Proc. Natl. Acad. Sci. USA
Vol. 83, pp. 1213-1216, March 1986
Biochemistry
Earthworms produce a collagen-like substance detected by the
garter snake vomeronasal system
(Lumbricus/Thamnophis/glycoprotein/chemoattractant)
DONALD M. KIRSCHENBAUM*t, NANCY SCHULMANt, AND MIMI HALPERNt
Departments of *Biochemistry and tAnatomy and Cell Biology, Downstate Medical Center, 450 Clarkson Avenue, Brooklyn, NY 11203
Communicated by Chandler McC. Brooks, October 15, 1985
ABSTRACT Earthworms (Lumbricus terrestris) produce a components of much lower molecular weight, contained
chemical substance that is readily detected by and serves as neutral carbohydrate but was devoid of chemoattractant (16).
an attractant for garter snakes (Thamnophis sirtalis). This
chemoattractant is sensed by the vomeronasal system ofsnakes. Snakes were able to discriminate the chemoattractant in
Amino acid analysis of the chemoattractant revealed a high EWW (pH 5.8) from solvent after the EWW was heated for
hydroxyproline/proline ratio and large amounts of serine and up to 120 min in a boiling water bath, but not after the EWW
threonine. More than one-third of the residues were glycine. No was heated for 240 min in a boiling water bath. When a 0.1 M
hydroxylysine and no cysteine were present. Carbohydrate HCl solution of EWW was heated in a boiling water bath,
analyses revealed a high content of galactose (11% by weight) chemoattractant activity was lost within 15 min. However,
and smaller amounts of fucose, mannose, glucose, N-acetylglu- even after a 0.1 M HCl solution of EWW had been heated for
cosamine, and N-acetylgalactosamine. These results were very 3 hr in a boiling water bath, an F2 peak could still be detected
similar to results reported for the amino acid composition and by absorption at 280 nm, carbohydrate assay, and reaction
carbohydrate content of earthworm cuticle collagen and gela- with ninhydrin, although the amount of material detected in
tin. A preparation of purified earthworm cuticle collagen the F2 fraction was much less than originally present at zero
proved to be a potent source of chemoattractant for garter time (17). When a 0.1 M NaOH solution of EWW was heated
in a boiling water bath, chemoattractant power was lost in
snakes. Further, it was not possible to prepare chemoattractant <15 min (17). Concomitant with the loss of the chemoat-
from decuticlized earthworms. These results strongly suggest tractant was the disappearance of the F2 peak, as shown by
that a component of the earthworm chemoattractant for snakes the absence of 280 nm absorption, carbohydrate, and ninhy-
is structurally related to earthworm cuticle collagen. drin-positive material (17). The F2 material was thus very
much more labile in alkali than in acid. After alkali treatment,
Earthworms (Lumbricus terrestris) are a favorite prey of the F4 peak increased, as measured by 280 nm absorption,
garter snakes of the genus Thamnophis and form a major carbohydrate content, and ninhydrin-positive material.
component of their natural diet (1, 2). Garter snakes respond
to objects coated with warm-water washings from earth- This paper presents evidence relative to the structure of
worms by rapid tongue-flicking and an attack (1-3), thus this vomeronasally detected chemoattractant for garter
suggesting that these washings contain a chemoattractant for snakes.
the snake. In addition, garter snakes can be trained to follow
trails of earthworm washings to receive bits of worm as MATERIALS AND METHODS
rewards (4, 5). Garter snakes use their vomeronasal systems,
but apparently do not require a functional olfactory system, Sources. Dialysis tubing was purchased from Thomas.
to detect and appropriately respond to this chemoattractant Prior to use, it was boiled in doubly glass-distilled water
contained in earthworm washings (3, 6-8). Substances that (ddH2O) for 30 min and stored in ddH2O with 0.02% sodium
stimulate the vomeronasal organs are typically naturally azide. Sephadex G-75 and Blue Dextran were purchased from
occurring complex large molecules that have been difficult to Pharmacia. AcA 34 and AcA 44 were purchased from LKB.
characterize chemically (9-14). Bovine serum albumin, ovalbumin, myoglobin, and phenyl-
methylsulfonyl fluoride were purchased from Sigma. 2,4-
Earthworm washings (EWW), prepared as described by Dinitrophenylalanine was prepared in this laboratory. Frac-
Wilde (3) and Burghardt (1), have been found to contain a tions were eluted with the aid ofan ISCO model 312 pump and
chemoattractant for garter snakes that is water-soluble, collected with an ISCO Retriever III fraction collector.
absorbs at 280 nm, contains carbohydrate, and gives a
positive reaction with ninhydrin and positive Lowry and Preparation of a Solution of Chemoattractant from EWW.
Bradford tests for protein (15-17). EWW was prepared (1, 3) from 6 g of earthworms (L.
terrestris; purchased from Connecticut Valley Biological
A solution of EWW that had been dialyzed, lyophilized, Supply, South Hampton, MA) suspended in 20 ml of ddH2O
and reconstituted in 0.15 M NaCl to 5 times its original at 60°C and was centrifuged at ==550 x g for 20 min at room
concentration was separated on various calibrated gel-exclu- temperature. The EWW was dialyzed (Mr 3500 cutoff)
sion columns (16). When 0.15 M NaCl was used as eluent, overnight against ddH20 at room temperature. The dialyzed
two peaks were detected by 280 nm absorption (16). The first solution of EWW was lyophilized, and 15 mg of the resulting
peak, F2, comprising components of Mr > 65,000, contained powder dissolved in 2 ml of 0.15 M NaCl to -5 times its
all the chemoattractant. This peak also contained neutral original concentration. This sample was placed on a 1.6 x
carbohydrate and ninhydrin-positive material and gave pos- 42.5 cm column of AcA 44 that had been calibrated with Blue
itive tests for protein. The second peak, F4, comprising Dextran, bovine serum albumin, ovalbumin, myoglobin, and
dinitrophenylalanine. Fractions (2 ml) were collected and
assayed for A230 and carbohydrates (18).
The publication costs of this article were defrayed in part by page charge Abbreviations: EWW, earthworm wash; ddH2O, doubly glass-dis-
payment. This article must therefore be hereby marked "advertisement" tilled water.
in accordance with 18 U.S.C. §1734 solely to indicate this fact. tDeceased October 4, 1985.
1213
/*F2-D\L/11214 Biochemistry: Kirschenbaum et al.*\A/n|v\ERSTWEH.zt6G<X0eDlvSo1YaW~,Pi/Ia7Ln=-~\by_~fclorsptuoieFJhdfn.IaiGmg;.isProc. Natl. Acad. Sci. USA 83 (1986)
dn///|-~~o~~~k)Amino Acid and Carbohydrate Analyses. The high molec- C * a (I) Chain AL
ular weight fractions in the first major peak that was eluted * F2-DL
from the AcA 44 column, the only fractions that contain 2
chemoattractant, were pooled, dialyzed, lyophilized, and
Hi
hydrolyzed in 6 M HCl for 24 hr at 110TC. The amino acid
analysis of the hydrolysate was done using a Beckman model
120C. The hydroxyproline assays (19) were done on hydro-
ANWCuil W
RESIDUES PER 1000 RESIDUES PH t / X
1. Star diagrams comparing the amino acid contents of
~ ~~~\A\ / ~~~~-
\~~ XHYP /dialyzed, lyophilized F2 (F2-DL) with earthworm cuticle collagen (22)
AX// \/ -71 / I ~7~0~~Ala1 / W> / (A), with earthworm cuticle gelatin (21) (B), and with the al(I) chain
/\//LPC \v l S \ PC \\IL human placental collagen (23) (C3). The amino acids, as indicated
\ each radius, are grouped as follows: glycine (GLY) to isoleucine
(ILE) are the neutral amino acids; phenylalanine (PHE) and tyrosine
(TYR) are the UV-absorbing aromatic amino acids (tryptophan was
RESIDUES PER 1000 RESIDUES not determined); methionine (MET) and cysteine (CYS) are the
l sulfur-containing amino acids; hydroxyproline (HYP) and proline
~~~~~~~~~(PRaOr)e the imino acids that are one of the hallmarks of collagen;
B
acid (ASP) and glutamic acid (GLU) are the acidic amino
/ \ ~~~~~~acids; serine (SER) and threomine (THR) are the hydroxy amino
arginine (ARG), lysine (LYS), and histidine (HIS) are the basic
; ~\~a\mino acids. In B, the proline content of earthworm gelatin was
to be the same as found in earthworm collagen. Note that
in A-C are scaled to 10%o for glycine (GLY/lO) and 50%o for
/Awls/1 a _ \ \ \ ~~~~~~hydroxyproline (HYP/2). In C, plots are scaled to 50%o for proline
/ / t / \_/61 \ LEU X (PRO/2) and alaniine (ALA/2).
//R/\/Ss it\ X \ \IL~~v lypsrepaared tby heeatisng at 1300C for 4 hr. The hydrolysis
and analysis ofEWW and F2 for carbohydrates was done by
1 R. Clamp of the University of Bristol (Bristol, England).
_ /YR of Earthworm Cuticle. Earthworm cuticle
SER// X ^/ w / / was prepared as described by Murray et al. (20), in
a/t\GLU/ /,jC40& / X / / presence ofthe protease inhibitor phenylmethylsulfonyl
Another preparation of earthworm cuticle collagen
was made by the method of Watson (21). The homogeneity of
.eachpreparation was demonstrated by the constancy of the
22) after two separate
~~~~hydroxyproline/protein ratio (20,
~~~reprecipitations for each preparation.
diPreparation of Decuticlized Earthworms. Earthworms were
placed in anhydrous ether for 15 m( m21T)h.e worms were
on eaancd htdrhcuitieecldesiwerre removedmu,ch like peeling off
60 (Isa)ausage casingT. dehceutic iz eeartdhowrmwers e then
\ E/W\Gelatin7 / placed in ddH2aOSP)at gluCand a solution of EWW was prepared
so as described above for intact worms.
90 Behavioral Testing. The testing procedure has been de-
RESIDUES PER 1000 RESIDUES scribed in detail (15). A trial, 2 min in duration, consisted of
the simultaneous presentation of an EWW sample and a
ddH20 or another control sample on Coplin jar covers placed
Biochemistry: Kirschenbaum et al. EWW EWWI EWW2 DECUT EWCI EWC2Proc. Natl. Acad. Sci. USA 83 (1986) 1215
in a presentation tray. The number of tongue flicks directed 48 - FIG. 3. Mean tongue-flick interest scores of snakes presented
to each side of the presentation tray was recorded, and the with earthworm washings (EWW), earthworm washings prepared
time spent in the vicinity of each side of the presentation tray w from worms previously soaked in 60°C ddH20 for 1 min (EWW1),
was timed. For each trial, a tongue-flick interest score 0o 40- earthworm washings prepared from worms previously soaked twice
[number of tongue flicks multiplied by the time (in seconds) (for 1 min each time) in 60°C ddH20 (EWW2), earthworm washings
spent at the sample, divided by the total time spent at the tray] U) prepared from decuticlized earthworms (DECUT), and two different
was determined for the EWW and for the control sample. solutions of purified earthworm cuticle collagen (EWC1 and EWC2).
I- Hatched bars represent scores for test samples (mean ± SEM, n =
Snakes were judged to discriminate test samples from 6); open bars represent control (ddH2O) scores. Asterisks indicate
control samples if the tongue-flick interest score for the test cr32- test scores significantly different (P < 0.05) from control.
sample was significantly higher than the tongue-flick interest
score for the control sample, using a t-test for matched z dinitrophenylalanine markers. A 4-fold higher galactose con-
samples. A 5% level of confidence was used throughout.
o24- centration (38 ,g/ml) was found in F2 than in F4 (10 Ag/ml).
RESULTS
LL A saline solution of purified earthworm cuticle collagen
Amino acid analysis of dialyzed, lyophilized chemoattractant was a strong chemoattractant for garter snakes (Fig. 3). All
isolated by gel-filtration chromatography from EWW (Fig. 1) w the chemoattractant in the cuticle collagen solution was
revealed hydroxyproline in a quantity 1l6-fold greater than :' 16- found in F2; F4 had no chemoattractant activity as determined
proline, when compared on a residues per 1000 residues z by our testing procedures. EWW obtained from decuticlized
basis. Serine and threonine were present in large amounts. 0 earthworms contained no chemoattractant (Fig. 3). Earth-
No hydroxylysine and no cysteine were present. More than
one-third of the residues were glycine. These results suggest- ~ +wormsusedtoprepareEWWcouldnotbeusedtopreparemore chemoattractant solution after two 1-min washes at
ed that the chemoattractant activity resided in an invertebrate
collagen-like molecule. The results from the amino acid 60°C (Fig. 3).
analysis were compared with previously reported analyses of
earthworm cuticle collagen (22), earthworm cuticle gelatin DISCUSSION
(21), and an al(I) chain of human placental collagen (23). The
star diagrams in Fig. 1 illustrate these comparisons. A strong Our results show that a chemoattractant for snakes in EWW
similarity between earthworm cuticle collagen, earthworm is similar in amino acid composition and carbohydrate
cuticle gelatin, and dialyzed, lyophilized F2 (F2-DL) is seen content to earthworm cuticle collagen (22, 24, 25) and gelatin
(Fig. 1 A and B). In contrast, comparison of amino acid (21). Further support for the idea that a chemoattractant in
profiles for the al(I) chain of a vertebrate collagen and for
F2-DL (Fig. 1C) reveals little similarity.
An analysis of dialyzed, lyophilized EWW and a similarly
prepared F2 fraction showed that the former contained 12.3%
galactose, equivalent to 82% of the total carbohydrate pres-
ent, while the latter had 11.1% galactose, equivalent to 74%
of the total carbohydrate present. Earthworm cuticle colla-
gen has been reported to have 10-14% galactose as the major
carbohydrate (24, 25), equivalent to =90%o ofthe total neutral
sugar present (25).
Gel-exclusion chromatography of earthworm cuticle
collagen on an AcA 44 column produced an elution profile
(Fig. 2) very similar to the one obtained in our previous
studies of EWW (16). High molecular weight material (F2)
was eluted close to the Blue Dextran marker and ahead of the
bovine serum albumin marker, and lower molecular weight
material (F4) was eluted between the myoglobin and
DB ALB O\VA MYO DNPA
*
1.2-
E
C
0rn) 0.9
C-)
I-I
o 0.3
C,)
40 50 60 70
ELUTION VOLUME (ml)
FIG. 2. A solution of 15 mg of lyophilized cuticle, prepared by the modified technique of Murray et al. (20), in 2 ml of 0.15 M NaCl was
fractionated on a 1.6 X 42.5 cm column of AcA 44, using 0.15 M NaCl as eluent. Elution positions of markers are shown by arrows. DB, Blue
Dextran; ALB, bovine serum albumin; OVA, ovalbumin; MYO, myoglobin; DNPA, dinitrophenylalanine.
1216 Biochemistry: Kirschenbaum et al. Proc. Natl. Acad Sci. USA 83 (1986)
Table 1. Comparison of earthworm cuticle collagen and garter snake chemoattractant obtained from earthworm
Earthworm cuticle collagen Earthworm chemoattractant
Precipitated with 20%o-saturated ammonium sulfate (20) Precipitated with 90- to 100%-saturated ammonium sulfate (16)
Hydroxyproline/proline ratio 10 (22) Hydroxyproline/proline ratio 415
High Thr + Ser content (22) High Thr + Ser content
No hydroxylysine (22) Little if any (0-1 residue) hydroxylysine
Carbohydrate present; galactose (10-14% by weight) the major Carbohydrate present; galactose (11%) the major carbohydrate;
carbohydrate (24, 25); small amounts of galactosamine and small amounts of fucose, mannose, glucose, galactosamine,
glucosamine (25) glucosamine, N-acetylgalactosamine, and N-acetylglucosamine
3-Elimination of carbohydrate (25) (3-Elimination of carbohydrate (proposed)
Small amount of ninhydrin-reactive material Small amount of ninhydrin-reactive material (16)
Cuticle solution yields an F2 fraction (Fig. 2) F2 fraction has all chemoattractant activity (16)
EWW is earthworm cuticle collagen-related or gelatin-related Scholar of the State University of New York. This research was
is obtained from the observations that solutions of purified supported by National Institutes of Health Grant NS11713.
earthworm cuticle collagen contain a chemoattractant and
that chemoattractant cannot be obtained from washes of 1. Burghardt, G. M. (1966) Psychon. Sci. 4, 37-38.
earthworms whose cuticles have been removed. 2. Burghardt, G. M. (1970) in Advances in Chemoreception, eds.
We had previously observed a rapid loss of chemoattract- Johnston, J. W., Jr., Moulton, D. R. & Turk, A. (Appleton,
ant 280 nm absorption, carbohydrate content, and ninhydrin- New York), Vol. 1, pp. 241-308.
positive material in EWW treated in hot alkali (17). Muir and
Lee (25) and Spiro and Bhoyroo (26) have shown that 3. Wilde, W. S. (1938) J. Exp. Zool. 77, 445-465.
earthworm cuticle collagen has at least 23 attachment sites 4. Kubie, J. L. & Halpern, M. (1975) J. Comp. Physiol. Psychol.
per 1000 residues for O-glycosidically bound carbohydrate on
serine and threonine residues. If these are distributed widely, 89, 667-674.
as suggested by Maser and Rice (27), throughout the collagen 5. Kubie, J. L. & Halpern, M. (1978) J. Comp. Physiol. Psychol.
structure, then the high molecular weight material F2, when
treated with hot 0.1 M NaOH, should form low molecular 92, 362-373.
weight material (F4) due to (3-elimination and hydrolytic cleav- 6. Burghardt, G. M. & Hess, E. H. (1968) J. Comp. Physiol.
age. The consequences of these cleavages should be the loss of
280 nm absorption in F2 and an increase in F4. The increase in Psychol. 66, 289-295.
ninhydrin-positive material in F4 following hot alkali treatment 7. Burghardt G. M. & Pruitt G. H. (1975) Physiol. Behav. 14,
of EWW may be due to generation of amino groups by peptide
bond hydrolysis and to cleavage ofthe amide groups formed by 185-194.
8. Kubie, J. L. & Halpern, M. (1979) J. Comp. Physiol. Psychol.
,(3elimination, to yield ammonia (17).
93, 648-667.
From our accumulated data and inferences from this 9. Singer, A. G., Macrides, F. & Agosta, W. C. (1979) in Chemical
information, we suggest that EWW contains a component of
Signals Vertebrates and Invertebrates, eds. Muller-Schwarze, D.
earthworm cuticle collagen/gelatin that acts as a chemoat- & Silverstein, R. M. (Plenum, New York) pp. 365-375.
tractant for snakes. It is known that earthworm cuticle 10. Novotny, M., Jorgenson, J. W., Carmack, M., Wilson, S. R.,
collagen is precipitated at a lower concentration of ammoni- Boyse, E. A., Yamazaki, K., Wilson, M., Beamer, W. &
um sulfate (20% saturation; ref. 20) than is our chemoattract- Whitten, W. K. in Chemical Signals Vertebrates and Inverte-
ant (90-100%o saturation; ref. 16), and this is to be expected brates, eds. Muller-Schwarze, D. & Silverstein, R. M. (Plen-
if earthworm chemoattractant is a fragment of cuticle um, New York), pp 377-390.
collagen/gelatin. Also, the solution of chemoattractant is pre- 11. Vandenbergh, J. C., Whitsett, J. M. & Lombardi, J. R. (1975)
pared at 60°C, much higher than the denaturing temperature of J. Reprod. Fertil. 43, 515-523.
earthworm cuticle collagen (22°C; ref. 28). From the chemical 12. Singer, A. G., Clancy, A. N., Macrides, F. & Agosta, W. C.
characteristics mentioned, and additional ones listed in Table 1, (1984) Physiol. Behav. 33, 639-644.
we believe that the F2 chemoattractant is chemically and 13. Singer, A. G., Clancy, A. N., Macrides, F. & Agosta, W. C.
structurally related to earthworm cuticle collagen/gelatin ob- (1984) Physiol. Behav. 33, 645-652.
14. Clancy, A. N., Macrides, F., Singer, A. G. & Agosta, W. C.
tained by solubilizing earthworm cuticle at 60°C. (1984) Physiol. Behav. 33, 653-660.
The earthworm produces a cuticle collagen/gelatin that 15. Reformato, L. S., Kirschenbaum, D. M. & Halpern, M. (1983)
Pharmacol. Biochem. Behav. 18, 247-254.
contains an agent detectable by the vomeronasal apparatus of 16. Halpern, M., Schulman, N., Scribani, L. & Kirschenbaum,
the garter snake. This agent is an attractant that enables D. M. (1984) Pharmacol. Biochem. Behav. 21, 655-662.
snakes to locate their prey. 17. Kirschenbaum, D. M., Schulman, N., Yao, P. T. & Halpern,
M. (1985) Comp. Biochem. Physiol. B 82, 447-453.
We thank Drs. L. Kesner, R. Carty, and J. Kang and Ms. Patti T. 18. Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A. &
Yao for their assistance; Mr. T. Housley and Dr. M. Tanzer Smith, F. (1956) Anal. Chem. 28, 355-358.
(University of Connecticut Health Science Center) for confirmatory 19. Woessner, J. F., Jr. (1961) Arch. Biochem. Biophys. 93, 440-447.
amino acid analyses; and Prof. J. R. Clamp (Bristol Royal Infirmary) 20. Murray, L. S., Waite, J. H., Tanzer, M. L. & Hauschka,
for the carbohydrate analyses. D.M.K. was a Faculty Exchange P. V. (1982) Methods Enzymol. 82, 65-95.
21. Watson, M. R. (1958) Biochem. J. 68, 416-419.
22. Goldstein, A. & Adams, E. J. (1970) J. Biol. Chem. 245,
5478-5483.
23. Miller, E. J. & Gay, S. T. (1982) Methods Enzymol. 82, 3-31.
24. Lee, Y. C. & Lang, D. J. (1968) J. Biol. Chem. 243, 677-680.
25. Muir, L. & Lee, Y. C. (1969) J. Biol. Chem. 244, 2343-2349.
26. Spiro, R. G. & Bhoyroo, V. D. (1980) J. Biol. Chem. 255,
5347-5354.
27. Maser, M. & Rice, R. V. (1963) Biochim. Biophys. Acta 74,
283-294.
28. Josse, J. & Harrington, W. F. (1964) J. Mol. Biol. 9, 269-287.