INTRODUCTION
Entamoeba histolytica is the causative agent of amoebiasis. Symptomatic intestinal or liver infection affects about 50 million individuals mainly in under-developed countries, resulting in 40–100 thousand deaths per year (WHO, 1997; Petri et al. Reference Petri, Haque, Lyerly and Vines2000). Asymptomatic infections may occur and, although their frequency is unknown, most of them have been attributed to E. dispar, a non-invasive species which is morphologically similar to E. histolytica (Diamond and Clark, Reference Diamond and Clark1993; WHO, 1997).
The absence of pathognomonic clinical signs for amoebiasis, the possibility of asymptomatic infections having an invasive course, and the treatment-associated risks make the differential diagnosis of these species relevant. In this context, special attention is given to identifying lectins and species-specific gene sequences (Mirelman et al. Reference Mirelman, Nuchamowitz and Stolarsk1997; Haque et al. Reference Haque, Mollah, Ali, Alam, Eubanks, Lyerly and Petri2000; Petri et al. Reference Petri, Haque, Lyerly and Vines2000; Tanyuksel and Petri, Reference Tanyuksel and Petri2003). Gal/GalNac lectin has been widely used as the target in ELISA. However, the sensitivity of the method is not so good when compared to PCR (Roy et al. Reference Roy, Kabir, Mondal, Ali, Petri and Haque2005). Several candidate gene sequences for differential diagnosis between E. histolytica and E. dispar have been reported (Tannich et al. Reference Tannich, Bruchhaus, Walter and Horstmann1991; Troll et al. Reference Troll, Marti and Weiss1997; Zengzhu et al. Reference Zengzhu, Bracha, Nuchamowitz, Cheng-I and Mirelman1999; Khairnar and Parija, Reference Khairnar and Parija2007). Obtaining high-quality DNA, however, is a crucial stage in PCR of fecal samples. Preservatives, fecal substances and even the polysaccharides of cyst walls can act as inhibitors in the sample, contributing to reducing the sensitivity of the technique. Some commercial kits developed to reduce this interference are available. Nevertheless, they are expensive for routine use in under-developed countries, where the disease is more common. In the present work, a new simple, fast and cheap technique of DNA extraction from fecal samples followed by a PCR for an episomal target are presented as an alternative to detect and differentiate E. histolytica and E. dispar in feces.
MATERIALS AND METHODS
Cultures
Five E. histolytica strains (ICB-452, HM1-IMSS, ICB-32, HK-9, and EGG) and 5 E. dispar strains (VEJ, JCAO, JPPC, WIL, and ADO), previously identified through zymodeme (Barral de Martinez et al. Reference Barral de Martinez, Gomes, Viana, Romanha and Silva1996) and PCR (Gomes et al. Reference Gomes, Pesquero, Furst, Valle, Pesquero and Silva1999), were prepared for direct DNA extraction or for spiking fecal samples to be used as control. Trophozoites were harvested from culture tubes, washed twice in PBS (Phosphate-buffered saline) pH 7·2, and then quantified in a Neubauer chamber.
Fecal samples
Fecal samples were obtained from laboratories in the metropolitan area of Belo Horizonte and day-care centres in Uberlândia, both in the state of Minas Gerais, Brazil. Fecal material was submitted to passive sedimentation (Hoffman et al. Reference Hoffman, Pons and Janer1934) or formalin-ether concentration techniques (Young et al. Reference Young, Bullock, Melvin and Spruill1979) for subsequent microscopic evaluation. The sediments with cysts of E. histolytica/E. dispar were each washed once with PBS pH 7·2, and the resulting pellet suspended in 0·5 ml of PBS. Then 100 μl of the sediment were transferred to 1·5 ml microtubes for subsequent DNA extraction. The samples obtained from formalin-ether concentration were additionally washed twice in PBS pH 7·2 (13 000 g for 1 min) to remove the excess formalin and ether before DNA extraction.
Two sediments that were negative for all enteroparasites were processed as described above and used as negative control samples for fecal DNA extractions. To obtain positive control samples, aliquots of the negative sediments were spiked with 5×102 trophozoites of E. histolytica or E. dispar. Ten spiked sediments were prepared for each method of DNA extraction.
DNA isolation
DNA extractions were performed with the phenol-chloroform method, QIAamp DNA stool minikit (QIAGEN), and with an adapted method of alkaline lysis (Sambrook et al. Reference Sambrook, Fritsch, Maniatis and Nolan1989). Each method was tested in 10 samples containing only trophozoites, 10 trophozoite-spiked fecal sediments, 39 fecal sediments microscopically positive for E. histolytica/E. dispar and 2 negative controls.
In the phenol-chloroform protocol, samples were previously subjected to 3 cycles of freezing in liquid nitrogen and thawing in a thermomixer at 99°C, and then 100 μl of Triton X-100 were added. After incubation at 90°C for 2 h, proteinase K (5 mg/ml) was added, and the sample was incubated again at 60°C for 2 h. DNA was extracted with phenol-chloroform (Ausubel et al. Reference Ausubel, Brent, Kingston, Moore, Seidman, Smith and Truhl1994).
DNA extracted with the QIAamp DNA stool minikit (QIAGEN) was performed after the same procedure of freezing and thawing described above, following the manufacturer's instructions.
In the alkaline lysis method, each sample was previously incubated with 100 μl of 1% polyvinylpolypyrrolidone at 99°C for 10 min. Then 200 μl of lysis buffer (50 mm glucose, 25 mm Tris-HCl, pH 8·0, 10 mm EDTA) were added, followed by incubation for 5 min at room temperature. A mixture of 400 μl of NaOH 0·2 m in 1% of SDS was added to the microtube, and 3 cycles of freezing in liquid nitrogen and thawing in a thermomixer at 99°C were performed again. DNA was extracted by adding 300 μl of 7·5 m ammonium acetate and precipitated by adding 0·6 volumes of isopropanol. The DNA pellet was washed in 75% ethanol and re-suspended in 20 μl of ultra-pure sterilized water.
When not used immediately for PCR, the DNA extracted in any of the protocols was stored at −20°C.
Primers and PCR conditions
A 308 bp-gene segment of SSU rRNA was used as the PCR target (Blessmann et al. Reference Blessmann, Buss, Nu, Dinh, Ngo, Van, Abd-Alla, Jackson, Ravdin and Tannich2002). The segments have similar sequences in both Entamoeba species, but present differences that permitted delineation of a specific sense primer for E. histolytica (5′GTACAAAATGGCCAATTCATTCAACG) and a specific sense primer for E. dispar (5′GTACAAAGTGGCCAATTTATGTAAGCA). The anti-sense primer is the same for both species (5′GAATTGATTTTACTCAACTCTAGAG). Each specific PCR was carried out in a 10 μl volume containing 1 μl of 10X PCR buffer (100 mm Tris-HCl, pH 8·4, 500 mm KCl, 15 mm MgCl2), 0·2 U of Taq DNA polymerase (Phoneutria), 5 pmol of each primer, 2·5 mm dNTPs, and 3 μl of fecal DNA. Reactions were performed with 50 cycles of 95°C for 30 s; 52°C for 30 s and 72°C for 45 s. Amplification products were subjected to electrophoresis in 4% polyacrylamide gel and stained with silver.
Alkaline-lysis processed samples that were initially negative in PCR were re-submitted to the reaction after diluting the templates 10 times and adding PCR facilitator BSA (10 μg) to the reaction.
To evaluate the presence of PCR inhibitors in other negative samples, the PCR was repeated with templates spiked with DNA from 5×102 trophozoites.
RESULTS
All 10 DNA templates obtained directly from 5×102 trophozoites yielded positive amplification of SSU rRNA products in PCRs, regardless of the method used for extraction. The same efficiency was observed when the DNA used in PCR was obtained from fecal sediments spiked with 5×102E. histolytica/E. dispar trophozoites and extracted by alkaline lysis or QIAamp kit. On the other hand, the phenol-chloroform method presented very low efficiency, since only 2 out of 10 spiked fecal sediments submitted to it presented signs of amplification after the PCR. Therefore, phenol-chloroform was not used for further DNA extraction using fecal sediments containing cysts.
Fecal sediments processed by the QIAamp DNA stool mini kit or alkaline lysis protocol provided DNA samples that yielded satisfactory results in PCR (Fig. 1). Out of 39 fecal sediments microscopically positive for E. histolytica/E. dispar cysts, 35 were positive in PCR with DNA extracted by at least one of the methods. Thirty-four QIAamp-processed samples were identified as E. dispar and only 1 as E. histolytica. The 4 negative samples yielded positive results in PCR after they had been spiked with DNA of E. histolytica.
Polyacrylamide gel electrophoresis of PCR products of DNA extracted from 6 different fecal samples of Entamoeba dispar. Lanes 1-6, PCR carried out with DNA extracted with the QIAamp DNA stool minikit. Lanes 7-12, PCR carried out with DNA extracted by the alkaline lysis method. C, negative control; M, molecular marker 100 bp DNA ladder.
PCR using 39 DNA samples extracted by the alkaline lysis identified 31 as E. dispar and 1 sample as E. histolytica. However, 20 of them were positive only after the dilution of the DNA template and addition of BSA. Three samples that remained negative were spiked with DNA of E. histolytica, and no amplification was obtained whatsoever.
No differences in PCRs results were observed among DNA templates obtained from fecal sediments processed by passive sedimentation or formalin-ether concentration.
DISCUSSION
After E. dispar was recognized as an organism that infects humans, the improvements of the techniques to differentiate this species from E. histolytica gained an increasing importance (WHO, 1997). Identification of gene sequences through PCR is a promising method due to its sensitivity and specificity (Mirelman et al. Reference Mirelman, Nuchamowitz and Stolarsk1997; Troll et al. Reference Troll, Marti and Weiss1997; Paglia and Visca, Reference Paglia and Visca2004; Qvarnstrom et al. Reference Qvarnstrom, James, Xayavong, Holloway, Visvesvara, Sriram and Silva2005). Obtaining high-quality DNA is fundamental in the PCR of fecal samples. Most methodologies described are laborious or use kits that make diagnosis very expensive.
In the present work, we evaluated a modified method of alkaline lysis to extract fecal DNA and amplify an episomal target to identify E. histolytica and E. dispar. Two additional methods for DNA extraction, phenol-chloroform and QIAamp DNA stool mini kit, were also tested for comparison.
The use of the phenol-chloroform method to process fecal sediments spiked with trophozoites was not satisfactory, since most of the samples were negative in PCR. Nevertheless, DNA from cultures isolated by this method has yielded positive amplifications of the investigated targets. This result suggests that the DNA extracted from feces was impure, carrying PCR-inhibiting substances that could not be eliminated by the method.
The QIAamp DNA stool mini kit has been used in several studies that describe identification of E. histolytica and E. dispar by PCR (Blessmann et al. Reference Blessmann, Buss, Nu, Dinh, Ngo, Van, Abd-Alla, Jackson, Ravdin and Tannich2002; Paglia and Visca, Reference Paglia and Visca2004; Roy et al. Reference Roy, Kabir, Mondal, Ali, Petri and Haque2005). In the present study, this method yielded amplifications in 35 of the 39 positive samples. The 4 negative samples spiked with DNA from E. histolytica yielded positive amplifications, indicating the absence of inhibitors. This result suggests that cysts in fecal sediments may belong to another species rather than E. histolytica/E. dispar or that the amount of parasites has not reached the detection threshold for the conventional PCR, which is less sensitive than the real-time PCR (Qvarnstrom et al. Reference Qvarnstrom, James, Xayavong, Holloway, Visvesvara, Sriram and Silva2005; Roy et al. Reference Roy, Kabir, Mondal, Ali, Petri and Haque2005).
Real-time PCR was the technique originally used with the primers described here (Blessmann et al. Reference Blessmann, Buss, Nu, Dinh, Ngo, Van, Abd-Alla, Jackson, Ravdin and Tannich2002). However, the cost limits their use in under-developed countries, and the same can be said about the kits for DNA extraction. Because of that, we propose a cheaper extraction method and the conventional PCR whose amplification target is present in 200 copies, contributing to enhance the efficiency of detection.
We adapted the plasmid DNA extraction method of alkaline lysis (Sambrook et al. Reference Sambrook, Fritsch, Maniatis and Nolan1989) in order to isolate fecal DNA. At the present time, the efficiency of the alkaline lysis protocol was similar to the extraction performed with the kit. Interestingly, some alkaline lysis-processed samples resulted in more intense bands at electrophoresis than the products obtained in PCR using QIAamp-processed templates. The high concentration of NaOH used in this technique possibly enhances cyst rupture besides facilitating extraction of rDNA, which is extra-chromosomal in E. histolytica (Bhattacharya et al. Reference Bhattacharya, Som and Bhattacharya1998). Furthermore, the protocol included a PVPP treatment step, a procedure that contributes to the removal of potential inhibitors of PCR present in feces (Anderson and Lebepe-Mazur, Reference Anderson and Lebepe-Mazur2003).
However, 23 alkaline lysis-processed samples were negative for PCR, suggesting interference by inhibitors. They were then diluted 10 times and BSA was added into the PCR reaction. These procedures dramatically decreased the inhibitory effect, and amplification occurred in 20 of the previously negative samples. In the other 3 samples, spiking with DNA of E. histolytica did not result in PCR positivity, indicating that inhibitors could not be eliminated. In spite of this, PCR using DNA extracted with alkaline lysis presented a good concordance (positive=91·4% and negative=100%) with PCR using DNA extracted with the kit.
In preliminary studies, the QIAamp DNA stool minikit and alkaline lysis method were used without the initial freezing/thawing procedure, and a low PCR yield was observed (data not shown). Microscopic analysis of fecal material after the treatment for lysis revealed the presence of intact cysts, and thus the initial freezing and thawing step was included in all protocols. After that, an increased number of positive reactions for PCR were observed. It is likely that the amounts of DNA and trophozoites in the samples used were not enough or, when present, DNA might have been degraded by the cyst concentration process. To evaluate if cyst rupture could be accomplished with dry ice, it was used as an alternative to liquid nitrogen in 7 fecal sediments treated by the alkaline lysis method (data not shown). Microscopic examination also confirmed cyst rupture, and PCR amplified the target segment as well. Thus, rapid freezing at −80°C is as efficient for cyst rupture as liquid nitrogen, and can be used in laboratories where liquid nitrogen is not available.
Concerning the method for concentrating fecal sediments, our findings indicated that formalin-ether concentration can be used to process feces before DNA extraction, since the PCR of the samples prepared by this method presented the same positivity as those prepared by passive sedimentation. The possibility of DNA degradation by formalin, as suggested by some authors (Ramos et al. Reference Ramos, Zurabian, Morán, Ramiro, Gómez, Clark, Melendro, García and Ximénez1999), was minimized in our experiments, since the contact with the substance was short (2 min). Furthermore, the pre-washing treatment before DNA extraction must have contributed to the elimination of the inhibitory effect of the formalin. Formalin-ether concentration is faster, produces a cleaner material, and allows a better concentration of cysts for microscopic analysis. These advantages can reduce the time for processing fecal material for DNA extraction and PCR identification.
In summary, the alkaline lysis protocol described here is a feasible method for fecal DNA extraction and, when combined with PCR to amplify the SSU rRNA target, could be an alternative method of moderate cost for the identification of E. histolytica and E. dispar in feces.
