Glomerular Organization of the Antennal Lobes of the Diamondback Moth, Plutella xylostella L.

The antennal lobe of the moth brain is the primary olfactory center processing information concerning pheromones and plant odors. Plutella xylostella is a major worldwide pest of cruciferous vegetables and its behavior is highly dependent on their olfactory system. However, detailed knowledge of the anatomy and function of the P. xylostella olfactory system remains limited. In the present study, we present the 3-Dimentional (3-D) map of the antennal lobe of P. xylostella, based on confocal microscopic analysis of glomerular segmentation and Neurobiotin backfills of Olfactory Receptor Neurons (ORNs). We identified 74–76 ordinary glomeruli and a macroglomerular complex (MGC) situated at the entrance of the antennal nerve in males. The MGC contained three glomeruli. The volumes of glomeruli in males ranged from 305.83 ± 129.53 to 25440.00 ± 1377.67 μm3. In females, 74–77 glomeruli were found, with the largest glomerulus ELG being situated at the entrance of the antennal nerve. The volumes of glomeruli in females ranged from 802.17 ± 95.68 to 8142.17 ± 509.46 μm3. Sexual dimorphism was observed in anomalous supernumerary, anomalous missing, shape, size, and array of several of the identified glomeruli in both sexes. All glomeruli, except one in the antennal lobe (AL), received projections of antennal ORNs. The glomeruli PV1 in both sexes received input from the labial palp nerve and was assumed as the labial pit organ glomerulus (LPOG). These results provide a foundation for better understanding of coding mechanisms of odors in this important pest insect.


INTRODUCTION
In insects, the olfactory system plays a highly important role in detecting odorants involved in sexual communication, social integration, host recognition, and escaping from enemies over a distance (Lopes et al., 2002;Gill et al., 2013;Sun et al., 2014;Xu et al., 2016). The antennae are the primary olfactory organ and bear several types of sensilla (Galizia and Rossler, 2010;Yan et al., 2014Yan et al., , 2017a. The most prominent sensillum types have olfactory functions and contain olfactory receptor neurons (ORNs), which send projections directly to the antennal lobe (AL), the primary olfactory center of the insect brain.
Typically, the insect AL is composed of many spheroidal neuropilar units, called glomeruli that house synaptic contacts between receptor axons and AL interneurons (Hansson and Stensmyr, 2011). The arrangement, size, and number of AL glomeruli are species-specific, allowing the identification of individual glomeruli according to the size, shape, and its position. The number of glomeruli in the AL varies from fewer than 15 glomeruli in Scaphoideus titanus (Rossi Stacconi et al., 2014) to more than 1,000 in locusts (Rospars, 1988). Most insect species studied to date have 40-160 glomeruli arranged in a single or double layer (Rospars, 1983;Hansson and Anton, 2000). In many insect species, sexual dimorphism with respect to specific glomeruli is observed, e.g., the AL of male Lepidoptera contains enlarged glomeruli that form the macroglomerular complex (MGC) at the entrance of the antennal nerve into the Al in Mamestra brassicae (Rospars, 1983) and Trichoplusia ni (Todd et al., 1992). All ORNs that express the same specific odorant receptor converge onto the same glomerulus (Vosshall and Wong, 2000). Odor representation from ORNs to projection neurons (PNs) is generally conserved . Different PN classes target dendrites to distinct olfactory glomeruli, whereas PNs of the same class exhibit indistinguishable anatomical and physiological properties . In the Drosophila olfactory circuit, 50 ORN classes and 50 PN classes form synaptic connections in 50 glomerular compartments in the antennal lobe, each of which represents a discrete olfactory information-processing channel . The one ORN class to one glomerulus and to one PN class relationship in the Drosophila olfactory system is likely an extreme situation as, in other species, this may vary as the glomeruli numbers vary dramatically. In male moths, the MGC receives and processes information regarding the female sex pheromone, calcium imaging in Heliothis virescens confirms that sex pheromone responses are restricted to the male-specific MGC, plant odors to ordinary glomeruli (Galizia et al., 2000), and sex pheromone-specific receptor neurons arborize in the MGC (Hansson et al., 1992;Ochieng et al., 1995).
The diamondback moth (DBM; Plutella xylostella; Lepidoptera: Plutellidae) is a major worldwide pest of cruciferous vegetables. Because of the intensive and extensive application of insecticides for the control of P. xylostella, the species has become one of the most resistant insect pests in the world (Sarfraz et al., 2005;Furlong et al., 2013;Wang et al., 2013;Li et al., 2016). This resistance has prompted increasing efforts to identify new, natural approaches to control this insect. One of these approaches focuses on understanding the plant chemistry that plays a major role in the moth acceptance or rejection of host plants. However, relatively little is known about the olfactory pathways in the central nervous system of this insect. To better understand the general organization and the functional significance of AL glomeruli in odor processing, we generated the first digital 3-D reconstruction of the AL in P. xylostella through Confocal Laser Scanning Microscopy (CLSM) reconstruction of the AL and backfilling of the antennal neurons. The atlas will serve as a template for future analysis of physiological responses in morphologically identified glomeruli in this pest insect.

Insects
Plutella xylostella were reared at the temperature of 25 • C, relative humidity of 75% and photoperiod of (L:D = 14:10) in the Insect Neuroethology & Sensory Biology laboratory, Shanxi Agricultural University. Pupae were individually segregated into test tubes according to sex, and recently emerged (1-2-days-old) female and male adults of P. xylostella were used in this study.

Palp and Antennal Backfills
Six male and six female moths were anesthetized with carbon dioxide and immobilized using double-sided adhesive tape, and the ventral part of the head was exposed. The antennae or the palp of one side of the moth were cut near the base, and the remainder of the segments were inserted for 30 min in an appropriately sized micropipette tip filled with 2% Neurobiotin (Neurobiotin Tracer, Vector Laboratories, Burlingame, USA) solution. After the micropipette was removed, the preparations were then incubated at 4 • C for 2 h in a humidified and dark box for diffusion of the dye. Subsequently, the brains were dissected out in 4% paraformaldehyde in phosphate buffer (PB, 0.1 M, pH 7.2) under a dissecting microscope, and the isolated brains were fixed in the same fixative at 4 • C for 1-2 days for batch processing. After rinsing three times with 0.1% Triton X-100 in 0.1 M PB, the brains were incubated in 2 µg/ml Alexa-488 Streptavidin (Molecular Probes, Life Technologies, USA) solution in PBS for 1-2 days and then in 10 µg/ml Propidium Iodide (Sigma) solution for 30 min at room temperature. The brains were dehydrated in an ascending ethanol series of 30, 50, 75, 95, and 100% alcohol and then embedded in DPX (Sigma) after a brief xylene transition.

Confocal Laser Scanning Microscopy
Twelve whole-mounts of brains were imaged with a Zeiss LSM 710 confocal laser scanning microscope equipped with multiple laser lines that permitted the visualization of structures labeled with Alexa-488 and labeling of Propidium iodide used in this study. CLSM imaging was conducted using 488-nm excitation wavelength for the Alexa-488 Streptavidin labeling, and the Alexa-488 signal was collected between 490 and 560 nm. The propidium iodide signals, which label the nucleus of neurons, were excited with a 561 nm laser line and collected between 567 and 596 nm. For overview scanning of the whole brain and detailed scanning of the antennal lobe, a 40× (NA1.3) oil objective was used with a sampling rate matching the Nyquist sampling rate with a pinhole setting of 1 Airy Unit. Image data were captured as serial stacks (pixel dwelling time of 0.39 µs, pinhole 1 AU, step size 0.4-0.45 µm). Generally, each brain requires an image stack of 300-500 slices (depending on the mounting orientation of the brain) of images with 2048 × 2018 pixels [with a pixel dimension of 0.104 × 0.104 × 0.4 µm (x, y, z)].

3-D Reconstructions and Identification of Glomeruli
All confocal image stacks were viewed and processed with the 3-D reconstruction software Imaris (vers: 8.4.0, Bitplane, Zurich, Switzerland). Glomeruli in the ALs were entirely demarcated by manually tracing the outline of each glomerulus in every other section of the stack file. The outlines of the glomeruli could be easily defined by a combination of many characteristics in the images: auto-fluorescence of the antennal lobe structure, propidium iodide-labeled glia cells that usually surround the glomeruli (Figures 1A,C) (Yan et al., 2017b) and signals from the Neurobiotin backfills of the ORNs (Figures 1B,C). Such manually delineated lines were subsequently used to generate surface rendered glomeruli structure, and the software generated the volume as well as sphericity of each individual glomerulus ( Figures 1D-F). Two analysts independently analyzed each data set to reduce possible errors and bias. Glomeruli were identified according to their location, shape, and size. The glomerular nomenclature used was adopted from previous publications of other groups (Ignell et al., 2005;Solari et al., 2016). Namely, each glomerulus was marked by one or two capital letters indicating the general position: anterior (A), posterior (P), ventral (V),   For clarity, the entire antennal lobe was separated into 7 consecutive arbitrarily reconstructed series (S1-S7). This panel shows the front 4 series (last 3 series will be demonstrated in Figure 4): (A1-A3) The first level, S1, where was the most anterior layer. The dorsal (D), lateral (L), medial (M), and central (C). Letters were followed by numbers to indicate glomerulus presentation order from the most anterior to the most posterior in the same region. Additionally, the glomeruli were identified using several structural landmarks such as (1) general contour structure of the AL; (2) the entrance of the antennal nerves; (3) the brain orientations, and (4) specific, easily identifiable glomeruli such as putative MGC, enlarged glomeruli in the female, ELG, PV1, etc. The same nomenclature structure was used for both the male and female glomeruli. Although we did not systematically compare the glomeruli from the male and female AL, every attempt was made to have glomeruli in both sexes in similar positions and sizes to have comparable glomeruli annotation. This visualization based matching presented some inherited limitations that the same glomeruli name may not present the homologous glomeruli. Every glomerulus was assigned with a random color for the purpose of visualization. The MGC received served as an essential orientation landmark for the male AL. Final output data were stored in TIFF format and annotated using Photoshop CS software (Adobe, San Jose, CA, USA).

Statistical Analyses
The following parameters were measured: total number of glomeruli and volume of glomeruli of the AL. The volume of each individual glomerulus was calculated with the SPSS statistical software package (ver. 13.0, SPSS incorporated, Chicago, Illinois), and values are presented as the mean ± SD.

Three-Dimensional Reconstruction of the AL
The ALs of P. xylostella were mostly sphere-shaped structures located in the front most part of the brain. Numerous glomeruli were found in the AL (Supplementary Videos 1-3). Generally, the majority of the glomeruli were spherical or elliptical in shape. The sphericities (a measure of how round a structure is) of the glomeruli were in the range of 0.7-0.96 (while a perfect sphere is 1). Only few glomeruli per antennal lobe displayed irregularities with a lower sphericity (Figure 2). The glomeruli were mostly located at the periphery of the lobe in a single layer and generally, the anterior glomeruli were more densely packed than the posterior glomeruli (Figures 3-8). Most glomeruli were uniform in shape, size, and relative position, based on the comparison of different ALs and therefore could be identified and recognized individually. Only glomeruli four and five displayed some anomalies (missing or extra) ( Table 3). The glomeruli were named and color-coded as descripted previously. Each identified glomerulus was named and numbered in the order in which the layers appear from the most anterior to the most posterior.

Male ALs
A putative Macro-Glomerular Complex (MGC) structure and 74-76 ordinary glomeruli were found in the six male ALs studied (

The MGC
The putative MGC structure was found close to the entrance of the antennal nerve of male P. xylostella. It contained three glomeruli: the cumulus and glomeruli a and b (Figures 3D1-3, 4A1-3, B1-3, 5D). They were clearly separated from the array of ordinary glomeruli, based on their shape, volume, and location. As the most anterior glomerulus of the putative MGC, the cumulus exhibited a cylindrical shape, but the ventral part was flattened ( Figure 3C3). The cumulus was the largest glomerulus in the AL, and the volume was 25440.00 ± 1377.67 µm 3 (Table 1). Glomerulus a of the putative MGC positioned posteriorly from the cumulus, and the volume was 12540.00 ± 968.50 µm 3 . Glomerulus b of the putative MGC positioned posteriorly from glomerulus a and was 16640.00 ± 1818.79 µm 3 in volume ( Figure 5D; Table 1). All three glomeruli in the MGC received projections of ORNs from the antenna (Figures 3D2,  4A2,B2). The cumulus and glomeruli a and b in MGC were regarded as very important landmark glomeruli.

Ordinary Glomeruli
The maximum number of ordinary glomeruli was 76, which was found in one male AL. M5, V3, PV2 and PV6 were missing in some ALs within our samples; M5 was found in five of the six ALs, V3 was found in four of the six ALs, PV2 was found in two of the six ALs, and PV6 was found in three of the six ALs. The other 72 glomeruli could be systematically identified in all six ALs (Table 3). Seven glomeruli in males had volumes smaller than 1000 µm 3 (Figure 9). Glomerulus AL2 was the smallest ordinary glomerulus, with a volume of 305.83 ± 129.53 µm 3 (Table 1), and this glomerulus was located in the anterior-lateral region of the AL (Figures 3B1-3,  5A). Most glomeruli showed volumes between 1000 and 3000 µm 3 , with 32 glomeruli volumes between 1000 and 2000 µm 3 , and 23 between 2000 and 3000 µm 3 in males. In males, 17 glomeruli had volumes larger than 3000 µm 3 , including 14 ordinary glomeruli and the 3 glomeruli in the MGC (Figure 9). Glomerulus PL2 was the largest ordinary glomerulus, with a volume of 9166.50 ± 1631.38 µm 3 ( Table 1). PL2 neighbored the MGC and was located in the posterior-lateral region of the AL (Figures 3D1-3, 5D). Antennal backfills revealed staining in many, but not in all glomeruli. Glomerulus PV1 was found without ORN branches ( Figure 3C2). Back filling of the labial palps with Neurobiotin performed similarly, revealed only one glomerulus (PV1) was innervated by sensory neurons from the labial palps ( Figure 10B). This glomerulus had a relatively large size (4573.00 ± 658.85 µm 3 ) and was located at the ventral side of the ALs (Figures 5A,F; Table 1). It is worth to note that the back filling of the labial palps nerve labels the PV1 of boths side of the brain, therefore, unlike the other glomeruli, PV1 receives information from both labial palps. The glomeruli AL2, PL2 and PV1 were regarded as landmark glomeruli in male ALs.

Female ALs
In the six female ALs, 74-77 glomeruli were found ( Table 2; Supplementary Video 3). Similar to the male, there was a slight variation in glomeruli numbers: 77 glomeruli were found in the ALs of one female. M3 was found in three of the six AL, M6 was found in two of the six ALs, AM4, M5, and PL3 were found in five of the six ALs, respectively. Nevertheless, 72 glomeruli could be systematically identified in all six ALs (Table 3). Four glomeruli in females had volumes smaller than 1000 µm 3 (Figure 9). AD3 was the smallest glomerulus, and the volume was 802.17 ± 95.68 µm 3 ( Table 2), and this glomerulus was located in the anterior-dorsal region of the AL (Figures 6A1-3, 8A). In females, 24 glomeruli had volumes between 1,000 and 2,000 µm 3 ; the volumes for 26 glomeruli were between 2,000 and 3,000 µm 3 ; and an additional 23 glomeruli had volumes larger than 3,000 µm 3 (Figure 9). Corresponding to the location of the MGC in the male, the largest glomerulus ELG was located in the anterior-lateral region of the AL and close to the entrance of the antennal nerve (see ELG in Figures 6B1-3, 8A), and the volume of this glomerulus was 8,142.17 ± 509.46 µm 3 ( Table 2). All the glomeruli received antennal receptor neurons with the exception of PV1 glomerulus (Figure 7B2), which received neurons from the labial palps, similar to the males ( Figure 10C). The glomerulus was large (4,961.33 ± 705.50 µm 3 ) and was located at the ventral side of the AL (Figures 8D,F; Table 2). The glomeruli ELG, AD3, and PV1 were regarded as landmark glomeruli in female ALs.

Landmark Glomeruli
There were several distinct glomeruli in both sexes which were used as landmarks for glomeruli identification and matching. For the male AL, in addition to the above-mentioned putative MGC, glomeruli AL2, PV1, PL2, AV3, AL4, PC1, PC4, and PD9 could be easily identified either by their marked locations, shapes or sizes. For example, the glomerulus AV3 was a flattened, elliptical sphere, located in the anterior-lateral region of the AL and under the MGC (Figure 5A). AL4 neighbored the MGC, and was larger than other ordinary glomeruli in this region ( Figure 5A). PC1 was a slender cylinder, and located in the posterior-lateral region of the AL (Figure 5C). PC4 was a long elliptical sphere, located in the posterior-central region of the AL (Figure 5C). The larger glomerulus PD9 was located in the most posterior region of the AL, and close to the MGC ( Figure 5C).
In the female ALs, in addition to the morphologically evident glomeruli ELG, AD3, and PV1, glomeruli AC2, AV5, PL5, and PL1 were also easily recognizable. AC2 was oblong in shape, and located in the most anterior region of the AL (Figure 8A). AV5 and PL5 were among the largest glomeruli in the ventral regions of the AL (Figures 8A,D). PL1 was the third largest glomerulus, and located in the posterior-central region of the AL (Figure 8C).

Comparison of Male and Female ALs
Comparing the glomerular organization of the AL of male and female P. xylostella demonstrated both sexes possessed approximately the same number of AL glomeruli. Within the same sex, most glomeruli proved to be highly consistent in size, shape and position in the ALs. Nevertheless, a few anomalous glomeruli were missing in some ALs of both sexes (Table 3). Additionally, although the majority of glomeruli were between 1,000 and 3,000 µm 3 in volumes in both sexes, there are some size distribution differences in the male and female glomeruli (Figure 9). The peak number of glomeruli in the males is with a volume between 1,000 and 2,000 µm 3 while, in the females, the peak number of the glomeruli was with a volume between 2,000 and 3,000 µm 3 (Figure 9).
Some special glomeruli were present in both sexes in the same position. For example, the largest glomeruli MGC in males and ELG in females occupy the same domain in both sexes. The second largest glomerulus PL2 in males and PL5 in females was situated under the MGC and ELG, respectively (Figures 5D, 8D;  Tables 1, 2). The smallest glomerulus AL2 in males and AD3 in females was located in the anterior region of the AL and near to the entrance of the antennal nerve (Figures 5A, 8A;  Tables 1, 2). Furthermore, PV1 was the only glomerulus not innervated by the antennal nerve, but by the labial palpus nerve in both sexes (Figure 10). This glomerulus was relatively large and located at the ventral side of the AL (Figures 5F, 8F). The glomerulus PV1 in the male was smaller than that in the females. Additionally, unlike other glomeruli in the AL which received neuronal input from ipsilateral antennal nerves, this glomeruli received innervation from both sides of the labial palpus nerve (Figure 10). Underlined texts indicate missing/abnormal glomeruli.
FIGURE 9 | Size distributions of the glomeruli in the male and female of P. xylostella. Note the male has the most glomeruli in the volume of 1-2 × 1000 µm 3 while the female has the most glomeruli with the volume of 2-3 × 1000 µm 3 .

DISCUSSION
In this paper, we presented a complete 3-D reconstruction of the glomerular organization of P. xylostella, based on the systematic anatomical matching of glomeruli within and between the sexes. We found most glomeruli were apparently isomorphic between sexes. However, significant differences occurred in the presence, absence, sizes, or locations of some glomeruli.

Number of Glomeruli in P. xylostella
Similar to other investigated insect species, the glomeruli are distributed around a central fiber core Hildebrand, 1992, 2000;Berg et al., 2002;Greiner et al., 2004). The number of glomeruli can vary greatly depending on the species (Rospars, 1988;Huetteroth and Schachtner, 2005); in Lepidoptera, the number of glomeruli ranges from 60 to 70 (Rospars, 1983;Hildebrand, 1992, 2000;Berg et al., 2002;Sadek et al., 2002;Greiner et al., 2004;Masante-Roca et al., 2005;Namiki et al., 2014). The number of glomeruli found in P. xylostella is slightly higher than in the other moth species previously mentioned, which is not surprising because these species belong to very distantly related families. However, even closely related species have significant differences in the number of glomeruli. Generally, one glomerulus is the target site for one ORN type (Couto et al., 2005). The number of glomeruli is different in different insect species because they possess different lifestyles. When comparing two closely related species, Cydia molesta has fewer glomeruli than Lobesia botrana, and the former is an oligophagous insect species, while the latter is a polyphagous species; therefore, C. molesta may require a narrower range of olfactory cues to identify their hosts than that of L. botrana (Varela et al., 2009). When insects lack glomeruli caused by some factors, e.g., mutagenesis causes, these insects lose corresponding olfactory functions. Orco mutant ants (disrupted orco, a gene required for the function of all ORs) lack most of the ∼500 glomeruli found in wild-type ants and are unable to perceive pheromones (Trible et al., 2017). Generally, insects with complex lifestyles (e.g., social insects) have relatively large numbers of glomeruli and use a large variety of chemical cues (Galizia et al., 1999;Kleineidam et al., 2005). P. xylostella are oligophagous insects (Renwick et al., 2006), and they have a relatively narrow range of olfactory cues. Thus, P. xylostella has lower numbers of glomeruli than those of the social insects studied to date. Nevertheless, comparing with other moths studied so far, P. xylostella shows a higher number of glomeruli suggesting a neuronal base for a more complex olfactory behavior.

Enlarged Glomeruli in the Female P. xylostella
In the female P. xylostella the largest glomerulus, ELG, was found at the entrance of the antennal nerve, a location similar to the one found in males MGC in the antennal lobe. However, this glomerulus was not a complex structure and was much smaller than that of the male MGC. This morphologically distinct glomerulus is not universally present in the female AL. It has previously been described in some female moths, such as Bombyx mori (Koontz and Schneider, 1987), M. sexta (Rospars and Hildebrand, 2000), H. virescens (Berg et al., 2002) but not in S. littoralis (Ochieng et al., 1995), and C. molesta (Varela et al., 2009). Physiological characterization of a PN innervating this glomerulus showed that it processes a self-released sex pheromone in addition to plant volatiles (Ljungberg et al., 1993;Reisenman et al., 2004;Trona et al., 2010). The similarity of the ELG suggests a similar role. It remains to be clarified in P. xylostella.

Other Glomeruli Between the Sexes
Male moths must locate a mating partner via the dedicated sex pheromones cues. Mated females must identify the suitable oviposition sites via plant volatiles. Therefore, it is not surprising the glomerular structure differs between the sexes. The specific ordinary glomeruli AL2 and PL2 were found close to the MGC. Due to the closeness with the putative MGC of these specific glomeruli, we speculate that these glomeruli might receive and process information regarding the female sex pheromone. In the female ALs, the specific ordinary glomeruli AD3 and PL5 were located near the ELG, and might be responsible for receiving/processing certain female specific behaviors-relative chemical cues (e.g., plant volatiles). More functional studies of these specific ordinary glomeruli are required to elucidate their true functions. Glomerulus PV1 did not receive projections of the antennal ORNs but received input from the labial palpus nerve, and these glomeruli were located at the ventral of the AL in both sexes and assumed to be the labial pit organ glomerulus (LPOG). The LPOG is found in some species, e.g., the moth Rhodogastria (Bogner et al., 1986), Pieris rapae (Lee and Altner, 1986), M. sexta (Kent et al., 1986), mosquitoes (Anton et al., 2003), C. molesta (Varela et al., 2009), and Helicoverpa armigera (Zhao et al., 2013(Zhao et al., , 2016. The LPOG is specialized in sensing CO 2 (Bogner et al., 1986;Stange, 1992;Guerenstein and Hildebrand, 2008;Ning et al., 2016). Glomerulus PV1 in ALs of P. xylostella is believed to process information on CO 2 levels in the environment. Further functional studies (either single sensillum recording or Ca imaging) are needed to verify if this is the case in this insect species.
In conclusion, we provided a 3-D reconstruction of the glomerular structure in the AL of P. xylostella. This result provides a foundation for further studying of the olfactory information processing in this important economical pest.

AUTHOR CONTRIBUTIONS
XS, CH, and XY conceived and designed the study. XY and ZW acquired and analyzed the data. XY, JX, and CD analyzed and interpreted the data. XY and XS wrote the manuscript. CH provided research funding.