¹Chair of Genetics, University of Erlangen-Nürnberg, Staudtstr. 5, D-91058 Erlangen, Germany

²Department of Pediatrics, Hematology and Oncology, Children's Hospital, University of Gieen, Feulgenstr. 12, D-35392 Gieen, Germany

³University Hospital, Hematology and Oncology, University of Göttingen, Robert-Kochstr. 40, D-37075 Göttingen, Germany

⁴Department of Pediatrics, University of Erlangen-Nürnberg, Loschgestr. 15, D-91054 Erlangen, Germany

Correspondence to: Rolf Marschalek, Chair of Genetics, University of Erlangen-Nürnberg, Staudtstr. 5, D-91058 Erlangen, Germany

Some chromosomal translocations involved in the origin of leukemias and lymphomas are due to malfunctions of the recombinatorial machinery of immunoglobulin and T-cell receptor-genes. This mechanism has also been proposed for translocations t(4;11)(q21;q23), which are regularly associated with acute pro-B cell leukemias in early childhood. Here, reciprocal chromosomal breakpoints in primary biopsy material of fourteen t(4;11)-leukemia patients were analysed. In all cases, duplications, deletions and inversions of less than a few hundred nucleotides indicative of malfunctioning DNA repair mechanisms were observed. We concluded that these translocation events were initiated by several DNA strand breaks on both participating chromosomes and subsequent DNA repair by `error-prone-repair' mechanisms, but not by the action of recombinases of the immune system.

MLL gene; AF-4/FEL gene; leukemia; chromosomal translocation t(4; 11); recombination

Introduction

Certain chromosomal translocations are known to be the initial steps of the malignant transformation of hematopoietic cells leading to the development of myeloid and lymphoblastic leukemias and lymphomas (Rabbitts, 1994; Barr, 1998). For other translocations, this is a reasonable assumption still awaiting definitive proof. Among the latter are a number of translocations to the MLL gene (Ziemin van der Poel, 1991; Djabali et al., 1992; Gu et al., 1992a; Tkachuk et al., 1992) located on the long arm of human chromosome 11 (band q23). More than 30 different chromosomal translocations to the 11q23 region have so far been identified, all of which are associated with hematologic malignancies, and many of the translocation partner loci have been analysed at the molecular level (Bernard and Berger, 1995; Marschalek et al., 1997). In particular, translocations t(4;11) are regularly associated with high-risk infant acute pro-B lymphocytic leukemias and have been proposed to initiate the development of these leukemias (Ford et al., 1993; Bernard and Berger, 1995; Canaani et al., 1995; Marschalek et al., 1997). Thus, the MLL gene is highly promiscuous in its recombinatorial activity, and it is an important objective to identify the molecular mechanisms leading to these translocations.

The `recombinases' of the immune system gene rate certain chromosomal translocations associated with leukemias and lymphomas as accidents of the rearrangements of lymphocyte antigen receptor genes (immunoglobulin receptor genes; T-cell receptor genes). However, in those cases, one of the partner loci is an antigen receptor gene (Hartl and Lipp, 1987), and specific target sequences for the `recombinase' together with non-encoded P- and N-nucleotide insertions are usually found at the translocation breakpoints. The same is true for the SIL-TAL deletion involved in a frequent form of T-cell leukemias, where nonamer and heptamer sequences were identified at the chromosomal breakpoints (Breit et al., 1993). None of these characteristic signs were observed for t(4;11) translocations in leukemia-derived cell lines and a limited number of primary blasts (Marschalek et al., 1995, 1997; Reichel et al., 1998). From the limited evidence to date it was concluded, that the illegitimate recombination leading to these translocations was probably initiated by a severe DNA damage and subsequent defects in the DNA repair mechanism (Reichel et al., 1998).

To verify and extend these initial observations and to ascertain, that they were not peculiar properties of a small number of leukemia-derived cell lines but also present in primary leukemic blasts, chromosomal breakpoints were investigated here in the genomic DNA from an extended collection of primary biopsy material from patients with t(4;11) leukemia. The analysis was greatly facilitated by the recent improvement of our knowledge of the breakpoint cluster regions of the MLL and AF-4 genes (Gu et al., 1994; Marschalek et al., 1995; Nilson et al., 1997; Reichel et al., 1998). Based on this extended knowledge, it was possible to design a novel set of several dozen specific oligonucleotides and thus to amplify and sequence both reciprocal breakpoints for each patient. Unexpectedly, the analysis not only confirmed the previous report describing leukemia-derived cell lines (Reichel et al., 1998), but it provided further evidence allowing us to specify the type of DNA repair involved. Characteristic signs for one of the several known DNA repair mechanisms, the `error-prone-repair' (EPR) pathway, were consistently observed. A set of rules governing the t(4;11) translocation is proposed which constitute the `DNA damage repair model of chromosomal translocations'. This model may also apply to other chromosomal translocations, that are not produced by recombinases of the immune system (Rabbitts, 1994; Barr, 1998).

Results

Rapid mapping of one of the two translocation breakpoints in primary blasts using a novel approach

A novel approach allowing a rapid identification of one of the two translocation break points based on a combination of long-range PCR and subsequent detection of fluorescence-labelled PCR restriction fragments has recently been published (Repp et al., 1995; Materials and methods). Using this technique, the breakpoints on the derivative 11 (der11) chromosome were cloned for patients AJ000167, AJ000168, AJ000169, AJ000171, AJ000173, AJ000178 and AJ000180, and the breakpoints on the derivative 4 (der4) chromosome for patients AJ000166, AJ000174, AJ000175, AJ000177 and AJ000178 (Figure 1). However, until now this method was limited to provide information for only one of the two derivative chromosomes for each patient, because only a limited subset of PCR primers was available. Using both recently published and unpublished sequence information about the breakpoint cluster region of the AF-4 gene (Nilson et al., 1997; Reichel et al., 1998; E Gillert, unpublished data), it was now possible to select a suitable primer pair to amplify the corresponding reciprocal breakpoint for 10/14 patients (Figure 1).

Mapping of both translocation breakpoints for three additional patients by a conventional approach

For three additional patients an approximate map location of the translocation was obtained by RT - PCR analysis of break-region transcripts (F Griesinger, E Gillert, unpublished data). The MLL/AF-4 mRNA of patient PCB contained a fusion of MLL exon 11 to AF-4 exon 4 (nomenclature according Nilson et al., 1996, 1997; Marschalek et al., 1997). The AF-4/MLL mRNA of the same patient showed a fusion of AF-4 exon 3 to MLL exon 12. Thus, the chromosomal break points were located in intron 11 of the MLL gene and intron 3 of the AF-4 gene. The MLL/AF-4 mRNA of patient CB contained a fusion of MLL exon 11 to AF-4 exon 6. The AF-4/MLL mRNA of this patient showed a fusion of AF-4 exon 6 to MLL exon 12. Thus, the breakpoints were located in intron 11 of the MLL gene and intron 5 of the AF-4 gene. The MLL/AF-4 mRNA of patient HH contained a fusion of MLL exon 10 to AF-4 exon 6, whereas the AF-4/MLL mRNA showed a fusion of AF-4 exon 5 to MLL exon 12. Thus, the breakpoints were located in intron 10 of the MLL gene and intron 5 of the AF-4 gene. Appropriate primer pairs were used to amplify the genomic breakpoints on both derivative chromosomes for all three patients (Figure 1).

Characteristic duplications, deletions and inversions are present at all chromosomal breakpoints

Using the approach outlined above, both reciprocal break points and importantly the sequences of the corresponding regions from the intact second allele were amplified for each patient (Figure 2). Electrophoretic analysis of the amplified sequences (Figure 2) revealed striking deviations from the expected length for some of the amplified fragments, pointing to a deviation from a strictly balanced mechanism of translocation. All PCR amplimers were cloned and sequenced, and a quantitative length analysis (Table 1) further supported this conclusion. In all cases the sum of the sequences from the derivative chromosomes clearly deviated from the sum of the corresponding non-rearranged sequences.

The final structure of each pair of reciprocal breakpoints was compared with the corresponding non-rearranged sequences for each patient and the most plausible interpretation is given in Figure 3. In none of these cases a simple reciprocal and balanced cross-over mechanism could have produced the observed final state. Each of these translocation events was accompanied by deletions, duplications and/or inversions of germline DNA sequences. Moreover, filler DNA and mini-direct repeats were identified at the breakpoint junctions (Table 2).

Frequently, the translocation event resulted in the duplication of germline DNA sequences either from the MLL (5/14 cases) or from the AF-4 breakpoint cluster region (7/14 cases). The length of duplicated germline DNA in both derivative breakpoint fragments ranged between seven and 324 basepairs (bp). Deletions of germline DNA sequences of MLL (8/14 cases) and AF-4 (7/14 cases) were also observed. The range of deleted areas differed between 5 - 1014 bp. The inversion of germline DNA sequences was found only in the derivative 11 breakpoint fragment of patient AJ000180, where a 43 bp inversion of AF-4 germline sequences was identified (Genbank accession number AJ1000180). In this case the reciprocal derivative breakpoint has not yet been cloned. However, the inversion observed for this patient was similar to the ones already described for the leukemia-derived RS4;11 and LLM cell lines carrying a t(4;11) translocation (Reichel et al., 1998).

Filler-DNA sequences and mini-direct-repeats are present at several chromosomal fusion points

Short single-stranded DNA molecules consistently found in the nuclei of normal cells are used as bridging molecules in DNA repair processes and are commonly designated `filler-DNA' fragments (Plesner et al., 1987). Their use by the DNA repair machinery leads to a characteristic incorporation of short non- encoded nucleotide sequences. Such filler fragments were observed at the chromosomal junctions of either derivative 4 (AJ000168, AJ000174, AJ000177) or derivative 11 break points (AJ000167, AJ000177, patient HH), and thus in a significant fraction of our patient sample (Tables 2 and 3).

Mini-direct-repeats are short sequences of typically 3 - 5 nucleotides found at recombination points. They are present on both participating partner fragments, but only once on the rearranged chromosome, and thus their ancestor in the final product can no longer be determined unambiguously. Such mini-direct-repeats of 3 - 4 nucleotides were also found at the breakpoints of 5 of our patients (AJ000166, AJ000171, AJ000173, AJ000174, and AJ000178; Tables 2 and 3). The presence of these characteristic filler-DNA sequences and mini-direct-repeats indicates the participation of error-prone DNA repair mechanisms (EPR) in the generation of those translocations.

Observations excluding other potential mechanisms of recombination

Alternative mechanisms of recombination have been considered for the present group of patients. These included homologous interchromosomal Alu-Alu recombination (Chen et al., 1989; Onno et al., 1992), recombination by the machinery of the immune system (Haluska et al., 1990; Gu et al., 1992b), and recombination subsequent to the inhibition of topoisomerase II (Negrini et al., 1993; Aplan et al., 1996). None of our cases showed evidence for Alu-Alu recombination. All of our sequences were extensively scrutinized by suitable computer algorithms for consensus sequences characteristic of the action of immune recombinases and topoisomerase II, as well as first and second order variants of these consensus sequences. Some consensus sequences and variants thereof were observed in both breakpoint cluster regions of the AF-4 and MLL genes. However, importantly, their position did not coincide with any of the observed translocation breakpoints. Thus we conclude that these mechanisms are unlikely to have contributed to the translocations presented above.

Discussion

The data gathered here led to three major conclusions and a set of rules governing the origin of t(4;11) translocations. In addition, they invited a few final comments. The three main conclusions were:

1. At the fine-structure level all t(4;11) translocations in our sample were reciprocal but not balanced. This is in contrast to a published report, in which reciprocal t(4;11) translocations were described as `balanced' (Raimondi et al., 1995). The term `balanced' conventionally means: `without obvious loss of parental sequences at the breakpoints'. This terminology was based on cytogenetic observations, whereas the data reported here were gene rated by using molecular genetic methods capable of revealing small inversions, duplications, insertions and deletions below the threshold of detection of cytogenetic methods. Thus, the difference is only semantic and no true contradiction exists between these two separate sets of observations.

2. The characteristic structural features of the translocation breakpoints reported here are not a property of a select small subset of t(4;11) leukemias that can be expanded into stable lines. They are also not an artefact introduced by the process of establishing stable lines. The same characteristic features reported before for stable leukemia-derived cell-lines with t(4;11)-translocation (Reichel et al., 1998) were observed here for a dozen primary samples from leukemia patients, that had not been subjected to the selective pressures inherent in the establishment of stable lines. The new observations reported here allowed us to rule out those valid two alternatives. Thus, the breakpoint structures reported previously for stable lines and here for primary biopsy material probably reflect the true genomic state in primary blasts. Therefore, they allow us to draw legitimate conclusions about the mechanisms generating these translocations.

3. The translocations most likely were produced by the action of a particular type of DNA-repair mechanism, the so-called `error-prone repair mechanism' (EPR; Roth et al., 1985; Roth and Wilson, 1986). In general, mammalian cells can chose among several different DNA repair mechanisms, one of which is the EPR mechanism. However, this mechanism is not usually the first choice, and mammalian cells tend to use other mechanisms first. Only if those alternatives are not available they switch to the EPR mechanism. In the previous publication dealing with leukemia-derived cell lines (Reichel et al., 1998) it had been concluded, that a DNA repair mechanism was involved in the generation of t(4;11) translocations. On the basis of additional findings reported here (presence of `filler nucleotides' and `mini direct repeats' at the translocation breakpoint in a significant fraction of our cases, which are indicative of the action of the EPR-system) we now conclude, that the new data presented here are most consistent with the participation of the EPR mechanism.

From the combined data reported here and in the previous publication (Reichel et al., 1998) the following set of rules governing the origin of t(4;11) translocations was extracted to formulate the `DNA damage repair model of chromosomal translocations':

• Chromosomes 4 and 11 participating in the translocation most likely undergo multiple DNA breaks before the actual recombination occurs. A minimum of two single-strand or two double-strand breaks on one of the two participating chromosomes and one break on the other are required to initiate the illegitimate recombination event. We cannot rule out the alternative, that breaks initially occur on only one of the two chromosomes and broken ends from this chromosome invade the other.

• Numerous DNA breaks on the participating chromosomes generate multiple DNA ends. In conjunction with the DNA repair machinery these in turn produce the deletions, duplications and inversions of DNA sequences observed at the translocation break points.

• A symmetry-relation between the location of the breakpoints on chromosomes 4 and 11 became apparent upon close analysis of the data (Figure 1). The closer the break point was located to the centromeric side on chromosome 11, the more it approached the telomeric side on chromosome 4 and vice versa. The apparent existence of this symmetry relation suggests the requirement for a defined and reproducible spatial positioning of chromosomes 4 and 11 needed to produce the translocation.

• The presence of filler DNA and mini direct repeats at the breakpoints suggests, that the EPR mechanism rather than the double strand break repair (DSBR) mechanism participates in producing the translocation.

• The recombination event does not involve the action of recombinases of the immune system. Heptamer and nonamer signal sequences, obligatory target sites for the recombinases of the immune system, usually are not observed at or near the recombination sites. Similarly, homologous recombination mediated by middle repetitive sequence elements such as Alu elements usually is also not observed for t(4;11) translocations, although such sequences are present in the MLL gene.

An important open question addresses the nature of the DNA damage in the MLL and AF-4 genes that triggers the repair activity and thus the chromosomal translocation. No comprehensive list of all agents exists that might possibly contribute to these processes, but radiation, infectious agents, chemical mutagens and free radicals are probably among them (Greaves, 1997). Recent data indicate that inhibitors of topoisomerase II are capable of inducing DNA breaks at preferred positions in the breakpoint cluster regions of the MLL and AML-1 genes (Aplan et al., 1996; Stanulla et al., 1997a). This might provide one explanation for the tight clustering of translocation breakpoints within the MLL and AF-4 genes. Recent data further suggest the existence of a broad range of agents capable of activating signal pathways, that disrupt chromatin and lead to apoptosis if allowed to proceed to completion. When interrupted prematurely, these signal pathways may produce translocations and thus allow the cell to escape the apoptotic fate (Stanulla et al., 1997b). Conceivably, these agents cleave chromatin at preferred sites, which may provide another explanation for the observed tight clustering of the translocation breakpoints in both genes. A third explanantion proposes that only translocations occuring in these breakpoint cluster regions produce chimeric proteins that confer a selectable advantage. Translocations may also accur elsewhere with equal frequencies, but these would not lead to the outgrowth of leukemic clones (Marschalek et al., 1995, 1997).

Some of the most interesting questions remaining are, which DNA repair processes are involved and how they lead to chromosomal translocations. The defining properties of the EPR mechanism are: the use of mini-direct repeats, the use of single strand filler DNA molecules and the process of filling-in protruding single-strand ends and joining them by the so-called `non-homologous end joining' (NHEJ; Roth et al., 1995; Roth and Wilson, 1996) mechanism. All cases of our collection had properties pointing to the EPR mechanism. `Non-encoded' filler nucleotides were observed for 6/14 patients (AJ000167, AJ000168, AJ000174, AJ000177 and HH); mini direct repeats of 3 - 4 nucleotides for 5/14 patients (AJ000166, AJ000171, AJ000173, AJ000174 and AJ000178; Tables 2 and 3). The remaining breakpoints in our collection can all be explained by the NHEJ-mechanism.

Finally, the analysis of chromosomal breakpoints in t(4;11) cells was greatly facilitated by the improved knowledge of the breakpoint cluster regions of the MLL and AF-4 genes (Gu et al., 1994; Marschalek et al., 1995; Nilson et al., 1997; Reichel et al., 1998). Together with the new techniques described above (long range PCR, fluorescence detection of restriction polymorphic fragments; (Repp et al., 1995; Leis et al., 1998) this combination of improvements enabled us to amplify and analyse chromosomal breakpoints in a reasonably short time. The resulting genomic data represent an individual finger print for each patient and have recently become very useful as diagnostic tools and for the customized monitoring of `minimal residual disease' for each individual patient.

Materials and methods

Patient material

Biopsy material from patients AJ000166, AJ000167, AJ000168, AJ000169, AJ000171, AJ000173, AJ000174, AJ000175, AJ000177, AJ000178; AJ000180, CPB, CB and HH was collected at the time of diagnosis. Patients AJ000166 to AJ000180 were enrolled in one of the two German multicenter ALL-therapy studies BFM90 and CoALL90. DNA material from Patient PCB was obtained from the Department of Pediatrics, University of Erlangen-Nürnberg. DNA material from Patients CB and HH was obtained from the University Hospital of the University of Göttingen, Department of Hematology and Oncology. All patient data are summarized in Table 4.

Oligonucleotides

Breakpoint fragments were amplified using MLL and AF-4 specific oligonucleotides as listed in Table 5 (3, directed to the telomer; 5, directed to the centromer). Oligonucleotide positions are abitrarily starting either with the first nucleotide of intron 8 of the MLL (Z69751) or of intron 3 of the AF-4 gene (X83606)).

DNA purification

Genomic DNA from 2´10⁵ - 2´10⁷ cells was purified by ion exchange chromatography (Qiagen Genomic tips 100/G; Qiagen Ltd, Hilden, Germany) according to the manufacturer's recommendations. The DNA was dissolved in 100 l TE buffer (pH 8.0) and its concentration was determined by absorbance at 260 nm. In general, 30 g of high molecular weight DNA (50 - 100 kb) were obtained from 10⁷ viable cells.

Amplification and analysis of genomic DNA fragments spanning translocation breakpoints

Long range PCR experiments were performed as published (Leis et al., 1998). The resulting breakpoint amplimers were completely digested by Dde1 or Tru91 and end-labelled with fluoresceine-11-UTP. Restriction polymorphic DNA fragments were detected as described (Repp et al., 1995) using the GeneScan 672 software package (ABI/PE). All DNA fragments presumably spanning chromosomal breakpoints were sequenced. These data were compared with the known wildtype sequences of the breakpoint cluster regions of the MLL and AF-4 genes (Gu et al., 1994; Marschalek et al., 1995; Nilson et al., 1997; Reichel et al., 1998). Reciprocal break points were then amplified with appropriate reciprocal primer pairs. All sequence data were deposited in public databases.

Genbank accesssion numbers

U04737, X83604, AJ000166 to AJ000180, Y16596-Y15599, AJ235330 to AJ235380.

Acknowledgements

This study was supported by research grants SFB 466C4 and SFB 466C5 from the Deutsche Forschungsgemeinschaft (DFG) to RM/JG and GHF/RM; and research grant 96.047.1 from the Wilhelm Sander Stiftung to RM, JG and GHF. RM was supported by a Career Development Award from the Ria Freifrau von Fritsch Stiftung. Support by the J and F Marohn Foundation is greatfully acknowledged. E Gillert, T Leis, R Repp and M Reichel contributed equally to this work

Aplan PD, Chervinsky DS, Stanulla M and Burhans WC. (1996). Blood 87, 2649-2658. MEDLINE

Barr F. (1998). Nature Gen. 19, 121-124.

Bernard OA and Berger R. (1995). Genes Chromosomes Cancer 13, 75-85. MEDLINE

Breit TM, Mol EJ, Wolvers-Tettero IL, Ludwig WD, van Wering ER and van Dongen JJ. (1993). J. Exp. Med. 177, 965-977. MEDLINE

Canaani E, Nowell PC and Croce CM. (1995). Adv. Cancer Res. 66, 213-234. MEDLINE

Chen SJ, Chen Z, d'Auriol L, Le Coniat M, Grausz D and Berger R. (1989). Oncogene 2, 195-202.

Djabali M, Selleri L, Parry P, Bower M, Young BD and Evans GA. (1992). Nature Gen. 2, 113-118.

Ford AM, Ridge SA, Cabrera ME, Mahmoud H, Steel CM, Chan LC and Greaves M. (1993). Nature 363, 358-360. MEDLINE

Greaves MF. (1997). Lancet 349, 344-349. MEDLINE

Gu Y, Nakamura T, Alder H, Prasad R, Canaani O, Cimino G, Croce CM and Canaani E. (1992a). Cell 71, 701-708. MEDLINE

Gu Y, Cimino G, Alder H, Nakamura T, Prasad R, Canaani O, Moir DT, Jones C, Nowell PC, Croce CM and Canaani E. (1992b). Proc. Natl. Acad. Sci. USA 89, 10464-10468. MEDLINE

Gu Y, Alder H, Nakamura T, Schichman SA, Prasad R, Canaani O, Saito H, Croce CM and Canaani E. (1994). Cancer Res. 54, 2327-2330.

Haluska FG, Finger LR, Kagan J and Croce CM. (1990). Molecular Genetics in Cancer Diagnosis. Cossman J (ed.). Elsevier, New York, pp. 143-162.

Hartl P and Lipp M. (1987). Mol. Cell. Biol. 7, 2037-2045. MEDLINE

Leis T, Repp R, Borkhardt A, Metzler M, Schläger F, Harbott J and Lampert F. (1998). Leukemia 12, 758-763. MEDLINE

Marschalek R, Greil J, Löchner K, Nilson I, Siegler G, Zweckbronner I, Beck JD and Fey GH. (1995). Brit. J. Haematol. 90, 308-320.

Marschalek R, Nilson I, Löchner K, Greim R, Siegler G, Greil J, Beck JD and Fey GH. (1997). Leuk. Lymphoma 27, 417-428. MEDLINE

Negrini M, Felix CA, Martin C, Lange BJ, Nakamura T, Canaani E, Croce CM. (1993). Cancer Res. 53, 4489-4492. MEDLINE

Nilson I, Löchner K, Siegler G, Greil J, Beck JD, Fey GH and Marschalek R. (1996). Br. J. Haemotol. 93, 966-972.

Nilson I, Reichel M, Ennas MG, Greim R, Knorr C, Siegler G, Greil J, Fey GH and Marschalek R. (1997). Br. J. Haematol. 98, 157-169. MEDLINE

Onno M, Nakamura T, Hillova J and Hill M. (1992). Oncogene 12, 2519-2523.

Plessner P, Goodchild J, Kalckar HM and Zamecnik PC. (1987). Proc. Natl. Acad. Sci. USA 84, 1936-1939. MEDLINE

Rabbitts TH. (1994). Nature 372, 143-149. MEDLINE

Raimondi SC, Frestedt JL, Pui CH, Downing JR, Head DR, Kersey JH and Behm FG. (1995). Blood 86, 1881-1886. MEDLINE

Reichel M, Gillert E, Nilson I, Siegler G, Greil J, Fey GH and Marschalek R. (1998). Oncogene 17, 3035-3044. MEDLINE

Repp R, Borkhardt A, Haupt E, Kreuder J, Brettreich S, Hammermann J, Nishida K, Harbott J and Lampert F. (1995). Leukemia 9, 210-215. MEDLINE

Roth DB, Porter TN and Wilson JH. (1985). Mol. Cell. Biol. 5, 2599-2607. MEDLINE

Roth DB and Wilson JH. (1986). Proc. Natl. Acad. Sci. USA 6, 4295-4304.

Stanulla M, Wang J, Chervinsky DS and Aplan PD. (1997a). Leukemia 11, 490-496.

Stanulla M, Wang J, Chervinsky DS, Thandla S and Aplan PD. (1997b). Mol. Cell. Biol. 17, 4070-4079. MEDLINE

Tkachuk DC, Kohler S and Cleary ML. (1992). Cell 71, 691-700. MEDLINE

Ziemin van der Poel S, McCabe NR, Gill HJ, Espinosa III R, Patel Y, Harden A, Rubinelli P, Smith SD, LeBeau MM, Rowley JD and Diaz MO. (1991). Proc. Natl. Acad. Sci. USA 88, 10735-10739. MEDLINE

Figure 1 Location of chromosomal breakpoints in primary material of patients with t(4;11) leukemias. Top: MLL gene. Bottom: AF-4 gene. Narrow bars: introns; broad boxes: exons. bcr: break-point cluster region. A restriction map is shown for both breakpoint cluster regions. B, E, H: restriction cleavage sites for the enzymes BamHI, EcoRI and HindIII. Sizes (kb) refer to genomic BamHI fragments. Breakpoints from patients (circled numbers or letters) were analysed as described in the text and are connected by lines between both breakpoint cluster regions. Numbering of exons of the MLL gene according to Nilson et al. (1996) and Marschalek et al. (1997); of the AF-4 gene according to Nilson et al. (1997)

Figure 2 Breakpoint fragments generated by genomic PCR. PCR amplimers were generated with genomic DNA from each patient (patient codes in the lower left corner). For each patient PCR reactions were performed using allele-specific sets of oligonucleotides for the wt11, der11, der4 and wt4 chromosomes, respectively. Marker M: lambda phage DNA digested with PstI. Sizes in bp. Primer sets used for the amplification process are specified in the Materials and methods. Asterisk: presumably non-specific amplification products

Figure 3 Interpretation of fine structures of translocation breakpoints. A code identifying each patient is given in the upper left corner of each panel. Top and bottom: alleles prior to the translocation event; recombined alleles are shown in the middle. Arrows: double-strand or single-strand DNA breaks. Broad boxes: double-stranded DNA. Narrow boxes: single-stranded fragments or single-stranded overhangs. Fragments presumed to be created by the DNA damage are labeled in alphabetical order. Sizes of duplicated or deleted fragments are indicated. Nucleotides between boxes represent `filler' nucleotides

Table 1 Length of amplification products [bp] from 14 individual patients with translocation t(4;11)

Table 2 Comparison of chromosomal breakpoints of 14 individual patients with translocation t(4;11)

Table 3 Filler-DNA and mini-direct-repeats at chromosomal breakpoints

Table 4 Patient data

Table 5 Oligonucleotides used for the amplification of chromosomal breakpoints