Work with me here: variations in genome content and emerging genetic tools in Entamoeba histolytica

Abstract

Entamoeba histolytica is the causative agent of amoebiasis, a significant source of morbidity and mortality in developing nations. Despite this, E. histolytica is understudied, leading few treatment options and a poor understanding of pathogenesis. Genetic tools have historically been limited. By applying modern approaches, it was recently revealed that the genome is aneuploid. Interestingly, gene expression levels do not correlate with ploidy, potentially highlighting the importance of RNAi in gene regulation. Characterization of the RNAi pathway has led to potent tools for targeted gene knockdown, and the advent of RNAi-based forward genetics. CRISPR/Cas tools for editing the endogenous genome are an exciting possibility on the horizon. We celebrate the gains that have made E. histolytica tractable and anticipate continued advances.

Amoebiasis

Entamoeba histolytica is the causative agent of amoebiasis. This organism spreads through fecal-oral transmission. The hardy, dormant, cyst stage can contaminate water or food. After accidental ingestion of the cyst stage, excystation occurs in the small intestine. Active, motile amoebic trophozoites (“amoebae”) colonize the large intestine, and both cysts and trophozoites are passed in the feces. E. histolytica is most commonly found in developing nations, infecting approximately 50 million people per year [1]. The infection has a wide range of clinical symptoms: asymptomatic infection, diarrhea, bloody diarrhea (dysentery), or fatal abscesses in other organs that arise after amoebae spread beyond the intestine. Children in endemic areas suffer a high burden of morbidity and mortality owing to this infection. In one recent study of diarrheal illness in children, E. histolytica infection was associated with the highest likelihood of death [2]. E. histolytica is estimated to cause 15,500 deaths per year among children and 67,900 deaths per year amongst adults [3]. Despite its impact on human health, E. histolytica is dramatically understudied. Treatment choices are limited and the basic mechanism of disease is not well understood.

The E. histolytica Genome

Annotations of the Genome

E. histolytica is generally tetraploid and has ~75% A+T content [4,5] (Table 1). The genome is estimated to comprise 31–35 linear chromosomes greater than 300 kb [4] and approximately 200 episomal 25 kb DNA circles encoding rRNA [6]. The genome was first annotated in 2005 with whole-genome shotgun sequencing [5]. The reference strain that was sequenced was HM-1:IMSS [ATCC number 30459], which was originally isolated from a patient with dysentery and is used in the vast majority of laboratory research studies. Because of the presence of repetitive tRNA arrays, LINEs and SINEs, assembly was a challenge. The final annotation contained 888 contigs [5], a much higher number than the estimated number of chromosomes; thus the assembly was incomplete. The genome size was determined to be 23.8 MB, with 9928 predicted genes [5]. Of these genes, 25% contained introns, with 6% containing multiple introns [5]. Interestingly, there was potential evidence for lateral gene transfer, with 96 genes that may have been acquired from bacteria, encoding metabolic enzymes or unknown proteins [5]. There was evidence for expansion of gene families including cysteine proteases, vesicle transport machinery (Rab and Arf family proteins), proteins that modulate the actin cytoskeleton (Rho GTPases, RhoGAPs, RhoGEFs), and ~ 270 putative kinases [5]. Fitting with its parasitic lifestyle, genes needed for the de novo synthesis of thymidylate, pyrimidine, purines, and fatty acids were absent [5]. Mitochondrial DNA was absent, fitting with the presence of mitosomes that lack DNA, and the absence of canonical mitochondria in this organism [5].

Table 1.

The E. histolytica genome and genetic tools at a glance.

Property	Notes	Reference
% A+T	~75%	[5]
Ploidy	Generally tetraploid, with variable ploidy at chromosomal and sub-chromosomal levels	[8]
Chromosomes	~31–35 linear chromosomes and ~ 200 episomal circles	[4][81]
Size	~26.8 MB	[8]
Genes	~8734	[8]
Introns	~25% of genes	[5]
Genetic Tool	Available in E. histolytica?	Reference
Transient transfection	Yes, plasmid episomes	[47,49]
Stable transfection	Yes, plasmid episomes with Neo or Hyg resistance genes	[48,50]
Overexpression	Yes, by using endogenous promoters, T7 promoter, or by raising selective drug concentration	[52]
Overexpression-based forward genetics	Yes	[77,78]
Regulatable expression	Yes, plasmid episomes (destabilization domains or tet- induction)	[51,61,62]
Integration into genome	No
Gene knockouts	No
Endogenous gene tagging	No
RNAi	Yes, modern trigger approach; older G3, antisense and shRNA approaches	[54,66,67,82]
RNAi-based forward genetics	Yes, genome-wide RNAi library and proof-of-concept genetic screen	[76]
Regulatable RNAi	No
CRISPR/Cas9	No, only one study showing editing of a plasmid episome	[79]
CRISPRi	No

The next annotation included 1496 scaffolds with 8201 predicted genes and uncovered new structural findings in 40% of the identified genes [7]. The genome size was reduced from the original estimate of ~23 MB to ~20 MB [7]. Genes were given descriptive names and were assigned GO terms [7]. This annotation found that 20% of the E. histolytica genome contains transposable elements and further helped to identify 460 new putative protein-encoding genes that were not seen in the original annotation of the genome [7].

The third, and most recent, annotation compared 11 clinical isolates and HM-1:IMSS. PacBio SMRT sequencing and Hi-C were used to obtain high coverage sequence of the HM-1:IMSS reference strain [8]. A final 26.8 MB genome was formed with 38 scaffolds and 8734 protein-encoding genes (Figure 1) [8]. Further characterization suggested that the genome is aneuploid, in that while it is generally tetraploid, sometimes ploidy is greater at sub-chromosomal and chromosomal levels (Figure 2a – c). Among the clinical isolates, there were distinct differences in ploidy that did not correlate with virulence. Ploidy was examined as the recent clinical cultures were adapted to maintenance in vitro. With in vitro culture, there were changes in ploidy, albeit subtle in some cases [8]. Interestingly, despite the noted variation in ploidy between strains and conditions, gene expression levels were essentially equivalent between strains and thus did not correlate with ploidy [8]. This contrasts with other organisms such as Leishmania, where higher ploidy does correlate with higher gene expression [9]. The relative stability of gene expression in E. histolytica may be due to the endogenous RNAi pathway that modulates gene expression at the post-transcriptional level (see below).

Figure 1. Overview of the current best annotation of the E. histolytica genome.

PacBio SMRT sequencing and Hi-C were used to sequence and scaffold the HM-1:IMSS reference strain [8]. Shown is a circus plot with newly defined scaffolds (grey; scale of 1:10 kb), prior scaffolds from previous genome annotations (purple), coding sequences (CDS) in the +orientation (green) and the -orientation (dark orange), tRNA (black), single copy genes (light orange), PacBio read coverage (black), and Illumina read coverage (blue). Reprinted with permission from [8].

Figure 2. Variability of the E. histolytica ploidy and the number of nuclei per cell.

(a) The ploidy pattern is generally tetraploid, but is variable at the chromosomal and sub-chromosomal levels. Shown is the ploidy of HM-1:IMSS Clone 6 2015 mock-12w. This strain was derived from HM-1:IMSS Clone 6 2001, maintained in vitro for 15 years, transfected with an empty vector and cultured for 12 weeks in the presence of G418 and tetracycline. (b) Magnification of the region boxed in magenta in panel A. A cyan rectangle shows scaffold 9. Within this scaffold, the first half is tetraploid and the second half is pentaploid (magenta arrow). (c) Schematic representation of scaffold 9 ploidy, showing the presence of 4 full-length (cyan) chromosomes, and one partial (magenta) chromosome that represents the second half of the chromosome. (d) DAPI staining analysis of xenic and axenic cultures of the strains 2592100 and DS4–868 shows that the nuclear content differs between xenic and axenic conditions. On the x-axis, DAPI fluorescence is on a logarithmic scale; the y-axis shows the number of nuclei as a fraction of the highest number of nuclei in each sub-class. Panels A-C are reprinted with permission from [8] and panel D is reprinted with permission from [37].

Species and strain differences

Prior to the 1990s, E. histolytica and the closely related species E. dispar were thought to be the same organism, exhibiting virulent and avirulent phenotypes respectively [10]. E. dispar is now recognized as a distinct species that is rarely virulent. However, E. dispar does have the capacity for virulence. E. dispar can be invasive and is sometimes detected in symptomatic infections, including liver abscesses [11–13]. E. dispar is capable of killing mammalian cells [14], but was not able to degrade the mucus layer in an intestinal explant model [15]. This supports a model where virulence is context-dependent, since E. dispar is less able to breach the mucus layer, but appears otherwise capable of causing tissue damage.

Besides E. dispar, other Entamoeba species that have been studied include E. moshkovskii and E. invadens, the latter a pathogen of reptiles. Comparing these four species shows they share 4704 gene families which include 21741 genes [16]. Among gene families unique to E. histolytica, 3 out of 4 encode surface proteins, the largest of which includes many BspA-like proteins [16,17]. BspA-like proteins have been characterized as surface proteins in E. histolytica previously [18,19] and are important in other species for adherence [20–22]. Also unique to E. histolytica are 18 Ariel1 proteins (surface antigen family proteins [23]), two orthologous serine-rich proteins, 3 peroxiredoxins, 7 cysteine proteases, and 12 members of the AIG1 family (proteins with the AIG1 domain, pfam04548) [16]. E. dispar has 13 distinct members of the AIG1 family [16]. E. dispar lacks the virulence factor cysteine protease CP-A5 as a functional protein and has fewer cysteine proteases than E. histolytica [16,17]. Despite many studies and clear evidence for genotypic differences between species, it has not been possible to define the salient differences that explain the virulence of these different species, likely because the differences are complex and dependent upon the genome context as a whole.

Prior to sequencing of the E. histolytica genome, differences in genetic markers were noted among isolates taken from individuals with asymptomatic, diarrhea/dysentery, or liver infection [24–26]. More recently, the presence of an AIG1 family gene, EHI_ 176590, was found to correlate with the virulence of isolates, when the HM-1:IMSS strain was compared to the KU27 and KU50 strains that were isolated from asymptomatic infections [27]. EHI_176590 is involved in surface formation protrusions and when overexpressed, adhesion was enhanced [27]. Interestingly, another AIG1 family gene, EHI_180390, was identified as differentially expressed between HM-1:IMSS and the Rahman strain that was isolated from an asymptomatic infection, as well as differentially expressed between different derivatives of HM-1:IMSS (see below) [28]. Other differences have been noted between HM-1:IMSS and Rahman [29]. Superoxide dismutase, peroxiredoxin, grainin 1, and grainin 2 are more highly expressed in HM-1:IMSS [29]. Proteins involved in adherence including SREHP, KERP1, and STIRPs, and cysteine protease CP-A5, are also more highly expressed in HM-1:IMSS [30]. Despite the many comparisons of strains with different virulence, key genetic determinants that explain virulence are not evident. Similar to differences among species, the differences among strains are complex and the influence of individual genes on virulence likely depends on the context of the strain as a whole.

Virulence can also change following maintenance of strains in culture. In a recent study, multiple clinical isolates of varying virulence were compared [31]. There were clear genetic profile differences that correlated with the virulence of these strains [31]. Encouragingly, these differences persisted even following prolonged axenization [31]. Despite this encouraging evidence of stability of virulence in vitro, historically, virulence changes have been noted following in vitro culture of the HM-1:IMSS strain. Many studies have compared gene expression levels in cell lines derived from HM-1:IMSS, and gene expression in HM-1:IMSS itself has also been evaluated in different conditions [32]. An AIG1 family gene, EHI_180390, was identified as differentially expressed between HM-1:IMSS and the less virulent derivative, UG10 [28]. In another study, under hyperoxia (exposure to 100% oxygen) conditions, genes involved in heat shock response were more highly expressed in HM-1:IMSS compared to a less virulent derivative strain [33]. Historically, two strains called “A” and “B” that are derivatives of HM-1:IMSS were separately maintained in culture. The A strain became less virulent, but there were no obvious genotype differences between A and B [34]. More recently, clonal lines derived from A and B were characterized as “pathogenic” if they led to large, sustained liver abscesses in a gerbil model, and “non-pathogenic” if they led to small, quickly resolved lesions [35]. Interestingly, in comparing the ability to induce abscesses with other conventional markers of virulence (e.g., erythrophagocytosis, cysteine peptidase activity, motility, cytopathic activity, and hemolytic activity), no correlations were noted [35]. In examining gene expression, there were 76 differentially expressed genes between A and B clonal lines, and 19 differentially expressed genes between B clonal lines that were pathogenic vs. non-pathogenic [35]. These findings suggest that there are multiple mechanisms for loss of virulence in culture, given the relatively minor differences between these strains [35]. These conclusions are further supported by a study that further examined some of the differentially expressed genes [36].

Ploidy and Maintenance of the Genome

As previously discussed, E. histolytica may be aneuploid, typically maintaining a tetraploid state with some instances of higher ploidy [8]. While ploidy of clinical isolates was not directly related to virulence, ploidy differences can occur after adaptation to in vitro culture [8]. Ploidy differences have also been noted in xenic vs. axenic conditions. Interestingly, recently axenized cultures had more nuclei and higher DNA content (Figure 2d) [37]. Two of the isolates examined had ~96–97% of cells containing one nucleus in xenic conditions, but in axenic conditions, 84–85% of cells contained one nucleus, and 13–15% of cells contained two nuclei [37]. Upon axenization, DNA content increased by 10-fold compared to xenic conditions [37]. These changes appeared reversible, since when cultures were returned to xenic conditions, a gradual drop in the number of nuclei occurred [37]. Notably, for the HM-1:IMSS strain, ~87% of cells contain one nucleus, ~11% contain two nuclei, and ~2% contain more than two nuclei [37]. Thus, not just variable ploidy at chromosomal and sub-chromosomal levels is a consideration, but a significant number of cells in the axenic conditions used for laboratory research studies contain more than one nucleus. The variations in ploidy may be due to a missing cell cycle checkpoint [38]. E. histolytica can undergo multiple rounds of S phase without completing cell division, and the frequency of this abnormal activation of S phase increases with time in culture [38]. The ability to tolerate increased ploidy and/or increased nuclei might tie back to the ability of E. histolytica to encyst as part of its lifecycle, since cysts have four 1–2n nuclei [4].

DNA replication and repair pathways are relevant to maintenance of ploidy. E. histolytica has homologs for DNA polymerases [alpha, delta, and epsilon] of polymerase family B and lesional repair polymerases, including one from family A, and Rev 1 and Rev3 [39]. There are five polymerase families encoded by bacteriophage [phi] and DNA transposons [5,7,39]. While it is not yet known if homologous recombination occurs in E. histolytica, there are homologous recombination related genes such as Mlh1, Rad51, Rad21, and Msh2, along with meiotic genes such as Mnd1, Spo11, and Dmc1 [40–44]. Homologous recombination-related genes are upregulated after exposure to UV-radiation [45] and upon growth stress [43]. A PCR-based method may have detected evidence of homologous recombination between chromosomal inverted repeats under growth stress conditions [43]. The possibility of meiosis in E. histolytica is supported by a short-read mapping study wherein polymorphism patterns suggested a decline in linkage that could have been due to homologous recombination events [46].

Genetic Tools

Transfection and Overexpression Approaches

A major advancement was the establishment of transient and stable transfection of E. histolytica with exogenous plasmid DNA (Table 1). Transient transfection was initially demonstrated following electroporation, while later studies established that lipofection was also effective [47,48]. When present on a plasmid, the reporters chloramphenicol acetyltransferase (CAT) and firefly luciferase were successfully expressed only when they were flanked by upstream and downstream sequences from highly expressed E. histolytica genes [47,49]. Stable transfection was demonstrated by including the neomycin resistance (Neo) marker gene in the plasmid construct and exposing transfectants to geneticin (G418) selection, typically at 6 μg/ml [50]. Neo or hygromycin (Hyg) resistance have since been widely used to obtain stable transfectants [51]. Following selection, plasmids are stably maintained as episomes indefinitely and replicate autonomously in proportion to antibiotic pressure [52]. Notably, stably transfected plasmids such as those used in [52] are able to replicate despite their lack of endogenous E. histolytica origins of replication. Even the endogenous circular chromosomes lack fixed origins and instead replication begins at multiple sites [53]. Increasing the concentration of the selective antibiotic is commonly used as a method to increase the level of exogenous gene expression, for example by increasing G418 from 6 μg/ml up to 50 μg/ml or more [52]. When antibiotic selection is discontinued, plasmids are generally still maintained for weeks to months before eventual loss [54].

While stable episomes have been useful tools, to date, there has been no successful manipulation of the endogenous genome. Stable integration of exogenous sequences has not been demonstrated, nor editing of endogenous sequences. Thus, overexpression has been a common approach to interrogate gene functions, and this is sometimes combined with dominant negative versions of the gene under study (for example, [55]). A widely used plasmid for overexpression studies, pEhEx, uses upstream and downstream regulatory elements from the cysteine synthase (CS) gene to flank the insert site for gene of interest, and it also includes Neo for stable transfection [56,57]. Other derivatives of this construct are also widely used, such as pKT-3M, which contains a 3xMyc epitope, and pKT-MG, which contains 3xMyc fused to GFP for protein tagging [57]. More recent approaches for protein tagging have included the use of a HALO tag [58,59] or a codon-optimized SNAP tag for protein localization [60].

Regulatable Expression of Exogenous Genes

For regulatable expression of exogenous genes, tetracycline (Tet) induction and destabilization domain approaches have been successful. In a two-plasmid approach for tetracycline regulatable expression, the tet-repressor tetR is present on one plasmid with a Hyg marker, and the second plasmid contains a Neo marker and the gene of interest with the tet operator (tet-o) positioned optimally upstream of the transcriptional start site [51]. A one-plasmid approach contains both tetR and the gene of interest on the same plasmid, with a hygromycin marker [61]. For both approaches, induction kinetics are similar [51,61].

In contrast with tet regulation at the level of transcription, in the destabilization domain approach, regulation at the protein level can be achieved. Destabilization domains have been successfully used in many organisms. Fusion of a destabilization domain to a gene of interest causes proteasomal degradation of the tagged protein, while the addition of a stabilizing compound prevents this degradation. Two destabilization domains have been evaluated in E. histolytica, ddFKBP and ddDHFR [62]. Both were appropriately responsive to their respective stabilizing compounds, and are compatible with N- or C-terminal tagging [62]. One point to consider is that when stabilizing compounds are not added, expression is slightly leaky, since in this scenario the tagged protein should be degraded, but is sometimes still detected at a low level [62]. However, it is apparent that the amount of protein expressed upon induction is substantially more than the basal leakiness of this system. Similarly, leakiness is a known challenge for tet-regulation in many systems.

Early RNAi Tools

Historically, the existence of an RNAi pathway was uncovered by an approach that unintentionally knocked down amoebapore A expression (Figure 3a – b) [63]. Amoebapore A encodes a protein with pore-forming activity that has historically been considered to be a secreted cytotoxin, owing to its pore-forming activity [64]. The discovery of RNAi knockdown in E. histolytica happened in the course of expressing amoebapore A in the antisense orientation, using 5’ and 3’ regulatory elements from the ribosomal protein RP L-21 gene. Expression of amoebapore A was downregulated, and this persisted even after loss of the plasmid. The mechanism by which amoebapore A was knocked down was not clear. Extending the approach towards other genes, other genes could not be successfully silenced, unless the same plasmid was used to express the new target gene in the background of the existing amoebapore A mutant strain. A clonal line derived from the amoebapore A mutant line was named “G3”, and has since been used as a tool to knockdown additional genes (Figure 3b). A significant limitation of the G3 amoebapore A mutant, and all mutants made using this approach, is that the expression of numerous off-target genes is also impacted [63] (see also dataset DS_1909b3c782 at amoebadb.org).

Figure 3. Comparison of commonly used E. histolytica RNAi tools.

(a) Schematic representation of pAP2-R2, with either the amoebapore A gene in reverse orientation (blue box), or the gene of interest (GOI) in reverse orientation (orange box). pAP2-R2 also contains a Neo selectable marker. Note that the circular plasmid is used for transfection, but is represented in a linear view for illustration purposes. (b) Stable transfection with pAP2-R2-amoebapore A leads to knockdown of amoebapore A, together with alterations in the expression of other genes (see text). A clonal line derived from this knockdown mutant was named “G3.” Stable transfection of G3 with pAP2-R2-GOI leads to knockdown of the GOI, together with amoebapore A, and alterations in the expression of other genes. (c) Schematic representation of pTrigger, with a 132 bp fragment of an endogenously silenced gene (“trigger”) upstream of a 500 bp fragment of a GOI. pTrigger also contains a Neo selectable marker. Note that the circular plasmid is used for transfection, but is represented in a linear view for illustration purposes. (d) Stable transfection with pTrigger-GOI leads to knockdown of the GOI.

Since the establishment of the G3 strain, a variety of alternative RNAi based tools have been used. One alternative method is expressing double stranded RNA (dsRNA) in the HT115 E. coli strain, which lacks the ability to degrade dsRNA, and either feeding the bacteria to E. histolytica or harvesting the dsRNA and incubating it with E. histolytica [65]. Expression of antisense RNA, by expressing target genes in the reverse orientation, or expression of short hairpin RNA have also been used [66,67].

The “Trigger” RNAi Tool

A key study isolated and characterized E. histolytica small RNAs (sRNA) that were associated with Argonaute (Ago) proteins, which are central to the RNAi pathway [68]. sRNAs were 27 nucleotides long, were polyphosphorylated and had a 5’G predominance [68]. sRNAs tended to map towards the 5’ end of genes (Figure 4a) [68]. There was an inverse correlation between abundance of antisense sRNAs mapping to a gene sequence and the gene expression level (Figure 4b) [68]. Genes with a high number of antisense sRNAs tended to be silenced or expressed at lower levels [68]. Importantly, sRNAs are relevant to prior E. histolytica RNAi approaches. In the G3 strain, 27nt sRNAs mapped to amoebapore A, and when additional genes were silenced in this background, new sRNAs mapped to the newly silenced genes [69]. There is no obvious Dicer homologue in E. histolytica, but a potential noncanonical Dicer and Argonaute homologues have been characterized [70,71]. There are three Ago proteins with distinct subcellular localization, with Ago2–2 localizing to the nucleus [71]. All E. histolytica Ago proteins contain Paz/Piwi domains that are essential for sRNA binding [71]. Thus, it has been proposed that the nuclear Ago2–2 may bind sRNA in the cytoplasm and bring it into the nucleus to mediate transcriptional gene silencing.

Figure 4. Endogenous antisense sRNAs negatively correlate with gene expression.

(a) Representation of a subset of E. histolytica genes (gray arrows) and the positions of mapped sRNAs (boxes). sRNAs that map antisense to genes are represented by boxes below genes and sRNAs that map sense to genes are represented by boxes above genes. sRNAs cloned in a 5’-P dependent manner are in black; sRNAs cloned in a 5’-P independent manner are in red; sRNAs cloned after immunoprecipitation with E. histolytica Piwi-related protein (EhPiwi-rp) are in blue. (b) Northern blotting analysis to detect mRNA and sRNA expression levels from a subset of genes in samples from three different E. histolytica strains, HM-1:IMSS, 200:NIH, and Rahman. The same blot was stripped and probed for each different sRNA (sRNA names are shown, per the naming in [68]). RT-PCR analysis was carried out to detect the expression of each corresponding gene. Reprinted with permission from [68].

Because endogenous small RNAs are polyphosphorylated, it was initially a challenge to exploit this pathway for knockdown of a gene of interest. It is not clear how these polyphoshorylated small RNAs are generated endogenously in E. histolytica, but in C. elegans, primary siRNAs arise from Dicer cleavage and are not polyphoshorylated, while secondary siRNAs arise from synthesis by RNA-directed RNA polymerase and are polyphosphorylated [72]. An approach was later devised that taps into endogenous sRNAs (Figure 3c – d). This approach involves using a 132 bp fragment of an endogenously silenced gene, taken from the 5’ end of the gene, where antisense endogenous sRNAs are mapped (Figure 3c) [54]. This gene fragment is referred to as the “trigger.” Fusing this trigger sequence to a ~500 bp fragment of a gene of interest, and expressing this exogenously, leads to highly effective silencing of the gene of interest (Figure 3d) [54]. This appears to occur via spreading of silencing driven by endogenous antisense sRNAs that map to the trigger sequence [54]. Following the use of trigger-mediated knockdown, 27 nt sRNAs can be detected that map to the newly silenced gene of interest. The trigger system leads to potent, long term silencing of the gene of interest [54]. This is particularly effective when clonal lines are obtained from heterogeneous transfectants, since there is clone-to-clone variation in the level of knockdown [73]. It also mediates epigenetic changes to the target, as knockdown persists even after removal of the plasmid, evidenced by the deposition of a repressive histone mark at the endogenous gene sequence [74]. The trigger approach provides a significant advantage over the G3 approach, since it is now possible to knockdown genes without having to do so in a background in which amoebapore A, and many other genes, are already impacted. Finally, it is also possible to knockdown two genes simultaneously using this approach [75].

The trigger system has since enabled RNAi based forward genetics in E. histolytica (Figure 5) [76]. Previously, forward genetic approaches were limited to the use of a plasmid-based overexpression library [77,78]. The trigger strategy was adapted for genome-wide RNAi knockdown, which involved the cloning of optimally sized sheared genomic fragments into the trigger plasmid (Figure 5a – b) to generate an RNAi library with full genomic coverage (Figure 5c) [76]. A pilot screen selecting for growth mutants validated the library and the approach [76]. The library was designed to accommodate Illumina sequencing of the gene fragment, so that deep sequencing can be used to identify and quantify gene sequences whose representation is changed following a genetic screen or selection [76]. The availability of a genome-wide knockdown approach for forward genetics significantly improves the tractability of this organism.

Figure 5. RNAi tools have enabled forward genetics in E. histolytica.

(a) The pTrigger RNAi library (pTrigger-library) contains random fragments of gDNA (represented by different colors) cloned into the pTrigger plasmid. After stable transfection, selection or screening can be performed, and Illumina deep sequencing is used to identify gene fragments and their quantitative representation. Non-selected transfectants (reference) serve as a control. (b) Schematic representation of pTrigger-library, shown with three representative gDNA fragments. The Neo marker enables selection for stable transfectants. gDNA fragments are flanked by custom Illumina TruSeq adaptor sequences (IA) to facilitate amplicon sequencing. These adaptors were designed such that after DNA isolation from transfected amoebae, only one round of PCR is needed to add the remaining Illumina sequences (shown in dark gray) and barcodes for Illumina deep sequencing analysis. PCR products are represented by lines, with colors corresponding to the gDNA fragment (green, yellow, or blue), the IA sequences (light gray), and the full Illumina sequences/barcodes (dark gray). (c) Visual representation of gDNA fragment coverage in the library, over a 100 kb of contig DS571145. The positions of genes are shown in black and green boxes; unique gDNA fragments are represented by lines. Fragments that map sense to genes are shown in black, and fragments that map antisense to genes are shown in green. Reprinted with permission from [76].

CRISPR/CasS Gene Editing

CRISPR/Cas9 is a powerful tool for precise genome editing, among other uses. To date, use of CRISPR/Cas9 in E. histolytica is limited to a single proof of concept study that demonstrated successful editing of an episome [79]. Cas9 expression was achieved by coupling Cas9 to a destabilization domain in order to regulate expression [79]. This inducible approach was necessary to overcome the toxicity of constitutive Cas9 expression, which was evidenced by the failure to recover transfectants when using Cas9 constructs lacking the destabilization domain [79]. To assay for Cas9 mediated editing activity, a luciferase expression construct containing a premature stop codon (pDeadLuc) and a repair template (pDonorLuc) were transiently transfected in amoebae that stably expressed Cas9. Luciferase activity was evident only in amoebae transfected all three constructs [79]. A PCR assay for luciferase repair demonstrated the presence of the recombined, repaired luciferase template [79]. The occurrence of recombination-mediated repair of the luciferase gene fits with the above described presence of homologous recombination related genes in E. histolytica, and the previous evidence for the potential occurrence of homologous recombination [54,79]. The major limitation of the CRISPR/Cas9 proof of concept study is that only an episomal sequence was edited [79]. Since there is currently no means of manipulating the amoebic genome, a system like CRISPR/Cas9 is sorely needed to target endogenous sequences for editing.

Priorities for Tool Development

Currently, a major limitation is the inability to knockout genes or directly edit the genome. Establishing CRISPR/Cas9 for precise editing of the endogenous genome is a key priority. This might prove challenging due to the A+T-richness of the genome, because Cas9 uses a G-rich PAM, though exploring other Cas variants such as Cas12a with a T-rich PAM could overcome this concern [80]. The ploidy of the E. histolytica genome could make it challenging to fully knockout a gene of interest, therefore strategies to select for homozygosity will be needed, particularly given the limited number of established selectable markers. Despite the challenges, establishment of CRISPR/Cas systems would vastly improve the tractability of this organism. Additionally, related tools for CRISPRi-mediated gene knockdown would be beneficial. Although RNAi is very effective in E. histolytica, CRISPRi could be advantageous for its potentially greater specificity. CRISPRi approaches use a 20 nt fragment of the target gene, which is much smaller than the ~500 bp fragment that is needed for trigger RNAi, and thus CRISPRi is generally less likely to impact off-target genes.

Aside from critical need for tools to edit the endogenous genome, additional needs include regulatable expression tools and modern tags for protein localization. While destabilization domains and tet-regulation have proven useful, both approaches can be leaky. Better tools for regulation of expression could be applied to existing overexpression and RNAi tools. Currently the trigger RNAi approach leads to constitutive gene knockdown, making it challenging to study genes that impact fitness. In addition to regulatable tools, there is a clear need for modern tools for protein tagging. GFP tags have been historically used for protein localization in E. histolytica. However, owing to the microaerophilic metabolism of E. histolytica, and the need for oxygen for GFP fluorescence, signal intensity is typically dim. Many other alternatives to GFP are now available that are more suitable for low oxygen conditions.

Concluding Remarks

Recent years have painted a clearer picture of the E. histolytica genome, uncovering that it is likely aneuploid. Ploidy varies between clinical isolates, and changes when isolates are axenized. Interestingly, despite the having variable ploidy, gene expression levels are generally consistent, potentially pointing to the key importance of RNAi in post-transcriptional gene regulation. Many studies have compared the E. histolytica genome with related species, or have compared the sequences of different clinical isolates. While some differences have been uncovered, the salient differences that clearly explain the differential virulence of different species or different isolates have not been defined, likely because these differences are complex and dependent upon the genome context as a whole.

Stable and transient transfection with plasmid episomes has been an established approach for many years. Overexpression, sometimes using dominant negative mutants, has been used to infer gene function. In more recent years, the advent of tools to exploit the endogenous E. histolytica RNAi pathway has enabled potent knockdown of individual genes and opened up the possibility of performing genome-wide RNAi knockdown screens. The possibility of applying CRISPR/Cas tools for gene editing in E. histolytica is very exciting. The successful editing of an episomal plasmid using CRISPR/Cas9 provides an encouraging proof-of-concept in this regard.

E. histolytica remains dramatically understudied, even relative to other parasites, though it heavily impacts human health. While it is clearly tractable, with many tools described here, the continued application of modern approaches is necessary. Much remains to be understood regarding the complexity of the aneuploid genome, and many gains could be made in the understanding of the organism and its virulence by developing tools for precise gene editing and regulatable gene expression. We celebrate the gains that have been made and look forward to continued advances in the E. histolytica genome and its associated genetic tools.

Glossary

Cyst

The infectious, dormant stage of E. histolytica, that has a cyst wall made of chitin and contains four nuclei. Excystation is the process whereby a cyst differentiates to produce eight active trophozoites. Encystation is the process whereby a trophozoite differentiates to produce a cyst.

Trophozoite

The actively dividing form of E. histolytica. It has an amoeboid morphology and is also referred to as an amoeba.

Tetraploid

Having a genome content of 4n.

Ploidy

The number of copies of the genome that are present in a single cell. A ploidy of 1n corresponds to a single copy of the genome.

Aneuploid

For an organism that has a genome containing multiple chromosomes, an aneuploid state is the condition of having unequal numbers of chromosomes. This could mean some specific chromosomes have extra copies, or are missing.

Axenization

Clinical isolates of E. histolytica are initially mixed cultures containing multiple different organisms. These cultures are called xenic. The process of separating E. histolytica from the other organisms is called axenization, and this produces an axenic culture.

Transfection

The process of introducing foreign DNA into a eukaryotic organism. Transfection can be transient, where the introduced DNA is temporarily present in the organism, but is then lost. Transfection can also be stable, in which the introduced DNA is maintained. Stable transfection typically involves the use of drug resistance marker genes, so that drug selection can be used to obtain cells that maintain the introduced DNA sequences.

Plasmid

A circular DNA molecule, typically 5–10 kb in size, that is convenient for molecular cloning techniques in order to introduce DNA sequences of interest.

Episome

A fragment of foreign DNA that is introduced into a eukaryotic organism and that is replicated and maintained without integration into the endogenous genome.

Tetracycline induction

A tool for transcriptional regulation of gene expression, that makes use of the tetracycline repressor from bacteria, that binds to a specific tet operator sequence when tetracycline in absent. Positioning the operator upstream of the transcription start site leads to a system where in the absence of tetracycline, the tet repressor blocks transcription of the gene, while in the presence of tetracycline, transcription occurs.

Destabilization domain

A tool for post-translational regulation of gene expression. Fusing the destabilization domain to a gene sequence leads to proteosomal degradation of the expressed fusion protein. When the specific stabilizing compound is added, which binds the destabilization domain, proteosomal degradation ceases.

Leaky expression

Aberrant expression of a gene in a regulatable expression system.

RNAi

The process of RNA interference, a post-transcriptional process in which a small RNA binds to the argonaute protein, the nuclease Dicer is recruited, and the mRNA complementary to the small RNA is cleaved by Dicer and subsequently degraded by other nucleases.

CRISPR/Cas9

A bacterial system that has been modified for gene editing in many organisms. The nuclease Cas9 binds to an sgRNA sequence that leads to cleavage of the gene sequence complementary to the PAM component of the sgRNA.

CRISPRi

A modified CRISPR/Cas9, in which the nuclease activity of Cas9 is inactivated to create dead Cas9 (dCas9). This enables dCas9 to block the transcription of a gene that is complementary to an sgRNA PAM. dCas9 is sometimes fused to additional transcriptional repressor domains to further inhibit transcription.