Synopsis of the aurantiactinomyxon collective group (Cnidaria, Myxozoa), with a discussion on the validity of morphotype definition and demise of guyenotia

Aurantiactinomyxon is one of the most diverse myxozoan collective groups, comprising types that mostly infect freshwater and marine oligochaetes belonging to the family Naididae Ehrenberg, 1828, but also Lumbriculidae Claus, 1872. In this study, a comprehensive revision of all known aurantiactinomyxon types is performed and highlights the fallibility of using the form and length of the valvular processes as main criterion for differentiating among style-less actinospore morphotypes. The demise of the guyenotia collective group is proposed based on the ambiguous features of several types that allow conformity with both the aurantiactinomyxon and guyenotia definitions. Nonetheless, the information presently available clearly shows that a general shift is needed in our approach to actinospore grouping, which should probably be based on actinospore functionality relative to environment and host ecology, rather than on morphology. Life cycle studies based on experimental transmission and molecular inferences of the 18S rDNA have linked aurantiactinomyxon (including former guyenotia) to myxozoans belonging to a diverse array of genera, including Chloromyxum, Henneguya, Hoferellus, Myxobolus, Paramyxidium, Thelohanellus and Zschokkella. This undoubtedly shows a high capacity of the aurantiactinomyxon morphotype to promote infection in intrinsically distinct vertebrate hosts and environmental habitats, consequently increasing interest in its study for attaining a better understanding of myxozoan-host interactions. The identification of novel and known types, however, is impeded by the lack of concise information allowing a comprehensive analysis of biological, morphological, and molecular criteria. In this sense, the compilation of data presented in this study will ultimately help researchers seeking to perform reliable identifications.


Introduction
Infections of aquatic oligochaetes by actinospores were first reported by Š tolc (1899), who created the Actinomyxidia to encompass hexactinomyxon, synactinomyxon and triactinomyxon types that the author found infecting tubificids in Czechia. Over time, recognition of the homology between these organisms and myxozoans, led to the relocation of Actinomyxidia to the phylum Myxozoa Grassé, 1970, which became divided into the classes Myxosporea Bütschli 1881 (fish parasites producing myxospores) and Actinosporea Noble, 1980 (worm parasites producing actinospores). In 1984, the ground-breaking discovery that Myxobolus cerebralis Hofer, 1903 develops triactinomyxon actinospores in the gut epithelium of the oligochaete Tubifex tubifex (Müller) (Wolf & Markiw, 1984), showed that myxozoan life cycles comprise both myxospore and actinospore phases, with members of the classes Myxosporea and Actinosporea representing morphologically distinct phases of the same species. This led to a major taxonomic revision, with Kent et al. (1994) proposing the demise of the class Actinosporea, and the use of its generic names as vernacular designations for actinospore morphotypes established within distinct collective groups.
Currently, there are ca. 20 valid actinospore collective groups (Lom & Dyková, 2006;Rangel et al., 2011;Milanin et al., 2017;Atkinson et al., 2019;Rocha et al., 2019aRocha et al., , b, 2020, with aurantiactinomyxon being one of the most diverse. This collective group was first described by Janiszewska (1957), who defined its actinospores having a style-less epispore, with three equal processes that curve downwards and embrace with their whole base the epispore cavity. Lom & Dyková (2006) updated the definition, and described aurantiactinomyxon as having three stout, semicircularly curved, leaf-like valvular processes attached to an ellipsoidal body with protruding polar capsules at the apex and containing a sporoplasm with many secondary cells. To date, 61 aurantiactinomyxon types have been reported based on these definitions (Table 1), with differentiation between types mostly relying on morphometric comparisons. Molecular data of the 18S rDNA is available for only 20 types. Another eight 18S rDNA sequences are available in GenBank but constitute unpublished submissions to the NCBI database; while the sequences with GenBank accession numbers MN294775 and MN294776 appear identified as aurantiactinomyxon in the database but have been published as belonging to the presently demised echinactinomyxon collective group (see Rocha et al., 2019a;Gao et al., 2021).
Assigning novel types to a specific collective group can be complicated given the overlapping definitions of several groups that share main morphological features, such as the formation of single spores versus multi-spore cages, and presence/absence of style or valvular processes. In fact, a boundary-less ''continuum of form'' has been suggested to exist between collective groups that differ based solely on the form of a specific morphological character, as is the case of aurantiactinomyxon, echinactinomyxon Janiszewska, 1957, guyenotia Naville, 1930, neoactinomyxum Granata, 1922and raabeia Janiszewska, 1955 (see Hallett et al., 2006;Atkinson, 2011). Differentiation between these style-less morphotypes is based on the shape and length of the valvular processes, which are traditionally defined as being long and straight in echinactinomyxon, curved in raabeia, leaf-like and curved downwards in aurantiactinomyxon, digitiform in guyenotia, and short and spherical in neoactinomyxum (see review in Lom & Dyková, 2006). Although the original definition of aurantiactinomyxon also stated that the valvular processes embraced with their whole base the epispore cavity (Janiszewska, 1957), this criterion has been widely disregarded by researchers (see types in Burtle et al., 1991;Bellerud, 1993;Yokoyama et al., 1993;Yokoyama, 1997;Trouillier et al., 1996;Hallett et al., 1997;El-Mansy et al., 1998a, b;Székely et al., 1998;Ö zer et al., 2002a;Rosser et al., 2014;Milanin et al., 2018), not being included in the updated definition by Lom & Dyková (2006).
Recently, Rocha et al. (2019a) showed that the shape of the valvular processes is a morphological character too variable for distinguishing between raabeia and echinactinomyxon, based on the observation of a type producing actinospores with long valvular processes that were either curved or straight. Consequently, the demise of the echinactinomyxon collective group was proposed and the definition of raabeia was updated to encompass actinospores having straight valvular processes (Rocha et al., 2019a). Similarly, many of the aurantiactinomyxon types included in this synopsis display ambiguous features allowing conformity with the definitions of other collective groups. For instance, Aurantiactinomyxon janiszewskai and the Aurantiactinomyxon type of Xi et al., 2013 have long valvular processes that best resemble those of raabeia (Bellerud, 1993;Xi et al., 2013). Although the processes of the first curve downwards as traditionally described for aurantiactinomyxon, the second is depicted has having straight processes and would probably be better allocated to raabeia. The boundary between these collective groups is further blurred by the report of aberrant spores displaying unequal and different-shaped caudal processes, as is the case of the Raabeia type 4 of Ö zer  Aurantiactinomyxon stellans Marques, 1984 Burtle et al., 1991;Bellerud et al., 1992Bellerud et al., , 1995Pote et al., 1992Pote et al., , 2000Styer et al., 1991Styer et al., , 1992 Dero digitata  52.0 ± 9.9 (37.4-63.0) 5.9 ± 0.0 (5.9-5.9) --8 - Bellerud, 1993               c Originally suggested to potentially be Aurantiactinomyxon pavinsis, but 18S rDNA does not match the sequence obtained by Holzer et al. (2004) from the type-host.
d Presents a case of polymorphism, in which two different phenotypes of Aurantiactinomyxon share one genotype. e Type 1 was previously described as Aurantiactinomyxon type 4 by Borkhanuddin (2013), with slightly distinct measurements. et al., (see Ö zer & Wootten, 2002. The distinction between aurantiactinomyxon and neoactinomyxum is also tenuous, given that several aurantiactinomyxon types have short valvular processes, which only differ from neoactinomyxum by being triangular or rounded with slightly pointed ends rather than completely spherical (see types in Hallett et al., 1997;Negredo & Mulcahy, 2001;Oumona et al., 2003;Székely et al., 2000Székely et al., , 2003Székely et al., , 2004Xi et al., 2015;Zhao et al., 2016). It should be noted that a few neoactinomyxum have been reported to have triangular valvular processes [see types described by Borkhanuddin et al. (2014) and Xi et al. (2015)]. However, it is with guyenotia that the lack of a distinctive boundary is most evident. Several aurantiactinomyxon types described in the literature have digitiform valvular processes that conform with the definition of guyenotia, including the Aurantiactinomyxon type of Burtle et al., 1991 Hallett et al., 2006. The inclusion of types with digitiform processes within the aurantiactinomyxon collective group was previously noticed by Xiao & Desser (1998), who suggested they should be transferred to guyenotia. However, the fallibility of this morphological criterion has led authors to compare aurantiactinomyxon and guyenotia interchangeably (see Burtle et al., 1991;Eszterbauer et al., 2006;Xi et al., 2013). Moreover, Hallett et al. (2002) proved that a single aurantiactinomyxon type can produce actinospores with different process length and shape, having observed two distinct phenotypes associated with the same genotype: one displaying swollen, leaf-like processes with either pointed or rounded ends, and the other having elongated, digitiform-like processes. This clearly shows that there is no real boundary between aurantiactinomyxon and guyenotia. Consequently, the demise of the guyenotia collective group is here proposed, with the transference of its types to aurantiactinomyxon. Original names are retained so as not to increase confusion. The decision to invalidate the oldest group rather than the most recent one relates to the low number of guyenotia that have been reported in the literature; only 5 types of guyenotia have been described versus the 61 types of aurantiactinomyxon that are presently known (see Table 1). Accordingly, aurantiactinomyxon is tentatively defined as having a spherical, subspherical, cylindrical or triangular actinospore body with 3 polar capsules protruding from the apex. Three equally sized lateroposterior valvular processes arise from the actinospore body without a style, curving downwards and tapering to a rounded or pointed end, being leaf-like, propellerlike, digitiform or triangular. Nonetheless, this should be regarded as a temporary definition, given that the increase of our knowledge of actinospore biodiversity will undoubtedly blur even more the boundaries between aurantiactinomyxon, raabeia, and even neoactinomyxum. Overall, this ''continuum of form'' demonstrates that a general shift is needed in our approach to actinospore grouping (Atkinson, 2011), which should probably be based on actinospore functionality relative to environment and host ecology, rather than on morphology.
The great majority of aurantiactinomyxon types reported in the literature infect freshwater oligochaetes belonging to the family Naididae Ehrenberg, 1828 [currently includes members of the former Tubificidae (Erséus et al., 2008)], with reports mainly from the species Branchiura sowerbyi Beddard, but also T. tubifex, Limnodrilus hoffmeisteri Claparède, and Dero digitata (Müller), and less frequently from Lophochaeta ignota Š tolc, and members of the genera Nais Müller and Pristina Ehrenberg. A few types have their oligochaete hosts identified only up to the genusor family-level (see Marques, 1984;Grossheider & Körting, 1992;Benajiba & Marques, 1993), while a few others lack host information (see El-Mansy et al., 1998a;Oumouna et al., 2003;Hallett et al., 2006). Only the three Aurantiactinomyxon types described by Hallett et al. (1997), and the Aurantiactinomyxon type of Rocha et al., 2019, are known to occur in the marine environment, parasitizing naidid oligochaetes belonging to the genera Limnodriloides Pierantoni, Pacifidrilus Erséus, and Tubificoides Lastočkin. The only exceptions to the usage of naidids as hosts are Aurantiactinomyxon pavinsis, widely reported from the freshwater lumbriculid Stylodrilus heringianus Claparède (see Marques, 1984;Oumouna et al., 2003;Holzer et al., 2004;Marcucci et al., 2009) Xiao & Desser, 1998 was also reported from L. variegatus (Xiao & Desser, 1998). A few types have been reported from more than a single host species: Aurantiactinomyxon raabei junioris, Aurantiactinomyxon minor, Aurantiactinomyxon of El-Matbouli et al., 1992, Aurantiactinomyxon of Benajiba & Marques, 1993, and Aurantiactinomyxon of Székely et al., 1998 supposedly infect more than a single naidid species (Table 1), while Aurantiactinomyxon pavinsis and the Aurantiactinomyxon type of McGeorge et al., 1997 have been reported from both naidids and lumbriculids ( Table 1). Considering that these reports are not backed-up by molecular data, Rocha et al. (2019c) suggested that aurantiactinomyxon might be host specific, further proposing that actinospores of new isolates be identified through a comprehensive morphological and biological comparison to known types sharing the same annelid host.
Individual prevalence of infection of aurantiactinomyxon types is typically low, ranging from 0.01% to 1.5% in wild environments, and from 0.26% to 4.6% in surveys performed from fish farms, though there is evidence of significant spatial and temporal variations (see El-Mansy et al., 1998a,b;Ö zer et al., 2002b;Eszterbauer et al., 2006) that probably reflect host genetics, proximity, and habitat preferences, as well as abiotic factors (see Alexander et al., 2015 and references therein). Higher prevalence of infection has been reported when considering the number of infected individuals within only a specific host species, rather than in relation to the annelids population that was sampled (see Székely et al., 2000;Negredo & Mulcahy, 2001), or when pooling all aurantiactinomyxon types occurring in a single annelid species to determine the prevalence of infection of the collective group in a specific sampling site (see El-Mansy et al., 1998a,b Lin et al., 1999;Rosser et al., 2014), so that molecular information can only be found for the myxosporean stage. In turn, no sequence is available for the myxosporean stage of T. hovorkai, which accounts for two distinct actinospore stage sequences in GenBank. Anderson et al. (2000) reported a single 710 bp 18S rDNA sequence (AJ133419) obtained from both myxosporean and actinosporean stages of T. hovorkai. Actinospores were retrieved from infections in B. sowerbyi and were identified by the authors as belonging to the Aurantiactinomyxon type 2 of Yokoyama et al., 1993, previously reported to be the life cycle counterpart of T. hovorkai based on experimental transmission (see Yokoyama et al., 1993;Yokoyama, 1997;Székely et al., 1998). Later, Eszterbauer et al. (2006) obtained two similarly sized sequences (DQ231153 with 817 bp and DQ231154 with 785 bp) from aurantiactinomyxon actinospores in B. sowerbyi that were reported to match unpublished sequences of T. hovorkai obtained by the authors during a previous experimental infection study, though being morphologically and genetically different from the Aurantiactinomyxon type 2 of Yokoyama et al., 1993(see Yokoyama, 1997Yokoyama et al., 1993;Székely et al., 1998).
Presently, both these aurantiactinomyxon types remain identified as life cycle counterparts of T. hovorkai, being included as such in Table 1. A more comprehensive and clear understanding of the diversity of this collective group is necessary to help clarify important interactions with annelid hosts and involvement in myxozoan life cycles. The description of novel types and re-description of known types that remain without molecular data, namely those comprising reports from several hosts, will surely contribute towards this aim. Thus far, molecular-based studies are limited by the paucity of available data but have shown that morphologically similar aurantiactinomyxon actinospores may be distantly related (Rocha et al., 2019c), in the same manner that morphologically different actinospores can share the same genotype (see Hallett et al., 2002;Eszterbauer et al., 2006;Zhao et al., 2016). Consequently, the combined analysis of biological, morphological, and molecular criteria is imperative for performing reliable type identification (Rocha et al., 2019c). This task is significantly hampered by the difficulty in obtaining earlier reports, and due to imprecision and confusion of information in the literature.
In this study, a comprehensive summary of the biological characters and morphometry of all types described within the aurantiactinomyxon group and former guyenotia is provided as an important tool for researchers working in this field. Sixty-six types were counted, with data from original descriptions and subsequent reports. Aurantiactinomyxon eiseniellae Ormières & Frézil, 1969 was not included in the count, as Marques (1984) transferred this type to the neoactinomyxum collective group. Morphometric characters include actinospore body length and width, length and width of valvular processes, length and width of polar capsules, and number of secondary cells. Number of coils of polar tubules was not included, given that this information is available only for the Aurantiactinomyxon of Székely et al., 1998 (3-4), Aurantiactinomyxon of Xiao & Desser, 1998 (3-4), Aurantiactinomyxon of Rocha et al., 2019c (4-5), and the Guyenotia of Xiao & Desser, 1998 (3-4) (Borzák et al., 2021;Rocha et al., 2019c;Székely et al., 1998;Xiao & Desser, 1998). Information on host, locality and availability of molecular data is also provided.