Human Pax-5 C-terminal isoforms possess distinct transactivation properties and are differentially modulated in normal and malignant B cells

In this study, we set out to investigate the possible existence of isoforms in the C-terminal region of the human Pax-5 gene. We report the identification, characterization and expression patterns of these novel Pax-5 isoforms from healthy B lymphocytes as well as B cell lines and primary lymphoma samples. Furthermore, we demonstrate that these isoforms are translated into proteins that possess different TA properties and are generated in response to different cell activation pathways. NJ) density gradient centrifugation. CD19 positive cells were isolated from PBMCs with an autoMACS (Miltenyi Biotec, Auburn, CA) apparatus using a CD19 positive selection cocktail. Specific B cell sub-populations were isolated by flow cytometry (FACSCalibur, BD Biosciences, Mississauga, ON) cell sorting using CD10, CD19, CD20 and CD38-labeled antibodies. Patient lymph nodes biopsies were obtained with informed consent and subsequently selected from a library of B cell lymphoma diagnoses from the Pathology Department of the Dr. L. Hospital. Murine cells were obtained from the excision of spleen and thymus of Balb/C mice (Charles River, Montréal, QC). Cells were either left untreated or treated (as described in figure legends) with either phorbol 12-myristate 13-acetate (PMA, at 20 ng/ml; Sigma, Oakville, ON) / Ionomycin (I, at 1 µM; Calbiochem, LaJolla, CA) or with phytohemagglutinin (PHA, at 3 µ g/ml; Sigma). For time course experiments, Raji cells were synchronized through serum deprivation using culture medium containing 1% FBS for 16 hours.


Introduction
The development of B lymphocytes initiates from a common hematopoietic stem cell progenitor and proceeds according to a specific and coordinated process, which results in the production of mature B cells and antibody secreting plasma cells. This differentiation program is tightly regulated by the precise deployment of stage-and lineage-specific B lymphoid transcription factors that include EBF, E2A, and PAX-5 or the B cell-specific activator protein (BSAP) among others (1,2).
In this study, we set out to investigate the possible existence of isoforms in the Cterminal region of the human Pax-5 gene. We report the identification, characterization and expression patterns of these novel Pax-5 isoforms from healthy B lymphocytes as well as B cell lines and primary lymphoma samples. Furthermore, we demonstrate that these isoforms are translated into proteins that possess different TA properties and are generated in response to different cell activation pathways. experiments included BSAP 5' and BSAP 3' (binding regions exon 1 and 10 respectively) which amplify full-length human Pax-5 (1200 base pair fragment) and a nested pair of primers Exon 5 and Exon 10 AS (binding regions exon 5 and 10 respectively). The primers amplifying full-length murine Pax-5 correspond to the oligos identified mBSAP 5' and mBSAP 3' where the nested oligos are mExon 5 and mExon 10 AS. Each RT-PCR reaction was performed in at least two separate experiments. As controls, primer sets were used to amplify DNA fragments from either the G 6 PDH gene or the β-tubulin gene (Table 1).

Plasmids and cloning.
The plasmids CD19-luciferase (26) was kindly provided by M. Busslinger  instructions. Briefly, cells were seeded into 6 well plates (2 x 10 5 /well) 48 h pretransfection where they received a total of 2µg of plasmid DNA that was complexed with the provided transfection reagent. Cell extracts were then performed accordingly to the analysis intended. Reporter gene assays were conducted using the Dual-

Identification of novel Pax-5 isoforms in normal human B-cells.
Our studies of Pax-5 gene expression in normal and malignant B cells lead us to investigate possible alterations in the 3' coding region of the Pax-5 mRNA. A cDNA library was constructed from a pool of CD19+ cells isolated from the peripheral blood of 13 healthy donors. The library was screened by RT-PCR using Pax-5specific primers that span exons 5 through 10. On the basis of cDNA length, six distinct variants of Pax-5 cDNA clones were identified (Fig. 1A). Sequence analysis confirmed that one group of Pax-5 cDNA clones was identical in length to the published coding sequence for the full-length (FL) human Pax-5 mRNA (herein referred to a Pax-5 FL). Interestingly, the remaining five groups correspond to novel Pax-5 cDNAs with discrete gaps in the 3' coding region of the gene sequence. These gaps correspond precisely to deletions of exon 7, exon 8, exon 9, exons 7 and 8 or  Table 2).

Expression of alternatively spliced Pax-5 isoforms in human B cells.
To investigate the Pax-5 isoform expression patterns from individual donor samples, specific primer pairs were designed to enable the amplification of 1) the entire coding region of the Pax-5 cDNA and 2) a nested product spanning exon 5 through exon 10.
For this experiment, primary CD19+ B cells from 13 healthy donors were analyzed individually ( Fig. 2A). As a control, cloned FL Pax-5 cDNA was amplified in parallel and yielded an anticipated 602 bp product containing the entire coding sequences of exons 6 through 10. The nested RT-PCR of donor samples revealed distinct amplification products that were shorter in length compared to the control Pax-5 FL.
In addition, the isoform expression profiles were different in nearly each individual, and no correlation could be established between the observed Pax-5 isoform expression patterns and either donor age, gender or race (data not shown).
We then set out to investigate the Pax-5 isoform expression profiles in various B cell lines. RT-PCR analysis was used to examine the isoform expression profiles from immature (Nalm6 and REH /lanes 1 and 2 respectively) as well as mature (TA cell line and Raji /lanes 3 and 4 respectively) cells from the B cell lineage (Fig. 2B).
Multiple Pax-5 amplicons that were shorter than the amplified Pax-5 FL control were observed. Interestingly, the Pax-5 FL variant did not appear to be the predominant isoform in some of the cell lines tested.
Since several studies have shown that aberrant expression of Pax-5 is implicated in the development of a certain type of B cell lymphomas, namely lymphoplasmocytoid lymphoma, we investigated Pax-5 isoform expression in RNA extracts from lymph nodes of patients diagnosed with B cell lymphoma using nested RT-PCR. As depicted in Figure 2C These results suggest that differences in Pax-5 isoform expression exist between human and mouse hosts.

Pax-5 isoforms are translated in human B cells.
To confirm the anticipated putative protein sequences from the human Pax-5 isoforms, we performed Western blot analysis on total protein extracts from 293T cells transiently transfected with the cloned human Pax-5 isoforms (

Pax-5 isoforms have DNA-binding ability and transcriptional activation properties.
To further assess the role and function of the lower MW PAX-5 protein variants, we first investigated the DNA-binding capacity of these proteins through electrophoretic mobility shift assays (EMSA). Nuclear extracts from transiently transfected 293T cells with the various human PAX-5 isoforms were incubated with a Cy-5-labeled DNA probe corresponding to the PAX-5 consensus binding sequence. The formation of multiple shifted PAX-5/DNA complexes was observed (lanes 3 to 8) (Fig. 4A).
The specificity of these complexes was confirmed using a 100-fold molar excess of unlabelled PAX-5 consensus DNA probe (lanes 9 to 14). Conversely, the PAX-5/DNA complexes were preserved in the presence a 100-fold excess of a non-specific DNA probe corresponding to the AP-1 consensus motif (lanes 15 to 20). These results strongly suggest that each of these human PAX-5 isoforms possess DNAbinding properties similar to PAX-5 FL.
Given the fact that all tested PAX-5 isoforms seem to recognize and bind a specific DNA consensus motif, we analyzed their TA potential using a luciferase-based reporter assay. 293T cells were co-transfected with each isoform construct individually, along with DNA from pRL (Renilla luciferase construct) and CD19-Luc (a PAX-5 inducible reporter construct). The cells were then lysed, and the transcriptional activation activity was assessed (Fig. 4B). Interestingly, all PAX-5 isoforms induced reporter gene activity. Further comparison revealed that PAX-5 FL had the highest activity with a nearly 16-fold induction of activity compared to noninduced reporter controls. Both PAX-5∆8 and PAX-5∆9 proteins produced a 14-fold increase of transcriptional activity whereas PAX-5∆7/8/9, PAX-5∆7/8 and PAX-5∆7 produced progressively lower levels of induction of 7.5-, 5-and 2.5-fold, respectively.

Cell activation modulates Pax-5 isoform expression in human B cells.
PAX-5 is an important regulator of B cell development and its expression is developmentally regulated accordingly. As previous studies were not designed to examine the expression patterns of the C-terminus isoforms, we sought to investigate the regulating parameters governing the expression of the different Pax-5 isoforms.
First, we set out to determine if, individually, these isoforms are differentially Thus, it appears from these results that the TD of PAX-5 is composed of at least two separate sub-domains delimited by exons 8 and 9 rather than a single domain as previously described (16). The modular organization of the TD is reminiscent of the transactivation domains of rel, (37) and AhR (38,39) where the deployment of various modules allows for increased functionality in cellular processes. Therefore, it appears that the generation of a large number of isoforms Alternatively-spliced Pax-5 isoforms have also been described in murine tissue (22). reported an increase in Pax-5 isoform expression in multiple myeloma patients following treatment with anti-cancer agents. Clearly, these treatments produced changes in cell activation states, not unlike PHA or P/I, which influenced the isoform expression pattern. Given its central role of in B cell activation, it seems likely that the Pax-5 gene, itself, could be influenced by intracellular transduction events.
Taken together, these data strongly suggest that Pax-5 isoform deployment is a stoichiometrically dynamic process that represents real-time responses to physiological stimuli. profiles. Taken together, these observations suggest a carefully controlled mechanism for the production of these Pax-5 isoforms, rather than a random or aberrant process.
Interestingly, we note that upon the comparison of both transcription and translation of Pax-5 isoforms in a given cell type, we observed that the expressed products of one did not always reflect the pattern in the other. This could be explained by the fact that not all mRNA are tunnelled into protein synthesis. In some cases, mRNA transcripts are pooled and influence self-regulating mechanisms as aptamers and riboswitches (47-49). These results could also be a consequence of a bias amplification by PCR of shorter fragments. Further investigation of these observations will bring more insight into these issues. We believe that this further argues in favor of a tightly regulated mechanism for the deployment of Pax-5 isoforms.
Alternative splicing is being increasingly accepted as an important means for the expansion of the variability of gene products and associated functions. The deployment of protein isoforms is likely an evolutionary adaptation as a mean to increase the functionality of a given gene, which provides a rapid response to cellular requirements during development and other processes such as cellular proliferation, differentiation and apoptosis. In light of our findings, we can thus anticipate that the finding of multiple human Pax-5 isoforms will help elucidate BSAP's role and complexity in transcriptional regulation as well as in B cell processes and development. The further characterization of the various human Pax-5 isoforms will bring new insight to BSAP's function in transcriptional regulation, in cellular processes and development, and also in Pax-5-related oncogenesis.