Parasite load effects on sex ratio, size, survival and mating fitness of Heleidomermis magnapapula in Culicoides sonorensis

The C. sonorensis were a colony (AA strain) kindly provided by the USDA Arthropod-Borne Animal Diseases Laboratory in Laramie, Wyoming. The H. magnapapula were started by nematodes emerging from naturally parasitized C. sonorensis that were collected from dairy wastewater ponds in western Riverside County, California (Mullens and Velten, 1994). For routine maintenance, H. magnapapula was maintained in its midge host, C. sonorensis. The colonies were reared at 28 °C at a 13L:11D photoperiod in enamel pans (42-cm-long, 28-cm-wide and 5-cm-deep) of nutrient-rich water in the laboratory (Fig. 1A) (Mullens and Velten, 1994; Luhring and Mullens, 1997). The pans contained two completely saturated polyester batting pads (3-cm thick) with room around the pads’ perimeter for water circulation stimulated by air bubblers. Pads extended to the water surface, were densely fibrous, and provided a substrate for a rich microbial community (especially bacteria, algae, and fungi) to provide food for the host larvae.

Figure 1:

Culicoides sonorensis has four larval instars (L1, L2, L3 and L4). Host larvae were exposed to the nematodes at 7 d (late L2 and early L3) by placing 16 fully gravid female mermithids (just beginning to produce live J2) onto the batting surface in each pan. This resulted in approximately 14,400 J2 to attack the 3,000–4,000 host larvae/pan (Luhring and Mullens, 1997). The late L4 hosts were sieved from the pans 5 to 9 days later and were held in 9-cm glass Petri dishes of dechlorinated tap water for adult nematode emergence. While the L4 larvae were in the Petri dishes, the bacterial-feeding nematode Panagrellus redivivus was offered as food (Mullens and Velten, 1994). Adult mermithids, particularly females, sometimes could be seen even with the naked eye through the midge host cuticle before emergence (Fig. 1B). After nematode emergence, nematodes were separated from remaining L4 hosts using an 80-mesh sieve, which only the nematodes could crawl through. We also reared C. sonorensis in non-infected pans using the same larval rearing methods, but with no exposure to mermithids.

A small-scale version of the above rearing conditions was used. Six-day-old C. sonorensis larvae (late L2 and early L3) were harvested from the routine non-infected rearing pans using a fine mesh sieve. Approximately 250 host larvae were placed into each of several 9-cm glass Petri dishes containing a disc of polyester batting and 50 ml of nutrient-rich rearing pan water from non-infested host pans. A single fully gravid H. magnapapula female (just beginning to produce live J2) was added to a dish. Based on her length, which is positively correlated with fecundity (Mullens and Velten, 1994), the number of J2 she held could be estimated: in general, an average female produced about 900 J2 (Luhring and Mullens, 1997). This yielded an average parasite load of about 3.5 nematodes per midge host and about 75% parasitism (Luhring and Mullens, 1997). The range of parasite loads was substantial, however, and allowed us to select midge hosts with a large range of numbers of J2, as detailed below.

After exposure to the J2 for approximately 24 h, hosts again were sieved from the Petri dishes. Individual host midge larvae were placed on a microscope slide in a drop of dechlorinated water and a cover slip was placed on top. The amount of water was adjusted as needed. A tissue was used to wick water from under the cover slip or a pipette was used to add small amounts of water to the cover slip edge. This method, after some practice, could gently press the midge larva between the slide and cover slip, restricting its movements without causing injury. The live host larva on the slide was placed under a phase contrast microscope and examined at 40–100×, where J2 could be seen through the translucent host cuticle (Mullens and Luhring, 1997). This allowed us to generate hosts with known nematode parasite loads for further study. The sex of host Culicoides larvae cannot presently be determined morphologically.

Midge larvae with known nematode loads, including unparasitized controls, were reared individually in 24-well plastic sorting trays with 16-mm-diam wells (Fig. 1D). Each well had a piece of folded filter paper, and nutrient-rich pan water (as described above) was added and then replaced every 2–3 days from uninfested routine rearing pans. The filter paper provided a substrate for the larva to occupy, a site for pupation, and also served to support growth of microorganisms to mimic the conditions of the regular rearing pans.

Wells were checked daily to see if both the host larva and any unemerged nematodes still inside died, if midges pupated, or if adult nematode parasites emerged (which kills the host insect). If adult nematode parasites successfully emerged, their number and sex were recorded. If the host midge larvae died prior to emergence of parasites, the hosts were dissected to determine the number and sex (if sufficiently developed) of unemerged mermithids inside the host body. The dissections were done by first cutting off the midge larva's head at the base of the head capsule. The midge body contents then were extruded in a drop of water on a glass microscope slide by gently holding the posterior tip of the host body using fine forceps, and then using a second pair of forceps to compress the body partially, sliding the second forceps toward the anterior. Adult male and female nematodes are easily differentiated: males are thinner, shorter and more translucent, lack a vagina, and have a distinctly curved caudal tip and spicules (Poinar and Mullens, 1987) (Fig. 1C). Nematodes and body contents easily came out through the open anterior end of the body cavity.

Host midge larvae (L4) that had been exposed in the routine parasite production pans were harvested as usual (5–9 days after exposure to J2). Those hosts were then held individually in dechlorinated tap water in 96-well plastic ELISA plates with 7-mm-diam wells. They were checked daily until host death, pupation, or nematode parasite emergence. Number and sex of emerging nematodes were recorded for each midge host yielding parasites. After adult nematode emergence (and invariably the death of the host), remains of the host midge larva and emerged nematodes were placed into hot (60 to 70 °C) water. These measures denatured proteins and helped prevent any changes in nematode shape, while causing them to straighten out, making measurements (using an ocular micrometer) easier. Nematodes and midge host remains were placed into a vial with 2% formalin and 3% glycerin, which kept them in good condition until measurements could be taken. These were done using a wet mount on a glass slide, with specimens under a cover slip and fluid level maintained to avoid deforming the specimen. The host larval head capsule was then measured at its widest point. Nematode width for males was measured at two points: the narrowest at the posterior end near the spicule and the widest at the widest point near the anterior end. Females were measured at the posterior end of the ovary and at the widest point in the middle, posterior to the vagina.

The two width measurements were averaged to yield an estimate of width (diameter), and the volume of the nematode thus was estimated using the formula for the volume of a cylinder (π r²h). If more than five males emerged from a single host, five representative males were selected and measured for the data set. Formal randomized choice of the selected males was not done, but males from the same host were generally the same size and similar in appearance.

Parasitized L4 hosts were held individually, as described above, in wells of a 96-well plastic plate, and were checked daily for mermithid emergence. In the rare cases where both sexes emerged from the same host, all of those nematodes were excluded from mating experiments. The nematodes used for these mating fitness studies thus were virgins, having had no prior access to the opposite sex.

Each nematode was then placed together with one member of the opposite sex in a well of a 24- well sorting tray (16-mm-diam wells). Males were always less than 24 h old when paired with a potential mate. Virgin females were less abundant, and thus were up to 48 h old when paired. Upon introduction to the wells, the sexes were gently moved into contact with each other using a small metal dental pick, to encourage mating. Typically, the nematodes immediately coiled around each other as soon as they made contact. Following being placed together, the nematodes were held together at room temperature (24 °C). Survival and egg hatch of nematodes were checked daily (sometimes only one day on weekends) for seven days. Presence of J2 (evidence of successful insemination) was recorded; the typical period for H. magnapapula embryogenesis at that temperature is about 3.5 days (Mullens et al., 1995). Dead nematodes were recorded as well; female nematodes typically die within 1–2 days of giving birth if inseminated (Mullens and Velten, 1994). Seven days after pairing, nematodes were killed in hot water and preserved. Both males and females were measured for length and average width, as described above.

If H. magnapapula emerge successfully, they begin to do so at 10 to 11 days after infection (Mullens and Velten, 1994). Table 1 shows the number and time of death of parasitized hosts (and unparasitized controls) that died without either the midge host pupating or adult parasites emerging. This is termed premature death, in which both the insect and nematodes inside it die. A weighted mean time of death was calculated first for each parasite load category. The weighted means were not analyzed statistically, but provided a useful overview of when premature death occurred.

Only five unparasitized host larvae (controls) died prematurely, with a weighted mean time of death of 10.6 days after the J2 exposures in infected hosts. This was too few to analyze statistically. Among parasitized midge hosts, the time of premature death tended to decrease with parasite load and was tested using chi-square (χ²) analysis. Based on our observations on numbers of mermithids emerging from parasitized midges collected in the past in the field (Paine and Mullens, 1994; Mullens and Luhring, 1998), a load of 1 was common and considered “light”, a load of 2 to 3 was also common and was considered “moderate”, and a load of 4+ was uncommon and considered “heavy.”

In order to place sufficient numbers of premature deaths in categories to facilitate χ² analysis (traditionally ≥ 5/cell) (McDonald, 2014), deaths were pooled into load categories of 1, 2 to 3, and 4+, that were matched against four time categories in a 3 × 4 contingency table. The χ² value was 13.84 (p < 0.05). So, as parasite loads for parasitized hosts increased, time of premature host death decreased significantly.

Fig. 2 shows the fate of midge hosts and nematodes, based on the visual assessment of how many J2 had initially penetrated the late L2 or early L3 host larva. Control midges had 90% pupation and 10% died before pupation. In marked contrast, 12.2% of parasitized midges (all loads) pupated and 87.8% of parasitized hosts (all loads) died, either prematurely or when nematodes emerged successfully from the host larva. The cell frequencies (parasitized vs control, pupated or died) were compared using a 2 × 2 χ² test. Parasitism reduced host survival very significantly χ² = 156.07, p < 0.001).

Figure 2:

Among parasitized midge larvae, in which at least one J2 was observed immediately after exposure, linear regression analysis was conducted to test whether parasite load (as the independent variable) influenced each of the three parameters of interest as a dependent variable (i.e., whether the slope of the regression differed from zero). Sample sizes per parasite load category ranged from 11 to 95 (Fig. 2). In the first regression, the percentage of midges yielding some nematodes that emerged successfully from the larval host ranged from 57.9% (load = 2) to 29.7% (load = 10+) and was not influenced by parasite load (t = −1.17, p = 0.28).

The other two dependent variables were affected substantially by parasite load. The percentage of midges pupating was reduced at higher nematode parasite loads (t = 4.83, p < 0.01). For hosts where a single J2 was observed immediately after exposure, 27.4% pupated. In general, for loads of ≥ 4, pupation was below 11.5%, and no hosts with ≥ 10 parasites survived to pupate. The percentage of hosts dying prematurely, without nematodes emerging, increased with parasite load (t = 3.51, p < 0.01). For loads ≥ 8, premature death occurred in the majority of hosts (Fig. 2).

Comparing visual estimates of parasite load with midge dissection data, the initial visual assessments of J2 per host usually were the same or more than subsequent nematode counts derived from midge dissection after death or live nematode emergence (in 92% of cases). Some larval hosts, however, yielded slightly more parasites than had been seen initially through the host cuticle by visual microscopic observation. That is, we had not seen every J2, and the visual estimates had been slightly conservative. In 13 of 158 determinations (8%), more nematodes were present by emergence or dissection than had been counted by visual examination through the host cuticle immediately after exposure. This source of error (undercounting) was evenly scattered through the range of nematode parasite loads (five cases with loads of 1 to 3, four cases with loads of 4 to 6, and four cases with loads > 7). The visual estimates were accurate predictors of dissected J2 numbers overall, and results were very similar (compare Table 1 using dissection of dead hosts and Fig. 2 using visual load estimates).

Looking more closely only at midge hosts from which live nematodes emerged, it was possible to compare initial visual load estimates with the number of live parasites that emerged for each parasite load level. This provides an estimate of nematode survival as a function of load (Fig. 3). In this calculation, for midges containing a single nematode (J2 counted 24 h after host exposure), successful emergence constituted 100% survival, although even single parasites caused some increase in premature host mortality (Fig. 2). For hosts that were initially judged to contain multiple nematode parasites, nematode parasite survival estimates ranged from 93% (for an estimated initial load of two) to 75 to 80% (for initial loads ≥ 8). Thus, there appeared to be a slight decline in survival as a function of initial nematode density.

Figure 3:

To test this, where we had adequate numbers of mermithids for analysis, a χ² analysis (Table 2) was performed to examine number of nematodes alive versus number missing and presumed dead in designated host-load categories. Categories were pooled to yield a better sample size (> 10 in this case) per group (Table 2). For hosts harboring multiple J2 initially, load did not significantly impact eventual successful emergence of adult H. magnapapula (p > 0.4).

When midge hosts were penetrated by a single J2 (load observed immediately after exposure) that emerged successfully, 83.7% were female (n = 31 out of 37) and 16.3% (6 out of 37) were male. For individual loads of 2 to 9 nematodes per host, the large majority (97% of hosts) yielded exclusively males.

The few exceptions are listed as follows: 1) one midge host with an initial load of two yielded a single female nematode (one died sometime during development); 2) one midge host with an initial load of three yielded one male and one female (and one died); and 3) one host with an initial load of six nematodes yielded two males and one female (and three nematodes died). In subsequent fertility testing, only single females emerging from a host, or males emerging with other males from a host (known virgins), were used.

One possible influence on nematode development and survival is variation in host size. In the present system, our host size measure was larval head capsule width near the base. Head width for the four host larval instars, taken from uninfected UCR colony material, is shown in Fig. 4. Following Dyar's law (Dyar, 1890), dimensions of relatively sclerotized body parts (such as head capsules) in insects usually separate into discrete groupings by instar, and this is true for larvae of C. sonorensis, including the UCR colony (Barnard and Jones, 1980; Abubekerov and Mullens, 2018). The L4 head capsule sizes typically ranged from about 150–170 μ.

Figure 4:

Midge host larvae cannot be sexed based on external morphology, but the pupae can (Shults et al., 2016). Based on head capsule width measurements from cast larval skins and differences in the shape of the tip of the resulting pupal abdomen between males and females, female midge hosts (n = 50) had average L4 head widths of 164 ± 1 μ, while males (n = 40) had average L4 head widths of 161 ± 2 μ and did not differ significantly using a two-sample t test (t = 1.19; p > 0.2). It was not possible to successfully separate the host sexes based on larval head capsule width.

Examining a plot of midge host size (L4 head width) versus nematode parasite load, host size did appear to have a positive influence on H. magnapapula across the range of parasite loads. Hosts yielding a single live adult H. magnapapula averaged about 125 μ in head width, while the few hosts yielding 9–10 mermithid adults averaged about 170 μ (Fig. 5).

Figure 5:

Increasing parasite loads had a substantial negative impact on average size of the emerging males (Fig. 6A). For midge hosts yielding a single live nematode male, those males averaged about 0.016 mm³ in volume. In contrast, average male volume was less than half of this number (< 0.008 mm³) for hosts yielding ≥ 7 males. A multiple regression was conducted with average individual male volume as the dependent variable, and the independent variables of host size (head capsule width) and the parasite load as independent variables. This was done for the entire range of parasite loads observed. In a general linear model (GLM), the most important factor influencing average individual male volume was load (a negative influence) (F = 8.89; p < 0.0001), followed by host size (a positive influence) (F = 10.61; p < 0.01), and the interaction of load × size was not significant (F = 1.68; p > 0.1).

Figure 6:

The increase of nematode parasite loads varied with total volume of nematodes emerging per midge host (Fig. 6B). It is important to keep in mind that it was the larger host larvae that tended to support successful emergence of larger numbers of nematodes (Fig. 5). That said, single successful male emergences yielded a nematode volume of 0.02 mm³, but ≥ 7 males per host, while smaller as individuals, produced a cumulative nematode volume of > 0.05 mm³ per host. A multiple regression was conducted with total male volume as the dependent variable, with independent variables of host size (head capsule width) and parasite load. This was done for the entire range of parasite loads observed. In this generalized linear model (GLM), the positive effect of load was most significant (F = 23.6; p < 0.0001), followed by a very significant positive effect of host size (F = 13.45; p < 0.001), and the load × size interaction (F = 3.49; p < 0.01).

The frequency distribution for the range of nematode sizes used in the individual mating pair experiment is shown in Fig. 7. Males ranged in size from 0.0025 to 0.0334 mm³ (a 13.4 × range). Females ranged in size from 0.0121 to 0.1110 mm³ (a 9.2 × range) (Table 3).

Figure 7:

Over all ranges of respective sizes in the pairing of individual virgin males and individual virgin females, 118 of 313 pairs (37.7%) resulted in insemination, as assessed by egg hatch. Internal hatching of eggs within the nematode female's body was almost always either 0% or 100%. There were two exceptions: one individual female had about 25% of the egg complement hatch, and another had about 75% of its complement hatch. Thus, 116 out of 118 (98.3%) inseminated females had all their eggs hatch.

A preliminary Kruskal-Wallis test was utilized first as a rough indicator of the effect of size on fertility. This was done by dividing the male or female nematodes into two groups based on whether their eggs hatched or not. Males that successfully mated had a 35% larger median volume (0.0131 mm³) versus males that had not mated successfully (0.0097 mm³) (H = 15.11, df =1, p < 0.001). Fertile females, in contrast, actually were 12% smaller (median 0.0398 mm³) than infertile females (median 0.0451 mm³) (H = 4.22, df = 1, p = 0.04). However, even the very smallest nematode individuals of each sex were among the pairs that were fertile, so there seemed to be no strict lower limit on mating success. Looking separately and across the range of mate sizes at the 50 smallest nematode males (< 0.005 mm), nine mated successfully (18%). For the 50 smallest nematode females (< 0.025 mm³), 18 mated successfully (36%). The smallest fertile male was 0.0037 mm³, while the smallest fertile female was 0.0147 mm³.

But this separate analysis of the sexes did not account for the likely simultaneous and interacting influence of the size of the mate on fertility. For this more comprehensive analysis, the H. magnapapula were sorted into size classes based on volume, with five categories for males and four for females (Table 3).

Logistic regression analysis was conducted, with female fertility (positive or negative) as a variable dependent on male and female size categories (volume) and the number of males per host. Results are shown in Table 4. Parameter estimates for female nematode volume (χ² = −0.005; p > 0.1) and number of male nematodes per host (χ² = −0.008; p > 0.1) were negative, but were not statistically significant. On the other hand, male nematode size (volume) had a very large positive effect on female nematode fertility (χ² = 0.034; p < 0.001). Larger nematode males were much better able to inseminate females, while all sizes of females were still able to be inseminated.

Male nematode survival in water (all pairs) is shown in Fig. 8. A GLM analysis (ANOVA) utilized the five male size categories as the independent variable and the number of days each survived as the dependent variable. Large males lived longer (F = 9.85; p < 0.001). The smaller males (size categories 1 and 2) lived an average of about 4.5 to 5 days, while larger males (size categories 3 to 5) lived an average of 6 to 6.5 days. Female nematode survival was not assessed, since females die about 3.5 days after mating, after producing J2 (Mullens and Velten, 1994). Unmated females can survive in water up to two weeks at 23 °C (Mullens and Velten, 1994), but our study did not follow females for that long.

Figure 8: