Multi-sites cleavage of leukemogenic AML1-ETO fusion protein by caspase-3 and its contribution to increased apoptotic sensitivity


Leukemia-associated fusion protein AML1-ETO is a product of the chromosome translocation (8;21) frequently occurred in acute myeloid leukemia (AML). The fusion oncoprotein blocks leukemic cell differentiation, and it also induces growth arrest with increased sensitivity to apoptosis induction. Such dichotomous functions make it difficult to clarify the role of AML1-ETO in leukemogenesis. Here, we systematically showed that constitutively and overexpressed AML1-ETO protein was cleaved to four fragments of 70, 49, 40 and 25 kDa by activated caspase-3 during apoptosis induction by extrinsic mitochondrial and death receptor signaling pathways. The in vitro proteolytic system combined with MALDI-TOF/TOF mass spectrometer confirmed that AML1-ETO and wild-type ETO but not RUNX1 (AML1) proteins were direct substrates of apoptosis executioner caspase-3. Site-directed mutagenesis analyses identified two nonclassical aspartates (TMPD188 and LLLD368) as caspase-3-targeted sites in the AML1-ETO sequence. When these two aspartates were mutated into alanines, more intriguingly, the apoptosis-amplified action of AML1-ETO induction completely disappeared, while inducible expression of the caspase-3-cleaved 70 kDa fragment of AML1-ETO after tetracycline removal is sufficient to enhance apoptotic sensitivity. Further investigations on the potential in vivo effects of such a cleavage and its possible role in leukemogenesis would provide new insights for understanding the biology and treatment of AML1-ETO-associated leukemia.


Acute myeloid leukemia (AML) is characterized by the blockage of the maturation/differentiation of myeloid progenitor cells at different stages with the uncontrolled proliferation and survival advantage of malignant leukemic cells. Leukemogenesis has been attributed to acquired genetic changes, especially specific reciprocal chromosome translocations.1 The t(8;21) is the most frequent translocation in AML, which is associated with 12% of de novo AML cases and up to 40% of AML with granulocytic differentiation, M2 subtype by French–American–British classification. It creates a fusion protein that comprises the conserved runt homology domain (RHD) from the hematopoietic transcription factor RUNX1 (also known as AML1, CBFα2 or PEBP2αB) joined to the majority of the ETO (eight-twenty one, also known as RUNX1T1 or MTG8) repressor protein.2, 3 Various kinds of mouse models that had been engineered to study AML1-ETO function failed to reliably produce overt leukemia in mice and thus it was widely accepted that additional genetic events, such as mutations of the Wilms' tumor gene WT1, Jak2, c-kit, FLT3 are required for the onset of AML,4, 5, 6, 7 which makes it difficult to clarify the role of AML1-ETO in leukemogenesis.

Almost all reports documented that AML1-ETO can impair differentiation of hematopoietic cells.8, 9, 10 However, inducible expression of AML1-ETO in leukemic U937 cell line was shown to induce growth arrest,11 which is associated with AML1-ETO-induced connexin-43 up-expression.12, 13 Also, loss of the negative cell cycle regulator p21WAF1 gene facilitates AML1-ETO-induced leukemogenesis.14 On the other hand, AML1-ETO endows leukemic cells with susceptibility to apoptosis induced by NSC606985 (a rarely studied camptothecin analog),15 anti-Fas antibody, ultraviolent (UV)16 as well as oridonin (a compound extracted from medicinal herbs).17 Recently, AML1-ETO protein was also shown to be degraded in apoptosis induced by eriocalyxin B and oridonin.17, 18 In addition, apoptosis-independent degradation of AML1-ETO was also reported.19 The exact mechanisms and biological significance of AML1-ETO degradation remain to be further explored.

In this work, we report that AML1-ETO and wild-type ETO rather than wild-type RUNX1 are cleaved directly by caspase-3, an apoptosis-critical executing enzyme.20 Furthermore, two nonclassical sites of AML1-ETO specifically cleaved by caspase-3 are identified through site-directed mutagenesis and a matrix-assisted laser desorption ionization time of flight time of flight (MALDI-TOF/TOF) mass spectrometer. More intriguingly, we show that the proteolytic cleavage of AML1-ETO is essential for the apoptosis-enhancing effect of AML1-ETO protein identified previously.16 These data would shed new insights for understanding the complicated biology and treatment of AML1-ETO-associated leukemia.

Materials and methods

Leukemic cell lines and apoptosis induction

All leukemic cell lines, including Kasumi-1 and SKNO-1, promonocytic leukemia U937-derived U937 T and U937-A/E 9/14/18, were grown in RPMI-1640 medium (Sigma-Aldrich, St Louis, MO, USA) supplemented with 10% heat-inactivated fetal calf serum (Gibco BRL, Gaithersburg, ML, USA) in 5% CO2/95% air humidified atmosphere at 37 °C. For SKNO-1 cells, 10 ng ml−1 of granulocyte–macrophage colony-stimulating factor was added. For AML1-ETO induction, 5 μM ponasterone A (Invitrogen, Groningen, The Netherlands) was added to U937-A/E 9/14/18 cells, which was a generous gift from Dr. Michael Lübbert in University of Freiburg Medical Center (Freiburg, Germany).21 For apoptosis induction, cells were treated with mouse monoclonal anti-Fas IgM antibody CH11 (Upstate, Waltham, MA, USA), NSC606985 (kindly provided by National Cancer Institute Anticancer Drug Screen standard agent database, dissolved in double distilled water as a 25 μM stock solution) or were exposed to a germicidal lamp providing predominantly 254-nm UV-C light (Philips TUV G30T8 30 W bulb) with or without caspase-3 inhibitor Z-DEVD-fmk and pan-caspase inhibitor Z-VAD-fmk (BD Biosciences, San Diego, CA, USA) dissolved in DMSO before use. Annexin V assay was performed on a flow cytometry (Beckman Coulter, Miami, FL, USA) according to instructions provided by the ApoAlert Annexin V kit (Clontech, Palo Alto, CA, USA).


Human AML1-ETO cDNA was amplified from U937-A/E 9/14/18 cells by reverse transcription (RT)–PCR and then cloned into pGEX-4T2 vector (Amersham Biosciences, Buckinghamshire, England) to construct GST-tagged AML1-ETO protein-expressing plasmid or into pCMV4 vector (Sigma-Aldrich) and pEGFP-C1 vector (Clontech) to generate pCMV4-AML1-ETO and pEGFP-AML1-ETO. All point mutants of AML1-ETO (D99A, D133A, D171A, D188A, D192A, D368A, D376A, D383A, D473A, D549A) were created on pGEX-4T2-AML1-ETO plasmid utilizing QuikChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA). Full-length (wild-type and mutants) and fragments (1–188 and 189–752) of AML1-ETO cDNA with EGFP tag were amplified by PCR from pEGFP-AML1-ETO plasmids and subcloned into pTRE2hyg expression vector (Clontech) to form the corresponding plasmids. The pGEX-4T2-RUNX1 and pGEX-4T2-ETO vectors were subcloned from the pCMV5-RUNX1, pCMV5-ETO vectors which were generous gifts from Dr. Hiebert.22 To obtain active His-tagged caspase-3, human caspase-3 cDNA lacking N-terminal 28 amino acid residues was amplified from U937 cells and subcloned into pQE30 (Qiagen, Valencia, CA, USA) vector according to Lee et al.23 The sequences of all cDNA inserts of plasmids were confirmed by sequencing.

Immunoprecipitation (IP) of FLAG-AML1-ETO protein

As described previously, 293T cells, which were cultured in DMEM medium (Sigma-Aldrich), were transiently transfected with pCMV4-FLAG-AML1-ETO (1 μg) expression vectors by Polyfect (Qiagen).24 Forty-eight hours later, whole-cell extracts were prepared in 300 μl of lysis buffer (50 mM HEPES, 50 mM NaCl, 0.1% Tween 20, 10% glycerol, 20 mM sodium pyrophosphate, 1 mM dithiothreitol, plus protease inhibitors) and incubated overnight with anti-FLAG M2 antibody (Sigma-Aldrich) at 4 °C. After IP, the beads were intensively washed and the precipitated AML1-ETO protein was gotten.

Purification of recombinant caspase-3 and in vitro proteolysis of AML1-ETO protein

His-tagged caspase-3 protein were expressed in bacteria BL21 (DE3) by induction with 1 mM isopropylthiogalactopyranoside (IPTG) at 30 °C and purified from the cytosol of the expressed cells by affinity chromatography on Ni-NTA-agarose (Qiagen). The GST-AML1-ETO or its mutants were expressed in BL21 by induction with IPTG at 28 °C and purified using Bulk and RediPack GST Purification Modules (Amersham Biosciences). Immunoprecipitated FLAG-AML1-ETO protein from transfected 293T cells or purified bacterially expressed GST-AML1-ETO protein and its mutants were incubated with purified recombinant caspase-3 (0.2 μg) in 100 mM HEPES (pH 7.2) containing 10 mM DTT and 10% (v/v) glycerol at 25 °C for 120 min. The reaction was stopped by the addition of an equal volume of SDS–PAGE sample buffer and then was subjected to western blot.

Mass spectrometer and database search for cleaved fragments

The protein spots were cut out of gels using Gelpix Spot-Excision Robot (Genetix, Hampshire, UK), and then washed three times with Milli-Q water. The gel pieces were dried in a vacuum, and the proteins were digested overnight in 10 ml trypsin (10 ng ml−1, Trypsin Gold, mass spectrometry grade, Promega, Madison, WI, USA) in 25 mM ammonium bicarbonate at 37 °C. The peptide fragments were extracted with 0.2% trifluoroacetic acid (TFA) for 30 min, applied onto the C18 resin and then desalted with 0.2% TFA. Finally, the tryptic peptide mixtures were recovered with 5 ml elution solution containing 50% acetonitrile (ACN)/0.1% TFA by centrifugation (1750 g, 15 s). Tryptic peptides were lyophilized and resuspended in 2 μl matrix solution containing 10 mg ml−1 α-cyano-4-hydroxycinnaminc acid (CHCA) prepared in 50% ACN/0.1% TFA, and 0.7 μl sample was spotted onto the MALDI sample target plate. Peptide mass spectra were obtained on a MALDI-TOF/TOF mass spectrometer (4700 Proteomics Analyzer, Applied Biosystems, Foster, CA, USA). Before the real sample acquisition, six calibrated spots were used for signal and parameter optimization. The peptide mass fingerprinting (PMFs) were obtained in the mass range between 800 and 4000 Da with approximately 5000 laser shots. To obtain the spectra with a mass accuracy of less than 25 ppm, trypsin and keratin autolytic peaks were used for internal calibration of the mass spectra. Protein identification was processed and analyzed by searching the Swiss-Prot and NCBI nr protein database using the MASCOT search engine of Matrix Science that integrated in the Global Protein Server (GPS) Workstation. The mass tolerance, the most important parameter, was limited to 50 ppm. The results from the MS spectra were accepted as a good identification when the GPS score confidence was higher than 95%.

Establishment of stable transformants

To generate U937-AML1-ETO stable transformants, 1 × 107 of U937T cells were washed in RPMI 1640 medium and resuspended in 0.2 ml of Isceve's Modified Dulbecco's medium without fetal bovine serum (FBS), pTRE2hygAML1-ETO (wild-type and mutants) plasmids (20 μg) in 20 μl ddH2O was transferred to electroporation cuvette with a 0.4 cm gap (Bio-Rad, Hercules, CA, USA). Electroporation was performed using a Gene-Pulser II (Bio-Rad) at 170 V and 960 μF. The samples were then transferred to complete RPMI-1640 medium. Twenty-four hours later, 1 μg ml−1 of tetracycline, 0.5 μg ml−1 of puromycin and 500 mg ml−1 of hygromycin B (Clontech) were added and cells were continued to incubate at 37 °C in 5% CO2. Positive polyclonal populations were identified based on the induction of EGFP-tagged protein expression following tetracycline withdrawal by western blot and flow cytometry.

Fluorescence-activated cell sorting (FACS)

Cell sorting was performed using a Mo-Flo DakoCytomation (Fort Collins, CO, USA) equipped with an Omnichrome series 43 air-cooled 488 nm argon laser at 100 mW of power. The EGFP signal was collected in the FL1 channel through a 530/40 bandpass filter. Cells in the gate were displayed in a single-parameter histogram for the EGFP and collected. Post-sorting cells were collected in RPMI-1640 medium with 10% FBS at a concentration of 50 000 cells per well. Cells were collected onto slides by cytospin (Shandon, Runcorn, UK) and observed on a Bio-Rad MRC-1024 laser scanning confocal microscope equipped with a Zeiss × 60 objective (Olympus, Tokyo, Japan).

Western blot

Protein extracts were equally loaded on 10–12% SDS–polyacrylamide gel, and transferred to nitrocellulose membrane (Amersham Bioscience). The blots were stained with 0.2% Ponceau S red to ensure equal protein loading. After blocking with 5% nonfat milk in PBS, the membranes were incubated with antibodies against RUNX1, ETO (C-20) (Santa Cruz, CA, USA), active caspase-3, caspase-8, poly-ADP ribose polymerase (PARP), protein kinase Cδ (PKCδ), GFP (Cell Signaling, Beverly, MA, USA) and GST (Amersham Bioscience), followed by horseradish perioxidase (HRP)-linked secondary antibodies (Cell Signaling). Detection was performed by chemiluminescence phototope-HRP kit (Cell Signaling).

Statistical analysis

The Student's t-test was used to compare the difference between two different groups. A value of P<0.05 was considered to be statistically significant.


AML1-ETO protein is cleaved into four fragments in apoptotic leukemic cells induced by intrinsic mitochondrial and extrinsic death receptor signaling pathways

U937-A/E 9/14/18 subclone, which was generated by introducing ecdysone-inducible two-vector system into human leukemic U937 cells,12, 21 conditionally expressed the AML1-ETO protein when it was treated by ponasterone A (Figure 1a). Consistent with that reported previously,16 NSC606985 and UV significantly triggered U937-A/E 9/14/18 cells to undergo apoptosis in the presence of AML1-ETO induction, which slightly but statistically significantly increased ‘spontaneous’ apoptosis in U937-A/E 9/14/18 cells (Figure 1b). To our surprise, four small fragments respectively with molecular weight of about 70, 49, 40 and 25 kDa could be detected by anti-ETO antibody during apoptosis (Figure 1a, top panel). The intensities of these fragments were closely related to apoptotic degree: low amount of fragments were seen in spontaneous apoptotic cells; UV light, which presented the stronger apoptosis induction than NSC606985 in spite of the presence or absence of AML1-ETO induction, induced the most significant cleavage of AML1-ETO protein. When the blot was stripped and reprobed with anti-RUNX1 N-terminal antibody, only one 20 kDa of fragment was detected besides intact 85 kDa AML1-ETO protein (Figure 1a, bottom panel).

Figure 1

AML1-ETO protein is cleaved during apoptosis induced by NSC606985 and UV. After pretreatment with or without 5 μM ponasterone A for 48 h, U937-A/E 9/14/18 cells were incubated with 50 nM of NSC606985 (NSC) for 24 h or irradiated by 400 J m−2 UV followed by 10 h culture. Then AML1-ETO protein was detected by western blot with antibodies against ETO (top, a) and N-terminus of RUNX1 respectively (bottom, a). The open triangles point to fragments with the indicated masses of AML1-ETO. (b) For assessment of apoptosis, annexin V assay was performed by flow cytometry. The numbers represent the mean±s.d. of the annexin V+ apoptotic cells % of triplicate samples in an independent experiment. Symbols * represent P<0.05 compared with the corresponding ponasterone A-untreated cells, and & or # represent P<0.01 for the comparison between untreated and NSC606985-treated cells or UV-treated and NSC606985-treated cells in the presence of ponasterone A. All experiments were repeated three times with similar results.

Apoptosis can be initiated by the intrinsic mitochondrial pathway and transmembrane death receptors-related extrinsic pathway.25 As documented,15, 16 NSC606985 and UV-induced apoptosis is mainly mediated by the intrinsic mitochondrial pathway. To understand if AML1-ETO cleavage also occurs in apoptosis triggered by death receptor, U937-A/E 9/14/18 cells with 48 h AML1-ETO induction were incubated with anti-Fas antibody CH11, which induced caspase-8/3 activation and PKCδ cleavage (Figure 2a), indicating the induction of apoptosis. Meanwhile, AML1-ETO was also cleaved into four fragments (Figure 2a), which were the same as those in NSC606985 and UV-induced apoptotic cells (Figure 1a).

Figure 2

The inhibition of caspase-3 blocks the cleavage of AML1-ETO protein in apoptotic U937-A/E 9/14/18 cells and Kasumi-1 cells. (a) After pretreatment with 5 μM ponasterone A for 48 h, U937-A/E 9/14/18 cells were incubated with or without 40 μM Z-DEVD-fmk or Z-VAD-fmk for one hour. Then, cells were exposed to 50 ng ml−1 CH11 for additional 6 h. The indicated proteins were detected by western blot, in which anti-ETO antibody was used for AML1-ETO assay. ΔPKCδ indicates cleaved 41 kDa fragment of PKCδ protein. ΔCaspase points to active fragments of caspase-3 or caspase-8. The open triangles point to fragments with the indicated masses of AML1-ETO. (b and c) Kasumi-1 and SKNO-1 cells were irradiated without or with 400 J m−2 UV (b) or after pre-incubated with or without 40 μM Z-DEVD-fmk for 1 h, Kasumi-1 cells were irradiated with 400 J m−2 UV (c). Eight hours later, AML1-ETO protein was detected by anti-ETO antibody. Symbol * represents positive U937-A/E 9/14/18 cell lysate treated with 50 ng ml−1 CH11 for 12 h after pretreatment of 5 μM ponasterone A for 48 h (c). The blot in panel (b) was also tested for active caspase-3 and poly-ADP ribose polymerase (PARP). ΔPARP indicates cleaved 89 kDa fragment of PARP protein. All experiments were repeated for three times and the same results were obtained.

To exclude the possibility of artifact associated with engineered transfected cell model, we also tested two leukemic cell lines from patients with t(8;21), Kasumi-1 and SKNO-1 cells,26, 27 which expressed endogenous AML1-ETO fusion protein. Towards this end, these two cell lines were radiated by UV light, which could effectively induce apoptosis, as evidenced by caspase-3 activation and cleavage of caspase-3-specific substrate PARP (Figure 2b) and annexin V assay (data not shown). Upon apoptosis induction, the endogenously expressed AML1-ETO protein was also cleaved into four fragments that could be blotted by anti-ETO antibody (Figure 2b), like those seen in U937-A/E 9/14/18 subclone with AML1-ETO induction. Of note, ponasterone A did not induce Kasumi-1 and SKNO-1 cells to undergo apoptosis. Correspondingly, the compound also did not induce cleavage of the endogenous AML1-ETO protein (data not shown), supporting that cleavage of AML1-ETO protein in U937-A/E 9/14/18 as seen above was specific for apoptosis induction but not due to effects of ponasterone A.

Active caspase-3 induces proteolytic cleavage of AML1-ETO protein

Considering that all three kinds of apoptosis induction tested in this work were accompanied with caspase-3 activation, we extrapolated that AML1-ETO cleavage could be performed by caspase-3. To address this, the cell-permeable caspase-3 inhibitor Z-DEVD-fmk and pan-caspases inhibitor Z-VAD-fmk were used to effectively inhibit CH11-induced caspase-3/8 activation and thus apoptosis in U937-A/E 9/14/18 cells with AML1-ETO induction (Figure 2a). Accordingly, AML1-ETO cleavage was also effectively blocked in this system (Figure 2a). Moreover, Z-DEVD-fmk also inhibited UV light-induced cleavage of AML1-ETO protein in Kasumi-1 cells, except a 49 kDa fragment that became stronger in the presence of caspase-3 inhibition (Figure 2c).

The experiments performed above strongly proposed that cleavage of AML1-ETO protein is mediated by active caspase-3 during apoptosis. To determine whether AML1-ETO is in fact a substrate for caspase-3, human recombinant active caspase-3 was incubated in vitro with immunoprecipitated FLAG-AML1-ETO protein from transfected 293T cells or bacterially expressed GST-AML1-ETO protein. The results demonstrated that the recombinant active caspase-3 only cleaved FLAG-AML1-ETO protein to a 70 kDa fragment (Figure 3a), and GST-AML1-ETO protein into two bands respectively with 70 kDa and 49 kDa (Figure 3b), as determined by immunoblot with anti-ETO antibody. When anti-GST N-terminus antibody was used to blot, moreover, GST-AML1-ETO protein was cleaved into 68 kDa and 46 kDa fragments (Figure 3c), which paralleled to two bands of C-terminus blotted by anti-ETO antibody (Figure 3b). Furthermore, the samples in Figure 3b were electrophoresized and stained with Coommassie Brilliant Blue. Then three bands respectively with about 115 kDa, 70 kDa and 46 kDa were analyzed by mass spectrometry. As expected, band presented the full-length AML1-ETO, while bands and corresponded respectively to C-terminal and N-terminal of AML1-ETO protein (Figure 3d). When Z-DEVD-fmk was added into these systems, all cleaved fragments disappeared (Figures 3a, b and d), indicating the specificity of caspase-3-cleaved AML1-ETO protein.

Figure 3

In vitro cleavage of AML1-ETO protein by recombinant caspase-3. (a–c) 293T cells were transfected by FLAG-AML1-ETO. Twenty-four hours later, FLAG-AML1-ETO protein was immunoprecipitated against anti-FLAG antibody. The precipitated FLAG-AML1-ETO (a) or bacterially expressed GST-AML1-ETO protein (0.1 mg, b and c) was incubated with or without recombinant active caspase-3 (0.2 μg) in the presence and absence of 4 mM Z-DEVD-fmk at 25 °C for 4 h. The samples were analyzed by immunoblotting with antibodies against ETO and active caspase-3 (a and b) or GST (c). Symbol * represents the same positive control as Figure 2. The open triangles point to fragments with the indicated masses of AML1-ETO. (d) Bacterially expressed GST-AML1-ETO protein (0.1 mg) was incubated without or with recombinant active caspase-3 (0.2 μg) in the presence or absence of 4 mM Z-DEVD-fmk at 25 °C for 4 h. Then, the mixtures were denatured and separated on SDS–polyacrylamide gels, followed by staining with Coomassie Brilliant Blue. The full-length GST-AML1-ETO protein () and two fragments ( and ) displayed differentially between lane 2 and lane 3 were cut out for identification by MALDI-TOF/TOF MS. Asterisks indicate AML1-ETO-matched peptides, while numbers and arrows point to exact sequences matched to the peptides of AML1-ETO protein.

Identification of caspase-3-targeted sites in AML1-ETO protein

To map the cleavage sites of AML1-ETO by caspase-3, we investigated whether wild-type RUNX1 and ETO are cleaved by caspase-3. As shown in Figure 4a, recombinant active caspase-3 degraded the bacterially expressed GST-ETO protein to two bands respectively with 70 kDa and 49 kDa, which were the same as the cleaved AML1-ETO fragments (Figure 3b). However, purified GST-RUNX1 protein was not disrupted by caspase-3 (Figure 4b). These results suggested that ETO protein was also a substrate of caspase-3. In other words, the caspase-3-cleaved sites in AML1-ETO protein should be located on the ETO portion of the fusion protein.

Figure 4

Wild-type ETO but not RUNX1 protein is cleaved by in vitro recombinant caspase-3. Bacterially expressed GST-ETO (0.1 mg, a) or RUNX1 (0.1 mg, b) was incubated without or with recombinant active caspase-3 (0.2 μg) at 25 °C for 4 h. The samples were detected by western blot with antibodies respectively against ETO (a) and GST (b). The open triangles point to fragments with the indicated masses of ETO. All experiments were repeated for three times and the same results were obtained.

Substantial studies established that caspase-3 substrates are recognized and cleaved due to presence of the consensus DXXD sequence (here ‘D’ is aspartate and X is any amino acid).28, 29 Inspection of AML1-ETO sequence revealed the presence of several classical (DXXD) and nonclassical (XXXD) motifs. According to molecular weight of the AML1-ETO fragments cleaved by caspase-3, 10 aspartates in AML1-ETO were respectively mutated to alanines (A), in which D99, 133 and 171 are located in RUNX1 moiety and other 7 (D188, 192, 368, 376, 383, 473 and D549) in ETO part (Figure 5a). The purified GST-tagged wild-type AML1-ETO and the mutant proteins were incubated with recombinant active caspase-3 and then analyzed by immunoblots. The results showed that among these mutants, D188A mutant could not be cleaved into a 70 kDa fragment, while D368A mutant failed to be cut into a 49 kDa fragment by caspase-3. Furthermore, the D188A and D368A double mutant was almost not cleaved at all by caspase-3 (Figure 5b). Collectively, D188 and 368 are essential sites for caspase-3-catalyzed proteolysis of AML1-ETO protein.

Figure 5

Identification of caspase-3-targeted sites of AML1-ETO protein. (a) Schematic diagram of human AML1-ETO protein. Numbers on the top represent sites of amino acid residues of the full-length AML1-ETO protein. The mutated sites (D → A) are shown on the bottom. Two identified caspase-3-cleaved sites are shown with its four amino acid sequences in the bold. (b) Bacterially expressed GST-tagged wild-type (WT) and mutated AML1-ETO protein (0.1 mg each) were incubated without or with recombinant active caspase-3 (0.2 μg) at 25 °C for 4 h. The mixtures were detected by western blot with anti-ETO antibody. The open triangles point to fragments with the indicated masses of AML1-ETO. All experiments were repeated for three times with the same results.

Proteolytic cleavage of AML1-ETO protein by caspase-3 amplifies apoptosis induction

To determine the functional significances of the AML1-ETO cleavage, U937T cell lines with inducible expression of wild-type, mutated (D368A and D188, 368A) AML1-ETO as well as its two fragments (1–188 and 189–752) were generated. For this purpose, these EGFP-tagged cDNAs were inserted into plasmid pTRE2hyg, a tetracycline-responsive expression vector, followed by transfection into the U937T cells, the latter containing stably transfected pUHD-tTA (tetracycline-responsive transcription activator), whose expression is under the control of tetracycline.30 In principle, the expression of tTA and targeted proteins should be extremely low in the tetracycline-containing medium. In the absence of tetracycline, on the contrary, tTA activates its own promoter to produce more tTA, the latter inducing targeted protein expression. As shown in Figure 6a, all the stable transformants were induced to express EGFP or EGFP-tagged AML1-ETO or its mutants upon tetracycline withdrawal for 5–7 days, as evidenced by western blots with anti-GFP antibody. Furthermore, these stable transformants were subjected to flow sorting based on the fluorescence of EGFP to remove some EGFP-free cells. All post-sorting cells were EGFP-positive with over 90% of viability. Like empty-EGFP transformant, moreover, fragment 1–188 of AML1-ETO was in the cytoplasm due to the absence of the nuclear localization signal (NLS, Figure 5a). Other transformants, including full-length wild-type or mutated AML1-ETO and its 189–752 fragment, were localized into nuclei (Figure 6b).

Figure 6

Generation of U937T cells with conditional expression of AML1-ETO or its mutants and effects of three inducible expressions on apoptosis induction. (a and b) U937T cells transfected with wild-type (WT) AML1-ETO and its mutants as indicated were grown in the tetracycline-free medium for days as indicated, and western blots were performed with anti-GFP antibody (a), and flow sorted stable cells were collected onto slides by cytospin and visualized by confocal microscope (b). (c and d) After grown in tetracycline-free medium for seven days, the flow sorted U937T cells were treated by 200 nM of NSC606985 for 24 h or 50 ng ml−1 CH11 for 12 h. Then, annexin V assay was performed on flow cytometry (c) and annexin V+ apoptotic cells % were shown (d). Each column and bar represents the mean and s.d. of triplicates in an independent experiment. Symbols # and * represent P<0.01 compared with vehicle cells and WT cells, respectively.

As reported11, 17 and shown in Figure 1b, AML1-ETO induction endows leukemic cells with susceptibility to apoptosis. Here, we investigated whether the cleavage of AML1-ETO by caspase-3 contributes to this event. For this, these sorted U937T transformants were treated with NSC606985 at 200 nM for 24 h or 50 ng ml−1 CH11 for 12 h, and annexin V+ apoptotic cells were measured. Consistent with our previous report,16 induction of wild-type AML1-ETO significantly increased NSC606985 or CH11-induced apoptosis (Figures 6c and d). More intriguingly, such an enhanced activity of AML1-ETO was lost under the D188, 368A double mutations (Figures 6c and d), which could not be cleaved by caspase-3 (Figure 5). Furthermore, the increased apoptotic sensitivity could still be seen in the induction of mutant D368A (Figures 6c and d). Because D368A AML1-ETO mutant was still cleaved into a 70 kDa of fragment (Figure 5b), we postulated that this fragment possibly mediates apoptosis-enhancing effect of AML1-ETO. Expectedly, fragment 1–188 of AML1-ETO failed to impinge on apoptosis induction (data not shown), while fragment 189–752 of AML1-ETO still increased NSC606985/CH11-induced apoptosis to the similar degree to wild-type AML1-ETO (Figures 6c and d).


In most cases, the activated caspase-3 protein is a critical executioner for apoptosis through inducing proteolytic cleavage of many substrates, although the caspases-independent apoptosis was also proposed.31 Because of wide employment of various methods, in particular a systematic proteome analysis of apoptotic cells, a bewildering number of substrate of caspase-3 were identified over the past 10 years, as summarized by Fischer et al.29 In this work, we showed that AML1-ETO protein, in spite of either inducible overexpression or constitutive expression, were cleaved to four fragments with different molecular weights during apoptosis induction by both intrinsic mitochondrial and extrinsic death receptor. When caspase-3 action was blocked by its inhibitors, these cleaved fragments disappeared, strongly suggesting that caspase-3 possibly directly or indirectly contributes to apoptosis-related cleavage. More recently, Chen and coworkers also reported that the degradation of AML1-ETO oncoprotein was paralleled to caspase-3 activation in apoptotic Kasumi-1 cells induced by eriocalyxin B and oridonin.17, 18 Based on these discoveries, we further used an in vitro system to identify potential cleavage sites in AML1-ETO with human recombinant active caspase-3. The results revealed that the recombinant active caspase-3 was able to cleave both immunoprecipitated FLAG-AML1-ETO and bacterially expressed GST-AML1-ETO protein into one or two fragments, which also existed in ex vivo system and was blocked by caspase-3 inhibitor. Moreover, mass spectrometry analysis confirmed that these fragments derived from AML1-ETO protein. All these experiments strongly supported that AML1-ETO protein is a direct caspase-3 target.

AML1-ETO protein encompasses the 177 amino acid residues of N-terminus of RUNX1 and nearly intact ETO protein.3, 32 Hence, we also tested whether caspase-3 cleaves RUNX1 and/or ETO, although a fusion gene product often exhibits a unique structure and biological activity, both of which cannot be compared to one of parental proteins. Our results showed that ETO but not RUNX1 was also a caspase-3 target because ETO could be cleaved by recombinant caspase-3 into 70 and 49 kDa of fragments, which is consistent to in vitro cleavage of AML1-ETO protein. Indeed, site-directed mutagenesis analysis showed that alanine for aspartate at 99, 133 and 171 in RUNX1 moiety of AML1-ETO fusion protein did not impact cleavage of caspase-3 on the fusion protein. Further, we identified that aspartates at 188 (TMPD) and 368 (LLLD) in ETO part are two effective caspase-3-targeted sites. Here we could not identify other two sites for generation of two small fragments (40 and 25 kDa), although more mutated sites were given. Actually, two smaller AML1-ETO fragments with 40 and 25 kDa that existed in apoptotic cells could not be seen in our in vitro proteolytic system. Although this was possibly due to the lower action of in vitro recombinant caspase-3, we could not exclude the possibility that other protease(s) also have something to do with apoptosis-related cleavage into these two small fragments of AML1-ETO protein. This notion could also be supported by the fact that caspase-3 inhibitor did not block the generation of a 49 kDa fragment by UV-induced apoptosis in Kasumi-1 cells.

The degradation of substrates of caspase-3, such as inhibitor of caspase-activated DNase (ICAD),33 play roles in the key morphological alteration, apoptosis-signaling pathway and their packaging into apoptotic body for phagocytosis by macrophage. For many of the identified substrates, however, the functional consequences of their cleavage are unknown. From a clinical standpoint, AML with t(8;21) or AML1-ETO transcript has a better prognosis with the application of high-dose cytarabine than most other types of AML.34 We speculated that this is related to apoptosis-enhancing effect of AML1-ETO fusion protein, as shown here and previously.16, 17 We tried to test whether AML1-ETO cleavage by caspase-3 affects its apoptosis-enhancing effect. As for this, we generated a series of cell lines with tetracycline withdrawal-induced expression of wild-type AML1-ETO and its mutants. The results showed that the increased apoptotic sensitivity could still be seen in the induction of D368A mutated AML1-ETO, while it was completely disappeared in the presence of D188, 368A double mutations. Based on the fact that D188, 368A double mutant could not be cleaved by caspase-3 while D368A mutant could still be cleaved into a 70 kDa of fragment, we postulated that this fragment possibly mediates apoptosis-enhancing effect of AML1-ETO. Indeed, fragment 189–752 of AML1-ETO still increased NSC606985/CH11-induced apoptosis to a similar degree with wild-type AML1-ETO induction. Therefore, we concluded that caspase-3-cleaved AML1-ETO fragment mediates apoptosis-enhanced action of the fusion protein whose mechanisms remain to be investigated.

The fact that the t(8;21) is the sole cytogenetic abnormality in the majority of AML1-ETO-positive AML cases suggests that AML1-ETO plays a critical role both in the distinctive phenotype and in the establishment of the leukemia clone. However, the mechanism by which AML1-ETO initiates or contributes to AML transformation is not clearly established. As described in the Introduction, various kinds of models including transgenic models, inducible systems and bone marrow transplant strategies proposed that sole expression of AML1-ETO failed to produce overt leukemia in the mice, suggesting that AML1-ETO-associated leukemogenesis might also require a secondary event or ‘hit’ for AML1-ETO-positive cells to adopt leukemogenic behavior. On the other hand, the importance of caspase activation under nonapoptotic conditions is widely recognized, although the mechanisms regulating caspases in nonapoptotic contexts are largely unknown. For instance, caspase activity is also required for various physiological functions beyond cell death, including sperm individualization,35 neural stem cell differentiation,36 osteogenic differentiation of bone marrow stromal stem cells,37 skeletal muscle differentiation, erythrocyte, keratinocyte, lens differentiation and T- and B-cell proliferation.38 Erythropoietin was also shown to activate caspase-3 during erythroid differentiation in TF-1 cells, which provides a mechanism for the first time how cells avoid DNA fragmentation with activated caspase-3.39 Therefore, caspase-3-cleaved fragments in nonapoptotic conditions and their roles in leukemogenesis are open to be explored. In fact, the truncation of AML1-ETO C-terminal N-CoR (nuclear receptor co-repressor)/SMRT (silencing mediator for retinoic acid and thyroid hormone receptor)-interacting domain was shown more recently to eliminate the cellular proliferation defect associated with the full-length oncoprotein, and thus transforms AML1-ETO into a potent leukemogenic protein.40 Because the enhancement of apoptosis does not favor the generation of leukemia harboring the t(8;21) translocation, we proposed that it deserves to be investigated in the future whether caspase-3 targeted sites develop mutation in t(8;21)-related leukemia. This prediction could be supported by a recent report,41 which identified a previously unknown alternatively spliced isoform of the AML1-ETO transcript (AML1-ETO9a) encoding a C-terminally truncated AML1-ETO protein, leading to rapid development of leukemia in a mouse retroviral transduction-transplantation model.

Taken together, our results demonstrated that the activated caspase-3 can cleave AML1-ETO and wild-type ETO proteins at least at two different sites, and caspase-3-cleaved AML1-ETO fragments contribute to apoptosis-enhanced activity of the oncoprotein. Further investigations on the potential in vivo effects of such a cleavage and its possible role in leukemogenesis would provide new understanding for molecular mechanisms of AML-ETO-associated leukemogenesis.


  1. 1

    Bohlander SK . Fusion genes in leukemia: an emerging network. Cytogenet Cell Genet 2000; 91: 52–56.

  2. 2

    Ferrara F, Del Vecchio L . Acute myeloid leukemia with t(8;21)/AML1/ETO: a distinct biological and clinical entity. Haematologica 2002; 87: 306–319.

  3. 3

    Peterson LF, Zhang DE . The 8;21 translocation in leukemogenesis. Oncogene 2004; 23: 4255–4262.

  4. 4

    Nishida S, Hosen N, Shirakata T, Kanato K, Yanagihara M, Nakatsuka S et al. AML1-ETO rapidly induces acute myeloblastic leukemia in cooperation with the Wilms tumor gene, WT1. Blood 2006; 107: 3303–3312.

  5. 5

    Schnittger S, Bacher U, Kern W, Haferlach C, Haferlach T . JAK2 seems to be a typical cooperating mutation in therapy-related t(8;21)/AML1-ETO-positive AML. Leukemia 2007; 21: 183–184.

  6. 6

    Schnittger S, Kohl TM, Haferlach T, Kern W, Hiddemann W, Spiekermann K et al. KIT-D816 mutations in AML1-ETO-positive AML are associated with impaired event-free and overall survival. Blood 2006; 107: 1791–1799.

  7. 7

    Peterson LF, Boyapati A, Ahn EY, Biggs JR, Okumura AJ, Lo MC et al. Acute myeloid leukemia with the 8q22;21q22 translocation: secondary mutational events and alternative t(8;21) transcripts. Blood 2007; 110: 799–805.

  8. 8

    Ahn MY, Huang G, Bae SC, Wee HJ, Kim WY, Ito Y . Negative regulation of granulocytic differentiation in the myeloid precursor cell line 32Dcl3 by ear-2, a mammalian homolog of Drosophila seven-up, and a chimeric leukemogenic gene, AML1/ETO. Proc Natl Acad Sci USA 1998; 95: 1812–1817.

  9. 9

    Kohzaki H, Ito K, Huang G, Wee HJ, Murakami Y, Ito Y . Block of granulocytic differentiation of 32Dcl3 cells by AML1/ETO (MTG8) but not by highly expressed Bcl-2. Oncogene 1999; 18: 4055–4062.

  10. 10

    Tonks A, Tonks AJ, Pearn L, Pearce L, Hoy T, Couzens S et al. Expression of AML1-ETO in human myelomonocytic cells selectively inhibits granulocytic differentiation and promotes their self-renewal. Leukemia 2004; 18: 1238–1245.

  11. 11

    Burel SA, Harakawa N, Zhou L, Pabst T, Tenen DG, Zhang DE . Dichotomy of AML1-ETO functions: growth arrest versus block of differentiation. Mol Cell Biol 2001; 21: 5577–5590.

  12. 12

    Li X, Xu YB, Wang Q, Lu Y, Zheng Y, Wang YC et al. Leukemogenic AML1-ETO fusion protein upregulates expression of connexin 43: the role in AML1-ETO-induced growth arrest in leukemic cells. J Cell Physiol 2006; 208: 594–601.

  13. 13

    Gao FH, Wang Q, Wu YL, Li X, Zhao KW, Chen GQ . c-Jun N-terminal kinase mediates AML1-ETO protein-induced connexin-43 expression. Biochem Biophys Res Commun 2007; 356: 505–511.

  14. 14

    Peterson LF, Yan M, Zhang DE . The p21Waf1 pathway is involved in blocking leukemogenesis by the t(8;21) fusion protein AML1-ETO. Blood 2007; 109: 4392–4398.

  15. 15

    Song MG, Gao SM, Du KM, Xu M, Yu Y, Zhou YH et al. Nanomolar concentration of NSC606985, a camptothecin analog, induces leukemic-cell apoptosis through protein kinase Cdelta-dependent mechanisms. Blood 2005; 105: 3714–3721.

  16. 16

    Lu Y, Xu YB, Yuan TT, Song MG, Lubbert M, Fliegauf M et al. Inducible expression of AML1-ETO fusion protein endows leukemic cells with susceptibility to extrinsic and intrinsic apoptosis. Leukemia 2006; 20: 987–993.

  17. 17

    Zhou GB, Kang H, Wang L, Gao L, Liu P, Xie J et al. Oridonin, a diterpenoid extracted from medicinal herbs, targets AML1-ETO fusion protein and shows potent antitumor activity with low adverse effects on t(8;21) leukemia in vitro and in vivo. Blood 2006; 109: 3441–3450.

  18. 18

    Wang L, Zhao WL, Yan JS, Liu P, Sun HP, Zhou GB et al. Eriocalyxin B induces apoptosis of t(8;21) leukemia cells through NF-kappaB and MAPK signaling pathways and triggers degradation of AML1-ETO oncoprotein in a caspase-3-dependent manner. Cell Death Differ 2007; 14: 306–317.

  19. 19

    Yang G, Thompson MA, Brandt SJ, Hiebert SW . Histone deacetylase inhibitors induce the degradation of the t(8;21) fusion oncoprotein. Oncogene 2007; 26: 91–101.

  20. 20

    Reed JC, Pellecchia M . Apoptosis-based therapies for hematologic malignancies. Blood 2005; 106: 408–418.

  21. 21

    Fliegauf M, Stock M, Berg T, Lubbert M . Williams-Beuren syndrome critical region-5/non-T-cell activation linker: a novel target gene of AML1/ETO. Oncogene 2004; 23: 9070–9081.

  22. 22

    Meyers S, Lenny N, Hiebert SW . The t(8;21) fusion protein interferes with AML-1B-dependent transcriptional activation. Mol Cell Biol 1995; 15: 1974–1982.

  23. 23

    Lee KK, Ohyama T, Yajima N, Tsubuki S, Yonehara S . MST, a physiological caspase substrate, highly sensitizes apoptosis both upstream and downstream of caspase activation. J Biol Chem 2001; 276: 19276–19285.

  24. 24

    Zhao KW, Li X, Zhao Q, Huang Y, Li D, Peng ZG et al. Protein kinase Cdelta mediates retinoic acid and phorbol myristate acetate-induced phospholipid scramblase 1 gene expression: its role in leukemic cell differentiation. Blood 2004; 104: 3731–3738.

  25. 25

    Schimmer AD, Hedley DW, Penn LZ, Minden MD . Receptor- and mitochondrial-mediated apoptosis in acute leukemia: a translational view. Blood 2001; 98: 3541–3553.

  26. 26

    Asou H, Tashiro S, Hamamoto K, Otsuji A, Kita K, Kamada N . Establishment of a human acute myeloid leukemia cell line (Kasumi-1) with 8;21 chromosome translocation. Blood 1991; 77: 2031–2036.

  27. 27

    Matozaki S, Nakagawa T, Kawaguchi R, Aozaki R, Tsutsumi M, Murayama T et al. Establishment of a myeloid leukaemic cell line (SKNO-1) from a patient with t(8;21) who acquired monosomy 17 during disease progression. Br J Haematol 1995; 89: 805–811.

  28. 28

    Zhan Q, Jin S, Ng B, Plisket J, Shangary S, Rathi A et al. Caspase-3 mediated cleavage of BRCA1 during UV-induced apoptosis. Oncogene 2002; 21: 5335–5345.

  29. 29

    Fischer U, Janicke RU, Schulze-Osthoff K . Many cuts to ruin: a comprehensive update of caspase substrates. Cell Death Differ 2003; 10: 76–100.

  30. 30

    Boer J, Bonten-Surtel J, Grosveld G . Overexpression of the nucleoporin CAN/NUP214 induces growth arrest, nucleocytoplasmic transport defects, and apoptosis. Mol Cell Biol 1998; 18: 1236–1247.

  31. 31

    Cande C, Cecconi F, Dessen P, Kroemer G . Apoptosis-inducing factor (AIF): key to the conserved caspase-independent pathways of cell death? J Cell Sci 2002; 115: 4727–4734.

  32. 32

    Licht JD . AML1 and the AML1-ETO fusion protein in the pathogenesis of t(8;21) AML. Oncogene 2001; 20: 5660–5679.

  33. 33

    Sakahira H, Enari M, Nagata S . Cleavage of CAD inhibitor in CAD activation and DNA degradation during apoptosis. Nature 1998; 391: 96–99.

  34. 34

    Bloomfield CD, Lawrence D, Byrd JC, Carroll A, Pettenati MJ, Tantravahi R et al. Frequency of prolonged remission duration after high-dose cytarabine intensification in acute myeloid leukemia varies by cytogenetic subtype. Cancer Res 1998; 58: 4173–4179.

  35. 35

    Arama E, Agapite J, Steller H . Caspase activity and a specific cytochrome C are required for sperm differentiation in Drosophila. Dev Cell 2003; 4: 687–697.

  36. 36

    Fernando P, Brunette S, Megeney LA . Neural stem cell differentiation is dependent upon endogenous caspase 3 activity. FASEB J 2005; 19: 1671–1673.

  37. 37

    Miura M, Chen XD, Allen MR, Bi Y, Gronthos S, Seo BM et al. A crucial role of caspase-3 in osteogenic differentiation of bone marrow stromal stem cells. J Clin Invest 2004; 114: 1704–1713.

  38. 38

    Schwerk C, Schulze-Osthoff K . Non-apoptotic functions of caspases in cellular proliferation and differentiation. Biochem Pharmacol 2003; 66: 1453–1458.

  39. 39

    Lui JC, Kong SK . Erythropoietin activates caspase-3 and downregulates CAD during erythroid differentiation in TF-1 cells—a protection mechanism against DNA fragmentation. FEBS Lett 2006; 580: 1965–1970.

  40. 40

    Yan M, Burel SA, Peterson LF, Kanbe E, Iwasaki H, Boyapati A et al. Deletion of an AML1-ETO C-terminal NcoR/SMRT-interacting region strongly induces leukemia development. Proc Natl Acad Sci USA 2004; 101: 17186–17191.

  41. 41

    Yan M, Kanbe E, Peterson LF, Boyapati A, Miao Y, Wang Y et al. A previously unidentified alternatively spliced isoform of t(8;21) transcript promotes leukemogenesis. Nat Med 2006; 12: 945–949.

Download references


We appreciate Dr Lübbert M, Dr Tenen DG and Dr Hiebert SW for providing us U937-A/E 9/14/18, U937T cell line and some plasmids, respectively. This work was supported in part by National Key Program (973) for Basic Research of China (NO2002CB512805), National Natural Science Foundation (30500257, 30630034), Grants from Science and Technology Committee of Shanghai (05JC14032). Dr. GQ Chen is a Chang Jiang Scholar of Ministry of Education of China, and is supported by Shanghai Ling-Jun Talent Program.

Author information

Correspondence to G-Q Chen.

Rights and permissions

Reprints and Permissions

About this article


  • AML1-ETO
  • apoptosis
  • leukemia
  • caspase-3
  • proteolytic cleavage

Further reading