Between extreme simplification and ideal optimization: antennal sensilla morphology of miniaturized Megaphragma wasps (Hymenoptera: Trichogrammatidae)

Material

Adult Megaphragma amalphitanum Viggiani, 1997, Megaphragma mymaripenne Timberlake, 1924 and Megaphragma caribea Delvare, 1993 were reared from eggs of H. haemorrhoidalis. For the study of gross morphology, specimens were fixed in alcoholic Bouin; for the study of sensilla ultrastructure, specimens were fixed in 2% glutaraldehyde (GA, EMS) solution in sodium 0.1 M cacodylate buffer pH = 7.2 with subsequent postfixation in 1% osmium oxide (EMS) solution in the same buffer.

Scanning electron microscopy

The fixed material was gradually dehydrated through a series of ethyl alcohols (GA fixed material—30%, 50%, 70%, 95% ethyl alcohol, each change for 30 min, 100%—two changes for 30 min; alcoholic Bouin fixed material—same protocol starting from the 70% ethyl alcohol step) and then acetone (100%, two changes for 15 min), critical point dried (Hitachi HCP-2) and sputtered with gold (Giko IB-3). The specimens were studied and photographed using Jeol JSM-6380 with a 5 megapixel digital camera and FEI Inspect F50 with a 4 megapixel digital camera.

Morphometry

All measurements were performed on the SEM images, using measurement tool in the Fiji package of ImageJ. Normality test, descriptive statistic, ANOVA, SMA and OLS were performed using R software.

Megaphragma amalphitanum antennae

Antennae of M. amalphatanum are geniculate at the scape-pedicel joint (Fig. 1). Male and female antennae are almost identical in shape and composition, containing the following antennomeres: scape including radicula, pedicel, and flagellum consisting of 1 ringlike anellus and three flagellomeres, the 1st flagellomere forming the funicle and the 2nd and 3rd flagellomeres forming the club (sometimes this structure is also called “clava”) (Fig. 2). Thus, antenna of this species is composed of five segments with two pronounced sub-segments, the radicula and anellus. The radicula forms the base of the scape and is inserted in the head via torulus, the basal socket joint upon which the radicula is articulated and which allows antennal movement in all directions. Antennal segments are elongated and cylindrical; the last flagellomere is slightly tapered. Female antennae measure 144 ± 9 µm (mean ± sd) in length; male antennae are 136 ± 10 µm long (Table S1).

Sensilla chaetica, aporous (ChS-AP)

These sensilla are tapered to a blunt aporous tip and inserted in a cuticular socket. Their surface is longitudinally fluted and bears no pores. They are typically protruding (Fig. 3B). Three ChS-AP appear on the scape, five on pedicel and one on the 1st flagellomere in the female antenna. Overall disposition remains the same for the male specimens, except for the 1st flagellomere, where one more sensillum was observed (Fig. 2). ChS-AP are 11.5 ± 2.62 µm in length and 0.9 ± 0.16 µm in diameter in females and 9.89 ± 2.88 µm in length, 0.71 ± 0.11 µm in diameter in males (Table 1). According to the differences in sizes on different segments of the antenna, ChS-AP may be further divided into subtypes. In females, chaetica sensilla on the 1st flagellomere are significantly longer than on the pedicel and scape. By contrast, males had the longest and thickest ChS-AP on the scape, whereas difference between ChS-AP on the 1st flagellomere and pedicel was insignificant. When comparing sizes between sexes, we estimated that ChS-AP on the 1st flagellomere are significantly longer and thicker in females (Table S2).

Table 1:

Megaphragma amalphitanum, M. mymaripenne and M. caribea numbers per antenna and sizes of antennal sensilla (mean ± sd), measurements are given in µm. For additional details, see Table S3.

Sensilla type	Parameter	Megaphragma amalphitanum		M. mymaripenne	M. caribea
		Female	Male	Female	Female	Male
ChS-AP	Number	9	10	9	8	8
	Length	11.5 ± 2.62	9.89 ± 2.88	14.2 ± 2.03	8.04 ± 1.49	6.28 ± 0.99
	Diameter	0.9 ± 0.16	0.71 ± 0.11	1.3 ± 2.68	0.75 ± 0.12	0.67 ± 0.13
TS1-AP	Number	17	14	22	18	25
	Length	12.7 ± 5.42	7.63 ± 1.59	13.68 ± 4.52	8.41 ± 3.54	5.16 ± 1.29
	Diameter	0.82 ± 0.14	0.65 ± 0.09	0.85 ± 0.14	0.72 ± 0.13	0.62 ± 0.12
TS-UP	Number	2	3	2	2	3
	Length	37.1 ± 5.59	33.6 ± 4.46	29.93 ± 4.78	33.03 ± 5.97	20.42 ± 2.87
	Diameter	1.36 ± 0.23	1.34 ± 0.16	1.36 ± 0.19	1.19 ± 0.12	1.28 ± 0.2
SS	Number
	Length	7 ± 0.72	5.05 ± 0.48	5.78 ± 1.48	4.78 ± 0.83	4.58 ± 1.2
	Diameter	0.93 ± 0.11	0.76 ± 0.09	0.86 ± 0.2	0.99 ± 0.11	1 ± 0.15
TS2-AP	Number	9	9	9	9	9
	Length	0.88 ± 0.22	1.03 ± 0.36	0.93 ± 0.24	0.89 ± 0.21	0.97 ± 0.23
	Diameter	0.55 ± 0.14	0.53 ± 0.12	0.51 ± 0.07	0.57 ± 0.03	0.61 ± 0.27
BS	Number	1	–	1	1	–
	Length	17.1 ± 2.04	–	14.8 ± 2.19	11.77 ± 2	–
	Diameter	1.96 ± 0.19	–	1.85 ± 0.29	1.53 ± 0.21	–
PS	Number	2	–	2	2	–
	Length	39.9 ± 4.15	–	29.77 ± 3.84	43,324	–
	Diameter	2.48 ± 0.6	–	2.22 ± 0.51	1.87 ± 0.19	–
	Tip length	11.7 ± 4.32	–	9.17 ± 4.08	6.83 ± 1.86	–
MPS	Number	2	2	2	2	2
	Length	37.3 ± 3.58	21.9 ± 5.61	31.94 ± 3.57	25.74 ± 2.29	21.55 ± 2.5
	Diameter	2.62 ± 0.27	1.82 ± 0.38	2.45 ± 0.5	2.3 ± 0.15	1.73 ± 0.35
	Tip length	3.69 ± 1.36	3.76 ± 4.51	6.66 ± 4.77	1.8 ± 1.16	21.55 ± 2.5

DOI: 10.7717/peerj.6005/table-1

Sensilla trichodea, aporous type 1(TS1-AP)

TS1-AP are slender, with an acute tip and a smooth surface lacking grooves and pores. They arise directly from the antenna without any kind of socket or rim, and usually are directed along its surface (Fig. 4C). It is the most abundant type of sensilla in both sexes. In females, four are found on the 1st, eleven on the 2nd, and two on the 3rd flagellomeres. Males have three TS1-AP on the 1st flagellomere and eleven on the 2nd flagellomere; they lack TS1-AP on the 3rd flagellomere (Fig. 2). TS1-AP measure 12.7 ± 5.42 µm in length and 0.82 ± 0.14 µm in diameter in females and 7.63 ± 1.59 µm in length, 0.65 ± 0.09 µm in diameter in males (Table 1). In both sexes, the sensilla are significantly longer on each subsequent flagellomere (sensilla on the 1st Fl. <2nd Fl. <3d Fl. for females and 1st Fl. <2nd Fl. for males). Females demonstrated significantly thicker and longer TS1-AP on the 1st and 2nd flagellomeres than males (Table S2).

Placoid sensilla (PS)

These large, conspicuous sensilla are unique to females and occur on the 3rd flagellomere, opposing each other at its dorsal and ventral surfaces (Fig. 4B). No pores were observed even at a magnification of 130,000×. PS are fused with the antennal surface and form large detached protrusions at the tip of antenna (Fig. 2). They measure 39.9 ± 4.15 µm in length and 2.48 ± 0.6 µm in width; the protrusions are 11.7 ± 4.32 µm long (Table 1).

Megaphragma caribea antennae

This species demonstrates fused flagellomeres in both sexes, without any visible edges between the 1st and 2nd flagellomeres and retaining only remnants of them between the 2nd and 3rd flagellomeres (Fig. 5). Additionally, the pedicel of males appears to be identical to that of females. Otherwise, however, the antennae of have a similar overall shape and composition to those of M. amalphitanum (Fig. 1). The male and female antennae are 120 ± 9 µm and 120 ± 7 µm long, respectively (Table S1).

Morphological types of antennal sensilla and their distribution

There are 44 sensilla on the female antenna and 49 on the male antenna (12 studied specimens). The morphological types of sensilla are identical to those in M. amalphitanum (Fig. 6). No difference was observed between the two species in overall sensilla structure, presence of socket, or sensilla surface properties, such as grooves and pores. The numbers and distribution of antennal sensilla in are similar to those of M. amalphitanum, except as follows. The fused 1st and 2nd flagellomeres lack ChS-AP in both males and females. Females have 18 TS1-AP whereas males have 25. Both sexes have an additional SS on the distal edge of the 2nd flagellomere (Figs. 5 and 6). It appears that this sensillum is significantly shorter than SS on the 3rd flagellomere in females (Table S2). The sizes of the antennal sensilla and the results of ANOVA are provided in the tables (Table 1, Tables S2–S4).

Megaphragma mymaripenne antennae

Since M. mymaripenne is a parthenogenetic species, the data was obtained for the females of this species only. Antennae of M. mymaripenne are identical in shape and composition to those of M. amalphitanum females (Fig. 1) and are 140 ± 9 µm long (Table S1).

Antennal size and gross morphology

The overall composition of antennae in M. amalphitanum, M. caribea and M. mymaripenne is typical for most groups of parasitoid hymenopterans. All three species have three flagellomeres (fused in M. caribea males and females), as in the females of several Trichogramma species (Olson & Andow, 1993; Amornsak, Cribb & Gordh, 1998; Consoli, Kitajima & Parra, 1999; Zhang et al., 2012; Van der Woude & Smid, 2016). Most Trichogrammatidae and Mymaridae, and most larger parasitoid wasps in other families of parasitoid wasps have significantly more flagellar segments; e.g., Pteromalus puparum (Linnaeus, 1758) (Hymenoptera: Pteromalidae) has 12 flagellomeres, and M. croceipes and Cotesia marginiventris (Cresson, 1865) (Hymenoptera: Braconidae) have 16 flagellomeres (Das et al., 2011). The observed reduction in the number of flagellomeres is a common feature among miniaturized insects, known also in some of the smallest species of Ptiliidae and Corylophidae (Coleoptera) (Polilov, 2015).

The antennae of M. caribea are considerably different from those of M. mymaripenne and M. amalphitanum: in M. caribea flagellomeres are fused in both sexes (Figs. 1C, 6A–6H). In female egg parasitoids, the last flagellomeres are sometimes fused into a club (Consoli, Parra & Zucchi, 2010), e.g., in Amitus spiniferus (Brèthes, 1914) (Hymenoptera: Platygasroidea) (Isidoro, Romani & Bin, 2001). By contrast, Ooencyrtus phongi Trjapitzin, Myartseva & Kostjukov, 1977 (Hymenoptera: Encyrtidae), Metaphycus parasaissetiae Zhang and Huang, 2007 (Hymenoptera: Encyrtidae), Trichogramma dendrolimi Matsumura, 1926 and T. australicum Girault, 1912 (Hymenoptera: Trichogrammatidae) have a single club segment in males, which is separated into flagellomeres in females (Amornsak, Cribb & Gordh, 1998; Xi et al., 2011; Zhang et al., 2012; Zhou et al., 2013b). In the case of M. caribea, both sexes have a fused club, which is uncommon.

The antennal lengths of the three species are diminutive, from 120 µm in M. caribea to 144 in M. amalphitanum females (Table S1), and minute Chalcidoidea species have antennae that are 1.5–4 times longer (Amornsak, Cribb & Gordh, 1998; Van Baaren et al., 1999; Zhou et al., 2013a; Zhou et al., 2013b). Interestingly, in the shortest winged insect, the parasitoid wasp Kikiki huna Huber, 2000 (Hymenoptera: Mymaridae) (Huber & Noyes, 2013), antennal length is more than 151 µm, which makes the antennae of Megaphragma possibly the smallest functional parasitoid antennae ever described.

Number of antennal sensilla

The number of antennal sensilla significantly decreases with the body size in parasitoid wasps (Fig. 7A, Table 2). T. australicum have 3–6 times as many sensilla as the studied species (Amornsak, Cribb & Gordh, 1998), Pteromalidae species have 23 times as many as the studied species (Onagbola & Fadamiro, 2008; Dweck, 2009), and some Braconidae species have hundreds of times as many (Gao, Luo & Hammond, 2007; Xi et al., 2010; Das et al., 2011). The number of antennal sensilla recorded in the studied species is the smallest known among parasitoid wasps, except for the highly reduced male of Dicopomorpha echmepterygis (Hymenoptera: Mymaridae) (Mockford, 1997). Even the minute K. huna, the smallest winged insect (body length = 158–190 µm) appears to have about twice as many antennal sensilla as the species treated in our study, judging by SEM photographs (Huber & Noyes, 2013). Thus, it is very likely that M. amalphitanum, M. mymaripenne and M. caribea are nearing the lowest number of antennal sensilla required for a highly sensitive parasitoid antenna.

The variation in number of sensilla also seems to be reduced, the smaller the species becomes. Thus, in the three species studied the numbers of antennal sensilla are completely invariable within species (sexual dimorphism being taken into account). By contrast, in larger parasitoid wasps the numbers of sensilla vary significantly within species; e.g., in O. phongi (Xi et al., 2011) the number of trichoid sensilla of type 1 is 1,549 ± 67.9 (mean ± se, n = 6). Even in the tiny Trichogramma australicum, T., galloi Zucchi, 1988, T. pretiosum Riley, 1879, T. nubilae Ertle and Davis, 1974, and T. evanescens (Hymenoptera: Trichogrammatidae) the numbers of antennal sensilla significantly vary within species, and for several types of sensilla the correlation with the body size was shown (Olson & Andow, 1993; Amornsak, Cribb & Gordh, 1998; Consoli, Kitajima & Parra, 1999; Van der Woude & Smid, 2016).

Morphological types of antennal sensilla

The terminology used to describe insect antennal sensilla is inconsistent and confusing. For instance, there are at least four terms describing one of the antennal sensilla types in Trichogramma (Ruschioni et al., 2012), one based on the external appearance of the sensillum (“Falcate Sensilla”, Amornsak, Cribb & Gordh, 1998, “Multiporous Pitted Sensilla Trichoid C”, Olson & Andow, 1993), another on its supposed functions (“Multiporous Gustatory Sensilla”, Isidoro et al., 1996), and in one case these sensilla are designated with a letter (“Sensilla type i”, Voegelé et al., 1975). In this study, we define and name sensilla types based on the morphological characteristics of sensilla as they appear when observed under a scanning electron microscope (SEM). The comparison of the terms adopted in this study with the terminology of earlier studies on parasitoid wasps is given in Table S5.

It is common for chalcidoid wasp males to have 1–3 types of sensilla fewer than females (Barlin, Vinson & Piper, 1981; Amornsak, Cribb & Gordh, 1998; Van Baaren et al., 1999; Xi et al., 2011; Zhang et al., 2012; Zhou et al., 2013b; Namikawa & Amornsak, 2016), although in several cases the same number of types was reported for both sexes (Onagbola & Fadamiro, 2008; Jun et al., 2013). It is likely that the additional types of sensilla found in females are required for the biological tasks exclusive for that sex, e.g., searching for host eggs and evaluating their suitability for parasitism. In comparison to males, female M. amalphitanum and M. caribea have two more types of sensilla (BS and PS).

We found no correlation between the number of antennal sensilla types and body size in our data or when these were combined with the data on larger Chalcidoidea and Braconidae (Fig. 7B, Table 2). The need for heterogeneous and varied information about the environment necessitates a diversity of receptors to detect that information, no matter what the size of the organism.

The possible functions of the sensilla in our work have been derived from morphological features, such as the overall shape of sensillum, presence and number of pores, presence of a basal socket, and position of the sensillum on the antenna (Van Baaren et al., 2007). Comparison with earlier data was based on the external morphology of sensilla as seen in SEM photographs.

Distribution

Within each of the species studied, sensilla distribution was found to be invariable. Between the studied species, significant differences were observed only on the flagellum so the scape and pedicel are omitted in Fig. 6, showing the distribution of antennal sensilla in M. mymaripenne and M. caribea, since these segments are identical in all three studied species. Terminal flagellar segments appear to be the most species-specific parts of the antenna in Chalcidoidea (Barlin, Vinson & Piper, 1981). Constant sensilla distribution within species and constancy of interspecies differences make the distribution of antennal sensilla a reliable distinguishing trait for species identification.

Sizes of antennal sensilla

The antennal sensillum sizes vary moderately between the studied species (Table 1, Table S3). It has been found that the types of sensilla which appear on several flagellomeres (such as ChS-AP, TS1-AP, TS2-AP and SS) tend to differ considerably in length and, less often, in diameter between different flagellomeres in specimens of the same species and sex (Tables S2–S4). Therefore, interspecies and intersex comparisons for these types of sensilla have been made separately for each flagellomere.

Females tend to have considerably longer and wider sensilla than males in both M. amalphitanum and M. caribea (Tables S2 and S3). In three comparable studies, treating Lysiphlebus fabarum (Marshall, 1986), O. phongi, and T. australicum, where measurements were provided separately for each sex (Amornsak, Cribb & Gordh, 1998; Xi et al., 2010; Xi et al., 2011), no such clear tendency in size among specimens of the same species and sex was observed. In other detailed studies on the antennal sensilla of parasitoid wasps the data on the sizes of sensilla were not specified between sexes for most types of sensilla.

In both sexes M. amalphitanum have sensilla longer than in M. caribea, while M. mymaripenne tend to have medium sensillum length. Sensillum diameters are prone to vary insignificantly, especially in males (Tables S1 and S2).

The data on three types of sensilla, chemoreceptors (MPS), mechanoreceptors (ChS-AP) and proprioceptors (TS2-AP) in the studied species were combined with data on larger parasitoid wasps. Possible homologues of these sensillum types in each study have been deduced from morphological traits, relative positions, and distribution of sensilla.

Only TS2-AP length and width have been found to decrease significantly with the body size (Figs. 8A and 8B; Table 2). TS2-AP have been shown to inform parasitoid wasps on the relative position of their antenna through changes in the scapal-head angle: TS2-AP on the scape touch the socket of antenna as it moves and deviate to different degrees from their initial orientations (Schmidt & Smith, 1986; Consoli, Parra & Zucchi, 2010). It seems that the same mechanism is implemented in the scape-pedicel joint, the second location of these sensilla. Thus, the sizes of TS2-AP should depend on the size of the antennal socket and the distance between the scape and pedicel, which decrease with the body size. Additionally, TS2-AP appear to be remarkably shorter in Megaphragma species than in larger parasitoid wasps, in which they are 4–17 times longer.

MPS length and ChS-AP shows correlation between variables and difference of slope from zero only in OLS regression analysis. Also, the regression slope values were relatively small (0.39 and 0.25, respectively). Therefore, we suppose that either there is only weak correlation of these parameters with the body length, or that it is absent.

Miniaturization of sense organs in parasitoid wasps

Our results demonstrate a strong correlation between the number of antennal sensilla and the body size in parasitoid wasps. These findings are comparable to the findings on insect compound eyes, as the decrease in body size also leads to a considerable decrease in the number of ommatidia (Polilov, 2016). Thus, species of genus Megaphragma have about 30 ommatidia per eye (M. carribea 32 ± 3; M. mymaripenne 29 ± 1; M. amalphitanum 29 ± 1). Ommatidia of the minute T. evanescens have inner ultrastructure comparable to Megaphragma (Makarova, Polilov & Fischer, 2015), however, the number of ommatidia in their eye is about 4.5 times as high (Fischer, Müller & Meyer-Rochow, 2011).

Sense organ sensitivity in insects with moderate differences in body size is affected mostly by their lifestyles (Stöckl et al., 2016). Insects can have a 30 times difference in body size and 1,000 times difference in the body volume; it seems logical to suggest that the difference in the sensitivity of their sensory organs should also be considerable though there is no definite evidence for this. While relatively reduced sensitivity might be true for vision (Palavalli-Nettimi & Narendra, 2018), Megaphragma species seem to have a level of olfactory and gustatory sensitivity close to that of larger parasitoid wasps. This needs to be further investigated to make more accurate statements about the retention of sensitivity in a miniaturized antenna.

The miniaturization of insect sensory organs results in the reduction of the relative volume of the structures which process sensory information (Makarova & Polilov, 2013a), e.g., the complexity of antennal sensory system was shown to correlate with the size of antennal lobes and individual glomeruli (Kelber, Rössler & Kleineidam, 2009; Mysore, Shyamala & Rodrigues, 2010), while the number of ommatidia in a compound eye was shown to correlate with the size of the optic lobes (Power, 1943; Rein, Zöckler & Heisenberg, 1999; Gronenberg & Hölldobler, 1999). Interestingly, the decrease of the relative volume was shown to be different for each of the optic ganglia (Makarova & Polilov, 2013b).

Despite a dramatic reduction in the number of antennal sensilla and number of ommatidia in miniaturized parasitoid wasps, the sizes of these sense structures did not show any strict correlation with the body size. It appears that the limit of eye miniaturization is set by the diameter of one facet, which determines the power of the facet lens. The smallest facet diameter in insects was recorded in T. evanescens (6.39 ± 0.33 µm; Fischer, Müller & Meyer-Rochow, 2011), while in smaller Megaphragma it was measured as 8.1 ± 0.3 µm (Makarova, Polilov & Fischer, 2015). Antennal sensilla sizes are also mostly uncorrelated with the body size in parasitoid wasps. However, it seems that Megaphragma have larger relative sensilla diameters than most other chalcidoids (from 0.32 µm (TS2-AP) to 2.6 µm (TS1-AP)). It appears that sensilla and ommatidia are hard to miniaturize because of the complexity of their ultrastructure, which imposes the limits of their reduction.