INTRODUCTION
Protozoan parasites of the genus Sarcocystis (Apicomplexa, Sarcocystidae) have a two-host life cycle with mainly herbivores and omnivores as intermediate hosts and carnivores as definitive hosts. Cervid species like red deer, reindeer, roe deer and moose may each act as intermediate hosts for several Sarcocystis species, as revealed by differences in sarcocyst morphology and/or DNA sequences (Dahlgren, Reference Dahlgren2010). For many years, molecular delimitation of Sarcocystis species has been based almost solely on nucleotide sequences from the nuclear ribosomal DNA unit, particularly the 18S or small subunit (ssu) ribosomal RNA (rRNA) gene, and to a lesser extent the 28S rRNA gene and the internal transcribed spacer 1 (ITS1) region (see Gjerde, Reference Gjerde2013b ). Recently, however, the mitochondrial protein-coding cytochrome c oxidase subunit I gene (cox1) was established as a new marker for delimitation of Sarcocystis species in cervids, cattle and sheep (Gjerde, Reference Gjerde2013b ), but this gene may also be useful for differentiating between Sarcocystis species in other hosts.
In a previous study (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ), free-ranging red deer (Cervus elaphus) from western Norway were found to host five different Sarcocystis species as determined from their sarcocyst morphology and near complete ssu rRNA gene sequences, i.e. Sarcocystis hjorti, Sarcocystis hardangeri, Sarcocystis ovalis, Sarcocystis rangiferi and Sarcocystis tarandi. There were some doubts, however, concerning the assignment of the latter two taxa to two species previously characterized from reindeer. Thus, isolates/clones of each of these sarcocyst types from both hosts showed a moderate sequence variation (0–1·1%) at the ssu rRNA gene, and the within-host and between-host sequence variation for each sarcocyst type overlapped. Moreover, both in the original (see Fig. 10 in Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ) and in subsequent (Gjerde, Reference Gjerde2012, Reference Gjerde2013b ) phylogenetic analyses, ssu rRNA gene sequences of S. tarandi from reindeer were interspersed with sequences of the similar species in red deer, suggesting that a single species of this type infected both hosts. All sequences of S. rangiferi from reindeer, on the other hand, clustered in a monophyletic group, whereas those obtained from the similar species in red deer formed two other clusters together with sequences from an unnamed Sarcocystis sp. in sika deer, indicating the presence of at least two separate species infecting reindeer and red deer/sika deer, respectively (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ; Gjerde, Reference Gjerde2012, Reference Gjerde2013b ). In an attempt to resolve this question, the ITS1 region was also examined in the original study, but these sequences showed an even higher degree of within-host variation (0–12·8%), and clustered in a similar manner to the ssu rRNA gene sequences (see Fig. 11 in Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ). Moreover, partial sequences of the 28S rRNA gene were also generated from a few isolates of each species from both hosts, but these sequences also displayed some intraspecific variation, as well as a high similarity between the S. tarandi-type isolates and a slight difference between the S. rangiferi-type isolates from the two hosts (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ; unpublished data).
However, when ∼1 kb long sequences of the mitochondrial cox1 were obtained and used in phylogenetic analyses (Gjerde, Reference Gjerde2013b ), the seven characterized isolates of the S. rangiferi-like species in red deer and the six isolates of S. rangiferi from reindeer formed two separate monophyletic clusters, as did the seven isolates of the S. tarandi-like species in red deer and the eight isolates of S. tarandi from reindeer, respectively. Moreover, there was less sequence divergence between the isolates of each type from the same host, than between isolates from different hosts, particularly for the S. rangiferi-type. Each of these two species in red deer therefore seemed to be different from the two similar species in reindeer, and were provisionally referred to as Sarcocystis cf. rangiferi and Sarcocystis cf. tarandi in that paper (Gjerde, Reference Gjerde2013b ), pending an even more thorough comparison using more isolates of each taxon. The principal aim of the present study was therefore to obtain partial cox1 sequences of additional isolates of each of these taxa from both red deer and reindeer, and use the new sequences together with the previously generated cox1 sequences to establish the boundaries of these species through sequence comparisons and phylogenetic analyses. Provided that the results showed that the two species in red deer were distinct from those in reindeer, a second aim was to redefine these species by using previous and new data, and collate the complete species descriptions in a single paper for future reference. The cox1 sequences previously obtained from a limited number of isolates of each Sarcocystis species (Gjerde, Reference Gjerde2013b ), seemed, on the other hand, to confirm that the three other species reported from red deer (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ) were indeed conspecific with S. hardangeri in reindeer and with S. hjorti and S. ovalis in moose, respectively. Nevertheless, the present study also aimed at corroborating these findings by comparing more isolates of these taxa at cox1 in order to make an updated review of all known Sarcocystis species in red deer.
MATERIALS AND METHODS
Host origin and isolation of sarcocysts
The host origin, identity and number of sarcocyst isolates of different types examined in the present study are given in Table 1. The main focus was on cysts of the S. rangiferi- and S. tarandi-types from red deer and reindeer and cysts of the S. hjorti- and S. ovalis-types from red deer and moose. However, for comparative purposes some cysts of Sarcocystis alces from moose and of Sarcocystis capreolicanis, Sarcocystis gracilis and Sarcocystis silva from roe deer were also included. All sarcocysts were isolated during 2005–2010 from the oesophagus, diaphragm and/or heart of semi-domesticated reindeer (Rangifer tarandus) in northern Norway (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2007), free-ranging moose (Alces alces) in south-eastern Norway (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2008, Reference Dahlgren and Gjerde2010b ; Gjerde and Dahlgren, Reference Gjerde and Dahlgren2010), free-ranging roe deer (Capreolus capreolus) in south-eastern Norway (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2009; Gjerde, Reference Gjerde2012), and free-ranging red deer (C. elaphus) in western Norway (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ). The red deer were adult animals from Rogaland and Hordaland counties in western Norway, which had been killed in September–October, 2006–2008 during the annual hunting season as described previously (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ). The reindeer were semi-domesticated adult animals that had been slaughtered in 2005 and 2006 at abattoirs in Nord-Trøndelag, Nordland, Troms, and Finnmark counties in northern Norway as described for some of this material previously (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2007, Reference Dahlgren and Gjerde2010a ).
Overview of Sarcocystis isolates (sarcocysts) from red deer (Ce), reindeer (Rt), roe deer (Cc) and moose (Aa) from which partial cox1 sequences were obtained in the present study and their associated GenBank accession numbers. The ratio between the number of isolates examined of each species and the number of haplotypes (Ha) found is also given, both for sequences obtained in the present (A) and a previous study (B) a , as well as for both studies combined (A+B)
a Data from Table 2 in Gjerde (Reference Gjerde2013b ), in which S. elongata was designated S. cf. tarandi and S. truncata was designated S. cf. rangiferi.
b–e Reverse primers used together with forward primer SF1 to amplify the various isolates: bSR8D, cSR9, dSR10, eCOIRm. Some isolates were amplified and sequenced twice with different reverse primers as indicated by two superscript letters; only the longest sequence in such pairs was submitted to GenBank.
f Haplotypes only from moose in Norway; a different haplotype from a Canadian moose is not included.
g Two additional haplotypes have been found for S. silva in moose (Gjerde, Reference Gjerde2013b ).
h The haplotype of S. tarandi in red deer was identical to a haplotype in reindeer.
All muscle samples were frozen shortly after collection and/or arrival at the laboratory and stored at −20 °C for up to 2 years before examination. Samples from different cervid hosts were kept and processed separately. Upon thawing, the samples were examined under a stereo microscope and individual sarcocysts were excised and identified to cyst type/species under a light microscope as described previously (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2007; Gjerde, Reference Gjerde2013b ). After cyst isolation and preliminary species identification, individual sarcocysts were placed in 20 μL of distilled water in labelled 1·5 mL micro-centrifuge tubes and kept frozen at −20 °C until DNA isolation weeks to years later.
DNA isolation, PCR and sequencing
The majority of the sarcocyst isolates used in this study had been kept frozen for 5–7 years before genomic DNA was extracted in January–April 2013 from individual sarcocysts using QIAmp DNA Mini Kit (Qiagen, Germany) according to the manufacturer's tissue protocol (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2007). The remaining DNA samples had been extracted from sarcocysts in the same manner in 2005–2008 in connection with the characterization of the ssu rRNA gene as reported previously (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2007, 2008, 2009, 2010a).
A fragment of cox1 was amplified using forward primer SF1 and one of the following reverse primers: SR8D (1029 bp long sequence, without primers), SR9 (1038 bp), SR10 (1092 bp) or COIRm (1095 bp), depending on species and isolate as detailed in Table 1. For some species and isolates reverse primers SR4, CoxS1R and SR5 were also attempted initially. Primer sequences have been published before (Gjerde, Reference Gjerde2013b ), except those of the two newly designed reverse primers SR9 (5′-atatccataccrccattgcccat-3′) and SR10 (5′-aacgacagcacaaaatggaagtg-3′). Each PCR reaction mixture contained 2·5 μL of the DNA solution, 18·75 μL HotStarTaq Master Mix (Qiagen GmbH, Germany), 10 pmol of each primer, 6 μg of bovine serum albumin, and RNase-free water to make a final volume of 37·5 μL. A negative control was included in each PCR run. PCR reactions were carried out in a Bio-Rad T100 thermal cycler (Bio-Rad Laboratories Inc., USA). Cycling conditions were: initial Hot Start at 95 °C for 15 min, followed by 45 cycles of denaturation at 94 °C for 30 s, annealing at 52–53 °C for 30 s (temperature depending on primer pair), extension at 72 °C for 90 s; and final extension at 72 °C for 10 min.
PCR products were evaluated, purified and sequenced and the resulting sequences were assembled as previously described (Gjerde, Reference Gjerde2013b ). All newly obtained cox1 sequences of different species and isolates were compared pair-wise with each other and with those previously obtained from cox1 of Toxoplasma gondii, Neospora caninum, Hammondia heydorni and Hammondia triffittae (Gjerde, Reference Gjerde2013a ) and various Sarcocystis spp. (Gjerde, Reference Gjerde2013b ) using the programme Nucleotide BLAST (Basic Local Alignment Search Tool) of the National Center for Biotechnology Information (NCBI) (http://blast.ncbi.nlm.nih.gov/). The software package DnaSP (DNA Sequence Polymorphism) version 5.10.01 (Librado and Rozas, Reference Librado and Rozas2009) was used for the analysis of nucleotide polymorphisms (sequence variation) among the cox1 sequences, but these comparisons were confined to nucleotides 1–1020, which were present in all sequences. The analyses included enumeration of variable (polymorphic) sites and haplotypes among the sequences, and estimation of different measures of sequence divergence within and between species.
Phylogenetic analyses
Phylogenetic analyses were conducted on cox1 sequences by means of the MEGA5 (v5.10) software (Tamura et al. Reference Tamura, Peterson, Peterson, Stecher, Nei and Kumar2011). A total of 308 nucleotide sequences from 25 species were included in the analyses. They comprised the 144 new sequences from 10 Sarcocystis species generated in the present study as listed in Table 1 with their GenBank accession numbers. Furthermore, they comprised the 155 sequences of 22 Sarcocystis spp. (GenBank accession numbers KC209578–KC209732) from the previous study (see Table 2 in Gjerde, Reference Gjerde2013b for details about these species/isolates), and sequences of the following six species: H. triffittae (JX473247–JX473249), H. heydorni (JX473250–JX473251), N. caninum (JX473252), T. gondii (JX473253) (see Gjerde, Reference Gjerde2013a ); Eimeria necatrix (HQ702482) and Eimeria tenella (HQ702484). The new partial cox1 nucleotide sequences of various Sarcocystis spp. could easily be aligned against each other and against the codon-based multiple sequence alignment generated previously (Gjerde, Reference Gjerde2013b ), which resulted in a multiple alignment with no gaps, comprising individual cox1 sequences of different Sarcocystis spp. varying in length between 1020–1095 nucleotides. Since all positions containing gaps and missing data were eliminated in the analyses, there were a total of 1020 positions in the final dataset. Phylogenetic trees were reconstructed by the neighbour-joining (NJ) method (Saitou and Nei, Reference Saitou and Nei1987) using the p-distance model and 10 000 bootstrap replications (Felsenstein, Reference Felsenstein1985). All codon positions were used. The two Eimeria spp. from chickens (family: Eimeriidae) were used as outgroup species to root the trees. Other tree-building methods (maximum likelihood, maximum parsimony) were also tested for comparison.
RESULTS
PCR amplification and sequence comparisons
Initially, attempts were made to PCR-amplify isolates of all species, except those of the S. ovalis-type, using primer pair SF1/SR8D. This primer pair was found to work very well with isolates of S. alces, S. capreolicanis, S. gracilis, S. hjorti, S. silva and the S. rangiferi-like species from red deer, and also fairly well with some isolates of S. rangiferi from reindeer and some isolates of S. cf. tarandi from red deer. However, this primer pair resulted in very poor amplification of isolates of S. tarandi, and these products could not be used for sequencing. Three other reverse primers, SR4, CoxS1R and SR5 were then attempted with isolates of S. tarandi, but with similar poor results. This prompted the design of a new reverse primer, SR9, which in combination with forward primer SF1 resulted in excellent amplification and high-quality sequences not only of S. tarandi, but also of S. rangiferi, S. cf. tarandi and S. cf. rangiferi. However, many isolates of the latter three species had already been satisfactorily amplified with SR8D, and were therefore not amplified again with primer SR9. As regards the isolates of S. ovalis-type from red deer and moose, some of them were initially satisfactorily amplified with primer pair SF1/COIRm, but a more specific reverse primer, SR10, was subsequently designed and used for some isolates with excellent results. The latter primer targeted nearly the same sequence region as primer COIRm, but was designed from a sequence of S. ovalis covering the complete 3′ end of cox1 (data not shown), and was therefore more specific than COIRm, which had primarily been designed for members of the Toxoplasmatinae (Gjerde, Reference Gjerde2013a ).
In total, new partial cox1 sequences, 1029–1095 bp in length (primers not included) were obtained from a total of 144 sarcocyst isolates, which after sequence comparisons using BLAST, could be assigned to 10 different Sarcocystis species as listed in Table 1. These sequences were submitted to GenBank and have been issued accession numbers KF241309–KF241452. Sequence comparisons using BLAST and DnaSP revealed considerable differences between the species concerning intraspecific nucleotide diversity as partly reflected in the relative number of haplotypes, which has been summarized in Table 1, both for the new sequences, the previous sequences (Gjerde, Reference Gjerde2013b ), and for all available cox1 sequences of these species. The sequences of S. rangiferi and S. tarandi displayed more polymorphic sites (42 and 40 of 1020 sites, respectively) than those of other species (Sarcocystis tenella, 35 sites; S. cf. tarandi, 27 sites; S. silva, 20 sites; S. cf. rangiferi, 19 sites; the remaining species, 0–18 polymorphic sites), but these numbers are also influenced by the total number and host origin of the isolates examined of each species.
Altogether 14 sarcocyst isolates from red deer could be assigned to S. cf. rangiferi, including three isolates from the oesophagus/diaphragm that had been tentatively identified microscopically as belonging to S. cf. tarandi, and one correctly identified isolate from cardiac muscle. Similarly, a total of 27 sarcocyst isolates from reindeer could be assigned to S. rangiferi based on cox1 sequences, including three isolates from the oesophagus/diaphragm and one isolate from cardiac muscle that had been identified microscopically as sarcocysts of S. tarandi. A comparison of the sequences of S. cf. rangiferi from red deer with those of S. rangiferi from reindeer, revealed a sequence identity of ∼93%, both for the new sets of sequences (14 vs 27), and for all available sequences (21 vs 33) of these taxa, and there were 58 fixed nucleotide differences (nucleotide sites where all sequences of one population differed from all sequences of the other population) between the sets comprising all sequences. Furthermore, the 21 sequences of S. cf. rangiferi from red deer shared ∼96% identity (30 fixed nucleotide differences) with the five available sequences of S. silva (from moose and roe deer), whereas the 33 sequences of S. rangiferi from reindeer shared ∼92% identity (62 fixed nucleotide differences) with those of S. silva. In contrast, the sequence identities within each of these three taxa were >99%.
A total of 27 sarcocyst isolates from reindeer and one isolate from red deer could be assigned to S. tarandi. The single isolate of this species from red deer (Ce34.14; GenBank number KF241438) was completely identical to one isolate of S. tarandi from reindeer and shared >99% identity with the other reindeer isolates of this species. The sarcocyst of this particular isolate from red deer had been subjected to DNA extraction concurrently with nine other sarcocysts of the S. tarandi-type from this host, but this isolate showed a much weaker band on agarose gel than the nine other isolates following simultaneous PCR-amplification with primer pair SF1/SR8D. The isolate was subsequently amplified and sequenced twice on two separate occasions using reverse primer SR9, and two identical sequences were obtained. Hence, this isolate from red deer has been included with the sequences of S. tarandi from reindeer in the sequence comparisons, which showed 98·7–100% sequence identity among all 36 isolates of this type examined in the present (28 isolates) and the previous (eight isolates) study.
A total of 18 sarcocyst isolates from red deer could be assigned to S. cf. tarandi based on cox1 sequences. When the seven previous isolates of this type were also included in the comparisons, the sequence identity among all 25 isolates was 98·5–100%. In contrast, the sequence identity between all available sequences of this type in red deer and all the above-mentioned sequences of S. tarandi was 96·9–97·6%, corresponding to differences between the two populations in 24–32 out of 1020 nucleotides (on average 27·6 nucleotides or 97·3% identity), including 14 fixed nucleotide differences. Most of the nucleotide differences were due to synonymous substitutions (silent mutations), and hence the two taxa differed at only two of 340 inferred amino acid residues.
The characterization of 13 new isolates of S. hjorti, 11 from red deer and two from moose, revealed some known and some new haplotypes compared with the nine isolates from the previous investigation (Gjerde, Reference Gjerde2013b ). None of the haplotypes in red deer were completely identical to any of the haplotypes in moose, but the differences between the sequences from either host were rather small, i.e. in 4–7 out of 1020 nucleotides (on average 5·3 nucleotides), and there was only one fixed nucleotide difference between the two populations.
The characterization of eight previously identified (by using the ssu rRNA gene) and four new isolates of the S. ovalis-type from red deer revealed that they all belonged to S. ovalis, whereas no isolates of the morphologically indistinguishable species S. hardangeri were found. Hence, only isolates of S. ovalis from moose were examined for comparison. The 12 newly examined isolates of S. ovalis from red deer belonged to the same two cox1 haplotypes as the five previously examined isolates from this host (Gjerde, Reference Gjerde2013b ). A single nucleotide substitution separated the most common haplotype (14 isolates) from the rarer haplotype, which were found in three isolates from the same host animal (isolates Ce8.1, Ce8.2, Ce8.3; GenBank numbers KC209652, KF241356–KF241357), suggesting that these three sarcocysts had developed from the same infection event. The 16 new isolates of S. ovalis from moose were identical (confirmed by amplifying and sequencing some isolates a second time using new primer solutions, as well as doing it simultaneously with the slightly different isolates from red deer), and also identical to six previous isolates from this host (Gjerde, Reference Gjerde2013b ). All 22 isolates from Norwegian moose differed at the same two or three nucleotide positions from the abovementioned 17 isolates of S. ovalis from red deer.
The characterization of additional isolates of S. alces from moose and of S. capreolicanis, S. gracilis and S. silva from roe deer, recovered new haplotypes of the latter two species (Table 1), but the new types shared >99% sequence identity with the previously known haplotypes of each species. The sequencing results confirmed that all of these isolates had been correctly assigned to species based on sarcocyst morphology and host origin.
Phylogenetic analyses
In the phylogenetic analysis using the NJ method, all taxa (sequence types) represented by two or more sequences formed monophyletic clusters, which have been collapsed in the NJ tree shown in Fig. 1. Such monophyletic clusters were also obtained when other tree-building methods (maximum likelihood, maximum parsimony) were used with the appropriate settings (data not shown). Of particular interest for the objective of this study was the fact that there was maximum support for the placement of all 25 S. tarandi-like isolates from red deer (=Sarcocystis elongata; see next section) as a sister taxon to all 35 isolates of S. tarandi from reindeer and the single isolate of this species from red deer. Likewise, there was maximum support for the placement of all 21 S. rangiferi-like isolates from red deer (=Sarcocystis truncata; see next section) as a sister taxon to the five isolates of S. silva from roe deer and moose, while both of these taxa formed a sister group to all 33 isolates of S. rangiferi from reindeer. Furthermore, all 16 isolates of S. hjorti from red deer formed a monophyletic group together with the six isolates of this species from moose, and all 17 isolates of S. ovalis from red deer formed a monophyletic group together with 22 isolates of this species derived from Norwegian moose and one isolate from a Canadian moose. Similarly, all nine isolates of S. hardangeri that were characterized in the recent study (Gjerde, Reference Gjerde2013b ), including one from red deer and eight from reindeer, formed a monophyletic group, which was sister taxon to the S. ovalis group.
Phylogenetic tree for members of the Sarcocystidae based on 308 partial sequences of cox1 and inferred using the neighbour-joining method. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (10 000 replicates) is shown next to the branches. The evolutionary distances were computed using the p-distance method and are in the units of the number of base differences per site. Subtrees formed by two or more sequences of the same species have been collapsed, but the number of sequences included is given in parentheses behind the taxon names. The intermediate hosts from which sarcocysts were obtained for DNA isolation (or for infection of definitive hosts: two isolates of S. gracilis from foxes and two isolates of S. ovalis from a magpie) are also given, using the following abbreviations: Aa = Alces alces (moose); Bt = Bos taurus (cattle); Cc = Capreolus capreolus (roe deer); Ce = Cervus elaphus (red deer); Oa = Ovis aries (sheep); Rt = Rangifer tarandus (reindeer).
As regards the overall phylogenetic relationships of the various taxa to each other, all Sarcocystis spp. formed a sister clade to the four members of the Toxoplasmatinae (T. gondii, N. caninum, H. heydorni, H. triffittae), and there was maximum bootstrap support for the placement of the various Sarcocystis spp. into three major clades, corresponding to their known or presumed definitive hosts, i.e. corvid birds (S. hardangeri, S. ovalis, Sarcocystis oviformis); canids (S. alces, Sarcocystis alceslatrans, S. capreolicanis, Sarcocystis cruzi, S. gracilis, Sarcocystis grueneri, S. hjorti, Sarcocystis rangi, Sarcocystis tarandivulpes, S. tenella); and felids/unknown (Sarcocystis gigantea, Sarcocystis hirsuta/S. elongata, S. rangiferi, Sarcocystis scandinavica, S. silva, Sarcocystis sinensis, S. tarandi, S. truncata) (Gjerde, Reference Gjerde2013b ). The relationships between some of the canine-transmitted species were, however, not clearly resolved in the phylogenetic analyses.
TAXONOMIC SUMMARY
The sequence comparisons and phylogenetic analyses based on partial cox1 sequences reported in the previous section clearly show that S. cf. rangiferi in red deer is different from S. rangiferi in reindeer, and is apparently more closely related to S. silva in roe deer and moose. The cox1 data also strongly suggest that S. cf. tarandi in red deer is different from S. tarandi, which is mainly found in reindeer. Hence, S. cf. rangiferi and S. cf. tarandi in red deer will be re-described in the following as Sarcocystis truncata n. sp. and Sarcocystis elongata n. sp., respectively, and their biological, morphological and molecular features will be summarized based on the results of the present study and those reported previously (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ; Gjerde, Reference Gjerde2013b ).
Name: Sarcocystis elongata (syn. S. tarandi, S. cf. tarandi)
Type intermediate host: Red deer (C. elaphus).
Type definitive host: Unknown, but possibly felids based on phylogenetic placement and epidemiological data.
Type locality: Norway; Hordaland and Rogaland counties.
Site of infection in intermediate host: Mainly in skeletal muscles and oesophagus, occasionally in the heart.
Description of sarcocysts: Cysts in diaphragm and oesophagus slender and spindle-shaped with tapering pointed ends, measuring 1·0–2·3×0·07–0·15 mm (including protrusions). Due to their small diameter, cysts may be difficult to detect macroscopically. Cyst surface covered by densely packed, upright, thin, finger-like protrusions, 7–8 μ m long and 1·5–2 μ m wide, with a round to polygonal outline, giving the cysts a thick, finely striated wall. Protrusions distributed in rows in a hexagonal pattern across cyst surface. Cysts in cardiac muscle, 0·2–0·3×0·06–0·13 mm in size, with about 5 μm long, finger-like protrusions. Such cysts are morphologically indistinguishable from those of S. truncata, making molecular identification of cardiac cysts necessary. The sarcocyst morphology as seen by light microscopy (LM) and scanning electron microscopy (SEM) is depicted in Figs 3a–d and 4 in Dahlgren and Gjerde (Reference Dahlgren and Gjerde2010a ).
Type specimens: Type material, consisting of a sarcocyst excised from muscular tissue of red deer, and photographs from LM and SEM examinations of the wall of other sarcocysts, have been deposited at the National History Museum, Oslo, Norway, with collection number NHMO-Prot00007 (originally deposited under the name S. tarandi).
Molecular characteristics: Ten nucleotide sequences (1727–1734 bp) of the near complete ssu rRNA gene differed from each other by 0·1–1·1%, and differed from 10 corresponding sequences of S. tarandi from reindeer (with 0·1–1·0% intraspecific variation) by 0·2–1·2%. Ten nucleotide sequences (∼460–488 bp) of the ITS1 region differed from each other by 0–7·7%, and differed from 10 similar sequences of S. tarandi from reindeer (with 0·2–7·6% intraspecific variation) by 0·3–10·3%. Twenty-five partial cox1 nucleotide sequences (1020–1038 bp) differed from each other by 0–1·5%, and differed from 36 sequences of S. tarandi by 2·4–3·1%.
Nucleotide sequences deposited in GenBank: GQ251011–GQ251020 (ssu rRNA gene; five clones from each of two cyst isolates from two red deer); GQ251001–GQ251010 (ITS1; five clones from each of two cyst isolates from two red deer); KC209705–KC209711 and KF241312–KF241329 (cox1; 25 sequences from 25 cyst isolates from 12 red deer).
Etymology: The species is named from its elongated, slender sarcocysts with pointed tips.
Name: Sarcocystis truncata (syn. S. rangiferi, S. cf. rangiferi)
Type intermediate host: Red deer (C. elaphus).
Type definitive host: Unknown, but possibly felids based on phylogenetic placement and epidemiological data.
Type locality: Norway; Hordaland and Rogaland counties.
Site of infection in intermediate host: Mainly in skeletal muscles and oesophagus, occasionally in the heart.
Description of sarcocysts: Cysts in diaphragm and oesophagus cigar-shaped to ellipsoidal with rounded tips, measuring 0·7–1·7×0·16–0·4 mm; mature cysts macroscopically visible. Cyst surface covered by densely packed, upright, fairly thick, finger-like protrusions, about 7–8 μm long and about 3 μm wide, with a round to polygonal cross-section, giving the cyst a thick, striated wall. Protrusions regularly distributed in rows in a hexagonal pattern across the surface. Cysts and their host cells do not seem to become enclosed by fibrous material. Cysts in cardiac muscle, 0·2–0·3×0·06–0·13 mm in size, with about 5 μ m long, finger-like protrusions; morphologically indistinguishable from those of S. elongata, making molecular identification necessary. The sarcocyst morphology as seen by LM and SEM is depicted in Figs 3e, 5 and 6 in Dahlgren and Gjerde (Reference Dahlgren and Gjerde2010a ). The cyst from cardiac muscle shown in Fig. 3e (isolate Ce34.24) was found in the present study to belong to S. truncata.
Type specimens: Type material, consisting of a sarcocyst excised from muscular tissue of red deer, and photographs from LM and SEM examinations of the wall of other sarcocysts, have been deposited at the National History Museum, Oslo, Norway, with collection number NHMO-Prot00009 (originally deposited under the name S. rangiferi).
Molecular characteristics: Ten nucleotide sequences (1734–1739 bp) of the near complete ssu rRNA differed from each other by 0–1·0%, and differed from 10 corresponding sequences of S. rangiferi from reindeer (with 0–1·1% intraspecific variation) by 0·9–1·8%.Ten nucleotide sequences (∼434–460 bp) of the ITS1 region differed from each other by 0·2–12·8%, and differed from 10 corresponding sequences of S. rangiferi from reindeer (with 0·2–10·0% intraspecific variation) by 5·3–14·1%. Twenty-one partial cox1 nucleotide sequences (1029–1095 bp) differed from each other by <1%, and differed from 33 sequences of S. rangiferi from reindeer by ∼7% and from five sequences of S. silva by ∼4%.
Nucleotide sequences deposited in GenBank: GQ251021–GQ251030 (ssu rRNA gene; five clones from each of two cyst isolates from two red deer); GQ250991–251000 (ITS1; five clones from each of two cyst isolates from two red deer); KC209677–KC209683 and KF241439–KF241452 (cox1; 21 sequences from 21 cyst isolates from 14 red deer).
Etymology: The species is named from its thick, cigar-shaped sarcocysts with blunt, truncated tips.
DISCUSSION
The present investigation has confirmed and expanded the findings in the previous study employing partial cox1 sequences for delimitation of Sarcocystis spp. in red deer and other hosts (Gjerde, Reference Gjerde2013b ), and, in so doing, it has contradicted some of the original species designations of sarcocysts in Norwegian red deer, which were based on sarcocyst morphology and nucleotide sequences of the nuclear ribosomal DNA unit (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ). In the following, the characteristics and relationships of the six Sarcocystis spp. in red deer to each other and to similar species in other cervid hosts will be discussed.
Sarcocystis truncata/S. rangiferi/S. silva
Based on cox1 sequences from many isolates, S. truncata of red deer seems to be clearly different from S. rangiferi in reindeer, and this species is apparently more closely related to S. silva in roe deer and moose than to S. rangiferi. Such a relationship was also suggested by phylogenetic reconstructions using ssu rRNA gene sequences, but only when more isolates of S. silva had been characterized at the latter gene (Gjerde, Reference Gjerde2012, Reference Gjerde2013b ) than in the original study (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ). In the previous phylogenetic analyses using ssu rRNA gene sequences (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ; Gjerde, Reference Gjerde2012, Reference Gjerde2013b ), isolates of S. truncata seemed to be even more closely related to an unnamed Sarcocystis sp. from sika deer (Cervus nippon) in Japan than to S. silva and S. rangiferi, which suggests that these Cervus spp. share the same Sarcocystis sp. Whether this is so, might be determined by sequencing cox1 of this Sarcocystis sp. in Japanese sika deer.
The sarcocysts of S. truncata in red deer are quite similar to those of S. rangiferi in reindeer as regards the general size and shape of the cysts and their protrusions (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ), but they do not seem to become encapsulated by a fairly thick layer of host-derived fibrous material, which is a consistent feature of S. rangiferi sarcocysts in reindeer (Gjerde, Reference Gjerde1984a , Reference Gjerde1985b , Reference Gjerde1986). This difference supports the conclusion from this study that they are separate species, but it could also be due to differences in host reactions between red deer and reindeer. However, a similar encapsulation also occurs around cysts of S. hardangeri in both red deer and reindeer (Gjerde, Reference Gjerde1984b , Reference Gjerde c , Reference Gjerde1985c ) and around cysts of S. ovalis in both red deer and moose (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2008, Reference Dahlgren and Gjerde2010a ), suggesting that this type of host reaction is induced by particular Sarcocystis spp. irrespective of their hosts. The sarcocysts of S. truncata have the same shape and the same type of surface protrusions as sarcocysts of S. silva in roe deer, but only a few cysts of the latter species have so far been examined morphologically, and only by LM (Gjerde, Reference Gjerde2012). The unnamed Sarcocystis sp. from sika deer that was related to S. truncata based on its ssu rRNA gene sequences was not described morphologically in connection with the molecular characterization, but the Type 1 cysts described from other sika deer in Japan by Narisawa et al. (Reference Narisawa, Yokoi, Kawai, Sakui and Sugawara2008) are consistent with those of S. truncata.
Sarcocystis elongata/S. tarandi
In previous phylogenetic analyses using ITS1 and/or ssu rRNA gene sequences, isolates of the S. tarandi-like species in red deer have clustered together with and been interspersed with isolates of S. tarandi from reindeer (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ; Gjerde, Reference Gjerde2012, Reference Gjerde2013b ). Hence, the isolates from red deer and reindeer seemed to be conspecific. However, based on cox1 sequences from a considerable number of isolates from many different animals, it seems reasonable to conclude that two separate, but closely related species exist, and that the erection of a new species, S. elongata, is justified. The distinction between S. elongata and S. tarandi was not as clear-cut, however, as that between S. truncata and S. rangiferi, as also suggested by the differences in the clustering pattern of the ssu rRNA gene and ITS1 sequences between these species pairs (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ). Nevertheless, the intraspecific sequence variation among cox1 sequences assigned to either S. elongata or S. tarandi did not overlap with the small interspecific sequence variation between the two taxa, and the phylogenetic analyses consistently placed the isolates of each taxon in separate monophyletic clusters. The finding of only a single, but typical, isolate of S. tarandi in red deer, also suggests that there are two distinct populations, or species, with different affinity and infectivity for red deer and reindeer, respectively, i.e. S. elongata occurring only/mainly in red deer and S. tarandi occurring mainly in reindeer. Since the sarcocysts of S. elongata are morphologically indistinguishable from those of S. tarandi (Gjerde, Reference Gjerde1984a , Reference Gjerde1985d , Reference Gjerde1986; Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ), and since S. tarandi may occur in red deer, a reliable species diagnosis of this sarcocyst type in red deer can only be achieved by molecular methods, and currently only by using cox1 as the genetic marker.
Sarcocystis elongata/S. truncata
By microscopy, four small cysts isolated from the heart of three of 37 red deer were identified as belonging to either S. elongata (S. cf. tarandi) or S. truncata (S. cf. rangiferi) due to the presence of upright finger-like protrusions (see Fig. 3d–e in Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ), but only one of these cysts were examined by molecular methods (partial ssu rRNA gene sequence) in the first study and identified as a cyst of S. elongata. In the recent study (Gjerde, Reference Gjerde2013b ), this identification was confirmed by using cox1 sequences (GenBank number KC209707), whereas two additional cysts (GenBank numbers KC209680–KC209681) were identified as belonging to S. truncata. In the present study, the fourth sarcocyst of this type was examined, and found to belong to S. truncata (GenBank number KF241452). Thus, both S. elongata and S. truncata may occur in cardiac muscle, but with similarly sized sarcocysts and with indistinguishable protrusions since the growth of the cysts is limited by the small size of cardiac muscle cells. In skeletal muscle, on the other hand, mature cysts of S. elongata and S. truncata are clearly different in both size and shape, as reflected in the proposed new species names. Young and small cysts of S. truncata might, however, be confused with mature cysts of S. elongata when examined by LM. Thus, three sarcocysts from skeletal muscle that had been preliminary identified as S. elongata based on their morphology, turned out to belong to S. truncata from their cox1 sequences. Such confusion may also occur between young sarcocysts of S. rangiferi and mature cysts of S. tarandi in reindeer.
It will often be necessary to distinguish between these two cyst types when examining muscle samples from red deer and reindeer, since individual red deer usually are concurrently infected with S. elongata and S. truncata (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ), while individual reindeer are usually concurrently infected with S. rangiferi and S. tarandi (Gjerde, Reference Gjerde1984a ), rather than hosting only one species in each pair. This infection pattern suggests that these four Sarcocystis spp. use the same definitive host, and this notion has been further supported by their genetic similarity and phylogenetic placement. No definitive hosts have so far been determined for these species, but their phylogenetic placement and common occurrence suggest that felids like domestic cats and lynx (Lynx lynx) may act in this capacity (Dahlgren, Reference Dahlgren2010). The cox1 primers and reference sequences generated in the previous (Gjerde, Reference Gjerde2013b ) and the present study may facilitate the molecular identification of the oocyst/sporocyst stage of these species, and thus the determination of their definitive hosts.
Species resembling S. elongata and S. truncata with respect to sarcocyst morphology have been reported from several cervid hosts in different countries as discussed in detail previously (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ). Of particular interest is the S. truncata-like species in red deer referred to as Sarcocystis cf. hofmanni by Wesemeier and Sedlaczek (Reference Wesemeier and Sedlaczek1995) and by Kutkienė (Reference Kutkienė2003), and the S. elongata-like species from sika deer depicted in Fig. 1 of the paper by Saito et al. (Reference Saito, Itagaki, Shibata and Itagaki1995). The relationships between S. elongata and S. truncata and the other similar species in different cervids can only be revealed through a molecular characterization of the other taxa.
Sarcocystis hjorti
Sarcocystis hjorti, having small sarcocysts with delicate hair-like protrusions, was the most prevalent species in Norwegian red deer (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ), and was found to be identical at the ssu rRNA gene with Sarcocystis sp. Type E previously detected in moose (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2008). The latter sequence had been obtained from a small cyst in cardiac muscle with no visible protrusions, but when the smallest visible cysts in the diaphragm of several other moose were specifically targeted for isolation in a subsequent study, several small cysts of S. hjorti with typical hair-like protrusions were found (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010b ). Thus, in moose, cysts of S. hjorti might be difficult to discern due to the concurrent presence of numerous cysts of S. alces, which look similar in situ, at least when not yet fully developed. The conspecificity of S. hjorti-like cysts in red deer and moose was confirmed by comparing cox1 sequences in the recent study (Gjerde, Reference Gjerde2013b ), and that finding was further supported by examining 13 more isolates from red deer and moose in the present study (Table 1). Again, the phylogenetic analyses placed all the isolates from both cervids in a monophyletic group (Fig. 1), confirming that S. hjorti uses both red deer and moose as intermediate hosts. Experimental infections have shown that S. hjorti uses foxes as definitive hosts (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010b ), which had been suggested by its common occurrence in red deer and moose and its phylogenetic placement (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a , Reference Dahlgren and Gjerde b ). Sarcocysts of S. hjorti have the same type of delicate, hair-like protrusions as sarcocysts of S. rangi in reindeer (Gjerde, Reference Gjerde1984c , Reference Gjerde1986; Dahlgren et al. Reference Dahlgren, Gjerde, Skirnisson and Gudmundsdottir2007), S. alceslatrans in moose (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2008; Dahlgren, Reference Dahlgren2010) and S. capreolicanis in roe deer (Gjerde, Reference Gjerde2012), and these species also cluster together in phylogenetic analyses using either ssu rRNA gene sequences or cox1 sequences, suggesting a very close relationship among these four canine-transmitted Sarcocystis species of cervids. Recently, S. hjorti has been detected by molecular methods in red deer and moose in Lithuania (Prakas and Butkauskas, Reference Prakas and Butkauskas2012) and in red deer in Switzerland (Stephan et al. Reference Stephan, Loretz, Eggenberger, Grest, Basso and Grimm2012).
Sarcocystis ovalis/S. hardangeri
In the first study of Sarcocystis spp. in red deer, partial or complete ssu rRNA gene sequences were obtained from 13 ovoid sarcocysts, and 12 of these sequences were near identical with that of S. ovalis in moose, whereas one sequence was near identical with S. hardangeri in reindeer (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ). The species assignments of all of these 13 cyst isolates from reed deer have subsequently been confirmed by using cox1 sequences, either in the previous (Gjerde, Reference Gjerde2013b ) or in the present study. Moreover, four additional cyst isolates of the S. ovalis-type were sequenced in this study and found to belong to S. ovalis. Thus, S. ovalis seems to be the more common of these two species in red deer, at least in western Norway. At cox1, there were consistent differences at two–three nucleotide positions between isolates of S. ovalis from the two hosts, but this might be due to sampling of geographically separated host populations, i.e. red deer in western Norway and moose in south-eastern Norway. The same might apply to the small nucleotide differences found between the isolates of S. hjorti in red deer and moose, respectively. However, the European magpie (Pica pica) has been found to act as a definitive host for S. ovalis from moose, and other corvid birds, particularly the hooded crow (Corvus cornix) and the common raven (Corvus corax) are assumed to act as additional definitive hosts for this species and the related species S. hardangeri and S. oviformis (Gjerde and Dahlgren, Reference Gjerde and Dahlgren2010). Thus, corvid birds might possibly disseminate these Sarcocystis spp. across considerable distances and between geographically separated deer populations.
Sarcocystis cervicanis
Still another Sarcocystis sp. has been recorded in red deer and elk in other countries, having been referred to as an unnamed Sarcocystis sp. (Entzeroth et al. Reference Entzeroth, Neméseri and Scholtyseck1983), Sarcocystis cervicanis (Hernández-Rodríguez et al. Reference Hernández-Rodríguez, Martínez-Gómez, Navarrete and Acosta-Garcia1981), Sarcocystis wapiti (Speer and Dubey, Reference Speer and Dubey1982) and Sarcocystis cf. grueneri (Wesemeier and Sedlaczek, Reference Wesemeier and Sedlaczek1995). The latter name stems from the fact that this species is morphologically indistinguishable from S. grueneri of reindeer (Gjerde, Reference Gjerde1985a , Reference Gjerde1986), which is very common in this host in Norway (Gjerde, Reference Gjerde1985a ; Dahlgren and Gjerde, Reference Dahlgren and Gjerde2007). The fact that such a species has not yet been found in Norwegian red deer, might indicate that it is different from S. grueneri in reindeer, but this can only be determined by a molecular characterization of this species in red deer.
CONCLUSIONS
The findings of the present and the preceding study (Gjerde, Reference Gjerde2013b ) using cox1 sequences for species delimitation, have in part confirmed and in part contradicted the findings of the original study on Sarcocystis spp. in Norwegian red deer (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ). Altogether six Sarcocystis spp. have now been recorded in this host. Two of these species (S. hjorti, S. ovalis) seem to be fairly common in both red deer and moose, two species (S. hardangeri, S. tarandi) seem to mainly occur in reindeer, while two species (S. elongata, S. truncata) are common in red deer, but have so far not been recorded in other hosts. Thus, at least four of the six species recorded in Norwegian red deer are not strictly intermediate host specific, and the claim set forth in the title of the previous paper (Dahlgren and Gjerde, Reference Dahlgren and Gjerde2010a ) is therefore still valid, and has even been further substantiated in the present study using cox1 sequences. However, the two renamed species were not identical to their counterparts in reindeer as suggested from their sequence similarity at the ssu rRNA gene and ITS1 locus.
This study has also demonstrated that mitochondrial cox1 sequences of appropriate length are better able to discriminate between some closely related and thus recently diverged Sarcocystis species in cervids than nucleotide sequences of the nuclear ribosomal DNA unit. Hence, it is recommended that cox1 sequences are obtained in future molecular characterizations of different Sarcocystis spp. in various hosts, but not to the exclusion of the ssu rRNA gene, but as a supplement to this locus. It may, however, be better to use cox1 as a genetic marker for molecular identification of oocysts/sporocysts in the intestines/feces of naturally or experimentally infected definitive hosts of various Sarcocystis spp., particularly when trying to identify species with a considerable sequence variation at the ssu rRNA gene.
ACKNOWLEDGEMENTS
The author would like to acknowledge the significant contribution of Stina S. Dahlgren in the previously published initial characterization of Sarcocystis spp. in Norwegian red deer.
