Staphylococcus aureus induces cell-surface expression of immune stimulatory NKG2D ligands on human monocytes

extracellular SE-AF647,

Staphylococcus aureus is among the leading causes of bacterial infections worldwide. The pathogenicity and establishment of S. aureus infections are tightly linked to its ability to modulate host immunity. Persistent infections are often associated with mutant staphylococcal strains that have decreased susceptibility to antibiotics; however, little is known about how these mutations influence bacterial interaction with the host immune system. Here, we discovered that clinical S. aureus isolates activate human monocytes, leading to cell-surface expression of immune stimulatory natural killer group 2D (NKG2D) ligands on the monocytes. We found that expression of the NKG2D ligand ULBP2 (UL16-binding protein 2) is associated with bacterial degradability and phagolysosomal activity. Moreover, S. aureus-induced ULBP2 expression was linked to altered host cell metabolism, including increased cytoplasmic (iso)citrate levels, reduced glycolytic flux, and functional mitochondrial activity. Interestingly, we found that the ability of S. aureus to induce ULBP2 and proinflammatory cytokines in human monocytes depends on a functional ClpP protease in S. aureus. These findings indicate that S. aureus activates ULBP2 in human monocytes through immunometabolic mechanisms and reveal that clpP inactivation may function as a potential immune evasion mechanism. Our results provide critical insight into the interplay between the host immune system and S. aureus that has evolved under the dual selective pressure of host immune responses and antibiotic treatment. Our discovery of an immune stimulatory pathway consisting of human monocytebased defense against S. aureus suggests that targeting the NKG2D pathway holds potential for managing persistent staphylococcal infections.
The commensal bacterium Staphylococcus aureus is a major cause of infections ranging from superficial skin infections to life-threatening diseases like sepsis, endocarditis, and osteomyelitis (1). Treatment of S. aureus mainly relies on antibiotics, a strategy that is challenged by the ability of S. aureus to develop antibiotic resistance (2). The most crucial antimicrobial resistance issue in S. aureus has been the evolution and dissemination of methicillin-resistant S. aureus (MRSA) strains (3). MRSA infections are typically treated with last-line antibiotics, such as daptomycin or vancomycin (4,5). The emergence of daptomycin-resistant S. aureus isolates is associated with spontaneous genetic changes that also seem to impact bacterial virulence and enhance the risk of chronic infections (4,6). Thus, the dual selective pressure of daptomycin and the immune system may select for mutations that render the bacterium less susceptible to the antibiotic while increasing immune evasion and persistent infections. The NKG2D receptor-ligand system is a key immune sensing mechanism involved in detection of infection, inflammation, and malignant cells (7,8). The activating NKG2D receptor is expressed in a number of immune effector cells, in particular CD8 1 T cells and natural killer cells (9). NKG2D interacts with eight different NKG2D ligands that can be induced on the cell surface of stressed cells (10,11). NKG2D ligands are divided into two families: the major histocompatibility complex class I-related family containing polymorphic MICA and MICB and the UL16-binding protein family consisting of ULBP1-6 (9, 10). NKG2D ligandreceptor engagement leads to cytokine release, degranulation, immune cell activation, and influence specificity of the immune response (7,12). Cancer and virus-infected cells can evade immune recognition by obstructing cell-surface expression of NKG2D ligands (13,14). We previously showed that N-glycosylation is important for cell-surface expression of MICA and that U21 expressed by herpesvirus-7 blocked allele-specific MICA surface expression through altered N-glycosylation (15,16). Monocytes are involved in the early defense against a number of microbial infections (17,18). The response to S. aureus is mediated through phagocytosis, intracellular degradation, and cytokine secretion (19)(20)(21)(22). Human monocytes express Toll-like receptor 2 (TLR2) that is involved in the staphylococcal response, and stimulation by S. aureus or purified peptidoglycan (PGN) induces expression of tumor necrosis factor a, IL-1b, IL-6, and IL-10 (23)(24)(25)(26). Phagocytosis and intracellular degradation release bacterial components, which are sensed by a number of cytoplasmic receptors (27). Because of antibiotic exposure, many staphylococcal strains carry cell wall modifications like increased PGN crosslinking and O-acetylation that have profound effects on degradability, inflammation, and immunity (28,29).
In this study, we investigated immunity to a series of isogenic MRSA strains isolated from the blood of a patient with persistent bloodstream infection undergoing daptomycin treatment. Daptomycin targets the bacterial cell envelope, and during treatment the infecting MRSA isolate acquired a number of single-nucleotide polymorphism (SNPs) that rendered the new strains less susceptible to daptomycin (6,30). We find that human monocytes in response to S. aureus induce expression of NKG2D ligands through increased transcription. Mechanistically, we show that the up-regulation of the NKG2D ligand ULBP2 associates with lysozyme susceptibility of S. aureus and phagolysosomal activity in monocytes. Daptomycin exposure led to mutations in the rpoB and clpP genes, resulting in a thickened bacterial cell wall and increased resistance to lysosomal degradation, and we show that a mutation in the highly conserved S. aureus clpP gene hampered ULBP2-mediated immune activation and secretion of proinflammatory cytokines. Finally, we show that S. aureus mediates a metabolic change of monocytes characterized by lowered glycolysis and elevated intracellular level of (iso) citrate that is likely causally involved in NKG2D ligand regulation, because citrate and inhibitors of glycolysis directly induce ULBP2 expression in monocytes. Our study describes induction of NKG2D ligands by S. aureus on human monocytes and furthermore connects S. aureus with an altered metabolic phenotype causing surface expression of ULBP2 in human monocytes.

S. aureus induces cell-surface expression of NKG2D ligands on human monocytes
We investigated NKG2D ligand expression in response to five clinical MRSA strains, SADR-1-SADR-5 (6). The strains developed reduced susceptibility to daptomycin via acquisition of SNPs in rpoB, clpP, clpX, mprF, and other genes (Fig. 1A). We previously described the contribution of the individual SNPs to antibiotic resistance and fitness of strains SADR-1-SADR-4 (30). SADR-5 was isolated from the same patient and shares the SNP in rpoB; however, it has four additional SNPs not present in SADR-2-SADR-4 (Fig. 1A).
We stimulated human monocytes (THP1 cells) with UVkilled S. aureus SADR-1-SADR-5 and assessed NKG2D ligand surface expression by flow cytometry after 12, 24, and 48 h. Upon S. aureus exposure, we found that ULBP2 was induced on the surface of THP1 with expression increasing over time (Fig. 1B). Interestingly, SADR-1 and SADR-5 induced a pronounced ULBP2 response compared with SADR-2-SADR-4 ( Fig. 1, B and C). In addition, we examined ULBP2 expression after exposure to another MRSA strain USA300JE2 and found this was as potent as SADR-1 and SADR-5 in inducing ULBP2 in THP1 cells (Fig. 1, B and C). Furthermore, USA300JE2 and SADR-1 induced ULBP2 in a dose-dependent manner (Fig.   S1A). Of note, expression of the NKG2D ligands MICA/B, ULBP1, ULBP3, and ULBP4 was not affected by exposure to any of the S. aureus strains in THP1 cells (Fig. S1, B and C). ULBP2 mRNA levels were significantly induced in monocytes exposed to SADR-1, SADR-5, and USA300JE2 for 24 h, whereas induction was lower in monocytes exposed to SADR-2-SADR-4 ( Fig. 1D), corresponding to the observed ULBP2 surface expression (Fig. 1, B and C). To further verify the NKG2D ligand response to S. aureus, we investigated NKG2D ligand expression in primary monocytes. As shown in Fig. 1E, ULBP2 transcription was increased in SADR-1-treated primary monocytes after 4 and 8 h. Interestingly, we saw that also MICA/B transcription and surface expression were increased in these cells (Fig. S1, D and E), whereas no surface expression was detected for ULBP1, ULBP3, and ULBP4 (Fig. S1F). The absence of MICA/B expression in THP1 cells is likely explained by a specific defect in these cells, in line with previous findings (31). Moreover, we saw a high donor variation in expression of the individual NKG2D ligands, an observation previously described in human monocytes (32).
We tested ULBP2 expression in response to live SADR-1 and USA300JE2. Consistent with our data using UV-killed strains, live S. aureus significantly increased expression of ULBP2 on both the surface (Fig. 1F) and mRNA levels (Fig. 1G). The response was even more potent for live USA300JE2 compared with UV-killed (Fig. 1G), likely because of bacteria proliferating within the monocyte culture medium in accordance with the observed dose-dependent ULBP2 induction by UV-killed strains (Fig. S1A). Together, these data show that S. aureus induces transcriptional expression of NKG2D ligands, ULBP2 in THP1 and primary monocytes, and ULBP2 and MICA/B in primary monocytes.
We next investigated the functional outcome of the S. aureus-mediated ULBP2 expression. First, we used a soluble NKG2D-Fc receptor to assess whether S. aureus-induced ULBP2 expression promoted binding to the NKG2D receptor. As shown in Fig. 2A, the soluble NKG2D receptor interacted with monocytes in a pattern that correlated with ULBP2 expression (Figs. 1C and 2B).
The data demonstrate that S. aureus induces ULBP2 expression on THP1 cells and that the mutations in rpoB and clpP acquired in SADR-2 during exposure to daptomycin reduce the ability to induce ULBP2. Mutations in rpoB are associated with the global emergence of stable, multidrug-resistant staphylococcal lineages (33,34), whereas the highly conserved clpP is important for stress tolerance and virulence. Interestingly, ClpP exerts many of its functions through interaction with ClpX (35). We find it intriguing that these mutations also impact the ability of S. aureus to cause immune activation by modifying surface expression of ULBP2 on monocytes and therefore focused primarily on SADR-1 and SADR-2 in the remainder of this study.
SADR-2 fails to induce NKG2D-mediated immune response compared with SADR-1 Next, we examined whether induced ULBP2 resulted in signaling through the NKG2D receptor. We used an established 2B4 cell line that expresses GFP in response to NFAT (nuclear factor of activated T cell) activity, and the reporter cells were further transfected with or without human NKG2D/DAP10-CD3z complex; these cells thus function as reporter cells that become GFP-positive upon NKG2D ligation (36)(37)(38). THP1 cells were exposed to SADR-1 or SADR-2 for 48 h to allow surface expression of ULBP2 and were then co-cultivated with 2B4_NKG2D pos or 2B4_NKG2D neg cells. Monocytes exposed to SADR-1 induced a significantly stronger signal through NKG2D than cells exposed to SADR-2, whereas no GFP expression was observed in 2B4_NKG2D neg control cells (Fig. 2C). Pretreatment of monocytes with a soluble NKG2D-Fc receptor completely abrogated the GFP expression, confirming that the activation was mediated by NKG2D (Fig. S2A). Expression of human NKG2D on 2B4_NKG2D pos cells was confirmed (Fig. S2B), and specificity of ULBP2 interaction with NKG2D was confirmed by blocking ULBP2 prior to analysis of NKG2D-fc interaction (Fig. S2C). S. aureus induces cell-surface expression of ULBP2 on human monocytes. A, schematic overview of the clinical, isogenic S. aureus strains SADR-1-SADR-5 used in this study. Arrows represent acquired SNPs and decreasing daptomycin susceptibility that the isolates have gained at each step in their developmental path. SADR-1 is the infecting isolate. SADR-2 and SADR-3 have decreased susceptibility to daptomycin, whereas SADR-4 and SADR-5 are daptomycin-resistant. Resequencing isolates revealed a mutation in clpP G94D in SADR-2 not previously reported (6,30). B, human THP1 cells were treated with SADR-1-SADR-5, USA300JE2, or PBS (untreated). Surface expression of ULBP2 was analyzed after 12, 24, and 48 h by flow cytometry. C, THP1 cells were treated with SADR-1-SADR-5 or USA300JE2. Surface expression of ULBP2 and isotype background were analyzed after 48 h by flow cytometry. Dot plots show percentage ULBP2 surface expression (upper row) and isotype staining (lower row) and are representative of three independent experiments. D, THP1 cells were treated with SADR-1-SADR-5 or USA300JE2. Total RNA was purified after 12 and 24 h and analyzed for ULBP2 expression by quantitative real-time PCR. E, human primary monocytes were treated with SADR-1. Total RNA was purified after 4 and 8 h and analyzed for ULBP2 expression by quantitative real-time PCR. F, THP1 cells were treated with UV-killed or live SADR-1 or USA300JE2. Surface expression of ULBP2 was analyzed after 24 and 48 h by flow cytometry. G, THP1 cells were treated with UV-killed or live SADR-1 or USA300JE2. Total RNA was purified after 24 h and analyzed for ULBP2 expression by quantitative real-time PCR. The data are presented as follows: B and F show surface expression as MFI 6 S.D. of data from three independent experiments; D and G show data from four or three independent experiments, respectively, presented as fold change 6 S.D. of ULBP2 mRNA normalized to HKG; E shows data from three independent experiments (two for untreated sample) presented as ULBP2 mRNA normalized to HKG 6 S.D. Statistical analysis was performed by two-way ANOVA with Tukey's multiple comparisons test for data in B, D, and, F and one-way ANOVA with Dunnett's multiple comparisons test for data in G. Statistics are presented relative to untreated control. FSC, forward scatter; ns, not significant.
Following ligand interaction, the NKG2D receptor is downregulated from the cell surface (39)(40)(41). We therefore investigated NKG2D receptor expression on primary natural killer and CD8 1 T cells after co-cultivation with SADR-1-or SADR-2-pretreated monocytes. Prior to co-cultivation, monocytes were incubated with either soluble NKG2D-Fc or IgG1-Fc control. Consistent with the NKG2D signaling data, SADR-1 induced a more potent NKG2D down-modulation than SADR-2 for all tested Effector:Target (E:T) ratios (Fig. 2D). Collectively, these data show that SADR-1 induces a potent, functional NKG2D ligand response compared with SADR-2, highlighting that the immune stimulatory structures on SADR-1 facilitating ULBP2 expression are altered by the mutations in SADR-2.

SADR-2 accumulates intracellularly in human monocytes compared with SADR-1
Myeloid responses to bacteria are orchestrated by signals from surface receptors, as well as through phagocytosis and degradation (27,42). Phagocytosis and intracellular degradation initiate inflammasome activation and secretion of mature IL-1b (43,44). Because S. aureus-induced ULBP2 surface expression occurred relatively late after stimulation (Fig. 1B), we hypothesized that ULBP2 induction by S. aureus was linked to phagocytosis and degradation. We therefore investigated cellular localization of SADR-1 and SADR-2 in THP1 cells by confocal microscopy. This was done by labeling S. aureus isolates with Alexa Fluor 647-conjugated succinimidyl-ester (AF-SADR) prior to stimulating monocytes. After 48 h, both SADR- Figure 2. SADR-2 fails to induce NKG2D-mediated immune response compared with SADR-1. THP1 cells were treated with SADR-1-SADR-5 or USA300JE2. Functional ULBP2 expression was assessed by interaction with AF647-conjugated NKG2D-Fc after 48 h by flow cytometry, presented as follows: A, dot plots showing forward-side scatter and percentage gated cells (top row), percentage Iso-fc surface interaction (middle row), and NKG2D-Fc surface interaction (bottom row), representative for three independent experiments; and B, bar graph showing NKG2D-Fc surface interaction as MFI of data from three independent experiments. C, NKG2D signaling was assessed by co-cultivation of effector cells (THP1 cells) with target cells (2B4_NKG2D neg or 2B4_NKG2D pos cells) at the indicated effector-target (E:T) ratios for 14-16 h prior to analysis by flow cytometry. Before co-cultivation effector cells were treated with SADR-1 or SADR-2 for 48 h, and target cells were labeled with the membrane dye DID. Signaling through the NKG2D receptor upon interaction with ULBP2 was assessed as GFP expression by flowcytometry. Graph depicts percentage GFP expression from three independent experiments. D, NKG2D down-modulation was assessed by co-cultivation of effector cells (THP1 cells) with target cells (NK/CD8 1 T cells) at the indicated effector-target (E:T) ratios for 2.5 h prior to analysis by flow cytometry. Prior to co-cultivation, effector cells were treated with SADR-1 or SADR-2 for 48 h and treated with NKG2D-Fc (to block NKG2D ligands) or IgG1-Fc (control). Surface expression of NKG2D on target cells was measured by flow cytometry. Graph depicts surface NKG2D expression relative to NKG2D blocking on target cells from three donors in independent experiments. Statistical analysis was performed by one-way ANOVA with Dunnett's multiple comparisons test for data in B, two-way ANOVA with Tukey's multiple comparisons test for data in C, and two-way ANOVA with Dunnett's multiple comparisons test for data in D. Statistics are presented relative to untreated control and SADR-2 in C. SCC, side scatter; FSC, forward scatter; ns, not significant.
1 (AF-SADR-1) and SADR-2 (AF-SADR-2) were observed inside the monocytes, with relatively more SADR-2 than SADR-1 being visible (Fig. 3A). Next, we used flow cytometry to quantify intracellular S. aureus and found that the percentage of intracellular AF-SADR-2 was substantially higher compared with AF-SADR-1 at 24 h and significantly increased at 48 h (Fig. 3, B and C). This was not an artifact of unequal labeling of SADR-1 and SADR-2, because analysis of labeled and unlabeled S. aureus confirmed staining of the entire bacterial populations (Fig. S3A). Internalization of S. aureus was confirmed by analyzing the cell culture medium for labeled S. aureus at 48 h after exposure. Few bacteria were detected extracellularly (Fig. S3B), suggesting that SADR-1 is internalized and steadily degraded by the monocytes, whereas SADR-2 is internalized but accumulates intracellularly. Finally, we assessed intracellular association of all S. aureus strains by flow cytometry, showing significant accumulation of AF-SADR-3, AF-SADR-4, and interestingly AF-USA300JE2, whereas AF-SADR-5 resembled AF-SADR-1 (Fig. 3C). Labeling of S. aureus and bacteria in cell culture medium was verified by flow cytometry (Fig. S3C).

SADR-2 is more resistant to lysozyme-mediated degradation compared with SADR-1
Alterations in the staphylococcal cell wall facilitated by antibiotic treatment can alter susceptibility to lysozyme and phagosomal degradation (45)(46)(47), leading to fewer bacterial cell wall components released intracellularly, thus reducing inflammatory signaling and immune activation (43,44). We therefore examined whether exposure to daptomycin had rendered SADR-2 resistant to degradation by lysozyme, an integral phagolysosomal enzyme. We cultivated SADR-1 and SADR-2 in the presence of lysozyme and subsequently monitored S. aureus membrane integrity through bacterial density (A 600 ) measurements. The optical density of the SADR-1 culture dropped (a sign of cell lysis) following addition of lysozyme (Fig. 4A). The optical density of the SADR-2 culture remained stable, and after 24 h the difference in bacterial growth was significantly reduced for SADR-1 compared with SADR-2, indicating that the acquired mutations in SADR-2 have made this isolate less susceptible to lysozyme (Fig. 4A).
We next investigated whether the induction of ULBP2 was associated with phagosomal degradation of S. aureus. First, we blocked degradation using bafilomycin, a selective vacuolar proton pump inhibitor (48). Bafilomycin treatment resulted in significantly reduced ULBP2 transcription, which was more pronounced for cells exposed to SADR-1 than SADR-2 ( Fig.  4B). Similarly, the specific lysozyme inhibitor TriNAg (43,49,50) significantly decreased ULBP2 transcription in monocytes exposed to SADR-1, whereas ULBP2 transcription was only slightly reduced in cells exposed to SADR-2 ( Fig. 4C). Cellular viability after inhibitor stimulation for 24 h was verified by annexin V and PI stain (Fig. S4). Collectively, these data indicate that phagosomal degradation of S. aureus is required for induction of ULBP2 expression and that SADR-2 is more resistant to degradation by lysozyme than SADR-1. A, THP1 cells were treated with AF647-labeled UV-SA (AF-SADR-1 or AF-SADR-2) for 48 h. Intracellular localization of labeled UV-SA was visualized by confocal microscopy. The cell surface was labeled with AF488-CD45 (green), and the nucleus was stained with Hoechst (blue) and AF647-labeled UV-SADR (magenta). The rows show overview of cells (top row), selected Z-plan projected as x axis versus y axis (middle row), and 3D structure presented as maximal intensity projection (bottom row). B, THP1 cells were treated with AF-SADR-1 or AF-SADR-2 for 24 or 48 h. Internalized bacteria were assessed by flow cytometry. Dot plots show percentage of internalized AF647-labeled UV-SA associated with viable THP1 cells. The data in A and B are representative of three independent experiments. C, THP1 cells were treated with AF-SADR-1-SADR-5 or AF-USA300JE2 for 48 h. Internalized bacteria were assessed by flow cytometry. Bar graph shows percentages of internalized AF647-labeled UV-SA associated with viable THP1 cells from three independent experiments. Statistical analysis was performed by one-way ANOVA with Dunnett's multiple comparisons test and presented relative to untreated control for data in C. FSC, forward scatter; ns, not significant.
Stimulation of TLR2 on human monocytes can facilitate IL-6 and IL-10 release, whereas IL-1b secretion is linked to staphylococcal degradability (lysozyme sensitivity), release of inflammasome-activating components, and caspase 1 activity (18,24,25,43,47,51). We therefore tested secretion of IL-1b, IL-6, and IL-10 in culture supernatants from monocytes exposed to SADR-1 or SADR-2. We found that only SADR-1 induced cytokine responses in the monocytes (Fig. 4, D-F). IL-1b secretion was rapidly induced and continued to increase during 48 h of stimulation (Fig. 4D), whereas secretion of IL-6 and IL-10 was initiated after 24 h (Fig. 4, E and F). These findings were supported by a significantly higher NF-kB activity in THP1-NF-kB-GFP reporter cells (52) responding to SADR-1 compared with SADR-2 ( Fig. 4, G and H). Investigating NF-kB activity in response to all S. aureus isolates further revealed that SADR-3 and SADR-4 resembled SADR-2, whereas SADR-5 and USA300JE2 similar to SADR-1 increased activity significantly ( Fig. 4H). NF-kB is important for production of pro-IL-b, whereas secretion of mature IL-1b depends on caspase 1 activity (43). To this end, we tested involvement of inflammasome activity on S. aureus-induced ULBP2 surface expression and cytokine secretion through treatment with the caspase 1 inhibitor, ac-YVAD-cmk (YVAD) (53). As expected, IL-1b secretion mediated by SADR-1 stimulation was significantly inhibited by YVAD (Fig. 4I). On the other hand, SADR-1-induced ULBP2 surface expression was significantly up-regulated by YVAD treatment for 48 h (Fig. 4J). This was in line with IL-6 ( Fig. 4K) and IL-10 ( Fig. 4L) secretion by SADR-1-stimulated cells, suggesting that ULBP2 surface expression is not dependent on inflammasome activation but is induced by a similar route as IL-6 and IL-10. Together, these data suggest that SADR-1 is more sensitive to degradation than SADR-2, resulting in higher immune activity of monocytes through both inflammasome-dependent (IL-1b) and -independent (ULBP2, IL-6, and IL-10) pathways.

Accumulation of intracellular (iso)citrate induces ULBP2 expression
Citrate is an integral tricarboxylic acid (TCA) cycle metabolite that is essential for supplying immune cells with energy and metabolites (54)(55)(56). Citrate is readily converted to its isomer, isocitrate, through aconitase activity (55). In phagocytic cells, (iso)citrate and derived molecules are important for the production of reactive oxygen species, NO, and prostaglandins making it a key metabolite in microbial immunity (57)(58)(59)(60). In activated myeloid cells, induced activation of the citrate carrier is known to increase transport of (iso)citrate from mitochondria to the cytosol in exchange for malate (57,59,60). Lipopolysaccharide and S. aureus cause intracellular citrate increase in macrophages that results in inflammasome activation and IL-1b secretion (44).
To investigate a potential link between (iso)citrate buildup and ULBP2 expression, we treated monocytes with citrate alone or combined with SADR-1. We observed that citrate induced surface expression of ULBP2, and this effect was not additive with SADR-1 exposure (Fig. 5D), implying that SADR-1 and citrate induce ULBP2 through a similar signaling pathway. In agreement with this, both citrate and SADR-1 increased ULBP2 transcript levels ( Fig. 5E). Interestingly, the citrate-mediated ULBP2 gene transcription occurred more rapidly than observed for SADR-1-exposed cells (Fig.  5E). The citrate analog hydroxycitrate (HC), has previously been used to inhibit the action of citrate (63,64). In line with this, HC significantly inhibited SADR-1-mediated ULBP2 surface expression (Fig. 5F). Integrity and viability of the cells was verified by PI stain (Fig. S5). Together, these data indicate that the SADR-1-mediated (iso)citrate accumulation is linked to degradation of SADR-1, altered intracellular metabolism, and ULBP2 expression.

Inhibition of the glycolytic flow increases ULBP2 expression
Breakdown of S. aureus in phagosomes results in PGN degradation and release of intracellular GlcNAc, which induces a potent IL-1b response in lipopolysaccharide-primed macrophages (44). Because ULBP2 induction correlated with the ability to degrade S. aureus, we speculated whether GlcNAc released from S. aureus associated with ULBP2 expression. We found that GlcNAc treatment induced cell-surface expression of ULBP2 to some extent ( Fig. 6A and Fig. S6A).
Cytoplasmic GlcNAc was shown to interact specifically with hexokinase, an essential enzyme in the glycolytic pathway . SADR-2 is more resistant to lysozyme-mediated degradation compared with SADR-1. A, SADR-1 and SADR-2 were grown to early stationary phase. The bacterial culture were then divided and mixed with either 300 mg/ml lysozyme or TSB as a control (0 h). Bacterial density was assessed by A 600 measurement after 18, 21, and 24 h. The graphs show measurements from three independent experiments. B, THP1 cells were treated with SADR-1 or SADR-2, alone or combined with bafilomycin (50 nM). Total RNA was purified after 24 h and analyzed for ULBP2 expression by quantitative real-time PCR. Bar graphs in B and C show data from three independent experiments presented as fold change 6 S.D. of ULBP2 mRNA normalized to the HKG and relative to the untreated, bafilomycin, or TriNAg controls. C, THP1 cells were treated with SADR-1 or SADR-2 alone or combined with TriNAg (0.5 mg/ml). Total RNA was purified after 24 h and analyzed for ULBP2 expression by quantitative real-time PCR. D and F, THP1 cells were treated with SADR-1 or SADR-2 for 48 h. Cytokines were measured in culture supernatants by ELISA at 1, 3, 6, 12, 24, and 48 h. Graphs show concentrations of secreted cytokines in supernatants from three independent experiments in IL-1b (D), IL-6 (E), and IL-10 (F). G and H, THP1-NF-kB-GFP reporter cells were treated with indicated concentrations of SADR-1 or SADR-2 for 24 h in G and SADR-1-SADR-5 or USA300JE2 (1:20) in H. NF-kB activity was assessed as GFP expression by flow cytometry. Graphs depict MFI of GFP expression from three independent experiments as fold change 6 S.D. I-L, THP1 cells were treated with SADR-1 or SADR-2 for 24 and 48 h alone or in combination with YVAD (25, 50, or 100 mM). Concentration of cytokines were measured in culture supernatants by ELISA in I, K, and L and ULBP2 surface expression by flow cytometry presented as MFI in J. Graphs show data collected from three independent experiments IL-1b (I), surface ULBP2 (J), IL-6 (K), and IL-10 (L). Statistical analysis was performed by two-way ANOVA with Tukey's multiple comparisons test for data in A-G and I-L, and one-way ANOVA with Dunnett's multiple comparisons test for data in H. Statistics are presented relative to untreated control and SADR-2 in D-F, relative to untreated in H, and relative to mock-treated (-) in I-L. ns, not significant. resulting in reduced glycolytic flow (44). Furthermore, accumulated cytosolic citrate blocks the glycolytic pathway by inhibiting phosphofructokinase (44,65). This prompted us to investigate ULBP2 expression in response to the glycolytic inhibitor 2-deoxy-D-glucose (2DG) (15,66). Similar to citrate treatment, 2DG significantly induced ULBP2 expression (Fig. 6B). In addition, combined treatment of 2DG and S. aureus resulted in higher ULBP2 expression compared with S. aureus exposure alone (Fig. 6B). For SADR-2-SADR-4, addition of 2DG increased ULBP2 induction to a level similar to SADR-1, SADR-5, and USA300JE2 (Fig. 6B). 2DG also blocks the N-glycosylation pathway, which we have previously shown is important for surface expression of the NKG2D ligand MICA (15,16,66). Addition of D-mannose rescues the inhibition of N-glycosylation by 2DG (15). We confirmed by D-mannose treatment that altered N-glycosylation was not causing the regulatory effect on ULBP2 expression by 2DG (Fig. S6B). Similar to S. aureus-mediated ULBP2 regulation, 2DG was found to increase ULBP2 by increased gene transcription (Fig. 6C). Like citrate, 2DG treatment induced ULBP2 mRNA more rapidly than SADR-1 (Figs. 5E and 6C). These data suggest that the immediate effect of glycolytic inhibition is related to ULBP2 expression, and this mechanism is delayed in S. aureusexposed cells because of the initial need for phagocytosis and degradation.

S. aureus reduces glycolysis in human monocytes
Glycolysis and mitochondrial oxidative phosphorylation are important for the proinflammatory cytokine response mediated by TLR2 stimulation in myeloid cells, and mitochondrial activity seems essential for phagocytosis and bacterial degradation (67,68). Given that ULBP2 expression associates with staphylococcal degradability and is induced by direct inhibition of glycolysis, e.g. by increased intracellular GlcNAc or citrate treatment, it is likely that S. aureus phagocytosis and degradation ultimately impairs glycolysis. We assessed the glycolytic flow and capacity in monocytes after exposure to SADR-1 or SADR-2 by performing a glycolytic stress test using Seahorse extracellular flux analysis. The cells were glucose-starved for 1 h before glucose was added to measure the current reliance of Bar graph shows surface expression as MFI 6 S.D. of data from three independent experiments. Statistical analysis was performed by two-way ANOVA with Tukey's multiple comparisons test for data in A-C and F, one-way ANOVA with Tukey's multiple comparisons test for data in D, and two-way ANOVA with Dunnett's multiple comparisons test for data in E. Statistics are presented relative to untreated control in D and E and relative to untreated or SADR-1 (2) in F. Note that untreated and SADR-1-treated (6, 12, and 24 h) samples in E are also presented as controls in Fig. 6C (because citrate and 2DG treatment were performed in the same assay). ns, not significant.
glycolysis. We observed a decrease in glycolysis in SADR-1-or SADR-2-exposed cells compared with untreated cells at 24 h (Fig. 7, A and B). Furthermore, the glycolytic capacity and glycolytic reserve were increased in response to SADR-1 treatment (Fig. 7, A and B). In agreement with previous findings (67), this shows an increased capacity in S. aureus-stimulated cells to utilize glycolysis when mitochondrial activity is blocked (by addition of oligomycin). In contrast SADR-2 only slightly induced the glycolytic capacity (Fig. 7B).
To investigate glucose metabolism in S. aureus-exposed monocytes, we exposed cells to SADR-1 or SADR-2 in the presence of 13 C-labeled glucose. The results showed a significant reduction in labeling of central metabolites in both the first and second turns of the TCA cycle upon stimulation with SADR-1 compared with untreated or SADR-2-stimulated cells (Fig. 7, C and D). A similar although less pronounced labeling pattern was observed after 12 h (Fig. S7, A  and B). The glycolytic contribution to the TCA cycle is therefore decreased after SADR-1 exposure. To study the effect of SADR-1 and SADR-2 on aerobic glycolysis, we measured intracellular lactate levels and saw a significant reduction in intracellular lactate levels after 24 h (Fig. 7E). The metabolic tracing studies are thus in line with the observed reduced glycolysis after SADR-1 exposure.

S. aureus-induced ULBP2 depends on functional mitochondria
To maintain ATP production, the cells can adapt to inhibited glycolysis by increasing mitochondrial activity (69). Because SADR-1 induced ULBP2 in a manner similar to glycolytic inhibitors, we examined whether ULBP2 induction by SADR-1 depended on functional mitochondria. Uncoupling the mitochondrial membrane potential with the ionophore, FCCP resulted in inhibition of ULBP2 by SADR-1 (Fig. 7F). We further tested etomoxir, an inhibitor of mitochondrial fatty acid import (CPT1a inhibitor), which was recently shown to act as a general inhibitor of mitochondrial function (70). Similar to FCCP, etomoxir impaired SADR-1-induced ULBP2 expression (Fig. 7G). Cellular viability was verified by PI stain after 48 h of incubation with FCCP (Fig. S7D) or etomoxir (Fig. S7E). Together, this suggests that SADR-1 promotes ULBP2 expression in monocytes in a mitochondrial-dependent manner.
Collectively, these data indicate that SADR-1 activates monocyte expression of ULBP2 after bacterial phagocytosis and degradation, which inhibits glycolysis and thereby promotes mitochondrial-dependent ULBP2 expression. On the other hand, SADR-2 is less degradable, releasing low amounts of intracellular signaling molecules, and hence accumulates silently intracellularly. . Surface expression of ULBP2 was analyzed after 48 h by flow cytometry. C, THP1 cells were treated with SADR-1, 2DG (10 mM), or combined SADR-1 and 2DG. Total RNA was purified after 3, 6, 12, and 24 h and analyzed for ULBP2 expression by quantitative real-time PCR. Bar graph shows data from three independent experiments, presented as fold change 6 S.D. of ULBP2 mRNA normalized to the HKG. Data in A and B show surface expression as MFI 6 S.D. of data from three independent experiments. Statistical analysis was performed by oneway ANOVA with Tukey's multiple comparisons test for data in A, two-way ANOVA with Sidaks's multiple comparisons test for data presented in B, and two-way ANOVA with Dunnett's multiple comparisons test for data presented in C. Statistics are presented relative to untreated control in C. Note that untreated and SADR-1-treated (6, 12, and 24 h) samples in C are also presented as controls in Fig. 5E (because citrate and 2DG treatment were performed in the same assay). ns, not significant.

Reversion of clpP in SADR-2 restores its immune stimulatory phenotype
To determine whether the hampered immune stimulatory phenotype of SADR-2 compared with SADR-1 was related to the SNPs in rpoB and/or clpP, four new S. aureus isolates were generated. First, the mutant clpP allele present in SADR-2 was reverted to the WT clpP in SADR-1. Two biological independent clones confirmed not to harbor unrelated secondary mutations were selected for further analysis (SADR-2 clpP_rev_A and SADR-2 clpP_rev_B ). Second, the mutant alleles of rpoB and clpP present in SADR-2 were separately introduced in SADR-1 (SADR-1 rpoB_mut and SADR-1 clpP_mut ). Of note, SADR-1 rpoB_mut Figure 7. S. aureus-induced ULBP2 depends on functional mitochondria and inhibited glycolysis. A and B, THP1 cells were treated with SADR-1 or SADR-2. Glycolytic stress test was measured by Seahorse analysis after 24 h. The data are presented as the fluctuation in ECAR (A) and the glycolytic capacity, glycolytic reserve, and glycolysis measured presented as ECAR normalized to protein content (B, described in detail in "Materials and methods"). C-E, THP1 cells were treated with SADR-1 or SADR-2 in medium containing 13 C-labeled glucose. The processing of labeled glucose was analyzed by GC-MS to determine the percentage of 13 C distribution in intracellular metabolites. The data are presented as percentages of 13 C-labeled metabolite (described in detail in "Materials and methods") for 13 C incorporation in key TCA cycle metabolites after 24 h from the first turn in the cycle (C), from the second turn (D), and 13 C incorporation in intracellular lactate (E). F, THP1 cells were treated with SADR-1 alone or combined with FCCP (2,5 mM). Surface expression of ULBP2 was analyzed after 48 h by flow cytometry. G, THP1 cells were treated with SADR-1 alone or combined with etomoxir (400 mM). Surface expression of ULBP2 was analyzed after 48 h by flow cytometry. A and B show ECAR 6 S.E. from eight independent experiments. C-E show percentages of 13 C labeling 6 S.E. from four independent experiments. Bar graphs in F and G show data from three independent experiments presented as fold change 6 S.D. of ULBP2 surface expression relative to the untreated, FCCP, or etomoxir controls. Statistical analysis was performed by two-way ANOVA with Tukey's multiple comparisons test for data in B-E, and unpaired two-tailed t test in F and G. ns, not significant. was a clean mutant, whereas SADR-1 clpP_mut carried two missense mutations in oleate hydratase and endonuclease mut S2 in addition to the clpP mutation.
We investigated ULBP2 mRNA and surface expression in THP1 cells after stimulation with the new S. aureus isolates. Interestingly, we found that SADR-2 with the reverted clpP (SADR-2 clpP_rev_A and SADR-2 clpP_rev_B ) similar to SADR-1 significantly up-regulated ULBP2 on the surface (Fig. 8, A and B) and mRNA level in THP1 cells (Fig. 8C). Moreover, introducing the clpP mutation in SADR-1 (SADR-1 clpP_mut ) did not induce ULBP2, whereas the rpoB mutation in SADR-1 (SADR-1 rpoB_mut ) had no influence on the ability of SADR-1 to increase ULBP2 expression (Fig. 8, A-C). Finally, no other NKG2D ligands than ULBP2 were induced on the THP1 cell surface in response to either S. aureus isolate (Fig. S8A).
To assess the functional outcome of the clpP and rpoB mutations on monocyte immune activity, we investigated NKG2D receptor interaction by the S. aureus-regulated ULBP2. We found an interaction pattern of the NKG2D receptor similar to the ULBP2 expression pattern, showing that SADR-1, SADR-2 clpP_rev_A , and SADR-2 clpP_rev_B , and SADR-1 rpoB_mut significantly increased NKG2D interaction compared with SADR-2 and SADR-1 clpP_mut (Fig. 8, A and D).

Discussion
We show that NKG2D ligands are induced in human monocytes in response to S. aureus. We demonstrate that ULBP2 expression is mediated through enhanced gene expression related to intracellular degradation of S. aureus. Degradation of S. aureus and increased cytoplasmic (iso)citrate associates with ULBP2 expression. Furthermore, our results suggest that SADR-1-induced ULBP2 is linked to a shift in the intracellular metabolism from glycolysis to mitochondrial oxidation. Regulation of NKG2D ligands have been coupled to DNA damage and reactive oxygen species production often mediated by external stressors (9, 12, 71), but it has not previously been linked to an immune-metabolic mechanism in human monocytes. Previous studies have associated TLR signaling with NKG2D ligand expression on human monocytes and exposure to S. aureus in mouse macrophages (40,72,73). Our data suggest that ULBP2 is induced through an inflammasome-independent pathway. We furthermore show that S. aureus-mediated ULBP2 expression in THP1 cells is abrogated by clpP mutation but not rpoB.
Exposure to antibiotics can alter cell wall structure and intracellular degradation of S. aureus that potentiates and directs the immune response (43,44,47,74,75). We found that SADR-2 was more resistant to lysozyme-mediated degradation than SADR-1. SADR-2 has a structurally altered cell wall compared with SADR-1, likely because of the mutations in rpoB and clpP causing reduced daptomycin susceptibility and virulence (30). Lysozyme targets PGN (46), and pathogenic S. aureus can modify their PGN structure via O-acetylation through O-acetylase A (OatA), activity making the bacteria resistant to lysozyme (45,46). Daptomycin exposure has not previously been linked to alterations in PGN. However, the stress regulator Spx that is up-regulated in SADR-2-SADR-4 is important for regulating OatA activity in other Gram-positive bacteria (30,76), indicating a possible link between daptomycin-mediated cell wall alterations and lysozyme sensitivity via OatA activity.
GlcNAc is one of the main components in the cross-linked network of carbohydrates present in bacterial PGN and is released upon intracellular degradation of S. aureus (28). When investigating the direct influence of GlcNAc on ULBP2 expression, we observed a modest induction of ULBP2. This is likely explained by a limited cellular uptake of GlcNAc even when complexed with transfecting agent or prolonged exposure to increase cytosolic delivery, as previously described (44). An important effect of cytosolic GlcNAc is inhibition of glycolysis (44,77). Likewise, we found that the glucose analog 2DG, which is efficiently taken up via glucose transporters and blocks the glycolytic pathway (15,66), significantly increased ULBP2 transcript and surface expression.
When investigating metabolic activity in the monocytes responding to S. aureus, we found that in particular SADR-1 reduced glycolysis but increased the cellular glycolytic capacity, the latter likely induced by an early (surface interactioninduced) glycolytic activity, corresponding with intracellular lactate levels at 12 h (Fig. S7C) and previous findings (67). In addition the intracellular isocitrate levels were increased after SADR-1 exposure, and there was no apparent increase in citrate levels. In other studies we have found that the intracellular citrate concentration is strongly regulated, so increased citrate production can mainly be measured by increased flux and not by increased concentration. 8 Anyway, citrate is likely involved in ULBP2 expression because it significantly induced ULBP2. In support of this, the citrate inhibitor hydroxycitrate potently inhibited SADR-1-induced ULBP2. Citrate is also an effective inhibitor of glycolysis (44,65), and we hypothesize that the combined action of intracellular released GlcNAc and citrate causes the observed reduced glycolysis and SADR-1-induced ULBP2. Cellular metabolism is a constant balance between glycolytic and mitochondrial activity. To this end, hampered glycolysis will result in an increased reliance on mitochondrial activity, necessary for bacterial killing and ULBP2 expression. Recent studies showed that mitochondrial activity is linked to staphylococcal degradation and killing in human monocytes (67,68).
SADR-2 was selected in vivo and differs from SADR-1 by only two single SNPs in the rpoB and clpP genes that are linked to daptomycin susceptibility and staphylococcal stress control and virulence, respectively (6,30,33,34). However, it is not clear how these mutations influence immunity to S. aureus (30,78). Our data indicate that clpP but not rpoB hampers human monocytic immunity to SADR-2, suggesting a new immune evasive mechanism of S. aureus. It has been suggested that phagolysosomal resistant S. aureus accumulate silently intracellularly because of the absence of released inflammatory signals (29). We show that both SADR-1 and SADR-2 are recognized and internalized by human monocytes. However, SADR-1 induced NF-kB activity, IL-1b, IL-6, and IL-10, as well as a robust ULBP2 response, whereas significantly lower NF-kB, nearly no cytokine, and nonfunctional ULBP2 induction was detected in response to SADR-2. Surface recognition primes cells for production of pro-IL-1b, whereas intracellular inflammasome activity is necessary for secretion of mature IL-1b (27,79). Our results suggest that resistant SADR-2 does not fully activate either the surface or intracellular inflammatory signaling cascade necessary for NF-kB activity, cytokine secretion, or ULBP2 induction and thereby accumulates silently in human monocytes.
SADR-2-SADR-4 are likely all less sensitive to degradation because of vital alterations in structural features of their cell wall, possibly mediated by an abrogated clpP-clpX pathway and increased Spx activity (30,35,76), seemingly making these strains able to silently persist inside human monocytes. SADR-5 is highly interesting because this isolate also carries the rpoB A477D mutation, but in contrast to SADR-2-SADR-4, SADR-5 induces NF-kB activity and ULBP2 on the surface of THP1 cells, further supporting our findings that the clpP mutation and likely hampered clpP-clpX pathway in SADR-2 are essential for the increased resistance to lysozyme observed in this isolate. Involvement of the clpX and the clpP-clpX pathway, as well as intracellular degradability and accumulation, however, requires further experimental verification. Together, these findings highlight that NKG2D-mediated immunity is activated upon staphylococcal infections through an inflammasome-independent and imposed inflammatory shift in the intracellular metabolism of the phagocytosing monocytes, thus defining a target for future anti-staphylococcal therapy that can be used against persistent S. aureus infections.

Sequencing bacterial strains
Whole genome sequencing (WGS) was performed on SADR-1 and SADR-2 by isolating genomic DNA from overnight culture in brain-heart infusion (BHI) broth (Oxoid) with the DNeasy blood and tissue kit (Qiagen). Before DNA extraction, harvested cells were treated with lysostaphin (Ambi), added to the Gram-positive lysis buffer (final concentration 100 mg/ml), and incubated at 37°C for 30 min. WGS was performed on an Illumina NextSeq platform using Nextera XT libraries (150-bp paired ends) according to the manufacturer's instructions. SNP/indel calling were performed using the Snippy Pipeline version 3.0 and using SADR-1 and SADR-2 genomes assembled with skesa (81) as references. Potential structural variants were investigated by comparing the coverage of all reads, split reads, or discordant reads of SADR-1 and SADR-2 strains aligned onto the SADR-1 genome assembly. Sequencing reads were mapped with bwa mem v0.7.17-r1188 (82). Split and discordant mapped reads were extracted from the bam alignment files using the extractSplitReads_BwaMem script and the samtools command "samtools view -b -F 1294," respectively. The different bam alignment files of the different strains were then compared using Artemis software (83) to identify potential structural variants.
WGS revealed that SADR-2 harbors a G281A (G94D) mutation in clpP that was not reported in previous studies (6,30). To test for the presence of the clpP mutation in SADR-1-SADR-5, the clpP gene was PCR-amplified using primers: 59-GAT-AGGTGGCTATCAAGCG-39 and 59-CGAGTCAGCTAGT-GGTCCG-39 (purchased from Eurofins MWG Operon). The Qiagen purified PCR products were sequenced at Eurofins MWG Operon using primer 59-AATTCCTACAGTTATGA-39. Analysis of the sequences was performed using the free software CLC sequence viewer, version 8.0.

Construction of rpoB and clpP mutants by allelic exchange
Allelic exchange experiments were performed using shuttle vector pIMAY-Z (84) and using the method described in Ref. 85 with minor modifications. Full-length rpoB and clpP sequences corresponding to the SADR-1 and SADR-2 alleles were obtained by performing PCR with Phusion High-Fidelity DNA polymerase (New England Biolabs) and primer pairs pIM_rpob-F/R and pIM_clpP_F/R. Gel-purified amplicons were then ligated to pIMAY-Z plasmid using seamless ligation cloning extract (SLiCE) cloning (86) and transformed into Escherichia coli strain IM08B (84). The presence of a cloned insert was confirmed by colony PCR using primers pIM_MCS_F and pIM_MCS_R. Purified plasmids were then electroporated into S. aureus strains SADR-1 or SADR-2, and the transformations were plated onto BHI agar plates supplemented with chloramphenicol (Cm) at 10 mg/liter and X-gal (Melford) at 100 mg/l. The plates were incubated 48 h at 30°C; the blue colonies were then picked and grown in BHI broth at 37°C without Cm selection pressure overnight to allow the loss of the pIMAY-Z thermosensitive plasmid. White colonies with double cross-overs leading to allelic replacement were isolated directly onto X-gal BHI plates (introduction of clpP-D94G mutation into SADR-2) or BHI plated supplemented with 1 mg/liter of rifampicin (introduction of rpoB-A477D into SADR-1) or supplemented with 50 mg/liter of oxacillin (introduction of clpP-G94D into SADR-1). Clones were colony-purified onto BHI plates before glycerol storage and extraction of genomic DNA. To validate the allelic exchange procedure, the whole genome sequences of all reconstructed strains were determined with the Illumina Nextseq 500 sequencer and using Nextera XT paired-end libraries (2 3 150 bp reads). Sequencing reads of all mutant strains were mapped to SADR-1 and SADR-2 genomes using Snippy (version 4.6) to confirm the introduced mutations (https://github.com/tseemann/ snippy). The primers used were as follows: pIM_rpob_F, CTCACTAAAGGGAACAAAAGCTGGGTACCTTGGCA-GGTCAAGTTGTCCAATATG; pIM_rpob_R, CGACTCA-CTATAGGGCGAATTGGAGCTCTTAATCAGTAACTTC-TTTTTGTGTTTCAGGAG; pIM_clpP_F, CCTCACTAAA-GGGAACAAAAGCTGGGTACCATATTTCCGCTCAAAG-TAT; pIM_clpP_R, CGACTCACTATAGGGCGAATTGGA-GCTCTATCAGTTTACAAAGGAAAG; pIM-MCS_F, AAT-ACCTGTGACGGAAGATCACTTCG; and pIM_MCS_R, TACATGTCAAGAATAAACTGCCAAAGC.
Preparation of live (live-SA) and UV-killed S. aureus (UV-SA) S. aureus was inoculated from overnight TSA plates in fresh TSB to A 600 ; 0,03 and grown in Erlenmeyer flasks to the early stationary phase (A 600 = 5-6 or 5-7 h of incubation). The bacteria were washed and resuspended in PBS to A 600 = 1. Live-SA were added directly to cell cultures. For preparing UV-SA, 5 ml of live-SA was transferred to Petri dishes and subjected to pulsed UV radiation of 10.000 mJ/cm2 for 120 s (monochromatic wavelength of 254 nm; CL-1000 cross-linker; UVP, Cambridge, UK). Bacterial death was verified by plating on TSA plates and incubation at 37°C overnight.

NKG2D signaling assay
Effector cells (THP1) were treated with UV-SA for 48 h before the cells were blocked with NKG2D-Fc (R&D Systems, 1299-NK) or control IgG1-Fc (R&D Systems, 110-HG) for 30 min at 4°C. Target cells (2B4_NKG2D neg and 2B4_NKG2D pos ) were stained with the Vybrant cell-labeling kit, color DID (1 ml/ ml to 1 3 10 6 cells/ml) (Molecular Probes, V22889). Target and effector cells were mixed at the indicated ratios and co-cultured at 37°C with 5% CO 2 for 14-16 h before analysis by flow cytometry.

Seahorse glycolytic function analysis
Glycolytic capacity of THP1 cells was assessed by the Seahorse XFe96 extracellular flux analyzer (Seahorse Biosciences-Agilent Technologies) using the standard glycolysis stress test. THP1 cells were treated with UV-SA or PBS (untreated) and incubated at 37°C with 5% CO 2 for 24 h. The cells were washed in unbuffered Dulbecco's modified Eagle's medium (Sigma, D5030) supplemented with 2 mM glutamine (Sigma, G3126), pH 7.4. The cells were equilibrated for 1 h at 37°C in a CO 2free incubator. The extracellular acidification rate (ECAR) measurement cycle consisted of 3 min of mixing and 3 min of measurement of acidification. Analysis was initiated by three baseline ECAR measurement cycles, followed by the sequential injection of glucose (10 mM, Sigma, 49159), oligomycin (1 mM, Sigma, 75351), and 2DG (50 mM, Sigma, D6134) with two ECAR measurement cycles in between each injection and three final measurement cycles. ECAR was recorded and calculated by the Seahorse XFe96 software, Wave. Subsequently the protein content was measured for each well using the Pierce BCA assay with BSA as standard. Based on the ECAR measurements, the glycolytic parameters were calculated as follows: (a) glycolysis (last measurement before glucose injection subtracted from last measurement before oligomycin injection), (b) glycolytic capacity (last measurement before glucose injection subtracted from last measurement before 2DG injection), and (c) glycolytic reserve (glycolysis subtracted from glycolytic capacity). The results are presented as mean values 6 S.E. (87).
Metabolism of 13 C-labeled glucose THP1 cells were incubated in medium containing uniformly labeled glucose ([U-13 C]glucose (99%) 13 C-enriched; CLM-1396-10, Cambridge Isotopes Laboratories Inc.). The cells were plated in unlabeled glucose-free RPMI 1640 medium (Gibco, 11879-020) with 10% FBS (Sigma, F9665), 2 mM L-glutamine (Sigma, G7513), 2 mM penicillin and streptomycin (Sigma, P4333), and enriched with 10 mM ([U-13 C]glucose), treated with UV-SA or PBS (untreated), and incubated at 37°C with 5% CO 2 for 12 or 24 h. The cells were washed, extracted in 70% ethanol, and centrifuged at 20,000 3 g for 20 min (4°C) to separate the soluble extract from the insoluble components. Cell extracts were lyophilized and reconstituted in water for biochemical analyses. Protein content was determined in the dissolved pellets as previously described (87,88). Extract samples were adjusted to pH 1-2 with HCl and evaporated to dryness under nitrogen flow. Analytes were extracted into an organic phase (96% ethanol/benzene) followed by derivatization with 14% N,N-dimethylformamide (DMF), 86% N-tertbutyldimethylsilyl-N-methyltrifluoroacetamide (MTBSTFA) with a modified procedure from (89). Standards containing unlabeled metabolites of interest and cell extracts were separated and analyzed in a gas chromatograph (Agilent Technologies 7820A chromatograph, J&W GC column HP-5MS, parts no. 19091S-433) coupled to a mass spectrometer (Agilent Technologies, 5977E). The isotopic enrichment of the metabolites of interest was corrected for natural abundance of 13 C using the unlabeled standards and calculated according to Ref. 90. The data are presented as labeling (%) of M 1 X, where M is the mass of the unlabeled molecule, and X is the number of labeled C-atoms in a given metabolite (87).

Cytokine measurements
Cytokine levels in supernatants from THP1 cells treated with UV-SA were assessed at 1-48 h after treatment. Standard ELISA was used for measuring human IL-1b (R&D Systems, DY201), human IL-6 (R&D Systems, DY206) and human IL-10 (R&D Systems, DY217B). The samples were run on a BioTek PowerWave instrument and analyzed using KC4v3.0 software.

Metabolomics
Intracellular levels of metabolites from THP1 cells were analyzed by GC-MS. 3.6 3 10 6 THP1 cells were treated with UV-SA for 6 or 12 h. The cells were washed, flash-frozen in liquid nitrogen, and stored at 280°C. Extraction of metabolites was performed as described in Ref. 91. In brief, metabolites were extracted by two snap freezes in methanol and one snap freeze in MilliQ water. The methanol and water extracts were pooled and dried under nitrogen flow before derivatized with methyl chloroformate to obtain volatile products. The samples were analyzed by GC-MS at MS-Omics Aps. Derivatization and analysis were performed using a slightly modified version of the protocol described in Ref. 92. A mixed pooled sample was made from small aliquots of each individual sample to ensure quality control. Raw GC-MS data were processed using the data software PARADISe described in Ref. 93.

Confocal microscopy and phagocytosis analysis
UV-SA strains were labeled with AF647-conjugated succinimidyl-ester (SE-AF647; Molecular Probes, A-20006). Bacteria were centrifuged at 10,000 3 g 10 min 4°C and resuspended in original volume of sodium bicarbonate buffer (pH 8.5). SE-AF647 was added in a final concentration of 9 ng/ml and bacteria were incubated at 4°C for 1 h with agitation. Stained bacteria were washed and resuspended to original volume in PBS. THP1 cells were treated with stained UV-SA for 24-48 h. Human cells were washed twice in PBS with 2% FBS. For flow cytometry, the cells were analyzed directly, whereas for microscopy, they were stained with 1 mg/ml Hoechst 33342 (Thermo Scientific, 62249) at 37°C for 15 min and stained with AF488conjugated CD45 (Biolegend, 304019) at 4°C for 30 min. Finally, the samples were washed and resuspended in 60 ml of PBS. Confocal images were taken using a Carl Zeiss LSM780 confocal system with a Plan-Apochromat 633/1.4 oil-immersion objective. The 405-, 505-, and 633-nm lasers were used to excite Hoechst 33342, AF488-CD45, and SE-AF647, respectively. Fluorescent signals were collected using a fMBS 405/ 505c beam splitter and acquired in the same track to minimize cell movement during acquisition. Images were processed using the software Zen Blue version 2.3. Fluorescent signal from AF488-CD45 was subtracted from the Hoechst signal during image processing to allow for a separate nuclear signal.

GlcNAc treatment
Treatment of THP1 cells with GlcNAc was done by cultivating cells in standard RPMI further supplemented with GlcNAc (25 mM or 50 mM) for up to 3 weeks prior to UV-SA stimulation and analysis or with inspiration from Wolf et al. (44). GlcNAc (Sigma, catalog no. A3286) was dissolved in fresh Optimem (Gibco, 31985-062) to 1 M, pH ;7.4. 100 ml of GlcNAc solution was then mixed with 3 ml of FuGENE (Promega, Wisconsin, United States, E2313) and incubated for 1-1.5 h at room temperature. The FuGENE-complexed GlcNAc was added to cells (50 ml/ml) and incubated for 24 h at 37°C with 5% CO 2 .

Statistical analysis
Data preparation and statistical analysis were performed using the software GraphPad Prism, version 8.01 (GraphPad Software). Statistical analysis was performed and presented as stated in the figure legends, and the level of statistical significance was determined by a p value of ,0.05 and presented as follows: *, p , 0.05; **, p , 0.01; ***, p , 0.001; and ****, p , 0.0001.

Ethics statement
Primary cells were isolated from buffy coats from healthy human volunteer donors, obtained from the State Hospital (Copenhagen, Denmark). Informed written consents were obtained from all donors in accordance with legislation and guidelines from the local ethics committee (Region Hovedstaden, Denmark).

Data availability
Supporting information accompanies this article online. Data will be shared upon request to the corresponding author, Søren Skov.
Acknowledgments-We thank Dr. Peter Steinberger (Institute of Immunology, Medical University of Vienna, Vienna, Austria) for THP1-NF-kB-GFP reporter cells and Dr. Chiwen Chang (Department of Pathology, University of Cambridge, Cambridge, UK) for providing 2B4 and 2B4_NKG2D cells. We acknowledge the Core Facility for Integrated Microscopy (Faculty of Health and Medical Sciences, University of Copenhagen, Denmark) for guidance in confocal microscopy imaging. We acknowledge MS-Omics Aps