Rapid initiation of ion transport occurs in the lower Malpighian tubule of the insect Rhodnius prolixus following feeding in vivo or stimulation with 5-hydroxytryptamine (5-HT) in vitro. Using the electron microscope, we have conducted a morphometric analysis of cells in the lowest one-third of the lower tubule, demonstrating that 5-HT also induces mitochondrial movement and microvillar growth simultaneously with, but independent of, the onset of ion transport.

Mitochondria move from a position below the cell cortex to one inside the microvilli within 10 min of stimulation with 5-HT, resulting in an 8- to 10-fold increase in the volume of mitochondria within the microvilli. Previous findings indicated that mitochondrial movement is dependent on actin-containing microfilaments, but not microtubules. As the mitochondria enter the microvillus, the core microfilaments are reorganized into a sheath of microfilaments, which extends closely parallel to the outer mitochondrial membrane down into the cell interior. This sheath of microfilaments is also observed around mitochondria in the axopods. We suggest that the core microfilaments are responsible for mitochondrial movement into the microvilli and axopods. Stimulation with 5-HT induces a shift in mitochondrial configuration from orthodox to condensed, indicating a possible increase in oxidative phosphorylation. Following stimulation, the microvilli grow about 3 × in volume and 2·5 × in surface area. These increases are more than can be accounted for by mitochondrial invasion and must involve the addition of new membrane and microfilament polymerization. The observed changes - microvillar growth, insertion of additional membrane, activation and movement of mitochondria adjacent to the ion transport membrane - are described in the light of their significance in ion transport. A simple model is proposed which unifies the observed ultrastructural changes and known ion movements in the lower tubule.

One of the earliest reports correlating changes in membrane surface area and arrangement with the initiation of ion transport was that of Wigglesworth (1931a). In a lightmicroscope study of the lower Malpighian tubule of the insect Rhodnius prolixus, Wigglesworth demonstrated that the length of the luminal microvilli increased over a period of hours, in association with the onset of feeding-induced diuresis. This finding was later corroborated in an early electron-microscope study antedating the use of glutaraldehyde as a fixative (Wigglesworth & Salpeter, 1962). The mechanism by which stimulation of membrane growth and change in function occurred could not be thoroughly explored in modern terms at that time, but the system clearly presents an elegant experimental model for these questions now.

Since the time of these original experiments, the physiology of Rhodnius Malpighian tubules has been intensively investigated, particularly with regard to in vitro techniques of stimulation and quantitation of transport (see Maddrell, 1971 for review). This progress has been paralleled by general improvements in electron-microscope technology and increasing current interest in membrane growth, membrane function and cytoskeletal-membrane relationships in microvilli (reviewed by Kenny & Booth, 1978; Satir, 1977). We were, therefore, prompted to reexamine the phenomena of microvillar growth and the correlated onset of ion transport in this tissue at the electron-microscope level, using stereological techniques, with an emphasis on cyto-skeletal and membrane events, under defined conditions of in vitro stimulation of the tubule.

In an earlier study we have qualitatively shown that mitochondrial movement into the microvilli accompanies stimulation (Bradley & Satir, 1979b). We found that 5-hydroxytryptamine (5-HT)-stimulated mitochondrial movement is blocked if tubules are treated for 15 min with 10μM cytochalasin B prior to stimulation, but proceeds even in the absence of microtubules in cells treated with colchicine (Bradley & Satir, 1979b). These results led us to propose that mitochondrial movement into the microvilli involves an interaction between the microvillar core microfilaments, which were shown to label with heavy meromyosin, and the mitochondrial outer membrane.

In the present report, we present a detailed quantitative study of the stimulatory effect of 5-HT on the cells of the lower tubule; effects including not only the abovementioned mitochondrial movement but membrane growth, mitochondrial configurational changes, and microvillar growth as well. We have clarified the time course with which these membrane and cytoskeletal modifications occur in correlation with 5-HT stimulation of the initiation of ion transport. On the basis of these findings, we present a model of cytoskeletal-membrane and membrane-membrane interactions which promote the initiation of ion transport.

Insect material

The fifth instar Rhodnius prolixus used for these experiments came either from a colony established by Lauren Zarate at the Department of Entomology and Parasitology, University of California, Berkeley or from a colony which we initiated with animals which were a gift from her. The colony was maintained at 27 °C, 40-60% relative humidity, using a 12-12 h light-dark regime. Blood meals were taken on rabbits.

Ultrastructural observations were made both on insects which had been starved 1 month and those recently fed. All tubules used for in vitro experiments came from animals which had been starved for at least 1 month. All experiments were conducted on the lowest one-third of the lower tubule.

Experimental solutions

Insect Ringer based on that used by Maddrell (1969) was made up to the following final concentrations: 11 nw K.C1, 129 mM NaCl, 8·5 mM MgCl2, 2 mM CaCl2> 34 mM glucose. In later experiments 10 mM Hepes (pH 6·9) was added with identical results. Solutions containing higher potassium ion concentrations were made by substituting additional KC1 for NaCl, such that the final KC1 concentrations were either 35 mM or 45 mM. Some solutions also contained 5 × 10−5 M 5-hydroxytryptamine creatine sulphate (5-HT). Solutions were stored frozen and brought to room temperature before use.

Maddrell & Gee (1974) have demonstrated that high levels of K+(ca. 60 mM) cause neuro-haemal sites in Rhodnius to release diuretic hormone, which can affect Malpighian tubule function. To avoid the possibility of such complications, all tubules used for in vitro studies were dissected in low (11 mM) KC1 Ringer without 5-HT, regardless of later treatment. After dissection from the animal, tubules were placed in fresh Ringer with K+ and 5-HT concentrations appropriate for the experiment.

Dissection procedures and preparation for electron microscopy

Dissections were conducted on insects pinned to the wax-covered bottoms of Petri dishes filled either with fixative (in situ fixation) or insect Ringer (for in vitro studies). The abdominal cuticle was cut open and the one-third of the lower tubule closest to the rectum was removed. Tubules used for in vitro experiments were transferred immediately to hanging drops of insect Ringer positioned on the inside surface of Petri dish covers, which in turn were placed over Petri dish bottoms filled with distilled water. The hanging-drop technique has been shown to maintain ion transport in other insect osmoregulatory organs for at least 3 h (Bradley & Phillips, 1977), although the experimental period for this study was usually only 10 min. With distilled water in the dish bottom, the air surrounding the hanging drop is saturated with water vapour and no measurable change in osmotic concentration of the Ringer occurs (Bradley & Phillips, 1977)-

At the end of the 10-min experimental period, or as appropriate, the tubules were removed from the hanging drop and placed immediately in 4% glutaraldehyde in 0·1 M cacodylate buffer (pH 7·3) containing 1% sucrose for 1 h. This was followed by washing in 0·1 M cacodylate buffer, further fixation in 1% OsO4 in 0·1 M cacodylate buffer, 10 min en bloc staining in 1% uranyl acetate in 70% ethanol, ethanol-gradient dehydration and embedding in Epon 812. Sections were cut on a Reichert ultramicrotome, stained with uranyl acetate in 50% ethanol followed by Reynolds lead stain and observed in a Siemens 101 electron microscope.

Morphometry

Malpighian tubules used for the morphometric studies were chosen at random from the fixations which we had performed. Sections were cut, placed on 100-mesh grids and, after staining, were first observed at a magnification of 285 ×. At this magnification the ultrastructural parameters later measured were not discernible and therefore did not influence the choice of sections. All the sections (approx. 30) on 2 grids were examined and the 3 exhibiting the greatest area of tissue unobscured by grid bars were chosen for photography. The choice of section and its orientation in the picture were therefore dependent on how the section fell on the grid and independent of the physiological state of the tissue.

The apical microvillar border visible in the selected sections was photographed at 6000 x magnification. Three sections, rather than one, were photographed in order to increase the area of microvillar border sampled. The sections were cut from the same area of the block and although never identical were therefore not independent samples. They were statistically treated as serial samples of the tissue. All morphometric measurements were made on prints with a final magnification of 16500 x. The volume density of mitochondria in the microvilli was determined by the ‘differential point counting’ method of Weibel (1969). Because of the anisotropic nature of the microvillar border, a point grid which was arranged as the points of intersection of a triangular lattice with 3 angles of orientation was used (Sitte, 1967). Such a pattern essentially eliminates the directional bias introduced by the parallel microvilli. The points were spaced such that the distance between them (1 cm) was greater than the diameter of the microvilli; thus the points were independent of one another with regard to the orientation and size of the measured structures (Weibel, 1969). The point grid was placed at random on the micrographs and the number of points falling on the microvilli (a) as well as those falling on the mitochondria within the microvilli (b) was recorded. The ratio b/a is the volume density (V,) of mitochondria in the microvilli, a dimensionless value which can be compared between various samples of different size.

A similar grid was made using r-cm-long lines which were similarly arranged in a triangular lattice. The number of intersections of the lines with the microvillar membrane is proportional to microvillar surface area, while the ends of the lines were used as points in a differential point counting method of estimating microvillar volume (Weibel, 1969).

As the plane of the section varies from a position perpendicular to the plane of the luminal surface in the region measured, the apparent microvillar volume density increases. This effect, however, is consequential only at angles which deviate greatly from the vertical. For example, if the angle of section is 30° from the perpendicular, the area of microvillar intersection increases by only a factor of 1·15. We therefore discarded micrographs of sections in which the plane of section noticeably deviated from one perpendicular to the luminal surface.

The longer edge of the grid used for measurements of microvillar volume and surface was placed along the border of the cell at the base of the microvilli. The grid therefore came to rest over the microvillar regions of the cell, extending into the lumen a distance equivalent to 3·8 μm in the original tissue. This distance is sufficient to count essentially all the microvilli, while excluding the longer axopods and other microvilli further out in the lumen which originate from the opposite wall of the lumen.

Appropriate, non-overlapping grid locations were marked on the micrographs starting at the left edge. The location of the grid was therefore determined by the position of the lumen in the micrograph, a factor which was determined in a non-biased manner as described above. Once the grid locations had been marked, 3 of them were chosen using the throw of a die, and used for morphometric analysis.

Measurements of the outer circumference and inner circumference (luminal surface) of the tubule were made on 0·5-μm-thick sections of the tubules, taken from the same blocks as the previous thin sections, stained with toluidine blue and photographed using bright-field illumination on a Zeiss photomicroscope. The circumferences of interest were marked on prints with a final magnification of 443 ×, and measured with a precision map measurer (opisometer).

Statistical analysis of the data was performed using a Student’s t test to test for significant differences of the mean. Morphometric values are expressed as the mean plus or minus the standard error (sample size).

The ultrastructure of the apical region of the unstimulated lower tubule

The Malpighian tubules of Rhodnius are differentiated morphologically and physiologically into an upper and lower tubule (Fig. 1). The upper tubule, the region most distal from the rectum, produces an isosmotic filtrate of the blood by means of active ion transport (Ramsay, 1953; Maddrell, 1969). The lower half of the lower tubule is the site of KC1 resorption from the primary urine during the rapid diuresis following a blood meal (Wigglesworth & Salpeter, 1962; Maddrell, 1978a). Between blood meals (the starved condition) the lower tubule is inactive.

Aspects of the apical surfaces of lower Malpighian tubule cells from animals starved for 1 month prior to dissection are shown in Figs. 2−6 and 7 A. Because ion transport under these conditions is minimal (Maddrell & Phillips, 1975), we refer to this as the unstimulated tubule. The cells of the lower tubule present a highly convoluted membrane to the lumen and possess 2 classes of cell extensions, microvilli (Figs. 2−5) and axopods (Fig. 6). The microvilli contain bundles of 6-nm microfilaments (Figs. 3−5) that run parallel to the long axis of the microvillus (Fig. 2) and occasionally also contain extensions of endoplasmic reticulum (ER) (Fig. 4), as observed previously in insects (Satir & Stuart, 1965; Wigglesworth & Salpeter, 1962).

In cross-section, the microfilaments in the microvilli are evenly spaced (Figs. 3, 5). When extensions of ER are also present in the microvillus, the microfilaments are arranged in a ring and lie closer to the plasma membrane than to the ER membrane (Fig. 4). This variability in arrangement suggests that the microfilaments can readily change relationships with one another in the microvillus.

Below the microvilli lies a 0 · 8-μ m-wide cell cortex, which stains more uniformly and somewhat more densely than other portions of the cytoplasm (Figs. 2, 7A). In this region the microfilament bundles spread out from the base of the microvilli into a loosely organized terminal web. There appear to be 2 populations of mitochondria in the lower tubules, a basal population that is intimately associated with the basal plasma membrane and an apical population that lies in the region immediately below the cortex in the non-stimulated tubules. A few mitochondria penetrate into the cortex in the non-stimulated state (Fig. 7A), very infrequently extending through the cortex into a microvillus.

The axopods differ from the microvilli in that they are larger and relatively less frequent along the Malpighian tubule and because they contain an array of microtubules. They have been described in detail in an earlier study (Bradley & Satir, 1979a). In addition to the microtubules, axopods contain some microfilaments, most of which lie close to the axopod membrane (Fig. 6). In the unstimulated tubule, mitochondria are seldom if ever present in the axopods.

Fig. 1 is a diagrammatic representation of the lower Malpighian tubule indicating the relative size and distributions of microvilli, mitochondria, and axopods in the unstimulated tubule.

Ultrastructural modifications associated with stimulation by feeding in vivo and 5-HT in vitro

The Malpighian tubules of animals which had been starved for 1 month could be stimulated by placing the insect on a rabbit for half an hour, during which time it imbibed a blood meal. Twenty minutes after feeding was terminated, the insects were dissected under the fixative solution. This protocol was selected because it is known that such time is sufficient for feeding-induced diuresis, and consequently ion transport in the tubules, to reach a maximum (Maddrell & Phillips, 1975). The lower Malpighian tubules from these larvae (Fig. 7B) showed considerable ultrastructural modifications compared to the starved larvae. One such obvious modification is that many branches from the apical mitochondria had moved into the microvilli. Notice also in Fig. 7B the whorl of rough ER, a configuration frequently observed in stimulated tubules.

The stimulation of ion transport in the tubule by release of diuretic hormone after feeding can be duplicated in vitro by the use of 5-HT (Maddrell, Pilcher & Gardiner, 1969; Maddrell & Phillips, 1975). Control tubules incubated 10 min in vitro in the absence of 5-HT (Fig. 7 c) were essentially identical ultrastructurally to tubules fixed in vivo (Fig. 7 A). Tubules treated with 5 × 10−5 M 5-HT for 10 min in vitro (Fig. 70) showed significant ultrastructural modifications, including the movements of branches of mitochondria into the microvilli, and a concomitant growth of the microvilli, essentially identical to the ultrastructural changes induced by feeding.

In both the controls and the 5-HT-stimulated tubules, the experimental period of incubation prior to fixation was normally 10 min. During this 10-min stimulation period, the mitochondria move into the tips of the microvilli, a total distance of about 7 μ m. In 3 tubules where only 5 min of stimulation was used, 1 of the tubules showed mitochondria fully extended, the other 2 were indistinguishable from controls. This suggests that the movement of the mitochondria occurs quite rapidly after stimulation so that intermediate states are not easily observed. The time course for the initiation of movement averages between 5 and 10 min, a time interval identical to that shown for 5-HT stimulation of KC1 transport in this tissue (Maddrell & Phillips, 1975).

An additional consequence of 5-HT stimulation can be observed in Figs. 8 A, B. The mitochondria in the non-stimulated state are in the orthodox configuration (Hackenbrock, 1966, 1968), that is, the mitochondrial matrix is moderately stained and fills most of the volume inside the outer mitochondrial membrane (Fig. 8 A). After 5-HT stimulation (Fig. 8 B) the mitochondria assume the condensed configuration, characterized by more intense staining of the matrix, shrinkage of the matrix away from the outer mitochondrial membrane, and increased space between the membranes of the cristae.

Mitochondrial-cytoskeletal relationships

In axopods which contain mitochondria, the longer axes of the mitochondria are aligned parallel to the microtubules in the microtubular array (Fig. 9). Occasionally, bridges can be seen between microtubules and the outer mitochondrial membrane, such as those shown in Fig. 9 (inset), near the base of an axopod. This parallel alignment of microtubules with the longer axes of mitochondria is also prevalent in the cell interior. Microtubules in the cell body and near the base of the axopods may form transient bridges with the outer mitochondrial membrane and influence the path of mitochondrial movement. Nevertheless, in the vast majority of cases where mitochondria occur in axopods, the mitochondria are more closely associated with microfilaments of the axopod than with microtubules (for example, in Fig. 11).

In the microvilli that contain no microtubules the mitochondria are always associated with a sheath of microfilaments which runs up the microvillus, parallel to both the mitochondrial outer membrane and the plasma membrane of the microvillus. Very few microfilaments remain associated with ER (Fig. 13). In cross-sections (Figs. 10, 13), the appearance, spacing, and location of the sheath around the mitochondrion is quite uniform. The microfilaments lie characteristically in a single row, within about 6–8 nm of the outer mitochondrial membrane. Such microfilament association with the mitochondrial membrane is apparently very stable for it can be observed even in microvilli which have been osmotically disrupted (Fig. 12). In contrast, the association between the sheath and the microvillar membrane is much less close and is disrupted by osmotic shock.

The mitochondrial-microfilament association continues below the microvilli into the cytoplasm as well. Fig. 14 shows microfilaments arranged around mitochondria both just above and below the bases of the microvilli, in a section cut tangentially to the cell surface. A longitudinal view of the same cell region in Fig. 15 shows that the sheath of microfilaments extends into the cell and now runs parallel to the outer mitochondrial membrane for a considerable distance. Thus, the mitochondrial penetration into the microvillus has been accompanied by a considerable ordering and alignment of microfilaments, well into the cortical region of the cell.

In summary, the ultrastructural evidence points to a strong association, both in microvilli and in axopods containing mitochondria, between the mitochondrial outer membrane and a neighbouring array of microfilaments. In a separate study (Bradley & Satir, 1979b) we have shown that the microfilaments decorate with heavy meromyosin (HMM) and therefore are presumed to contain actin, and that movement of the mitochondria into the microvilli is prevented by cytochalasin B but not by colchicine. Together with the morphological evidence presented here, this suggests that the microfilaments of the sheath may have an active role in producing the observed mitochondrial translocation upon stimulation.

Morphometric analysis

In order to evaluate quantitatively the relationship between 5-HT stimulation and cellular reorganization, we employed morphometric measurements of mitochondrial entry into microvilli and of coordinate increases in microvillar volume and surface area. The effects of solutions containing concentrations of KC1 high enough to inhibit ion transport (> 30 mM KC1) (Maddrell & Phillips, 1977) were examined with regard to cellular reorganization.

The degree of mitochondrial entry into the microvilli was determined by measuring changes in the mitochondrial volume density (Vv) in the microvilli, a value equivalent to the percentage of the volume of the microvillar space occupied by the mitochondria.

As shown in the first column of Table 1, non-stimulated tubules fixed either in situ, or after incubation in vitro in solutions containing no 5-HT, have mean mitochondrial percentage volumes of 0 – 3%, i.e. the microvilli contain essentially no mitochondria. There are no statistically significant differences (P > 0 · 1) among the unstimulated preparations regardless of K+ concentration in the test solution. In solutions containing 5-HT, however, mitochondrial percentage volume increases very significantly (P < 0 · 00 1) to a mean value of approximately 20%, reflecting the massive entry of mitochondria into the apical microvilli. This result is again identical in solutions of low (11 mM) and high (35 mM) K+ concentration (P > 0 · 5).

Measurements were also made of changes in microvillar volume and surface area in response to 5-HT. Microvillar volume and surface area are measured relative to a reference length of luminal border along the base of the microvilli. It was therefore important to demonstrate that the increase in microvillar surface area was truly due to changes in that parameter and not to changes in the reference surface, i.e. a decrease in the luminal circumference. This latter possibility might result from shrinkage of the whole tubules or a swelling of the cells which would reduce the size of the luminal space. To investigate these possibilities, we used thick sections to measure the circumference of the same tubules used for the morphometric determinations. The outer circumference of the cellular portion of the tubule was measured by following a contour just inside the extracellular basal lamina. A line at the base of the microvilli outlined the luminal circumference. Any shrinkage of the whole tubule would reduce both circumferences, while cellular swelling with resultant luminal constriction would affect primarily the inner circumference.

Table 2 shows the results of these measurements. No significant differences in the outer or inner diameters of the tubules were found either in response to 5-HT (P > 0 · 5) or K+ concentration (P > 0 · 2). These data establish the validity of our morphometric procedures, indicating that measured variations in microvillar volume and surface area upon stimulation are not due to a change in the reference surface, i.e. the luminal circumference.

Table 1, column 2 shows measurements of specific microvillar volume (Vv), i.e. volume of microvilli per given surface of lumen adjacent to the cell, after various treatments. Addition of 5 × 10−5 M 5-HT produces a statistically highly significant ∼ 3-fold increase in microvillar volume as compared to controls (P > 0 · 001). If the increase in microvillar volume were entirely due to the mitochondria entering the microvilli upon stimulation, a 3-fold increase in microvillar volume would result in 66% of the microvillar space being occupied by mitochondria. Since the observed percentage of the microvilli occupied by mitochondria is about 20% (Table 1), the increase in volume cannot be accounted for solely by mitochondrial entry and additional mechanisms of microvillar growth must be postulated. Like the mitochondrial movement, the 5-HT-stimulated increase in microvillar volume is not significantly different in high levels of external K+ (P > 0 · 2).

The surface area of microvilli (surface density, Sv) (Weibel, 1969) was measured on the same micrographs as the microvillar volume density. Tubules in 11 mM K+ Ringer and stimulated with 5-HT have about 2 · 5 times the microvillar surface area of controls. These differences are significant (P< 0·03) (Table 2). No significant differences in surface area were found between in situ and non-stimulated in vitro fixed tubules or between tubules in different K+ concentrations (all P values > 0 · 4).

To be absolutely certain that we were above the level of potassium necessary to inhibit transport (Maddrell & Phillips, 1977), 2 tubules were bathed in even higher K+ levels (45 mM). These also responded to 5-HT with rapid microvillar growth and mitochondrial movement. It is clear, therefore, that levels of K+ sufficient to reduce ion transport substantially according to Maddrell & Phillips (1977) do not affect (1) the entry of mitochondria into the microvilli; (2) the final mitochondrial volume within the microvilli; (3) the increase in microvillar volume and surface area.

Mitochondrial movement

Earlier workers conducting studies on the ultrastructure of Malpighian tubules (Beams, Tahmishian & Devine, 1955) suggested that the mitochondria seemed to ‘flow’ (their quotation marks) into the microvilli. This suggestion was based entirely on the elongate appearance of the mitochondria, not on observations of changes in position with time. More recently, it has been reported that in some insects the mitochondria in secretory Malpighian tubules are, over a period of hours, gradually retracted from the microvilli during the pupal stage of development (Berendes & Willart, 1971 ; Byers, 1971 ; Ryerse, 1977). This morphological change, which can be induced in vitro by ecdysterone, an insect molting hormone, is correlated with a period of greatly reduced urine production (Ryerse, 1978 a, b). R. prolixus, a hemipteran insect, does not have a pupal stage and the microvilli in the upper tubule of Rhodnius contain mitochondria throughout the life stages that have been examined (Wigglesworth & Salpeter, 1962; personal observation). The resorptive lower half of the tubule, the region used for this study, becomes active during rapid diuresis following feeding (Maddrell & Phillips, 1975). Mitochondrial movement in the lower tubule differs from that described in other insect Malpighian tubules because (1) mitochondrial movement can be stimulated in vitro by 5-HT alone, in the absence of hormones which regulate development, and (2) the mitochondria show rapid movements apically, into the tips of the microvilli, upon stimulation. This movement is rapid when compared to other hormonally-induced morphological changes, but as a form of motility, the rate of mitochondrial movement is fairly slow (7 μ m in 5 – 10 min) - about 2 – 3 times faster than the rate of chromosome movement in anaphase (Nicklas, 1965).

As the mitochondria enter the microvilli, the spatial arrangement of microfilaments in the microvillar core changes such that a sheath of microfilaments comes to lie between the mitochondria and the nearby plasma membrane. The observation that the microfilaments extending down from the microvilli run parallel to the mitochondria, even into the cell body below the microvilli, suggests that the mitochondria may be guided along pre-existing microfilament ‘tracks’ which are continuous with the microfilament bundles of the microvilli. We have recently shown that the microvillar microfilaments, which have diameters identical to actin filaments in other systems, will label with the myosin fragment HMM, presumably because they contain actin (Bradley & Satir, 1979b). Many, if not all, of the microfilaments label with HMM ‘arrowheads’ pointing toward the cell body. This raises the possibility that the mitochondria may be moved into the microvilli by active sliding along the microfilaments in a manner similar to actin-myosin sliding in muscle.

We have found that mitochondria lie adjacent to the microtubules in the cell body and are aligned parallel to the microtubule arrays in axopods after stimulation. Occasionally, bridges which presumably maintain these associations are observed between microtubules and the outer mitochondrial membrane. In axopods, however, mitochondria are invariably part of a mitochondrial-microfilament-plasma membrane association, even if microtubules are present nearby. Microtubules are apparently not directly involved in this mitochondrial movement, since mitochondria can enter microvilli where no microtubules are seen (this study) and depolymerization of microtubules with colchicine does not block 5-HT-stimulated mitochondrial movement (Bradley & Satir, 1979b). This clearly differentiates the mechanism of mitochondrial movement into microvilli in the lower tubule from that postulated for mitochondrial movement for nerve axons, where microtubules are implicated on the basis of morphometric studies (Smith, Jarlfors & Cayer, 1977) and the blockage of mitochondrial movement by vinblastine and colchicine (Friede & Khang-Cheng, 1977). The microtubule arrays in the cell cortex of the lower tubule may serve as a structural framework to which mitochondria are attached in the non-stimulated condition and to which they return upon leaving the microvilli.

The morphological association of mitochondria and microfilaments reported in this study and the inhibition of mitochondrial movement by cytochalasin B (Bradley & Satir, 1979b) support the hypothesis that actin microfilaments play a key role in the production of mitochondrial movement in the lower tubule. It would be interesting to know the localization of myosin in this tissue. The mechanism by which the mitochondria are retracted from the microvilli following stimulation is also unknown.

Microvillar growth

We have shown that the microvillar growth described by Wigglesworth (1931a) and Wigglesworth & Salpeter (1962) several hours after feeding, can also be induced in 10 min in vitro by 5-HT. This growth involves not only a 3-fold increase in microvillar volume, but a 2’4-fold increase in microvillar membrane surface area as well. Such a rapid increase in surface area must occur by incorporation of membrane from non-microvillar regions. Since the lateral cell membrane is thought to be stabilized by the presence of septate junctions, the most likely source would seem to be intracellular membrane.

This form of membrane incorporation, which may be a general mechanism of membrane modification and growth (Satir, 1974), has been observed in other ion transport systems. In salt-stressed ducks, the nasal salt gland displays increases in salt transport and Na+-K+-ATPase concentration over a period of days (Ernst, Goertemiller & Ellis, 1967). It has recently been shown that additional Na+-K+-ATPase is inserted into the lateral cell membranes after being synthesized in the ER and glycosylated in the Golgi apparatus (Hossler, Sarras & Barrnett, 1978). A similar pathway of membrane incorporation is thought to occur in frog oxyntic cells where cyclic AMP causes a simultaneous increase in apical surface membrane and decrease in intracellular tubular and vesicular membrane (Carlisle, Chew & Hersey, 1978). The increase in microvillar surface area which we detect in the lower tubule may reflect the incorporation into the apical membrane of intracellular membrane stores containing transport proteins. The nature of these proteins, and the location and site of synthesis of the intracellular membrane stores, remains at present undetermined.

The longer microvilli from stimulated tubules contain microfilaments along their entire length, indicating that the microfilament core bundle lengthens either by continuing microfilament polymerization, sliding within the bundle, or entry of preformed microfilaments from the terminal web. One possible mechanism of microvillar growth, namely volume displacement by entering mitochondria, has been ruled out by the morphometric observations. To account for the observed 3-fold increase in microvillar volume by mitochondrial entry alone, the final percentage volume of microvilli occupied by mitochondria would need to be about 3 times greater than the observed value.

Mitochondrial configuration

We have shown that the apical mitochondria in the lower tubule undergo a change in configuration from orthodox to condensed upon 5-HT stimulation. This may reflect a change in the mitochondria to state HI (fast oxidative phosphorylation) or state V (anaerobiosis) (Hackenbrock, 1968). Either state would indicate that oxygen consumption by the tissue increases upon 5-HT stimulation. In further study we hope to distinguish between the possible states. If the mitochondria are indeed in state HI this will enable us to monitor not only changes in mitochondrial position in relation to ion transport, but rates of mitochondrial phosphorylation as well.

Changes in cellular ATP consumption and mitochondrial configuration in response to hormone or cyclic AMP stimulation, have also been reported for other systems. Thyroxin induces an increase in oxygen consumption, ATP hydrolysis and the concentration of the ion transport enzyme Na+-K+-ATPase in a wide variety of mammalian tissues (Philipson & Edelman, 1977; Smith & Edelman, 1979). In frog oxyntic cells, administration of cyclic AMP initiates HC1 secretion and a mitochondrial configurational change similar to that which we observe in the lower tubule (Carlisle et al. 1978).

Relationship between the observed morphological changes and the initiation of ion transport

Two features of cellular morphology which appear to have general significance in ion-transporting tissues are: (1) a specialized surface area of the cell membrane enlarged by microvilli or infoldings, and (2) the placement of mitochondria close to this membrane. We have demonstrated large and rapid changes in the lower Malpighian tubule of Rhodnius with respect to both of these features upon treatment with 5-HT.

Evidence to date points to a close link in this tissue between the initiation of mitochondrial movement and microvillar growth which we describe and the initiation of KC1 transport. Maddrell & Phillips (1975) have shown that the lower Malpighian tubules of Rhodnius are stimulated to initiate ion transport in vivo by a diuretic hormone released soon after initiation of feeding. We have shown that microvillar growth and mitochondrial movement also occur in vivo, following a blood meal, within the time period during which the initiation of ion transport is known to occur. In the preparations in vitro no ion transport occurs in the absence of a stimulant. Upon addition of 5-HT or diuretic hormone, KC1 transport could be demonstrated within 5-10 min (Maddrell & Phillips, 1975). We have shown that mitochondrial movement and microvillar growth also depend on 5-HT in vitro. These morphological effects occur simultaneously and with the same time course in vitro, as the initiation of ion transport, i.e. within 10 min (Maddrell & Phillips, 1975), although we have not yet demonstrated simultaneous changes in a single preparation.

One of our observations, however, clearly indicates that despite the common time course of stimulation, mitochondrial placement does not by itself control the initiation of ion transport in the lower tubule. Maddrell & Phillips (1977) showed that potassium ion concentrations of 30 mM or above in the bathing medium inhibited 5-HT-stimu-lated KC1 resorption in the lower tubule. They pointed out that such a system served as a negative feedback which controls blood potassium levels. We have demonstrated in this study that 35-45 mM KC1 does not inhibit 5-HT-stimulated mitochondrial movement and microvillar growth, showing that the cell can continue to respond morphologically to 5-HT while ion transport is blocked.

We suggest that the increase in microvillar volume and surface area, as well as the movement of the mitochondria into the microvilli, are longer time course cellular adjustments in response to 5-HT. These prepare the cell for rapid initiation of ion transport, i.e. they are necessary but not sufficient for initiation. Yet another cellular control mechanism, which is sensitive to external basal K+, controls the instantaneous transport rate. It seems sensible to regulate the instantaneous rate of ion transport by constant feedback monitoring of blood K+ instead of by the 5-HT (or diuretic hormone) level in the blood, as long as these are above a threshold level necessary to activate the cell.

A model coordinating observations on ion transport, mitochondrial configuration and morphological reorganization in the lower tubule

Fig. 16 schematically illustrates the direction of known ion movements across the apical and basal membranes of the lower tubule. Potassium ions are resorbed from the urine following stimulation, i.e. they are transported across both the apical and basal membranes of the cells of the tubule (Fig. 16). We have therefore shown only the potassium ion movements. Wigglesworth (1931b) found that urine was acidified during the course of passage through the lower tubule, going from a pH of ∼ 7 · 2 to 6 · 6. We have represented this as a transport of hydrogen ions into the lumen as proposed by Miles (1966). The pharmacology of the events following 5-HT stimulation in the lower tubule is unknown, but in the upper tubule and in another very closely related insect ion-transporting epithelium (Carausius salivary gland) 5-HT binds to the basal membrane (Maddrell, Pilcher & Gardiner, 1971) and acts via intracellular increases in cyclic AMP (Berridge, 1975 ; Maddrell et al. 1971) and influxes of extracellular calcium (Prince & Berridge, 1973). We have drawn Fig. 16 assuming that this is the case for the lower tubules.

We propose a model (Fig. 17) to explain how such stimulatory and transport events can be related to morphological changes in mitochondrial movement and configuration and to microvillar growth as described in this study. We suggest that 5-HT binding to the basal membrane leads to the transmission of a transcellular signal, possibly an increase in cyclic AMP. This signal could result in an increase in the permeability of the apical membrane to Ca2+. The subsequent rise of intracellular Ca2+ would probably be restricted to the apical microvilli and cell cortex since the distance that free Ca2+ can move in insect cells is quite limited (Rose & Lowenstein, 1975).

Transient increases in intracellular Ca2+ are known to activate (1) membrane fusion (Satir & Oberg, 1978; Tilney, Kiehart, Sardet & Tilney, 1978) and (2) actin-myosin-based cell movements (Bendall, 1974; Mooseker, 1976; Rodewald, Newman & Karnovsky, 1976). In the lower tubule, local calcium influx might promote fusion of internal membrane stores with the microvillar plasma membrane (Fig. 17, b) leading to an increased microvillar surface. A secondary possible result of fusion might be the insertion of new ion transport ‘pumps’ into the apical membrane. Simultaneously, the microvillar core microfilaments would be activated by an increase in cytoplasmic Ca2+ to initiate mitochondrial movement into the microvilli, thereby positioning the mitochondria closer to the apical ion-transporting membrane.

The transport of K+ across the luminal membrane is against its chemical gradient. Presumably, the energy needed to drive this transport is derived from ATP supplied by the now nearby mitochondria (Fig. 17, c). The presence of an ATP-consuming process (the K+ pump in the plasma membrane) in close proximity to the source of ATP leads to a maximization both of transport rates and of ATP utilization, promoting stimulation of oxidative phosphorylation in the mitochondria. We believe that the change in mitochondrial configuration to the condensed state is evidence of this sequence of events.

The transport of K+ into the cell of the lower tubule must occur either by means of symport with an anion or antiport with another cation. Since the osmotic concentration of fluid passing through the lower tubule drops dramatically during K+ resorption (Maddrell & Phillips. 1975), it is clear that the majority of K ions resorbed are accompanied by an anion, probably Cl. Chloride concentration values for fluid collected immediately after leaving the lower tubule are not available. Urine excreted by the insect has passed through the rectum and may be somewhat modified by that organ. However, if this fluid is collected during rapid diuresis when its passage through the rectum is swift, ionic analysis indicates that 83% (Maddrell, 1978 b) to 96% (Maddrell & Phillips, 1975) of the resorbed K+ is accompanied by Cl. The quantities of K+ not accompanied by Cl must be electrically balanced by the movement of other ions. In view of Wigglesworth’s (1931 b) findings that urine passing through the lower tubule is acidified, we propose that some of these K+ are exchanged for H* in a mechanism similar to the Na+/H+ antiport observed in the apical membranes of the vertebrate proximal tubule and small intestine (Murer, Hopfer & Kinne, 1976). By exchanging H+ for K+, the antiport would effectively raise the pH of the microvillar cytoplasm. We suggest that this leads to increased actin polymerization, which in conjunction with the increased microvillar membrane derived from exocytosis, produces the observed microvillar elongation and growth.

The involvement of intracellular pH in controlling actin polymerization has been demonstrated in Dictyostelium cytoplasmic extracts (Condeelis & Taylor, 1977), echinoderm sperm (Tilney et al. 1978), and the microvilli of sea-urchin eggs (Begg & Rebbun, 1979; Spudich & Spudich, 1979). In each of these systems an increase in actin polymerization following a rise in pH is thought to result from the dissociation of actin-binding proteins from G-actin, allowing the polymerization to F-actin. One such protein, profilin, has been isolated and characterized (Tilney, 1976). During the formation of the acrosome in echinoderm sperm, the sequence of ion changes is rather similar to that proposed here. The reaction is initiated by the Ca2+-dependent exocytosis of the acrosomal vesicle (Tilney et al. 1978) which fuses with the cell membrane and contributes membrane for acrosomal elongation. This acrosomal lengthening is presumably itself caused by a subsequent rise in pH leading to microfilament polymerization from the actomere (Tilney, 1978; Tilney et al. 1978).

In this study we have demonstrated that although high levels of external K+ may block K+ transport across the epithelium, they clearly do not interfere with 5-HT-stimulated microvillar growth and mitochondrial movement or configurational change. It may be that high external K+ interferes with the exit of K+ across the basal mem brane (Fig. 17), perhaps through its influence on the resting potential of this membrane.

While our model contains several intermediate steps which have not been unambiguously demonstrated, the lower tubule is clearly a valuable system in which to investigate the cellular mechanisms which link hormone stimulation of epithelia with changes in membrane structure and composition, ion transport, organelle distribution, and cytoskeletal control of these processes. Our model attempts to rationalize the diverse responses of the cell that we have documented by postulating a few, simple common mechanisms that produce the effects.

In further study we can clarify and extend our model by asking, for example, whether: (1) we can induce mitochondrial movement and microvillar growth in unstimulated tubules by direct treatment of the apical surface with Ca2+ plus appropriate ionophores; (2) urine acidification and microvillar growth are simultaneously blocked when the tubule lumen is perfused with K+-free fluid; (3) increased ATP utilization accompanies stimulation ; and (4) manipulating the resting potential of the basal membrane with low chloride media has the same effect as high K+.

This work was supported by a grant from the USPHS (HL 22560) to P. Satir and NIH postdoctoral fellowship Am 05499 to T. Bradley.

Beams
,
H. W.
,
Tahmishian
,
T. N.
&
Devine
,
R. L.
(
1955
).
Electron microscope studies on the cells of the Malpighian tubules of the grasshopper (Orthopiera, Acrididae)
.
jf. biophys. biochem. Cytol
.
1
(
3
),
197
202
.
Begg
,
D. A.
&
Rebbun
,
L. I.
(
1979
).
pH regulates the polymerization of actin in the sea urchin egg cortex
.
J. Cell Biol
.
83
,
241
248
.
Bendall
,
J. R.
(
1974
).
Muscles, Molecules and Movement
.
N.Y
. :
Elsevier
.
Berendes
,
H. D.
&
Willart
,
E.
(
1971
).
Ecdysone-related changes at nuclear and cytoplasmic level of Malpighian tubule cells in Drosophila
.
J. Insect Physiol
.
17
,
2337
2350
.
Berridge
,
M. J.
(
1975
).
The interaction of cyclic nucleotide and calcium in the control of cellular activity
.
Adv. eye. nuc. Res
.
6
,
1
98
.
Bradley
,
T. J.
&
Phillips
,
J. E.
(
1977
).
The location and mechanism of hyperosmotic fluid secretion in the rectum of the saline-water mosquito larvae, Aedes taeniorhynchus
.
J. exp. Biol
.
66
,
111
126
.
Bradley
,
T. J.
&
Satir
,
P.
(
1979a
).
Insect axopods
.
J. Cell Sci
.
35
,
165
175
.
Bradley
,
T. J.
&
Satir
,
P.
(
1979b
).
Microfilament-associated mitochondrial movement
.
J. supramolec. Struct
.
12
(
2
),
165
175
.
Byers
,
J. R.
(
1971
).
Metamorphosis of perirectal Malpighian tubules of the mealworm Tenebrio molitor. I. Histology and histochemistry
.
Can. J. Zool
.
49
,
823
832
.
Carlisle
,
K. S.
,
Chew
,
C. S.
&
Hersey
,
S. J.
(
1978
).
Ultrastructural changes and cyclic AMP in frog oxyntic cells
.
J. Cell Biol
.
76
,
32
42
.
Condeelis
,
J. S.
&
Taylor
,
D. L.
(
1977
).
The contractile basis of amoeboid movement. V. The control of gelation, solation, and contraction in extracts from Dictyostelium discoideum
.
J. Cell Biol
.
901
927
.
Ernst
,
S. A.
,
Goertemiller
,
C. C.
&
Ellis
,
R. A.
(
1967
).
The effect of salt regimens on the development of (Na+-K+)-dependent ATPase activity during the growth of salt gLonds in ducklings
.
Biochem. biophys. Acta
135
,
682
692
.
Friede
,
R. L.
&
Khang-Cheng
,
Ho
. (
1977
).
The relation of axonal transport of mitochondria with microtubules and other axoplasmic organelles
.
J. Physiol
.
265
,
507
519
.
Hackenbrock
,
C. R.
(
1966
).
Ultrastructural basis for metabolically linked mechanical activity in mitochondria. I. Reversible ultrastructural changes with change in metabolic steady state in isolated liver mitochondria
.
J. Cell Biol
.
30
,
269
297
.
Hackenbrock
,
C. R.
(
1968
).
Ultrastructural basis for metabolically linked mechanical activity in mitochondria. I. Electron transport-linked ultrastructural changes in mitochondria
.
J. Cell Biol
.
37
,
345
369
.
Hossler
,
F. E.
,
Sarras
,
M. P.
&
Barrnett
,
R. J.
(
1978
).
Ouabain binding during plasma membrane biogenesis in duck salt gLond
,
J. Cell Sci
.
31
,
179
198
.
Kenny
,
A. J.
&
Booth
,
A. G.
(
1978
).
Microvilli: Their ultrastructure, enzymology and molecular organization
.
In Essays in Biochemistry
, vol.
14
, (ed.
P. N.
Campbell
&
W. N.
Aldridge
), pp.
1
44
.
London and New York
:
Academic Press
.
Maddrell
,
S. H. P.
(
1969
).
Secretion by the Malpighian tubules of Rhodnius. The movements of ions and water
.
J. exp. Biol
.
51
,
71
97
.
Maddrell
,
S. H. P.
(
1971
).
The mechanisms of insect excretory systems
.
Adv. Insect Physiol
.
8
,
199
331
.
Maddrell
,
S. H. P.
(
1978a
).
Malpighian tubules
.
In Transport of Ions and Water in Animal Tissues
(ed.
B. L.
Gupta
,
R. B.
Moreton
,
J. L.
Oschman
&
B. J.
Wall
), pp.
541
569
.
London
:
Academic Press
.
Maddrell
,
S. H. P.
(
1978b
).
Physiological discontinuity in an epithelium with an apparently uniform structure
.
J. exp. Biol
.
75
,
133
145
.
Maddrell
,
S. H. P.
&
Gee
,
J. D.
(
1974
).
Potassium-induced release of the diuretic hormones of Rhodnius prolixus and Glossina austeni: Ca2+-dependence, time course and localization of neurohaemal areas
.
J. exp. Biol
.
61
,
155
171
.
Maddrell
,
S. H. P.
&
Phillips
,
J. E.
(
1975
).
Secretion of hypo-osmotic fluid by the lower Malpighian tubules of Rhodnius prolixus
.
J. exp. Biol
.
62
,
671
683
.
Maddrell
,
S. H. P.
&
Phillips
,
J. E.
(
1977
).
Regulation of absorption in insect excretory systems
.
In Comparative Physiology
(ed.
K.
Schmidt-Nielsen
,
L.
Bolis
&
S. H. P.
Maddrell
), pp.
179
185
.
Cambridge University Press
.
Maddrell
,
S. H. P.
,
Pilcher
,
D. F. M.
&
Gardiner
,
B. O. C.
(
1969
).
Stimulatory effect of 5-hydroxytryptamine (serotonin) on secretion by Malpighian tubules of insects
.
Nature, Lond
.
222
,
784
785
.
Maddrell
,
S. H. P.
,
Pilcher
,
D. F. M.
&
Gardiner
,
B. O. C.
(
1971
).
Pharmacology of the Malpighian tubules of Rhodnius and Carausius-. The structure-activity relationship of tryptamine analogues and the role of cyclic AMP
.
J. exp. Biol
.
54
,
779
804
.
Miles
,
P. W.
(
1966
).
A modification of Wigglesworth’s model for the excretion of uric acid in insects in the light of modern hypotheses of ion transport
,
J. theor. Biol
.
12
,
130
132
.
Mooseker
,
M. S.
(
1976
).
Brush border motility. Microvillar contraction in Triton-treated brush borders isolated from intestinal epithelium
,
J. Cell Biol
.
71
(
2
),
417
433
.
Murer
,
H.
,
Hopper
,
U.
&
Kinne
,
R.
(
1976
).
Sodium/proton antiport in brush-border-membrane vesicles isolated from rat small intestine and kidney
.
Biochem. J
.
154
,
597
604
.
Nicklas
,
R. B.
(
1965
).
Chromosome velocity during mitosis as a function of chromosome size and position
.
JC Cell Biol
.
65
,
119
136
.
Philipson
,
K. D.
&
Edelman
,
I. S.
(
1977
).
Characteristics of thyroid-stimulated N+-K+-ATPase of rat heart
.
Am. J. Physiol
.
232
,
C202
206
.
Prince
,
W. T.
&
Berridge
,
M. J.
(
1973
).
The role of calcium in the action of 5-HT and cyclic AMP on salivary gLonds
,
J. exp. Biol
.
58
,
367
384
.
Ramsay
,
J. A.
(
1953
).
Active transport of potassium by the Malpighian tubules of insects
.
J, exp. Biol
.
30
,
358
369
.
Rodewald
,
R.
,
Newman
,
S. B.
&
Karnovsky
,
M. J.
(
1976
).
Contraction of isolated brush borders from the intestinal epithelium
,
J. Cell Biol
.
70
,
541
554
.
Rose
,
B.
&
Loewenstein
,
W. R.
(
1975
).
Permeability of cell junction depends on local cytoplasmic calcium activity
.
Nature, Lond
.
254
,
250
252
.
Ryerse
,
J. S.
(
1977
).
Control of mitochondrial movement during development of insect Malpighian tubules
.
Proc, microscop. Soc. Can
.
4
,
48
49
.
Ryerse
,
J. S.
(
1978a
).
Ecdysterone switches off fluid secretion at pupation in insect Malpighian tubules
.
Nature, Lond
.
271
,
745
746
.
Ryerse
,
J. S.
(
1978b
).
Developmental changes in Malpighian tubule fluid transport
,
J. Insect Physiol
.
24
,
315
319
.
Satir
,
B. H.
(
1974
).
Membrane events during the secretory process
.
In S.E.B. Symposium
, vol.
28
(ed.
M. A.
Sleigh
&
D. H.
Jennings
), pp.
399
418
.
Cambridge University Press
.
Satir
,
B. H.
&
Oberg
,
S. G.
(
1978
).
Paramecium fusion rosettes: possible function as Cas+ gates
.
Science, N. J
.
199
,
536
538
.
Satir
,
P.
(
1977
).
Microvilli and cilia : Surface specializations of mammalian cells
.
In Mammalian Cell Membranes
, vol.
2
(ed.
G. A.
Jamieson
&
D. A.
Robinson
), pp.
323
340
.
London
:
Butterworth
.
Satir
,
P.
&
Stuart
,
A. M.
(
1965
).
A new apical microtubule-associated organelle in the sternal gLond of Zootermopis nevadinsis (Hagen), Isoptera
.
J. Cell Biol
.
24
(
2
),
227
283
.
Sitte
,
H.
(
1967
).
In Quantitative Methods in Morphology
(ed.
E. R.
Weibel
&
H.
Elias
), pp.
167198
.
Berlin
:
Springer
.
Smith
,
D. S.
,
Jarlfors
,
M.
&
Cayer
,
M. L.
(
1977
).
Structural cross-bridges between microtubules and mitochondria in central axons of an insect (Periplaneta americana)
.
J. Cell Sci
.
27
,
255
272
.
Smith
,
T. J.
&
Edelman
,
I. S.
(
1979
).
The role of sodium transport in thyroid thermogenesis
.
Fed. Proc. Fedn Am. Socs exp. Biol
.
38
,
2150
2153
.
Spudich
,
A.
&
Spudich
,
J. A.
(
1979
).
Actin in Triton-treated cortical preparations of unfertilized and fertilized sea urchin eggs
.
J. Cell Biol
.
82
,
212
226
.
Tilney
,
L. G.
(
1976
).
The polymerization of actin. III. Aggregates of non-filamentous actin and its associated proteins: a storage form of actin
.
J. Cell Biol
.
69
,
73
89
.
Tilney
,
L. G.
(
1978
).
The polymerization of actin. V. A new organelle, the actomere, which initiates the assembly of actin filaments in Thyone sperm
.
J. Cell Biol
.
77
,
551
564
.
Tilney
,
L. G.
,
Kiehart
,
D. P.
,
Sardet
,
C.
&
Tilney
,
M.
(
1978
).
Polymerization of actin. IV. Role of Ca2+ and H+ in the assembly of actin and in membrane fusion in the acrosomal reaction of echinoderm sperm
.
J. Cell Biol
.
77
,
536
550
.
Weibel
,
E. R.
(
1969
).
Stereological principles for morphometry in electron microscope cytology
.
Int. Rev. Cytol
.
26
,
235
302
.
Wigglesworth
,
V. B.
(
1931a
).
The physiology of excretion in a bloodsucking insect, Rhodnius prolixus (Hemiptera, Reduviidae). II. Anatomy and histology of the excretory systems
.
J. exp. Biol
.
8
,
428
442
.
Wigglesworth
,
V. B.
(
1931b
).
The physiology of excretion in a blood-sucking insect, Rhodnius prolixus (Hemiptera, Reduviidae). III. The mechanism of uric acid excretion
.
J. exp. Biol
.
8
,
443
451
.
Wigglesworth
,
V. B.
&
Salpeter
,
M. M.
(
1962
).
Histology of the Malpighian tubules of Rhodnius prolixus Stal (Hemiptera)
.
J. Insect Physiol
.
8
299
307
.