The different types of human neutrophil granules (azurophil, specific, and gelatinase granules) are formed sequentially during maturation of neutrophils from the promyelocyte stage to the band cell stage. The promyelocytic HL-60 cells can maturate to segmented granulocytes but are incapable of activating the transcription of any known intragranular protein, normally located in specific or gelatinase granules. To study the sorting of granule proteins during maturation, we transfected HL-60 cells with the specific granule protein NGAL, inserted under control of a cytomegalovirus promoter. We previously showed that NGAL is sorted to azurophil granules and colocalizes with myeloperoxidase in undifferentiated HL-60 cells. We show here that, when such transfected HL-60 cells differentiate into granulocytes, newly synthesized NGAL is not retained in granules but is constitutively secreted. This indicates that highly specific mechanisms must exist that are responsible for diverting transport vesicles into storage granules, and that HL-60 cells not only lack the ability to activate transcription of specific granule proteins, but also lose the ability to form storage granules during maturation.
THE HUMAN NEUTROPHIL contains several subsets of granules that each contribute to specific functions of the cell by serving as stores of proteins that are either liberated from the cell during exocytosis of granules, in the case of intragranular proteins, or become incorporated into the plasma membrane, in the case of granule membrane proteins.1 The granules of human neutrophils are classified as peroxidase-positive (azurophil or primary) or peroxidase-negative (specific or secondary) depending on their content of myeloperoxidase (MPO).2 These subsets can be further classified on the basis of their content of other proteins. Peroxidase-positive granules may be divided into defensin-rich and defensin-poor granules,3 and peroxidase-negative granules can be classified into three subsets on the basis of their content of lactoferrin and gelatinase, ie, granules that contain lactoferrin but no gelatinase, granules that contain both lactoferrin and gelatinase, and granules that contain gelatinase but no or very little lactoferrin.4 Whereas the first two are referred to as specific granules because lactoferrin is a time-honored marker for specific granules, the latter are known as gelatinase granules, or tertiary granules.4,5 These subsets differ not only in their intragranular content, but also regarding their mobilization during neutrophil activation.5-7 Therefore, the structural heterogeneity of granules is functionally important.
Tight control of protein targeting to the correct granule subsets during formation of granules is fundamental to this structural and functional granule heterogeneity. However, with the excessive number of different granule subsets, such targeting can hardly rely on specific recognition signals present in each protein. We have forwarded the hypothesis that such specific targeting signals do not exist in the human neutrophil, and that all proteins that are retained in storage granules will end up in the same type of granules, if synthesized at the same time.8,9 Consequently, the heterogeneity of granule subsets is explained by differences in the time of biosynthesis of the individual granule proteins. This hypothesis is supported by the observations that azurophil granule proteins are synthesized at the promyelocyte stage and that specific granule proteins are synthesized at the myelocyte/metamyelocyte stage, whereas gelatinase is synthesized mainly at the band cell stage.8,10 Direct evidence in support of this hypothesis was recently obtained by the demonstration that NGAL, a newly described specific granule protein,11,12 became targeted to azurophil granules and colocalized with MPO when transfected into the promyelocytic HL-60 cells.13
HL-60 cells can be induced to maturate into granulocytes. In the course of this, they acquire the ability to undergo a respiratory burst, as measured by Nitroblue tetrazolium (NBT) reduction.14 This functional maturation is characterized by the acquisition of the membrane proteins, the flavocytochrome b55814,15 and the adhesion protein Mac-1 (αmβ2 ).16 Despite the ability to activate the synthesis of these membrane proteins that are localized in the membrane of peroxidase-negative granules and secretory vesicles in the normal neutrophil,17,18 HL-60 cells are unable to activate the synthesis of proteins localized inside peroxidase-negative granules, and peroxidase-negative granules have not been identified in differentiated HL-60 cells.19,20
To investigate whether, in addition to the block in the synthesis of peroxidase-negative granule proteins, there is also a block in the formation of granules in differentiated HL-60 cells, we decided to examine the sorting of NGAL in HL-60 cells during maturation of these cells into granulocytes. HL-60 cells, stably transfected with NGAL under control of a cytomegalovirus (CMV) promoter, were driven into granulocytic differentiation by combined treatment with dimethyl sulfoxide (DMSO) and retinoic acid (RA), and the synthesis and sorting of NGAL were examined by pulse-chase studies of newly synthesized protein.
MATERIALS AND METHODS
Materials.RPMI-1640 culture medium, methionine-deficient medium, fetal calf serum (FCS), L-glutamine, streptomycin, and penicillin were purchased from GIBCO-BRL (Gaithersburg, MD). RA (all-trans), NBT, and phorbol myristate acetate (PMA) were from Sigma (St Louis, MO). DMSO was purchased from Merck (Darmstadt, Germany). L-[35S]-methionine (1,175 Ci/mmol) was from New England Nuclear-DuPont (Boston, MA), and Amplify was from Amersham Corp (Arlington Heights, IL). Protein A-Sepharose CL4-B and CNBr-activated Sepharose 4B were purchased from Pharmacia (Uppsala, Sweden). Diisopropyl fluorophosphate (DFP) was from Aldrich Chemical Company Inc. (Steinheim, Germany) PMA was dissolved in DMSO at a concentration of 1 mg/mL and was used to stimulate the cells at a final concentration of 200 ng/mL. RA was prepared as a 10-mmol/L stock solution in DMSO and was used at a final concentration of 1 μmol/L in culture. Rabbit anti-MPO antibody (DAKO A398), anti-CD11b monoclonal antibody (MoAb; DAKO M741), rabbit anti–β2-microglobulin antibody (DAKO A072), rabbit Ig fraction (DAKO X903) and fluorescent-conjugated rabbit antimouse antibody (DAKO F313) were purchased from DAKO (Glostrup, Denmark). Purified mouse IgG1 was from Becton Dickinson (San Francisco, CA).
|Undifferentiated HL-60 Cells||Differentiated HL-60 Cells|
|Morphology||PM, 98%; M + MM, 2%; B + S, 0%||PM, 28%; M + MM, 56%; B + S, 16%|
|NBT-test (% positive cells)||5.7 ± 2.2 (n = 12)||70.5 ± 14.6 (n = 16)|
|CD11b (MFI)||3.8 ± 4.5 (n = 12)||251 ± 128 (n = 12)|
|Undifferentiated HL-60 Cells||Differentiated HL-60 Cells|
|Morphology||PM, 98%; M + MM, 2%; B + S, 0%||PM, 28%; M + MM, 56%; B + S, 16%|
|NBT-test (% positive cells)||5.7 ± 2.2 (n = 12)||70.5 ± 14.6 (n = 16)|
|CD11b (MFI)||3.8 ± 4.5 (n = 12)||251 ± 128 (n = 12)|
Differentiated cells (DMSO [1.3%] and 1 μmol/L RA for 3 to 4 days) and undifferentiated cells were analyzed in parallel for nuclear morphology in cytospin preparations and classified as promyelocytes (PM), myelocytes and metamyelocytes (M + MM), and band and segmented cells (B + S). Results are expressed as the mean percentage of cells showing the corresponding morphological features in four independent experiments. The ability to mount a respiratory burst was assessed by NBT reduction and given as percentage of cells containing blue intracellular formazan deposits after 30 minutes of stimulation by 0.2 μg/mL PMA at 37°C. Cell-surface expression of CD11b was determined by FACS analysis and expressed as MFI. The differentiation of wild-type and transfected HL-60 cells was identical, and results from both cell types were combined in the statistical analysis of the data. Values are mean ± SD of the number of experiments shown in parentheses.
Cells and culture conditions.HL-60 cells were purchased from American Type Culture Collection (Rockville, MD). HL-60 cell subclones, transfected with the cDNA encoding for the specific protein NGAL under control of the constitutively active CMV promoter, as previously described,13 were expanded in culture in parallel with wild-type HL-60 cells. Three different clones, referred to as clones A, B, and C, respectively, as well as wild-type HL-60 cells were used in the experiments. The cells were maintained in RPMI-1640 medium supplemented with 10% (vol/vol) heat-inactivated FCS, 2 mmol/L L-glutamine, 100 U/mL penicillin/streptomycin (complete medium), and 1 mg/mL Geneticin (GIBCO-BRL) except for wild-type HL-60 cells, in a humidified atmosphere of air:CO2 (19:1). Cell cultures were passaged 3 times weekly to maintain a cell density between 2 × 105 and 6 × 105 cells/mL.
Induction of differentiation.Exponentially growing HL-60 cells were harvested by centrifugation and resuspended at a density of 5 × 105 cells/mL in complete RPMI culture medium supplemented with 1.3% (vol/vol) DMSO and 1 μmol/L RA for 3 to 4 days. The medium was renewed after 2 days of differentiation. The combined treatment of HL-60 cells with RA and another polar solvent, N,N-dimethylformamide, has previously been shown to result in maximal granulocytic differentiation.15 Viability was assessed by the ability of the cells to exclude 0.1% trypan-blue dye.
Assessment of cell differentiation.The extent of differentiation/maturation was assessed by monitoring growth arrest, morphological changes, and capacity for superoxide generation as assessed by the NBT test and by quantitation of a differentiation marker (CD11b antigen) by FACScan (Becton Dickinson, Mountain View, CA) analysis.
Growth of HL-60 cells was assessed by counting the number of trypan-blue–negative cells by light microscopy. The differentiation response was also apparent on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) protein profiles (Coomassie blue) of undifferentiated versus the differentiated HL-60 cells as previously reported (Meyer and Howard21 and data not shown). For the morphological assessment of the cells, 105 cells were harvested, centrifuged, and resuspended in 0.1 mL of phosphate-buffered saline (PBS). Slides were prepared by centrifugation of the cells in a cytospin centrifuge, stained with May-Grünwald-Giemsa stain, and examined by light microscopy under oil immersion at 1,000-fold magnification.
NBT dye reduction.Differentiation was confirmed by counting cells capable of reducing NBT in response to PMA stimulation. A total of 1 × 105 cells were harvested by centrifugation and resuspended in 0.1 mL serum-free RPMI-1640 without phenol red, containing 0.1% NBT and 200 ng/mL PMA in a 96-well U-bottomed microtiter plate. The reaction mixture was incubated for 30 minutes at 37°C in an humidified atmosphere containing 5% CO2. The reaction was stopped by placing the plate on ice. After centrifugation, the cells were resuspended in 0.1 mL of PBS, transferred onto slides by cytospin, fixed in methanol, and stained with safranine as previously described.22 Two hundred cells were examined by microscopy for the presence of blue formazan deposits, and the number of positive cells was expressed as percentage of total number of cells.
Flow cytometry.Quantitation of CD11b antigen expression was performed using an FACScan flow cytometer (Becton Dickinson) as previously described.23 Briefly, cells at 107/mL (50 μL) in PBS were mixed with 50 μL primary MoAb at saturating concentration (5 μg/mL anti-CD11b or 1 μg/mL control mouse IgG1) in microtiter plates and were incubated for 30 minutes at 4°C. The cells were then washed twice, suspended in 50 μL buffer containing 30 μg/mL fluorescein-conjugated rabbit antimouse antibody. After incubation for 30 minutes at 4°C, the cells were washed 3 times, fixed in PBS supplemented with 1% formaldehyde, and diluted 5 times in PBS with 1% formaldehyde just before immunofluorescence flow cytometry analysis. At least 5,000 cells were analyzed for each conditions. All washes were performed in PBS supplemented with 0.5% bovine serum albumin at 4°C by centrifugation of microtiter plates at 200g for 5 minutes. Negative controls were obtained using antibodies without relevant specificity as primary antibody. Data are expressed as mean fluorescence intensity (MFI). Specific fluorescence was obtained by subtracting the background MFI obtained by labeling with irrelevant antibody.
Biosynthetic labeling, chase experiments, and immunoprecipitation.Biosynthetic labeling of NGAL-transfected HL-60 cells was performed as previously described.13 Cells suspended at 2.5 × 106 cells/mL in methionine-free medium containing 10% dialyzed, heat-inactivated FCS were incubated for 60 minutes at 37°C, in a humidified atmosphere of air:CO2 (19:1). The cells were then pulsed for 3 hours by the addition of 50 μCi/mL of [35S]-methionine. In some experiments, cells were incubated with 25 mmol/L DFP during the last 30 minutes of the pulse-labeling. A total of 100 μL of cells was withdrawn and applied to a Whatman filter disk (Whatman, Maidstone, UK) for the determination of the total amount of radioactivity incorporated in Trichloroacetic acid (TCA)-precipitable proteins.24 The pulse was stopped by pelleting the cells. The supernatant was kept on ice until immunoprecipitation was performed. The cell pellet was washed once, and biosynthetically labeled proteins were chased by resuspending the cells at 106 cells/mL in complete medium and withdrawing 2 mL for immunoprecipitation at timed intervals. Cells in the samples were pelleted and solubilized in radioimmunoprecipitation assay buffer as described.13 Immunoprecipitations were performed in parallel on the cell lysates and on the corresponding volumes of culture medium, sequentially with affinity-purified rabbit anti-NGAL antibodies (2.5 μg/mL) followed by protein A-Sepharose (4 mg/mL),13 affinity-purified rabbit anti-MPO, and anti–β2-microglobulin antibodies coupled to CNBr-activated Sepharose 4B (25 μL/mL).8 The immunoprecipitates were washed, resuspended in 40 μL of Laemmli SDS-sample buffer,25 boiled, and centrifuged. Twenty microliters of the supernatant was analyzed by electrophoresis. Wild-type HL-60 cells were pulse-labeled in parallel to NGAL-transfected cells, and immunoprecipitation was performed on cell lysates and culture medium as a negative control for NGAL (data not shown).
SDS-PAGE and fluorography.SDS-PAGE was performed on 0.75-mm-thick mini-gels containing 10% polyacrylamide under reducing conditions.25 Gels were then stained with Coomassie blue, destained, treated with Amplify, and dried on filter paper. Fluorography was performed by exposure to Kodak X-Omat AR film (Eastman Kodak, Rochester, NY) at −80°C for 20 hours to 7 days.
To quantify radioactivity of NGAL, MPO, and HLA, individual bands localized on the dried gel by fluorography were excised and treated overnight at 37°C in 1 mL of a solution containing 30% H2O2 and 6% NH3; 5 mL of scintillation liquid was added for counting.26
Enzyme-linked immunosorbent assay (ELISA).NGAL and MPO in cells and culture medium were assayed by ELISA as previously described12,27 and expressed in micrograms or nanograms per 106 cells. Release from undifferentiated and differentiated HL-60 cells was expressed as the amount of marker present in the culture medium in the percentage of the total amount present in cells and culture medium.
Immunocytochemistry.Cytospin preparations of transfected (undifferentiated and differentiated) HL-60 cells were fixed in 4% formaldehyde in 0.1 mol/L phosphate buffer, permeabilized with 1% Triton X-100 and labeled with either rabbit anti-NGAL antibody (2.5 μg/mL) or rabbit anti-MPO (76 μg/mL) as previously described.8,13 Nonimmune serum (16 μg/mL) was used as a negative control for rabbit antibodies. The antibody was visualized by the alkaline phosphatase-antialkaline phosphatase detection kit (DAKO).
Immunoelectron microscopy.HL-60 cells were fixed with a mixture of 0.5% (vol/vol) glutaraldehyde and 4% (wt/vol) paraformaldehyde in 0.1 mol/L phosphate buffer (pH, 7.2) for 1 hour and pelleted in 10% (wt/vol) gelatine in PBS. For the localization of NGAL, ultrathin frozen sections were incubated at room temperature with rabbit anti-NGAL (1/250) followed by incubation with goat antirabbit IgG linked to 10 nm gold (Amersham Netherland, s-Hertogenbosch, The Netherlands). For the localization of cytochrome b558 , sections were incubated successively with MoAb CLB-48 (1/10),28 rabbit antimouse IgG (1/80), and goat antirabbit IgG linked to 10 nm gold. All incubations were for 1 hour. For controls, the primary antibody was replaced by nonrelevant murine or rabbit antiserum. After immunolabeling, the cryosections were embedded in a mixture of methylcellulose and uranyl acetate and examined with a Phillips CM10 electron microscope (Phillips, Eindhoven, The Netherlands).
Differentiation/maturation of transfected HL-60 cells. Differentiation into the granulocyte lineage and maturation of HL-60 was achieved by combined DMSO (1.3% vol/vol) and RA (1 μmol/L) treatment for 3 to 4 days and was assessed by changes in nuclear morphology, ability to mount a respiratory burst, acquisition of CD11b, and downregulation of MPO synthesis.
After induction, most cells matured to myelocytes or beyond (metamyelocytes, band and segmented), as observed by morphological inspection of May-Grünwald-Giemsa–stained cytospin preparations, and acquired the ability to mount a respiratory burst, as evidenced by PMA-induced reduction of NBT (Table 1). As shown in Table 1, the MFI of cells labeled with an MoAb against CD11b increased from 3.8 to 251 as further proof of maturation of the HL-60 cells. The biosynthesis of an endogenous azurophil granule protein, MPO, and of the plasma membrane HLA-complex was investigated in undifferentiated and differentiated cells. As shown in Fig 1, in agreement with earlier observations,29 differentiation results in complete shutdown of MPO synthesis (synthesis in differentiated cells 3% ± 1.1% of level in undifferentiated cells; mean of 7 experiments), whereas only a partial reduction was observed in the synthesis of HLA (synthesis in differentiated cells 68% ± 23% of level in undifferentiated cells; mean of 7 experiments).
Sorting of NGAL in undifferentiated HL-60 cells.We have shown before that NGAL localizes to azurophil granules, when synthesized in undifferentiated HL-60 cells, but will be proteolytically degraded.13 This can be observed in the upper panel of Fig 2A where newly synthesized, cell-associated, NGAL is broken down and eventually disappears. The decrease in amount of NGAL in undifferentiated cells cannot be caused by secretion, because NGAL recovered from the medium was intact throughout the chase period (Fig 2A, middle panel). Furthermore, when cells were treated with 25 mmol/L of the serine protease inhibitor DFP before the chase period, no degradation occurred, and the signal stayed constant during the chase (Fig 3A).
When the fraction of NGAL secreted into medium is measured by ELISA and expressed as the percentage of total NGAL in medium and cells, the degradation of NGAL in cells leads to an overestimation of the release as the percentage of the total. The figure for NGAL secretion to medium in undifferentiated cells as measured by ELISA is 41% (Fig 4). When the sorting of newly synthesized NGAL is evaluated, the effect of degradation is partly corrected by expressing the amount of biosynthetically labeled NGAL present in medium by the end of the chase as the percentage of total labeled NGAL present at the start of chase (Fig 2A lower panel). This figure amounts to 35% ± 5% (mean ± SD; n = 4) of NGAL being secreted, thus arguing for an effective sorting of NGAL to azurophil granules in undifferentiated cells that is almost as effective as the sorting of MPO, of which 12% was found in the medium (Fig 4).
NGAL and MPO in undifferentiated and differentiated HL-60 cells.When the HL-60 cells differentiate, they reduce their content of MPO from 1.24 μg/106 cells to 0.57 μg/106 cells, as observed from Fig 4. This can be explained by one round of cell division after the stop of biosynthesis.2 In contrast, the cellular content of NGAL is reduced from 33.6 ng/106 cells in undifferentiated cells to 4.1 ng/106 cells in differentiated cells (of which some are still undifferentiated; see Table 1). We have shown that NGAL and MPO localize in the same granules in undifferentiated cells.13 Consequently, the loss of previously synthesized NGAL that occurs during differentiation cannot be caused by exocytosis of azurophil granules, because this would affect MPO to the same extent, but must be a result of degradation. This is clearly shown in Fig 5, which shows that undifferentiated cells stain well for both MPO and NGAL, whereas differentiated cells have mainly lost their staining for NGAL but retain the staining for MPO.
Sorting of NGAL in differentiated HL-60 cells.The biosynthetic pulse-chase experiment on NGAL in differentiated cells (Fig 2B) shows that newly synthesized NGAL is not retained in the differentiated cells. Furthermore, in contrast to undifferentiated cells, in which NGAL was broken down in the cells, there is no sign of degradation of newly synthesized NGAL in differentiated cells. When the fraction of newly synthesized NGAL, accumulating in medium during the 8 hours of chase is expressed as the percentage of total synthesized NGAL, the figure amounts to 77% ± 8% (mean, n ; Fig 2B). When quantitated as the total amount of NGAL by ELISA, 82% of NGAL is secreted in medium (Fig 4).
We addressed the possibility that the decrease of intracellular NGAL in differentiated cells could be caused by an increased intragranular degradation potency activated by differentiation and, in contrast to undifferentiated cells, leaving no detectable proteolysis intermediates on gels. Therefore, pulse-chase experiments were performed on differentiated cells treated with DFP under the same conditions as were shown to inhibit the degradation of NGAL in undifferentiated HL-60 (Fig 3A). In contrast to undifferentiated HL-60 cells, 25 mmol/L DFP had no effect on the decrease in intracellular NGAL in differentiated HL-60 cells (Fig 3B), thus strongly indicating that disappearance of newly synthesized NGAL from differentiated cells is due to secretion. This finding is also confirmed by the paucity of NGAL in differentiated cells as visualized by immunostaining (Fig 5) and by electron microscopy, where NGAL staining was observed only by 4 to 5 gold particles per cell section, localized to endoplasmic reticulum and to the Golgi apparatus (Fig 6A). Exceptionally, one NGAL gold particle was observed in a granule.
Therefore, it is concluded that, when a specific granule protein, in casu NGAL, is artificially transcribed in HL-60 cells that have been differentiated along the granulocytic lineage and are unable to transcribe endogenous specific granule proteins, the transfected specific granule protein cannot be retained in granules but is secreted into the medium. This is in contrast to undifferentiated cells, which effectively retain the specific granule protein. This indicates that these cells, in addition to being unable to transcribe endogenous specific granule proteins, also lose the ability to form storage granules as they differentiate.
Finally, to further address this conclusion, we examined differentiated HL-60 cells for the localization of a membrane protein that, in normal neutrophils, is associated with the membrane of peroxidase-negative granules. This was accomplished by immunoelectron microscopy using an antibody that recognizes the flavocytochrome b558.28 As expected, no labeling was observed in undifferentiated HL-60 cells, but abundant labeling was present on the cell surface of differentiated (NGAL-transfected) HL-60 cells (Fig 6B). In addition to the major labeling of the cell surface, some labeling was observed in vesicles localized underneath the cell membrane. Labeling was also always observed over Golgi stacks but was never observed over granules.
Introducing the cDNA of NGAL, a specific granule protein, into HL-60 cells has allowed us to study the sorting of NGAL in relation to the maturation of the promyelocytic HL-60 cells. In their undifferentiated form, HL-60 cells are structurally similar to their normal counterpart, the myeloblast/early promyelocyte; however, the maturation of HL-60 cells into granulocytes is characterized by a significant departure from the route taken by normal granulocyte precursors with regards to formation of peroxidase-negative granules.20 Transcription of genes for proteins located inside peroxidase-negative granules, which normally takes place at the myelocyte, metamyelocyte, and band cell stages, does not occur in HL-60 cells.19,20
A central question is whether these cells, in addition to lacking the ability to activate the synthesis of peroxidase-negative granule proteins, also lose the ability to form storage granules or whether storage granules can be detected if synthesis of a protein, normally located inside these granules is induced artificially by transfecting the corresponding cDNA, which is transcriptionally controlled by a constitutively active promoter. It is known that transfection of cells with von Willebrand factor can induce formation of Weibel-Palade bodies,30 the structure that holds this protein in endothelial cells, and that transfection of lymphocytes with caveolin induces formation of caveolae.31 The results presented here show that induction of NGAL synthesis in HL-60 cells does not lead to formation of peroxidase-negative granules.
We have previously shown that NGAL becomes targeted to azurophil granules and colocalizes with MPO in undifferentiated HL-60 cells,13 thus supporting our hypothesis that all granule proteins synthesized at the same time will localize to the same type of granules.32 It is apparent that the ability to divert NGAL from the “default” route of constitutive secretion into storage granules is lost when HL-60 cells maturate into myelocytes and metamyelocytes. Thus, the noted lack of peroxidase-negative granules in HL-60 cells20 cannot be explained solely by the lack of content but is also caused by the loss of the ability to form granules. This loss of ability to form granules clearly does not affect the ability to form transport vesicles capable of transporting newly synthesized protein out of the cell.
It was shown that cytochrome b558 , which in the normal neutrophil is confined largely to the membrane of peroxidase-negative granules,28,33-35 was localized almost exclusively to the plasma membrane of the differentiated HL-60 cells, in addition to being localized to structures involved in synthesis and routing of proteins. This further testifies to the inability of differentiated HL-60 cells to form granules and argues that, when the route to storage granules is blocked, the default route goes to the surface of the cell, resulting in secretion of soluble proteins (NGAL) and incorporation of membrane proteins into the plasma membrane (cytochrome b558 ).
It is generally assumed that storage granules form by aggregation of smaller immature granules that bud off from the trans-Golgi network.2,36 It can be deduced that such immature granules are fundamentally different from the transport vesicles that mediate constitutive secretion of protein, because the immature granules have a tendency for spontaneous homotypic fusion but need a signal for fusion with the plasma membrane (a feature that characterizes them as regulated storage granules), whereas transport vesicles lack the ability for homotypic fusion but fuse spontaneously with the plasma membrane. The biochemical basis for this fundamentally different behavior of vesicles budding off the Golgi is unknown, but our results show that the biochemical basis for forming storage granules is lost in HL-60 cells, as these mature beyond the promyelocyte stage.
Another important observation that can be deduced from our results is that degradation of proteins occurs in azurophil granules or their precursors. It has been known for a long time that many of the proteins located in azurophil granules are processed from proforms to their mature form after sorting to storage granules in the trans-Golgi network. This has been shown for MPO,37-39 defensins,40 and cathepsin G.41 This processing is important for functional maturation of the proteins. When NGAL becomes localized to azurophil granules, it is clearly broken down, as previously shown by pulse-chase labeling13 and as evidenced by the disappearance of NGAL staining from granules during maturation of HL-60 cells. This degradation is most likely the result of a serine protease, because that degradation can be blocked by DFP. In contrast, DFP has no effect on the disappearance of newly synthesized NGAL that is formed in more mature HL-60 cells that have stopped forming azurophil granules (as evidenced by the lack of MPO biosynthesis). NGAL in these cells is diverted away from the degradative compartment and eventually is recovered intact in the medium. Because NGAL that escapes sorting to granules in undifferentiated HL-60 cells is also recovered intact in the medium, it can be concluded that the azurophil granules, or their precursors, are a major compartment for proteolytic modification and degradation of proteins that are targeted to these granules. This conclusion is in accordance with the observation that the proteolytic processing of the 75-amino acid proform of defensins only takes places in myeloid cells, whereas the intact propeptide is targeted to granules but is not processed when the protein is expressed in the AtT-20 pituitary adenoma cell line.42
A key questions exposed by our results but left to be answered is why do HL-60 cells lose the ability to make storage granules as these cells maturate? Because formation of storage granules is dependent on protein synthesis as opposed to formation of transport vesicles mediating constitutive secretion,43 it is possible that HL-60 cells lose the ability to form storage granules because formation of proteins essential for formation of storage granules is downregulated along with the downregulation of azurophil granule protein synthesis.
V.L.C. is supported by a postdoctoral fellowship from Association pour la Recherche sur le Cancer, and from the Amalie Jørgensen Fund. N.B. is supported by grants from The Alfred Benzon Foundation, The Danish Cancer Society, and The Danish Medical Research Council.
Address reprint requests to Niels Borregaard, MD, PhD, The Granulocyte Research Laboratory, Rigshospitalet L-4042, 9 Blegdamsvej, DK-2100 Copenhagen Ø, Denmark.