Fig 1.
ERG gene and its transcript variants.
(A) Schematic representation of ERG gene and its transcript variants according to the current reference sequences available in NCBI Reference Sequence Database. Exons (represented by boxes) are numbered from 5’ to 3’ within the ERG gene (RefSeq ID NG_029732.1). Alternative exon numbering used by Owczarek et al. and by Bohne et al.[12, 14] is also shown. Transcript variants are annotated with RefSeq accession numbers and names for the encoded protein isoforms commonly used in the literature (ERG3 for NM_182918.3 and ERG2 for NM_004449.4); coding exons are in green, non-coding in grey. Recurrent types of intragenic deletions are displayed in red. Forward and reverse primers (blue triangles) used for amplification of full-length coding regions are positioned across all variants which they can theoretically amplify. (B) Schematic representation of three successfully amplified and cloned ERG transcript variants. (C) Probe sets from HG_U95Av2 Affymetrix gene expression array annotated to ERG gene exons. The scheme of probe sets mapping was adapted from UCSC Genome Browser (Kent WJ, Sugnet CW, Furey TS, Roskin KM, Pringle TH, Zahler AM, Haussler D. The human genome browser at UCSC. Genome Res. 2002 Jun;12(6):996–1006).
Fig 2.
Amplification, cloning and translation of ERG isoforms.
A) Analysis of PCR-amplified ERG isoforms by electrophoresis on agarose gel. PCR products with length corresponding to ERG3 coding region (1458bp) were present in all samples. Calculated length of amplified coding region of predicted aberrant ERG isoforms was 775bp in NALM6 and ALL samples 1,2,4 and 5 (bearing the deletion of ERG exons 7–13) and 880bp in ALL sample 3 (bearing the deletion of ERG exons 7–11). PCR products corresponding to predicted aberrant ERG isoforms were present in all ERGdel-positive samples and in none ERGdel-negative samples (B-other cases: ALL-6 and ALL-7, hyperdiploid cases: ALL-8 and ALL-9, ETV6/RUNX1-positive case—ALL-10). (B) Schematic representation of ERG transcript variants cloned from NALM6 and of predicted encoded proteins. Regions encoded by alternative frames are displayed in grey. (C) Analysis of ERGaber proteins synthesized from PCR-prepared DNA templates by in vitro transcription/translation (T/T) assay. Proteins were detected by western blot using following antibodies: Anti-ERG antibody EPR3863 (Ab-N), Erg-1/2/3 Antibody C-17 (Ab-int), Erg-1/2/3 Antibody C-20 (Ab-C). In vitro T/T reaction without any DNA template served as a negative control (Neg. Ctrl.). ERGaberN was detected by Ab-N only and ERGaberC by Ab-C only. Ab-int showed only unspecific binding to proteins present in reticulocyte lysate used for T/T assay.
Fig 3.
Analysis of subcellular localization of ERG isoforms by western blot and confocal microscopy.
(A) Protein lysates from HeLa and HEK293T cells (10μg and 2μg, respectively) transiently transfected by ERG3 (3), ERG3var (v) and ERGaber (a) in pcDNA3.1 vector or by empty vector (-) were analyzed by western blot to determine subcellular localization of individual ERG isoforms (left panel). ERG3 and ERG3var were detected dominantly in the nuclear fraction of protein lysate (N) in both cell lines, while ERGaberN was found in both cytoplasmic (C) and nuclear fractions in HeLa cells and dominantly in cytoplasmic fraction in HEK293T cells. ERGaberC was not detected in any cell line using these protein loads. Using higher load of protein lysate (30μg), sensitive visualization kit and longer exposition to X-ray films ERGaberC was detected in nuclear fraction of HEK293T but not HeLa cells (right panel, black arrow points to corresponding protein band). TBP and GAPDH proteins were used to control protein load and separation of cellular fractions. (B) HEK293T cells were transiently transfected by ERG3, ERG3var and ERGaber isoforms in pcDNA3.1 vector or by empty vector. Forty-eight hours after transfection the presence and the subcellular localization of ERG isoforms was analyzed by confocal microscopy using Ab-N antibody. Nuclei were stained by DAPI. The scale bar represents 10μm.
Fig 4.
ERG protein expression in NALM6 cells and in patient samples.
The presence of ERGaberN and ERGaberC proteins in nuclear (N) and cytoplasmic (C) fractions of protein lysates from NALM6 cell line (panel A), ERG-del-positive (n = 3) and ERGdel-negative (two B-other– ALL-7, ALL-12; one ETV6/RUNX1-positive—ALL-10) ALL cases (panel B, C and D) was analyzed by western blot. ERGaberN and ERGaberC proteins synthesized by T/T assay served as positive controls (corresponding bands are highlighted by white arrows). Any protein bands possibly corresponding to ERGaberN or ERGaberC were detected neither in nuclear nor in cytoplasmic protein fractions of NALM6 nor in ALL samples. Two bands were detected in cytoplasmic fraction of NALM6 cells (within 15-25kDA molecular weight region) using Ab-int antibody which, however, lacks reactivity against ERGaberN and ERGaberC. TBP and GAPDH proteins were used to control protein load and separation of cellular fractions. Protein loads: 10μg of nuclear and cytoplasmic proteins of NALM6; 10–20μg of nuclear and 25–43μg of cytoplasmic proteins of ALL cases. Panels B, C and D: individual lanes were cut from the scans of X-ray films and grouped. Full scans of X-ray films are available in Fig C in S1 File. These contain additional lanes which were excluded from the analysis because of protein degradation or subclonality of ERGdel.
Fig 5.
Expression of probe sets annotated to ERG coding region in ALL subgroups.
Normalized expression levels (Y axis) of 4 probe sets annotated to ERG exons in 11 ALL subgroups (X axis): ERGdel enriched B-other ALL (NOVEL, n = 14), B-other ALL with normal karyotype (NORMAL, n = 11), high hyperdiploid ALL (>50, n = 63), ALL with 47–50 chromosomes (47–50, n = 21), ALL with pseudodiploid karyotype (PSEUDO, n = 25), ALL with <47 chromosomes (HYPODIP, n = 9), ETV6-RUNX1-positive ALL (ETV6-RUNX1, n = 79), TCF3-PBX1-positive ALL (TCF3-PBX1, n = 27), MLL-rearranged ALL (MLL, n = 20), BCR-ABL1-positive ALL (BCR-ABL1, n = 15) an T-ALL (n = 43). Boxes extend from the 25th to 75th percentiles, the lines in the middle represent medians, and the whiskers extend to minimal and maximal values. ALL subgroups labeled with asterisks have significantly different expression compared to ERGdel associated B-other ALL subgroup: * −0.01 < p ≤ 0.05; ** −0.001 < p ≤ 0.01; *** −p ≤ 0.0001 (Mann-Whitney U test).
Fig 6.
Expression of physiological ERG isoforms in ALL subgroups.
Normalized expression levels (Y axis) of physiological ERG isoforms with (ERG+10) and without (ERG-10) exon 10 and their sum (ERGtotal) in 21 ERGdel-negative B-other ALL cases (ERGdel-), 8 ERGdel-positive B-other ALL cases (ERGdel+) and 13 B-other ALL cases with ERGdel at sublocnal level (ERGdel+sub). Horizontal lines represent means (ERG+10: 34.4±21.4, 20.6±13.0 and 41.0±17.6 for ERGdel-, ERGdel+ and ERGdel+sub, respectively; ERG-10: 34.3±22.1, 12.6±4.9 and 23.3±11.2 for ERGdel-, ERGdel+ and ERGdel+sub, respectively; ERGtotal: 68.7±42.5, 33.2±16.7 and 64.3±25.5 for ERGdel-, ERGdel+ and ERGdel+sub, respectively). Mann-Whitney U test was used to compare two groups.