the paracellular pathway in the lepidopteran larval midgut: modulation by intracellular mediators

gy, Part A 144 (2006) 464–473www.elsevier.com/locate/cbpa

Comparative Biochemistry and Physiolo

The paracellular pathway in the lepidopteran larval midgut:Modulation by intracellular mediators

L. Fiandra ⁎, M. Casartelli, B. Giordana

Department of Biology, University of Milan, Via Celoria 26, 20133 Milano, Italy

Received 27 December 2005; received in revised form 23 March 2006; accepted 11 April 2006Available online 28 April 2006

Abstract

The features of the paracellular pathway, an important route for the transfer of ions and molecules in epithelia, are in insects still poorlyinvestigated and it has not yet been elucidated how the septate junction (SJ) acts as a transepithelial barrier. In this study, some properties of theparacellular pathway of Bombyx mori larval midgut, isolated in Ussing chambers, were determined and the modulation of SJ permeability byintracellular events disclosed. Diffusion potentials evoked by transepithelial gradients of different salts indicated that the junction bore weak negativecharges and that the paracellular pathway was selective with respect to ion charge and size. In standard conditions, the transepithelial resistance was28.2±2.1Ω cm2, a value indicating that the midgut is a low resistance epithelium. The modulation of midgut SJ by typical enhancers of mammaliantight junction permeability known to act on the cytoskeleton was studied by measuring the shunt resistance and the lumen-to-haemolymph flux ofsucrose. An increase of the intracellular level of cAMP and Ca2+ caused a significant decrease of the shunt resistance and an increase of SJpermeability. The attenuation of Ca2+ effect in the presence of the calcium channel blocker nifedipine indicated that the influx of external Ca2+ into thecytoplasm was important for the opening of the SJ, as well as the release of Ca2+ from the intracellular stores.© 2006 Elsevier Inc. All rights reserved.

Keywords: Bombyx mori; cAMP; Calcium; Midgut; Ion selectivity; Septate junction; Sucrose flux; Transepithelial resistance

1. Introduction

The septate junction (SJ) connecting the adjacent cells inseveral kinds of epithelia acts as a permeability restriction indifferent invertebrate phyla: Porifera, Coelenterata, Platyhel-minthes, Annelida, Mollusca, Arthropoda, Echinodermata andCephalochordata (Gerencser, 1982; Green and Bergquist, 1982;Welsch, 1983; Bleher andMachado, 2004; Hori, 2005). In insects,in particular, SJ has been extensively investigated both at amorphological (Noirot and Noirot-Timothée, 1998) and, morerecently, at amolecular level (Tepass et al., 2001; Behr et al., 2003).

However, junction permeability properties have been thor-oughly characterized only for vertebrate tight junction (TJ),while the functional features of the paracellular pathway ininsects are still poorly known. In fact, only few data are availableabout its role in the permeation of ions and small organic

⁎ Corresponding author. Dipartimento di Biologia, Università degli Studi diMilano, Via Celoria 26— 20133Milano, Italy. Tel.: +39 02 5031 4748; fax: +3902 5031 4802.

E-mail address: [email protected] (L. Fiandra).

1095-6433/$ - see front matter © 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.cbpa.2006.04.011

molecules. In the larval rectal epithelium of Aeshna cyanea,lanthanum ions (La3+) were unable to cross the entire length ofthe pleated SJ, suggesting a strictly occlusive role for thisjunction (Kukulies and Komnick, 1983). The septate junctionsof an epidermal cell line (UMBGE-4) derived from the cock-roach Blatella germanica were extensively permeated bylanthanum, although a partial resistance to the complete diffu-sion of the molecule across the cell layer was observed also inthis experimental model (Reis e Sousa et al., 1993). Conversely,a study on theMalpighian tubules of Rhodnius prolixus (Skaer etal., 1987) evidenced a fair permeability of the SJ to a largevariety of substances, such as La3+ (180 Da), sucrose (342 Da)and polyethylene glycol (PEG) (4000 Da). The permeation ofthese solutes was strictly related to the shape of the molecule,since PEG, formed by a straight chain of a considerably lowcross-sectional area, permeated the tubule epithelium consis-tently faster than the ring structured sucrose, despite its highermolecular weight. A significant permeation of large moleculeswas also observed in the gut of Schistocerca gregaria, wheresmooth SJs allowed the passage of a molecule as large as inulin(approximately 5000 Da) (Zhu et al., 2001). These data suggest

mailto:[email protected]

http://dx.doi.org/10.1016/j.cbpa.2006.04.011

465L. Fiandra et al. / Comparative Biochemistry and Physiology, Part A 144 (2006) 464–473

that the paracellular pathway in insects might be rather per-meable to molecules with a high molecular weight and implythat SJ could not perform the same occluding function of TJ.According to Lane and Flores (1988), SJs might have anadhesive role because of their strict association with actinfilaments, but ten years later Noirot and Noirot-Timothée (1998)again raised the debate on SJ function and claimed that anadhesion role, likely present, could be a by-product of a primaryocclusive function, and suggested that the number and disposalof septa could determine the higher or lower permeability of thejunction, which would thus be tight or leaky, respectively.However, the SJ appears to be a less efficient device than thetight junction, needing a considerable extension to produce agood tightness (Noirot and Noirot-Timothée, 1998).

The efficacy of orally administered insecticidal peptides tar-geting haemocoelic receptors is strictly related to their ability tocross the insect intestinal epithelium. The paracellular pathwaymay represent an interesting route for the delivery of biologicallyactive molecules of relatively small dimensions, and an accuratefunctional study of SJ permeability of the insect midgut is a firstnecessary step, that may later allow identification of suitablestrategies to increase the delivery of insecticides through thisroute. Indeed, a large number ofmolecules are able to increase theparacellular permeability in mammals, primarily those interactingwith the cytoskeletal elements connected to the junctional com-

Fig. 1. Transepithelial potentials generated in B. mori midgut by the addition of 50 mluminal solution. Typical experiments are shown.

plex (Madara et al., 1986; Balda et al., 1991; Tai et al., 1996; Pérezet al., 1997; Karczewski and Groot, 2000).

Aim of the study here presented was to disclose some pro-perties of the paracellular pathway inBombyx mori larval midgutand to investigate if SJ permeability could be modulated. Themidgut epithelium of lepidopteran larvae is composed by typicalabsorptive columnar cells and by highly specialized goblet cells.A vacuolar-type proton ATPase coupled to a K+/2H+ antiporteris expressed in the apical membranes lining goblet cells largecavity (Wieczorek et al., 1991, 1999, 2003). The electrogenicactivity of the V-ATPase generates a large transapical electricalpotential difference, cell interior negative, with values rangingbetween 140 and 240 mV (Dow and Peacock, 1989; Moffett andKoch, 1992). The cooperative activity of the V-ATPase and ofthe K+/2H+ antiporter is responsible for the passive re-entry ofthe protons into the goblet cell cytoplasm following the largeelectrochemical gradient, and for the active secretion of K+ intothe midgut lumen. A detailed analysis of the influence ofhaemolymph K+ concentration and of anoxia on the conduc-tance of the larval midgut epithelium was performed by Moffett(1980) inManduca sexta. K+ entry into the goblet cells from thehaemolymph is ensured by different specific channels present inthe basal membrane (Moffett and Lewis, 1990; Moffett andKoch, 1991; Zeiske, 1992). The V-ATPase is also responsiblefor the high alkaline pH present in the midgut lumen, reaching

M KCl (a), NaCl (b), (TMA)Cl (c), KGluconate (d) or NaGluconate (e) to the

Fig. 2. Transepithelial diffusion potential generated in B. mori midgut by theaddition of 50 mM KCl, NaCl or (TMA)Cl to the luminal solution. Data are themeans±SEM of 4 different experiments.

Fig. 3. Vt (○) and Rt (●) values with time in the presence of O2 or N2. Midgutswere incubated in the following buffers (mM): 20 KGluconate, 1 CaCl2, 4.8MgSO4, 240 sucrose, 5 Tris adjusted to pH 7 in the haemolymph compartmentor 5 CAPS adjusted to pH 10 in the luminal one. A typical experiment is shown.

466 L. Fiandra et al. / Comparative Biochemistry and Physiology, Part A 144 (2006) 464–473

values of 12 in the anterior-middle region (Dow, 1992; Azumaet al., 1995).

The electrogenic activity of the proton pump and the highluminal K+ concentration provide the electrochemical gradientfor K+ influx across the apical membrane of columnar cells thatdrives amino acid accumulation by means of specific K+/aminoacid symporters (Giordana et al., 1998; Castagna et al., 1998;Casartelli et al., 2001). Therefore, apical K+ channels in colum-nar cells are not expressed or are inactive in feeding larvae(Peyronnet et al., 2004), a channel activity becoming evidentonly in moulting or starved larvae, or in larvae exposed to par-ticular environmental conditions. Cl− permeability across thecolumnar cell apical membrane is exclusively due to a Cl−/HCO3

− exchanger (Chao et al., 1989). Finally, the very steep pHgradient across the intestinal epithelium strongly suggests a lackof proton channels in the apical membranes. Therefore, when theproton pump is not active, the major part of the ion movementsshould take place across the SJ.

We studied ionmovements in themidgut in vitro by recordingthe transepithelial potentials generated by salt gradients and thevariation of the transepithelial resistance induced by Ca2+ orcAMP. The modulation of the paracellular flux of sucrose, anextracellular marker not transported by midgut cells (Giordanaand Sacchi, 1977), was also investigated.

2. Materials and methods

2.1. Materials

[6,6′(n)-3H]Sucrose (16.8 Ci/mmol) was supplied by Amer-sham Bioscience, Italy. Cytochalasin D, dibutyryl-cAMP,nifedipine, thapsigargin and the other reagents were purchasedfrom Sigma-Aldrich, Italy.

2.2. Experimental animals

Larvae of Bombyx mori polyhybrids adapted to the artificialdiet were reared under controlled conditions (25±1 °C, 65–70%RH, 12L:12D photoperiod) on the artificial diet formulated byC.R.A.–I.S.Z.A.–Specialized Section for Sericulture, Padua, Italy.

2.3. Isolation of B. mori midgut in Ussing chambers

Larvae were anaesthetized with CO2 and sacrificed on thefourth day of the last instar. The anterior-middle region of themidgut was severed from the anaesthetized larvae, deprived of theperitrophic membrane and intestinal contents, laid on a piece ofthin cotton gauze and cut longitudinally. The cotton gauze, nec-essary tomaintain the tissue extended, presented very largemeshesand offered no restriction to the permeation of molecules. Thedissected midgut was thenmounted as a flat sheet between Ussingchambers (World Precision Instruments, Berlin, Germany) with anexposed surface of 19.6 mm2, bathed with buffer solutions (seefigure legends andSections 2.4 and 2.6) and then connected for themeasurement of the transepithelial voltage (Vt) and/or of theelectrical resistance. Solutionswere circulated by gas influx (100%O2) and maintained at 25 °C in water-jacketed reservoirs. Theircomposition differed according to the experimental condition.

2.4. Measurements of SJ ion selectivity

The midguts were incubated in the following solution (inmM): 1 CaCl2, 4.8 MgSO4, 300 sucrose, 5 Tris adjusted to pH 7in the haemolymph compartment or 5 CAPS adjusted to pH 10 inthe luminal compartment. After 15–20 min necessary for theequilibration of the tissue, the luminal buffer was diluted 1:1with the above solution added with 100 mM KCl, NaCl, (TMA)Cl, KGluconate or NaGluconate to generate a salt gradientdirected from the lumen to the haemolymph compartment.Osmolarity was kept constant by modifying sucrose concentra-tion. Since the transepithelial voltage of the midgut epithelium instandard conditions is lumen positive (see Figs. 3 and 4), for allthe recordings the positive pole was conventionally located inthe lumen side of the epithelium.

2.5. Electrophysiological measurements

The transepithelial voltage across the tissue was measuredcontinuously using a voltage-clamp device (DVC-1000, World

Fig. 4. Vt (○) and Rt (●) values with time in the presence or in the absence of K+.Midguts were incubated in the following buffers (mM): 1 CaCl2, 4.8 MgSO4,280 sucrose, 5 Tris adjusted to pH 7 in the haemolymph compartment or 5 CAPSadjusted to pH 10 in the luminal one, with or without 40 mM KGluconate. Atypical experiment is shown.


Precision Instruments, Berlin, Germany) connected to the bathsolutions via Ag–AgCl voltage electrodes in series with agarbridges (3 M KCl, 5.5% Agar). All Vt measurements were per-formed assuming the lumen side as the positive pole. A currentpulse generator was connected to the solutions by means of Ag–AgCl current electrodes in series with agar bridges (3 M KCl,5.5% Agar). Variations of Vt (ΔVt) induced by a 1 s current pulseof 459.2 μA/cm2 (I ) were recorded and the transepithelialelectrical resistance (Rt) was calculated by the Ohm law ΔVt /Iratio (Ω cm2). In order to operate in experimental conditions inwhich the main component of the tissue resistance is representedby the shunt resistance (Rsh) (Pannabecker et al., 1992), theelectrogenic V-ATPase was inactivated by bubbling the solutionswith N2 or by incubating the midguts in the absence of K+. Sincethe voltage electrodes are inserted in the luminal and haemolymph

Fig. 5. Effect of 80 μM cytochalasin D (CD) on Vt (○) and Rsh (●) values with time. Mto pH 7 in the haemolymph compartment or 5 CAPS adjusted to pH 10 in the luminalrecorded after 35 min in the absence (control) or in the presence of CD.

compartments of the Ussing chambers at a distance of 0.5 cmfrom the midgut, the resistance values recorded were corrected bysubtraction of the solution resistance (Rsol=ΔVsol / I ).

The effect on Rsh of different concentrations of Ca2+, of the

Ca2+-channel inhibitor nifedipine, of dibutyryl-cAMP, ofcytochalasin D (CD) or of thapsigargin, inhibitor of the intra-cellular Ca2+-ATPase, was recorded. Since CD and nifedipinewere dissolved in DMSO and thapsigargin in ethanol, controlexperiments performed with the same amounts of DMSO orethanol (less than 0.5%) showed that they did not alter theelectrophysiological properties of the tissue. The compositionof the incubation media varied according to the different ex-perimental conditions and is reported in the legend of the figure.Variations of Ca2+ concentration were performed maintainingconstant the Ca2+/Mg2+ ratio (=4.8). Osmolarity was alwayskept constant by modifying sucrose concentration.

2.6. Flux measurements

Flux values were determined by incubating the midgut in aphysiological solution with the following composition (in mM):20KGluconate, 240 trehalose, 1 sucrose, 5 Tris adjusted to pH 7 inthe hemolymph compartment or 5 CAPS adjusted to pH 10 in theluminal one. Sucrose influx (lumen to hemolymph, Jl–h) wasmeasured by adding 2 μCi/mL 3H-sucrose to the luminal solution.A 100 μL sample was withdrawn from the haemolymph side aftera 20-min equilibration period and a second onewas taken from thesame compartment after further 40 min. At the same times, sam-ples were withdrawn also from the luminal side to keep constantthe hydrostatic pressure and to measure the total radioactivitypresent in the luminal solution. Radioactivity was measured in aliquid scintillation counter (Tri-Carb Packard, 1600CA). Fluxesweremeasured in the presence of different concentrations of CaCl2or in the presence of 2 mM dibutyryl-cAMP, according to theexperimental condition. Variations of Ca2+ concentration were

edia composition was (mM): 1 CaCl2, 4.8 MgSO4, 280 sucrose, 5 Tris adjustedone. A typical experiment is shown. Insert: means±SEM of at least 4 Rsh values

Fig. 6. Effect of 2 mM db-cAMP on Vt (○) and Rsh (●) values with time. Media composition is reported in Fig. 5. A typical experiment is shown. Insert: means±SEMof at least 3 Rsh values in the absence or in the presence of db-cAMP. **Pb0.01 (t-Student test).


performed maintaining constant the Ca2+/Mg2+ ratio (=4.8).Osmolarity was kept constant by modifying the trehalose con-centration. During the entire experimental period, Vt was recordedin order to monitor tissue vitality.

3. Results

3.1. Septate junction selectivity to ions

The selectivity of the septate junctions to ions of differentcharge and size was studied bymeasuring the potentials generatedby salt gradients directed from the lumen to the haemolymphcompartment. To operate in the most simplified experimentalcondition, larval midguts were initially incubated in a K+-freesolution: the absence of K+ in the haemolymph side depresses V-ATPase activity, which is strictly linked to that of the 2H+/K+

antiporter, and the transepithelial potential difference approaches

Fig. 7. Effect of an increase of the external Ca2+ concentration on Vt (○) and Rsh (●) iis reported in Fig. 5. CaCl2 was added to both media to reach a final concentration

0 mV within 15–30 min. The addition of 50 mM KCl, NaCl,(TMA)Cl, KGluconate or NaGluconate to the luminal solutiongenerated a variation of Vt according to ions diffusion from theluminal to the haemolymph side (Fig. 1) through the paracellularpathway. Addition of KCl induced the rapid appearance of alumen-negativeVt (see Section 2.4 for the location of the referenceelectrode), which indicated a permeation of K+ across the junctionfaster than that of the anion. The lumen-negative polarization ofthe epithelium was immediately followed by a slow depolariza-tion, likely caused by Cl− flux (Fig. 1a). The NaCl diffusionpotential was similar to that induced by the KCl gradient, but theabsolute value of the Vt generated by Na+ permeation across thejunction was lower (Fig. 1b). Addition of (TMA)Cl to the luminalside induced a very small lumen negative polarization, immedi-ately followed by the appearance of a positive potential, in ag-reement with the slow diffusion of chloride ions from the luminalto the hemolymph side (Fig. 1c). The means of the diffusion

n the absence (a) or in the presence (b) of 200 μM nifedipine. Media compositionof 6 mM. A typical experiment is shown.

Fig. 8. Effect of Ca2+ concentration on Rsh. Means±SEM of at least 4recordings. ***Pb0.001 for 6 mM Ca2+ vs. 1 mM Ca2+; §§§Pb0.001 for 6 mMCa2+ plus nifedipine vs. 6 mM Ca2+ (t-test).

Fig. 9. Effect of 20 μM thapsigargin on Vt (○) and Rsh (●) values with time.Media composition is reported in Fig. 5. A typical experiment is shown.


potential (4 experiments) obtained with KCl, NaCl and (TMA)Cltransepithelial gradient confirmed a different permeabilitybetween Na+ and K+ (Pb0.05) through the junction, while, asexpected, TMA+ permeability was about two-fold lower than thatof the alkaline cations (Fig. 2). The transepithelial potentialsevoked by the addition of KGluconate and NaGluconate (Fig. 1dand e) differed from those observed with KCl and NaCl gradients:within 1 min a new steady-state was reached since Gluconate−,unlike Cl−, was unable to cross the epithelium. The equilibriumpotential reached by NaGluconate addition was about threefoldlower than that caused by KGluconate, emphasizing the differentpermeability of the epithelium to the two cations. Hence, lepi-dopteran midgut SJs are selective for monovalent cations with thefollowing rank order: K+NNa+NNTMA+. The high permeabilityof the paracellular pathway to cations suggests the presence ofnegative charges along the junction. Nonetheless, the passage of asmall anion like chloride indicates that within the SJ the strengthof the overall negative charges is relatively weak.

3.2. Transepithelial resistance measurements and shuntresistance (Rsh) evaluation

The cellular resistance (Rc) in B. mori larval midgut is domi-nated by the electromotive force (Ec) generated by the activity ofthe V-ATPase. The paracellular pathway between the adjacentcells is usually referred to as shunt pathway, and the resistance ofthis route to the passive permeation of ions and charged mole-cules is the shunt resistance (Rsh).

The transepithelial resistance (Rt) can be easily determinedby measuring the variation of Vt induced by constant tran-sepithelial current pulses. The Rt value measured in controlconditions was 28.2±2.1Ω cm2 (mean±SEM, 12 experiments)and indicates that B. morimidgut is a low resistance epithelium.

According to the theoretical model and the experimental datadiscussed by Pannabecker et al. (1992) for Aedes aegypti Mal-pighian tubules, in those epithelia that are characterized by ahighly active electrogenic pump, the inhibition of Ec causes alarge increase of Rc, and in this condition the Rt value approachesthat of Rsh. Therefore, an inhibition of the dominant active trans-

cellular transport, obtained by blocking V-ATPase activity, willallow an estimation of midgut Rsh. To this end, the proton pumpwas inhibited by substituting N2 for O2 in the circulatingsolutions. A typical experiment is shown in Fig. 3, reporting Vtand Rt values with time. As expected, N2 induced a voltagedepolarization toward 0 mV and a parallel increase of Rt. ThemaximalRt value achieved after 35min should almost correspondto that of Rsh (Pannabecker et al., 1992). The effect of anoxia wasreversible, since reintroduction of O2 in the incubation mediareadily restored a positive Vt, and the resistance values returned tothe initial ones. The estimated Rsh mean value from 14 ex-periments was 54.4±2.7 Ω cm2 (mean±SEM).

However, the stress condition to which the epithelium is ex-posed during the experiment, due to the inhibition of the oxidativemetabolism, could per se modify the SJ permeability. Therefore,we blocked the electrogenic activity of the V-ATPase by simplyremoving K+ from the incubation solutions. The lack of K+ at thebasolateral side had the same inhibitory effect of N2 on Ec, causinga decrease of Vt and an increment of Rt (Fig. 4). After 20 min, theresistance was steady and the value was 61.9±4.4Ω cm2 (mean±SEM, 14 experiments), not significantly different from thatobtained by incubating the midgut with N2. The addition ofpotassium to both the luminal and hemolymph compartmentsrestored the pump activity, followed again by the positive polar-ization of the tissue and the reduction of the epithelial resistance.

In all following experiments,Rsh values and its variations weremeasured by inhibiting the pump activity through a removal ofK+

rather than by perfusing the bathing solutions with N2.

3.3. Effect of cytochalasin D on Rsh

Rsh represents a valuable parameter to investigate the effect onthe paracellular pathway of cytochalasin D, an enhancer typicallyactive on the vertebrate tight junction. Fig. 5 reports Vt and Rsh

values registered inB.morimidgut before and after the addition of80 μM CD in both the luminal and haemolymph solutions. Ex-posure of the tissues to CD for approximately 35 min did notinduce any effect either on electrical resistance or Vt. The meanRsh values (4 experiments, values recorded after 35 min) in theabsence (control) or in the presence of CD is reported in the insert

Fig. 10. Effect of 2 mM db-cAMP and 6 mM Ca2+ on lumen-to-haemolymphsucrose flux. Means±SEM of at least 4 replicates. *Pb0.05, ***Pb0.001 (t-test), vs. control.


of the figure. The increment of CD concentration up to 300 μMand the prolongation of the incubation period up to 60 min con-firmed the complete lack of effect on the paracellular resistance(data not shown).

3.4. Effect of dibutyryl-cAMP and Ca2+ on Rsh

To increase the intracellular cAMP concentration in enter-ocytes, we exposed both sides of the midgut to 2 mM dibutyryl-cAMP (db-cAMP), a permeant and relatively stable cAMPanalogue. When db-cAMP was added to the media, Rsh readilydeclined by approximately 10 Ω cm2, while the transepithelialpotentialwas only slightlymodified (Fig. 6). As shown in the insertof Fig. 6, the mean Rsh values (3 experiments) in the presence ofdb-cAMP were significantly different from those of controls.

As extensively reported in the literature, an increase ofcytosolic Ca2+ concentration affects tight junction permeability,inducing again a decrease of the tissue electrical resistance. Thetypical experiment reported in Fig. 7a shows that the increasefrom 1 to 6 mM of Ca2+ concentration in the incubation mediumcaused a decrease of the paracellular resistance. The mean of Rsh

decrease (4 experiments) was 50% of control value (Fig. 8). Theaddition of the Ca2+ channel inhibitor nifedipine (200 μM) tothe incubation media significantly attenuated the effect on Rsh ofextracellular Ca2+ (Figs. 7b and 8), an indication that this cationpartly enters the cells via nifedipine-sensitive channels.

The role of intracellular calcium concentration on lepidopteranSJ modulation was further confirmed by incubating the midgutswith 20 μM thapsigargin, a specific inhibitor of Ca2+ re-uptakeinto the intracellular stores. The gradual increase of cytosolic Ca2+

level induced by thapsigargin caused a slow decrease (Fig. 9) ofmidgut Rsh of 16.8±0.70 Ω cm2 (mean±SEM, 3 experiments)after 60 min of incubation.

3.5. Effect of dibutyryl-cAMP and Ca2+ on the lumen tohaemolymph flux of sucrose

Mannitol, the classical marker of the paracellular route inmammalian intestine, cannot be used to trace the same pathway

in B. mori midgut, because it is transported, like othermonosaccharides, at the apical membrane of columnar cells(Casartelli M., personal communication). Conversely, sucrosedoes not enter B. morimidgut cells (Giordana and Sacchi, 1977)and crosses the midgut epithelium in M. sexta (Koch andMoffett, 1987). It can therefore be considered a suitable markerto study the paracellular permeability.

Sucrose lumen-to-haemolymph fluxes were measured inmidguts incubated with a medium containing 1 mM Ca2+: theaddition of db-cAMP significantly increased the Jl–h values andan even larger effect was observed with the increase of externalCa2+ concentration (Fig. 10).

4. Discussion

Although insect septate junctions are morphologically (Noirotand Noirot-Timothée, 1998) and molecularly (Tepass et al., 2001;Behr et al., 2003) well-known, their permeability properties havebeen poorly considered and their selectivity features are stillcontroversial (Kukulies and Komnick, 1983; Skaer et al., 1987;Noirot and Noirot-Timothée, 1998; Zhu et al., 2001). Septatejunctions of Aedes aegyptiMalpighian tubules, which represent apassive transport pathway for the secretion of Cl− (Yu andBeyenbach, 2001), seem to prevent the passage of cations, actingas a route with a positive electric field of low strength (Williamsand Beyenbach, 1984). Taking into account that the apical mem-branes of the lepidopteran midgut epithelium are almost imper-meable to K+ (see Introduction) and that Cl− conductance acrosscell membranes is almost negligible (Chao et al., 1989), thediffusion potentials generated across B. mori larval midgut by saltgradients (Figs. 1 and 2) indicate that the physiological mono-valent cations can rapidly diffuse through the SJ, suggesting thepresence of negative charges along the junction. However, themoderate permeability to the small anion chloride agrees with anegative electric field of relatively weak strength. In dog proximalrenal tubule and in rabbit ileum, the paracellular pathway shows aslight cation selectivity, which suggests that the tight junction ispredominantly lined by negative charges (Fanning et al., 1999).Therefore, the charge selectivity of lepidopteran midgut SJ iscomparable to that of mammalian TJ. Midgut SJ appears alsohighly selective with respect to the size of the permeant ion.Although the permeability of K+ compared to that of Na+, whenpermeable Cl− is the counterion, is only slightly different (Figs.1a, b and 2), the equilibrium potentials generated by the twocations when the counterion is the impermeant Gluconate− (Fig.1d and e) clearly evidences their distinct diffusion rates. Indeed,K+ crystal radius (1.33 Å) is greater than that of Na+ (0.95 Å), butthe hydrated radius of the two ions is 3.31 Å and 3.58 Årespectively (Nightingale, 1959). The permeability to a cationwith a larger molecular weight like TMA+ is considerably lowerthan that observed for K+ and Na+ (Figs. 1c and 2). Hence, cationdiffusion across the junction is inversely related to theirdimensions: K+NNa+NNTMA+. That SJ discriminates ionsaccording to their size is further supported by the low permeabilityof the paracellular route to the large anion Gluconate−.

An estimation of ion mobility through the epithelial barrierand the junction can be obtained from electrophysiological


measurements. The equivalent model of electrolyte transport inB.mori larval midgut, fairly different from that of Aedes aegyptianterior stomach (Onken et al., 2006), appears quite similar to thatproposed for theMalpighian tubules of mosquito (Pannabecker etal., 1992; Beyenbach et al., 2000), similarly dominated by anapical highly active electrogenic pump. The Rt measured incontrol conditions indicates that the midgut epithelium oflepidopteran larvae has a low electrical resistance, due to theextremely high electromotive force represented by the massiveproton transport mediated by the V-ATPase. According toPannabecker et al. (1992), in this type of epithelium, the electricalresistance of the paracellular pathway can be estimated byabolishing the electromotive force (Figs. 3 and 4). In fact, ionconductance across lepidopteran midgut cell membranes isalmost negligible when the activity of the proton pump is blocked(Moffett, 1980; Chao et al., 1989; Peyronnet et al., 2004). Even ifwe cannot exclude that a small cellular component could still bepresent after V-ATPase inhibition, it is reasonable to assume thatthe maximal value of the resistance reached once that Ec is totallyabolished has to be ascribed mainly to Rsh.

Many studies have characterized in mammals TJ permeabil-ity and its modulation, and Rsh was measured as a determinantparameter to investigate the effect of a number of enhancers(Madara et al., 1986; Pérez et al., 1997; Karczewski and Groot,2000). In insects, an electrophysiological study performed in1993 with isolated perfused Malpighian tubules evidenced forthe first time that SJ can be hormonally modulated (Pannabeckeret al., 1993). Further recent investigations (Yu and Beyenbach,2002) clarified the mode of action of leucokinin-VIII in AedesaegyptiMalpighian tubules, showing that the signal transductionpathway activated by the neuropeptide involved changes inintracellular Ca2+ concentration. Ca2+ was in part released fromthe intracellular stores, but the influx of extracellular Ca2+ intothe cytoplasm was also a necessary cofactor for the modulationof SJ. In B. mori larval midgut, Ca2+ and cAMP, both typicalenhancers of the permeability of mammalian TJ (Pérez et al.,1997; Karczewski and Groot, 2000), are able to modify the shuntpermeability. Either the increment of extracellular Ca2+

concentration or the exposure of the tissue to cAMP induces areduction of Rsh (Figs. 6–8) and an increase of sucrosepermeability (Fig. 10). The partial inhibition of calcium effectin the presence of the Ca2+-channel inhibitor nifedipine (Fig. 7b)confirms that the influx of Ca2+, increasing the cationconcentration in the cytosol, is important for the opening ofthe paracellular pathway. Besides, release of intracellularcalcium is also relevant, as indicated by the reduction of Rsh inthe presence of thapsigargin, which increases cytosolic Ca2+ byinhibiting calcium sequestration within the stores (Fig. 9).

In mammals, TJ regulation by intracellular mediators isexerted through a modification of the interactions between thecytoskeleton components, mediated by the PKA and PKCsystems (Balda et al., 1991; Tai et al., 1996; Pérez et al., 1997;Karczewski and Groot, 2000). It is now established that theeffect is due to a phosphorylation of myosin light chain(Yamaguchi et al., 1991; Turner and Madara, 1995; Hecht etal., 1996; Turner et al., 1997), that leads to the contraction of theactin–myosin bundles interacting with the peripheral ZO-

proteins, connected to the transmembrane tight junction proteins.With this model as a reference frame, it is reasonable to speculatethat in B. mori midgut the increase of SJ permeability observedin response to cAMP and Ca2+ could depend on thephosphorylation of the cytoskeleton filaments connected to thejunctional proteins. In the insect midgut, actin microfilamentsdeparting from the microvillar rootlets branch and connect withone another, forming a scaffold which descends deep into thecytoplasm, making contacts with the lateral membrane at thejunctional level (Dallai et al., 1998). Actin filaments seem tointeract directly with the membrane (Dallai et al., 1998), butinterposition of proteins like α-actinin and vinculin (Colombo etal., 1993) or spectrin (Bonfanti et al., 1992), all shown to bepresent in lepidopteran enterocytes, should also be considered.The identification of myosin II in association with the SJ ofManduca sexta larval midgut (Bonfanti et al., 1992) supports thehypothesis of a kinase-mediated effect on SJ permeabilityleading to a controlled contraction of the cytoskeleton scaffold.

The exposure of mammalian intestinal epithelium to cytocha-lasin D typically causes a decrease of the transepithelial electricalresistance and an increase of the paracellular permeability tomannitol (Madara et al., 1986) as a result of the depolymerizationof the actin filaments of the perijunctional ring (Lamaze et al.,1997). In B. mori larval midgut exposed to CD nomodification ofRsh was recorded (Fig. 5). Besides, we do not have a directevidence that the drug have induced a depolymerization of theactinic cytoskeleton, although CD did cause a disruption of themicrofilaments (Lane and Flores, 1990) and a consequentalterations of SJ intramembranous protein distribution (Laneand Flores, 1988) at least in cockroach accessory glands. So, wemust take into account the possibility that in lepidopteran midgutcytochalasin D might have been unable to reach the intracellulartargets in a sufficient amount. A proof of the effective ability ofCD to depolymerize actin filaments in midgut cells will benecessary to elucidate this crucial aspect.

In conclusion, the findings here presented describe someproperties of the SJ of an insect larval midgut, proving that,despite its low transepithelial electrical resistance, the lepidop-teran midgut still performs the typical barrier function byexerting a high selectivity with respect to ion charge and size.Moreover, it is here reported for the first time that also in insectenterocytes intracellular mechanisms can modulate the perme-ability of the paracellular pathway.

Acknowledgements

We are indebted to Prof. GiulianoMeyer for helpful commentsand suggestions on the manuscript. This work was supported bythe Italian Ministry of Education University and Research (FIRB,project no. RBNE01YXA8-001; COFIN 2004, project no.2004077251).

References

Azuma, M., Harvey, W.R., Wieczorek, H., 1995. Stoichiomentry of K+/H+

antiport helps to explain extracellular pH 11 in a model epithelium. FEBSLett. 361, 153–156.


Balda, M.S., Gonzalez-Mariscal, L., Contreras, R.G., Macias-Silva, M.E.,Torres-Marquez, J.A., Garcia-Sainz, J.A., Cereijido, M., 1991. Assemblyand sealing of tight junctions: possible participation of G-proteins,phospholipase C, protein kinase C and calmodulin. J. Membr. Biol. 122,193–202.

Behr, M., Riedel, D., Schuh, R., 2003. The claudin-like Megatrachea is essentialin septate junctions for the epithelial barrier function in Drosophila. Dev.Cell 5, 611–620.

Beyenbach, K.W., Pannabecker, T.L., Nagel, W., 2000. Central role of the apicalmembrane H+-ATPase in electrogenesis and epithelial transport inmalpighian tubules. J. Exp. Biol. 203, 1459–1468.

Bleher, R., Machado, J., 2004. Paracellular pathway in the shell epithelium ofAnodonta cygnea. J. Exp. Zool. 301, 419–427.

Bonfanti, P., Colombo, A., Heintzelman, M.B., Mooseker, M.S., Camatini, M.,1992. The molecular architecture of an insect midgut brush bordercytoskeleton. Eur. J. Cell Biol. 57, 298–307.

Casartelli, M., Fiandra, L., Parenti, P., Giordana, B., 2001. Multiple transportpathways for dibasic amino acids in the larval midgut of the silkwormBombyx mori. Insect Biochem. Mol. Biol. 31, 621–632.

Castagna, M., Shayakul, C., Trotti, D., Sacchi, V.F., Harvey, W.R., 1998.Cloning and characterization of a potassium-coupled amino acid transporter.Proc. Natl. Acad. Sci. U. S. A. 95, 5395–5400.

Chao, A.C., Koch, A.R., Moffett, D.F., 1989. Active chloride transport in theisolated posterior midgut of the tobacco hornworm (Manduca sexta). Am. J.Physiol. 258, R112–R119.

Colombo, A., Bonfanti, P., Camatini, M., 1993. Actin, α-actin, and vinculin areassociated with septate junctions in Insecta. Cell Motil. Cytoskelet. 26,205–213.

Dallai, R., Lupetti, P., Lane, N.J., 1998. The organization of actin in the apicalregion of insect midgut cells after deeep etching. J. Struct. Biol. 122,283–292.

Dow, J.A.T., 1992. pH gradients in lepidopteran midgut. J. Exp. Biol. 172,355–375.

Dow, J.A.T., Peacock, J.M., 1989. Microelectrode evidence for the electricalisolation of goblet cell cavities in Manduca sexta middle midgut. J. Exp.Biol. 143, 101–114.

Fanning, A.S., Mitic, L.L., Anderson, J.M., 1999. Transmembrane proteins inthe tight junction barrier. J. Am. Soc. Nephrol. 10, 1337–1345.

Gerencser, G.A., 1982. Paracellular transport characteristic of Aplysia julianaintestine. Comp. Biochem. Physiol. 72A, 721–725.

Giordana, B., Sacchi, V.F., 1977. Extracellular space values and intracellularionic concentrations in the isolated midgut of Philosamia cynthia andBombyx mori. Experientia 33, 1065–1066.

Giordana, B., Leonardi, M.G., Casartelli, M., Consonni, P., Parenti, P., 1998.K+-amino acid symport of Bombyx mori larval midgut: a system operative inextreme conditions. Am. J. Physiol. 274, R1361–R1371.

Green, C.R., Bergquist, P.R., 1982. Phylogenetic relationships within theinvertebrata in relation to the structure of septate junctions and thedevelopment of ‘occluding’ junctional types. J. Cell Sci. 53, 279–305.

Hecht, G., Pestic, L., Nikcevic, G., Koutsouris, A., Tripuraneni, J., Lorimer, D.D.,Nowak, G., Guerriero Jr., V., Elson, E.L., Lanerolle, P.D., 1996. Expression ofcatalytic domain of myosin light chain kinase increases paracellularpermeability. Am. J. Physiol. 271, C1678–C1684.

Hori, I., 2005. Formation of the septate junction in regenerating planariangastrodermis. J. Morphol. 192, 205–215.

Karczewski, J., Groot, J., 2000. Molecular physiology and pathophysiology oftight junctions III. Tight junction regulation by intracellular messengers:differences in response within and between epithelia. Am. J. Physiol. 279,G660–G665.

Koch, A., Moffett, D.F., 1987. Kinetics of extracellular solute movement in theisolated midgut of tobacco hornworm (Manduca sexta). J. Exp. Biol. 133,199–214.

Kukulies, J., Komnick, H., 1983. Plasma membranes, cell junctions and cuticleof the rectal chloride epithelia of the larval dragonfly Aeshna cyanea. J. CellSci. 59, 159–182.

Lamaze, C., Fujimoto, L.M., Yin, H.L., Schmid, S.L., 1997. The actincytoskeleton is required for receptor-mediated endocytosis in mammaliancells. J. Biol. Chem. 272, 20332–20335.

Lane, N.J., Flores, V., 1988. Actin filaments are associated with the septatejunctions of invertebrates. Tissue Cell 20, 211–217.

Lane, N.J., Flores, V., 1990. The role of cytoskeletal components in the main-tenance of intercellular junctions in an insect. Cell Tissue Res. 262, 373–385.

Madara, J.L., Barenberg, D., Carlson, S., 1986. Effects of cytochalasin D onoccluding junctions of intestinal absorptive cells: further evidence that thecytoskeleton may influence paracellular permeability and junctional chargeselectivity. J. Cell Biol. 102, 2125–2136.

Moffett, D.F., 1980. Voltage–current relation and K+ transport in tobaccohornworm (Manduca sexta) midgut. J. Membr. Biol. 54, 213–219.

Moffett, D.F., Koch, A., 1991. Lidocaine and barium distinguish separate routesof transbasal K+ uptake in the posterior midgut of the tobacco hornworm(Manduca sexta). J. Exp. Biol. 157, 243–256.

Moffett, D.F., Koch, A., 1992. Driving forces and pathways for H+ and K+

transport in insect midgut goblet cells. J. Exp. Biol. 172, 403–415.Moffett, D.F., Lewis, S.A., 1990. Cation channels of insect midgut goblet cells:

conductance diversity and Ba2+ activation. Biophys. J. 57, 85–97.Nightingale, E.R., 1959. Phenomenological theory of ion solvation. Effective

radii of hydrated ions. J. Phys. Chem. 63, 1381–1387.Noirot, C., Noirot-Timothée, C., 1998. Cell associations. Microsc. Anat.

Invertebr. 11A, 27–49.Onken, H., Moffett, S.B., Moffett, D.F., 2006. The isolated anterior stomach of

larval mosquitoes (Aedes aegypti): voltage-clamp measurements with atubular epithelium. Comp. Biochem. Physiol. 143A, 24–34.

Pannabecker, T.L., Aneshansley, D.J., Beyenbach, K.W., 1992. Uniqueelectrophysiological effects of dinitrophenol in Malpighian tubules. Am. J.Physiol. 263, R609–R614.

Pannabecker, T.L., Hayes, T.K., Beyenbach, K.W., 1993. Regulation of epithelialshunt conductance by the peptide leucokinin. J. Membr. Biol. 132, 63–76.

Pérez, M., Barber, A., Ponz, F., 1997. Modulation of intestinal paracellularpermeability by intracellular mediators and cytoskeleton. Can. J. Physiol.Pharm. 75, 287–292.

Peyronnet, O., Noulin, J.F., Laprade, R., Schwartz, J.L., 2004. Patch-clampstudy of the apical membrane of the midgut ofManduca sexta larvae: directdemonstration of endogenous channels and effect of a Bacillus thuringiensistoxin. J. Insect Physiol. 50, 791–803.

Reis e Sousa, C., Howard, J.E., Hartley, R., Earley, F.G.P., Djamgoz, M.B.A.,1993. An insect epidermal cell line (UMBGE-4): structural and electro-physiological characterization. Comp. Biochem. Physiol. 106A, 759–767.

Skaer, H., Le, B., Maddrell, S.H.P., Harrison, J.B., 1987. The permeabilityproperties of septate junction in Malpighian tubules of Rhodnius. J. Cell Sci.88, 251–267.

Tai, Y.-H., Flick, J., Levine, S.A., Madara, J.L., Sharp, G.W.G., Donowitz, M.,1996. Regulation of tight junction resistance in T84 monolayers by elevationin intracellular Ca2+: a protein kinase C effect. J. Membr. Biol. 149, 71–79.

Tepass, U., Tanentzapf, G., Ward, R., Fehon, R., 2001. Epithelial cell polarityand cell junctions in Drosophila. Annu. Rev. Genet. 35, 747–784.

Turner, J.R., Madara, J.L., 1995. Physiological regulation of intestinal epithelialtight junction as a consequence of Na+-coupled nutrient transport.Gastroenterology 109, 1391–1396.

Turner, J.R., Rill, B.K., Carlson, S.L., Carnes, D., Kerner, R., Mrsny, R.J.,Madara, J.L., 1997. Physiological regulation of epithelial tight junctions isassociated with myosin light-chain phosphorylation. Am. J. Physiol. 273,C1378–C1385.

Welsch, U., 1983. Septate junctions with paired septa in the cephalochordateBranchiostoma lanceolatum. Cell Tissue Res. 229, 175–181.

Wieczorek, H., Putzenlechner, M., Zeiske, W., Klein, U., 1991. A vacuolar-typeproton pump energizes H+/K+-antiport in an animal plasma membrane.J. Biol. Chem. 266, 15340–15347.

Wieczorek, H., Brown, D., Grinstein, S., Ehrenfeld, J., Harvey, W.R., 1999.Animal plasma membrane energization by proton-motive V-ATPase.Bioessays 21, 637–648.

Wieczorek, H., Huss, M., Merzendorfer, H., Reineke, S., Vitavska, O., Zeiske,W., 2003. The insect plasma membrane H+ V-ATPase: intra-, inter-, andsupramolecular aspects. J. Bioenerg. Biomembranes 35, 359–366.

Williams, J.C., Beyenbach, K.W., 1984. Differential effects of secretagogues onthe electrophysiology of the Malpighian tubules of the yellow fevermosquito. J. Comp. Physiol., B 154, 301–309.


Yamaguchi, Y., Dalle-Molle, E., Hardison, W.G., 1991. Vasopressin and A-23187 stimulate phosphorylation of myosin light chain-1 in isolated rathepatocytes. Am. J. Physiol. 261, G312–G319.

Yu, M.J., Beyenbach, K.W., 2001. Leucokinin and the modulation of the shuntpathway in Malpighian tubules. J. Insect Physiol. 47, 263–276.

Yu, M.J., Beyenbach, K.W., 2002. Leucokinin activates Ca2+-dependent signalpathway in principal cells of Aedes aegypti Malpighian tubules. Am. J.Physiol. 283, F499–F508.

Zeiske, W., 1992. Insect ion homeostasis. J. Exp. Biol. 172, 323–334.Zhu, W., Vandingenem, A., Huybrechts, R., Vercammen, T., Baggerman, G., De

Loof, A., Poulos, C.P., Valentza, A., Breuer, M., 2001. Proteoliticbreakdown of the Neb-trypsin modulating oostatic factor (Neb-TMOF) inthe haemolymph of different insects and its gut epithelial transport. J. InsectPhysiol. 47, 1235–1242.

the paracellular pathway in the lepidopteran larval midgut: modulation by intracellular mediators

Documents