Identification of Type 4B Secretion System Substrates That Are Conserved among Coxiella burnetii Genomes and Promote Intracellular Growth

ABSTRACT

Coxiella burnetii is a Gram-negative pathogen that infects a variety of mammalian hosts. Infection of domesticated ewes can cause fetal abortion, whereas acute human infection normally manifests as the flu-like illness Q fever. Successful host infection requires replication of the pathogen within the lysosomal Coxiella-containing vacuole (CCV). The bacterium encodes a type 4B secretion system (T4BSS) that delivers effector proteins into the host cell. Disruption of C. burnetii T4BSS effector export abrogates CCV biogenesis and bacterial replication. Over 150 C. burnetii T4BSS substrates have been designated often based on heterologous protein translocation by the Legionella pneumophila T4BSS. Cross-genome comparisons predict that many of these T4BSS substrates are truncated or absent in the acute-disease reference strain C. burnetii Nine Mile. This study investigated the function of 32 proteins conserved among diverse C. burnetii genomes that are reported to be T4BSS substrates. Despite being previously designated T4BSS substrates, many of the proteins were not translocated by C. burnetii when expressed fused to the CyaA or BlaM reporter tags. CRISPR interference (CRISPRi) indicated that of the validated C. burnetii T4BSS substrates, CBU0122, CBU1752, CBU1825, and CBU2007 promote C. burnetii replication in THP-1 cells and CCV biogenesis in Vero cells. When expressed in HeLa cells tagged at its C or N terminus with mCherry, CBU0122 localized to the CCV membrane and the mitochondria, respectively. Collectively, these data further define the repertoire of bona fide C. burnetii T4BSS substrates.

IMPORTANCE Coxiella burnetii secretes effector proteins via a T4BSS that are required for successful infection. Over 150 C. burnetii proteins are reported to be T4BSS substrates and often by default considered putative effectors, but few have assigned functions. Many C. burnetii proteins were designated T4BSS substrates using heterologous secretion assays in L. pneumophila and/or have coding sequences that are absent or pseudogenized in clinically relevant C. burnetii strains. This study examined 32 previously reported T4BSS substrates that are conserved among C. burnetii genomes. Of the proteins tested that were previously designated T4BSS substrates using L. pneumophila, most were not exported by C. burnetii. Several T4BSS substrates that were validated in C. burnetii also promoted pathogen intracellular replication and one trafficked to late endosomes and the mitochondria in a manner suggestive of effector activity. This study identified several bona fide C. burnetii T4BSS substrates and further refined the methodological criteria for their designation.

INTRODUCTION

Coxiella burnetii is a Gram-negative intracellular pathogen capable of infecting a broad range of host organisms. Symptomatic human infections often present as an acute flu-like illness, termed Q fever, or less commonly as chronic focalized infections that manifest as endocarditis or hepatitis. Human exposures to C. burnetii are frequently traced to foods, liquids, or aerosols contaminated with by-products from infected livestock (1, 2). C. burnetii colonizes a broad range of mammalian hosts, including domesticated goats and sheep, where pathogen tropism for placental tissue coincides with fetal death and abortion. During the late stationary phase of growth, C. burnetii differentiates into a metabolically inactive small-cell variant form that exhibits pronounced stability in the extracellular environment. Birth products and other materials, including milk and feces, shed by infected animals can contain high numbers of C. burnetii organisms that are capable of causing disease in susceptible hosts (1, 3).

C. burnetii colonizes a variety of host cell types, including epithelial cells, dendritic cells, and resting macrophages (4, 5). Pathogen colonization of alveolar macrophages following inhalation of aerosolized C. burnetii is linked to acute human disease (6). During early colonization of a host cell, C. burnetii resides at neutral pH within the lumen of a nascent Coxiella-containing vacuole (CCV) that is labeled with early endosomal proteins, including Rab5 GTPase and EEA.1. Maturation of the early CCV through the endosomal pathway gives rise to vacuole acidification, exchange of Rab5 for Rab7 GTPase, and activation of C. burnetii metabolism (7, 8). Complete maturation of the CCV into an expansive, replication-competent niche requires bacterial protein synthesis and activation of the C. burnetii type 4B secretion system (T4BSS), which translocates effector proteins from the bacterium into the host cell (9, 10). C. burnetii T4BSS effectors target eukaryotic host cell pathways to promote CCV maturation and bacterial replication, modulate host immune defenses, and maintain host cell viability during the pathogen’s prolonged intracellular developmental cycle (11). Thus, functional characterization of T4BSS effectors is critical to understanding mechanisms of C. burnetii pathogenesis.

Our current understanding of C. burnetii T4BSS structure and function is largely derived from investigations of a related T4BSS found in Legionella pneumophila encoded by dot/icm (defect in organelle trafficking/intracellular multiplication) genes (12, 13). Dot/Icm T4BSSs are activated during intracellular growth, where the small amounts of native protein secreted are often refractory to methods of direct detection. Consequently, C. burnetii T4BSS effector discovery has relied heavily on genetic screens to probe C. burnetii genomes for open reading frames (ORFs) that encode peptide sequences capable of driving T4BSS-dependent export. For these genetic screens, candidate T4BSS substrate ORFs are commonly cloned into plasmids engineered for constitutive expression of proteins N-terminally fused to adenylate cyclase (CyaA) or beta-lactamase (BlaM) enzymatic reporter tags (14–16). Because methods for transformation of L. pneumophila with expression vectors preceded the development of similar genetic tools for C. burnetii, many seminal investigations expressed C. burnetii proteins in L. pneumophila to test for T4BSS-dependent export of candidate proteins. Indeed, a majority of the ~150 reported C. burnetii T4BSS substrates were originally identified using L. pneumophila (11). C. burnetii encodes homologs of many L. pneumophila Dot/Icm proteins that are required for T4BSS-dependent secretion (9, 10, 17–20), and early studies showed that many of the T4BSS substrates identified with L. pneumophila are secreted by the C. burnetii T4BSS (10, 11, 21). However, there are also important differences between these two evolutionarily related T4BSSs, according to more recent studies showing that several proteins exported in a T4BSS-dependent manner by L. pneumophila are not secreted by the C. burnetii T4BSS (14, 21–23).

C. burnetii strains exhibit significant genetic diversity and virulence phenotypes ranging from mild to severe in animal models of acute human disease (24, 25). Evidence suggests that genotype-specific T4BSS effector repertoires contribute to pathotype variation among C. burnetii strains (26). Genetic screens for C. burnetii T4BSS substrates have predominately referenced sequences from the genome of the highly virulent Nine Mile phase I (NMI) strain RSA493 (11). Intergenomic comparisons with less virulent strains such as K Q154 or Dugway 5J108-111 show that the Nine Mile genome contains many frameshift mutations that create premature stop codons within ORFs reported to encode T4BSS substrates (11, 15, 16, 27). The abundance of T4BSS substrate polymorphisms encoded among C. burnetii genomes, along with inconsistencies between T4BSS translocation results reported for L. pneumophila compared to C. burnetii, complicates efforts to identify the functional repertoire of C. burnetii T4BSS effectors. Acknowledging these challenges, this study used C. burnetii genetic tools to analyze a group of T4BSS substrates with uncharacterized functions that are conserved among a diverse set of C. burnetii genomes. Interestingly, C. burnetii did not efficiently export many proteins that were previously reported to be T4BSS substrates of L. pneumophila. Four proteins robustly translocated by C. burnetii in a T4BSS-dependent manner were also required for maximal intracellular replication of the bacterium. Collectively, the findings of this report help to further characterize C. burnetii T4BSS substrate functions and refine methods used to investigate C. burnetii T4BSS effectors.

RESULTS

Selection of proteins for analysis.

We previously compared the amino acid sequences of 143 reported C. burnetii T4BSS substrates between genomes belonging to the highly virulent Nine Mile phase I RSA493 and Henzerling RSA331 strains, the moderately virulent G Q212 strain, and minimally virulent Dugway 5J108-111 and K Q154 strains (11). T4BSS substrates encoded by sequences conserved among these C. burnetii genomes were considered most likely to encode effectors important for C. burnetii host cell parasitism. Consequently, we investigated the functions of 32 ORFs from the Nine Mile genome with predicted amino acid sequences highly conserved among the analyzed genomes (Table 1). At the onset of the study, effector activity had not been established for any of these proteins. According to annotations of the Nine Mile genome and BLASTP analysis, 15 of the ORFs selected encode hypothetical proteins unique to Coxiella species that do not have homologs found in other organisms. Three ORFs encoded homologs of conserved bacterial proteins of unknown function. Proteins with annotated functions include a bacterial serine/threonine kinase domain protein (CBU0175), coenzyme PQQ synthesis protein C (CBU0637), glyoxalase/bleomycin resistance protein (CBU0773), and ribosomal protein S18 alanine acetyltransferase (CBU0801). Also included were two uncharacterized ankyrin repeat proteins, CBU0201/AnkC and CBU1268/AnkN, along with CBU1634, which exhibits 31% sequence identity to the L. pneumophila T4BSS component IcmQ (19). None of the ORFs selected for analysis were annotated with predicted functions in eukaryotic cell biology (BLASTP [www.microbesonline.org]).

TABLE 1.

Conserved C. burnetii proteins selected for analysis

ORFa	Nameb	Protein annotationc	T4BSS translocation datad	Growth phenotypee	Reference(s)
cbu0122		Hypothetical protein*	Cb, +; Lp, −		22, 39
cbu0175		Ser/Thr protein kinase domain protein	Cb, ND; Lp, +		14
cbu0201	AnkC	Ankyrin repeat protein	Cb, ND; Lp, +		15, 16
cbu0270		Short-chain alcohol dehydrogenase	Cb, ND; Lp, +		39
cbu0410		Hypothetical membrane spanning protein*	Cb, IcmS dependent; Lp, +		14, 17
cbu0469		Hypothetical cytosolic protein*	Cb, ND; Lp, +		39
cbu0590		Hypothetical cytosolic protein*	Cb, ND; Lp, +		39
cbu0637		Coenzyme PQQ synthesis protein C	Cb, ND; Lp, +		39
cbu0773		Glyoxalase/bleomycin resistance protein	Cb, ND; Lp, +		39
cbu0801		Ribosomal-protein-S18-alanine acetyltransferase	Cb, ND; Lp, +		14
cbu0937	CirC/MceB	Hypothetical protein* | signal peptide cleavage site (+1): 24	Cb, ND; Lp, +	Growth defect	14
cbu0978		Hypothetical membrane associated protein*	Cb, ND; Lp, +	Growth defect	40
cbu1079		Conserved bacterial protein of unknown function	Cb, ND; Lp, +	Growth defect	39
cbu1198		Hypothetical cytosolic protein*	Cb, ND; Lp, +	No growth defect	39
cbu1268	AnkN	Ankyrin repeat protein	Cb, ND; Lp, +	Growth defect	16
cbu1349		Hypothetical protein*	Cb, ND; Lp, +		14
cbu1387		Hypothetical cytosolic protein*	Cb, ND; Lp, +	Growth defect	40
cbu1425	MceC	17-kDa common-antigen with lipoprotein signal peptide	Cb, ND; Lp, +	Growth defect	14
cbu1434		Conserved bacterial protein of unknown function	Cb, ND; Lp, +		39
cbu1566		Conserved bacterial protein of unknown function	Cb, ND; Lp, +		39
cbu1594	MceD	GatB/Yqey domain protein	Cb, ND; Lp, +	Growth defect	39
cbu1634	IcmQ	T4BSS structural membrane protein	Cb, ND; Lp, +		40
cbu1677	MceE	Hemerythrin HHE cation binding domain-containing protein	Cb, ND; Lp, +		39
cbu1752		Hypothetical protein*	Cb, +; Lp, −	Growth defect	22
cbu1754		Hypothetical protein*	Cb, −; Lp, +	Growth defect	18, 39
cbu1769		Alpha/beta hydrolase	Cb, ND; Lp, +		14
cbu1789		SMP-30/CGR1 family protein	Cb, ND; Lp, +		39
cbu1794		Hypothetical protein*	Cb, ND; Lp, +		40
cbu1819		Hypothetical membrane associated protein*	Cb, IcmS independent; Lp, ND		23
cbu1825		Hypothetical protein*	Cb, IcmS inhibited; Lp, +		10, 14, 17
cbu2007		Hypothetical protein*	Cb, ND; Lp, +		39
cbu2078		Filamentation induced by cAMP protein	Cb, ND; Lp, +	No growth defect	14

^aConserved C. burnetii NMI RSA493 open reading frames selected for this study.

^bPredicted functions annotated in the C. burnetii Nine Mile RSA493 genome. Asterisks indicate proteins unique to Coxiella species.

^cPublished information for C. burnetii (Cb) and L. pneumophila (Lp), summarized as positive (+) or negative (−) T4BSS export, IcmS dependency, or no data available (ND).

^dPredicted protein function annotated in the NCBI database and Sec signal peptides predicted by SignalP 6.0.

^eIntracellular growth phenotype reported for C. burnetii strains harboring mutations in the indicated ORF.

Validation of protein translocation by the Coxiella T4BSS.

C. burnetii T4BSS-dependent export was previously demonstrated for CBU0122, CBU0410, CBU1752, CBU1819, and CBU1825 (17, 22, 23). Translocation assays were conducted for the remaining 27 ORFs that had been identified using L. pneumophila and had not been validated as substrates of the C. burnetii T4BSS. Proteins fused at their N termini to the CyaA adenylate cyclase reporter were expressed in wild-type C. burnetii or the secretion-negative C. burnetii ΔicmD mutant (9) (see Fig. S1A in the supplemental material). Translocation of the CyaA fusion proteins was assessed by measuring cytosolic cyclic AMP (cAMP) concentrations in cell lysates from THP-1 cells at 48 h postinfection (hpi) (Fig. 1). Of the 27 proteins tested, only CBU1198 and CBU2007 were efficiently translocated by C. burnetii, as judged by a >2.5-fold increase in cAMP concentration relative to the empty vector control (CyaA). It was unexpected that so few of the proteins tested were not translocated by wild-type C. burnetii. Previously, we reported the existence of C. burnetii IcmS-inhibited substrates that are translocated by the C. burnetii ΔicmS mutant and not wild-type C. burnetii (17). To assess if any of the proteins in the current study were IcmS-inhibited substrates, the translocation assays were repeated using the C. burnetii ΔicmS mutant (Fig. 1). Expression of the CyaA fusion proteins by the C. burnetii ΔicmS mutant was confirmed by immunoblotting (Fig. S1B). The positive-control protein CvpB was secreted as previously reported (17), but none of the other proteins tested were exported by the C. burnetii ΔicmS mutant.

FIG 1.

C. burnetii T4BSS-dependent translocation of CyaA tagged proteins. THP-1 cells were infected for 48 h with wild-type C. burnetii or the ΔicmS or ΔicmD mutant strains expressing the indicated proteins fused to CyaA. Increases in cAMP concentration are expressed as fold change relative to the negative control cells infected with C. burnetii expressing CyaA alone (CyaA). CvpB was included as a positive control, and the cutoff for positive secretion (dotted line) was set to 2.5 times the cAMP concentration measured for the CyaA control cells. The histogram depicts means and standard deviations measured in two independent experiments conducted in triplicate.

Many of the proteins not secreted by C. burnetii in the CyaA assays were previously reported to be exported in a T4BSS-dependent manner by L. pneumophila when expressed fused to the BlaM reporter tag. To determine if CyaA and BlaM constructs produced different results, C. burnetii T4BSS-dependent translocation was investigated using the BlaM reporter system. C. burnetii strains that expressed BlaM constructs were successfully generated for 13 of the proteins that were not secreted when fused to CyaA (Fig. S1C). In THP-1 macrophages infected for 24 h, 48 h, or 72 h, the translocation of the BlaM-tagged proteins or BlaM alone was evaluated by loading cells with the fluorescent substrate CCF4-AM. Cells were also infected with C. burnetii strains expressing CBUA0012 or CBUA0015 fused to BlaM as negative and positive secretion controls, respectively (Fig. 2A). The fluorescence emission of CCF4-AM shifts from green to blue wavelengths when it is cleaved by BlaM. To normalize for cell-to-cell differences in CCF4-AM loading, the ratio of blue to green fluorescence was calculated for each cell (Fig. S2) along with the mean blue/green ratio for each group of cells (Fig. 2B). The BlaM translocation data broadly confirmed the CyaA secretion results with a minor but notable difference observed for CBU1387. Cells infected with C. burnetii expressing BlaM-CBU1387 consistently exhibited a small increase in mean blue/green ratio beginning at 48 hpi and exceeding the threshold required for positive secretion at 72 hpi (17, 18). When manually inspected under the microscope, the emission of blue fluorescence indicative of positive translocation was apparent in cells from the CBUA0015 positive-control group but not in cells from the CBU1387 or CBUA0012 negative-control groups (Fig. 2A). Comparison of blue/green ratios for individual cells at 72 hpi showed that ~90% (497/563) of the cells in the CBUA0015 group exhibited greater blue/green ratios than the single largest blue/green ratio measured for any cell in the CBU1387 group (Fig. S2). Results from the CyaA translocation assays indicated that CBU1198 and CBU2007 are robustly translocated C. burnetii T4BSS substrates. The absence of secretion and inefficient secretion of CBU1387 observed in CyaA and BlaM assays, respectively, suggested that CBU1387 is not productively exported by C. burnetii under the conditions tested.

FIG 2.

C. burnetii T4BSS-dependent translocation of BlaM-tagged proteins. THP-1 cells were left uninfected (UI) or infected for 24, 48, or 72 h with wild-type C. burnetii expressing the indicated proteins fused to BlaM or BlaM alone (BlaM). When delivered to the host cell cytoplasm, the BlaM enzymatic reporter converts CCF4-AM (green) to CCF4 (blue). (A) Representative fluorescence micrographs of THP-1 cells infected with C. burnetii strains expressing the indicated BlaM fusion proteins. (B) Quantitation of BlaM translocation. The histogram depicts the means and standard deviations of the ratio between CCF4 and CCF4-AM signals (Blue:Green) calculated for each cell (n > 500) within a sample group. Asterisks indicate values significantly greater than the cutoff for positive secretion (dotted line), which was set at 2.5 times the mean blue/green ratio measured for the negative-control cells infected with C. burnetii expressing BlaM alone (*, P < 0.05; ***, P < 0.001). Results are representative of three independent experiments.

Intracellular growth requirements for C. burnetii T4BSS substrates.

Of the 32 proteins analyzed in this study, only CBU0122, CBU0410, CBU1198, CBU1752, CBU1819, CBU1825, and CBU2007 were validated here or previously as bona fide C. burnetii T4BSS substrates. C. burnetii intracellular growth requirements for these T4BSS substrates were investigated using an isopropyl-β-d-1-thiogalactopyranoside (IPTG)-inducible CRISPR interference (CRISPRi) system to suppress transcription of T4BSS substrate genes (28). C. burnetii CRISPRi strains induced with IPTG expressed a catalytically inactive form of the Cas9 protein (dCas9) along with a single guide RNA (sgRNA) for targeting of dCas9 to template DNA. Binding of dCas9 to template DNA sterically hinders RNA polymerase binding and interferes with transcription of the targeted gene. For each of the T4BSS substrates, three C. burnetii CRISPRi strains were generated that expressed different sgRNA sequences for the targeted gene (Table S2.) For positive controls of growth impairment, a C. burnetii CRISPRi strain was included that expressed sgRNA targeting the essential T4BSS structural gene icmD (28), along with three different CRISPRi strains that targeted cvpB, an effector required for intracellular growth (23, 29, 30). A C. burnetii strain that expressed dCas9 but did not target sgRNA was used as a negative control (28). In C. burnetii strains treated with IPTG, the expression of CRISPRi constructs was confirmed by immunoblotting of dCas9 (Fig. S3A) and the suppression of targeted gene transcripts was confirmed by reverse transcription-quantitative PCR (RT-qPCR) of isolated RNA (Fig. S3B). Growth of CRISPRi strains was compared in the presence or absence of IPTG. As expected, treatment with IPTG did not impair the growth of any CRISPRi strains in defined acidified citrate cysteine medium (ACCM-D) broth culture (Fig. S3C). In THP-1 macrophages at 6 dpi, treatment with IPTG decreased replication by CRISPRi strains expressing targeting sgRNAs cbu0122_1 and _3, cbu0410_1, cbu1752_1 to _3, cbu1819_1, cbu1825_1 and _2, and cbu2007_1 and _2 as well as the positive-control sgRNAs icmD and cvpB_1 to _3 (Fig. 3). IPTG induction did not reduce replication of CRISPRi strains that expressed sgRNA targeting cbu1198 or lacked targeting sgRNA. The role of the T4BSS substrates in intracellular growth was also investigated by examining whether the C. burnetii CRISPRi strains produced replication-permissive CCVs in cells treated with IPTG (Fig. 4A). Vero cells infected with the CRISPR strains were cultured for 5 days in growth media treated with IPTG or left untreated, and then the area occupied by C. burnetii within each CCV (CCV colony) and the number of CCV colonies per cell was measured by fluorescence microscopy coupled with computational analysis (Fig. 4B; Fig. S4A and B). IPTG induction decreased the mean size of CCV colonies produced by CRISPRi constructs expressing the sgRNAs cbu0122_1 and _3, cbu0410_1, cbu1198_2, cbu1752_1 and _2, cbu1825_1, and cbu2007_1 to _3. C. burnetii CRISPRi strains expressing sgRNAs targeting cbu0410 or cbu1819 or no sgRNA did not exhibit significantly impaired CCV formation. None of the C. burnetii CRISPRi test strains exhibited an increase in the number of CCVs per cell when treated with IPTG. In the control groups, CRISPRi of cvpB_1 to _3 or icmD decreased the mean size of CCV colonies produced (Fig. 4B) and increased the percentage of cells with two or more colonies (Fig. S4B). These defects in bacterial replication and CCV fusion are consistent with phenotypes previously reported for C. burnetii cvpB or icmD mutant strains (9, 23, 28, 29). Decreased CCV size in Vero cells correlated with reduced bacterial replication in THP-1 cells for CRISPRi cbu0122_1 and _3, cbu1752_1 and _2, cbu1825_1, and cbu2007_1 and _2 strains. Collectively, these data indicated that CBU0122, CBU1752, CBU1825, and CBU2007 are C. burnetii T4BSS substrates that promote intracellular growth. Furthermore, our findings for CBU1752 concurred with another recent study showing that it promotes C. burnetii growth in HeLa cells (31).

FIG 3.

CRISPRi of several validated T4BSS substrates impairs C. burnetii replication. THP-1 macrophages were infected with C. burnetii strains engineered for IPTG-inducible CRISPRi with the indicated targeting sgRNAs or no target (no sgRNA). Bacterial genome equivalents (GEs) were measured on the day of infection (0 dpi) or 6 days postinfection in cells treated with IPTG (6 dpi IPTG+) or left untreated (6 dpi IPTG−). The plot depicts the means and standard deviations from three independent experiments conducted in duplicate. Asterisks indicate statistically significant differences between GEs from IPTG+ and IPTG− cells at 6 dpi (*, P < 0.05; ***, P < 0.001).

FIG 4.

CRISPRi of several validated T4BSS substrates impairs C. burnetii CCV formation. Vero cells left untreated (IPTG−) or treated with IPTG (IPTG+) were infected for 5 days with C. burnetii strains engineered for IPTG-inducible CRISPRi with the indicated targeting sgRNAs or no target (no sgRNA ctrl). (A) Representative fluorescence micrographs showing CCVs in Vero cells stained with anti-C. burnetii guinea pig serum and AF488 secondary antibody (green), Hoechst (blue), and Cell Mask deep red (magenta). (B) Quantitation of CCV colony size formed by C. burnetii CRISPRi strains. The change in CCV colony size (IPTG+/IPTG−) in response to induction of CRISPRi was calculated for infected cells (n > 350). The plot depicts the means and standard deviations of IPTG+/IPTG− measurements from three independent experiments. Asterisks indicate significant differences in CCV colony size between IPTG+ and IPTG− cells (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

Subcellular localization of T4BSS substrates.

C. burnetii T4BSS effector proteins often localize to specific subcellular compartments when ectopically expressed in mammalian cells (10, 21, 23, 29, 30). Indeed, the family of Coxiella vacuolar protein (Cvp) effectors localize to the membrane of late endosomes and the CCV (23). Here, we examined whether any of the T4BSS substrates found to promote Coxiella burnetii intracellular growth exhibited similar subcellular localization. HeLa cells expressing CBU0122, CBU0410, CBU1752, CBU1819, CBU1825, or CBU2007 with mCherry fused at the C or N terminus were fixed and stained with an antibody recognizing the late endosomal protein CD63 (Fig. 5; Fig. S5A). Of the proteins examined, only CBU0122 fused at its C terminus to mCherry (CBU0122-C-mCherry) localized with CD63 on late endosomes and the CCV. Indeed, CBU0122-C-mCherry localization with CD63 was observed in approximately 80% of the cells with the remaining cells exhibiting diffuse perinuclear localization. Intriguingly, when CBU0122 was expressed fused at its N terminus to mCherry (mCherry-N-CBU0122), only ~25% of the cells exhibited CD63 localization, while nearly 70% of the cells exhibited perinuclear tubular localization to structures that were labeled with the mitochondrial outer membrane protein Tom20. Cells expressing mCherry-N-CBU0122 that localized to CD63 or to Tom20 were observed within the same field of view, and the localization phenotypes did not correlate specifically to infected or uninfected cells (Fig. S5A). Because CBU0122 resembled a Cvp that traffics to the mitochondria, it was termed CvpM. Last, HeLa cells that expressed CBU1752, CBU1819, CBU1825, or CBU2007 fused at the N or C terminus to mCherry exhibited diffuse, cytosolic localization that resembled mCherry alone or nonspecific puncta that did not label with CD63 or Tom20 (Fig. S5B). CBU0410 fused at the N or C terminus to mCherry was not consistently expressed by HeLa cells.

FIG 5.

Subcellular localization of ectopically expressed CBU0122 fused at its N or C terminus to mCherry. (A) Representative fluorescence micrographs of HeLa cells at 5 dpi expressing CBU0122 fused at the C terminus (CBU0122-C-mCherry) or N terminus (mCherry-N-CBU0122) to mCherry (red). Cells were fixed and stained with antibodies against CD62 or Tom20 (green) and Hoechst (blue). Asterisks designate CCVs. Bar, 10 μm. (B) Quantitation of CBU0122-C-mCherry or mCherry-N-CBU0122 subcellular localization. The percentage of cells (n > 100) that exhibit CBU0122 localized to CD63 or Tom20 to labeled structures or diffuse perinuclear fluorescence localization was calculated. The plot depicts the means and standard deviations from three independent experiments.

DISCUSSION

Central to the pathogenesis of many Gram-negative bacteria is the secretion of effector proteins that modulate eukaryotic host systems to promote infection. The specialized translocation systems used for delivery distinguish bacterial effectors from other classes of exported virulence factors, such as toxins (32). C. burnetii delivers effectors with a T4BSS homologous to the L. pneumophila Dot/Icm secretion apparatus. Researchers have reported T4BSS-dependent secretion of ~150 C. burnetii proteins that are presumed to function as effectors (11). The majority of these C. burnetii proteins were designated T4BSS substrates based on heterologous secretion by L. pneumophila and have not been evaluated for secretion by C. burnetii. Here, C. burnetii genetic tools were used to investigate a group of proteins reported to be T4BSS substrates that are conserved among a diverse set of C. burnetii genomes but lack defined effector activities. Robust secretion of proteins by the C. burnetii T4BSS along with impaired intracellular growth by C. burnetii in response to transcriptional repression of coding genes indicated that CBU0122/CvpM, CBU1752, CBU1825, and CBU2007 could be T4BSS effectors. However, demonstrating that T4BSS substrates are functional effectors typically requires evidence of biochemical activity within a eukaryotic system (32). Distinct subcellular localization of CBU0122/CvpM to late endosomal membranes and/or the mitochondria is suggestive of activity in mammalian cells and supports the notion that CBU0122/CvpM is a bona fide effector. CBU1752, CBU1825, and CBU2007 were not designated bona fide effectors because they did not exhibit distinct localization in HeLa cells, whereas the T4BSS substrates CBU1198 and CBU1819 were not deemed effectors because they did not promote intracellular growth of C. burnetii or exhibit specific localization in HeLa cells. Intriguingly, the CyaA and BlaM assays performed in this study also indicated that C. burnetii does not secrete many proteins that were previously reported to be T4BSS substrates based on their secretion by the L. pneumophila T4BSS. CyaA and BlaM translocation reporter assays have been used to identify C. burnetii T4BSS substrates, because the detection of native proteins is extremely difficult (10, 11, 21). However, these reporter assays could result in the mischaracterization of proteins for a variety of reasons, including interference by the CyaA or BlaM tags, anomalies arising from the overexpression or heterologous expression of proteins, and/or misinterpretation of translocation reporter data.

Functional similarities between the C. burnetii and L. pneumophila T4BSSs are well established (11). The C. burnetii Nine Mile genome encodes homologs of 23 of the 26 L. pneumophila Dot/Icm proteins, with amino acid sequence identity ranging from 22 to 66% (19, 20, 27, 33). Previous studies reported numerous instances where T4BSS substrates recognized by L. pneumophila are also exported by the C. burnetii T4BSS (10, 11, 21). This is consistent with the observation that defects in L. pneumophila strains containing isogenic mutations in icmS, icmW, icmT, or dotB can be rescued by complementation with the cognate C. burnetii gene (19, 20). Conversely, although CBU1634 exhibits 31% sequence identity to L. pneumophila IcmQ, impaired growth of the L. pneumophila icmQ mutant is not rescued by expression of cbu1634 (19). Other functional differences between the C. burnetii and L. pneumophila T4BSSs are observed by comparison of the T4BSS coupling protein (T4CP) complex. Studies with L. pneumophila first showed that export of certain effectors requires a T4CP composed of DotL, DotM, DotN, IcmS, IcmW, and the adaptor protein LvgA (34–36). Examination of effector secretion by the L. pneumophila ΔicmS mutant showed that there are IcmS-dependent substrates that require IcmS for export and IcmS-independent substrates that do not (37, 38). C. burnetii exhibits a third class of IcmS-inhibited substrates that are exported by the C. burnetii ΔicmS mutant and wild-type L. pneumophila but not wild-type C. burnetii (17). We examined whether any of the proteins tested in this study were IcmS-inhibited T4BSS substrates, but none were exported by the C. burnetii ΔicmS mutant. A subcomplex of the T4CP composed of the chaperone pair IcmSW and adaptor protein LvgA contains binding sites that interact with effectors (34–36). C. burnetii does not encode a primary sequence homolog of LvgA, but a recent report showed that CBU1754 is a functionally analogous component of the C. burnetii T4CP (18). Impaired growth of a L. pneumophila ΔlvgA mutant strain is rescued by expression of cbu1754, but expression of lvgA does not rescue impaired growth of a C. burnetii cbu1754::transposon (Tn) mutant. Despite having severe intracellular growth defects, the C. burnetii cbu1754::Tn mutant remains competent for translocation of IcmS-dependent and IcmS-independent substrates (18). Secretion of IcmS-inhibited substrates by the C. burnetii cbu1754::Tn mutant was not reported, and the relevance to C. burnetii biology of this class of T4BSS substrates is unclear. We previously showed that CBU1825 is an IcmS-inhibited substrate (17) and show here that CBU1825 promotes C. burnetii replication and CCV formation. However, the lack of secretion by wild-type C. burnetii suggests that CBU1825 is not an effector. Additional investigation is needed to understand the potential significance of C. burnetii IcmS-inhibited T4BSS substrates.

A primary criterion for effectors is their translocation by the Dot/Icm T4BSS from the cytoplasm of C. burnetii into host cells. Of the 27 proteins analyzed that were previously identified as substrates of the L. pneumophila T4BSS, only two were efficiently translocated in this study by the C. burnetii T4BSS in CyaA translocation assays. Many of the proteins that exhibited no secretion in C. burnetii were designated T4BSS substrates in BlaM translocation assays where 5% or less of infected cells exhibited positive secretion (14, 39). This suggested that the BlaM reporter system might enable detection of proteins efficiently translocated into only a small fraction of cells. However, our BlaM assays confirmed inefficient translocation by C. burnetii of 14 proteins that were not secreted in the CyaA assays and demonstrated that CyaA and BlaM assays produce comparable results consistent with previous reports (18). CBU1387-BlaM was consistently exported by C. burnetii at levels greater than the negative control, but relative to the positive control and other validated T4BSS substrates, the amount of CBU1387-BlaM detected was unusually low. We concluded that CBU1387 along with CBU1634, CBU1754, and CBU1825 is not secreted by wild-type C. burnetii, in contrast with previous work showing that they are robustly secreted by the L. pneumophila T4BSS (14, 39, 40). The lack of secretion of CBU1634 and CBU1754 by C. burnetii is consistent with reports indicating that CBU1634 functions as IcmQ in the C. burnetii T4BSS (19, 20) and CBU1754 is a LvgA-like component of the C. burnetii T4CP (18). This indicates that certain components of the C. burnetii T4BSS apparatus can be aberrantly translocated when heterologously expressed in L. pneumophila. Similarly, the secretion of CBU1387 by L. pneumophila (40) or CBU1825 by the C. burnetii ΔicmS mutant (17) but not by wild-type C. burnetii could indicate that CBU1387 and CBU1825 are components of the T4BSS apparatus and also explain the C. burnetii intracellular growth requirements reported here for CBU1825 and previously for CBU1387 (31).

According to the CyaA and BlaM translocation assays in this study, CBU0937, CBU1425, CBU1594, and CBU1677 are not translocated by the C. burnetii T4BSS. These proteins were previously designated T4BSS substrates because cells infected with L. pneumophila expressing CBU0937, CBU1425, CBU1594, or CBU1677 fused to BlaM exhibited 2%, 1%, 5%, or 5% positive secretion, respectively (14, 39). Based on impaired growth of a C. burnetii cbu0937::Tn mutant strain, CBU0937 was named CirC (Coxiella effector for intracellular replication C) (39). More recently, CBU0937/CirC, CBU1425, CBU1594, and CBU1677 were detected by mass spectrometry in mitochondria isolated from cells infected with C. burnetii and designated MceB to MceE (mitochondrial Coxiella effector proteins B to E), respectively (41). BLASTP analysis indicates that CBU0937/CirC/MceB is a hypothetical protein with homologs present in many other species of Proteobacteria: CBU1425/MceC is a surface antigen with ~40% sequence identity to a conserved 17-kDa lipoprotein produced by Rickettsia, CBU1594/MceD contains a bacterial GatB/Yqey domain protein involved in tRNA metabolism, and CBU1677/MceE contains a hemerythrin HHE cation binding domain-containing protein that binds oxygen and is commonly found in anaerobic and microaerophilic organisms (42). In contrast, a survey of the literature reveals that the majority of validated C. burnetii T4BSS effectors are annotated hypothetical proteins unique to C. burnetii genomes (11, 14, 21, 23, 39). CBU0937/CirC/MceB and CBU1425/MceC were originally selected as candidate T4BSS substrates based on bacterial two-hybrid experiments with DotF, which interacts with several L. pneumophila effectors (14, 39, 43). More recently, cryo-electron microscopy (cryo-EM) projections have shown that DotF is an inner-membrane-associated component of the T4BSS core complex and spans the periplasm to interact with lipoproteins DotC and DotD in the outer membrane (13, 35, 44). CBU0937/CirC/MceB and CBU1425/MceC are present in C. burnetii outer membrane extracts and harbor signal (SPI) and lipoprotein signal (SPII) peptides, respectively, predicted to direct their transport to the periplasm by the Sec translocon (SignalP 6.0) (45). Thus, within the periplasm or outer membrane of C. burnetii interactions could occur between DotF and CBU0937/CirC/MceB or CBU1425/MceC. Efforts to decipher the apparent discrepancies between functions reported for CBU0937/CirC/MceB, CBU1425/MceC, CBU1594/MceD, and CBU1677/MceE illustrate the challenges associated with the authentication and functional characterization of C. burnetii T4BSS effectors.

MATERIALS AND METHODS

Bacterial strains and mammalian cell lines.

Bacterial strains and plasmids used in this study are listed in Table S1. C. burnetii Nine Mile phase II (RSA439, clone 4, NMII) genetic transformants were axenically cultured in ACCM-D as previously described (46). Bacterial stocks were maintained in ACCM-D containing 10% dimethyl sulfoxide and frozen at −80°C. Escherichia coli Stellar (TaKaRa) and E. coli PIR2 (Invitrogen) cells were used for recombinant DNA procedures and cultivated in Luria-Bertani broth or Terrific broth. HeLa (CCL-2; American Type Culture Collection [ATCC]) human cervical epithelial cells were cultured in Dulbecco modified Eagle medium (DMEM) (Thermo Fisher) containing 10% fetal bovine serum (FBS) (Thermo Fisher) at 37°C and 5% CO₂. THP-1 macrophages (TIB-202; ATCC) and African green monkey kidney (Vero) cells (CCL-81; ATCC) were maintained in RPMI 1640 medium containing 10% FBS at 37°C and 5% CO₂. C. burnetii CRISPRi strains were generated by electroporating C. burnetii Nine Mile phase II grown in ACCM-D with pB-CRISPRi and pTnS2::1169P-tnsABCD plasmids (Table S1) and selecting transformants in ACCM-D lacking proline, as previously described (9, 28).

Plasmids.

Plasmids used in this study are listed in Table S1. Oligonucleotides purchased from Integrated DNA Technologies are listed in Table S2. PCRs were carried out using AccuPrime Pfx or Taq (Thermo Fisher), purified using the NucleoSpin gel and PCR cleanup kit (TaKaRa, San Jose, CA), and cloned into plasmids using the In-Fusion HD cloning system (TaKaRa). Cloning reaction products were transformed into E. coli Stellar cells (TaKaRa) or PIR2 cells (Invitrogen) for plasmids with the R6K origin of replication. Oligonucleotides encoding targeting sgRNA were cloned into plasmid backbones with the NEBridge Golden Gate assembly kit (New England Biolabs [NEB]) as previously described (28).

Translocation assays.

Genes encoding reported T4BSS substrates were amplified by PCR from Nine Mile phase II genomic DNA using the oligonucleotide primers listed in Table S2. PCR products were inserted into the unique SalI site of pJB-CAT-CyaA or pJB-CAT-BlaM by In-Fusion (TaKaRa) to produce the translocation reporter plasmids listed in Table S1. Translocation assays were performed with THP-1 macrophage-like cells in 24-well plates (5 × 10⁵ cells per well) infected at a multiplicity of infection (MOI) of ~50 in RPMI medium plus 10% FBS with C. burnetii transformants expressing CyaA or BlaM fusion proteins, as previously described (17, 23). To assess CyaA translocation, the concentration of cAMP in cells lysed with 50 mM HCl and 0.1% Triton X-100 was measured using the cAMP enzyme immunoassay (GE Healthcare) (23). To assess BlaM translocation, infected cells in glass-bottom Sensoplates (Greiner) were loaded for 1 h with CCF4-AM (LiveBLAzer-FRET B/G loading kit; Invitrogen) in solution with 15 mM probenecid (Sigma) (17, 21). Cells were replenished with fresh medium and immediately imaged on a Nikon DS-Qi2 camera (Nikon Instruments Inc.) and X-Cite XLED1 excitation source UV/visible-light (UVV) light-emitting diode (LED) module (Excelitas Technologies Corp., Waltham, MA) with ET405/20x excitation and ET460/40m or ET525/36m emission filters (Chroma, Bellows Falls, VT). Image segmentation and signal quantitation were performed using CellProfiler as previously described (17). CyaA or BlaM fusion proteins were considered secreted if signals exceeded 2.5 times the signal produced by CyaA or BlaM alone, respectively, in accordance with previously published methods (9, 17, 18, 21).

Measurement of C. burnetii intracellular growth.

To measure intracellular replication, THP-1 cells (5 × 10⁵ per well) in 24-well plates (Corning) were infected with C. burnetii CRISPRi strains in RPMI with 10% FBS by centrifugation of plates at 500 × g for 30 min using an MOI of approximately 0.1. Following culture in the presence or absence of 0.5 mM IPTG for 3 h (0 dpi) or 6 dpi, cells were lysed in 0.5 mL of buffer containing 0.05% trypsin, 0.5 mM EDTA, and 20 mM Tris-HCl (pH 8) with shaking at 200 rpm. Lysates in microcentrifuge tubes were boiled with beads with vigorous shaking for 10 min, centrifuged briefly to pellet the beads, and diluted 1:10 for analysis of C. burnetii genomic equivalents by quantitative PCR (qPCR) as previously described using a probe and primers specific to groEL (8, 47). For quantitation of CCV morphology, Vero cells (6 × 10³ per well) in 96-well plates (Corning) were infected at an MOI of 100 with C. burnetii CRISPRi strains in RPMI with 10% FBS and centrifuged at 500 × g for 30 min. Infected-cell monolayers were cultured in growth media left untreated or treated with 0.5 mM IPTG for 5 days, then washed with phosphate-buffered saline (PBS), detached with trypsin, and resuspended in RPMI with 10% FBS (total volume, 150 μL). A portion of the trypsinized cell suspension (25 μL) was transferred to a fresh 96-well μ-Plate (Ibidi, Fitchburg, WI) containing 150 μL RPMI with 10% FBS with or without IPTG and incubated for 4 h to allow cell attachment. Cell monolayers were fixed overnight in PBS with 4% paraformaldehyde (PFA) at 4°C, washed three times in PBS, blocked for 10 min in PBS with 5% FBS, and stained with fluorescent markers as described below. CellProfiler was used to quantitate the area and number of CCV colonies formed by C. burnetii CRISPRi strains in Vero cells, as previously described (48).

Cell transfection, staining, and imaging.

HeLa cells infected with C. burnetii for 72 h and seeded on 24-well plastic-bottom μ-Plates (Ibidi) were transfected using X-tremeGENE 9 (Roche) with 1 mg of the pT-Rex-DEST30/N-mCherry or C-mCherry constructs and then incubated with fresh growth medium for 24 h. Cells were fixed for 30 min in PBS containing 4% paraformaldehyde, washed in PBS, and permeabilized with PBS containing 0.1% Triton X-100 and 5% FBS. Cell staining was conducted with the following antibodies and fluorescent dyes: Tom20 (Santa Cruz, 1776), CD63 (BD Biosciences, 556019), Alexa Fluor 488 goat anti-mouse (A11029) and goat anti-rabbit (A11034) immunoglobulin, Hoechst 33342 (62249), HCS CellMask deep red stain (H32721), and guinea pig anti-C. burnetii serum (9) (all from Thermo Fisher). Images were collected using a Nikon Eclipse Ti2 inverted epifluorescence microscope fitted with a Nikon DS-Qi2 camera (Nikon Instruments Inc.) and a Lumencor Sola pad excitation source (Lumencor, Beaverton, OR).

Statistical analysis.

Statistical analyses were conducted using Prism software (GraphPad Software, Inc.) to perform the unpaired Student's t test or one-way analysis of variance (ANOVA) using Tukey’s posttest.