Dissecting the Role of the N-Terminal Region of the Escherichia coli Global Transcription Factor FNR

ABSTRACT The role of the N-terminal region of the transcription factor FNR, which immediately precedes the first ligand (Cys20) of the [4Fe-4S] cluster, was investigated. We found that truncation mutants that removed residues 2 to 16 and 2 to 17 had wild-type levels of FNR protein but surprisingly altered O2 regulation.

trypsin digestion, suggesting that this region is either unstructured or surface exposed (17).
To dissect the role of the first 19 amino acid residues of FNR, we constructed a series of N-terminal truncation mutants (Fig. 1B) by site-directed mutagenesis of pPK823 (13) and studied the properties of these mutant proteins by measuring the ß-galactosidase activity produced from strains containing lacZ transcriptional fusions to FNR-dependent promoters. The amount of ß-galactosidase activity produced from anaerobically grown strains containing either dmsAЈ-lacZ (data not shown) or narGЈ-lacZ (Fig. 2) showed that removal of residues 2 and 3 from FNR (FNR⌬2-3) had no effect on transcription activation of either the narG or the dmsA promoter, whereas removal of residues 2 to 8 (FNR⌬2-8) reduced FNR activity ϳ8-fold. However, elimination of just seven more residues (FNR⌬2-15) restored FNR activity to levels 1.5-fold greater than those for full-length FNR. This unexpected observation led us to further characterize the role of residues 8 through 19.
Truncation of residues 2 to 10, 2 to 12, or 2 to 13 of FNR caused a complete loss of activation of the narG promoter (Fig.  2). However, elimination of just one additional amino acid, Gly14, partially restored transcription activity such that the activity was now 50% relative to that of the full-length protein.
Unexpectedly, FNR protein levels in strains containing truncation of residues 2 to 10, 2 to 12, or 2 to 13 were found to be greatly reduced (Ͼ20-fold) by Western blot analysis (Fig. 2), and FNR transcription activity was largely correlated with the amount of FNR protein (Fig. 2). It seems unlikely that these truncation mutants have accelerated the ClpXP-dependent proteolysis of FNR, since neither FNR⌬2-12 nor FNR⌬2-13 contains the N-terminal ClpXP binding site (residues 5 to 11), and the second ClpXP binding site (residues 249 and 250) is not sufficient to target FNR for ClpXP-dependent proteolysis (16). It also seems unlikely that decreased transcription can explain the reduction in protein levels, since transcription of these mutant genes is driven by plasmid sequences and should be unaffected by the truncation mutations. Thus, we suggest that perhaps another proteolytic site becomes unmasked in these mutants, resulting in their proteolysis.
In contrast, truncation of amino acids 2 to 16 and 2 to 17 ( Fig. 2) showed slightly increased levels of ß-galactosidase activity expressed from the narG promoter-lacZ fusion, similar to that found with the truncation of residues 2 to 15. In addition, the FNR protein levels in these truncation mutants were the same as those in full-length FNR, indicating that the element that leads to the posttranscriptional reduction in FNR protein levels was removed.
The additional removal of Ile18 (FNR⌬2-18) decreased FNR activity ϳ3-fold (Fig. 2), with a corresponding decrease in FNR protein levels, similar to what was found for FNR⌬2-14, whereas the additional elimination of His19 (Fig. 2), adjacent to the first cluster ligand, Cys20, abolished FNR activity, although protein levels were reduced only ϳ 3-fold. The complete loss in FNR activity of FNR⌬2-19 is similar to what has been observed for FNR mutants containing substitutions for the Cys ligands. This led us to consider that Cys20 was removed by methionine aminopeptidase because of its preference for cleaving methionine as well as the following amino acid residue when it is Ala, Cys, or Ser (4). In support of the idea that His19 has a less critical function in FNR than Cys20, we found that replacement of His19 in FNR⌬2-18 with Val does not change the activity of this mutant significantly (Fig. 2). However, the basic side chain of His19 is likely to play a small role since replacement with Tyr but not the similarly charged amino acid Arg decreases FNR activity about twofold (15). Therefore, there seems to be only a slight preference for a basic residue immediately preceding Cys20.
Since previous studies of FNR identified amino acid substitutions in the N-terminal region that allowed FNR to be active in the presence of O 2 (FNR* mutants) (1,11,15), we investigated whether any of the truncations altered the response of FNR to O 2 by assaying FNR activity under aerobic growth conditions. As expected, the previously characterized FNR* mutants FNR(D154A) and FNR(L28H) showed increased expression of the narG (Fig. 3A) promoter and another FNRdependent promoter, P ydfZ (6) (Fig. 3B), under aerobic growth conditions. The truncation mutant of FNR residues 2 to 15 also increased the activity of FNR under aerobic conditions. To test whether this mutant bypassed the need for the Fe-S cluster analogous to FNR(D154A) or altered cluster stability comparably to FNR(L28H) (2), we combined the truncation mutant   2. Effect of N-terminal truncation mutants on FNR activity and protein levels. ß-Galactosidase activity (reported in Miller units) was determined as previously described (13) from derivatives of strain RZ8480 (narGЈ-lacZ), transformed with vector plasmid (pET11a) or plasmid derivatives encoding N-terminal truncation mutants, which were grown under anaerobic conditions in M9 minimal medium with 0.2% (wt/vol) glucose, 10 M ferric ammonium citrate, 1.4 mM KNO 3 , and 0.2 M ammonium molybdate. The values shown for ␤-galactosidase activity (gray bars) are means of results from three independent experiments, expressed as percentages relative to the activity for full-length FNR (the average activity of wild-type FNR [expressed from pET11a] was 250 Miller units, and the average error from triplicate samples was less than 10%). Cells were also subjected to Western blot analysis (16) to measure the amount of full-length FNR and various FNR truncation mutants after growth under the same conditions as those for the ß-galactosidase assays. Proteins were detected with UV, and the amounts of FNR were quantified using Molecular Dynamics ImageQuant software and expressed as percentages relative to the level for full-length FNR (black bars). Values shown are means of results from duplicate experiments. VOL. 190, 2008 NOTES 8231 with either C122A, which abolishes the binding of the [4Fe-4S] 2ϩ cluster of the wild-type protein (8), or D154A, which alters dimerization (13). The change to Ala122 completely eliminated the activity of the truncation mutant FNR⌬2-15 under aerobic conditions, suggesting that the truncation does not bypass the need for the Fe-S cluster. Rather, these data suggest that FNR⌬2-15 has altered cluster stability, similar to FNR(L28H) (2). Consistent with this finding, the change to Ala at residue 154 had no additional effect on the truncation mutant, indicating a mechanism independent of increasing dimerization of apo-FNR (11,13). Further analysis of other truncation mutants between FNR⌬2-15 and FNR⌬2-19 indicated that only truncations from residues 2 and 15 to 2 and 17 showed increased activity in the presence of O 2 (Fig. 4). The FNR⌬2-15 protein was isolated under anaerobic conditions, and analysis of the visible absorbance spectrum confirmed that it had the same type of Fe-S cluster as the fulllength protein (data not shown). However, the cluster appeared to be more stable to O 2 than the wild-type protein, since the Fe-S absorbance spectrum showed little change after 5 min (data not shown). The increased stability of the cluster is reminiscent of the FNR(L28H) mutant (2). To determine if removal of the N terminus caused any significant conformational changes in the rest of the protein, the isolated protein was treated with trypsin under anaerobic conditions and analyzed by mass spectrometry. Only one major cleavage product was observed with FNR⌬2-15, which corresponds to the cleavage product at Arg247 observed with the full-length protein ( Table 1). The other two residues that lead to trypsin digestion products in the full-length protein, Arg9 and Arg10, are not present in the truncation mutant, and accordingly, these sites of cleavage were not detected. Thus, we conclude that the overall structure of FNR⌬2-15 is not significantly altered. In addition, this mutant protein may prove useful for crystallography experiments since removal of this protease-accessible region may improve crystallization of FNR.
The MS data also showed that isolated FNR⌬2-15 contains only residues 17 to 250 of FNR, indicating that the N-terminal Met and Cys16 residues have been removed (Table 1), most likely as a result of cleavage from methionine aminopeptidase because of its preference for cleaving Cys following the Nterminal Met residue (4). A parallel set of truncation mutants that initiated with Met-Val was constructed to evaluate any possible effects of the first set of mutants resulting from cleavage of the second amino acid. Overall, the insertion of Val did not alter the activity or protein levels of the truncation mutants except for FNR⌬2-19 (described above) and FNR⌬2-15. Insertion of a Val residue preceding Cys16 [FNR⌬2-14(G15V)] resulted in a mutant protein in which both the protein level and the ß-galactosidase activity produced from the narGЈ-lacZ fusion under anaerobic conditions are similar to those for fulllength protein (Fig. 2), suggesting that removal of Cys16 in FNR⌬2-15 (shown by mass spectrometry; see above), FNR⌬2-16, and FNR⌬2-17 may partially explain their increased activity. In support of this notion, replacement of Cys16 with Val in FNR⌬2-15 has activity similar to that of FNR⌬2-15 (which lacks Cys16 because of processing) under anaerobic conditions (data not shown). This region also seems important in protein stability, since replacement of Gly15 with Val in FNR⌬2-14 results in greater levels of FNR protein (Fig. 2) than in FNR⌬2-14.
It is also interesting to note that the N-terminal region is conserved in closely related FNR orthologs (Fig. 1A, rows 1 to 6), suggesting a similar function. However, this degree of conservation is not observed in orthologs from more-distantly related species (Fig. 1A, rows 7 to 10). In particular, the Pseudo-