The expression of type 3 fimbriae under the oxygen-limiting

Environmental stimulus that regulates the expression of type 3 fimbriae is yet to be identified. A previous study has demonstrated that K. pneumoniae grown in glycerol minimal medium possess stronger type 3 fimbrial activity than that grown in glucose minimal medium (275). It suggests that carbon sources may affect the expression of type 3 fimbriae. Thus, analysis of the type 3 fimbriae expression was performed while K. pneumoniae CG43S3 grown in M9 minimal medium supplemented with 0.4% glucose, glycerol, mannose, galactose, arabinose, or lactose.

The bacteria were grown at 37^oC with agitation for 16 h and then subjected to Western blot analysis using MrkA antiserum. However, no apparent difference in MrkA production was noted (Fig. 4.15A).

Both the expression of type 1 and MR/P fimbriae have been reported to be increased during reduced oxygenation (173). The MR/P fimbriae do not express when P. mirabilis is grown in highly aerated broth culture, while increases in broth volume

results activation of the fimbriation (173). It is known that culturing condition with poor aeration results in a reduction of oxygen (173), and good aeration results from constant agitation and a large surface-to-volume ratio of the culture medium to the air;

while poor aeration results from culturing a large volume in a small tube. To determine if oxygen level could affect the expression of type 3 fimbriae, we cultured bacteria in upright (rather than tilted) culture tubes or added a mineral oil overlay, which prevents direct oxygen exchange between broth and air. Overnight K.

pneumoniae CG43S3 culture was 1:100 in volume subcultured into 1-, 3-, or 5-ml

fresh LB broth and then subjected to incubation for another 20 h. As shown in Fig.

4.15B, the type 3 fimbriae expression was relatively low in the 1- and 3-ml cultures.

In upright tube with 5-ml culture, the MrkA amount was obviously increased, and overlay with a mineral oil further increase the MrkA production. Although the oxygen level of the cultures remains to be determined, this result implies that the availability of oxygen affects the expression of type 3 fimbriae.

The reduced oxygenation effect on type 3 fimbriae expression was also analyzed in K. pneumoniae NTUH-K2044, and the CG43S3 mutant lacking fur or mrkI (Fig 4.15C). Compared to CG43S3, the expression level of MrkA in NTUH-K2044 was relatively low, suggesting the two clinical isolates regulate the type 3 fimbriae expression in different manners. The reduced oxygenation also caused a slight

increase of the MrkA production in NTUH-K2044. In the highly aerated culturing condition, either mrkI- or fur- deletion abolished the type 3 fimbriae expression;

however the fur-deletion effect was partially restored upon reducing the aeration (Fig.

4.15C). These results suggest that the availability of oxygen affects type 3 fimbriae expression and also plays a role in the Fur-MrkI regulatory pathway.

4.4. Discussion

The expression of fimbrial genes are usually controlled by their adjacent genes encoding transcriptional regulators, such as PapB/I for P fimbriae (17, 94, 125, 141), FimW/Y/Z for Salmonella type 1 fimbriae (261), and MrpJ for MR/P fimbriae (173, 182). In K. pneumoniae CG43, deletion of mrkI, located downstream to the type 3 fimbrial genes mrkABCDF, resulted in a significant decrease of MrkA production (Fig.

4.3). The promoter-reporter assay also showed that the mrkA promoter activity was decreased by the mrkI-deletion (Fig. 4.5), suggesting that MrkI positively regulates the expression of type 3 fimbriae at transcriptional level.

Interestingly, a slight increase of type 1 fimbriae expression was also found in the ΔmrkI strain, which is assessed by mannose-sensitive yeast agglutination, Western blot analysis against FimA antiserum, and orientation analysis of the fim switch (Wei-Yun, Cheng, unpublished data). In addition, the deletion of mrkA caused an increase of type 1 fimbriae expression by an unknown mechanism (273). Whether MrkI plays a regulatory role in this counter-expression remains to be investigated.

Besides, a Western blot analysis using FimA antiserum revealed that the deletion of fur from K. pneumoniae CG43S3 has no apparent effect on type 1 fimbriae expression.

Sequence analysis of MrkI revealed a LuxR-type transcriptional factor with an N-terminal regulatory domain and a C-terminal DNA-binding domain. Activation of the LuxR type regulator could be achieved by one of four mechanisms: (i) two-component system regulators that activated by phosphorylation on an aspartate residue (28, 202); (ii) regulators which are activated, or in very rare cases repressed, when bound to quorum-sensing molecules such as N-acyl homoserine lactones (235);

(iii) autonomous effecter domain regulators, without a regulatory domain (77); (iiii) Multiple ligand-binding regulators (270). The sequence alignment and point mutation study showed that MrkI may be activated via phosphorylation at its D56 residue for the expression of the type 3 fimbriae (Fig. 4.6). In order to carry out an electrophoresis mobility shift assay (EMSA), the recombinant MrkI fused with MBP-tag has been constructed to resolve the problem of the aggregates resulted from the recombinant MrkI proteins fused with 6xHis-tag or GST-tag. The current data provided by Dr. Ching-Ting Lin (School of Chinese Medicine, China Medical University) indicated that this recombinant protein was able to bind PmrkA only after the addition of the phosphodonor acetyl-phosphate in the reaction mixture. These results suggest that the phosphorylated MrkI directly regulates the expression of type 3 fimbriae. MrkI appeared to be an orphan regulator since no sensor kinase encoding gene could be found in the adjacent region. The MrkI cognate sensor remains to be

identified.

As shown in Fig. 4.2, the RT-PCR analysis revealed that mrkHIJ could be transcribed in a polycistronic mRNA. However, the deletion of mrkI or mrkJ resulted in an opposite effect on MrkA production (Fig. 4.3). This may result from additionally different regulation of mrkJ transcription by another promoter located in the 143-bp intergenic region between mrkI and mrkJ. The deletion of mrkH did not affect the type 3 fimbriae expression (Fig. 4.3); however, the overproduction of MrkH increased the amount of MrkA (Fig. 4.8). This suggested that, in particular growth conditions, MrkH also plays a role in the regulation of expression of type 3 fimbriae. An optimal growth condition for assessing the mrkH-deletion effect remains to be shown. As shown in Fig. 4.7 and Fig. 4.8, the R111 residue, the N-terminus of MrkH, and the C-terminal PilZ domain are all required for the MrkH-activated type 3 fimbriae expression.

The quantitative real-time-PCR (qRT-PCR) analysis indicated that the induced expression of MrkH significantly activated the mRNA level of mrkA (Dr. Ching-Ting Lin, unpublished data), suggesting that MrkH activates the expression at transcriptional level. The PilZ domain proteins have been reported to bind to c-di-GMP and then exert their function by protein-protein interaction (34, 122). In

Xanthomonas, the binding between PilZ domain proteins to an ATPase (PilB) and an

EAL domain protein (FimX) regulate the type IV pilus biosynthesis (115). If protein-protein interactions occur between MrkH, MrkI, and MrkJ for the regulation of type 3 fimbriae expression awaits to be studied.

As shown in Fig. 4.9C and Fig. 4.10, the promoter activity of mrkA and mrkH were decreased by the fur-deletion. To ascertain if Fur could directly interact with the promoter regions of mrkA and mrkH, an EMSA has been performed by Dr.

Ching-Ting Lin. The purified recombinant Fur protein was found to be able to bind to PmrkH but not to PmrkA, suggesting a direct Fur regulation on the mrkHIJ operon.

Besides, the qRT-PCR analysis showed the expression of mrkA, mrkH, and mrkI, but not mrkJ, were reduced by the fur-deletion, and the deletion effects could be complemented by introducing a fur-expressing plasmid into the Δfur strain (Dr.

Ching-Ting Lin, unpublished data). These findings suggest that Fur indirectly regulates the type 3 fimbriae expression via MrkI.

We have previously demonstrated that the CPS biosynthesis of K. pneumoniae CG43S3 is regulated in coordination by multiple regulators RcsB, RcsA, RmpA, RmpA2, and Fur (53, 171, 183-185). Since Fur is also an activator for the expression of type 3 fimbriae, a cross-regulation by Fur on the expression of CPS and type 3

fimbriae could be predicted. To further indentify regulators involved in this cross-regulation, Western blot analysis using MrkA antiserum was performed among the individual strains lacking CPS regulatory genes including rcsB, rcsA, kvhA, kvgA, rmpA, rmpA2, or rpoS. However, only a slight decrease (approximate 43%) of MrkA

production was found in the ΔrcsB strain (Fig. 4.16). Whether RcsB is involved in the regulation of type 3 fimbriae expression remains to be studied. Besides, the CPS biosynthesis was not affected by the mrkI-deletion (data not shown).

Fur has been implicated in iron uptake and metabolism, oxidative stress response, colonization, and virulence in many bacteria (41). Although Fur is predicted to exert similar function in K. pneumoniae, we have shown that, besides iron-uptake systems (185), Fur also participates in the regulation of CPS biosynthesis (53, 185) and type 3 fimbriae expression. Since iron-uptake systems, CPS, and type 3 fimbriae are well-known bacterial virulence factors, our findings suggest that Fur plays an important role in K. pneumoniae pathogenicity.

Iron availability influences the activity of the Fur protein as well as the transcription of fur. Fe²⁺-Fur is an autorepressor, reducing fur expression in response to iron (68, 69, 124, 262). In E. coli, the expression of fur has been demonstrated to be modulated by many regulators including OxyR (309), SoxS (335), CRP (68), RstAB

(152), and ArcA (35). Transcription of fur is also activated in response to an elevated cellular c-di-GMP level (209). Since Fur acts as an activator for type 3 fimbriae in K.

pneumoniae, regulators that modulate the expression of fur may subsequently affect

the type 3 fimbriae expression. However, no obvious effect on the type 3 fimbriae expression was found for the K. pneumoniae strains lacking rstA, rstB, or soxRS.

Artificial manipulation of the cellular c-di-GMP content by the overproduction of GGDEF domain proteins has been reported to strongly stimulated the synthesis of adhesins and biofilm matrix components, whereas overproduction of EAL domain proteins produced the opposite phenotypes (64, 76, 122, 151, 250, 256, 257, 293, 322).

We have also found that the induced-expression of YdeH c-di-GMP cyclase in K.

pneumoniae CG43S3 activated the expression of type 3 fimbriae (Fig. 4.12). The

qRT-PCR analyses also showed that the mRNA level of fur, mrkA, mrkH, mrkI, and mrkJ were significantly increased upon the induced expression of YdeH (Dr.

Ching-Ting Lin, unpublished data). In addition, overexpression of MrkH, the PilZ domain protein, affects the type 3 fimbriae expression. These suggest c-di-GMP is involved in the regulation of type 3 fimbriae expression. Nevertheless, the exact mechanism awaits investigation.

By analyzing the genome sequence of K. pneumoniae NTUH-K2044, ORFs

encoding 11 GGDEF, 10 EAL, and 5 GGDEF-EAL domain proteins were found (Fig.

1.1). To our knowledge, most of these proteins have not been studied in K.

pneumoniae yet, except MrkJ (153), FimK (290), YjcC (170), and BlrP1 (22), which

are EAL domain proteins. A slight increase of MrkA production was found in the ΔmrkJ strain (Fig. 4.3), and the deletion of fimK has been shown to activate the expression of type 1 fimbriae but not type 3 fimbriae (290). The MrkA amount was obviously increased in the K. pneumoniae CG43S3 ΔyjcC strain which is highly susceptible to oxidative stress (Jing-Rou Hwang, unpublished data). BlrP1 is a light-regulated PDE, and its crystal structure complexed with c-di-GMP has been determined (22); however its biological role is unknown. Although the PDE activities of FimK and YjcC have not been determined yet, these EAL domain proteins seem to play differential roles in fimbriae expressions in K. pneumoniae.

Since oxidative stress has been shown to affect the expression of yjcC and fur (41, 170, 309), the type 3 fimbriae expression may also be regulated. This is supported by the study that K. pneumoniae OxyR, a central regulator for oxidative stress response, affected the bacterial colonization (123). As shown in Fig 4.11 and 4.15B, availability of iron and oxygen were found to activate the expression of type 3 fimbriae. To our knowledge, this is the first study to show environmental stimuli for type 3 fimbriae expression. Deletion of fur abolished the MrkA production regardless

of the iron availability (Fig 4.11). However, the fur-deletion effect on MrkA production was restored in an oxygen-limiting condition (Fig. 4.15C). Expression of fur has been shown to be controlled by ArcA (35), the primary regulator for the

transition to anaerobiosis (109). Whether ArcA involves in the regulation of type 3 fimbriae expression remains to be studied. qRT-PCR analysis showed that, under highly aerated culturing condition, the mRNA level of mrkI was dramatically reduced in fur-deletion mutant; however this deletion effect was no longer observed in oxygen-limiting conditions (Dr. Ching-Ting Lin, unpublished data). This result suggested that anaerobic regulator such as ArcA and FNR may activate the expression of mrkI in oxygen-limiting conditions, and the possibility remains to be studied.

In summary, a model depicted as shown in Fig 4.17 is concluded. In K.

pneumoniae CG43, the response regulator MrkI directly activates the expression of

type 3 fimbriae upon phosphorylation of its D56 residue and auto-activates the expression of the mrkHIJ opeorn. Expression of MrkH increases the type 3 fimbriae expression through unknown mechanism, while MrkJ decreases the cellular level of c-di-GMP to repress the expression of type 3 fimbriae. The ferric uptake regulator Fur acts as an activator for the type 3 fimbriae expression through indirect activation of the expression of mrkI. The second messenger c-di-GMP activates the expression of fur, the mrkHIJ operon, and the type 3 fimbrial genes. The expression of type 3

fimbriae is activated in oxygen-limiting conditions, and an unknown regulator may activate the expression of mrkI to modulate the type 3 fimbriae expression during reduced oxygenation.

1 Kb mrkA

mrkF

(4559) (4555)

(4554) (4552)

(4551) (4550)

Type 3 fimbrial gene cluster mrkI

(locus tag) (4567) (4569)

fimB fimE fimAICDFGHK

Type 1 fimbrial gene cluster mrkJ

mrkH mrkDCB

Fig. 4.1. Schematic gene organization of a chromosomal region encoding K.

pneumoniae type 3 and type 1 fimbriae. The designation and the locus tag (KP1_number) of the ORFs are indicated. The three ORFs encoding putative regulatory proteins are shown in black.

mrkD

Fig. 4.2. The transcription units of mrkH, mrkI, and mrkJ defined by RT-PCR.

The genetic organization of the downstream genes of the type 3 fimbrial gene cluster (mrkABCDF, and only mrkD and mrkF are shown as gray arrows) in K. pneumoniae is shown. The designation and the locus tag (KP1_number) of the ORFs are indicated.

The three ORFs encoding putative regulatory proteins are shown in black. The upper panel of the figure shows the DNA sequence upstream of the mrkH gene. The start codon of mrkH and the stop codon of KP4550 are shown in bold. The predicted -10, -35, and transcriptional factor binding boxes are indicated. Primers pcc319 and pcc320 used for the promoter-reporter construct are indicated by vertical arrows. The lower part of the figure shows the RT-PCR results by ethidium bromide–stained agarose gel. Panels a–f show the corresponding PCR products for primers located at ORF or the junction between ORFs. Lanes 1, RT-PCR products; 2, RT-PCR without reverse transcriptase, as a negative control; 3, PCR with genomic DNA as a template, as a positive control. Arrowheads indicate the expected sizes of RT-PCR products.

WT ΔmrkH ΔmrkI ΔmrkJ CCW40 CCW41

MrkA

1 1.19 ND 1.67 ND 0.92

fold

Fig. 4.3. Deletion of mrkI decreases the expressions of type 3 fimbriae. K.

pnuemoniae CG43S3 (WT, wild-type), its isogenic gene-deletion strains (ΔmrkH, ΔmrkI, and ΔmrkJ), and the mrkI-complement strain CCW41 as well as the control strain CCW40 were grown overnight at 37^oC with agitation in LB broth. Bacterial total protein, approximately five micrograms per lane, was separated by SDS-PAGE and then subjected to Western blot analysis using MrkA antiserum. The MrkA protein is indicated by an arrow. The fold change of MrkA amount calculated by ImageJ software is also shown. ND, not determined.

+1 mrkA

Long Universal Primer Nested PCR

Primary PCR

Fig. 4.4. Identification of mrkA transcription start site by 5’-RACE. (A) Electrophoresis of the 5’-RACE PCR products. M, DNA molecular size markers. The templates used in each PCR reaction include the cDNA from K. pneumoniae CG43S3 (Primary PCR) (lane 1), reverse transcription reaction mixture without transcriptase as a negative control (lane 2), or one hundred-fold diluted primary PCR mixture (Nested PCR) (lane 3). The arrows indicate the expected sizes of the PCR products. (B) Schematic representation of the mrkA loci and the 5’-RACE experimental design. The large arrow represents MrkA open reading frame. Relative position of the primers and expected sizes of the products in Primary and Nested PCR are indicated. The mrkA transcriptional start site is marked as +1. The potential -10, -35, ribosomal binding site, and the translational start site are underlined.

placZ15

Fig. 4.5. Deletion of mrkI decreased the transcription of mrkA. The β-galactosidase activities of K. pneumoniae CG43S3ΔlacZ (ΔlacZ) and its isogenic mrkI deletion mutant (ΔlacZ ΔmrkI) carrying each of the reporter plasmids pmrkA-P1, pmrkA-P2, or pmrkA-P3 were determined from log-phased cultures grown in LB broth. The results are shown as average of the triplicate samples. Error bars indicate standard deviations. *, P < 0.0001 compared with ΔlacZ [placZ15].

(A)

Fig. 4.6. MrkI is probably a response regulator activated by phosphorylation. (A) Sequences of MrkI and LuxR-type transcriptional regulators NarL, BvgA, and RcsB were aligned by Vector NTI software. The conserved aspartate (D56) residue of MrkI as a putative target site for phosphorylation is indicated by an arrow. (B) D56 is important for MrkI functionality. K. pneumoniae CG43S3 (WT, wild-type), the ΔmrkI strain, and the mutant strains expressing MrkID56E (D56E) or MrkID56A (D56A) were grown overnight at 37^oC with agitation in LB broth. Bacterial total protein, approximately five micrograms per lane, was separated by SDS-PAGE and then subjected to Western blot analysis using MrkA antiserum. The MrkA protein is indicated by an arrow.

1. K. pneumoniae NTUH-K2044 MrkH 2. P. aeruginosa PAO1 PA3353 3. P. putida kt2440 PP4397 4. E. coli MG1655 YcgR 1. K. pneumoniae NTUH-K2044 MrkH

2. P. aeruginosa PAO1 PA3353 3. P. putida kt2440 PP4397 4. E. coli MG1655 YcgR

Fig. 4.7. Amino acid sequence alignment of PilZ domain proteins. Sequences of the PilZ domain proteins, including MrkH, PA3353, PP4397, and YcgR (27, 258), were aligned by the Vector NTI software. The conserved RxxxR motif and the D/NxSxGG motif are underlined (x, any residue). The critical lysine residue involving in c-di-GMP binding activity of YcgR (258) is indicated by an arrow.

pETQ pMrkH pMrkH* pPilZ pPilZ* pMrkH_N

MrkA

Fig. 4.8. MrkH-mediated activation of type 3 fimbriae expression. K. pneumoniae CG43S3 carrying expression plasmids, as shown in the upper panel, were grown in LB broth at 37^oC with agitation. When the bacterial growth reached mid-log phase, expression of the recombinant protein was induced by addition of 0.5 mM IPTG, and then subject to additional 3 h incubation. Bacterial total protein, approximately five micrograms per lane, was separated by SDS-PAGE and then subjected to Western blot analysis using MrkA antiserum. The MrkA protein is indicated by an arrow. pETQ, the vector only control.

(A) P_mrkA ATATTTGTCGGCGAATAAATAGCATTCTTTGACGCCGATA

β-galactosidase activity (Miller units) 0

β-galactosidase activity (Miller units) 0

Fig. 4.9. Deletion of fur repressed the expression of type 3 fimbriae. (A) The predicted Fur-binding sequences on the promoter regions of mrkA and mrkH. The alignment with the 19-bp Fur box (w = A or T) is shown. (B) Anti-MrkA Western blot analysis of the total protein, approximately five micrograms per lane, isolated from K.

pneumoniae CG43S3 strains. WT, wild-type. M, protein molecular size markers. The MrkA protein is indicated by an arrow. (C) Assessment of mrkA transcription using a promoter-reporter system. The β-galactosidase activities of K. pneumoniae CG43S3ΔlacZ and its isogenic deletion mutants (ΔmrkI and Δfur) respectively carrying the reporter plasmid pmrkA-P2 were determined from log-phased cultures grown in LB broth. The results are shown as average of the triplicate samples. Error bars indicate standard deviations. *, P < 0.0001 compared with ΔlacZ [pmrkA-P2].

β-galactosidase activity (Miller units) 0 20 40 60 80 100 120 140 160

WT Δfur ΔrcsA ΔrcsB ΔmrkH ΔmrkI ΔlacZ (P_mrkH::lacZ)

* *

Fig. 4.10. The promoter activity of the upstream region of mrkH was regulated by Fur and MrkI. The β-galactosidase activities of K. pneumoniae CG43S3 ΔlacZ or its isogenic deletion strains respectively lacking fur, rcsA, rcsB, mrkH, and mrkI

Fig. 4.10. The promoter activity of the upstream region of mrkH was regulated by Fur and MrkI. The β-galactosidase activities of K. pneumoniae CG43S3 ΔlacZ or its isogenic deletion strains respectively lacking fur, rcsA, rcsB, mrkH, and mrkI