Identification of dihydropyrimidine dehydrogenase as a virulence factor essential for the survival of Entamoeba histolytica in glucose-poor environments

Authors


Summary

Adaptation to nutritional changes is a key feature for successful survival of a pathogen within its host. The protozoan parasite Entamoeba histolytica normally colonizes the human colon and in rare occasions, this parasite spread to distant organs, such as the liver. E. histolytica obtains most of its energy from the fermentation of glucose into ethanol. In this study, we were intrigued to know how this parasite reacts to changes in glucose availability and we addressed this issue by performing a DNA microarray analysis of gene expression. Results show that parasites that were adapted to growth in absence of glucose increased their virulence and altered the transcription of several genes. One of these genes is the dihydropyrimidine dehydrogenase (DPD), which is involved in degradation of pyrimidines. We showed that this gene is crucial for the parasite's growth when the availability of glucose is limited. These data contribute to our understanding of the parasite's ability to survive in glucose-poor environments and reveal a new role for the DPD enzyme.

Introduction

Amebiasis is caused by the enteric protozoa, Entamoeba histolytica. The World Health Organization estimates that amebiasis is one of the three most common causes of death from parasitic disease and is responsible for about 100 000 deaths annually (WHO, 1997). Worldwide, it is estimated that about 50 million patients, mostly from developing countries, of both sexes and all ages, suffer from amebiasis (WHO/PAHO/UNESCO, 1997; Gilchrist and Petri, 1999). Results from previous studies have shown that anaerobic glycolysis occurs in E. histolytica, and most of the enzymes that constitute this glycolytic pathway have been characterized (Beanan and Bailey, 1995; Saavedra et al., 2005). Lacking any functional mitochondria, E. histolytica obtains most of its energy from glucose fermentation rather than the Krebs cycle and oxidative phosphorylation. In its pyruvate to ethanol fermentation pathway, aldehyde dehydrogenase and alcohol dehydrogenase convert acetyl-CoA to acetaldehyde, which is then reduced to ethanol (Ginger, 2006; Espinosa et al., 2009). E. histolytica lives in the colon, a niche where the amount of available glucose for fermentation is usually small (around 0.2 g kg−1 tissue) due to the high absorptive capacity of the glucose transporters in the small intestine (Cummings and Macfarlane, 1997; Kellett et al., 2008; Hirayama et al., 2009). On rare occasions, it has been reported that E. histolytica trophozoites leave the colon and migrate to the liver. In this organ, the concentration of glucose was estimated to be twice that of perfusing blood (around 2.0 g kg−1 tissue) (Appelboom et al., 1959; Stanley, 2003; Gunst and Van den Berghe, 2010). This latter finding prompted us to investigate the mechanism of the parasite's adaptation to long-term glucose starvation (LTGS) in its colonic niche, and the parasite's ability to cope with the glucose shock during its migration to the liver. In order to investigate these two biological phenomena, E. histolytica trophozoites that were routinely grown using high-glucose Diamond's TYI-S-33 medium (HGM; final glucose concentration 7.35 g l−1) (Clark and Diamond, 2002) were cultivated in Diamond's TYI-S-33 medium without glucose added (LGM; final glucose concentration 0.15 g l−1). Although most trophozoites did not survive in LGM, some trophozoites did survive and were adapted to grow under this condition of glucose starvation. These adapted trophozoites were also more virulent than trophozoites which were cultured in HGM. A DNA microarray analysis of gene expression was conducted on trophozoites which were cultivated in LGM. The analysis was repeated in glucose-starved trophozoites after they had been re-exposed to HGM for 3 h. The results of these analyses revealed that LTGS modulated the transcription of various genes, such as virulence-related genes and genes that encode for amino acid degradation enzymes, amylases, glycolytic enzymes and heat shock proteins. In this report, we describe an unexpected role for dihydropyrimidine dehydrogenase (DPD), a gene which encodes for an enzyme that is involved in pyrimidine metabolism, as being essential for the adaptation of E. histolytica to LGM.

Results

Cultivation of E. histolytica trophozoites under conditions of a LTGS

In order to investigate the response of the parasite to glucose starvation, wild-type E. histolytica trophozoites, which were previously cultivated in HGM, were glucose-starved for 72 h in LGM. In agreement with the results of our previous study, 12 h of glucose starvation exerted no significant effect on the viability of the wild-type parasites (Tovy et al., 2010) (Fig. 1). Extending the duration of the glucose starvation beyond 12 h caused the death of most parasites. Nevertheless, a small number of parasites did survive and became adapted to LGM. These adapted parasites were cultivated in LGM for 1 month before being investigated. To check whether the reduced amount of glucose in LGM has any effect on cellular energy levels, we measured the intracellular ATP levels using a luciferase method. We observed that trophozoites cultivated in HGM have twice the amount of ATP compared with that of the trophozoites cultivated in LGM (Table S1).

Figure 1.

Adaptation of E. histolytica trophozoites to long-term glucose starvation. E. histolytica trophozoites strain HM1:IMSS were grown in LGM for 12, 24, 48 and 72 h. ‘WT’ are the wild-type trophozoites. ‘AS-DPD’ are trophozoites with a downregulated DPD (Fig. 5A). ‘AS-Cont’ are trophozoites that were transfected with a control antisense vector (Fig. 5B). The number of trophozoites at the beginning of each experiment was at 100%. Data are expressed as the mean ± standard deviation of seven independent experiments.

Effect of LTGS on E. histolytica virulence

In an attempt to establish whether LTGS could alter E. histolytica virulence, we conducted a series of virulence tests. The results revealed that trophozoites cultivated in LGM were able to ingest or lyse erythrocytes at the same rate as trophozoites that were cultured in a HGM (data not shown). Nevertheless, LTGS increased the trophozoite's ability to adhere and destroy HeLa cells (Fig. 2). Interestingly, when the glucose-starved trophozoites were re-exposed to HGM for 3 h, their increased virulence phenotype was partially reversed.

Figure 2.

The effect of long-term glucose starvation on E. histolytica (A) cytopathic activity and (B) adhesion. The cytopathic activity and adhesion of E. histolytica trophozoites were tested under three conditions: ‘control’ – E. histolytica trophozoites that were grown in HGM; ‘glucose starvation’ – glucose-starved E. histolytica trophozoites that were kept in LGM for 1 month; and ‘supplying glucose to glucose-starved trophozoites’ – glucose-starved trophozoites after they had been re-exposed to HGM for 3 h. Data are expressed as the mean ± standard deviation of at least three independent experiments. The value of the intact HeLa cell monolayer before each experiment was set at 100%.

Glucose-starved trophozoites display global modulation in gene transcription

In an attempt to find the strategy that E. histolytica develops in order to survive LTGS and the response of glucose-starved trophozoites when glucose becomes again available, we performed a DNA microarray analysis of gene expression. For this analysis, we used the EH-IP2008 version of microarray (Santi-Rocca et al., 2012) which contains the entire E. histolytica genome. Three independent analyses were conducted and corresponded to three biological replicates. The microarray analysis was conducted as previously published (Santi-Rocca et al., 2012), and only genes whose levels of transcription changed by at least twofold (P ≤ 0.05) in each of the three biological replicates were deemed as being up- or downregulated (Tables 1-4). Genes that coded for hypothetical proteins are presented in Table S2. In total, during LTGS there was an upregulation in the transcription of 32 genes and downregulation in the transcription of 24 genes. Supplying glucose to the glucose-starved trophozoites by their exposure to HGM resulted in an upregulation in the transcription of 30 genes and downregulation in the transcription of 34 genes. Our findings show that LTGS upregulated the transcription of various genes, most of which (15 genes) were classified as virulence-related genes (Table 1), such as genes encoding for cysteine proteinase 4, Gal/GalNAc lectin, pore-forming peptides, lysozyme and surface antigen ARIEL1. In addition, we found that glucose starvation increased the transcription of the gene that encoded DPD, three amylase-coding genes, two genes that encoded amino acid degradation enzymes, two genes that encoded ribosylation factors and two genes that encoded ribonucleases. We also found that glucose starvation downregulated the transcription of four virulence-related genes that encoded mostly cysteine proteinases, four genes that encoded heat shock proteins, two genes that encoded alcohol dehydrogenases, three genes that encoded P-glycoprotein-6, three actin binding proteins and one gene that encoded a long-chain fatty acid ligase (Table 2). Re-exposure of the glucose-starved trophozoites to HGM for 3 h resulted in upregulation in the transcription of genes that encoded cysteine proteinases, glycoproteins and two heat shock factors (Table 3). This re-exposure to HGM also resulted in downregulation in the transcription of 34 genes of which one-third were classified as virulence-related genes (Table 4). Figure 3 lists the 17 E. histolytica genes whose transcription was most modulated by the changes in the concentration of glucose in the culture medium. Interestingly, more than a half of these genes are associated with E. histolytica virulence. To verify the data that were obtained by the microarray analysis, we performed a Northern blot analysis on two representative genes: DPD (EHI_012980) and HSP101 (EHI_183680) (Fig. 4A). The results of this analysis confirmed that LTGS and re-exposure to HGM modulated the transcription of DPD and HSP101 in E. histolytica (Tables 1-4).

Figure 3.

The most responsive glucose genes. A list of genes whose expression was modulated both in glucose-starved trophozoites and when trophozoites were re-incubated in HGM. The genes were classified according to their respective annotations. The fold change in gene transcription is displayed on the y-axis. A positive fold change indicates upregulation in the transcription of a gene, whereas a negative fold change indicates downregulation in the transcription of a gene compared with the control (trophozoites that were grown in HGM).

Figure 4.

Verification of the data that were obtained from the Affymetrix-based microarray by (A) Northern blot analysis and (B) Western blot analysis. The analyses were performed for two genes: the gene that encoded dihydropyrimidine dehydrogenase (DPD) (EHI_012980) and the gene that encoded HSP101 (EHI_183680). Parasites growth conditions were the same as those described in the legend to Fig. 2. The nuclear protein, EhMLBP, was used as an internal marker for the quality the cytoplasmic and nucleic lysates. The ‘control’ value was set at 100%.

Table 1. Genes that were upregulated in glucose-starved trophozoites
DescriptionGene IDFold changeRaw P-value
Surface antigen ariel1EHI_1313606.24.1E-08
ADP ribosylation factorEHI_1899604.49.2E-06
Dihydropyrimidine dehydrogenaseEHI_0129803.95.2E-05
LysozymeEHI_1991103.61.2E-04
Galactose-inhibitable lectin small subunitEHI_1598703.41.5E-04
Galactose-inhibitable lectinEHI_0583303.15.5E-05
Beta/alpha amylase precursorEHI_1535903.05.5E-04
Nonpathogenic pore-forming peptide precursorEHI_1693502.93.6E-05
CP-A4 cysteine proteinaseEHI_0505702.91.3E-03
Surface antigen ariel1EHI_1864702.73.2E-04
RibonucleaseEHI_1693002.72.4E-04
RibonucleaseEHI_1563102.51.5E-04
Ras family GTPaseEHI_0937602.53.1E-05
Lysozyme (N acetylmuramidase)EHI_1768202.46.8E-04
Ras family GTPaseEHI_0747502.43.8E-03
Methionine gamma lyaseEHI_1446102.49.8E-03
ARIEL1160.m000872.47.7E-04
Pore-forming peptide amoebapore A precursorEHI_1594802.38.6E-03
Surface antigen ariel1EHI_1698002.31.1E-03
Aspartate ammonia lyaseEHI_0822702.31.3E-02
Sulfate adenylyltransferaseEHI_1971602.29.2E-03
Surface antigen ariel1EHI_1728502.25.4E-04
Ganglioside gm2 activator proteinEHI_1518002.27.3E-05
FerredoxinEHI_1986702.23.6E-03
Variant surface proteinEHI_1286002.22.0E-02
Histone H4EHI_0232302.24.2E-05
Alpha amylase family proteinEHI_0233602.14.9E-03
S phase kinase associated protein 1AEHI_1349602.11.8E-05
Endo 1,4 beta xylanaseEHI_0962802.11.5E-03
ADP ribosylation factorEHI_0734802.13.3E-04
Dual specificity protein phosphataseEHI_1754902.11.4E-04
Alpha amylase family proteinEHI_1528802.12.2E-02
LysozymeEHI_0965702.07.8E-04
Table 2. Genes that were downregulated in glucose-starved trophozoites
DescriptionGene ID1/fold changeRaw P-value
Cysteine proteinase 2 precursorEHI_1326500.351.6E-03
Myosin heavy chainEHI_1101800.361.9E-03
Cysteine proteinaseEHI_0108500.362.4E-03
Glutamate synthase small subunitEHI_1105200.362.3E-03
CP-A7 cysteine proteinaseEHI_0396100.373.0E-03
Cysteine proteaseEHI_1440400.383.9E-03
L myo-inositol 1 phosphate synthaseEHI_0707200.399.3E-04
P glycoprotein 6EHI_1012300.421.3E-03
P glycoprotein 6EHI_0506400.422.4E-03
Long chain fatty acid CoA ligaseEHI_0793000.429.2E-03
P glycoprotein 6EHI_0811300.425.1E-03
Phosphoserine aminotransferase (serC)EHI_0263600.424.5E-03
Heat shock proteinEHI_1565600.441.0E-03
Actin binding proteinEHI_0940600.461.7E-03
Aldehyde alcohol dehydrogenase 2EHI_0242400.461.2E-02
Heat shock protein 101EHI_1836800.471.1E-03
Aldehyde alcohol dehydrogenase 2EHI_1609400.478.3E-03
NAD(P) transhydrogenase subunit alphaEHI_0140300.474.6E-06
Actin binding proteinEHI_1046300.481.8E-03
70 kDa peptidyl prolyl isomeraseEHI_1788500.488.0E-03
ActobindinEHI_0390200.482.9E-02
Rab family GTPaseEHI_0596700.482.7E-02
RhoGAP domain-containing proteinEHI_0495700.487.7E-04
Heat shock protein 101EHI_0764800.505.1E-03
Table 3. Genes that were upregulated by making glucose available to glucose-starved trophozoites
DescriptionGene IDFold changeRaw P-value
Specifically modified by glucose addition   
Cyst wall-specific glycoprotein JacobEHI_0289304.971.48343E-09
Cyst wall-specific glycoprotein JacobEHI_1363603.902.3913E-07
SulfotransferaseEHI_1973403.253.89816E-06
P glycoprotein 2EHI_1863502.668.50163E-07
Maltose O acetyltransferaseEHI_0296102.560.000126312
Myb like DNA binding domain-containing proteinEHI_0980702.563.3778E-05
TRNA nucleotidyltransferaseEHI_1670702.513.13194E-05
Protein kinase domain-containing proteinEHI_1099502.488.33705E-05
Phospholipase C precursorEHI_0880802.452.92472E-05
Zinc finger proteinEHI_1050802.380.000246688
Aspartate aminotransferaseEHI_0060802.382.04727E-06
Heat shock protein 70EHI_1554902.361.03253E-05
Phospholipase C precursorEHI_1032502.318.39923E-05
RNA modification enzymes, MiaB familyEHI_1418302.281.45914E-05
Phospholipase C precursorEHI_0677102.280.000238399
Transporter, major facilitator familyEHI_0081102.270.004448217
UBX domain proteinEHI_0741602.260.000164597
Membrane bound O acyltransferase (MBOAT) familyEHI_1527902.223.90746E-05
Ubiquitin activating enzyme E1 1EHI_0386902.220.000070198
Heat shock protein 70EHI_1803802.183.86517E-05
Leucine-rich repeat protein, BspA familyEHI_1484602.179.10774E-05
Acetyltransferase, GNAT familyEHI_1670802.130.001079861
Tyrosine kinaseEHI_0736602.090.000895998
CyclinEHI_1664202.026.85508E-05
Tyrosine kinaseEHI_1144202.020.000600703
Genes downregulated in glucose-starved trophozoites   
CP-A2 cysteine proteinaseEHI_1326502.800.000436015
CP-A7 cysteine proteinaseEHI_0396102.780.000510018
Long chain fatty acid CoA ligaseEHI_0793002.596.15449E-06
Cysteine proteinaseEHI_1440402.470.00187067
Cysteine proteinaseEHI_0108502.190.002942847
Table 4. Genes that were downregulated by making glucose available to glucose-starved trophozoites
DescriptionGene ID1/fold changeRaw P-value
Specifically modified by glucose addition   
Aspartate ammonia lyaseEHI_1503900.360.000377637
Serine threonine isoleucine-rich proteinEHI_0043400.360.000587878
Galactose binding lectin 35 kDa subunitEHI_0278000.360.000548566
Galactose-inhibitable lectin 35 kDa subunit precursorEHI_1834000.370.000775087
Protein kinase, putativeEHI_0119200.380.000354959
Alcohol dehydrogenase (adh2)EHI_1504900.390.001443214
Serine threonine isoleucine-rich proteinEHI_0257000.390.001559507
Molybdenum cofactor sulfuraseEHI_1946000.400.00181573
Galactose-specific adhesin light subunitEHI_0496900.416.45754E-05
Gal/GalNAc lectin heavy subunit213.m000610.430.000714561
Pyruvate:ferredoxin oxidoreductaseEHI_0510600.440.000900345
Serine threonine isoleucine-rich proteinEHI_0123300.440.005778511
DnaJ family proteinEHI_0229500.460.00215779
Gal/GalNAc lectin Igl153.m001710.470.002410569
Gal/GalNAc lectin heavy subunitEHI_0466500.470.000246107
Gal/GalNAc lectin Igl2119.m001180.470.000610285
Immuno-dominant variable surface antigenEHI_0153800.475.52883E-05
Fatty acid elongaseEHI_1128700.470.002403954
Kinase, PfkB familyEHI_1571200.470.004719816
Genes that were upregulated in glucose-starved trophozoites   
ADP ribosylation factorEHI_1899600.172.18667E-09
Dihydropyrimidine dehydrogenaseEHI_0129800.310.000100142
Galactose-inhibitable lectin, putativeEHI_0583300.357.99384E-07
Surface antigen ariel1EHI_1313600.353.74846E-05
Aspartate ammonia lyaseEHI_0822700.364.82649E-05
Nonpathogenic pore-forming peptide precursorEHI_1693500.420.002680838
Surface antigen ariel1EHI_1864700.456.80294E-05
ADP ribosylation factorEHI_0734800.470.003199305
Surface antigen ariel1EHI_1698000.480.000347271
Beta/alpha amylase precursorEHI_1535900.491.94917E-05
LysozymeEHI_1991100.490.01105852
Alpha amylase family proteinEHI_1528800.510.005630962
Genes that were downregulated in glucose-starved trophozoites   
Aldehyde alcohol dehydrogenase 2EHI_0242400.410.002433724
Rab family GTPaseEHI_0596700.450.006774084
Aldehyde alcohol dehydrogenase 2EHI_1609400.460.004420993

In order to determine the intracellular location of DPD protein, we performed a Western blot analysis on E. histolytica cytoplasmic and nucleic lysates using a polyclonal DPD antibody which we generated in immunized mice (Fig. 4B). The results of this analysis indicate that E. histolytica DPD is a cytoplasmic protein, and this location is the same as that of human DPD protein (Lu et al., 1992) and murine DPD protein (Porsin et al., 2003). Interestingly, the relative densitometry values for DPD expression in both the Northern and Western blots show that LTGS resulted in a more than threefold increase in DPD expression.

DPD is essential for the adaption of E. histolytica to LTGS

The discovery that E. histolytica trophozoites are able to adapt to the low glucose-containing environment encouraged us to seek those gene(s) which make this adaptation possible. DPD was chosen as the candidate gene because its expression was found to be significantly increased both in long-term glucose-starved trophozoites (Table 1, Fig. 4) and in short-term glucose-starved trophozoites (Tovy et al., 2011). Moreover, its expression was found to be significantly decreased when trophozoites cultivated in LGM were transferred to HGM (Table 4, Fig. 4). To test whether DPD gene expression is crucial for the adaption to glucose starvation, we decided to downregulate its expression. This downregulation of DPD expression was performed using antisense technology that was published previously by Ankri et al. (1999a). Since DPD expression is highly upregulated by glucose starvation, replacing of the original EhRPg34 promoter with DPD's own promoter allowed us to create a glucose starvation-inducible antisense vector (Fig. 5A). In order to verify whether the activity of the DPD is increased by glucose starvation, a CAT reporter gene assay was performed (Fig. 6). Accordingly, we found that DPD's promoter is indeed active and that its activity is induced in LGM (almost 4.5 times higher than that found in HGM). The results of the Northern and Western blot analysis revealed that DPD expression was indeed downregulated in glucose starvation when using the glucose starvation-inducible antisense approach (Fig. 7). The relative densitometry values of the Western blot analysis showed that the extent of DPD downregulation that was caused by glucose starvation was almost two times greater than that found in HGM (3.2 times less expression versus 1.8 times less expression). Once achieving the DPD-downregulated trophozoites, we then repeated the viability assay under conditions of glucose starvation in order to test their ability to adapt to this condition (Fig. 1). We found that the DPD-downregulated trophozoites are more sensitive to glucose starvation than trophozoites that were transfected with an empty antisense vector (AS control; Fig. 5B). After 48 h of glucose starvation, we found that 50% of the control trophozoites were still alive whereas only 12% of the DPD-downregulated trophozoites were alive (Fig. 1). In contrast, we found that the DPD-downregulated trophozoites and the AS control trophozoites have the same generation time when cultivated in HGM (8.7 h, data not shown).

Figure 5.

Schema of (A) DPD antisense inducible vector (AS-DPD) and (B) control plasmid (AS-Cont). The DPD gene was cloned into the pSA plasmid in its antisense orientation under the control of its own promoter.

Figure 6.

CAT reporter gene assay. The CAT gene was cloned into a pJST4 expression vector under the control of the DPD promoter. The expression of CAT gene was tested by Western blot analysis under two conditions: ‘control’ – HGM and ‘glucose starvation’ – LGM.

Figure 7.

(A) Northern blot analysis and (B) Western blot analysis on trophozoites with a downregulated DPD gene. ‘AS-DPD’ – E. histolytica trophozoites with a downregulated DPD gene. ‘AS-Cont’ – E. histolytica trophozoites transfected with a control antisense vector. The ‘AS-Cont’ value was set at 100%.

Entamoeba histolytica DPD use NADH but not NADPH as cofactor

DPD is the rate-limiting enzyme in the first three steps of pyrimidine catabolism, in which uracil and thymine are converted into β-alanine or β-aminoisobutyrate and use NADH or NADPH as cofactors (Fig. 8). Although sequencing of the E. histolytica genome has revealed the existence of a DPD homologue (Anderson and Loftus, 2005), nothing is known about its characteristics. In order to obtain information about this homologue, we measured the activity of DPD using a previously described DPD activity assay (West, 1991). Our results showed that E. histolytica DPD use NADH (specific activity of 1.4 μmol NADH min−1 mg−1) but not NADPH as a cofactor (no activity was detected with NADPH). Interestingly, we also found that DPD activity was markedly increased (1.5-fold) by glucose starvation (Fig. 9A). This result is consistent with our finding that DPD expression is upregulated (3.8-fold) in glucose-starved trophozoites (Fig. 4). In addition, during glucose starvation, DPD activity in the downregulated DPD trophozoites was 40% of that of the trophozoites that were transfected with an empty vector (Fig. 9B).

Figure 8.

Metabolic pathway of pyrimidine catabolism. Asterisks represent genes that encode for enzymes which were previously identified in the E. histolytica genome according to Anderson and Loftus (2005).

Figure 9.

DPD specific activity in (A) wild-type trophozoites and (B) DPD-downregulated trophozoites under glucose starvation. In A, two conditions were tested: ‘control’ and ‘glucose starvation’ which were the same as those described in the legend to Fig. 6. In B, only ‘glucose starvation’ condition was tested. ‘AS-DPD’ – trophozoites with a downregulated DPD gene. ‘AS-Cont’ – trophozoites transfected with an antisense control vector. The respective values of the specific activity of the ‘control’ in (A) (1.4 μmol NADH min−1 mg−1) and ‘AS-Cont’ in (B) (2.0 μmol NADH min−1 mg−1) were set at 100%. Data are expressed as mean ± standard deviation of at least three independent experiments.

Discussion

In individuals with amebiasis, E. histolytica trophozoites live in the colon under conditions where glucose availability is relatively scanty (Cummings and Macfarlane, 1997; Kellett et al., 2008). The ability of E. histolytica trophozoites to live under such conditions may explain the successful adaptation of the parasite to LGM.

To date, a limited number of studies have investigated the effect of glucose starvation on the virulence of eukaryotic pathogens (Cullen and Sprague, 2000; Eichinger, 2001; Fang et al., 2004; Tovy et al., 2011). In the present study, we reported that LTGS upregulated the transcription of CP-A4 and underlay the ability of trophozoites to destroy HeLa cells. This result is consistent with the findings of He et al. (2010), who reported that CP-A4 is necessary for invasive amebiasis, and Gilchrist et al. (2006), who found that CP-A4 was upregulated in a mouse model of amebiasis.

The GalNAc lectin is a major virulence complex that is responsible for the protozoan's interaction with host tissues (Petri et al., 2002). The lectin light chain has been previously reported to be involved in the parasite's virulence (Ankri et al., 1999b). Indeed, one of the lectin light chain subunits, Lgl1, had a fundamental role in cell lysis in short-term glucose-starved trophozoites (Tovy et al., 2011). In this study, we also found that two galactose-inhibitable lectin light subunit coding genes were upregulated in the glucose-starved E. histolytica trophozoites, and eight genes, which were mainly those which encoded its light chain, were downregulated during re-incubation with HGM. These findings can explain the increased adherence of long-term glucose-starved trophozoites to HeLa cells and the loss of increased adhesion of these long-term glucose-starved trophozoites when re-exposed to HGM.

To our surprise, we also found that LTGS upregulated several genes which encode for surface antigen ARIEL1. This antigen was shown to only exist in E. histolytica and to be expressed in particular in trophozoites which were harvested from hamster liver abscesses. Ariel is absent in the nonpathogenic strain Entamoeba dispar (MacFarlane and Singh, 2006; Santi-Rocca et al., 2008). The exact role of Ariel1 as a virulence factor and its contribution to the growth of E. histolytica in LGM will need further investigation.

Although the cultivation of E. histolytica in this study was performed in axenic culture, we found that LTGS upregulated several genes that encode for proteins that are associated with bacterial lysis, for example lysozymes and the pore-forming peptides, such as amoebapore A (Leippe and Herbst, 2004). This result suggests that a low glucose-containing environment elicits the parasite to use bacteria as a nutrient source (Bracha et al., 1982; Mirelman, 1987).

It has been recently demonstrated that EhCaBP1, a calcium binding protein (Jain et al., 2008), the C2 domain-containing protein kinase, EhC2PK (Somlata et al., 2012) and the hypothetical protein (EHI_160980) (King et al., 2012) are important regulators of erythrophagocytosis. In addition, the major haemolytic activity of E. histolytica has been associated with an alkaline phospholipase A activity (Vargas-Villarreal et al., 1998). None of these proteins had their expression differentially expressed in trophozoites cultivated in LGM, which may explain why LTGS had no effect on the erythrophagocytosis and haemolytic activity of the parasite. We have recently shown that short-term glucose starvation (STGS) enhances the haemolytic activity of the parasite (Tovy et al., 2011). In contrast, we showed here that LTGS had no effect on their haemolytic activity. These apparently conflicting results may be explained by the fact that STGS and LTGS represent two independent physiological conditions for the parasite as illustrated by the induction of lysine-rich protein 1 by STGS (Tovy et al., 2011) but not by LTGS.

The sequencing of E. histolytica's genome has allowed a reconstruction of its metabolic pathways, and it appears that amino acids play a bigger role in energy metabolism than previously thought (Anderson and Loftus, 2005). In our study, we found that LTGS caused significant upregulation in the transcription of two genes that encode enzymes which degrade amino acids, namely methionine gamma-lyase and aspartate ammonia lyase. Our finding is in agreement with that of Zuo and Coombs, who found increased consumption of several amino acids by E. histolytica and Entamoeba invadens when they were kept in a low glucose-containing growth medium (Zuo and Coombs, 1995), and Liebeke et al. (2011), who found that the consumption of amino acids by glucose-starved Staphylococcus aureus was high.

Previous studies on glucose-starved Saccharomyces cerevisiae and Escherichia coli revealed that these cells exhibited induction of heat shock-responsive genes and enhanced resistance to heat shock stress (Jenkins et al., 1988; Boender et al., 2011). To our surprise, we found that trophozoites cultivated in LGM had the transcription of several heat shock proteins downregulated despite the fact that these cells were under energetic stress according to the measure of their intracellular ATP content. Although we do not have an explanation for this phenomenon, it is interesting to notice that similar results have been reported for glucose-starved Plasmodium falciparum (Fang et al., 2004) and in nucleated red blood cells from rainbow trout (Currie et al., 1999).

In this study, we found that DPD is essential for the adaptation of the parasite to LTGS. The results of previous studies have shown that DPD's expression is upregulated in trophozoites which were isolated from the colon of E. histolytica-infected mice (Gilchrist et al., 2006) in human colorectal cancer and in toxic stress (Shirota et al., 2002; Williams et al., 2003). These latter findings led us to hypothesize that the DPD gene may be crucial for cell viability and survival. Since the end-product of the three-step enzymatic pathway of uracil degradation is β-alanine (Gojkovic et al., 2003), we speculate that β-alanine may have an important role in glucose starvation. Unlike α-alanine, β-alanine is not required for protein biosynthesis in cell, but has several other functions. In bacteria, β-alanine is associated with pantothenate (vitamin B5) and coenzyme A synthesis (Webb et al., 2004; Zrenner et al., 2009). In plants, β-alanine is thought to be involved in the adaptation to environmental stress (Mayer et al., 1990; Moran et al., 2000; Kaplan et al., 2004; Rizhsky et al., 2004). Although the exact role of β-alanine in environmental stress is not clear, it has been suggested that it has chaperone activity and acts as a γ-aminobutyric acid (GABA)-like signalling molecule (Fouad and Rathinasabapathi, 2006). In line with this idea, GABA was reported recently to increase during encystation induced by LGM in E. invadens (Jeelani et al., 2012). Further analysis of the β-alanine and GABA levels in E. histolytica trophozoites cultivated in LGM or in HGM will be necessary to better understand the involvement of these molecules in the adaptation of the parasite to different glucose environments.

In this study, we determined the activity of E. histolytica DPD, and found that NADH is used as a cofactor but not NADPH. Although this characteristic makes E. histolytica DPD different from the NADPH-dependent DPD in most mammals and some bacteria, a novel NADH-dependent DPD was recently discovered in E. coli (Hidese et al., 2011) suggesting that E. histolytica DPD is not an exception. The difference observed between the increase in DPD expression and DPD activity suggests that this latter value is underestimated and that some unknown limiting cofactors are required for optimal activity.

To summarize, we found in this study a new role for E. histolytica DPD in promoting the survival of E. histolytica trophozoites in a glucose-poor environment. Based on the result of our expression array analysis, it is well likely that an orchestra of genes is involved in the adaptation of the parasite to LTGS. Further studies will be needed to elucidate their individual role in this process. In the future, it will be interesting to investigate whether DPD inhibitors, which are widely prescribed to reduce the incidence of hand–foot syndrome in cancer patients treated with fluoropyrimidine-based therapy (Yen-Revollo et al., 2008), represent a promising treatment for amebiasis.

Experimental procedures

Ethics statement

Animal work has been approved by the Technion Animal Inspection Committee (Approval Number: IL0670611). The protocol adhered to the National Research Council (US) Guide for the Care and Use of Laboratory Animals.

Microorganisms and their cultivation

Entamoeba histolytica trophozoites strain HM1:IMSS were maintained under axenic conditions in Diamond's TYI-S-33 medium (HGM, final glucose concentration 7.35 g l−1) at 37°C. The same batch of bovine serum (BEIT HAEMEK) and BBL Biosate Peptone (Becton Dickinson) was used in this study to prevent fluctuation in the final concentration of glucose. Trophozoites in the exponential phase of growth were used in all experiments. For obtaining populations of E. histolytica trophozoites that could survive under LTGS condition, trophozoites in exponential phase were transferred to a Diamond's TYI-S-33 medium without glucose added (LGM, final glucose concentration 0.15 g l−1). This medium was replaced every 2 days until the establishment of a trophozoite population adapted to LGM. This selection process took around 2 weeks. These adapted trophozoites were cultivated for 1 month in the LGM before being analysed. In order to simulate the glucose shock during the parasite's migration to the liver, the glucose-starved trophozoites were incubated for 3 h in HGM Diamond's TYI-S-33 medium.

Measurement of glucose concentration in the Diamond's TYI-S-33 medium

Glucose concentrations in HGM and LGM Diamond's TYI-S-33 medium were measured using the Glucose Flex® reagent cartridge (Dimension® clinical chemistry system) according to the manufacturer's instructions.

Measurement of ATP levels

ATP levels in trophozoites cultivated in HGM or LGM were measured using the Promega's CellTiter-Glo® Luminescent Cell Viability Assay according to the manufacturer's instructions.

Viability assay under glucose-starved conditions

Trophozoites (1 × 106) which were previously cultivated under axenic conditions in HGM at 37°C were incubated in LGM for 12, 24, 48 and 72 h. For each time point, an aliquot of the trophozoite culture was stained with eosin (0.1% final concentration), and the number of living trophozoites was counted in a counting chamber under a light microscope.

Determination of cytopathic activity

The destruction rate of a HeLa cell monolayer by trophozoites was determined using a previously described protocol (Ankri et al., 1999a). Briefly, E. histolytica trophozoites (2.5 × 105) were incubated with a HeLa cell monolayer in 24-well culture plates at 37°C for 60 min. The incubation was stopped by placing the plates on ice and then washing of the plates several times with ice-cold phosphate-buffered saline (PBS) to remove the non-adhered HeLa cells. The adherent HeLa cells were stained with methylene blue (0.1% in 0.1 M borate buffer, pH 8.7). The incorporated dye was extracted from the stained cells by adding 1 ml of 0.1 M HCl at 37°C for 30 min. The colour was read in a spectrophotometer at OD660. The destruction rate of the cells was expressed as a function of the amount of dye that was extracted from the monolayer of HeLa cells.

Determination of adhesion

The adhesion of trophozoites to a HeLa cell monolayer was measured using a previously described protocol (Ankri et al., 1999a). Briefly, HeLa monolayers were first fixed with 5% formaldehyde. The fixed cells were washed twice with PBS, incubated with 250 mM glycine for 30 min at 37°C and then washed again with PBS. Trophozoites (2 × 105) were added to the wells that contained the fixed monolayers in 1 ml of serum-free Dulbecco's modified Eagle medium (DMEM), and then incubated at 37°C for 30 min. After gentle decanting (twice) of the non-adhered trophozoites with warmed (37°C) DMEM, the adherent trophozoites were washed using 200 mM galactose in ice-cold DMEM. The number of adherent trophozoites was counted in a counting chamber under a light microscope.

Microarray gene expression and statistical analysis

Two million trophozoites growing in either HGM, LGM or in a LGM and then for 3 h in HGM were recovered by centrifugation (550 g) and RNA purification and the microarrays experiments were carried using a previously published protocol (Santi-Rocca et al., 2008). Three independent biological replicates were performed, each of them having two technical replicates hybridized in dye swap. Comparison of microarray data was performed from standard growth versus starved growth or starved growth versus replenishment with glucose. Statistical analyses were carried out with the R software (http://www.R-project.org) and Bioconductor packages (http://www.bioconductor.org). A lowess normalization was first performed on all spots using the normalizeWithinArrays function of the limma package. The statistical test of differential expression was carried out using the Varmixt package. We thus considered those genes biologically relevant as being differentially expressed and presenting raw P ≤ 0.05 and fold change (FC) ≥ 2. The complete experimental details and data sets are available online at http://www.ebi.ac.uk/arrayexpress/ with the Accession Link: E-MTAB-1187. Differentially expressed genes were thus further annotated including BLAST hits against GenBank RefSeq, KEGG pathways, Gene Ontologies and AmoebaDB version 1.3.

Northern blot analysis

Total RNA was extracted from the trophozoites using the TRI Reagent® RNA isolation Reagent (Sigma). Briefly, the RNA samples (10 μg) were separated on a 1% agarose–0.3% formaldehyde gel in MOPS buffer (0.2 M morpholinepropanesulfonic acid, 50 mM sodium acetate and 5 mM EDTA, pH 7.0), and then blotted to GeneScreen membranes (NEN Bioproducts, Boston, MA). The RNA was cross-linked to the membrane by ultraviolet (UV) light (1200 J cm−1) in a UV Stratalinker (Stratagene), and then dried at 80°C for 2 h. The membranes were washed in hybridization buffer (0.5 M NaP buffer, 7% SDS and 1 mM EDTA, pH 7.2) and then blocked with 100 μg ml−1 salmon sperm DNA at 60°C for 1.5 h. Probes were labelled with radiolabelled (α32P)-dCTP using the Random primer DNA labelling mix (Biological Industries, Kibbutz Beit Haemek, Israel), and then cleaned on a G-50 column (GE Healthcare). Hybridization with the probes was performed overnight at 60°C. Each membrane was then washed several times at 60°C with washing buffer 1 (5% SDS, 40 mM NaP buffer and 1 nM EDTA, pH 7.2), and then with washing buffer 2 (1% SDS, 40 mM NaP buffer and 1 mM EDTA, pH 7.2). The membranes were then exposed to X-ray film (Fujifilm). Relative densitometry values were quantified using Image J software (NIH).

Western blot analysis

Entamoeba histolytica cytoplasmic and nuclear lysates were prepared using a previously described protocol (Lavi et al., 2008). The proteins in each lysate were separated on 10% polyacrylamide SDS-PAGE gel, and then transferred to a nitrocellulose membrane. The membrane was exposed to Ponceau S (Sigma) in order to verify the efficiency of the transfer. The blots were first blocked using 5% bovine serum albumin, and then incubated with a mouse polyclonal DPD antibody (1:500), which we generated in four mice which were immunized three times with a DPD recombinant protein. After the incubation with the DPD antibody, the blots were incubated with a HRP-conjugated goat anti-mouse antibody (1:10 000) (Jackson ImmunoResearch), and then developed by enhanced chemiluminescence. A mouse monoclonal actin antibody (Santa Cruz Biotechnology) was used as the control. The specificity of the generated DPD antibody was confirmed by Western blot analysis on recombinant protein GST-DPD (Fig. S1). Relative densitometry values were quantified using Image J software (NIH).

Construction of inducible antisense plasmid for DPD downregulation

For the construction of the inducible antisense plasmid, the DPD promoter sequence (350 bp upstream to the DPD gene) from the E. histolytica genome was amplified by PCR using the sense primer GGATCCAATAGTTTAAACCCGTTTT and the antisense primer AGCAAGAATTAATACCTTTCTTATTCC. This product was then cloned into the pGEM-T Easy vector (Promega), digested with BamHI/NotI and ligated to the previously reported plasmid pSA after removal of its insert, the EhRPg34 promoter, by BamHI/NotI digestion (Ankri et al., 1999a). In order to obtain the antisense DPD sequence, part of the DPD gene sequence (first 1000 bp) from the E. histolytica genome was amplified by polymerase chain reaction (PCR) using the sense primer ATGTGTGATATTGAACTTGTTGAC and the antisense primer CAAGTCTTTGTGGAAGAACTTTTGGAA. This product was then cloned into the pGEM-T Easy vector, digested with NotI and ligated to the DPD promoter that contained pSA vector, which was also digested by NotI. The antisense orientation of the DPD gene was verified by PCR amplification using the sense primer of the DPD promoter and the sense primer of the DPD gene.

CAT reporter gene assay

For the construction of the inducible antisense plasmid, the DPD promoter sequence (350 bp upstream to the DPD gene) from the E. histolytica genome was amplified by PCR using the sense primer GGATCCAATAGTTTAAACCCGTTTT and the antisense primer GGTACCAGCAAGAATTAATAC. This product was then cloned into the pGEM-T Easy vector (Promega), digested with SacI/KpnI and the ligated into the pJST4 expression vector (kindly provided by Prof. Anuradha Lohia, Department of Biochemistry, Bose Institute, India) after removal of its original actin promoter by SacI/KpnI digestion. In order to obtain the chloramphenicol acetyltransferase (CAT) sequence, the CAT sequence from the pEhNeo/CAT vector was amplified by PCR (Isakov et al., 2008) using the sense primer GGTACCATGGAGAAAAAAATCACTG and the antisense primer AGATCTTTACGCCCCGCC. This product was then cloned into the pGEM-T Easy vector, digested with KpnI/BglII and then ligated to the DPD promoter that contained the pJST4 expression vector, which was also digested by KpnI/BglII. Expression of the CAT gene was monitored under standard condition and in glucose-starved conditions by a Western blot analysis using a rabbit polyclonal CAT antibody (1:1000) (Sigma-Aldrich). A mouse monoclonal actin antibody (Santa Cruz Biotechnology) was used as the control.

DPD activity assay

The activity of DPD enzyme was determined spectrophotometrically by monitoring for 5 min at OD340 the decrease in absorbance that occurs when NADH is converted to NAD+ as reported previously (West, 1991). The reaction mixture contained 1 mM DTT, 0.2 mM NADH, 150 mM uracil and an E. histolytica cytoplasmic extract in PBS buffer (pH 7.4) in a final volume of 1 ml. Each measurement was performed on freshly prepared lysate since DPD activity was found to decrease over time (data not shown). The reaction was initiated by the addition of the uracil and run against a blank that comprised the identical reaction mixture without uracil. Specific activity of DPD was expressed as μmol NADH min−1 mg−1. The respective values of the specific activity of the ‘Control’ in Fig. 9A (1.4 μmol NADH min−1 mg−1) and ‘AS-Cont’ in Fig. 9B (2.0 μmol NADH min−1 mg−1) were set at 100%. The same experimental procedure was followed to measure the activity of DPD with NADPH as cofactor, except that 0.2 mM NADH was replaced by 0.2 mM NADPH.

Acknowledgements

This study was supported by grants from the Israel Science Foundation, the Rappaport Family Institute for Research in the Medical Sciences and the Deutsche Forschungsgemeinschaft (DFG) to S. A. and by a grant to N. G. from the French National Research Agency (MIE-08, Intestinalamibe). The authors also wish to thank Dr M.A. Dillies for her help in the statistical analysis of the results, Prof. Ami Aronheim for his help in the measure of intracellular ATP, Drs Marielle Kaplan and Tatiana Koutsenko for the measure of glucose in the medium and Dr Arieh Bomzon, ConsulWrite (http://www.consulwrite.com), for his editorial assistance in preparing this manuscript.

Ancillary