A genome-wide CRISPR screen defines host determinants of early Brucella infection in human macrophage-like cells

ABSTRACT

Brucella spp. are widespread intracellular animal pathogens that cause brucellosis, a significant zoonosis. Despite the global impact of brucellosis on animal and human health, the host genes that support Brucella infection remain incompletely defined. To address this knowledge gap, we developed a flow cytometry-based infection assay with fluorescent Brucella and performed a genome-wide CRISPR-Cas9 loss-of-function screen in human macrophage-like cells. Disruption of >150 host genes significantly reduced intracellular B. abortus signal at 3 hours post-infection. In addition to recovering known host factors, the screen revealed previously unappreciated genes linked to endosomal trafficking, cytoskeletal remodeling, and lipid homeostasis. The screen was robust, as validation within these functional categories confirmed that the small GTPase RAB14, the Src-family kinase regulator CSK, and the phospholipid flippase subunit TMEM30A support the B. abortus and B. ovis infection process at a post-entry step. Gene set enrichment analysis further identified positive regulators of mTORC1 signaling as host factors. This result was validated by genetic disruption of LAMTOR2 and AKT1, and pharmacologic inhibition of AKT1. Together, these data indicate that the AKT-Ragulator-mTORC1 axis contributes to establishing a permissive intracellular niche. Finally, to assess whether these host requirements extend beyond Brucella, we examined infection by the unrelated intracellular pathogen Mycobacterium abscessus. CSK, AKT1, and LAMTOR2 were required for efficient M. abscessus infection, whereas RAB14 was dispensable. Together, these results define host genes that impact Brucella infection and distinguish shared versus pathogen-specific host dependencies exploited by intracellular bacteria.

INTRODUCTION

Brucella spp. are intracellular pathogens that cause brucellosis, a globally distributed zoonosis that inflicts a significant burden on human and animal health (1, 2). Following host cell entry through actin-dependent mechanisms (3), Brucella localizes to a membrane-bound compartment known as the Brucella-containing vacuole (BCV), which undergoes a defined series of intracellular maturation steps. Newly formed BCVs initially traffic along the endocytic pathway, transiently acquiring early (Rab5 and EEA1) and late (Rab7 and LAMP1) endosomal markers while undergoing progressive acidification (4–7). Endosomal Brucella-containing vacuole (eBCV) acidification provides a cue that induces the VirB Type IV secretion system (T4SS), enabling delivery of bacterial effectors that remodel vesicular trafficking to support establishment of a replicative vacuole (rBCV) that is derived from the endoplasmic reticulum (8–12).

Multiple host surface molecules and membrane microdomains have been implicated in Brucella uptake, including sialic acid-containing receptors, components of the extracellular matrix (13), scavenger receptors (14, 15), complement receptors, cellular prion protein (16), and other lipid raft-associated molecules (17). Host complement proteins and antibodies can further modulate the route of entry and influence infection outcomes (17–21). These results indicate that Brucella infection mechanisms will vary depending on the host cell type, host exposure history, and the physiological state of the bacterium. Given the complexity and context dependence of Brucella entry, a more integrated understanding of the host determinants that facilitate infection is needed. Genetic approaches have begun to address this gap. For example, RNA interference (RNAi)–based screens in both invertebrate (22) and mammalian (23) host cells have highlighted roles for endoplasmic reticulum (ER) stress responses, autophagy, and vesicle trafficking in successful Brucella infections. Yet many questions remain regarding the host pathways that are required for Brucella to initiate a productive eBCV.

To build on this prior work, we conducted a genome-wide CRISPR–Cas9 loss-of-function screen to identify host genes required for Brucella infection of macrophage-like cells. Using a pooled knockout library in THP-1 cells and a flow cytometric readout of infection with highly fluorescent Brucella strains, we distinguished infected from uninfected cells by fluorescence-activated cell sorting (FACS) and quantified guide RNA representation between these populations by amplicon sequencing. Our results provide a genome-scale framework for understanding early stages of Brucella infection of human macrophage-like cells. This study confirms several known host factors and identifies previously unrecognized host genes and pathways involved in endosomal trafficking, cytoskeletal remodeling, lipid metabolism, and mTOR signaling that support the early stages of mammalian cell infection by Brucella and other intracellular pathogens.

RESULTS

Development of a flow cytometry-based approach to study early Brucella infection

To discriminate between infected and uninfected THP-1 macrophage-like cells by flow cytometry, we engineered highly fluorescent Brucella strains by integrating the mNeonGreen gene at the chromosomal glmS locus. We piloted this infection assay using B. ovis, a largely ovine-restricted pathogen that poses no zoonotic risk. Fluorescent B. ovis exhibited wild-type (WT) growth kinetics in axenic culture (Fig. 1A) and was readily visualized inside THP-1 cells by fluorescence microscopy. A multiplicity of infection (MOI) of 1,000 provided a clear and reproducible separation between infected and uninfected host cell populations at a 3 hours timepoint (Fig. 1B; Fig. S1A). A shorter infection time (2 hours) did not yield a sufficiently high enough number of infected cells to conduct a flow cytometry-based genetic screen (Fig. 1C; Fig. S1B). Importantly, there was no impact on THP-1 host viability following infection with B. ovis at an MOI of 1,000 for 3 hours (Fig. 1D). We therefore proceeded to assess THP-1 infection by B. abortus::mNeonGreen using the same infection conditions. Flow cytometry showed that 50% of THP-1 cells were infected with B. abortus (Fig. 1E). We concluded that our infection protocol enabled sufficient B. abortus entry in a high enough percentage of cells to conduct a flow cytometry-based screen for host genes that impact the early stages of B. abortus infection.

Fig 1.

Flow cytometry enables detection of early Brucella infection in THP-1 macrophages. (A) Growth curves of WT and mNeonGreen-expressing B. ovis in rich medium over 72 hours post-inoculation. (B) Percentage of THP-1 macrophages infected with mNeonGreen B. ovis after 3 hours of infection at MOIs of 100 or 1,000, as assessed by flow cytometry. (C) Percentage of THP-1 macrophages infected with mNeonGreen B. ovis after 2 or 3 hours of infection at an MOI of 1,000. (D) Representative flow cytometry plots gated on live, single THP-1 macrophages. Cells were stained with Zombie NIR viability dye and infected with mNeonGreen B. ovis (MOI 1,000). The left panel shows a control population of heat-treated, uninfected macrophages. Percentages of live/dead (Zombie NIR) and infected (mNeonGreen-positive) cells are indicated. (E) Representative flow cytometry plots of THP-1 macrophages infected with mNeonGreen B. abortus (MOI 1,000), showing a distinct population of mNeonGreen-positive infected cells. Data in panels B and C represent means ± standard deviation (SD) from three biological replicates. Statistical significance was determined by one-way ANOVA followed by Dunnett’s multiple comparison test (****, P < 0.0001).

CRISPR-Cas9 loss-of-function screen identifies host genes required for Brucella infection of macrophages

We generated a genome-wide loss-of-function library in immortalized, Cas9-expressing THP-1 monocytes using a pooled lentiviral set of single-guide RNAs (sgRNAs) targeting human genes (24) (Fig. 2A; see Materials and Methods for details on library design and quality metrics). To identify host genes that facilitate B. abortus entry and establishment of the eBCV, we infected a PMA-differentiated THP-1 CRISPR library with fluorescently labeled B. abortus. At 3 hours post-infection, host cells were separated by FACS into the top 10% mNeonGreen-positive (infected) and bottom 10% mNeonGreen-negative (uninfected) populations (Fig. 2A). Genomic DNA was extracted from both groups, and sgRNAs were PCR-amplified and sequenced. Relative sgRNA abundance between the two sorted populations was compared to identify candidate host genes that impacted Brucella infection.

Fig 2.

RAB14 is a host factor that facilitates Brucella infection of macrophages. (A) Schematic of the genome-wide loss-of-function screen. A CRISPR-Cas9 pooled sgRNA library was introduced into Cas9-expressing THP-1 macrophage-like cells, followed by infection with mNeonGreen-expressing Brucella for 3 hours. Infected (mNeonGreen⁺) and uninfected (mNeonGreen⁻) populations were separated by FACS, and the abundance of sgRNAs in each population was quantified by sequencing. (B) Plot of gene enrichment P values from the B. abortus CRISPR-Cas9 screen, calculated using the α-robust rank aggregation (α-RRA) algorithm in MAGeCK. P-value threshold of 10⁻² (1/P of 10²) is drawn as a dotted line. The small GTPases RAB14 and RAB5C are highlighted in red. (C) Validation of RAB14 as a host factor. THP-1 macrophages expressing a RAB14-targeting sgRNA were infected with mNeonGreen B. abortus or B. ovis at an MOI of 1,000. The percentage of infected cells was quantified by flow cytometry and normalized to non-targeting control (NTC). Data represent means ± SD from three biological replicates. Statistical significance was determined by one-way ANOVA followed by Dunnett’s multiple comparison test (*, P < 0.05; **, P < 0.0025). Editing efficiency for both sgRNAs >60%. (D) Latex bead uptake assay in NTC and RAB14-deficient macrophages. Cells were incubated with yellow-green fluorescent beads (bead:cell ratio = 10) for 3 hours. Representative flow cytometry plots show comparable bead uptake between genotypes, indicating that phagocytic capacity is preserved in RAB14-deficient cells.

Host gene disruptions associated with increased intracellular Brucella burden were identified (Table S1), but gene set enrichment analysis (GSEA) of these putative restriction factors did not yield significant pathway-level enrichment. Furthermore, comparison of these putative restriction genes in our data set to a previous siRNA screen of B. abortus infection in HeLa cells (23) revealed no overlap, with the exception of NME8. This limited concordance likely reflects differences in cell type (THP-1 macrophages versus HeLa epithelial cells), infection conditions, and readout timing and suggests that restriction mechanisms may be more context- or cell-type-specific than host-permissive pathways.

We therefore focused on host genes whose disruption significantly reduced intracellular B. abortus burden. Specifically, we identified over 150 host genes whose disruption was associated with reduced levels of B. abortus infection based on the following criteria: (i) detection of at least two sgRNAs in the data set, (ii) an experimental log selection ratio greater than 1, and (iii) experimental P value less than 0.01 (Table S1). Several hits correspond to host factors previously implicated in Brucella-containing vacuole (BCV) biogenesis or maintenance, including calnexin (CANX) (10), CDC42 (3), and the early endosomal GTPase RAB5C (25, 26), underscoring the robustness of our screen and the biological relevance of our screen results. In addition, we identified CYFIP1, a core component of the WAVE regulatory complex that promotes actin polymerization at the plasma membrane (27); ARPC2, a subunit of the Arp2/3 complex that drives actin nucleation; and TBC1D10B, a GTPase-activating protein involved in endosomal recycling (28). In contrast to host restriction factors discussed above, these pathogen supportive genes were previously implicated in B. abortus infection in a prior siRNA screen in HeLa cells (23), lending support to our loss-of-function results.

Endosomal trafficking and recycling genes supporting early Brucella infection: a role for RAB14

As described above, the endosome-associated small GTPase RAB5C and the GAP protein TBC1D10B were identified as host factors that promote early B. abortus infection (Table S1; Fig. 2B). Additional hits in genes involved in endosomal processes included ARF4, which regulates recycling endosome integrity (29) (Table S1). We further identified DNAJC13, a chaperone that alters the spatial organization of endosomal subdomains involved in recycling and degradation when disrupted (30, 31). A top hit in the B. abortus screen was the GTPase RAB14, which coordinates membrane trafficking at the interface of early and recycling endosomes and the trans-Golgi network (32–34) (Fig. 2B). To validate that RAB14 plays a role in infection, we generated RAB14-deficient THP-1 macrophages using two distinct sgRNAs and observed significantly reduced B. abortus intracellular signal compared to NTC cells (Fig. 2C; Fig. S2A). RAB14 disruption also diminished B. ovis infection (Fig. 2C), showing RAB14 to be a host factor for both the smooth zoonotic strain B. abortus and the naturally rough ovine-restricted pathogen B. ovis. To test whether the reduced Brucella infection observed in RAB14 mutant cells was due to a general defect in phagocytosis, we incubated the mutant THP-1 macrophages with fluorescent 1 µm latex beads. Bead uptake was comparable to non-targeting control cells (Fig. 2D), indicating that basic phagocytic capacity was intact. Thus, the infection defect in RAB14-deficient cells is not a result of impaired general phagocytosis.

Expanding understanding of cytoskeleton regulatory genes that promote Brucella infection

GSEA (35) of the ranked list of host mutants with reduced B. abortus infection revealed expected enrichment of functional categories involved in cytoskeletal and cell motility regulation (Table S1). Among these genes was C-terminal Src kinase (CSK) (Fig. 3A), a regulator that links G protein signaling to actin remodeling (36) and that has been implicated in dengue virus replication (37) and in the internalization of Pseudomonas aeruginosa by mammalian cells (38). To validate CSK as a host factor that supports Brucella infection, we generated a CSK mutant cell line. Infection assays using fluorescent B. abortus and B. ovis revealed significantly reduced mNeonGreen signal in the CSK mutant lines compared to control cells (Fig. 3B; Fig. S2B), supporting the conclusion that CSK facilitates infection by both Brucella species. Again, uptake of fluorescent 1 µm latex beads was comparable between mutant and control cells (Fig. 3C; Fig. S2D), indicating that basic phagocytic function was preserved in a CSK mutant.

Fig 3.

CSK and TMEM30A promote Brucella infection of THP-1 macrophages. (A and D) Statistical enrichment plots (1/P; dotted line) calculated using the α-RRA algorithm in MAGeCK. CSK (A) and TMEM30A (D) met the criteria for B. abortus host factors (≥1 log₂ fold change, ≥2 independent sgRNAs, P < 0.01). (B and E) Validation of CSK (B) and TMEM30A (E) as host factors. THP-1 cells expressing sgRNAs targeting either gene were infected with mNeonGreen B. ovis or B. abortus at an MOI of 1,000 for 3 hours. The percentage of infected macrophages was quantified by flow cytometry and normalized to the NTC. Data represent mean ± SD from three biological replicates. Significance in panels B and E was assessed by an unpaired t-test (***, P < 0.001). (C and F) Phagocytosis assays using yellow-green fluorescent latex beads (bead:cell ratio = 10) in THP-1 cells expressing NTC or gene-targeting sgRNAs. The percentage of bead-positive cells was quantified by flow cytometry and normalized to NTC. Editing efficiency—CSK: 61%; TMEM30A: 71%.

Several additional cytoskeleton-related genes were identified whose disruption significantly reduced intracellular bacterial signal, including ELMO2, which facilitates actin remodeling during phagocytosis (39). Both CYFIP1 (described above) and WASF2 mutants showed diminished infection in the screen (Table S1). These two genes encode core components of the WAVE complex that drive actin polymerization during dynamic membrane remodeling events (40). Other cytoskeleton/cell motility-related hits included WDR1, an actin disassembly factor that enhances cofilin-mediated filament turnover and cell motility (41). Loss of NHLRC2, which regulates RhoA-Rac1-dependent actin polymerization during phagocytosis (42), also resulted in decreased THP-1 infection. These results are consistent with the known role of actin-driven uptake in Brucella entry (3, 43, 44) and expand our view of actin polymerization and disassembly processes that support the early stages of macrophage infection by Brucella.

Host lipid metabolism/trafficking as a determinant of Brucella intracellular infection

Our screen also identified several host genes involved in lipid metabolism and trafficking that are required to support B. abortus infection in THP-1 cells. Among the top hits in this functional category was TMEM30A, which encodes a subunit of phospholipid flippases responsible for maintaining membrane lipid asymmetry (45) (Fig. 3D). TMEM30A was recently identified as a host factor for Mycobacterium bovis BCG (46), SARS-CoV-2 (47, 48), and murine norovirus (49) infection, indicating this gene has an important role in mammalian host cell interactions with both bacterial and viral pathogens. To validate the role of this gene in Brucella infection, we generated a TMEM30A mutant THP-1 cell line. Compared to the non-targeting control, the TMEM30A mutant showed significantly reduced intracellular signals for B. abortus and B. ovis (Fig. 3E; Fig. S2C), supporting the screen results. Uptake of fluorescent latex beads was comparable between mutant and control lines (Fig. 3F; Fig. S2D), showing that TMEM30A disruption does not impair general phagocytic capacity.

Several other lipid-related genes that were disrupted and significantly reduced B. abortus infection were identified (Table S1). Among these were PGS1, previously identified as a host factor in an siRNA knockdown screen in HeLa cells infected with B. abortus (23), and AGPAT5, both of which participate in ER-localized glycerophospholipid metabolism. Disrupting PTPMT1, a mitochondrial phosphatase required for cardiolipin biosynthesis, DOLK, which produces dolichyl monophosphate for ER-associated glycosylation, and PIP4K2A, a phosphatidylinositol phosphate kinase involved in phosphoinositide signaling and membrane trafficking, also significantly reduced early B. abortus infection of THP-1 cells (Table S1). These results provide evidence that host lipid metabolism and homeostasis, in particular ER-localized lipid metabolic processes, play a key role in supporting intracellular Brucella infection.

mTORC1 signaling processes promote B. abortus macrophage infection

Genes categorized as positive regulators of mTOR Complex 1 (TORC1) signaling were significantly enriched among host mutants that showed reduced B. abortus infection (Table S1). Although the TORC1 complex itself has not been previously implicated in Brucella pathogenesis, the broader mTOR signaling axis is a central regulator of macrophage functions (e.g., polarization, inflammatory signaling, metabolism, and autophagy) that are likely to influence Brucella fitness within the host cell environment (50–53). Prompted by this observation, we sought to validate the role of TORC1 signaling-associated genes in promoting early Brucella infection. Among the enriched hits, we prioritized LAMTOR2 and AKT1 for functional validation (Fig. 4A; Table S1). Notably, LAMTOR2 was previously identified as a significant hit in an siRNA screen of B. abortus infection in HeLa cells (23), providing independent support for its role in Brucella-host interactions. THP-1 macrophages with CRISPR-mediated disruptions in either LAMTOR2 or AKT1 were infected with fluorescent B. abortus or B. ovis, both of which showed significantly reduced intracellular signal at 3 hours compared to non-targeting control cells (Fig. 4B and C; Fig. S3A and B). Latex bead uptake was comparable to control cells in both mutants (Fig. 4D; Fig. S3C), indicating that this reduction does not reflect a general phagocytic defect.

Fig 4.

TORC1 signaling genes support efficient Brucella infection of macrophages. (A) Plot showing gene enrichment P values for the B. abortus CRISPR-Cas9 screen, ranked using the α-RRA algorithm in MAGeCK (P-value threshold of 10⁻² (1/P of 10²) is drawn as a dotted line). TORC1-related genes LAMTOR2 and AKT1 are highlighted. (B and C) Validation of AKT1 and LAMTOR2 as host factors. THP-1 cells expressing sgRNAs targeting either gene were infected with mNeonGreen B. abortus (B) or B. ovis (C) at an MOI of 1,000 for 3 hours. The percentage of infected macrophages was quantified by flow cytometry and normalized to the NTC. Data represent mean ± SD from three biological replicates. Significance in panels B and C was assessed using one-way ANOVA with Dunnett’s multiple comparison test (****, P < 0.0001; ***, P < 0.001; **, P < 0.01; *, P < 0.05). Editing efficiency—AKT1: 63%; LAMTOR2: 87%. (D) Phagocytosis assay in THP-1 macrophages expressing NTC, AKT1, or LAMTOR2 sgRNAs. Cells were incubated with yellow-green fluorescent latex beads (bead:cell ratio = 10) for 3 hours, and bead-positive cells were quantified by flow cytometry. Data are normalized to the mean uptake in NTC cells. (E) THP-1 viability following treatment with increasing concentrations of AKT inhibitor X for 48 hours, assessed by XTT assay and normalized to untreated controls. (F) Axenic growth of B. abortus in broth culture in the presence of AKT inhibitor X. Cultures were incubated for 48 hours, and OD₆₀₀ values were normalized to untreated samples. (G) THP-1 macrophages pretreated with 10 µM AKT inhibitor X showed reduced B. abortus infection, as measured by flow cytometry. Data represent mean ± SD from three biological replicates; unpaired t-test (P < 0.01).

As an orthologous approach to explore the role of AKT in promoting infection, we chemically inhibited AKT activity in THP-1 macrophages using Akt inhibitor X (54). Concentrations above 25 µM impaired host cell viability (Fig. 4E), and doses above 50 µM inhibited Brucella growth in axenic culture (Fig. 4F). However, treatment of THP-1 cells with a sub-toxic, sub-inhibitory dose (10 µM) prior to infection significantly reduced intracellular B. abortus levels as measured by flow cytometry (Fig. 4G; Fig. S3D), providing further evidence that AKT activity promotes early Brucella infection. In addition to LAMTOR2 and AKT1, other significant screen hits with known functional links to TORC1/mTOR signaling included WDFY3 (ALFY), AMBRA1, HDAC3, SLC7A5, DYRK1A, RPTOR, and SIRT6. WDFY3 is reported to localize to RAB5- and EEA1-positive early endosomes in HeLa cells and is highly enriched at dynamic cellular protrusions (55), which may explain its role in supporting early stages of Brucella infection. Taken together, our results indicate that TORC1 signaling plays an important role in early host-Brucella interactions.

Validated host factor mutations do not impair Brucella uptake

Although latex bead experiments confirm that general phagocytic capacity is intact in all mutant lines, they do not directly test Brucella entry. To address this, we measured B. abortus internalization at 30 minutes post-infection across all six validated mutant lines. The fraction of mNeonGreen-positive macrophages did not differ significantly from the non-targeting control in any line at this early timepoint (P > 0.05 for all; Fig. S5A), and viable intracellular bacteria enumerated by gentamicin-protection colony-forming unit (CFU) at 30 minutes were likewise equivalent to control (P > 0.05; Fig. S5B). Together, these results indicate that RAB14, CSK, TMEM30A, AKT1, and LAMTOR2 are dispensable for Brucella entry into THP-1 macrophages and support the conclusion that these factors act downstream of entry during early BCV maturation.

We also assessed bacterial viability at the 3-hour screening timepoint by gentamicin-protection CFU in all mutant lines. Viable intracellular bacteria were not significantly different from control in any line (P > 0.05; Fig. S5C), though we note that with three biological replicates, this assay has limited power to resolve differences smaller than approximately threefold. Modest effects on viable burden (in line with those quantified by our flow cytometry experiments) cannot be formally excluded. Given that mNeonGreen has a reported pKa near 5.7 (56), it is also possible that altered vacuolar acidification in mutant macrophages reduces per-bacterium fluorescence signal independently of bacterial number, contributing to the flow cytometry phenotype observed at 3 hours. Distinguishing these contributions will require fluorescence-independent viability assays at later post-infection timepoints in future work.

Comparative analysis of host factors in Brucella and Mycobacterium abscessus macrophage infection

To test whether validated host genes that support Brucella infection also contribute to infection by an unrelated intracellular pathogen, we challenged the individual mutant THP-1 cell lines with Mycobacterium abscessus (Mab), an opportunistic pathogen of increasing concern in people with chronic lung disease (57–59). We infected mutant and non-targeting control THP-1 cells with a previously optimized fluorescent Mab strain (ATCC 19977) that constitutively expresses mEmerald GFP (60) and quantified infection at 4 hours post-infection. Disruption of CSK, AKT1, or LAMTOR2 significantly reduced Mab infection relative to controls, whereas RAB14 disruption had no effect (Fig. 5; Fig. S4). Together, these results indicate that CSK, AKT1, and LAMTOR2 support infection by unrelated intracellular pathogens in a THP-1 macrophage model, while RAB14 appears to be Brucella-specific.

Fig 5.

Select THP-1 loss-of-function mutants show reduced susceptibility to Mycobacterium abscessus infection. THP-1 macrophages expressing sgRNAs targeting AKT1, LAMTOR2, CSK, or RAB14, along with a NTC, were infected with mEmerald-expressing M. abscessus at an MOI of 5 for 4 hours. The percentage of infected cells was quantified by flow cytometry and normalized to NTC. Data represent means ± SD from three biological replicates. Statistical significance was determined using one-way ANOVA followed by Dunnett’s multiple comparison test (****, P < 0.0001; not significant (ns)).

DISCUSSION

To dissect host-pathogen interactions during Brucella infection, we developed a flow cytometry-based assay and performed a genome-wide CRISPR-Cas9 loss-of-function screen in human THP-1 macrophage-like cells infected with B. abortus. By coupling fluorescence-based discrimination of infected and uninfected cells with sorting at an early timepoint, we identified host genes that support the earliest steps of Brucella infection, including entry and early establishment of the BCV. The 3-hour fluorescence readout used for our study integrates multiple cellular events, so screen hits could in principle reflect defects at entry or at any downstream step. For the six candidates we selected for validation, direct measurement of bacterial uptake and viable intracellular burden at early timepoints confirmed that the observed phenotypes reflect post-entry defects rather than impaired internalization; other screen hits may well include genes whose disruption affects Brucella entry specifically. Recovery of known regulators of BCV biogenesis, including CANX, CDC42, RAB5, and CYFIP1 (3, 10, 23, 25, 26), validates the screen and establishes a foundation for interpreting the novel candidates. Follow-up experiments uncovered new genes and pathways that support Brucella infection of mammalian hosts, including host factors shared across B. abortus, B. ovis, and the unrelated intracellular pathogen M. abscessus. Together, these results lay the groundwork for mechanistic studies of the distinct host pathways discussed below.

Identification of a new Rab GTPase that supports early Brucella infection: RAB14

Consistent with the known role of RAB5 in the formation of nascent BCVs (25, 26), our screen identified the isoform RAB5C as a host gene that promotes early B. abortus infection. The small GTPase RAB14, previously proposed as a possible host factor (61), was also identified as a significant hit. RAB14 coordinates membrane trafficking between early and recycling endosomes and the Golgi, and our results demonstrate that it supports the early stages of infection of THP-1 macrophages by both B. abortus and B. ovis. In other bacterial and fungal pathogens, including Mycobacterium tuberculosis and Candida albicans, RAB14 is implicated in delaying endosomal maturation, thereby maintaining a vacuolar environment that favors intracellular persistence (62, 63). Our results raise the possibility that Brucella spp. similarly exploit RAB14-regulated pathways to stabilize the eBCV and prevent premature maturation or lysosomal fusion prior to ER remodeling.

RAB14 modulation, however, has pathogen-specific consequences. In Legionella pneumophila, RAB14 localizes to the Legionella-containing vacuole (LCV) around 1 hour post-infection (64), and RAB14 silencing increases intracellular bacterial growth (65). In contrast, Klebsiella pneumoniae infection in cells expressing dominant-negative RAB14 shows a significant decrease in intracellular bacterial uptake (66). RAB14 has also been shown to facilitate intracellular replication of Chlamydia trachomatis by mediating the delivery of sphingolipids to chlamydial inclusions (67). Whether such a sphingolipid delivery mechanism impacts the development of the BCV during Brucella infection remains an open question. Adding to this repertoire of context-dependent functions, we found that RAB14 was dispensable for infection by Mycobacterium abscessus, indicating that RAB14 is not universally required for intracellular bacterial growth or restriction. Together, our data and prior work suggest that RAB14 differentially interacts with intracellular bacteria to either enhance or restrict uptake and growth. This specificity almost certainly reflects important differences in intracellular trafficking dynamics between Brucella spp., M. abscessus, and other bacterial species.

The absence of late endosomal and lysosomal markers such as RAB7 and LAMP1 among our top candidates supports the idea that our 3 hours post-infection screening window is enriched for host factors involved in the earliest stages of infection (specifically, eBCV formation and early trafficking events). Future studies that track RAB14 dynamics on BCVs and dissect its effector networks will clarify whether RAB14 primarily sustains a pro-entry niche or plays other roles in modulating BCV interaction with other compartments.

CSK as a conserved determinant of early infection

CSK is a negative regulator of Src-family kinases (SFKs) that maintains SFKs in an inactive conformation via C-terminal phosphorylation (68). SFKs, in turn, influence multiple processes relevant to pathogen entry, including actin remodeling, endocytic routing, and inflammatory signaling. Our discovery that CSK disruption reduces early infection by B. abortus, B. ovis, and M. abscessus suggests that precise SFK tuning by CSK favors the successful establishment of intracellular niches for both Brucella and M. abscessus. This conclusion is consistent with prior work linking CSK to microbial entry phenotypes. For example, CSK perturbation has been associated with reduced internalization of Pseudomonas aeruginosa (38) and with effects on viral replication (37). CSK-deficient THP-1 cells retained normal uptake of fluorescent beads, arguing against a global defect in phagocytosis and instead pointing to signaling abnormalities that specifically affect BCV establishment. Given the known roles of CSK and SFKs in host cell biology, defective coordination of early endosomal trafficking remains a likely mechanism underlying the attenuated infection observed in CSK mutant lines (69, 70). One possibility is that loss of CSK derepresses SFK activity in a manner that promotes bactericidal outcomes, for example, by enhancing endosome-lysosome fusion, but this remains to be tested experimentally.

On TORC1 signaling and Brucella infection

GSEA implicated mTORC1-related processes as important for the early stages of Brucella infection (Table S1); genes with connections to TORC1 signaling, including WDFY3, AMBRA1, HDAC3, SLC7A5, DYRK1A, RPTOR, and SIRT6, emerged as significant host factor hits in our screen. AKT–mTORC1 signaling is a central regulator of macrophage metabolism, polarization, and inflammatory programs (50, 53, 71, 72), all of which can condition the intracellular niche. Consistent with this, pharmacologic inhibition of AKT1 at a non-toxic, sub-host inhibitory dose reduced early B. abortus infection, mirroring the reduction observed in AKT1-deficient THP-1 cells and aligning with prior work showing that select intracellular bacteria and viruses exploit AKT signaling to support survival or replication (73, 74). AKT1-mutant THP-1 cells were also less susceptible to M. abscessus infection, suggesting that AKT1-dependent signaling promotes the establishment of a permissive environment for disparate intracellular bacterial pathogens. It remains an open question whether AKT1 helps to promote a Warburg-like metabolic state that was previously reported to support Brucella infection (75), or whether signaling differences between AKT isoforms (AKT1 versus AKT2) differentially shape macrophage inflammatory responses during infection (76).

In addition to AKT1, we validated LAMTOR2 as a host factor for B. abortus, B. ovis, and M. abscessus infection of THP-1 macrophages. This protein is a core component of the Ragulator complex (77), where it coordinates amino acid sensing with ERK/MAPK signaling (78, 79), influencing both mTORC1 activation and lysosomal function (80). Notably, one of the top hits in our screen, DYRK1A, was recently shown to interact with the TSC1-TSC2 complex, where it phosphorylates TSC2 to promote TORC1 activation (81). Thus, loss of DYRK1A function may reduce Brucella infection by dampening TSC–mTORC1 signaling in macrophages. Consistent with a nutrient-sensing contribution to this signaling axis, we also identified SLC7A5 (LAT1) as a host factor (Table S1). This large neutral amino acid transporter mediates leucine uptake and is required for full mTORC1 activation and metabolic reprogramming in immune cells (82–84). Altogether, these results point to a broad network of regulators that converge on mTORC1 to influence a Brucella-permissive intracellular state.

Contrasting with our results, previous infection studies in LAMTOR2⁻/⁻ macrophages reported increased intracellular bacterial loads 6–24 hours after infection with Salmonella enterica serovar Typhimurium or Klebsiella pneumoniae (85, 86). In those systems, LAMTOR2 thus appears to limit intracellular replication. Our result that LAMTOR2 disruption reduces early infection by two Brucella species and by M. abscessus indicates that LAMTOR2 can act as either a pro-infection host factor or a restriction factor depending on the organism, cell type, and stage of infection.

It is possible that Brucella depends on Ragulator-dependent lysosomal cues during eBCV maturation (e.g., controlled changes in luminal pH, ion composition, or positioning of BCVs within the endolysosomal network) that are required to trigger virB (Type IV secretion system) expression and subsequent ER remodeling without committing to terminal lysosomal fusion. In this model, loss of LAMTOR2 and other positive regulators of TORC1 signaling could misregulate these cues, diverting BCVs into more bactericidal pathways. Dissecting whether and how Ragulator-mTORC1 and AKT1 integrate metabolic, lysosomal, and signaling inputs with other validated host factors at early timepoints will be essential to understand why mTORC1 pathway components are required for efficient infection by Brucella and M. abscessus. We note that the contribution of TORC1 signaling is supported here by genetic disruption of multiple pathway components and by pharmacological AKT inhibition, but direct pharmacological inhibition of mTORC1 and direct measurement of infection-induced mTOR activity were not performed in this study; both are important directions for future mechanistic work.

Altogether, our functional genetic approach identified and validated distinct host cell processes that are required for Brucella infection and may be potential host targets to limit diverse intracellular pathogens.

MATERIALS AND METHODS

Bacterial strains and growth conditions

Brucella ovis 25840 and Brucella abortus 2308 were grown on tryptic soy agar (TSA; Difco Laboratories) plates supplemented with 5% (vol/vol) sheep blood (Quad Five) or in brucella broth (BB; Difco Laboratories) dissolved in milliQ water for liquid cultures. B. ovis and B. abortus cells were incubated at 37°C with 5% (vol/vol) CO₂ supplementation. Smooth Mab (ATCC 19977) cultures were grown aerobically in Middlebrook 7H9 medium supplemented with 10% Middlebrook OADC (oleic acid, dextrose, catalase, and bovine albumin) at 37°C. mEmerald GFP-expressing strains were propagated in medium containing 5 µg/mL zeocin (Invivogen). All Escherichia coli strains were grown in liquid LB or on LB solidified with 1.5% (wt/vol) agar. Stbl3 or Top10 strains were incubated at 30°C or 37°C, respectively. Strain information is available in Table S2.

Construction of Brucella mNeonGreen strains

B. abortus and B. ovis strains that constitutively express mNeonGreen fluorescent protein were generated from the wild-type B. abortus 2308 and wild-type B. ovis ATCC 25840 parent strains by integration of a mini Tn7 cassette expressing mNeonGreen from the Paph promoter at the glmS site. This plasmid (pUC18-mTN7K-mNeonGreen) was generated by inserting a Paph-mNeonGreen fusion into the MCS of pUC18-mTn7K. The pUC18-mTn7-mNeonGreen-kanamycin plasmid was co-conjugated into B. abortus with the helper plasmid pTNS3 expressing the Tn7 integrase gene as previously described for the construction of lux-expressing Brucella strains (87). B. abortus and B. ovis colonies carrying the integrated mTn7-mNeonGreen construct were selected on TSA blood plates containing 50 µg mL⁻¹ kanamycin. Strains and plasmids are available in Table S2.

THP-1 cell culture

THP-1 macrophage-like cells were grown to a maximum titer of 1 × 10⁶ cells mL⁻¹ in complete RPMI 1640 medium supplemented with 2 mM glutamine (GIBCO), and 10% heat-inactivated fetal bovine serum (vol/vol) (HI FBS) (HyClone).

THP-1 infections and treatments

THP-1 Brucella infection

THP-1 cells were seeded at a titer of 1 × 10⁶ cells per well in six-well plates, and phorbol myristate acetate (PMA) was added at a final concentration of 50 ng µL⁻¹ to induce differentiation into macrophage-like cells for 48–72 hours prior to infection. B. ovis and B. abortus strains expressing mNeonGreen were harvested from 48-hour-old plates, resuspended in RPMI containing 10%HI FBS, and added to tissue culture plates on the day of infection at a MOI of 1,000. Plates were centrifuged for 5 minutes at 150 × g and incubated for 3 hours at 37°C in 5% (vol/vol) CO₂. The medium was removed, and fresh medium containing 50 µg mL⁻¹ gentamicin was supplied, followed by a 30-minute incubation. Following gentamicin incubation, the medium was removed, and cells were washed with PBS. Cells were treated with TrypLE Express (Gibco) at 37°C for 15 minutes and gently lifted and harvested. Harvested cells were centrifuged at 500 × g and washed once with PBS. For viability analysis, the supernatant was aspirated, and cells were incubated for 15 minutes at room temperature in the dark with Zombie NIR Fixable Viability Dye (BioLegend) diluted in PBS. Stained cells were washed with Cell Staining Buffer (BioLegend), fixed in 3.2% paraformaldehyde for 20 minutes at room temperature, and washed three times with PBS. After completing a confirm-kill protocol, macrophage infection was quantified using the Attune CytPix benchtop analyzer (ThermoFisher).

THP-1 Brucella 30-minute uptake assay

To measure early bacterial uptake, THP-1 cells were seeded, differentiated, and infected with mNeonGreen-expressing B. abortus as described above for THP-1 Brucella infection, with the following modifications. Cells were differentiated with PMA (50 ng/µL) for 48 hours, and following the 5-minute, 150 × g centrifugation, infections were incubated for 30 minutes (rather than 3 hours) at 37°C in 5% (vol/vol) CO₂. Extracellular bacteria were then killed by a 10-minute incubation in fresh medium containing 50 µg/mL gentamicin (rather than 30 minutes). Cells were washed with PBS, detached with TrypLE Express (37°C, 15 minutes), harvested, stained with Zombie NIR Fixable Viability Dye (BioLegend), fixed in 3.2% paraformaldehyde, and the percentage of mNeonGreen-positive macrophages within the live, single-cell gate was quantified on the Attune CytPix benchtop analyzer (ThermoFisher) as described above.

Gentamicin-protection CFU assay

To enumerate viable intracellular bacteria, THP-1 cells were seeded at a titer of 1 × 10⁵ cells per well in 96-well plates, and PMA was added at a final concentration of 50 ng/µL to induce differentiation into macrophage-like cells for 48 hours prior to infection. mNeonGreen-expressing B. abortus was harvested from 48-hour-old plates, resuspended in RPMI containing 10% HI FBS, and added at an MOI of 1,000. Plates were centrifuged for 5 minutes at 150 × g and incubated for 30 minutes or 3 hours at 37°C in 5% (vol/vol) CO₂. The medium was removed and replaced with fresh medium containing 50 µg/mL gentamicin to kill extracellular bacteria, followed by a 10-minute incubation (for 30-minute infections) or a 30-minute incubation (for 3-hour infections). Following gentamicin incubation, the medium was removed, and cells were washed with PBS. Macrophages were lysed by incubation in water for 10 minutes until lysis occurred, and lysates were serially diluted 10-fold in PBS. 5 µL of each dilution was spotted onto TSA plates containing 50 µg/mL kanamycin, and CFUs per mL were enumerated after 3 days of incubation.

THP-1 chemical pretreatment

THP-1 cells were seeded at 1 × 10⁶ cells per well in six-well plates and treated with 50 ng/µL PMA to induce differentiation into macrophage-like cells for 48–72 hours prior to infection. Twenty-four hours before infection, AKT inhibitor X (Cayman Chemical) was added to the culture media at a final concentration of 10 µM and incubated at 37°C with 5% CO₂. On the day of infection, mNeonGreen-expressing Brucella abortus cells were resuspended from 48-hour-old plates in RPMI supplemented with 10% (vol/vol) heat-inactivated FBS. Media from plates were removed and replaced with fresh media containing B. abortus at a MOI of 1,000. Plates were centrifuged at 150 × g for 5 minutes, then incubated for 3 hours at 37°C with 5% CO₂. Following infection, the media were removed, the cells were washed with PBS, and the macrophages were detached using TrypLE Express (Gibco) at 37°C for 15 minutes. Lifted cells were gently harvested, centrifuged at 500 × g, and washed once with PBS. Cells were then fixed in 3.2% paraformaldehyde for 20 minutes at room temperature, followed by three PBS washes. After completing a confirm-kill protocol, macrophage infection was quantified using the Attune CytPix benchtop analyzer (ThermoFisher). Cell viability was assessed as described above.

THP-1 Mycobacterium abscessus infection

THP-1 cells were seeded at 1 × 10⁶ cells per well in six-well plates and treated with 50 ng/µL PMA for 48–72 hours to induce differentiation into macrophage-like cells, as described above. Log-phase M. abscessus cultures were resuspended in RPMI supplemented with 10% (vol/vol) heat-inactivated FBS, and large clumps were removed by centrifugation at 800 × g for 5 minutes. The resulting single-cell suspension was used to infect THP-1 macrophages at an MOI of 5. Infected plates were centrifuged at 150 × g for 5 minutes to promote contact and incubated for 4 hours at 37°C with 5% CO₂. Following infection, the media were removed, and the cells were washed with PBS. Macrophages were detached using TrypLE Express (Gibco) at 37°C for 15 minutes, then gently harvested, centrifuged at 500 × g, and washed once with PBS. For viability assessment, cells were incubated with Zombie NIR fixable viability dye (BioLegend) diluted in PBS for 15 minutes at room temperature in the dark, followed by a wash with cell staining buffer (BioLegend). Cells were then fixed in 3.2% paraformaldehyde for 20 minutes and washed three times with PBS. After completing a confirm-kill protocol, infection and viability were quantified using the Attune CytPix benchtop analyzer (ThermoFisher).

Fluorescent bead uptake assay

A total of 1 µm carboxylate-modified yellow-green fluorescent polystyrene latex beads (Sigma-Aldrich) were diluted in PBS to 1 × 10⁷ beads per mL. Beads were incubated with THP-1 cells for 3 hours. Following bead exposure, macrophages were washed with PBS and fixed in 3.2% paraformaldehyde solution for 20 minutes. The percentage of macrophages with yellow-green fluorescence latex beads was quantified by flow cytometry using the Attune CytPix benchtop analyzer.

CRISPR screen and analysis

We used the Human Brunello CRISPR knockout pooled guide library (Addgene #73178; gift of David Root and John Doench) (24) to generate a THP-1 gene knockout pool. Briefly, lentiviral particles were produced and used to transduce THP-1 cells stably expressing Cas9 at low MOI to favor single-sgRNA integration. Over 99.2% of sgRNAs were detected in the resulting mutant pool used for infection. The pool had a low Gini index (0.091), indicating good evenness across the mutant library (average mapped read depth of ~540 reads/sgRNA).

This THP-1 mutant library was infected with a B. abortus strain that constitutively expressed mNeonGreen from a Tn7 construct integrated at the glmS locus; infections were at an MOI of 1,000 for 3 hours. After 3 hours, macrophages were washed with PBS, and then fixed with 3.2% paraformaldehyde. The THP-1 cells were then sorted in BSL3 containment using a Bio-Rad S3e cell sorter to isolate mNeonGreen+ (i.e., infected) and mNeonGreen− (i.e., uninfected) macrophages. A total of 1.5–2.5 × 10⁷ cells were sorted into each bin from two replicate samples. Genomic DNA was isolated from each sorted population using Qiagen DNeasy kits after reversing DNA cross-links following a 55°C incubation overnight. Amplification of sgRNAs by PCR was performed as previously described (88) using Illumina-compatible primers from IDT, and amplicons were sequenced on an Illumina NovaSeq (B. abortus reads). Amplicon sequencing depth was assessed for the B. abortus experiment (~15 million mapped reads per sample). Sequenced reads were trimmed to remove any adapter sequence and to adjust for the p5 primer stagger. We used model-based analysis of genome-wide CRISPR-Cas9 knockout (MAGeCK) version 0.9.5.0 (89) to count reads with matches to the Brunello sgRNA library index without allowing for any mismatch. Subsequent sgRNA counts were median normalized in MAGeCK to account for variable sequencing depth. To test for sgRNA and gene enrichment, we used the test command in MAGeCK to compare the distribution of sgRNAs in the mNeonGreen + and mNeonGreen− bins. To identify host factors that influence Brucella infection, we used the following criteria: (i) detection of at least two sgRNAs in the data set, (ii) an experimental log selection ratio greater than 1, and (iii) experimental P value less than 0.01.

Gene set enrichment analysis

GSEA was performed in R v4.4.3 using the packages fgsea and msigdbr. All pathway-level GSEA analyses of host factors were based on the B. abortus screen data set. Gene-level statistics were imported from a tab-delimited file containing gene symbol, P value, and log₂ fold change. To generate a directional pre-ranked list, we computed a gene score defined as (−log₁₀ P value) × (−log₂ fold change), so that genes whose disruption most strongly depleted infection (large negative [Neg LogFC] and small P value) received large positive scores.

Open gene sets for Homo sapiens were obtained from MSigDB via msigdbr, using the Gene Ontology Biological Process collection (C5, GO:BP) and Reactome canonical pathways (C2, REACTOME). For each collection, we required gene sets to contain between 15 and 500 genes. One-sided multilevel GSEA was then performed with fgseaMultilevel, testing the positive tail of the ranked list (“pos”; enrichment among host factors) and the negative tail (“neg”; enrichment among putative restriction factors) separately for each collection. This procedure returns, for every gene set, a normalized enrichment score (NES), nominal P value, Benjamini-Hochberg-adjusted P value (FDR), gene set size, and leading-edge subset. Results from all four analyses (GO:BP-pos, GO:BP-neg, Reactome-pos, Reactome-neg) were combined into summary tables and filtered by FDR. Using an arbitrary significance threshold of adjusted P value (FDR) < 0.1, no gene sets were significantly enriched in the negative (“restriction factor”) tail; thus, only host factor-enriched categories with at least one leading-edge gene (CORE_ENRICHMENT = TRUE) are reported in Table S1.

Generation of targeted CRISPR knockout lines

Individual sgRNAs were cloned into the sgOPTI vector (Addgene plasmid #85681, from the Eric Lander lab) as previously described (88). Briefly, annealed sgRNA oligonucleotides were phosphorylated and ligated into BsmBI-digested, dephosphorylated sgOPTI-Puromycin-Ampicillin (PA) plasmids. Transformants were selected on ampicillin-containing plates, and at least one sgRNA construct per target gene was packaged into lentivirus for transduction into Cas9-expressing THP-1 cells. Following puromycin selection, genomic DNA was extracted from transductants. Target regions were PCR-amplified and submitted to the Michigan State University Genomics Core for Sanger sequencing. Editing efficiency and indel size were quantified from the Sanger sequence data using Tracking of Indels by Decomposition (TIDE) analysis (90). For each gene, a single knockout line with >60% editing efficiency was selected for downstream experiments. sgRNA sequences and TIDE primers are listed in Table S3.

Colorimetric THP1 XTT cell proliferation assay

THP-1 cells were seeded at a titer of 1 × 10⁵ cells per well in RPMI supplemented with 10% (vol/vol) FBS and containing 50 ng µL⁻¹ PMA to induce differentiation into macrophage-like cells in 96-well plates (Corning) for 48–72 hours. Varying concentrations of compounds were prepared by hand and transferred to plates containing THP1 cells with final concentrations of each compound (suspended in RPMI and 10% [vol/vol] FBS) ranging from 1 µM to 100 µM. After incubation, 50 µL XTT (sodium 3′- [1-(phenyl aminocarbonyl)−3,4-tetrazolium]-bis (4-methoxy-6-nitro) benzene sulfonic acid hydrate) (Roche) was added to each well and incubated for 4 h at 37°C and 5% (vol/vol) CO₂. Following incubation, the absorbance of the reaction product was measured at 570 nm and 650 nm.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/iai.00117-26.

Supplemental figures. iai.00117-26-s0001.pdf.

Fig. S1 to S5.

Table S1. iai.00117-26-s0002.xlsx.

Screen results.

Table S2. iai.00117-26-s0003.xlsx.

Strains and plasmids.

Table S3. iai.00117-26-s0004.xlsx.

sgRNA sequences.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.