Urocanic acid is a major chemoattractant for the skin-penetrating parasitic nematode Strongyloides stercoralis

Abstract

Host-seeking behavior by parasitic nematodes relies heavily on chemical cues emanating from potential hosts. Nonspecific cues for Strongyloides stercoralis, a nematode that infects humans and a few other mammals, include carbon dioxide and sodium chloride; however, the characteristic species specificity of this parasite suggested the existence of other, more specific cues. Here we show that the infective larva of S. stercoralis is strongly attracted to an extract of mammalian skin and that the active component in this skin extract is urocanic acid. Urocanic acid, a histidine metabolite, is particularly abundant in mammalian skin and skin secretions, suggesting that it serves as an attractant specific to mammalian hosts. The attractant activity of urocanic acid is suppressed by divalent metal ions, suggesting a possible strategy for preventing infection.

From both a medical and socioeconomic standpoint, skin-penetrating nematodes are among the most important helminth parasites of humans (1–3). The hookworms Ancylostoma duodenale and Nector americanus parasitize >600 million people globally and contribute significantly to widespread iron-deficiency anemia in the tropical and subtropical world. Hookworms are the causative agents of ill health, diminished work capacity, depressed physical and cognitive development in children, and poor school performance. Consequently, these parasites have a major socioeconomic and medical impact (2, 4). The threadworm Strongyloides stercoralis parasitizes 300 million people globally (5). It is sometimes an HIV-associated nematode. In patients who are immunosuppressed, either as a result of intercurrent disease or intentionally, as a consequence of treatment, the parasite can multiply internally to fatal levels of abundance. Thus, fatalities occur even in sophisticated hospital facilities in highly developed countries (6, 7).

Even though soilborne, skin-penetrating nematode parasites cause ill health throughout much of the world and have the adverse socioeconomic effects described above, little is known about the sequence of biological and physicochemical events that constitute the infective process (1), the process one would wish to interrupt to prevent infection. The host-finding behavior of the free-living infective larvae of the two major hookworms of humans has only recently been analyzed and described in detail (8); the host-provided chemical signals involved in skin penetration have been identified, but those involved in host-finding remain unknown, as do the signal transduction systems resulting in parasite growth and development. These systems, however, are being described in S. stercoralis, a parasitic nematode that is more easily maintained in the laboratory. Neurons that detect thermal gradients and some chemical signals have been identified and are considered to be important in host-finding (9, 10); however, the specific chemical signals that attract skin-penetrating infective larvae to the host have yet to be determined. Here we report the isolation and identification of a chemoattractant that, when provided either as a natural product or as a reagent chemical, attracts infective larvae in an in vitro system.

Results

Because dogs, along with humans and other primates, are natural hosts of S. stercoralis, the infective larvae of this parasite are attracted to a crude aqueous extract of canine skin. Inspection of the tracks left by larvae as they migrate toward a well containing crude or purified attractant showed that, even from a distance of 30 mm, migration is strongly directed toward the sample well (Fig. 1). Fig. 2 shows a family of dose–response curves illustrating larval attraction to such an extract at a series of dilutions. Preliminary experiments showed that attractant activity was retained after heating of the attractant for 10 min at 80°C or digestion with trypsin (0.25% for 30 min). Attractant activity was recovered in the filtrate after ultrafiltration through Amicon membranes (Millipore, Billerica, MA) with nominal molecular weight cut-offs from 30,000 to 500 Da. In these preliminary experiments, the larvae also responded positively to an aqueous extract of skin from gerbils, which are a permissive experimental host for S. stercoralis; felines, in contrast, are rarely infected by S. stercoralis, and larvae were not attracted to an extract of feline origin.

Fig. 1.

The chemoattractant assay. (A) Diagram of the plates used to assay chemoattraction. (B) Tracks made by L3 larvae in response to unfractionated canine skin extract.

Fig. 2.

Dose–response curves for chemoattraction, using serial dilutions of unfractionated canine skin extract. Dilutions are indicated on the graph.

The chemoattractant was isolated from the crude extract of canine skin by three successive chromatographic steps: anion exchange (Fig. 3A), gel filtration (Fig. 3B), and hydrophobic adsorption (Fig. 3C). The purified material consisted of a single component on RP-HPLC (Fig. 3D). Analysis by GC-MS showed a major component at 138 Da, and comparison of its mass spectrum with the National Institute for Standards and Technology database identified it as urocanic acid (Fig. 4A). The purified attractant and a commercial sample of urocanic acid (Sigma–Aldrich, St. Louis, MO) showed identical UV absorbance spectra (Fig. 4B). ¹H NMR (Fig. 4C) was consistent with published results (11). A second component, of mass 94 Da, was also found in the isolated chemoattractant and was identified as 3-aminopyridine (data not shown). Because the chemoattractant activity behaved as an anionic compound, it was concluded that this second component was a contaminant.

Fig. 3.

Chromatographic isolation and analysis of chemoattractant. (A) Anion-exchange chromatography of the crude extract on DEAE-cellulose. The chemoattractant was predominantly in fractions 25–35. (B) Gel filtration chromatography of DEAE fractions 25–35 on Sephadex G25. The chemoattractant was predominantly in fractions 55–70. (C) Hydrophobic adsorption chromatography of G25 fractions 55–70 on AmberChrom CG-71S. The chemoattractant was found in fractions 66–90. (D) Analytical RP-HPLC of the purified chemoattractant after chromatography on AmberChrom CG-71S. All fractions throughout the chemoattractant peak showed a single component, eluting at 200 sec.

Fig. 4.

Spectroscopic characterization of the isolated chemoattractant. (A) Mass spectrum of isolated chemoattractant (Lower) and reference spectrum of urocanic acid (Upper). (Inset) Structure of urocanic acid [3-(4-imidazolyl)acrylic acid]. The relative intensity, I, of the 39-atomic mass unit (amu) fragment (NCHN) is lower in the experimental spectrum than in the reference spectrum. (B) UV absorbance spectra of reagent-grade urocanic acid (upper trace) and chemoattractant-containing fractions from the CG71 column (lower traces). (C) ¹H NMR spectrum of the purified chemoattractant. Chemical shifts are stated in ppm relative to tetramethylsilane. The major resonances are assigned based on published data (11): imidazole C2, 8.6373 ppm; imidazole C5, 7.6405 ppm; C_α, 7.3010–7.1796 ppm; C_β, 6.5340–6.4798 ppm.

Because urocanic acid is known to bind metal ions (12), chemoattractant activity was assayed in the presence of equimolar calcium, magnesium, and manganese. All three of these physiologically important metals inhibited chemoattraction: the response of larvae to urocanic acid was reduced by ≈50% by equimolar magnesium and by ≈75% by equimolar calcium or manganese (Fig. 5). Conversely, the response to purified urocanic acid (both isolated from skin and in the form of commercial preparations) was elevated by pretreatment with Chelex to remove residual metal ions. A maximal response, equivalent to the response to unfractionated skin extract, was observed at 150–200 mM urocanic acid (Fig. 6). Migration is strongly directed toward the sample well (Fig. 7), as is the case when the unfractionated extract is used (Fig. 1B).

Fig. 5.

Inhibition of the chemoattractant activity of 150 mM urocanic acid by equimolar metal ions, assayed as described in Fig. 1 and in the text. Filled squares, no metal added; open triangles, equimolar MgCl₂; crosses, equimolar MnCl₂; open circles, CaCl₂.

Fig. 6.

Dose–response curves for chemoattraction to Chelex-treated urocanic acid. Filled squares, control (deionized water). Urocanic acid concentrations are indicated as follows: filled triangles, 25 mM; crosses, 50 mM; open circles, 100 mM; filled circles, 150 mM; open squares, 200 mM.

Fig. 7.

Tracks made by larvae in response to Chelex-treated urocanic acid. As in Fig. 1, the larvae were placed at the left and right edges of the plate; the well at the top contained urocanic acid, whereas the control well contained deionized water.

Discussion

S. stercoralis infective larvae respond positively to several physicochemical attractants. They are attracted to warmth, showing a strong response to temperatures in the range of avian and mammalian body temperature (9). They also respond positively to 3.3–4% carbon dioxide (13) and to sodium chloride (10), both nonspecific attractants produced by terrestrial vertebrates. Although urocanic acid is too widespread a component of mammalian skin to be a host-specific attractant, it may well bias penetration behavior to a limited group of mammals in preference to other species and, in combination with other physical and/or chemical factors, it could be one factor in determining host specificity. Whether host specificity is determined by external physicochemical attractants, internal nutritional requirements, or other physiological factors or by immunological compatibility remains unknown for helminth parasites.

Urocanic acid is produced by deamination of histidine (14) and is abundant in mammalian skin (15). The concentration in the stratum corneum varies between different individuals and at different sites on the body. Most of the reported values are in the range of 6–12 nmol/cm², except for the sole of the foot, where the observed concentration was ≈60 nmol/cm² (16). Given the approximate thickness of the stratum corneum as 10 μm, the values for human skin correspond to 6–12 mM and 60 mM, respectively. In isolating urocanic acid from canine skin, the amount recovered was ≈20 nmol/cm², and it is likely that a significant fraction was lost during purification. Thus, the concentrations of pure urocanic acid that elicit a response in the bioassay are within the physiological range for sloughed-off epidermis. For a soil-dwelling parasite, the high concentrations on the sole of the foot may be particularly relevant to the infective process, attracting larvae to unprotected skin in contact with the ground.

The photochemical activity of urocanic acid has been described by numerous investigators and is metal-dependent (12), which suggested that its chemoattractant activity might also be metal-dependent. Somewhat surprisingly, the converse was found to be true: the activity of both reagent urocanic acid and urocanic acid isolated from skin extract was increased by removal of contaminating metal ions, which resulted in chemoattractant activity equal to that of the crude attractant. Thus, the active form of the attractant appears to be free, rather than metal-bound, urocanic acid. It is possible that, during purification, removal of other metal-binding compounds from the crude extract increased the availability of uncomplexed metal ions, thus promoting the formation of urocanic acid–metal complexes.

The observation that metal ions inhibit chemoattraction suggests that it may be possible to develop an inexpensive, practical, topical preventative for use on exposed body surfaces in persons at risk of infection with the skin-penetrating larvae of S. stercoralis. Such a strategy would be consistent with the current emphasis on “appropriate technology” to prevent parasitic disease in poor developing countries, as advocated by leading nongovernmental organizations such as the Carter Center, the Gates Foundation, and the World Health Organization.

Materials and Methods

Parasites.

Parasites were obtained as described by Forbes et al. (10).

Preparation of the Extract.

Canine skin from laboratory animals euthanized for purposes of another investigation was generously provided by Mark Haskins (School of Veterinary Medicine, University of Pennsylvania) through an Animal Byproducts Transfer Protocol, as approved by the Institutional Animal Care and Use Committee of the University of Pennsylvania. A 15 × 25-cm rectangle of skin was excised, and the outer layer was removed in strips, which were reduced to a paste with a meat grinder. Portions (5 g each) of the paste were extracted with 10 ml chloroform/methanol/water, 2:1:0.2 (vol/vol) (17). Phase separation was accelerated by centrifugation at 1,800 × g for 15 min. The aqueous phase was partially dried by rotary evaporation to remove residual organic solvent, then lyophilized to dryness. The lyophilized material was redissolved with deionized water, using 0.2 ml/g of macerated skin.

Isolation of the Attractant.

Redissolved extract (50 ml) was incubated with 20 g Amberlite XAD-16 (Sigma–Aldrich) to remove hydrophobic components, without loss of chemoattractant activity. Subsequent chromatographic steps were performed at 4°C, and fractions were screened for attractant activity as described below. The Amberlite-treated extract was loaded onto a DEAE-cellulose column (DE-52, 2.5 × 22 cm; Whatman, Florham Park, NJ) and eluted with a 350 × 350-ml gradient of 0–0.4 M triethylammonium bicarbonate at 40 ml/h; fractions were collected at 10-min intervals. Active fractions were pooled, lyophilized, redissolved in 2.5 ml of deionized water, and applied to a 1.5 × 55-cm column of Sephadex G25-Fine (Amersham Pharmacia Biotech, Piscataway, NJ). The column was eluted with deionized water at 12 ml/h, and fractions were collected at 10-min intervals. Active fractions were again lyophilized, redissolved in 1.5 ml of deionized water, and applied to a 1.5 × 54-cm column of AmberChrom CG-71S (Tosoh Bioscience, Tokyo, Japan). The column was eluted with deionized water at 40 ml/h, and fractions were collected at 2-min intervals.

Analytical Methods.

RP-HPLC was performed using a polar-embedded reverse-phase column (RP-Amide C16, 5 μm, 4.6 × 150 mm; Supelco, Bellefonte, PA); elution was isocratic at 1 ml/min, using deionized water as the mobile phase. ¹H NMR spectroscopy was performed at the Center for Biomedical NMR of the University of Pennsylvania School of Medicine, under the direction of Krzysztoff P. Wroblewski, using a Bruker (Billerica, MA) AMX-500 spectrometer at 500 MHz. Analysis by GC-MS was performed by M-Scan (West Chester, PA).

Reagent-grade urocanic acid (Sigma–Aldrich) was suspended in deionized water and neutralized with 1 M Tris(hydroxymethyl) aminomethane base. The final concentration was determined from its absorbance at 277 nm, using the extinction coefficient 18,800 M⁻¹cm⁻¹ (18). To remove trace divalent metal ions, the solution was incubated with 0.2 g Chelex-100 resin (BioRad, Hercules, CA) per ml for 4 h.

Assay for Chemoattractant Activity.

Assays were performed in 60 × 15-mm plastic Petri dishes containing 5 ml of 0.7% Bacto-Agar (Becton Dickinson, Franklin Lakes, NJ) in deionized water. Two wells, 3 mm in diameter, were punched into the agar layer, opposite each other near the edge of the dish and ≈40 mm apart (Fig. 1A). The wells were filled with 15 μl of either test or control solution, and the solutions were allowed to diffuse into the agar for 2 h. Thereafter, two groups of three larvae were introduced into each dish, as shown in Fig. 1; one group was placed about halfway between the two wells near one side of the dish, and the other group was placed opposite the first group, near the other side of the dish, and again approximately halfway between the two wells. The distance between the placement spots and each of the wells was ≈35 mm. The presence of larvae in either of the wells was recorded at 4-min intervals for 28 min. Their response was considered positive when the larvae entered the sample well and remained there during successive observations. Duplicate plates were used for each assay; thus, each experimental point in an assay represents the behavior of 12 larvae. An unfractionated aqueous extract of canine skin was used as a positive control in all assays. To check for external factors that might influence the direction of migration, plates were set up without attractant and in various orientations, and the migration of L3 larvae was monitored. When the tracks were to be recorded photographically, a filter membrane (black, gridded HABG, 47 mm; Millipore) was placed on top of the agar layer. Holes were punched through the membrane over the wells in the agar, and the filter was allowed to absorb the fluid layer. Larvae were then placed on the filter as described above, and their now easily visible tracks were observed to confirm a chemotactic response and were documented photographically (Fig. 1B).