Abnormal V(D)J Recombination of T Cell Receptor β Locus in SMAR1 Transgenic Mice*

Scaffold/matrix-associated region-1-binding protein (SMAR1) specifically interacts with the MARβ sequence, which is located 400-bp upstream of the murine TCRβ enhancer and is highly expressed during the DP stage of thymocyte development. To further analyze the functions of SMAR1, transgenic mice were generated that express SMAR1 in a tissue-independent manner. SMAR1-overexpressing mice exhibit severely altered frequency of the T cells expressing commonly used Vβs (Vβ5.1/5.2 and Vβ8.1/8.2/8.3). The rearrangements of Vβ5.1/5.2, Vβ8.1/8.2/8.3 loci are also reduced in SMAR1 transgenic mice. The T cells in SMAR1 transgenic mice exhibit a mild perturbation at the early DN stage. SMAR1 transgenic mice exhibit hypercellular lymph nodes and spleen accompanied with prominent architectural defects in these organs. These results indicate that SMAR1 plays an important role in the regulation of T cell development as well as V(D)J recombination besides maintaining the architecture of the lymphoid organs.

The immunoglobulin (Ig) and T cell receptor (TCR) 1 genes are assembled by a complex process of V(D)J recombination to generate antigen receptor diversity. In this process, the antigen binding domains of Ig and TCR chains that are encoded by germline variable (V), diversity (D), and joining (J) gene segments are rearranged to produce functional B or T cell receptors (1)(2)(3). For IgH and TCR␤ chains, V, D, and J gene segments are involved whereas for IgL and TCR␣ chains, V and J gene segments take part in the recombination event (4 -6). V(D)J recombination is known to be stage-specific within a given lymphoid lineage, and the antigen receptors are expressed from only one of the two alleles; the phenomenon called allelic exclusion (7,8). At the TCR␤ locus, the recombination occurs during DN (CD4 Ϫ CD8 Ϫ CD44 Ϫ CD25 ϩ ) stage of thymocyte development (9). V(D)J recombination is mediated by lymphoid and non-lymphoid-specific factors as well as cis regulatory elements (10). The lymphoid-specific factors of the recombination machinery include recombination activating genes (RAG) 1 and 2, and terminal deoxynucleotidyltransferase (TdT) (6, 10, and 11). Besides lymphoid-specific factors, nonlymphoid-restricted factors have also been implicated in the regulation of the recombination process. Among these, proteins that are involved in DNA double-strand break repair also play a critical role in the recombination events. These include Ku-80, XRCC4 protein, the catalytic subunit of DNA-dependent protein kinase (DNA-PK), and poly(ADP-ribose) polymerase (PARP) (12)(13)(14).
In addition to activating transcription of germline gene segments, cis regulatory elements also participate in recombination (8,15). Recently, it was reported that deletion of the germline promoter PD␤1, upstream of the D␤1 gene segment in the murine TCR␤ locus (16,17) not only resulted in reduction of D␤1 germline transcription but also reduced the rearrangement of D␤1 (15,18). Interestingly, the germline transcription and rearrangement of D␤2, J␤2, and V␤ gene segments remain unaffected by the deletion of the PD␤1 promoter. Similarly, the enhancer at the TCR␣ locus, E␣, is essential for rearrangement of all J␣ gene segments, whereas for proximal J␣ gene segments T early ␣ promoter is required (19). Transcriptional enhancers such as E␣, E␤, E, and E are also shown to be essential for recombination of their respective genomic loci as well as minilocus recombination substrates (20,21).
Besides promoters and enhancers, MARs (matrix-associated regions) are the cis regulatory elements that contribute in the regulation of TCR␤ gene rearrangement and transcription. MARs, though being distinct from promoters and enhancers, are often closely associated with these regulatory elements (22,23). MARs are commonly located at the boundaries of transcription units often flanking enhancer-like regulatory sequences (24), particularly those lying in the antigen receptor gene loci, such as the Ig light chain gene locus, IgH heavy chain gene locus (25), TCR␤ gene locus (26), TCR␣/␦ gene locus (27), and CD8␣ gene locus (28). MARs help the cell type-specific expression of genes by residing close to enhancers as well as by synergistically acting with them (29), thus maintaining some chromatin domains in condensed inactive structures and others in decondensed transcriptionally active structures. MARs flanking the enhancers in the immunoglobulin or T cell receptor genes are known to function in association with their specific MAR-binding proteins such as Cux (30) and SATB1 (31). Earlier, we reported a novel MAR-binding protein, SMAR1, interacting with MAR␤ (32), a MAR flanking 5Ј-end of the TCR␤ enhancer (E␤) (33). SMAR1 is the first known MARbinding protein that functions as a candidate tumor suppressor by interacting with and activating p53, thereby regulating the cell cycle (34).
To investigate the role of SMAR1 at the TCR␤ locus, we have generated SMAR1 transgenic mice. Here, we show that over-expression of SMAR1 severely reduces the frequency of cells expressing the commonly used V␤s (V␤5.1/5.2, V␤8.1/8.2, and V␤8.3) including V␤13. Accompanied with reduced transcription, there is a significant impairment in V(D)J gene rearrangement of these specific V␤s. In vitro results demonstrate that SMAR1 strongly reduces transcription mediated by the E␤ enhancer. A mild perturbation in the thymocyte development is observed at the early DN stage in SMAR1 transgenic mice. Taken together, the data indicate that SMAR1 regulates V(D)J recombination by modulating the function of the TCR␤ enhancer.

EXPERIMENTAL PROCEDURES
Generation of SMAR1 Transgenic Mice-For generating SMAR1 S transgenic mice, SMAR1-expressing vector, pBK-CMV-SMAR1 S was used where SMAR1 is under the control of the CMV promoter (Fig. 1B). The construct was first linearized with MluI and ApaLI, giving rise to a 3.8-kb fragment that contains the CMV promoter along with 1.8 kb of a full-length alternatively spliced form of SMAR1 (SMAR1 S ). The insert was purified on agarose gel, followed by DNA extraction with phenolchloroform and ethanol precipitation. The DNA pellet was resuspended at a final concentration of 4 g/ml. DNA was microinjected into the fertilized pronucleus of fertilized eggs derived from F-2 generation of (C57BL/6 X SJL) mice at 4 ng/l concentration using a Nikon micromanipulator. After injection, the eggs were transferred into the oviduct of a 0.5-day pseudopregnant foster CD-1 female mice.
Analysis of smar1 Transgene by Southern Hybridization-To identify the presence of SMAR1, tail DNA was prepared from mice following standard protocols. Briefly, a 2-mm tail biopsy was incubated in high salt digestion buffer containing 50 mM Tris-HCl, 100 mM EDTA, 0.5% SDS, and 0.5 mg/ml proteinase K for 16 h at 55°C. The lysates were processed for DNA isolation by phenol-chloroform extraction followed by ethanol precipitation. For Southern analysis, 15-20 g of DNA were subjected to restriction digestion with BamHI-HindIII or EcoRI-HindIII enzymes. The digested DNAs were fractionated on a 0.9% agarose gel, followed by denaturation and neutralization. DNA was transferred to Zeta-probe filters (Bio-Rad) under vacuum (Bio-Rad 785 model vacuum blotter). The filters were prehybridized for 4 h, followed by overnight hybridization with the CMV probe. Filters were washed twice for 10 min in 2ϫ SSC and 0.1% SDS at 65°C. Hybridization signals were detected by phosphorimaging and autoradiography.
RT-PCR assays were done as described previously (34) except that the cDNA was amplified for 27-30 cycles (94°C for 1 min, 65°C for 1 min, 72°C for 1 min). RT-PCR products were then separated on a 1.2% agarose gel and visualized by staining with ethidium bromide. The band intensities corresponding to the RT-PCR products were quantified using a phosphorimager (Bio-Rad) and normalized with respect to the ␤-actin product.
Probes and Primers-The DNA probe (CMV probe) used to screen the transgenic mice was obtained by restriction digestion of pBK-CMV-S-MAR1 S plasmid with BamHI-NdeI that generated a 0.6-kb fragment specific only for the CMV promoter. For TCR␤ gene rearrangement PCRs, the probes used were: TCR␤1 probe corresponding to the J␤1.6 fragment, obtained by the PCR amplification with the D␤1-J␤1 primers; TCR␤2 probe corresponding to J␤2.6 fragment resulting from the PCR amplification with the D␤2-J␤2 primers; and Thy1 probe generated by isolating the PCR fragment obtained from amplification with Thy1 primers. The PCR amplification of the indicated probes was followed by their gel purification (Qiagen gel extraction kit). All the probes were labeled using the Random Primed labeling kit (Roche Applied Science) and [␣- 32  PCR Amplification for Detection of TCR␤ Rearrangement-PCRs for TCR␤ rearrangements along with Thy1 as a control were performed using genomic DNA of thymus from control and transgenic mice. The primers and PCR conditions used were as described under "Experimental Procedures" (34). A brief initial denaturation at 95°C for 2 min was followed by 25 cycles with the following steps: 94°C for 1 min, annealing at 56°C for 2 min, extended at 68°C for 3 min. PCR products were separated on a 1.2% agarose gel, blotted onto Zeta-probe membrane (Bio-Rad) followed by Southern analysis. The PCR blots were first hybridized with either the TCR␤1 or TCR␤2 probe, then stripped and rehybridized with Thy1 probe. The blots were exposed as described (35,36); that is, for 30 min (DJ␤), 2 h (VDJ␤), and 1 h (Thy 1).
Purification of Double-Negative Thymocytes-Total thymocytes were washed with Dulbecco's modified Eagle's medium containing 10% fetal calf serum. The cells were spun at 1000 rpm for 5 min at 4°C, and the supernatant was discarded. The blocking of cells was performed with rat serum followed by incubation on ice for 20 min. The contaminating double-positive thymocytes were eliminated by magnetic depletion with BD TM IMAG anti-mouse CD4 or CD8 antibodies. The purified DN (lineage-negative) thymocytes were then stained with CD44-FITC (BD PharMingen) and CD25-PE. The cells were washed with FACS buffer followed by fixation with 1% p-formaldehyde. The cells were then analyzed by FACS as mentioned before.
Northern Blot Hybridization-Total cellular RNA was isolated from thymi of transgenic as well as control mice by a one-step acid guanidine isothiocyanate-phenol method using TRIzol reagent (Invitrogen), precipitated with ethanol, and quantitated by spectrophotometry. Twenty micrograms of RNA were electrophoresed on a 1% formaldehyde denaturing gel and blotted onto Zeta-probe membrane followed by Northern analysis. The blots were then probed with V␤4, V␤5.1, V␤8.1, V␤8.2, V␤8.3, and ␤-actin probes. The probes for all the V␤s were isolated from respective plasmids clones by EcoRI-BamHI digestions.
Luciferase Reporter Assays-Transient transfection assays were performed in T cell line (4980) using luciferase reporter constructs V␤13-E␤, V␤13-MAR␤, V␤13-MAR␤-E␤ that were driven by the V␤13 promoter (53). V␤13-E␤ contains the 550-bp core E␤ enhancer whereas V␤13-MAR␤ contains a 170-mer sequence of MAR␤ upstream of the reporter gene. In V␤13-MAR␤-E␤ construct, there is an additional 1-kb fragment within which MAR␤ and E␤ spans for about 170 and 550 bp, respectively (33). For V␤14 clones, a 1.2-kb BglI-BglII fragment that spans the V␤14 promoter and was isolated and blunt-ended using Klenow enzyme. The fragment was then cloned either in the vector alone or upstream of either E␤ enhancer or MAR␤-E␤. The cells were transfected by Lipofectamine, employing variable amounts of expression plasmids containing the luciferase reporter gene. Two micrograms of pRL-CMV (Renilla luciferase reporter DNA) were included in all transfections and used to normalize the transfection efficiency. Lucif-erase activity was assessed using the dual luciferase assay reporter kit (Promega), according to the manufacturer's instructions, and the luciferase activity was measured using Fluoroskan Ascent Luminometer (Labsystems). For all the luciferase assays, the data shown are the mean Ϯ S.D. of three independent experiments.

Analysis of the Transgene in Mice
Expressing SMAR1-Previously, we have shown that within a 100-kb region of the TCR␤ locus, there are 11 hypersensitive sites ( Fig. 1A) (26). Among these sites, HS1/MAR␤ is a major DNase I hypersensitive site that is strongly induced during the CD4 Ϫ CD8 Ϫ (DN) to CD4 ϩ CD8 ϩ (DP) stage of thymocyte differentiation (26,33). MAR␤, a matrix-associated region located 400-bp upstream of the TCR␤ enhancer (E␤), associates with two MAR-binding proteins SMAR1 and Cux ( Fig. 1A and Refs. 32 and 33). During transition from DN to DP stage of thymocyte development, there is a halt in TCR␤ gene rearrangement at the DP stage during which MAR␤ is strongly induced, and a parallel upregulation in SMAR1 expression was observed (26,32). To investigate the role of SMAR1 in V(D)J recombination and T cell development, we generated SMAR1 transgenic mice using a eukaryotic expression vector carrying SMAR1 cDNA as described under "Experimental Procedures" (Fig. 1B). SMAR1 transgene incorporation into the genome was confirmed by Southern blot analysis (Fig. 1C). Upon hybridization with CMV probe, one major band of 3.8 -4.0 kb was generated. Among a number of positive lines, two independent SMAR1 transgenic lines named T-12 and T-15 were used for breeding purposes (Fig. 1C, lanes 2 and 6). Because most results were similar in both transgenic lines, we have presented the data obtained from the T-12 line only. Although the thymus of the transgenic mice was slightly smaller than the littermate mice, both spleen and lymph nodes were comparatively much larger in the trans-genic mice. To verify the expression of SMAR1 at mRNA level, total RNA was isolated from thymus, spleen, and lymph nodes of both transgenic and littermate control mice and processed for RT-PCR analysis (Fig. 1D). Densitometric analysis demonstrated that compared with the littermate normal, SMAR1 transgenic mice exhibited a 3-5-fold increase in the smar1 s transcript in thymus, spleen, and lymph nodes ( Fig. 1D) indicating a higher expression of SMAR1 in transgenic mice. As a control for equal template concentrations used in these two samples, cDNA from both transgenic and normal littermate were processed for RT-PCR using Thy-1 expression. There was no apparent difference in the expression of Thy-1 between these mice (Fig. 1E). Other tissues from these mice were also checked for SMAR1 expression. Among the other organs checked, higher expression was observed both in kidney and testis (data not shown).
SMAR1 Transgenic Mice Exhibit Organomegaly-To observe the phenotypic effect of overexpression of SMAR1 in mice, we analyzed transgenic mice of various age groups. At birth and until three months of age, wild-type and transgenic mice were indistinguishable. The survival of transgenic mice appeared similar to that of their normal counterparts. After 6 -8 weeks, the transgenic mice displayed splenomegaly and marked lymphadenopathy but the average body weight remained the same as that of wild type. By 8 weeks, all the lymph nodes isolated from the transgenic mice were 3-5-fold larger than the control mice. In particular, the lymph nodes isolated from the inguinal region of SMAR1 transgenic mice (1 cm) were five times bigger than that of the control (2 mm) ( Fig. 2A). Compared with control, the size of the spleen was also increased by 1.5-2.5-fold in transgenic mice (Fig. 2, E and G). There was no appreciable size difference in other organs of transgenic and control mice indicating that overexpression of SMAR1 might be showing the 3.7-kb SMAR1 and 0.6-kb CMV promoter. The DNA fragment containing CMV-SMAR1 was injected into the pronucleus of fertilized mouse eggs to generate SMAR1 transgenic mice as described under "Experimental Procedures." A 0.6-kb BamHI-NdeI probe corresponding to the CMV promoter was used as a probe for screening transgenic mice. C, Southern blot analysis of tail DNA for identifying smar1 transgene upon digestion of genomic DNA with EcoRI plus HindIII and hybridization with the CMV probe. The blot shows that a variety of transgenic lines, including T-12 and T-15, were used for all experimental purposes. The CMV probe hybridized with a major band of about 3.8 kb, confirming the presence of the CMV promoter along with the smar1 transgene. D, RT-PCR analysis of smar1-specific mRNA (upper panels) in thymus, lymph nodes, and spleen from SMAR1 non-transgenic littermate (LM) or transgenic (Tg) mice. SMAR1-specific primers were used as discussed under "Experimental Procedures." The lower band corresponds to the RT-PCR product of ␤-actin, used as a loading control. E, RT-PCR analysis of Thy-1-specific mRNA in thymus, lymph nodes, and spleen from SMAR1 non-transgenic littermate (LM) or transgenic (Tg) mice.

SMAR1 Controls V(D)J Recombination and T Cell Development
deleterious for the normal development of lymph nodes and spleen.
The histological sections of lymph node depicted severe hyperplasia and infiltration of lymphocytes (Fig. 2, B-D). Moreover, in the lymph nodes, the normal distribution of the lymphoid cells within the germinal center was altered. The overall network within the medullar region in transgenic mice was irregular and more compact compared with normal mice, thus exhibiting a distinct loss in normal nodal architecture (Fig. 2, B-D). Interestingly, lymph node sections in transgenic mice exhibited greater numbers of lymphoid cells in the form of round patches. Compared with the littermate control, spleen in transgenic mice also showed high infiltration of lymphocytes, a characteristic of hyperplasia (Fig. 2, F and H). Thus, overexpression of SMAR1 results in architectural alteration in spleen as well as lymph nodes accompanied by their significant enlargement.
SMAR1 Overexpression Mildly Perturbs T Cell Development at Early DN Stage-Earlier, we have shown that SMAR1 expression is higher during the DP stage compared with either the DN or SP stage T cells (32). We reasoned that overexpression of SMAR1 in transgenic mice may further elucidate its role in the T cell development. To assess the development of T cells in the thymus of transgenic mice, thymocytes were stained for the surface expression of CD4/CD8 and analyzed by flow cytometry. Compared with the littermate control, in SMAR1 transgenic mice the percentage of cells in the DP (CD4 ϩ CD8 ϩ ) compartment remained unchanged (ϳ73%) (Fig. 3, A-C). There was no significant difference in the population of T cells expressing CD4 ϩ SP or CD8 ϩ SP cells in transgenic mice (Fig. 3, A-C). Interestingly, in transgenic mice both CD4 ϩ and CD8 ϩ single-positive T cells were reduced by 1.5-and 1.8-fold, respectively (Fig. 3, D and E). Thus, overexpression of SMAR1 perturbed maturation of either CD4 ϩ or CD8 ϩ T cells in thymus. Since V(D)J recombination of TCR␤ locus occurs in the DN subset of thymocyte population (CD4 Ϫ CD8 Ϫ CD44 Ϫ CD25 ϩ ) (12), the expression profile of CD44 and CD25 was studied. The DN cells were purified as described under "Experimental Procedures." CD44 versus CD25 FACS profiles of purified DN thymocytes revealed about a 2-fold increase in the population of CD44 ϩ CD25 ϩ cells in SMAR1 transgenic mice compared with the control mice (Fig. 3, F-H). Accumulation of such a significant population of CD44 ϩ CD25 ϩ cells in the thymus suggests that overexpression of SMAR1 moderately perturbs the T cell development at an early DN stage.
SMAR1 Affects T Cell Maturation-Because SMAR1 transgenic mice exhibit a defect in the T cell development at an early DN stage, it is possible that its overexpression may affect the

SMAR1 Controls V(D)J Recombination and T Cell Development
maturation of T cells. The expression of early T cell markers was checked by FACS analysis of T cells isolated from lymph nodes of either transgenic or littermate control mice. Interestingly, there is a significant increase in the frequency of T cells expressing CD4 ϩ CD62L ϩ (7.3-fold) (Fig.  4, (Fig. 4, D-F), and CD4 ϩ CD44 ϩ (2.5fold) markers (Fig. 4, G-I). An equivalent increase in the population of CD4 ϩ CD45RB Ϫ and CD4 ϩ CD44 Ϫ T cells as well suggesting that there is an increase in the number of CD4 ϩ T cells in transgenic mice compared with the littermate control. In other words, even though there is a selective increase in the population of T cells expressing the early markers suggestive of a mild perturbation in the T cell maturation, the increased CD4 ϩ T cell number is consistent with the increased hypercellularity observed in lymph node and spleen of transgenic mice.
Altered Frequency of T Cells Expressing Specific V␤s in Transgenic Mice-SMAR1 is highly expressed in DP thymocytes (32), and thus, it is possible that its binding to MAR␤ (induced during DN to DP transition) might control the V(D)J recombination at the TCR␤ locus in a stage-specific manner. Because overexpression of SMAR1 results in decreased population of mature T cells, T lymphocytes from thymus and lymph nodes were subjected to surface staining with a panel of V␤specific antibodies using mouse V␤ TCR screening panel (BD PharMingen). Analysis of V␤ profile in thymus from transgenic mice reveal that there was a substantial decrease in the frequencies of T cells expressing V␤5.1/5.2 (3.7-fold), V␤8.1/8.2 (4.4-fold), V␤8.3 (5.7-fold), and V␤13 (3.7-fold) compared with the control (Fig. 5, A-D). A similar decrease in the frequency of T cells expressing V␤5.1/5.2 (3.0-fold), V␤8.1/8.2 (3.6-fold), V␤8.3 (7.5-fold), and V␤13 (3.4-fold) was detected in lymph nodes from transgenic mice than that of littermate control (Fig.  5, E-H). In addition, there was a marginal decrease in the majority of the V␤-specific T cells that include V␤s 9, 10 b , 11, and 12 (data not shown). No significant decrease in the number of T cells expressing V␤2, V␤3, V␤14, and V␤17 a was observed in thymus as well as in lymph nodes of transgenic mice (data not shown).
These results suggest that upon overexpression of SMAR1, there is a severe perturbation in the frequency of T cells expressing the V␤s that are frequently used in TCR␤ gene recombination. The reduction in the number of commonly used V␤ T cells was more prominent in lymph nodes than that of thymus.

Impaired V(D)J␤ Rearrangement in SMAR1 Transgenic
Mice-Overexpression of SMAR1 leads to a decrease in the number of T cells expressing frequently used V␤s both in thymus as well as lymph nodes. To find out the effect of SMAR1 on TCR␤ gene rearrangement and to quantitatively test the differences in the recombination of V, D, and J gene segments, PCR assays were carried out. Genomic DNA from thymus of transgenic mice as well as the littermate controls were used as templates. The primers used for PCR amplification are depicted in Fig. 5A and are described under "Experimental Procedures." PCR products were gel-fractionated followed by Southern blot analysis. Hybridizations were performed either with D␤1-J␤1 (TCR␤1) or D␤2-J␤2 (TCR␤2) probes (Fig. 6A) to analyze the effect of SMAR1 on either V-D␤1-J␤1 or V-D␤2-J␤2 rearrangements, respectively. SMAR1 transgenic mice exhibited no change in the rearranged products corresponding to all possible D␤1 to J␤1.1-1.5 (Fig. 6B) as well as D␤2 to J␤2.1-J␤2.6 in thymus (Fig. 6C). Consistent with our FACS analysis of T cells from thymus, the V to DJ␤ rearrangement of frequently used V␤5.1, 5.2, 8.1, 8.2, and 8.3 genes are significantly reduced compared with the wild-type control mice. Depending on the specific V␤s mentioned above, there was about a 10 -20fold lower amount of recombined products in SMAR-Tg mice samples (Fig. 6, B and C). Interestingly, there was no significant difference in the rearrangement pattern of other V␤s such as V␤4 and V␤6 (Fig. 6, B and C, left two panels). Together, these findings demonstrated that SMAR1 overexpression selectively impaired V(D)J recombination of commonly used V␤s. As an internal control, Thy-1 primers were used under identical PCR conditions. To obtain a linear relationship between the amounts of input genomic DNA and the PCR product, the PCR assay was performed with serially diluted template DNA from thymus of both the mice (LM and Tg) with a limited number of PCR cycles (30) (Fig. 6D). This allowed semi-quantitative comparison of Thy-1 levels found in normal littermate and the SMAR1 transgenic. The level of Thy-1 as a marker of T cells was the same for both normal littermate and SMAR1 transgenic (Fig. 6D), indicating that there was a significant decrease in the levels of rearrangement between LM and Tg for the V␤ 5.1, V␤ 5.2, V␤ 8.1, V␤8.2, and V␤ 8.3 genes.

FIG. 4. FACS profile of T cells from lymph nodes expressing developmental markers.
A and B, lymph nodes cells from either control littermate (LM) or transgenic (Tg) mice were stained with either CD4 and CD62-L. C, the plot shows the number of CD4 and CD62-L cells in control littermate or transgenic mice respectively. D and E, lymph node cells from either LM or Tg were stained with CD4 CD45RB and were processed for FACS analysis. F, the result of the same was plotted as bar graph representing the number of CD4 CD45RB cells.

SMAR1 Acts as a Transcriptional
Repressor-The locus accessibility generally correlates with the transcription of the particular genes. Decreased V(D)J␤ rearrangement in SMAR1 transgenic mice could be attributed to poor accessibility of the TCR␤ locus to the recombination machinery. Because there was a significant decrease in the frequency as well as rearrangement of T cells expressing the commonly used V␤s, as shown by FACS and genomic PCR analysis, (Figs. 5 and 6), it is possible that the defect could be at the level of poor transcription at the locus. To check the levels of transcription for the specific V␤s, the expression levels of transcripts of V␤4, V␤5.1, V␤8.2, V␤8.3, and C␤1/C␤2 were verified by Northern blot analysis. Total RNA was isolated from thymocytes of littermate control and SMAR1-Tg mice, and the respective probes were used to detect the transcripts. Compared with the littermate control, there was no significant difference in the level of the transcripts for V␤4 (Fig. 7, left panel). As expected, there was a strong reduction in the transcript levels for V␤5.2 (3.2fold), V␤8.2 (5.3-fold), and V␤8.3 (3.1-fold) (Fig. 7). Thus, the constitutive expression of SMAR1 reduced the transcription of specific V␤s that are used frequently. On analyzing the expression of C␤1 and C␤2 transcripts in transgenic mice compared with the control mice (Fig. 7, right two panels) no apparent difference could be detected between the two. The reason for this unaltered expression of C␤1/C␤2 transcripts in SMAR1 transgenic mice could be because of compensatory effects introduced by those V␤s that are not affected in the SMAR1 transgenic mice. ␤-Actin PCR was done for normalization of the templates used in the Northern analysis. Thus, the data suggest that diminished rearrangement of the commonly used V␤s correlates with their reduced transcription as well as frequency of the T cells expressing the specific V␤s.
SMAR1 Represses Transcription through MAR␤-Earlier, we have shown that MAR␤, the binding site of SMAR1, functions as a silencer of E␤-mediated transcription (33). The factors contributing to this silencing function were unknown. Since SMAR1 binds to MAR␤, it is possible that binding of SMAR1 to MAR␤ might result in transcriptional repression of E␤ enhancer. To find out the specific role of SMAR1 in transcription, transient transfection assays were performed in T cell line (4980) using luciferase reporter constructs pGL2 vector, pGL2-MAR␤, pGL2-E␤, and pGL2-MAR␤-E␤ that are driven either by V␤13 promoter (33) (Fig. 8A) or by V␤14 promoter (Fig. 8B). Details of the V␤14 constructs are under "Experimental Procedures." MAR␤ alone contained only the 170-mer AT-rich sequence that binds to SMAR1 (32). No appreciable transcriptional activity was observed when the cells were transfected with either the control pGL2 vector or pGL2-MAR␤ either in the absence or in the presence of SMAR1 (Fig.  8C, lanes a-d). In the presence of increasing amounts of SMAR1, there was no change in the transcriptional activity. In the presence of E␤ enhancer alone there was about a 5.5-fold higher transcriptional activity (Fig. 8C, lane e). Upon increasing the amount of SMAR1, we could see a very small reduction in the transcriptional activity (Fig. 8C, lanes f and g) indicating that presence of SMAR1 did not change the transcriptional activity in the presence of enhancer alone. When the pGL2-MAR␤-E␤-containing plasmid driven by the V␤13 promoter was transfected, transcriptional activity was about 6.5-fold higher than that of the basal promoter (Fig. 8D, lanes a and b). Importantly, on cotransfecting with an increasing amount of SMAR1 reporter plasmid along with pGL2-MAR␤-E␤, a drastic reduction in the transcriptional activity was observed (Fig. 8D,  lanes c-f). In the presence of the highest SMAR1 concentration, the transcriptional activity came down almost near to basal levels of transcription. Thus the effect of SMAR1 on the V␤13 promoter correlates with the decreased transcription along with reduced frequency of V␤13-expressing T cells in the transgenic mice. Because the frequency of V␤14 increases in transgenic mice, the effect of SMAR1 was then analyzed on the V␤14 promoter in the presence of the E␤ enhancer. In the presence of the V␤14-E␤ enhancer alone, there was about 7-fold higher transcriptional activity (Fig. 8E, lane b). Upon increasing the amount of SMAR1, we could see a very minor reduction in the transcriptional activity (Fig. 8E, lanes c and d) compared with the V␤13 promoter. At the highest concentration of SMAR1, there was only a 1.2-fold of decrease in the transcription indicating that the presence of SMAR1 did not appreciably change the transcriptional activity in the presence of enhancer alone. As expected, there was a minor effect on the E␤ transcription when SMAR1 was cotransfected along with V␤14-MAR␤-E␤ (Fig. 8E, lanes e-g). Thus, the transcription silencing activity of specific V␤s corresponded with their decreased frequency in transgenic mice of respective T cells. These data suggest that the repressor activity of SMAR1 depends on the presence of MAR␤ upstream of the E␤ enhancer through which it controls the E␤ transcription, which in turn regulates the transcription of specific V␤ promoters. Because SMAR1 is overexpressed during the DP stage of thymocyte development (32), it is possible that during this stage when HS1/MAR␤ is induced (32), SMAR1 is directly recruited at the site, in turn inhibiting E␤ enhancer-mediated transcription. DISCUSSION The cis regulatory elements have been known to control the recombination and transcription of antigen receptor genes (8,37). T cell receptor ␤ (TCR␤) enhancer, E␤, is one such cis element that has been extensively studied with respect to its role in the regulation of V(D)J recombination and transcription by both deletion and transgenic mouse studies. The role of enhancer in the maintenance of local accessibility is critical and in its absence, recombination is halted resulting in blockage of lymphoid cell division (38 -40). The function of any enhancer is controlled by various positive and negative regulatory elements (acting either in cis or trans) that may either promote or inhibit its activity. MARs are one of the cis elements that closely associate with promoters and enhancers and are often found to flank antigen receptor genes (27). MARs either promote (41) or inhibit transcription (33) depending on the context. Earlier, we have reported identification of a new MAR (MAR␤), a 170-bp AT-rich sequence, present at 400-bp proximal to the E␤ enhancer (26,33). MAR␤ is one of the 11 DNase I hypersensitive sites that are induced during the DP stage of T cell development (26). Our report showed that MAR␤ provides a docking site for a novel MARBP, SMAR1 (32) along with two other well known MARBPs, Cux/CDP and SATB1. SMAR1 is abundantly expressed in the DP thymocytes, correlating with high accessibility of MAR␤ as well as the halting of V(D)J recombination at the DP stage, thereby suggesting a role for SMAR1 in the early T cell development (32,33). To delineate a specific role for SMAR1 in T cell development as well as in the regulation of V(D)J recombination, transgenic mice were generated that express SMAR1 in a tissue-independent manner.
Our data show for the first time that the MAR␤-binding protein, SMAR1, is directly associated with the recombination of antigen receptor genes in the T cells. The results from our analysis of the stage-specific thymocyte population indicate that overexpression of SMAR1 perturbs the T cell development during transition from DN to DP stage. Thus, SMAR1 overexpression results in generation of T cells that tend to remain in the early stage. Even though SMAR1 transgenic mice exhibit a selective increase in the population of T cells expressing early markers such as CD4/CD62-L, CD4/CD45RB, and CD4/CD44 but on analyzing the CD4 ϩ CD45RB Ϫ and CD4 ϩ CD44 Ϫ population of T cells, it appears that indeed there is an increase in the number of CD4 ϩ T cells. The phenotype of SMAR1 transgenic mice exhibit marked organomegaly of lymph nodes and spleen accompanied by follicular hyperplasia, suggestive of a hyper-responsive immune system. This increased cellularity of the lymphoid organs can be attributed to the increase in the CD4 ϩ T cell population observed in these organs.
MAR sequences are distributed every 10 kb of the human genome, and MAR-binding proteins upon interaction to these MAR sequences alter the chromatin structure. Earlier it was shown that the accessibility of the receptor loci alters V(D)J recombination (16). MARs and other cis elements have been implicated in maintaining locus accessibility (26, 33, and 42). Because SMAR1 is expressed more at the DP stage of T cell development, we analyzed the effect of the SMAR1 transgene on the antigen receptor gene loci in terms of its usage and recombination. The most critical feature exhibited by SMAR1 transgenic mice is the decreased frequency of T cells expressing commonly used V␤s, particularly, V␤5.1, 5.2, 8.1, 8.2, and 8.3 in thymus and lymph nodes. Interestingly, these V␤s are present in more than one-third of the entire population of peripheral T cells in normal mice. In addition, SMAR1 also significantly reduces the frequency of other V␤s including V␤9, V␤10 b , V␤11, V␤12, and V␤13. Importantly, the decreased frequency of T cells expressing the commonly used V␤s correlated with their diminished somatic gene rearrangement in the transgenic mice compared with the littermate control. There was no gross change in the recombination frequency of D␤ to J␤, indicating that the overexpression of SMAR1 affects recombination only from V to DJ but not from D to J locus. Because transcriptional promoters have been implicated as essential regulators of V(D)J recombination (43), it is possible that SMAR1 either directly or indirectly (by recruiting other factors) might control the enhancer function that in turn regulates the promoter activity of individual V␤s, thereby controlling recombination as well as transcription at the locus. The impairment in the rearrangement of the frequently used V␤s was reflected by their reduced transcription at mRNA level. This implies that the reduced transcription observed is caused by a lower number of T cells expressing the specific V␤s in transgenic mice. Thus, after the incorporation of the SMAR1 transgene, the frequency as well as transcription of the commonly utilized V␤s are affected, thereby resulting in their reduced rearrangement.
Earlier, it was demonstrated in reporter gene assays that MAR␤ silences the TCR␤ enhancer-mediated transcription (33) suggesting, therefore, involvement of associated trans factors that may interact with enhancer and repress transcription. SMAR1 exhibits significant sequence homology with the MAR binding domain of SATB1, Cut repeats of Cux and SATB1, and  (4980) were transfected with reporter constructs pGL2 or pGL2-E␤ or pGL2-MAR␤-E␤ in the absence or in the presence of increasing amounts of SMAR1. The lysates were made as described under "Experimental Procedures," and luciferase assays were performed. For MAR␤ or E␤ constructs, either 5 or 10 g of SMAR1-expressing plasmid DNA was used. D, for cotransfection with V␤13-MAR␤-E␤, 1-10 g of SMAR1 plasmid was used. Lanes a and b show relative luciferase activity in the absence of SMAR1 whereas lanes d-f exhibit transcription in the presence of increasing amounts of SMAR1. E, similarly, 5 ϫ 10 6 P4980 cells were cotransfected with either V␤14-E␤ or V␤14-MAR␤-E␤ alone or along with increasing amounts of SMAR1 (5-10 g). Lanes b and e show relative luciferase activity in the absence of SMAR1 whereas lanes c and d and f and g exhibit transcription in the presence of increasing amounts of SMAR1. The S.D. value was calculated from a minimum of three independent results done in triplicate. The mean Ϯ S.D. used, were generated from three independent experiments. *, p Ͻ 0.001. with the tetramerization domain of Bright (32). SATB1 is a more T cell-specific protein; Bright being specific for B cells whereas Cux exhibits ubiquitous expression. Cux and SATB1 have been known to function as strong transcriptional repressors (30,31) whereas Bright functions as a transcriptional activator (44). Both Cux and SATB1 proteins play significant role in various processes such as chromatin remodeling, tissuespecific gene regulation, and cell cycle progression, specifying cell fates during cell development and differentiation as well as tumor-specific metabolism (45)(46)(47)(48). Because SMAR1 shares homology with transcriptional repressors Cux and SATB1 and all of them interact with MAR␤, it is possible that SMAR1 might function in a coordinated manner through interaction with these proteins at MAR␤, thereby inducing the silencing function of MAR␤. To decipher the role of SMAR1 with respect to E␤-mediated transcription, transient transfection assays were performed on the DN T cell line using V␤13-or V␤14driven luciferase reporter constructs containing either E␤ or MAR␤ or MAR␤-E␤. Interestingly, these results show that SMAR1 upon interaction with MAR␤ reduces the E␤-mediated transcription of the promoter of that TCR V␤ whose frequency as well as rearrangement is decreased in transgenic mice. These results can be extrapolated to those frequently used V␤s that were significantly reduced in SMAR1 transgenic mice. On the other hand, there is no significant affect on the transcription of that V␤ whose frequency is increased in SMAR1 transgenic mice. Overall the data point to the critical role played by SMAR1 in regulating E␤-mediated transcription of specific V␤s.
Our results regarding the transcriptional repressor role played by SMAR1 through interaction with MAR␤ is supported by our recent observation wherein we show that SMAR1 physically associates with negative regulator Cux, especially through its arginine-serine-rich (RS) domain, and both the proteins synergistically function to repress E␤ mediated transcription (42). Overexpression of both SMAR1 and Cux results in chromatin modulation of MAR␤ thereby increasing its accessibility to DNase I both in T as well as non-T cells (42). Both the proteins form a ternary complex with MAR␤ and negatively regulate the transcription mediated by E␤ enhancer. Previously, it was shown that at the DP stage of T cell development, V(D)J recombination is halted and during this stage MAR␤ is induced. Moreover, at the DP stage, SMAR1 is abundantly expressed (32), and recently it was reported that expression of CDP/Cux p75, a spliced variant, was at a higher level in the thymus in CD4 ϩ /CD8 ϩ T cells (49). Interestingly, both SMAR1 and Cux were shown to interact with MAR␤ in DP T cells (42). Therefore, in a time-specific window of T cell development, the cellular machinery recruits negative regulators in the form of SMAR1 and Cux at MAR␤ that may either independently or in a concerted manner regulate T cell development through maintaining locus accessibility of the particular gene segments and in effect control the V(D)J recombination via E␤ regulation.
The mechanistic model for SMAR1-mediated perturbation of V(D)J recombination can be hypothesized in two independent ways. One, SMAR1 might function as a potent repressor controlling the E␤ enhancer (Fig. 7) function together with negative regulator Cux/CDP protein through recruitment of HDACs. In fact, we have recently found that SMAR1 recruits HDAC1 at the cyclin D1 promoter resulting in strong repression of the promoter. 2 It was found that in various breast cancer cell lines where SMAR1 is less expressed, cyclin D1 expression is induced. Thus, SMAR1 might function as a re-pressor for other genes as well and function through recruitment of HDAC1. The second model could be that SMAR1 works through the MARs present next to V gene segments as shown in Fig. 9. In support of this hypothesis, recently it was reported that MAR sequences are distributed more frequently in the V regions (48), particularly in the 5Ј-region of commonly used V␤s such as V␤8.1, V␤8.2, and V␤5.1. In the mouse genome, V␤5.1, V␤8.2, and V␤8.3 gene segments exist in close proximity within a region of 6.0 kb. Interestingly, these MARs are shown to be putative sites for Cux/CDP binding. Since SMAR1 and Cux are recruited together through their direct association with MAR sequences, it is possible that upon overexpression of SMAR1 in transgenic mice, these two MARs-binding proteins together alter chromatin at these MARs and thus perturb the recombination events at this loci. Thus, SMAR1 may either directly or indirectly in the form of a complex with enhancer cross-talk with the MAR sites close to the specific V␤ sequences mentioned. Such combinatorial effects may thus perturb the recombination of the specific V␤s that has frequent usage during recombination (Fig. 9).