Cross-Activation of Two Nitrogenase Gene Clusters by CnfR1 or CnfR2 in the Cyanobacterium Anabaena variabilis.

In Anabaena variabilis, the nif1 genes, which are activated by CnfR1, produce a Mo-nitrogenase that is expressed only in heterocysts. Similarly, the nif2 genes, which are activated by CnfR2, make a Mo-nitrogenase that is expressed only in anaerobic vegetative cells. However, CnfR1, when it was expressed in anaerobic vegetative cells under the control of the cnfR2 promoter or from the Co2+-inducible coaT promoter, activated the expression of both nifB1 and nifB2. Activation of nifB2, but not nifB1, by CnfR1 required NtcA. Thus, expression of the nif1 system requires no heterocyst-specific factor other than CnfR1. In contrast, CnfR2, when it was expressed in heterocysts under the control of the cnfR1 promoter or from the coaT promoter, did not activate the expression of nifB1 or nifB2. Thus, activation of the nif2 system in anaerobic vegetative cells by CnfR2 requires additional factors absent in heterocysts. CnfR2 made from the coaT promoter activated nifB2 expression in anaerobic vegetative cells grown with fixed nitrogen; however, oxygen inhibited CnfR2 activation of nifB2 expression. In contrast, activation of nifB1 and nifB2 by CnfR1 was unaffected by oxygen. CnfR1, which does not activate the nifB2 promoter in heterocysts, activated the expression of the entire nif2 gene cluster from a nifB2::nifB1::nifB2 hybrid promoter in heterocysts, producing functional Nif2 nitrogenase in heterocysts. However, activity was poor compared to the normal Nif1 nitrogenase. Expression of the nif2 cluster in anaerobic vegetative cells of Nostoc sp. PCC 7120, a strain lacking the nif2 nitrogenase, resulted in expression of the nif2 genes but weak nitrogenase activity. IMPORTANCE Cyanobacterial nitrogen fixation is important in the global nitrogen cycle, in oceanic productivity, and in many plant and fungal symbioses. While the proteins that mediate nitrogen fixation have been well characterized, the regulation of this complex and expensive process is poorly understood in cyanobacteria. Using a genetic approach, we have characterized unique and overlapping functions for two homologous transcriptional activators CnfR1 and CnfR2 that activate two distinct nitrogenases in a single organism. We found that CnfR1 is promiscuous in its ability to activate both nitrogenase systems, whereas CnfR2 depends on additional cellular factors; thus, it activates only one nitrogenase system.

INTRODUCTION

Biological nitrogen fixation is the reduction of atmospheric dinitrogen to ammonium, mediated by the oxygen-sensitive enzyme nitrogenase. Nitrogen fixation is found only in some bacterial and archaeal strains, but it is essential for the productivity of the oceans and is important for the growth of many plants that benefit from nitrogen fixation by rhizosphere bacteria or by symbiotic associations with nitrogen-fixing bacteria (1, 2). Cyanobacteria play a major role in nitrogen productivity in the oceans and form symbiotic associations with many plants and fungi (3, 4).

All cyanobacteria produce oxygen from photosynthesis; therefore, nitrogen-fixing strains must protect nitrogenase from oxygen. Nitrogen fixation is separated from photosynthesis either spatially in specialized microoxic cells or temporally when cells are not producing oxygen by photosynthesis (5–8). Cyanobacteria in the Nostocaceae differentiate specialized nitrogen-fixing cells called heterocysts but only under conditions of nitrogen starvation. Heterocysts provide a microoxic environment to protect nitrogenase, allowing the cyanobacterial filament to fix nitrogen in air (9–12). The microoxic environment of the heterocysts results from their cell wall structure, their lack of oxygen-evolving photosystem II, and their high rate of respiration (13–17). All heterocyst-forming cyanobacteria express a Mo-nitrogenase in heterocysts, but some strains also have a heterocyst-specific V-nitrogenase that is made only under conditions of Mo deficiency (18, 19).

The genes for nitrogen fixation in cyanobacteria are well conserved and typically include a few clusters of nif genes, some of them comprising over a dozen genes (20). These large clusters of genes have been best characterized in Anabaena variabilis ATCC 29413 (19), Leptolyngbya boryana (21, 22), and Cyanothece sp. (7, 23). A. variabilis is unusual because it has two nif-encoded Mo-nitrogenases. The nif1 genes are expressed only in heterocysts, while the nif2 genes are expressed only in anaerobic vegetative cells (19, 20, 24). The expression of the nif1 genes depends on the differentiation of heterocysts, a complex developmental process that requires a carefully regulated cascade of regulatory factors, including NtcA, HetR, NrrA, DevH, and other proteins (9, 11, 12, 25). NtcA, which senses changes in the cellular carbon-to-nitrogen ratio, is important not only in heterocyst differentiation but also in the global cellular nitrogen response (26, 27). While the nif1 and nif2 gene clusters are expressed in different cell types and under different conditions, they have a very similar gene organization comprising a large cluster of contiguous genes, nifB, nifS, nifU, nifH, nifD, nifK, nifE, nifN, nifX, nifW, hesA, and hesB, and a few genes of unknown function. The nif1 and nif2 gene clusters are primarily under the control of the promoter upstream of nifB, although nifU1 and nifE1 have weak promoters within the genes and hesA1 has its own strong promoter (20, 28, 29). A major difference between the two clusters is the interruption of the nifD1 gene by an 11-kb element that must be excised during heterocyst differentiation (30).

The large cyanobacterial nif clusters are regulated by the transcriptional activator CnfR. CnfR proteins all share a C-terminal XRE-type DNA-binding domain and two N-terminal 4Fe-4S-binding sites, similar to bacterial ferredoxins (21, 31–33). The cnfR gene was first described in Nostoc sp. PCC 7120 as patB, a gene that affected heterocyst pattern formation (32). In Nostoc sp. PCC 7120, the patB (cnfR) gene is expressed exclusively in heterocysts, and a deletion mutant grew very poorly in the absence of fixed nitrogen (31). The cnfR1 gene of A. variabilis is very similar to cnfR (patB) in Nostoc sp. PCC 7120. In the nonheterocystous strain L. boryana, deletion of either the conserved CnfR N-terminal iron-sulfur cluster-binding region or the C-terminal DNA-binding domain abolished expression of nifB (21, 22).

In A. variabilis, CnfR1 activates expression of the nifB1 promoter only in heterocysts, and CnfR2 activates expression of the nifB2 promoter only in anaerobic vegetative cells (33). Deletion of cnfR1 or cnfR2 prevents the expression of nifB1 or nifB2, respectively. Loss of cnfR1 has little effect on the expression of hesA1, which has its own promoter, whereas hesA2 expression is driven by the nifB2 promoter and is therefore inhibited by loss of cnfR2 (29, 33). Expression of cnfR2 in anaerobic vegetative cells requires NtcA; hence, an ntcA mutant does not make the Nif2 nitrogenase in anaerobic vegetative cells (20, 33, 34).

A region 300 to 600 bp upstream of the transcription start sites of nifB1 and nifB2 and upstream of the promoter of nifB in L. boryana have three conserved cis-acting elements that are thought to be the binding sites for CnfR proteins (22, 35). In A. variabilis, the loss of either the first or the second conserved site decreases transcription of nifB1 by about 50%, and loss of both of these conserved sites leads to a 75% reduction in niB1 expression (35). The transcription start sites for both the nifB1 and nifB2 promoters have a canonical “extended −10 promoter” (35, 36) but no −35 region, indicating that they are a type 2 promoter, which often requires an alternative sigma factor (37).

Several cyanobacterial sigma factors respond to various types of stress (38–40). In Nostoc sp. PCC 7120, SigB and SigC play a role in the expression of some NtcA-dependent, nitrogen-responsive genes (37), and sigC, sigE, and sigG expression are increased in heterocysts (41, 42). Both SigC (42) and SigE (43) are involved in the expression of nif genes in heterocysts, but they are not essential. HetR, the central activator of heterocyst differentiation, increases the expression of SigC. Overall, the data indicate that sigB, sigC, sigD, and sigE likely have roles, which may overlap, in the expression of nitrogen-regulated genes and possibly in nitrogen fixation.

We are interested in the mechanism of CnfR-mediated activation of expression of nifB and the possible role of other factors in nifB transcription. Since A. variabilis has two CnfR proteins that act in different cell types to activate two nifB promoters, we determined whether these two CnfR proteins could activate the noncognate nifB promoter, how cell type affected nifB activation, and how environmental conditions or factors, such as NtcA, NrrA, or HetR, affected CnfR-mediated activation of nifB1 and nifB2. We determined how hybrid CnfR1-CnfR2 proteins and a hybrid nifB2-nifB1 promoter affected nifB transcription. In addition, we showed that the nif2 genes of A. variabilis were expressed in Nostoc sp. PCC 7120, producing the Nif2 nitrogenase in anerobic vegetative cells.

RESULTS

CnfR proteins.

CnfR proteins are present in all sequenced nitrogen-fixing cyanobacteria, and strains with two Mo-nitrogenase gene clusters have two copies of cnfR (Fig. S1 and Table S1 in the supplemental material). Cyanobacterial CnfR proteins cluster into four phylogenetic groups, including one group of CnfR1 proteins in heterocyst-forming cyanobacteria and another that includes the CnfR2 proteins. CnfR2 proteins are more closely related to the CnfR protein in Chroococcidiopsis thermalis PCC 7203 than to CnfR proteins in heterocystous cyanobacteria lacking a nif2 system (Fig. S1) (33). The sequence similarities between CnfR proteins in the most divergent strains are about 60%. In A. variabilis, the CnfR1 and CnfR2 proteins share 61% identity and 80% overall similarity, but they are more similar to each other than to the CnfR proteins of most nonheterocystous cyanobacteria (Fig. S1). All CnfR proteins have an N-terminal Fe-S-binding domain and a C-terminal helix-turn-helix (HTH) domain. The C-terminal HTH domains of CnfR1 and CnfR2 in A. variabilis share 87% similarity, while the N-terminal Fe-S-binding domains share only 73% similarity (33).

The role of the DNA-binding domains in activation of nifB1 and nifB2 in heterocysts.

In A. variabilis, CnfR1 activates nifB1, the first gene in the large nif cluster, in heterocysts, while CnfR2 activates nifB2 in anaerobic vegetative cells starved for fixed nitrogen. Therefore, the promoters that drive expression of cnfR1 versus cnfR2 are only activated in heterocysts or anaerobic vegetative cells, respectively (33). We were interested in the ability of CnfR1, CnfR2, and hybrid CnfR1-CnfR2 proteins to cross regulate expression of nifB1 and nifB2; therefore, we expressed them from promoters that drove their expression under conditions in which cnfR1 and cnfR2 are not normally produced.

CnfR1 made under the control of its own promoter in heterocysts (strain BP920) at 24 h after nitrogen step-down activated the expression of nifB1 more than 200-fold compared to the ΔcnfR1/ΔcnfR2 control strain (BP894) but did not activate the expression of nifB2 (Fig. 1A). CnfR2 made under the control of the cnfR1 promoter in heterocysts (BP870) at 24 h after nitrogen step-down was induced 5-fold compared to anaerobic vegetative cells (Fig. S2) but failed to activate expression of either nifB1 or nifB2 (Fig. 1A). Replacement of the C-terminal HTH DNA-binding domain of CnfR2 by the HTH domain of CnfR1 (CnfR2-HTHCnfR1) in BP910 did not allow CnfR2 to activate nifB1 or nifB2 in heterocysts (Fig. 1A). This suggests that the difference in the DNA-binding domains of CnfR1 versus CnfR2 was not responsible for the inability of CnfR2 to activate nifB1 or nifB2 in heterocysts.

FIG 1

Cell-type-specific expression of nifB1 and nifB2 by CnfR1 and CnfR2, respectively. Expression of nifB1 or nifB2 by CnfR1, CnfR2, or a hybrid CnfR protein, driven by either the cnfR1 or cnfR2 promoter as indicated, was determined by RT-qPCR. (A, C) Cells were grown 24 h under −N +O2 conditions, leading to formation of heterocysts that activate the cnfR1 promoter. (B) Cells were grown 6 h under −N −O2 conditions, leading to vegetative cells that activate the cnfR2 promoter. Expression of nifB1 and nifB2 was normalized to rnpB, and the values on the y axis represent the relative log2 fold differences in expression. Nitrogenase activity (N2'ase) is expressed as nmoles of ethylene OD720−1 min−1. Statistical analysis: graphs, ***, P < 0.001; nitrogenase, all differences in values were significant (P < 0.001) except the two labeled nonsignificant (ns; P ≥ 0.05). The horizontal bars below the P values provide statistical comparisons for the means of the two values immediately below the ends of the bar and do not include values between these two.

The role of the DNA-binding domains in activation of nifB1 and nifB2 in vegetative cells.

CnfR2 made under the control of its own promoter (BP921) in anaerobic vegetative cells 6 h after nitrogen step-down activated the expression of nifB2 several thousand-fold (quantification cycle [ΔΔ] = 14) compared to the ΔcnfR1/ΔcnfR2 control strain (BP894). In contrast, CnfR2 did not activate the expression of nifB1 (Fig. 1B). Replacement of the HTH domain of CnfR2 by the HTH domain of CnfR1 (CnfR2-HTHCnfR1) in BP873 decreased the expression of nifB2 about 4-fold but did not increase expression of nifB1 (Fig. 1B). Expression of CnfR1, driven by the cnfR2 promoter (BP871), in anaerobic vegetative cells 6 h after nitrogen stepdown was induced 3-fold compared to cells grown aerobically for 24 h (Fig. S2), but it strongly activated the expression not only of nifB1 but also of nifB2 (Fig. 1B). Expression of CnfR1-HTHCnfR2 driven by the cnfR2 promoter (BP907) increased both nifB1 and nifB2 expression about 3-fold compared to CnfR1 driven by the cnfR2 promoter (Fig. 1B). Thus, CnfR2 can activate only nifB2 in anaerobic vegetative cells, while CnfR1 can activate nifB1 and nifB2. Replacement of the DNA-binding domain of CnfR1 with the similar region of CnfR2 improved activation of both the nifB1 and nifB2 promoters by CnfR1 in anaerobic vegetative cells, where it is not normally expressed.

Expression of nitrogenase in vegetative cells.

Although nifB1 was expressed in anaerobic vegetative cells of BP871 and BP907 at 6 h after nitrogen step-down (Fig. 1B), the 11-kb excision element interrupting the nifD1 gene would not have been excised; hence, no Nif1 nitrogenase would be made. However, activation of nifB2 by CnfR1 (BP871) or CnfR1-HTHCnfR2 (BP907) resulted in Nif2 nitrogenase activity, although activity was lower than in cells that activated nifB2 using CnfR2 (BP921) or CnfR2-HTHCnfR1 (BP873) (Fig. 1B). This decrease could be due to missing proteins whose genes are likely not activated by CnfR1 or due to the production of noncognate Nif1 proteins upstream of the 11-kb excision element in nifD1, that is, NifB1, NifS1, NifU1, NifH1, and possibly a truncated NifD1 in addition to the normal NifB2, NifS2, NifU2, NifH2, and NifD2, which may decrease the activity of the Nif2 nitrogenase.

Effect of hybrid CnfR1/CnfR2 proteins on nifB1 and nifB2 expression in heterocysts.

Since CnfR1 activates nifB1 in heterocysts, but CnfR2 does not, we determined the ability of CnfR1-CnfR2 hybrid proteins to activate nifB1 in heterocysts. Hybrid proteins that contained mostly CnfR2 sequences (BP870, BP910, BP950, and BP951) showed no activation of nifB1 in heterocysts and no significant nitrogenase activity (Fig. 1C). However, BP952, with a larger region from CnfR1, increased nifB1 expression at least 10-fold compared to BP951, which had a smaller region from CnfR1 (Fig. 1C). BP952 also produced high levels of nitrogenase activity, suggesting that the larger CnfR1 region included in BP952 compared to BP951 plays a role in the specific activation of nifB1. While CnfR2 (BP870) poorly induced the expression of nifB2 in heterocysts, two hybrids that included the N-terminal Fe-S domain of CnfR1 (BP950 and BP951) improved nifB2 expression in heterocysts about 4-fold. However, the hybrid CnfR1-CnfR2 protein that most increased the expression of nifB1 (BP952) decreased the expression of nifB2. It appears that the N-terminal half of CnfR1 is important in determining the specificity of activation of nifB1 versus nifB2.

Expression of cnfR1 and cnfR2 under the control of the coaT promoter.

The previous experiments that substituted the promoters of cnfR1 and cnfR2 for each other provided information on the activation of nifB1 and nifB2 expression only in cell types in which these cnfR1 and cnfR2 promoters were expressed. However, we could not determine the effects of fixed nitrogen on the activation of nifB1 or nifB2, since neither the cnfR1 nor the cnfR2 promoter is expressed in cells grown with fixed nitrogen; therefore, we tested the coaT promoter, which can be induced by Co2+ or Zn2+ (44–46). We constructed a strain of A. variabilis with P::lacZ (BP1141) and measured β-galactosidase activity 2 h after the addition of various concentrations of Co2+ or Zn2+ (Fig. 2A and B). Concentrations of either metal in the range of 2 to 10 μM induced lacZ; however, Co2+ induced about 10-fold more β-galactosidase than did Zn2+. We induced strains with P::cnfR1 (BP1142) or with P::cnfR2 (BP1143) with 2 μM Co2+, which resulted in a 20- to 50-fold increase in the expression of cnfR1 or cnfR2 (Fig. 2C).

FIG 2

Induction of the P promoter by Co2+ or Zn2+. (A) Expression of β-galactosidase in strain BP1141 containing P::lacZ after induction with Co2+ for 2 h. (B) The same as A but induced with Zn2+. (C) Expression of cnfR1 or cnfR2 was determined by RT-qPCR. P::cnfR1 (BP1142) was grown for 24 h under −N +O2 conditions and induced with 2 μM Co2+ for 4 h, and P::cnfR2 (BP1143) was grown for 4 h under −N −O2 conditions and simultaneously induced with 2 μM Co2+. Expression of cnfR1 and cnfR2 was normalized to rnpB, and the values on the y axis represent the relative log2 fold differences in expression. P values of ≥0.05 are not shown; **, P < 0.01 or ***, P < 0.001. The horizontal bars below the P values provide statistical comparisons for the means of the two values immediately below the ends of the bar and do not include values between these two. The units for β-galactosidase activity are described in Materials and Methods.

Induction of cnfR1 in strain BP1142 by the addition of Co2+ for 4 h starting 20 h after the removal of fixed nitrogen (when heterocysts had already formed) induced about 8-fold more nifB1 (Fig. 3A) than what was observed in a strain in which cnfR1 was expressed from its native promoter (compare BP1142 in Fig. 3A with BP920 in Fig. 1A). CnfR1 activated nifB1 in cells grown with fixed nitrogen (Fig. 3A), indicating that nifB1 was expressed in vegetative cells. However, nifB1 expression in filaments with heterocysts (24 h, +O2 −N) was about 2-fold higher than in filaments without heterocysts (24 h, +O2 +N) (Fig. 3A). Since heterocysts comprise only about 10% of the cells, a 2-fold increase that results from activation only in heterocysts is about a 20-fold increase compared to vegetative cells. Expression of CnfR1 4 h after Co2+ induction in vegetative cells grown with or without fixed nitrogen and with or without oxygen resulted in activation of nifB1 expression (Fig. 3A). While the presence of fixed nitrogen had little effect on the levels of nifB1, oxygen decreased expression of nifB1 about 3- or 4-fold. CnfR1 also activated expression of nifB1 in an ntcA mutant (BP7142) (Fig. 3A), which lacks this key regulatory protein in nitrogen metabolism that is required for heterocyst differentiation (47). Thus, CnfR1 can activate nifB1 in vegetative cells under aerobic or anaerobic conditions in the presence of fixed nitrogen, and this activation does not require NtcA.

FIG 3

Expression of nifB1 and nifB2 by CnfR1 or CnfR2 under various environmental conditions. Expression of nifB1 (A) or nifB2 (B) in P::cnfR1 (BP1142) and nifB1 (C) or nifB2 (D) in P::cnfR2 (BP1143) were determined by RT-qPCR in cells grown for 4 h under ±N ±O2 conditions or for 24 h under −N +O2 conditions (to induce heterocysts). The P promoter was induced by the addition of Co2+ for 4 h before acetylene reduction assays, and cells were harvested for RNA. The P::cnfR1 ΔntcA strain was BP7142, and the P::cnfR2 ΔntcA strain was BP7143. Expression of nifB1 and nifB2 was normalized to rnpB, and the values on the y axis represent the relative log2 fold differences in expression. Nitrogenase activity (N2'ase) is expressed as nmoles of ethylene OD720−1 min−1. Note that nitrogenase data for panels A and B and for C and D are identical because they are from the same experiments. P values of ≥0.05 are not shown; **, P < 0.01 or ***, P < 0.001. The horizontal bars below the P values provide statistical comparisons for the means of the two values immediately below the ends of the bar and do not include values between these two.

In contrast, expression of CnfR1 at 4 h after Co2+ induction in vegetative cells grown with or without fixed nitrogen and with or without oxygen resulted in activation of nifB2 only in cells starved for fixed nitrogen, while oxygen had relatively little effect on the expression of nifB2 (Fig. 3B). CnfR1 failed to activate the expression of nifB2 in an ntcA mutant (BP7142), indicating that NtcA was required for activation of nifB2 by CnfR1 and that repression of ntcA expression by fixed nitrogen prevented nifB2 activation in cells grown with fixed nitrogen.

The experiments described above demonstrated that CnfR2 could not activate nifB1 in heterocysts or vegetative cells (BP870, Fig. 1A; BP921, Fig. 1B), and this was confirmed by the lack of expression of nifB1 in Co2+-induced cells expressing CnfR2 from the coaT promoter in BP1143 (Fig. 3C). However, 4 h after Co2+ induction, expression of CnfR2 in cells grown with or without fixed nitrogen resulted in strong activation of nifB2 only in anaerobic cells (Fig. 3D). In anaerobic cells, fixed nitrogen decreased nifB2 expression about 5-fold. While NtcA is required for the expression of cnfR2 (33), CnfR2 made under the control of the coaT promoter in an ntcA mutant activated the expression of nifB2, although expression was decreased severalfold compared to the ntcA+ strain. Although NtcA was not required for nifB2 expression in anaerobic cells grown without fixed nitrogen, there was no nitrogenase activity (Fig. 3D; Table 1).

TABLE 1

Nif2 nitrogenase activity in ntcA, hetR, or nrrA mutants

Strain	Mutationa	Growth conditions	Nitrogenase: nmoles ethylene OD720−1 min−1
BP1143	WT	−O2 −N, 4 h	0.96 ± 0.05
	WT	−O2 +N, 4 h	<0.01
BP7143	ntcA	−O2 −N, 4 h	<0.01
	ntcA	−O2 +N, 4 h	<0.01
BP1107	hetR	−O2 −N, 6 h	0.69 ± 0.03
	hetR	−O2 +N, 6 h	0.01 ± 0.002
BP1108	nrrA	−O2 −N, 6 h	0.61 ± 0.04
	nrrA	−O2 +N, 6 h	<0.01

WT, wild-type.

Nif2 nitrogenase activity in hetR or nrrA mutants.

Several regulatory proteins regulate the response to starvation for fixed nitrogen, including NtcA, NrrA, and HetR. We showed that a ntcA mutant reduced the expression of nifB2 by CnfR2 but did not abolish expression; however, the mutant did not fix nitrogen (Fig. 3D; Table 1). To determine whether NrrA or HetR affected Nif2 expression, we constructed mutants lacking these genes. The hetR mutant (BP1107), starved for fixed nitrogen for 6 h under anaerobic conditions, fixed nitrogen well compared to cells grown with fixed nitrogen (Table 1). The nrrA mutant (BP1108) gave similar results (Table 1). Hence, neither NrrA nor HetR is essential for the expression of the Nif2 system.

FIG 4

Expression of a hybrid nifB2::nifB1::nifB2 promoter by CnfR1 or CnfR2 in vegetative cells. (A) Organization of the SM54 hybrid nifB2::nifB1::nifB2 promoter fused to lacZ. CS1, CS2, and CS3 are highly conserved regions. The nifB1 sequence in the red box with the −10 region is indicated by the red portion of SM54; tss, transcription start site. (B, C) Expression of a hybrid nifB2::nifB1::nifB2 promoter fused to lacZ by CnfR1 in strain BP2197 (B) or CnfR2 in strain BP3197 (C) was determined by β-galactosidase assays in cells grown for 4 h under ±N ±O2 conditions. The P promoter was induced by the addition of Co2+ for 4 h before β-galactosidase assays. Statistical analysis: ***, P < 0.001; **, P < 0.01; *, P < 0.05. Nonsignificant comparative values (P ≥ 0.05) are not labeled. The horizontal bars below the P values provide statistical comparisons for the means of the two values immediately below the ends of the bar and do not include values between these two.

Expression of a hybrid nifB2::nifB1::nifB2 promoter by CnfR1 or CnfR2 in vegetative cells.

The nifB1 and nifB2 promoter regions share several highly conserved regions, CS1, CS2, and CS3 (Fig. 4A) (22, 35). We have shown that a hybrid nifB2::nifB1::nifB2 promoter that is mostly nifB2 but has a short nifB1-specific region near the transcription start site (SM54) (Fig. 4A) can be activated by CnfR1 in heterocysts and by CnfR2 in anaerobic vegetative cells (35). Here, we determined whether oxygen or fixed nitrogen affected the activation of the nifB2::nifB1::nifB2 hybrid promoter driving the expression of lacZ in vegetative cells by CnfR1 or CnfR2 made from the Co2+-inducible coaT promoter. The effects of oxygen or fixed nitrogen on the pattern of expression of the nifB2::nifB1::nifB2 promoter were very similar to nifB1 or nifB2. CnfR1 activated the nifB2::nifB1::nifB2 promoter in vegetative cells with or without fixed nitrogen or oxygen; however, the highest levels of expression of lacZ were in vegetative cells grown anaerobically without fixed nitrogen (Fig. 4B). This pattern is similar to the activation of nifB1 by CnfR1 made under the control of the coaT promoter under the same conditions (Fig. 3A) and different from the pattern of activation of nifB2 by CnfR1 made under the control of the coaT promoter (Fig. 3B). Even though the hybrid promoter is mostly nifB2, its activation by CnfR1 in vegetative cells did not show the requirement for NtcA that was evident for activation of nifB2 by CnfR1 (Fig. 3B). CnfR1 activation of the hybrid promoter was not affected by fixed nitrogen, whereas CnfR1 activation of the nifB2 promoter was affected by nitrogen. The only difference between the two promoters is the 50-bp nifB1 region near the transcription start site.

CnfR2 activated the nifB2::nifB1::nifB2 promoter in vegetative cells with or without fixed nitrogen but failed to activate the promoter in the presence of oxygen (Fig. 4C); however, the highest levels of expression were for cells grown without oxygen and fixed nitrogen (Fig. 4C). This pattern for activation of the nifB2::nifB1::nifB2 promoter is similar to the activation of nifB2 by CnfR2 made under the control of the coaT promoter under the same conditions (Fig. 3D), which is not surprising since most of the promoter is nifB2.

Expression of the Nif2 nitrogenase from the hybrid nifB2::nifB1::nifB2 promoter in heterocysts.

The Nif1 nitrogenase functions well in heterocysts; therefore, it was difficult to determine whether the Nif2 nitrogenase could function in heterocysts unless the Nif1 nitrogenase was absent. Nif2 expression in heterocysts could not be determined using the native nifB2 promoter, which is not expressed in heterocysts, but could be determined if nifB2 was expressed under the control of the nifB2::nifB1::nifB2 hybrid promoter. In strain BP1101, nifB2 is driven by the nifB2::nifB1::nifB2 hybrid promoter in JE9, a strain with a large deletion of nifDKE1, eliminating the Nif1 nitrogenase. While neither CnfR1 nor CnfR2 expressed in heterocysts 24 h after nitrogen step-down induced the expression of nifB2 from its native promoter (see BP910 and BP870 in Fig. 1A), CnfR1 made in heterocysts from its own promoter activated nifB2 expression from the nifB2::nifB1::nifB2 hybrid promoter. In the absence of heterocysts (24 h +O2 +N or 6 h +O2 −N), there was no expression of nifB2 from the hybrid promoter (Fig. 5A). While active Nif2 nitrogenase was synthesized in heterocysts of BP1101, nitrogenase activity was significantly lower in BP1101 than in wild-type strain FD (Fig. 5B).

FIG 5

Expression of nifB2 and the Nif2 nitrogenase in heterocysts. (A) Expression in strain BP1101 of nifB2 under the control of the hybrid nifB2::nifB1::nifB2 promoter was determined by RT-qPCR in cells grown for 24 h under −N +O2 conditions (with heterocysts), 24 h under +N +O2 conditions (no heterocysts), or 6 h under −N +O2 conditions (no heterocysts). Expression of nifB2 was normalized to rnpB, and the values on the y axis represent the relative log2 fold differences in expression. (B) Nitrogenase activity (N2'ase), expressed as nmoles of ethylene OD720−1 min−1, was determined for wild-type strain FD, JE9 (nifDKE1 deletion strain), and BP1101 grown for 24 h under −N +O2 conditions. Statistical analysis: ***, P < 0.001; **, P < 0.01. The horizontal bars below the P values provide statistical comparisons for the means of the two values immediately below the ends of the bar and do not include values between these two.

CnfR1 activation of the nifZ2 gene.

We showed previously that genes far downstream of nifB2, including nifEN2 (a fused gene) and hesA2, were not expressed in a cnfR2 mutant, suggesting that they are under the control of the nifB2 promoter (33). The nifZT2 genes, which are divergently transcribed from nifB2, were also not expressed in a cnfR2 mutant. Since expression of CnfR1 from the cnfR2 promoter (BP871) in anaerobic vegetative cells at 6 h after nitrogen step-down induced strong expression of nifB2 (Fig. 1B), we measured CnfR1-induced expression of nifEN2, hesA2, and nifZ2. Expression of all three genes was activated by CnfR1 in anaerobic vegetative cells nearly as strongly as by CnfR2 in the wild-type strain FD (Fig. 6). Since nifZ2 is divergently transcribed from nifB2, it must have its own promoter, which is activated by CnfR1 as effectively as is the nifB2 promoter. Thus, all the nif2 genes can be activated by CnfR2 or CnfR1 (if expressed) in anaerobic vegetative cells.

FIG 6

Expression of nifEN2, hesA2, and nifZ2. Expression of nifEN2, hesA2, and nifZ2 by CnfR2 (FD), by CnfR1 expressed from the cnfR2 promoter (BP871), and with neither CnfR1 nor CnfR2 (BP894) was determined by RT-qPCR in cells grown for 6 h under −N −O2 conditions. Expression of nifEN2, hesA2, and nifZ2 was normalized to rnpB, and the values on the y axis represent the relative log2 fold differences in expression. Statistical analysis: ***, P < 0.001; **, P < 0.01; *, P < 0.05. The P values for data for BP894 compared to the other two strains were <0.001. The horizontal bars below the P values provide statistical comparisons for the means of the two values immediately below the ends of the bar and do not include values between these two.

Nif2 nitrogenase expression in Nostoc sp. PCC 7120.

We have shown previously that CnfR2 expressed from a plasmid in Nostoc sp. PCC 7120 activated expression of the nifB2 promoter in anaerobic vegetative cells (33). Using a fosmid from a genomic library for A. variabilis that has the entire nif2 cluster (Fig. 7A), we constructed a mobilizable plasmid with this nif2 cluster, which was integrated into the chromosome of Nostoc sp. PCC 7120, creating strain BP893. In anaerobic vegetative cells, BP893 expressed cnfR2 at nearly the same levels as the A. variabilis FD control strain; however, expression of nifB2 and nifH2 was reduced significantly in BP893 (Fig. 7B). Nitrogenase activity, which was absent in anaerobic vegetative cells of the Nostoc sp. PCC 7120 parent strain (data not shown), was present in BP893; however, activity was reduced to about 5% of the activity of the A. variabilis FD control strain (Fig. 7C).

FIG 7

Expression of nif2 genes and nitrogenase activity in Nostoc sp. PCC 7120. (A) The nif2 genes in BP893. (B) Expression of cnfR2, nifB2, and nifH2 in wild-type strain FD and in BP893, a strain of Nostoc sp. PCC 7120 with the large nif2 cluster integrated in the chromosome, was determined by RT-qPCR in cells grown for 6 h under −N −O2 conditions. Expression of cnfR2, nifB2, and nifH2 was normalized to rnpB, and the values on the y axis represent the relative log2 fold differences in expression. (C) Nitrogenase activity was determined for wild-type strain FD and BP893 grown for 6 h under −N −O2 conditions. Statistical analysis: ***, P < 0.001; **, P < 0.01; ns, P > 0.05. The horizontal bars below the P values provide statistical comparisons for the means of the two values immediately below the ends of the bar and do not include values between these two.

DISCUSSION

CnfR proteins, which have an N-terminal Fe-S-binding domain and a C-terminal HTH domain, are present in all sequenced nitrogen-fixing cyanobacteria, and strains with two Mo-nitrogenase gene clusters have two copies of cnfR (20, 21, 33). In A. variabilis, the CnfR1 and CnfR2 proteins that control expression of nifB1 and nifB2, respectively, are more similar to each other than to the CnfR proteins of most nonheterocystous cyanobacteria. CnfR2, made only in anaerobic vegetative cells, cannot activate nifB1 (33); however, when CnfR2 was expressed in heterocysts under the control of the cnfR1 promoter, it did not activate expression of nifB1 or nifB2 (Fig. 1A). This suggests that a factor present in anaerobic vegetative cells, but not in heterocysts, is essential for CnfR2 activation of nifB2. In contrast, when CnfR1, which is normally expressed only in heterocysts, was made in anaerobic vegetative cells under the control of the cnfR2 promoter, it activated expression of both nifB1 and nifB2 (Fig. 1B). Therefore, CnfR1 requires no additional heterocyst-specific factor for activation of nifB1 expression, whereas CnfR2 does require additional factor(s) to activate nifB2 expression in anaerobic vegetative cells.

It seemed likely that differences in the HTH DNA-binding domains of CnfR1 and CnfR2 specified their binding to the nifB1 and nifB2 promoters, respectively. However, exchanging these HTH domains had little effect on the ability of either protein to activate its noncognate promoter when expressed in either heterocysts or in anaerobic vegetative cells (Fig. 1A and B). However, CnfR1-HTHCnfR2, expressed in anaerobic vegetative cells under the control of the cnfR2 promoter, increased expression of nifB1 and nifB2 compared to CnfR1, suggesting that the HTH domain of CnfR2 binds better to both promoters, perhaps because the HTH domain interacts with another transcription factor unique to anaerobic vegetative cells. Therefore, while the DNA-binding domains of CnfR1 versus CnfR2 do not provide specificity for activation of their cognate promoters, they do impact the efficacy of activation. Other CnfR1-CnfR2 hybrid proteins that were mostly CnfR2 did not induce nifB1 expression in heterocysts, but hybrids that included longer regions of the N-terminal half of CnfR1 marginally increased nifB1 expression compared with those with a shorter N-terminal region from CnfR1 (Fig. 1C). While none of the CnfR1-CnfR2 hybrid proteins activated expression of nifB1 or nifB2 in heterocysts well, the N-terminal half of CnfR1 plays a role in determining the specificity of activation of nifB1 versus nifB2.

Exchanging promoters allowed us to determine whether CnfR1 and CnfR2 functioned in anaerobic vegetative cells and in heterocysts, respectively. However, to determine the role of environmental factors, including fixed nitrogen and oxygen, in the activation of the nifB1 and nifB2 promoters, we needed to control the expression of cnfR1 or cnfR2. Expression of cnfR1 or cnfR2 from the Co2+-inducible coaT promoter (46, 48) led to a 20- to 50-fold induction by Co2+ (Fig. 2C). Co2+-induced expression of CnfR1 from the coaT promoter activated the nifB1 promoter in vegetative cells, and expression was somewhat sensitive to oxygen but insensitive to fixed nitrogen (Fig. 3A). In contrast, CnfR1 activated the nifB2 promoter only in cells starved for fixed nitrogen, but oxygen had little effect (Fig. 3B).

CnfR2 expressed from the coaT promoter activated the nifB2 promoter strongly only in anaerobic vegetative cells but did not activate nifB1 in heterocysts or aerobic vegetative cells (Fig. 3C and D). Although the nif2 genes were expressed in anaerobic vegetative cells with fixed nitrogen, nitrogenase activity was poor (Fig. 3D), which suggests that other proteins absent in nitrogen-replete cells (e.g., NtcA) support nitrogenase activity, and/or posttranslational modifications inhibited nitrogenase activity, as has been shown for the Nif1 nitrogenase (49, 50). Oxygen inhibited CnfR2 activation of nifB2 expression (Fig. 3D). Since nifB2 transcripts were not affected by oxygen in cells expressing CnfR1, oxygen either inhibits CnfR2 directly, or another transcription factor made only in anaerobic vegetative cells is required for nifB2 expression. While CnfR2 may be inactivated by oxygen, the fact that CnfR2 cannot activate expression of nifB2 in heterocysts, which have a low oxygen tension, suggests that a factor induced in vegetative cells by anaerobiosis might be required for CnfR2 activation of nifB2.

NtcA is a key regulatory protein in nitrogen metabolism that is required for heterocyst differentiation (47). Therefore, neither cnfR1 nor nifB1 is normally expressed in an ntcA mutant because it has no heterocysts. Using the coaT promoter, we expressed CnfR1 in an ntcA mutant and found that it activated the expression of nifB1 but not nifB2 in anaerobic vegetative cells (Fig. 3A and B). This is consistent with the fact that nifB2 was not expressed in cells grown with fixed nitrogen, which represses expression of NtcA (47). NtcA is required for expression of cnfR2 (33); however, CnfR2 made under the control of the coaT promoter in an ntcA mutant activated strong expression of nifB2 (Fig. 3D), but there was no nitrogenase activity. Thus, NtcA is important for the Nif1 nitrogenase because it is required for heterocyst formation, and cnfR1 and nifB1 are expressed only in heterocysts. In contrast, NtcA has a more direct role in the Nif2 system since it is required for expression of cnfR2 and for active nitrogenase, but it is not essential for transcription of nifB2.

We were interested in the role of the nifB1 and nifB2 promoters in cell-type-specific expression of nif1 versus nif2. The nifB1 and nifB2 promoter regions share several highly conserved regions, CS1, CS2, and CS3 (22, 35). We previously described a hybrid nifB2::nifB1::nifB2 promoter that is primarily nifB2, but has a short region of nifB1 around the transcription start site (Fig. 4A) that can be activated by CnfR1 in heterocysts and by CnfR2 in anaerobic vegetative cells (35). When the hybrid promoter was activated by CnfR1 or CnfR2 made in all cells from the coaT promoter, expression levels for the hybrid promoter were similar to those for nifB1 or nifB2, respectively (Fig. 4B and C). However, unlike the nifB2 promoter, activation of the hybrid promoter by CnfR1 in vegetative cells did not require NtcA. Similarly, unlike the nifB2 promoter, CnfR1 activation of the hybrid promoter was not affected by fixed nitrogen, which decreases NtcA expression. Thus, the NtcA requirement for activation of nifB2 by CnfR1 in vegetative cells is associated with the nifB2 region around the transcription start site, which is missing in the hybrid promoter. CnfR2 activation of the nifB2::nifB1::nifB2 promoter, like the nifB2 promoter, occurred only in anaerobic vegetative cells, with or without fixed nitrogen, although fixed nitrogen decreased transcription of both promoters. This implies that the transcription start site region of nifB2 (missing in the hybrid promoter) is not involved in the requirement for anaerobic conditions for activation by CnfR2 but is required for NtcA-dependent nitrogen control of nifB2 activation by CnfR1.

Although the Nif2 nitrogenase normally functions in anoxic vegetative cells, the nifB2::nifB1::nifB2 hybrid promoter provided a means to determine whether the Nif2 nitrogenase could function in the microoxic environment of the heterocyst. In a nifDK1-deletion strain, we found that expression of the nif2 genes, under the control of the hybrid promoter, was activated by CnfR1 in heterocysts and produced active nitrogenase (Fig. 5). The nif2 cluster that was expressed from the hybrid promoter has nifB2, nifS2, nifU2, nifH2, nifD2, nifK2, nifEN2, nifX2, nifW2, hesA2, and hesB2 as well as several conserved nif2 genes of unknown function. However, the nif2 cluster lacks nifZ2 and nifT2, which are transcribed divergently from nifB2 (19) and lack the hybrid promoter, which is separated from nifZT2 by the integrated plasmid. The nifV, nifZ1, and nifT1 genes are distant from the primary nif1 cluster (19, 51), so their expression would be unaffected by the nifDK1 deletion, and NifV, NifZ1, and NifT1 likely functioned to support Nif2 nitrogenase synthesis in heterocysts, although possibly not as well as the Nif2 proteins.

Since the region between the two divergent nif2 operons (nifZT2 and the large cluster that begins with nifB2) has only one apparent binding site for CnfR proteins (20, 22, 35), CnfR2 controls expression of all the nif2 genes using this intergenic region. When expressed in anaerobic vegetative cells, CnfR1 (and CnfR2) can control the expression of all nif2 genes using the single regulatory region between nifZ2 and nifB2. (Fig. 6).

Transferring the genes to fix nitrogen to nonnitrogen-fixing strains is an important first step in the long-term goal of creating nitrogen-fixing plants. However, many genes, some unknown, are required for the synthesis of a functional nitrogenase. The compact nif2 cluster of A. variabilis has all the nif genes and cnfR2. We transferred the nif2 cluster to Nostoc sp. PCC 7120 where Nif2 nitrogenase was made, allowing the strain to fix nitrogen in anaerobic vegetative cells (Fig. 7). Similarly, a large chromosomal region with 25 nif and nif-related genes from the cyanobacterium L. boryana was integrated into the genome of the unicellular cyanobacterium Synechocystis 6803, which has no nitrogenase genes (52). We saw about 5% of the normal A. variabilis nitrogenase activity in vegetative cells for the Nostoc sp. PCC 7120 strain with the nif2 genes, which was much more than the 0.26% nitrogenase reported for Synechocystis 6803 compared to L. boryana, the source of the nif genes (52). It is apparent that even among cyanobacteria, the ability to express cyanobacterial nif genes and make a functional nitrogenase depends on unknown factors that are important for optimal function.

MATERIALS AND METHODS

Construction of hybrid cnfR1/cnfR2 strains.

Hybrid cnfR1/cnfR2 strains were constructed by integration of plasmids into the chromosome of BP894 (cnfR1, cnfR2 double deletion mutant). The details of the construction of all plasmids are included in Table S2 in the supplemental material, and primers used in the construction are listed in Table S3. Plasmids pBP870-pBP873, pBP907, pBP910, pBP920, pBP921, pBP950, pBP951, and pBP952 were conjugated into BP894 (cnfR1, cnfR2 double deletion mutant), as described previously (53), where they integrated by single recombination into the cnfR1 or cnfR2 promoter regions to create strains BP870, BP871, BP872, BP873, BP907, BP910, BP920, BP921, BP950, and BP951. The integration of the plasmids was verified to be a promoter cross by PCR.

Construction of Co2+-inducible strains.

Co2+-inducible lacZ, cnfR1, or cnfR2 genes in plasmids pBP1141, pBP1142, and pBP1143 were constructed by fusion PCR, as described in Table S2 in the supplemental material. Plasmid pBP1141 was conjugated into FD, and plasmids pBP1142 and pBP1143 were conjugated into both BP894 and MM3, an ntcA mutant (34), where they integrated by single recombination into the frtA region to create strains BP1141 (P::lacZ in FD), BP1142 (P::cnfR1 in BP894), BP1143 (P::cnfR2 in BP894), BP7142 (P::cnfR1 in MM3), and BP7143 (P::cnfR2 in MM3). The integration of the plasmids in frtA was verified by PCR.

Construction of hybrid nifB promoter strains.

Since cyanobacterial strains BP1142 and BP1143 were already Nmr and had the P::fusions integrated into the frtA region, a different antibiotic reporter and integration region was required to insert the hybrid pnifB2::nifB1::nifB2 promoter into these strains. Plasmid pBP1195 was engineered as described in Table S2 to be Emr and to allow integration into the modA region. The pnifB2::nifB2::nifB2 promoter was inserted into pBP1195, as described in Table S2, to create pBP1197. Plasmid pBP1197 was conjugated into BP1142 and BP1143 where it integrated via single recombination into the modA region to create cyanobacterial strains BP2197 and BP3197, respectively. The integration of the plasmid in the modA gene was verified by PCR. The hybrid pnifB2::nifB1::nifB2 promoter was extended about 870 bp into the nifB2 gene, as described in Table S2, to create plasmid pBP1101. Plasmid pBP1101 was conjugated into JE9 (xisA-nifE1 deletion mutant [54]), where it integrated by single recombination in the nifB2 gene to drive expression of the nifBSUHDKENXW2-hesAB2 operon, thereby creating BP1101. The integration of the hybrid promoter upstream of the structural nif2 genes was verified by PCR.

Construction of hetR and nrrA mutants.

Plasmids pBP1107 and pBP1108 were constructed as described in Table S2 to delete the hetR and nrrA genes, respectively. Plasmids pBP1107 and pBP1108 were conjugated into strain FD where they integrated by single recombination into the hetR or nrrA regions. Single recombinants were selected by Emr, followed by sacB selection on 10% sucrose plates for double recombinants (55, 56) to create cyanobacterial strains BP1107 and BP1108. PCR was used to verify that there were no wild-type copies of the genes.

Construction of Nostoc sp. PCC 7120 with the nif2 cluster.

We created plasmid pBP716 by inserting a 1.9-kb SmaI fragment containing the aadA cassette (SprSmr) from pRL5801 into the SmaI site of pRL2948a. The Escherichia coli strain containing pAAWZ1787 (a fosmid from the Joint Genome Institute [JGI] containing the nifB2 cluster from genes ava_4241–ava_4266) was electroporated with pKM208. pKM208 contains all the genes necessary for recombineering under the Ptac promoter; red and gam genes are repressed by lacI (57). The plasmid also has a temperature-sensitive origin of replication, which allows the plasmid to be cured from the strain by growth at 37°C. A 2.1-kb PCR product generated from pBP716 using primers FosOriTSp-L2 and FosOriTSp-R2 containing the oriT site (required for conjugation of the plasmid to cyanobacteria) and the SprSmr cassette was treated with DpnI to remove plasmid DNA and recombineered into pAAWZ1787 at the Cmr cassette using pKM208 by electroporation. The E. coli strain containing pAAWZ1787 and pKM208 was treated with isopropyl-β-d-thiogalactopyranoside (IPTG) just before electroporation to activate the recombination genes. The recombineered plasmid, pBP890, containing the nif2 gene cluster in addition to oriT and the SprSmr cassette was screened for replacement of the Cmr cassette with the SprSmr cassette. Plasmid pBP890 was electroporated into HB101 pRL528 and then conjugated into BP291 (a Nostoc sp. PCC 7120 strain containing the frtRABC operon that allows the strain to use fructose as a carbon source [58]), as described previously (53), creating strain BP893. The presence of the nif2 region in BP893 was verified by PCR.

Aerobic nitrogen step-down experiments.

Aerobic cultures were grown for about 10 generations in an 8-fold dilution of AA medium (AA/8) (59) with 5 mM NH4Cl and 10 mM N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid (TES), pH 7.2, and then diluted 1:100 in the same medium and grown to an optical density at 720 nm (OD720) of 0.1 to 0.2. Cultures were washed and diluted with AA/8 to an OD720 of 0.1 to achieve nitrogen step-down. At least three 25-ml biological replicates of each culture were grown for 24 h at 30°C with shaking and illumination at 80 to 100 μE m−2 s−1 in 125-ml flasks.

Anaerobic nitrogen step-down experiments.

For anaerobic nitrogen step-down experiments, cells were grown in air in the light with shaking in AA/8 with 5.0 mM fructose, 5.0 mM NH4Cl, and 10 mM TES, pH 7.2, to an OD720 of 0.5 to 0.6. Cells were washed with AA/8 and resuspended in AA/8 with 10 mM fructose and 10 μM dichlorophenyldimethylurea (DCMU; to inhibit oxygen evolution from photosystem II) to an OD720 of 0.6. At least three, 8-ml biological replicates were aliquoted into 16-ml Hungate tubes for each condition tested; 5 mM NH4Cl and 10 mM TES, pH 7.2, (+N condition) were added to the samples, when applicable, and samples were flushed with dinitrogen for 3 min. The cultures were then incubated for 4 to 6 h anaerobically at 30°C with illumination at 80 to 100 μE m−2 s−1, and induction of nitrogenase activity was verified by acetylene reduction assays.

Expression of genes from the coaT promoter.

The cnfR genes under the control of the coaT promoter were grown in Co2+-free BG-11o (60) medium containing 5 mM NH4Cl and 10 mM TES, pH 7.2, 5 μg ml−1 neomycin, and 10 μg l−1 cobalamin (vitamin B12) (because cobalt is required for cobalamin synthesis) and then diluted 1:100 in conditioned medium (medium further depleted of cobalt by prior growth of strain FD in Co2+-free BG-110 medium lacking fixed N) and grown to an OD720 of 0.1 to 0.2. Cells were washed free of fixed nitrogen and diluted with conditioned Co2+-free BG-110 containing 10 μg l−1 cobalamin and lacking fixed nitrogen to an OD720 of 0.1. At least three 25-ml biological replicates of each culture were grown for 20 h at 30°C with shaking and illumination at 80 to 100 μE m−2 s−1 in 125-ml flasks. After confirmation of heterocyst formation at 20 h after nitrogen step-down, 2 μM CoCl2 was added to some cultures for 4 h to induce the cnfR genes driven by the coaT promoter. AA/8 medium could not be used for induction of the coaT promoter because the EDTA in the medium caused chelation of the Co2+.

For anaerobic nitrogen step-down experiments, cells were grown aerobically, as described above, in Co2+-free BG-110 medium containing 5 mM NH4Cl and 10 mM TES, pH 7.2, 5 mM fructose, 5 μg ml−1 neomycin, and 10 μg l−1 cobalamin to an OD720 of 0.5 to 0.6. Cultures were diluted 1:100 in the same medium and grown to an OD720 of 0.3 to 0.4. Cells were washed free of nitrogen with conditioned Co2+-free BG-110 lacking fixed nitrogen and resuspended in conditioned cobalt-free BG-110 (−N) containing 10 μg ml−1 cobalamin, 10 mM fructose, and 10 μM DCMU to an OD720 of 0.6. At least three, 8-ml biological replicates were aliquoted into 16-ml Hungate tubes for each condition tested; 5 mM NH4Cl and 10 mM TES, pH 7.2 (+N condition), and 2 μM CoCl2 (to induce P) were added to samples, when applicable, and samples were flushed with dinitrogen for 3 min. The cultures were then incubated for 4 h anaerobically at 30°C with illumination at 80 to 100 μE m−2 s−1, and induction of nif genes was verified by acetylene reduction assays.

Acetylene reduction assays.

Acetylene reduction assays have been described previously (33). For acetylene reduction of anaerobic cultures, 1 ml of acetylene gas was added to Hungate tubes 30 min before the end of the anaerobic incubation period. For aerobic cultures, 1 ml of acetylene gas was added to 5-ml cultures stoppered in 16-ml Hungate tubes and incubated for 1 h at 30°C, with illumination at 80 to 100 μE m−2 s−1. Samples (250 μl) of headspace gas were removed via a gas-tight needle/syringe and injected into a Shimadzu gas chromatograph equipped with a 6-foot Porapak N column. The column temperature was 75°C.

β-Galactosidase assays.

β-Galactosidase assays were performed in 96-well microtiter plates using 250 μl of sample for at least 3 biological replicates, with quadruple technical replicates for each biological replicate. β-Galactosidase assays were performed as previously described (35). OD420 and OD665 measurements were taken to determine the amount of 2-nitrophenol (OD420) and to correct for chlorophyll and light scattering from permeabilized cells (OD665). Calculations were performed using the following equation, developed empirically: β-galactosidase = 1,000 × ([OD420 – (1.58 × OD665)]/[OD720 × time of assay (in minutes)]). The values from the quadruple technical replicates were averaged for each of the three biological replicates. These three averages for the biological replicates were used to calculate the average, and the standard deviation is shown in the graphs as error bars.

Aerobic cultures of BP1141, BP2197, and BP3197 were grown in Co2+-free BG-11 medium and induced as described above for expression of the coaT promoter for about 10 generations in Co2+-free BG-11 medium containing, 10 μg l−1 cobalamin, and antibiotics and then diluted 1:100 in the same medium and grown to an OD720 of 0.1 to 0.2. For BP1141, various concentrations of CoCl2 (0 to 20 μM) and ZnCl2 (0 to 100 μM) were added to four 2-ml biological replicates of BP1141, which were grown for 2 h at 30°C with shaking and illumination at 80 to 100 μE m−2 s−1 in 12-well plates before the β-galactosidase assays. For BP2197 and BP3197, CoCl2 at 2 μM (to induce P) and 5 mM NH4Cl and 10 mM TES, pH 7.2 (+N condition), when applicable, were added to samples. Cultures were grown for 4 h at 30°C with shaking and illumination at 80 to 100 μE m−2 s−1 before the β-galactosidase assays.

For anaerobic nitrogen step-down experiments, cells of BP2197 and BP3197 were grown in Co2+-free BG-11 medium and induced as described above for expression of the coaT promoter. At least three, 8-ml biological replicates were aliquoted into 16-ml Hungate tubes for each condition tested; 2 μM CoCl2 (to induce P) and 5 mM NH4Cl and 10 mM TES, pH 7.2 (+N condition), when applicable, were added to samples, and samples were flushed with dinitrogen for 3 min and incubated for 4 h at 30°C with illumination at 80 to 100 μE m−2 s−1 before the β-galactosidase assays.

RNA isolation and reverse transcriptase-quantitative PCR.

RNA isolation and reverse transcriptase-quantitative PCR (RT-qPCR) were performed as described previously (28, 33, 58). RNA was isolated using TRI Reagent (Sigma) and subjected to DNase digestion (Turbo DNA-free kit; Ambion). A total of 40 ng of RNA was converted to cDNA in a 10-μl reaction mixture using iScript reverse transcription supermix for RT-qPCR (Bio-Rad). The cDNA was diluted 1:20 to 0.2 ng μl−1. The qPCRs used 0.8 ng of cDNA in a 10-μl reaction mixture with 5 pmol of gene-specific primers and 1× SsoAdvanced SYBR green supermix (Bio-Rad). All genes were compared to the housekeeping gene rnpB to determine Δ values.

Statistical analysis.

Data are shown as the mean ± standard deviation. The significance of the differences between the means for two values was analyzed using an unpaired, two-tailed Student’s t test, with P values of < 0.05 considered statistically significant.

Data availability.

All processed data (e.g., relative expression and β-galactosidase values) used in the experiments are provided here. Additional raw data may be requested from the authors.