Hypertonic water reabsorption with a parallel-current system via the glandular and saccular renal tubules of Ruditapes philippinarum Free

Bivalves live in a very wide range of environmental osmolality, from freshwater to seawater. It is known that marine bivalves are euryhaline (Robertson, 1964), and accordingly, the osmolarity of the haemolymph of Mytilus is virtually in equilibrium with the external seawater (Robertson, 1953). Although the osmolality of the haemolymph of freshwater bivalves is higher than the osmolality of freshwater (Robertson, 1964), the osmolality of Anodonta cygnea haemolymph has remained at around 40 mOsm kg⁻¹ H₂O (Potts, 1954). A few brackish water bivalves are also known to use hyperosmotic regulation of the haemolymph to maintain a higher osmolality, compared with that of the environmental fluid. However, the exocrine system in the bivalves does not have a large ability for osmotic regulation of the body fluid (up to 100 mOsm kg⁻¹ H₂O) compared with the seawater (around 1000 mOsm kg⁻¹ H₂O) (Robertson, 1964; Burton, 1983). Extensively examined around the end of the 19th century, the excretory system in bivalves consists of the nephridium (kidney) and the heart, including the pericardium (Martin and Harrison, 1966; Bayne, 1976). Primary filtration occurs in the podocytes in the pericardial glands of the pericardium, or in the wall of the auricle (Morse, 1987; Andrews and Jennings, 1993). Absorption and secretion might be attributed to the kidney (Martin and Harrison, 1966). It has been reported that the kidney is the site of reabsorption of salts and water in bivalves living in both seawater and freshwater (Picken, 1937; Tiffany, 1974). However, there have only been a few studies conducted on the function of the renal tubules and their relationship to osmotic regulation (Robertson, 1964; Seo et al., 2021).

The anatomical structure of the kidney varies in bivalves (Andrews, 1988; Andrews and Jennings, 1993; Mackie, 1984; Martin and Harrison, 1966). For example, Mytilus has a straight-sac-shaped kidney with a single layer of epithelial cells, which produces an almost isotonic urine compared with seawater. The function of the U-shaped kidney in marine bivalves has not been well documented to date. The U-shaped kidneys of Anodonta and Nodularia, both freshwater mussels, produce slightly hypotonic urine compared with the haemolymph, using counter-current renal tubules (Andrews, 1988; Burton, 1983; Robertson, 1953; Wakashin et al., 2018; Seo et al., 2021). Clams such as Mercenaria mercenaria and Meretrix lusoria have two types of renal tubules, the glandular tubule (GT) and the saccular tubule (ST). The epithelium is well developed compared with those of Mytilus and Nodularia (Andrews and Jennings, 1993; MacKenzie et al., 2002; Ikeda, 2012). The clam Ruditapes philippinarum can live in brackish water in a range of salinity from 20 to 30‰. This species has a fairly high tolerance to changes in salinity (Carregosa et al., 2014; Elston et al., 2003; Matsuda et al., 2008). It can be considered that the kidney of the clam must play an important role in the osmotic regulation of the haemolymph. As far as we know, no in vivo studies have been performed on the function of the tubules in osmotic regulation in R. philippinarum. We supposed that the proximal GT and the distal ST comprise a counter-current system that produces hypoosmotic urine, in order to maintain the osmolality of the haemolymph. Owing to the need to cover the wide range of salinity that occurs in brackish water, it is expected that the capacity of the renal system in R. philippinarum will be higher than that of the freshwater bivalves. In order to test this hypothesis, we examined the anatomical structure of the excretory system, including the kidney and the heart, using Haematoxylin & Eosin (HE) stained slices, three-dimensional T₁-weighted magnetic resonance imaging (3D T_1w-MRI) and X-ray micro computed tomography (micro-CT), and built a schema of the excretory system based on the anatomical results. The function of the renal tubules was studied by investigating the accumulation of manganese ion (Mn²⁺) in the clam immersed in seawater containing Mn²⁺. We measured T_1w-MRI signals and the T₁ relaxation rate (1/T₁=R₁) of the renal tubules in the clam, and evaluated the Mn²⁺ concentration using the R₁ of the urine in the renal tubules. We also examined changes in the intratubular Mn²⁺ concentration using 3D T_1w-MRI. In addition, we measured the haemolymph flow in the vessels and renal tubules, and estimated the molecular size of the filtration that occurs in the kidney. From these results, we have added functional information on the schema, and proposed a new model for the production of hypotonic urine in the clam by a parallel-current system consisting of the GT and the ST. Finally, we tested the parallel-current system using clams exposed in seawater in a range of salinity from 21.0 to 41.0‰. From the increase of R₁ shown in the ST, we confirmed the increase in osmolarity in the ST by the parallel-current system in seawater over a hypoosmotic range.

The Ruditapes philippinarum (A. Adams & Reeve 1850) used in this study were supplied by Marue-Suisan Co., Ltd (Tahara, Aichi, Japan). These clams were cultivated using the on-bottom culture method on a tidal flat in Mikawa Bay, Aichi, and collected in April, June, September and October 2023. The temperature and salinity of the seawater measured 1 m from the bottom on 8 April, 22 April, 3 June, 23 September and 21 October were 14°C/32‰, 16°C/32‰, 18°C/31‰, 27°C/32‰ and 22°C/31‰, respectively (Aichi Fisheries Research Institute, 2024). After collection, the clams were kept in seawater at 15–20°C/33.4‰ in an indoor circulating tank for 2–3 days. Then, the clams were packed in a box without seawater, cooled by ice, and transported to the laboratory using a refrigerated transport service, maintained at 10°C, within a time period of 16 h (Cool Ta-Q-BIN, Yamato Transport Co., Ltd, Tokyo, Japan). At the laboratory, the clams were stored in a refrigerator at 10°C. At 6 to 12 h before the experiments, 10–20 clams were taken out from the refrigerator, and kept in a 5 liter bath in aerated synthetic seawater (salinity 36‰) at room temperature (20°C), and the clams were used for experiments within the same day. All of the in vivo experiments finished within 3 days for each batch of clams. The synthetic seawater was made by dissolving a synthetic seawater mixture (Marine Merit), containing NaCl, KCl, MgCl₂, Na₂SO₄, SrCl₂, CaCl₂, pH buffer and trace elements from seawater, supplied by Matsuda Inc. (Osaka, Japan) with distilled water, and the salinity was adjusted with distilled water. A total of 120 clams were used in this MRI study. The length of the mussels was 31.4±1.2 mm (mean±s.d.). All of the animal experiments in this study were carried out under the rules and regulations of the Guiding Principles for the Care and Use of Animals set by the Physiological Society of Japan, and approved by the Animal Research Councils at Dokkyo University School of Medicine.

The MRI examinations of R. philippinarum conducted in this study were performed using procedures noted in a previous report (Seo et al., 2014a, 2016). In brief, the clams were placed in a plastic tube (inner diameter of 22.5 mm), and each clam was positioned in place using a piece of elastic silicone strip, which was inserted at the hinge position of the shell. The clams were immersed in 15 ml of synthetic seawater without aeration, and the temperature was kept at 20°C. Seawater was exchangeable through another tube set in the bottom of the tube holding the clam. The ¹H MR images were obtained by a magnetic resonance imaging system (AVANCE III, Bruker Biospin, Ettlingen, Baden-Württemberg, Germany) with ParaVison operating software (version 5.1) and a 7 T vertical superconducting magnet equipped with an active shielded gradient (micro2.5) and a 25-mm ¹H linear radiofrequency coil.

The heartbeat and flow of the haemolymph were imaged by the inflow effect of T_1w-MRI (Bock et al., 2001; Seo et al., 2014b). The typical transverse imaging parameters were as follows: 24×24 mm field of view (FOV) with a 64×32 data matrix and a slice thickness of 1 mm, 10 ms relaxation delay (T_R), 3.5 ms echo-time (T_E), 22.5 deg flip angle (FA) and 1 accumulation. Each MRI measurement session consisted of 128 images obtained every 0.32 s. Sagittal images were obtained every 0.64 s by 48×24 mm FOV with a 128×64 data matrix and a slice thickness of 1 mm. Image intensity increases proportionally to [1−exp(−T_Rkv)], where v is the flow velocity and k is a constant that depends on the flip angle and the T₁ of the haemolymph. Therefore, an increase in the T_1w-MR image intensity is roughly proportional to the flow velocity (Seo et al., 2014b). Image intensity is presented as relative intensity compared with that of the seawater.

The direction and velocity were measured by phase-contrast gradient echo sequences (PC-MRI) (Lotz et al., 2002), using transverse slices with a voxel resolution of 100×100 µm and a slice thickness of 1 mm with a combination of T_R=100 ms, T_E=8 ms and FA=45 deg. Eight pairs of velocity encoding gradients were used, with a strength corresponding to a velocity from −15 to 15 mm s⁻¹ with a 3.8 mm s⁻¹ step, and a total image acquisition time of 10 min.

T₁ relaxation time was measured by a two-dimensional saturation–recovery imaging method with five relaxation delays from 0.1 to 4 s. The pixel size was 190×190 µm and the slice thickness was 1 mm. The total image acquisition time was 9 min 21 s.

The whole structure of the kidney, heart and vessels was measured by 3D T_1w-MRI. The typical parameters used were a combination of T_R=50 ms, T_E=2.5 ms and FA=45 deg, the total image acquisition time was 7 min 40 s, and image voxel size was 180×180×180 µm. A high-resolution in vivo 3D T_1w-MRI was obtained under MgCl₂ anaesthesia, using combinations of T_R=50 or 75 ms, T_E=3 ms and FA=45 deg, a total image acquisition time of 4 or 6 h, and an image voxel size of 90×90×90 µm. In the histological examinations, high-resolution 3D T_1w-MRI was measured with a combination of T_R=100 ms, T_E=4 ms and FA=45 deg, and an image voxel size of 60×60×60 µm. The T_1w-MR image intensity is proportional to [1−exp(−T_RR₁)]/[1−cos(FA)exp(−T_RR₁)]. The image intensity was compared with that of a reference containing 0.5 mmol l⁻¹ MnCl₂ solution, where R₁ is the T₁ relaxation rate (R₁=1/T₁). The R₁ value depends on the concentration of Mn²⁺ ([Mn²⁺]): R₁=R₀+K[Mn²⁺], where R₀ and K are the intrinsic R₁ of the urine and the relaxivity value of Mn²⁺ (6.3 s⁻¹ mmol⁻¹ l), respecitvely. Therefore, any increase in the T_1w-MR image intensity is roughly proportional to K[Mn²⁺] (Wakashin et al., 2018). Maximum intensity projection (MIP) and 3D reconstruction were obtained by a 3D analysis package in ParaVison software.

Experiments on the accumulation of Mn²⁺ in the kidney started from the immersion in seawater containing 1 to 50 µmol l⁻¹ MnCl₂ (nacali, Kyoto, Japan) for 2–22 h. Then, the clams were anaesthetised using 4% MgCl₂ in order to minimize motion artifacts, and set in the tube for the MR experiments. One set of MRI experiments consisted of (1) setting of the orientation of clams by scout T_1w-MR images, and the adjustment of the MRI parameters (10 min), (2) measurement of the heart rate (5 min), (3) measurement of the T₁ relaxation time of the kidney in two slices (20 min), and (4) measurement of the whole structure in the clam by 3D T_1w-MRI (10 min). When necessary, the flow of the haemolymph was measured by PC-MRI (5–20 min). The measurements were usually finished within 1 h, which can minimize the effects of the anaerobic conditions. In nine clams, we also obtained high-resolution in vivo 3D T_1w-MRI images thereafter.

In order to estimate the molecular size of the filtration that occurs in the kidney, four tracers were used: gadolinium-diethylenetriamine pentaacetic acid (GdDTPA; 0.55 kDa, Magnevist, Schering, Berlin, Germany), CH3-DTPA-Gd (2 kDa, NMS-60, Nihon Medi-Physics, Chiba, Japan), and two polylysine Gd conjugates (22 and 225 kDa, PLL-Gd, BioPAL, Worcester, MA, USA). Under MgCl₂ anaesthesia, 0.1 ml of 1 or 2 mmol l⁻¹ MR tracers were injected into the foot of the clam using a 30 G needle (Nipro, Osaka, Japan). The clams were immersed in normal seawater for 15 min, and then were anaesthetised again for the MRI experiments. Injection of the MR tracers was confirmed by noting a higher signal intensity of the visceral mass in scout T_1w-MR images. One set of the MRI experiments was the same as the Mn²⁺ accumulation experiments.

The acute effects of the salinity of the seawater on Mn²⁺ accumulation in the kidney and heart rate started from the time of immersion in seawater at a salinity of 21.0 to 41.0‰ at 20°C for 6 to 12 h. The clams were then immersed in the seawater containing 30 µmol l⁻¹ MnCl₂ for 2–7 h. Finally, the clams were anaesthetised by 4% MgCl₂ in order to minimize motion artifacts, and set in the tube for the MR experiments. One set of MRI experiments consisted of (1) setting of the orientation of the clams by scout T_1w-MR images, and the adjustment of the MRI parameters (10 min), (2) measurement of the heart rate (5 min), (3) measurement of the T₁ relaxation time of the kidney in two slices (20 min), and (4) measurement of the whole structure in the clam by 3D T_1w-MRI (10 min).

Anatomical information of the clam was obtained using a micro-CT imager (Rigaku R-mCT2, Rigaku, Tokyo, Japan). The clams were fixed with 4% paraformaldehyde (PFA) and immersed in 3% iodine potassium iodide (I₂KI) for more than 2 weeks (Sakurai and Ikeda, 2019). Then, after washing with pure water for a day, the clams were imaged in pure water with a source voltage of 90 kV and a source current of 160 µA. The 3D CT images were FOV=5×5×5 mm with a voxel size of 10×10×10 µm, and a total image acquisition time of 3 min.

The clams were fixed with PFA for the histological examinations, and the shells were decalcified using 0.5 mol l⁻¹ EDTA. Then, tissues were embedded in paraffin wax after dehydration. The paraffin sections were prepared using a slice thickness of 10 µm. The sections were stained with Haematoxylin & Eosin (HE). Images were detected using a BZ-X710 microscope in image-stitching mode using the BZ-X viewer and BZ-X Analyzer software (Keyence, Osaka, Japan), and a BX63 microscope in image-stitching mode using cellSens imaging software (Olympus, Tokyo, Japan). Low magnitude images were detected using a stereomicroscope (SZX16, Olympus) with a WRAYCAM-NOA2000 digital camera operated by MicroStudio software (WRAYMER, Osaka, Japan).

Observed values were presented as means±s.d. with the number of samples (n). Significant differences between two means were tested using a two-sample two-sided t-test employing Excel 2021 (Microsoft, Redmond, WA, USA). A statistical difference was detected by employing one-way ANOVA for more than three groups, and we also detected differences compared with the control using Dunnett's post hoc test in Easy R Statistics software (EZR v2.9-2) with R (The R Foundation for Statistical Computing Platform, v.4.3.2) (Kanda, 2013). P-values less than 0.05 were regarded as significant. The sample size was estimated using Easy R Statistics software (EZR v2.9-2). In cases where the differences were 2.0, 1.5 and 1.0 s.d., the minimum sample sizes were 4, 7 and 16, respectively, in order to discriminate between the two groups with a two-sided α error of 0.05 and a power of 0.8.

The anatomy of R. philippinarum was observed using MRI, micro-CT and light microscopy. The position of the kidneys is shown in maximum intensity projection (MIP) images of 3D T_1w-MRI of an Mn²⁺ incubated clam (Fig. 1A,B). Kidneys are depicted as a high signal intensity owing to the accumulation of Mn²⁺. The kidneys were located between the heart and the posterior adductor muscle (PAM). The anterior end of the kidneys was located on the ventral side of the heart, and the dorsal side of the foot. The rectum passed through the dorsal side of the kidneys. The foot and the posterior retractor muscles supported the ventral side of the kidney. Please see Movie 1 for the positions of the kidneys, digestive glands and the rectum. As shown in the 3D constructed images (Fig. 1C), the shape of the kidney wall was lobular. The total volume of the kidney incubated in 30 µmol l⁻¹ Mn²⁺ was 30.1±7.1 µl as estimated from nine clams collected in April and June 2023.

The important anatomical elements of the exocrine system were summarized in the schema of a coronal oblique section (Fig. 2). The kidney consists of two types of tubules: the glandular tubule (GT) and the saccular tubule (ST) (Fig. 3A). The epithelial cells of the ST were cuboidal, and the basement membrane lay on the outer surface of the ST facing the interstitial space (Fig. 3B). The anterior end of the ST connected with the auricle directly without any epithelium (Fig. 3C,G). Then, as shown in Fig. 3G, the ST formed a tubular structure, and the GT appeared inside of the ST at around the beginning of the visceral sinus (Fig. 3D,G, 0.38 mm). The intratubular space was enlarged in the middle part of the kidney, and the visceral sinus passed through between the right and left GTs (Figs 1A and 3A). Then, the ST merged with the GT at the posterior end of the kidney near the anterior side of the PAM (Fig. 3F). The posterior end of the ST was located on the dorsal side of the PAM (Fig. 1B). As far as we know, having searched in our histological studies, there are no outlets like an excretory pore.

The GT received haemolymph from the pedal sinus via the visceral sinus, and there was no Keber's valve between the sinuses (Fig. 3C,D,G). The GT runs in parallel to the inner side of the ST (Fig. 3A,G). The basement membrane lay on the inner surface of the GT, and the epithelial cells were cuboidal, facing the lumen of the ST (Fig. 3B). The intratubular space of the GT was enlarged in the middle part of the kidney, and merged with the ST at the posterior end of the kidney (Fig. 3F). In the middle part of the GT, the renal duct appeared inside of the GT (Fig. 3D,G), and ran to the anterior direction until the excretory pore opening in the mantle cavity (Fig. 3G, 0.17 mm). The inside of the renal duct was covered by well-developed cilia for the whole length of the duct (Fig. 3E).

The molecular size of filtration of the kidney was estimated using four MR tracers and 17 clams collected in June, September and October 2023. After immersion in 1 mmol l⁻¹ GdDTPA containing seawater, there was no detectable increase in the image intensity of the kidney (Fig. 4A). So, the clam could not take up GdDTPA from the seawater. When GdDTPA was injected in the interstitial space of the visceral tissue, the kidney was enhanced at a higher signal intensity owing to accumulation of GdDTPA (Fig. 4B). The kidney of the clam could concentrate CH3-DTPA-Gd (Fig. 4C), 22 kDa PLL-Gd and even as much as 225 kDa PLL-Gd (Fig. 4D). Increases in the R₁ of the kidney (around 1 s⁻¹) also supported the accumulation of the MR tracers.

In order to detect the concentration dependency of Mn²⁺ accumulation, the clams were immersed in seawater containing 1 to 50 µmol l⁻¹ MnCl₂. Using 86 kidneys of 43 clams collected in April and June 2023, the T₁ relaxation rate (R₁=1/T₁) of the ST near the posterior end of the kidney was measured 2 h after exposure to MnCl₂, because R₁ values reached a plateau level after 1 h of Mn²⁺ exposure in preliminary experiments. As shown in Fig. 5A, R₁ was Mn²⁺ concentration dependent and elevated, and significantly increased at a Mn²⁺ concentration of seawater ([Mn²⁺]_SW) of 3 µmol l⁻¹ and above (P<0.05), and tended to reach a plateau at over 30 µmol l⁻¹ (around 3 s⁻¹). [Mn²⁺] in the kidneys was estimated from the R₁ values (Fig. 5B). In a range of [Mn²⁺]_SW 3 µmol l⁻¹ and above, [Mn²⁺] was 11.5±7.9-fold higher (mean±s.d., n=58) in the kidney compared with the ambient [Mn²⁺]_SW. The estimated [Mn²⁺] was close to 400 µmol l⁻¹ at a [Mn²⁺]_SW of 30 µmol l⁻¹, and then the concentration rate tended to decrease at 50 µmol l⁻¹ Mn²⁺ in seawater.

In order to detect changes in Mn²⁺ concentration in the kidney, the in vivo high-resolution 3D T_1w-MR images were measured using nine clams collected in April and June 2023. A typical coronal-oblique image and transverse images after immersion of 30 µmol l⁻¹ Mn²⁺ are shown in Fig. 6. As shown in the coronal-oblique image, image intensity of the ST was higher than that of the GT, and the image intensity of the GT increased progressively from the anterior to posterior side of the kidney (see Movie 2). Typical changes in image intensities of the GT and ST after immersion in 30, 10 and 3 µmol l⁻¹ Mn²⁺ are shown in Fig. 7A–C. The statistical results of 3D-MR image intensity of the ST and GT in the four slices from 22 clams collected in April and June 2023 are shown in Table 1. The image intensity in the ST increased sharply in the anterior part of the ST in 30 and 10 µmol l⁻¹ Mn²⁺ (P<0.05). Then, the image intensity was almost constant until the posterior end of the ST. From R₁ values obtained in the same clam shown in Fig. 7A–C, the Mn²⁺ concentrations were estimated as 550, 100 and 20 µmol l⁻¹ for 30, 10 and 3 µmol l⁻¹ [Mn²⁺] in seawater, respectively. The image intensities in the GT were lower than those in the ST (P<0.05); they increased continuously, and reached a similar image intensity as the ST at the posterior end of the kidney (P>0.05) in all of the Mn²⁺ concentrations (Table 1).

In order to detect abrupt changes in the haemolymph flow during heartbeat, flow images were obtained using T_1w-MRI. The heart rate of the clam, with and without MgCl₂ anaesthesia, was 11.6±5.8 and 24.6±7.5 beats min⁻¹ (mean±s.d. of seven clams collected in April and June 2023), respectively. Typical changes in the MRI image intensity during heartbeats are shown in Fig. 8. The anatomical structure was presented by 3D T_1w-MRI images of the anaesthetized clams incubated with Mn²⁺ and MgCl₂ (Fig. 8A). Owing to the suppression of the heartbeat by anaesthesia, the sinuses were depicted in a negative contrast, compared with the digestive gland and the kidney, which were depicted as a higher signal intensity because of the accumulation of Mn²⁺. In the mid-sagittal slice of 2D T_1w-MRI of a clam without anaesthesia, the haemolymph flow into the slice plane from outside was detected as a higher signal intensity (Fig. 8B). The visceral sinus ran through the kidney positioned on the ventral side of the heart (Fig. 8A,B). Changes in the MR image intensity during heartbeats and the position of the regions of interest are shown in Fig. 8C. From the peak-to-peak interval, the heart rate was 25±4.9 beats min⁻¹ (mean±s.d., 15 heartbeat intervals). Peaks of image intensity of the anterior aorta corresponded to the ejection of the haemolymph in the systolic phase of the heart. Peaks and bottoms of image intensity of the auricle corresponded to the systolic and diastolic phase of the ventricle, respectively. The timing of the peaks of the pedal sinus, the visceral sinus and the auricles were almost synchronized with the peak of the aortic ejection. During the foot extension, flows of the aorta and the visceral sinus were maintained, and although the flow of the pedal sinus tended to decrease, it did not stop. The posterior-directed flow in the visceral sinus was confirmed by the PC-MRI, but no urinary flow (≥3.8 mm s⁻¹) was detected in the renal tubules (Fig. 8D).

The effect of salinity of the seawater on kidney function was estimated by Mn²⁺ accumulation in the kidney in seawater at a range of salinity from 21.0 to 41.0‰ at 20°C. Using 92 kidneys of 48 clams collected in April, June, September and October 2023, the T₁ relaxation rate (R₁=1/T₁) in the ST near the posterior end of the kidney and heart rate was measured. As shown in Fig. 9A, the R₁ showed a significant difference at a range of salinity from 21.0 to 41.0‰ (P<0.05) and furthermore, a significant difference including the control seawater, as shown by one-way ANOVA (P<0.05). Using Dunnett's post hoc test, R₁ significantly increased at a salinity range from 26.0 to 36.0‰ and 41.0‰, compared with the R₁ of the control seawater (P<0.05). As shown in Fig. 9B, no significant differences in heart rate were demonstrated at a range of salinity from 21.0 to 41.0‰ (P>0.05), and furthermore, no significant differences were shown when control seawater was included, as shown by one-way ANOVA (P>0.05).

Even though the anatomical structure of the kidney varies in the bivalves, as shown in fig. 12-120 in Ruppert et al. (2015), the kidney consists of a single renal tubule connecting from the inlet in the pericardium to the outlet of the excretory pore, and the direction of the urine in the renal tubule is one way from the inlet to the outlet (Andrews, 1988; Andrews and Jennings, 1993; Mackie, 1984; Martin and Harrison, 1966). The kidney of R. philippinarum has a different structure, which has never been reported in the literature. We found that it consists of two tubules running in parallel – the glandular tubule (GT) and the saccular tubule (ST) (Fig. 2) – with the inlets of the ST and the GT opening from the auricle and the visceral sinus, respectively. The GT ran parallel inside of the ST to the posterior direction, and the ST and the GT merged at the posterior end of the kidney. The renal duct divided in the middle of the GT, and ran in the anterior direction inside the GT, and the excretory pore opened in the mantle cavity (Fig. 2). As shown in Fig. 8C, the haemolymph flows in the pedal sinus and the visceral sinus were pulsative flows synchronized with the ventricular contraction. Therefore, it is not likely that the urine in the GT could flow in the opposite direction to the visceral sinus. It is also likely that the luminal fluid in the ST could not flow in the opposite direction to the auricle. We did not detect any luminal flow in the renal tubules by the PC-MRI (Fig. 8D), but changes in the Mn²⁺ concentration in the GT and ST supported the posterior-directed flows in these renal tubules (Fig. 7). There was no filtration epithelium in the beginning of the GT and ST (Fig. 3). Filtration was detected by using tracer molecules such as indigo sulfonate (0.34 kDa), which appeared in the urine in Anodonta, Cardium and Tellina (Martin, 1983). Furthermore, injected inulin (5.5 kDa) has been detected in the urine in Anodonta cygnea (Potts, 1954). More recently, the molecular weight cut-off (MWCO) for filtration by the auricular wall of M. galloprovincialis was estimated at around 0.5 kDa (Wakashin et al., 2019), and freshwater mussels have larger MWCO values, such as 22 kDa for Nodularia douglasiae (Seo et al., 2021). Compared with these MWCO values, 225 kDa is a molecular size too large for filtration. Therefore, the result of injected 225 kDa PLL-Gd concentrated in the kidney (Fig. 4D) must support the concept that the renal ducts received haemolymph without filtration from the auricle and the visceral sinus. One remaining question is how is the influx of haemocytes into the kidney is eliminated without filtration epithelium. The number of haemocytes of R. philippinarum is relatively small at around 500–2500 cells µl⁻¹ compared with bivalves filtrating by the auricular wall, such as Mytilus edulis (4.2±1.75×10⁶ cells µl⁻¹) (da Silva et al., 2008; Oubella et al., 1993; Renwrantz et al., 2013). The GT was shown to be divided perpendicularly from the visceral sinus, so that the haemocytes are eliminated by the plasma skimming effect of the flow in the visceral sinus (Schmid-Schönbein, 1996). Therefore, the well-developed cilia shown over the whole length of the inside of the renal duct could manage the haemocytes (Fig. 3E). As a working hypothesis, we originally supposed that the GT and the ST comprised a counter-current system like the U-shaped kidney. However, the GT and the ST formed a parallel-current system, and merged with each other at the posterior end of the kidney.

The function of the renal tubules was assessed by accumulation of Mn²⁺ in the clams immersed in seawater containing Mn²⁺. [Mn²⁺] was maximally concentrated 12-fold in the posterior end of the ST, compared with the ambient seawater, and the maximum [Mn²⁺] was close to 400 µmol l⁻¹ (Fig. 5B). These values are similar to those observed in M. galloprovincialis (Wakashin et al., 2018). The 400 µmol l⁻¹ value must be the upper limit owing to the toxic effects of Mn²⁺ that appeared with concentrations of more than 100 µmol l⁻¹ (Bayne, 1976). The M. galloprovincialis kidney is a long straight-sac almost the same longitudinal length of the gills, but showed almost the same [Mn²⁺] in the kidney even after transient changes owing to urine excretion (fig. 7 in Wakashin et al., 2018). However, in R. philippinarum, there was a static gradient of [Mn²⁺] in the renal tubules (Figs 6 and 7).

Based on the T_1w-MR image intensities, [Mn²⁺] in the ST increased sharply in the initial part of the ST, and was almost constant thereafter until the posterior end of the ST. In order for Mn²⁺ to be concentrated in the luminal fluid in the ST, one possibility is the secretion of Mn²⁺ by the epithelial cells into the ST. As shown in Fig. 3G, the luminal space in the initial part of the ST is narrow, so that it can build up Mn²⁺ efficiently. For this purpose, epithelial cells should be distributed uniformly on the whole wall of the ST. However, the epithelial cells were distributed on the wall facing the interstitial space or the pericardial cavity, not on the wall facing the auricle (Fig. 3G). In addition, as shown in Fig. 5, the fold changes of [Mn²⁺] were constant from 1 to 30 µmol l⁻¹ [Mn²⁺]_SW. This constant behaviour would not be expected in the case of a specific substrate transporter, and should show a decreased [Mn²⁺] less than the affinity of the Mn²⁺ transporter. It is also true that the four different MR tracers were concentrated in the ST even though those tracers have different chemical structures and properties (Fig. 4). Therefore, the epithelial cells must absorb water from the haemolymph that enters from the auricle, so that [Mn²⁺] was increased by 12-fold until the maximum level of the water reabsorption of the epithelial cells. This does not mean that the osmolality of the luminal fluid increased by 12-fold, because useful solutes such as amino acids and glucose should be reabsorbed by transporters, and by water following the osmosis. Furthermore, non-absorbed solutes such as waste molecules should remain in the luminal fluid, so it is expected that the osmolality of the luminal fluid in the ST would increase. The upper limit might be 100 mOsm kg⁻¹ H₂O, which can be generated by the active transport system using ATP (Guyton and Hall, 2006). The epithelial cells in the middle and posterior parts of the ST may have the same function, and function to maintain the hyperosmotic luminal fluid in the ST.

Based on the T_1w-MR image intensities, [Mn²⁺] in the GT increased linearly, and reached the same [Mn²⁺] level in the ST at the posterior end of the kidney (Fig. 7). The increased ratio depends on the length from the initial part of the GT, and does not depend on the [Mn²⁺] in the seawater. Therefore, this increase of [Mn²⁺] might not be due to active transport of Mn²⁺ by the epithelial cells of the GT. The apical side of the epithelial cells faced the lumen of the ST, and the basal side faced the inner lumen of the GT (Fig. 3B). This orientation is similar to the podocyte in the auricles and pericardial glands in M. edulis and A. cygnea (Andrews and Jennings, 1993; Morse, 1987). Therefore, it might be possible that a leaky epithelium would transport water from the GT to the ST. The driving force might be the osmotic gradient of the luminal fluid in the ST. If the epithelium could not transfer Mn²⁺, more than 90% of the water would be reabsorbed from the urine in the GT. However, it is likely that some fraction of Mn²⁺ and waste molecules also transfer from the ST to the GT via the epithelium of the GT owing to their concentration gradient.

We have postulated a parallel-current system model. This model can maintain or increase the osmolality of the haemolymph, and can excrete excess amounts of water in the urine, as follows (Fig. 10). (1) A small amount of the haemolymph enters the ST from the auricle, and in the initial part of the ST, water is reabsorbed by isotonic osmosis with absorption of solutes, so that the concentration of non-absorbed waste increases. As a result, osmotic pressure in the luminal fluid of the ST becomes higher than that of the circulating haemolymph. (2) A larger amount of the haemolymph enters the GT from the visceral sinus. (3) Water is transported by osmosis or filtered by the osmotic pressure gradient to the ST lumen via the epithelium of the GT. (4) The reabsorbed water from the GT to the ST will be reabsorbed to the interstitial space by the epithelium of the ST. The surface area of the epithelium of the ST and the volume of luminal fluid in the ST are much larger than those of the GT, so that the osmolality of the luminal fluid in the ST could be almost constant. As a result of reabsorption by the osmosis of a hypertonic fluid to the haemolymph from the ST, the osmolality of the haemolymph will be increased. (5) In regard to waste molecules, the concentration is increased in the initial part of the ST so that the waste molecules might be moved to the GT owing to the concentration gradient. (6) From the middle of the GT, the renal duct is divided. The osmolality of the luminal fluid in this part is lower than that of the ST. Thus, the urine must have a lower osmolality than the reabsorbed fluid from the ST. Therefore, urine could excrete any excess amount of water from the haemolymph. In order to confirm this model, it will be necessary to detect transporters and Na⁺/K⁺-ATPase in the epithelia of the ST and GT in a future study.

The R. philippinarum clam lives in brackish water, and has suffered from mass mortality in their culture areas owing to effects of low salinity after heavy rains and severe weather, such as typhoons (Elston et al., 2003; Matsuda et al., 2008; Carregosa et al., 2014). Even in low salinity seawater (15 to 10‰), R. philippinarum can survive for a few days. Indeed, the clam can maintain osmolality of the haemolymph higher than that of the seawater for 36 h (Matsuda et al., 2008). When the clam is exposed to a lower salinity seawater, water will enter into the clam body owing to the osmotic gradient, and that should dilute the haemolymph. Therefore, it is necessary to excrete an excess amount of water from the haemolymph. This is similar to freshwater bivalves, such as Anodonta cygnaea (Picken, 1937), which produce hypoosmotic urine, which can be considered to function in a manner similar to the haemolymph in the clams we studied. However, there is a lack of information and experimental data on urine in marine and brackish water bivalves (Burton, 1983). Therefore, in order to test this hypothesis, the acute effect on the R₁ of the ST was observed in a range of salinity from 21.0 to 41.0‰. As shown in Fig. 9A, the R₁ of the ST significantly increased at the lower side of salinity from 26.0 to 36.0‰, compared with the R₁ of the control seawater (P<0.05), and the R₁ tended to decrease to that of the control seawater in the higher salinity range, exceeding 38.5‰. Because there were no significant differences in the heart rate (Fig. 9B), the observed changes in R₁ might reflect a response in the renal tubule to changes in the salinity. Therefore, the proposed parallel-current system of the kidney would certainly control osmolarity in the ST depending on the salinity of the seawater, and may concentrate haemolymph under lower salinity conditions. Future studies are required to conduct direct measurements of the osmolarity of the urine in the ST (Picken, 1937), and studies on the electrolytic composition of the urine might be useful to identify transporters working to control the osmolarity in the ST using the parallel-current system against the low salinity of seawater.

In summary, at the beginning of this study, we supposed that the proximal GT and the distal ST consisted of a counter-current system, which would make hypoosmotic urine to maintain the osmolality of the haemolymph. The results of this experiment denied that anatomical working hypothesis. We found that the ST and GT form a parallel-current system. The ST creates an osmotic pressure gradient for reabsorbing hypertonic water from the GT. The renal duct divides from the middle of the GT, and excretes any excess amount of water. As a result, the clam could maintain the osmolality of the haemolymph, even in low salinity seawater. The larger epithelial surface area and volume of the ST suggest a higher and more stable transportability of the hypertonic water than the U-shaped counter-current tubules in Nodularia douglasiae (Seo et al., 2021). The parallel-current system is a new renal system, compared with the single-sac-shaped kidney seen in the marine bivalves and the U-shaped kidney in the freshwater bivalves, and might be a specialized renal tubular system for the bivalves that live in brackish water.

The authors express their sincere thanks to Mr D. Takahashi (Marue Suisan) for providing detailed information on the clam, and Dr Y. Kamei, Ms M. Saida, Ms C. Ichikawa and Ms M. Asao (Optics and Imaging Facility, Trans-Scale Biology Center, National Institute for Basic Biology) for their technical support related to the micro-CT and light microscopes. We also thank Dr D. Gross, Dr V. Lehman and Dr T. Oerther (Bruker Biospin), as well as Ms Y. Imaizumi-Ohashi and Ms M. Yokoi-Hayakawa (Dokkyo Medical University) for their technical assistance.