An ECF41 Family σ Factor Controls Motility and Biogenesis of Lateral Flagella in Azospirillum brasilense Sp245

ABSTRACT

ECF41 is a large family of bacterial extracytoplasmic function (ECF) σ factors. Their role in bacterial physiology or behavior, however, is not known. One of the 10 ECF σ factors encoded in the genome of Azospirillum brasilense Sp245, RpoE10, exhibits features characteristic of the typical ECF41-type σ factors. Inactivation of rpoE10 in A. brasilense Sp245 led to an increase in motility that could be complemented by the expression of rpoE10. By comparing the number of lateral flagella, transcriptome, and proteome of A. brasilense Sp245 with those of its rpoE10::km mutant, we show here that this ECF41-type σ factor is involved in the negative regulation of swimming motility and biogenesis of lateral flagella of A. brasilense Sp245. The genome of A. brasilense Sp245 also encodes two OmpR-type regulators (LafR1 and LafR2) and three flagellins, including Laf1, the major flagellin of lateral flagella. Elevated levels of laf1 transcripts and Laf1 protein in the rpoE10::km mutant indicated that RpoE10 negatively regulates the expression of Laf1. The elevated level of LafR1 in the rpoE10::km mutant indicated that LafR1 is also negatively regulated by RpoE10. The loss of motility and Laf1 in the lafR1::km mutant, complemented by lafR1 expression, showed that LafR1 is a positive regulator of Laf1 and motility in A. brasilense. In addition, upregulation of laf1::lacZ and lafR1::lacZ fusions by RpoE10 and downregulation of the laf1::lacZ fusion by LafR1 suggest that RpoE10 negatively regulates swimming motility and the expression of LafR1 and Laf1. However, LafR1 positively regulates the swimming motility and Laf1 expression.

IMPORTANCE Among extracytoplasmic function (ECF) σ factors, ECF41-type σ factors are unique due to the presence of a large C-terminal extension in place of a cognate anti-σ factor, which regulates their activity. Despite their wide distribution and abundance in bacterial genomes, their physiological or behavioral roles are not known. We show here an indirect negative role of an ECF41-type of σ factor in the expression of lateral flagellar genes and motility in A. brasilense. This study suggests that the motility of A. brasilense might be controlled by a regulatory cascade involving RpoE10, an unknown repressor, LafR1, and lateral flagellar genes, including that encoding Laf1.

INTRODUCTION

Bacteria sense fluctuations in their external environment and respond either by expressing genes required for adapting to the altered environmental condition or by moving toward or away from the environmental stimulus (1). The locomotion ability of bacteria provides them with a survival advantage under favorable or unfavorable conditions. Most bacteria depend on flagellar motility and chemotaxis to move along chemical gradients (2–5). Escherichia coli and Salmonella enterica possess a single type of peritrichous flagellum that promotes swimming as well as swarming motility (6). Other bacteria, such as Aeromonas hydrophila, Vibrio parahaemolyticus, Rhodospirillum centenum, and Azospirillum brasilense, possess two types of flagella: a polar flagellum and multiple lateral flagella (7–9). The polar flagellum plays the main role in swimming through a liquid medium, whereas lateral flagella facilitate swarming through a viscous medium or over surfaces. An increase in the viscosity of the medium creates hindrance in the rotation of the polar flagellum, leading to the expression of genes involved in the synthesis of lateral flagella (10). While the polar flagellum is produced under all growth conditions, the synthesis of lateral flagella is inducible. Thus, under conditions that lead to swarming, both lateral flagella and the polar flagellum are produced. The expression of genes involved in the synthesis of flagella is controlled by a transcriptional hierarchy (11–13), which includes early, middle, and late genes that are expressed sequentially by specific transcriptional regulators and σ factors (6, 14). The assembly of flagellar components is also ordered (12).

Adaptation to a change in the bacterial environment is brought about by the expression of new sets of genes, initiated at promoter sequences that are specifically recognized by the sigma (σ) factor of an RNA polymerase. Based on their structure, promoter consensus, and mechanism of transcription initiation, σ factors are divided into two major families, the σ⁵⁴ family (represented by RpoN) and the σ⁷⁰ family. The σ factors of the σ⁷⁰ family are further divided into four groups based on their sequence, domain architecture, and function (15, 16). The primary σ factor (RpoD) belongs to group 1 and contains four highly conserved domains (designated σ¹ through σ⁴) along with a nonconserved region (15). Group 2 σ factors are closely related to group 1 σ factors but are not essential for growth. The σ factors of group 3 lack the σ¹ domain and control cellular processes such as sporulation, flagellum biosynthesis, or the heat shock response (RpoH) (17). Group 4 constitutes the largest and most diverse group of σ factors, which regulate the cellular response to extracellular stimuli and hence are known as extracytoplasmic function (ECF) or RpoE σ factors (18–20). In contrast to other σ⁷⁰ family members, the ECF σ factors contain only two of the four conserved domains, σ² and σ⁴, which are sufficient for promoter recognition and interaction with the core enzyme.

Based on sequence similarity, domain architecture of cognate anti-σ factors, genomic context conservation, and potential target promoter motifs, the ECF σ factor protein family has been classified into more than 40 distinct major groups (17). Typically, the activity of ECF σ factors is regulated by their cognate anti-σ factors, which are often cotranscribed as part of the same operon (17–20). Anti-σ factors physically bind and sequester their cognate ECF σ factors until an intracellular or extracellular stimulus leads to an alteration of their conformation or their proteolytic degradation, resulting in the release of the ECF σ factor to interact with RNA polymerase and to initiate transcription (20, 21). However, several ECF σ factor groups are neither cotranscribed with their cognate anti-σ factors nor known to interact with any anti-σ factor encoded elsewhere in the genome. ECF41 is one among such groups of ECF σ factors that lack cognate anti-σ factors. Instead of being regulated by anti-σ factors, ECF41 σ factors contain a C-terminal protein domain fused with the σ⁴ domain of the ECF that is thought to provide the necessary regulatory function (17, 19, 22, 23).

Azospirillum brasilense is a plant growth-promoting rhizobacterium that colonizes the roots of many grasses and promotes their growth by producing phytohormones and fixing atmospheric nitrogen (24). It shows motility via a dual flagellar system comprised of a constitutive polar flagellum and several inducible lateral flagella (8). While the polar flagellum is mainly responsible for swimming in liquid media, lateral flagella mediate swarming across surfaces or in a viscous medium (25–27). Nearly all the genes involved in the synthesis of lateral flagella are located on chromid 4, whereas those involved in the synthesis of the polar flagellum occur as large clusters on the chromosome and small clusters on chromids 1 and 3.

Flagellins are the major proteins that constitute polar as well as lateral flagella. The genome of A. brasilense Sp245 carries genes encoding three flagellins: fliC1 (AZOBR_p1160045) on chromid 1, fliC2 (AZOBR_p340061) on chromid 3, and laf1 (AZOBR_p410056) on chromid 4 (28). The gene encoding the Laf1 flagellin from A. brasilense was previously isolated and used for the construction of a laf1 mutant, which lacked lateral flagella but formed polar flagellum (29). Therefore, a laf1::km mutant lacks swarming on a semisolid surface but is still able to swim. The expression of laf1 in A. brasilense is induced on solid medium or when the rotation of the polar flagellum is hindered (9). σ⁵⁴ was also shown to affect the motility of A. brasilense, as an rpoN::km mutant of A. brasilense Sp7 completely lacked lateral as well polar flagella and was nonmotile (30).

While examining the genomic organization of genes encoding ECF σ factors in A. brasilense Sp245, we found that the gene encoding an ECF σ factor (AZOBR_p470050, RpoE10) is not accompanied by an anti-σ factor. Further, the protein predicted by this gene contains an extension of about 100 amino acids at its C terminus, similar to other ECF41-type σ factors (31, 32). Although these types of σ factors are encoded in the genomes of many bacteria, their role in bacterial physiology or behavior has not yet been elucidated. In this study, we provide evidence that an ECF41 σ factor homolog negatively regulates the motility and biogenesis of lateral flagella in A. brasilense Sp245. We also show that an OmpR-type master regulator (LafR1), which appears to be negatively regulated by RpoE10, controls the expression of the Laf1 flagellin and motility in A. brasilense Sp245.

RESULTS

RpoE10 of Azospirillum brasilense belongs to the ECF41 family of σ factors.

The genome of A. brasilense Sp245 encodes 23 σ factors, including 1 RpoD, 1 RpoN, 5 RpoH, 6 RpoI, and 10 RpoE factors. A dendrogram based on the deduced amino acid sequences of the 23 σ factors (see Fig. S1 in the supplemental material) shows that RpoE10 belongs to the group 4 σ factors, which include RpoE and FecI; however, RpoE10 exhibits distinct differences from other members of this family. Alignment of the deduced amino acid sequences of the 10 RpoE σ factors of A. brasilense shows that RpoE10 harbors a C-terminal extension of 119 amino acids, which is characteristic of the ECF41 family of ECF σ factors (see Fig. S2 in the supplemental material). We aligned the deduced amino acid sequence of RpoE10 with the sequences of the previously characterized ECF41 family σ factors of Rhodobacter sphaeroides, Bacillus licheniformis, and Mycobacterium tuberculosis (22, 31) and found that RpoE10 of A. brasilense was more closely related to the ECF41 σ factors from R. sphaeroides (RSP_0607) and B. licheniformis (BLi04371) and more distantly related to the ECF41 σ factor SigJ (Rv3328c) from M. tuberculosis (Fig. 1A). These in silico analyses show that RpoE10 of A. brasilense Sp245 belongs to the ECF41 family of σ factors.

FIG 1.

(A) Alignment of the deduced amino acid sequence of A. brasilense RpoE10 (AZOBR_p470050) with ECF41 from Bacillus licheniformis (Bli_04371), SigJ of Mycobacterium tuberculosis (Rv3328c), and ECF41 from Rhodobacter sphaeroides (Rsp_06070), showing the σ² and σ⁴ regions. Domains are marked σ², σ⁴, and C-terminal extension. The alignment also shows the signature motifs characteristic of ECF41-type sigma factors (WLPEP, DGGGR, and NPDK). (B) Organization of the gene encoding RpoE10 in A. brasilense Sp245. Identities of proteins encoded by each ORF are shown within the arrows (LafR1, regulator of lateral flagella; ABM, antibiotic biosynthesis monooxygenase; FliG, flagellar motor switch protein; and HP, hypothetical protein). The distance (27 nucleotides [NT]) between rpoE10 and its upstream gene is indicated below the junction of the two genes. The agarose gel shows RT-PCR amplicons of 832 bp obtained with primers RT1 and RT2 (lane 2), of 336 bp obtained with primers RT3 and RT4 (lane 3), of 986 bp with primers RT5 and RT6 (lane 4), and of 1,322 bp with primers RT3 and RT6 (lane 5), no amplicon with primers RT1 and RT4 (lane 6) using A. brasilense total RNA as a template, and an amplicon of 1,322 bp with primers RT3 and RT6 (lane 7) used as a positive control using genomic DNA. Lane 8 shows a control lacking reverse transcriptase. Lane 1 shows a 100-bp ladder.

A gene encoding antibiotic biosynthesis monooxygenase (abm) is cotranscribed with rpoE10.

We next analyzed the genomic organization of the gene encoding RpoE10 in A. brasilense, which has a tight linkage with a putative gene encoding antibiotic biosynthesis monooxygenase (ABM) located 27 nucleotides upstream of rpoE10 (Fig. 1B). When the genomic context of rpoE10 was analyzed in other members of the genus Azospirillum, all the strains of A. brasilense showed linkage with a gene encoding ABM, while in A. lipoferum and Azospirillum sp. strain B510, genes encoding alkyl hydroxyperoxide reductase (AhpD) and carboxymuconolactone decarboxylase (CMD), respectively, were located upstream of rpoE10 orthologs (see Fig. S3 in the supplemental material). Except for A. brasilense, A. lipoferum, A. thiophilum, and Azospirillum B510, rpoE10 orthologs were absent in all other species of Azospirillum. To determine if abm and rpoE10 are cotranscribed, we performed reverse transcriptase PCR (RT-PCR) of A. brasilense Sp245 RNA using primers spanning both genes. We obtained PCR amplification of cDNAs of the expected sizes using primer RT3 with primer RT6 (primers are listed in Table 1) and total RNA of A. brasilense Sp245 as the template (Fig. 1B), confirming that abm and rpoE10 are cotranscribed and are part of a bicistronic operon. A gene encoding a putative transcription regulator, which we named LafR1 (lateral flagellum regulator 1) based on evidence obtained in this study (see “Expression of the LafR1 regulator is strongly upregulated in the rpoE10::km mutant” below), is located upstream of abm but not cotranscribed with abm, as shown by the failure to obtain any amplicon with the primer pair RT1 and RT4 although the lafR1 cDNA was readily amplified with primer pair RT1 and RT2 (Fig. 1B). From these results, we conclude that abm and rpoE10 are cotranscribed and that the gene encoding LafR1is not part of this operon.

TABLE 1.

Primers used in this study

Primer	Sequence (5′ to 3′)a
ProRpoE10_FP	CCCAAGCTTCGTTTGGGCGTGGGGGTG
ProRpoE10_RP	GCTCTAGAGGTCCTGCCGCCTTGCCA
RpoE10_OE_FP	AAACTGCAGATGGACGACATCACCACCG
RpoE10_OE_RP	CCCAAGCTTTCAGTTCAGCTCGATTTTG
RpoE10_OE(DEL1)_RP	CCCAAGCTTTCACCGCGCTGGCAATGACATC
RpoE10_OE(DEL1)_RP	CCCAAGCTTTCACCGCGCTGGCAATGACATC
RpoE10_AFP	AACTGCAGATGCGTGCTGTTCTGGTTG
RpoE10_ARP	GAAGATCTGGGTCGGTGGTGATGTCG
RpoE10_BFP	GAAGATCTAAATCGAGCTGAACTGAGCC
RpoE10_BRP	CGGAATTCGCGTGTCGCTCCAGGCG
RpoE10_GSP1	TGAGTTGCTCTTGCGTCTC
RpoE10_GSP2	CCCTCGAAGTGCTCCTTGG
RpoE10_GSP3	GTCGAGGCTGCTATGGATGC
Oligo(dT) anchor primer	GACCACGCGTATCGATGTCGACTTTTTTTTTTTTTTTTV
PCR anchor primer	GACCACGCGTATCGATGTCGAC
M13 forward primer	GTTTTCCCAGTCACGAC
RpoD_RTF	GCTCATGGACGGCATGGGTC
RpoD_RTR	CTGCTCGGAAGACGACGTATC
Prolaf1_FP	CCCAAGCTTGGCGCCTTAATTTCAAAGTC
Prolaf1_RP	GCTCTAGAGTGTATCTCCATTCGCGAG
Prolaf1_RT_FP	CATCGTCGAGAACCTGCAGACG
Prolaf1_RT_RP	TCCTGGCGCTGGAAGCTGATGT
ProFliG_FP	CCCAAGCTTCAGATTGTCGAAGGCCATC
ProFliG_RP	GCTCTAGACGAAATATTACTAAAATCCATTG
ProFlgB_FP	CCCAAGCTTCAAGGGGCCGCGCCCGTC
ProFlgB_RP	GCTCTAGACGAAATATTACTAAAATCC
LafR1_AFP	AACTGCAGTCTGCCGCCGCGCCAAC
LafR1_ARP	GAAGATCTACCAGAACAGCACGCATGG
LafR1_BFP	GAAGATCTATCGAGACCGTCTACGGC
LafR1_BRP	CGGAATTCGAGCCGCCTGGATATGCTC
LafR1_FP	CGGAATTCATGCGTGCTGTTCTGGTTG
LafR1_RP	CGGGATCCCAGATCGTGTCCGCCCTCA
PROLafR2_FP	CCCAAGCTTCGTTTCCTCTCTCCCGTTC
PROLafR2_RP	GCTCTAGAGGGATTCTCCGCAACAGGG
PROLafR1_FP	CCCAAGCTTACGCGCAGGAGTTGCGTGG
PROLafR1_RP	GCTCTAGAGGGGTTCTCCGCAGCATC

^aUnderlined sequences are the restriction endonuclease sites used for cloning of inserts.

The promoter of the abm-rpoE10 operon is similar to the ECF41 σ factor promoter consensus.

To define the promoter elements that control transcription of the abm-rpoE10 transcript, we first determined the transcription start site (TSS) by 5′ rapid amplification of cDNA ends (RACE) with RNA extracted from A. brasilense Sp245. We identified a “C” as a TSS, which is located 15 nucleotides upstream of the start codon ATG (Fig. 2A). Upstream of the TSS, we identified a bipartite sequence motif with similarity to the typical ECF σ factor-dependent promoter elements (5). The promoter contains a TGTCAC sequence as a −35 element, which is identical to that found upstream of the genes encoding the ECF41 σ factor in R. sphaeroides (RSP_0606), B. licheniformis (ydfG), and M. tuberculosis (sigJ) (31, 33). The abm-rpoE10 promoter also contains TGTCAT as a −10 element, which is similar to the CGTCAT present upstream of the ECF41 σ factor in R. sphaeroides. The −10 and −35 elements are separated by 17 nucleotides in the consensus sequence of ECF41 homologs from both bacteria (Fig. 2B). The predicted promoter elements of rpoE10 were, therefore, highly similar to the consensus of the promoters of the genes encoding ECF σ factors.

FIG 2.

(A) Chromatogram indicating the transcription start site of rpoE10 determined by 5′ RACE, showing a “C” as a transcriptional start point encircled in red. (B) Nucleotide sequence of the intergenic regions between the stop codon of lafR1 and the start codon of abm of A. brasilense Sp245, showing −10 and −35 hexamers of the promoter. The sequences below the promoter region of the abm-rpoE10 operon of A. brasilense show the promoter regions of ECF41-type σ factors of R. sphaeroides, B. licheniformis, and M. tuberculosis, revealing conservation of promoter sequences. (C) Schematic representation of the RpoE10 σ factor and its deletion/mutant derivatives. RpoE10 shows the characteristic motifs of ECF41-type σ factors: the WLPEP motif located between the σ² and σ⁴ regions and other two motifs (DGGGR and NPDKV) located in the C-terminal extension. The first deletion derivative (DEL1) contained the RpoE10 sequence up to DGGGR, while the second deletion derivative (DEL2) excluded the DGGGR motif. (D) Effect of deletion and mutation of the DGGGR and NPDKV motifs at the C terminus of RpoE10 on the expression of an abm-rpoE10::lacZ promoter fusion (pAPD12) in A. brasilense Sp245 expressing full-length RpoE10 (pAPD8) and its two deletion derivatives, RpoE10(DEL1) (pAPD9) and RpoE10(DEL2) (pAPD10). Error bars show standard deviations (SD) for triplicates of three independent experiments. Data (n = 3) were analyzed by using the SPSS17 software and subjected to one general-linear-model analysis of variance (ANOVA). The differences between means were compared, and letters above the bars indicate the result of Tukey’s multiple-comparison test (different letters indicate P values of <0.05).

Effect of truncation of the C-terminal extension of RpoE10 on the activation of the abm-rpoE10 promoter.

To compare the regulation of the abm-rpoE10 promoter with that of other ECF41 target promoters of R. sphaeroides and B. licheniformis, we constructed an abm-rpoE10::lacZ fusion in the broad-host-range vector pCZ750 (designated pAPD12), which was conjugatively mobilized in A. brasilense Sp245. Since ECF41 σ factors have been shown to regulate their promoters, the full-length rpoE10 was also cloned in the same isopropyl-β-d-thiogalactopyranoside (IPTG)-inducible vector (designated pAPD8) and mobilized in A. brasilense Sp245 harboring pAPD12. We did not find any significant increase in the β-galactosidase activity from an abm promoter even after inducing the expression of RpoE10 with 1 mM IPTG (Fig. 2D). Since truncation of C-terminal residues beyond the DGGGK motif was shown to increase the promoter activity of the genes encoding the ECF41 σ factors of R. sphaeroides and B. licheniformis, we also constructed genes expressing two different truncated derivatives of RpoE10 in a broad-host-range expression vector. RpoE10(DEL1), cloned in pAPD9, carries a deletion up to (but excluding) the DGGGR motif, while RpoE10(DEL2), cloned in pAPD10, includes the DGGGR residues (Fig. 2C). When we mobilized these recombinant plasmids in A. brasilense Sp245 harboring an abm-rpoE10::lacZ fusion, we found that expression of RpoE10(DEL1) from pAPD9 increased β-galactosidase activity by almost 15 times after induction with 1 mM IPTG. In contrast, expression of RpoE10(DEL2) from pAPD10 did not result in any promoter activity (Fig. 2D). This observation indicates that RpoE10 in its native form does not activate its own promoter (upstream of abm). However, in an altered form it activates its promoter, demonstrating an autoregulation similar to that exhibited by several other ECF σ factors.

Inactivation of rpoE10 in A. brasilense Sp245 increases motility.

To understand the biological role of RpoE10 in A. brasilense, the gene encoding this protein was inactivated by replacing its internal region with a kanamycin resistance gene cassette and placing rpoE10::km into the A. brasilense genome by allele replacement. The rpoE10::km mutant showed no difference in growth rate from A. brasilense Sp245 in either minimal malate (MM) medium or Luria-Bertani (LB) medium, nor did these strains show any notable difference in their sensitivity to salt stress or different oxidative stresses (data not shown). When the swarming motility of the rpoE10::km mutant was compared with that of its parent in nutrient broth (NB) containing 0.6% agar, the diameter of the swarm halo did not show any major difference even after 72 h (Fig. 3A). However, there was a dramatic difference in the colony morphology between the mutant and the parent on minimal medium containing 0.6% agar-agar and triphenyl tetrazolium chloride (TTC) (50 μg/ml) (Fig. 3B). Colonies of the parent were circular and raised with entire margins, whereas those of the rpoE10::km mutant were hyperwrinkled and umbonate with irregular margins. The swimming motility of the rpoE10::km mutant in soft agar (0.3% agar-agar in MM or NB) after 72 h was considerably higher than that of its parent (Fig. 3C). The sizes of the haloes of the parent and the rpoE10::km mutant were 2.0 ± 0.3 cm and 5.3 ± 0.5 cm, respectively, indicating that the swimming motility of the mutant was almost 2.5-fold greater than that of the parent. The swimming motility of the complemented strain was similar to that of the parent. This observation indicates that RpoE10 negatively regulates the motility of A. brasilense Sp245.

FIG 3.

(A) Comparison of the swarming motilities of A. brasilense Sp245 and the rpoE10::km mutant on 0.6% NB agar medium. (B) Comparison of the colony morphologies of A. brasilense Sp245, the rpoE10::km mutant, and the rpoE10::km (rpoE10↑) complemented strain on 0.6% MM agar medium. Triphenyl tetrazolium chloride (TTC) (50 μg/ml) was used to improve visualization of the colony morphology. (C) Comparison of the swimming motilities of A. brasilense Sp245, the rpoE10::km mutant, and the rpoE10::km (rpoE10↑) complemented strain on 0.3% MM agar medium.

Identification of differentially expressed transcripts in the rpoE10::km mutant by RNA-seq.

To understand the role of RpoE10 in A. brasilense, we used transcriptome sequencing (RNA-seq) to analyze the differences in the transcriptomes of the rpoE10::km mutant and its parent. Only significantly and markedly differentially expressed transcripts (P < 0.05 by analysis of variance [ANOVA]; maximum fold change, ≥2.0) were considered candidate transcripts for informatics analysis. A total of 208 transcripts were differentially expressed at least 2-fold with P values of <0.05. Out of these, 114 transcripts were downregulated and 94 transcripts were upregulated (see Table S1 in the supplemental material). Among the differentially expressed subset of transcripts in the rpoE10::km mutant, 23 genes that were highly upregulated (6- to 10-fold) are known to be involved in the biogenesis of lateral flagella (Table 2; see Fig. S5A in the supplemental material). In addition, we also found that the genes encoding the sugar ABC transporter (RbsA, RbsB, and RbsC) and 2-ketoglutarate semialdehyde dehydrogenase were downregulated in the rpoE10::km mutant. However, we could not detect the transcript of the abm gene, which is located upstream of the rpoE10 gene. This is possibly due to the autoregulation of the abm-rpoE10 operon by RpoE10.

TABLE 2.

Identification of a subset of differentially expressed transcripts in the rpoE10::km mutant using RNA-seq

Gene annotation IDa	Encoded protein	Log2 fold expression change	P value
AZOBR_p410056	Laf1, flagellin	10.00094	1.21E−45
AZOBR_p410057	FlaF, putative modulator of flagellin synthesis	8.196866	9.34E−30
AZOBR_p410058	FlbT, flagellum biosynthesis repressor protein	8.219169	1.16E−28
AZOBR_p410059	FlaG, hypothetical protein	4.241586	0.02626
AZOBR_p410060	Hypothetical protein	7.007155	2.99E−12
AZOBR_p410066	FlgB, flagellar basal-body rod protein	8.6885	1.02E−56
AZOBR_p410067	FlgC, flagellar basal-body rod protein	7.810572	3.51E−11
AZOBR_p410068	FliE, flagellar basal body protein	7.865424	1.10E−08
AZOBR_p410069	FliQ, flagellar biosynthesis protein	7.623516	7.39E−13
AZOBR_p410071	FlhA, flagellar export pore protein	7.427027	0.005911
AZOBR_p410072	FliR, flagellar export pore protein	7.198581	5.26E−14
AZOBR_p410073	FlhB, flagellar export pore protein	6.846883	3.76E−06
AZOBR_p410096	FlgK, flagellar hook-associated protein	6.196993	0.011561
AZOBR_p410097	FlgE, flagellar hook protein	6.735556	0.008521
AZOBR_p470048	LafR, cell cycle transcriptional regulator	8.9485	8.73E−54
AZOBR_p470052	FliG, flagellar motor switch protein	6.72792	1.81E−06
AZOBR_p470055	FliN, flagellar motor switch protein	7.141596	7.41E−05
AZOBR_p470056	FliH, flagellar assembly protein	7.312883	7.97E−11
AZOBR_p470059	FliM, flagellar motor switch protein	7.791814	1.39E−18
AZOBR_p470060	Hypothetical protein	7.67948	1.56E−30
AZOBR_p470061	FliK, putative flagellar hook length control protein	7.204571	7.61E−07
AZOBR_p470062	FlgD, flagellar hook capping protein	6.66652	2.11E−07
AZOBR_p470073	FlgH, flagellar L-ring protein FlgH	7.303781	1.97E−11
AZOBR_p470074	FlgA, flagellar basal-body P-ring formation protein	7.419726	4.07E−20
AZOBR_p470075	FlgG, flagellar basal-body rod protein	6.75704	0.006882
AZOBR_p470076	FlgF, flagellar basal-body rod protein	6.621594	0.008787
AZOBR_p440159	2-Ketoglutarate semialdehyde dehydrogenase	−2.76553	0.005302
AZOBR_p440155	Sugar ABC transporter, permease component (RbsC-like)	−3.3599	7.26E−04
AZOBR_p440158	Sugar ABC transporter, periplasmic-binding component (RbsB-like)	−3.52774	3.55E−06
AZOBR_p440156	Sugar ABC transporter, permease component (RbsC-like)	−3.68281	1.20E−06
AZOBR_p440157	Sugar ABC transporter, ATP-binding component (RbsA-like)	−3.8609	2.33E−09

^aBold indicates genes whose protein products were also upregulated in the rpoE10::km mutant.

All 23 highly upregulated genes in the rpoE10::km mutant are located on chromid 4 of the A. brasilense Sp245 genome (Fig. S5B) and organized primarily in 6 clusters: the flhA fliR flhB cluster, the flgB flgC fliE fliQ cluster, the flgI flgL flgK flgE cluster, the laf1 flaF flbT flaG cluster, the flgH flgA flgG flgF cluster, and the fliP fliN fliH flhF fliF flhM cluster (Fig. S5C). However, fliL and fliG were monocistronic. Upregulation of several genes related to the biogenesis of lateral flagella in the rpoE10::km mutant suggested that RpoE10 is involved in the negative regulation of genes involved in the biogenesis of lateral flagella.

Production of lateral flagella and Laf1 protein is significantly upregulated in the rpoE10::km mutant.

Comparison of the morphology of the cells of A. brasilense Sp245 and the rpoE10::km mutant under transmission electron microscopy revealed that A. brasilense Sp245 produces only one polar flagellum, whereas the rpoE10::km mutant produces several (relatively thinner) lateral flagella in addition to the polar flagellum (see Fig. S4 in the supplemental material). We also analyzed the differences between extracellular proteins of A. brasilense Sp245 and the rpoE10::km mutant by SDS-PAGE. We also included an rpoN::km mutant of A. brasilense Sp245, which does not produce either a polar flagellum or lateral flagella (30). Extracellular proteins were obtained after thorough vortexing of the overnight-grown cultures of A. brasilense Sp245, the rpoE10::km mutant, and the rpoN::km mutant, followed by precipitation with polyethylene glycol (PEG) 6000 and NaCl, and the resulting proteins were resolved by SDS-PAGE. Figure 4A shows one prominent protein band of about 45 kDa, which was present in the rpoE10::km mutant but absent in both A. brasilense Sp245 and the rpoN::km mutant. This protein was identified as Laf1 by matrix-assisted laser desorption ionization (MALDI) tandem mass spectrometry (MS/MS) (Fig. 4B). A distinct protein with an approximate molecular mass of 100 kDa was present in the lanes containing extracellular proteins from A. brasilense Sp245 and the rpoE10::km mutant but was absent in the lane containing extracellular proteins from the rpoN::km mutant. Since the characteristic flagellin (Fla) of the polar flagellum of A. brasilense (29, 34) is ca. 100 kDa and is not produced by the rpoN::km mutant, we considered the 100-kDa protein to be Fla. The presence of the 100-kDa protein band in A. brasilense Sp245 and the rpoE10::km mutant with almost equal intensity showed that production of the polar flagellum is not affected in the rpoE10::km mutant.

FIG 4.

(A) Extracellular proteins obtained from the supernatant of the rpoN::km mutant of A. brasilense Sp245 (lane1), A. brasilense Sp245 (lane 2), and the rpoE10::km mutant (lane 3). M, protein molecular weight marker. The extracellular proteins were separated by SDS-PAGE and stained with Coomassie brilliant blue. The positions of the flagellins Fla and Laf1 (polar and lateral, respectively) are indicated at the sides of the images. (B) Amino acid sequence of the Laf1 (AZOBR_p410056) protein, showing sequences of the peptides (underlined) identified by MALDI-TOF/TOF obtained after tryptic digestion of the distinctly upregulated (45-kDa) protein of the rpoE10::km mutant.

Expression of laf1::lacZ, flhA::lacZ, flgB::lacZ, and fliG::lacZ is upregulated in the rpoE10::km mutant.

Data obtained by RNA-seq analysis of the upregulated genes were validated by constructing transcriptional lacZ fusions with the promoters of fliG, flgB, and laf1 as representatives of the upregulated genes. While fliG is a solo gene, flgB and laf1 are the first genes of the flgB flgC fliE fliQ cluster and the laf1 flaF flbT flaG cluster, respectively. Figure 5 shows that expression of all three reporters was upregulated more than 12-fold in the rpoE10::km mutant, with those containing the laf1 and lafR1 promoters being maximally upregulated. These results validate the results of RNA-seq analysis, confirming negative regulation of fliG, flgB, and laf1 by RpoE10.

FIG 5.

Comparison of the β-galactosidase activities from laf1::lacZ, flgB::lacZ, and fliG::lacZ fusions in A. brasilense Sp245 and its rpoE10::km mutant. Error bars show standard deviations (SD) for triplicates of three independent experiments. Data (n = 3) were analyzed using the SPSS17 software and subjected to one general-linear-model analysis of variance (ANOVA). The differences between means were compared, and letters above the bars indicate the result of Tukey’s multiple-comparison test (different letters indicate P values of <0.05).

Identification of differentially expressed proteins in the rpoE10::km mutant.

To confirm that the RNA-seq data on upregulated genes related to the biogenesis of lateral flagella are also corroborated by proteomic analysis, we carried out a comparison of the proteomes of A. brasilense Sp245 and the rpoE10::km mutant grown in MM medium. Liquid chromatography (LC)-MS-based analysis revealed that a large number of proteins were differentially expressed in the rpoE10::km mutant. Only significantly and markedly differentially expressed proteins (P < 0.05 by analysis of variance; maximum fold change, ≥1.5) were considered candidate proteins for informatics analysis. Accordingly, a total of 296 proteins were differentially expressed with P values of <0.05. Out of these, 171 were downregulated in the rpoE10::km mutant and 125 were upregulated (Table S1). There were 13 lateral flagellar proteins that were upregulated in the rpoE10::km mutant (Table 3). Our RNA-seq data also showed upregulation of their transcript levels in the rpoE10::km mutant (Table 2). These proteins included Laf1, FlaF, FlbT, FlgC, FliE, FlhB, FlgE, FliN, FlgD, FlgH, FlgG, FliH, and FliM. Some flagellar genes (including flaG, fliQ, flhA, fliR, fliG, flgA, and flgF) that were upregulated in the RNA-seq analysis did not show upregulation of their encoded proteins. Some upregulated proteins whose genes were not found in the list of upregulated transcripts included FliL, FlgN and FlgI. The proteomic analysis further supports the role of RpoE10 as a negative regulator of lateral flagellum biosynthesis genes.

TABLE 3.

Identification of a subset of differentially expressed proteins in the rpoE10::km mutant using proteomics data

Gene annotation IDa	Description	Abundance ratio	Log2 abundance ratio
AZOBR_p410056	Laf1, flagellin	21.929	4.45
AZOBR_p410057	FlaF, flagellar basal body protein	28.639	4.84
AZOBR_p410058	FlbT, flagellum biosynthesis repressor protein	10.133	3.34
AZOBR_p410067	FlgC, flagellar basal-body rod protein	3.576	1.84
AZOBR_p410068	FliE, flagellar hook–basal-body complex protein	10.57	3.4
AZOBR_p410073	FlhB, flagellar biosynthetic protein	8.402	3.07
AZOBR_p410084	FlgI, flagellar P-ring protein	4.94	2.3
AZOBR_p410092	FlgN, flagellar biosynthesis protein	11.019	3.46
AZOBR_p410097	FlgE, flagellar hook protein	36.005	5.17
AZOBR_p470055	FliN, flagellar motor switch protein	37.638	5.23
AZOBR_p470056	FliH, flagellar assembly protein	13.034	3.7
AZOBR_p470059	FliM, flagellar motor switch protein	3.514	1.81
AZOBR_p470062	FlgD, Basal-body rod modification protein	7.057	2.82
AZOBR_p470073	FlgH, flagellar L-ring protein	32.229	5.01
AZOBR_p470075	FlgG, flagellar basal-body rod protein	26.326	4.72
A0A0P0FEY1	FliL, flagellar basal body protein	12.127	3.6
A0A560BCI3	MotE, flagellar motility protein (MotC chaperone)	7.831	2.97

^aBold indicates proteins whose transcripts were also upregulated in the rpoE10::km mutant.

Expression of the LafR1 regulator is strongly upregulated in the rpoE10::km mutant.

The second most conspicuously upregulated gene (after laf1) in the transcriptome of the rpoE10::km mutant was AZOBR_p470048, which encodes a putative homolog of the OmpR type of master regulators, such as Rem (SMc03046) in Sinorhizobium meliloti (35), Ftc (BMEII0158) in Brucella melitensis (36), and LafR (blr6846) in Bradyrhizobium diazoefficiens (37) (see Fig. S7 in the supplemental material). Similar to its orthologs in other bacteria, AZOBR_p470048 possesses a signal receiver (REC) domain and a winged helix (HTH) domain (38). Examination of the genome of A. brasilense Sp245 revealed two copies of the lafR gene; one (lafR1, AZOBR_p470048) was located upstream of the abm-rpoE10 operon, and the other (lafR2, AZOBR_p410055) was present upstream of the laf1 gene. Both copies of lafR were located on chromid 4 of the A. brasilense Sp245 genome. Although their amino acid sequences are 90% identical, they exhibit major differences at their C termini (see Fig. S7 in the supplemental material).

To evaluate the effect of RpoE10 on the two paralogs of lafR, we constructed lacZ fusions with the promoters of both the lafR paralogs and quantified β-galactosidase activity in A. brasilense Sp245 and the rpoE10::km mutant. Figure 6A shows that the expression of lafR1 (AZOBR_p470048) was upregulated by almost 5-fold in the rpoE10::km mutant, but that of lafR2 (AZOBR_p410055) was only marginally upregulated. Examination of the transcript levels of lafR1, lafR2, and laf1 in A. brasilense and the rpoE10::km mutant by real-time PCR also showed that lafR1and laf1 transcripts were strongly upregulated in the rpoE10::km mutant (Fig. 6B). Thus, out of the two LafR paralogs encoded in the genome of A. brasilense, the expression of lafR1 was regulated negatively by RpoE10, but the expression of lafR2 was not dependent on RpoE10.

FIG 6.

(A) Comparison of the β-galactosidase activities from lafR1::lacZ and lafR2::lacZ fusions in A. brasilense Sp245 and its rpoE10::km mutant. Each bar represents mean and standard deviations (SD) of triplicates in two independent experiments. Data (n = 3) were analyzed using the SPSS17 software and subjected to one general-linear-model analysis of variance (ANOVA). The differences between means were compared, and letters above the bars indicate the result of Tukey’s multiple-comparison test (different letters indicate P values of <0.05). (B) Relative transcript levels of laf1, lafR1, and lafR2 in the rpoE10::km mutant. Error bars show SD of triplicates from three independent experiments, and differences between means were compared. Letters above the bars indicate the result of Tukey’s multiple-comparison test (different letters indicate P values of <0.05).

LafR1 regulates the expression of Laf1.

Since the expression of both Laf1 and LafR1 is strongly upregulated in the rpoE10::km mutant, we aimed to examine if the expression of laf1 is regulated by LafR1. For this, we inactivated the lafR1 gene in A. brasilense Sp245 by inserting a kanamycin (Km) resistance gene cassette and replaced the mutated allele into the genome by homologous recombination. Comparison of the β-galactosidase activities from the laf1::lacZ fusion in A. brasilense Sp245 and the lafR::km mutant showed almost complete loss of activity of the laf1::lacZ fusion in the lafR1::km mutant. β-Galactosidase activity was restored to nearly that of the parent in the complemented strain, indicating that LafR1 regulates the expression of laf1. To examine this further, we compared the extracellular proteins of A. brasilense Sp245 with those of the lafR1::km mutant and the complemented strain. While the lafR1::km mutant and wild-type strain failed to synthesize Laf1 flagellin, Laf1 production was gained by introducing a broad-host-range plasmid expressing the lafR1 gene under the control of an IPTG-inducible lacUV5 promoter (Fig. 7A and B), indicating that the defect in laf1 synthesis was a consequence of the disruption of lafR1.

FIG 7.

(A) Comparison of the β-galactosidase activities from the laf1::lacZ fusion in A. brasilense, the lafR1::km mutant, and the lafR1::km mutant overexpressing lafR1 (pAPD11). Error bars show standard deviations (SD) for triplicates of two independent experiments. Data (n = 3) were analyzed by using the SPSS17 software and were subjected to one general-linear-model analysis of variance (ANOVA). The differences between means were compared, and letters above the bars indicate the result of Tukey’s multiple-comparison test (different letters indicate P values of <0.05). (B) Extracellular proteins of A. brasilense Sp245 (lane1), the lafR1::km mutant (lane 2), and the lafR1::km mutant overexpressing lafR1 (pAPD11), showing overexpression of Laf1 in the lafR1::km mutant. The extracellular proteins were separated by SDS-PAGE and stained with Coomassie brilliant blue. The positions of the flagellins Fla and Laf1 (polar and lateral, respectively) are indicated at the sides of the images. (C) Comparison of the motilities of the lafR1::km mutant and the lafR1::km mutant overexpressing lafR1 (pAPD11) with that of A. brasilense Sp245 (Fig. 3C). The arrow indicates complementation via expression of lafR1 cloned in a broad-host-range vector.

If LafR1 regulates the expression of flagellar genes, including laf1, in A. brasilense Sp245, then the expression of lafR1 in the lafR1::km mutant should increase the motility of the mutant. Figure 7C shows that the diameter of the swimming halo of the lafR1::km mutant increased from 1 ± 0.2 cm to about 6 ± 0.5 cm due to the trans complementation of the lafR1::km mutant with a broad-host-range plasmid expressing lafR1 under the control of an IPTG-inducible promoter, indicating that lafR1 expression is sufficient to produce functional lateral flagella. Taken together, these observations demonstrate that LafR1 of A. brasilense is a positive regulator of motility and of the Laf1 flagellin.

DISCUSSION

On the basis of the presence of characteristic −10 and −35 elements in the promoter of rpoE10, the lack of a gene encoding an anti-σ factor in the vicinity of rpoE10, and the presence of a C-terminal SnoaL_2 domain in the RpoE10 protein, we show that RpoE10 of A. brasilense belongs to the ECF41 type of σ factors. We also show that RpoE10 negatively regulates flagellar motility of A. brasilense by controlling the expression of several genes involved in the synthesis of lateral flagella. Although alternative sigma factors σ⁵⁴ (RpoN) and σ²⁸ (FliA) regulate the expression of flagellar genes in several bacteria (12), including A. brasilense (30), the role of an ECF41-type σ factor in controlling the biogenesis of lateral flagella has not previously been demonstrated. R. sphaeroides and B. licheniformis both harbor an ECF41-type σ factor, but control of any distinct phenotype by this factor has not been shown (31, 32). The only target promoter of the ECF41 σ factors in these two bacteria is the gene or operon encoding the ECF41 σ factor itself (31). One reason for the inability of ECF41 σ factor mutants of R. sphaeroides and B. licheniformis to show any phenotype (39, 40) might be their inability to synthesize lateral flagella, as they possess only the polar flagellum. M. tuberculosis also produces two ECF41 σ factors (σ^I and σ^J), but the gene encoding σ^I is the only known target of σ^J (33). Azospirillum amazonense produces only a polar flagellum and hence encodes neither proteins involved in the biogenesis of lateral flagella nor an ECF41-type σ factor (41).

Similar to the case for the previously described ECF41 σ factors of R. sphaeroides and B. licheniformis, the gene encoding RpoE10 of A. brasilense is not organized together with an anti-σ factor-encoding gene. Instead, it is located downstream of a gene encoding an antibiotic monooxygenase (abm). Here, we show that the abm gene is cotranscribed with rpoE10. Linkage of an antibiotic monooxygenase gene with an ECF41 σ-encoding gene has not been described in any other bacterium to date. The genes encoding ECF41 σ factors are genomically associated with the genes encoding carboxymuconolactone decarboxylases, oxidoreductases, or epimerases (COE) in other bacteria, but the functional connection between COE genes and ECF41 σ factor is not known (31). Similarly, the functional connection between antibiotic monooxygenase and ECF41 σ factor in A. brasilense is also intriguing. Attempts to study the ability of ECF41 σ factor to activate the promoter of its adjacent gene in R. sphaeroides and B. licheniformis showed that full-length ECF41 σ factor was highly inefficient in activating the target promoter and that truncation of the ECF41 σ factor up to the DGGGR motif of the C-terminal extension led to only about 2-fold activation of their promoters (32). Further truncation of the C-terminal extension in R. sphaeroides and B. licheniformis led to the complete loss of ECF41 σ factor activity. In A. brasilense, we also show that full-length RpoE10 has negligible ability to activate the abm promoter. In contrast, truncation of the C terminus of RpoE10 up to DGGGR motif (DEL1) increased the abm promoter activity by almost 15-fold. The deletion of the C-terminal extension of RpoE10 including the DGGGR motif (DEL2) led to the complete loss of its ability to activate its target promoter. Thus, the distal part of the C-terminal extension of RpoE10 inhibits its own activity.

A dramatic difference in the size of haloes formed by A. brasilense Sp245 and the rpoE10::km mutant on 0.3% agar in MM medium but the lack of any major difference in the size of their haloes on NB medium containing 0.6% agar indicated that the inactivation of rpoE10 in A. brasilense did not affect the swarming motility but enhanced its swimming ability. It seems that under conditions suited for swarming, there may not be any difference in the formation of lateral flagella between A. brasilense Sp245 and the rpoE10::km mutant. However, under conditions suited for swimming, formation of lateral flagella improves the swimming ability of the rpoE10::km mutant. Thus, the ability of the rpoE10::km mutant to form a considerably larger halo in 0.3% agar, the increased number of lateral flagella, and elevated levels of Laf1 flagellin in this mutant in the liquid medium indicate that RpoE10 plays a negative role in the synthesis of lateral flagella and swimming motility in liquid or semisolid medium. Although several bacteria that possess a dual flagellar system use polar flagella exclusively for swimming in liquid media and lateral flagella for swarming on surfaces, A. brasilense Sp245, like Bradyrhizobium diazoefficiens, seems to use both flagellar systems to increase its swimming efficiency in media of low viscosity (10, 42). A. brasilense is also considered a robust swarmer (10) because of its ability to use lateral flagella for swimming through more-viscous media (8).

Comparison of the transcriptome and proteome of the rpoE10::km mutant with those of its parent A. brasilense Sp245 showed that most of the core set of genes and proteins involved in the biogenesis of lateral flagella were negatively regulated by RpoE10. These proteins included flagellin (Laf1), hook-capping protein (FlgD), and those involved in the formation of the hook-filament junction (FlgK), hook (FlgE), rod (FlgB, FlgC, FlgG, and FlgF), L ring (FlgH), C ring (FliG, FliM, and FliN), and the export apparatus (FlhA, FlhB, FliI, FliP, FliR, and FliQ). laf1 is the first gene of the tricistronic laf1-flaF-flbT operon in A. brasilense Sp245. Along with Laf1, FlaF and FlbT are also upregulated in the rpoE10::km mutant. Both of the latter proteins are known to control a regulatory checkpoint in Caulobacter crescentus and B. melitensis that links hook assembly with the translation of flagellin (43–45). The upregulated proteins in the rpoE10::km mutant also include FliL, an inner membrane protein of the basal body of the flagellum that interacts with the motor components of the flagellar motor to enable effective swimming against fluids with increased viscosity (46, 47). FliL is thought to sense the torque applied to the basal body due to the alterations in medium viscosity. Two paralogs of FliL are found in B. diazoefficiens; one is required for swimming and the other for controlling the synthesis of lateral flagella (47).

We noticed a striking resemblance between the mechanism by which flagellar genes and motility are regulated by the ECF σ factor RpoE1 of B. melitensis and regulation of RpoE10 of A. brasilense Sp245. Inactivation of rpoE1 in B. melitensis leads to overproduction of the flagellar hook protein FlgE, suggesting that the flagellar synthesis in B. melitensis is downregulated by RpoE1 (48). The rpoE1 mutant of B. melitensis also overexpressed flagellar genes fliF, fliC, flaF, and flbT, demonstrating the role of RpoE1 as a flagellar repressor (48). Increased promoter activity of the flagellar master regulator FtcR in the rpoE1 mutant of B. melitensis also suggested that RpoE1 acts upstream of ftcR (48). Since expression of RpoE1 downregulates the expression of flagellar genes and expression of FtcR upregulates the expression of flagellar genes in B. melitensis, the repression of flagellar genes was thought to be due to the repression of ftcR. Based on these observations, RpoE1 was thought to act as a repressor of the flagellar system in B. melitensis (48).

The location of the gene encoding LafR1 (a homolog of FtcR) in the vicinity of rpoE10 and the lateral flagellar gene cluster suggested that LafR1 might be involved in the synthesis of lateral flagella in A. brasilense. Loss of motility in the lafR1::km mutant and its restoration via complementation with the lafR1 gene indicated that LafR1 positively affects the motility of A. brasilense. Further, a low level of expression of Laf1 in A. brasilense Sp245 and the lafR1::km mutant and a distinct increase in Laf1 content upon expression of lafR1 showed that LafR1 is a positive regulator of Laf1. The strong upregulation of both lafR1 and laf1 in the rpoE10::km mutant indicated a negative effect of RpoE10 on the expression of LafR1. Since σ factors are not known to act as negative regulators, it is likely that RpoE1 of B. melitensis encodes a repressor which binds upstream of ftcR to keep its expression repressed. RpoE10 might exert its negative effect indirectly. RpoE10 may regulate the expression of a repressor, which keeps the expression of lafR1 repressed by binding to a repressor-binding site in the upstream region of lafR1. An absence of such a repressor in the rpoE10 mutant is expected to induce the expression of LafR1 and other flagellar proteins, including Laf1. Keeping in view the possible role of the C-terminal extension of the ECF41 σ factor as an anti-sigma factor (18, 19) that may respond to some environmental signal to bring about change in its conformation, RpoE10 might assume two alternative conformations, one allowing its activity and the other preventing it. It is likely that the viscosity of the medium affects the activity of the RpoE10, leading to differences in the synthesis of lateral flagella and the consequent motility. Identification of the repressor of lafR1 and elucidation of the mechanism by which the C-terminal extension of RpoE10 senses the viscosity of the medium will lead to a better understanding of novel mechanisms of regulation of the biogenesis of lateral flagella in A. brasilense Sp245.

MATERIALS AND METHODS

Bacterial strains, plasmids, chemicals, and growth conditions.

E. coli DH5α and E. coli S17.1 were grown in LB medium at 37°C. A. brasilense Sp245 was grown in MM medium or LB medium at 30°C. All the chemicals used for culturing bacteria were from Hi-media (Mumbai, India), chemicals used in stress assays were purchased from Sigma-Aldrich, and enzymes used for DNA manipulation and cloning were from New England Biolabs. All strains and plasmids used in this work are listed in Table 4. All primers used in this work are listed in Table 1.

TABLE 4.

Bacterial strains and plasmids used

Bacterial strain or plasmid	Relevant properties	Reference or source
Strains
E. coli
DH5α	ΔlacU169 hsdR17 recA1 endA1 gyrA96 thi-1 relA1	Gibco-BRL
S17-1	Smr recA thi pro hsdRRP4-2(Tc::Mu; Km::Tn7)	49
A. brasilense
Sp245	Wild-type strain	50
Sp245 rpoE10::km	Sp245 mutant having inactivation of the rpoE10 gene by the insertion of a kanamycin resistance gene cassette from pUC4K	This work
Sp245 lafR1::km	Sp245 mutant having inactivation of the lafR1 gene by the insertion of a kanamycin resistance gene cassette from pUC4K	This work
Sp245 rpoN::km	Sp245 mutant having inactivation of the rpoN gene by the insertion of a kanamycin resistance gene cassette from pUC4K	51
Plasmids
pSUP202	ColE1 replicon, mobilizable, suicide vector for A. brasilense, Apr Cmr Tcr	49
pUC4K	Vector containing kanamycin resistance gene cassette	GE Healthcare
pCZ750	pFAJ1700 containing the KpnI-AscI lacZ gene from the pCZ367 plasmid; Tetr Amp	52
pAPD1	rpoE10 disruption plasmid harboring kanamycin resistance gene (apt) used for generating the rpoE10::km mutant	This work
pAPD2	lafR1 disruption plasmid harboring kanamycin resistance gene (apt) used for generating the lafR1::km mutant	This work
pMMB206	Cmr; broad-host-range, low-copy-number expression vector	53
pAPD3	laf1 promoter of A. brasilense Sp245 cloned in pCZ750 vector	This work
pAPD4	fliB promoter of A. brasilense Sp245 cloned in pCZ750 vector	This work
pAPD5	fliG promoter of A. brasilense Sp245 cloned in pCZ750 vector	This work
pAPD6	lafR1 promoter of A. brasilense Sp245 cloned in pCZ750 vector	This work
pAPD7	lafR2 promoter of A. brasilense Sp245 cloned in pCZ750 vector	This work
pAPD8	rpoE10 gene of A. brasilense Sp245 cloned in pMMB206 vector	This work
pAPD9	rpoE10DEL1 gene of A. brasilense Sp245 cloned in pMMB206 vector	This work
pAPD10	rpoE10DEL2 gene of A. brasilense Sp245 cloned in pMMB206 vector	This work
pAPD11	lafR1 gene of A. brasilense Sp245 cloned in pMMB206 vector	This work
pAPD12	abm-rpoE10 promoter of A. brasilense Sp245 cloned in pCZ750 vector	This work

Gene organization and sequence comparisons.

The genomic organization of rpoE10 was examined by analyzing the arrangement of the flanking genes to construct a genetic map using Vector NTI software (Invitrogen). The sequences of RpoE10 orthologs from other bacteria and RpoE paralogs present in A. brasilense were retrieved from the NCBI database (https://www.ncbi.nlm.nih.gov/) and RpoE10 homologs at the EBI server (http://www.ebi.ac.uk) and aligned by using ClustalW software. A DNA pattern search tool available at RSAT-DNA-pattern server (54, 55) was used to identify putative RpoE10-binding sites upstream of the protein-coding genes in the A. brasilense Sp245 genome. A consensus pattern of ECF41 binding sites, TGTCAC-N_16/17/18-T/CGTC, was used as a template for an in silico search of similar motifs in the promoter regions on the six extrachromosomal replicons and one chromosomal replicon in the A. brasilense Sp245 genome.

Insertional inactivation of rpoE10 and lafR1.

Primers were designed to amplify the rpoE10 open reading frame (ORF) along with its flanking region in two parts, amplicons A and B. Appropriate restriction sites were introduced in the primers for cloning of amplicons in vector pSUP202. Amplicon A, which consists of 21 nucleotides of the 5′ region of rpoE10 with its upstream flanking region up to 900 bp, was amplified using primers RpoE10_AFP and RpoE10_ARP. Amplicon B, which consists of 21 nucleotides of the 3′ region of rpoE10 with its downstream flanking region up to 900 bp, was amplified using primers RpoE10_BFP and RpoE10_BRP. The two amplicons (A and B) were ligated into pSUP202, followed by insertion of a kanamycin resistance gene cassette (derived from pUC4K) between amplicons A and B to disrupt the rpoE10 gene in pSUP202. The rpoE10 disruption plasmid (pAPD1) was mobilized in A. brasilense Sp245 via Escherichia coli S17.1. Exconjugants were selected on kanamycin plates, and the knockout mutant (rpoE10::km mutant) was confirmed by PCR amplification using gene-specific primers from rpoE10 (56). A lafR1::km mutant was also generated by inserting a kanamycin resistance gene cassette in between the two amplicons of lafR1 flanking sequences cloned in pSUP202 (pAPD2) using the above-described method.

Cloning of rpoE10 and its deletion derivatives in a low-copy-number, broad-host-range expression vector, pMMB206.

To examine the effect of expression of a wild copy of the rpoE10 gene in the rpoE10::km mutant, a wild-type copy of the rpoE10 gene was supplied to the rpoE10::km mutant by cloning the entire coding region of rpoE10 in an expression vector (57). The gene encoding RpoE10 was amplified by PCR using DreamTaq DNA polymerase (Fermentas) and primer pair RpoE10-F′ and RpoE10-R′, having EcoRI and HindIII restriction overhangs in their 5′ ends, respectively. The gel-purified PCR product was digested using EcoRI and HindIII, purified again, and ligated with compatible ends downstream of the IPTG-inducible lacUV5 promoter region in a broad-host-range expression vector, pMMB206. The resulting plasmid (pAPD8) was conjugatively mobilized into A. brasilense, and exconjugants were selected on plates containing chloramphenicol. The RpoE10 deletion derivatives were cloned using the same method described above, and recombinants were named pAPD9 (DEL1) and pAPD10 (DEL2). The gene encoding LafR1 was PCR amplified by using LafR1-F and LafR1-R, having EcoRI and HindIII restriction overhangs in their 5′ ends, respectively, and ligated in the pMMB206 expression vector, and recombinants were named pAPD11.

Determination of transcription start site by 5′ RACE.

The 5′ end of the mRNA of rpoE10 was identified by 5′ RACE as described earlier (56). Briefly, 1 μg RNA was reverse transcribed to cDNA using gene-specific reverse primer 1 (GSP1), followed by cDNA purification and poly(dA) tailing. Poly(dA)-tailed cDNA was then PCR amplified using GSP2 and oligo(dT) anchor primers. The amplicons so obtained were further used as a template in the next round of nested PCR using anchor and GSP3 primers. The final PCR product was then cloned in pGEM-T Easy vector (Promega), and nucleotide sequences were determined by the chain termination method.

RNA purification and RNA-seq.

Cultures of A. brasilense and the rpoE10::km mutant were inoculated in MM medium and allowed to grow up to an optical density at 600 nm (OD₆₀₀) of ∼1. Cultures were harvested for RNA extraction by the TRIzol method, followed by treatment with DNase I (NEB) for 30 min at 37°C and heat inactivation at 65°C for 10 min. The quality and quantity of RNA were checked by denaturing formaldehyde-agarose gel electrophoresis (58) and by using a NanoDrop (ND-1000), respectively. RNA samples were subjected to standard paired-end RNA-Seq library preparation and sequencing using the Illumina platform. Samples were sequenced using 101-bp paired-end module sequencing. Sequencing reads obtained for each library were subjected to quality control using the NGSQC tool kit (59) to get high-quality (HQ) filtered reads. Gene ontology (GO) annotation was assigned using UniProt (http://www.uniprot.org/), and annotation results were categorized. Fold expression of transcripts for each condition was further obtained by calculating the log₂ of reads per kilobase per million (RPKM) values. The fold change of transcripts in mutants was calculated by subtracting log₂ the expression value of the mutant sample from that of the standard sample. Transcripts with at least a 2-fold change (P ≥ 2.0) were considered differentially expressed. The P values for the statistical significance of the fold change were adjusted for multiple testing with the Benjamini-Hochberg correction for controlling the false-discovery rate, accepting a maximum of 2% false discoveries (P ≥ 2.0).

Identification of differentially expressed proteins.

Cells were pelleted and processed for proteomic analysis. Protein samples (100 μg/25 μl) were taken, reduced with 5 mM Tris(2-carboxyethyl)phosphine hydrochloride (TCEP), alkylated with 50 mM iodoacetamide, and then digested with trypsin (1:50 trypsin/lysate ratio) for 16 h at 37°C. The digests were cleaned using a C₁₈ silica cartridge to remove salt and then dried in a SpeedVac. The dried pellet was resuspended in buffer A (5% acetonitrile, 0.1% formic acid). All the experiments were performed using the EASY-nLC 1000 system (Thermo Fisher Scientific) coupled to Thermo Q-Exactive equipped with a nano-electrospray ion source. Peptide mixtures (1.0 μg) were resolved using a 60-cm Viper column (360-μm outer diameter, 75-μm inner diameter, 10-μm tip) filled with 3.0 μm of C₁₈ resin (Dr. Maisch GmbH, Germany). The peptides were loaded with buffer A and eluted with a 0 to 40% gradient of buffer B (95% acetonitrile, 0.1% formic acid) at a flow rate of 300 ml/min for 100 min. MS data were acquired using a data-dependent top10 method dynamically choosing the most abundant precursor ions from the survey scan. Raw files generated after processing of the samples were analyzed with Proteome Discoverer (v2.2) against the UniProt A. brasilense reference proteome database. For SEQUEST and AMANDA searches, the precursor and fragment mass tolerances were set at 10 ppm and 0.5 Da, respectively. Carbamidomethyl on cysteine was considered a fixed modification and oxidation of methionine and N-terminal acetylation were considered variable modifications for database searches. Both the peptide spectrum match and protein false-discovery rate were set to 0.01.

Construction of laf1::lacZ, flgB::lacZ, fliG::lacZ, lafR1::lacZ, and lafR2::lacZ mutants.

The DNA regions located upstream of the start codons of the abm (300 bp), laf1 (450 bp), flhA (350 bp), flgB (450 bp), fliG (470 bp), lafR1 (450 bp), and lafR2 (400 bp) genes were PCR amplified. The PCR conditions were as follows: initial denaturation at 95°C for 2 min, denaturation 95°C for 30 s, annealing 45°C for 30 s, and extension at 72°C for 30 s up to 30 cycles and a final extension at 72°C for 5 min. Primers used for these amplifications are described in Table 1. The amplicons were inserted in the pCZ750 vector (52) using XbaI and HindIII to construct promoter::lacZ transcriptional fusions. Constructs were confirmed by sequencing and mobilized into A. brasilense and the rpoE10::km mutant via biparental conjugation using E. coli S17-1 as the donor strain. Exconjugants were selected on tetracycline-containing plates.

β-Galactosidase assay.

Cultures of A. brasilense Sp245 and its mutants harboring the laf1::lacZ, flgB::lacZ, fliG::lacZ, lafR1::lacZ, or lafR2::lacZ fusion were inoculated in three flasks each containing 25 ml MM medium and allowed to grow up to an OD₆₀₀ of ∼1. Cells were pelleted at 6,000 rpm and resuspended in lysis buffer (50 mM phosphate buffer [pH 7.0], 0.1% SDS, 0.27% β-mercaptoethanol, and 100 μl chloroform). The β-galactosidase activity in the supernatant of the cell lysate was measured as described previously (60) and expressed as Miller units by using the following formula: 1,000 × (OD₄₂₀ × 1.75 − OD₅₅₀)/time of reaction (in minutes) × volume of culture assayed.

RT-PCR.

cDNA was synthesized from 1 μg RNA using the Qiagen RNase minikit. The purity of RNA was checked by performing PCR with RNA samples and genomic DNA using specific primers to amplify a housekeeping gene (rpoD, encoding σ⁻⁷⁰ factor). Amplification of an rpoD-specific amplicon with DNA and absence of any amplicon with the purified RNA confirmed that the extracted RNA was free from any DNA contamination (data not shown). Relative expression of genes was quantified by real-time PCR using SYBR green I (Roche) in the Light Cycler 480 II instrument. The real-time PCR was carried out according to the manufacturer’s instructions (Roche) as follows. The real-time PCR mixture consisted of 5 μl of 2× Light Cycler 480 SYBR green I, 0.5 μM each primer, and 1 μl (2 to 5 ng) of cDNA. The cycling conditions comprised an initial incubation step at 95°C for 5 min, followed by 45 cycles of amplification for 10 s at 95°C, 10 s at 62°C (single acquisition), and 12 s at 72°C. The final cooling step was performed at 40°C for 30 s. The housekeeping gene (rpoD) primers RpoD _RTF and RTR were used for real-time PCR to set an endogenous control (61).

Analysis of flagellar proteins.

Cultures of A. brasilense and the rpoE10::km mutant were grown up to mid-log phase, and an equal amount of culture was vortexed vigorously for 5 min. The supernatant was decanted into fresh tubes, and proteins in the cell-free supernatant were precipitated by adding 1.5% PEG 6000 and 0.75% NaCl and incubating at 4°C for 2 h. This solution was then centrifuged at 13,000 × g for 40 min, the supernatant discarded, and the pellet dried at room temperature and then dissolved in phosphate-buffered saline (PBS). This PBS suspension was then mixed with 2× loading dye, and sample was loaded for 12% SDS-PAGE (62). The excised protein band was digested with trypsin as per the manufacturer’s instructions (GE Healthcare, USA). The mass of the resulting peptides was analyzed by MALDI-TOF/TOF (ABI USA). The peptide mass list obtained from the analyzer was then used for MASCOT search in the NCBI database to identify the protein.

Motility assay.

Cultures of A. brasilense and the rpoE10::km mutant were grown overnight, and secondary cultures were inoculated in MM medium and grown to an OD₆₀₀ of ∼1. Equal-cell-density cultures of both the strains were centrifuged at 10,000 × g for 2 min at 4°C. The supernatant was discarded and the pellet washed twice with 0.85% saline. The pellet was then resuspended in 100 μl of MM medium. For comparing swarming motility, a 2-μl drop of each strain was placed in the center of NB containing 0.6% agar in triplicates and incubated at 30°C in a humidified incubator for 72 h. For comparing swimming motility, a 2-μl drop of each strain was placed on MM medium with 0.3% agar in two independent experiments in triplicate. The diameter of the halo on each plate was measured after 72 h.

Electron microscopy of flagella.

Freshly streaked cells of A. brasilense Sp245 and its rpoE10::km mutant were inoculated into 5 ml MM medium and incubated at 30°C with shaking. The next day, the cultures were diluted 100-fold into 20 ml of the same medium and further allowed to grow to late log phase. Cultures (1 ml) were harvested at 1,500 rpm for 6 min (to avoid breakage of flagella), and the resulting pellet was resuspended in 200 μl of Milli Q water. Carbon-coated copper grids were glow discharged for 50 to 60 s at 20-mA current using the GloQube glow discharge system (Quorum Technologies). Samples (3.0 μl) were adsorbed on freshly glow-discharged carbon-coated grids for 20 min. Excess sample was taken out using a filter paper, and negative staining was performed using 2% uranyl acetate. The images were captured at room temperature at a magnification of ×26,500 using a 120-kV Tecnai T12 transmission electron microscope equipped with side-mounted Olympus Velita (2 K × 2 K) charge-coupled device (CCD) camera.