Membrane-bound carboxypeptidase E facilitates the entry of eosinophil cationic protein into neuroendocrine cells

Abstract

ECP (eosinophil cationic protein) is a major component of eosinophil granule proteins, and is used as a clinical biomarker for asthma and allergic inflammatory disease. ECP has been implicated in damage to the cell membrane of many tissue types, but the mechanism is not well known. In the present study, mECP-eGFP-6H, a recombinant fusion protein containing mature ECP (mECP), enhanced green fluorescence protein (eGFP) and a His(6) tag (6H), has been expressed, purified and added to GH3 neuroendocrine cells to study the internalization ability of ECP. We found that mECP-eGFP-6H entered into GH3 neuroendocrine cells and inhibited the growth of the cells with an IC(50) of 0.8 microM. By yeast two-hybrid screening and immunoprecipitation, we have identified a specific protein-protein interaction between mECP and CPE (carboxypeptidase E), a well characterized metalloprotease. Further in vivo yeast two-hybrid screening has also revealed that residues 318-387 located in a region of unknown function in mature CPE are indispensable for association with mECP. In addition, the uptake of mECP-eGFP-6H is suppressed by dominant-negative expression of the recycling defect mutant pre-pro-HA-CPE(S471A,E472A) in GH3 cells, suggesting that the entry of mECP-eGFP-6H is associated with the recycling of CPE in GH3 cells. Taken together, we have demonstrated that CPE possesses a novel function to facilitate the entry of ECP to neuroendocrine cells, and such an endocytotic process allows the cytotoxic ECP to inhibit growth of the target cells.

Figure 1. Suppression of GH3 cell growth in the presence of mECP–eGFP–6H

(A) The molecular mass of mECP–eGFP–6H fusion protein (lane 1) and eGFP-6H (lane 2) are about 42 kDa and 27 kDa respectively. (B) Growth of GH3 cells was monitored by the MTT assay; the percentages of viable cells at various concentrations of RNase A (•), eGFP-6H (○) and mECP–eGFP–6H (▾) were plotted. Each experiment was carried out in triplicate and results are means±S.D.

Figure 2. GH3 cells treated with RITC-labelled mECP–eGFP–6H and RITC-labelled RNase A

GH3 cells were seeded on to six-well plates and grown on coverslips at 40000 cells/well. After 24 h, 0.5 μM mECP–eGFP–6H and RITC-labelled RNase A were added to the culture medium respectively, and the cells were fixed and detected by fluorescence microscopy at the time indicated and nuclei of the cells were stained with DAPI (4′,6-diamidino-2-phenylindole; blue). (A) GH3 cells treated with RITC-labelled RNase A for 3 h. No RITC-labelled RNase A was detected. (B) GH3 cells treated with mECP–eGFP–6H for 0 and 3 h. mECP–eGFP–6H was visualized (green). (C) Relative amounts of mECP–eGFP–6H in GH3 cells at different incubation time were detected by Western blot employing monoclonal anti-ECP109 antibody.

Figure 3. Effects on mECP–eGFP–6H uptake into GH3 cells by lysosomotrophic and secretagogue agents

Lanes 1, 2, 3 in (A) and (B) were GH3 cells treated with medium containing 0.5 μM mECP–eGFP–6H for 1, 2 and 3 h respectively. (A) GH3 cells were pre-treated with 15 mM NH₄Cl in culture medium for 1 h at 37 °C under 5% CO₂, and then treated with 0.5 μM mECP–eGFP–6H fusion protein for 1 h (lane 4), 2 h (lane 5) and 3 h (lane 6). (B) GH3 cells were pre-treated with 5 μM forskolin in culture medium for 1 h at 37 °C under 5% CO₂, and then treated with 0.5 μM mECP–eGFP–6H fusion protein for 1 h (lane 4), 2 h (lane 5) and 3 h (lane 6). The uptake of mECP–eGFP–6H was detected by Western blotting using anti-ECP109 monoclonal antibody.

Figure 4. Yeast two-hybrid screen using mECP as bait and a human brain cDNA library identified CPEc

mECP was cloned into bait vector pGBKT7. Yeast strain AH109 was co-transformed with mCPE–pGADT7 or CPEc–pGADT7. (A) Transformants were spread on to SD-Ade-His-Leu-Trp medium plates. Vigorous growth was observed for cells co-expressing the mECP bait plasmid and two CPE clones at the high-strength selection plate. 1, mCPE; 2, CPEc. (B) The expression of mECP–BD and CPE–AD fusion proteins in yeast cells was detected by Western blotting with anti-c-myc and anti-HA antibody respectively. mCPE–AD (upper panel, lane 1), CPEc–AD (upper panel, lane 2) and mECP–BD (lower panel, lanes 1 and 2) fusion proteins were expressed in yeast cells.

Figure 5. ECP is associated with CPE in vitro

The direct interaction between CPE and ECP was confirmed further by immunoprecipitation using purified mECP–eGFP–6H fusion protein and GH3 cell lysates with monoclonal anti-CPE antibody (lanes 1–3) or anti-His₆ antibody (lanes 4–7). The cell lysates were collected, and 100 μg of lysate proteins were pre-mixed with 1 μg of eGFP–6H (lane 1) or 1 μg of mECP–eGFP–6H fusion protein (lane 2) in immunoprecipitation buffer. Lane 3 is GH3 cell lysate loading control. Immunoprecipitation was performed using a rabbit polyclonal anti-eGFP antibody and Protein A–Sepharose, subjecting samples to SDS/PAGE and transferring on to PVDF membrane. Lanes 4 and 5, immunoprecipitation efficiency control of mECP–eGFP–6H or eGFP–6H with anti-eGFP antibody respectively. Lanes 6 and 7, the input protein of mECP–eGFP–6H or eGFP–6H.

Figure 6. Investigation of the interacting region on CPE

(A) CPEc–pGADT7 was used as template to generate seven nested deletion mutants (left-hand panel). The numbers at both ends of each fragment represent the positions of the terminal amino acid residues in the CPE deletion mutants. Interactions were indicated by growth (+) or no growth (−) on SD-Ade-His-Leu-Trp medium plates (right-hand panel). (B) The expression of AD–CPEc and AD-nested CPE deletion mutant fusion proteins was detected by Western blotting using anti-HA antibody (upper left panel). Lanes 1 and 9, AD–CPEc; lanes 2–8: nested 1–7 respectively. The expression of BD–mECP (lanes 1–8) or BD–hRNase A (lane 9) fusion proteins was detected by Western blotting with anti-c-myc antibody (lower left panel). The interacting strength between mECP and CPEc nested deletion mutants was monitored with the Gal4 reporter gene expression by β-galactosidase activity assay (right-hand panel). (C) Amino acid sequence alignment of human CPE 318–387 with duck CPD domain II (CPD D2; Protein Data Bank code 1QMU), human CPM (EC 3.4.17.12) and human CPN (EC 3.4.17.3).

Figure 7. Internalization of mECP–eGFP–6H fusion protein is prevented by expression of the pre-pro-HA–CPE_S471A,E472A cytoplasmic tail mutation protein in GH3 cells

(A) pcDNA3 vector (lane 1) and pre-pro-HA–CPE_S471A,E472A–pcDNA3 (lane 2) were transfected into GH3 cells. The expression of recombinant protein was analysed by SDS/PAGE (upper panel) or Western blotting using monoclonal anti-CPE antibody (lower panel). (B) GH3 cells transfected with pcDNA3 or pre-pro-HA–CPE_S471A,E472A–pcDNA3 (CPE_DN–pcDNA3) were incubated with 0.5 μM mECP–eGFP–6H for the indicated times. The mECP–eGFP–6H fusion proteins in GH3 cells were detected by Western blotting employing anti-ECP109 monoclonal antibody (top panel). The relative amount of mECP–eGFP–6H was quantitatively analysed with α-tubulin as an internal control (bottom panel). Black bars show results of GH3 cells transfected with pcDNA3 vector only. Grey bars show results of GH3 cells transfected with pre-pro-HA–CPE_S471A,E472A–pcDNA3.