Identification of genes affecting alginate biosynthesis in Pseudomonas fluorescens by screening a transposon insertion library

Abstract

Background

Polysaccharides often are necessary components of bacterial biofilms and capsules. Production of these biopolymers constitutes a drain on key components in the central carbon metabolism, but so far little is known concerning if and how the cells divide their resources between cell growth and production of exopolysaccharides. Alginate is an industrially important linear polysaccharide synthesized from fructose 6-phosphate by several bacterial species. The aim of this study was to identify genes that are necessary for obtaining a normal level of alginate production in alginate-producing Pseudomonas fluorescens.

Results

Polysaccharide biosynthesis is costly, since it utilizes nucleotide sugars and sequesters carbon. Consequently, transcription of the genes necessary for polysaccharide biosynthesis is usually tightly regulated. In this study we used an engineered P. fluorescens SBW25 derivative where all genes encoding the proteins needed for biosynthesis of alginate from fructose 6-phosphate and export of the polymer are expressed from inducible Pm promoters. In this way we would avoid identification of genes merely involved in regulating the expression of the alginate biosynthetic genes. The engineered strain was subjected to random transposon mutagenesis and a library of about 11500 mutants was screened for strains with altered alginate production. Identified inactivated genes were mainly found to encode proteins involved in metabolic pathways related to uptake and utilization of carbon, nitrogen and phosphor sources, biosynthesis of purine and tryptophan and peptidoglycan recycling.

Conclusions

The majority of the identified mutants resulted in diminished alginate biosynthesis while cell yield in most cases were less affected. In some cases, however, a higher final cell yield were measured. The data indicate that when the supplies of fructose 6-phosphate or GTP are diminished, less alginate is produced. This should be taken into account when bacterial strains are designed for industrial polysaccharide production.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-016-3467-7) contains supplementary material, which is available to authorized users.

Background

Linear polysaccharides composed of mannuronic and guluronic acid residues that may be O-acetylated, are denoted alginate. These polymers are synthesized by brown and some red algae and by bacterial species belonging to the genera Azotobacter and Pseudomonas. Alginates manufactured from brown algae are currently used in diverse industrial and pharmaceutical applications. However, alginates produced by bacteria can more easily be tailored to obtain the compositions desired for the more high-value end of the alginate market [1], and this has motivated our studies on alginate-producing bacteria.

Production of a secreted polysaccharide imposes a drain on the cell’s carbon and energy sources, and thus the biosynthesis is usually tightly regulated under natural conditions. In batch cultures, alginate-producing P. fluorescens mutants display a reduced cell yield compared to the corresponding non-alginate producing strains [2]. Bacterial alginate production is controlled by the alternative sigma factor AlgU and is usually turned off in Pseudomonas spp. Induction of alginate biosynthesis results in a proteolytic cascade that finally cleaves the AlgU anti-sigma factor MucA, leading to transcription of the genes in the alg operon [3].

In the first steps of bacterial alginate biosynthesis fructose 6-phosphate (Fru6P) is converted to GDP-mannuronic acid by the concerted action of AlgA, AlgC and AlgD. GDP-mannuronic acid is then polymerized to polymannuronic acid by Alg8 and the copolymerase Alg44. Together with AlgG, AlgX, AlgK and AlgE these form a protein complex that transports the alginate out of the cell as depicted in Fig. 1a [4]. AlgG also epimerizes some M-residues to G, while AlgI, AlgJ, AlgF and AlgX are needed to O-acetylate some of the M-residues. The alginate lyase AlgL removes alginate molecules that have been released to the periplasm [5]. Twelve of the thirteen genes directly involved in alginate biosynthesis are found in the alg operon, while the last, algC, is found elsewhere on the chromosome. This gene organization is found in all characterized alginate-producing bacteria. In addition to Fru6P and GTP, dimeric cyclic di-GMP (c-di-GMP) is needed for bacterial alginate biosynthesis [6, 7].

Fig. 1.

The relationship between alginate biosynthesis and the cellular metabolism in P. fluorescens. a The proteins and metabolites needed for alginate biosynthesis. b A simplified model of the cell’s metabolism highlighting the processes identified in the present study as being important for full alginate biosynthesis levels. The genes discussed in the paper are highlighted in yellow. The Entner-Doudoroff pathway and the oxidative part of the pentose phosphate pathway are indicated by red arrows, and the non-oxidative part of the pentose phosphate pathway with purple arrows. Green arrows indicate other pathways competing with accumulation of the three metabolites Fru6P, GTP and c-di-GMP, while blue arrows indicate pathways that would increase the synthesis of one of these three metabolites. Each arrow may represent several enzymatic steps. Abbreviations: OM: Outer membrane, IM: Inner membrane, M: mannuronic acid residue, G: guluronic acid residue, Ac: Acetyl, TCA: Tricarboxylic acid cycle, PP: the non-oxidative part of the pentose phosphate pathway, GN6P: Glucosamine 6-phosphate, PG: Peptidoglycan, G6P: Glucose 6-phosphate, 6PG: 6-phosphogluconate, Pyr: Pyruvate, ILV: Isoleucine Leucine Valine, B5: Pantothenate, Trp: Tryptophan, PRPP: Phosphoribosyl pyrophosphate, R5P: Ribose 5-phosphate, E4P: Erythrose 4-phosphate

Recently we showed that the alginate synthesis rate is not proportional to the number of alginate biosynthetic complexes, indicating that there must be some kind of metabolic control as well [4]. In a recent transposon screen, some genes affecting AlgU-regulation were identified in P. aeruginosa [8]. However, the aim of the present study was to identify genes and pathways that influence alginate biosynthesis indirectly by perturbing the cell’s metabolism. An alginate-producing P. fluorescens strain in which the alg operon and algC is under control of the inducible Pm promoter was constructed and subjected to transposon mutagenesis. The Pm promoter and its activator XylS originally controls expression of the genes of the meta-cleavage pathway of aromatic hydrocarbons on the Pseudomonas putida plasmid pWW0 [9]. We have earlier shown that the Pm promoter and the weaker Pm promoter derivative Pm-G5 are useful for obtaining different levels of controlled gene expression in P. fluorescens [5]. About 11500 insertion mutants were screened with respect to growth and alginate biosynthesis, and the inactivated genes in mutants displaying altered alginate yields were identified. The results supported our hypothesis that further levels of post-translational regulation exist, allowing the cell to prioritize basic cellular metabolism over alginate biosynthesis.

Results and discussion

Construction of a P. fluorescens strain in which the alginate biosynthesis genes are controlled by the inducible Pm promoter

In order to avoid re-identification of the genes already known to directly regulate expression of the structural alginate biosynthetic genes, a derivative of P. fluorescens SBW25 designated strain MS1 was constructed (Fig. 2a). In this strain the naturally regulated algD promoter (which controls expression of the alg operon) was substituted with the wild-type Pm promoter. xylS, encoding the activator protein needed for expression from the Pm promoter, was inserted upstream of Pm. Then algC was inactivated by an in-frame deletion followed by a chromosomal insertion of a transposon containing a new algC copy expressed from a mutant version of Pm (PmG5) [5, 10]. This strain, designated MS2, produces only a small amount of alginate in the absence of Pm induction due to the low uninduced activity of PmG5.

Fig. 2.

Genotypes for selected genetic constructs used in this study. a Strain MS1 in which the Pm promoter and the gene encoding XylS is inserted between the promoter and start codon of algD. b Strain MS2 in which a transposon expressing algC from PmG5 is inserted into PFLU2944 in an algC ⁻ derivative of MS1. c Map of the transposon TnMS11 used for mutagenesis in this study. d Strain HE230 in which the gene encoding XylS and the PmG5 promoter is inserted between the promoter and start codon of algC in SBW25mucA. Inactivation of mucA confers a high level expression from wild type PalgD. Relevant promoters, and the two restriction sites used for sequencing are displayed above each map-line. The alg-genes are coloured to match Fig. 1, other P. fluorescens genes flanking the genes of interest are coloured blue, and heterologous genes and elements are coloured green. I and O denote the minitransposon ends

Alginate production has been reported to affect cell yield in P. fluorescens [2], and it was also possible that m-toluic acid would have an effect on growth. This was tested by cultivating the non-alginate producing wild type strain SBW25 and strain MS2 in Biolector® for three days in 0.5 x PIA supplemented with glycerol as carbon source. Growth rate and cell yield was significantly lower for the induced strain MS2 relative to the non-alginate producing strain, while no effect was seen by cultivating SBW25 in the presence or absence of 0.5 mM m-toluic acid (Additional file 1: Figure S1).

The transposon carrying algC was found to disrupt PFLU2944, which is the last gene in an operon encoding a putative ABC transporter (Fig. 2b). In the presence of the Pm/PmG5 inducer (m-toluate), the alginate production of strain MS2 was similar to that of strain MS1 (results not shown).

Construction of a transposon insertion library and screening with respect to alginate synthesis

The transposon-containing suicide vector pMS11 (Fig. 2c) was used for mutagenesis of strain MS2. Nearly 11500 insertion mutants were picked robotically from the original agar medium plates and cultivated in 96-deep-well microtiter plates containing 0.5x liquid PIA with glycerol and m-toluate. After three days, cell densities and alginate production were measured. The initial screen was followed by a rescreen of primary candidates and 184 mutants were found to produce less than 50% (163 mutants) or more than 110% alginate (21 mutants) when compared to the parent strain. The transposon insertion sites in all these mutants were determined by DNA sequencing, leading to identification of 134 different genes belonging to most of the main cellular functions (results not shown). Of these genes only ten were known alginate biosynthesis structural genes, while one was xylS, the positive regulator of Pm expression. These results show that about 92% of the identified genes are not directly associated with alginate synthesis. The screen did not cover all relevant genes in the genome, since insertions in algG, algF and algI (members of the alg operon) were not found.

Evaluation of the mutants to select candidates for further studies

Sequenced mutants with altered alginate phenotypes were cultivated in triplicates in 96-deep-well microtiter plates in three different media; 0.5xPIA with glycerol and 0.5xDEF4 with fructose or glycerol as carbon sources (7 g/L), and 0.5 mM m-toluate to induce alginate production. In the DEF4 media ammonium is the only nitrogen source, while PIA contains peptone that may be used as both nitrogen and carbon source. Furthermore, DEF4 contains more phosphate than PIA. The alginate yield from the control strain (MS2) was significantly higher in the DEF4 media, about 3 g/L compared to about 1 g/L in PIA, which resulted in better accuracy of the data in DEF4 for low alginate producers.

Results for mutants displaying significantly altered alginate production levels in at least one of the three media, are shown in Table 1. Significant changes were defined as less than 50% or more than 110% of the alginate production of the parent strain, and 36% of the retested gene-inactivation mutants did not meet this criterion. No mutant produced more alginate than the control strain in all three media. Mutants with insertions in alginate biosynthetic genes and xylS did, as expected, not produce alginate and are not included in Table 1. When several mutants had the same gene inactivated and displayed similar phenotypes, results from only one of them are shown in Table 1. For mutants where genes involved in glycerol utilization, amino acid biosynthesis or phosphate uptake had been inactivated, one would expect that the observed effects on biomass and alginate yield should be media dependent. As shown in Table 1 this was the case for most genes belonging to these categories.

Table 1.

Identified mutants and their growth yield and alginate production in the three media^a

Number of independent transposon mutants	Gene ID	Gene	Function	Growth (G) and Alginate production (A) in different mediaa
				PIA Gly				0.5xDEF4 Gly				0.5xDEF4 Fru
				G		A		G		A		G		A
				%	SD	%	SD	%	SD	%	SD	%	SD	%	SD
	SBW25 WT			132	6	0	25	86	0	0	0	350	6	1	6
	Control			100	6	100	13	100	2	100	13	100	5	100	3
1	PFLU0460	aceE1	Energy production and conversion	66	2	63	2	57	4	19	4	293	20	0	4
1	PFLU3193	aceE2		108	4	48	2	98	6	89	9	130	3	106	19
1	PFLU5345	cioB		151	18	69	3	102	1	65	5	129	7	119	13
2	PFLU3801	ftsK	Cell cycle control, cell division, chromosome partitioning	38	0	17	5	72	14	4	3	425	42	3	3
	PFLU3801	ftsK		53	7	39	8	102	2	86	8	16	1	7	4
1	PFLU5304			107	6	47	7	100	6	96	8	130	18	101	12
1	PFLU1384		Amino acid transport and metabolism	108	3	18	3	71	1	81	33	148	45	95	6
1	PFLU2019			126	27	0	3	116	14	0	1	338	14	0	0
1	PFLU2124			127	4	40	7	102	2	88	7	294	30	55	27
1	PFLU3475			106	2	45	11	98	7	88	9	132	5	99	32
1	PFLU3887			104	4	12	11	93	6	72	16	132	28	53	5
1	PFLU5797	ilvD		69	22	64	2	60	10	17	1	338	31	2	3
1	PFLU4188	trpF		100	2	34	3	105	5	0	0	343	3	0	5
1	PFLU5559	trpD		62	5	58	11	106	9	9	12	344	23	0	1
2	PFLU5561	trpE		102	2	55	4	140	1	0	3	4	14	1	2
1	PFLU0612	purH	Nucleotide transport and metabolism	44	5	4	2	81	3	0	0	317	13	0	0
1	PFLU4183	purF		55	2	48	21	81	2	9	6	344	11	4	2
4	PFLU5034	purL		50	1	43	10	70	1	1	3	348	7	0	2
2	PFLU6054	purK		84	4	2	5	113	8	0	4	343	3	1	4
1	PFLU6055	purE		88	7	45	25	87	5	37	5	368	9	0	2
2	PFLU5396	amn		79	2	29	17	106	2	66	9	135	16	62	18
2	PFLU1142	glpK	Carbohydrate transport and metabolism	52	2	66	39	19	7	2	4	214	9	52	40
3	PFLU1143	glpF		59	2	4	2	18	3	0	2	128	10	80	4
1	PFLU3030	paaF		109	5	33	7	94	4	84	18	122	7	118	16
2	PFLU3365	treZ		107	2	62	12	97	1	98	3	132	17	132	4
3	PFLU4630	acnA		144	8	73	14	92	7	68	23	125	3	113	20
1	PFLU4949	pykA		107	4	41	14	51	3	132	5	160	3	85	12
1	PFLU0416	hemE	Coenzyme transport and metabolism	147	5	68	8	98	2	66	14	117	11	136	12
1	PFLU5820	nudH	Translation, ribosomal structure and biogenesis	77	3	23	12	78	0	127	5	65	23	31	3
1		23S rRNA		152	10	52	16	98	1	94	14	130	15	142	24
1	PFLU3173		Transcription	92	8	10	3	93	2	89	4	276	0	67	7
1	PFLU3307			165	18	74	11	103	2	88	28	137	10	110	8
1	PFLU4259			125	2	62	1	75	5	84	11	153	20	111	5
1	PFLU4774			102	10	7	3	91	12	75	7	110	3	79	26
1	PFLU5984	dut	Replication, recombination and repair	114	6	45	6	47	3	62	7	149	5	86	1
1	PFLU0013	htrB	Cell wall/ membrane/ envelope biogenesis	117	7	40	5	78	5	62	3	123	24	110	18
1	PFLU1562	nagZ		67	1	5	2	89	2	60	20	186	9	80	7
1	PFLU4993	ampG		58	2	4	2	87	1	53	10	198	6	83	17
1	PFLU5439	mpl		129	6	36	5	103	1	80	8	106	4	46	6
2	PFLU5545			62	3	8	2	99	9	90	11	204	1	65	6
2	PFLU5546	anmK		54	4	3	2	86	4	68	11	192	1	59	26
3	PFLU5573	amgK		95	30	45	10	23	2	15	4	192	18	46	38
1	PFLU5461	ispA		88	5	2	6	85	4	46	8	282	31	15	7
1	PFLU4418	fleN	Cell motility	88	1	43	4	80	4	104	29	80	1	40	12
1	PFLU4439	fliF		140	8	12	2	97	3	0	0	342	3	0	0
1	PFLU4448	fliC		111	8	38	2	87	5	102	7	171	3	132	6
1	PFLU0870	tldD	Posttransla-tional modification, protein turnover, chaperones	111	3	26	5	103	1	88	6	137	6	111	16
4	PFLU2032	prc		100	2	37	6	81	3	59	6	124	7	26	6
2	PFLU2614	sohB		98	3	11	3	96	1	73	8	108	7	109	12
1	PFLU3805	clpA		147	7	60	7	104	1	127	20	256	8	124	21
3	PFLU4383			118	6	21	8	87	22	51	5	298	8	38	13
1	PFLU5007	dsbC		102	7	13	6	97	2	79	5	120	10	84	2
2	PFLU5911	ppx	Inorganic ion transport and metabolism	50	2	82	38	73	3	67	28	90	13	76	2
	PFLU5911	ppx		54	1	0	0	73	4	101	7	107	5	98	15
1	PFLU0511	rsgA	General function prediction only	108	5	86	11	96	4	101	32	104	12	124	26
1	PFLU2104			149	9	66	18	94	1	65	13	117	7	136	15
1	PFLU2996			134	10	66	7	104	3	74	12	122	9	132	5
1	PFLU3202			108	3	76	10	81	3	47	6	403	7	22	3
1	PFLU3391			51	25	58	8	68	1	95	7	118	0	129	13
1	PFLU3411			105	6	50	9	101	2	70	29	135	2	137	19
1	PFLU3456			104	4	47	5	103	0	107	2	151	13	117	6
1	PFLU1883		Function unknown	69	0	0	0	74	14	66	14	712	268	0	0
1	PFLU1995			149	7	85	9	104	1	97	7	158	18	113	17
1	PFLU4517			129	3	82	6	96	3	28	5	377	18	2	6
1	PFLU5579	apaG		39	2	47	21	104	5	4	3	175	32	18	7
1	PFLU2489			93	10	120	51	69	3	75	8	324	40	23	9
1	PFLU5377			107	8	8	3	97	4	76	5	145	1	92	20
1	Upstream PFLU2629			107	6	9	3	106	4	84	13	300	16	106	7
1	Upstream PFLU3162			122	2	77	20	76	1	88	12	280	36	96	15
1	Upstream PFLU3931			97	2	35	3	75	2	3	3	327	11	12	5
1	PFLU2519		Pseudogene	114	4	148	26	102	3	98	6	337	9	89	13
1	PFLU0259	ompR	Signal transduction mechanisms	89	5	7	1	101	4	70	12	127	17	101	12
2	PFLU0461	glnE		85	14	10	4	120	4	104	13	317	11	67	1
1	PFLU4125A			119	4	164	14	112	0	86	4	158	5	72	2
4	PFLU5236	cbrA		140	1	34	7	105	4	55	7	319	5	66	21
2	PFLU5237	cbrB		118	1	15	3	109	3	23	3	483	24	31	2
2	PFLU5819	ptsP		80	12	8	3	90	6	79	15	49	1	5	5
1	PFLU6039	phoB		78	1	44	20	102	2	90	16	110	7	84	21
1	PFLU6040	phoR		25	0	0	0	97	12	87	0	99	22	100	23
1	PFLU2808			100	2	11	3	110	4	75	5	255	52	19	11
1	PFLU3002		Intracellular trafficking, secretion, and vesicular transport	153	6	48	15	93	5	83	6	141	9	94	5
1	PFLU3951			106	6	8	1	105	2	83	1	110	5	101	20
1	PFLU5567			98	9	61	5	0	0	0	0	59	1	2	4

a: The strains were cultivated in microtiter plates for three days before cell and alginate yield were measured. The mutants shown are those that displayed significantly different alginate production levels in at least one of the three tested media. Data are not shown for strains with transposon insertions in the genes encoded by the alginate operon or in algC. The Table shows how many independent transposon insertions mutants that were identified for each gene, the gene identifier, the gene name, and which functional group the corresponding protein is assigned to. Growth above 125% and alginate production above 110% are marked using bold types, growth and alginate production between 10 and 50% are marked using italics, and growth and alginate production below 10% are written in bold italics. Three biological replicates were cultivated for each strain, and the results are given as percent (%) of the values obtained from the control strain MS2. Standard deviations for the three replicates are shown in the columns to the right (SD)

It is probable that in many cases the phenotype observed in a transposon insertion mutant is caused directly by inactivation of the identified gene. However, polar effects (particularly in operons) and unrelated, spontaneous mutations can certainly not be excluded. For those genes where several independent transposon insertion mutants were identified, it is more likely that the observed phenotype is caused by the observed transposon insertion. The same argument may be used when several genes encoding proteins in the same metabolic pathway have been identified. In addition, 18 of the identified genes were chosen to be complemented either by expressing the wild type gene on a transposon or by adding the lacking metabolite. The transposons were constructed and transferred to the mutant strains, and both the mutant strains and the complemented strains were cultivated in two new growth experiments (Tables 2 and 3). Two of the 18 mutants could not be complemented and are not discussed further. These results show that the phenotypes of 16 out of 18 (89%) tested mutants can be explained by the transposon insertions only.

Table 2.

Growth and alginate production of mutants using medium supplements or complementing transposons^a

Inactivated gene	Supplement/comple-menting gene (s)b	PIA		0.5xDEF4 Glycerol		0.5xDEF4 Fructose
		Growthc	Alginate	Growthc	Alginate	Growth	Alginate
wt		100	100	100	100	100	100
trpF		65	0	145	0	291	38
	tryptophane	70	139	78	60	130	45
	trpF	88	121	88	94	153	106
trpD		56	0	1	0	3	0
	tryptophane	68	261	68	80	71	75
	trpD	85	142	114	23	242	38
	trpDC	89	85	81	91	152	120
purH		29	0	1	0	21	0
	Adenine, thiamine	51	188	14	33	39	21
	purH	90	100	91	88	163	97
purE		33	0	0	0	1	0
	Adenine, thiamine	56	91	17	50	43	29
	purE	57	0	1	0	1	0
purL		19	0	0	0	1	0
	Adenine, thiamine	52	124	12	48	38	30
ilvD		77	18	2	2	4	0
	ilvD	91	82	91	109	222	97
aceEI		33	58	6	20	13	7
	aceEI	96	42	85	96	244	123
PFLU3030		88	0	80	89	104	103
	PFLU3030	97	142	89	123	101	111
dsbC		108	52	111	89	149	115
	dsbC	105	127	105	117	94	92
sohB		100	9	100	50	479	−2
	sohB	91	109	104	106	109	100
nagZ		53	0	139	93	115	75
	nagZ	87	145	84	120	289	79
anmK		32	0	124	98	197	99
	anmK	84	118	86	135	111	128
ispA		82	0	100	97	121	115
	ispA	92	212	88	117	405	77
cbrB		83	0	110	63	374	46
	cbrB	91	118	91	95	213	107
PFLU3887		91	67	98	103	105	110
	PFLU3887	90	33	109	103	112	103
PFLU5567		87	64	2	6	141	30
	PFLU5567	93	103	1	16	6	10

a: The strains were grown in deep-well plates containing the indicated media for four days before cell and alginate yield were measured. b: empty field denotes no supplement or complementing vector. c: Values are given as percentage of the control strain (SBW25 MS1 ΔalgC:: TnKB61). Actual values for the control strain were (growth [OD₆₆₀]/alginate [g/L]): PIA: 0.492/0.33, DEF4 glycerol: 0.850/1.72, DEF4 fructose: 0.308/3.08. Growth above 125% and alginate production above 110% are marked using bold types, growth and alginate production between 10 and 50% are marked using italics, and growth and alginate production below 10% are written in bold italics

Table 3.

Effect of PhoBR disruptions on P. fluorescens growth and alginate biosynthesis

Strain	Growth (OD600)	Alginate (g/l)
SBW25mucAHE230	2.5+/−0.24	4.3+/−0.89
SBW25mucAHE230 ΔphoR	2.7+/−0.33	3.6+/−0.36
SBW25mucAHE230 ΔphoB	1.4+/−0.12	0.0+/−0.0
SBW25mucAHE230 ΔphoR:: TnTK5	2.4+/−0.09	4.0+/−0.24
SBW25mucAHE230 ΔphoR:: TnTK7	2.1+/−0.10	4.3+/−0.66
SBW25mucAHE230 ΔphoB:: TnTK6	1.5+/−0.20	1.2+/−0.56
SBW25mucAHE230 ΔphoB:: TnTK7	2.0+/−0.13	5.5+/−0.12

a: The cells were grown for 72 h in shaking flasks using DEF3 medium with 20 g/l glycerol, 1 μM phosphate and 0.5 mM m-toluate. Average values from three independent experiments are shown

Alginate biosynthesis requires a functional biosynthetic complex, Fru6P and a dimeric form of c-di-GMP (Fig. 1a). Interestingly, the majority of those mutants that reproducibly produced less alginate were assigned to the groups involved in uptake and metabolism of carbohydrates, amino acids and nucleotides (Table 1). In addition four genes encoding proteins involved in protein modification were identified. Fig. 1b summarizes how the pathways identified in the current study might influence alginate yield, and these genes and pathways are discussed in more detail below.

Alginate production is influenced by signal transduction systems involved in carbon, nitrogen and phosphor metabolism

Four different signal transduction systems, CbrAB, NtrBC, PTS^Ntr, and PhoBR, were identified in the screen by using the criteria of either complementation or identification of several independent mutants in specific genes or pathways. The CbrAB two-component system has been described in several species of Pseudomonas as sensors and regulators of genes involved in utilization of different carbon and nitrogen sources, and has been proposed as sensors for the C/N balance in the cell [11, 12]. It has been shown that CbrB activates the expression of non-coding RNAs that relieve the catabolite repression otherwise exerted by Crc [13]. In P. putida, inactivation of cbrB also affected stress responses and biofilm development [14]. Our results show that the identified cbrB mutant produces less alginate (0-63%) than the otherwise isogenic control strain in all three media (Table 2). The mutant could be complemented by introducing a transposon-encoded copy of cbrB (Table 2). The effect of inactivating cbrA was, however, less pronounced, and might be caused by a polar effect on cbrB (Table 1). In P. putida, a cbrB mutant was shown to be unable to use some amino acids as carbon source, and to have an increased expression level of some of the genes encoding proteins involved in the Entner-Doudoroff pathway [14]. If the consequences of inactivating cbrB is similar in P. fluorescens, these two effects alone might explain the observed growth and alginate yields for the cbrB mutants, by reducing the net flow to Fru6P (Fig. 1b). However, given the known pleiotropic nature of a cbrB mutation, this probably is not the full explanation.

NtrBC is known to be an important response regulator system for bacterial nitrogen sensing, and has been found to interact with the CbrAB system [14]. GlnE is needed for the posttranscriptional activation of glutamine synthase, which is a part of the NtrC regulatory cascade [15]. It has been shown that inactivation of this gene lowered the pool of Fru6P in Corynebacterium glutamicum [16]. Consistent with this the alginate yield was significantly lower in PIA and in DEF4 with fructose for both glnE mutants (Table 1).

Glutamine and α-ketoglutarate are used by the NtrC-cascade to sense the carbon and nitrogen status of the cell, and these metabolites were recently found to affect the phosphorylation rate of the nitrogen-related phosphoenolpyruvate phosphotransferase system (PTS^Ntr) in E. coli [17]. PTS^Ntr is also known to form a link between carbon and nitrogen metabolism in pseudomonads [18]. While fructose is probably imported and phosphorylated by a PTS in P. fluorescens, glycerol is taken up through a transport and kinase system and is fed into the central metabolism as triose phosphates [19]. PtsP (EI^Ntr) is the first protein in the nitrogen-related phosphate relay, and the two ptsP mutants identified in the current study produced low amounts of alginate both in PIA (24 and 8%) and in DEF4 with fructose (14 and 5%). An earlier study has shown that a ptsP mutant of P. putida produces less polyhydroxyalkanoate than the wild type, and it was suggested that such a mutant would behave as if there was a carbon limitation [20]. A similar argument could be used to explain the lower yield of alginate in our ptsP mutant. Recently it was also shown that inactivation of ptsP in P. aeruginosa decreases the level of c-di-GMP [21].

The response regulator PhoB and the histidine kinase PhoR control the Pho-regulon, which covers a major pathway for bacterial adaptation to phosphate starvation. PhoB may also be activated by other kinases [22]. Since phoB and phoR form an operon, new in-frame deletion mutants for each of these genes were constructed in the alginate-producing strain SBW25mucAHE230 (Fig. 2d). This strain was chosen because our standard gene recombination vector could not be used in the tetracycline-resistant strain MS2. The wild-type genes were cloned both individually and as an operon on transposons, and these transposons were used to complement the deletion mutants. The new phoR mutant behaved similarly to the wild type strain, while the phoB deletion resulted in lower cell yield and no alginate production when cultivated in DEF3 with reduced phosphate concentration (1 μM) (Table 3). Both traits were restored by chromosomal insertion of a transposon encoding both phoB and phoR, while chromosomal insertion of a transposon encoding phoB only partially regained alginate production and normal growth. Lack of PhoB will lead to decreased phosphate uptake under phosphate-limiting conditions, and this may result in less trinucleotides [23]. Furthermore, in Pseudomonas aeruginosa the AlgQ (AlgR2), has been shown to regulate the production of GTP through its positive regulatory effect on transcription of ndk, and Ndk is required for alginate production [24]. AlgQ is an anti-sigma-70 factor and has been shown to positively regulate alginate production [25], possibly by increasing the amount of RNAP available for the alternative sigma-factor AlgU. Transcription of algQ is positively regulated by PhoB [24]. In our strain, transcription of the alginate biosynthetic genes depends on the Pm promoter, which in turn depends on the sigma factors RpoH and RpoS for transcription [26]. Thus, it is possible that AlgQ may have a positive effect on expression from Pm. If that is the case, this might also explain the lack of alginate production in the phoB mutant when grown in a low phosphate medium.

Inactivation of certain genes involved in cell wall metabolism and vitamin biosynthesis leads to decreased alginate yield

In the present screen, insertions in five of the nine genes known to be involved in peptidoglycan recycling in Pseudomonas [27] were identified as having a negative impact on alginate biosynthesis (mpl, ampG, anmK, amgK and nagZ). The absence of Mpl, which is involved in recycling of the peptide part of peptidoglycan, only slightly decreased the alginate production. However, absence of any of the other four identified enzymes, AmpG, AnmK, AmgK or NagZ, resulted in very low alginate production in the PIA medium and reduced alginate yield in the DEF4 media (Table 1). The sugar phosphates used for peptidoglycan synthesis either originates from peptidoglycan recycling or is synthesized from Fru6P (Fig. 1b). Since Fru6P is also a precursor for alginate, depletion of this phosphorylated sugar would be expected to cause decreased alginate yield [2]. The nagZ and anmK genes were cloned on transposons, and shown to complement the deficiency in alginate production in the corresponding insertion mutants (Table 2).

Three of the identified genes, aceE1, ilvD and ispA were linked to pyruvate metabolism (Fig. 1b). aceE1 encodes a component of pyruvate dehydrogenase, which is an essential part of the central carbon metabolism. The viability of this mutant might be explained by the presence of other genes encoding AceE-like proteins in P. fluorescens. However, the aceE1 mutant grew more slowly than strain MS2, and hardly produced any alginate. ilvD encodes a dihydroxy-acid dehydratase that participates in the biosynthesis of branched amino acids and in the biosynthesis of pantothenate (vitamin B5) and coenzyme A. The ilvD mutant displayed a similar phenotype as the aceE1 strain in all three media (Table 2). The ispA mutant would be expected to have defects in the biosynthesis of isoprenoids, which would affect the biosynthesis of ubiquinone and the cell membrane. This mutant produced very low amounts of alginate when grown in PIA, while the phenotypes in the DEF4 media were more similar to the control strain (Table 2). All three mutants were complemented when the corresponding wild type genes were expressed from transposons (Table 2). Disruption of a pathway may often result in an increased flow to the immediate precursor for the missing enzyme, since the cell will perceive a lack of the end product. In the ispA and ilvD mutants this would lead to consumption of pyruvate, which then would have to be replenished by increasing the flow through the Entner-Doudoroff pathway (Fig. 1b). Pantothenate (needed for CoA) and ubiquinone are necessary for the anabolism and energy production of the cell, and the medium-dependent defects in growth and alginate yield displayed by the mutants might be caused by a lower content of these vitamins in peptone (PIA) compared to yeast extract (DEF4).

Deficiencies in purine or tryptophan biosynthesis reduce alginate yield

Eleven of the mutants identified in the screen turned out to have insertions in genes needed for purine biosynthesis (purHFLKE and amn). GTP is required for alginate biosynthesis as a precursor for both GDP-mannuronic acid and the signal molecule c-di-GMP (Fig. 1b). Three of the identified purine biosynthesis mutants (purE, purH and purL) were retested in deep-well plate cultivations and grew poorly in all media (Table 2). The purH strain was complemented when wild-type purH was expressed from a transposon, while the purE mutant was not complemented by expressing purE. This might, however, result from a polar effect on the downstream purK gene. Addition of adenine and thiamine to the media increased both growth and alginate yield for all three mutants (Table 2), strongly suggesting that the observed phenotypes were caused by deficiencies in the purine synthesis pathway.

In eight of the sequenced mutants, the transposon had disrupted genes putatively involved in amino acid biosynthesis (Table 1). Three of these, trpDEF, were genes involved in tryptophan synthesis. The mutants with insertions in trpD and trpF were investigated further and both could be complemented by inserting an intact corresponding gene on a transposon (Table 2). Furthermore, addition of tryptophan to the growth medium restored normal growth and alginate yield in both mutants (Table 2).

Both tryptophan and purine synthesis are linked to Fru6P through the pentose phosphate pathway (Fig. 1b). Defects in these biosynthetic pathways might affect alginate synthesis negatively by increasing the need for phosphoribosyl pyrophosphate (PRPP), and thus increase the flow from Fru6P to this intermediate. Since GTP is necessary for alginate biosynthesis, the observed phenotypes might also be caused by an insufficient supply of purines. Our results are corroborated by other studies demonstrating that de novo synthesis of purine is necessary for biofilm formation in P. fluorescens [28], and that tryptophan is important for biofilm formation in Salmonella enterica [29].

Disruption of several genes encoding proteins involved in protein folding and modification result in reduced alginate yield

Prc is a protease known to affect alginate biosynthesis in some mucA mutants of P. aeruginosa, and has been proposed to indirectly participate in alginate biosynthetic gene activation through MucA cleavage induced by cell wall stress [30, 31]. However, in our strain both algC and the alg operon are controlled by the Pm promoter, not by the endogenous AlgU-MucA-regulated promoters. Still, four independent prc mutants were identified as displaying a reduced alginate yield (Table 1). Our results therefore show that in P. fluorescens a prc mutation negatively affects alginate biosynthesis even in a mucA ⁺ strain. In addition the screen identified another peptidase belonging to the same family, SohB, which also negatively affected alginate yield when inactivated. This phenotype was complemented by a transposon expressing sohB (Table 2). It is unknown which proteins, apart from MucA, is the target of these two proteases in P. fluorescens.

Two genes encoding proteins involved in protein folding were identified in the screen as producing less alginate than the control (Table 1). PFLU4383 encodes a putative thiol:disulfide interchange protein and is located upstream of and partly overlapping dsbG, encoding another disulfide isomerase. Three independent inactivations of PFLU4383 were identified. PFLU5007 encodes the disulfide isomerase DsbC and its phenotype was complemented by a transposon-encoded copy of the gene (Table 2). A mutant of P. aeruginosa with transposon-inactivated dsbC was recently found to display a non-mucoid phenotype [32], indicating that DsbC is needed for normal levels of alginate production in both species. The results suggest that full alginate production in these media depend on correct folding of some proteins. It remains unknown which proteins need these isomerases for correct folding.

Conclusion

In an earlier study, it was shown that inactivation of glucose-6-phosphate dehydrogenase increased alginate yield when glycerol was used as carbon source, and this indicated that the availability of Fru6P may be one limiting factor to sustain high level alginate production [2]. Furthermore, it has been shown that the number of alginate biosynthetic complexes are not influenced by the absence of precursors for alginate synthesis [4], indicating that these complexes are not destabilized in the absence of polymer synthesis. The aim of screening a transposon insertion library, was to discover genes and metabolic pathways that indirectly influence alginate production in P. fluorescens. The main conclusion of our data is that alginate biosynthesis depends on sufficient levels of Fru6P, GTP and c-di-GMP (Fig. 1b). Inactivation of genes in several systems sensing the carbon/nitrogen ratio resulted in mutants that produce less alginate than the parent strain, and this further indicates that alginate production might be down-regulated as a response to a perceived carbon limitation. A majority of the analysed mutants displayed a significantly decreased alginate yield, while the cell yield was less affected, and in some cases even increased. This suggests that when P. fluorescens is facing certain nutrient limitations, less alginate is produced.

Methods

Growth of bacteria

E. coli and P. fluorescens (Table 4) were routinely cultivated in L broth or on L agar at 37 °C or 30 °C, respectively [33]. P. fluorescens was also grown in PIA medium [33], DEF4 medium [34] and DEF3 medium with low phosphate: KH₂PO₄ 0,14 mg/L, KCl 0.36 g/L, NH₄Cl 2.21 g/L, citric acid · H₂O 0.9 g/L, ferric citrate 0.02 g/L, H₃BO₃ 0.001 g/L, MnCl₂ · 4H₂O 0.005 g/L, EDTA · 2H₂O 0.0039 g/L, CuCl₂ · 2H₂O 0.0005 g/L, Na₂Mo₄O₄ · 2H₂O 0.0008 g/L, CoCl₂ · 6H₂O 0.0008 g/L, Zn (CH₃COO)₂ · 2H₂O 0.0027 g/L, NaCl 1.56 g/L, MgSO₄ · 7H₂O 0.57 g/L, MOPS 10 g/L. For precultures, 0.39 g/L yeast extract was added to the DEF4 medium. The pH of DEF3 and DEF4 was adjusted to 7.0. Carbon sources – fructose or glycerol – were added to 20 g/L. Antibiotics used: ampicillin (Ap, 200 mg/L), tetracycline (Tc, 15 mg/L), apramycin (Am, 25 mg/L), kanamycin (Km, 50 mg/L). For growth in microtiter plates and micro bioreactors (BioLector®), half the concentrations of the media containing 7 g/L carbon source was used, and the cultures were incubated at 25 °C as detailed previously [34]. For some experiments adenine (0.8 mM), thiamine (0.05 mM), or tryptophan (2.5 mM) were added as medium supplements. For growth studies in Biolector® microreactors the cultivations were performed in M2P-labs FlowerPlate® BOH with 1 ml medium per reactor. The cultivations were started (3 vol-% inoculum) from L broth precultures cultivated at 30 °C for 18 h. The BOH plates were incubated at 25 °C, 1300 rpm with 3 mm orbital movement at 80% humidity. pH, dissolved oxygen and biomass were measured automatically every hour by the Biolector system. The biomass measured by the Biolectors Photomultiplier was calibrated by offline optical density measurements using a standard spectrophotometer.

Table 4.

Bacterial strains^a and plasmids used in this study

Strains	Description	Reference
E. coli S17-1 (λpir)	λpir (for replication of oriR6K-plasmids) recA, thi pro hsdR-M + RP4 2-Tc::Mu-Km::Tn7TpRSMR	[36]
P. fluorescens SBW25	Non-mucoid P. fluorescens wild type	[37]
SBW25MS1	Derivative of SBW25 where the Pm promoter is inserted directly upstream of algD using pMS9.	This study
SBW25MS1 ΔalgC:: TnKB61	Derivative of SBW25 MS1 where algC has been deleted utilizing pKB22, and a copy of algC controlled by the PmG5 promoter has been introduced on a transposon inserted into gene PFLU2944.	This study
SBW25mucAHE230	Alginate-producing derivative of SBW25 encoding a defect MucA and where the expression of algC is controlled by the PmG5 promoter	H. Ertesvåg, unpublished
SBW25mucAHE230ΔphoB	Derivative of SBW25mucAHE230 where an in-frame deletion in phoB was introduced utilizing pTK10.	This study
SBW25mucAHE230ΔphoR	Derivative of SBW25mucAHE230 where an in-frame deletion in phoR was introduced utilizing pTK9.	This study
Plasmids
pKD20	pUT based transposon vector containing PmG5. Apr, Kmr.	[5]
pLitmus28Tc	High copy number cloning vector. Tcr, Apr	[5]
pMG48	RK2-based gene replacement vector. lacZ + , Tcr, Apr	[33]
pMC1	RK2-based gene replacement vector for replacing the DNA sequence upstream of algD with the Pm-promoter. lacZ + , Tcr, Apr	[10]
pKB22	Gene replacement vector for creating an algC-deletion. lacZ + , Tcr, Apr	[5]
pKB60	Transposon vector. Contains the transposon TnKB60 with algC under the control of PmG5. Apr, Kmr	[5]
pYQ1	pUT based transposon vector containing PmG5. Amr, Kmr.	[2]
pEM1	Derivative of pLitmus28Tc containing part of the transposon from pKD20. Apr	[2]
pKB61	Derivative of pKB60 where a 1.7 kb AvrII-NcoI DNA fragment encoding Kmr and most of XylS was exchanged with a 2.5 kb AvrII-NotI DNA fragment containing tetAR. Tcr, Apr	This study
pMS9	Derivative of pMC1 where a 0.7 kb SbfI-NotI DNA fragment containing a gene upstream of algD was exchanged with a PCR product containing the 0.8 kb sequence directly upstream of algD. lacZ + , Tcr, Apr	This study
pMS2	Derivative of pLitmus28Tc where the tetAR genes were exchanged with a 3.4 kb BamHI-fragment from pKD20 containing the minitransposon and oriR6K. Kmr, Apr	This study
pMS10	Derivative of pMS2 where a 0.5 kb BsiWI-EcoRI-fragment containing oriR6K was deleted and the 1.9 kb NotI-PstI fragment encoding XylS was exchanged with a 0.4 kb PCR product encoding oriR6K. Kmr, Apr	This study
pMS11	Derivative of pKD20 where a 3.7 kb BssHII-SfiI-fragment was exchanged with a 1.5 kb BssHII-SfiI-fragment containing oriR6K from pMS10. Kmr, Apr	This study
pTK1	Derivative of pEM1 in which a 2.2 kb PCR-amplified NdeI-NotI DNA fragment encoding phoBR from P. fluorescens was inserted. Apr.	This study
pTK3	Derivative of pTK1 in which an inserted 2.2 kb PCR-amplified NcoI-PspOMI DNA fragment from P. fluorescens including the first 46 nt of phoB replaced most of the phoB gene. Apr.	This study
pTK4	Derivative of pTK1 from which a 0.9 kb BstEII-BsaBI DNA fragment encoding most of phoR was deleted. Apr.	This study
pTK5	Derivative of pKD20 in which a 1.5 kb NdeI-NotI PCR fragment from P. fluorescens containing phoR was inserted. Kmr.	This study
pTK6	Derivative of pKD20 in which a 1.1 kb NcoI-NotI DNA fragment from pTK4 containing phoB was inserted. Kmr.	This study
pTK7	Derivative of pKD20 in which a 2.2 kb NdeI-NotI PCR fragment from P. fluorescens containing phoBR was inserted. Kmr.	This study
pTK8	Derivative of pTK1 from which a 0.3 kb BstEII DNA fragment was deleted, creating an in-frame deletion in phoR. Apr.	This study
pTK9	Derivative of pMG48 in which a 3.0 kb NcoI-NotI DNA fragment from pTK8 was inserted, containing a deletion in phoR. Apr, Tcr.	This study
pTK10	Derivative of pMG48 in which a 3.1 kb NcoI-NotI DNA fragment from pTK3 was inserted, containing a deletion in phoB. Apr, Tcr.	This study
pYQ1 trpF	Derivative of pYQ1 in which a 0.7 kb NdeI-NotI PCR fragment encoding TrpF was inserted. Amr.	This study
pYQ1 trpD	Derivative of pYQ1 in which a 1.1 kb NdeI-NotI PCR fragment encoding TrpD was inserted. Amr.	This study
pYQ1 trpDC	Derivative of pYQ1 in which a 1.9 kb NdeI-NotI PCR fragment encoding TrpDC was inserted. Amr.	This study
pYQ1 purH	Derivative of pYQ1 in which a 1.6 kb NdeI-NotI PCR fragment encoding PurH was inserted. Amr.	This study
pYQ1 purE	Derivative of pYQ1 in which a 0.5 kb NdeI-NotI PCR fragment encoding PurE was inserted. Amr.	This study
pYQ1 ilvD	Derivative of pYQ1 in which a 2.1 kb NdeI-NotI PCR fragment encoding IlvD was inserted. Amr.	This study
pYQ1 aceEI	Derivative of pYQ1 in which a 2.7 kb NdeI-NotI PCR fragment encoding AceE1 was inserted. Amr.	This study
pYQ1 PFLU3030	Derivative of pYQ1 in which a 1.0 kb NdeI-NotI PCR fragment encoding PFLU3030 was inserted. Amr.	This study
pYQ1 PFLU3887	Derivative of pYQ1 in which a 1.0 kb NdeI-NotI PCR fragment encoding PFLU3887 was inserted. Amr.	This study
pYQ1 PFLU5567	Derivative of pYQ1 in which a 1.2 kb NdeI-NotI PCR fragment encoding PFLU5567 was inserted. Amr.	This study
pYQ1 dsbC	Derivative of pYQ1 in which a 0.9 kb NdeI-NotI PCR fragment encoding DsbC was inserted. Amr.	This study
pYQ1 sohB	Derivative of pYQ1 in which a 1.1 kb NdeI-NotI PCR fragment encoding SohB was inserted. Amr.	This study
pYQ1 nagZ	Derivative of pYQ1 in which a 1.1 kb NdeI-NotI PCR fragment encoding NagZ was inserted. Amr.	This study
pYQ1 anmK	Derivative of pYQ1 in which a 1.5 kb NdeI-NotI PCR fragment encoding AnmK was inserted. Amr.	This study
pYQ1 ispA	Derivative of pYQ1 in which a 0.9 kb NdeI-NotI PCR fragment encoding IspA was inserted. Amr.	This study
pYQ1 cbrB	Derivative of pYQ1 in which a 1.4 kb NdeI-NotI PCR fragment encoding CbrB was inserted. Amr.	This study

a: Mutant strains complemented with transposons are not included in the Table

Analyses of alginate and growth

The cultures were incubated for three to four days before the cell density and alginate yield were assayed. Enzymatic measurements of alginate production were performed as described earlier [2, 35]. Briefly, the cell free medium were treated with a mixture of an M-specific and a G-specific alginate lyase, and OD₂₃₀ before and after the reaction were measured using a Beckman Coulter robotic liquid handling work station with a Paradigm microplate reader.

Construction of the transposon vector and the transposon insertion library

Cloning, transformation, conjugation and gene deletions were performed as described earlier [33]. The plasmids and transposons used and constructed in this study are described in Table 4, while the primer sequences are found in Additional file 2: Table S1. PCR was performed using the Expand High Fidelity kit (Roche). PCR-amplified genes were confirmed by sequencing. Transposon insertions were to be identified by sequencing, so a transposon vector that would allow easy cloning of the transposon insertion site in E. coli was constructed and designated pMS11 (Table 4, Fig. 2c). The vector contains a derivative of the Tn5 minitransposon that comprises oriR6K and a gene encoding kanamycin resistance within the transposon boundaries. The transposon contains single sites for the restriction enzymes SacI and EcoRI close to the ends of the transposon. pMS11 was propagated in E. coli S17-1 λpir that encodes the Pir protein necessary for R6K-replication. pMS11 was transferred to P. fluorescens by conjugation, and conjugants were selected on PIA containing kanamycin. Colonies were picked using a Genetix QPixII colony picking robot and transferred to 384 well plates with 0.5 x PIA and Km, and incubated at 25 °C overnight before glycerol was added to 15% v/v and the plates were stored at −80 °C.

Identification of transposon insertion sites

Genomic DNA was isolated from mutants of interest. For some mutants the transposon insertion site was identified by direct sequencing using this DNA as the template and the primer MS11 Ori (Additional Table S1). For sequencing on genomic DNA, 5 μg DNA, 50 pmol sequencing primer, 8 μl 2.5x BigDye Terminator Ready Reaction Mix v1.1 (Applied Biosystems) and water to 20 μl was mixed. The reaction was subjected to sixty cycles of 30 s denaturation at 95 °C, 30 s annealing at 52 °C, and four minutes elongation at 60 °C. Alternatively, the DNA flanking the transposon insertion site was cloned by restricting genomic DNA isolated from a transposon mutant with SacI or EcoRI. The fragments were circularized by ligation, and the ligation mixture was transformed into E. coli S17-1 λpir and selected for resistance to kanamycin. Sequencing the resulting plasmids provided better quality sequences than by sequencing directly on genomic DNA. The transposon insertion points were identified by comparing the obtained sequences to the genome sequence (GenBank Accession number AM181176).

Additional file

Additional file 1: Figure S1. ^{(75.5KB, pptx)}

Growth profiles of Pseudomonas fluorescence SBW25 and MS2 cultivated in 0.5 x PIA. (PPTX 75 kb)

Additional file 2: Table S1. ^{(32.6KB, xls)}

Primers used in the study. (XLS 30 kb)