The Human Caspase-2 Gene Alternative Promoters, Pre-mRNA Splicing and AUG Usage Direct Isoform-specific Expression

8/17/2019 The Human Caspase-2 Gene Alternative Promoters, Pre-mRNA Splicing and AUG Usage Direct Isoform-specific Exp…

1/12

The human caspase-2 gene: alternative promoters, pre-mRNA splicing andAUG usage direct isoform-specific expression

Emmanuelle Logette, Anne Wotawa, Ste ´ phanie Solier, Lydie Desoche, Eric Solary and LaurentCorcos*

Inserm U517, Faculty of Medicine, 7 Bd Jeanne d’Arc, 21000 Dijon, France

Caspases have been shown to play important roles inapoptotic cell death, cytokine maturation and celldifferentiation. However, the transcriptional regulationof the corresponding CASP genes remains poorly known.

We describe a 5.1 kb fragment located upstream of thefirst translated exon in the human CASP-2 gene, which isknown to encode caspase-2L and -2S protein isoforms.Transient transfection experiments, together with tran-scription start site mapping and transcript analysis,demonstrate that each caspase mRNA is initiated fromseparate promoter regions, and produced from alternativesplicing events in these regions. The CASP-2L promoter ismuch stronger than the CASP-2S promoter, in goodagreement with the respective transcript levels of the twocaspases. In addition, several in-frame translational startsites can be identified for each isoform, one of which iscommon to both, present in the second common exon, andused efficiently. Surprisingly, the short isoform may also

be initiated at a downstream AUG codon within the sameexon. Thus, promoter strength, alternative transcriptionalinitiation and 50-splicing events regulate the expression of the main caspase-2 isoforms that may be translated fromalternative translation initiation codons.Oncogene (2003) 22, 935–946. doi:10.1038/sj.onc.1206172

Keywords: caspase-2 gene; casp-2S; casp-2L; AUG;mRNA splicing; transcriptional regulation

Introduction

A family of cysteine proteases, known as caspases, playsan important role in many forms of cell death byapoptosis and proinflammatory cytokine maturation(Thornberry and Lazebnik, 1998). So far, 14 caspaseshave been identified in mammalian cells, of which 12human enzymes are known (Nicholson, 1999). Thesecond identified mammalian caspase, known as cas-pase-2 (Alnemri et al ., 1996), was initially referred to asNedd-2 in mice (Kumar et al ., 1992) and Ich-1(interleukin-lb converting enzyme homolog 1) in humans

(Wang et al ., 1994). Nedd-2 was originally identified as adevelopmentally regulated gene in mouse brain (Kumaret al ., 1992). The human CASP-2 gene, Ich-1, encodes atleast two different forms of mRNA species derived from

alternative splicing in the 30

region (Wang et al ., 1994).Overexpression of the long isoform caspase-2L wasshown to induce programmed cell death (Kumar et al .,1994), whereas its decreased expression by antisensetechnology delayed apoptosis in response to variousstimuli (Kumar, 1995; Troy et al ., 2000; Droin et al .,2001a). Conversely, overexpression of the short isoform,caspase-2S, could suppress mammalian cell death(Wang et al ., 1994; Kumar, 1995; Droin et al ., 2001b).Experiments using caspase-2-knockout mice showedthat oocytes were resistant to cell death induced bycytotoxic drugs and that B lymphoblasts were resistantto perforin- and granzyme B-induced apoptosis whereascell death was accelerated in some neuronal populations

(Bergeron et al ., 1998; Morita et al ., 2001). Altogether,these observations indicate that caspase-2 may act bothas a positive and a negative regulator of programmedcell death, depending on cell lineage, development stageand stimulus.

Inclusion or skipping of a 61-base pair (bp) exonin the 30-end of casp-2 pre-mRNA leads to the forma-tion of two mRNAs encoding the short and the longisoform of caspase-2, respectively (Wang et al ., 1994;Cote et al ., 2001a). Insertion of the alternative exon 9introduces a premature stop codon, which leads toproduction of caspase-2S (Wang et al ., 1994; Cote et al .,2001b). In addition, the two mRNAs differ at their

5

0

- end, upstream of the first exon common to bothisoforms, which suggests the existence of distincttranscriptional initiation sites and/or alternativesplicing events in this region. The initial survey of CASP-2 gene expression showed that the long isoform isthe dominant form expressed in most tissues examinedwhereas the short isoform is predominantly expressed inheart, brain and skeletal muscle (Wang et al ., 1994;Kumar et al ., 1997). In the course of rat braindevelopment, casp-2S mRNA reaches similar amountsas casp-2L mRNA, and remains at this level in the adulttissue (Kojima et al ., 1998). Additional caspase-2isoforms may also be derived from CASP-2 genealternative splicing. For example, we have identified a

truncated isoform, casp-2L-Pro, whose overexpression

Received 10 July 2002; revised 1 October 2002; accepted 22

October 2002

*Correspondence: L Corcos;E-mail: [email protected]

Oncogene (2003) 22, 935–946

& 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00

www.nature.com/onc

O

N C O G

E N O M I C S


2/12

negatively interferes with apoptotic pathways (Droinet al ., 2000).

Caspases are synthesized as inactive proenzymes orprocaspases (Nicholson, 1999). These proenzymes mustbe cleaved at key conserved aspartate residues to beactivated. Upon activation, the 48-kDa procaspase-2 is

processed in two subunits of 18 and 14 kDa (Li et al .,1997; Colussi et al ., 1998a; Paroni et al ., 2001). Thisactivation occurs in many cell types in response tovarious apoptotic stimuli, including growth factorwithdrawal, cytotoxic drug or death receptor ligandtreatment, and antigen receptor ligation (Harvey et al .,1996; Stefanis et al ., 1998; Droin et al ., 2001b). Golgin-160 has been identified as a true caspase-2 substrate(Mancini et al ., 2000). However, the exact function of this enzyme in apoptosis and other biological processesis not completely understood. For example, whethercaspase-2 acts upstream or downstream in the caspasecascade remains a matter of controversy despite the fact

that it belongs to the group of caspases with a longprodomain that function mainly as initiators of theproteolytic cascade (Nicholson, 1999; Paroni et al .,2001). Procaspase-2 can undergo oligomerization andautoactivation through its prodomain (Butt et al ., 1998;Colussi et al ., 1998b). Through its large amino-terminalcaspase recruitment domain (CARD), it interacts withthe adaptor protein RAIDD (RIP-associated Ichl-homologous protein with a death domain) followingengagement of some death receptors (Ahmad et al .,1997; Duan and Dixit, 1997; Droin et al ., 2001a).Caspase-2L overexpression can trigger the translocationof Bid to the mitochondria and cytochrome c releasefrom this organelle (Sato et al ., 1997; Swanton et al .,

1999; Paroni et al ., 2001; Guo et al ., 2002; Paroni et al .,2002). Caspase-2 also plays a role as an effector caspasefollowing activation by other proteases, such as caspase-3 or the serine protease granzyme B (Allet et al ., 1996;Ahmad et al ., 1997; Li et al ., 1997; Thornberry et al .,1997; Slee et al ., 1999).

Procaspase-2 was identified in the cytosol and severalcell compartments, including the mitochondria (Susinet al ., 1999), the nucleus (Colussi et al ., 1998a, b;Shikama et al ., 2001) and the Golgi apparatus (Manciniet al ., 2000) depending on the differentiation status of the cells (Sordet et al ., 2002). The proenzyme could beactivated in these various compartments, whereas the

active enzyme could undergo translocation in responseto apoptotic signals (Susin et al ., 1999) and trigger celldeath-related mitochondrial events either directly (Guoet al ., 2002; Lassus et al ., 2002; Robertson et al ., 2002)or from the nucleus without relocalizing into thecytoplasm (Paroni et al ., 2002). Caspase-2 activity maybe modulated by various interacting proteins, includingproapoptotic caspase adaptor protein (PACAP) thatspecifically interacts with procaspase-2L (Bonfoco et al .,2001) and ISBP (IchlS-binding protein) that interactswith procaspase-2S (Ito et al ., 2000).

Very little information is available concerning themechanisms of CASP-2 gene regulation. Upregulationof casp-2L mRNA was observed in the rat brain

following transient ischemia (Asahi et al ., 1997) whereas

previous results from our group have suggested thatcasp-2S mRNA was upregulated in tumor cells exposedto the cytotoxic drug etoposide (Droin et al ., 1998). Thetranscriptional promoters of rat CASP-3 and CASP-9genes have been recently described (Nishiyama et al .,2001; Liu et al ., 2002), but no human CASP gene

promoter has been described so far. In the present study,we demonstrate that several mechanisms contribute tothe synthesis of different caspase-2 isoforms, includingtwo alternative splicing events at the pre-mRNA leveland the use of distinct AUG initiating codons for eachisoform with the short protein using two translationinitiation sites. We also provide the first description of a5.1 kb sequence located upstream of the first translatedexon in a human caspase gene.

RESULTS

Casp-2L and Casp-2S cDNA sequences

Alignment of originally described human casp-2L (Ich-l L) and casp-2S (Ich-ls) cDNA sequences (AccessionU13021 and U13022) shows a perfect identity fromposition 69 to position 929 in casp-2L sequence. Then, a61-bp insertion is found in casp-2S cDNA. Subsequentsequences show a nearly perfect match until position1550 in the casp-2L sequence, corresponding to the end(position 1456) of casp-2S cDNA, with a single basedifference at position 1255 in the casp-2L sequence(Figure 1a). The 61-bp insertion in the casp-2S cDNAhas been previously shown to result from an alternativesplicing event whereby exon 9 is inserted specifically in

casp-2S mRNA (Wang et al ., 1994).Five variant (V1–V5) cDNA sequences have now

been reported (Accession NM32982, NM001224, NM032983, NM 032984, XM 036975, respectively). Three of them (V1, V3 and V5) are related to casp-2L (Figure 1b)whereas the two other (V2 and V4) are related to casp-2S (Figure 1c). Most of the point mutations described inthese variants were identified in only one transcriptsequence. However, some mutations could be moreconsistent, for example, at position 1255 in the 30

untranslated region of casp-2L.Compared to the casp-2L cDNA sequence, V1

includes a 103 bp 50 extension, a 1381 bp 30- extension

and three point mutations (positions 130, 476 and 1255);V5 is characterized by a 52 bp 50-extension, a 1362 bp 30

-extension and one point mutation (position 1255); V3contains the 103 bp 50-extension described in V1, a 1380bp 30-extension, two point mutations (positions 768 and1255) and a 217 bp internal deletion (Figure 1b). Wehave previously isolated and described a truncated casp-2L-like isoform, casp-2L-Pro, which corresponds to apartial V3-like mRNA (Sato et al ., 1997). A survey of the EST databases suggests that casp-2L transcriptscould initiate at least 7 bp upstream of V1 50-end (BG397039, AW 503239) or even further upstream, at 69 bpof V1 50-end (AL582071).

V2 cDNA sequence is 100% identical with casp-2S

with a 765 bp 30

-extension, whereas V4, which contains

Characterization of the 50-end of the human CASP-2 gene

E Logette et al

936

Oncogene


3/12

the same 765 bp 30-extension, is characterized by a 217bp and a 60 bp internal deletions, two point mutationsat positions 908 and 1032 and one single base deletion atposition 1807 (Figure 1b). The survey of EST databasesdoes not bring in new information on the structure of casp-2S cDNA.

Procaspase-2L and procaspase-2S amino-acid sequences

It has been reported previously, based on protein

molecular weight and software-based analysis, thattranslation of casp-2L and casp-2S mRNAs was notinterrupted by the same stop codon. Insertion of 61bases in casp-2S mRNA induces a premature arrest intranslation, leading to a shorter protein (Wang et al .,1994). However, the site of translation initiationremained poorly defined.

The originally described casp-2L (Ich-1L) cDNAencodes a 435 amino-acid protein known as procas-pase-2L (Figure 2a). The common amino acids betweenthe originally described sequence and those deducedfrom V1 and V5 variants are either 100% identical (V5)or differ by only one amino acid at position 159 (L to Vin V1). The protein sequence deduced from V3 is

strikingly different from the previous proteins in its first

N-terminal 19 amino acids, which then differs from theoriginally described procaspase-2L sequence by threeamino-acid changes (Figure 2a).

Since the 50-extension in V1 and V5 variants adds anupstream in phase ATG codon, the correspondingproteins may be longer, including 452 amino acids.Upon analysis of V1 and V5 cDNAs, it appeared thatcasp-2L mRNA contained three in-frame AUG initia-tion codons (Figure 2a). None of these AUGs appar-ently matched the consensus ACCAUGG sequence

identified by Kozak (1986). In order to determine whichAUG was used by the translation machinery, wemutated each of them in V5 cDNA and in vitroexpressed these mutated cDNAs using rabbit reticulo-cyte lysates. Mutations in either of the first two ATGshad no effect on the amount or on the apparentmolecular weight of the protein. In contrast, a mutationin the third ATG completely abolished procaspase-2Lexpression in reticulocyte lysates. Thus, the initial ATGcodon in casp-2L mRNA is the third one. In addition,the ability of procaspase-2L to induce apoptosis wasdetermined following transient transfection. As ex-pected, overexpression of wild-type (wt) procaspase-2Linduced roughly 15% apoptosis, a level similar to that

attained upon treatment of the cells with anti-Fas

Figure 1 Schematic representation of casp-2L and casp-2S cDNA sequences. (a) casp-2L and casp-2S cDNAs differ in their 50-portion(non coding exon 1) and by insertion of the 61 bp exon 9 in the casp-2S cDNA sequence. (b) Initially reported casp-2L cDNA sequence,its subsequently described variants (V1, V3 and V5) and ESTs from the corresponding genomic region. ( c) Initially reported casp-2S cDNA sequence and subsequently described variants (V2 and V4). Arrows with numbers on initially described sequences indicate thepositions at which sequence divergence is observed. Arrows in variants and ESTs represent 5 0 or 30 extensions with numbers indicatingtheir length in base numbers. The dotted lines in V3 and V4 represent internal deletions and numbers indicate their position. A single

base deletion occurs at position 1807 in V2 cDNA as compared to casp-2ScDNA


E Logette et al

937

Oncogene


4/12

receptor antibody in the presence of cycloheximide(21% apoptosis) used as a positive control in theseexperiments. An additive effect (31% apoptosis) wasobserved in procaspase-2L-transfected cells treated byFas receptor activating agents (data not shown).

The 312 amino-acid procaspase-2S sequence wasinitially deduced from casp-2S (Ich-1S) cDNA(Figure 2b). The amino-acid sequence deduced fromV2 cDNA is 100% identical. The sequence deducedfrom V4 cDNA is strikingly different in its first N-terminal 19 amino acids, then shows two amino-acidchanges within its region of overlap with the previouslydescribed sequences and lastly extends the proteinsequence by 106 amino acids (Figure 2b). Analysis of casp-2S cDNAs identified two in-frame ATG initiationcodons (Figure 2b). It had been suggested that transla-tion of this isoform was initiated at the first one.Reticulocyte lysate-driven translation of casp-2S mRNA

showed production of a doublet of roughly 33–34 kDa(Figure 2b). Mutation of the first ATG prevented thesynthesis of the largest protein product, whereasmutation of the second ATG prevented the productionof the shorter peptide. These results indicate that bothATGs function as protein initiation sites for procaspase-2S synthesis. These observations were confirmed bytransient transfection of cDNA expression vectors inHCT116 cells (Figure 2d). The lower molecular weightprotein was more abundant than the largest polypep-tide, indicating that in the living cell, the 33KDaprocaspase-2S is the main protein of these isoforms.Procaspase-2S overexpression did not induce apoptosis,as expected, and no protection against apoptosis

induction by Fas receptor activation was observed,

either with wt or the M1 and M2 mutants (data notshown).

Isolation of the 50-flanking region of the CASP-2 gene

The analysis of casp-2 cDNAs combined with data fromhuman genome databases indicated that the CASP-2gene might contain at least 12 exons. These include twodistinct exons 1 as caspase-2L and caspase-2S isoformsdiffer at the level of their first exon. Since the 5 0-portionof the CASP-2 gene and its 50-flanking region ingenomic DNA had not been reported, we performed aseries of experiments to isolate these regions. Startingfrom exon 2 that contains both caspase-2L and caspase-2S ATG start codons, three steps were required to clonea 5.1 kb sequence from human genomic DNA. First, byusing the Genome Walker kit from Clontech, weobtained a 620 bp fragment including 100 bp of perfect

match with the known cDNA sequence from the CASP-2 gene exon 2 and 520 bp of unknown sequence(Figure 3a). This fragment was subsequently used as aprobe to screen a genomic library, in combination with afull-length cDNA probe. Nine recombinant phages wereselected, which permitted to identify a 4-kb-longsequence at the 50-end of the gene. This sequence wasnearly identical with that of the human clone AC073342,with the exception of 2 bp from the 3 0-portion of thefragment. Lastly, another 1140 bp fragment was isolatedby direct PCR from human genomic DNA. In total,about 5.1 kb of sequence was isolated, upstream of exon2 of the CASP-2 gene (Figure 3a).

Analysis of this 5.1 kb sequence identified a 1883 bp

intron between a 69 bp fragment identical with the first

Figure 2 Alternative translation initiation of casp-2S mRNA at two AUG start sites. (a) Schematic alignment of protein sequencesderived from initially described casp-2L cDNA (Procasp2L) and the three variants described in Figure 1. (b) Schematic alignment of

protein sequences derived from initially described casp-2S cDNA (Procasp2S) and the two variants depicted in Figure 1. M indicates amethionine and its associated number relates to the various mutants tested in panels c and d. Other letters refer to amino-aciddifferences in some variants: V5- and V2-derived proteins are 100% identical with procaspase-2L and procaspase-2S, respectively,whereas V3 and V4 contain internal deletions. (c) Expression of wild-type and mutated V5 casp-2L cDNA in rabbit reticulocyte lysates.The three methionine residues (M) indicated in panel a were successively mutated. ( d) Expression of wild-type and mutated V2 casp-2S cDNA in rabbit reticulocyte lysates. The methionine residues were either independently (M1, M2) or simultaneously (M1+M2)mutated. Controls correspond to empty plasmids. The lower panel shows the Western blot analysis from transfected HCT116 cellextracts


E Logette et al

938

Oncogene


5/12

noncoding exon in casp-2S mRNA and the second exoncontaining the two functional ATG start codons forcaspase-2S translation. The 140 bp, noncoding firstexon of casp-2L was located a further 1 kb distanceupstream. These results indicate that, in additionto exon 9 alternative splicing, two other distinct splic-

ing events may take place in the 50-part of the geneto produce either casp-2L or casp-2S mRNAs(Figure 3b).

Analysis of the 50-end of casp-2L and casp-2S mRNAs

Software analyses identified putative transcription startsites at positions 2615 (casp-2S ) and 3310 bp (casp-2L) (Figure 4a). In order to confirm the existence of active transcription start sites around positions 3310and 2615, we subcloned 100 bp fragments encompass-ing these sites in the promoterless pGL3 enhancervector. In this vector, insertion of a transcription start

site is required to drive luciferase gene expression underthe control of the SV40 enhancer. The DNA segmentencompassing the 3310 position is extremely active atinducing luciferase transcription (Figure 4b), whereasthe DNA segment encompassing the 2615 position didnot permit to obtain luciferase expression above back-ground. It should be noted that, in this case, thebackground luciferase level obtained with the controlvector was quite high, which may have hampered ourability to detect the activity of casp-2S start site.

We then attempted to determine experimentally thelocation of the 50-ends of casp-2L and casp-2S mRNAs.For this purpose, we first used 50-RT–RACE–PCR withvarious specific primers from within exon 1S (casp-2S )

or exon 2 (casp-2L) to amplify the 50

-flanking region of

the casp-2 mRNA isoforms (Figure 4a). This experimentextended V1 cDNA sequence by 21 bp. Since we usedmRNA capped site-dependent amplification, identifica-tion of several extensions suggested that casp-2Ltranscripts could originate at more than one 50-site. AnEST survey also identified several sequences, the largest

one (AW500166) extending the V1cDNA sequence by62 bases. The existence of several transcription start siteswas confirmed by performing an RNase protectionassay using a probe encompassing positions 3320 to3020 bp (Figure 4c). According to this experiment, thetwo most efficiently used 50-transcription sites (3195and 3175) map, respectively, 115 and 135 bp upstreamof the 50-end of V1 (3310). Several other bands wereidentified by the RNase protection assay, suggestingthat other initiation sites may exist (Figure 4c, d).

The 50 RT–RACE–PCR analysis extended V2 cDNAsequence by 89 bp, indicating that neither the initiallydescribed casp-2S sequence nor V2 cDNA corresponded

to full-length transcripts (Figure 4a). RNase protectionassays demonstrated that the main transcription startsite was located at position 2585 (Figure 4c, d). NoEST sequence has been described to cover the 5 0-end of casp-2S cDNA.

Altogether, these analyses demonstrate that severaltranscription start sites within a limited promoter regionexist for casp-2L mRNA, whereas transcription of casp-2S mRNA starts from a major site.

Transcriptional activity of the promoter regions

To identify the putative promoter regions, we firstconducted software analyses. A TATA box was located

at position 2635, upstream of the first base of exon 2.

Figure 3 The 50 flanking region of the CASP-2 gene. (a) Isolation of 5115 bp in the 50-flanking region of the CASP-2 gene. First,genome walking yielded a 620 bp fragment. This fragment and the full-length casp-2S cDNA were subsequently used as probes (doublelines) to screen a lambda DashII phage library. The largest fragment (4 kb) was isolated from a single recombinant phage. Anadditional 1140 bp sequence was isolated by PCR from genomic DNA. ( b) Genomic organization of the 50-flanking region of theCASP-2 gene. The first exons of casp-2L (E1C2L) and casp-2S (E1C2S) mRNAs, as well as the second common exon (E2) are shown asboxes. The first introns for casp-2L (I1C2L, 3 kb) and casp-2S (I1C2S, 1.88kb) are shown as dotted lines, respectively. Numbersindicate the base positions, using the first base of exon 2 as position 1


E Logette et al

939

Oncogene


6/12

Although less clearly defined, a TATA-less region wasalso identified roughly 700 bp upstream (Figure 4a).Then we used a luciferase reporter assay to evaluate thetranscriptional activity of the CASP-2 gene 50-flankingregion. We first tested a 1.4 kb fragment starting at3970 and ending at position 2595, 40 bases down-stream from the putative TATA box and 20 basesdownstream of the casp-2S +1 site identified bysoftware analysis (Figure 5a). Initial experiments wereperformed in human HCT116 colon carcinoma cells(Figure 5b). This fragment demonstrated a high level of

luciferase activity that was roughly 50% that of the

SV40 promoter in the pGL3pm vector. Deletion of 650bp from the 50-end led to a 32% decrease in activity. Aset of additional 50 deletions, which removed the casp-2L+1 region but retained various lengths of the casp-2S promoter region, were produced. Removal of 800, 933,1100 and 1250 bp led to a strong decrease in luciferaseactivity, by up to 92%, which was on average 5 timesabove background (pGL3basic vector). These observa-tions agree to the notion that casp-2S promoter strengthis much weaker than that of casp-2L. On the other hand,deletion of 430 bp, 30 from casp-2L first exon, which

removed casp-2S +1 region, strongly increased the

Figure 4 Determination of the transcription start sites in CASP-2 gene. (a) 5 0 RT-RACE-PCR analysis. Specific primers (arrowheads)from within exon 2 (casp-2L) or E1C2S (casp-2S) were used to amplify the 50-flanking region of the CASP-2 gene. This methodextended the initially described E1C2L by an additional 21 bp, whereas E1C2S was extended by another 89 bp. An EST furtherextended the 50-end of E1C2L by 62 bp. The position of putative transcription start sites, as determined by software predictions, isidentified as large open arrows at positions -2615 and -3310 bp. ( b) Functional analysis of putative transcription start sites. The pGL3vector was used for measuring luciferase activity after transient transfection in HCT116 cells. pGL3b: promoter-less luciferase vector;pGL3E: pGL3enhancer luciferase vector. Fragments (100 bp) of the 5 0-region encompassing the putative transcription start sites wereinserted into the pGL3E vector and assayed for luciferase activity in HCT116 cells. These regions covered positions 3371/3271 and2720/2620 for the putative transcription start sites of casp-2L (‘+1 2L’) and casp-2S (‘+1 2S’), respectively. Basal activities of pGL3b and pGL3E vectors were determined as controls. Numbers represent the mean 7 s.d. of at least three independent experiments.(c, d) RNase Protection Assay. Two probes (shown as A and B on panel d; double lines) were used for this assay. These probesoverlapped with a portion of E1C2L and E1C2S and the putative transcription start sites for each mRNA, respectively. The reactionproducts were analysed by denaturing gel electrophoresis (panel c). Arrows in panels c and d show the positions of the transcriptionstart sites. The lengths of the elongated mRNAs are indicated on the lower part of panel d and the length of the main elongatedfragments is underlined


E Logette et al

940

Oncogene


7/12

luciferase activity, since this activity was almost twicethat of pGL3pm vector. Lastly, two large fragments of 0.8 and 1.8 kb including the software-determined +1position for casp-2L mRNA transcription demonstrated

a functional activity similar to that of the first testedfragment. All these constructs were tested in two othercell lines, namely the K562 human leukemic cell line(Figure 5c) and the Hela cervix carcinoma cell line(Figure 5d). The relative strength of the variousconstructs was comparable in the three cell lines,although their activity ratios to that of the pGL3pmvector were somewhat different.

Discussion

The aim of the present study was to obtain a

comprehensive picture of the human CASP-2 gene

structure and regulation. Although the mouse gene hasbeen invalidated several years ago (Bergeron et al .,1998), structural and functional information on thehuman gene has remained limited. The combination of

previously published data (Wang et al ., 1994; Cote et al .,200la, b) and the present study shows that at least twoalternative splicing events are required to generate thepreviously described long and short isoforms, withseveral potential transcriptional start sites for eachmRNA. In addition, we show that two alternativetranslation initiation sites may be used to generateprocaspases-2S. Thus, both alternative promoter/spli-cing and alternative initiation of translation regulateexpression of the main caspase-2 isoforms.

The human CASP-2 gene includes 12 exons. The firstsplicing event takes place at the 50-end of the gene anddetermines which noncoding exon 1 will be included incasp-2 mRNA. At least two distinct mRNAs can be

generated by alternative splicing of these noncoding

Figure 5 Functional analysis of CASP-2 gene promoter fragments. (a) A schematic representation of CASP-2 promoter-luciferase(LUC) constructs in the pGL3b vector is shown. The positions of 5 0- and 30-ends of the fragments are indicated. Del 1, 2, 3, 4 and 5plasmids are 650, 800, 922, 1100 and 1250 bp 5 0-deletions from fragment A, respectively. Del 6 is 650 bp 5 0-deletion and a 430 bp 30-deletion from fragment A. Del 7 extends fragment A by 1145 bp in 50, and removes 676 bp 30 from fragment A. Del 8 is a 1005 bp 5 0-deletion from del 7. (b, c – d) Luciferase activity was measured after transient transfection in HCT116 (panel b), K562 (panel c) and Hela(panel d) cells, respectively, and compared to the activity of the SV40 promoter (pGL3pm). The luciferase activity is expressed aspercentage of that measured with the so-called fragment A. Numbers represent the mean 7 s.d. of at least three independentexperiments. Luciferase activity of Del 1–5 plasmids was on average 5–8 times above background (pgL3 basic vector)


E Logette et al

941

Oncogene


8/12

exons to a common acceptor site located in the secondexon (the first translated one). Inclusion of the most 5 0

exon 1 is used to generate casp-2L mRNA, whichrequires another splicing event leading to exon 9skipping. The resulting mRNA encodes the long isoformof caspase-2 (Wang et al ., 1994). Inclusion of the second

downstream noncoding first exon 1 (exon 2 of the gene)is associated with that of the 61-bp exon 9 in themRNA, which modifies the open reading frame andgenerates a premature termination codon. A 100 bpelement present in intron 10 has been shown to promotecasp-2S exon 9 inclusion by acting as a decoy splice siteacceptor and proteins that bind to this sequence caneither promote or decrease exon 9 inclusion (Cote et al .,2001a, b). A schematic organization of the CASP-2 genelocus encompassing all exons and introns is shown inFigure 6.

One of the main biological functions of caspases is tomediate the proteolysis that characterizes programmed

cell death. This death process is under complexregulation in which alternative pre-mRNA splicingplays an important role by generating protein isoformsoften harboring opposite activities (for review, Jiang andWu, 1999). For example, differential splicing of the Bcl-x gene that encodes members of the Bcl-2 protein familygenerates several protein variants including Bcl-xl thatdisplays an antiapoptotic activity and Bcl-xS, whichcounteracts the effects of Bcl-x

l (Boise et al ., 1993).

Characterization of the 50-region of the mouse bcl-xgene has identified five distinct promoters whose tissue-specific selection influences the outcome of the splicingprocess (Pecci et al ., 2001).

Alternative splicing of caspase-2 exon 9 was described

initially (Wang et al ., 1994). We show here thatalternative splicing that takes place in the 50 untrans-lated portion of the gene also contributes to theregulation of the caspase-2L/caspase-2S ratio. Interest-ingly, the two main splicing events that characterizeCASP-2 gene regulation can occur independently sincewe have shown recently that the cytotoxic drug etopo-side specifically induces exon 9 inclusion without

influencing the splicing of the noncoding exon 1 in the50 portion of the gene. This effect induces a decrease incasp-2L transcript and protein levels (Wotawa et al .,2002).

The use of alternative translation initiation at multipleAUG start sites is another mechanism that possibly

contributes to the synthesis of multiple protein isoforms.Such an alternative translation initiation was demon-strated to occur on several viral and eucaryotic mRNAs(for review, Willis, 1999). In the present study, we usedexpression vectors mutated at each of the potentialAUG start sites to show that caspase-2S can besynthesized from both the first and the second in -frameAUGs from its second exon. This raises the possibilitythat two closely related caspase-2S isoforms might beexpressed in vivo. Accordingly, transfection of wild-typeand mutated casp-2S cDNAs in colon cancer cells canlead to synthesis of one or both isoforms, depending onthe presence or absence of the two AUG codons. On the

other hand, the first AUG from exon 2 is the only oneidentified to be a functional start site for the caspase-2Lisoform. Thus, in addition to alternative splicing of pre-mRNA, alternative translation initiation contributes tothe synthesis of various caspase-2 isoforms by generat-ing two caspase-2S variants.

The structure of the 50-end and the 50-flanking regionof the CASP-2 gene may account for the differentialexpression of casp-2L and casp-2S mRNAs. Bothnoncoding first exons for casp-2L and casp-2S, in virtueof their differences in sequence and length, may play arole in the respective amounts and/or stability of thetranscripts (Pecci et al ., 2001). In addition, RNaseprotection assays identified several transcription start

sites for the production of these transcripts. Theseinclude one major transcriptional start site for casp-2S mRNA and several start sites for casp-2L mRNA,among which two major sites could be found. Transienttransfection with serial 50 deletions, with the largestremoving the start site used to generate casp-2L mRNA,demonstrated that the casp-2L promoter region has amuch stronger transcriptional activity than that of

Figure 6 Schematic representation of the CASP-2 gene and its two main isoforms. Representation of CASP-2L and CASP-2S exon/intron organization as well as presumptive splicing events. The gene contains 12 exons, among which the first exon differs between thetwo isoforms and is noncoding, and the ninth is incorporated only in casp-2S mRNA. This schematic is adapted with permission from

the Ensembl genome group (www.ensembl .org)


E Logette et al

942

Oncogene


9/12

casp-2S, in good agreement with the respective expres-sion levels of the caspase isoforms (Wang et al ., 1994).The relative activity of promoter constructs are compar-able among several cell lines of distinct tissue origin.However, some quantitative differences can be ob-served, indicating that the relevant trans-acting factors

might differ in their content or activity between cells.Recent studies have revealed novel insights in the

biological functions of caspase-2, including the ability of the protein to regulate cell death from the nucleus(Paroni et al ., 2001; Lassus et al ., 2002; Robertson et al .,2002) and the enzyme interaction with cyclin D3 thatconnects cell proliferation and cell death machineries(Mendelsohn et al ., 2002). We have shown here thatalternative transcription and translation of the CASP-2gene might potentially give rise to a series of variantproteins. Future studies may delineate the biologicalfunctions of these isoforms, the role of specifictranscription factors in CASP-2 gene expression and

the influence of pathophysiological conditions uponCASP-2 gene regulation.

Materials and methods

Cell lines and culture

The human HCT116 colon carcinoma cells and the humanleukemic HL60, U937 and K562 cell lines were obtained fromthe ATCC (Rockville, MD, USA). The Hela cell line wasobtained from the ECACC (Salisbury, UK). Cell lines weremaintained in Eagle’s minimum essential medium (HCT116and Hela) and RPMI 1640 (other cell lines) supplemented with10% fetal calf serum, 2 mm l-glutamine, penicillin (l00 U/ml)and streptomycin (l00 mg/ml) at 371C in an atmosphere of 95%

air and 5% CO2.

In vitro translation and Western blotting

In vitro translation was performed with a coupled transcrip-tion/translation reticulocyte lysate system (Promega, Char-bonnières, France) according to the manufacturer’s protocol.The desired cDNA sequences were cloned into the PCR-TOPO2.1 or TOPO-II plasmid in front of the T7 promoter(Invitrogen, Paisley, Scotland).

Subconfluent HCT116 cells were harvested, washed twice inPBS and lysed in boiling buffer (1% SDS, 1 mm Na3VO4,10 mm Tris pH 7.4) containing 0.1 mm PMSF, 10mg/mlaprotinin and 5 mg/ml leupeptin for 10min at 41C. Equalamounts of proteins (or 10 ml reticulocyte lysate) were boiled in

Laemli buffer for 5 min, separated by SDS–PAGE using 12%polyacrylamide gels and electroblotted onto PVDF mem-branes (BioRad, Ivry-sur-Seine, France). The membranes wereincubated overnight at 41C with an anti-human procaspase-2monoclonal antibody (Pharmingen, Pont de Claye, France)that recognizes both the long and short isoforms of theprotein, an anti-human PARP antibody (Roche, Meylan,France) or with an anti-human HSC70 (SantaCruz/ Tebu, LePerrey en Yvelines, France) as a loading control. The blotswere revealed using an Enhanced Chemioluminescence detec-tion kit (Amersham) by autoradiography.

Apoptosis determination

HCT116 cells were treated with the CH11 anti-Fas receptor

antibody (50 ng/ml) in the presence of cycloheximide (CHX,

0.8 mg/ml), conditions known to induce apoptosis (Micheauet al ., 1999), for 24 h. Transfected cells (see below) were treatedsimilarly 24 h after transfection. Apoptosis was determinedby Hoechst 33342 (1m g/ml) staining the cells for 30 min at371C, and fluorescence microscopy analysis of 200 cells percondition.

Library screening and genome walking

A lambda DashII genomic library (Stratagene, La Jolla, CA,USA) was plated and transferred onto Hybond-N nylonmembranes (Amersham, Les Ulis, France) and was screenedunder conventional hybridization procedures. Extension fromgenomic DNA was performed with the Genome walker kit(Clontech/Ozyme, St Quentin en Yvelines, France) usingdownstream primers from the 50-end of the casp-2 cDNAsequence and adapter primers. A partial genomic sequencecorresponding to the first intron of the CASP-2 gene wasoriginally isolated and submitted to Genbank (Accession #AF283569). The complete sequence of the isolated genomicfragments was determined, and found to match nearlyperfectly with the sequence from BAC clone AC073342. Twoadjacent base pairs from the downstream portion of the firstintron of CASP-2 gene were found to be divergent, possiblycorresponding to a polymorphism.

RNA preparation and RT–PCR analysis

Total RNA was isolated with the ‘nucleospin RNA’ extractionkit (Macherey-Nagel, Hoerdt, France). RNA (500 ng) wasreverse transcribed and then amplified in one step with the‘QIAGEN OneStep RT–PCR’ kit (Qiagen, Courtabeuf,France). The identity of each PCR-generated product withits corresponding cDNA was confirmed by sequencing. Senseand antisense primers used to amplify human cDNAs were asfollows: casp-2L forward primer: 50-ATGGCCGCTGA-CAGGGGACGC-30, reverse primer: 50 GGCAGCAAGTT-GAGGAGTTC 30 and casp-2S forward primer: 50

GGGAGGGAACGATTTAAGGA-30, reverse primer: 50

CCAGAAGATGTTCTAACAATTCGCTC 30. The casp-2Land casp-2S cDNAs were amplified for 25 and 32 PCR cycles,respectively.

50 RACE–PCR

The 50-ends of casp-2 mRNAs were isolated by 50 RACE usingthe ‘FirstChoice RLM-RACE’ Kit (Ambion, Montrouge,France). Total RNA was isolated by cesium chloridecentrifugation, and 10 mg RNA were treated by phosphataseto remove free 50-phosphates from rRNA, tRNA, DNA anddegraded mRNA, leaving intact the cap structure from intact50-mRNA ends. RNA was then treated with a pyropho-

sphatase to remove the cap structure from full-length mRNAs.A 45-base RNA adapter oligonucleotide was ligated to thedecapped mRNA population using a T4 RNA ligase. Arandom primer reverse transcription reaction and nested PCRwith the adapter primer and a specific primer were used toamplify 50-ends of casp-2 transcripts. The amplificationproducts were analysed on 1% agarose gels with ethidiumbromide staining and sequenced. The specific primers used toamplify the 50-ends of casp-2 mRNA and their positions areindicated in Table 1.

RNase protection assays (RPA)

The transcription start points were determined by RPA usingthe RiboQuant Ribonuclease Protection Assay Kit (Pharmin-

gen). The cDNA fragments of interest were amplified by PCR


E Logette et al

943

Oncogene


10/12

and subcloned into the PCR TOPO-II plasmid (Invitrogen).The plasmid template, which contains the T7 bacteriophagepromoter, was used for the synthesis of a radiolabeledantisense RNA probe. Two probes were constructed: the first300 nt cRNA probe encompassing nucleotides(nt) 3320 to3020, which includes part of the first noncoding exon of casp-2L mRNA, and the second 810 nt cRNA probe encompassingnt 2720 to 1910, which includes part of the first noncodingexon of casp-2S mRNA. The probes were then hybridizedovernight in excess to 20 mg of total RNA in solution, afterwhich the free probe and other single-stranded RNAs weredigested by RNases. After proteinase K treatment, theremaining RNase-protected probes were purified and resolvedon denaturing 6% polyacrylamide gels.

Plasmid vectors

Putative promoter-containing fragments were inserted into thepGL3-basic (pGL3b, Promega) vector. Putative transcrip-tional start site-containing fragments were cloned into thepGL3-enhancer (pGL3E) vector. Caspase expression vectorswere constructed by inserting cDNAs into the pTarget plasmid(Promega). Progressive deletion constructs of the CASP-2 gene50-flanking region were produced by PCR. Sequences of primers used are shown in Table 1. These primers also containKpn1 and Xho1 or Bgl II restriction enzyme sites to facilitatesubsequent cloning steps. All plasmids were purified with‘Nucleobond ax’ extraction kits (Macherey-Nagel) and verified

by sequencing.

Transient transfections

Functional promoter activity of the 50-flanking region of theCASP-2 gene was analysed using pGL3 luciferase-reportervectors. The pRSV-b-galactosidase reporter vector was used asan internal control for normalization of transfection efficien-cies. HCT116 and K562 cell lines were seeded into 24-wellplates at a density of 100 000 cells/well 16 h before transfectionby the ‘Lipofectamine plus reagent’ (Invitrogen) according tothe manufacturer’s instructions. A measure of 100 ng of eachtest plasmid and 20ng of pRSV-b-galactosidase reportervector were used for each experiment. U937, HL60 and Helacells were transfected under the same conditions with the ‘X-tremeGENE Q2 transfection reagent’ kit (Roche). After 24 h,cells were harvested and lysates were analysed for luciferase

and b-galactosidase activities using the ‘Luciferase AssayReagent’ and the ‘b-galactosidase Enzyme Assay System’(Promega) and a luminometer (Lumat LB 9507, EG&GBerthold). Caspase expression vectors were introduced intothe cells under the same conditions. Cells were lysed 24 h aftertransfection and protein expression was analysed by Westernblotting.

Site-directed mutagenesis

Point mutations were introduced into caspase-2 expressionvectors with the ‘QuickChange site-directed mutagenesis’ kit(Stratagene) according to the manufacturer’s instructions.Expression vectors (40 ng) were used with the relevantoligonucleotides and PCR was performed for 16 cycles. The

mutations were detected by sequencing. Primers used for

Table 1 Primers used in the indicated experiments

Position Sequence Forward (FP)/reverse (RP)

Primers used to construct deletion fragments5115 50-GGGGTACCGGGACAATGAAGCCTTTGAA-3 0 FP4110 50-GGGGTACCGCGACCAATTGTGTCTACAGAG-3 0 FP3970 50-CTCGAGCTAGAGAGCAGTCTCCAGTG-3 0 FP3320 50-CTCGAGTTTGTCTGTCCGCCGAGCAC-3 0 FP3271 50-CCGCTCGAGCAGGCGCAGAGCTCTAGCGG-3 0 RP3170 50-CTCGAGTCTGAGGGGAGGGATGTGGG-30 FP3048 50-TTCCAGCACAAGGAGCTGAT-3 0 FP3020 50-TACTTACCTGCGTCCCCTGTCTCGAG-3 0 RP2595 50-TCGATACAAATCAAATGCACATC-3 0 RP1883 50-GGGGTACCGGTTAGTATCCCAGCTGCAGTGTTA-3 0 FP+1 50-ACAGATAAACGGTAGAAGGAAAGGCCTCGAGCGG-3 0 RP+15 50-GCCACACACTCCCAATATCC-3 0 RP

Primers used for 50 RT–RACE–PCR

2002 50-GAAGATCTTCGGAAGAAATCTGCTGCACAA-3 0 RP1910 50-GTGCAGCAGATTTCTTCCATTAG-3 0 RP+15 50-GCCACACACTCCCAATATCC-3 0 RP

Primers used for synthesis of RPA probes

3320 50-CTCGAGTTTGTCTGTCCGCCGAGCAC-3 0 FP3020 50-TACTTACCTGCGTCCCCTGTCTCGAG-3 0 RP2720 50-CGGGGTACCTCTTGGGTTCTCTTGAAGAT-3 0 FP1910 -50-GTGCAGCAGATTTCTTCCATTAG-3 0 RP

Primers used for pGL3enhancer constructions

3371 50-CGGGGTACCGCCCTCGCTCTTTCCCCGC-3 0 FP3271 50-CCGCTCGAGCAGGCGCAGAGCTCTAGCGG-3 0 RP2720 50-CGGGGTACCTCTTGGGTTCTCTTGAAGAT-3 0 FP2620 50-CCGCTCGAGGAGCACCGAGGTCTGCAGAA-3 0 RP


E Logette et al

944

Oncogene


11/12

mutagenesis were as follows, with mutated bases bolded andunderlined: V2ATG1, 50 ATTGGGAGTGTGTGGCTGG-CATCCTCATCATCAGG-3 0, V2ATG2, 50 -TCATCACCT-TGGAATGGAGGGAGCTCATCCAGGCC-30, V5ATG1, 50-GGGAAAAGCGGGAAAGGGCGGCGCCGAGCG-3 0, V5-ATG2, 50-CAGCACAAGGAGCTGAGGGCCGCTGACA-

G G-30

, V5ATG3, 50

-GAGTGTGTGGCAGGCATCCTCAT-CATCAGG-30.

AcknowledgementsWe thank members from the Inserm U517 for numerousencouraging discussions. E Logette and S Solier wererecipients at fellowships from the French Ministry of Researchand Technology and from the Medical Research Foundation,respectively. This work was supported by grants from the

Inserm, the Ligue Nationale Franc ¸aise de la Recherche Centrele Cancer and the Conseil Regional de Bourgogne.

References

Ahmad M, Srinivasula SM, Wang L, Talaniann RV, LitwackG, Fernandes-Alnemri T and Alnemri ES. (1997). CancerRes., 4, 615–619.

Allet B, Hochmann A, Martinou I, Berger A, Missotten M,Antonsson B, Sadoul R, Martinou JC and Bernasconi L.(1996). J. Cell. Biol., 135, 479–486.

Alnemri ES, Livingston DJ, Nicholson DW, Salvesen G,Thornberry NA, Wong WW and Yuan J (1996). Cell , 87,171.

Asahi M, Hoshimaru M, Uemura Y, Tokime T, Kojima M,Ohtsuka T, Matsuura N, Aoki T, Shibahara K andKikuchi H. (1997). J. Cereb. Blood Flow Metab., 17,11–18.

Bergeron L, Perez GI, Macdonald G, Shi L, Sun Y, JuriscovaA, Varmuza S, Latham KE, Flaws JA, Salter JC, Hara H,Moskowitz MA, Li E, Greenberg A, Tilly JL and Yuan J.(1998). Genes Dev., 12, 1304–1314.

Boise LH, Gonzalez-Garcia M, Postema CE, Ding L, LindstenT, Turka LA, Mao X, Nunez G and Thompson CB. (1993).Cell , 74, 597–608.

Bonfoco E, Li E, Kolbinger F and Cooper NR. (2001). J. Biol.Chem., 276, 29242–29250.

Butt AJ, Harvey NL, Parasivam G and Kumar S. (1998). J.Biol. Chem., 273, 6763–6768.

Colussi PA, Harvey NL and Kumar S. (1998a). J. Biol. Chem.,273, 24535–24542.

Colussi P, Harvey NL, Shearwin-Whyatt LM and Kumar S.(1998b). J. Biol. Chem., 273, 26566–26570.

Cote J, Dupuis S, Jiang Z and Wu JY. (2001a). Proc. Natl.Acad. Sci. USA, 98, 938–943.

Cote J, Dupuis S and Wu JY. (2001b). J. Biol. Chem. 276,8535–8543.

Droin N, Beauchemin M, Solary E and Bertrand R. (2000).Cancer Res., 60, 7039–7047.

Droin N, Bichat F, Rebe C, Wotawa A, Sordet O, HammanA, Bertrand R and Solary E. (2001a). Blood , 97, 1835–1844.

Droin N, Dubrez L, Eymin B, Renvoize C, Breard J,Dimanche-Boitrel MT and Solary E. (1998). Oncogene, 16,

2885–2894.Droin N, Rebe C, Bichat F, Hammann A, Bertrand R andSolary E. (2001b). Oncogene 20, 260–269.

Duan H and Dixit VM. (1997). Nature 385, 86–89.Guo Y, Srinivasula SM, Druilhe A, Fernandes-Alnemri T

and Alnemri ES. (2002). J. Biol. Chem., 277, 13430–13437.Harvey N, Trapani JA, Fernandes-Alnemri T, Litwack G,

Alnemri ES and Kumar S. (1996). Genes Cells, 1, 673–685.Ito A, Uehara T and Nomura Y. (2000). FEBS Lett., 470, 360–

364.Jiang ZH and Wu JY. (1999). Proc. Soc. Exp. Biol. Med ., 220,

64–72.Kojima M, Asahi M, Kikuchi H, Hashimoto N, Noda M and

Hoshimaru M. (1998). Neurosci. Res., 31, 211–217.Kozak C. (1986). Cell , 44, 283–292.

Kumar S. (1995). FEBS Lett., 368, 69–72.

Kumar S, Kinoshita M and Noda M. (1997). Leukemia, 11,S3: 385–386.

Kumar S, Kinoshita M, Noda M, Copeland NG and JenkinsNA. (1994). Genes Dev., 8, 1613–1626.

Kumar S, Tomooka Y and Noda M. (1992). Biochem. Biophys.Res. Commun., 185, 1155–1161.

Lassus P, Opitz-Araya X and Lazebnik Y. (2002). Science,297, 1352–1354.

Li H, Bergeron L, Cryns V, Pasternack MS, Zhu H, Shi L,Greenberg A and Yuan J. (1997). J. Biol. Chem., 272, 21010– 21017.

Liu W, Wang G and Yakovlev AG. (2002). J. Biol Chem., 277,8273–8278.

Mancini M, Machamer CE, Roy S, Nicholson DW, Thorn-berry NA, Casciola-Rosen LA and Rosen A. (2000). J. Cell Biol., 149, 603–612.

Mendelsohn AR, Hamer JD, Wang ZB and Brent R. (2002).Proc. Natl. Acad. Sci. USA, 99, 6871–6876.

Micheau O, Solary E, Hammann A and Dimanche-BoitrelMT. (1999). J. Biol. Chem., 274, 7987–7992.

Morita Y, Maravei DV, Bergeron L, Wang S, Perez GI,Tsutsumi O, Taketani Y, Asano M, Horai R, Korsmeyer SJ,Iwakura Y, Yuan J and Tilly JL. (2001). Cell Death Differ.,8, 614–620.

Nicholson DW. (1999). Cell Death Differ., 6, 1028–1042.Nishiyama J, Yi X, Venkatachalam MA and Dong Z. (2001).

Biochem. J., 360, 49–56.Paroni G, Henderson C, Schneider C and Brancolini C. (2001).

J. Biol. Chem., 276, 21907–21915.Paroni G, Henderson C, Schneider C and Brancolini C. (2002).

J. Biol. Chem. 277, 15147–15161.Pecci A, Viegas LR, Baranao JL and Beato M. (2001). J. Biol.

Chem., 276, 21062–61069.Robertson JD, Enoksson M, Suomela M, Zhivotovsky B and

Orrenius S. (2002). J. Biol. Chem., 277, 29803–29809.Sato N, Milligan CE, Uchiyama Y and Oppenheim RW.

(1997). Gene, 202, 127–132.Shikama Y, Mami U, Miyashita T and Yamada M. (2001).

Exp. Cell Res., 264, 315–325.Slee EA, Harte MT, Kluck RM, Wolf B, Casiano CA,Newmeyer DD, Wang HG, Reed JC, Nicholsmon DW,Alnemri ES, Green DR and Martin SJ. (1999). J. Cell. Biol.,144, 281–292.

Sordet O, Rebe C, Dubrez-Daloz L, Boudard D and Solary E.(2002). Leukemia, 16, 1569–1570.

Stefanis L, Troy CM, Qi H, Shelanski ML and Greene LA.(1998). J. Neurosci., 18, 9204–9215.

Susin SA, Lorenzo HK, Zamzami N, Marzo I, Brenner C,Larochette N, Prevost MC, Alzari PM and Kroemer G.(1999). J. Exp. Med., 189, 381–394.

Swanton E, Savory P, Cosulich S, Clarke P and Woodman P.(1999). Oncogene 18, 1781–1787.

Thornberry NA, Rano TA, Peterson EP, Rasper DM, Timkey

T, Garcia-Calvo M, Houtzager VM, Nordstrom PA, Ro S,


E Logette et al

945

Oncogene


12/12

Vaillancourt JP, Chapman KT and Nicholson DW. (1997).J. Biol. Chem., 272, 17907–17911.

Thornberry NA and Lazebnik Y. (1998). Science, 281, 1312– 1316.

Troy CM, Rabacchi SA, Friedman WJ, Frappier TF, BrownK and Shelanski ML. (2000). J. Neurosci., 20, 1386–1392.

Wang L, Miura M, Bergeron L, Zhu H and Yuan J. (1994).Cell , 78, 739–750.

Willis AE. (1999). Int. J. Biochem. Cell. Biol., 31, 73–86.Wotawa A, Solier S, Logette E, Solary E and Corcos L. (2002).

Cancer Lett., 185, 181–189.


E Logette et al

946

Oncogene

Documents

The Human Caspase-2 Gene Alternative Promoters, Pre-mRNA Splicing and AUG Usage Direct Isoform-specific Expression