Entamoeba histolytica modulates a complex repertoire of novel genes in response to oxidative and nitrosative stresses: implications for amebic pathogenesis


  • João B. Vicente,

    1. Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, CA 94305-5107, USA.
    2. Instituto de Tecnologia Química e Biológica, Universidade Nova de Lisboa, Avenue da República (EAN), 2781-901 Oeiras, Portugal.
    Search for more papers by this author
  • Gretchen M. Ehrenkaufer,

    1. Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, CA 94305-5107, USA.
    Search for more papers by this author
  • Lígia M. Saraiva,

    1. Instituto de Tecnologia Química e Biológica, Universidade Nova de Lisboa, Avenue da República (EAN), 2781-901 Oeiras, Portugal.
    Search for more papers by this author
  • Miguel Teixeira,

    1. Instituto de Tecnologia Química e Biológica, Universidade Nova de Lisboa, Avenue da República (EAN), 2781-901 Oeiras, Portugal.
    Search for more papers by this author
  • Upinder Singh

    Corresponding author
    1. Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, CA 94305-5107, USA.
    2. Division of Infectious Diseases, Department of Internal Medicine, Stanford University, Stanford, CA 94305-5107, USA.
      *E-mail usingh@stanford.edu; Tel. (+1) 650 723 4045; Fax (+1) 650 724 3892.
    Search for more papers by this author

*E-mail usingh@stanford.edu; Tel. (+1) 650 723 4045; Fax (+1) 650 724 3892.


Upon host infection, the protozoan parasite Entamoeba histolytica is confronted with reactive oxygen and nitrogen species and must survive these stresses in order to cause invasive disease. We analysed the parasite's response to oxidative and nitrosative stresses, probing the transcriptional changes of trophozoites of a pathogenic strain after a 60 min exposure to H2O2 (1 mM) or a NO donor (dipropylenetriamine-NONOate, 200 μM), using whole-genome DNA microarrays. Genes encoding reactive oxygen and nitrogen species detoxification enzymes had high transcriptional levels under basal conditions and upon exposure to both stresses. On a whole-genome level, there was significant modulation of gene expression by H2O2 (286 genes regulated) and dipropylenetriamine-NONOate (1036 genes regulated) with a significant overlap of genes modulated under both conditions (164 genes). A number of transcriptionally regulated genes were in signalling/regulatory and repair/metabolic pathways. However, the majority of genes with altered transcription encode unknown proteins, suggesting as yet unraveled response pathways in E. histolytica. Trophozoites of a non-pathogenic E. histolytica strain had a significantly muted transcriptional response to H2O2 compared with the pathogenic strain, hinting that differential response to oxidative stress may be one factor that contributes to the pathogenic potential of E. histolytica.


Infection by Entamoeba histolytica, the causative agent of amoebiasis, is a global health problem, which affects 500 million people worldwide (Stanley, 2003). Most commonly, this pathogen causes haemorrhagic dysentery and liver abscesses. During tissue invasion, E. histolytica adapts to changing oxygen tensions as it goes from the anaerobic colonic lumen to an oxygen-rich environment in the colonic tissue (Stanley, 2003). Additionally, the parasite must cope with cytotoxic reactive oxygen species (ROS) and reactive nitrogen species (RNS) that are produced and released by activated phagocytes that are attracted to the site of infection (MacMicking et al., 1997; Bogdan et al., 2000; Stanley, 2003). Therefore, a significant contribution to E. histolytica's pathogenic potential is likely to be due to its ability to cope with oxidative and nitrosative stresses generated during tissue invasion.

The cellular components targeted by ROS and RNS include proteins (metal cofactors, thiolate side-chains, tyrosine and methionine residues), nucleic acids and lipids (Halliwell and Gutteridge, 2007). Common defence strategies against oxidative and nitrosative stresses include detoxification enzymes and repair systems that enable cells to resist RNS and ROS (Justino et al., 2005; Vandenbroucke et al., 2008). Several microbial transcription factors and regulons, which are involved in the response to both oxidative and nitrosative stresses as well as in the transition from anaerobic metabolism to aerobic metabolism, have redox-sensitive active sites that are modified and/or damaged by both ROS and RNS. Not surprisingly, given the overlap in the types of damage caused by ROS and RNS, common mechanisms exist to deal with these stressors. These include the Crp-Fnr superfamily of transcriptional regulators, which respond both to nitrosative and oxidative stresses (Korner et al., 2003), as other transcriptional regulators do, e.g. in Escherichia coli: NsrR, OxyR, SoxRS, MetR, ferric uptake regulator and NorR, regulating a wide range of cellular processes (Spiro, 2006). A survey of the genomes of the parasitic protists E. histolytica (Loftus et al., 2005), Giardia lamblia (Morrison et al., 2007) and Trichomonas vaginalis (Carlton et al., 2007) revealed the absence of homologues of any of the above-mentioned transcriptional regulators. In contrast, genes coding for detoxification systems for ROS and RNS are present in the genomes of these anaerobic protists. Some of these genes may have been acquired by lateral gene transfer from prokaryotes (Andersson et al., 2003; 2006). The E. histolytica genome has four genes encoding flavodiiron proteins (FDPs), enzymes endowed with oxygen and/or nitric oxide reductase activity that are widespread in prokaryotes (Saraiva et al., 2004; Kurtz, 2007), and have been studied in the protozoa T. vaginalis (Sarti et al., 2004) and G. lamblia (Di Matteo et al., 2008). E. histolytica's genome also contains genes encoding other enzymes involved in the detoxification of ROS, including peroxiredoxin, rubrerythrin, hybrid-cluster protein and superoxide dismutase (SOD). Peroxiredoxin constitutes a major defence against oxidative stress as it is induced by a high-oxygen environment (Akbar et al., 2004) and trichostatin A (Isakov et al., 2008), and contributes to E. histolytica's virulence (Davis et al., 2006; Sen et al., 2007). Although peroxiredoxin and SOD are ubiquitous in all domains of life, FDPs, rubrerythrin and hybrid-cluster proteins have thus far been identified only in prokaryotes and in these anaerobic protists.

Whole-genome expression profiling has been used to assess the effects of oxidative and nitrosative stresses in diverse eukaryotes and prokaryotes (Thum and Bauersachs, 2007; Vandenbroucke et al., 2008). A recent meta-analysis of microarray data performed to assess the common denominators in the oxidative stress response across different domains of life revealed that there are both strong species-specific responses and common strategies for diverse organisms to cope with this challenge (Vandenbroucke et al., 2008). Microarray technology has been used in Entamoeba to investigate a wide variety of biological questions, including virulence (MacFarlane and Singh, 2006; Davis et al., 2007), host colonic and hepatic invasion (Gilchrist et al., 2006; Santi-Rocca et al., 2008) and development (Ehrenkaufer et al., 2007). In order to determine the molecular mechanisms by which E. histolytica trophozoites respond when challenged with oxidative and nitrosative stresses, we used whole-genome expression profiling using a short oligonucleotide microarray containing 9435 of the annotated 9938 amebic genes. Our results demonstrated a significant transcriptional response of E. histolytica HM-1:IMSS, a canonical virulent strain, to H2O2 (286 genes regulated), NO (1036 genes regulated) and a significant overlap among the genes responsive to both conditions (164 genes). To further identify which components of these response mechanisms may be correlated with E. histolytica's virulence potential, the response to oxidative stress was assessed for a non-pathogenic strain, E. histolytica Rahman. In contrast to the observations for the virulent HM-1:IMSS strain, the Rahman strain had fewer transcriptional changes and the overall fold-changes for the regulated genes were significantly lower. Overall, our results provide insights into the molecular network regulating adaptation to oxidative and nitrosative stresses in E. histolytica and suggest that one important difference between virulent and non-virulent amebae is their ability to deal with the stresses encountered during host invasion.

Results and discussion

Sensitivity of E. histolytica trophozoites to oxidative and nitrosative stresses

To determine the response of E. histolytica to nitrosative and oxidative stresses, trophozoites from the HM-1:IMSS strain were exposed to dipropylenetriamine (DPTA)-NONOate (nitric oxide releaser) or hydrogen peroxide (H2O2). We analysed the viability of E. histolytica trophozoites in varying concentrations of H2O2 and DPTA-NONOate in order to identify conditions in which E. histolytica trophozoites stressed but still ≥ 90% viable. A 1 h exposure to 1 mM H2O2 or 200 μM DPTA-NONOate resulted in a significant fraction of the cells being stressed (as judged by rounded morphology), but only a few dead cells (≤ 5% and ≤ 10% respectively), based on Trypan blue staining (data not shown). We tested the sensitivity of both the virulent HM-1:IMSS and the non-virulent Rahman strains to 1 mM H2O2, and found a similar percentage of dead cells (data not shown). The two stress agents are differently released into the cultures: addition of hydrogen peroxide results in immediate exposure to the added concentration, whereas DPTA-NONOate is a slow releaser of nitric oxide. The tested concentration of H2O2 (1 mM) is within physiologically relevant concentrations found in the gastrointestinal lumen (Mayol et al., 2006). At the chosen concentration (200 μM), DPTA-NONOate releases NO at a rate of ∼25 nM NO s−1, which after 1 h of exposure would accumulate to a maximum of ∼82 μM of NO in the medium, a concentration above physiological concentrations (Halliwell and Gutteridge, 2007). It is, however, likely that the NO concentration experienced by E. histolytica under the tested conditions is lower than that released, due to the general reactivity of nitric oxide within biological milieu. Both NO and H2O2 may react with components of the TYI-S-33 medium, such as serum components and cysteine, in which E. histolytica is cultured. However, no defined medium lacking serum or cysteine has been developed in which E. histolytica can be reliably grown (Gillin and Diamond, 1980; Diamond and Cunnick, 1991). Thus, the TYI-S-33 medium, in which parasite growth is well standardized and robust and in which many of the previous transcriptome analyses done previously have been performed, is the best option for our studies.

Expression analysis of E. histolytica strains challenged with oxidative or nitrosative stress

Affymetrix whole-genome microarrays were used to determine the global transcriptional changes in E. histolytica HM-1:IMSS trophozoites upon exposure to 1 mM hydrogen peroxide for 1 h. Data from three independent H2O2-exposed cultures were compared with those from E. histolytica HM-1:IMSS trophozoites from standard axenic culture conditions. All sets of arrays displayed high correlation values (0.94–0.98) (Table 1). Genes with a twofold or greater change and false discovery rate (FDR) < 0.05 were considered differentially expressed: 184 genes were upregulated by H2O2 and 102 genes were downregulated (Tables 2 and 3; Table S2).

Table 1.  (A) Correlations between individual DNA microarrays for E. histolytica HM-1:IMSS and Rahman under each experimental condition; (B) averaged correlations between DNA microarrays for each experimental condition of E. histolytica HM-1:IMSS. Thumbnail image of
Table 2.  Genes upregulated by hydrogen peroxide in E. histolytica HM-1:IMSS.
Probe IDAccession numberDescriptionBaseline expression levelFold-changeP-valueRegulated under DPTARegulated under other conditions
  1. The probe ID, accession number, description, baseline expression level, fold-change, P-value and regulation under DPTA or other conditions are shown. The 30 most highly regulated genes are listed. HS, heat shock, adapted from Hackney et al. (2007); cysts, E. histolytica cysts, adapted from Ehrenkaufer et al. (2007).

879.m00008_atXM_642785Hypothetical protein0.07189.60.000
194.m00102_s_atXM_645780Hypothetical protein0.06127.10.001+HS
654.m00032_x_atXM_642891Hypothetical protein1.39114.90.023+
363.m00049_x_atXM_643869Hypothetical protein0.1789.050.015+HS
256.m00084_x_atXM_644865Hypothetical protein0.1286.640.014+HS
256.m00083_x_atXM_644864Hypothetical protein1.2282.850.014+
692.m00024_s_atXM_642872Hypothetical protein0.0576.140.000+
248.m00060_s_atXM_644979Hypothetical protein0.0571.780.000+
266.m00066_s_atXM_644755Hypothetical protein0.1353.250.002+
194.m00123_x_atXM_645776dUTP diphosphatase, putative0.1150.950.011+
397.m00055_s_atXM_643659Hypothetical protein0.1842.560.020+
654.m00031_s_atXM_642894Hypothetical protein0.0641.050.000+HS
397.m00062_x_atXM_643664Hypothetical protein1.3537.880.034+
692.m00026_s_atXM_642873Hypothetical protein0.0635.780.001+
711.m00021_x_atXM_642861Hypothetical protein0.1033.680.018+HS
375.m00057_x_atXM_643790Deoxyuridine 5′-triphosphate nucleotidohydrolase0.0521.750.001+
397.m00054_x_atXM_643658Hypothetical protein0.0615.740.001+
344.m00044_atXM_644000Hypothetical protein0.2415.10.025+
245.m00039_x_atXM_645025Hypothetical protein0.4714.390.026+Cysts
219.m00107_atXM_645351Cell division control protein 42, putative0.1011.340.003HS
267.m00069_atXM_644746Hypothetical protein0.1011.20.034
358.m00033_atXM_643905Hypothetical protein0.6411.050.012+Cysts
194.m00103_atXM_645781Hypothetical protein0.0710.440.005+
8.m00393_atXM_651687Late competence protein, putative0.258.850.027HS + cysts
356.m00029_s_atXM_643911Hypothetical protein0.187.680.021+HS
20.m00272_atXM_650868Conserved hypothetical protein0.136.830.046+
36.m00211_atXM_650002Hypothetical protein0.096.780.009+Cysts
35.m00253_atXM_650038Iron-sulfur flavoprotein, putative2.166.70.006Cysts
312.m00037_atXM_644279Iron-sulfur flavoprotein, putative0.996.620.005HS + cysts
1.m00705_atXM_652401BspA-like leucine-rich repeat protein, putative0.276.560.004HS
Table 3.  Genes downregulated by hydrogen peroxide in E. histolytica HM-1:IMSS.
Probe IDAccession numberDescriptionBaseline expression levelFold-changeP-valueRegulated under DPTARegulated under other conditions
  1. The probe ID, accession number, description, baseline expression level, fold-change, P-value and regulation under DPTA or other conditions are shown. The 30 most highly regulated genes are listed. HS, heat shock, adapted from Hackney et al. (2007); colitis, ameba from a mouse model of amebic colitis, adapted from Gilchrist et al. (2006); trophs, E. histolytica trophozoites, adapted from Ehrenkaufer et al. (2007).

172.m00078_atXM_646199Hypothetical protein0.86−9.090.000+HS
537.m00017_x_atXM_643063AIG1 family protein, putative4.58−7.410.039+
2.m00528_atXM_652349Hypothetical protein0.35−6.80.019+HS
3.m00563_atXM_652186Hypothetical protein6.39−5.750.002+
86.m00158_s_atXM_648244Hypothetical protein0.46−5.590.039+HS
226.m00092_atXM_645246Rab family GTPase0.26−4.830.000+
233.m00105_atXM_645153Hypothetical protein6.57−4.410.004+
9.m00372_x_atXM_651612Hypothetical protein0.32−4.180.004+HS
87.m00165_atXM_648195Hypothetical protein0.21−4.130.009+
87.m00154_atXM_648212Formate nitrite transporter family protein, putative0.78−4.020.001Colitis
31.m00209_x_atXM_650257Conserved hypothetical protein0.25−3.790.019+
66.m00150_atXM_648887High mobility group protein, putative0.39−3.770.006+HS
343.m00064_x_atXM_644015WD repeat protein0.21−3.760.039+
125.m00091_atXM_647175Cyclin, putative0.71−3.690.006+HS
286.m00057_atXM_644512Mitotic inducer phosphatase, putative0.21−3.310.000+HS
37.m00216_atXM_649962Hypothetical protein0.79−3.260.018+
56.m00175_atXM_649238Elongation factor 1 beta, putative12.26−3.150.038
46.m00250_atXM_649627Hypothetical protein3.53−3.070.005
19.m00301_atXM_650878Hypothetical protein2.32−3.030.015+HS
408.m00045_s_atXM_643583Putative GTPase2.43−3.020.006+HS
80.m00159_atXM_648429Hypothetical protein7.76−2.990.004+
1.m00606_s_atXM_652473Hypothetical protein9.86−2.910.006+HS
40.m00246_s_atXM_649855rRNA biogenesis protein, putative4.83−2.890.041+HS
95.m00147_atXM_647953PfkB family carbohydrate kinase, putative6.85−2.890.005HS
51.m00159_s_atXM_649424Hypothetical protein0.30−2.880.031+
221.m00089_s_atXM_645315Hypothetical protein3.74−2.850.027+
127.m00147_atXM_647143Predicted protein1.33−2.820.019+HS
4.m00641_atXM_652060tRNA-specific adenosine deaminase, putative13.13−2.80.040+
34.m00273_atXM_650116Rab family GTPase4.84−2.770.048Trophs
13.m00334_atXM_651296Conserved hypothetical protein2.85−2.670.027+Trophs

Entamoeba histolytica HM-1:IMSS trophozoites were also subjected to nitrosative stress by addition of the NO releaser DPTA-NONOate (200 μM, 1 h) and resulting transcriptional changes assayed. Three arrays from independent parasite cultures challenged with DPTA-NONOate were performed; all arrays displayed good correlation values [0.94–0.97 (Table 1)]. Although E. histolytica displayed similar percentages of cell death in H2O2 compared with NO, a substantially greater number of genes were transcriptionally regulated by nitrosative stress compared with oxidative stress. Using the same statistical criteria applied above (twofold or greater change and FDR < 0.05), 443 genes were upregulated and 593 genes downregulated by DPTA-NONOate (Tables 4 and 5, and Table S3). To confirm the array results, seven genes were selected for semiquantitative RT-PCR analysis. In every case the data from the microarray analyses were confirmed (Fig. 1).

Table 4.  Genes upregulated by DPTA-NONOate in E. histolytica HM-1:IMSS.
Probe IDAccession numberDescriptionBaseline expression levelFold-changeP-valueRegulated under other conditions
  1. The probe ID, accession number, description, baseline expression level, fold-change, P-value and regulation under other conditions are shown. The 30 most highly regulated genes are listed. HS, heat shock, adapted from Hackney et al. (2007); colitis, ameba from a mouse model of amebic colitis, adapted from Gilchrist et al. (2006); cysts, E. histolytica cysts, adapted from Ehrenkaufer et al. (2007).

194.m00102_s_atXM_645780Hypothetical protein0.06380.20.009HS
363.m00049_x_atXM_643869Hypothetical protein0.17305.50.002
256.m00084_x_atXM_644865Hypothetical protein0.12252.80.002
248.m00060_s_atXM_644979Hypothetical protein0.05240.90.006
692.m00024_s_atXM_642872Hypothetical protein0.05219.50.003
194.m00123_x_atXM_645776dUTP diphosphatase, putative0.11191.50.003
266.m00066_s_atXM_644755Hypothetical protein0.13176.20.004
654.m00032_x_atXM_642891Hypothetical protein1.39162.90.015
654.m00031_s_atXM_642894Hypothetical protein0.06130.40.008
256.m00083_x_atXM_644864Hypothetical protein1.22123.60.006
397.m00055_s_atXM_643659Hypothetical protein0.18119.30.004
692.m00026_s_atXM_642873Hypothetical protein0.06107.70.010
375.m00057_x_atXM_643790Deoxyuridine 5′-triphosphate nucleotidohydrolase0.0582.540.020
711.m00021_x_atXM_642861Hypothetical protein0.1081.620.006
397.m00062_x_atXM_643664Hypothetical protein1.3558.470.019
397.m00054_x_atXM_643658Hypothetical protein0.0657.660.020
36.m00211_atXM_650002Hypothetical protein0.0949.280.005
245.m00039_x_atXM_645025Hypothetical protein0.4722.340.014
301.m00039_s_atXM_644356Heat shock protein 70, putative0.3021.140.003
81.m00150_s_atXM_648418Heat shock protein 101, putative1.5420.480.001
450.m00030_atXM_643384Hypothetical protein0.2019.960.027Colitis
356.m00029_s_atXM_643911Hypothetical protein0.1816.20.008
7.m00453_s_atXM_651737Hypothetical protein0.3014.840.001
344.m00044_atXM_644000Hypothetical protein0.2413.530.033Cysts
295.m00030_atXM_644416Conserved hypothetical protein0.1613.280.048
134.m00124_atXM_646949Heat shock protein, Hsp20 family, putative1.4412.290.016
33.m00209_x_atXM_650174Hypothetical protein0.1012.10.004
141.m00082_atXM_646820Protein kinase, putative0.5412.090.027
20.m00272_x_atXM_650868Conserved hypothetical protein0.179.860.043
796.m00013_s_atXM_642828Conserved hypothetical protein0.239.690.015
Table 5.  Genes downregulated by DPTA-NONOate in E. histolytica HM-1:IMSS.
Probe IDAccession numberDescriptionBaseline expression levelFold-changeP-valueRegulated under other conditions
  1. The probe ID, accession number, description, baseline expression level, fold-change, P-value and regulation under other conditions are shown. The 30 most highly regulated genes are listed. HS, heat shock, adapted from Hackney et al. (2007); colitis, ameba from a mouse model of amebic colitis, adapted from Gilchrist et al. (2006); trophs, E. histolytica trophozoites, adapted from Ehrenkaufer et al. (2007).

4.m00678_s_atXM_652013Hypothetical protein41.94−49.020.002
172.m00078_atXM_646199Hypothetical protein0.86−14.430.001HS
99.m00179_atXM_647858Inositol polyphosphate kinase, putative4.53−11.480.001
341.m00039_s_atXM_644035Hypothetical protein2.41−10.590.015
229.m00063_atXM_645215Hypothetical protein5.04−9.350.019
221.m00089_s_atXM_645315Hypothetical protein3.74−9.010.002
13.m00349_atXM_651311Protein kinase, putative2.17−8.850.004
65.m00147_atXM_648922Hypothetical protein2.93−8.470.003Trophs + HS
255.m00049_atXM_644881Conserved hypothetical protein1.25−8.20.022
93.m00151_atXM_648019WD repeat protein1.99−8.20.004
22.m00263_atXM_650739Hypothetical protein1.42−7.940.019
74.m00199_atXM_648611Hypothetical protein2.53−7.870.001
232.m00071_atXM_645176Hypothetical protein1.15−7.460.004
4.m00607_atXM_652082Protein kinase, putative3.19−7.460.002HS
408.m00044_atXM_643584Hypothetical protein0.95−7.460.043
86.m00158_s_atXM_648244DEAD DEAH box helicase, putative0.46−7.350.044HS
32.m00207_atXM_650196Zinc finger protein, putative12.62−6.940.009
32.m00202_s_atXM_650191Hypothetical protein0.72−6.90.008
4.m00640_atXM_652059Conserved hypothetical protein3.07−6.90.000
113.m00152_atXM_647500BspA-like leucine-rich repeat protein, putative1.09−6.850.015Trophs
289.m00071_s_atXM_644481Putative GTPase7.01−6.290.000
82.m00139_s_atXM_648379Hypothetical protein0.59−6.210.005
408.m00045_s_atXM_643583Hypothetical protein2.43−6.170.007HS
2.m00528_atXM_652349Protein phosphatise, putative0.35−5.990.028HS
201.m00110_atXM_645662Hypothetical protein1.35−5.920.006HS
12.m00322_atXM_651338Hypothetical protein3.67−5.810.030HS
48.m00216_s_atXM_649527Hypothetical protein3.05−5.750.019Colitis
98.m00148_s_atXM_647890Hypothetical protein1.02−5.520.001
66.m00150_atXM_648887Hypothetical protein0.39−5.490.000HS
816.m00009_atXM_642817Hypothetical protein4.41−5.380.049HS
Figure 1.

Semiquantitative RT-PCR analysis of selected genes for validation of array results. Up – genes induced by exposure to the corresponding stress; down – genes repressed by exposure to the corresponding stress.

Genes encoding known detoxification systems in E. histolytica trophozoites are not significantly modulated by oxidative or nitrosative stress

In E. histolytica HM-1:IMSS strain growing under standard axenic culture conditions, the basal transcription levels of genes encoding the enzymatic ROS and RNS detoxification systems are generally high (Fig. 2). Putative detoxification systems for reactive oxygen and nitrogen species, which were identified in the genome of E. histolytica, are depicted in Fig. 3. FDPs have nitric oxide (NO) and/or molecular oxygen (O2) reductase activities, although the molecular basis for substrate selectivity remains elusive (Vicente et al., 2007). Out of the four copies of genes encoding FDPs, two (6.m00467 and 155.m00084) have much higher transcription levels than the other two homologues (65.m00171 and 146.m00121) (Fig. 2A). The gene encoding SOD (384.m00041), the sole known superoxide-detoxifying enzyme in E. histolytica's genome, displays high transcription levels in all conditions tested with no significant changes under either stress condition (Fig. 2B). The same is observed for gene products involved in hydrogen peroxide detoxification, as the three distinct scavenging systems identified in its genome also display high and almost invariable expression levels: peroxiredoxin (176.m00112), rubrerythrin (131.m00144) and hybrid-cluster protein (8.m00410) (Fig. 2B). Although these genes do not display significant transcriptional changes upon H2O2 or NO exposure, basal expression levels of some of these genes do vary across Entamoeba: the FDP-encoding 6.m00467 gene, which is the most highly expressed FDP in E. histolytica HM-1:IMSS, has a lower expression in the non-virulent species, Entamoeba dispar, and the gene coding for peroxiredoxin, has reduced expression in both E. histolytica Rahman and E. dispar (Davis et al., 2006; MacFarlane and Singh, 2006). Moreover, a proteomic analysis revealed that E. histolytica SOD and peroxiredoxin have significantly higher expression levels in the virulent E. histolytica HM-1:IMSS strain as compared with the non-virulent Rahman strain (Davis et al., 2006), indicating a possible contribution by these proteins to amebic virulence.

Figure 2.

Expression levels of genes encoding identified detoxification pathways for reactive oxygen and nitrogen species for unchallenged cultures of E. histolytica strains HM-1:IMSS (pathogenic) and Rahman (non-pathogenic), and for cultures challenged with hydrogen peroxide and the nitric oxide releaser DPTA-NONOate.

Figure 3.

Putative detoxification pathways for reactive oxygen and nitrogen species identified in the genome of E. histolytica HM-1:IMSS. ROS depicted as the sequential one-electron-reduced intermediates of oxygen reduction. Rbr, rubrerythrin, hydrogen peroxide reductase; HCP, hybrid-cluster protein; Prx, peroxiredoxin.

Although the majority of genes encoding the ROS and RNS stress-detoxifying proteins were not significantly changed in parasites exposed to oxidative and nitrosative stresses, some transcriptional changes were noted. Three genes encoding iron-sulfur flavoproteins (35.m00253, 312.m00037 and 41.m00244) that are proposed to be related to the oxidative stress response, but whose function remains undetermined (Cruz and Ferry, 2006), were upregulated by H2O2. Two homologous genes (187.m00073 and 646.m00021) encoding iron-sulfur flavoproteins were induced by nitrosative stress. Additionally, one of the four FDPs (146.m00121) was upregulated by NO (4.8-fold). Besides FDPs, no other known NO-detoxifying enzyme has thus far been identified in E. histolytica's genome.

Aside from these limited changes, most of the genes in the ROS and RNS detoxification pathways were not transcriptionally regulated under a number of different conditions that model the host–pathogen interaction (parasites colonizing the mouse intestine) (Gilchrist et al., 2006) or trophozoite to cyst stage conversion (Ehrenkaufer et al., 2007). Thus, contrary to what has been observed in many prokaryotes and a few eukaryotes (Paget and Buttner, 2003; Rodionov et al., 2005), the transcriptional response at the detoxification level appears constitutive as E. histolytica trophozoites have a number of scavenging enzymes that are robustly expressed even under basal conditions.

Genes upregulated in E. histolytica HM-1:IMSS in response to oxidative stress

Out of the 185 genes upregulated by oxidative stress, 107 (58%) are annotated as encoding hypothetical proteins. The remaining genes code for proteins with roles in signalling/regulatory processes, metabolic/repair processes, energy metabolism, stress response and transport (Fig. 4). The categories comprising the largest numbers of upregulated genes include the repair systems and signalling/regulatory pathways, which are detailed below.

Figure 4.

Transcriptional profiles of E. histolytica HM-1:IMSS exposed to oxidative and nitrosative stresses. Genes were grouped according to putative functions inferred from the respective annotations.

Response to DNA damage.  DNA is a known cellular target of oxidative stress (reviewed in D'Autreaux and Toledano, 2007) and 8% of the genes upregulated by H2O2 in E. histolytica encode putative nucleic acids metabolism/repair proteins (Fig. 4, Table 2). Two genes encoding deoxyuridine 5′-triphosphate nucleotidohydrolase (194.m00123 and 375.m00057), considered to be essential for DNA integrity (Nguyen et al., 2005), are among the genes most highly upregulated by H2O2 (51- and 22-fold respectively). A putative homologue of polynucleotide kinase-3-phosphatase (1.m00709, fourfold) is involved in the repair of nicks and gaps in damaged DNA, including oxidatively generated DNA strand breaks (Betti et al., 2001; Blondal et al., 2005). A similar function in the repair of oxidative damage to DNA has been attributed to the large family of MutS repair proteins (Chang et al., 2002; Khil and Camerini-Otero, 2002; Dzierzbicki et al., 2004). A gene coding for a homologue of MutS DNA mismatch repair proteins in E. histolytica (115.m00144) was upregulated by oxidative stress (twofold). We also observed induction by hydrogen peroxide of a gene encoding a homologue of DEAD DEAH box helicase (27.m00240, threefold). The homologous gene from the pathogenic fungus Candida albicans is regulated by Cap1p, a transcription factor involved in oxidative stress tolerance (Wang et al., 2006).

Response to protein and lipid damage.  Hydrogen peroxide oxidatively damages proteins, mainly by reacting with thiol groups from cysteine side-chains, and also with redox cofactors such as metal centres. Moreover, the hydroxyl radical generated in the reaction between H2O2 and free Fe2+ further reacts with amino acid residues, namely methionine (Halliwell and Gutteridge, 2007). In E. histolytica HM-1:IMSS challenged with H2O2, 4% of the upregulated genes encode homologues of proteins involved in the repair or degradation of misfolded proteins (Fig. 4 and Table 2). Genes encoding homologues of chaperone-like heat-shock proteins were upregulated by oxidative stress, for example, HSP40/DnaJ (21.m00247) that stimulates the ATPase activity of HSP70 chaperones (Qiu et al., 2006) and HSP101 (64.m00187). An ubiquitin-conjugating enzyme (142.m00162), which may mark misfolded proteins for degradation, was also upregulated. A homologue of peptidyl-prolyl cis-trans isomerase (75.m00189), a protein with a role in the repair of oxidatively damaged proteins from plants (Shapiguzov et al., 2006) and mammalian cells (Santos et al., 2000; Hong et al., 2002) was also upregulated.

Oxidative damage exerted on lipids by H2O2 and other ROS may result in loss of membrane integrity (Colles and Chisolm, 2000; Halliwell and Gutteridge, 2007). Accordingly, some of the genes upregulated by H2O2 have a role in lipid metabolism: phosphatidylcholine transport-like protein (99.m00180) and a phospholipid-transporting P-type ATPase (75.m00173), homologous to aminophospholipid translocases. Increased expression of the gene coding for glucosamine-6-phosphate N-acetyltransferase (34.m00243) suggests a requirement for cell wall repair and/or the assembly of cell wall proteins (Hurtado-Guerrero et al., 2007). It has been reported that a mutant strain of C. albicans with a deletion in this gene displays decreased virulence (Mio et al., 2000).

Signalling and regulatory pathways.  The largest group of genes upregulated by H2O2 (20%) is that encoding proteins that may be related to signalling pathways including protein kinases, phosphatases and acetyltransferases. E. histolytica possesses an intricate phosphorylation network involving a large number of kinases, which have roles in diverse cell processes (Anamika et al., 2008), including effects on parasite virulence (Batista Ede and de Souza, 2004; Beck et al., 2005). Other genes that may be involved in signal transduction mechanisms include those that encode GTPases and related proteins, such as the G protein regulator phosducin (407.m00055) (2.4-fold). The two genes coding for Rab GTPases, RabI1 (20.m00304, 3.6-fold) and RabM1 (103.m00161, twofold) (Saito-Nakano et al., 2005) are homologues of Rab1B and Rab15, respectively, involved in transport and endocytosis. A small GTPase CDC42 (encoded by 219.m00107, 11.3-fold) has a human homologue that regulates adherence and membrane permeability (Broman et al., 2007) and more recently has been proposed to have a role in regulating protein ubiquitination (Shen et al., 2008). In addition, ArfA3 (147.m00113) (Clark et al., 2007), an ADP ribosylation factor GTPase, was induced by H2O2 (2.6-fold). ADP ribosylation factor GTPases participate in the regulation of organelle structure and vesicular trafficking, besides participating in diverse cellular functions, such as cytokinesis, endocytosis, phagocytosis and cell adhesion (D'Souza-Schorey and Chavrier, 2006). We also observed induction of a gene encoding a putative copine (333.m00053) (2.1-fold), whose homologue from Dictyostelium discoideum plays a role in cytokinesis and contractile vacuole function (Damer et al., 2007). An AIG1 plant-like antibacterial protein (451.m00039) was induced (2.1-fold) and was also significantly upregulated upon E. histolytica invasion of the mouse intestine (Gilchrist et al., 2006). Two genes encoding BspA-like leucine-rich proteins (1.m00705 and 310.m00070) were induced by H2O2 (6.6- and 2.4-fold respectively). Homologues of these proteins from Tannerella forsythia have been reported to modulate the host response by interfering with interleukin-8 expression and ultimately contributing to invasion of the epithelial barrier (Inagaki et al., 2006; Onishi et al., 2008). In summary, E. histolytica responds to oxidative stress by inducing genes coding for a multitude of signalling/regulatory systems, with roles in diverse cellular functions.

Other pathways.  As E. histolytica redox homeostasis is sustained by free (homo)cysteine, in the absence of glutathione or any related metabolic enzymes, it is noteworthy that sulfur amino-acid metabolism enzymes are regulated by oxidative stress. The gene encoding the MGL1 isotype of methionine-γ-lyase (MGL1, 395.m00028), which in eukaryotes has thus far only been found in plants (Rebeille et al., 2006), E. histolytica (Tokoro et al., 2003) and T. vaginalis (Nozaki et al., 2005), is upregulated (3.6-fold) by H2O2. This enzyme catalyses elimination reactions with a wide range of substrates, such as methionine (homo)cysteine, and substituted (homo)serine homologues (Nozaki et al., 2005). Moreover, the products of methionine degradation by MGL not only supply the energy metabolism, but may also be implicated in amebic pathogenicity as thiols and hydrogen sulfide can become toxic for the host cells by permeating the membrane barrier and interfering with host signalling systems (Nozaki et al., 2005).

Genes downregulated in E. histolytica HM-1:IMSS in response to oxidative stress

A total of 102 genes were downregulated by H2O2 exposure, 52 (51%) of which encode hypothetical proteins (Table 3; Table S2). As displayed in Fig. 4, the majority of the E. histolytica HM-1:IMSS genes repressed by oxidative stress appear to be involved in the same general processes whose components were upregulated, namely repair systems (mostly for nucleic acids), signalling pathways and regulatory mechanisms. The numerous genes possibly involved in signalling and regulatory processes encode a number of putative GTPases, one of which (64.m00149) encodes a protein with RhoGEF and ArfGAP domains and has a homologue proposed to be involved in D. discoideum development (Mondal et al., 2007).

Genes upregulated in E. histolytica HM-1:IMSS in response to nitrosative stress

From the 443 genes upregulated by nitrosative stress, 248 (56%) are annotated as encoding unknown hypothetical proteins. Similar to what was observed with oxidative stress, the largest groups of genes induced by nitrosative stress are those related with signalling/regulatory processes and repair systems for nucleic acids, proteins and lipids.

Response to DNA damage.  Upon exposure to NO, 33 upregulated genes (7%) encode proteins involved in the metabolism and/or repair of nucleic acids (Fig. 4). This reflects the direct damage exerted by RNS on nucleic acids, either NO directly or NO-derived species, such as peroxynitrite (Halliwell and Gutteridge, 2007). In Salmonella enterica, NO impairs DNA replication and arrests growth (Schapiro et al., 2003). Two genes encoding deoxyuridine 5′-triphosphate nucleotidohydrolase (194.m00123 and 375.m00057) are among the genes most highly upregulated by NO (192- and 82-fold respectively), and are essential enzymes for DNA integrity (Nguyen et al., 2005). A homologue of MutS DNA mismatch repair proteins (115.m00144) was also upregulated by nitrosative stress (3.3-fold). We observed induction of genes encoding a panoply of DNA repair enzymes, such as DNA excision repair protein (117.m00190) (fourfold), Rad3 DNA repair helicase (197.m00081) (2.1-fold) and Rad50 (102.m00081) (twofold).

Response to protein and lipid damage.  NO and its derived species can be harmful to proteins involved in all cellular processes, both by reacting with specific amino acids and/or with redox active cofactors, mostly metal centres. The reaction of RNS with amino acids results in S-nitrosylation or nitration of the corresponding side-chains. NO and other RNS can also bind transiently or permanently to protein metal cofactors and either inhibit or definitely inactivate its function. The resulting modifications can be damaging to the overall folding and structural arrangements of proteins, which was confirmed by the significant number of E. histolytica genes with these functions that were upregulated by NO exposure (5% of all upregulated genes by NO) (Fig. 4). The vast majority of these genes encode either heat-shock proteins, [HSP20 (134.m00124) (12.3-fold), three homologues of HSP40 (12.m00313, 28.m00338 and 28.m00311) (5.6-, 3.8- and 3.3-fold respectively), two of HSP70 (301.m00039 and 584.m00019) (21.1- and 2.3-fold), and three of HSP101 (81.m00150, 64.m00178 and 64.m00187) (20.5-, 8.8- and 6.1-fold respectively)], or ubiquitin-conjugating proteins, which are all involved in degradation/repair of misfolded proteins and often responsive to many types of cellular stress.

Other major targets of NO and RNS reactivity are lipids. While this interaction may damage lipids, NO and RNS may also act as antioxidants, reacting with lipid radicals generated by oxidative stress, thus blocking and terminating harmful lipid radical chain reactions (Halliwell and Gutteridge, 2007). The low polarity of NO allows it to freely diffuse through membranes. Damage to lipids cannot only affect metabolic processes, but most importantly may result in cell membrane permeability. The effect of nitrosative stress on lipids in E. histolytica is directly observable by the number of upregulated genes encoding proteins related to lipid metabolism (3% of all induced genes): a PCTP-like protein (99.m00180) (2.8-fold), a myotubularin lipid phosphatase (35.m00216) (2.6-fold) and a phospholipid-transporting P-type ATPase (75.m00173) (2.6-fold). In addition to affecting the membranous environment, NO reacts directly with membrane proteins, such as ion channels and transport proteins, thus disturbing ion homeostasis. Two genes encoding putative importins (310.m00066 and 1.m00747) were induced by nitrosative stress (4.0- and 2.1-fold respectively). The nuclear transport pathway mediated by importin in human cells has recently been shown to be impaired by nitrosative stress (Qu et al., 2007). A putative sodium/proton antiporter (152.m00122) was upregulated by NO (3.5-fold). Nitric oxide has been shown to inhibit Na+/H+ exchange activity, mediated by the cyclic GMP signal transduction pathway (Gill et al., 2002). Genes encoding homologues of amino acid transporters (2.m00499, 46.m00238 and 82.m00146) were also induced by nitrosative stress.

Signalling and regulatory processes induced by nitrosative stress.  The largest groups of regulated genes (18%) are those involved in signalling and regulation of cellular processes (Fig. 4). Part of this group overlaps with genes found to be upregulated by oxidative stress. From the 78 genes in this group, at least 21 encode putative protein kinases, 7 encode protein phosphatases and 5 encode acetyltransferases. A significant number of GTPases (four Rab-type, three Rho family, one Rap Ran GTPase-activating protein and two Rab GTPase-activating proteins) and zinc finger proteins were also upregulated by nitrosative stress. Reversible inhibition of DNA-binding zinc-containing proteins by nitrosative stress has been thought to be part of NO-related DNA replication inhibition in pathogenic bacteria (Schapiro et al., 2003).

Other pathways.  A putative FAD-binding NADH oxidoreductase (328.m00064) was induced (2.4-fold) by NO. Homologues of this enzyme are primary electron carriers in electron transfer chains with NO-detoxifying activity (Saraiva et al., 2004). A nitroreductase (13.m00321) was induced upon NO exposure (3.9-fold). Nitroreductases are flavoproteins that catalyse the reduction of nitro groups in a wide range of substrates and are associated with resistance to the antiparasitic antibiotic metronidazole (Mendz and Megraud, 2002). The gene that codes for MGL1 (395.m00028) was also upregulated by nitrosative stress (4.4-fold). The products of methionine degradation by MGL may permeate the host cells membrane barrier and interfere with signalling systems (Nozaki et al., 2005). Interestingly, three genes encoding putative cysteine proteases were induced by NO (24.m00271, 180.m00101 and 97.m00133) (2.0-, 2.1- and 2.1-fold respectively). Cysteine proteases are an important virulence factor for E. histolytica, and are involved in cytotoxicity and colonic invasion (Stanley, 2003).

Genes downregulated by nitrosative stress

A surprisingly large number of genes had decreased expression upon exposure to DPTA-NONOate: 592 genes of which 366 (62%) are annotated as unknown hypothetical proteins (Table 4; Table S3). The profile of downregulated genes and their putative functions is shown in Fig. 4. There is a marked repression of genes involved in the metabolism of nucleic acids (6%), repair and degradation of misfolded proteins (2%) and lipids (2%), and in transport (2%). The largest category of genes is that involved in cell signalling and regulatory processes (21%).

A significant downregulation of genes encoding proteins involved in RNA metabolism was observed. Genes encoding RNA polymerase subunits were repressed by nitrosative stress (64.m00185, 73.m00160, 59.m00197, 51.m00170 and 406.m00050) (4.0-, 2.1-, 3.2-, 2.5- and 3.2-fold respectively), suggesting that the overall transcription may be slowed down. Exposure to nitric oxide led to repression of genes coding for putative proteins related to ubiquitination (three cullin homologues, one ubiquitin and one ubiquitin-conjugating enzyme) and ribosome-related proteins. We also observed downregulation of genes encoding proteins involved in (glyco)lipid metabolism and glycosylation, such as phosphatidylinositol-glycan biosynthesis class C protein (80.m00142) (3.4-fold), N-acetylglucosaminyl transferase (32.m00239) (3.3-fold) and N-acetylglucosaminyl-phosphatidylinositol de-N-acetylase (52.m00150) (2.5-fold). These genes are likely to participate in the biosynthesis of glycosylphosphatidylinositol anchors, which have been proposed to be involved in the regulation of cell growth, endocytosis and the adhesion to target cells by E. histolytica (Vats et al., 2005). The bulk of genes repressed by NO that encode putative transport proteins comprise those coding for importin subunits.

From the extended list of NO-repressed genes encoding putative signalling/regulatory proteins, 21 code for protein kinases and 23 for GTPases (9 Rab-type, 6 Ras-type and 8 Rho-type). There is also downregulation of 10 genes coding for zinc finger proteins, 6 for WD repeat proteins and 10 for leucine-rich proteins. Significantly, exposure to NO resulted in repression of gene 43.m00187 (2.4-fold). This gene encodes a key regulatory protein, phosphatidylinositol-3,4,5-trisphosphate 3-phosphatase (PTEN), which controls diverse cellular processes as the antagonist of phosphatidylinositol 3-kinase (constituting the PTEN/PIK signalling pathway). E. histolytica phosphatidylinositol 3-kinase inhibition has been shown to impair proliferation, encystation and autophagy (Picazarri et al., 2008).

Common transcriptional responses of E. histolytica HM-1:IMSS to oxidative and nitrosative stresses

We observed a substantial overlap in genes transcriptionally regulated by oxidative and nitrosative stresses with 102 genes upregulated under both stress conditions, including those genes that are the most induced under each of those conditions (Fig. 5) (Table S5). Of these 102 genes, 63 encode unknown hypothetical proteins. The remaining ones are distributed according to the function profiles observed for the individual stresses (Fig. 4). The observation of commonly induced repair system-encoding genes by both stress types reflects the fact that nucleic acids and proteins are among the cell components that suffer similar damaging reactions upon exposure to both ROS and RNS (Halliwell and Gutteridge, 2007). Two of the genes showing the highest induction folds by both stress types are deoxyuridine 5′-triphosphate nucleotidohydrolase (194.m00123 and 375.m00057). As mentioned above, this enzyme is involved in nucleotide metabolism and contributes to DNA integrity and has been previously assessed as a potential target for antiparasitic drugs (Nguyen et al., 2005). Two genes (1.m00709 and 115.m00144) are homologues of a polynucleotide kinase-3-phosphatase and a MutS DNA mismatch repair protein respectively, and are both involved in the repair of oxidatively damaged DNA (Betti et al., 2001; Dzierzbicki et al., 2004). Both stresses induced genes encoding systems involved in the degradation and repair of misfolded proteins such as heat-shock proteins HSP101 and DnaJ/HSP40 (64.m00187 and 21.m00247) and an ubiquitin-conjugating enzyme (142.m00162).

Figure 5.

Venn diagrams depicting the overlap between the transcriptional changes observed upon exposing E. histolytica HM-1:IMSS to oxidative and nitrosative stresses.

We observed induction under oxidative and nitrosative stresses of genes involved in lipid metabolism, transport and glycosylation. Two genes encoding proteins involved in lipid metabolism were phospholipid-transporting P-type ATPase (75.m00173) and phosphatidylcholine transfer protein (99.m00180). An aminophospholipid translocase (75.m00173), upon nitrosative inhibition in apoptotic cells, leads to an accumulation of extracellular phosphatidylserine that marks cells for macrophage engulfment (Tyurina et al., 2007). E. histolytica has been reported to recognize externalized phosphatidylserine on the surface of host cells and target these cells for phagocytosis (Huston et al., 2003; Boettner et al., 2005). Genes encoding putative transport proteins (mainly ion transporters) were induced both by oxidative and nitrosative stresses, consistent with the observation that nitrosative and oxidative stresses disturb ion homeostasis (Beausejour et al., 2007; Orsenigo et al., 2007). A dTDP-D-glucose 4,6-dehydratase (116.m00108) was upregulated by both types of stresses; this enzyme is related to glycosylation in pathogenic bacteria (Allard et al., 2001). Also induced by both stresses was a glucosamine-6-phosphate N-acetyltransferase (34.m00243) whose homologous gene in C. albicans has a role in virulence (Mio et al., 2000).

It is worth noting that all four protein families that respond to oxidative stress in all eukaryotic kingdoms (Vandenbroucke et al., 2008) (heat-shock proteins, ubiquitin-conjugating enzymes, kinases and small GTPases) have homologues that were transcriptionally regulated in E. histolytica by both oxidative and nitrosative stresses: HSP101 (64.m00187), an ubiquitin-conjugating enzyme (142.m00162), three protein kinases (199.m00096, 46.m00221 and 275.m00123) and one Rab family GTPase (20.m00304).

In previous sections we emphasized the induction of a gene encoding MGL (395.m00028) by both stress types, as its methionine degradation products may contribute to the permeation and disruption of the host epithelial barrier by E. histolytica. MGL is an attractive drug target that is being actively pursued (Ali and Nozaki, 2007; Sato et al., 2008).

The majority of the genes most significantly repressed by each individual stress were downregulated under both stresses (Table S5). Sixty-two genes were downregulated by both stresses, 35 (56%) of which encode unknown hypothetical proteins. The remaining genes encode putative proteins mostly involved in signalling and regulatory mechanisms, and also include a few genes encoding nucleic acid metabolism proteins. No genes encoding repair systems for misfolded proteins or lipids were commonly downregulated under both stresses, although both stresses did regulate genes in these pathways. In summary, the overlap between genes responsive to oxidative and nitrosative stresses by E. histolytica is significant and shows common strategies to overcome the cytotoxicity of reactive oxygen and nitrogen species.

We also analysed the overlap between the transcriptional responses of E. histolytica to oxidative or nitrosative stress and other conditions. As expected, some genes regulated by H2O2 were also regulated by cyst conversion, indicating that these genes respond to multiple stress conditions (Weber et al., 2006; Ehrenkaufer et al., 2007; Hackney et al., 2007). A significant fraction of the genes repressed by H2O2 (47%) was also repressed in response to heat shock (Hackney et al., 2007; Weber et al., 2006). Of the 443 genes upregulated by DPTA-NONOate, 50 were also induced by heat shock, some of which encode putative repair systems for damaged nucleic acids and proteins, 25 were also upregulated in cysts, but only a few were induced upon colonization of the mouse intestine (Gilchrist et al., 2006). Almost half of the 592 genes downregulated by nitrosative stress were also downregulated by heat shock, a much smaller fraction (10%) was downregulated in cysts, and 12 genes were downregulated in an in vivo mouse colitis model (Table 5 and Table S3). The limited overlap between genes regulated by oxidative and nitrosative stresses and the changes seen during colonization of the mouse colon and hepatic invasion deserve further comment (Gilchrist et al., 2006; Santi-Rocca et al., 2008). Whether this represents technical issues (differences in time points of colonic animal model, which represents more colonization than invasion, or differences in the arrays used for the liver invasion model compared with these studies) or biological differences (animal models are more complex and difficult to compare directly with studies utilizing in vitro methods) is not currently clear. However, as the genes and pathways transcriptionally modulated by oxidative or nitrosative stress overlap significantly with changes seen in multiple other systems using the same stress conditions, the approach with the in vitro model of stress exposure is identifying conserved mechanisms of stress response between Entamoeba and other systems.

Differential response to oxidative stress may contribute to the decreased virulence phenotype of the E. histolytica Rahman strain

In order to determine whether virulent and non-virulent amebic strains had differing responses to oxidative stress, we characterized the transcriptional changes of the non-pathogenic E. histolytica Rahman strain to hydrogen peroxide. It has previously been demonstrated that the non-virulent E. histolytica Rahman is more susceptible to hydrogen peroxide than the pathogenic E. histolytica HM-1:IMSS and that higher levels of peroxiredoxin in the virulent strain contribute to E. histolytica's virulence (Davis et al., 2006). Three arrays using RNA from E. histolytica Rahman exposed to 1 mM H2O2 for 1 h were performed and displayed good correlation values (0.97–0.99) (Table 1). Expression was compared with array data from E. histolytica Rahman under standard axenic culture conditions (Ehrenkaufer et al., 2007). Using the same fold-change criteria described in the Experimental procedures (twofold or greater change and FDR < 0.5), a total of 153 genes were upregulated by H2O2 in Ehistolytica Rahman and 65 genes were downregulated (Tables 6 and 7; Table S4).

Table 6.  Genes upregulated by hydrogen peroxide in E. histolytica Rahman.
Probe IDAccession numberDescriptionBaseline expression levelFold-changeP-valueRegulated in HM-1:IMSSFold-change in HM-1:IMSS
  1. The probe ID, accession number, description, baseline expression level, fold-change and P-value are shown. The 30 most highly regulated genes are listed. If the gene is regulated in HM-1:IMSS, then its fold-change is listed.

301.m00039_s_atXM_644356Heat shock protein 70, putative0.307.370.005 
205.m00100_s_atXM_645582Hypothetical protein8.196.90.013 
134.m00124_atXM_646949Heat shock protein, Hsp20 family, putative1.446.620.006 
606.m00014_s_atXM_642953Hypothetical protein0.286.590.043 
64.m00187_s_atXM_648976Heat shock protein 101, putative36.556.530.015+2.42
8.m00393_atXM_651687Late competence protein, putative0.256.480.010+8.85
264.m00070_x_atXM_644777Hypothetical protein0.056.310.030 
181.m00068_s_atXM_646040Hsp101-related protein18.456.10.009 
562.m00023_atXM_643023Protein kinase, putative1.185.740.022 
256.m00083_x_atXM_644864Hypothetical protein1.225.720.050+82.85
188.m00103_atXM_645894Hypothetical protein0.485.680.038 
451.m00037_s_atXM_643383Hypothetical protein0.075.590.016 
81.m00150_s_atXM_648418Heat shock protein 101, putative1.545.510.001 
110.m00118_x_atXM_647568Rho family GTPase0.115.470.030 
64.m00178_s_atXM_648975Heat shock protein 101, putative3.415.240.011 
654.m00032_x_atXM_642891Hypothetical protein1.395.010.047+114.90
42.m00175_atXM_649778Hypothetical protein0.964.920.018 
363.m00049_x_atXM_643869Hypothetical protein0.174.90.030+89.05
344.m00045_atXM_644001Hypothetical protein0.604.820.008+6.21
442.m00023_atXM_643432Hypothetical protein0.694.750.015+4.49
50.m00195_s_atXM_649449Hypothetical protein0.054.730.018 
30.m00249_atXM_6502771-O-acylceramide synthase, putative0.404.630.026 
39.m00253_atXM_649881CXXC-rich protein8.574.560.009 
168.m00119_s_atpseudogene, N-acetylmuraminidase0.354.550.029 
167.m00116_x_atXM_646306Hypothetical protein0.084.480.014 
458.m00058_atXM_643333Hypothetical protein0.144.450.017 
6.m00424_atXM_651810Hypothetical protein0.224.260.012+3.00
70.m00178_s_atXM_648752Hypothetical protein0.243.960.027 
245.m00039_x_atXM_645025Hypothetical protein0.473.840.040+14.39
7.m00453_s_atXM_651737Hypothetical protein0.303.70.021+3.46
Table 7.  Genes downregulated by hydrogen peroxide in E. histolytica Rahman.
Probe IDAccession numberDescriptionBaseline expression levelFold-changeP-valueRegulated in HM-1:IMSSFold-change in HM-1:IMSS
  1. The probe ID, accession number, description, baseline expression level, fold-change and P-value are shown. The 30 most highly regulated genes are listed. If the gene is regulated in HM-1:IMSS, then its fold-change is listed.

50.m00168_atXM_649471Hypothetical protein107.85−30.490.012 
233.m00105_atXM_645153Hypothetical protein6.57−4.130.005+−4.41
214.m00072_atXM_645429Hypothetical protein1.02−3.690.015 
249.m00072_atXM_644965Hypothetical protein0.24−3.50.042 
459.m00030_atXM_643329Hypothetical protein74.08−3.40.034 
25.m00245_atXM_650545Conserved hypothetical protein0.35−3.190.006 
67.m00091_x_atXM_648866Protein kinase, putative6.08−3.10.046 
113.m00152_atXM_647500Hypothetical protein1.09−2.920.002 
224.m00085_atXM_645268Cullin, putative0.44−2.820.033 
41.m00219_s_atXM_649804ABC transporter, putative47.84−2.740.006 
380.m00029_atXM_643755Hypothetical protein0.09−2.720.019 
54.m00183_atXM_649345Hypothetical protein28.38−2.650.025+−2.30
129.m00151_atXM_647089Hypothetical protein0.08−2.560.015 
264.m00067_atXM_644788Hypothetical protein2.21−2.560.034 
234.m00047_atXM_645144Hypothetical protein1.04−2.510.007 
232.m00071_atXM_645176Hypothetical protein1.15−2.490.022 
41.m00243_s_atXM_649788Hypothetical protein5.24−2.460.000 
310.m00064_atXM_644295RNA polymerase I largest subunit, putative26.67−2.430.009 
103.m00165_x_atXM_647745Hypothetical protein0.14−2.40.034 
291.m00043_atXM_644448Leucine-rich repeat protein0.78−2.380.041 
26.m00293_atXM_650512Poly(A) polymerase, putative1.35−2.370.029 
247.m00075_atXM_644998LIM domain protein1.74−2.330.007 
2.m00522_atXM_652343U6 snRNA-associated Sm-like protein, putative0.63−2.320.008 
9.m00420_atXM_651594Carbonic anhydrase, putative2.67−2.320.031 
129.m00135_atXM_647073PH domain protein4.92−2.310.002 
34.m00252_atXM_650085Conserved hypothetical protein0.29−2.30.030 
25.m00242_atXM_650542Hypothetical protein7.15−2.290.011 
277.m00057_atXM_644616Actobindin homologue, putative126.03−2.280.007 
108.m00108_atXM_647607Hypothetical protein0.50−2.260.019 
59.m00197_x_atXM_649134DNA-directed RNA polymerase subunitN, putative7.02−2.250.023 

Overall, E. histolytica Rahman had a decreased repertoire of transcriptional changes in response to oxidative stress – both in terms of the numbers of genes regulated and the magnitude of their regulation (Fig. 6 and Tables 6 and 7; Table S4). Of the genes upregulated in E. histolytica Rahman, only 36 (24%) were also upregulated in the E. histolytica HM-1:IMSS strain under the same conditions (Fig. 6). Furthermore, only 20% of the genes induced by oxidative stress in HM-1:IMSS strain also upregulated in Rahman. Importantly, even for the genes that overlapped in their expression patterns between E. histolytica HM-1:IMSS and Rahman, the extent of upregulation was much higher in the pathogenic strain E. histolytica HM-1:IMSS (Table S6). Given the degree of genetic identity between E. histolytica HM-1:IMSS and E. histolytica Rahman (Shah et al., 2005), the limited similarity of the response to the same stress condition was unexpected. Indeed, previous comparisons of the two strains under standard culture conditions have identified a limited number of genes with lower expression in Rahman compared with the HM-1:IMSS strain (Davis et al., 2006; MacFarlane and Singh, 2006). The more robust transcriptional response of the virulent strain may contribute to its decreased sensitivity to oxidative stress. Furthermore, the differential response may signify that other genes exclusively upregulated in the pathogenic E. histolytica HM-1:IMSS strain possibly contribute to this strain's virulence potential. These include highly induced genes such as those encoding deoxyuridine 5′-triphosphate nucleotidohydrolase (194.m00123 and 375.m00057), aminophospholipid translocase (75.m00173), dTDP-D-glucose 4,6-dehydratase (116.m00108) and MGL (395.m00028). These genes are highly upregulated by both oxidative and nitrosative stresses in the HM-1:IMSS strain but are not regulated at all in the Rahman strain. Moreover, some of these genes have roles in the pathogenicity of other organisms and have been tested as potential novel drug targets (Nguyen et al., 2005; Ali and Nozaki, 2007; Sato et al., 2008).

Figure 6.

Venn diagrams depicting the overlap between the transcriptional changes observed upon exposing a pathogenic and a non-pathogenic E. histolytica strain to oxidative stress.

Of the genes induced in E. histolytica Rahman by oxidative stress, 93 (61%) encode unknown hypothetical proteins. Some genes with known functions in response to oxidative stress were regulated in the Rahman strain. One such is a UDP-glucose 4-epimerase (226.m00073) whose homologue from C. albicans contributes to morphology and cell-wall integrity and gene silencing results in fungi that are more susceptible to H2O2-derived oxidative stress (Singh et al., 2007). The remaining genes displayed a different profile from those of the E. histolytica HM-1:IMSS strain challenged with oxidative stress (Fig. 4). The number of genes encoding nucleic acids metabolism/repair proteins that were induced by oxidative stress in E. histolytica Rahman was much lower than the number of such genes upregulated by oxidative stress in the pathogenic HM-1:IMSS strain (Fig. 4). Notably, no induction of ubiquitin-conjugating enzymes was observed, contrary to what was observed in E. histolytica HM-1:IMSS.

Overall, it appears that the Rahman strain may lack the transcriptional regulatory mechanisms for coping with oxidative damage. During tissue invasion, trophozoites are exposed to an oxygen-rich environment and the Rahman strain's limited ability to cope with oxidative stress may contribute to its avirulent phenotype.


Upon invasion of the host intestinal epithelium, E. histolytica trophozoites are confronted with varying oxygen tensions and cytotoxic reactive oxygen and nitrogen species. The transcriptional changes in E. histolytica HM-1:IMSS upon exposure to oxidative and nitrosative stresses were extensive both in the numbers of regulated genes and the fold-changes of these genes. A significant fraction of the genes modulated by both stresses codes for unknown proteins, which may constitute response mechanisms yet to be unraveled. Among the genes regulated by H2O2 exposure, we identified genes encoding members of four protein families proposed to compose the core of oxidative stress response in eukaryotes: heat-shock proteins, ubiquitin-conjugating enzymes (misfolded protein degradation and repair), protein kinases and small GTPases (signalling and regulation). Strikingly, genes coding for members of these four families were also induced by nitrosative stress and a significant fraction of these genes responded to both stress types. The common responses thus reflect the overlapping regulatory mechanisms to both stresses by E. histolytica.

Following the premise that an important component of E. histolytica's pathogenic potential is related to its resistance to ROS cytotoxicity and RNS cytotoxicity, we identified a number of genes responsive to either or both stresses, which may contribute to this organism's virulence. Furthermore, we demonstrated that the non-pathogenic E. histolytica Rahman strain had a marked difference in response to oxidative stress. The differential transcriptional regulation observed for both strains upon exposure to the same oxidative stress conditions suggests that Rahman may experience a higher degree of oxidative damage and these changes could contribute to a decreased virulence phenotype of the Rahman strain.

Overall, our work demonstrates that: (i) response to oxidative and nitrosative stresses is modulated by a large and complex network of genes in E. histolytica; (ii) a number of known virulence genes are regulated by these stresses; (iii) the decreased virulence phenotype of the non-pathogenic E. histolytica Rahman may be in part due to its limited response to oxidative stress and (iv) some genes responding to these stress pathways may represent important novel drug targets.

Experimental procedures

E. histolytica strains and culture methods

Entamoeba histolytica HM-1:IMSS (pathogenic, ATCC 30459) and E. histolytica Rahman (non-pathogenic, ATCC 30886) (Ankri et al., 1999; Dvorak et al., 2003) were obtained from ATCC, strain identity confirmed by PCR and RFLP of known genomic loci (Clark and Diamond, 1993) and grown axenically in TYI-S-33 medium at 36.5°C, as previously described (Diamond et al., 1978).

Sensitivity of E. histolytica trophozoites to oxidative and nitrosative stresses

To determine the sensitivity of E. histolytica trophozoites to NO and oxidative stresses, parasites from the HM-1:IMSS strain were seeded into 48-well plates (2 × 104 cells per well), each well filled with growth medium and individually sealed with parafilm. After 16–18 h, cells in mid-log phase (50–70% confluent) were exposed to DPTA-NONOate 100 μM to 1 mM, nitric oxide releaser with a half-time 180 min−1 at 37°C (Nittler et al., 2005), or hydrogen peroxide (H2O2, 100 μM to 5 mM) for 1–8 h. At regular intervals, the percentage of rounded-up parasites and the number of cells that stained with Trypan blue were determined. For microarray experiments, E. histolytica trophozoites were grown in capped 16 ml Falcon glass tubes to mid-log phase, the medium changed and culture tubes capped. In the case of the nitrosative stress assays, the tube cap included a rubber seal that could be punctured without significantly compromising the anaerobic conditions within the culture. Two hours after replacing the medium, DPTA and H2O2 were added, cultures incubated for 60 min, observed to assess the cell morphology, chilled for 5 min and parasites harvested (spun at 1000 g, 4°C, 5 min). The supernatant was removed, the cells re-suspended in trypan blue and the percentage of dead cells assessed.

Isolation of RNA and microarray hybridization

RNA was isolated with Trizol reagent (Invitrogen) according to the manufacturer's protocol, purified with a Qiagen RNeasy kit, and microarray hybridization performed at the Stanford University Protein and Nucleic Acids facility (http://cmgm.stanford.edu/pan/) using previously published protocols (Ehrenkaufer et al., 2007). A custom-generated Affymetrix platform microarray described in Gilchrist et al. (2006), with probe sets that represents 9435 open reading frames, was used for all studies. A fraction of probe sets is predicted to cross-hybridize with multiple genes and are annotated as follows: probe sets labelled (_at) represent a single gene; probe sets labelled (_x_at) have at least one probe that may cross-hybridize with another gene(s); probe sets labelled (_s_at) are those in which all the probes for a given gene cross-hybridize with another gene(s). Probes for intergenic non-coding regions were excluded from all analyses. The arrays were scanned after hybridization and the probe intensities were calculated using Affymetrix GCOS software (http://www.affymetrix.com/products/software/specific/gcos.affx).

Microarray data normalization and analysis

Analysis was performed as in Ehrenkaufer et al. (2007). A minimum of three arrays were used for each condition. Standard correlation coefficients were calculated in Genespring. Normalized expression values for each probe set were obtained from raw probe intensities in R 2.2.0 downloaded from the BioConductor project (http://www.bioconductor.org), using robust multiarray averaging with correction for oligosequence (gcRMA) (Wu et al., 2004). To identify differentially expressed genes, we used local pooled error testing along with Benjamini-Hochberg multiple test correction (Benjamini and Hochberg, 1995). In addition, fold-change was calculated in Genespring GX (http://www.chem.agilent.com/scripts/pds.asp?lpage=27881). The threshold for a probe set to be considered differentially expressed was set at a twofold change with an FDR of < 0.05. Differentially expressed genes that were not annotated as hypothetical proteins were grouped according to their putative functions.

Semiquantitative RT-PCR

Total RNA from E. histolytica HM-1:IMSS and Rahman trophozoites was isolated using Trizol reagent (Invitrogen) and further purified with a Qiagen RNeasy kit. cDNA was synthesized from total RNA (2.5 μg) with the Universal riboclone cDNA system (Promega), following the manufacturer's instructions. The cDNA samples were quantified in a Nanodrop spectrophotometer (NanoDrop Technologies, LLC) and PCR reactions performed with 100 ng of cDNA (15 cycles at 55°C plus 15 cycles at 50°C for every gene, except for gene 879.m00008, which was amplified with 15 cycles at 50°C plus 15 cycles at 57°C). The PCR products were fractionated on 2% agarose gels and analysed with a Kodak Digital Science electrophoresis documentation and analysis system 120. The primers used are listed in Table S1.


We gratefully acknowledge all members of the Singh and Teixeira laboratories for helpful discussions. This work is supported by FCT Grant SFRH/BPD/26895/2006 for J.B.V., NIH Grants AI-068899 and AI-069382 for G.M.E., FCT Project Grant POCI/SAU-IMI/56088/2004 to L.M.S., FCT Project Grant PTDC/BIA-PRO/67267/2006 to M.T. and a grant from the NIAID (AI-053724) to U.S. We wish to thank Lígia S. Nobre (ITQB, Portugal) for the support with the semiquantitative RT-PCRs. All array data are available at the GEO database website at NCBI (Accession Number GSE11811).