Early diagnosis and prompt treatment are essential for optimal management of typhoid patients, and therefore a sensitive and specific PCR assay is useful. Although many
Salmonella Typhi PCR-based assays were studied in last decades, none has been brought to clinical use due to the limitations caused by low bactereamic level in typhoid patients. PCR sensitivity is directly related to the actual number of colony forming units found in the blood [
3]. Typical clinical typhoid samples have <1 bacteria/ml of blood [
11,
19]. As a result, PCR offers only limited potential for typhoid diagnostics unless a new sample preparation method is developed. A DNA or bacterial capture system or even a culture enrichment step prior to amplification could improve PCR sensitivity. Using a culture enrichment step prior to amplification Teh
et al. demonstrated that the 5 hour broth culture enrichment improved PCR sensitivity by 10 times for spiked blood, and 100 times for spiked stool samples [
23]. We have recently developed a blood culture PCR assay which could detect as few as 0.75 bacteria per millilitre of blood following 3 hours of culture enrichment [
22]. But like conventional blood culture, the PCR assay following bacterial culture enrichment still has some practical limitations including availability of culture facilities in disease endemic areas and bacteria being unculturable due to the use of antibiotics. Therefore, a PCR assay method that can be used directly on clinical blood samples is still a better choice. Collecting and then extracting DNA from a large volume of blood may theoretically improve PCR assays, but the presence of dominating human DNA causes false-positive PCR signals due to the non-specific binding of primers and false-negative results due to reduced sensitivity. Currently two protocols have been commercially marketed for reducing the presence of dominating human DNA:MolYsis (Molzym GmbH & Co. KG, Bremen, Germany) and Pureprove (SIRS-Lab GmbH, Jena, Germany). The former uses a chaotropic buffer containing guanidine hydrochloride to selectively lyze human cells. The released eukaryotic DNA was subsequently degraded by the addition of a chaotropic resistant nuclease, and the intact bacterial cells were collected for bacterial DNA purification. For the latter, total genomic (i.e. human and bacterial) DNA is first conventionally extracted specific bacterial genomic DNA is then isolated using a DNA binding protein that recognizes unmethylated CpG motifs predominantly present in bacterial and fungal genomes at significantly lower frequencies than in human genomes. Horz
et al. recently compared these two protocols for DNA preparation, and found that both were able to substantially reduce the human background DNA in most of the cases but loss of bacterial DNA was also a problem [
24]. For example, with MolYsis the recovery of
P. gingivalis in periodontal samples was less than 20%. As mentioned above, the MolYsis protocol uses a guanidine hydrochloride chaotropic buffer for lysis of human/animal cells, and bacteria with a thin or labile cell wall (e.g. members of the genus
Treponema) or those exposed to cell wall-active antibiotics and/or human immunosystems or bacteria even devoid of a cell wall (e.g.
Mycoplasma,
Chlamydia) are also lyzed. Therefore, the MolYsis protocol may not be suitable for use in the diagnosis of Gram-negative sepsis. They also found, on the other hand, the Pureprove technology was not efficient in removal of human DNA. This is because it is based on selective binding of bacterial DNA to a DNA binding protein that recognizes unmethylated CpG motifs. CpG islands and motifs are not distributed equally over the entire human genome, and fragments without any 5′-methylcytosin are present in every preparation and will mix with prokaryotic DNA. In addition, as cytosine methylation is negatively correlated with gene expression, in regions of active genes the CpG dinucleotides-motifs are also de-methylated and, if fragmentized, could possibly mix with prokaryotic DNA as well. It should be emphasized that in Horz
et al. study, the mean absolute bacterial gene copy numbers was 2.23×10
10 (saliva and supragingival plaque collected by cotton tampon) and 1.37x10
10 (pooled subgingivalsulcus fluid) [
24]. Such a high level of bacterial template in a blood sample from a patient with typhoid is unlikely. Therefore, the fact that typical clinical typhoid samples contain <1 organism/ml of blood highlights the big challenges faced for developing PCR diagnostics for typhoid.
Bile is a widely used ingredient in culture media for blood culture and the bile resistance of
Salmonella is well-known. This could be exploited for developing typhoid diagnostics as bile can selectively lyze human blood cells and release both human DNA and intracellular bacteria without damaging bacterial cells [
22,
25]. In the present study, we have investigated the use of ox bile and micrococcal nuclease for selective lysis of blood cells and removal of background human DNA. We tested the use of micrococcal nuclease for removal of the released human DNA in the presence and absence of ox bile, and found that micrococcal nuclease is bile resistant and active at a wide range of bile concentrations used up to 9%. This suggests that micrococcal nuclease could be used to degrade human DNA while ox bile is added for the lysis of blood cells, potentially simplifying the design of typhoid diagnostics. We have tried two STEM protocols for removal of background human DNA and confirmed that both methods can substantially reduce background human DNA in the DNA preparations. As a result, the percentage of bacterial DNA in the preparation was increased. When the DNA prepared by STEM was used for PCR assay, we have consistently found that PCR sensitivity increased by at least 1,000 fold, compared with that using conventionally prepared DNA (which contained a large amount of background human DNA). These promising results indicate that STEM should now be investigated in field trials and to this end the approach is currently being applied in an endemic setting. Although the present study aims to improve current diagnostics for typhoid, the method described herein can be generally applied to sample preparation for diagnosis of infections with other ox bile resistant bacterial and fungal pathogens, in particular, where the low level of a pathogen presents a problem for current molecular diagnostics.
In conclusion, we describe herein a novel method, which we have termed “the selective target DNA enrichment method” (STEM). STEM can enrich bacteria from large volume samples, remove background human DNA, and enhance the sensitivity of PCR detection for Salmonella Typhi in blood samples. Therefore, this novel method of sample preparation offers a better option for improved typhoid PCR assays directly using clinical specimens in diagnosis of this globally important infection disease which we believe could be of importance in improving clinical care and providing effective evaluation of novel vaccines.