2B4, the natural killer and T cell immunoglobulin superfamily surface protein, is a ligand for CD48

Abstract

2B4 is a cell surface glycoprotein related to CD2 and implicated in the regulation of natural killer and T lymphocyte function. A recombinant protein containing the extracellular region of mouse (m)2B4 attached to avidin-coated fluorescent beads bound to rodent cells, and binding was completely blocked by CD48 monoclonal antibodies (mAbs). Using surface plasmon resonance, we showed that purified soluble mCD48 bound m2B4 with a six- to ninefold higher affinity (Kd approximately 16 microM at 37 degreesC) than its other ligand, CD2. Human CD48 bound human 2B4 with a similar affinity (Kd approximately 8 microM). The finding of an additional ligand for CD48 provides an explanation for distinct functional effects observed on perturbing CD2 and CD48 with mAbs or by genetic manipulation.

Figure 1

CD2 IgSF subfamily. (A) Schematic representation of mouse CD2, CD48, 2B4, and Ly-9 and human CD58, CDw150 (SLAM), and CD84. (B) Recombinant chimeric m2B4-CD4d3+4 protein. (C) The 2B4-CD4d3+4 fluorescent beads used for ligand identification.

Figure 2

m2B4-CD4d3+4 fluorescent beads bind to rodent CD48 on cells. m2B4-CD4d 3+4 fluorescent beads bind to LPS (A and D) and Con A (B)– activated and resting (C) mouse spleen cells and a rat basophilic leukemia cell line (E). Preincubation of cells with mCD48 mAbs OX78 (A and B) or HM48-1 (C) or rCD48 mAb (E) blocked binding to the level seen with the negative control, rCD5-CD4d3+4. Control mCD2 mAb (A and B), mCD3 mAb 2C11 (C), or rCD147 mAb (E) did not block binding. Preincubation of m2B4-CD4d3+4 fluorescent beads with 2B4 mAb but not a CD4d3+4 mAb partially blocks binding to cells (D).

Figure 2

m2B4-CD4d3+4 fluorescent beads bind to rodent CD48 on cells. m2B4-CD4d 3+4 fluorescent beads bind to LPS (A and D) and Con A (B)– activated and resting (C) mouse spleen cells and a rat basophilic leukemia cell line (E). Preincubation of cells with mCD48 mAbs OX78 (A and B) or HM48-1 (C) or rCD48 mAb (E) blocked binding to the level seen with the negative control, rCD5-CD4d3+4. Control mCD2 mAb (A and B), mCD3 mAb 2C11 (C), or rCD147 mAb (E) did not block binding. Preincubation of m2B4-CD4d3+4 fluorescent beads with 2B4 mAb but not a CD4d3+4 mAb partially blocks binding to cells (D).

Figure 3

h2B4-CD4d3+4 fluorescent beads bind to hCD48 on cells. (A) h2B4-CD4d3+4 fluorescent beads bind to PBMCs. Preincubation of cells with hCD48 mAb 6.28 but not MEM 102 blocked binding to the level seen with the negative control, rCD4d3+4. (B) CD48 mAbs 6.28 and MEM 102 bind to the cells in A as detected with a fluorescent- labeled second mAb.

Figure 4

Affinity and dissociation rate of soluble mCD48 binding to m2B4-CD4d3+4. (A) Soluble mCD48 was injected at the indicated concentration through flow cells with immobilized m2B4-CD4d3+4 (3,733 RU) or as a negative control, rCD5-CD4d3+4 (1,857 RU), at 37°C. (B) The difference between the response at equilibrium in the m2B4-CD4d3+4 and control flow cells is plotted against the mCD48 concentration. A K _d = 15 μM and maximum binding of 1,197–1,225 RU were calculated by nonlinear curve fitting of the Langmuir binding isotherm (line) to data (circles) from A with negative control subtracted. (C) Scatchard plot of data in B plus data for mCD48 binding mCD2. (D) Soluble mCD48 (23.4 μM) was injected over immobilized m2B4-CD4d3+4 at high (1,060 RU) and low (572 RU) levels, mCD2 (557 RU), and rCD5-CD4d3+4 (1,083 RU). A k _off = 3 s⁻¹ for 2B4, 22 s⁻¹ for mCD2, and 41 s⁻¹ for rCD5-CD4d3+4 was calculated by exponential decay curve fitting (line) to dissociation data (symbols). (E) In the experiment shown in D, soluble mCD48 (23.4 μM) was injected over immobilized m2B4-CD4d3+4 (1,060 RU) before and after saturation with 2B4 mAb.

Figure 5

Affinity and dissociation rate of soluble hCD48 binding to h2B4-CD4d3+4. (A) Soluble hCD48 was injected at the indicated concentration through flow cells with immobilized h2B4-CD4d3+4 (2,166 RU) or as a negative control, CD4d3+4 (2,112 RU), at 37°C. (B) The difference between the response at equilibrium in the h2B4-CD4d3+4 and control flow cells is plotted against the hCD48 concentration. A K _d = 8 μM and maximum binding of 945–959 RU were calculated by nonlinear curve fitting of the Langmuir binding isotherm (line) to data (circles) from A with negative control subtracted. (C) Scatchard plot of data in B. (D) Soluble hCD48 (14.4 μM) was injected over immobilized h2B4-CD4d3+4 at high (1,073 RU) and low levels (578 RU), and CD4d3+4 (1,782 RU). A k _off = 5 s⁻¹ for 2B4 high, 7 s⁻¹ for h2B4 low, and 41 s⁻¹ for CD4d3+4 was calculated by exponential decay curve fitting (line) to dissociation data (symbols).

Figure 5

Affinity and dissociation rate of soluble hCD48 binding to h2B4-CD4d3+4. (A) Soluble hCD48 was injected at the indicated concentration through flow cells with immobilized h2B4-CD4d3+4 (2,166 RU) or as a negative control, CD4d3+4 (2,112 RU), at 37°C. (B) The difference between the response at equilibrium in the h2B4-CD4d3+4 and control flow cells is plotted against the hCD48 concentration. A K _d = 8 μM and maximum binding of 945–959 RU were calculated by nonlinear curve fitting of the Langmuir binding isotherm (line) to data (circles) from A with negative control subtracted. (C) Scatchard plot of data in B. (D) Soluble hCD48 (14.4 μM) was injected over immobilized h2B4-CD4d3+4 at high (1,073 RU) and low levels (578 RU), and CD4d3+4 (1,782 RU). A k _off = 5 s⁻¹ for 2B4 high, 7 s⁻¹ for h2B4 low, and 41 s⁻¹ for CD4d3+4 was calculated by exponential decay curve fitting (line) to dissociation data (symbols).