Functional Overlap of hetP and hetZ in Regulation of Heterocyst Differentiation in Anabaena sp. Strain PCC 7120

ABSTRACT

HetR plays a key role in regulation of heterocyst differentiation and patterning in Anabaena. It directly regulates genes involved in heterocyst differentiation (such as hetP and hetZ), genes involved in pattern formation (patA), and many others. In this study, we investigated the functional relationship of hetP and hetZ and their role in HetR-dependent cell differentiation. Coexpression of hetP and hetZ from the promoter of ntcA, which encodes the global nitrogen regulator, enabled a hetR mutant to form heterocysts with low aerobic nitrogenase activity. Overexpression of hetZ restored heterocyst differentiation in a hetP mutant and vice versa. Overexpression of hetR restored heterocyst formation in either a hetP or a hetZ mutant but not in a hetZ hetP double mutant. The functional overlap of hetP and hetZ was further confirmed by reverse transcription-quantitative PCR (RT-qPCR) and transcriptomic analyses of their effects on gene expression. In addition, yeast two-hybrid and pulldown assays showed the interaction of HetZ with HetR. HetP and HetZ are proposed as the two major factors that control heterocyst formation in response to upregulation of hetR.

IMPORTANCE Heterocyst-forming cyanobacteria contribute significantly to N₂ fixation in marine, freshwater, and terrestrial ecosystems. Formation of heterocysts enables this group of cyanobacteria to fix N₂ efficiently under aerobic conditions. HetR, HetP, and HetZ are among the most important factors involved in heterocyst differentiation. We present evidence for the functional overlap of hetP and hetZ and for the key role of the HetR-HetP/HetZ circuit in regulation of heterocyst differentiation. The regulatory mechanism based on HetR, HetP, and HetZ is probably conserved in all heterocyst-forming cyanobacteria.

INTRODUCTION

In the history of Earth, cyanobacteria were the first group of oxygenic photosynthetic microorganisms (1, 2). Due to limited nitrogen sources in the environment, nif genes for N₂ fixation arose in cyanobacteria (3). As cyanobacteria flourished in the ancient water bodies, the atmospheric oxygen accumulated and rapidly reached a level that compromised nitrogenase activities in unicellular and filamentous N₂-fixing species. Under such a selection pressure, specialized N₂-fixing cells called heterocysts occurred in certain filamentous species (4). Such cells were differentiated from a small percentage of vegetative cells in defined patterns, creating micro-oxic conditions for nitrogenase. Now, heterocyst-forming cyanobacteria are important contributors to N₂ fixation in marine, freshwater, and terrestrial ecological systems (5–7).

Anabaena/Nostoc sp. strain PCC 7120 (here Anabaena 7120) responds to nitrogen deficiency by producing single heterocysts at semiregular intervals along filaments. It is the strain used most often in studies of cell differentiation and patterning in cyanobacteria. Based on the studies in Anabaena 7120, NtcA and HetR have been shown to be the two key regulators in control of heterocyst differentiation (8, 9). Compared to NtcA, the global nitrogen regulator, HetR is more specifically required for heterocyst differentiation, and overexpression of hetR leads to production of multiple contiguous heterocysts. Therefore, HetR is often referred to as the master regulator of heterocyst differentiation. A comparison of the genome-wide transcription dynamics in the wild-type (WT) strain of Anabaena 7120 and a hetR mutant led to identification of at least 209 transcriptional start sites that are directly or indirectly controlled by HetR (10). The HetR homodimer (11, 12) or homotetramer (13) binds to consensus recognition sites upstream of hetP (12, 14), hetZ (15), hepA (16), patA (15, 17, 18), and other genes (16–18), exhibiting 4 modes of gene regulation (18).

Of the genes regulated by HetR, hetP and hetZ were first identified by transposon mutagenesis in Anabaena 7120, and the initial transposon insertion mutants showed no heterocyst differentiation (19, 20). Later, mutants generated by insertion or replacement with an antibiotic resistance cassette showed greatly retarded heterocyst differentiation (14, 20). After nitrogen step-down, hetZ starts to be upregulated at about 3 h and is expressed mainly in differentiating cells (20). Similarly, hetP is also activated in patterned cells at the early stage of heterocyst differentiation (14, 19). Overexpression of hetP in Anabaena 7120 led to the multiple-contiguous-heterocyst phenotype (19), but overexpression of hetZ did not (20). Ectopic expression of hetP enabled a hetR deletion mutant to form heterocyst-like cells with anoxic nitrogen-fixing activity after 50 h of nitrogen step-down (14). In addition, HetP also promotes commitment after induction of heterocyst differentiation and may interact with its homologs (Asl1930, Alr2902, and Alr3234) that delay the commitment or inhibit the development (21).

The current information indicates that HetR controls heterocyst differentiation and patterning by regulating hetP, hetZ, patA, and other genes. Of the HetR-dependent genes, hetP and hetZ appear to be upregulated simultaneously and are both required for initiation of heterocyst differentiation; however, their relationship and their respective roles in cell differentiation have not been clarified. In this study, we found the evidence for the functional overlap of hetP and hetZ and the interaction of HetZ with HetR. Taking the results together, we propose the HetR-HetP/HetZ regulatory circuit as a part of the core machinery for control of heterocyst differentiation.

RESULTS

Formation of heterocysts with aerobic nitrogenase activity in hetR mutants.

hetZ and hetP are two genes that are directly regulated by HetR and required for heterocyst differentiation. After nitrogen step-down, they are upregulated immediately after hetR is turned on (see Fig. S1 in the supplemental material). It has been reported that expression of hetP from P_petE, a copper-regulated promoter, enabled a hetR deletion mutant to form heterocyst-like cells and that the anaerobic N₂-fixing activity was detectable after 50 h of nitrogen step-down (14). Instead of P_petE, we employed the promoter of ntcA, the global nitrogen regulator gene, to express hetP and hetZ in a hetR::C.CE2 mutant (15). The hetR mutant produced no heterocysts after nitrogen step-down. Our constructs, thereby expressing P_ntcA-hetZ or P_ntcA-hetP, did not enable the hetR mutant to produce heterocysts or proheterocysts after nitrogen step-down. However, coexpression of hetZ and hetP led to heterocyst differentiation at a frequency of 1.0% ± 0.3% and aerobic nitrogenase activity (1.06 ± 0.08 μmol mg chlorophyll a [Chla]⁻¹ h⁻¹ [average over 6 h]) at 24 h (Fig. 1; Table 1). Of 150 heterocysts that we observed, all were located at terminal positions. In comparison to the hetR mutant, the hetZ- and hetP-coexpressing strain showed very slow but evident growth in BG11₀ (without nitrate) under aerobic conditions. Prolonged nitrogen step-down (up to 96 h) did not substantially change the heterocyst frequency.

FIG 1.

Heterocyst formation in hetR mutants of Anabaena 7120 expressing hetZ, hetP, or both from P_ntcA. Filaments were stained with alcian blue at 24 h or 96 h after nitrogen step-down. Heterocysts and proheterocysts are indicated by arrows.

TABLE 1.

Nitrogenase activities and heterocyst frequencies of Anabaena strains

Relevant genotype	Time (h)	Nitrogenase activity (μmol C2H4 · mg Chla−1 · h−1)a		Heterocyst frequency (%)b	Diazotrophic growthc
		Anoxic	Aerobic
WT	24	15.37 ± 3.80	9.06 ± 2.03	10.8 ± 0.7	+++++
	48	17.87 ± 1.51	8.68 ± 2.57	12.9 ± 0.5
WT + PntcA-hetZ-hetP	24	ND	ND	13.0 ± 0.5	++++
	48	ND	ND	23.5 ± 1.6
hetR::C.CE2	24	0	0	0	−
	48	0	0	0
hetR::C.CE2 + PntcA-hetZ	24	0	0	0	−
	48	0	0	0
hetR::C.CE2 + PntcA-hetP	24	0	0	0	−
	48	0	0	0
hetR::C.CE2 + PntcA-hetZ-hetP	24	1.53 ± 0.60	1.06 ± 0.08	1.0 ± 0.3	+
	48	0.92 ± 0.17	1.03 ± 0.27	1.5 ± 0.2
ΔhetP	24	0	0	0	−
	48	3.30 ± 1.92	1.55 ± 0.49	1.9 ± 0.3
ΔhetP + PhetZ-hetZ	24	1.47 ± 0.65	0.52 ± 0.42	7.0 ± 0.4	++
	48	14.95 ± 2.01	26.04 ± 7.37	10.9 ± 0.6
ΔhetP + PhetR-hetR	24	9.55 ± 1.64	8.07 ± 2.17	10.3 ± 1.3	++
	48	35.46 ± 6.90	24.84 ± 4.94	17.0 ± 4.4
hetZ::Tn5-1087b	24	0	0	0	−
	48	0	0	0
hetZ::Tn5-1087b + PhetP-hetP	24	3.54 ± 0.46	0.32 ± 0.05	6.7 ± 0.5	+++
	48	18.97 ± 1.91	4.15 ± 0.22	18.1 ± 1.7
hetZ::Tn5-1087b + PhetR-hetR	24	4.97 ± 0.11	2.43 ± 0.56	6.8 ± 0.3	++
	48	21.36 ± 5.32	30.52 ± 6.38	15.2 ± 4.2
hetZ::C.K2 ΔhetP	24	0	0	0	−
	48	0	0	0
hetZ::C.K2 ΔhetP + PhetR-hetR	24	0	0	0	−
	48	0	0	0

^aAverage ± standard deviation over 6 h. ND, not determined.

^bAverage ± standard deviation. Both heterocysts and proheterocysts were included.

^c+, growth; −, no growth.

To confirm the partial bypass of hetR function by hetZ and hetP, we also constructed a hetR deletion (ΔhetR) mutant and coexpressed hetZ and hetP in the mutant. At 24 h after nitrogen step-down, coexpression of hetP and hetZ from P_ntcA did not restore heterocyst formation in the ΔhetR mutant; at 96 h, heterocysts (1.1% ± 0.1%) were formed at terminal positions (Fig. 1) with low aerobic nitrogenase activity (1.67 ± 0.03 μmol mg Chla⁻¹ h⁻¹ [average over 6 h]).

Functional overlap of hetZ and hetP.

Because ectopic expression of hetZ and hetP could partially bypass the hetR mutation, we wondered whether overexpression of hetR could bypass a hetZ, hetP, or hetZ hetP (double) mutation. Inactivation of hetZ (20) or deletion of hetP (14) abolished or greatly impeded heterocyst formation in Anabaena 7120. At 48 h after nitrogen step-down, the hetZ::Tn5-1087b (here referred to as hetZ::Tn5) mutant showed no heterocyst differentiation, while the ΔhetP mutant produced 1.9% heterocysts (Fig. 2; Table 1). Unlike the hetZ::Tn5 mutant, the hetZ::C.K2 mutant showed significantly delayed heterocyst differentiation (20). Both the hetZ::C.K2 and hetZ::Tn5 mutations could be complemented by hetZ (15). We also inactivated hetZ with C.K2 in the ΔhetP mutant. The resulting hetZ::C.K2 ΔhetP double mutant showed no heterocyst formation (Fig. 2; Table 1). hetZ::C.K2 in combination with the hetP mutation was used to make the double mutation discernible in phenotype from the two original single mutations (hetZ::C.K2 and ΔhetP). hetR with its native promoter cloned in a pDU1-based plasmid (denoted P_hetR-hetR) was introduced into the Anabaena hetZ::Tn5, ΔhetP, and hetZ::C.K2 ΔhetP mutants. Heterocyst formation was restored in the hetZ::Tn5 and ΔhetP mutants but not in the hetZ::C.K2 ΔhetP double mutant (Fig. 2; Table 1). In the hetZ mutant overexpressing hetR, multiple contiguous heterocysts (Mch) were formed (occasionally, we found cell division within heterocysts). In BG11₀, the hetZ::Tn5 and ΔhetP mutants with P_hetR-hetR showed very slow growth, and the hetZ::C.K2 ΔhetP double mutant with P_hetR-hetR showed no growth.

FIG 2.

Bypass of hetP mutation by hetR or hetZ, and of hetZ mutation by hetR or hetP, on a pDU1-based plasmid. Filaments were stained with alcian blue at 24 h or 48 h after nitrogen step-down. Heterocysts and proheterocysts are indicated by arrows. Tn5-1087b is indicated as Tn5.

One explanation for the bypass of the hetP or hetZ mutation by overexpression of hetR is that hetP and hetZ functionally overlap each other. To test this hypothesis, we introduced hetZ cloned in a pDU1-based plasmid into ΔhetP or hetP in the plasmid into the hetZ::Tn5 mutant. Heterocyst formation was restored either way, but in comparison to the case for the wild type, the diazotrophic growth was significantly slowed and the occurrence of nitrogenase activity was delayed (Fig. 2; Table 1).

Using reverse transcription-quantitative PCR (RT-qPCR), we confirmed the overexpression of hetR, hetP, and hetZ in the strains used as described above (Fig. 3; see Fig. S2 in the supplemental material). Evidently, overexpression of hetR caused overexpression of hetP in the hetZ::Tn5 mutant and overexpression of hetZ in the ΔhetP mutant. Formation of Mch in the hetZ mutant overexpressing hetR is consistent with the phenotype caused by overexpression of hetP (19); in contrast, overexpression of hetZ would not cause such a phenotype (20).

FIG 3.

RT-qPCR analyses of relative mRNA levels of hetP, hetZ, and hetR in Anabaena 7120 strains. +N, in BG11; −N, after nitrogen step-down. P_ntcA-hetP was carried on pHB4908, P_ntcA-hetZ was carried on pHB4909, P_ntcA-hetZ-hetP was carried on pHB4910, P_hetP-hetP was carried on pHB4550, P_hetZ-hetZ was carried on pHB1462, and P_hetR-hetR was carried on pHB4023. Strains and plasmids are listed in Table S1 in the supplemental material.

Effects of hetZ and hetP on gene expression.

To compare the roles of hetP and hetZ in heterocyst differentiation, we examined the expression of hepA (22), hepB (23), hglD (24), devB (25), and hetN (26, 27) in Anabaena 7120 and the derivative hetZ::Tn5-1087b, ΔhetP, and hetZ::C.K2 ΔhetP strains at 6 h after nitrogen step-down (Fig. 4). hepA and hepB are required for formation of the envelope polysaccharide layer, devB and hglD are required for the glycolipid layer, and hetN is required for maintenance of the heterocyst pattern. Of these genes, hepA, hglD, and hetN were significantly downregulated by the hetP and hetZ mutations, hepB was slightly downregulated by the hetZ mutation, and devB was downregulated by the hetP hetZ double mutation but not by the hetP or hetZ single mutation. Apparently, hetP and hetZ are both required for the expression of some of these genes but not for that of the others.

FIG 4.

RT-qPCR analyses of gene expression in Anabaena 7120 strains at 6 h after nitrogen step-down.

We further employed transcriptome sequencing (RNA-seq) to compare gene expression in the hetP and hetZ mutants to that in the wild type (Table 2). Seventy-four genes showed more-than-2-fold changes (>2 or <0.5) in mRNA level in the hetP mutant, and 71 genes did so in the hetZ mutant. Fifty-four genes showed similar changes in expression in the two mutants. These 54 genes are involved in formation of the envelope polysaccharide layer (glycosyl transferase and transporter), formation of the glycolipid layer (glycolipid synthetase and transporter), heterocyst respiration (cytochrome c oxidase), nitrogen fixation (nitrogenase/nitrogen fixation proteins and Hes proteins), hydrogen uptake (hydrogenase), electron transfer (flavoprotein, ferredoxin, and rubrerythrin), gene regulation (PatB), pattern formation (HetN), and unknown functions (hypothetical or unknown proteins). They are directly or indirectly regulated by HetP and HetZ. For example, the expression of nitrogenase and hydrogenase genes is probably dependent on the formation of mature heterocysts rather than the functions of hetP and hetZ. There are also some genes that showed significant changes in expression only in the hetP or hetZ mutant relative to the wild type levels, indicating the differences between hetP and hetZ in function.

TABLE 2.

Differential gene expression in Anabaena 7120 WT and the hetZ and hetP mutants at 24 h after nitrogen step-down

Gene	Description of product	hetZa mutant		hetPb mutant
		hetZ/WT expression ratio	P value	hetP/WT expression ratio	P value
all0178	Flavoprotein	0.1636	0.0004	0.2491	0.0053
all0664	WD-40 repeat protein	—c	—	0.1210	1.56E−5
all0665	Unknown protein	—	—	0.1655	0.0002
all0666	Unknown protein	—	—	0.0061	9.88E−17
all0667	Unknown protein	—	—	0.0074	7.78E−18
all0687	Uptake hydrogenase large subunit; HupL	0.0610	0.0009	0.0581	0.0009
all0688	Uptake hydrogenase small subunit; HupS	0.0940	0.0011	0.0409	8.82E−5
all1424	Unknown protein	0.0459	0.0001	0.0208	1.79E−5
all1430	Heterocyst ferredoxin; FdxH	0.0087	1.17E−10	0.0000	2.69E−12
all1431	HesB	0.0000	1.56E−6	0.0225	3.54E−5
all1432	HesA	0.0055	5.36E−8	0.0472	2.51E−5
all1433	Nitrogen fixation protein; NifW	0.0000	0.0002	0.0000	0.0002
all1435	Hypothetical protein	0.0000	9.77E−7	0.0222	2.24E−5
all1436	Nitrogen fixation protein; NifX	0.0000	1.56E−6	0.0225	3.54E−5
all1437	Nitrogenase molybdenum-iron protein; NifN	0.0057	4.81E−9	0.0111	2.17E−8
all1438	Nitrogen Fe/Mo cofactor biosynthesis E	0.0502	1.32E−6	0.0442	8.04E−7
all1439	Hypothetical protein	0.0000	3.93E−5	0.0000	4.00E−5
all1440	Nitrogenase molybdenum-iron protein beta chain	0.0000	4.09E−21	0.0000	1.77E−21
all1454	Nitrogenase molybdenum-iron protein alpha chain; NifD	0.0011	2.00E−14	0.0006	3.85E−15
all1455	Nitrogenase iron protein; NifH	0.0023	1.08E−13	0.0007	8.96E−16
all1456	Nitrogen fixation protein; NifU	0.0000	2.11E−11	0.0122	3.92E−9
all1457	Nitrogenase cofactor synthesis protein; NifS	0.0015	4.54E−16	0.0071	5.40E−14
all1516	Ferredoxin-like protein; FdxN	0.0000	0.0001	0.0000	0.0001
all1517	Nitrogen fixation protein; NifB	0.0020	1.28E−17	0.0031	8.20E−17
all2512	Transcriptional regulator; PatB	0.1155	0.0016	0.0834	0.0004
all2868	Unknown protein	0.1205	0.0004	—	—
all3793	Unknown protein	—	—	5.1142	0.0049
all4460	Unknown protein	0.1839	0.0038	0.0989	0.0001
all4617	Hypothetical protein	—	—	0.0725	0.0006
all5341	Probable glycosyl transferase	0.0322	0.0003	0.0153	9.55E−5
all5342	Unknown protein	—	—	0.0000	5.21E−5
all5347	Heterocyst specific ABC transporter, membrane fusion protein DevB homolog	0.0362	0.0013	0.0427	0.0011
all7607	Unknown protein	0.1467	0.0085	—	—
all8046	Unknown protein	—	—	3.8541	0.0043
alr0267	Unknown protein	0.0223	1.20E−6	0.0464	1.68E−5
alr0611	Nitrate transport ATP-binding protein; NrtD	0.2148	0.0010	—	—
alr0663	Alpha-amylase family protein	—	—	0.0044	9.00E−7
alr0668	Hypothetical protein	—	—	0.0012	1.16E−25
alr0669	Unknown protein	—	—	0.0004	1.25E−29
alr0671	WD-repeat protein	—	—	0.0002	7.07E−22
alr0672	Similar to vanadium chloroperoxidase	—	—	0.0021	1.55E−6
alr0691	Hypothetical protein	0.0000	0.0051	—	—
alr0874	Nitrogenase reductase; NifH2	0.0000	0.0014	—	—
alr1174	Rubrerythrin	0.2882	0.0070	0.2157	0.0013
alr1404	Serine acetyltransferase	0.0274	0.0037	—	—
alr1407	Homocitrate synthase; NifV1	0.1552	0.0020	0.0899	0.0001
alr1555	Unknown protein	—	—	0.2331	0.0091
alr1810	Unknown protein	—	—	0.2422	0.0078
alr2463	Unknown protein	—	—	0.0754	0.0013
alr2464	Unknown protein	0.2562	0.0067	0.1837	0.0010
alr2514	Cytochrome c oxidase subunit II	0.0968	0.0004	0.0634	5.64E−5
alr2515	Cytochrome c oxidase subunit I	0.0446	6.42E−7	0.0450	3.78E−7
alr2516	Cytochrome c oxidase subunit III	0.0000	3.53E−7	0.0098	2.42E−6
alr2517	Hypothetical protein	0.1001	0.0097	0.0512	0.0025
alr2518	Hypothetical protein	0.0219	1.05E−5	0.0000	4.39E−7
alr2520	Hypothetical protein	0.1481	0.0058	0.0985	0.0028
alr2729	Hypothetical protein	—	—	0.0260	0.0038
alr2730	Hypothetical protein	0.0577	3.96E−5	0.0684	0.0001
alr2731	Cytochrome c oxidase subunit II	0.0647	4.40E−6	0.0541	1.70E−6
alr2732	Cytochrome c oxidase subunit I	0.0384	3.06E−8	0.0461	1.95E−7
alr2818	Heterocyst differentiation protein; HetP	—	—	0.0000	0.0069
alr2822	Hypothetical protein	0.1126	0.0054	0.1236	0.0048
alr2823	Hypothetical protein	0.0769	0.0027	—	—
alr2824	Hypothetical protein	0.0150	0.0001	0.1208	0.0100
alr2826	Hypothetical protein	0.1226	0.0040	0.0306	3.06E−5
alr2828	Unknown protein	0.0000	0.0067	0.0000	0.0062
alr2831	Probable NAD(P)-dependent oxidoreductase	0.1411	0.0009	0.2059	0.0043
alr2833	Hypothetical protein	0.0364	0.0002	0.0547	0.0005
alr2835	ABC transporter, ATP-binding protein; HepA	0.0319	0.0003	0.0944	0.0025
alr2836	Glycosyltransferase	0.0555	0.0030	0.0165	0.0006
alr2839	Glycosyltransferase	0.0566	0.0094	—	—
alr2841	Unknown protein	0.1392	0.0018	0.0928	0.0004
alr3059	Similar to polysaccharide export protein	3.4301	0.0032	3.9741	0.0014
alr3676	Unknown protein	0.2635	0.0064	—	—
alr4456	Tropinone reductase homolog	5.2357	0.0015	—	—
alr4706	Unknown protein	—	—	0.2965	0.0098
alr4777	Hypothetical protein	—	—	0.2403	0.0086
alr4938	Hypothetical protein	0.1463	0.0010	—	—
alr4984	Unknown protein	0.0000	0.0006	—	—
alr5294	Probable cation efflux system protein	4.1869	0.0025	—	—
alr5351	Heterocyst glycolipid synthase	0.0239	0.0014	0.0041	7.99E−5
alr5352	Unknown protein	0.0000	0.0077	0.0000	0.0077
alr5353	Hypothetical protein	0.1564	0.0019	0.1376	0.0014
alr5354	Heterocyst glycolipid synthase; HglD	0.0651	0.0013	0.0000	1.55E−5
alr5355	Heterocyst glycolipid synthase; HglC	0.0739	0.0001	0.0104	7.98E−7
alr5356	Hypothetical protein	0.0000	0.0035	—	—
alr5358	Ketoacyl reductase; HetN	0.0709	4.43E−6	0.0664	4.36E−6
asl0026	Unknown protein	0.0953	0.0024	—	—
asl0662	Unknown protein	—	—	0.0007	3.39E−21
asl1434	Hypothetical protein	0.0000	0.0068	0.0000	0.0067
asr0670	Unknown protein	—	—	0.0015	1.03E−15
asr0905	Hypothetical protein	—	—	0.1534	0.0021
asr4612	Unknown protein	—	—	0.0504	0.0044

^ahetZ::Tn5.

^bΔhetP.

^c—, 0.5 < ratio < 2.

Interaction of HetZ with HetR.

HetR appears to control cell differentiation through HetP, HetZ, and other proteins. We wondered if these proteins interact with each other. A yeast two-hybrid system was employed to test the interactions between HetR, HetZ, and HetP. In the presence of the empty prey vector (pGADT7), hetP cloned in the bait vector (pGBKT7) was able to activate the reporter genes in yeast (Fig. 5A). Therefore, we tested possible interactions with HetP by cloning hetP in the prey vector. As seen in Fig. 5A, HetR showed positive interaction with HetZ, but neither HetR nor HetZ showed interaction with HetP. The self-activation of HetP in the yeast system implied that HetP may have DNA-binding activity. In vitro, HetP fused to EF-Ts indeed bound to DNA fragments without apparent sequence specificity, while EF-Ts did not (see Fig. S3 in the supplemental material).

FIG 5.

Yeast two-hybrid tests and pulldown assays showing the interaction of HetR with HetZ but not HetP. (A) Yeast two-hybrid tests showing the interaction of HetR with HetZ. pGBKT7 and pGADT7 are bait and prey vectors, respectively; the combinations pGBKT7-Lam/pGADT7-T and pGBKT7-53/pGADT7-T are negative and positive controls, respectively. Growth on an SD/Ade/His/Leu/Trp plate and hydrolysis of X-α-Gal into a blue product indicate positive protein-protein interaction or autonomous activating activity of a protein. (B) Pulldown assays showing the interaction of HetR with HetZ. Proteins were separated by SDS-PAGE (left) and analyzed by Western blotting detection (right) using anti-MBP monoclonal antibody. Lanes: 1, MBP-HetR; 2, GST-HetZ pls GST·Bind resin plus MBP-HetR; 3, GST-HetZ plus GST·Bind resin plus MBP; 4, GST plus GST·Bind resin plus MBP-HetR; 5, GST plus GST·Bind resin plus MBP; 6, MBP.

The HetR-HetZ interaction was confirmed by a pulldown experiment. A glutathione S-transferase (GST)–HetZ fusion protein bound to resin could pull down a maltose-binding protein (MBP)–HetR fusion, while GST and MBP showed no interaction with each other or the fusion proteins (Fig. 5B).

DISCUSSION

Key regulators of prokaryotic cell differentiation, such as CtrA in Caulobacter crescentus (27), Spo0A in Bacillus subtilis (28), and AdpA in Streptomyces griseus (29), can directly regulate many genes or a few genes and then indirectly regulate many other genes. In Anabaena 7120, HetR directly regulates only a few genes involved in heterocyst differentiation, including hetP and hetZ. It appears that hetP and hetZ mediate the control by HetR over many other genes for heterocyst formation. In this study, we investigated the relationship between hetP and hetZ and the role of HetP/HetZ in control of cell differentiation.

Coexpression of hetZ and hetP from P_ntcA enabled a hetR-null mutant to form heterocysts with low aerobic nitrogenase activity at terminal positions, but expression of either hetZ or hetP from the same promoter did not restore heterocyst formation in the hetR background. This result indicates that hetZ and hetP differ from each other in function or are both essential for certain functions. On the other hand, when overexpressed, they can bypass each other's mutation. This suggests that hetZ and hetP overlap in function. RT-qPCR and RNA-seq analyses of the differential gene expression in their mutants and the wild type supported this conclusion. Consistently, overexpression of hetR bypassed the hetZ mutation by upregulating hetP and bypassed the hetP mutation by upregulating hetZ. Overexpression of hetR, however, was unable to bypass the hetZ hetP double mutation, indicating that the two functionally overlapping genes as a whole are essential for heterocyst differentiation.

Coexpression of hetZ and hetP from P_ntcA partially bypassed the function of HetR but apparently did not circumvent the HetR-dependent expression of patA (18): all heterocysts formed in the hetR mutant coexpressing hetZ and hetP were located at the terminal positions, reminiscent of the phenotype of a patA mutant. Even though hepA, a gene involved in the formation of envelope polysaccharide layer, is also dependent on HetR (16), the terminal heterocysts were stained by alcian blue (Fig. 1), indicating the presence of envelope polysaccharides. As shown in a previous report, lowered expression of hepA may not necessarily abolish the formation of polysaccharide layer (30). In addition to HetR-dependent gene expression, there must be other HetR-dependent processes that could not be bypassed by hetZ and hetP, for example, suppression of certain genes by HetR (16, 18), mutual dependence of hetR and ntcA (31), etc.

HetP and HetZ, proteins of 159 and 401 amino acids (aa), respectively, show no similarity to each other. A HetP-green fluorescent protein (GFP) fusion was localized in the cytoplasm in proheterocysts but focalized near the cell poles in mature heterocysts (32); a HetZ-GFP fusion is also found in the cytoplasm of differentiating cells (see Fig. S4 in the supplemental material). Although no functions can be predicted from the amino acid sequence, HetP showed DNA-binding activity in electrophoretic mobility shift assays (EMSA) (see Fig. S3 in the supplemental material). Because the DNA-binding activity appeared to be nonspecific, we speculate that HetP may be involved in the organization of chromosomal DNA regions, altering the accessibility of DNA to regulatory factors. Also, HetZ was predicted to be a DNA-binding protein (20), and our analyses of HetZ point to its role in control of DNA conformation (S. Wang et al., unpublished data). These observations are consistent with their functional overlap in regulation of cell differentiation. Meanwhile, the functional difference between HetP and HetZ could be due to their ways of interaction with DNA and their interactions with different proteins. It was reported that HetP may interact with its homologs (Asl1930, Alr2902, and Alr3234), modulating heterocyst commitment (21), while HetZ interacts with the master regulator HetR (Fig. 5).

Genes involved in heterocyst differentiation could be directly or indirectly regulated by NtcA, HetR, or both (10, 33). Among those genes directly regulated by HetR, hetP and hetZ appear to be the two major mediators in transmitting the signal of HetR accumulation to many other genes involved in heterocyst development. Thus, HetR-HetP/HetZ constitutes a key part in the gene regulation cascade and plays a central role in control of heterocyst differentiation. When overproduced, HetP and HetZ bypass each other's functions in cell differentiation; however, at the wild-type level, they differ significantly in effects on the regulation of many genes, and therefore they are both required for heterocyst differentiation.

MATERIALS AND METHODS

General.

Anabaena 7120 and derivatives, listed in Table S1 in the supplemental material, were cultured in BG11 medium (34) in the light (30 μE m⁻² s⁻¹) on a rotary shaker. Erythromycin (5 μg ml⁻¹), neomycin (20 μg ml⁻¹), or spectinomycin (10 μg ml⁻¹) was added to the medium as appropriate. For induction of heterocysts or growth in nitrogen-free medium, Anabaena strains grown to mid-logarithmic phase were washed 3 times with BG11₀ (without nitrate) and resuspended in the same medium. Growth curves were made by measuring the turbidity (optical density at 730 nm [OD₇₃₀]) of cultures (starting from an OD₇₃₀ of ≈0.05) every day for 8 days. Heterocyst frequencies (>300 cells counted per sample) were analyzed using 3 parallel cultures. Data are means ± standard deviations. Microscopy and alcian blue staining of the heterocyst polysaccharide layer were performed as previously described (35, 36). Cells for extraction of RNA were cultured with air bubbling before and after nitrogen step-down.

Plasmid construction.

The plasmid construction processes are described briefly below and in detail in Table S1 in the supplemental material. DNA fragments cloned by PCR were confirmed by sequencing.

(i) Plasmids for expressing hetZ and hetP from P_ntcA in hetR mutants.

An 813-bp PCR fragment containing P_ntcA (with the Shine-Dalgarno sequence for translation) was cloned into pTA2 (Toyobo), producing pHB4875. A 1,229-bp fragment containing hetZ was cloned into pMD18-T (TaKaRa), producing pHB4226. The second fragment was excised with NdeI/BamHI and cloned downstream of P_ntcA, with hetZ fused at the NdeI site (CATATG) to the ribosome-binding site at the end of P_ntcA, producing pHB4905. The P_ntcA-hetZ fragment was then cloned into pHB3729 (15), a derivative of pRL25C (37), producing pHB4909. Similarly, P_ntcA-hetP was generated and cloned into the pRL25C-derived vector, producing pHB4908. P_ntcA-hetZ-hetP was generated by overlap PCR (38) and cloned into pTA2, generating pHB4907, and then excised and cloned into pRL25C-derived vector, generating pHB4910.

(ii) Plasmids for expressing hetR or hetP from the indigenous promoter.

A 2,050-bp fragment containing the promoter and coding region of hetR was cloned into a plasmid derived from pRL25C, producing pHB4023. A 1,644-bp fragment containing the promoter and coding region of hetP was cloned into the pRL25C-derived plasmid, producing pHB4550.

(iii) Plasmids for replacing hetP or hetR with C.CE2 in Anabaena 7120.

The replaced region of hetP was the same as that deleted in the ΔhetP mutant reported by Higa and Callahan (14). An 1,149-bp PCR fragment which ends 15 bp upstream from hetP was cloned into pMD18-T, and a 935-bp PCR fragment which begins 17 bp downstream of hetP was cloned into pTA2. The second fragment was excised and inserted downstream of the first fragment in the same orientation in pMD18-T. Cassette C.CE2 excised from pRL598 (39) was inserted between the two fragments, and the resulting structure was then excised and cloned into pRL277 (40), producing pHB4526. The plasmid pHB4930 for replacing hetR with C.CE2 was generated in a similar way.

(iv) Plasmids for yeast two-hybrid tests.

hetZ excised from pHB4226 with NdeI/EcoRI was cloned into pGADT7 (Clontech), the prey vector, producing pHB4349, or was cloned into pGBKT7 (Clontech), the bait vector, producing pHB4350. Similarly, hetR was cloned into NdeI/XhoI-cut pGADT7 and pGBKT7, producing pHB4378 and pHB4379; hetP was cloned into NdeI/EcoRI-cut pGADT7 and pGBKT7, producing pHB4589 and pHB4554.

(v) Plasmids for overproducing recombinant proteins in Escherichia coli.

hetZ was cloned into pET21b (Novagen), producing pHB3142, or cloned into pET41a (Novagen) in translational fusion with a GST-encoding sequence, producing pHB4376. hetR was cloned into pET28a (Novagen) with an MBP-encoding sequence, producing pHB5040. A tsf-hetP translational fusion gene (encoding EF-Ts–HetP) was generated by overlap PCR using a PCR fragment carrying tsf from Escherichia coli BL21(DE3) and a PCR fragment carrying hetP from Anabaena 7120. tsf-hetP was cloned into pMD18-T and then recloned into pET21b, producing pHB5018.

Construction of Anabaena strains.

Plasmids were introduced into Anabaena 7120 and mutants by conjugation (41). Homologous double-crossover recombinants were generated based on positive selection with sacB (42). The plasmid pHB4526 was introduced into the wild type to replace hetP with cassette C.CE2 (43), generating the hetP deletion mutant; the plasmid pHB214 (20) was introduced into the ΔhetP strain to interrupt hetZ with the kanamycin/neomycin resistance cassette C.K2 (43). Similarly, the plasmid pHB4930 was used to generate the hetR deletion mutant. The complete segregation of mutants was confirmed by PCR. Anabaena strains are listed in Table S1.

Measurements of nitrogenase activity.

Nitrogenase activities were measured as described by Ernst et al. (44) with modifications. Anabaena cells grown in BG11 were washed 3 times with BG11₀ and resuspended in BG11₀, and 5 ml of cells was placed into a 20-ml stoppered vessel and incubated on a shaker in the light (30 μE m⁻² s⁻¹) at 30°C for 24 h or 48 h. For assays of anoxic nitrogenase activities, 3-(3,4-dichlorophenyl)-1,1-dimethyl urea (DCMU) was added to the cell suspension at a final concentration of 10 μM and the air in the vessel was replaced with argon. For aerobic nitrogenase activities, the cell suspension was used under aerobic conditions without addition of DCMU. Cells were incubated for 6 h without acetylene and for an additional 6 h after a 2-ml portion of acetylene was injected into the vessel. Acetylene reduction was assayed using gas chromatography. The nitrogenase activity was evaluated as the average rate of acetylene reduction (micromoles per milligram of Chla per hour) over 6 h. Data are means ± standard deviations of measurements using 3 parallel cultures.

RNA sequencing and differential expression analyses.

Anabaena 7120 was cultured in BG11 with air bubbling to an OD₇₃₀ of ≈0.8, washed three times with BG11₀, resuspended in the same medium to an OD₇₃₀ of about 0.8, and aerated. Total RNA was extracted by using TRIzol reagent (Invitrogen) from Anabaena strains 7120 and DRHB4526 (ΔhetP) and the hetZ::Tn5-1087b mutant at 24 h after nitrogen step-down with two independent biological repeats. DNA was eliminated with RNase-free DNase I (Promega). The quantity and quality of total RNA were evaluated using RNA electropherograms. To avoid biased depletion of coding RNAs, total RNA were sequenced without rRNA elimination. Strand-specific RNA-seq libraries were prepared using the Illumina small RNA sample preparation kit according to the manufacturer's protocol. RNA sequencing was carried out using the HiSeq 2500 sequencing instrument (Illumina) to generate paired-end reads with length of 125 bp.

A total of six RNA-seq data sets were obtained from two biological replicates of the three strains. Each data set contained over 100 million reads, over 2 million of which were retained for the following expression analyses after removing rRNA and tRNA reads. Reads were trimmed to 50 bp, since we found that this trimming increased the percentage of alignment. Reads were aligned against the Cyanobase (http://genome.microbedb.jp/cyanobase) genome sequence for Anabaena 7120 (assembly ID of GCA_000009705.1) using TopHat (45). The resulting BAM files were processed by bam2rpkm (http://bam2rpkm.sourceforge.net/) to create both raw read counts and reads per kilobase of exon per million mapped reads (RPKM values). Differential gene expression analysis was performed by using the DESeq package (46), which uses raw read counts as input. Fold changes of >2 or <0.5 with a threshold P value of <0.01 were defined as differentially expressed.

RT-qPCR.

Reverse transcription-quantitative PCR (RT-qPCR) was performed as previously described (47, 48). rnpB was used as the internal control. PCR primers (indicated with “RT” in the name) are listed in Table S1. Data are presented as means ± standard deviations resulting from tests of triplicate samples.

EMSA.

Expression of His₆-tagged EF-Ts-HetP and EF-Ts was induced in E. coli BL21(DE3) containing pHB5018 and pHB3831, respectively, purified from soluble proteins using the His.Bind purification kit (Novagen) and desalted using Microcon YM-3 (Millipore). DNA fragments F1 to F8 were generated by PCR using primer pairs all2928-F1-f/all2928-F1-r, all2928-F2-f/all2928-F2-r, all2928-F3-f/all2928-F3-r, alr0286-F4-f/alr0286-F4-r, alr0286-F5-f/alr0286-F5-r, alr0286-F6-f/alr0286-F6-r, all2928-F7-f/all2928-F1-r, and alr0286-F5-f/all2928-F8-r, respectively (Table S1). Biotin labeling of these DNA fragments was performed by second-round PCR using bio-f and each reverse primer (-r) mentioned above. PCR products were purified after agarose gel electrophoresis. Electrophoretic mobility shift assays (EMSA) were performed as described by Du et al. (15) with modifications. The mixture was incubated at 25°C for 20 min and separated by electrophoresis with a 6% nondenaturing polyacrylamide gel at 4°C. Labeled DNA fragments were then transferred onto a Hybond-N⁺ membrane and visualized.

Yeast two-hybrid tests.

The Matchmaker two-hybrid kit (Clontech) was used for yeast two-hybrid tests. Bait and prey plasmids were transformed into Saccharomyces cerevisiae strain AH109 using the polyethylene glycol-lithium acetate method (49). Screening of interaction clones was performed according to the manufacturer's instructions. Transformants were first selected on synthetic defined medium plates with Leu and Trp (SD/Leu/Trp plates) and confirmed by PCR; they then were streaked onto SD/Ade/His/Leu/Trp plates to check interactions at high stringency. To further test the interactions in yeast, transformants grown in liquid SD/Leu/Trp medium at 30°C were diluted 10-fold and allowed to grow to mid-log phase (OD₆₀₀ ≈ 0.5); 1-μl aliquots then were patched on SD/Ade/His/Leu/Trp plates with 5-bromo-4-chloro-3-indolyl-α-d-galactopyranoside (X-α-Gal) and incubated at 30°C for 3 days (50).

Pulldown assays.

His₆-tagged glutathione S-transferase (GST)–HetZ, GST, maltose-binding protein (MBP)–HetR, and MBP were purified from soluble proteins of E. coli BL21(DE3) containing pHB4376, pET41a, pHB5040, or pET28a-MBP using the His.Bind purification kit and desalted using Microcon YM-3.

To assay the interaction between HetZ and HetR, a GST-HetZ fusion or GST (as the negative control) was bound to Pierce glutathione-agarose (Thermo Fisher). The agarose beads were incubated with 100 μg of purified MBP-HetR fusion or MBP (as the negative control) in 0.5 ml of phosphate-buffered saline (PBS) (10 mM Na₂HPO₄, 20 mM NaCl, 68 mM KCl, 1.76 mM KH₂PO₄, pH 7.4) at 4°C for 1 h and washed with 1 ml of PBS buffer 10 times. The retained proteins were boiled off the beads in 100 μl of sample buffer for 10 min. Proteins were separated by SDS-PAGE (10%) and electroblotted onto an NC filter. MBP-HetR and MBP were detected with an anti-MBP monoclonal antibody (Abcam) and a horseradish peroxidase (HRP)-conjugated secondary antibody specific for mouse IgG (Thermo Fisher) and visualized by chemiluminescence generated by HRP in the presence of a luminol/enhancer solution and stable peroxide solution (Thermo Fisher). GE ImageQuant LAS 4000 (GE Healthcare) was used to photograph blots.

Accession number(s).

The RNA-seq data have been deposited in the ArrayExpress database at EMBL-EBI (www.ebi.ac.uk/arrayexpress) under accession number E-MTAB-6459.