Insect-Like Characteristics of the Malpighian Tubules of a Non-Insect: Fluid Secretion in the Centipede Lithobius Forficatus (Myriapoda: Chilopoda)

Since the pioneer work of Ramsay (1954) on isolated insect Malpighian tubules, the transport mechanisms of the Malpighian tubules of a number of insect species have been characterized (see reviews by Maddrell, 1980; Phillips, 1981; Bradley, 1985). Knowledge of transport characteristics of Malpighian tubules in other tracheates derives from a single study on a myriapod, the diplopod Glomeris marginata (Farquharson, 1974a,b,c). The differences from insect Malpighian tubules were attributed to evolutionary specialization (e.g. Phillips, 1981; Nicholls, 1985). To evaluate the specificity of fluid secretion mechanisms in insects and other tracheates, however, more data are needed.

This study describes the fluid secretion of the Malpighian tubules of another myriapod, the chilopod Lithobius forficatus L. Each of its two Malpighian tubules is almost twice as long as the animal itself (Rilling, 1968). Each consists of a single, uniform cell layer without a muscular sheath (Füller, 1966). Micropuncture studies on whole animals showed that the tubules release a potassium-enriched, hypertonic fluid into the rectum (Wenning, 1978, 1979).

Mechanisms of urine formation in L. forficatus were examined in isolated Malpighian tubules with respect to ion requirement, cellular and transepithelial potentials, dependence on external osmolality and pharmacology of transportactive drugs and insect diuretic factors. The results are discussed with respect to the specificity and evolution of fluid secretion mechanisms in the Malpighian tubules of tracheates. Some of the results have appeared in abstract form (Wenning, 1989).

Lithobius forficatus (average body mass 130 mg) were collected locally. The animals were kept in containers filled with leaves and moist soil in a room at 16 °C at high relative humidity and a light:dark cycle of 12h:12h.

Preparation of the tubules and flow measurements were carried out at room temperature (20–24 °C). Animals were decapitated and dissected in artificial haemolymph. The two Malpighian tubules were freed from the few tracheoles holding them in place and severed near the ampulla. To test for regional differences in the Malpighian tubules, all experiments were carried out on upper and lower tubule portions separately. Each tubule was cut into two portions - an ‘upper tubule’ (distal portion), containing crystals in its lumen, and a ‘lower tubule’ of wider diameter (proximal to the ampulla), containing clear fluid. Upper and lower tubules were of approximately equal length. Preparation time was 30–45 min.

The experimental design for flow measurements was that originally described by Ramsay (1954). In brief, the upstream end of the respective tubule portion was tied with superthin surgical thread (Leonhard Klein, Germany), submerged in a droplet (200 μl) of bathing medium in a wax-lined dish and covered with liquid paraffin. The open end was drawn into the oil. Fluid production was calculated by measuring the diameter of urine droplets with a calibrated eyepiece graticule. Fluid production is given in nanolitres per unit time and millimetres effective tubule length, i.e. the tubule part submerged in the bathing medium. In some experiments, the tubule fluid was analyzed for its osmotic and ionic concentration.

The composition of the artificial haemolymph was based on data of Wenning (1978): (mmol l⁻¹) Na⁺, 190; K⁺, 6; Cl⁻, 204; Ca²⁺, 2; Mg²*, 2; trehalose, 10; buffered with 10 mmol l⁻¹ Hepes (pH 7·0). Osmolality was 410mosmolkg⁻¹H₂O To test for its effects on urine production, the osmolality was decreased by diluting artificial haemolymph with Hepes buffer (10 mmol l⁻¹; pH 7·0) or increased by adding sucrose or NaCl Decreases in NaCl concentration without changes in osmolality were achieved by replacing NaCl with isotonic sucrose solution. [Na⁺] was decreased by replacing NaCl with choline chloride, keeping [Cl⁻] constant. Higher K⁺ concentrations were achieved at the expense of NaCl.

Osmolality was determined with a Clifton nanolitre osmometer (Clifton Technical Physics, New York, USA). The error was ±3mosmolkg⁻¹H₂O. Na⁺ and K⁺ concentrations were determined using an ultramicro double-beam flame spectrophotometer (W. Hampel, Frankfurt; developed by Müller, 1958; modified by Malnic et al. 1964). The error was ±3 mmol l⁻¹ Na⁺ and ±0.3 mmol l⁻¹ K⁺.

For electrophysiological recordings, a chamber of 1·6 ml was used. The downstream end of the tubule was placed in a Vaseline trough filled with oil. The reference electrode (chlorided silver wire) was in the bath. To determine the cellular potential (basolateral E_m), the cells were impaled with glass microelectrodes (35–45 M Ω, filled with a mixture of 4 mol l⁻¹ potassium acetate and 20mmoll⁻¹ KCl). To determine the transepithelial potential (E_TEP), a low-resistance (≈10 M Ω) glass microelectrode was compensated for its junction potential in the bath and then moved into the urine droplet.

Absolute flow rates among individual tubules varied. A control rate was established for each tubule by averaging the flow rate in artificial haemolymph measured for 60–120min at intervals of 15–20min. This rate was taken as 100%. The bathing medium was then exchanged for the test solution. The new flow rate was determined for 60–80min and expressed as a percentage of the control rate. To test whether effects were reversible, the bathing medium was replaced with artificial haemolymph and the flow rate was again determined for 60min. Values are given as mean ±S.E.M., with the number of tubules indicated for each set of experiments. Appropriate t-tests (paired, unpaired) were used to assess the significance of different treatments.

No time dependence of flow rate was observed (see Fig. 1), indicating that the difference between maintenance and experimental temperature (see above) did not affect the results.

Although the ultrastructure of the epithelium of the Malpighian tubules of L. forficatus is uniform (Füller, 1966), the upper and lower parts of the tubules differed in appearance owing to the presence of crystals in the lumen of the ‘upper’ tubule. The ‘lower’ tubule was more transparent. As in some insects (Maddrell and Phillips, 1975; Cooper et al. 1989), upper and lower portions exhibited differences in their physiological properties.

The outer diameter was about 50 μm at the proximal end of the tubule and widened to about 90 μm at the entrance to the ampulla. The flow rate in the lower part of a tubule was significantly higher than in the upper part (Fig. 1, P=0·022, paired t-test; see also Table 1). Isolated tubules continued to produce urine for 24 h.

Both upper and lower tubules secreted a potassium-enriched fluid but the K⁺ concentration was significantly higher in the fluid in the lower tubule. Na⁺ was the main cation and its concentration was similar in the fluid in both tubule portions (Fig. 2; Table 1). At a concentration of 170–180 mmol l⁻¹, Cl⁻ was the main anion in the fluid released from the ampullae in vivo (Wenning, 1978) and from isolated whole tubules (A. Wenning, unpublished results). The tubule fluid was hypertonic to the bathing medium in upper tubules by 9·4mosmolkg⁻¹ H₂O and in lower tubules by 12mosmolkg⁻¹ H₂O (Fig. 2). Apparently, little or no reabsorption occurred along the tubules as it does, for example, in Rhodnius prolixus (Maddrell and Phillips, 1975)..

Basolateral E_m was similar in upper and lower tubules (–39·2 ±1·8mV and –38·1 ±1·7 mV, respectively), while E_TEP was significantly more positive in lower than in upper tubules (±12.3 ±2mV versus ±5.5 ±0.8mV, P=0.006) (Fig. 2). The method used to measure E_TEP might diminish its true value, owing to leakage pathways (discussed by Aneshansley et al. 1989), although with tubule portions 10–14 mm long and with 2 mm between urine droplet and haemolymph, electrical isolation should be adequate.

Flow rate in isolated Malpighian tubules was inversely correlated with external osmolality (Fig. 3), but upper and lower tubules differed significantly in their sensitivity. With decreasing external osmolality, the flow rate in upper tubules was more enhanced than that in lower tubules. With increasing external osmolality, achieved by adding NaCl, the flow rate in upper tubules was unaffected, whereas it was greatly reduced in lower tubules (Fig. 3, squares). An increase in osmolality at constant ionic strength caused a similar reduction of flow rate in upper and lower tubules (Fig. 3. diamonds).

The increase of flow rate in dilute media was due to hypotonicity rather than to the concomitant decrease in ionic strength, because a decrease in external [K⁺] or [Na⁺] at constant osmolality resulted in a decrease in flow rate (Fig. 4A,B). At 50 mmol l⁻¹ Na⁺ and at constant osmolality, fluid production ceased in both tubule portions (Fig. 4A). However, a comparison of the decline of flow rates in upper and lower tubules (Fig. 4A) shows that upper tubules were less affected when [Na⁺] was lowered to 140–170 mmol l⁻¹. The strong inverse relationship between flow rate and osmolality in upper tubules (Fig. 3) seemed to be due to their relative insensitivity - compared with lower tubules - to a decrease in external Na⁺ concentration. This was demonstrated by lowering either [NaCl] or [Na⁺] (Fig. 4A). Lower tubules were more sensitive to decreases in Na⁺ concentration in the bathing medium (Fig. 4A) and the increase in flow rate with decreasing osmolality (Fig. 3) was therefore not as steep as that in upper tubules.

Although sensitive to hypertonicity at constant ionic strength (Fig. 3, diamonds), fluid production in upper tubules did not decrease with increasing external [NaCl] (Fig. 4A), indicating an antagonistic effect of the increases in NaCl and osmolality. In lower tubules, a decrease in flow rate occurred regardless of whether sucrose or NaCl was added (Fig. 4A), illustrating the marked effect of hypertonicity.

Lowering [K⁺] from 6 to 3 mmol l⁻¹ caused the fluid production to drop to 30% of the control rate in lower tubules; upper tubules were unaffected (Fig. 4B). Fluid production continued in nominally 0 mmol l⁻¹ K⁺. The increase in [K⁺] at the expense of [Na⁺], however, had little effect on fluid production in upper tubules, but decreased it in lower tubules (Fig. 4B). This was not due to the simultaneous decrease in external [Na⁺], since an increase in [KCl] with no change in [NaCl] had the same effect (not illustrated).

Furosemide, an inhibitor of Na⁺/K⁺/2Cl⁻ cotransport in a variety of transporting epithelia (reviewed by Greger, 1985), blocked fluid production completely in upper and lower tubules at 5 ×10⁻⁴moll⁻¹ (Fig. 5). When furosemide was removed, the flow rate did not return to its control value.

Ouabain, an inhibitor of Mg²⁺-dependent Na⁺/K⁺-ATPase, did not block fluid production completely (Fig. 6). Sensitivity differed in upper and lower tubules. At concentrations between 5 ×10⁻⁷ and 5 ×10⁻⁵mol l⁻¹, ouabain was ineffective in upper tubules and had only minor effects in lower tubules (Fig. 6, and unpublished data). At 5 ×10⁻⁵mol l⁻¹, fluid production was inhibited after wash-out in lower tubules and was stimulated in upper tubules (Fig. 6). At 10⁻³mol l⁻¹, ouabain clearly inhibited urine formation in lower tubules, whereas the effects on upper tubules varied (Fig. 6).

Dibutyryl cyclic AMP (dbcAMP), applied in combination with the phosphodiesterase blocker isobutyl methylxanthine (IBMX) (both at 10⁻³mol l⁻¹), caused a marked decrease in flow rate (Fig. 7), which returned to its normal value in upper tubules after wash-out.

Synthetic arginine-vasopressin-like diuretic hormone (DH) of Locusta migratoria (Proux et al. 1987) increased fluid production in isolated L. forficatus tubules at 10⁻⁷mol l⁻¹ (Fig. 8), which is 100 times the effective dose for L. migratoria (Proux et al. 1988). Locust DH at 10⁻⁸mol l⁻¹ was less effective (not shown). The response to peptides was, however, transient and lasted for 20min. The locust diuretic hormone (DH) is the antiparallel dimer of a nonapeptide (Fl) (Proux et al. 1987). Fl at 10⁻⁵mol l⁻¹ enhanced the flow rate in lower and upper tubules (Fig. 8).

The question of whether Na⁺ and/or K⁺ transport increase with increased fluid production was investigated by measuring Na⁺ and K⁺ concentrations in the tubule fluid before and after they had been stimulated either by locust DH or by lowering the osmolality (Table 1). Na⁺ and K⁺ secretion rates were calculated using flow rates determined simultaneously. Application of DH had no effect on Na⁺ and K⁺ concentrations of the tubule fluid in either portion, and consequently did not affect the Na⁺/K⁺ ratio. Absolute Na⁺ and K⁺ secretion rates increased significantly in the upper tubules, owing to the stimulation of fluid production. A slight increase was observed in the lower tubules. Upon lowering the external osmolality to 160 mosmol kg⁻¹ H₂O, Na⁺ and K⁺ concentrations in the tubule fluid decreased. The Na⁺/K⁺ ratio was unaffected in lower tubules but decreased significantly in upper tubules, indicating that K⁺ transport was stimulated more than Na⁺ transport in the upper tubules.

The luminal fluid in isolated upper and lower portions of the Malpighian tubules of the centipede L. forficatus is K⁺-enriched, slightly hypertonic and electropositive with respect to artificial haemolymph (Fig. 2; Table 1), indicating active urine formation. This is also true for most insects. In contrast, fluids produced by the Malpighian tubules of the millipede Glomeris marginata and the dragonfly Libellula quadrimaculata remain similar in composition to the bathing fluid over a wide range of external Na⁺ and K⁺ concentrations, and the transepithelial potential was reported to be 0 mV (Farquharson, 1974b; Nicholls, 1985).

It has been established that an apical cation pump is the driving force for urine formation in insect Malpighian tubules (Berridge, 1968). This pump enriches the tubule fluid with K⁺, which is the major intracellular cation, but can also transport Na⁺, as was first shown in a blood-sucking insect during diuresis (Gee, 1975). As a result of its ability to transport either Na⁺ or K⁺, depending on their respective intracellular concentrations, the pump was named the ‘common cation pump’ (Maddrell, 1977). Both Na⁺ and K⁺ are actively transported across most insect Malpighian tubules (see O’Donnell and Maddrell, 1984), generating a lumenpositive transepithelial potential (Sawyer and Beyenbach, 1985). Although the tubule fluid is enriched with K⁺ in the chilopod L. forficatus, K⁺ is not the prime mover of fluid secretion and an increase in [K⁺] does not enhance secretion (Fig. 4 A), as it does in most insects (Bradley, 1985; Morgan and Mordue, 1981; Baldrick et al. 1988; Van Kerkhove et al. 1989).

In contrast, Na⁺ is essential for fluid secretion in both upper and lower portions of isolated Malpighian tubules of L. forficatus (Fig. 4A), in the pill millipede (Farquharson, 1974b) and in a dragonfly (Nicholls, 1985), the only known example among insects. Elevated NaCl concentrations, however, affect only the lower tubules (Fig. 4A), where they cause fluid secretion to decrease. Since both portions are equally sensitive to media made hypertonic by addition of sucrose (Fig. 3), elevated Na⁺ concentrations might enhance fluid secretion in upper tubules, which is then counteracted by the inevitable increase in osmolality.

Electrochemical gradients for K⁺ and Na⁺ were calculated by using the Nernst equation. To determine the transepithelial electrochemical gradients, the measured concentrations of Na⁺ and K⁺ in the bath and in the tubule fluid were used. The basolateral and apical electrochemical gradients were estimated from assumed intracellular concentrations. Owing to the negative potential of the tubule cells, the electrochemical gradient for Na⁺ is directed into the cell and apical transport of Na⁺ must be active in L. forficatus. Paracellular Na⁺ transport as a major factor in urine formation - albeit possible in upper tubules (Fig. 2) - is unlikely, since furosemide blocks fluid secretion completely (Fig. 5, see also below). During diuresis, Na⁺ transport is stimulated more than K⁺ transport in the Malpighian tubules of blood-sucking insects (Gee, 1975; Maddrell, 1969; Williams et al. 1983; Williams and Beyenbach, 1983) and in the nephridia of the blood-sucking medicinal leech Hirudo medicinalis (Zerbst-Boroffka et al. 1982). Elevated rates of Na⁺ transport ensure efficient clearance of the major cation of the ingested meal. The rate of Na⁺ transport (without diminishing K⁺ transport) is also increased in other insect species during diuresis: twofold in a desert beetle (Nicolson and Hanrahan, 1986) and fivefold in the house cricket (Spring and Hazelton. 1987). stimulated with corpora cardiaca, brain and/or prothoracic ganglion homogenates. In L. forficatus, the Na⁺/K⁺ ratio does not change significantly in the presence of locust DH (Table 1), indicating that both ions are transported at the same rate under these conditions. When fluid production is stimulated by lowering the external osmolality, K⁺ secretion is enhanced more than Na⁺ secretion in upper, but not in lower, tubules. Apical electrochemical gradients for K⁺ were estimated by using the measured concentrations of the tubule fluid and assumed intracellular concentrations. Transcellular K⁺ transport is against the electrochemical gradient (Fig. 2), but - as calculated using the Nernst equation - apical K⁺ transport would be downhill when intracellular [K⁺] exceeds 85 mmol I⁻¹ in upper and 150 mmol l⁻¹ in lower tubules. Intracellular K⁺ concentrations between 95 and 140 mmol I⁻¹ have been measured in insect tubule cells (Baldrick et al. 1988).

The urine flow rate in upper tubules is largely unaffected by changes in Na⁺ concentration between 140 and 350 mmol l⁻¹ and K⁺ concentration between 3 and 21 mmol l⁻¹. Lower tubules secrete comparatively less fluid when the K⁺ concentration is varied (between 3 and 21 mmol l⁻¹) and have a more negative transepithelial potential difference with respect to the lumen (Fig. 2). This makes them more ‘insect-like’ than upper tubules, which resemble those parts of the Malpighian tubules that have been studied in the pill millipede (lower portions, see Farquharson, 1974a).

Ouabain, at a concentration of 10⁻⁵ mol l⁻¹, inhibits fluid production in isolated Malpighian tubules of the pill millipede (Farquharson, 1974b). It has no effect on fluid secretion in most insect species (Bradley, 1985), with the noteworthy exception of Drosophila hydei (Wessing et al. 1987). At high concentrations, it even stimulates urine flow in Rhodnius prolixus (Maddrell and Overton, 1988). In L. forficatus, ouabain has been tested at four different concentrations. Upper and lower tubules differ in their sensitivity to ouabain. Compared to their respective control periods, upper tubules have consistently higher flow rates than lower tubules (Fig. 6; and unpublished data). At 10⁻³mol l⁻¹, ouabain inhibits fluid production in lower tubules dramatically, although not completely. Its effect on upper tubules varies greatly.

An important mechanism for Na⁺ and Cl⁻ entry into the cells of many transporting epithelia is Na⁺/K⁺/2Cl⁻ cotransport, which is specifically blocked by furosemide and bumetanide. Furosemide inhibits fluid secretion in isolated Malpighian tubules of L. forficatus (Fig. 5), Rhodnius prolixus (O’Donnell and Maddrell, 1984), Drosophila hydei (Wessing et al. 1987) and Locusta migratoria (Baldrick et al. 1988). At high concentrations (10⁻³ mol l⁻¹), bumetanide (not tested in L. forficatus) inhibits fluid secretion by the Malpighian tubules of ants (Verhulst et al. 1988). In the centipede L. forficatus, the entry of Na⁺ and Cl⁻ into the epithelial cells might be by Na⁺/CU cotransport, since fluid secretion continues in nominally 0 mmol l⁻¹ K⁺ but ceases in 50 mmol l⁻¹ Na⁺. Na⁺/Cl⁻ cotransport was proposed by Nellans et al. (1973) and also suggested for Rhodnius prolixus (O’Donnell and Maddrell, 1984).

In insects, blood-borne diuretic and antidiuretic peptides modulate urine formation and urine flow of the Malpighian tubules (Maddrell, 1963; cf. Spring, 1990; Spring et al. 1988; Hayes et al. 1989). The arginine-vasopressin-like diuretic hormone from L. migratoria, which has been isolated and sequenced (Proux et al. 1987), doubles the initial rate of fluid secretion in isolated tubules of L. forficatus (Fig. 8), but the effect is transient. Whether and where a diuretic factor is produced is currently being investigated in L. forficatus.

At high concentrations (10⁻³mol l⁻¹), dbcAMP inhibits fluid secretion in L. forficatus (Fig. 7) and in isolated tubules of the pill millipede G. marginata, the other myriapod species studied (Farquharson, 1974a). At the same high concentrations, dbcAMP mimics the action of diuretic hormones on insect Malpighian tubules (Spring, 1990). However, the maximal effect of dbcAMP in mimicking the action of synthetic locust DH in locusts occurred at a concentration of 10⁻⁵ mol l⁻¹ and that of IBMX at 2 ×10⁻⁴mol l⁻¹ (Proux and Hérault, 1988). High concentrations - as used in most studies (see above) - might lead to a ‘run-away’ of cyclic AMP metabolism. As a consequence, tubule stimulation might be slowed down owing to a down-regulation of intracellular cyclic AMP concentration or to a negative effect on other second-messenger systems. In tubule cells, where intracellular cyclic AMP levels were high, fluid secretion increased only slightly (Proux, 1991).

Fluid secretion by insect and myriapod Malpighian tubules is inversely correlated with medium osmolality (Phillips, 1981; Bradley, 1985; Nicholls, 1985; Farquharson, 1974b; Fig. 3). The underlying mechanism, however, is not clear (Bradley, 1985), especially the contribution of the transport systems involved in cell volume regulation (see, for example, Lang et al. 1990). In L. forficatus, the inverse correlation between flow rate and osmolality is important for water regulation in the intact animal. The hindgut of L. forficatus is, unlike that of many terrestrial insects (Bradley, 1985), incapable of reabsorbing fluid hypotonic with respect to the haemolymph (Wenning, 1978, 1979). Under these circumstances, drastically reduced urine production is essential for water conservation. Upon dehydration, haemolymph osmolality cannot be held at its normal value (350–400 mosmol kg⁻¹ H₂O; Wenning, 1978; Riddle, 1985) and may increase to 800mosmolkg⁻¹ H₂O (Riddle, 1985). The limited ability to conserve water imposes ecological restrictions on activity (Rilling, 1968; Wenning, 1978; Riddle, 1985). In the intact animal, rehydration may be facilitated through uptake of water vapour by coxal glands located on the last four hindlegs (Rosenberg, 1983; Rosenberg and Bajorat, 1983).

Although it is unable to reabsorb hypotonic fluid, the rectum of L. forficatus selectively reabsorbs ions (Wenning, 1978). This capacity appears to be essential for ion homeostasis, since no net ion reabsorption occurs along the Malpighian tubules.

In conclusion, isolated Malpighian tubules of most insects produce a K⁺-enriched, hypertonic fluid and the transepithelial potential difference is positive with respect to the haemolymph. The same is true for the non-insect tracheate L. forficatus, a chilopod. The ability to enrich K⁺ in the urine appears to be a common feature of tracheate Malpighian tubules rather than being unique to ‘higher’ insects (discussed in Nicholls, 1985). Furthermore, the urine flow rate in insect tubules is controlled by diuretic factors and locust DH enhances fluid secretion in L. forficatus. Unlike that in the majority of insects (Bradley, 1985; Baldrick et al. 1988; Van Kerkhove et al. 1989), however, fluid secretion in the Malpighian tubules of L. forficatus is strongly dependent on [Na⁺] and independent of [K⁺], Along with a haemolymph with Na⁺ as the main cation, these are features shared with the ‘primitive’ dragonfly (Nicholls, 1985) and the pill millipede (Farquharson, 1974b). The ‘common cation pump’ (Maddrell, 1977) and the coupled entry of Na⁺ and Cl⁻ through the basolateral membrane appear to be early evolutionary characteristics of tracheate Malpighian tubules, modified by diet or by the ionic concentrations in the haemolymph. While the rectum in L. forficatus is important for ion regulation, the Malpighian tubules contribute greatly to volume regulation.

We thank our colleagues and students who helped us to catch the quick centipedes used in this study. Ionic concentrations were determined at the Freie Universität Berlin and we thank Dr I. Zerbst-Boroffka for her hospitality. Dr E. Florey commented on the manuscript and Ms M. A. Cahill corrected the English text.