The Azotobacter vinelandii AlgU regulon during vegetative growth and encysting conditions: A proteomic approach

Abstract

In the Pseduomonadacea family, the extracytoplasmic function sigma factor AlgU is crucial to withstand adverse conditions. Azotobacter vinelandii, a closed relative of Pseudomonas aeruginosa, has been a model for cellular differentiation in Gram-negative bacteria since it forms desiccation-resistant cysts. Previous work demonstrated the essential role of AlgU to withstand oxidative stress and on A. vinelandii differentiation, particularly for the positive control of alginate production. In this study, the AlgU regulon was dissected by a proteomic approach under vegetative growing conditions and upon encystment induction. Our results revealed several molecular targets that explained the requirement of this sigma factor during oxidative stress and extended its role in alginate production. Furthermore, we demonstrate that AlgU was necessary to produce alkyl resorcinols, a type of aromatic lipids that conform the cell membrane of the differentiated cell. AlgU was also found to positively regulate stress resistance proteins such as OsmC, LEA-1, or proteins involved in trehalose synthesis. A position-specific scoring-matrix (PSSM) was generated based on the consensus sequence recognized by AlgU in P. aeruginosa, which allowed the identification of direct AlgU targets in the A. vinelandii genome. This work further expands our knowledge about the function of the ECF sigma factor AlgU in A. vinelandii and contributes to explains its key regulatory role under adverse conditions.

Introduction

Azotobacter vinelandii, a member of the Psedomonadaceae family, is a free-living bacterium having a strict aerobic metabolism during vegetative growth with the particular capacity of fixing nitrogen in aerobiosis [1]. In contrast to Pseudomonas species, Azotobacter undergoes a differentiation process that culminates with the formation of cysts able to resist desiccation [2,3]. During A. vinelandii differentiation, several morphological and metabolic changes are observed, including the reduction in nitrogen fixation, the switch to anaerobic metabolism, or the loss of the flagella. It has also been reported that during this process, the bacterium replaces the phospholipids of the membrane by aromatic lipids, called alkylresorcinols (ARs) and alkyl pyrones [4], while granules of the polyester poly-hydroxy butyrate (PHB) accumulate in the central body as a reservoir of carbon and energy. The cell is surrounded by two layers, which are mainly composed of alginate, proteins, and ARs [2,3]; these layers are essential for the desiccation resistance of the mature cyst [5].

Several regulators have been identified to be essential for successful cyst formation in A. vinelandii. One of these regulators is the sigma factor AlgU, belonging to the family of extracytoplasmic function (ECF) sigma factors, and homolog to the stress response RpoE sigma factor of Escherichia coli [5,6]. These types of sigma factors induce gene expression in response to specific environmental stimuli [7]. The activity of ECF sigma factors is often controlled by a cytoplasmic membrane-bound anti-sigma factor that sequesters the ECF sigma factor [8]. Under inducing conditions, the ECF sigma factor is released, allowing it to interact with the RNA polymerase, activating gene expression. In P. aeruginosa, AlgU is normally sequestered by its cognate anti-sigma factor MucA. Proteolysis of MucA is initiated by the AlgW protease located in the periplasm, followed by cleavage by the MucP protease located in the inner membrane. MucA degradation is completed in the cytoplasm by AAA+ proteases such as ClpXP [9–16]. One of the signals triggering this proteolytic cascade is the accumulation of misfolded proteins in the periplasm. In P. aeruginosa AlgU directs RNA polymerase to activate the expression of a plethora of genes, including the alginate biosynthetic genes, encoded in an operon headed by algD [17].

In A. vinelandii, AlgU is also essential for expressing the alginate biosynthetic genes algD and algC [5,18–21]. In addition, AlgU was shown to be necessary for oxidative stress resistance [5] and for expression of cydR [22], encoding a repressor of cydAB genes (required for aerotolerant nitrogen fixation) and flhDC, which encode master regulators of flagella biogenesis [22].

To elucidate the global molecular changes that occur during encystment, in a previous work we determined the proteome of the A. vinelandii cells undergoing differentiation [23]. We identified proteins differentially expressed during encystment induction with respect to vegetative cells. We found modifications in the abundance of proteins involved in nitrogen fixation, flagella synthesis and cell division, which agrees with the biochemical, morphological and physiological changes known to take place during this process [23].

In the current work, proteins under the control of the sigma factor AlgU in vegetative or encysting conditions were identified using a proteomic approach. They include proteins for processes previously known to be dependent on AlgU, but our study also revealed that AlgU was necessary to produce ARs and trehalose or to positively regulate stress resistance proteins such as OsmC, MdaB, and LEA-1. A position-specific scoring matrix (PSSM) was generated based on the consensus sequence recognized by AlgU in P. aeruginosa, allowing it to confirm some of the genes positively regulated by AlgU in A. vinelandii.

Materials and methods

Strains and cultivation conditions

The A. vinelandii wild-type AEIV strain (also named E strain) [24] and its algU or mucA derivative mutants, named AEalgU and AEA8 [5,21], respectively, were used in this study. A. vinelandii cells were routinely cultivated in Burk’s medium; sucrose (20 gL^-1) was used as a carbon source (Burk’s-sucrose medium) [1]. The composition of the culture medium has been previously reported [25].

For encystment induction, the A. vinelandii wild-type strain and its algU derivative mutant AEalgU, was cultured for 48 h in Burk’s-sucrose medium. The cells were harvested and washed three times with Burk’s solution (Burk’s medium without carbon source). The cells were then resuspended in fresh medium supplemented with 0.2% v/v of n-butanol as the sole carbon source and incubated for 48 h at 30°C.

Analytical methods

Protein concentration was quantitated as described previously [26]. The activity of β-glucuronidase in cells grown in liquid medium was determined as reported before [27] with some modifications; A. vinelandii cells were permeabilized using 0.013%(W/V) lysozyme and incubating at 37°C/5 min, followed by the addition of 0.13% (V/V) Triton before the enzymatic assay. One U corresponds to 1 nmol of O-nitrophenyl-β-D-glucuronide hydrolyzed per minute per μe of protein. Quantification of ARs and ARs qualitative visualization on agar plates was conducted using the Fast Blue colorimetric assay [28]. Details of the adapted methods for A. vinelandii are described elsewhere [29]. All experiments were conducted in triplicates; the results presented are the averages of the independent runs. Statistical analysis was carried out using a Student’s t-test (p = 0.05).

Construction of an algC-gusA transcriptional fusion

A fragment of 306 bp, spanning a region from nucleotides -342 to -36 with respect to the ATG translation initiation codon of algC, was PCR amplified using oligonucleotides algC-gusF and algC-gusR (S1 Table). This fragment was subsequently cloned into plasmid pUMATcgusAT [30] as an XbaI-EcoRI fragment (the restriction sites were included in the corresponding oligonucleotides), thus generating plasmid pCN62. Strain AEIV was transformed with plasmid pCN62, previously linearized with NdeI endonuclease, and transformants Tc^r were selected. The strain generated, carrying an algC-gusA transcriptional fusion, was named CNL35.

Cell fractionation and protein sample preparation

The proteomic analysis was conducted using soluble protein extracts derived from three independent cultures (biological replicates) from the wild-type strain and from mutant algU, grown under vegetative and encysting conditions. Then, the AlgU regulon was obtained by comparing the proteins expressed in the algU mutant with those expressed in the wild-type strain in both conditions, vegetative and encystment [23]. Cells from vegetative (24 h/Burk´s-sucrose medium) and encysting conditions (48 h/ Burk´s-butanol medium) were harvested by centrifugation at 1900 x g for 10 min at 4°C, washed, and resuspended in 10mM of sodium-phosphate buffer of pH 7.4. Cell fractionation and sample preparation were conducted as described [23].

LC-MS/MS analysis and identification/quantification of proteins

A detailed description of the analysis and protein quantification procedures has been previously reported [23]. LC-MS/MS analysis was performed at the IRCM Proteomics Discovery Platform of the Montreal Clinical Research Institute as described previously. Raw files obtained from Orbitrap Q-Exactive spectrometer were acquired using Mascot 2.3 (Matrix Science) against a database of A. vinelandii DJ strain from NCBI (taxon identifier 322710). Data analysis was performed using Scaffold (http://www.proteomesoftware.com/products/scaffold/download/) [31]. Protein expression in the algU mutant was considered significantly different only if protein ratios differed more than two-fold with respect to the expression observed in the wild-type strain AEIV. The Student’s t-test was also done for each sample set having three different biological replicas and a threshold level of 0.05 was considered for selecting the proteins.

Protein functional classification analysis

Protein functional classification and KEGG pathways analysis of differentially modulated proteins was conducted using the Kyoto Encyclopedia of Genes and Genomes database (www.genome.jp/kegg) [32]. The String 10 database was used to determine protein–protein interaction network [33].

Quantitative analysis of mRNA levels

Strains AEIV, AEalgU and AEA8 were cultured in Burk´s-sucrose or in Burk´s-butanol medium for 24 or 48 h, respectively. Cells were collected by centrifugation, and the total RNA was extracted as described [34]. Details of DNA contamination removal, cDNA synthesis and qPCR amplification conditions are reported elsewhere [25]. qPCR assays were performed with a Light Cycler 480 II instrument (Roche), using the Maxima TM SYBR Green/ROX qPCR Master Mix (2X) kit (Thermo Scientific). The sequences of the primer pairs used are listed in S1 Table. Three biological replicates (independent cell cultures) were performed, with three technical replicates for each one. Similar results were obtained for the transcription of all measured genes in the repetitions. Relative mRNA transcript levels were determined in relation to gyrA (Avin15810) mRNA, as reported previously [25]. A non-template control of each reaction was included for each gene. The quantification technique used to analyse the generated data was the 2^-Δ,ΔCT method reported previously [35].

Identification of orthologous genes

Orthologous genes were defined using “bidirectional best hits” criteria where reciprocal best hits were identified by pairwise comparisons using BLAST (BLASTP version 2.12.0+) [36]. The results were filtered using a cut off E-value of 1x10⁵ and a query and subject coverage of at least 50%.

Operon predictions

Operon predictions were performed using the method previously reported [37], which is based on the intergenic distance between codirectional transcribed genes, and the functional relationship of their protein products defined in the STRING database [38].

In silico prediction of AlgU-dependent promoters

Experimentally determined AlgU-dependent promoters in P. aeruginosa (S2 Table) were used as a reference to construct a Position-Specific Scoring Matrix (PSSM) for the de novo motif detection using ad hoc developed PERL program [39–47]. Our PSSM was based on the tight consensus sequence [(-35) GAACTT-N16/17-(-10)TCtgA (highly conserved residues in bold capital letters; conserved residues in capital letters)] reported for most of the AlgU-dependent promoters. Our PSSM used only three different values: 1, 0.9, and 0.81 associated with invariant, well-conserved, and conserved nucleotides within the promoter sequences, respectively. We only considered as likely promoters those whose -10 and -35 consensus boxes were separated by 16 or 17 nt. For our AlgU-dependent promoter search, we scanned the first 250 nt upstream of every gene regardless of the gene position within their corresponding operons. This consideration allowed us to identify internal promoter sequences within coding regions. The promoter score was obtained by multiplying the values associated with the nucleotides of the region analysed, based on our PSSM in such a way that sequences with 100% conserved nucleotides obtained a score of 1, sequences with a substitution of one nucleotide in one position obtained a score of 0.9 and sequences with one substitution in a highly conserved region or two substitutions in conserved regions, obtained a score of 0.81. The search was performed for the P. aeruginosa PAO1 (tax id: 208964), A. vinelandii DJ (tax id: 322710), Azotobacter chroococcum Ac-8003 (tax id: 1328314) and Pseudomonas fluorescens SBW25 (tax id: 216595) genomes. The latter two served as comparative genomes for AlgU-predicted promoters in A. vinelandii and P. aeruginosa, respectively.

Results and discussion

The activity of AlgU increases upon encystment induction

The essential role of the sigma factor AlgU during A. vinelandii encystment has been clearly demonstrated in previous works from our laboratory [5,22]. As the activity of this type of sigma factor is highly regulated at the protein level, we reasoned that such activity would be increased upon encystment induction. To explore this assumption, we constructed a PalgC-gusA transcriptional fusion since expression of algC was previously shown to be under the direct control of AlgU [19]. The activity of AlgU was estimated during vegetative growth in Burk’s-sucrose liquid medium and during encysting conditions, induced with 0.2% n-butanol. As a negative control, this transcriptional PalgC-gusA fusion was tested in an algU^- genetic background. In vegetative conditions, at 24 h, the activity of AlgU reached the highest point. Thereafter the activity slightly decreased at 48 h (Fig 1A). As expected, we detected a basal level of PalgC activity in the negative control that did not significantly change along the growth curve. 24 h after induction of encystment, the activity of AlgU doubled and remained high for the following 24 h; then, it gradually dropped to levels like those observed at the end of vegetative growth (Fig 1B). Therefore, samples from 24 h for the vegetative condition and from 48 h for the encystment condition, were used for further analysis of the AlgU regulon.

Fig 1. Activity of the AlgU sigma factor in A. vinelandii.

A transcriptional PalgC-gusA fusion was used to estimate the activity of AlgU in the background of the wild-type strain AEIV (solid line) or in the algU mutant (dotted line). Cells were cultured in minimum Burk’s medium supplemented with sucrose (vegetative growth) or n-butanol (encystment-inducing conditions) as the sole carbon source for the indicated time. The bars of standard deviation from three independent experiments are shown.

The AlgU proteome during vegetative growth

To compare the total proteomic profile in cytoplasmic fraction of the algU mutant, we have analyzed the protein samples through Hybrid Quadrupole-Orbitrap Mass Spectrometer, which combines quadruple precursor ion selection with high-resolution, accurate-mass (HRAM) Orbitrap detection. The total expressed proteins in the absence of AlgU (S3 Table) were compared to the previously reported proteins present in the wild-type strain AEIV [23]. After evaluating the data generated from orbitrap, 126 proteins were found to be differentially expressed in the algU mutant, among which, 50 proteins were downregulated (S4 Table), and 76 proteins were upregulated (S5 Table), after 24 hours of vegetative growth, as compared to the wild-type strain.

With the aim of understanding the roles of the differentially expressed proteins due to the absence of AlgU under vegetative conditions, we mapped the genes encoding the differentially expressed proteins to their corresponding terms in the KEGG database. We identified 20 main pathways for the 126 differentially expressed proteins (Fig 2A). Besides the group of proteins of unknown function, the most represented groups corresponded to energy, carbohydrate and amino acid metabolism.

Fig 2. KEGG pathway enrichment of proteins of the AlgU regulon in A. vinelandii.

The analysis was conducted for the differentially expressed proteins in the absence of the sigma factor AlgU under vegetative (A) or encystment-induced (B) conditions. Positive and negative axes represent the numbers of up- or downregulated proteins, respectively.

Protein networks generated by String 10 software, identified 81 protein interactions (for a confidence interaction score of 0.7), for the 126 proteins. Interaction nodes include those for proteins related to central and lipid metabolism, flagella biogenesis, trehalose synthesis, among others (S1 Fig).

Analysis of proteins under the control of AlgU during vegetative conditions

The abundance of the corresponding mRNAs for 6 proteins found under the control of AlgU was evaluated by qPCR. We reasoned that accumulation of such mRNAs would be diminished in the algU mutant but would be increased in a mutant lacking the anti-sigma factor MucA, thus showing elevated AlgU activity. The results are presented in Table 1.

Table 1. Relative mRNA levels of some genes encoding proteins under the positive control of AlgU.

Protein name	Gene	locus Tag	Relative mRNA levels*
			algU/wt	mucA/wt
Vegetative conditions
OsmC-like protein	osmC	Avin_01800	0.09 ± 0.001	20 ± 1.4
Quinone NAD(P)H:oxidoreductase MdaB	mdaB	Avin_46260	1.35 ± 0.18	1.03 ± 0.2
Alginate lyase; Poly(beta-D-mannuronate) lyase	algL	Avin_10900	0.2 ± 0.01	7 ± 0.85
Transport-associated protein		Avin_43670	0.05 ± 0.007	102 ± 22
Fe-S protein assembly chaperone HscA	hscA	Avin_40360	0.57 ± 0.09	34 ± 4.5
Organic solvent tolerance ABC efflux transporter, substrate binding protein		Avin_12870	0.84 ± 0.15	0.87 ± 0.10
Encysting conditions
Hemerythrin HHE cation binding protein		Avin_00300	0.69 ± 0.09	107 ± 18
Conserved hypothetical protein		Avin_11100	0.076 ± 0.001	114 ± 23
Conserved hypothetical protein		Avin_11110	0.23 ± 0.02	25 ± 2.3
Acetoacetyl-CoA thiolase		Avin_16380	1.194 ± 0.22	5.7 ± 0.77
Malto-oligosyltrehalose trehalohydrolase	treZ	Avin_24900	0.91 ± 0.16	13 ± 3.0
Encystment and alginate biosynthesis response regulator; AlgR	algR	Avin_47610	0.54 ± 0.09	19 ± 1.8

* qPCR assay was conducted in triplicates. The mean of independent experiments (biological replicates) is shown along with the standard deviation.

As our key target was to study the proteins under AlgU regulation, we were very interested in the proteins which are most affected (missing) due to the algU mutation. We have found 18 proteins that completely disappeared due to the mutation in algU during vegetative growth as compared to wild type (Tables 2 and S4).

Table 2. Selected proteins under the positive control of AlgU under vegetative growing conditions.

No.	Protein	Gene	Fold change algU/wt	locus Tag	AlgU promoter prediction
					Type*	Sequence	Score
1	Nitrogen fixation protein orf9, ClpX	clpXnif	0	Avin_01700	TU hit	GCAGTTctaccaggtctgtaac.TCcgA	0.81
2	OsmC-like protein	osmC	0	Avin_01800	Motif	GAACTTgtccgtcccgctcccc.TCtgA	1
3	phosphoenolpyruvate carboxykinase	pckA	0.3	Avin_05450	Motif	GAACCTttgccggactggcggc.TCctA	0.9
4	alginate lyase; Poly(beta-D-mannuronate lyase	algL	0	Avin_10900	TU hit	GAACTAtttcggagaaagtatt.TCttA	0.9
5	GDP-mannose 6-dehydrogenase	algD	0	Avin_10970	Motif	GAACTAtttcggagaaagtatt.TCttA	0.9
6	Organic solvent tolerance ABC efflux transporter, substrate binding protein		0	Avin_12870	TU hit	GGACTCccgaggacttgatgaacTCcaA	0.81
7	Cytochrome b/b6	petB	0.2	Avin_13070	Motif	GAACCTgccggtacctccgcat.TCctA	0.9
8	Type I fatty acid synthease ArsA	arsA	0	Avin_29560
9	3-isopropylmalate dehydratase large subunit LeuC	leuC	0.4	Avin_34280	Motif	GAATTTgagttattccctgcaagTCccT	0.81
10	AmrZ-like DNA binding protein	amrZ	0	Avin_34410	Motif	GCACTCtataacaactcaccgg.TCgaA	0.81
11	Acetate kinase	ackA-1	0	Avin_34560
12	septum site-determining protein MinD	minD	0.1	Avin_35120	Motif	GAACTGgggcgaggccgtgcacaTCcaC	0.81
13	Phosphoribosylglycinamide formyltransferase	purT	0.3	Avin_39610	Motif	GGACGTacgcatgttgaaggaaaTCcc	0.81
14	Fe-S protein assembly chaperone HscA	hscA	0	Avin_40360	TU hit	GAGCTTttatggcccgctccgg.TCgaA	0.81
15	Transport-associated protein		0	Avin_43670	Motif	GAACTTaccaagtcccgagggagTCcaT	0.9
16	Quinone NAD(PH:oxidoreductase MdaB	mdaB	0	Avin_46260

* Genes showing predicted AlgU-dependent promoters (Motif) or genes located in an operon under the control of AlgU (TU hit) are indicated.

As anticipated, some proteins for alginate biosynthesis were found to be down-affected in the algU mutant, including the GDP-mannose 6-dehydrogenase (AlgD), which is the enzyme catalyzing the key step in alginate biosynthesis or AlgL, an alginate lyase that also serves as part of the multi-protein alginate-secretion complex [48–50]. As expected, the abundance of the algL mRNA in the algU mutant was 5-fold reduced, but it was increased 7-fold in the mucA genetic background, as revealed by qPCR (Table 1).

Besides these, the OsmC-like protein was found to be absent in the algU mutant. Osmotically inducible protein C (OsmC) is a protein found in E. coli during stress conditions such as salt stress [51] or in a medium of low osmotic pressure [52]. In oxidative stress OsmC-induced cells were found to be highly viable whereas an osmC mutant showed more sensitivity to butanol in the exponential growth phase and to H₂O₂ and butanol in the stationary phase [53], indicating that OsmC may play a role as a scavenger for specific ROS [51]. In P. aeruginosa OsmC is part of the AlgU regulon [41]. In A. vinelandii we have shown that expression of OsmC is completely dependent on AlgU during both, vegetative and encysting conditions (see below). Accordingly, the osmC mRNA was 10-fold reduced in the algU mutant, but 20-fold higher in the mucA mutant (Table 1). As in P. aeruginosa, transcription of osmC seems to be under direct regulation of AlgU as we found a potential AlgU promoter in its regulatory region (Table 2). Description of AlgU promoters’ prediction in the genome of A. vinelandii is detailed in the last two sections of Results.

The quinone oxidoreductase MdaB protein was absent in the algU mutant. This protein was first identified as a modulator of drug activity in E. coli [54]. Quinone oxidoreductases were shown to reduce quinone substrates via an intensive two-electron mechanism which play key roles to maintain a pool of reduced quinols that contribute to antioxidant defense in E. coli and P. aeruginosa [55–57]. MdaB of P. aeruginosa shares 70% identity with its orthologous in A. vinelandii. Accumulation of the mdaB mRNA was not diminished in the algU mutant (Table 1), nor affected in the mucA mutant implying that the effect of AlgU on the expression of MdaB might be post-transcriptionally.

An iron-sulfur protein assembly chaperone HscA was undetected in the algU mutant. HscA is an Hsp70 class molecular chaperone previously described in many bacteria, including E. coli, P. aeruginosa and A. vinelandii [58–61]. In A. vinelandii the hscA gene is a component of the isc operon which is responsible for Fe-S cluster biogenesis and helps in the maturation of [2Fe–2S] proteins [58]. Several genes of the isc operon, including hscA, were found to be essential for vegetative growth in the OP genetic background. The OP strain is a naturally occurring algU mutant. As we detected no HscA protein in the algU mutant derived from strain AEIV, it is likely to propose a basal expression level of the isc operon undetectable by our proteomic approach. Indeed, qPCR assay showed that in the algU mutant, the levels of the hscA mRNA were diminished by about 50%, but its accumulation increased 34-fold in the background of mutant mucA, indicating a positive dependence on AlgU (Table 1). This agrees with a predicted AlgU promoter directing the transcription of the hcsA containing operon (Table 2). Another important function of this operon is to defend against oxidative stress [61–64]. All the above three proteins (OsmC, MdaB and HscA) related to counteracting the stress response, and found under the control of AlgU, explain the previous observation of the reduced survival of an algU mutant in oxidative stress conditions [5].

Another protein absent in the algU mutant is an ABC efflux transporter protein (Avin12870). ATP binding cassette (ABC) efflux transporters proteins use energy to remove solvents from the cells to the outer medium. ABC transporters translocate a wide variety of substrates, including amino acids, peptides, ions, sugars, toxins, lipids and drugs in several bacteria including Pseudomonas and E. coli [65]. These proteins discharge the toxic compounds from the cell to the external medium which is a relevant mechanism in the solvent-tolerance of bacteria [66]. Of note, differences in accumulation of the Avin_12870 mRNA was not detected in either, the algU or the mucA mutant with respect to the WT strain (Table 1), even though an AlgU dependent promoter was detected for its transcriptional unit.

Our analysis also revealed a transport associated protein (Avin_43670) which has a BON domain. This domain was found in an OsmY protein of E. coli, that was reported to protect the cell against stress, especially during osmotic shock by contacting the phospholipid interfaces surrounding the periplasmic space [67]. In A. vinelandii the Avin_43670 transport-associated protein may also help to perform these functions by binding or interacting with the phospholipid membrane. The Avin_43670 gene shows a putative AlgU promoter implying a direct transcriptional regulation. In agreement with this, accumulation of the corresponding mRNA was diminished in the algU mutant, whereas it showed a strong upregulation, of about 102-fold in the mucA mutant.

The A. vinelandii AlgU regulon during encystment

As described in Materials and Methods, the proteome of AlgU was also determined at 48 h of encystment induction. The total expressed proteins in the algU mutant were obtained (S6 Table), and their abundance was compared to those of the wild-type strain [23]. We found 305 proteins downregulated (S7 Table) and 184 ones upregulated (S8 Table) in the algU genetic background. The vast number of proteins whose expression was affected in the absence of AlgU might reflect the central role of this sigma factor during A. vinelandii encystment.

Based on the KEGG database, the differentially expressed proteins due to the algU mutation are involved in several functions, including carbohydrate and amino acid metabolism, translation and signal transduction, among others (Fig 2B).

Protein interaction networks were generated by String 10 software (confidence cutoff of 0.7), for proteins downregulated (S2 Fig) or upregulated (S3 Fig) in the absence of AlgU. The number of interaction nodes for both data sets was significantly higher than expected, revealing a connection among the identified proteins. This analysis also confirmed the positive role of AlgU in oxidative phosphorylation, amino acid metabolism, synthesis of glycogen and trehalose or alginate production, among others, as revealed by the cluster of identified functional groups. Furthermore, a negative role of AlgU in amino acids and aminoacyl-tRNA biosynthesis was also identified (S2 and S3 Figs).

Analysis of proteins under the control of AlgU during encystment-induced conditions

After 48 h of encystment induction, 183 proteins were not detected in the algU mutant when compared to the wild-type strain (S7 Table). Strikingly, 50 of such proteins were strongly expressed during encystment in the wild-type strain, as compared to its own vegetative growth (Tables 3 and S7) [23], revealing AlgU targets specifically associated to this differentiation process. The expression levels of six of these proteins were also examined by qPCR (Table 1) and the results are discussed below.

Table 3. Selected proteins positively control by AlgU during encysting conditions.

No.	Protein	Gene	Fold change algU/wt	locus Tag*	AlgU promoter prediction
					Type**	Sequence	Score
1	Hemerythrin HHE cation binding protein		0	Avin_00300
2	OsmC-like protein	osmC	0	Avin_01800	Motif	GAACTTgtccgtcccgctcccc.TCtgA	1
3	CsbD-like stress response protein		0	Avin_02330
4	Non-heme chloroperoxidase (abhydrolase_1 family)		0	Avin_02370	Motif	GCACTAacgcccggcggcaaag.TCgtA	0.81
5	AlgC	algC	0.3	Avin_02910	Motif	GAACTCcgcgccgtcccggcca.TCcaA	0.9
6	RimK	rimK	0	Avin_05140
7	Polysaccharide export protein		0	Avin_05380
8	30S ribosomal protein S3	rpsC	0	Avin_06310	TU hit	GAAATGactattggtgtaatcggTCgtA	0.81
9	30S ribosomal protein S14		0	Avin_06380	TU hit	GAACTCgacagtatgctggttgcTCagA	0.81
10	30S ribosomal protein S5	rpsE	0	Avin_06420	TU hit	GAACTCgacagtatgctggttgcTCagA	0.81
11	Thiolase		0	Avin_07370
12	TPP-dependent dehydrogenase, E1 component alpha subunit, E1_dh family		0	Avin_10770
13	Alginate lyase; Poly(beta-D-mannuronate) lyase	algL	0.04	Avin_10900	TU hit	GAACTAtttcggagaaagtatt.TCttA	0.9
14	Alginate biosynthetic protein AlgK	algK	0	Avin_10940	TU hit	GAACTAtttcggagaaagtatt.TCttA	0.9
15	Alginate biosynthesis protein Alg44	alg44	0	Avin_10950	TU hit	GAACTAtttcggagaaagtatt.TCttA	0.9
16	GDP-mannose 6-dehydrogenase	algD	0	Avin_10970	Motif	GAACTAtttcggagaaagtatt.TCttA	0.9
17	LEA-1 protein	lea-1	0.2	Avin_11010
18	Conserved hypothetical protein		0	Avin_11100	Motif	GAACTTtcaaagatcgggcggatTCtaC	0.9
19	Conserved hypothetical protein		0.1	Avin_11110	Motif	GAACTCtatccggccatgcaag.TCgtA	0.9
20	Leucyl aminopeptidase	pepA	0	Avin_11650
21	MraZ-family protein	mraZ	0.3	Avin_13160
22	DNA topoisomerase I	topA	0	Avin_14470
23	Glycerophosphoryl diester phosphodiesterase		0	Avin_14680
24	UDP-Glycosyltransferase/glycogen phosphorylase		0	Avin_17290
25	Metallo-beta-lactamase family protein		0	Avin_17550
26	CHAD domain superfamily protein		0	Avin_17860
27	Conserved hypothetical protein		0	Avin_18160
28	Threonyl-tRNA synthetase	thrS	0	Avin_20420
29	Thiolase protein		0	Avin_22240
30	Aminotransferase class-III protein		0	Avin_22800
31	2-methylcitrate synthase	prpC	0	Avin_23220
32	Zinc-containing alcohol dehydrogenase superfamily		0	Avin_24840
33	Glycogen debranching enzyme	glgX	0.1	Avin_24860
34	4-alpha-glucanotransferase	malQ	0.1	Avin_24890
35	Malto-oligosyltrehalose trehalohydrolase	treZ	0	Avin_24900
36	Acyl-activating enzyme		0	Avin_25250
37	Imidazolonepropionase	hutI	0	Avin_26160
38	1,4-alpha-glucan branching enzyme	glgB	0	Avin_27990	TU hit	GAACTTtttccacgcatccgccaTCggA	1
39	NADH-quinone oxidoreductase, chain I	nuoI	0	Avin_28510	Motif	GAACTGgccgatggctaccata.TCgaA	0.9
40	ABC-type antimicrobial peptide transport system, ATPase component		0	Avin_29510
41	Type III polyketide synthase	arsC	0	Avin_29530	TU hit	GAACTGgtgatcaccgcgagtc.TCgaG	0.81
42	Type III polyketide synthase	arsB	0	Avin_29550	Motif	GAACTGgtgatcaccgcgagtc.TCgaG	0.81
43	Type I fatty acid synthase ArsA	arsA	0	Avin_29560
44	AmrZ-like DNA binding protein	amrZ	0	Avin_34410	Motif	GCACTCtataacaactcaccgg.TCgaA	0.81
45	Cation/acetate symporter ActP	actP	0	Avin_35030
46	Ribosomal protein L19	rplS	0	Avin_39530
47	Cysteine desulfurase IscS	iscS	0	Avin_40400	TU hit	GAGCTTttatggcccgctccgg.TCgaA	0.81
48	Ribosomal protein L27	rpmA	0	Avin_40780	TU hit	TAACTTtcctgggtatttccctgTCttA	0.81
49	Isochorismatase hydrolase		0	Avin_42350
50	FabI-like dehydrogenase/reductase		0	Avin_44040
51	glycogen phosphorylase		0	Avin_45650
52	Encystment and alginate biosynthesis response regulator; AlgR	algR	0.3	Avin_47610
53	Threonine dehydratase	ilvA	0	Avin_48270	Motif	GAACTCcagttcgacgacgttgtTCagA	0.9
54	Phosphate transport system regulatory protein PhoU	phoU	0.3	Avin_48560	TU hit	GAAATAtgcgtcgctctgaacg.TCggA	0.81
55	Putative glutamine synthetase	spuB	0	Avin_48860
56	HMG-CoA lyase-like protein		0	Avin_49510
57	Secreted mannuronan C-5 epimerase	algE3	0	Avin_51170
58	Secreted mannuronan C-5 epimerase	algE6	0	Avin_51230	Motif	GGACGTtctcgcccgctcctcttTCcaA	0.81
59	Calcium-binding protein		0	Avin_51240
60	Multidrug efflux pump membrane fusion protein		0	Avin_51550
61	Multidrug efflux pump RND-family transporter protein		0	Avin_51560

* Proteins detected to be induced during encystment are in italics.

** Genes showing predicted AlgU-dependent promoters (Motif) or genes located in an operon under the control of AlgU (TU hit) are indicated.

The abundance of several alginate biosynthetic proteins was reduced in the algU mutant, confirming the positive role of AlgU for alginate production during cyst formation. AlgD, AlgE3 and AlgE6 (Secreted mannuronan C-5 epimerase) were totally missing in the algU mutant along with ORF Avin51240, encoding a calcium-binding protein important for the activity of the mannuronan C-5 alginate epimerases. AlgL and AlgC were 0.05 and 0.3-fold downregulated as compared to wild type, respectively. The response regulator AlgR, which is essential for cyst formation [68], was also found to be 0.4-fold downregulated in mutant algU. The AlgU-dependent transcription of algR was further demonstrated by qPCR as algR mRNA levels were reduced in the algU mutant but were enhanced 19-fold in the mucA mutant (Table 1). In P. aeruginosa AlgU recognizes the promoter of algR for transcription initiation [42,43]. In A. vinelandii, however, the positive effect of AlgU on algR might be indirect as the algR promoter does not show consensus sequences recognized by this sigma factor (Table 3) [68].

The hemerythrin HHE cation binding protein (Avin_00300), strongly expressed during encystment of the wild-type strain [23,30], was undetectable in the algU mutant. This effect seems to be at the transcriptional level as the amount of the corresponding mRNA was reduced in the algU mutant, but it was 107-fold higher in the mucA mutant (Table 1). Hemerythrin is a non-heme, iron-containing protein that binds to oxygen [69]; it has been proposed to regulate interactions between cellular enzymes and oxygen or to be involved in the transport of oxygen within the cell [70,71]. However, its exact role in A. vinelandii differentiation remains unknown.

Other proteins involved in synthesizing trehalose such as GlgX, MalQ and TreZ (Avin_24860, Avin_24890 and Avin_24900, respectively), and previously shown to be induced during encystment, were also downregulated in the algU mutant. The control of these genes by AlgU might be complex, involving different layers of regulation, as the levels of treZ mRNA was not significantly reduced in the absence of this sigma factor, but it was 13-fold higher in the mucA genetic background. Late embryogenesis abundant (LEA) proteins conform a large family associated with resistance to abiotic stress. A. vinelandii cysts expresses LEA-1 protein essential for the survival of the differentiated cell in dry conditions and high temperatures [72]. Expression of LEA-1 was AlgU dependent, as its accumulation was 5-fold reduced in the algU mutant when compared to the wild-type strain.

In several Pseudomonas spp. and in E. coli, the protein RimK is involved in modifying the ribosomal protein RpsF, affecting the translation of several key genes necessary to survive adverse conditions [73,74]. In A. vinelandii RimK was previously shown to be strongly induced during encystment [23]. Our proteomics results indicate that its expression depends on AlgU as in its absence, RimK (Avin_05140) was undetectable. The reported role of RimK in controlling the activity of ribosomes may be related to the strong downregulation of multiple 50S and 30S ribosomal proteins in the absence of AlgU, highlighting the role of this sigma factor on protein translation (S7 Table and S2 Fig).

It is worth mentioning that the regulon detected for AlgU during encystment contains a total of 49 hypothetical proteins, which implies the existence of additional cellular processes yet to be defined under the control of this sigma factor. Two hypothetical proteins (Avin_11100 and Avin_ 11110), upregulated during encystment, were missing in the algU mutant. This effect seems to be at the transcriptional level as the amount of the corresponding mRNAs was negligible in the algU mutant, but it was strongly enhanced in the mucA genetic background (Table 1). This result agrees with the presence of an AlgU-dependent promoter in their regulatory regions (Table 3).

AlgU controls ARs production

Upon encystment, 95% of the cell membrane phospholipids are replaced by the phenolic lipids alkyl-resorcinols (ARs) and alkyl-pyrones [4]. Interestingly, expression of proteins ArsA, ArsB and ArsC for the synthesis of these lipids was totally suppressed in the algU mutant, implying that cell membrane phospholipids replacement does not occur in the absence of AlgU (Tables 3 and S7). Indeed, ARs in encystment-induced cells were not detected in the algU mutant, in contrast to the wild-type strain (Fig 3A). Accordingly, the synthesis of fatty acids seems to remain active in the algU mutant, as suggested by the up regulation of some proteins involved in this pathway (Avin_14930, Avin_15000, Avin_29050 and Avin_44250) (S8 Table and S3 Fig).

Fig 3. The production of ARs is impaired in the absence of AlgU.

A. Staining with Fast Blue B of ARs produced by strain AEIV (WT) and by its derivative algU mutant under encysting conditions. Cells were grown on Burk’s-butanol for 48 h before staining. B. ARs quantification under vegetative growth. The wild-type strains AEIV, ATCC 9046 and ATCC 12518, and the AEIV derivative mutant algU were cultivated in Burk’s-sucrose medium for 24 h prior to ARs extraction. The bars of standard deviation from three independent experiments are shown. The asterisks denote statistical significance (unpaired Student’s t-test, *P<0.05; **P<0.01) when compared to the wild-type strain AEIV.

Previous reports indicated the production of ARs in glucose aging cultures of A. vinelandii strain ATCC12837 [75]. Similarly, quantification of ARs production by the wild-type strain AEIV at 24 h of vegetative growth, confirmed the presence of these phenolic lipids (Fig 3B). The same was true for A. vinelandii wild-type strains ATCC 9046 and ATCC 12518. Of note, these two strains produced ARs in a lower amount when compared to strain AEIV. This result agrees with our proteomic data indicating the expression of ArsA in WT vegetative cells. The regulation of ARs by AlgU also occurs during vegetative growing conditions; ArsA was undetectable in mutant algU (Table 2) and correlated with the absence of these lipids in this genetic background (Fig 3B). Furthermore, q-PCR suggested that the positive effect of AlgU on ARs production occurs at the transcriptional level since mRNA accumulation of arsA and arsB was abrogated in the absence of AlgU under both, vegetative or encystment-induced conditions (0.0015 ± 1x10^-4 and 0.0008 ± 1.6x10^-5 for arsA and arsB, respectively), when compared to the wild-type strain.

Identification of AlgU binding motifs

To identify potential targets directly regulated by AlgU in A. vinelandii, AlgU-dependent promoters were predicted based on the consensus sequence recognized by this sigma factor in P. aeruginosa. A. vinelandii and P. aeruginosa are phylogenetically related and their AlgU binding motifs, so far reported, are very well conserved [6,19,76].

A Position-Specific Scoring Matrix (PSSM), was developed as described in Materials and methods, using as a reference experimentally determined P. aeruginosa AlgU promoters previously reported (i.e., by mapping the 5’ end of the corresponding mRNA) (S2 Table) [39–47]. This PSSM was used to search the genome of P. aeruginosa PAO1. A total of 134 AlgU promoters with a score ≥ 0.9 were predicted (motifs showing invariant or well-conserved consensus sequences), and 486 with a score of 0.81 (motifs showing conserved consensus sequences) (S9 Table).

The P. aeruginosa AlgU PSSM was subsequently used for searching AlgU binding sites in the genome of A. vinelandii. Predicted AlgU biding sites with a score ≥ 0.9 were found in the regulatory region of 117 genes, a number like that found for P. aeruginosa. 420 genes showed a predicted AlgU-dependent promoter with a score of 0.81 (S10 Table).

For comparative purposes, the prediction of potential AlgU promoters was extended for the P. fluorescens SBW25 and A. chroococcum Ac-8003 genomes (S11 and S12 Tables). The size of the predicted regulons was conserved, with about 100 motifs with a score ≥ 0.9. Interestingly, the genes encoding an OsmC-like protein, a peptidyl-prolyl cis-trans isomerase, a transaldolase TalB, the RpoH sigma factor, or genes required for trehalose synthesis or alginate production, were among the many genes conserved in the four bacteria (S13 Table). Of note, some of them were previously shown to be part of the primary AlgU regulon of P. aeruginosa PA14 [77]. A sequence logo for the AlgU promoters identified in each bacterium was generated and as expected, it reflected the original promoter sequences used as input to develop the PSSM (Figs 4 and S4).

Fig 4. Sequence logo for AlgU DNA binding motifs.

A PSSM was used to identify AlgU binding motifs in the genome of P. aeruginosa (A) or A. vinelandii (B) (see Materials and Methods section for details). The predicted AlgU binding motifs are shown with 16 (upper panels) or 17 (lower panels) bp spacers, between the -10 and -35 boxes. The relative sizes of the letters represent their frequency in the sequences.

The primary AlgU regulon of A. vinelandii

A comparative analysis between the predicted AlgU targets in the genome of A. vinelandii (S10 Table) and the gene products identified by our proteomic approach, under the positive control of this sigma factor (S4 and S7 Tables), allowed us to define the primary AlgU regulon of A. vinelandii, i.e., genes directly regulated by AlgU under our tested conditions (vegetative growing conditions or during encystment). A total of 47 genes showed AlgU binding motifs in their regulatory region, while 53 genes were present in poly-cistronic operons with predicted AlgU promoters (S14 Table). This primary regulon was compared to that defined for P. aeruginosa PA14 strain upon heat shock exposure (50°C/5 min) [77]. Nine genes were common between these two regulons and comprise osmC; genes for alginate production (alg); fkbA, encoding a peptidyl-prolyl cis-trans isomerase; pckA, encoding a phosphoenolpyruvate carboxykinase; and glgB, encoding a 1,4-alpha-glucan branching enzyme, involved in trehalose or glycogen synthesis. These core genes might contribute to the survival of P. aeruginosa after a heat shock or upon A. vinelandii encystment, which is triggered by adverse conditions. The reduced number of common genes between the primary AlgU regulons in A. vinelandii and P. aeruginosa PA14, might be derived from the different conditions employed to define these regulons, and reveals the existence of specialized regulons under the control of the same sigma factor. In agreement with this assumption, orthologs in PA14 of 46 genes of the primary AlgU regulon in A. vinelandii showed conserved AlgU-binding motifs (S15 Table). However, they were not identified as part of the primary AlgU regulon in P. aeruginosa upon heat shock [77], implying that the description of the AlgU regulon is far from being complete. On the other hand, 22 genes of the primary AlgU regulon of A. vinelandii were exclusive of this organism; as anticipated, some of them encoding functions only found in the Azotobacter but not in the Pseudomonas genus. They include genes for ARs production (arsB and arsC) and nitrogen fixation (i.e., nifD, nifH), for the extracellular modification of alginate (algE6) but also include genes encoding a CRISPR-associated protein (Avin_17200) or proteins of unknown function.

Concluding remarks

In summary, we have reported the AlgU regulon during both, vegetative and encysting conditions in A. vinelandii. Although we detected molecular targets that explained processes previously documented under the control of AlgU (such as flagella biogenesis, alginate production or oxidative stress resistance), this work further expands our knowledge about the function of this sigma factor in A. vinelandii. AlgU is required for losing the flagella during the early steps of differentiation and agrees with the increase in the activity of this sigma factor (Fig 1). However, our data indicate that AlgU is also needed for the metabolic switch that takes place upon encystment. The algU mutant could not produce ARs (Fig 3), implying that the replacement of the phospholipids of the cell membrane does not occur. A total of 337 proteins were found under the positive control of AlgU (S4 and S7 Table). Among these, the corresponding genes of 100 proteins showed predicted AlgU promoters in their regulatory region. The presence of AlgU binding motifs for some orthologous of P. aeruginosa or P. fluorescens was conserved but others were exclusive of A. chroococcum and A. vinelandii suggesting that the AlgU regulon is flexible and optimized for each bacterium and in response to the environmental or culture condition.

Supporting information

S1 Fig. Visualization of protein-protein interaction network by String of AlgU controled proteins during vegetative conditions.

126 proteins with altered expression in the absence of the sigma factor AlgU, were analized. Interaction nodes such as those constituted by proteins involved in lipids metabolism (green circle), central metabolism (blue circle), flagella biogenesis and motility (cyan circle), trehalose synthesis (black circle) and enzymes for alginate production (pink circle) are indicated. Disconnected nodes are hided; the network was generated using an interaction score of 0.7.

(PDF)

S2 Fig. Visualization of protein-protein interaction network generated by String of proteins positively controled by AlgU during encystment.

305 down-represented proteins in the absence of the sigma factor AlgU, during encysting conditions were analized. Interaction nodes such as those constituted by proteins involved in ribosome assembly (red circle), nitrogen fixation (blue circle), amino acid metabolism (cyan circle), respiration (black circle), central metabolism (green circle) and enzymes for alginate (pink circle) or trehalose (yellow circle) production are indicated. Disconnected nodes are hided; the network was generated using an interaction score of 0.7.

(PDF)

S3 Fig. Visualization of protein-protein interaction network by String of proteins negatively controled by AlgU.

184 up-regulated proteins in the absence of the sigma factor AlgU, during encysting conditions were analyzed. Interaction nodes such as those constituted by proteins involved in amino acids (green circle), aminoacyl-tRNA (cyan circle), or fatty acid (black circle) biosynthesis are indicated. Disconnected nodes are hided; the network was generated using an interaction score of 0.7.

(PDF)

S4 Fig. AlgU sigma factor binding motifs.

A Position-Specific Scoring Matrix was used to identify AlgU binding motifs in the genome of P. fluorescens (A) or A. chroococcum (B) (see Materials and Methods section for details). The predicted AlgU binding motifs with are shown with 16 (upper panels) or 17 (lower panels) bp spacers between the -10 and -35 boxes.

(PDF)

S1 Table. DNA sequences of the primer pairs used in the present work.

(XLSX)

S2 Table. List of experimentally determined AlgU promoters of P. aeruginosa that served as a reference for generating the Position-Specific Scoring Matrix.

(XLSX)

S3 Table. Total proteins identified in the algU mutant under vegetative growth as compared to wild-type strain.

(XLSX)

S4 Table. List of proteins downregulated in the absence of AlgU as compared to the wild-type strain, under vegetative growing conditions.

(XLSX)

S5 Table. List of proteins upregulated in the absence of AlgU as compared to the wild-type strain, under vegetative growing conditions.

(XLSX)

S6 Table. Total proteins identified in the algU mutant under encystment-induced conditions as compared to the wild-type strain.

(XLSX)

S7 Table. List of proteins downregulated in the absence of AlgU as compared to the wild-type strain, under encystment-induced conditions.

(XLSX)

S8 Table. List of proteins upregulated in the absence of AlgU as compared to the wild-type strain, under encystment-induced conditions.

(XLSX)

S9 Table. Predicted AlgU-dependent promoters in the P. aeruginosa PAO1 genome.

(XLSX)

S10 Table. Predicted AlgU-dependent promoters in the A. vinelandii DJ genome.

(XLSX)

S11 Table. Predicted AlgU-dependent promoters in the P. fluorescens SBW25 genome.

(XLSX)

S12 Table. Predicted AlgU-dependent promoters in the A. chroococcum Ac-8003 genome.

(XLSX)

S13 Table. Comparative analysis of the predicted AlgU-dependent promoters in A. vinelandii with their orthologous in A. chroococcum, P. aeruginosa and P. fluorescens.

(XLSX)

S14 Table. The primary AlgU regulon of A. vinelandii and a comparative analysis with the primary regulon reported for P. aeruginosa PA14.

(XLSX)

S15 Table. Predicted AlgU dependent promoters in the genome of P. aeruginosa PA14 and PAO1.

(XLSX)

Data Availability

All relevant data are within the manuscript and its Supporting Information files.

Funding Statement

This work was supported by a grant from Programa de Apoyo a Proyectos de Investigación e Innovación Tecnológica, UNAM (PAPIIT) IN209521 to CN. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.