To the Editor:
Dierlamm et al1 reported in the June 1 issue ofBlood that an apoptosis inhibitor gene, API2 at 11q21, and a novel gene, MLT at 18q21, are involved in t(11;18)(q21;q21) associated with mucosa-associated lymphoid tissue (MALT) lymphoma.1 We recently reported the results of a breakpoint analysis of an MALT lymphoma with t(11;18)(q21;q21) by means of fluorescence in situ hybridization (FISH) analysis using YAC contig at 18q21, which identified a YAC clone, y789F3, encompassing the breakpoint region.2 Our further analysis of t(11;18)(q21;q21) with the YAC clone has identified a novel gene at 18q21, which was found to be derived from the same gene as the MLT gene.3 This gene was designated as MALT1 in accordance with the recommendation by the genome nomenclature committee.
We first would like to point out that there is some confusion regarding the name of the 11q21 gene. Dierlamm et al1 described the 11q21 gene as API2 in their report, but the sequence shown in their article represents that of API1 rather than API2 as assessed by the GenBank registry (GenBank Accession No. NM001165). The other three names already designated to the 11q21 gene, namely c-IAP2,4HIAP1,5 and MIHC,6 were correctly used in Dierlamm’s article and, again, the sequence of these genes is identical to that of API1. We have also confirmed the presence of API1(c-IAP2)-MALT1/MLT fusion transcripts by using reverse transcription-polymerase chain analysis (RT-PCR) for our series of MALT lymphomas with t(11;18) as noted in our report.3 To avoid any further confusion, we strongly suggest that API2 should be corrected to API1.
We next would like to point out a few but significant differences between Dierlamm et al’s results1 and ours3regarding the analysis of the MALT1/MLT gene. In the former, it is mentioned that the MLT transcript is approximately 3.0 kb and the predicted MLT protein consists of 729 amino acids (data not shown in the article).1 Our analysis, on the other hand, indicated that the MALT1 transcripts are 4.5 kb and 3.1 kb and the open reading frame of MALT1 predicted a protein of 813 amino acids.3 The Northern blot analysis with MALT lymphoma with t(11;18)(q21;q21) demonstrated additional aberrant signals in the range from 6.4 kb to 9.4 kb.3 Furthermore, the deletion of 11 amino acids has been noted in our MALT1 sequence (API2-MLT amino acids 625 to 635; GRTDEAVECTE).1,3 In our analysis, 3 out of 5 MALT1 cDNA clones did not contain the 11 amino acids, whereas the remaining 2 cDNA clones did (unpublished results, March 1999). We tentatively chose the cDNA sequence lacking the 11 amino acids for the MALT1 sequence,3 and consider it most likely that the difference between the two sequences is due to alternative splicing of the MALT1/MLT gene. Although the true size of MALT1/MLT amino acids remained to be determined, awareness of the above-mentioned differences should be helpful for readers who plan to perform molecular genetic analysis of the t(11;18) translocation. Further studies are warranted to clarify the essential role of API1-MALT1/MLT fusion protein in the molecular pathogenesis of MALT lymphomas.
Response
There was indeed confusion possible about the nomenclature of theAPI1/API2 genes and, therefore, it is useful to summarize the data. The cloning of API1/API2 was first reported by Rothe et al,1-1 who used the names c-IAP1 and c-IAP2.Liston et al1-2 independently reported the cloning of the same proteins named, respectively, hiap-1 and hiap-2.It should be noted that Hiap-1 corresponds to c-IAP2 andhiap-2 is identical to c-IAP1, which we think is the source of the confusion. The same year, Uren et al1-3 clonedMIHB and MIHC, which are identical to, respectively,c-IAP1 and c-IAP2. The Genome Database always listed the following entries: GDB :9848645 − API1 − alias apoptosis inhibitor 1 − alias MIHB − aliascIAP1 − alias hiap-2 and GDB :9848647 −AP12 − alias apoptosis inhibitor 2 − alias MIHC − alias cIAP2 − alias hiap-1. We thus propose that theAPI2 nomenclature, based on original data and the GDB gene entries, is followed; consequently the t(11;18) generates anAPI2/MLT fusion.
We have obtained from the GDB Gene Nomenclature Committee MLTas the official gene symbol for the gene on 18q21 involved in the t(11;18) found in marginal zone B-cell lymphoma of the MALT type. The t(11;18) breakpoint occurs in different exons of MLT1-4(Baens and Marynen, submitted) and we also showed that differential splicing results in the absence of the GRTDEAVECTE sequence in the MLT protein derived from 3 out of 11 cDNAs we sequenced (Baens and Marynen, submitted). Eight cDNAs contained this exon, as does the EST sequence for clone zq90d12.r1 (GenBank AA214173). Alternative splicing resulting in different protein isoforms is a frequent event and we have no functional data allowing to ‘tentatively choose the cDNA sequence lacking this exon.’ As we were the first to publish the API2/MLT fusion,1-4 theMLT symbol was accepted by the Gene Nomenclature Committee and since there is no reason to assign different gene symbols to different splice variants of the same gene, we propose that Suzuki et al avoid any confusion as happened for API1/API2 and consequently adoptMLT as the gene symbol for the gene involved in the t(11;18).
We performed Northern analyses with the MLT probe on total RNA and detected a major transcript of 3.0 kb. This transcript must correspond to the 3.1 transcript of Suzuki et al. These Northern blots also show a weak band at 4.5 kb. Whether this corresponds to a transcript identical to the longer one seen by Suzuki et al remains to be investigated. Similarly, it remains to be investigated whether the longer open reading frame reported by these authors results from alternative splicing, the use of a different promotor, or from experimental differences, and we agree that this could be relevant with regard to the biology of the MLT protein. In this regard we would like to point out that we presented evidence for the fact that the relevant fusion resulting from the t(11;18) is the one generating anAPI2-MLT mRNA.1-4 In fact, data presented by Akagi et al1-5 seem to support this claim, as in one of their cases the t(11;18) is accompanied by a deletion of 18q sequences, which most probably contains the 5′ part of the MLT gene.
