Department of Hematology, Research Department 2, University Hospital, Lund, Sweden
Correspondence: Elinor Bülow, M.D., Research Department 2, E-blocket, University Hospital, S-221 85 Lund, Sweden. E-mail: elinor.bulow{at}hematologi.lu.se
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: azurophil granule • chimera • neutrophil • secretion • stability
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The specific bactericidal effect of BPI toward gram-negative bacteria is associated with a strong attraction for their outer envelope LPS [13 ]. This attraction explains in part the LPS-neutralizing activity of BPI [14 ]. The whole BPI molecule is not needed for the bactericidal and the LPS-neutralizing effects; a 25-kDa amino-terminal fragment promotes both activities [4 , 15 ], whereas the carboxy-terminal domain confers an opsonic function of holo-BPI [16 ] and has some LPS-neutralizing activity [4 ]. The carboxy-terminal half may also serve as an anchor in the membrane because of its prominent hydrophobicity [17 ] supported by the membrane-associated localization of BPI in the azurophil granules [18 ]. In addition, this location may protect against degradation, but either or both of the halves of BPI may be necessary for maintaining the stability of the molecule during processing, sorting, and storage. LBP, a circulating structural homolog of BPI, is normally produced by the liver [5 ] and exerts its function by binding to LPS and facilitating the delivery of LPS to CD14 on monocytes/macrophages [19 ]. In contrast, BPI neutralizes the biological effects of LPS. Despite the opposite responses to endotoxin, BPI and LBP share 44% sequence homology and both bind with their amino-terminal half to lipid A of LPS [20 21 22 ].
In this work we have investigated the role of the two domains of BPI in targeting for storage in granules. To this end, we expressed human BPI and LBP in the rat basophilic leukemia (RBL) [23 ] and the murine myeloid 32D [24 ] cell lines to investigate the processing and sorting for storage. These cell lines have previously been utilized for studies of the processing of neutrophil granule proteins [7 , 9 , 11 , 12 ]. Our aim was to examine the role of the amino-terminal and carboxy-terminal domains of BPI for protein stability and targeting for storage. Thus, we determined the consequences of deletion of the carboxy-terminal or the amino-terminal half, respectively, for posttranslational processing, targeting for storage, and stability in myeloid cells. Two chimeric proteins, BPI/LBP and LBP/BPI, that contain the amino- and the carboxy-terminal domain of BPI or LBP were also examined. Our results suggest a function of the carboxy-terminal half of BPI in the intracellular stability of the holoprotein and a potential role for the amino-terminal half of BPI in sorting.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
cDNA, mutagenesis, and construction of expression vectors
The human cDNA for BPI was generously provided by Dr. P.
Elsbach, New York University, New York, NY. All 5’- and 3’-noncoding
sequences were removed, the "Kozak" consensus leader sequence for
maximal translational efficiency [25
] was introduced,
and the flanking restriction enzyme sites HindIII and NotI were
added for subsequent cloning into plasmid. To this end polymerase chain
reactions (PCR) were performed with full-length BPI cDNA as template in
a 20-cycle reaction with upstream primer
5’-GACTTCAGAAGCTTCCGCCACCATGAGAGAGAACATGGCCAGGGGC
and downstream primer
5’-GACTTCAGGCGGCCGCTCATTTATAGACAACGTCTGC
(start and stop codons in boldface and restriction enzyme sites
underlined). The resulting product was cloned into the expression
vectors pcDNA3 and pMPSV-H for construction of the expression vectors
pcDNA3-BPIwt and pMPSV-H-BPIwt. Control sequencing showed preserved
integrity of the reading frame.
The amino-terminal half and three other carboxy-terminal deletion mutants of BPI were formed by PCR amplifications of BPI cDNA with the following downstream primers: 5’-GACTTCAGGCGGCCGCTCATGGAGGTGCCACCAGACCATAG for introduction of a stop codon (boldface) after Pro213 (numbered from the first amino acid of the mature protein), 5’-GACTTCAGGCGGCCGCTCACTCAGGTAGGAAGGTTCCAAAG for introduction of a stop codon after Glu304, 5’-GACTTCAGGCGGCCGCTCACAGGAGCAGCCTATCCAGC for introduction of a stop codon after Leu386, and 5’-GACTTCAGGCGGCCGCTCAGAGAGGGAAGCCTTTCTGTAG for introduction of a stop codon after Leu427. The deletion mutants were cloned into the pcDNA3 and pMPSV-H vectors to create the expression vectors pcDNA3-BPI1–213, -BPI1–304, -BPI1–386, and -BPI1–427, as well as pMPSV-H-BPI1–213, -BPI1–304, -BPI1–386, and -BPI1–427.
The carboxy-terminal region of BPI (BPI210–456) was made with full-length BPI cDNA as a template in a two-step "spliced overlap extension" PCR [26 ]. In the first reaction two separate amplifications produced two fragments of BPI, one amino-terminal fragment including the signal peptide, and one fragment with the carboxy-terminal part of the protein. In designing the primers, the "Kozak" consensus leader sequence as well as the flanking restriction enzyme sites HindIII and NotI were included. The PCR primers in the two amplifications were upstream 5’-GACTTCAGAAGCTTCCGCCACCATGAGAGAGAACATGGCCAGGGGC (primer 1) plus downstream 5’-CGTGGTTGCTGGAGGTGCCACGGCCGCTGTCACGGCGGTGC (primer 2), and upstream 5’-GTGGCACCTCCAGCAACCAC (primer 3) plus downstream 5’-GACTTCAGGCGGCCGCTCATTTATAGACAACGTCTGC (primer 4), respectively. The PCR products were isolated on agarose gel, mixed, and subjected to a second splicing PCR amplification with primers 1 and 4, thus creating BPI lacking the amino acids 1–209 (BPI210–456). The resulting PCR product was cloned into plasmid (pcDNA3) to create the expression vector pcDNA3-BPI210–456. All PCRs were performed in a Perkin-Elmer 480 Thermal Cycler using Pfu polymerase (Stratagene, La Jolla, CA) according to the manufacturer’s instructions.
The plasmids p4161(B-L), p4160(L-B), and pLBP encoding the fusion proteins BPI1–199/LBP198–456, LBP1–197/BPI200–456, and the protein LBP, respectively, were provided by XOMA (US) LLC [27 ]. The inserts were cut out from the plasmids with the restriction enzymes SalI and XhoI. The 5’-end produced by SalI was made blunt with the Klenow enzyme. The inserts were then ligated to pcDNA3 that had been cut with EcoRV (blunt end) and with XhoI, to create the expression vectors pcDNA3-BPI/LBP, -LBP/BPI, and -LBP.
The cDNA constructs were investigated by in vitro translation and were found to encode for proteins of the expected molecular size as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The cDNA constructs used in this work are shown in Figure 1 .
 View larger version (34K): [in a new window] |
Figure 1. Schematic view of the cDNA constructs used. The numbers indicate the
first and last amino acid in the respective peptide, numbered from the
first amino acid of the mature protein. SP, signal peptide.
|
Transfection
The RBL-1 cells were transfected with pcDNA3 constructs through
the use of the Bio-Rad Electroporation Apparatus (Bio-Rad, Hercules,
CA), with electrical settings of 960 µF and 260 V as described
previously [29
]. After electroporation, 2 mg/mL
geneticin was added to select for recombinant clones containing the
geneticin-resistant gene of pcDNA3. The eucaryotic expression vector
pMPSV-H contains a promoter from myeloproliferative sarcoma virus but
confers no selection marker for eucaryotic cells. Therefore, the 32D
cl3 cells were cotransfected with a pMPSV-H construct and the pRC/CMV
vector or transfected with pcDNA3 constructs with the same electrical
settings as above, and selected with 1 mg/mL geneticin. Individual
antibiotic-resistant cell clones were selected, expanded in suspension
cultures, and screened for the expression of the transfected protein by
biosynthetic radiolabeling. The 32D cell clones of BPI, LBP, BPI/LBP,
and LBP/BPI (see Fig. 7B
) are, however, not-clonal antibiotic-resistant
transfectants.
 View larger version (46K): [in a new window] |
Figure 7. Differences in processing of BPIwt, LBP, BPI/LBP, and
LBP/BPI. RBL cells (A) and 32D cells (B) transfected with cDNA for
BPIwt, human LBP, the chimeric BPI/LBP, or LBP/BPI were
incubated with
[35S]methionine/[35S]cysteine for 30 min,
followed by chase of the label for up to 5 h. At depicted time
points, 20 x 106 cells were removed, and after lysis,
subjected to immunoprecipitation with anti-LBP (LBP, BPI/LBP) or
anti-BPI (BPIwt, LBP/BPI). In addition, the incubation
medium was also investigated at each chase time. Immunoprecipitates
were analyzed as described in the legend to Figure 2
. The fluorograms
were exposed 12, 11, 12, and 16 days, respectively (A), and 18, 18, 13,
and 13 days, respectively (B).
|
Subcellular fractionation
Subcellular fractionation was performed as described
[29
]. Thus, the postnuclear cell homogenate was
centrifuged in a Percoll density gradient, after which nine fractions
were collected with all the cytosol in fraction 9. The distribution of
lysosomes and Golgi elements in the gradient was determined by assaying
ß-hexosaminidase and galactosyl transferase as described elsewhere
[30
, 31
]. The peak activities of
ß-hexosaminidase and galactosyl transferase in subcellular fractions
from RBL cells were localized in fractions 1–2 (containing the dense
cytoplasmic organelles referred to in this work) and 5–8,
respectively, and from the 32D cells in fractions 2 and 6, respectively
[28
].
Immunoprecipitation
Before immunoprecipitation, whole cells were solubilized in a
lysis buffer consisting of 1 M NaCl, 50 mM Tris-HCl, pH 8.0, 0.5%
Triton X-100 with added protease inhibitors (Complete), frozen and
thawed three times, whereas culture medium or Percoll-containing
subcellular fractions were solubilized in radioimmune precipitation
buffer (RIPA) [29
], also including protease inhibitors.
Biosynthetically radiolabeled BPIwt, truncated forms of BPI and LBP/BPI
were immunoprecipitated with polyclonal anti-BPI while BPI/LBP and LBP
were precipitated with polyclonal anti-LBP, followed by electrophoretic
analysis and fluorography as described before [29
].
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
BPIwt is visualized as a single 55-kDa band (Fig. 2 ). A minor decrease of the intracellular product is seen in 32D cells (Fig. 2B) but not in RBL cells (Fig. 2A) during a 3-h chase of the radiolabeled BPI. Thus RBL cells retain synthesized BPI more efficiently than 32D cells. This is at least in part due to minimal secretion in RBL cells, whereas 32D cells secrete a large amount of newly synthesized BPI. Similar differences were observed for other neutrophil granule proteins expressed in these cells [32 , 33 ].
 View larger version (41K): [in a new window] |
Figure 2. Differences in stability and secretion of BPIwt and
BPI1–213. RBL cells (A) and 32D cells (B) transfected with
cDNA for human BPIwt or BPI1–213 were
incubated with
[35S]methionine/[35S]cysteine for 30 min,
followed by chase of the label for 1 and 3 h. At depicted time
points, 40 x 106 cells were removed and, after lysis,
subjected to immunoprecipitation with anti-BPI. In addition, BPI was
also precipitated from the incubation medium at each time point. The
immunoprecipitates were analyzed by SDS-PAGE in a 10–20% gradient gel
followed by fluorography. The fluorograms were exposed for 6 weeks
(BPIwt) and 15 days (BPI1–213) in panel A, and
3 days and 33 days, respectively, in panel B. The positions of
BPIwt and BPI1–213 are indicated to the right
with arrows.
|
To determine whether newly synthesized BPI1–213 is degraded in the ER as a consequence of misfolding, pulse-chase experiments were carried out in the presence of BFA. BFA induces a disassembly of the Golgi complex, thereby preventing ER-Golgi transport [34 , 35 ]. The degradation of BPI1–213 is clearly diminished in both RBL (Fig. 3A ) and 32D cells (Fig. 3B) in the presence of BFA, and the lower-molecular-weight forms of intracellular BPI1–213 are not visible, indicating that the degradation is inhibited. The secretion of BPI1–213 is abolished by BFA. These results indicate that the degradation of BPI1–213 does not occur to a major extent in the ER, but rather after that the intracellular pathway for storage is separated from that for constitutive secretion. To determine the role of lysosomal degradation for instability, the cells were incubated with chloroquine or NH4Cl, thus blocking lysosomal proteolysis by raising the pH. Results from pulse-chase experiments carried out in the presence of either agent showed diminished degradation of BPI1–213 during chase of the radiolabeled product in both cell lines (Fig. 3) . However, all degradation was not abolished because a lower-molecular-weight form was still produced, indicating a pH-insensitive step. The difference in stability of BPI1–213 in RBL and 32D cells may reflect a difference in the proteolytic equipment of the two cell lines. In any case, these results indicate that the instability of BPI1–213 is caused by lysosomal/granule degradation.
 View larger version (71K): [in a new window] |
Figure 3. The effect of BFA, NH4Cl, and chloroquine on the processing
of BPI1–213. RBL cells (A) and 32D cells (B) transfected
with cDNA for BPI1–213 were incubated with
[35S]methionine/[35S]cysteine for 30 min,
followed by chase of the label for up to 3 h (controls). Separate
experiments were carried out with 5 µg/mL brefeldin A, 10 mmol/L
NH4Cl or 1 µmol/L chloroquine, with the agents present
during starvation (30 min), pulse-labeling, and chase of the label. At
indicated time points, 40 x 106 cells were removed
and, after lysis, subjected to immunoprecipitation with anti-BPI. In
addition, BPI1–213 were also precipitated from the
incubation at each time point. Immunoprecipitates were analyzed as
described in legend to Figure 2
. The fluorograms were exposed for
28–32 days. The position of BPI1–213 is indicated to the
right with arrows.
|
 View larger version (52K): [in a new window] |
Figure 4. Targeting of BPIwt to dense organelles. RBL cells (A)
and 32D cells (B) transfected with cDNA for human BPIwt
were incubated with
[35S]methionine/[35S]cysteine for 30 min
followed by chase of the label for 1 and 3 h. At these time
points, 100 x 106 cells were homogenized, after which
subcellular fractionation of the postnuclear supernatant was performed
by centrifugation in Percoll, with subsequent collection of nine
subcellular fractions, fraction no. 9 containing all cytosol. The
fractions were lysed and subjected to immunoprecipitation with
anti-BPI. Immunoprecipitates were analyzed as described in the legend
to Figure 2
. The fluorograms were exposed for 6 and 8 days,
respectively. The position of BPIwt is indicated with an
arrow to the right. Peak activities of ß-hexosaminidase and
galactosyl transferase were localized in fractions 1–2 and 5–8,
respectively, in RBL cells and in fractions 2 and 6, respectively, in
32D cl3 cells. Peak activities of ß-hexosaminidase and galactosyl
transferase indicate the position of lysosomes and Golgi elements,
respectively.
|
 View larger version (51K): [in a new window] |
Figure 5. Targeting of the amino-terminal half of BPI,
BPI1–213, to dense organelles. 32D cells (A) and RBL cells
(B) transfected with BPI1–213 were incubated with
[35S]methionine/[35S]cysteine for 1 h
and 30 min, respectively, followed by chase of the label for 1 and
3 h or 30 and 60 min, respectively, as indicated. RBL cells, but
not 32D cells, were preincubated, pulse-labeled, and chased in the
presence of 1 µmol/L chloroquine. At depicted time points, 100 x 106 cells were withdrawn and subjected to homogenization
and subcellular fractionation as discribed in the legend to Figure 4
.
The fluorograms were exposed for 33 days. The position of
BPI1–213 is indicated to the right with arrows.
|
BPI210–456 is highly unstable in both RBL and 32D cells (Fig. 6 ). BPI210–456 is unlikely to be transferred to dense cytoplasmic organelles because the instability of the mutant protein is unchanged in the presence of chloroquine, NH4Cl, or BFA in both the RBL (Fig. 6A , compare with the results for BPI1–213 in Fig. 3 ) and the 32D cell line (data not shown). Therefore, BPI210–456 seems to be retained in the ER and degraded, most likely in the proteasome. This conclusion is supported by the finding of diminished degradation of BPI210–456 by N-acetyl-leu-leu-norleucinal (ALLN) in RBL (Fig. 6A) and 32D cells (data not shown), a cysteine protease inhibitor that blocks the 20S proteasome and ER degradation at the same time [36 , 37 ]. Moreover, a highly specific inhibitor of the 20S proteasome, lactacystin [36 , 38 ], diminishes the degradation of BPI210–456 in both RBL and 32D cells (Fig. 6B) , indicating retention in ER and subsequent proteasome degradation of BPI210–456. Some secretion of BPI210–456 becomes visible in RBL cells when the degradation of the mutant protein is diminished by lactacystin.
 View larger version (41K): [in a new window] |
Figure 6. Effects of BFA, NH4Cl, chloroquine, ALLN, and lactacystin
on the processing of BPI210–456. RBL cells (A) and also
32D cells (B), transfected with cDNA for BPI210–456 were
incubated with
[35S]methionine/[35S]cysteine for 30 min,
followed by chase of the label for indicated time periods (controls).
Separate experiments were carried out with 5 µg/mL brefeldin A, 10
mmol/L NH4Cl, 1 µmol/L chloroquine, 50 µmol/L ALLN, or
10 µmol/L lactacystin present during starvation, pulse-labeling, and
chase of the label. At indicated time points, 20 x
106 cells were removed and, after lysis, subjected to
immunoprecipitation with anti-BPI. In addition, BPI21-–456
was also precipitated from the incubation medium at each chase time
point. Immunoprecipitates were analyzed as described in the legend to
Figure 2
. The position of BPI210–456 is indicated with
arrows. The fluorograms were exposed for 8–9 days (A) and 18 days
(B).
|
 View larger version (21K): [in a new window] |
Figure 8. Targeting of LBP, LBP/BPI, and BPI/LBP to dense cytoplasmic organelles.
RBL cells transfected with cDNA for human LBP (A), LBP/BPI (B), or
BPI/LBP (C) were incubated with
[35S]methionine/[35S]cysteine for 30 min,
followed by chase of the label for 1 and 3 h. At times indicated,
100 x 106 cells were homogenized, after which
subcellular fractionation was performed, with subsequent collection of
nine subcellular fractions, fraction no. 9 containing all cytosol. The
fractions were lysed and subjected to immunoprecipitation with anti-LBP
(LBP, BPI/LBP) or anti-BPI (LBP/BPI). Immunoprecipitates were analyzed
as described in the legend to Figure 2
. The fluorograms were exposed
for 14 days. The position of respective protein is indicated with an
arrow to the right. Peak activities of ß-hexosaminidase, fractions
1–2, and galactosyl transferase, fractions 5–8, indicate the position
of lysosomes and Golgi elements, respectively.
|
 View larger version (26K): [in a new window] |
Figure 9. Summary of the results with full-length and chimeric BPI and LBP. A
comparison of the stability, sorting, and secretion of
BPIwt, BPI1–213, BPI210–456, LBP,
BPI/LBP, and LBP/BPI in the RBL and 32D cell lines. The degree of
stability, the amount of protein transferred to dense organelles
(sorting) and the constitutive secretion is depicted. ND, not
determined.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Stored BPI is very strongly associated with the granules and the most extensive solubilization of BPI has been achieved by extraction of cells with sulfuric acid [1 , 40 ]. A comparison by Western blotting between BPI extracted from RBL and 32D cells by sulfuric acid and the lysis buffer used in the present work (0.5% Triton X-100, 1 M NaCl) showed that sulfuric acid indeed gave the most thorough extraction (data not shown). However, sulfuric acid extracts are not appropriate for our purpose because some BPI will be lost during the dialysis of the samples carried out to allow the immunoprecipitation to take place. The density gradient fractions cannot be extracted with sulfuric acid because of agglutination of the Percoll. Therefore, the results might underestimate to some extent total stored amount of BPI. But only a minor decrease, if any, occurred of radiolabeled BPI during pulse-chase experiments carried out in both cell lines (Figs. 2 and 7) . Therefore, the recovery of the label seems to be unchanged and we conclude that the analysis reflects all the radiolabeled BPI present during the course of the experiments.
The myeloid cell lines as models for processing and sorting
The rodent myeloid cell lines RBL [23
] and 32D
[24
, 41
] have been much utilized for
studies of processing and sorting of human neutrophil granule proteins
after transfection of their corresponding cDNAs [reviewed in ref. 42
]. The RBL cells have cytoplasmic granules that are rich in serine
proteases similar to azurophil granules of human neutrophils. These
cells may therefore be suitable for studies of processing and sorting
of neutrophil granule proteins such as serine proteases and BPI. The
32D cells can differentiate toward neutrophils during incubation with
granulocyte colony-stimulating factor [41
]. The
cytoplasmic granules of 32D cells have an abnormal, vacuole-like
appearance but can still function as phagolysosomes [24
,
43
]. Thus, these cell lines are convenient for studies of
processing and sorting of proteins destined for storage in myeloid
granules. One major difference between these cells is seen in the
efficiency of protein retention from secretion; RBL cells are more
efficient than 32D cells in retaining a protein as opposed to
constitutive secretion. Because these rodent cell lines do not
correspond entirely to normal myeloid progenitor cells, additional
studies will eventually be needed in normal hematopoietic progenitor
cells by use of viral transfer of genes encoding granule proteins and
their variants. The present study does not allow the exact localization
of BPI, LBP, truncated variants, and chimeras targeted for storage in
dense cytoplasmic organelles, including the granules of RBL cells and
the vacuole-like granules of 32D cells, which may be heterogeneous. In
this respect too, results from studies in normal progenitor cells,
where feasible, should focus on subcellular localization in azurophil
and specific granules as well as other organelles.
Both halves of BPI are necessary for stability but the
amino-terminal half alone can be sorted for storage
Unlike most soluble proteins, BPI has a unique elongated
shape with a pseudo-twofold symmetry [44
]. Thus, it is
organized in an amino-terminal and one carboxy-terminal domain (which
correspond in this study to BPI1–213 and BPI210–456, respectively)
connected by a proline-rich linker of 21 amino acids. Both the amino-
and the carboxy-terminal domains contain a barrel formed by two
-helices and an antiparallel ß-sheet. The two barrels are
connected by a smaller central antiparallel ß-sheet formed from the
end and the beginning of each domain [44
]. The
three-dimensional structure of BPI is likely to determine the
stability, since neither the amino- or carboxy-terminal domains of BPI
were stable when expressed in RBL or 32D cells. An important role for
the central ß-sheet of BPI in stabilization of the molecule was
suggested by the finding of increased stability of deletion variants of
BPI retaining the ß-sheet, compared with BPI1–213, lacking it.
The amino-terminal half of BPI (BPI1–213) accumulated in the dense cytoplasmic organelles indicating that the carboxy-terminal half may not be required for targeting. However, in contrast to full-length BPI, the accumulated amino-terminal half is unstable after transfer to dense organelles, indicating a requirement for the carboxy-terminal half to protect against degradation. An alternative possibility is that BPI may normally be protected against degradation as a result of its peripheral localization within azurophil granules [18 ]. In contrast, BPI1–193 [40 ] as well as rBPI23 and rBPI21 [45 ] are produced routinely as stable entities in non-myeloid cells. The stability may be explained by constitutive secretion and absence of the proteases that are abundant in myeloid cells. The carboxy-terminal half of BPI is hydrophobic and might mediate an association with the membrane. Likewise, although the amino-terminal domain of BPI binds to LPS of gram-negative bacteria, the carboxy-terminal domain of holo-BPI promotes opsonophagocytosis of the bacteria by mediating binding to the surface of the phagocyte [16 ]. The carboxy-terminal domain of BPI alone (BPI210–456) does not appear to folded in a manner that allows release from the ER. Thus, this half of BPI was degraded by a proteasome-mediated activity indicated by the increased stability in the presence of lactacystin, a specific proteasome inhibitor. Therefore, it is not possible to determine whether BPI210–456 also would accumulate in dense organelles if it could escape from the ER.
Our data suggest that the myeloid-derived protein BPI is better adapted than the secretory protein LBP for sorting for storage in myeloid cells because it is more stable than LBP. Because of the close relationship between BPI and LBP, the two-domain organization of BPI provides a model also for LBP [46 ] and the homologous organization of BPI and LBP is apparently conserved between different species [47 ]. The results from studies of the processing and sorting of chimeras between BPI and LBP further emphasized that the amino-terminal and the carboxy-terminal halves may have separate functions (Fig. 9) . Both hybrids are, like full-length BPI and LBP, targeted for storage. However, LBP/BPI is more stable than full-length LBP (or BPI/LBP), indicating that the carboxy-terminal domain of BPI may be adapted to confer protection against degradation in the storage compartment. Moreover, the amino-terminal domain of BPI may confer targeting for storage because BPI/LBP displayed lower secretion than LBP or LBP/BPI in both cell lines.
Mechanisms for targeting for storage in myeloid cells
Neutrophil granules are produced by fusion of coated vesicles
derived from the trans-Golgi network. Sorting of proteins may depend on
their conformation and aggregation. Studies have indicated that a
mature conformation is compatible with storage in dense cytoplasmic
organelles of hematopoietic serine proteases [11
,
42
]. In this case, removal of the amino-terminal
activation peptide, the carboxy-terminal prodomain, or the
glycosylation sites does not interfere with targeting in 32D and RBL
cells. However, the elimination of the propeptide of proMPO blocks
sorting for storage in 32D cells [9
]. Thus, in this
case, a propeptide seems to carry a sorting signal for targeting or
affect the physical state of proMPO that normally allows targeting.
Likewise, the propeptide of defensins seems necessary to facilitate
targeting for storage in dense organelles of 32D cells
[8
]. Unlike most neutrophil granule proteins, BPI is not
modified by proteolytic trimming during subcellular sorting. Our
results suggest that the protein conformation may be a determinant for
sorting to storage granules to avoid constitutive secretion. Obviously,
the conformation of the amino-terminal half of BPI when expressed alone
(BPI1–213) allows targeting for storage but not stability. This
observation exemplifies that successful storage depends on both
targeting and stability at the storage site. The stability of BPI
during sorting and storage was found to require the carboxy-terminal
domain. Thus, the hydrophobic carboxy-terminal domain of BPI plays a
role in opsonization extracellularly and a role for prevention of
autodigestion in the intracellular environment. The segregation of
neutrophil proteins into various storage compartments (granules) is
probably a requisite for coexistence of biologically active
constituents and the timed expression of the genes for the granule
proteins coordinates the composition of each granule compartment so as
to promote the stability of the included proteins [reviewed in ref. 6
]. Our data suggest that, in addition to timed gene expression, the
conformation of BPI is also important for long-term storage in
granules.
Cell-specific targeting for storage
Myeloid cells are specialized in sorting for storage in granule
compartments and may have an ability to segregate even non-myeloid
proteins into regulated secretion (e.g., to granules). We show here
that not only BPI but also the normally secretory LBP was efficiently
sorted for storage in a myeloid cell line. Therefore, this result
suggests that non-myeloid proteins can indeed be targeted to a myeloid
cell storage compartment, indicating that sorting for storage may be
cell-specific. The targeting of LBP may, however, be explained by a
conformation that is closely similar to that of BPI, thus allowing
sorting in myeloid cells. Moreover, even if heterologously expressed
non-myeloid proteins are sorted for storage in a granule compartment of
myeloid cells, they will not necessarily be resistant against
degradation during storage.
Finally, what are the implications of these results for the understanding of determinants for sorting in myeloid cells to avoid constitutive secretion? In particular, our findings illuminate functions of the BPI domains. BPI has two functionally distinct domains with separate roles extra- and intracellularly. We propose a critical stability role for the hydrophobic carboxy-terminal domain during sorting for storage and a potential role for the amino-terminal domain in sorting.
rdsdistrikt.
We thank Dr. Peter Elsbach for the cDNA of BPI and XOMA (US) LLC, for
LBP, BPI/LBP, and LBP/BPI plasmids and the LBP antibody. We are
grateful to Ann-Maj Persson for expert technical assistence. Received February 24, 2000; revised May 9, 2000; accepted May 10, 2000.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  Y. Gao, M. Hansson, J. Calafat, H. Tapper, and I. Olsson Sorting soluble tumor necrosis factor (TNF) receptor for storage and regulated secretion in hematopoietic cells J. Leukoc. Biol., October 1, 2004; 76(4): 876 - 885. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Alexander, J. Bramson, R. Foley, and Z. Xing Protection from endotoxemia by adenoviral-mediated gene transfer of human bactericidal/permeability-increasing protein Blood, January 1, 2004; 103(1): 93 - 99. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Rosen, Y. Gao, E. Johnsson, and I. Olsson Artificially controlled aggregation of proteins and targeting in hematopoietic cells J. Leukoc. Biol., November 1, 2003; 74(5): 800 - 809. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Gao, H. Rosen, E. Johnsson, J. Calafat, H. Tapper, and I. Olsson Sorting of soluble TNF-receptor for granule storage in hematopoietic cells as a principle for targeting of selected proteins to inflamed sites Blood, July 15, 2003; 102(2): 682 - 688. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Bulow, N. Bengtsson, J. Calafat, U. Gullberg, and I. Olsson Sorting of neutrophil-specific granule protein human cathelicidin, hCAP-18, when constitutively expressed in myeloid cells J. Leukoc. Biol., July 1, 2002; 72(1): 147 - 153. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Bulow, W. M. Nauseef, M. Goedken, S. McCormick, J. Calafat, U. Gullberg, and I. Olsson Sorting for storage in myeloid cells of nonmyeloid proteins and chimeras with the propeptide of myeloperoxidase precursor J. Leukoc. Biol., February 1, 2002; 71(2): 279 - 288. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
