Catabolite repression control (Crc) is a post-transcriptional global regulator of carbon metabolism that is widely distributed in many Pseudomonas species (1–4) and in some related bacterial genera such as Acinetobacter (5). Crc has been found responsible for the post-transcriptional levels of numerous genes mostly relating to the catabolic pathways (6), by directly or indirectly affecting their expression (7, 8). In Pseudomonasputida, Crc was found to play an important role in the metabolism growing optimization in complete medium by proteomic and transcriptomic analyses (6). Moreover, Crc can modulate arsenic susceptibility and biofilm formation (9, 10), indicating that it might affect special processes in the opportunistic pathogen Pseudomonas aeruginosa, some of which may play important role in its virulence and toxic resistance. The crc-knockout (Δcrc) strain was defective in many physiological processes in vivo, such as type III secretion, motility, expression of quorum sensing regulated virulence factors, and was less virulent in a Dictyostelium discoideum model (7). Moreover, this strain was more susceptible to β-lactams, aminoglycosides, fosfomycin, and rifampin, because it expresses at higher levels their membrane transporters OprD and GlpT, whose original function is the transport of basic amino acids and glycerol-3-phosphate, respectively (7). Therefore, Crc has been suggested as nonlethal targets useful for developing novel antimicrobials (7, 11).
Crc can inhibit translation initiation by binding the A-rich region, which is next to the ribosomal binding site, leading to the arrest of ribosome and target mRNA binding (8). The regulation of metabolism by Crc is reached by the binding between Crc and an A-rich motif on the 5′-end of the target mRNAs. The activity of Crc is modulated by the small RNA CrcZ in P.aeruginosa, a small 407-nt RNA containing five catabolite activity (CA) motifs that are able to remove Crc from amiE mRNA in vitro, which acts as an antagonist of Crc by binding and titrating it (12). The transcription of CrcZ is activated by a carbon source-sensitive two-component system CbrAB (13), which provides a linkage between this sensor system and the global regulation exerted by Crc (12). In P.putida, Crc can also bind to the 5′-end of the alkS and benR mRNAs (the pathway gene for the alkane and benzoate pathway, respectively), inhibiting their translation in vivo (3, 14). The band-shift assays using variant alkS mRNAs showed that Crc specifically binds to a short unpaired AU-rich sequence motif in the coding sequence of alkS mRNA and prevents the formation of active ribosomal initiation complex and a similar Crc-binding site was localized at benR mRNA, upstream of the Shine-Dalgarno sequence (8).
Unveiling the molecular mechanisms of Crc is important for developing novel antibiotics as a potential target to reduce both resistance and virulence of bacterial pathogens. In this study, we solved high-resolution crystal structure of Crc from P. aeruginosa and performed comprehensive structural analysis and comparison with its homologs. This is the first report on the structure of Crc in Pseudomonas family. The structure can be described as 11 β-sheet inside and 13 helix outside. By aligning to the structure of other families, Crc is classified into a member exodeoxyribonuclease III and the active site is discovered. In structural respect, this study revealed some clues for deeply understanding the mechanism and function of Crc.
Abbreviations Crc, catabolite repression control; PDB, protein data bank; RMSD, root mean square deviation; CD, circular dichroism.
Materials and Methods
Cloning, Protein Expression, and Purification
The full-length of crc gene from P. aeruginosa PAO1 genomic DNA was amplified by PCR using a forward primer containing a BamHI restriction site (GCC GGATCC ATGCGGATCATCAGTGTGA) and a reverse primer containing an XhoI restriction site (GCC CTCGAG TCAGATGCTCAACTGCCAG). The amplified insert of crc gene was cloned into the expression vector pET28at_plus vector with a cleavable N-terminal His6 tag (constructed by our lab) and transformed into E. coli DH5α cloning strain and plated onto Luria-Bertani (LB) kanamycin plates. The plasmid was isolated and transformed into an E. coli BL21 DE3 star expression strain (Invitrogen, USA). The protein was purified as previously described (15). Prior to crystallization, Crc was concentrated to 18 mg mL−1.
Protein Crystallization, Data Collection, Structure Determination, and Refinement
Initial crystallization conditions of His–Crc were obtained through utilization of several sparse matrix screens (Hampton Research, USA) with the sitting drop vapor diffusion method at room temperature using the sample mixture and well buffer mixed at a 1:1 volume ratio. The crystal quality was optimized by adjusting the concentration of the precipitant and buffer. The best crystal was obtained in solution 0.1 M HEPES (pH 7.5), 10% (w/v) polyethylene glycol (PEG) 6000, and 5% (v/v) (+/−)-2-methyl-2,4-pentanediol after 3–4 days.
Before data collection, crystals were soaked for 10 Sec in a cryoprotectant consisting of 20% PEG 400 in the crystal mother liquor and then flash-frozen in liquid nitrogen. Diffraction data were collected on the beamline station BL17U of Shanghai Synchrotron Radiation Facility (SSRF). The temperature was held at 100 K by liquid nitrogen during data collection. Data were processed with the program HKL2000 (16). Data statistics are summarized in Table 1.
Table 1. Data collection and refinement statistics
The values in parenthesis means those for the highest resolution shell.
The initial phases were calculated using the program PHASER (17). The phases from PHASER and the structure factors from HKL2000 were combined in automated refinement procedure (ARP)/wARP (18) and 271 of 313 residues were traced. Some biases were reduced manually in COOT (19). The program Phenix.refine (20) was used to refine the structure and append the water molecules. PROCHECK (21) was used to validate the structure. The final refinement statistics was listed in Table 1. The program PyMOL (http://www.pymol.sourceforge.net/) was used to prepare structural figures.
Circular Dichroism Spectroscopy
Purified Crc protein was loaded onto a Superdex 75 column (GE Healthcare) equilibrated with phosphate-buffered saline buffer (pH 8.0). The elution containing Crc was subsequently concentrated to 1.65 mg mL−1. The circular dichroism (CD) spectra of Crc were measured on the 4B8 station of Beijing Synchrotron Radiation Facility at 1 nm bandwidth with a 1 nm step resolution from 180 to 260 nm. The data were averaged over eight accumulations. The thermal denaturation curve of Crc was measured by monitoring the change in absorption from 250 to 180 nm at increasing temperatures (35–75 °C) using a temperature controller. The temperature increased at a step of 10 °C. The data are processed by CDtool (22) and further analyzed by Dichr web (http://dichroweb.cryst.bbk.ac.uk/html/home.shtml).
Protein Structure Accession Number
The atomic coordinates of Crc have been deposited in the Protein Data Bank (PDB) under accession code 4F1R.
Overall Structure of Crc and Its Comparison with the Eukaryotic and Prokaryotic Homologs
The structure of Crc from P. aeruginosa was solved by molecular replacement using the predicted DNA uridine endonuclease (PDB ID: 2JC5) from Neisseria meningitidis (23), as the searching model and was refined to a final R/Rfree factor of 0.20/0.25 at 2.20 Å. The asymmetric unit contains one Crc molecule with overall dimensions of ∼74 Å × 74 Å × 124 Å comprising 259 residues and a total of 23 water molecules. A Ramachandran plot calculated for the final model showed that 97.3% of the amino acid residues is in most favored regions, 2.7% is in additional allowed regions, and no residues are in disallowed regions (Table 1).
Crc is a globular α/β protein and is made up of two similar overall topology halves, comprising five/six-stranded β-sheet surrounded by α-helices (Figs. 1A and 1B). One-half comprises the residues 1–74 and 231–259, whereas the amino acids 75–230 constitute the other half. A four-layered α/β-sandwich is formed together. One β-sheet is composed of the strands β1–β4, β10, and β11, whereas the other β-sheet is formed by β5–β9 with the arrangement β3↓–β4↑–β2↓–β1↓–β11↑–β10↓ and β5↓–β6↑–β7↓–β8↓–β9↑, respectively. Seven β-strands are parallel with each other including β1, β2, β3, β5, β7, β8, and β10, and antiparallel with β4, β6, β9, and β11. The five- and six-stranded β-sheets of the two halves are flanked by two longer α-helices on one side while other shorter helices surrounded.
A structural similarity search of Crc was performed using the DALI web server (http://ekhidna.biocenter.helsinki./dali_server). Although they shared relatively low sequence identities (22.0–36.5%), Crc showed high similarity with its eukaryotic and protokaryotic homologs, including apurinic/apyrimidinic (AP) endonucleases from Neisseria meningitidis (nApe), zebrafish (zApe) and human (hApe1), DNA uridine endonuclease Mth212 from Methanothermobacter thermautotrophicus, and the exonucleases from Archaeoglobus fulgidus (Af_Exo) and E. coli (Ec_Exo III) (Table 2). A clustal X (version 1.81) structure-based sequence alignment for Crc and these homologs above is shown in Fig. 2. Here, identical residues are highlighted. Only hApe and zAPe own one long insertion among members located in the N terminator, consisting of about 60 amino acids. It is noteworthy that the identical conserved residues are nearly located in the two five/six-stranded β strands and the loops in the forward surface formed by the two halves (Fig. 2).
Table 2. Structure and sequence comparisons of Crc with several eukaryotic and protokaryotic nucleases whose structures are currently available
Sequence identity (%)
The number of Cα atoms used in the calculation is indicated in parentheses.
To study the secondary structure of Crc in solution, CD spectroscopy was applied to probe its solution conformation at 25 °C. The CD spectrum of His-tagged Crc reveals a canonical curve with maximum at 222 nm and minima at 195 nm (Supporting Information Fig. S1A), suggesting that Crc is composed of α-helices and β-strands with nearly the same proportion in crystal. The normalized value for ellipticity at 195 nm is 10.55 × 103 Deg cm2 dmol−1 residue−1 and 222 nm is −8.67 × 103 Deg cm2 dmol−1 residue−1, which confirms calculated 27% helices and 23% strands. To evaluate the thermostability of Crc in solution, we monitored the CD signal from 250 to 180 nm in a thermal changing experiment. The thermal changing curves of Crc suggest that Crc maintains a thermostable fold in solution (Supporting Information Fig. S1B). Below 45 °C, there are little changes in absorption from 250 to 180 nm. But at the temperature of 55 °C, there is a sharp absorption lost at 195 and 222 nm indicating partial secondary structure was destroyed by thermal changes. Crc is stable below 45 °C but becomes unstable since 55 °C.
Potential Active Pocket of Crc
Comparison of Crc with hApe1-DNA and Mth212-DNA complexes showed they shared both same folding pattern of both halves and similar grooves formed by the two halves (Figs. 3A and 3B). The active site of hApe1 and Mth212 is located at the interface of the two halves and involves conserved residues in the shadow of several loops and β-strand (Fig. 3C).
Structural arrangement of the conserved residues in the canonical catalytic site of hApe1 (Asp70, Glu96, Asn212, Asp283, Asp308, and His309) and Mth212 (Asn12, Glu38, Asn153, Asp222, Asp247, and His248) is highly similar with that in Crc, which comprises Asn9, Asp35, Tyr150, Asp220, Gln245, and His246 (Figs. 2 and 3C). These highly conserved residues can form similar positive pocket in the potential active site of Crc for RNA binding. Moreover, the residues Asp70, Glu96, Tyr171, and His309 in hApe1 and Asn12, Glu38 and Asp247 in Mth212, which constitute the Mg2+ and Mn2+ binding pocket, respectively, correspond to the residues Asn9, Asp35, Tyr150, and His246 in Crc, which are also likely involved in metal binding. The bound Mg2+ or Mn2+ has been suggested to help orient the phosphate group in the active site and to polarize the scissile P-O3′ bond as well as to stabilize the transition state (24–26).Therefore, the divalent metal ions (such as Mg2+ and Mn2+) may play similar role in Crc–RNA interaction, although no metal ions can be observed in the present crystal structure.
Implication for RNA-Binding Region of Crc
In hAPE1–DNA complex, APE1 stabilization of the kinked abasic DNA is mediated by residues emanating from four loops and one α-helix while there were three DNA-binding loops in Mth212–DNA complex (Figs. 3A and 3B). Similarly, there are five loops: L1 (β7–α), L2 (β7–η4), L3 (η6–η7), L4 (η7–β7), and L5 (β11–β11), and one helix η7 at the interface of the two halves of Crc (Figs. 1A and 2), where most residues are conserved or identical conserved. The residues, such as Tyr150 from L2, Asp220 from L4, Gln245 and His246 from L5, are involved in formation of the active site as mentioned above.
In Crc structure, these flexible loops constitute the positively charged protuberances and especially L5 form part of a positive groove, which are most likely to mediate recognition and binding of negative substrate RNA (Fig. 4C). The hAPE1–DNA complex structure showed that APE1 uses a rigid, preformed, positively charged surface to kink the DNA helix and engulf the AP-DNA strand (Figs. 3A and 4A). hAPE1 inserts loops into both the DNA major and minor grooves and binds a flipped-out AP site in a pocket that excludes DNA bases and racemized β-anomer AP sites (26). In wild-type Mth212–DNA complex structure, one Mth212 molecule binds both of one dsDNA oligonucleotide's ends (Figs. 3B and 4B). Both Mth212 molecules show the same DNA binding surface area with each DNA blunt end located at the active site. The side chain of Arg209 inserts into the DNA to form base stack interactions, which is similar to hApe1 (24). Moreover, remarkable conformational changes were found in hApe1 or Mth212 after bound to DNA: a shift of up to 5 Å occurred in Loop 1 into the 3′- or 5′-direction, respectively, which allowed additional contacts to the DNA backbone, and the binding to the DNA major groove was enhanced (24, 26). Considering their significant similarity, conformational changes in these loops of Crc are also required for efficient RNA binding.
As the active pocket in Crc is very similar to those in hAPE1 and Mth212, we can reasonably assume they share similar substrate-binding mode and an interesting problem that what caused their distinct substrate specificity (RNA vs. dsDNA binding) was raised. Close inspection of their electrostatic potential mapped onto the surface revealed the obvious difference of the surface charge distribution, especially in substrate-binding regions. There are the positively charged protuberances surrounding DNA-binding region on the hAPE1 and Mth212 surface, involved in forming large catalytic pocket, where the substrate dsDNA can be bound efficiently (Figs. 4A and 4B). However, there is only a small positive patch on the potential RNA-binding protuberance of Crc, which contains the conserved active pocket surrounded by negative amino acids (Fig. 4C). In this context, the smaller positive groove may only attract single-strand RNA, especially the A-rich motifs from CrcZ or the target mRNAs, and cannot accommodate dsDNA with larger negative charges. Besides, different conformational changes in these loops of Crc and the two homologs when bound to their respective substrates may also lead to the difference in substrate specificity.
Recently, extensive biochemical studies have shown Crc a RNA-binding protein with strong affinities to the A-rich region in vitro (8, 12), meanwhile, its mechanism of action in signal transduction pathway has not been well defined. In this study, based on the significant similarity of the structures and key residues in active sites between Crc and several known endonucleases, Crc is reasonably classified into a member exodeoxyribonuclease III, although its preferred substrate is RNA but not DNA. However, its role as a global regulator that controls the metabolism of carbon sources and catabolite repression may be more complicated in signal transduction, not being limited to its nuclease activity.
In P. putida, Crc is proposed to impede the formation of 30S-tRNAfMet-RNA ternary complex at the alkS mRNA by a direct steric hindrance (8). The AUG codon and the 3′-end of the mRNA could not be positioned correctly into the ribosomal track normally by the prevention of Crc, leading the codon–anticodon could not be formed. Conversely, the initial binding between ribosome and the BenR mRNA is prevented by the inhibited binding of ribosomal protein S1 by Crc competition (14). In P. aeruginosa, CrcZ and Crc were suggested as being not the only catabolite responsive elements in amidase regulation, as the existence of a second Crc-like protein may exist for the global function of the Cbr/Crc signal transduction pathway (12). Our structure showed Crc possesses a long positively charged groove for target RNA binding. Moreover, Crc is likely operating in different ways due to the difference of target mRNA structures and the binding site of Crc next to the start codon AUG. Based on its high-resolution crystal structure and the potential active site, combined with its important role in affecting the antibiotics resistance and toxicity of P. aeruginosa, the rational design of the inhibitors of Crc will be performed in the future to screen the drugs with antivirulence and antiresistance activities.
The authors are grateful to the staff of the beamline station BL17U of SSRF for providing technical support and for many fruitful discussions. This work was supported by the grants from the National Natural Science Foundation of China (10979005 and 31200552) and the National Basic Research Program of China (2009CB918600 and 2012CB917203).