INFECTION AND IMMUNITY, Mar. 1992, P. 1018-10230019-9567/92/031018-06$02.00/0Copyright © 1992, American Society for Microbiology

Hyaluronidase from Infective Ancylostoma Hookworm Larvaeand Its Possible Function as a Virulence Factor in Tissue

Invasion and in Cutaneous Larva MigransP. J. HOTEZ, 12,3* S. NARASIMHAN,4 J. HAGGERTY,5 L. MILSTONE,5

V. BHOPALE,6 G. A. SCHAD,6 AND F. F. RICHARDS"7

Yale MacArthur Center for Molecular Parasitology' and Departments of Pediatrics, 2 Epidemiology and Public Health,3Molecular Biophysics and Biochemistry,4 Dermnatology,' and Internal Medicine,7 Yale University School ofMedicine,

New Haven, Connecticut 06510, and Department ofPathobiology, University of PennsylvaniaSchool of Veterinary Medicine, Philadelphia, Pennsylvania 191046

Received 5 August 1991/Accepted 19 December 1991

During skin penetration, infective hookworm larvae encounter hyaluronic acid as they migrate betweenepidermal keratinocytes and through the ground substance of the dermis. A hyaluronidase would facilitatepassage through the epidermis and dermis during larval invasion. Zoonotic hookworm larvae of the genus

Ancylostoma were shown to contain a hyaluronidase activity that migrated on modified sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE) hyaluronic acid gels with an apparent Mr of 49,000. Asecond form with an M, of 87,000 was also identified. The major etiologic agent of cutaneous larva migrans,A. braziliense, was shown to have the greatest enzyme activity, hydrolyzing up to 3.3 Kg of hyaluronic acid per

h per ,ug of total parasite protein at pH 6.0, whereas A. caninum and A. tubaeforme each had much less enzymeactivity. The differences in enzyme activities between species correlated with differences in the intensities of thelytic zones at 49 and 87 kDa on SDS-PAGE hyaluronic acid gels. Hookworm hyaluronidase activity exhibiteda broad pH optimum between 6.0 and 8.0 and did not hydrolyze chondroitin sulfate, two features that suggestthat the hookworm enzyme is more like the invertebrate leech hyaluronidase than mammalian testicular or

lysosomal hyaluronidase. Larvae of A. braziliense were shown to release hyaluronidase activity and degraderadiolabeled hyaluronic acid in vitro. Gold sodium thiomalate was identified as an enzyme inhibitor. Thehyaluronidase is the second major virulence factor that we have identified from infective hookworm larvae.

Clinically, hookworm infection is one of the most impor-tant helminthiasis affecting humans (10, 13). Transmission tohumans occurs when the third-stage larvae enter through theskin. Parasite-derived enzymes, including a 68-kDa metallo-protease that we recently identified from Ancylostoma duo-denale and A. caninum (14), appear to facilitate this process.In addition to a protease, the early literature describes a

so-called "spreading factor" from infective helminth larvaethat causes dissemination of dyes that were previouslyinjected into skin (17, 18). The spreading activity was

postulated to be a hydrolytic enzyme that depolymerizes themucopolysaccharide hyaluronic acid, a major component ofthe dermal ground substance (17, 19, 30).

Hyaluronic acid not only constitutes the ground substanceof the dermis, but also functions as a cell adhesion molecule(22). Cell-to-cell adhesion is mediated by a hyaluronic acidreceptor that is preferentially expressed on proliferatingepithelial cells, such as the keratinocytes in the basal layer ofthe epidermis (1, 31). Hyaluronic acid may function as a

bridge connecting epidermal cells expressing the receptor. Ahyaluronidase from infective hookworm larvae would serve

to facilitate migration through the ground substance of thedermis and act as a chemical scissors to break apart cellswithin the basal layer of the epidermis. We developed assaysfor hyaluronidase activity to show that infective hookwormlarvae of the genus Ancylostoma produce 49- and 87-kDahyaluronidases that are related. The enzyme activity isgreatest in A. braziliense, the species that migrates exten-

* Corresponding author.

sively in the basal layer of the epidermis to cause cutaneouslarva migrans (7, 27-29).

MATERIALS AND METHODS

Collection of hookworm larvae. Feces containing embryo-nated eggs ofA. braziliense orA. tubaeforme were obtainedfrom an experimentally infected cat, and eggs ofA. caninumwere obtained from an experimentally infected dog. Third-stage larvae were reared in coprocultures prepared withbone charcoal or by a modified Harada-Mori filter paperculture method. Larvae were recovered from the coprocul-tures with a Baermann apparatus. The sedimented larvaewere washed with phosphate-buffered saline (PBS) or waterby using low-speed centrifugation prior to biochemical anal-ysis. The washings from the final centrifugation were savedas a control. Living larvae were either used directly or storedfrozen. Homogenates were prepared by grinding the larvaein either PBS or sodium acetate buffer (pH 6.0) at 4°C in a

glass homogenizer. A soluble extract from A. braziliensewas also prepared by centrifuging the high-ionic-strengthhomogenate for 10 min at 12,000 x g in a microcentrifugeand recovering the supernatant.

Hyaluronidase assays. (i) Liquid-phase hyaluronic acid deg-radation. Purified hyaluronic acid from human umbilicalcord (Sigma Chemical Co., St. Louis, Mo.) was suspendedin 0.05 M sodium acetate buffer (pH 6.0) at a final concen-

tration of 0.5 mg/ml and stored at 4°C. Enzyme activity wasassayed in a final volume of 100 ,ul containing 8 ,u ofsubstrate (final concentration, 40 ,ug of hyaluronic acid perml) in 0.05 M sodium acetate (pH 6.0). The reaction was

1018

Vol. 60, No. 3

on June 19, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

HYALURONIDASE FROM INFECTIVE HOOKWORM LARVAE 1019

initiated by adding a 1- to 5-pl aliquot of larval homogenateand allowed to proceed for 16 to 24 h at 37°C. The assayswere also done at different pHs in either 0.05 M sodiumacetate buffer or Tris hydrochloride buffer. At the end ofthe incubation period, the reaction was stopped by addinga solution containing the dye 1-ethyl-2-[3-(1-ethyl-naphthol[1,2d] - thiazolin - 2 - ylidene) - 2 - [methyl - propenyl]naphthol[1,2d] thiazolium bromide (Stains-All; Sigma Chemical Co.)that interacts with hyaluronic acid but not with its productsof digestion by hyaluronidase (5). Approximately 17 mg ofStains-All powder was mixed in a solution containing 50%formamide and 0.06% glacial acetic acid immediately prior touse (the dye decomposes in light). The assay was stopped byadding 0.9 ml of the Stains-All solution, and then the A64owas measured immediately. As controls, the Stains-Allsolution was added either to sodium acetate buffer alone orto buffer containing 40 ,ug of hyaluronic acid per ml. Theenzyme activity was expressed as micrograms of hyaluronicacid hydrolyzed per hour per microgram of protein.

Since some mammalian hyaluronidases, e.g., testicularhyaluronidase, also cleave glycosidic bonds in chondroitinand chondroitin sulfate, the enzyme assays were also donewith the latter as a substrate. Like hyaluronic acid, chon-droitin sulfate but not its degradation products will also bindthe Stains-All dye. The sodium salt of chondroitin sulfate A(containing some chondroitin sulfate C) from bovine trachea(Sigma Chemical Co.) was suspended in 0.05 M sodiumacetate, and the assays were done as described above.Hyaluronidase from bovine testes (Sigma Chemical Co.) wasused as a positive control.

(ii) SDS-PAGE hyaluronic acid gels. Larval homogenates insodium dodecyl sulfate-polyacrylamide gel electrophoresis(SDS-PAGE) sample buffer (65 mM Tris hydrochloride [pH6.8], 10% glycerol, 3% SDS) were applied to a polyacryl-amide gel containing hyaluronic acid (8). The procedure wasmodified to incorporate SDS into the gel, thereby allowingseparation of the enzymatic activities as a function ofmolecular weight. The separating gel contained 2.5 ml of 1.5M Tris hydrochloride (pH 8.8) and 0.4% SDS, 3.3 ml of28.4% acrylamide and 1.6% bisacrylamide (final acrylamideconcentration was equal to 11%), 0.45 ml of 0.5-mg/mlhyaluronic acid substrate, 0.75 ml of 6.4% dimethylamino-propionitrile, 2.0 ml of water, 0.01 ml of TEMED(N,N,N',N'-tetramethylethylenediamine), and 0.06 ml of0.1-mg/ml ammonium persulfate. The samples were appliedwithout heating to the gels and run in one dimension undernonreducing conditions. Low voltage (less than 100 V) wasused to prevent heating. Gels that were either 0.75 or 1.0 mmthick were used. The gel was washed once in 2.5% TritonX-100 for 40 min and then washed in 0.05 M sodium acetatebuffer (pH 6.0) for 1 h. The gel was then incubated in freshbuffer overnight at 37°C. The gel was covered to preventexposure to light and incubated first in 50% formamide for 30min at room temperature and then for 2 h in 1 mg ofStains-All (diluted 20-fold in 50% formamide) per ml. Bandsof enzyme activity were visualized after placing the gel inwater for 30 to 60 min. The lytic zones of activity were

initially pink against a blue background but rapidly fadedwith exposure to light. The gel was either immediatelyphotographed or kept covered prior to drying.

(iii) SDS-PAGE autoradiography of radiolabeled hyaluronicacid. Radiolabeled hyaluronic acid was prepared from[3H]glucosamine-labeled glycosaminoglycan synthesized byhuman keratinocytes (16). The radiolabeled hyaluronic acidwas purified by elution through DEAE-Sephacel with 0.2 MNaCl in 8 M urea and then passage through a column (100 by

1.5 cm) containing Sepharose 4B. The [3H]hyaluronic acidwas collected in the void volume fraction. Digestion of asample of this material with commercial Streptomyces hy-aluronidase revealed that it was greater than 99% hyaluronicacid. The [3H]hyaluronic acid was mixed with 90 ,ug ofhuman umbilical cord hyaluronic acid prior to ethanol pre-cipitation. The final precipitate was resuspended in Tris-saline (20 mM Tris hydrochloride buffer (pH 8.2) containing150 mM NaCI) at a concentration of 0.03 mg of hyaluronicacid per ml. The specific radioactivity was approximately27,000 cpm/,ug of hyaluronate.

(iv) Degradation by living larvae. To observe hyaluronicacid digestion by living hookworm larvae, approximately8,000 larvae of A. braziliense in a volume of 100 ,ul of PBScontaining antibiotics (1,000 U of penicillin per ml and 1 mgof streptomycin per ml) were incubated with 150 ,ul of[3H]hyaluronic acid in Tris-saline buffer at 37°C. Aliquots of25 pl were removed at 0, 1, 3, and 6 h and placed inSDS-PAGE sample buffer with 0.1% 2-mercaptoethanol. Asa negative control, the [3H]hyaluronic acid was also incu-bated in the presence of 100 ,ul of larval washings andincubated under identical conditions. As a second negativecontrol, an aliquot of [3H]hyaluronic acid was aliquoteddirectly into SDS sample buffer. The samples containingpartially degraded [3H]hyaluronic acid were applied onto anSDS-6% polyacrylamide gel with a 3% stacking gel. Afterelectrophoresis, the gel was fixed and prepared for autora-diography by soaking the gel in Autofluor (National Diag-nostics, Somerville, N.J.).

(v) Degradation by soluble larval extract. Aliquots of theA.braziliense soluble extract containing approximately 16 ,ug ofprotein were incubated with 15 ,ul of radiolabeled hyaluronicacid and 2.5 ,ug of cold carrier hyaluronic acid in a finalreaction volume of 50 pl for 0 to 18 h at 37°C in 0.05 Msodium acetate buffer (pH 6.0). A negative control reactionwas also done in the absence of extract. The reactions werestopped by the addition of 10 ,ul of SDS-PAGE sample buffercontaining 1 mM dithiothreitol and 0.1% 2-mercaptoethanoland electrophoresed on a 7.5% polyacrylamide gel. Afterelectrophoresis, the gel was fixed in 40% methanol-10%acetic acid for 1 h prior to autoradiography.To determine the effect of the inhibitor gold sodium

thiomalate (Myocrisin; Merck Sharpe & Dohme, WestPoint, Pa.) on hookworm hyaluronidase, quantities of theinhibitor ranging from 0.001 mg (0.05 mM) to 1.0 mg (55 mM)were preincubated with 16 ,ug of soluble A. brazilienseextracts for 10 min at 37°C prior to incubation with radiola-beled hyaluronic acid and cold carrier hyaluronic acid for 18h.

RESULTS

To identify a hyaluronidase activity from hookworm lar-vae, we used a sensitive assay based on the observation thathyaluronic acid, but not its products of degradation, inter-acts with the carbocyanine dye Stains-All to shift the wave-length of maximal absorbance in the visible spectrum of thedye toward longer wavelengths (5). We found that differ-ences between the substrate and products of the reactionwere detected only by the addition of a small amount ofglacial acetic acid to the dye. Hyaluronidase activities weredetected in all three species of infective hookworm larvaeafter overnight incubations of larval homogenates with thesubstrate (Table 1). Homogenates of A. braziliense larvaehad the greatest enzyme activity compared with homoge-nates of the dog hookworm A. caninum and another cat

VOL. 60, 1992

on June 19, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

1020 HOTEZ ET AL.

TABLE 1. Hyaluronidase activities of Ancylostomalarval homogenatesa

Species Expt 1 Expt 2 Expt 3 Expt 4 Expt 5

A. braziliense NDb >0.2 >0.6 3.3 2.2A. caninum 0.07 0.06 0.1 ND NDA. tubaeforme ND 0.1 ND ND ND

a Numbers show micrograms of hyaluronic acid hydrolyzed per hour permicrogram of parasite protein. All experiments were done at 37°C and pH 6using approximately 4 ,ug of substrate. When approximately 100% of thesubstrate was hydrolyzed over the time course of the experiment, the enzymeactivity is expressed as a minimum value.

b ND, not done.

hookworm, A. tubaeforme, catalyzing the hydrolysis of upto 3.3 ,ug of hyaluronic acid per h per ,ug of parasite protein.Some variation was noted between experiments, possiblyreflecting the variations in the larval age in culture at the timewhen the larvae were collected.To determine whether the enzyme activity was specific for

cleaving hyaluronic acid, we also tested another glycos-aminoglycan, chondroitin sulfate A, as a possible substrate.Some mammalian hydrolases will typically cleave glycosidicbonds in chondroitin and chondroitin sulfate, whereas someinvertebrate hyaluronidases such as that described from theleech will function as an endoglucuronidase that specificallyhydrolyzes hyaluronic acid (33). As shown in Table 2,chondroitinase activity was found in the commercial prepa-ration of bovine testicular hyaluronidase, but not in A.braziliense.

Hyaluronidase activities from A. braziliense and A. cani-num were examined as a function of pH. Activities fromboth species showed a broad range of optimal pH between 6and 8, declining sharply under either more acidic or alkalineconditions (Fig. 1). The overall activity of the A. caninumhyaluronidase was again substantially less than that of theA.braziliense hyaluronidase. The remainder of the hyaluroni-dase studies from hookworm larvae were done at pH 6.To characterize the hyaluronidase activity further, we

separated larval homogenate proteins by SDS-PAGE undernonreducing conditions in a polyacrylamide gel that con-tained the hyaluronic acid substrate. This technique allowedthe detection of hyaluronidase activity within the gel, basedon the principles that (i) the polyacrylamide matrix immobi-lizes the high-molecular-weight substrate and prevents itsmigration in an electrical field, (ii) the gel is incubated underconditions that allow degradation of the substrate by theenzyme, and (iii) the substrate within the matrix stains withthe carbocyanine dye Stains-All to create a blue backgroundon which a pink zone of hyaluronidase activity will appear

TABLE 2. Glycosaminoglycan substrate specificityof A. braziliense hyaluronidase

Source of enzyme % Hydrolysis'(Ag) Hyaluronic acid Chondroitin sulfate

Bovine testes (5.0)b 96.4 ± 0.0 36.5 ± 1.0A. braziliense (0.1) 14.3 ± 4.7 0.0 ± 1.4A. braziliense (1.2) 91.5 + 0.8 0.0 ± 0.6

a Numbers refer to the percentage of substrate hydrolyzed over 20 h at 37°Cat pH 6. Approximately 8 ,ug of chondroitin sulfate A or 4 pLg of hyaluronicacid was used as the substrate. Each experiment was done in duplicate.

b Commercially prepared bovine testicular hyaluronidase (Sigma ChemicalCo.), reported to have a specific activity of 290 U/mg of solid.

5C)

4

-oa)NC 32-0

-o

CM

0- A. brazilienseo--O A. caninum

2 3 4 5 6 7 8 9 10 3

pH4 5 6 7 89

pH

FIG. 1. pH optima of Ancylostoma hyaluronidase activities.Approximately 0.06 ,ug of A. braziliense larval homogenate proteinin PBS was placed in a tube containing 4 ,ug of hyaluronic acid in 100,ul of either 0.05 M sodium acetate buffer (pH 3 to 8) or 0.05 M Trishydrochloride buffer (pH 9) and incubated at 37°C. After 14 h, 0.9 mlof the Stains-All solution was added and the A640 was determined.Approximately 1.96 ,ug of A. caninum larval homogenate proteinwas placed in a tube containing 4 ,ug of hyaluronic acid in 100 ofeither 0.05 M sodium acetate buffer (pH 4 to 6) or 0.1 M Trishydrochloride buffer (pH 7 to 9) and incubated at 37°C. After 22 h,0.9 ml of the Stains-All solution was added and the A64o wasdetermined. A second, expanded scale for the pH optimum of A.caninum hyaluronidase activity is also shown.

(8). The polyacrylamide gel detection of hyaluronidase ac-tivity was modified to incorporate SDS within the gel andrunning buffer, thus allowing the parasite proteins to sepa-rate by apparent molecular mass. Incubation of the gel in thenonionic detergent Triton X-100 facilitates the displacementof the SDS and allows the enzyme to renature within the gel(14), thereby permitting visualization of the hyaluronidaseactivities that separate by apparent molecular weight.The larval homogenates of both A. braziliense and A.

caninum exhibited a major band of hyaluronidase activityhaving an Mr of 87,000 (Fig. 2). When the same amount ofparasite protein was applied to the gel, the intensity of A.braziliense activity was much greater than that of A. cani-num and reflected the differences in enzyme activity seen inTable 1. A second band of activity having an Mr of 49,000was also seen in larval homogenates of A. caninum. Toexplore the relation of the 87-kDa form of the hyaluronidaseto the 49-kDa form, we ran a second SDS hyaluronidase

...... ... t .. ..~k

Ab Ak

FIG. 2. SDS-PAGE hyaluronic acid gel. Approximately 9 ,ug ofeitherA. braziliense (A.b.) orA. caninum (A.c.) larval homogenateprotein in PBS (prepared by grinding larvae in a glass homogenizerand then rapid freezing in dry ice) was thawed and applied inSDS-PAGE sample buffer (without 2-mercaptoethanol) to a nonre-

ducing gel containing hyaluronic acid. After completion, the gel waswashed in 2.5% Triton X-100 and then 0.05 M sodium acetate buffer(pH 6.0). The gel was incubated overnight in the same buffer at 37°C.Hyaluronidase activity is shown as a clear zone after incubation inStains-All and then destaining in water.

0.10

0.08

0.06

-0.04

0.02

_ 0.0010

INFECT. IMMUN.

on June 19, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

HYALURONIDASE FROM INFECTIVE HOOKWORM LARVAE 1021

2O kDa-

'$ I

0 1 3 6 0 1 3 6 HoursHA 11*9-a .

A.b. A.C. AltFIG. 3. SDS-PAGE hyaluronic acid gel. Larval homogenates

containing 10 ,ug of A. braziliense (A.b.), 30 pg of A. caninum(A.c.), or 6 ,ug of A. tubaeforme (A.t.) protein in 0.1 M Trishydrochloride buffer (pH 7.5) (prepared by grinding larvae in a glasshomogenizer and then freezing and thawing them several times) wasapplied in SDS-PAGE sample buffer (without 2-mercaptoethanol) toa nonreducing gel containing hyaluronic acid. After completion, thegel was washed in 2.5% Triton X-100 and then 0.05 M sodiumacetate buffer (pH 6.0). The gel was incubated overnight in the samebuffer at 37°C. Hyaluronidase activity is shown as a clear zone afterincubation in Stains-All and then destaining in water.

activity gel after the larval homogenates had been frozen andthawed in low-ionic-strength buffer (0.05 M sodium acetate).As seen in Fig. 3, the hyaluronidase activities from A.braziliense, A. caninum, and A. tubaeforme each migratedwith an Mr of 49,000, suggesting that this form of the enzymeis derived from a higher-molecular-weight precursor. Unlikethe 87-kDa form, the 49-kDa form is highly resistant tofurther biochemical manipulation. The shift in molecularweight may reflect dimer formation. The presence of SDS inthe gel did not denature the enzyme activity after reactiva-tion with Triton X-100, although the possibility remains thatwe did not detect other hyaluronidases that were otherwisedenatured irreversibly by detergent. In some cases, pinkbands would appear on the gel that were not associated withhyaluronidase activity, especially at the bromophenol bluedye front and with one commercial preparation of molecularweight markers. To confirm that the 49-kDa band reflectedhyaluronidase activity, we heat inactivated the sample toshow that disappearance of the band correlated with a loss inactivity (data not shown).To determine whether living larvae release the hyaluroni-

dase activity in vitro, we incubated them with radiolabeledhyaluronic acid at 37°C. The reaction mixture was subse-quently sampled at hourly intervals prior to application ontoSDS-PAGE and then autoradiography (Fig. 4). The larvaebegan to hydrolyze the substrate by 1 h and went on tocompletely degrade the hyaluronic acid. Intermediate prod-ucts of digestion were not seen on autoradiography, suggest-ing that the larvae hydrolyzed the substrate to very lowmolecular weight sugars. No hyaluronidase activity wasidentified in the negative control larval washings over thesame incubation period.The radiolabeled hyaluronic acid was also enzymatically

digested by soluble extracts of A. braziliense. Extractscontaining approximately 16 jig of protein hydrolyzed thesubstrate in a time-dependent manner and proceeded tocompletion by 18 h (data not shown). To identify a potentialinhibitor of the hyaluronidase from hookworm larvae, werepeated the experiment by preincubating the soluble ex-tracts with various concentrations of gold sodium thiomalate(Myocrisin) for 10 min at 37°C (Fig. 5). This compound wasshown previously to inhibit sperm hyaluronidase (17, 18). Ata concentration of 1 mg/ml (33 mM), gold sodium thiomalate

FIG. 4. Degradation of hyaluronic acid by living A. brazilienselarvae. Approximately 8,000 living third-stage larvae (L3) in 100 ,ulof PBS (containing antibiotics) were added to 150 ,ul of approxi-mately 4.5 pg of [3H]hyaluronic acid (approximate specific radioac-tivity of 27,000 cpm/,ug) in Tris-saline buffer (20 mM Tris hydrochlo-ride [pH 8.2], 150 mM NaCl) at 37°C. Aliquots of 25 p,l wereremoved at 0, 1, 3, and 6 h and added to SDS-PAGE sample buffer(containing 2-mercaptoethanol). Larval washings (W) in an identicalvolume were used as controls. Samples were subjected to autoradi-ography after application to a 6% polyacrylamide separating gel (3%stacking gel). HA, hyaluronic acid.

completely inhibited the enzymatic digestion of radiolabeledhyaluronic acid by A. braziliense soluble extract over 18 h.Partial inhibition was apparent when the inhibitor was usedat a concentration between 5 and 33 mM, while there waslittle or no inhibition below 5 mM.

DISCUSSIONOur results suggest that in addition to a protease, hook-

worm larvae of the genus Ancylostoma produce a secondenzyme that may function as a virulence factor in larvalinvasion, namely, a hyaluronidase. We identified both 87-and 49-kDa forms of the enzyme, each found in A. brazil-iense,A. caninum, and A. tubaeforme. The greatest enzymeactivity was found in A. braziliense, the species classicallyidentified as the cause of cutaneous larva migrans (7, 27-29).The carbocyanine dye binding method for hyaluronidase

detection (5) was sensitive enough to measure the enzymaticactivity found in 60 ng of total protein from A. brazilienselarval homogenates and was more sensitive than turbidimet-ric measurements (6) or colorimetric methods for the esti-

.nht>jor - cs * Ca co

Soncerlraslon ° ° o c o o

(i"g. °

i

c

,G

FIG. 5. Inhibition of A. braziliense hyaluronidase by gold so-dium thiomalate. Approximately 16 ,ug-of larval soluble extract in0.5 M sodium acetate buffer (pH 6.0) (prepared by grinding thelarvae in a glass homogenizer and then centrifuging in a microcen-trifuge for 10 min at 12,000 x g at 4°C and recovering the superna-tant) was preincubated with various concentrations of gold sodiumthiomalate for 10 min at 37°C. The reaction was initiated by theaddition of approximately 0.45 p.g of [3H]hyaluronic acid (approxi-mate specific radioactivity of 27,000 cpml,ug) and 2.5 pLg of coldcarrier hyaluronic acid and allowed to proceed for 18 h at 37°C. Acontrol reaction was done in the presence or absence of larvalsoluble extract and the absence of inhibitor. The reactions werestopped by the addition of 10 pLl of SDS-PAGE sample buffer(containing 2-mercaptoethanol) and applied on an SDS-7.5% poly-acrylamide gel prior to autoradiography.

L3 w

VOL. 60, 1992

on June 19, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

1022 HOTEZ ET AL.

mation of N-acetylamino sugars (25). Hookworm larvalhyaluronidase activity was not detected by either of theselatter methods (data not shown), an observation in agree-ment with a previous study (17).Hyaluronidase pH optima are characteristically sensitive

to changes in salt concentration and protein concentrationand to oligomerization of the subunits (12). We found thatthe pH optima of the hyaluronidases fromA. braziliense andA. caninum were similar to that described from the leechHirudo medicinalis, which has activity in the neutral pHrange (33), but were different from the pH optima reportedfor mammalian hyaluronidases such as sperm hyaluronidase(12), which are active in the acid pH range. The absence ofchondroitinase activity in hookworm larval homogenatesfurther suggests that the hookworm hyaluronidase belongsto the class of endoglucuronidases that characterize theinvertebrate leech hyaluronidase but not to a class of mam-malian hyaluronidases (33).The intensity of larval hyaluronidase activities seen on

SDS-polyacrylamide gels correlated with the quantitativeactivities measured in liquid-phase assay. Just as a larvalhomogenate of A. braziliense had the greatest quantitativeenzyme activity, it also produced the greatest amount ofsubstrate lysis on activity gels. The activity zone migratedwith an apparent Mr of 87,000. The other zoonotic hook-worm species also produced lytic zones at this molecularweight but with less enzyme activity. The Mr of the Ancy-lostoma species hyaluronidase is similar to that described forthe two monomeric forms (alpha and beta) of sperm hyal-uronidase having Mr values of 89,600 and 81,200, respec-tively (12), and for the bacteriophage-associated hyaluroni-dase that was reported to be a glycoprotein with an Mr of71,000 (4). Interestingly, the open reading frame of a genecorresponding to the streptococcal hyaluronidase predicts apolypeptide with an Mr of 40,000, suggesting that the enzymeis heavily glycosylated (15).A second form of the enzyme, from Ancylostoma spp.

having an Mr of 49,000, was also seen. We believe that thetwo forms are related and that possibly the 49-kDa form wasderived from the 87-kDa form through degradation by pro-teolysis or gluconylysis. The evidence for degradation is that(i) repeated freezing and thawing of the larval homogenateresulted in the appearance of the 49-kDa form and (ii)preparation of the homogenate at nonphysiological highionic strength or by rapid freezing in dry ice sometimesprevented the appearance of the 49-kDa form. Hyaluroni-dases are reported to be destabilized at low ionic strengthand by manipulations such as shaking and dialysis (12). Asan alternative explanation, some hyaluronidases can oligo-merize by forming intermolecular disulfide bonds to producethe oligomer series run m, 2m, 4m, 6m, etc. (11, 12). Thus,the 49-kDa hookworm hyaluronidase may represent the truemonomeric form of the enzyme that forms an 87-kDa dimer.We have recently found that during purification, only the49-kDa form can be recovered; this form also binds tohyaluronic acid affinity resins.Our finding that living larvae of A. braziliense release a

hyaluronidase suggests that the enzyme may function duringthe initial stages of parasite invasion through human skin.Hookworm larvae first encounter hyaluronic acid shortlyafter entry into the skin as they try to negotiate passagebetween the keratinocytes located in the basal layer of theepidermis. Hyaluronic acid mediates both the attachmentbetween keratinocytes in this deep epidermal layer and theattachment of keratinocytes to the basement membrane ofthe epidermis-dermis junction (1). Release of a hyaluroni-

dase by invading hookworm larvae would facilitate bothpassage between these cells and access to the basementmembrane. The observation that A. braziliense has thegreatest hyaluronidase enzyme activity is consistent with itsability to migrate through the deeper layers of the epidermiswhere these larvae are characteristically found in histo-pathologic sections of skin in patients with cutaneous larvamigrans (29). In contrast, the zoonotic dog hookworm A.caninum which by comparison contains much less hyalu-ronidase activity is an uncommon cause of cutaneous larvamigrans (21). Most members of the genus Ancylostoma,including A. caninum, A. tubaeforme, and A. duodenale,can infect their host by either the oral or percutaneous route(26), whereas A. braziliense is thought to be an obligatelyskin-penetrating parasite. The differences in hyaluronidaseactivity may also underlie these differences in migratorybehavior, whereas the proteases of these species are similar(14).

Hyaluronic acid is also a major component of the groundsubstance of the dermis. Cinematographic studies of hook-worm larvae migrating through skin show that the dermisoffers little resistance to advancing larvae (20); this is possi-bly related to a hyaluronidase secreted by these parasites.Our finding that gold sodium thiomalate, a low-molecular-weight inhibitor of mammalian hyalurondase (23, 24), alsoinhibits the hyaluronidase from A. braziliense will be usefulas a tool to explore the function of this enzyme as a virulencefactor in ancylostomiasis. We are also exploring the hypoth-esis that hyaluronidases are virulence factors for a number ofinfectious agents that invade connective tissue includingbacteria (4, 9, 15, 32), protozoa, helminths (17, 18), andarthropods (2, 3), as well as metastasizing tumor cells.

ACKNOWLEDGMENTSThis work was supported by the Consortium on the Biology of

Parasitic Diseases of the MacArthur Foundation, by the CharlesHood Foundation, and by Public Health Service grants AI-22662,AI-08614, 1-P30-HD27757-01, and TDRU AI-28778 from the Na-tional Institutes of Health. P.J.H. is a Pfizer postdoctoral fellow anda Culpeper Foundation medical science scholar. P.J.H. and F.F.R.are investigators of the MacArthur Foundation. P.J.H. is also aninvestigator of the Child Health Research Center.

REFERENCES1. Alho, A. M., and C. B. Underhill. 1989. The hyaluronate

receptor is preferentially expressed on proliferating epithelialcells. J. Cell Biol. 108:1557-1565.

2. Allalouf, D., A. Ber, and J. Ishay. 1972. Hyaluronidase activityof extracts of venom sacs of a number of Vespinae (Hy-menoptera). Comp. Biochem. Physiol. 43B:119-123.

3. Barker, S. A., S. I. Bayyuk, J. S. Brimacombe, and D. J. Palmer.1963. Characterization of the products of the action of beevenom hyaluronidase. Nature (London) 199:693-694.

4. Benchetrit, L. C., E. D. Gray, R. D. Edstrom, and L. W.Wannamaker. 1978. Purification and characterization of a hy-aluronidase associated with a temperate bacteriophage of groupA, type 49 streptococci. J. Bacteriol. 134:221-228.

5. Benchetrit, L. C., S. L. Pahuja, E. D. Gray, and R. D. Edstrom.1977. A sensitive method for the assay of hyaluronidase activ-ity. Anal. Biochem. 79:431-437.

6. DiFerrante, N. 1956. Turbidimetric measurement of acid muco-polysaccharides and hyaluronidase activity. J. Biol. Chem.220:303-306.

7. Dove, W. E. 1932. Further studies on Ancylostoma brazilienseand the etiology of creeping eruption. Am. J. Hyg. 15:664-711.

8. Fiszer-Szafarz, B. 1984. Hyaluronidase polymorphism detectedby polyacrylamide gel electrophoresis. Application to hyaluron-idases from bacteria, slime molds, bee and snake venoms,bovine testes, rat liver lysosomes, and human serum. Anal.

INFECT. IMMUN.

on June 19, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

HYALURONIDASE FROM INFECTIVE HOOKWORM LARVAE 1023

Biochem. 143:76-81.9. Fitzgerald, T. J., and L. A. Repesh. 1987. The hyaluronidase

associated with Treponema pallidum facilitates treponemal dis-semination. Infect. Immun. 55:1023-1028.

10. Gilles, H. M. 1985. Hookworm infection and anemia. Rev.Infect. Dis. 7:111-118.

11. Harrison, R. A. P. 1988. Hyaluronidase in ram semen, quanti-tative determination, and isolation of multiple forms. Biochem.J. 252:865-874.

12. Harrison, R. A. P. 1988. Preliminary characterization of themultiple forms of ram sperm hyaluronidase. Biochem. J. 252:875-882.

13. Hotez, P. J. 1989. Hookworm disease in children. Pediatr.Infect. Dis. J. 8:516-520.

14. Hotez, P., J. Haggerty, J. Hawdon, L. Milstone, H. R. Gamble,G. A. Schad, and F. F. Richards. 1990. Metalloproteases frominfective Ancylostoma hookworm larvae and their possiblefunctions in tissue invasion and ecdysis. Infect. Immun. 58:3883-3892.

15. Hynes, W. L., and J. J. Ferretti. 1989. Sequence analysis andexpression in Escherichia coli of the hyaluronidase gene ofStreptococcus pyogenes bacteriophage H4489A. Infect. Im-mun. 57:533-539.

16. Lamberg, J. I., S. H. Yuspa, and V. C. Hascall. 1986. Synthesisof hyaluronic acid and synthesis of proteoglycans is increasedwhen cultured mouse epidermal cells differentiate. J. Invest.Dermatol. 86:659-667.

17. Lee, C. L., and R. M. Lewert. 1957. Studies on the presence ofmucopolysaccharidases in penetrating helminth larvae. J. In-fect. Dis. 100:287-294.

18. Lewert, R. M., and C. L. Lee. 1954. Studies on the passage ofhelminth larvae through host tissues. J. Infect. Dis. 95:18-51.

19. Lincicome, D. R. 1953. A streptococcal decapsulation test fordetection of hyaluronidase activity in animal parasites. Exp.Parasitol. 2:333-347.

20. Matthews, B. E. 1975. Mechanism of skin penetration byAncy-lostoma tubaeforme larvae. Parasitology 70:25-38.

21. Miller, T. A. 1979. Hookworm infection in man. Adv. Parasitol.17:315-383.

22. Miyake, K., C. B. Underhill, J. Lesley, and P. W. Kincade. 1990.Hyaluronate can function as a cell adhesion molecule and CD44participates in hyaluronate recognition. J. Exp. Med. 172:69-75.

23. Perreault, S., L. J. D. Zaneveld, and B. J. Rogers. 1980.Inhibition of fertilization in the hamster by sodium aurothio-malate, a hyaluronidase inhibitor. J. Reprod. Fertil. 60:461-467.

24. Reddy, J. M., C. Joyce, and L. J. D. Zaneveld. 1980. Role ofhyaluronidase in fertilization: the antifertility activity of myo-crisin, a nontoxic hyaluronidase inhibitor. J. Androl. 1:28-32.

25. Ressig, J., J. L. Strominger, and L. F. Leloir. 1955. A modifiedcolorimetric method for the estimation ofN-acetylamino sugars.J. Biol. Chem. 217:959-966.

26. Schad, G. A. 1979. Ancyostoma duodenale: maintenancethrough six generations in helminth naive pups. Exp. Parasitol.47:246-253.

27. Shelmire, B. 1928. Experimental creeping eruption from a catand dog hookworm (A. braziliense). JAMA 91:938-944.

28. Stone, 0. J., and C. J. Willis. 1967. Cutaneous hookwormreservoir. J. Invest. Dermatol. 49:237-239.

29. Sulica, V. I., B. Berberian, and G. F. Kao. 1988. Histopathologicfindings of cutaneous larva migrans. J. Cutaneous Pathol. 15:346.

30. Tamaki, T. 1961. Studies on hyaluronidase of Ancylostomacaninum larvae. Jpn. J. Parasitol. 10:692-696. (Reference notseen, but quoted in Y. Komiya and K. Yasuraoka. 1966. Thebiology of hookworms. Prog. Med. Parasitol. Jpn., vol. III.)

31. Tammi, R., J. A. Ripelino, R. U. Margolis, and M. Tammi. 1988.Localization of epidermal hyaluronic acid using the hyaluronatebinding region of cartilage proteoglycan as a specific probe. J.Invest. Dermatol. 90:412-414.

32. Unsworth, P. F. 1989. Hyaluronidase production in Streptococ-cus milleri in relation to infection. J. Clin. Pathol. 42:506-510.

33. Yuki, H., and W. H. Fishman. 1963. Purification and character-ization of leech hyaluronic acid-endo-,-glucuronidase. J. Biol.Chem. 238:1877-1879.

VOL. 60, 1992

on June 19, 2018 by guesthttp://iai.asm

.org/D

ownloaded from