IscR Regulation of Capsular Polysaccharide Biosynthesis and Iron-Acquisition Systems in Klebsiella pneumoniae CG43

IscR, an Fe–S cluster-containing transcriptional factor, regulates genes involved in various cellular processes. In response to environmental stimuli such as oxidative stress and iron levels, IscR switches between its holo and apo forms to regulate various targets. IscR binding sequences are classified into two types: the type 1 IscR box that is specific for holo-IscR binding, and the type 2 IscR box that binds holo- and apo-IscR. Studying Klebsiella pneumoniae CG43S3, we have previously shown that iron availability regulates capsular polysaccharide (CPS) biosynthesis and iron-acquisition systems. The present study investigated whether IscR is involved in this regulation. Compared with that in CG43S3, the amount of CPS was decreased in AP001 (ΔiscR) or AP002 (iscR 3CA), a CG43S3-derived strain expressing mutated IscR mimicked apo-IscR, suggesting that only holo-IscR activates CPS biosynthesis. Furthermore, a promoter-reporter assay verified that the transcription of cps genes was reduced in AP001 and AP002. Purified IscR::His6, but not IscR3CA::His6, was also found to bind the predicted type 1 IscR box specifically in the cps promoter. Furthermore, reduced siderophore production was observed in AP004 (Δfur-ΔiscR) but not in AP005 (Δfur-iscR 3CA), implying that apo-IscR activates iron acquisition. Compared with those in AP004, mRNA levels of three putative iron acquisition systems (fhu, iuc, and sit) were increased in AP005, and both purified IscR::His6 and IscR3CA::His6 bound the predicted type 2 IscR box in the fhuA, iucA, and sitA promoters, whereas IscR3CA::His6 displayed a lower affinity. Finally, we analyzed the effect of external iron levels on iscR expression. The transcription of iscR was increased under iron-depleted conditions as well as in AP001 and AP002, suggesting an auto-repression exerted by apo-IscR. Our results show that in K. pneumoniae, IscR plays a dual role in the regulation of CPS biosynthesis and iron-acquisition systems in response to environmental iron availability.


Introduction
Klebsiella pneumoniae is a rod-shaped, gram-negative bacterium that causes community-acquired diseases including pneumonia, bacteremia, septicemia, and urinary and respiratory tract infections that occurr particularly in immune-compromised patients [1]. In Asian countries, especially in Taiwan and Korea, K. pneumoniae is the predominant pathogen responsible for pyogenic liver abscesses in diabetic patients [2,3,4]. Among the virulence factors identified in K. pneumoniae, capsular polysaccharide (CPS) is considered the major determinant for K. pneumoniae infections. Pyogenic liver abscess isolates often carry heavy CPS loads that protect bacteria from phagocytosis and from being killed by serum factors [5,6]. Apart from its anti-phagocytic function, Klebsiella CPS also promotes bacterial colonization and biofilm formation at infection sites [7,8,9].
Our previous studies have demonstrated that CPS biosynthesis in K. pneumoniae is repressed in iron-replete conditions, and this regulation is controlled by an iron uptake regulator (Fur) [10]. Under iron-replete conditions, dimeric Fur in complex with Fe(II) indirectly activates K. pneumoniae CPS biosynthesis through transcriptional factors RmpA and RcsA and a small non-coding RNA, RyhB [10,11]. The transcription of cps genes is directly regulated by RmpA and RcsA but appears to be indirectly regulated by RyhB. These findings indicate that environmental iron availability influences K. pneumoniae CPS biosynthesis through multiple regulators.
To maintain iron homeostasis, Fur acts as a master regulator to control iron transport, storage, and metabolism in many gramnegative bacteria including K. pneumoniae [11][12][13]. We have previously reported that Fur directly represses at least six of the eight iron acquisition systems in K. pneumoniae CG43S3 [10]. In addition to Fur, the transcriptional regulator IscR plays a crucial role in iron metabolism. IscR regulates the biosynthesis of Fe-S clusters, which are key cofactors of proteins intervening in various cellular processes in bacteria [12,13]. Fe-S clusters can be generally classified into two types, rhombic [2Fe-2S] and cubic [4Fe-4S], which have either ferrous (Fe 2+ ) or ferric (Fe 3+ ) iron and sulphide (S 22 ) [14,15]. IscR is itself a [2Fe-2S] cluster-containing protein encoded by the first gene of the iscRSUA operon. The switch between the [2Fe-2S] holo and apo forms of IscR is believed to be influenced by environmental conditions such as oxidative and nitric oxide stress and cellular iron levels [13,16,17,18]. Moreover, holo-and apo-IscR have been shown to regulate different target genes, suggesting that the presence of the [2Fe-2S] cluster affects the regulatory specificity of IscR [18,19,20,21]. Transcriptomic analysis has identified 40 genes in 20 predicted operons, which are regulated by IscR under aerobic and anaerobic conditions in Escherichia coli [19]. This analysis has also revealed two classes of IscR binding sites (IscR boxes). Type 1 IscR box consists of a 25-bp sequence interacted with holo-IscR, whereas type 2 IscR box consists of a 26-bp sequence interacted with apo-IscR [19]. Furthermore, a detailed analysis of the type 2 IscR box has verified an IscR binding motif for both holo and apo-IscR binding [21].
In this study, we investigated whether IscR participates in the regulation of CPS biosynthesis and the expression of iron acquisition systems in K. pneumoniae. We also analysed the expression of iscR in response to various iron levels.

IscR activates K. pneumoniae CPS biosynthesis in an Fe-S cluster-dependent manner
To study whether IscR regulates K. pneumoniae CPS biosynthesis, we determined the amounts of K2 CPS in CG43S3 (wild type [WT]) and AP001 (DiscR) strains. Compared with the WT, AP001 produced significantly lower amounts of CPS (Fig. 1A), suggesting that IscR activates the biosynthesis of CPS. In K. pneumoniae, IscR contains three highly conserved cysteine residues (C92, C98, and C104 in E. coli IscR) which are thought to coordinate the [2Fe-2S] cluster [20].
To investigate the role of the [2Fe-2S] cluster in IscR regulation of CPS biosynthesis, we created an iscR mutant AP002 (iscR 3CA ) by replacing the three cysteines with alanines and tested whether this mutant, which is predicted to encode an IscR lacking an Fe-S cluster, affected CPS biosynthesis. As shown in Fig. 1A, we found that the amount of CPS decreased in the AP002 strain compared with that in the WT, indicating that the regulation of IscR required the [2Fe-2S] cluster. Moreover, iscR and iscR 3CA were respectively cloned into pACYC184, to yield pIscR and pIscR 3CA , for complementation analysis. Compared with AP001 [pA-CYC184], AP001 [pIscR] produced a significantly higher amount of CPS, whereas the introduction of pIscR 3CA into the AP001 strain did not change the CPS amount (Fig. 1B). These results confirmed that IscR has a positive role in the regulation of CPS biosynthesis and that the presence of the [2Fe-2S] cluster of IscR is essential for this regulation. On the other hand, the CPS amount appeared to obviously increased in AP001 [pIscR], compared with that in WT [pACYC184] (Fig. 1B), which may result from multicopy plasmids are used for complementation. Therefore, we also used single copy constructs to complement the iscR-deletion (Methods S1), and the result showed that the expression of iscR, but not iscR 3CA , could restore the CPS biosynthesis (Fig. S1).

IscR directly binds the promoter of galF
For further investigation of the mechanism of IscR regulation on cps transcription, the sequence of the putative IscR binding site was manually analysed in the three promoter regions of the K2 cps gene cluster. As shown in Fig. 2A, we found a putative type 1 IscR box with 52% (13/25 bp) homology to the consensus sequence located between 2173 bp and 2197 bp relative to the translational start codon of galF (orf1 in the K2 cps gene cluster). In addition, the putative IscR binding sequence in P galF was highly homologous to the IscR-binding motif (59-AxxxCCxxAxxxxxxx-TAxGGxxxT-39) reported by Nesbit et al. [21]. However, no typical IscR binding site was found in the upstream sequence of wzi (orf3 in the K2 cps gene cluster) or manC (orf16 in the K2 cps gene cluster), suggesting that IscR indirectly regulates the promoter activities of orf3-15 and orf16-17, which remains to be studied. On the other hand, we hypothesised that IscR binds directly to the promoter region of galF to activate gene transcription, and we confirmed this by performing an electrophoretic mobility shift assay (EMSA). As shown in the upper panel of Fig. 2B, purified recombinant IscR::His 6 protein was able to bind PgalF-1 but not PgalF-2, in which the region containing the putative IscR binding site was deleted. In addition, compared with that of IscR::His 6 , the recombinant [2Fe-2S] clusterless IscR 3CA ::-His 6 had reduced PgalF-1 binding activity. Furthermore, no obvious interaction between the recombinant IscR proteins and PgalF-1*, the galF promoter lacking only the 25-bp predicted IscR box, was found (the lower panel of Fig. 2B). Besides, PgalF-1 and PgalF-1* DNA showed a slightly different mobility in the gel. These results suggested a direct interaction between IscR and the galF promoter and that the [2Fe-2S] cluster of IscR plays a crucial role in this interaction. On the contrary, we also analysed whether recombinant IscR::His 6 could bind the promoter regions of wzi and manC. As expected, EMSA showed no obvious DNA-protein complex (data not shown).

Effect of IscR on normal human serum resistance
Because CPS acts as a protectant for K. pneumoniae against serum factors, we hypothesize that through modulation of CPS levels, IscR affects the ability of K. pneumoniae to resist the bactericidal effects of serum. To test this hypothesis, we treated K. pneumoniae strains with 75% normal human serum and determined their survival rates. Compared with the WT, the AP001 and AP002 strains had a slightly reduced survival rate (Fig. 3A), implying a positive role for [2Fe-2S]-IscR in the serum resistance of K. pneumoniae. To confirm this result further, we performed a complementation study. As shown in Fig. 3B, the introduction of pIscR, but not pACYC184 or pIscR 3CA , into the AP001 strain increased the bacterial survival rate, to a similar level compared with that of WT [pACYC184], after serum treatment. These results supported the hypothesis that [2Fe-2S]-IscR activates the expression of CPS to increase K. pneumoniae resistance to normal human serum.

IscR has a regulatory role in iron acquisition systems
In E. coli, both Fur and IscR play important roles in the maintenance of cellular iron homeostasis [12,13]. To analyse whether IscR regulates iron acquisition in K. pneumoniae, we performed a chrome azurol S (CAS) assay to assess siderophore CPS levels in WT carrying pACYC184 and AP001 carrying pACYC184, pIscR, or pIscR 3CA were determined in LB. Bacterial glucuronic acid content was determined after 16 h of growth. (C) b-Galactosidase activities of K. pneumoniae AP006 and isogenic strains (AP007 and AP008) carrying the reporter plasmid pOrf12 (P orf1-2 ::lacZ), pOrf315 (P orf3-15 ::lacZ), or pOrf1617 (P orf16-17 ::lacZ) were determined using log-phase cultures grown in LB medium. Error bars indicate standard deviations. *P,0.01 compared with the indicated groups. doi:10.1371/journal.pone.0107812.g001  Fig. 4A, no apparent siderophore secretion was detected in the WT, AP001, or AP002 strains. Moreover, as in our previous report [10], deletion of fur clearly increased halo formation on the CAS plate. However, the halo was reduced in the AP003 (Dfur) strain background by the further deletion of iscR, indicating the positive role of IscR in iron acquisition. Furthermore, no obvious difference in siderophore secretion was found between the AP003 and AP005 (Dfur-iscR 3CA ) strains, suggesting that the IscR regulation of iron acquisition does not require the [2Fe-2S] cluster.
To verify whether apo-IscR activates iron acquisition, we introduced pACYC184, pIscR, or pIscR 3CA into the AP004 (Dfur-DiscR) strain and performed a CAS assay. As shown in Fig. 4B, the introduction of both pIscR and pIscR 3CA increased the halo phenotype on the CAS plate compared with that of the vector-only control. These results confirmed that IscR activates siderophore secretion in a [2Fe-2S] cluster-independent manner.
To further investigate the regulatory effect of IscR on iron acquisition, we used quantitative reverse transcription polymerase chain reaction (qRT-PCR) to measure the expression of genes corresponding to the eight putative iron acquisition systems in the indicated K. pneumoniae strains. As shown in Table 1, messenger RNA (mRNA) levels of genes (fhuA, iucA, and sitA) corresponding to three iron acquisition systems were increased more than 2-fold in the AP005 strain as compared with that in the AP004 strain. To further confirm this result, pACYC184, pIscR 3CA , or pIscR were respectively introduced into the AP004 strain, to avoid the effects of Fur, and the transcription of fhuA, iucA, and sitA were measured. The introduction of pIscR 3CA into AP004 apparently increased the transcription of fhuA, iucA, and sitA compared with that in the AP004 strain carrying pACYC184 only (Table 1). These results implied that apo-IscR activates the transcription of fhu, iuc, and sit to increase iron acquisition in K. pneumoniae. Besides, the introduction of pIscR into AP004 also slightly increased transcription of fhu, iuc, and sit (Table 1).
IscR 3CA directly binds the promoter region of fhuA, iucA, and sitA Apo-IscR has been demonstrated to bind the type 2 IscR box in IscR-regulated promoter sequences directly in E. coli [19]. Analysis of the promoter regions of fhuA, iucA, and sitA revealed consensus sequences of the E. coli type 2 IscR box. As shown in Fig. 5A, the predicted type 2 IscR boxes are located at 2154 to 2130 relative to the translation start site of fhuA and 267 to 243 relative to the translation start site of iucA. The predicted type 2 IscR boxes in P fhuA and P iucA have 50% (13/26 bp) and 46% (12/ 26 bp) homology, respectively, with the consensus sequence. In addition, two putative type 2 IscR boxes (R1 and R2) located at 2112 to 287 and at 253 to 228 relative to the translation start site of sitA were found in P sitA . The R1 and R2 sites contain 50% (13/26 bp) and 61.5% (16/26 bp) homology, respectively, with the consensus sequence.
To verify whether apo-IscR binds to these predicted type 2 IscR boxes, we performed an EMSA. As shown in Fig. 5B-D, both the purified IscR::His 6 and IscR 3CA ::His 6 were able to bind with the promoter regions of fhuA, iucA, and sitA, and IscR::His 6 appeared to contain higher binding activities. Furthermore, IscR 3CA ::His 6 did not bind PfhuA-2 and PiucA-2, which lacked a region containing a putative IscR box ( Fig. 6B-C). We also found that IscR 3CA ::His 6 did not bind PiucA-2, which contained the R1 site but not the R2 site (Fig. 6D). To further confirm the importance of these predicted IscR boxes, recombinant IscR proteins were respectively interacted with these promoters lacking only the 26-bp predicted type 2 IscR box (PfhuA-1*, PiucA-1*, and PiucA-1*), and no obvious interaction was found. These results suggested that apo-IscR interacts directly with the promoters of fhuA, iucA, and sitA via the predicted type 2 IscR boxes.

Regulatory control of iscR transcription in K. pneumoniae
To analyse whether environmental iron availability affects K. pneumoniae iscR expression, we grew AP006 in Luria-Bertani (LB) broth supplemented with increasing amounts of the iron chelator 2, 2-dipyridyl (Dip) and monitored the promoter activity of iscR box on PgalF-1* is indicated by a caret. (B) EMSA of IscR recombinant proteins and various DNA fragments of the upstream regions of galF. Different concentrations of purified IscR::His 6 or IscR 3CA ::His 6 were incubated with 5 ng of DNA fragments, as indicated in the margin. After incubation at room temperature for 30 min, the mixtures were resolved on a 5% non-denaturing polyacrylamide gel. The gel was stained with SYBR Green I dye and photographed. doi:10.1371/journal.pone.0107812.g002 Figure 3. Deletion effect of iscR on K. pneumoniae susceptibility to normal human serum. The susceptibility to normal human serum of each bacterial mutant (A) and the complement strains (B) indicated in the margin was determined. Bacterial serum resistance was determined using log-phase cultures grown in LB medium. *P,0.01 compared with the indicated groups. doi:10.1371/journal.pone.0107812.g003 using a LacZ reporter system [23]. As shown in Fig. 6A, the addition of 250 or 500 mM Dip to the growth medium increased iscR promoter (P iscR ) activity by approximately 2-fold and 3.9-fold, respectively, indicating that the transcription of the iscRSUA operon was activated by iron limitation.
In K. pneumoniae, Fur and RyhB reportedly play crucial roles in gene regulation in response to cellular iron levels [10,11]. Thus, we investigated whether Fur and RyhB regulate the activity of P iscR . As shown in Fig. 6B, the deletion of fur and the further deletion of ryhB in AP006 strain had no obvious effects on P iscR activity, whereas the deletion of iscR in AP006 strain activated P iscR activity by approximately 4.5-fold. In E. coli, IscR has been demonstrated to exert negative auto-regulation which requires the [2Fe-2S] cluster [20]. Thus, we measured P iscR activity in an AP006-derived strain, AP008 (DlacZ-iscR 3CA ), expressing a mutated IscR predicted to be defective in cluster binding [20]. As shown in Fig. 6B, the P iscR activity in AP008 was increased approximately 4.5-fold. This increase was comparable to that in AP007 (DlacZ-DiscR). Our results suggested that IscR inhibits the  Table 1. qRT-PCR analyses of the expression of iron-acquisition genes in K. pneumoniae strains. transcription of the iscRSUA operon in a [2Fe-2S] clusterdependent manner in K. pneumoniae.
To investigate whether IscR is the sole regulator of iscRSUA transcription in response to iron availability, we monitored P iscR activity in the AP007 and AP008 strains in LB broth containing various levels of iron. As shown in Fig. 6C, the P iscR activity in AP006 was activated in LB broth supplemented with Dip, and the further addition of 100 or 250 mM FeSO 4 reversed the activation. In the AP007 and AP008 strains, P iscR activity was increased compared with that of AP006. Nevertheless, the addition of Dip still activated P iscR activity, suggesting the presence of unknown factors, and the further addition of FeSO 4 restored the effect.
To analyse whether Fur is responsible for this regulation, we measured P iscR activity in AP011 (DlacZ-Dfur-DiscR) compared with that of AP007 at various iron levels and noted no obvious effect. These results indicated that in addition to IscR and Fur, other factors modulate the transcription of the iscRSUA operon in response to environmental iron availability.

Discussion
Clinically isolated K. pneumoniae strains usually carry large amounts of CPS to resist engulfment by phagocytes and serum bactericidal factors [6,24]. Therefore, tightly controlling CPS biosynthesis is critical for successful infection by K. pneumoniae [10,11,25]. We have previously shown that Fur represses the expression of mucoid factors RmpA and RcsA as well as the small RNA RyhB in response to environmental iron to decrease CPS biosynthesis indirectly in K. pneumoniae [10,11]. In this study, we focused on IscR, a central regulator of iron metabolism, to analyse more thoroughly how external iron affects CPS biosynthesis. Our data indicated that IscR activates CPS biosynthesis in a Fe-S cluster-dependent manner ( Fig. 1A and B). Moreover, IscR positively regulates the transcription of three transcription units in the cps gene cluster (Fig. 1C). Purified IscR::His 6 also directly interacts with the promoter of orf1-2, possibly through the predicted type 1 IscR box (Fig. 2). These findings indicated that Fur and IscR exert negative and positive regulation, respectively, on CPS biosynthesis in response to external iron.
However, we hypothesize herein that Fur plays a major regulatory role because (i) K. pneumoniae grown under ironreplete conditions displayed decreased CPS levels, (ii) the AP004 strain produced elevated amounts of CPS compared with that of the WT and in the AP004 background, expression of IscR or IscR 3CA did not cause an obvious effect on CPS amount under iron-replete or iron-limited conditions (data not shown), and (iii) all three transcription units of the cps genes were obviously increased after fur deletion [10]. This result supported the notion that Fur plays a major regulatory role in the regulation of CPS biosynthesis. Nevertheless, the contribution of IscR could also be an important part of the iron network. On one hand, in iron-replete conditions, Fur indirectly represses the CPS production, however CPS is still needed for K. pneumoniae survival inside the host, so IscR may contribute in maintaining a higher basal expression level of the genes involved in CPS biosynthesis. On the other hand, in ironlimited conditions, the transcription of iron-acquisition genes is increased not only due to the de-repression of apo-Fur but also the activation of apo-IscR. Besides, in addition to iron level, oxidative stress has been demonstrated to signal Fur and IscR, and the reversible interconversion of Fe-S clusters makes them exquisite sensors of such stress [13,18,26]. Thus, the modulation of CPS levels by Fur and IscR under oxidative stress conditions could be predicted, but this hypothesis remains to be elucidated.
A BLAST search identified eight putative iron acquisition systems in the genome of K. pneumoniae CG43S3. Moreover, Fur directly represses the transcription of genes corresponding to six iron acquisition systems (iucA, iroB, entC, hmuR, feoA, and fecA) under iron-replete conditions [10]. In this study, we have shown that IscR directly activates the transcription of fhuA, sitA, and iucA (Table 1 and Fig. 5). Although fur deletion also led to increased expression of these three iron acquisition systems, an in vivo Fur titration assay demonstrated that Fur did not interact with the promoters of fhuA and sitA and also had a relatively lower binding affinity to iucA [10]. These findings indicated that Fur and IscR orchestrate the expression of iron acquisition systems in K. pneumoniae. In addition, Fur is suggested to play a major regulatory role because deletion of fur, but not iscR deletion, activated the secretion of siderophores, and the effect mediated by IscR was observed only in the fur-deleted background (Fig. 4).
Because IscR and Fur are both iron-responsive regulators which share strong overlap among downstream targets, we next investigated whether IscR and Fur are cross-regulated. In E. coli, the expression of the iscR is controlled not by Fur but by IscR itself [19,20]. In K. pneumoniae, we found that P iscR activity is clearly activated by IscR but not influenced by Fur (Fig. 6), which is consistent to the finding in E. coli. Moreover, deletion of iscR did not affect the transcription of fur (data not shown), suggesting that in K. pneumoniae, IscR and Fur do not share cross-regulation. Furthermore, E. coli RyhB reportedly binds directly to the upstream region of iscS mRNA to decrease the stability of iscRSUA but not iscR mRNA, thereby leading to a stable secondary iscR structure and resulting in active translation [27]. Then, the increased IscR likely causes auto-repression of P iscR activity.
An analysis of the upstream region of iscS in K. pneumoniae, also identified a conserved sequence paired with RyhB (data not shown). However, the regulatory effect on P iscR activity mediated by RyhB was not obvious under our assay conditions (Fig. 6B). On the contrary, as shown in Fig. 6C, P iscR activity in AP007 and AP008 was still activated by iron depletion, which prompted us to verify whether IscR is the sole iron-responsive regulator that controls P iscR activity. However, in AP007, the addition of FeSO 4 to iron-depleted medium still led to a reduction in P iscR activity. These results suggested that an unknown regulator, beside of IscR and Fur, represses the transcription of iscR in response to external iron. In addition to IscR, the [2Fe-2S] cluster is critical for regulation mediated by FNR and SoxR [27]. However, sequence analysis of the promoter region of iscR revealed no typical FNR and SoxR binding sites. On the contrary, we found a putative binding site of SoxS [28], an oxidative transcriptional regulator activated by SoxR, in P iscR . This putative SoxS binding site displays 79% (15/19 bp) homology with the consensus sequence Figure 5. IscR 3CA ::His 6 binds directly to P fhuA , P iucA , and P sitA . (A) DNA sequence alignment between the E. coli type 2 IscR box and the putative IscR binding sequence in the upstream regions of fhuA, iucA, and sitA. Positions identical to the consensus sequences are bolded. Diagrammatic representation of the fhuA (B), iucA (C), and sitA (D) loci. The large arrows represent the open reading frames. The primer sets used in PCR amplification of the DNA probes are indicated, and the numbers denote the DNA amplified length. The predicted IscR boxes is deleted and indicated by a caret. The grey boxes indicate the predicted type 2 IscR box. Different concentrations of purified IscR 3CA ::His 6 were incubated with 5 ng of various DNA fragments of the upstream regions of indicated genes. Following incubation at room temperature for 30 min, the mixtures were analysed on a 5% non-denaturing polyacrylamide gel. The gel was stained with SYBR Green I dye and photographed. doi:10.1371/journal.pone.0107812.g005 and is located at position 2144 to 2126 relative to the translation start site of iscR. Therefore, we hypothesized that SoxS may be involved in the regulation of iscR in response to oxidative stress and will investigate this possibility in future studies.
IscR differs from other known Fe-S cluster-containing transcription factors such as FNR and SoxR because both apo-and holo-IscR regulate transcription and exhibit different DNA binding specificities [29]. Structural and biochemical studies have suggested that the ligation of the [2Fe-2S] cluster broadens the DNA binding specificity of IscR, thereby allowing holo-IscR to bind both type 1 and type 2 boxes, whereas apo-IscR binds only the type 2 box [29]. In the K. pneumoniae K2 cps gene cluster, we found a type 1 box in the promoter region of galF, and purified IscR::His 6 could bind this motif in vitro (Fig. 2). In addition and as expected, the clusterless IscR 3CA :His 6 , mimicking apo-IscR, showed no obvious binding affinity to P galF . On the contrary, all three promoters of the iron acquisition genes regulated by IscR contain predicted type 2 boxes, and both IscR::His 6 and IscR 3CA :His 6 appeared to bind these boxes (Fig. 5). Although apo-and holo-IscR have been demonstrated to bind the type 2 box with similarly high affinity [21], EMSA revealed that IscR 3CA :His 6 displayed weaker binding (Fig. 5). In E. coli, the Fe-S cluster status of IscR is a key variable that regulates gene expression in response to iron availability [12]. Our results suggested that the transcription of cps and iron acquisition genes is regulated by not only the level of IscR but also the cellular ratio of apo-and holo-IscR.
Although IscR was first discovered as an auto-repressor of the isc operon, it is now known to be a global regulator that influences the expression of ,40 genes in E. coli and ,67 genes in Vibrio vulnificus [19,30]. Although the regulon of IscR in K. pneumoniae has not been identified, our study provided evidence that IscR regulates CPS biosynthesis and iron acquisition, which are also regulated by Fur. Moreover, previous studies have shown that Fur controls type 3 fimbriae expression and biofilm formation in K. pneumoniae [31]. The possibility that IscR also participates in the regulation of these phenotypes is currently under investigation.
After infection, bacterial pathogens frequently encounter iron starvation and oxidative/nitric oxide stress conditions that are highly detrimental for maintaining Fe-S cluster homeostasis [32,33,34]. These conditions are predicted to influence the transcription of genes regulated by IscR. Thus, in response to these stress conditions, IscR may regulate CPS biosynthesis, iron acquisition systems, and other virulence factors in K. pneumoniae to facilitate bacterial persistence in the host. In this study, we demonstrated that external iron levels regulate CPS biosynthesis and iron acquisition systems through IscR in K. pneumoniae and proposed a working model (Fig. 7). In response to iron availability, IscR and Fur control the expression of downstream targets in a parallel and cooperative manner, which is predicted to play a crucial regulatory role during infection.

Bacterial strains, plasmids, and media
Bacterial strains and plasmids used in this study are listed in Table 2. Primers used in this study are list in Table 3. Bacteria were routinely cultured at 37uC in LB medium supplemented with appropriate antibiotics. The antibiotics used include ampicillin (100 mg/ml), kanamycin (25 mg/ml), streptomycin (500 mg/ml), and tetracycline (12.5 mg/ml).

Construction of the deletion of iscR mutants
Specific gene deletion of iscR was introduced into K. pneumoniae CG43S3 using an allelic exchange strategy as previously described [35]. In brief, two approximately 1000 bp DNA fragments flanking both sides of iscR were cloned into the suicide vector pKAS46 [36], a suicide vector containing rpsL, which allows positive selection with streptomycin for vector loss. The resulting plasmid was then mobilized from E. coli S17-1lpir [37] to K. pneumoniae CG43S3 or CG43S3-derived strains by conjugation. The transconjugants, with the plasmid integrated into the chromosome via homologous recombination, were selected with ampicillin and kanamycin on M9 agar plates. Several of the colonies were grown in LB broth supplemented with 500 mg/mL of streptomycin to log phase at 37uC and then spread onto an LB agar plate containing 500 mg/mL of streptomycin. The streptomycin-resistant and kanamycin-sensitive colonies were selected, and the deletion was verified by PCR and Southern hybridization (data not shown). The resulting K. pneumoniae mutants are listed in Table 2.

Construction of the pIscR complementation plasmid and the pIscR 3CA mutant plasmid
To obtain the complementation plasmid (pIscR), a DNA fragment containing the promoter and coding sequence of iscR was amplified by PCR using the primer pair GT138/GT139 (Table 3) and cloned into the pACYC184 shuttle vector. The pIscR 3CA plasmid, which carried the mutant allele encoding IscR with the C92A, C98A, and C104A mutations, was constructed using the inverse-PCR method. Briefly, the pIscR plasmid was used as the PCR template to generate the mutant allele with the primer pair GT206/GT207 ( Table 3). The recovered PCR product was treated with DpnI for 2 h, subjected to T4 polynucleotide kinase treatment, and self-ligated with T4 DNA ligase. The ligation product was transformed into E. coli DH5a. The pIscR 3CA plasmid was subsequently confirmed by sequence analysis.
Construction of a K. pneumoniae iscR 3CA mutant A DNA fragment carrying iscR and approximately 1000 bp adjacent regions on either side was amplified by PCR using primer pairs GT241/GT242 (Table 3) and cloned into yT&A. The resulting plasmid was used as the template for the inverse-PCR with the primer pair GT206/GT207 (Table 3) to generate a mutant iscR allele encoding the C92A, C98A and C104A mutations. Subsequently, the mutant allele of iscR was subcloned into pKAS46 and confirmed by DNA sequencing. Then, the plasmid was mobilized from E. coli S17-1 lpir to the K. pneumoniae AP001 strain by conjugation, and the subsequent selection was performed as described above.

Extraction and quantification of CPS
CPS was extracted and quantified as previously described [38]. The glucuronic acid content, represents the amount of K. pneumoniae K2 CPS, was determined from a standard curve of glucuronic acid (Sigma-Aldrich) and expressed as micrograms per 10 9 c.f.u. [39].

Measurement of promoter activity
The promoter-reporter plasmids, pOrf12, pOrf315, pOrf1617, and piscRZ15 were individually mobilized into K. pneumoniae strains by conjugation from E. coli S17-1 lpir. The bacteria were grown to logarithmic phase in LB broth or indicated medium, and the b-galactosidase activity was measured as previously described [23].

Bacterial survival in serum
Normal human serum, pooled from healthy volunteers, was divided into equal volumes and stored at 270uC before use. Bacterial survival in serum was determined as previously described [35]. First, the bacteria were grown to log phase in LB broth and the viable bacterial concentration was adjusted to 1610 6 c.f.u./ml. Next, 1 ml of the cultures was washed twice using phosphatebuffered saline (PBS) and resuspended in 1 ml PBS. A mixture containing 250 ml of the cell suspension and 750 ml of pooled human serum was incubated at 37uC for 15 min. The number of viable bacteria was then determined by plate counting. The survival rate was expressed as the number of viable bacteria treated with human serum compared with the number of viable bacteria pretreatment. The 0% survival of K. pneumoniae AP012 (DgalU) served as a negative control.
Purification of IscR::His 6 and IscR 3CA ::His 6 The coding regions of iscR and iscR 3CA were amplified using the primer pair GT215/GT216 (Table 3) and cloned into the NheI/XhoI site in pET30b (Novagen, 205 Madison, Wis). The resulting plasmids (pET30b-IscR and pET30b-IscR 3CA , respectively) were then transformed into E. coli BL21(DE3)[pLysS] (Invitrogen, USA), and overproduction of the recombinant proteins IscR::His 6 and IscR 3CA ::His 6 , respectively, were induced by the addition of 1 mM IPTG for 3 h at 37uC. The cell pellets were washed and resuspended in cold binding buffer (20 mM sodium phosphate, 0.5 M NaCl, 5 mM imidazole, pH 7.4). The cells were then broken by sonication and the cell pellets were removed by centrifugation at 14000 rpm for 10 min at 4uC. The recombinant proteins were then purified from the soluble fraction of the total cell lysate by affinity chromatography using His-Bind resin (Novagen, Madison, Wis) according to the manufacturer's instructions. The nonbinding proteins were washed away using binding buffer and the recombinant proteins were eluted by elution buffer (20 mM sodium phosphate, 0.5 M NaCl, 500 mM imidazole, pH 7.4). Finally, the purified proteins were dialyzed against TEGX buffer [20 mM Tris-HCl (pH 8.0), 0.5 mM EDTA (pH 8.0), 10% (v/v) glycerol, 0.2% Triton-X 100] containing 0.1 mM NaCl at 4uC overnight. Subsequently, the dialyzed protein was checked for purity by SDS-PAGE and stored for up to two weeks at 4uC. The purified protein was transparent and no obvious precipitation was observed after storage.
For the EMSA, the purified IscR::His 6 and IscR 3CA ::His 6 proteins were incubated with 5-ng DNA in a 10 ml solution containing 4 mM Tris-HCl (pH 7.4), 10 mM KCl, 100 mM dithiothreitol, and 10 mg/ml BSA at 37uC for 30 min. The samples were then loaded onto a native gel of 5% nondenaturing polyacrylamide in 0.56 TB buffer (45 mM Tris-HCl, pH 8.0, 45 mM boric acid). Gels were electrophoresed with a 20-mA current at 4uC and then stained with SYBR Green I dye (Invitrogen). The assay was repeated in at least 3 independent experiments.

qRT-PCR
Total RNAs were isolated from early-exponential-phase grown bacteria cells by use of the RNeasy midi-column (QIAGEN) according to the manufacturer's instructions. RNA was DNasetreated with RNase-free DNase I (MoBioPlus) to eliminate DNA contamination. RNA of 100-ng was reverse-transcribed with the Transcriptor First Strand cDNA Synthesis Kit (Roche) using random primers. qRT-PCR was performed in a Roche Characterization of IscR in Klebsiella pneumoniae LightCycler 1.5 Instrument using LightCycler TaqMan Master (Roche). Primers and probes were designed for selected target sequences using Universal ProbeLibrary Assay Design Center (Roche-applied science) and listed in Table 3. Data were analyzed using the real time PCR software of Roche LightCycler 1.5 Instrument. Relative gene expressions were quantified using the comparative threshold cycle 2 2DDCT method with 23S rRNA as the endogenous reference.

CAS assay
The CAS assay was performed according to the method described by Schwyn and Neilands [40]. Each of the bacterial strain was grown overnight in LB medium, and then 5 ml of culture was added onto a CAS agar plate. After 24 h incubation at 37uC, the effects of the bacterial siderophore production could be observed. Siderophore production was apparent as a halo around the colonies; the absence of a halo indicated the inability to produce siderophores.

Statistical methods
An unpaired t-test was used to determine the statistical significance and values of P,0.01 were considered significant. The results of CPS quantification, b-galactosidase activity, serum survival rate, and qRT-PCR analysis were performed in triplicate and independently repeated at least three times, and the mean activity and standard deviation are presented.

Ethics statement
For isolation of normal human serum from healthy volunteers, the procedure and the respective consent documents were approved by the Ethics Committee of the China Medical University Hospital, Taichung, Taiwan. All healthy volunteers provided written informed consent. Figure S1 Single-copy complementation of iscR but not iscR 3CA in the AP001 strain restores native production levels of CPS. CPS levels of the K. pneumoniae strains, as indicated, grown in LB broth were determined as described in Materials and Methods (*P,0.01).

(TIF)
Methods S1 Supporting Materials and Methods.