Isolation of Mutants Sensitive to Diphenyl-Ether Herbicides.

To identify the gene encoding Protox in the cyanobacterium Synechocystis sp. PCC6803, we screened a mutant pool in which the cyanobacterial genes were randomly disrupted by insertion of the kanamycin resistance gene (15). Because the oxidation of Protogen IX is an obligatory step in tetrapyrrole synthesis, the disruption of the gene encoding Protox was predicted to be lethal, unless redundant genes encode Protox in Synechocystis sp. PCC6803. Therefore, before screening, we first introduced the hemY gene of Arabidopsis thaliana (D83139) into the WT strain of Synechocystis sp. PCC6803. Consequently, this transgenic cyanobacterium (designated as the AT strain) had both Arabidopsis hemY-type Protox and the endogenous cyanobacterial Protox. We then introduced mutations by transposon-mediated random insertion of the kanamycin resistance cassette into the chromosome of the AT strain. We predicted that disruption of the cyanobacterial Protox gene would not cause severe growth retardation of the cell, because the gene disruption should be complemented by the already-introduced Arabidopsis hemY gene. Therefore, we hypothesized that the disruption of the endogenous Protox gene would render the cells sensitive to diphenyl-ether herbicides, because the cells were predicted to become dependent on the Arabidopsis Protox that is sensitive to these herbicides. To the best of our knowledge, the use of a specific inhibitor with genetic precomplementation is a unique screening strategy for essential enzymes. This strategy may therefore be applicable to the identification of a variety of genes encoding essential enzymes, as long as another enzyme with altered sensitivity to the selection marker is available.

We screened 6,686 transformants and successfully isolated nine mutant lines that were sensitive to the diphenyl-ether herbicide, acifluorfen. We determined the insertion sites of the kanamycin resistance cassette in these lines by inverse PCR. We identified seven lines containing the kanamycin resistance cassette at 21 bp upstream of the start of the slr1790 gene ORF (BAA16808.1), suggesting that these lines were clonal. Among all of the lines isolated, these exhibited the strongest herbicide sensitivity. The other lines, which showed sensitivity to the herbicide, contained the kanamycin resistant cassette in the slr0844 locus, encoding NADH dehydrogenase I chain L (BAA10530.1). These lines were characterized by slower growth in the presence of acifluorfen, but unlike the lines with a mutation upstream of the slr1790 gene, they survived in the presence of acifluorfen.

Targeted Disruption of the slr1790 Gene.

To investigate the function of the protein encoded by the slr1790 gene, we disrupted the slr1790 gene in the WT and AT strains by targeted insertion of the kanamycin resistance cassette. We designated these disruptants as WTSK and ATSK, respectively (Fig. S1A). Because Synechocystis is polyploid, we assessed the segregation of the mutant genome copy in the mutant strains by PCR after several rounds of selection on kanamycin-containing agar plates. Genomic DNA was extracted from cells and amplified by PCR, using primer pairs that hybridized to the upstream and downstream regions of the slr1790 gene. Amplification of genomic DNA extracted from the WT and the AT strains generated a PCR fragment of 1.9 kbp, which corresponds to the size of the intact slr1790 gene (Fig. S1B). In contrast, PCR using genomic DNA isolated from the ATSK strain yielded a 3.2-kbp PCR fragment, which corresponds to the size of the WT fragment (1.9 kb) plus the inserted kanamycin resistance cassette (1.3 kb). Our results demonstrate that the disrupted slr1790 gene segregated in the ATSK strain. On the other hand, PCR using genomic DNA isolated from the WTSK strain resulted in two fragments of 1.9 kbp and 3.2 kbp, suggesting that segregation of the tagged slr1790 gene was incomplete. It is likely that the slr1790 gene is essential for the growth of the cells and that disruption of this gene is lethal.

Effect of Herbicides on Cell Growth.

Supporting our hypothesis that the slr1790 gene encodes an essential protein, the WTSK strain with heteroplasmic DNA (Fig. S1B) showed retarded growth compared with the WT strain (Fig. 1A). In contrast, the ATSK strain, in which the slr1790 gene appeared to be completely knocked-out, survives like the WT and AT strains (Fig. 1A). These results indicate that the Arabidopsis hemY gene can compensate the loss of the slr1790 gene and can restore the growth of the mutant to the level comparable to that of WT (Fig. 1A).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Growth of the slr1790 insertional mutants in the presence and absence of the herbicide. The WT, AT, WTSK, and ATSK lines were grown in liquid media in the absence (A) and presence (B) of 100 μM acifluorfen under continuous illumination (20 μmol photons m⁻² s⁻¹).

At the present stage, a possibility cannot be excluded that the slr1790 gene has an unidentified role other than being involved in Protox activity and that the Arabidopsis hemY gene coincidently compensated this unidentified function. To clarify this point, we fed acifluorfen to the mutant strains to inhibit Protox activity of the hemY gene product. In the presence of 100 μM acifluorfen, the growth of the ATSK strain was significantly retarded compared with the WT, AT, and the WTSK strains (Fig. 1B). These results demonstrate that Protox activity of the hemY gene product is necessary to compensate the loss of the slr1790 gene.

To further substantiate that acifluorfen specifically inhibited Protox activity in the ATSK strain, we tested the effects of two other classes of Protox inhibitors, pyraflufen-ethyl and flumioxazin (Table 1). Although the AT strain was insensitive to the diphenylether herbicides when used in a concentration of up to 10⁻⁴ M, growth of the ATSK strain was severely affected with I₅₀ values of 10^−4.5 to 10^−6.7 M (Table 1). These results are consistent with our hypothesis that the ATSK strain is dependent on the HemY-encoded Protox because of the lack of the cyanobacterial-type Protox. However, we could not rule out a possibility that the slr1790 gene is involved in scavenging reactive oxygen species (ROS) and that the disruption of the slr1790 locus conferred sensitivity to ROS, which might be generated by a small amount of Proto IX or Protogen IX accumulation. Therefore, we tested the effect on cell growth in the presence of diquat, a generator of ROS in cells. In contrast to diphenylether herbicides, there was no difference in the growth of the AT strain and the ATSK strain on diquat-containing cultures. Therefore, our data support the hypothesis that the disruption of the slr1790 gene abolished the endogenous Protox activity, resulting in growth retardation of the WTSK strain and the ATSK strain in the presence of the diphenylether herbicide acifluorfen.

Analysis of Pigments by HPLC.

We examined the levels of chlorophyll a and intermediates of chlorophyll biosynthesis in WT, AT, and the insertional mutants by HPLC (Table 2) in the presence and absence of acifluorfen. Although the WTSK strain was heteroplasmic, this strain accumulated a substantial amount of Proto IX, and the WT strain failed to accumulate detectable levels of Proto IX. It should be noted that we could not detect Protogen IX in the cell extracts of the WTSK strain. One might expect that inhibition of Protox activity may lead to accumulation of Protogen IX, the substrate of the Protox enzyme. However, it has been reported that inhibition of Protox usually results in the detection of only Proto IX, as Protogen IX is readily oxidized by oxygen or by other nonspecific cellular activities within the cells, or during the procedure of pigment analysis (13, 16–18). Thus, we assumed that the level of Proto IX detected by our HPLC system most likely represented that of Protogen IX plus Proto IX within the cells.

In the WTSK strain, the levels of chlorophyll a substantially reduced to nearly one-half of the WT strain both in the absence and presence of acifluorfen (Table 2). These data are in accordance with the increase in the Proto IX level in this strain. Similarly, the ATSK strain accumulated a higher level of Proto IX (890 pmol/10⁸ cells) than any other strains tested in the presence of acifluorfen, which indicates that metabolism of Proto(gen) IX is severely impaired in this strain. It was rather unexpected that the same strain accumulated a moderate amount of Proto(gen) IX (76 pmol/10⁸ cells) in the absence of acifluorfen. We are speculating that tetrapyrrole synthesis is partially stalled because of the presence of HemY in the ATSK strain. Slr1790 may be in tight contact with the subsequent enzymes such as ferrochelatase or Mg-chelatase, but HemY may not have such a contact with subsequent enzymes. As a result, efficient channeling of Proto IX might have been compromised in the ATSK strain. The level of chlorophyll a in the ATSK strain was close to that of WT. In the AT strain, we observed a substantial accumulation of Proto IX (197 pmol/10⁸ cells) and reduction in the chlorophyll contents in the presence of acifluorfen. This observation can be explained by assuming that the inhibited HemY is competing with endogenous HemJ for access to docking sites on Mg-chelatase and ferrochelatase. Collectively, we conclude that the slr1790 disruptants are defective in their native Protox activity.

Phylogenetic Analysis of the Slr1790 Sequences.

The slr1790 gene encodes a protein of 193 amino acid residues with five putative transmembrane helices. The amino acid sequence shares homology to the NuoM (ND4) subunit of the NADH dehydrogenase complex I from various organisms, including E. coli (CAA48372), human (ABX40271), and Synechocystis sp. PCC6803 (K05575) (Fig. S2). This subunit is proposed to be involved in proton translocation across membranes (19). However, the Slr1790 sequences are about half of the NuoM sequences in length, and a few conserved regions of the NuoM sequences are missing in the Slr1790 sequences, including the essential residues for the function of the NuoM protein (19)(Fig. S2). Therefore, it is unlikely that the protein encoded by slr1790 possesses the same function as NuoM.

The slr1790 homologs are highly conserved in most cyanobacteria (Fig. S3). Among 30 cyanobacterial genomes that have been sequenced at the time of writing this article, 23 cyanobacterial species contain slr1790 homologs, but no hemG or hemY homologs are present (Table S1). In contrast, two Prochlorococcus species (Prochlorococcus marinus str. MIT 9215 and Prochlorococcus marinus str. MIT 9515) only contain hemG homologs, and four other cyanobacteria [Synechococcus sp. JA-2–3B'a(2-13), Synechococcus sp. JA-3–3Ab, Thermosynechococcus elongatus BP-1, and Trichodesmium erythraeum IMS101] only have hemY homologs. Gloeobacter violaceus PCC 7421 is an exception, because it contains both slr1790 and hemY homologs. It is noteworthy that except for this organism, distribution of slr1790, hemG, and hemY homologs in cyanobacteria were exclusive to each other (Table S1). We examined whether the distribution of any other genes in cyanobacteria was mutually exclusive to that of hemG and hemY using the genome comparison tool CCCT (20). We found that slr1790 is the only gene that shows such an exclusive distribution pattern. Thus, these genome analyses also support our genetic data, suggesting that hemG, hemY, and slr1790 encode proteins that share the same function.

The slr1790 homologs were found in a variety of bacterial species, including Bacteroidetes, α-, β-, ε-, and δ-proteobacteria (Fig. S4). In proteobacteria, except for γ-proteobacteria, slr1790 homologs seemed to dominate over the other two genes. In contrast, many Chloroflexi species contain hemY homologs. The hemG homologs were mainly found in γ-proteobacteria, in which distributions of slr1790 and hemG homologs were mutually exclusive. Likewise, distribution of slr1790 and hemY homologs in the Bacteroidetes/Chlorobi was mutually exclusive (Table S2). Within this group of bacteria, those which have more than one hemY, hemG, and slr1790 homolog seem to be an exception. For example, only Chlorobium phaeobacteroides DSM 266 contains both hemG and hemY homologs among the Bacteroidetes/Chlorobi group (Table S2). Of 17 sequenced Bacteroidetes/Chlorobi genomes, 7 are lacking slr1790, hemG, or hemY homologs (Table S2), indicating that there is at least one yet unidentified gene encoding Protox in these species.

Protox Activity in Recombinant Rhodobacter Protein Similar to the Cyanobacterial Slr1790 Proteins.

To examine whether Slr1790 protein possess Protox activity in vitro, we attempted to obtain recombinant Slr1790 protein by expressing the Synechocystis slr1790 gene in E. coli or using the Baculovirus expression system in insect cells. However, the protein did not accumulate to a detectable level in these systems by unidentified reasons. Therefore, we attempted to express an slr1790 homolog from another organism to examine if it encodes a protein with Protox activity. Because Gomelsky and Kaplan had reported that they successfully overexpressed a Rhodobacter sphaeroides homolog of slr1790 (hereafter referred to as the Rs-slr1790 gene) in E. coli (21), we began with this gene to examine Protox activity. The E. coli transformant harboring the Rhodobacter homolog successfully produced a chimeric recombinant protein fused with the maltose-binding protein (MBP) tag (hereafter referred to as Rs-Slr1790 protein). It should be noted that Rs-Slr1790 was solubilized in the presence of 0.04% (wt/vol) Tween 20 when the protein was fused with the MBP tag. Similar observations were reported with other putative membrane-spanning proteins (e.g., refs. 22, 23). When we added this protein into a reaction mixture containing Protogen IX, the pigment was converted to Proto IX and emitted strong fluorescence at 633 nm (Fig. 2). In contrast, addition of the MBP tag alone to the reaction mixture did not significantly increase the rate of the Protogen IX-to-Proto IX conversion when it was compared with the auto-oxidation rate of Protogen IX (Fig. 2). These results clearly demonstrate that Rs-Slr1790 possess Protox activity.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Protox activity of Rs-slr1790 recombinant protein. Protox assays are shown for purified recombinant Rs-slr1790 protein fused with an MBP tag (open squares), purified MBP fused with the α-subunit of β-galactosidase (crosses), and a buffer control without protein (filled squares). Each reaction mixture was incubated at 25 °C as described in Materials and Methods. The production of Proto IX was monitored by fluorescence with a fluorometric plate reader equipped with 405 nm (excitation) and 633 nm (emission) bandpass filters. Error bars indicate SD of duplicate measurements.