Molecular and Biochemical Parasitology 105 (2000) 39 – 49 www.elsevier.com/locate/parasitology
Transport of model peptides across Ascaris suum cuticle Barbara A. Sheehy, Norman F.H. Ho, Philip S. Burton, Jeffrey S. Day, Timothy G. Geary, David P. Thompson * Pharmacia and Upjohn, Inc., Mailstop 7923 -25 -13, Kalamazoo, MI 49001, USA Received 21 May 1999; received in revised form 9 August 1999; accepted 10 August 1999
Abstract Several FMRFamide-related peptides (FaRPs) found in nematodes exert potent excitatory or inhibitory effects on the somatic musculature of Ascaris suum and other nematode species when injected into the pseudocoelom or applied directly to isolated neuromuscular preparations. These peptides, however, generally fail to induce detectable effects on the neuromusculature when applied externally to intact nematodes. The apparent lack of activity for these peptides when administered externally in whole-organism assays is likely a function of both absorption and metabolism. To delineate the factors that govern transport of peptides across the cuticle/hypodermis complex of nematodes, we measured the rates of absorption of a series of structurally related model peptides using isolated cuticle/hypodermis segments from A. suum and two-chamber diffusion cells. [14C]-Labeled peptides were prepared from D-phenylalanine, with the amide nitrogens sequentially methylated to give AcfNH2, Acf3NH2, Acf(NMef)2NH2 and Ac(NMef)3NHMe. These model peptides were designed to allow systematic analysis of the influence of peptide size, hydrogen bonding and lipophilicity on transport. Results of these studies show that, within this series, permeability across the cuticle increases with addition of each methyl group. The permeability coefficient of Ac(NMef)3NHMe, with four methyl groups, was 10-fold greater than that of the smaller peptide, AcfNH2, even though both peptides contain five hydrogen bonds. When compared with vertebrate membranes, transport of the model peptides across A. suum cuticle was about 10-fold slower. A biophysical model for transcuticular transport of peptides predicted that nematode FaRPs, which are larger, less methylated and less lipophilic than the model peptides tested, would not be absorbed across the cuticle of nematodes. This prediction was confirmed for the excitatory FaRP, AF2 (KHEYLRFamide), which did not diffuse across the cuticle/hypodermis complex, but diffused rapidly across lipid-extracted cuticle preparations. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Ascaris suum; Nematode; Peptides; FMRFamide-related peptides; Transport; Cuticle; FMRFamide-related peptides
1. Introduction * Corresponding author. Tel.: +1-616-833-1939; fax: + 1616-833-8350. E-mail address:
[email protected] (D.P. Thompson)
Several FMRFamide-related peptides (FaRPs), including KHEYLRFamide (AF2) and SDPNFLRFamide (PF1), found in nematodes exert potent excitatory or inhibitory effects on the somatic
0166-6851/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 6 - 6 8 5 1 ( 9 9 ) 0 0 1 6 1 - 9
40
B.A. Sheehy et al. / Molecular and Biochemical Parasitology 105 (2000) 39–49
musculature of Ascaris suum and other nematode species when injected into the pseudocoelom or applied directly to isolated neuromuscular preparations [1 – 3]. These peptides, however, generally fail to induce detectable effects on the neuromusculature when applied externally to intact nematodes. This apparent lack of activity in whole-organism assays is likely a function of both metabolism and absorption. That is, nematode tissues contain enzymes that rapidly degrade FaRPs [4] and, based on a biophysical model for absorption of small organic molecules across the cuticles of A. suum [5,6] and another gastrointestinal nematode, Haemonchus contortus [7], it is predicted that peptides would not be absorbed. However, direct measurements of transport rates for small peptides across nematode cuticles have not been reported. To delineate the factors that govern transport of peptides across the cuticle/hypocuticle complex of nematodes, we measured the rates of absorption of a series of structurally related model peptides and the excitatory FaRP, AF2, using isolated cuticle/hypocuticle and lipid-extracted cuticle segments of A. suum and twochamber diffusion cells. [14C]-Labeled model peptides were prepared from D-phenylalanine, with the amide nitrogens sequentially methylated to give AcfNH2, Acf3NH2, Acf(NMef)2NH2 and Ac(NMef)3NHMe (Fig. 1). These model peptides were designed to allow systematic analysis of the influence of peptide size, hydrogen bonding and lipophilicity on transport without the complication of metabolism [8 – 12]. The rates of transport of [125I]AF2 across isolated cuticle/ hypocuticle and lipid-extracted cuticle were also determined to test the predictions derived from studies on the model peptides, using a derivative of this small neuropeptide that is found in nematodes, including A. suum [13]. Results of these studies were compared with those from analogous studies, conducted previously, that measured rates of transport of the same model peptides across Caco-2 cell monolayers, a model for vertebrate intestinal epithelial transport [8 – 12].
2. Materials and methods
2.1. Model peptides and AF2 A series of model peptides (Fig. 1) and their [14C]-labeled derivatives (specific activity, 15 mCi/mmol) were prepared from D-phenylalanine as previously described [8,9]. [125I]AF2 (KHE*YLRFamide) was synthesized by the nboc method and radiolabeled using chloramineT (Amersham Pharmacia Biotech, Arlington Heights, IL) to 2000 Ci/mmol. Selected physicochemical properties for each model peptide and AF2 are listed in Table 1.
2.2. Parasites, culture medium and cuticle preparations Adult female A. suum, 25–30 cm in length, were collected from swine at a commercial abattoir and stored in Ascaris Ringers Solution (ARS) under air at 37°C. ARS was prepared with the following salts: KCl, 1.83 g/l; CaCl2·2H2O, 1.73 g/l; MgCl2·6H2O, 1.99 g/l; NaCl, 0.23 g/l; Na acetate, 10.25 g/l. The solution was buffered with 5 mM Hepes to pH 7.4. Isolated cuticle preparations were obtained from the parasites within 20–24 h after collection. Parasites selected for dissection were motile and exhibited no signs of tissue necrosis. Starting about 6 cm from either end, 2–3 cm cylindrical coronal sections were cut from the mid-body region of the parasites. Precise cuts were made along one lateral line, allowing the cylindrical sections to be opened. Ovary and digestive tract remnants were removed with forceps, and a dulled-edge scalpel was used to scrape away muscle and most of the hypodermal tissue, leaving only the cuticle and hypocuticle intact [5]. The remaining tissue, colorless and translucent, was rinsed three times in ARS. These tissue segments were stored in ARS at 37°C and used within 1 h of preparation. Some of the scraped cuticle segments were further processed for use in studies to delineate the collagen barrier of the cuticle. Following the third rinse in ARS, these segments were extracted overnight in 20 ml ice-cold chloroform/methanol
B.A. Sheehy et al. / Molecular and Biochemical Parasitology 105 (2000) 39–49
41
Fig. 1. Structures of model peptides and AF2. Methyl groups on the model peptides that contribute to faster transport kinetics are shaded. Abbreviations: Ac-f-NH2, acetamido-D-phenylalanylcarboxamide; Ac-f2-NH2, acetamido-D-phenylalanyl-D-phenylalanylcarboxamide; Ac-f3-NH2, acetamido-D-phenylalanyl-D-phenylalanyl-D-phenylalanylcarboxamide; Ac-f2-(NMe-f)NH2, acetamido-Dphenylalanyl-D-phenylalanyl-N-methyl-D-phenylalanylcarboxamide; Ac-f-(NMe-f)2-NH2, acetamido-D-phenylalanyl-N-methyl-Dphenylalanyl-N-methyl-phenylalanylcarboxamide; Ac-(NMe-f)3-NH2, acetamido-N-methyl-D-phenylalanyl-N-methyl-D-phenylalanyl-N-methyl-D-phenylalanylcarboxamide; Ac-(NMe-f)3-NHMe, acetamido-N-methyl-D-phenylalanyl-N-methyl-D-phenylalanylN-methyl-D-phenylalanyl-N-methylcarboxamide; AF2, Lys-His-Glu-Tyr-Leu-Arg-Phe amide.
42
D-phenylalanyl
oligomer peptides and AF2
Peptide
Molecular weight
Molecular radiusa (A, )
Aqueous diffusion coefficient (×10−6 cm s−1)
Number of Hbonds
Log PC, n-octanol/ waterb
Log PC, n-heptane/ glycolc
1. 2. 3. 4. 5. 6. 7. 8.
206 353 500 514 528 542 556 991
4.41 5.28 5.92 5.98 6.03 6.09 6.14 7.32
7.40 6.19 5.51 5.46 5.41 5.36 5.32 3.30
5 7 9 8 7 6 5 30
0.05 1.19 2.30 2.63 2.53 2.92 3.24 −9.35d
−5.46 −6.52 −7.10 −6.28 −5.14 −4.20 −2.86 NDe
AcfNH2 Acf2NH2 Acf3NH2 Acf2(NMef)NH2 Acf(NMef)2NH2 Ac(NMef)3NHMe Ac(NMef)3NHMe KHEYLRF-NH2 a
Calculated using the Stokes–Einstein equation [5,9]. Log of partition coefficient in n-octanol/water [9]. c Log of partition coefficient in n-heptane/glycol [12]. d Calculated at pH 7.4 using CLOGP programs. e Not determined. b
B.A. Sheehy et al. / Molecular and Biochemical Parasitology 105 (2000) 39–49
Table 1 Physicochemical properties of model
B.A. Sheehy et al. / Molecular and Biochemical Parasitology 105 (2000) 39–49
(2:1). The segments were rinsed three times in 20 ml cold ARS and stored in ARS at 4°C until use. This process removed lipids from the tissue, thereby eliminating the lipoidal barrier to transcuticular diffusion [5].
2.3. In 6itro transcuticular kinetic studies Studies to determine the transport kinetics of the four model peptides tested and AF2 across cuticle preparations were performed at 37°C, using two-cell diffusion systems (Crown Glass, Philadelphia, PA). Cuticle preparations were mounted between the windows of a matching pair of diffusion cells. The system was then placed onto motor-driven magnetic stirring units (Crown Glass). M-9 buffer (3.5 ml) was added to the receiver side (the side facing the hypocuticle) of each system. M-9 buffer was prepared with the following salts: Na2HPO4, 6.0 g/l; KH2PO4, 3.0 g/l; NaCl, 5.0 g/l; MgSO4·H2O, 0.25 g/l; pH 7.4. Radiolabeled peptide (1 mCi/ml final concentration) plus unlabeled peptide (25 mM final concentration) was added to the donor cell (epicuticle side) of each cell system immediately prior to each experiment. Both the donor and receiver solution volumes were 3.5 ml and the diffusional area was 0.636 cm2. Contents of the diffusion cells were stirred with bar magnets throughout the experiment to minimize the effects of the aqueous boundary layer on permeation kinetics [5]. The entire system was allowed to equilibrate for 5 min, then three 10 ml samples were obtained from both the donor and receiver chambers. Samples from studies using the model peptides were spotted onto a gridded filtermat for analysis by liquid scintillation counting (LKB Betaplate Model 1205; Pharmacia, Turku, Finland). Subsequent samples were collected and spotted onto the filtermat periodically over 24 h. For plotting purposes and determination of flux rates, only data collected during the first 6 – 8 h were used. This ensured that, for both lipid-extracted and unextracted preparations, data were from linear portions of the plots. Samples collected during testing of [125I]AF2 were pipetted into 20 ml plastic vials and counted using a Beckmann-5500B counter (Beckmann Instruments, Inc., Fullerton, CA). To
43
determine if leakage or other edge effects had occurred, residual solution in the donor chamber was removed after each assay (after the 24 h sample was collected) and replaced with 3.5 ml M-9 containing 0.5% blue dextran, a hydrophilic, high molecular weight dye that does not diffuse across the lipid-extracted or unextracted cuticle [5]. The receiver chamber was observed after 24 h for evidence of dye accumulation (lower limit of blue dextran detection was 0.00025%); data from systems that showed evidence of leakage were not used.
2.4. High-performance liquid chromatography radiodetection of model peptides To determine if the peptides AcfNH2 or Acf(NMef)2NH2 were degraded during transport studies, high-performance liquid chromatography (HPLC) assays were carried out on a Varian Vista Series Model 5000 HPLC (Varian, Walnut Creek, CA) equipped with a LDC/Milton Roy (Riviera Beach, FL) SpectroMonitor D variable-wavelength UV detector, a Varian Model 9090 autosampler and a Radiomatic FLO-ONE/Beta Model CR radiodetector (Radiomatic, Meriden, CT). Ten-microliter samples from donor and receiver chambers were diluted to 50 ml with distilled water and then 30 ml of the diluted sample was injected onto a Hypersil column (ODS C-18, 5 mm, 250× 4.6 mm2; Alltech Associates, Inc., Deerfield, IL). The mobile phase was 17% acetonitrile, 83% buffer (1 g/l NaH2PO4·H20, pH 7.5) for assaying AcfNH2 and 50% acetonitrile, 50% buffer for Acf(NMef)2NH2. Mobile phase flow rate was 1 ml/min and was mixed with Radiomatic FLO-SCINT II (Radiomatic) scintillation cocktail at a ratio of 4:1. UV absorbance was measured at 205 nm with retention times of 7.4 min for AcfNH2 and 5.7 min for Acf(NMef)2NH2. 3. Results
3.1. In 6itro transport kinetics HPLC/radiodetection analyses of AcfNH2 and Acf(NMef)2NH2 revealed no evidence for peptide
44
B.A. Sheehy et al. / Molecular and Biochemical Parasitology 105 (2000) 39–49
degradation during the course of 24 h transport kinetic studies. Therefore, samples examined by liquid scintillation counting only were used to determine transport kinetics for these and each of the other peptides tested. In vitro transport studies were carried out in cuticle/hypocuticle (non-lipid extracted) and cuticle (lipid extracted) tissue preparations. Fig. 2 shows the time course of the concentration of Ac(NMef)3NHMe appearing in the receiver chamber with respect to the initial concentration in the donor chamber, CR(t)/CD(0). The general kinetic profiles of Ac(NMef)3NHMe are typical of the profiles for the other peptides tested, in which transport across the cuticle/hypocuticle was slower than transport across the lipid-extracted cuticle. When there was a lag time to steady-state linear kinetics, it occurred with non-lipid-extracted cuticle. In the case of the nematode peptide, AF2, the number of counts in the receiver chamber did not increase significantly over the course of 9 or 24 h incubations when unextracted cuticle/hypodermis segments were used, indicating that significant transport across the lipid-containing cuticle did not occur (Fig. 2). In contrast, AF2 diffused across lipid-extracted cuticle segments at a rate of
4.3× 10 − 6 cm s − 1, or about one-third as fast as Acf(NMef)2NH2. These results indicate that the lipoidal hypocuticle is the rate-limiting barrier to AF2 transport, and that diffusion through the aqueous pores of the collagen cuticle occurs at a rate consistent with the restrictions imposed by the size (7.32 A, radius) of this peptide.
3.2. Peptide permeability coefficients for the cuticle and the hypocuticle To gain mechanistic insights to the passive diffusional processes, the steady-state kinetic data were treated to yield permeability coefficients. The effective permeability coefficient was calculated by Pe =
VR DCR ACD(0) Dt
(1)
where Pe is the effective permeability coefficient, VR the volume of the receiver chamber, A the cross-sectional area of the cuticle/hypodermis or lipid-extracted cuticle tissue, CD(0) the initial donor concentration, and DCR/Dt the linear change in receiver concentration with time, found by least-square analysis [5]. The effective permeability coefficient is further described by the general expression for the cuticle/hypodermis
Fig. 2. Transport kinetics of Ac(NMef)3NHMe (left) and AF2 (right) across cuticle/hypocuticle ( ) and lipid-extracted cuticle () from A. suum. Lines were derived by linear regression analysis, using data collected during the first 8 h. Data collected at 21 and 24 h timepoints not shown.
B.A. Sheehy et al. / Molecular and Biochemical Parasitology 105 (2000) 39–49
45
Table 2 Permeability coefficients of model peptides and AF2 across isolated cuticle preparations of A. suum Peptide
Permeability coefficient (×10−6 cm s−1) Cuticle
Cuticle/hypocuticle 1. AcfNH2
3. Acf3NH2
5. Acf(NMef)2NH2
7. Ac(NMef)3NHMe
0.9 1.4 0.6 0.8 0.9 90.3 0.5 0.3 0.4 0.49 0.2 2.9 2.1 2.0 2.8 2.59 0.4 10.2 4.5 5.4 6.7 92.5
33.8 9 1.4
NDb 14.3 12.8 18.4 9.8 13.8 9 3.1 21.6 18.1 11.4 17.0 94.2 3.6 5.3 4.4 3.7 4.2 90.8
0
b
Hypocuticlea
33.7 32.1 35.5
8. KHEYLRF-NH2 (AF2)
a
% Hypocuticle-controlled
0.92
97.2
0.4
98
3.1
82
11.1
61
0
100
Calculated using Eq. (2). Not determined; estimated value based on Stokes radius is 15×10−6 cm s−1.
Pe =
1 (2/Paq)+(1/Pcuticle) +(1/Phypocuticle)
(2)
where Paq is the permeability coefficient of the aqueous boundary layer, Pcuticle the permeability coefficient of the collagen cuticle, and Phypocuticle the permeability coefficient of the hypodermal tissue which underlies the cuticle. The physical interpretation of this expression is that, in the cuticle/hypodermis preparation, there are four diffusional transport barriers in series: the two aqueous boundary layers at the epicuticle (donor) and hypocuticle (receiver) sides of the preparation, the collagen cuticle, and the lipoidal hypodermal tissue associated with the hypocuticle. Under well-stirred conditions, the diffusional barriers of the aqueous boundary layers are insignificant [5] and can be ignored in the data analyses. By extracting the lipid from the cuticle/
hypodermis tissue, using chloroform/methanol, this equation simplifies to Pe for the cuticle: Pe =
1 (2/Paq)+ (1/Pcuticle)
(3)
After calculating Pcuticle, Phypocuticle is readily delineated using Eq. (2). For the model peptides, the permeability coefficients for the cuticle/hypocuticle in Table 2 ranged from 0.9× 10 − 6 to 6.7× 10 − 6 cm s − 1; among the peptides, the ascending rank order being Acf3NH2 B AcfNH2 B N-methylated peptides. Among the N-methylated peptides, increasing the number of N-methyls resulted in increased permeability coefficients. In contrast, the permeability coefficients for the lipid-extracted cuticle were 33×10 − 6 cm s − 1 for AcfNH2 and about 15×10 − 6 cm s − 1 for the
46
B.A. Sheehy et al. / Molecular and Biochemical Parasitology 105 (2000) 39–49
N-methylated peptides. Here, permeation across the water-filled, collagen matrix of the cuticle was influenced by molecular size-restricted diffusion across the porous matrix, which was previously determined to be 15 A, in effective pore radius [5]. For instance, AcfNH2 (4.4 A, radius) was faster than the larger N-methyl peptides (6.1 A, radius), which were faster than AF2 (7.3 A, radius). With the effective permeability coefficient values for the cuticle/hypocuticle and cuticle embodied in Eqs. (2) and (3), the permeability coefficients for the lipoidal hypocuticle layer were delineated. It is noted that the permeability coefficient of 2.2 × 10 − 3 cm s − 1 for the aqueous boundary layer on each side of the tissue preparations was sufficiently large so that the aqueous boundary layers contributed insignificantly to the overall mass transfer process involving the cuticle/ hypocuticle or cuticle. The Phypocuticle of AcfNH2 was 0.9×10 − 6 cm s − 1 and significantly smaller than the N-methylated peptides. Additions in Nmethyl groups to Acf3NH2 resulted in increased permeability coefficients. Given the estimates of permeability coefficients for the cuticle/hypocuticle and the lipid-extracted cuticle, one can readily determine the extent to which the collagen cuticle or lipoidal hypocuticle is controlling the overall transport process. It is informative, for instance, that the transport of AcfNH2 across the cuticle matrix is fairly rapid and the rate-determining barrier is almost entirely the hypocuticle layer. On the other hand, the larger and more lipophilic N-methyl peptides traversed the cuticle matrix slower than AcfNH2, and the hypocuticle faster than AcfNH2. The net result is that the hypocuticle is clearly the principal barrier to peptide diffusion, with the collagen cuticle becoming increasingly restrictive with increasing peptide size.
4. Discussion Previous mass transport studies using A. suum [5,6] and other nematodes [7,14] have established the principle that the cuticle/hypodermis complex provides a pathway for the absorption of anthelmintics and a wide range of other lipophilic
permeants that is quantitatively more important than the intestine. These studies have demonstrated that, for lipophilic molecules of molecular weight B 2000, lipid components in the hypodermis/cuticle complex form the rate-limiting barrier to diffusion. Results from our studies using unextracted hypodermis/cuticle tissue from A. suum indicate that the lipoidal biophase also forms the rate-determining barrier to peptide transport across this structure. Permeation of the hypocuticle appears to involve partitioning at the water/ lipid interface (i.e. at the base of aqueous pores in the collagen matrix of the cuticle) followed by diffusion through the lipoidal biophase. There is no evidence for the existence of a carrier-mediated transport mechanism at the hypocuticle membrane for the peptides tested. Following extraction of the lipids from this tissue, the findings are consistent with the principles of molecular size-restricted diffusion across the aqueous channels of the porous collagen matrix of the cuticle [5]. Our results further suggest that, for the model peptides tested, the principles that govern in vitro permeation across the hypocuticle are similar to those established for the lipid membranes in vertebrate cell monolayers [9]. They involve, principally, the interplay of lipophilicity and hydrogen bond potential. In Tables 1 and 2, it is apparent that molecular size is not a major determinant of the rate of transport across the hypocuticle. This is best evidenced by the observation that the Phypocuticle value for Ac(NMef)3NHMe is threefold greater than that for Acf(NMef)2NH2, even though these two model peptides have comparable molecular weights (MWs) and radii, while Pe for AcfNH2 (MW= 206) is much lower than those for the larger peptides Acf(NMef)2NH2 (MW= 528) or Ac(NMef)3NHMe (MW= 556). A similar pattern was shown in analogous studies using these model peptides and the Caco-2 cell system [9]. It was suggested that this pattern could be due to the effects of increased alkylation of the amide bonds [9]. That is, a strong correlation exists between the permeability across Caco-2 cell monolayers (and A. suum hypocuticle) and the number of methyl groups attached to nitrogens in the backbone of the peptide.
B.A. Sheehy et al. / Molecular and Biochemical Parasitology 105 (2000) 39–49
To examine the influence of lipophilicity, Phypocuticle is plotted against the partition coefficient (log PC, n-octanol/water) in Fig. 3. The partition coefficient is the fraction of molecules that distribute in an organic phase (n-octanol in this case) versus an aqueous phase (water in this case), and provides an estimate of how readily a molecule will penetrate a lipoidal membrane. There is essentially no increase in permeability coefficients between AcfNH2 and the more lipophilic Acf3NH2, although there is an approximately 50-fold increase in log PC between them. Upon increasing the extent of N-methylation of Acf3NH2, there is a proportional increase in Phypocuticle with lipophilicity. In general, the trends mimic the profile for the Caco-2 cell monolayer system. Although lipophilicity is an important factor, it alone is not sufficient to account for all the effects of N-methyl addition on membrane transport of peptides [9]. Desolvation of the amide backbone and other hydrogen bonding groups in the molecule is also a requirement. Following methylation, NH groups are no longer available to form hydrogen bonds with water, thereby reducing the energy required for transport of the peptide across lipoidal membranes. Energy is required to break strong hydro-
47
gen bonds in the aqueous phase before the peptide is able to reach the required transition state for transfer from the external, aqueous medium to the hydrophobic lipid interior of the hypocuticle [5]. Therefore, it follows that removal of hydrogen bond potential lowers the amount of energy required for the peptide to enter the lipid phase. Although the hydrogen bond number may explain the increase in Pe observed between Acf(NMef)2NH2 and Ac(NMef)3NHMe, it alone cannot describe all the permeability results reported here. For example, both the smallest peptide, AcfNH2, and Ac(NMef)3NHMe, which has four more methyl groups and two additional amino acids, contain five potential hydrogen bonds. If their transport was governed solely by hydrogen bond number, then the amount of energy required to break these bonds prior to transport across the cuticle would be equivalent for these peptides. However, the larger peptide crosses the cuticle 10-fold faster than the smaller peptide. Furthermore, Acf(NMef)2NH2, with seven potential hydrogen bonds, has two more amino acids and methyl groups than AcfNH2, yet traverses the cuticle threefold faster. These observations are explicable if it is recognized that not all hydrogen bonds are energetically equivalent. Simply count-
Fig. 3. Relationship between permeability coefficients for model peptides across A. suum cuticle/hypocuticle ( ) and Caco-2 cell monolayer (), and log PC values for n-octanol/water (left) and n-heptane/ethylene glycol (right). Numbers by data points refer to model peptides shown in Tables 1 and 2 and Fig. 1. Lines were derived by linear regression analysis.
48
B.A. Sheehy et al. / Molecular and Biochemical Parasitology 105 (2000) 39–49
ing potential hydrogen bonds is a useful strategy for demonstrating the concept, but it naively assumes that all such bonds are energetically equally solvated. In reality, it is well recognized that hydrogen bond strength depends on a variety of factors including heteroatom type, along with inductive and steric hindrances. A better model for estimating the desolvation contribution to solute transport is the n-heptane/ ethylene glycol partition coefficient system. It places emphasis on the desolvation required of the various polar groups in the peptide, rather than on the hydrophobic effect [10 – 12]. This system also appears from our results to be more representative of the barrier to transport imposed by the nematode hypocuticle, where hydrophobicity does not appear to be as integral to peptide transport. As seen with the Caco-2 cells in Fig. 3, the peptides permeate the hypocuticle more readily (i.e. log Phypocuticle becomes less negative) with increasing methylation and accompanying reduction of hydrogen bonds (i.e. as log PC, n-heptane/ ethylene glycol, becomes less negative). The plots are both linear and appear to be approximately parallel. The perceived trend toward lower permeability across the hypodermis/cuticle (i.e. relative to the Caco-2 cell monolayer) with larger peptides may be attributable to the expected increasing influence of size-restriction imposed by the collagen barrier of the cuticle [5,7,14]. For all of the peptides, transport across the hypocuticle occurs slower than that across the Caco-2 cells by about an order of magnitude, in terms of permeability coefficients. Although we do not know the basis for this difference, it may be partially due to the greater thickness of the lipid barrier presented by the nematode hypocuticle. Based on micrographs of A. suum body wall sections stained with osmium vapor or Nile red, the lipoidal barrier of the hypocuticle is approximately 10 mm thick [5]. This is about 103-fold greater than the combined apical and basolateral membranes of the Caco-2 cell monolayers [8]. AF2 and other FMRFamide-related neuropeptides isolated thus far from nematodes contain six to 15 amino acids, are less methylated and far less lipophilic than the model peptides used in this study (Table 1). Based on our results showing that
AF2 does not traverse the isolated cuticle/hypocuticle complex of A. suum, and theoretical considerations from extrapolation of transport kinetics observed for the model peptides tested, it is predicted that none of the nematode FaRPs identified thus far would be absorbed across the cuticle of A. suum. In contrast, all of the anthelmintics currently marketed are lipophilic molecules capable of diffusing rapidly across the cuticles of all nematode species studied, including A. suum, H. contortus and Trichostrongylus colubriformis [5–7,14]. This probably accounts for the easily detectable and often extremely rapid effects of these compounds on the motility of intact parasites during in vitro exposure. Conversely, it is likely that the inability of FaRPs to traverse the cuticle of nematodes contributes significantly to their apparent inactivity when tested using intact organisms. The present study, however, does not account for the possibility that FaRPs may be absorbed across the cuticles of other species, including larval stages, where the hypodermis/cuticle complex is often much thinner. Finally, in the context of this study, it is useful to consider the experimental compound, PF1022A, a cyclic depsipeptide of fungal origin [15] that possesses remarkably potent actions of rapid onset against some nematode species during in vitro exposure [16,17]. PF1022A contains four amide bonds, each of which is N-methylated. The only other polar functionalities in this molecule are four ester bonds, which are not very significant from a permeability perspective. Furthermore, the molecule has no charge, but numerous hydrophobic residues. These overall structural characteristics are reminiscent of those in cyclosporin A. Cyclosporin A is a neutral, cyclic undecapeptide, comprised of hydrophobic amino acids, with all but two of the amide bonds Nalkylated. The remaining two amides are oriented in such a way as to facilitate intramolecular hydrogen bond formation [11]. This further reduces the desolvation energy associated with movement from an aqueous environment into a lipoidal membrane. Consistent with these structural features, both PF1022A and cyclosporin A are readily absorbed by nematodes (unpublished observations). With respect to our conclusions
B.A. Sheehy et al. / Molecular and Biochemical Parasitology 105 (2000) 39–49
regarding structure/transport relationships for peptidic molecules, they illustrate precisely the sort of structure that one would propose for a permeable peptide (i.e. relative to their native, linear sequences).
References [1] Cowden C, Stretton AOW, Davis RE. AF1, a sequenced bioactive neuropeptide isolated from the nematode Ascaris suum. Neuron 1989;2:1465–73. [2] Bowman JW, Winterrowd CA, Friedman AR, Thompson DP, Klein RD, Davis JP, Maule AG, Blair KL, Geary TG. Nitric oxide mediates the inhibitory effects of SDPNFLRFamide, a nematode FMRFamide-related neuropeptide, in Ascaris suum. Journal of Neurophysiology 1995;74:1880 – 8. [3] Maule AG, Geary TG, Bowman JW, Shaw C, Halton DW, Thompson DP. The pharmacology of nematode FMRFamide-related peptides. Parasitology Today 1996;12:351 – 7. [4] Sajid M, Isaac RF. Metabolism of AF1 (Lys-Arg-GluPhe-Ile-Arg-Phe-NH2) in the nematode, Ascaris suum. Molecular and Biochemical Parasitology 1996;75:159–68. [5] Ho NFH, Geary TG, Raub TJ, Barsuhn CL, Thompson DP. Biophysical transport properties of the cuticle of Ascaris suum. Molecular and Biochemical Parasitology 1990;41:153 – 66. [6] Ho NFH, Geary TG, Barsuhn CL, Sims SM, Thompson DP. Mechanistic studies in the transcuticular delivery of antiparasitic drugs II: ex vivo/in vitro correlation of solute transport by Ascaris suum. Molecular and Biochemical Parasitology 1992;52:1–14. [7] Ho NFH, Sims SM, Vidmar TJ, Day JS, Barsuhn CL, Thomas EM, Geary TG, Thompson DP. Theoretical perspectives on anthelmintic drug discovery: interplay of transport kinetics, physicochemical properties, and in vitro activity of anthelmintic drugs. Journal of Pharmaceutical Sciences 1994;83:1052–9.
.
49
[8] Conradi RA, Hilgers AR, Ho NFH, Burton PS. The influence of peptide structure on transport across caco-2 cells. Pharmaceutical Research 1991;8(12):1453– 60. [9] Conradi RA, Hilgers AR, Ho NFH, Burton PS. The influence of peptide structure on transport across caco-2 cells II: peptide bond modification which results in improved permeability. Pharmaceutical Research 1992;9(3):435 – 9. [10] Burton PS, Conradi RA, Hilgers AR, Ho NFH, Maggiora LL. The relationship between peptide structure and transport across epithelial cell monolayers. Journal of Controlled Release 1992;19:87 – 98. [11] Conradi RA, Burton PS, Borchardt RT. Physicochemical and biological factors that influence a drug’s cellular permeability by passive diffusion. In: Pliska V, Testa B, VandeWaterbeemd H, Mannhold R, Kubinyi H, Timmerman H, editors. Lipophilicity in Drug Action and Toxicology. Weinheim: VCH Publishers, 1996. [12] Paterson AD, Conradi RA, Hilgers AR, Vidmar TJ, Burton PS. A non-aqueous partitioning systemfor predicting the oral absorption potential of peptides. Quantitative Structure – Activity Relationships 1994;13:4 – 10. [13] Cowden C, Stretton AOW. AF2, an Ascaris neuropeptide: isolation, sequence and bioactivity. Peptides 1993;14:423 – 30. [14] Thompson DP, Ho NFH, Sims SM, Geary TG. Mechanistic approaches to quantitate anthelmintic absorption by gastrointestinal nematodes. Parasitology Today 1993;9:31 – 5. [15] Sasaki T, Takagi M, Yaguchi T, Miyadoh S, Okada T, Koyama M. A new anthelmintic cyclodepsipeptide, PF1022A. Journal of Antibiotics 1992;45:692 – 7. [16] Terada M, Ishih A, Tungthrongchitr M, Sano M, Shomura T. Effects of PF1022A on developing larvae of Angiostrongylus costaricensis in mice, with special reference to route, dose and formulation. Japanese Journal of Parasitology 1993;42:199 – 210. [17] Conder GA, Johnson SS, Nowakowski DS, Dutton FE, Nelson SJ, Thomas EM, Davis JP, Thompson DP. Anthelmintic profile of the cyclodepsipeptide PF1022A in in vitro and in vivo models. Journal of Antibiotics 1995;48:820 – 3.
