MORPHOLOGICAL AND MOLECULAR CHARACTERIZATION OF HEPATOZOON SPECIES INFECTING FROGS AND SNAKES ACROSS THE CENTRAL AND EASTERN UNITED STATES
ABSTRACT
The genus Hepatozoon Miller, 1908 (Apicomplexa: Adeleorina) includes vector-borne, intracellular blood parasites that infect a wide range of vertebrate hosts, including frogs and snakes. Species identification of bloodstream forms is difficult because there are usually not many morphological characters to distinguish species and gamonts of genetically different isolates. Traditionally, Hepatozoon spp. have been distinguished by characters such as pathology to host erythrocytes and/or developmental stages in the invertebrate vector. However, recent molecular studies are finding that these distinctions do not correlate with gene sequence data. Specifically, this is the case for 2 closely related Hepatozoon species that infect North American anurans, Hepatozoon catesbianae and Hepatozoon clamatae. Gamonts infecting erythrocytes of these 2 species are morphologically indistinguishable, and traditionally they have been differentiated based on whether they fragment the host erythrocyte nucleus; however, recent genetic studies indicate that this character does not correlate with specific genotypes. In this study, we sampled frogs and snakes from the east central United States for Hepatozoon spp. and compared their effects on the host erythrocytes with genotype. Hepatozoon spp. infections were morphologically characterized with microscopy and molecularly characterized with Sanger sequencing at 3 loci (COIII, ITS-1, and 18S rDNA). We found that individuals of 3 ranid species (Rana catesbeiana, Rana clamitans, and Rana sphenocephala) were infected with Hepatozoon species. Of those, only individuals of R. clamitans were infected with Hepatozoon spp. that fragmented the host erythrocyte nuclei. As previously reported, mixed infections, determined with both microscope observation of fragmentation characters and Sanger sequencing, were common and obfuscated species identification and the usefulness of the fragmentation character in differentiating H. catesbianae and H. clamatae infecting North American anurans. We recommend that future studies aim to establish better definite links between cellular fragmentation characteristics and DNA sequences to differentiate these 2 species. We also report and characterize Hepatozoon cf. sipedon from 3 snake species. Infected erythrocytes in all 3 snake species displayed variation in the extent of cytoplasm clearing. Sequences from these 3 snakes were identical at ITS-1 and 18S rDNA (COIII was sequenced from only 1 isolate). In our 18S rDNA phylogeny, Hepatozoon spp. infecting frogs were in a single clade, whereas Hepatozoon spp. infecting snakes were found in multiple clades with Hepatozoon spp. that infect other hosts, including lizards, small mammals, and frogs. This study adds to a growing number of studies that indicate snakes are capturing Hepatozoon spp. from their prey, and we discuss the implications of these host captures for the life cycle evolution of Hepatozoon spp. infecting snakes.
Photomicrographs of Hepatozoon spp. cells. (A) Hepatozoon sp. gamont infecting the frog Rana sphenocephala. Note that the nucleus of the host red blood cell is not fragmented. (B) Hepatozoon sp. gamont infecting the frog Rana clamitans. Note the fragmented nucleus of the host red blood cell. (C) Hepatozoon sp. immature gamont/merozoite infecting R. sphenocephala. (D) Hepatozoon sp. infecting the snake Thamnophis proximus with cleared cytoplasm. Scale bars = 10 µm. Color version available online.
Photomicrographs of Hepatozoon spp. infecting green frogs Rana clamitans, showing the variation in the effects on the host erythrocyte nuclei. (A) The unfragmented nucleus is displaced by the gamont and slightly elongated. (B) The nucleus is fragmented into many smaller pieces, often connected by visible strands. (C) The nucleus fragmentation appears incomplete and has split into only 2 lobes. (D) Nuclei appear distorted, losing their shape and seeming to fill available cell space. Scale bar = 10 µm. Color version available online.
Photomicrographs showing variation and possible progression of cytoplasm clearing in snake erythrocytes infected with Hepatozoon cf. sipedon. (A) No cytoplasm clearing. The apparent rip of the cell is most likely an artifact of slide preparation. (B) Cytoplasm clearing is noticeable but not extensive. (C) Cytoplasm is almost fully cleared, but the cytoplasm remains opaque around the erythrocyte nucleus and gamont. (D) Cytoplasm is fully cleared, and the erythrocyte nucleus and gamont are forced very close together. A and C are from Thamnophis proximus, B is from Nerodia rhombifer, and D is from Agkistrodon piscivorus; however, examples of all degrees of cytoplasm clearing were observed in all 3 snake species. Scale bar = 10 µm. Color version available online.
Maximum likelihood phylogeny of partial cytochrome oxidase subunit III (COIII) nucleotide gene sequences (629 bp, −Ln = −2,102.34) using the GTR+G model in MEGA 11 (Tamura et al., 2021). Values at nodes represent bootstrap support values, assessed with 1,000 replications. Clades are colored according to their genetic similarity to genotypes identified by Léveillé et al. (2021), which are taxa with GenBank accession numbers. The Hepatozoon cf. sipedon COIII sequence added in this study is shown in purple. Symbols next to taxa labels indicate the effect(s) on the host red blood cell nuclei observed on blood smears from the frog which the sequence originated. Open circles represent unfragmented nuclei, and closed diamonds indicate fragmented nuclei, which here include extensively fragmented nuclei, 2-lobed nuclei, and/or distorted nuclei. Asterisks indicate sequences from Léveillé et al. (2021) that were identified to species by sequence data (not fragmentation character) and obtained from multiple frogs, some of which were single and mixed infections, yet produced identical sequences. Color version available online.
Maximum likelihood phylogeny of the internal transcribed spacer region 1 (163 bp, −Ln = −270.06) using the T92 model in MEGA 11 (Tamura et al., 2021). Values at nodes represent bootstrap support values, assessed with 1,000 replications. Genotypes of frogs are colored according to their identity at the COIII locus, following Léveillé et al. (2021; Fig. 4). Hepatozoon cf. sipedon sequences added in this study are shown in purple. Symbols next to taxa labels indicate the effect(s) on the host red blood cell nuclei observed on blood smears. Open circles represent unfragmented nuclei, and closed diamonds indicate fragmented nuclei, which here include extensively fragmented nuclei, 2-lobed nuclei, and/or distorted nuclei. Color version available online.
Maximum likelihood phylogeny of the partial 18S rDNA sequences (1,094 positions, −Ln = −5,285.96) using the HKY+G+I model in MEGA 11 (Tamura et al., 2021). Values at nodes represent bootstrap support values, assessed with 1,000 replications. Genotypes of frog Hepatozoon spp. added in this study are colored according to their identity at the COIII locus, if possible, following Léveillé et al. (2021; Fig. 4). Hepatozoon cf. sipedon sequences added in this study are shown in purple. Three clades are highlighted with brackets and icons showing their hosts; the first includes Hepatozoon spp. circulating in the blood of snakes and lizards, the second includes snakes, lizards, and mammals, and the third includes snakes and frogs. The presence of snakes Hepatozoon spp. in all 3 of these clades has interesting implications for Hepatozoon spp. life cycle evolution. Color version available online.
Contributor Notes