JOURNAL OF CLINICAL MICROBIOLOGY, Apr. 2009, p. 1050–1057 0095-1137/09/$08.00⫹0 doi:10.1128/JCM.02242-08 Copyright © 2009, American Society for Microbiology. All Rights Reserved.
Vol. 47, No. 4
Truncated Human Cytidylate-Phosphate-Deoxyguanylate-Binding Protein for Improved Nucleic Acid Amplification Technique-Based Detection of Bacterial Species in Human Samples䌤 Svea Sachse,1 Eberhard Straube,1* Marc Lehmann,2 Michael Bauer,3 Stefan Russwurm,2,3 and Karl-Hermann Schmidt1 Institute of Medical Microbiology, Friedrich-Schiller-University, Jena, Germany1; SIRS-Lab GmbH, Jena, Germany2; and Department of Anaesthesiology and Intensive Care Therapy, Friedrich-Schiller-University, Jena, Germany3 Received 21 November 2008/Returned for modification 9 January 2009/Accepted 24 January 2009
A trunk of human cytidylate-phosphate-deoxyguanylate-binding protein/CXXC finger protein 1 (CFP1), immobilized onto an aminohexyl-Sepharose column, can be used as a preanalytical tool for the selective enrichment of bacterial DNA from mixed solutions with high amounts of human background DNA for nucleic acid amplification technique-based detection of pathogens. The transcriptional activator protein exhibits a high affinity for nonmethylated CpG dinucleotide motifs, which are differentially distributed in prokaryotic and higher eukaryotic genomes. The feasibility of the affinity chromatography (AC) step was tested with DNA from severely septic patients. AC using 16S rRNA gene primers substantially increased PCR sensitivity. Approximately 90% of eukaryotic DNA was removed, which significantly increased the signal-to-noise ratio. Threshold cycle values revealed that sensitivity was elevated at least 10-fold. The change in the ratio of bacterial DNA to human DNA increased from 26% to 74% the likelihood of culture-independent PCR-based identification of bacterial presence. Compared to the results seen with blood culture (which is the clinical gold standard for systemic infections, exhibiting 28% positives), the combination of AC and PCR achieves a significant increase in sensitivity and contributes to shortening the time to results for the initiation of guided antibiotic therapy. delay of adequate antibiotic treatment (12). Moreover, the rise in the rate of infective diseases and the liberal use of broad-spectrum antibiotics has led to increasing resistance problems in intensive-care units and the likelihood of misdirected long-term therapies, while new classes of antibiotics are rare (31). Thus, identification of the causative pathogen(s) and antibiotic resistances is required forthwith. Nucleic acid amplification techniques (NAT) (e.g., PCR) applied to systemic infection diagnostics allow for a more rapid (within several hours) result for target and resistance detection compared to culture-based methods. Free bacterial DNA, as well as DNA from adherent, phagocytosed, or free intact and nonintact pathogens, is detected, while cultural methods contribute only to the detection of viable cells. However, molecular diagnostics for the culture-independent identification of the infectious stimuli still pose significant problems. The high sensitivity of detection is decreased by factors such as high fractions of eukaryotic bulk DNA, salts, hemin, and other blood ingredients, most of the latter of which should be effectively removed by affinity chromatography (AC) steps during sample preparation. PCR is, despite all inherent limitations, ready to be introduced in routine diagnostics and currently reflects the most promising avenue to decrease the time to results. Although there is a broad consensus that NAT may eliminate the above-mentioned drawbacks and improve diagnostics, i.e., in cases of polymicrobial infections and of those caused by fastidious, multiresistant, and noncultivable strains (5, 9, 21, 30), the minute quantities of pathogen genome copies compared to the huge eukaryotic DNA background within clinical samples result in significant signal-to-noise problems. Thus, preanalytical strategies to alter
The routine methods utilized in clinical microbiology laboratories, such as the demonstration of the presence of pathogens in samples from patients suspected of systemic infections, are predominantly culture based and exhibit drawbacks due to antibiotic treatment of the patient prior to sample withdrawal (e.g., from blood, wound swabs, or cerebrospinal fluid), low abundance of causative agents in exclusive samples, and, frequently, noncultivable or growth-repressed organisms. The gold standard, blood culture (BC), for example, returns negative results for 80% to 90% of all invasive infection incidents even when the presence of an infection is obvious from the medical history and additional clinical diagnostics. Cultural results usually take periods of 24 to 72 h to be obtained, whereas a sample can be reliably declared negative within up to 7 days’ incubation (26, 27). These results are therefore only the basis for further microbial diagnostics, e.g., species differentiation and/or generation of antibacterial susceptibility profiles, which are also laborious and time-consuming processes. The derivation of an antibiotic therapy from the results obtained with the gold standard (e.g., in the case of sepsis) within the first “golden hours” would determine the course and prognosis for the case (17); however, such an approach is currently not feasible, while adequate (directed) and early antibiotic therapies are mandatory for the avoidance of mortal outcomes (7, 8, 11, 28). In the case of sepsis, an increase of mortality of 7% to 8% per hour was proven after * Corresponding author. Mailing address: Institute of Medical Microbiology, Friedrich-Schiller-University, Erlanger Allee 101, D-07747 Jena, Germany. Phone: 49 (0) 3641 93 93 501. Fax: 49 (0) 3641 93 34 74. E-mail:
[email protected]. 䌤 Published ahead of print on 4 February 2009. 1050
VOL. 47, 2009
IMPROVED NAT-BASED DETECTION OF BACTERIAL SPECIES
this ratio are likely to increase the sensitivity and specificity of PCR-based assays. Discrimination between “self” DNA and foreign DNA is achieved in vivo by immunocompetent cells via species-specific cytidylate-phosphate-deoxyguanylate (CpG) motif recognition, as described for, e.g., human Toll-like receptor 9 (hTLR9) (1) and the human CpG-binding protein (hCGBP) (also known as CXXC finger protein 1 [CFP1]) (29). The latter transcriptional activator avidly binds nonmethylated CpG motifs by recognition of the sequence [A/C]CG[A/C] (16) with an even higher number of potential binding sites than hTLR9 (1). Methylation of the 5⬘ CpG sequence is an epigenetic modification of vertebrate DNA, and such motif clusters are colocalized within regions of silenced promoters (13, 15). Compared to those of vertebrate DNA, nonmethylated CpG dinucleotide patterns occur in microbial nucleic acids with a frequency 20-fold higher. In this communication, the expression of a truncated derivative of hCGBP/CFP1, its immobilization onto aminohexylSepharose, and further use as an AC-based preanalysis tool for the selective enrichment of bacterial DNA from mixed solutions with high amounts of human background DNA are described. A significant improvement of the sensitivity of NAT-based downstream pathogen detection was achieved in samples derived from patients suffering from severe sepsis.
MATERIALS AND METHODS Bacterial strains, plasmids, and chemicals. Competent cells of the Escherichia coli K12 strains JM109 (Promega, Mannheim, Germany) and EZ (Qiagen, Hilden, Germany) were used for subcloning and maintenance of plasmids. E. coli M15 (pREP4; Qiagen) was used for the expression of recombinant proteins. The plasmids pDrive (Qiagen) and pCR2.1 (Invitrogen, Karlsruhe, Germany) served as vectors for TA subcloning of the PCR fragments. Expression of recombinant proteins was completed using vector pQE9 (Qiagen). The plasmid pUC18emmC, which carries the gene encoding the M protein of the group C streptococcal strain 25287 (MC) (10), was used for binding experiments. Calf thymus DNA was purchased from Serva (Heidelberg, Germany), while all other chemicals were obtained from Sigma (Deisenhofen, Germany) unless stated otherwise. Preparation of total DNA from whole blood. Total DNA was isolated from buffy coats from voluntary donors who displayed no signs of infections (for spiking experiments) and patients with severe sepsis. The isolated samples were prepared from 5 ml of anticoagulated EDTA whole blood after dextran (5%) density gradient centrifugation on the basis of the assumption that the majority of bacteria is associated with or within phagocytic cells. The cells were resuspended in 200 l of TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7). A 50-l volume of a mixture of lysozyme (20 mg/ml) and mutanolysin (5 mg/ml) (Sigma) was subsequently added, and the suspension was incubated at 37°C for 90 min. A final incubation with 50 l of 10% sodium dodecyl sulfate (SDS) and 50 l of proteinase K (20 mg/ml) at 50°C for 2 h followed and resulted in complete lysis. The volume was adjusted to 500 l with TE buffer, and DNA isolation was performed by phenol-chloroform-isoamyl alcohol extraction (23). A final precipitation was done with the addition of a 0.7 volume of ice-cold isopropanol and subsequent centrifugation at 15,000 ⫻ g for 30 min. The pellet was washed twice in 70% ice-cold ethanol and vacuum dried. The lyophilisates were resuspended in 110 l of TE buffer, and the DNA concentration was determined in 10 l of the solution via A260/280 (Ultrospec 3000 spectrophotometer; Amersham Pharmacia). On average, a concentration of 0.1 g of total DNA per l was obtained. Plasmid isolation, cloning, and expression of the truncated CpG-binding protein P181. Plasmid DNA was isolated using a plasmid Mini kit from Qiagen. Extraction of bacterial DNA was performed by enzymatic lysis at 37°C for 30 min followed by digestion with proteinase K at 50°C for 1 h. The obtained DNA pellet was subjected to phenol-chloroform extraction and subsequent ethanol precipitation. RNA isolation and cDNA production were completed using human synovial fibroblasts and an RNeasy Kit (Qiagen) according to the manufacturer’s instructions. PCR was performed with the primers P756fw and P756rv (Table 1),
1051
TABLE 1. Primers used for PCR and sequencing Primer
Sequence
P756fw ...............5⬘-CCATGGGTGGAGGGCGCAAGAGGCCTG-3⬘ P756rv................5⬘-CAGATCTGTATCGCTCCTCCTTCTTCTTCTCAGAC-3⬘ P181fw ...............5⬘-GGATCCGGTGGAGGGCGCAAGAGGCCTG-3⬘ P181rv................5⬘-AAGCTTAGAGGTAGGTCCTCATCTGAG-3⬘ M13fw................5⬘-GTAAAACGACGGCCAGTG-3⬘ M13rv ................5⬘-CAGGAAACAGCTATGAC-3⬘ 16Sfw .................5⬘-CCAGCAGCCGCGGTAATACG-3⬘ 16Srv..................5⬘-TAAGGGCCATGAKGAYTTGAC-3⬘ irp2fw.................5⬘-GGCTGATTACCAACAATTACGC-3⬘ irp2rv .................5⬘-GGCTGAGCATTAACTGGTTCC-3⬘ hlafw ..................5⬘-CTGGTCAATATAGAGTTTATAGCGAAG-3⬘ hlarv...................5⬘-ATGCAATTGGTAATCAT CACGAAC-3⬘ rpoDfw ..............5⬘-GATCTTCAGTACCTTGCGGATCTTG-3⬘ rpoDrv ...............5⬘-TGATTTCCATCGCCAAGAAGTACAC-3⬘
derived from the sequence of hCGBP/CFP1 (29) (NCBI accession no. AF149758), and the cDNA of hCGBP/CFP1 as the template revealed a 756-bp fragment, which was ligated into the vector pDrive (pDrivecg756) and transformed into EZ-competent cells. The correctness of clone pDrivecg756 was verified by sequencing, and the clone was used as a source for designing further truncated gene fragments. An hCGBP/CFP1 cDNA fragment encoding amino acids 106 to 286 (termed P181; Fig. 1) was amplified with primers P181fw and P181rv (Table 1) and the plasmid pDrivecg756 as the template. After being subcloned in plasmid pCR2.1 and digested with BamHI and HindIII, the insert was ligated into pQE9 expression vector, resulting in pQE9p181. This construct allowed for expression of an N-terminal His-tag fusion protein. pQE9p181 was transformed into E. coli expression strain M15 (pREP4). Overnight cultures (2 ml) of the clone were inoculated in 2⫻ Luria-Bertani (LB) broth (200 ml) supplemented with 100 g of ampicillin and 25 g of kanamycin per ml. The bacteria were incubated at 37°C while being shaken (150 rpm) for approximately 4 h until the optical density at 600 nm reached 0.6. After addition of isopropyl-ß-D-thiogalactopyranoside at a final concentration of 2 mM, incubation continued for a further 12 h under the aforementioned conditions. The majority of the P181 produced was found in inclusion bodies and was soluble in buffers containing 6 M guanidinium-HCl. For the preparation of the recombinant protein, the bacterial sediment of the 400-ml culture was dissolved in 10 ml of lysis buffer (0.05 M phosphate, 6 M guanidinium-HCl, pH 6.5), precipitated by centrifugation at 10,000 ⫻ g, and applied onto a column containing 6 ml of nickel-nitrilotriacetic acid agarose. The column was washed with buffer (0.05 M phosphate, 8 M urea, pH 6.3). To elute P181, a gradient from pH 6.3 to 4.0 was applied by the addition of glacial acetic acid to the washing buffer. Fractions containing P181 were collected and dialyzed using distilled water. Under these conditions, P181 was renatured and did not precipitate. SDS-polyacrylamide gel electrophoresis (SDS-PAGE) was performed using a Mini-Protean II electrophoresis chamber followed by staining with Coomassie blue R250. Electrophoretic mobility shift assay. The discriminative DNA-binding capability of P181 was qualitatively determined using methylated as well as nonmethylated DNA. Complete methylation of pUC18emmC (10 g) was performed with an SssI methylase kit from New England Biolabs (Frankfurt, Germany). Methylated and nonmethylated plasmid pUC18emmC (0.3 g each), dissolved in 20 l of TE buffer (pH 7.2), was incubated with increasing amounts (5 to 20 g) of P181 in water, and the binding reaction was carried out for 10 min at room temperature. The reaction products were applied onto 1.5% agarose gels and subjected to horizontal electrophoresis (100 V for 1 h) followed by ethidium bromide staining. Binding characteristics of aminohexyl-Sepharose-immobilized P181 for affinity chromatography. P181 was immobilized on aminohexyl-Sepharose (GE Healthcare, Freiburg, Germany) by the glutaraldehyde method described in reference 3, resulting in a spacer of 11 C atoms that allowed the exposure of the protein at the outer surface of Sepharose beads. P181 (2 mg) was coupled onto 1 ml of a Sepharose bead matrix. Human serum albumin-Sepharose (HSASepharose) columns served as negative controls. Conjugates were washed with TE buffer (pH 7.2) containing 1 M NaCl and stored in 20% ethanol at 4°C. For DNA binding, a 100 l (wet volume) of P181-Sepharose was filled in small spin columns (Qiagen) and equilibrated with water or TE buffer at pH 7.2. The appropriate DNA samples (50 l) were applied onto each column to assess the competitive binding behaviors of pro- and eukaryotic DNA fractions. Washing and elution were performed by intermittent additions of 100-l portions of TE buffer with incremental increases in NaCl concentrations. For PCR
1052
SACHSE ET AL.
J. CLIN. MICROBIOL.
FIG. 1. Characteristics of recombinant P181. (A) P181 comprises amino acids 106 to 286, including the CXXC zinc finger domain of mature human DNA-binding protein 1 (hCGBP/CFP1) (left box, PHD type zinc finger; middle [dark gray] box, CXXC zinc finger; right box, coiled-coil region). (B) SDS-PAGE analysis of 0.11 g of recombinant P181 (Coomassie blue staining). M, molecular mass marker.
analysis of the eluates, each fraction was precipitated with isopropanol in the presence of 0.3 M sodium acetate and redissolved in 30 l of water without further determination of the DNA concentration. The 50-l PCR volumes consisted of 5 l of 25 mM MgCl2, 5 l of 2 mM deoxynucleoside triphosphate mix, 1 l of Taq polymerase (Fermentas GmbH, St. Leon-Rot, Germany) (1 unit), 5 l of 10⫻ PCR buffer with (NH4)2SO4 (Fermentas), 1 l each (10 pmol each) of 16S rRNA gene-specific forward and reverse primers (Table 1), 5 l of fractioned and unfractioned DNA (before AC), respectively, as the template, and 27 l of DNA-free and DNase-free water. PCR experiments were performed as follows: 1 cycle of 94°C for 6 min, 35 cycles of 94°C for 30 s, 50 to 60°C for 30 s, and 72°C for 2 min, and 1 cycle of 72°C for 7 min. All steps were performed using a Mastercycler ep gradient S system (Eppendorf AG, Hamburg, Germany). The samples were analyzed on 2% agarose gels. As a positive control, 5 l of Staphylococcus aureus DNA (0.0005 g/l) (and, as a negative control, 5 l of water) was added instead of donor DNA. DNA labeling and quantitative binding experiments. For labeling of plasmid or bacterial chromosomal DNA, the fragment was digested with SalI and sticky ends were filled up with Klenow fragment (Hybaid AGS, Heidelberg, Germany) in the presence of [␣-32P]dCTP according to standard protocols. The labeled DNA fragments were separated from immobilized nucleotides on a Sephadex G-50 column (GE Healthcare, Freiburg, Germany) after termination of the reaction by heating for 20 min to 70°C. To assess the feasibility of the relative enrichment by P181 of bacterial DNA through the use of a mixture of pro- and eukaryotic DNA, quantitative binding studies were conducted using radiolabeled bacterial DNA with human DNA as the competitor. [␣-32P]dCTP-labeled DNA (0.02 g) derived from Streptococcus pyogenes BK 42440 was mixed with a molar excess of 50 g of human DNA in 250 l of TE buffer. A 240-l volume of the mixture was applied to 200 l of P181-Sepharose and incubated for 10 min at room temperature. Subsequently, the column was washed twice with 200 l of TE buffer (pH 7.2) and the bound DNA was eluted twice with 200 l of 0.7 M NaCl in TE buffer under the elution conditions given above. The DNA concentration was determined spectrophotometrically (see above), and 2 min of scintillation (LS 6000 TA system; Beckman, Krefeld, Germany) of bound and unbound fractions was performed. Six independent experiments were performed to calculate the standard deviation (see Fig. 4). Relative enrichment of bacterial DNA as assessed by real-time PCR of bacterial target sequences. In additional experiments, mixed pro- and eukaryotic DNA templates with or without P181-AC were investigated by real-time PCR as a typical readout for culture-independent identification of potential pathogens. Samples of 0.02 g of E. coli DNA and 50 g of human DNA in 300 l of TE buffer were prepared in triplicate experiments. One aliquot was directly subjected to isopropanol precipitation. The second aliquot was passed through a 200-l P181-Sepharose column. Flowthrough and the eluate at 0.7 M NaCl were collected in a 300-l buffer volume followed by DNA precipitation with a 0.7 volume of isopropanol. After washing with ice-cold 70% ethanol, the DNA pellet was dissolved in 30 l of water. For real-time PCR, a negative control experiment using DNA-free water was
performed analogously to the procedure followed with the DNA samples to determine a bacterial DNA handling threshold (cutoff). No-template controls were included. The detection was based on fluorescence due to insertion of SYBR green into double-stranded DNA. A 25-l reaction volume consisted of 0.2 g of total genomic DNA in 10 l of water, 12.5 l of 2⫻ QuantiTect SYBR green PCR master mix (Qiagen), and 1.25 l (10 pmol final concentration) of each of the forward and reverse primers (Table 1). All steps were performed in duplicate on a Rotor-Gene RG-3000 quantitative PCR (qPCR) device (Corbett Life Science, Sydney, Australia). An initial denaturation of DNA was carried out at 94°C for 15 min followed by 45 cycles of 94°C for 30 s, 50°C for 30 s, and 72°C for 1 min. The calculation was done using Rotor-Gene 6 software. Proof of concept for patients with severe sepsis. Within a pilot study, 20 randomly selected patients fulfilling American College of Chest Physicians/ Society of Critical Care Medicine consensus criteria for severe sepsis or septic shock were enrolled after informed consent was obtained from the patient or a legal representative in accordance with the Declaration of Helsinki. Negative controls with samples of donors devoid of signs of infection were eschewed in favor of a direct comparison of results obtained with samples pretreated by AC and with non-AC-treated samples to BC results. A total of 39 EDTA wholeblood samples were assayed comparatively by means of conventional BC, 16S rRNA gene PCR with subsequent amplicon sequencing, and clinical microbiological routine diagnostics. Case study results were additionally supported by PCR with specific primers (for rpoD, encoding the RNA polymerase sigma factor, and hla, encoding alphahemolysin) when multiple 16S rRNA gene PCR amplicons revealed evaluable results. The samples were processed with and without P181-AC prior to PCR analysis. For BC, 5-ml aliquots of whole blood were taken for aerobic and anaerobic cultures by the use of a BacT/Alert system (BioMerieux, Marcy l’Etoile, France). Aliquots from positive BCs were examined by Gram staining, subcultivated, and typed according to standard protocols. AC was done with 5 g of total DNA in 50 l of water (isolated from buffy coats as outlined above) applied to a 100-l P181-Sepharose spin column. The column was treated stepwise twice with a 100-l washing mixture (10 mM Tris, 10 mM EDTA, pH 7.5) and elution buffer (0.5 M NaCl or 1 M NaCl). The eluted DNA was subsequently precipitated with isopropanol and dissolved in 30 l of water. 16S rRNA gene PCR was done as described above. All PCR amplicons obtained from patient samples were subjected to sequencing in both sense and antisense directions performed with a BigDye Terminator v1.1 cycle sequencing kit and an ABI Prism 310 sequencer (Applied Biosystems, Foster City, CA) according to standard protocols. Sequencing was performed as follows: 1 l of forward and reverse primers (10 pmol each), 5 l of PCR amplicon purified from the gel with an Invisorb spin DNA extraction kit (Invitek, Berlin, Germany), and 4 l of sequencing kit material were merged. The following program was executed: 1 cycle at 96°C for 1 min and 25 cycles of 96°C for 10 s, 50°C for 5 s, and 60°C for 4 min. The electropherograms obtained were analyzed using sequencing analysis software (version 3.7; Applied Biosystems). The results were aligned and examined by GenBank NCBI genetic sequence database searching. Species-specific PCR experiments were performed as follows: the 25-l PCR
VOL. 47, 2009
IMPROVED NAT-BASED DETECTION OF BACTERIAL SPECIES
1053
volumes consisted of 0.25 l of 50 mM MgCl2, 2.5 l of 10⫻ AccuPrime buffer II, including deoxynucleoside triphosphate mix (Invitrogen), 0.25 l of AccuPrime Taq polymerase (Invitrogen) (5 units/l), 1 l each (10 pmol each) of hla- and rpoD-specific forward and reverse primers (Table 1), 0.2 g of template DNA, and 19 l of DNA- and DNase-free water. The following PCR program was utilized: 1 cycle of 94°C for 3 min, 30 cycles of 94°C for 45 s, 55°C for 30 s, and 72°C for 45 s, and 1 cycle of 72°C for 2 min. All steps were performed using a personal Mastercycler system (Eppendorf AG, Hamburg, Germany).
RESULTS Cloning, expression, and characteristics of recombinant P181. P181 is a truncated protein covering amino acids 106 to 286 of hCGBP/CFP1 (Fig. 1A). The protein contains the zinc-finger domain of CXXC, a highly conserved DNA-binding protein selective for nonmethylated CpG motifs. P181 was N-terminally His tagged and expressed from plasmid pQE9 in E. coli in the form of inclusion bodies. After the inclusion bodies were dissolved in Tris buffer containing 6 M guanidinium-HCl, the protein was purified by chelate chromatography using nickel-nitrilotriacetic acid agarose under denaturing conditions, eluted from the column with a pH gradient of 6.3 to 4.0 in a sodium phosphate-sodium acetate buffer mixture containing 8 M urea, and refolded by dialysis using distilled water. The protein could be stored in water at ⫺20°C for more than 6 months while still retaining its binding characteristics. A molecular mass of 21.8 kDa was calculated from the amino acid sequence, while SDS-PAGE determined a molecular mass of 24 kDa due to the His-tag fusion (Fig. 1B). The isoelectric point of pH 10.4 (theoretical value) was consistent with the high amounts of 31 positively charged amino acids predicted from the gene sequence. Binding behavior of P181. Binding of nonmethylated CpG motifs is a characteristic feature of hCGBP/CFP1 and reflects the basic principle for the relative enrichment of bacterial DNA in this AC approach. Thus, a first set of experiments was conducted to assess whether P181 can distinguish between methylated and nonmethylated DNA in mixed solutions. P181 retarded the migration of the nonmethylated pUC18emmC plasmid (pUC18 with the MC protein gene of group C streptococci included) in mobility shift assays, while the electrophoretic mobility of the methylated plasmid was not affected (Fig. 2A). To test for nonspecific binding, P181-free HSASepharose columns were used as negative controls in binding experiments and documented the lack of nonspecific retardation by the columns (Fig. 2B). To specify the binding characteristics of P181, 100 l of mixed DNA containing a molar excess of eukaryotic calf thymus DNA (2 g) and the pUC18emmC plasmid (0.025 g) were loaded onto a P181-Sepharose column. Bound DNA was eluted using incrementally increased NaCl concentrations. The DNA content of the elution fractions was monitored via A254, and aliquots were subjected to PCR using the primers M13fw and M13rv (Table 1) to amplify plasmid DNA. The majority of calf thymus DNA was eluted at NaCl concentrations of ⬍0.3 M, and nonmethylated pUC18emmC eluted at ⱖ0.4 M NaCl, indicating preferential binding of nonmethylated plasmid DNA by immobilized P181 (Fig. 3). To quantify the competitive binding of P181, radioactively labeled genomic S. pyogenes BK 42440 DNA was loaded onto the prepared columns together with a 2,500-fold excess of
FIG. 2. Binding of P181 to prokaryotic DNA as assessed by electrophoretic mobility shift analysis and lack of prokaryotic DNA retardation by HSA-Sepharose. (A) Retarded electrophoretic migration of nonmethylated plasmid pUC18emmC (second and fourth lanes) containing the gene encoding the M protein of the group C streptococcal strain 25287 in the presence of P181. Plasmid DNA (0.15 g) was incubated in each binding reaction with either 10 g (first, second, and fifth lanes) or 5 g (third and fourth lanes) of P181 at room temperature for 30 min and was subjected to electrophoresis on a 1.5% agarose gel stained with ethidium bromide. Any retardation that occurred was observed upon methylation of the plasmid (first and third lanes), as is consistent with maintained recognition of nonmethylated CpG motifs by the truncated protein. The fifth lane presents results for nonmethylated pUC18emmC DNA in the absence of P181 (control); the sixth lane represents the molecular mass marker. (B) Proof of nonbinding of prokaryotic DNA on HSA-Sepharose columns (in the absence of P181). A mixture of 25 g of human DNA spiked with 0.02 g of S. aureus chromosomal DNA in 100 l of water was applied to 100 l of HSA-Sepharose. PCR with the obtained fractions was performed to detect the 16S rRNA gene. Lane 1, flowthrough; lanes 2 and 3, fractions occurring after washing with buffer devoid of NaCl (10 mM Tris, 10 mM EDTA, pH 7.5); lanes 4 and 5, fractions occurring after washing with buffer containing 1 M NaCl; lane 6, total DNA loaded onto HSA-Sepharose column; lane 7, positive control with S. aureus DNA; lane 8, negative control; lane M, pGEM molecular mass standard.
human DNA, and effluent and bound fractions were monitored by scintillation to assess the recovery rates. Despite the molar excess of competing human DNA, about 90% of the human DNA was lost in flowthrough or wash fractions but 60% to 70% of bacterial DNA was retained in the eluent. Only 10% of the eukaryotic DNA coeluted with the majority of the bacterial nucleic acids (Fig. 4). Thus, the fractionation step performed using P181-coated Sepharose columns strongly improved the ratio between the bacterial and human DNA levels in favor of bacterial DNA. Use of prokaryote-enriched DNA as a template for downstream NAT. To test whether the P181-mediated enrichment of minute quantities of prokaryotic DNA from mixed samples containing a huge bulk of eukaryotic host DNA improved qPCR-based pathogen detection, genomic E. coli DNA was used as a prokaryotic target in a molar excess of human bulk DNA with and without subsequent P181-AC. The strain-specific irp2 gene (encoding iron-repressible protein 2) was amplified (Table 1). A flat increase of the qPCR SYBR green fluorescence of AC-treated fractions was observed (Fig. 5). The calculation of threshold cycle (CT) values as determined
1054
SACHSE ET AL.
J. CLIN. MICROBIOL.
FIG. 3. Separation of nonmethylated plasmid DNA from an excess of eukaryotic DNA by P181-AC. A molar excess of 2 g of calf thymus DNA spiked with 0.025 g of pUC18emmC plasmid DNA was loaded onto P181-Sepharose columns and eluted by incremental increases of [NaCl] in 10 mM Tris-HCl elution buffer (pH 7.0). The majority of calf thymus DNA in the eluate appeared at concentrations below 0.3 M NaCl. Elution of plasmid DNA started at 0.4 M NaCl, as shown by a peak of A254 in fraction 21, which was confirmed by PCR using plasmid-specific primers M13fw and M13rv as outlined in Materials and Methods.
by the AC procedure performed previously revealed an increase in sensitivity greater than 10-fold. Proof of concept in samples obtained from septic patients with or without a positive BC. For about 50% of the DNA samples from buffy coats, the results revealed significant overlaps of amplicon sequences. However, the share of positive PCR results with primers targeting 16S rRNA gene regions rose from 26% to 74%, as determined using P181-AC (Fig. 6A). Causative pathogens from two selected patients suffering from severe sepsis could be detected only after enrichment by P181-AC (Fig. 6B, lanes 3, 4, and 9) and the results would have provided therapy-relevant information for the physician in
FIG. 4. Binding of 0.02 g of 32P-labeled S. pyogenes BK 42440 DNA to 200 l of P181-Sepharose in competition with 50 g of human DNA. bact., bacterial.
charge. Neither BCs nor PCR with total DNA as the template (Fig. 6B, lanes 5 and 11) yielded positive results in these cases. Five blood samples were taken from selected patient I within 14 days after admission, and all of them tested BC negative. Sequencing of the appropriate amplicons from samples 2, 3,
FIG. 5. Detection of E. coli PCR targets within a molar excess of human DNA via qPCR using irp2-specific primers (Table 1). A mixture of genomic E. coli (0.02 g) and human DNA (50 g) was used as a template. Flowthrough and starting material (the DNA mixture prior to P181-AC) showed high CT values due to low target concentrations and high background DNA charges. The AC elution fraction exhibited a significantly lower CT value indicative of an improved signal-to-noise ratio. The calculated CT values showed ⬎10-fold-higher sensitivity after P181-AC.
VOL. 47, 2009
IMPROVED NAT-BASED DETECTION OF BACTERIAL SPECIES
1055
FIG. 6. Application of P181-AC for the improvement of PCR-based identification of bacterial presence in buffy coats from 3.5 ml of EDTA whole-blood samples from severely septic patients. (A) Positive PCR results with 16S rRNA gene primers (see Table 1) rose from 26% to 74%. (B) PCR analysis after enrichment of bacterial DNA via P181-AC. Both selected patients had consistently negative BCs. Bacterial presence in the enriched DNA fractions was analyzed by 16S rRNA gene qPCR. Lanes 1 and 7, flowthrough; lanes 2 and 8, washing fractions (buffer without NaCl); lanes 3 and 9, washing fractions (buffer containing 0.5 M NaCl); lanes 4 and 10, washing fractions (buffer containing 1.0 M NaCl); lanes 5 and 11, original samples applied to the P181-AC column; lanes 6 and 12, inhibition controls (sample DNA [without P181-AC] plus S. aureus DNA; lane 13, positive control (S. aureus DNA); lane 14, washing buffer-negative control (no-template control); lane M, pGEM molecular mass marker. Patient I tested positive for P. aeruginosa by sequencing of the 16S rRNA gene amplicons and rpoD-specific PCR; patient II tested positive for S. aureus (confirmed by hla-specific PCR as outlined in Results).
and 4 revealed the closest alignments with the Pseudomonas aeruginosa 16S rRNA gene after a GenBank search. A subsequent rpoD-specific PCR (see Table 1) delivered P. aeruginosa bacteremia results consistent with the clinical data (data not shown). The results of sequencing of the universal PCR amplicon in the sample from the second patient selected, patient II (Fig. 6B, lane 9), were not analyzable. Based on the clinical diagnosis of ventilator-associated pneumonia with S. aureus (documented by the results of an earlier bronchoalveolar lavage), enriched DNA was used for S. aureus-specific amplification of hla and confirmed the presence of S. aureus DNA within the enriched buffy coat DNA (data not shown). DISCUSSION In this study, a trunk of hCGBP/CFP1 was immobilized on a Sepharose column and allowed for the relative enrichment of bacterial DNA from mixtures of eu- and prokaryotic genomic DNA. CpG motif recognition by P181 is essentially similar to endogenous pattern recognition of bacterial DNA by hTLR9. CG dinucleotides were found with a frequency of 1:16 in prokaryotic DNA and are approximately one/fourth less frequent in human DNA (14). Both TLR9 and hCGBP/CFP1 proteins bind nonmethylated CpG motifs with deviating preferences: the consensus motifs are GTCGTT for hTLR9 (1) and [A/C ]CG[A/C] for hCGBP/CFP1 (16), the latter motif appearing statistically more often within the genomes due to the 5⬘- and 3⬘-flanking variable nucleotides. Differing immunostimulatory properties of genomic DNA from bacterial species (including gram-positive and -negative
pathogens, e.g., S. aureus and E. coli) have been tentatively attributed to their CpG motif content (18). The idea of the significance of those differences has been supported by the finding that certain species genomes (from, e.g., the sepsiscausative pathogen Burkholderia cepacia) display significant overrepresentations of immunostimulatory CpG motifs which cause intense inflammatory responses (4). Our own in silico analyses, applying the algorithm in search of the frequency of total (methylated and nonmethylated) [A/C]CG[A/C] motifs within the whole genomes of gram-positive and -negative bacterial species, representing about 90% of the main sepsis-causative pathogens, revealed a medial frequency of one CG motif per 138 bp (average, 1:138). Beside rare outliners with particular high or low motif frequencies (e.g., Clostridium perfringens [ATCC 13124] at 1:804 and Burkholderia sp. at 1:27, confirming the above-mentioned observation [4]), the motifs were assumed to be predominantly nonmethylated. However, the C. perfringens genome also exhibits a significantly higher content of nonmethylated motifs than the human genome: the total motif content of chromosomes 1 to 22, X, and Y has been calculated to amount to 1:443, of which about 80% are methylated (13), which results in a low frequency of nonmethylated [A/C]CG[A/C] motifs of 1:2,216. Epigenetic modifications, including methylation of CpG, have been previously shown to deviate not only in mammalian cancer cells (19). Cytosine can be subjected to variable methylation in bacteria, e.g., under stress conditions in phage-infected streptococci, although to a substantially lesser extent (6). Those differing species- and stage-dependent methylation grades as well as deviations in overall GC contents must consequently affect the
1056
SACHSE ET AL.
number of nonmethylated CpG motifs accessible for the respective DNA-binding proteins. However, P181-AC performed with genomic DNA of many bacterial species with differing target motif contents never failed to improve downstream NAT detection due to unsatisfying pathogen DNA enrichments, which could be attributed to low target motif numbers. The feasibility of P181-AC as a pretreatment tool for DNA templates prior to NAT detection was shown by its ability to remove human DNA (and further potential PCR inhibitors) and simultaneously to enrich prokaryotic DNA regardless of the bacterial species previously spiked. qPCR confirmed a substantial increase in sensitivity after P181-AC of spiked samples. The level of removal of eukaryotic DNA was about 90% (as shown by AC of radiolabeled DNA on immobilized P181), which significantly decreased the signal-to-noise ratio. The CT values seen after P181-AC revealed sensitivity at least 10-fold higher. Standard PCR amplification using 16S rRNA gene primers with DNA isolated from blood samples of severely septic patients also showed increased sensitivity. However, sequencing often revealed mixed infections, which hampered species identification for 51% of the clinical samples, although the length of the generated amplicons was chosen in accordance with the fact that the size might have an effect on contamination: most contaminating DNA derives from nonviable organisms, which could imply that bigger amplicons are less sensitive with respect to contamination than small amplicons (25). Additionally, although amplicon sizes of only 500 bp are usually sufficient for identification of a clinical isolate, longer gene sequences deliver results with greater accuracy (20). However, the inability to analyze these PCR results suggests a need for improvement of the detection methods (e.g., via optimized targets for universal primers and/or various primer designs and detection of genus- or species-specific targets). Sequencing of PCR amplicons as a prerequisite for routine pathogen identification is generally not recommended due to the frequency of nonanalyzable results. The crucial benefits of the tested preanalytical tool, however, are in fact the increase in the number of positive (clinically valid) patient test results and the receipt of therapydirected information. Moreover, a significantly shortened time to results compared to the time required for the current gold standard can be expected. Within a working day, cell lysis, total DNA isolation, pathogen DNA enrichment, and subsequent NAT can be performed. Although it is clear that these cultureindependent approaches are at present not rapid enough to postpone the first dose of empirical antibiotics until identification of a presumably causative organism, they have the potential to minimize the time requirements for substantial readjustment or de-escalation of antibiotic therapies (2, 21). Meanwhile, P181-AC was applied to detect bacterial infections in the ascites of patients suspected of suffering from spontaneous bacterial peritonitis and led to an increase in the sensitivity of detection by a factor of 40 (22). The consequence of increased sensitivity, however, is the rise of false-positive results due to, e.g., contaminants from host flora and associated consumables or introduced via applying routine sample withdrawal techniques. The data pertaining to the origin and clinical significance of those “falsepositive” samples are often ambiguous and might result from
J. CLIN. MICROBIOL.
yet-unknown host-pathogen interactions (24). Consequently, the usage of broad-range primers should be carefully balanced in favor of the detection of particular species. Therefore, the pathogenetic significance of the results has to be confirmed by examination of further clinical data and should not be the sole reason for any therapeutical decisions. P181-based target amplification might significantly enhance downstream NAT sensitivities. For clinical applications, there is a necessity for it to be integrated into an assay system that combines cell disruption, total DNA isolation, and NAT detection, offering high negative and positive predictive values despite high analytical sensitivities and low detection thresholds. ACKNOWLEDGMENTS The project was supported by grants from the Thu ¨ringer Ministerium fu ¨r Wirtschaft, Arbeit und Infrastruktur (TMWAI 2001 FE 0283/ 2004 FE 0115) and Thu ¨ringer Ministerium fu ¨r Wissenschaft, Forschung und Kunst (TMWFK B309-00014). We thank J. Ro ¨del, Jena, Germany, for providing fibroblast cDNA, the senior physician F. Bloos and the study nurses P. Bloos, A. Braune, and U. Redlich for blood withdrawal at the Department of Anesthesiology and Intensive Care Therapy, Friedrich-Schiller-University, Jena, Germany, and I. Walz, D. Hoffmann, and D. Peter for skilled technical assistance. We thank R. Schmitz (SirsLab GmbH) for careful reviewing the manuscript. REFERENCES 1. Bauer, S., C. J. Kirschning, H. Hacker, V. Redecke, S. Hausmann, S. Akira, H. Wagner, and G. B. Lipford. 2001. Human TLR9 confers responsiveness to bacterial DNA via species-specific CpG motif recognition. Proc. Natl. Acad. Sci. USA 98:9237–9242. 2. Bochud, P. Y., M. Bonten, O. Marchetti, and T. Calandra. 2004. Antimicrobial therapy for patients with severe sepsis and septic shock: an evidencebased review. Crit. Care Med. 32(Suppl.):S495–S512. 3. Cambiaso, C. L., A. Goffinet, J. P. Vaerman, and J. F. Heremans. 1975. Glutaraldehyde activated aminohexyl derivative of Sepharose 4B as a new versatile immunosorbent. Immunochemistry 12:273–278. 4. Coenye, T., and P. Vandamme. 2005. Overrepresentation of immunostimulatory CpG motifs in Burkholderia genomes. J. Cyst. Fibros. 4:193–196. 5. Domann, E., G. Hong, C. Imirzaglioglu, S. Turschner, J. Ku ¨hle, C. Watzel, T. Hain, H. Hossain, and T. Chakraborty. 2003. Culture-independent identification of pathogenic bacteria and polymicrobial infections in the genitourinary tract of renal transplant recipients. J. Clin. Microbiol. 41:5500–5510. 6. Euler, C. W., P. A. Ryan, J. M. Martin, and V. A. Fischetti. 2007. M.SpyI, a DNA methyltransferase encoded on a mefA chimeric element, modifies the genome of Streptococcus pyogenes. J. Bacteriol. 189:1044–1054. 7. Fine, J. M., M. J. Fine, D. Galusha, M. Petrillo, and T. P. Meehan. 2002. Patient and hospital characteristics associated with recommended processes of care for elderly patients hospitalized with pneumonia: results from the Medicare quality indicator system pneumonia module. Arch. Intern. Med. 162:827–833. 8. Garnacho-Montero, J., J. L. Garcia-Garmendia, A. Barrero-Almodovar, F. J. Jimenez-Jimenez, C. Perez-Paredes, and C. Ortiz-Leyba. 2003. Impact of adequate empirical antibiotic therapy on the outcome of patients admitted to the intensive care unit with sepsis. Crit. Care Med. 31:2742–2751. 9. Gasanov, U., D. Hughes, and P. M. Hansbro. 2005. Methods for the isolation and identification of Listeria spp. and Listeria monocytogenes: a review. FEMS Microbiol. Rev. 29:851–875. 10. Geyer, A., A. Roth, S. Vettermann, E. Gu ¨nther, A. Groh, E. Straube, and K. H. Schmidt. 1999. M protein of a Streptococcus dysgalactiae human wound isolate shows multiple binding to different plasma proteins and shares epitopes with keratin and human cartilage. FEMS Immunol. Med. Microbiol. 26:11–24. 11. Ibrahim, E. H., G. Sherman, S. Ward, V. J. Fraser, and M. H. Kollef. 2000. The influence of inadequate antimicrobial treatment of bloodstream infections on patient outcomes in the ICU setting. Chest 118:146–155. 12. Iregui, M., S. Ward, G. Sherman, V. J. Fraser, and M. H. Kollef. 2002. Clinical importance of delays in the initiation of appropriate antibiotic treatment for ventilator-associated pneumonia. Chest 122:262–268. 13. Klose, R. J., and A. P. Bird. 2006. Genomic DNA methylation: the mark and its mediators. Trends Biochem. Sci. 31:89–97. 14. Krieg, A. M., A.-K. Yi, S. Matson, T. J. Waldschmidt, G. A. Bishop, R. Teasdale, G. A. Koretzky, and D. M. Klinman. 1995. CpG motifs in bacterial DNA trigger direct B-cell activation. Nature 374:546–549.
VOL. 47, 2009
IMPROVED NAT-BASED DETECTION OF BACTERIAL SPECIES
15. Larsen, F., G. Gundersen, R. Lopez, and H. Prydz. 1992. CpG islands as gene markers in the human genome. Genomics 13:1095–1107. 16. Lee, J.-H., K. S. Voo, and D. G. Skalnik. 2001. Identification and characterization of the DNA binding domain of CpG-binding protein. J. Biol. Chem. 276:44669–44676. 17. Meehan, T. P., M. J. Fine, H. M. Krumholz, J. D. Scinto, D. H. Galusha, J. T. Mockalis, G. F. Weber, M. K. Petrillo, P. M. Houck, and J. M. Fine. 1997. Quality of care, process, and outcomes in elderly patients with pneumonia. JAMA 278:2080–2084. 18. Neujahr, D. C., C. F. Reich, and D. S. Pisetsky. 1999. Immunostimulatory properties of genomic DNA from different bacterial species. Immunobiology 200:106–119. 19. Nishiyama, R., L. Qi, M. Lacey, and M. Ehrlich. 2005. Both hypomethylation and hypermethylation in a 0.2-kb region of a DNA repeat in cancer. Mol. Cancer Res. 11:617–626. 20. Patel, J. B. 2001. 16S rRNA gene sequencing for bacterial pathogen identification in the clinical laboratory. Mol. Diagn. 6:313–321. 21. Peters, R. P., M. A. van Agtmael, S. A. Danner, P. H. Savelkoul, and C. M. Vandenbroucke-Grauls. 2004. New developments in the diagnosis of bloodstream infections. Lancet Infect. Dis. 4:751–760. 22. Sachse, S., S. Assefa, K.-H. Schmidt, M. Lehmann, A. Herrmann, R. P. H. Schmitz, E. Straube, and A. Stallmach. 2008. NAT detection of bacterial pathogens in ascites of patients with cirrhosis of the liver outclasses ascites cultures. Int. J. Med. Microbiol. 298(Suppl. 45):13. 23. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
1057
24. Schrenzel, J. 2007. Clinical relevance of new diagnostic methods for bloodstream infections. Int. J. Antimicrob. Agents 30:S2–S6. 25. Schuurman, T., R. F. de Boer, A. M. Kooistra-Smid, and A. A. van Zwet. 2004. Prospective study of use of PCR amplification and sequencing of 16S ribosomal DNA from cerebrospinal fluid for diagnosis of bacterial meningitis in a clinical setting. J. Clin. Microbiol. 42:734–740. 26. Shang, S., G. Chen, Y. Wu, L. Du, and Z. Zhao. 2005. Rapid diagnosis of bacterial sepsis with PCR amplification and microarray hybridization in 16S-rRNA gene. Pediatr. Res. 58:143–148. 27. Tong, M. Q., S. Q. Shang, Y. D. Wu, and Z. Y. Zhao. 2004. Rapid diagnosis of neonatal sepsis by 16S-rRNA genes PCR amplification and gene chip hybridization. Zhonghua Er Ke Za Zhi 42:663–667. (In Chinese.) 28. Valle´s, J., J. Rello, A. Ochagavia, J. Garnacho, and M. A. Alcala. 2003. Community-acquired bloodstream infection in critically ill adult patients: impact of shock and inappropriate antibiotic therapy on survival. Chest 123:1615–1624. 29. Voo, K. S., D. L. Carlone, B. M. Jacobsen, A. Flodin, and D. G. Skalnik. 2000. Cloning of a mammalian transcriptional activator that binds unmethylated CpG motifs and shares a CXXC domain with DNA methyltransferase, human trithorax, and methyl-CpG binding domain protein 1. Mol. Cell. Biol. 20:2108–2121. 30. Wellinghausen, N., B. Wirths, A. Essig, and L. Wassill. 2004. Evaluation of the Hyplex BloodScreen multiplex PCR–enzyme-linked immunosorbent assay system for direct identification of gram-positive cocci and gram-negative bacilli from positive blood cultures. J. Clin. Microbiol. 42:3147–3152. 31. Wenzel, R. P. 2004. The antibiotic pipeline—challenges, costs, and values. N. Engl. J. Med. 351:523–526.
