The gastric sieve of penaeid shrimp species is a sub-micrometer nutrient filter

The Pacific whiteleg shrimp [Penaeus (Litopenaeus) vannamei] and the giant tiger shrimp (Penaeus monodon) (called the black tiger shrimp in Thailand) are the most commonly cultivated shrimp species in Thailand and constitute an important export product (Engle et al., 2017). The shrimp culture industry is continually threatened by emerging diseases caused by pathogens such as viruses, bacteria and protozoa (Thitamadee et al., 2016). Many of these pathogens target vital organs such as those of the alimentary tract. This consists of the mouth, esophagus, stomach (anterior and posterior stomach chambers, ASC and PSC), hepatopancreas (HP, combined functions similar to the liver and pancreas in vertebrates and sometimes called the midgut gland), anterior midgut cecum (AMC), midgut (MG or intestine), hindgut, posterior midgut cecum (PMC) and anus.

In shrimp defense against pathogens, the cuticle (a non-living external layer containing the polymer chitin) is the primary physical barrier. Cuticular structures are hard when calcified but soft and flexible when not. For example, the un-calcified cuticle between abdominal segments (articular elements) is soft to allow flexing, in contrast to the generally hardened, relatively inflexible shell. For the following description of the shrimp alimentary system, we use the terminology of Bell and Lightner (1988). Unlike that of vertebrates, the shrimp stomach is of ectodermal embryonic origin, and so the cuticle from the external body surface is continuous with the inner surface of the esophagus, the stomach and the hindgut near the anus (Meiss and Norman, 1977). When shrimp molt (undergo ecdysis), the cuticle of the outer surface, the mouth, the esophagus, the stomach and the end of the hindgut is sloughed and replaced by a previously generated underlying cuticle. Some parts of the stomach cuticle are soft (un-calcified) while others (e.g. ossicles to grind food) are hardened (calcified). However, similar to the vertebrate intestine, the epithelium of the homologous shrimp MG and its associated organs (i.e. HP, AMC, MG and PMC) are of endodermal origin and their lumens are not covered with cuticle. Thus, there is a transition in the alimentary system from the cuticle-covered stomach to the subsequently uncovered regions up to the end near the anus, where the cuticular covering resumes. At the same time, the epithelial cells of the MG produce a multilayered peritrophic membrane (PTM) that forms a chitinous layer that surrounds the MG contents and has an exclusion diameter of approximately 20 nm (Martin and Hose, 2010).

At the anterior end of the shrimp body, the transition from cuticle to no cuticle takes place at the dorsal, posterior end of the ASC and at the posterior end of the PSC or pyloric (from the Ancient Greek pulōrós or ‘gatekeeper’) region where the gastric sieve (GS) is located. Similar structures are widely reported in crustacean species. After mechanical and chemical digestion in the ASC, relatively large, liquid-borne particles pass from the dorsal end of the ASC directly into the intestine, without passing into the PSC. The PSC is located ventro-medially to the ASC and is separated from it by a relatively narrow, longitudinal slit in the ventro-medial ASC wall that leads first into a dorsal sub-chamber (PSCD) and then via another narrowing into a ventral sub-chamber (PSCV) where the GS is located. Entry to the PSCV is partially obstructed by longitudinal brush-like setae extending up from the dorsal midline of the underlying GS that is an integral part of the PSCV. Together, the narrow slit and the brush-like setae extending from the GS and from the adjacent stomach walls serve as a primary filtration apparatus that allows entry of only liquids and relatively small particles into the PSCV, which serves as a kind of secondary filtration chamber where the GS is located.

Basically, the GS serves as a size-partitioning structure that allows liquids containing dissolved nutrients and very fine particles from masticated and pre-digested stomach contents to enter the HP via right and left hepatopancreatic ducts, where further digestion and transfer of nutrients into the hemolymph takes place. Particles too large to pass through the GC are conducted out of the PSCV through a separate channel to join the coarser material going directly into the intestine from the dorsal region of the ASC.

This process has been described in detail for the river prawn Macrobrachium carcinus (Order Decapoda) (Lima et al., 2016) and the mountain shrimp Anaspides tasmaniae (Order Syncardia) (Wallis and Macmillan, 1998). Structural and morphological investigation of the GS has also been reported for many decapod species, e.g. Penaeus (Marsupenaeus) japonicus (Lin, 1996), Pleoticus muelleri (Díaz et al., 2008), Nematocarcinus lanceopes, Notocrangon antarcticus and Chorismus antarcticus (Muhammad et al., 2012), P. monodon and Metapenaeus ensis (Lin, 2000) and M. carcinus (Lima, et al., 2016). All these studies reported the same pattern of GS organization as an inverted-V, dome-like structure covered with numerous stacks of overlapping fine setae and located at the ventro-medial side of the PSC.

Recently, the pore size (exclusion limit) of the GS in sub-adult P. monodon has been determined using various sizes of inert digestibility particles. Particles larger than 3 µm were rapidly excluded while those smaller than 400 nm stayed longer in the digestive tract (Wade et al., 2018), suggesting that the exclusion limit was smaller than approximately 3 µm. In addition, it is widely believed among shrimp pathologists (T. W. Flegel, personal communication) that the GS is normally able to exclude bacteria from the HP because histological and ultrastructural examination of the HP tissue of normal (healthy) shrimp does not reveal the presence of bacteria in the HP lumen even when they are abundant in the stomach and MG of the same specimens, as previously reported (Hopkin and Nott, 1980). Also, when P. monodon was challenged with Vibrio spp. by immersion and oral intubation, only soluble digested bacterial antigen (i.e. no intact bacteria) could be detected in the HP (Alday-Sanz et al., 2002). In contrast, anal intubation with Vibrio spp. has been shown to result in HP tubule infection and disease (Song et al., 1993), suggesting that bacterial introduction with pressure against the internal anatomical structures and mechanics resulted in disruption of the normal exclusion process.

Because of the information summarized above, we hypothesized that the shrimp GS has the capability to exclude particles 1 µm or larger from entering the HP and that this hypothesis could be tested by using polystyrene microspheres and by ultrastructural examination of the GS. Fluorescent polystyrene microsphere tracking has been widely used for many purposes such as determining the size of bacteria and phytoplankton that rotifers consume (Ooms-Wilms et al., 1995; Agasild and Nõges, 2005), tracing the transition in planktonic food webs (Cole et al., 2013), investigating the existing functions and development of the GS in different larval stages of the Eastern spiny lobster Sagmariasus verreauxi (Simon et al., 2012) and assisting in biomedical and nanoparticle studies (Jiang et al., 2017).

The aim of our study was to investigate the exclusion limit of the GS in two species of penaeid shrimp (P. monodon and P. vannamei) using a combination of approaches including inspection of gross morphology, histology and visualization of fed fluorescent microbeads. We also compared the structure of the GS between P. monodon and P. vannamei to look for possible differences between the species and at different life stages in P. vannamei. Our experiments revealed that the GS exclusion limit was less than 1 µm, that the sieve pore diameter was approximately 0.2–0.7 µm, and that there were no significant differences in the GS of P. monodon and that of P. vannamei or among different developmental stages in P. vannamei. The results have ramifications for shrimp nutrition and disease control and provide useful basic knowledge that may also be valuable in the design of novel microfilters for medical and industrial use.

All the experiments and procedures used with shrimp in this study were approved by the Faculty of Science, Mahidol University Animal Care and Use Committee SCMU-ACUC, Faculty of Science, Mahidol University, Thailand (no. MUSC61-007-409).

Broodstock specimens (average body mass 50–60 g) and juveniles (average body mass 5–8 g) of the whiteleg shrimp, Penaeus vannamei Boone 1931, and the giant tiger shrimp, Penaeus monodon Fabricius 1798, were obtained from local farms in Nakhon Nayok province and Chonburi province, Thailand. They were transported to and maintained in aquaria containing artificial seawater (Marinium, Mariscience Int. Co. Ltd) at 20 ppt and at ambient temperature (27–29°C) with sufficient aeration at Mahidol University, Phayathai campus, Bangkok, Thailand. The shrimp were fed with commercial food pellets (Betagro) twice per day and excess feed was removed twice daily. Shrimp were used for experiments within 7 days of their arrival at the laboratory. Feed was withheld for 3 h prior to experiments.

Broodstock shrimp of both species were chosen for gross anatomical and morphological investigations because their GS was significantly larger than that in the smaller shrimp. For gross anatomical and histological investigations of the gastrointestinal tract (including the GS), shrimp were anesthetized and killed following the AVMA Guidelines for the Euthanasia of Animals (Leary et al., 2013). Briefly, shrimp were submerged in ice for 15 min until no movement was observed. Then, the stomach and HP were removed and fixed in 2.5% glutaraldehyde with 4% paraformaldehyde in 0.1 mol l⁻¹ phosphate buffer (PB) solution for scanning electron microscopy (SEM), and Davidson's fixative solution for routine histology. For histology, tissues were fixed for 24 h in Davidson's fixative and processed for sectioning and staining with hematoxylin and eosin as previously described (Bell and Lightner, 1988), and they were examined using a light microscope (Leica DM750) with a digital camera (Leica ICC50 HD).

After fixing with 2.5% glutaraldehyde with 4% paraformaldehyde in 0.1 mol l⁻¹ PB solution overnight, tissues were washed 3 times (15 min each) with 0.1 mol l⁻¹ PB, pH 7.4 at 4°C. They were post-fixed with 1% osmium tetroxide (OsO₄) in 0.1 mol l⁻¹ PB at 4°C for 2 h followed by gentle washing 3 times with 0.1 mol l⁻¹ PB at 4°C. Next, they were dehydrated through increasing concentrations of ethanol (30%, 50%, 70%, 80%, 90% and 95%), ending with absolute ethanol at 4°C for 15 min (2 times each). The specimens were dried with a critical drying point machine (Hitachi HCP-2) under liquid CO₂ and mounted on stubs with double-coated conductive carbon tape. Finally, they were coated with platinum and palladium in a HitachiE-120 ion sputter device and observed under a Hitachi scanning electron microscope S-2500.

Commercial food pellets (Betagro) (1 g) were immersed in 100 µl of fluorescent bead solution (either 0.1 or 1 µm diameter, for groups 1 and 2, respectively; see below) (Thermo Fisher Scientific) at a concentration of 1×10⁶ beads ml⁻¹ in 0.1 mol l⁻¹ phosphate-buffered saline (PBS) for 3 h. Control food pellets (group 3) were immersed in PBS only. Thereafter, all were left to dry at room temperature for at least 5 h. Each shrimp was fed with three pellets in separate aquaria.

Juvenile shrimp of both species were selected for the feeding experiments using food pellets containing 0.1 µm fluorescent microbeads. After feed administration, translocation of the fluorescent microbeads within individual shrimp was examined using a fluorescence stereomicroscope (Macro Zoom Imaging System MVX10, Olympus) at 4× magnification and at 3, 6, 10 and 100 min post-feeding. Results from different time points were compared with those of shrimp given the control feed with no fluorescent microbeads (group 3).

Juvenile shrimp were divided into three groups (3 replicates per group) and fed with commercial food pellets containing either 0.1 or 1 µm fluorescent microbeads (test shrimp groups 1 and 2, respectively) or no microbeads (i.e. normal food pellet control, group 3). Each shrimp was killed on ice 5 min after feeding, then fixed with 4% paraformaldehyde for 24 h. Tissues were immersed in sucrose solution of increasing concentration (10%, 20% and 30% for 5 h, 5 h and overnight, respectively) at 4°C. Once the tissue was confirmed to sink, it was embedded in OCT compound and cut with a cryostat microtome (Leica) to 7 µm thickness. Each tissue sample was examined under a fluorescence microscope BX53 (Olympus). Additional shrimp were similarly fed, anesthetized and prepared for observation by SEM as described above.

The gross anatomy and histology of the stomach structures in penaeid shrimp are shown in Fig. 1, using a specimen of P. monodon as an example. In Fig. 1A, the ASC has been opened to expose the slit-like portal leading into the PSC, most of which is overlaid by the HP tissue in this image. The PSC consists of a dorsal sub-chamber (PSCD) and a ventral sub-chamber (PSCV) that cannot be seen in Fig. 1A but contains the GS, the position of which is shown with a dashed outline. In Fig. 1B, a lateral view with half of the HP removed shows the whole PSC and its relationship to the ASC. During eating, liquid containing large particles passes from the dorsal side of the ASC directly into the MG, which is homologous to the vertebrate intestine. Liquid with dissolved nutrients and smaller particles passes into the PSCD and then into the PSCV where the GS is located. The GS selectively filters liquids containing nutrients and sub-micrometer particles (see later) that are shunted ventrally into the HP via hepatopancreatic ducts, the left one of which is indicated in Fig. 1B. Particles that cannot pass through the filter are shunted dorsally into the MG to join material coming directly from the AMC and destined for excretion. In Fig. 1C, the PSC has been cut open and the PSCD and PSCV walls have been folded back to reveal the GS that is an integral part of the base of the PSCV. A spearhead-like structure called the posterior median plate (PMP) is situated at the posterior end of the GS, serving as the route for material excluded by the filter that is shunted dorsally into the MG. These structures are shown at higher magnification in Fig. 1D.

Fig. 2A shows a photomicrograph of a frozen tissue cross-section of the cephalothorax through the PSC, surrounded by HP tissue. It reveals the relationship between the dorsal chamber (PSCD) and ventral chamber (PSCV), the latter containing the GS with the closely adjacent PSCV wall. The enlargement in Fig. 2B clearly reveals that both the chamber wall (labeled upper ampullary half, UAH) and the GS (labeled lower ampullary half, LAH) are covered with a carpet of setae. It also clearly shows the longitudinal inter-setal grooves (LSG) that underlie the setal layers. Note also that the dorsal, midline crest of the GS has an extension that is spike like in cross-section with a smooth base and short stalk that then expands into a brush-like head. In Fig. 2C, a similar section through the PSC of a different specimen shows a nascent, replacement GS forming below the existing one (i.e. in preparation for the next molt).

In addition to frozen tissue sections, some specimens were embedded in epoxy resin for preparation of semi-thin sections (Fig. 3) that gave more detailed images of the PSCV structure. In Fig. 3A, the central spike of the GS crest can be seen to extend into the slit between the PSCD and PSCV, together with its covering of setal extensions and with other long setae on either side that arise directly from the top of the GS on either side of the midline. The adjacent PSCV wall or UAH is also covered with long setae. An enlarged image of the LSG (Fig. 3B) shows that they lie at the base of the overlapping ranks of setae on the GS surface (labeled LAH). The higher magnification in Fig. 3B shows that all the setae have sub-branches (setulae) that extend at an outward, dorso-lateral angle with respect to the surface of the GS. The particles found within the LSG are sub-micrometer in diameter.

SEM photomicrographs of the morphology of GS structures are shown in Fig. 4 for juvenile P. vannamei (Fig. 4A–I) and broodstock P. monodon (Fig. 4J–O). In Fig. 4A, the PSC of P. vannamei has been cut open and the setae-covered walls of the dorsal (PSCD) and ventral (PSCV) chambers have been folded back to reveal the underlying GS as an elongated, inverted-V-like dome arising medially from the bottom of the PSCV at its posterior end. In Fig. 4B, the carpet-like, overlapping layers of setae on the GS surface can be seen and are progressively enlarged in Fig. 4C–E to clearly reveal the dorso-lateral branches (setulae) that were seen near the limit of light microscope resolution in the semi-thin sections in Fig. 3C. The space between these secondary setulae was approximately 0.2–0.7 µm (Fig. 3D).

In Fig. 4F, the rows of setae along the midline crest of the GS (cf. Fig. 3A by light microscopy) can be seen from the left side only, with the matching structures on the right side of the GS in shadow. They consist of two distinctive types. Both have secondary setulae, but one setal type has hooked ends (Fig. 4H) and the other not. In reference to Fig. 3A, the hooked setae (Fig. 4H) arise directly from the upper surface of the GS on either side of the midline, while the non-hooked ones (Fig. 4G) arise from the enlarged end of the smooth extension from the GS midline. At the posterior end of the GS, there is a spearhead-shaped PMP (Fig. 4I) with long setae, presumed to direct the flow of GS-excluded particles into the MG (see above). Similar structures to these were seen in SEM photomicrographs of P. monodon (i.e. branching setae covering the GS in Fig. 4K–N and a PMP in Fig. 4J). One difference from P. vannamei was that the tips of the primary setae in P. monodon were not straight but looped back and enmeshed in the setal mat.

Details of the LSG by SEM (Fig. 4M–O) are shown for P. monodon only, but they are similar in P. vannamei. In Fig. 4M, the GS was cut open to reveal the LSG and the overlying layers of setae, portions of which are enlarged in Fig. 4N and O. These images can be compared with the photomicrographs from frozen tissue sections (Fig. 2) and semi-thin sections (Fig. 3). The transverse section of the GS passing through the mid-part shows overlapping rows of primary setae aligned to form the surface of the GS (Fig. 4M,N). They arise from single triangular heads (triangular pyramid, TP; Fig. 4O) attached to the ‘setal core’ (Co; Fig. 4O) that consists of fused setae, together forming the dorsal wall of one LSG and the ventral wall of another (Fig. 4N, see also the semi-thin sections in Fig. 3B,C). The TP is also lined with setulae along its length, similar to the elongated setal arms that extend from the TP (Fig. 4O).

From all of the information above, we propose a model for food transportation through the GS, as illustrated in Fig. 5. The arrangement and stacking of the primary setae (PS) and its secondary setulae (SS) show that the gaps between the PS and SS are approximately 1.5–1.9 µm and 0.2–0.7 µm, respectively. The green sphere represents a food particle capable of passing through the PS and SS gaps to collect in the LSG and move in a posterior direction towards the HP.

Microbead fluorescence could be observed through the translucent tissue from the mouth to the anus (Fig. 6). The signal could be seen in the ASC and PSC 3, 6 and 10 min after feeding (Fig. 6A–C). The signal was also visible in the surrounding areas from 6 min (Fig. 6B, arrow) to 10 min (Fig. 6C), indicating that the microbeads had also moved into the HP. Fluorescence was initially detected in the area of the MG at 6 min (Fig. 6B) and strongly detected at 10 min (Fig. 6C, arrow). By 100 min after feeding, fluorescence had decreased in the ASC, PSC and HP, and was most prominent in the MG and HG regions (Fig. 6D, arrows). The control group fed with normal food pellets did not show any fluorescence in the gastrointestinal tract (Fig. 6E).

After shrimp were fed for 5 min with the 0.1 µm fluorescent microbeads, a fluorescence signal was found in the dorsal and ventral chambers of the PSC (Fig. 7A,B). At higher magnification, fluorescence was detected in the PSCV in the narrow space between the setae-covered wall of the PSCV and the GS surface (Fig. 7B′) and also in a hepatopancreatic duct (Fig. 7C). Fluorescence was also seen in HP tubule lumens (Fig. 7D,F,G) and within the cytoplasm of HP tubule epithelial cells (Fig. 7E).

By contrast, fluorescence from the 1 µm microbeads was found only in the PSCD lumen and in the PSCV luminal space between the stomach wall and GS surface (Fig. 8A,B). No fluorescence was seen in the LSG, in the HP tubule lumens or in the cytoplasm of HP tubule epithelial cells (Fig. 8C,D). Nor was there any microbead fluorescence in the HP, either in the tubule lumens or in the tubule epithelial cells (Fig. 8E,F). The control shrimp fed with normal pellets showed a degree of autofluorescence from the feed in the PSCV (Fig. 8G), but none in the LSG or HP (Fig. 8H,I).

The GS of penaeid shrimp is located at the base of the PSC and is connected to the HP via posterio-lateral gastro-hepatopancreatic ducts. The GS is a non-living, V-shaped, dome-like, cuticular structure covered with comb-like rows of setae with dorso-lateral branches that arise at an approximate right-angle from fused setae that form the walls of the LSG of the GS. Food particles larger than 1 µm cannot pass through the setal layers into the LSG and are instead transported directly to the MG via the PMP. In contrast, food particles substantially smaller than 1 µm move down through the setal matt and enter the LSG to be transported to the HP via the hepatopancreatic ducts (Fig. 5).

Our results for the structure of the penaeid shrimp ASC, PSC and GS are similar to those previously reported for P. japonicus (Lin, 1996) and for other crustacean species, including the spider crab, Notomithrax ursus (Woods, 1995), the Brazilian native prawn, M. carcinus (Lima et al., 2016) and three Antarctic shrimp species (N. lanceopes, N. antarcticus and C. antarcticus) (Storch et al., 2001). Most of these reports focused mainly on the relationship between the structure and dietary habits of each species. However, the overall GS structure in these various decapods was very similar even though their diets differed, as observed earlier for three species of southern African decapods (Schaefer, 1970).

Despite these earlier reports, little work has been done on the size exclusion limit (SEL) for GS conservation over the life span of crustaceans, particularly the important aquatic penaeid species, i.e. P. vannamei. For example, the width of setae covering the GS was proposed to be approximately 1 µm in adult P. japonicus (Lin, 1996). Recently, the investigation of Wade et al. (2018) reported that the GS in sub-adult P. monodon allowed particles below 400 nm to pass through into the HP lobules. However, larger molecules (>3 µm) were excluded by the GS and consequently excreted within 1 h. Also, the pore size of the GS of the spiny lobster (S. verreauxi) was found to be smaller than 1 µm (Simon et al., 2012). These pieces of evidence support our finding that the pore size of GS in two penaeid shrimps was consistent within and even across the species. Here, our results using SEM and fluorescent microbeads support our hypothesis that the SEL of the GS of both species is substantially less than 1 µm, which may explain why whole bacterial cells are rarely present in the lumen of hepatopancreatic tubules in healthy shrimp. This finding has important implications for preparation of shrimp feeds that contain microencapsulated nutritional additives such as vitamins, which have to escape the digestive process in the stomach and enter the HP as quickly as possible. Our results suggest that the size of these particles should be substantially under 1 µm or that they should be designed to enter the PSCV and disintegrate there to enter the HP. Otherwise, larger feed particles would have to be reduced to an appropriate size by the stomach or risk being directed to the MG.

A previous study (Simon et al., 2012) reported that the SEL of the GS of different larval development stages (well developed at stages 3–4) and of juvenile lobster, S. verreauxi, remains morphologically unchanged, meaning that once the GS is fully formed, it has a constant filtering capability in healthy shrimp. Interestingly, Gophen and Geller (1984) also found that, by using spherical plastic beads with bacteria, the SEL of the GS of the crustacean Daphnia spp. was determined to be approximately 0.4–0.7 µm diameter. This supports our hypothesis that the GS can exclude whole bacteria from the HP.

In the digestive mechanism of crustaceans, movement of the stomach chambers is controlled by a nerve plexus for muscles that allow squeezing, cutting and grinding. The resulting small amplitude pulsating movement in the PSC results in a ‘pumping’ movement that propels food into the MG (McGaw and Curtis, 2013). The muscle attached to the external lateral wall of the PSC is a key modulator of the GS filtering function. Relaxation and contraction of this muscle result in fine particles and liquids moving into the space between the setae-covered surfaces of the PSCV wall and GS. When the two surfaces are pressed together, liquids and particles passing through are squeezed through the setal mat into the LSG (Schaefer, 1970). The action is facilitated by the flexible cuticle lining of the PSCV and its muscle control, but the precise mechanism remains unclear. For example, our frozen cross-sections of the GS of P. monodon and P. vannamei show relatively large spaces behind the chitinous wall of the PVSC and under the layer of LSG in the GS. Bell and Lightner (1988), in reference to their photomicrograph of the PSCV, describe the space below the LGS as a preparation artifact. However, our frozen tissue sections suggest instead that these spaces are real. This contention is supported by Fig. 2C, where the active GS is separated by such a space from a replacement GS that has formed in preparation for the next molt. Thus, we hypothesize that these spaces are filled with gas and play an important role in the mechanical ‘pumping’ action necessary for the filtering function of the GS. Moreover, it is clear that the chitin-based structure of the GS is a complex, non-living cuticular structure and that its mechanical function does not require attachment to a living layer of underlying epidermal cells. The mechanics of the GS could inspire the construction of a novel sub-micrometer filter that can resist plugging.

The appropriate timing of food passage in juvenile shrimp from the mouth to the anus and the retention of digested food particles in the stomach and alimentary tract are essential for nutrient absorption with maximum efficiency. However, the transit time varies among Decapod species, and might be influenced by the size and type of consumed food, the size and activity of an animal, and any number of other environmental factors including salinity and temperature. A slow transit rate allows more time for nutrient absorption (McGaw and Curtis, 2013). In the adult shore crab, Carcinus maenas, the microbeads were retained for up to 2–3 weeks (Watts et al., 2014). However, our study found that the fluorescence in the HP rose and then fell by 100 min after feeding, suggesting that the microbeads ingested by the HP tubule epithelial cells of juvenile shrimp were very much reduced in signal. This is consistent with gut passage times of 1–2 h for other penaeid shrimp species (Beseres et al., 2005; Pollack et al., 2006). For sub-adult P. monodon, the signal of particles (≤100 µm) in the digestive tract emerged within 1 h and was cleared within 5 h (Wade et al., 2018). Aside from being expelled in feces, there are other possible reasons for the loss of fluorescence from the HP, especially from the cytoplasm of tubule epithelial cells. One is possible translocation from HP cells to the hemolymph and then to other parts of the body such as the ovaries and gills, as reported in C. maenas (Farrell and Nelson, 2013; Watts et al., 2014). Alternatively, they might be fragmented into even smaller micro- or nano-particles as reported for millimeter diameter microplastic beads in Antarctic krill (Dawson et al., 2018). It is also possible that engulfed particles may end up in feces after ejection from the engulfing cells or from sloughing of those cells, as has been reported for B-cells that accumulate indigestible materials in vacuoles before they slough into the tubule lumen, to be replaced by new epithelial cells arising in the E-cell region of the HP (Hopkin and Nott, 1980). This is consistent with the report that 0.1 µm microspheres fed to larval lobster were discarded approximately 18–24 h after feeding, during B-cell extrusion. Similarly, B-cells were found to be released after 18 h in juvenile lobster (Simon, 2009). However, there was a small detectable signal in the ASC, PSC and HP at 100 min after feeding, suggesting that at least some of the microbeads might have been trapped inside these complex structures for longer periods of time. Such a long duration of particle retention agrees with a previous finding that small indigestible particles that migrated through into the digestive gland could stay there for up to 24 h (Wade et al., 2018).

In conclusion, understanding the morphology, organization and SEL of the GS in penaeid shrimps has twofold benefits. Firstly, it indicates the involvement of the GS in preventing potentially pathogenic bacteria from entering the HP. However, some bacteria in the stomach can produce toxins that are capable of passing through an intact GS to damage or destroy viral organs (Lightner, 1988; Aguirre-Guzmán et al., 2004). A good current example is acute hepatopancreatic necrosis disease caused by isolates of Vibrio species that colonize the shrimp stomach, where they produce Pir-like toxins that pass through the GS and damage HP tubule epithelial cells (Thitamadee et al., 2016). Alternatively, bacteria in the stomach may produce hydrolytic enzymes (Tzuc et al., 2014) that may compromise the GS filtering capacity, facilitating entry of foreign materials into the HP. Secondly, knowledge that the GS is a sub-micrometer filter means that it may serve as a selective route for the introduction of nano-carriers into the shrimp HP and hemolymph for various purposes, ranging from nutrition to disease prevention and control.

We extend our gratitude to Dr Kallaya Sritunyalucksana-Dangtip, Dr Charoonroj Chotwiwatthanakun, Dr Jirawat Saetan and Dr Tatpong Tulyananda for their kindness and help. We would also like to thank Prof. T. W. Flegel (National Center for Genetic Engineering and Biotechnology - BIOTEC, National Science and Technology Development Agency - NSTDA) for assistance in preparing and editing the manuscript. Finally, we also thank the reviewers and editor for providing valuable comments and suggestions that improved the manuscript.