Dr Susan E Caldwell Publication Samples

Uploaded on

Collection of some biomedical manuscripts published in the peer-reviewed medical literature, authored by Susan E Caldwell, PhD.

Collection of some biomedical manuscripts published in the peer-reviewed medical literature, authored by Susan E Caldwell, PhD.

  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
    Be the first to like this
No Downloads


Total Views
On Slideshare
From Embeds
Number of Embeds



Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

    No notes for slide


  • 1. Proc. Nati. Acad. Sci. USA Vol. 86, pp. 2844-2848, April 1989 Medical Sciences GLQ223: An inhibitor of human immunodeficiency virus replication in acutely and chronically infected cells of lymphocyte and mononuclear phagocyte lineage (AIDS/macrophages/T lymphocytes/antiviral compound) MICHAEL S. MCGRATHaOb, Kou M. HWANGc"d, SUSAN E. CALDWELLc, ISABELLE GASTONb, KA-CHEUNG LUKc, PAUL WUc'd VALERIE L. NGb e, SUZANNE CROWEb f, JAN DANIELSc, JANE MARSHb, TRCY DEINHARTC, PATRICIA V. LEKASb, JOANN C. VENNARIc, HIN-WING YEUNGg, AND JEFFREY D. LIFSONc'h aDivision of AIDS/Oncology, Department of Medicine, and 'Department of Laboratory Medicine, University of California-San Francisco, San Francisco, CA 94110; bAIDS/Immunobiology Research Laboratory, and fDivision of Infectious Diseases, Department of Medicine, San Francisco General Hospital, San Francisco, CA 94110; Divisions of cCellular Immunology and dMedicinal Biochemistry, Genelabs Incorporated, Redwood City, CA 94063; and gDepartment of Biochemistry and Chinese Medicinal Material Research Center, Chinese University of Hong Kong, Shaitin, NT, Hong Kong Communicated by Leonard A. Herzenberg, December 27, 1988 ABSTRACT GLQ223 is a highly purified, formulated the etiologic agent for AIDS, in cells of both lymphoid and preparation of trichosanthin, a 26-kDa plant-derived ribo- mononuclear phagocyte cell lineages in vitro. some-inactivating protein with potent inhibitory activity against human immunodeficiency virus (HIV) in vitro. The MATERIALS AND METHODS compound produced concentration-dependent inhibition of Source and Purification of GLQ223. Root tubers of T. HIV replication in acutely infected cultures of T-lymphoblas- kirilowii were obtained from southern China. GLQ223 was toid cells (VB cell line). Treatment with GLQ223 selectively purified from an aqueous extract of homogenized tuber reduced levels of detectable viral proteins compared to total material by using modifications of a previously described cellular protein synthesis and produced a selective decrease in procedure (1). Purity ofthe material obtained was determined levels'of viral RNA relative to total cellular RNA in acutely to be >98% by laser densitometric scanning of Coomassie infected cells. Substantial inhibition of viral replication was blue-stained gels (NaDodSO4/PAGE) and by size-exclusion observed at concentrations of GLQ223 that showed little HPLC analysis (data not shown). Authenticity of the material inhibition of parallel uninfected cultures. Selective anti-HIV obtained was confirmed by Western blot analysis with rabbit activity was also observed in cultures of primary monocyte/ antiserum raised against a reference preparation and by macrophages chronically infected with HIV in vitro. When N-terminal amino acid sequence analysis (data not shown). freshly drawn blood samples from HIV-infected patients were Cell Lines and Preparation of Primary Cell Populations. The treated with a single 3-hr exposure to GLQ223, HIV replication VB cell line (5) was used for acute infectivity assays; cells was blocked for at least 5 days in subsequently cultured were maintained in RPMI-1640 supplemented with 10% monocyte/macrophages, without further treatment. The anti- (vol/vol) heat inactivated fetal calf serum. Primary peripheral HIV activity of GLQ223 in both acutely and chronically blood-derived monocyte/macrophages were prepared and infected cells and its activity in cells of both lymphoid and cultured from either HIV-seronegative healthy donors or mononuclear phagocytic lineage make it an interesting candi- from donors known to be infected with HIV by using date as a potential therapeutic agent in HIV infection and procedures described in detail elsewhere (6). The research AIDS. protocol was approved by the Human Subjects Committee, University of California, San Francisco. Bioassays for Anti-HIV Activity. T-cell assays: Acutely GLQ223 is a highly purified, formulated preparation of infected cells. The VB cell line was used to assess the effects trichosanthin, a 26-kDa basic protein isolated from root of GLQ223 on primary HIV infection of T-lymphoblastoid tubers of Trichosanthes kirilowii (1, 20). Based on structural cells. Briefly, cells were inoculated with a titered cryopre- and functional properties, the protein belongs to the family of served virus stock [isolate HIV-lDV (6)] at a multiplicity of single-chain ribosome-inactivating proteins, which inhibit in infection of -0.005. Cells were incubated at a density of 1- vitro translation in cell-free systems (2). Trichosanthin has 5 x 107 cells per ml with the inoculum for 60 min at 370C to been reported to show a 56% similarity at the amino acid level permit adsorption of viral particles and then washed to with the A chain of ricin toxin when both identical and remove unbound virus. Cells were then resuspended at 1.0 x conservative residues are considered (2) and shows substan- 105 cells per ml in RPMI-1640 supplemented with 10% fetal tial amino acid similarity with several other single-chain calf serum and cultured in 24-well culture plates (1 ml per ribosome-inactivating proteins (Michael Piatak, personal well) with or without GLQ223 added to the desired concen- communication). Partially purified preparations of trichosan- tration for the duration of culture. Cells were cultured for 4 thin have been administered in China, in single doses from 5 days, when supernatants were harvested for quantitation of to 12.5 mg, as a midtrimester abortifacient (3) and, in multiple HIV p24 antigen content by using a commercially available dose regimens involving 5-12 mg per dose, for treatment of capture immunoassay (Coulter). In this assay system, at the trophoblastic tumors (4). In general, these doses have been multiplicity of infection used, virally mediated cytopathic reported to be both effective for the intended clinical indi- effects and HIV p24 levels peak at day 4; anti-HIV activity cation and well tolerated (3, 4). was thus evaluated at the time of maximum virus production. In this communication, we report that GLQ223 selectively inhibits replication of human immunodeficiency virus (HIV), Abbreviations: HIV, human immunodeficiency virus; AZT, 3'- azido-3'-deoxythymidine. The publication costs of this article were defrayed in part by page charge hTo whom reprint requests should be addressed at: Division of payment. This article must therefore be hereby marked "advertisement" Cellular Immunology, Genelabs Incorporated, 505 Penobscot in accordance with 18 U.S.C. §1734 solely to indicate this fact. Drive, Redwood City, CA 94063. 2844
  • 2. Medical Sciences: McGrath et al. Proc. Natl. Acad. Sci. USA 86 (1989) 2845 Treated and untreated, infected and uninfected cultures (i.e., concentration-dependent inhibition of HIV replication as mea- four different types of cultures) were also plated in 96-well sured in this acute primary infection assay system. Inhibition plates for determination of incorporation of [3H]leucine and of HIV replication was observed by measurement of super- [3H]thymidine into trichloroacetic acid-precipitable protein natant p24 content (Table 1) and by decreased viral cytopathic and cellular DNA, respectively, as determined by scintilla- effects (data not shown). The majority of HIV p24 production tion counting of harvested cells. As the extent of viral was abolished at relatively low concentrations of GLQ223 (16- cytopathology by day 4 was too great to allow reliable 63 ng/ml), with essentially complete inhibition seen at higher quantitation of [3H]leucine and [3H]thymidine incorporation concentrations (1-2 ,ug/ml). Greater than 75% inhibition was or meaningful comparison to treated cultures, cells for observed at concentrations that produced little inhibition incorporation studies were harvested on day 3. (<20%) of total cellular protein synthesis or cellular DNA For RNA studies, cells were inoculated at a multiplicity of synthesis in infected cells or parallel cultures ofuninfected VB infection of 0.03 and then cultured in 75-cm2 culture flasks in cells, but inhibition of these parameters was seen at higher the continuous presence of the desired concentration of concentrations with 3 days of continuous exposure to GLQ223 GLQ223. Aliquots of these cultures were harvested at 48 hr, (Tables 1 and 2). the cells were lysed, and RNA was extracted (7) for analysis A similar acute infection experiment was used to assess of viral and cellular RNA by Northern blot (8). Cell-free effects of GLQ223 on HIV replication at the RNA level. Cells culture supernatants were harvested for determination of were harvested for RNA extraction at day 2 of culture to HIV p24 content at the time RNA was prepared. [3H]Leucine prevent the extensive RNA degradation observed at later and [3H]thymidine incorporation were determined for ali- time points in untreated cultures, presumably as a conse- quots of the cultures, plated as above in 96-well plates, pulsed quence of advanced viral cytopathic effects and cell lysis. with [3H]leucine or [3H]thymidine, and harvested at 48 hr. Concentration-dependent inhibition of HIV replication was Macrophage assays: Cells chronically infected in vitro again observed, as measured by supernatant HIV p24 levels with exogenous virus. Monocyte/macrophages were isolated (Fig. 1) and decreased virally induced cytopathology (data from peripheral blood of HIV-seronegative healthy donors, not shown). Essentially complete inhibition of HIV p24 infected in vitro with HIV (HIV-lDV isolate; multiplicity of production was observed at concentrations of GLQ223 that infection of 1.0), and cultured as described (6). HIV antigen did not measurably inhibit total cellular protein synthesis or expression was assessed by flow cytometric analysis. Cells cellular DNA synthesis in infected cells or parallel cultures of were stained with a p24-specific murine monoclonal antibody uninfected VB cells in this 2-day assay (Fig. 1C). GLQ223 after using an isotype-matched monoclonal antibody of treatment also resulted in a selective proportional decrease in irrelevant specificity to set the threshold for background the amount of viral RNA present in treated cells relative to fluorescence. The percentage of cells with HIV-specific total RNA extracted from the cells (Fig. 1 A and B). RNA (above background-fluorescence channel 150) fluorescence from infected untreated cells probed with a 32P-labeled was then determined. Cultures were used for studies de- nick-translated full-length HIV DNA probe derived from scribed below only if the percentage of HIV antigen-express- pHXB-2 (9) showed a strong signal with characteristically ing cells, as determined by flow cytometric analysis, was sized hybridizing bands corresponding to full-length (=9.5 .30%. After establishing the extent of infection, cells were kilobases) and spliced forms [4.3 and =2.0 kilobases (10)] of treated with a 3-hr pulsed exposure to GLQ223. Alternatively, HIV RNA (Fig. LA, lane 1), with the hybridizing species of cells received continuous exposure to 3'-azido-3'-deoxythy- less than 2 kiobases probably representing degraded viral midine (AZT, 40 p.M; Burroughs Wellcome). Treated cells message. No hybridization of the HIV probe to RNA from were maintained in suspension culture in Teflon jars (PTFE; uninfected cells was seen, regardless of GLQ223 treatment Savillex, Minnetonka, MN) as described. After 4 days of (Fig. LA, lanes 3 and 4). GLQ223 treatment resulted in a culture, HIV antigen expression was assessed in GLQ223- decrease in the amount of viral RNA detectable by hybrid- treated cultures and buffer-treated control cultures by flow ization (Fig. LA, compare lanes 1 and 2), despite equivalent cytometric analysis (6). amounts of total RNA visible in each lane in the ethidium Macrophage assays: Cells infected in vivo. Experiments bromide-stained gel prior to transfer (data not shown). In were also conducted with monocyte/macrophages infected contrast, GLQ223 treatment did not affect levels of message in vivo with HIV. Freshly drawn EDTA anti-coagulated detectable by hybridization with a full-length y-actin cDNA whole blood from HIV-seropositive patients was treated for probe from pHF A-1 (11) in either HIV-infected treated cells 3 hr at 37°C with 500 ng of GLQ223 per ml or sham-treated (Fig. 1B, lanes 1 and 2) or uninfected treated cells (Fig. 1B, with buffer. Mononuclear cells were then isolated over lanes 3 and 4). Ficoll/Hypaque, with extensive washing, and placed in culture in Teflon jars. Five days later, HIV antigen expres- Table 1. Inhibition of viral replication by GLQ223 in sion was quantitated by flow cytometric analysis (6). T-lymphoblastoid cells: 96-hr continuous exposure RESULTS GLQ223 treat- GLQ223 was tested for inhibition of HIV replication in a ment, ,ug/ml p24, ng/ml x 10-2 p24, % control variety of assays using both primary and transformed target 0 18.5 ± 1.4 100 ± 8 cells and cells representing both lymphoid and monocytoid 0.016 4.9 ± 1.6 27 ± 9 lineages. The compound was tested in assays of both acute 0.031 3.7 ± 1.4 20 + 8 and chronic HIV infection and in experiments using primary 0.063 2.3 ± 1.2 13 ± 6 cells from several distinct in vivo-infected donors, each 0.126 1.4 ± 0.7 6± 4 presumably infected with a different strain of HIV. 0.251 0.7 ± 1.4 4± 2 Cultures of the highly susceptible T-lymphoblastoid cell line 0.502 0.3 ± 1.2 2_ 1 VB were inoculated with HIV and then treated with GLQ223. 1.005 <0.3 0 Infection and viral replication were monitored in the cultures 2.010 <0.3 0 by following the appearance of characteristic HIV-induced HIV p24 production is shown relative to untreated infected control cytopathic changes, by measuring HIV p24 in culture super- cultures as a function of GLQ223 concentration for acutely infected natants, and by assessing production of viral antigens by cells of the T-cell line VB. Cells were continuously exposed to immunofluorescence analysis. As shown in Table 1, in mul- GLQ223 for 96 hr. p24 values are means ± SD for eight experiments tiple experiments, GLQ223 treatment reproducibly resulted in at each concentration.
  • 3. 2846 Medical Sciences: McGrath et al. Proc. Natl. Acad. Sci. USA 86 (1989) Table 2. Effect of GLQ223 on [3H]thymidine and [3H]leucine A B incorporation in uninfected and acutely HIV-infected VB cultures: kb kb 72-hr continuous exposure 9.5 - 0 9.5 - 7.5 7.5 - GLQ223 treatment, [3H]Thymidine, [3H]Leucine, 4.4 - 4.4 - /Lg/ml % control % control to 2.4 rn Uninfected cells 1.4 - 1.4- - 0 100 ± 4 100 ± 9 0.012 87 ± 3 85 ± 8 0.24- 1 0.24- 0.025 91 ± 3 80± 8 0.049 86 ± 4 83 ± 2 0.098 84 ± 3 83 ± 4 0.1% 84 ± 2 90 ± 13 0.393 80 ± 5 71 ± 4 1 2 3 4 1 2 3 4 0.785 73 ± 2 61 ± 5 1.570 60 ± 0 48 ± 4 C Untreated: 106.2 ng/ml 3.140 46 ± 2 33 ± 3 120 / HIV-infected cells 0 78 ± 5 89 ± 6 100 E 0.012 90 ± 3 104 ± 10 80 0- 0.025 89 ± 3 98 ± 7 60 0.049 91 ± 4 99± 5 40 0.098 89 ± 2 98 ± 5 I 20 0.1% 0.393 88 ± 3 87 ± 3 93 ± 9 95 ± 5 v .01 I...1 .1 10 0.785 75 ± 4 62 ± 4 GLQ223 (ig/ml) 1.570 69 ± 3 50 ± 3 3.140 54 ± 2 45 ± 4 FIG. 1. GLQ223 treatment selectively decreases relative levels Cells (2 x 104 per well) were seeded in 96-well plates and then of HIV RNA: 48-hr continuous exposure. (A) HIV probe. No pulsed either with 1 gCi (1 Ci = 37 GBq) of (3H]leucine or hybridization of the HIV probe to RNA from uninfected cells that [3H]thymidine 8 hr prior to harvesting cells at 72 hr of culture. were untreated (lane 3) or treated with GLQ223 at 3.14 Ag/ml (lane Labeled precursor incorporated into trichloroacetic acid-pre- 4) was seen, whereas a strong signal with characteristically sized cipitable cell-associated macromolecules was determined by scintil- bands corresponding to full length and spliced forms of HIV RNA lation counting. Values shown are the means SD for six experi- was seen in untreated infected cells (lane 1). GLQ223 treatment at ments. Results are normalized to counts obtained for untreated 3.14 Ag/ml (lane 2) resulted in a decrease in the amount of viral RNA uninfected control cultures ([3H]thymidine control value = 136,500 detectable by hybridization. (B) y-Actin probe. GLQ223 treatment 5007; [3H]leucine control value = 7897 691). did not affect levels of y-actin message detectable by hybridization in untreated infected cells (lane 1), infected cells treated with GLQ223 treatment of monocyte/macrophages chronically GLQ223 at 3.14 ,ug/ml (lane 2), untreated uninfected cells (lane 3), infected with HIV in vitro also inhibited HIV replication, as and uninfected cells treated with GLQ223 at 3.14 j.g/ml (lane 4). The assessed by flow cytometric analysis 4 days after a short similarity in size of the hybridizing bands for y-actin and spliced HIV message is coincidental and does not represent cross-hybridization; exposure of infected cells to GLQ223. Cells were infected with HIV-lDv and then cultured for 20 days prior to either a despite hybridization of the y-actin probe to RNA from uninfected cells (lanes 3 and 4), there is no hybridization of the HIV probe to the pulsed 3-hr exposure to GLQ223 or initiation of continuous same RNA preparations (lanes 3 and 4 in A). (C) Aliquots of the exposure to AZT (40 AM). After a subsequent 4 days of cultures used for RNA studies were also tested for supernatant p24 culture, flow cytometric analysis was performed. A repre- content at 48 hr. p24 values are single determinations at each sentative experiment is shown in Fig. 2, where untreated concentration; variability of the assay employed is <10% for repli- cultures (Fig. 2A) showed 34% of the cells displaying HIV- cate samples. [3H]Leucine incorporation values were as follows: specific fluorescence (solid lines) above the background uninfected untreated cells, 19,673 3137; uninfected cells (3.14 ± fluorescence threshold set with irrelevant control antibody ,ug/ml GLQ223), 20,285 2019; infected untreated cells, 15,325 ± (dashed lines; background threshold set at fluorescence 1116; infected cells (3.14 /hg/ml GLQ223), 15,220 ± 482. channel 150). No HIV-specific fluorescence was detectable [3H]Thymidine incorporation values were as follows: uninfected untreated cells, 54,840 5763; uninfected cells (3.14 ,ug/ml ± in GLQ223-treated cells (Fig. 2B). In contrast to the complete GLQ223), 60,600 2589; infected untreated cells 32,889 2995; ± ± abrogation of HIV antigen expression seen in GLQ223- infected cells (3.14 ,ug/ml GLQ223), 55,021 2905. (Mean values ± treated cultures, AZT-treated cultures (Fig. 2C) showed 36% SD in cpm; n = 4.) of cells displaying HIV-specific fluorescence. No change in viability was seen in GLQ223-treated cultures at the time of extensive washing, cells were cultured with no further analysis, 4 days after treatment. However, 2 weeks following exposure to GLQ223. HIV antigen expression was quanti- treatment, an =40% loss of cellular viability was noted by tated on day 5 of culture by flow cytometric analysis. As flow cytometry in GLQ223-treated, in vitro-infected cultures shown in Table 3, this single pulsed treatment with GLQ223 when HIV antigen expression remained undetectable (data completely abrogated HIV antigen expression by monocyte/ not shown). Identically treated uninfected cultures studied in inacrophages from five of eight evaluable patients (quantifi- parallel showed no change in cellular viability. able HIV-specific immunofluorescence detectable in parallel Additional studies were conducted using monocyte/mac- untreated cultures) and substantially inhibited expression in rophages isolated from individuals infected with HIV in vivo. the other three cultures. The treatment did not affect cellular In an ex vivo experimental system designed to simulate as viability at day 5 (data not shown). closely as possible the potential in vivo administration of GLQ223, freshly drawn EDTA anticoagulated whole blood DISCUSSION from HIV-infected patients was treated with GLQ223 (500 HIV is the etiologic agent for AIDS and a spectrum of related ng/ml) or sham-treated with buffer for 3 hr at 370C. After disorders. Although a variety of potentially vulnerable steps isolation of mononuclear cells by density centrifugation with in the viral life cycle could theoretically serve as targets for
  • 4. Medical Sciences: McGrath et al. Proc. Natl. Acad. Sci. USA 86 (1989) 2847 200 rA cyte lineage actually appear to be a major in vivo reservoir for HIV (6, 12), as is the case for most other retrolentiviruses (13, 14). Therefore, agents that are not active in mononuclear 100 phagocytes are unlikely to exert significant clinical effects, no matter how promising the activity observed in vitro in T-cell-based assay systems. Furthermore, monocyte/macro- phages appear to be relatively more resistant to the cyto- CIO 1200 400 600 800 pathic effects of HIV infection than T cells, more readily T- 200 a allowing establishment of chronic infection with potential for 0x persistence of an in vivo viral reservoir. The importance of a) chronic infection of cells, especially cells of the mononuclear .0 E 100 phagocytic lineage, in the in vivo pathobiology of HIV z infection may represent an intrinsic limitation to the ultimate clinical efficacy of nucleoside analogs such as AZT. As AZT 0 appears to act by inhibiting reverse transcription, an early 1 200 400 600 800 event in the HIV life cycle, it would be expected to have little 200 C or no effect on chronically infected cells presumably produc- ing viral proteins from integrated proviral DNA. In vitro studies show this to be the case in both macrophages and T 100 cells (Fig. 2; M.S.M., unpublished observations; ref. 15). :.1, GLQ223 inhibits HIV replication in both T cells and monocyte/macrophages and shows activity in in vitro assays 'I -- of both acute and chronic infection. Inhibition of HIV repli- 1 200 400 600 800 cation is manifest as a selective inhibition of levels of viral Fluorescence antigen and viral RNA relative to host cell protein synthesis and total cellular RNA levels in treated cells. GLQ223 shows FIG. 2. Selective inhibition of HIV replication in chronically anti-HIV activity in in vitro assays against both cells infected infected monocyte/macrophage cultures. Flow cytometric analysis in vitro with the virus and against cells infected with HIV in (linear fluorescence scale) for HIV p24 expression in cytoplasm of vivo. Activity of GLQ223 against in vivo-infected cells from infected cells is shown for GLQ223 and AZT-treated monocyte/ different patients (Table 3) demonstrates the in vitro activity of macrophage cultures chronically infected, in vitro, with HIV. Cells the compound against a variety of presumptively distinct virus were either untreated (A), pulse treated with GLQ223 (500 ng/ml for 3 hr followed by washing, without further GLQ223 exposure) (B), or strains. Typically, monocytes freshly isolated from HIV- treated continuously with 40 AM AZT (C). infected patients do not express detectable levels of HIV antigens. Brief in vitro cultivation of these cells allows HIV therapeutic intervention, the in vivo pathobiology of HIV expression in a subpopulation of monocyte-derived cells from infection will probably determine the ultimate efficacy of such donors; expressed viral antigens are readily detectable by potential therapies. CD4-expressing lymphoid cells are sus- flow cytometric analysis (M.S.M., unpublished observations). ceptible to HIV infection, and depletion of CD4-bearing T A pulsed exposure of whole peripheral blood from HIV- lymphocytes is both the hallmark of HIV-associated disease infected patients to GLQ223 resulted in essentially complete and the apparent cause of many of the clinical consequences suppression of viral antigen expression by subsequently cul- of HIV infection. However, cells of the mononuclear phago- tured cells from a majority of donors. No viral antigens were detectable 5 days after a single brief exposure to the com- Table 3. Effect of pulsed GLQ223 treatment of whole blood on pound. In preliminary studies, a single 3-hr pulsed exposure to HIV expression in subsequently cultured monocyte/macrophages GLQ223 inhibits expression of HIV antigens in subsequently from infected donors cultured treated infected cells for up to 4 weeks. As a decrease Integrated total in cellular viability is noted 2-3 weeks after treatment, when % p24-expressing HIV viral antigen expression remains undetectable, but not at cells* fluorescencet earlier time points (4-5 days after treatment) or in identically Patient Control Treated Control Treated % inhibition treated uninfected cultures, GLQ223 appears to exert a selec- tive inhibitory activity that initially suppresses viral replication 1 6.3 0 3502 0 100 and ultimately selectively kills HIV-infected monocyte/ 2 2.1 0.1 1158 71 95 macrophages. Despite the fact that freshly isolated peripheral 3 5.7 0 3156 0 100 blood monocytes from HIV-infected subjects do not typically 4 3.1 0 1813 0 100 express viral antigens, treatment of freshly drawn whole blood 5 3.0 0 1797 0 100 from such subjects with GLQ223 prevented expression of HIV 6 5.3 0.9 3332 613 82 antigens by subsequently isolated and cultured cells, suggest- 7 7.1 0 4371 0 100 ing that the compound may be capable of acting on at least 8 4.4 0.8 3158 584 83 some infected cell populations that do not express detectable The results shown are for representative experiments involving viral antigens. Based on its potent anti-HIV activity in both T eight different donors. Under the culture conditions employed (no cells and monocyte/macrophages, as well as its activity in exposure of cells to mitogens or other activation stimuli), p24- assays of both acute and chronic infection, GLQ223 is an expressing cells were observed only among cells in the macrophage, interesting agent for evaluation for anti-HIV activity in vivo. but not lymphocyte, gating region. Animal testing will permit assessment of any toxicities asso- *Percent p24-expressing cells indicates the percentage of cells in ciated with the compound, including any toxic effects related cultures expressing p24-specific (above background-fluorescence channel 150) fluorescence on day 5 of culture. to immunogenicity. Although essentially complete inhibition tIntegrated total HIV fluorescence refers to the integrated area ofthe of viral antigen expression was observed, in vitro, at concen- fluorescence histogram obtained with the p24-specific antibody, trations of GLQ223 that did not appear to inhibit incorporation above fluorescence channel 150, and reflects both the percentage of of [3H]leucine or [3H]thymidine, higher concentrations of the p24-expressing cells and the intensity of fluorescence seen for these compound did inhibit [3H]leucine and [3H]thymidine incorpo- cells. ration, especially when present continuously in culture for
  • 5. 2848 Medical Sciences: McGrath et al. Proc. Natl. Acad. Sci. USA 86 (1989) several days. However, relatively brief, pulsed exposure to We thank Dr. Leonard A. Herzenberg (Stanford University) for the compound appears to be sufficient for anti-HIV effects and thoughtful review of the manuscript, Drs. Gregory Reyes and seems to minimize any inhibitory effects on noninfected cells, Michael Piatak (Genelabs) for helpful discussions, Dr. Holly Abrams in vitro, a finding of potential relevance to minimizing any (Genelabs) for providing virus stocks used in some studies, and the observed in vivo toxicities. Genelabs Process Development Group for providing purified GLQ223. The mechanism of action whereby GLQ223 inhibits HIV replication is currently unknown. Although the compound 1. Yeung, H. W., Wong, D. M., Ng, T. B. & Li, W. W. (1985) belongs to the family of single-chain ribosome-inactivating Int. J. Pept. Protein Res. 27, 325-333. proteins, it has not been established whether the anti-HIV 2. Zhang, X. & Wang, J. (1986) Nature (London) 321, 477-478. activity is directly related to this property. Two general 3. Cheng, K. F. (1982) Obstet. Gynecol. 59, 494-498. classes of mechanisms can be envisioned. In the first, the 4. Chan, Y., Tong, M. K. & Lau, M. J. (1982) Shanghai J. Chin. selective antiviral activity of the compound may be due to Med. 3, 30-31 (in Chinese). selective binding or uptake by virally infected cells (16, 17). 5. Lifson, J. D., Reyes, G. R., McGrath, M. S., Stein, B. S. & Once inside the infected cells, the compound may exert Engleman, E. G. (1986) Science 232, 1123-1127. 6. Crowe, S., Mills, J. & McGrath, M. (1987) AIDS Res. Hum. nonspecific effects through catalytic inactivation of ribo- Retrov. 3, 135-145. somes (18, 19). In the second class of mechanism, selective 7. Chomczynski, P. & Sacchi, N. (1987) Anal. Biochem. 162,156- binding or uptake by infected cells may or may not occur, but 159. the anti-HIV activity of the compound may be attributed to 8. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular differential effects on viral as opposed to host cell nucleic Cloning: A Laboratory Manual (Cold Spring Harbor Lab., Cold acid or protein synthesis, processing, or stability. Treatment Spring Harbor, NY), pp. 202-203. of acutely infected T-lymphoblastoid cells with GLQ223 9. Fisher, A. G., Collati, E., Ratner, L., Gallo, R. C. & Wong- Staal, F. (1985) Nature (London) 316, 262-265. resulted in decreased levels of viral proteins at concentra- 10. Muesing, M. A., Smith, D. H., Cabridilla, C. D., Benton, tions of GLQ223 that did not affect cellular protein synthesis C. V., Lasky, L. A. & Capon, D. J. (1985) Nature (London) (Fig. 1C), at a time when immunofluorescence analysis 313, 450-457. showed -75% of untreated infected cells expressing viral 11. Gunning, P., Ponte, P., Okayama, H., Engel, J., Blau, H. & antigens (48 hr after infection; immunofluorescence data not Kedes, L. (1983) Mol. Cell. Biol. 3, 787-795. shown). As selective binding or uptake by infected cells 12. Gartner, S., Markovits, P., Markovits, D. M., Kaplan, M. H., followed by nonspecific ribosomal inactivation might be Gallo, R. C. & Popovic, M. (1986) Science 233, 215-219. 13. Haase, A. T. (1986) Nature. 322, 130-136. expected to result in a measurable inhibition of total cellular 14. Narayan, O., Kennedy-Stoskopf, S. & Zink, M. C. (1988) Ann. protein synthesis under these conditions, this observation Neurol. (Suppl.) 23, S95-S100. argues against a mechanism based on nonspecific inhibition 15. Perno, C.-F., Yarchoan, R., Cooney, D. A., Hartman, N. R., of protein synthesis through ribosomal inactivation. Treated Gartner, S., Popovic, M., Hao, Z., Gerrard, T. L., Wilson, cells showed a selective proportional decrease in the amount Y. A., Johns, D. G. & Broder, S. (1988) J. Exp. Med. 168, of viral RNA relative to total cellular RNA at 48 hr of culture 1111-1125. (when -75% of untreated cells were productively infected; 16. Fernandez-Puentes, C. & Carrasco, L. (1980) Cell 20, 769-775. data not shown), with no effect on levels of RNA encoding 17. Foa-Tomasi, L., Campadelli-Fiume, G., Barbieri, L. & Stirpe, F. (1982) Arch. Virol. 74, 322-332. the cellular gene for -actin, suggesting a possible selective 18. Endo, Y. & Tsurugi, K. (1987) J. Biol. Chem. 262, 8128-8130. effect of GLQ223 on viral nucleic acid synthesis, processing, 19. Stirpe, F., Bailey, S., Miller, S. P. & Bodley, J. W. (1988) or stability. Additional studies are needed to clarify the Nucleic Acids Res. 16, 1349-1357. mechanism that underlies the observed selective anti-HIV 20. Wang, Y., Qian, R. Q., Gu, Z. W., Jin, S. W., Zhang, L. Q., activity of GLQ223 and other related compounds that show Xia, Z. X., Tian, G. Y. & Ni, C. Z. (1986) Pure Appl. Chem. similar activity in vitro. 58, 789-798.
  • 6. Coregulation of NADPH Oxidase Activation and Phosphorylation of a 48-kD Protein(s) by a Cytosolic Factor Defective in Autosomal Recessive Chronic Granulomatous Disease Susan E. Caldwell, Charles E. McCall, Cynthia L. Hendricks, Peter A. Leone, David A. Bass, and Unda C. McPhail Departments ofBiochemistry and Medicine, Bowman Gray School ofMedicine of Wake Forest University, Winston-Salem, North Carolina 27103 Abstract the generation of toxic oxygen radicals by the PMN. The en- zyme in PMN that triggers oxygen radical production is an The mechanisms regulating activation of the respiratory burst activatable NADPH oxidase, whose mechanism of activation enzyme, NADPH oxidase, of human neutrophils (PMN) are is still unknown. The active enzyme appears to be a multicom- not yet understood, but protein phosphorylation may play a ponent complex, most likely consisting of a flavoprotein and a role. We have utilized a defect in a cytosolic factor required for unique low potential cytochrome, cytochrome b559 (2). NADPH oxidase activation observed in two patients with the Until recently, activation of NADPH oxidase could only autosomal recessive form of chronic granulomatous disease be achieved by stimulation of intact PMN with a wide variety (CGD) to examine the role of protein phosphorylation in acti- of agents, including opsonized particles, phorbol esters, che- vation of NADPH oxidase in a cell-free system. motactic peptides, and certain detergents (2). One characteris- NADPH oxidase could be activated by SDS in reconstitu- tic these stimuli share is the ability to trigger activation of tion mixtures of cytosolic and membrane subcellular fractions protein kinases inside the cell resulting in the phosphorylation from normal PMN, and SDS also enhanced phosphorylation of endogenous proteins (3-8). Protein kinase C (PKC)' has of at least 16 cytosolic and 14 membrane-associated proteins. been of particular interest as an intermediate. Direct activators However, subcellular fractions from CGD PMN plus SDS ex- of PKC, such as phorbol esters and cell-permeable diacylglyc- pressed little NADPH oxidase activity, and phosphorylation erols (9), trigger oxidase activation in the intact cell by an of a 48-kD protein(s) was selectively defective. The membrane apparent PKC-dependent mechanism (10-13). In addition, fraction from CGD cells could be activated for NADPH oxi- stimuli, such as concanavalin A and chemotactic peptides, dase when mixed with normal cytosol and phosphorylation of induce the release of endogenous stores of diacylglycerol from the 48-kD protein(s) was restored. In contrast, the membrane membrane phospholipids and thus could activate PKC (14). fraction from normal cells expressed almost no NADPH oxi- Further evidence supporting a role for protein phosphory- dase activity when mixed with CGD cytosol, and phosphoryla- lation in the activation of NADPH oxidase is given by recent tion of the 48-kD protein(s) was again markedly decreased. reports of a phosphorylation defect in the PMN from several Protein kinase C (PKC) activity in PMN from the two patients patients with chronic granulomatous disease (CGD), observed appeared to be normal, suggesting that a deficiency of PKC is by stimulation of the intact cell (15-17). CGD is an inherited not the cause of the defective 48-kD protein phosphorylation disease characterized by increased susceptibility to bacterial and that the cytosolic factor is not PKC. infection, in which the only known functional defect is the These results demonstrate that the cytosolic factor re- inability oftheir phagocytic cells to produce oxygen radicals in quired for activation of NADPH oxidase also regulates phos- response to cell stimulation. Two modes of inheritance of the phorylation of a specific protein, or family of proteins, at 48 disease are observed, X-linked recessive and autosomal reces- kD. Although the nature of this protein(s) is still unknown, it sive (18, 19). Recently, the gene defect in the X-linked form may be related to the functional and phosphorylation defects has been described (20) and the gene product appears to be a present in CGD PMN and to the activation of NADPH oxi- subunit of the cytochrome b559 (21, 22). PMN from most pa- dase in the cell-free system. tients with this form of the disease lack the characteristic spec- trum of the cytochrome (23). The gene defect in the autosomal Introduction recessive form has yet to be identified. However, the phosphor- ylation defect in CGD has been attributed solely to the auto- Human neutrophils (PMN) play a critical role in host defense somal recessive form (15), although, this has been disputed against microorganisms and in the inflammatory response (1). (16, 24). One of the primary mechanisms for effecting these reactions is Recently, a cell-free system for activation of NADPH oxi- dase has been developed by ourselves (25) and others (26-28). Address reprint requests to Dr. McPhail, Department of Biochemistry, Activation requires the apparent interaction of cytosolic and Bowman Gray School of Medicine of Wake Forest University, 300 S. membrane cellular cofactors and can be achieved by the addi- Hawthorne Rd., Winston-Salem, NC 27103. tion of an amphiphile, either arachidonic acid or SDS (29), to Published in abstract form in 1987. (Clin. Res. 35:655a.) the isolated cell fractions. The identities of the cytosolic and Received for publication 9 January 1987 and in revised form 5 membrane cofactors are unknown, although at least one nec- October 1987. J. Clin. Invest. 1. Abbreviations used in this paper: CB, cytochalasin B; CGD, chronic X The American Society for Clinical Investigation, Inc. granulomatous disease; DFP, diisopropylfluorophosphate; EP, extract- 0021-9738/88/05/1485/12 $2.00 able particulate fraction; NEP, nonextractable particulate fraction; Volume 8 1, May 1988, 1485-1496 PKC, protein kinase C; PMA, phorbol myristate acetate. Cell-free Protein Phosphorylation and Neutrophil Oxidase Activation 1485
  • 7. essary membrane component is cytochrome b559 (30). In addi- Subcellularfractionation. Purified PMN were suspended at 5 X 107 tion, a recent abstract reports that the cytosolic factor activity PMN/ml in Hanks' balanced salt solution containing 4.2 mM sodium is absent in a CGD patient with the autosomal recessive bicarbonate and 10 mM Hepes, pH 7.4 (HBSS). To inhibit proteolysis form (31). (38), cells were treated with 2 mM diisopropylfluorophosphate (DFP) It is not yet clear if a protein kinase is involved in oxidase for 5 min at 4VC, and then washed with a 15-fold volume of ice-cold HBSS. PMN were resuspended at 1 X 108/ml in a sonication buffer activation in this system. Arachidonate can activate PKC (32), containing 0.34 M (11 %, wt/vol) sucrose, 10 mM Hepes, 1 mM EGTA, but inhibition or depletion of PKC does not block arachidon- 1 mM NaN3, 130 mM NaCl, 100 mM NaF, and 10 mM sodium ate-mediated activation of NADPH oxidase in the cell-free pyrophosphate. In some experiments, in which NADPH oxidase activ- system (30, 33, 34). Contradictory reports as to the require- ity only was measured, the DFP treatment was omitted, 0.5 mM PMSF ment for ATP in the system have appeared (30, 35). Finally, it was included, and NaF and pyrophosphate were omitted. These alter- has not yet been determined if arachidonate or SDS can in- ations had no effect on the final oxidase activity obtained. l-ml ali- duce phosphorylation of endogenous protein substrates in quots of cells were sonicated in 5-s bursts in a melting ice bath with a conjunction with activation of NADPH oxidase in the cell-free setting of 1.5 using a sonicator (model W-220; Heat Systems-Ultra- system. sonics, Inc., Farmingdale, NY) equipped with a stepped microtip. Cell In this report, the role of protein phosphorylation in the breakage of 90% was achieved as monitored by phase microscopy. - Sonicates were centrifuged at 800 g for 10 min at 40C to pellet activation of NADPH oxidase in the cell-free system is exam- unbroken cells and nuclei. Supernatants were fractionated by discon- ined. Using both normal PMN and PMN lacking cytosolic tinuous sucrose density gradient centrifugation. In the first experi- factor activity from patients with the autosomal recessive form ment, supernatants in the 11% sucrose buffer were layered over 40% of CGD, we have compared the ability of SDS to induce oxi- (wt/vol) sucrose in a ratio of 2:1 (vol/vol), as described by Gabig et al. dase activation and phosphorylation of endogenous proteins (39). In subsequent experiments, a 15% (wt/vol) sucrose interface layer in cell-free mixtures of cytosolic and membrane fractions. Re- was included to more easily separate membrane and cytosolic fractions sults suggest that the cytosolic factor required for NADPH (40), the final ratio being 2: 1:1 (vol/vol) of 11, 15, and 40%, respec- oxidase activation in the cell-free system is also specifically tively. Gradients were centrifuged at 4°C (SW 60.1 rotor; Beckman required for the SDS-dependent phosphorylation of a 48-kD Instruments, Inc., Palo Alto, CA) at 150,000 g for 30 min (39). After protein, or family of proteins. centrifugation, the top layer of each discontinuous gradient (crude cytosol) was collected and centrifuged again at 150,000 g in a Beckman Type 50 rotor for 60 min at 4°C to pellet any contaminating particu- Methods late material (cleared cytosol). The white band at the 1 1%/40% or the 15%/40% interface from each discontinuous gradient and most of the Isolation of cells. PMN (> 95% purity) were purified from human 40% sucrose layer was collected (membrane fraction). The discontin- peripheral blood by dextran sedimentation, Ficoll-Hypaque centrifu- uous gradients also yielded granule fractions as pellets, which were gation, and hypotonic lysis as described (36). Two patients with auto- resuspended in sonication buffer to the original volume layered onto somal recessive CGD were studied, one of which (S.S.) is a 27-yr-old the gradient. All fractions were either used immediately or stored male reported previously (37). The other (S.H.) is a 25-yr-old female at -700C. patient of Dr. Michael Cohen (University of North Carolina School of Gradient fractions were analyzed for purity by measuring the dis- Medicine, Chapel Hill, NC) and was diagnosed based on clinical his- tribution of the following marker enzymes: lactic dehydrogenase tory and neutrophil oxidative functional studies (defective in vitro (LDH; 41) for cytosol, alkaline phosphatase (42) for plasma mem- bactericidal activity and the absence of a luminol-dependent chemilu- brane, vitamin B12-binding protein (43) for specific granules, lysozyme minescence response to phorbol myristate acetate, PMA, or opsonized (44) for both specific and azurophil granules, and myeloperoxidase zymosan). (MPO; 45) for azurophil granules. Table I summarizes the distribution Table L Subcellular Marker Enzyme Distribution in Fractions Isolated by Discontinuous Sucrose Gradient Fractionation % Distribution of enzyme* Alkaline B12-binding Fraction assayedt LDH phosphatase protein Lysozyme MPO 11%/40% gradient Cytosolic 95.9 0.9 ND 0 7.1 Membrane 2.5 93.8 ND 8.7 3.7 Granule 2.0 5.3 ND 91.3 89.2 1 1 %/ 15%/40% gradient Cytosolic 86.7 1.1 2.4 5.3 2.9 15% layer 8.3 7.3 0.4 0.1 0.4 Membrane 1.4 83.2 12.1 8.2 3.7 Granule 0.6 5.6 85.1 86.3 92.7 * Marker enzyme levels were determined in subcellular fractions as described in Methods. Numbers given are the percentage of total activity re- covered in each fraction for one experiment with the 11 %/40% gradient and the mean of two experiments with the 11 %/ 1 5%/40% gradient. In the 1I%/1 5%/40% gradient, total recovery for each marker (mean of two experiments) was: LDH, 85.7%; alkaline phosphatase, 97.4%; vitamin B12-binding protein, 79.9%; lysozyme, 102.9%; MPO, 68.2%; and protein (46), 87.3%. Distribution of protein in the 11%/40% gradient was: cy- tosolic, 50.4%; membrane, 7.5%; and granule, 42.2%. Protein distribution in the 1I%/l15%/40% gradient was: cytosolic, 43.8%; 15% layer, 5.3%; membrane, 7.8%; and granule, 41.4%. * Normal PMN were sonicated and fractionated by either an 1l%/40% or an 1l%/15%/40% discontin- uous sucrose gradient as described in Methods. Fractions were assayed immediately for LDH and stored at -70'C before the remainder of the marker enzyme assays. 1486 Caldwell, McCall, Hendricks, Leone, Bass, and McPhail
  • 8. of these markers. In one experiment with the 1 1%/40% gradient, the rotor. Supernatants (cytosolic fractions) were reserved and pellets cleared cytosol fraction contained 95.9% of the recovered LDH, 0.9% (membrane fractions) were resuspended by sonication in a small vol- of alkaline phosphatase, and a mean of 3.5% of lysozyme and MPO. ume (< 1 ml) of sonication buffer. Part of each sample (200-300 Ml) The membrane fraction contained 93.8% of the recovered alkaline was solubilized in electrophoresis sample buffer (50), boiled for 2 min phosphatase, 2.5% of LDH, 8.7% of lysozyme, and 3.7% of MPO. In in a water bath, and stored for no more than 3 d at -70'C before the 1 1%/1 5%/40% gradient procedure, the cleared cytosol contained analysis by SDS-polyacrylamide gel electrophoresis, as described 86.7% of the recovered LDH, 1.1% of alkaline phosphatase, 2.4% of below. The remainder of each sample was frozen at -70'C without B12-binding protein, 5.3% of lysozyme, and 2.0% of MPO. The mem- solubilization; these nondenatured fractions were assayed for protein brane fraction contained 83.2% of the recovered alkaline phosphatase, by Peterson's modification (51) of the method of Lowry et al. (46). 1.4% of LDH, 12.1% of B,2-binding protein, 8.2% of lysozyme, and Endogenous protein phosphorylation in intact PMN. Cells at 1 3.7% of MPO. The granule fractions from both gradient procedures X 108/ml from a normal donor and a CGD patient (S.S.) were sus- contained > 85% of the recovered granule enzyme markers. The dis- pended in buffer (4) containing 6 mM Hepes/Tris (pH 7.4), 0.15 M continuous sucrose gradient fractionation procedure, therefore, was NaCl, 10 mM glucose, 5 mM KCI, 1 mM MgCI2, 0.25 mM CaCI2, and considered to yield cytosolic and membrane fractions of acceptable 32p; (1 mCi/1-2 X 0I cells; New England Nuclear). PMN were incu- purity for use in the cell-free oxidase and phosphorylation systems. bated at 370C for 30 min, washed, and resuspended at 5 X 107/ml in NADPH oxidase assay in the cell-free system. Assays measured the the Hepes/Tris buffer without 32p;. Cells were prewarmed for 5 min at SOD-inhibitable reduction of cytochrome c, and were performed as 370C in the presence of 10-1 M cytochalasin B (CB), then treated with previously described (25, 36) with slight modifications. The assay either dimethylsulfoxide (DMSO) or 100 ng/ml PMA for 30 s. Reac- mixture consisted of 0.048 mM potassium phosphate (pH 7.0), 0.076 tions were terminated by addition of a 15-fold excess volume of ice- mM cytochrome c (type III), 1.0 mM EGTA, 10.0 MM flavin adenine cold incubation buffer containing 1.0 M NaFand 10 mM EDTA. Cells nucleotide (all from Sigma Chemical Co., St. Louis, MO), 7.5 mM were washed and then treated with DFP and sonicated as described MgCl2, and either a 2:1 or 4:1 (vol/vol) ratio of cytosol to membrane above. Sonicates were centrifuged at 500 g for 10 min at 4°C, and fraction. Cell equivalence in the assay was 1.5 X 107/ml of cytosol and postnuclear supernatants were centrifuged at 100,000 g for 1 h at 4°C either 0.75 or 1.5 X 10'/ml of membrane fraction. The final assay in a fixed angle rotor. Supernatants (cytosolic fractions) were removed volume was 0.105 ml. Activation of NADPH oxidase was achieved by and stored at -70°C: one aliquot was solubilized in electrophoresis addition of 0 (H20 control) to 175 MM SDS (Bio-Rad Laboratories, sample buffer as described above and the remainder analyzed for pro- Richmond, CA) and incubation at room temperature for 4 min. The tein content (51). concentration of SDS yielding optimal activation of NADPH oxidase SDS-polyacrylamide gel electrophoresis and autoradiography. Re- in reconstitution mixtures of control fractions varied from 120 to 175 duced, denatured samples (30-100 Mg protein/lane) were electropho- MM and was determined in each experiment. Equal aliquots of assay resed on 8-15% gradient polyacrylamide slab gels (1.5 X 140 X 320 mixture were placed into two cuvettes: the reference cuvette, contain- mm) using the discontinuous buffer system of Laemmli (50). Except ing 48 Mg/ml SOD (Diagnostic Data, Inc., Mountain View, CA), and where noted, within each gel, equal amounts of protein were loaded the sample cuvette, containing H20. NADPH oxidase activity was onto gel lanes. After electrophoresis, gels were silver stained by the measured by addition of 0.19 mM NADPH to both cuvettes, and the method of Wray et al. (52), photographed, and then dried between two change in absorbance at 550 nm recorded by a dual beam spectropho- sheets of cellophane using a slab gel dryer (model 443; Bio-Rad). Each tometer (Cary 2390; Varian Instruments, Sunnyvale, CA). Initial rates dried gel was exposed, typically for 1-3 d, to preflashed Kodak X- were used for calculations and activity was expressed as nmol/min/mg Omat x-ray film with a DuPont Cronex Lightning Plus intensifying membrane protein (46) using an extinction coefficient of 21 mM-' screen at -70°C (53) to generate autoradiograms of the radioactive cm-' for cytochrome c (47). Assays were performed on both fresh and phosphoproteins. Autoradiograms were scanned with a laser densi- frozen cell fractions with no differences in results obtained. tometer (UltroScan XL) and scans were analyzed by the software Assay for cytochrome b559. The method used was a modification of package (GelScan XL; LKB-Produkter AB, Bromma, Sweden). published procedures (48, 49).PMN from normal donors and the PKC assay. PKC activity and distribution in normal PMN and autosomal recessive CGD patients were sonicated and fractionated by PMN from the patients with CGD were determined in crude subcellu- discontinuous sucrose density gradient centrifugation with sodium py- lar fractions as previously described (10). Briefly, PMN at 5 X 107/ml rophosphate and NaF omitted from, and PMSF included in, the soni- in HBSS were exposed, in the presence of I0-' M CB at 370C, to either cation buffer. The membrane and granule fractions were analyzed for solvent alone (DMSO), 10-6 M fMLP, or 100 ng/ml PMA for 30 s and cytochrome b559 content by difference spectroscopy after dithionite reactions were terminated by addition of a 10-fold excess of ice-cold titration of an aerobic sample. Fractions were diluted 1:1 with 2X HBSS. After centrifugation, PMN were resuspended in an extraction buffer, yielding final concentrations of 0.25 M potassium phosphate buffer containing 50 mM Tris-HCI (pH 7.5), 50 mM 2-mercaptoeth- (pH 7.5), 0.12 mM sodium deoxycholate, 5.0 mM NaN3, and 0.5 mM anol, 1 mM PMSF, and 2 mM EGTA. PMN were sonicated as de- H202. The reduced minus oxidized spectrum of each diluted fraction scribed above, centrifuged at 500 g, and postnuclear supernatants were was determined on a Cary 2390 spectrophotometer after complete centrifuged at 150,000 g for 90 min. Supernatants (cytosolic fractions) reduction of the sample cuvette by 6-10 mM anaerobically prepared were collected; pellets were resuspended by sonication in extraction dithionite. Amounts of cytochrome b559 were expressed as pmol/108 buffer containing 1.0% Triton X-100 (Sigma Chemical Co.) and cen- cell equivalents using an extinction coefficient of 106 mM-' cm-' at trifuged at 150,000 g for 90 min. Supernatants, the extractable particu- the Soret peak (49). late fractions (EPs), were saved and pellets, the non-extractable partic- Cell-free protein phosphorylation. Cell-free mixtures for phosphor- ulate fractions (NEPs), were resuspended in extraction buffer contain- ylation studies consisted of 1.4 mM MgCl2, 6.0 X 10' cell equivalents/ ing 0.1% Triton X-100 and saved. Fractions were stored at 0°C and ml each of cytosolic fraction and membrane fraction, and 51 MCi/ml assayed within 2 d of preparation. [.y-32P]ATP (12.88 Ci/mmol; New England Nuclear, Boston, MA). Protein kinase assay mixtures (10) contained, in a final volume of SDS at a final concentration of 460 MM or an equal volume of distilled 0.25 ml: 35 mM Tris-HCl (pH 7.5), 10 mM MgCl2, 160 Mg/ml histone water was added and the samples were incubated 5 min at room tem- type HI (type III, Sigma), 50 MM ATP with 2 X 106 cpm [y-32P]ATP, perature. The higher concentration of SDS was used to keep the ratio 0.4 mM EGTA, 0.2 mM PMSF, 10 mM 2-mercaptoethanol, 0.01% of SDS to cellular material similar to that used in the procedure for Triton X-100, and 50 Ml diluted subcellular fraction. Cytosolic frac- activation of NADPH oxidase. Reactions were stopped by adding a tions were diluted 10-fold, EP fractions 20-fold, and NEP fractions two-fold volume of ice-cold sonication buffer, which also reduced the 2-fold. Assays were performed in the presence or absence of activators viscosity of the mixtures for subsequent centrifugation. Samples were (0.6 mM added CaC12, 20 Mg/ml phosphatidylserine, and 2 ,ug/ml then centrifuged at 150,000 g for 12 h at 40C in a Beckman Type 50 diolein) at 30°C for either 5 min (cytosolic fractions), 60 min (EP Cell-free Protein Phosphorylation and Neutrophil Oxidase Activation 1487
  • 9. fractions), or 30 min (NEP fractions). Reactions were stopped with normal, n = 4). In one experiment (not shown), an excess of 20% (wt/vol) TCA, precipitates collected and counted, and activity was CGD cytosol (5.6 X 107 cell equivalents/ml versus the usual expressed as pmol 32p incorporated/min per 107 cell equivalents. Pre- 1.5 X 107/ml) was mixed with normal membrane fraction to vious studies in normal PMN (10, 54) had demonstrated that protein test the possibility of a quantitative defect. The markedly de- kinase activity in cytosolic and EP fractions was completely dependent creased level of activation of NADPH oxidase was not en- on the presence of phosphatidylserine, with no activity above back- hanced, suggesting that the cytosolic factor is not present in ground in the presence of calcium alone. Activity in NEP fractions showed no dependence on added Ca2l and lipids (10). decreased amounts, at least to the level of sensitivity of our detection system. Results To further exclude an oxidase defect in the CGD mem- brane, membrane and granule fractions from both patients Activation of NADPH oxidase in a cell-free system from nor- were examined for their content of cytochrome b559. No cy- mal PMN and PMN from patients with autosomal recessive tochrome b was detected in cytosolic fractions from normal CGD. Activation of NADPH oxidase in a cell-free system was PMN. The level of cytochrome b in normal PMN was 495±80 measured using cell fractions from normal PMN and PMN pmol/ 108 cells (mean±SD, n = 7) and the cytochrome was from two patients with the autosomal recessive form of CGD. distributed such that 32% was present in the membrane frac- As shown in Table II, substantial activation of NADPH oxidase tion and 68% in the granule fraction. Both patients had normal by SDS occurred using reconstitution mixtures of normal (S.H.) or slightly elevated (S.S.) levels of cytochrome b (S.H., membrane and cytosolic fractions. In contrast, mixtures of 492 pmol/108 cells; S.S., 757±3 pmol/108 cells, mean±SD, CGD membrane and cytosolic fractions had very low levels of n = 2) and a similar distribution as a paired normal control. NADPH oxidase activity in the presence of SDS. If corrected Taken together, these results indicate that the PMN of the two for activity observed in the absence of SDS in reconstitution patients with autosomal recessive CGD do not have a defect in mixtures of normal cell fractions (2.5±0.7 nmol O/min per cytochrome b559 and, instead, are defective in the cytosolic mg, mean±SEM, n = 6), oxidase activity induced in CGD factor required for the activation of NADPH oxidase in the reconstituted fractions was only 2.1% of the normal control. cell-free system. This defect in the ability to activate NADPH oxidase appeared Protein phosphorylation in the cell-free system for NADPH to reside in the cytosolic fractions from the patients with CGD, oxidase activation. The cell-free system for activation of since (a) reconstitution mixtures of CGD cytosol with normal NADPH oxidase was examined for protein phosphorylation membrane supported very little activation of the oxidase by changes that might correlate with oxidase activation. Recon- SDS, and (b) the CGD membrane fraction in combination stitution mixtures of membrane and cytosolic fractions iso- with normal cytosol did allow normal levels of NADPH oxi- lated from normal PMN and the PMN of the two patients with dase activation to occur. The membrane fractions of the two autosomal recessive CGD were incubated with or without SDS patients showed equal ability to support NADPH oxidase ac- in the presence of [y-32P]ATP and then subjected to ultracen- tivation in the presence of normal cytosol (S.H., 84% of activ- trifugation to allow separate analysis of phosphoproteins in the ity seen with paired normal membrane, n = 2; S.S., 106% of two fractions. Representative results obtained with reseparated cytosolic fractions are shown in Fig. 1 and with reseparated membrane fractions in Fig. 2. The left panel of each figure shows the silver-stained protein pattern of the gel and the right Table II. Deficient Cytosolic Factor for Activation of NADPH panel shows the autoradiogram obtained from the same gel. Oxidase in Two Patients with Autosomal Recessive CGD No apparent differences in the pattern of silver-stained pro- Reconstitution mixture* NADPH oxidase activity$ teins was observed, comparing fractions obtained from the normal donor with those from the CGD patient. nmol 0j/min/mg As many as 24 phosphoproteins were detected in cytosolic membrane protein fractions following reconstitution of normal fractions in the Normal cytosol + normal membrane 159.0±52.7 (6) presence of SDS (Fig. 1 B, lane 1), of which 16 were phos- CGD cytosol + CGD membrane 5.8±0.4 (3) phorylated more intensely and 3 were dephosphorylated by Normal cytosol + CGD membrane 153.4±62.4 (6) the SDS treatment (compare to lane 2). An increase in phos- CGD cytosol + normal membrane 8.2±0.8 (4) phorylation of a 48,000-D protein(s) (arrow) was observed in cytosolic fractions from the normal reconstitution mixture by * Cytosolic and membrane fractions were isolated from two patients SDS treatment (lanes I and 2), but not in cytosolic fractions (S.S. and S.H.) with the autosomal recessive form of chronic granulo- from the CGD reconstitution mixture (lanes 3 and 4). This matous disease and from six normal donors as described in Methods. was the only obvious phosphorylation defect observed in the One fractionation with CGD (S.S) and one with a normal donor uti- cytosolic fraction from the CGD cell-free reconstitution. lized the 1 1%/40% (wt/vol) discontinuous sucrose gradient. The re- Crossover mixing of normal cytosol with CGD membrane maining fractionations (one with S.S., one with S.H., and five nor- fraction (lanes 5 and 6) restored the SDS-dependent phosphor- mal donors) utilized the 1I%/15%/40% (wt/vol) gradient separation. ylation of the 48-kD protein(s), suggesting that any CGD * NADPH oxidase activation was achieved by an optimal concentra- tion of SDS (120-175 MM) and activity was measured as described in membrane components participating in the phosphorylation Methods. Data are given as the mean±SEM with the number of ex- reaction are functional. In contrast, when the CGD cytosolic periments performed, each done in triplicate on a separate day, in fraction was mixed with normal membrane fraction (lanes 7 parentheses. At n = 3, two experiments were with S.S. fractions and and 8), SDS induced no increase in background phosphoryla- one was with S.H. fractions. In all other instances involving CGD tion in the 48-50-kD region of the gel. These results indicate fractions, two experiments were performed with S.H. fractions and that a cytosolic factor necessary for phosphorylation of a the remainder were with S.S. fractions. 48-kD phosphoprotein(s) is either defective or absent in the 1488 Caldwell, McCall, Hendricks, Leone, Bass, and McPhail
  • 10. A. Cytosolic Fractions-Silver Stain - I U a qw %, I 9 Al *i: _. I 1. .i. I I1 -~ ~ . U .I U -200 -116 -92.5 -66 -45 B. Cytosolic Fractions-Autoradiogram Ia 4 9 4.t i i - 200 - 116 - 92.5 -66 4 ii s^ i^L ^ ^ -~45 1 4 * in WY Ii i ~~~~~~~~j~31 a! S 4 I __ _ _ b-21, ~~~~~-1 .5 1*tj * mm":t* *511 -'Wi -~ ~ ~ ~21.5 * , <IP -4 lii ±n t J Irn.W, v C ,Qcz2+ / ±z1. CtrI.-C CGD---C 0t - CGD-C Ctrl.-C CGD-C Ctrl.-C CGD-C CtrI.-M CGD M CGD M (trl.-N1 Ctrl.-M CGD-M CGD-M Ctrl.-M Figure 1. Phosphorylation of normal and CGD PMN cytosolic pro- and indirect autoradiography. Ctrl.-C/Ctrl.-M, control cytosol + con- teins in the cell-free system. PMN from a normal donor and a CGD trol membrane; CGD-C/CGD-M, CGD cytosol + CGD membrane; patient (S.S.) were fractionated; cytosolic and membrane fractions Ctrl.-C/CGD-M, control cytosol + CGD membrane; COD-C/Ctrl.- were mixed with 50 MCi [y-32P]ATP in the presence of 460 gM SDS M, CGD cytosol + control membrane. Arrow marks the position of (+) or of H20 as solvent control (-); and fractions were reseparated the 48-kd phosphoprotein(s). Molecular weight markers are in kilo- as described in Methods. Cytosolic fractions (31 ug protein) were an- daltons. (A) Silver-stained polyacrylamide gel. (B) Corresponding au- alyzed by SDS-polyacrylamide gel electrophoresis, silver staining, toradiogram. CGD cytosolic fraction, or that the 48-kD phosphoprotein(s) is increased phosphorylation and three had decreased phosphor- not present in the cytosblic fraction from the CGD PMN. ylation compared with the same fraction isolated from a re- Analysis of membrane fractions, reisolated after reconsti- constitution mixture incubated in the absence of SDS (lane I tution of normal membrane and cytosolic fractions in the pres- vs lane 2). Analogous to the results with the cytosolic fractions, ence of SDS, by SDS-polyacrylamide gel electrophoresis and phosphorylation of a 48-kD protein(s) (arrow) was markedly autoradiography showed that 32 phosphoproteins were - enhanced by SDS treatment in the mixture of normal cytosolic present (Fig. 2 B, lane 1). Of these phosphoproteins, 14 had and membrane fractions (lanes 1 and 2), but not during re- Cell-free Protein Phosphorylation and Neutrophil Oxidase Activation 1489
  • 11. A. Membrane Fractions-Silver Stain B. Membrane Fractions - Autoradiogram f.i so 4s Os6 -200 _. E4 ilA. i.. W -i F:n am - 200 -116 ,-92.5 -116 1 -92.5 i I i -66 -66 I. I -45 - 45 -31 -31 w.l ^ A. w A. A. ,w ~~~~~21.5 - _; L I -21.5 - - 1!iF A:: z t s14 ::ts: -14 _IIX + -/ + -I +±/=I _ i+ / + Ctrl.- C CGD-C Ctrl.-C CGD- C Ctrl.-C CGD-C Ctrl.-C CGD-C Ctrl.-M CGD-M CGD-M Ctrl.-M Ctrl.-M CGD-M CGD-M Ctrl.-M Figure 2. Phosphorylation of normal and CGD PMN membrane viations and methods are as described in the legend to Fig. 1. The proteins in the cell-free system. Membrane fractions (70.3 ,Ag pro- arrow marks the position of the 48-kd phosphoprotein(s). Molecular tein) reseparated from cytosolic fractions in the reconstitution experi- weight markers are in kilodaltons. (A) Silver-stained polyacrylamide ment described in Fig. 1 were analyzed by SDS-polyacrylamide gel gel. (B) Corresponding autoradiogram. electrophoresis, silver staining, and indirect autoradiography. Abbre- constitution of the CGD cytosol and membrane fraction (lanes tient having autosomal recessive CGD, were subjected to den- 3 and 4). Reconstitution of CGD membrane fractions with sitometric analysis. Phosphoproteins in the region of 48 kD normal cytosol fully restored SDS-dependent phosphorylation (arrows) are shown in Fig. 3. In the presence of SDS, phosphor- of the 48-kD phosphoprotein(s) (lanes 5 and 6). In contrast, ylation of a 48-kD protein(s) as a result of reconstitution of minimal phosphorylation of the 48-kD phosphoprotein(s) by normal cytosol with normal membrane fractions was ob- SDS occurred when the CGD cytosolic fraction was reconsti- served. No phosphorylation of that protein(s) was observed tuted with normal membrane fraction (lanes 7 and 8), further during reconstitution of CGD cytosol with CGD membrane indicating the requirement for a cytosolic component that is fractions in the presence of SDS. Reconstitution of normal not functional in autosomal recessive CGD PMN. cytosol with CGD membrane fractions and SDS restored Autoradiograms of membrane fractions from two addi- phosphorylation of the 48-kD protein(s). In contrast, reconsti- tional cell-free reconstitution experiments, one from each pa- tution of CGD cytosol with normal membrane fractions and 1490 Caldwell, McCall, Hendricks, Leone, Bass, and McPhail
  • 12. 0 0 C 0 .0 Figure 3. Densitometry scans of autoradiograms of membrane fractions from normal and CGD PMN following reconstitution with SDS in the cell-free sys- tem. Unstimulated PMN from two normal donors and two patients with autosomal recessive CGD (S.S. and S.H.) were fractionated; cytosolic and membrane fractions were reconstituted in the presence of [y-32P]- ATP and 460 MM SDS and reseparated as described in Methods. The experiment with S.S. is separate from the one described in Fig. 1. Membrane fractions (S.S., 100 ,ug protein; S.H., 41 gg protein) were sub- jected to SDS-polyacrylamide gel electrophoresis, silver-staining, and autoradiography. Autoradiograms were scanned by a soft laser densitometer in the region of the 48 kD phosphoprotein(s) (arrows). In all scans the bottom of the gel is to the left. (A) CGD 1 (S.S.). (B) CGD 2 (S.H.). (Top panels) Nor- mal cytosol + normal membrane; (second panels) CGD cytosol + CGD membrane; (third panels) nor- mal cytosol + CGD membrane; (bottom panels) Rf CGD cytosol + normal membrane. SDS resulted in minimal, or no, phosphorylation in that re- reconstitution mixtures of normal cell fractions and the lack of gion of the gel. phosphorylation in mixtures of CGD cell fractions was ob- The area under the curve at 48 kD on the scans shown in served in both cytosolic and membrane fractions, it is possible Fig. 3 and on scans of the lanes containing fractions isolated that the same protein(s) is distributed between both fractions. from SDS-containing mixtures in the autoradiogram shown in Cytosolic and membrane fractions reisolated from the normal Fig. 2 was determined. If the area for the reconstitution mix- and CGD reconstitution mixtures described in Figs. 1 and 2 ture of normal cytosol and CGD membrane is set at 100%, the were analyzed on the same gel to test this possibility. The summarized data were as follows: normal cytosol + normal results are shown in Fig. 4. The position of the 48-kD phos- membrane = 75±22; CGD cytosol + CGD membrane phoprotein(s) can be identified by the increase in phosphoryla- = 1.5±1.5; CGD cytosol + normal membrane = 14±8 (per- tion induced by SDS in the normal cell fractions and the ab- cent, mean±SEM, n = 3). The inability of the CGD cytosol sence of that increase in the fractions from the CGD PMN. and membrane to phosphorylate the membrane-associated The 48-kD phosphoprotein(s) (arrow) in both cytosolic and 48-kd protein(s), and the markedly decreased ability of the membrane fractions comigrated (lanes I and 2 from left), sup- CGD cytosol to support phosphorylation when mixed with porting the possibility that they are the same protein or family normal membrane thus was observed in three separate experi- of proteins. ments. Taken together, these results indicate that a cytosolic It has been reported previously that neutrophils from pa- factor activity necessary for SDS-dependent activation of tients with the autosomal recessive form of CGD demonstrate NADPH oxidase and defective in two patients with autosomal a defect in phosphorylation of a protein(s) in the 47-48-kD recessive CGD, is also necessary to support SDS-dependent region when intact cells are exposed to a stimulus (15, 17). It is phosphorylation of a 48-kD phosphoprotein(s). thus possible that the 48-kD protein(s) observed in our cell-free Since phosphorylation of a 48-kD protein(s) by SDS in system is the same as that in the intact cell system. We tested Cell-free Protein Phosphorylation and Neutrophil Oxidase Activation 1491
  • 13. g V.- '.f' +SDS A. Normal .. W,: .i',; 1R a. -d -SDS * + SDS C CGD NI 4w ee a: - .: -SD.-.r 1CM,40.' / ~ ;Aw' ::c 0. SD: Figure 4. Comigration in SDS-polyacrylamide gel of the 48 kd phos- -200 -116 -92.5 -66 -45 -31 -21.5 -14 phoprotein(s) in both cytosolic and membrane fractions following re- constitution with SDS in the cell-free system. Cytosolic (C, 31 yg protein) and membrane (M, 70 Mig protein) fractions from the experi- ment described in Fig. 1 were electrophoresed on the same polyacryl- amide gel and autoradiography was performed. +SDS, fractions re- constituted in the presence of 460 MM SDS; -SDS, fractions recon- stituted in the presence of H20 (solvent control); Normal, PMN fractions from a normal donor; CGD, PMN fractions from a patient with autosomal recessive CGD (S.S.). Molecular weight markers are in kilodaltons. The arrow indicates the 48-kD position. that possibility by analyzing, in the same gel, phosphoprotein .t Normal / i Recon. Intact .! I . A. I d I .1 Um i? ~ &i 0< W I 11 .i .'1 rI I.-le.' Co cI) Recon. CGD 1? ° ^ I' Figure 5. Comparative phosphorylation patterns in cytosolic frac- . *g Ha. A* *: ~ 1 Ft .i s 'I * * tions from cell-free reconstitution mixtures and from intact normal and CGD PMN. Cytosolic fractions from the reconstitution experi- .w. ment described in Fig. 1 (31 Mg protein; Recon.) and from intact nor- mal or CGD PMN that had been loaded with 32P; and exposed to ei- ther DMSO or 100 ng/ml PMA for 30 s (64 Mg protein; Intact) were analyzed in the same SDS-polyacrylamide gel and subjected to auto- radiography. SDS, +460 MAM SDS; H20, solvent control; Normal, fractions isolated from normal PMN; CGD, fractions isolated from a patient (S.S.) with autosomal recessive CGD. Molecular weight markers are in kilodaltons. Arrow marks the position of the 48-kD phosphoprotein(s). 48-kD protein(s) was observed in cytosolic fractions from :Siia A: ..S j t IF a as sX * us. A; (0 :* .F. _ _ < Intact - 200 -116 -92.5 -66 -45 -31 -21.5 -14 patterns of cytosolic fractions prepared from cell-free reconsti- CGD cells. In addition, the 48-kD phosphoprotein(s) in cyto- tution mixtures incubated with [y-32P]ATP and cytosolic frac- solic fractions from normal reconstitution mixtures and from tions prepared from resting or PMA-stimulated 32Pi-labeled normal intact cells appeared to comigrate, supporting the pos- intact cells (Fig. 5). Interestingly, the phosphoprotein patterns sibility that they are the same protein(s). of the cell-free versus the intact cell cytosolic fractions were PKC activity in autosomal recessive CGD PMN. It was quite dissimilar. However, a defect in both SDS-dependent possible that the lack of phosphorylation of the 48-kD phos- phosphorylation of a 48-kD protein(s) (arrow) in reconstitu- phoprotein(s) and the absence of NADPH oxidase activity tion mixtures and in PMA-stimulated phosphorylation of a were due to deficient levels of PKC. The activity and distribu- 1492 Caldwell, McCall, Hendricks, Leone, Bass, and McPhail
  • 14. tion of PKC were determined in crude subcellular fractions zyme complex at the molecular level is being made (20-22). from unstimulated and stimulated PMN of normal controls Recently, a cell-free system for activation of NADPH oxidase and the two patients with autosomal recessive CGD. As shown was developed by us (25) and others (26-28) and can now be in Fig. 6, the level and distribution of PKC activity in cytosolic utilized to more directly study the mechanisms involved. Of and particulate EP and NEP fractions was similar in normal particular interest is the role of a newly discovered cytosolic and CGD PMN, under all conditions examined. PMA stimu- cellular cofactor required for activation of the oxidase in this lation of PMN from both patients with CGD induced the system (25-28). Protein phosphorylation has been implicated redistribution and loss of PKC activity similar to that observed as a mechanism ofoxidase activation (3-17) and phosphoryla- in each paired normal control (Fig. 6, B). The effect of fMLP tion has not been directly studied in the cell-free system. stimulation was examined in one patient with CGD and the Therefore, we have examined whether or not phosphorylation normal increase in particulate-associated (EP and NEP) kinase of endogenous proteins correlated with the cell-free activation activity was observed. These studies suggest that PKC is not of NADPH oxidase and have investigated the role of the cyto- defective in the two patients with CGD. solic factor in the regulation of the observed phosphorylation events. Discussion Activation of NADPH oxidase in the cell-free system cor- related specifically with phosphorylation ofa 48-kD protein(s). The mechanisms that regulate NADPH oxidase, the respira- NADPH oxidase activation was achieved by the addition of tory burst enzyme in human PMN, are not yet well under- SDS to a reconstituted mixture of cytosolic and membrane stood; although, progress in elucidating the nature of the en- fractions from unstimulated normal PMN, in agreement with previous reports in guinea pig macrophages (29) and human PMN (33), and similar to results obtained with arachidonic 800 A. DMSO Normal 1 Normal 2 acid instead of SDS as the activator (25-28). SDS also induced CGD 2 the phosphorylation or dephosphorylation of at least 19 cyto- solic and 17 membrane-associated proteins, including a 48-kD ] Cytood protein(s) seen in both fractions, out of 56 phosphoproteins observed in the reconstituted cell-free system. Thus, SDS ap- ~~~NEP 2W pears capable of activating a protein kinase(s), as well as either activating a phosphatase(s) or inhibiting a protein kinase(s). B. PMA Although the identity of the kinase or kinases affected by SDS is not yet known, we have found that SDS can activate PKC in fat Normala IOr D Normal2 CGD 2 cytosolic fractions from human PMN as measured by exoge- nous histone phosphorylation (data not shown). Therefore, at 1i400 least some of the phosphorylation reactions observed may be 2- catalyzed by PKC. .-G 200 Phosphorylation of the 48-kD protein(s) by SDS was mark- E C. fMLP edly decreased in cell-free reconstitution mixtures of cytosolic Normal 1 CGD 1 and membrane fractions from PMN of two patients with au- tosomal recessive CGD. This was the only phosphorylation 600 defect noted and it correlated with the dramatic decrease in the 400 ability of SDS to activate NADPH oxidase in the same frac- tions. The defects in activation of NADPH oxidase and in 200 phosphorylation of the 48-kD protein(s) could both be attrib- 0 uted to a defect in the cytosolic fraction from the CGD PMN. Fraction Assayed The membrane fraction from CGD PMN appeared to be nor- mal, in that it supported normal levels of both phosphoryla- Figure 6. PKC activity and distribution in subcellular fr-actions, from tion of the membrane-associated 48-kD protein(s) and activa- normal and CGD PMN. PKC activity was measured in subcellular tion of NADPH oxidase when reconstituted with a normal fr-actions isolated from normal or CGD PMN that had been treated with either DMSO (A), 100 ng/ml PMA (B) or IO' M fMLP (C) for cytosolic fraction. In addition, reconstitution of CGD cyto- 30 s at 370C as described in Methods. Fractions assayed were the cy- solic fraction with normal membrane fraction mimicked the tosolic (o, Cytosol), the detergent extract of the particulate fraction defects in phosphorylation of the 48-kD protein and in activa- (Es, EP), and the resuspended residual particulate pellet from the de- tion of NADPH oxidase seen with the reconstituted CGD tergent extract (-, NEP). Activity shown is the mean of triplicate de- fractions. Thus, the cytosolic factor required for activation of terminations on each fraction and is given as net activity (the activity NADPH oxidase in the cell-free system is defective in PMN of in the presence of calcium and lipids minus the activity in the ab- two patients with the autosomal recessive form of CGD, as sence of calcium and lipids) for the cytosolic and EP fractions and as originally suggested by Curnutte et al. (31), and this defect activity in the absence of calcium and lipids for the NEP fractions correlates with defective phosphorylation of a 48-kD pro- (see Methods). Activity in the absence of calcium and lipids in the tein(s). EP and cytosolic fractions was not different in the fractions from normal and CGD PMN and was < 91 pmol/min per I07 PMN in cy- These results clearly indicate that the cytosolic factor re- tosolic and 17 pmol/min per 107 PMN in EP fractions. Two paired quired for activation of NADPH oxidase also is required for experiments, each with PMN from a different normal and a different phosphorylation of the 48-kD protein(s). Our results do not patient with autosomal recessive CGD, were performed: CGD 1, differentiate between two possible explanations for the rela- S.S.; CGD 2, S.H. tionship between the cytosolic factor and the 48-kD protein(s). Cell-free Protein Phosphorylation and Neutrophil OxidaseActivation 1493
  • 15. The cytosolic factor could be the 48-kD phosphoprotein(s) and 48-kd protein(s) observed by the various groups and to deter- its presence in the membrane due to partial translocation in- mine if all groups are studying the same family of proteins. duced by SDS. If so, then the presence of the 48-kD phospho- The relationship of the phosphorylation of the 48-kD pro- protein in the CGD membrane fraction could be caused by tein(s) to the activation ofNADPH oxidase in the intact cell or translocation from the normal cytosolic fraction during re- by SDS in the cell-free system is still not clear. Phosphoryla- constitution. In fact, the 48-kD protein(s) in normal mem- tion of a protein or proteins at 46-48 kD has been correlated brane fractions comigrated with the 48-kD protein(s) in nor- with activation of the respiratory burst in intact cells by several mal cytosolic fractions, supporting the possibility that they are laboratories (8, 56-58). However, other data dissociates phos- the same protein(s) and could either have dual localization or phorylation of a 48-kD protein(s) from stimulation of oxida- be translocated by SDS treatment. Observations consistent tive metabolism (59-61). With the exception of Ohtsuka et al. with the possibility that translocation of the cytosolic factor (56), these studies have utilized one-dimensional gel separa- occurs are that, under conditions that prime intact normal tion systems and, if the 48-kD region contains a family of PMN or macrophages, NADPH oxidase in membrane frac- proteins (16, 17, 24), such discrepancies might be expected. tions can be activated in the cell-free system without the addi- A requirement for a protein kinase activity in the SDS-me- tion of cytosolic fractions (25, 29). diated activation of NADPH oxidase in the cell-free system It is also possible that the cytosolic factor regulates phos- has not yet been demonstrated nor conclusively eliminated. phorylation of the 48-kD protein(s) because the factor is itself a Oxidase activation by SDS or arachidonate does not require kinase or it activates a kinase. Our results suggest that the exogenously added ATP (30, 35) and depletion of endogenous cytosolic factor is not PKC, since PKC activity appeared to be ATP has been reported by two groups to inhibit (35, 55) and present at normal levels and to respond normally to stimula- by another to have no effect (30) on oxidase activation. PKC tion by either PMA or fMLP in CGD PMN. However, we has been shown to activate NADPH oxidase in a cell-free sys- cannot exclude the possibility of a more subtle defect in PKC tem (62); however, the mechanism utilized by PKC appears to activity or regulation not detected by our assay procedure. be different than that utilized by SDS and arachidonate (33) That the cytosolic factor is not PKC is also supported by others and PKC does not seem to be required for the latter com- using inhibitors (33, 34, 55) or column fractionation (30) to pounds to exert their activating effect (30, 33, 34). These re- dissociate PKC activity from the oxidase activating factor in sults and our own suggest that if a kinase is involved in the normal PMN cytosol. It is interesting that 46-48-kD proteins SDS-activated cell-free system, it is not PKC. in guinea pig (56) and human (57) PMN have been reported to Based on the present state of knowledge and the work re- be substrates for PKC or its catalytic subunit, although it is not ported here, at least three possibilities for the role of the 48-kD yet known if these and the protein(s) observed here are the protein(s) in the activation of NADPH oxidase are apparent. same. Further understanding of the relationship between the The 48-kD protein(s) may be unrelated to oxidase activity in 48-kD protein, the cytosolic factor required for NADPH oxi- PMN, but its phosphorylation may depend upon a normal dase activation, and PKC or other kinases will necessitate puri- cytosolic factor activity. Alternatively, the 48-kD protein(s) fication ofthe components and more rigorous physical charac- may be a component(s) of NADPH oxidase; a protein of that terization of the 48-kD protein or proteins. molecular weight appears, albeit variably, in preparations of A defect in phosphorylation of a 47-48-kD protein, or the purified oxidase flavoprotein (63). Finally, the 48-kD pro- family of proteins, in PMN from patients with CGD has been tein(s) could be a regulatory molecule(s), influencing one or observed previously by stimulation of intact cells (15-17). more of several pathways (2, 11, 18, 19, 33, 36) of activation of Segal et al. (15), using one-dimensional polyacrylamide gel oxidative metabolism. Further efforts will be made to distin- electrophoretic separation, reported that the defect was re- guish between these possibilities and to resolve the current stricted to patients with the autosomal recessive form of the discrepancies in the literature by fully characterizing the disease. However, Hayakawa et al. (16) and Okamura et al. 48-kD protein(s). (17), using two-dimensional gel electrophoresis, found that both forms of the disease expressed defects in phosphorylation Acknowledgments of a 48-kD family of proteins and suggested that a more pro- nounced defect is present in the autosomal recessive form. A We thank Sue Cousart, William Crone, and Gary Nabors for their third group (24), using two-dimensional gel electrophoresis, excellent technical assistance and Susan Britt for valuable secretarial reported no defect in protein phosphorylation in six patients help. We especially thank Dr. Mike Cohen for providing access to his with CGD (presumably representing both forms of inheri- patient with autosomal recessive CGD. tance). Our results, using one-dimensional gel electrophoresis, This work was supported in part by National Institutes of Health indicate, for the first time, that a defect in phosphorylation ofa grants AI-22564, AI-10732, AI-09169, AI-14929, and HL-29293. 48-kD protein(s) can be observed in a cell-free system from PMN of patients with the autosomal recessive form of CGD References and relate the defect to a deficient cytosolic factor required for activation of NADPH oxidase. The 48-kD protein(s) observed 1. Babior, B. M. 1978. Oxygen-dependent microbial killing by in our cell-free system by SDS treatment comigrated with a phagocytes. N. Engl. J. Med. 298:659-668. 48-kD protein observed by stimulating intact PMN with PMA, 2. McPhail, L. C., and R. Snyderman. 1984. Mechanisms of regu- consistent with the possibility that the phosphorylation defect lating the respiratory burst in leukocytes. In Regulation of Leukocyte Function. R. Snyderman, editor. Plenum Publishing Corp., New in intact CGD PMN is the same as in the cell-free system. York. 247-281. However, further analysis by two-dimensional gel electropho- 3. Schneider, C., M. Zanetti, and D. Romeo. 1981. Surface-reactive resis, peptide mapping, and phosphoamino acid determina- stimuli selectively increase protein phosphorylation in human neutro- tion will be necessary to fully identify and characterize the phils. FEBS (Fed. Eur. BioL. Soc.) Lett. 127:4-8. 1494 Caldwell, McCall, Hendricks, Leone, Bass, and McPhail
  • 16. 4. Andrews, P. C., and B. M. Babior. 1983. Endogenous protein 21. Dinauer, M. C., S. H. Orkin, R. Brown, A. J. Jesaitis, and C. A. phosphorylation by resting and activated human neutrophils. Blood. Parkos. 1987. The glycoprotein encoded by the X-linked chronic 61:333-340. granulomatous disease locus is a component of the neutrophil cy- 5. Andrews, P. C., and B. M. Babior. 1984. Phosphorylation of tochrome b complex. Nature (Lond.). 327:717-720. cytosolic proteins by resting and activated human neutrophils. Blood. 22. Teahan, C., P. Rowe, P. Parker, N. Totty, and A. W. Segal. 64:883-890. 1987. The X-linked chronic granulomatous disease gene codes for the 6. White, J. R., C.-K. Huang, J. M. Hill, Jr., P. H. Naccache, E. L. B-chain of cytochrome b-245. Nature (Lond.). 327:720-72 1. Becker, and R. I. Sha'afi. 1984. Effect of phorbol 12-myristate 13-ace- 23. Segal, A. W., A. R. Cross, R. C. Garcia, N. Borregaard, N. H. tate and its analogue 4-a-phorbol 12,13-didecanoate on protein phos- Valerius, J. F. Soothill, and 0. T. G. Jones. 1983. Absence of cy- phorylation and lysosomal enzyme release in rabbit neutrophils. J. tochrome b-245 in chronic granulomatous disease. A multicenter Euro- Biol. Chem. 259:8605-8611. pean evaluation of its incidence and relevance. N. Engl. J. Med. 7. Okamura, N., S. Ohashi, N. Nagahisa, and S. Ishibashi. 1984. 308:245-251. Changes in protein phosphorylation in guinea pig polymorphonuclear 24. Ishii, E., K. Irita, I. Fujita, K. Takeshige, M. Kobayashi, T. leukocytes by treatment with membrane-perturbing agents which Usui, and K. Ueda. 1986. Protein phosphorylation ofneutrophils from stimulate superoxide anion production. Arch. Biochem. Biophys. normal children and patients with chronic granulomatous disease. 228:270-277. Eur. J. Pediatr. 145:22-26. 8. Pontremoli, S., E. Melloni, F. Salamino, B. Sparatore, M. Mi- 25. McPhail, L. C., P. S. Shirley, C. C. Clayton, and R. Snyderman. chetti, 0. Sacco, and B. L.Horecker. 1986. Phosphorylation of pro- 1985. Activation of the respiratory burst enzyme from human neutro- teins in human neutrophils activated with phorbol myristate acetate or phils in a cell-free system. Evidence for a soluble cofactor. J. Clin. with chemotactic factor. Arch. Biochem. Biophys. 250:23-29. Invest. 75:1735-1739. 9. Nishizuka, Y. 1984. The role of protein kinase C in cell surface 26. Bromberg, Y., and E. Pick. 1984. Unsaturated fatty acids stim- signal transduction and tumour promotion. Nature (Lond.). 308:693- ulate NADPH-dependent superoxide production by cell-free system 698. derived from macrophages. Cell. Immunol. 88:213-221. 10. Wolfson, M., L. C. McPhail, V. N. Nasrallah, and R. Snyder- 27. Heyneman, R. A., and R. E. Vercauteren. 1984. Activation ofa man. 1985. Phorbol myristate acetate mediates redistribution of pro- NADPH oxidase from horse polymorphonuclear leucocytes in a cell- tein kinase C in human neutrophils: potential role in the activation of free system. J. Leukocyte Biol. 36:751-759. the respiratory burst enzyme. J. Immunol. 135:2057-2062. 28. Curnutte, J. T. 1985. Activation of human neutrophil nicotin- 11. Gerard, C., L. C. McPhail, A. Marfat, N. P. Stimler-Gerard, amide adenine dinucleotide phosphate, reduced (triphosphopyridine D. A. Bass, and C. E. McCall. 1986. Role of protein kinases in stimula- nucleotide, reduced) oxidase by arachidonic acid in a cell-free system. tion of human polymorphonuclear leukocyte oxidative metabolism by J. Clin. Invest. 75:1740-1743. various agonists. Differential effects of a novel protein kinase inhibitor. 29. Bromberg, Y., and E. Pick. 1985. Activation of NADPH-de- J. Clin. Invest. 77:61-65. pendent superoxide production in a cell-free system by sodium dode- 12. Gennaro, R., C. Florio, and D. Romeo. 1986. Co-activation of cyl sulfate. J. Biol. Chem. 260:13539-13545. protein kinase C and NADPH oxidase in the plasma membrane of 30. Curnutte, J. T., R. Kuver, and P. J. Scott. 1987. Activation of neutrophil cytoplasts. Biochem. Biophys. Res. Commun. 134:305- neutrophil NADPH oxidase in a cell-free system. Partial purification 312. of components and characterization of the activation process. J. Biol. 13. Cox, C. C., R. W. Dougherty, B. R. Ganong, R. M. Bell, J. E. Chem. 262:5563-5569. Niedel, and R. Snyderman. 1986. Differential stimulation of the respi- 31. Cumutte, J. T., P. J. Scott, R. Kuver, and R. Berkow. 1986. ratory burst and lysosomal enzyme secretion in human polymorpho- NADPH oxidase activation cofactor (ACF): partial purification and nuclear leukocytes by synthetic diacylglycerols. J. Immunol. absent activity in a patient with chronic granulomatous disease (CGD). 136:4611-4616. Clin. Res. 34:455a. (Abstr.) 14. Rider, L. G., and J. E. Niedel. 1987. Diacylglycerol accumula- 32. McPhail, L. C., C. C. Clayton, and R. Snyderman. 1984. A tion and superoxide anion production in stimulated human neutro- potential second messenger role for unsaturated fatty acids: activation phils. J. Biol. Chem. 262:5603-5608. of Ca2'-dependent protein kinase. Science (Wash. DC). 224:622-625. 15. Segal, A. W., P. G. Heyworth, S. Cockcroft, and M. M. Bar- 33. Cox, J. A., A. Y. Jeng, P. M. Blumberg, and A. I. Tauber. 1987. rowman. 1985. Stimulated neutrophils from patients with autosomal Comparison of subcellular activation of the human neutrophil recessive chronic granulomatous disease fail to phosphorylate a NADPH-oxidase by arachidonic acid, sodium dodecyl sulfate (SDS), M,-44,000 protein. Nature (Lond.). 316:547-549. and phorbol myristate acetate (PMA). J. Immunol. 138:1884-1888. 16. Hayakawa, T., K. Suzuki, S. Suzuki, P. C. Andrews, and B. M. 34. McPhail, L. C., C. L. Hendricks, M. C. Seeds, and D. A. Bass. Babior. 1986. A possible role for protein phosphorylation in the acti- 1986. A possible role for calmodulin in activation of the respiratory vation of the respiratory burst in human neutrophils. Evidence from burst enzyme of human neutrophils. Clin. Res. 34:465a. (Abstr.) studies with cells from patients with chronic granulomatous disease. J. 35. Clark, R. A., D. G. Leidal, D. W. Pearson, and W. M. Nauseef. Biol. Chem. 261:9109-9115. 1987. NADPH oxidase of human neutrophils. Subcellular localization 17. Okamura, N., J. T. Curnutte, and B. M. Babior. 1987. Phos- and characterization of an arachidonate-activatable superoxide-gener- phoproteins and the activation of the neutrophil respiratory burst oxi- ating system. J. Biol. Chem. 262:4065-4074. dase. Fed. Proc. 46:1952. (Abstr.) 36. McPhail, L. C., and R. Snyderman. 1983. Activation of the 18. Tauber, A. I., N. Borregaard, E. Simons, and J. Wright. 1983. respiratory burst enzyme in human polymorphonuclear leukocytes by Chronic granulomatous disease: a syndrome of phagocyte oxidase de- chemoattractants and other soluble stimuli. Evidence that the same ficiencies. Medicine (Baltimore). 62:286-309. oxidase is activated by different transductional mechanisms. J. Clin. 19. Curnutte, J. T., and B. M. Babior. 1987. Chronic granuloma- Invest. 72:192-200. tous disease. In Advances in Human Genetics. Vol. 16. H. Harris and 37. Sanders, D. Y., M. R. Cooper, C. E. McCall, and L. R. DeCha- K. Hirschhorn, editors. Plenum Publishing Corp., New York. telet. 1972. Chronic granulomatous disease of childhood with onset of 229-297. symptoms at age 11 years. J. Pediatr. 80:104-106. 20. Royer-Pokora, B., L. M. Kunkel, A. P. Monoaco, S. C. Goff, 38. Amrein, P. C., and T. P. Stossel. 1980. Prevention of degrada- P. E. Newburger, R. L. Baehner, F. S. Cole, J. T. Curnutte, and S. H. tion of human polymorphonuclear leucocyte proteins by diisopropyl- Orkin. 1986. Cloning the gene for an inherited human disorder- fluorophosphate. Blood. 56:442-447. chronic granulomatous disease-on the basis of its chromosomal lo- 39. Gabig, T. G. D. English, L. P. Akard, and M. J. Schell. 1987. cation. Nature (Lon4.). 322:32-38. Regulation of neutrophil NADPH oxidase activation in a cell-free Cell-free Protein Phosphorylation and Neutrophil Oxidase Activation 1495
  • 17. system by guanine nucleotides and fluoride. Evidence for participation 53. Laskey, R., and A. D. Mills. 1977. Enhanced autoradiographic of a pertussis and cholera toxin-insensitive G protein. J. Biol. Chem. detection of 32P and 1251 using intensifying screens and hypersensitized 262:1685-1690. film. FEBS (Fed. Eur. Biol. Soc.) Lett. 82:314-316. 40. Huey, R., and T. E. Hugh. 1985. Characterization of a C5a 54. Pike, M. C., L. Jakoi, L. C. McPhail, and R. Snyderman. 1986. receptor on human polymorphonuclear leukocytes (PMN). J. Im- Chemoattractant-mediated stimulation of the respiratory burst in munol. 135:2063-2068. human polymorphonuclear leukocytes may require appearance of 41. Wroblewski, F., and J. S. LaDue. 1955. Lactic dehydrogenase protein kinase activity in the cells' particulate fraction. Blood. 67:909- activity in blood. Proc. Soc. Exp. Biol. Med. 90:210-213. 913. 42. DeChatelet, L. R., J. V. Volk, C. E. McCall, and M. R. Cooper. 55. Seifert, R., and G. Schultz. 1987. Fatty-acid-induced activation 1971. Studies on leukocyte phosphatases. II. Inhibition of leukocyte of NADPH oxidase in plasma membranes of human neutrophils de- alkaline phosphatase by amino acids and its reversal by zinc. Clin. pends on neutrophil cytosol and is potentiated by stable guanine nu- Chem. 17:210-213. cleotides. Eur. J. Biochem. 162:563-569. 43. Gottlieb, C., K. Lau, L. R. Wasserman, and V. Herbert. 1965. 56. Ohtsuka, T., N. Okamura, and S. Ishibashi. 1986. Involvement Rapid charcoal assay for intrinsic factor (IF), gastric juice unsaturated of protein kinase C in the phosphorylation of 46 kDa proteins which B12 binding capacity, antibody to IF, and serum unsaturated B12 bind- are phosphorylated in parallel with activation of NADPH oxidase in ing capacity. Blood. 25:875-883. intact guinea-pig polymorphonuclear leukocytes. Biochim. Biophys. 44. Zeya, H. I., and J. K. Spitznagel. 1971. Characterization of Acta. 888:332-337. cationic protein-bearing granules of polymorphonuclear leukocytes. 57. Pontremoli, S., E. Melloni, M. Michetti, B. Sparatore, F. Sala- Lab. Invest. 24:229-236. mino, 0. Sacco, and B. L. Horecker. 1987. Phosphorylation and pro- 45. Migler, R., and L. R. DeChatelet. 1978. Human eosinophilic teolytic modification of specific cytoskeletal proteins in human neutro- peroxidase: Biochemical characterization. Biochem. Med. 19:16-26. phils stimulated by phorbol 12-myristate 13-acetate. Proc. Nati. Acad. 46. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. Sci. USA. 84:3604-3608. 1951. Protein measurement with the Folin phenol reagent. J. Biol. 58. Heyworth, P. G., and A. W. Segal. 1986. Further evidence for Chem. 193:265-275. the involvement of a phosphoprotein in the respiratory burst oxidase 47. Massey, V. 1959. The microestimation of succinate and the of human neutrophils. Biochem. J. 239:723-73 1. extinction coefficient of cytochrome c. Biochim. Biophys. Acta. 59. Caldwell, S. E., D. A. Bass, L. C. McPhail, and C. E. McCall. 34:255-256. 1987. Phosphorylation of a 48 kD protein, defective in chronic granu- 48. Gabig, T. G., E. W. Schervish, and J. T. Santinga. 1982. Func- lomatous disease (CGD) neutrophils, is not required for activation of tional relationship of the cytochrome b to the superoxide-generating the respiratory burst. Fed. Proc. 46:1033. (Abstr.) oxidase of human neutrophils. J. Biol. Chem. 257:4114-4119. 60. Sha'afi, R. I., T. F. P. Molski, and C.-K. Huang. 1987. Dissocia- 49. Ohno, Y., E. S. Buescher, R. Roberts, J. A. Metcalf, and J. I. tion of the 47 kDa protein phosphorylation from degranulation and Gallin. 1986. Reevaluation of cytochrome b and flavin adenine dinu- superoxide production in neutrophils. Fed. Proc. 46:1037. (Abstr.) cleotide in neutrophils from patients with chronic granulomatous dis- 61. Reibman, J., H. M. Korchak, K. A. Haines, L. B. Vosshall, and ease and description of a family with probable autosomal recessive G. Weissmann. 1987. Phosphorylation of the 47 kDa target protein is inheritance of cytochrome b deficiency. Blood. 67:1132-1138. not sufficient for superoxide anion production in neutrophils. Clin. 50. Laemmli, U. K. 1970. Cleavage of structural proteins during Res. 35:537a. (Abstr.) assembly of the head of bacteriophage T4. Nature (Lond.). 227:680- 62. Cox, J. A., A. Y. Jeng, N. A. Sharkey, P. M. Blumberg, and A. I. 685. Tauber. 1985. Activation of the human neutrophil nicotinamide ade- 51. Peterson, G. L. 1977. A simplification of the protein assay nine dinucleotide phosphate (NADPH)-oxidase by protein kinase C. J. method of Lowry et al. which is more generally applicable. Anal. Bio- Clin. Invest. 76:1932-1938. chem. 83:346-356. 63. Glass, G. A., D. M. DeLisle, P. DeTogni, T. G. Gabig, B. H. 52. Wray, W., T. Boulikas, V. P. Wray, and R. Hancock. 1981. Magee, M. Markert, and B. M. Babior. 1986. The respiratory burst Silver staining of proteins in polyacrylamide gels. Anal. Biochem. oxidase of human neutrophils. Further studies of the purified enzyme. 118:197-203. J. Biol. Chem. 261:13247-13251. 1496 Caldwell, McCall, Hendricks, Leone, Bass, and McPhail
  • 18. 0022-1767/88/14010-3560$02.00/0 THE JOURNALIMMUNOLOGY OF Vol. 140.3560-3567. No. 10. May 15.1988 Copyright 0 1988 by The American Assodation of Immunologists Printed ( n U . S . A . ALTERATIONS IN CELL PROTEINPHOSPHORYLATION INHUMAN NEUTROPHILS EXPOSED TO INFLUENZAAVIRUS A Possible Mechanism for Depressed Cellular End-Stage Functions’ SUSAN E. CALDWELL,2’ LISA F. CASSIDY,+ AND JON S. ABRAMSON3* From the Departments of ‘Medicine, ‘Microbiology and Immunology, and ‘Pediatrics, Wake Forest University Medical Center, Winston-Salem.NC 27103 When polymorphonuclear leukocytes (PMN) are caused minimal changes in cAMP levels, and similar exposed to most harvests of influenza A virus (de- increases occurred in cAMP levels upon FMLP stim- pressing virus, DV) for 20 min, chemotactic, secre- ulation, tory, and oxidative functions are depressed upon To determine whether DV interferes with trans- subsequent exposure to soluble or particulate stim- membrane signaling, the effect of influenza virus uli. Otherharvests of influenza A virus (non-DV)do on P N transmembrane potential was studied by M not alter these activities. The DV-induced changes using a fluorescent cyanine dye. Transmembrane in multiple functions suggest the virus may inter- potential changes were greater in P N exposed to M fere with steps involved in P N activation. Because M DV than tonon-DV or buffer: however, subsequent some of these steps may be regulated by protein stimulation with FMLP caused equivalent changes phosphorylation, we examined the effect of non-DV in transmembrane potential. and DV on cellular protein phosphorylation. Our data show that protein phosphorylation in P N loaded with 32P-labeled M inorganic orthophos- P N is induced by DV and non-DV infection: upon M phate were exposedto non-DV, DV, or buffer for 30 subsequent stimulation with FMLP or PMA, there is min: cells were then treatedwith buffer, FMLP ( inhibited cellular phosphorylation onlyin P N pre- M M), or PMA (100 ng/ml) for30 s. Samples weresoni- viously exposed to DV. Studies on cAMP levels and cated and centrifuged: cytosolic andparticulate transmembrane potential changes suggest these pa- fractions were analyzed by SDS-PAGE and autora- rameters are not involved in DV-induced depression diography. Exposure PMN to either non-DV or DV of of phosphorylation. The role of intracellular [Ca”] caused phosphorylation of several cell proteins. in inhibiting DV-induced phosphorylation remains However, when DV-treated P N were then stimu- M to be clarified. lated with FMLP or PMA, further phosphorylation was inhibited compared to non-DV- or buffer-treated cells, This suggests that DV-induced depression of Exposure of PMN4to most harvests of influenza A virus PMN end-stage functions may be due to changes in causes inhibition of in vitro end-stage PMN functions cell protein phosphorylation. upon stimulation with secondarysoluble and particulate DV could interfere with phosphorylation oP N fM stimuli (1-7). Those depressed functions include respi- proteins by altering protein kinase activity. We ratory burst activity, secretion, and chemotaxis. In addi- therefore examined the influence of non-DV and DV tion, the harvestsof virus which inhibit these functions on some parameters that could affect kinase func- also cause disruptionof lysosome-phagosome fusion (4); tion. P N intracellular [Ca”] was monitored by us- M these viruses aredesignated DV. Maximal depression of ing the fluorescent Ca2+ indicator, Indo 1, and cAMP PMN function occurs within 30 min of exposure ofPMN levels weremeasured by RIA. P N treated with DV M to the virus. The inhibition of PMN function caused by alone or DV plus FMLP had higher intracellular DV is therefore occurring while virusparticles are [CA2+)than P N similarly treated with non-DV or M undergoing endocytosis into thePMN (4.6) before the but buffer. Exposure oP N to non-DV, DV. or buffer fM production of viral proteins (8). Occasional harvests of influenza virus do not interfere Received for publication December 1, 1987. with lysosome-phagosome function or PMN end-stage Accepted for publicatton February 16. 1988. functions:theseharvestsare designated non-DV (4). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked Studies have shown that at least as many non-DV parti- aduertfsernent in accordance with 18 U.S.C. Section 1734 solely to indi- cles are taken up PMN when compared with DV, and into cate this fact. that DV, but not non-DV, causes inhibition of fluid- and ‘ This study was supported in part by Grants AI-20506, and receptor-mediated endocytosis of PMN (7). AI-09169, Thus, the dif- HL-29293. J. S . A. is the recipient of Research Career Development Award AI-00670 from the National lnstitutes of Health. This work has been fering effectsthat DV and non-DV have on PMN function published in abstract form, entitled “Alterations in Protein Phosphoryla- not simply due to increased numbers of DV particles is tion During Influenza Virus Infection of Neutrophils May Induce De- pressed Cellular End-Stage Functions”[Clin. Res. 35:470A. 1987). taken upby the cell. Indeed, the cumulative datasuggest Present address: Genelabs. lnc., 505 Penobsot Dr.. Redwood City, CA 94063. Abbreviations used in this paper: PMN. polymorphonuclear leukocyte: Author to whom reprint requests should addressed: J o n S . Abram- be CB, cytochalasin 9; DV. depressing virus; CGD. chronic granulomatous son. M.D.. Department of Pediatrics. Wake Forest University Medical disease; non-DV, non-depressing virus: Di0C5(3). 3.3-dipentyloxacarbo- Center, 300 S . Hawthorne Road. Winston-Salem, NC 27103. cyanine: N,. guanine nucleotide protein:NI. non-infected. 3560
  • 19. VIRUS-INDUCED CHANGES IN NEUTROPHIL PROTEIN PHOSPHORYLATION 356 1 that influenza virus is interfering with one ormore steps of the buffy coat component cells Lymphocyte Separation Medium on (Bionetics, Inc., Bethesda, MD). and hypotonic lysis of E as previously in theactivation pathway of PMN. described (3). Cells were resuspended in HBSS (pH 7.4) with 1.3 mM Protein phosphorylation may have a major regulator ea2+. mM MgZ+, 0.1 % gelatin unless noted otherwise. 0.9 and role in activation of PMN responses to stimulating agents Preparation o influenza virus. Influenza A virus was harvested f from allantoic fluid of embryonated chicken eggs (19). purified by such as thechemotactic factor, FMLP, or the tumor pro- centrifugationon 15 to 60% (w/w) linear sucrose gradients, and moter, PMA. Indeed, phosphoproteins in uninfectedPMN dialyzed against PBS (10 mM sodium phosphate in 0.9%NaCl, pH undergo numerouschangesinphosphorylation upon 7.4) overnight. A recombinant virus (X-47) containing the internal stimulation of the cells with FMLP, PMA, and other ag- proteins of influenza A/Puerto Rico/8/34 (H,NI) virus and the sur- face glycoproteins of influenza A/Victoria/3/75 (&&?) virus was onists (9-18). changes which precede such responses a s used in all experiments. Most harvests of the X-47 virus inhibited activation of the respiratory burst, chemotaxis, and de- oxidative, secretory, and chemotactic activities inPMN (DV).Several granulation. harvests of this virus did not depress PMN functions (non-DV). Determination of the 50%egg infectivity dose. hemagglutinin and Recently, phosphorylation of a 48-kDa protein in hu- neuraminidase activities, total protein content, andSDS-PAGE pat- man PMN has beencorrelatedwithactivation of the terns have not established the difference between D and non-DV V respiratory burst enzyme, NADPH oxidase,instudies harvests (our unpublished data). The amount of DV and non-DV used in each experiment was adjusted to achieve equal virus protein done with PMN from patients with CGD (15.17). Such concentration (100 pg viral protein/l x 10' PMN unless otherwise cells cannot mounta respiratory burst upon stimulation: stated) as determined by the method of Lowry et al. (20). Each virus therefore, a phosphorylation defect in PMN from CGD harvest was aliquoted and stored a t -70°C until use. patients could indicateaspecificphosphoproteinin- Protein phosphorylation in influenza virus-infected PMN. Pro- tein phosphorylation experiments were modifications of those de- volved in regulating the respiratory burst. A phosphopro- scribed by s. E. Caldwell et al. (see footnote4) for uninfected, intact tein with a m.w. of 44,000 (15) or 48,000 (17), phos- PMN. Purified PMN (5 X lo7 PMN/sample) were suspended a t a phorylated in response to various stimuli in control PMN, concentration of 1 X 10' PMN/ml in phosphate-free buffer contain- ing 6 mM HEPES-Tris (pH 7.4), 0.15 mM NaCI, 1 0 mM glucose, 5 is not phosphorylated in PMN from patients with CGD. mM KCI, 1 mM MgCl,, 0.25 mM CaCI2(10).and 32P, 1 75 &i/5 X 1O7 ( We have shown that, in a cell-free, reconstituted system PMN: New England Nuclear. Boston, MA) and incubated for30 min derived from PMN from a patient with CGD, a 48-kDa a t 37°C with constant gentle agitation. Non-DV or DV in PBS (40 pg protein is not phosphorylated upon activation of the sys- viral protein/5 x 1O7 PMN) or PES alone was added to PMN so that cells were in a total volume of 0.6 to 0.7 ml, depending upon virus tem with SDS.4 Additionally, phosphorylation of a 48- concentration. The incubation was continued with gentle agitation kDa protein present in purified membranes from CGD for 30 min a t 37°C: PMN were then further diluted with warmed PMN was restored in the presence of normal cytosol phosphate-free buffer to a volume of 1.0 ml (5 X lo7 cells/ml) and stimulated for 30 secat 37°C with FMLP M) without CB. or under conditions in which a respiratory burst was also with PMA ( 100 ng/ml). Reactions were stopped with a 15-fold volume generated. These results further correlate phosphoryla- of ice-cold phosphate-free buffer also containing mM EDTA and 10 tion of this protein withrespiratory burst activation, 100 mM NaF as previously described (12). PMN were washed and incubated on ice for 5 min a t a concentration of 5 X IO7 cells/ml in although a causal relationship has not yet been demon- 2 mM diisopropylfluorophosphate to inhibitproteases (21) and strated. In the studies described here, depressionof phos- washed once more the samecold buffer. phorylation of the 48-kDa cytosolic protein in influenza PMN were resuspended in 0.7 ml sonication buffer containing virus-treated PMN occurs in parallel with depression of 0.34 M sucrose, 1 mM EDTA, 1 mM EGTA, 10 mM sodium pyro- phosphate, and 100 mM NaF. Cells were sonicated in a melting ice PMN oxidative metabolism. bath, with a model W-220 sonicator (Heat Systems-Ultrasonics,Inc.. Inasmuch as exposure of PMN to influenza virus is Farmingdale, NY) fittedwith a stepped microtip. Sonication was known to depress multiple end-stagePMN responses [ 1- done in 5- or 10-s bursts for a total of 25 s . and breakage was monitored by phase microscopy. Centrifugation a t 500 x g for 10 7). the effect of DV and non-DV on PMN protein phos- min a t 4°C removed intact cells and nuclei; the supernatant was phorylation was studied. DV, but not non-DV, was found centrifuged a t 100.000 X g for 1 h a t 4°C in a Beckman type 50 to cause marked inhibition in protein phosphorylation rotor. Resulting cytosolic fractions (supernatants] were removed and when cells were stimulated with FMLP or PMA. Further the pelleted particulate fractionsresuspended in 0.4 ml of the same sucrose buffer sonication. Cytosols and particulate fractions by were studies were then done to examine effect of influenza the divided into roughly equal aliquots: one aliquot of each was frozen virus on other parameters which could affect protein as native protein a t -70°C and later assayed for protein concentra- phosphorylation: intracellular [Ca"'], cAMP levels, and tion by Peterson's modification (22) of the method of Lowry et al. (20).These aliquots were also used for liquid scintillation counting PMN transmembrane potential.Interferencewith the to monitor 32Pincorporation. The remainder of each sample was first two parameters could directly affect protein phos- solubilized immediately by adding Laemmli electrophoresis sample phorylation in virus-infectedPMN by influencing activity buffer (23) and boiling in a water bath 2 min. Solubilized fractions for of the Ca2+- or CAMP-dependentprotein kinases. Changes were stored a t -70°C for no more than 3 days, and then analyzed by SDS-PAGE and indirect autoradiography, described below. in transmembrane signaling could indirectly influence SDS-PAGE and indirect autoradiography. Reduced, denatured protein phosphorylation by affecting steps in cell acti- pg samples (30 to 70 protein/lane) were electrophoresed by using the vation before phosphorylation. Thedata suggest DV does discontinuous buffer system of Laemmii (23) on linear 8 to 15% gradient slab gels a s described by S . E. Caldwell et al. (see footnote not alter protein phosphorylation affecting cAMP lev- by 4). Electrophoresis was carried out for 12 h a t 10°C. by using equal els or transmembrane potential. However, the roleof amounts of protein in eachgel lane within a n experiment. Gels were intracellular[Ca2+]in DV-induced changes in protein silver-stained (24) to detect total protein, and then dried between phosphorylation requires further study. two sheets of cellophane by using a model 443 slab gel dryer (Bio- Rad: Rockville Centre, NY). Dried gels were exposed to pre-flashed Kodak X-Omat x-ray film in the presence of a Dupont Cronex Light- MATERIALS AND METHODS ning Plus intensifying screen a t -70°C for 1 to 3 days to generate Isolation o P M N . Purified human PMN (2 97%) were obtained f autoradiograms (25).Autoradiograms were scanned with a n Ultro- from heparinized whole blood by dextran sedimentation, separation scan XL laser densitometer(LKB. Bromma, Sweden). P M N intracellular [Ca"]. Intracellular [Ca2'] was measuredin 4Caldwell. S . E., C. E. McCall, C. L. Hendricks. P. A. Leone, D A. . PMN washed in HBSS without ea2+or Mg2' and then resuspended Bass, and L. C. McPhail. Reconstitution of defect in phosphorylation of in HBSS with 1.6 mM CaC12.The acetoxymethyl ester of the fluores- a 48kd protein and NADPH oxidase activity in neutrophils from a patient cent compound, Indo 1 (Molecular Probes, Junction City, OR; kept with chronic granulomatous disease. J. Clin. Inuest. Submittedfor pub- a s a 1 mM stock solution in DMSO at -70°C). was added to PMN (1 Iication. X 10' cells/ml) at afinal concentration of 2 X M. and the cells
  • 20. 3562 VIRUS-INDUCED CHANGES IN NEUTROPHIL PROTEIN PHOSPHORYLATION were incubated at 25OC for 20 min. PMN were washed twice in ice- cold buffer and resuspendeda t a concentration of 1 X 10' PMN/ml in HBSS with 1.6 mMCaCI, (in certain experiments 2 mMEGTA was added to the buffer to remove extracellular Ca2+).Immediately after washing PMN were exposed to PBS. non-DV. or DV for 20 min a t 25°C. with or without subsequent stimulation with FMLP (5 X 10-'M) without CB for 10 min. Fluorescence intensity was monitored -116.2 continuously on a Spex Fluorolog spectrofluorometer (Metuchen.N J ) with a n excitation wavelength of 355 nm and an emission wave- length of either 410 or 490 nm. Intracellular [Ca2+] was estimated -92 using thefollowing formula as described by Grynkiewicz et al. (26): ICs"] = Kd (R-R,,./R,-R)(Sfz/Sba): where R = the ratioof fluores- cence of 410/490. Kd = 250 nM. R,,, = 0.12. R,, = 2.6 and Sf2/Sb2 -66.2 = 2.01. CAMP levels. cAMP levels were measured in PMN (2 X lo7 cells/ ml) exposed to non-DV, DV. or buffer for 0 to 20 min a t 25'C with or without subsequent stimulation with FMLP (1 X 1O' M) without - CB for 0 to 1 min. At various times. aliquots of cells were removed and cAMP levels determined usinga RIA kit (New England Nuclear) -45 as described by Smolen et al. (27). This methodallowsreliable as measurements of cAMP with time intervals short as 5 s. Transmembrane potential. Membranepotential changes were monitored as described by Seligmann and Gallin (28)with a spectro- fluorometric technique utilizing the cyaninefluorescent dye. Di0C5(3)(Molecular Probes). PMN (5 X 10' cells/mi) were incubated -31 with Di0C5(3)(5 x M) at 25°C for 5 min with constant stirring: this timeperiod allowed the fluorescence intensity to reach a steady state. Cells were then exposed to PBS. non-DV. or DV for 20 min a t 25°C. with or without subsequent stimulation with FMLP (5 X lo-' c 4 M) without CB for 10 min. Fluorescence emission was monitored on a continuous recorder attachedto a Spex Fluorolog spectroflu- orometer with a n excitation wavelength of 470 nm and emission -21.5 wavelength of 495 nm. Statfstfcal eualuatlons. Statistical evaluations were done by us- c ing Student's t-test. RESULTS Protein phosphorylation in exposed to influenza PMN virus alone. To determine what effect exposure PMNof to virus alone has on protein phosphorylation in PMN, I - + L- +, ,- +, cells were loaded with 32P,and incubated with buffer, non-DV, or DV. PMN were also stimulated with FMLP or NI DVNon-DV Flgure 1 . Phosphorylation ofPMN cytosolic proteins in cells exposed PMA within each experimentas a positive control. Cells to influenza virus without or with subsequent stimulation with FMLP. were then fractionated. and cytosolic and particulate PMN were prelabeled with"P, treated with buffer, non-DV. or DV for 30 min and then wlth I - M FMLP or solvent (DMSO)for 30 s. and fraction- O' fractions analyzed by SDS-PAGE. autoradiography, and ated a s described in Materlals and Methods. An autoradiogram from a scanning densitometry. Figure 1 is a n autoradiogram representative ( n = 3)SDS-PAGE of PMN cytosollcproteins is shown. An from a representative experiment, showing rg of protein was analyzed in each gel lane. NI. the effect of aliquot of 50noninfected [buffer-treated]PMN: Non-DV. cytosolscytosols derived from derived influenza virus alone on cytosolic phosphoproteins. Ex- from non-DV-treatedPMN; DV. cytosols derived from DV-treated cells: -. PMN: +. cytosols fromcells posure of cells to bothnon-DV (third lane) and (fijth cytosols fromsolvent [DMSO)-treated control DV stimulated with IO-'M FMLP. M, markers, to the right. are in kilodaltons. lane)-stimulated phosphorylationof some proteins when Arrowhead shows the position of the 48 kDa phosphoproteln. compared with cytosols from buffer-treated cells (first lane). Protein phosphorylation o influenza virus-treated f Superimposed densitometry scans of thefirst and third PMN induced by secondary stimulation with FMLP. The lanes of Figure 1. comparing buffer and non-DV treated autoradiogram in Figure 1 also illustrates the effect on cytosolic fractions (Fig. 2A). show that non-DV induced protein phosphorylation of PMN exposed to influenza phosphorylation of several mid to high m.w. proteins virus followed by stimulation with M FMLP. Com- (primarily those peaks on the right half of the scans). parison of each set of cells before and after FMLP stim- Figure 2B shows scansof cytosols from buffer- and DV- ulation shows that the agonist induces similar, though treated PMN in Figure 1 (first and fifth lanes). The scansnot identical, patterns of phosphorylation in buffer and clearly show DV also causes phosphorylation of several non-DV-treated PMN cytosols (second andfourth lanes). proteins. The differences in phosphorylation induced by This effect is also shown in densitometryscans (of Fig. non-DV and DV are shown in Figure which compares 1) in Figure 3A. (cytosols from buffer-treated cells with- 2C, scans of the representative autoradiogram lanes (Fig. 1. out and with FMLP stimulation). PMN pretreated with third andfijth lanes). Although the scans are nearly non-DV. with or without subsequent FMLP stimulation, identical in the low m.w.region of the autoradiogram (left showed equivalent or increased protein phosphorylation end of the scans), some mid to high m.w. proteins are as compared to buffer-treated cells (Fig. 3 B ) . In contrast, phosphorylated more intensely in PMN exposed to DV phosphorylation of cytosolic proteins in FMLP-stimulated compared with non-DV (Figs. 1 and 2 ) . Similar results cells was partially inhibited pretreatment of cells with by were obtained with particulate fractions: Le.. both DV DV (Fig. 3C). One of the proteins with inhibited phos- and non-DV caused phosphorylation of some proteins in phorylation in DV-treated PMN cytosols (arrows, Fig. 3 ) particulate fractions (data not shown). was the 48-kDa phosphoprotein. Similar overall effects
  • 21. VIRUS-INDUCED CHANGES IN NEUTROPHILPROTEIN PHOSPHORYLATION 3563 A A s C (d e 0 () I D d p i L I C C F f g u r e2. Effect of influenza virus alone on phosphorylation of cyto- Figure 3. Effect of FMLP stimulation on phosphorylation of cytosolic sollc proteins in unstimulated PMN. Densitometry scans of the first, third, proteins in PMN exposed to influenza virus. Densltometry scans of all and fifth lanes of Figure 1 are superimposed to show the effect of virus lanes in Figure 1 are superimposed to illustrate the effect of FMLP on phosphorylation. A , cytosols from PMN exposed to buffer (-) are stimulation on phosphorylation of cytosolic proteins. A , Cytosols from compared with those from PMN exposed to non-DV (- - - - -): B, cytosols buffer-treated PMN without ()- and with [- - - - -) FMLP stimulation: B. from PMN exposed to buffer (-) are compared with those from PMN cytosols from non-DV treated PMN without (-) and with (- - - - - ) treated with DV (- - - - -); C. cytosols from non-DV-treated PMN (-) are FMLP treatment; C . cytosols from DV-treated PMN without (-)and with compared with those from DV-treated cells (- - - - -1. Top. top of autoradi- (- - - - -1 FMLP treatment. Top. top of autoradtogram. Arrows show the ogram. Arrows show the position of the 48 kDa protein. position of the 48 kDa phosphoprotein. ofFMLP stimulation were observed in particulate frac- PMN (second and fourth lanes). contrast, and in In agree- tions (data not shown). ment with the results described above with FMLP stim- Protein phosphorylation o influenza virus-treated f ulation, phosphorylation was depressed after PMA treat- PMN induced by secondaryStimulation with PMA. The ment only in cytosolic proteins from DV-treated PMN. inhibition of protein phosphorylation in DV-treated PMN Again, one of the phosphoproteins with inhibited phos- upon FMLP stimulation was a striking observation, and phorylation was the 48-kDa protein (arrowhead).These therefore other studies were done to determine whether effects are illustrated clearly in thesuperimposed densi- a different secondary stimulus (Le., PMA) caused a simi- tometry scans shown in Figure 5 (scans derived from lar effect. A representative autoradiogram of cytosolic autoradiogram in Fig. 4).The degree of inhibition of phosphoproteins from buffer, non-DV, and DV-treated phosphorylation in DV-treated PMN cytosols in this ex- PMN without and with 100 ng/ml PMA stimulation is periment (Fig. 5 C ) was even more dramatic than that shown inFigure 4. As with FMLP treatment, the patterns observed in the experiment described in Figure 1, in and degree of PMA-induced phosphorylation were similar which FMLP was the stimulus.Indeed, with PMA stimu- for cytosolic proteins from buffer and non-DV-treated lation, very few proteins were phosphorylated above rest-
  • 22. -200 A A -116.2 -92 -66.2 15 31 n B nI -21.5 -14.4 - + - + - + " U NI Non-DV DV FLgure 4. Phosphorylatlon of PMN cytosolic proteins after exposure of cells to influenzavlrus and subsequent stimulation with PMA. PMN were prelabeled. stlmulated wlth 100 ng/ml PMA for 30 s. and analyzed as described in the legend to Figure 1. An autoradiogram from a represent- ative experiment I n = 3)is shown. An allquot of 46 fig of protein was analyzed in each gel lane. NI. Cytosols from noninfected (buffer-treated] PMN: Non-DV. cytosols from non-DV-treated PMN: DV. cytosols from DV- -. treated cells: cytosols from control PMN treated wlth solvent (DMSO): +. cytosols from PMA-stlmulated cells. M. markers. to the rlght. are in kllodaltons. Arrow shows theposition of the 48 kDa phosphoproteln. ing levels in DV-treated PMN. Particulate fractions exhib- ited similar changes in phosphorylation with treat- PMA ment (data not shown). Figure 5. Effect of PMA stlmulatlon on phosphorylation of PMN cyto- Studies were done to examinewhether the differences sollc proteins following exposure to influenza virus. Densitometry scans noted in the amount of phosphorylation of proteins in of the autoradiogram Inexposed to buffer wlthout 1 ease of comparison. A. Cytosols from PMN Flgure 4 are superimposed for - ( and with (- - - - -) DV- vs non-DV-treated PMN. with or without secondary PMA stlmulatlon: E. cytosols from non-DV-treated PMN without (-) stimulation, was due to different amounts of radioactivity and with (- - - - -1 PMA treatment: C. cytosols from DV-treated PMN with- out () - and with (- - - - -) PMA treatment. Top, top of autoradiogram. in thevarious cell preparations. These studies madeuse Arrows show the positlon of the 48 kDa phosphoprotein. of our observation that in all the gel experiments (>20) that we have done using the method described in this Intracellular [Ca2']. The effect of influenza virus,with paper, there.had been several proteins that appeared to or without subsequent stimulation with FMLP. on intra- have constant amountsof radioactivity regardless of the cellular [Ca"'] in PMN was measured using the fluores- stimuli the cells had been exposed to. Scans doneon the cent probe, Indo 1. The peak increase in intracellular autoradiograms of these particular proteins revealed that [Ca"] occurred by 10 s in PMN exposed to DV or non-DV: the area under the curve was similar in a given gel for Ca2+levels returned to base line by 20 min. When PMN each of the conditions (data notshown). Thesedata sug- were exposed to virus for 20 min and subsequently stim- gest that differencesin 32P, uptake into the cell and ulated with FMLP for 10 min. the maximal increase in loading onto the various lanes of the gel does not account intracellular [Ca"] occurred by 10 sec. andwas still for the differences in protein phosphorylation in DV- greater than control levels at 15 min. treated versus non-DV-treated PMN. There wasa significant ( p< 0.01) increase in intracel-
  • 23. VIRUS-INDUCED CHANGES IN NEUTROPHIL PROTEIN PHOSPHORYLATION 3565 lular [Ca'+] in PMN treated with DV or non-DV over that observed in buffer-treatedcells (Table I). When PMN were preincubated with DV for 20 min and then stimulated with FMLP, intracellular [Ca"] increased significantly( p < 0.01) over levels seen in non-DV or buffer-treated PMN. Theestimatedpeakintracellular [Ca"'], calculated by using the formulaof Grynkiewicz et al.(261 is shown for PMN with or without virusand FMLP treatment [Table I). When EGTA was preincubated with PMN before treat- ment with virus or FMLP, the immediate [( 10 s) rise in intracellular [Ca"] upon subsequent exposure to virus alone or after subsequent stimulation with FMLP was decreased 15 to 25%. cAMP levels. Both DV and non-DV caused very little increase in cAMP levels in PMN, as measured with RIA. cAMP concentrations in response FMLP stimulation of to J ~ m e( m i d PMN preincubated with virus or buffer for 20 min was Figure 6. Transmembrane potential changes in buffer and influenza measured at 0 , 5 . 10, 15.20.30 and60 s. The maximum virus-treatedPMN before and after FMLP stimulation. The kinetics of the transmembrane potential changes in PMN exposed to buffer [A-A), cAMP level occurred within 15 s, and there was no sig- non-DV 1 -. ( or DV (a"0) for 20 min followed by FMLP treatment nificantdifferencebetweencontrol and virus-treated for 10 mln are shown. DiOC,(3) was used to measure transmembrane cells at anyof the time points (data shown). not potential changes as described in the Materlals and Methods. Data are Transmembrane potential. The effectof influenza vi- the mean f SEM (bars)from four experiments. At 0 . 1non-DV ( p < 0.05 cence intensity was greater for DV than for buffer or min. the fluores- rus or buffer on transmembrane potential ofPMN was and 0.07.respectively).No significant differences were notedat any other measured by monitoring fluorescence intensity in PMN time point. by using DiOC5(3) (28).The peak changes in fluorescence intensity induced by non-DV or DV occurred within 10 s tein phosphorylation upon stimulation with either ago- and returned to base line by 1 min (Fig. 6). The peak nist. To date, thebiophysical difference between DV and changeinfluorescenceintensity was greater in PMN non-DV has not been elucidated, however, the use of exposed to DV than to non-DV or buffer. However, no non-DV in these experiments provides a useful control differences in the fluorescence intensity changes were fordetermining if the effect of influenzaviruson a seen when PMN were incubated with non-DV, DV, or particular PMN activation step might be responsible for buffer for 20 min and then stimulated with FMLP for 10 the depression of end-stage functions. For example, if min. the changes in protein phosphorylation that occurred upon secondary stimulation DV and non-DV pretreated of DISCUSSION PMN were identical, then a cause-effect relationship be- tweenaltered protein phosphorylation and depressed To our knowledge, this report constitutes the first ob- end-stage PMN functions would be very unlikely. Al- servation of early (i.e-, 30-min postexposure) alterations though DV-induced inhibition of phosphorylation corre- in cellular protein phosphorylation induced in immune lates nicely with depression of PMN end-stage functions, cells by viruses. In experiments described here, several a direct cause-effect relationship remains to proven. be proteins underwent phosphorylation (few, if any, were It is possible that DV alters phosphorylation of certain dephosphorylated] when PMN were exposed to non-DV or regulatory proteins, which could then suppress cellular DV alone for 30 min. Subsequent stimulation of these responses to subsequent stimulation. example, virus- For cells with FMLP or PMA resulted in similar patterns of induced changes in phosphorylation of N, proteins could phosphorylation in buffer and non-DV-treated PMN. In cause inhibition of end-stage responses similar to that contrast, DV-treated PMN showed inhibited levels of pro- seen in DV-treated cells. Specifically, studies have shown TABLE I that pertussis toxin, which inhibits N, protein in PMN, the Peak Lntracellular [Ca2+jin resting and stlmulatedP M N causes depression of FMLP-induced chemotactic, secre- Stimulant Fluorescence tory, and oxidative functions in PMN, while not inhibitlng Ratio at 10 Peak [Caz+lb phagocytosis (29-34). DV also inhibits chemotactic, se- Sa (nM) 1S t 2nd 410/490 nm cretory, and oxidative functions in vitro, but does not Buffer None 0.8 f 0.10 190 alter the phagocytic response (3-7). Inhibited phos- DV None 1.3 ? 0.06' 456 phorylation of N,-type proteins is unlikely to be the com- Non-DV None 1.2 f 0.06' 388 plete explanation for depression of end-stage functions Buffer FMLP 1 . 3 f 0.05 388 by DV, because the depressed phosphorylation occurs DV FMLP 1.5 f 0.05d 630 FMLP Non-DV 1.3 f 0.06 388 both with FMLP and PMA treatment, whereas FMLP "Peak rises in intracellular [Ca2'1 occurred at the 10-s time point. (reviewed in Ref. 35).but not PMA (29, 33, 36). has been Incubation conditions and methods for measuring intracellular [Ca"+-] associated withN, protein-mediated responses. Addition- were as described in Materials and Methods. Data are the means f SEM ally, the role of phosphorylation in controlling N, protein- for four experiments. Peak [Ca2'1 was estimated by using the formula described by Gryn- mediated steps remains speculative. kfewicz et al. (26). The 48-kDa phosphoprotein is another possible regu- 'p < 0.01 when cells treated with DV are compared with non-DV or latory protein through which DV could interfere with buffer. p < 0.01 when cells treated with DV plus FMLP are compared with PMN end-stage functions, particularly actlvation of the non-DV or buffer plus FMLP. respiratory burst. (We can identify this protein in our
  • 24. 3566 VIRUS-INDUCED CHANGES IN NEUTROPHIL PROTEIN PHOSPHORYLATION autoradiograms because experiments recently of reported cells stimulated with FMLP (44, 45). The enhanced rise in which PMN from a CGD patients failed to phosphoryl- in intracellular [Ca"+] cells preincubated with DV was in ate this protein, both in a cell-free system and in intact not due to a higher base line [Ca"'], because cells stimu- PMN (see footnote 4).) The finding that phosphorylation lated with either virus have equivalent intracellular [Ca"'] of the 48-kDa protein in DV-treated cells, in response to by 20 min (Table I). The initial rise in intracellular [Ca"] FMLP or PMA treatment, is inhibited is particularly in- (i.e., the increase which occurs within the first 10 s) teresting, since this proteindefectively phosphorylated caused by virus alone or virus plusFMLP came mainly is by in PMN from CGD patients. Inasmuch as the only func- from internal sources, inasmuchas addition of EGTA to tional defect described in PMN from CGD patients is the the cell suspension caused only a small decrease in the inability to mount a respiratory burst, this phosphopro- rise of intracellular [Ca2+] upon cell stimulation. Similar tein has recently been regarded as potentially important findings have been reported for PMN stimulated with in regulation of oxidative metabolism in normal PMN. FMLP (45). Sawyer et al. (46) have recently shown that, However, a definitive role for the 48-kDa phosphoprotein, upon stimulation ofPMN with opsonized zymosan, there or any other phosphoprotein, in the regulation of a spe- are regional differences in the subcellular localization of cific PMN end-stage response remains to proven. be Ca2+.We plan to do similar studies to determine what Our data are consistent with the idea that this protein role, if any, the augmented rise in intracellular [Ca"'] has may participate in regulation of the respiratory burst. in DV-induced disruption of PMN functions. The evidence for this is that I ) exposure of PMN to either An increase in cAMP levels in response to stimulation virus stimulated some phosphorylation of this protein, ofPMN with FMLP does not appear to be a fundamental a n event accompanied by a respiratory burst (4, 5, 19); requirement for activating the cell (47), but may have a 2) non-DV-treated PMN, when stimulated by FMLP or role in modulating the cell's response to FMLP (47.48) PMA. underwent normal phosphorylation of this protein and in activating CAMP-dependent protein kinases. Ex- under conditions which also elicit normal respiratory posure ofPMN to virus alone had very little effect on burst activation (5, 37); and 3) in contrast, DV-treated cAMP levels. Upon subsequent stimulation of these cells PMN, upon stimulation,exhibitdepressedrespiratory with FMLP, both the kinetics and degree of increase in burst activity (3-5, 37) and also show decreased phos- cAMP levels were similar in buffer and virus-treated phorylation of this protein. The inhibited phosphoryla- PMN, suggesting that virus was not causing PMN dys- tion of multiple cellular proteins by DV suggests that: 1) function by altering CAMPlevels. DV may depress PMN end-stage functions through alter- PMN transmembrane potential changes occur before ations in phosphorylation of several regulatory proteins; end-stage functions, and are thought to be due to ionic and 2) the actual step(s) at which the virus interferes fluxes acrosscell membranes (28, 49). The initial to (0.1 with PMN functions occurs prior to protein phosphoryl- 0.2 min) change in transmembrane potential was greater ation. in PMN exposed to DV than in cells treated with non-DV The effect of DV on intracellular [Ca"], cAMP levels, or buffer. However, when cells preincubated with virus and transmembrane signaling was also examined, be- or buffer were subsequently stimulated with FMLP, the cause various studies suggest that these steps are impor- responses did not differ significantly. One explanation tant in regulating cellular protein phosphorylation (re- for the finding that caused a greater initial change in DV viewed in Refs. 38 and 39). PMN have both Ca2+-and transmembrane potential could bethat thevirus induces CAMP-dependent protein kinases which mediate phos- permeability changes in PMN membranes. A precedent phorylation of multiple proteins upon cell stimulation. for this is that Sendai virus,which is structurally similar For example, recent studies suggest that protein kinase to influenza virus, is known to cause leakage of intracel- C is involved in the selective phosphorylation of PMN lular metabolites from cells (50).Although influenza vi- proteins stimulated with FMLP or PMA (40). and that rus was not found to cause such membrane permeability Ca"' participates in the regulation of this kinase by in- changes (50),those studies were not done using PMN; ducing its translocation from the cytosol to the mem- thus, depending oncell type, permeability changes might brane (41, 42). The activation of protein kinase C and still be possible. Our own data using PMN preincubated other kinases appears to dependent upon transmem- with carboxyfluorescein diacetate or carboxy-4' ,5'-di- be brane signalling that occurs when a ligand binds to its methylfluorescein diacetate, followed by exposure to in- receptor (38,39). fluenza virus did not, however, show leakage of these Studies were done to determine if DV interferes with dyes out of the cell (our unpublished data). Regardless of the rise in intracellular [Ca"'], which occurs when PMN the mechanism by which DV induces the change trans- in are stimulated with FMLP (discussed inRefs. 43 and 44). membrane potential,the data from experiments by using The findingthat virus alone caused a rise in intracellular FMLP a s a secondary stimulus suggest DV is not depress- [ca"'] was not surprising, since influenza virus can ini- ing PMN end-stage functions by interfering with trans- tiate a n oxidative burst in PMN (4, 5, 19). However, the membrane potential changes. data showingincreasedintracellular [Ca"'] in FMLP- The dataobtained in this studyshow that both DV and stimulated cells preincubated withDV, as compared with non-DV induce protein phosphorylation in PMN, but that non-DV or buffer, were unexpected. Although it is pos- only DV suppresses subsequent phosphorylation upon sible that a marked increasein intracellular [Ca"+] could treatment with a secondary stimulus. Although cAMP inhibit cell activation, we are unaware evidence in the levels and transmembrane potential changes do not ap- of literature which supports this hypothesis. Additionally, pear to be causing this effect, the role of intracellular the estimated threefold increase of intracellular [Ca"] [Ca"'] is still unclear. Further work is also needed to above base line in cells treated withDV followed byFMLP determine 1) which, if any, of the proteins which undergo is within the range reported by other investigators for altered phosphorylation are important in causing abnor-
  • 25. VIRUS-INDUCED CHANGES IN NEUTROPHIL PROTEIN PHOSPHORYLATION 3567 mal PMN end-stage responses, and 2) the mechanism by tection of 32Pand lZ51 using intensifying screens and hypersensitized film. F E E S Lett. 82:314. which DV inhibits cellular protein phosphorylation. 26. Grynkiewicz. G ., M. Poenie, and R. Y.Tsien. 1985. A new generation of Ca2+ indicators with greatly improved fluorescence properties. J . Biol. Chem. 260:3440. Acknow'edgments* We thank Drs' David Bass and 27. Smolen. J. E.,H. M. Korchak, and G . Weissman. 1980.Increased Douglas Lylesfor helpful discussions, Mrs. Jannell Rowe levels of cyclic adenosine-3',5'-monophosphate in human polymor- and Mrs. Lisa Winkler for technicalassistance.and phonuclearleukocytesaftersurfacestimulation. J . Clin.Invest. 65: 1077. MrsSandy Alexander for secretarial assistance. 28. Seligmann. B. E ,and J. I. Gallin. 1980. Use of lipophilic probes of . membrane potential to assess human neutrophil activation. Abnor- REFERENCES mality in chronic granulomatous disease. CIin. Invest. 66:493. J. 29. Becker, E L.. J. C. Kermode, P. H. Naccache, R. Yassin. M. L. . 1 Larson, H. E ,and R. Blades. 1976. Impairmentof human polymor- . Marsh, J. J. Munoz, and R. I. Sha'afi. 1985.Theinhibition of phonuclear leucocyte function by influenza virus. Lancet 1:283. neutrophil granule enzyme secretion and chemotaxis by pertussis 2 . Ruutu. P., A. Vaheri, and T. U. Kosunen. 1977. Depression of toxin. J . Cell Biol. 100:1641. human neutrophil motility by influenza virus in vitro. Scand. J . " 30. Lad. P. M.. C. V. Olson, I. S . Grewal, and S . J. Scott. 1985. A Immunol. 6~897. pertussis toxin-sensitive GTP-binding protein in the human neutro- 3. Abramson, J. S . . E. L. Mills, G . S . Giebink, and P. G . Quie. 1982 phil regulates multiple receptors, calcium mobilization, and lectin- Depression of monocyte and polymorphonuclear leukocyteoxidative induced capping.Proc. Natl. Acad. Sci. USA 82:8643. metabolism and bactericidal capacity by influenza A virus. Infect. 3 1. Krause, K.-H., W. Schlegel, C. B. Wollheim. T. Anderson. F. A. Immun. 35:350. Waldvogel. and P. D. Lew. 1985. Chemotactic peptide activation of 4. Abramson, J. S., J. C.Lewis. D. S . Lyles. K. A. Heller, E. L. Mills. human neutrophils andHL-60 cells. Pertussis toxin reveals correla- and D A. Bass. 1982. Inhibitionof neutrophil lysosome-phagosome . tion between inositol trisphosphate generation, calcium ion tran- fusionassociatedwithinfluenzavirusinfection invitro: role in depressed bactericidal activity.J . Clin. Invest. 69:1393. sients. and cellular activation. . Clin. Invest.76: 1348. J 32. Verghese, M. W.. K. Fox, L. C. McPhail. and R. Snyderman. 1985. 5. Abramson, J. S.. D. S . Lyles, K. A. Heller, and D. A. Bass.1982. Influenza A virus induced polymorphonuclear leukocyte dysfunction. Chemoattractant-elicited alterations of CAMP levels in human poly- Infect. Immun. 37:794. morphonuclearleukocytesrequire a Ca2+-dependentmechanism 6. Abramson. J. S., J. W.Parce, J. C. Lewis. D. S. Lyles, E L. Mills.. which is independent of transmembrane activation of adenylate R. D. Nelson, and D. A. Bass. 1984. Characterization the effect of of cyclase. J. Biol. Chem. 260:6769. influenzavirusonpolymorphonuclearleukocytemembranere- 33. Barrowman. M. M., S . Cockcroft. and B. D. Gomperts. 1986. Two sponses. Blood 64: 131. roles for guanine nucleotides in the stimulus-secretion sequence of 7. Abramson. J. s., J. G. Wheeler, J. W. Parce. M. J. Rowe, D. s. neutrophils. Nature 319:504. Lyles, M. Seeds, and D.A. Bass. 1986. Suppression of endocytosis 34. Lad. P. M.,C. V. Olson. and I. S . Grewal. 1986. A step sensitive to in neutrophils by influenza A virus in vitro.J.Infect. Dis. 154:456. pertussis toxin and phobol ester in human neutrophils regulates 8. Abramson. J. S , F. Cassidy, and D. S Lyles. 1987. Synthesisof . L. . chemotaxis and capping but not phagocytosis. FEBS Lett. 200:91. influenza A virus and endogenous proteins in polymorphonuclear 35. Snyderman, R.. C. D. Smith, and M. W. Verghese. 1986. Model for leukocytes. Pediatr. Res. 21:308A [Abstr.). leukocyte regulation chemoattractant receptors. roles a guanine by of 9. Schneider, C., M. Zanetti, and D. Romeo. 1981. Surface-reactive nucleotide regulatory protein and. polyphosphoinositide metabolism. " . stimuli selectively increase protein phosphorylation in human neu- J . Leukocyte-Biol. 401785. trophils. F E E S Lett. 127:4. 36. Verghese, M. W.. C. D. Smith, and R. Snyderman. 1985. Potential 1 . Andrews, P. C., and B. M. Babior. 1983. Endogenous protein phos- 0 role for a guanine nucleotideregulatory piotein in chemoattractant phorylation by restingandactivatedhumanneutrophils. Blood receptor mediated polyphosphoinositide metabolism, Ca'+ mobiliza- 61:333. tion and cellular responses by leukocytes. Biochem. Biophys. Res. 11. Huang, C. K., J. M. Hill, Jr.. B.-J. Bormann. W. M. Mackin, and E . Commun. 127:450. L. Becker. 1983. Endogenous substrates for cyclic AMP-dependent 37. Abramson, J. S., G . L. Wiegand, and D. S . Lyles. 1985. Neuramin- and calcium-dependent protein phosphorylation in rabbit peritoneal idase activity is not the causeof influenza virus-induced neutrophil neutrophils. Biochim. Biophys. Acta 760: 126. dysfunction. J . Clin. Mlcrobtol. 22:129. 12. Andrews, P. C., and B. M. Babior. 1984. Phosphorylationof cytosolic 38. Nishizuka, Y. 1984. The role of protein kinase C in surface signal cell proteins by resting and activated human neutrophils. Blood 64:883. transduction and tumour promotion. Nature 308:693. 13. White, J. R.. C.-K. Huang, J. M. Hill, Jr.. P. H. Naccache. E. L. 39. Ashendel, C. L. 1985. The phorbol ester receptor: a phospholipid- Becker, and R. I. Sha'afi. 1984. Effect of phorbol 12-myristate 13- regulated protein kinase.Biochim. Biophys. Acta 822:219. acetate and its analogue 4-a-phorbol 12.13-didecanoate on protein 40. Pontremoli. S., E Melloni, F. Salamino, B. Sparatore. M. Michetti, . phosphorylation and lysosomal enzyme release in rabbit neutrophils. 0 Sacco, and B. L. Horecker. 1986. Phosphorylation of proteins in J.BioL Chem. 259:8605. human neutrophils activated with phorbol myristate acetateor with 14. Okamura. N.. S . Ohashi, N. Nagahisa. and S . Ishibashi. 1983. chemotactic factor. Arch. Biochem. Biophys.250:23. Changes in protein phosphorylation in guinea pig polymorphonu- 41. May, W. S.. Jr.. N. Sahyoune. M. Wolf, and P. Cuatrecasas. 1985. clearleukocytes by treatment with membrane-perturbing agents Role of intracellular calciummobilization in the regulationof mem- which stimulate superoxide anion production.Arch. Biochem. Bio- brane processes.Nature 31 7:549. p h y s . 228:270. 42. Wolf, M.. H. Levine 111. W. S . May, Jr.. P. Cuatrecasas, and N. 15. Segal, A. W.. P G . Heyworth, s. Cockcroft, and M. M. Barrowman. . 1985. Stimulated neutrophils from patients with autosomal recessive Sahyoune. 1985. A model for intracellular translocation of proteln chronic granulomatous disease fail to phosphorylate a M,-44,000 kinaseC involving synergism between CaZ' and phorbol esters. protein. Nature 316:547. Nature 31 7:546. 16 Ishii, E , K. . Irita. I. Fujita, K. Takeshige, M. Kobayashi, T. Usui. and K. Ueda. 1986. Protein phosphorylation of neutrophils from 43. Pozzan, T.. D. P. Lew. C. B. Wollheim, and R. Y Tsien. 1983. Is . cytosolic ionized calcium regulatlng neutrophil activation? Science normal children and patients with chronic granulomatous disease. 221:1413. Eur. J. Pediatr. 1 4 5 ~ 2 2 . 44. White. J. R.. P. H. Naccache, T. F. P. Molski. P. Borgeat, and R. I. 17 Hayakawa, T., K. Suzuki. S. Suzuki, P. C. Andrews, and B. M. Sha'afi. 1983. Direct demonstration of increased intracellular con- Babior. 1986. A possible role forproteinphosphorylationin the centration of free calcium in rabbit and human neutrophils following activation of the respiratory burst in human neutrophils. Evidence stimulation by chemotactic factor. Biochem. Biophys. Res. Com- from studies with cells from patients with chronic granulomatous mun. 113:44. disease. J . Biol. Chem. 261:9109. 45. Sklar. L. A., and 2. G . Oades. 1985. Signal transductionand Hgand- 18. Heyworth, P. G.. and A. W. Segal. 1986. Further evidence for the receptor dynamics in the neutrophil. Ca2+modulation and restora- involvement of a phosphoprotein in the respiratory burst oxidase of tion. J . Biol. Chem. 260: 1 1468. human neutrophils. Bfochem. J. 239:723. 46. Sawyer, D W.. J. A. Sullivan, and G . L. Mandell. 1985. Intracellular . 19. Mills, E. L.. Y. Debets-Ossenkopp, H. A. Verbrugh, and J. Verhoef. free calciumlocalization in neutrophils during phagocytosis. Science 1981. Initiation of the respiratory burst of human neutrophils by 2303363. influenza vlrus. Infect. Immun. 32:1200. 47. Simchowitz, L.. I. Spilberg. and J.P. Atkinson. 1983. Evidence that 20. Lowry. D. H., N. J. Rosebrough, A. L. Farr, and R.J.Randall. 1951, Protein measurements with the Folin phenol reagent.J . Biol. Chem. the functional responses human neutrophils occur independently of of transient elevations in cyclic AMP levels. J . Cyclic Nucleotide 193:265. Protein Phosphor. Res.9:35. 21. Amrein. P. C.. and T. P. Stossel. 1980. Prevention of degradation of 48. Grady, P. G.. and L.L. Thomas. 1986. Characterization of cyclic- human polymorphonuclear leukocytes proteins by diisopropylfluo- nucleotide phosphodiesteraseactivitiesinrestingand N-formyl- rophosphate. Blood 46:442. methionyl-leucylphenylalanine-stimulated human neutrophils. 22. Peterson, G . L. 1977. A simplification of the protein assay method Biochfm. Biophys. Acta 885:282. of Lowry e t a l . which is more generally applicable. Anal. Biochem. 49. Kitagawa. S., and R. B. Johnston, J r . 1985. Relationship between 83:346. membrane potential changes and superoxide-releasing capacity in 23. Laemmli. U. K. 1970. Cleavage of structural proteins during assem- resident and activated mouse peritoneal macrophages. Immunol. J. bly of the headof bacteriophage T4.Nature 227:680. 135:3417. 24. Wray. W.. T. Boulikas, V. P. Wray, and R. Hancock. 1981. Silver 50. Foster, K. A.. K. Gill, K. J. Micklem, and C. A. Pasternak. 1980. staining of proteins in polyacrylamide gels. Anal. Biochem.1 18:197. Survey of virally mediated permeability changes. Biochem. J . 25. Laskey, R., and A. D. Mills. 1977. Enhanced autoradiographic de- 190:639.
  • 26. THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 256, No. 10, Issue of May 25, pp. 4838-4842, 1981 Printed in U. S.A. Interaction of Sendai Virus Proteins with the Cytoplasmic Surface of Erythrocyte Membranes following Viral Envelope Fusion* (Received for publication, October 30, 1980) Susan E. Caldwell and Douglas S. Lyles From the Department of Microbiology and Immunology, The Bowman Gray School ofMedicine of Wake Forest University, Winston-Salem, North Carolina 27103 Radiolabeled Sendai virus was allowed to fuse with brane or matrix protein, which coats the inner surface of the human erythrocytes and inside out vesicles were pre- viral envelope. The viral nucleocapsid proteins include NP, pared from the erythrocyte membranes. Viral proteins the major nucleocapsid protein, and P and L, probably sub- incorporated into inside out vesicles were released units of the virion RNA polymerase (nomenclature discussed from the membrane following treatment with various in Scheid and Choppin, 1974, 1977). agents which perturb protein structure. Disruption Sendai virus fuses readily with erythrocytes, thus incorpo- products were analyzed by sucrose gradient centrifu- rating viral proteins into the erythrocyte membrane (Bachi et gation and sodium dodecyl sulfate-polyacrylamide gel al., 1973; Buechi and Bachi, 1979). Virus envelope fusion with electrophoresis to identify the viral proteins. The re- erythrocyte membranes provides a model for virus penetra- sults indicated that 6 M guanidine, 40 mm lithium diio- tion. The advantage in using erythrocyte membranes is that dosalicylate (pH 11-13), and 4 M potassium thiocyanate they can be made into inside out vesicles containing viral Downloaded from www.jbc.org by on May 29, 2009 removed viral nucleocapsid and M proteins from the vesicles. Urea (6.0 M) and 5 mi p-chloromercuriben- proteins. These IO1 vesicles have all of the internal viral zenesulfonate removed only viral nucleocapsid pro- proteins on their external (cytoplasmic) surface (Lyles, 1979), teins. KCI (1.0 or 0.1 M), 5 or 0.5 mm sodium phosphate, where the mechanisms of membrane attachment of these 10 mm EDTA (pH 10), 10 mm lithium diiodosalicylate, 1 proteins can be investigated. The associations characterized M urea, 0.5 mw ATP, or freeze-thaw cycles did not in the erythrocyte model should also be relevant to the release any viral proteins. By contrast, treatment of mechanism of virus maturation by budding from the host inside out vesicles with Triton X-100 solubilizes the plasma membrane. In the process of budding, the viral gly- lipid bilayer, releasing viral integral membrane glyco- coproteins are inserted into the host membrane biosyntheti- proteins and leaving viral nucleocapsids intact. Nucleo- cally (rather than by fusion); the nucleocapsid then associates capsids isolated from virions do not adsorb nonspecif- with areas of membrane containing viral envelope components ically to control inside out vesicles. These data imply prior to evagination of the membrane to form the mature that Sendai viral nueleocapsid and M proteins interact virion (Choppin and Compans, 1975). extensively with the cytoplasmic surface of infected The forces binding proteins to cell membranes can be cell membranes, in the manner of peripheral membrane characterized in terms of the difficulty of removal of such proteins, whereas the glycoproteins, HN and F1, behave proteins from the membranes. Membranes may be treated as integral membrane proteins. with various agents which perturb membrane structure but leave the lipid bilayer relatively intact. The proteins removed from the membranes by these agents are functionally defined Sendai virus, a parainfluenza virus, penetrates into the host as peripheral membrane proteins. This technique has been cell by a mechanism which probably involves fusion of the developed by Steck and others in studies of erythrocyte mem- viral envelope with the membrane of the host cell (discussed brane proteins (reviewed by Steck, 1974a). Some peripheral in Scheid and Choppin, 1976). Sendai virus has two glycopro- proteins are easily eluted, some are more difficult to elute, and teins on the surface of the envelope. One, designated HN, the integral membrane proteins do not elute from the mem- possesses both hemagglutinating and neuraminidase activities brane without disrupting the lipid bilayer. In the present and allows the virus to attach to receptors on the host cell work, it is shown that the Sendai virus nucleocapsid and M membrane (Tozawa et al., 1973;, Scheid and Choppin, 1974). proteins behave as peripheral membrane proteins in the eryth- The other glycoprotein, F, enables the attached virus to fuse rocyte membrane, while HN and F behave as integral mem- with the host cell (Homma and Ohuchi, 1973; Scheid and brane proteins. Choppin, 1974). Virions grown in most tissue culture cells MATERIALS AND METHODS contain F in the form of an inactive precursor, F0, which can Virus and Cells-The Z strain of Sendai virus was grown in MDBK be activated in vitro by trypsinization. Sendai virus also cells with Dulbecco's minimal essential medium supplemented with contains a nonglycosylated envelope protein, M, the mem- 10% newborn calf serum (Flow Laboratories) in the presence of [35S]methionine (10 OCi/ml, New England Nuclear) (1 Ci = 3.7 x 1010 * This research was supported by a grant from the North Carolina Bq) and purified as described (Lamb and Choppin, 1977). The F. United Way and Forsyth Cancer Service and by Grant AI 15892 from glycoprotein was activated with tosylphenylalanyl chloromethyl ke- the National Institutes of Health. Preliminary results of this work tone-treated trypsin (5 ,ug/ml; trypsin-TPCK, Worthington) for 10 were presented at the 1980 meeting of the American Society for min at 37 'C as described (Scheid and Choppin, 1974). Microbiology (Caldwell, S., and Lyles, D. S. (1980) Abstr. Annu. Meet. Erythrocyte Membranes Containing Viral Proteins-The fusion Am. Soc. Microbiol. 265). The costs of publication of this article were of radiolabeled Sendai virus with human erythrocytes and the pro- defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 'The abbreviations used are: IO, inside out; SDS, sodium dodecyl U.S.C. Section 1734 solely to indicate this fact. sulfate. 4838
  • 27. Sendai Virus Protein Interaction with Inside Out Vesicles duction of IO vesicles from these membranes has been described in detail (Lyles, 1979) and is briefly discussed under "Results." Elution Procedure-IO vesicles containing viral proteins were pelleted for 1 h at 15,000 rpm in a Sorvall SS 34 rotor. Each resulting pellet was resuspended in 0.5 ml of a given eluting agent in 5 mm sodium phosphate, pH 8.0 (58P buffer), unless otherwise specified, and incubated on ice for 30 min. The preparation was then further diluted with 0.5 ml of buffer to reduce the density of concentrated solutions and was layered over a linear gradient of 15-45% (w/w) sucrose over a 60% sucrose cushion, all in 10 mM phosphate buffer, pH 7.4. The gradients were centrifuged at 40,000 rpm for 2 h in an SW 50L rotor. Each gradient was then fractionated in 20-drop frac- tions; 25-p1l aliquots of each fraction were counted by liquid scintilla- tion counting and peak fractions were pooled. These pooled fractions were precipitated with 10% (w/w) trichloroacetic acid (Fisher) after 100 pl of bovine serum albumin (1 mg/ml) was added to each as a carrier. Samples were spun at 2,000 rpm for 30 min and then washed with 10% (w/w) trichloroacetic acid and finally with butanol (Fisher). Samples were evaporated to dryness under a thin stream of nitrogen. Nucleocapsid Purification and Association with IO Vesicles- Radiolabeled virus was added to 2% Triton X-100 in 10 mM sodium phosphate (pH 7.4) and potassium chloride (0.0, 0.1, or 1.0 M) and incubated on ice for 30 min. The preparations were then spun through discontinuous gradients consisting of 2 ml each of 15 and 45% sucrose and 0.3 ml of 60% sucrose (all w/w) in an SW 50L rotor at 40,000 rpm for 2 h. The bottom 1.0 ml of each gradient was collected and dialyzed overnight in 10 mm phosphate buffer, pH 7.4. Nucleocapsid preparations (0.75-3.0 ml) were then added to a 3- 2- 68 26 .cl 0, 0 9 A 25. 20- A 5. 6 -1 /t 0 6 6 *0 Q... 0 0 0.. 0 4839 o..... 0 Downloaded from www.jbc.org by on May 29, 2009 pellet of IO vesicles which did not contain viral protein. The mixture was incubated on ice 5 min and then spun through a linear 15-45% 4 ~~~~~~~~4 (w/w) sucrose gradient with a 60% sucrose cushion for 2.5 h at 40,000 rpm in an SW 50L rotor. Gradients were fractionated in 20-drop 3 5 7 FRACTION NO. 9 11 13 top 3 5 7 9 FRACTION NO. I, 13 top fractions; 25-p1 aliquots were counted by liquid scintiUation counting, and peak fractions were pooled and precipitated with trichloroacetic FIG. 1. Sedimentation analysis of disruption products from acid as described for the elution procedure. SDS-polyacrylamide gel inside out vesicles containing Sendai viral proteins. Inside out electrophoresis was performed with the samples as described below. vesicles containing Sendai viral proteins labeled with [nS]methionine SDS-Polyacrylamide Gel Electrophoresis-The pooled, trichlo- were incubated in the indicated solutions and centrifuged on a sucrose roacetic acid-precipitated samples were electrophoresed on 7.5% poly- gradient as described under "Materials and Methods." The gradients acrylamide slab gels containing SDS as described (Laemrnli, 1970). were collected from the bottom. The radioactivity of an aliquot of The gels were processed for fluorography as described (Bonner and each fraction was determined by liquid scintillation counting and was Laskey, 1974) and exposed to Kodak SB-5 x-ray film. plotted on the ordinate. O-- -0, control vesicles incubated in 5 mM sodium phosphate, pH 8.0; * -*, vesicles incubated in: A, 1.0 M RESULTS KCI; B, 5 mm sodium phosphate, then frozen and thawed 3 times; C, 4 M KSCN; D, 40 mM lithium diiodosalicylate; E, 5 mM p-chloromer- Radioactive Sendai virus was allowed to fuse with human curibenzenesulfonate; F, 6 M urea. erythrocytes, thus incorporating viral proteins into the cell membranes. The erythrocytes were lysed in 5 mi sodium agents tested, none released rapidly sedimenting material such phosphate, pH 8.0. IO vesicles were prepared by incubation as, for example, intact viral nucleocapsids. in low ionic strength buffer, which causes the membrane to Peak fractions from Fig. 1 were pooled and analyzed by invaginate, followed by gentle homogenization to generate IO SDS-polyacrylamide gel electrophoresis with fluorography to vesicles from the invaginations. The IO vesicles were purified detect the labeled viral proteins, as shown in Figs. 2 and 3. from unsealed membrane fragments by centrifugation on dex- Note that, as shown previously, each viral structural protein tran gradients, which separates membranes based on their is present in control 10 vesicles. In addition, a proteolytic osmotic properties (Steck, 1974a, 1974b). These vesicles have fragment of NP, designated NP', is occasionally observed both been shown to have the internal viral proteins on their exter- in virions and in IO vesicles, presumably due to low levels of nal (cytoplasmic) surface (Lyles, 1979). The extent of associ- endogenous proteases in the preparation (Lyles, 1979). As ation of these viral proteins with the vesicle membrane can be shown in Fig. 2, vesicles treated with 40 mm lithium diiodo- examined by treating vesicles with a variety of perturbants salicylate, 6 M guanidine, 4 M KSCN and high pH (pH 12-13) which disrupt membrane protein structure, leaving the lipid had little nucleocapsid (NP) or M proteins, which were found bilayer intact. The IO vesicles were treated with various in the top fractions of the gradient. Viral glycoproteins (HN eluting agents, and the disruption products were analyzed by and F1) remained associated with the vesicle fraction after sedimentation in sucrose gradients (Fig. 1). treatment, as is typical of integral membrane proteins. Treat- Some treatments which elute erythrocyte membrane pro- ment of vesicle preparations with high pH could conceivably teins (Steck and Yu, 1973) did not elute viral proteins (Fig. 1, hydrolyze phospholipids in the membrane; however, Steck A and B), such as 1 M KCI or 3 freeze-thaw cycles in 5 mM and Yu (1973) have shown that this does not occur with brief sodium phosphate, pH 8.0. JO vesicles treated with these treatments at low temperature. agents sedimented to density equilibrium in the middle of the Electrophoretic analysis of IO vesicles containing viral pro- gradient similar to the control, untreated vesicles. Other re- teins treated with 5 mmp-chloromercuribenzenesulfonate and agents, such as 4 M potassium thiocyanate, 40 mm lithium 6 M urea is shown in Fig. 3. These agents removed only viral diiodosalicylate, 6 M urea, and 5 mM p-chloromercuriben- nucleocapsid proteins, leaving the M protein associated with zenesulfonate eluted viral proteins from the membrane (Fig. the vesicle fraction. Lower concentrations of reagents de- 1, C-F). Slowly sedimenting proteins released from the mem- scribed in Figs. 2 and 3, such as 10 mM lithium diiodosalicylate, brane remained at the top of the gradient, while other proteins 1 M urea, or pH 10, did not elute any viral proteins under the remained associated with the vesicle fraction. Of all the re- same experimental conditions (data not shown). Electropho-
  • 28. 4840 Sendai Virus Protein Interaction uith Inside Out Vesicles 4. -44 ,-- NP- __ am_, S ow** I h Fi Downloaded from www.jbc.org by on May 29, 2009 F0- *#4 NP- - F- NP I;,_ (_ S no 40o _W 09 m6 ^ - 40 *-t X S be'. P 95 *:~ S^ C V T V T V T V T C V T V T V T GUAN KSCN LIS pH 12 pH 13 PCMBS UREA FIG. 2 (left). Reagents that remove both nucleocapsid and m 13, 0.1 N NaOH. proteins from inside out vesicles containing Sendai viral pro- FIG. 3 (right). Reagents which selectively remove the nu- teins. Inside out xesicles containing Sendai viral proteins labeled cleocapsid from inside out vesicles containing Sendai viral with [PSjmethionine were incubated with the indicated reagents, and proteins. Inside out vesicles containing Sendai viral proteins labeled subjected to sucrose gradient centrifugation as in Fig. 1. Peak fractions with [ 8]methionine were incubated with the indicated reagents and were pooled, acid-precipitated, and subjected to SDS-polxacrllam,,lde suhjeeted to sucrose gradient centrifugation. P-eak fractions were gel electrophoresis. Fluorographs of the dried gels are shown. C, pooled, precipitated, and subjected to SDS-polvacrvlamide gel elec- control vesicles incubated in 5 mM sodium phosphate, pH 8.0. V, trophoresis as in Fig. 2. Fluorographs of the dried gels are shown. V, pooled fractions containing treated vesicles; T, pooled fractions near pooled fractions containing treated vesicles; T, pooled fractions near the top of the gradient; GUAN, 6 M guanidine; KSCN, 4 M KSCN; the top of the gradient; PCM1BS, 5 mM p-chloroniercuribenzenesul- LIS, 40 mM lithium diiodosalicylate; pH 12, 50 mM glycine, pH 12; pH fonate: lREA, 6 M urea. retic analvsis of vesicles treated with agents that do not release when similar 10 vesicle preparations are treated with Triton vliral proteins confirmed that the content of viral proteins in X-100, a nonionic detergent which solubilizes the lipid bilayer the vlesicles was unchanged compared to untreated vesicles. and leaves protein structure largely intact. Fig. 4A shows the Quantitation of the amount of NP or M remaining associated analysis of Triton-treated IO vesicles by sucrose gradient with the vesicle fraction following elution with the agents centrifugation. The fluorogram in Fig. 4B shows that the described in Figs. 2 and 3 by densitometry revealed that 70- integral membrane proteins HN and F, remain at the top of 90%2c of NP and M could be removed from the membrane. The the gradient, nucleocapsid proteins sediment to the bottom in remaining NP and M may be due to unfused virions trapped the form of intact nucleocapsids, and, depending upon the salt within the 10 vesicles and carried through the preparation concentration, M protein is distributed between the top and (Lyles, 1979). bottom fractions. This disruption pattern (Fig. 4) resembles The eluting agents shown in Figs. 2 and 3 removed periph- that obtained when intact virions are treated with Triton, as eral viral proteins from the vesicle membrane, leaving the described by Scheid et al. (1972). integral membrane glvcoproteins associated with the vesicle Control experiments tested whether viral nucleocapsids as- lipid bilayer. By contrast, complementarv results are observed sociate nonspecificallv with IO erythrocyte membranes. Nu-
  • 29. Sendai Virus Protein Interaction with In side Out Vesicles 4841 m ! rP ws a amide gel electrophoresis. Agents eluting peripheral proteins 1 from erythrocvte membranes and leaving the lipid bilayer intact (reviewed by Steck, 1974a) selectively remove the nu- cleocapsid and M proteins; the viral glvcoproteins remain associated with the lipid bilayer, as would be expected for integral membrane proteins (Figs. 2 and 3). Several treatments FEs 2- that remove peripheral proteins from erythrocyte membranes x p resulted in no elution of viral proteins, such as extremes of H N- l * ionic strength (0.5 mm sodium phosphate, 1.0 M KCl), 10 mM EDTA, and pH 10. Unexpectedly, I M KCl did not elute viral - @ o/.~ -m . Vw proteins despite its ability to dissociate M protein from nu- 5 ~ ~~~~~~ 10 NP- cleocapsids following detergent extraction of virions (Scheid 15 NP'I *ww et al., 1972). No elution of viral proteins was observed follow- 5 10 ing freezing and thawing of 10 vesicles, despite changes in- FRACTION fop ~- 0 * * * duced in the association of nucleocapsids with the viral en- velope by freezing and thawing (Homma et al., 1976; Kim et al. 1979). C T 1 The ability to remove the nucleocapsid and M proteins from 10 vesicles (Figs. 2 and 3) implies that their membrane FIG. 4. Analysis of disruption products fr inside out vesicles containing Sendai viral pro teins eaithmTnitof attachment involves the maintenance of protein structure. In X-100. Inside out vesicles containing Sendai virn al proteins labeled contrast, disruption of interactions occurring in the aqueous with [ 'S]methionine were incubated in the presen ce of 1V. Triton X- phase does not dissociate the interactions in the lipid phase 100 in 0.1 M KCl on ice for 5 min and centrifuged oin sucrose gradients that bind HN and F, to the membrane bilayer. This is also as described under "Materials and Methods." A, an alysis of disruption true of integral proteins of the erythrocvte membrane (Steck, Downloaded from www.jbc.org by on May 29, 2009 products bv sucrose gradient centrifugation. The gradients were cl lected from the bottom. The ordinate is a plot of t the radloactixvitv in 1974a). t is not known with what membrane components an aliquot of each fraction as determined bx liquid scintillation xiral proteins are associated on the cytoplasmic surface of the counting. O --, control vesicles incubated in imM 5 soedium pho)s- 10 vesicles. It appears that there is little or no nonspecific phate, pH 8.0; v-S,xesicles incubated in TrittoIi X-100. B, fluo- association of nucleocapsids with the cytoplasmic surface of rograph of dried gel is shown. SDS-polvacrxlamide gel electrophoresis the membrane since nucleocapsids isolated from virions under was performed on peak fractions which were pooLE ed and acid precip- a variety of conditions do not bind to J0 vesicles lacking v iral itated: C, control vesicles incubated in 5 mM sodi urn phosphate, pH 8.0; T, pooled fractions near the top of the gr-adient, V pooled proteins. An association between NP and M of Sendai virus fractions of treated vesicles; B, pooled fractions fi'rom the bottom of has been identified by treatment of intact virions with chem- the gradient. ical crosslinking reagents (Markwell and Fox, 1980). Such an association could account in part for the membrane interac- tions observed here. One or both of these components may cleocapsids were purified from intact viions and were added also interact with other proteins in the membrane, such as to JOvesicles which did not contain vira il proteins. The the cytoplasmically exposed portion of HN or F, (Lyles, 1979), mixture was then analyzed by sedimentatioi n in sucrose gra- as hypothesized for this and other virus types (Blough and dients to detect association of labeled nucleo )capsids with the Tiffanv, 1975; Simons and Garoff, 1980). In addition, these vesicle fraction. Nucleocapsids were preparec in the presence components could interact directil with the membrane lipids of 1 M KCl, 0.1 M KCl, or 10 mm phosphatE buffer, pH 7.4. in a conformation-dependent manner. We have not found The salt concentration influences the externt of M protein conditions which allow removal of intact nucleocapsids from association with nucleocapsids; 1 M KCI re,sults in little M the membranes without disruption of the lipid bilayer, as with being present, and low salt concentration resu ilts in nucleocap- Triton X-100. This suggests that the forces binding nucleo- sids containing greater amounts of M protei in (Scheid et al., capsids to membranes are similar in magnitude to those 1972). Nucleocapsid preparations were also added to similar responsible for nucleocapsid integrity. IO vesicles in different salt concentrations (I 0.0, 0.1, or 1.0 M Some intriguing possibilities are suggested by these results. KCl in 10 mM sodium phosphate) since the conformation of Since nucleocapsids remain associated with the membrane the nucleocapsid has been shown to be depe tndent upon salt after fusion of the viral envelope with the cell membrane, the concentration (Heggeness et al., 1980). Sucro se gradient anal- initial transcription events in virus replication may occur on ysis of nucleocapsids prepared in 1 M KCI or 1 L0 mM phosphate membrane-bound nucleocapsids or, alternatively, an addi- buffer which were added to 0 vesicles not containing viral tional step after viral envelope fusion may be necessary for proteins demonstrated that less than 10%8c of nucleocapsid releasing nucleocapsids into the cytoplasm. Our preliminary protein associated nonspecifically with IO v esicles (data not data favor the former possibility, suggesting that nucleocap- shown). sids remain membrane-bound following penetration of Sendai virus into cell types other than erythrocytes. 10 vesicles DISCUSSION containing viral proteins also provide a model for the struc- A model for the plasma membrane of Send lai virus-infected tural organization of the viral envelope. It has been suggested cells is generated by allowing the envelopes c :f mature virions recently (Li and Fox, 1980) that M is a peripheral membrane grown in the presence of [ 'S]methionine to fuse with eryth- protein in virus envelopes, based upon disruption of virions rocyte membranes. 10 vesicles, prepared firom such mem- with lithium diiodosalicylate. Likewise, in electron micro- branes, have viral nucleocapsid and M piroteins on their graphs, nucleocapsids are frequently observed to be periph- external (cytoplasmic) surface (Lyles, 1979) The extent of erallv associated with the viral envelope (Kim et al., 1979) association of these proteins with the vesidlEes was examined and with cell membranes following viral envelope fusion (Bue- by treatment with agents which dissociate pro teins from mem- chi and Bachi, 1979). Mechanisms of membrane attachment branes. The disruption products were analy ,zed by centrifu- similar to those studied here may also be involved in virus gation on sucrose gradients and identified by SDS-polvacryl- maturation by budding from host cell membranes, in which
  • 30. 4842 Sendai Virus Protein Interaction with Inside Out Vesicles nucleocapsids and M protein associate with the cytoplasmic Kim, J., Hama, K., Miyake, Y., and Okada, Y. (1979) Virology 95, surface of the host plasma membrane. 523-535 Laemmli, U. K. (1970) Nature (Lond.) 227, 680-685 REFERENCES Lamb, R. A., and Choppin, P. W. (1977) Virology 81, 382-397 Li, J. K.-K., Miyakawa, T., and Fox, C. F. (1980) J. Virol. 34, 268-271 Bachi, T., Aguet, M., and Howe, C. (1973) J. Virol. 11, 1004-1012 Lyles, D. S. (1979) Proc. Natl. Acad. Sci. U. S. A. 76, 5621-5625 Blough, H. A., and Tiffany, J. M. (1975) Curr. Top. Microbiol. Markwell, M. A. K., and Fox, C. F. (1980) J. Virol. 33, 152-166 Immunol. 70, 1-30 Scheid, A., Caliguiri, L. A., Compans, R. W., and Choppin, P. W. Bonner, W. M., and Laskey, R. A. (1974) Eur. J. Biochem. 46, 83-88 (1972) Virology 50, 640-652 Buechi, M., and Bachi, T. (1979) J. Cell Biol. 83, 338-347 Scheid, A., and Choppin, P. W. (1974) Virology 57, 475-490 Choppin, P. W., and Compans, R. W. (1975) in Comprehensive Scheid, A., and Choppin, P. W. (1976) Virology 69, 265-277 Virology (Fraenkel-Conrat, H., and Wagner, R. R., eds) Vol. 4, pp. Scheid, A., and Choppin, P. W. (1977) Virology 80, 54-66 95-178, Plenum Press, New York Simons, K., and Garoff, H. (1980) J. Gen. Virol. 50, 1-21 Heggeness, M. H., Scheid, A., and Choppin, P. W. (1980) Proc. Natl. Steck, T. L. (1974a) J. Cell Biol. 62, 1-19 Acad. Sci. U. S. A. 77, 2631-2635 Steck, T. L. (1974b) Methods Membr. Biol. 2, 245-281 Homma, M., and Ohuchi, M. (1973) J. Virol. 12, 1457-1465 Steck, T. L., and Yu, J. (1973) J. Supramol. Struct. 1, 220-232 Homma, M., Shimizu, K., Shimizu, Y. K., and Ishida, N. (1976) Tozawa, H., Watanabe, M., and Ishida, N. (1973) Virology 55, 242- Virology 71, 41-47 253 Downloaded from www.jbc.org by on May 29, 2009