Expression from DIF1-motif promoters of hetR and patS is dependent on HetZ and modulated by PatU3 during heterocyst differentiation.

HetR and PatS/PatX-derived peptides are the activator and diffusible inhibitor for cell differentiation and patterning in heterocyst-forming cyanobacteria. HetR regulates target genes via HetR-recognition sites. However, some genes (such as patS/patX) upregulated at the early stage of heterocyst differentiation possess DIF1 (or DIF+) motif (TCCGGA) promoters rather than HetR-recognition sites; hetR possesses both predicted regulatory elements. How HetR controls heterocyst-specific expression from DIF1 motif promoters remains to be answered. This study presents evidence that the expression from DIF1 motif promoters of hetR, patS and patX is more directly dependent on hetZ, a gene regulated by HetR via a HetR-recognition site. The HetR-binding site upstream of hetR is not required for the autoregulation of hetR. PatU3 (3' portion of PatU) that interacts with HetZ may modulate the expression of hetR, hetZ and patS. These findings contribute to understanding of the mutual regulation of hetR, hetZ-patU and patS/patX in a large group of multicellular cyanobacteria.

Introduction

Cyanobacteria were the first group of microorganisms that performed oxygenic photosynthesis [1, 2]. In the early earth environment, nitrogen nutrient was a limiting factor for propagation of microbes. Under this selective pressure, nif genes spread among bacteria, and some cyanobacteria acquired the N2 fixation capability. With the rise of atmospheric oxygen, certain filamentous species developed the capability to form specialized N2-fixing cells, called heterocysts, to protect nitrogenase from inactivation by oxygen [3-5]. Nowadays, heterocyst-forming cyanobacteria contribute significantly to nitrogen fixation in the earth’s biosphere [6-8]. In species from different genera of heterocyst-forming cyanobacteria, heterocysts are differentiated at one end, two ends, or intercalary positions of filaments [9]. Anabaena sp. PCC 7120 (hereafter Anabaena 7120) was derived from a species that produces semi-regularly spaced single heterocysts along non-branched filaments in response to nitrogen stepdown. It is the most often used model strain for molecular studies on heterocyst-related topics [10]. Other species used in such studies include Anabaena variabilis [11, 12], Nostoc punctiforme [13, 14], Nostoc ellipsosporum [15], etc.

Heterocyst differentiation and pattern formation largely depend on the key regulator HetR [16] and RGSGR-containing peptides, which are derived from PatS [17, 18], PatX [19] or HetN [20], representing an example of the most ancient activator-inhibitor (reaction-diffusion) patterning processes [21-23]. In Anabaena 7120, PatS is the main source of morphogen for de novo pattern formation [18], while HetN is required for maintenance of the pattern [24]. HetR is the only known target of RGSGR-containing peptides [25], and it binds to consensus recognition sites upstream of hetP [26, 27], hetZ [28] and several other genes, including its own encoding gene [29-31]. Among these genes, hetZ is involved in control of heterocyst differentiation at an early stage [32], and hetP is required for commitment to heterocyst differentiation [33]. hetZ and hetP functionally overlap with each other, and co-expression of these two genes was shown to restore heterocyst formation in hetR-minus mutants [34]. In a different substrain of Anabaena 7120, expression of hetZ alone restored heterocyst formation in a hetR-deletion mutant [35]. The variable requirement for hetP expression may depend on differences in genetic backgrounds of substrains [36]. hetP and hetZ are both upregulated in differentiating cells, as a result of the accumulation of HetR [26, 28]. patS is also upregulated in differentiating cells [17], but no consensus recognition site for HetR is present in the sequence upstream of patS.

Immediately downstream of hetZ in many filamentous cyanobacteria is a gene called patU; these two genes, together with hetR, are listed among the core set of genes for filamentous species [32, 37]. In Anabaena 7120, patU is split into patU5 and patU3 [32]. hetZ and patU3 play opposite roles in heterocyst differentiation: hetZ promotes, while patU3 inhibits [32].

Before the consensus HetR-recognition sequence was identified, DIF+ (later called DIF1) motif (TCCGGA) had been bioinformatically identified in sequences upstream of hetR and several other genes in Anabaena 7120 [38]. The role of DIF1 motif in heterocyst-specific expression was shown with the promoter of nsiR (a heterocyst-specific non-coding RNA) [38] and a synthetic minimal promoter [39]. More recently, the DIF1 motif was proposed as a consensus regulatory sequence (centered at -35 region) for patS and patX in heterocyst-forming cyanobacteria [19]. However, there are two questions to be answered. (1) What is the role of the predicted DIF1 motif promoters in upregulated expression of hetR, patS and patX? This must be examined experimentally. In particular, HetR-recognition site and DIF1 motif are both present upstream of hetR. (2) Which of HetR, HetZ and HetP is required for the regulation of DIF1-motif promoters? In Anabaena 7120, deletion of hetZ blocked the induced expression of hetR, hetP and patS, whereas hetP showed no effects on these genes [35]. This result excluded HetP as the factor for inducing the expression of hetR and patS; however, because hetR was not expressed in the hetZ mutant, which of HetR and HetZ is required for the upregulated expression of hetR and patS remained unclear. Earlier, the expression from DIF1 motif promoters had been shown to be dependent on a functional hetR [38, 40], but hetZ was not expressed in the hetR mutant either.

To elucidate the role of HetR and HetZ in control of DIF1 motif promoters during heterocyst differentiation, it is necessary to produce heterocysts without HetR or HetZ. In this study, we tested the expression from P and P in heterocysts without HetR and the role of DIF1 motif in expression of hetR and patS. We found that HetZ plays a more direct role in control of these promoters than HetR and that the expression of hetR and patS is mainly dependent on the DIF1-motif promoter sequences. In addition, PatU3 that interacts with HetZ may modulate the expression of hetR, hetZ and patS.

Materials and methods

General

Anabaena 7120 and derivatives (S1 Table) were cultured in BG11 medium in the light of 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 nitrogen stepdown, Anabaena 7120 grown in BG11 (OD730, 0.7~0.9) was collected by centrifugation, washed 3 times with BG110 (without nitrate) and resuspended in the same medium for 24 or 48 hours as indicated.

Microscopic observations

Microscopy was performed as previously described [41]. Photomicrographs were captured using an Olympus BX41 microscope (Olympus Corp., Tokyo, Japan) equipped with a JVC 3 CCD colour video camera (TK-C1381) (Victor Company of Japan Ltd, Tokyo, Japan). The GFP fluorescence was observed using a Sapphire GFP filter set (Exciter D395/40, dichroic 425DCLP, and emitter D510/40) (Chroma Technology Corp., Brattleboro, USA); autofluorescence was observed using the red long pass WG fluorescence cube (BP 510–550, BA590) from Olympus.

Construction of plasmids and Anabaena strains

Plasmid construction processes are described in S1 Table in the supplemental materials. DNA fragments cloned by PCR were confirmed by sequencing.

Plasmids were introduced into Anabaena 7120 and mutants by conjugation [42]. Homologous double-crossover recombinants were generated based on positive selection with sacB [43]. The complete segregation of mutants was confirmed by PCR. Anabaena strains and primers are listed in S1 Table.

Transcription analyses

Total RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, USA), and the residual DNA was removed with DNase RQ1 (Promega, Madison, USA). Reverse transcription was performed with the PrimeScript reverse transcription system (Takara, Dalian, China). RT-qPCR analyses were conducted as we described before [34]. rnpB (RNase P subunit B) was used as the internal control. PCR primers (indicated with ‘RT’ in name) are listed in S1 Table. Data are means ± SD produced from 3 technical or biological repeats as indicated.

Promoter activities were visualized using gfp (green fluorescence protein) as the reporter gene. Relative copy numbers of zeta- or pDU1-based plasmids (relative to rnpB) were evaluated by quantitative PCR as described in the reference [31] using primers gfp-1/gfp-2, pDU1-1/pDU1-2 and rnpB-1/rnpB-2 listed in S1 Table.

Rapid Amplification of cDNA Ends (RACE)

cDNA was synthesized with the SMART RACE cDNA amplification kit (Clontech, TaKaRa Bio., Otsu, Japan) using random primers. The 5′ end DNA fragments were generated by nested PCR as described by Zhang et al. [32], using universal primer/hetR-race-1 and nested universal primer/hetR-race-2 as the primers for 2 rounds of PCR. The universal primer and nested universal primer were provided with the SMART RACE cDNA amplification kit; hetR-race-1 and hetR-race-2 are listed in S1 Table. Transcription start points were determined based on sequencing of RACE products. Two biological repeats showed similar results.

Western blot analysis

Anabaena 7120 was deprived of fixed nitrogen for 24 h, harvested by centrifugation, washed with 20 mM Tris-HCl (pH 8.0) containing 1 mM PMSF and resuspended in the same buffer. Cells were broken with a French press (SCIENTZ, Ningbo, China) at 240 MPa (cell pressure) and centrifuged at 12,000 × g for 15 min. The supernatant was used as cell extracts for the Western blot analysis.

Proteins were separated by 12% SDS-PAGE and electro-blotted onto NC filters. HetR and HetZ were detected with rabbit antiserum against purified HetR or HetZ overproduced in Escherichia coli, visualized using alkaline phosphatase-conjugated secondary antibody specific for rabbit IgG (Thermo Scientific, Waltham, USA) with NBT and BCIP as substrates. Two biological repeats showed similar results.

Results

Upregulated expression from P and P in hetR-minus heterocysts

In a hetR-minus mutant, heterocyst differentiation is not initiated, and genes otherwise specifically expressed in heterocysts are mostly not upregulated after nitrogen stepdown. Such genes could be directly or indirectly regulated by HetR. Under our conditions, co-expression of hetZ and hetP from P (rather than expression of hetZ or hetP alone from the same promoter) enabled the hetR mutant, 7120hetR::C.CE2, to form functional heterocysts at the ends of filaments [34]. Such a phenotype was probably due to the lack of expression of patA, a gene required for heterocyst formation at intercalary positions, in vegetative cells of the hetR mutant [31]. Formation of functional hetR-minus heterocysts [34] indicated that genes required for the function of heterocysts could be properly expressed without HetR, but it gave no information about the regulation of P and P in (pro)heterocysts. Using gfp (green fluorescence protein) as the reporter gene, we tested the promoters of hetR and patS in (pro)heterocysts without HetR.

Plasmids carrying P-hetZ-hetP and the structure ‘Ω-promoter-gfp’ (the Ω cassette terminates background transcription, ref. [44]) were constructed and introduced into the hetR mutant. The tested promoters included P, P, P and P. hepB and hglD are genes involved in the formation of heterocyst envelope polysaccharide layer and glycolipid layer respectively, therefore P and P were included as the controls for heterocyst-specific expression [41]. Without a functional hetR, overexpression of hetZ and hetP led to heterocyst formation at the ends of filaments and upregulated expression of gfp from P, P, P and P in heterocysts relative to that in vegetative cells (Fig 1). Clearly, HetR is not essential for the expression from all these promoters.

Fig 1

Light (I), autofluorescence (II) and GFP fluorescence (III) photomicrographs of Anabaena 7120 hetR::C.CE2 harboring plasmids with P-hetP-hetZ and Ω-promoter-gfp.

On pHB6316, pHB6226, pHB6317 and pHB6318, gfp was expressed from P, P, P and P, respectively. Photomicrographs were taken at 24 h after nitrogen stepdown. Solid arrowheads point to heterocysts.

Upregulation of patS in heterocysts depends on the DIF1 motif and hetZ

Like hetR, hetZ and hetP, patS is upregulated in Anabaena 7120 shortly after nitrogen stepdown (S1A Fig). In the hetR mutant, patS could be upregulated by overexpression of hetZ rather than hetP (S1B Fig). Consistently, patS was upregulated in a ΔhetP mutant but not in a ΔhetZ mutant [35] or a hetZ::Tn5-1087b mutant [32] of Anabaena 7120. These results implied that the upregulation of patS is dependent on HetZ rather than HetR.

To confirm the role of HetZ in expression of patS, we further generated a partial deletion mutant, 7120hetZdel4-201, of Anabaena 7120 with 66 amino acids near the N-terminus of HetZ deleted in frame while preserving the putative promoter internal to hetZ serving patU5-patU3 [32]. This mutant showed no morphologically discernible heterocyst differentiation but formed some cells with less autofluorescence after nitrogen stepdown. These cells initiated differentiation, but the differentiation process ceased at the very early stage. The decreased autofluorescence was due to the degradation of phycobilisomes [45]. A non-replicative plasmid (pHB6069) containing P (-1070 ~ +48 relative to the translational start site of patS) upstream of gfp was integrated into the genomes of Anabaena 7120 and the derivative strain 7120hetZdel4-201 via homologous single-crossover recombination. Anabaena 7120::pHB6069 showed moderate expression of gfp specifically in (pro)heterocysts, whereas 7120hetZdel4-201::pHB6069 showed much weaker (but visible) expression of gfp in differentiating cells (Fig 2A).

Fig 2

Light (I), autofluorescence (II) and GFP fluorescence (III) photomicrographs of Anabaena 7120 and 7120hetZ del4-201, with gfp expressed from the full-length promoter or DIF1-motif promoter of patS.

Photomicrographs were taken at 24 h after nitrogen stepdown. Solid and empty arrowheads point to heterocysts and differentiating cells, respectively. Means ± SD are relative copy numbers of plasmids (relative to the copy number of rnpB in the genome). (A) Expression of gfp from the full-length patS promoter in the genome. The plasmid pHB6069 with P-gfp was integrated into the chromosome of Anabaena 7120 and the hetZ mutant via homologous single-crossover recombination. In the schematic diagram for the structure of full-length P fused to gfp, the bent line with an empty arrowhead indicates the transcription start point of the DIF1-motif promoter. (B) Expression of gfp from the minimal DIF1-motif promoter on zeta-based plasmids in Anabaena 7120 and the hetZ mutant. pHB6486 and pHB6458 are plasmids with the minimal DIF1-motif promoter of patS, in which TCCGGA was substituted or not. The stem-loop structure stands for the transcription terminator at the end of Ω cassette.

Employing gfp as a reporter gene in Anabaena 7120, we delimited the promoter of patS to the region -662 ~ -457 upstream of the start codon (S2 Fig, see photomicrographs for expression of gfp from fragments i, ii and iii). In this region, there is a DIF1 motif (TCCGGA) located 35 bp upstream of the tsp (transcriptional start point) -580 of patS [39]. We constructed a zeta-based plasmid with the minimal DIF1-motif promoter (a 41-bp fragment) positioned upstream of gfp (pHB6458) and a similar plasmid with TCCGGA replaced with GATATC (pHB6486). GFP was expressed in (pro)heterocysts of Anabaena 7120 [pHB6458] but not in differentiating cells of 7120hetZdel4-201 carrying the same plasmid; substitutions at TCCGGA abolished the expression of gfp in the wild-type strain (Fig 2B). These results established that activation of patS in (pro)heterocysts largely depends on HetZ and the DIF1-motif promoter. Similarly, expression from the DIF1-motif promoter of patX is also dependent on the function of hetZ (S3 Fig).

Upregulation of hetR in heterocysts depends on the DIF1 motif and hetZ

As shown with RT-qPCR, hetR was upregulated in the hetZ mutant 7120hetZdel4-201 at 6 h after nitrogen stepdown (S4 Fig). However, the expression of hetR in hetZ mutants was probably not patterned [32].

hetR is an autoregulated gene [46], and a potential HetR-binding site has been identified upstream of the tsp -271 (for heterocyst-specific expression) [28, 31]. Upstream of the same tsp, there is also a potential DIF1-motif promoter [38]. To clarify the role of the HetR-binding site and the DIF1 motif in regulation of hetR, we compared the expression of gfp from the promoter (-695 ~ -250 relative to the translational start site) of hetR and the same DNA fragment without the HetR-binding site or the DIF1 motif. Expression from the promoter of hetR was upregulated in (pro)heterocysts of Anabaena 7120, and the upregulated expression was abolished by substitutions at the DIF1 motif but not at the HetR-binding site (Fig 3).

Fig 3

Light (I) and GFP fluorescence (II) photomicrographs of Anabaena 7120 derivatives with gfp expressed from the wild-type or mutated promoter of hetR.

pHB6321: with the wild-type promoter (-695 ~ -250) of hetR; pHB6322: with GGGN (potential HetR-binding site) in the promoter of hetR substituted with AAAN; pHB6323: with TCCGGA (DIF1 motif) in the promoter of hetR substituted with CAATTG. Photomicrographs were taken at 24 h after nitrogen stepdown. Solid arrowheads point to heterocysts; means ± SD are relative copy numbers of plasmids.

To confirm the role of the DIF1 motif in heterocyst-specific expression of hetR, we constructed a zeta-based plasmid with the minimal DIF1-motif promoter (a 40-bp fragment) upstream of gfp (pHB6821) and introduced the plasmid into Anabaena 7120 and the hetZ mutant. As shown in Fig 4, GFP was expressed in (pro)heterocysts in Anabaena 7120 [pHB6821] but barely expressed in differentiating cells of the hetZ mutant. The copy numbers of zeta-based plasmids showed some changes in different strains but were still comparable with each other. Apparently, the upregulated expression of hetR in (pro)heterocysts is also mediated by HetZ via the DIF1 motif promoter.

Fig 4

Expression of gfp from the minimal DIF1-motif promoter of hetR on a zeta-based plasmid in Anabaena 7120 and 7120hetZdel4-201.

Top: the minimal sequence of DIF1-motif promoter cloned upstream of gfp in pHB6821. Photographs: light (I), autofluorescence (II) and GFP fluorescence (III) photomicrographs of Anabaena 7120 [pHB6821] and 7120hetZdel4-201 [pHB6821] at 24 h after nitrogen stepdown. hetZ4-201, 7120hetZdel4-201. Solid and empty arrowheads point to heterocysts and differentiating cells; relative copy numbers of plasmids are indicated as means ± SD.

We further generated a mutant of Anabaena 7120, P-DIF1-, with the DIF1 motif substituted with GATATC in the chromosomal DNA. Compared to the wild type, the P-DIF1- strain showed delayed heterocyst differentiation and lowered heterocyst frequency (Fig 5). Using RACE-PCR, we confirmed that the tsp at nucleotide -272 (-271 in previous reports [16, 47]) upstream of hetR disappeared in P-DIF1-. Clearly, the DIF1 motif is required for the heterocyst-specific expression of hetR and normal heterocyst differentiation.

Fig 5

Differences between Anabaena 7120 and the P-DIF1- strain in heterocyst differentiation and expression of hetR.

(A) Photomicrographs of Anabaena 7120 and the P-DIF1- strain at 24 h and 48 h after nitrogen stepdown. Frequencies of heterocysts/proheterocysts are indicated. (B) A stretch of sequence upstream of hetR, including the DIF1 motif, potential HetR-binding sequence and the tsp at -272.

PatU3 interacts with HetZ and modulates the expression of patS and hetR

hetZ and patU3 play opposite roles in heterocyst differentiation, whereas patU5 (which lies between hetZ and patU3) is not involved in heterocyst differentiation [32]. Employing the yeast two-hybrid system, we showed that PatU3 may interact with HetZ (Fig 6A-i); by a pull-down experiment, we confirmed the interaction between the two proteins (Fig 6B). As indicated in the two-hybrid assay, HetZ without the C-terminal portion no longer interacted with PatU3 (Fig 6A-ii).

Fig 6

Interaction of PatU3 with HetZ.

(A) The hetZ-patU5-patU3 region. The bent line with an arrowhead indicates the tsp upstream of hetZ-patU5-patU3. Additional tsps within hetZ for patU5-patU3 are not indicated. (B) Yeast two-hybrid assays of the interaction between PatU3 and HetZ. i) 1, pGBKT7-Lam + pGADT7-T, as the negative control; 2, pGBKT7-53 + pGADT7-T, as the positive control; 3, pGBKT7-PatU3 + pGADT7-HetZ. ii) 1, pGBKT7-PatU3 + pGADT7-HetZ[2–144]; 2, pGBKT7-PatU3 + pGADT7-HetZ[145–288]; 3, pGBKT7-PatU3 + pGADT7-HetZ[289–401]. Bracketed numbers (amino acid residue no.) indicate the portion deleted from HetZ (full length: 401 aa). (C) Pull-down assays of the interaction. Proteins were separated by SDS-PAGE (I) and analyzed with Western blot detection using anti-HA monoclonal antibody (II). 1, EF-Ts(HA)-HetZ; 2, MBP-PatU3 + MBP·Bind resin + EF-Ts(HA)-HetZ; 3, MBP-PatU3 + MBP·Bind resin + EF-Ts(HA); 4, MBP + MBP·Bind resin + EF-Ts(HA)-HetZ; 5, MBP + MBP·Bind resin + EF-Ts(HA); 6, EF-Ts(HA). (D) RT-qPCR analysis of mRNA abundance of hetR, patS and hetZ in Anabaena 7120 and the patU3::C.K4 mutant at 6 h after nitrogen stepdown. Data are means ± SD of 3 technical replicates. Asterisks indicate significant changes in mRNA abundance of hetR, hetZ and patS in the patU3 mutant compared to that in the wild type.

The interaction between PatU3 and HetZ may modulate HetZ-dependent gene expression. Based on RT-qPCR analysis, we compared the expression of hetR and patS in the wild type and the 7120patU3::C.K4 strain at 6 h after nitrogen stepdown (Fig 6C). The mRNA level of patS was greatly increased in the patU3 mutant relative to the wild type level, whereas that of hetR was slightly increased. Increased expression of patS probably inhibited the transcription of hetZ in the mutant (P-gfp in the mutant had shown a similar result, see ref. 32). However, the patU3::C.K4 mutation did not change the abundance of proteins HetR and HetZ in Anabaena filaments (S5 Fig).

Discussion

HetR and PatS-derived peptides are key players for heterocyst differentiation and patterning in Anabaena 7120. How their encoding genes are regulated is an important question for understanding the molecular mechanism of the differentiation/patterning process. In this study, we showed that the DIF1 motif plays an important role in regulation of these genes and that the expression from DIF1 promoters depends on the function of hetZ.

HetR is often considered as the master regulator of heterocyst differentiation, and it directly regulates the expression of hetP [26] and hetZ [28] in developing heterocysts via HetR-recognition sequences and is required for the expression of patA in vegetative cells [31]. Whether HetR directly regulates the expression of patS and its own gene was a problem to be clarified. By examining gene expression in hetR-minus heterocysts, we were able to show that HetR is non-essential for the upregulated expression from promoters of hetR and patS during heterocyst differentiation. Therefore, HetR may control the expression of these genes through other regulatory factors, such as HetZ.

First, we showed that DIF1-motif promoters are responsible for the upregulation of hetR, patS and patX in (pro)heterocysts. Substitutions at the DIF1 motif greatly reduced the transcription activity of P; a mutant of Anabaena 7120 with the DIF1 motif of hetR substituted in the genome showed no transcription from the tsp -272 (or -271, the heterocyst-specific tsp in the wild type [47]). Upstream of hetR, there is also a potential HetR-recognition site, but that site was shown to be not required for the upregulated expression. Second, we showed that hetZ is required for the upregulated expression from these DIF1-motif promoters. gfp fused to minimal DIF1-motif promoters of hetR, patS and patX was specifically expressed in (pro)heterocysts in the wild type, and the expression was greatly weakened in the 7120hetZdel4-201 strain. These results indicated that HetZ directly or indirectly regulates the expression of these genes via DIF1 motif promoters.

For the results we presented, two points need to be addressed in particular. (1) How to explain the upregulation of P in a hetR-minus background? Because HetR and the global nitrogen regulator NtcA are dependent on each other for upregulated expression during heterocyst differentiation [48], lack of HetR would keep ntcA from being upregulated. For this question, we think that NtcA and HetR do not directly regulate each other’s encoding gene. In at least one substrain of Anabaena 7120, NrrA mediates the regulation of hetR by NtcA [49, 50]. Actually, formation of functional heterocysts in the hetR mutant with P-hetZ-hetP implied that genes regulated by NtcA were properly expressed in developing cells. Presumptively, the expression of hetZ and hetP from P allowed sufficient expression of NtcA in those developing cells (the relationship between hetZ/hetP and ntcA awaits investigation), and NtcA in turn enhanced the expression of P-hetZ-hetP and indirectly upregulated P. (2) How to explain the differentiating cells in the 7120delhetZ4-201 mutant? In the hetZ mutant generated with Anabaena 7120 in our laboratory (substrain IHB), we found that the mRNA level of hetR was increased after nitrogen stepdown as in the wild type (S4 Fig), therefore the expression of hetR could have initiated cell differentiation that ceased at the very early stage in less regular pattern (consistent with the low expression of patS). This is a difference between the hetZ mutant generated in this study and that reported by Videau et al [35].

As a gene directly regulated by HetR, hetZ is involved in initiation of heterocyst differentiation and regulation of patS/patX and hetR. patX is not required for de novo heterocyst patterning in Anabaena 7120 (Du Y, Gao H and Xu X, unpublished), but its counterparts in most other heterocyst-forming species may play a role in heterocyst patterning. Therefore, HetR, HetZ and PatS/PatX form the core regulatory circuit in most heterocyst-forming cyanobacteria. This conclusion is important, because HetZ may provide an additional site for modulation of the expression of patS/patX, the sources of diffusible inhibitors for de novo pattern formation. PatU3 is a candidate for the modulator. It interacts with HetZ and somehow modulates the expression of hetR, hetZ and patS (Fig 6). Presumptively, interaction with PatU3 can regulate the cellular concentration of free HetZ therefore modulate HetZ-dependent gene expression. Alternatively, PatU3 may have additional functions that indirectly affect the expression of these genes. The core regulatory circuit of heterocyst differentiation in Anabaena 7120 is summarized in Fig 7. This coordination scenario involving multiple activating/inhibiting factors may help to refine the current models [51, 52] for heterocyst differentiation and patterning.

Fig 7

A schematic diagram showing the core regulatory circuit for heterocyst differentiation in Anabaena 7120.

Dark lines with a solid arrowhead indicate gene expression, processing of peptide or cell differentiation; grey lines with an open arrowhead (+) or T-shaped end (-) indicate activation or inhibition of gene expression or protein activity (solid lines for confirmed direct interaction/regulation, dashed lines for direct or indirect regulation). The diamond-Y fork indicates protein-protein interaction. Thickness and darkness of lines roughly indicate the strength of interaction/regulation. Processing of PatS, regulation of hetP and regulation of hetZ are described in references 18, 26 and 28. HetN (for maintenance of heterocyst pattern), PatA (for heterocyst formation at intercalary positions), PatX (not required for heterocyst patterning in Anabaena 7120), and other factors that affect heterocyst differentiation/patterning, are not shown here.

RT-qPCR analyses showing the upregulation of patS in Anabaena 7120 and the relationship between the expression hetZ and patS in a hetR-minus background.

(PDF)

Click here for additional data file.

Expression of gfp from fragments upstream of patS on a pDU1-based plasmid in Anabaena 7120.

(PDF)

Click here for additional data file.

Light (I), autofluorescence (II) and GFP fluorescence (III) photomicrographs showing the expression of gfp from the DIF1-motif promoter of P in Anabaena 7120 and 7120hetZdel4-201.

(PDF)

Click here for additional data file.

RT-qPCR analysis of the expression of patS and hetR in the wild type and the mutant 7120hetZdel4-201 at 0 and 6 h after nitrogen stepdown.

(PDF)

Click here for additional data file.

Western blot detection of HetR and HetZ in the wild type and the patU3 mutant of Anabaena 7120 at 24 h after nitrogen stepdown.

(PDF)

Click here for additional data file.

Anabaena strains, plasmids and primers.

(PDF)

Click here for additional data file.

(PDF)

Click here for additional data file.