in a reaction of Cooke C. R. Alwine J.C. 1998; 4: PubMed Scopus Google Scholar). to at for processed and as for polyadenylation by of specific oligoribonucleotides as into the polyadenylation reaction and to as a and on the and in the A of the in and is found in representing the USE of COL1A2 the sequence A also and the sequence in to for and USE 3: USE complement in a and in with bovine and in the in of to which of and the at for the addition of of bovine to the the on the and with the The and with with and into of and into at for 5 The and at for the cell the to the for of as and the of the The determined by the for at by I and the and on as COL1A2, COL2A1, and COL1A1 by The COL2A1 and COL1A1 and recognition to insertion into whereas the COL1A2 and recognition and into with at The both and for expression of by the the and the sequences by for in a of reaction for 35 of and with an step of on a in are to in The COL1A2 and the the for of and with a the a for at The into E. and to expression in polyadenylation The COL1A2, the USE and the by the of the of as previously A. J. 1993; 14: Google Scholar). a an upstream containing the as the COL1A2, and the COL1A2 downstream containing the sequence and as the downstream for the second The COL1A2 type upstream for the second A the second as the and COL1A2 type upstream and downstream the the second into and The of in Wilusz J. T. Mol. Cell. Biol. 1988; 8: PubMed Scopus (56) Google Scholar). the of adenovirus type 5 sequences to into at and a containing the The to which a USE of the by a two-step reaction the A. J. 1993; 14: Google Scholar). The first reaction a and to a that a USE insertion upstream of the AAUAAA The second reaction the and the The of this reaction with and as a to all COL1A2 into of of downstream of a and upstream of a bovine growth hormone polyadenylation also of the gene by the and by auxiliary polyadenylation elements known as upstream efficiency elements (USEs) have been and characterized in the SV40 late polyadenylation signal (21Schek N. Cooke C. Alwine J.C. Mol. Cell. Biol. 1992; 12: 5386-5393Crossref PubMed Scopus (71) Google Scholar). determined that these elements as efficiency elements in the SV40 since their in polyadenylation (21Schek N. Cooke C. Alwine J.C. Mol. Cell. Biol. 1992; 12: 5386-5393Crossref PubMed Scopus (71) Google Scholar). has also been shown that a USE the SV40 signal can the human immunodeficiency virus USE in 3′ end in A. Zeichner S. Carswell S. Alwine J.C. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 2108-2112Crossref PubMed Scopus (89) Google Scholar). SV40 and human immunodeficiency virus have polyadenylation signals. has not previously been determined USEs be to a weak polyadenylation as the adenovirus polyadenylation to enhance 3′ processing efficiency. A USE SV40 the sequence upstream of the AAUAAA in the polyadenylation signal to both and and the to in that the of a USE in polyadenylation efficiency as with These data that insertion of a USE can polyadenylation efficiency of a weak processing suggest that USEs can poly(A) site and suggest that USEs may be found in not genes. is important, therefore, to mammalian polyadenylation signals for USE elements. A of cellular polyadenylation signals revealed that elements the USE present in SV40 can also be found in that similar to the consensus and within of the We to on three collagen genes since their 3′-UTRs are highly conserved regulatory the of the SV40 late polyadenylation signal to three human collagen genes for type I and and type II The polyadenylation signals and the USE elements We to these USEs present in the collagen 3′-UTRs 3′ end processing efficiency the SV40 shown that polyadenylation containing an SV40 be specifically by oligoribonucleotides representing the USE Alwine J.C. Genes Dev. 1994; 8: PubMed Scopus Google Scholar). a of each collagen gene 3′-UTR containing the polyadenylation signals by of human for the polyadenylation by in vitro transcription in the of In vitro coupled cleavage and polyadenylation and COL1A2 representing a USE to COL1A2 a also to the The for COL1A2 are shown as an of a in vitro polyadenylation The second in a reaction in the of competitor oligoribonucleotides and that the COL1A2 efficiently in in vitro found with COL1A1 and COL2A1 and their specific oligoribonucleotides not In each the specific USE whereas the effect on polyadenylation (see and data not effect also a for each not of the is the of the in B. oligoribonucleotides representing the collagen USEs can in this a COL1A1 can with a COL1A2 data not Taken together, these data suggest that the oligoribonucleotides specifically and a important for polyadenylation and suggest that the to the SV40 identified in the collagen 3′-UTR is Because of the processing efficiency the COL1A2 we to on that polyadenylation signal in We also by the high degree of sequence conservation of this signal (see We a of the COL1A2 USEs with to the of the USEs to in in USEs as as two at a and three at a A by a in a region upstream of the polyadenylation each by in in the of and to in vitro polyadenylation on The data multiple in vitro and are in Mutation of USE effect on polyadenylation. Mutation of both USEs and also a of USEs and USEs and polyadenylation efficiency to of type A the USE region effect on polyadenylation. We that of the USEs is for COL1A2 polyadenylation that of USE in with at one other USE to the most decreases in polyadenylation. Because all of have been we found important to We COL1A2 type and into downstream of a and upstream of a poly(A) polyadenylation signals have been previously to for a type of auxiliary sequences in the gene S. H.L. M. Proudfoot N.J. Nucleic Acids Res. 1997; 25: PubMed Scopus Google Scholar). A on the downstream of the poly(A) signal also present for in The into and to an transcript the as a are shown in and the of all and are shown in 4 B. The that of the COL1A2 polyadenylation signal of the polyadenylation site the COL1A2 signal type a the USE mutations this and The as the of the to at the COL1A2 site relative to those at the poly(A) site 4 A that the COL1A2 polyadenylation signals the polyadenylation whereas a small that the polyadenylation signal The overall in the in data with the in the USE are in as with in USE and mutations effect on of the COL1A2 polyadenylation whereas USE polyadenylation efficiency to of Mutation of the USEs in also polyadenylation efficiency to of Close of the COL1A2 mRNA sequence revealed an unusual that are in two polyadenylation signals within of each other (see the composition and of the downstream binding site known as the relative to the might that poly(A) signal be of poly(A) signal the of which poly(A) signal the major site of we to cleavage a of that poly(A) signal is the major site of whereas poly(A) signal is the site 5 a of poly(A) signal to polyadenylation now to poly(A) signal suggested that USEs and sequence of the AAUAAA relative to the DSE influences poly(A) signal in this We to of the USEs in with the poly(A) signal 3′ end In vitro of the poly(A) signal consensus AAUAAA to not the overall polyadenylation efficiency (see In in of poly(A) site effect on overall that in with the and USE mutations polyadenylation to of type (see 4 and and 5 Taken together, these data demonstrate that USEs influence 3′ end efficiency in the COL1A2 In this we have identified auxiliary 3′ end processing elements in highly conserved regions of the 3′-UTRs of human collagen genes. These elements polyadenylation in vitro and in In addition, COL1A2 has the unusual feature of polyadenylation one of which and suggests a mechanism for poly(A) signal (see These into regulation of collagen gene expression that of type I and type II collagen genes all have highly conserved the of the collagen 3′-UTRs is by their The 3′-UTRs likely contain important regulatory sequences that influence polyadenylation site and may also influence the of the mRNA. of two cis-acting elements AAUAAA and the downstream by the polyadenylation factors and is the key of efficiency. In the of 3′-UTRs multiple polyadenylation auxiliary elements may be to polyadenylation. The 3′-UTRs of these three collagen genes likely auxiliary to the assembly of polyadenylation Interestingly, within the conserved regions of these 3′-UTRs elements containing with the USE auxiliary polyadenylation elements of We that these USEs in the collagen 3′-UTRs as auxiliary polyadenylation efficiency elements and that these USEs play an important role in an polyadenylation in data suggest that the USEs might be most important for poly(A) signal since of the USEs polyadenylation at that site both poly(A) signals are in vitro data also the are not as not has been in 3′ end is both to the other major processing and and also to mRNA transcription (for review, see Refs. 6Proudfoot N. Trends Biochem. Sci. 2000; 25: 290-293Abstract Full Text Full Text PDF PubMed Scopus (140) Google Scholar, 9Zhao J. Hyman L. Moore C. Microbiol. Mol. Biol. Rev. 1999; 63: 405-445Crossref PubMed Google Scholar, and N.J. A. M.J. Cell. 2002; Full Text Full Text PDF PubMed Scopus Google Scholar). likely in most of cis- and trans-acting signals. is to that the in data most closely regulation at the cellular level and the nature of these The polyadenylation signals present in the COL1A2 3′-UTR are data demonstrate that poly(A) signal is the major site of whereas poly(A) signal is to a (see 5 These data suggest a shown in The of the signals a between binding to poly(A) signal binding to poly(A) signal Mutation of the AAUAAA in poly(A) signal of poly(A) signal (see 5 A and These data suggest that the two polyadenylation signals are in with each other. may not the between and at the for poly(A) signals and may suggest that are at the DSE for poly(A) signal These data demonstrate a that can with a mechanism for poly(A) signal to be this be to polyadenylation efficiency at signals. We the of the Lutz, and for and on the