Figure

Fig. 2.1. Deletion effects of ugd, wza and rcsB genes on K. pneumoniae K2 CPS production and resistance to polymyxin B

(A) Comparison of colony morphology. The K. pneumoniae strains were streaked on an LB agar plate, incubated at 37℃ overnight and photographed. (B) Sedimentation test. The strains were cultured overnight in LB broth at 37 ℃ and subjected to centrifugation at 4,000 ×g for 5 min. Quantification of K2 CPS amounts of each strain is shown below the figure. Values are shown as averages ± standard deviations from triplicate samples. (C) Polymyxin resistance assay. The log-phased cultures of K. pneumoniae CG43S3, Δugd, Δwza or ΔrcsB mutants were challenged with 1 or 2 units/mL of polymyxin B. (D) Polymyxin resistance assay. The log-phased culture of K. pneumoniae strains were challenged with 2 or 4 units/mL of polymyxin B. The survival rates are shown as the average ± standard deviations from triplicate samples. *, P < 0.01 compared with the parental strain CG43S3. #, P < 0.01 compared with each strain carrying pRK415.

(A)

Polymyxin B resistance assay (2 units/mL)

LB-culture LB (1 mM Fe³⁺) culture

% Relative survival

Fig. 2.2. Involvement of K. pneumoniae pmrF gene in polymyxin B resistance and intra-macrophage survival

(A) The log-phased cultures of K. pneumoniae CG43S3, the ΔpmrF mutant or ΔpmrF carrying pRK415-PmrF were grown in LB or LB supplemented with 1mM Fe³⁺ and then challenged with 2 units/mL of polymyxin B. The survival rates are shown as the average ± standard deviations from triplicate samples. (B) The survival rates of K. pneumoniae CG43S3ΔrcsB, the isogenic ΔpmrFΔrcsB mutant, and ΔpmrFΔrcsB mutant strain carrying the complementation plasmid pRK415-PmrF within the mouse macrophage RAW264.7 were determined. The results shown are relative survival rates which were calculated from the viable colony counts of intracellular bacteria divided by individual original inoculums. Values are shown as the average of five replicas. Error bars, standard deviations. *, P < 0.01 compared with the parental strain CG43S3. #, P < 0.01 compared with ΔpmrF mutant strain.

(A)

Polymyxin B resistance assay (2 units/mL)

% Relative survival

Polymyxin B resistance assay (2 units/mL)

#

Fig. 2.3. Effects of K. pneumoniae pmrA, pmrD and phoP deletion and complementation in polymyxin B resistance and intra-macrophage survival

(A) The log-phased cultures of K. pneumoniae CG43S3, the ΔpmrA, ΔpmrD or ΔphoP mutants were grown in LB, LB supplemented with 10 mM Mg²⁺ or LB supplemented with 1mM Fe³⁺

and then challenged with 2 units/mL of polymyxin B. The survival rates are shown as the average ± standard deviations from triplicate samples. (B) The log-phased cultures of K.

pneumoniae CG43S3 carrying pRK415, the ΔpmrAΔphoP mutant strains carrying pRK415, pRK415-PhoP or pRK415-PmrA were grown in LB and challenged with 2 units/mL of polymyxin B. The survival rates are shown as the average ± standard deviations from triplicate samples. (C) The survival rates of K. pneumoniae CG43S3ΔrcsB, the isogenic ΔpmrAΔrcsB, ΔphoPΔrcsB and ΔpmrDΔrcsB mutants, and each mutant strain carrying the complementation plasmids pRK415-PmrA, pRK415-PhoP or pRK415-PmrD within the mouse macrophage RAW264.7 were determined. The results shown are relative survival rates which were calculated from the viable colony counts of intracellular bacteria divided by individual original inoculums. Values are shown as the average of five replicas. Error bars, standard deviations. *, P < 0.01 compared with the parental strain CG43S3 or CG43S3ΔrcsB. #, P < 0.01 compared with ΔpmrAΔphoP[pRK415]. &, P < 0.01 compared with ΔphoPΔrcsB or ΔpmrDΔrcsB.

(A)

Fig. 2.4. Schematic representation of pmrH and pmrD loci and determination of K.

pneumoniae P_pmrH::lacZ and P_pmrD::lacZ activity

(A) Diagrammatic representation of the pmrH and pmrD loci. The large arrows represent the open reading frames. The relative positions of the primer sets used in PCR-amplification of the DNA fragments encompassing the PpmrH and PpmrD regions are indicated, and the numbers denote the relative positions to the translational start site. The name and approximate size of the DNA probes used in electro-mobility shift assay (EMSA) are shown on the left. The dashed boxes indicate the predicted PhoP and PmrA binding sequences and the alignment result is shown below. The identical nucleotide sequences are underlined. HP, hypothetical protein. (B) The β-galactosidase activities of log-phased cultures of K. pneumoniae strains carrying placZ15-PpmrH grown in conditions indicated in the figure were determined and expressed as Miller units. The data shown were the average ± standard deviations from triplicate samples. *, P < 0.01 compared with CG43S3ΔlacZ or ΔpmrDΔlacZ mutant grown in LB medium. #, P <

0.01 compared with CG43S3ΔlacZ grown in LB medium supplemented with ferric ions. &, P

< 0.01 compared with CG43S3ΔlacZ grown in LB medium. (C) The β-galactosidase activities of log-phased cultures of K. pneumoniae strains carrying placZ15-PpmrD grown in conditions indicated in the figure were determined and expressed as Miller units. The data shown were the average ± standard deviations from triplicate samples. *, P < 0.01 compared with CG43S3ΔlacZ or ΔpmrAΔlacZ mutant grown in LB medium. #, P < 0.01 compared with CG43S3ΔlacZ grown in LB medium.

(A)

(B)

Fig. 2.5. Binding of His-PhoP and His-PhoPN149 to PpmrD

(A) Specific binding of recombinant His-PhoP protein to the putative pmrD promoter. EMSA was performed by using the ³²P-labeled DNA probe of PpmrD incubated with increasing amounts of the His-PhoP (lanes 2 to 5), with 40 pmole of His-PhoP plus increasing amounts of the unlabeled PpmrD DNA (specific competitor, lane 6 to 9), or with excess amounts of non-specific competitor DNA (lane 10 and 11). The amounts of recombinant proteins and DNA probes used are indicated in the figure. (B) EMSA was performed with 0, 4 or 40 pmole of His-PhoP (lanes 1 to 3), His-PhoPN149 (lanes 4 to 6) or 100 pmole of BSA (lane 7). The arrows indicate the PhoP/PpmrD complex and free probe of PpmrD.

(A)

(B)

(C)

(D)

Fig. 2.6. Klebsiella PmrD interacts with PmrA to prevent dephosphorylation

(A) Bacterial two-hybrid analysis of PmrD/PmrA interaction in vivo. The E. coli XL1-Blue cells co-transformed with various combinations of pTRG and pBT-derived plasmids were diluted serially and spotted onto the indicator plate. The bacterial growth after 36h was investigated and photographed. Combinations of plasmids carried by each strain were indicated above the figure. (B) Klebsiella PmrD prevents the dephosphorylation of PmrA by its cognate sensor protein. The phosphorylation state of the recombinant His-PmrA protein was monitored upon the addition of the sensor protein His-PmrBC276 in the presence (PmrD) or absence (-) of purified His-PmrD protein at specific time points as indicated. The arrows indicate phospho-PmrA (P-PmrA) and phospho-PmrBC276 (P-PmrBC276). (C) Kinase/phosphatase assay was carried out using the recombinant His-PmrA (final concentration 5 μM) and His-PmrBC276 (final concentration 2.5 μM) in the presence (PmrD) or absence (-) of the recombintant His-PmrD protein (final concentration 5 μM). The small cationic proteins RNase A and cytochrome C were introduced individually as a negative control at a final concentration of 5 μM. (D) Autokinase assay of the recombinant His-PmrBC276 (final concentration 2.5 μM) was performed in the presence (PmrD) or absence (-) of the recombintant His-PmrD protein (final concentration 5 μM).

(A)

Fig. 2.7. Identification of critical residues involved in Klebsiella PmrD functioning

(A) Sequence alignment of PmrD homologues from different bacterial species. The predicted secondary structural of Klebsiella PmrD was shown above the figure. Residues selected for the construction of plasmids encoding PmrD with point mutations were indicated by arrowheads.

(B) The log-phased cultures of K. pneumoniae CG43S3ΔpmrD mutant strain carrying pRK415 (vector control), pRK415-PmrD (PmrD) or pRK415-PmrD derived plasmid encoding PmrD with each point mutation were grown in LB and then challenged with 4 units/mL of polymyxin B. The survival rates are shown as the average ± standard deviations from triplicate samples. *, P < 0.01 compared with each strain carrying pRK415-PmrD. (C) The phosphorylation state of the recombinant His-PmrA protein was monitored upon the addition of the sensor protein His-PmrBC276 in the absence (-) or presence (PmrD or PmrDC35S) of purified His-PmrD or His-PmrDC35S proteins at specific time points as indicated. The arrows indicate phospho-PmrA (P-PmrA) and phospho-PmrBC276 (P-PmrBC276) respectively.

Fig. 2.8. A model illustrating the regulation of polymyxin B resistance in K. pneumoniae by PhoP/PhoQ, PmrD and PmrA/PmrB

When environmental magnesium concentrations are low (micro-molar), the PhoP/PhoQ system is turned on, resulting in the expression of pmrD through a direct binding of PhoP to PpmrD. The PmrD protein then binds to the phosphorylated form of PmrA, which can also be generated when environmental ferric ion concentrations are high (sub-milimolar). Either PmrA or PhoP can activate the expression of pmr genes, eventually rendering the bacteria more resistant to polymyxin B.

CHAPTER 3

Functional Characterization of The Two-component

System RstA/RstB in Klebsiella pneumoniae CG43

3.1 Abstract

The two component system (2CS) RstA/RstB has been reported as a member of another 2CS PhoP/PhoQ and has been implicated in the regulation of bacterial acid tolerance response, iron transport and the resistance against digestive tract stresses among several enteric bacteria. Firstly, to investigate the regulation of rstA expression in Klebsiella pneumoniae, the PrstA::lacZ transcriptional fusion harboring a DNA fragment encompassing the putative rstA promoter region was constructed. Expression of PrstA::lacZ was reduced in ΔphoP or ΔrstA mutant strains, indicating a positive regulatory role of PhoP and RstA.

Subsequently, to explore the functional role of RstA/RstB in K. pneumoniae, subtractive cDNA hybridization was performed, yielding a total of 11 RstA-activated genes and 19 RstA-repressed genes involved in various cellular functions. Comparative phenotype analyses, including the bacterial growth under iron depletion/repletion conditions, lead resistance, siderophore production, acid stress response and resistance to bile salts, of the wild-type, ΔrstA, ΔrstB and ΔrstAΔrstB mutant strains were carried out. None of the assays yielded results in which the deletion mutants exhibited behaviors apparently different from those observed in the parental strain. This indicates the functional role of RstA/RstB still requires further investigation.

3.2 Introduction

The signal transduction cascade governed by 2CS has been shown to be important in the expression of virulence factors during infections or involved in bacterial survival in hostile environments (12). For example, PhoP/PhoQ, perhaps the most completely characterized 2CS to date, governs the Mg²⁺ regulon and a repertoire of virulence genes in Salmonella enterica (8, 90, 164), and is also required for virulence in Erwinia chrysanthemi (149), Neisseria meningitides (179), Yersinia pestis (101), Pseudomonas aeruginosa (82, 83) and Photorhabdus luminescens (49). In addition to the typical mode of regulation, PhoP-regulated genes can also be expressed when activated by another 2CS PmrA/PmrB through the connector protein PmrD, which in turn activated the bacterial resistance to polymyxin B (59, 117).

Compared with the well-documented characterization of the 2CS PmrA/PmrB, the functional role of RstA/RstB, another 2CS regulated by PhoP/PhoQ, has remained relatively unclear and explored mostly by genome-wide studies (99, 100, 264). Previously, rstA was identified as a multicopy suppressor for an essential ATPase encoding gene in Escherichia coli K12 (29), and recently a more complex network involving YeaZ and YgiD has implied a functional role of RstA between DNA metabolism and cell division (96). RstA/RstB was also speculated to be involved in bacterial acid resistance since asr, which encodes acid shock RNA (214, 228), was identified as one of its downstream targets by genomic SELEX search (183). In S. enterica, RstA was found to be able to induce RpoS degradation (28) and modulated Fur activity via the activation of the ferrous iron transporter FeoB (111). More recently, Salmonella PhoP/PhoQ was found to either activate iron transporter encoding gene feoB through RstA or to induce the expression of mgtA encoding the magnesium transporter in response to different environmental signals (41). Moreover, a comprehensive study on mutations of all 2CS response regulator encoding genes in Yersinia pesudotuberculosis has indicated the involvement of RstA in bacterial resistance to the stresses encountered in the digestive tract (69).

Sequence analysis of the rstAB locus in different bacterial species showed that the gene organization of K. pneumoniae rstAB was similar to its counterpart in S. enterica in which the two genes were separated rather than physically linked, as was the case in E. coli. To search for the downstream genes of RstA/RstB in K. pneumoniae, subtractive cDNA hybridization was preformed to identify cDNA fragments specifically present in the wild-type strain or in

the isogenic rstA mutant strain lacking the response regulator encoding gene. Phenotypic comparison of the parental strain to the K. pneumoniae ΔrstA, ΔrstB mutant which lacked the sensor kinase encoding gene and the ΔrstAΔrstB double mutant strain was performed in order to validate the functional role of RstA/RstB.

3.3 Results

3.3.1 Regulation of rstA expression in K. pneumoniae CG43

A close examination of the K. pneumoniae rstAB region has revealed a PhoP box-like sequence as identified in E. coli (120) and S. enterica (220). In addition, sequence analysis has indicated the presence of the RstA box recently identified by a genomic SELEX search (183) (Fig. 3.1A). To investigate if the expression of K. pneumoniae rstA was subjected to the regulation by PhoP/PhoQ and RstA/RstB, reporter plasmids pHY048, pHY050 and pHY053 each carried a transcriptional fusion of the rstA upstream region to a lacZ gene were constructed. Results from the reporter assay in Fig. 3.1B indicated that RstA and PhoP were both required in the activation of rstA expression, as shown by the reduction of promoter activity in the ΔrstA, ΔphoP or ΔrstAΔphoP mutant strains carrying pHY048 or pHY050. The promoter activity in strains harboring pHY053, which did not contain the PhoP box-like sequence, was largely reduced suggesting a major role of PhoP in rstA expression.

Nevertheless, the activity in strains carrying pHY053 was not reduced to levels exhibited by strains with vector alone. As a result, the role of other factors such RstA could not be underestimated. The notion that the activities of strain carrying pHY048 or pHY053 was not abolished in ΔrstAΔphoP mutant further supported the hypothesis that factors other than PhoP or RstA may still be involved in rstA expression in K. pneumoniae.

3.3.2 Identification of RstA-regulated genes by subtractive cDNA hybridization

To explore the functional role of RstA/RstB in K. pneumoniae, subtractive cDNA hybridization was anticipated to search for genes expressed only in the parental strain or in the ΔrstA mutant. Since the promoter activity of strains carrying pHY048 in M9 medium was more than twice as those in LB medium (Fig. 3.2A), it would be conceivable that the a greater number of RstA-regulated genes would be recovered in this condition. Therefore the experiment was performed by using total RNA extracted from M9-grown cultures. The result of second PCR amplification was shown in Fig. 3.2B indicating either the RstA-repressed genes (lane 1) or RstA-activated genes (lane2), each representing differentially-expressed DNA sequences specifically present in the ΔrstA mutant or in the parental strain. The distinct PCR-amplicon patterns also indicated a successful hybridization procedure. Subsequently, the DNA fragments were cloned into yT&A, subjected to sequencing determination, and the sequences were compared in the currently available genome sequences of K. pneumoniae

NTUH-K2044 (262), MGH78578 (http://genome.wustl.edu/), and 342 (74) by BLAST (http://www.ncbi.nlm.nih.gov). The genes were classified according to their predicted functions as shown in Table 3.1 (for RstA-activated genes) and Table 3.2 (for RstA-repressed genes).

3.3.3 Sequence analysis of RstA-regulated genes

The results from nucleotide sequence determination are shown in Table 3.1 and Table 3.2. A total of 11 RstA-activated genes were identified, and a search in the documents has revealed a wide range of cellular functions involved, including fecE encoding an iron dicitrate transport ATP-binding protein (19, 211), yieP encoding an ion channel protein, two open reading frames each with a conserved GGDEF or EAL domain involved in cyclic di-GMP metabolism (98, 206, 210, 234), gabP encoding an RpoS-dependent γ-aminobutyrate transport protein (21, 208), yifP encoding a putative esterase, ybeF encoding a LysR family transcriptional regulator, mnmA encoding a tRNA methyltransferase involved in E. coli tRNA thiolation (114, 182), tldD encoding a DNA gyrase modulator involved in the peptide antibiotic microcin processing and antidote degradation (1, 173) and two open reading frames encoding hypothetical proteins with unknown functions.

Sequence analysis of the retrieved 19 DNA fragments representing the RstA-repressed genes has also revealed the involvement of numerous cellular functions such as iroN encoding an enterochelin and dihydrobenzoic acid receptor (33, 190, 221), pbrR encoding a regulator for the bacterial resistance to heavy metal (16, 112, 194, 232), four genes encoding integral membrane protein or transporter proteins such as mgtE encoding a magnesium transporter (3, 97, 110), hmuR encoding an outer membrane receptor for hemoglobin uptake (216, 238), an open reading frame encoding a fimbrial protein, an open reading frame encoding a β-lactamase-like protein, an open reading frame encoding a putative glycosyl transferase, gpmB encoding a phosphoglycerate mutase (231), gshB encoding a glutathione synthetase (141, 184), an open reading frame encoding a putative acetyltransferase, dnaK encoding a molecular chaperone (79), ppk encoding a polyphosphate kinase (26), yeaG encoding an RpoS-dependent serine kinase belonging to the stress response regulon (109, 233, 249) and two open reading frames encoding hypothetical proteins with unknown functions.

Based on the findings from cDNA subtraction and previous studies (41, 69, 111, 214), it was of interest to investigate if K. pneumoniae RstA/RstB played a role in bacterial growth under iron-depletion/repletion conditions, resistance to lead, siderophore biosynthesis and resistances to stress encountered in the digestive tract.

3.3.4 Effect of iron on the growth of K. pneumoniae strains in rich and minimal medium It has been reported that the two component system RstA/RstB is involved in iron transport in S. enterica (111). Furthermore, subtractive cDNA hybridization has revealed that an iron transporter is among RstA-activated genes. Therefore, the growth of K. pneumoniae wild type along with ΔrstA, ΔrstB, and ΔrstAΔrstB mutant strains under the iron depletion or repletion conditions was investigated. As shown in Fig. 3.3A, strains grown in LB medium reached an OD600 of as high as 1.2 in 6 hours while those grown in LB containing 2’,2’-dipyridyl only reached an OD600 of 0.7. The addition of 2’,2’-dipyridyl resulted in an overall slower bacterial growth in LB medium. To further verify the effect of iron on bacterial growth, strains were grown in M9 minimal medium for the growth curve measurement. As shown in Fig. 3.3B, the strains reach an OD600 of 1.0 in 7 hours when grown in M9 medium while the addition of Fe²⁺ to the medium resulted in a higher OD600 of 1.2. The growth rates of the strains were similar; neither the iron depletion nor repletion significantly altered bacterial growth.

3.3.5 Effect of lead on the growth of K. pneumoniae strains

Initially a disc assay was employed to investigate bacterial resistance to lead; however, when lead nitrate discs were placed on LB, M9 or HMM (heavy metal MOPS) agar plates (132), the lead appeared to react with compounds such as phosphates in the medium and precipitates formed in rings around the discs rapidly. As a result, the growth inhibition assay was performed in HMM broth (132). Results in Fig. 3.4A show that there is evident growth inhibition at and beyond 10 μM Pb²⁺ for K. pneumoniae CG43-101, a derivative of K.

pneumoniae CG43 deprived of a large resident plasmid pLVPK carrying the lead resistance pbr gene cluster (37). When the microtiter plate assay was implemented with the wild type strain CG43S3, the ΔrstA, ΔrstB and ΔrstAΔrstB mutants, more gradual growth inhibition patterns were observed (Fig. 3.4B). The four strains, while differing slightly in their growths at the range of lead concentrations tested, did not show as clear and abrupt inhibition patterns as observed in CG43-101. At concentrations below 10 μM Pb²⁺, several of the strains reach relative growths greater than one, signifying no growth inhibition at these concentrations.

Furthermore, the ΔrstB mutant strain exhibited growth comparable to that of the wild type, while ΔrstA and ΔrstAΔrstB mutants displayed slightly lower relative growths in medium supplemented with the same concentrations of lead nitrate. In the presence of 20 μM Pb²⁺, ΔrstB and ΔrstAΔrstB mutants exhibited a relative growth generally higher than the parental

strain CG43S3 while in the presence of 40 μM Pb²⁺, a notable growth inhibition in all strains was observed. At the highest concentration of 80 μM Pb²⁺, relative growths of around 0.3 were observed. Based on the result of cDNA subtraction, deletion of rstA was expected to enhance the bacterial resistance to lead. However, the finding that deletion of rstA exerted no

strain CG43S3 while in the presence of 40 μM Pb²⁺, a notable growth inhibition in all strains was observed. At the highest concentration of 80 μM Pb²⁺, relative growths of around 0.3 were observed. Based on the result of cDNA subtraction, deletion of rstA was expected to enhance the bacterial resistance to lead. However, the finding that deletion of rstA exerted no