Serpin 2a Is Induced in Activated Macrophages and Conjugates to a Ubiquitin Homolog

Cells and mice

RAW264.7 macrophages (American Type Culture Collection (ATCC), Manassas VA) were cultured in RPMI (BioWhittaker, Gaithersburg, MD) with 10% FCS (HyClone Laboratories, Logan, UT), glutamine, penicillin, and streptomycin (Life Technologies, Gaithersburg, MD) at 37^οC and 5% CO₂. Female ICR and C57BL/6 mice were purchased from Charles River Breeding Laboratories (Wilmington, MA).

Bacteria and infection

Mycobacterium bovis BCG (strain Pasteur; ATCC) was a gift from Dr. S. Smith (University of Washington, Seattle, WA). BCG was grown in Proskauer-Beck medium with aeration to 5 × 10⁷ CFU/ml and stored in aliquots at −70^οC. To infect mice, an aliquot of BCG was thawed, sonicated three times for 30 s in a water bath sonicator, and diluted in PBS. Mice were injected i.p. with 5 × 10⁶ CFU. Listeria monocytogenes, strain 10403S, was grown in trypticase soy broth (Difco, Detroit, MI). Listeria in log phase were diluted in PBS, and 1 × 10³ CFU were injected i.p. S. typhimurium SL3261 (attenuated DL1344 ΔaroAhisGsylrpsL) was a gift from Dr. B. Cookson (University of Washington). Salmonella growing at log phase in LB (Difco) were diluted in PBS, and 1 × 10⁵ CFU/mouse were injected i.p.

Macrophages and dendritic cells

Activated macrophages were harvested by peritoneal lavage with PBS 12 days after infection with BCG and 5 days after infection with Listeria or Salmonella. Resident peritoneal macrophages were harvested from uninfected ICR mice. Unless otherwise noted, macrophages were plate adhered for 2 h and then washed several times with PBS to remove nonadherent cells. Cells remaining were >95% macrophages by visual inspection. Bone marrow dendritic cells were generated by a modification of the method of Inaba et al. (15). Briefly, bone marrow cells from C57BL/6 mice were cultured with 20 ng/ml rGM-CSF (R&D Systems, Minneapolis, MN) for 7 days. Dendritic cells were then cultured in medium or in medium with 100 ng/ml LPS (List Biological Laboratories, Campbell, CA) for 24 h; stained with Abs to CD11b, CD11c, and IA^b; and purified by sorting on a FACSVantage (BD Biosciences, San Jose, CA). Unstimulated dendritic cells were sorted as CD11b⁺CD11c⁺IA^b− cells, whereas LPS-stimulated dendritic cells were sorted as CD11b⁺CD11c⁺IA^b+ cells, using class II MHC up-regulation as a marker of activation.

cDNA arrays

Macrophages from day 12 after BCG infection were lysed in Trizol (Life Technologies), and then total RNA was isolated according to manufacturer’s instructions. mRNA was then purified using two rounds of oligo(dT) cellulose columns (Pharmacia, Piscataway, NJ); 5 μg of this mRNA were used as a template to generate an unamplified oligo(dT)-primed cDNA library in the pSPORT plasmid vector according to the manufacturer’s directions (Life Technologies). Then 9200 individual clones in E. coli were grown and stored in 384-well plates. These clones were then spotted using a Q-bot onto 20- × 20-cm nylon membranes in duplicate, with the central spot of every grid of 9 containing a plasmid with a control cDNA. The colonies were lysed on the membrane using proteinase K, and the plasmid DNA was denatured, neutralized, and cross-linked to the membrane using a Stratalinker (Stratagene, La Jolla, CA). Duplicate membranes were probed with ³²P-labeled first strand cDNA generated from mRNA from resident peritoneal or BCG macrophages primed with a mixture of oligo(dT) and random hexamer primers. Control mRNA was added to the labeling reactions to control for labeling and hybridization efficiency between probes and membranes, respectively. Hybridization was detected using a Molecular Dynamics (Sunnyvale, CA) PhosphorImager, and the location and intensity of individual spots were determined using SFV software provided by Dr. R. Baumgartner. Average background for the membrane was subtracted from the spot intensities, and intensities were normalized for average control spot intensity. Duplicate spots on each membrane for each clone were averaged, and data were discarded if the duplicate intensities varied by >50%. Clones that met this criterion and that showed a >2-fold difference in intensity were sequenced. Sequences were analyzed by searching the nonredundant and EST databases at GenBank using the Advanced Blast search algorithm.

Northern blot

Total RNA from resident peritoneal and BCG-activated macrophages was isolated as above. For in vitro activated macrophages, resident peritoneal macrophages were plate adhered overnight, nonadherent cells were removed, and cells were activated with 100 ng/ml LPS or 10 U/ml IFN-γ (R&D Systems). After 4 h (LPS) or 48 h (IFN-γ), cells were lysed, and RNA was prepared as above. Northern blot analysis of 10 μg total RNA was conducted using standard methods. The Northern blot was probed with a random primed, [³²P]dCTP (New England Nuclear, Boston, MA)-labeled probe corresponding to the entire cDNA for spi2a or for EF1α as a housekeeping gene.

Expression constructs and transfection

All expression constructs were in the EF6/V5-His-TOPO vector (Invitrogen, San Diego, CA). The spi2a-HA construct was generated by amplifying the spi2a open reading frame using a forward primer encoding an HA tag N-terminal to the cDNA start (forward primer, CGGAATTCATGTACCCATACGACGTCCCAGACTACGCTGCTGGTGTCTCCCCTGCTGTC; reverse primer, CGGGATCCTCACTGTCCAATCAGGCATAG) with the cDNA array clone as template. The spi2a-HisHA expression construct was generated with the identical reverse primer to the spi2a-HA construct and with a forward primer encoding both the 6-residue histidine tag (His tag) and the HA tag (forward primer, ATGCATCATCACCATCACCATTACCCATACGACGTCCCAGACTACGCTGCTGGTGTCTCCCCTGCTGTC). To generate the ISG15-V5 construct, the open reading frame of ISG15 was amplified from cDNA made from total RNA of RAW264.7 cells treated for 4 h with 100 ng/ml LPS (List) using a forward primer encoding a V5 tag N-terminal to the start codon (forward primer, ATGGGTAAGCCTATCCCTAACCCTCTCCTCGGTCTCGATTCTACGGCCTGGGACCTAAAGGTG; reverse primer, TTAGGCACACTGGTCCCCTCC). All constructs were verified by sequence analysis. Ten micrograms DNA were transiently transfected into between 10 and 50 × 10⁶ RAW264.7 cells by electroporation (16). Cells were plate adhered overnight and used for subsequent experiments. To generate stable clones, transient transfectants were selected in medium with 5 μg/ml blasticidin (Invitrogen) for 10 days and then cloned by limiting dilution.

Western blot

Approximately 5 × 10⁵ transiently transfected cells or stable clones were plated in wells of 24-well dishes. In some cases, cells were treated overnight with 100 ng/ml Salmonella minnesota LPS (List). The cells were lysed in 10 mM HEPES, pH 7.4, with 150 mM NaCl and 1% Triton X-100 containing leupeptin (1 μM, Boehringer Mannheim, Indianapolis, IN), aprotinin (1/100, Sigma, St. Louis, MO), and PMSF (1 mM, Boehringer Mannheim). The extracts were spun at 15,000 rpm for 15 min to remove nuclei, and then the supernatant removed for analysis by SDS-PAGE. Resident peritoneal macrophages were plated at ∼1 × 10⁶ macrophages per well in 24-well tissue culture dishes. After overnight adherence, nonadherent cells were washed away, and then cells were cultured with medium alone or medium with LPS at 100 ng/ml or IFN-γ at 10 U/ml (R&D Systems). At the indicated times after activation, the cells were lysed, and cytoplasmic extracts generated as above. For the experiment involving in vivo activated macrophages, macrophages from uninfected mice or those from BCG-infected mice were plate adhered for 2 h before removal of nonadherent cells and lysis. Extracts were prepared as above. A 25-μg protein sample was used per lane for this experiment. For detection of the tagged proteins, the anti-HA.11 mAb (Covance, Princeton, NJ) or the anti-PK mAb to the V5 tag (Serotech, Raleigh, NC) was used as suggested by manufacturer with anti-mouse HRP (Zymed Laboratories, San Francisco, CA) and detected using ECL Plus (Amersham, Arlington, Heights, IL). To detect ISG15, Western blots were probed with affinity purified rabbit polyclonal antiserum to ISG15 (14) used at 1 μg/ml and anti-rabbit HRP (Zymed Laboratories). To detect spi2a, Western blots were probed with affinity-purified rabbit polyclonal antiserum generated to recombinant spi2a (E. C. Morris, T. Dafforn, S. L. Forsyth, A. J. Horvath, L. Hampson, I. N. Hampson, R. W. Carrell and P. B. Coughlin, manuscript in preparation) used at 1/1000 dilution followed by anti-rabbit HRP.

Semiquantitative RT-PCR

Approximately 1 μg total RNA from macrophage populations was reverse transcribed using Superscript II reverse transcriptase and oligo(dT) primers (Life Technologies). Serial 1/3 dilutions of cDNA were amplified with primers to murine hypoxanthine phosphoribosyltransferase (HPRT) to standardize between cDNA samples (forward primer, GATACAGGCCAGACTTTGTTG; reverse primer, GGTAGGCTGGCCTATAGGCT). Matched 3-fold dilutions of cDNA from each sample were then amplified with primers to spi2a (forward primer, GGAATGGCAGGTGTCGGATG; reverse primer, GGTCAGGAACCTGATTTCGTC). These primers were chosen to minimize cross-hybridization with other serpins that may be expressed in macrophages; the forward primer encompasses the reactive site loop of spi2a, and the reverse primer is in the 3′-UTR. Amplified products were separated on 1% agarose gels and visualized with ethidium bromide staining.

Real time PCR

cDNA from dendritic cells purified by cell sorting was generated as above and amplified with probe and primer sets for murine HPRT and spi2a as indicated below using TaqMan Universal PCR master mix and an ABI Prism 7700 Sequence Detector (Applied Biosystems, Foster City, CA) following the manufacturer’s instructions. Genomic DNA contamination was measured by including template that had been mock reverse transcribed and at all times accounted for <10% of the signal. Fold induction was determined from C_t values normalized for HPRT expression and then normalized to the value derived from medium-treated dendritic cells. Primers and probes used were as follows. Spi2a: forward, CCAAATGGTGAGGGTGCTTCT; reverse, GCATAGCGGATCACCAAAACA; probe, CCCAACGGCTGGAATCTAAGCGTTTAT. HPRT primers: forward, TGGAAAGAATGTCTTGATTGTTGAA; reverse, AGCTTGCAACCTTAACCATTTTG; probe, CAAACTTTGCTTTCCCTGGTTAAGCAGTACAGC.

Luciferase assays

RAW 264.7 cells (1 × 10⁷) were transiently transfected by electroporation as above with 10 μg pXP-2 plasmid (17) containing −259 to +46 nucleotides of the spi2a promoter driving the firefly luciferase gene (18) and plated into one 96-well dish. Cells were adhered overnight, washed once with PBS, and treated as indicated with IFN-γ at 10 U/ml (R&D Systems) or medium for 8 h. Bacteria or their products were then added for 14 h at the indicated final concentration. All stimuli except LPS and S. minnesota were treated with polymyxin B (10 μg/ml) for 1 h before addition to cells to ensure that the data did not result from LPS contamination. Cells were lysed, and luciferase activity was read using the Luciferase Reporter Assay System (Promega, Madison, WI) according to the manufacturer’s instructions. Heat-killed Mycobacterium tuberculosis H37Rv (100 μg/ml) and M. tuberculosis products lipoarabinomannan (AraLAM; 10 μg/ml), mycolylarabinogalactan-peptidoglycan complex (mAGP; 100 μg/ml), and soluble cell wall proteins (SCWP; 1 μg/ml) were obtained from J. Belisle (National Institute of Allergy and Infectious Diseases, Bethesda, MD) through the TB Research Materials and Vaccine Testing Contract. Heat-killed Staphylococcus aureus (clinical isolate) was used at 2 × 10⁶ bacteria/well, and heat-killed S. minnnesota R595 (ATCC) was used at 2 × 10³ bacteria per well. Zymosan (3 × 10⁶ particles/well) was obtained from Molecular Probes (Eugene, OR), S. minnnesota LPS (100 ng/ml) was obtained from List, Staphylococcus aureus peptidoglycan (10 μg/ml) was from Fluka (Buchs, Switzerland), and the synthetic lipopeptide PAM₃CSK₄ (100 ng/ml) was from Boehringer Mannheim.

Purification of spi2a-HisHA and identification of complexed proteins

For affinity purification, CNBr-activated Sepharose beads conjugated to the HA.11 mAb at 1 mg Ab/ml beads following the manufacturer’s instructions (Pharmacia) were used followed by Ni²⁺ beads (Invitrogen). Cytoplasmic extracts generated in 20 mM phosphate buffer, pH 7.8, with 150 mM NaCl, 1% Triton X-100 (lysis buffer), and protease inhibitors as above were pooled from 2 × 10⁹ RAW264.7 macrophages transiently transfected with the spi2a-HisHA construct. The extracts were incubated with HA.11 beads for 1 h at 4°C with rocking. The HA.11 beads were washed three times with lysis buffer without Triton X-100 and eluted by boiling in this buffer with 0.3% SDS to prevent precipitation of proteins. This eluate was then incubated with Ni²⁺ beads at room temperature for 1 h, washed five times in 20 mM phosphate buffer, pH 7.8, with 150 mM NaCl and then three times with 20 mM phosphate buffer, pH 6.0, with 150 mM NaCl. The protein was eluted by boiling in SDS sample buffer containing 2-ME. Ninety-five percent of the eluted protein was run in one lane of a 10% SDS-PAGE gel for silver stain, and the remaining 5% was run in one lane for Western blotting with the HA.11 Ab. Silver staining was performed as described (19).

To prepare samples for tandem mass spectrometry, silver-stained bands were excised from the gel, cut into ∼1-mm cubes, and subjected to in-gel tryptic digests (19). Peptides generated were identified by nanoscale microcapillary liquid chromatography-tandem mass spectrometry techniques using an LCQ Classic ion trap mass spectrometer (ThermoFinnigan, San Jose, CA) (20). Spectra were searched against the OWL nonredundant protein sequence database as well as the EST databases using the program SEQUEST (21), which matches theoretical and acquired tandem mass spectra.

Spi2a is up-regulated in macrophages during bacterial infection in vivo

We used cDNA arrays to identify a large number of genes that are specifically induced in murine peritoneal macrophages during in vivo infection with BCG (data not shown). One of these genes encoded serpin 2a (spi2a), a member of the serine protease inhibitor (serpin) family. By Northern blot analysis, mRNA for spi2a was not detected in resident peritoneal macrophages from uninfected mice but was detected strongly in macrophages from mice infected for 12 days with BCG (Fig. 1⇓A). The strong induction in BCG-activated macrophages was also seen at the protein level; Fig. 1⇓B shows a Western blot of macrophage lysates probed with polyclonal antiserum recognizing spi2a. The specificity of the polyclonal antiserum to spi2a is also shown in Fig. 1⇓B. The polyclonal serum detects HA-tagged spi2a when transfected into the macrophage cell line, RAW264.7, but not in vector control-transfected cells, which do not express any endogenous spi2a mRNA (Fig. 1⇓B and data not shown). The band detected by the polyclonal serum is identical with that detected by an anti-HA Ab on the same blot after stripping (Fig. 1⇓B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Spi2a is up-regulated in macrophages induced in vivo by bacterial infection. A, Top, Northern blot analysis of total RNA from resident peritoneal macrophages from uninfected mice (RPMφ) and peritoneal macrophages from mice infected 12 days previously with BCG (BCGMφ) probed with the complete cDNA for spi2a; bottom, the same blot stripped and reprobed for the housekeeping gene EF1α. B, Top, Western blot analysis of spi2a expression in cytoplasmic extracts from peritoneal macrophages as in A or from vector control (control) and HA-tagged spi2a (spi2a-NHA) expression construct-transfected RAW264.7 macrophages. spi2a protein is detected using polyclonal antiserum generated to recombinant spi2a protein. Bottom, The same blot stripped and reprobed with a mAb to the HA epitope tag. The slower mobility of the transfected spi2a in comparison with endogenous spi2a is due to the addition of the HA epitope tag. C, Semiquantitative RT-PCR analysis of cDNA generated from total RNA from resident peritoneal macrophages, macrophages isolated 5 days after Salmonella and Listeria infection, and 12 days after BCG infection. Three-fold serial dilutions of cDNA were matched for HPRT expression (bottom) and then amplified with primers for spi2a (top). PCR products were separated on agarose gels and detected by ethidium bromide staining.

We investigated whether induction of spi2a was unique to infection with Mycobacterium or whether infection with other intracellular pathogens would also induce this transcript. Mice were infected i.p. with L. monocytogenes, a Gram-positive bacterium, or an attenuated strain of S. typhimurium, a Gram-negative bacterium, and peritoneal macrophages were isolated after 5 days. spi2a mRNA levels were assessed by semiquantitative RT-PCR and compared with resident peritoneal macrophages and BCG macrophages. Fig. 1⇑C shows that both infection with S. typhimurium and infection with L. monocytogenes induced spi2a mRNA. This suggests that induction of spi2a is a general response of macrophages to in vivo infection with intracellular bacteria and that it is not specific to mycobacterial infections.

spi2a is induced in vitro by bacterial products and IFN-γ

We assessed whether the induction of spi2a was in response to the bacteria directly or indirectly via a cytokine induced in vivo by bacterial infection. To address this, we initially performed Northern blot analysis on total RNA from resident peritoneal macrophages that were treated in vitro with LPS, a molecule from the coat of Gram-negative bacteria, or with IFN-γ, a cytokine that is secreted by NK cells and T cells in response to bacterial infection. These molecules are both potent activators of macrophages although they act on distinct pathways. Both LPS treatment for 4 h and IFN-γ treatment for 48 h induced spi2a mRNA in resident peritoneal macrophages to a greater extent than culture alone (Fig. 2⇓A). We also investigated whether spi2a is similarly regulated in myeloid dendritic cells, a cell highly related to the macrophage. Real time PCR analysis showed that spi2a mRNA is also induced by LPS treatment in bone marrow-derived dendritic cells, a stimulation that also induces maturation of these cells and up-regulation of cell surface class II MHC levels (Fig. 2⇓B). These data suggest that spi2a is regulated similarly in macrophages and dendritic cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

spi2a is induced in macrophages and dendritic cells by in vitro treatment with bacterial products and IFN-γ. A, Northern blot analysis of resident peritoneal macrophages that had been plate adhered (medium) or activated with LPS for 4 h (LPS 4 h) or IFN-γ for 48 h (IFN-γ 48 h). Spi2a and EF1α transcripts were detected as in Fig. 1⇑A. B, Real time PCR analysis of spi2a expression in bone marrow-derived dendritic cells. Dendritic cells from a 7-day culture of GM-CSF-treated bone marrow were either cultured in medium or with 100 ng/ml LPS for 24 h and then sorted as CD11b⁺CD11c⁺ cells that were either IA^b− (medium) or IA^b+ (LPS). cDNA from these populations was subjected to real time PCR using primer/probe sets specific for spi2a and HPRT and normalized to HPRT levels. Data are represented as the fold induction of spi2a over medium-treated cells and are the average of triplicate samples ± SD. C, RAW264.7 cells were transfected with a plasmid containing bases −259 to +46 of the spi2a promoter driving the firefly luciferase gene. The cells were stimulated with bacteria or their products as indicated for 14 h and lysed, and the luciferase activity was measured. The amount of each stimuli used is listed in Materials and Methods. The activity of the promoter is represented as the fold induction over untreated cells (none) on the y-axis. D, Cells were transfected and treated as in B but were first pretreated with IFN-γ (10 U/ml) for 8 h before the addition of stimuli. The data in C and D are expressed as the mean of three independent experiments ± SEM. staph, Staphylococcus aureus; MTb, M. tuberculosis.

The finding that LPS, a Gram-negative bacterial product, can activate transcription of spi2a in vitro led us to ask whether other classes of bacteria and their products can also induce spi2a. To do so, we used the RAW264.7 macrophage cell line transfected with a plasmid containing base pairs −259 to +46 of the spi2a promoter driving the firefly luciferase gene as a reporter (18). LPS treatment of RAW264.7 cells induces spi2a mRNA as is seen in resident peritoneal macrophages (data not shown). A variety of whole bacteria as well as purified bacterial cell wall components and yeast cell walls directly induced the spi2a promoter in vitro (Fig. 2⇑C). This induction varied between 2.5-fold (araLAM) and 4.6-fold (zymosan) over basal levels. As found with in vivo infection, Gram-negative (S. minnesota) and Gram-positive (Staphylococcus aureus) bacteria as well as Mycobacterium (M. tuberculosis) activated this promoter. The bacterial products tested were derived from Gram-negative bacteria (LPS), Gram-positive bacteria (peptidoglycan) and Mycobacterium (araLAM, MAGP, SCWP). The doses used in this experiment have been previously shown to induce maximal production of TNF-α from RAW264.7 cells (data not shown), and all stimuli except LPS and Salmonella were treated with polymyxin B to ensure the induction did not result from LPS contamination.

IFN-γ is known to sensitize macrophages to respond to bacterial stimuli. Therefore, we were interested in how the presence of IFN-γ affected the ability of macrophages to induce spi2a in response to these stimuli. To test this, we pretreated RAW264.7 macrophages transfected with the spi2a promoter luciferase construct with IFN-γ and then activated with bacteria or bacterial components as in Fig. 2⇑C before assaying luciferase activity. As shown in Fig. 2⇑D, IFN-γ alone induces the spi2a promoter in this system as it does in primary macrophages (Fig. 2⇑A). The induction by IFN-γ is 16-fold in the RAW264.7 system, which is greater than the extent of induction by bacterial products alone seen in Fig. 2⇑C. Interestingly, pretreatment with IFN-γ caused a large increase in the induction of the spi2a promoter by the bacteria and their components (Fig. 2⇑D). This induction ranged from 27-fold (MAGP) to 68-fold (LPS) over background and was much greater than the additive effects of IFN-γ and the bacteria/components alone. These results were confirmed when looking at protein production by primary macrophages treated in vitro with LPS either with or without IFN-γ pretreatment for 14 h (data not shown). Therefore, as with other proteins important in the macrophage response to pathogens such as inducible NO synthase (iNOS) and IL-12, there is profound induction of both spi2a mRNA and protein by a combination of IFN-γ and bacterial products.

Identification of spi2a-ISG15 conjugates in activated macrophages

While studying the expression of spi2a protein in the RAW264.7 macrophage cell line, we observed that higher molecular mass forms of spi2a were detected (Fig. 3⇓A). An N-terminally HA-tagged version of spi2a (spi2a-HA) was transiently expressed in the RAW264.7 macrophage cell line, cytoplasmic extracts were separated by SDS-PAGE, and spi2a was detected with a mAb to the HA tag. The majority of the spi2a-HA was present in a band with the electrophoretic mobility predicted for the tagged protein, 50 kDa (Fig. 3⇓A). Slower migrating forms of ∼65 and 80 kDa were also detected. The 42-kDa band corresponds to a spi2a-HA degradation product.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

spi2a is present in higher molecular mass conjugates containing ISG15 in RAW264.7 macrophages. A, Cytoplasmic extract from RAW264.7 cells transiently transfected with an expression construct for spi2a-HA was analyzed by Western blot after separation by SDS-PAGE. spi2a expression was detected using a mAb to the HA tag. B, spi2a protein was affinity purified from RAW264.7 cells transiently transfected with spi2a-HisHA using Sepharose beads conjugated to an anti-HA mAb followed by Ni²⁺ beads to bind the six-His tag. The eluate from the Ni²⁺ beads was separated by SDS-PAGE and either silver stained (left) or Western blotted (right) with detection by the anti-HA Ab. The two bands designated with arrows were identified by both Western blot and silver stain. The two bands running at 65 and 80 kDa (arrows) were cut out of the gel, digested with trypsin, and subjected to tandem mass spectrometry to determine the peptide sequence. C, Amino acid sequence of ISG15 with the three tryptic peptides identified by mass spectrometry of the 80-kDa band (upper arrow, B) shown in the boxes.

To identify the protein components of the 65- and 80-kDa complexes, we purified spi2a from transiently transfected RAW264.7 cells using a version of spi2a with both an N-terminal HA and a six-histidine tag (spi2a-HisHA). Spi2a-HisHA was first affinity purified from transiently transfected RAW264.7 cells with Sepharose beads conjugated to a mAb to the HA tag, eluted, and then affinity purified with Ni²⁺ beads that bind the His tag. The protein was eluted from the Ni²⁺ beads, separated by SDS-PAGE, and silver stained (Fig. 3⇑B). Two of the silver-stained bands that matched precisely with the 65- and 80-kDa forms of spi2a detected by Western blotting (Fig. 3⇑B, arrows) were cut out, and the protein content was identified by tandem mass spectrometry. The 80-kDa band contained only two proteins, spi2a and murine ISG15, also known as ubiquitin cross-reactive protein; Fig. 3⇑C shows the sequence of ISG15 with the three independently identified peptides boxed. The ISG15 peptides identified covered 30.6% of the protein sequence. The 65-kDa band also contained spi2a and ISG15, as well as an apparent contaminant, TCP-1γ, a 66-kDa chaperone.

ISG15 is a 15-kDa protein that contains two ubiquitin homology domains. It has been shown to conjugate to cellular proteins in a process analogous to that for ubiquitin, using homologous, but distinct, enzymes (22, 23, 24). Thus, the 65-kDa form of spi2a likely represents spi2a (50 kDa) covalently bound to one molecule of ISG15, whereas the 80-kDa form represents spi2a bound to two molecules of ISG15. The other bands seen on the silver-stained gel contained serum proteins, including serum albumin (69 kDa) and α₂-macroglobulin (165 kDa), and were most likely derived from the FCS in the tissue culture medium.

To confirm that the spi2a 65 and 80 kDa bands contained ISG15, we affinity purified spi2a-HisHA from RAW264.7 cells and probed Western blots with Abs to the HA tag or to endogenous ISG15. Polyclonal Abs raised against purified human ISG15 identified the 65-kDa and 80-kDa bands that also reacted with the anti-HA Ab (Fig. 4⇓) but did not label the 50-kDa band that contained the unconjugated spi2a.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

ISG15 Abs detect the 65- and 80-kDa forms of spi2a. Protein from RAW264.7 cells transiently transfected with vector alone or vector encoding spi2a-HisHA was affinity purified as in Fig. 3⇑ and separated by SDS-PAGE; Western blots were probed with affinity-purified polyclonal serum to ISG15 (right). This blot was then stripped and reprobed with an anti-HA mAb (left).

We next coexpressed spi2a-HA and V5-tagged ISG15 (ISG15-V5) in RAW264.7 cells and assayed for the presence of ISG15-V5-spi2a conjugates. The 65- and 80-kDa spi2a conjugates were detected when spi2a-HA was transfected alone or cotransfected with the ISG15-V5 construct (Fig. 5⇓, top). As expected, the conjugates were slightly larger in the cells cotransfected with the ISG15 containing the 14-aa V5 tag. A replicate blot was probed with Abs to the V5 tag present on ISG15. When ISG15-V5 was transfected alone, we detected a ladder of bands that presumably corresponds to ISG15 conjugated to a variety of cellular proteins, including spi2a (Fig. 5⇓, middle). When spi2a-HA and ISG15-V5 were coexpressed, ISG15 was found in four principal bands, two of which corresponded to the spi2a complexes detected with the HA Ab. It is striking that overexpression of spi2a along with ISG15 caused the redistribution of ISG15 into four bands as opposed to the numerous bands seen when ISG15-V5 was expressed alone. This may reflect a rate-limiting step in the conjugation process exposed by overexpression of both proteins. The lower panel in Fig. 5⇓ shows that unconjugated ISG15-V5 was not limiting regardless of whether spi2a was overexpressed.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Epitope-tagged ISG15 conjugates to spi2a. RAW264.7 cells were transiently transfected with vector alone, vector encoding spi2a-HA, and/or ISG15-V5. Cytoplasmic extracts were separated by SDS-PAGE, and Western blots were probed with anti-HA (top) or anti-V5 (middle and bottom) mAbs. The bottom panel is a shorter run of the middle panel to show free ISG15-V5 protein.

ISG15 expression is induced during macrophage activation

Macrophages are strongly activated by bacterial DNA, and transient transfection using bacterial vectors results in the enhanced expression of many proteins associated with the activated state (25). We therefore examined the conjugation of spi2a and ISG15 in stably transfected macrophages, where the transient activation induced by the bacterial plasmid DNA has subsided. Surprisingly, stable clones of RAW264.7 cells expressing spi2a-HA did not contain the 65- and 80-kDa spi2a-ISG15 conjugates that were present in transiently transfected cells (Fig. 6⇓A). However, LPS induced these conjugates in the macrophages (Fig. 6⇓A), implying that LPS had induced the expression of ISG15, which could now conjugate with stably transfected spi2a. This was confirmed by Western blot analysis demonstrating that ISG15 was essentially undetectable in untreated RAW264.7 cells and was very strongly induced after LPS stimulation (Fig. 6⇓B). The hypothesis that bacterial DNA was inducing ISG15 during transient transfections was similarly confirmed (Fig. 6⇓B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Spi2a-ISG15 conjugates and ISG15 are induced by transfection and LPS treatment. A, Cytoplasmic extracts from RAW264.7 cells either transiently transfected with the spi2a-HA construct or stable clones made with vector alone or the spi2a-HA construct were analyzed as in Fig. 1⇑A. The stable clones were either untreated or treated overnight with 100 ng/ml LPS. B, Cytoplasmic extracts from RAW264.7 cells that were untreated, transiently transfected with empty vector (+transfection), or treated with 100 ng/ml LPS overnight were analyzed by Western blot with affinity-purified polyclonal Abs to ISG15.

Although the RAW 264.7 cell line is useful for transfection studies, it was important to demonstrate the regulation of ISG15 in primary macrophages, both in vitro and in vivo. LPS induced ISG15 in resident peritoneal macrophages; the protein was detected within 1 h and reached maximal levels of expression 4 h after stimulation with LPS (Fig. 7⇓A). LPS also induced the expression of spi2a in resident peritoneal macrophages with similar kinetics to ISG15. Interestingly, although IFN-γ alone strongly induced spi2a, it had no effect on the expression of ISG15 (Fig. 7⇓A), demonstrating that the regulation of these two proteins could be uncoupled.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

ISG15 and spi2a are induced in in vitro and in vivo activated primary murine macrophages. A, Cytoplasmic extracts were generated from resident peritoneal macrophages that were treated with 100 ng/ml LPS or 10 U/ml IFN-γ from 0 to 24 h as indicated. ISG15 was detected as in Fig. 4⇑ and spi2a was detected as in Fig. 1B⇑. B, Cytoplasmic extracts from peritoneal macrophages from uninfected mice (RPMφ) or mice infected 12 days previously with BCG (BCGφ) were analyzed as in A.

We identified spi2a as a protein induced in response to macrophage activation by BCG infection in vivo (Fig. 1⇑), and we therefore investigated whether ISG15 is also induced under these conditions. Infection of mice i.p. with BCG resulted in the very strong induction of both spi2a and ISG15 in macrophages in vivo (Fig. 7⇑B). This demonstrates that the induction of ISG15 and spi2a in vitro accurately reflects the more complex process of macrophage activation during bacterial infection in vivo.