Phylogenetic relationships, expanded diversity and distribution of Crassiphiala (Digenea, Diplostomidae), an agent of black spot disease in fish

Abstract

Crassiphiala is a monotypic genus of diplostomid digeneans, and is the type-genus of the subfamily Crassiphialinae. The type-species Crassiphiala bulboglossa parasitizes kingfishers in the Nearctic and has a Neascus-type metacercaria that encysts on fish intermediate hosts, often causing black spot disease. While recent molecular phylogenetic studies included some members of the Crassiphialinae, no DNA sequence data of Crassiphiala is currently available. Our molecular and morphological study of adult and larval crassiphialines from the Americas revealed the presence of at least 3 lineages of Crassiphiala from the Nearctic and 2 lineages from the Neotropics. This is the first record of Crassiphiala from the Neotropics. Herein, we provide the first molecular phylogeny of the Diplostomoidea that includes Crassiphiala. Our data revealed 0.2–2.4% divergence among 28S sequences and 11–19.8% among CO1 sequences of lineages of Crassiphiala. The results of our analyses did not support the monophyly of Crassiphialinae. Our results clearly demonstrated that the diversity of Crassiphiala has been underestimated.

Introduction

Crassiphiala Van Haitsma, 1925 (Diplostomidae: Crassiphialinae) is a monotypic genus of diplostomid digeneans parasitic in kingfishers (Alecedinidae Rafinesque) (Dubois 1968). The type-species Crassiphiala bulboglossa Van Haitsma, 1925 was described from the intestine of the belted kingfisher Megaceryle alcyon (Linnaeus) from Michigan, USA (Van Haitsma 1925) and since then reported only in the Nearctic (Preble and Harwood 1944; Dubois and Rausch 1948; Hoffman 1956; Dubois 1969; Boyd and Fry 1971; Scott 1984; Niewiadomska 2002; Muzzall et al 2011). The life cycle of C. bulboglossa is similar to members of the genus Uvulifer Yamaguti, 1934 and includes planorbid snails and fishes as intermediate hosts (Hoffman 1956). Notably, C. bulboglossa has a Neascus-type metacercaria that normally encysts in fish skin and is often melanized by the fish host. This infection is often referred to as the “black spot disease” (Hunter 1933; Hoffman 1956; Niewiadomska 2002; McAllister et al. 2013). Adult Crassiphiala are characterized, among other features, by a large holdfast organ, rudimentary or absent ventral sucker and the absence of an ejaculatory pouch (Niewiadomska 2002).

Crassiphiala is the type-genus of the subfamily Crassiphialinae Sudarikov, 1960. The most recent revision of the Crassiphialinae by Niewiadomska (2002) included 15 genera; however, adults of only 4 of these genera have been included in prior molecular phylogenetic analyses based on the 28S rDNA gene. The molecular phylogenetic studies that included more than 3 genera of the Crassiphialinae have shown mixed support for the monophyly of the subfamily (see Blasco-Costa and Locke 2017; Hernández-Mena et al. 2017; López-Jiménez et al. 2017; Achatz et al. 2019). Despite very weak support or the lack of support evident from their phylogenetic trees, some authors have repeatedly suggested that this subfamily may warrant elevation to family (Blasco-Costa and Locke 2017; Hernández-Mena et al. 2017; Locke et al. 2018). Moreover, no prior molecular phylogenetic study included representatives of Crassiphiala, the type genus of the Crassiphialinae. The purpose of this study is the demonstration of the phylogenetic placement of Crassiphiala using DNA sequence data for the first time as well as the presence of this genus in South America. Formal morphological and taxonomic descriptions of the genetic lineages presented herein will be published separately.

Materials and Methods

We obtained adult specimens of Crassiphiala from intestines of M. alcyon collected in Clearwater (one specimen) and St. Louis (one specimen) Counties in Minnesota, USA (collecting permit MB072162–0) and a single ringed kingfisher Megaceryle torquata (Linnaeus) collected in Pantanal, Fazenda Retiro Novo, Municipality of Poconé, Mato Grosso State, Brazil (collecting permit 10698 approved by the Instituto Chico Mendes de Conservação da Biodiversidade), using corresponding federal and state collecting permits. In addition, metacercaria of Crassiphiala were collected from the skin and fins of yellow perch Perca flavescens Mitchill from Cass County, Minnesota USA, central mudminnow Umbra limi Kirtland from Hubbard County, Minnesota and Phoxinus eos Cope from St. Louis County, Minnesota, USA. Live digeneans removed from the hosts were briefly rinsed in saline, killed with hot 10% ethanol and preserved in 70% ethanol. One sample was obtained from a frozen carcass of M. alcyon that died after flying into a glass window. These specimens were directly fixed in 70% ethanol. In total, we obtained 5 adult specimens of Crassiphiala lineage 1 from M. alcyon collected in the St. Louis Co., MN; a single adult specimen of Crassiphiala lineage 2 from M. alcyon collected in the Clearwater Co., MN; one metacercaria of Crassiphiala lineage 3 from P. eos in St. Louis County; several dozen adult specimens of Crassiphiala lineage 4 and 52 specimens of Crassiphiala lineage 5 from M. torquata in Pantanal.

Specimens for light microscopy were stained with aqueous alum carmine according to Lutz et al. (2017) and studied using a DIC-equipped Olympus BX51 compound microscope (Tokyo, Japan) with a digital imaging system. Vouchers of adult specimens of genetic lineages 1, 4 and 5 are deposited in the collection of the Harold W. Manter Laboratory (HWML), University of Nebraska State Museum, Lincoln, NE, U.S.A. (Table 1). The lineage 2 was represented in our material by a single adult specimen which was used for DNA extract and the lineage 3 was a metacercaria sequenced as a part of an unrelated study.

Table 1.

List of Crassiphiala samples used in the phylogenetic analyses of 28S rDNA and COI mtDNA including sample size (n), Harold W. Manter Laboratory (HWML) voucher numbers, their host species, geographical origin of material and GenBank accession numbers.

Digenean taxa	Host species	Geographic origin	Museum No.	Accession numbers
				28S	COI
Crassiphiala lineage 1 (n=2)	Megaceryle alcyon	USA	HWML-XXXX	MN193951, MN193952	XXXX
Crassiphiala lineage 2 (n=1)	M. alcyon	USA	-	MN193953	XXXX
Crassiphiala lineage 2 (n=1)	Phoxinus eos	USA	-	MN193954	XXXX
Crassiphiala lineage 2 (n=2)	Umbra limi	USA	-	MN193955	XXXX, XXXX
Crassiphiala lineage 3 (n=1)	Perca flavescens	USA	-	MN193956	XXXX
Crassiphiala lineage 4 (n=3)	Megaceryle torquata	Brazil	HWML-XXXX	MN193957–MN193959	XXXX, XXXX
Crassiphiala lineage 5 (n=2)	M. torquata	Brazil	HWML-XXXX	MN193960	XXXX, XXXX

Genomic DNA was extracted according to the protocol described by Tkach and Pawlowski (1999). A fragment of the 5’ end of 28S was amplified by polymerase chain reactions (PCR) as described in Tkach et al. (2003). A fragment of the cytochrome c oxidase (CO1) gene was amplified using the previously published forward primer Cox1_Schist_5’ and reverse primers acox650R and JB5 (Lockye r et al. 2003 3; Derycke et al. 2005; Kudlai et al. 2015). In some cases, CO1 was amplified in 2 overlapping fragments using a combination of published primers and new internal primers designed for this study by TJA. The forward primer Cox1_Schist_5’ was used with new reverse primer BS_CO1_IntR (5’–TAA TAC GAC TCA CTA TAA AAA AAA MAM AGA AGA RAA MAC MGT AGT AAT–3’); the new forward primer BS_CO1_IntF (5’–ATT AAC CCT CAC TAA ATG ATT TTT TTY TTT YTR ATG CC–3’) was used with the reverse primer acox650R. The underlined portions indicate a shortened T3 and T7 tail sequences. PCRs of CO1 were performed in a total volume of 25 μl using New England Biolabs (Ipswich, MA, USA) One-Taq quick load PCR mix according to the manufacturer’s protocol. The thermocycling protocol for CO1 was as follows: 30 sec denaturation hold at 94 °C; 40 cycles of 25 sec at 94 °C, 30 sec at 45 °C, 1 min at 68 °C; and 5 min extension hold at 68 °C.

PCR products were purified using ExoSap PCR clean-up enzymatic kit from Affymetrix (Santa Clara, CA, USA) following the manufacturer’s protocol. PCR products were cycle-sequenced directly using BrightDye® Terminator Cycle Sequencing Kit (MCLAB, California, USA) chemistry, alcohol precipitated and run on an ABI 3130 automated capillary sequencer (Thermo Fisher Scientific, Waltham, MA, USA).

Sequencing reactions of 28S were carried out as described in Achatz et al. (2019). The PCR primers were used for sequencing of CO1 PCR reactions. In addition, the shortened T3 tail (5’–ATT AAC CCT CAC TAA A–3’) and shortened T7 tail (5’–TAA TAC GAC TCA CTA TA–3’) primers from Van Steenkiste et al. (2015) were used for sequencing of the PCR reactions prepared with BS_CO1_IntF and BS_CO1_IntR primers. Contiguous sequences were assembled using Sequencher version 4.2 software (GeneCodes Corp., Ann Arbor, Michigan, USA). Newly generated sequences are deposited in the GenBank (Table 1).

Phylogenetic relationships of Crassiphiala were analyzed using 28S and CO1 datasets as separate alignments. Three CO1 sequences (two from the lineage 2 and one from the lineage 4) were much shorter than the rest and therefore not included in the alignment, although they were submitted to the GenBank. These shorter sequences were identical to their longer counterparts.

Newly obtained and previously published sequences were aligned using ClustalW implemented in Mega7 (Kumar et al. 2016); both alignments were trimmed to the length of the shortest sequence. The cyathocotylid Suchocyathocotyle crocodili (Yamaguti, 1954) (GenBank accession MK650450) was selected as outgroup in the 28S analysis based on the topology presented by Achatz et al. (2019). Uvulifer sp. (GenBank accession MF124281; Blasco-Costa and Locke, 2017) was selected as outgroup in the CO1 analysis based on the results of our 28S analysis and genetic distances.

The 28S alignment included newly generated sequences of 6 taxa of Crassiphiala and previously published sequences of 18 members of the Diplostomidae Poirier, 1886, 1 member of the Proterodiplostomidae Dubois, 1936 and 12 members of the Strigeidae Railliet, 1919. The CO1 alignment included newly generated sequences of 7 taxa of Crassiphiala. Phylogenetic analyses were conducted as described by Achatz et al. (2019). The trees were visualized in FigTree ver. 1.4 software (Rambaut 2016) and annotated in Adobe Illustrator®.

Results

Upon trimming to the length of the shortest sequence the 28S alignment was 1,137 bp long. Similar to the results of several previous studies (e.g. Blasco-Costa and Locke 2017; Hernández-Mena et al. 2017; Achatz et al. 2019) the phylogenetic tree resulting from the BI analysis demonstrated the non-monophyletic nature of the Diplostomidae and Strigeidae. Members of the Diplostomidae formed 6 clades: 1) Crassiphialinae clade 1 which included Crassiphiala + Uvulifer (100%), 2) Crassiphialinae clade 2 which included Bolbophorus Dubois, 1935 + Ornithodiplostomum Dubois, 1936 + Posthodiplostomum Dubois, 1936 (88%), 3) Hysteromorpha Lutz, 1931, 4) Austrodiplostomum Szidat & Nani, 1951 + Diplostomum von Nordmann, 1832 + Tylodelphys Diesing, 1850 (100%), 5) Neodiplostomum Railliet, 1919 and 6) Alaria Schrank, 1788 (99%). All sequenced lineages of Crassiphiala formed a strongly supported clade (98%); Crassiphiala lineage 4 from Brazil formed a sister group to all other Crassiphiala isolates (Fig. 1). Members of the Strigeidae formed 2 strongly supported clades. The first clade (100%) included Apharyngostrigea Ciurea, 1927 + Strigea Abildgaard, 1790 + Apatemon Szidat, 1928+ Australapatemon Sudarikov, 1959; the second clade (98%) included Cardiocephaloides Sudarikov, 1959 + Cotylurus Szidat, 1928 + Ichthyocotylurus Odening, 1969.

Fig. 1.

Phylogenetic interrelationships among 38 diplostomoidean taxa including Crassiphiala based on Bayesian Inference (BI) analysis of partial 28S rRNA gene sequences. Members of the subfamily Crassiphialinae as currently recognized are indicated by the shaded rectangles. Bayesian Inference posterior probability values lower than 80% are not shown. New sequences obtained in this study are marked by an asterisk. Scale bar indicates number of substitutions per site. Abbreviations: BR – Brazil, USA – United States of America.

The internal interrelationships among available isolates of Crassiphiala were studied using the 392 bp long CO1 alignment. While the overall topology in the CO1 tree was similar to that of the Crassiphiala clade in the 28S tree, the much more variable CO1 sequences provided added resolution in form of the well-supported cluster of Crassiphiala lineages 3 and 5 which was unresolved in the 28S tree. Crassiphiala lineage 4 formed a sister group with all other Crassiphiala isolates, although with a somewhat lower support than in the 28S gene tree (Fig. 2).

Fig. 2.

Phylogenetic interrelationships among Crassiphiala lineages based on Bayesian Inference (BI) analysis of partial CO1 mtDNA sequences. Bayesian Inference posterior probability values lower than 80% are not shown. All sequences of Crassiphiala included in our analysis are new and marked by an asterisk. Scale bar indicates number of substitutions per site. Abbreviations: BR – Brazil, USA – United States of America.

Pairwise nucleotide comparisons of 28S sequences among all unique Crassiphiala isolates are provided in Table 2. The divergence in 28S sequences of Crassiphiala lineages was generally low (0.2–2.4%). One of the 28S sequences of Crassiphiala lineage 4 (MN193957) had a single mixed base (adenine or guanine), while the 2 other isolates of Crassiphiala lineage 4 (MN193958, MN193959) had only guanine in this position. No other variation within lineages was detected in sequences of 28S.

Table 2.

Pairwise comparisons of partial sequences of the 28S rRNA gene between lineages of Crassiphiala included in this study. Percentage differences are given above diagonal and the number of variable nucleotide positions is given below the diagonal. The 28S results are based on a 1,132 bp long alignment.

	Cr. 1 MN193952	Cr. 2 MN193953	Cr.. 3 MN193956	Cr. 4 MN193959	Cr. 5 MN193960
Crassiphiala lineage 1 MN193952 USA	-	0.9%	1.1%	2.2%	1.1%
Crassiphiala lineage 2 MN193953USA	10	-	0.2%	2.2%	0.2%
Crassiphiala lineage 3 MN193956 USA	12	2	-	2.4%	0.4%
Crassiphiala lineage 4 MN193959 Brazil	25	25	27	-	2.4%
Crassiphiala lineage 5 MN193960 Brazil	12	2	4	27	-

Pairwise nucleotide comparisons of CO1 sequences among all unique Crassiphiala isolates are provided in Table 3. The CO1 sequences showed a much greater divergence among lineages (11–19.8%) than in 28S. There were no differences between the CO1 sequences of Crassiphiala lineage 4.

Table 3.

Pairwise comparisons of partial sequences of the COI mtDNA gene between lineages of Crassiphiala included in this study. Percentage differences are given above diagonal and the number of variable nucleotide positions is given below the diagonal. Results are based on a 435 bp long alignment.

	Cr. 1 XXXX	Cr. 2 XXXX	Cr. 2 XXXX	Cr. 3 XXXX	Cr. 4 XXXX	Cr. 5 XXXX	Cr. 5. XXXX
Crassiphiala lineage 1 XXXX USA	-	12.4%	12.2%	16.3%	15.6%	14.3%	14.3%
Crassiphiala lineage 2 XXXX USA	54	-	0.2%	13.1%	16.1%	11.3%	11.5%
Crassiphiala lineage 2 XXXX USA	53	1	-	12.9%	15.9%	11.0%	11.3%
Crassiphiala lineage 3 XXXX USA	71	57	56	-	19.8%	12.9%	13.1%
Crassiphiala lineage 4 XXXX Brazil	68	70	69	86	-	17.0%	17.2%
Crassiphiala lineage 5 XXXX Brazil	62	49	48	56	74	-	0.7%
Crassiphiala lineage 5 XXXX Brazil	62	50	49	57	75	3	-

Discussion

The morphology of the adult Crassiphiala specimens included in our study conforms closely to the diagnosis of the genus provided by Niewiadomska (2002) (Fig. 3).

Fig. 3.

Adult specimens of Crassiphiala from Megaceryle torquata from Pantanal, Brazil. (a) Ventral view of whole mount of Crassiphiala lineage 4. (b) Ventral view of the prosoma of Crassiphiala lineage 4. (c) Lateral view of posterior body end of Crassiphiala lineage 4. (d) Ventral view of the prosoma of Crassiphiala lineage 5. (e) Ventral view of whole mount of Crassiphiala lineage 5. Scale bars: (a) = 500 μm; (b) = 100 μm; (c) = 100 μm; (d) = 100 μm; (e) = 300 μm.

While some authors (Blasco-Costa and Locke 2017; Hernández-Mena et al. 2017; Locke et al. 2018) have noted that the Crassiphialinae may be elevated to the family level, our analysis of 28S did not support the monophyly of the Crassiphialinae with 2 clades comprising members of this subfamily being branches of a large polytomy. This is in concordance with the phylogenetic data and branch support of corresponding clades in some of the recent works, e. g., Blasco-Costa and Locke (2017) and Hernández-Mena et al. (2017) which reported a low level of support for the Crassiphialinae. Although the authors of those publications suggested the monophyly of the Crassiphialinae, their phylogenetic trees did not provide a sufficient evidence for such conclusions. The content and systematics of the Crassiphialinae need to be carefully re-evaluated based on a detailed morphological study and additional phylogenetic analyses of its constituent taxa which is outside of the scope of this work. Based on the data obtained in the present work, particularly the demonstrated non-monophyly of the Crassiphialinae, we do not see a sufficient ground for elevating its status to the family level. Moreover, the content of the Crassiphialinae as currently recognized (Niewiadomska 2002) needs to be revised; likely, only Crassiphiala and Uvulifer should remain in the subfamily. However, a more detailed analysis involving a greater diversity of crassiphialine taxa and a thorough morphological study is needed to adequately address this question.

Despite the relatively low number of Crassiphiala lineages, our analyses (Figs. 1, 2) allowed for an interesting observation that the phylogenetic relationships within the genus do not follow the geographic origin of the samples. One of the lineages from Brazil (Crassiphiala lineage 4) appeared on the trees as the sister group to all other members of the genus while the branch that included the other Brazilian lineage (Crassiphiala lineage 5) was nested among North American isolates. The reasons for this pattern are not clear at this time. One explanation may be the relatively old evolutionary origin of Crassiphiala which allowed for transcontinental spread (in both directions). Another explanation could be based on the partial overlap of the geographic distribution of the typically North American M. alcyon with several species of kingfishers broadly distributed in the Central and South America.

This study is the first to generate DNA sequence data of adult specimens of Crassiphiala and the first to report Crassiphiala in the Neotropics. Our results demonstrated the presence of at least 5 lineages of Crassiphiala in the Nearctic and Neotropics. This indicates that the diversity of Crassiphiala was seriously underestimated and allows us to hypothesize that additional species belonging to this genus are likely to be discovered in future studies. Central and South America hold a greater potential in this respect due to the more diverse fauna of kingfishers.