Profiling Schistosoma mansoni development using Serial Analysis of Gene Expression (SAGE)

Abstract

Despite the widespread use of chemotherapy and other control strategies over the past 50 years, transmission rates for schistosomiasis have changed little. Regardless of the approach used, future control efforts will require a more complete understanding of fundamental parasite biology. Schistosomes undergo complex development involving an alteration of parasite generations within a mammalian and freshwater molluscan host in the completion of its lifecycle. Little is known about factors controlling schistosome development, but understanding these processes may facilitate the discovery of new control methods. Therefore, our goal in this study is to determine global developmentally-regulated and stage-specific gene expression in Schistosoma mansoni using Serial Analysis of Gene Expression (SAGE). We present a preliminary analysis of genes expressed during development and sexual differentiation in the mammalian host and during early larval development in the snail host. A number of novel, differentially expressed genes have been identified, both within and between the different developmental stages found in the mammalian and snail hosts.

1. Introduction

Schistosomiasis is an important tropical parasitic disease that results from infection by various species of the blood fluke Schistosoma. More than 200 million people are estimated to be infected in more than 70 countries (Hotez et al., 2006). Currently, schistosomiasis treatment is based on a single agent, praziquantel, and there is evidence supporting the development of drug resistance in this parasite (Fenwick et al., 2003). Slow progress in vaccine development (Pearce, 2003) suggests we are limited by our lack of knowledge on how schistosomes evade the host immune response and persist in their host. Clearly more detailed information about the basic biology of schistosomes is necessary to facilitate vaccine and novel drug design (Colley et al., 2001; Pearce and MacDonald, 2002) or to discover new approaches to disrupting parasite development throughout its lifecycle.

As genome sequencing of S. mansoni is nearing completion, functional genomic studies will be essential to capitalize on the massive genomic database being generated (Wilson et al., 2006). Analysis of global gene expression will complement the genome sequencing effort through the identification of a substantial portion of all transcribed genes, thus providing a novel source of dynamic genome annotation. It is anticipated that these studies will provide important insights into mechanisms controlling development and host adaptation of this significant human parasite.

Methods for assessing gene expression on a genomic scale include DNA microarrays, Serial Analysis of Gene Expression (SAGE) (Velculescu et al., 2000; Velculescu et al., 1995; Velculescu et al., 1997), Massively Parallel Signature Sequencing (MPSS) (Brenner et al., 2000), differential display reverse transcriptase polymerase chain reaction (DD RT-PCR) (Liang et al., 1992; Liang and Pardee, 1992), and subtraction hybridization (Jiang et al., 2000; Jiang et al., 1995). The variety of gene expression profiling techniques facilitates adaptation to nearly any organism or question of interest.

The use of DNA microarrays has been widely applied to schistosomes and is considered the method-of-choice for “global” or genome-wide evaluation of gene expression among the mammalian host stages (Fitzpatrick et al., 2004; Fitzpatrick et al., 2005; Chai et al., 2006; DeMarco et al., 2006; Fitzpatrick and Hoffmann 2006; Gobert et al., 2006; Moertel et al., 2006) and those involved in larval development within the snail host (Vermeire et al., 2006). However, a major drawback of microarrays is their dependence upon pre-selection of probes. In our case, this would require a priori knowledge of expressed genes in S. mansoni (including sequence data for probe design), construction of microarrays containing every predicted open reading frame in the (currently incomplete) S. mansoni genome, or construction of probes from an exhaustively sampled cDNA library.

In contrast, SAGE requires no a priori knowledge of the genome, differentially expressed genes, or gene sequences. In SAGE, a short sequence tag (21 bp) from a unique position of an mRNA molecule is used to uniquely identify the source gene from within the genome (Velculescu et al., 1995). Sequence tags are isolated from the mRNA pool of a cell and are linked together to form long concatenated molecules that are cloned and sequenced. The population of tags defines patterns of expression of individual genes (Fig. S1). Quantification of all tags provides a relative measure of gene expression (i.e., mRNA abundance). SAGE thus provides both the identity of expressed genes and levels of their expression. The material benefits of SAGE are its simplicity, low cost, and efficiency.

The utility of SAGE and other functional genomics approaches are greatly increased if complete genome sequences are known or large amounts of genomic shotgun sequences are made available. In this way, the ongoing S. mansoni genome project has augmented the power of functional genomics investigations of S. mansoni’s biology, development, adaptation and sexual differentiation (Wilson et al., 2006). Since SAGE identifies complete mRNA transcripts from 21 bp tags, measures of relative expression can be directly connected to genome sequences or predicted ORFs within the genome (Velculescu et al., 1995; Velculescu et al., 1997). Such a combination of SAGE expression data and a known genome sequence provides the ability to confirm ORF predictions, detect unpredicted ORFs, and annotate a complete genome with gene expression information. However, Long-SAGE is equally promising for parasites lacking available EST or genome information (Coyne et al., 2004). High-throughput methods exist to identify full transcripts from SAGE tag sequences (Chen et al., 2000, 2002a,b) and, in some cases, SAGE tag sequences have been used to prime 5′ Rapid Amplification of cDNA Ends (RACE) for isolation of nearly complete transcripts (AGM, unpublished).

SAGE has been used to identify nitric oxide responsive changes in gene expression in S. mansoni following a 3 hour exposure to sodium nitroprusside, a nitric oxide donor (Messerli et al., 2006). Nitric oxide is a gaseous intercellular signaling molecule that also plays a role in host-parasite relations. Adult worms showed only a modest response to nitric oxide, with 13 SAGE tags upregulated 2-fold or greater in treated parasites. In addition, a total of 8 tags were uniquely present in the sodium nitroprusside-treated library and 21 tags uniquely present in the control library. Overall, this study indicated that nitric oxide does not rapidly induce large-scale changes in schistosome gene expression after a 3 hour exposure, but that expression of particular genes of interest, such as superoxide dismutase, cyclic nucleotide phosphodiesterase and calcineurin, appears to respond to nitric oxide.

In this study, SAGE was used to profile gene expression during development and sexual differentiation of S. mansoni. Ten stage-specific SAGE libraries have been constructed and sequenced and a preliminary analysis is presented.

2. Materials and methods

2.1. Parasite material used in SAGE library construction

Cercariae were obtained from infected Biomphalaria glabrata snails obtained the Biomedical Research Institute in Rockville, MD. Juvenile liver worms were collected by perfusion of NIH-Swiss mice 23 days after percutaneous infection of the tail skin by cercariae for 1 hr in water containing 500 S. mansoni cercariae (Lewis, 1998). Adult worms were collected 7–8 weeks after percutaneous infection of NIH-Swiss mice with 150–200 cercariae. To obtain adult worms from single-sex infections, NIH-Swiss mice were infected with 500 cercariae obtained from a single snail infected with a single miracidium hatched from eggs isolated from infected mouse livers (Lewis, 1998). A single miracidium infecting a snail will produce a clonal population of cercariae that are all of the same sex and infections resulting from these cercariae will produce all male or all female parasites (Kunz, 2001). Worms were identified as male or females based on known characteristics; males from single-sex infections closely resemble those from mixed sex infections (and females will be absent), while females from single-sex infections are stunted in size and exhibit an immature reproductive system (Kunz, 2001). This study was approved by the Institutional Animal Care and Use Committee of Illinois State University (08–2002; DHHS animal welfare assurance number A3762-01).

Free-living miracidia and in vitro cultured mother sporocysts were obtained as previously described by Yoshino and Laursen (1995). S. mansoni eggs were recovered from homogenates of livers removed from mice at 7–8 weeks post-infection. Miracidia were hatched from the eggs in sterile artificial pond water and concentrated on ice in conical polypropylene tubes. Cold-immobilized miracidia were either centrifuged and harvested for RNA immediately or transferred to a 24-well culture plate to permit transformation into primary or mother sporocysts. Sporocysts were cultured for 6 or 20 days at 26°C in either sporocyst medium (Ivanchenko et al., 1999) or sporocyst media ‘conditioned’ with B. glabrata embryonic (Bge) cell factors (Vermeire et al., 2004). Culture media in all wells of sporocysts were changed at two-day intervals. On 6d and 20d developmental time points, total RNA from all cultured sporocysts was isolated as described below.

2.2. RNA isolation and SAGE library construction and sequencing

Total RNA was isolated with Trizol reagent (Roche Biochemicals, Indianapolis, IN) according to manufacturer’s protocols. The I-SAGE Long kit (Invitrogen, Carlsbad, CA) was used to construct LongSAGE libraries according to manufacturer’s protocols, with the exception of the use of pGEM3Z as the cloning vector. Plasmid sequencing templates were prepared from 1.2 ml cultures using alkaline lysis as performed by a RevPrep Orbit robotic workstation (GeneMachines, San Carlos, CA). Sequencing reactions were set up on a Tecan Miniprep 75 workstation (Tecan US, Durham NC) and performed in an ABI 9600 PCR machine. Recombinant DNA sequencing was performed at the Marine Biological Laboratory’s Josephine Bay Paul Center on an ABI 3730xl capillary DNA sequencer.

2.3. SAGE data analysis

Sequences collected were analyzed with software created at the MBL specifically for SAGE analysis. The SAGE software parses out individual SAGE tag sequences from the DNA sequencing results according to the SAGE sequence grammar, excludes tags with sequence ambiguities, and reduces all SAGE tags to a look-up table of unique SAGE tag sequences and their observed frequencies among all of the S. mansoni SAGE libraries. SAGE tags not detected more than once in at least one SAGE library were excluded from analyses as putative sequencing error, unless the tag sequence had a perfect match to all available genomic and EST data. The unique tag sequences were mapped to available S. mansoni DNA sequences to determine the identity of expressed genes. These included ongoing TIGR/Sanger genome assemblies, TIGR S. mansoni gene index sequences (http://www.tigr.org/tdb/tgi), sequences from the S. mansoni EST project (http://verjo18.iq.usp.br/schisto/), and S. mansoni entries in GenBank. From this collective data, we produced a reference cDNA/EST database to aid gene identification and annotation. Genes with significant differential expression between the SAGE libraries were additionally annotated by BLAST against the S. mansoni GeneDB (http://www.genedb.org/genedb/smansoni).

SAGE tags were scored for differential expression among the two libraries using the R-statistic (Stekel et al., 2000). Higher scores represent a greater deviation from the null hypothesis of equal frequencies, while scores close to zero represent near constitutive expression. Hierarchical clustering, as performed by Eisen’s CLUSTER package (Eisen et al., 1998), was additionally used to determine patterns of correlated expression among SAGE libraries. For these cluster analyses, SAGE data was log transformed and median centered for both individual tags and libraries. Clustering was performed using Pearson’s correlation (i.e. centered) as a distance metric and the average linkage algorithm.

2.4. Semi-quantitative RT-PCR validation of SAGE tag frequencies

Complementary DNA (cDNA) was synthesized from the ten different parasite life stags using 1μg of RNA and oligo (dT)₂₀ primer with 200 U of SuperScript II reverse transcriptase in 20 μl reactions according to manufacturer’s protocol (Invitrogen, Carlsbad, CA). Semi-quantitative PCR reactions were performed using, 1 μl of cDNA template, 1.25 mM dNTPs, 5 pmol/μl gene specific primers (Supplemental Table S1), and 1 U Taq DNA polymerase (Promega, Madison, WI) in 50 μl reactions. The number of cycles used for each semi-quantitative PCR reaction was individually optimized to fall within the exponential phase of template amplification by checking the product on 1.2 % agarose gels every cycle of the PCR (between cycles 10–38).

3. Results and discussion

3.1. SAGE library construction

Construction, sequencing, and preliminary analysis of ten S. mansoni SAGE libraries profiling the development and sexual differentiation of the juvenile to adult worm stages in the mammalian host, and the intramolluscan development of the miracidium to 20-day cultured sporocysts have been completed. More than 60,000 SAGE tags were generated from each library constructed from sub-adult liver parasites (SA_LIVER), adult male and female worms from both single sex (AM_SS and AF_SS) and mixed sex (AM_MS and AF_MS) infections, miracidia (MIRA), and sporocysts cultured for 6 days in media pre-conditioned by Bge cells (6D_S_COND) or in media without additions (6D_S_UN). In addition, significant numbers of SAGE tags also were generated from sporocysts cultured for 20 days in media pre-conditioned by Bge cells (20D_S_COND) or in media without additions (20D_S_UN) (Table 1). A total of 28,837 unique SAGE tag sequences were sampled (Table 2), many of which varied in frequency among the SAGE libraries. 28,837 unique tags potentially represents the total number of unique transcripts in S. mansoni. However, based on gene numbers obtained from fully sequenced genomes of other invertebrate metazoans such as Caenorhabditis elegans and Drosophila melanogastor, it is estimated that S. mansoni probably has from 15,000–18,000 genes (Verjovski-Almeida et al., 2003). Most likely, the discrepancy in estimated gene numbers represents artifacts in the production of SAGE tags, including those containing sequencing errors or those resulting from incomplete NlaIII digestion of cDNAs, producing multiple tags from a single transcript. These tags, when identified, are indicated as ‘alternative sense’ (AS) SAGE tags. SAGE tags mapped to a site proximal to the 3′ end of a putative transcript, either shown by the presence of a poly-A tail or a predicted poly-adenylation sequence, are designated ‘primary sense’ (PS) SAGE tags.

TABLE 1.

SAGE tag statistics for completed libraries

SAGE Library	Acronym	Total Tags Sampled	Putative Sequencing Error Tags	Total Tags Used in Analyses
Adult male – single sex infection	AM_SS	74,549	7,446	67,103
Adult female – single sex infection	AF_SS	74,427	6,304	68,123
Adult male –mixed sex infection	AM_MS	75,804	7,477	68,327
Adult female – mixed sex infection	AF_MS	75,456	7,676	67,780
Sub-adult liver stage	SA_LIVER	78,076	6,861	71,215
Miracidia	MIRA	76,098	7,648	68,450
6d sporocysts (un-cond.)	6D_S_UN	75,659	7,615	68,044
6d sporocysts (cond.)	6D_S_CON	68,603	8,432	60,171
20d sporocysts (un-cond.)	20D_S_UN	34,574	3,890	30,684
20d sporocysts (cond.)	20D_S_CON	59,865	7,199	52,666

TABLE 2.

Ability to assign SAGE tags to putative genes and transcripts. The genomic data represent the v2.0 assembly contigs while the transcript data represent our database of cDNA and EST sequences (see Methods). We did not attempt to predict open reading frames (ORFs) for the putative genome sequence contigs, but instead used them as a reference for the uniqueness of SAGE tag sequences within the genome.

Statistic	Genomic	Transcript
Total Number of Unique Tag Sequences	28,837
Tags that Do Not Map to the Contigs	11.91%	35.66%
Tags that Map to One Location in the Contigs	78.23%	55.99%
Tags that Map to Multiple Locations in the Contigs	9.85%	8.35%
Tags Mapped to an Open Reading Frame	0.00%	32.10%
Tags Not Mapped to an Open Reading Frame (UK)	100.00%	67.90%

Mapping the SAGE tags to the cDNA/EST database can be used to identify the gene of origin of the tag. When SAGE tags are mapped to the assembled S. mansoni cDNA/EST (i.e., transcript) database, ~56% of the tags map to a unique transcript contig, ~36% did not map to any transcript contig, and ~8% mapped to multiple transcript contigs (Table 2). Overall, ~32% of the SAGE tags mapped to an open reading frame, while 68% either mapped to transcript contigs with no obvious open reading frame or to no transcript contig at all (Table 2). The fraction of SAGE tags that do not map to any transcript contig may represent tag sequencing errors or, perhaps more likely, considering the results relative to the draft genome sequence (see below), may be produced from transcripts for which the region of the transcript producing the SAGE was not captured in an EST. The vast majority of transcript sequences in the S. mansoni transcript database are ESTs that were produced using the ORESTES method (Verjovski-Almeida et al., 2003), which generates partial open reading frame sequences. Full-length ESTs and/or complete annotation of the S. mansoni genomic sequence will be required to identify the origin of the majority of the SAGE tags. However, the usefulness of the SAGE tag remains, whether its gene of origin is known or not. Analysis of the presence and abundance of SAGE tag can be made without knowledge of the origin and SAGE tags having a particular expression pattern can be used to identify (and then clone) genes displaying a particular expression pattern (see below).

Approximately 78% of SAGE tags mapped to a unique site in the S. mansoni genomic sequence (assembly v2.0, September, 2005), while 12% did not map to the genome (Table 2). The latter group of SAGE tags either originated from an unsequenced region of the genome or represent SAGE-tag sequencing artifacts. Approximately 10% of the SAGE tags map to multiple locations in the genome, either originating from identical sequences in gene families or in unrelated genes or represent SAGE sequencing artifacts (Table 2). Linkage of SAGE tags to predicted or known open reading frames will be accomplished with complete annotation of the S. mansoni genome sequence and in the immediate term will aid in the annotation of the S. mansoni genome. In addition, a completed genome sequence with accompanying gene/transcript models will greatly aid in evaluation of alternate sense tags (incomplete NlaIII digestion) and the presence and abundance of antisense transcripts, which appear to be quite prevalent and may be functionally significant in Plasmodium and Toxoplasma (Panankar et al., 2001; Radke et al., 2005) and have been shown to occur in Schistosoma (Smyth et al., 2003).

3.2. Ontology of transcripts identified by SAGE tags

The assignment of all open reading frames in transcripts from PS SAGE tags (R>0), differentially expressed tags (assigned from PS tags R>4) in mammalian and snail stages into the three different GO categories (molecular function, biological process, cellular localization) was performed using Goblet program (http://goblet.molgen.mpg.de; Groth et al., 2004). Of the 2500 total sequences (R>0), 57 % are assigned a molecular function (GO: 0003674), 44 % assigned to a biological process (GO: 0008150), and 30 % assigned to a cellular component (GO: 0005575). The majority of proteins receive a functional assignment as binding activity (70%), physiological process (85%), and to the cellular compartment (74%) (Fig. S2). Approximately 7 % and 4 % of the total proteins are found to be differentially expressed in mammalian and snail stages, respectively. Gene ontology analysis of these differentially expressed proteins showed a three fold increase in proteins involved in signal transduction and cell communication, a four fold increase in proteins involved in development, in mammalian stages compared to snail stages (Fig. S2). On the other hand, proteins involved in the regulation of physiological process and complex formation exhibited increases of 12 fold and 3 fold, respectively, in the snail stages compared to mammalian stages (Fig. S2).

3.3. Expression analysis

Similar to results from microarrays, SAGE tags can be clustered based on their abundance in tag libraries constructed from different tissues, treatments, or developmental stages and displayed as a heatmap. Hierarchical clustering of 502 differentially expressed (R>4) PS tags is represented in Figure 1, and clearly illustrates global changes in gene expression levels as the parasite undergoes maturation and development throughout its lifecycle. A variety of stage-associated differentially expressed genes were identified by Blast sequence analyses and examples are shown in Tables 3 and 4. However, the data also show that approximately ~59% of mammalian stage PS tags and ~47% of snail stage PS tags represent unknown or hypothetical gene sequences, which underscores the fact that we still lack a significant amount of information on a large proportion of differentially expressed genes potentially involved in worm development.

Fig. 1.

Hierarchical clustering of 502, differentially expressed (R > 4) PS tags from all stages profiled was done using Cluster 3.00 and Treeview 1.0.13 programs (Stanford MicroArray Database). Red indicates up regulation and green indicates down regulation relative to the median abundance (black) for each tag. Abbreviations used for stage SAGE libraries are sub-adult liver, SA_Liver; adult female and male from single-sexed infections, AF_SS and AM_SS; adult male and female from mixed sex infections, AM_MS and AF_MS; miracidia, MIRA; 6-day mother sporocysts without or with Bge-conditioned media, 6D_S_UN and 6D_S_COND; and 20-day sporocyst without or with Bge-conditioned media 20D_S_UN and 20D_S_COND.

TABLE 3.

Differentially-expressed transcripts (log-likelihood statistic R > 4) between adult male/single-sexed (AM_SS), adult female/single-sexed (AF_SS), and subadult/liver (SA_Liver) Schistosoma mansoni SAGE libraries; PS SAGE tags only; see Fig. 1.

AM_SS	AF_SS	SA_Liver	# Genes	Examples
↑			10	GST-28 kDa, Prx1, CytC ox; actin; fatty acid binding protein
	↑		11	Ubq-CytC Red, Fructose aldolase, purine nucleoside phosphorylase
		↑	30	Immunophilin, cyclophilin, ferritin, HSP 70, lamin B receptor
↑	↑		15	Saposin type B, serpin, β-xylosidase
	↑	↑	6	RNA Pol B transcription factor, leucine rich protein
↑		↑	6	Receptor activated PKC

TABLE 4.

Differentially-expressed transcripts (log-likelihood statistic R > 4) between miracidial (MIRA), 6-day cultured mother sporocysts (6D_S_UN), and 20-day cultured sporocysts (20D_S_UN); see Fig. 1.

MIRA	6D_S_UN	20D_S_UN	# Genes	Examples
↑			27	P40 egg antigen, Ca-binding protein (SME16), histone H1, PEPCK, GPx, proteorhodopsin, CP391S (S. japonicum egg protein)
	↑		4	RNA polB, TF, 14-3-3 epsilon2
		↑	2	SM28 GST
	↑	↑	17	Stathmin-like protein (SPO-1), Sm26 GST, fatty acid-binding protein, calreticulin (Sm4), 4 ribosomal-related genes, S. japonicum egg protein CP391B
↑	↑		3	Trypsin-like protein

SAGE profiling will be useful to monitor specific aspects of parasite biology. For instance, development across different stages and in different host environments could be investigated by profiling Homeobox (Hox) genes, which are involved in controlling pattern during development (McGinnis and Krumlauf, 1992). Because their expression is tightly regulated and their transcripts are not abundant, Hox transcripts would make good tests for the sensitivity and specificity of our SAGE analysis. Several S. mansoni Hox genes have been identified, including orthologues of Drosophila labial (SmHox1), deformed (SmHox4), and abdominal A (SmHox8) (Pierce et al., 2005). SAGE tags for these Hox genes as well as the related nuclear receptor co-repressor S. mansoni Fushi Tarazu-Factor 1 interacting protein (Oger et al., 2006) were identified and their expression pattern analyzed. While the expression of Fushi Tarazu-Factor 1 interacting protein was seen in all developmental stages analyzed, as was suggested previously (Oger et al., 2006), it was found to vary from a low of 0.0044% in AF_SS (3 tags in 68,123 total tags) to 0.01519% in 20D_S_COND (8 tags out of 52,666 total tags), a difference of 3.5 fold (Fig. 2). Hox genes showed varying levels of low and stage-specific expression (Fig. 2). For example, tags for abdominal A (SmHox8) were only identified in SA_Liver, AF_SS, and AM_MS worms, while deformed (SmHox4) displayed a similar expression pattern in mammalian stages and it was also present in 6-day cultured sporocysts. For comparison, the SAGE-tag frequency determined for two highly expressed genes, beta-tubulin and GAPDH, were found to be expressed at more than 10 and nearly 100 times higher, respectively, than the Hox genes.

Fig. 2.

SAGE profiling of Hox genes during S. mansoni lifecycle. A. SAGE tags from all 10 profiled stages for S. mansoni Fushi Tarazu-Factor 1 interacting protein (SmFIP-1), labial (SmHox1), deformed (SmHox4), and abdominal A (SmHox8) are shown. In this figure, and in Figures 4 and 5, for each gene examined the stages are shown in the following order: 1–5, mammalian host stages (closed bars), 1: sub-adult liver; 2: adult males from single-sex infections; 3: adult females from single-sex infections; 4: adult males from mixed sex infections; 5 adult females from mixed sex infections; 6: miracidia (open bar); 7–10 invertebrate host stages (gray bars), 7: 6-day mother sporocysts without Bge-conditioned media; 8: 6-day mother sporocysts with Bge-conditioned media; 9: 20-day sporocysts without Bge-conditioned media; and 10: 20-day sporocysts with Bge-conditioned media. B. SAGE tags for two constitutive, highly expressed genes, beta-tubulin and glyceraldehyde 3-phosphate dehydrogenase (GAPDH), are shown. Tag libraries as in A.

In order to verify that SAGE tag frequencies represent true variations in transcript abundance, several SAGE tags showing differential expression were further investigated by semi-quantitative reverse transcription PCR (RT-PCR). In all tags examined, the relative expression levels indicated by the SAGE results were reflected in the RT-PCR (Fig. 3). This relationship was consistent for low abundance tags, 0-~20 tags per library, to highly abundant tags, 100 to >400 tags per library. For example, SAGE results for a putative cationic amino acid transporter (SAGE tag 13441) indicate very low expression in levels in AM__SS and AM_MS stages and at significantly higher levels in AF_SS, but to be completely absent from AF_MS and sub-adult liver parasites; the same pattern was seen in semi-quantitative RT-PCR analysis.

Fig. 3.

Semi-quantitative reverse-transcription – PCR analysis of SAGE tags. Identity of the SAGE tags and the number of cycles used in the PCR are: 2, GAPDH, 21 cycles; 13441, high affinity cationic amino acid transporter, 36 cycles; 15463, purine nucleoside phosphorylase, 35 cycles; 25004, lamin B receptor, 36 cycles; 10596, histone H3.3, 33 cycles; 4, fatty acid binding protein, 21 cycles; 10, 28 kDa glutathione S-transferase, 23 cycles; 1885, similar to nuclear ribonucleoprotein K, 36 cycles; and 684, similar to nucleoside diphosphate kinase, 30 cycles.

Known, female-specific transcripts were clearly identified by SAGE. For example, tyrosinase 1 and 2 (Fitzpatrick et al., 2006) and FS800 (Reis et al., 1989) were preferentially expressed in females from mixed sex infections (Table 5). The low number of SAGE tags in male SAGE libraries most likely represents the presence of small amounts of contaminating female worm fragments in the male worms. The SAGE tags for these genes were almost entirely absent in males and females from single-sex infections. However, the very low number of SAGE tags found in these stage libraries may support the hypothesis others have suggested that worms developing in single-sex infections do not have a committed transcriptome, instead one that may represent the primordial hermaphroditic state of trematodes (Fitzpatrick and Hoffmann, 2006). Of note, a novel tetraspanin was identified that was found to have nearly 20 times higher expression in AF_MS than other stages. The expression of this transcript is strongly induced by worm pairing, suggesting that it plays a role in male-female interactions. Tetraspanins are surface proteins with four membrane-spanning domains that form molecular complexes with each other and other surface proteins and function in the maintenance of cell membrane integrity and in cell-cell interactions (Boucheix and Rubinstein, 2001). Tetraspanins in S. mansoni have recently been shown to be promising vaccine candidates (Tran et al., 2006).

Table 5.

Female-specific SAGE tags. Group I, transcripts encode genes encoding structural proteins of the eggshell; Group II, transcripts encode enzymes involved in formation of the egg shell; Group III, female-specific transcripts encoding proteins not directly involved in eggshell formation; and Group IV, partial list of female-specific SAGE tags with unknown function. Abbreviations are: subadult/liver worms (SA_Liver); adult male worms/single-sex infection (AM_SS); adult female worms/single-sex infection (AF_SS); adult male worms/mixed sex infection (AM_MS); and adult female worms/mixed sex infection (AF_MS).

Tag ID	SA LIVER	AM SS	AF SS	AM MS	AF MS	Identification	R-value
5962	1	6	3	84	1022	Eggshell protein precursor p14	627	I
5116	1	5	0	34	684	Eggshell protein precursor p19	434
10392	0	1	0	19	327	Eggshell protein precursor p34	209
8450	0	4	0	12	215	Eggshell protein precursor p48	133
9799	0	2	0	12	208	Eggshell protein precursor	131
10548	0	0	0	0	92	Eggshell protein precursor	65
10396	0	0	0	4	86	Eggshell protein precursor	56
10409	0	0	0	4	79	Eggshell protein precursor	52
21221	0	0	0	0	67	Eggshell protein precursor	47
10423	0	0	0	4	52	Tyrosinase 1	33	II
11230	0	0	0	0	14	Tyrosinase 2	10
1610	2	6	1	5	108	Tetraspanin	60	III
8987	0	1	0	5	62	Pro-His-rich protein	38
21936	0	0	1	4	20	Female-specific protein 800	11
10435	0	0	0	8	162	SJCHGC09430 protein	106	IV
8056	0	1	0	2	94	SJCHGC06578 protein	62
10397	0	0	0	1	53	SJCHGC05655 protein	36
11223	0	0	0	2	52	SJCHGC05655 protein	34
10688	0	0	0	5	51	SJCHGC09200 protein	32
10611	0	0	0	5	49	SJCHGC09200 protein	31
21702	0	0	0	2	47	SJCHGC05655 protein	31
10545	0	0	0	3	43	SJCHGC01837 protein	28
22694	0	0	1	4	34	SJCHGC02864 protein	20
10687	0	0	0	0	32	SJCHGC06422 protein	23

Egg formation in S. mansoni involves the action of two main cell types; the fertilized ovum and 25–30 vitelline cells combine in the ootype of the female worm to produce the egg (Kunz, 2001; Hoffmann, 2004). Vitelline cells are produced in the vitellaria of females and provide components for synthesis of the eggshell, which consists of a number of eggshell precursor proteins that are cross-linked by tyrosinase to form a melanized, proteinacious network (Ebersberger et al., 2005). SAGE tags from transcripts encoding eggshell precursors are found to be both highly and specifically expressed in AF_MS worms (Table 5). The expression of the eggshell precursors and associated proteins accounts for ~4% of the total transcripts in AF_MS worms. The eggshell cross-linking enzymes tyrosinase1 and 2 were also found to be AF_MS-specific transcripts (Table 5).

For molluscan stages of S. mansoni, marked differential expression of transcripts between the freeliving miracidium and parasitic mother sporocyst stage also have been noted, especially for several “egg proteins”. For example, transcripts for SME 16 (calcium binding protein), ESP3-6 (S. mansoni egg secreted protein 3–6), and CP391S (S. japonicum egg protein) are exclusively produced by miracidia, whereas S. japonicum egg protein CP391B transcripts are highly expressed only in the sporocyst stages. One assumes that such tightly-regulated, stage-specific expression at high steady-state levels is associated with larval functions that are critical to normal growth and development. However, slow progress in the development of tools for manipulating specific gene expression in intact worms has impeded research efforts aimed at addressing important functional genomics questions. Currently RNA interference (RNAi) methods being applied to S. mansoni mammalian (Skelly et al., 2003; Correnti et al., 2005; Sayed et al., 2006) and snail (Boyle et al., 2003; Dinguirard and Yoshino, 2006) stages represent the most promising approach to date for gene manipulation in these parasites, although advances in schistosome transgenic technology recently have been reported (Grevelding, 2006; Kines et al., 2006; Brindley and Pearce, 2007) showing potential for application to functional gene studies.

One of the main uses of SAGE will undoubtedly be to assign potential function to genes encoding proteins of unknown function and to discover novel genes. For example, Table 5 has a partial list of SAGE tags showing highly AF_MS-specific expression, with most of these genes being completely silent in AF_SS, male and juvenile worms. Many of these SAGE tags originate from transcripts with moderate to high levels of expression, but their sequences and/or protein products have not been previously identified. Several of these SAGE tags mapped to ESTs that had significant similarity to predicted proteins of unknown function in S. japonicum. Although the precise role of proteins encoded in these transcripts remains unknown, we now have a list of genes likely to play key roles in female schistosome biology that can be investigated in situ hybridization, RNA silencing, and two-hybrid technologies in future studies.

SAGE data will provide clues as to how different stages accomplish routine biochemical functions across different stages and in different host environments. For example, profiling lactate dehydrogenase (Fig. 4) supports the idea that parasites in the mammalian host use lactic acid fermentation more extensively than do parasites in their invertebrate host; lactate dehydrogenase tags are significantly higher in mammalian lifecycle stages than in snail stages. Other glycolytic enzymes profiled show a generally constitutive expression pattern. However, role of oxygen in adult metabolism is less well defined. While it has been reported that no increase in lactate production is seen in worms incubated under anaerobic conditions (Bueding, 1950), other studies indicate that oxygen may be essential, especially if worms are to maintain high egg production levels, and that adult worms possess a functional tricarboxcylic acid cycle and respiratory chain (Coles, 1972). Profiling SAGE tags for the tricarboxcylic acid enzyme, succinate dehydrogenase, and the terminal oxidase beta subunit clearly indicate that the expression of genes encoding enzymes of oxidative respiration is equal or higher in mammalian-stage parasites than in snail-stage parasites (Fig. 4) suggesting that aerobic energy production is significant in adult worms in the mammalian host.

Fig. 4.

SAGE profiling energy metabolism during S. mansoni development. Stage-specific SAGE-tag libraries are shown as indicated in Figure 2. GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

Redox biology is an important biochemical process in all aerobic organisms. Schistosomes are thought to be especially sensitive to oxidative stress and antioxidant enzymes have been shown to play important roles in the host-parasite interaction (LoVerde, 1998; Sayed et al., 2006). Figure 5 illustrates the tag expression patterns for a number of antioxidant enzymes. Although we cannot comment on the expression of antioxidants during the entire developmental cycle, a number of interesting points can be made. Peroxiredoxins appear to be the main hydrogen peroxide-reducing activity in the parasite (Sayed et al., 2006). Cytoplasmic peroxiredoxin 2 and mitochondrial peroxiredoxin 3 appear to be constitutively expressed in the stages profiled. However, the most abundant form in adult worms, peroxiredoxin 1, has a highly specific expression, being nearly absent in liver worms and sporocysts. In addition, although peroxiredoxin 1 expression appears to be negligible in 6-day cultured sporocysts by SAGE analysis, recent quantitative PCR studies have shown a temporal expression pattern for sporocyst peroxiredoxin 1 and 2, in which larval steady-state transcript levels increase at 2–3 days in culture, after which time transcript levels begin to fall off (Vermeire et al., in revision). Moreover, sporocyst peroxiredoxin gene expression is stimulated by exogenous H₂O₂ treatment, suggesting that gene expression may be regulated by H₂O₂-responsive elements. Also of note, although peroxiredoxin 1 has been shown to be present in schistosome eggs (Williams et al., 2001) no peroxiredoxin 1 tags were present in the miracidial library. This suggests that the expression of peroxiredoxin 1, and perhaps other secreted egg proteins, occurs in extra-miracidial tissues within the egg (Williams et al., 2001) and conforms to the pattern of peroxiredoxin expression recently seen in S. japonicum (Kumagai et al., 2006). Glutathione peroxidase is another hydrogen peroxide-reducing enzyme in the parasite. Although both glutathione peroxidase 1 and 2 appear to be generally constitutively expressed, sexually mature females and miracidia have significantly elevated levels of glutathione peroxidase 1. Sexually mature females also have very high levels of the secretory form of superoxide dismutase; the bar for adult female worms from mixed sex infections does not accurately indicate the SAGE tag percent, which is 0.71%, or 20 times the level found in adult male worms from mixed sex infections. Although egg-producing females may be under significant oxidative stress due to egg production, not all redox enzymes are elevated in these worms.

Fig. 5.

SAGE profiling redox enzymes during S. mansoni development. Stage-specific SAGE-tag libraries are shown as indicated in Figure 2. Abbreviations used: Prx, peroxiredoxin; GPx, glutathione peroxidase; SOD, superoxide dismutase; Sec-SOD, secretory SOD; GST, glutathione S-transferase. The bar for adult female worms from mixed sex infections does not accurately indicate the SAGE tag percent, which is 0.71%, or 20 times the level found in adult male worms from mixed sex infections.

Another, general pattern that emerges from profiling antioxidant enzymes is the unusual finding that sets of genes appear to be coordinately expressed in the mammalian compared to molluscan stages of the lifecycle. This expression pattern is shown in broad strokes in Figure 1, with significant numbers of genes following a sporocyst-specific expression pattern or a mammalian stage-specific expression pattern. In Figure 5, the expression of the Cu/Zn-superoxide dismutase and 26 kDa glutathione S-transferase are similar and increased during the early larval stages of development, while Mn-superoxide dismutase and omega-glutathione S-transferase are coordinated and increased during the juvenile and adult stages of the lifecycle. We also have evidence for a fourth, snail-stage specific peroxiredoxin, whose expression essentially mirrors that of peroxiredoxin 1 (unpublished). Whether enzymes are selectively expressed because of differential activity in the homeothermic vs. the poikilothermic host remains to be determined.

4. Summary and future prospects

In this report we present a preliminary analysis of SAGE profiling the S. mansoni lifecycle. We have shown that SAGE is sensitive and that tag frequency accurately reflects the expression pattern of a variety of genes. We are able to monitor the expression of genes involved in development and have identified of a number of novel transcripts differentially expressed in sexually mature female parasites including a novel tetraspanin. We present results showing the utility of SAGE to profile house-keeping functions and biochemical pathways of interest. However, a significant number of goals remain to be accomplished. Many SAGE tags identify transcripts with unknown function. These may be truly new, previously uncharacterized transcripts/proteins or they are transcripts with short or no ORFs. Many tags did not map to any transcript, those identified as UK tags. This analysis only included PS tags, a fraction of the total number of tags. Assignment of identity of the UK tags will require coordination with genome annotation and prediction of all ORFs in the S. mansoni genome. We have begun to construct tag libraries to profile the transition from cercariae to lung stage parasites as well as the egg. Completion and analysis of these libraries will allow for rapid profiling of gene expression during the entire lifecycle of S. mansoni.

Supplementary Material

01

Supplemental Fig. S1. Illustration of SAGE protocol using NlaIII as the anchoring enzyme and MmeI as the tagging enzyme (LongSAGE). In this diagram, MmeI cleaves downstream from its recognition site (TCCRAC), creating a 21 bp tag that includes the NlaIII recognition site (CATG) and 17 bp of 3′-adjacent sequence after removal of the linker by NlaIII digestion.

02

Supplemental Fig. S2. Gene ontology (GO) assignment. Gene ontology of all proteins (assigned from PS SAGE tags R > 0) [closed bars], differentially expressed mammalian-stage proteins (assigned from PS tags R > 4) [gray bars], and differentially expressed snail-stage proteins (assigned from PS tags R > 4) [open bars] are plotted against the percentage of the protein distribution in each of the GO categories.

Abbreviations

SAGE

serial analysis of gene expression

Bge

Biomphalaria glabrata embryonic cells

SA_LIVER

sub-adult liver parasites

AM_SS

adult male from single sex infections

AF_SS

adult female worms from single sex infections

AM_MS

adult male worms from mixed sex infections

AF_MS

adult female worms from mixed sex infections

MIRA

miracidia

6D_S_COND

sporocysts cultured for 6 days in media pre-conditioned by Bge cells

6D_S_UN

sporocysts cultured for 6 days in media without additions

20D_S_COND

sporocysts cultured for 20 days in media pre-conditioned by Bge cells

20D_S_UN

sporocysts cultured for 20 days in media without additions

PS SAGE tags

primary sense SAGE tags

Hox

homeobox genes

RT-PCR

semi-quantitative reverse transcription PCR