|Home | About | Journals | Submit | Contact Us | Français|
The unicellular eukaryote Entamoeba histolytica is a human parasite that causes amebic dysentery and liver abscess. A genome-wide analysis of gene expression modulated by intestinal colonization and invasion identified an upregulated transcript that encoded a putative high-mobility-group box (HMGB) protein, EhHMGB1. We tested if EhHMGB1 encoded a functional HMGB protein and determined its role in control of parasite gene expression. Recombinant EhHMGB1 was able to bend DNA in vitro, a characteristic of HMGB proteins. Core conserved residues required for DNA bending activity in other HMGB proteins were demonstrated by mutational analysis to be essential for EhHMGB1 activity. EhHMGB1 was also able to enhance the binding of human p53 to its cognate DNA sequence in vitro, which is expected for an HMGB1 protein. Confocal microscopy, using antibodies against the recombinant protein, confirmed its nuclear localization. Overexpression of EhHMGB1 in HM1:IMSS trophozoites led to modulation of 33 transcripts involved in a variety of cellular functions. Of these, 20 were also modulated at either day 1 or day 29 in the mouse model of intestinal amebiasis. Notably, four transcripts with known roles in virulence, including two encoding Gal/GalNAc lectin light chains, were modulated in response to EhHMGB1 overexpression. We concluded that EhHMGB1 was a bona fide HMGB protein with the capacity to recapitulate part of the modulation of parasite gene expression seen during adaptation to the host intestine.
Entamoeba histolytica is the causative agent of amebiasis and prevails in areas of poor sanitation. The organism is estimated to be responsible for 50 million infections and 100,000 deaths each year. Infection can lead to amebic dysentery, resulting from trophozoites invading the intestinal wall. Amebic liver abscess and other extraintestinal lesions can result through the spread of trophozoites from the intestine into the bloodstream. The host and parasite factors determining infection outcome have yet to be fully elucidated.
We hypothesize that alteration in transcription of certain crucial genes may contribute to the expression of the virulence phenotype. Distinct patterns of E. histolytica gene expression have been observed under a variety of experimental conditions (3, 6, 14, 18, 19, 23, 24, 27, 33, 43, 44, 67). Previously we catalogued the gene expression profile of E. histolytica associated with amebic colitis in the murine model (27). mRNA transcripts more abundant in vivo (mouse colon) than in vitro (laboratory culture) included putative DNA/RNA regulatory factors, representing a pool of potential phenotype-specific transcription factors. The transcript of the locus XM_652200/EHI_093800 was upregulated more than twofold at day 1 of amebic colitis. This gene showed significant sequence similarity to the high-mobility-group box chromosomal protein 1 (HMGB1).
HMGB1 is an abundant nonhistone nuclear protein and is a member of the HMG superfamily. The protein is highly conserved in all metazoans, plants, and yeasts and in the parasites Trypanosoma cruzi (49), Schistosoma mansoni (29), and Plasmodium falciparum (13). Mammalian HMGB1 is composed of two homologous HMG boxes and a highly acidic C-terminal tail. It is an L-shaped structure composed of alpha helices which bind DNA in the minor groove, resulting in a widened minor groove and a significant bend in the helix (50).
HMGB proteins recognize and bind a variety of DNA structures in a conformation-dependent manner including stem-loops, B-Z junctions, palindromes, four-way junctions, and single-stranded and cruciform DNA (16). They contain one or several copies of the HMG box DNA binding motif and have been shown to enhance sequence-specific DNA binding by a number of proteins involved in a variety of cellular functions, including nucleosome remodeling, transcription, recombination, and repair (1, 53, 61).
In the present study we characterized EhHMGB1 and confirmed that it functionally was an HMGB protein. Overexpression of this gene changed the levels of 20 of the transcripts that were previously reported as modulated in the mouse model of amebiasis. Markedly, the transcripts of Lgl3 and Lgl4, members of the gene family that encode the light subunit of the Gal/GalNAc lectin, an important virulence factor in E. histolytica, were upregulated, suggesting an important role for this protein in modulating the virulence properties of E. histolytica.
E. histolytica strain HM1:IMSS trophozoites were grown at 37°C in TYI-S-33 medium containing penicillin (100 U/ml) and streptomycin (100 μg/ml) (Gibco/BRL) (22). Stable transfection of E. histolytica trophozoites was achieved by lipofection as previously described (5, 51). Briefly, the amebae were washed and suspended (2.2 × 105 amebae per ml) in medium 199 (Invitrogen, California) supplemented with 5.7 mM cysteine, 1 mM ascorbic acid, 25 mM HEPES, pH 6.8 (M199s). Ten micrograms of plasmid was added to a tube containing 30 μl of Superfect (Qiagen) and incubated at room temperature for 20 to 30 min before being mixed with 1.8 ml amebae in M199s medium. This transfection mixture was incubated for 3 h at 37°C and then added to a flask containing prewarmed TYI medium. Transfected amebae were selected with hygromycin (15 μg/ml).
The entire 342-bp EhHMGB1 open reading frame (ORF) sequence was PCR amplified from HM1:IMSS genomic DNA using primers EhHMGB-1 and EhHMGB-2 (see Table S1 in the supplemental material) and cloned into pCR2.1-TOPO (Invitrogen). pEhHMGB1 was generated by replacing the Tetr gene from pGIR308 with the EhHMGB1 ORF as an XbaI fragment (55). The E. histolytica and Escherichia coli expression vectors were generated using either N-terminal Myc-His6 (primers EhHMGB 3 and 4) or Strep II-Myc (primers EhHMGB 5 and 6) epitope tags. The PCR products were cloned into a Gateway entry vector, pENTR/SD/D/TOPO (Invitrogen). The LR Clonase enzyme (Invitrogen) was used to transfer the recombinant EhHMGB1 cassette from the entry clone into the destination vectors pDEST14 and pAH-DEST for expression in E. coli and E. histolytica, respectively. The C-terminal acid tail deletion mutant was generated using primers EhHMGB 5 and 7. The point mutants were generated as described previously (32).
In short, a pair of complementary primers (EhHMGB 8 and 9 for the T34G mutation and EhHMGB 10 and 11 for the F56G mutation) carrying the desired base substitutions were used in a PCR with the entry clone carrying the recombinant wild-type EhHMGB1 cassette. The Accuzyme mixture (Bioline) was used to generate the PCR products followed by DpnI digestion and transformation. The expression of all the recombinant proteins (E. coli and E. histolytica) was confirmed by Western blotting and/or immunofluorescence and/or quantitative PCR.
The expression of the His6-tagged fusion protein was induced in E. coli BL21(DE3) cells with 400 μM IPTG (isopropyl-β-d-thiogalactopyranoside). The His6-tagged recombinant protein was affinity purified using nickel-chelate resin according to the manufacturer's directions (Qiagen). The Strep-tagged recombinant protein was purified using Strep-Tactin spin columns (IBA GmbH, Germany) per the manufacturer's instructions. Protein estimation was done using the bicinchoninic acid protein assay kit (Pierce).
Typically 1 ml of Trizol (Invitrogen) was used to extract RNA from 2 × 106 trophozoites. RNeasy (Qiagen) columns were used to enrich RNA greater than 200 nucleotides in length. Quantitative reverse transcription-real-time PCR (qRT-PCR) was used to independently measure mRNA abundance in amebae transfected with pEhHMGB1 or pGIR308 vector alone. The cDNA was synthesized using Superscript II (Invitrogen) and subjected to 40 cycles of amplification using HotStar Taq (Qiagen). Primers were designed to amplify 100 to 300 bp using genomic sequences from the E. histolytica Genome Sequencing Project (http://www.tigr.org/tdb/e2k1/eha1/ and http://pathema.jcvi.org/cgi-bin/Entamoeba/PathemaHomePage.cgi) and the Primer3 program (57). Continuous Sybr green I (Molecular Probes) monitoring during amplification using the MJR Opticon II machine was done according to the manufacturer's recommendations. All real-time amplification reactions were performed in triplicate, and the resulting fluorescent values were averaged. In all qRT-PCR experiments the cycle threshold values (the cycle number at which fluorescence exceeds the threshold value) were linked to the quantity of initial DNA after calibration of the effectiveness of the amplifying primer pair (28, 65). The relatively invariant fdx transcript was used to compensate for the variation in the amount of amebic mRNA isolated (see Table S1 in the supplemental material).
The Gateway construct expressing recombinant EhHMGB1 was introduced into HM1:IMSS trophozoites by transfection, and the transfected amebae were prepared for confocal microscopy as described previously (8). The coverslips were incubated in 20% goat serum and 5% bovine serum albumin (Sigma) in phosphate-buffered saline for 1 h at 37°C to block the nonspecific binding. After incubation with a 1:200 dilution of anti-Myc antibody (Santa Cruz) for 1 h at room temperature, the coverslips were washed three times before the addition of Cya-3-conjugated donkey anti-mouse secondary antibody (Santa Cruz). The nuclei were stained with DAPI (4′,6′-diamidino-2-phenylindole) before mounting. Confocal images were visualized using a Zeiss LSM 510 laser scanning microscope (Carl Zeiss, Inc.).
The circularization assay was essentially carried out as described previously (59). Briefly, the 32P-end-labeled 123-bp DNA fragments (Invitrogen) were preincubated at RT for 20 min with appropriate amounts of recombinant wild-type or mutant EhHMGB1 (0.25 to 4 μM) in T4 DNA ligase buffer (New England Biolabs [NEB]) in a 20-μl reaction volume. The DNA was then ligated with T4 DNA ligase (NEB; 40 units per reaction) at room temperature for 60 min. The linear DNA fragments were then digested with exonuclease III (NEB; 30 units per reaction) for 30 min at 37°C. Deproteinized DNA samples were then loaded onto 6% polyacrylamide gels in 0.5× Tris-borate-EDTA. The gels were dried and subjected to autoradiography.
Quality control of RNA samples was performed by use of the Agilent Bioanalyser Nano Assay. The standard protocol for hybridization of eukaryotic mRNA to Affymetrix arrays was followed. Two micrograms of total RNA was used for cDNA and subsequent biotinylated cRNA synthesis. This labeled RNA probe was hybridized to the Affymetrix custom array previously designed using information generated from the E. histolytica genome sequencing project (release date 8 December 2004) as described previously (27, 41). The Affymetrix probes were mapped to the new Genome Assembly and recognized 6,385 of the reannotated ORFs (78% of E. histolytica ORF 8197; http://pathema.jcvi.org/). The ORF probe sets were preferentially selected from the 600 bases proximal to the 3′ end of the E. histolytica sequences. The arrays were scanned with an Affymetrix Gene Chip scanner 7G, and report files were generated to determine the percentage of present calls of each array. The detection calls (present, marginal, or absent) for each probe set were obtained using the GCOS system (Affymetrix). Only genes with at least one “present” call were used in assessment of the data. Raw data from the arrays were normalized at the probe level by the gcRMA algorithm and then log2 transformed (36). Microarray experiments were performed in duplicate on biological replicates of RNA samples isolated in tandem, and a separate aliquot of this RNA was used to validate the array data. Control experiments were done with RNA from trophozoites harboring the parent vector alone expressing Tetr transcript and no amebic gene.
The data set used in this analysis was that of the reannotated E. histolytica genome of Caler et al. publicly available at http://pathema.tigr.org and GenBank (AAFB00000000).
Microarray data analysis was performed using the Array Data Analysis and Management System (VBI) (http://pathport.vbi.vt.edu/main/microarray-tool.php). The system uses tools such as Bioconductor (26) for analysis of the data. Statistical significance was determined for the microarray data using the LIMMA programs (as described in Results) (58, 64). Statistically significant P values were adjusted using the Benjamini and Hochberg method (9, 35). Statistical significance was determined for the qRT-PCR results using Student's t test.
The association of p53 with its target DNA was tested essentially as described by McKinney and Prives (47). Human recombinant p53 was purchased from Protein One Inc. Oligonucleotides carrying the p53 recognition site were hybridized to generate the double-stranded GADD45 p53 recognition sequence (see Table S1 in the supplemental material). The reaction mixture (20 μl) contained 20 mM HEPES (pH 7.9), 25 mM KCl, 0.5 mM EDTA, 10% glycerol, 2 mM MgCl2, 2.2 mM spermidine, 0.025% NP-40, 75 ng single-stranded DNA, and 0.15 nmol of labeled oligonucleotides. One hundred fifty nanograms of recombinant p53 and increasing amounts of recombinant EhHMGB1 or its mutants were added to the respective tubes and incubated at room temperature for 30 min. The reaction mixture was then separated by electrophoresis on a 4% native polyacrylamide gel containing 0.5× Tris-borate-EDTA.
The Entamoeba histolytica HM1:IMSS genome sequence released in 2005 (41) and updated by Caler et al. (http://pathema.jcvi.org) was the basis for annotation. The presence of the HMGB domain in EhHMGB1 (XM_652200) was validated using NCBI's conserved domain database. ORFs modulated by microarray analysis were analyzed for conserved domains using BLAST, PROSITE, or Pfam. In addition, each sequence underwent BLASTP analysis using the nonredundant protein sequence database to identify potential homologs. The EhHMGB1 sequence was compared to other reported HMGB1 proteins using the CLUSTAL W alignment tool (34).
The complete microarray data set has been deposited in GEO (Gene Expression Omnibus) under accession number GSE12204 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE12204).
EhHMGB1 (locus EHI_093800) was significantly upregulated in trophozoites isolated from the murine model of amebic colitis (27). The EBI-ClustalW tool was used to align the sequence of EhHMGB1 with known HMGB proteins (Fig. (Fig.1).1). Significant homology with HMGB proteins from a diverse range of species was observed. In particular, EhHMGB1 shared sequence similarity with HMGB proteins from Plasmodium falciparum (53% and 58% with PfHMGB1and PfHMGB2, respectively), Schistosoma mansoni (40%), and Homo sapiens (50%) (data not shown). Two residues have been predicted to be crucial to determine structural DNA specificity (13, 50), namely, a serine at position 10 and a hydrophobic residue at position 32 according to residue numbering of Drosophila melanogaster HMG-D. The corresponding conserved residues in EhHMGB1 were threonine at position 34 and phenylalanine at position 56 (Fig. (Fig.1).1). EhHMGB1 also had the acidic C-terminal tail seen in other eukaryotes. Thus, all these computational analyses supported the idea that EhHMGB1 was a bona fide HMGB protein. When EhHMGB1 sequence was used as a query against the E. histolytica database via BLASTP (NCBI), six proteins showed alignment. However, only one other protein (XP_653586) shared significant sequence similarity with EhHMGB1. Two HMGB proteins which were also modulated in the mouse model of amebiasis (EHI_086110 and EHI_087410) had little similarity to EhHMGB1 (27). Interestingly, analyses by ourselves and others have indicated that E. histolytica has only single HMG box-containing proteins (37).
HMGB proteins have been reported to bind and bend DNA sequences in a non-sequence-specific manner (59). A T4 DNA ligase-mediated DNA circularization assay has been successfully used to demonstrate the presence of functional HMG domains capable of bending DNA (59). We tested the ability of His6-tagged EhHMGB1 to bend a 123-bp DNA fragment sufficiently for T4 DNA ligase to circularize it. This resulted in the appearance of exonuclease III-resistant bands, indicating that they consisted of circularized DNA fragments (Fig. (Fig.2).2). In contrast exonuclease III digested most of the DNA in the reaction mixture lacking EhHMGB1 (Fig. (Fig.2A,2A, lane 3). The DNA bending activity was not an artifact of the His tag, as Strep-Myc-tagged EhHMGB1 gave similar results (Fig. (Fig.2B,2B, lanes 4 and 5). In addition, circularization was not observed in the presence of a His6-tagged non-HMG box-containing protein, His6-TMK54 (data not shown) (8). Comparison of a large number of HMG box domains identified two highly conserved residues—a serine at position 10 and a hydrophobic residue at position 32 (according to the residue numbering of Drosophila HMG-D ). The corresponding residues of EhHMGB1 (threonine 34 and phenylalanine 56) were mutated to glycine as described in Materials and Methods, and the resulting double mutant was tested for DNA bending activity. Mutations in the two conserved residues dramatically inhibited DNA bending (Fig. (Fig.2B,2B, lanes 8 and 9).
As expected, deletion of the C-terminal acidic tail did not affect the ability of EhHMGB1 to bend the DNA as seen by the production of circular products (Fig. (Fig.2B,2B, lanes 6 and 7). In fact this mutant appears to enhance the formation of circular products by at least 1.5-fold compared to wild-type recombinant protein (our unpublished densitometry analysis). We concluded that one or both of the conserved residues, but not the acidic tail, were essential for DNA bending function.
Although extracellular release of HMGB proteins under certain conditions has been reported (20, 42), most of the characterized HMGB proteins are localized in the nucleus (10, 30). To determine the subcellular location of EhHMGB1, trophozoites transfected with the plasmid expressing the Strep-Myc-tagged recombinant EhHMGB1 were examined by confocal microscopy. A clear staining of the nuclei was seen with anti-c-myc antibody (Fig. (Fig.3).3). This result was confirmed by the use of an anti-Strep antibody (data not shown). The control experiments, performed with secondary antibody alone or using the amebae transfected with vector alone, did not show any nuclear staining (data not shown). We concluded that EhHMGB1 is a nuclear protein.
Although several HMGB proteins from protozoa and helminths have been cloned and characterized (13, 21, 29), their role in gene regulation remains unknown. In metazoan eukaryotes the HMGB protein has been shown to promote the assembly of transcription factor complexes on specific DNA targets (1, 10, 39). The DNA binding activity of p53 is central to its biological function as a tumor suppressor in higher eukaryotes and is stimulated by its interaction with covalent as well as noncovalent modifiers (16, 40). Native HMGB1 protein purified from HeLa cells as well as its His6-tagged recombinant version enhanced p53 DNA binding in vitro (38). We tested the capacity of EhHMGB1 to enhance the binding of recombinant human p53 to its target sequence in vitro by an EMSA (47). The reaction mixture contained labeled oligonucleotides carrying a p53 binding site and recombinant p53 protein. Recombinant p53 by itself did not show significant binding under these conditions, but the addition of EhHMGB1 greatly augmented binding of p53 to its target DNA (Fig. (Fig.4).4). The double mutant, as expected, did not enhance the binding of p53 to its target sequence. Interestingly, the Δ-Acid mutant, which was able to support DNA bending, did not enhance binding of p53. EhHMGB1 and more prominently the Δ-Acid mutant, but not the double mutant, showed nonspecific binding to the probe at the concentrations used (Fig. (Fig.4).4). We concluded that wild-type EhHMGB1 shared the capacity of human HMGB1 to augment the binding of certain transcription factors to DNA.
Gene expression in E. histolytica trophozoites overexpressing recombinant EhHMGB1 was compared to that of amebae transfected with an empty vector by Affymetrix gene chips. EhHMGB1 was found to be expressed at least 100-fold over the basal level in these transfected trophozoites. Overexpression of EhHMGB1 resulted in the modulation of 33 unique transcripts. The putative functions of 31 of the 33 modulated transcripts were assigned by bioinformatics analysis (see Table S2 in the supplemental material). Of these, 20 were also modulated, although in an opposite direction, at day 1 or day 29 in the mouse model of intestinal amebiasis (Fig. (Fig.5).5). Four genes, known to be involved in virulence, were modulated, including genes coding for two of the five lectin light subunits (46) and one of the cysteine proteinases (EhCP-A7) (62). A fourth transcript, encoding a potential enterotoxic peptide, was also changed (see Table S2 in the supplemental material). These microarray results were verified by qRT-PCR (data not shown).
We have shown that EhHMGB1 is a homolog of mammalian HMGB protein. EhHMGB1 was able to bend DNA, which is a characteristic feature of all HMGB proteins. Mutation of two key conserved residues in EhHMGB1, which are required for DNA bending in other HMGB proteins, inhibited its ability to bend DNA, whereas deletion of the C-terminal acidic tail did not. EhHMGB1 enhanced binding of human p53 to its target DNA and, as expected, was located in the parasite nucleus. Overexpression of EhHMGB1 resulted in modulation of 33 transcripts involved in a variety of cellular functions, including 20 altered in amebae invading the mouse intestine. Most notable was the modulation of parasite genes encoding two of the light subunits of the Gal/GalNAc inhibitable lectin.
HMG proteins are one of the most evolutionarily conserved proteins in eukaryotes. They are abundantly expressed nonhistone proteins which play a role in nucleosome remodeling (39). All members of the HMG superfamily share a common structural motif called the HMG box. The HMGB proteins typically interact with specific DNA structures rather than sequence. They have been divided into three families: HMGA proteins, which interact with an AT hook; HMGN proteins, which interact with the nucleosomes; and HMGB (15). Structurally the HMGB domain consists of approximately 80 amino acids that fold into three alpha helices forming an L-shaped structure (7, 68). The vertebrate HMGB proteins have two consecutive L-shaped domains due to the presence of two HMG boxes. They also possess basic N termini and long acidic C-terminal tails. The two P. falciparum proteins which have been studied (PfHMGB1 and PfHMGB2) contain only one HMGB domain (13). In contrast, the six HMGB proteins that have been identified in S. mansoni and Schistosoma japonicum (21, 29) contain two HMGBs similar to mammalian HMGB proteins. The domain organization and amino acid sequence of EhHMGB1, which carries a single HMGB domain, is most similar to the Plasmodium HMGBs. Although both single and double HMGB domain-containing proteins have the ability to bend and stabilize DNA, it has been suggested that proteins with two HMGBs have a greater affinity for DNA, but the significance of an increased affinity is not known (45).
In metazoan eukaryotes the amino acids serine at position 10 and a hydrophobic residue at position 32 (domain numbering based on HMG-D) were predicted to be crucial for HMGB function by mutational and bioinformatics analysis (13, 50). In EhHMGB1 the equivalent residues are conserved (threonine, a conserved substitution, and phenylalanine, a hydrophobic residue). Point mutations of both these residues resulted in the loss of DNA bending ability, a proof of involvement of one or both of these residues in HMGB function (Fig. (Fig.2B).2B). This allowed us to assign EhHMGB1 as a classical HMGB family transcription factor.
PfHMGB1 and PfHMGB2 lack the characteristic C-terminal acidic tail of HMGB proteins in metazoan eukaryotes. The EhHMGB1 C-terminal acidic tail is comprised of nine acidic residues interrupted by a glycine and a serine residue. The exact role of the acidic tail in HMGB proteins is unknown, although it has been suggested that it interacts with the positive charges of histones (63). The much longer acidic tail of human HMGB1 has been shown to be repressive for DNA bending (31, 59). However, the deletion of the short C-terminal acid tail of SmHMGB1 did not affect the ability of this protein to support the ligation of a 123-bp DNA fragment (21). Similarly, the deletion of the EhHMGB1's acidic C-terminal residues did not change the function of the protein in ligase-mediated circularization assays. This mutant appeared in fact to enhance the formation of circular DNA, as seen from the densitometry analysis.
HMGB proteins are known to facilitate the binding of a variety of transcription factors to their cognate DNA binding sites (1). Proteins shown to physically interact with HMGBs include the HOX, p73, Rel, nuclear hormone receptor, and p53 transcription factors (2, 38, 47, 60, 66, 69). HMGB proteins operate as chaperones and augment the binding of p53 to linear DNA by presenting prebent DNA for p53 binding (47). It has also been suggested that HMGB remains associated with DNA only transiently and is released when the cognate transcription factor has bound (44). The wild-type EhHMGB1 possessed the capability to enhance the binding of human p53 to its target sequence in a manner similar to that of human HMGB1 (Fig. (Fig.4).4). This suggests that EhHMGB1 could provide an active architectural role by interacting with many potential E. histolytica transcription factors including the homolog of human p53, Ehp53 (17, 41, 48).
The acid tail of EhHMGB1 may have an important role in this process because the Δ-Acid mutant, although not defective in DNA bending, was defective in the enhancement of p53 transcription factor binding. Therefore, this suggests that the absence of an acid tail traps EhHMGB1 into a stable DNA complex and thus prevents the binding of other transcription factors (1). Bonaldi et al. have previously shown that the mutant HMGB1 lacking the acidic tail had 100-fold-increased DNA binding affinity but completely froze ACF (ATP utilizing chromatin assembly and remodeling factor)-mediated nucleosome sliding (11). Thus, the presence of the acidic tail seems to be important for transient interaction of HMGB proteins with DNA. As expected, the double mutant with mutations in the conserved amino acids Thr34Ala and Phe56Ala failed to bend DNA or to enhance the binding of p53. Future studies will be required to determine the biological role of EhHMGB1 in the assembly of E. histolytica enhanceosomes.
Microarray analysis of cells overexpressing EhHMGB1 showed a substantial overlap with the transcriptome profile of trophozoites isolated from the mouse model of amebiasis (Fig. (Fig.5A).5A). However, these transcripts were modulated in opposite directions (Fig. (Fig.5B).5B). One explanation for this unexpected result is the 100-fold overexpression of EhHMGB1.
Overexpression of EhHMGB1 alone without cooperating transcription factors could lead to a dominant-negative effect or another type of dysregulated gene expression. Nevertheless, these data can still provide an important insight into the EhHMGB1 regulon (see Table S2 in the supplemental material). These changes suggest that EhHMGB1 could be part of the trophozoite's response to sudden changes in the environment when it encounters the host. One could anticipate the need for the parasite to quickly adjust to changes in carbon source and the presence of existing microbial flora in the host gut as well as to host defense mechanisms. The EhHMGB1-modulated genes involved in cellular metabolism belong to various pathways including carbon metabolism. In addition, genes implicated in virulence were altered. Of particular interest were four members of the E. histolytica AIG (avrRpt2-induced gene) gene family, which are similar to plant AIG genes involved in bacterial resistance. The function of AIG genes in E. histolytica still needs to be defined (56). The most striking observation was upregulation of the transcripts encoding light subunits of the Gal/GalNAc inhibitable lectin that mediates contact-dependent cytolysis of host cells, an important hallmark of amebiasis. The Gal/GalNAc lectin is composed of a transmembrane heavy subunit (170 kDa) linked by disulfide bonds to the light subunit (31 to 35 kDa) and associated with an intermediate subunit (150 kDa). EhHMGB1 overexpression resulted in a significant increase in the levels of two of the five homologous genes that encode the light subunits Lgl3 and Lgl4 (4, 12, 25, 46, 52, 54).
To our knowledge, this is the first report for a unicellular parasite of a role for an HMGB protein in the regulation of parasite genes involved in pathogenicity. This suggests a role of EhHMGB1 in parasite adaptation to, and destruction of, the host intestine.
This work was supported by NIH grant AI-37941 to W.A.P.
We thank Ellyn Moore, Sarah K. Connell, and Christina Bousquet for excellent technical assistance.
Published ahead of print on 25 July 2008.
†Supplemental material for this article may be found at http://ec.asm.org/.