|Home | About | Journals | Submit | Contact Us | Français|
We identified in the Entamoeba histolytica genome a family of over 80 putative transmembrane kinases (TMKs). The TMK extracellular domains had significant similarity to the intermediate subunit (Igl) of the parasite Gal/GalNAc lectin. The closest homolog to the E. histolytica TMK kinase domain was a cytoplasmic dual-specificity kinase, SplA, from Dictyostelium discoideum. Sequence analysis of the TMK family demonstrated similarities to both serine/threonine and tyrosine kinases. TMK genes from each of six phylogenetic groups were expressed as mRNA in trophozoites, as assessed by spotted oligoarray and real-time PCR assays, suggesting nonredundant functions of the TMK groups for sensing and responding to extracellular stimuli. Additionally, we observed changes in the expression profile of the TMKs in continuous culture. Antisera produced against the conserved kinase domain identified proteins of the expected molecular masses of the expressed TMKs. Confocal microscopy with anti-TMK kinase antibodies revealed a focal distribution of the TMKs on the cytoplasmic face of the trophozoite plasma membrane. We conclude that E. histolytica expresses members of each subgroup of TMKs. The presence of multiple receptor kinases in the plasma membrane offers for the first time a potential explanation of the ability of the parasite to respond to the changing environment of the host.
The Gal/GalNAc lectin of Entamoeba histolytica mediates parasite adherence to the host and signals the initiation of cytolysis (41, 44, 45, 49). It is a heterotrimer consisting of covalently linked heavy (Hgl) and light (Lgl) subunits with a noncovalently linked intermediate (Igl) subunit (9, 36, 37, 43, 46). The Igl subunit of the Gal/GalNAc lectin has two known family members, Igl1 and Igl2. The Igl subunit has sequence similarity to the variant surface protein (VSP) of Giardia. We have previously identified a large number of proteins in the genome of E. histolytica containing CXXC motifs similar to those of Igl (8). Here we show that these CXXC-rich proteins form a large family of E. histolytica transmembrane kinases (TMKs) with highly variable extracellular domains homologous to Igl and VSPs of Giardia and with cytoplasmic kinase domains.
Amebic trophozoites have been demonstrated to persist in humans for longer than 6 months (21, 22). This prolonged period of infection suggests that the amebae evade the immune system. Other protozoan parasites, such as Plasmodium, Giardia, and Trypanosoma brucei, are also able to infect the host for long periods in spite of inducing robust immune responses. The mechanism(s) of persistence of these organisms is thought in part to be due to the variation of surface proteins. Plasmodium falciparum has three families of var genes that are independently expressed (29). The highest variation rate of these families is 2% per generation (52). Giardia encodes a family of 100 to 150 VSPs whose surface expression changes at a rate of one variation every 5 to 13 generations (38). T. brucei has a family of over 1,000 variant surface glycoproteins that change at a rate of 10−2 to 10−7 variations per generation (13, 51).
The discovery of the large family of CXXC-containing TMKs is of interest not only for their potential role in antigenic variation but also for their role in cell signaling. E. histolytica must respond to a wide variety of environmental stimuli as it excysts into a trophozoite in the intestinal lumen and enters the host by invasion of the intestinal mucosal epithelium. Invasion involves attaching to the epithelium and responding to that attachment event through signaling events via the E. histolytica Gal/GalNAc adherence lectin that lead to host cell killing. The changing host environment should necessitate having a variety of ways of sensing and responding to the host.
Here we report sequence and expression analysis of the TMKs in laboratory-cultured trophozoites. An oligoarray and real-time PCR were used to measure the expression in cultured trophozoites of the TMK genes. We demonstrate that there are six families of TMKs, with each having one or more family members expressed. In addition, anti-TMK antibodies were used to localize the TMKs to the plasma membrane of trophozoites, consistent with their proposed function in sensing the environment.
The genes were identified in the 7X assembly available from The Institute for Genomic Research (TIGR) and Sanger sequencing centers (http://www.tigr.org/tdb/e2k1/eha1 and http://www.sanger.ac.uk/Projects/E_histolytica/) by searching the database for homologs of Igl1. Genes with high sequence similarity to Igl1 were used to search the database and identify additional family members. Additionally, the Sanger assembly was translated in all six reading frames, and genes were identified by sequence similarity to known genes in the National Center for Biotechnology Information (NCBI) database. These sequences were then screened for genes containing sequences for three or more CXXC motifs, or kinase domains. Genes containing sequences for CXXC motifs but not kinase domains, transmembrane domains, or signal peptides were eliminated from the data set.
Genes were identified by sequence similarity to genes for amoebapores, cysteine proteinases, and the Gal/GalNAc lectin Igl, Lgl, and Hgl subunits. Additionally, genes were identified by examination of the translated Sanger assembly, which had been annotated to known genes in the NCBI database. Phagocytosis genes and control genes were similarly identified. BspA genes were identified in the translated Sanger assembly and then identified by sequence similarity in the TIGR assembly.
A 260-amino-acid alignment of the kinase domains of the TMK proteins was made to Hanks's kinase alignment (Protein Kinase Resource [http://pkr.sdsc.edu/html/pk_classification/pk_catalytic/pk_hanks_class.html]) using CLUSTALX (20, 61). One representative per family, called the query panel of kinases, was employed (http://pkr.sdsc.edu/html/pk_classification/pk_catalytic/query_panel.html). The alignment was manually optimized using Genedoc (39), and then sequences were analyzed using the PHYLIP v3.6 package (15) and bootstrapped using Seqboot, Protdist, Neighbor, and Consense. A subset of the sequences were then bootstrapped using Seqboot, Protpars, and Consense. The TMKs were broken into groups based on signature motifs found in the kinase domains and aligned using CLUSTALX and with manual adjustments using Genedoc.
Oligonucleotide probes typically ranging from 50 to 60 bases, and optimized for standard hybridization conditions, were designed using Array Designer 2.0 software (Premier Biosoft International, Palo Alto, CA). The selected probes were then analyzed by BLAST against the 7X assembly of the E. histolytica genome at both TIGR and Sanger. Probes were redesigned if they contained more than 75% sequence similarity with other target sequences or had a continuous stretch of complementary sequence exceeding 15 bases (28). In some cases it was not possible to design gene-specific probes. The actin probe was predicted to hybridize to several actin genes, the Jacob probe to all three Jacob genes, the EHCP1/2 probe to genes for both E. histolytica cysteine proteases 1 and 2 (EHCP1 and EHCP2), the Hgl family probes to all five Hgl genes but not homolog Sp1, and the Hgl1/5 probe to Hgl1 and Hgl5 genes. The Hgl1/5, Hgl2, Hgl3, and Hgl4 probes were more than 75% similar.
The oligoarray had probes to genes for amoebapores A, B, and C and homologs (32, 67), BspA homologs (25), actin, intergenic regions, L37a from mouse and human, chitin synthase (10), chitinase (11), Jacob (16), Jessie1-3 (65), EHCPs (6), EhRabs (53, 54), Vps26, Vps35, glycerate dehydrogenase (3), methionine gamma-lyase (62), phosphoglycerate dehydrogenase (2), Ebp1 and Ebp2 (55), L10 (7), ribosomal gene Sa, indigoidine synthase homolog (50), Hgl1-5 (35, 48, 58), Igl1 and Igl2 (8), Lgl1-6 (36, 59), Sp1 (an Hgl homolog), ferredoxin (26), Ariel1 (34), an HMW1 homolog (19), serine-rich E. histolytica protein (56), TMK genes, and other hypothetical surface genes (Table S1 in the supplemental material). TMK genes for which we did not generate a specific probe are shown in Table S2. Jacob is an amebic cyst wall glycoprotein expressed during encystation (16). EHCP1 and EHCP2 are highly homologous cysteine proteinases (6). Sp1 is a homolog of Hgl, recently identified in the TIGR database (B. Mann, personal communication).
A 200-nmol quantity of each probe (typically ranging from 50 to 60 bases optimized for standard hybridization conditions) was synthesized on an ABI 3900 DNA synthesizer (Applied Biosystems, Foster City, CA). The probe oligonucleotides were dissolved in 50% dimethyl sulfoxide at a concentration of 0.25 mg/ml and arranged in 96-well microtiter plates. The panel of probes, including control (housekeeping) oligonucleotides, was printed with two spot replicates on Corning UltraGAPS coated slides using an Affymetrix 417 arrayer (Affymetrix, Santa Clara, CA). Slide quality control was analyzed by hybridizing two randomly selected slides per batch (up to 40 slides/batch) with Cy3-labeled universal oligonucleotide probes. The hybridized slides were then washed and scanned with a ScanArray 4000 scanner (PerkinElmer Life Sciences Inc., Boston, MA). If the spotting quality standards were met, the batch was deemed satisfactory for analysis.
Trophozoites of E. histolytica strain HM1:1MSS were grown axenically at 37°C in TYI-S-33 medium (12) with 100 U/ml of penicillin and 100 μg/ml of streptomycin sulfate (Invitrogen, Carlsbad, CA). For growth curve analysis, amebae were grown until they became nonadherent but still viable (144 h), seeded into T25 flasks with 300,000 ameba per flask (Corning Life Sciences, Corning, NY), and grown for 12 to 144 h.
Human erythrocytes were isolated using Mono-Poly resolving medium (ICN Biomedicals, Aurora, OH) according to the manufacturer's directions. Erythrocytes were washed twice in 10 mM HEPES (pH 7.0), 140 mM sodium chloride, and 0.1% bovine serum albumin and then resuspended in the same buffer until use. One million log-phase trophozoites were grown in 50 ml of TYI-S-33 medium for 24 h in the presence or absence of 24 million erythrocytes per ml of medium.
Amebae were lysed with 2 ml of buffer RLT containing β-mercaptoethanol (the first component of the RNeasy kit from Qiagen, Valencia, CA). Samples were processed immediately or flash-frozen in liquid nitrogen and stored at −80°C until processing using QIAshredders, followed by the RNeasy mini kit, including all optional steps and a 5-min incubation with buffer RWI. Samples were treated on the columns with RNase-free DNase from Qiagen according to the manufacturer's directions (Qiagen). Samples were analyzed for residual DNA contamination by PCR using primers for Jacob (conditions are described below). Samples that contained residual DNA were retreated with DNase I (Roche, Indianapolis, IN) for 1 h at 37°C in a 100-μl total volume with 10 μl of 10× DNase I buffer (100 μM Tris [pH 7.5], 25 mM MgCl2 and 5 mM CaCl2) and 3 μl DNase I, repurified on RNeasy columns and rescreened for residual DNA contamination. The Agilent BioAnalyzer (Agilent Technologies, Palo Alto, CA) was used to assess RNA quality. The results were inspected to ensure that both ribosomal peaks were intact and that no degradation had occurred. Acceptable 260/280 ratios ranged from 1.8 to 2.1.
DNA oligoarray assay of gene expression used cohybridization of two fluorescently labeled cDNA targets, prepared from different samples. For routine oligoarray expression analysis, a previously described (24) indirect labeling procedure was used. Approximately 10 μg of RNA per sample and random hexamers were used for synthesis of cDNA containing amino-allyl-labeled nucleotides. The newly synthesized cDNA was then labeled by a covalent coupling of an appropriate cyanine fluor (CyDye postlabeling reactive dye pack; Amersham Biosciences Corp., Piscataway, NJ). In a typical oligoarray assay, the cDNA of one preparation (control) was labeled with Cy5, while the second cDNA (experiment) was labeled with Cy3. Both reactions were purified with a QIAquick PCR purification kit (Qiagen) for removal of the uncoupled dye. The labeling efficiencies of the purified target preparations were examined by spectrophotometry, as well as by calculations of the mass of cDNA and Cy5 or Cy3 dye incorporation. The nucleotide-to-dye molecular ratios were considered suitable for oligoarray experiments with a ratio of less then 50 nucleotides/dye molecule. Both targets were equalized based on the total amount of dyes incorporated before hybridization (24). The samples were mixed, dried by speed vac, and dissolved in hybridization buffer solution (50% formamide, 5x SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate] and 0.1% sodium dodecyl sulfate [SDS]). The cDNA-containing hybridization solution was then denatured, applied to the oligoarray (prehybridized in 5x SSC, 0.1% SDS and 1% bovine serum albumin), and hybridized at 42°C for 18 h. Following 5-min washes in 2× SSC-0.1% SDS and in 0.1× SSC, the slide was scanned using a ScanArray 4000 scanner (PerkinElmer, Wellesley, MA). Both Cy5 and Cy3 images of one experiment were analyzed with QuantArray 3.0 microarray analysis software. Normalization to median between both channels was used.
Real-time PCR (RT-PCR) primers (Table (Table1)1) were designed using Beacon Designer 2.0 (Premier Biosoft International, Palo Alto, CA). RT-PCR primers and oligoarray probes were designed independently; thus, the PCR fragment and the oligoarray probe represented different regions of the same gene. Each primer was analyzed against the TIGR E. histolytica database, and any primer that had significant sequence similarity to multiple genes was rejected. Thus, both the forward and reverse primers were specific for one gene, except actin, Jacob, and Hgl, which detected all family members and/or alleles in the genome. Optimal annealing conditions (determined by gradient PCR) were used to ensure specificity, and any PCR primer pair that produced more than one melt peak was discarded. PCR products that produced single melt peaks were analyzed by gel electrophoresis in 1.5% agarose-Tris-borate-EDTA, and if multiple bands were observed, the primer pair was discarded. Finally, all PCR products were sequenced using the forward amplification primer to verify specificity.
RNA was reverse transcribed using iQsuperscript (Bio-Rad, Hercules, CA) according to the manufacturer's directions. cDNA levels were measured using RiboGreen (Molecular Probes, Eugene, OR) with DNA of known quantity as a standard in a SPECTRAmax Gemini EM fluorescent plate reader, according to the manufacturer's directions. Before proceeding to analysis of cDNA samples, a no-reverse-transcriptase control PCR was done using primers to the cyst-specific gene Jacob to verify that there was no residual DNA contamination of the samples. All samples were analyzed in duplicate, and all time points were analyzed in triplicate. Each time point was thus represented in six wells during the real-time PCR assays, and the six wells were averaged after normalization to the RNA polymerase II gene's average (47). Two sequentially performed growth curves were analyzed in the real-time PCR assays to ensure reproducibility. All real-time PCR assays were quantitative to allow direct comparison of gene expression levels.
A PCR master mix consisted of 1,100 μl of iQSYBRGreen super mix (Bio-Rad, Hercules, CA), 1,100 μl of distilled H2O, 88 μl of forward primer (50 pmol/μl), and 88 μl of reverse primer (50 pmol/μl). To each well containing 2 μl of cDNA was added 25 μl of master mix. Duplicate assays were performed on each sample. Each assay included standards, no-DNA-control wells, and no-RT-control wells. The cycling conditions were 95°C for 5 min; 30 cycles of 95°C for 30 s, annealing for 30 s (see Table Table11 for annealing temperatures), and 72°C for 30 s; and 1 cycle of 72°C for 2 min 30 s followed by a 90-step melt curve increasing 0.2°C with a 5-s hold.
The kinase region of Tmk96 (ΔTMK96) was PCR amplified with the primers 5′-CAATTTAGAGAAGGAATTCCT-3′ (5′ primer) and 5′-TCACATTAATTGAAGATGTTTTAAAACAACA-3′ (3′ primer). This 1,000-bp fragment was cloned into TOPO NT/T7 (Invitrogen), in frame with an amino-terminal six-His tag via TA cloning. Bacteria were grown at 37°C to an optical density at 600 nm of 0.5 and induced with isopropyl-β-d-thiogalactopyranoside for 4 h, and the recombinant protein was purified with nickel agarose beads (Qiagen). Antibodies were raised to this purified recombinant protein by a 90-day protocol including three inoculations of New Zealand White rabbits with recombinant ΔTMK96 (Covance, Princeton, NJ), and the antibodies were purified from serum with a protein A column.
Soluble proteins were extracted from amebae by harvesting 5 × 107 trophozoites by incubation on ice for 10 min, followed by centrifugation (200 × g at 4°C for 5 min). The amebae were lysed in 10 mM sodium phosphate buffer with protease inhibitor cocktail I (Sigma, St. Louis, MO) per the manufacturer's directions. Membranes were then cleared by centrifugation (100,000 × g at 4°C for 1 h). Whole-cell lysates were prepared by sonication of 106 amebae in three 5-min pulses on ice. Large intact particles were eliminated by centrifugation (20,000 × g at 4°C for 30 min). All samples were then separated on 10% polyacrylamide gels and then electrotransferred to polyvinylidene difluoride membranes (Millipore, Bedford, MA). Nonspecific binding was blocked by preincubation with Tris-buffered saline with 5% bovine serum albumin and 0.05% Tween 20 (TBST). In order to detect proteins on the blot, either anti-ΔTMK96 rabbit serum (a dilution of 1:5,000) or preimmune serum (1:2,500) was added in TBST for 1 h at room temperature. Interactions were detected by the addition of peroxidase-conjugated goat anti-rabbit IgG (Sigma) and development with ECL (Amersham) per the manufacturer's directions.
E. histolytica trophozoites (106) were bound to glass coverslips in a 24-well plate for 30 min at 37°C in TYI-S-33 medium. Adherent amebae were washed twice in phosphate-buffered saline (PBS) and fixed in 3% paraformaldehyde for 30 min at room temperature. To make the plasma membrane permeable, amebae were treated with 0.2% Triton X-100 in PBS for 1 min. Nonspecific binding was blocked by incubation with 20% goat serum and 5% bovine serum albumin (Sigma) in PBS for 1 h at 37°C. After incubation with either the anti-ΔTMK96 rabbit polyclonal antibody (200 μg/ml) or anti-Gal/GalNAc lectin antibody (6 μg/ml) for 1 h at 37°C, the coverslips were washed twice before Cy3-conjugated goat anti-rabbit secondary antibodies (Jackson Laboratories, Bar Harbor, ME) were added at a 1:160 dilution for 1 h at 37°C. The coverslips were washed twice and mounted on slides with Gel/Mount (Biomeda, Foster City, CA). Confocal images were visualized using a Zeiss LSM 510 laser scanning microscope (Carl Zeiss, Inc., Thornwood, NY).
Previously we reported a family of CXXC-containing proteins with sequence similarity to Igl1 and Igl2 (8). With the completion of the genome project, it became apparent that the majority of these CXXC-containing proteins were members of an approximately 80-gene family with intracellular kinase domains. The extracellular domains shared sequence similarity to Igl and to VSPs of Giardia. The kinase domain of these proteins in most cases contained all five conserved kinase motifs: the key conserved glycine-rich motif, the K residue, and the HRDL, DFG, and APE motifs (Table (Table2).2). However, a few lacked the glycine-rich motif (Tmk01, Tmk24, Tmk45, and Tmk73), which helps coordinate the second phosphate of ATP (14). The glycine, K, and DFG motifs are implicated in the binding and orientation of ATP. The HRDL motif is involved in catalysis. The APE motif is responsible for anchoring the substrate and thereby influencing the specificity of interaction with the protein that is being phosphorylated (14, 20, 60). Despite the conservation of the essential kinase motifs, overall the kinase domains were divergent from other known kinases. The TMK family branched closely with the other protein kinase group IX (OPK IX), which are the TGFβ receptor and activin family of serine/threonine kinases, OPK VIII, which are a Raf family of serine/threonine kinases, and OPK XII, which are the casein kinase I family of kinases (Fig. (Fig.1).1). It was not possible to determine by sequence analysis alone if the TMKs were serine/threonine or tyrosine kinases. Sp1A from Dictyostelium discoideum was the most closely related kinase identified by sequence similarity in the NCBI database that has been functionally characterized. The Sp1A kinase has been shown to be a dual-specificity kinase that phosphorylates both tyrosine and serine/threonine residues (40). The Sp1A kinase was found to be phylogenetically within the TMK family (Fig. (Fig.1).1). Unlike the TMK family, SplA is a cytoplasmic kinase rather than a transmembrane kinase. Two TMKs (Tmk58 and Tmk89) clearly did not group with the TMK family but grouped closely with other protein kinase groups VI and VII (data not shown), suggesting that these may represent serine/threonine transmembrane kinases that have evolved separately from the TMK family.
The TMK family members were grouped based on sequence and phylogenetic analysis of their cytoplasmic kinase domains. A specific signature motif between the conserved HRDL and APE motifs was identified for each family (Table (Table2).2). The motif was 15 to 24 amino acids C-terminal to the histidine in the HRDL motif. The motif sequences were CC(I/V)KITDFGTSR (group A), KLTDFGS(A/S)R (group B), C(A/G)KLTDFGTC (group C), PITAKVTDFGTS (group D1), V(T/V)(C/X)KV(T/S)DFGTS (group D2), AKLSDFGTSR (group E), and VKVSDFGLS with a conserved tryptophan two residues N-terminal to the APE motif (group F) (Table (Table22 and Fig. Fig.2).2). Tmk58 and Tmk89, which did not group with the rest of the TMKs, contained an ITDFGLAKK motif. Group G TMKs lacked a conserved motif in the kinase domain and also lacked one or more of the conserved kinase domain motifs. Sequence similarity between family members was not limited to the kinase domain. Additional motifs were found in some families that can be used to identify group members between the cytoplasmic and kinase domain (Table (Table22 footnotes). Additionally, group D TMKs had a serine-rich region N-terminal to their kinase domain and a serine/threonine region C-terminal to their kinase domain. Serine/threonine rich regions are often found in kinases and typically regulate kinase activity (30). The kinase domains of many of the E. histolytica TMKs were a mix of serine/threonine and tyrosine kinase signature motifs, further complicating efforts to predict their kinase activity (Table (Table3).3). However, the sequence similarity to SplA of D. discoideum suggests that they likewise may be dual-specificity kinases.
An overview of the sequence motifs of the different subfamilies of kinases is shown in Table Table22 (all domains except kinase domains), Table Table33 (kinase domains), and Fig. Fig.2.2. The TMKs had extracellular domains of 36 to 2,117 amino acids (with 11 to 98% similarity between family members), a transmembrane domain, and a kinase domain. Groups B, C, and D shared significant sequence similarity in their extracellular domain to Igl1 (Fig. (Fig.2).2). Most extracellular domains were rich in CXC/CXXC/CXXXC motifs. Almost all of the TMKs had at least one CXXCXXGYY motif beginning approximately 25 amino acids N-terminal to the transmembrane domain. Like Igl1 most of the TMKs had many additional CXXCXXGYY motifs (Table (Table22 and Fig. Fig.2).2). A very similar C(D/E)XCXXG(Y/F)(Y)(G) motif was found in Igl, VSP of Giardia (1, 38), and laminin LE domains (27, 57, 63). In VSP of Giardia the CXXC motifs have not been crystallized but have been shown to bind zinc (38), to have their N terminus at the host parasite interface (38), and have nonreactive cysteines (42). In laminin the LE domains form a rod-like structure of mini-globular folds (57, 63). This is consistent with the CXXCXXGYY motifs forming a linear array of mini-globular folds. By analogy to the laminin LE domains, the TMKs may have a rod-like structure.
To characterize the expression profile of these genes, we constructed an oligoarray of the TMK genes, Gal/GalNAc lectin genes, and other putative surface virulence genes. Few changes in gene expression were seen when early phase (12 h) was compared with mid-log phase (48 h) and late log phase (96 h) was compared with nonadherent phase (144 h) (Table (Table4).4). When mid-log-phase (48 h) and late-log-phase (96 h) amebic cultures were compared, very little change was seen (data not shown). actin was clearly growth regulated (Table (Table4),4), as were the Gal/GalNAc lectin hgl genes, many but not all of which decreased significantly during late log phase (Table (Table44 and data not shown). This decrease is consistent with previous observations of hgl1, hgl2, and hgl3 gene expression (48). Expression of known genes was, in general, consistent with RT-PCR, Western, and/or Northern analysis (data not shown). We concluded that E. histolytica did not appear to growth-phase regulate expression of most putative virulence genes when grown under lab culture conditions.
The expression profile of 67 of the TMK genes during mid-log phase (72 h) is shown in Fig. Fig.3.3. At this time point, 19 genes (group A, tmk61, tmk65, and tmk72; group B2, tmk02, tmk08, and tmk74; group B3, tmk21 and tmk28; group C, tmk39 and tmk63; group D1, tmk40 and tmk56; group D2, tmk19, tmk44 and tmk46; group E, tmk22 and tmk54; group F, tmk59; group G, tmk06) (Fig. (Fig.3)3) showed hybridization values in trophozoites significantly greater than that for the cyst-specific transcript jacob (P < 0.05).
Real-time PCR was conducted on jacob, tmk31, tmk80, tmk96, and tmk98, all of which were predicted by oligoarray not to be expressed, and tmk19, tmk21, tmk63, tmk65, tmk71, tmk75, sa, hgl, and actin, all of which were predicted to be expressed. Gene expression was monitored sequentially in trophozoites in laboratory culture over a 12-day period. To allow comparison between time points, results were normalized to the average of three RNA polymerase II genes (Fig. (Fig.4).4). All real-time PCR results of the TMK genes were consistent with the oligoarray results. Significant variations in expression during laboratory culture were observed for tmk19, tmk63, and tmk79 (P < 0.01) (Fig. (Fig.4D,4D, ,4F,4F, and and4I4I).
One million trophozoites were grown in 50 ml of medium with or without a vast excess of erythrocytes (24 million/ml of medium) for 24 h. We did not observe significant changes in TMK gene expression during erythrophagocytosis (data not shown).
The kinase domain of Tmk96 was expressed in Escherichia coli and used to generate polyclonal antibodies. Multiple trophozoite proteins were detected with polyclonal anti-kinase domain antisera (Fig. (Fig.5A).5A). No bands were observed with preimmune sera (data not shown). The recognition of multiple proteins by the antisera raised against the Tmk96 kinase domain was not surprising given that the kinase domain is conserved between different TMKs. The predicted TMKs vary in size from 482 to 2,577 amino acids, and Tmk96 is in the largest subfamily (group B) of TMKs with multiple members expressed. Of the TMK genes that we detected with expression significantly above jacob the size range was from 686 to 2,577 amino acids (78 kDa to 294 kDa not accounting for potential posttranslational modifications). The smallest proteins observed on Western blots were of a mass consistent with that of a typical kinase domain (270 amino acids) and may represent TMKs that do not have an extracellular domain or whose extracellular domain has been cleaved off.
Permeabilized amebae showed a focal plasma membrane staining pattern with anti-TMK kinase domain antibodies (Fig. (Fig.5B).5B). No staining was seen with nonpermeabilized cells or with permeabilized cells stained with preimmune sera (data not shown). We concluded that the kinase domain was on the cytoplasmic side of the plasma membrane. The TMKs therefore appeared to be typical type I transmembrane proteins with an amino-terminal signal sequence and a predicted transmembrane domain preceding the kinase domain. The focal staining pattern contrasted with the uniform plasma membrane staining pattern seen with anti-Gal/GalNAc lectin antisera (Fig. (Fig.5C5C).
The most important finding of this work is the identification of a large family of over 80 transmembrane kinases in E. histolytica. Although eukaryotic-type transmembrane kinases are found in organisms from bacteria to humans (4, 64, 68), large families of TMKs have been previously described only in multicellular organisms (20). For example, we were not able to find any TMKs in the partially completed genome of the ameba D. discoideum (17), and only a few TMKs were found in the apicomplexan P. falciparum (4). The TMKs were also found in Entamoeba invadens, Entamoeba dispar, and Entamoeba moshkovskii, indicating that this family is conserved within Entamoeba (data not shown) (66). Multiple TMKs were observed to be expressed at the mRNA and protein levels, consistent with nonoverlapping biological functions for individual members of the TMK family.
The existence of multiple different extracellular domains of the TMKs suggests that each interacts with the host environment and signals into the parasite in distinct ways. The extracellular domains of the group B, C, and D TMKs had sequence similarity to Igl of E. histolytica (8), laminin LE domains (27, 57, 63), and VSPs of Giardia lamblia (1, 38). The sequence similarity is largely limited to a repeated CXXCXXGYY motif. In laminin the LE domains function as mini-globular folds arranged in tandem to form a rod-like structure. If the CXXCXXGYY motifs in the TMKs take on a similar conformation, then these motifs may function to help the extracellular domain of the TMKs project off the surface of the cell in a pilus-like manner. This may make them available for interaction with host cell factors. In each TMK subfamily there was one or more members expressed; however, most family members appeared not to be expressed under the conditions of laboratory culture. By RT-PCR we did observe that the expression of some TMKs varied between growth curves. This indicates that the expression of these genes may be dynamic. Whether the TMKs share with the Giardia VSPs the process of antigenic variation under different biological conditions remains to be determined.
The most significant feature of the TMKs is the kinase domain that, with the exception of two TMKs, is distinct from other known kinases. It is not possible, based on sequence analysis, to predict activity, as most have similarity to both the serine/threonine and tyrosine kinases. Interestingly, a closely related kinase, SplA from D. discoideum, is a dual-specificity kinase with both tyrosine and serine/threonine kinase activity (40). All of the essential kinase motifs were conserved, suggesting that these are functional kinases. We were not able to demonstrate kinase activity when the kinase domain was expressed in E. coli. Since all of the functional residues were conserved, the most likely explanation for this is that the kinase is not functional in E. coli or that the kinase domain is not able to phosphorylate the substrates we have used. It has been previously shown that some kinases were not functional when expressed in E. coli (18). Additionally, even if the kinase is functional, identification of a substrate is often the rate-limiting step in characterizing a kinase (5, 23). Further experimentation will be necessary to identify the substrate or interacting partners of the TMKs and determine if the TMKs are serine/threonine and/or tyrosine kinases. Phylogenetic and sequence analysis shows that there are six subfamilies of kinases with distinct motifs within the kinase domains. Some families had additional conserved motifs outside of the kinase domain. This would imply that the subfamilies may represent functionally different families of kinases in sensing (differences in extracellular domains) and signaling (differences in kinase domains).
The focal staining pattern of the TMKs distinctly contrasts with the uniform plasma membrane staining pattern seen with the Gal/GalNAc lectin of E. histolytica or VSPs in G. lamblia, both of which lack cytoplasmic kinase domains (38, 42, 45). This localization suggests that the TMKs form a focal multimolecular signaling complex in the plasma membrane (31, 33).
In conclusion, the work presented here may begin to explain how E. histolytica is able to persist in the host for long periods of time despite immune surveillance, as well as sensing and responding to host stimuli. The large families of TMKs described here could serve in both biological sensing and antigenic variation. The distinct extracellular and kinase domains of the TMKs suggest that each TMK may sense or interact with different host factors and cause a distinct signaling event in response to that environmental cue.
We thank Aaron J. Mackey and William R. Pearson for guidance with the bioinformatics analyses and Brendan Loftus and Neil Hall for access to the E. histolytica genome sequencing project data at the TIGR and Sanger sequencing centers. Barbara Mann provided the Sp1 sequence.
This study was supported by NIH grant AI26649 to W.A.P. B.D. was supported by the Biomolecular Research Facility of the University of Virginia. T.N. was supported by a grant for Precursory Research for Embryonic Science and Technology, Japan Science and Technology Agency, a Grant-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Culture, Sports, Science and Technology of Japan (16017307, 16044259, 15590378), and a grant from the Japan Health Sciences Foundation.
†Supplemental material for this article may be found at http://ec.asm.org/.