Broad infectivity of Leidynema appendiculatum (Nematoda: Oxyurida: Thelastomatidae) parasite of the smokybrown cockroach Periplaneta fuliginosa (Blattodea: Blattidae)

Abstract

Host specificity of parasites is important for the understanding of evolutionary strategies of parasitism that would be a basis of predictions of the disease expansion when parasitized hosts invade new environments. The nematode order Oxyurida is an interesting parasite group for studying the evolution of parasitism as it includes parasites of both invertebrates and vertebrates. In our survey, we found that the smokybrown cockroach Periplaneta fuliginosa was primarily infected with only one nematode species Leidynema appendiculatum. In two cases, L. appendiculatum was isolated from two additional cockroach species Pycnoscelus surinamensis, sold in Japan as a reptile food, and Blatta lateralis, captured in the field and cultured in the laboratory. Inoculation of L. appendiculatum into three additional cockroach species P. japonica, Blattella nipponica, and P. surinamensis also resulted in parasitism. Infection prevalence was high, and timing of postembryonic development from hatched nematode larva to mature adult in these hosts was identical with that in P. fuliginosa. While ecological interactions strongly determine the host range, such broad infectivity is still possible in this parasitic nematode.

INTRODUCTION

The host specificity of parasites is important as it reflects the evolutionary strategy of parasitism. The nematode order Oxyurida consists of a number of families that are divided into two superfamilies. The first superfamily Oxyuroidea, which consists of vertebrate parasitic nematodes (also called pinworms that include the human parasite Enterobius vermicularis), and second Thelastomatoidea, which consists of invertebrate parasitic nematodes (Carreno, ; Nadler et al., ). Nematodes belonging to the order Oxyurida share a simple infection cycle: Nematode eggs laid by adult females are deposited within their host feces and released by defecation, next ingestion of eggs by new host individuals leads to infection (Adamson, ; Ozawa et al., ). While Oxyurida parasites sometimes cause diseases in humans, such as colitis, perianal abscess, and ectopic infections in females (Ariyarathenam et al., ; St Georgiev, ), these conditions are not regarded as serious compared to infectious diseases caused by another gastrointestinal worms including Onchocerciasis, Elephantiasis, and Ancylostomiasis (Bungiro & Cappello, ; Pilotte, Unnasch, & Williams, ). Moreover, Oxyurida parasites in invertebrates are seemingly harmless to their hosts (Adamson, ; Ozawa et al., ), but consequences of carriage are largely unknown. Cospeciation is well conserved in vertebrates and pinworms (Falk & Perkins, ; Hugo, ; Murata, Hasegawa, Nakano, Noda, & Yanai, ). However, a wide variety of nematode species have been reported from diverse invertebrate hosts and that do not seemingly reflect host/parasite cospeciation events (Adamson & van Waerebeke, ; Jex, Schneider, & Cribb, ; Jex, Schneider, Rose, & Cribb, ). The evolutionary strategies of parasitism might differ between Oxyuroidea and Thelastomatoidea.

The family Thelastomatidae have been reported for more than forty Blattodea species (Adamson & van Waerebeke, ; Ozawa et al., , ; Sriwati, Ozawa, Morffe, & Hasegawa, ). We reported earlier for the first time that the smokybrown cockroach Periplaneta fuliginosa in Japan was infected with only one nematode species Leidynema appendiculatum with high prevalence (Ozawa et al., ). However, L. appendiculatum is reported as a cosmopolitan nematode species, which has been isolated from many Blattaria hosts known as sanitary pests, including P. americana (American cockroach), P. australasiae (Australian cockroach), and Blatta orientalis (Oriental cockroach) (Adamson & Noble, ; Blanco, Lax, Dueñas, Gardenal, & Doucet, ; Connor & Adamson, ; Shah, ). As genotypic characterizations have not been conducted, it is still doubtful whether these nematodes were L. appendiculatum or cryptic species.

Here, we show the broad infectivity of L. appendiculatum through natural and artificial infection of five cockroach species within three families and two suborders. L. appendiculatum was basically isolated from P. fuliginosa collected in three area of Japan with high prevalence. Infection prevalence, intensity, and infrapopulation of L. appendiculatum in five host cockroach species were similar. Moreover, the developmental timing of this nematode from hatched larvae to matured adult was identical. While ecological interactions with P. fuliginosa seemingly strongly determine the host range of L. appendiculatum, such broad infectivity is still possible in this parasitic nematode.

MATERIALS AND METHODS

Cockroach strains and rearing

The following laboratory‐culturing cockroach strains (seven strains from five species in three genus, two subfamily) used in this experiment were reared as described by Ozawa et al., , (1) Periplaneta fuliginosa EE, (2) P. fuliginosa UF, (3) P. japonica Miyoshi, (4) Pycnoscelus surinamensis Pet, (5) P. surinamensis Yaedake, (6) Blattella nipponica CU, (7) Blatta lateralis KX.

We used two independent strains of P. fuliginosa: P. fuliginosa EE strain was supplied by the Earth Environmental Service (Ako, Hyogo Prefecture, Japan) and reared since 2012. The P. fuliginosa UF strain, also laboratory‐culturing strain, has been established and reared for more than 60 years in the Philip G. Koehler laboratory, University of Florida. P. japonica, the Japanese cockroach or Yamato cockroach, is a native of Japan and belonging to the same Periplaneta genus with P. fuliginosa. The P. japonica Miyoshi strain was established from 10 adult males and 10 females collected at the Miyoshi city (Aichi Prefecture, Japan) in June 2012. The P. surinamensis Pet strain, sold for pet reptile food, was obtained by Japanese online shop (no detailed information obtained). The P. surinamensis Yaedake strain was established from three adult females hand‐picked at the Yaedake (Okinawa prefecture, Japan) in March 2013. B. nipponica is also a native of Japan. The B. nipponica CU was collected at the Chubu University campus (Kasugai, Japan) during April to September 2014 and 2017; because this species was impossible to be reared in laboratory for over a year, all individuals were collected and reared only temporary before experiments. The B. lateralis KX strain was established from three adult females collected at Knoxville (Tennessee, USA) in July 2013.

Nematode‐free cockroach strains were established as follows: oothecae oviposited were collected, surface wiped with 70% ETOH, and kept in a plastic dish until hatching. Hatched nymphs were transferred into the plastic cage and reared. P. japonica Miyoshi, B. nipponica CU, and P. surinamensis Yaedake were originally nematode‐free. Individuals of the established nematode‐free strains were often dissected to confirm the absence of parasitic nematodes.

Cockroach dissection and parasitic nematode observation

The hindgut of cockroach adult male and female was extracted, placed in a Syracuse watch glass containing cockroach Ringer's solution (NaCl 9.32 g, KCl 0.77 g, NaHCO₃ 0.18 g, NaH₂PO₄ 0.01 g, CaCl₂ 0.5 g in 1 L of distilled water), and split longitudinally with tweezers to release nematodes. Nematode species, sex, and stage were observed under a stereomicroscope (SMZ600, Nikon, Japan). When detailed morphological information was needed, nematodes were picked up with a mouth pipette, transferred onto an agar pad (Shaham, ), covered, and sealed with a silicon grease‐rimmed coverslip for viewing by Nomarski DIC optics (Eclipse E600, Nikon).

Because we found that four cockroach strains (P. fuliginosa EE, P. fuliginosa UF, P. surinamensis Pet, and B. lateralis KX) were originally infected with L. appendiculatum, we analyzed the infrapopulation of nematode male, female, and juvenile by the software Quantitative Parasitology 3.0 (Rózsa, Reiczigel, & Majoros, ).

Molecular identification and phylogenetic analysis

Genomic DNA from a single female nematode was extracted using Qiagen DNeasy Blood & Tissue Kit (Qiagen, USA) following manufacturer's instructions. The D2/D3 expansion segment of 28S ribosomal RNA gene (D2/D3 LSU) and partial fragment of small subunit of ribosomal RNA gene (SSU) were amplified using the universal primers D2A (5′‐ACA AGT ACC GTG AGG GAA AGT TG‐3′) and D3B (5′‐TCG GAA GGA ACC AGC TAC TA‐3′) (Nunn, ), and nSSU_F_07 (5′‐AAA GAT TAA GCC ATG CAT G‐3′) and nSSU_R_26 (5′‐CAT TCT TGG CAA ATG CTT TCG‐3′) (Blaxter et al., , The Blaxter Lab website: http://www.nematodes.org/research/barcoding/sourhope/nemoprimers.shtml). The amplified DNA fragment was purified from agarose gels with NucleoSpin^® Gel and PCR Clean‐up (MACHEREY‐NAGEL, Germany). Samples were submitted to Hokkaido System Science Co., Sapporo, Japan, for sequencing from both strands, using the same PCR primers. Sequences were deposited in GenBank NCBI (http://www.ncbi.nlm.nih.gov/genbank/).

For the phylogenetic analysis, several sequences from Thelastomatidae were selected: L. appendiculatum JQ343844, EU365630, and KC540759. L. portentosae GQ401114 and EF180073 were selected as outgroup. ClustalW multiple alignment was conducted in BioEdit version 7.2.6 (Hall, ), and sequence alignments were trimmed automatically by trimAI with default setting (Capella‐Gutiérrez, Silla‐Martínez, & Gabaldón, ). Phylogenetic trees were constructed from evolutionary distances using the maximum likelihood (ML) method using the Mega 6.0 software (Tamura, Stecher, Peterson, Filipski, & Kumar, ). Based model was D2/D3 LSU: Hasegawa‐Kishino‐Yano model (Hasegawa, Kishino, & Yano, ); SSU: Jukes‐Cantor model (Jukes & Cantor, ). Phylogenetic robustness was inferred by bootstrap analysis using 1,000 iterations (Felsenstein, ).

Pairwise comparisons of % differences (D) between each sequence combination were performed using the formula D = (M/L) × 100 (Chilton, Gasser, & Beveridge, ), where M is the number of alignment positions at which the two sequences have a base in common, and L is the total number of alignment positions.

Artificial infection test of parasitic nematodes

Artificial infection experiments were performed in four nematode‐free cockroach strains, P. fuliginosa EE, P. japonica Miyoshi, P. surinamensis Yaedake, B. nipponica CU. L. appendiculatum adult females were collected from nematode‐infected P. fuliginosa EE as described above. The transferred L. appendiculatum gravid females were placed in a Syracuse watch glass containing cockroach Ringer's solution and cut in the center by sterilized scalpel blade to release eggs (~150 eggs per individual). Released eggs were kept in the cockroach Ringer's solution for 14 days at 25°C until all reached to the L2 resting stage. The last instar of nematode‐free cockroaches was fasted for 3 days before artificial infection. Ten fasted cockroaches were reared in a cylindrical plastic case (130 mm diameter × 225 mm height) with 0.4 g of the bait mixed with about 500 of the L2 resting stage nematode eggs. After artificial infection, cockroaches were fed with bait without nematode eggs ad libitum. To observe the nematode prevalence (% of the infected cockroaches among all cockroaches examined), intensity (mean number of nematodes in the infected cockroaches, excluding the number “zero” of uninfected host), and developmental stages, one or two cockroaches of each group from P. fuliginosa EE were dissected 1, 3, 7, 14, 21, 28, and 35 days after infection, or one or two cockroaches of each group from P. japonica, B. nipponica, and P. surinamensis were dissected every week (from 7 days until 35 days after infection). All nematodes isolated in these experiments were mounted on the agar pad (described above) and observed by Nomarski DIC optics (Eclipse E600, Nikon).

After experimental infection, the rest of P. japonica was pooled and reared for several generations, and then checked the infrapopulation of L. appendiculatum male, female, and juvenile by the Quantitative Parasitology 3.0 (Rózsa et al., ).

RESULTS

Two exceptional cases of L. appendiculatum hosts

We dissected 14 species of cockroaches collected in the field, and as well laboratory‐cultured strains. Basically L. appendiculatum were isolated only from P. fuliginosa in our experiments. We always isolated only L. appendiculatum from laboratory‐cultured (Ozawa et al., ) and field‐captured (Table S1) P. fuliginosa with high prevalence. In this study, we showed that the two different laboratory strains of P. fuliginosa EE (Japan) and UF (USA) were infected with L. appendiculatum with 100% prevalence (Table ). As these two P. fuliginosa strains were established independently and maintained for a long period in different countries and laboratories, we conclude that this parasitic association of L. appendiculatum with P. fuliginosa is quite stable.

Infrapopulation of Leidynema appendiculatum in five cockroach strains

Host cockroach	L. appendiculatum infection data
Strain and sex1	N	Body size2	Stage	Prevalence (Cl)3	Mean intensity (Cl)4	Median intensity (CI)5	V/M ratio6
P.f. EE Male	10	25.7 ± 1.4	♂	50.0 (22.3–77.8)	1.00 (uncertain)	1.0 (1–1)	0.56
♀	70.0 (38.1–91.3)	2.43 (1.43–4.57)	2.0 (1–7)	2.62
J	100 (70.9–100)	7.60 (3.40–15.30)	4.0 (1–20)	11.73
P.f. EE Female	15	26.6 ± 1.8	♂	86.7 (60.3–97.6)	1.15 (1.00–1.46)	1.0 (1–1)	0.43
♀	100 (77.8–100)	9.20 (5.6–19.00)	8.0 (3–12)	14.57
J	93.3 (69.8–99.7)	15.43 (10.71–19.79)	15.0 (8–24)	5.94
P.f. UF Male	20	27.2 ± 1.7	♂	70.0 (47.5–86.0)	1.14 (1.00–1.43)	1.0 (1–1)	0.61
♀	70.0 (47.5–86.0)	2.14 (1.43–2.93)	1.5 (1–3)	1.72
J	50.0 (29.3–70.7)	2.70 (1.90–3.80)	2.0 (2–5)	2.28
P.f. UF Female	22	29.4 ± 1.9	♂	81.8 (61.1–93.5)	1.06 (1.00–1.17)	1.0 (1–1)	0.25
♀	90.9 (70.9–98.3)	4.45 (3.65–5.25)	4.0 (3–5)	1.24
J	45.5 (20.8–71.8)	3.60 (2.20–5.10)	3.0 (1–6)	3.06
P.s. Pet Female	10	18.5 ± 2.5	♂	90.0 (55.4–99.5)	1.00 (uncertain)	1.0 (1–1)	0.11
♀	100 (70.9–100)	3.50 (2.70–4.30)	3.0 (2–5)	0.59
J	100 (70.9–100)	18.90 (12.40–28.80)	13.5 (7–39)	10.05
B.l. KX Male	18	21.8 ± 1.4	♂	88.9 (67.0–98.0)	1.00 (uncertain)	1.0 (1–1)	0.12
♀	94.4 (72.9–99.7)	2.65 (2.06–3.12)	3.0 (2–3)	0.67
J	55.6 (33.0–76.4)	3.70 (2.30–5.60)	2.0 (2–8)	3.86
B.l. KX Female	19	24.3 ± 1.5	♂	78.9 (55.4–92.5)	1.00 (uncertain)	1.0 (1–1)	0.22
♀	100 (82.5–100)	4.47 (3.37–5.74)	4.0 (3–5)	1.75
J	94.7 (67.2–99.9)	11.39 (7.83–16.22)	9.0 (3–13)	8.24
P.j. AI Male	22	24.9 ± 2.1	♂	45.5 (26.1–66.2)	1.00 (uncertain)	1.0 (1–1)	0.57
♀	59.1 (38.3–77.8)	2.23 (1.62–2.92)	2.0 (1–3)	1.62
J	72.7 (50.0–87.4)	4.19 (2.69–5.94)	3.5 (1–5)	3.86
P.j. AI Female	20	24.8 ± 2.9	♂	40.0 (20.1–62.8)	1.00 (uncertain)	1.0 (1–1)	0.63
♀	65.0 (42.4–83.3)	2.15 (1.46–2.85)	1.0 (1–4)	1.68
J	75.0 (52.6–89.6)	8.27 (5.27–11.47)	6.0 (3–14)	7.14

¹P.f. EE, P. fuliginosa EE strain. P.f. UF, P. fuliginosa UF strain. P.s. Pet, P. surinamensis Pet strain. B.l. KX, B. lateralis KX strain. P.j. AI, P. japonica artificial infection strain. All individuals examined in this experiment were adult and 100% nematode infection prevalence except P.j. AI female (91%).

²Host cockroach body size, average ± SD (mm).

³Sterner's exact method, confidence limits for the population prevalence (95% Confidence limits).

⁴Bootstrap (BCa) method, confidence limits for the mean intensity (95% Confidence limits), except for adult male nematode from P.f. EE Male (93.8% confidence level).

⁵Exact confidence limits for the median intensity (95% Confidence limits).

⁶Variance to mean ratio is calculated including uninfected hosts as well.

Interestingly, we also found two rare cases in the cockroach species P. surinamensis and B. lateralis. We established three strains of P. surinamensis; the first strain P. surinamensis Pailand was infected with unknown nematode species Suifunema sp. with almost 100% of prevalence (data not shown). The second strain, P. surinamensis Yaedake, was free from nematode, but the third P. surinamensis Pet was infected with L. appendiculatum with 100% prevalence. We established one strain of B. lateralis KX and found that was also infected with L. appendiculatum with 100% prevalence. We confirmed that these nematodes were L. appendiculatum from morphological characteristics of the adult male and female (data not shown).

Molecular identification and phylogenetic analysis of L. appendiculatum

We obtained ribosomal DNA sequence data, D2/D3 LSU and SSU from L. appendiculatum isolated from cockroach strains used in these experiments including P. fuliginosa CU (Ozawa et al., ), P. fuliginosa EE, P. fuliginosa UF, P. surinamensis Pet, and B. lateralis KX. We also obtained sequence data for L. appendiculatum Tokyo 20140825SB, isolated from P. fuliginosa captured in Minato city, Tokyo in 2014 (Table S1). These data were deposited in the NCBI GenBank (accession numbers, Table ). Several published sequences of Leidynema spp. were included in this phylogenetic analysis. All SSU sequences of L. appendiculatum were 100% identical (Figure a), whereas D2/D3 LSU sequences showed unique polymorphisms that divided into two clades. One clade contained L. appendiculatum isolated from cockroaches in Japan. The other clade included cockroaches from the United States, Russia, and Argentina (Figure b). The hosts of Japanese L. appendiculatum are P. fuliginosa EE, P. fuliginosa CU, P. fuliginosa Tokyo 140825SB, and P. surinamensis Pet. The host species in other countries are P. fuliginosa UF (USA), B. lateralis KX (USA), P. americana (Russia), and P. americana (Argentina) (Table ). Pairwise sequence differences between the two clades were 1.5%, although 0.0% within the clades (Table ).

Leidynema strains information used in phylogenetic analysis

Nematode	Host cockroach	Original Country	D2/D3 Accession No.	SSU Accession No.	References
L. appendiculatum EE	Periplaneta fuliginosa EE	Japan	KY057026	KY057032	This study
L. appendiculatum UF	P. fuliginosa UF	USA	KY057030	KY057034	This study
L. appendiculatum Pet	Pycnoscelus surinamensis pet	Japan	KY057029	KY057033	This study
L. appendiculatum KX	Blatta lateralis KX	USA	KY057027	KY057031	This study
L. appendiculatum Tokyo	P. fuliginosa Tokyo	Japan	KY057028	KY057036	This study
L. appendiculatum CU	P. fuliginosa CU	Japan	KC540759	KY057035	Ozawa et al. ()
L. appendiculatum	P. americana	Argentina	JQ343844	–	Blanco et al. ()
L. appendiculatum	P. americana	Russia	EU365630	–	Spiridonov, direct submission
L. portentosae	Gromphadorhina portentosa	–	GQ401114	EF180073	Nadler et al. ()
					Spiridonov et al. (2009)

Maximum likelihood (ML) trees inferred from (a) 18S SSU rDNA and (b) D2D3 LSU rDNA for Leidynema appendiculatum isolated from cockroach hosts. Leidynema portentosae was used as outgroup. Values at the nodes correspond to bootstrap probability

Pairwise differences (%) in the D2/D3 sequence between nine samples of the genus Leidynema (Leidy, 1850) Chitwood, 1932

	1	2	3	4	5	6	7	8	9
1. L. appendiculatum CU (Japan) KC540759	–
2. L. appendiculatum EE (Japan) KY057026	0.0	–
3. L. appendiculatum Pet (Japan) KY057029	0.0	0.0	–
4. L. appendiculatum Tokyo (Japan) KY057028	0.0	0.0	0.0	–
5. L. appendiculatum UF (USA) KY057030	1.5	1.5	1.5	1.5	–
6. L. appendiculatum KX (USA) KY057027	1.5	1.5	1.5	1.5	0.0	–
7. L. appendiculatum (Russia) EU365630	1.5	1.5	1.5	1.5	0.0	0.0	–
8. L. appendiculatum (Argentina) JQ343844	1.5	1.5	1.5	1.5	0.0	0.0	0.0	–
9. L. portentosae GQ401114	9.1	9.1	9.1	9.1	9.4	9.4	9.4	9	–

Infrapopulation of L. appendiculatum in different cockroach strains

We analyzed the prevalence, intensity, and infrapopulation structure of L. appendiculatum males, females, and juveniles in four cockroach strains: P. fuliginosa EE, P. fuliginosa UF, P. surinamensis Pet, and B. lateralis KX. All cockroaches in the laboratory‐cultured strains were infected with L. appendiculatum with 100% prevalence. We also confirmed the infrapopulation structures were similar in each cockroach strain. First, the distributions of female and juvenile nematodes are positive skewed (Mean > Median); second, the number of nematode males was basically one or zero; third, the mean and median intensity of nematodes in female hosts is higher than that in male hosts; and fourth, there was no difference in nematode prevalence between male and female cockroach hosts (Table ).

Artificial infection of L. appendiculatum in P. fuliginosa

Next, we inoculated L. appendiculatum into P. fuliginosa to monitor nematode prevalence and developmental timing from the L2 stage egg to mature adult. Just 1 day after infection, several nematode larvae appeared in the cockroach hindgut (Figure a). After 3 days of infection, differences of male/female morphology were not evident at these larval stages. Seven days postinfection, few nematodes started molting. After molting, morphological differences between male and female appeared. The spicule and copulatory apparatus appeared in males’ round‐shaped tails that resembles tails of adult males (Figure e). Female tail shape became long and filiform (Figure b). After 14 days of infection, nematodes started molting (Figure S1A–D) and sexual dimorphism became prominent despite lack of sexually maturity (Figure c,f). At this timing, 16 nematodes were molting and showing female character, and four were male character. The rest of the nematodes (N = 31) were still larvae and without sexual dimorphisms. At 21 days postinfection, reproduction systems of the nematodes were maturing or matured (Figure S1E–H). At 28 days after infection, males and females had reached adulthood, and eggs were visible in adult female uteri (Figure d,g). At 35 days after infection, a small number of larvae appeared in the cockroach hindgut that seemed to be a new infection with a second‐generation nematodes.

Developmental stage of L. appendiculatum. (a) Larva, 1 day after infection. (b) Female larva, after 14 days of infection. (c) Female larva, after 21 days of infection. (d) Matured adult female, after 28 days of infection. (e) Male larva, after 14 days of infection. (f) Male larva, after 21 days of infection. (g) Matured adult male, after 28 days of infection. Scale bars, (a, e) 100 μm, (b, c, f, g) 200 μm, (d) 500 μm

Figure a shows the infection prevalence of L. appendiculatum and mean intensity (mean number of nematodes in the infected cockroaches, excluding the number “zero” of uninfected host) of male, female, and juvenile nematodes in a cockroach hindgut at 1 to 35 days after infection. We counted male/female when nematodes appeared sexually dimorphic (at or after the 7th day of infection, Figure b,e) and classified as juvenile when the dimorphism was not clear (before 7th day of infection, Figure a). As indicated in Figure , several hatched vermiform larvae appeared immediately from the 1st day, and the first sexual dimorphism appeared in males on the 7th day. The mean intensity of males was two on the 7th day, but mostly did not exceed one after 21 days of infection. The average nematode prevalence over the course of the experiment (from 1st day to 35th day) was 78% (N = 76).

Nematode intensity (%) (bars, left y‐axis) and infection prevalence (lines, right y‐axis) in each cockroach host, (a) P. fuliginosa, (b) P. japonica, (c) B. nipponica, and (d) P. surinamensis

Artificial infection of L. appendiculatum in other cockroaches

Next, we inoculated L. appendiculatum into the cockroach species, P. japonica, B. nipponica, and P. surinamensis, which belong respectively to three families (Blattidae, Blattellidae, Blaberidae) and two suborders (Blattoidea, Blaberoidea). We observed the nematode number, developmental stage, and sex. Figure b–d shows the infection prevalence of L. appendiculatum (%) and mean intensity (uninfected “zero” data were excluded) of male, female, and juvenile nematodes in a cockroach hindgut at 7 to 35 days after infection.

Although average nematode prevalence over the course of the study was lower in P. japonica (65%, N = 40), B. nipponica (53%, N = 66), and P. surinamensis (51%, N = 51) when compared with that in P. fuliginosa (78%, N = 76), L. appendiculatum successfully infected all three hosts. The developmental timings were the same as that observed in P. fuliginosa (Figure b–d). The mean intensity of male, female, and juvenile nematodes in cockroach hindgut at each timing was similar, although no juvenile nematodes were observed in B. nipponica at 28 and 35 days postinfection (Figure c).

After these artificial infection experiments, the remaining P. japonica were combined and reared for several generations under the same conditions as our laboratory‐cultured strains (see Materials and Methods) and then checked for infection prevalence. The population of L. appendiculatum in adult male and female cockroaches was assessed. Infection prevalence was 91% (N = 22) in adult males and 90% (N = 20) in adult females, and infection intensity was similar to the other cockroaches in cultured in the laboratory (Table ).

DISCUSSION

We have dissected 14 cockroach species including P. fuliginosa (Smokybrown cockroach), P. japonica (Yamato cockroach), P. americana (American cockroach), P. australasiae (Australian cockroach), Blattella germanica (German cockroach), B. nipponica (Japanese forest roach), B. lituricollis, Blatta lateralis, Pycnoscelus surinamensis (Surinam cockroach), P. indicus (Indicus cockroach), Onychostylus pallidiolus, Lobopterella dimidiatipes, Opisthoplatia orientalis, Panesthia angustipennis spadica (data not shown). These species were collected in the field as well as cultured in the laboratory. Although we have not finished identification and description of all parasitic nematodes yet, host specificity of the nematodes belonging to the family Thelastomatidae was seemingly high. Species combinations of cockroach host and parasitic nematode were almost fixed except for two exceptions described in this work. We always isolated L. appendiculatum from P. fuliginosa with high prevalence. L. appendiculatum was not isolated from the other cockroaches with the two exceptions; one of our three P. surinamensis strains and one of our one B. lateralis strain (Table ). Captured B. lateralis were concluded to be originally infected with L. appendiculatum rather than to be contaminated during rearing in laboratory. More investigations are needed to understand the patterns of infection of the parasitic nematodes in B. lateralis in the field. Our finding that B. lateralis can act as a host of parasitic nematodes might be a case showing the broad host range of L. appendiculatum.

In the case of B. lateralis, we only checked one strain for these experiments. However, we did check more than ten individuals of wild P. surinamensis as well as laboratory‐cultured strains. All parasitic nematodes isolated from P. surinamensis were unknown species Suifunema sp. (data not shown) except for one laboratory‐cultured strain P. surinamensis Pet (Table ). As this cockroach reproduces parthenogenetically, grows, and reproduces quickly, it is easy to maintain without any special condition. Therefore in Japan, P. surinamensis is popular for pet reptile breeders as a food of their animals. As P. fuliginosa is the most widely distributed cockroach in Japan, there are likely many chances for contamination with the parasitic nematode from P. fuliginosa if the cockroaches are cultured in environments of general households or pet shops.

We present here that L. appendiculatum was capable of infecting three cockroach species, P. japonica, B. nipponica, and P. surinamensis. The infectivity of L. appendiculatum is quite broad and includes five cockroach species, three families in two suborders (Figure ). P. japonica is a Japanese domestic species and lives mainly in northern area moving between indoors to outdoors (Tanaka & Tanaka, ; Tanaka & Uemura, ). P. surinamensis is estimated to have originated in the Indo‐Malayan region and is now distributed worldwide in tropical and subtropical regions (Kramer & Brenner, ). We found P. japonica and P. surinamensis were infected with specific nematode Protrellus sp. and Suifunema sp., respectively, with high prevalence (data not shown). B. nipponica is a Japanese domestic cockroach and morphologically resembles B. germanica, the world sanitary pest, however, with differences in ecological traits; B. nipponica mainly lives in the outdoors under fallen leaves in forests and grassland of southwestern Japan (Asahina, ; Tsuji, ). Interestingly, B. nipponica is always free from parasitic nematodes.

Phylogenetic relation of five cockroach species examined in these experiments, constructed based on Inward, Beccaloni, & Eggleton (). These species belong to three families (Blattidae, Blattellidae, Blaberidae) and two suborders (Blattoidea, Blaberoidea). Blattodea was believed to separate with Mantodea from common ancestor between 250 and 200 mya. Two suborders, Blattoidea and Blaberoidea, were thought to separate between 200 and 150 mya (Misof et al. ). American cockroach P. americana another major sanitary pest and reported as a host of L. appendiculatum also belongs to the family Blattidae (Grimaldi & Engel, )

We also confirmed that L. appendiculatum, when inoculated artificially into B. nipponica could persist in the host hindgut and lay eggs but could not be re‐infected. This is likely because B. nipponica does not eat feces. An infection cycle for parasitic nematodes might not be established. This result matches with our data that B. nipponica captured in the field is not infected with parasitic nematodes at all. Therefore, we conclude that L. appendiculatum is not a native parasite of B. nipponica. It was surprising that L. appendiculatum can infect a cockroach species that has a different ecological and systematic position than P. fuliginosa.

Although infection prevalence of L. appendiculatum within these three cockroaches was less than that in P. fuliginosa, values exceeded 50% and the developmental timing of nematode maturation resembled that seen in P. fuliginosa (Figure ). We believe that there was no difference in infection opportunities as all cockroaches consumed all of the given bait within a few days. These results might suggest that L. appendiculatum has evolved as a parasite of P. fuliginosa, but that its infection mechanism makes it capable of infecting other cockroach species accidentally when exposed.

The remarkable features of L. appendiculatum we could reveal in these experiments are (1) it has a broad infectivity and (2) it has a strong association with a cockroach species that is currently spreading and inhabiting worldwide. P. fuliginosa inhabits all prefectures of Japan, although mainly central and western Japan (Nakano, ; The Japanese Society of Pestology, ; Tsuji, ). As the cockroaches examined in this experiment originally live in different habitats, there would be few opportunities to replace their original parasite in nature. The field survey revealed host sharing of thelastomatid nematodes and suggested that host specificity is dictated largely not by their taxonomy but by the host ecology (Jex et al., ,b, ). As the habitat of thelastomatid nematodes is absolutely limited to the host cockroach intestine, parasite and host have cospeciated over a long historical period to form a tight association. On the other hand, this nematode group has a simple parasitic ecology; eggs are discharged together with the host feces and infect new host individuals directly by oral infection without intermediate hosts. Because of such simple mechanisms of nematode infections, interspecific host exchange might occur more frequently in nature than we thought and that might have promoted its broad infectivity. As we confirmed that nematode‐free P. japonica was infected with L. appendiculatum when cultured together with L. appendiculatum‐infected P. fuliginosa (data not shown), nematode exchange between species could be established by feeding feces of other cockroach species. Recently, surveys of parasitic nematodes in native and invasive cockroaches in Galápagos Islands were reported. L. appendiculatum was isolated from invasive species P. americana, P. australasiae, and P. surinamensis with relatively low prevalence (14%, 31%, and 6%, respectively; Sinnott, Carreno, & Herrera, ). Cephalobellus ovumglutinosus, first described from worldwide distributed species B. germanica (van Waerebeke, ), was isolated from endemic and invasive cockroaches (Sinnott et al., ). As there is no record of P. fuliginosa in these islands, these cockroaches can be also considered nematode carriers contributing to their spreading.

Adamson proposed that ancestors of oxyurid nematodes were diversifying in invertebrates long before the parasites of vertebrates, and that they probably arose in diplopods or blattarians (Adamson, , ). Cospeciation of host and parasite was well conserved in primates and pinworm nematodes (Hugo, ). Host specificity influences the expansion of parasitic diseases when the host invades a new environment. Thus, understanding host range might contribute valuable information (Agosta, Janz, & Brooks, ). For example, the human pinworm Enterobius vermicularis only causes itching in the anal area of its native human host, but can be fatal in chimpanzees Pan troglodytes (Murata et al., ). In addition, parasitic nematodes greatly affect the composition of microbial communities in cockroaches. It is obvious that there are direct and indirect interactions between hosts and parasites (Vicente, Ozawa, & Hasegawa, ). This nematode is also a good model system for understanding the mechanisms and evolution of mutualistic relationships between parasites and hosts.

ACKNOWLEDGMENTS

We thank Professor Philip G. Koehler (University of Florida, USA) for providing cockroach strains; Professor Tsutomu Tanabe (Kumamoto University, Japan) and Mr. Kenichi Nakano (Minato city, Japan) for their help of cockroach sampling; Assistant Professor Nathan Weyand (Ohio University, USA) and Lecturer and PI Ryoji Shinya (Meiji University, Japan), Dr. Cláudia S.L. Vicente (University of Évora, Portugal) for reviewing this manuscript. This work was supported by the research fund from the Institute of the Biological Function, Chubu University (to KH), and JSPS KAKENHI Grant number 15J09961 (to SO).

CONFLICT OF INTEREST

None Declared.

AUTHOR CONTRIBUTION

KH and SO planned experiments; SO and KH performed experiments; KH and SO analyzed data; KH contributed reagents and material; KH and SO wrote the study. All authors read and approved the final manuscript.