A highly sensitive quantitative real-time PCR assay based on the groEL gene of contemporary Thai strains of Orientia tsutsugamushi

Partial nucleotide sequences (459 bp) of the groEL gene (encoding the 60-kDa heat shock protein, HSP60) from 23 contemporary isolates of Orientia tsutsugamushi isolated from patients with acute scrub typhus in Thailand were compared with 16 reference strain sequences to evaluate the potential of groEL as a conserved and representative target for molecular diagnostics.. Overall nucleotide identity within all available O. tsutsugamushi isolates (n = 39) was 98.8% (range: 95.0–100), reflecting a high degree of conservation; nucleotide identities were 67.5% and 65.6%, respectively, when typhus and spotted fever group rickettsiae were included.. A highly sensitive and quantitative real-time PCR assay was designed and evaluated using 61 samples, including buffy coats from patients in Thailand and Laos. Reliable and accurate quantitation of bacterial loads allows further investigation of other diagnostic methods and may lead to an improved understanding of the pathophysiology of acute scrub typhus, an important but under-recognized disease.


Introduction
The Rickettsiaceae family consists of a group of highly fastidious, obligate intracellular Gram-negative organisms. They are divided into three groups, based on antigenic reactivity-the scrub typhus group, typhus group (TG) and spotted fever group (SFG). Scrub typhus, caused by Orientia tsutsugamushi, and murine typhus, caused by Rickettsia typhi, are the most common forms of Typhus in rural Thailand and Laos, accounting for 20-30% of undifferentiated fevers [1][2][3][4].
Problems in under-recognition of rickettsial illnesses, mainly due to diagnostic difficulties, lead to delay and errors in patient management. The development of rapid, inexpensive and accurate diagnostic methods is necessary, both to improve diagnosis and to promote awareness of these potentially serious but treatable diseases in highly populous rural areas of Southeast Asia.
With the increasing availability of gene sequences, allowing the exploitation of more gene-based targets, molecular assays have been developed and evaluated for the diagnosis of acute scrub typhus. A common target gene used in nested conventional as well as real-time PCR assays, encodes the 56-kDa outer membrane protein [5,6]. Another target gene encodes the 47-kDa outer membrane protein, used in a real-time PCR assay [7].
Quoted sensitivities for the nested 56-kDa assays range from 62% to 90%, with specificities approaching 100% when compared to the reference standard immunofluorescence assay (IFA) [5,6,8,9]. Recent studies have characterized the 60-kDa heat shock protein GroEL of a-proteobacteriaceae as a molecular indicator of various forms of cellular stress. GroEL production is upregulated during the early period of infection, leading to highlevel expression of essential proteins in eubacterial genomes and in eukaryotic organelles [10,11]. The most prominent protein of Rickettsia conorii (SFG), revealed by two-dimensional PAGE proteomic analysis and reacting with antibodies in rabbit and patient sera, was a 60-kDa protein identified as GroEL [12].
The corresponding gene has been proposed as a target for molecular diagnostics for differentiation between members of the genus Rickettsia [13] and the family Anaplasmata-ceae [14]. Recently, Park et al. described the use of a conventional duplex PCR assay, based on the groEL gene, for the detection of rickettsiae and the identification of O. tsutsugamushi [15]. This assay was evaluated using a limited number of O. tsutsugamushi reference type strains, including strains Karp, Kato, Kawasaki, Gilliam and Boryong.
The groEL nucleotide sequences of 23 contemporary in vitro isolates of O. tsutsugamushi isolated from patients with scrub typhus in Thailand were determined and the corresponding amino acid sequences deduced. The sequences were compared with those of reference strains to evaluate the potential of groEL as a conserved and representative molecular target. Using these data, together with previously published nucleotide sequence information, a novel and highly sensitive real-time PCR assay was developed for the detection and quantitation of O. tsutsugamushi.

Clinical samples
The O. tsutsugamushi isolates cultured in this study (Table 1) were collected from scrub typhus patients (in 5 mL of full blood containing EDTA) at two sites in Thailand during 2004-2005. Previous antibiotic use was an exclusion criterion for sample collection. The isolates were cultivated in VERO cell monolayers in 25-cm 2 polystyrene tissue culture flasks (Becton Dickinson, Franklin Lakes, NJ, USA) containing RPMI-1640 medium supplemented with L-glutamine, HEPES (2 mM) and fetal bovine serum (FCS Gold; PAA, Laboratories GmbH, Pasching, Austria) (10%, v/v)). Cultures of O. tsutsugamushi were incubated at 35°C in a 5% CO 2 atmosphere [16]. When the cytopathic plaque-formation reached 90-100% confluency of the whole monolayer, the cells were harvested, pelleted and stored at )80°C.
The buffy coat samples (from 5 mL of full blood containing EDTA) for real-time PCR were collected from patients with scrub typhus, who gave informed consent, at Udon Thani Hospital, north-eastern Thailand, and at Mahosot Hospital, Vientiane, Laos PDR.
The present study was approved by the Ministry of Public Health, Royal Government of Thailand (Thailand), the Faculty of Medical Sciences Ethical Review Committee, the National University of Laos (PDR Laos) and the Oxford Tropical Research Ethics Committee (OXTREC, UK). The O. tsutsugamushi isolates cultured in this study (Table 1) were collected from scrub typhus patients (5 ml full blood in EDTA), which gave informed consent, at two sites in Thailand during 2004-2005.

Conventional PCR
DNA was extracted with the Wizard SV Genomic DNA purification system (Promega, Madison, WI, USA). Amplification of the partial groEL gene was performed using PCR with the previously described [13] primers 5¢-GTTGAAGTT/AG TTAAAGG-3¢ (forward) and 5¢-TTTTTCTTTT/ATCATAAT C-3¢ (reverse), generating a product of 534-546 bp. A PCR reaction mix consisted of 50 ng of template DNA, 20 nmol of each primer, 1 U of Taq DNA polymerase, 1.5 mM MgCl 2 and distilled water in a total volume of 20 lL.  Table 1). The reference nucleotide sequences were downloaded from GenBank.

Real-time PCR
On the basis of alignments of sequences determined from conventional PCR products and sequences available from GenBank, a set of specific primers for the generation of a 160-bp amplicon of the groEL gene of O. tsutsugamushi was designed using PrimerSelect To determine detection limits of the assay, plasmids containing the amplified regions of groEL (O. tsutsugamushi UT176 strain) were generated by ligation into pGEM-T Easy Vectors (Promega, USA) and transfered by transformation into Escherichia coli, cultured overnight in a shaking incubator at 37°C in Luria Bertani broth and followed by plasmid extraction using the QIAprep Spin Miniprep Kit (Qiagen, Valencia, CA, USA).
The plasmids were purified and linearized by restriction enzyme digestion with pST1 (Promega, Madison, WI, USA). Linearized DNA was quantified using the Quant-iT PICO Green dsDNA Assay Kit (Invitrogen, Carlsbad, CA, USA). Ten-fold dilution series were used as external controls, and the theoretical number of plasmid copies and corresponding reaction efficiencies were calculated (Rotor-Gene software, Version 6.0; Corbett Research, Australia). Real-time PCR was performed with duplicates of each serial dilution to create a standard curve (Fig. 1).
Comparison of percentage identities of groEL gene sequences among antigenic groups showed that Thai strains and non-Thai strains shared a mean nucleotide identity of 96.7%. When Thai O. tsutsugamushi strains were compared with TG and SFG strains, the percentage identity levels were found to be 67.5% and 65.6%, with similar values for non-Thai Orientia strains when compared to TG and SFG strains, i.e. 67.2% and 65.7%, respectively (Table 3).
Deduced groEL amino acid sequences The Thai O. tsutsugamushi isolates also demonstrated a high overall mean intragroup identity of 98.6% in the amino acid sequences as compared to 89.9% for the non-Thai O. tsutsugamushi isolates ( Table 2). Further subgroup analysis was based on the full open reading frame (ORF) sequences of the more variable 56-kD outer membrane protein derived from 23 isolates [18] and demonstrated a dominance of the Karp subtype in Thailand (15 of the total 23 isolates) with 99.7% average amino acid identity followed by the Gilliam subgroup (7/23) with 99.6% and the TA716-like group (1/23) with 98.9% amino acid identity (data not shown) Comparison of deduced GroEL sequences of STG with those of TG and SFG isolates revealed low amino acid identity values of 28.6% and 29.2%, respectively (Table 3).

Real-time PCR
The assay repeatedly demonstrated a detection limit of < 3 copies/lL of O. tsutsugamushi, using serial dilutions of linearized plasmids (Fig. 1). Amplicons with appropriate melting temperatures (average, 84.6°C; range, 84.3-85.1°C) were produced from all O. tsutsugamushi plasmids, isolates and clinical samples. All templates derived from clinical isolates and strains belonging to the TG and SFG repeatedly and reliably led to negative results.
Quantitative data obtained with the clinical buffy coat samples, including two samples from Laotian patients, demonstrated copy numbers ranging from 2 to 31 668 copies/lL, with a median value of 64 copies/lL of buffy coat. These values represent bacterial loads in admission samples, corresponding to a median "days of fever" time of 6 (5 -10) (interquartile range). One sample (UT530) was a clear outlier, both for buffy coat and isolation quantitation, as it demonstrated high bacterial loads in both samples; 28 237/lL of buffy coat and 1 059 061/lL of VERO cell culture at 100% infection of cells as determined by IFA. In cell culture sam- ples, copy numbers ranged between 5.3 · 10 3 and 1.4 · 10 6 copies/lL (median 7.0 · 10 4 ) of DNA extract (Table 4).

Discussion
The results presented here demonstrate a high level of conservation among the groEL nucleotide and correspond-ing amino acid sequences of contemporary Thai and non-Thai reference isolates of O. tsutsugamushi. The high mean nucleotide intragroup identities among Thai O. tsutsugamushi isolates can be attributed to the similarities between the two main subgroups, Karp and Gilliam. The groEL gene is highly conserved but sufficiently variable to form the basis for genetic target design allowing differentiation of the genera Orientia and Rickettsia, as sequence analysis demonstrated 99.5% identity within the current 23 Thai isolates, and 98.8% identity for all available groEL sequences, including those deposited in GenBank. By comparison, a gene analysis based on sequences encoding the 56-kDa outer membrane protein, covering the full ORF of approximately 1600 bp, demonstrated only 80% identity within the same 23 isolates, underlining the high variability and limitations of this gene as a target for molecular diagnostic assays [18]. At present, only limited DNA sequence data covering the full ORF of the gene encoding the 47-kDa transmembrane protein of O. tsutsugamushi are available, but these data and preliminary sequencing results of strains accross Asia (data not shown) are indicative of a high level of conservation.
To date, only the real-time PCR assay based on the 16S rRNA gene has been evaluated with a large number of clinical samples, and it has demonstrated a diagnostic sensitivity of 45%, using full blood samples drawn upon admission and IFA as a reference standard.
A recently described nested PCR assay showed an increased sensitivity of 82% [21], using buffy coat as a sample specimen. Both assays had excellent specificities and used reference standard IFA as a comparator, but with different cut-off levels for positivity. As O. tsutsugamushi is an obligate intracellular organism disseminating within white blood cells, this assay was based on buffy coat specimens, with the expectation of a concentration effect with higher bacterial loads. A prospective evaluation of the groEL-based real-time PCR assay presented here is underway.
On the basis of isolation and 16S rRNA gene sequencing, Karp and Kawasaki strains were found in the samples examined, and no Kato strains were included. Manosroi et al. [8] have previously described Karp and Kato strains in the same region, although this was based on a nested PCR assay with use of strain-specific nested primer sets. It is possible that these primers designed for the hypervariable 56-kDa protein-encoding gene could be less specific in detecting Gilliam strains in Thailand. In a recent study, the current group analysed the same 23 isolates by full open reading frame sequencing of the highly specific 56-kDa protein, and revealed the predominance of Karp strains and a substantial presence of Gilliam strains [18]. This new information adds to the existing groEL characterization data of five non-Thai isolates [13], suggesting that it is an ideal target for the development of molecular diagnostic assays for O. tsutsugamushi.
Stover et al. [22] first described the high degree of homology between the (formerly known) Rickettsia tsutsugamushi proteins Stp11 and Sta58 and the E. coli proteins GroES and GroEL, respectively, and the family of primor-  dial heat shock proteins designated Hsp10 and Hsp60. Although the sequence homology between the Sta58 antigen and the Hsp60 protein family was striking, the antigenic distinction among other bacterial Hsp60 homologues highlighted the uniqueness of this target, suggesting that it may be both a potentially protective antigen and a useful diagnostic reagent for scrub typhus. Park et al. [14,15] from Korea took further advantage of these features and incorporated the genetic information for identification, differentiation and characterization within the Rickettsiae and Anaplasmatacae. The resulting real-time PCR assay proved to be highly sensitive and specific for all tested isolates of O. tsutsugamushi. As scrub typhus is endemic in many resource-poor developing countries, groEL is a suitable candidate for the application of molecular methods in settings where the costs of establishing a real-time thermocycler are prohibitively high, but alternative methods, e.g. loop-amplified isothermal PCR, an accurate and relatively inexpensive technique, could be used [23,24]. In addition, the reliable and accurate quantitation of bacterial loads allows further investigation of other diagnostic methods and may lead to an improved understanding of the pathophysiology of this important neglected disease. Validation and evaluation in clinical settings in the field are underway.