Using the electron microscope we have found axopods, a cell organelle previously undescribed in multicellular animals, in the lower Malpighian tubule of the insect, Rhodnius prolixus. The axopods, which are 0 · 2 to 0 · 8 μm in diameter and 10 or more μm in length, derive from the luminal surface of the tubule and contain an array of 1 to about 46 microtubules each. These microtubules arise within the cell near the cell junctions or near clumps of mitochondria.

Uric acid crystals which occur naturally in the lower tubule have been observed to move down the tubule under experimental conditions where peristalsis and fluid secretion can be ruled out. We suggest that the axopods are motile and serve to transport the crystals along the narrow tubule lumen. Since cilia are not found on somatic cells of arthropods, we suggest that axopods have evolved in the lower tubule to perform a function analogous to a ciliated epithelium in other animals.

Many cells possess fingerlike, motile extensions of the cell membrane that contain microtubules. Cilia and flagella, the most widely occurring microtubule-containing cell extensions, found in nearly all the animal phyla, are composed of an axoneme of 9 doublet microtubules arranged around a central singlet pair. In the arthropods, motile flagella occur only in the sperm, although certain sense organs contain centriolar derivations. A second motile microtubule-containing structure, so far reported only in the protozoa (Tucker, 1977), is the axopod. In contrast to cilia, the microtubules of an axopod are singlets and they do not arise from a centriole or basal body. Accordingly, the axopodial microtubule array does not exhibit 9-fold symmetry, but instead consists of relatively large numbers of regularly spaced microtubules.

While examining the ultrastructure of the lower Malpighian tubule of the bloodsucking insect Rhodnius prolixus, we were surprised to observe membrane-bounded microtubule-containing structures in the tubule lumen. As described below, we have determined that these structures derive from the apical surfaces of the cells of the lower tubule and that they are not centriole-related. On this and other morphological grounds they are apparently insect axopods. Further, these axopods may be motile and play an important role in transporting material down the Malpighian tubule. We are able to demonstrate directed movement of uric acid crystals in the tubule lumen and to provide some indications that such movement is caused by the axopods.

Material

The Rhodnius proUxus used for this research were taken either from a colony established by Dr Lauren Zarate of the Department of Entomology and Parasitology at the University of California, Berkeley, or from our own colony established with animals which were a gift from Dr Zarate. The colony was maintained at 27 °C, 40–60% relative humidity using a 12:12 light-dark regime, with all blood meals from rabbits. All experiments were carried out on fifth instar larvae which had been held without a meal for 1 month prior to use.

Light microscopy

Lower Malpighian tubules were removed from larvae which had been pinned to the wax-covered bottom of Petri dishes filled with insect Ringer (all values in mM: 11 KC1, 129 NaCl, 8 · 5 MgCI2, 2 CaCl2, 34 glucose) based on a formula used by Maddrell (1969). Portions of the one-third of the lower Malpighian tubule closest to the rectum were placed in a depression slide partly filled with insect Ringer. Care was taken to place the tubule close to the air-Ringer interface so that an oxygen supply was available to the tissue when a coverslip was placed over the depression. The tubules were observed using a Zeiss Photomicroscope 1 equipped with Nomarski interference-contrast optics.

Electron microscopy

Lower Malpighian tubules were either fixed in situ by dissecting larvae in 4 % glutaraldehyde in 0 ·1 M cacodylate buffer or they were removed from larvae which had been pinned under insect Ringer, and placed in a hanging drop preparation following the procedure of Bradley & Phillips (1977). This technique has been shown to sustain ion transport for 3 h, although the longest experimental period used in this study was 10 min. The observations on axopods discussed were conducted during a study of hormone regulation of mitochondrial movements in the lower tubule (Bradley & Satir, in preparation); therefore, the hanging drops were composed of insect Ringer containing either 11 or 35 mM KC1 with and without 5-hydr0xytryptamine (5-HT). The various treatments were not found to change the basic axopod structure, except that 5-HT does cause mitochondria to move into the axopods. After the to min treatment, the tubules were placed in 4% glutaraldehyde in 0 · 1 M cacodylate buffer (pH 7 · 4) followed by 1 % OsO4 in 0 ·1 M cacodylate buffer. After 10 min of en bloc staining in 1 % uranyl acetate in 70 % ethanol, the tubules were dehydrated in an ethanol gradient and embedded in Epon 812. Sections were stained with uranyl acetate in 50 % ethanol followed by Reynolds lead stain and observed and photographed on a Siemens 101 electron microscope.

Morphology

The lower Malpighian tubule of Rhodnius, when viewed in the electron microscope, contains 2 clearly different types of organelles in the lumen (Fig. 1). The first are microvilli, as described by Wigglesworth & Salpeter (1962), which arise from the apical surface of the cell, vary from 2 to 5 μm in length, contain microfilaments and may also contain mitochondria or extensions of endoplasmic reticulum. The ultrastructure of the microvilli, as well as the factors which control their size and the movement of mitochondria into them are discussed elsewhere (Bradley & Satir, in preparation). The second structures observed in the lumen are cross-sections of what appear to be microtubule-containing organelles. These structures are the axopods, so called for reasons presented briefly in the Introduction and discussed below. The axopods vary in diameter from 0 · 2 to 0 · 8 μm and may contain from 1 to about 46 microtubules (Figs. 1-3). The microtubules are arranged in an inexact array, with a mean distance between adjacent microtubules of about 50 nm (range 49–65 nm). Bridges between microtubules which could maintain this array are not preserved by our present methods. In addition to the microtubular array, microfilaments can be observed in the axopods, particularly near the membrane (Fig. 2). The axopod membranes are typical unit membranes that resemble those of the adjacent microvilli. The exterior surface of both membranes shows an electron-dense coat.

Symbiotic protozoa are not uncommon in insect Malpighian tubules (Grell, 1973). We looked for such organisms in the lumen to explain the possible origin of the axopods. These were not observed. Instead, we found that the bases of the axopods were located in among the microvillar bases at the luminal surface of the tubule cells (Fig. 4). The axopods are not derived from intracellular symbionts either. Since organelles the size of mitochondria can move back and forth from cytoplasm to axopod, we conclude that the axopods are extensions of the cytoplasm of the Malpighian tubule cell.

The arrangement of microtubules in the axopods of Rhodnius does not show a 9 + 2 or related configuration (Figs. 1, 2). Even cross-sections made very near the base show neither a basal body nor arrangement of microtubules similar to that in ciliary derivatives (Fig. 3). To be certain that a basal body was not being overlooked, we examined serial sections made through the bases of axopods. Figs. 5-7 are 3 such micrographs, taken from a series of serial sections. Fig. 5 shows the base of an axopod and a portion of the adjacent apical region of the cell. Figs. 6 and 7 show the portion of the cell from which the axopod is arising 2 sections above and below Fig. 5. No basal body can be observed in the section which shows the place of attachment of the axopod, nor in the section on either side in the series. Instead, the microtubules which enter the axopods are observed to be extensions of a broad array of microtubules which penetrate deep into the cell. These microtubule arrays can be followed for considerable distances into the cell and are observed to terminate either near the septate or continuous junctions between cells (Fig. 5) or in clusters of mitochondria (Fig. 3). No specific end-on connexion of the microtubules to any other cell structure has been observed. The individual microtubules in the axopods of Rhodnius are probably quite long since they can be followed for considerable distances, with the maximum length observed in a single section being 4 μm.

The axopods differ from the axopods of protozoa in that as they extend outward from the cell surface they branch, yielding cross-sectional views of many sizes. Although we have not yet followed individual microtubules through this progression, it seems likely that the microtubules continue without interruption from the cell body out into the axopods, splitting into ever smaller groups as the axopod branches. Nothing is at present known about the tips of the axopods and whether the microtubules interact with the membrane at that point. No intimate association of the microtubules and axopod membrane has been observed along its length.

Uric acid crystal movement

The upper Malpighian tubule of Rhodnius is the site of haemolymph filtrate formation by active secretion of potassium and sodium chloride (Wigglesworth, 1931a; Maddrell, 1969). Uric acid, the primary form of nitrogenous waste in this animal, enters the upper tubule lumen as well, but in soluble form, presumably as the potassium salt (Wigglesworth, 1931c). Uric acid crystals precipitate in the lower tubule as a result of acidification of the urine, releasing potassium and water for resorption if needed. The lower tubule, therefore, contains an accumulation of large uric acid crystals which are easily visualized in the light microscope.

While performing dissections in Rhodnius which had been submerged in insect Ringer, we observed rapid crystal movements in intact tubules in situ, while in severed tubules, which retained their tracheal attachments, such movement resulted in crystals being dumped out of the open end of the tubule. The observation of crystal movement was unexpected. The flow of fluid in the tubules should be minimal, since the animals had been starved for 1 month and fluid secretion under these conditions is repressed (Maddrell, 1963). We therefore investigated the mechanism of crystal movement more carefully. Portions of lower tubules from the region nearest the rectum were cut into short lengths and placed in insect Ringer (see Methods). These in vitro preparations were open at both ends and isolated from fluid movements produced in other regions (e.g. secretion by the upper tubule; peristalsis in the rectum). While spiral muscles are occasionally seen on the portion of the Malpighian tubule nearest the rectum, crystal movements are observed in other regions of the tubule as well, and no contractile movements were observed during the in vitro experiments.

Under these conditions, in many preparations, crystals could still be seen to move very rapidly in one direction. The crystal movement is steady and rapid with slight irregularities in the speed of individual crystals, suggesting that as the crystals are carried in a current of fluid, they bump into the walls of the tubule as they pass along it. The tubules had to float freely in the medium so that no kinks were produced in the lumen, otherwise only that portion ‘downstream’ from the blockage showed crystal movement. Another critical factor for optimal movement was proximity to the air/ Ringer interface, presumably because of an oxygen requirement. Like most insect tissues, the tubules are tracheated in vivo, and therefore are physiologically adapted to a steady high oxygen concentration.

When the lumen is filled with crystals, a stream of crystals flows down the tubule en masse. Figs. 8 and 9, taken 5 s apart, show the forward edge of a mass of crystals moving along a tubule in vitro. The mean speed calculated from these 2 micrographs and a third not shown was 0 · 28 mM/min. Since the lower tubule is about 11 mM long, a crystal could traverse its length at this speed in about 39 min. Crystals have been observed to flow along a tubule for many minutes, with one end of the tubule segment being emptied of crystals while a steady stream of crystals pours out of the opposite cut end. It is not yet known whether the direction of movement corresponds to that of normal ‘downstream’ flow in the tubule in vivo.

An indication of the nature of crystal movement when the lumen is relatively empty of crystals is shown in Figs. 10-13. Figs. 10-12 are light micrographs of crystal movement within the lumen, taken at 5-s intervals. The large arrow points to a stationary crystal in the lumen and the movement of a small group of crystals relative to that point can be observed. The rate of crystal movement here is about 0 · 40 mM/min, 45% faster than that of crystal masses (see above). In Fig. 13, a photograph of the same region taken within 1 min of the time Fig. 12 was taken, all the crystals except for the stationary one have been propelled out of this region of the tubule. By focusing up and down, one can determine that most of the stationary crystals are not in the open lumen but are lodged between the microvilli near their bases.

Several points argue against peristalsis being the motive force for crystal movement:

(1) Crystal movement in the isolated tubule is non-pulsatile. Peristaltic movements are generally pulsatile. (2) In sequential micrographs taken many seconds apart (i.e. Figs. 8-9; 10-12) no tissue movement can be observed, indicating that no peristalsis, tubule bending or contraction has occurred. Similarly, because the lower tubule is not a secretory region of the tubule, but rather a resorptive one (Maddrell, 1978), the flow of crystals out of the end of the tubule cannot be explained by the known ion and fluid transport capabilities of the tubule.

We have observed the microvillar masses which can be discerned in Nomarski optics. The microvilli do not move in a coordinated manner or for extensive distances during crystal movement in the lumen. When a crystal is lodged among the microvilli it is stationary. The microvilli exhibit a slight waving motion, which seems insufficient to explain the crystal movement. This microvillar movement was observed by Wigglesworth (1931 b) as well, who considered it to be a passive response to fluid movements.

High-resolution observations of axopod movement in the light microscope have been difficult because of the thickness of the in vitro tubule preparation. We have sometimes been able to see movement of fingerlike projections in the lumen, while microvillar masses are relatively stationary. We believe, on the basis of their size and shape, that these structures are the axopods. However, we cannot completely rule out the possibility that they represent motile bacteria or unusually long microvilli, or that these presumed axopod movements are due to Brownian movement. The flow of crystals on the other hand is clearly not due to Brownian movement because (1) it is highly directional, and (2) while the smaller crystals display Brownian movements both inside and outside the tubule, the larger crystals which exhibit jiggling and unidirectional flow in the lumen are perfectly still when placed on a slide in fluid outside the tubule.

We conclude that the force propelling the uric acid crystals in the lower tubule is sensitive to tissue oxygen supply but independent of tissue contraction. Crystal movement occurs only in the open lumen, not among the microvilli. Fluid secretion has not been observed in this tissue and seems incapable of explaining unidirectional flow in a tubule open at both ends. Since the long axopods we have described here occur in the open lumen where crystal movement is seen to take place, and contain distinctive arrangements of microtubules and microfilaments associated with many forms of cell motility, we believe that there is a strong likelihood that the crystals are propelled by beating movements of the axopods.

To our knowledge this paper is the first report of an axopod in the tissue of a multicellular animal. The luminal border of the lower Malpighian tubule of the insect Rhodnius prolixus is an important ion-transporting surface that has received previous attention. The epithelial cell microvilli have been described by Wigglesworth (1931 b) and Wigglesworth & Salpeter (1962). The latter description is an early electronmicroscope study that antedates glutaraldehyde fixation. The axopods were not observed, possibly because microtubules were not preserved. Using fixed tissue under the light microscope, Wigglesworth (1931 b) did observe occasional long microvilli which extended past the masses of shorter microvilli and into the centre of the lumen. Although the limitation of resolution makes the details of identification obscure, these were probably the axopods. Wigglesworth also observed that feeding in-vivo caused the microvilli in the lower tubule to elongate. We can induce a similar microvillar growth in vitro using 5-HT (unpublished observations). It is not known whether the axopods also change in length upon stimulation.

The most distinctive characteristic of the axopods is the microtubule array they contain. In addition, Rhodnius axopods also contain microfilaments which run close to and parallel to the plasma membrane. We have shown that the microtubule arrays in Rhodnius axopods share many characteristics with protozoan axopods. The microtubules are all singlets, are evenly spaced in cross-section and do not show a pattern in any way similar to the 9 + 2 pattern of cilia and flagella. No basal body structure exists at the base of the axopod. Instead, the microtubules extend as an array deep into the cell, terminating near a cell junction or mitochondrial cluster. We feel that these characteristics justify our claim that the structures are not ciliary in derivation but should be classified with the axopods of protozoa.

In protozoan axopods, the microtubules serve as a framework along which material is transported (cf. Edds, 1975). In Rhodnius axopods, the microtubules similarly show interactions with mitochondria moving along them (Bradley & Satir, in press).

We have presented a considerable body of evidence which suggests that the Rhodnius axopods are motile. While we have observed objects moving in the lumen of the tubule which are the shape and size of the axopods, we cannot be certain that these movements are active. The most powerful evidence that some structural element in the lumen is motile is provided by a natural marker in the lumen, the uric acid crystals. These crystals can be observed to move rapidly along the lumen in isolated tubules under conditions where peristalsis and fluid flow due to secretion can be ruled out. Protozoan axopods are also thought to be motile or contractile in certain circumstances (Davidson, 1975).

The lower Malpighian tubule of Rhodnius is known to be both the site of KC1 resorption from, and acidification of, the primary urine secreted by the upper tubule. Maddrell (1978) has recently shown that specifically the lower one third of the lower tubule is the site of KC1 resorption and is very impermeable to water. It is this region to which we have confined our experiments and in which we have found the axopods. While Maddrell did not assign a direct function to the region closest to the upper tubule, we would suggest that this is where the urine acidification occurs, since the uric acid crystals first appear at this location (Wigglesworth, 1931 c). Presumably the osmotic concentration of the urine is reduced as crystallization occurs, and water is then passively resorbed. In other insects, salt and water resorption occur in the rectum (Phillips, 1970) which can expel solid waste by peristalsis. In Rhodnius the lower Malpighian tubule supercedes the rectum as the major site of ion intake during rapid diuresis. Maddrefl & Phillips (1975) suggest that the tubule is better suited for this purpose than the rectum because of the extensive membrane surface presented by the microvilli, and the narrow lumen which assures rapid flow and minimal unstirred layers. As a consequence of this system, however, the tubule lumen becomes filled with large crystals which can easily lodge and block fluid flow and in which peristalsis as it occurs is weak. In most biological systems where solids are propelled in narrow channels lacking peristalsis, cilia serve to loosen packed material and move suspended solids (e.g. fallopian tube, lung trachioles, molluscan gill surfaces) (Satir, 1977). The arthropods, however, do not possess cilia in any of their somatic cells. We suggest that Rhodnius has evolved the axopods, which are presumably found only in the lower tubule where the uric acid crystals are located, as a form of parallel evolution to cilia.

In this regard, it would be interesting to know whether the microtubules present in the Rhodnius axopods are directly involved in the production of movement, as are ciliary microtubules, or serve only as a framework (Edds, 1975) to which microfilament (i.e. actin-myosin)-generated or other force is applied. Such information would shed some light on how Rhodnius axopods might have evolved. Microfilament-powered movement would suggest that microvilli which already show motile capabilities (see Mooseker, 1976) and insect axopods are intimately related. Microtubule-driven motility would suggest a different evolutionary, as well as physiological, mechanism possibly involving dynein-like molecules. In order to begin answering these questions we are presently using heavy meromyosin to determine whether the microfilaments present in the axopods are composed of actin and what effect cytochalasins, colchicine and dynein antagonists have on axopod movement and crystal transport.

The authors would like to thank Drs Ralph Smith and Richard Eakin for valuable discussions during the course of the work. We also are grateful to Dr Simon Maddrell for kindly supplying a preprint of one of his publications. Part of these studies were conducted at the Department of Physiology-Anatomy, University of California, Berkeley, U.S.A. This work was supported by a grant from the USPHS (HL 22560). T. Bradley is the recipient of NIH postdoctoral fellowship AM 05499.

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
.
Davidson
,
L. A.
(
1975
).
Studies of the Actinopods Heterophrys marina and Cilophrys marina
.
Ph.D. Thesis
,
University of California, Berkeley
.
Edds
,
K. T.
(
1975
).
Motility in Echinosphaerium nucleofilum I. An analysis of particle motions in the axopodia and a direct test of the involvement of the axoneme
.
J. Cell Biol
.
66
,
145
155
.
Grell
,
K. G.
(
1973
).
Protozoology
.
Berlin
:
Springer-Verlag
.
Maddrell
,
S. H. P.
(
1963
).
Excretion in the blood sucking bug Rhodniusprolixus Stal. I The control of diuresis
.
J. exp. Biol
.
40
,
247
256
.
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.
(
1978
).
Physiological discontinuity in an epithelium with an apparently uniform structure
.
J. exp. Biol
.
75
,
133
145
.
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
.
Mooseker
,
M. S.
(
1976
).
Brush border motility. Microvillar contraction in triton-treated brush borders isolated from intestinal epithelia
.
J. Cell Biol
.
71
,
417
433
.
Phillips
,
J. E.
(
1970
).
Apparent transport of water by insect excretory systems
.
Am. Zool
.
10
,
413
436
.
Satir
,
P.
(
1977
).
Microvilli
and cilia
:
Surface specializations of mammalian cells
.
In Mammalian Cell Membranes
, vol.
2 (ed
.
G. A.
Jamieson
&
D. M.
Robinson
), pp.
323
353
.
London
:
Butterworth
.
Tucker
,
J. B.
(
1977
).
Shape and pattern specification during microtubule bundle assembly
.
Nature, Lond
.
266
(
5597
),
22
26
.
Wigglesworth
,
V. B.
(
1931a
).
The physiology of excretion in a blood-sucking insect, Rhodnius prolixus (Hemiptera, Reduviidae). I. Composition of the urine
.
J. exp. Biol
.
8
,
411
427
.
Wigglesworth
,
V. B.
(
1931b
).
The physiology of excretion in a blood-sucking insect, Rhodnius prolixus (Hemiptera, Reduviidae). II. Anatomy and histology of the excretory system
.
J. exp. Biol
.
8
,
428
442
.
Wigglesworth
,
V. B.
(
1931c
).
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
45 1
.
Wigglesworth
,
V. B.
&
Salpeter
,
M. M.
(
1962
).
Histology of the Malpighian tubules in Rhodnius prolixus Stal. (Hemiptera). J’
.
Insect Physiol
.
8
,
299
307
.