Changes in Permeability of Brush Border Membrane Vesicles from Spodoptera littoralis Midgut Induced by Insecticidal Crystal Proteins from Bacillus thuringiensis

Abstract

Bacillus thuringiensis insecticidal crystal proteins (ICPs) are thought to induce pore formation in midgut cell membranes of susceptible insects. Cry1Ca, which is significantly active in Spodoptera littoralis, made brush border membrane vesicles permeable to KCl (osmotic swelling was monitored by the light scattering technique); the marginally active ICPs Cry1Aa, Cry1Ab, and Cry1Ac did not.

Different strains of Bacillus thuringiensis produce insecticidal crystal proteins (ICPs) that are highly and specifically toxic for some insect species. The mode of action can be summarized as follows. Susceptible larvae ingest ICPs, which are converted to active toxins by solubilization and proteolytic processing in the insect midgut. The “activated” ICPs bind to specific receptors in the apical membrane of the midgut cells, and they produce pores, which leads to colloid osmotic lysis of the cells and insect death (4). Brush border membrane vesicles (BBMVs) from the midguts of target insects have been used in an assay to monitor osmotic swelling (1, 9) based on changes in scattered-light intensity (SLI). A good relationship has been found between ICP toxicity and the ability to make BBMVs leaky. In a hyperosmotic medium, the vesicles initially shrink. Subsequently, if the substances in the medium can cross the membrane, water follows and the vesicles swell. Swelling decreases the SLI.

The present work was carried out with the Egyptian cotton leafworm (Spodoptera littoralis Boisduval), an important agricultural pest. The relationship between vesicle leakiness and ICP toxicity in vivo was studied with three marginally active ICPs (Cry1Aa, Cry1Ab, and Cry1Ac) and a much more active ICP (Cry1Ca). The 50% lethal concentration of Cry1Ca is at least 10 times lower than that of Cry1A proteins (2a, 8).

The data reported below are means ± standard errors of the means, and n indicates the number of determinations. Values were compared by using a one-way analysis of variance test with Dunnett posttests. Differences were considered statistically significant if P < 0.05. Calculations were performed with the Prism computer program (GraphPad Software, Inc.).

Permeability characteristics of BBMVs from S. littoralis.

The assay used was adapted from the technique developed by Carroll and Ellar (1). BBMVs prepared (10) from guts of last-instar S. littoralis larvae were suspended in BBMV buffer (17 mM Tris-HCl, pH 7.5) at a concentration of 0.6 mg of protein/ml and incubated overnight at 4°C. A 4-ml quartz cuvette containing a 3-ml sample of BBMVs was placed in the sample compartment of a luminescence spectrophotometer (model LS-5B; Perkin-Elmer Co.) at room temperature. The intensity of 450-nm light 90° from a 450-nm incident beam was recorded. Stock solutions (2.5 M) of sucrose, KCl, potassium gluconate, and NaCl were prepared in BBMV buffer. Hyperosmotic conditions were generated by adding 0.1 ml of a concentrated stock solution. BBMVs from S. littoralis were incubated with different solutes (final concentration, 80 mM) in order to determine which solute produced the greatest initial shrinking (i.e., the greatest increase in the SLI).

The maximum increase in SLI was obtained with KCl (39% ± 10%; n = 27). The effects of potassium gluconate and NaCl were less pronounced than the effect of KCl (12% ± 8% [n = 3] and 14% ± 9% [n = 3], respectively). BBMVs kept for more than 24 h at 4°C did not continue to respond to a hyperosmotic challenge, suggesting that degradation occurred. Surprisingly, incubation with sucrose in the medium (n = 4) did not change the SLI significantly. This effect may have been related to changes in optical density and artifacts associated with vesicle motion and aggregation. Examples of changes that can be induced by sucrose include changes in light emission due to variation in the refractive index and volume-independent scattering changes (3).

In a control experiment the addition of 0.1 ml of BBMV buffer alone caused a small decrease in the SLI (4% ± 1%; n = 5; P < 0.05).

The SLI of a BBMV sample was stable for at least 0.5 h (n = 3), but in the presence of hyperosmotic challenge with KCl, the SLI slowly decreased with time. This probably reflects the leakage of ions entering the vesicles through channels; when the slope of the linear regression line obtained in the presence of 6 mM BaCl₂ (Ba²⁺ is a K⁺ channel blocker) was examined, it was found that Ba²⁺ virtually completely inhibited this decrease (Fig. 1A).

FIG. 1.

Time course of SLI. S. littoralis midgut BBMVs were shrunken by exposure to a hyperosmotic KCl solution (data not shown). A subsequent decrease (from time zero to 25 min) in SLI was observed due to KCl (and water) that leaked into the vesicles. The values shown are means ± standard errors of the means (n = 3). (A) Swelling of the vesicles monitored in the absence (⧫) or in the presence of Ba²⁺ (○) or nystatin (▿). Incubation of BBMVs with nystatin in the absence of a hyperosmotic KCl solution was used as a control (▵). (B) Swelling accelerated in the presence of Cry1Ca (0.9 nmol/mg of BBMV protein) (▿), but not in the presence of Cry1Aa (▵), Cry1Ab (○), or Cry1Ac (◊). The Cry1Ab and Cry1Ac curves coincide with the relative SLI in the presence of KCl alone (see panel A [⧫]).

A 25-mg/ml (27 mM) stock solution of a pore-forming agent, nystatin (Sigma), was prepared in dimethyl sulfoxide. BBMVs were allowed to stabilize for 2 min in a hypertonic KCl solution prior to addition of nystatin (final concentration, 0.17 μmol/mg of BBMV protein). An abrupt 18% ± 0.1% decrease in the SLI was observed when nystatin was added (n = 3) (Fig. 1A). Dimethyl sulfoxide alone (n = 2) (data not shown) had no effect (there was no significant difference in the parameters of linear regression). When nystatin was used without prior KCl treatment (n = 3) (Fig. 1A), it had no effect.

In summary, a hypertonic KCl solution produced an increase in the SLI caused by BBMVs from S. littoralis, as observed previously in Manduca sexta (1) and Bombyx mori (9); a sudden increase occurred just after salt addition, followed by slow decrease due to KCl (and water) that slowly leaked in. The results obtained with KCl in combination with Ba²⁺ (which prevented K⁺ from leaking in) or nystatin (which allowed faster swelling) corroborate the suitability of using BBMVs in permeability assays. The occurrence of channels in BBMVs could be due to contamination with other membranes. Although the presence of K⁺ channels in the apical membrane of an epithelium actively secreting K⁺ is not expected, the possibility cannot be eliminated a priori. Indeed, it has been suggested that such channels are present in Spodoptera frugiperda (5).

ICP effects.

Activated ICPs (Cry1Aa, Cry1Ab, Cry1Ac, and Cry1Ca) were prepared as described previously (8), diluted in 20 mM Tris-HCl–150 mM NaCl (pH 8.6) (ICP buffer), and added 2 min after exposure of the BBMVs to a hypertonic KCl solution. Figure 1B shows the relative SLI of samples in the presence of the ICPs. Adding ICP buffer alone (n = 3) had no effect (data not shown). The results revealed a direct relationship with toxicity. Thus, the marginally active ICPs Cry1Aa (n = 4), Cry1Ab (n = 3), and Cry1Ac (n = 3) at a concentration of 35 μg/ml (0.9 nmol of ICP/mg of BBMV protein) did not seem to accelerate swelling. In contrast addition of Cry1Aa resulted in an increase in the SLI. Only Cry1Ca (n = 3) significantly accelerated the decrease in SLI compared with exposure to hypertonic KCl alone, suggesting that there was an increase in K⁺ permeability. The effects were slower than those of nystatin; at least 5 min of incubation with Cry1Ca was necessary before a significant difference could be detected. A 10-fold-lower concentration of Cry1Ca (0.09 nmol of ICP/mg of BBMV protein; n = 3) had no effect compared to the control (data not shown). The results obtained suggest that at the concentrations tested, Cry1Ca, but not Cry1Aa, Cry1Ab, or Cry1Ac, made S. littoralis BBMVs permeable to KCl. We cannot eliminate the possibility, however, that (much) higher doses of Cry1A may result in pore formation. These findings support the hypothesis that toxic ICPs increase membrane permeability (4). The protection from KCl entry by Cry1Aa might be due to some interference of this protein with K⁺ or Cl⁻ channels. The slower effects of Cry1Ca compared to the effects of nystatin could indicate that the mode of action of this ICP is more complex. Fast effects of Cry1Ac (1) and Cry1Ca (6) on M. sexta have been reported, but there may be differences between ICP-insect interactions.

Cry1A ICPs are only marginally active against Spodoptera species. Nevertheless, at least some Cry1A toxins bind to BBMVs from Spodoptera spp. with high affinity. For example, Cry1Ac is virtually inactive against S. frugiperda, while it binds with high affinity to S. frugiperda BBMVs (2); Cry1Ab is only marginally active against S. littoralis, while it binds to S. littoralis BBMVs with high affinity (7a). For Cry1Aa, which also is only marginally active against S. littoralis, contradictory binding data for this species have been reported. Whereas a virtually complete absence of binding to BBMVs has been reported (8), binding to different BBMV proteins has been observed in a ligand blot (7). Perhaps the BBMV proteins are not accessible in the native membranes. We are not aware of experiments in which binding of Cry1Ac to native S. littoralis BBMVs has been examined.

The presence of receptors (at least receptors for Cry1Ab) and the results of our swelling experiments suggest that the very low levels of activity of Cry1A ICPs in S. littoralis are probably due to an inability to carry out the pore formation step, even if some ICP binding can take place. The marginal toxicity of Cry1A ICPs for other Spodoptera species may be explained in the same way.