The Primate Lentiviral Receptor Bonzo/STRL33 Is Coordinately Regulated with CCR5 and Its Expression Pattern Is Conserved Between Human and Mouse

Generation of mAb against Bonzo

An mAb against Bonzo/STRL33 was raised at R&D Systems (Minneapolis, MN) by immunizing BALB/c mice with a syngeneic mouse myeloma (NSO) transfected with full-length human Bonzo/STRL-33 with a polyhistidine fused to the C-terminal end of the receptor sequence. The polyhis serves as an epitope tag that can be used to identify (via intracellular staining in FACS or via Western blot) transfected clones that appear to be expressing high levels of the gene of interest. An immunization protocol (24) for soluble protein was adapted for use with whole cells as the immunogen. The priming immunization was performed by mixing the cell suspension in PBS with an equal volume of emulsified MPL/TDM adjuvant (Ribi); subsequent boosts used cells in PBS alone. Following immunization, lymph node cells were used for polyethylene glycol-mediated fusion following conventional protocols. After 7 days of culture, supernatants were screened for Abs that could bind to paraformaldehyde-fixed NSO/STRL-33/polyHis cells used for immunization. Cultures that were positive in this primary screen were then tested for binding to NSO cells that had been transfected with an irrelevant gene (GDF-9) also expressed as a polyhis construct. One clone was chosen and subcloned based on strong binding of its supernatant to the unfixed relevant transfectants. This hybridoma secretes an IgG2b κ mAb that was purified and used in subsequent experiments. This Ab is designated MAB699.

Preparation of human PBMC and resting T cells

PBMC were separated from buffy coats of healthy donors (New York Blood Bank, New York, NY) through Ficoll-Hypaque (Pharmacia, Uppsala, Sweden). Purification of T cells was performed as previously described (25). Briefly, monocytes were first removed from PBMC by plastic adherence for 2 h at 37°C. Nonadherent cells were incubated with anti-CD4 or anti-CD8 conjugated with Dynabeads (Dynal, Oslo, Norway) at a 1:4 target/bead ratio. The bead-bound cells were recovered using a magnet (Dynal) washed at least four times to remove unbound cells. The CD4⁺ or CD8⁺ cells were detached from the beads using Detachabead according to the manufacturer’s instructions (Dynal). These cells were then incubated with anti-HLA-DR Ab followed by Dynabeads conjugated with goat anti-mouse IgG for magnetic removal of preactivated T cells and contaminating dendritic cells or macrophages.

Media, reagents, and T cell cultures

The culture medium used in all experiments was RPMI (Life Technologies, Grand Island, NY) supplemented with 10% FCS (HyClone, Logan, UT), penicillin (50 U/ml; Life Technologies), streptomycin (50 μg/ml), sodium pyruvate (1 mM; Life Technologies), and glutamine (2 mM; Life Technologies). T cell lines were prepared by activation of purified resting T cells with allogeneic PBMC and were treated with 50 μg/ml mitomycin C (Sigma) for 30 min at 37°C and 5 μg/ml PHA (Sigma). Cells were split 3 days postactivation, expanded, and maintained in culture medium supplemented with 200 U/ml recombinant IL-2 (Chiron). T cell lines were maintained by restimulating the cells every 2 wk with PHA and mitomycin C-treated allogeneic PBMCs. Culture of T cells with cytokines has been previously described (26, 27). Cytokines IFN-γ, IL-4, IL-7, IL-12, and IL-15 and chemokines RANTES, macrophage inflammatory protein-1α (MIP-1α), MIP-1β, and stromal cell-derived factor 1α (SDF-1α) were all obtained from R&D Systems.

Ab staining and FACS analyses

Cells were incubated with the relevant Ab on ice for 30 min in PBS buffer with 2% FCS and 0.1% sodium azide. For staining of Bonzo, cells were incubated with anti-Bonzo mAb at 3 μg/ml; after two washes, cells were incubated with goat anti-mouse IgG conjugated to PE or tricolor (TC) (Caltag, South San Francisco, CA). The cells were washed twice again and blocked with excess mouse IgG (100 μg/ml) followed by staining with directly conjugated Abs against the relevant cell surface molecules. The Abs used for staining were PE, FITC, TC, or peridin chlorophyll protein conjugates of anti-human CCR5 and CXCR4; anti-mouse CD3, CD4, CD8, CD44, CD45RB, and γδTCR (PharMingen, San Diego, CA); anti-human CD4 and CD45RO, and secondary Abs goat anti-mouse PE or FITC (Caltag); anti-human CD3, CD8, CD14, CD16, CD19, CD45RA, HLA-DR, and γδTCR (all from Becton Dickinson, Palo Alto, CA); or anti-human CCR6 (R&D Systems). Staining was analyzed on a FACScan using CellQuest software (Becton Dickinson). Live cells were gated based on forward and side scatter. Intracellular staining was performed using Cytofic/Cytoperm solution according to the manufacturer’s protocol (PharMingen). To perform FACS analyses on fresh human thymocytes, thymi were obtained from 7- to 9-mo-old pediatric heart surgery cases. Thymic tissue was disrupted by mincing and forcing through stainless steel mesh. Thymocytes were incubated twice at 37°C for 30 min in complete medium to remove adherent cells. Single-cell suspensions were then placed on ice and stained for FACS analysis as described above.

Gene targeting in embryonic stem cells and generation of mice

A 129/Sv mouse genomic DNA library was screened with human Bonzo cDNA and a 16-kb DNA fragment was isolated. Sequence analysis identified a 1.2-kb intronless open reading frame (ORF) homologous to that of the human Bonzo gene. To generate an EGFP (Clontech) knockin targeting vector, three DNA fragments were sequentially inserted into a pBS-KS⁺ plasmid: 1) a GFP expressing cassette followed by SV40 poly(A)n sequence, 2) a neomycin resistance cassette (neo^R) flanked with loxP sites (28), and 3) a HSV thymidine kinase cassette (HSV-TK). Subsequently, an 8-kb ApaI-NotI genomic fragment downstream of the Bonzo ORF was inserted between the neo^R and HSV-TK cassettes, whereas a 1.6-kb genomic fragment upstream of the Bonzo ORF and containing the 5′-untranslated region was inserted 5′ of the GFP ORF. The resulting targeting vector was linearized with ClaI and electroporated into 129/Sv-derived E14 embryonic stem (ES) cells. G418-resistant ES cell clones were then screened for homologous recombination. To eliminate possible interference from the neo gene, correctly targeted clones were electroporated with a Cre recombinase-expressing vector (pCMV-Cre) to delete the neo^R cassette, which was confirmed by sensitivity to G418 and by PCR or Southern blotting for the absence of the neo^R-coding region. The resulting clones were microinjected into C57BL/6 blastocysts. Chimeric mice were mated with wild-type C57BL/6 mice to produce heterozygous progeny. Six- to 8-wk-old littermates from the mating of heterozygous mice were then analyzed.

RT-PCR analysis

Total RNA was extracted from mouse lymph nodes and spleen using Trizol reagent (Life Technologies). The RNA was further treated with RNase-free DNase I (Roche, Indianapolis, IN). RNA (1 μg) was used in an Access RT-PCR system (Promega, Madison, WI) with 50 pmol of sense (GCT TGC TCA TTT GGG TG) and antisense (CGC CGC GTC GAC CTT CTC TAA GTG TGG CAA GGC) Bonzo primers. RT was performed for 45 min at 48°C, and cDNA amplification was conducted for 40 cycles at 94°C for 30 s, 58°C for 1 min, and 68°C for 2 min. To exclude possible contamination with genomic DNA, control reactions in which reverse transcriptase was omitted were performed in parallel.

Fluorescence microscopy

Mouse organs were fixed in 4% formaldehyde for at least 12 h and incubated in 15% sucrose in PBS for 1 h. The samples were embedded in HistoPrep (Fisher, Fairlawn, NJ) and later frozen with dry ice. Samples were cut into 15-μm sections with a cryostat. Fluorescence microscopy was used to visualize GFP expression.

Bonzo expression pattern on PBMC

An mAb reacting against human Bonzo was obtained as described in Materials and Methods. This Ab binds specifically to HOS cells that stably express the Bonzo cDNA (Fig. 1⇓a) (20). Similar results were obtained using Bonzo-transduced NIH-3T3 cells or a mouse thymoma line (data not shown). In addition, 293T cells were transfected with a plasmid encoding Bonzo-GFP fusion and stained with anti-Bonzo Ab. Only Bonzo-GFP-expressing cells were positively stained with anti-Bonzo mAb, and this correlated with GFP expression. To exclude the cross-reactivity of the Ab with other chemokine receptors, we also stained HOS cell lines stably expressing CCR1, CCR3, CCR4, CCR7, CCR8, APJ, and V28. None of these cell lines was stained with anti-Bonzo mAb (data not shown). We then determined the expression pattern of Bonzo on PBMC of normal donors using multicolor FACS analysis. Expression of Bonzo was found to segregate primarily to CD3⁺ T cells (Fig. 2⇓a). Both CD4⁺ and CD8⁺ T cell subsets expressed Bonzo (Fig. 2⇓, b and c), although the percentage of Bonzo⁺ cells was always higher within the CD8⁺ subset in samples obtained from different donors (data not shown). T cells can also be subdivided into memory and naive subsets based on expression of CD45 isoforms, CD45RO⁺RA⁻ and CD45RO⁻RA⁺, respectively (29). Bonzo was exclusively expressed on CD45RO⁺ T cells (Fig. 2⇓d). About 30–40% of γδ T cells also expressed Bonzo (Fig. 2⇓f). The few CD3-negative cells that expressed Bonzo were CD16⁺ NK cells (Fig. 2⇓e), and no expression was detected on B cells, monocytes, or dendritic cells (Fig. 2⇓, g, h, and i, respectively).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Bonzo Ab specifically stains transfected cells. a, HOS cells stably expressing Bonzo were generated by retroviral transduction of the Bonzo or CCR5 cDNA using the pBabe vector (20 ). Cells were stained with the anti-Bonzo mAb followed by a PE-conjugated goat-anti-mouse Ab. b, 293T cells were transfected with a pcDNA3 plasmid encoding either GFP or Bonzo-GFP fusion protein. Cells were then stained, 48 h post-transfection, with Bonzo Ab as described above.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Expression of Bonzo on PBMC. Freshly isolated PBMC were stained with anti-Bonzo mAb followed by the PE-conjugated secondary Ab and were subsequently stained with PE-conjugated Abs specific for other cell surface molecules. Three donors gave similar results, and data from one donor are shown. Numbers in quadrants are percent positives. Expression of Bonzo on a) T cells (CD3⁺); b and c) CD4⁺ and CD8⁺ T cell subsets; d) memory (CD4RO⁺) and naive (CD45RO⁻) T cells (gated for CD3⁺ cells); e) NK cells (CD16⁺); f) γδ T cells (gated on CD3⁺ cells); g) B cells (CD19⁺); h, monocyte expression (CD14⁺); and i) dendritic cells. The dendritic cells were identified as follows: PBMC were first stained with anti-Bonzo Ab as described above and then with FITC-conjugated anti-CD3, -CD14, and -CD19 and TC-conjugated anti-HLA-DR Abs. A gate was set to exclude FITC-positive cells (T cells, B cells, and monocytes). The remaining HLA-DR⁺ cells identify DC.

It has been reported that expression of some chemokine receptors, such as the major HIV coreceptor CCR5, is also restricted to the memory subset of T cells (30). As shown in Fig. 3⇓, the expression profile of CCR5 on resting T cells very closely parallels that of Bonzo. Indeed, costaining of PBMC with Abs specific for Bonzo and CCR5 showed that most Bonzo-positive cells were also CCR5⁺ (Fig. 4⇓). PBMC from a healthy donor with a mutation in the CCR5 gene that prevents its cell surface expression (31) exhibited normal expression of Bonzo (Fig. 3⇓b), demonstrating that absence of CCR5 from the cell surface does not negatively influence Bonzo expression. Most Bonzo-positive resting T cells also expressed CCR6 (Fig. 4⇓), whereas expression of Bonzo and CXCR4 was inversely correlated (Fig. 4⇓). This inverse expression pattern reflects lower levels of CXCR4 on memory T cells (30) (data not shown), which preferentially express Bonzo (Fig. 2⇑d).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Bonzo and CCR5 are expressed on similar T cell subsets. PBMC were stained with mAbs against either Bonzo or CCR5 and then with secondary PE-conjugated Ab followed by CD3-TC and FITC-conjugated anti-CD8, -CD4, or -CD45RO mAbs. Gating was restricted to CD3⁺ cells. a, CCR5 expression on T cell subsets and comparative staining of Bonzo (bottom panel). b, Bonzo expression of T cell subsets from a CCR5-null individual.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Coexpression of Bonzo with CCR5. PBMC were stained with Bonzo as previously described, followed by PE-conjugated anti-CCR5, -CCR6, or -CXCR4 and TC-conjugated anti-CD3 mAbs. Profiles are gated on CD3⁺ cells.

We also analyzed the expression of Bonzo within the thymus and secondary lymphoid organs, adenoids and tonsils. Bonzo, CCR5, and CCR6 expression was either undetectable or was present on <1% of thymocytes, whereas the majority of the thymocytes expressed CXCR4 (data not shown). Expression profiles of Bonzo, CCR5, and CCR6 on the lymphocytes isolated from tonsils or adenoids were similar to those observed on PBMC, although a greater proportion of T cells was positive, and most of these cells expressed CD45RO⁺ as well as the early activation marker CD69 (data not shown).

Regulation of Bonzo expression by cytokine stimulation of resting T cells

The concordant subset distribution of Bonzo and CCR5 suggested that the surface expression of these molecules may be regulated by similar mechanisms. It has been noted that IL-2 or IL-15 stimulation of human T cells can up-regulate some of the CC chemokine receptors (27, 32, 33, 34). We asked whether Bonzo is similarly regulated by cytokines on resting human T cells. CD4⁺ and CD8⁺ resting T cells were purified from PBMC and cultured for 8 days in the presence of various cytokines. Culture of CD4⁺ or CD8⁺ T cells with IL-2 or IL-15 resulted in up-regulated expression of Bonzo exclusively on CD45RO⁺ memory T cells (Fig. 5⇓), with expression on CD8⁺ T cells usually greater than that on CD4⁺ T cells. This result is consistent with the staining pattern observed on freshly isolated PBMC. Little expression was observed on CD4⁺ T cells cultured with IL-4 or IL-7 (Fig. 5⇓), and none was found with IL-6, IL-12, or IFN-γ (data not shown). However, significant up-regulation of Bonzo was observed on CD8⁺ T cells cultured with IL-7 (Fig. 5⇓, lower panel). IL-2 and IL-15 also up-regulated CCR5 and CCR6 expression on CD4⁺ T cells, and most Bonzo-expressing cells also coexpressed these CC chemokine receptors (Fig. 6⇓). Similar results were obtained with CD8⁺ T cells (data not shown). This result recapitulates the coordinate expression profile of Bonzo with CCR5 or CCR6 in freshly isolated PBMC (Fig. 4⇑). Conversely, Bonzo-positive cells had slightly lower CXCR4 expression than Bonzo-negative cells after stimulation with IL-2 or IL-15 (Fig. 6⇓). CXCR4 expression was significantly up-regulated in the presence of IL-4, and there was also a moderate increase in response to IL-2 or IL-15, in accordance with previous reports (35, 36, 37).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Cytokine-mediated regulation of Bonzo and CCR5 on resting CD4⁺ and CD8⁺ T cells. Purified CD4⁺ or CD8⁺ resting T cells were cultured for 8 days in the presence of the cytokines IL-2 (200 U/ml), IL-4 (20 ng/ml), IL-7 (10 ng/ml), and IL-15 (10 ng/ml). Cells were then stained with mAbs specific for Bonzo and for CD45RO and CD4 or CD8 (top and lower panels, respectively).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Cytokine-mediated coregulation of Bonzo and chemokine receptors CCR5, CCR6, or CXCR4 on resting T cells. Purified CD4⁺ T cells were cultured with cytokines as described in Fig. 5⇑. Cells were stained for Bonzo, CCR5, CCR6, or CXCR4 expression as described in Fig. 4⇑.

Post-transcriptional regulation of Bonzo cell surface expression

Most chemokine receptors contain a DRY sequence motif that appears to be required for coupling to G proteins (38). In contrast to other CC and CXC-family chemokine receptors, Bonzo possesses a noncanonic DRY box sequence at the second intracellular loop. This raised the possibility that in response to its physiologic ligand Bonzo may not signal in a similar fashion to the other chemokine receptors. Recently, it has been reported that chemokine receptors are down-regulated through ligand-mediated endocytosis (39, 40, 41, 42). Therefore, we reasoned that if any of the CCR5 ligands also bound to Bonzo, we may be able to detect this through down-regulation of Bonzo in the absence of detectable signaling. Fig. 7⇓a shows that none of the known CCR5 ligands (MIP-1α, MIP-1β, and RANTES) had any effect on Bonzo expression, whereas they completely down-modulated CCR5 at the same concentrations. As expected, CXCR4 was down-modulated by its ligand SDF-1α, but not by CCR5 ligands (Fig. 7⇓a). Interestingly, stimulation of cells with phorbol ester (PMA) also did not have any effect on Bonzo expression, although CCR5 and CXCR4 expression was completely down-modulated (Fig. 7⇓b). Similarly, upon restimulation of T cell lines with mitogen, CCR5 expression was down-modulated, but that of Bonzo was unaffected (Fig. 7⇓c). To determine whether Bonzo is premade, stored in intracellular compartments, and released to cell surface upon signaling, we performed intracellular staining on resting or cytokine-stimulated primary T cells with anti-Bonzo Ab. No difference was observed between cell surface staining alone or in conjunction with intracellular staining (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Effects of chemokines, PMA, and mitogen stimulation on Bonzo, CCR5, and CXCR4 cell surface expression. To maximize expression levels, CD4⁺ cells cultured with IL-15 for 12 days were used in the Bonzo and CCR5 down-regulation experiments, and IL-4-cultured cells were used for examining CXCR4 down-regulation. a, Cells were cultured for 24 h in the presence of the chemokines RANTES, MIP-1α, MIP-1β, or SDF-1α (1 μg/ml each) before staining. b, Cells were stimulated with 15 ng/ml PMA for 12 h and stained for chemokine receptor expression. c, Activated CD4⁺ human T cell lines maintained in IL-2 were stained for chemokine receptor expression 2 wk poststimulation (upper panel) or 3days after restimulation through TCR (lower panel).

Targeted disruption and knockin of EGFP into the mouse Bonzo gene locus

To gain insight into the expression pattern and function of Bonzo in mice, we used a gene-targeting strategy that replaced the Bonzo-coding sequence with that of EGFP (Fig. 8⇓A) Germline transmission was obtained with two independent ES cell clones. Deletion of the Bonzo gene was verified by Southern blot analysis (Fig. 8⇓B). In an RT-PCR assay, a 3′ portion of Bonzo mRNA was amplified from total RNA derived from lymph node and spleen of both Bonzo^+/− (Bz^+/−) and Bonzo^−/− (Bz^−/−) mice. No amplification was observed from Bz^−/− mice (Fig. 8⇓C). Mating of heterozygous mice gave rise to Bz^−/− mice in Mendelian proportions. Bz^−/− mice were phenotypically indistinguishable from Bz^+/+ and Bz^+/− littermates in a specific pathogen-free environment. Histologic analysis of organs displayed no morphologic difference among the three genotypes. Flow cytometric analysis of cells from lymphoid organs showed no alteration in various cell populations (data not shown). Proliferation of T cells from Bz^+/− or Bz^−/− mice was assessed in response to different concentrations of anti-CD3 Ab in the presence or the absence of anti-CD28 Ab. No difference was observed between Bz^−/+ and Bz^−/− cells (data not shown). Nucleoprotein-keyhole limpet hemocyanin-immunized Bz^+/− and Bz^−/− mice also displayed similar levels of nucleoprotein-specific IgM and IgG1 production (data not shown), suggesting that T cell help and B cell function were normal in Bonzo-deficient mice.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

Targeted knockin of EGFP in the mouse Bonzo gene. A, From the top, the wild-type Bonzo genomic structure, the targeting construct, and the targeted Bonzo/EGFP knockin structure, respectively. The arrows above each gene depict the directions of transcription. ▵, loxP sites. H, HindIII; A, ApaI; RV, EcoRI; C, ClaI. B, Southern blot analysis of progeny of heterozygous intercrossing. The probe used is depicted by the solid bar beneath the targeted structure in A. The wild-type allele yields a 7-kb EcoRV genomic fragment, whereas the targeted allele yields a 6.6-kb fragment. C, RT-PCR analysis of expression of Bonzo mRNA in mutant mice. The primers used are described in Materials and Methods. In a TAE gel, a 590-bp DNA fragment is amplified from wide-type (top panel, lane 2) and heterozygous (lane 3) mice, but is absent in Bz^−/− mice (lane 4). A similar RT-PCR assay was conducted for GADPH mRNA, as a loading control (lower panel).

Tracing mouse Bonzo expression in EGFP knockin mice

The in-frame substitution of EGFP for Bonzo yields a sensitive tool to trace Bonzo expression and regulation in vivo. Histologic analysis did not show any GFP expression in nonlymphoid tissues. Strong green fluorescence was detected in spleen and lymphoid nodes. In spleen, GFP-positive cells were clustered in the periarteriolar lymphoid sheath area in the white pulp, where T cells reside (Fig. 9⇓). Notably, there was no difference in expression pattern between Bz^+/− and Bz^−/− mice.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 9.

Expression of GFP in spleen sections of Bonzo knockin mice. The section shows GFP fluorescence in T cell-rich periarteriolar lymphoid sheath area from Bonzo knockin mice.

We next performed FACS analysis of GFP expression in different lymphoid subsets. GFP expression was undetectable in B cells, granulocytes, monocytes or macrophage, and dendritic cells (data not shown). In thymus, no expression was observed in CD4⁺CD8⁺ cells or double-negative thymocytes. Only a small subset of single-positive cells showed strong expression (data not shown).

Within the lymph nodes and spleen almost all GFP⁺ cells were CD3⁺ T cells (Fig. 10⇓a). The mean fluorescence of GFP-expressing Bz^−/− cells was slightly higher than that of Bz^+/− cells, most likely due to GFP expression from both targeted alleles (data not shown). Few CD4⁺ T cells expressed GFP, whereas the majority of CD8⁺ T cells were positive, with a broad distribution of intensity of GFP expression. We next analyzed the distribution of GFP expression on naive vs memory T cells in Bonzo knockin mice. CD8⁺ cells were all CD44⁺, and it was difficult to discriminate naive vs memory cells (Fig. 10⇓b). γδ T cells from lymphoid organs or resident within the skin, reproductive tract, or gut also expressed GFP (Fig. 10⇓c). Interestingly, GFP expression on CD4⁺ murine T cells was restricted to the CD44⁺ and CD45RB^low subset (Fig. 10⇓d), suggesting that, as in human CD4⁺ T cells, Bonzo is primarily expressed in the memory subset of mouse CD4⁺ T cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 10.

GFP expression on lymphocyte subsets from Bonzo knockin mice. FACS analysis of T cell subsets from lymph nodes of Bonzo-GFP knockin mice was performed. a, GFP expression on CD3⁺ T cells; b, cells stained with anti-mouse CD8-TC and CD44-PE, gated on CD8⁺ T cells; c, IEL stained with PE-conjugated anti-TCRγδ Ab; d, GFP expression on naive and memory mouse CD4 T cells. Lymph node cells stained with CD4-TC and with either CD44-PE or CD45RB-PE and analyzed after gating was set on CD4⁺ cells.

Finally we determined whether, similar to human Bonzo, expression of mouse Bonzo can be modulated by the presence of cytokines. Indeed, we observed dramatic up-regulation of GFP expression after 10 days of culture with IL-2, but less with IL-7 and none with IL-4. These results suggest that cytokine-mediated regulation of murine Bonzo follows the pattern of its human counterpart (Fig. 11⇓).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 11.

Regulation of mouse Bonzo expression by cytokines. CD4⁺ T cells were purified from lymph nodes of Bz^+/− and Bz^−/− mice. Cells (4 × 10⁵) were treated with cytokines (IL-2, IL-7, and IL-4) or activated by plate-bound CD3 Ab and assayed for GFP expression 3 and 10 days later. Most GFP-expressing cells were CD44^high. The result is plotted as the percentage of CD4⁺ T cells that are GFP positive.