Cytophysiologic effects and molecular inhibition of a functional actin specific ADP ribosyltransferase CDT from clostridium difficile 3

Chapter Molecular Characterization of a binary cdt toxin genes from a variant Clostridium difficile strain with truncated pathogenicity locus 3.1. Introduction Several approaches have been used to identify virulence proteins in recent years. For example, random transposon mutagenesis was useful through creation of Pseudomonas aeruginosa mutants which were screened for virulence reduction (Rahme, Tan et al. 1997; Jander, Rahme et al. 2000). Gaining increasing application is nucleotide microarray chip technology (Lan and Reeves 2000) which allows identification of differential genes between pathogenically diverse organisms. However, availability of genetic sequences is necessary for inclusion into the array. More recently, phenotype microarray has allowed comparison of differential proteome expression amongst six strains (Bochner, Gadzinski et al. 2001). As knowledge in phenotypic variations has become popular in understanding disease processes and formulation of directed defense, detection of virulence genes by phenotype analysis between wild-type and isogenic knock-out strains for example, has become promising. Again, differences in protein products is dictated by gene diversity and this highlights the importance on our continued discovery of genetic sequences. Recently, genomic subtraction (GS) between virulent P. aeruginosa pathogen PA14 and avirulent PA01 revealed Yersinia pestis ybtQ virulence homolog in P. aeruginosa using G. mellonella and burned mouse model (Sawada, Kokeguchi et al. 1999; Choi, Sifri et al. 2002). Moreso, Sawada et al. (1999) detected insertion sequence IS1598 involved in necrotic abscess formation among virulent Porphyromonas gingivalis strains after failed attempts to identify virulence determinants from strain differences using biochemical means (Neiders, Chen et al. 1989). Use of this technique has likewise resulted in rapid isolation of gene islands and pathogenes between closely related organisms (Klee, Nassif et al. 2000; Choi, Sifri et al. 2002). In this study, we have adopted a similar approach and have localized 19126-specific 62 virulence gene portions which to our knowledge is the first report of genomic subtraction administered between two C. difficile strains. Putative pathogenes were concentrated as genomic library probe through elimination of closely homologous DNA from two sets of genome that have hybridized. The process has led to the identification of several virulence genes and characterization of variant forms of cdt whose detection was reported at about 6% of C. difficile clinical isolates (Stubbs, Rupnik et al. 2000; Geric, Rupnik et al. 2004; Goncalves, Decre et al. 2004). This has illustrated the efficiency of the technique in enriching specific genomic subset involved in pathogenicity and those encoding unknown or hypothetical proteins with possible novel identity. Furthermore, we have explored cdt gene expression since knowledge on this aspect is lacking, unlike the PaLoc toxin genes whose expression has been well-characterized to follow both mono- and polycistronic transcription with higher expression of downstream mRNA (tcdA>tcdB mRNA). Finally, we have studied the functional role of several conserved amino acid residues in CDTa confirming identity of the genes isolated. 3.2. Results 3.2.1. Isolation of putative 19126-specific virulence DNA Initially, we performed genomic subtraction between pathogenic (19126) and nonpathogenic (11186) C. difficile strains to derive virulence gene fragments. Identity of DNA source strains were first ascertained by detecting a portion of toxin B gene, tcdB using colony PCR (see Table 2.2 for primers used). Results showed the presence of 1362 bp segment in genomes of known C. difficile pathogens ATCC 43596 and 19126 which was not amplified from 11186 (Fig. 3.1). Using enzyme immunoassay, toxin A was produced by 19126 (OD450=0.589) and 20309 (0.446) which are higher than the >0.200 cut-off for positive result but not 11186 (0.036). These indicated the presence of PaLoc-encoded toxins in 19126 and absence in 11186. Enrichment of 19126 DNA was achieved by allowing its reassociation with excess of sheared, biotinylated subtractor DNA from 11186 (Fig. 2.1). The streptavidin-bound biotinylated 63 kb 1.6 M 1.3 Figure 3.1. Characterization of C. difficile reference strains for the presence of tcdB gene portion using colony PCR. The PaLoc gene was amplified from chromosomal DNA of ATCC 43596 (lane1), CCUG 19126 (lane 2) and VPI 11186 (lane 3). M, kb plus DNA ladder (Gibco BRL). DNA species (single-stranded, homoduplexes and heteroduplexes) were removed by organic phase extraction with streptavidin while unbound DNA subjected to more rounds of subtraction cycle. Rounds 1-4 extracts contained amplicons of wide-ranged sizes whereas fifth round products were limited to 100 to 300 bp, suggesting more complete range of target DNA template until the 4th round. Using colony and dot blot hybridization, pathogenic ATCC43596 and 48 out of 292 library clones reacted with the probe but not 11186 and E. coli containing pUC18, SK1200 (Table 2.1)(Fig. 3.2A,B,C,E). Accordingly, clinical isolate CD108, screened as non-PaLoc containing was not recognized by the probe (Fig. 3.2E). These indicate efficient enrichment through successive subtractive cycles and have shown probe identity to PaLoc toxigenic elements and other putative pathogenes. 3.2.2. Identification of insert fragments with putative virulence function Nineteen representative plasmids with inserts ranging from 100 bp to kb fragments were sequenced (GenBank accession no. CC927338-CC927348) and submitted to NCBI BLAST programs for homology search (Table 3.1). Majority of clone inserts at 42% showed identity to 64 Figure 3.2. Colony hybridization showing autoradiogram of CCUG 19126 genomic library clones detected by round subtraction product. A-C, discs blotted with 292 colonies starting from slot A1 up to C93. For all discs, slots 101-104 contained the following colonies: SK1200-JM109 carrying pUC18 (negative control), ATCC 43596 (Positive control), CCUG 19126 (test), and VPI 11186 (test) shown by arrows on disc A. Disc A had colonies 1-100 exclusively, disc B with colonies 101-200, and disc C had colonies 201-292. D, template grid used for colony blotting. E, dot blot of C. difficile genome probed with round subtraction product: 1-SK1200, 2-CD108, 3-CCUG 8884, 4-VPI 11186, 5-CCUG 4938, 6-ATCC 43596, 7-CCUG 19126. 65 Table 3.1. Protein similarities of CCUG 19126 library inserts Length of fragment (bp) Predicted protein homologies Clone Organism E- value Percent identities GenBank access. no. NP779526 GS05 382 Phage-related protein Xylella fastidiosa Temecula1 2.E-01 68% GS10 129 Toxin B Clostridium difficile 1.E-23 100% AF217292 GS41 348 Toxin A locus CDTOXA, X5179, AA1-142 Clostridium difficile 2.E-10 100% CAA36093 TcdE locus CDI011301, AJ011301 Clostridium difficile 5.E-09 90% CAC19892 GS65 560 Hypothetical protein Deinococcus radiodurans 5.E-09 41% D75542 GS68 725 Carbamoyl-phosphate synthetase subunit Clostridium perfringens 6.E-04 91% NP563488 GS80 573 CDT binding component Clostridium difficile 3.E-92 98% AAB67305 Iota toxin component Ib Clostridium perfringens 5.E-75 75% CAA51960 GS101 230 SocE-csgA suppressor Myxococcus xanthus 5.E-05 53% AAF91388 GS104 398 Toxin B Clostridium difficile 1.E-59 100% AF217292 GS110 127 Tox protein DT-201 Corynebacterium diphtheriae 2.E-06 100% AAA72620 GS128 187 S-layer protein Clostridium difficile 1.E-01 75% CAC35720 GS157 303 Alpha-hemolysin Aeromonas hydrophila 3.E-04 65% AAB81227 GS159 293 Hypothetical protein Clostridium perfringens 2.E-04 68% Q8XM08 GS166 194 HD superfamily hydrolase, HD-GYP domain Clostridium acetobutylicum 2.E-08 92% NP347489 GS194 635 Unknown Pasteurella multocida 1.E-01 76% NP245838 GS201 165 Catalase Agrobacterium tumefaciens 3.E-01 94% NP535120 GS213 537 Hypothetical protein Bacillus megaterium 4.E-08 65% NP799510 GS237 142 Cat-2 catalase Zea mays 5.E-08 93% S71455 GS241 164 DnaK heat shock protein Clostridium acetobutylicum 5.E-02 56% NP347113 GS272 288 Hypothetical protein Escherichia coli 4.E-02 42% NP308728 bacterial virulence homologs with clones GS10 and GS104 containing portions of tcdB covering an average 112 amino acid residues, while GS41 carries portions of tcdA at amino acids 112-142 and tcdE at amino acids 135-165. On the other hand, 32% matched with unknown, hypothetical or phage-associated factors and 26% with housekeeping proteins. Although the proportion of identified DNA here is small relative to complete genome sequences, similarity in categorical identity are reflective of those in many genome projects like in Clostridium perfringens and E.coli K-12 where 38% of ORFs had homology to hypothetical or unclassified proteins and 87.8% to known factors (Blattner, Plunkett et al. 1997; Shimizu, Ohtani et al. 2002). Detection of several virulence-encoding fragments is expected as our library was probed with nucleotides which have been potentially rid of strain-specific complementary duplexes that are likely maintenance genes. 66 3.2.3. CCUG 19126 and CCUG 20309 encodes variant forms of cdt To further support applicability of genomic subtraction, we validated the identity of GS80 insert by attempting to capture and functionally characterize the complete cdt in 19126 and other C. difficile strains. The clone has a 573 bp insert of 75% identity to iota toxin component Ib of C. perfringens (nt 638-786) and 98% to C. difficile CD196 ADP-ribosyltransferase binding component, CDTb with extensive coverage of 148 amino acid residues (nt 639-787) (Perelle, Gibert et al. 1997a). Based on CD196 nucleotide sequence, primer pairs were designed to detect cdt from various C. difficile strains. Our preliminary survey of toxin A-producing hospital isolates yielded 13 toxinotypes of PaLoc and cdt gene variants with none of the complete cdt. Strain 19126 encodes a 1,282 bp truncated cdt (GenBank accession no. AY341253). Sequence identity encompass 533 bases downstream of cdtA start site and 3’end of cdtB punctuated with a large block deletion of 1,958 bp (Fig. 3.3). The incomplete orf was sequenced from pDA579 in clone SK1222 (Table 2.1). Among the reference strains tested including ATCC 43596, nonpathogenic VPI 11186 and VPI 8884, only CCUG 20309 contained the full cdtA, cdtB and binary genes of 0.9, 1.8 and 3.2 kb amplicon sizes, respectively (Fig. 3.3). cdtA is 1,392 bp long encoding a 463-amino acid protein (53 kDa, pl of 8.81), whereas cdtB has 2,631 bp encoding a polypeptide of 876 amino acid residues (99 kDa, pl of 4.74). The higher prevalence of C. difficile with truncated cdt (40%) over the complete cdt toxinotype is reflected on our survey. In comparison to CD196, additional nucleotides (ACCAGAAGA) were located 165 bp downstream of cdtA translational start site. This resulted in the replacement of Ser55 by Arg55, Pro56, Glu57 and Asp58 resulting in conservative deduced amino acid substitutions. Immediately upstream lies the putative cleavage site (Lys42-Val43) that is essential for the release of proposed cdtA N-terminal transmembrane signal peptide (Klein, Kanehisha et al. 1985; Perelle, Gibert et al. 1997a). A similar cleavage site was found in cdtB (Lys42-Glu43). 67 Figure 3.3. Comparative genetic map of CCUG 19126 and CCUG 20309 cdt genes. Double arrow heads indicate amplicon position and sizes with corresponding bands on 1% agarose gel (M, DNA ladder; lanes 1-3, from template 20309; lanes 4, and 6, from templates VPI 10463, 11186 and 19126, respectively. Single arrow head points to the direction and location of primers (see Table 2.2). Block circles show the location of homologous direct repeats (underlined) in 20309 cdtA (TATACAAAACAAATTATTTAA), 20309 cdtB (ACTACAAATTATTCCCATACA), and 19126 truncated cdt (TATACAAGACAAATTATTACCATACA). Dashed lines show the extent of cdt deletion (not drawn to scale). 68 3.2.4. Analysis of cdt regulatory region Primer extension generated a first strand cDNA product (52 bp) which terminated at the 5’ transcription initiation site (TSS) that was mapped to an adenine residue at nt. -24, that is 25 bp and 14 bp upstream of start codon and RBS, respectively (Fig. 3.4)(Angeles, Leong et al. 2004). The 542 bp sequence upstream of cdtA ATG start site (GenBank accession no. AY029209) showed several features including a ribosomal binding site (RBS) located bp upstream of start site (Fig. 3.4). An RBS was also found nucleotides upstream of cdtB that is conserved in iota Ib. Inverted repeats of 11 bp and bp in length were also located 47 bp and 325 bp respectively, upstream of start site. Two putative promoter regions were detected with one -10 region located 34 bp upstream of start site, separated from the -35 region by 15 nucleotides while the other at 128 bp upstream of start site has -10 and -35 regions with 18 intergenic spaces (Fig. 3.4). The – 10 consensus promoter sequence at nt –33 to –38 (TTCAAG) was located bases whereas the – 35 site at nt –54 to –59 (TATAAT) is 32 bases upstream of TSS (Fig. 3.4)(Table 3.2). The –45 AT-rich region upstream of promoter which is conserved in Gram-positive bacteria was also identified at nt. –69 to –80. This 0.58 kb regulatory region was also detected in 19126 while the 0.9 kb cdtB downstream region containing inverted repeats at nt 161-173 and 186-198 was not present. Comparison with truncated cdt revealed a single copy of 10 bp direct repeat (ACAAATTATT) in place of deleted block also found flanking the deletion region in 20309 cdt (Fig. 3.3). Such intergenic repeat sequences may represent insertion or deletion site remnants of transposable DNA elements mediating mutation through gene transfer or recombination. Manifestations exist in 19126 cdt as intermittent deletion, base substitution and insertion that resulted in premature termination (TGA) at the 69th codon. 3.2.5. Growth dependent transcription of cdt 69 GAACCATCTCTTTTTTTATACAAAAAAAGTAGTTCCTAAGAAT -310 CCTCTATA TCTCTTTAAAATATT -160 CAGTTGTTATTTTGTACTGACATATCATATAAATACATATTTT -117 -35 -10 TATGATATATAGTTACATATTTTATGAAATTTATATAAAAAAT -74 TCTTATTTAGATTATATAATCTAAATAAATTAAAGTTCAAGAG -31 -35 -10 Start cdtA +1 TTAATTAAACTAATATTGGGAGGGAGAATAAATGAAAAAATTT 12 AGGAAACAT TGATGCAACATTGA 1383 Stop TACCTTAA tattttttcacataaataatttaatatttttcaa Start cdtB atttaaggAGGAGAaaca ATGAAAATACAAATGAGGAATAAA 24 Figure 3.4. Characteristic features of 20309 cdt regulatory sequences. The initiation (ATG) and termination (TAA) codons are labeled. Putative RBS sequences are shown in bold, putative promoters are italicized and underlined and the transcription initiation site is italicized and labeled (+1). Arrows indicate the direction and position of inverted repeats. Intergenic sequence between cdtA and cdtB are in lower case and flanked with spaces. Numerical designation on the right indicates sequence position of the last nucleotide in the line. 70 Table 3.2. Comparison of clostridial promoter sequences with bacterial consensus DNA Intergene space Bacterial source Gene -35 -10 (bp) Reference Gram positive E. coli Clostridia spp. C. difficile cdtA C. difficile tcdA C. difficile tcdB C. difficile tcdD C. difficile C. perfringens C. botulinum tcdE Ia botA Ttgaca* Ttgaca* Ttgaca* TTGTTA TATAAT TTAACA TTTACA TTTACA TTAGCA TTTACA TATGTC TTTACA TGCACA TTGTCAT TTAACC TATAAT TATAAT TAtAAT* TATAAA TTCAAG TTATCT CTCCTT GTCTTT TATAGT TTATTC TATTTT TTATTG TCTAAT TATAAT TATGTT 17 18 17 18 15 20 17 17 17 21 14 20 20 17 18 Graves & Rabinowitz, 1986 Hawley & McClure, 1983 Young et al., 1989 This study This study Sauerborn et al., 1990 Dupuy & Sonenshein, 1998 Dupuy & Sonenshein, 1998 Song & Faust, 1998 von Eichel-Streiber et al., 1992 Hammond et al., 1997 Hundsberger et al., 1997 Sauerborn et al., 1990 Perelle et al., 1993 Binz et al., 1990 *lower case letters represent less conserved sequences To characterize transcription pattern of cdt, gene expression of the complete and truncated form were initially compared using RT-PCR. At OD600, C. difficile growth phases followed a typical sigmoid curve with the early log phase observed between to 10 h proceeding to peak at 19 h (Fig. 3.5A). Preliminary control assays proved experimental validity by showing no amplicon when total RNA template was digested or no first-strand synthesis was performed while similar treatments without DNAse I digestion generated PCR products (Fig. 3.5B). Temporal expression at growth points revealed that cdtA is transcribed from the exponential phase between the 8th and 12th h, which waned starting from stationary phase (Fig. 3.5C,D). Similar expression pattern was observed for cdtB and higher transcription of the truncated form which seemingly persisted up to the 24th h (Fig. 3.5E,F). In addition, primer pairs flanking both cdtA and cdtB regions of 20309 produced a 690 bp amplicon at the exponential phase (Fig. 3.5G). Taken together, such synchronous transcription suggests possible expression of cdt locus as a bicistronic operon controlled by similar if not identical regulatory elements. 71 The geometric MFIs of variably treated and stained cells were compared. CDT-treated double-stained cells registered 29.6% increase in green (from 121.91+ 16.2 to 157.94+ 8.3) and 24.9% decrease in red signal (from 116.38+ 3.6 to 93.14+ 5.9), a trend reflected in single stained cells with 45.4% increase in green (from 102.98+ 4.4 to 149.78+ 3.9) and 63.5% reduction in red signal (from 187.79+ 12.8 to 114.83+ 15.1) (Fig. 4.4A,B). The G:F-actin ratio ~1 for untreated cells increased 2-fold after treatment. All comparisons were significant at p[...]... concentrations Lanes 1 -3, 4-6, 7-9, 10-12, 13- 15 and 16-18 contained wild-type CDTa, CDTaY344N, CDTaY344P, CDTaR345P, CDTaS388H and CDTaE 43 0A, respectively at 2, 6 and 12 nCi concentrations of [32 P]NAD C NAD photoaffinity labeling of CDTa from SK1214, CDTa from SK12 03, CDTaY344N, CDTaY344P, CDTaR345P, CDTaS388H and CDTaE 43 0A (lanes 1-7) M, BenchMark protein ladder (Invitrogen) 81 3. 3 Discussion Global increase... profile in 12% SDS-PAGE (left) and phosphorscreen autoradiogram showing ARTase reactions of CDTa variants Lane 1, contained reaction mixture with purified wild-type CDTa from SK1214; lane 2 with SK12 03; lane 3, with no CDT; lanes 4-8, with CDTaY344N, CDTaY344P, CDTaR345P, CDTaS388H and CDTaE 43 0A B, protein profile (upper panel) and autoradiogram (lower panel) showing ARTase reactions at increasing NAD... containing CDTaY344P showed gradual increase in ADP- ribose labeling of actin as the wild-type (Fig 3. 10B) UV Photolabeling of CDTa with [32 P]NAD was performed to further explore mechanistic basis for the attenuation or inhibition of ARTase activities by variant CDTa Binding of NAD to CDTaR345P, CDTaS388H and CDTaE 43 0A was beyond the detection limit suggesting that inhibition in ARTase activities is largely... SDS-PAGE Lanes 3- 7 had mutant CDTa expressed from mutated pDA577 including pDA587, pDA594, pDA582, pDA586 and pDA589, respectively (see Table 2.1) M, high range protein molecular weight standards (Gibco BRL) 79 SK1214 (wild-type) 1018 AAT TTA ACT GTA TAT AGA AGA TCT GCT CCT 34 0 N L T V Y R R S A P SK1 238 (Y344N) 1018 AAT TTA ACT GTA AAT AGA AGA TCT GCT CCT R R S A P 34 0 N L T V N SK1252 (Y344P) 1018 AAT... 1018 AAT TTA ACT GTA CCT AGA AGA TCT GCT CCT R R S A P 34 0 N L T V P SK1228 (R345P) 1018 AAT TTA ACT GTA TAT CCA AGA TCT GCT CCT R S A P 34 0 N L T V Y P SK1214 (wild-type) 1147 TAT CCA AAC TTT ATT AGT ACT AGT ATT GGT 38 3 Y P N F I S T S I G SK1 236 (S388H) 1147 TAT CCA AAC TTT ATT CAT ACT AGT ATT GGT H T S I G 38 3 Y P N F I SK1214 (wild-type) 1270 GGT TAT GCA GGT GAA TAT GAA GTG CTT TTA 424 G Y A G E Y... illustrated in sequence electropherograms (Fig 3. 9) 3. 2.8 Conserved ADP- ribosyltransferase residues are essential for enzymatic function We then investigated the ability of CDTa and its variant forms to mediate direct hydrolysis of [32 P]NAD and attachment thereafter of radiolabeled ADP- ribose moiety to muscle G -actin in an in vitro ADP- ribosylation assay Modified toxins include CDTaY344N and CDTaY344P... changes in expression pattern in vivo where influences of complex environmental and host factors are involved Binary CDT and constructed mutant forms of CDTa were then expressed for functional characterization Using in vitro biochemical assay, the actin- specific ADPRT action of CDTa was demonstrated Identified were important residues for ADP- ribosyltransferase activity namely Arg345, Ser388 and Glu 430 ... Tyr344 was replaced with asparagine and proline, respectively; CDTaR345P whose Arg345 was replaced with proline, CDTaS388H whose Ser388 was replaced with histidine and CDTaE 43 0A which had a glutamic acid to alanine substitution in residue 430 (Fig 3. 9) Not observed in control and other test lanes, multiple trial phosphorscreen images for wild-type and CDTaY344Ptreated lanes, consistently showed a single... toxin Arg349 which increased the affinity to actin substrate through binding to phosphates and adenine ring of the NAD (Han, Craig et al 1999) In the crystal structure of diphtheria toxin, 2 NAD phosphates bind Arg349 and Arg400 in the catalytic domain (Bell and Eisenberg, 1996) Substitution of equivalent Arg349 to lysine or alanine in iota Ia and C2 exotoxins lead to decrease in ADP- ribosylation and. .. ADP- ribosyltransferase assay on mammalian actin from equimolar total cellular lysate protein nv, HCT 116 treated with novobiocin inhibitor-CDTa mixture; sk, treated with SK12 03 eluate; m, protein ladder; hc, HCT 116 treated with CDTa; he, HepG2 treated with CDTa; and n1, N1E-115 treated with CDTa 94 4.2.2 CDT induced adverse morphological changes and increased G:F -actin ratio Actin rearrangement was . -10 -35 TCTTATTTAGATTA TATAAT CTAAATAAATTAAAG TTCAAG AG -31 - 10 TTAATT A AACTAATATTGGGAGGGAGAATAAATGAAAAAATTT 12 AGGAAACAT TGATGCAACATTGA 138 3 TACCTTAA tattttttcacataaataatttaatatttttcaa atttaaggAGGAGAaaca. GAACCATCTCTTTTTTTATACAAAAAAAGTAGTTCCTAAGAAT - 31 0 CCTCTATA TCTCTTTAAAATATT - 160 CAG TTGTTA TTTTGTACTGACATATCA TATAAA TACATATTTT -117 TATGATATATAGTTACATATTTTATGAAATTTATATAAAAAAT - 74 -35 -10 -35 TCTTATTTAGATTA TATAAT CTAAATAAATTAAAG TTCAAG AG. (TATACAAAACAAATTATT TAA), 2 030 9 cdtB (ACTACAAATTATTCCCATACA), and 19126 truncated cdt (TATACAAGACAAATTATT ACCATACA). Dashed lines show the extent of cdt deletion (not drawn to scale). 69 3. 2.4. Analysis of cdt