BAFF controls B cell metabolic fitness through a PKCβ- and Akt-dependent mechanism

Incubation of mature B cells with BAFF drastically changes the state of B cell metabolism. Treatment with BAFF increases B cell size, cellular protein content, and mitochondrial membrane potential (Fig. 1, A–C). BAFF-treated cells show changes in gene transcription that direct these cells toward production of proteins required for glycolytic metabolism and cell cycle progression. Analysis of polyribosome-associated and, hence, actively translated mRNAs revealed two large mRNA clusters that were particularly prominent in BAFF-treated B cells (Fig. 2, A–C; and Table S1). The first cluster contains mRNAs collectively related to glycolysis. The second cluster overrepresented in BAFF-stimulated B cells contains mRNAs for proteins controlling cell cycle, chromosome condensation, and mitosis. The BAFF-induced up-regulation of cell cycle progression proteins such as cyclin D and cyclin E, Cdk4, Mcm2 and 3, the proliferation marker Ki67, and Survivin was independently confirmed by protein expression analysis (Fig. 3 A). Analysis of total RNA from BAFF-treated cells showed that the abundance of these transcripts reflects overall changes in the pattern of gene expression rather then BAFF-mediated recruitment of the selected mRNAs into polyribosomes (Table S2).

Stimulation with BAFF resulted not only in the transcription and translation of genes required for cell cycle progression but also in phosphorylation of the key cell cycle controlling Rb protein (Fig. 3 A, bottom), which is prerequisite for the release of E2F and cell cycle entry into S phase (22). Despite the substantial accumulation and modification of cell cycle–controlling proteins, BAFF does not induce B cell proliferation in vitro (Fig. 3 B). However, preincubation of B cells with BAFF accelerates proliferation in response to BCR stimulation compared with BCR-only triggered cells (Fig. 3 B). This result points to a role of BAFF in maintaining B cells in a state of immediate responsiveness to antigenic stimulation.

Increased protein synthesis in response to BAFF suggests that BAFF controls activation of proteins required for translation. Indeed, treatment with BAFF caused phosphorylation of eukaryotic translation initiation factor 4E (eIF4E) and its inhibitor, eIF4E-binding protein 1 (4E-BP1; Fig. 4 A). Both are required for active protein synthesis: phosphorylation of eIF4E increases its binding to capped mRNAs, whereas 4E-BP phosphorylation disrupts its binding to eIF4E. Treatment with BAFF also leads to phosphorylation of S6 ribosomal protein, a hallmark of active protein synthesis (23–25).

Phosphorylation of 4E-BP, which is known to be controlled by Akt and Pim-2 (19), suggested an involvement of these kinases in BAFF signaling. Activation of Akt requires its recruitment to the plasma membrane, where it binds to the lipid second messenger phosphatidyl inositol-3,4,5-triphosphate via its pleckstrin homology domain. This induces a conformational change, which allows for Akt phosphorylation at threonine 308 by PDK1 and at serine 473 by a second kinase, whose identity is still controversial (26, 27).

We found that treatment with BAFF led to a rapid phosphorylation of Akt at both S473 and T308, which, after a transitory decline after 1 h of stimulation, regained a stably high level by 24 h and beyond (Fig. 4 B and not depicted). Activation of Akt by BAFF was associated with the phosphorylation of several Akt targets known to control cell survival and metabolism (Fig. 4 C). First, treatment with BAFF lead to phosphorylation of the GTPase-activating protein tuberous sclerosis complex 2 (TSC2), which promotes protein synthesis through activation of mammalian target of rapamycin (28, 29). Second, BAFF induced the phosphorylation and, hence, inactivation of glycogen synthase kinase–3 (GSK-3). GSK-3 has been found to cause apoptosis by inducing MCL1 degradation and compromising mitochondrial membrane integrity (30). Finally, stimulation of B cells with BAFF led to phosphorylation of FoxO1 and its subsequent degradation. Members of the FoxO family of transcription factors are known to play a key role in the regulation of genes required for cell cycle progression, glycolytic metabolism, and survival (31, 32).

Growth factor stimulation of hematopoietic cells induces Pim-2 expression, which confers resistance to atrophy and cell death (33). Treatment with BAFF strongly induced all three isoforms of Pim-2 within 24 h of stimulation (Fig. 4 D).

Akt-activation requires its binding to specific phosphoinositides, which are generated by PI3K at the plasma membrane (27). The predominant form of PI3K in B cells consists of the regulatory subunit p85α and the catalytic subunit p110δ (34). We found that treatment with BAFF led to tyrosine phosphorylation of a p85-associated protein with a molecular weight of approximately p110 (Fig. 5 A). Sequence analysis of a protein that co-migrated with phospho-p110 confirmed that this protein represents the catalytic subunit p110δ (unpublished data). It is therefore possible that BAFF activates PI3K by inducing the tyrosine phosphorylation of its catalytic subunit. A similar phenomenon has previously been reported in B cell lines stimulated through the BCR (35, 36).

The activation of PI3K by BAFF is important for the regulation of Akt activation and B cell survival. Treatment of B cells with the PI3K-specific inhibitor LY294002 diminished BAFF-induced activation of Akt and phosphorylation of Akt target proteins (Fig. 5 B and Fig. S1). Consistently, BAFF-mediated B cell survival was severely impaired by concentrations of LY294002, which abolish Akt activity (Fig. 5 C). The half-life of the phosphoinositide products of PI3K is limited by the lipid phosphatase and tensin homologue deleted on chromosome 10 (PTEN), but BAFF did not appear to affect PTEN activity (Fig. S2), as judged by phosphorylation of the inhibitory PTEN phosphorylation site serine 380 (37).

Akt activation requires two sequential events. Initially, Akt binding to the phosphoinositide products of PI3K induces a conformational change in the protein. This allows for its phosphorylation at two essential sites: T308 in the activation loop is phosphorylated by PDK1, whereas the identity of the kinase that phosphorylates the C-terminal S473 residue remains to be determined (26, 27). In search of the kinase, which is responsible for Akt S473 phosphorylation and B cell survival in response to BAFF, we tested the role of PKCβ in BAFF-mediated signaling. This choice was based on two major reasons. First, coincubation with recombinant PKCβ leads to Akt phosphorylation at S473 but not T308 (38). Second, PKCβ deficiency in mice causes a decline in B cell survival in vitro and in vivo (39, 40).

We initially assessed the ability of BAFF to activate PKCβ by analysis of BAFF-induced PKCβ translocation to the plasma membrane. This translocation is an activation hallmark for phospholipid-dependent classic PKCs, including PKCβ (41–43), and triggering of B cells through the BCR leads to translocation of PKCβ into specialized membrane microdomains called lipid rafts (44). We found that treatment of B cells with BAFF increased the amount of membrane-associated PKCβ (Fig. 6 A). A similar level of BAFF-induced membrane translocation was observed for Akt (Fig. 6 A). Notably, membrane translocation of PKCβ in response to BAFF occurred outside lipid rafts (Fig. S3). Because lipid rafts of BAFF-triggered B cells did also not contain Akt or PDK1, we conclude that BAFF-induced Akt activation is a raft-independent event.

In addition to PKCβ translocation to the plasma membrane, BAFF promotes the association of PKCβ with Akt (Fig. 6 B). This association does not depend on PI3K activity (Fig. S4), but it is functionally important, as BAFF-induced Akt phosphorylation at S473 was greatly reduced in PKCβ-deficient B cells (Fig. 6 C). In contrast, Akt phosphorylation at the PDK1 target site T308 was similar between wild-type and mutant cells. The functional link between PKCβ and Akt is specific to BAFF signaling, as stimulation through the BCR induced Akt phosphorylation in PKCβ-deficient B cells at wild-type levels (Fig. 6 D).

PKCβ-deficient B cells exhibit poor survival in vitro in the absence of stimuli (Fig. 7 A) (40). Treatment with BAFF promotes the viability of mutant B cells, although they fail to reach wild-type survival levels regardless of the BAFF dose and the duration of treatment. Although PKCβ-deficient cells were partially responsive to the survival action of BAFF, its ability to support B cell growth in vitro was severed by lack of PKCβ (Fig. 7, A and B). These results suggest that the PKCβ-mediated Akt signaling element contributes mostly to BAFF-mediated B cell fitness rather than mere survival. Altered cellular responses of PKCβ-deficient B cells were not caused by defective surface expression of BAFF-R (Fig. 7 C), arguing for a cell-intrinsic BAFF signaling defect in the absence of PKCβ.

The involvement of PKCβ in BAFF-mediated signaling is indirectly supported by the pattern of changes in the PKCβ-deficient, peripheral B cell compartment. In the absence of BAFF signaling, B cell maturation arrests between the T1 (CD21^lo) and the T2 (CD21^hi) stages (7, 10–12). PKCβ deficiency impairs B cell maturation in a fashion similar to failed BAFF signaling, albeit at a lower magnitude. In addition, the size of the peripheral PKCβ-deficient B cell compartment falls to ∼70% of wild-type levels (39), and immature B cells prevail over mature B cells in the spleen of mutant mice, as judged by surface IgM and IgD expression (Fig. 7 D). Collectively, these observations suggest a partial defect in BAFF-mediated signaling and peripheral B cell maturation in the absence of PKCβ.

BAFF's prosurvival function has been linked to its ability to activate the processing of the NF-κB2 protein p100 into the transcriptionally active form p52 (15, 16). We find that BAFF-induced p100-p52 processing was not affected by the absence of PKCβ (Fig. 8 A). We have previously shown that BAFF controls B cell survival through cytoplasmic retention of the proapoptotic kinase PKCδ (18). However, the ratio of cytoplasmic versus nuclear PKCδ was not altered by PKCβ deficiency (Fig. 8 B). We conclude that the BAFF-mediated regulation of NF-κB and PKCδ occurs independent of PKCβ.

Our findings show that like cytokines such as IL-3 and IL- 7, which promote survival of hematopoietic cells, BAFF supports B cell survival by increasing metabolic fitness. Treatment of B cells with BAFF induces transcription of mRNAs that encode components of carbohydrate metabolism. The BAFF-induced metabolic bias toward glycolysis might be especially important for B cell survival in lymphoid organs and at inflammatory sites where oxygen tension is low compared with arterial blood (45, 46). In this environment, activation of glycolysis will provide energy to sustain the active protein synthesis and cell growth caused by BAFF. Indeed, prolonged BAFF treatment led to expression of the hypoxia-inducible factor α (unpublished data).

Normally, accumulation of proteins, particular those controlling cell cycle, and increase in cell volume precedes cell division. However, the lack of BrdU incorporation in BAFF-treated cells shows that BAFF does not induce DNA-replication, thus precluding cell division. Previous findings show persistent expression of Cdk inhibitor proteins p18 and p27 in BAFF-treated cells (47). Therefore, it is likely that high expression of these and possibly other cell cycle inhibitors prevents the proliferation of BAFF-treated B cells in the absence of an antigenic signal.

BAFF-stimulated cells enter BCR-induced proliferation more readily than untreated cells. This result suggests that the accumulation of cell cycle–controlling proteins in response to BAFF prepares resting B cells for an immediate immune response upon antigenic challenge. It is also attractive to speculate that, in response to BAFF, B cells might establish a storage pool of certain cell cycle–controlling mRNAs or proteins, which could be used for several rounds of cell division. Such a mechanism would help explain the astonishingly short replication time (∼7 h) of B cells at the height of the germinal center response (48, 49).

Akt activation requires two potentially independent pathways. The PI3K-dependent pathway is shared both by BAFF-R and BCR. This feature of the pathway makes it particularly important in the regulation of B cell immunity and explains the poor survival of B cells in the absence of the p85 regulatory subunit of PI3K (50, 51). The pathway that involves PKCβ appears to be BAFF specific. Thus, PKCβ deficiency impairs phosphorylation of Akt on activating serine 473 in response to BAFF but not BCR stimulation. The residual BAFF-induced Akt S473 phosphorylation, which we observe in the absence of PKCβ, indicates an ability of other kinases to partially adopt this role. This would be expected, as a plethora of kinases has previously been implicated in mediating Akt S473 phosphorylation (52).

Increased translocation of PKCβ to the plasma membrane and PKCβ association with Akt in response to BAFF suggest a direct involvement of PKCβ in BAFF signaling. Furthermore, PKCβ dependency of Akt phosphorylation at S473 but not T308 agrees with in vitro data on Akt phosphorylation by PKCβ (38) and suggests a direct action of PKCβ on S473 of Akt upon BAFF stimulation.

Involvement of PKCβ in BAFF and BCR signaling provides a mechanistic explanation for the poor survival and altered peripheral maturation of PKCβ-deficient B cells in vitro and in vivo. However, the relatively mild reduction of B cell numbers in the absence of PKCβ argues in favor of additional PKCβ-independent signaling pathways initiated by BAFF-R and/or BCR. Signaling from both receptors induces activation of NF-κB, albeit through distinct mechanisms. Although PKCβ has an important function in canonical NF-κB activation upon BCR-triggering, BAFF-induced NF-κB2 processing is independent of PKCβ. Another PKC family member, PKCδ, also plays an important role in the regulation of B cell survival, but it promotes cell death rather than survival. The proapoptotic potential of PKCδ is contained by BAFF, which prevents its accumulation in the nucleus. This BAFF-dependent survival mechanism can also function in the absence of PKCβ.

BAFF-mediated B cell survival likely represents the collective outcome of several BAFF-induced signaling features, including NF-κB activation, cytoplasmic retention of PKCδ, and Akt activation. It is thus conceivable that the loss of a single signaling branch can to some degree be compensated for. In this case, the loss of PKCβ would be expected to cause some damage to B cell survival, but it would be less severe than the complete abrogation of BAFF signaling. This could explain the more dramatic changes to the B cell compartment in BAFF- or BAFF-R–deficient mice than in PKCβ knockouts. Although PKCβ-deficient B cells were partially responsive to the survival action of BAFF, they appeared to be largely refractive to BAFF-mediated cell growth. This cellular process is closely associated with Akt- and mammalian target of rapamycin–dependent signaling pathways and represents an important distinction between growth factor–mediated cellular fitness on the one hand and mere cell survival on the other (19, 53). The latter can also be achieved through an altered ratio in the expression of pro- and antiapoptotic proteins, but, in this context, cells become progressively smaller (54). It appears that, perhaps in contrast to BAFF-induced NF-κB activation and cytoplasmic retention of PKCδ, which mediate B cell survival (55), PKCβ-dependent Akt activation preferentially regulates the fitness facet of BAFF-mediated cellular responses.

Our findings may have practical implications. There is increasing evidence that BAFF plays an important role in the control of autoreactive B cells (8, 9, 13, 14, 56, 57). Therefore, identification of Akt and PKCβ as components of BAFF signaling may suggest novel ways of pharmacological intervention in B cell–mediated autoimmune disorders.

Mice were housed in the Laboratory Animal Research Center at the Rockefeller University under specific pathogen-free conditions. C57BL/6 mice were used as a source of wild-type B cells. PKCβ-deficient mice on a C57BL/6 genetic background were used for analysis (39, 40). Mice were used at ages of 8–12 wk. Protocols were approved by the Institutional Animal Care and Use Committee at the Rockefeller University.

Resting mature CD43⁻, CD62L⁺ B cells were isolated from spleen and lymph nodes by MACS (Miltenyi Biotec) according to the manufacturer's instructions. The purity of mature B cells was routinely >95%. Cells were cultured for 12–16 h in RPMI 1640 medium supplemented with 10% FBS, 2 mM l-glutamine, 50 μM 2-mercapto-ethanol, and 100 U/ml penicillin/streptomycin at 37°C in a CO₂ incubator to reduce the background BAFF signaling caused by endogenous BAFF in vivo. This preincubation did not alter B cell size. After preincubation, live cells were separated over a Ficoll gradient (Cedar Lanes), and their purity was assessed by trypan blue exclusion. Cells were stimulated in culture medium using 25 ng/ml BAFF (R&D Systems) or 1.3 μg/ml F(ab′)₂ fragment goat anti–mouse IgM (Jackson ImmunoResearch Laboratories). LY294002 (Calbiochem) was used at a 10-μM concentration.

Cells were analyzed on a flow cytometer (FACSCalibur; Becton Dickinson) using CellQuest software (BD Biosciences). For measurement of mitochondrial membrane potential, cells were used after 12–16 h of preincubation without BAFF (time point 0) or after 72 h of BAFF stimulation. Cells were incubated in culture medium with or without BAFF and containing the potentiometric dye tetramethylrhodamine ethyl ester (TMRE; Invitrogen) at a 100-nM concentration. Carbonyl cyanide m-chlorophenyl hydrazone (Sigma-Aldrich) was added to control samples at a concentration of 50 μM as an uncoupler of the mitochondrial respiratory chain. After 30 min of incubation at 37°C in a CO₂ incubator, cells were immediately analyzed by FACS. For BrdU incorporation, B cells were cultured in the presence of 10 μM BrdU, and fixed and stained using the FITC BrdU Flow Kit (BD Biosciences). For cell death analysis, an aliquot of ∼10⁵ cells was removed from the B cell culture at various time points ranging from 0 to 96 h. TO-PRO-3 (Invitrogen) was added at a concentration of 10 nM to distinguish live (TO-PRO-3⁻) and dead (TO-PRO-3⁺) cells in the culture samples. Upon TO-PRO-3 addition, cells were immediately analyzed by FACS. B cell surface marker expression was analyzed using antibodies against IgD, CD21/35, B220 (BD Biosciences), BAFF-R (R&D Systems), or a F(ab′)₂ fragment of goat anti–mouse IgM (Jackson ImmunoResearch Laboratories), as previously described (58).

Cells were lysed in buffer containing 150 mM NaCl, 20 mM Tris-HCl, pH 7.5, 1 mM EDTA, 10% glycerol, 1% NP-40, 10 μM NaF, 2 mM Na₃VO₄, and protease inhibitor cocktail (Sigma-Aldrich). Lysates were resolved by SDS-PAGE and transferred onto Immobilon-P membrane (Millipore) using standard procedures. Membranes were developed with the following antibodies against: pS6 (S235/236), S6, peIF4E (S209), eIF4E, p4E-BP1 (S65), 4E-BP1, pAkt (S473), pAkt (T308), pGSK-3β (S9), GSK-3β, pFoxO1 (S256), FoxO1, pTSC2 (T1462), pPTEN (S380), PTEN, and Erk (all obtained from Cell Signaling Technology); Pim-2, Akt, TSC2, and PKCβ (all obtained from Santa Cruz Biotechnology, Inc.); p85, p110δ, and pY (4G10; all obtained from Upstate Biotechnology); and tubulin (Sigma-Aldrich). Where applicable, membranes were first developed with phosphospecific antibodies, and stripped and reprobed with the respective control antibody. For Akt immunoprecipitation, lysates were precleared with goat-serum and protein G–Sepharose (GE Healthcare), followed by incubation with anti-Akt antibody overnight. Anti–ferritin heavy chain antibody (Santa Cruz Biotechnology, Inc.) was used as a goat immunoglobulin control. Immune complexes were precipitated with protein G–Sepharose, washed five times in lysis buffer, and eluted in SDS sample buffer for SDS-PAGE. For p85 immunoprecipitation, lysates were precleared with rabbit immunoglobulin and protein A–Sepharose (GE Healthcare), followed by incubation with p85 antibody–agarose conjugate overnight. Immune complexes were washed and eluted as described earlier in this section. For analysis of NF-κB2 p100 processing and for the expression of cell cycle proteins, cells were lysed in a buffer containing 350 mM NaCl, 20 mM Hepes/NaOH, pH 7.9, 1 mM MgCl₂, 0.2 mM EDTA, 0.1 mM EGTA, 1% NP-40, 20% glycerol, 10 μM NaF, 2 mM Na₃VO₄, and protease inhibitor cocktail (Sigma-Aldrich). Lysates were analyzed by SDS-PAGE, and membranes were incubated with antibodies against NF-κB2 p52 (Upstate Biotechnology) or cyclin D2, cyclin E, Cdk4, and Mcm3 (all obtained from Santa Cruz Biotechnology, Inc.); Mcm2 and Rb (obtained from BD Biosciences); Ki67 (DakoCytomation); and Survivin (Chemicon). Signal quantification was done using NIH Image 1.63 software.

Lipid rafts were prepared as described previously (59). The purity of the preparation was assessed by the presence of Lyn and the lipid raft marker ganglioside GM1, which was detected using horseradish peroxidase–coupled cholera toxin B subunit (Sigma-Aldrich). Proteins were detected using antibodies against PKCβ, Akt, and PDK1 (all obtained from Santa Cruz Biotechnology, Inc.) and Lyn (a gift from C. Lowell, University of California, San Francisco, San Francisco, CA).

Fractionation of primary mouse B cells into cytoplasmic and nuclear extracts was performed as described previously (18). Lysates were analyzed for expression of PKCδ using an antibody from Santa Cruz Biotechnology, Inc. Fraction purity was assessed by Western blot analysis using tubulin as cytoplasmic and lamin B (Santa Cruz Biotechnology, Inc.) as nuclear markers, respectively. For preparation of cytoplasmic and membrane fractions, cells were disrupted by hypotonic lysis for 15 min on ice in a buffer containing 10 mM Hepes, pH 7.4, 10 mM KCl, 1.5 mM MgCl₂, 0.1 mM EDTA, 10 μM NaF, 2 mM Na₃VO₄, and protease inhibitor cocktail (Sigma-Aldrich). Nuclei were removed by centrifugation at 400 g for 10 min. The supernatant was then centrifuged at 95,000 g for 1 h in a SW55Ti rotor (Beckman Coulter). The supernatant was removed and saved as the cytoplasmic fraction. The pellet containing the membrane fraction was resolved in SDS sample buffer. Fraction purity was assessed by Western blot analysis using tubulin as cytoplasmic and Lyn as membrane markers, respectively.

Polyribosomes were purified essentially as described previously (60) from unstimulated cells and those that had been cultured in the presence of BAFF for 36 h. In brief, cells were incubated in culture medium containing 100 μg/ml cycloheximide (CHX; Sigma-Aldrich) for 15 min, washed three times in PBS containing CHX, and lysed in 10 mM Hepes-KOH, pH 7.4, 150 mM KCl, 5 mM MgCl₂, 1% NP-40, 0.5 mM DTT, 100 μg/ml CHX, 200 U/ml RNAsin (Promega), 100 U/ml SUPERase (Ambion), and EDTA-free protease inhibitor cocktail (Roche). All chemicals were of molecular biology–grade purity and nuclease free. Nuclei were removed from the extracts by centrifugation at 2,000 g for 10 min. An aliquot of the lysate was saved for preparation of total RNA. The remaining lysate was loaded onto a 20–50% wt/wt linear density sucrose gradient in 10 mM HEPES-KOH, pH 7.4, 150 mM KCl, and 5 mM MgCl₂. Gradients were centrifuged for 2 h at 40,000 g at 4°C in a SW41 rotor (Beckman Coulter). Fractions of 0.7-ml volume were collected with continuous monitoring at 254 nm using an ISCO UA-6 UV detector. To identify polyribosome-containing fractions, the content of ribosomal S6 protein in the fractions after trichloroacetic acid precipitation was determined by Western blot. For RNA isolation, polysome-containing fractions were pooled, and RNA was isolated from total and polysomal samples using TRIzol LS (Invitrogen) according to the manufacturer's protocol.

Quality of RNA was confirmed before labeling by analyzing 20–50 ng of each sample using the RNA 6000 NanoAssay and a Bioanalyzer 2100 (Agilent Technologies). All samples had a 28S/18S ribosomal peak ratio of 1.8–2 and were considered suitable for labeling. 2 μg of total RNA was used for cDNA synthesis using an oligo(dT)-T7 primer and the SuperScript Double-Stranded cDNA Synthesis Kit (Invitrogen). Synthesis, linear amplification, and labeling of cRNA were accomplished by transcription in vitro using the MessageAmp aRNA Kit (Ambion) and biotinylated nucleotides (Enzo Diagnostics). 10 μg of labeled and fragmented cRNA were hybridized to the mouse genome MOE430 2.0 array (Affymetrix), which interrogates ∼39,000 transcripts at 45°C for 16 h. Automated washing and staining were performed using the Fluidics Station 400 (Affymetrix), according to the manufacturer's protocols. Finally, chips were scanned with a high numerical aperture and flying objective lens in the scanner (GS3000; Affymetrix). Raw expression data were analyzed using Microarray Analysis software (version 5.1; Affymetrix). The complete datasets can be accessed at http://www.ncbi.nlm.nih.gov/geo under accession no. GSE5985. Data were normalized to a target intensity of 500 to account for differences in global chip intensity. Genes, which were differentially expressed between unstimulated and BAFF-treated cells, were initially identified using GCOS software. 377 transcripts showed a change of twofold or more (Table S2) and were selected for further analysis. Approximately three quarters of these genes were up-regulated upon BAFF stimulation, whereas one quarter was down-regulated. For a biological interpretation of the differentially regulated genes, the GoMiner gene ontology (GO) tool was used. GO is a genome project that describes gene products in terms of their associated biological processes, cellular components, and molecular functions (61). GO categories are organized in directed acyclic graphs, a kind of hierarchy in which one category can have more than one “parent.” GoMiner identifies GO categories that are over- or underrepresented in lists of genes of interest, such as differentially expressed genes from a microarray experiment, and calculates statistical significance as the one-sided nominal unadjusted p-value from Fisher's exact test (62). We report only those BAFF-induced biological processes in which the p-value was <0.001.

Gel-resolved proteins were digested with trypsin, batch purified on a reversed-phase microtip, and resulting peptide pools were individually analyzed by matrix-assisted laser desorption/ionization reflectron time-of-flight (MALDI-reTOF) mass spectrometry (MS; ultraflex TOF/TOF; Bruker Daltronics Inc.) for peptide mass fingerprinting (PMF), as previously described (63). Selected peptide ions (m/z) were taken to search a nonredundant (NR) protein database (3,245,378 entries on 28 January 2006; National Center for Biotechnology Information) using the PeptideSearch algorithm (developed by Matthias Mann, Max-Planck-Institute for Biochemistry, Martinsried, Germany; an updated version of this program is currently available as PepSea from Applied Biosystems/MDS SCIEX). A molecular mass range up to twice the apparent molecular weight (as estimated from electrophoretic relative mobility) was covered, with a mass accuracy restriction of <35 ppm and a maximum of one missed cleavage site allowed per peptide. To confirm PMF results with scores ≤40, mass spectrometric sequencing of selected peptides was done by MALDI-TOF/TOF (MS/MS) analysis on the same prepared samples using the ultraflex instrument in “LIFT” mode. Fragment ion spectra were taken to search the NR protein database using the Mascot MS/MS ion search program (version 2.0.04 for Microsoft Windows; Matrix Science Ltd.) (64). Any tentative confirmation (Mascot score ≥30) of a PMF result thus obtained was verified by comparing the computer-generated fragment ion series of the predicted tryptic peptide with the experimental MS/MS data.

Fig. S1 shows BAFF-induced Akt phosphorylation in the absence or presence of LY294002. Fig. S2 shows PTEN phosphorylation at serine 380, as well as total PTEN content in cell extracts prepared at various time points of BAFF stimulation. In Fig. S3, the distribution of PKCβ, Akt, and PDK1 into lipid rafts was analyzed before and after BAFF stimulation. Lipid raft–containing fractions are marked by the presence of Lyn and GM1. Fig. S4 shows BAFF-induced PKCβ-Akt association, which is not affected by the presence of LY294002. Table S1 lists the gene IDs for the GO categories presented in Fig. 2 (B and C). Table S2 contains the complete list of Affymetrix probe sets that showed a fold change of two or more in BAFF-stimulated versus unstimulated cells. Results are given both for RNA samples isolated from polysomes and analyzed in Fig. 2 (sheet 1 = polysome), as well as RNA samples derived from total cell lysates (sheet 2 = total).

We thank the Genomics Core Laboratory at the Memorial Sloan-Kettering Cancer Center for Affymetrix microarray analysis, Arpi Nazarian for help with mass spectrometric analysis, members of the Robert Darnell lab for advice on polysome purification and generous sharing of equipment, Marc-Werner Dobenecker for advice on lipid raft isolation, Sara Buonomo for discussion and reagents, and Eva Besmer for critical reading of the manuscript.

This work was supported by National Institutes of Health grant AI050867 (A. Tarakhovsky), the Irene Diamond Fund/Professorship Program (A. Tarakhovsky), the Boehringer Ingelheim Fonds (A. Patke), the S.L.E. Foundation (I. Mecklenbräuker), and National Cancer Institute Cancer Center support grant P30 CA08748 (P. Tempst).

The authors have no conflicting financial interests.