Functional Analysis of cis-Acting Elements Regulating the Alternative Splicing of Human CFTR Exon 9

© 1999 Oxford University Press Human Molecular Genetics, 1999, Vol. 8, No. 132339–2349

d

eddT5ly

lynalmul-ms

ic-ed

ts)ct

sG

eci-ticsis

ARTICLE

Functional analysis of cis -acting elements regulatingthe alternative splicing of human CFTR exon 9Martina Niksic +, Maurizio Romano +, Emanuele Buratti, Franco Pagani andFrancisco E. Baralle §

International Centre for Genetic Engineering and Biotechnology (ICGEB), Padriciano 99, 34012, Trieste, Italy

Received August 6, 1999; Revised and Accepted September 30, 1999 DDBJ/EMBL/GenBank accession nos AF176095 and AF176096

The rate of exon 9 exclusion from the cystic fibrosis transmembrane conductance regulator (CFTR) mRNA isassociated with monosymptomatic forms of cystic fibrosis. Exon 9 alternative splicing is modulated by apolymorphic polythymidine tract within its 3 ′′′′ splice site. We have generated a minigene carrying human CFTRexon 9 with its flanking intronic sequences and set up an in vivo model to study the cis -acting DNA elementswhich modulate its splicing. Transfections into human cell lines showed that T5, but not T9 or T7 alleles,significantly increases the alternative splicing of exon 9. Moreover, we found that another polymorphic locusjuxtaposed upstream of the T tract, and constituted by (TG) n repeats, can further modulate exon 9 skippingbut only when activated by the T5 allele. Then, we extended our studies to the mouse CFTR exon 9 which doesnot show alternative splicing. Comparison of human and mouse introns 8 and 9 revealed a low homologybetween the two sequences and the absence of the human polymorphic loci within the mouse intron 3 ′′′′ splicesite. We have tested a series of constructs where the whole human exon 9 with its flanking intronic sequenceswas replaced partially or completely by the murine counterpart. The transfections of these constructs inhuman and murine cell lines reveal that also sequences of the downstream intron 9 affect exon 9 definitionand co-modulate, with the UG/U 3 ′′′′ splice site sequences, the extent of exon 9 skipping in CFTR mRNA.

INTRODUCTION

Mutations in cystic fibrosis transmembrane conductanceregulator (CFTR) gene lead to dysfunction of lung, sweatglands, testis, ovary, intestine and pancreas (1). The clinicalmanifestations of the disease are widely variable, ranging fromsevere pulmonary disease with pancreatic insufficiency to mildpulmonary disease and pancreatic sufficiency (2). Moreover,mutations in the CFTR gene have also been found in patientsthat have normal lung function but show other clinical signssuch as congenital absence of vas deferens (3), nasal polyposis(4), bronchiectasis (5,6) and bronchopulmonary allergic aspergil-losis (7). Interestingly, genotype–phenotype correlations haveshown that there is a strong association of the CFTR geneIVS8-T5 allele with male infertility caused by congenital bilat-eral absence of vas deferens (CBAVD) (8–11), and possibly withother monosymptomatic forms of CFTR, such as bronchiectasis(12) and chronic idiopathic pancreatitis (13).

Genetic studies have found that a polymorphic polypyrimidine(Tn) locus located within the 3′ splice site of IVS8 in the CFTRgene is associated with a variable efficiency of exon 9 splicing(14).

There are three alleles at this locus, with 5, 7 or 9 thymidines(T5, T7 and T9, respectively), that affect the efficiency of

CFTR exon 9 inclusion in mRNA which is directly correlatewith the length of the polypyrimidine tract (8).

Several studies performed on patient genotypes confirmthe association between CBAVD and IVS8 T5 allele, anfound that both in nasal epithelium and in vas deferens theallele is associated with a high proportion of alternativespliced CFTR mRNA lacking exon 9 (11,15,16).

However, the fact that the T5 allele is not found exclusivein patients suffering from CBAVD but is also present ihealthy subjects indicates that it is a mutation with partipenetrance (9). Therefore, it has been suggested that the sitaneous presence of other mutations and/or polymorphiscould explain the partial penetrance of the Tn allele. In partular, it was hypothesized that another polymorphic locus bason TG repeats (with alleles ranging from 9 to 13 repeaplaced immediately upstream of the polypyrimidine tramight affect the efficiency of exon 9 inclusion in the CFTRmRNA, in combination or independently from the Tn locu(10). The TG and T tracts are present in the pre-mRNA (as Uand U) within the 3′ splice site of exon 9 and hence could hava direct involvement in exon recognition. Although the assoation of the polymorphic Tn locus and the monosymptomaforms of cystic fibrosis is quite clear, the exact molecular ba

+These authors contributed equally to this work§To whom correspondence should be addressed. Tel: +39 040 3757337; Fax: +39 040 3757361; Email: [email protected]

2340 Human Molecular Genetics, 1999, Vol. 8, No. 13

dd

atee

g- or

T

onse

es-s-e

iny-he

9

of the CFTR exon 9 alternative splicing is still unknown. Inthis paper we have used anin vivo model to investigate thecis-acting RNA elements which modulate the splicing efficiencyof CFTR exon 9.

RESULTS

Tn tract and alternative splicing of CFTR exon 9

In order to explore which are thecis-elements that control thesplicing of human exon 9 and modulate its splicing efficiency,we initially generated a plasmid containing a minigene withthe human CFTR exon 9 and its intron boundaries (Fig. 1a).The human epithelial lung cell line NCI-H322 was chosenbecause of its tissue origin and because it has a reasonablelevel of endogeneous CFTR expression (data not shown).

To analyse the effects of the polypyrimidine tract length onalternative splicing we designed three constructs where thenumber of TG dinucleotide repeats was fixed at 11, whereasthe number of polypyrimidines was variable and included five(pTG11/T5), seven (pTG11/T7) or nine (pTG11/T9) thymidines(Fig. 2a).

Transient transfections into the NCI-H322 cell line followeby RT–PCR using primers specific for our minigene showethat the constructs carrying T9 or T7 genotypes genermRNA with 100% exon 9 inclusion, whereas only T5 genotypactivates alternative splicing of exon 9 (Fig. 2b).

To investigate whether the activation of alternative splicinobserved with the T5 allele is regulated through a sequencespacing-specific effect we generated mutants in which thetract was completely deleted, leaving the (TG)11 repeats adjacentto the 3′ splice site of exon 9 (p∆T), or replaced by unrelatedsequences, i.e. a non-polypyrimidine A tract, creating twconstructs carrying five adenines (pTG11/A5) or seveadenines (pTG11/A7) (Fig. 2a). Figure 2b shows that all theconstructs transfected into the NCI-H322 cell line lead to exprsion of mRNA excluding exon 9, demonstrating that the preence of a polythymidine tract in CFTR IVS 8 is essential for thutilization of the exon 9 3′ splice site. Secondly, it is also shownthat a stretch of polypyrimidines alternating with purines (withthe TG tract) is not equivalent to a functional continuous polpyrimidine tract, contrary to what has been observed for tapolipoprotein AII gene (17).

Figure 1. Schematic representation of human and mouse minigenes. (a) Human exon 9 and its flanking intronic sequences cloned into theNdeI site of pSV-mEDAplasmid. (b) Mouse exon 9 and its flanking intronic sequences cloned into theNdeI site of pSV-mEDA plasmid. The sizes of CFTR exon 9, CFTR introns 8 andand EDA introns are shown as well as the position of relevant restriction enzyme target sites. Dotted lines depict exon 9 alternative splicing. +1 EDA and –1 EDA,fibronectin EDA exons of the pSV-mEDA plasmid.

Human Molecular Genetics, 1999, Vol. 8, No. 132341

islyse

tots

Aione

heggain

weheat

ofthe

gesNA

dive.

propor

T/TG repeats interplay and alternative splicing of CFTRexon 9

Previous observations in human subjects by Cuppenset al.(10) suggested that the TG dinucleotide repeats could affectthe alternative splicing of exon 9 even in a way independent ofthe allele found at the T locus. We have analysed how thelength of the TG tract might regulate the efficiency of exoninclusion in our minigene system, where the T5 allele waslinked to 11, 12 or 13 TG repeats, respectively (Fig. 3a).

Transient transfections of these constructs into the NCI-H322 cell line show that for the T5 allele the proportion ofexon 9 exclusion is further modulated by the number of TGrepeats, since the longer the TG tract is the higher is theproportion of transcripts devoid of exon 9 (Fig. 3b).

On the other hand, when the T7 allele is used there is noactivation of exon 9 skipping regardless of the number of TGrepeats (Fig. 3b). These results suggest that the TG repeatscannot directly control exon 9 skipping, i.e. independentlyfrom the T tract, but they can only modulate alternativesplicing when this is activated by the T5 allele.

To improve the sensitivity of the system and to explore thestructural reasons for this poor exon definition as well as thepossible effects the TG repeats if masked by othercis-elements,we introduced a mutation in the 5′ splice site of exon 9 todecrease its definition and hence increase the exclusion rate ofthis exon. Transfections with constructs carrying the mutatedexon 9 5′ splice site showed again that activation of alternativesplicing occurs only with the T5 allele, and that the number ofTG repeats acts in synergy with the T5 allele to drive exclusionof exon 9 (Fig. 4).

These findings show that the exclusion rate of exon 9inversely proportional to the length of the T tract and directproportional to the number of TG repeats. Moreover, theexperiments indicate that the 3′ splice site affects exon 9 definitionmore significantly than the 5′ splice site.

Secondary structure of human CFTR IVS8 3′′′′ splice site

The secondary structure of the pre-mRNA has been shownstrongly influence the interaction with splicing factors and iprocessing (18). We have analysed experimentally the RNsecondary structure of exon 9 to test whether the compositand length of the TG/T polymorphic loci might affect thsecondary structure of the 3′ splice site of IVS8.

Single- and double-strand specific RNases confirmed tstructure predicted by the energy minimization modellinthrough the program RNAfold (19–21) (Fig. 5a). The resultinmodel structure suggests that the CFTR transcripts contfour loop regions (I, II, III and IV) with the (UG)11/U7 regionactively involved in the formation of domain II by fixing thestructure of this stem–loop (which also contains the 3′ splicesite). To evaluate the role of U7 sequence on this structurealso performed a similar analysis on a mutant carrying t(UG)11/U5 polymorphism. These results (Fig. 5b) suggest thloss of two thymidine residues might weaken the stabilitythe stem, as evidenced by the presence of V1 cleavages inU7 tract which are absent in the U5 structure. No other chancould be detected in the cleavage pattern of the rest of the R(data not shown).

Figure 2. Effect of the T tract on alternative splicing of human exon 9. (a) Sequence of intron 8 3′ end and exon 9 splice junction of constructs carrying 11 fixeTG repeats juxtaposed to nine, seven or five thymidines [p(TG)11/Tn]. In p(TG)11/A7 and p(TG)11/A5 constructs, seven or five adenines replaced seven or fthymidines, respectively. In p(TG)11/∆T construct the T tract was completely deleted. (b) Analysis of pre-mRNA splicing of human constructs in NCI-H322 cellsThe size of PCR products including (409 bp) and excluding (226 bp) exon 9 are indicated. Amplicons were separated on a 1.5% (w/v) agarose gel. Thetionof exon 9 exclusion was determined by densitometry using the total densitometric units of 9+ and 9– amplicons as 100%. p(TG)11/T5, 4%.


to-us

lezeines ame

reon oftric

Cis-acting elements regulating mouse and human exon 9definition

Interestingly, the comparison between human and murineCFTR splice variants showed that in mouse the alternativesplicing of CFTR exon 9 is not present (22).

To investigate the molecular basis of this difference, themurine genomic regions spanning exons 8–9 and exons 9–10were amplified by PCR and the 3′ and 5′ splice junctions andflanking intronic sequences of mouse exon 9 were sequenced.

Two main differences were immediately apparent. First, thesizes of the human and murine IVS8 and IVS9 are significantlydifferent (Fig. 6a) and, secondly, the alignment of mouse andhuman exon 9 with its flanking intronic sequences shows thatmouse IVS8 3′ splice site does not contain the human TG/Tpolymorphic sequences (Fig. 6b). These structural differenceswere consistent with the results obtained in this study indicatingthat the human TG and T polymorphic loci can regulate thealternative splicing of human exon 9. To experimentallyconfirm this observation, the mouse CFTR exon 9 and itsflanking sequences was cloned in the same minigene constructused for the human CFTR exon 9 (Fig. 1b). A series ofconstructs was designed, in which murine IVS8 was replacedwith human IVS8 carrying the (TG)11/T5 or (TG)11/T7 genotypes(Fig. 7a). These human/mouse hybrid constructs, as well as thehuman exon 9 constructs, were transiently transfected either

into human hepatocarcinoma Hep3B or into mouse hepacarcinoma N-Muli cell lines, which express endogeneo

Figure 3. Effect of TG/T interplay on alternative splicing of human exon 9. (a) Sequence of intron 8 3′ end and exon 9 splice junction of constructs carrying variabnumbers of TG repeats juxtaposed to seven or five thymidines [p(TG)11/Tn]. (b) Analysis of pre-mRNA splicing of human constructs in NCI-H322 cells. The siof the PCR products including (409 bp) and excluding (226 bp) exon 9 are indicated. Activation of exon 9 skipping is observed only when five thymidrepresent. PCR products were separated on 1.5% agarose gel. The proportion of exon 9 exclusion was determined by densitometry using the total densitotric unitsof 9+ and 9– amplicons as 100%. p(TG)11/T5, 4%; p(TG)12/T5, 10%; p(TG)13/T5, 18%.

Figure 4. Effect of 5′ splice site mutation on alternative splicing of CFTRexon 9. The influence of exon definition on 3′ splice site recognition wasinvestigated through mutagenesis of the 5′ splice site of exon 9 by replacingguanine in position +1 within intron 9 with an adenine (AAG/gtag→AAG/atag).Analysis of pre-mRNA splicing of human constructs mutagenized at the 5′ splicesite shows activation of exon 9 skipping only when five thymidines apresent. PCR products were separated on 1.5% agarose gel. The proportiexon 9 exclusion was determined by densitometry using the total densitomeunits of 9+ and 9– amplicons as 100%. p(TG)11/T5-5′A, 18%; p(TG)12/T5-5′A,52%; p(TG)13/T5-5′A, 75%.


te of

cal

t

reer,

.er.8n

t aheof, it

ro

uenci

CFTR mRNA (data not shown). Figure 7b shows that thehuman constructs in the mouse cell line displayed activation ofalternative splicing when the T5 allele was present. The sameconstructs in the human cell line showed the activation ofCFTR exon 9 alternative splicing with a significantly lowerrate of exclusion (Fig. 7b). On the other hand, the mouse minigenedid not present exon 9 skipping in both the human and mousecell lines (Fig. 7b). The transfections of the phTG11T7/M/mand phTG11T5/M/m hybrid minigenes resulted in 100% inclusionof exon 9 both in the human and in mouse cell lines, regardlessthe number of Ts within the polypyrimidine tract (Fig. 7b).Together, these results suggest that in humans the polymorphicloci at the 3′ end of intron 8 might be critical but not sufficientto regulate the exon 9 alternative splicing.

In order to define the localization of other regulatorysequences, we generated further human/mouse hybrids where themurine sequences progressively replaced the human counterpart.Transient transfections of the phTG11T7/H/m and phTG11T5/H/mconstructs again resulted in 100% exon 9 inclusion (Fig. 8b).Conversely, transfections of pm/H/h construct resulted in100% exon 9 exclusion (Fig. 8b). The weakness of the mouse3′ splice junction was confirmed through the transfections ofpm/H/m construct that showed activation of a cryptic splicesite 118 bp within the human exon (Fig. 8b), whose upstream

sequences resemble a canonical 3′ splice site better than thatpresent within mouse IVS8 (Fig. 6b, nt 275).

These experiments exclude exon 9 sequences as the sihuman/mouse differences for regulatorycis-acting elementsand point out to the human CFTR IVS9 5′ splice site and itsflanking intronic sequences as another regulatory region critifor exon 9 splicing. Furthermore, they indicate that the 5′ spliceregion of human CFTR IVS9 might exert an inhibitory effecon exon 9 recognition.

DISCUSSION

Human CFTR exon 9 and its flanking introns apparently shamost of the canonical features of splicing sites (23). Howevthe polypyrimidine tract at the 3′ splice site is composedexclusively by thymidines and is polymorphic for its lengthMoreover, it is placed immediately downstream of anothpolymorphic locus made of a variable number of TG repeats

We have shown that this peculiar architecture of the intron3′ splice site modulates the alternative splicing of humaCFTR exon 9. In our minigene system, we have found thastretch of five, but not seven or nine, thymidines activates texon 9 skipping. This is consistent with the relevancepolypyrimidine tract length observed in other genes. In fact

Figure 5. Enzymatic determination of the RNA secondary structure of the human CFTR intron 8 polymorphic loci. (a) Computer prediction of the secondary stuctuemodels of the (UG)11/U7 region. The four loops (I, II, III and IV) are shown and the (UG)11/U7 region is actively involved in the formation of domain II, which alscontains the 3′ splice site splice site. (b) Enzymatic analysis of RNA templates of (UG)11/U7 and (UG)11/U5 genotypes.In vitro transcribed RNAs were enzymaticallydigested with S1 nuclease and T1 and V1 RNases and reverse transcribed. The reverse-transcribed products were separated on polyacrylamide seqng gels.Target sites of S1 nuclease and of T1 and V1 RNases digestions are shown.


inl

2vedites

of intron. The

e

was shown previously that the longer the tract consisting ofpyrimidines (in particular, thymidines) is, the more efficient isthe inclusion of the downstream exon (24). In addition, wehave demonstrated that the TG tract is unable to complementthe polythymidine function and seems to have a detrimentaleffect for exon 9 inclusion when associated with the T5 allele(Fig. 3b). Therefore, within the 3′ splice site of CFTR exon 9,

the effects of the TG tract are different from those reportedother contexts where it is able to form a fully functionapolypyrimidine tract (17,25).

In our system, the proportion of exon 9 skipping in NCI-H32does not reach the reported levels of exon 9 exclusion obserin the individuals with the T5 allele (11,15,26,29) althoughhas the same qualitative trend. The quantitative differenc

Figure 6. Comparison of human and mouse genomic regions spanning exons 8–10. (a) Schematic representation of mouse and human exons 8–10. Sizes of intronsare shown. (b) Alignment of mouse and human exon 9 (183 bp) with its flanking sequences of intron 8 (157 bp; GenBank accession no. AF176095) and9 (209 bp; GenBank accession no. AF176096). Upper line, human sequence; lower line, mouse sequence; uppercase, exon 9; lowercase, introns 8 and 9asteriskdenotes the guanine nucleotide in position +1 within human intron 9 replaced with an adenine nucleotide in constructs mutagenized at the 5′ splice site. Bases inbold indicate the human cryptic 5′ splice site activated after mutagenesis of the wild-type 5′ splice site (CCAG/GCAAGatag). Superlined bases, 118 bp within thhuman exon (nt 275, AG/AT), indicate the cryptic 3′ splice site activated in transfection of pm/H/m construct.


ng

lyncheionattal

eedisnegasfgth

ty)

aneight-r

is

n 9;igel. The

may be due to the variations in cell- and/or tissue-specifictrans-acting factors involved in the regulation of the CFTRexon 9 splicing. This is consistent with the reported cell type- anddevelopment-specific variations of the total and relativeamounts of splicing factors, such as SR (27) and hnRNPproteins (28,29). However, it should be noted that in ourconstruct, exon 9 and its intronic region are flanked by unre-lated genomic sequences lacking CFTR distant exons and,possibly, some other crucialcis-acting element. Hence, wecannot rule out that the minigene context might influence thelevel of exon 9 exclusion. In any case, the TG/T constructsshow a very clear qualitative difference due only to the poly-morphic loci as they differ only in the length of the TG and/orT tract making them a suitable system for studying the alterna-tive splicing of CFTR exon 9.

It is surprising that Hep3B and N-muli give such a strikingdifference in efficiency of CFTR exon 9 alternative splicing(Fig. 7b), considering that both cells have the same tissue oforigin. It is unlikely that we are seeing species-specific interactionsof the murine and human splicing factors because they arehighly homologous (30). It is possible that the cells representdifferent stages of hepatocyte differentiation and may havequalitative and quantitative differences in the set oftrans-acting factors.

Our studies then confirm previous observations by Cuppenset al.(10) based on genotype–phenotype correlations about therelevance of TG repeats on the modulation of exon 9 alternativesplicing. However, our findings strongly suggest that the twoTG and T polymorphic loci do not work independently from

each other and that their function is not only to define spacibetween the branch point and the 3′ splice site. The otherobservation that further supports this hypothesis is that if onthe space that these polymorphic loci create between brasite and the 3′ splice site was important, we would havexpected to obtain the same proportion of exon 9 excluswith different TG/T genotypes. However, we have found ththe behaviour of two genotypes which have the same tonumber of 29 bases, i.e. (TG)11/T7 and (TG)12/T5, is different:the (TG)12/T5 but not the (TG)11/T7 combination activatesalternative splicing (Fig. 3b).

Since the polymorphic T tract is placed adjacently to th3′ splice site of exon 9, two possible models could be proposto explain the activation of the exon 9 alternative splicing. Itplausible that the decrease in the length of the polypyrimiditract might reduce its ability to interact with one of the splicinfactors which recognizes the polypyrimidine tract, suchPTB, U2AF65 or hRNPC (24,31) lowering the efficiency o3′ splice site selection. On the other hand, changes in the lenof the polypyrimidine tract might change the binding specificior increase the binding affinity of some splicing factor(sresulting in a low recognition of intron 8–exon 9 junction.

Recent studies demonstrated that hnRNPA1 can exerteffect at the 3′ splice site, promoting exon skipping (32), so wcan hypothesize that the decrease in length of the T tract mreduce the affinity for a splicing factor recognizing the polypyrimidine tract and this, in turn, might allow a novel ostronger binding of hnRNP A1 protein to the 3′ splice site ofexon 9, resulting in the activation of alternative splicing. Th

Figure 7. Analysis of pre-mRNA splicing from human/mouse hybrid constructs. (a) Schematic representation of human/mouse hybrids. White box, human exoblack box, mouse exon 9; dotted lines, human intron 8 or 9; solid lines, mouse introns. h(TG)11-T7 and h(TG)11-T5 refer to the genotypes of the polymorphic locat the 3′ end of human intron 8. (b) Analysis of pre-mRNA splicing of human/mouse hybrid constructs. PCR products were separated on a 1.5% agaroseproportion of exon 9 exclusion was determined by densitometry using the total densitometric units of 9+ and 9– amplicons as 100%. Hep3B: p(TG)11/T5, 3%;p(TG)12/T5, 7%; p(TG)13/T5, 11%. N-Muli: p(TG)11/T5, 54%; p(TG)12/T5, 72%; p(TG)13/T5, 100%.


hely

ofl

n 9T5

ofR

thtory

Inherartsue

ta-ns

s:

ittheofr

that

atasese-hicy

ndinary

giny intedannTRes

isrred

9,

d

hypothesis is supported by the observation that the 3′ splicejunction of exon 9 (Fig. 6b) has a high homology with thehnRNPA1 consensus target sequence (TAGGGACTTAGGGT)(33). Moreover, the comparison between human and mousesequences (Fig. 6b) shows that the mouse 3′ splice site alsoshares the high homology with the hnRNPA1 consensussequence (Fig. 6b) and, according to our human/mouse hybridstudies, it is poorly recognized when placed into the humanminigene context (Fig. 7b).

Following this reasoning, we are tempted to speculate thatvariations in the accessibility of the CFTR exon 9 3′ splice siteto splicing factors, such as hnRNPA1, might affect its splicingefficiency. Furthermore, experimental studies of the secondarystructure of the 3′ end of intron 8 are consistent with thehypothesis that variations in length of the poly(T) tract mightaffect the accessibility of some splicing factors to the 3′ splicesite.

The transfection experiments suggest that decreasing thenumber of thymidines within the polypyrimidine tract up tofive is critical and provides the basis for a lower exon recognitionat its 3′ splice site. We have tested the overall influence of exondefinition on 3′ splice site recognition, mutagenizing the5′ splice site of exon 9. Interestingly, the mutation in the 5′ splicesite increased the rate of exon skipping only when the T5 allelewas present. In this case the proportion of exon exclusion washigher than that of the corresponding wild-type 5′ splice siteconstructs and it was directly proportional to the length of the

TG tract. These experiments indicate that (i) the function of tpolymorphic polypyrimidine tract might be to rescue a poordefined exon, at least in its 3′ splice site where five thymidines seemto be the threshold for exon 9 recognition; (ii) the contribution5′ splice site to exon 9 definition is significantly less criticathan that of the 3′ splice site.

The fact that the human T tract juxtaposed to mouse exowas not able to activate alternative splicing even whenallele was present (Fig. 7b) led us to consider the existenceother inhibitorycis-acting sequences within the human CFTminigene. To identify the localization of the othercis-elementswe progressively replaced human with mouse sequences.

The first information obtained through the transfections wihuman/mouse hybrids was that there are not different regulaelements within human exon 9, as compared with mouse.fact, on one hand, exon 9 skipping was not observed eitwhen mouse exon 9/intron 9 replaced the human counterp(Fig. 7b) or when mouse intron 9 replaced its human homolog(Fig. 8b). On the other hand, when mouse intron 8 was juxposed to human exon 9/intron 9, the resulting transfectioshowed 100% exon 9 exclusion (Fig. 8b).

It was possible to exclude the 5′ splice site as anothercis-actingelement determinant in exon 9 definition for two reason(i) when we mutagenized the human exon 9 5′ splice site, themodulation of alternative splicing was not affected;continued to be activated uniquely by the presence of five Ts at3′ end of intron 8, but not by seven or nine Ts; (ii) comparisonthe mouse and human 5′ splice site sequences by computeanalysis showed that the human exon 9 5′ splice site is closer tothe consensus than the mouse 5′ splice site.

Subsequent analysis of the human/mouse hybrids showsthe 269 bases at the 5′ region of human intron 9 have a role inthe human-specific exon skipping. Figure 8b shows thhuman intron 9 exerts a negative effect for exon 9 inclusionit leads to the complete exon 9 skipping when it is linked to thmouse 3′ splice site (construct pm/H/h). In our model, it seemthat the exon 9 definition is decreased through two indpendent mechanisms: one is determined by the polymorploci at the 3′ end of intron 8 and the other is controlled bsequences placed within the 269 bp at the 5′ region of humanintron 9, whose function is currently under study.

In conclusion, the divergences of the intronic sequences athe different regulation of exon 9 alternative splicing observedhuman and mouse might be interpreted through an evolutionpoint of view. In fact, Rozmahelet al. (34) have found that afragment of CFTR gene spanning exon 9 and its flankinintrons and polymorphic loci is present in multiple copiesthe human genome, whereas it is present in only single copdifferent Old World monkeys. These results have suggesthat an ancestral retrotransposition event followed byamplification of the integration site occurred in the humagenome (34). Our studies have now found that mouse CFIVS 8 does not present the human polymorphic loci and donot show alternative splicing of exon 9. Therefore, itconceivable that insertion and rearrangement events occuin intervening sequences surrounding human CFTR exonplacing fortuitous newcis-acting elements in the proximity ofits splice junction which lowered exon definition and providethe basis for exon 9 skipping in CFTR mRNA.

Figure 8. Analysis of pre-mRNA splicing from human/mouse hybrid constructs.(a) Schematic representation of human/mouse hybrids. White box, humanexon 9; black box, mouse exon 9; dotted lines, human intron 8 or 9; solid lines,mouse introns; h(TG)11-T7 and h(TG)11-T5, genotype of the polymorphic lociat 3′ end of human intron 8. (b) Effect of transient transfection of human/mousehybrid constructs into Hep3B and N-Muli cell lines. Transient transfections ofthe phTG11T7/H/m and phTG11T5/H/m constructs show 100% exon 9 inclusion.Transfection of pm/H/h construct shows 100% exon 9 exclusion. The transfectionof pm/H/m construct shows activation of a cryptic splice site 118 bp within thehuman exon (nt 275, Fig. 6b). The sizes of PCR product including and excludingexon 9 are shown. PCR products were separated on a 1.5% agarose gel.


ine

ingp

aas

ed

ndnth

arerech-chthdy.24ted

eH-nt

e-

ir

,

ofv)

asin

tes).

tita-35

%nt

nti-e

8)

MATERIALS AND METHODS

Constructs

The human CFTR genomic region including intron 8 (221 bpat the 3′ end), exon 9 (183 bp) and intron 9 (269 bp at the5′ end) was amplified by PCR (94°C 30 s, 60°C 30 s, 72°C60 s, 35 cycles) using the following oligonucleotides, thatinclude anNdeI target site: hcfIVS8 dir, 5′-ttttcatatggggccgct-ctaggacttgataatgggcaaatatctta-3′; hcfIVS9 rev, 5′-cccctcgacca-tatgctcgccatgtgcaagatacag-3′. The PCR product wasNdeI-cut andligated into a pBluescript KS plasmid (Stratagene, La Jolla, CA),previously mutagenized by deletion of theXbaI–XhoI fragmentand insertion of anNdeI site through two complementarysynthetic oligonucleotides within theNotI site.

To permit subsequent clonings, anEcoRI target site wasintroduced through a two-step PCR overlap extension method(35) by replacing an adenine with a cytosine at position +15within exon 9. Sequencing excluded the presence of mutationswithin the insert.

To create Tn and (TG)m alleles, anXbaI–EcoRI cassette wasgenerated by PCR using a common sense primer (5′-catat-ggggccgctctagga-3′) and antisense primer (5′-aaagaattcccc-aaatccctgttaaaaaaacacacacacacacacacacacacatcaaaaataaaagatg-agtt-3′) where A and/or CA number was changed according tothe desired genotype. Identity of each construct was confirmedby sequencing.

To generate expression vectors, pBluescript human exon 9inserts wereNdeI cut and ligated within the EDA intron of thepSV-mEDANdeI-digested vector, described previously (36).

In order to study the influence of exon definition on 3′ splicesite recognition, we mutagenized the 5′ splice site of exon 9 byreplacing guanine in position +1 within intron 9 with anadenine (AAG/gtag→AAG/atag) using a two-step PCRoverlap extension method. After mutagenesis, the new 5′ splicesite selected for exon 9 splicing was placed 5 bases upstream ofthe wild-type site (CCAG/GCAAGatag) (Fig. 6b).

Mouse exons 8–9 and exons 9–10 were amplified by long-range PCR (Expand Long Template; Boehringer Mannheim,Wielandstrasse, Germany) according to the manufacturer’s instruc-tions using genomic DNA of CD1 strain as template and thefollowing oligonucleotides: 8dir, 5′-agtataacttaatgaccacaggca-taatc-3′; 9rev, 5′-tccagtagatccagtaatagccaacatctc-3′; 9dir, 5′-gagatgttggctattactggatctactgga-3′; 10rev, 5′-atctgtactcatcatag-gaaacaccaaaga-3′.

The 2.2 kb exon 8–9 and the 6 kb exon 9–10 fragments wereblunt-end cloned into pBluescript and exon and their junctionswere sequenced.

The mouse CFTR genomic region including intron 8 flankingsequences (157 bp at the 3′ end), exon 9 (183 bp) and intron 9flanking sequences (209 bp at the 5′ end) was amplified byPCR (94°C 30 s, 60°C 30 s, 72°C 60 s, 35 cycles) using thefollowing oligonucleotides: mCF8i dir, 5′-ttttcatatgtctagaaacc-atgtgctttatagt-3′, that includes theNdeI and XbaI target sites;mCF9i rev, 5′-aaaacatatgataggttatccaatcttaagtgatcagttctaaacacg-tgta-3′, that includes theNdeI target site.

The PCR product wasNdeI-cut and ligated into the previouslydescribed pBluescriptNdeI-digested vector. AnEcoRI targetsite was also introduced in mouse minigene through a two-step

PCR overlap extension by replacing an adenine with a cytosat position +15 within exon 9.

The mouse minigene wasNdeI-excised from pBluescript andtransferred into the EDA intron of the pSV-mEDANdeI-digestedvector.

Human/mouse hybrids were generated by exchanghuman/mouseXbaI–EcoRI cassettes or through a two-stePCR overlap extension using previous constructs astemplate. Before expression, the identity of all constructs wchecked by sequencing.

Transfections

Human lung carcinoma NCI-H322 cell line (ECACC95111734) was maintained in culture in RPMI supplementwith 10% fetal calf serum, 50µg/ml gentamicin and 4 mMglutamine. Human hepatocarcinoma Hep3B cell line amouse hepatocarcinoma N-Muli cell line were grown iDulbecco’s modified Eagle’s medium supplemented wi4.5 g/l glucose, 10% fetal calf serum, 50µg/ml gentamicin and4 mM glutamine.

The DNA used for transfections was purified with JetStcolumns (Genomed, Wielandstrasse, Germany). Cells wtransfected with the calcium phosphate co-precipitation tenique. Five micrograms of DNA construct were used for eatransfection and 200 ng of a plasmid carrying human growhormone (hGH) under CMV promoter control were includeas an internal control to normalize transfection efficiencAfter 12 h the medium was replaced with fresh medium andh later the cultures were terminated. The RNA was extracby the method of Chomczynski and Sacchi (37), whilst thmedium was used to assay the hGH production by ELISA (hGELISA; Boehringer Mannheim). Each transfection experimewas repeated at least three times.

Poly(dT) cDNA was synthesized using MoMuLV reverstranscriptase (Gibco BRL, Grand Island, NY). Oligonucleotides specific for pSVEDA constructs, mFN-alphaglob d(cactgcctgctggtgacgtac) and mFN-alphaglob rev (5′-tgggcggccag-ggtcacggc-3′), were used for PCR amplifications (35 cycles45 s at 94°C, 30 s at 60°C, 30 s at 72°C), using 2 U ofTaqDNA polymerase (Boehringer Mannheim).

PCRs were optimized to remain in the exponential rangeamplification and products were fractionated in 1.5% (w/agarose gels. Densitometric analysis of PCR amplicons wcarried out using the Macintosh version of the public domaNIH Image 1.62 program (developed at the US National Instituof Health and available at http://rsb.info.nih.gov/ nih-imageDensitometry data were confirmed by Phospho-imager quantion of radioactive PCRs. Radioactive PCRs were carried out forcycles (45 s at 94°C, 30 s at 60°C, 30 s at 72°C) in 50µl reactionvolumes containing 0.1µl [α-32P]dCTP (1µCi). PCR productswere purified with Microspin S-200-HR columns (AmershamPharmacia Biotech, Little Chalfont, UK) and analyzed on 6denaturing polyacrylamide gels. A phosphor-imager (InstaImager; Packard Instrument, Meriden, CT) was used to quatate PCR amplifications normalized for the C/G content in thPCR-product sequence.

Splice site prediction was carried out using Neural Network (3at http://www.fruitfly.org/seq_tools/splice.html and Splice View(39) at http://l25.itba.mi.cnr.it/~webgene/wwwspliceview.html .


e.ia

on-

ic),

.,non

.,J.,andin

isasal

.ith

, N.,, M.sis.

y

G.s-

S.,

of

, A.,nd

ane

.,)tic

.,

sis

t, S.,B.s inripts

rd

, V.,in

uees

Expression of the CFTR in the cell lines used for our study

Endogenous CFTR expression in NCI-H322, Hep3B and N-Mulicell lines was assayed by RT–PCR using species-specificprimers. cDNA synthesis was carried out with randomhexamer primers. Primers specific for human (dir, 5′-gaagtagt-gatggagaatgt-3′; rev, 5′-agaatgaaattcttccactgtgct-3′) and mouse(dir, 5′-agtataacttaatgaccacaggcataatc-3′; rev, 5′-gcttaataattc-cctctgaagctt-3′) were used to amplify CFTR mRNA spanningexons 8–10, by PCR (35 cycles, 94°C min, 58°C 1 min, 72°C1 min). Semi-nested PCRs were, if necessary, carried out usingdir primers and rev primers (human: rev, 5′-cttgctcgttgacctc-cactca-3′; mouse: rev, 5′-atccttgcacgctgacctcca-3′) spanningexons 8–11 following amplification of exons 8–10. Expressionof CFTR was detected in the NCI-H322 cell line after oneround of PCR, whereas in Hep3B and N-Muli it was detectedafter the semi-nested PCR.

RNA structural mapping

In order to probe the RNA secondary structure of the humanexon 9 3′ splice site we used single- and double-strand specificRNases.

RNA was transcribed using T7 RNA polymerase (Stratagene)from the pBluescript construct containing the human CFTRexon 9 with its flanking intronic sequences, after linearizationwith BamHI. Reaction mixes (100µl final volume) contained1 µg of RNA and 0.02 U RNase V1 (Amersham PharmaciaBiotech) or 0.5 U of RNase T1 (Sigma-Aldrich, Steinheim,Germany) in buffer A (10 mM Tris pH 7.5, 10 mM MgCl2,50 mM KCl), or 20 U S1 nuclease (Amersham PharmaciaBiotech) in buffer B (buffer A plus 1 mM ZnSO4). The RNAwas digested at 30°C for 15 min in a water bath. A controlaliquot of RNA without the addition of RNase was processedsimultaneously with the digested samples. Reactions werestopped by extraction with phenol–chloroform and the aqueousphase was extracted again with chloroform and ethanol precipi-tated in 0.3 M potassium acetate. The pellet was resuspendedin 3 µl of water and RNase cleavage sites were identified byprimer extension with an end-labelled oligonucleotide primer:h9rl, 5′-tccagcaaccgccaacaactg-3′. In a total volume of 5µl,10 ng of 32P-labelled primer were hybridized to the resus-pended RNA in RT buffer (50 mM Tris pH 8.3, 3 mM MgCl2,75 mM KCl). The solution was heated at 65°C for 5 min andallowed to cool for 5 min at room temperature. To each reactionwe then added 15µl of a solution in RT buffer containing0.2 U of MoMuLV reverse transcriptase (Gibco BRL), 2µl of5 mM dNTP mix and 2µl of 0.1 M DTT. The mixture was heldat 42°C for 30 min and then 2µg of RNase A were added andthe mixture was incubated for a further 30 min at 37°C. It wasthen phenol–chloroform extracted and ethanol precipitated in0.3 M potassium acetate. The samples (enzymatically digestedRNA, control reaction and a sequencing reaction using the sameprimer as the RT reaction) were loaded on a 6% polyacrylamidegel which was subsequently dried and exposed to Kodak X-OmatAR films for 12–24 h.

The theoretical prediction of the RNA secondary structure ofthe human exon 9 3′ splice site was performed using the Mfoldprogram (21,40).

ACKNOWLEDGEMENTS

We thank Roberto Marcucci for his skilful technical assistancThis work was supported by grants from the region Friuli-VenezGiulia (no. 119/EC.FIN), from the I.R.C.S.S. Burlo Garofal(Progetto Finalizzato; no. 1327) and from Telethon Onlus Foudation, Italy (no. E.1038).

REFERENCES1. Welsh, M.J., Tsui, L.-C., Boat, T.F. and Beaudet, A.L. (1995) Cyst

fibrosis. In Scriver, C.R., Beaudet, A.L., Sly, W.S., and Valle, D. (edsThe Metabolic and Molecular Bases of Inherited Diseases. 7th edn.McGraw-Hill, New York, NY, pp. 3799–3876.

2. Kerem, E., Corey, M., Kerem, B.S., Rommens, J., Markiewicz, DLevison, H., Tsui, L.C. and Durie, P. (1990) The relation betweegenotype and phenotype in cystic fibrosis—analysis of the most commmutation (delta F508).N. Engl. J. Med., 323, 1517–1522.

3. Dork, T., Dworniczak, B., Aulehla-Scholz, C., Wieczorek, D., Bohm, IMayerova, A., Seydewitz, H.H., Nieschlag, E., Meschede, D., Horst,Pander, H.J., Sperling, H., Ratjen, F., Passarge, E., Schmidtke, J.Stuhrmann, M. (1997) Distinct spectrum of CFTR gene mutationscongenital absence of vas deferens.Hum. Genet., 100, 365–377.

4. Irving, R.M., McMahon, R., Clark, R. and Jones, N.S. (1997) Cystic fibrostransmembrane conductance regulator gene mutations in severe npolyposis.Clin. Otolaryngol., 22, 519–521.

5. Pignatti, P.F., Bombieri, C., Marigo, C., Benetazzo, M. and Luisetti, M(1995) Increased incidence of cystic fibrosis gene mutations in adults wdisseminated bronchiectasis.Hum. Mol. Genet., 4, 635–639.

6. Girodon, E., Cazeneuve, C., Lebargy, F., Chinet, T., Costes, B., GhanemMartin, J., Lemay, S., Scheid, P., Housset, B., Bignon, J. and Goossens(1997) CFTR gene mutations in adults with disseminated bronchiectaEur. J. Hum. Genet., 5, 149–155.

7. Cockrill, B.A. and Hales, C.A. (1999) Allergic bronchopulmonaraspergillosis.Annu. Rev. Med., 50, 303–316.

8. Chu, C.S., Trapnell, B.C., Curristin, S., Cutting, G.R. and Crystal, R.(1993) Genetic basis of variable exon 9 skipping in cystic fibrosis tranmembrane conductance regulator mRNA.Nature Genet., 3, 151–156.

9. Chillon, M., Casals, T., Mercier, B., Bassas, L., Lissens, W., Silber,Romey, M.C., Ruiz-Romero, J., Verlingue, C., Claustres, M.et al. (1995)Mutations in the cystic fibrosis gene in patients with congenital absencethe vas deferens.N. Engl. J. Med., 332, 1475–1480.

10. Cuppens, H., Lin, W., Jaspers, M., Costes, B., Teng, H., VankeerberghenJorissen, M., Droogmans, G., Reynaert, I., Goossens, M., Nilius, B. aCassiman, J.J. (1998) Polyvariant mutant cystic fibrosis transmembrconductance regulator genes. The polymorphic (Tg)m locus explains thepartial penetrance of the T5 polymorphism as a disease mutation.J. Clin.Invest., 101, 487–496.

11. Rave-Harel, N., Kerem, E., Nissim-Rafinia, M., Madjar, I., Goshen, RAugarten, A., Rahat, A., Hurwitz, A., Darvasi, A. and Kerem, B. (1997The molecular basis of partial penetrance of splicing mutations in cysfibrosis.Am. J. Hum. Genet., 60, 87–94.

12. Pignatti, P.F., Bombieri, C., Benetazzo, M., Casartelli, A., Trabetti, EGile, L.S., Martinati, L.C., Boner, A.L. and Luisetti, M. (1996) CFTRgene variant IVS8-5T in disseminated bronchiectasis.Am. J. Hum. Genet.,58, 889–892.

13. Taylor, C.J. (1999) Chronic pancreatitis and mutations of the cystic fibrogene.Gut, 44, 8–9.

14. Chu, C.S., Trapnell, B.C., Murtagh, J.J.Jr, Moss, J., Dalemans, W., JallaMercenier, A., Pavirani, A., Lecocq, J.P., Cutting, G.R., Guggino, W.and Crystal, R.G. (1991) Variable deletion of exon 9 coding sequencecystic fibrosis transmembrane conductance regulator gene mRNA transcin normal bronchial epithelium.EMBO J., 10, 1355–1363.

15. Mak, V., Jarvi, K.A., Zielenski, J., Durie, P. and Tsui, L.C. (1997) Higheproportion of intact exon 9 CFTR mRNA in nasal epithelium comparewith vas deferens.Hum. Mol. Genet., 6, 2099–2107.

16. Larriba, S., Bassas, L., Gimenez, J., Ramos, M.D., Segura, A., NunesEstivill, X. and Casals, T. (1998) Testicular CFTR splice variantspatients with congenital absence of the vas deferens.Hum. Mol. Genet., 7,1739–1743.

17. Shelley, C.S. and Baralle, F.E. (1987) Deletion analysis of a uniq3′ splice site indicates that alternating guanine and thymine residurepresent an efficient splicing signal.Nucleic Acids Res., 15, 3787–3799.


ntiales.

A.R.NA

ales

g

ndle

eic

An.

an

in

r-In

18. Muro, A.F., Caputi, M., Pariyarath, R., Pagani, F., Buratti, E. and Baralle, F.E.(1999) Regulation of fibronectin EDA exon alternative splicing: possiblerole of RNA secondary structure for enhancer display.Mol. Cell Biol., 19,2657–2671.

19. Zuker, M. (1989) Computer prediction of RNA structure.MethodsEnzymol., 180, 262–288.

20. Zuker, M. (1994) Prediction of RNA secondary structure by energyminimization.Methods Mol. Biol., 25, 267–294.

21. Zuker, M. and Jacobson, A.B. (1998) Using reliability information toannotate RNA secondary structures.RNA, 4, 669–679.

22. Delaney, S.J., Rich, D.P., Thomson, S.A., Hargrave, M.R., Lovelock, P.K.,Welsh, M.J. and Wainwright, B.J. (1993) Cystic fibrosis transmembraneconductance regulator splice variants are not conserved and fail to producechloride channels.Nature Genet., 4, 426–431.

23. Lopez, A.J. (1998) Alternative splicing of pre-mRNA: developmentalconsequences and mechanisms of regulation.Annu. Rev. Genet., 32, 279–305.

24. Norton, P.A. (1994) Polypyrimidine tract sequences direct selection ofalternative branch sites and influence protein binding.Nucleic Acids Res.,22, 3854–3860.

25. Coolidge, C.J., Seely, R.J. and Patton, J.G. (1997) Functional analysis ofthe polypyrimidine tract in pre-mRNA splicing.Nucleic Acids Res., 25,888–896.

26. Teng, H., Jorissen, M., Van Poppel, H., Legius, E., Cassiman, J.J. andCuppens, H. (1997) Increased proportion of exon 9 alternatively splicedCFTR transcripts in vas deferens compared with nasal epithelial cells.Hum. Mol. Genet., 6, 85–90.

27. Zahler, A.M., Lane, W.S., Stolk, J.A. and Roth, M.B. (1992) SR proteins:a conserved family of pre-mRNA splicing factors.Genes Dev., 6, 837–847.

28. Dreyfuss, G., Matunis, M.J., Pinol-Roma, S. and Burd, C.G. (1993)hnRNP proteins and the biogenesis of pre-mRNA.Annu. Rev. Biochem.,62, 289–321.

29. Faura, M., Renau-Piqueras, J., Bachs, O. and Bosser, R. (1995) Differedistribution of heterogeneous nuclear ribonucleoproteins in rat tissuBiochem. Biophys. Res. Commun., 217, 554–560.

30. Hanamura, A., Caceres, J.F., Mayeda, A., Franza, B.R.Jr and Krainer,(1998) Regulated tissue-specific expression of antagonistic pre-mRsplicing factors.RNA, 4, 430–444.

31. Roscigno, R.F., Weiner, M. and Garcia-Blanco, M.A. (1993) A mutationanalysis of the polypyrimidine tract of introns. Effects of sequence differencin pyrimidine tracts on splicing.J. Biol. Chem., 268, 11222–11229.

32. Bai, Y., Lee, D., Yu, T. and Chasin, L.A. (1999) Control of 3′ splice site choicein vivoby ASF/SF2 and hnRNP A1.Nucleic Acids Res., 27, 1126–1134.

33. Varani, G. and Nagai, K. (1998) RNA recognition by RNP proteins durinRNA processing.Annu. Rev. Biophys. Biomol. Struct., 27, 407–445.

34. Rozmahel, R., Heng, H.H., Duncan, A.M., Shi, X.M., Rommens, J.M. aTsui, L.C. (1997) Amplification of CFTR exon 9 sequences to multiplocations in the human genome.Genomics, 45, 554–561.

35. McPherson, M.J., Quirke, P. and Taylor, G.R. (1991)PCR. A PracticalApproach. Oxford University Press, New York, NY.

36. Muro, A.F., Iaconcig, A. and Baralle, F.E. (1998) Regulation of thfibronectin EDA exon alternative splicing. Cooperative role of the exonenhancer element and the 5′ splicing site.FEBS Lett., 437, 137–141.

37. Chomczynski, P. and Sacchi, N. (1987) Single-step method of RNisolation by acid guanidinium thiocyanate–phenol–chloroform extractioAnal. Biochem., 162, 156–159.

38. Brunak, S., Engelbrecht, J. and Knudsen, S. (1991) Prediction of hummRNA donor and acceptor sites from the DNA sequence.J. Mol. Biol.,220, 49–65.

39. Rogozin, I.B. and Milanesi, L. (1997) Analysis of donor splice signalsdifferent organisms.J. Mol. Evol., 45, 50–59.

40. Zuker, M., Mathews, D.H. and Turner, D.H. (1999) Algorithms and themodynamics for RNA secondary structure prediction: a practical guide.Barciszewski, J. and Clark, B.F.C. (eds),RNA Biochemistry andBiotechnology. Kluwer Academic Publishers, Dordrecht, The Netherlands.

Functional Analysis of cis-Acting Elements Regulating the Alternative Splicing of Human CFTR Exon 9

Documents

Transcript of Functional Analysis of cis-Acting Elements Regulating the Alternative Splicing of Human CFTR Exon 9