Hypusination Orchestrates the Antimicrobial Response of Macrophages

SUMMARY Innate responses of myeloid cells defend against pathogenic bacteria via inducible effectors. Deoxyhypusine synthase (DHPS) catalyzes the transfer of the N-moiety of spermidine to the lysine-50 residue of eukaryotic translation initiation factor 5A (EIF5A) to form the amino acid hypusine. Hypusinated EIF5A (EIF5AHyp) transports specific mRNAs to ribosomes for translation. We show that DHPS is induced in macrophages by two gastrointestinal pathogens, Helicobacter pylori and Citrobacter rodentium, resulting in enhanced hypusination of EIF5A. EIF5AHyp was also increased in gastric macrophages from patients with H. pylori gastritis. Furthermore, we identify the bacteria-induced immune effectors regulated by hypusination. This set of proteins includes essential constituents of antimicrobial response and autophagy. Mice with myeloid cell-specific deletion of Dhps exhibit reduced EIF5AHyp in macrophages and increased bacterial burden and inflammation. Thus, regulation of translation through hypusination is a critical hallmark of the defense of eukaryotic hosts against pathogenic bacteria.


INTRODUCTION
The mammalian gastrointestinal tract is the home of the largest population of myeloid cells (Lee et al., 1985). These resident and infiltrating cells respond to inflammatory signals and to foreign antigens, thus exhibiting enhanced antimicrobial abilities that are needed for the clearance of pathogens (Hardbower et al., 2017;Niess et al., 2005;Wennerås et al., 2000). In this context, deciphering the molecular mechanisms that regulate the innate response of these immune cells is essential for a better understanding of the protective immunity against gastrointestinal pathogens.
The arginase-ornithine decarboxylase (ODC) metabolic pathway plays an essential role in the orchestration of the activation of macrophages (Latour et al., 2020). Arginase activity converts arginine into urea and ornithine, with the latter being metabolized by ODC into putrescine (Pegg, 2009). This first and shortest polyamine is then sequentially converted by spermidine synthase and spermine synthase into spermidine and spermine, respectively (Latour et al., 2020;Pegg, 2016). We have implicated putrescine in the pathophysiology of infection with Helicobacter pylori (Hardbower et al., 2017), a specific pathogen of the stomach responsible for diseases ranging from gastritis to gastric cancer (Peek et al., 2010). Using mice with myeloid-specific Odc knockout, we found that endogenous putrescine alters histone methylation/acetylation and dampens the transcription of the genes encoding proinflammatory markers (Hardbower et al., 2017). This results in the limitation of inflammation and the exacerbation of H. pylori burden (Hardbower et al., 2017). In addition, we have established that spermine reduces arginine uptake, which is needed for nitric oxide (NO) synthase 2 (NOS2) translation and NO-dependent killing of H. pylori, thus impairing immunity to the bacterium (Bussière et al., 2005;Chaturvedi et al., 2010). However, the importance of spermidine for the outcome of the infection remains unknown.
Spermidine has been shown to improve cardiac function, to be neuroprotective, and to dampen inflammation (Madeo et al., 2018), notably by stimulating autophagy and mitophagy (Eisenberg et al., 2009(Eisenberg et al., , 2016Qi et al., 2016). Importantly, spermidine is required for the synthesis of the non-proteinogenic amino acid hypusine. The only known protein that contains hypusine is the eukaryotic translation initiation factor 5A (EIF5A; Park et al., 1981;Park and Wolff, 2018). Two isoforms of EIF5A exist: EIF5A1, referred to as eIF5A in this study, which is expressed universally; and eIF5A2, which is mainly present in testes, brain, and tumor cells (Jenkins et al., 2001;Nakanishi and Cleveland, 2016). Hypusine is formed by the action of two enzymes: deoxyhypusine synthase (DHPS), which transfers the Nterminal moiety of spermidine to the lysine-50 residue in EIF5A; and deoxyhypusine hydroxylase (DOHH) that completes hypusine biosynthesis by hydroxylation of the side chain ( Figure 1A; Joe et al., 1995;Park et al., 2006). This post-translational modification of EIF5A, called hypusination, has been reported to be essential for two molecular events. On the one hand, EIF5A Hyp binds specific mRNAs that contain a 5′-AAAUGU-3′ consensus sequence (Maier et al., 2010;Xu and Chen, 2001;Xu et al., 2004), and the EIF5A Hyp / mRNA complex is then translocated to the cytoplasm by the nuclear transporter exportin-1 (Aksu et al., 2016), reaches ribosomes, and allows for translation (Park and Wolff, 2018). On the other hand, hypusinated EIF5A alleviates ribosome pausing during the translation of peptides enriched in diprolyl, diglycyl motifs and in basic and acidic amino acids (Gutierrez et al., 2013;Pelechano and Alepuz, 2017;Schuller et al., 2017). Hypusination is crucial for embryonic development, cell differentiation, and proliferation (Ganapathi et al., 2019;Nishimura et al., 2005;Park et al., 1997;Sievert et al., 2014). When small interfering RNA (siRNA) against EIF5A or mice with specific deletion of Dhps in pancreatic islet β cells were used, it was shown that hypusinated EIF5A leads to the development of glucose intolerance in inflammatory mouse models of diabetes, notably by supporting NOS2 translation (Levasseur et al., 2019;Maier et al., 2010). In addition, hypusination in macrophages supports the tricarboxylic acid (TCA) cycle and oxidative phosphorylation by promoting the translation of mitochondrial proteins and promotes alternative (M2) activation (Puleston et al., 2019). However, the role of hypusination in the innate response of macrophages during bacterial infection has not been determined. We hypothesized that protein translation by the DHPS/EIF5A Hyp pathway is critical for the appropriate response of macrophages to invaders.
In this report, we show that the enzyme DHPS is induced by H. pylori in macrophages. This results in increased hypusination of EIF5A in vitro and in mouse and human tissues with H. pylori gastritis. Using mice with specific deletion of Dhps in myeloid cells, and macrophages treated with the DHPS inhibitor N 1 -guanyl-1,7-diaminoheptane (GC7), we identified a series of inducible proteins regulated by hypusination in infected macrophages. These effectors are notably involved in the antimicrobial response and autophagy. Deletion of DHPS in myeloid cells resulted in higher survival of phagocytized bacteria and increased colonization in vivo by the gastric pathogen H. pylori and by the colonic pathogen Citrobacter rodentium, the rodent equivalent of the human pathogen enteropathogenic Escherichia coli. Thus, in addition to the classical transcriptional response of macrophages, we emphasize herein that the regulation of translation through hypusination is a critical hallmark of the defense of the mammalian host against pathogenic bacteria.

Pathogenic Bacteria Stimulate DHPS Induction
We used the Lyz2-Cre driver to generate mice with specific deletion of Dhps in the myeloid cell lineage, including macrophages, dendritic cells, and granulocytes. Based on the conceptual framework of the role of hypusination in innate immunity, we first assessed the macrophage response to H. pylori infection. In response to this pathogen, bone-marrowderived macrophages (BMmacs) from Dhps fl/fl mice exhibited a significant increase in Dhps mRNA expression, whereas the Dohh level was not affected by the infection ( Figure 1B). The expression of Dhps in control or infected macrophages from Dhps fl/fl ;Lyz2 cre/cre mice (hereafter referred to as Dhps Δmye mice) was significantly reduced compared with cells from Dhps fl/fl or Dhps fl/fl ;Lyz2 cre/+ mice ( Figure 1B). Similarly, the H. pylori-induced DHPS protein expression in BMmacs from Dhps fl/fl and Dhps fl/fl ;Lyz2 cre/+ mice was markedly attenuated in Dhps Δmye cells ( Figure 1C). By densitometry, we found that the level of DHPS was reduced in H. pylori-stimulated macrophages from Dhps Δmye mice to the same level as unstimulated controls ( Figure 1D). In addition, the level of Dhps mRNA, but not Dohh mRNA ( Figure 1E), as well as DHPS protein ( Figures 1F and 1G), was increased in the murine macrophage cell line RAW 264.7 infected with H. pylori. This induction was reduced by blocking phosphatidylinositol 3-kinase (PI3K) with LY294002 (Figures S1A and S1B); however, inhibitors of nuclear factor κB (NF-κB), mitogen-activated protein kinase 14 (MAPK14), MAPK1/3, or MAP2K1 had no effect on H. pylori-induced DHPS protein expression (Figures S1A and S1B), indicating that signaling through PI3K is involved in stimulation of DHPS in infected macrophages. In addition, expression of Dhps mRNA ( Figure S1C) and DHPS protein ( Figure S1D) was also induced by the colonic pathogen C. rodentium in BMmacs from Dhps fl/fl and Dhps fl/fl ;Lyz2 cre/+ mice and markedly reduced in Dhps Δmye BMmacs. Lastly, we observed that DHPS was also induced by H. pylori and C. rodentium in bone-marrow-derived dendritic cells (BMDCs; Figure S1E).
Using a specific antibody (Nishiki et al., 2013), we found that the level of EIF5A Hyp was enhanced in BMmacs from Dhps fl/fl mice infected with H. pylori compared with uninfected cells and was absent in uninfected and infected Dhps Δmye macrophages ( Figures 1H and 1I). Similarly, H. pylori-induced hypusination in RAW 264.7 cells was significantly reduced by the DHPS inhibitor GC7 (Figures 1J and 1K). Moreover, we found hypusination of EIF5A was significantly induced by H. pylori in BMmacs from Odc fl/fl mice and from animals with specific Odc deletion in myeloid cells (Odc Δmye ; Hardbower et al., 2017;Figures S1F and S1G). While EIF5A Hyp was reduced in BMmacs from Odc Δmye mice compared with Odc fl/fl mice, more than 50% of the EIF5A Hyp remained in infected cells (Figures S1F and S1G), explaining why increased immune effectors could still be identified in BMmacs from Odc Δmye mice lacking the transcriptional inhibitor putrescine (Hardbower et al., 2017).
These data demonstrate that bacterial infection of macrophages leads to enhanced hypusination, notably through the inducible expression of DHPS.

Increased Hypusination in H. pylori-Infected Patients
To determine whether DHPS is expressed in gastric macrophages in humans, we assessed gastric biopsies from uninfected individuals and H. pylori-infected patients with gastritis.
The number of cells with double staining for EIF5A Hyp and CD68 was significantly increased in tissues from infected patients compared with individuals without infection (Figures 2A and 2B).

Hypusination Regulates Macrophage Function
To determine the role of hypusination in the innate response of macrophages to bacterial infection, we analyzed the proteome of BMmacs from three different Dhps fl/fl and Dhps Δmye mice. These cells were stimulated or not with H. pylori, and an isobaric tag for relative and absolute quantification (iTRAQ)-based proteomics approach was utilized. We identified 2,994 proteins, comprising 2,895 murine proteins and 99 H. pylori proteins. There were 130 and 191 proteins induced by H. pylori infection in BMmacs from Dhps fl/fl and Dhps Δmye mice, respectively (Table S1). The most induced proteins in Dhps fl/fl BMmacs included (1) classical antibacterial effectors IRG1, NOS2, and ASSY; (2) the proteins IFIH1, ISG20, GBP2/4, and IFIT2/3 that are involved in the antiviral response; (3) the proinflammatory cytokines IL1A, IL1B, and CCL5; (4) the inflammasome constituent NLRP3; and (5) the mediators of autophagy IIGP1, CLEC4E, and SQSTM ( Figure 3A; Table S1). Remarkably, there were also proteins less expressed in infected BMmacs from Dhps Δmye mice compared with Dhps fl/fl animals; these included IRG1, NOS2, ASSY, IL1A, CCL5, IIGP1, CLEC4E, IRGM1, and SQSTM ( Figure 3A; Table S1). We then used targeted approaches to verify our findings. We found by western blot that the level of the proteins NOS2, IRG1, and SQSTM were reduced in BMmacs from Dhps Δmye mice infected with H. pylori compared with infected Dhps fl/fl -derived BMmacs ( Figure 3B), thus validating the results obtained by proteomic analysis. We also confirmed that de novo translation of NOS2 was reduced in BMmacs from Dhps Δmye mice compared with Dhps fl/fl macrophages ( Figure S2). The production of NO 2 − , the stable metabolite of NO derived from NOS2 activity, was also reduced in infected Dhps Δmye BMmacs compared with cells from Dhps fl/fl mice ( Figure   S3A). Importantly, the mRNA expression of the genes encoding these proteins was not affected by DHPS deletion ( Figure 3C), further demonstrating that DHPS regulates only protein translation in infected macrophages. We also found that the expression of genes encoding for the chemokine CXCL1, the classical M1 marker IL-6, and the M2 marker arginase-1 was increased in response to H. pylori in both genotypes, but was not differentially expressed between them ( Figure S3D), indicating that Dhps deletion in macrophages does not affect the inducible transcriptome in response to bacteria.
To confirm this first proteomic analysis, we performed a second iTRAQ experiment using a new set of BMmacs from one Dhps fl/fl mouse and one Dhps Δmye mouse. In this experiment, there were 248 proteins significantly upregulated by H. pylori in Dhps fl/fl BMmacs (Table   S2), and 78.5% of the induced proteins identified in the first experiment were also found in this independent study. Strikingly, the main H. pylori-stimulated immune mediators affected by Dhps deletion in the first experiment were also downregulated in BMmacs from Dhps Δmye mice in this second experiment ( Figure S4A). Additionally, we confirmed by immunoblotting that the immune effectors NOS2, IRG1, and SQSTM were induced in Dhps fl/fl BMmacs infected with C. rodentium compared with control cells and that their expression was reduced in infected Dhps Δmye BMmacs ( Figure S4B), evidencing that the regulation of specific proteins by hypusination occurs in response to different pathogens. In this analysis, there were 27 and 13 proteins up-and downregulated, respectively, in uninfected Dhps Δmye BMmacs compared with Dhps fl/fl cells (Table S2), but there was no major regulator of the antimicrobial activity of macrophages in the uninfected cells.
The differential proteome dataset from H. pylori-infected BMmacs from Dhps fl/fl and Dhps Δmye mice was subjected to ingenuity pathway analysis (IPA) to determine biological functions and the important pathways regulated by hypusination in macrophages. The complete analysis is presented in Table S3. Strikingly, when we focused on the functions related to immune pathways, we found that the categories linked to macrophage activation and antimicrobial responses were significantly affected by Dhps deletion ( Figure 3D); this was associated with an increase in "viral infection" and "growth of bacteria" pathways ( Figure 3D). Overall, there was also a dampening of the pathways related to inflammation, adhesion, and death of immune cells ( Figure 3D). In addition, we used IPA to identify the biochemical networks that were modulated in infected macrophages from Dhps Δmye versus Dhps fl/fl mice. The highest scored networks were "hematological system development and function, inflammatory response, tissue morphology" (score 29; Figure S5A) and "antimicrobial response, connective tissue disorders, inflammatory response" (score 26; Figure S5B), further indicating that the specific alterations in protein expression regulated by DHPS lead to enhanced host defenses.
Next, we assessed the effect of hypusination on the macrophage proteome using a pharmacological approach. RAW 264.7 cells were treated or not with the DHPS inhibitor GC7 for 2 h prior to infection with H. pylori PMSS1. After 24 h, the proteome was similarly analyzed using the iTRAQ-based experimental approach. We quantified 3,556 proteins, comprising 3,507 murine proteins and 49 H. pylori proteins. There were 57 and 58 proteins significantly upregulated in response to H. pylori in cells without and with GC7 (Table S4). We found a series of proteins more highly expressed in H. pylori-infected macrophages compared with infected cells treated with GC7, including many of the same proteins affected by Dhps deletion in BMmacs, e.g., IRG1, NOS2, EHD1, and IRGM1 ( Figure 4A; Table S4).
Of note, 58 proteins were induced by the GC7 in uninfected cells, whereas 158 were downregulated by DHPS inhibition; the proteins that were the most inhibited were ribosomal proteins (Table S4). We confirmed by immunoblotting that the proteins NOS2, IRG1, and IRGM1 were less induced in response to H. pylori in RAW 264.7 cells treated with GC7 compared with infected macrophages without GC7 ( Figure 4B), while their mRNA levels were not affected ( Figure 4C). In addition, the NO metabolite NO 2 − was also produced in lower amounts by infected macrophages treated with GC7 ( Figure S3B); GC7 also reduced H. pylori-induced autophagy in RAW 264.7 cells ( Figure S3C). Lastly, the immune pathways involved in the activation and antimicrobial effects of macrophages were downregulated by GC7 in H. pylori-infected RAW 264.7 macrophages ( Figure 4D; Table   S5). Interestingly, the pathways predicted to impact cell death, including by autophagy, were also negatively affected by the inhibition of hypusination ( Figure 4D; Table S5).
These data suggest that hypusination supports the antimicrobial protein expression profile of macrophages.

mRNA and Protein Sequence Analysis of the Effectors Regulated by Hypusination in Macrophages
Hypusinated EIF5A binds preferentially to mRNAs that contains an 5′-AAAUGU-3′ consensus sequence (Maier et al., 2010;Xu and Chen, 2001). For this in silico analysis, we focused on the 48 proteins downregulated by at least 1.3-fold in infected Dhps Δmye BMmacs and RAW 264.7 cells infected with H. pylori and treated with GC7. Among them, we found this binding site in the coding region of 23 of the proteins regulated by hypusination, and 20 more displayed the similar sequence with only one substitution (Table S6).
In addition, EIF5A Hyp improves ribosome pausing during the translation of peptides enriched in diprolyl, diglycyl motifs and in basic and acidic amino acids (Gutierrez et al., 2013;Pelechano and Alepuz, 2017;Schuller et al., 2017). We identified 31 proteins with at least one diprolyl motif and 36 with one diglycyl motif (Table S6); 22 proteins exhibited at least one of each motif. Furthermore, we also identified 34 proteins containing at least one tripeptide whose translation is recognized to be facilitated by hypusinated EIF5A (Pelechano and Alepuz, 2017;Schuller et al., 2017). Only the proteins IFIT3, CCL5, and CLEC4E did not exhibit diprolyl, diglycyl, or EIF5A-regulated tripeptide motifs, but each of these did show the 5′-AAAUGU-3′ sequence (± 1 substitution).
Thus, all the innate effectors identified in this study to be affected by hypusination in macrophages exhibit at least one feature of EIF5A Hyp -regulated proteins, validating our proteomic findings.

Regulation of the Mucosal Antimicrobial Response by DHPS
Our data indicate that the induction of DHPS in macrophages supports the translation of inducible proteins involved in the defense against pathogens. To verify this postulate, we tested the involvement of hypusination in the antimicrobial effect of macrophages. We observed increased H. pylori survival in BMmacs from Dhps Δmye mice compared with Dhps fl/fl macrophages ( Figure 5A) and in RAW 264.7 cells pre-treated with GC7 compared with untreated macrophages ( Figure 5B).
Next, to assess the effect of hypusination in infection in vivo, we used a well-established model of H. pylori infection in mice (Chaturvedi et al., 2010;Hardbower et al., 2016Hardbower et al., , 2017. The level of gastric colonization by H. pylori at 4 weeks and 8 weeks post-infection was significantly enhanced by 1.1 and 0.8 log, respectively, in animals lacking Dhps in myeloid cells compared with Dhps fl/fl mice ( Figure 5C). Concomitantly, hematoxylin and eosin (H&E) images showed more infiltration of immune cells in the gastric mucosa and more gastric hyperplasia in H. pylori-infected Dhps Δmye mice ( Figure 5D). Scoring for acute and chronic inflammation in the gastric antrum and corpus (Chaturvedi et al., 2010;Hardbower et al., 2016Hardbower et al., , 2017 confirmed that gastritis was significantly higher in infected Dhps Δmye mice than Dhps fl/fl mice at both time points ( Figure 5E). This was confirmed by increased expression of the genes encoding for the chemokines CXCL1 and CXCL2 and for the T cell effectors IL-17 and IL-10 in the gastric mucosa of infected Dhps Δmye mice ( Figure S6A).
However, the inducible expression of the genes encoding for NOS2, IL-1β, and interferon (IFN)-γ was not affected by Dhps deletion ( Figure S6A).
We then confirmed the role of DHPS in the antibacterial immunity of the gastrointestinal tract using the colonic pathogen C. rodentium. Similar to the result with H. pylori in the stomach, mice with specific deletion of Dhps in myeloid cells exhibited increased colonization by C. rodentium ( Figure 6A). There was also a significant increase in colon weight ( Figure 6B) and histologic colitis in Dhps Δmye mice ( Figures 6C and 6D). Cxcl1, Il1b, Ifng, and Il17a mRNA levels were more abundant in the colon of C. rodentiuminfected Dhps Δmye mice compared with Dhps fl/fl mice ( Figure S6B).  Figures 7C and 7D). Note that EIF5A Hyp was also present in other cells from the gastric mucosa of H. pylori-infected mice and from the colon of animals infected with C. rodentium, but these levels were not altered by myeloid cell Dhps deletion ( Figure 7C). Of interest, the levels of NOS2, which have been found to be regulated by hypusination, were also increased in CD68 + macrophages in H. pyloriand C. rodentium-infected Dhps fl/fl mice and were markedly reduced in Dhps Δmye mice ( Figures 7A-7D).
Lastly, we assessed the effect of Dhps deletion in macrophages on the gastric metabolomic signatures. There were 267 positively charged and 341 negatively charged metabolites significantly affected in infected Dhps Δmye mice compared with Dhps fl/fl mice ( Figure S7).
As an example, the level of spermidine, the substrate for DHPS, was increased in the stomach of Dhps Δmye mice. Conversely, the NOS2 product citrulline, and the oxidized from of glutathione, which is a reliable indicator of oxidative stress, were less abundant in the Dhps Δmye animals ( Figure S7). The analysis of the metabolomic pathways (Table S7) evidenced that the citrulline-NO cycle was significantly affected by infection in Dhps fl/fl mice, but not in Dhps Δmye mice. We also found that the TCA cycle was affected by Dhps deletion in myeloid cells (Table S7).

DISCUSSION
Hypusination of EIF5A is a unique mechanism that controls the translation of specific proteins. In this study, we have utilized an unbiased approach using specific genetic deletion of Dhps and a DHPS inhibitor to probe the role of the spermidine-EIF5A Hyp axis in macrophage responses to pathogenic bacteria. Increased DHPS expression and levels of EIF5A Hyp were observed in macrophages infected ex vivo with H. pylori and in gastric CD68 + cells from H. pyloriand C. rodentium-infected mice. This results in the translation of immune effectors principally involved in microbial clearance and cell death, as underlined by the IPA of the proteomic data. Importantly, these results were obtained with the human gastric pathogen H. pylori and the murine colon bacterial pathogen C. rodentium, providing strong evidence that our finding is not limited to one pathogen but concerns the global innate response of macrophages. Lastly, our data showing increased colonization of Dhps Δmye mice by both pathogens endorse the in vivo relevance of the paramount role of hypusination in the first line of defense of macrophages against bacterial infection of the gastrointestinal mucosa. Our present findings emphasize one more step in the complex regulation of the immune response by polyamines: first, putrescine regulates the transcription of innate effectors (Hardbower et al., 2017); and second, spermidine controls translation through hypusination, as shown herein. Of importance, we found that hypusination is reduced in infected BMmacs from Odc Δmye mice, but not completely suppressed, consistent with the partial reduction of spermidine concentration by Odc deletion in macrophages (Hardbower et al., 2017). This indicates that ODC-derived putrescine affects transcription, but does not sufficiently alter translation by modulation of hypusination. Taken together, our data herein and our prior study (Hardbower et al., 2017) illustrate a complex immunologic rheostat for macrophage innate immune effectors, where putrescine from ODC regulates transcription, and spermidine, acting through DHPS/hypusination, regulates translation, providing robust biological control.
Previous reports have evidenced that EIF5A knockdown using small hairpin RNA in cervical cancer cells (Mémin et al., 2014), pancreatic ductal adenocarcinoma cells (Fujimura et al., 2015), and HeLa cells (Mandal et al., 2016) regulates proteins involved in migration, invasion, proliferation, endoplasmic reticulum (ER) stress, and/or unfolded protein response. Moreover, mitochondrial enzymes and proteins involved in the TCA cycle are affected by GC7 in M0 and M2 macrophages and by Eif5a and Dhps silencing in murine embryonic fibroblasts (Puleston et al., 2019). We have identified a series of bacteria-induced proteins that are regulated by DHPS function in macrophages.
Interestingly, it has also been shown that hypusination of EIF5A is enhanced in alternatively activated macrophages, in which the arginase-ODC metabolic pathway is strongly activated, but not during the classical macrophage activation with lipopolysaccharide (LPS)/IFN-γ (Puleston et al., 2019). However, in the context of infected macrophages, we propose now that DHPS induction supports increased hypusination of EIF5A and translation of antimicrobial factors. Note that in this prior study (Puleston et al., 2019), the authors did not identify the proteins involved in the antimicrobial response since they were not induced by M2 stimulation. In this previous work, the authors demonstrated that the TCA cycle was dampened in M0, M1, and M2 macrophages treated with GC7 (Puleston et al., 2019). Accordingly, our proteomic analysis of RAW 264.7 cells ± GC7 evidenced that the enzymes SDHB, SUCB1, SUCB2, ODPA, and ODPB, which all belong to the TCA cycle, are downregulated by GC7 in unstimulated cells (Table S4). However, since these enzymes were also reduced by H. pylori infection, we suggest that the control of infection by hypusination is not dependent on the TCA cycle.
In our study, numerous immune effectors regulated by hypusination are involved in the host defense against pathogens. Notably, the protein levels of NOS2, ASSY, and IRG1 were markedly reduced during H. pylori infection in Dhps Δmye macrophages compared with Dhps fl/fl and in GC7-treated macrophages. NOS2 expression is a hallmark of macrophage activation by pathogen-associated molecular patterns and proinflammatory cytokines (MacMicking et al., 1997) and has been already reported to be regulated by hypusination in pancreatic β cells (Maier et al., 2010) and in murine macrophages stimulated with the mycobacterial cell wall component trehalose-6,6′-dimycolate (Lee et al., 2016). NOS2derived high-output NO production has been implicated in the killing of various parasites and bacteria (Duleu et al., 2004;El Kasmi et al., 2008;Qualls et al., 2012), including H. pylori (Gobert et al., 2001). Similarly, ASSY converts citrulline into arginine, the NOS2 substrate, to increase the NO-dependent killing of bacteria (Qualls et al., 2012). In addition, the innate effector IRG1 has been shown to be induced in RAW 264.7 cells and primary macrophages stimulated by LPS and IFN-γ (Strelko et al., 2011), tumor necrosis factor alpha, Toll-like receptor (TLR) 2, or TLR4 signaling (Degrandi et al., 2009) and leads to the synthesis of itaconic acid (Basler et al., 2006;Degrandi et al., 2009;Strelko et al., 2011). This metabolite inhibits isocitrate lyase, a key enzyme that is necessary for bacterial growth (Michelucci et al., 2013). Interestingly, it has been recently shown that the CRISPR-Cas9 deletion of both NOS2 and IRG1 reduced the killing of Legionella pneumophila by murine BMmacs stimulated by IFN-γ (Price et al., 2019). In addition, the maximal killing of L. pneumophila by macrophages required the expression of four other immune effectors: CASP11, NOX2, IRGM3, and IRGM1 (Price et al., 2019); this last protein was also found to be regulated by hypusination in our experiments. Thus, our data highlight that hypusination is a major regulator of host defense by affecting the translation of multiple antimicrobial effectors in macrophages; this postulate was supported by our finding that bacterial survival is increased in macrophages with DHPS knockout or inhibition. Of note, we also observed that hypusination regulates mediators involved in antiviral response, including the receptor of viral nucleic acids IFIH1, the antiviral exoribonuclease ISG20, the GTPases GBP2 and GBP4, and the antiviral proteins RRSAD2, IFIT2, and IFIT3, further indicating that hypusination is a general control mechanism of the overall innate response of macrophages.
The abundance of the proteins IIGP1, IRGM1, CLEC4E, and SQSTM, which are all implicated in autophagy, was affected by DHPS deletion or inhibition in infected macrophages, and we also found that GC7 reduced H. pylori-mediated autophagy.
Remarkably, these effectors also play a fundamental role in the destruction of microorganisms (Al-Zeer et al., 2009;MacMicking et al., 2003;Pahari et al., 2019;Watson et al., 2012). Intriguingly, spermidine, the DHPS substrate, supports autophagy in various cells including human peripheral blood mononuclear cells (Eisenberg et al., 2009) or in a human colon cancer cell line (Morselli et al., 2011). In the same way, mice treated with spermidine exhibit prolonged lifespan and improved cardiomyocyte function due to increased cardiac autophagic flux (Eisenberg et al., 2016). Furthermore, it has been recently observed that hypusinated EIF5A controls the synthesis of the transcription factor EB, a master regulator of genes encoding mediators of autophagy, in naive B cells (Zhang and Simon, 2020). This transcription factor was not evidenced to be regulated by infection and/or inhibition of DHPS in our study, suggesting that EIF5A Hyp may regulate autophagy by various pathways. Nonetheless, we contend that hypusination is also involved in the autophagy-dependent clearance of pathogenic bacteria by macrophages.
Polyamine metabolism plays a fundamental role in the outcome of bacterial infections, including those of the gastrointestinal tract (Chaturvedi et al., 2010;Gobert et al., 2018;Hardbower et al., 2017). This report highlights that the spermidine-hypusine metabolic pathway represents an unrecognized foundation of the antimicrobial armature. It is important to underline that reduced levels of EIF5A Hyp in tissue macrophages led not only to increased colonization by H. pylori and C. rodentium but also to enhanced inflammation. We therefore propose that the loss of innate host defense in Dhps Δmye macrophages led to the uncontrolled development of infection, which results in increased inflammation. Epithelial responses and chronic inflammation are also critical for the outcome of gastrointestinal infections. Hypusination has been shown to support the development of cancerous cells (Nakanishi and Cleveland, 2016); in contrast, it has also been recently evidenced that the polyamine-EIF5A Hyp metabolic pathway exhibits tumor suppressor functions in a mouse lymphoma model (Scuoppo et al., 2012). Given that H. pylori is the main causative agent for the development of gastric cancer (Parsonnet et al., 1991), the relative functional contribution of hypusination in epithelial, myeloid, and T cells in host defense against carcinogenetic events deserves further investigation and is underway in our laboratory.

STAR★METHODS RESOURCE AVAILABILITY
Lead contact-Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Keith T. Wilson (keith.wilson@vumc.org).
Materials availability-Dhps fl/fl ;Lyz2 cre/+ and Dhps fl/fl ;Lyz2 cre/cre mice were generated for this study and are available upon request.
Data and code availability-The mass spectrometry proteomics data regarding BMmacs (Tables S1 and S2) and RAW 264.7 cells (Table S4) have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository (Perez-Riverol et al., 2019), and the accession numbers are ProteomeXchange Consortium: PXD018187 and ProteomeXchange Consortium: PXD010082, respectively. The accession number for the metabolomics data reported in this paper is EMBL-EBI MetaboLights: MTBLS607.

EXPERIMENTAL MODEL AND SUBJECT DETAILS
Human tissues-Biopsies from gastric tissues were obtained from human subjects in Colombia as described (Mera et al., 2018), under protocols approved by the ethics committees of the local hospitals and of the Universidad del Valle in Cali, Colombia, as well as the Institutional Review Board at Vanderbilt University. We used biopsies from 5 uninfected subjects (4 females and 1 male) aged 26-64 years (mean ± standard deviation: 47.2 ± 13.6 years) and 9 H. pylori-infected patients (5 females and 4 males) aged 26-45 years (mean ± standard deviation: 34.6 ± 6.1 years).  (1966). To achieve specific deletion of Dhps in myeloid cells, we crossed C57BL/6J Dhps fl/fl mice, in which exons 2 to 7 of the Dhps gene were flanked by Cre recombinase recognition sequences (Levasseur et al., 2019), to mice with the myeloid driver Lyz2 cre . Dhps fl/fl ;Lyz2 cre/+ mice were then crossed to each other to generate littermate Dhps fl/fl ;Lyz2 cre/cre , Dhps fl/fl ;Lyz2 cre/+ and Dhps fl/fl mice in expected proportions (25%, 52% and 23%, respectively). Thereafter, we used littermates, namely Dhps fl/fl mice as controls, Dhps fl/fl ;Lyz2 cre/+ mice, and Dhps fl/fl ;Lyz2 cre/cre mice, which we termed Dhps Δmye mice. We also used Odc fl/fl and Odc Δmye mice that are bred in our animal facility (Hardbower et al., 2017). Adult mice (8-12 weeks) of both sexes were used for isolation of BMmacs. Only male mice (8-12 weeks) were used for infection with H. pylori and C. rodentium.
Bacteria-The cagA + H. pylori strains PMSS1 (Arnold et al., 2011) and C. rodentium DBS100 (Barthold et al., 1976) were grown on Tryptic Soy agar plates containing 10% sheep blood and MacConkey agar plates, respectively. Bacteria were harvested directly from plates to infect the cells. From the plates, H. pylori and C. rodentium were also grown in Brucella broth containing 10% FBS and in Luria-Bertani broth liquid medium overnight, respectively; they were then diluted to an A 600 nm of 0.1 in these media, grown, and harvested at the exponential phase to infect mice.

METHOD DETAILS
Infections-C57BL/6J, Dhps fl/fl , Dhps fl/fl ;Lyz2 cre/+ , Dhps fl/fl ;Lyz2 cre/cre , Odc fl/fl , and Odc Δmye mice were bred in our animal facility. Mice were fed ad libitum with regular 5L0D chow (LabDiet). For infections, male mice at 8-12 weeks of age were randomly distributed in the different groups and infected by oral gavage with i) 10 9 H. pylori PMSS1 in 0.2 mL Brucella broth, two times every 2 days , or ii) 5×10 8 C. rodentium DBS100 in in 0.2 mL Luria-Bertani broth (Gobert et al., 2018;Hardbower et al., 2016). Control mice were gavaged with the corresponding broth alone. In both models, mice were monitored daily and those that showed extreme distress, became moribund, or lost more than 20% of initial body weight were euthanized. Animals were sacrificed after 4 or 8 weeks (H. pylori) and 14 days (C. rodentium). Stomachs and colons were harvested and analyzed as described (Chaturvedi et al., 2010;Gobert et al., 2018;Hardbower et al., 2016Hardbower et al., , 2017. The number of bacteria in gastric and colon tissues was determined by counting the CFUs after plating serial dilutions of homogenized tissues.
To determine the survival of bacteria in macrophages, cocultures were washed thoroughly with PBS after 24 h, incubated 1 h with 200 μg/ml gentamicin, and lysed in 0.1% saponin for 30 min at 37°C (Gobert et al., 2014). The number of bacteria in each lysate was determined by counting the CFUs after plating serial dilutions on blood agar plates.
Analysis of mRNA levels-Total RNA was isolated using the RNeasy Mini kit (QIAGEN). Reverse transcription was performed using Superscript II Reverse Transcriptase and Oligo dT (Invitrogen). mRNAs were amplified by real-time PCR using the iQ ™ SYBR Green kit (Bio-Rad) (Gobert et al., 2014) and the primers listed in Key Resources Table. Proteomics and IPA-Proteomic analysis was performed as described (Gobert et al., 2019;Noto et al., 2019). After infections, macrophages were lysed in 50 mM Tris-HCl pH 7.6, 150 mM NaCl, 1% NP-40, and 2 mM EDTA and protein concentrations were determined by BCA. Protein samples were precipitated with ice-cold acetone overnight at −20°C. Precipitates were washed with cold acetone, dried, and reconstituted in 8 M urea in 250 mM TEAB (pH 8.0). Samples were reduced with 50 mM TCEP, alkylated with 200 mM MMTS, and diluted with TEAB to obtain a final solution containing 2 M urea, and digested with sequencing-grade trypsin overnight. Quantitative proteomics analysis was performed using iTRAQ. For 25 μg of protein, 1 unit of 4-plex iTRAQ labeling reagent (SCIEX) was used. Labeling reagent was reconstituted in ethanol such that each protein sample was labeled at a final concentration of 90% ethanol, and labeling was performed for 2 h. Fourplex iTRAQ experiments were conducted. The resulting labeled peptides were then desalted by a modified Stage-tip method. Following desalting, peptides were reconstituted in 0.1% formic acid, and peptides were loaded onto a self-packed biphasic C18/SCX MudPIT column using a helium-pressurized cell. Using a Dionex Ultimate 3000 nanoLC and autosampler, MudPIT analysis was performed with a 15-step salt pulse gradient. Following each salt pulse delivered by the autosampler, peptides were gradient-eluted from the reverse analytical column at a flow rate of 350 nl/min. Mobile phase solvents consisted of 0.1% formic acid, 99.9% water (solvent A) and 0.1% formic acid, 99.9% acetonitrile (solvent B). For the peptides from the first 13 SCX fractions, the reverse phase gradient consisted of 2%-50% B in 83 min, followed by a 10 min equilibration at 2% B. For the last 2 SCX-eluted peptide fractions, the peptides were eluted from the reverse phase analytical column using a gradient of 2%-98%B in 83 min, followed by a 10 min equilibration at 2% B. A Q Exactive Plus or HF mass spectrometer (Thermo Scientific), equipped with a nanoelectrospray ionization source, was used to mass analyze the eluting peptides. The Q Exactive instrument was operated in data-dependent mode acquiring HCD MS/MS after each MS1 scan on the 15 most abundant ions using an MS2 target of 10 5 ions. The HCD-normalized collision energy was set to 30, dynamic exclusion was set to 30 s, and peptide match and isotope exclusion were enabled.
Peptide/protein identifications and quantitative analysis were performed using Spectrum Mill (Agilent). MS/MS spectra were searched against a subset of the UniProt KB protein database containing Mus musculus and Helicobacter pylori 26695 proteins. Autovalidation procedures in Spectrum Mill were used to filter the data to < 1% false discovery rates at the protein and peptide level. Log2 protein ratios were fit to a normal distribution using non-linear (least-squares) regression. The calculated mean derived from the Gaussian fit was used to normalize individual log2 ratios for each quantified protein. The normalized log2 ratios were then fit to a normal distribution, and the mean and standard deviation values derived from the Gaussian fit of the normalized ratios were used to calculate p values. Subsequently, p values were corrected for multiple comparisons by the Benjamini-Hochberg method.
IPA software (QIAGEN) was used for the functional interpretation of differential expression results obtained from the aforementioned proteomic analyses. The pathways related to Diseases and Functions and the protein interaction networks were generated. The statistical significance for each assignment was expressed by a corresponding p values calculated using Fisher's exact test.
Untargeted Metabolomics-Gastric tissues were homogenized by sonication in water:methanol (9:1) containing 50 mM ammonium acetate (pH~6) to yield a tissue density of 50 mg/ml. A portion of the homogenate (100 μL) was combined with HPLC-grade methanol (300 μL), vortexed vigorously, and centrifuged at 10,000 × g. An aliquot of the supernatant (100 μL) was diluted with an equal volume of HPLC-grade acetonitrile. Discovery metabolomics data were acquired using a Vanquish ultrahigh performance liquid chromatography (UHPLC) system interfaced to a Q Exactive HF quadrupole/orbitrap mass spectrometer (Thermo Fisher Scientific). Each sample was injected twice. The first injection was made in positive ESI mode; the second injection was made in negative mode. A Zic-cHILIC analytical column (3 μm, 2.1 × 150 mm; Merck SeQuant) was used for all chromatographic separations. Mobile phases were made up of 0.2% acetic acid and 15 mM ammonium acetate in (A) H 2 O:CH 3 CN (9:1) and in (B) CH 3 CN:CH 3 OH:H 2 O (90:5:5). The flow rate was maintained at 300 μl/min, the total chromatographic run time was 20 min, and the sample injection volume was 10 μl. Mass spectra were acquired over a precursor ion scan range of m/z 100 to 1,200 at a resolving power of 30,000 using the following ESI source parameters: spray voltage 5 kV (3 kV in negative mode); capillary temperature 300°C; S-lens RF level 60 V; N2 sheath gas 40; N2 auxiliary gas 10; auxiliary gas temperature 100°C. MS/MS spectra were acquired for the top-five most abundant precursor ions with an MS/MS AGC target of 10 5 , a maximum MS/MS injection time of 100 ms, and a normalized collision energy of 30. Chromatographic alignment, peak picking, and statistical comparisons were performed using XCMS (https://xcmsonline.scripps.edu).

QUANTIFICATION AND STATISTICAL ANALYSIS
Statistical analyses were performed with GraphPad Prism 8.4.3 and the significance level was set as p < 0.05. All the data shown represent the mean ± SEM. Data that were not normally distributed according to the D'Agostino & Pearson normality test were log or square root transformed. Student's t test or ANOVA with the Tukey test were used to determine significant differences between two groups or to analyze significant differences among multiple test groups, respectively. All statistical tests were two-sided.

Supplementary Material
Refer to Web version on PubMed Central for supplementary material.

Pathogenic bacteria induce DHPS and hypusination of EIF5A in macrophages
• Hypusinated EIF5A controls the translation of numerous antimicrobial effectors • Macrophage DHPS is required to clear gastrointestinal pathogens   The heatmap depicts the normalized fold change compared with uninfected cells of the 92 proteins induced significantly by at least 1.3-fold in infected RAW 264.7 macrophages, treated or not with GC7. For each protein, we calculated the ratio between the fold increase in infected macrophages and in infected macrophages + GC7. (B) Levels of NOS2, IRG1, IRGM1, and β-actin proteins in RAW 264.7 cells ± GC7 infected or not with Hp for 24 h; these immunoblots are representative data of three to six independent experiments. (C) Expression of the genes encoding NOS2, IRG1, and IRGM1 was determined in RAW 264.7 macrophages ± GC7 after a 6 h infection. Bars represent mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001; n = 3-6 independent experiments. (D) Diseases and functions pathways were determined by IPA using the differential proteome dataset obtained from RAW 264.7 cells. The full lists of pathways are shown in Table S5. See also Figure S5 and Tables S4 and S5. Gastric colonization was determined by culture (C). H&E staining of the stomach (D) was used to generate the inflammation score of the gastric tissues (E). In all the panels, the mean ± SEM is shown; *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Scale bars in (D), 50 μm. See also Figures S9 and S10 and Table S7. Gobert et al. Page 27 Cell Rep. Author manuscript; available in PMC 2021 January 18.

Author Manuscript
Author Manuscript Author Manuscript Author Manuscript  (A). Colon was weighed (B). H&E staining of the colon (C) was used to generate the score for colitis (D); the inflammation scores in uninfected Dhps fl/fl and Dhps Δmye mice were 0 in stomach and colon. In all the panels, the mean ± SEM is depicted; *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Scale bars in (C), 50 μm. See also Figure S9. (A-D) The gastric tissues of Dhps fl/fl and Dhps Δmye mice infected for 8 weeks with Hp PMSS1 (A) or the colon of these animals infected with C. rodentium for 14 days (C) were immunostained for CD68 and EIF5A Hyp , as well as CD68 and NOS2. Scale bars, 50 μm. Quantification of CD68 + /EIF5A Hyp + cells and CD68 + /NOS2 + cells among the total number of CD68 + cells is shown in (B) and (D). The immunofluorescence images are representative of staining performed on two uninfected animals and five infected mice in each group, and the same number of animals was used for quantification. Bars represent mean ± SEM in all graphs; **p < 0.01 and ****p < 0.0001.