A new internal-ribosome-entry-site motif potentiates XIAP ... - Nature

brief communications

A new internal-ribosome-entry-site motif potentiates XIAP- mediated cytoprotection Martin Holcik*, Charles Lefebvre†, Chiaoli Yeh‡, Terry Chow‡ and Robert G. Korneluk*†§¶ *Solange Gauthier-Karsh Molecular Genetics Laboratory, Children’s Hospital of Eastern Ontario, 401 Smyth Road, Ottawa, Ontario K1H 8L1, Canada †Apoptogen Inc., Ottawa, Ontario K1H 8M5, Canada ‡Department of Oncology, McGill University, Montreal PQ H3G 1A4, Canada §Department of Microbiology and Immunology, University of Ottawa, Ottawa, Ontario K1H 8M5, Canada ¶e-mail: [email protected]

tionally regulated, including an unusually long 5′ untranslated region (UTR) (>5.5 kilobases (kb) for murine and >1.6 kb for human XIAP transcripts) with predicted complex secondary structure and numerous potential translation start sites upstream of the authentic initiation codon. This UTR would be expected to present a significant obstacle to efficient translation by conventional ribosome scanning7. An alternative mechanism of translation initiation, mediated through the IRES, has been identified in picornaviruses and in a few cellular mRNAs8. Thus we tested whether the 5′ UTR of XIAP mRNA could be involved in translation initiation from reporter-based bicistronic mRNA transcripts encoding β-galactosidase and chloramphenicol aceytyltransferase (CAT) (for example, see ref. 9). (Translation of β-galactosidase is driven by the 5′ mRNA methylguanosine cap.) Both human and mouse XIAP 5′ UTRs directed translation of the second cistron (encoding CAT) at 150fold higher levels than those produced without the 5′ UTR or with the 5′ UTR in reverse orientation, suggesting the presence of an IRES (Fig. 1a). No activity was detected when using the identical DNA segments cloned into a promoterless construct, confirming

rogrammed cell death (apoptosis) plays a critical part in regulating cell turnover during embryogenesis, metamorphosis, tissue homeostasis and viral infection1. Dysregulation of apoptosis occurs in such pathologies as cancer, autoimmunity, immunodeficiency and neurodegeneration. Proteins of the inhibitor-ofapoptosis (IAP) family are intrinsic cellular suppressors of apoptosis and are represented by highly conserved members found from insect viruses to mammals2–4. The most potent mammalian IAP is the X-linked IAP, or XIAP5, whose mechanism of action involves direct inhibition of caspases 3 and 7, key proteases of the apoptotic cascade6. Cellular control of XIAP expression should be fundamental to a cell’s ability to modulate its responses to apoptotic stimuli. However, XIAP messenger RNA is expressed in most tissues and cells at fairly constant levels5, indicating that translational control of XIAP levels may be an important regulatory mechanism. Here we characterize the primary genomic structure and function of XIAP, and show that XIAP expression is controlled at the translational level, specifically through an internal ribosome-entry site (IRES). Several features of XIAP mRNA indicate that it may be transla-

P

a

βgal

CMV

LR

CAT

b

polyA

pβgal/CAT

βgal

CMV

pβgal/hUTR/CAT

polyA

pβgal/5'(–162)/CAT pβgal/5'(–83)/CAT pβgal/3'(–34)/CAT pβgal/3'(–47)/CAT pβgal/∆(–162;–47)/CAT

–1 (AUG) –1395

–1 hUTR

–1

CAT

–1(AUG)

UGUUCUCUUUUU

–1007

–1395

–1007

hUTR

–1007

–1 (AUG)

–1

pβgal/hRTU/CAT

pβgal/mRTU/CAT

CAT

Xhol

–1007

pβgal/hUTR/CAT

pβgal/mUTR/CAT

LR

pβgal/CAT

Xhol

polyA

–162 –83 –1007 –1007 –1007

–1 –1 –34 –47 –34

pCATbasic/hUTR 0

20 40 60 80 100 120 140 160

Relative CAT activity

pβgal/Py-Pu/CAT pβgal/Py(–46,45AA)/CAT pβgal/Py(–44,43AA)/CAT pβgal/Py(–42,41AA)/CAT pβgal/Py(–40,39AA)/CAT pβgal/Py(–38,37AA)/CAT pβgal/Py(–36,35AA)/CAT

–1007 –1007 –1007 –1007 –1007 –1007 –1007

AGAAGAGAAAAA AAUUCUCUUUUU UGAACUCUUUUU UGUUAACUUUUU UGUUCUAAUUUU UGUUCUCUAAUU UGUUCUCUUUAA CAT

polyA

–1

CAT

polyA

–1007

CAT

polyA

CAT

polyA

pCMV/CAT MONO/hUTR/CAT MONO/hRTU/CAT MONO/Py(–42,41AA)/CAT

–1007 hUTR –1

hRTU

–1007 hUTR

–1 X

0

20

40

60

80

100

120

Relative CAT activity

Figure 1 The 5′ UTR of mouse and human XIAP mRNA contains functional IRES elements. a, DNA segments corresponding to the indicated regions of the 5′ UTR of human (h) or mouse (m) XIAP transcripts were inserted (in the indicated directions) into the XhoI site of the linker region (LR) of the bicistronic plasmid pβgal/CAT (where βgal is β-galactosidase); HeLa cells were transfected with these plasmids (2 µg each). The promoterless CAT reporter plasmid pCATbasic/UTR was constructed by inserting the indicated 5′ UTR region into the pCATbasic vector, and in this case HeLa cells were co-transfected with both the pCATbasic/hUTR and the pCMVβ plasmids as described in Methods (only the construct with the human 5′ UTR is shown). After 24 h, cell extracts were prepared and the βgal12 and CAT13 activities were determined. Relative CAT activity was calculated by normalizing with βgal activity. Expression of the CAT cistron from the pβgal/CAT construct was set as 1.

The bars represent the average ± s.d. of five independent transfections. Identical results were obtained with NIH3T3 cells (data not shown). b, Deletion and mutational analysis. DNA segments corresponding to indicated regions of the 5′ UTR of the human XIAP transcript were inserted into the XhoI site of the LR of the pβgal/CAT plasmid. The small black boxes indicate the position of the PPT. Plasmids with mutated PPT were constructed by PCR-directed mutagenesis. The sequence of the PPT is indicated for each construct; base substitutions are underlined. Monocistronic plasmids (pMONO) were constructed by deleting the βgal cistron (NotI site) from respective plasmids. Cell culture and determination of βgal and CAT activity were done as in a. The bars represent the average ± s.d. of three independent transfections. Py, pyrimidine substitution.

190

NATURE CELL BIOLOGY | VOL 1 | JULY 1999 | cellbio.nature.com

© 1999 Macmillan Magazines Ltd

brief communications a

mRNA

b

c HeLa

0.5

1.0

2

1

0

0.5 Dose (Gy)

1.0

150

100

50

0

Control

pCMV2A

Serum starvation

Figure 2 XIAP-IRES-directed translation is resistant to cellular stress. a, XIAP mRNA and protein levels in the non-small-cell lung carcinoma cell line H661 were analysed by northern and western blot analysis following low-dose γ-irradiation. Inset shows representative blots. Changes in XIAP levels were analysed densitometrically and normalized for levels of GAPDH (mRNA) or total protein loaded (average ± s.d. of three experiments). b, HeLa cells were co-transfected with plasmids pβgal/5′(268)/CAT (2 µg) together with pCMV2A (2 µg) or with the control plasmid pcDNA3 (2 µg), and the βgal and CAT activities were determined 48 h post-transfection. For the irradiation experiment, H661 cells were transfected with plasmid pβgal/5′(-268)/ CAT (5 µg). After the transfection procedure, the cells were incubated for 12 h and then irradiated with 60Co γ-rays (1.0 Gy). The βgal and CAT activities were determined

that the 5′ UTR regions do not contain cryptic promoters. To determine which part of the 5′ UTR is responsible for translation initiation, we generated constructs containing deletions of the human XIAP 5′ UTR (Fig. 1b). The region that retained full IRES activity was the nucleotide segment –162 to –1 upstream of the initiation codon; this segment was as effective as the larger 5′ UTR. The smallest construct contained only 83 nucleotides of the 5′ UTR, but its activity was 25% that of the full-length UTR, being 30fold higher than that of the bicistronic reporter containing no IRES. In monocistronic plasmids, the 5′ UTR in the sense orientation did not reduce translation of the reporter gene. However, when the IRES was in the antisense orientation or was mutated (see below), expression was substantially reduced, indicating that XIAP translation may be fully dependent on the IRES (Fig. 1b). A polypyrimidine tract (PPT) is located 34 nucleotides upstream of the XIAP mRNA initiation codon. IRES elements of picornavirus contain a similarly positioned PPT but this sequence has not been found in the cellular IRES motifs described so far. Therefore, we Table 1 Relative levels of XIAP protein and mRNA in response to stress Relative levels of XIAP* Cell line

Stress trigger Protein

LacZ XIAP IRES.XIAP

100

βgal CAT Survival (per cent of control)

0

200

3

0

H661

Protein Activity of reporter relative to control

Relative amount of protein or mRNA

4

mRNA

H661 (non-small-cell carcinoma, lung) 1.0 Gy

3.33±0.05

0.85±0.07

H460 (non-small-cell carcinoma, lung) 1.0 Gy

0.34±0.14

0.76±0.11

H520 (squamous carcinoma, lung)

1.0 Gy

1.41±0.25

0.76±0.06

SKOV3 (adenocarcinoma, ovary)

1.0 Gy

0.73±0.07

0.83±0.19

HeLa (cervical carcinoma)

72 h SW

2.03±0.10

0.91±0.07

293 (human embryonic kidney)

72 h SW

2.17±0.32

0.80±0.16

HEL299 (human embryonic lung)

72 h SW

1.95±0.04

0.83±0.09

SH-SY5Y (neuroblastoma)

72 h SW

1.75±0.32

0.78±0.16

* Endogenous levels of XIAP protein (western blot) and mRNA (northern blot or ribonuclease protection assay) were analysed following either 1.0 Gy ionizing irradiation or serum withdrawal (SW) for 72 h. Changes in XIAP levels were analysed densitometrically, normalized to levels of GAPDH (mRNA) or total protein loaded, and compared with untreated controls which were set as 1 (average ± s.d.).

Serum 1.0 Gy starvation

80 60 40 20 0

0

24 48 72 Time of serum deprivation (h)

12 h post-irradiation. For the serum-deprivation experiment, HeLa cells were transfected with plasmid pβgal/5′(-268)/CAT (2 µg). The transfected cells were deprived of serum for 24 h and the βgal and CAT activities determined. Expression of each reporter cistron assayed in the control transfection was set at 100% in each case. Bars represent the average ± s.d. of three independent transfections. c, HeLa cells were transfected with 2 µg of plasmid pCI-LacZ, pCI-XIAP or pCI-IRES/XIAP. After 24 h, cells were washed with PBS and grown in fresh serum-free media. The viability of the cells at different time intervals was assessed by a colorimetric assay (see Methods). Bars represent the percentage of surviving cells ± s.d. of three independent experiments done in triplicate.

determined whether the PPT is important for XIAP IRES function. Although the sequence between the initiation codon and the PPT is dispensable for XIAP IRES function, deletion of this tract abolished IRES activity completely (Fig. 1b). A similar result was found for picornavirus IRES elements, where the PPT is essential for ribosome binding10. We next deleted the –162 to –47 nucleotide segment to determine whether the PPT itself is sufficient for IRES function. This deletion, however, completely abolished IRES activity, showing that although the PPT sequence is necessary, it is not sufficient for internal translation initiation. To study the sequence specificity of the PPT, we tested several base-substitution mutants (Fig. 1b). The substitution of purines for pyrimidines markedly reduced IRES activity, indicating the need for the PPT. Thus a specific sequence within the PPT may be critical for XIAP activity. Significantly, the XIAP IRES is the first cellular IRES to be shown to possess a functional PPT. We then studied the physiological relevance of IRES-mediated XIAP translation under different cellular stresses. Low-dose irradiation resulted in a 3.5-fold upregulation of XIAP protein amounts in the non-small-cell lung carcinoma cell line H661, whereas the level of XIAP mRNA remained unchanged (Fig. 2a). The expression pattern of other IAP genes (HIAP1, HIAP, and NAIP) did not change (data not shown). We proposed that XIAP expression is translationally upregulated through the IRES after low-dose irradiation. To test this hypothesis, we transfected H661 cells with the bicistronic reporter and measured the translation of both cistrons after irradiation. The expression of β-galactosidase was reduced to 51% of that in non-irradiated cells, but the expression of the IRESdriven CAT reporter was increased (Fig. 2b), indicating that, in the irradiated cells, the upregulation of endogenous XIAP is mediated by the IRES sequence. Other cellular stresses, such as poliovirus infection, growth arrest, hypoxia or heat shock, also inhibit cap-dependent, but not IRES-dependent, translation11. We determined whether the XIAP IRES element is functional during overexpression of viral protease 2A or serum withdrawal (Fig. 2b). In both cases, the translation of the cap-dependent reporter was reduced whereas the translation of the XIAP-IRES-driven reporter remained unchanged. Overexpression of XIAP protects cells against apoptosis triggered by various



191

brief communications stimuli, including serum withdrawal4. In these experiments, however, only the coding region of the XIAP transcript was used. We reasoned that if the translation of XIAP is mediated by the IRES element located within its 5′ UTR, the overexpression of transcript containing IRES should offer greater protection than that seen with the XIAP coding complementary DNA alone. Indeed, at all time points following transfection and serum starvation of HeLa cells, the XIAP IRES protected cells from apoptosis more efficiently than did the XIAP coding region alone (Fig. 2c). The amount of XIAP protein produced from the XIAP IRES construct during serum starvation exceeded that of the construct with XIAP coding region alone, although the levels of XIAP mRNA transcribed from both plasmids were the same (data not shown). What is the biological significance of the internal initiation of translation of XIAP? To address this issue, we tested several cell lines for their ability to upregulate the translation of XIAP mRNA in response to cellular stress. IRES-mediated translational upregulation of endogenous XIAP without concomitant upregulation of XIAP mRNA varied in response to different triggers of cellular stress in different cell lines (Table 1). Serum deprivation upregulated XIAP protein in all cell lines tested, but low-dose irradiation induced XIAP translation in only one (H661), possibly two (H520), lines. H661 was the most resistant of the four cell lines to radiationinduced apoptosis (data not shown). A general cellular response to apoptotic stress is the shut-off of cap-dependent protein synthesis. The presence of the IRES would allow for the continuous production of XIAP, even under stress conditions. The degree of responsiveness of the IRES element would dictate the cellular threshold of responsiveness to apoptotic stimuli. This threshold would be set according to the cell’s intrinsic properties or could be manipulated by external stimulation of IRES-mediated XIAP translation. In either case, the induction of XIAP protein might be beneficial to the cell’s survival under acute but transient apoptotic conditions. h

Methods Plasmid construction.

The basic bicistronic vector pβgal/CAT was constructed by inserting the β-galactosidase gene (NotI fragment) from plasmid pCMVβ (Clontech) and CAT gene (XbaI–BamHI fragment) from plasmid pCATbasic (Promega) into the linker region of plasmid pcDNA3 (Invitrogen). Two cistrons are separated by 100 base pairs of the intercistronic linker region containing a unique XhoI site. The expression of bicistronic mRNA is driven by a cytomegalovirus (CMV) promoter. Monocistronic plasmids were constructed by deleting the βgal cistron (NotI fragment) from respective plasmids. The promoterless CAT reporter plasmid pCATbasic/UTR was constructed by inserting the indicated 5′ UTR into the pCATbasic vector. The expression plasmid pCI-IRES/XIAP was constructed by inserting the 1kb 5′ UTR of XIAP in front of the XIAP coding region in the plasmid pCI (Invitrogen). The XIAP 5′ UTR elements of human and mouse XIAP were obtained by reverse transcription with polymerase chain reaction using human and mouse fetal liver Marathon-Ready cDNAs (Clontech) and XIAP primers containing a XhoI site. 5′ UTR clones were inserted into the XhoI site of the intercistronic linker region of plasmid pβgal/CAT. Plasmids with a mutated PPT were constructed by PCR-directed mutagenesis. The orientation and the proper sequence of the 5′ UTR fragments were confirmed by sequencing.

Cell culture and transient DNA transfections. NIH3T3, HeLa, HEL299, 293 and SH-SY5Y cells were cultivated in DMEM medium; H460, H520, H661 and SKOV3 cells were grown in RMPI medium. All media were supplemented with 10% fetal calf serum (FCS) and antibiotics. Transient DNA transfections were done using lipofectamine reagent (Gibco; HeLa

and NIH3T3 cells) or the SuperFect transfection reagent (Qiagen; H661 cells) and the manufacturers’ recommended procedures. Briefly, cells were seeded at a density of 1 × 105 per 35-mm well and transfected 24 h later in serum-free OPTI-MEM medium with 2 µg DNA and 10 µl lipofectamine or 30 µl SuperFect per well. The transfection mixture was replaced 4 h later with DMEM supplemented with 10% FCS. For serum-deprivation experiments, the cells were washed with PBS 24 h post-transfection and serum-free DMEM was used for subsequent growth. Cells were collected 48 h post-transfection and the cell extracts analysed for βgal and CAT activities. For the irradiation experiment, the cells were incubated for 12 h after transfection and then irradiated with 60Co γ-rays at a dose rate of ~1.5 Gy min–1; the βgal and CAT activities were determined 12 h post-irradiation.

Northern and western blot analysis. Total RNA was prepared by guanidine isothiocyanate/phenol-chloroform extraction using the TRIzol reagent (Gibco) according to the manufacturer’s protocol. RNA was denatured in formamide and separated on 0.8% agarose gel. RNA was then transferred onto a nylon membrane (Biodyne) and hybridized with a XIAP DNA probe labelled with 32P using the Rediprime random primer labelling kit (Amersham). Membranes were exposed onto an X-ray film (Kodak) overnight using an intensifying screen (Amersham). Total cell protein extracts were prepared in 20 mM Tris-Cl (pH 7.5), 5 mM EDTA and 1 mM phenylmethylsulphonyl fluoride (PMSF) by sonication and were cleared by centrifugation at 10,000g for 10 min. The supernatants were loaded onto nitrocellulose membrane using the Bio-Rad slot blot apparatus. Membranes were probed with the rabbit polyclonal anti-human-XIAP antibody.

β-Galactosidase and CAT analysis. Transiently transfected cells were collected in PBS 48 h post-transfection and cell extracts were prepared by the freeze–thaw method as described12. βGal enzymatic activity in cell extracts was determined by a spectrophotometric assay using o-nitrophenol-β-D-galactoside12; CAT activity was determined by a liquid scintillation method as described13.

Cell-death and cell-survival assays.

HeLa cells were seeded at a density of 3 × 105 cells per 35-mm well and transfected as described above. 24 h after transfection, cells were trypsinized and plated on 96-well plates at a density of 3 × 103 cells per well. Cells were washed with serum-free DMEM 24 h post-trypsinization and were subsequently grown in serum-free DMEM. Viability of the cells at different time intervals was assessed by the colorimetric assay based on the cleavage of the tetrazolium salt WST-1 (Boehringer Mannheim) by mitochondrial dehydrogenases in viable cells, using the manufacturer’s protocol. The fractions of surviving cells were calculated from three separate experiments done in triplicate. RECEIVED 12 FEBRUARY 1999; REVISED 22 APRIL 1999; ACCEPTED 5 MAY 1999; PUBLISHED 18 JUNE 1999.

1. 2. 3. 4. 5. 6. 7.

White, E. Genes Dev. 10, 1–15 (1996). Deveraux, Q. L. & Reed, J. C. Genes Dev. 13, 239–252 (1999). LaCasse, E. C., Baird, S., Korneluk, R. G. & MacKenzie, A. E. Oncogene 17, 3247–3259 (1998). Liston, P., Young, S. S., Mackenzie, A. E. & Korneluk, R. G. Apoptosis 2, 423–441 (1997). Liston, P. et al. Nature 379, 349–353 (1996). Deveraux, Q. L., Takahashi, R., Salvesen, G. S. & Reed, J. C. Nature 388, 300–304 (1997). Merrick, W. C. & Hershey, J. W. B. in Translational Control (eds Hershey, J. W. B., Mathews, M. B. & Sonenberg, N.) 31–69 (Cold Spring Harb. Lab. Press, Cold Spring Harbor, 1996). 8. Sachs, A. B., Sarnow, P. & Hentze, M. W. Cell 89, 831–838 (1997). 9. Pelletier, J. & Sonenberg, N. Nature 334, 320–325 (1988). 10. Ehrenfeld, E. in Translational Control (eds Hershey, J. W. B., Mathews, M. B. & Sonenberg, N.) 549–573 (Cold Spring Harb. Lab. Press, Cold Spring Harbor, 1996). 11. Sonenberg, N. in Translational Control (eds Hershey, J. W. B., Mathews, M. B. & Sonenberg, N.) 245–269 (Cold Spring Harb. Lab. Press, Cold Spring Harbor, 1996). 12. MacGregor, G. R., Nolan, G. P., Fiering, S., Roederer, M. & Herzenberg, L. A. in Methods in Molecular Biology (eds Murray, E. J. & Walker, J. M.) 217–235 (Humana, Clifton, New Jersey, 1991). 13. Seed, B. & Sheen, J. Y. Gene 67, 271–277 (1988). ACKNOWLEDGEMENTS We thank G. Belsham and P. Liston for plasmids pCMV-2A and pCI-lacZ; J.-Y. Xuan and C. Neville for sequencing; and R. Farahani and M. Legace for the mouse and human XIAP genomic clones. We acknowledge support from the Medical Research Council (MRC) of Canada, the Canadian Networks of Centers of Excellence and the Howard Hughes Medical Institute (HHMI). M.H. is a recipient of an MRC Postdoctoral Fellowship. R.G.K. is a recipient of an MRC Senior Scientist Award, a Fellow of the Royal Society of Canada, and an HHMI International Research Scholar. Correspondence and requests for materials should be addressed to R.G.K.

192

