Identification and Characterization of the Alternative σ²⁸ Factor in Treponema denticola

ABSTRACT

FliA (also known as σ²⁸), a member of the bacterial σ⁷⁰ family of transcription factors, directs RNA polymerase to flagellar late (class 3) promoters and initiates transcription. FliA has been studied in several bacteria, yet its role in spirochetes has not been established. In this report, we identify and functionally characterize a FliA homolog (TDE2683) in the oral spirochete Treponema denticola. Computational, genetic, and biochemical analyses demonstrated that TDE2683 has a structure similar to that of the σ²⁸ of Escherichia coli, binds to σ²⁸-dependent promoters, and can functionally replace the σ²⁸ of E. coli. However, unlike its counterparts from other bacteria, TDE2683 cannot be deleted, suggesting its essential role in the survival of T. denticola. In vitro site-directed mutagenesis revealed that E221 and V231, two conserved residues in the σ₄ region of σ²⁸, are indispensable for the binding activity of TDE2683 to the σ²⁸-dependent promoter. We then mutated these two residues in T. denticola and found that the mutations impair the expression of flagellin and chemotaxis genes and bacterial motility as well. Cryo-electron tomography analysis further revealed that the mutations disrupt the flagellar symmetry (i.e., number and placement) of T. denticola. Collectively, these results indicate that TDE2683 is a σ²⁸ transcription factor that regulates the class 3 gene expression and controls the flagellar symmetry of T. denticola. To the best of our knowledge, this is the first report establishing the functionality of FliA in spirochetes.

IMPORTANCE Spirochetes are a group of medically important but understudied bacteria. One of the unique aspects of spirochetes is that they have periplasmic flagella (PF, also known as endoflagella) which give rise to their unique spiral shape and distinct swimming behaviors and play a critical role in the pathophysiology of spirochetes. PF are structurally similar to external flagella, but the underpinning mechanism that regulates PF biosynthesis and assembly remains largely unknown. By using the oral spirochete Treponema denticola as a model, this report provides several lines of evidence that FliA, a σ²⁸ transcriptional factor, regulates the late flagellin gene (class 3) expression, PF assembly, and flagellar symmetry as well, which provides insights into flagellar regulation and opens an avenue to investigate the role of σ²⁸ in spirochetes.

INTRODUCTION

Multisubunit (α₂ββ′ω) DNA-dependent RNA polymerase (RNAP) is the core enzyme for transcription initiation, the first step of gene expression in cells (1, 2). The core enzyme is evolutionarily conserved; however, bacterial RNAP is unable to specifically initiate transcription alone and thus depends on several sigma factors (σs) for transcription initiation. Mechanistically, σs recruit the RNAP complex to a promoter by recognizing specific consensus sequences and then form a holoenzyme to initiate transcription by isomerizing the RNAP-σ-promoter complex from a closed to an open state (1, 3–5). Based on their distinct structures and mechanisms, bacterial σs are classified into two main families: σ⁵⁴- and σ⁷⁰-type factors. The latter is further divided into four phylogenetically and structurally distinct groups (1, 6). Group 1 σs contain domains σ₁, σ₂, σ₃, and σ₄; group 2 σs are similar to group 1 but they lack the σ_1.1 subdomain; group 3 σs harbor domains σ₂, σ₃, and σ₄; and group 4, or extracytoplasmic function (ECF), σs have only domains σ₂ and σ₄. The domains σ₂, σ₃, and σ₄, but not σ₁, all possess a helix-turn-helix (HTH) DNA-binding module that recognizes different consensus sequences in a promoter: σ₄, −35 motif; σ₃, extended −10 motif; and σ₂, −10 and discriminator motifs. Group 1 σs (e.g., σ⁷⁰) are also known as primary or housekeeping σs because they are essential for bacterial growth. In contrast, the other three groups are not essential and thus are often referred to as alternative σs (1, 6).

FliA (also known as σ²⁸ or σ^F in Escherichia coli and Salmonella enterica serovar Typhimurium and σ^D in Bacillus subtilis) is the most widely distributed alternative σ factor that targets RNAP to flagellar late (class 3) promoters in a wide range of motile bacteria (1, 2, 6). It has been functionally and structurally characterized in the paradigm models of E. coli and S. Typhimurium (7, 8). FliA lacks the σ₁ domain but aligns well with the σ₂, σ₃, and σ₄ domains of group 1 σs (8, 9). FliA targets the RNAP complex to specific promoters and forms a holoenzyme to initiate the transcription of class 3 genes, such as fliC and tar (8, 10). Therefore, FliA is often known as a flagellum-specific σ factor. The activity of FliA is regulated in response to flagellar assembly by FlgM, an anti-σ²⁸ factor, which sequesters FliA by forming a 1:1 complex in the cytoplasm. Upon completion of the flagellar hook, the hook-basal body complex (HBB) provides an export route for FlgM. As FlgM exits the cell, FliA is free to act (11–13). A similar regulatory scheme is also found in other motile bacteria, such as Campylobacter jejuni (14), Helicobacter pylori (15, 16), Pseudomonas aeruginosa (17), Vibrio cholerae (18), and B. subtilis (2, 19). In these bacteria, deletions of fliA often lead to mutants that are nonflagellated and nonmotile, highlighting its essential role in bacterial flagellar synthesis and motility (16, 18, 20–22).

Spirochetes are a group of bacteria that can be readily recognized by their flat-wave or coiled-cell morphology and distinct form of corkscrew-like motility (23–25). They are responsible for several human diseases, including Lyme disease (Borrelia burgdorferi), syphilis (Treponema pallidum), leptospirosis (Leptospira interrogans), and periodontal disease (Treponema denticola) (26–29). Unlike external flagellates, spirochetes swim by means of rotating two bundles of periplasmic flagella (PF) that reside between the outer membrane and cell cylinder (23, 30). The number and length of PF vary from species to species. In general, spirochetal PF are structurally similar to the flagella of other bacteria, as they consist of a basal body-motor complex, a hook, and a filament (31). However, accumulating evidence suggests that spirochetes have evolved unique and complex mechanisms to regulate their flagellar gene expression and assembly. For example, B. burgdorferi has no homologs of FliA or FlgM and primarily depends on the housekeeping σ⁷⁰ to control its flagellar and chemotaxis gene expression (23). Most spirochetes have multiple flagellin genes (23, 30). For instance, the flagellar filaments of T. denticola, T. pallidum, and Brachyspira hyodysenteriae are composed of at least one sheath protein (FlaA) and three core FlaB proteins (32–36). Of note, FlaBs are homologs of FliC. FlaA proteins are unique to spirochetes and do not share sequence similarity to bacterial flagellin proteins. In addition, FlaA proteins are not exported through the flagellar type III secretion system (fT3SS) as other flagellar proteins are. Instead, they are likely exported to the periplasmic space by the type II secretion pathway (30, 34). Interestingly, our previous studies indicate that the genes encoding these proteins are regulated by different promoters (32, 35). For instance, a conserved σ²⁸-dependent promoter is mapped upstream of flaB2, and a σ⁷⁰-dependent promoter is mapped upstream of flaA and flaB1 of T. denticola (32). A similar scenario is also found in B. hyodysenteriae and T. pallidum (37–39). However, while σ²⁸-dependent promoters have been reported in several spirochetes (35, 40, 41), σ²⁸ transcription factors have not yet been functionally characterized. In this study, TDE2683, a homolog of FliA, was first identified in T. denticola, a keystone pathogen of human periodontitis (29), and then its function as a σ²⁸ was established using an approach combining bioinformatics, genetics, biochemistry, and cryo-electron tomography (cryo-ET). This report sets a benchmark for investigating the role of σ²⁸ in spirochetes.

RESULTS

Classification of σs in T. denticola.

The genome of T. denticola ATCC 35405 encodes at least seven σs, including TDE0070 (275 amino acids [aa]; σ⁷⁰/RpoD), TDE0091 (180 aa; σ²⁴/RpoE), TDE0937 (286 aa; σ⁷⁰/RpoD), TDE1346 (619 aa; σ⁷⁰/RpoD), TDE2320 (213 aa; σ²⁴), TDE2404 (463 aa; σ⁵⁴/RpoN), and TDE2683 (262 aa; σ²⁷/WhiG) (42). BLAST search and sequence comparisons revealed that, among the three putative σ⁷⁰ factors, TDE1346 is a bona fide σ⁷⁰ because it contains all four functional domains (σ₁, σ₂, σ₃, and σ₄) in the group 1 σs (Fig. 1; also, see Fig. S3 in the supplemental material). TDE0070 and TDE0937 harbor σ₂, σ₃, σ₄, and one part of σ₁ (σ_1.2). A BLAST search showed that TDE0937 has 30% sequence identity and 53% similarity to the E. coli σ³² (RpoH), which regulates the expression of heat shock protein (43). TDE0070 is a homolog of RpoS; e.g., it has 35% shared sequence identity and 59% similarity to the RpoS of B. burgdorferi. RpoS is an alternative σ factor that regulates general stress responses such as oxidation, starvation, and change of host (44, 45). For instance, B. burgdorferi RpoS is a gatekeeper that regulates the expression of genes for stress responses and tick-to-mammal transmission (46). Therefore, TDE0070 likely functions as an alternative σ to control the stress response of T. denticola. TDE0091 contains only the σ₂ and σ₄ domains and thus likely belongs to the group 4 ECF σs (Fig. 1). Both TDE2320 and TDE2683 belong to the group 3 σs, which have σ₂, σ₃, and σ₄ domains; however, one part of σ₄ (σ_4.2) is absent in TDE2320 (Fig. 1 and Fig. S3). In contrast, TDE2683 contains all the domains that are possessed by the group 3 σs, such as σ²⁸. Therefore, between these two σs, TDE2683 more likely functions as σ²⁸.

FIG 1.

Schematic illustration of σs in T. denticola. These factors are annotated in the genome of T. denticola (42). Individual domains in these σ factors were identified by BLAST (Fig. S3) and SMART searches and then aligned to their counterparts (i.e., σ⁷⁰, σ²⁸, and σ⁵⁴) from E. coli and other bacteria (1, 4–6, 8, 10, 90, 91).

Identification of a FliA homolog in T. denticola.

TDE2683 consists of 262 aa with a predicted molecular weight (MW) of 29.95 kDa. It resides in a gene cluster composed of eight open reading frames (ORFs): TDE2687 (HisC), TDE2686 (FlhF), TDE2685 (FlhG), TDE2684 (hypothetical protein [HP]), TDE2683 (WhiG/σ²⁷), TDE2682 (FapA), TDE2681 (HP), and TDE2680 (HP) (Fig. 2). HisC belongs to the family of histidinol phosphate aminotransferases (47); FlhF and FlhG are two small GTPases that regulate flagellar number and placement in polar flagellates such as C. jejuni and B. burgdorferi (48, 49). A PSI-BLAST search revealed that the N terminus of TDE2684 (222 aa) shares 35% (21 of 60 aa) sequence identity with B. burgdorferi Hfq (50), a global regulatory RNA-binding protein, and TDE2682 is a homolog of FapA (flagellar assembly protein A), which controls the flagellar localization of Vibrio vulnificus (51, 52). Co-reverse transcription-PCR (RT-PCR) analysis showed that TDE2683, four genes upstream, and three genes downstream are cotranscribed, but not TDE2679 and TDE2689, two genes in the opposite direction (Fig. 2A and B). We then mapped the transcription start site (TSS) using 5′ rapid amplification of cDNA ends (RACE) and found that the −10 and −35 regions contain a well-conserved σ⁷⁰-dependent promoter (TTGCCA-N₁₇-TATAAT-N₆-A) which is 23 bp from the start codon of TDE2687 (Fig. 2), suggesting that this gene cluster is controlled by σ⁷⁰.

FIG 2.

TDE2683 is located at a large gene cluster regulated by a σ⁷⁰ promoter. (A) Diagram showing the genes adjacent to TDE2683. Arrows represent the relative positions and orientations of RT-PCR primers that span the intergenic regions between individual ORFs. (B) RT-PCR analysis. For each pair of primers, chromosomal DNA was used as a positive control. The numbers below the primers are predicted sizes of RT-PCR and PCR products. (C) 5′ RACE analysis. The arrow shows the sequencing direction, and an asterisk indicates the transcriptional start site. (D) Sequence comparison between the E. coli σ⁷⁰ promoter and the promoter sequence identified upstream of TDE2687 (designated P_TDE2687).

TDE2683 is annotated as σ²⁷/WhiG in the genome of T. denticola (42). Recent studies showed that WhiG, an ortholog of bacterial FliA or σ²⁸, regulates multicellular differentiation (e.g., sporulation) of Streptomyces (53, 54). In addition, domain composition analysis showed that TDE2683 belongs to the group 3 σs (Fig. 1 and Fig. S3). Therefore, we reasoned that TDE2683 is a FliA homolog. To test this hypothesis, we first searched the NCBI protein database using TDE2683 as a query and found that it belongs to the family of WhiG/FliA (TIG02479). We then carried out multisequence alignment (Fig. 3A), including FliA proteins of E. coli and S. Typhimurium and SigD (an ortholog of FliA) of Bacillus subtilis (19), and found that TDE2683 shares 39 to 44% sequence identity with these three proteins. Previous phylogenetic and structural studies showed that the family of σ²⁸ factors contains three invariable residues, i.e., D81, R84, and R91 in the FliA of E. coli (FliA_Ec), which form a binding site for −10 regions (GCCGATAA) in σ²⁸-dependent promoters, such as fliC and tar promoters in E. coli (8, 10, 55, 56). Like FliA_Ec, TDE2683 also has these three invariable residues (D95, R98, and R105). The structure of FliA_Ec and S. Typhimurium FliA (FliA_St) was solved (8). Using these two proteins as a structural template, we conducted homolog modeling and found that TDE2683 has a structure similar to that of FliA_Ec and FliA_St (Fig. 3B); i.e., they all possess 10 α-helixes (H₁ to H₁₀) that form three functional domains, σ₂ (H₁ to H₃), σ₃ (H₄ to H₆), and σ₄ (H₇ to H₁₀) (Fig. 3B), linked by variable disordered loops (8). Together, sequence and structural comparison analyses indicate that TDE2683 is a FliA homolog, and it is referred to here as FliA_Td.

FIG 3.

TDE2683 is a homolog of FliA. (A) Sequence alignment of FliA homologs. TDE2683 (WP_002667112) is aligned with E. coli FliA (NP_416432), S. Typhimurium FliA (NP_460909), and B. subtilis SigD (WP_187704716). Only partial alignment is shown. Asterisks represent three invariable residues in the family of FliA. The residues in red are to be mutated. The alignment was analyzed using Clustal Omega. (B) Homology modeling shows that TDE2683 (FliA_Td) has structural topology similar to that of the FliA proteins (σ²⁸) of E. coli and S. Typhimurium. The figure was generated by the Swiss-Model server using FliA_Ec (PDB ID 1SC5) (8) as a template.

FliA_Td can substitute for the function of FliA_Ec.

σ²⁸ factors are evolutionarily conserved and functionally interchangeable across different bacterial species (2, 18, 57). For example, SigD, a σ²⁸ ortholog in B. subtilis, can substitute for the function of FliA_Ec (19). To determine if the same scenario occurs with FliA_Td, a heterologous gene replacement was conducted. To this end, we replaced the entire fliA_Ec in frame with fliA_Td, and the resultant strain was designated the cfliA_Td strain. As a control, we constructed a fliA_Ec deletion (ΔfliA_Ec) mutant. Both ΔfliA_Ec and cfliA_Td were confirmed by PCR and immunoblotting using a specific antibody against FliA_Td. Immunoblotting analyses showed that fliA_Td was successfully expressed in the cfliA_Td strain (Fig. 4A). Flagellar staining and swimming plate assays showed that the ΔfliA_Ec mutant was unable to assemble flagellar filaments and failed to swim out in 0.4% soft-agar plates. In contrast, the cfliA_Td strain was still motile and able to assemble flagella (Fig. 4B and C), suggesting that FliA_Td can substitute, at least in part, for the function of FliA_Ec in E. coli. This study aligns with the above-described sequence and structural analyses and provides the first line of experimental evidence that FliA_Td has a function similar to that of FliA_EC.

FIG 4.

FliA_Td substitutes for the function of FliA_Ec in E. coli. (A) Detection of FliA_Td in E. coli by immunoblotting analysis. For this experiment, equivalent amounts (~10 μg) of whole-cell lysates of three E. coli strains were analyzed by SDS-PAGE and then probed with specific antibodies against E. coli GroEL (αGroEL) and T. denticola FliA (αFliA_Td). GroEL was used as a loading control. WT, E. coli DH5α; ΔfliA_Ec, fliA deletion mutant of E. coli; cfliA_Td, engineered E. coli strain in which the entire fliA_Ec gene was replaced in frame with fliA_Td. (B) Visualization of E. coli flagella. E. coli cells were first stained with Ryu as previously described (82) and then visualized under dark-field illumination at ×100 magnification using a Zeiss Axiostar Plus microscope. (C) Swimming plate assay. This assay was carried out on swimming plates containing 1% tryptone, 1% NaCl, and 0.4% Bacto agar, as previously documented (87). The plates were incubated at 37°C overnight.

FliA_Td binds to σ²⁸-dependent promoters.

σ²⁸ recognizes a promoter through interactions with specific consensus sequences at −10 and −35 regions (1, 58, 59). P_fliC, a promoter that initiates the expression of fliC in E. coli and S. Typhimurium, is a well-characterized σ²⁸-dependent promoter (60, 61). We recently found that the flaB2 gene of T. denticola is controlled by the σ²⁸-dependent promoter P_flaB2 (32). Both promoters contain a conserved σ²⁸ binding site at their −10 (CGGATAA) and −35 (TA/TAA) regions. Using these two promoters as surrogates, we carried out electrophoretic mobility shift assays (EMSA) to determine if FliA_Td binds to σ²⁸-dependent promoters. EMSA showed that the recombinant FliA_Td protein (rFliA_Td) bound to biotin-labeled P_fliC and P_flaB2 in a dose-dependent manner, which was competed out upon addition of unlabeled DNA probes (Fig. 5A and B). To further confirm that the observed binding is specific, we mutated E221 and V231, two conserved residues in the σ₄ domain of σ²⁸ factors (55, 56), by using site-directed mutagenesis and found that two mutated recombinant proteins (E221D and V231E) were unable to bind to P_fliC and P_flaB2 (Fig. 5C and D). This study provides the second line of experimental evidence that FliA_Td binds to the σ²⁸-dependent promoters.

FIG 5.

EMSA shows that FliA_Td binds to two σ²⁸-dependent promoters. (A) EMSA using E. coli fliC promoter (P_fliC) as a DNA probe. For this assay, biotin-labeled P_fliC (Bio-P_fliC) was incubated with increasing amounts of wild-type FliA_Td recombinant protein (rFliA_Td) and titrated with increasing concentrations of unlabeled P_fliC (Unl-P_fliC). (B to D) EMSA using T. denticola flaB2 promoter (P_flaB2) as a DNA probe. Biotin-labeled P_flaB2 (Bio-P_flaB2) was incubated with increasing amounts of rFliA_Td (B) or two mutated recombinant proteins, rFliA^E221D (C) and rFliA^V231E (D), and competed out with increasing concentrations of unlabeled P_flaB2 (Unl-P_flaB2).

FliA_Td controls the expression of flagellar filament and chemotaxis genes in T. denticola.

To investigate the role of FliA_Td, we sought to delete this gene in T. denticola via allelic exchange. Despite different constructs (deletions versus insertions), antibiotic resistant markers (erythromycin versus gentamicin), and transformations (electroporation versus heat shock) attempted, we were unable to delete fliA_Td, suggesting that it is essential for T. denticola. To circumvent this issue, we mutated E221 and V231 of FliA_Td in T. denticola, because the above-described EMSA study shows that these two residues are essential for its binding activity to the σ²⁸-dependnet promoter. To this end, we replaced wild-type fliA_Td in frame with two mutated genes (E221D and V231E) along with an erythromycin cassette (ermB) (62). Erythromycin-resistant colonies that appeared on the plates were first screened by PCR for the presence of ermB. Positive colonies were then subjected to PCR and DNA sequencing to confirm two site-directed mutations in fliA_Td. The results showed that E221 and V231 were mutated as designed, and the two obtained mutants were designated FliA_Td^E221D and FliA_Td^V231E mutants. To investigate the role of FliA_Td, we first carried out swimming plate assays and found that the swimming rings (n = 12 plates) formed by the FliA_Td^E221D (11.67 ± 0.61 mm) and FliA_Td^V231E (14.37 ± 0.68 mm) mutants are significantly (P < 0.05) smaller than those of the wild type (19.42 ± 0.25 mm) but larger than those of the Δtap1 strain (4.38 ± 0.13 mm), a previously constructed nonmotile mutant (63) (Fig. 6A), indicating that the motility of these two mutants is impaired.

FIG 6.

Characterizations of two fliA_Td site-directed mutants of T. denticola. (A) Swimming plate assay of WT, FliA_Td^E221D, and FliA_Td^V231E strains. This assay was carried out on 0.35% agarose plates containing TYGVS medium diluted 1:5 with PBS. The plates were incubated anaerobically at 37°C for 5 days. A nonmotile Δtap1 mutant was used as a control to determine initial inoculum sizes. (B) Immunoblotting analysis of WT, FliA_Td^E221D, and FliA_Td^V231E strains. Equivalent amounts of whole-cell lysates were analyzed by SDS-PAGE and then probed with specific antibodies to DnaK, FlaA, and FlaBs. DnaK was used as a loading control. (C to F) qRT-PCR analysis of WT, FliA_Td^E221D, and FliA_Td^V231E strains. For this experiment, the levels of four flagellar filament genes (flaA, flaB1, flaB2, and flaB3), two chemotaxis genes (cheA and cheY), two flagellar genes (flgE and fliG), TDE2682 (fapA), and TDE1346 (σ⁷⁰) were measured by qRT-PCR, as previously described (33). The dnaK gene transcript was used as an internal control to normalize the readouts of qRT-PCR. The results are expressed as the level of individual gene transcripts in two mutants relative to that in WT. *, P < 0.05.

We next assessed the role of FliA_Td in the expression of flagellin genes in T. denticola using immunoblotting and quantitative RT-PCR (qRT-PCR) analyses. Notably, unlike E. coli and S. Typhimurium, which have a single flagellin protein, T. denticola has multiple flagellin proteins, including FlaB1, FlaB2, and FlaB3, and a sheath protein, FlaA (33, 42, 64). Immunoblotting analysis revealed that the levels of FlaA and three FlaBs in both the FliA_Td^E221D and FliA_Td^V231E mutants were significantly decreased (Fig. 6B). Densitometry analysis showed that compared to the wild type, FlaA and three FlaB proteins in these two mutants decreased ~40% and ~90%, respectively. qRT-PCR analysis further demonstrated that the reduction of FlaA and three FlaBs in these two mutants mainly occurred at the transcriptional level (Fig. 6C). In addition to the four flagellar filament genes, qRT-PCR analysis showed that the transcription of cheA and cheY, two chemotaxis genes, in FliA_Td^E221D and FliA_Td^V231E was also decreased (Fig. 6B). We also measured the transcription of flgE (hook) and fliG (C ring), two genes that belong to class 2 in the flagellar transcriptional regulatory cascade (13), and found that their transcripts had no significant changes in the two mutants (Fig. 6D). These results indicate that FliA_Td controls the expression of class 3 flagellin and chemotaxis genes in T. denticola. In addition, qRT-PCR analysis demonstrated that the transcription of TDE2682 (fapA), a gene immediately downstream of fliA_Td, was not affected (Fig. 6E), indicating that the insertion of ermB has no polar effect on its downstream gene expression in these two mutants.

FliA_Td controls the flagellar symmetry of T. denticola.

The above-described immunoblotting analysis showed that the levels of four flagellar filament proteins are significantly impaired in both FliA_Td^E221D and FliA_Td^V231E. To determine if this impairment affects flagellar filament assembly, we conducted in situ whole-cell cryo-ET analysis. Cryo-ET revealed that two mutants are still able to assemble PF, even though the levels of the four flagellar filament proteins were substantially attenuated (Fig. 6B). Interestingly, the number and placement of PF in these two mutants became abnormal (Fig. 7 and Table 1). The wild-type cells uniformly have bipolar flagella (two PF at both cell poles), which form a ribbon-like structure wrapping around the cell cylinder (Fig. 7A). In contrast, two mutants were diverse in terms of flagellar number and placement (Fig. 7B to F); e.g., some mutant cells had multiple PF at one end (lophotrichous) (Fig. 7D) or a single flagellum at both ends (amphitrichous) (Fig. 7B). Some mutant cells even had lateral rather than polar flagella (Fig. 7B). In addition, the PF in these two mutants became disarrayed and were unable to form a ribbon-like structure (Fig. 7). Moreover, cells of the FliA_Td^E221D (9.7 ± 1.2 μm, n = 3 cells) and FliA_Td^V231E (10.1 ± 4.1 μm, n = 4 cells) strains appeared to be longer than wild-type cells (7.1 ± 0.30 μm, n = 3 cells). These results indicate that mutations of FliA_Td somehow disrupt the symmetric arrangement of PF and alter the cell length of T. denticola.

FIG 7.

Whole-cell cryo-ET analysis of T. denticola wild type and two mutants. (A) Projection image of a WT cell; (B to F) projection images of FliA_Td^E221D and FliA_Td^V231E mutant cells. PF originating from one end are colored in yellow, and those from the opposite end are in orange. The boxed images in panel E and F show the detailed structure of PF and flagellar protrusions. PF, periplasmic flagella; OM, outer membrane; IM, inner membrane. Each panel was generated by combining multiple high-resolution cryo-ET images; the image contrast is different among these data.

TABLE 1.

Whole-cell cryo-ET analysis of T. denticola wild type and two site-directed mutants

Strain and cell	Cell length (μm)	Flagellar length (μm)	Flagellar no.a
WT
1	6.9	3.4	2
		3.5
		4.6	2
		4.6
2	7.3	6.1	2
		4.8
		2.1	2
		2.1
3	7.0	5.0	2
		5.5
		2.5	2
		3.5
Avg	7.1	4.0
FliAE221D mutant
1	11.1	4.0	2
		4.9
		—b	—
2	9.2	6.3	4
		6.7
		6.3
		5.6
3	8.8	3.5	1
		4.7	1
Avg	9.7	3.5
FliAV231E mutant
1	13.0	5.0	1
		2.9	1
2	7.2	2.9	2
		3.3
		3.3	2
		3.2
3	10.2	2.2	2
		7.2
		—	—
4	10.7	3.1	1
		—	—
Avg	10.3	2.1

^aNumber of flagellar filaments visualized at the cell poles.

^b—, indicates that there is no PF detected by cryo-EM.

DISCUSSION

Due to high sequence similarity, σ²⁸ factors such as FliA are often misannotated as other members in the family of σ⁷⁰ factors (1, 6). This probably explains why TDE2683 is annotated as σ²⁷/WhiG (42). By combining bioinformatics, heterologous complementation, biochemistry, site-directed mutagenesis, and cryo-ET, this report provides several lines of coherent evidence that TDE2683 functions as a flagellum-specific σ²⁸ that regulates the expression of flagellar class 3 genes in T. denticola. To the best of our knowledge, this is the first report to establish the role of FliA as a flagellum-specific σ²⁸ in spirochetes. In the flagellar transcription hierarchy of E. coli and S. Typhimurium, FliA and FlgM act collaboratively to regulate flagellar late-gene transcription (65, 66). Identification of FliA_Td indicates that a similar regulatory mechanism exists in T. denticola, yet there has been no FlgM annotated in its genome. We searched the genome of T. denticola for FlgM homologs. A BLASTP search did not find any, but a DELTA-BLAST search revealed that TDE0201, a small hypothetical protein of 93 aa, has 17% shared sequence identity to E. coli FlgM. We are currently conducting biochemical and loss-of-function studies to determine if TDE0201 functions as FlgM. We then extended our search for the presence of FliA and FlgM to other spirochetes and found their homologs in most spirochetes, including T. pallidum, L. interrogans, and B. hyodysenteriae, suggesting that these spirochetes may also employ FliA and FlgM to regulate their flagellar biosynthesis. In line with our previous reports, there are neither FliA nor FlgM homologs identified in the genus Borrelia, supporting the idea that this group of spirochetes is an outlier that has evolved unique mechanisms to regulate their flagellar synthesis and assembly (23, 67, 68).

In some bacterial pathogens, FliA is required for virulence and in vivo fitness (69–73). However, to the best of our knowledge, there has been no prior report that FliA, an alternative σ factor, is essential for bacterial growth in vitro. It was therefore totally unexpected that we were unable to delete fliA_Td via allelic exchange despite the fact that different methods were attempted. In contrast, we were able to delete TDE2682, a gene immediately downstream of fliA_Td. In addition, we were able to replace fliA_Td in frame with two site-directed mutants, fliA_Td^E221D and fliA_Td^V231E, via allelic exchange. The success of these two experiments ruled out the possibility that the inability to delete fliA_Td is due to a polar effect on downstream gene expression caused by insertions of antibiotic cassettes. Moreover, the family of σ²⁸ factors contains three invariable residues, i.e., D81, R84, and R91 in FliA_Ec and D95, R98, and R105 in FliA_Td (8, 9). Previous mutagenesis and structural analyses demonstrate that these three residues form a binding site for −10 regions (GCCGATAA) in σ²⁸-dependent promoters, such as fliC and tar promoters in E. coli (55). Thus, they are absolutely required for the function of σ²⁸ factors. We also attempted to mutate these three residues in T. denticola; however, this attempt failed. By using the same method, we successfully mutated E221 and V231, two residues in the σ₄ domain that binds to −35 regions in σ²⁸-dependent promoters. Previous studies showed that substitutions of these two residues partly impair the function of FliA_Ec and FliA_St (55, 56). A similar scenario might occur in FliA_Td, which allows us to mutate these two residues in T. denticola. Taking these results together, we posit that fliA_Td is essential for T. denticola growth. If so, how does this occur? T. denticola has two long PF that arise from the cell poles and extend along the cell cylinder. Due to this unique structure, T. denticola must couple its flagellar biosynthesis and elongation to cell division. It is possible that FliA_Td also functions as a regulator to synchronize these two processes. If this is the case, deletion of fliA_Td may perturb this synchronization, in turn arresting cell division and/or leading to cell lysis. In line with this speculation, we found that two site-directed mutants of fliA_Td grew slightly more slowly than the wild type (Fig. S4) and had disarrayed and misplaced PF, some of which even protruded from the cell membrane (Fig. 7F). Flagellar protrusions may disrupt the cell membrane and lead to cell lysis. We are currently attempting to construct a conditional knockout of fliA_Td using a thiamine-inducible promoter we previously identified (74). If this undertaking is successful, this mutant will allow us to decipher the role of FliA_Td in regulation of flagellar biosynthesis and cell division of T. denticola.

As mentioned in the introduction, the flagellar filaments of T. denticola are composed of at least four proteins: FlaA, FlaB1, FlaB2, and FlaB3. Among these four proteins, the gene encoding FlaB2 is regulated by a σ²⁸-dependent promoter (32). Consistent with this result, we found that the expression level of flaB2 is impaired in the FliA_Td^E221D and FliA_Td^V231E mutants. Surprisingly, immunoblotting and qRT-PCR analyses also showed that the levels of flaA, flaB1, and flaB3, three genes controlled by a σ⁷⁰-dependent promoter, are also attenuated in these two mutants (Fig. 6B and C). We repeated these two experiments using three sets of samples that were harvested at different growth phases of T. denticola and obtained similar results. FliA_Td^E221D and FliA_Td^V231E were constructed by replacing fliA_Td in frame with two mutated genes along with ermB, which is used as a selection marker. It is therefore possible that insertion of ermB has a polar effect on downstream gene expression that, in turn, affects the expression of three flagellin genes. To rule out this possibility, we measured the expression of TDE2682, a gene immediately downstream of fliA_Td, by qRT-PCR and found that it remained unchanged in two mutants (Fig. 6E). We also measured the expression of TDE1346, a putative σ⁷⁰ factor, by qRT-PCR and found that it was not affected either (Fig. 6F). WhiG is a sporulation-specific σ²⁸ of Streptomyces (53). Two recent reports showed that the signaling molecule cyclic di-GMP (c-di-GMP) controls the differentiation of hyphae into spores by arming RsiG, an anti-σ of WhiG (53, 54). FliA_Td and WhiG share 49% sequence identity. In addition, T. denticola has a complex c-di-GMP signaling pathway, and previous studies by us and others showed that c-di-GMP plays a critical role in the motility and virulence of T. denticola (42, 75–77). Thus, it is possible that there is a cross talk between the c-di-GMP signaling pathway and FliA_Td. If this is true, mutations in FliA_Td may perturb the signaling pathway of c-di-GMP that in turn affects the expression of flaA, flaB1, and flaB3 either directly or indirectly. Alternatively, FliA_Td may control the expression of these three genes by modulating other signaling pathways, such as bacterial two-component systems (TCS). For instance, C. jejuni and H. pylori have no FlhC and FlhD, two master regulators in the flagellar regulatory transcription cascade, and have evolved to control their flagellar gene expression using TCS (22). Like C. jejuni and H. pylori, FlhC and FlhD are absent in T. denticola (42). The genome of T. denticola encodes at least six pairs of TCS, including six histidine kinases and six response regulators, and their roles remain largely unknown. It is possible that there is an interplay between FliA_Td and some of these TCS signaling pathways. We are currently investigating these possibilities.

MATERIALS AND METHODS

Bacterial strains, culture conditions, and oligonucleotide primers.

T. denticola ATCC 35405 (wild type [WT]) was used in this study (42). Cells were grown in tryptone-yeast extract-gelatin-volatile fatty acids-serum (TYGVS) medium at 37°C in an anaerobic chamber in the presence of 85% nitrogen, 5% carbon dioxide, and 5% hydrogen (78). T. denticola isogenic mutants were grown with erythromycin (50 μg/mL) or/and gentamicin (20 μg/mL). Escherichia coli DH5α strain (New England Biolabs, Ipswich, MA) was used for DNA cloning; BL21-CodonPlus (DE3)-RIL (Agilent, Santa Clara, CA) was used for preparing recombinant proteins. E. coli strains were grown in lysogeny broth (LB) supplemented with appropriate concentrations of antibiotics for selective pressure as needed: erythromycin (400 μg/mL), kanamycin (50 μg/mL), chloramphenicol (50 μg/mL), and ampicillin (100 μg/mL). The oligonucleotide primers for PCR and RT-PCR used in this study are listed in Table S1. These primers were synthesized at IDT (Integrated DNA Technologies, Coralville, IA).

Gel electrophoresis and immunoblots.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting analyses were carried out as previously described (33, 79). For immunoblotting, T. denticola cells were harvested at the mid-logarithmic phase (~5 × 10⁸ cells/mL). Equal amounts of whole-cell lysates (~5 μg) were separated on SDS-PAGE and then transferred to polyvinylidene difluoride (PVDF) membranes. Immunoblots were probed with specific antibodies against T. pallidum FlaBs and T. denticola FlaA and DnaK, as described in previous studies (64, 80). The antibody against T. denticola TDE2683 was generated in this study (see below for details). E. coli GroEL antibody was purchased from Abcam (Cambridge, United Kingdom). Immunoblots were developed using a horseradish peroxidase-conjugated secondary antibody with an enhanced chemiluminescence (ECL) luminol assay, and signals were quantified using the Molecular Imager ChemiDoc system with the Image Lab software (Bio-Rad Laboratories, Hercules, CA) (79).

Preparation of TDE2683 recombinant protein and antiserum.

The gene encoding the full-length TDE2683 protein was PCR amplified with primers P₁/P₂ using Pfx DNA polymerase (Life Technologies, Grand Island, NY) and then cloned into the pET200/D-TOPO expression vector (Life Technologies), which encodes a six-histidine-tag (6×His) at the N terminus. The resulting plasmid was then transformed into BL21-CodonPlus (DE3)-RIL (Agilent). The expression of TDE2683 were induced using 1 mM isopropyl β-D-1-thiogalactosidase (IPTG). For antiserum production, recombinant TDE2683 protein was purified using nickel-nitrilotriacetic acid (Ni-NTA) agarose (Qiagen, Valencia, CA) under denaturing conditions. For EMSA, recombinant TDE2683 protein was purified using Ni-NTA agarose under native conditions. The purified proteins were then dialyzed in a buffer containing 20 mM Tris-HCl buffer (pH 8) at 4°C overnight using 3.0-kDa-molecular-weight-cutoff Spectra/Por dialysis bags (Spectrum Laboratories, Rancho Dominguez, CA). The concentrations of purified proteins were determined using a Bio-Rad protein assay kit (Bio-Rad). For antibody production, 5 mg of purified recombinant TDE2683 was used to immunize two rats following a standard immunization procedure. The primers for preparation of recombinant proteins are listed in Table S1.

Heterologous replacement of fliA_Ec with TDE2683.

We first in frame replaced fliA_Ec with TDE2683 via allelic exchange using the vector illustrated in Fig. S1A. This vector was constructed by two-step PCR and DNA cloning. Herein, the fliA_Ec upstream region (UR) and the full-length TDE2683 gene were first PCR amplified with primers P₃/P₄ and P₅/P₆, respectively, and then fused with primers P₃/P₆, generating UR-TDE2683, which was then cloned into the pGEM-T Easy vector (Promega). Then, an erythromycin cassette (erm) (62, 81) and the fliA_Ec downstream region (DR) were PCR amplified with primers P₇/P₈ and P₉/P₁₀, respectively. The resultant DR fragment was then cloned into the pGEM-T Easy vector (Promega, Madison, WI), which was then released and ligated to the 3′ end of UR-TDE2683 at an engineered XbaI cut site. Then, the erm cassette was inserted into the obtained vector at an engineered XbaI cut site, generating the vector TDE2683-erm. The obtained construct was linearized and transformed into E. coli DH5α to replace fliA_Ec. The resultant erythromycin-resistant colonies were first screened for the presence of TDE2683 by PCR and then confirmed by immunoblotting with TDE2683 antibody. In parallel, a mutant with a deletion of fliA_Ec (ΔfliA_Ec) was constructed by replacing fliA_Ec in frame with a kanamycin cassette (kan), as depicted in Fig. S1B. The primers used here are listed in Table S1.

Flagella staining.

A previously described Ryu staining method was used to visualize E. coli flagella (82). In brief, E. coli cultures were grown at 37°C for 2 days without shaking. The cell culture was transferred onto a microscope slide using a loop and then covered with a coverslip. A drop of Ryu stain (Remel, San Diego, CA) was applied to the edge of the coverslip, and the slide was incubated at room temperature for 15 min. The cells were visualized under dark-field or phase-contrast microscopy using a Zeiss Axio Imager.A2 microscope.

EMSA.

EMSA were carried out as previously described with some modifications (83). In brief, the T. denticola flaB2 promoter (P_flaB2) and E. coli fliC promoter (P_fliC) (55, 60) were PCR amplified using P₁₈/P₁₉ and P₂₀/P₂₁, respectively. The obtained amplicons were labeled with biotin using a biotin DecaLabel DNA labeling kit (Thermo Fisher Scientific, Waltham, MA) to generate two DNA probes. EMSA were performed by coincubating individual biotin-labeled DNA probes (2 nM) with increasing amounts of recombinant TDE2683 (6.4 μM, 12.9 μM, and 25.8 μM), or two mutated recombinant proteins (E221D and V231E), in a reaction buffer (10 mM Tris-HCl, pH 7.5) at room temperature for 1 h. For each reaction, only one labeled DNA probe and one type of recombinant protein were included. For competition assays, increasing amounts of unlabeled DNA probes (20 nM, 40 nM, 200 nM, and 2 μM) were added to the reaction mixtures. Following the reactions, DNA loading dye was added. The resulting samples were then run on 4 to 20% Mini-Protean TGX gels (Bio-Rad Laboratories), which were then transferred to nylon membranes and cross-linked using UV light. Detections were performed using a chemiluminescent nucleic acid detection module kit (Thermo Fisher Scientific). The primers for the two DNA probes are listed in Table S1.

Site-directed mutagenesis.

Site-directed mutagenesis was performed using a Q5 site-directed mutagenesis kit (New England Biolabs) according to the manufacturer’s instruction. The plasmid that expresses TDE2683 recombinant protein was used as a template to replace Glu221 and Val231 with Asp and Glu, respectively, using primers P₂₂/P₂₃ and P₂₄/P₂₅. The resultant mutations were confirmed by DNA sequencing analysis.

Construction of two TDE2683 site-directed mutants in T. denticola.

The vector shown in Fig. S2 was constructed to replace the wild-type TDE2683 with two genes with site-directed mutations at amino acid E221 or V231 by using two-step PCR and DNA cloning, as previously documented (84). In brief, to make this construct, the downstream region of TDE2683 and the mutated gene, either E221D or V231E, were PCR amplified with primers P₂₉/P₃₀ and P₅/P₂₆, respectively, and then fused with primers P₅/P₃₀, generating fliA_Td-3′, which was then cloned into the pMD19 T-vector (TaKaRa Bio USA, Inc., Mountain View, CA). An erythromycin cassette (ermB) (62) was PCR amplified with primers P₂₇/P₂₈ and cloned into the pGEM-T Easy vector (Promega), which was then released and inserted into fliA_Td-3′ at a NotI cut site, generating TDE2683*-ermB, which was transformed into T. denticola wild-type competent cells via heat shock; cells were then plated on soft agars with erythromycin, as previously described (85). Site-directed mutations were confirmed by PCR, followed by DNA sequencing.

RNA preparation, RT-PCR, qRT-PCR, and 5′ RACE.

RNA isolation was performed as previously described (33, 79, 86). Briefly, T. denticola cells were harvested at the mid-logarithmic phase (~5 × 10⁸ cells/mL). Total RNA was extracted using TRI reagent (Sigma-Aldrich), following the manufacturer's instructions. The resultant samples were treated with Turbo DNase I (Thermo Fisher Scientific) at 37°C for 2 h to eliminate genomic DNA contamination. The resultant RNA samples were re-extracted using acid phenol-chloroform (Ambion, Austin, TX), then precipitated in isopropanol, and finally washed once with 70% ethanol. The RNA pellets were resuspended in RNase-free water. cDNA was generated from the purified RNA (1 μg) using a SuperScript IV VILO cDNA synthesis kit (Thermo Fisher Scientific). qRT-PCR was performed using iQ SYBR green supermix and a QuantStudio 3 system (Applied Biosystems, Waltham, MA). The housekeeping gene dnaK (TDE0628) of T. denticola was used as an internal control to normalize the qRT-PCR data, as previously described (33). The results were expressed as the normalized difference in threshold cycle (ΔΔC_T) between the wild-type and the mutant strains. 5′ RACE analysis was performed using the SMARTer RACE 5′/3′ kit (TaKaRa Bio USA), according to the manufacturer’s protocol. The primers for RT-PCR, qRT-PCR, and 5′ RACE are listed in Table S1.

Bacterial swimming plate assays.

Swimming plate assays of E. coli were performed on soft-agar plates containing 1% tryptone, 1% NaCl, and 0.4% Bacto agar, as previously described (87). In brief, overnight cultures were diluted to 1:100 and incubated at 37°C for 8 h with shaking. One microliter of the culture was inoculated onto the soft-agar plates and incubated at 37°C for 14 h. Swimming plate assays of T. denticola were performed as previously described (79). Briefly, 3 μL of cultures (10⁹ cells/mL) were inoculated onto 0.35% agarose plates containing TYGVS medium diluted 1:5 with phosphate-buffered saline (PBS). The plates were incubated anaerobically at 37°C for 5 days to allow the cells to swim. The diameters of the swim rings were measured in millimeters. As a negative control, a Δtap1 strain, a previously constructed T. denticola nonmotile mutant, was included to determine the initial inoculum size (63). The average diameters of each strain were calculated from four independent plates; the results are presented as means and standard errors of the means (SEM). The data were statistically analyzed by one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparison at a P value of <0.01.

Cryo-ET sample preparation, data collection, and image processing.

The frozen-hydrated specimens of T. denticola were prepared as previously described (33). Briefly, T. denticola cultures were mixed with 10-nm-colloidal-gold solutions and then deposited on a freshly glow-discharged, holey carbon grid for about 1 min. The grids were blotted with a piece of filter paper for ~4 s and then rapidly plunged into liquid ethane using a gravity-driven plunger apparatus. Frozen-hydrated specimens of T. denticola were transferred to a 300-kV Krios electron microscope (Thermo Fisher Scientific) equipped with a field emission gun and a K2 direct detection detector (Gatan). To generate three-dimensional (3D) reconstructions of whole bacterial cells, SerialEM (88) was used to acquire multiple tilt series along the cell at ×4,500 magnification. The pixel size at the specimen level is 8.2 Å. All tilt series were collected in the low-dose mode with ~10-μm defocus. A total dose of 50 e⁻/Ų is distributed among 35 tilt images covering angles from −51° to +51° at tilt steps of 3°. The tilt series were aligned and reconstructed by IMOD (89). Multiple reconstructions from different segments of the same cell were integrated into one whole-cell map. In total, we generated 10 whole-cell reconstructions from the wild-type strain (3 cells) and two TDE2683 mutants (7 cells). IMOD (89) was used to take 2D snapshots of the reconstructions.

Bioinformatics and statistical analyses.

Sequence alignment analyses were performed using Clustal Omega, and homology modeling was conducted on the Swiss-Model server. Statistical significance was analyzed by one-way ANOVA followed by Tukey’s multiple comparison at a P value of <0.05.