A Mouse Model for the Study of SYK Function through Chemical Genetics Demonstrates SYK-Dependent Signaling through the B Cell Receptor, but Not TLR4

SYK is a 72 kDa protein-tyrosine kinase originally identified as a component of the signal transduction machinery that couples the BCR to downstream signaling pathways (1–4). Most all BCR-initiated biochemical changes that take place are dependent on the expression of SYK (5, 6). SYK also is required for signaling through the FcεRI, FcγRI, FcγRIIA, FcγRIIIA, and FcαRI Ig receptors; the GPVI collagen receptor on platelets; NK cell–activating receptors; and many others (7). Most receptors coupled to the activation of SYK share in common a subunit or associated protein that contains at least one ITAM. For the BCR, these are CD79A and CD79B. The phosphorylation of an ITAM on a pair of tyrosines creates a binding site for the tandem pair of SYK SH2 domains, leading to the binding and activation of the kinase (8).

Although a role for SYK in signaling through receptors with ITAMs is well established, numerous other studies have identified or suggested roles for the kinase in signaling through less conventional modes. For example, several C-type lectin receptors contain only one half of an ITAM (a hemi-ITAM) and function as dimers to generate a single SYK-interacting motif (7, 9). Other receptors that lack ITAMs entirely reportedly interact with and signal via the kinase. Examples include the G-CSF receptor, TNF receptor, EPO receptor, BAFF receptor, integrin αIIbβ3, and TLR4 (10–18). SYK has been reported to interact directly with TLR4 to mediate certain LPS-signaling pathways in neutrophils and monocytes (17) and, in B cells, to mediate activation through TLR4 by a mechanism that requires the BCR complex (18).

Ascribing specific functions to a particular protein kinase can at times be difficult, especially for studies employing primary cells. Such studies often rely on the use of small molecule inhibitors. For protein kinases, these are mostly compounds that target the ATP-binding site (19). As such, these inhibitors are prone to exhibiting off-target effects. Such effects can arise from the interactions of the inhibitor with kinases other than the intended target or from interactions with one or more nonkinase targets (20). CRISPR/Cas9-mediated knockout technologies in which a specific kinase gene is disrupted work well for cultured cells but are not as well suited for studies in isolated primary cells where clones of gene-targeted cells can be difficult to generate. In contrast, gene knockouts at the cellular or whole organism levels can yield phenotypes distinct from those seen in inhibitor studies. Such differences can be due, for example, to compensatory changes in gene expression that occur in the knockout cells or to the disruption of protein–protein interactions in cases in which a targeted kinase is normally present as a component of a multi-subunit complex (21). Gene disruptions may also affect the development and/or survival of specific cell populations. In the specific example of Syk, its deletion from mice results in perinatal lethality (6, 22), and B cells fail to develop in immune systems reconstituted from SYK-deficient progenitors (6, 22).

Chemical genetics using analogue-sensitive kinases offer an alternative for the analysis of kinase-specific functions. In this approach, a mutation made in an amino acid within the ATP-binding site opens access to an additional pocket that allows the kinase to bind small molecule inhibitors, which contain bulky substituents that prevent them from binding with high affinity to the wild-type enzyme (23). Application of such an orthogonal inhibitor to cells expressing only the analogue-sensitive kinase will selectively block its activity, whereas cells expressing the wild-type kinase serve as controls for detecting any possible off-target effects of the inhibitor. We had developed previously an analogue-sensitive form of SYK (SYK-AQL) (24). In this study, we generated a strain of mice in which the endogenous Syk gene is replaced by one coding for SYK-AQL. These animals are viable, B cell development is largely restored, and IgM⁺IgD⁺ B cells populate the spleen. Activation through the BCR of SYK-AQL–expressing B cells, but not of Syk-expressing wild-type B cells, is blocked by an inhibitor that selectively targets the engineered kinase. However, the activation of B cells by LPS is not affected by the orthogonal inhibitor, indicating a lack of SYK dependency for LPS signaling.

C57BL/6N mice in which the endogenous Syk gene was replaced by one coding for SYK-AQL were generated by inGenious Targeting Laboratory (Ronkonkoma, NY). iTL IC1 (C57BL/6N) embryonic stem cells were targeted using the vector shown in Fig. 1A. Targeted cells were microinjected into BALB/c blastocysts. Chimeras with a high percentage of black coat color were mated with wild-type C57BL/6N mice to generate F1 heterozygous offspring. F1 heterozygotes were mated with C57BL/6 FLP mice to delete the Neo gene. Mice were genotyped by PCR using the primers 5′-CTCTTAACCTCACTCTTGCATGCG-3′ and 5′-CTTCAGGATTTTCACAGCCACG-3′ to distinguish between endogenous Syk and the knocked-in allele, which retains one set of LoxP-Frt sites. A BAC transgene coding for SYK-AQL was generated by the Transgenic Vectors Core, Hope Center for Neurological Disorders, Department of Neurology of the Washington University School of Medicine. The mutation was engineered using GalK recombineering (25) of BAC no. RP24-177L1 (Invitrogen, Carlsbad, CA). Pronuclear injections to generate C57BL/6N founders were carried out in the Purdue Transgenic Mouse Core Facility of the Purdue Center for Cancer Research. The incorporated BAC was identified using PCR primers 5′-GATGAATGCTCATCCGGAGTTCCGTATGGC-3′ and 5′-ACAAACGGCATGATGAACCTGAATCGCCAG-3′ to detect the chloramphenicol resistance gene. The Syk or Syk-AQL genes were detected using the primers 5′-GGCGAACGTCATGCAGCAG-3′ and 5′-CTGTTCTGCTGCAGGTACTTG-3′. PCR products were digested with PvuII to generate a smaller fragment selectively from the mutant Syk-AQL allele to distinguish it from endogenous Syk. Mice containing the incorporated BAC constructs were mated with the heterozygous knock-in animals to generate animals homozygous for both the knocked-in Syk-AQL allele and for the BAC Syk-AQL sequences. All animal studies were performed in accordance with the animal protocol procedures approved by the Institutional Animal Care and Use Committee of Purdue University.

Mouse splenocytes were isolated from 3- to 4 mo-old male and female mice and examined on a BD LSRFortessa flow cytometry cell analyzer after staining with Abs against B220 (FITC rat anti-mouse CD45, clone 30-F11; BD Biosciences), IgM (allophycocyanin rat anti-mouse IgM, clone II/41; BD Biosciences), and IgD (FITC rat anti-mouse IgD, clone 11-26c.2a; BD Biosciences). B cells were enriched from splenocyte preparations by negative selection using a B Cell Isolation Kit (Miltenyi Biotec). B cells were activated by treatment with 10 μg/ml goat anti-mouse IgM F(ab′)₂ (31178; Invitrogen) or 25 μg/ml LPS (L-4005; Sigma-Aldrich) for 48 h. Activated cells were examined by light microscopy or by flow cytometry using FITC-labeled Ab against CD86 (clone GL-1; MilliporeSigma). Cells were first gated on size to identify live cells for analysis. Cell viability was determined using the exclusion of trypan blue as an indicator of live cells. DNA synthesis in activated cells was measured by flow cytometry using the Click-iT EdU Alexa Fluor 488 Imaging Kit (Thermo Fisher Scientific, Waltham, MA). In some experiments, cells were treated with 1-(1,1-dimethylethyl)-3-[(3-methylphenyl)methyl]-[¹H]-pyrazolo[3,4-d]pyrimidin-4-amine (3-MB-PP1) or 1-(1,1-dimethylethyl)-3-(1-naphthalenylmethyl)-[¹H]-pyrazolo[3,4-d]pyrimidin-4-amine (1-NM-PP1) at a concentration of 10 μM. 1-NM-PP1 and 3-MB-PP1 were obtained from Cayman Chemical (Ann Arbor, MI).

Cells activated by treatment with goat anti-mouse IgM F(ab′)₂ for 5 min at 37°C were lysed in ice-cold 20 mM Tris-HCl (pH 8), 137 mM NaCl, 1% NP-40, 10% glycerol, 2 mM Na₃VO₄, and 1× protease inhibitor mixture (65621; Abcam) for 15 min. Lysates were centrifuged at 16,000 × g for 10 min to clear cell debris. Equal amounts of soluble protein per lane were separated by NaDodSO₄-polyacrylamide gel electrophoresis, transferred to a PDVF membrane, and immunoblotted using Abs against phosphotyrosine (4G10; MilliporeSigma) or SYK (D3Z1E; Cell Signaling Technology).

Generation of a mouse in which the Syk gene was replaced by one coding for SYK-AQL was accomplished using the targeting vector shown in Fig. 1A (26). In this animal, exon 10 contains the three mutations needed to produce SYK-AQL. Unfortunately, homozygous knock-in animals died shortly after birth despite a successful germline transmission of the mutant alleles. Such perinatal lethality occurs also in SYK-deficient mice (6, 22). Genotyping confirmed the presence of the Syk-AQL allele and absence of the wild-type Syk allele from the nonviable pups (Fig. 1B, lane 5). However, analysis by Western blotting of fetal liver cells from homozygous knock-in mice failed to detect any SYK or SYK-AQL protein, whereas cells from heterozygous animals exhibited one half the normal level of kinase (26). Thus, there was no apparent expression of protein from the knocked-in Syk-AQL allele.

To compensate for the lack of kinase expression in the knock-in animals, we generated a line of mice containing a bacterial artificial chromosome (BAC) encompassing the Syk locus but containing the R428Q, M429L, and M442A mutations found in SYK-AQL. For genotyping purposes, the mutant allele could be distinguished from wild-type Syk by the presence of a restriction site for PvuII that was introduced as result of the mutations (Fig. 1C). These BAC transgenic mice were then crossed with the knock-in mice that contained the unexpressed Syk-AQL alleles to generate animals homozygous for the BAC transgene but lacking both copies of Syk. Western blotting analysis of the lysates of isolated spleen B cells from homozygous mice revealed that SYK-AQL was present at a level ∼50% of that observed for SYK in B cells from control animals (Fig. 1D). These mice were viable, produced normal numbers of offspring, and suffered from no obvious pathological conditions. These animals are referred to in this study as Syk-AQL mice.

A major defect in immune cells reconstituted in lethally irradiated mice from SYK-deficient hematopoietic stem cells is a block in the pro- to pre-B cell and immature to mature B cell transitions (6, 22). As a consequence, B cells are absent from the spleens of radiation chimeric mice. We compared the spleens of Syk-AQL mice to those from wild-type mice for the presence of B cells to determine if SYK-AQL could rescue B cell development. The percentages of IgM⁺ (Fig. 2A), IgD⁺ (Fig. 2B), and B220⁺ IgM⁺ B cells (Fig. 2C, 2D) in the spleens of Syk-AQL mice were similar to those found in spleens from wild-type animals. Thus, B cell development was largely restored in mice expressing SYK-AQL in the absence of SYK. The most notable difference in the B cell populations between wild-type and Syk-AQL mice was the presence of fewer IgD^high and IgM^low B cells in the spleens of the Syk-AQL animals (Fig. 2D).

Analogue-sensitive kinases are designed to bind orthogonal inhibitors that fit exclusively into the active sites of the engineered enzymes but not those of the native kinases (23). This allows for the specific assignment of functions to the kinase of interest as possible off-target effects of the inhibitor are controlled for by the insensitivity of the wild-type enzyme to the inhibitor. To confirm this, we compared BCR signaling in wild-type versus Syk-AQL B cells in the presence or absence of the inhibitor. In response to cross-linking the BCR with anti‐IgM F(ab′)₂, B cells from both control and Syk-AQL animals increased in size and formed homotypic adhesions, resulting in the appearance of large clusters of cells characteristic of activated B cells (Fig. 3A, 3B). Treatment of cells with 3-MB-PP1, an inhibitor of analogue-sensitive kinases (27) shown previously to inhibit SYK-AQL (28), completely blocked the phenotypic changes induced by BCR engagement in Syk-AQL B cells but not wild-type B cells. 1-MN-PP1, another orthogonal inhibitor of analogue-sensitive kinases (29), had a similar effect in that it selectively blocked the activation of cells expressing SYK-AQL. The activation of B cells by cross-linking the BCR leads to the phosphorylation of multiple intracellular proteins in a manner dependent on the activity of SYK. The treatment of Syk-AQL B cells, but not wild-type B cells, with 1-NM-PP1 blocked this protein-tyrosine phosphorylation induced by anti-IgM F(ab′)₂ as determined by Western blotting of cell lysates with Abs against phosphotyrosine (Fig. 3C). Thus, SYK-AQL supports the activation of SYK-deficient B cells through the BCR, and this function can be selectively blocked by small molecule inhibitors that selectively target and inhibit the mutant enzyme.

We further examined cells for the appearance of CD86 and increased DNA synthesis in response to BCR engagement. CD86 is a cell surface marker found on activated B cells. Cross-linking the BCR of wild-type and SYK-AQL B cells with anti-IgM F(ab′)₂ induced the appearance on activated cells of CD86 to similar extents (Fig. 4). Treatment with 3-MB-PP1 completely blocked the appearance of CD86 on B cells from Syk-AQL mice but did not affect the wild-type B cells (Fig. 4). Treatment of both wild-type and Syk-AQL B cells with anti‐IgM F(ab′)₂ induced cell proliferation as measured by the incorporation of EdU into DNA (Fig. 5). 3-MB-PP1 did not inhibit BCR-induced DNA synthesis in wild-type cells but completely blocked DNA synthesis in Syk-AQL cells. Together, these results confirmed that the kinase activity of SYK is essential for the activation of primary B cells through the BCR.

We then compared the activation of wild-type and Syk-AQL B cells by LPS, which signals through a different mechanism that uses TLR4 (30). We first examined the viability of activated cells because SYK has been reported to be essential for the survival of B cells responding to LPS (18). The percentage of viable cells measured 48 h following treatment with LPS was not significantly different between wild-type cells and cells expressing SYK-AQL (Fig. 6A). As expected, the treatment of wild-type cells with 3-MB-PP1 had no effect on cell survival. Interestingly, however, 3-MB-PP1 also had no effect on the viability of LPS-activated cells expressing SYK-AQL. Thus, the activity of SYK was not important for the survival of LPS-treated cells in this model system.

To confirm that B cells of both types became activated in response to LPS, we monitored the expression of CD86. LPS treatment resulted in the expression of CD86 on splenic B cells from wild-type and Syk-AQL mice to similar extents (Fig. 6). Treatment of cells with 3-MB-PP1 had no effect on LPS-stimulated CD86 expression in either cell type.

We then examined DNA synthesis as a measure of cell proliferation. LPS enhanced the incorporation of EdU into DNA in both wild-type and Syk-AQL B cells to similar extents (Fig. 7). 3-MB-PP1 had no effect on LPS-induced cell proliferation in either wild-type or Syk-AQL B cells (Fig. 7). Thus, the activation of neither wild-type B cells nor SYK-AQL–expressing B cells by LPS was negatively influenced by the SYK-AQL inhibitor. Thus, in this model, the activation and survival of primary B cells in response to LPS is independent of SYK.

Analogue-sensitive kinases are generated by replacing the “gatekeeper” amino acid located within the ATP-binding pocket with a residue that has a smaller side chain (23). We had found previously that substitution of the gatekeeper methionine in SYK with either glycine [SYK(M442G)] or alanine [SYK(M442A)] resulted in enzymes and with little or no activity (24). We could restore activity to SYK(M442A), however, by replacing two additional amino acids whose side chains were located near the gatekeeper residue with the 2 aa found in the corresponding locations in SRC. The resulting mutant [SYK(R428Q/M429L/M442A)], termed SYK-AQL, was able to replace SYK in restoring BCR signaling to SYK-deficient DT40 B cells (24), mediating the changes in the cytoskeleton and enhancing cell adhesion when expressed in MD-MBA-231 breast cancer cells (31) and promoting autophagic clearance of stress granules in MCF7 cells (28). We felt, therefore, that this mutant enzyme would be well suited as a replacement for the wild-type kinase in a knock-in animal. Consistent with this assumption, most aspects of B cell development were retained in mice expressing SYK-AQL, but lacking SYK, including the restoration of nearly normal levels of IgM⁺ IgD⁺ B cells to the spleen. We observed a decrease in IgM^low IgD^high B cells in the spleens of Syk-AQL mice, which suggests either a decrease in B cell differentiation or a decrease in the survival of mature, follicular B cells. This may be related to the lower level of kinase, which is expressed at a level ∼50% of that of the wild-type enzyme, and/or to intrinsic differences in the activities of the wild-type and analogue-sensitive enzymes (32, 33). The B cells that developed in Syk-AQL mice, however, retained the ability to be stimulated through both surface IgM and TLR4.

One advantage of using an analogue-sensitive kinase is the ability to selectively inhibit its activity with a specially designed inhibitor that does not block the activity of the wild-type enzyme. Indeed, 3-MB-PP1 blocked the activation of B cells through the BCR selectively in cells expressing SYK-AQL. Thus, B cell activation is dependent on the activity of SYK. This result is consistent with many other studies in both cell lines (5) and in B cells isolated from mice in which the gene for Syk was inducibly deleted that indicate an essential function for the kinase in BCR-mediated activation (32, 33). These results indicate that cells isolated from Syk-AQL animals should be useful for the analysis of SYK-specific functions.

We then tested a possible role for SYK in the activation of B cells by LPS. Wild-type and SYK-AQL–expressing cells responded similarly to LPS by proliferating and upregulating the expression of CD86. However, in this case, the orthogonal SYK-AQL inhibitor had no effect on the activation or survival of either cell type. This indicates that, in this model system, the activation of primary B cells by LPS is not dependent on the activity of SYK. This observation is somewhat surprising in light of a recent study of SYK-deficient B cells that were generated from tamoxifen-treated mice that contain a loxP-flanked Syk allele and a tamoxifen-inducible Cre recombinase (18). Stimulation with LPS of the SYK-deficient B cells that survived the tamoxifen treatment led to extensive cell death unless an antiapoptotic protein (BCL-XL) was ectopically expressed. In our study, even though we used a high concentration of orthogonal inhibitor (10 μM) to ensure complete inhibition of SYK-AQL, we did not observe cell death in response to LPS in inhibitor-treated cells. Thus, one difference between these two B cell model systems may lie in the relative dependencies of the tested B cell populations on SYK activity for survival when cells are treated with LPS because the Syk-AQL mice contain proportionally more immature and marginal zone B cells than mature follicular B cells. Also, in this chemical genetic model, the kinase itself is retained during inhibition such that any scaffolding functions it might have also would be retained. Such scaffolding functions have been described for the related SYK family kinase, ZAP-70 (34).

In conclusion, we have generated a mouse that lacks the endogenous Syk gene and contains instead a gene coding for an analogue-sensitive mutant of SYK that can be inhibited selectively through the use of small molecule inhibitors that are readily available commercially. SYK-AQL supports both B cell development in vivo and B cell activation through the BCR in vitro in the absence of SYK. The BCR-dependent activation of B cells is totally dependent on the activity of the kinase. The activation of primary B cells from Syk-AQL mice by LPS proceeds normally but is not dependent on the kinase activity of SYK. Thus, the pathway that couples TLR4 to the expression of CD86, which requires MyD88 (35), is intact in the absence of the SYK activity.

The authors have no financial conflicts of interest.