ABSTRACT
Solute concentrations were measured in the frontal bodies and hypopharyngeal bladder surfaces, which form the absorption system of Arenivaga. Measured values were compared with the concentrations necessary to lower relative humidity to the absorption threshold, 81% R.H.
The predominant inorganic solutes, measured by instrumental neutron activation analysis, were Na, K, and Cl. Their concentrations were 2–3 orders of magnitude below those necessary for absorption.
Quantities of a variety of hydrophilic organic molecules were determined with fluorometric reagents. Concentrations of free amino acids, peptides, reducing sugars and polyhydroxyl alcohols were negligible in frontal bodies and on bladder surfaces.
Frontal bodies did not appear to be active sites of protein or polysaccharide synthesis, because they did not accumulate significant amounts of radio-labelled amino acids and glucose injected into the hemolymph.
The results of this and previous studies indicate strongly that absorption in Arenivaga differs markedly from the solute-dependent schemes which have been proposed for other arthropods.
INTRODUCTION
The desert cockroach is the only insect in which buccal structures are known to be involved in water vapour absorption (O’Donnell 1977a, 1980). In relative humidities (RH) above a threshold value of 81 %, vapour condenses onto a fluid layer covering two protrusible bladders, which are lateral diverticula of the hypopharynx (Figure 1). The fluid is held in the interstices of fine cuticular hairs which densely cover the bladder surface. It is composed of condensate, and the product of a pair of structures called frontal bodies which are situated beneath the frons and connected to the bladders by a groove in the epipharynx (O’Donnell 1977b, 1981). Interrupting this connection causes the associated bladder to become dry. The fluid is necessary for absorption because vapour does not condense onto dry bladders (O’Donnell, 1977a).
Recent studies in tenebrionids and acarines have suggested that the necessary reduction in the thermodynamic activity of water is established by high solute concentrations at, or near, the absorption site. Ramsay (1964) measured potassium chloride concentrations in excess of 2 M in the Malpighian tubular fluid of Tenebrio molitor. The freezing point depression of the perirectal fluid was 8 °C (Ramsay, 1964; Machin, 1979b), and gradients of osmotic pressure were present, increasing posteriorly along the rectum and radially towards the Malpighian tubules. On the absorbing surfaces of dried animals, crystalline deposits near the salivary glands of ticks contain significant amounts of potassium (Rudolph & Knulle, 1974). Large quantities of potassium chloride are associated with the pre-coxal glands of dried mites (Wharton & Furumizo, 1977).
The predominance of solute-dependent models of absorption and epithelial transport (Diamond and Bossert, 1967) in other species suggested that structures associated with absorption in Arenivaga should be analysed to determine if high solute concentrations exist. Many suggestions have been made concerning the properties of solutes which could actively be transported to lower water’s thermodynamic activity. Inorganic ions, low molecular weight organic molecules and macromolecules have all been proposed (Machin, 1979a; Noble-Nesbitt, 1977, 1978). Madrell (1971) suggested that such substances must be highly soluble in water and possess many sites at which hydrogen bonding can occur so as to reduce water activity. For absorption in Thermobia, at humidities as low as 45 % RH, he considered it doubtful that any non-toxic inorganic molecule could be used. Even for absorption by T. molitor, against much smaller gradients, active transport of KC1 by the leptophragmata must be achieved exclusive of water move-ment from other tissues which would dilute the established gradient. Cells would also have to resist membrane depolarization in such a high extracellular potassium environment (Machin, 1979a).
Certain organic molecules, on the other hand, would not lead to toxic effects and ionic imbalance, and may also be tolerated by living cells at sufficiently high concentrations (Machin, 1979a). Glycerol was suggested as a possibility by J. Diamond (personal communication to House, 1974) and T. Weis-Fogh (personal communication to Noble-Nesbitt, 1977).
Water activity could also be lowered by high concentrations of macromolecules, especially those which have large osmotic coefficients and show strongly non-ideal solute behaviour (Machin, 1979a). Water activity could be reduced either directly by the macromolecules, or indirectly by their enzymatic breakdown to produce a high concentration of smaller molecules. High concentrations of amino acids and saccharides might be produced by de-polymerization in the cuticle (Locke, 1964). It has also been suggested that the sub-cuticular mucopolysaccharide in the rectum of Thermobia might be a component of the absorption mechanism (Noble-Nesbitt, 1978). Reabsorption would facilitate removal of the entrained water.
Solute concentrations sufficient to lower water activity to the levels necessary for absorption are extremely high. For example, 11700 mm of an ideal solute will equilibrate with 81% RH, the threshold for absorption, by Arenivaga. For comparison, saturated KC1 (4600 mm) equilibrates with 85 % RH, and saturated NaCl (6250 mm) equilibrates with 75·5% RH (Weast, 1968). It therefore seemed feasible to test a solute-coupled hypothesis of water vapour absorption in Arenivaga through measurement of organic and inorganic solute concentrations in the frontal bodies and bladder fluid.
MATERIALS AND METHODS
Inorganic solutes were analysed by multi-element instrumental neutron activation analysis (INAA), and by atomic absorption spectrophotometry (AAS). Hydrophilic organic molecules were detected by fluorescent labelling techniques for amino acids, peptides, reducing sugars and polyhydroxyl alcohols. Macromolecules were detected through fluorescent labelling of active groups, and through incorporation of radio-labelled precursors.
Sample preparation
Frontal bodies and hypopharyngeal bladders were excised from absorbing animals which had been fast-frozen during absorption and lyophilized (O’Donnell, 1980, 1981).
For comparing levels of inorganic elements in bladder fluid and in bladder tissue, bladders were everted by squeezing the abdomen and washed, before freezing, with 5 ml of water forcibly ejected from a syringe through a 26 gauge needle. Fluid was also collected on Millipore strips (0·1 μm pore size) which had been moistened and applied to the surface of protruded bladders.
Samples were weighed in platinum boats over silica gel to the nearest 0·1 μg using an electronic microbalance (Mettler ME22).
Instrumental neutron activation analysis
Lyophilized samples, or Millipore strips used in collecting bladder fluid, were placed in polythene vials and irradiated in the low flux SLOWPOKE-2 (Safe Low Power Critical Experiment) nuclear reactor (Atomic Energy Canada Ltd., Commercial Products, Ottawa) at the University of Toronto. Details of the reactor are described by Kay et al. (1973) and its use in multi-element analysis by Hancock (1976). Irradiation at 20 kW at the number 1 site in the reactor was equivalent to a neutron flux of 1012 neutrons/cm2 s−1. For common biological elements 20 kW irradiation for 10 minutes was sufficient. For greater sensitivity, and especially for potassium, samples were also irradiated at 5 kW for 16 h.
Atomic absorption spectrophotometry
Measurements of potassium levels by INAA were corroborated by atomic absorption spectrophotometry. Frontal bodies were dissected from fast frozen animals, macerated with tungsten needles and centrifuged in flame-sealed micropipettes containing immersion oil. Fluid volumes were calculated from drop diameters in oil, prior to addition of 1 ml distilled water. The spectrophotometer, a Varian Techtron AA6 (Springvale, Australia), was calibrated with solutions of Analytical Reagent grade potassium chloride.
Analysis of primary amines
Amino acids in frontal bodies and bladder fluid were detected as the highly fluorescent compounds produced upon reaction with O-pthalaldehyde (OPA; Roth, 1971). A stock solution of OPA was prepared as a 15 mg % solution (w/v) in glass distilled water with a drop of 1 N NaOH added for every 200 ml. The reagent solution, prepared just before use, consisted of 20 ml stock OPA and 0·01 ml 2-mercaptoethanol (Stephens et al., 1978). For peptides, sensitivity was maximized by adding 0· 01 ml triethanolamine to 15 mg % OPA stock. Standard solutions were applied to a strip of Millipore filter (1 cm × 0· 5– 1· 0 mm) and allowed to dry. OPA reagent (0·5–1·0 μl) was then pipetted onto the same region of the strip. For peptides, a 10% aqueous solution of triethanolamine was applied after the OPA solution. For determination of amino acids in bladder fluid, OPA reagent was applied directly to the surface of protruded bladders of animals maintained in humid conditions in transparent chambers. Alternatively, OPA was applied to 0·1 μm pore size Millipore strips which had been moistened and applied one or more times to the bladder surface (O’Donnell, 1980). Filter strips were examined under a long wave ultraviolet lamp. The limit of detection was determined for several solutions of amino acids and peptides applied to the strips.
Analysis of reducing sugars
DANSYL hydrazine (i-naptholenesulfonyl 5-dimethylamino hydrazine; Sigma, St Louis, Missouri) was used as a sensitive fluorometric reagent for detection of reducing sugars. Over 90% of a glucose standard is converted to glucose DANSYL hydrazone (Avigad, 1977). Because DANSYL hydrazine is itself fluorescent, reaction products were separated by thin layer chromatography (TLC). ‘
As a standard, glucose (0·05–2·0 μmol in 100 μl water) was added to 100 μl of 0·1 M/1 acetic acid and 200 μl of DANSYL hydrazine in a 2 ml plastic vial, heated to 80 °C for 10 min, then cooled to room temperature and spotted on the plate.
Reducing sugars in lyophilized frontal bodies were extracted in 5–10 μl of water. Tissues were either macerated in Reacti-Vials (Pierce Chemical Co., Rockford, Illinois) using a glass rod ground to fit the conical vial, or were ultrasonicated, frozen, then thawed, three times. Volumes were reduced to 0·5–1·0 μl by evaporation under partial vacuum, then mixed with 2 μl acetic acid (0·05 M) and 2 μl DANSYL hydrazine (1 % in ethanol). Samples were heated, and in some cases volumes reduced to less than 1 μl by evaporation. Proteins were first precipitated from hemolymph samples, either by heating (70 °C for 5 min) or by addition of an equal volume of 0·1 M acetic acid. Samples (1 μl) were spotted onto high performance thin layer chromatographic plates (HPTLC) under a stream of warm air. Detection limit of glucose by HPTCL plates developed in chloroform-ethyl acetate – 1% boric acid (3:5:2; Avigad, 1977) was 30 pmol. 1 μl aliquots of the reagent mixture were spotted on the HPTLC plate under a stream of water air. The detection limit was extended by applying the super-natant collected from the centrifuged homogenate of as many as ten frontal bodies to one spot on the plate.
Analysis of polyhydroxy alcohols
DANSYL hydrazine has also been used to detect aldehydes produced by periodate oxidation of glycoproteins in polyacrylamide gels (Eckhardt, Hayer & Goldstein, 1976). A method was developed to produce polyhydroxy alcohols by periodate oxidation. Frontal bodies were homogenized and added to an equal volume of aqueous periodate (0·5%), then heated at 75 °C for 30 min. A small volume of 0·5% sodium metabisulfite in 5 % aqueous acetic acid was added to destroy excess periodate.as standards. Visual detection limits on HPTLC plates were 1 and 10 nmol for mannitol and glycerol, respectively.
The mixture was then treated as for reducing sugars. Mannitol and glycerol were used as standards. Visual detection limits on HPTLC plates were 1 and 10 nmol for mannitol and glycerol, respectively.
Incorporation of 14C-labelled compounds
5 μl each of 14C-labelled glucose, alanine and an amino acid mixture were injected into the haemocoel of desiccated animals through the arthrodial membrane between the forelegs. Animals were placed in glass vials which were humidified by plugging with moistened cotton wool. They were fast-frozen at intervals after injection of the label, and then decapitated, lyophilized, dissected with tungsten needles, and weighed. Tissues were placed in 1 ml of Protosol (New England Nuclear Corp.) and heated to 55 °C for 18–24 h in glass scintillation vials with polythene-lined aluminium screw caps. Nine ml of scintillation fluid (6 g diphenyloxazole/litre toluene) were added and samples analysed by LSC. Aliquots of hemolymph (usually 5 μl) were collected in micropipettes from an incision in the arthrodial membrane between the forelegs. The haemolymph was dispensed into 1 ml of Protosol and the pipette broken and included in the vial. Washed bladders were prepared as above.
Bladder fluid was also collected on Millipore strips after injection of 14C-labelled glucose and amino acid mixture.
RESULTS
Inorganic solutes
Instrumental neutron activation analysis showed that inorganic material contributes only a small part to the total dry weight of frontal bodies and bladders, the principal elements (Na, K, Cl) all constituting less than 4% dry weight (Table 1). The basis for the apparent increase in the proportion of each of these elements in washed bladders, and estimated elemental concentrations (Table 3) will be discussed. Calcium, magnesium and bromine were present in trace amounts, less than 0·06 % of total dry weight for each. Iodine was present in significant quantities in bladders from absorbing animals (0·07%) and in washed bladders (0·04%) but was undetectable in frontal bodies. No elements were detectable in Millipore strips applied to the bladder surface.
Potassium concentrations in fluid from frontal bodies were measured by AAS, which gave values of 99 ± 49 mequiv/1 (X±SD ; n = 5). The large deviation of individual values from the mean is possibly due to errors inherent in manipulating small samples (5–23 nl) and in separating the frontal bodies from the surrounding muscle tissue.
INAA was used to determine the maximum volume of fluid held in spaces between hairs on the bladder surface. The quantity of the manganese chloride marker solution retained on the bladder surface was a constant proportion of the dry weight of the bladder. Calculations indicated that 6·2 ± 0·1 nl/μg dry weight were retained on the bladder. For a typical animal weighing 364 mg, a maximum of 273 ± 6 nl were retained on both bladders, or 137 nl/bladder.
Organic solutes
No amino acids were detectable by OPA applied either directly to the bladder of absorbing animals or to Millipore strips used to collect fluid from the surface. The maximum undetectable concentrations of amino acids and peptides present in the bladder fluid were therefore estimated by determining the sensitivity of OPA to amino acids, peptides or proteins applied to the bladder surface. Application of OPA onto the bladder surface within 5 s of a preceding application of several nanoliters of 10 mm glycine, 0·1 mg/1 bovine serum albumin, or 1 mg/ml trypsin by micropipette showed clear, visible fluorescence. OPA was an extremely sensitive detector of amino acids, peptides or proteins on Millipore strips (Table 2) which had been used to collect fluid after application of known quantities of the experimental solutions to the bladder surface.
There were no reducing sugars in the frontal bodies detectable by DANSYL hydrazine. The detectable limit of glucose ( < 30 pmol/spot) was used to estimate the maximum undetectable amount of reducing sugar present in frontal bodies, 3–30 pmol. The technique readily detected a sugar with the same Rf as glucose (0·18) in muscle tissue. Reducing sugars were also detectable in haemolymph.
Any polyhydroxyl alcohols present in the frontal bodies were also at concentrations below detectable limits. Limits of detection of glycerol and mannitol were 10 and i nmol/spot, respectively.
It was concluded that amino acids, peptides, proteins and reducing sugars were undetectable because they were either absent, or present in only trace amounts, and not because of deficiencies in the analytical techniques.
Calculation of solute concentrations
Solute concentrations were calculated from the dry weights (Tables 1, 2) by estimating the appropriate solvent volumes. Measurements of a marker solution indicated that for each microgram dry weight of bladder tissue, 6·2 nl of an aqueous solution is retained by the cuticular hairs. As an example, the 801 μg sample of bladder tissue (pooled from 20 absorbing animals) would retain a maximum of 801 × 6·2 = 4966 nl of fluid. It was assumed that all solutes were dissolved in this estimated volume, and that none were intrinsic to bladder tissue ; estimated concentrations are therefore maximums.
Incorporation of 14C-labelled compounds
Following their injection into the haemocoel, there was no significant incorporation of 14C-labelled glucose, alanine or amino acid mixture into the frontal bodies or bladders (Fig. 2). Approximately the same level of each compound was present in washed and unwashed bladders (Fig. 2b), indicating that most of the compound was present in bladder tissue rather than bladder fluid. Concentrations were below those in haemolymph or muscle. Quantities of 14C-labelled amino acids in bladder fluid collected on Millipore strips within 2 days of injection (35–43 cpm/strip; 46 strips applied 1–5 times to bladders; n = 6 animals) were not significantly elevated above background levels of radioactivity (36 cpm/strip).
DISCUSSION
The absorption mechanism of Arenivaga differs fundamentally from the solute dependent systems which have been proposed for tenebrionids (Machin, 1979 a, b) and acariñes (Rudolph & Knulle, 1974; Wharton & Furumizo, 1977).
A comparison of maximal solute concentrations in frontal bodies and bladder fluid shows that concentrations are insufficient to lower water activity to the absorption threshold, 81% RH (Table 3), even when the most unfavourable allowances for experimental error are used.
Inorganic solute concentrations are unexpectedly higher in washed bladders than in bladders from absorbing animals, presumably because haemolymph tended to adhere to the inner surface of bladders forcibly everted before freezing and lyophilizing. Haemolymph chloride concentration (i28meq/l; Edney, 1966) represents 6–7% of the dry weight of haemolymph, considerably more than is present in bladders (1·2%) and frontal bodies (0·8%). Because the calculated concentrations in bladder fluid assume all of each element is in the bladder fluid, whereas in fact most is in the tissue, the concentrations in Table 3 are overestimates.
Corroboration of the estimated potassium concentration from the % dry weight determined by INAA (107 mm) is provided by the similar value obtained by atomic absorption spectrophotometry of wet tissue extracted from fast frozen animals.
The unexpectedly high amount of iodine in bladder samples (0·04–0·07% dry weight) is also more likely contained in the cuticle, rather than in the fluid layer. Iodinated amino acids are accumulated in cuticle (Limpel & Casida, 1957a, b), and may be involved in cuticle hardening (Tong & Chaikoff, 1961). If the iodine were in ionic form, its maximum concentration would be 0·5 and 0·2 mm in absorbing and washed bladders, respectively.
Nor can organic solutes lower water activity sufficiently for absorption. Only 2–4% of the total dry weight of bladders and frontal bodies is inorganic; the major proportion is organic. However, virtually all of this consists of chitin and structural proteins, which have been demonstrated histochemically (O’Donnell, 1981). In bladder fluid and frontal bodies, concentrations of free amino acids, glucose, or their derivatives, as indicated by direct chemical analysis and by accumulation of the corresponding 14C-labelled compounds, are negligible. Concentrations of the labelled molecules per microgram dry weight of frontal bodies or bladders was generally one-fifth that in the haemolymph (Fig. 2), suggesting that there is no synthesis of proteins, glycoproteins or long-chain carbohydrates within the frontal bodies during absorption.
These results support the hypothesis, based on ultrastructural analysis and measurements of internal osmolalities (O’Donnell, 1980, 1981) that the frontal bodies do not function as secretory structures or salt glands, but as hydrostatic pumps which produce an ultrafiltrate of the haemolymph. The frontal body plate, across which filtration occurs, has a very low permeability to molecules as small as glucose (O’Donnell, 1981). It is therefore unlikely that macromolecules could be transferred out of the frontal bodies in high enough concentrations to form a solution of reduced water activity.
Measurements of osmolalities within the frontal bodies have been reported elsewhere (O’Donnell, 1981); they support the hypothesis that these structures do not produce a solution which is sufficiently concentrated to directly absorb water vapour. Measured values (645 mOsm/kg; O’Donnell, 1981) exceed the value approximated by summing the solute concentrations (275–300 mOsm/kg), presumably because of the effects of proteins or other macromolecular components on ice crystal formation and therefore, freezing point depression.
It may be argued that as yet unconsidered solutes, particularly those of low molecular weight, are involved during absorption Arenivaga. However, osmolalities within frontal bodies, the source of bladder fluid, indicate that solute concentrations are not elevated. Moreover, many solutes can be excluded on the basis of their physical properties and physiological effects. Many of the alcohols are either too toxic or too volatile to be present in a solution of low vapour pressure (Machin, 1979a). Although amino acid concentrations of 0·29–2·43 g % occur in insect haemolymph (Wyatt, 1961), most of the amino acids are too insoluble to be present in the concentrations required for water vapour absorption. Concentrations at saturation range from 4250 mm for glycine and 2250 mm for alanine, to as low as 70 mm for aspartic acid (Weast, 1968). A saturated solution of proline, the most soluble amino acid (saturated at 12400 mm) could absorb water vapour from humidities as low as 78% RH (Machin, 1979a).
A number of aspects of the behaviour of the absorption system are inconsistent with the lowering of water activity by high solute concentrations, and therefore support the results above. The drying of the bladder surface which follows interrupting the supply of frontal body fluid to the bladders is incompatible with a solute dependent absorption mechanism. A suitably concentrated solution on the bladder surface would not dry out, but would equilibrate with ambient humidity.
A further observation concerns the appearance of dried bladders. No crystalline deposits are evident; such deposits have been found on the absorbing surfaces of acariñes (Rudolph & Knulle, 1974; Wharton & Furumizo, 1977) and are cited as evidence for the use of concentrated solutions during absorption.
Clearly, the results presented here strongly imply that fluid, produced by the frontal bodies and conveyed to the bladder surface, plays a subordinate role in the mechanism of water vapour absorption, perhaps aiding condensate removal so as to facilitate continuous absorption. The primary function of reducing water activity may, therefore, reside in the cuticular hairs which cover the bladder surface. The water affinity of the hairs is currently under study.
ACKNOWLEDGEMENTS
I am grateful to A. J. Forester, J. Machin, and S. H. P. Maddrell for their critical readings of the manuscript. INAA measurements were performed by R. Hancock. This research was supported by a grant from the Natural Sciences and Engineering Research Council (Canada) to J. Machin.