Our study on myxozoan minicollagen genes revealed that the gene repertoire is less reduced as previously indicated [17]. Our mining of myxozoan transcriptomes and genomes supported the existence of Ncol-4 by seven newly identified myxosporean Ncol-4 sequences, which showed high sequence similarities and close phylogenetic relationships to Myxobolus pendula Ncol-4 identified by [20]. We revealed the presence of a fifth myxozoan minicollagen, Ncol-5, which we determined in 11 myxosporean species. Myxozoan Ncol-5 is structurally similar to myxozoan Ncol-1 and Ncol3 in the size of the central GlyXY (collagen) domain region surrounded by single CRD at both ends. Ncol-5 may be classified as a group one minicollagen based on the character of the CRDs [11]. The character of both Ncol-5 domain regions differs from that of most cnidarians and that of myxozoan Ncol-1,2 and 3, which are composed of polyprolines. Ncol-5 does not include proline-rich domains but regions rich in serine and glycine (Additional file 4). Similar modifications of the polyproline domain can be found also in other cnidarians. Cubozoans Morbakka virulenta Ncol-7 and Alatina moseri Ncol-2 have modifications in the first polyproline region (between CRD and GlyXY domain) characterized by glycine-rich region followed by alanine-proline repeats. Hydra vulgaris Ncol-15–17 contains a relatively long region with a reduced number of proline amino acids and a number of serines and glycines. More importantly, five identified minicollagens of P. hydriforme Ncol-7–11 contain polyproline regions that are non-canonical with a low number of prolines and abundance of glycine, serine, and alanine amino acids (Additional file 4).
Reconstruction of the evolutionary history and the phylogenetic relationships of the minicollagen gene family is challenging. The characteristics of minicollagen genes make the construction of sequence alignments complicated, with regions containing ambiguously aligned amino acid positions. Therefore, we decided to exclude the polyproline domains from the dataset because these regions were highly variable with many ambiguously aligned sites, even in closely related species. Despite that, we observed a high degree of instability in the phylogenetic trees as a consequence of the low phylogenetic signal of these protein-coding genes. These inconsistencies in the tree topologies resulting from different phylogenetic methods of tree reconstruction demonstrated that the character of cnidarian minicollagens does not allow us to reveal its evolution with high reliability.
We proved that wide cnidarian minicollagen taxon sampling performed in our work did not influence previously determined relationships of three myxozoan minicollagens (Ncol-1,2,3) and minicollagens of Polypodium hydriforme [17, 20]. Although BI analysis gave a more conservative illustration of the phylogeny of minicollagen genes with polytomies at weakly supported nodes, our ML analysis supported a close relationship of myxozoan Ncol-1 and P. hydriforme Ncol-1–4, myxozoan Ncol-2 and P. hydriforme Ncol-5 and 6, and myxozoan Ncol-3 and P. hydriforme Ncol-11. However, we did not reveal a close relationship between myxozoan Ncol-4 and Polypodium Ncol-7,8,9 that was shown by [20]. Our BI analysis showed a close position of myxozoan Ncol-5 and Polypodium Ncol-10 that strengthens the view of the common evolutionary history of the Myxozoa and P. hydriforme [27].
We were not able to determine myxozoan Ncol-4 and Ncol-5 in malacosporeans, myxozoans with soft-wall malacospore, and parasitism in bryozoan hosts, which could suggest that these two minicollagens are unique only for myxosporeans. Ncol-4 and Ncol-5 may either evolved after myxosporean-malacosporean split from an ancestral minicollagen orthologue or malacosporeans lost these genes during the evolution. The absence of Ncol-4 and Ncol-5 in malacosporeans may also be explained by the character of available data (March 2020) because ESTs may not include all minicollagens present in the genome of Buddenbrockia plumatellae and Tetracapsuloides bryosalmonae, respectively.
Newly identified Ncol-4 sequences from seven myxosporeans revealed high sequence similarity with Ncol-4 of Myxobolus pendula first identified by [20], with an exceptional short tripeptide GlyXY, lack of polyproline behind GlyXY, and a unique C-terminal CRD domain interrupted by a polyproline stretch. Our analysis also associated Ncol-2-like minicollagen of Ceratonova shasta identified in the proteome of this myxosporean by [36] with Ncol-4 cluster of myxozoan minicollagens. A greater sampling of myxozoan Ncol-4 sequences helped to better identify the structure of myxozoan Ncol-4. In six myxosporeans (including M. pendula), the tripeptide domain includes only four repeats of GlyXY rather than seven as documented in [20]. Alanine is lacking in the fifth repeat of the GlyXY, replacing the glycine residue [20] but it is the first amino acid of the modified polyproline domain. Myxosporeans Kudoa iwatai and Enteromyxum leei present five and six GlyXY repeats, respectively. These two myxosporeans are also the only ones with proline residues behind the GlyXY domain of Ncol-4. The modified C-terminal Ncol-4 polyproline domain of other myxosporeans contains predominantly glycine and serine residues, similarly as myxozoan Ncol-5, which may suggest their common evolutionary origin, indicated by their separated clustering from all other cnidarian minicollagens. Moreover, the E. leei N-terminal polyproline domain lost all proline residues and includes also mostly glycine and serine amono acids. Interestingly, P. hydriforme Ncol-7–9 have only two (non-proline) amino acid residues between GlyXY and C-terminal CRD. This atypical polyproline domain character may suggest close relationships of P. hydriforme Ncol-7–9 and myxozoan Ncol-4 but our phylogenetic analysis did not suggest this scenario in contrast with [20] phylogenetic reconstruction.
We conclude that the myxozoan minicollagen repertoire is similar to that known from other cnidarian groups but has two types unique to myxozoans, more specifically myxosporeans (Ncol-4 and Ncol-5). Our analysis suggested five clusters of minicollagen homologues in Myxozoa as well as in Scyphozoa and Cubozoa, six clusters in Hydrozoa and Polypodium, and only three clusters in Anthozoa. The evolutionary reconstruction showed that the 21 known minicollagen types reported from Hydra [22] are a result of recent evolutionary multiplication of ancestral types of the minicollagen, in different clusters. For example, H. vulgaris Ncol-3,4,8,9,13,14 or Ncol-15–17 share a common recent ancestor as well as paralogs of P. hydriforme Ncol-1–4 or Ncol-7–9. The number of these main minicollagen clusters is thus similar in all medusozoan groups including parasitic Myxozoa. Multiplication and the resulting variety of minicollagen homologues in Hydra and other cnidarians may be in agreement with the hypothesis that the diversity of minicollagen is linked to the nematocyst diversity, which is the highest in Hydrozoa [11]. Only a single nematocyst type has been reported from Myxozoa and this may be the reason for the limited multiplication of the minicollagen types within the myxozoan minicollagen gene repertoire (only one duplication in S. molnari).
Our phylogenetic analysis of the minicollagen genes supports a close relationship between the Myxozoa and P. hydriforme [27]. Three out of five myxozoan minicollagen homologues showed close relationships with Polypodium minicollagens in ML analysis. Newly identified myxozoan minicollagen Ncol-4 and Ncol-5 were not associated with the evolution of any cnidarian homologue. This is in disagreement with [20] who revealed close relation of Myxobolus pendula Ncol-4 to P. hydriforme Ncol-7–9. Our analysis does not include polyproline domains in the alignment, because we found them too variable to be unambiguously aligned. However, if we kept these regions and performed the phylogenetic analysis, both, ML and BI did not support the close relationship of these homologues (data not shown). Therefore, we may suppose that wider taxon sampling of myxozoan Ncol-4 may influence its phylogenetic position. Myxozoan Ncol-4 represents a rather independent minicollagen lineage with no closely related cnidarian taxa. P. hydriforme Ncol-7–9 clustered in the large group of minicollagens representing all cnidarian taxa also including all three anthozoan clusters of minicollagen homologues. We may expect the existence of one more myxozoan minicollagen type that would cluster within the above mention group of P. hydriforme minicollagens. Alternatively, this putatively missing myxozoan minicollagen may be the result of gene loss during the evolution of myxozoan parasitic lifestyle.
Minicollagen genes were found to be organized into clusters in the genome with collinear expression in representatives of Anthozoa, Hydrozoa, Cubozoa, and Scyphozoa [8, 11, 37]. We identified myxozoan Ncol-1 and Ncol-4 on the single contig in a very close distance that supports a similar character of the gene organization in Myxozoa (Fig. 4). The close physical proximity of Ncol-1 and Ncol-4 on the genome level suggests an ancestral gene duplication event and subsequent neofunctionalization of either Ncol-1 or Ncol-4. However, the other three myxozoan minicollagens are very likely not organized in clusters, which is indicated by the observed long-distance between Ncol-2 and Ncol-3 in the genome of Thelohanellus kitauei. Interestingly, we identified rearrangement of the gene order in Ncol-1 and Ncol-4 gene cluster and different transcription directions, potentially suggesting diverse genomic organization and noncollinear expression compared to cnidarian minicollagen gene clusters [37].
Our newly generated transcriptome of Nephrocystidium pickii represents novel transcriptomic data of myxozoan extrasporogonial stages for which a great multiplication of parasite is typical and the spore formation was never observed. Moreover, these stages of N. pickii replete the cytoplasm of endothelial cells of the host glomerular capillary and thus representing an atypical myxozoan xenoma-like intracellular development stage. The finding of minicollagen gene expression in these extrasporogonic myxozoan stages was surprising. We detected two minicollagen genes Ncol-3 and Ncol-5 in the transcriptome of the xenoma stage of Nephrocystidium pickii as well as Ncol-1,3,5 in the transcriptome of blood stages of Sphaerospora molnari. We assume that the expression levels of these minicollagens will be very low in these stages and increase later in spore-forming stages. However, myxozoan Ncol-3 and Ncol-5 might be the earliest expressed minicollagens in the myxozoan development, with basal levels of gene transcription detectable in the non-spore-forming stages. Sequence differences in the Ncol-3 and Ncol-5 of N. pickii and Myxidium lieberkuehni supported SSU rDNA-based analysis of [32] that proven non-conspecificity of pike parasites M. lieberkuehni and N. pickii.
The lack of proline residues together with a high number of glycine and serine residues between CRDs and the tripeptide domain documented in Ncol-5 and partially also in Ncol-4 may be an adaptation of myxozoan minicollagen genes to some specific function of these parasitic cnidarians. [38] reported a contraction of the myxozoan polar tubule after its release from the myxozoan polar capsule, causing the spore to be moved closer to the host surface for entry of the infective sporoplasm. Such polar tubule elasticity is a unique myxozoan feature unknown in nematocyst tubes of free-living cnidarians. We hypothesize that the contraction of the polar tubule may be enabled by the character of myxozoan Ncol-4 and Ncol-5 that include glycine-rich domains instead of polyproline ones. Glycine-glutamine-rich domain is part of elastic protein Cnidoin that was discovered in the proteome of Hydra nematocysts [22] and its elastic sequence is homologous to the glycine-rich region of the spider silk protein Spidroin-2 [39]. Both Cnidoin and myxozoan Ncol-4 and Ncol-5 reveal the glycine-rich region next to CRD, which is supposed to be involved in network formation with other minicollagens [14]. The putative elastic glycine-rich domain of the myxozoan Ncol-4 and Ncol-5 minicollagens may have an elastic function enabling the polar tubule to contract after its evertion.
Myxozaons are unique in the formation of resistant spores, a feature not known from any free-living cnidarian. Thus, we can also hypothesize, that atypical minicollagens Ncol-5, as well as, Ncol-4 may be involved in the synthesis of the spore valve, whose constitution is unknown. The presence of minicollagens in the spore valve wall could be expected as the spore valves similar to polar capsules must resist high pressure, which is caused by the water environment and enhanced when sinking in the water column after spore release from the host. Minicollagens can form a distinct pattern that can provide the tensile strength to withstand the high osmotic pressure [40], a feature that structural components of the spore valves should possess to survive at different depths. The absence of Ncol-4 and Ncol-5 in malacosporeans discussed above, thus maybe associated with a soft-walled character of their spore valves. Malacosporeans lacking Ncol-4 and Ncol-5 may evolve different components of spore valves that are not so resistant to water pressure, which may have consequences in the ecological limits of their distribution limited only to the freshwater environment with relatively shallow waters.